(ToniefT 'dllis7Cl^^it'l 
 
 [) 
 
 WitK part of tKe Irxcome 
 of tKe 
 
 NON-RESIDENT LECTURESHIP IN CHEMISTRY 
 
QD 412.n''iS56""'™"'*>"-"'"^ 
 
 Mili^ °'^^"''^ Chemistry of nitrogen, 
 
 3 1924 004 040 790 
 
Cornell University 
 Library 
 
 The original of this bool< is in 
 the Cornell University Library. 
 
 There are no known copyright restrictions in 
 the United States on the use of the text. 
 
 http://www.archive.org/details/cu31924004040790 
 
THE 
 
 ORGANIC CHEMISTRY OF NITROGEN 
 
HENRY FROWDE, M.A. 
 
 PUBLISHKR TO THE UNIVERSITY OF 0XK0I1I> 
 
 LONDON, EDINBURGH, NEW YORK 
 
 TORONTO AND MELBOURNE 
 
THE 
 
 ORGANIC CHEMISTRY 
 OF NITROGEN 
 
 BY 
 
 NEVIL VINCENT SIDGWICK 
 
 M.A. (OxoN.). ScD. (Tubingen) 
 
 FELLOW OF LINCOLN COLLEGE, OXFORD 
 
 OXFORD 
 AT THE CLARENDON PRESS 
 
 1910 
 
7 
 
 OXFOED 
 
 PEINTED AT THE CLARENDON PRESS 
 
 BY HORACE HART, M.A. 
 
 PRINTER TO THE UNIVERSITY 
 
PREFACE 
 
 This book originated in lectures given to more advanced 
 students in Oxford. Its object is thus primarily educational, 
 and it is in no sense intended as a work of reference. I have 
 endeavoured to give an account of the present state of knowledge 
 on those parts of the subject which are of the greatest theoretical 
 interest, and at the same time to avoid overloading the text with 
 the names of less important substances. In dealing with the 
 vast group of heterocyclic compounds, I have thought it better 
 not to attempt even an enumeration of all the known types of 
 ring ; but I have selected a few of the more important, and 
 discussed these in detail. I have assumed throughout that the 
 reader has at least an elementary knowledge of organic and of 
 general chemistry. 
 
 It is becoming generally recognized that organic chemistry 
 cannot be treated satisfactorily without reference to those 
 questions of physical chemistry which it involves. To attempt 
 a separation of the two is to refuse all the assistance which can 
 be derived from what is really the quantitative side of chemistry. 
 The various physical questions are therefore discussed as they 
 arise. A full treatment of the phenomena of tautomerism would 
 have required too great an interruption of the main current of 
 thought ; but I have tried to indicate the more important points 
 in which they are illustrated by the bodies under consideration. 
 The dynamics of organic reactions is a field which, in spite of the 
 increasing amount of attention recently devoted to it, is still very 
 largely unexplored ; and yet it is of the utmost value for eluci- 
 dating the mechanism of chemical change. I have therefore 
 made the references to investigations of the velocity of reaction 
 
vi Preface 
 
 as complete as I could, and the methods of analysis adopted in 
 
 each case have been described. 
 
 I have to acknowledge my obligations to several of the 
 established textbooks of organic chemistry, and above all to the 
 great Lehrbuch of Meyer and Jacobson. I have also made great 
 use of Richter's Lehrbuch, and in one or two places I have quoted 
 his admirably concise summaries of the relations of a compHcated 
 group of substances. The various monographs dealing with 
 special branches of the subject which I have consulted are 
 referred to in their places. 
 
 To my colleagues in Oxford I am indebted for help on various 
 points, more especially dealing with physical questions, and in 
 particular to Mr. D. H. Nagel, for his assistance and advice as the 
 book was passing through the press. I wish to express my 
 heartiest thanks to Mr. H. T. Tizard, who has read the whole 
 work in manuscript, and whose constant suggestions and criticisms 
 have been of the greatest value to me. 
 
CONTENTS 
 
 DIVISION I 
 
 COMPOUNDS WITH NO NITROGEN DIRECTLY ATTACHED 
 
 TO CARBON 
 
 PAGE 
 
 Chapter I. Esters of nitrous and nitric acids 5 
 
 Esters of nitrous acid (5). Baeyer and Villiger's theory of their re- 
 actions (6). 
 
 Esters of nitric acid (8). Nitroglycerine (9). Nitrocellulose : nitro- 
 explosives (10). Acyl nitrates (12). 
 
 DIVISION II 
 
 BODIES CONTAINING ONE NITROGEN ATOM ATTACHED 
 
 TO CARBON 
 
 Chapter II. Amines 15 
 
 Methods of formation (15). Properties (18). Separation (19). Chemical 
 properties (20). 
 
 Quaternary ammonium compounds (22). Structure (23). Basicity (24). 
 Stereochemistry of pentavalent nitrogen (27). Substituted alkylamines 
 (31). Amino-alcohols (32). Amino-ketones (34). 
 
 Amino-acids (34). Preparation (35). Properties (36). Polypeptides : 
 structure of the proteins (38). 
 
 Benzylamine bases (43). 
 
 Chapter III. Aromatic amines 45 
 
 Preparation (45). Properties (46). 
 
 Substituted anilines. Mixed amines (48). Purely aromatic amines (50). 
 Substitution : migration of substituents (51). Halogen derivatives (54). 
 Sulphonic acids (55). Nitranilines (55). Aminophenols (56). Homologues 
 of aniline (57). 
 
 Benzidine derivatives (58). 
 
 Amino-di- and tri-phenyl-methanes (59). Constitution of the rosaniline 
 dyes (61). 
 
 Diamines. Alkylene diamines (68). Aromatic diamines (70). Quinone 
 imines and diimines (72). 
 
 Chapter IV. Amides 76 
 
 Formation (76). Properties (79). Hofmann reaction (82). Amides of 
 dibasic acids (84). Anilides (86). Thioamides (88). Imides (90). 
 Amido-chlorides, imido-chlorides, imido-ethers (92). 
 
viii Contents 
 
 PAGE 
 
 Chapter V. Derivatives of hydroxylamine ^® 
 
 a- and ^-hydroxylamines (96). Keactions of ^-aryl-hydroxylamines (98). 
 
 Amine-oxides or oxy-amines (101). Hydroxamic acids (103). Amidoximes 
 
 (105). 
 Oximes (106). Stereoisomerism of the oximes (110). Stereo-hindrance 
 
 (117). Oximes of benzil (118). Beckmann reaction (119). 
 
 Chapter VI. Nitroso-oompounds 122 
 
 Fatty nitroso-compounds. Formation (122). Secondary derivatives 
 (126). Stability (126). Nitrosites, pseudo-nitrosites, nitrosates (128). 
 
 Aromatic nitroso-compounds. Formation (131). Properties (182). 
 Nitrosolic acids (183). 
 
 Quinone oximes or nitrosophenols (185). 
 
 Chapter VII. Nitro-compounds ........ 139 
 
 Nitroparaffins. Formation (139). Properties (141). Structure (142). 
 Phenyl nitromethane (143). a-Nitro-ketones (145). Nitrolic acids (146). 
 Pseudonitrols (147). Poly-nitroparaffins (148). Nitroform and its salts 
 (150). 
 
 Aromatic nitro-compounds. Formation (153). Keduction (155). Oxi- 
 dation of amines (162). Properties of nitro-compounds (164). Nitronic 
 acids (165). Nitrophenols (169). Colour and constitution (171). Picric 
 acid (177). Nitro-diphenylamines (178). 
 
 Chapter VIII. Carbonic acid derivatives 179 
 
 Carbamic acid (179). Carbamic chloride (180). Urethanes (181). 
 Thiocarbamic acid (182). Imino-dicarboxylic acid (183). 
 
 Urea. Formation (184). Properties (185). Alkyl ureas (187). Thio- 
 ureas (189). Carbodiimide derivatives (192). Guanidine (198). 
 
 Chapter IX. Cyanogen compounds 195 
 
 Cyanogen (195). Prussia acid (197). Benzoin synthesis (200). Salts 
 of prussic acid (201). 
 
 Nitriles (208). Formation (203). Properties (204). Polymerization 
 (205). Iso-cyanides (207). Constitution of prussic acid and its salts (209). 
 
 Cyanic acid (214). Cyanogen halides (215). Cyanamide (216). Cyanic 
 esters (217). Phenyl isocyanate (219). Thiocyanic acid (220). 
 
 Fulminic acid (223). Polymers (228). Nitrile-oxides (229). 
 
 Tricyanogen compounds (231). Cyanuric acid (233). Cyamelide (234). 
 Thiocyanuric compounds (235). Melamines (236). Summary of poly- 
 merization products (236). 
 
Contents ix 
 
 DIVISION III 
 
 COMPOUNDS CONTAINING AN OPEN CHAIN OF TWO OR MORE 
 
 NITROGEN ATOMS 
 
 PAGE 
 
 Chapter X. Hydrazine derivatives 241 
 
 Alkyl-hydrazines (241). Aromatic hydrazines (242). Formation (242). 
 Properties (243). Tetra-aryl-hydrazines (245). Acid hydrazides (246). 
 Hydrazones (247). Osazones (249). Formation of hydrazones from diazo- 
 eompounds (250). Hydrazo-compounds (254). 
 
 Chapter XI. Diazo-compounds . . 256 
 
 Formation (257). Properties (259). Diazo-reactions (260). Constitu- 
 tion (262). Diazonium (265). Diazotates (267). Syn- and anti-diazo- 
 compounds (268). Diazo-hydrates and nitrosamines (270). Solid 
 diazonium salts (272). Diazo-cyanides (273). Theory of the diazo- 
 reactions (275). Summary (278). 
 
 Chapter XII. Azo-compounds, azoxy-compounds, nitramines . 280 
 
 Azo-compounds. Formation (280). Properties (281). Azomethane 
 
 (281). Methods of dyeing (282). Aminoazo-compounds (284). Properties 
 
 (285). Oxy-azo-compounds (287). Constitution (288). 
 Azoxy-compounds (292). Liquid crystals (298). 
 Nitrosamine8-^293}r Nitramines (295). Nitrimines (297). Isonitra- 
 
 mines {^T)^ 
 
 Chapter XIII. Derivatives of carbonic acid 289 
 
 Semicarbazide (299). Thiosemicarbazide (300). Azo-dicarboxylic acid 
 (801). Nitroso-derivatives (802). Nitro-derivatives (803). 
 
 Chapter XIV. Compounds containing a chain of three or more 
 
 nitrogen atoms 305 
 
 Diazoamino-compounds (805). Fatty diazoamino-compounds (306). 
 
 Reactions (308). Hydrazoamino-compounds (311). 
 
 Buzane or tetrazane derivatives (313). Diazo-hydrazides (313). Bis 
 
 diazoamino-compounds (314). 
 
 Chapter XV. Uric acid derivatives 315 
 
 Ureides (815). Barbituric acid derivatives (816). Uric acid (317). 
 Constitution (318). Syntheses (320). Purine (322). Syntheses of purine 
 derivatives (323). 
 
X Contents 
 
 DIVISION IV 
 KING COMPOUNDS 
 
 FAOE 
 
 Chapter XVI. 3-rings and 4-rings 331 
 
 I. CaN. Ethylene imine (331). 
 
 II. CNj. Diazomethane group. Diazomethane (332). Potassium 
 methane-diazotate (832). Eeactions of diazomethane (333). Diazoacetic 
 ester. Preparation (335). Properties (386). Pseudo-phenyl-acetic acid 
 (339). Polymerizations of diazoacetic ester (341). Diazomethane di- 
 sulphonic acid (344). 
 
 III. N3. Azides. Preparation (345). Properties (346). Eeduction : 
 phenyl-triazene (347). Syntheses with phenyl azide (348). 
 
 IV. 4-rings. Trimethylene imine (350). Stability of polymethylene 
 imines (350). 
 
 Chapter XVII. 5-rings 353 
 
 Pyrrol group. Syntheses (353). Orientation (855). Properties (357). 
 Pyrrol character (357). Derivatives (360). Potassium pyrrol (360). 
 Alkyl pyrrols (360). Halogen derivatives (361). Nitroso- and nitro- 
 pyrrols (361). Amino- and azo-pyrrols (362). Carboxylic acids (362). 
 Keduced pyrrol derivatives (363). 
 
 Indol group. History (365). Indigo derivatives (365). Constitution 
 of indigo (366). Indol (869). Alkyl-indols (369). Indoxyl (870). Oxindol, 
 dioxindo], isatin (371). Indigo (372). Syntheses of indigo (374). Com- 
 mercial syntheses : Heumann's method (375). Subsequent developments 
 (379). Sandmeyer's synthesis (380). 
 
 Chapter XVIII. 6-rings 383 
 
 I. Pyridine. Constitution (383). Syntheses (384). Properties (386). 
 Substitution (387). Pyridinium compounds (388). Orientation (389). 
 Acid chlorides (390). Amino-pyridines (390). Piperidine (391). 
 
 Pyridine alkaloids. Coniine (392). Piperine (396). Nicotine (397). 
 
 II. Quinoline. Syntheses (400). Skraup reaction (402). Properties 
 of quinoline (408). Derivatives (403). 
 
 III. Isoquinoline (405). 
 
ABBREVIATIONS 
 
 Am, Ch. J. American Chemical Journal. 
 
 Ann. Liebig's Annalen der Chemie. 
 
 Ann. Chim. Phys. Annales de Chimie et Physique. 
 
 Arch. d. Pharm. Archiv der Pharmacie. 
 
 Atti. B. Atti della Eeale Accademia dei Lincei. 
 
 Ber. Berichte der deutschen chemischen Gesellschaft. 
 
 Brit. Ass. Beports. Eeports of the British Association. 
 
 Bull. Soc. Bulletin de la Societe Chimique de France. 
 
 C. Chemisches Centralblatt. 
 
 Gk. Neivs. Chemical News. 
 
 G. B. Comptes Eendus des Seances de I'Academie des Sciences. 
 
 Gass. Gazzetta Chimica Italiana. 
 
 J. Am. GJi. Soc. Journal of the American Chemical Society. 
 
 •7. G. S. Journal of the Chemical Society. 
 
 ■/. Ghim. Phys. Journal de Chimie physique. 
 
 J. pr. Gh. Journal ftlr praktische Chemie. 
 
 J. Soc. Gliem. Ind. Journal of the Society of Chemical Industry. 
 
 Mon. Monatshefte fur Chemie. 
 
 Phil. Mag. Philosophical Magazine. 
 
 Pogg. Ann. Poggendorff's Annalen der Physik und Chemie. 
 
 Proc. Garni. Phil. Soc. Proceedings of the Cambridge Philosophical Society. 
 
 Proc. G. S. Proceedings of the Chemical Society. 
 
 Bee. Trav. Eecueil des travaux chimiques des Pays-Bas. 
 
 Z. f. angew. Ghem. Zeitschrift filr angewandte Chemie. 
 
 Z. f. EleTdroehem. Zeitschrift fiir Elektrochemie. 
 
 Zeit. Farb. u. Text. Zeitschrift fur Farb- und Textilindustrie. 
 
 Z. Ph. Gh. Zeitschrift fiir Physikalische Chemie. 
 
 Alk = Alkyl,C,,H2„ + i. 
 
 Ar = Aryl, aromatic radical, as CH3'CgH4-. 
 
 Ac = Acyl, acid radical, R'CO-. 
 
 <f> = Phenyl, CeH,--. 
 
 M ■ = Atom of monovalent metal, as K or Na. 
 
 Me = Methyl, CH3-. 
 
 Et = Ethyl, C2H5-. 
 
 CX = Concentration of X (gram-molecules per litre). 
 
THE ORGANIC CHEMISTRY OF NITROGEN 
 
 The organic nitrogen derivatives Ml naturally into four divisions : — 
 
 1. Those in which the carbon is not attached to nitrogen directly, but 
 indirectly, through oxygen. This group is practically confined to the esters 
 of nitrous and nitric acids. (The a-hydroxylamines, E-O-NHj, belong 
 strictly speaking to this class, but it is more convenient to discuss them 
 along with the other hydroxylamine derivatives.) 
 
 2. Bodies containing one or more nitrogen atoms attached to carbon, 
 but not to one another, and not forming part of a closed ring. These form 
 the most important and fundamental class. 
 
 3. Bodies containing two or more nitrogen atoms attached to one another 
 in an open chain. 
 
 4. Compoimds with closed rings containing one or more nitrogen atoms. 
 Of this enormous class only a selection of the more important types will be 
 discussed. 
 
DIVISION I 
 
 COMPOUNDS WITH NO NITROGEN DIRECTLY 
 ATTACHED TO CARBON 
 
CHAPTER I 
 ESTERS OF NITROUS AND NITRIC ACIDS 
 
 ESTERS OF NITROUS ACID 
 
 These bodies are formed — 
 
 1. By the action of nitrous fumes (a mixture of nitrogen peroxide, trioxide, 
 and nitric oxide) on the alcohols. A modification of this method' is to pass 
 the vapour of nitrosyl chloride NOCl into a mixture of the required alcohol 
 and pyridine in molecular proportions. The action of the pyridine is merely 
 to remove hydrochloric acid. 
 
 0=NC1 + HOR = HCl + ONOR. 
 
 2. By the action of sodium nitrite on a mixture of the alcohol and 
 sulphuric acid. 
 
 3. As by-products in the preparation of the nitro-para£Qns by the action 
 of silver nitrite on the alkyl iodides. 
 
 This curious instance of a tautomeric reaction will be discussed later, 
 in dealing with the nitro-paraffins. At present it may be regarded as a simple 
 double decomposition : — 
 
 Ag(N02) + C2H5I = Agl + C2H5(NOjJ. 
 
 The nitrous esters are volatile pleasant smelling liquids, which boil at 
 a much lower temperature than the corresponding alcohols. Thus methyl 
 nitrite is a gas, boiling at - 17° ; ethyl nitrite boUs at + 17°, and normal 
 propyl nitrite at + 57°. On reduction they are split up, with the formation 
 of an alcohol and either hydroxylamine or ammonia. This shows that the 
 nitrogen is not attached to carbon directly, but indirectly, through oxygen, 
 as CH3-0-N=0. If it were attached to the carbon it would remain there 
 on reduction, as it does in the nitro-compounds. 
 
 The nitrous esters undergo a singular reaction '^ when treated with excess 
 of zinc or magnesium alkyl halide. An addition-product is formed, which 
 breaks up on treatment with water to give the alcohol of the ester and 
 a /3-dialkyl hydroxylamine : the -N=0 group of the nitrous acid being con- 
 verted into -N=(Alk)2. 
 
 /OZnl 
 ' E-0-N=0 -I- 2 ZnAJkl -♦ R-0-N=(Alk)2 -> R-O-H + H0-N(Alk)2. 
 
 \ZnI 
 Among the esters in general those of nitrous acid occupy quite a unique 
 position in respect of the ease with which they are both formed and decom- 
 
 > liouveanlt, Wahl, C. 03. ii. 338. " Bewad, Ber. 40. 8065 (1907). 
 
 I 1175 B 
 
6 Nitrous Esters 
 
 posed. The formation of an ester from an alcohol and an acid is usually 
 a comparatively slow reaction, which may take several hours or even days. 
 But if a solution of benzyl alcohol (or amyl alcohol) in water is acidified, 
 and then treated with sodium nitrite solution, it instantly becomes milky 
 through separation of the nitrous ester. These esters can also be saponified 
 by strong acids with remarkable rapidity. Indeed the ease with which they 
 break up is shown by the fact that they can be used" instead of free nitrous 
 acid, for example in diazotizing ; in the same way ethyl nitrite acts on hydrazme 
 in alkaline solution to form hydrazoic acid NgH.' 
 
 To explain these and certain other reactions of the nitrous ester Baeyer 
 and Villiger^ have proposed the following theory. They assume that nitrous 
 acid readily forms unstable addition-compounds in which the nitrogen is 
 pentad. Thus it adds on water to form a compound analogous to phosphorous 
 acid : — 
 
 /OH 
 
 0=n-oh + hoh = 0=n-oh. 
 
 \h 
 
 The hydroxyls of this acid can be replaced by alkoxyl or by the anions of 
 acids (including the peroxides), giving rise to such compounds as 
 
 /OSO2OH 
 
 /OC,H, 
 
 /OS020H 
 
 0=N-:6h: 
 
 0=I(-;OHi 
 
 Nh i 
 
 Nh J 
 
 o=N-;oc,H, 
 
 Nh 
 
 These unstable compounds then break up again according to the following 
 rule : if both hydroxyl and alkoxyl are present, the hydroxyl goes out with 
 the hydrogen as water, but if the anion of an acid is present, this remains 
 while the hydroxyl or alkoxyl splits off, as shown by the dotted lines. 
 
 This hypothesis explains many remarkable reactions of the nitrites. The 
 rapidity of formation of the esters is due to the occurrence of an addition- 
 reaction, which can take place instantaneously (like most addition-reactions 
 in which trivalent nitrogen becomes pentavalent), followed by the loss of 
 water : — 
 
 /OC2H5 
 0=N-OH -f HOC2H5 = 0=N-OH = O^N-OCaHg + H^O. 
 
 On the other hand the saponification by acids goes thus : — 
 
 /OC2H5 
 O^N-OCaHj -f HO-SO^OH = 0=N-OS020H 
 
 \h 
 
 = C^H^OH + O^N-OSO^OH -^ ^=^^0^0^' 
 
 In accordance with the rule H and OC2H5 go out to form alcohol, while 
 a tautomeric form of nitrosulphonic acid remains. This is broken up by 
 water into sulphuric and nitrous acids, so that if a substance is present which 
 can react with nitrous acid it will do so ; hence the use of nitrous esters in 
 the preparation of diazo-compounds. 
 
 ' Thiele, Ber. 41. 2681 (1908) ; Stolid, ib. 2811. > Ber. 34. 765 (1901). 
 
Nitrous Esters: Theory of Reactions 7 
 
 The hypothesis also explains an interesting series of reactions of the nitrites 
 with hydrogen peroxide and its organic derivatives. It is found that while 
 (1) ethyl nitrite and hydrogen peroxide give alcohol and nitric acid, (2) ethyl 
 hydroperoxide, CaHgO-OH, and nitrous acid give ethyl nitrate and water. 
 Now if the hydrogen peroxide broke up, in adding on to the nitrous acid, into 
 two hydroxyls, one would get in both reactions the same addition-product : — 
 
 (1) 0=N-OC,H, + (OH), ^Qjj 
 
 QC2H5 ^0=N-OH 
 
 (2) 0=N-OH + V y \oc,H, 
 
 and hence the same ultimate product. The fact that the products are different 
 shows that the peroxide breaks up (like the sulphuric acid in the previous 
 case) into H + 0-OH, which gives : — 
 
 /OC2H5 
 
 (1) 0=n-oc2h5 + h-ooh = 0=n-0-oh = 0=n-ooh + c2h5oh. 
 
 \h 
 
 This again breaks up, in accordance with the rule that OOH like the anion 
 of an acid remains attached to the nitrogen, into alcohol and a peracid 
 0=N-0-OH, which then changes into nitric acid : the ultimate products 
 from the nitrite and hydrogen peroxide being nitric acid and alcohol, as 
 experiment shows they are. On the other hand nitrous acid and ethyl 
 
 hydroperoxide react thus: — 
 
 /OOC2H5 
 (2) 0=N-OH + H-OOCoHg = 0=N-OH = 0=N-OOC2H5 + HgO 
 
 \h 
 
 giving the ester of the peracid, which changes into ethyl nitrate : and this 
 again is in accordance with experiment. This view is strikingly confirmed 
 by the action of ethyl hydroperoxide on amyl nitrite. If the peroxide acted 
 only as an oxidizing agent it would of course produce amyl nitrate ; whereas 
 if it split up into H + O-OCgHg it should give ethyl nitrate thus : — 
 
 /OCgHu 
 
 0=N-0C5Hii + HOOC2H5 = 0=N-H 
 
 -> O^N-O-OC^H, -* 0=N<g^^jj^. 
 
 As a fact only ethyl nitrate was obtained. 
 
 This theory is not definitely established, but it certainly appears to offer 
 a plausible explanation of the very peculiar behaviour of nitrous acid and 
 its esters. 
 
 The rate of formation and of hydrolysis of the alkyl nitrites has been 
 measured by W. M. Kscher,' the amount of ester being determined by 
 titration. He finds that these reactions though very rapid are not instan- 
 taneous. Ethyl nitrite is hydrolysed by water, but not completely so ; thus 
 an N/43 solution of the ester is 84 per cent, hydrolysed. In presence of 
 hydrochloric acid the hydrolysis is practically complete, but the first titration 
 was foimd to be about 6 per cent, higher than that obtained after five minutes, 
 which was also the final value. This indicates a very high velocity, which 
 
 » Z. PA. Ch. 65. 61 (1908). 
 
 b2 
 
8 Nitrous Esters 
 
 may be compared with that of the hydrolysis of triphenyl-methyl chloride 
 to the carbinol. The reactions seem to be of the same order as in other 
 cases, since the usual expression 
 
 ^alcohol "^ ^acid 
 
 p p = const. 
 
 '^water ^ *^ester 
 
 was fomid to hold good. 
 
 On the other hand soda and sodium ethylate act on the ester compara- 
 tively slowly. The velocity constant obtained with aqueous soda is only about 
 three times as great as that for ethyl acetate ; while with sodium ethylate 
 in absolute alcohol the velocity is only about one-fortieth of this, and is 
 actually only a tenth of that observed with the ethyl ester of ethyl sul- 
 phonic acid. 
 
 The esters of nitrous acid undergo a remarkable condensation, in presence 
 of sodium or potassium ethylate, with compounds containing an acidic methy- 
 lene group, such as phenyl-acetic ester, giving the salt of an oxime ' : — 
 
 CjHs-ONO + ^.CHa-COaCaHs + KOC2H5 = 2 C2H5OH + ^-C<^o c H ' 
 
 Baeyer and Villiger^ have also drawn attention to a singular analogy 
 between nitrous acid and triphenyl-carbinol (C5H5)3COH. They both form 
 esters with alcohols with unusual ease. They both combine with sulphurous 
 acid to give sulphonic acids, NO-SOgH, ^gC-SOgH. The carbinol with aniline 
 forms an anilide ^gC-NH^, as nitrous acid with secondary amines forms 
 a nitrosamine ON-NEj. With phenols it condenses to oxy-tetraphenyl- 
 methane ^gCCgH^OH, as nitrous acid does to a nitrosophenol, ONCgH^OH. 
 Triphenyl-methyl chloride, like nitrosyl chloride, gives double salts with 
 metallic chlorides. Finally, corresponding to the free monovalent radicle NO 
 there is (probably) the free monovalent radicle ^gC, Gomberg's triphenyl- 
 methyl. 
 
 ESTERS OF NITRIC ACID 
 
 The esters of nitric acid, like those of other mineral acids, can be prepared 
 by the direct action of the acid on the alcohol; but in this case special 
 precautions are required. The nitric acid generally contains nitrous acid, 
 and in any case this is likely to be formed by the reduction of some of the 
 nitric acid by the alcohol; and in the presence of nitrous acid the ester is 
 violently oxidized, which diminishes the yield and may easily lead to violent 
 explosions. To avoid this danger the acid may be previously purified by 
 distillation in vacuo with concentrated sulphuric acid, and then the alcohol 
 dropped into it, the temperature not being allowed to rise above 5° ; ' or, as in 
 the preparation of the aromatic nitro-compounds, the alcohol may be dropped 
 into a mixture of nitric and sulphuric acids kept at 0°. 
 
 A more usual method is to add some substance which vsdll destroy the 
 nitrous acid as fast as it is formed. The substance employed is urea, which 
 
 » W. Wislicenus, GrUtzner, Ber. 42. 1930 (1909). a Ber. 35. 3019 (1902). 
 
 * Bouveault, Wahl, G. 03. ii. 338. 
 
Nitric Esters 9 
 
 has the advantage of giving with nitrous acid only nitrogen, carbon dioxide, 
 and water : 
 
 0=C(NH2)2 + 2 0=NOH = COg + 2 N^ + 3 H^O. 
 
 The nitric acid is first boiled with urea, and then more urea is added and the 
 alcohol dropped in. 
 
 The nitric esters of the lower alcohols are colourless, pleasant smelling 
 liqiiids, whose boiling-points are a good deal higher than those of the corre- 
 sponding nitrites, and not far removed from those of the alcohols themselves : — 
 
 Alcohol Nitrate Nitrite 
 
 methyl 
 
 -f65° 
 
 4-66° 
 
 -12° 
 
 ethyl 
 
 78° 
 
 87° 
 
 -t-17° 
 
 n-propyl 
 
 97° 
 
 110-5° 
 
 57° 
 
 i-propyl 
 
 81° 
 
 101-5° 
 
 89° 
 
 n-butyl 
 
 117° 
 
 136° 
 
 75° 
 
 When set on fire they burn with a white fiame, and when heated above their 
 boUing-points they explode violently. On reduction with tin and hydrochloric 
 acid they give the alcohol and hydroxylamine, showing that the nitrogen is 
 
 attached to the carbon through oxygen : CHg-O-N^^ : this is also proved by 
 
 the ease with which they are saponified by acids to alcohol and nitric acid, 
 though these primary products react with one another to some extent, forming 
 oxidation products of alcohol, such as aldehyde, and nitrous acid. The same 
 thing happens on alkaline saponification ' ; and in this case it is probable 
 that a certain amount of ethyl hydroperoxide C2H50-OH is formed. They 
 condense (like the nitrites) in presence of sodium or potassium ethylate, 
 with bodies containing an acidic methylene group, to form isonitro-com- 
 pounds '' : — 
 
 EtO-NOa + NaOEt + ^-CHa-COaEt = 2 EtOH + <^-C<(^q "^^ . 
 
 Among the nitric esters some are of great practical importance— certain 
 polynitrates of the polyatomic alcohols. The simplest of these is glyceryl 
 
 trinitrate 
 
 ? 
 
 !H,-0-NO, 
 
 CH-0-NO„ 
 
 I ^ 
 
 CH2-O-NO2 
 
 commonly, though wrongly, called nitroglycerine. It is formed by the action 
 of a mixture of nitric and sulphuric acids on glycerine. It is an oily liquid, 
 without colour or smell, which solidifies at +8°. It is scarcely soluble in 
 water. On saponification " the NO3 groups are readUy split off as nitric acid, 
 but they oxidize the glycerine, and are themselves reduced to nitrous acid. If 
 the ester is set. on fire it burns away rapidly but quietly : but under certain 
 definite conditions it can be made to explode with great violence. 
 
 1 Gutraann, Ber. 41. 2052 (1908) ; Tor Carlson, C. 08. i. 938 ; Cambi, C. 09. u. 693. 
 
 2 W. Wislicenus, Endrea, Ser. 35. 1755 (1902) ; Wialicenua, Grtttzner, Ber. 42. 1930 (1909). 
 » See Kobertson, Ch. News, 99. 289 (1909). 
 
10 Nitric Esters 
 
 Another class of nitric esters used for explosives consists of the mtro- 
 celluloses. Cellulose is a carbohydrate which forms the chief constituent of 
 the walls of plant cells. Its molecular weight is unknown, but is very 
 high, and it is quite possible that under this name are included several 
 allied but distinct substances. Common forms of it are cotton-wool and 
 Swedish filter-paper. Cellulose contains hydroxyl gi-oups, and therefore 
 when treated with strong nitric acid or a mixture of nitric and sulphuric 
 acids it is converted into a series of nitric esters, whose constitution is 
 proved by the fact that they give nitric acid when treated with alkalies, and 
 that under the action of reducing agents they split off their nitrogen and 
 re-generate cellulose.' The nitration may occur in several stages,' and in order 
 to designate the products the molecular formula of cellulose is assumed to 
 be C13H20O10 (it is of course really a high polymer of this) ; and to this 
 their formulae are referred. Thus we have cellulose dinitrate, Ci2HigOg(N03)2, 
 tetranitrate, Ci2Hi60e(N03)4, and hexanitrate, Ci2HiA(N03)6. Higher nitration- 
 products than the hexanitrate cannot be obtained. These various stages can 
 be separated by means of their solubility in certain solvents, such as alcohol, 
 ether, and ethyl acetate ; but it is probable that the products so obtained 
 are themselves mixtures and not chemical individuals. They all possess the 
 property, which is characteristic of the organic polynitrates in general, of being 
 highly explosive. 
 
 The less nitrated compounds, the di- and tetranitrates, which are got by 
 treating cotton-wool with a mixture of nitric and sulphuric acids of a definite 
 strength, are soluble in a mixture of alcohol and ether: and this solution is 
 known as coUodion. If the solvent is allowed to evaporate the nitrocellulose 
 remains behind as a firm and continuous film. It is on this property that 
 its use in surgery and in the older wet-plate process of photography depends. 
 A solution of this feebly nitrated cellulose in molten camphor constitutes 
 the celluloid which is used nowadays for so many purposes. It has the 
 advantage that it can be worked into almost any form, but its use is somewhat 
 dangerous, since, as we should expect from its composition, it catches fire very 
 readily and burns with great violence. 
 
 The more highly nitrated celluloses, the tetra- to the hexanitrate, form 
 the well-known explosive gun-cotton. With comparatively few and unimportant 
 exceptions all modern explosives are made out of nitrocellulose and nitro- 
 glycerine. The history of the development of this branch of applied chemistry 
 has been described by Will," from whose paper the following facts are 
 taken. 
 
 In 1846 Schonbein discovered that cotton, when treated with a mixture 
 of nitric and sulphuric acids, was converted without any change in its external 
 appearance into a powerful explosive, which had many advantages over 
 gunpowder. It was more stable, it burnt without smoke, and it had a much 
 
 '■ Some so-called nitrocelluloaes are, however, possibly nitrates of oxy-celluloae, an oxidation- 
 product of cellulose. Vignon, C. 03. i. 1081. 
 
 ^ Cf. Hake, Bele, Joarn. Soe. Chem. Ind. 28. 457 (1909). 
 s Ber. 37. 268. 
 
Nitro-eooplosives 11 
 
 greater explosive power. The first difficulty met with in its preparation was 
 the complete removal of the nitrating acids : for if these are not removed, the 
 gun-cotton rapidly decomposes, not unfrequently with violent explosions. 
 This was got over by Abel's method (1865) of grinding up the nitrocellulose 
 with water and then compressing it. 
 
 The next difficulty was firing the gun-cotton. The ordinary method of 
 firing by a train of gunpowder through the touch-hole frequently failed to do 
 more than make it burn away rapidly. Nobel overcame this in 1864 by using 
 a small initial charge of black powder, fired by a silver fulminate detonator. 
 He soon found that the detonator alone was sufficient, without the initial 
 charge. The same method could also be used to fire nitroglycerine, which 
 had been discovered about the same time as gun-cotton (Sobrero 1848), but for 
 nearly twenty years afterwards had not been employed as an explosive, owing 
 to the peculiar difficulties which it offered, from the fact of its being a liquid. 
 In 1865, however, Nobel showed that 'Baeselgur', a fine porous infusorial 
 earth, would absorb three times its weight of nitroglycerine to form a substance 
 of the constituency of putty, which was easily packed into cartridges, and 
 was much less sensitive to blows than nitroglycerine itself. This he called 
 dynamite. It at once came into use on an enormous scale for blasting 
 purposes, as is shown by the following figures, which give the total production 
 of dynamite in the world : 
 
 1867 .... 11 tons 
 
 1874 .... 3,000 tons 
 1899 .... 62,150 tons. 
 
 The Kieselgur which is used in dynamite to give a solid consistency to 
 the nitroglycerine is of course a perfectly inert substance and diminishes 
 correspondingly the energy of the explosive. Hence Nobel endeavoured to 
 replace it by some solid explosive. In 1878 he found that if collodion, 
 a solution of the less nitrated cellulose in ether and alcohol, is mixed with 
 nitroglycerine and the solvent allowed to evaporate, an indiarubber-like mass 
 of great explosive energy is formed. This constituted his 'gelatinized dynamite', 
 which soon replaced ordinary dynamite to a very lai-ge extent. Shortly after- 
 wards he found that nitroglycerine would 'gelatinize' with nitrocellulose 
 without the employment of any solvent at all. The cellulose does not dissolve, 
 but it absorbs the nitroglycerine and swells up in the same sort of way that 
 gelatine does when treated with cold water. Ballistite and cordite are explosives 
 prepared in this way. 
 
 The use of these explosives for guns was at first impossible from their very 
 high rate of combustion. In order to secure the maximum velocity for a given 
 strength of barrel it is necessary that the pressure of the gases should not be 
 put on too suddenly, but should continue to rise as long as it can act on the 
 projectile ; in other words the charge must go on burning as long as the shot 
 is in the gun. With the increase in the size of ordnance difficulties arose in 
 this way even with the old black powder, and the rate of combustion of 
 nitrocellulose is far higher. It was therefore necessary to find some means 
 for making the rate of combustion slower, and, moreover, for regulating it to 
 
12 Nitric Esters 
 
 the particular value required for each gun. Now it has been shown that 
 an explosion started in a piece of explosive gelatine proceeds in concentric 
 spheres from the point of origin, and hence it is obvious that if the substance 
 is cut up into thin strips the rate of decomposition wiU be diminished. In 
 this way the difficulties were finally overcome, and it was found possible 
 to use the nitro-powders for cannon, as well as for mining purposes. It is to be 
 noticed that the gelatinized powders are peculiarly suitable for modification 
 in this way : not only because their consistency makes it easy to give them 
 any required form, but also, which is even more important, because this form is 
 retained during the explosion ; whereas powders of a more or less crystalline 
 character, like black powder, are broken up by the pressure of the gases into _ 
 fine particles, and thus the effect of the form which is originally given to 
 them is largely destroyed. 
 
 Acyl Nitrates 
 
 A recently discovered class ' of nitric acid derivatives are the nitrates of the 
 acid radicles, such as acetyl nitrate, CHg-CO-ONOj, and benzoyl nitrate,- 
 CoHj-CO-ONOg. They are made by the action of nitrogen pentoxide on the 
 acid anhydride, or of silver nitrate on the acid chloride, in the cold. Acetyl 
 nitrate is a fuming colourless liquid, boiling at 22° under a pressure of 77 mm. 
 Benzoyl nitrate is a yellow oU. 
 
 They resemble the acid chlorides in being violently decomposed by water 
 into the organic acid and nitric acid. They explode on sudden heating, and 
 are very powerful nitrating agents. 
 
 Another mixed anhydride of acetic and nitric acids is the so-called diacetyl- 
 ortho-nitric acid,' (CH3-CO-0)2-N(OH)3, which is formed by the action of acetic 
 anhydride on nitric acid. It is a liquid boiling at 128°, and is also a powerful 
 nitrating agent. 
 
 ' Francis, /. C. S. 1906. 1; Ber. 39. 3798; Butler, Ber. 39. 3804 (1906); Pictet, Khotinaky, 
 C. B. 144. 210 ; Ber. 40. 1163 (1907). 
 
 = Piotet, C. 02. ii. 1438 ; 03. ii. 419, 1108; Pictet, Genequand, Ber. 35. 2526. 
 
DIVISION II 
 
 BODIES CONTAINING ONE NITROGEN ATOM 
 ATTACHED TO CARBON 
 
CHAPTEE II 
 
 AMINES 
 
 All these bodies may be regarded as derived from ammonia or hydro- 
 xylamine by the replacement of hydrogen by organic radicals : or from 
 nitrous or nitric acid by the replacement of the hydroxyl. The various 
 classes will be dealt with in order, beginning with the most highly reduced, 
 that is, the derivatives of ammonia, the amines and amides. 
 
 By replacing one, two, or three of the hydrogen atoms in ammonia by 
 hydrocarbon radicals we obtain the three classes of primary, secondary, and 
 tertiary amines. Moreover the nitrogen may become pentad, and so we may 
 have a fourth hydrocarbon group attached, giving rise to the quaternary 
 ammonium compounds. 
 
 The first amine to be discovered was aniline, prepared by Unverdorben' 
 in 1826, by the distillation of indigo ; its constitution was made out by 
 Hofmann''' in 1843. Of the fatty amines, the primary were discovered by 
 Wurtz ' in 1848, who prepared them by the hydrolysis of the isoeyanates : 
 and the three other classes by Hofmann,* by the action of the alkyl halides 
 on ammonia. 
 
 Methods of Formation 
 
 1. Hofmann's method, by heating the alkyl halides with alcoholic 
 ammonia : — 
 
 C2H5I + NH3 = CaHgNHa-HI 
 
 C2H5I + C2H5NH2 = (C2H5)2NH-HI 
 C2H5I + (C,H5)2NH = (C2H5)3N.HI 
 C2H5I + (C2H,)3N = (C^Hs),^! 
 
 This method cannot be used with aromatic halides on accoimt of the 
 firmness with which the halogen is attached : unless negative groups (such 
 as nitro-groups or other halogen atoms) are present, whereby this firmness 
 is diminished. 
 
 Both primary and secondary alkyl iodides can be used in this reaction ; 
 but tertiary cannot, as they give no amine, but lose hydriodic acid and form 
 alkylenes. 
 
 This reaction gives rise to a mixture of all the four classes of derivatives, 
 for which special methods of separation are required, which will be discussed 
 later. The proportions in which the various possible products are formed 
 
 1 Pogg. Am. 8. 397. ' Ann. 47. 37. ' Ann. 71. 330. 
 
 * Ann. 73. 91 (1850); 74. 159 (1850); 78. 253; 79. 16 (1851). 
 
16 Amines 
 
 constitutes a very complicated physico-chemical problem/ since they are 
 determined partly by the relative velocities of reaction of the intermediate 
 products, and partly by the solubilities of the substances in the solvent 
 employed. In some cases a suitable selection of conditions enables one to 
 obtain mainly the particular base required; but often (as in this case of the 
 ethyl amines) a mixture of all possible substances is got, which must after- 
 wards be separated by chemical means. 
 
 A modification of this method is to heat the alcohols to 250-260° with 
 the compound of zinc chloride and ammonia ^ : — 
 
 E-OH + NHg = ENH2 + H2O. 
 The aromatic amines can thus be obtained from the phenols at about 330 . 
 This reaction is used for preparing amines of the naphthalene series. 
 
 It is also possible to prepare methylamine by the direct alkylation of 
 ammonia in aqueous solution with methyl sulphate in presence of alkali.^ 
 
 2. There are a series of reactions which depend on starting with an 
 imido-compound whose imide hydrogen atom can be replaced by a metal. 
 If this metal is replaced by an alkyl group and the body so formed is 
 saponified, an amine is obtained. There are three important reactions of 
 this kind. 
 
 a. By boiling the isocyanic esters with potash. Isocyanic acid is the 
 imide of carbonic acid, as urea is its amide : — 
 
 ^0~NH or H-N=C=0. 
 
 Isocyanic acid 
 
 Silver (iso)cyanate acts on ethyl iodide to give ethyl isocyanate, and this 
 hydrolyses to ethylamine and carbon dioxide : — 
 
 C2H5-N=C=0 + H2O = CaHg-NHg + CO2. 
 
 It was by this reaction that Wurtz discovered the fatty amines in 1848. 
 
 6. Similarly the isocyanides when treated with alkali give the amine and 
 formic acid : — - 
 
 CgHs-N^C + H2O = CaHj-NHa + H-COOH. 
 
 c. A method much used in recent years for obtaining amines which are 
 otherwise difficult to prepare, is Gabriel's phthalimide reaction. Phthalimide 
 is prepared from phthalic anhydride and ammonia gas. If it is dissolved in 
 absolute alcohol and treated with alcoholic potash, potassium phthalimide is 
 precipitated. If this is treated with, say, ethylene dibromide in molecular 
 proportions, the metal is replaced by the group CHa-CHjBr: — 
 
 a CO, ^/CO, 
 
 ^^>N.K + Br-CH^-CH^Br = ^^^^>N-CH2CH2Br + KBr. 
 
 ' Pinner, Franz, Ber. 38. 1589 (1907). ^ Merz, Gasiorowski, Ber. VI. 623 (1886). 
 
 » Burmann, Bull. Soc. [3] 35. 801 (1906). 
 
 ^OH 
 C=0 
 \0H 
 
 /NH2 
 C^O 
 
 \NH2 
 
 Carbonic acid 
 
 Urea 
 

 Alkylamines : Formation 17 
 
 When the product is treated with fuming hydrochloric acid it splits up into 
 phthalic acid and brom-ethylamine : — 
 
 ^O. ^"v'COOH 
 
 .OO'^^"^^'"^^''"^'' ^ ^ ^^^ = CCcO-OH ^ H2N.CH,.CH,.Br. 
 
 3. Eeduction of bodies containing nitrogen doubly or triply linked to 
 carbon. (This naturally cannot be used to prepare aromatic amines.) 
 
 a. Eeduction of nitriles: most conveniently with sodium and alcohol. 
 (Mendius reaction.) 
 
 CaHg.CfcN + 4 H = CaHj-CHg-NHg. 
 
 It may also be carried out' by passing the vapour of the nitrile mixed 
 with hydrogen over powdered nickel at 180-220°, whereby a mixture of 
 primary, secondary, and tertiary amines and ammonia is formed, the secondary 
 amine predominating. 
 
 6. Reduction of the oximes : — 
 
 Ch'>C=NOH + 4 H = c§'>CHNH2 + H^O. 
 
 This can also be done'' by means of hydrogen gas in presence of heated 
 nickel or copper. 
 
 A reaction which may be compared with this, though it does not properly 
 belong to this group, is the rediiction of the aldehyde-ammonias, which are 
 really a-oxy-amines : — 
 
 CH3C<ff -»> CH3C-OH -^ CH3C-H ■ 
 
 This is used to some extent commercially,' the aldehyde being mixed with 
 aqueous ammonia and the solution electrolysed. 
 
 c. Reduction of hydrazones: for example that of acetone: — 
 
 (CH3)2C=N-NH0 + 4H = (CH3)2CH-NH2 + HgN^. 
 
 It is a general rule that whenever a body containing a chain of two 
 nitrogen atoms is reduced, they separate from one another. 
 
 4. Reduction of nitro-compounds. This is of the greatest importance in 
 the aromatic series, but it can also be used for the fatty compounds: — 
 
 RNO2 -* RNH2. 
 
 5. Finally, there are two very peculiar reactions leading from the amide 
 or azide of an acid to an amine with one atom of carbon less. Their 
 mechanism will be discussed in detail later. 
 
 a. The Hofmann reaction,* starting with the amide. This is dissolved in 
 bromine, and then distUled with excess of potash. There are various inter- 
 mediate stages,' but the result is that the CO group of the amide is oxidized 
 by the bromine to COg and eliminated : — 
 
 R-CO-NHa -t- Brj + HgO = RNH^ + CO^ -|- 2HBr. 
 The primary amine goes over mixed with some ammonia, but free from 
 
 ' Sabatier, Senderens, C. B. 140. 482 (1906). ^ MaUhe, C. M. 140. 1691 ; 141. 113 (1905). 
 
 ^ Knndsen, 0. 03. ii. 271 ; 04. i. 134. * Ber. IS. 762 (1882). = See p. 82. 
 
18 
 
 Amines 
 
 secondary and tertiary amines. It is collected in hydrochloric acid, and 
 purified by extraction with absolute alcohol, in which ammonium chloride is 
 practically insoluble. The yield in the lower series is excellent, but in the 
 higher is bad, as the bromine then removes hydrogen from the amine to form 
 the nitrile (reversal of the Mendius reaction). 
 
 K.CH,.NH, + 4Br 
 
 EC=N + 4HBr. 
 
 i. The Curtius reaction ' is similar to this. If the acid azide is boiled with 
 hydrochloric acid, it breaks up into nitrogen, carbonic acid, and the amine : — 
 E-CO-Ng + H2O = E-NHa + CO2 + Ng. 
 
 It has been shown by Forster'' that this reaction is probably due to the 
 intermediate formation of an isocyanate, which can actually be isolated in 
 some cases : — E-CO-N, 
 
 Nj + E.N=::Ct=0 
 
 E-NHa + CO2 
 
 Properties 
 
 The boiling-points of some of the simpler amines are given in the 
 following table ' : — 
 
 
 Mono- 
 
 Di- 
 
 Tri-amine. 
 
 methyl 
 
 - 7° 
 
 + 7° 
 
 + 3-5° 
 
 ethyl 
 
 + 19° 
 
 + 56° 
 
 + 90° 
 
 propyl 
 
 49° 
 
 110° 
 
 156° 
 
 isopropyl 
 
 32° 
 
 84° 
 
 — 
 
 isoamyl 
 
 95° 
 
 187° 
 
 235° 
 
 n-octyl 
 
 176° 
 
 297° 
 
 366° 
 
 The lower members are very soluble in water, and have a smell resem- 
 bling that of ammonia, though often very unpleasant. They are not easy to 
 distinguish from ammonia except by the fact that they burn : it was in this 
 way that Wurtz discovered that the gas he obtained from ethyl isocyanate 
 was not ammonia, as he had for some time supposed. As the molecular 
 weight increases, the smell and the solubility in water diminish. The 
 alkylamines are distinct bases,* with an alkaline reaction to litmus, and they 
 absorb carbon dioxide from the air. These properties are not shared by the 
 aromatic amines, which are much less basic. In fact the successive intro- 
 duction of alkyl groups into ammonia increases the basicity, while that of 
 aromatic groups diminishes it. 
 
 The alkylamines precipitate the hydroxides of many metals from solutions 
 of their salts, but not always the same metals as ammonia. They form stable 
 salts with mineral acids. It is remarkable that Hantzsch'^ has found by the 
 boiling-point method that the molecular weight of dimethyl-ammonium 
 chloride, (CH3)2NH2C1, dissolved in chloroform is in a 1 per cent, solution 
 
 1 Ber. 27. 778 (1894) ; 29. 1166 (1896). » J. C. S. 1009. 433. 
 
 ' Meyer, Jacobson, Lehrb. I. 1. 864. 
 
 * At low temperatures (—75°) they seem to be capable of forming acid salts with two or three 
 molecules of haloid acid. See Korozynski, Ber. 41. 4379 (1908); Eaufler, Kunz, Ber. 42 
 385 (1909). 5 Ser. 38. 1045 (1905). 
 
Alkylamines : Properties 19 
 
 three times, and in a 3 per cent, solution four times that required by the 
 simple formula. 
 
 Many amines also combine with water to form stable hydrates (of the 
 type RNH3OH, corresponding to NH4OH) which can be dried with potassium 
 carbonate without breaking up. These hydrates are generally oily liquids, 
 which only lose water when treated with potash or barium oxide. 
 
 Methods of separation of the various classes of amines 
 
 The quaternary compounds are easily removed by distilling the mixed 
 salts with potash. The primary, secondary, and tertiary amines pass over, 
 while the quaternary hydroxides, not being volatile, remain behind. 
 
 A general method for the separation of primary, secondary, and tertiary 
 amines is that of Hinsberg,' which depends on their behaviour with benzene- 
 sulphonic chloride, CgHsSOaCl. The mixed bases are treated with an equi- 
 valent quantity of this reagent, and shaken with potash till the smell of the 
 chloride disappears. The acid chloride reacts with the amines as it does with 
 ammonia, the chlorine being eliminated along with the hydrogen attached 
 to the nitrogen (the potash assists the reaction just as it does in the Schotten- 
 Baumann method of benzoylation). As there is no such hydrogen in the 
 tertiary amines the sulphonic chloride does not react with them at all. With 
 the primary and secondary amines it replaces one H by 0SO2, giving with 
 the primary ENH-SOj^, and with the secondary Ea^'SOa^. This last is 
 insoluble in potash, while the derivative of the primary amine, which still 
 has a hydrogen atom attached to the nitrogen, is soluble, since the strongly 
 negative phenyl-sulphonyl group makes this hydrogen acidic. Thus on 
 extracting the alkaline solution with ether, the tertiary amine is removed 
 along with the phenyl-sulphonyl derivative of the secondary, from which it 
 can be separated by distillation ; while the derivative of the primary remains 
 in solution, and can be precipitated by acid. The amines can be recovered 
 from these derivatives by heating with strong acid.'^ 
 
 There is no other general method' of separation, but there are a variety 
 of special methods for use in particular cases, of which the best known is 
 that employed by Hofmann for separating the ethylamines. It is remarkable 
 that although the boiling-points of the three ethylamines lie so far apart 
 (19°, 56°, and 90°) Hofmann found that it was impossible to separate them by 
 fractional distillation even when he used more than a kilogram of the mixed 
 amines. This must be due to their vapour pressures falling very gradually 
 with the temperature. 
 
 To separate them, the dry mixture of bases is heated with dry ethyl 
 
 oxalate. Triethylamine does not react, and can be distilled off. The other 
 
 ,. X, , .,. CO-NEta 
 
 two react in different ways, diethylamme givmg diethyl-oxamethane, | 
 
 OO'OEt 
 
 1 Ber. as. 2962 (1890). " Cf. Hinsberg, Kessler, Ber. 38. 906 (1905). 
 
 ' For a method of identifying the various classes of amines by the action of Grignard's reagent, 
 see Sudborongh, Hibbert, /. C. S. 1909. 477. Tor another method, see y. Braiin, Ber. 41. 
 2156 (1908). 
 
20 Amines 
 
 and ethylamine diethyl-oxamide, ^^'^HEt rpjie reason for this difference 
 
 CONHEt 
 is unknown. (The best way of remembering it is to observe that both the 
 products have two ethyl groups attached to nitrogen.) On extracting the 
 mass with water, the diethyl oxamide dissolves, while the ester does not ; 
 and if the aqueous solution is boiled with potash it gives oxalic acid and 
 ethylamine, while from the residue the diethylamine can be obtained by a 
 similar process of saponification. 
 
 Chemical properties of the alkylamines 
 
 The reactions of the four classes are in general very different, since they 
 largely depend on the hydrogen remaining attached to the nitrogen; thus in 
 many eases tertiary amines do not react at all. 
 
 This hydrogen on the nitrogen can be replaced by sodium or potassium, 
 giving rather unstable compounds. It can also be replaced by acyl groups, 
 forming substituted amides, on treatment with acid anhydrides or chlorides : — 
 
 CjHsNHa + (CH3CO)20 = C2H5NHCOCH3 + CH3COOH. 
 This is also true of the aromatic aniines. 
 
 Primary and secondary fatty amines react with carbon bisulphide to give 
 alkyl-sulphocarbamic acids : — 
 
 C^Hs-NH, , C=S _ c^Hs-NH \ 
 
 the acid forming a salt with a second molecule of the amine. Aromatic 
 amines behave in a different way. These salts, if they are formed from a 
 primary amine, as in the above example, break up when treated with mercuric 
 chloride thus : — 
 
 C,H,NH x^g _ C,H,N=C=S 
 
 giving an amine and an isothiocyanate or mustard oil, which can be detected 
 by its peculiar smell. This is Hofmann's ' mustard oil ' reaction for detecting 
 primaiy amines. 
 
 Another test of Hofmann's for primary amines, which goes equally well 
 with alkyl and aryl amines, is the isonitrile reaction. The amine is heated 
 with chloroform and alcoholic potash, when the isonitrile is formed, which is 
 recognized by its smell : — 
 
 CaHgNHj + CHCI3 = CaHgN^C + 3HC1. 
 A fraction of a drop of amine is suf&cient for this test. 
 
 The behaviour of the three classes of amines with nitrous acid is very 
 characteristic. As usual, the nitrous acid, HO-N==0, can react with either 
 end of its molecule, the hydroxyl going out as water with one hydrogen atom 
 or the oxygen atom with two. 
 
 The primary alkylamines form water, nitrogen, and the alcohol : — ■ 
 C2H5NH2 + 0=:NOH = CaHgOH + N^ + H^O. 
 A nitrite is of course formed as an intermediate product, and it has recently 
 
Alkylamines : Properties 21 
 
 been shown by Wallach ' that these nitrites are much more stable than is 
 generally supposed, and if prepared in absolutely neutral solution can often be 
 isolated, especially in the case of the polymethylene derivatives ; though they 
 are rapidly decomposed by small quantities of acid. The alcohol produced is 
 often an isomer of what one would expect : for example, a secondary instead 
 of a primary. This peculiar change has been investigated by Henry.^ He 
 finds that normal propylamine gives 58 per cent, of isopropyl alcohol and 
 only 42 per cent, of the normal alcohol ; (CH3)2CH-CH2-NH2 gives 75 per cent, 
 of trimethyl carbinol (CH3)3C-OH, and 25 per cent, of isobutyl alcohol ; while 
 (CH3)3COH is wholly converted into C2H5(CH3)2COH. In other words, the 
 larger the number of methyl groups attached to the carbon atom next but 
 one to the amino group the greater the extent to which this change takes 
 place. 
 
 In the polymethylene derivatives the ring is often affected. Thus tri- 
 methylene-amine gives allyl alcohol ' : — 
 
 I />CH.NH2 -^ CHa^^CHOHa-OH. 
 Ga..2 
 
 More commonly the ring increases in size, a 3-ring becoming a 4-ring ' : — 
 
 ^^ ^CH.CH,NH2 ^ 1^ _ I ^:^^ 
 
 
 or a 5-ring a 6-ring ° 
 
 S- 
 
 CH2-CH2/^^*^^^"^^^ "" CH2.CH2CH.OH- 
 
 On the other hand a 6-ring will apparently not go into a 7-ring ; for Wallach 
 finds " that when cyclohexylamine CeH11.CH2.CH2.NH2 is treated with nitrous 
 acid the ring is not altered, the product consisting of a mixture of the alcohols 
 CfiHu-CHaCHaOH and CeHij.CHOHCHg, together with a small quantity of 
 the olefine CoHii.CH=CH2. 
 
 Secondary amines with nitrous acid give nitrosamines : — 
 
 (C2H5)2]Sr.H + HO-NO = H2O + (C2H5)2N.NO. 
 
 Here also a nitrite is no doubt an intermediate product ; and in one case, 
 that of di-isopropylamine, the salt [(CH3)2CH-]2NH2.0.NO has been isolated. 
 
 The tertiary amines, having no hydrogen on the nitrogen, do not react 
 with nitrous acid at all. 
 
 The amine hydrogen of the primary and secondary amines can be replaced 
 by chlorine on treatment with bleaching powder, giving such compounds 
 as CH;,.NCl2, which must have its chlorine attached to nitrogen since with 
 zinc methyl it gives trimethylamine. These bodies are explosive, but much 
 less so than nitrogen chloride, and they can often be distilled unchanged. 
 The chlorine is very loosely attached to the nitrogen, as is shown by the 
 
 ' C. 07. ii. 54. ^ 0. B. 145. 899, 1247 (1907). ' Kiahner, C. 05. i. 1708. 
 
 » Demjanow, Ber. 40. 4393 (1907). ' Wallach, C. 07. ii. 54. 
 
 « Ann. 359. 287 (1908). 
 1175 
 
22 Amines 
 
 fact that on warming with concentrated hydrochloric acid it is replaced by 
 hydrogen, the amine being reproduced : — 
 
 C5H11NHCI + HCl = CI2 + C5H11NH2. 
 The hydrochloric acid here behaves as a reducing agent, like hydriodic acid. 
 Other compounds with chlorine attached to trivalent nitrogen are reduced by 
 hydrochloric acid in the same way. 
 
 When the haloid salts of the amines are heated to a high temperature, 
 alkyl halide is split oif. If there are several alkyls present including methyl, 
 it is always the methyl which comes off. This fact is made use of in 
 preparing methyl chloride on the large scale from the trimethylamine of 
 the beet sugar residues. 
 
 Of the individual alkylamines methylamine, the simplest, can be made 
 by passing a mixture of hydrocyanic acid and hydrogen over platinum black 
 at 110°. 
 
 The one with the longest chain is heptadegylamine, C17H35NH2. It melts 
 at 49° and boils at 335°. It is as basic as the lower amines. Its alcoholic 
 solution absorbs carbon dioxide from the air, and has a strong alkaline reaction. 
 But its hydrochloride, though easily soluble in alcohol, is insoluble in water, 
 whereas the hydrochlorides of the lower amines are deliquescent. 
 
 QUATEENAEY AMMONIUM COMPOUNDS 
 
 They were discovered by Hofmann as the final product of the action of 
 the alkyl halides on alcoholic ammonia. They may be made by acting with 
 the alkyl halides on the tertiary amines. The reaction goes with very 
 different ease in different cases : thus trimethylamine combines with methyl 
 chloride at the ordinary temperature with evolution of heat, but it will not 
 combine with ethyl chloride at the ordinary temperature at all. The velocity 
 of reaction of triethylamine with ethyl iodide to form tetra-ethyl-ammonium 
 iodide was investigated in great detail by Menschutkin.' He found that 
 it was enormously affected by the solvent. Thus it was 720 times as great 
 in acetophenone as in hexane. 
 
 The other salts are generally got from the iodide by treatment with the 
 silver salt of the acid in question. They mostly crystallize well. The 
 haloid salts all break up on heating into tertiary amine and alkyl halide. 
 If one of the alkyls is methyl, it is always methyl halide that is split off, 
 as happens with the salts of the other amines also. 
 
 The quaternary hydroxides are strong bases. It is often said that they 
 are so strong that they cannot be prepared from their salts by the action 
 of potash. This is not true ; only as both products, the quaternary hydroxide 
 and the potassium salt, are soluble, it is not easy to isolate them in this 
 way. It is however possible in some cases to get the hydroxide to crystallize 
 out. The usual method of preparation is to treat the iodide with moist 
 silver oxide : — 
 
 NMeJ + AgOH = NMe^OH + Agl. 
 
 The silver iodide is filtered off and the solution evaporated. 
 
 ' Z. Ph. Ch. 6. 41. (1890). 
 
Quaternary Ammonium Compounds 23 
 
 They resemble potash to an extraordinary degree. They form deliquescent 
 crystalline masses ; their aqueous solution turns red litmus blue, and absorbs 
 carbon dioxide readily. It dissolves the skin and saponifies fats. 
 
 When heated the hydroxides are converted into tertiary amines, one alkyl 
 being split off. If this is a methyl, it appears as methyl alcohol ; but in all 
 other cases the alcohol which may be supposed to be formed breaks up further 
 into water and alkylene, for example : — 
 
 (CH3)3N<gg2-^^3 = (CH3)3N + CH,:CH, + H^O. 
 
 It is to be noticed that if one of the alkyls is a methyl, it remains attached 
 to the nitrogen ; whereas, as we have seen, when the hydroxyl is replaced 
 by halogen, the methyl group is the first to come off. 
 
 As there is a certain amount of evidence for the existence of free 
 ' ammonium ', NH4, in ammonium amalgam, attempts have been made to 
 isolate tetramethyl-ammonium, N(CHg)4, which we should expect to be more 
 stable. These have not succeeded, but it has been shown ^ that if a solution 
 of tetramethyl-ammonium chloride (free from alkali) in liquid ammonia is 
 electrolysed, deep blue layers are formed at the cathode, similar to those 
 produced by dissolving sodium in liquid ammonia. In the latter case the 
 blue compound, which is formed without evolution of hydrogen, is probably 
 the substituted ammonium, NHgNa. Hence it is probable that in the former 
 case the blue substance is the analogous tetramethyl-ammonium, ^(CHg)^. 
 Similarly, it is found ^ that metallic caesium dissolves in liquid methylamine 
 at very low temperatures to form a blue liquid, no doubt containing CH3NH2CS ; 
 but when the temperature rises to — 20° the colour disappears and hydrogen 
 is evolved, the substituted metalamide NHCHgCs being left behind. 
 
 A question which may be conveniently discussed in connexion with the 
 quaternary ammonium compounds is that of the structure of ammonium 
 compounds and of pentavalent nitrogen derivatives in general. The earlier 
 supporters of the theory of constant valency held that nitrogen was triad, 
 and explained the existence of the quaternary derivatives by calling them 
 molecular compounds. This view was generally abandoned when in 1876 
 V. Meyer and Lecco ' showed that the same body was obtained from 
 trimethylamine and ethyl iodide as from ethyl-dimethylamine and methyl 
 iodide. It was then recognized that nitrogen in these compounds had a valency 
 of five. Recently however some chemists, especially Werner,* have introduced 
 a view not unlike that of molecular compounds. Werner supposes that in 
 ammonium chloride, for example, the whole hydrochloric acid molecule is 
 attached to the ammonia by a new kind of bond ; and he writes it NH3 . . . HCl, 
 
 [XT. /TT"1 
 ■jt./>N\tt CI ; the group enclosed in the bracket is the 
 
 cation: the dotted line in this last formula is intended to show that one of 
 the four groups is attached to the nitrogen in a different way. All these 
 
 ' Palmaer, Z.f. MeMrochem. 8. 729 (1902). " Eengade, C. R. 140. 246 (1905). 
 
 ' Ann. 180. 173. 
 
 * Cf. Ann. 322. 276 (1902) ; Lehrb. d. Sfereochemie , p. 310 (1904). Hantzaoh also inclines to 
 the same view ; Ber. 38. 2161 (1905). 
 
 c 2 
 
24 Amines 
 
 suggestions of new species of valencies are to be regarded with grave suspicion. 
 They are usually advanced to account for some class of reactions which is 
 not yet thoroughly understood, and we should be cautious of accepting them 
 until there is a great mass of evidence in their support, and until we are 
 quite certain that we cannot explain the facts without them. It is safe to 
 say that at present of all the various new types of linkage which Werner, 
 Baeyer, and others have offered to us, not one has yet made good its claims. 
 (This does not, of course, apply to Thiele's theory of conjugate links, which 
 is only an extension of the ordinary structural theory.) 
 
 At the same time it is worth considering the evidence which has been 
 adduced in support of this view of the structure of the pentavalent nitrogen 
 compounds. Hantzsch ' has argued in its favour from the behaviour of the 
 dibromides of the tertiary amines. Trimethylamine combines with bromine 
 to form a dibromide (CH3)3NBr3. When this is treated with potash, we 
 
 should expect it to give first the compound (CH3)3N<^g > a well-known 
 
 body, obtained from the amine-oxide (CH3)gN=0, and convertible into it by 
 the further action of alkali. But the dibromide with potash does not give 
 this body, nor the amine-oxide. It breaks up into trimethylamine, potassium 
 bromide, and potassium hypobromite. Hantzsch regards this as a proof that 
 
 /Br 
 
 it cannot have the formula (CH3)3N\g , and he writes it on Werner's model 
 
 I Qjj^/N\-gj. MBr, whereby the action of potash on it — 
 
 [(CH3)3NBr]Br -» [(CH3)3NBr]OH -* (CH3)3N 4- HOBr 
 appears analogous to the action of potash on trimethylamine hydrobromide : — 
 [(CH3)3NH]Br -^ [(CH3)3NH]OH -* (CH3)3N + HOH. 
 
 This is ingenious, but we can surely explain the reaction without having 
 recourse to this new hypothesis. These dibromides readily break up again 
 into their components, the tertiary amine and bromine, as Hantzsch admits, 
 and therefore the body might react with potash in two different ways : — 
 
 (CH3)3NBr,^(^^^^^^^Br 
 
 ^(CH3)3N -1- Br^ -^ (CH3)3N -t- KBr + KOBr. 
 
 Which of these two reactions actually occurs will depend on their relative 
 velocities, as to which nothing is known. As we find that the second takes 
 place, this must be the quickest. The assumption of a new type of valency 
 seems quite unnecessary. 
 
 Another argument of more value is derived from the influence of the 
 successive introduction of alkyls on the basicity of the amine hydrates. This 
 quantity requires some explanation. The statements to be found in the 
 textbooks as to the strength of organic bases are often misleading. They 
 are frequently based on quahtative evidence, such as the readiness of formation 
 of salts with weak acids, where many other factors are concerned besides 
 
 ' Ber. 38. 2154 (1905). 
 
Structure of Ammonium Compounds 25 
 
 the actual strength of the base ; and even where they depend on a measure- 
 ment of the concentration of the hydroxyl ions, as in Bredig's work, they 
 are affected by a disturbing influence which for a long time escaped notice. 
 A primary, secondary, or tertiary amine, E3N, when dissolved in water under- 
 goes two changes. It takes up water to form the quaternary hydroxide, and 
 this then dissociates electrolytically : — 
 
 E3N + H,0 ±j E3N<oH ^ E3NH' + OH'. 
 X y z 
 
 Now if X be the concentration of the amine E3N, y of the ammonium 
 hydrate K3NHOH, and s of the ions, we know that 
 
 - = constant, = 6. 
 
 y 
 
 Further, from Ostwald's law — = K. 
 
 y 
 
 The real strength of the base is given by K. But in the ordinary methods 
 for determining basicity we measure z, the concentration of the ions, in 
 a solution in which the total concentration of the base in all three forms 
 is known, the existence of the first form, the non-hydrated tertiary base E3N, 
 being neglected. The value of the strength calculated from this is not the 
 
 true K, but what may be called the apparent strength, Kj, = 
 
 X + y 
 
 But since x = ty, Kj 
 
 hy + y 
 
 «'2 
 
 = K 
 
 y{p + 1) 
 1 
 & + l' 
 
 Hence the apparent strength of the base is smaller than the real strength, 
 and the more so the greater the proportion of anhydrous amine present. 
 When therefore the strength of an organic base (other than quaternary) is 
 spoken of, what is meant is the product of two factors, the hydration constant 
 and the dissociation constant. And until recently there was no method 
 known for determining either of these two factors separately. The measure- 
 ment of the ionization is no use, as the hydration factor comes in to the 
 same extent at every dilution. The determination of the partition-coeificient 
 between water and air or an organic solvent has been suggested ; but this 
 does not help us either. The concentration of the E3N in the other solvent 
 is proportional to that of the E3N in the water, and therefore also to that 
 of the E3NHOH: which leaves us where we were before. The problem 
 has at length been solved by Moore.' He points out that all methods of 
 solution must fail, in which the observations are all made at the same 
 temperature. If however we measure the partition-coefiicients and the degree 
 of dissociation at different temperatures, and if we further make the assumption, 
 which for small differences of temperature is justifiable, that the temperature- 
 
 ' J. C. 8. 1907. 1373. 
 
26 Amines 
 
 coefficients of these two quantities are constant, we obtain a sufficient number 
 of equations to solve the problem. In this way we can determine for any 
 amine (and the method applies equally to all pseudo-acids and pseudo-bases, 
 and to such bodies as carbonic acid and the lactones) the proportion present 
 in the non-hydrated form, and the true dissociation-constant. So far, the 
 necessary data have only been obtained for three substances, piperidine (which 
 does not concern us at present), ammonia, and triethylamine. The results 
 show that in an aqueous solution of ammonia at 20°, about two-thirds is 
 present as non-hydrated NHg, the rest being mainly NH4OH and a small 
 quantity of ions; in the case of triethylamine only about a third is in the 
 anhydrous form, as (€2115)3]^. From these data the true dissociation constants 
 of these two bases can be calculated, and are found to be at 20°: — 
 
 Ammonia K = 5-23 x 10-« 
 
 Triethylamine 64 x lO-^. 
 
 That is to say, the introduction of three ethyl groups into ammonia has 
 increased the constant only twelve-fold. We should therefore expect that the 
 introduction of a fourth ethyl group, in tetra-ethyl-ammonium hydroxide, 
 would involve a further increase only to about 150 x 10~^, which would 
 mean that it was still a very weak base. But as a fact the quaternary 
 hydroxide is an excessively strong base, comparable to potash, for which K 
 is too high to be measured and is certainly greater than 1. It therefore 
 appears that the first three ethyl groups produce only a small increase in 
 the basicity of ammonium hydrate, while that produced by the fourth is 
 enormous. It has long been known that the quaternary bases were far 
 stronger than the others, but until the true dissociation-constants had been 
 determined it was possible that this might be due to the unknown quantity 
 of anhydrous base present in the latter cases but not in the former. This 
 is no longer possible ; and the sudden increase in the quaternary bases 
 certainly seems to point to a difference between their constitution and that 
 of the hydroxides of the primary, secondary, and tertiary amines. If we are 
 not prepared to accept a formula such as Werner's, there are two possible 
 explanations. One is that the data on which these calculations are based 
 (especially those for triethylamine) may be incorrect. But even if they are 
 to be accepted, and this sudden change of basicity really occurs, it can be 
 explained on the assumption of analogous structures for the tertiary and 
 quaternary hydroxides, and without having recourse to any new theories of 
 valency. If we say that the five valencies of pentavalent nitrogen are all 
 equal, this only means that the relation between any five groups and the 
 nitrogen atom is determined by those groups themselves, so that (excluding 
 stereoisomerism) isomeric arrangements are impossible. It is nevertheless 
 quite conceivable, and indeed probable, that three of the nitrogen valencies 
 (those of trivalent nitrogen) are essentially different from the other two : and 
 that when successive alkyl groups are introduced into ammonium hydrate 
 they fill up these three places first. If so, the introduction of a fourth group, 
 which must now take up one of the other two positions, may be expected 
 to produce an effect on the molecule different from that produced by any 
 of the other three. 
 
Solubility of Alkylamines 27 
 
 Another phenomenon which is no doubt connected with the formation of 
 hydrates is the remarkable effect of temperature on the solubility of certain 
 amines in water.^ When a liquid is not completely miscible with water, it 
 is almost invariably found that with a rise of temperature the solubilities 
 both of the liquid in water and of water in the liquid increase, until a point 
 is reached at which they become completely miscible. But there are certain 
 amines (triethylamine, diethylamine, /3-collidine) for which these relations are 
 reversed. At low temperatures they are completely miscible with water, but 
 on warming, separation into two layers takes place, and with a further rise 
 of temperature the solubility of each substance in the other diminishes ; 
 although there are indications that at a still higher temperature their mutual 
 solubilities begin to increase again. In the case of nicotine ^ both phenomena 
 are combined. Below 60° and above 210° nicotine is miscible with water in all 
 proportions : but between these two temperatures it is only partially soluble. 
 The fact that these peculiar phenomena occur to a marked extent only with 
 substances of the same chemical type suggests that they depend on a chemical 
 reaction ; and it is fairly obvious that this is the formation of a hydrate. On 
 the analogy of the quaternary hydroxides we should expect triethyl-ammonium 
 hydroxide, EtsNHOH, to be excessively soluble in water, while the anhydrous 
 triethylamine, containing no hydroxyl, should be comparatively insoluble. 
 A rise of temperature will no doubt increase both these solubilities, as this 
 is its effect in all normal cases : but at the same time it will increase the 
 proportion of anhydrous to hydrated base. Indeed Moore has shown for 
 triethylamine that the proportion of base present as EtsN is at 15° 30 per 
 cent., and at 35° over 50 per cent. Thus the total effect of raising the 
 temperature is greatly to increase the proportion of the much less soluble 
 form ; and this may well overcome the other effect (on the solubility of that 
 form) and cause the total solubUity to diminish. On the other hand, at a 
 much higher temperature, where the dissociation into anhydrous base and 
 water is practically complete, and it is really the solubility of the anhydrous 
 base alone that is being measured, the second effect may be expected to 
 predominate, and the solubility to begin to increase again, as happens in the 
 case of nicotine. This explanation is much strengthened by the fact that 
 the only liquids other than amines which show, though to a much smaller 
 extent, a marked diminution of solubility with rise of temperature, are certain 
 ketones and lactones, which from their structure should be capable, though in 
 a much lower degree, of forming hydrates by the oxygen becoming tetravalent. 
 
 STEEEOCHEMISTEY OF PENTAVALENT NITEOGEN 
 
 The quaternary ammonium compounds" are of special interest from their 
 bearing on the stereochemistry of nitrogen. As early as 1877 Wislicenus 
 pointed out that van 't Hoff's theory of asymmetric carbon should be capable 
 
 » Rothmund, Z. Ph. Ch. 26. 433 (1898). Lattey, Phil. Mag. [6] 10. 397 (1906) ; J. C. S. 
 1907. 1959. FlaBchner, Z. Ph. Ch. 62. 493 (1908). 
 ' Hudson, Z. Ph. Ch. 47. 113 (1904). 
 ' Cf. Wedekind, Zur Stereochemie des fimfwertigen SticJcstoffes (Leipzig, 1907). 
 
28 Amines 
 
 of extension to the case of pentavalent nitrogen. But for twenty years all 
 efforts to discover cases of stereoisomerism among the pentavalent nitrogen 
 compounds were unsuccessful. (The case of the trivalent nitrogen derivatives 
 will be discussed under the oximes.) In 1891 Le Bel ^ claimed to have obtained 
 an optically active form of isobutyl-propyl-ethyl-methyl-ammonium chloride 
 by the action of moulds ; but no one else has been able to repeat his work. 
 In 1899 Wedekind ^ obtained two inactive but apparently stereoisomeric forms 
 of benzyl-allyl-methyl-phenyl-ammonium bromide, differing in melting-points 
 and physical properties generally. As no other analogous compounds could 
 be obtained in isomeric forms, and as further neither of these two isomers 
 could be converted into the other, doubt was thrown on the work, and it 
 was repeated by Hantzsch and Horn,' who confirmed Wedekind's conclusions. 
 Eecently, however, both Wedekind ■* himself and H. C. Jones ^^ have shown that 
 the second of these supposed isomers is really benzyl-phenyl-dimethyl-ammonium 
 bromide, the allyl group being expelled by the methyl in the course of the 
 preparation. 
 
 There have been several other supposed cases of this kind, of inactive 
 isomerism of completely asymmetric ammonium compounds, depending on 
 the order in which the various groups are introduced ; but they have all been 
 shown to be erroneous, although such bodies often give two dimorphic crystalline 
 forms." 
 
 The first optically active nitrogen compounds were obtained by Pope and 
 Peachey' in 1899. Wedekind had previously tried to resolve his benzyl- 
 phenyl-allyl-methyl-ammonium salt (the true one) by fractional crystallization 
 of its salts with active acids, but in vain. Pope and Peachey employed the 
 same method in non-aqueous solvents, in order to avoid the effect of disso- 
 ciation, and they used dextro-camphor-sulphonic acid and its bromo-derivative, 
 which have the advantage of being both powerfully active and also very strong 
 acids. The iodide of Wedekind's base, dissolved in a mixture of acetone and 
 ethyl acetate, was treated with the silver salt of this acid, and thus converted 
 into its d-camphor sulphonate. This salt was then fractionally crystallized 
 from the same solvent, and so separated into the less soluble salt of the 
 dextro-base and the more soluble salt of the laevo-base. These bodies on 
 treatment with potassium iodide or bromide are converted into the iodide or 
 bromide, which are found to be optically active in solution. In these latter 
 salts, for instance in the bromide 
 
 C3H/ ^' 
 there is no possible asymmetric carbon atom ; and thus it is conclusively 
 proved that the phenomenon of optical activity in solution is not confined 
 to carbon, but is shared by nitrogen as well. 
 
 ' G. B. 112. 724. ' Ser. 32. 517. ' Ber. 35. 883 (1902). 
 
 * Ber. 39. 481 (1906). » /. C. S. 1905. 1721. 
 
 « Cf . Le Bel, J. CUm. Phys. 2. S40 (1904) ; Kipping, J. C. S. 985. 628 ; Schryver aud CoUie, 
 Ch. News, 63. 174. 
 
 ' Proc. C. S. 15. 192 ; J. C. S. 1899. 1127. 
 
Stereochemistry of Pentavalent Nitrogen 29 
 
 It is to be noticed that the activity is retained in the platinum and silver 
 double salts. 
 
 Since 1899 there have been many attempts to prepare the active forms of 
 ammonium compounds of this type. Some of them have failed, but a good 
 many have been successful, so that by now we have more than a dozen 
 quaternary ammonium compounds known in their two active forms. It is 
 evident that all bodies of the type K1E2E3E4NX are capable of giving two 
 active modifications, but in many cases one changes into the other too easily 
 for separation to be possible. It should also be possible to get activity when 
 two of the four hydrocarbon groups are replaced by the two parts of a carbon 
 ring of which the nitrogen forms part ; this has recently been realized by 
 Buckney' and Wedekind,^ who have separated the two forms of the salts of 
 aUyl-kairolinium (allyl-methyl-tetrahydro-quinolinium) hydroxide : — 
 
 N-OH 
 
 CH3 C3H5 
 
 In solution the asymmetric bodies lose their activity more or less rapidly, a 
 phenomenon originally attributed to the change of one antimer into the other, 
 and known as auto-racemization. This has been investigated by Wedekind,'' 
 who found it to be due to a monomolecular reaction. It proceeds much 
 faster in chloroform than in alcohol. It is affected by the nature of the acid 
 radical ; the free base (hydroxide), the chloride, bromide, and iodide all change, 
 the iodide the quickest ; while the nitrate, the camphor-sulphonate, and 
 apparently the fluoride, do so much more slowly or not at all. It is also of 
 course affected by the nature of the base. In the series which Wedekind 
 investigated, the alkyl-phenyl-methyl-benzyl-ammonium compounds, the race- 
 mization is much more rapid when the alkyl is isobutyl than when it is 
 allyl or normal propyl. 
 
 Later work** has shown that the primary cause of this change is the 
 dissociation of the quaternary salt into tertiary amine and alkyl halide: — 
 
 E^NBr -^ E3N + EBr. 
 
 This dissociation can be followed by means of the fall in the amount of 
 halogen ion, by titrating the solution from time to time with silver. It is 
 found that the diminution of activity is proportional to the fall in the amount 
 of silver solution required, showing that the undissociated salt is wholly in 
 the original active form. Measurements of the molecular weight by the 
 cryoscopic method, and of the conductivity, confirm these results. The 
 temperature coefficient of the velocity constant in these cases is imusually 
 high, being between 4 and 5 for 10°. The solvents used were in most cases 
 
 > Proe. Canib. Phil. Soc. 14. 177 ; C. 07. ii. 820. " Ber. 40. 4450 (1907). 
 
 ^ Stereochemie, p. 72 ; Z.f. Meklroehem. 12. 330 (1906). 
 
 ' V. Halban, Z. f. Elektrochem. 13. 57 (1907) ; Ber. 41. 2417 ; E. Wedekind, Paschke, ib. 
 2659 (1908). 
 
30 Amines 
 
 chloroform and bromoform ; in solvents of greater ionizing power, such as 
 alcohol and water, the dissociation does not seem to take place. 
 
 At the same time it may be pointed out that though these results show that 
 the primary reaction is a dissociation, a true racemic change must ultimately 
 ensue. The dissociation is reversible, as an amount of salt varying in different 
 cases from 6 to 40 per cent, was found to remain undissociated at equilibrium. 
 Hence the tertiary amine and the alkyl halide must recombine: and as they 
 are themselves inactive, they must produce a racemic mixture of the dextro and 
 laevo salt. It is thus to be expected that the activity will ultimately disappear 
 entirely, while the sUver titre wUl not. 
 
 A very remarkable extension of the conditions of activity of pentavalent 
 nitrogen has recently been made by Meisenheimer.^ He prepared methyl-ethyl 
 
 Me\ Me\ ^-g 
 
 aniline oxide, a basic substance of the formula Et-N=0 or Et-N<^QTj, which 
 
 Me\ QTT 
 forms salts with acids of the type Et-N\p, . The d-brom-camphor-sulphonate 
 
 of this base was divided by fractional crystallization into two parts, from 
 which two active chlorides (of the above formula) were prepared. This was 
 so far a new discovery, that it was the first case of an active nitrogen compound 
 in which two of the five different groups attached to the nitrogen were negative. 
 A far more unexpected result is that if this salt is treated in solution with 
 baryta, whereby it is practically entirely converted into the weak base E^N^O 
 
 or R;,N<\p,TT, the activity remains. On either formula there are only four 
 
 different groups attached to the nitrogen, and yet the compound is active. 
 
 The stereochemical analogy between carbon and nitrogen has been carried 
 beyond the simple case of a single asymmetric nitrogen atom. E. and O. 
 Wedekind have shown that the compound obtained from methyl iodide and 
 trimethylene-di-ethylaniline 
 
 Et\ /Et 
 
 N_CH,-CH,-CH,-N-0 , 
 Me/ \j. / \Me 
 
 which contains two asymmetric nitrogens with symmetry in the molecule, 
 
 occurs in two inactive modifications, differing in physical properties. These 
 
 must correspond to racemic and mesotartaric acids : one of them should be 
 
 capable of being broken up into two active antimers, but attempts to carry 
 
 this out have failed. 
 
 A further development is due to Aschan.'' He acted (1) on ethylene 
 
 dipiperide with trimethylene bromide and (2) on trimethylene dipiperide with 
 
 ethylene dibromide. The products, which must both be ethylene-trimethylene- 
 
 dipiperidinium bromide : — 
 
 /CH,-CHAj./CH2 CH^Xj^/CH^-CH^x 
 
 ^^2\CH2-CH./|\CH2-CH2-CH/^|'\CH2-CH/*^^2 
 
 " Br Br 
 
 ' Ber. 41. 3966 (1908). " Z. Ph. Ch. 46. 293 (1903). 
 
Stereochemistry of Pentavalent Nitrogen 31 
 
 were found to possess different physical properties, and neither of them could be 
 separated into active forms. This is most easily explained as a case of cis-trans- 
 isomerism, like that of the hexahydro-terephthalic acids : — 
 CgHjo^Br C,Hi„^/Br 
 
 N N 
 
 I 
 and CHj 
 
 CH„ CH, 
 
 ^2 
 
 CH2 CII2 
 
 N 
 C,H,o^^Br Br^VsHio 
 
 The various stereo-formulae which have been proposed for the pentavalent 
 nitrogen atom are still too doubtful to be worth discussing. We must content 
 ourselves at present with observing the facts : which are, to put them briefly, 
 that a nitrogen atom with five different groups can exist in two optically active 
 forms, and offers in general the same possibilities of isomerism as a carbon atom 
 with four different groups. Inactive isomers cannot be produced unless there 
 are two such asymmetric nitrogen atoms : a difference in the order of intro- 
 duction of the groups on a single nitrogen atom is not sufficient to cause 
 isomerism. 
 
 On the other hand, as Meisenheimer's work shows, a nitrogen atom may be 
 active when two of the groups attached to it are the same, if these are negative 
 groups. To explain this, he assumes that four of the groups attached to a 
 pentavalent nitrogen atom occupy positions analogous to the four groups on a 
 carbon atom, allowing of two enantiomorphic arrangements when all four are 
 different, while the fifth (ionizable) group is mobile, and takes up the preferential 
 position determined by the other four. Thus stereoisomerism is possible if one 
 of these four groups is the same as the ionizable group, as it is in his base, 
 but not if two positive groups (hydrocarbon radicals) are identical, which also 
 
 ,QTT 
 
 agrees with experience. On this view the two hydroxyl groups in E3N\qtt 
 
 would be different, and this may account for the fact that these bodies behave 
 as monacid bases. ^ 
 
 SUBSTITUTED ALKYLAMINES 
 Various amines containing substituted alkyl groups exist, most of which 
 are not of great importance. Thus instead of alkyl radicals we may have 
 their chlorine substitution-products attached to the nitrogen. These bodies 
 may be made in various ways. Thus y-chloro-butylamine is got by the reduction 
 of -y-chloro-butyronitrile :— 
 
 CH2CICH2CH2CN -> CH2C1CH2-CH2CH2NH2. 
 It is an unstable compound, which can only exist as a salt. When the free 
 base is liberated it changes spontaneously into the hydrochloride of pyrrolidine 
 (tetrahydro-pyrrol) :— CHo-CH^x 
 
 CH2-CH2/ 
 
 ■ For a case of optical activity due in part to a trivalent nitrogen atom see p. 116. 
 
32 Amines 
 
 The hydroxy-derivatives (amino-alcohols) are of some interest. The most 
 easily obtained are those a-compounds which have the OH and the NHg 
 attached to the same terminal carbon atom. These are the aldehyde-ammonias. 
 They are formed by most (not all) fatty aldehydes with great ease : with the 
 lower members by passing ammonia gas into the ethereal solution of the 
 aldehyde, with the higher by treating the aldehyde with aqueous ammonia. 
 It is to be noticed that the ketones do not give such compounds: in fact 
 their formation depends on the presence of a CHO group attached to a primary 
 alkyl radical : — 
 
 H 
 
 CH3.C<o + NH, = CHa-C-OH. 
 
 NH2 
 They split off the nitrogen as ammonia with remarkable ease in presence 
 of dilute acids, re-forming the aldehydes. They can be reduced to alkyl- 
 amines, and they polymerize readily. Thus ordinary aldehyde-ammonia when 
 exposed to the air forms a substance resembling aldehyde resin. 
 
 The other oxy-amines, which have the OH and the NHg on different 
 carbon atoms, are sharply distinguished from the aldehyde-ammonias by the 
 fact that the nitrogen is firmly attached. They are strong stable bases. They 
 are of interest from their close relationship to certain natural products. The 
 dialkyl -amino-alcohols ^ and their esters are powerful anaesthetics. 
 
 The amino-alcohols are formed : — 
 
 (1) By the action of hal-hydrines such as glycol chlorhydrine on ammonia : — 
 
 CH2OHCH2CI + NH;j = CH2OHCH2NH2 -f HCl. 
 This reaction of course gives rise to all four classes. 
 
 (2) By the action of ammonia on the alkylene oxides : — 
 
 CH2-CH2 + NH;, = HOCH2CH2NH2. 
 
 This reaction gives the primary, secondary, and tertiary bases. 
 
 (3) By Gabriel's phthalimide reaction. Thus potassium phthaUmide and 
 
 ethylene dibromide give CeH4\QQ/N-CH2-CH2-Br, and when this is heated 
 
 with dilute sulphuric acid to 200-220° phthalic acid is split off, and at the 
 same time the bromine is replaced by hydroxyl, giving CH20H-CH2NH2. 
 
 This body CH20H-CH2NH2, oxyethylamine, amino-alcohol, or ethanolamine, 
 is best obtained by the action of ethylene oxide on strong ammonia. It is 
 remarkable'' for its volatility. The reaction-mixture can be separated com- 
 pletely by fractionation into water and the primary, secondary, and tertiary 
 bases : — 
 
 HOCH2CH2NH2, (HOCH2CH2)2NH, (HO0H2-CH2)3N. 
 
 Ethanolamine is a colourless viscous liquid, of strong alkaline reaction, 
 which boils at 171°. It absorbs carbon dioxide from the air, and forms stable 
 salts. It is the mother substance of choline : — 
 
 nw'^^/CH2-CH20H 
 
 C. 06. ii. 471, 472, 932. '' Knorr, Ber. 30. 909 (1897). 
 
Choline, Taurine 33 
 
 a product of the decomposition of lecithine. Lecithine is a compound of 
 choline, glycerine-phosphoric acid, and stearic and palmitic acids, which is 
 widely distributed in the animal and vegetable kingdoms, occurring in the 
 brain, the nerves, in blood corpuscles, and in yolk of egg, as well as in the seeds 
 of many plants. 
 
 The constitution of choline is shown by its synthesis from trimethylamine 
 and ethylene oxide in concentrated aqueous solution : — 
 
 A more complicated substitution-product of ethylamine of some importance 
 is taurine. It is the classical example of the danger of the method which is 
 universally employed, of determining oxygen in organic analysis by difference. 
 It is got fiom taurocholic acid, a substance occurring in various animal 
 secretions. Taurine was discovered by Gmelin in 1824, and was supposed 
 to be a compound of carbon, hydrogen, oxygen, and nitrogen, with the formula 
 C2H7NO5. Twenty-two years later Eedtenbacher found that it contained 
 sulphur, and determined its amount. It then appeared that its formula is 
 C2H7NSO3. But as the atomic weight of sulphur is nearly twice that of 
 oxygen, the proportions of carbon, hydrogen, and nitrogen required by the two 
 formulae are the same within the limits of experimental error. 
 
 The constitutional formula of taurine is : — 
 
 CHa-NHa 
 
 CH2SO3H 
 It is amino-ethyl sulphonic acid, as was shown by Kolbe, who synthesized it by 
 the following method. When alcohol is treated with sulphur trioxide it gives 
 ' carbyl sulphate ', which is converted by cold water into ethionic acid, a 
 sulphonic acid of ethyl sulphuric acid : — 
 
 CH,— O— SOo CH2-OSO2OH 
 
 i + H,0 
 
 . ? 
 
 CH2-SO2-O ' CH2SO2OH 
 
 When ethionic acid is boiled with water, the sulphuric acid group, being 
 
 attached to carbon by the oxygen link, is split off, while the sulphonic group, 
 
 being joined directly through sulphur, remains, and we get oxy-ethyl sulphonic 
 
 acid, or isethionic acid, CHjOH-CHj-SOg-OH. When this is treated with 
 
 phosphorus pentachloride, both hydroxyls are replaced by chlorine, but by the 
 
 action of water on the product the acid chlorine is again replaced by hydroxy!, 
 
 while the other chlorine is unaffected, giving chlorethyl sulphonic acid, 
 
 CH2C1-CH2-S02-OH ; and on heating this with ammonia, the chlorine is 
 
 replaced by NH2, to form amino-ethyl sulphonic acid, which is taurine. 
 
 An aqueous solution of taurine has a neutral reaction, obviously from the 
 
 formation of an internal salt — 
 
 CH2-NH;, 
 
 ? 
 
 CH2-SO2 
 
 In the presence of metallic oxides it behaves as a weak acid and forms salts. 
 It is reconverted by nitrous acid into isethionic acid. 
 
34 Amines 
 
 Amino-Jcetones 
 These bodies' can be made from the phthalimide derivatives of the fatty- 
 acids by acting on their chlorides with benzene and aluminium chloride, 
 and saponifying the products :— 
 
 HO.CO.(CH,)„.N<gg>CeH, -^ Cl-CO-(CH2)„-N<^g>C6H, 
 
 -^ </..CO.(CH,)„.N<co>CeH, -* ./..CO-CCHal.-NH,. 
 
 Their behaviour depends mainly on the distance between the amino-groups and 
 the carbonyl. The a-compounds, such as amino-acetophenone ^-CO-CHg-NHg, 
 are stable only as salts, and when set free oxidize spontaneously in the air 
 to pyrazines: — 
 
 (^H, Jo.0 + = HC, ,J.0 + ^ H,0. 
 
 NH2 
 
 \N^ 
 
 The y8-compounds are stable. The y- and 8- lose water as soon as they are set 
 free, forming tetrahydro-pyrrols and tetrahydro-pyridines respectively : e. g. : — 
 
 0C-NH \ 
 0.COCH3.CH2CH2NH, -^ WXnjj >CH 
 
 But the corresponding formation of a 7-ring compound from e-amino-capro- 
 phenone, 0-CO-(CH2)5-NH2, does not take place. These reactions resemble those 
 of the hydroxy-acids, and suggest that the amino-ketones are really enols, 
 ^■C(OH);CH-(CH2)„-NH2 : a conclusion supported by the fact that many bodies 
 of this class wUl form salts with alkalies, which are in some cases stable in 
 aqueous solution.' 
 
 AMINO-ACIDS 
 
 The carboxyl derivatives of the alkylamines, the amino-acids, are of great 
 importance from their close relationship to various products of organized life. 
 They are found in the juices of plants and in animals, and constitute the 
 greater part of the decomposition-products of albuminoid substances — white 
 of egg, casein, blood-fibrin, silk, gelatine, &c.— whether by decay or by the 
 action of acids or alkalies. It is therefore evident that they form constituent 
 parts of the molecules of these bodies of fundamental physiological importance, 
 the investigation of whose structure is one of the highest aims of organic 
 chemistry. 
 
 For this reason the amino-acids have for a long time attracted a considerable 
 amount of attention; and in particular since 1899 they have been studied 
 in great detail by Emil Fischer, with his customary ingenuity and success. 
 He has attacked the problem both from the analytical and from the synthetical 
 
 1 Gabriel, Ser. 40. 2649 (1907) ; 41. 1127, 2010 (1908). 
 
 2 WiUstatter, Bode, Ser. 33. 411 ; v. Miller, Eohde, ib. 3222 (1900) ; Rabe, Schneider Set- 
 41. 872 (1908). ' 
 
Amino-aeids 35 
 
 side. He has discovered new methods of synthesizing the amino-acids and 
 of building up more complicated molecules from them, and at the same time 
 he has greatly extended our knowledge of the decomposition-products of 
 the proteins. 
 
 For their formation almost any of the regular methods for making amines 
 can be used. The following are the most important: — 
 
 1. Gabriel's phthalimide reaction, using a halogen-substituted ester such 
 as /3-iodo-propionic, which gives /8-amino-propionic acid : — 
 
 C6H4<Qo>NCH2.CH2.COOEt -^ HaNCHa-CH^COOH. 
 
 An important extension of this method has been made by Sorensen,' He 
 treats potassium phthalimide with bromo-malonic ester, giving 
 
 ^e-li^XCO/-'^ "yNCOOEt 
 H 
 
 The remaining hydrogen of the methylene group can then be replaced by 
 sodium, and this, on treatment with alkyl halide, by alkyl. Tor example, 
 with benzyl chloride the benzyl derivative is formed, and this on boiling with 
 aqueous potash has the ester groups saponified and the ring broken, with the 
 formation of the substituted phthalamic acid : — 
 
 /C0\ /COOEt /COOH p^^„ 
 
 CoHiXco/N-CxCOOEt -* CeH,\co.NHC<^ggg ■ 
 
 ^^2-^ CH2-0 
 
 When this substance is heated with concentrated hydrochloric acid the amide 
 is saponified, and at the same time one of the two a-carboxyls is split off, 
 leaving in this case j8-phenyl-a-amino -propionic acid or phenyl-alanine, 
 
 0CH2CHNH2COOH. 
 Since any required alkyl can be introduced into the malonic ester residue, 
 this method enables us to prepare any mono-substituted glycoUic acid. 
 
 2. Fischer has recently ^ developed a still simpler way of making the amino- 
 acids through malonic acid. The required alkyl-malonic acid, E-CH(C00H)2, 
 is brominated, which gives an almost quantitative yield of the mono-brom- 
 derivative K-CBr(C00H)2 . This loses COg on heating, to form the a-brom-acid 
 K-GHBr-COOH, which is converted into the a-amino-acid by ammonia. 
 
 3. From the keto-acids by conversion into oximes or hydrazones and 
 
 reduction :— 
 
 ^ >C=NOH -* >CH-NH2 
 >C=0 
 
 ~-*\C=N-NH0 -^ >CH-NH2 + HgN^ 
 
 4. By the reduction of nitrile-acids (Mendius reaction). An examjsle 
 of the combination of the last two methods is afforded by Fischer's synthesis 
 of a-e-diamino-caproic acid, which is identical with lysine, a product of the 
 decomposition of proteins. Trimethylene chlorobromide, CHgCl-CHg-CHaBr, 
 
 1 C. 03. ii. 33. ^ Fischer, Schmitz, Ber. 39. 351 (1906). 
 
36 Amines 
 
 when treated with potassium cyanide ' under proper conditions yields g-chloro- 
 butyroniti-ile (trimethylene cyanchloride), CHjCl-CHa-CHjCN. This con- 
 denses with sodium-malonic ester ^ to CN-CH2-CH2-CH2.CH-(COOEt)2. This 
 was the body with which Fischer' started. Nitrous acid acts upon it to 
 split off one carboxethyl group and form the isonitroso-compound (oxime), 
 which is saponified and treated with sodium and alcohol. Both the cyanide 
 and the oxime groups are thus reduced, giving a-e-diamino-caproic acid : — 
 CNCH2CH2CH2C(NOH)COOEt -> H2NCH2CH2CH2CH2CH(NH2)COOH. 
 5. The a-amino-acids are got by the Strecker reaction, the action of ammonia 
 on cyanhydrins, or of prussic acid on aldehyde-ammonias :■ — 
 
 The amino-nitriles so formed give amino-acids on hydrolysis. 
 
 Fischer has shown that by starting with a hydroxy-aldehyde it is possible 
 by this reaction to prepare the oxy-amino acids, which also occur among the 
 decomposition-products of proteins. Thus from glycollie aldehyde, CH20H-CH0, 
 he prepared a-amino-/3-oxy-propionic acid,* CH20H-CHNH2-COOH, identical 
 with serine, a product of the hydrolysis of silk. In the same way from arabi- 
 nose,= CH20H(CHOH)3CHO, he got the acid CH20H(CHOH).jCHNH2COOH, 
 which on reduction gives its aldehyde, CH20H-(CHOH)3-CHNH2-CHO,= iden- 
 tical with glucosamine, a body obtained from certain fungi. 
 
 The amino-acids are crystalline substances, generally easily soluble in 
 water, but only slightly in alcohol and ether. If the NH2 is in the a-position 
 they have a sweet taste ' ; in the /3-position this is very slight, in the y it 
 is absent. It is remarkable that their power of acting as foods for microscopic 
 fungi* changes in a similar way, though in the opposite direction. Moulds 
 grow well in solutions of y-amino-acids, less well in /?-, and hardly at all in a-. 
 
 They have singularly high melting-points, glycocoll, for example, melting 
 above 220°. They form salts both with acids and bases. There is reason 
 to think that the free acids exist only in the form of intramolecular salts 
 such as CH2-NH3 
 
 0=0 O 
 
 This structure is indicated by their neutral reaction, their high melting-points, 
 and their insolubility in alcohol and ether. In the case of some of their 
 derivatives an analogous structure can be proved. Thus betaine, which is 
 trimethyl-glycocoU, and may have either of the two formulae 
 
 Mm 
 
 CH2-N(OH3)2 OH2-N-CH3 
 I or i^^CHo, 
 
 COOCH3 0=0 o 
 
 1 Gabriel, Ber. 23. 1771 (1890). Blank, Ser. 25. 3041 (1892). 
 
 3 C. 02. i. 985. i C. 02. i. 762. 
 
 •'■ Fischer, Leuchs, Ber. 35. 3787 (1902). = Ber. 36. 24 (1903). 
 
 ' Ber. 35. 2660 (1902). « Bm: 35. 2289. 
 
Amino-acids 37 
 
 is shown to have the second formula by the ease with which it splits off 
 trimethylamine. 
 
 When heated with barium oxide the amino-acids split off COj and form 
 amines:— CHaNH^COOH = CO2 + CH3NH2. 
 
 When the acids are treated with nitrous acid, the normal reaction of a 
 primary alkylamine takes place, and the NHg is replaced by hydroxyl. But 
 if their esters are treated with nitrous acid, by acting on the ester-hydro- 
 chloride with potassium nitrite, a so-called fatty diazo-compound is produced, 
 which is not a true diazo-compound at all, but contains a ring of two nitrogen 
 atoms and one carbon. Thus glycocoll yields diazo-acetic ester :- — 
 
 CHa-NH^ + 0=NOH HCk(j^ 
 COOC2H5 COOC2H5 
 
 -I- 2 H2O. 
 
 They are easily reduced by sodium amalgam to amino-aldehydes, the amino- 
 group apparently promoting the reduction in the same way as the hydroxyl 
 group in hydroxy-acids.' 
 
 As an example of the properties of an amino-acid we may take those of 
 the simplest member of the series, glycocoll, glycine, or amino-acetic acid, 
 CH2NH2-COOH. It was originally obtained by hydrolysing glue with baryta 
 or dilute sulphuric acid, whence the name (yXuKu's sweet and KoXKa glue). It 
 occurs in nature as such, in mussels, more often in the form of derivatives, 
 such as betaine, in sugar-beet and other plants, and hippuric acid, in the urine 
 of herbivora. It is best obtained from hippuric acid (benzoyl-glycocoll), by 
 hydrolysing it with dilute acid to benzoic acid and glycocoll: — 
 
 ^•C0NHCH2C00H -f- H2O = 0COOH + NH2CH2COOH. 
 
 It may also be obtained from chloracetic acid and ammonia, or, by Strecker's 
 reaction, from formaldehyde, prussic acid, and ammonia : — 
 
 _T /H /H heated /S. 
 
 HC<^ + HCN -^ HC-OH -> HC-NH2 > HC-NH2 . 
 
 ^ \CN (NH3) \CN with BaO \COOH 
 
 It is also formed from cyanogen by the action of hydriodic acid, a peculiar 
 reaction in which one CN group is reduced while the other is saponified : — 
 
 C=N CH2NH2 
 
 1^^-.4HI.2H20 = V^;^ ^.2l2-.NH3. 
 
 Olyeocoll is a crystalline substance with a sweet taste. It turns brown at 228° 
 and melts at 232-236°— nearly 220° higher than acetic acid. 
 
 If treated with alcohol and hydrochloric acid it is converted into its ethyl 
 ester hydrochloride, CH2NH3ClCOOEt. This is a crystalline body melting 
 at 144°, which can be sublimed by cautious heating. If it is treated with 
 potassium nitrite in aqueous solution, diazoacetic ester, CHNa-COOEt, is 
 precipitated as a yeUow oil: this was the first known fatty diazo-compound, 
 discovered by Curtius^ in 1883. The intermediate nitrite of glycocoll ester, 
 CH2(NH30NO)COOEt, can be obtained by treating the hydrochloride suspended 
 
 1 E. Fischer, Ber. 41. 1019 (1908) ; of. Neuberg, ib. 956. " Ber. 16. 2230. 
 
 1175 D 
 
38 Amines 
 
 in dry ether vdth silver nitrite. It is unstable, and readily loses water to 
 give diazoacetic ester. 
 
 The hydrochloride of the ester when distilled with sodium carbonate 
 undergoes a remarkable intramolecular change, giving propylamine: — 
 
 COOC2H5 C2H5 ' 
 
 This is analogous to the formation of stilbene on heating phenyl fumarate 
 or phenyl cinnamate * : — 
 
 CH.COO0 CH-0 CH-0 
 
 ^H.CO.O^= ^^^ - ^H.CO.O0 = ^^«^^^H.0- 
 
 Free glycocoU ester is got from the hydrochloride by treatment with silver 
 
 oxide, or more conveniently, as Fischer has shown, with strong aqueous soda at 
 
 a low temperature. It is an unstable liquid, B. Pt. 148-9°, with a strong basic 
 
 reaction. It decomposes spontaneously, and is saponified by water. 
 
 TJie polypeptides and tJie structure of tlie Proteins ^ 
 
 The main interest of the amino-acids arises from their relationship to the 
 proteins. The proteins are an extremely important class of substances derived 
 from living matter, and forming the chief constituents of protoplasm, as well as 
 other animal and vegetable substances. They are very numerous, and are 
 obtained from the most various sources ; they include such bodies as gelatine, 
 white of egg, casein (from milk), globulin (from blood serum), whUe others are 
 got from horn, silk, and so forth. They are all decomposed by boiling with 
 acids and in other ways, and by far the most important products of their 
 decomposition are in aU cases certain amino-acids. The first step towards the 
 investigation of their structure is the separation and identification of the 
 products of their hydrolysis. This is a work of unusual difficulty. An 
 excessively complicated mixture of substances is obtained in every case (thus 
 gelatine gives glycocoU, alanine, pyrrolidine-carboxyUc acid, leucine, aspartic 
 acid, glutamic acid, phenyl-alanine, amino-valeric and amino-butyric acids, 
 together with oxy-amino- and diamino-acids and other substances). It is 
 obviously not easy to separate and identify the constituents of such a mixture, 
 many of them only occurring in very small quantity : and the difficulty is 
 increased by 'the properties of the amino-acids, which are insoluble in ether, 
 are soluble both in acids and alkaHes, and have no definite melting-points. 
 E. Fischer has, however, devised a method by which they can be separated with 
 comparative ease, by making use of the fact that the esters of the amino-acids 
 are volatile (at any rate under greatly reduced pressure) without decomposition, 
 and that their boiling-points lie far apart. As this method has been of the 
 utmost service we may consider in detail the separation of the products of the 
 
 • Anschiitz, Ber. 18. 1945 (1885). 
 
 '' See the numerous papers of E. Fischer and iiis pupils in the Berichte from 1899 onwards, and 
 in the Zeitschr. f. physiologiiche Chemie for the same years. Only the more important references 
 are given in the text. Fischer has summed up the work up to the end of 1905 in a lecture, Ber. 39. 
 530. Cf. also Kossel, Ber. 34. 3214 (1901), and Cohen, Organic Chemistry (1907), chap. xi. 
 
Polypeptides : Structure of the Proteins 89 
 
 decomposition of casein, which is typical of the other bodies investigated. 
 Casein is the protein of milk, and is made by precipitating milk with hydro- 
 chloric acid. 
 
 A kilogram^ of casein was hydrolysed by prolonged boiling with pure 
 concentrated hydrochloric acid. From the product the glutamic acid (a-amino- 
 glutaric acid, COOHCHNHa-CHa-CHa-COOH) was separated as the slightly 
 soluble hydrochloride. The residue, containing the amino-acids, was evaporated 
 down, stirred up with absolute alcohol, and saturated with gaseous hydrochloric 
 acid. This converts the amino-acids into their ethyl esters, which of course 
 combine with the hydrochloric acid (owing to the presence of the NHg) to form 
 hydrochlorides. The product was then treated with excess of strong caustic 
 soda to remove the hydrochloric acid and liberate the free esters. By keeping 
 the temperature low saponification of the esters is avoided. The liquid was 
 saturated with potassium carbonate to diminish the solubility of the esters, 
 which can then be extracted with ether. The mixed esters were then frac- 
 tionated under 8-15 mm. pressure, and thus separated into 8 fractions, boiling 
 from 40° to 160°, of which the largest (165 gr. from 1 kgr. casein) boiled at 
 80-85°. For the further separation of the esters in each fraction it was 
 saponified with baryta, and the amino-acids separated by means of their copper 
 salts and in other ways. 
 
 This method has been applied with excellent results to the hydrolysis of 
 a large number of similar substances, such as silk, white of egg, horn, 
 gelatine, &c. In some cases it is necessary to conduct the distUlation of the 
 mixed esters at pressures not exceeding 1 mm. For this purpose the ordinary 
 water-pump is quite useless, as it will not produce a higher vacuum than 10 to 
 15 mm. ; and even a mercury-pump is not much good, because in such cases of 
 the distillation of a mixture of substances, where there is always a certain 
 amount of decomposition, the gases and volatile bodies so produced tend to 
 increase the pressure. To avoid these difficulties Fischer ^ employs a powerful 
 air-pump, such as is used for exhausting the globes of incandescent lamps, and 
 interposes between the pump and the receiver a condenser cooled with liquid 
 air, by which the more volatile decomposition-products are prevented from 
 getting into the pump. By these means he is able to conduct the distillation 
 under a pressure of a third of a mm. or less. The importance of this becomes 
 evident when one considers that Krafft has found that a diminution of pressure 
 from 15 mm. (the lowest that a water-pump ordinarily gives) to the vacuum 
 of the cathode light (^V to ^^ mm.) lowers the boiling-point by 70-100 degrees. 
 
 A further difficulty in identifying the amino-acids obtained in this way from 
 natural sources arises from the fact that they nearly all contain asymmetric 
 carbon and occur in the active forms, whereas the synthetic substances with 
 which they are compared are of course inactive, and are usually racemic 
 compounds, so that their physical constants are difierent from those of the 
 active modifications. For a satisfactory identification it is necessary to split 
 the racemic form into its active components. This is not easy, because the 
 moulds, as we have seen, do not as a rule grow easily in the solutions of these 
 
 1 C. 01. ii. 691. " Ber. 35. 2158 (1902). 
 
 d2 
 
40 Amines 
 
 acids, and their feebly basic and feebly acidic character makes it difBcult to form 
 their salts either with active acids or with active bases. Fischer finds ' that by 
 treating the amino-acids in alkaline solution with benzoyl chloride (Schotten- 
 Baumann reaction) they give the much more acidic benzoyl derivatives (e. g. 
 ^•CONHCHjCOOH) which form stable salts with active bases such as brucine, 
 which can then be separated into their active forms by fractional crystallization. 
 The corresponding formyl derivatives '' are even better for this purpose, as they 
 are more easily produced (merely by boiling the amino-acids with anhydrous 
 formic acid) and are more easily saponified again after separation, so that there 
 is no danger of racemization. 
 
 All the proteins give very similar products on hydrolysis. The most 
 important are the following: — 
 
 Monamino-acids. 
 
 Glycine or glycocoU, a-amino-acetic acid, CHaNHj-COOH. 
 
 Alanine, a-amino-propionic acid, CH3-CH2NH2-COOH. 
 
 Phenyl alanine, a-amino-/3-phenyl-propionic acid, ^CHa-CHNHjCOOH. 
 
 Valine, a-amino-isovaleric acid, (CH3)2CHCHNH2COOH. 
 
 Leucine, a-amino-isocaproic acid, (CH3)2CH-CH2-CHNH2-COOH. 
 
 Aspartic acid, aminosuccinic acid, COOH-CH2CHNH2-COOH. 
 
 Glutamic acid, a-amino-glutaric, COOHCH2CH2CHNH2COOH. 
 
 Hydroxy-amino-acids. 
 
 Serine, a-amino-/3-hydroxy-propionic acid, CH20H-CHNH2-COOH. 
 Isoserine, /3-amino-a-hydroxy-propionic acid, CH2NH2-CHOH-COOH. 
 TjTosine, a-amino-jJ-oxyphenyl-propionie acid, HOC6H4-CH2-CHNH2-COOH. 
 
 Diamino-acids. 
 
 Ornithine, a-6-diamino-valeric acid, CH2NH2(CH2)2CHNH2COOH. 
 Lysine, a-e-diamino-caproic acid, CH2NH2(CH2)3CHNH2-COOH. 
 Arginine, a-amino-6-guanido-valeric acid, 
 
 ^^)>CNHCH2(CH2)2CHNH2COOH. 
 
 Heterocyclic compounds. 
 ■H2C — CII2 
 Proline, a-pyrrolidine-carboxylie acid, tt X CH-COOIJ ' 
 
 HOHC-CH2 
 
 Oxyproline, hydroxy-pyrrolidine-carboxylic acid, „ X l^-o- -^/-./-.tt • 
 
 iloO\^L>rl-OOOH 
 NH 
 
 
 CCHaCHNH^COOH 
 
 NH 
 
 ■ Ser. 32. 2461, 3638 (1899); 33. 2370 (1900). 
 ' Fischer, Warburg, Ber. 38. 3997 (1905). 
 
Polypeptides: Structure of the Proteins 41 
 
 There are also one or two sulphur compounds, such as cystine, a-amino- 
 /3-thiolactie acid, CHaSHCHNHa-COOH, which is the sulphur analogue of 
 serine. The occurrence of the two pyrrolidine derivatives is remarkable, 
 especially as they are found to be present in almost all proteins. It is possible 
 that they are only produced by secondary reactions, as they have an obvious 
 relation to amino-acids : thus proline would be produced by the loss of water 
 from a-amino-6-oxy-valeric acid, (^jj -CH-COOH 
 
 I ' ^NH, , 
 
 CHa-CHjOH 
 
 but there is reason to think that this is not the case. It is especially to be 
 noticed that with the single exception of isoserine, all these bodies are a-amino- 
 acids, or are simply derived from them. 
 
 Since the proteins are so readily split up into these acids, it is natural to 
 expect that we may be able to build up the proteins from them. It is to this 
 end that the second part of Fischer's work is directed. We can hardly expect 
 to be able at first to form the proteins themselves ; for they are the most 
 complicated compounds of the whole group, and have molecular weights which 
 are supposed to be about 15,000, and are certainly enormous. We should 
 rather try to meet the synthesis half-way, so to speak, and to build up some of 
 the simpler earlier decomposition-products of the proteins. Such bodies are 
 the peptones. They are the simplest products of the hydrolysis of the proteins 
 by the gastric juice and certain other enzymes. They resemble the proteins in 
 general behaviour, but are very soluble in water, acids, and alkalies, and are 
 not coagulated by the ordinary methods. They have molecular weights of 
 about 600 — that is to say high, but not excessively so. They have an obviously 
 close relationship to the proteins from which they are produced, and to the 
 amino-acids into which they are converted by further hydrolysis. The problem 
 which Fischer set himself was the formation of bodies resembling the peptones 
 from the a-amino acids. Now it has long been known ' that these acids can be 
 converted into complex substances of high molecular weight. Thus glycocoU 
 when heated with glycerine gives a so-called anhydride, resembling horn : and 
 there are many other such cases. But the products are always amorphous 
 bodies of indefinite character : the ' brutal ' reactions, as Fischer calls them, by 
 which they are produced throw no light on their structure : and nothing 
 is known as to their relationship to the natural proteins. It is clear that the 
 only way of arriving at certain results is to discover methods by which the 
 molecules of various amino-acids can be linked up together in successive 
 stages to definite chemical individuals, whose structure can be satisfactorily 
 determined. 
 
 As to the nature of the linkage there can be little doubt. It must be 
 formed by loss of water, and it must be easily broken up again by hydrolysis, 
 under the influence of mineral acids. This indicates that the bodies are of the 
 nature of amides. Two molecules of amino-acid can obviously form an amide, 
 which will still have a carboxyl group at one end, and an NHg at the other ; so 
 that the product can again form an amide with itself, or with another molecule 
 
 ' Cf. Fischer, Ber. 34. 2868 (1901). 
 
42 Amines 
 
 of amino-acid. In fact, as far as the formulae go, there is no limit to the 
 
 process ; molecules of any size can be built up, with the general formula 
 
 E-CHNH2C0(NHCHEC0)^-NHCHRC00H. To bodies of this type Fischer has 
 
 given the name of polypeptides. He has worked throughout on the idea, which 
 
 all his discoveries have confirmed, that the natural peptones are really mixtures 
 
 of polypeptides. He has elaborated a series of methods for their synthesis, of 
 
 which the most important are the following. 
 
 The a-amino-acids, and still more easily their esters, can be made to give 
 
 bimolecular anhydrides analogous to glycide, the anhydride of glyoollic acid. 
 
 Thus glycine yields glycine anhydride, which has been shown to be a piperazine 
 
 NHCHg.CO 
 derivative, diketo-piperazine, of the formula I I . It is, in fact, an 
 
 ^^ COCH2NH 
 
 intramolecular double amide. If this is treated with hydrochloric acid (or 
 
 more conveniently with alkali'), it takes up water and the ring is broken, with 
 
 the formation of the simple anhydride, NHj-CHa-CO-NH-CHa-COOH, which 
 
 may be called glycyl-glycine, and is the simplest polypeptide. In this way the 
 
 dipeptide of an amino-acid can be made, but for the further lengthening of 
 
 the chain other methods have to be used. One of these is to treat the dipeptide 
 
 with a-chloracyl chloride, which can be obtained by the Zelinsky method of 
 
 chlorinating (or brominating) the acid in the presence of phosphorus. This 
 
 adds on the chloracyl group to the NHj. For instance, glycyl-glycine with 
 
 chloracetyl chloride gives CHaClCONHCHs-CO-NHCHaCOOH, and when 
 
 this is treated with ammonia the chlorine is replaced by NHg, giving diglycyl- 
 
 glycine, NH2CH2CONHCH2CONHCH2COOH. This can then be combined 
 
 with chloracetyl chloride again, and a fourth glycyl group introduced, 
 
 and so on. 
 
 Fischer has subsequently improved this method by the discovery^ that 
 the amino-acids themselves can be converted into their acid chlorides, 
 E-CH(NH3C1)C0C1, by the action of phosphorus pentachloride in acetyl chloride 
 solution. This reaction cannot be applied to the polypeptides themselves, but 
 their chlor- or bromacyl derivatives, like the chloracetyl derivative of glycyl- 
 glycine mentioned above, give acid chlorides in this way, and these of course 
 condense with the NH^ of a polypeptide, whereby the process of building up 
 long chains is considerably simplified. 
 
 For the sake of simplicity, the glycine compounds have been used as 
 examples of these synthetical methods ; but Fischer's work has been extended 
 to a large number of other naturally occurring amino-acids, and in this way 
 he has produced more than a hundred polypeptides, of the most various kinds. 
 The largest hitherto prepared is an octodeca-peptide, containing eighteen amino- 
 acid residues.' Its name is 1-leucyl-triglycyl-l-leucyl-triglycyl-l-leucyl-octo- 
 glycyl-glycine, its formula is 
 
 NH2CH(C,H9)CO-(NHCH2CO)3.NH-CH(C4H9)CO(NHCH2-CO)3NH.CH(C,H9)GO 
 (NHCH2-CO)8NH-CH2COOH, (C^HgoO^Nis), -, 
 
 and its molecular weight 1213. It is the compound of largest molecular 
 
 1 Ser. 38. 607 (1905). 2 ^^^ 33. 2917 (1905). 
 
 ' £er. 40. 489 (1907). 
 
Polypeptides : Structure of the Proteins 43 
 
 weight known, whose constitution is definitely ascertained. It is soluble in 
 100 parts of water, and very easily in strong acids. 
 
 If Fischer's view as to the structure of the peptones is correct, these higher 
 polypeptides ought to resemble them in properties : and to a very great extent 
 this is the case. The typical reactions of the peptones are the biuret reaction, 
 the formation of a red colour with potash and copper sulphate : the precipitation 
 by phosphotungstic acid: and the hydrolysis to amino-acids by trypsin, the 
 pancreatic ferment. All these reactions are given by many of the higher 
 polypeptides. At the same time we must remember how extremely compli- 
 cated the question is. It is probable that none of the natural peptones are 
 single chemical individuals: many of them are no doubt mixtures of a great 
 many compounds. But apart from this, the pure peptones must each be made 
 up of ten or more residues of several different amino-acids, and we do not yet 
 know at all exactly the proportions of these different acids which go to make 
 up their structure, still less the order in which they are put together: and 
 yet these two factors must have a profound influence on. the properties of the 
 product. It is to be hoped that by the study of a large number of polypeptides 
 we may get some guiding principles as to the relations between their structure 
 and their properties : but this must be a matter of time. Some indications 
 have already appeared. Fischer and Abderhalden have shown that the 
 susceptibility to ferments is much greater if the bodies are composed of the 
 naturally occurring stereo-forms of the amino-acids. It also seems that 
 the solubility, which is generally less with the artificial polypeptides than with 
 the natural peptones, is increased with the use of the active modifications, 
 with the employment of a variety of different amino-acids, and after reaching 
 a minimum with an increasing length of chain, begins to increase again. 
 
 One of the most hopeful signs for the success of this investigation is the 
 recent discovery that polypeptides identical with those prepared synthetically 
 occur among the products of the hydrolysis of the natural proteins under 
 suitable conditions. Thus fibroin ' (from silk) gives glycyl-d-alanine, as well 
 as a tetra-peptide, composed of two glycine molecules, one of d-alanine, and one 
 of I-tyrosine ; and there are several other similar cases.^ This makes it 
 practically certain that structures of the polypeptide type form part of the 
 protein molecule; though it is not proved that they are wholly built up on 
 this type. Fischer suggests that the piperazine ring (of the double anhydrides) 
 may be present, and also that the hydroxyl groups of the oxy-amino-acids may 
 play a part in the formation of the molecule. All these questions can only be 
 settled by much further investigation. 
 
 BENZYLAMINE BASES 
 
 This group consists of those aromatic derivatives which have the amino- 
 group in the side chain. They are prepared by the same methods as the fatty 
 amines : by the action of ammonia on the halides, by the saponification of 
 
 ' Abderhalden, Ber. 39. 752, 2315 (1906) ; C. 07. ii. 1533. 
 
 ' Cf. Hougounenq, Morel, Bull. Soc. [4] 3. 1146 ; C. B. 148. 236 (1909) ; Abderhalden, C. 
 09. i. 1246 : ii. 1754. 
 
44 Amines 
 
 isocyanates, by Gabriel's phtbalimide reaction, &c. They may also be got 
 by the reduction of the nitriles, which is often most conveniently carried out 
 by combining them with hydrogen sulphide to the thio-amides, and then 
 reducing these, which gives the amine and hydrogen sulphide : — 
 ^■CN -* ^-CS-NHa -> ^-CHa-NHa + HgS. 
 
 In general properties they closely resemble the alkylamines. They have a 
 smell Kke that of ammonia, and are easily soluble in water. The aqueous 
 solution has an alkaline reaction and absorbs carbon dioxide from the air. Like 
 the alkylamines they give alcohols with nitrous acid, and not diazo-compounds, 
 and with carbon bisulphide they form dithiocarbamates. 
 
 The acidic phenyl group exerts an influence on their basicity, though from 
 its greater distance much less than in aniline. Thus the (apparent) strength 
 of benzylamine is about equal to that of ammonia, that is to say much less 
 than that of methylamine, but more than that of aniline. 
 
CHAPTER HI 
 
 AROMATIC AMINES 
 
 The aromatic amines properly so called are those bodies in which the 
 amino-group is directly attached to the benzene nucleus. They are formed by 
 the reduction of nitro-compounds, and as the latter are easily obtained by direct 
 nitration, the aromatic amines are much more accessible than the fatty, and in 
 consequence are far better known. Indeed there is scarcely any class of bodies 
 which has been investigated with so much industry and so much success. 
 This work has of course been stimulated in no small degree by the value of 
 the aniline derivatives as dyes. 
 
 The basis of our knowledge of aniline and its allies was laid down by 
 Hofmann in 1846-1851 in his ' Contributions to our knowledge of the Volatile 
 Organic Bases '.^ ' Considering the vei-y modest means which were then at the 
 disposal of the organic chemist, and moreover that he was not able to obtain 
 benzene, much less aniline, as a commercial product, it is astounding what 
 an immense number of facts Hofmann discovered which are still amongst the 
 most important in the whole subject.' '^ 
 
 The simplest of the aromatic amines, aniline, is also the most important. 
 Thousands of kilograms of it are made in the great German factories every 
 day. The annual production of aniline and similar bases in 1892 was 8,000 tons. 
 
 Aniline was first obtained by Unverdorben in 1826 by the distillation of 
 indigo. In 1840 Fritzsche prepared it by distilling indigo with potash, and 
 called it aniline from anil, the Spanish word for indigo. In 1843 Hofmann 
 showed that this base obtained from indigo was identical with that which 
 Eunge in 1834 had found in coal tar, and with that which Zinin ^ had obtained 
 in 1842 by the reduction of nitrobenzene. This last method of preparation 
 proves its formula. 
 
 Aniline only occurs in very small quantity in coal tar, so this source, 
 which was at first adopted, was very soon abandoned in favour of the 
 reduction of nitrobenzene, which is now the sole method of preparation 
 employed. 
 
 As reducing agents several substances have been adopted. Zinin used an 
 alcoholic solution of ammonium sulphide, which is still used in certain cases, 
 where we want to reduce one only of several nitro-groups. The reducing agent 
 used in commerce is iron in presence of hydrochloric, or in more recent times 
 of acetic acid. It is not necessary to take more than 5 per cent, of the 
 calculated quantity of acid in either case. With hydrochloric acid it appears 
 that ferrous chloride is formed, which then acts as a hydrogen carrier, so to 
 speak, the rest of the iron being converted into ferric hydrate. With acetic 
 
 ' Ann. 57-79. ^ Meyer, Jacobson, Leitrh. ii. 172. " Ami. 44. 286. 
 
46 Aromatic Amines 
 
 acid the iron forms ferric acetate, and the hydrogen evolved reduces some of 
 
 the nitrobenzene to aniline ; but at the high temperature employed (the 
 
 mixture is heated with steam) the salt breaks up into ferric hydrate which 
 
 is precipitated, and acetic acid which serves to dissolve more iron. In either 
 
 case the reaction may be represented as consisting only in the oxidation of the 
 
 iron to ferric hydrate and the simultaneous reduction of the nitrobenzene to 
 
 aniline : — . „ 
 
 2 Fe + 0NO2 + 4 H2O = 2 Fe(0H)3 + ^NH,. 
 
 When the reaction is finished, the liquid is neutralized with lime and 
 
 distilled with steam. The aniline, which forms the lower layer of the 
 
 distillate, is run off, and the upper layer, consisting of an aqueous solution 
 
 of aniline, is used to feed the boilers which supply the steam for further 
 
 distillations. 
 
 In the laboratory the reduction is generally done with tin or stannous 
 chloride, in presence of excess of strong hydrochloric acid. 
 
 Other methods for preparing aromatic amines are: — 
 
 (2) Heating the halogen derivatives of the aromatic hydrocarbons with the 
 compound of calcium chloride and ammonia. This requires a very high 
 temperature (360-370°), and gives a bad yield, owing to the firmness with 
 which the halogens are attached to the benzene nucleus.* If there are nitro- 
 groups on the same nucleus in the ortho or para position to the halogen, the 
 reaction goes much more easily. Nitro-groups in the meta position have not 
 this effect. 
 
 (3) By heating the phenols with the compound of zinc chloride and 
 ammonia or with ammonium chloride to above 300°. In the case of certain 
 negatively substituted naphthols it is found that their ethers react quite easily 
 at moderate temperatures with alkylamines to form substituted naphthyla- 
 mines ^ : for example : — 
 
 CioHgNOaBr-OEt + EtNHj = CioHsNOgBrNHEt + HOEt. 
 This resembles the formation of amides from the esters and the amines, and is 
 an example of the analogy between the phenols and naphthols and the acids. 
 
 Freshly prepared aniline is a colourless oily liquid, which soon turns 
 reddish brown when exposed to the air: but this depends on the formation 
 of very small quantities of oxidation-products, which may be removed by 
 treatment with acetone, with which they combine.'' If perfectly pure, aniline 
 remains colourless even when exposed to the air for a long time. It melts 
 at - 8°, and boils at 183-184°. At 25° it dissolves 5 per cent, of water : while 
 water at 16° dissolves 3 per cent, of aniline. It is easily soluble in the aqueous 
 solution of its hydrochloride. 
 
 Many of its properties are identical with those of the alkylamines : but 
 in many important points its behaviour is different. In particular its basicity 
 is weakened by the acidic character of the phenyl group, which is also shown, 
 for example, in the comparison of phenol and alcohol. It dissolves in mineral 
 
 ' A lower temperature is sufficient, and a much better yield obtained, if the haloid compound 
 ia heated with ammonia under pressure, in presence of a copper salt. C. 08. ii. 1221 ; 09. i. 
 475, 600. 2 Meldola, J. C. S. 1906. 1434. 
 
 3 Cf. Hantzsoh, Freese, Ber. 27. 2966 (1894). 
 
Properties of Aromatic Amines 47 
 
 acids (and also in organic acids) to form stable salts such as i^NHgCl. But 
 in contrast to the fatty amines it has no alkaline reaction to litmus or phenol- 
 phthalein : it cannot absorb carbon dioxide : and its mineral acid salts have 
 an acid reaction. It is, however, sufSciently basic to precipitate the hydroxides 
 of zinc, aluminium, and ferric iron from solutions of their salts. 
 
 It may be detected by giving a deep bluish violet with bleaching powder, 
 and a fine blue colour with potassium bichromate in presence of strong 
 sulphuric acid. 
 
 The aromatic amines, like the fatty, can have the hydrogen attached to 
 nitrogen replaced by alkali metals, and by alkyl, aryl, or acyl groups. Sodium 
 and potassium dissolve in hot aniline with evolution of hydrogen, to give 
 compounds like 0-NHK and ^-NKg. Other such derivatives will be described 
 later. 
 
 One of the most characteristic differences between the aromatic and the 
 fatty amines is that the aromatic with nitrous acid give the diazo-compounds : — 
 
 ^•NHa + ONOH = HgO + ^-NgOH. 
 The diazo-compounds are of the utmost importance, both theoretical and 
 practical. They are capable of more varied reactions than almost any other 
 class of bodies, and they form an intermediate stage in the preparation of 
 by far the greater number of aniline dyes. Moreover they have given rise 
 to a series of investigations which have been of the highest value in extending 
 our knowledge of the phenomena of tautomerism. 
 
 The characteristic reaction of the primary alkylamines with nitrous acid 
 (giving nitrogen and the alcohol) occurs also with the aromatic amines, the 
 diazo-compounds breaking on warming with the greatest ease to give phenols, 
 with evolution of nitrogen. 
 
 This formation of diazo-compounds gives us a means of determining 
 nitrous acid volumetrically.' The nitrous acid solution is titrated against 
 aniline hydrochloride solution at — 10°, potassium iodide and starch being 
 used as indicator. As long as aniline is present, the nitrous acid is all used 
 up in diazotizing it ; but as soon as it is gone, the iodine is liberated and 
 colours the starch blue. The strength of the aniline solution is determined by 
 titrating it under the same conditions against potassium nitrite. 
 
 Another reaction characteristic of the aromatic amines is with carbon 
 bisulphide. The alkylamines react with it very readily in the cold to give 
 dithiocarbamie salts: — jj 
 
 CHoNH, 0=8 CH3N<^^_c, 
 
 CH3NH2 S CH3NH3-S/ 
 
 The aromatic amines do not as a rule react in the cold at all. They do so, 
 however, readily on heating, but then they split off hydrogen sulphide, and 
 form disubstituted thioureas : — 
 
 ^ S + s=c=s = ^ ;c=s + H^s, 
 
 1 Vignon, Bay, C. B. 135. 507 ; C. 02. ii. 1094. 
 
48 Aromatic Amines 
 
 the product in this case being generally known as thiocarbanilide. The 
 reaction is greatly hastened by dissolving a small quantity of sulphur in the 
 mixture : but how this acts is not known. 
 
 The aromatic amines have a curious reaction with sulphurous acid.' If 
 they are treated with sodium hydrogen sulphite ammonia is expelled, and 
 an acid sulphite (not sulphonate) is formed : — 
 
 E-NHg + NaHSOg = K-O-SOgNa + NH3. 
 These acid sulphites are readily saponified (e.g. by alkalies) to the corre- 
 sponding phenols, which shows that the sulphur is attached to carbon through 
 oxygen. Conversely, the phenols are converted into amines by treatment with 
 ammonium sulphite and ammonia, no doubt with intermediate formation of the 
 sulphite esters, so that the reaction is reversible throughout : — 
 Bisulphite Alkali 
 
 Amine * Ester Phenol. 
 
 ■< < 
 
 NH3 NaHSOg 
 
 Fatty amines and alcohols do not give these reactions at all ; and even the 
 benzene derivatives do so much less readily than those of naphthalene. 
 
 Aniline is very susceptible to oxidation and gives various products, which 
 will be discussed later, in dealing with the general question of the oxidation 
 of amines. 
 
 On the other hand it resists reducing agents strongly, as is shown by its 
 ready formation from nitrobenzene. It cannot be converted into its hexahydro- 
 compound, though this body exists and can be prepared by indirect methods. 
 
 SUBSTITUTED ANILINES 
 
 The very numerous substitution-products of aniline fall into two main 
 divisions : — 
 
 A. Those in which the hydrogen of the NHj is replaced : these again are 
 divided into 
 
 1. The alkyl substitution-products (mixed amines). 
 
 2. The aryl substitution-products (purely aromatic amines). 
 
 3. (The acyl-anilines : these are amides, and as such will be considered 
 
 later.) 
 
 B. Those in which the hydrogen attached to the nucleus is replaced. 
 
 A. 1. The mixed alkyl-aryl-amines may be secondary, tertiary, or quater- 
 nary. They are made as in the fatty series by heating aniline with the alkyl 
 halides : commercially by heating it in an autoclave with the necessary alcohol 
 and hydrochloric or sulphuric acid to 180-200°, when the nascent alkyl ester 
 reacts. In this way the tertiary mixed amines, such as dimethyl-aniline, can 
 be got in the pure state, but not the secondary, as even if the calculated 
 quantities are used, some tertiary amine is formed, and some aniline is not 
 acted on. Hence to isolate the secondary compound some method of separation 
 must be adopted. The simplest is by means of nitrous acid. The mixed bases 
 are treated at a low temperature in hydrochloric acid solution with sodium 
 
 ' Bucheior, J.pr. Oh. [2].69. 49 (0. 04. i. 811). 
 
Mixed Amines 49 
 
 nitrite. The primary base gives a diazonium chloride, which is soluble in the 
 acid liquid : the tertiary gives the hydrochloride of the nitroso-dorivative, 
 ON-CeH^-NEaHCl, v?hich is also soluble. The secondary gives, as an alkyla- 
 mine would, a true nitrosamine, ^-NR-NO, which, being no longer basic, is 
 insoluble. This is separated and reduced, whereby the nitroso-group is 
 removed, and the secondary amine (^NH-E is re-generated. 
 
 Owing to this difficulty it is often better to use an indirect method of 
 preparation for the secondary amines. Thus the sodium derivative of aceta- 
 nilide has the sodium replaced by alkyl on treatment with alkyl halide, and 
 the acetyl group can be split off from the product by boiling with potash : — 
 
 i>-<m^^' - <^-N<A?k^^^ - <t><Mk + CH3.C00H. 
 
 There are various other methods of alkylation, among which the introduction 
 of methyl groups by means of formaldehyde may be mentioned. Formaldehyde 
 combines^ with aniline to give the so-called anhydro-formaldehyde-aniline, 
 (}>-'N—CH.2, which is reduced by tin and hydrochloric acid to mono-methyl- 
 aniline, ^-NH-CHg. This again combines with formaldehyde in hydrochloric 
 acid solution to give 0-N{CH3)CH2Cl, which is converted by reduction into 
 dimethyl-aniline. In some cases '' it is sufficient to heat the hydrochloride of 
 the base with formaldehyde alone to 120-160° under pressure ; the aldehyde 
 acts as its own reducing agent, being partly oxidized to carbon dioxide and 
 partly reduced to methyl alcohol. (Compare the analogous behaviour of 
 benzaldehyde in presence of alkali, giving its acid and its alcohol.) 
 
 The secondary and tertiary mixed amines are oily liquids, which distil 
 unchanged. They have no alkaline reaction, as the influence of the phenyl 
 group still outweighs that of the aJkyl. They give salts with one equivalent 
 of acid, and have many of the properties of the purely fatty compounds. 
 
 Many tertiary mixed amines differ from the tertiary alkylamines in reacting 
 with extraordinary ease with nitrous acid. They have of course no hydrogen 
 on the nitrogen to go out with the hydroxyl of the nitrous acid : but the 
 presence of the dialkyl-amino-group renders the para hydrogen on the benzene 
 nucleus sufficiently mobile to react. 
 
 (CH3)2]SrCI>H + HONO = H2O + {CH3)2NCI>NO. 
 
 There are many other instances in which this para hydrogen atom shows 
 
 itself to be extremely reactive : it can combine with many bodies which do 
 
 not react with benzene itself. Thus with benzaldehyde it gives malachite green 
 
 ^•CHO + 2 HCI>N(CH3)2 = 0-CH[-ON(CH3)2]2 -f H^O. 
 
 This reaction can be made to take place with benzene, but less easUy. 
 
 With carbonyl chloride it gives a ketone (Michler's ketone) 
 
 CO-Cla + 2 H-C3-N(CH3)2 = CO[-C>N(CH3)2]2 + 2 HCl. 
 In the same way it will condense ' through the para hydrogen with oxalic ester 
 and with ketones. 
 
 The hydrochlorides of the mixed amines undergo a remarkable change on 
 
 ' C. Goldsclimidt, C. OS. i. 227. 
 
 ' EschweUer, Ber. 38. 880; Koeppen, Ber. 38. 882 (1905). 
 
 = Guyot, C. E. 144. 1051 (1907) ; Guyot, Haller, ib. 947. 
 
50 Aromatic Amines 
 
 heating, which was discovered by Hofmann. If they are heated in an open 
 tube in a current of hydrochloric acid gas, the alkyls are split off as halides : 
 thus the final products of the action of hydrochloric acid on dimethyl-aniline are 
 aniline hydrochloride and methyl chloride. But if the salt is heated alone 
 in a tube to 300°, though this decomposition may very likely occur first, the 
 reaction goes further and the alkyl is replaced : not, however, on the nitrogen, 
 but on the nucleus. Thus : — 
 
 C,,H,NHC,H,.HC1 = G,^i<^^\^QY 
 
 ■^O-^-^S •^l-l-l^2-»~^5'- 
 
 The most important of these mixed amines is dimethyl-aniline, the mother 
 substance of a whole series of valuable dyes, such as auramine, crystal-violet, 
 methylene blue, malachite green, &c. It has an extraordinary power of 
 forming condensation-products, owing to the mobility of the para hydrogen. 
 Thus if it is gently oxidized, for example, with chloranU (tetrachloro-quinone, 
 CgCl^Oa) or copper chloride, it is converted into methyl violet, a mixture of the 
 hydrochlorides of the penta and hexa-methyl derivatives of triamino-triphenyl- 
 carbinol, HO-C(CeH4NH2)3. The linking methane carbon undoubtedly comes 
 from the separation of one of the methyl groups as formaldehyde. If manga- 
 nese dioxide and sulphuric acid are used as the oxidizing agents, formaldehyde 
 can actually be detected. 
 
 The quaternary halides are got from the tertiary bases by the action of 
 alkyl halide. But this reaction sometimes occurs with great difficulty, and 
 sometimes not at all. Thus Fischer ' has shown that anihnes in which both the 
 ortho-positions to the nitrogen are replaced, such as 
 
 I^H, NH, l^H, 
 
 CHg-^CHg CH3-ppCH3 CHg-j^Br, 
 
 CH3 
 though they form tertiary bases easily, will not form quaternary compounds 
 at all. This is obviously a case of stereo-hindrance, and is analogous to the 
 inactivity of the di-ortho-substituted benzoic acids and their esters, acid 
 chlorides, amides, and nitriles. 
 
 When the quaternary salts are heated above 300°, the alkyl groups, as has 
 been mentioned, migrate to the ring, occupying first the ortho positions, and 
 then the para. Thus trimethyl-phenyl-ammonium iodide, ^-NMegl, gives 
 dimethyl-toluidine, methyl-xylidine, and finally symmetrical amino-trimethyl- 
 benzene, or mesidine. 
 
 The quaternary hydroxides, unlike the primary, secondary, and tertiary 
 bases, are strongly basic caustic substances. 
 
 A. II. Purely aromatic amines 
 
 Diphenylamine, N^jH, was discovered by Hofmann in 1864. It is of 
 commercial importance as the source of certain dyes. It is made by heating 
 aniline with its hydrochloride to 200°. 
 
 ^NHg -f- 0NH3CI = (^aNH + NH4CI. 
 
 1 Ber. 33. 345, 1967 (1900). 
 
Purely Aromatic Amines 51 
 
 It can also be made by heating the phenols with aniline in the presence of zinc 
 chloride, calcium chloride, or antimony trichloride. This latter method is 
 useful for preparing unsymmetrical secondary amines, with substituted nuclei, 
 
 Diphenylamine melts at 54° and boils at 302°. It is scarcely soluble in 
 water. Its basic character is much weakened by the presence of the second 
 phenyl group ; it gives salts with strong acids, but they are at once decomposed 
 by water. 
 
 Its solution in concentrated sulphuric acid gives a deep blue colour in 
 presence of a trace of nitric acid, which affords a very delicate test for the 
 latter. It has been shown ^ that this reaction is due to the formation of 
 tetraphenyl-hydrazine, <p^-'^^^, by the oxidation of the diphenylamine. 
 
 The purely aromatic tertiary amines are difficult to prepare and are little 
 known. Triphenylamine, N^g, is got by dissolving sodium in boiling diphenyl- 
 amine and treating the product with phenyl bromide. It is not a base at all. 
 It gives no salt with hydrochloric acid even when the gas is passed into its 
 benzene solution. This is a remarkable proof of the negative nature of the 
 phenyl group as compared with alkyl. If we replace the hydrogen of ammonia 
 by alkyl groups, we get a strong base, which is stronger the more alkyls it 
 contains : while the introduction of aryl (e. g. phenyl) groups produces a con- 
 tinuous diminution of the basic character. Phenylamine (aniline) has no 
 alkaline reaction and does not absorb carbon dioxide : diphenylamine gives 
 salts only with strong acids, and these are decomposed by water : triphenylamine 
 gives no salts at all. 
 
 B. Aniline derivatives substituted in the nucleus 
 
 These may be obtained by direct substitution, or in other ways. When 
 they are obtained by direct substitution, the substituents occupy the ortho and 
 para positions, in accordance with Crum Brown's rule. (In some cases, where 
 the reaction is carried on in strong acid solution, the meta position is taken 
 up as well : this must be due to the fact that it is not the aniline which is then 
 being substituted, but its salt.) It is to be observed that in many cases of the 
 direct substitution of aniline it can be proved that the substituent first enters 
 the amino-group, and then migrates to the ring : and it is quite probable that 
 this happens in all cases." It seems, in fact, to be a general rule that a sub- 
 stituent in the amino-group of aniline (as in 0-NHX) is capable of migrating 
 to the ring, taking up mainly the para position, and to a less extent the ortho. 
 Thus if aniline is treated with bleaching powder, the chloramine, (/)-NHCl, is 
 formed : this changes with the greatest ease into para- and a little ortho- 
 chloraniline, NHCl Ij[H, NHg 
 
 6 - -^ (T 
 
 €1. 
 
 I 
 CI 
 
 These can be converted by bleaching powder into chloramines, which go over 
 
 1 Wieland, Gambarjan, Ber. 39. 1499 (1906). 
 
 ' Cf. Blanksma, Bee. Trav. 21. 269 (C 02. ii. 513). 
 
52 Aromatic Amines 
 
 into o-p-dichloraniline ; and this again gives a chloramine, which changes into 
 symmetrical trichloraniline : NHg 
 
 Cl-p|-Cl. 
 
 CI 
 Finally, the chloramine prepared from this tri-chloro-compound, CgHgClj-NHCl, 
 is quite a stable body. The chlorine cannot any longer migrate from the 
 nitrogen to the nucleus, because the para and both the ortho positions are 
 occupied, and it is not able to go into the meta. 
 
 The rate of change of acetyl-chloranilide into p-chlor-acetanilide, 
 ^•NCl-COCHa -> ClCeH^NHCOCHa, 
 has been measured by Blanksma,' the amount of unchanged substance being 
 determined by titration with potassium iodide and thiosulphate. The reaction 
 is found to be monomolecular ; it' takes place more rapidly in acetic acid or in 
 alcohol than in water ; and it requires the presence of hydrochloric acid as 
 a catalyst. The velocity constant is approximately proportional to the square 
 of the concentration of the hydrochloric acid. The reaction is also catalysed 
 by light. 
 
 This behaviour is not confined by any means to the halogens but is common to 
 nearly all the N-substitution-products of aniline. It is worth while to collect the 
 various instances in which it has been proved to occur, as otherwise the general 
 nature of the reaction is liable to be overlooked. 
 
 Taking the typical form 
 
 CeH5.N<| -> X.C,H,.NH„ 
 
 X may belong to any of the following classes : — 
 
 1. Alkyl: as in the Hofmann reaction, e.g. methyl-aniline -^ toluidine. 
 i. Chlorine or bromine, as in the above example. 
 
 3. Hydroxyl : thus ^-phenyl-hydroxylamine gives p-amino-phenol : 
 
 CeHXH"" -* HO.C,H,.NH,. 
 
 4. NO. Thus methyl-phenyl-nitrosamine gives p-nitroso-methyl-aniline : 
 
 <^6il5-J>i\CH3 ^ ^b^AnhCHs 
 
 5. NOg. Phenyl-nitramine, the so-called diazo-benzenic acid, ^-NH-NO,,, 
 gives a mixture of mainly para- with some ortho-nitraniline, NOg-CgH^-NH.,. 
 
 6. SO3H : as in ^NHSOgH -» HSOgCoH^NHo (o -f- p). " 
 
 7. Acyl. This only occurs with the diacyl-compounds. But diacetaniUde, 
 0-N(CO-CH.j)2, gives the isomeric acetyl-amino-benzophenone, 
 
 CHgCOCeH^NHCOCKj. 
 
 8. NHg. Phenyl-hydrazine, ^NHNH^, can be converted into para-pheny- 
 lene-diamine, H2N-CjH4-NH2. 
 
 9. -N=::N-0 : as in the conversion of the diazo-amino-compounds into the 
 amino-azo: ^-NH-N^N-^ -» H^N'OeH^-N^N-^. 
 
 ■ llec. Trav. 21. 366 ; 22. 290 (C. 03. i. 141, ii. 241). 
 
Migration of Suhstituents in Aniline 53 
 
 10. -NH0 and } 
 
 11. -CeH^NH^.J 
 
 These two together form the most striking illustration : but it is often formulated 
 
 in a way that conceals its real bearing. Hydrazobenzene, 0NH-NH^, a product 
 
 of the alkaline reduction of nitrobenzene, is very readily acted on by acids, 
 
 giving first a semidine, by the migration of the -NH0 to the para position on 
 
 the other nucleus:— H<3-NH-NH0 -» 0NH-<3-NH2. 
 
 The product is still a substituted aniline, and so a second change occurs, the 
 
 substituent C(;H4NH2 migrating on to the nucleus, forming diamino-diphenyl or 
 
 benzidine:— HC>NH<3NH2 -* H^NO^-O-NHa. 
 
 These cases will serve to show how very general the reaction is. It is 
 
 remarkable that in nearly all of them a body of feebly basic or neutral or acidic 
 
 character is converted into one of more pronounced basicity. 
 
 The property of giving on direct substitution ortho and para derivatives 
 is common to all derivatives of benzene CgHg-X in which the compound HX 
 cannot be directly oxidized to HOX. This is Crum Brown's rule. Since in the 
 case of aniline we find that the substituent first enters the side chain and then 
 passes on to the nucleus, we may ask whether the same is the case in all bodies 
 belonging to this class. The most important suhstituents X of the ortho and 
 para class are NHj, OH, the hydrocarbon radicals, and the halogens. In the 
 case of the halogens, this hypothesis is only possible if they assume a higher 
 valency in the intermediate compound : and of this there is no evidence. In 
 the case of the hydrocarbon radicals, though there is no intrinsic improbability 
 of a reaction of this kind, there is also no evidence of its occurrence. With 
 hydroxyl, on the other hand, there is direct evidence, though it is not so 
 wide as with NH2. Thus potassium phenyl sulphate is completely converted 
 by heating in a sealed tube into potassium phenol-sulphonate : — 
 
 CcHsOSOjOK -> KOSOgCsH^OH : 
 and sodium phenyl carbonate if heated with soda to 200° gives sodium 
 salicylate:— CoHg-OCOONa -^ NaOCOCeH^-OH. 
 
 Again, in the formation of oxyazo-compounds from diazo-derivatives and 
 phenols, there is reason to think that a similar change takes place : — 
 
 ^•Na-OH + HO.0 -» 0-N2-O-^ -^ ^-Na-CeH^-OH. 
 
 Whether similar intermediate compounds are formed in the case of other 
 substitutions of phenol still remains to be proved. 
 
 In the direct substitution of aniline and similar bodies, it is often necessai'y 
 to ' protect ' the NHg group, as it is called, against secondary reactions. For 
 example, aniline is very sensitive to oxidizing agents, and therefore, if it is 
 treated with a halogen or with nitric acid, the molecule is liable to be completely 
 broken up or to undergo complicated reactions, unless special precautions are 
 taken. To avoid this, the amine group is protected commonly by acetylation. 
 Thus, if aniline is to be nitrated, it is first converted into acetanilide, and then 
 this is nitrated, and the nitro-aeetanilide is then saponified by boiling with 
 potash, whereby the acyl group is removed and nitraniline obtained. But 
 1175 E 
 
54 Armnatic Amines 
 
 in many cases the presence of a strong acid is in itself a sufficient protection, 
 forming the more stable aniline salt. This, however, is liable to affect the 
 position of the substituent. For example, whereas acetanilide on nitration gives 
 the ortho and -feeta-nitro-products, as we should expect, and aniline itself does 
 the same if nitrated with nitric acid alone at very low temperatures, if it is 
 treated with nitric acid in presence of a very large excess of sulphuric acid, 
 it gives mainly meta-nitraniline. In this case of course it is not aniline which is 
 being nitrated, but aniline sulphate. 
 
 The influence of substituents on the basicity of aniline has been investigated 
 by Farmer and Warth.^ They determined the degree of hydrolysis of the 
 hydrochlorides by shaking the aqueous solution with benzene, and observing 
 the concentration of the base in the benzene layer. From this the concentration 
 of the free base in the water can be calculated, and hence the hydrolysis. 
 This does not give the real but the apparent dissociation-constant, as explained 
 above : that is, it gives the value of the real K -f- (1 + V), where 6 is the 
 (unknown) hydration-constant. It is, no doubt, owing to this complication 
 that the results show a certain irregularity. The general conclusions are that 
 the ortho position gives the greatest effect, and the meta the least ; and as 
 regards the nature of the substituents, the effect is in the order (strongly 
 negative) NOj, CO2H, N=N^, Br, CI, CH3, CH3O (weakly positive). 
 
 Halogen derivatives 
 
 Aniline, like phenol, can be chlorinated and brominated with great ease, 
 by the action of chlorine or bromine on an aqueous solution of an aniline 
 salt. Iodine can also substitute directly, since the excess of aniline removes the 
 hydriodic acid formed. The halogens can also replace one another on the ring 
 with singular ease : thus if para-bromanUine is chlorinated,' both trichloranUine 
 and chlor-dibromanUine are formed, some of the bromine of the original 
 compound being expelled by chlorine, and then attacking more of the com- 
 pound. 
 
 It is remarkable that chlorine and bromine have very little action on 
 aniline in strong sulphuric acid solution ; and as far as they do react, produce 
 meta-substitution-products. 
 
 In aqueous solution the final product of the action of the halogens is the 
 symmetrical (di-ortho-para-)tri-halide. To get the lower products (mono- and 
 di-haloid) it is best to treat acetanilide suspended in water with chlorine or 
 bromine (or if iodine is to be introduced, to treat it with iodine monochloride). 
 The first halogen atom goes mainly to the ortho position, the second mainly 
 to the para. In this way by starting with aniline it is not possible to get 
 beyond the tri-substitution-product : but by starting with meta-brom- or di- 
 meta-brom-aniline the tetra- or penta-brom compound can be obtained, the three 
 vacant places 2, 4, and 6 being filled up. 
 
 The diminution of basicity on introducing halogen atoms is shown by the 
 fact that while the mono-halogen anilines still give salts stable to water, 
 
 1 J. C. S. 1904. 1713. ^ Eeed, Orton, /. C. S. 1907. 1543. 
 
Halogen Derivatives of Aniline 55 
 
 the di-halogen compounds give salts which, if the acid is volatile, are largely- 
 decomposed on evaporating their aqueous solutions (i. e. which are highly 
 hydrolysed) : while the tri-halogen derivatives give no salts at all. 
 
 Sulplionic acids of aniline 
 
 Aniline is easily sulphonated by concentrated or slightly fuming sulphuric 
 acid, giving first the para-mono-sulphonic acid, and then the ortho-para- 
 disulphonic acid. The others may be got by indirect methods: for example 
 the meta by reducing meta-nitrobenzene sulphonic acid. They may also be 
 prepared by heating the haloid aryl-sulphonic acids with alcoholic ammonia to 
 160-180°, as the strongly negative sulphonic group (like the nitro-group) makes 
 the halogen more mobile. 
 
 The mono-sulphonic acids of aniline are colourless bodies which crystallize 
 well. Unlike the benzene sulphonic acids, which are excessively soluble in 
 water, they only dissolve sparingly. Their melting-points are also much 
 higher : thus while benzene sulphonic acid melts at 50°, aniline para-sulphonic 
 acid blackens without melting at 280-300°. They form salts with bases but 
 not with acids, and hence dissolve in aqueous alkalies and are reprecipitated 
 by acids. All these peculiarities indicate that they are intramolecular salts, e. g. 
 
 WO , 
 
 as in the parallel case of the amino-acids. This is confirmed by the fact that 
 they cannot be acetylated, while their sodium salts, which must have a free 
 NHg group, can. 
 
 Nitranilines 
 
 Aniline can be nitrated directly if it is dissolved in concentrated sulphuric 
 acid, cooled to 0°, and the calculated quantity of nitric acid dissolved in a 
 great excess of sulphuric acid slowly added. These precautions are necessary 
 in order to prevent the oxidation of the aniline. The product is a mixture 
 of meta and para-nitraniline with a little ortho. Alkyl and acyl anilines 
 behave in the same way.^ 
 
 The nitranilines may also be got by the partial reduction of the poly- 
 nitro-compounds, as in the preparation of meta-nitranUine from meta-dinitro- 
 benzene with ammonium sulphide. 
 
 The great effect of the nitro-group in weakening the basicity of the NHj 
 and its dependence on position were shown by Lellmann'' by a rough but 
 ingenious method. He took an equal weight (0-5171 gr.) of the hydro- 
 chlorides of each of the three nitranilines, boiled it up with 27 c.c. of water, 
 and evaporated the solution to dryness at 75°. The loss of hydrochloric acid 
 gives an approximate measure of the degree of hydrolysis. He found that 
 of the ortho-nitraniline salt, 63-8 per cent, was decomposed under these 
 
 » Tingle, Blanck, Am. Ch. J. 36. 605 (C. 07. i. 632). ' Ber. 17. 2719 (1884). 
 
 E 2 
 
0' 
 
 56 Aromatic Amines 
 
 conditions, of the meta 34, and of the para 13- 1. A more exact deter- 
 mination ' of the strength of these bases, by the increase of solubility of slightly 
 soluble acids (cinnamic and p-nitro-benzoic), shows their dissociation constants 
 
 to be 
 
 ortho 0-01 X 10-1^ 
 
 meta 4-0 x 10-^2 
 
 para 1-0 x IQ-i^. 
 
 Htfdroxy-anUines, or aminqphenols 
 
 They are obtained by the reduction of the nitrophenols and of the azo- 
 phenols or oxy-azo-compounds. The latter are often used for preparing the 
 more complicated derivatives, as they are easy to make by coupling the phenols 
 with the diazo-compounds. Thus if sulphanilic acid is diazotized and coupled 
 with, say, para-cresol, an oxy-azo body is formed which on reduction breaks 
 up, as usual, between the two nitrogen atoms, giving amino-cresol : — 
 
 OH OH 
 
 >jN=N.C,H,.S03H _ QNH, ^ h,N.C,H,.S03H. 
 
 CHg CHg 
 
 They are also formed by intramolecular rearrangement from the /3-phenyl- 
 hydroxylamines : CeHjNHOH -> HOC0H4NH2. Hence they are often pro- 
 duced in reactions which would be expected to lead to the hydroxylamines, 
 as in the electrolytic reduction of the aromatic nitro-compounds. 
 
 Their properties are both basic and acidic. They form stable salts with 
 acids, and also dissolve in aqueous alkalies. The ortho-compounds condense 
 very readily to form a second (nitrogenous) ring, giving, for example, with acid 
 anhydrides the so-called anhydro-bases, or benzoxazoles : — 
 
 0-2§^ + V^H, = CtS>C-CH3 + HO.COCH3. 
 
 ^^OCOCHa 
 The para-aminophenols, on the other hand, are readily oxidized to quinones. 
 These two reactions are characteristic of the two classes of ortho and para 
 di-derivatives of benzene respectively. 
 
 The oxygen esters of the ortho-aminophenols undergo a remarkable intra- 
 molecular change,'' the acyl group migrating from the oxygen to the nitrogen, 
 with the formation of a urethane : for example : — 
 
 f^OCO-OEt AOH 
 
 UNH2 ~* IjNHCOOEt' 
 
 The change takes place in acid solution, and its velocity can be measm'ed 
 by means of the conductivity.' The salt is partly hydrolysed in solution 
 into free base and acid, and it is found that it is only the free base (the 
 hydrolysed part) which changes, and that for this part the reaction is mono- 
 molecular. Einhorn and Pfyl* have shown that it is only the ortho-amino- 
 phenol derivatives which undergo this change, and not the meta or para. 
 
 » Lowenherz, Z. Ph. Ch. 25. 385 (1898). ' Eansom, Ber. 33. 199 (1900). 
 
 ' Stieglitz, Upson, Am. Ch. J. 31. 468 (C. 04. ii. 94). * Ann. 311. 34 (190C). 
 
Aminophenols 57 
 
 That is, the acyl group will not travel further than to an NH2 attached to 
 the next carbon on the ring. But if the NHg is on a side chain, the acyl will 
 go further. Thus ' a benzoyl group will go from the oxygen to the nitrogen of 
 salicylamide (i. e. to a nitrogen attached to the next carbon atom but two) : — 
 
 
 o o 
 
 and Auwers^ finds a similar migration to the nitrogen of a substituted 
 ortho-hydroxy-benzylamine : — 
 
 Brf^CHaBr + H^N^ 
 IJO-CO-0 
 
 ^'-05S?* - HB. 
 
 I 
 
 Br 
 
 Br-ACH^N^-CO-^ 
 UOH 
 I 
 Br 
 
 Br 
 
 The intermediate compound in this case has not been isolated, but there can 
 be no doubt that it is formed. The reaction takes place more easily the more 
 negative the acyl group and the more basic the nitrogen. 
 
 These facts indicate that the nitrogen attached to the carbon of the side 
 chain is nearer for the purposes of reaction than that on the meta carbon 
 atom of the ring. (Compare the fact that meta-(iso)-phthalic acid will not 
 give an anhydride though the corresponding open-chain acid, glutaric acid, 
 does so readily.) The reactivity of the side-chain compound is certainly not 
 due to any peculiarity of the open-chain compounds as such: for no such 
 migration of acyl groups occurs with the purely fatty derivatives, the 0-acyl 
 
 CH -Nil 
 
 esters of the amino-alcohols, for example ^J' ^ J' ,, being perfectly stable. 
 
 CH2-0-CO-9> 
 
 Nor does it occur even with the aromatic side-chain derivatives, if the amino- 
 group is on the nucleus and the hydroxyl on the side chain, as in [Jijij 
 whose 0-acyl ester is also quite stable. 
 
 Homologttes of aniline 
 
 The homologues of aniline (with the alkyl groups attached to the nucleus) 
 can be made by reducing the nitro-derivatives of the homologues of benzene, 
 in cases where these last can be easily obtained, and will give the required 
 nitro-derivatives. Or the corresponding phenols may be heated with the 
 compound of zinc or calcium chloride and ammonia. 
 
 Another method is to introduce the alkyls first into the amino-group of 
 aniline, and then transfer them to the nucleus by means of the Hofmann 
 reaction (heating the hydrochlorides to 250-350°). The alkyls can only be put 
 into the ortho and para positions in this way : but by starting mth sym- 
 metrical meta-xyUdine (1:3:5), where the two meta positions are already 
 occupied, Hofmann prepared amino-pentamethyl-benzene. 
 
 Their behaviour is in general that of aniline, except in so far as the side 
 chains interfere: thus para-acettoluidide on nitration gives not the para but the 
 ortho nitro-compound. 
 ■ McConnan, Titherley, J. 0. S. 1906. 1318. ' Ann. 332. 159 (1904) ; 364. 147 (1909). 
 
58 Aromatic Amines 
 
 Like aniline itself they tend to form quinones on oxidation; and this 
 tendency is so strong that if there is a methyl group in the para position 
 to the NHa, it is often eliminated. For example, mesidine is very easily 
 converted into meta-xyloquinone : — 
 
 NH2 O 
 
 CH3\ I /CH3 ^ CH3V " /CH3 
 
 I Jl 
 
 CH3 O 
 
 In some cases the side chains have unexpected effects, as is shown in the 
 
 formation of nitroso-derivatives from the tertiary amines. Dimethyl-aniline 
 
 of course gives the para-nitroso-compound, 0— N< >NMe^. Dimethyl-para- 
 
 toluidine, CH^< >NMe;,, gives no nitroso-derivative at all. This is to be 
 
 expected, since the para position is occupied. Dimethyl-meta-toluidine, 
 
 where the para position is free, behaves like aniline, and gives a para-nitroso- 
 
 0-]Sr<ri>NMe2 , , , .,. HONMe, 
 
 compound, ] . But dimethyl-ortho-toluidine, 1 , 
 
 CH3 OH 3 
 
 though it has the para position open, gives no nitroso-derivative. Nor is this 
 the only reaction in which dimethyl-ortho-toluidine shows that for some 
 unknown reason the para hydrogen is unusually firmly attached to the nucleus. 
 It refuses to give it up either to couple with diazo-compounds or to condense 
 with aldehydes: reactions which occur with dimethyl-anUine itself with the 
 greatest ease. 
 
 The oxidation of the amines leads to a variety of products, and has been 
 investigated in great detail ; but it will be more convenient to deal with it 
 later,' in connexion with the analogous question of the reduction of the nitro- 
 compounds. 
 
 BENZIDINE DEKIVATIVES 
 
 The diamines of diphenyl are of interest from the peculiar reaction in 
 which they are produced by intramolecular change from the hydrazo- 
 compounds : 
 H-<3-NH-NH<3-H -^ H-<0-NHO-NH, -* H,NC^-<I>-NH„ 
 
 amino-diphenylamine or semidine being first formed. As usual, it is normally 
 the para position that is occupied, di-para-diamino-diphenyl or benzidine being 
 produced. But it is not necessary that the final product should: be a benzidine. 
 The reaction may stop at the first stage, giving a semidine, or the second 
 stage may take place, but in a different way, the substituent (CgH4-NH2) 
 migrating not to the para but to the ortho position, 
 
 / — S-NH NH2 
 
 ^ — \ y — ^IjjH ^^ ^ ^* ^® usually written I • 
 
 This ortho- derivative is known as diphenyline. Which of the three possible 
 derivatives is formed depends on the substituents present in the two nuclei: 
 
 '■ See p. 162. 
 
Benzidine Derivatives 59 
 
 in the case of hydrazo-benzene itself the products are mainly benzidine, a 
 little diphenyline, and no semidine. 
 
 Many of the salts are remarkable for their insolubility in water, and this 
 has been made use of for the quantitative determination of certain acids, such 
 as sulphuric and tiingstic' 
 
 AMINO-DI- AND TRI-PHENYL-METHANES 
 
 Some of the amino-derivatives of di- and tri-phenyl-methane are of great 
 practical importance as dyes ; and they are also of theoretical interest in 
 respect both of their mode of preparation and of their structure. 
 
 The only important class of diphenyl-methane dyes is the auramines. 
 They are obtained from a body which is also a valuable so>irce of the triphenyl- 
 methane dyes, tetramethyl-diamino-benzophenone, known as Michler's ketone. 
 This body is prepared by saturating dimethyl-aniline with phosgene gas : — 
 
 /CI HCD-NMea /C^-lSKe^ 
 
 C=0 + = C=0 + 2 HCL 
 
 \C1 H<3-NMe2 \O-NMe2 
 
 This ketone is extraordinarily reactive, as is also the benzhydrol (secondary 
 alcohol) HC(OH) (C6H4-NMe2)2 obtained from it by reduction. The ketone 
 condenses with ammonia to give an imine 
 
 Me2N0>-C-O-NMe2 
 NH 
 whose hydrochloric acid salt is auramine, the purest yellow dye known. 
 
 Of more importance are the triphenyl-methane dyes, which include the 
 earliest anUine dyes made on the commercial scale. In 1843 Hofmann observed 
 that the so-called aniline oil could be converted by oxidation into coloured 
 substances. The first dye to be put on the market was W. H. Perkin's 
 ' Mauve ' in 1857 : this was an alcoholic solution of the product of treating 
 aniline sulphate with potassium chromate. In 1859 fuchsin was made in 
 France, by oxidizing aniline oil with stannic chloride. 
 
 The processes by which the triphenyl-methane dyes are made are very 
 various. They are formed in many quite unexpected reactions, in which 
 aniline or an alkyl aniline is oxidized in presence of some compound which 
 can supply the linking methane carbon atom. So great is the tendency to form 
 compounds of this type, that even if dimethyl-aniline is oxidized alone, one 
 methyl group is split off, as has already been mentioned, to give the necessary 
 link. 
 
 They can also be made by starting with benzaldehyde, whose CHO group 
 supplies the methane carbon. If this is heated with dimethyl-aniline, it 
 condenses with it with loss of water, giving tetramethyl-diamino-triphenyl- 
 methane, the leuco-base of malachite green : — 
 
 HC>-NMe2 /C3-NMe2 
 
 0-c=o -f = H2O + 0-c; 
 
 j^ HC>-NMe2 jl.\C>-NMe2 
 
 1 Friedheim, Nydegger, C. 07. i. 504 ; v. Knorre, ib. 993. 
 
60 Aromatic Amines 
 
 or a mixture of para-toluidine and aniline may be oxidized, which gives 
 para-leucaniline. 
 
 H.,N-C:>-CH, + -» H^N-O-Cf 
 
 Or thej' can be prepared from Michler's ketone, or better from its reduction- 
 product the benzhydrol. This condenses with aniline to give tetramethyl- 
 triamino-triphenyl-methane : — 
 
 H H 
 
 Me,N-C>-Q-<Z>-NMe2 + H-0-NH„ = MoaN-O-C-C^NMeo. 
 
 
 
 NH, 
 These bodies also appear to exchange their nuclei with remarkable ease, 
 considering that this implies the breakage of the link between carbon and 
 carbon. Thus' para-diamino-triphenyl-methane, if it is heated with a large 
 excess of ortho-toluidine and its hydrochloride, goes over into diamino-ditolyl- 
 phenyl-methane, which when heated with excess of aniline and its hydrochloride 
 goes back to the triphenyl-methane derivative : — 
 
 0.CH(CeH,NH,), + 2 CoHXSril ^ ^•CH(c,H3<g|p^ + 2 C^H^NH,. 
 
 The di- and tri-amino-triplienyl-methanes, which are made by the methods 
 described above and by many othei-s, are not themselves dyes. They are 
 not even coloured. They are the so-called leuco-bases. On oxidation they 
 readily give compounds whose composition is that of the leuco-bases plus 
 one atom of oxygen. These bodies have the hydrogen on the methane 
 carbon replaced by hydroxyl : they are carbinols or tertiary alcohols, e. g. 
 (Me^N( ^-):vCOH. These again are not dyes, but they form salts with acids 
 which are. The colour of the dyes so formed depends on their constitution, 
 and can be altered at will within certain rather wide limits by the successive 
 introduction of various groups. The main facts with regard to the changes 
 of colour so produced are as follows : — 
 
 The amino-triphenyl-methanes do not yield dyes at all on oxidation 
 (though they may give coloured substances) unless they contain at least two 
 amino-groups in the para position. Di-para-diamino-triphenyl-carbinol is a 
 violet dye, which, however, is not sufficiently fast to be used commercially. 
 The tri-para-triamino-compound is pai-a-leucaniline ; on oxidation it gives 
 para-rosanUine, the carbinol, whose hydrochloric acid salt is para-fuchsin, a 
 valuable red dye. 
 
 The introduction of substituents such as NO, into the nucleus has a 
 distinct though not a very great effect on the colour. Noelting and Gerlinger " 
 have examined the influence on the colour of malachite green (the salt of 
 
 ' Vongerichten, Weilinger, 0. 04. ii. 226. 
 
 2 Ber. 39. 2041 (1906). Cf. Keitzenatein, Kunge, /. pr. Ch. [2] 71. 57 ; Keitzenstein, 
 
 Eothschild, 73. 192; Keitzenstein, Sohwerdt, 75. 369 (C. 05. i. 1020; 06. i. 1168; 07. ii. 1412); 
 
 Bielecki, Koloniew, C. 08. ii. 876 ; Finger, /. pr. Ch. [2] 79. 492 (C. 09. ii. 362). 
 
Triphenyl-methane dyes 61 
 
 tetramethyl-diamino-triphenyl-carbinol (Me2N-C8H4-)2C(OH)C|)H5 of substituents 
 introduced into the nucleus which does not contain an amino-group. They 
 find that if these are in the ortho position the colour is changed towards blue, 
 and if in the para, towards yellow: in the meta, little effect is produced. 
 Of the substituents examined, chlorine had the greatest effect, then NOj, then 
 methyl, and least of all methoxyl, CH3O, which was very feeble. 
 
 The introduction of hydrocarbon radicals into the amino-groups, which is 
 easier, and therefore of greater commercial value, has more influence on the 
 colour. The simple diamino-compounds (all substituents are para unless 
 otherwise stated) are red or violet, but on alkylation they give valuable green 
 dyes. The red fuchsin (triamino-) becomes increasingly violet as successive 
 alkyl groups are introduced : while the introduction of phenyl groups causes the 
 colour to pass over into blue. 
 
 But it is remarkable that if an acid radical such as acetyl is introduced 
 into an amino-group, or if it is converted into a quaternary salt by the addition 
 of alkyl halide, its influence on the colour is entirely destroyed. Thus hexa- 
 methyl-para-fuchsin, whose leuco-base is (MejN-CgH^-jgCH, is crystal violet, an 
 important dye. On combination with methyl iodide the leuco-base gives 
 (Me2N-C(;H4-)2CH-C8H4-[NMe3lJ, and this body on oxidation gives a green dye, 
 that is, a dye which has the same colour as it would have if the group in 
 brackets were replaced by hydrogen, which would give malachite green. 
 
 Constitution of tlie rosaniline dyes 
 
 This includes two distinct questions, firstly the constitution of the leuco- 
 bases, which has been satisfactorily made out, and secondly that of the salts of 
 their oxidation-products, the true dyes, which is more open to dispute. 
 
 The nomenclature of the bodies is rather confusing. There are two groups 
 of rosaniline compounds, one with 19 and the other with 20 carbon atoms in 
 the molecule. In each we have to distinguish the leuco-base, its oxidation 
 product (the carbinol), and the salt (usually the hydrochloride) of the latter, 
 which is the true dye. The names are : — 
 
 Leuco-base Para-leucaniline Leucaniline 
 
 Carbinol Para-rosaniline Eosaniline 
 
 Salt (dye) Para-fuchsin Fuchsin 
 
 The first question is the constitution of the leuco-bases. Eosaniline was 
 isolated and analysed by Hofmann in 1862-3, but our knowledge of its 
 constitution is due to the work of Emil and Otto Fischer in 1876-8. Eosani- 
 line had been shown to contain three amino-groups, because it gave a compound 
 with nitrous acid which on warming lost all its nitrogen to form a body with 
 three hydroxyl groups. Leucaniline gave a diazo-compound which when treated 
 with absolute alcohol was converted into an unknown hydrocarbon CjoHig. 
 
 E. and 0. Fischer repeated this experiment with para-leucaniline, the next 
 lower homologue of leucaniline, containing one CH2 less. This gave the 
 hydrocarbon CigHie, which they found to be identical with triphenyl-methane. 
 
62 Aromatic Amines 
 
 Thus the key to the solution of the problem was discovered : the compounds 
 were shown to be triphenyl-methane derivatives, and para-leucaniline to be 
 triamino-triphenyl-methane. It remained only to find the position of the 
 amino-groups. 
 
 Now benzaldehyde condenses with aniline to give a diamino-triphenyl- 
 methane:- ^.(xjjq ^ 2CoHo-NH2 = (p-CIliG.li^NB.^)^ + H^O. 
 
 If this is diazotized the two amino-groups are replaced by hydroxyl, giving 
 dioxy-triphenyl-methane : and when this is melted with potash it splits off the 
 phenyl group to form dioxy-benzophenone, CO(C5H4-OH)2. This compound 
 was proved to have the two hydroxyls in the para position by the following 
 synthesis. 
 
 Benzaldehyde condenses in presence of potassium cyanide to benzoin, which 
 on oxidation yields benzil ; this when boiled with potash undergoes a remark- 
 able rearrangement and forms benzilic acid, which on oxidation loses COj and 
 forms benzophenone : — 
 
 0-CHO + CHO-0 -» 0CHOHCO-0 -* 0COCO-0 
 
 Benzoin Benzil 
 
 -» 0-C(OH)0 -> ^-CO-^. 
 
 COOH 
 Benzilic acid Benzophenone 
 
 Starting with anisaldehyde (para-methoxy-benzaldehyde) a precisely similar 
 series of reactions can be performed: — 
 
 CH,0<3CH0 + CH0O>0CH:) -^ CH30OCH0HC0<O0CH3 
 
 Anisoin 
 
 -> CH30OC0C0<30CH, -» CH..OC>C(OH)-C>OCH, 
 Anisil ^QQjj 
 
 Anisilic acid 
 ■^ CHoOC^COO'OCHs, 
 
 giving ultimately the dimethyl-ether of para-dioxy-benzophenone. This ether 
 is found to yield and to be formed from the same dioxy-benzophenone as that 
 got from the diamino-triphenyl-methane. Hence this diamino-compound, the 
 condensation-product of aniline and benzaldehyde, must have the two amino- 
 groups in the para position. 
 
 Now para-nitro-benzaldehyde condenses with aniline in the same way as 
 benzaldehyde itself. The product must have the two amino-groups on two 
 of the nuclei in the para position, and also the nitro-group in the para position 
 on the third, thus :- NHoC^CHONH^ 
 
 
 
 N02 
 
 When this compound has the nitro-group reduced to NHg, it gives para- 
 leucaniline, which thus is proved to have all the three amino-groups in the 
 para position, and to be tri-para-triamino-tri-phenyl-methane. 
 
 Finally, if para-nitro-meta-toluie aldehyde is condensed with aniline and 
 
Constitution of Rosaniline Dyes 63 
 
 the product reduced, leucaniline is formed, which therefore has the 
 formula : — jj 
 
 H2]sr-<o-c-c:>-NH2 
 
 0- 
 
 ^ -CH3 
 
 Thus the formulae of the leuco-bases are determined. Those of the dye-bases 
 (carbinols) are derived from them by substituting hydroxyl for the hydrogen 
 attached to the methane carbon. 
 
 We now come to the much more difficult problem of the constitution of the 
 salts of the carbinols, which are the real dyes, para-fuchsin and fuchsin. Take 
 the case of the simplest of these bodies, para-fuchsin. Para-leucanUine is 
 proved to be tri-para-triamino-triphenyl-methane. When this is oxidized 
 it gives the corresponding carbinol, which is certainly HO-C(CgH4-NH2)3 : 
 this is para-rosaniline. On treatment with hydrochloric acid it forms the dye 
 para-fuchsin. Analysis shows that one molecule of the carbinol reacts with 
 one molecule of acid, and one molecule of water is eliminated. The resulting 
 compound may be formulated in three ways. 
 
 1. The hydroxyl of the carbinol may be directly replaced by chlorine. This 
 would give a haloid ester, in the same way that alcohol gives ethyl chloride. 
 This is Eosenstiehl's formula. 
 
 2. The hydrochloric acid may add on to one amino-group, and the NH3CI so 
 produced may split ofif water with the COH, the nitrogen becoming linked with 
 the methane carbon. (0. and E. Fischer.) 
 
 3. The addition of acid and loss of water may occur in the same way as in 
 the last case, but instead of the nitrogen becoming linked to the methane 
 carbon, the ring may become quinoid, and be doubly linked on the one side 
 to the methane carbon and on the other to the dyad group ^NHaCl. (Nietzki.) 
 
 I. I^fH^NHa II. NH2NH2 III. NH2NH2 
 
 0^0 0^0 0^0 
 
 I I II 
 
 
 
 -NH2C1 j^JJ^^^J 
 
 NH2 
 
 Eosenstiehl 0. and E. Fischer Nietzki 
 
 Formulae 2 and 3 differ in the same way as do the old and the new formulae 
 for quinone (the peroxide and the diktoene): and as the former has been 
 abandoned in favour of the latter in the case of quinone, we may provisionally 
 leave the corresponding formula (2) out of consideration, although, as will be 
 seen later, there is some reason to think that these dyes may not be strictly 
 analogous to the simple quinones. The real question is between Eosenstiehl's 
 and Metzki's formulae. Eosenstiehl's theory, representing the salt as an ester, 
 gives a peculiarly simple explanation of its formation from aniline and carbon 
 tetrachloride 3 ^NHg + CCI4 = (NHaCoHJ^CCl -1- 3 HCl. 
 
64 Aromatic Amines 
 
 But the balance of evidence is strongly in favour of Nietzki's view. For 
 example, if para-fuchsin is warmed with potassium cyanide and alcohol, it gives 
 rosaniline hydrocyanide, which is certainly a nitrile, having the CN attached 
 to carbon thus (NH2CgH4)3C-CN. This at first sight appears in favour of 
 Rosenstiehl's view, which represents it as a simple replacement of chlorine 
 by cyanogen. But if so, the product ought to be similar in properties to the 
 fuchsin, which it is not. The cyanide is colourless, gives colourless salts 
 (of the NHj groups) with acids, and does not decompose, as fuchsin does, with 
 bases into the carbinol and an alkaline salt. Hence it is clear that the two 
 compounds are not analogous in structure, and that in para-fuchsin the chlorine 
 is not on the methane carbon. 
 
 It has been urged in favour of Eosenstiehl's formula that para-fuchsin can 
 take up three more molecules of hydrochloric acid to give a compound the 
 composition of which is that of the carbinol + 4 HCl - HjO : and that 
 therefore the three nitrogen atoms must be trivalent, and that one cannot 
 already have formed a salt. But this reaction is really one of the strongest 
 arguments on the other side. For in the first place the acid salt so produced 
 is colourless (or practically so) and hence cannot retain the structure of the 
 original dye. In the second place it is not formed instantaneously but slowly. 
 Now if its formation consisted merely in the addition of hydrochloric acid to an 
 amino-group, this should occur at once : for there is no known case where 
 the addition of acid to a tertiary nitrogen atom is slow : and also the product 
 should be similar to the original dye. But if Nietzki's formula is accepted, we 
 should expect the formation of this acid salt to be slow, and to be accompanied 
 by a loss of colour. For on this formula the dye-salt can take up two molecules 
 of hydrochloric acid at once : — 
 
 (H,NCeH4),=C=CcH^=NH„Cl -^ (ClH3NCeHJ,,=C=CoH,NH2Cl 
 but the third acid molecule can only be taken up by destroying the quinoid 
 
 p XT 'M'tT pi 
 
 grouping, e.g. to form (ClHgNCeHJa^CxQj * ^ , which must take time, 
 
 as does the formation of the quinoid itself, and must destroy the colour, as it 
 destroys the quinoid structure. 
 
 Further, there is no reason on Eosenstiehl's hypothesis why the bodies 
 should be coloured at all. The carbinol and the cyanide, whose stincture is 
 strictly analogous to that which he attributes to the dye, are colourless. But 
 the quinoid ring is constantly found to be accompanied by colour, both in the 
 quinones themselves, and in at least the majority of their derivatives. The 
 tendency of bodies of this type to be coloured is well illustrated by two hydro- 
 carbons discovered by Thiele, which have a close analogy to Nietzki's fuchsin 
 formula (III), as far as the carbon skeleton is concerned, and about whose 
 structure there can be no doubt.' They are tetraphenyl-para-xylylene (I) and 
 diphenyl-fulvene (II). 
 
 <f>-C-<f> 
 
 II. HC— CH 
 
 II II 
 
 HC^/CH 
 
 C 
 
 <t>-G-<f> 
 
 III. NH,C1 
 
 II ^ 
 
 9 
 
 A 
 
 NH^C^H, C.H^NH^ 
 
 ' Ser. 33. 666 (1900) ; 37. 1465 (1904). 
 
Coiistitution of Rosaniline Dyes 65 
 
 The two hydrocarbons are brilliantly coloured. The first crystallizes in 
 needles of the colour of powdered potassium bichromate, and gives deep yellow 
 or orange solutions. The fulvene derivative forms deep red prisms. The 
 analogy between these three compounds shows that we have good reason 
 to expect a body with Nietzki's formula to be coloured. Diphenyl-fulvene is 
 also of interest as bearing on the question of the formula of quinone and its 
 derivatives. The supporters of the older peroxide formula held that in quinone 
 the nucleus retained its aromatic structm-e, whereas on the diketone theory 
 it is supposed that this structure is destroyed, and replaced by an arrangement 
 of ordinary double bonds. Now in diphenyl-fulvene there is just such an 
 arrangement of double bonds : and a peroxide-like linking, much more an 
 aromatic structure, is inconceivable : and yet the characteristic brilliant colour 
 remains. This question, which is still very obscure, will be further discussed 
 in dealing with the nitrophenols. 
 
 But the most conclusive evidence for Nietzki's theory is derived from the 
 physico-chemical investigation of the change of the coloured rosaniline deriva- 
 tives into the colourless. Hantzsch' has shown that the *fuchsins and the 
 triphenyl-methane dyes in general are electrolytically dissociated in water 
 to a high degree. This in itself is evidence that they are true salts of penta- 
 valent nitrogen and not esters. Again, we know that when para-fuchsin is 
 treated with alkali it is converted into the carbinol, para-rosaniline. On 
 Eosenstiehl's formula this is only the replacement of the chlorine by hydroxyl. 
 But on the quinoid formula it must occur in two stages, a quaternary hydroxide, 
 the true dye-base, being formed first : — 
 
 NH.,C6H,-C-C6H,NH2 NH^C.H.-C-CeH.NH^ NH^CeH.-CJj-CeH.NH, 
 
 Oi - Oi - ■ 
 
 II II I 
 
 NH2C1 NH20H NH2 
 
 Para-fuchsin True dye-base Para-rosaniline 
 
 Hantzsch has shown that these intermediate compounds really exist. If 
 a dilute (red) solution of para-fuchsin is treated with an equivalent of soda, 
 the electric conductivity is at first very nearly the sum of those of the two 
 solutions separately. Since the solution must contain the ions of sodium 
 chloride, it follows that it must also contain the ions of a highly dissociated 
 ammonium hydroxide. At the same time the solution is still coloured. But 
 on standing, and more rapidly the higher the temperature, the conductivity 
 diminishes, and vrith it the colour: till finally, after some hours at low 
 temperatures, you get a colourless solution, whose conductivity is that of the 
 sodium chloride it contains. This can only be explained in one way. The first 
 product of the action of the soda on the para-fuchsin must be the formation 
 of a highly dissociated coloured hydroxide, which then changes on standing 
 into the colourless non-dissociated carbinol. Hence the fuchsin is not the 
 chloride of the carbinol, but of an isomeric basic hydroxide, the true dye-base 
 
 ' Hantzeeh, Osswald, Ber. 33. 278 (1900). 
 
66 Aromatic Amines 
 
 required by Nietzki's formula. The earbinol belongs to the class of bodies, 
 recently found to be of considerable extent, known as pseudobases : which are 
 not themselves basic, but in the presence of an acid are converted into the salts 
 of an isomeric (tautomeric) base. 
 
 Eecently a case has been discovered^ in which two forms can actually 
 be isolated. Hexamethyl-triamino-diphenyl-naphthyl-carbinol when precipitated 
 from the salt by alkali and recrystallized from xylene forms dark green crystals 
 melting at 260°, which yield red or red-violet solutions, and give the bright 
 blue colour of the dye with dilute acids at once. If the body is recrystallized 
 several times from ether it is converted into a colourless substance melting 
 at 153°, which forms colourless solutions, and only gives the blue colour with 
 acids on warming. The two have the same composition, and it is obvious 
 that the colourless body is the earbinol. If, as Noelting and Philipp assume, 
 the other isomer is the dye-base, we may write the formulae of the two : — 
 
 aH,.N{CH,), CeH,.N(CH3)2 
 
 HOC-<CI>-N(CH3)2 C=O=N(CH3)20H 
 
 CioHe-N(CH3), CioHe-NCCH.)^ 
 
 Colourless: M. Pt. 153". Coloured: M. Ft. 260°. 
 
 Carbinol. Dye-base. 
 
 But as Willstatter '' has pointed out, the high-melting isomer has a different 
 colour from the dye, and therefore must have a different constitution. What 
 this may be, we are scarcely yet in a position to consider : but the body 
 is evidently more closely related to the dye than is the carbinol, since it is 
 converted into it more easily. 
 
 In Hantzsch's work, the carbinol gradually separated out during the 
 experiment : and it might be objected that this separation was the cause of 
 the diminution of conductivity and colour. But similar results are obtained by 
 using solutions so dilute that the carbinol remains in solution, so that this 
 objection falls to the ground. By observing the conductivity at definite 
 intervals of time it is possible to measure the rate at which the dye-base goes 
 over into the carbinol. From the results of Hantzsch and Osswald Gerlinger "* 
 showed that the reaction was bimolecular ; that is, the rate of change was 
 proportional to the square of the concentration of the dye-base. By using the 
 colour as a measure of the amount of dye-base present, it has been shown' 
 that in the presence of a large excess of alkali (whose concentration may be 
 taken as constant during an experiment) the reaction is monomolecular, and 
 the constant proportional to the amount of alkali. Hence it follows that the 
 rate of production of carbinol is proportional to the product of the concentrations 
 of the cation of the base and of the hydroxyl. Hence if the hydroxyl remains 
 constant, the reaction is monomolecular : if the two vary together, as in 
 Hantzsch's experiments, it is bimolecular. This seems to be the general 
 rule: that where a dissociated substance changes tautomerically into an un- 
 
 1 Noelting, Philipp, Ber. 41. 579 (1908). » Ber. 41. 1459 (1908). 3 Ber. 37. 3958 (1904). 
 » Sidgwiok, Moore, Z. Pli. C/i. 58.385 (1907); J.C.S. 1909. 889; Sidgwick, Kivett, ib. 899. 
 
Constitution of Rosaniline Dyes 67 
 
 dissociated, the velocity is proportional to the product of the concentrations 
 of the two ions. It has been observed by Walker and Hambly ^ for the change 
 of ammonium cyanate into urea, and by Dimroth'' for a similar change in 
 certain triazole derivatives. It is natural to conclude that in such cases the 
 ions combine directly to the undissoeiated tautomer ; and that in the case of 
 these dyes, for example, it is the ions and not the undissoeiated dye-base that 
 go over into the carbinol. But this conclusion is not certain. The results are 
 equally compatible " with its being the undissoeiated base and that alone which 
 undergoes the change. For however strong the base is there must be some of it 
 undissoeiated : and the concentration of this part will by the law of mass-action 
 be proportional to the product of the concentrations of the two ions. If, 
 therefore, it is only the undissoeiated base that reacts, the rate of change will 
 be proportional to its concentration, and so also to that of the ions : which is 
 precisely what is found. The results are therefore equally compatible with 
 either view, and at present it does not seem possible to decide definitely 
 between them. 
 
 The rosaniline dyes seem to have a certain tendency to go back from 
 the coloured quinoid form to the colourless non-quinoid form at very low 
 temperatures : for it has been shown * that the alcoholic solutions of the 
 rosaniline dyes (including crystal violet, but not malachite green), if they 
 are cooled with liquid air, become nearly colourless. If an aqueous solution 
 is treated with sulphur dioxide, it loses its colour, owing to the formation 
 of an acid sulphite ' ; but if an aldehyde is added, this removes the sulphurous 
 acid, and the colour is restored. This is Sehiff's test for aldehydes. 
 
 Nietzki's theory as to the constitution of these dyes is entirely confirmed 
 by the behaviour of their salts, as has already been mentioned. Of these 
 salts various and somewhat conflicting accounts have been given, but the 
 main facts are as follows.'' The compounds all give at least three series of 
 salts, though they cannot in all eases be isolated. (1) The carbinol combines 
 in the cold with as many molecules of a monobasic acid as there are nitrogen 
 atoms in the molecule, to form a colourless salt. This is obviously (to take 
 for simplicity the case of a diamino-body) ^•COH(CcH4-NH2-HX)2. In solution 
 this gradually changes with loss of water (and probably also of one of its acid 
 molecules) into the dye itself (2), for example H2N-C6H4-C0=CoH4=NH2X. In 
 presence of excess of acid this takes up at once another molecule of acid, without 
 change of colour in the solution, forming (3) XNH3-CoH4-C0=CeH4=NH2X, 
 and then slowly a third molecule, by the destruction of the quinoid grouping 
 
 and hence also of the colour, giving (4) XWHa'CgHi-C'CsH^-NHaX. Of these 
 
 X 
 salts all except (3) have been isolated. The last (4) is orange in the solid 
 state, but its colour is infinitesimal in comparison with that of the dye. 
 
 1 /. C. S. 1895. 746. ^ ^»»- 335. i. (1904). 
 
 = Goldschmidt, Z.f. EleUrochem. 11. 5 (C. 05. i. 451). 
 
 * Sohmidlin, C. R. 139. 731 (C. 05. i. 96). ' Diirrschnabel, Weil, Ber. 38. 3492 (1905). 
 
 ' Lamprecht, Weil, Ber. 37. 3058 (1904) ; 38. 270 (1905) ; Hantzsoh, Ber. 33. 753 (1900). 
 
68 Aromatic Amines 
 
 If a solution of fuchsin is treated with alkali and at once extracted with 
 ether, a brown substance dissolves in the ether which is a quinone imine, 
 (NH2'CoH4)2C=CeH4=NH, known from its discoverer as the Homolka base. 
 Baeyer and Villiger' have urged that this is the only immediate product of 
 the action of alkali on the dye, and that it takes up water to form the carbinol, 
 without any true ammonium base being formed at all. But this would not 
 explain^ the high conductivity of the freshly prepared solution, nor the fact 
 that the completely alkylated dyes (such as malachite gi-een and crystal 
 violet) behave in precisely the same way, though they cannot possibly form 
 an imine. The imine is no doubt formed in the aqueous solution to a small 
 extent by the loss of water from the true dye-base, and being much more 
 soluble in ether is removed by it, and then the disturbed equilibrium restored 
 by the production of more imine, which is again removed, and so on : just as 
 ammonia can be removed from its aqueous solution, for example by a stream 
 of air, in the anhydrous form of NH;j . 
 
 DIAMINES 
 
 Some of these bodies have already been mentioned, for example the diamino- 
 acids among the products of the hydrolysis of proteins: and the triphenyl- 
 methane dyes are of course di- and tri-amino-compounds. But the important 
 reactions of all these bodies are closely analogous to those of the simple mon- 
 amino-compounds, and therefore they have been discussed along with the latter. 
 
 The aliphatic diamines differ very greatly from the aromatic, since in 
 addition to the differences already pointed out between the alkyl- and arylamines 
 the interaction of the two amino-groups (always an important class of reactions 
 of di-substitution-products) is much affected by the greater rigidity which 
 results when the intervening carbon atoms form part of a benzene ring. 
 
 Aliphatic diamines 
 Methods of preparation. 
 
 1. From the di-halogen-derivatives of the hydrocarbons which have the 
 halogens on different carbon atoms, by treatment with ammonia, generally 
 by heating in sealed tubes with alcoholic ammonia to 100° : — 
 
 CH^Br NH. CH,NH„ 
 
 1 ' + ' = I ^ ^ + 2 HBr. 
 CH^Br NH,, CH2NH0 
 
 In this reaction secondary and tertiary derivatives are formed at the same 
 
 time, either with open chains as NHa-CH^-CHj-NHiCHa-CHa-NHj, or with 
 
 closed chains, as NH ' /NH, and the separation of these products is 
 
 ^CHoCH./ 
 difficult. 
 
 2. By the reduction with sodium in alcoholic solution of the nitriles of the 
 dibasic acids, such as succinonitrile, 
 
 ? 
 
 H^-CSN ^ ^ jj ^ (^H^-CH^-NH^ 
 CHa-CEN "^ - CH2CH2NH2" 
 
 ' Bei: 37. 2848 (1904). = Hantzsoh, Ber. 37. 3434 (1904). 
 
Aliphatic Diamines 69 
 
 This reaction gives bad yields of the lower members of the series owing to 
 the formation of ring compounds, such as pyrrolidine and piperidine ; but 
 this ring-formation diminishes, and the yield improves, as the number of 
 carbon atoms increases.' 
 
 3. By Gabriel's phthalimide reaction : — 
 
 ^CO. 
 
 CeH,^^>NK + BrCH,.CH,.CH,.Br + KN<gg>CeH, 
 
 ■^co 
 
 CO 
 
 4. By a modification of Kolbe's electro-synthesis. Just as acetic acid on 
 electrolysis gives hydrogen at the cathode and a mixture of carbon dioxide 
 and ethane at the anode : — 
 
 CH3.COOH _ H + 2 ro + 9^« 
 
 CH3COOH - ^2 + ^ <^02 + ^jj , 
 
 so "^ if an amino-acid such as glycocoU is used, a diamine is produced : — 
 
 H,N.CH,.COOH _ „ , o po ^ QH.-NH, 
 H2NCH2COOH - tl.2+ 2L0., + ^jj^j^jj • 
 
 5. The substituted «-diamines may be got by the action of secondary amines 
 on oxymethylene (formaldehyde) : — 
 
 The alkylene diamines are colourless liquids or solids of low melting-point ' ; 
 it is to be noticed that their melting-points, like those of the dibasic acids, 
 rise and fall alternately, so that a compound with an odd number of carbon 
 a,toms always melts at a lower temperature than one with one carbon atom 
 less : the boiling-points, however, rise continuously. (The melting-points are : 
 C2, -f 8-5°; C3, Uquid; C4, 27°; Cj, liquid; Cg, 42°.) They are easily soluble 
 in water, and are strong diacid bases, attracting carbon dioxide from the air. 
 They have a great afiBnity for water; some of them form hydrates with 
 constant boiling-points, which will not give up their water except on heating 
 for several hours with freshly fused potash, or on repeated distillation over 
 sodium : in this they resemble hydrazine. 
 
 Nitrous acid converts them into glycols, or sometimes, by loss of water 
 from the latter, into unsaturated alcohols. Ethylene diamine itself is converted 
 by nitrous acid into ethylene oxide. 
 
 As in the monamines, the amine hydrogen can be replaced by acyl groups. 
 
 Thus, on shaking the alkaline solution with benzoyl chloride, characteristic 
 
 , CH2.NHC0<^ 
 dibenzoyl derivatives are produced, such as I • This is the Schotten- 
 
 CH2*NB['C0^ 
 
 Baumann reaction, which is of great importance for isolating and identifying 
 
 various classes of bodies containing an NH group. 
 
 |,v. Eraun, Miiller, jBer. 38. 2203 (1905). = Lilienfeld, C. 04. i. 133 
 
 ' Kaufler, C. OX. i. 610. 
 1175 ■ P 
 
70 Diamines 
 
 The hydrochlorides of the diamines when heated lose NH4CI to give cyclic 
 imines, but with very different ease, and not always in the same way. 
 Pentamethylene-diamine reacts very readily, giving piperidine : — 
 
 ^CHa-CHa-NHa-HCl ^CHj-CH^. 
 
 CH2 = CH2 >NHHC1 + NH4CI. 
 
 ■^^CHaCHa-NHg-HCl ^CHsCH/ 
 
 Tetramethylene-diamine gives the corresponding pyrrolidine (tetrahydro-pyrrol) 
 
 I /NH. Trimethylene-diamine gives a certain quantity of trimethylene- 
 
 CH2"CH2 
 
 imine, CH2 /NH, but produces mainly picolines (methyl-pyridines) by 
 
 ^CH2^ 
 a more complicated reaction. Ethylene-diamine does not form any ethylene- 
 
 T^\ 
 
 /NH, although this body exists, but only diethylene-diimine, 
 CH2 
 
 ^^CH2CH2. 
 NH '>NH. This is obviously a question of strain in Baeyer's sense. 
 
 ^CHg-CHa 
 the larger rings being the most easily formed.' 
 
 Some of these diamines are among the ptomaines found in the animal 
 body a short time after death. Such are tetramethylene-diamine or putrescine 
 and pentamethylene-diamine or cadaverine. They are probably derived from 
 the diamino-acids which are among the constituents of the proteins. 
 
 Aromatic diamines 
 
 These are most easily formed by the reduction of the dinitro-compounds 
 or the nitro-amines with tin or stannous chloride and hydrochloric acid. 
 They can also be got by the reduction of the aminoazo-bodies, which gives 
 a diamine and a monamine : — 
 
 NH2-C3-N=N-<p + 2 H2 = NH2-<p-NH2 + H2NO ■ 
 
 CH3 CHg CH, CH3 
 
 They are colourless crystalline bodies, which are much more soluble in 
 water than the aniline bases. They form salts with two equivalents of acid. 
 They are very easily oxidized ; their solutions are readily acted on by air and 
 light. But the solid diamines and their salts are stable and may be kept 
 without change. 
 
 Their chemical behaviour depends largely on the relative position of the 
 two amino-groups. 
 
 The ortho-diamines are remarkable for their great tendency to form ortho- 
 condensation-products ; the two nitrogens joining up through the other body 
 with which they react to form a new ring. Thus they combine with acids 
 to form the anhydro-bases, cyclic amidines or imidazoles : — 
 
 0:Sg:+«S>c.cH.-Q:5!2i«-«H. 
 
 ' See under trimethylens-imine. 
 
Aromatic Diamines 71 
 
 These bodies are strong bases, and are thereby distinguished from the true 
 amides, such as C6H4(NH-CO-CH3)2, which are formed with acids by the 
 meta- and para-diamines. Such amides are also formed by ortho-diamines if 
 they are treated with acid anhydrides instead of acids, but they are unstable, 
 and when heated above their melting-points readily pass into the anhydro-bases. 
 
 Similar anhydro-bases have been obtained' from the fatty diamines also, 
 by heating their hydrochlorides with sodium acetate. In this case a mixture 
 of the anhydro-base and the true amide is produced. The proportion of 
 anhydro-base is greater the less the distance between the amino-groups. 
 Thus trimethylene-diamine gives mainly the anhydro-base, tetramethylene- 
 diamine mainly the amide. These results further support the strain theory, 
 and they at the same time illustrate the greater rigidity of the benzene ring 
 as compared with an open chain ; for the aromatic bodies corresponding to 
 tri- and tetramethylene-diamine, namely meta- and para-phenylene-diamine, 
 do not form anhydro-bases at all. 
 
 The ortho-diamines form similar ring compounds with nitrous acid, the 
 
 azimides:— i^-NH, , H0\^^ |^-NH\^^ „„_ 
 
 U-NH, + 0>N = U-N^N + 2 H,0. 
 
 These are colourless excessively stable substances, which can be boiled with 
 acid or alkali or heated to a high temperature without decomposition. They 
 are thus sharply distinguished from the diazo-compounds, and more particularly 
 
 from the enormously explosive azoimide, H-N<^ll , which contains the same 
 
 chain of three nitrogen atoms. This difference also is no doubt due to the 
 strain in the ring. 
 
 In the same way the ortho-diamines condense with the ortho-(a)-diketones,. 
 giving azines or quinoxaUnes : — 
 
 
 
 These are formed very easily, and may be used to characterize the ortho- 
 diamines. The alcoholic solution of the diamine is boiled with an acetic 
 acid solution of phenanthrene quinone, when the phenanthrazine 
 
 O^N 
 
 ■ ID 
 
 separates at once, and may be identified by its melting-point and by analysis. 
 
 Finally, ortho-phenylene-diamine is capable of an ortho-condensation with 
 itself, forming an azine. If its concentrated hydrochloric acid solution is 
 warmed with ferric chloride, long deep red needles of diamino-phenazine 
 separate :- H-f^-NH, ^ r^^f NppNH, . 
 
 1 Haga, Majima, 0. 03. i. 702. 
 f2 
 
72 Diamines 
 
 The meta-diamines are incapable of these condensations to anhydro-bases, 
 azimides, azines^ and so forth. Their most characteristic behaviour is with 
 nitrous acid. In strong hydrochloric acid solution, if the nitrous acid is 
 
 always kept in excess, a normal tetrazo-compound, rTjtr^Viij is formed. But 
 
 if the neutral solution of the hydrochloride is treated with sodium nitrite, 
 a brown dye of the aminoazo-class is produced : — 
 
 0-N=NCl + H<3-NH2 -^ O-N^N-O-NHg. 
 NH2 NH2 
 
 This is Bismarck brown or Manchester brown, the earliest azo-dye (1866). 
 Owing to this reaction a solution of meta-phenylene-diamine is turned deep 
 yellow by a trace of nitrous acid, a very delicate test for the latter. (It is 
 to be remembered that the best organic test for nitrous acid is meta-phenylene- 
 diamine, and for nitric acid, di-phenylamine.) 
 
 The para-phenylene-diamines are distinguished by the ease with which they 
 form quinones : — NHg O 
 
 I II 
 
 NH2 o 
 Also when treated in dilute acid solution with hydrogen sulphide and ferric 
 chloride they form violet or blue dyes (Lauth's violet). 
 
 It is to be noticed that the reactions of the ortho-diamines in giving 
 ortho'condensation-products, and of the para- in giving quinones, are character- 
 istic of the ortho- and para-di-substitution-products respectively in many other 
 cases as well. 
 
 Compounds containing more than two amino-groups attached to a benzene 
 nucleus are not of great importance. They may be prepared by nitrating the 
 aoetyl-derivatives of the diamines and reducing. In this way compounds up 
 to penta-amino-benzene, Cr,H(NH2)5, have been made. Attempts to prepare 
 hexa-amino-benzene, C6(NH2)c, have not been successful. Trinitro-triamino- 
 benzene can be made, but when it is reduced one nitrogen is split off and 
 penta-amino-benzene formed. This looseness of attachment of the nitrogen 
 is characteristic of these poly-amines. Thus triamino-mesitylene, when its 
 hydrochloride is boiled for four hours with acetic acid, has one NHg replaced 
 by hydroxyl,' 
 
 C6(CH3)3(NH2)3 -^ Ce(CH3)3(NH2)20H. 
 
 Also, as the number of amino-groups increases, the compounds become 
 more soluble in Water, and more unstable and readily oxidized. 
 
 QUINONE IMINES AND DIIMINES 
 
 These form a group of compounds closely allied to the diamines, which 
 are of considerable theoretical importance. They are derived from the quinones 
 by replacing the oxygen by NH or substituted NH. The first to be prepared 
 
 ' Wenzel, C. 02. i. 185. 
 
Quinone Imines and Diimines 73 
 
 were the chlorimine C1-N=<^=0 and the dichloriniine C1N=<3=NC]. 
 They are got by oxidizing para-amino-phenol and para-phenylene-diamine with 
 bleaching powder. They are respectively yellow and colourless crystalline 
 bodies ; they are volatile in steam, and are decomposed by dilute acids to form 
 ammonium salts and quinone. 
 
 Many vain attempts were made to obtain the mother substances, the 
 monimine 0=C6H4=NH, and the diimine HN=CoH4=NH, until recently 
 Willstatter ' succeeded in preparing them in two ways: firstly, by treating 
 the dichlorimine with hydrochloric acid, which acts as a reducing agent (like 
 hydriodic acid), as it does in other cases with chlorine attached to trivalent 
 nitrogen : probably an intermediate addition-compound is formed : — 
 
 CeH,(=NCl)2 -» C6H,(=NHCl2)2 -* CeH^l^NH)^ + 2 Cl^ ; 
 secondly,' by oxidizing the amino-phenol or the/phenylene-diamine in ethereal 
 solution vsdth silver oxide, or in some cases with lead dioxide. 
 
 The bodies are very difficult to isolate, and can only be prepared if the 
 materials are absolutely dry. The best known is the diimine 
 
 NH 
 
 01- 
 
 NH 
 
 It is a colourless crystalline substance, whose solutions are colourless at first, 
 though they soon become coloured owing to decomposition : its molecular 
 weight in acetone solution (as determined by the boiling-point) is that of the 
 simple formula. It is explosive. It is easily reduced to para-phenylene- 
 diamine, and when treated with dilute acids it breaks up into ammonia and 
 quinone. It is not acidic but is weakly basic, forming a colourless hydrochloride. 
 
 The monimine, O^CgH^— NH, is similar, and is also colourless ; but it is even 
 less stable, and explodes spontaneously in a few minutes. 
 
 By the oxidation of the methyl^-phenylene-diamines Willstatter " obtained 
 the methyl- and dimethyl-dumines, which likewise are colourless and explosive : — 
 CH3N=C6H4=NH and CH3N=C6H4=NCH3 . 
 
 The absence of colour in these compounds is important, in view of their 
 undoubtedly quinoid structure. It is commonly assumed that all quinoid 
 compounds are coloured, like the quinones themselves : and there is even a 
 tendency to suppose that nearly all coloured aromatic derivatives contain a 
 quinoid ring. But these views require to be reconsidered in view of the recent 
 discoveries on the one hand of quinoid substances (like the imines) which 
 are colourless, and on the other of coloured bodies closely resembling the 
 (supposed quinoid) aromatic coloured compounds, which cannot contain a 
 quinoid ring because they are not aromatic derivatives at all. 
 
 The causes which determine the absence of colour among these quinone 
 derivatives are not understood. It seems certain that the simple quinoid 
 
 ' Willstatter, Mayer, Ber. 37. 1494 (1904). 
 
 = WUlatatter, Pfannenstiehl, Ber. 37. 4605 (1904). = Ber. 38. 2244 (1905). 
 
74 Imines 
 
 compounds form two series, one coloured, like the quinones themselves, and 
 the other not. On the other hand, these bodies are capable of a further increase 
 of colour, which must be accompanied by some structural change. In the 
 triphenyl-methane series, for example, the feebly coloured Homolka bases go 
 over on treatment with acids into the brilliantly coloured dyes. This is 
 generally represented as the passage of trivalent into pentavalent nitrogen : — 
 (H2N.CoHJ,C=0=NH -^ (H2N.CoHJ2C=0=NH2Cl. 
 
 But there is much evidence to show that the change of valency of the nitrogen 
 is not sufficient to determine colour : thus the diimine salts are colourless. It 
 is clear that the influence of a pentavalent nitrogen atom on the colour 
 largely depends on whether it still has a hydrogen atom attached to it : if 
 it has, then the efPect on the colour is much the same as if it was trivalent ; if it 
 has not, then its effect is usually quite different.' The conversion of the group 
 -NXg, where X is a hydrocarbon radical, into -NXgHCl does not seem in 
 general to influence the colour, but its conversion into -NX3CI does so in 
 a very marked way. Of this we have an example in the quinone imine 
 derivatives which we have just been considering, and another (curiously in 
 the opposite direction) in the rosaniline dyes, where the change from -NEtg 
 to -NEt^HCl is without effect, while the change to -NEtgCl destroys the 
 influence of this group on the colour altogether. This is one of the phenomena 
 which suggest that there is a difference in constitution between bodies of 
 the type E3NHX and those of the type E^NX." 
 
 We must therefore look for some other change of structure in the con- 
 version of the Homolka bases into the rosaniline dyes. In this connexion 
 certain facts recently discovered by WUlstatter' with regard to the quinone- 
 diimines are of great interest. If unsymmetrical dialkyl-^phenylene-diamine 
 is oxidized, brilliantly coloured substances are produced, known from their 
 discoverer as Wurster's salts, which were supposed to be true diimine derivatives, 
 e. g. HN=CqH4=N(CII3)2C1. Further investigation has shown, however, that 
 these bodies contain an atom of hydrogen more than this : they are bimolecular 
 compounds (analogous to the quinhydrones) of one molecule of the oxidation 
 product with one molecule of the unoxidized diamine, and may be written 
 
 NH2CI NH., 
 
 II ^ I ' 
 
 Q6H4 C'eH^ , 
 
 N(CH3)2C1 N(CH3)2 
 
 the dotted lines indicating some unknown kind of linkage. If they are further 
 oxidized to the true diimines the colour disappears. WUlstatter suggests that 
 this peculiar linkage, which he calls meri-quinoid (partially quinoid), is the 
 cause of colour in bodies of this type. An analogous explanation would hold 
 for the triphenyl-methane dyes, the true dye having a linkage between the 
 quinoid nucleus and one of the other nuclei, which would be absent in the 
 Homolka base. 
 
 ' Cf. Pringsheim, Ber. 38. 3354 (1905). = See p. 26. 
 
 = Willstiitter, Piocard, Ber. 41. 1458, 3245 (1908) : 42. 1902 (1909). 
 
Quinone Imines and Diimines 75 
 
 When o-phenylene-diamine is oxidized with silver oxide or lead dioxide it 
 
 appears to give an o-quinone diinaine,' [ t-iiTTTi though this is too unstable to 
 
 be isolated. It gives a yellow solution with a red fluorescence (ortho-benzo- 
 quinone itself is red and not yellow). 
 
 Benzidine (diamino-diphenyl) seems to give an analogous compound,'' 
 HN=CD=CD=NH. This is brown and gives dark yellow solutions. The 
 chlor-imines, which are more stable and can be isolated, are simUar." But the 
 diphenoquinone, 0=C6H4=C6H4=0, from which it is derived is itself red in one 
 form, and generally resembles ortho- rather than para-benzoquinone. 
 
 The analogous salts from tetramethyl-benzidine, where the nitrogen is 
 pentavalent and has no hydrogen attached to it, form two series, one yellow 
 or red, the other green. The red salts probably have the formula 
 
 (*^H3)|^N=C„H,=CeH,=N<gH3)2 . 
 
 The benzidine compounds are able, like the jp-phenylene diamines, to give 
 coloured ' meri-quinoid ' oxidation products.* 
 
 ' WiUstiitter, Pfannenstiehl, Ser. 38. 2348 (1906). 
 
 - WiUstatter, Kalb, Ber. 37. 3761 (1904); 38. 1232 (1905). 
 
 3 Schlenk, Ann. 363. 313 (1908). * Schlenk, 1. 1;. 
 
CHAPTER IV 
 
 AMIDES 
 
 As the amines are formed by replacing the hydrogen of ammonia by 
 hydrocarbon radicals, so by replacing it by acid radicals the amides are 
 produced : or we may say that the amines are alcohols whose hydroxyl is 
 replaced by NHj, and the amides acids whose hydroxyl is replaced by NHg. 
 As in the case of the amines, primary, secondary, and tertiary amides exist ; 
 but the last two classes are of small importance. Quaternary amides are 
 unknown. 
 
 Methods of formation 
 
 1. By heating the ammonium salts of the acids : — 
 
 CH3C<C.NH, = Bf.O + CH3-C<gjj^- 
 This reaction is exactly parallel to the formation of ester from acid and alcohol, 
 as is evident if they are written : — 
 
 (a) CH3.C< gjj + HNH^ = H,0 + CH3-C<gjj_^ • 
 
 (6) CH3C<gjj + HO-C^Hs = H^O + GK^-C^^^^^ ■ 
 
 The similarity is not merely formal. Goldschmidt' has investigated the 
 rate of formation of substituted amides (anilides) and filnds that it follows 
 the same laws as esterification. In both cases the reaction in the absence 
 of a strong acid is bimolecular, being auto-catalysed by the hydrogen ion of 
 the organic acid itself. In the presence of a strong acid it is highly catalysed, 
 and then becomes monomolecular, as the disappearance of the organic acid 
 no longer diminishes the concentration of the hydrogen ion. 
 
 This work has been confirmed by Menschutkin,^ who has investigated the 
 influence of the nature of the organic acid on the velocity, and on the per- 
 centage of amide formed when equilibrium is reached. (These experiments, 
 like Goldschmidt's, were carried out by dissolving the acid in a large excess 
 of the amine, under which conditions the reaction is far from complete.) 
 Menschutkin's general conclusions are that the velocity is greatest with 
 normal fatty acids. The presence of branch chains diminishes it, and the 
 more so the nearer they are to the carboxyl. With the true aromatic acids 
 (where the carboxyl is directly attached to the nucleus) the velocity is very 
 small, as it is with all acids which have the carboxyl attached to a tertiary 
 carbon atom. Side chains on the nucleus diminish the velocity greatly in the 
 ortho position : in the meta and para they sometimes increase it. Aromatic 
 acids whose carboxyl is in the side chain behave like fatty acids. 
 
 1 Z. Ph. Ch. 24. 35S (1897). ' C. 03. i. 1121 ; ii. 324. 
 
Formation of Amides : Stereo-hindrance 77 
 
 The case of the ortho and di-ortho-substituted benzoic acids is of interest. 
 V. Meyer has laid down the general rule that the reactions of benzoic acid 
 and its derivatives are in general delayed or prevented by the presence of 
 two (and in less degree of one) ortho-substituents. In accordance with this 
 Menschutkin finds that the rate of formation of the amides is diminished by 
 the presence of an ortho-substituent : the case of two he has not examined. 
 But it does not follow that the percentage of amide present at equilibrium 
 will be diminished. It is characteristic of stereo-hindrance in general, and 
 of this form of it in particular, that where it delays a reaction it also delays the 
 reverse action. Thus it hinders not only the formation of the esters but also 
 their saponification. In fact stereo-hindrance is of the nature of a negative 
 catalyst. It acts not by causing instability of the product (as a purely 
 chemical influence does) but by making a particular group difficult of access. 
 Hence not only is the group slow to react, but the product of its reaction is 
 slow to decompose. Since, however, the proportions at equilibrium do not 
 depend on either of these reaction-velocities separately, but on the ratio between 
 them, they are much less affected by the stereo-hindrance, and need not 
 necessarily be affected by it at all. These general principles are well illustrated 
 by the case under consideration, that of the amides. That the velocity of 
 hydrolysis of benzamide is greatly diminished by ortho-substituents has been 
 shown by Sudborough.^ The percentage of monobrom-benzamide hydrolysed 
 by boiling for 15 minutes with 30 per cent, sulphiuric acid was : ortho 25-07, 
 meta 99-2, para 80-7. With the di-ortho-substituted compounds the effect was 
 so great that a more energetic treatment was required. After heating for 
 7 hours to 160° with 75 per cent, sulphuric acid, di-o-dibrom-benzamide was 
 only hydrolysed to the extent of 11-56 per cent. : whereas under the same 
 conditions over 99 per cent, of o-j3-dibrom-benzamide is broken up. In the 
 same way Menschutkin's results show that while ortho-substituents greatly 
 diminish the initial velocity of amide-formation they have much less efifect 
 on the percentage formed at equilibrium, which may even be increased : the 
 following are the numbers obtained for o- and w-toluic acid, with ammonia 
 and dimethylamine respectively : — 
 
 
 
 With NH3 
 
 With 
 
 NH(CH3)2 
 
 
 Initial 
 
 Per cent, at 
 
 Initial 
 
 Per cent, at 
 
 
 velocity 
 
 equUibrium 
 
 velocity 
 
 equilibrium 
 
 o-toluic acid 
 
 29-1 
 
 83-4 
 
 19-4 
 
 74-3 
 
 Mj-toluic acid 
 
 56-4 
 
 78-5 
 
 26-4 
 
 64-1 
 
 The hydrolysis of amides both by acids and by alkalies in aqueous solution 
 has been measured by Crocker,'' who determines the extent to which the 
 reaction has gone by means of the conductivity. In both cases the reaction 
 is shown to be bimolecular, i. e. the velocity is proportional to the product of 
 the concentrations of the amide and the acid or the alkali. The catalytic effect 
 of hydroxyl ion is much greater than that of hydrogen ion. 
 
 ' /. C. S. 1897, 229. 
 
 ' Crocker, J. 0. S. 1907. 593 ; Crocker, Lowe, ib. 952. Cf. Davis, J. C. S. 1909. 1397. 
 
78 Amides 
 
 This reaction for forming the amides, by heating the ammonium salt, or 
 a mixture of the acid and the amine, is the one most commonly used. The dry 
 ammonium salt cannot be made by evaporating down the aqueous solution 
 without great loss, as it dissociates. It is therefore usually prepared by 
 saturating the anhydrous acid with ammonia or ammonium carbonate. The 
 dry salt can then be distilled : but if so, there is much loss through dissociation 
 into ammonia and acid. It is better to heat it in sealed tubes to about 230° for 
 several hours : from the product, in the case of the lower acids, the amide is 
 separated by distillation ; in the case of the higher, it is found crystallized out. 
 
 It was by this method that the first known amide, oxamide, was prepared by 
 Dumas in 1830. 
 
 2. From the esters by the action of aqueous ammonia : this was used by 
 Liebig in 1834 for making oxamide : — 
 
 0=COC2H5 NHg _ O^CNHa HO-CaHg 
 0=COC2H5 "^ NH,j "" 0=CNH2 "^ HOCaH,, * 
 This goes weU with the lower and more soluble esters, such as formic, acetic, 
 and oxalic, sometimes even in the cold : the higher esters must be heated in 
 sealed tubes, and the yields are not good. 
 
 An interesting case of the failure of this reaction, apparently for stereo- 
 chemical reasons, has been discovered by E. Fischer.' Malonic ester easily 
 forms its diamide, CH2(CO-NH2)2, when treated vdth aqueous ammonia in 
 the cold. The mono-alkyl-malonic esters form their diamides (such as 
 C3H7CH(CO-NH2)2) almost but not quite as easily. But the esters of dialkyl- 
 malonic acid, even of dimethyl-malonic acid, (CH3)2C(COOH)2, can only be got 
 to yield amides with the greatest difficulty, and sometimes not at all. This is 
 the more remarkable since the chloride of dimethyl-malonic acid reacts with 
 ammonia with the greatest ease to ,give a perfectly stable diamide, 
 (CH3)2C(CO-NH2)2. This case has an obvious resemblance to that of the 
 di-ortho-substituted benzoic acids, and may probably be explained in the 
 same way : — 
 
 ""^^^^ CH,ArJ^CH3 
 
 0\ ^'-'\ /O ^KJ-^ ^ 
 
 Fischer, however, suggests a different explanation. The monalkyl-malonic 
 esters, like malonic ester itself, contain an acidic hydrogen atom, while the 
 dialkyl do not : and this may possibly be the cause of the inactivity of the 
 latter. Fischer supposes that the ammonia first forms a salt with the tauto- 
 meric form of the ester p/0 
 
 CH.,-c/^OEt , 
 " V/OEt 
 
 which then loses alcohol to give the amide. Of com-se a dialkyl-malonic 
 ester could form no such derivative. The explanation is ingenious, but it 
 can hardly be maintained. In the analogous case of the benzoic acids no such 
 
 ■ Ber. 35. 844 (1902). 
 
Formation of Amides : Stereo-hindrance 79 
 
 explanation is possible, whereas stereo-hindrance is a well established influence, 
 which is quite sufficient to account for the facts. It seems probable that in 
 these cases of amide-formation there are two influences at work, a chemical 
 influence and that of stereo-hindrance. It is the latter, for example, which 
 prevents the formation' of the amide from the ester of trimethyl-acetic acid, 
 and the former which causes it to be produced easily from that of trichloracetic 
 acid. 
 
 acid chloride or acid 
 
 3. 
 
 The action of ammonia or amines on th 
 
 Tin A ' 
 
 iju 
 
 .1 lUC . 
 
 (CH3-CO)20 + 2NH3 = CHgCO-NHa + 
 
 
 CH3.CO.CI + 2NH3 = CH.CO.NH^ + 
 
 CHs-COONHi 
 NHiCl. 
 
 By this reaction Liebig and Wohler in 1832 prepared benzamide from benzoyl 
 chloride. With the higher acids it is better to use the acid chloride. 
 
 If one molecule of acid chloride is treated with one molecule of each of 
 two amines, the amide of the more negative amine is produced and the 
 hydrochloride of the more positive." 
 
 To this reaction is probably to be referred a singular method of preparing 
 amides discovered by Orton.' This consists in treating a dUute ammoniacal 
 solution of the acid with benzoyl chloride, and shaking with soda till the smell 
 of the benzoyl chloride disappears, when the amide separates mixed with a 
 certain quantity of benzamide. It seems probable that the benzoyl chloride 
 reacts with the acid either to give the acid chloride or to give the acid anhydride, 
 and then this acid chloride or anhydride acts on the ammonia to form the amide. 
 It is remarkable that unless excess of ammonia is used the amount of benzamide 
 formed is very small. 
 
 All these methods may be used for preparing alkylated amides by sub- 
 stituting alkylamines for ammonia, and for the aromatic derivatives by using 
 the anilines. 
 
 4. By the addition of water to the nitriles : — 
 
 CH3C=N + H2O = CHj-CONHa. 
 This may be done by dissolving them in concentrated sulphuric acid, by 
 shaking them with concentrated hydrochloric acid, or most easily by the 
 action of hydrogen peroxide in alkaline solution : — 
 
 C5Hii-C=N -f 2H2O2 = C5H11CO.NH2 + O2 -t- H2O. , 
 
 This is a remarkable reaction, as it appeai-s to be due to thejiasegat jvatej,. • 
 
 5. Alkylated amides can also be obtained by the action of acids on 
 isocyanic esters (as water gives amines): — 
 
 CH3-C00H + CaHjN^C^O = CHoCONHC^H, + CO2. 
 
 Properties of the amides 
 
 All the amides except formamide are crystalline solids. The lower are 
 excessively soluble in water and are deliquescent: they can be distilled un- 
 
 ' H. Meyer, Mon. 27. 31 {C. 06. i. 1235). 
 
 = Dains, J. Am. Cli. Soc. 28. 1183 (C. 06. ii. 1186). = J. C. S. 1901. 1352. 
 
80 Amides 
 
 changed. They generally have an unpleasant smell, but this is due to 
 impurities. Pure acetamide (like pure iodoform and pure skatole) has no 
 smell. 
 
 Though the amines have much lower boiling-points than the corre- 
 sponding alcohols, the amides boil at much higher temperatures than the 
 acids from which they are derived. Thus acetic acid boils at 118°, acetamide 
 at 223° : the difference diminishes in ascending the series. 
 
 The amides are comparable in constitution to the primary amines, but differ 
 from them markedly in behaviour. The electro-negative acyl group makes 
 them indifferent substances of neutral reaction. The basic character of the 
 NH2 group is not wholly destroyed, and they are able to form salts with 
 strong acids : but these have a strongly acid reaction and are highly hydro- 
 lysed. The alkylated amides are more basic and can give stable platini- 
 chlorides. In consequence of the more negative character of the amides as 
 compared with the amines, the hydrogen is more easily replaced by metals. 
 Thus an aqueous solution of an amide dissolves mercuric oxide, and on 
 evaporation salt-like bodies such as Hg(NH-CO-CH3)2 crystallize out. This is 
 the general behaviour of NH attached to negative groups, as shown in isocyanic 
 acid, H-N=C=0, and in phthalimide. 
 
 This formation of salts raises the question of the true formula of the 
 amides. Besides that which has been hitherto assumed, there is the possibility 
 of the tautomeric formula of an iso-amide or imino-hydrin : — 
 
 Normal amide Iso-amide 
 
 Alkyl derivatives of both forms are known. Those of the first are the normal 
 alkyl-amides, those of the second are the so-called imino-ethers. They can 
 both be obtained from, and converted into, the normal amides, whose 
 behaviour is thus to some extent tautomeric. But the balance of evidence 
 with regard to the free amides is strongly in favour of the first (normal) formula. 
 This conclusion is arrived at by Auwers ' from their cryoscopic behaviour ; by 
 Claisen "^ from a comparison with the oxymethylene compounds ; by Schmidt ' 
 from spectroscopic investigations ; and by Hantzsch * from the fact that they do 
 not form salts with ammonia in non-dissociating solvents, as bodies with an 
 acidic hydroxyl group would be expected to do. 
 
 The structure of the metallic derivatives is much less easy to determine. 
 It is quite conceivable that while the free amides have the normal structure, 
 
 their salts may be derived from the iso-formula E-C<^2tjj, the metal going, as 
 
 it so often does, into the more acidic position. The evidence is not conclusive. 
 The mercury compounds appear to be the normal N-salts, E-CO-NHhg." They 
 have the tendency, characteristic of compounds containing mercury joined to 
 nitrogen, not to dissociate in water ; and hence their solutions give very few of 
 the reactions of the mercuric ion. But this is very little guide to the constitution 
 
 ' Ber. 34. 3558 (1901). ^ Ann. 287. 360 (1895). = Ber. 36. 2459 (1903). 
 
 * Ber. 35. 226 (1902). <> Ley, Schafer, Ber. 35. 1309 (1902). 
 
Constitution of the Amides 81 
 
 of the salts of other metals, as we know that mercury has a very strong tendency 
 to combine with nitrogen. The silver salts generally give 0-ethers (imino- 
 ethers) when the metal is replaced by alkyl, and have therefore been assumed 
 
 to be iso-salts E-C<(j^jj^. But there is plenty of evidence to show how 
 untrustworthy this reaction (the replacement of the metal by alkyls) is for 
 determining the constitution of metallic derivatives. It seems probable that 
 the silver salts can exist in two isomeric forms. Thus Titherley ^ has shown 
 that besides the ordinary white silver benzamide, which gives imino-ethers, 
 
 and so may possibly be ^-C^j^g^, there is an orange silver salt, formed by 
 the action of alcoholic silver nitrate on potassium benzamide, which may be 
 
 <^-<HAg- 
 
 The structure of the alkaline salts is equally uncertain. Those of the 
 aromatic amides react readily with alkyl iodides to give 0-ethers, while those 
 of the fatty amides will hardly react at all. This certainly seems to indicate 
 that there is a difference of constitution between them : and Titherley considers 
 
 that the alkaline salts of the aromatic amides are 0-salts, Ar-C^jq^jj* while 
 
 those of the fatty amides are N-derivatives, as CHg-Cxj^gj^g^. But the 
 question cannot yet be regarded as settled. 
 
 One of the most marked differences between the amides and the amines 
 is that the nitrogen is much less firmly attached to the carbon in the amides. 
 (It is a general rule that an increase in the negative character of a molecule 
 weakens the linkages within it. Compare for the C-O-C grouping, the ethers, 
 esters, and acid anhydrides: for C-C, such cases as oxalic and trichloracetic 
 acids, &c.) The amides can be saponified even by heating with water, their 
 formation from acid and ammonia being, as has been mentioned, a reversible 
 reaction : and much more easily by acids and alkalies. This reaction has been 
 used by Ostwald to measure the concentration of hydrogen ion, that is, the 
 strength of acids. 
 
 A similar reaction occurs when the amides are heated with alcohols. In 
 this case the reaction may go in two ways : the alcohol radical may attach itself 
 either to the acid or to the nitrogen : — 
 
 CHgC'" + CHjOH'^ 
 
 ^^H^ ^ ^CH..CC ^ . NH, 
 
 ^0 ^CH3C<^H + C^H^^H, 
 
 -OC2H5 
 As a matter of fact both reactions occur. 
 
 The hydrolysis of the amides by certain natural ferments, such as pepsin 
 and trypsin, has already been mentioned in discussing the polypeptides ; it is 
 not confined to those amino-acid derivatives, but is shared by the simple amides.' 
 
 With dehydrating agents such as phosphorus pentoxide the amides give 
 
 ' .7. C. S. 1897. 468 ; 1901. 891. 
 
 ' Gonuermann, C. 02. i. 909 ; 03. i. 960. 
 
82 Amides 
 
 nitriles. This makes it possible to go both ways along the series of reactions :— 
 boiling water H2SO4 or H2O2 
 
 CH3C<gj,H, ^ : CH3<H2 ^ H2O - : CH3.C=N H- 2 H^O. 
 
 heat ^2^6 
 
 The simpler fatty amides, like acetamide, when acting as solvents behave 
 somewhat like water. Pure acetamide is a conductor of electricity, and is no 
 doubt ionized to a small extent. Solutions of salts in acetamide also conduct,' 
 though to a less extent than in water. In the same way the salts of weak bases 
 are 'amidolysed' by formamide, just as they are hydrolysed by water.^ For 
 example, antimony trichloride dissolved in formamide gives a white precipitate 
 composed of a mixture of SbCl^X, SbClX^, and SbXg, where X = NH-CHO. 
 So, too, it can combine with salts, acting like water of crystallization.' In 
 solution the amides are commonly associated, some of them even in water.* 
 
 The conversion of the amides by the action of alkali and bromine into 
 the amines has ah-eady been mentioned as a mode of preparing the latter. The 
 various stages of this singular reaction have now been made out. The amide 
 is first treated with bromine and potash (commercially chlorine and potash or 
 chlorine and soda are used), whereby it is converted into the monobromamide : — 
 
 CH:..C<gH, + B'-^ = CH3C<gjjg^ + HBr. 
 This is a very reactive substance. If it is treated with ammonia or aniline 
 a violent reaction occurs, and the amide is regenerated. The bromine atom 
 increases the acidity of the amide, so that the remaining hydrogen atom is 
 more easily replaced by metals. Hantzsch" has shown that the bromamide 
 is a pseudo-acid : that is to say, though it is not itself an acid it gives with 
 bases the salts of an acidic tautomeric form, the iso-amide. Thus the potash 
 with which it is treated in the Hofmann reaction forms the potassium salt 
 
 /OK 
 of the iso-bromamide, CHs-C^j^g^, The alkaline liquid is then distilled, 
 
 when the following changes take place. In the first place the potassium salt 
 contains the peculiar C=N linking, which also occurs in the oximes. Now the 
 oximes are capable of undergoing the Beckmann reaction, whereby they are 
 converted into the amides : — . - ■ --- - , 0..^-,^ _>,^, , , „ , 
 
 (b-C-K- HO-C-H 0=C-H ' ■ 
 
 ^1^ _,. II -> I . f'-'--'^ 
 
 Hp;N 0-N ^-N-H ^„,.,,,.r,;,-, 
 
 benzaldoxime giving fomialdshycle.i. The essence of this reaction, which 
 
 involves a very unusual insertion of a nitrogen atom into the middle of a 
 
 carbon chain, is that the two groups on the same side of the C=N change 
 
 places. We may suppose the same change to occur with the similarly 
 
 constituted potassium salt of the bromamide : — 
 
 CH.COK BrCOK ^^^ C=0 
 
 ^11 = II = KBr + ^„ II ■ 
 
 Bi-N-fcT CH;,N CHg-N 
 
 ' J. W. Walker, Johnson, /. C. S. 1905. 1597. 
 
 ' Bruni, Manuelli, Z.f. Elektrocliem. 11. 554 (C. 05. ii. 873). 
 
 = Mensohutkin, C. 09. i. 900. • Meldrnm, Turner, /. C. S. 1908. 876. 
 
 ' £er. 35. 226, 8579 (1902). 
 
Amides : Hqfmann Reaction 83 
 
 In the primary product the bromine and the potassium are on the same 
 carbon atom, and they therefore split off together as potassium bromide, 
 leaving an isocyanate. If the bromamide is treated with silver carbonate the 
 isocyanate can actually be isolated; but in the ordinary Hofmann reaction, 
 where excess of alkali is used, it cannot be detected, and its formation has 
 been denied. It has, however, been shown ^ that if the bromamide is treated 
 with one equivalent of alkali, and the solution distiUed with steam, the 
 isocyanate passes over. If the alkali is in excess, the isocyanate does not 
 appear as such, but is immediately transformed, through the intermediate 
 stage of the carbamate, into the amine and carbon dioxide : — 
 
 E.N=C!=0 + KOH -^ ENHCOOK -^ ENH^ + OO2. 
 At the same time (and especially if the alkali is not in excess) some of this 
 amine reacts with the unchanged isocyanate to form the di-substituted urea 
 C0(NHE)2. 
 
 Thus Hantzsch's discovery of the constitution of the bromamide salts 
 supplied the link necessary to the explanation of this remarkable reaction, 
 by establishing its exact analogy to the Beckmann reaction. To recapitulate, 
 the following changes take place, in which all the stages except the one in 
 brackets have been isolated : — 
 
 CH3-C=0 CHgC^O CHa-COK r BrCOK-i 
 — - ~* -'- — ~* ^LcHoN J 
 
 NH., NHBr BrN 
 
 ^■3-^ 
 
 0=0 KOC=0 
 
 CH.NH,. 
 CHa-N CHa-NH 
 
 There is a remarkable method, not fully understood, of forming the amide 
 of phenyl-acetic acid and its homologues. It consists in acting on the mixed 
 fatty-aromatic ketones with yellow ammonium sulphide. Thus aeetophenone is 
 converted into phenyl-acetamide : — 
 
 ^COCHg -» ^CHa-CO-NHa. 
 
 The secondary and tertiary amides, such as diacetamide, NH(CO-CH3)2, may 
 be got by boiling the primary amides with an acid anhydride : — 
 
 or by acting '' on the monamides with acid chlorides ; this method can be used 
 for preparing the mixed secondary amides : — 
 
 C.H.C<gH^ + CH3-C<gi = ^h;:^5nH + HCl, 
 
 or by heating a nitrile with an acid : — 
 
 CH3CEN + CHgCO-OH = (CH3C0),NH. 
 They are unstable bodies, which easily lose one acyl group on hydrolysis, 
 passing back into the primary amides. 
 
 1 Mohr, J. Pr. Ch. (2) 72. 297 ; 73. 177, 228 (C. 05. ii. 1534 ; 06. i. 1091, 1153). 
 2 Tarbourieoh, (J. 03. ii. 552, 712. 
 
84 Amides 
 
 AMIDES OF DIBASIC ACIDS 
 
 These may be of two kinds, according as one or both of the hydroxyls 
 are replaced by NHj. Bodies of the former kind, which still contain a 
 carboxyl group, are known as -amic acids (oxamic, succinamic, &c.). Thus 
 if acid ammonium oxalate is heated, the monamide, oxamic acid, is formed : — 
 0=gOH 0=COH 
 
 T = T + H„o. 
 
 0=CONH^ G^CNHg ^ 
 
 This is a white powder, slightly soluble in water. It is an acid, and hence can 
 
 give salts and esters. If its ethyl ester, known as oxamethane, is heated with 
 
 phosphorus pentoxide, it loses water and forms cyan-carbonic ester : — 
 
 O^CNH, GEN 
 
 1 ^ = T + H2O. 
 0=COEt 0=COEt ^ 
 
 On the other hand, if neutral ammonium oxalate is heated, or if ethyl 
 
 0=9NH2 
 oxalate is treated with ammonia, the diamide, oxamide, \ , is pro- 
 
 O—G'Pi Jig 
 
 duced. This is a crystalline almost insoluble powder, which on heating to 
 
 a rather high temperature sublimes partly unchanged, partly with decomposition 
 
 into water and cyanogen. This last change takes place readily and completely 
 
 if oxamide is heated with phosphorus pentoxide : — 
 
 0=CNH, OEN 
 
 ^2 
 
 = ? 
 
 I = T +2 H,0. 
 
 0=CNH2 CEN ' 
 
 The insolubility and high melting-point of oxamide point to its having 
 a high molecular weight: and this is confirmed by measurements of its 
 molecular weight in solution,' which give values nearly twice that of the 
 simple formula. 
 
 The diamide of succinic acid is remarkable for existing in two forms. 
 
 CTT •OO-'N^TT 
 
 When the ester is treated with ammonia the normal diamide, I ^ ^ , is 
 
 CHyCONHa' 
 
 produced. This body loses ammonia with singular ease to form a ring 
 
 compound, succinimide: — 
 
 I ^ ^NH^ = NH3 -f NH. 
 
 The ease with which this reaction takes place no doubt depends on the 
 stability of the 5-ring produced, and may be compared with the ease with 
 which succinic acid itself forms an anhydride. 
 
 The isomeric diamide is got by the action of ammonia on succinyl chloride. 
 
 CH2-C-NH2 
 The only structure possible for it is | ^0 ' ^^^'^^ indicates that the 
 
 CHa-C^Q 
 
 ' Maselli, Gas. 37. ii. 135 (C. 07. ii. 1234). 
 
Amides of Dibasic Acids 85 
 
 CH -C^^^^ 
 chloride is \ ^\0 , a formula supported also by other evidence. 
 
 In ordinary (ortho) phthalic acid the relative position of the two carboxyls 
 is the same as in succinic acid : — 
 
 K-^.COOH CH^COOH 
 
 UCOOH CH,.COOH- 
 
 o-phthalic acid Succinic acid 
 
 Accordingly the acid shows the same tendency to form an anhydride ; and here 
 too there is reason to think that the chloride has the unsymmetrical formula 
 
 v'x^/O • If so, the diamide got from it should also be unsymmetrical, and so 
 
 should the imide, phthalimide, which is readily obtained from the diamide : — 
 
 rr^l^^ _ Q-C<o i,3tead of (X )Sg^ ^ Pr^V 
 
 In this case only one of the two theoretically possible diamides is known, 
 and we cannot tell for certain which of the two formulae should be assigned 
 to it. Phthalimide forms a potassium derivative (which has already been 
 mentioned in discussing Gabriel's amine synthesis), and this when treated with 
 
 methyl iodide certainly gives a symmetrical methyl-phthalimide [ j ^NCHo, 
 
 which is in favour of the view that the imide, and therefore also the diamide, 
 has a symmetrical structure. On the other hand, we have by this time 
 plenty of evidence to show that conclusions as to the structure of tautomeric 
 compounds drawn from the behaviour of their metallic derivatives must be 
 received with great caution. The question must therefore be left open. 
 The most we can say is that in the case of phthalyl chloride the evidence 
 is in favour of an unsymmetrical, and in that of the diamide and imide it is 
 in favour, though more doubtfully, of a symmetrical structure. 
 
 CH2CONH2 
 The monamide of amino-succinic (aspartic) acid, I > which is 
 
 OHWHg'OO-OH 
 
 known as asparagine, occurs in nature in germinating seeds, in two forms, 
 a dextro and a laevo, which do not racemize but crystallize separately. The 
 interesting point about them is that while the ordinary (laevo) form is tasteless, 
 the dextro is sweet. Now it is only bodies which themselves contain an 
 asymmetric carbon atom — optically active bodies — which can react differently 
 with two optical antimers. Hence Pasteur has said that in this case the 
 nerve substance of the tongue acts as an optically active body. 
 
 The formation of cyclic diamides ^ from the dibasic acids and the phenylene 
 
 > K. Meyer, Ann. 347. 17 (1906). 
 1175 G 
 
86 Amides 
 
 diammes is of some interest. The ortho-diamines give diamides of the type 
 )E, with all the dibasic acids from oxalic, which gives a 6-ring, 
 
 a 
 
 NH-C<^ 
 
 NH-C<o 
 
 to sebacic, which gives a 14-ring. There is some indication that as the ring 
 gets larger it is formed with more difficulty. 
 
 Meta and para-diamines might form compounds of the same type, such as 
 
 m— 0=0 
 
 
 
 E 
 
 NH— C=0 
 
 But in no case are such bodies produced. Instead, only one NHg is attacked, 
 
 and a cyclic imide r^N<'pf^')E is formed. 
 
 
 ANILIDES 
 
 These are a class of substituted amides of considerable importance in which 
 one amide hydrogen is replaced by phenyl. They are derived from aniline in 
 the same way as the ordinary amides are derived from ammonia. 
 
 They are made by heating aniline with the acid. A salt is no doubt 
 formed at first, and then this breaks up with loss of water, as an ammonium 
 salt does to form an amide, but generally with much more ease. 
 
 CH3C<Q jjjjg^ = H^O + CH3-C<j^jj^ • 
 
 Menschutkin ' has measured the velocity of formation of acetanilide in 
 acetic acid. He finds it to be bimolecular in the absence of a catalyst ; but 
 in the presence of a halogen salt of aniline as catalyst it becomes mono- 
 molecular, the iodide having the greatest accelerating effect, and the chloride 
 the least. Ortho substituents on the aniline diminish the velocity, while meta 
 and para greatly increase it. H. Goldschmidt^ and his pupils have shown 
 that the same results are obtained if the reaction is carried on at 110° in 
 presence of excess of amine. They find that with different acids the ratio 
 of the velocities with any amine is approximately — 
 
 Formic Acetic Propionic Butyric Isobutyric 
 
 1000 1 0-5 0-3 01 
 
 The reaction of aniline with any acid is from two to three times as quick 
 as that of ortho-toluidine, 
 
 Formanilide is rapidly produced even if dilute formic acid is boiled with 
 aniline. To prepare acetanilide the anhydrous acid must be used. In making 
 this compound, even in the laboratory, it is convenient to use a reverse air 
 condenser with a side tube at the top. The liquid is boiled at such a rate 
 
 1 C. 06. i. 551 ; of. 03. ii. 324. 
 
 » Goldschmidt, Wachs, Z. Ph. Ch. 24. 353 (1897) ; Goldschmidt, Brauer, Ber. 39. 97 (1906). 
 
Anilides 87 
 
 that the top of the condenser keeps at 100-110°. The water formed in the 
 reaction distils over with a very small quantity of acetic acid ; the concen- 
 tration of the liquid is thus kept up, and the reaction goes much more rapidly. 
 This method is also used on the large scale. 
 
 A curious modification of this reaction is to use instead of the acid the 
 thio-acid, for example thioacetic acid, CHgCO-SH. The reaction is similar 
 to that of the acid itself, but sulphuretted hydrogen is formed instead of 
 water =— n r> 
 
 CH3.C<^jj +HNH0 = CH3.C<gjj^ + h,S, 
 
 and this is evolved, instead of remaining in solution and retarding the 
 reaction. 
 
 The anilides can also be formed with great ease by treating the acid chloride, 
 anhydride, or ester, or the amide, with aniline. In the case of the negatively 
 substituted anilines, which do not react easily with acyl anhydrides, the 
 formation is promoted by the presence of strong acids, such as sulphuric' 
 
 The anilides are colourless crystalline compounds, slightly soluble in water. 
 Like the amides they split off the acyl group when boiled with acids or 
 alkalies, but unlike the amides they will not do so with water alone. 
 
 The anilide and toluidide of formic aeid^ exist in two forms, one an oil 
 easily soluble in potash, and the other a solid insoluble in potash. It is 
 probable that these are derivatives of the two tautomeric amide formulae, 
 the oil being ArN=CHOH, and the solid ArNHCOH. 
 
 The anUides form metallic derivatives which, like those of the amides, 
 may have either of two tautomeric formulae. Thus formanilide gives with 
 alcoholic soda a precipitate of the mono-sodium derivative, which may have 
 
 the structure H-C<^i^j-|^ > since with methyl iodide it forms a compound 
 which must be H-C^jq- j njr , as on hydrolysis it yields methyl aniline. Silver 
 formanilide gives with methyl iodide an iso-(0-)-ether H-C<^„ . ^, but on the 
 
 other hand with benzyl chloride it gives an N-ether H-C^^, .,pTT .a, which 
 
 shows how untrustworthy such reactions are as proofs of constitution. 
 
 Alkyl-aniUdes can be made by treating the sodium-anUides with alkyl 
 iodide ; aryl-anilides, by acylating the secondary aniline bases, but this is 
 much less easy than acylating aniline itself. For example, diphenylamine 
 cannot like aniline be acylated by dilute formic acid, but only by concentrated : 
 and not by acetic acid at all, but only by the acid anhydride or chloride. 
 
 The diacyl-anilines are little known. They split off one acyl with extreme 
 ease. Thus diacetanilide {C'H.^-GO)^<l) (obtained from acetanilide and acetic 
 anhydride or chloride) is more than half hydrolysed into acetanUide and acetic 
 acid by boiling for an hour with N/100 soda. If they are heated alone for some 
 time to a rather high temperature, one of the acyl groups migrates to the ring, 
 
 1 Smith, Orton, /. 0. S. 1908. 1242 ; 1909. 1060. '' Orlow, C. 06. ii. 403. 
 
 g2 
 
-COCH3 
 
 
 
 -COCHg 
 
 
 
 
 
 
 88 Amides 
 
 diacetanilide, for example, giving ^-acetamino-acetophenone : — 
 
 
 
 CO-CHg 
 
 The acyl group goes mainly to the para but partly also to the ortho position.' 
 
 THIOAMIDES 
 
 These are amides in which the oxygen is replaced by sulphur. The simple 
 thioamides are little known. They may be made by acting on the amides 
 with phosphorus sulphide, or on the nitriles with hydrogen sulphide : — 
 
 CH,.C=N ^„ ^^S 
 
 -2 " ^NH,' 
 
 
 This reaction is formulated on the analogy of the amides. It is, however, 
 
 SH 
 more probable' that the thioamides have the tautomeric iso-formula E-C\jjjj- 
 
 This is indicated by a variety of reactions in which alkyl and other groups 
 become attached to the sulphur. Thus with alkyl halides they often give 
 
 sulphur ethers E-C\gjj • Again, with chloracetone they form thiazoles, joining 
 
 up through both sulphur and nitrogen, which could scarcely happen unless 
 both had hydrogen attached to them : — 
 
 .TJTT HOCCH, ^NCCH, 
 
 CH3C<^" + ^, H ' = CH3.C< II ^ + H^O + HCl. 
 ^SH CICH ^ ^S-CH 
 
 Thioformamide has recently been obtained by Willstatter and Wirth,' by 
 the action of phosphorus pentasulphide on formamide. It forms colourless 
 crystals melting at 28° to a yellow oil. It is unstable and forms an unstable 
 salt with one molecule of hydrochloric acid. It is tautomeric, reacting according 
 
 to the two formulae H-C<^jj^tt and H-C\jq-Tj • It has an acid reaction, while 
 
 its di-substitution-products, such as diphenyl-thio-formamide H-C\jq-j, , which 
 
 cannot assume the imide form, have not. This suggests that the free thio- 
 formamide consists largely of the imide form. It gives a remarkably stable 
 hydrate H-CS-NHa-HaO. 
 
 Thioformamide is very reactive, and combines, for example, with chloracet- 
 aldehyde to thiazol :— QH^Cl HS _^ CH— ^ 
 
 I + 
 
 CHO CM (JH CH' 
 
 and with bromethylamine to thiazoline : — 
 
 CHaBr HS QHa— S 
 
 1 4- I _ 1 I 4. 'NTT "Rv 
 
 CH2NH2 + OH - CH2 CH + J^i^4«r. 
 
 hn/ \]sr/ 
 
 • Chattaway, Proe. C. S. 1902. 173 ; 1903. 57, 124. " 
 
 ' But see Eiilmann, C. 07. ii. 1778. » Ber. 42. 1908 (1909). 
 
Thioamides 89 
 
 It might be expected that as formamide on heating gives carbon monoxide, 
 so thioformamide would give the unknown carbon monosulphide CS ; but this 
 is not formed, the main products of the decomposition being hydrogen sulphide 
 and prussic acid. 
 
 The thioanilides are better known. They are most conveniently made by 
 fusing the anilides for a short time with phosphorus pentasulphide. 
 
 Thioformanilide may also be made by the action of hydrogen sulphide on 
 phenyl isocyanide : — 
 
 CSN.<^ + H,S = H.C<|jj^ or H-C^^ . 
 
 ThiobenzaniUde is the product of the action of phosphorus pentasulphide 
 on benzophenone oxime, no doubt through the Beckmann reaction ' :— 
 
 N-OH N-SH N-^ 
 
 The thioanilides differ from the anUides in having an acid character. They 
 form alkaline salts which are not decomposed by water, and hence dissolve 
 in aqueous alkalies, though they are reprecipitated by carbon dioxide. This 
 might be due to the fact that sulphur is more acidic than oxygen, as is shown 
 by a comparison of the mercaptans with the alcohols. There seems, however, 
 to be no doubt that the sodium salts have the metal attached to sulphur ; for 
 whenever it is removed they always join up through the sulphur. Thus the 
 
 sodium salts with ethyl bromide give the thioethers CHg-C^jj^ . • The products 
 
 are proved to have this formula (1) by their being hydrolyzed by dilute acids 
 to aniline and the esters of the thioacids : — 
 
 CH3.C<|^* + H,0 = CH3-C<|^* + H,N.^, 
 
 and (2) by their giving amidines and mercaptans when treated with amines : — 
 
 CH3-C<|^* + H,N<^ = CH3-C<5;^^ + HSEt. 
 
 Thus in both reactions a compound is formed in which the ethyl is attached 
 to sulphur. 
 
 Another proof that the alkaline salts of the thioanilides have the metal 
 attached to sulphtir is that when they are oxidized v?ith potassium ferricyanide 
 a new ring is formed through the sulphur, giving a benzothiazol : — 
 
 We may therefore assume ^ that the salts have the formula E-C<(j^0- If the 
 
 thioanilides themselves are analogous to the anilides, they must be pseudo-acids : 
 that is, they must undergo a tautomeric change when they form salts : — 
 
 1 Ciusa, C. 07. i. 28. 
 
 " The evidence, however, is not conclusive. See Biilmann, Ann. 364. 314 (1909). 
 
90 Amides 
 
 That this is so is quite possible, especially in view of the fact that they 
 react with hydroxylamine to give anilido-oximes : — 
 
 
 CH; 
 
 IMIDES 
 
 C'Hg-C^jfjj^ + 
 
 H^S. 
 
 The 
 
 imides are 
 
 the cyclic secondary amides of the dibasic acids 
 
 
 
 T,/COOH , TT TST _^ 
 
 E<^HH. 
 
 
 Their formation is practically confined to those acids in which the two carboxyls 
 
 are separated by a chain of 2 or 3 carbon atoms ; that is, where the imide 
 
 contains a 5- or a 6-ring. This is analogous to the formation of the cyclic 
 
 anhydrides, which, however, are also produced by acids with still longer chains, 
 
 which do not give imides. It is an example of the truth of Baeyer's strain 
 
 theory, that the rings most easily formed are those containing 5 atoms, and next 
 
 those of 6. 
 
 CO 
 An imide of oxalic acid 1 ^NH has been described,' but its existence 
 
 is doubtful, and its structure still more so. Malonic acid gives no imide, nor 
 do adipic and the higher members of the series ; but succinic and glutaric acids 
 form them with great ease. 
 
 Among the phthalic acids it is only the ortho which gives an imide. 
 Iso-(meta)-phthalic acid, though it has Hke glutaric acid three carbon atoms 
 between the carboxyls, gives neither an imide nor an anhydride, evidently 
 because the rigid structui-e of the benzene ring holds the carboxyls further 
 apart than in an open-chain compound. This view is confirmed by the fact 
 that homophthalic acid, which like isophthalic acid has three carbons between 
 the carboxyls, but has one of these in a side chain, forms both an imide and 
 an anhydride. The structure of this imide is established by its conversion, 
 on treatment with phosphorus oxychloride and subsequent reduction, into 
 isoquinoline : — 
 
 ^O '^O CH^^ 
 
 Phthalimide Homophthalimide Isoquinoline 
 
 The imides in general are formed by heating the ammonium salts, the 
 diamides, or the monamides (-amic acids) of dibasic acids: — 
 
 I i:^2 = NH3 + I NH, 
 
 CH2-C<g^2 CH2-C<Q 
 
 or most conveniently by heating the anhydrides in a stream of ammonia. 
 ' Ost, Mente, Ber. 19. 3228 (1886). 
 
Imides 91 
 
 They may also be got by the partial hydration of the dicyanides,' such as 
 ethylene dicyanide. 
 
 They are easily hydrolysed by alkalies to give first the -amic acid, and 
 then the acid itself: — 
 
 (jIH^-C^^ _^ CH2COOH CH2COOH 
 
 CH2.C<Q ~^ CH2CONH2 ~* CHa-COOH* 
 
 On reduction succinimide is converted into pyrrol derivatives; thus on 
 distillation over zinc dust it gives pyrrol itself, a reaction which is most 
 clearly represented as taking place through the tautomeric dienolic form: — 
 
 This reaction proves that succinimide has the symmetrical formula given 
 
 above, and not the possible unsymmetrieal structure 1 ^ /O . 
 
 CHg-C^Q 
 
 The combined influences of the two carbonyl groups and of the ring structure 
 render the imide hydrogen distinctly acidic ; and succinimide forms a potassium 
 salt even with alcoholic potash. A silver salt is also known, whose electrical 
 behaviour is peculiar.^ The concentration of silver ion in its aqueous solution, 
 as measured by the E. M. F., is only half that required by the conductivity. 
 This indicates that it breaks up to give some other cations than silver. 
 
 CHg-C^O 
 Glutarimide CH2<f "^NH resembles succinimide. It can be obtained ' 
 
 by the oxidation of piperidine with hydrogen peroxide, and on distillation 
 over zinc dust it gives a small quantity of pyridine : — 
 
 ^CH^-CHa ^CHa-C^O ^CH-CH 
 
 CH2 >NH -> CH2 >NH -» CH >N. 
 
 \CH2-CH2 \CH2-C=0 \CH=CH 
 
 Piperidine Glutarimide Pyridine 
 
 Phthalimide, the imide of ortho-phthalic acid, is a body of great practical 
 importance, both for Gabriel's synthesis of amines, which has already been 
 so frequently mentioned, and also as an intermediate compound in the 
 commercial preparation of indigo. It can be made by acting with ammonia 
 on phthalic anhydride, and is also formed * by a curious intramolecular change 
 when o-cyan-benzoic acid is heated to its melting-point (180-190°) : — 
 
 1 Bogert, Eocles, /. Am. Ch. 80c. 24. 20 (C. 02. i. 711). 
 
 2 Ley, Schafer, Ber. 39. 1259 (1906). ' Wolffenstein, Ber. 25. 2777 (1891). 
 * Hoogewerff, van Dorp, Bee. Trav. 11. 81 (1892). 
 
92 Amides 
 
 Its imide hydrogen is sufficiently acidic for it to dissolve in aqueous potash. 
 If the potassium salt is treated with alkyl iodide, a symmetrical alkyl 
 
 _rL>N-Alk is produced. 
 Ho 
 
 Although no unsymmetrical isomeric phthalimide f ]_~/0 is known, its 
 
 alkyl derivatives can be obtained by treating the alkyl-phthalamio acids with 
 acetic anhydride, which acts merely as a dehydrating agent ^ : — 
 
 Cr^^o = H,o + (Yyo . 
 
 AMIDO-CHLOEIDES. IMIDO-CHLOEIDES. IMIDO-ETHEES 
 
 The last two of these classes form a group of substances which are 
 derived from the iso-amides, having the NH group attached to carbon by 
 a double bond. 
 
 The amido-cMorides themselves are derived from the normal amides, and 
 are obtained by acting on them with phosphorus pentachloride : — 
 
 CH3.C<gjj^ + PCI, = CH3-C<g^|j^ + POCI3, 
 
 the action being analogous to that of phosphorus pentachloride on aldehydes 
 
 and ketones. 
 
 The amido-chlorides lose hydrochloric acid with the greatest ease to form imido- 
 
 /Cl 
 chlorides, such as CH3-C<^-j^tt • These bodies are rather more stable, but they 
 
 readUy lose a second molecule of hydrochloric acid to form nitriles. Water 
 converts the imido-chlorides into amides and hydrochloric acid : — 
 
 CH3.C<g\j + HOH = HCl + CH3.C<gg -* CH3-0<gjj 
 
 ■2 
 
 The imido-chlorides are really the acid chlorides of the imido-acids or iso- 
 
 amides E-C<^-|^-pT, which, however, seem to be incapable of existence, and to pass 
 
 over as soon as they are formed into the more stable amides. 
 
 In the case of the a-oxy-acids, such as glycollic acid, it was formerly 
 supposed that the true imido-acids could be isolated. Glycollic acid gives a 
 normal amide CH2OHCONH2 ; but if its anhydride is treated with ammonia, 
 or if its imido-ether is saponified, an isomeric substance is obtained,^ which 
 
 was originally assumed to be the iso-amide or imidohydrine CHgOH-C'^S^" • 
 
 Analogous compounds have been obtained from other a-oxy-aoids. The subject 
 has been reinvestigated by Hantzsch and Voegelen,^ who have shown that 
 the properties of these substances are incompatible with this structure — and, 
 indeed, as far as one can see, with any structure. They cannot be converted 
 
 ' Hoogewerif, yan Dorp, Bee. Trav. 13. 93 (1894). 
 
 ' Eschweiler, Ber. 30. 998 (1897). ' Ber. 34. 3142 (1901). 
 
Amido- Chlorides, Imido- Chlorides 93 
 
 into the amides, a change which one would expect to occur with the utmost 
 ease, and their molecular weight is twice that required by the simple formula. 
 Their other properties are also most extraordinary. Though they are only weak 
 bases, and are not acidic at all, their solutions are very good conductors of 
 electricity, so that their electrolytic behaviour is that of salts. Moreover, 
 though they are fairly stable to acids, they are easily decomposed by certain 
 salts, such as calcium chloride, with the formation of a glycollate. No satis- 
 factory hypothesis has yet been proposed to account for these phenomena. 
 
 The esters of the imido-acids, the imMo-efhers, are much better known than 
 the acids themselves or their chlorides. The hydrochlorides of these imido- 
 ethers are formed by passing hydrochloric acid into a mixture of a nitrile and 
 an alcohol in molecular proportions, diluted with ether and well cooled. The 
 hydrochloride of a chlor-amido-ether is first produced : — 
 
 /NHa-HCl 
 CHs-Casr + C2H5OH + 2 HCl = CHa-C-OOaHg • 
 
 \oi 
 
 This is very unstable, and loses hydrochloric acid to form the hydrochloride 
 
 of an imido-ether GH.^-C^^^ Vr • The chlor-amido-ethers ' are much more 
 
 stable if the two hydrogens of the amino-group are replaced by two alkyls: 
 though even then they are decomposed by heat, giving first the imido-chloride 
 and one molecule of alkyl chloride, and then the nitrile and two molecules of 
 alkyl chloride : — 
 
 /NEtg-HCl ,-K!-pi. Txo^ 
 
 E-C-OC„H, = IS.Ci'^ff':^^^ + EtCl = EC=N + 2 EtCl + EtOH. 
 
 \C1 
 
 '%'-'-6 
 
 '^OC,H, 
 
 The hydrochlorides of the imido-ethers are crystalline bodies which on 
 heating lose alkyl chloride and form amides: — 
 
 CH3-C<gg^^^^ = CH3-C<g^2 + O2H5GI. 
 If treated with soda under ether they give the free imido-ethers, such as 
 CHg-C^^^jj , a liquid boiling at 97°. 
 
 The imido-ethers present a remarkable case of intramolecular rearrangement. 
 A variety of substituted imido-ethers are known, in which the imide hydrogen is 
 replaced by hydrocarbon radicals. They may be made from the simple ethers 
 by alkylation ; or by starting with an alkyl (or aryl) amide, converting this into 
 the imido-chloride, and treating this with sodium alkylate : — 
 
 TITTTA -KIA. NaOEt -KrA. 
 
 CH3-<H0 _^ CH3.</' . CH3.C<gg^jj^. 
 
 Now the imido-ethers behave on heating^ in different ways according as 
 the imide hydrogen is substituted or not. The simple ethers, where it is 
 not substituted, break up back again into nitrile and alcohol: — 
 
 CH3-C<gJjj = CH3C=N + C2H5OH. 
 ' V. Braun, Ber. 37. 2678, 2812 (1904). ' W. Wislioenus, Ber. 35. 164 (1902). 
 
94 Amides 
 
 If the imide hydrogen is replaced this reaction does not occur; but instead 
 of this, at about 200°, a rearrangement takes place, and a di-substituted 
 amide is formed: — /ff 
 
 This reaction is of particular interest as illustrating an intermediate stage 
 between an ordinary (isomeric) and a true tautomeric change. The imido- 
 acids (iso-amides) go over spontaneously into the normal amides: — 
 
 This is what is known as a tautomeric change, occurring spontaneously, and 
 consisting only in the passage of a hydrogen atom from one position to 
 another ; it is supposed to be due to the great ' mobility ' of the hydrogen. 
 It appears, however, that if the hydrogen atom is replaced by an alkyl 
 group, this can 'wander* also, only not so easily. Instead of changing at 
 the ordinary temperature the substance requires to be heated to about 200°. 
 Accordingly, we find, as we should expect, that a methyl group wanders 
 more easily than an ethyl, though the difference is not very great.^ It is 
 evident that the distinction between such a reaction as this and a true 
 tautomeric change is one of degree and not of kind, and that the phenomenon 
 of tautomerism is not confined to the case of the ' wandering ' hydrogen 
 atom, although it occurs in that case with peculiar ease. 
 
 The conversion of the imido-ethers into the substituted amides can be 
 brought about much more easily in the presence of an alkyl hahde, such as 
 ethyl iodide. It then takes place slowly at the ordinary temperature, and 
 rapidly at 100°. But here it can be shown that an intermediate compound 
 
 is formed: — /^ 
 
 ./t 
 
 N— Et N\"&f 
 
 CH3-C<g|^ + EtI = CH3-C^^\I = EtI + CH3.C<^ ''' 
 
 This is probable because, if the iodide of a different alkyl is used, an exchange 
 of alkyl groups sometimes occurs : — 
 
 CH3-C<q|j. + Mel = CH3-C^Q ^ + EtI. 
 
 It is further shown by the following facts. There is one case of an imido-ether 
 which changes into the amide below 100° without the addition of alkyl iodide. 
 This is the chlorethyl ether of isobenzamide : — 
 
 .^/OCH^-CH^-Cl _ .p^O 
 
 ^CHa-CHaCl 
 
 
 This is the solitary case where the addition of an alkyl halide is not required, 
 and the reason obviously is that the chlorethyl group is able to attach itself 
 to the nitrogen in exactly the same way as an alkyl halide, save that a ring 
 compound is produced: — 
 
 Lander, J. C. S. 1903. 406. 
 
Imido-ethers 95 
 
 If the original body is heated on a water bath it first melts, then solidifies, 
 and then melts again. If the solid, before it can melt again, is dissolved in 
 water and treated with silver nitrate, it gives a precipitate of silver chloride, 
 which it does not do before or after this point, showing that the hydrochloride 
 of the base is formed. Indeed, if the intermediate solid phase is rapidly cooled, 
 the ring compound can actually be isolated from it. 
 
 If the hydrochlorides of the imido-ethers are treated with excess of alcohol, 
 they first dissolve and then precipitate ammonium chloride, with the formation 
 of an ortho-ester of the original acid : — 
 
 With alcoholic ammonia, the imido-ethers react like an ordinary ester. The 
 ethoxy group is replaced by NH2, and an amidine is produced : — 
 
 EC<gJ^jj^ + NH3 = E-C<gH^ + HO-C^Hs. 
 
 The amidines can also be made by heating the amides in a stream of 
 hydrochloric acid gas: — 
 
 CH3.C<g^^ _ CH3.C<g^ 
 
 or by heating the nitriles with ammonium chloride: — 
 
 CH,-C=N _ p„ p^NH 
 -I- HgNH - '""a'-'^NHa ■ 
 
 They are at once amides and imides of the acid. They are strong monacid 
 bases, giving stable salts. The free amidines have an alkaline reaction, and 
 are easily hydrolysed into the acid and ammonia. 
 
CHAPTER V 
 
 HYDROXYLAMINE DERIVATIVES 
 
 These fall into three classes: — 
 
 A. Those in which one or more of the hydrogen atoms in hydroxylamine 
 are replaced by as many organic radicals : these are what are generally meant 
 by the hydroxylamine derivatives. 
 
 B. Derivatives of the hypothetical tautomer of hydroxylamine H3N— : 
 the amine-oxides. 
 
 C. Derivatives of hydroxylamine in which the two hydrogen atoms attached 
 to the nitrogen are replaced by a divalent radical, the oximes or isonitroso- 
 
 compounds ^>C=NOH. 
 
 The third group is by far the most numerous and important, but the 
 first two also contain some interesting compounds. 
 
 TEUE HYDEOXYLAMINE DEEIVATIVES 
 
 The mono-derivatives are of two kinds, according as the hydrogen replaced 
 is attached to nitrogen or oxygen. Those in which it is attached to oxygen, of 
 the type HjN-O-E, are known as a-hydroxylamine compounds, those in which it 
 is on the nitrogen as /3. 
 
 The a-hydroxylamines (0-ethers) are obtained from the oximes. If an 
 oxime is boiled with acids it goes back into hydroxylamine and the aldehyde 
 or ketone from which it was formed. If the oxime is treated with alkyl iodide 
 and sodium ethylate it is converted into its alkyl (0-) ether : — 
 
 0CH=NOH + C2H5I = HI + 0CH=::]SrOO2H5, 
 and when this ether is boiled with acids it splits in the same way as the 
 oxime itself, giving the 0-ether of hydroxylamine : — 
 
 ^•CH^NOCgHg + H2O = 0CHO + H2NOC2H5. 
 That the product, ethyl-a-hydroxylamine, is really an 0-ether is shown 
 by its splitting off ethyl chloride with hydrochloric acid at 150°, which would 
 not occur if the ethyl was attached to nitrogen. 
 
 The /3-hydroxylamines or N-ethers are got by the partial reduction of the 
 nitro-compounds. Thus nitromethane is converted by reduction with zinc 
 dust and water into /3-methyl-hydroxylamine : — 
 
 CH3.N<g + 2H2 = CH3N<^jj + H2O, 
 
 a reaction which proves the structure of the latter. The reduction may 
 also be effected by ammonium sulphide.' 
 
 /3-phenyl hydroxylamine, ^NHOH, is of interest because of the remarkable 
 
 > WiUstatter, Kubli, Ber. 41. 1936 (1908). 
 
fi-Hydroxylamines 97 
 
 changes which it can undergo. For a long time all attempts to prepare it 
 were unsuccessful, the product obtained being, most unexpectedly, j3-amino- 
 phenol. Thus one would expect it to be formed by boiling phenyl azide, 
 ^-Ng, with dilute acids, according to the equation 
 
 0-N/if + ? = 0-N<JL + N2 ; 
 \N OH ^OH 2> 
 
 but ^-aminophenol is produced instead, as it is also in the electrolytic reduction 
 of nitrobenzene. These facts show that /3-phenyl-hydroxylamine must be an 
 unstable body, readily changing into p-aminophenol : — 
 
 H ys 
 
 -0 ■ 
 
 This view has been fully confirmed by the work of Bamberger, who has 
 succeeded in preparing /3-phenyl-hydroxylamine in various ways, and has 
 investigated its properties in great detail. 
 
 In order to prepare it, nitrobenzene is covered with water or dissolved 
 in hot alcohol, and reduced with zinc dust, the reduction being promoted by 
 the presence of neutral salts such as calcium or ammonium chloride : — 
 
 0-N<g 
 
 /H 
 
 + 2 H2 = 0-N<gjj + H2O. 
 
 The solution must be neutral or aminophenol is formed ; this is why 
 the earlier attempts to prepare the body failed. The reaction is exactly 
 parallel to V. Meyer's preparation of the fatty /3-hydroxylamines by reduction 
 of the nitroparaffins. By this method of reduction with zinc dust in neutral 
 solution in the presence of a salt, any nitro-compound can be converted into 
 a /3-hydroxylamine. 
 
 They can also be obtained ' by the electrolytic reduction of the nitro- 
 compounds in the presence of acetic acid, the concentration of hydrogen ion 
 being kept low by the addition of sodium acetate. 
 
 Further, the y8-hydroxylamines are, as Bamberger has shown, the first 
 products which can be isolated of the oxidation of primary or secondary 
 amines with potassium permanganate, hydrogen peroxide, Caro's acid (sulpho- 
 
 mono-peracid, t^^^SOj, obtained by dissolving a persulphate in sulphuric 
 
 acid), and other substances. The final products of such oxidations are very 
 numerous and complicated, but they are all such as would be obtained from 
 the hydroxylamines, which can be isolated in many cases. It is probable 
 that the actual first oxidation product in the case of the primary and secondary 
 amines is, as it certainly is in that of the tertiary, the amine-oxide Ar-NHj : O or 
 Ara'NH : 0, and that this then changes into the tautomeric 6-hydroxylamine, e.g. 
 
 H-N + O = H-N=0 -^ ^>N-OH. 
 H/ H/ ^ 
 
 1 Brand, Ber. 38. 3076 (1905). 
 
98 Hydroxylamine Derivatives 
 
 /3-phenyl-hydroxylamine 0-NH-OH is a colourless crystalline substance 
 melting at 81°. It reduces ammoniacal silver solutions and Fehling's solution 
 in the cold. It is an unstable body in solution, and breaks up in a variety 
 of ways. In alkaline solution in the absence of air one part oxidizes another, 
 giving nitrosobenzene and aniline: — 
 
 2 0NHOH = ^-NO + 0-NH2 + HaO. 
 The nitrosobenzene so formed reacts with unchanged phenyl-hydroxylamine 
 to give azoxy benzene : — 
 
 0-NO + ^2/^"^ = 0-N— N-^ + H2O. 
 
 If the air is not excluded, the solution rapidly absorbs oxygen. If the 
 (neutral) aqueous solution is exposed to the air, the body is oxidized to 
 nitrosobenzene, which combines as before with unchanged hydroxylamine to 
 form azoxybenzene. At the same time, for each molecule of nitrosobenzene 
 formed, one molecule of hydrogen peroxide is produced. That is to say, 
 in this, as in other ' autoxidations ', for whatever reason, one half of the 
 oxygen goes to the substance oxidized, and the other half combines with 
 the water. 
 
 The alkaline solution' is much more easily oxidized by the air, so easily 
 that it can be used like p3rrogallate for removing oxygen in gasometry.^ In 
 this case only traces of hydrogen peroxide can be detected, while much 
 nitrobenzene is formed, the hydrogen peroxide being used up in presence of 
 the alkali to oxidize nitrosobenzene, the primary product, to nitrobenzene. 
 
 With aldehydes, the ;3-hydroxylamines form the N-ethers of the oximes :— 
 
 ArN<^jj + 0=CH-^ = Ar.N-CH-(^ + H^O. <^ '^^^'f 
 
 (Compare the strictly analogous reaction with nitrosobenzene, giving oxyazo- 
 benzene.) 
 
 When warmed with acids, they undergo the remarkable change already 
 mentioned into aminophenols. Bamberger' has investigated this change in 
 great detail, with a variety of hydroxylamines, and finds that the following 
 rules hold good. 
 
 Aryl hydroxylamines with the para position open are converted by sulphuric 
 acid (or by alum solution) into para-aminophenols ; while, if there is a halogen 
 in the para position, an ortho-aminophenol is formed : — 
 
 Br Br 
 
 Oh -O-oh- 
 
 *'<!Jh ™' 
 
 Sometimes these ortho-derivatives are produced even when the para position 
 is not occupied. In many cases, if sulphuric acid is used, a jJ-aminophenol 
 
 ' Bamberger, Ber. 33. 113 (]^900). 
 " Bamberger, Brady, Ber. 33. 271- 
 = Ber. 33. 8600 (1900). 
 
^-Aryl-hydroxylamines 99 
 
 9H 
 sulphonic acid, as Q'^OsH^ j^ obtained. If alcoholic sulphuric acid is used, 
 
 the p- and o-phenol ethers are obtained, although they are not formed by 
 treating the aminophenols with alcohol and sulphuric acid. In some cases 
 the formation of ^J-aminophenols is accompanied by that of the corresponding 
 hydroquinones, e. g. : — 
 
 OH OH 
 
 CHaO-CHg -* CHaO-CHg + CHgO-CHg, 
 NHOH NHa qh 
 
 and sometimes, too, if there is a para methyl group in the original hydroxyl- 
 amine, this is moved on to the next carbon atom: — 
 
 CH3 OH OH 
 
 (y-^ _CH3.Q ^CH3.Q 
 
 NHOH NH2 OH 
 
 The possibility that these hydroquinones are formed by the hydrolysis of 
 
 the aminophenols is excluded by direct experiment, which shows that this 
 
 hydrolysis does not occur. 
 
 These remarkable changes — and there are others which have not been 
 
 mentioned — Bamberger proposes to explain by the following hypothesis,* 
 
 which, though it cannot be regarded as established, is sufficiently ingenious 
 
 to be worth considering. 
 
 He supposes that the first action of the sulphuric acid solution on the 
 
 hydroxylamine is to remove water, giving the residue Ar-N<(, which is unstable, 
 
 and immediately reacts with the other substance, HX, which is present, adding 
 
 on H and X in one of two ways : — 
 
 H H\/X X 
 
 I- 0-^^ = • "• C)-Hx= iQl -Q . 
 
 /^\ /^\ /N\ NH NH, 
 
 H X 
 
 In the case we have been dealing with, of the action of dilute acid, the body 
 
 added on is generally water. This goes according to the first equation, giving 
 
 first an imino-quinole : — 
 
 H^/OH OH 
 
 ^ ^-^^ = Oi -* • 
 
 I II r 
 
 /N^ NH NH2 
 
 These imino-quinoles, with hydrogen still attached to the para carbon atom, are 
 unstable, and immediately change over to the aminophenols, and thus we get 
 the normal reaction, leading to the para-aminophenols. 
 
 If the para hydrogen is replaced by methyl, the imino-quinoles are more 
 
 ' Ber. 34. 61 (1901), Cf. Ber. 40. 1893 (1907). 
 
100 Hydroosylamine Derivatives 
 
 stable. They can then only change into aminophenols by the migration of the 
 methyl group, 
 
 CH3V-OH OH 
 
 fa SH, 
 
 and to a small extent this takes place. But the greater part undergoes 
 a different decomposition ; the NH is split off, and quinoles are formed. 
 These can actually be isolated, and are indeed the main products of the 
 reaction. 
 
 CH;,\/OH CHg^OH 
 
 iQl + H,o = iQi + NH3. . 
 
 II II 
 
 NH O 
 
 They are, however, capable of further change, the quinoid ring having a strong 
 tendency to go back to the normal aromatic structure. This can only happen 
 by the migration of one of the two groups, hydroxyl and methyl, attached to 
 the para carbon. Both of these groups can migrate. The methyl does so to 
 a certam extent in the course of the reaction in which the quinoles are formed, 
 with the production of a hydroquinone : — 
 
 CH.jv'OH OH 
 
 iQl -^«3-Q . 
 O OH 
 
 On the other hand, in the presence of alcoholic sulphuric acid, the hydroxyl 
 migrates, being etherified at the same time : — 
 
 CH3\yOH C'H3 
 
 iQl _. Eto-pj 
 A I 
 
 OEt 
 
 The formation of ortho-aminophenols can be explained in the same way, 
 except that the derivative of an ortho-quinone is produced : — 
 
 Br Br Br Br 
 
 Oh - 0.H - 0<?h ^ O.0H 
 
 NHOH N< ^■Q_ NH2 
 
 All these are examples of the second of the two possible reactions of the unsatu- 
 rated group Ar-N\. There are also cases of the first reaction. For instance, 
 phenyl hydroxylamine reacts with aniline to give 0- and ^J-amino-diphenyl- 
 
 amme:- 0.N< + H-CeH^-NH^ = ^Xc.H.-NH^ " 
 
 Again, tolyl hydroxylamine and phenol form p-oxy-phenyl-^-tolyl-amine : — 
 
 CH.CoH.N + H-CeHi-OH = CH3CeH,N<^jj ^-^ ■ 
 
^-Alkyl-hydroxylamines 101 
 
 The /3-/3-dialkyl-hydroxylamine8 (Alk)2N0H can be made by the oxidation 
 of the secondary amines,' and similar compounds are got from the cyclic 
 secondary amines, e. g. CgHmNOH from piperidine and hydrogen peroxide." 
 They are also formed in two singular reactions, by treating zinc alkyl or 
 magnesium alkyl halide with an alkyl nitrite or a nitroparafiSn.' With an 
 alkyl nitrite a substituted hydroxylamine is obtained which has the two 
 alkyls from the zinc alkyl ; and we may assume an intermediate product 
 to be formed which is then broken up by water : — 
 
 P„ ov /OZnCHs /OH 
 
 OgH„-0-N=0 + 2 Zn(CH,)a -» ca"zn>^^^3 "* CgHu-OH + N-CH3. 
 
 \GIi3 \Oxi3 
 
 With a nitroparaffin one of the alkyl groups is added on to the carbon 
 chain, apparently through the tautomeric isonitro form : — 
 
 CHg-CH^NOsj -* CH3-CH=N<^„ -* '^^3-OH.N<(.jj^.(3jj^ 
 
 NOH CH3CH2 
 This last substance was for some time supposed to be triethylamine oxide 
 (C2H6)3N=0. 
 
 AMINE-OXIDES OR OXY-AMINES 
 
 These bodies are derivatives of the hypothetical tautomeric form of hydroxyl- 
 amine HsN^O. 
 
 The first compounds of this type were prepared by Dunstan and Goulding,* 
 
 by the action of methyl iodide on hydroxylamine. The product is the hydriodide 
 
 /OTT 
 
 of trimethylamine oxide, (CH3)3Nw , which, when treated with potash, gives 
 
 the oxide (CH3)3N=0. If ethyl iodide is used instead of methyl iodide, the 
 true hydroxylamine derivative EtjN-OH is formed first, and this on further 
 treatment with ethyl iodide takes up another ethyl group, giving EtjN^O. 
 With propyl or isopropyl iodide the action stops at the first stage, with the 
 production of PrgN-OH. 
 
 A simpler method of preparation, also due to Dunstan, is to treat the 
 tertiary amines with hydrogen peroxide, when they are directly oxidized to 
 the amine-oxide :— Pr^N + = Pr3N=0. 
 
 These oxy-amines have not been obtained in the anhydrous state. They 
 separate from water in the form of hygroscopic crystals of a hydrate, e. g. 
 (CH3)3N=0, 2 H2O, which retains its water of crystallization with great 
 firmness. The solution has a strong alkaline reaction, and it does not reduce 
 Fehling's solution, showing that the body is not a true hydroxylamine 
 
 derivative. When treated with acid it gives a salt such as (CH3)3N\j , so 
 
 that it behaves as if one of the molecules of water was chemically united 
 
 > Mamlook, Wolffenstein, Ber. 33. 159 (1900) ; 34. 2499 (1901). 
 ' Wolffenstein, Haase, Ber. 37. 3228 (1904). 
 3 Bewad, J. pr. Ch. [2] 63. 94, 193 (1901) ; Ber. 40. 3065 (1907). 
 • J.C.8. 1899. 792, 1()04. 
 1175 H 
 
102 Amine-oxides 
 
 to form a dihydroxy-compound (CH3)3N<^qjj- Similarly, it combines with 
 
 /OPH" 
 
 methyl iodide to give the ether-salt (CH3)3N\j ^ • 
 
 When it is heated, trimethylamine-oxide breaks up with the loss of a methyl 
 group into dimethylamine and formaldehyde : — • 
 
 (CH3)3N=0 = (CH3)2NH + CH2O. 
 When reduced with zinc dust it gives trimethylamine, which is a proof that 
 all three methyl groups are directly attached to nitrogen. 
 
 An analogous series of compounds are the oxides of the aromatic amines, 
 prepared by Bamberger,' by direct oxidation. They are formed almost quanti- 
 tatively from any tertiary mixed amines, such as dimethyl-aniline, dimethyl- 
 toluidine, &c., by treatment with hydrogen peroxide or Caro's acid.^ They 
 are crystalline compounds (0N(CH3)2=O melts at 152°) which are excessively 
 hygroscopic, but apparently do not form definite hydrates. When heated 
 a,bove their melting-points they break up mainly into tertiary base and oxygen. 
 They have an alkaline reaction and form salts with acids. They resemble 
 Dunstan's alkyl compounds in their general behaviour, with one important 
 exception. Owing to the presence of the benzene ring they can form substitu- 
 tion-products, and do so with great ease. Thus they give sulphonic acids with 
 sulphurous acid and nitro-derivatives with nitrous acid. We may assume 
 that in these cases the salts are first produced, and then isomerize : — 
 H H SO3H 
 
 =0 ^ P .H.O. 
 
 CH,N=0 + H-SOaH CH3-N<g^jj CH3N 
 
 CH3 CH3 ' CH3 
 
 The positions taken up by the substituents are ortho and para. In the case 
 of the sulphonic acid it is mainly the ortho that is formed ; in the nitro- 
 compound about equal quantities of the two are obtained. The reaction is 
 parallel to the ordinary substitution of aniline ; but it is to be noticed that the 
 nitrogen is here pentavalent, which, however, seems to make no difference 
 to the result. 
 
 A third group of similar compounds has been prepared by Haase and 
 Wolffenstein by the action of hydrogen peroxide on the alkyl-piperidines : ' 
 ■e. g. from N-ethyl-piperidine the oxide 
 
 CH2 CH2 • 
 N=0 
 
 » Ber. 32. 342, 1882 (1899); 34. 12 (1901); 35. 1082 (1902). 
 
 " Simil.ir compounds can be prepared from the leuco-bases of the triphenyl methane dyes, such 
 .as malachite green and crystal violet. Bamberger, Rudolf, Ser. 41. 3290 (1908). 
 s Ber. 31. 1553 (1898); 38. 2607 (1899). 
 
Amine-oxides 103 
 
 Unlike the fatty, but like the aromatic derivatives, these bodies lose oxygen 
 vfhen distilled, re-forming the alkyl-piperidine. Indeed, the oxygen appears 
 to be much more loosely attached than in either of the other two series ; thus 
 they break up into alkyl-piperidine and oxygen with almost explosive violence 
 if heated in a current of hydrochloric acid, and they liberate iodine from 
 potassium iodide on v^arming. 
 
 As has already been mentioned, methyl-ethyl-aniline oxide, (f>(GS.^) (C2H5)N=0, 
 •CSM be broken up by the fractional crystallization of its d-brom-camphor 
 sulphonate, into two optically active forms, which retain their activity not 
 only when converted into the chlorides, but also in solution when treated 
 with baryta, i. e. in the form of the oxide or hydroxide/ 
 
 HYDKOXAMIC ACIDS 
 
 These bodies are related to the amides as hydroxylamine to ammonia, and 
 like the amides they have two possible formulae: — 
 
 In no instance have the two tautomers been separated, and it is even more 
 difficult here than in the case of the amides to decide which of the two 
 represents the actual structure ; there are arguments of no great weight on 
 -either side. 
 
 The hydroxamic acids are formed (like the amides) by the action of hydroxyl- 
 amine on the esters or on the acid anhydrides.^ In the latter case an acyl- 
 hydroxamic acid (generally known as a di-hydroxamic acid) is produced, which 
 is easily hydrolysed by alkalies to form the hydroxamic acid :— 
 
 CH3.C<o ^ ^'^^^ ^ ^^^-^NNHOCOCHa ^ ^^^-^^NHOH " 
 
 They are also formed (often even in the cold) by the action of hydroxylamine 
 •on the amides, ammonia being expelled. 
 
 They can further be obtained ' by oxidizing the aldoximes with Caro's acid : — 
 
 which resembles the oxidation of an aldehyde to an acid. Since the aldoximes 
 -are themselves formed by the oxidation of amines of the type E-CHj-NHg with 
 Caro's acid, the hydroxamic acids are among the products of oxidation of such 
 amines.* They are also formed from the nitroparaffins by intramolecular 
 ■change : for example, in small quantity by acidifying their alkaline solution " : — 
 
 R.CH=N<gjj -. E.C<ggjj 
 
 — a somewhat unusual migration of oxygen. In the same way the sodium 
 
 ^ Meisenheimer, Ber. 41. 3966 (1908). See above, p. 30. 
 
 " Jeanrenaud, Ber. 22. 1270 (1889). Cf. Morelli, C. 08. ii. 1019. 
 
 = Bamberger, Ber. 34. 2029 (1901). ' Bamberger, Ber. 36. 710 (1903). 
 
 " Bamberger, Rust, Ber. 35. 45 (1902). 
 
104 Hydroxamic Acids 
 
 salts of the nitroparaffins when treated with aeyl chloride form the di-hydrox- 
 amic acids ^ : — 
 
 CH3.CH=N<gj^^ -. CH3.CH=:N<Oco.jj -» Cn,G<g\^Q.^- 
 
 The hydroxamic acid chlorides (having of course the hydroxyl on the carbon 
 replaced by chlorine) can be got by the action of hydrochloric acid on the 
 alkyl-nitrolic acids '' : — 
 
 CH3-C<55g|j + HCl = CH3.C<g^Qjj + HNO,, 
 
 or by the action of chlorine on the aldoximes " : — 
 
 The alkyl-hydroxamic acids are neutral, while the aromatic have an acid 
 reaction. They all, however, form salts, the ferric salts having a characteristic 
 
 deep red colour. Form hydroxamic acid, H-C'^Tig-/^TT (obtained, for example, 
 
 from formic ester and hydroxylamine), melts at 81-82°, and explodes above 
 
 this temperature, decomposing into carbon monoxide and hydroxylamine. Its 
 
 /CI 
 chloride, H-C<^-xtqtt, also known as formyl chloride oxime, is closely related 
 
 to fulminic acid, from which it can be obtained by the action of hydrochloric 
 acid, and into which it is very easily converted: — 
 
 HC<^Qjj ^ H^^ + C=N.OH . 
 
 /OH 
 
 Di-acet-hydroxamic acid, C'H3-C^-»jq qq pxr , obtained from hydroxylamine 
 
 hydrochloride and acetic anhydride, is a strong acid, readily hydrolysed by 
 excess of alkali to acet-hydroxamic acid. But if it is treated with half 
 a molecular proportion of potassium carbonate it gives symmetrical dimethyl 
 urea. This change probably takes place (like the Hofmann reaction) through 
 the Beckmann rearrangement : — 
 
 CH3GOK CH,COOCOK 
 CH3.CO.0I - CH,1 - 0^:^^^^0 . CH3.CO.OK 
 
 2 CH3-N=C=0 + H2O = (CH3-NH)2C=0 + COj. 
 
 /OH 
 
 The aromatic hydroxamic acids, such as benzhydroxamic acid, 'i>-^\-vjrya (from 
 
 benzoyl chloride and hydroxylamine, or by the oxidation of benzaldoxime, &c.), 
 closely resemble the alkyl compounds. 
 
 The ethers (or esters) of the hydroxamic acids can exist in several isomeric 
 forms. For example, if formic ester is treated with an a-(0-)-hydrqxylamine, 
 it gives a hydroxamic ether in which the hydrogen of the hydroxylamine OH 
 is replaced * : — 
 
 HC<Ce + H,N.OEt -. H.C^jj.o^^ or H-^^^^ • 
 
 1 L. W. Jones, Am. Ch. J. 20. 1 (1898). ^ Wemer, Busa, Ber. 28. 1280 (1895). 
 
 » Wieland, Ber. 40. 1676 (1907). * Nef, Ber. 31. 2721 (1898) ; Biddle, Ann. 310. 9 (1899). 
 
Hydroxamic Acids 105 
 
 On the other hand, if benzimido-ether is treated with hydroxylamine, an ether 
 is formed in which the hydrogen of the other hydroxyl (COH) is replaced 
 
 (Here there is no tautomeric formula possible, as the ' wandering ' hydrogen 
 is replaced.) The ethers of this latter type all occur in two modifications 
 (when derived from the aromatic hydroxamic acids), which must be stereo- 
 isomers of the type of the syn- and anti-oximes : — 
 
 ^•COEt d)COEt 
 
 Msyn)jj^ll ^(anti) 11 ^^ • 
 
 This is confirmed by the fact that the a-(syn)-ethers, when treated with phosphorus 
 pentachloride, undergo the Beckmann change, and are converted into phenyl- 
 carbamic acids: — 
 
 d>.COEt HOCOEt 0:COEt 
 
 ^ 11 -^ II -» I , 
 
 HO-N 0N 0NH 
 
 while the /3-(anti)-ethers do not react in this way, but form phosphoric esters.' 
 
 AMIDOXIMES 
 
 These are related to the hydroxamic acids in the same way as the amidines 
 
 to the amides : they are oxy-amidines R-C<\|g^iT or R-C<^i^tt . The first 
 
 formula agrees best with their behaviour. The first member of the group 
 (formamidoxime or isuretin) was discovered by Lossen ; "^ the whole group was 
 subsequently investigated by Tiemann ' and his pupils. 
 
 Their method of preparation is similar to that of the amidines. The alkyl 
 derivatives are got by the action of hydroxylamine on the nitriles : — 
 
 EC=N + H2NOH = EC\NH^' 
 
 the aromatic by the action of hydroxylamine on the thioamides, imido-ethers, 
 and amidines. 
 
 They form stable salts with acids, unstable salts with bases. The free 
 amidoximes are easily hydrolysed, even by water alone, to hydroxylamine and 
 the amide. Their hydrochlorides are converted by sodium nitrite into the 
 amide and nitrous oxide : — 
 
 E-C<Jg^ + HNO, = E-C<gjj^ + N,0 + H^O. 
 
 They resemble the hydroxamic acids in many ways. When treated with 
 phosphorus pentachloride they are converted, by the Beckmann reaction, into 
 alkyl-ureas :- e.C-NH^ HOCNH^ _^ O^C-NH^ 
 
 HON ~* EN ~^ ENH 
 
 ' Werner, Ber. 29. 1146 (1896). ' Ann. Spl. 6. 234 (1868). 
 
 » Ber. 17. 126, 1685 (1884) ; 22. 2391 (1889), &c. 
 
106 Amidoximes 
 
 Formamidoxime HC^iJS^ , also known as isuretin, being an isomer of 
 urea, is a solid melting at 114-115°, which has a strong alkaline reaction. 
 
 Oxy-amidoximes or hydroxamoximes, ^^'C^jyfjj.oH' ^^^ ^^^° known, being 
 obtained, for example, from hydroxylamine and a hydroxamic chloride. 
 
 OXIMES OE ISONITEOSO-COMPOUNDS 
 
 The most important and numerous class of hydroxylamine derivatives are 
 the oximes, containing the group )'C=NOH . 
 
 There are two chief methods of preparing the oximes. The first and most 
 general is by the action of hydroxylamine on a compound containing a carbonyl 
 group — an aldehyde or a ketone. The reaction is quite general, and may be 
 taken as a criterion of an aldehydic or ketonic grouping. The number of 
 instances in which it does not occur is very small. 
 
 It is usually carried out by warming the aldehyde or ketone in aqueous 
 or dilute alcoholic solution with hydroxylamine hydrochloride and an equiva- 
 lent of sodium carbonate : — 
 
 EC<Q + H2NOH = H2O + E-C<^Qjj- 
 
 This generally suffices for aldehydes, but many ketones require a more energetic 
 treatment, such as heating in a sealed tube with excess of soda. 
 
 The velocity of formation of oximes by this method has been investigated 
 by Petrenko-Kritschenko and Kautscheff.' They used N/1000 solutions in 
 50 per cent, aqueous alcohol. They did not determine the velocity constant, 
 but only measured the amount of oxime formed in a given time by removing 
 a sample and titrating the unchanged hydroxylamine either with iodine, or 
 with hydrochloric acid, using p-nitro-phenol as indicator. Their chief results 
 were as follows: — 
 
 1. The order of velocities of formation of oxime with hydroxylamine is 
 roughly the same as that of the formation of hydrazones and of the bisulphite 
 compounds. 
 
 2. Among the ketones, those with a methyl attached to the carbonyl react 
 most rapidly, those with larger alkyls (especially secondary) more slowly. 
 
 3. Eing ketones (such as keto-cyclohexane) form oximes very rapidly, and 
 those with 6-rings very much quicker than those with 5- or 7-rings. 
 
 4. Aromatic ketones react more slowly than fatty, aromatic aldehydes more 
 quickly than fatty. 
 
 5. In the keto-acids, the substitution of COjEt for CH3 diminishes the 
 velocity, and the more so the further it is froni the carbonyl. Also the 
 replacement of ethyl for hydrogen diminishes the velocity. The reaction is 
 reversible. It is catalysed to some extent by acids, and much more so by 
 
 ' Ser. 39. 1452 (1906). Very similar results were obtained in aqueous solution by Stewart, 
 J.C.S. 1905. 410. 
 
 2 Aeree, Am. Ch. J. 39. 300 (C. 08. i. 1389). 
 
Oocimes : Formation 107 
 
 The second main method for preparing oximes is by the action of nitrous 
 acid on the methylene group CH2. This is the exact converse of the first :— 
 
 >C=0 + H2NOH ^ 
 
 >C:H, + 0:NOH 
 
 )C=NOH + HaO. 
 
 It is not all methylene groups which will react with nitrous acid, but only 
 those whose hydrogen has acquired a more or less acidic character from the 
 proximity of negative groups. Thus aceto-acetic ester gives the so-called 
 isonitroso-aceto-acetic ester : — 
 
 CHg-COCHa + 0:NOH _ CHgCOC^NOH 
 
 COOEt ~ COOEt ^ ' 
 
 The nomenclature of these bodies is rather confusing. It is to be noticed 
 that there ia no difference between the oxime and the isonitroso grouping. 
 The distinction has reference only to the way in which the body is derived 
 from its mother substance. Take acetone as an example. Its oxime, 
 CH3-C(:NOH)CH3, is derived from it by replacing the ketonic oxygen by 
 =NOH : while in isonitroso-acetone the =NOH replaces two hydrogens attached 
 to one carbon, CHg-CO-CHtNOH. Hence the difference of name indicates 
 a real difference of constitution ; but it would be equally correct to call 
 acetoxime isonitroso-propane, and isonitroso-acetone methyl-glyoxal monoxime. 
 
 There is a third method of preparing the oximes which is of great theoretical 
 interest. This is Scholl's reaction,' a modification of that of Friedel and Crafts. 
 It consists in treating an aromatic hydrocarbon with mercury fulminate in 
 presence of a mixture of hydrated and anhydrous aluminium chloride. We 
 may suppose that in this reaction the water liberates hydrochloric acid, which 
 sets free fulminic acid. This, as will be shown later, has the formula C=NOH. 
 It combines with more hydrochloric acid to give formyl chloride oxime 
 
 pi^C=NOH ; and this reacts with the aromatic hydrocarbon in the normal 
 
 Friedel and Crafts manner: — 
 
 CeHs-H + g>C=NOH = HCl + ^s^>C=NOH. 
 
 The actual product is therefore benzaldoxime. Unless a partially hydrated 
 aluminium chloride is used, the product body obtained is a dehydration-product 
 of benzaldoxime, benzonitrile. 
 
 The oximes can also be prepared by the reduction of the nitro-compounds 
 and by the oxidation of the amines. 
 
 The oximes are at once feebly basic and feebly acidic. They dissolve 
 in alkalies, and also form salts in ethereal solution with mineral acids. They 
 are, however, only in a secondary sense amphoteric electrolytes. The hydrogen 
 ion to which the acidic properties are due is derived directly from the oxime : — 
 
 K2C=N0H ^ EaC^NO' + H', 
 while the basic properties are due to the nitrogen becoming pentad, as in 
 the salt E2C=NH(0H)C1. 
 
 » Ber. 36. 10 (1903). 
 
108 Oximes 
 
 When the oximes are warmed with acids they break up again into 
 hydroxylamine and the aldehyde or ketone, the reaction by which they are 
 formed being reversible : — 
 
 ^23>C=0 + H^NOH :^ g][^3\,c^N0H + H^O. 
 
 In this particular case equilibrium is attained from either side (in presence 
 of one equivalent of mineral acid) when about two-thirds of the acetone has 
 been converted into oxime ; and this equilibrium is very little affected by 
 the temperature.* 
 
 The ketoximes when treated with sodium ethylate are converted into 
 ethers, such as (CHg)2C:N0-Et, which on hydrolysis with acids yield the 
 a-hydroxylamines, as HjNOEt.^ 
 
 The ketoximes are converted by acetic anhydride into their acetic esters, 
 the oxime hydrogen being replaced by acetyl : — 
 
 EaCiNOH -^ EaCrNOCOCHg. 
 
 But most aldoximes do not give this reaction, the anhydride acting merely 
 as a dehydrating agent and forming the nitrile : — 
 
 K-C<T^, 
 
 NOH = ^-^-^ + ^2°- 
 If treated with acetyl chloride many oximes undergo the remarkable 
 Beckmann reaction, the nitrogen inserting itself into the middle of the chain 
 to give a substituted amide : — 
 
 CH3CCH3 CHgCiO 
 
 NOH "*" HNCHg' 
 On reduction with sodium amalgam in dilute acetic acid,' or electrolyticaUy 
 in presence of strong sulphuric acid, or with calcium turnings in alkaline 
 solution in presence of mercuric chloride,* they are converted into amines. The 
 electrolytic method has been used commercially, for example, for the preparation 
 of isopropylamine ° : — 
 
 ci:>C=NOH + 2 H, = gg^Xgjj^ + H,0. 
 
 On oxidation (with Caro's acid) the oximes undergo two changes.* In the 
 first place the CH group is oxidized to COH, giving a hydroxamic acid : — 
 
 0-CH=NOH + O = <^C<^H^ . 
 
 This is the common behaviour of hydrogen attached to a carbon which carries 
 no other hydrogen, as in the oxidation of aldehydes to acids. At the same 
 time part of the oxime is oxidized in a different way, giving a nitro-compound :— 
 
 ^•<0H + = f< o -> 0-CH,N<g. 
 
 'N<OH 
 
 ' Francesconi, Milesi, C. 02. ii. 259. 
 
 ' The velocity of etherification of the oximes has been exaniined by Goldsohmidt, Z. f. 
 Elektrochem. 14. 681 (C. 08. ii. 1351). 
 
 ' Tafel, Pfeffermann, Ber. 35. 1510 (1902). • Beckmann, Ber. 38. 904 (1905). 
 
 ' C. 03. i. 1162. 6 Bamberger, Ber. 33. 1781 (1900). 
 
Oximes : Properties 109 
 
 Formaldoxime,^ if it is rapidly prepared, can be obtained in the monomolecular 
 form as a liquid, HaC^NOH, boiling at 84-5° ; but this rapidly changes to 
 a polymer, probably (H2CNOH)3, a gelatinous body insoluble in organic 
 solvents, but soluble in acids. If this polymer is heated it volatilizes, giving 
 a vapour whose density shows it to be monomolecular. If it is carefully 
 heated in a test tube, liquid drops of the simple oxime are deposited on the 
 sides of the tube ; but they change almost immediately into the solid polymer. 
 In this strong tendency to polymerize formaldoxime resembles formaldehyde itself. 
 
 On sudden heating formaldoxime breaks up into water and prussic acid. 
 
 Acetaldoxime, CHa-CHiNOH, shows signs of occurring in two modifications." 
 The ordinary form melts at 47° ; but if it is kept at or above this temperature 
 for some time, it does not freeze till much lower, and the longer it has been 
 heated the more supercooling is required. Even when the freezing has begun 
 the crystals only separate slowly, though they are always found to melt at 47°. 
 This behaviour suggests that the heated liquid may be a solution of the 
 ordinary form in one of lower melting-point. OH 
 
 If chloral is treated with hydroxylamine, an addition-product, CClg-CH , 
 
 ^NHOH 
 is first formed, which, however, readily loses water to give the oxime 
 CClg-CHiNOH. We may suppose that such an additive compound is the 
 first product in all cases of oxime formation, though it has not been isolated 
 in any other instance. 
 
 The oximes of diketones may be obtained in two ways. The diketone 
 may be treated with hydroxylamine, which gives the mono- or dioxime according 
 to the quantities employed ; or the monoxime may be made (as the isonitroso- 
 compound) by treating a mono-ketone with nitrous acid. In the latter case 
 the oxime (isonitroso) group can often turn out another group to make room 
 for itself. Thus, if methyl-aceto-acetic acid is treated with nitrous acid, the 
 monoxime of diacetyl (isonitroso-ethyl-methyl-ketone) is produced : — 
 CH3 CH3 
 
 CHgCO-CH + 0:NOH = CHo,COC=NOH + CO^ + H^O. 
 COOH 
 
 When the product is boiled with acids it splits off the NOH in the normal 
 manner, and diacetyl is formed : — 
 
 CH3-COC(:NOH)CH3 + H^O = CH3COCOCH3 + H2NOH. 
 Thus, by the successive action of nitrous acid and hydrolysis a CH2 group 
 is converted into CO, the nitrous acid being reduced to hydroxylamine. In 
 the same way aceto-acetic acid itself gives isonitroso-acetone and carbon dioxide, 
 while aceto-acetic ester does not split off the carboxyl, but forms its own 
 isonitroso-derivative : and acetone-dicarboxylic acid gives di-isonitroso-acetone : — 
 
 CHa-GOOH 0:NOH CH:NOH 
 
 CO + = CO +2 H2O -f 2 CO2. 
 
 CH2COOH 0:NOH CH:NOH 
 
 ' Cf. Dunstan, Bossi, J. C. S. 1898. 353. 
 
 ' Dunatan, Dymond, J. C. S. 1894. 206 ; Carveth, C. 98. ii. 178 ; Dutoit, Fath, C. 04. i. 256. 
 
110 Oocimes 
 
 It is to be noticed that all those compounds which contain at once 
 a carbonyl and a C=NOH group dissolve in alkali to form a deep yellow 
 solution, whereas if aU the keto-groups are converted into oxime, the alkaline 
 solution is colourless. Thus di-isonitroso-acetone, CH(:NOH)COCH=NOH, 
 gives a reddish yellow alkaline solution. Hydroxylamine converts this body 
 into tri-isonitroso-propane, CH(:NOH)C(:NOH)CH:NOH, whose alkaline solution 
 is colourless. There can be little doubt that this is due to the fact that the 
 bodies of the former class (a-keto-oximes) are pseudo-acids : that their alkaline 
 salts are derived from a tautomeric form. Thus the colourless iso-nitroso- 
 
 .TT 
 
 acetone is CH3-C0-C<^^^tt ; but in the alkaline salt some kind of interaction 
 
 must occur between the carbonyl and the oxime groups, which we may 
 
 CHs-C-O ) 
 represent by the formula I I l^K, as in the formation of the chromo- 
 
 salts of the a-nitroketones, the dinitroparafSns, and the nitrophenols. If 
 
 aU the ketone groups are converted into oxime groups this particular form 
 
 of tautomerisni is no longer possible. 
 
 In the aromatic series, however, the dioximes of a-diketones (a-quinones) 
 
 show a similar tautomeric behaviour. ' The a-quLnones of benzene, naphthalene, 
 
 and phenanthrene are themselves red, and their dioximes yellow. These 
 
 -C=N 
 dioximes all form anhydrides, I ,^.^0, which are colourless. Their ethers, 
 
 -C=NOAlk '^~^ 
 
 I , in the naphthalene series are yellow, and in the phenanthrene 
 
 -t^WOAIk _r=NONa 
 
 series colourless. The salts, as l_ , are dark red in the benzene 
 
 derivatives, yellow in those of naphthalene, and nearly colourless in those 
 of phenanthrene. Such facts as these show the weakness of the ' quinoid ' 
 theory of colour : for all these compounds, as usually written, contain the 
 same quinoid ring. We must suppose an equilibrium between two tautomeric 
 forms, one colourless, no doubt the simple quinoid form, and the other 
 coloured, in which the two oxime groups interact : — 
 
 Ar<^«^ ^ Ar<r^ 
 ^NOX ^NOX 
 
 Colourless Coloured 
 
 The first structure will be that of the anhydrides, and will predominate in 
 
 the free oximes : the salts will mainly consist of the second. 
 
 STEEEOISOMEEISM OF THE OXIMES 
 
 The study of the aromatic oximes led to the discoveiy of the stereoisomerism 
 of trivalent nitrogen." The question was opened by the discovery of V. Meyer 
 and Goldschmidt (1883-8) that benzil dioxime, ^•C:NOHC:NOH-^, exists in two 
 isomeric forms. The investigation of their properties showed that they must 
 
 ' Schmidt, Soil, Ber. 40. 2454 (1908) ; Hantzsch, Glover, ib. 4344. 
 
 ^ The earlier history of this subject is well given by Lachman, The Spirit of Organic 
 Chemistry (1899). 
 
Stereoisomermn of the Oximes 111 
 
 both have the same structural formula. They are both hydrolysed by aeida 
 to form benzil, and therefore have the carbon chain unaltered ; they both 
 
 yield on oxidation the same body, .1 T , and they both have the same 
 
 ^•C=N-0 ^ 
 
 molecular weight. It is interesting to notice that this was the first application 
 of Eaoult's cryoscopic method in organic chemistry — a method afterwards 
 improved by Beckmann in connexion with the same subject. 
 
 Since structural isomerism was excluded, it was natural to suggest stereo- 
 isomerism. But in benzU dioxime stereoisomerism of the carbon is impossible 
 unless we suppose that the two central carbon atoms are incapable of free 
 rotation about the line joining their centres of gravity. In that case there 
 are two possible arrangements : — 
 
 I and I 
 
 ^C=NOH ^/C=NOH 
 
 But van 't Hoff 's second law, which is supported by an enormous mass of 
 evidence, declares that with two singly linked carbon atoms free rotation 
 must be possible ; and no parallel case of stereoisomerism of this kind could 
 be adduced. If, however, the assumption is made, it is easy to show that 
 three isomeric dioximes of benzil and three monoximes are possible ; and 
 when, in 1889, V. Meyer and Auwers actually discovered a third dioxime 
 and a second monoxime, it seemed as if this hypothesis must be accepted. 
 
 Meanwhile, in 1886, Beckmann had found that benzaldoxime, 0-CH:NOH, 
 could be converted into an isomer. Here the explanation offered for the 
 benzil derivatives is impossible, because there is no second carbon atom ; and 
 hence chemists adopted the theory that the second oxime had the formula 
 d)-CH— NH. 
 
 This theory soon received apparent confirmation from the following 
 discovery. If the two benzaldoximes are treated with benzyl iodide and 
 alkali, they are converted into two isomeric benzyl ethers, which, according 
 to the received formulae of the oximes, should be 
 
 ^•CH:NOCH2-0 and ^-CH— N-CH^^, 
 
 and it was shown that they actually had these formulae by the fact that 
 on hydrolysis one gave a-benzyl-hydroxylamine, H2N-OCH2-0, and the other /3-, 
 
 JJJJ/CH2-0 
 
 But before long this view was overthrown by the work of Goldschmidt. 
 He investigated first the behaviour of the benzil oximes with phenyl isocyanate, 
 0-N:C:O. This body reacts readily both with OH and with NH groups. With 
 hydroxyl compounds it gives urethane derivatives ; for example, with alcohol, 
 phenyl urethane : — 
 
 + HO-Et - j,^o-^-"- 
 
112 Oximes 
 
 With inline groups it gives substituted ureas ; e. g. with diethylamine, diethyl- 
 phenyl urea:— ^.nh 0-NH 
 
 >C:0 = >C:0. 
 
 NEt2 Et^N 
 
 Hence phenyl isocyanate should react differently with the oximes according 
 as they are E-CH:NOH or RCH— NH. Goldschmidt found, however, that 
 
 the two isomeric benzil monoximes gave the same product with this reagent ; 
 and so also did the three benzil dioximes. This M^as in accordance with the 
 views of V. Meyer and Auwers, who regarded these isomers as structurally 
 identical. But on applying the same reagent to the two benzaldoximes, 
 Croldschmidt found that they too gave identical products, which was incon- 
 sistent with Beckmann's view that the benzaldoximes were structural isomers. 
 And Goldschmidt pointed out that his evidence was much stronger than 
 Beckmann's evidence on the other side ; for Beckmann had only shown 
 that the benzyl ethers were structurally different ; and we know of many 
 cases where the same mother substance gives two series of ethers. 
 
 Thus it was shown that the benzaldoximes, like the benzil oximes, were 
 structurally identical ; while the hypothesis of the absence of free rotation, 
 brought forward to explain the isomerism of the benzil derivatives — a hypo- 
 thesis which was even there of doubtful validity — was wholly inapplicable to 
 the case of the benzaldoximes. 
 
 At this point Hantzsch and Werner brought out a paper of fundamental 
 importance,^ in which they offered a solution of the whole problem which 
 has since been universally accepted. The novelty of tieir hypothesis consists 
 in this, that the stereoisomerism is referred not to the carbon, but to the 
 nitrogen, or rather to the two together. In the ordinary case of triad 
 nitrogen, where the nitrogen is united to three different monovalent groups, 
 it is most natural to suppose that the three valencies are equally distributed 
 in a plane, which would make stereoisomerism impossible. But there are 
 many compounds in which a triad nitrogen atom replaces a CH group, as in 
 
 CH CH N 
 
 the series III ^ Ml 111, or in benzene and pyridine. In such cases it seems 
 CH N N ^■' 
 
 possible that the three nitrogen valencies may be arranged in the same way 
 
 as the three free valencies of the carbon in CH ; that is, in the directions 
 
 from the centre towards three angles of a tetrahedron. This is the hypothesis 
 
 which Hantzsch and Werner suggest : they represent the three nitrogen 
 
 valencies as directed towards three angles of a tetrahedron, of which the 
 
 nitrogen atom itself occupies the fourth angle. Thus, a compound with 
 
 doubly linked triad nitrogen of the general formula tr/C'^-^'^ may occur 
 in two stereo-modifications : — 
 Xk 7Y 
 
 z 
 
 Ber. 23. 11 (1890). 
 
Stereoisomerism of the Oximes 113 
 
 which for the sake of brevity may be written 
 
 XGY XCY 
 
 It and II • 
 
 N-Z Z-N 
 
 The case is exactly parallel to that of the double-link carbon isomerism, as 
 in fumaric and maleic acids, where the general type is 
 
 a-C-b a-C-b 
 
 II and II • 
 aC-b bCa 
 
 The truth of this theory was finally established by three further discoveries. 
 
 In the first place, it was found that in many cases the benzoyl and acetyl 
 esters of the oximes can exist in three forms. One of these is the N-ester; 
 but the others are both 0-esters, and must be stereoisomeric, since there is 
 no other possibility. 
 
 Secondly, the occurrence of stereoisomeric oximes has been found to depend 
 on the two groups attached to the carbon being different. It is obvious that 
 on Hantzsch and Werner's scheme, if X and Y are identical, no isomerism 
 can occur, and the facts have been shown to agree with this. Even the slightest 
 difference between the two groups is sufBcient to cause isomerism, but unless 
 some such difference occurs, no isomers are obtained. Thus benzophenone oxime, 
 
 II , only occurs in one form, whereas of »-chloro-benzophenone oxime 
 
 NOH' ^ > ^ f 
 
 ^•C'CeH^Cl 0-C-C„H4Cl 
 
 there are two, 11 ^^^ ^^^^ ^^^ II 
 
 Thirdly, it has been found possible to assign to the various isomers 
 their stereo-formulae in a consistent manner. The considerations employed 
 in determining them are as follows : — 
 
 To begin with, a nomenclature is required for these compounds. For this 
 purpose Hantzsch and Werner use the prefixes syn- and anti-, the former 
 denoting that the hydroxyl is near to, and the latter that it is remote from, 
 that group (of X and Y) which immediately follows the prefix. Thus the body, 
 
 II ^ * ' , may be called syn-phenyl-tolyl-ketoxime, or anti-tolyl-phenyl- 
 
 ketoxime. When one of the gi-oups reacts with the hydroxyl, as in the 
 aldoximes (which can give nitriles), the name is so chosen that the reactive 
 form is called syn ; thus of the two benzaldoximes : — 
 
 (Reactive) ^-C-H (Not reactive) ^-C-H 
 
 N-OH HO-N 
 
 Benz-synaldoxime Benz-antialdoxime 
 
 (not anti-benzald oxime) (not syn-benzaldoxime) 
 
 The dioximes of the symmetrical diketones can occur in three forms, 
 distinguished as syn, anti, and amphi : — 
 
 syn -C C- anti -C-C- amphi -C C- 
 
 II 1 > II 11 ' II II • 
 
 NOH HO-N HON NOH NOH NOH 
 
 In determining which configuration is to be assigned to each of the two 
 
114 Oximes 
 
 isomers, we rely mainly on two reactions. The first of these, which applies 
 to the aldoximes alone, is the formation of nitriles. For example, the 
 two benzaldoximes can easily be converted into acetyl derivatives, which are 
 readily saponified, and so must have the acetyl attached to oxygen, not 
 nitrogen : and which can be readily turned into one another, and therefore 
 must be stereo- and not structural isomers. Their formulae are therefore: — 
 
 ^ II and ^ II 
 
 NOCOCH3 CH3COON 
 
 On gentle warming with sodium carbonate one of these esters forms 
 benzonitrile : — 
 
 II = 0-C=N + CHa-CO-OH, 
 
 while the other gives no nitrile, but only regenerates the original oxime 
 by saponification. Hence the first must be the syn-compound, derived from 
 benz-synaldoxime, since this has the hydrogen and the O-CO-CHg close together ; 
 while the second must be the anti-body. 
 
 With the ketoximes we cannot use this method, and we have to rely on the 
 Beckmann intramolecular transformation. In this strange reaction, which is 
 brought about, often at the ordinary temperature, by various reagents, of which 
 xicetyl chloride, phosphorus pentachloride, and concentrated sulphuric acid are 
 the most commonly used, the oxime is converted into a substituted amide. 
 Thus benzophenone oxime gives phenyl benzamide (benzanUide) : — 
 
 ^ ir = (ACONH-0. 
 N-OH ^ ^ 
 
 The reaction is most simply explained as consisting in an exchange of the 
 hydroxyl with one of the groups attached to the carbon : — 
 
 XCX XCOH XCO 
 
 II _» II _> I ■ 
 
 NOH NX NHX 
 
 If the two groups attached to the carbon are different, it can go in two ways : — 
 
 X.c-Y ^XCONHY 
 II 
 NOH-~-:^YCONHX 
 
 Thus, wherever stereoisomerism is possible, it is also possible to get two 
 products in the Beckmann reaction. It is evident that, if our interpretation 
 of the nature of the isomerism and of the nature of the Beckmann reaction 
 is correct, the group which becomes attached to the nitrogen is that which 
 was in the syn-position to the hydroxyl in the original oxime, e. g. : — 
 
 XCY XC-OH XC=0 
 
 II ^ II _* I . 
 
 N-OH NY NHY 
 
 X-C-Y HOCY 0:=CY 
 
 HO-N ~^ XN ~* NHX" 
 Now it is found that in all cases of stereoisomeric ketoximes one form gives 
 
Stereoisomerism of the Oaeimes 115 
 
 only one product in the Beckmann reaction, while the second form gives 
 some of this product, but mainly the other. This only means that the 
 second isomer is less stable than the first, and in the course of the reaction 
 is partly converted into it. It does not prevent us from drawing conclusions 
 as to the spatial configurations of the two forms from the products of the 
 reaction. For example, one of the phenyl- tolyl-ketoximes gives only toluic 
 anilide, whence we can infer that it has the hydroxyl and the phenyl in the 
 syn-position : — 
 
 CH3-OeH^-C-0 _^ CHa-CoH^-C-OH _^ CH^CeH.-C^O 
 NOH ~* N-0 ~* NH^ ' 
 
 while the other gives some of this, but mainly the toluidide of benzoic acid, 
 so that it must be the anti-phenyl compound : — 
 
 '^ ^ * ir ^ ir -^ r • 
 
 HON CHg-CeH^-N CHg-CeH^NH 
 
 By means of these two reactions — the formation of nitrile with the aldoximes 
 iind the Beckmann reaction with the ketoximes — we can assign stereo-formulae 
 to all the known isomeric oximes. 
 
 Now although there are no known cases of isomerism where the two groups 
 attached to the carbon are the same, yet there are many cases in which they 
 are different, where isomers cannot be obtained. And these cases occur not 
 at haphazard, but regularly, in particular classes of oximes. Thus, nearly 
 all aldoximes in which the CH:NOH group is directly attached to the benzene 
 nucleus give isomers ; but they are not obtained with the mixed (aryl-alkyl) 
 ketoximes, such as acetophenone oxime. This is no objection to the theory 
 of Hantzsch and Werner ; it indicates that one isomer is so unstable that 
 it changes spontaneously into the other : which is confirmed by the fact that 
 in all cases of this kind — where only one solid form is known — it gives only 
 a single product in the Beckmann reaction, showing that it consists wholly 
 of one isomer. Thus, when the mixed ketoximes are submitted to the 
 Beckmann reaction, it is always the aromatic group which migrates to the 
 nitrogen, proving the oxime is the syn-aromatic body ; e. g. that acetophenone 
 ^•C-CHg 
 
 *^"^'ho.n • 
 
 The unsymmetrical purely aliphatic ketoximes, which also give no isomers, 
 are oils ; when treated by Beckmann's method they give a mixture of the two 
 possible amides, and are therefore, no doubt, themselves mixtures of the isomeric 
 oximes. This affords a good example of the general rule for tautomeric 
 substances, first laid down by Knorr, that a solid tautomeric body must consist 
 wholly of one form, while a liquid tautomer must always be a mixture of the two. 
 
 The greater stability of one form of an unsymmetrical oxime is an indica- 
 tion that the hydroxyl of the oxime group is more strongly attracted by one 
 of the groups attached to the carbon than by the other ; and from an extensive 
 study of these compounds Hantzsch has compiled a list of radicals in the order 
 in which they attract the hydroxyls : so that if an oxime has attached to the 
 
116 Oocimes 
 
 carbon two of these groups, the more stable isomer will have the hydroxy! 
 on the same side as the radical higher on the list : — 
 
 1. -CHa'COOH (strongest attraction for hydroxyl). 
 
 2. -CHa-CHj-COOH. 
 
 3. -COOH. 
 
 4. -CeHg. 
 
 5. -CgH^X (meta or para). 
 
 6. -CO-^. 
 
 7. -CeH^X (ortho). 
 
 8. -C4H3S (thienyl, the thiophene residue). 
 
 9. -C„H2„ + i(t»>l). 
 10. -CH3. 
 
 In the case of any ketoxime, the chance of getting isomers is greater the 
 nearer the two radicals stand on the list, since the ' preferential ' attraction 
 of the hydroxyl is less. But it is to be noticed that this list applies only 
 to the oximes themselves. If the hydrogen of the hydroxyl is replaced, as 
 in the formation of an ester or a salt, these relations no longer hold. 
 
 Besides the above-mentioned cases of the purely aromatic ketoximes, some 
 of the more complicated fatty ketoximes give isomers, and also apparently 
 some quite simple fatty aldoximes. 
 
 Beckmann has recently announced ' the discovery of a third modification 
 of benzaldoxime and of certain other aromatic aldoximes. These forms are 
 very unstable, and readily change into the usual isomers. They have not yet 
 been fully investigated, and we may assume provisionally that they are only 
 cases of crystalline dimorphism, like that of acetophenone. 
 
 A remarkable confirmation of the theory of Hantzsch and Werner has been 
 afforded by the recent work of MUls and Bain.^ They prepared the oxime 
 of ^-keto-hexamethylene carboxylic acid : — 
 
 H COOH 
 
 HjC CH2 
 
 HgC CH2 . 
 
 C 
 II 
 NOH 
 
 The quinine salt of this oxime-acid was separated by fractional crystallization 
 into two parts. When one fraction was treated with soda and the quinine 
 removed by ether, the resulting solution (of the sodium salt) showed distinct 
 rotatory power, which, however, was destroyed by the addition of hydrochloric 
 acid. It was thus proved that the oxime can exist in two optically active forms. 
 This activity is only possible if the molecule has no plane of symmetry. 
 Now a plane perpendicular to that of the ring will pass through the carboxyl 
 and the hydrogen attached to the same carbon atom, and also through the 
 carbon and the nitrogen of the oxime group, the two valencies joining these last 
 
 ' Ber. 37. 3042 (1904.) ' Proc. V. S. 85. 177 (1909). 
 
Optically active Oocimes 117 
 
 two atoms lying in this plane. Hence the remaining group, the oxime 
 hydroxy], must lie outside it : or, in other words, the three valencies of the 
 nitrogen cannot lie in one plane. The asymmetry is analogous to that of 
 
 carbon compounds of the type f>C< >C=0< • 
 
 This discovery incidentally disposes of the view, to which some chemists 
 
 are still attached, that the isomeric oximes have the structure C\Jt . A body 
 of the formula H COOH 
 
 O NH 
 
 would have the hydrogen and the carboxyl, and also the oxygen and the 
 nitrogen of the oxime, in one plane, perpendicular to that of the ring. It would 
 thus admit of ' geometrical ' isomerism (like that of the hexahydro-terephthalic 
 acids), but not of optical activity, since it would be symmetrical. 
 
 The formation of oximes offers some remarkable instances of stereo-hindrance, 
 in which the carbonyl group of the ketone is, so to speak, blockaded by the 
 other parts of the molecule so that the hydroxylamine cannot get at it, or not 
 without difficulty. These occur where the carbonyl is directly attached to the 
 benzene nucleus, and may be regarded as special cases of stereo-hindrance 
 of the benzoic acid derivatives, as such ketones are really derived from benzoic 
 acid. A ketone of this kind, if it has both the ortho positions on the ring, 
 occupied, will not give an oxime at all. Thus acetomesitylene, 
 
 CH3 
 
 CH3OCOCH3, 
 
 CH3 
 
 it it is heated with hydroxylamine on the water bath, does not react. If the 
 
 two bodies are heated together in a sealed tube to 160°, they go at once to the 
 
 product of the Beckmann reaction, acetomesidide : — 
 
 CH3 
 CH3-O-NHC0CH3, 
 CH3 
 showing that at this temperature the oxime has a transient existence. Again, 
 in benzophenone, < >-C0-< >, the close proximity of the two phenyl groups 
 makes the carbonyl difficult of approach, and in order to form the oxime 
 the body must be heated on the water bath with hydroxylamine for a whole 
 day. If one ortho position on each ring is occupied no oxime can be isolated ; 
 there is no action with hydroxylamine except under very energetic treatment, 
 which converts the oxime as fast as it is produced into the Beckmann product : — 
 
 O-c C> -* O-C0NH<p 
 
 \nOh/ CH3 CHg' 
 
 CH3 CH3 
 1175 I 
 
118 Oscimes 
 
 Finally, if all the four ortho positions on the two rings are occupied by 
 methyl groups, CH3 CH3 
 
 O-co-O, 
 
 CH3 CH3 
 
 hydroxylamine has no action on the compound whatever. 
 
 As examples of the aromatic aldoximes, those of benzaldehyde may be 
 
 considered. j. n-n 
 
 9C'H 
 The a-form, benz-anti-aldoxime, II , is obtained directly. It melts 
 
 at 35°. If treated with hydrochloric acid in ether at 0° it forms its hydro- 
 chloride, melting at 193-195° ; but at the ordinary temperature the hydrochloride 
 of the syn-compound, melting at 66°. If the free a-oxime is treated with 
 hydrochloric acid or sulphuric acid it is converted (through the salt) into the 
 /3-form, the syn-aldoxime. 
 
 The y3-form, benz-syn-aldoxime, II „, is obtained from the a by one 
 
 of the above methods. It melts, when rapidly heated, at 128-130° ; if it is 
 slowly heated, or left for some time in contact with alcohol, or boiled with 
 ether, or distilled in vacuo, it goes over into the a. 
 
 Beckmann's third modification is got by fusing the a-form and cooling 
 rapidly to a low temperature, when it separates in crystals, melting at -I- 5°. 
 
 The a-oxime, when dissolved in alcoholic potash, gives with benzyl iodide 
 mainly a liquid benzyl ether, which on hydrolysis with hydrochloric acid 
 yields benzaldehyde and a-benzyl hydroxylamine, H2N-O-CH20, from which 
 bodies it also can be synthesized. It is, therefore, the 0-ether ^-CH^N-O-CHg^. 
 
 The yS-oxime with benzyl iodide gives mainly a solid benzyl ether, which 
 hydrolyses to benzaldehyde and /3-benzyl hydroxylamine, HNOH'CHj^, and 
 
 can be synthesized from them. It therefore has the formula 0-CH'] ^r _ 
 
 The oximes of anisaldehyde (p-methoxybenzaldehyde) are remarkable from 
 
 CHaOCeH^-CH 
 the fact that while the syn-compound is tasteless, the anti-form, II » 
 
 is intensely sweet. 
 
 As examples of more complicated derivatives, we may consider the oximes 
 
 of benzil, (p-CO-CO-(p. There are two monoximes and three dioximes, as 
 
 required by the theory. They all differ in their physical properties, and are 
 
 obtained from one another by the same methods as the two benzaldoximes. 
 
 Their structure is established by the following facts: — 
 
 1. They are all decomposed by hydrochloric acid to give hydroxylamine and 
 benzil, and therefore have the same carbon chain ^-C-C-^. 
 
 2. They yield both 0- and N-ethers and esters. But the isomerism of the 
 oximes is repeated in the 0-ethers, which shows that it is stereoisomerism, 
 since the only structural isomerism possible, that of the N-ethers, is excluded 
 Ly the independent existence of these compounds, 
 
 3. Further, the isomers can be converted into one another by all the means 
 usually effective in bringing about stereo-chemical change. 
 
 4. As regards the dioximes, their identity of structure is shown by their 
 
Oooimes of Benzyl 119 
 
 giving the same anhydride, and on oxidation with potassium femeyanide the 
 same peroxide :— <A.C— C(6 (i-C— C-(^ 
 
 II II or,-1 II n 
 
 N N aiia N N . 
 
 ^0/ (!) J 
 
 The formulae are determined mainly, as usual, by the Beckmann reaction. 
 The two monoximes are known as a and y : the three dioximes as a, /3, and y. 
 The y-monoxime in the Beckmann reaction gives benzoyl formanilide ; this 
 indicates that it is the syn-phenyl compound: — 
 
 d>.COG(f> d>CO.COH ^COCO 
 
 N-OH N-0 NH^ 
 
 The a-oxime is the unstable one, and has a great tendency to go over into 
 the y ; but by careful treatment it can be made to give dibenzamide, as would 
 be expected from the anti-phenyl compound : — 
 
 0-COC-0 HO-C-^ 0=C-^ 
 
 HO-N ~* fCON ~* 0CONH' 
 With the dioximes the case is more complicated, and the results are less 
 certain. The /3-dioxime in the Beckmann reaction gives oxaniUde ; it must 
 therefore be the anti-compound : — 
 
 <l)-G-G-(j> HOG-COH 0=C — 0=0 
 
 HO-N N-OH ~* ^-N N-^ ~* 0-NH NH^ ' 
 
 It is less easy to determine which of the other two is syn and which is amphi. 
 
 The a-dioxime gives dibenzenyl-azoxime, which has no phenyl on the nitrogen. 
 
 It should therefore be the syn-compound : — 
 
 H0\ O 
 
 0-C G-(p 0-C-OH N di-C/^N 
 
 ^11 ir -^ ^ II II -» ^ II II , 
 
 N-OH HON N C-^ N— C-0 
 
 while the y gives benzoyl phenyl-urea, and so should be amphi : — 
 
 rf>-C C-0 rf>-C-OH N-OH d)-COH N-0 6G=0 NH0 
 
 ^11 II ^ _» ^ II II -» ^ II II ^ _» ^ I I ^ . 
 
 N-OH N-OH N 0-0 N C-OH NH— C=0 
 
 This formula for the y-oxime is confirmed by the fact that it is the form produced 
 
 <p.GGO-<f> 
 by the direct action of hydroxylamine on the y-monoxime II , and 
 
 therefore we should expect that it must have at least one hydroxyl turned 
 outwards. 
 
 On the other hand, the y-oxime is more easily converted into the anhydride 
 than either of the other two ; and from this we should infer that it was the 
 syn-form :— 0.C C-0 d)-G—G-6 
 
 ^11 ir -^ ^ II ir . 
 
 NOH HON N N 
 
 ^0/ 
 
 The peculiar character of the Beckmann reaction has led to various attempts 
 being made to suggest a probable mechanism for it. It does not stand alone. 
 
 X-C-Y 
 There is a general tendency for compounds of the type II to go over into 
 
 i2 
 
120 Oocimes 
 
 ZCY 
 
 ^11 . Of this we have already met with at least three examples besides 
 
 that of the oximes, namely the Hofmann reaction (amines from amides), the 
 Curtius reaction (amines from azides), and the hydroxamic acids. In all these 
 cases the chain is first broken and then put together again in the same unusual 
 way. It is natural to suggest the formation of some intermediate product. 
 Wallach^ has proposed the series: — 
 
 CHq'C'CHo CHo'C'CHn CHq'C'OH 
 NOH N^ NCHg 
 
 This is the kind of explanation one looks for, but the point where the ring 
 is assumed to break seems improbable, and if the theory is applied to the 
 aromatic compounds it leads to a conclusion which has been shown to be 
 false. For whereas on the ordinary view of a mere change of positions 
 the oxime of a para-substituted benzophenone should give a para-anilide, 
 on Wallach's h3rpothesis it should give a meta-anilide : — 
 
 ^|~0x - xCTi "" ■ 
 
 Usual theory. 
 
 ^0" w -* xO;|Ox -* xCr|^Ox • 
 
 Wallach's theory. 
 
 Now di^p-dichloro-benzophenone oxime has been found to give an anilide 
 in which the chlorine is still in the para position,^ which is incompatible 
 with this view. 
 
 The reaction has been further examined,' in the case of the oximes of the 
 a-diketones, such as benzil, by Werner and his pupils. They used benzene 
 sulphonic chloride as the reagent to bring about the change, and showed 
 that it would act in presence of aqueous alkali, and even in pyridine solution. 
 They further showed that the reaction could take place in two ways ; for 
 example, with the (a) anti-phenyl monoxime of benzil : — 
 
 0-C-OH 
 
 /^ II 
 
 Type I (normal). 
 
 NOH 
 
 ^^ift + T^ Type II. 
 N OH 
 
 In the case of the monoximes of the cyclic diketones, such as nitroso- 
 /3-naphthol, benzene sulphonic chloride in pyridine solution always causes 
 a reaction of the second kind : — 
 NOH 
 
 fT.nV^ IT CH-CO-OH. 
 
 H 
 ' Ann. 346. 266. ' Montagne, Bee. Trav. 25. 376 (C. 07. i. 474). 
 
 ' Werner, Piguet, Ber. 37. 4295 (1904); Werner, Detaoheff, £er. 38. 69 (1905). 
 
JBechnann Reaction 121 
 
 In the same way the monoxime of phenanthrene quinone, which in acid 
 solution undergoes the normal Beckmann change (type I) :— 
 CjH^-C^O CoH4-C=0 
 
 II -> I >N, 
 
 CeH^-C^NOH CgHi-COH 
 
 in pyridine solution gives the reaction of type II: — 
 
 C„H,-C=0 _^ C6H4-C<gjj. 
 
 CfiHi-CtiNOH C6H,-C3>T 
 
 So too the a- (anti-phenyl) oxime of benzoin gives benzo-nitrile and benz- 
 aldehyde : — 
 
 </)-CCHOH-^ 0-C CHOH-0 A-C. 
 NOH N OH N ^ 
 
 while the ;3-oxime gives the normal reaction. This reaction is not confined 
 to the oximes of a-diketones, but occurs also, for example, with camphor 
 oxime. 
 
 The occurrence of this second type of Beckmann reaction seems to 
 indicate that an intermediate compound, such as WaUach supposes, is not 
 required ; but that the two groups on the same side of the C=N are 
 actually split off as such, and may either go back again in the reverse 
 order (normal reaction, type I), or may combine with one another (type II),' 
 
 The dynamics of the Beckmann reaction have been investigated by Sluiter,^ 
 who examined the change of acetophenone oxime into acetanUide in presence 
 of sulphuric acid : — 
 
 0CCHo 
 
 HON ^ 
 
 He finds that the reaction takes place in three stages (HX = acid) : — 
 
 I, R'C'Ivi R'G'Ri 
 
 II ^ + HX = II ^ + H„0. (Quick) 
 NOH N-X ^ ^ ' 
 
 II. RCE, ECX 
 
 II = II . (Slow) 
 
 N-X N-Ri ^ ^ 
 
 III. ECX RC=0 
 
 II + HjO = [ + HX. (Quick) 
 
 NRi ' NHEi ^^ ' 
 
 So that the velocity measured is that of reaction II. It was determined 
 (at 60-65°) by pouring the solution after a definite time into water, distUling 
 it, and titrating the acetic acid which came over. The reaction was shown 
 to be monomolecular, the temperature coefficient being about 3 for 10 degrees. 
 The velocity depends largely on the concentration of the sulphuric acid. Thus 
 at 60° an acid of 99-2 per cent, converts nearly the whole in 15 minutes, an 
 acid of 93-6 per cent, only half in 275 minutes. 
 
 ' A similar change in the case of diacetyl monoxime is given by Dials and Stern, Ber. 40. 1622, 
 1629 (1907). 
 
 2 Uec. Trav. 24. 372 (C. 05. ii. 1178). 
 
CHAPTER VI 
 
 NITROSO-COMPOUNDS 
 
 The nitroso-compounds, properly so called, contain the grouping -N=0 
 attached to carbon. (The nitrosamines, in which the -N=0 is joined to 
 nitrogen, will be dealt with later.) They are not a large group, but they 
 present many interesting peculiarities. 
 
 FATTY NITEOSO-DEKIVATIVES 
 
 There are two main methods of preparing them: first, by the oxidation 
 of amines, and secondly, by the reduction of nitro-compounds : — 
 C-NHa -> C-N:0 : C-NOg -* G-N:0. 
 
 It was for a long time supposed that the nitroso-compound could only 
 be obtained in these ways if there was no hydrogen attached to the same 
 carbon as the NHg or NO2 . V. Meyer, in fact, laid it down as a general rule 
 that nitroso-compounds containing the groups -CHj-NO or -CH-NO could not 
 exist, as they changed spontaneously into oximes -CH:NOH or =C:NOH. We 
 now know that there are some exceptions to this rule, which are of great 
 interest. But it still holds good in the main. The majority of known 
 nitroso-compounds are tertiary ; these will therefore be dealt with first, and 
 the exceptional cases of the secondary bodies later. 
 
 1. Oxidation of amines. 
 
 Bamberger has shown' that primary alkylamines, in which the NHj is 
 attached to a tertiary carbon atom, such as tertiary butylamine (CH3)3C-NH2, 
 when oxidized by Caro's acid (sulphomono-peracid) are converted into nitroso- 
 derivatives : — 
 
 (CH3)3C-NH2 -I- 20 = (CH3)3C-N:0 -f HgO. 
 
 This nitroso-butane is the simplest nitroso-compound known, and it exhibits 
 the characteristic properties of the group in a very marked degree. It forms 
 colourless prisms, and is the most volatile solid known ; this volatility 
 is a very general characteristic of the nitroso-compounds. If it is exposed 
 to the air for a short time it disappears entirely. If it is heated in an open 
 tube it vanishes, without melting or boiling, when the temperature reaches 76°, 
 and deposits in the cooler parts of the tube in the form of deep blue drops, 
 which soon turn into a colourless crystalline solid. If heated in a sealed 
 tube it melts at 76° to a deep blue oil. This of course means that under 
 atmospheric pressure its boiling-point is lower than its melting-point ; or, in 
 other words, the vapour pressure of the solid becomes equal to one atmosphere 
 at a temperature below the melting-point : hence it evaporates without melting. 
 But when it is heated in a closed tube, the pressure produced by its own 
 
 » Ber. 36. 685 (1903). 
 
Fatty Nitroso-compounds 123 
 
 vapour raises the boiling-point without appreciably affecting the melting-point, 
 till the boiling-point becomes higher than the melting-point, and the body 
 melts. If an ethereal solution of nitroso-butane is distilled, the whole of it 
 comes over with the ether vapour. 
 
 Nitroso-butane is easily soluble in organic solvents. The solutions are 
 at first, like the solid, colourless ; but they change gradually on standing, and 
 rapidly on warming, to a deep blue. If the molecular weight of the dissolved 
 substance is determined by the freezing-point method in the colourless 
 solution, it is found to be twice that required by the simple formula (CHgJaC-NO. 
 But as the colour grows the molecular weight diminishes, until, when the 
 colour has reached its deepest, it is only half what it was to begin with, and 
 now corresponds to the simple formula. This, again, is a general characteristic 
 of compounds containing a nitroso-group attached to carbon. It has long 
 been known that such bodies when solid are colourless, and when melted 
 or dissolved are blue or green. It has recently been shown that the 
 coloured form always has the simple molecular weight, while the colourless 
 form is a double polymer. In the case of nitroso-butane (and in many 
 other instances) the solution of the colourless (bimoleeular) solid is itself 
 at first colourless and bimoleeular. But on warming, or more slowly 
 even at the ordinary temperature, the colourless double molecules break 
 up into the blue simple molecules of C4H9NO. (Compare nitrogen peroxide, 
 where the coloured NO2 associates to the colourless double polymer N2O4.) 
 It is a remarkable fact that this dissociation of nitroso-butane is delayed by 
 exposure to sunlight. The colourless solution turns blue more rapidly in the 
 dark. This associating action of sunlight has been compared to the behaviour 
 of many unstable compounds, especially aldehydes, which polymerize much 
 more rapidly in the light than in the dark. 
 
 By the same method Bamberger has obtained nitroso-isopropyl acetone 
 by the oxidation of diacetone-amine, the product of the action of ammonia 
 on acetone : — 
 
 CHaXp/CHa-CO-CH, CHgXp/CHaCO.CHg 
 
 This resembles nitroso-butane in its general properties, though it is less 
 volatile. Its solution in organic solvents is colourless at first, but soon turns 
 blue. It is curious that this dissociation occurs much less readily in water. 
 The aqueous solution remains colourless for weeks, though it at once turns 
 blue on warming. It is also worth noticing that in the ease of this substance 
 the monomolecular form is more stable than in nitroso-butane. Nitroso- 
 isopropyl acetone — a colourless crystalline body, which must be the bimoleeular 
 form — melts at 75° to a blue (monomolecular) liquid, and if this is rapidly 
 cooled it remains a blue liquid long enough for its properties to be examined, 
 though it gradually changes into the colourless polymer. The white form 
 is almost insoluble in cold organic solvents, and is not volatile. The blue 
 unstable form is excessively soluble in organic solvents, and is volatile. 
 Corresponding to this difference in volatility is the fact that whereas colourless 
 solutions of this body have no smell, the blue solutions have a strong pene- 
 trating smell. 
 
124 Nitroso-compounds 
 
 Triphenyl-methylamine ^jC-NHg is not oxidized by Caro's acid at all. 
 This may be due to stereo-chemical causes, and is at any rate to be remembered 
 among the many peculiarities of the triphenyl-methyl compounds. 
 
 2. The second method of preparing fatty nitroso-compounds is by the 
 reduction of the nitro-bodies. Those fatty hydrocarbons which contain a 
 tertiary hydrogen atom readily give nitro-compounds on direct nitration, so that 
 the tertiary nitroparaffins are comparatively easy to prepare. For example, 
 
 di-isobutyl gives a nitro-compound of the formula (CH3)2C<^-j^(-.^* ^ ^ ^^> 
 
 and Piloty has shown ' that if this is treated with aluminium amalgam it is 
 reduced to the ;Q-hydroxylamine derivative 
 
 CNO2 -^ CNHOH, 
 which, when oxidized with potassium bichromate, yields the nitroso-compound 
 
 (CH3)2C\jx. J' ^' ^ ^'"^ . This is a colourless substance melting at 54° to 
 
 a deep blue liquid, and resembling nitroso-butane in general properties. 
 
 3. A singular method of preparing a di-nitroso-compound of the fatty series 
 is by the electrolysis of the sodium salt of the oxime of malonic ester ^ : — 
 
 (C02Et)2C-N:0 
 2(C0,Et),C=N0.Na^^^^^^^^r^^^. 
 
 The sodium goes to the cathode, while the residues liberated at the anode 
 combine in pairs, as in Kolbe's electro-synthesis from the fatty acids. As the 
 product is colourless, it is probable that the two nitroso-groups combine with 
 one another. 
 
 4. Nitroso-compounds are formed by the action of nitrous fumes on the 
 acyl derivatives of the fatty esters, especially when the acyl is attached to 
 a tertiary carbon atom.' The acyl group is expelled and the nitroso takes 
 its place, e. g. CHa-CO-CH-COaEt OiN-CH-COaEt 
 
 CH.COaEt ~* CHa-COaEt* 
 
 5. Piloty* has discovered a peculiar reaction of very general application, 
 which gives rise to the halogen derivatives of the nitroso-compounds. ° This 
 consists in treating a ketoxime with bromine in the presence of pjrridine. It is 
 possible, though there is no evidence of this, that a hypobromite is formed 
 as an intermediate product : — 
 
 (CH3)2CfcNOH + Br^ = HBr -f (CH3)2a=NOBr -^ (0^^\0<^q ■ 
 
 The action of the pyridine is not understood. The product, brom-nitroso- 
 propane, is a mobile blue liquid boiling at 83°, which has a strong smell 
 resembling at once bromine and acrolein. It seems to have no tendency 
 to form the colourless double polymer. 
 
 A remarkable example of the tendency of the NO group to polymerize 
 is furnished by Piloty's chlor-nitroso-derivatives of diketo-hexamethylene.° 
 The dioxune of para-diketo-hexamethylene (obtained from succino-succinic ester) 
 
 1 Ber. 31. 457 (1898). = Ulpiani, Eodano, C. 06. i. 449. 
 
 s Schmidt, Widmann, Ber. 42. 497, 1886 (1909). » Ber. 31. 452 (1898). 
 
 = Cf. Ponzio, C. 06. I. 1692. e Ber. 35. 3101 (1902). 
 
Fatty Nitroso-compounds 125 
 
 is converted by bromine and still more readily by chlorine into the corre- 
 sponding di-halogen-dinitroso-derivative : — 
 
 NOH CI NO 
 
 C C 
 
 CH2 CH2 ~* CH2 CH2 ' 
 
 T X 
 
 NOH CI NO 
 
 This body is a blue crystalline substance, very soluble in organic solvents, 
 melting at 108°. It changes slowly in alcohol, more rapidly in acetic acid 
 containing hydrochloric acid, into a colourless crystalline substance of the 
 same composition and the same molecular weight. This new isomer, when 
 dissolved in methyl alcohol or acetone, gives a colourless solution which 
 turns blue on heating, the colour disappearing again on cooling. The reactions 
 of this isomer are in general similar to those of the original substance. 
 
 A consideration of the structure of the original body will show at once 
 what the cause of this isomerism must be. This substance is obviously 
 capable of the same stereoisomerism as hexahydro-terephthalic acid. The 
 four valencies of the para-carbon atoms lie in one plane, and so two dichlor- 
 dinitroso-derivatives can exist, a cis-trans, in which the two nitroso-groups 
 are on opposite sides of the plane of the ring, and a cis, in which they are 
 both on the same side. On the ordinary convention we can represent them 
 on a plane thus : — 
 
 C1\/N:0 C1\/N:0 
 
 
 
 0:N^\C1 CF^N:0 
 
 Cis-trans. Cis. ' 
 
 the essential point being that in the cis-form the two NO groups are near 
 to one another, while in the cis-trans they are far apart. The blue crystalline 
 body which is the product of the original reaction is clearly the cis-trans. 
 
 The two nitroso-groups are too far apart to combine with one another, and 
 the body could only go into the colourless form by the combination of two 
 molecules. But as is shown in the case of brom-nitroso-propane, which is 
 a blue solid, there is less tendency to polymerization in these haloid derivatives 
 than in the simple nitroso-compounds. Hence no such polymerization occurs. 
 But when the body is allowed to stand in solution, it changes into the more 
 stable cis-form. This is the colourless isomer. In this the two nitroso- 
 groups are on the same side of the ring, and are therefore capable of combining. 
 We thus get the two blue nitroso-groups changing into the colourless NjOg 
 within the molecule : and this is why the colour in this case disappears without 
 an increase of molecular weight. That the product really has this formula 
 is shown by the fact that its solution turns blue on heating. This cannot 
 be due to the regeneration of the original cis-trans modification, because the 
 blue colour immediately disappears on cooling, whereas the cis-trans only 
 
126 Nitroso-compounds 
 
 changes slowly into the cis. It must be caused by a separation of the two 
 NO groups in the cis-form itself: — 
 
 Clx/N-0 
 
 
 Clx/N:0 
 
 
 
 
 ^ 
 
 
 
 Cl/^N-O 
 
 
 C1/^N:0 
 
 There is an obvious analogy between this case, where in one stereoisomer 
 the two groups interact, while in the other they do not, and that of fumaric 
 and maleic acids, where the cis-form (maleic acid) alone forms an anhydride. 
 
 The recently discovered secondary nitroso-compounds containing the group 
 
 ^C'yi^^ are obtained in two ways. 
 
 First, Piloty has shown' that acetaldoxime, when treated with chlorine, 
 undergoes the same change as a ketoxime, giving a chlor-nitroso-compound : — 
 
 CH3-C<gQjj + CI2 = HCl -I- CH3-C-N:0. 
 
 It is a colourless crystalline substance, which gives a colourless solution in 
 liquid hydrocyanic acid at - 10°, in which it is bimolecular. It melts at 65° 
 to a deep blue liquid, and its solutions in organic solvents at the ordinary 
 temperature are blue and monomolecular. If these solutions are warmed, 
 or if the substance is kept fused for some time, the colour disappears, and 
 the body is converted into the isomeric chloro-aldoxime, or hydroxamic acid 
 chloride : — 
 
 CHs-C-NiO -> OH •C<pf*-^ - 
 \C1 ^^ 
 
 Other aldoximes, such as benzaldoxime, do not give any such nitroso- 
 compound. When they are treated with chlorine a transient green or blue 
 colour is produced, showing that a nitroso-compound is formed, but this 
 immediately vanishes, with the formation of the chloro-oxime. It is clear 
 that the normal action of chlorine on an aldoxime is : — 
 
 In general the second reaction is so rapid that the nitroso-compound cannot 
 be obtained ; but in the particular case of acetaldoxime it is sufficiently slow 
 to enable us to isolate the intermediate stage. 
 
 The second method of preparing the secondary nitroso-compounds is by 
 the action of nitrous fumes on unsaturated hydrocarbons. But this will be 
 discussed later,^ in dealing with the rather complicated series of bodies to 
 which this reaction gives rise. 
 
 The characteristic tendency of the nitroso-compounds to polymerization, 
 with an accompanying loss of colour, a tendency which also occurs, as has 
 been remarked, with the same accompaniment in the case of nitrogen 
 peroxide, is not possessed by all the nitroso-bodies in an equal degree. On 
 
 1 Ber. 35. 3101 (1902). 2 p_ 128. 
 
Fatty Nitroso-compounds 127 
 
 the contrary, the relative stability of the mono- and bimolecukr forms varies 
 very greatly in different cases. We can arrange these compounds in a series ' 
 in the order of the increasing stability of the bimolecular form. In all cases 
 the monomolecxdar form is characterized by being coloured and soluble in 
 orgajiic solvents, the bimolecular being colourless and much less soluble. 
 At one end of the scale are certain cases, particularly of halogen derivatives, 
 where the bimolecular form is so unstable that it cannot be obtained at 
 all ; thus brom-nitroso-propane (the acetone derivative) is a blue oil ; and 
 even the pinacoline derivative, /CH 
 
 (CHgJaC-C-Br % 
 \N:0 
 
 which is solid, does not polymerize, but forms blue crystals. In the commoner 
 cases the solid substance is the colourless form (as nitroso-butane), but on 
 melting it turns blue. 
 
 The solutions are sometimes colourless below 0° (acetaldoxime compound), 
 but blue at the ordinary temperature: in other cases they are colourless 
 when freshly prepared, but turn blue on standing (nitroso-butane), while in 
 cis-dichloro-dinitroso-hexamethylene the colourless form is so stable that the 
 solution only goes blue on heating, and loses its colour immediately on cooling. 
 
 This brings us to a group of bodies, not hitherto mentioned, which were 
 previously supposed to belong to a different class, but which obviously only 
 continue the series of the nitroso-compounds. They are the so-called bis- 
 nitrosyl-compounds. If benzyl -hydroxy lamine, 0-CH2-NHOH, is oxidized ^ (which 
 should produce a nitroso-compound), bis-nitrosyl-benzyl ^•CH2-N2O2-CH2-0 is 
 obtained. This is a colourless crystalline body which it is very difficult to 
 break up at all, and hence it has been supposed that the N2O2 group 
 must have some peculiar structure. It behaves, however, in other ways 
 exactly like the colourless bimolecular nitroso-compounds ; and by prolonged 
 boiling with acetic and hydrochloric acids it can be split up into two molecules, 
 when it gives not the true nitroso-derivative, but benzaldoxime. This, however, 
 is to be expected. The vigorous action necessary to break up the firmly 
 united N2O2 group is sufficient to change the very unstable primary nitroso- 
 compoimd ^-CIIa-NO so formed into the isomeric benzaldoxime. In this body, 
 too, the N2O2 group is stable enough to give some evidence as to its structure. 
 
 We have hitherto assumed without any evidence that this is J. ' - The 
 
 behaviour of bis-nitrosyl-benzyl indicates that the two nitrogen atoms are 
 linked together, which is really all that is known about this grouping. For 
 if the body is treated with hydrochloric acid in chloroform solution, it is 
 converted into a mixture of benzoyl hydrazine, 0-CO-NII-NH2, and benzoyl- 
 benzal-hydrazine, 0-CO-NH-N:CH0. Hence the two nitrogens must be joined, 
 and the body may be written 
 
 0-CH2-N-N.CH2-0 0.CH2-N=N-CH2.0 
 
 II or II II 
 
 0-0 o 
 
 An example of a still more firmly united N2O2 group is afforded by 
 
 ' Piloty, JBer. 35. 3090 (1902). ' Behrend, Konig, Ann. 263. 212 (1891). 
 
128 Nitroso-compounds 
 
 a class of compounds prepared by Baeyer, by treating certain terpene deri- 
 vatives with nitrous acid. Thus from caron is obtained bis-nitrosyl-caron, 
 CioHi50N202-CioHi50. In this body the nitrogens are so firmly linked 
 together that they cannot be separated at all. If the compound is broken 
 up, it does not split between the nitrogen atoms, but between the 'S^O^ 
 group and the carbon, losing one caron residue and giving CioHijO-NgOg-H, 
 
 It follows from this gradation of properties that all the compounds dealt 
 with belong to the same class, and that the stability of the N2O2 group 
 increases by successive stages, though over a wide range. 
 
 We may consider here a group of substances whose formulae, reactions, 
 and nomenclature are rather confusing ; the nitrosites, pseudo-nitrosites, 
 nitrosates, and the like, bodies obtained for the most part by the addition 
 of the higher oxides of nitrogen (especially the trioxide and peroxide) to 
 compounds with a double carbon link.^ 
 
 These nitrous vapours act as if they consisted of the trioxide, which adds 
 itself on to the two doubly linked carbon atoms as NO + NOg. Of these two 
 groups the first must join directly through nitrogen to the carbon ; the second 
 may also do this, forming a nitro-group C-NOj, or it may join through oxygen, 
 forming a nitrous ester C-O-NO ; or this last may be the primary product, but 
 may be oxidized further to a nitric ester C-O-NOj. We thus get the following 
 groupings : — 
 
 i-r i-V- i-t- 
 
 NO NO2 NO ONO NO2 ONO2 
 
 Pseudo-nitrosites. Nitrosites. Wallach's Nitrosates. 
 
 -?-?- 
 
 NO2 ONO 
 
 Gabriel's 'Dinitriire', sometimes also called nitrosates. 
 The first two classes, being nitroso-compounds, may, and in fact generally do, 
 ■occur in the bimolecular form, e. g. : — 
 
 -C C- 
 
 NO, 
 
 N2O2 
 
 NO2 
 
 -C C- 
 
 Which of these bodies is obtained depends on the nature of the unsaturated 
 compound employed. If the nitrous fumes act on a simple ethylene hydro- 
 carbon,'^ or on a hydroaromatic compound, derivatives with the oxygen link 
 (nitrous or nitric esters, nitrosites or nitrosates) are formed : e. g. with ethylene 
 
 itself the nitrosite CH -CH 
 
 72,2. 
 
 NO ONO 
 
 1 See especially Wieland, Ann. 328. 154; 329. 225 (1908); Wieland, Bloch, Ann, 340. 
 63 (1905). 
 
 2 Wieland, Ann. 328. 154. s Schmidt, Ber. 35. 2S23 (1902). 
 
Nitrosites, Nitrosates, &,c. 129 
 
 The phenyl-ethylenes, especially those with positive substituents, such as 
 anethol (i)-methoxy-phenyl-ethylene), are able to attach both the nitrogen atoms 
 to carbon directly, with the formation of pseudo-nitrosites, such as 
 
 CH30-CeH4-CH-CH2N02 
 
 CH30-CcH4-CH-CH2-N02 
 
 With the accumulation of negative groups (especially if the other doubly 
 linked carbon is joined to carbonyl, as in benzalacetophenone, ^■CH=CH-CO-^), 
 though it is probable that these pseudo-nitrosites are the primary product, 
 they become unstable, and the bodies actually obtained are either the nitro- 
 
 0-<f!- 
 
 -CHR 
 
 oximes, ^ T , or their decomposition products, or unsaturated nitro- 
 
 ^•CH=CE 
 compounds, I . At the same time a certain quantity of nitro-nitrous 
 
 NOg 
 
 esters are formed, such as 1 1 . 
 
 ONO NOa 
 
 The main distinction in these various groupings depends on whether 
 the nitrogen is attached to carbon directly, or through oxygen. In the latter 
 case (nitrosites or nitrosates) the nitrogen so attached is always removed with 
 
 'H2 
 
 ONO NO ' 
 
 have the -O-NO group removed even by the action of alcohol. The nitrosates 
 or ' Dinitrure ' split off the nitrous ester group even more easily, losing nitrous 
 acid spontaneously, and thus passing into unsaturated nitro-ketones. The 
 result of this is that where these nitrosates are formed by the action of 
 nitrous fumes, there appears to be a direct nitration, the ketone being finally 
 converted into a nitro-ketone, though in fact an addition product is first formed : — 
 
 ArCH=CHCOE ArCH CHCO-R ArCH-CCOE 
 
 the greatest ease. Thus the true nitrosites, like that of ethylene, T .f_ T. ^ 
 
 ?' 
 
 IT -9. r + HNO, 
 
 ).2 ONO NO2 NO2 
 
 This suggests that similar intermediate addition products are formed in 
 other cases of nitration as well. 
 
 The pseudo-nitrosites (also known as a- nitrosites) in which both nitrogen 
 atoms are united to carbon, are more interesting. They are all bimolecular, 
 owing to the polymerization of the nitroso-group ; thus the styrol compound is : — 
 
 ^-CH-CHg-NOa 
 
 N2O2 
 
 0-CHCH2NO2 
 
 They are colourless crystalline compounds, very easily decomposed. They 
 give green solutions in hot acetic acid, owing to the production of the mono- 
 molecular form, e. g. ^-CH-CHa-NOg 
 
 NO 
 
130 
 
 Nitroso-conipounds 
 
 But on cooling, the solution loses its colour and deposits crystals of the 
 bimolecular form. 
 
 Their various decompositions all depend on two initial stages : — 
 
 1. If they are boiled with alkali they first go into the monomolecular 
 form, and then this, like all secondary nitroso-compounds in presence of 
 alkali, changes over into the oxime: — 
 
 0CH-CH2-NO2 0-CCH2-NO2 
 
 N:0 ~* NOH 
 
 2. Under other conditions, especially ia the presence of mineral acids, they 
 split off the whole N2O2 group as hj^onitrous acid (i. e. as nitrous oxide and 
 water), and give unsaturated nitro-compounds : — 
 
 H 
 Ar-CH-C-NOa 
 
 2 ArCH=CH-NO, 
 
 H 
 
 N2O2 
 H 
 
 Ar-CH-(^-N02 
 H 
 
 + N,0 + H,0 
 
 A remarkable reaction of the pseudo-nitrosites, which led to a misconception 
 as to their structure, is their conversion into the so-called glyoxime peroxide. 
 
 These ' peroxides ' contain the group 
 
 ArC-CR 
 
 II 11 , 
 N2O2 
 
 and as they are converted by 
 
 reduction with zinc and acetic acid into the dioximes of a-diketones, they were 
 formerly thought to have the two nitrogen atoms joined by a peroxide link : — 
 
 -C C- -C 0- 
 
 II II -» II II • 
 
 NOON NOH NOH 
 
 Later work,' however, has shown that they contain a 5-ring: — 
 
 -C C- 
 
 II |>0. - If 
 
 N-O-N ^' ■" 
 
 Thus other reducing agen,ts, such as tin and hydrochloric acid, convert them 
 into furazane derivatives : — 
 
 -C C- -C C- 
 
 N-O-N'^ N-O-N 
 
 Wieland has proved that the production of these peroxides from the 
 
 pseudo-nitrosites takes place through the nitro-oximes (reaction (1) above), which 
 
 yield peroxides if they are first dissolved in alkali (giving the aci-nitro-salt), 
 
 and then the solution is acidified. This explains the mechanism of the 
 
 reaction:— Ar-C C-R Ar-C— C-R 
 
 II II -» II T\n • 
 
 NOH N:0 N-O-N/0 
 
 HO 
 
 Wieland, Semper, Aim. 358. 36 (1908.) 
 
Nitrosites, Nitrosates, S^c. 131 
 
 But it is obvious that the nitro-oxime can exist in two stereoisomeric forms, 
 and of these it is the anti-aryl form which will give the peroxide. Now, according 
 to Hantzsch's investigations, the presence of positive substituents (such as 
 methoxyl, CHg-O-) on the aryl will favour this form, while their absence, or the 
 presence of negative substituents, makes the syn-aryl form, which cannot 
 undergo this change, the more stable. The facts bear out this view. If the aryl 
 group of the pseudo-nitrosite contains positive substituents, the peroxide is 
 formed at once, and the nitro-oxime cannot be isolated, whereas if these are 
 absent, the nitro-oxime can be isolated, and is often quite a stable substance. 
 
 As has been pointed out, Wieland finds that the more negative bodies will 
 often give no pseudo-nitrosites at all, but only the nitro-oximes. This probably 
 means that the pseudo-nitrosites, which we may assume to be the primary 
 products even in these cases, have less tendency to polymerize, and therefore 
 undergo the normal change of the monomolecular nitro-secondary nitroso- 
 compounds into the nitro-oxime. 
 
 AEOMATIC NITEOSO-COMPOUNDS 
 
 These, like the fatty derivatives, may be made either by the oxidation 
 of amines or by the reduction of nitro-compounds. The simplest member 
 of the group, nitrosobenzene, has long been known in solution. It was first 
 prepared by Baeyer in 1874, by the action of nitrosyl bromide, NOBr, on 
 mercury phenyl in benzene solution. But it was not isolated untU nearly 
 twenty years later. In 1893 Bamberger obtained it in the pure state by 
 various methods : by the oxidation of diazobenzene with potassium ferricyanide 
 or with potassium permanganate in alkaline solution, or better, by oxidizing 
 i3-phenyl hydroxylamine, 0NHOH, with cold chromic acid mixture. This 
 last is the method generally employed for preparing the aromatic nitroso- 
 compounds. The nitro-compound is reduced, commonly with zinc dust and 
 in neutral solution ; this converts it first into the nitroso-compound and 
 then into the /3-hydroxylamine : — 
 
 Ar-NOa -^ ArNO -> Ar-NHOH, 
 but the nitroso-group is so rapidly reduced further that it is practically 
 impossible to stop the reaction at this stage. It is therefore allowed to 
 go on another stage, giving the hydroxylamine derivative, and this is then 
 oxidized back to the nitroso-compound. The nitroso-derivatives are so reactive 
 that the experimental conditions, which differ somewhat in each case, must 
 be very exactly observed. 
 
 The reduction can be effected electrolytieally. Thus, if nitrobenzene is 
 electrolysed ^ in neutral solution under suitable conditions, a good yield of 
 nitrosobenzene is obtained. No diaphragm is used, and it seems that at the 
 cathode the nitrobenzene is reduced to /3-phenyl hydroxylamine, which is 
 reoxidized at the anode to nitrosobenzene. Nitroso-compounds are also formed 
 when an aromatic amine is oxidized with hydrogen peroxide or Caro's acid ; 
 but it is almost impossible in this case to isolate them. 
 
 ' Dieffenbaoh, C. 08. i. 911. 
 
132 Nitroso-compounds 
 
 Nitrosobenzene can also be obtained by acting with nitrosyl chloride on 
 phenyl magnesium bromide,' 0-MgBr + NOCl = 0-NO + MgBrCl. 
 
 The simplest and best known aromatic nitroso-derivative, nitrosobenzene, 
 0'N:O (whose properties may be taken as typical of the whole class), forms 
 large colourless crystals melting at 68° to an emerald green liquid. It is 
 remarkable for being extraordinarily volatile — even to some extent with ether 
 vapour. If its solution is distilled with steam, the whole of the nitroso- 
 benzene comes over in the first few minutes. The same colour phenomena 
 are observed with these bodies as with their fatty analogues ; and the same 
 changes in molecular weight have been found to accompany them, the substance 
 being bimolecular in colourless solutions (where these can be obtained) and 
 monomolecular when coloured. The aromatic nitroso-compounds show a great 
 stability of the monomolecular form, the solutions being usually dissociated at 
 the ordinary temperature. But it is to be noticed ^ that if there are two methyl 
 groups in the ortho position to the -N:0, as in nitrosomesitylene and nitroso- 
 meta-xylene, jj:0 N:0 
 
 CHj-pj-CHa CH3-A-CH0, 
 
 CH;, 
 
 the stereo-effect which they produce favours the stability of the bimolecular 
 form : it promotes association. The solutions of these two bodies in organic 
 solvents are of a very pale colour at the ordinary temperature, and contain 
 mainly double molecules ; they do not darken on standing but only on 
 warming, and gradually lose their colour again on cooling. 
 
 The great reactivity of the nitroso-group indicates that it is under great 
 strain ; and this is further shown by the fact that if nitrosobenzene is ignited 
 in oxygen under a pressure of 25 atmospheres it explodes. 
 
 Bamberger has examined the decomposition products formed under a variety 
 of conditions in great detail : as an example of the scale on which he worked it 
 may be mentioned that in a single set of experiments ' on the action of alkali at 
 100°, 1,330 giams of nitrosobenzene were heated with alkali under pressure in 
 19 champagne bottles, and 12 different decomposition products were isolated 
 and identified. Only the more important reactions of this substance can be 
 dealt with here. 
 
 On reduction it yields first /3-phenyl hydroxylamine and then aniline. 
 When treated with hydrazobenzene, it oxidizes it quantitatively to azobenzene, 
 being itself reduced to phenyl hydroxylamine : — 
 
 0-N:O -f 0-NH-NH-^ = ^NHOH -1- ^•N=N-^. 
 On oxidation it is converted into nitrobenzene. With hydroxylamine it gives 
 diazobenzene : — 
 
 ^•N:0 + H2NOH = 0N=NOH + H^O, 
 
 not, as originally stated by Bamberger, the anti, but the syn form.* 
 It combines ■with aniline to form azobenzene : — 
 
 0-N:O + HjN^ = ^•N=N-0 -1- H^O. 
 
 1 Oddo, Qaz. 39. i. 659 (1909). '' Bamberger, Ber. 34. 3877 (1901). 
 
 s Ber. 33. 1939. * Hantzsch, Ber. 38. 2056 (1905). 
 
Aromatic Nitroso-compounds 183 
 
 and with phenyl hydroxylamine very readily to give azoxybenzene : — 
 
 ^•N:0 + •^H>N.0 = 0-N— N-0 + HgO. 
 
 It decomposes of itself in benzene solution if exposed to the light, giving mainly 
 azoxybenzene, but also a variety of other products.' 
 
 The aromatic nitroso-derivatives which also contain an amine-group (the 
 two nitrogen atoms being, of course, joined to different carbon atoms and 
 not to one another) are obtained more easily by special methods. 
 
 Those with primary amine-groups are got from the nitrosophenols or 
 quinone oximes (a class of bodies which will be described later) by heating 
 them with ammonium chloride and ammonium acetate, the phenol hydroxyl 
 being replaced by NHj. Para-nitroso-anUine, 0:N-{ ^-NH,, which is 
 obtained in this way from benzoquinone monoxime (^-nitrosophenol), 
 
 0:N-O0H + NHg = 0:NC>NH2 + HjO, 
 forms steel-blue needles. It gives salts with acids, which are yellow in 
 solution, and also with potash. But if it is boiled with potash, the NHj is 
 split off as ammonia, and the nitrosophenol regenerated. 
 
 The ^-nitroso-derivatives of the secondary amines are got from the nitrosa- 
 m.ines (in which the nitroso-group is attached to nitrogen) by the action 
 of alcoholic hydrochloric acid, which causes the nitroso-group to migrate (in 
 the normal manner) from the nitrogen to the para position on the ring : — 
 
 HC3N<g^' ^ 0:N.<3.N<^H3. 
 
 They are green crystalline substances, more basic than the primary compounds, 
 which dissolve in alkalies and are reprecipitated by carbon dioxide. As 
 secondary amines they react with nitrous acid to give the nitroso-nitrosamines, 
 
 such as 0:N<ON<^^. 
 
 The corresponding derivatives of the tertiary amines are the best known 
 of all the nitroso-compounds. They are formed directly with great ease 
 by the action of nitrous acid on the tertiary amines. They are of commercial 
 importance for the production of certain dyes, such as methylene blue and 
 the indophenols. 
 
 Para-nitroso-dimethyl-aniline, 0:N-^ )-N< pjy^ , forms large green crystals. 
 
 The stability of the monomolecular form is shown by the fact that the 
 green colour does not perceptibly diminish if the toluene solution is cooled 
 to below — 100° with liquid air. The hydrochloride is a yellow salt, and gives 
 a deep yellow solution in water. On reduction it gives ^-amino- and on 
 oxidation p-nitro-dimethyl-aniline. On boiling with aqueous alkalies the amino- 
 group is split off as dimethylamine, and nitrosophenol is formed. 
 
 A remarkable compound of this class is benznitrosolic acid,'* 0-C<^„' „' 
 corresponding to the nitrolic acids E-C<^i^q|t. It is obtained by the action 
 
 1 Bamberger, Ber. 35. 1606 (1902). ^ Wieland, Bauer, Ber. 39. liSO (1906). 
 
 1175 s. 
 
134 Nitroso-compounds 
 
 of potash on the benz-oxy-amidines, which first give a red azo-cotapound, and 
 then break up to form the blue nitrosolic salt and an amidoxime : — 
 
 The total effect of these reactions is : — 
 
 2 -NHOH = -NHj + -N:0 + HgO, 
 which may be compared to : — 
 
 2 ^CHiO -^ 0-CH2OH + 0COOH. 
 
 Benznitrosolic acid forms an unstable green solution. It gives a red 
 silver salt. It is easily reduced to an amidoxime (-N:0 — > -NHj), and 
 oxidized to a nitrolic acid (-N:0 —* -NOj). When treated with mineral 
 acids it forms benzonitrile and nitrous acid, the reaction no doubt being : — 
 
 When the silver salt is treated with iodine, a quantitative yield of diphenyl- 
 glyoxime peroxide is produced ; this is no doubt due to the intermediate 
 formation of the peroxide of the nitrosolic acid: — 
 
 2 ^-^Slg ^ I^ = '^■<:n:N>C-^ + 2 Agl -> '^'l~^0 + 2 NO. 
 
 The simplest nitrosolic acid, methyl- nitrosolic acid, H-C"^-jtqtt, has recently 
 
 been obtained by Wieland and Hess,' by a method exactly analogous to that 
 given above for the benzyl derivative : by the action of potash on form-oxy- 
 amidoxime : — 
 
 Its salts resemble those of the benzyl compound in colour, but they are 
 more explosive. They are converted by the prolonged action of alkali into 
 the salts of prussic and nitrous acids: — 
 
 HC<^Qjj = HON + 0=NOH. 
 
 The free acid is green in solution, but crystallizes in a yellow bimolecular 
 form. It is very unstable, and decomposes into hyponitrous acid and fuhninic 
 acid : — 
 
 2 H-C^JJ-Jg = HON=NOH + 2 C=NOH, 
 
 which is analogous to the change of methyl-nitrolic acid into fulminic and 
 nitrous acids: — 
 
 HC<^q|j = HNO2 + C=NOH. 
 
 As methyl-nitrosolic acid contains the secondary nitroso-group, =C\tt' , we 
 should expect the hydrogen to migrate readily from the carbon to the nitroso- 
 
 1 Ber. 42. 4175 (1909). 
 
Aromatic Nitroso-compounds 135 
 
 group, which would give HON=C=NOH, the dioxime of carbon dioxide. This 
 change cannot, however, be brought about at all, owing, probably, to the 
 reluctance of the carbon atom to attach itself by two double links, which is 
 shown also, for example, in the difficulty of forming hydrocarbons (allenes) 
 of the type HaC^C^CEa. 
 
 QUINONE OXIMES OR NITEOSOPHENOLS 
 
 When hydroxylamine acts on a quinone, a monoxime is formed, such as 
 0:< >:NOH. These bodies are identical with the substances obtained by 
 treating phenols with nitrous acid, and we should therefore expect them 
 to be not oximes but nitrosophenols : — 
 
 H0<O-H + HONO = HOC>N:0 + H^O. 
 
 This latter structure is also supported by the fact that they can be made 
 by boiling the nitroso-tertiary aromatic amines with alkali, which replaces 
 the amine-group by hydroxyl: — 
 
 (CH3)2N-<3N0 + H2O = (CH3)2NH + HO-O-^O. 
 
 The behaviour of these bodies is obviously tautomeric, and it is by no 
 means easy to determine their real structure. The quinone oxime formula 
 is supported by their production from quinone and hydroxylamine, and also 
 by their conversion by the further action of hydroxylamine into the quinone 
 dioximes, as HON— < > =]S[OH. Their formation from phenol and nitrous 
 acid, and from nitroso-dimethyl-anUine, and (less certainly) their oxidation 
 to nitrophenols, support the nitrosophenol structure. It is at least probable 
 that their salts are derived from quinone oxime, for when the silver salt 
 is treated with methyl iodide it gives an O-ether,' which must have the 
 formula 0: ( ) :NOCHa, since it is also produced by the action of a-methyl 
 hydroxylamine on quinone: — 
 
 0:0:0 + H^N-OCHa = 0:O=N-0CH3 + HgO. 
 
 Moreover, the isomeric nitrosophenol ethers are also known,' being obtained 
 by the oxidation of the aminophenol ethers with Caro's acid : — 
 
 CHgO-CD-NHa -» CH30C>N:0. 
 
 We may therefore conclude that the salts are derived from the quinone 
 formula. It is commonly assumed that the free compound is also quinoid, 
 but this is much more open to dispute. The argument from the colour, which 
 is usually of great value in such cases, is here complicated by the tendency 
 of the nitroso-group to polymerize to colourless compounds. Nitrosophenol' 
 itself occurs in two crystalline forms, one colourless, the other brownish green. 
 Its solutions in water, alcohol, and ether are green, and in ether it has been 
 
 ' Bridge, Ann. 277. 79 (1893). ^ Baeyer, Ber. 35. 3034 (1902). 
 
 » Sluiter, Bee. Trav. 25. 8 (C. 06. i. 756). 
 
 k2 
 
136 Nitroso-compounds 
 
 shown to be monomolecular. In benzene, chloroform, and carbon bisulphide it 
 is yellow ; in benzene its molecular weight is one and a half times that required 
 by the simple formula. This behaviour is quite compatible with its being 
 a nitroso-compound. There is also evidence that in forming the (probably 
 quinoid) salts it undergoes an intramolecular change. It gives the ammonia 
 reaction ' which is characteristic of pseudo-acids, and the colour changes to 
 yellow. This would seem to show that the free compound has a different 
 structure from its salts, and hence is a nitrosophenol. 
 
 A case has recently been discovered " of a quinone oxime which is capable 
 of existing in two very definite modifications, which there is strong reason 
 to believe correspond to the two tautomeric formulae. If the monoethyl 
 ether of resorcin is treated with nitrous acid, a body is produced whose formula, 
 regarded as a quinone oxime, can be shown to be 
 
 
 A=NOH. 
 
 It is a golden-yellow crystalline soUd, melting at 146'. If it is recrystallized 
 from anhydrous non-ionizing solvents like benzene, toluene, or carbon bisul- 
 phide, it is converted into a green substance, strongly dichroic, and of a 
 different crystalline form. The green form changes back into the yellow at 
 130', or on recrystallizing from a dissociating solvent such as alcohol. The 
 change seems to be brought about by the solvent at once. Solutions in 
 benzene are always green, those in dissociating solvents yellow. Both forms 
 give the same sodium salt when treated with soda, and the alkaline solution 
 when acidified always precipitates the yellow modification. It is highly 
 probable that these two forms correspond to the two tautomeric formulae, 
 
 OH O 
 
 I II 
 
 ry^-.O and |A=NOII ; 
 
 C,H,0-U C^H.O-i; 
 
 and, further, that the unstable green form is the nitrosophenol (since green 
 is the characteristic colour of the aromatic nitroso-compounds) and the stable 
 yellow form the quinone oxime. 
 
 A somewhat similar case, but one which it is less easy to interpret, is that 
 of nitroso-orcin.' This body, if it is to be considered as a nitrosophenol, has 
 the structure N:0 
 
 OH 
 
 If its potassium salt is treated with acid below 0°, the free compound 
 separates as yellow crystals, decomposing at 16B°. If it is acidified when 
 warm, red crystals are obtained, which change into the yellow form at 124-125°. 
 The chemical behaviour of the two forms is the same, but they are not 
 
 ' Farmer, Hantzsoh, Ber. 32. 3101 (1899). " Henrioh, C. 04. ii. 1539. 
 
 ' Hantzsch, Sluiter, Ber. 39. 162 (1906). 
 
 CHg^^-OH 
 
Nitrosophenols 137 
 
 merely dimorphic, as the difference persists in solution (unlike nitroso- 
 resoroin). The yellow form gives an orange-yellow solution in water, whose 
 conductivity indicates a value of the dissociation constant K = 0-037. The 
 red form gives a dark orange-red solution, of a higher conductivity, for which 
 K = 0051. If they are quite pure the two solutions remain unchanged for 
 several days, but if not, their conductivities approach the same value. The 
 addition of alcohol — even as little as 10 per cent. — brings about equilibrium 
 from either side at once. If the solutions are evaporated m vacuo at the 
 ordinary temperature they both leave the same mixture of the two solid 
 forms ; but if evaporated at 60-70° they both leave only the yellow form. 
 The two forms give in all cases the same salt, even when treated with 
 liquid ammonia at -70°. From the solution of the potassium salt acids 
 precipitate at 0° the yellow form, at higher temperatures the red. The 
 solution of the potassium salt at high dilution, when treated with an exact 
 equivalent of hydrochloric acid, shows by its conductivity that it contains 
 only the yellow form. 
 
 These two forms are obviously tautomeric, the yellow being less acidic 
 than the red ; and they change over into one another in solution in the 
 usual way, the only remarkable fact being the unusually powerful catalytic 
 influence of alcohol. But their stability relations are very difficult to 
 explain. In the solid state the red form goes over into the yellow at 
 124°. Hence we should infer that the yeUow is the stable form at high 
 temperatures. On evaporating the aqueous solution at ordinary temperatures 
 the mixture of the two forms is obtained : at higher temperatures the yellow 
 alone. This again seems to show (though it is quite probable that equilibrium 
 is not maintained throughout the whole process) that the yeUow is more 
 stable when hot. On the other hand, if the salt is precipitated with acid 
 the red form is got at high temperatures and the yellow at low, which is 
 just the reverse of what one would expect. The only explanation which 
 seems possible is this. The form which is precipitated on acidification is 
 produced rapidly, and so is not necessarily the stable modification ; and it 
 is conceivable that the potassiiun salt may also exist in two forms, and that 
 the influence of temperature on their stability is the reverse of its influence 
 on that of the free hydrogen compound ; so that the potassium salt corre- 
 sponding to the red form predominates in the solution when it is hot, and 
 is therefore precipitated from such a solution by acids. But this seems 
 improbable. 
 
 It is still less easy to assign formulae to the two modifications with any 
 degree of probability. The body offers a great variety of tautomeric possi- 
 bilities, since it has two unsymmetrically arranged hydroxyl groups, and 
 since, further, besides the nitroso-oxime tautomerism, 
 
 (HO-C • ■ • C-N:0 -* 0=0 • • • C=]SrOH), 
 
 you may have the ordinary keto-enolic change (CH=C-OH — » CHj-CsO). 
 Almost any of these formulae are equally compatible with the observed 
 facts as to their tautomeric behaviour. It is worth noticing the suggestion 
 
138 Nitroso-compounds 
 
 of Kehrmann * that both forms are monoketones (quinone oximes), but that 
 the red is ortho-quinoid and the yellow para-quinoid : — 
 
 NOH NOH 
 
 Eed CH3^Q=0 . Yellow CH3-|Q-0H_ 
 
 OH O 
 
 We know that red and yellow are colours characteristic of the ortho- and 
 para-quinones respectively, both in the simple benzoquinones and in many 
 of their derivatives. 
 
 1 Ber. 29. 1417 (1899). 
 
CHAPTER VII 
 
 NITRO-COMPOUNDS 
 
 The nitro-compounds are those in which the hydrogen attached to carbon 
 
 is replaced by the group -N'^q. The aromatic nitro-compounds differ greatly 
 
 from the fatty, mainly because they are necessarily tertiary. Many of the 
 more important reactions of the primary and secondary fatty nitro-compounds 
 are those in which the hydrogen atoms attached to the same carbon as the 
 nitro-group take part ; and these, of course, cannot occur with the aromatic 
 derivatives. The two classes will therefore be dealt with separately. 
 
 FATTY NITEO-COMPOUNDS OR NITEOPARAFFINS 
 
 The fatty nitro-derivatives were not discovered till long after the aromatic 
 derivatives had become well known. They were first prepared by V. Meyer 
 in 1872, and shortly afterwards by Kolbe. 
 
 The method of direct nitration, which is always used in the aromatic 
 series, can only be employed with the paraf&ns in a limited number of cases. 
 It is possible in several compounds which contain tertiary hydrogen (such as 
 chloroform and isovaleric acid, (CH3)2CH-CH2-CH2-COOH) to replace this tertiary 
 hydrogen by NOg directly; and Konowalow has shown that direct nitration 
 is also possible in the case of the normal hydrocarbons from hexane upwards. 
 But it is remarkable that, whereas the method universally employed for nitrating 
 the aromatic hydrocarbons is to use concentrated nitric acid in the presence of 
 concentrated or even fuming sulphuric acid and to work at a low temperature 
 (generally below 0°), Konowalow found, by an exhaustive series of experiments, 
 that the most favourable conditions for nitrating the higher paraffins are to use 
 dilute nitric acid, and to heat to a rather high temperature (130-140°). Nitric 
 acid, of course, acts on organic compounds in two ways : it may nitrate them, 
 or it may oxidize them. Both of these reactions are promoted by using 
 a concentrated acid and by raising the temperature ; and the extent to which 
 these conditions can be employed in nitration is limited by the danger of 
 oxidation. It would seem, therefore, that in the case of the aromatic hydro- 
 carbons the influence of temperature on the nitration is less (or that of 
 concentration is greater) than in that of the paraffins ; and hence with the 
 former we can use a low temperature and a very concentrated acid, while 
 with the latter a high temperature is essential, and so to prevent oxidation 
 the acid must be diluted. 
 
 In certain cases, however,' it has been found possible to apply the aromatic 
 method to the paraffins as well, and by using fuming nitric acid to introduce, 
 
 ' Poni, C. 02. ii. 16, 
 
140 Nitro-campounds 
 
 for example, as many as three nitro-groups into isopentane, each group going 
 to a different carbon atom ; and recently Konowalow ' has shown that where 
 the hydrocarbon contains tertiary hydrogen atoms these are most readily 
 replaced when a dilute acid is used, whereas if a stronger acid is employed 
 the primary and secondary nitro-derivatives predominate. 
 
 The introduction of negative groups into the paraffins facilitates the 
 formation of nitro-corapounds. Thus, Konowalow ^ finds that by his method 
 of using a dilute acid and a high temperature the chlorine derivatives of the 
 hydrocarbons (such as butane and pentane) are more easily nitrated than the 
 hydrocarbons themselves. So, too, acetic ester ' can be directly nitrated at 30° 
 by pure nitric acid in presence of acetic anhydride. If the negative groups 
 are sufficiently powerful to produce an ' active methylene ', as in malonic acid, 
 nitration is still easier. Malonic ester ' can be nitrated by concentrated nitric 
 acid at the ordinary temperature, and its amide'' even by somewhat diluted 
 nitric acid without heating. 
 
 The more usual method of preparing the nitroparafSns is to heat the alkyl 
 halides with silver nitrite : — 
 
 CH3I + AgNOs = Agl + CHg-NOa. 
 
 In this reaction the isomeric nitrites are formed as well as the nitroparaffins, 
 in quantities depending on the nature of the alkyl halide employed. With 
 methyl iodide, practically only the nitro-compound is formed, and this is the 
 main product with primary halides containing up to 3 carbon atoms. But 
 secondary halides, and any halides containing more than 4 carbon atoms, 
 give very small yields of the nitro-compounds ; while tertiary halides give 
 almost entirely the nitrite. Mercurous nitrite behaves in the same way.' 
 
 Nitromethane may be prepared more easily by means of a reaction discovered 
 by Kolbe. This consists in treating chloracetic acid with potassium nitrite. 
 We may suppose that nitro-acetic acid is formed first, and that this at once 
 splits off carbon dioxide to form nitromethane : — 
 
 CH„C1 CHa-NOo 
 
 I -* I -* C0„ + CH,-NO,. 
 
 COOH COOH ' ' ' 
 
 A similar reaction is given' by the a-bromo-derivatives of propionic, butyric, 
 and heptoic acids, which, when treated with potassium nitrite, give a 50 per 
 cent, yield of the corresponding primaiy nitroparaffins. But if the a-bromine 
 atom is attached to a tertiary carbon, the secondary nitroparaffin is not produced, 
 but only a small quantity of the pseudo-nitrol : — 
 
 (CH3)2CBr-COONa + 2 NaNOj = (CH3)2C<^q + Na^COa + NaBr. 
 
 The nitroparaffins can be obtained from potassium nitrite still more easily 
 by acting on its aqueous solution with ethyl hydrogen sulphate,^ or better with 
 the universal methylating agent, methyl sulphate.' In both cases a certain 
 
 ' C. 06. ii. 318 2 0. 04. i. 1478. 
 
 ' Bouveault, Wahl, C. 04. ii. 640. ' Ulpiani, C. 03. ii. 848. 
 
 s Ratz, Mon. 25. 55 (C. 04. i. 1552). " Kay, Neogi, Proc. C. S. 23. 246 (1907). 
 
 ' Auger, Bull. Soc. [3] 23. 333 (C. 00. i. 1263). 
 
 " Eay, Neogi, J. C. S. 1006. 1900. » Walden, Ber. 40. 3214 (1907). 
 
Nitroparaffins : Preparation 141 
 
 quantity of the isomeric nitrite is formed at the same time. These instances 
 are worth noticing as being among the very few in which potassium nitrite 
 
 acts as if its formula was K-N^q, its normal reactions corresponding to the 
 structure K-O-NO. 
 
 A singular method of nitrating compounds which, like malonic ester, 
 contain an acidic methylene group, is to act on them with ethyl nitrate and 
 sodium ethylate ; ' for example, with benzyl cyanide : — 
 
 jj^>CH2 + Et-NOa + NaOEt = j^^CNOaNa + 2 EtOH . 
 
 If the product is boiled with soda,^ it splits oif the cyanide group as sodium 
 carbonate and ammonia, and gives phenyl nitromethane, ^-OHg-NOa. If this 
 body, or the sodium salt of the original nitrile, is heated with soda to a high 
 temperature, the nitro-group itself breaks off, and a good yield of stilbene 
 is obtained. This reaction follows naturally from the formula for the salts 
 of the nitroparaffins which will be established later: — 
 
 0-C-NOONa 
 Lt -^ 0-CH-NO-ONa -^ MB=GB.-(j>. 
 
 CN 
 
 It is remarkable that if tetraiodoethylene is treated with nitric acid it is 
 converted into nitro-triiodoethylene, Clg^CINOa.' 
 
 The structure of the nitroparaffins is proved by their yielding primary 
 amines on reduction. This shows that the nitrogen must be joined directly 
 to carbon ; if it was joined through oxygen it would split off on reduction to 
 give an alcohol and a reduction product of nitrous acid, as happens with 
 the isomeric nitrous esters. 
 
 The nitroparaffins are colourless liquids which are almost insoluble in water 
 and distil unchanged. Their boiling-points are much higher than those of 
 the isomeric nitrites, e.g. : — 
 
 Nitromethane ... B.Pt. 101° Methyl nitrite. .. . -12° 
 
 Nitroethane 113° Ethyl nitrite +16° 
 
 Thus, they can easUy be separated by distillation from the nitrites which are 
 formed at the same time. Nitromethane is remarkable ' for having a very high 
 specific inductive capacity, but at the same time a very low ionizing power. 
 
 On reduction the nitroparaffins yield first ^-hydroxylamines and then 
 primary amines. 
 
 All nitroparaffins which still have hydrogen attached to the same carbon 
 as the NO2 are capable of dissolving in alkalies to form salts. Thus all simple 
 primary and secondary nitroparaffins do so, but not the tertiary. Again, 
 the primary compounds, such as nitroethane, CH3-CH2-N02, when their 
 sodium salts are treated with bromine, give mono-bromo-derivatives, such as 
 CHj-CHBr-NOa. This body can also give a sodium salt, and when the salt is 
 
 * W. Wislicenus, Endres, Ber. 35. 1755 (1902); W. Wislicenus, Waldmliller, Ber. 41. 8334 
 (1908). 2 Wislicenus, Wren, Ber. 38. 502 (1905). 
 
 " Biltz, Kedesdy, Ber. 33. 2190 (1900) ; Nef, Ann. 298. 346 (1897). 
 
 * Bruner, Ber. 36. 3297 (1903). 
 
142 Nitro-compounds 
 
 treated with bromine another bromine atom is introduced, giving CH3-CBr2-N02 . 
 This substance is no longer capable of forming a sodium salt, because no 
 hydrogen remains attached to the carbon. These bromine compounds are 
 proved to have the halogen atom joined to carbon by their reaction veith zinc 
 ethyl : for example : — 
 
 CHa-CH-NO^ CHo-CH-NOa 
 
 I ^.Zn(CH3),^ ^1^^ ^ 
 
 In the same way the mono-bromo-derivative of a secondary nitroparaffin, as 
 
 /Br 
 (CH3)2C\jTQ , will not form salts with alkali. 
 
 These facts were discovered by V. Meyer, and he explained them by the 
 very natural assumption that the presence of the strongly negative nitro-group 
 renders the hydrogen attached to the same carbon atom (the a-carbon atom) 
 acidic ; but that this influence does not extend to the hydrogen attached to 
 other carbon atoms in the same molecule : so that if all the a-hydrogen is 
 replaced, the power of forming salts disappears. Subsequent researches have 
 shown, however, that this explanation is not correct. 
 
 In the first place, Nef found that if the solution of the salt of a nitroparaffin 
 is acidified, none of the nitroparaffin is regenerated, unless special precautions 
 are observed. This in itself tends to show that the process does not consist 
 merely in the replacement of the sodium by hydrogen ; for if it did, the 
 reaction should go quite easily. Moreover, the study of the decompositions 
 which occur throws considerable light on the structure of the salt. 
 
 Before Nef 's work V. Meyer had discovered that if the salt of a nitroparafiin 
 was heated with acids, it broke up into an acid and hydroxylamine : — 
 
 CHgCHa-NOa + H2O = GR^G<^-^ + H2NOH. 
 
 Nef finds that if the solution of the salt is added to dilute nitric acid at the 
 ordinary temperature, no nitro-compound is regenerated ; but practically the 
 whole is split up into an aldehyde (or a ketone, if a secondary nitroparaffin 
 is taken) and nitrous oxide. The formation of an aldehyde indicates that in 
 the salt the nitrogen is joined to carbon by a double bond ; and Nef suggests 
 the following explanation. He assumes that the salt is derived from a tautomeric 
 
 form of the nitro-compound and has the structure CH3-CH=N<^^„ . When 
 
 this is treated with acid the sodium is first replaced by hydrogen, and then 
 the product undergoes what he calls intramolecular oxidation : — 
 
 2 CH3-CH=N<^TT = 2 CH3CHO + if^^ , 
 
 the hyponitrous acid breaking up at once into nitrous oxide and water. 
 
 The next advance was due to HoUeman, who found that the colourless 
 
 <I>-CH2N02 
 
 m-nitrophenyl-nitromethane, 1 , gave a yellow sodium salt, and 
 
 JNO2 
 
 that when the solution of this salt was treated with an equivalent of 
 
 hydrochloric acid, it at first remained yellow and had a higher conductivity 
 
Phenyl Nitromethane 143 
 
 than that of the sodium chloride which it contained, indicating the presence 
 of another electrolyte. But on standing the solution gradually lost its colour, 
 and simultaneously the conductivity fell to that of the sodium chloride. This 
 shows that the first effect of adding acid is to produce a coloured acid substance, 
 which in time goes over into the colourless non-dissociated nitro-compound. 
 Hence the salt must be derived not from the nitro-compound but from 
 a tautomer. 
 
 Finally, Hantzsch succeeded in the case of phenyl nitromethane, ^-CHg-NOj, 
 and of its para-brom-derivative in isolating both the tautomeric forms. Phenyl 
 nitromethane dissolves in soda to form a salt. If carbon dioxide is passed 
 into the solution a yellow oil separates out slowly ; but if the solution is 
 treated with a mineral acid, a white crystalline precipitate forms at once. 
 Both these substances have the composition of ^-CHa-NOa ; and Hantzsch 
 has been able to show that the liquid precipitated by carbon dioxide is the 
 true nitro-compound ^-CHg-NOa, while the solid obtained with mineral acids 
 
 is the tautomeric isonitro-derivative 0-CH=N'\qtt • In the case of _p-brom- 
 
 phenyl-nitromethane both the isomers are solid. The true nitro-body, got 
 as before by ntteans of carbon dioxide, melts at 60°, while the isonitro- 
 compound, produced with hydrochloric acid, melts at 89-90°. 
 
 The arguments by which these formulae are established are as follows ; — 
 The solid form differs from the liquid in giving a reddish-brown colour 
 with ferric chloride. This is a well-known test for hydroxyl, applied by 
 W. Wislicenus in the case of the tautomeric formyl-phenyl-acetic esters. 
 
 Again, the solid reacts with phenyl isocyanate, a recognized test for 
 hydroxyl, while the liquid does not. The reaction is indeed abnormal; the 
 isocyanate takes up water from the nitro-compound, and is converted into 
 diphenyl urea: — 
 
 0-NH 
 2 0.N=C=O -1- H2O = >C:0 + CO2 ; 
 
 0NH 
 
 but it is evident that a compound with hydroxyl attached to the nitrogen 
 would lose water more easily than one without it. Further, the solid 
 modification dissolves in alkali at once, while the liquid only does so on 
 prolonged shaking. This is to be expected if the salt so produced is a direct 
 derivative of the hydroxyl form (the solid), whereas the true nitro-compound 
 (the liquid) cannot form a salt until it has undergone an intramolecular change. 
 
 Finally, the solid is a strong acid, which the liquid is not. 
 
 These facts show conclusively that the liquid is the true nitro-compound 
 
 ^•CHj-NOg, and the solid the hydroxylic isomer ^-CH^N^QTr • 
 
 Of the two, the liquid is the stable form, and the solid changes into it 
 on standing, either in the pure state or in solution. The change can be 
 followed by means of the ferric chloride reaction. In the pure state it is 
 complete in a few days in the cold. In solution the velocity depends on 
 the solvent. It is extraordinarily high in acetic acid ; in water it takes about 
 an hour at the ordinary temperature and four hours at 0°. In other solvents 
 
144 Nitro-compounds 
 
 it is slower, the order being : Water (greatest), alcohol, ether, benzene, chloroform 
 (least), the order of the dissociating power and the dielectric constant, as in 
 the case of formyl-phenyl -acetic ester.' 
 
 The nitroparaffins are typical instances of pseudo-acids. In fact it was 
 in this connexion that Hantzsch developed the theory of pseudo-acidity. 
 Pseudo-acids are a particular class of tautomeric substances distinguished by 
 the fact that one form is much more acidic than the other. Hence a pseudo- 
 acid exhibits all the characteristics of tautomeric substances. It changes 
 reversibly into its isomer at the ordinary temperature, and is therefore capable 
 of reacting according to both of the two isomeric formulae, being in fact in 
 the liquid state a mixture of the two forms in equilibrium. 
 
 Pseudo-acids are distinguished from other tautomers by the marked 
 difference in acidity between the two forms, from which it follows that one 
 is much more highly dissociated than the other. This enables us to apply 
 special means of investigation, in particular the reaction with indicators (or 
 any other test for hydrogen ion) and the conductivity methods. 
 
 With the nitroparaffins, as with most pseudo-acids, the non-acidic form is 
 the most stable in the free state. But this form is incapable of forming salts. 
 Hence, if the free compound is treated with a base, the salt formed is derived 
 from the other (acidic) modification : and when the solution of the salt is 
 acidified it is the unstable acidic form which is first produced. 
 
 To return to the behaviour of the nitroparaffins in general. As has been 
 stated, Nef showed that a solution of their salts when treated with acid gave 
 an aldehyde and nitrous oxide ; while V. Meyer found that, on heating with 
 acid, a fatty acid and hydroxylamine were produced. This last reaction was 
 explained by Bamberger,^ who showed that when the salt is treated with acid, 
 not only an aldehyde is produced, but also a hydroxamic acid, by a simple 
 intramolecular rearrangement : — 
 
 E.CH=N<gH->E.C<Ogjj- 
 If the salt is heated with acid, as in V. Meyer's experiment, the hydroxamic 
 acid breaks up further, like any other oxime, giving hydroxylamine and the 
 fatty acid: — 
 
 E-C<^Ojj + H^O = K-C<g^ + H^NOH,. 
 
 Bamberger also finds that in this formation of hydroxamic acid from the 
 isonitro-salt and acid, a transient blue colour is produced, indicating the 
 production of a nitroso-compound. This is probably a nitroso-alcohol, which 
 then changes, in accordance with the general tendency of secondary nitroso- 
 compounds, into the hydroxamic acid : — 
 
 /OH 
 
 -N:( 
 
 \h 
 
 With magnesium or zinc alkyl iodide, the nitroparaffins give secondary 
 hydroxylamines ' : — 
 
 ECHa-NO, + MgAlkl -» ECH^-N^^^ ■ 
 
 > W. WisUoenus,^«)i. 291. 178 (1896). "^ Ber. 35. 45 (1902). ' Bewad, Ber. 40. 3065 (1907). 
 
 ECH=N<gjj -* E.C^N:0 -» KC<2^jj • 
 
Nitroparaffins : Properties 145 
 
 If nitroethane is treated with mercuric chloride, a white salt, 
 
 is formed, which is not explosive. On the other hand, nitromethane gives 
 with mercuric chloride a yellow salt, which explodes with the utmost violence. 
 This remarkable difference in behaviour has been explained by Nef. He 
 analysed the explosive nitromethane salt, and found it to be free from hydrogen. 
 Further investigation showed it to be mercury fulminate. If we accept Nef s 
 formula for the fulminates, which is practically certain, the reaction is easily 
 accounted for. The first product is the mercuric salt of nitromethane; but 
 this loses water by a kind of intramolecular oxidation: — 
 
 H„C=N-0-hg 
 2 II « = C=N-0-hg + h^O (Hg = iHg), 
 
 and this is Nef s formula for the fulminate. The reaction is a strong confirma. 
 tion of Nef s views. It is obvious that no similar change is possible with 
 the nitroethane salt. 
 
 Nitromethane condenses with benzaldehyde in presence of sodium ethylate. 
 The primary product can be isolated, but it readily loses water to give 
 tt)-nitrostyrol : — 
 
 0CHO + CHg-NOg = 0CHCH2-NO2 -^ ^CH^CHNOg. 
 
 OH 
 This reaction is strictly analogous to the aldol condensation: — 
 
 CHg-CHO + CHg-CHO = GJi.,Q,B.CB.^-GRO -> CHoCH^CHCHO. 
 
 OH 
 
 Among the substitution-products of the nitroparaffins the a-nitro-Jcetones 
 are of some theoretical importance.'* Having the nitro-group attached to 
 the next carbon to the earbonyl they are analogous, in the enolic form, to 
 the ortho-nitrophenols, as the grouping -C(N02)=C-0H- occurs in both. They 
 are also analogous in behaviour, giving three series of salts, (1) colourless, 
 (2) yellow, (3) red. This difference of colour is not caused by water of 
 crystallization, and so must be due to a difference of structure. Now the 
 group -CO-CH-NOg- can change tautomerically into an acid form in two 
 different ways, giving either (1) C(0H)=CN02, or (2) C0-Ct=N00H. Both 
 of these forms are colourless in their derivatives, as is shown for (2) by the 
 colourless ester of dimethyl nitro-barbituric acid, which certainly has the 
 group =NO-OAlk ; and for (1) by the colourless nitrophenol esters, such as 
 
 CeHiwX . Hence the colourless salts of the nitro-ketones must correspond 
 
 to one or other of these two aci-forms. 
 
 In the coloured salts some further change must have taken place, and 
 in two ways, one for the yellow and another for the red. This evidently 
 implies the eo-operation of the keto-group, as its presence is necessary to 
 
 ' Holleraan, Eee. Trav. 23. 298 (C. 05. i. 89). ' Hantzsoh, Ber. 40. 1523 (1907). 
 
146 Nitro-compounds 
 
 the colour ; but exactly how this takes place is uncertain. It is possible 
 that the two forms are stereoisomers : — 
 
 -C— i;^ -C— N=0 
 
 syn I 
 
 " -N=0 -U— 
 
 or, more probably, they are structural isomers, one having one of the formulae 
 given above, and the other some such structure as 
 
 -g O 
 
 
 
 \ 
 
 =N-OK 
 
 The reactions of the nitroparaffins with nitrous acid are both of practical 
 and of theoretical importance. They are closely analogous to those of the 
 amines. 
 
 Primary nitro-compounds exchange the two hydrogen atoms for the oxime 
 group, giving nitrolic acids : — 
 
 CH,<5^ - «=N«« = CH3.<gf + H,0. 
 
 This is analogous to the formation of diazo-compounds from primary amines. 
 
 Secondary nitro-compounds exchange the one hydrogen for NO, giving 
 nitroso-nitro-derivatives, or pseudo-nitrols : — 
 
 CH^x p/H + HONO _ CH3\p/N:0 , „ f. 
 
 as secondary amines give nitrosamines. 
 
 Tertiary nitro-compounds, like tertiary amines, having no mobile hydrogen, 
 do not react with nitrous acid at all. 
 
 The behaviour of the Nitrolic Acids is in many respects remarkable. 
 Their structure is sufficiently established by the above method of formation, 
 and by the fact that they can be made by treating the di-brom-substitution- 
 product of a primary nitroparaffin with hydroxylamine : — 
 
 CH,.C<|^^^ + H,NOH ^ CH3-C<5;gJ^ + 2 HBr. 
 
 But their salts show an isomerism not unlike that of the salts of the nitro- 
 ketones. Hantzsch has shown,^ for example, that ethyl nitrolic acid gives rise 
 to three series of salts. It first produces very unstable colourless salts, which 
 
 are probably the true nitrolates, as CHg-C'^-jj^S^. Their instability may 
 
 well be due to their having the metal attached to the very weakly acid 
 oxime-group, and in close proximity to the strongly acidic NOj. They 
 pass over with the greatest ease into the second series, which are bright 
 red, and are known as erythro-salts : these are the usual product of the 
 action of alkalies on the nitrolic acids, if no special precautions are taken. 
 They regenerate the nitrolic acid when treated with acids. On heating, 
 or on exposure to sunlight, these red salts change further into a third 
 
 J Ber. 31. 2854 (1898) ; Hantzsch, Kanaeirski, Ber. 42. 889 (1909); Wieland, ib. 817. 
 
Nitrolic Acids 147 
 
 series, which like the first are colourless, and which were originally termed 
 leiico-salts. It is, however, more in accordance with analogy, as Hantzsch 
 points out, to apply this term to the first series of (unstable) colourless salts, 
 and to call the third series, which obviously are not true nitrolates at all, 
 isonitrolates. These isonitrolates form the most stable series. They cannot 
 be converted back into the red salts, and when treated with mineral acids 
 they do not regenerate the nitrolic acid, but lose nitrous acid and form 
 
 a polymer of the nitrile-oxide p-rr r / \ r. In the same way on heating 
 
 to 120° they break up into potassium nitrite and methyl isocyanate, CHg-N^C^O, 
 a product of the rearrangement of the nitrile-oxide. On reduction they give 
 aldehydes. They can be shown to have the simple molecular weight. Hence 
 
 the formula originally suggested by Hantzsch, CH3-C<^i^^ ^^-rr , is excluded, 
 
 as such a compound would be coloured owing to the nitroso-group. The 
 most probable structure is 
 
 
 KON-0< 
 
 Such a formula would account for their two chief decompositions : the compound 
 would break up by itself along the dotted line, giving potassium nitrite and the 
 nitrile-oxide; while on reduction hydrogen would | add on to the double link, 
 forming the imine CH3-CH=NH and then the aldehyde. The alternative formula, 
 
 CHa-^-N , 
 KON— O 
 suggested by Wieland differs little from this : but the stability of the compound 
 seems to point rather to a 5- than to a 4-ring. 
 
 The constitution of the red salts is uncertain ; but it is evident that both 
 the nitro- and the oxime-group take part in the salt formation ; and we may 
 
 write them CH3-C<(-j^^^ [ K, on the analogy of the salts of the dinitroparaffins, 
 
 ^<zi 
 
 K. 
 
 The Pseudo-nitrols, obtained by the action of nitrous acid on the secondary 
 nitroparafSns, give the usual colour changes of the nitroso-compounds. If 
 a solution of a secondary nitroparafSn is treated with nitrous acid, an intense 
 blue colour is produced, but this soon disappears, while the colourless solid 
 pseudo-nitrol separates out. If the solution is shaken with chloroform, the 
 pseudo-nitrol is extracted by it, and the chloroform forms a deep blue layer 
 at the bottom. This behaviour is strong evidence that the pseudo-nitrols 
 
 contain the nitroso-group, and have the structure EjCxjjq . This structure 
 
 has been disputed, but it has been established beyond doubt by Piloty.' Piloty, 
 as has been mentioned, obtained haloid nitroso-compounds by acting on the 
 oximes with bromine ; e. g. from acetoxime he prepared brom-nitroso-propane, 
 
 '■ Ber. 31. 452 (1898). 
 
148 Nitro- compounds 
 
 (GRr^^G<^-JL. This body cannot have the isomeric formula (CH3)2C=NOBr, 
 
 because the corresponding chlorine compound (CH^JaC^NOCl is known (from 
 the oxime and hjrpochlorous acid), and is a colourless pleasant smelling liquid of 
 high boiling-point ; whereas the brom-nitroso-compound is a bright blue liquid 
 of penetrating smell and low boiling-point. The latter must therefore be a true 
 nitroso-derivative. Now when it is treated with silver nitrite' it is converted 
 into isopropyl pseudo-nitrol, and this reaction can only take place in one way : — 
 
 (CH3),C<N0 " ""^"^^^ = (CH3)XnS^ + AgBr, 
 so that the structure of the pseudo-nitrols is no longer open to doubt. 
 
 These bodies show the characteristic behaviour of nitroso-compounds in 
 solution. At low temperatures the solution is colourless and the dissolved 
 substance bimolecular; at higher temperatures it turns blue, and the solute 
 is now monomolecular. 
 
 The behaviour of the three classes of nitroparaffins with nitrous acid has 
 been made use of by V. Meyer to distinguish primary, secondary, and tertiary 
 alcohols. The alcohol is converted into iodide by treatment with phosphorus 
 and iodine, and the iodide distilled with silver nitrite. This gives a mixture 
 of nitro-compound and nitrous ester ; but the latter does not interfere with 
 the reaction, and need not be removed. The product is shaken with a mixture 
 of potassium nitrite and potash, and sulphuric acid slowly dropped in. If the 
 original alcohol was primary, a red colour is produced (nitrolic salt), which 
 disappears when the solution becomes acid, and comes back again if it is made 
 alkaline. If it was secondary, a blue colour is formed, which disappears on 
 standing, and on shaking with chloroform goes over into the latter. If it was 
 tertiary, no colour appears at all. 
 
 This test can be used with all alcohols not containing more than five carbon 
 atoms in the molecule, and does not require more than 0-3-0-5 gr. iodide. 
 
 POLY-NITEOPAEAFFINS 
 Of these the most important are those in which the nitro-groups are all 
 attached to the same carbon atom. They may be obtained : 
 
 1. By oxidizing the pseudo-nitrols with chromic acid. 
 
 2. By the action of potassium nitrite on the mono-brom-mono-nitro-com- 
 pounds : — r>„ -vrp, 
 
 (CH3)2C<§o, + ^^^2 = (CH3),C<j;^^ -f KBr. 
 
 It is to be noticed that in this case potassium nitrite acts as a nitro-compound, 
 and not as a nitrite. 
 
 3. By the action of concentrated nitric acid on ketones, which are thus 
 broken up to form a dinitro-compound and an acid : — 
 
 /H 
 2 HNO3 + CaHjCO-CaHs = Q^B.^■<^-^^0^ + CH3COOH + H2O. 
 
 This reaction gives only the primary compounds. 
 
 ' Ber. 35. 3098 (1902). 
 
• Poly-nitraparqffim 149 
 
 The introduction of a second nitro-group greatly increases the tendency 
 ■of the body to go over into the aci-form ; and the primary dinitro-eompounds 
 are strong acids, existing in aqueous solution largely in the aci-form 
 
 from which the salts are derived. 
 
 Their behaviour on reduction is remarkable. They easily split off one nitrogen 
 to give an oxime, which on further reduction is converted into an amine ' — 
 
 E-CH(N02)2 -» K-CH:NOH -» ECH2NH2. 
 
 When heated with potash they give ammonia, potassium nitrite, and a fatty 
 acid. It has been argued ° from this looseness of attachment of the nitro- 
 gen, and especially from their behaviour on reduction, that both the nitrogen 
 
 atoms cannot be attached to carbon, and the formula E-CH^^ -J^^ has been 
 
 suggested for them. As against this we have to observe that in the aci-form 
 •of the mono-nitro-compounds the nitrogen is easily split off (giving an aldehyde 
 or ketone), as Nef showed, so that there is nothing surprising in the nitrogen 
 being easily removed when there are two nitro-groups present. Moreover, 
 now that the structure of the pseudo-nitrols is well established, their oxida- 
 tion to dinitro-eompounds makes it practically certain" that in the latter (at 
 any i;ate in the secondary derivatives) both the nitrogen atoms are joined 
 to carbon. 
 
 The salts of the dinitroparaffins have recently been investigated by 
 Hantzsch.* As we have seen, the mono-nitroparaffins give colourless salts 
 
 E-CH=N^Q-|yr (M = any monovalent metal), which are never known to occur 
 
 in more than one form. The primary dinitroparaf&ns E-CH(N02)2 also give 
 colourless salts, which no doubt have the analogous structure 
 
 But these are very unstable, and can only be obtained under special conditions. 
 The ordinary salts are different, and belong to two series, one yellow and the other 
 red (like the salts of the nitro-ketones and the nitrophenols). Which of these is 
 obtained depends on the temperature, solvent, metal and anion, and it is only in 
 a few cases that both forms have been isolated. They attain equilibrium in aqueous 
 solution at once, the yellow form predominating at low temperatures and the 
 red at high. In some cases the yellow can be crystallized from cold propyl 
 alcohol, and the red from hot. They are monomolecular in water, and the 
 conductivity of the solution is not affected by the change of colour ; also the 
 solid salts are anhydrous, so that the difference cannot be due to water of 
 crystallization. Hence they must be isomeric. Dinitro-ethane is shown to be 
 a pseudo-acid by the fact that it takes time to neutralize a base, and that the 
 neutralization is accompanied by a correspondingly slow change of colour. The 
 
 ' Ponzio, C. 02. i. 976. * Ponzio, C. 03. i. 865. 
 
 ' Scholl, C. 02. ii. 937. * Ber. *0. 1533 a907). 
 
 1175 L 
 
150 
 
 Nitro-compounds 
 
 ordinary aci-form, CH3'C<^u^l-j-]^|- , like the esters and salts of the mono- 
 nitro-compounds, even when negatively substituted, as in CH3-0<^-jtq_qtt and 
 ^'^"^NO-OH' "^^^^^ ^® colourless. Also the unstable colourless dinitro-salts 
 obviously occupy this formula. It is clear that the change in structure is due 
 to the tendency of the positive alkali metal to convert the anion of the salt 
 into as negative a form as possible. How much is possible will depend on 
 the nature of the compound : so that salts of different series will be obtained 
 in different cases. As in the previous instances, the structure of the coloured 
 salts is uncertain, but it is evident that both the nitro-groups must take part. 
 Setting aside stereoisomerism, the following structures are possible : — 
 
 0=N- 
 
 EO- 
 
 o=: 
 
 ■^^OM 
 
 O 
 
 N- O 
 
 E.C:^NOM E-C— ^NOM 
 
 In the dinitro-compounds we have been considering, the two nitro-groups are 
 symmetrically arranged with respect to the molecule. Where this is not the 
 case, the number of isomers is doubled. This is realized in the salts of 
 nitrophenyl-nitromethane NOaCcHi-CHg-NOa , where one nitro-group is on 
 the nucleus, and the other on the side chain. In this compound, besides the 
 colourless salt, which is unstable and no doubt has the normal iso-nitro formula 
 NOj-CjH^-CH^NOOM, there are no less than four series of coloured salts. 
 Of these one is yellow and another red ; from the similarity of colour to the 
 salts of the dinitroparaffins, these must have the metal connected to the 
 nitro-group in the side chain. Besides these, it is possible to prepare, usually 
 by heating the solid red salt, a series of violet salts, which go over into a green 
 series. These will obviously have the metal more closely related to the 
 aromatic nitro-group. In the case of the potassium and caesium salts of the 
 para compound, all four isomeric salts have been prepared, and three of them 
 are known in several other cases. The colour does not depend on the degree 
 of hydration, as this varies quite irregularly in the various series. The question 
 of the structure of the nucleus in such cases will be discussed later, in dealing 
 with the nitrophenols ; but it may be noticed that the ortho and meta com- 
 pounds have much less tendency to form these isomeric salts than the para, 
 and the meta least. 
 
 Of the trinitroparaffins the only one of any importance is nitroform 
 CH(N02)3. It is obtained by a rather peculiar reaction from fulminuric acid, 
 the isoamide of nitro-cyanacetic acid. If sodium fulminurate is treated with 
 a mixture of concentrated nitric and sulphuric acids, it is converted into 
 trinitro-acetonitrUe : — 
 
 ^ /OH ^^^ 
 
 NC-C— C<^g -* NC-C-NOa, 
 
 NO2 NO2 
 
 and this can be made to take up water and form the ammonium salt of 
 
Nitroform 151 
 
 tiinitro-acetic acid, which however (like nitro-acetic acid) is unstable, and 
 immediately loses carbon dioxide to give the ammonium salt of nitroform : — 
 
 = CO, + (N02),=C<gg2^ -> (NO,)FC=N<gj^jj^- 
 
 Nitroform is a colourless liquid, which solidifies below 15°. Its solutions 
 in non-ionizing solvents are also colourless. On the other hand, its aqueous 
 solution and that of its salts are yellow. Further, in aqueous solution nitro- 
 form is a strong acid, almost wholly dissociated: while on the addition of 
 concentrated hydrochloric acid (which must depress the dissociation) the 
 colour becomes paler. This behaviour indicates that nitroform can exist in 
 two tautomeric modifications : one colourless and undissociated, the other 
 yellow and dissociated. The first (pseudo-acid) form is, no doubt, the true 
 nitro-compound HC(N02)3. The coloured acidic form was originally assumed 
 by Hantzsch to be the iso-nitro-compound (N02)2C=NO-OH ; but we now 
 know that bodies of this type are, or at any rate can be, colourless, and it 
 is more in accordance with the recent work on this class of compounds to' 
 suppose that in the coloured form a second nitro-group in some way takes 
 part in the acidic structure. 
 
 The mercuric salt of nitroform' shows a very remarkable behaviour, to 
 which there is scarcely any parallel. It is easily formed by dissolving 
 mercuric oxide in nitroform, and is very soluble in many solvents, especially 
 in ether, being even deliquescent in ether vapour. Unlike the other nitroform 
 salts, which are yellow, the mercuric salt is colourless in the solid state, and 
 so is its solution in ether, in benzene and its homologues, in ethyl acetate, 
 ethyl oxalate, chloroform, carbon tetrachloride, and (still more remarkably)' 
 lactic acid. When dissolved in alcohols, fatty ketones, nitriles, chloracetic 
 or acetic acids, it is pale yellow, while in water and in pyridine it is deep 
 yellow. The depth of colour of equimolecular solutions (N/32) i^i water, 
 methyl alcohol, and ethyl alcohol is in the ratio 240 : 8 : 3. In other- 
 words, the mercuric salt has, roughly speaking, the same colour as free 
 nitroform in the same solvent, while the potassium salt (and the other salts)' 
 is always yellow. Thus we have: — 
 
 CH(N02)3. 
 
 Pure substance Colourless 
 
 Ethereal solution Colourless 
 
 Aqueous solution Yellow 
 
 The mercuric salt is not associated in ethereal solution, for it gives the 
 normal value of the molecular weight. We can only explain its behaviour 
 in the same way as that of nitroform itself. The solid salt and its solutions 
 in non-dissociating solvents must, like nitroform, contain three true nitro- 
 groups, and therefore have the mercury attached to carbon hg-C(N02)3 ; while 
 the solution in ionizing solvents must have a tautomeric form, with the 
 mercury attached in some way to oxygen. The mercury must migrate from 
 
 ' Ley, Kissel, Ber. 32. 1365 (1899) ; Ley, Ber. 38. 973 (1905). 
 l2 
 
 Hg salt. 
 
 K salt. 
 
 Colourless 
 
 Yellow 
 
 Colourless 
 
 (Insoluble) 
 
 Yellow 
 
 Yellow 
 
152 Nitro-compounds 
 
 one position to the other, according to the solvent, just as the hydrogen 
 does with a pseudo-acid, and hence mercuric nitroform should be called 
 a pseudo-salt. 
 
 There is evidence that even in aqueous solution the compound is not 
 wholly in the form of an oxygen salt. In the first place, the colour of the 
 solution is paler than that of an equivalent (N/32) solution of the potassium 
 salt, in the proportion 1 : 3-6. Further, experience with other acids where 
 no tautomeric change is possible, shows that the mercury salt of an acid 
 is dissociated to about the same extent as the free acid itself. Now aci- 
 nitroform (in aqueous solution) is about as strong an acid as trichloracetic. 
 If therefore the mercuric salt in aqueous solution were present wholly as 
 the 0-salt, its dissociation should be about equal to that of mercury trichlora- 
 cetate. It is found, however, to be much less, especially at high concentrations. 
 This confirms the view that the aqueous solution of the mercuric compound 
 contains a considerable percentage of the tautomeric undissociated form. 
 
 If silver nitroform is treated with methyl iodide, it gives the colourless 
 stable trinitro-ethane CH3-C(N02)3. Besides this we should expect, in view 
 of the behaviour of other poly-nitroparaffins, that there would be at least 
 one and possibly two isomeric forms of the alkyl derivatives. There is 
 first the simple aci-form (N02)2C=NO-OAlk, which should be colourless, and 
 secondly the coloured chromo-form of uncertain structure, where the second 
 nitro-group comes into play. Hantzsch and Caldwell^ have shown that, 
 though no isomeric methyl nitroform can be isolated, there is reason to 
 think that one is produced under certain circumstances. Silver nitroform 
 gives an addition-compound with methyl iodide at — 70°, of the composition 
 (N02)20(NO2Ag), 2 CH3I. If this is allowed to warm to the ordinary tempera- 
 ture either alone or in contact with anhydrous solvents, it changes explosively 
 into trinitro-ethane ; but if it is in contact with solvents containing water, it 
 gives a quantitative yield of free nitroform. And yet, if silver nitroform is 
 alkylated in water, alcohol, or ether, even at temperatures at which the addition- 
 product gives only nitroform, it produces only traces of free nitroform, and 
 an almost quantitative yield of trinitro-ethane. There seems to be only one 
 possible explanation of these singular results. In solution (the last case) silver 
 nitroform reacts in the form of ions. These can add the methyl group either 
 to the carbon or to the oxygen. As the body formed by the former reaction 
 (trinitro-ethane) is the more stable, it is the main product. But the solid 
 addition-product reacts only in the undissociated form, by intramolecular 
 change. The silver is attached to oxygen, and the methyl must therefore 
 take its place on the oxygen. This gives the aci-ester, which is very unstable, 
 and hence if water is present it is at once hydrolysed (like the corresponding 
 acl-esters of the nitrophenols) to give free nitroform. If water is absent, it 
 undergoes a further intramolecular change to give trinitro-ethane. 
 
 Hantzsch has recently drawn attention, in a short but extraordinarily 
 ingenious paper,' to the unexpectedly close analogy between the corresponding 
 derivatives of trinitro-methane and triphenyl-methane. Triphenyl-methane, 
 
 • Ber. 39. 2472 (1906). = Bar. 39. 2478 (1906). 
 
Nitrqform 153 
 
 like trinitro-methane (nitroform), is easily prepared j but it is impossible to 
 combine the two triphenyl-methyl groups into hexaphenyl-ethane ^gC-C^g ; or 
 if they can be so combined (for this is of course hotly disputed), at any rate 
 they are so loosely held together that they separate with very unusual ease. 
 To combine two trinitro-methyl groups into hexanitro-ethane (N02)3C-C(N02)3 
 is quite impossible. The iodo-derivative of nitroform (NOaJsC-I will react 
 readily with silver nitrite to give tetranitro-methane C{N02)4, but it will not 
 react with silver nitroform to give hexanitro-ethane at all. On the other hand, 
 as tetranitro-methane is well known and quite stable, so is tetraphenyl-methane. 
 Again, the groups ^gC- and (N02)3C- are remarkably similar in their 
 ' Halochromie ', that is, in their power of passing from a colourless undissociated 
 form to a coloured salt : though the salt formation is less marked in the phenyl 
 derivative, which is a weak base, and requires the presence of strong acids to 
 bring about the change. The resemblance is shown in the following table : — 
 
 Colourless. Coloured. 
 
 Pseudo-acid: CH(N02)3 True acid: C(N02)'3 + H" 
 
 Pseudo-base : HO-C^g [True base : HO' H- C^'g] 
 
 Pseudo-salt : hg-C(N02)3 True salt : C(N02)'3 + hg" 
 
 C1.C03 „ „ cr + c^g- 
 
 The last form may be supposed to occur in the coloured solution in sulphur 
 dioxide, and in the coloured double chlorides. The question as to the structural 
 change which is associated with the production of colour in the triphenyl- 
 methane derivatives is very obscure, and cannot be discussed here : the most 
 probable view with regard to the nitro-compounds (though this also is obscure) 
 has already been mentioned. But the analogy between these two apparently 
 remote groups of compounds is certainly worth noticing. 
 
 Tetranitro-methane C(N02)4 is colourless and insoluble in water. It cannot 
 form salts, as it has no hydrogen to be replaced ; it is much more stable than 
 nitroform, and can be distilled unchanged. 
 
 AEOMATIC NITEO-COMPOUNDS 
 
 The aromatic nitro-compounds cannot be prepared by the action of silver 
 nitrite on the haloid derivatives, as the halogen is too firmly attached to the 
 nucleus. On the other hand, the method of direct nitration can be employed 
 in this case with great effect, and this is of enormous importance, since the 
 nitro-group can be exchanged, by means of the diazo reaction, for almost any 
 other group. 
 
 This method was first used by Mitscherlich in 1834 ; he prepared nitro- 
 benzene by acting on benzene with fuming nitric acid. It was soon discovered 
 that an equally good yield may be obtained with much smaller quantities of 
 nitric acid, if the nitration is carried out in presence of concentrated sulphuric 
 acid, which takes up the water formed and causes the reaction to proceed to 
 an end. The method which is now employed in the great majority of cases 
 of nitration on the large scale, is to use the calculated quantity of nitric acid 
 (an expensive substance from the technical point of view) mixed with excess 
 
154 Nitro-coinpounds 
 
 of concentrated or fuming sulphuric acid, and not to allow the temperature 
 to rise above 0°. The reason for keeping the temperature low is that while 
 this has comparatively little effect on the nitration, it enormously diminishes 
 the oxidizing action of the nitric acid. The most striking example of this 
 is. that it is now found to be quite easy to nitrate aromatic aldehydes by 
 working below 0°, without oxidizing the very sensitive aldehyde group to 
 carboxyl. 
 
 In the laboratory the rather violent sulphuric acid is sometimes replaced 
 by acetic anhydride : or if a milder nitrating agent is required, the compound 
 may be dissolved in acetic acid and nitric acid run in. (In this way acetaldehyde 
 can be nitrated at — 16°.) 
 
 On the other hand, to make the poly-nitro-compounds, such as trinitro- 
 benzene, a mixture of nitric and fuming sulphuric acids is used, and the reaction 
 carried out at a high temperature. 
 
 The introduction of two nitro-groups into the nucleus is easy ; but the 
 introduction of a third is in most cases a matter of some difficulty ; and in 
 no case have more than three nitro-groups been put into the same nucleus. 
 
 The dynamics of the direct nitration of aromatic compounds have been 
 investigated by Martinsen.' He used a solution of the aromatic compound and 
 nitric acid in equivalent quantities (from half to tenth normal) in sulphuric 
 acid. The progress of the reaction was measured by extracting the nitro- 
 compound, or in some cases by determining the amount of free nitric acid 
 remaining. He finds that the reaction is of the second order : that is, the 
 velocity is proportional to the product of the concentrations of the nitric acid 
 and the aromatic compound. Nitrous acid has no catalytic effect ; but the 
 strength of the sulphuric acid has a great influence on the velocity, which 
 reaches a maximum for an acid of the composition H2SO4, 0-7 HgO, and falls 
 off both towards the pure acid and towards the mono-hydrate. The amount 
 of this influence varies with the nature of the substituent already present on 
 the nucleus, being especially high in the case of the carboxylic and sulphonic 
 acids. It is remarkable that where aqueous sulphuric acid is used, the, addition 
 of phosphorus pentoxide to this has no effect on the velocity. 
 
 The influence of substituents on the velocity of nitration is in the 
 following order : — 
 
 NO2 > SO3H > CO2H > CI < CH3 < O-CHs < OC2H5 < OH, 
 
 where chlorine is put in the middle, since its effect is always small, and 
 sometimes positive and sometimes negative: of the others, those to the left 
 have an increasing effect in making the velocity less, and those to the right 
 in making it greater. It is to be noticed that those substituents which 
 diminish the velocity orient the nitro-group in the meta position, those 
 which increase it in the ortho and para. 
 
 The immense increase in the facility of nitration produced by hydroxyl 
 is indicated by the fact that phenol can be nitrated by dilute nitric acid in 
 the cold. 
 
 If a homologue of benzene is heated with dilute nitric acid to a high 
 
 1 Z. Ph. Ch. 50. 385 (1905) ; 59. 605 (1907). 
 
Aromatic Nitro-compounds 155 
 
 temperature, the nitro-group goes only into the side chain, toluene, for 
 example, giving phenyl nitromethane, ^-CHaNOj. (The same thing occurs 
 to some extent when it is nitrated in acetic acid solution.') These are the 
 conditions which are found to be most suitable for the direct nitration of the 
 paraf&ns. 
 
 In comparison with the direct methods of preparing the aromatic nitro- 
 compounds, the indirect are of scarcely any practical importance. But one 
 or two are of sufficient theoretical interest to be worth mentioning. Aniline 
 may be converted into nitrobenzene by oxidation, for example with sodium 
 peroxide or potassium permanganate, and diazobenzene may be converted 
 into it by treatment with potassium nitrite and cuprous oxide. To carry out 
 this last reaction, which is due to Sandmeyer, the aniline is dissolved in 
 excess of acid and diazotized by the equivalent quantity of potassium nitrite. 
 As much more potassium nitrite is then added, and the solution of phenyl 
 diazonium nitrite is poured on to moist freshly prepared cuprous oxide. An 
 addition product of uncertain structure is produced, which breaks up into 
 cuprous oxide, nitrogen, and nitrobenzene : — 
 
 0-N2-ONO -> Na + ^-NOa. 
 
 The mononitro-derivatives of the aromatic hydrocarbons are yellowish or 
 colourless liquids or solids, which are nearly insoluble in water, but are easily 
 soluble in strong nitric acid, from which they are precipitated on dilution. 
 They are volatile in steam, and generally boil without decomposition. The 
 polynitro-derivatives, on the other hand, if they are heated under atmospheric 
 pressure, usually explode with more or less violence. 
 
 They often have strong and characteristic smells. That of nitrobenzene 
 is scarcely to be distinguished from that of benzaldehyde ; ^-nitro-toluene 
 has a smell resembling that of prussic acid ; and trinitro-tertiary-butyl-toluene, 
 
 CH3 
 NO^-ppNO^ 
 
 T^C(CH3)3' 
 
 NO2 
 has an especially powerful odour, and is used in commerce as a scent, under 
 the name of artificial musk. 
 
 The structure of these compounds, as true nitro-compounds with the 
 nitrogen attached to carbon, is shown by the fact that they cannot be saponified 
 as nitrites can (or only in particular cases and with difficulty), and that on 
 reduction they are ultimately converted into primary amines. 
 
 The rediiction of nitro-compounds, which has been examined in great 
 detail, is by no means a simple process. A variety of products is obtained, 
 according to the conditions of the reaction. 
 
 Speaking generally, the reduction proceeds in three ways, according as the 
 solution is alkaline, neutral, or acid. In all cases the ultimate product, if 
 the reducing agent is strong enough, is aniline. (For simplicity, the body 
 
 ' Konowalow, Gurewitsch, C. OS. ii. 818. 
 
156 Nitro-compounds 
 
 reduced is assumed to be nitrobenzene: the presence of substituents does- 
 not seriously affect the results.) 
 
 If it is reduced in alkaline solution (with stannous oxide, alcoholic potash, &c.) 
 the main products are those formed by the condensation of two molecules of 
 nitrobenzene : — 
 
 Azoxy- Azo- Hydrazo-benzene 
 
 The earlier stages of the reduction proceed with great ease ; it is sufficient 
 to boil nitrobenzene with alkali in order to convert it into azoxybenzene ; 
 but for the further reduction of this a stronger reducing agent must be used. 
 Where alcoholic potash or sodium methylate is used as the reducing agent, its 
 action depends on the temperature employed.' At low temperatures the 
 alcohol is oxidized to aldehyde, giving up two atoms of hydrogen. At higher 
 temperatures sodium formate is produced : this implies that the alcohol takes 
 up oxygen, but does not give off hydrogen : — 
 
 CH3OH + 20 = HCOOH + H2O. 
 Hence at high temperatures the hydrogenized compounds, such as the hydrazo- 
 and the amine, are not produced, but only those compounds which are formed 
 from the nitro by loss of oxygen alone, such as the azoxy and the azo. 
 
 Reduction in neutral solution leads, as has already been mentioned, to 
 the /3-hydroxylamine. It can be carried out by means of zinc dust or aluminium 
 amalgam and water, and is promoted (for some reason not yet understood) 
 by the presence of neutral salts, such as calcium or ammonium chloride. An 
 unusual reducing agent which has been used for this purpose is red phosphorus 
 and water.'' 
 
 Eeduction in acid solution leads normally to the formation of aniline ; 
 but under special conditions, particularly on electrolytic reduction, amino- 
 phenol is produced. 
 
 These are the main products of the reduction of aromatic nitro-compounds. 
 It is obvious that such peculiar results call for some explanation ; and in 
 particular the formation of the bodies with two benzene nuclei — the azo- 
 group — in alkaline solution, and practically only in alkaline solution, is most 
 remarkable if it were not so familiar, and is clearly due to the interaction of 
 the earlier reduction-products. The whole question has been investigated 
 in great detail by Haber for the special case of electrolytic reduction, which 
 is of great technical and also theoretical interest : and the conclusions which 
 he has estabUshed there are applicable in the main to the chemical methods 
 of reduction as well. The electrolytic method is peculiarly suitable for 
 investigation. It can be carried on both in acid and in alkaline solution, and 
 by varying the size of the electrodes and the current strength the reducing 
 power can be altered to any required extent. 
 
 The electrolytic reduction of nitro-compounds is always a secondary process ; 
 that is to say, the nitro-compounds are not themselves electrolytes, and the 
 reduction is entirely carried on by means of the hydrogen set free at the 
 
 ' Eotarski, C. 05. ii. 893. ^ Weyl, Ber. 39. 4340 (1906) ; 40. 970 (1907). 
 
Reduction of Nitro-compounds 157 
 
 cathode. We may consider that the electrolysis is a means of preparing 
 on the surface of the cathode a solution of hydrogen, the concentration of 
 which can be varied, as it depends on and is measured by the difference 
 of potential between the cathode and the solution. It is only limited by the 
 fact that at a certain concentration the hydrogen is given off at the cathode 
 in bubbles ; but even this limit can be altered by changing the material of 
 the cathode. 
 
 We are thus able, by using the electrolytic method, to obtain the same 
 variation of results as would be produced chemically by reducing agents of 
 greater or less strength, and this without the complications which are 
 necessarily introduced in the chemical method by changing the nature of 
 the substances present : while the strength of the reducing solution of hydrogen 
 can be determined from the cathode potential. 
 
 This last relation between the reducing power and the cathode potential 
 is discussed in Haber's original paper ' ; the main results which he arrives 
 at with respect to the reactions by which the various reduction-products are 
 formed, and the order of their formation, are as follows. 
 
 The first established product of the reduction of an aromatic nitro-compound, 
 whether in alkaline or in acid solution, is the nitroso-eompound ;-^ 
 
 0NO2 -> ^-NO. 
 There is, however, some reason to think that in certain cases there may 
 be a stage intermediate between these two. Meisenheimer ^ has shown that 
 para (and less easily ortho) dinitro-bodies can be made by careful reduction 
 in alkaline solution to yield red (the ortho blue-violet) salts of the type 
 KOON=< )=NOOK, which, when treated with acid, go over into the nitro- 
 
 nitroso-derivative :■ — 
 
 P „ /NO-OH _^ p TT /NO 
 
 '^6^*\]srooH ~* '^6ti4'\]sro2 ' 
 
 Their quinoid structure is supported by their colour, as ortho quinones are 
 always darker than para. Meta-dinitro-compounds do not give derivatives of 
 this kind. This initial stage is no doubt limited to those nitro-compounds 
 which can easily go over into the quinoid form; but in such cases it seems 
 probable that we begin with the three stages: — 
 
 -NO2 -^ =NOOH -> -NO. 
 
 But in general it may be assumed that the first reduction-product is under 
 all circumstances the nitroso-eompound. This, however, is much more rapidly 
 reduced than the nitro-body, as is shown by the fact that a much lower cathode 
 potential is sufficient for this purpose ; so that it can never attain more than 
 a very low concentration in the liquid. It has not been found possible to 
 isolate it, but the fact of its formation has been established by the following 
 experiment. If nitrobenzene is electrolysed in acid solution in the presence 
 of hydroxylamine and a-naphthylamine, the well-known azo-dye benzene-azo- 
 a-naphthylamine is produced. This dye is the result of coupling diazobenzene 
 
 > Z. Fh. Ch. 32. 198 (1900). 
 
 " Ber. 36. 4174 (1903) ; Meisenheimer, Patzig, Ber. 39. 2526 (1906). 
 
158 Nitro-compounds 
 
 with a-naphthylamine, and it is clear that the diazobenzene must be formed 
 by the action of the hydroxylamine on nitrosobenzene : — 
 
 0NO + H2NOH = ^-NaOH + H2O, 
 ^-N^OH + C10H7NH2 = ^•N=NCioHeNH2 + H2O. 
 
 The nitrosobenzene, however, has only a transient existence, and is rapidly 
 converted into the second reduction-product, /3-phenyl-hydroxylamine, ^-NHOH. 
 This, as has been shown, is a very reactive substance. In presence of acids 
 it is converted into p-aminophenol, which consequently is a product of 
 electrolytic reduction in acid but not in alkaline solution. 
 
 A reaction similar to this production of aminophenol takes place on 
 reduction with a metal such as tin and hydrochloric acid,' when nitrobenzene 
 gives a certain amount of ortho and para chlor-aniline : the hydroxylamine 
 is no doubt converted by the hydrochloric acid into the chloramine, which 
 then undergoes the usual change: — 
 
 C6H,.N<g^ -^ CeH,.N<g -* C,^,<^f^- 
 
 The yield of these bodies is increased by using a small quantity of the metal, 
 which hinders the further reduction of the hydroxylamine, and by working 
 at the boiling-point in presence of excess of hydrochloric acid, which promotes 
 the intramolecular change. 
 
 The phenyl-hydroxylamine is also capable of other reactions. It condenses 
 vrith the nitrosobenzene, the first reduction-product, to form azoxybenzene : — 
 
 0-N=O + tJ[>N-0 = 0-N— N-i^ + H2O. 
 
 ^^ \o/ 
 
 This reaction is of great importance, as it is almost the only source of all 
 the ' bimolecular ' reduction-products of the nitro-compounds, the azoxy-, azo-, 
 and hydrazo-compounds, and hence the amounts of these bodies which are 
 formed under any given circumstances of reduction — electrolytic or chemical — 
 will depend on the velocity with which this reaction takes place. It is found 
 that it proceeds very slowly in acid and much more rapidly in alkaline 
 solution, and this explains why it is mainly in alkaline solution that the azo- 
 group of compounds are obtained. The reason of this difference appears to 
 be" that it is only the ^-phenyl-hydroxylamine itself, 0-NHOH, which is able 
 to undergo this condensation, and not its salts, ^-NHOH-HX. Hence excess 
 of acid delays the reaction by forming the salt ; while it is found that the 
 presence of negative substituents on the nucleus promotes it by increasing 
 the hydrolysis. 
 
 The amount of azoxy-compound formed will obviously depend also on 
 the concentration of the nitrosobenzene in the solution ; and it is therefore 
 diminished by working with a high current density, which reduces the nitroso- 
 benzene to phenyl-hydroxylamine more rapidly, and consequently keeps its 
 concentration down. 
 
 ' Blankama, Mec. Trav. 25. 365 (C. 07. i. 468). 
 ' FlUrscheim, Simon, J. C. S. 1908. 1468. 
 
Reduction of Nitro-compounds 159 
 
 There is pnother reaction, also occurring only in alkaline solution, which 
 leads to the production of the azo-group of compounds ; and this is the 
 conversion of the phenyl-hydroxylamine in presence of alkali into aniline 
 and azoxybenzene : — 
 
 3 0-NHOH = 0.N— N-0 + rf).NH„ + 2 HgO. 
 
 \o/ 
 
 This affords another reason for the predominance of these derivatives in 
 alkaline reduction. 
 
 Finally, the phenyl-hydroxylamine is also reduced electrolytically to aniline ; 
 and in acid solution, where the coupling to azoxybenzene only occurs to a small 
 extent, this is the chief source of the aniline which is the main product of the 
 reduction. 
 
 But the list of reactions is not yet exhausted. We have seen how there 
 are formed nitrosobenzene, phenyl-hydroxylamine, and azoxybenzene : as well 
 as aminophenol and aniline, which do not react any further, and may therefore 
 foe set aside. 
 
 The azoxybenzene is further reduced by the electrolysis, not, as we might 
 expect, to azobenzene, but directly to hydrazobenzene, which, however, never 
 reaches more than a low concentration, as it undergoes in alkaline solution 
 two, and in acid solution three, further changes. In acid solution, and there 
 only, it is converted into semidine and benzidine : — 
 
 0NH-NH-0 -* H^N-O-NH-^ -» -H.^<;Z>-<Z>-^^i- 
 
 The semidine is to some extent oxidized by the nitrosobenzene to phenyl- 
 quinone-diimide ' : — 
 
 and this polymerizes to the dye emeraldine, of uncertain constitution, which 
 has actually been obtained by electrolytic reduction in weakly acid solution 
 under certain conditions. 
 
 The second change undergone by the hydrazobenzene, both in acid and 
 in alkaline solution, is its reduction to aniline. 
 
 The third change, which occurs in alkaline and to a small extent in acid 
 solution, is that it is oxidized by the unreduced nitrobenzene to azobenzene. 
 This is the source of the azobenzene which is formed in quantity in presence 
 of alkali, and in traces in presence of acid : — 
 
 2 0-NO2 + 3 ^-NHNH-^i = 3 H^O + 0N— N-^ + 3 ^-N^N-^. 
 
 These reactions are summarized in the following scheme. The vertical arrows 
 indicate direct reduction by the electrolytic hydrogen, the oblique arrows 
 chemical reactions. Those which occur only in acid solution are indicated by a 
 thick line, those which occur chiefly in acid or alkaline solution are respectively 
 distinguished by thick lines and dots. 
 
 ' Nover, Ber. 40. 288 (1907). 
 
160 
 
 Nitro-coinpounds 
 
 0-N=N-^ 
 
 NO 
 
 0-NHO: 
 
 
 0-NH2 
 
 HaNCeHvCeH^NH, 
 
 0-N=:C6H4=NH 
 
 Emeraldine 
 
 The important points to notice are these. The general scheme of the 
 reduction is the same whether the solution is acid or alkaline ; but the relative 
 quantities of the various products are very different in the two cases. The 
 chief difference arises through the behaviour of the phenyl-hydroxylamine. 
 In acid solution it is rapidly reduced to aniline and very slowly condensed 
 with nitrosobenzene to azoxybenzene. In alkaline solution it is only slowly 
 reduced to aniline, while it is rapidly condensed to azoxybenzene (the starting- 
 point of practically all the Ng compounds) and is also converted into the same 
 azoxybenzene (together with some aniline) by the direct action of the alkali, 
 a reaction which does not occur in acid solution at all. 
 
 These facts explain the general results of the chemical reduction of nitro- 
 benzene in acid alkaline and neutral solution respectively. They show why 
 the main product in acid solution is aniline, and in alkaline solution one or 
 other of the azo-group of bodies : while if phenyl-hydroxylamine is to be 
 obtained, the solution must be neutral, because an acid converts it into amino- 
 phenol, and an alkali promotes its condensation with nitrosobenzene to 
 azoxybenzene and its change by spontaneous oxidation and reduction into 
 azoxybenzene and aniline.' 
 
 The dynamics of the chemical reduction of nitro-compounds has been the 
 subject of a series of investigations by Goldschmidt and his pupUs.'' The 
 method used in all cases was to determine by titration from time to time 
 the fall in the concentration of the reducing agent. Goldschmidt showed 
 that in every case the reduction whose velocity is measured is that of the 
 nitro to nitroso, the reduction of the latter to the hydroxylamine being 
 infinitely rapid, and that of the hydroxylamine to the amine being always 
 much quicker than the first stage, and sometimes too quick to be measured 
 at all. 
 
 ^ For a fuller account of these phenomena see MoUer, EleMroehemische Beduktion der NitroTcorper 
 (Halle, 1903), and Haber's paper quoted above, together witt his other papers in the Zeitschr.f 
 EleJctrocliemie, especially 1897-8, p. 506. 
 
 2 Goldschmidt, Ingebrechtsen, Z. Ph. Ch. 48. 435 (1905); Goldschmidt, Sunde, ib. 56. 1; 
 Goldschmidt, Eckardt, ib. 56. 385 (1906). 
 
Reduction of Nitro-compounds 161 
 
 In the case of reduction by stannous halide in the presence of the haloid 
 acid, he finds that with a constant concentration bf the acid (say hydrochloric 
 acid) the reaction is of the second order, the rate of reduction being proportional 
 to the product of the concentrations of the nitro-compound and the stannous 
 salt. If the concentration of the acid is altered, the velocity constant changes 
 practically in the same proportion. It follows that the real reducing agent is 
 not the stannous chloride, but the complex, SnClaH, or its ion, SnClg. It is 
 probably the ion, since, if the hydrochloric acid is partly replaced by an 
 equivalent of sodium chloride, the velocity is unchanged, which indicates that 
 it is not the hydrogen ion which is important, but the chlorine ion. 
 
 If the stannous chloride and hydrochloric acid are replaced by stannous 
 hromide and hydrobromic acid, the velocity constants for practically all the 
 nitro-compounds examined become about eight times as large. 
 
 As regards the influence of the substituents in the nitrobenzene, it was 
 found that if they are of a reactive character (such as COOH, NHg), the velocity 
 of reduction is greatest in the ortho compounds, and about the same in the 
 meta and para. But a methyl group always diminishes the velocity, and to 
 about the same extent whatever its position. This is remarkable, because 
 the usual effect of a methyl group is to increase the reactivity of an organic 
 compound, as in the velocity of diazotization and of nitration. j3-Nitrophenol 
 shows a peculiar behaviour, being reduced with great slowness by stannous 
 halides, as it is by hydrogen sulphide. This is the more strange, as 
 29-mtrosophenol (quinone oxime) is reduced instantaneously. The nitro- 
 amines (as the nitranilines) occupy a peculiar position. Being weak bases, 
 they are present in the acid solution partly as such, partly as the ions of 
 their salts. It is found that the velocity of their reduction increases abnor- 
 mally fast with increasing concentration of the acid. As the greater concentra- 
 tion of the acid increases the amount present in the ionic form (by diminishing 
 the hydrolysis), it follows that the ion is reduced more rapidly than the free 
 base. This view is confirmed by the fact that in these cases, and in these 
 alone, the addition of sodium chloride, though it hastens the reduction, does 
 not do so to the same extent as an equivalent quantity of hydrochloric 
 acid. For though the chlorine ion increases the amount of the active 
 reducing agent SnCls, the salt formation from the nitraniline is not increased 
 by it, as it is by the hydrogen ion of the hydrochloric acid. The abnormal 
 effect of hydrochloric acid was not observed in the case of the nitrophenols 
 and the nitro-benzoic acids ; from which Goldschmidt concludes that for these 
 bodies the ion and the undissociated molecule have the same reduction velocities. 
 But it is very doubtful whether this conclusion is justified. Such weak acids 
 could hardly be expected to be dissociated to any appreciable extent in the 
 presence of excess of hydrochloric acid. 
 
 In the case of reduction by an alkaline stannous solution Goldschmidt 
 showed that such solutions contain the compound SnOgHNa, and that the 
 active agent is its ion SnOgH'. Allowing for this, the reaction is again 
 bimolecular, as required by the equation: — 
 
 ENO„ -1- SnOoH' = ENO + SnOoH'. 
 
162 Nitro-compounds 
 
 The further stages of the reduction proceed, as in acid solution, with much 
 greater velocity. 
 
 OXIDATION OF AMINES 
 
 This question has been postponed to this point because it is most easily 
 understood in connexion with the reduction of the nitro-compounds. It has 
 been worked out by Bamberger in a large number of cases in very great detail.' 
 We may confine our attention to the primary amines, and disregard the less 
 important oxidation products. The oxidizing agents used were commonly 
 Care's acid or hydrogen peroxide ; in some cases also potassium permanganate 
 and other substances. The precise agent employed has little effect on the 
 course of the action, though it influences the proportion of the various 
 products. 
 
 As has already been mentioned, when a tertiary amine is oxidized, it is 
 converted into an amine-oxide, such as ^-NMog:©. Bamberger assumes that 
 the same reaction occurs with a primary or secondary amine, but that in these 
 cases, as there is also a hydrogen atom attached to the nitrogen, this body 
 changes into the isomeric /3-hydroxylamine : — 
 
 E.N<|->E.N^H^E.N<H^. 
 
 At any rate the first oxidation-product which can be detected, both with fatty 
 and with aromatic amines, is the hydroxylamine derivative.^ The nature of 
 the further oxidation-products depends on whether the carbon attached to the 
 nitrogen is primary, secondary, or tertiary. If it is tertiary, as with tertiary 
 monobutylamine or aniline, the next stage is the formation of the nitroso- 
 compound : — 
 
 C-C-NH, -* C-C-N<5^„ -* C-C- 
 
 c/ - c/ "" c/ 
 
 The same thing may happen when the carbon is primary or secondary ; 
 
 but if so, the nitroso-compound cannot be isolated, going over at once into 
 
 an aldoxime or ketoxime : — 
 
 ^•CHa-NHa -^ 0CH2NHOH -» 0-CH2-N:O -^ 0CH:NOH 
 (CH3)2CHNH2 -^ {CH3)2CH-NHOH -> (CH3)2CH-N:0 -* (CH3)2C:NOH. 
 If the oxidation is carried further, the ketoxime is converted into an isonitro- 
 
 compound, which then changes into the normal nitro-compound : — 
 
 (CH3)2C:NOH -* (CH3),C:N<gjj -^ {GU^)^GB..m^. 
 
 • Ber. 31. 1522 (1898); 32. 342, 1675; 33. 17S1; 34. 2262; 35. 703, 714, 4293, 4299; 
 36. 6S5, mi, 710, 817, 3827, 3831 (1903) ; J. pr. Ch. [2] 68. 473, 480 (1903). The more 
 important references are italicized. 
 
 ' With a purely aromatic secondary amine, such as diphenylamine, it has been shown that 
 a still earlier oxidation product is formed, the tetra-aryl-hydraziue, 
 
 2.^2NH — » ^jN-N^j, 
 but this is easily broken up to give derivatives of the di-aryl-hydroxylamine <p.^OE.. Wieland 
 Gambarjan, Ber. 39. 1499 (1906). 
 
Oocidation of Amines 163 
 
 If the aldoxime (from an amine with primary carbon) is further oxidized, it 
 is partly converted in the same way into an isonitro-, and so into a nitro- 
 compound ; but some of it is converted into the hydroxamic acid. 
 
 The following schemes indicate the succession of reactions, the bodies not 
 actually isolated being put in brackets. 
 
 Oxidation of a primary amine with primary carbon, such as benzylamine 
 or ethylamine: — 
 
 ^-CHg-NHa -» r^.CHa-NHa-j -♦ ^.CH^-NHOH -^ ^-CHg-NiO 
 
 Oxidation of an amine with secondary carbon, as isopropylamine : — 
 (CH3)2CHNH2 -» r{CH3)2CH.NH2-| -> {CH3)2CH-NHOH -* (CH3)2CH-N:0 
 
 -» (CH3)2C:NOH -» ((CH3)2C:NOOH) -> (CH3)2CH-N02. 
 Oxidation of an amine with tertiary carbon, as tertiary butylamine: — 
 {CH3)3C.NH, -* r(CH3)3C.NH2-| -* (CH3)3C.NHOH 
 
 -» (CH3)3C.N:0 -* (CH3)3C.N02. 
 
 The velocity of oxidation of amines by alkaline permanganate solution has 
 been investigated by Vorlander,' though only in a qualitative way. He finds 
 that a tertiary amine reacts more rapidly than a secondary, and a secondary 
 than a primary ; and as regards the character of the radical attached to the 
 nitrogen, a primary carbon atom gives the highest velocity, and a tertiary 
 the lowest. 
 
 In the case of the aromatic amines, where the carbon is necessarily tertiary, 
 the same series of oxidation products is obtained as with tertiary butylamine : — 
 
 -NH2 -* -NHOH -> -NO -^ -NO2. 
 
 But in this case a further series of reactions take place, which are impossible 
 with the fatty compounds, the formation of aminophenols and of the bodies 
 of the azo-group. These are in many respects the same as in the reduction 
 of the aromatic nitro-compounds, the same original substances being present. 
 The aminophenol is also oxidized to quinone, and the semidine (p-amino- 
 diphenylamine), which is formed both from the hydrazobenzene and also by 
 the action of the phenyl-hydroxylamine on the aniline " : — 
 
 CeHs-NHOH + H^N-^ -» CeHX^il,. 
 is oxidized to phenyl-quinone diimide, which is further converted ' by a series 
 
 1 Arm. 345. 251 (1906). " Ber. 31. 1522 (1898). 
 
 = Nover, Ber. 40. 288 (1907). Cf. also Ostrogovioh, Silbermann, 0. 07. i. 1194, 
 
164 Nitro-compounds 
 
 of complicated reactions into emeraldine, aniline black, and other dyes. The 
 fuU scheme is as follows: — 
 
 '^■^^^ >.W<gg^-*CeH,<g 
 
 ^ ^ (^-NHa) ^TT _ /Indophenin 
 
 0-NHOH ^^^HX^h^ -^ HN=O=N-0 -> JMauveine 
 
 ^^■-\ ™ ^ llnduline, &c. 
 
 C&-N— N-d) ^^ „ , ,. 
 
 ^ \Q/ '^ Emeraldine 
 
 V 
 
 AnUine black, &c. 
 
 To return from this digression to the reactions of the aromatic nitro- 
 compounds. 
 
 The nitro-groups may be exchanged indirectly for almost any others by 
 reducing them to NHg and diazotizing. But a direct exchange is difficult, 
 especially in the case of the simple mono-nitro-derivatives. If there are 
 several nitro-groups present it can be e£fected more easily. Thus ortho and 
 para, but not meta, dinitrobenzene can have the NOg replaced by OH (with 
 the formation of potassium nitrite) on boiling with potash ; by ethoxyl (OC2H5) 
 on treatment with alcoholic potash ; and by NH2 or NH^ with ammonia 
 or aniline. This is only one of the many instances which show that the 
 effect of one substituent on another is much greater in the ortho or para than 
 in the meta position. There are other similar cases among the haloid nitro- 
 compounds. If more than two nitro-groups are present, these exchanges 
 can occur even when they are in the meta position, as in symmetrical 
 trinitrobenzene. 
 
 The presence of nitro-groups also exerts a great influence on the behaviour 
 of other substituents in the molecule. For example, the hydrogens attached 
 to the nucleus are more easily oxidized. Nitrobenzene, when heated with 
 solid potash, is converted into nitrophenol, while poly-nitro-compounds can 
 be oxidized to phenols with potassium ferricyanide. In the same way the 
 link between chlorine and the nucleus is very much weakened in the haloid 
 nitro-d erivatives. 
 
 The poly-nitro-derivatives of the aromatic hydrocarbons show in some 
 respects a peculiar behaviour. They have the power, which is shared by 
 some of their substitution-products, such as picric acid, of combining to form 
 crystalline addition-products with aromatic hydrocarbons like benzene, 
 naphthalene, and anthracene (but generally not with the homologues of these 
 hydrocarbons). Thus symmetrical trinitrobenzene gives a compound with 
 benzene of the composition C6H3(N02)8-C8H6 . These are, however, unstable 
 compounds, which readily give up the hydrocarbon again. 
 
 More stable addition-compounds are formed with amines' in particular 
 
 • Hepp, Ann. 215. 344 (1882) ; Eeverdin, Crepieux, Ber. 33. 2507 (1900) ; Sudborough, J. C. S. 
 
Nitronic Adds 165 
 
 by symmetrical trinitrobenzene. These are coloured bodies, which can 
 generally be reerystallized unchanged, and in some cases the amino-group can 
 even be acetylated without breaking up the compound. No satisfactory 
 formula has been proposed for these compounds, but they are probably 
 analogous to the alkaline derivatives described below. 
 
 The behaviour of the nitro-compounds with alkalies is very remarkable. 
 Mono-nitro-derivatives, such as nitrobenzene, are not affected by alkalies in 
 the cold. It is true that ordinary commercial nitrobenzene, if dissolved 
 in alcohol and treated with a drop of potash solution, gives a red colour. 
 But this is due to the presence of dinitrothiophene, 
 
 HC— C-NO, 
 
 II II ' 
 NO2C CH ' 
 
 which is formed from the thiophene in the original benzene. Pure nitrobenzene 
 gives no such colour.' 
 
 But the di- and tri-nitro-derivatives, even when pure, give colours with 
 alkali. Thus s-trinitrobenzene gives a red colour with alkali, and trinitro- 
 mesitylene dissolves in potash, forming a red solution. These phenomena point 
 clearly to the formation of salts; and V. Meyer" supposed that the presence 
 of the nitro-group rendered the hydrogen attached to the nucleus acidic, and 
 that in the red compound produced by treatment with potash this hydrogen 
 was replaced by potassium. 
 
 Lobry de Bruyn^ has, however, shown that this explanation cannot be 
 correct, since metallic potassium does not react with trinitrobenzene. He also- 
 succeeded in isolating one of these compounds. If a solution of trinitrobenzene 
 in methyl alcohol is treated with a molecular proportion of potash in concentrated 
 aqueous solution, red crystals separate out, whose composition is expressed by the 
 formula [C6H3(N02)3CH30K]2-H20. They are explosive, and on treatment 
 with acids regenerate trinitrobenzene. 
 
 The subject has been further investigated by Hantzsch,* who has isolated- 
 several of these compounds and has discussed their constitution. Since they 
 are formed by direct addition, and not by replacement of the hydrogen, he- 
 first assumed that the addition must take place to the nitro-group alone, in 
 this way: — 
 
 (N02)2C6H3N<^ ^(N02)2CeH3.N-OH, -N^OH, -N^OEt • 
 " \0H \ONa \ONa 
 
 Subsequently, however, it was urged by Meisenheimer ° that though this- 
 structure will explain many of the reactions of the compounds it wholly fail& 
 
 1901. 522; Hibbert, Sudborough, ib. 1003. 1334; Sudborough, Picton, ib. 1906. 583; Noelting, 
 Sommerhoff, Ber. 39. 76 (1906) ; Kremann, Mon. 25. 1215 ; C. 05. i. 161, 162. 
 
 ' This reaction can be employed to test benzene for its thiophene, its most common impurity. 
 A few CO. of benzene are warmed in a test tube with a mixture of nitric and sulphuric acids, and 
 the product is poured into water and washed. It is then dissolved in alcohol and a drop of 
 concentrated potash solution added. If the benzene was free from thiophene, the liquid remains 
 colourless; if not, it turns red. 
 
 ^ Ber. 2.1. 3153 (1894). ' Rec. Trav. 14. 89 (1895). 
 
 • Hantzsch, Kissel, Ber. 32. 3137 (1899). = Ann. 323. 219 (1902). 
 
 1175 M 
 
166 Nitro-compounds 
 
 to account for their colour. He therefore suggested that the alkoxy-group was 
 attached not to the nitrogen but to the para carbon atom of the benzene 
 ri°g' e. g- H\/0CH3 
 
 0=N-OK 
 This point he was able to establish in an ingenious way. If we combine 
 on the one hand trinitroanisol with potassium ethylate, and on the other 
 trinitrophenetol with potassium methylate, the products on Hantzsch's theory 
 wUl be different, having the formulae 
 
 C^CHa OC,H, 
 
 NO,-Q-NO, ^^^ NO,-Q-NO, 
 
 0=N<C?,H, 0-N(g^H3 
 
 On Meisenheimer's view the alkoxy-group attaches itself to one of the three 
 carbons which are in the para position to nitro-groups : i. e. either to one of 
 the CH groups or to the C-OCHg. If it should happen to be the latter, then 
 the structure of the product will be the same in both cases, namely 
 
 CH3OVOC2H5 1 
 
 N0,^Q|-N02 ^^ 
 
 0=N-OK 
 Experiment showed that the products of the two reactions were in fact identical, 
 so that it is certain that the alkoxy-group is attached to the para carbon and 
 not to the nitrogen. 
 
 In a more recent paper Hantzsch ^ has pointed out that while these results 
 are conclusive as far as they go, Meisenheimer's formula stUl fails to account 
 for the colour, since we know that both quinols and iso-nitro-salts are colourless. 
 It is evident that the colour depends in some way on the co-operation of 
 a second nitro-group, and therefore the formulae must be written (on the 
 analogy of the other coloured nitro-derivatives) 
 
 H^0CH3 
 
 NO,i . _,, 
 
 K. 
 -NO2' 
 
 
 Hantzsch has isolated several of these bodies, and also prepared derivatives 
 of them. The mother substance of the whole group is a body of the structure 
 
 Hv/OH 
 
 rvNo, 
 
 
 for which we may retain Hantzsch's name of nitronic acid. 
 
 Among the aromatic compounds the formation of derivatives of the nitronic 
 acids does not occur with the mono-nitro bodies, and only with some of the 
 
 ' Hantzsch, Pioton, Bcr. 42, 2119 (1909). 
 
Nitroiiic Adds 167 
 
 di-nitro ; it is most marked with the symmetrical tri-nitro-derivatives, the 
 stablest compounds being those obtained from symmetrical trinitrotoluene, 
 
 CH3 
 
 NO2-A-NO2. 
 
 In this case both the ester salt and the free ester acid can be isolated, and are 
 comparatively stable substances. 
 
 CHaOx ^0 
 
 The ester salt (CH3)(N02)2-CgH2=N-OK is got by treating the nitro-com- 
 pound with potassium methylate in methyl alcohol solution. It forms dark 
 violet crystals which explode violently on heating. Measurements of electrical 
 conductivity show that it is considerably hydrolysed in water. If its solution 
 in dilute methyl alcohol is treated with an equivalent of hydrochloric acid 
 
 CHgOx jy 
 
 at -5°, the free ester acid (CH3)(N02)2-CeH2=N-OH is obtained as a dark red 
 precipitate, also explosive, which is very slightly soluble in water, ether, and 
 benzene, but more so in acetic acid. A determination of its molecular weight 
 by the freezing-point of its acetic acid solution gave results agreeing with 
 the formula. Its electrical conductivity shows it to be a weak acid. It is 
 remarkable that it is not acted on by phosphorus pentachloride, but with 
 acetyl chloride it gives the corresponding acetate. 
 
 Hantzsch was unable to isolate either the free nitronic acids or their salts ; 
 but he obtained evidence of the existence of both classes of compounds in 
 solution. 
 
 The conductivity of a solution of s-trinitrobenzene in soda is found to be 
 rather less than that of the original soda solution (e. g. jui= 113-5 instead of 125). 
 This shows that a small amount of combination has taken place, as indeed 
 is evident from the red colour of the solution. When the solution is treated 
 with acids the colour remains, and the conductivity is found to be seven units 
 higher than that of the sodium chloride which the solution contains, so that 
 it is evident that there is present in the solution a small quantity of the 
 free nitronic acid, H^ /OH 
 
 '.-A-N02^ 
 
 ^N02' 
 
 The nitronic salts can be isolated in the case ot s-trinitrobenzoic acid, 
 
 COOH 
 
 NO2PPNO2, 
 
 NO2 
 a body whose behaviour is interesting in several ways. It is colourless, and 
 if its dilute (N/g2) solution is treated with one equivalent of soda, a colourless 
 solution of the ordinary benzoate is obtained. The further addition of soda 
 produces a deep red coloration, showing that a nitronic salt is formed. If 
 a whole second equivalent of soda is added, the conductivity of the solution 
 
 m2 
 
 N02-A-N02,jj_ 
 
168 Nitrocompounds 
 
 is found to be 220 ; whereas the sum of the conductivities for N/32 solutions 
 of the simple sodium salt and sodium hydrate is 282. This indicates that the 
 second molecule of soda has largely though not wholly combined ; or in other 
 words, that the nitronic salt is formed, but is to a considerable extent hydro- 
 lysed. That the solution does contain a di-sodium salt is shown by its giving 
 precipitates with the solutions of the salts of the heavy metals, such as lead 
 and silver. If the acid is neutralized with baryta instead of soda, the very 
 slightly soluble barium salt 
 
 H\/Oba 
 
 baO-CO-LLj.Q |ba 
 
 is formed as a dark reddish brown precipitate, which is highly explosive. 
 
 Trinitrobenzene can also form an addition-compound with potassium 
 cyanide, which is precipitated on mixing the concentrated solutions of its 
 components. It is a violet crystalline mass, which is very explosive. It 
 must have the formula 
 
 NC\ jy 
 
 (N02)2C6H3=N-OK. 
 
 With mineral acids it gives the insoluble free acid. It is remarkable that this 
 wUl not react either with phosphorus pentachloride or with acetyl chloride. 
 
 The formation of similar addition-products is perhaps indicated by the peculiar 
 behaviour of certain nitro-bodies in formic acid solution.' If the molecular 
 weight is determined cryoscopically in this solvent, it is found that whereas 
 the nitroparaffins, and those nitro-aryls (like trinitromesitylene) which have 
 no unreplaced hydrogen attached to the nucleus, give normal values, other 
 nitro-aryls give abnormally low values, indicating dissociation : and the more 
 so the larger the number of negative groups (NOj, CI, Br, &c.) they contain. 
 Somewhat similar resxilts are obtained in methyl cyanide solution. Now it 
 is scarcely possible that the nitro-aryls can dissociate (for example in the 
 aci-form) when the nitroparafBns do not ; but it is conceivable that they may 
 form dissociating compounds with the solvent. 
 
 A remarkable reaction occurring in this group is that of o-nitrotoluene 
 when heated with potash solution ; the methyl is oxidized by the NOj, with 
 the formation of anthranilic acid (o-amino-benzoic acid) : — 
 
 The haloid derivatives of the nitro-compounds afford a good example ot 
 the influence of nitro-groups on other substituents in the nucleus. They 
 may be made either by nitrating the haloid compounds, which gives the 
 ortho and para, or by chlorinating or brominating the nitro-compounds, which 
 gives the meta derivatives, in accordance with Crum Brown's rule. If the 
 halogen is in the ortho or para position, it is so loosely attached that it may 
 be exchanged for hydroxyl by boiling with alkali, or for NHj or NH^ by 
 heating with ammonia or aniUne. The meta haloid derivatives do not show 
 
 ' Bruni, Berti, Atti B. [5] 9. i. 273, 393; Bruni, Sala, Gaz. 34. ii. 479 (C. 00. i. 1258; 
 ii. 532; 05. i. 673) ; Ciuaa, C. 00. ii. 1051. 
 
Nitronic Adds 169 
 
 this mobility. This difference between the meta derivatives and the others 
 is of course common to many other classes of compounds as well. No 
 satisfactory explanation of it has been offered ; but Lapworth ^ has suggested 
 that the reactivity may be due to the intermediate formation of a compound 
 with a quinoid ring; and its failure in the meta series to the well-known 
 difficulty of formation of meta quinoid compounds. 
 
 If the haloid derivative contains two or more nitro-groups the halogen 
 will react with sodium malonic or sodium acetoacetic ester. There is, however, 
 one remarkable exception to this reactivity in dinitro-chloro-pseudocumene, 
 
 NO2-U-CH3' 
 CH3 
 which will not react even with ammonia or aniline. 
 
 The extreme case of the influence of nitro-groups in increasing the mobility 
 of the halogens is shown in picryl chloride (s-trinitro-chloro-benzene), 
 
 CI 
 
 NO^pj-NOa, 
 
 NO2 
 which exchanges its chlorine for hydroxyl on treatment with water, behaving, 
 in fact, as an acid chloride. 
 
 NITEOPHENOLS 
 
 Phenol can be nitrated, as it can be brominated, with extraordinary ease. 
 To introduce one nitro-group it is enough to act on it with dilute nitric acid in 
 the cold. If several nitro-groups are to be put in, a stronger acid must be 
 used ; and it is then found better to dissolve the phenol in concentrated 
 sulphuric acid, which converts it into phenol sulphonic acid, and to treat 
 this solution with nitric acid. In this way the violence of the reaction is 
 modified, and the nitration facilitated ; the sulphonic acid group is split off 
 again in the process. 
 
 The products of the direct nitration of phenol are (1) ortho and para- 
 nitrophenol, (2) o-o- and o-j)-dinitrophenol, (3) picric acid (s-trinitrophenol). 
 The substituents take up the same positions as in aniline. 
 
 The nitrophenols may also be obtained by various indirect methods, for 
 example by diazotizing the nitranilines and boiling with water. The nitro- 
 phenols may be prepared on the large scale by treating the hydrocarbons with 
 nitric acid in the presence of mercury. Under these circumstances the acid 
 oxidizes the hydrogen to hydroxyl, being partially reduced to nitrous acid." 
 
 They are colourless or yellow crystalline substances, whose salts are deep 
 yellow or red. They are much more acidic than the phenols, and they 
 decompose alkaline carbonates. 
 
 The nitrophenols have recently been the subject of a veiy important 
 
 » Froc. C. S. 14. 159 (1898). ^ Wolflfenstein, Boters, C. 08. i. 1005, 
 
170 Nitro-covipounds 
 
 series of investigations by Hantzsch and his pupils. Their change of colour 
 on forming salts suggests that they are probably pseudo-acids; but until 
 lately there was very little evidence to confirm this, practically the only tests 
 for pseudo-acids which they give being their conductivity in aqueous alcohoP 
 and in pyridine.^ The most satisfactory way of proving a difference of 
 constitution between the nearly colourless hydrogen compounds and their 
 strongly coloured salts would be to prepare isomeric esters, about whose 
 constitution there could be no doubt, which should show the same difference 
 of colour. This had never been done for any group of tautomeric substances, 
 until it was finally effected in the case of the nitrophenols by Hantzsch and 
 Gorke.' The ordinary nitrophenol ethers have long been known ; they are 
 true phenol ethers (as nitroanisol N02-C6H4-OCH3) and are colourless. But 
 besides these it was found that under special conditions isomeric ethers of 
 a deep red colour could be obtained. They are formed by the action of 
 alkyl iodides on the silver salts of the nitrophenols ; but there are great 
 difficulties in preparing them. All the materials must be purified with the 
 utmost care and completely dried. The action is carried on first at 0° and 
 then at the ordinary temperature. These precautions are necessary on account 
 of the instability of the red ethers. They change with the greatest ease into 
 the isomeric colourless ethers, and in the presence of traces of moisture they 
 saponify to the free nitrophenols with extraordinary rapidity. If the proper care 
 is taken, a mixture of the two isomeric ethers is obtained, which is separated 
 by fractional crystallization from ether or chloroform and ligroin, the red 
 ethers being the more soluble. These ethers are dark red in colour, the 
 isomeric ethers being quite colourless when pure. The red ethers, whose 
 melting-points are at least 20° lower than those of the leuco-compounds, 
 are much more soluble in all solvents, giving a colom-ed solution even in 
 water. They change over into the leuco-ethers even on standing in the solid 
 state, the chromo-trinitrophenol ether being completely changed in a few weeks. 
 In indifferent solvents the change is quicker, and in presence of hydrochloric 
 acid gas almost instantaneous. Hence it has not been found possible to obtain 
 the red ethers free from their leuco-isoifters ; the one which was obtained in the 
 highest degree of purity was the trinitrophenol derivative, of which a 1-5 per 
 cent, yield was got, still containing 10 per cent, of the leuco-ether. The red 
 ethers were shown by the raising of the boiling-point in chloroform solution to 
 be monomolecular. 
 
 Their saponification is extraordinarily rapid considering the relative stability 
 of the leuco-ethers, and is quicker the more acidic the nitrophenol. Thus the 
 trinitrophenol (picric acid) derivative is the most rapidly attacked, being rapidly 
 saponified by water alone ; the ortho-nitrophenol compound the slowest, but 
 even this is instantaneously decomposed by sodium carbonate solution, which 
 has no action on the leuco-ether. In this way the percentage of leuco-ether in 
 the compound was determined. 
 
 The existence of these easily saponified isomeric ethers explains why it is 
 
 often found that if the silver salt of a nitrophenol is alkylated without special 
 
 ' Ber. 36. 1001 (1902) ; cf. SiU, Z. Ph. Ch. 51. 589 (1905). 
 
 " Hantzsch, Caldwell, Z. Fh. Ch. 61. 227 (1908). ^ ^^ gg jq^s (1903). 
 
Nitrophenols : Isomeric Ethers 171 
 
 precautions the free nitrophenol is produced (compare the case of nitroform). It 
 is even possible that the chromo-ether is in all cases the sole primary product, 
 and that this then partly isomerizes to the leuco-ethers, and partly saponifies to 
 the nitrophenol. 
 
 The chromo-ether of o-nitrophenol is a dark red oil, which was purified from 
 the isomeric nitroanisol (CHgO-CgH^-NOa) by freezing out, though even then it 
 contained 20 per cent, of the latter. It does not change into the leuco-compound 
 on standing if it is quite dry. The s-trinitrophenol ether (which may be called 
 the true picric ester) loses its colour in ethereal solution in three to four weekS) 
 or, if hydrochloric acid gas is passed into the solution, in three minutes. In the 
 case of para-nitrophenol, the chromo-ether was shown to exist, as the product 
 was coloured and only lost its colour gradually on addition of acid, which would 
 destroy the colour of the salt at once. But it could not be isolated. 
 
 The existence of a chromo-ether of meta-nitrophenol could only be inferred 
 from the fact that in the alkylation free m-nitrophenol was produced. 
 
 Now there is no doubt that the leuco-ethers have the alkyl attached to the 
 phenol oxygen, as in CH30-CgH4'N02. The chromo-ethers, having the same 
 molecular weight, must have an isomeric structure, and must have the alkyl 
 attached to the nitro-group. Hantzsch originally suggested for them the 
 
 obvious quinoid formula 0=;<^=N^^ . But later work has shown that 
 
 it is probable that the quinoid group alone is not sufficient to account for 
 the colour, and we must therefore suppose in this case (as in those of the 
 nitro-ketones and the dinitro-parafSns) that the second substituent (here the 
 phenol hydroxyl) also comes into play. We are thus led to accept, provisionally 
 at any rate, a formula analogous to the old peroxide formula for quinone 
 (though without necessarily accepting that formula for quinone itself) : — 
 
 -O 
 
 
 
 
 The discovery of these chromo-nitrophenol ethers is of great importance 
 in its bearing on the question of the relation between constitution and colour.' 
 Without discussing this question in detaU, there are two points in it to be 
 noticed. In the first place, it was held formerly, especially by Ostwald and 
 his school, that ionization is sufficient of itself to account for a change of 
 colour ; and that the ions have or may have a colour different from that 
 of the undissociated molecule which produces them. But for several years 
 past there has been an accumulation of instances in which it has been proved 
 that the differently coloured ions had also a different constitution. Thus 
 the great majority of indicators have been shown either actually to be pseudo- 
 acids or pseudo-bases, or at any rate to possess structures which are characteristic 
 of these classes of bodies. Hantzsch, to whom a great part of the evidence 
 is due, has been led to the conclusion that all changes of colour are caused 
 by changes of constitution, and that ionization is altogether without influence 
 on colour. There is some reason to think that the truth lies between these 
 
 ' Hantzsch, Ber. 39. 1084 (1906). 
 
172 Nitro-compounds 
 
 two extremes, though much more on the side of Hantzsch ; and that ioniza- 
 tion may affect the depth of colour to some extent, but not its tint, so that 
 from the qualitative point of view (with which we are mainly concerned) 
 Hantzsch's doctrine may be accepted. Whenever we get a distinct change of 
 tint, and especially whenever we get the change of a colourless into a coloured 
 body or vice versa, we are justified in assuming that a change of constitution 
 has occurred. 
 
 The second question concerns the structures which are required to produce 
 colour. Witt's theory of chromophor and auxochrome (in its original form,^ 
 and not as subsequently modified to apply to dyes) is an excellent empirical 
 guide to this problem. According to this theory there are certain groups, 
 for example the azo-group, the quinone ring, and the nitro-group, which 
 tend to be coloured, and are known as chromophors ; and certain other groups, 
 the auxochromes, such as hydroxyl, methoxyl, NHj, &c. , which are not in 
 themselves coloured, but which, when they occur in the same molecule as 
 a chromophor, increase its colour to a greater or less extent. 
 
 A controversy has been going on lately between Hantzsch and Kauffmann 
 as to the relation between colour and constitution, which appears at first 
 sight to be a dispute as to the validity of Witt's theory, but is really concerned 
 only with the interpretation of the facts. To take a concrete instance : nitro- 
 benzene when pure is colourless; if a hydroxyl is introduced (giving nitro- 
 phenol) it becomes feebly coloured ; if the hydrogen is replaced by metal 
 it becomes strongly coloured. Witt would say that the nitro-group is 
 a chromophor, the hydroxyl a weak and the MO group a strong auxochrome. 
 This is merely a clear restatement of the facts. The question at issue is why 
 this is so. The older theory scarcely concerned itself with this : it considered 
 that the mere 'passive coexistence' of chromophor and auxochrome was 
 suificient to account for the colour. The modern defenders of the theory, 
 such as Kauffmann, conscious of a certain weakness at this point, speak of 
 a mysterious influence, not to be too closely examined, exerted by the auxo- 
 chrome on the delicate structure of the molecule. As against this view, 
 Hantzsch adopts a clear and definite standpoint. He holds in fact, though 
 not perhaps explicitly, that the colour is an additive property, in this sense, 
 that every group exerts its own influence on the absorption of light inde- 
 pendently of the other groups present. Thus the fact that nitrobenzene 
 is colourless shows that the group "^-NOa has no effect on visible light. The 
 same must be true of the group KO attached to the nucleus, since KO-0 has 
 no colour. Hence the body KO-CgH4-N02 must be colourless also, and as potas- 
 sium nitrophenol is coloured it cannot be represented by this formula. Since 
 a new effect — ^the production of colour — results from the simultaneous presence 
 of KO and NOg in benzene which does not occur with either of them separately, 
 a new structure must have been produced by the interaction of these two 
 groups (the chromophor and the auxochrome) with one another and with the 
 benzene nucleus, as may be represented, for example, by the quinoid formula 
 0=CgH4=NO-OK, or in other ways. The difference and the superiority of 
 
 ' Ber. 9. 522 (1876). 
 
Constitution and Colour 173 
 
 Hantzsch's view lie in the fact that he looks for definite changes of structure 
 as accompanying colour changes. The advantages of this are obvious. There 
 is abundant evidence that change of colour is an indication of change of 
 structure, and that it is one of the most valuable— just because it is one of the 
 most obvious — proofs that we possess of tautomeric change. We may therefore 
 disregard Kauffmann's view, and consider what light Hantzsch's theory throws 
 on the question of the structure of the nitrophenols and their derivatives. 
 
 All ' true (i. e. not aci-) nitrohydrocarbons, whether fatty or aromatic, 
 whether they contain one or several nitro-groups (as trinitrobenzene, nitroform, 
 tetranitromethane, &c.), have been shown to be colourless when pure, though 
 in some cases the purification is a matter of some difliculty. So are also all 
 substituted nitrobenzenes with unalterable substituents, i. e. substituents which 
 cannot go over into tautomeric forms, as OAlk, OAc, with one exception to 
 be discussed later. The only nitro-aromatic compounds which are coloured 
 are : many nitrophenols, some of their ethers, and all of their salts. The exis- 
 tence of the two series of ethers makes it certain that the nitrophenols are 
 
 tautomeric, giving derivatives of two forms, one CgH4<^-|^--. , colourless, and 
 
 the other, provisionally assigned the quinoid formula, Cl8H4\T , coloured. 
 
 Many free nitrophenols are either absolutely or almost absolutely colour- 
 less (p-nitrophenol, 2,4-dinitrophenol, 2,4,6-trinitrophenol) ; they are therefore 
 practically entirely the true nitro-form. Others, like o-nitrophenol, are quite 
 distinctly coloured, but much less so than their salts. This suggests that they 
 are partly in the chromo-form, but mainly in the other. To this there is 
 an apparent objection, for Knorr has shown that, except at its hypothetical 
 transition point, a solid tautomeric substance can only be in one form. But 
 this proof depends on the assumption that the two solid forms constitute 
 two distinct phases ; and it is possible for a solid solution of the two solid 
 forms to exist as a stable substance. Hantzsch assumes (with great probability) 
 that this happens with the two forms of the nitrophenols, as he had previously 
 done for the solid diazonium halides. The proportion of the two forms will 
 depend on the substance and also on the temperature : and it is found that 
 as a fact all nitrophenols in the solid state (as also in solution) get darker 
 on heating. 
 
 The salts of the nitrophenols with the alkalies and alkaline earths, which 
 are always much darker than the free hydrogen compounds, must be practically 
 entirely in the much more acidic chromo-form. These salts, like those of 
 the a-nitro-ketones and the dinitroparaffins, occur in two series,'* one yellow 
 and one red, this difference not depending on the state of hydration. Hence 
 the nitrophenols can give two series of chromo-derivatives, the salts belonging 
 to both series, the ethers, as far as they are yet known, only to one. 
 
 In solution the free nitrophenols are coloured, and hence the two forms 
 are present in equilibrium, the proportions depending largely on the solvent. 
 In this respect the order of solvents is the same as that observed by Wislicenus 
 and Claisen for the keto-enolic equilibrium : ligroin nearly colourless : (nearly 
 
 ' Hantzsch, Ber. 39. 1084 (1906). ' Hantzsch, Ber. 40. 330 (1907). 
 
174 Nitro-covipounds 
 
 all true nitro) — chloroform pale — benzene rather darker — ^alcohol and ether dis- 
 tinctly darker — water darker still. In other words, the chromo-form is favoured 
 by an increase of the dielectric constant, as it is by a rise of temperature. 
 
 In an aqueous solution of a nitrophenol we have the following substances 
 present : — 
 
 (1) Ar<^? ; (2) Ar< I ; (3) Ar< I ; (4) H", 
 
 omitting the ions of the practically undissociated true nitrophenol (1). Of 
 these the coloured substances are (2) and (3), while the conducting substances 
 are (3) and (4). Experiments on an aqueous solution of 2,4-dinitrophenol 
 showed that the dissociation (up to 30 per cent.) and the colour remained 
 proportional, showing that the amount of undissociated coloiu:ed aci-form 
 present was negligible. If a strong acid, such as hydrochloric acid, is added 
 to a solution of a nitrophenol in water, the consequent increase of hydrogen 
 ions will drive back the dissociation of the aci-form, and so in order to restore 
 equilibrium a corresponding quantity of this must go over into the coloiurless 
 form and the colour must diminish. Experiment showed that this was the 
 case. The colour of picric acid solution (at a dilution of 1,145 litres) diminishes 
 to 68 per cent, of its original value on adding half normal hydrochloric acid, 
 and to 20 per cent, if the acid is made ten times normal. Ortho-nitrophenol 
 also becomes paler on addition of acid, but it is still distinctly yeUow even 
 in concentrated hydrochloric acid, where it must be practically undissociated : 
 so that even in this case a perceptible quantity of the aci-form must be present. 
 
 The salts of the nitrophenols are sometimes yellow and sometimes red ; 
 in certain cases salts of an intermediate orange colour are obtained, but these 
 can be separated by careful recrystallization into a yellow and a red component. 
 These latter are probably in most cases only mixtures of the two salts ; but 
 sometimes their colour deepens on warming, and returns to its original shade on 
 cooling. In such instances it is probably a solid solution of one form in the 
 other.' It is often found, where both salts can be obtained from the same 
 alkali and the same nitrophenol, that they differ in the amount of water which 
 they contain ; but the fact that it is sometimes one form and sometimes the 
 other which is hydrated shows that the water is not necessary to either form, 
 but merely determines (in the same way that the alkali metal does) which is 
 the stable modification in that particular case. 
 
 One of these bodies, s-tribromo-meta-dinitro-phenol, 
 
 NO2 Br 
 
 Br-<~i-OH, 
 
 NO2 Br 
 
 gives two isomeric potassium salts, one yellow and one red, of the same 
 
 composition ; they give different solutions of their own respective colours, 
 
 but of the same molecular weight and conductivity. 
 
 1 Korczynski, Ber. 42. 167 (1909), finds that the red salts of the same nitrophenol with different 
 alkalies differ in depth of colour; and hence infers that they are themselves solid solutions of 
 a darker form in a lighter. 
 
Constitution and Colour 175 
 
 These two series of salts must have two different constitutional formulae. 
 The structure MO-C6H4-N02 is obviously excluded as colourless, since the con-e- 
 sponding ether is colourless. It is conceivable that one series might correspond 
 to the peroxide and the other to the diketone structure of quinone : — 
 
 n and [T| 
 
 To this there is the objection that such isomerism would be possible in other 
 quinone derivatives as well, while it would not be possible in fatty compounds. 
 Now this occurrence of two isomeric series of differently coloured salts has 
 not been observed with any quinone derivatives not containing nitrogen, 
 while on the other hand it is exactly repeated in the nitro-ketones and in the 
 dinitroparaffins. This would seem to show that it is of a kind which does 
 not depend on the peculiarities of the quinone ring. This points to stereo- 
 isomerism as the cause, and we already have a parallel case in the syn and 
 anti diazo-sulphonates, of which the former are dark in colour while the 
 latter are light. 
 
 As it is reasonable to assume a direct connexion between the nitro-group 
 and the phenol oxygen, we may take the peroxide formula of quinone as the 
 basis, and write the isomers 
 
 CgH^-O CeH^-O 
 
 a. "~^N-OK; /3. KO^N ' 
 
 II II 
 
 O 
 
 Now the red ethers, which must correspond to the red salts, go over very 
 easily into the colourless phenol ethers ; this suggests that they have the 
 OK group close to the phenol oxygen, and so have the a formula, in which 
 case the yellow salts would have the /3 structure. 
 
 These phenomena of isomerism among the nitrophenol salts occur in all 
 three series, ortho, meta, and para: indeed it is in the meta series that the 
 isomers have been separated. And unless we are to assume that the meta 
 salts have a structure entirely different from that of the ortho and para, we 
 must admit the possibility, which is generally denied, of meta quinones. But 
 the formulation which Hantzsch adopts gives us a way out of this difficulty. 
 There is no doubt that the quinones themselves and many of their simpler deri- 
 vatives, such as the imides and oximes, which are well known in the ortho 
 and para series, have never been prepared, and so far as we can see cannot 
 exist, in the meta series. But for such compounds, and especially for the 
 quinones themselves and their oximes, the diketone formula is far more 
 probable than the peroxide ; and it is quite conceivable that this is why 
 the meta derivatives do not exist in these cases, the grouping 
 
 X 
 
 a= 
 
 =x 
 
 which we should have to assume in them, being impossible. But if Hantzsch 
 
176 Nitro-compounds 
 
 is right in thinking that in the nitrophenol salts we have derivatives of the 
 peroxide formula, it does not in the least follow that the structure 
 
 of the meta derivatives of this type is impossible. If so, it is perhaps a mistake 
 to speak of such a formula as ' quinoid ', but the name has been accepted, and 
 it is a convenience to retain it, as long as the questions of structure are still 
 undecided. 
 
 This isomerism of the salts is not confined to those of the alkali metals. 
 It is observed in the thallium salts of picric acid ' ; and in the mercuric salts 
 very peculiar relations are found." Mercury is of course much less positive 
 than the alkali metals, and its salts are remarkable for their low degree of 
 ionization, frequently not greater than that of the acid from which they are 
 derived. Accordingly we find that the properties, and especially the colour, of 
 
 the true mercuric salts of the nitrophenols, such as (C6H4^q 2) Hg, resemble 
 
 those of the free nitrophenols themselves. They must therefore be mixtures 
 
 /TSro /NO-Ohg 
 
 or solid solutions of the normal and aci-form CfjH4<(Q V^ and C6H4^ I 
 
 But, as Dimroth has ahown, the mercuiy atom in the salts of aromatic 
 compounds has a strong tendency to migrate to the ring ; and in these 
 compounds this actually occurs. If they are boiled with water they split 
 ofi^ one molecule of free nitrophenol, and form deeply coloured anhydrides : — 
 
 ^Hg-C) 
 (CeH,<g02) Hg = CeH,<g2^ + CeH3 NO, 
 
 in which the mercury is attached on one side to the nitrp-group, and on the 
 other directly to the benzene ring. They are exactly analogous, for example, 
 to Dimroth's mercuri-benzoic anhydride 
 
 C6H.<gf>0. 
 
 They are non-electrolytes and give no reactions of the mercuric ion. If they 
 are treated with strong acid the ring breaks between the mercury and the 
 nitro-group, the mercury remaining attached to the nucleus, and a salt, such as 
 
 /NO, 
 CeHg-HgCl, 
 
 \0H 
 
 is formed, which is colourless or only faintly coloured. The same breaking 
 of the ring is produced by alkalies, giving, for example, 
 
 NOOM 
 
 ^HgOH 
 
 ' Eabe, Z. Ph. Ch. 38. 175 (1901). " Hantzsch, Auld, Ber. 39. 1105 (1902). 
 
Constitution and Colour I77 
 
 but these of course, are coloured, being the aci-salts of substituted nitro- 
 phenols The position of the mercury on the ring is by preference ortho 
 or para to the nitro-group, but it can be meta. 
 
 It has been mentioned that there is one exception to the rule that nitro- 
 compounds with unchangeable substituents are colourless. This is nitrohydro- 
 qmnone methyl ether,^ OCH. 
 
 
 
 ■3 
 
 1-NOo 
 
 OCH, 
 
 A body with this formula we should expect to be always colourless; and as 
 a matter of fact in most non-dissociating solvents it is so, but not in others. 
 Hantzsch has made a careful study of its properties in various solvents. In 
 dissociatmg solvents it gives strongly coloured (yellow) solutions, the colour 
 being roughly proportional to the dielectric constant. The molecular weight 
 IS practically normal both in the nearly colourless hexane solution and in the 
 deep yellow solution in methyl alcohol. The colour shows a shght tendency 
 to increase with the concentration, but this may be only experimental error. 
 The yeUow solutions are non-conductors, so that a dissociated form is not 
 produced. The colour seems to be unaffected by temperature. 
 
 It is evident that the colourless form is the true hydroquinone ether, whose 
 formula is given above. In solution this changes into a yellow isomer, the 
 equilibrmm between the two depending on the solvent, and the yellow form 
 being favoured by solvents of high dielectric capacity. But what this yellow 
 form may be it is at present impossible to say. 
 
 Picric acid, symmetrical trinitrophenol, has long been known. It is formed 
 by the action of nitric acid on many organic substances, such as indigo, aniline 
 resin, silk, and leather. It is generaUy prepared by the action of nitric acid 
 on a solution of phenol in concentrated sulphuric acid. It is the oldest artificial 
 dye. The free substance is pale yellow, but its aqueous solution and that 
 of its salts are much more deeply coloured. It is found ^ that its alcoholic 
 solution becomes much paler when it is cooled with liquid air, and that the 
 solid under these conditions is almost white. This indicates a displacement 
 of the equiHbrium between the two forms, and seems to show that the free 
 substance is a solid solution. It is remarkable that its solution in anhydrous 
 ether = is almost colourless, but turns yellow on adding a drop of water, a 
 fact which can be used to detect the presence of water in ether. This addition 
 of a trace of water also greatly increases the solubility in ether." Picric acid 
 can be used to dye animal fibres directly, but the colour is not very fast. 
 It is now practically abandoned as a dye, but is manufactured in hundreds 
 of tons for use as an explosive,^ especially in war, on account of the ease 
 with which it is prepared, and the fact that it is not liable to be exploded 
 by an accidental blow. Its salts explode when struck, but the free acid 
 requires a detonator. The so-called melinite bombs are filled with molten 
 
 1 Hantzsch, Ber. 40. 1556 (1907). Cf. Hantzsch, Staiger, Ber. 41. 1204 (1908) 
 
 ^ Eeiduschka, C. 07. i. 572. ^ Bougault, C. 03. ii. 565. * Cobet, o'o6 i 233 
 
 '- Will, Ber. 37. 294 (1904). 
 
178 Nitro-compounds 
 
 picric acid, and fitted with gun-cotton detonators. In the laboratory picric 
 acid is often used to identify bases, with which it forms well crystallized 
 and sparingly soluble salts, and also many hydrocarbons, with which it forms 
 crystalline addition compounds. A strong solution of sodium picrate may 
 be used as a qualitative test for potassium, giving a precipitate with a potas- 
 sium salt, owing to the fact that sodium picrate is twenty-six times as soluble 
 in cold water as the potassium salt. 
 
 Symmetrical trinitrobenzoic acid, which has already been referred to in 
 connexion with the nitronic acids, affords a striking instance of the inactivity 
 (due to stereo-hindrance) of the derivatives of the di-ortho-substituted benzoic 
 acids. Its chloride, CO-Cl 
 
 NOa-ppNOs,, 
 
 is by far the most stable acid chloride known. It is largely undecomposed 
 even after boUing with water for an hour. This is the more remarkable since 
 in picryl chloride, which differs from trinitrobenzoyl chloride only in not 
 having the CO between the chloriae and the nucleus, the effect of the nitro- 
 groups is to loosen the attachment of the chlorine so much that it is removed 
 by water. The stabUity of the acid chloride is of course due to its protection 
 by the nitro-groups, which prevent the water molecules from coming up to 
 react with it. But it is difficult to see why in picryl chloride the nitro-groups 
 should not exert a similar protective influence. 
 
 NITEO-DIPHENYLAMINES 
 
 The nitro- derivatives of diphenylamine closely resemble the nitrophenols 
 in their colour relations. They can be shown to be pseudo-acids by their 
 conductivity in pyridine, and while they themselves have only a pale colour, 
 their salts are deeply coloured. It has recently been shown by Hantzsch 
 and Opolski' that this analogy extends to their alkyl derivatives as well. 
 Thus hexanitro-diphenylamine, (N02)3CoH2-NH-C6H2(N02)3, and its methyl 
 derivative, (N02)3C6H2-N(CH3)-C6H2(N02)3, are only feebly coloured, while the 
 salts form two series, one red and the other violet. By a method exactly 
 similar to that used for making the chromo-ethers of the nitrophenols 
 {treatment of the silver salt with alkyl iodide at a low temperature in 
 complete absence of moisture) it yields an isomeric methyl ether, forming 
 black crystals which give a violet solution in benzene. This is only produced 
 in small quantity and is very unstable, going over into the normal isomer 
 when heated above its melting-point, and even at the ordinary temperature 
 in many solvents, and being almost instantaneously saponified by acids. 
 As with the nitrophenols, the chromo-ether has the lower melting-point 
 (chromo-ether 140°, leuco-ether 236-7°), but in this case it is less soluble 
 in all solvents than its leuco-isomer. It is to be noticed that in both cases 
 the colour of the chromo-ether is that of the more deeply coloured series 
 of salts. 
 
 ' Ber. 41. 1745 (1908). 
 
CHAPTER VIII 
 
 CARBONIC ACID DERIVATIVES 
 
 /OH 
 Caebonic acid, Q=0 , being a dibasic acid, can give : — 
 \0H 
 
 /NH, 
 
 1. A monamide CtO carbamic acid, which cannot exist in the free state, 
 
 \0H 
 but is known in the form of salts and esters, the latter being the so-called 
 urethanes. 
 
 2. A diamide C=0 carbamide or urea. 
 
 XNHa 
 
 /NH, 
 
 3. An amidine C— NH tautomeric with urea. There is evidence for the 
 
 \0H 
 existence of this form among the derivatives of urea. 
 
 4. A di-imide C!<^jTg; > ot which only a few derivatives are known. 
 
 /NH, 
 
 5. An amidine-amide Q=NH , the amidine of carbamic acid, which is 
 
 Xnh, 
 
 guanidine. 
 
 It can also form (6) an imide, CO NH, which is isoeyanic acid, but this is 
 
 more conveniently dealt with among the cyanogen compounds. 
 
 Those derivatives which contain oxygen can have this oxygen replaced by 
 sulphur ; and the thio-compounds so produced are of considerable importance. 
 
 Carbamic acid, the monamide of carbonic acid, occurs only in the form of 
 salts and esters. 
 
 The ammonium salt is produced by the direct union of carbon dioxide and 
 ammonia :— /NH, 
 
 C<" + 2 NH3 = (^O 
 " \ONH4 
 
 It is best obtained by passing carbon dioxide and ammonia simultaneously into 
 cooled absolute alcohol, when it is deposited as a crystalline powder. It occurs 
 in commercial ammonium carbonate, which is made by subliming a mixture of 
 ammonium sulphate and calcium carbonate ; that is, by the condensation of 
 a mixture of equal volumes of water vapour, carbon dioxide, and ammonia. If 
 they were completely condensed they would form ammonium carbonate, but 
 some of the water escapes. 
 
 A solution of ammonium carbamate when treated with dilute calcium chloride 
 
180 Carbonic Acid Derivatives 
 
 solution gives no precipitate at first ; but on standing, or more rapidly on 
 heating, calcium carbonate is thrown down, owing to the carbamate taking up 
 water. Under these conditions the hydrolysis is complete, as the carbonate is 
 removed from the solution as quickly as it is formed. But if the carbamate is 
 treated with water alone in the absence of the calcium salt, the reaction 
 is found ^ to be reversible, an equilibrium between carbamate and carbonate 
 being established. On treatment with mineral acids it is saponified at once, 
 giving an ammonium salt and carbon dioxide. On the other hand, as the 
 ammonium salt of a carboxylic acid, it breaks up when heated in a sealed tube 
 into its amide — urea — and water : — 
 
 /NH2 /NH„ 
 
 (^O = C=0 " + H2O. 
 
 \ONH4 \NH2 
 
 The acid chloride of carbamic acid, carbamic chloride, is obtained by passing 
 hydrochloric acid gas over heated metallic cyanates : — 
 
 0=C=NH + HCl = 0=C<^j^2, 
 
 or by passing carbonyl chloride over heated ammonium chloride : — 
 
 /CI /NH2 
 
 Ct=0 + NH3 = C=0 + HCl. 
 
 \ci \ci 
 
 It is a colourless liquid which boils at 61-62°, breaking up partially into cyanic 
 and hydrochloric acids, which recombine to form carbamic chloride in the 
 receiver ; but mainly into hydrochloric acid and a polymer of cyanic acid, 
 cyamelide. This latter decomposition occurs fairly soon if the chloride is 
 allowed to stand at the ordinary temperature. As an acid chloride it is violently 
 decomposed by water to give ammonium chloride and carbon dioxide ; with 
 ammonia or amines it forms ureas, and with alcohols, the esters of carbamic 
 acid or urethanes : — 
 
 /NH2 /NH2 
 
 C=0 + C2H5OH = C=0 + HCl. 
 
 \C1 \OC2H6 
 
 If the alcohol is treated with excess of carbamic chloride, this excess reacts with 
 the urethane as with an amine to give an allophanic ester : — 
 
 H2NCOCI + H2NCOOC2H5 = H2NCONHCOOC2H5 + HCl. 
 
 With benzene and aluminium chloride carbamic chloride reacts according to 
 the Friedel and Crafts method to form benzamide : — 
 
 H2NCOCI + CgHe = HgNCOCeHs + HCl. 
 The alkyl-substituted carbamic chlorides are got by passing carbonyl chloride 
 over heated amine hydrochlorides. When they are distilled over lime they 
 behave as carbamic chloride itself does when heated alone, and break up into 
 hydrochloric acid which combines with the lime, and isocyanic esters : — 
 
 CH3NHCOCI = HCl + CHgN^C^O. 
 The phenyl derivatives are obtained in the same way. 
 
 1 Maoleod, Haskins, C. 06. i. 1820. 
 
Urethanes 181 
 
 The esters of carbamic acid or ureflmms are obtained : — 
 
 1. From carbonic or chlorocarbonic ester and ammonia : — 
 
 /OCaHg /NHa 
 
 C^O + NH3 = q=0 + C2H5OH, 
 
 \OC2H5 \OC2H5 
 
 2. From the alcohol and cyanic acid : — 
 
 HN=C=0 + HOCHj = H2NC<2^ jj 
 
 3. By the action of alcohol on urea at a high temperature:— 
 
 /NH2 /NH2 
 
 C^O + HOC2H5 = C=0 + NHg. 
 
 \NH2 \OC2H5 
 
 This is a reversal of the ordinary formation of amides by the action of ammonia 
 on the esters. 
 ^ 4. They may be prepared, as has been shown, by treating carbamic chloride 
 with alcohol. 
 
 By substituting in these reactions amines for ammonia, or isocyanic esters 
 for cyanic acid, alkyl substituted urethanes can be obtained. 
 
 The simple urethanes are crystalline, the alkyl urethanes liquid. They all 
 boil without decomposition. 
 
 The simple urethanes give with potash potassium cyanate : — 
 
 /NH2 
 C=0 + KOH = KNCO + H2O + HOC2H5. 
 
 \OC2Hfl 
 
 When their solution in benzene is treated with sodium they give a quantitative 
 yield of sodium cyanate ^ : — 
 
 Na-NH-COOEt = NaNCO + HOEt. 
 The mono-alkyl urethanes, when treated with nitrous acid, give nitroso- 
 derivatives, such as nitroso-methyl-urethane, 
 
 C=0 ^" , 
 
 \OC2H5 
 
 according to the general reaction for secondary amines and amides. These 
 nitroso-urethanes show a remarkable behaviour on saponification, which can 
 best be explained by supposing that they first yield the very unstable nitroso- 
 primary amine : — 
 
 Q=.0 ^'^ + H2O = C2H5OH + CO2 + CHa-NHNO. 
 
 '■s 
 
 \OO2Hs 
 
 This methyl nitrosamine then breaks up according to the conditions of the 
 experiment to give either an open-chain diazo-compound CHs-NiNOK, or 
 diazomethane CH2N2, or methyl alcohol and nitrogen. 
 
 The mono-alkyl urethanes on treatment with anhydrous nitric acid are 
 nitrated in the NH group, and the products when treated with ammonia split 
 
 ' LeuchB, Geaeriok, Ber. 41. 4171 (1908). 
 1175 N 
 
182 Carbonic Acid Derivatives 
 
 up, regenerating urethanes, and forming nitramines, which, being comparatively 
 stable bodies, are not further decomposed : — 
 
 C^O^^^ + NH3 = C=0 + hnC^^ 
 
 \0C,H, \OC2H5 ^"^ 
 
 The simple urethanes give sodium derivatives in which the amide hydrogen 
 is replaced, such as NHNa-COOEt. These bodies' react with esters as amines, 
 giving amides, for example, with phenyl acetic ester : — 
 
 ^CHa-COOEt + NHNa-COOEt = ^CHj-CONH-CO-OEt + NaOEt. 
 
 The tendency to this reaction is so strong that it even takes place with haloid 
 esters, such as chloracetic ester, which forms CHaCl-CO-NH-CO-OEt, instead 
 of sodium chloride being eliminated, as we should expect. 
 
 Ordinary urethane, when treated with bromine^ in alkaline solution, gives 
 an amido-bromide C2H50-CO-NBr2, an unstable oil. In neutral solution 
 bromine has scarcely any action except in the presence of a carrier, such as 
 iron wire ; but in that case the bromine attacks the ethyl group, splitting off 
 bromal CBrgCHO. 
 
 The sulphur derivatives of carbamic acid can occur in various forms : thus 
 three mono-thio-derivatives are possible : — 
 
 /NH2 ^NH /NH2 
 
 C=S , C-SH , C-SH . 
 
 \0H \0H \0 
 
 The free acids are tautomeric, differing only in the position of the hydrogen 
 atoms and the double bond. They belong to the same highly tautomeric class 
 of bodies as the thio-acids and the thio-amides. Free thiocarbamic acid is 
 unstable, but its salts are known. The ammonium salt is formed by the 
 combination of carbon oxysulphide and ammonia. If, however, the hydrogen 
 atoms are replaced by alkyl groups, these groups are incapable of tautomerizing. 
 It is obvious that the monoalkyl derivatives can be of three kinds, according as 
 the alkyl is attached to nitrogen, sulphur, or oxygen. Of these three classes 
 two are known, the sulphur and oxygen esters. The nitrogen derivatives would 
 still be acids, and hence would share the instability of thiocarbamic acid itself. 
 
 The 0-esters are obtained from the dithio-carbonic esters by the action of 
 ammonia (normal formation of amides) : — 
 
 /SEt /NH2 
 
 C=S + NH3 = EtSH + (^S . 
 
 \0Et \OEt 
 
 Their constitution is shown by their breaking up when treated with aqueous 
 alkali to give alcohol and a thiocyanate. 
 
 The S-esters are got by the partial hydrolysis of thiocyanic esters with 
 alcoholic hydrochloric acid : — 
 
 EtS-C=N + H„0 = Et-SC-NH„. 
 2 II * 
 
 O 
 
 ' Diels, Heinteel, Ber. 38. 297 (1906). ' Diels, Ochs, Ber. 40. 4571 (1907). 
 
Thiocarbamic Adds 183 
 
 Their constitution is proved by their method of formation, and by the fact that 
 on saponification they yield mercaptans. 
 
 /NH2 ^NH 
 
 Dithio-carbamic acid CfcS or C-SH is obtained in the form of its 
 \SH \SH 
 
 ammonium salt by the combination of ammonia and carbon disulphide ; and this 
 salt on treatment with dilute sulphuric acid gives the free acid as an unstable 
 reddish oil. 
 
 A dialkyl derivative of this ammonium salt is formed by the action of 
 carbon disulphide on fatty amines : — 
 
 Q /NHEt 
 
 C<S + 2NH2Et = (ts 
 
 * NS-NHgEt 
 
 This reaction is generally said to distinguish the fatty amines from the aromatic, 
 which under these conditions give a diaryl thiourea (e. g. aniline gives thio- 
 «arbanilide) ; but it has recently been shown that some aromatic amines will 
 form dithiocarbamates in presence of ammonia.^ 
 
 Carbamic acid is the mono-carboxylic acid of ammonia, NHa-COOH. The 
 corresponding dicarboxylic acid, NH(C00H)2, the so-called immo-dicarboxylic 
 cKid, is also known in the form of certain derivatives.'' Its neutral esters are 
 obtained by the action of chloro-carbonic ester on the sodium urethanes : — 
 
 HN<S.OEt + Cl-COOEt = HN<gg:gg +NaCl. 
 
 This is a crystalline body, M.Pt. 50°, B.Pt. 215°. It is to be observed that this 
 •ester is more acidic than urethane. If it is treated with sodium urethane, the 
 sodium passes over to the nitrogen of the dicarboxylic ester.' 
 
 If this sodium imino-dicarboxylic ester is again treated with chloro-carbonic 
 ester, the last hydrogen on the nitrogen is replaced by carboxethyl, and nitrogen 
 tricarboxylic ester ' is produced. 
 
 /COOEt /CO-OEt 
 
 N-Na + ClCOOEt = N-COOEt + NaCl. 
 
 \C00Et \CO-OEt 
 
 This body, a liquid boiling at 146° under 12 mm. pressure, corresponds to a 
 triamide like triacetamide N(CO-CH3)3, and accordingly is very easily saponified 
 to alcohol, carbon dioxide, and imino-dicarboxylic ester. If it is treated with 
 anhydrous ferric chloride, it splits up in an unusual manner, to give carbox- 
 ethyl isocyanate and ethyl carbonate ° : — 
 
 /COOEt /OEt 
 
 N-COOEt = 0=C=NCOOEt -i- O^O . 
 
 \CO-OEt \OEt 
 
 /CO.'N'H 
 
 Themonamide of imino-dicarboxylic acid is allophanic acid HN\p(-. .-.tt^, 
 
 \COOH 
 
 or 
 
 C=0 (urea carboxylic acid), and its diamide is biuret HN<'p^ -ntttS 
 
 \NHC00H OUJNii2 
 
 which is one of the products of the action of heat on urea. 
 
 ' LosanitBch, Ber. 40. 2970 (1907). " Kraft, Ber. 23. 2786 (1890). 
 
 ' Diels, Ber. 38. 736 (1903). * Diels, Nawiasky, Ber. 37. 3672 (1904). 
 
 « Diels, Wolf, Ber. 39. 686 (1906). 
 
 n2 
 
184 Carbonic Acid Derivatives 
 
 UEEA 
 
 This body is memorable in the history of chemistry from the fact that its 
 formation from ammonium cyanate by Wohler in 1828 went far to break down 
 the barrier which had been set up between organic and inorganic chemistry. 
 It is of course incorrect to describe this as a synthesis of an organic substance 
 from inorganic materials ; but it was the first case in which a body admittedly 
 of an organic nature was produced from a material which was regarded as 
 belonging to the class of inorganic compounds. It is not a little remarkable 
 that John Davy had obtained urea several years before this by the action of 
 carbonyl chloride on ammonia, but had not recognized it. 
 
 Wohler's method, of heating ammonium cyanate, is still used for the prepara- 
 tion of urea. The reaction occurs on evaporating the aqueous solution ; but it 
 is reversible, and hence the conversion is never complete. It has been in- 
 vestigated in great detail by Walker and Hambly,' who find that the change of 
 ammonium cyanate into urea is much more rapid than the reverse change, 
 equilibrium being reached in a decinormal solution when there is about 95 per 
 cent, of urea and 5 per cent, of cyanate present. This equilibrium is little affected 
 by the temperature. Their method of titration was to treat a known volume of th6 
 solution with an excess of silver nitrate, whereby the cyanate was precipitated 
 as the almost insoluble salt, while the urea did not react. It was found that the 
 formation of urea from the cyanate was not, as might have been expected, mono- 
 molecular, but bimolecular, the velocity being proportional to the square of the 
 concentration of the ammonium cyanate. Further examination showed that 
 while the addition of indifferent salts such as potassium sulphate, or of ammonia, 
 had very little influence on it, it was greatly increased by adding either a cyanate 
 or an ammonium salt. This proves that the two molecular species, to the 
 product of whose concentration the velocity is proportional, are NH4' and CNO'. 
 The salt being highly dissociated at the dilutions employed, the concentration 
 of each of these ions will be nearly equal to that of the salt, when no other 
 substance is present, and therefore the variation in the velocity with the dilution 
 will be that required for a bimolecular reaction. The addition of a salt like 
 potassium sulphate will not affect the concentration of these ions, and will 
 therefore not affect the velocity either. On the other hand, if we add to the 
 solution either a cyanate or an ammonium salt, we shall increase the concentra- 
 tion of one or other of the two ions, and hence also the velocity of the reaction. 
 This explains also why it is that ammonia does not increase the velocity, while 
 an ammonium salt does so. The addition of ammonia does not appreciably 
 affect the concentration of the NH4 ions, because, as a very weak base, it is only 
 ionized to a minute extent, especially in the presence of an ammonium salt. 
 
 From these experiments Walker infers that it is the ions of the ammonium 
 cyanate which change into urea, and not the undissociated cyanate. But this 
 conclusion, though it was accepted for many years, is not really valid. If we 
 assume that the salt obeys Ostwald's law, we have in solution the equilibrium : — 
 
 ^ ^ ^NHjCNO = ^SUi ^ ^CNO" 
 ' J. C. S. 1805. 746; Walker, Appleyard, ib. 1896. 193; Walker, Kay, ib. 1897. 489. 
 
Urea : Formation 185 
 
 Now what Walker showed is that the rate of formation of urea is proportional 
 to the right-hand side of this equation. It follows that it must also be pro- 
 portional to the left-hand side : that is, to the concentration of the undissociated 
 portion of the ammonium cyanate. But this is what we should find if it was 
 only the undissociated portion, and not the ions, which formed urea. Moreover, 
 the addition of excess of either ion, i. e. of an ammonium salt or a cyanate, will 
 alter the concentration of the undissociated salt in the same proportion as it 
 alters the product of the concentration of the two ions. Thus the results 
 obtained by Walker are equally compatible either with the view that it is the 
 ions which undergo the change, or with the view that it is only the undissociated 
 part of the salt which does so. Subsequent investigations have shown that it is 
 on the whole more probable that the reaction is due to the undissociated portion 
 of the salt. 
 
 The substitution of alkyl ammonium cyanates^ for ammonium cyanate 
 affects the velocity constants of the reaction, generally diminishing them ; but 
 no regularities could be observed. 
 
 Another remarkable method " of making urea from inorganic constituents is 
 by heating a solution of carbon monoxide in ammoniacal cuprous chloride, when 
 metallic copper separates and urea is formed : — 
 
 CO -f 2 NHg + CU2CI2 = H2N-CO-NH2 -t- 2 HCl + 2 Cu. 
 
 There are two other methods of formation which are of interest as showing 
 the constitution of urea, firstly J. Davy's synthesis from phosgene and ammonia, 
 which is best carried out by acting with the phosgene on sodium phenate, so as 
 to give phenyl carbonate, and then warming the phenyl carbonate in a stream 
 
 of ammonia : — 
 
 /CI NaO0 /O0 
 
 C-0 + =2 NaCl + C=0 : 
 
 \C1 NaO^ \O0 
 
 /0(P /NH2 
 
 C=0 + 2 NH, = 2 HO0 + Q-0 : 
 
 \0<p " \NH2 
 
 and secondly, the partial saponification of cyanamide : — 
 
 Urea forms long colourless prisms, melting at 132-133° ; it is easily soluble in 
 alcohol, and very easily in water. It is a fairly strong monacid base, and forms 
 stable salts with acids, one of which, the nitrate, is often used for isolating it, 
 as it is only slightly soluble in water, and nearly insoluble in strong nitric acid. 
 
 Urea is the most important form in which nitrogen is excreted from the 
 animal organism, and it is also found ^ in various animal fluids, such as blood 
 and lymph, though generally only in small quantities. It has also been found 
 in certain plants.* 
 
 As an amide, urea is saponified by acids and alkalies, forming carbonic acid 
 
 ' Walker, Appleyard, J. C. S. 1896. 193. ^ Jouve, C. 99. i. 422. 
 
 ' Bamberger, Landsiedl, C. 03. ii. 66. 
 
 * For theories as to its origin in the animal body cf. Eppinger, 0. 05. ii. 151. 
 
186 Carbonic Add Derivatives 
 
 and ammonia. The velocity of this change has been examined with somewhat 
 unexpected results by Fawsitt.' He determined the extent to which the reaction 
 had proceeded by titrating the liquid with acid and methyl orange, which 
 measured the ammonia produced. According to the equations usually gfiven :— 
 CO(NH2)2 + 2 HCl + HjO = 2 NH^Cl + CO^, 
 CO(NH2)2 + 2NaOH = NajCO.. + 2NH3, 
 we should expect the reaction in both cases to be of the third order, where the 
 initial concentration of acid or alkali was that required by this equation. But 
 experiment showed that in both cases it was of the first order. Where acid 
 was used, it was found that its concentration had little influence on the velocity, 
 unless it was great, when the velocity was somewhat diminished. With alkalies^ 
 an increase in concentration somewhat increased the velocity. 
 
 The results obtained with acid can be explained as follows. We know from 
 Walker and Hambly's work that urea changes in aqueous solution into ammonium 
 cyanate, though this change stops when about 5 per cent, of the urea has gone 
 over. Now ammonium cyanate was found to be very rapidly saponified by acid 
 to ammonia and carbon dioxide. If, therefore, we suppose that the direct 
 saponification of the urea does not occur at all, or only slowly, we shall have 
 the urea changing over into cyanate, and this decomposing almost as fast as it 
 is formed. In this case, as long as the concentration of acid is sufBcient to 
 prevent the accumulation of cyanate, and thereby to stop the back action, it will 
 not affect the velocity, which wUl be that of the tautomeric change, and will be 
 proportional to the amount of urea present. We shall thus get the values 
 required for a monomolecular reaction, as is found to be the case. The diminu- 
 tion of velocity produced by a large quantity of acid is due to the conversion of 
 part of the urea into its salt, which does not undergo the change into ammonium 
 cyanate, and is thus withdrawn from the reaction. This view is confirmed by 
 the fact that the initial velocity (when there is practically no cyanate present) 
 of disappearance of urea is nearly the same in pure water as when acid is added, 
 and that it has the same temperature coefficient. 
 
 With alkali the case is different. The decomposition of the cyanate by 
 alkali is much slower than by acid, and hence this product accumulates in the 
 solution ; while at the same time, especially if the alkali is strong, a certain 
 amount of direct saponification of the urea takes place. 
 
 The alkyl ureas exhibit the same phenomena, but in this case the formation 
 of cyanate is more rapid, and a larger amount of this accumulates, so that the 
 results are more complicated. 
 
 When urea is strongly heated, it breaks up with the formation of ammonia, 
 carbon dioxide, biuret, and cyanuric acid : — 
 
 2 CO(NH2)2 = NH^CONH-CONH^ + NH3, 
 COCNHj)^ == HN:CO + NH3 : 
 the cyanic acid formed in the latter case polymerizing to cyanuric. 
 If it is heated with potash it forms mainly potassium cyanate. 
 If urea nitrate, CO(NH2)2-HN03, is treated with concentrated sulphuric acid, 
 it gives nitro-urea, NHj-CO-NH-NO^ , an acid stronger than acetic acid (no doubt 
 
 > Z. Ph. Ch. 41. 601 (1902 ■ ; J. C. S. 1904. 1581 ; 1905. 494. 
 
Urea : Properties 187 
 
 having the iso-nitro form), which on reduction yields amino-urea or semi- 
 carbazide NHa-CONH-NHa. 
 
 When treated with sodium nitrite or sodium hypobromite it is oxidized to 
 carbon dioxide and nitrogen. Both of these reactions are used for the quantita- 
 tive estimation of urea, the evolved nitrogen being measured directly. It is to 
 be observed that whereas with hypobromite the nitrogen evolved is that which 
 was contained in the urea : — 
 
 3 NaOBr + CO(NH2)2 = CO2 + N^ + 2 HjO + 3 NaBr, 
 twice this quantity is obtained by the other method, one half coming from the 
 urea and the other half from the nitrous acid : — 
 
 CO(NH2)2 + 2 HONO = CO2 + 2 Nj + 3 H2O. 
 
 Sodium hypobromite or hypochlorite is also able to act on urea in another 
 way,^ in accordance with the Hofmann reaction. Urea is an amide, and can 
 therefore give vnth hypochlorite the corresponding amine and carbon dioxide. 
 The amine formed in this case is hydrazine : — 
 
 H2NC=0 HaNCONa Cl-C-ONa C=0 CO, 
 
 I -* II -* II -> II -* '• 
 
 NH2 GIN HjN-N H2NN + HaN-NH^ 
 
 The amount of hydrazine formed is very small, as it is further oxidized to 
 nitrogen by the excess of hypochlorite. This can to some extent be prevented, 
 and the yield increased, by adding to the solution benzaldehyde, which removes 
 the hydrazine as the comparatively stable benzal-azine 0-CH— N-N— CH-^. 
 
 When treated with chlorine in aqueous solution, urea is converted into 
 a symmetrical dichloro-derivative, C0(NHC1)2. This body is broken up by acids 
 with formation of nitrogen chloride, NCI3, and is converted by ammonia into 
 the so-called diurea or para-urazine. This last reaction probably takes place 
 through the intermediate formation of monochloro-urea ^ : = 
 
 ^NHCl ^NH„ 
 
 CO _» CO : 
 
 ^NHCl ^NHCl 
 
 ^NHCl H2N^ ^NHNH\ 
 
 CO + ^00 = CO CO + 2HC1. 
 
 ^NH2 CINH/ ^NHNH/ 
 
 p-Urazine. 
 
 The alkyl derivatives of urea can be made in various ways : — 
 
 1. By Wehler's method, using mono- or di-alkyl ammonium cyanate ; the 
 di-derivatives so formed are of course unsymmetrical. 
 
 2. The isocyanie esters give with ammonia monalkyl ureas, with primary 
 and secondary amines symmetrical di- and tri-alkyl ureas : — 
 
 /NC H /NH-CaHg 
 
 C^^'^^iis + HNHCH3 = C=0 
 
 " \NHCH3 
 
 1 Schestakow, C. OS. i. 1227. 
 
 ' Chattaway, Chem. News, 98. 166 (1908) ; J. C. S. 1909. 235. Cf. Chattaway, Wiiusch, J. C. B. 
 1909. 129. 
 
188 Carbonic Add Derivatives 
 
 3. The tetra-alkyl derivatives may be obtained by the action of carbonyl 
 chloride on secondary amines. 
 
 Those derivatives which no longer contain an NHg group can generally be 
 distilled unchanged. Those which contain an NH group yield with nitrous acid 
 nitroso-compounds, which on reduction give semicarbazide derivatives : — 
 
 /NH-CH3 /N<^i? /N<rH' 
 
 \NH2 \NH2 \NH2 
 
 It is interesting to observe that mono-phenyl urea, owing to the negative 
 influence of the phenyl group, will not combine with nitric acid. 
 
 /NH-CN 
 Cyan-urea, C=0 , is obtained by treating cyan-guanidine (a polymeriza- 
 
 \NH2 
 
 tion product of cyanamide) with baryta : — 
 
 /NHCN /NHCN 
 
 C=NH + H^O = 0=0 + NH3. 
 
 Xnh^ XnHj 
 
 Like nitro-urea it is a strong acid, which expels carbon dioxide from carbonates. 
 When warmed with mineral acids it takes up one molecule of water to form the 
 corresponding amide biuret, 
 
 /NH-CONHg 
 C=0 
 
 Biuret gives a pink colour when treated with copper sulphate and alkali: 
 this is the so-called biuret reaction which is used as a test for urea and certain 
 amides, such as the polypeptides. It is due to the formation of a copper 
 derivative, in which one of the hydrogen atoms of an NHg group is replaced 
 by copper.' 
 
 The tautomeric form of urea, HO-C^-j^tt , the so-called iso-urea, is not 
 
 known in the free state, but only in certain derivatives. Its methyl ether, 
 
 CH30-C<^-|^TT , is obtained in the form of its hydrochloride when hydrochloric 
 
 acid is passed into a solution of cyanamide in methyl alcohol. If this hydro- 
 chloride is heated in aqueous solution, it breaks up into methyl chloride and 
 urea.'' In the analogous case of the 0-ethers of the dialkyl-ureas, the velocity 
 of the reaction, 
 
 Alk2N-C<^J^^^ = Alk2N-C<JH2 + CH3CI, 
 
 has been measured ' ; and it has been shown that it is probably the undissociated 
 hydrochloride which reacts, and not its ions. 
 
 1 Cf. Schiff, Ann. 299. 236 (1897). 
 
 2 Stieglitz, M'Kee, Ber. 33. 1517 (1900). 
 
 ' M«Kee, Am. Ch. J. 42. 1 (C. 09. ii. 1126). 
 
Thio-ureas 189 
 
 THIO-UEEAS 
 
 In considering the constitution of the amides it was pointed out that whUe 
 there is some evidence that they can react in accordance with the isoamide 
 
 formula E-C<(qtt, yet they must in general be regarded as having the true 
 
 amide structure K-C<^q ^. But in the case of the thio-amides there was much 
 
 stronger evidence for believing that they had the constitution I^'C\gTT • 
 Precisely similar relations hold in the case of the carbonic acid derivatives. 
 
 Urea itself behaves in accordance with the formula 0:C\iyjTT^ ; and there is very 
 
 ■NH, 
 
 little reason for adopting the iso-formula HO-O^iyTTT . But thio-urea frequently 
 
 reacts as if it had the formula HS-C<^TyTi:T^ ; and though we cannot be certain 
 
 that the free substance possesses this structure, it must at least be able to 
 assume it in many reactions. 
 
 Thio-urea is prepared like urea, by heating ammonium thiocyanate. The 
 change occurs much less readily than in the case of urea, and the proportion of 
 thio-urea present at equilibrium is very much smaller. In order to bring about 
 the change at all rapidly, the dry salt must be heated to 160-170°, and at this 
 temperature equilibrium is reached after about an hour, the amount of thio-urea 
 formed being only 25 per cent. The reaction has been examined by Eeynolds 
 and Werner,^ and more satisfactorily by Findlay.' Findlay has determined the 
 freezing-point curve for mixtures of thio-urea and ammonium thiocyanate, and 
 the proportions in the liquid at equilibrium. Ammonium thiocyanate melts at 
 149°, and thio-urea at a temperature above 177°, which cannot be exactly 
 determined owing to the tautomeric change. The freezing-point curve has the 
 usual form (like that of the benzaldoximes), consisting of two branches meeting 
 at a eutectic point (32 per cent, thio-urea) at 104-3°. The liquid at equilibrium 
 contains 25 per cent, thio-urea, independent of the temperature, showing that 
 the conversion of one form into the other is not accompanied by any appreciable 
 heat effect. This line cuts the thiocyanate branch of the curve at 114-115°, 
 indicating that this is the natural freezing-point, and that the thiocyanate is the 
 stable form. From the fact that the proportions at equilibrium correspond to 
 a compound CS(NH2)2-3 NH^CNS, Eeynolds and Werner assumed that such 
 a compound was formed ; but Findlay has been unable to discover any indications 
 of its existence. 
 
 The velocity of the change in the fused substance has been measured by 
 Waddell,' who removed portions from time to time and titrated them with 
 silver. He finds the reaction under these conditions (where the ionization is of 
 course very small) to be monomolecular in both directions. 
 
 In aqueous solution, ammonium thiocyanate has even less tendency to go 
 over into thio-urea than it has in the case of the melted salt. Dutoit and 
 Gagnaux* find that at 140-170° a decinormal solution of ammonitun thiocyanate 
 
 ' J. 0. S. 1903. 1. 2 lb. 1904. 403. ' J. Phys. Chem. 2. 626 (1898). 
 
 ' /. C&im. Pkys. 4. 268 (C. 06. ii. 675). 
 
190 Carbonic Acid Derivatives 
 
 forms no urea at all, while the reverse change is complete. This latter change 
 (of thio-urea into the thiocyanate) was found to be monomolecular, like that of 
 urea into ammonium cyanate, and to be retarded by the presence of soda but 
 hastened by that of sulphuric acid. 
 
 Thio-urea can also be obtained from cyanamide by the action of hydrogen 
 sulphide in the presence of a little ammonia : — 
 
 H2NC=N + H^S = S=C<^g2 . 
 
 Thio-urea is a crystalline substance, which melts if rapidly heated at about 
 172° ; its true melting-point cannot be determined, owing to the tautomeric 
 change. It sublimes in vacuo at 160°, condensing as ammonium thiocyanate. 
 It forms a comparatively stable salt with hydrochloric acid.' Its aqueous 
 solution reacts with mercuric oxide in the cold to give cyanamide. If it is 
 treated with cold permanganate solution, it is converted into urea. When 
 oxidized in acid solution it gives salts of an unstable disulphide 
 
 H„N-C-S-S-C-NH, 
 
 ^ II II ■ 
 
 NH NH 
 
 This reaction is a strong argument in favour of the formula HS-C^t^tt • 
 
 The alkyl derivatives of thio-urea are of two kinds, one having the alkyl 
 attached to sulphur, and the other to nitrogen. The mono-, di-, and tri-N-esters 
 maybe regarded as derived from either pseudo-form ; the S-esters must of course 
 be derived from the imide (iso-) structure. 
 
 The N-esters are obtained by the action of ammonia or amines on the 
 isothiocyanic esters or mustard oils :^ 
 
 NHC,H: 
 
 S^C^N-C^Hg + H^N-CHg = S=C<jJg^g 
 
 It is found that in this reaction the same product is formed from the methyl 
 ester and ethylamine. This has been used as an argument against the imide 
 formula, but it is not of much force. If the body contained an SH group, we 
 
 should expect ethyl isothiocyanate and methylamine to give HS-C^-|^TT_pTT , 
 
 while the methyl ester and ethylamine should give HS-C^T^-rrp tt . But 
 
 these bodies, differing only in the position of a hydrogen atom, would obviously 
 change into one another with the utmost ease. 
 
 Symmetrical diphenyl-thio-urea (thiocarbanilide) is readily obtained by the 
 action of carbon bisulphide on aniline : — 
 
 < * V4 = ^™? * ^^- 
 
 This reaction is characteristic of the aromatic amines, the fatty (and under 
 certain conditions the aromatic also : see p. 183) giving salts of dithiocarbamic 
 acid. 
 
 The preparation of thiocarbanilide requires several hours boiling ; but it may 
 be much shortened by the addition of certain catalytic agents. For this purpose 
 
 ' J. C. S. 1902, 79. 
 
Thio-ureas 191 
 
 powdered potash is sometimes used : or a few grams of sulphur may be dissolved 
 in the carbon bisulphide, when the reaction is complete in about an hour. 
 Hydrogen peroxide has a similar effect.' The reaction fails unexpectedly in the 
 case of certain amines ; but no regularities in this have been observed. 
 
 The mixed symmetrical di-derivatives of thio-urea of the type S=C<^jjjj'q®jj* 
 
 when heated are partially converted by a reversible reaction into a mixture of 
 the two simple di-derivatives SC(NH0)2 and SC(NH-C6H4X)2 (where X = CH3, 
 NO2, &c.). This is obviously due to their breaking up into amine and mustard 
 oil, which then recombine.^ 
 
 The N-substituted thio-ureas on treatment with mercuric oxide behave 
 differently according as they do or do not contain an intact NHg group. If they 
 do, they lose hydrogen sulphide and form cyanamide derivatives : — 
 
 S=C<g|^^^3 + HgO = HgS + H^O + C^^-^^s. 
 
 If they do not, they merely have the sulphur replaced by oxygen. Thus 
 thiocarbanilide when boiled with mercuric oxide in alcoholic solution gives 
 carbanilide or symmetrical diphenyl-urea : — 
 
 S=<|^ + HgO = 0=C<5;H^ + HgS. 
 
 It is, however, possible to bring about a change of the first kind, the removal 
 of hydrogen sulphide, with thiocarbanilide, by boiling it in benzene solution 
 with mercuric oxide. The product is then a derivative of the pseudo-form of 
 cyanamide, carbodiimide : — 
 
 ^<^<NH0 + ^SO = C<N0 + HgS + H,0. 
 The S-derivatives of thio-urea are obtained as hydriodides by treating thio- 
 urea with alkyl iodide : — 
 
 jjjj ^NH 
 
 S=C<{Jg2 + C2H5I = C-S-C^Hs-HI. 
 
 On treatment with alkali or silver oxide the free thioether is obtained. It is 
 a strongly basic substance, whose structure is shown by its giving a mercaptan 
 on saponification, and a sulphonic acid on oxidation. 
 
 The formation of these S-derivatives, which are known as pseudo- thio-ureas, 
 is an argument in favour of the imide formula of thio-urea. Similar compounds ' 
 are produced by the action of the esters of chlorocarbonic acid, or of acid 
 
 ^NH 
 chlorides, on thio-urea. Thus acetyl chloride gives a compound C-S-COCHj. 
 
 Xnh^ 
 
 These bodies are somewhat unstable, and if their hydrochlorides are heated, 
 the acyl group migrates to the nitrogen, giving an acyl thio-urea, such as 
 
 /NHCOCH, 
 
 (^S 
 \NH2 
 
 1 y. Braun, Beschke, Ber. 39. 4369 (1906). ' Hugershoff, Ber. 36. 1138 (1903). 
 
 ' A. E. Dixon, Hawthorne, J. C. S. 1907. 122. 
 
192 Carbonic Acid Derivatives 
 
 The derivatives of carhodiimide, the diimide of carbonic acid HN=C^NH, are 
 not of great importance. This formula is tautomeric with that of cyanamide, 
 and therefore the monalkyl cyanamides might be supposed to be mono-substitu- 
 tion products of carbodiimide. The evidence, however, is in favour of the true 
 cyanamide formula for these compounds, which wUl therefore be dealt vnth 
 among the cyanogen derivatives. 
 
 But there are a few bodies which undoubtedly possess this structure. One 
 of these has just been mentioned, and it is typical of the class. They are 
 formed, but only in certain cases, by the action of mercuric oxide on sym- 
 metrical di-derivatives of thio-urea. Thus dipropyl urea yields the compound 
 CgHYN^C^NCsH,. This is the only known alkyl-carbodiimide ; and in the 
 aromatic series only a few have been prepared. The diphenyl compound 
 C(=N0)2, obtained from thiocarbanilide, can be distilled under 31 mm. pressure 
 at 218°. The distillate forms a glassy mass which, by treatment with boUing 
 ligroin, can be divided into two parts, a soluble oil and an insoluble powder 
 melting at 158-160°. These two substances have the same composition, and 
 their chemical behaviour is identical. They are possibly stereoisomers, since 
 the body contains two N=C groups. 
 
 This substance is characterized, as one would expect, by a strong tendency 
 to form addition-compounds. When treated with alcohol and hydrochloric acid 
 it takes up water to give carbanilide CO(NH0)2. It combines with hydrogen 
 sulphide at the ordinary temperature to give thiocarbanilide, with aniline to 
 give triphenyl-guanidine : — 
 
 and with carbon bisulphide at 140° to give phenyl mustard oil (isothiocyanate) : — 
 - With phenol and nitrophenol it forms an ether of iso-urea : — 
 
 But it is remarkable that 2,4-dinitrophenol and picric acid add on in a different 
 way, forming a normal urea derivative ' : — 
 
 In all these reactions carbodiphenyl-itnide shows a close analogy to an iso- 
 cyanate, both containing the group ^C=N-Ar ; and this analogy extends to 
 their polymerization products, the diimide giving a melamine, i. e. a derivative 
 of the amide of cyanuric acid. 
 
 1 Busoh, Elume, Pungs, J. pr. CA. [2] 79. 513 (1909). 
 
Guanidine Compounds 193 
 
 GUANIDINE DEEIVATIVES 
 
 Guanidine, C— NH , is the amidine of carbamic acid. The name was given 
 
 to it by its discoverer, Strecker, who prepared it in 1861 from guanine, a 
 naturally occurring substance of the uric acid group. 
 
 Its constitution is shown by a variety of syntheses. It is formed by the 
 action of ammonia on carbonyl chloride, orthocarbonic ester C(OC2Hg)4, chloro- 
 picrin CCI3NO2, or cyanogen iodide. One method of preparation which illus- 
 trates its structure peculiarly well is by heating an alcoholic solution of 
 cyanamide with ammonium chloride : — 
 
 //N /NH2 
 
 The method adopted in practice for preparing guanidine is to heat ammonium 
 thiocyanate to 180-185° for twenty hours, when ammonium trithiocarbonate, 
 (NHJaCSg, sublimes, and nearly pure guanidine thiocyanate remains behind. 
 The reaction probably proceeds in several stages. Thio-urea is first formed, and 
 this breaks up at the high temperature into hydrogen sulphide and cyanamide. 
 The latter combines with ammonium thiocyanate to form guanidine thiocyanate: — 
 
 C<t;„ + NH4CNS = C=NH-HCNS, 
 
 while the hydrogen sulphide combines with more ammonium thiocyanate to 
 give the trithiocarbonate : — 
 
 NH4-S-C=N + 2 HjS = NH4-S-C-S-NH4. 
 
 S 
 
 Free guanidine is a colourless mass, strongly basic and very caustic. It 
 deliquesces in the air, and rapidly absorbs carbon dioxide from it. It forms 
 stable salts with one equivalent of acid, which are mostly soluble, except the 
 nitrate, which, like urea nitrate, is comparatively insoluble. 
 
 As the imide of urea it is converted into urea on treatment with baryta. 
 
 If its hydrochloride is heated to 180° it gives off ammonia and forms 
 biguanide (corresponding to biuret) : — 
 
 (^NH /NH2 
 
 \NH2 C=NH 
 
 = )NH + NH,. 
 
 /NH2 C=NH 
 
 (^NH \NH2 
 
 \NH2 
 
 On treatment with fuming nitric acid guanidine is converted into nitro- 
 
 /NH-N02 
 guanidine, C^^NH , a feebly acid substance which can be reduced first to 
 
 \nh„ 
 
 /NH-NHa 
 nitroso- and then to amino-guanidine, Ct=NH . This substance when, 
 
 \NH2 
 
194 Carbonic Add Derivatives 
 
 treated with nitrous acid yields the azide : — 
 
 N< II 
 /NH-NHj + HONO /' \N 
 
 C=NH = C=NH + 2 H2O . 
 
 In accordance with this formula it breaks up when treated with alkali into 
 cyanamide and hydrazoic acid HN3. 
 
 The substituted (alkyl or aiyl) guanidines may be obtained by acting on 
 cyanamide with amine hydrochlorides instead of ammonium chloride. Sym- 
 metrical triphenyl guanidine, 0N:C(NH^)2, is formed as a by-product in many 
 reactions in which the main product is symmetrical diphenyl urea or a similar 
 substance. Thus it is formed in the preparation of thiocarbanilide : — 
 
 /NH0 /NH0 
 
 (^S + H2N0 = HjS + C=N^ . 
 
 \NH0 \NH^ 
 
 There are two guanidine derivatives which are of great physiological im- 
 portance, creatin and creatinin. Creatin occurs in the muscles of mammalia, 
 a full-grown man containing about 90 grams. It can be synthesized by the 
 action of sarcosin (methyl-amino-acetic acid) on cyanamide, whereby its formula 
 is established :— 
 
 J, /N— CH2COOH 
 
 -■f^ + CHsNHCHoCOOH = C=NH 
 
 On boiling with water, especially in presence of acids, it loses water and goes 
 over into its cyclic amide, creatinin : — 
 
 /N(CH3).CH2 
 
 \NH CO 
 
 a substance which occurs to a small extent in urine. 
 
CHAPTEE IX 
 
 CYANOGEN COMPOUNDS 
 
 The derivatives of cyanogen may be classified as follows : — 
 
 I. The cyanides and isocyanides, E-C=N and E-N=C or E-N=C. 
 
 II. The compounds containing the group -(CNO). Of this group four 
 isomeric arrangements are possible, and representatives of all of them are 
 actually known : — 
 
 1. The cyanates E-0-C=]Sr. 
 
 2. The isocyanates E-N=C=:0. 
 
 3. The fulminates, E-0-N=C, which are the esters of the oxime of carbon 
 monoxide. 
 
 4. The nitrile-oxides -^ / • 
 
 In the first two of these groups the oxygen may be replaced by sulphur, 
 giving the thiocyanates, E-S-C=N, and the isothiocyanates or mustard oils, 
 E-N=C=S. 
 
 III. The tricyanogen derivatives. All the above-mentioned classes of com- 
 pounds readily form triple polymers containing the so-called tricyanogen ring 
 
 C 
 
 N N 
 I I 
 C 
 
 N 
 
 The first of the cyanogen compounds to be prepared was Prussian blue, 
 
 which was discovered by Diesbach at the beginning of the eighteenth century ; 
 
 but our real knowledge of their constitution dates from the work of Gay-Lussac 
 
 in 1815. Gay-Lussac showed that these bodies contain a radical composed of 
 
 carbon and nitrogen, which plays the part of an element. This radical, which 
 
 may be written -C=N, -N=C, or -N=C", resembles the halogens, especially in 
 
 forming a hydracid, and in the fact that the molecule of the free substance 
 
 consists of two such groups joined together, CN-CN being analogous to Cl-Cl. 
 
 CYANOGEN (CN)^. 
 
 Cyanogen is generally said to be formed by passing the electric arc between 
 carbon poles in an atmosphere of nitrogen. It has however been shown that it 
 cannot be prepared in this way. It follows from Nernst's theorem that a con- 
 siderable percentage of cyanogen must be formed at a temperature of 3,500°, but 
 it is apparently all decomposed in passing out of the arc* It is usually prepared 
 by heating mercuric cyanide, by which method it was first obtained by Gay- 
 Lussac in 1815. Another method of making it is by the action of potassium 
 
 1 Berthelot, C. R. 144. 354 (1907) ; Wallis, Aim. 345. 353 (1907) ; Wartenberg, Z. an. Ch. 
 52. 299 (JO. 07. i. 800). 
 
196 Cyanogen Compounds 
 
 cyanide on copper sulphate in solution. Cupric cyanide, or a double compound 
 of this with potassium cyanide, is first formed ; but on heating the solution this 
 breaks up into cuprous cyanide and cyanogen which is evolved. It is also 
 contained in coal gas, along with prussic acid, in sufficient quantity to supply 
 the whole commercial demand for cyanogen compounds ; and a proposal has 
 been made ' to utilize this source by passing the gas through a mixture of 
 a ferrous salt and lime, when the cyanogen is obtained as calcium ferrocyanide. 
 It has further been proposed to make it on the large scale by passing nitrogen over 
 iron covered with coke (and therefore saturated with carbon) at 1,500-1,800°.^ 
 There are three possible formulae for cyanogen : — 
 
 C=N N=C C=N 
 
 I , I , I . 
 
 C=N N=C N=C 
 
 Cyanogen is the only known substance of the molecular formula CjNj , and it is 
 
 proved to have the first of these structures by its formation from ammonium 
 
 oxalate or oxamide on treatment with phosphorus pentoxide : — 
 
 0=C-NH2 C=N 
 
 and from glyoxime (the oxime of glyoxal) on treatment with acetic anhydride : — 
 
 H-C=NOH C=N 
 
 I - 2 H„0 = I 
 
 H-C=NOH ^ C=N 
 
 Both of these reactions show that the two carbon atoms are linked together. 
 
 Cyanogen is a colourless gas of a peculiar smell ; it boils at -20-7° and melts 
 at -34 "4°. Iws extremely poisongus. It is very stable to heat, not being 
 affected even aT^SOO" ; abovethat temperature it slowly breaks up. Liquid 
 cyanogen has a very low electric conductivity, and practically no dissociating 
 power.' Cyanogen gas burns with a characteristic purple-mantled flame. One 
 volume of water dissolves four volumes of the gas, one volume of alcohol 
 twenty- three.'' If the saturated aqueous solution is left in contact with the gas, 
 it continues to absorb it slowly, as cyanogen is unstable in solution, and at once- 
 begins to decompose, depositing a brown amorphous mass known as azulmic 
 acid, while the solution contains ammonium oxalate, ammonium carbonate, 
 prussic acid, and urea. 
 
 As the nitrile of oxalic acid cyanogen is saponified by fuming hydrochloric 
 acid or by hydrogen peroxide to oxamide ; and like chlorine it is absorbed by 
 aqueous potash to give a mixture of the cyanide and cyanate : — 
 (CN)2 + 2K0H = KCN + KCNO + H^O. 
 
 In the presence of a trace of sodium ethylate, cyanogen is able to add itself 
 on to compounds containing an acidic methylene group.' Thus with acetoacetic 
 ester it gives at low temperatures a body 
 
 CHg-COCHCOOEt 
 
 C=N 
 
 1 Feld, Nausa, C. 03. i. 264. " Erlwein, C. 08. ii. 273. 
 
 » Centnerszwer, Z. Ph. Ch. 39. 217 (1901). * Berthelot, 0. 04. ii. 428. 
 
 5 \V. Traube, Ber. 31. 191, 2938 (1898). 
 
Cyanogen 197 
 
 while at higher temperatures one molecule of cyanogen reacts with two 
 molecules of the methylene compound, giving bodies of the type 
 
 R-CH-R 
 
 C=NH 
 
 I 
 C=NH 
 
 R-CH-R 
 
 Paracyanogen, a polymer of cyanogen of unknown molecular weight, is 
 
 formed as a by-product in the preparation of cyanogen by heating mercuric 
 
 cyanide, and is the form in which the whole of the cyanogen is obtained when 
 
 an aqueous solution of potassium cyanide is electrolysed. It is a brown 
 
 amorphous insoluble powder, which is reconverted into cyanogen gas at 860°. 
 
 HYDROCYANIC ACID 
 
 The simplest derivative of cyanogen is hydrocyanic or prussic acid, whose 
 formula is either H-C=N or H-N=C". It is a tautomeric substance : that is 
 to say it gives rise to derivatives of both forms. It is an extremely complicated 
 question which of these two formulae is to be adopted for the acid and for its 
 salts ; and as its solution depends on a comparison of their properties with 
 those of the organic derivatives — the nitriles and the isocyanides — the discussion 
 will be deferred until these bodies have been considered. 
 
 Prussic acid occurs in nature in the glucoside amygdalin, which is contained 
 in bitter almonds, and when it is hydrolysed — which can be done by means of 
 .a ferment also contained in the almonds — splits up into prussic acid, benz- 
 aldehyde, and sugar. In the free state it is found in the tree Pangium edule 
 which grows in Java. All parts of this tree contain free prussic acid, but 
 particularly the seeds ; and it is calculated that a full-grown tree contains at 
 least 350 grams. 
 
 It has recently been shown ' that prussic acid is much more widely distri- 
 buted in the vegetable world than had previously been supposed, occurring 
 mainly in the form of glucosides. There is some reason to think that it may be 
 ■produced by the action of the carbohydrates on the inorganic nitrates, whose 
 presence can often be detected in those plants which yield prussic acid." 
 
 Prussic acid was discovered by Scheele in 1782, and was first obtained in 
 the anhydrous state by Gay-Lussac in 1811. It is formed by the combination 
 of hydrogen and cyanogen at high temperatures, or under the influence of 
 the silent electric discharge ; or by passing sparks through a mixture of 
 nitrogen or ammonia and acetylene, or of carbon monoxide, nitrogen, and 
 hydrogen.' It can also be made* by passing a carefully dried mixture of 
 hydrogen, ammonia, and a volatile carbon compound (such as carbon monoxide, 
 
 1 Arragon, Guignard, C 06. ii. 1849 ; cf. Greshoff, Arch, der Pharm. 244. 397 (1906 
 Couperot, C. 09. i. 387. 
 
 2 Eavenna, Peli, C. 08. i. 654 ; cf. Plimmer, C. 05. i. 357. 
 
 ' Gruszkiewicz, Z.f. SleMrochem. 9. 83 (C. 03. i. 494) ; Hoyenuazm, C. 02. i. 525 ; Briner, 
 Durand, J. CMm. Fht/s. 7. 1 (C. 09. i. 1453). 
 * Woltereok, C. 04. i. 1306. 
 1175 O 
 
198 Cyanogen Compounds 
 
 carbon dioxide, acetylene, or the vapour of alcohol or petroleum) over heated 
 platinized pumice ; and this method has been used for making it on the large 
 scale. It is formed when ammonia comes in contact with red-hot carbon, as 
 in the manufacture of coal-gas, the conversion being complete at 1300°; and 
 by the action of a high temperature on many nitrogenous substances containing 
 carbon : thus it has been found ^ in cigar smoke to the extent of one milligram 
 to every 4 or 5 (Austrian) cigars. Its salts are produced when organic 
 nitrogenous substances are raised to a high temperature in contact with 
 potassium or sodium, or when metallic nitrates are heated with charcoal in 
 closed vessels.' It is also formed from many phenols (especially nitrophenols) 
 and similar compounds by boUing with dilute nitric acid.^ 
 
 Prussic acid is usually prepared by the action of dilute sulphuric acid on 
 potassium ferrocyanide : — 
 
 3 H2SO4 -t- 2 K^FeCyg = 3 KaSO^ + KaPe-FeCyo -f 6 HON. 
 It is also produced by the dry distillation of ammonium formate : — 
 
 H-C<g_ 
 
 jjjj^ = 2 H^O -(- H-C=N, 
 
 which has been used as an argument for the nitrile formula : and by the action 
 of chloroform and ammonia on alcoholic potash : — 
 
 CHCI3 -t- NH3 + 8K0H = H-N=C 4- 3KC1 + SHaO, 
 which is an equally weak argument in favour of the isocyanide formula. 
 
 Hydrocyanic acid is a colourless liquid with a smell like bitter almonds. 
 It is remarkable that different people differ more in their power of detecting 
 this smeU than any other. It burns with a violet flame. It is one of the 
 most powerful of poisons ; the best antidote to it is either hydrogen peroxide, 
 or inhaling air containing chlorine. 
 
 Liquid prussic acid has a very high dielectric constant, ■* higher indeed 
 than that of water ; in this it resembles the nitriles, for which also the values 
 of this constant are high, though not so high. It also, as we should expect 
 from this, has a high dissociating power ' ; and some salts, such as potassium 
 iodide, conduct better in prussic acid than in water. Cryoscopic measurements 
 have also shown that these salts are practically completely dissociated in this 
 solvent, while acids (such as trichloracetic and sulphuric) give normal values and 
 appear not to dissociate at all." 
 
 It has been shown by Nef ' that anhydrous prussic acid as usually prepared 
 contains a small quantity of ammonium cyanide, by which its properties are 
 sensibly modified. If the liquid is allowed to stand over phosphorus pentoxide, 
 and then distilled through tubes containing phosphorus pentoxide and warmed 
 to 40-50°, a pure acid is obtained. This boils at 25°, and solidifies on cooling 
 to a colourless crystalline mass melting at — 10° to — 12°. When completely 
 
 ' Habermann, C. 03. i. 53. ^ Miiller, C. 08. i. 1343. 
 
 " Seyewetz, Poizat, C. li. 148. 286, 1110 ; Bull. Soc. [4] 5. 489 (C. 09. i. 997). 
 * Schlundt, J. Fhy$. Chem. 5. 167 (C. 01. i. 1135). 
 
 " Centnerszwer, Z. Ph. Ch. 39. 220 (1901) ; KaMenberg, Schlundt, J. Phys. Chem. 6. 447 
 (C. 03. i. 1). 
 
 6 Lespieau, C. B. 140. 855 (1905). "> Ann. 287. 327 (1895). 
 
Prussic Add 199 
 
 purified in this way prussic acid is an extraordinarily stable substance. It may 
 be kept for months in sealed tubes without change, and it is not acted on at — 10° 
 by chloiine, hydrochloric acid, or ethyl hypochlorite. If the acid contains traces 
 of impurity, such as water or potassium cyanide, it rapidly turns brown, 
 forming the so-called azulmic compounds, about which very little is known. 
 The aqueous solution rapidly decomposes, especially if exposed to light, forming 
 a brown precipitate of azulmic acid, while ammonia, formic and oxalic acids, 
 and other substances remain in solution. In the presence of a trace of mineral 
 acid the aqueous solution is more stable. 
 
 On reduction prussic acid yields methylamine. If it is hydrolysed with 
 concentrated aqueous hydrochloric acid it gives formamide, while an alcoholic 
 solution of hydrochloric acid converts it into esters of formic acid. These 
 reactions are equally consistent with either of the two formulae. 
 
 It reacts with diazomethane, and according to v. Pechmann the product is 
 methyl cyanide : — 
 
 ChTiI + H-C=N = CH3-C=N + No. 
 
 This is commonly regarded as a strong argument for the nitrile formula, since 
 diazomethane is a general reagent for replacing hydrogen by methyl, and 
 especially since the reaction goes on at a low temperature. But it has recently 
 been found that methyl isocyanide is formed at the same time, which destroys 
 the force of this argument ; and indeed the nitrile itself is at least as likely to 
 be formed from H-N=C, as will be shown later. 
 
 Owing no doubt to its highly unsaturated character, prussic acid has a great 
 power of forming addition-compounds. Thus it combines with hydrochloric 
 acid in ethyl acetate solution, giving a body of the composition 2HGN-3HCL 
 The formula of this substance has been shown to be 
 
 ^-^^NH-CHCla' ^^^■ 
 It is dichlormethyl-formamidine hydrochloride. This is proved by its giving 
 with alcohol formamidine : — 
 
 ^NH-CHCl^ + ^^*^^ = ^^*^^ + ^^^^2 "^ ^-^^NH^' 
 
 and further by its reacting readily with benzene and aluminium chloride to 
 give diphenyl-formamidine :-^ 
 
 H<H.CHC1, + 2^«He = H.C<™cH^^ + 2HC1. 
 Hydrocyanic acid can also form a simpler compound with hydrochloric acid, 
 imino-formyl chloride, C1-C<^tt . This body cannot be isolated, but its forma- 
 tion is proved by Gattermann's reaction for the synthesis of aldehydes.' If an 
 aromatic hydrocarbon or a phenol or a phenol ether is treated with prussic acid 
 and hydrochloric acid, either alone, or in presence of a condensing agent such 
 as zinc chloride or aluminium chloride, it is converted into an aldimine, from 
 
 1 J3er. 31. 1140,, 1765 (1S98) ; Ann. 347. 3i7 (1906) ; 367. 313 (1907). 
 
 o2 
 
200 Cyanogen Compounds 
 
 which the aldehyde is easily obtained by heating with dilute acid. This can 
 only be explained by the intermediate formation of the imino-formyl chloride : — 
 
 CeHe + C1-C<5JH = HCl + CeH,.C<JJ^ -* NH3 + CeH,-C<g. 
 
 Prussic acid will combine with aldehydes and ketones to form the so-called 
 cyanhydrins, which are really the nitriles of the a-oxy-acids : — 
 
 (CH3)2C=0 + HON = (CH3),C<gJ. 
 
 A reaction which very probably depends on the formation of a cyanhydrin 
 is the benzoin synthesis. If benzaldehyde is heated in alcoholic solution with 
 a small quantity of potassium cyanide, it is converted into benzoin : — 
 
 0CHO + OHC-0 = ^CHOH-CO-^. 
 The same reaction occurs with many aromatic aldehydes (including furfurol), but 
 it does not take place in the fatty series. The action of the potassium cyanide 
 is catalytic : that is to say, it does not appear in the equation, and a small 
 quantity of it can effect a large amount of the reaction. The dynamics of this 
 change have been investigated by Stern.' The reaction was carried out in 
 sealed tubes at 60°, in 60 per cent, aqueous alcohol. In order to determine the 
 amount of change after a given time, the reaction was stopped by acidifying the 
 solution, the prussic acid and benzaldehyde were distilled off with steam, and 
 the benzoin was estimated with Fehling's solution. It was found that vsdth 
 a constant quantity of the catalytic agent, the rate of formation of benzoin was 
 proportional to the square of the concentration of the benzaldehyde, as we 
 should expect from the equation. If the amount of potassium cyanide is varied, 
 the bimolecular constant varies in proportion to it. Sodium or barium cyanide 
 has practically the same effect as potassium cyanide ; prussic acid itself does 
 not bring about the synthesis at all. This shows that the reaction depends on 
 the cyanogen ion. This was further proved by the fact that if silver or mercuric 
 cyanide is added as well as potassium cyanide, the constant obtained is the 
 same as that calculated for the excess of potassium cyanide after subtracting 
 the amount required for the complete formation of the double cyanide of 
 potassium and silver or potassium and mercury ; in other words, the part of the 
 cyanogen ion which has gone over into the complex ion has no influence. 
 
 The reaction is hastened by increasing the proportion of water in the solvent. 
 
 Various theories as to the mechanism of this reaction have been suggested ; 
 those which ascribe it either to prussic acid or to hydroxyl ion are shown by 
 these experiments to be untenable. It is certainly due to the potassium cyanide, 
 or to the cyanogen ion. The suggestion which best agrees with the facts is that 
 of Chalanay and Knoevenagel.'' They suppose that three successive reactions 
 take place : — 
 
 (1) 0-C<Q + KCN = ^•C<Q + HCN, 
 
 (2) 0-C<^ + HCN = '^■CH<gJ, 
 
 (3) ^•CH<gJ + 0-C<Q = KCN + 0-CHOH.CO-0. 
 
 1 Z. Ph. Ch. 50. 513 (1906). ' Ser. 25. 295 (1892). 
 
Prussia Add: Benzoin Synthesis 201 
 
 If we assume that reactions (1) and (2) are very rapid, and that the quantity of 
 the intermediate products formed is very small, it can easily be shown that the 
 rate of formation of benzoin is proportional to the concentration of the potassium 
 cyanide, and to the square of the concentration of the benzaldehyde. 
 
 Prussio acid can add on to the grouping C=C, but with much less ease than 
 to the grouping C=0, and only in particular cases ; whereas sulphurous acid 
 combines with either of these groups with equal ease. An example is the 
 formation of nitrile-acids from certain unsaturated acids: — 
 
 ? 
 
 H3 CH3 
 
 CH CHCN 
 
 II + HON = Y 
 
 CH (j)H, 
 
 COOH COOH 
 
 As this reaction is greatly hastened by the presence of primary or secondary 
 bases, it is probably due to the cyanogen ion.^ 
 
 If an aqueous solution of prussic acid, containing a little potassium carbonate, 
 is allowed to stand, it deposits azulmic acid ; and if it is now extracted with 
 ether, the ether is found to contain a polymer of prussic acid. This substance 
 on saponification with acid or alkali yields carbon dioxide, ammonia, and 
 glycocoU. It is, therefore, probably the nitrile of amino-malonic acid : — 
 
 H-C3T ?=N ?00H 
 
 H-C=N = H-CNH2 -^ H-CNH2 + 2 NH3 -* H2CNH2 + CO2. 
 
 TT_P='Nr I I I 
 
 u. o-i>i ^=^ GOOU. COOH 
 
 Salts of Prussic Acid 
 
 Prussic acid is a distinct acid, though a very weak one. Its dissociation 
 constant (K = 0-0013 x 10"") is less than a ten-thousandth of that of acetic acid 
 (18-0 X 10~'). It turns litmus a dark red, and forms salts, which, however, are 
 decomposed by carbonic acid. The salts of the alkalies and alkaline earths are 
 soluble in water, have an alkaline reaction, and are stable even at a red heat. 
 The salts of the heavy metals on the other hand are insoluble (except mercuric 
 cyanide), are not decomposed by any but strong acids, and break up at a red 
 heat into the metal and cyanogen. Mercuric cyanide has a remarkable behaviour, 
 owing to the fact that while it is soluble in water it is not ionized to a perceptible 
 extent, so that it is scarcely correct to call it a salt. Its solution does not 
 conduct electricity, and does not give the ordinary tests for mercury or for a 
 cyanide ; it gives no precipitate either with silver nitrate or with alkali.^ 
 
 The cyanides of the heavy metals all dissolve in potassium cyanide with the 
 formation of double cyanides. Many of these double cyanides do not give the 
 reactions either of cyanides or of the heavy metals which they contain. The 
 ferro- and ferricyanides, for example, are scarcely poisonous : on treatment with 
 acids in the cold they liberate ferro- and ferricyanic acids, H4PeCy6 and H3FeCy6 ; 
 and the iron is not precipitated by alkali or alkaline sulphide. All these facts 
 
 1 Knoevenagel, Ber. 37. 4065 (1904). 
 
 " It is, however, capable under certain conditions of giving tests for these ions. See Hoftnann, 
 Wagner, Ser. 41. 317 (1908). Cf. BoreUi, Oaz. 38. i. 361 (0. 08. ii. 288). 
 
202 Cyanogen Compounds 
 
 indicate that the iron is not present as cation but as part of the anion — of the 
 complex acid radical. Hence the ordinary tests for cyanion and for ferro- and 
 ferri-ion fail, because the solution does not contain these ions (to a measurable 
 extent), but only K' and FeCye'" or FeCyo"". 
 
 The double cyanide of potassium and silver, KAgCyg, is much used in 
 commerce, being the only soluble silver compound which gives satisfactory 
 results in silver plating. The reason for this is interesting. The only way to 
 obtain an electrolytic deposit of silver of the proper consistency is to employ 
 a bath in which the concentration of silver ions is very small. This condition 
 might be secured by using silver nitrate, if the solution were made sufiBciently 
 dilute; but since silver nitrate is practically completely dissociated, the con- 
 centration would have to be almost infinitesimal, and this would cause the bath 
 to have an enormous resistance and the electrolysis to go on very slowly. But 
 potassium argenticyanide, when dissolved in water, breaks up almost entirely 
 into the ions K' and AgCyg' ; while to an excessively small extent the latter 
 dissociates further into Ag' and CN'. It is therefore possible by using the 
 double cyanide to obtain a solution which conducts well, and contains a con- 
 siderable quantity of silver, and yet a very small concentration of silver ions. 
 As fast as the silver is deposited a fresh quantity of silver ions is formed, so as 
 to preserve the equilibrium ; and thus the bath does not become exhausted. 
 
 If an alternating current is passed through a potassium cyanide solution by 
 means of copper electrodes, the positive current carries copper ions into the 
 solution ; these ions combine with the cyanion to form complex anions, from 
 which, therefore, the copper is not deposited by the reverse current ; so that the 
 poles dissolve. But if the alternations of the current are very rapid, there is not 
 time for the copper ion to form the complex ion before the reverse current 
 brings it out of solution again. By observing the rate of alternation at which 
 the solution of the electrodes ceases, it is possible to determine the time required 
 for the formation of the complex ion. This is naturally less the more con- 
 centrated the potassium cyanide ; it lies between x^ro ^^d gocioo "^ * second.' 
 
 The reaction between potassium cyanide and hydrogen peroxide has been 
 investigated in detail by Orme Masson," who has shown that it proceeds in two 
 directions. First, the cyanide is oxidized to cyanate, and this is hydrolysed, 
 under the catalytic influence of the peroxide, to ammonia and carbonic acid. 
 At the same time a part of the cyanide is hydrolysed, without oxidation, to 
 ammonia and formic acid. The rate of oxidation is proportional to the product 
 of the concentrations of the peroxide and the cyanide. The most remarkable 
 fact is that the second reaction (the production of formate) always takes place to 
 the same extent, one-fifth of the cyanide being hydrolysed in this way, while 
 four-fifths are oxidized to cyanate. 
 
 » Le Blanc, Schick, Z. Ph. Ch. 46. 213 (1904) ; Le Blanc, Z. EUhtrochem. 11. 705 (C. 05. ii. 
 1619) ; Lob, Z. EleMrochem. 12. 79 (0. 06. i. 728). Cf. also Eucken, Z. Ph. Ch. 64. 562 (1908), 
 who finds the velocity constant for the dissociation of the AgCyj' ion to be greater than 7"5 x 10^° : 
 that is, if the AgCy/ ions were maintained at normal concentration, in one second more than 
 7-5 X 10^° gr. molecules would dissociate per litre. 
 
 2 J. C. 8. 1907. 1449. 
 
Nitriles: Formation 203 
 
 NITEILES 
 
 Prussia acid, in accordance with its tautomeric character, gives rise to two 
 classes of esters. The normal cyanides or nitriles are derived from the form 
 H-C=N. They may be regarded either as hydrocarbons in which one hydrogen 
 has been replaced by the monovalent CN group, or as acids in which the 
 carboxyl group has been converted into CN. Accordingly CH3-CN, for example, 
 may be called methyl cyanide on the analogy of methyl chloride, CHg-Cl, or 
 acetonitrile, after the acid into which it can be converted. 
 
 The nitriles were discovered in 1834, when Pelouze obtained propionitrile 
 by distilling barium ethyl sulphate with potassium cyanide. In 1847 Dumas 
 prepared acetonitrile by the action of phosphorus pentoxide on ammonium 
 acetate. Their reactions were investigated soon after this, mainly by Dumas 
 and by Frankland and Kolbe, who showed that on hydrolysis they yield acids 
 containing the same number of carbon atoms. This reaction establishes their 
 constitution. 
 
 Methods of formation 
 
 1. From potassium cyanide and alkyl iodide. This generally requires the 
 presence of a solvent. The alkyl iodide is dissolved in aqueous alcohol and the 
 cyanide added. The reaction takes place either in the cold or on warming ; 
 a certain amount of isocyanide is formed at the same time. In some cases, as 
 with tertiary butyl iodide, it is better instead of potassium cyanide to use the 
 double cyanide of potassium and mercury. 
 
 2. From the salts of the alkyl-sulphuric acids, by distillation with potassium 
 cyanide. Here, too, a little isocyanide is generally formed as a by-product, but 
 it can easily be destroyed by shaking with cold concentrated hydrochloric acid, 
 which immediately hydrolyses the isocyanide, while it leaves the nitrile practically 
 unaffected. 
 
 3. From cyanogen chloride and zinc alkyl. 
 
 4. By heating the isocyanides to about 250°, when they change over into 
 nitriles. 
 
 In all these reactions the carbon chain is lengthened ; but there are also 
 a series of methods of formation from the acids, i.e. without altering the carbon 
 chain : — 
 
 1. By converting the acid into its amide, and treating this with a dehydrating 
 agent : — 
 
 «-<NH, - «<H, -* ^■''^- 
 The dehydrating agent usually employed is phosphorus pentoxide ; but in 
 some cases it is found better to treat the amide in pyridine solution with 
 carbonyl chloride.^ The carbonyl chloride seems to act only by removing the 
 elements of water, and the pyridine facilitates the reaction by taking up the 
 hydrochloric acid formed : — 
 
 KCO-NHa + COCI2 = ECN + COg + 2 HCl. 
 
 1 Einhom, Mettler, Ber. 35. 3647 (1902). 
 
204 Cyanogen Compounds 
 
 The whole change from acid to nitrile may be carried out in one operation by 
 distilling the acid with potassium thiocyanate : — 
 
 CH3COOH + HCNS = CH3CONH2 + COS = CH3CN + HgS + CO2. 
 
 2. By distilling the acids in a stream of ammonia through a red-hot tube. 
 This explains the occurrence of nitriles in animal oil, and of acetonitrUe in 
 commercial benzene. 
 
 3. From the aldehydes by conversion into aldoximes and dehydration with 
 acetic anhydride : — 
 
 CH3-CH:N0H = H2O + CH3CN, 
 
 Another method by which an acid can be converted into a nitrile — in this 
 case containing an atom of carbon less — is the Hofmann reaction. In this 
 reaction, as has already been explained, the acid is converted into its amide, and 
 this is treated with bromine and potash. This gives primarily an amine vsdth 
 one atom of carbon less than the original acid ; but if the amine so formed 
 contains more than four carbon atoms, the bromine and potash act on it further 
 to oxidize it, removing hydrogen, and forming a nitrile : — 
 
 CgHii-CONHa -^ C^Hg-CHj-NHa -^ C^Hg-CN. 
 
 This last stage is really a reversal of the Mendius reaction, in which a nitrile is 
 reduced to a primary amine. 
 
 In the case of the aromatic compounds, the above methods require to be 
 somewhat modified. Chlorobenzene will only react with potassium cyanide 
 above 300°, and hence this method is not usually adopted for preparing the 
 nitrile. On the other hand benzoic acid, like acetic acid, readily yields its nitrile 
 when treated with potassium or lead thiocyanate ; and its amide does so on 
 treatment with carbonyl chloride in the presence of pyridine. But the simplest 
 and most usual method in the aromatic series is one which is not available 
 with the fatty compounds : it consists in treating a diazo-compound with 
 cuprous cyanide : — 
 
 0-N2-CN = 1^2+ 0-CN. 
 
 Froperties 
 
 The lower nitriles are colourless volatile liquids of a not unpleasant smell, 
 which are somewhat soluble in water. They are readily saponified by acids or 
 alkalies to form ammonia and the corresponding acid, a reaction equally im- 
 portant from the theoretical and from the practical point of view. In the 
 aromatic series a somewhat high temperature is required to bring about the 
 change. If the saponification is carried out with an alcoholic solution of hydro- 
 chloric or sulphuric acid, the ester is produced. 
 
 Kaufler ^ has determined the velocity of hydrolysis of 2,7-dicyan-naphthalene, 
 as an example of the occurrence of two successive similar reactions. He deduces 
 an expression for the amount of change, assuming each reaction to be mono- 
 molecular. The dicyanide was heated in amyl alcohol solution with excess of 
 alcoholic potash, and a stream of air passed through ; the rate of reaction was 
 
 1 Z. Ph.. Ch. 5S. 502 (1906). 
 
Prussic Acid: Properties 205 
 
 measured by the amount of ammonia evolved. The results agree fairly well 
 with the theory, the ratio of the two reaction constants being about 6:1. 
 
 In this saponification the triple link between the carbon and the nitrogen 
 is completely broken ; but there are many reactions in which it is only partially 
 destroyed. Thus a nitrile can take up one molecule of water to form an amide. 
 This may be brought about by heating it with water to 180°, or with alkaline 
 hydrogen peroxide to 40°. In the latter case, as has been pointed out, oxygen 
 is evolved, and the reaction appears to be due to nascent water : — 
 
 E-CN + H2O2 = ECONH2 + 0. 
 There are many other analogous reactions. If heated with acids the nitriles 
 form secondary amides, and with acid anhydrides, tertiary amides : — 
 
 .CO.CH3 _ CH .C0.N^<^^-^^3 
 
 CHg-CN + o<^g;^g3 = ch^-conO 
 
 Similarly with hydrogen sulphide they give thioamides. With hydrochloric 
 acid they form imino-chlorides, or in alcoholic solution imino-ethers : — 
 
 CH3-C=N + HCl = CH3C<^^jj. 
 
 With hydrobromic acid a different reaction occurs ; two molecules are taken up 
 instead of one, and an amido-bromide is formed : — 
 
 CH3C=N + 2 HBr = CHo-C<^^ . 
 
 Bromine first substitutes in the alkyl group, and then the hydrobromic acid so 
 produced adds on to form an imino-bromide : — 
 
 CaHgC^N + Bra = CaH^BrCHN + HBr = C2H4Br-C<5^- 
 
 Nascent hydrogen converts the nitriles into amines (Mendius reaction). 
 Hydroxylamine converts them into amidoximes : — 
 
 CH3C=N + H2NOH = CH3C<^g^. 
 This reaction goes peculiarly easily when there are halogen atoms in the 
 nitrile ^ : thus dichlor-acetonitrile forms CHCla-C^-xTTT . 
 
 All these reactions are typical of an unsaturated body. They are additive 
 reactions, and, as in all additive reactions, the molecule which is added on 
 divides into two or more parts, which then satisfy one or two of the three links 
 between the carbon and the nitrogen. The position which these groups take 
 up is governed by the respective affinities of the carbon and the nitrogen. 
 Oxygen always tends to attach itself to carbon, and hydrogen to nitrogen. 
 
 Another property which the nitriles share with all unsaturated bodies is the 
 tendency to polymerize — that is, to form addition-products with themselves. 
 These polymers can be formed by the combination of either two or three 
 molecules of the same or of different nitriles. The bimolecular compounds are 
 indifferent substances ; the trimolecular are strongly basic. They are all formed 
 by the action of sodium on the nitriles under different conditions. In ethereal 
 
 1 Steinkopf, Bohrmann, Ber. 40. 1633 (1907). 
 
206 Cyanogen Compounds 
 
 solution the bimolecular polymers are obtained ; but on treatment with sodium 
 in the absence of a solvent, or on heating with dry sodium ethylate to 140°, the 
 cyanides of primary alcoholic radicals produce the basic trimolecular polymers 
 known as cyanalkineS. 
 
 The first stage in the formation of these bodies appears to consist in the 
 formation of sodium derivatives of the nitriles. These have only been isolated 
 in the case of certain aromatic compounds, such as benzyl cyanide (phenyl- 
 acetonitrile), ^-CHg-CN, where the negative influence of the CN group is 
 strengthened by that of the phenyl. But there is evidence of their existence in 
 the fatty series : as in the production of butyronitrile by the action of sodium 
 on a mixture of acetonitrile and ethyl iodide : — 
 
 CHaNa-CN + C2H5I = C^Hj-CHa-CN + Nal. 
 
 In the production of a double polymer from acetonitrile by treatment with 
 sodium, marsh gas is evolved and sodium cyanide formed. This indicates that 
 one sodium atom turns out a methyl group from one molecule of acetonitrile, 
 forming sodium cyanide, while another displaces a hydrogen from a second 
 molecule to give CHgNa-CN, this hydrogen combining with the methyl to give 
 methane : — 
 
 2 Na + 2 CH3-CN = NaCN + CH^ + CHaNa-CN. 
 
 The sodium derivative now combines with another molecule of the nitrile thus : — 
 
 CH3C=N _ CHgC^N-Na 
 
 + CHaNa-CN ~ CHg-CN ' 
 
 and this on treatment with water yields the hydrogen compound T > 
 
 CIi2*CN 
 
 which is the imino-nitrile of acetoacetic acid. That this is the structure of 
 the polymer is shown by its hydrolysing with cold acid to the ketone-nitrile, 
 CH3-CO-CH2-CN, and with hot acid (through acetoacetic acid) to carbon dioxide 
 and acetone.^ 
 
 The formation of the trimolecular derivatives or cyanalkines is not under- 
 stood ; but their constitution has been made out, and they are found to be 
 amino-pyrimidine derivatives. Thus cyanethine (from propionitrile) is amino- 
 methyl-diethyl-pyrimidine 
 
 Q2H5 
 
 C 
 
 CHg-C^N 
 II I 
 H2NC C-CoHg 
 
 N 
 
 It is remarkable that while the alkyl cyanides give these pyrimidine deriva- 
 tives, the triple polymer which benzonitrile forms under the same conditions 
 (cyaphenine) has quite a different constitution, being a cyaauric compound with 
 the tricyanogen ring : — 
 
 1 Benzyl cyanide behaves in the same way. Atkinson, J. F. Thorpe, J. C. S. 1906. 1906. 
 
Polymerization of Prussic Acid 207 
 
 N 
 
 <p-(/\<t> 
 
 N N , 
 \/ 
 
 while benzoyl cyanide forms a triple polymer of a different type again, whose 
 formula is probably ' :— 
 
 ^•CON=C— C=:NCO-0 
 
 N=C-CO-0 ' 
 as well as a double polymer which seems to have the structure ^ : — 
 
 (j) C CN 
 
 — C — NO 
 
 The formation of polymers of the cyanalkine type is obviously impossible in 
 the last two cases, since there is no hydrogen attached to the carbon carrying 
 the CN group. 
 
 ISOCYANIDES, ISONITEILES, OE CAEBYLAMINES 
 
 The isoeyanides were first prepared by Gautier in 1866 by the action of 
 silver cyanide on the atkyl iodides. Shortly afterwards Hofmann obtained them 
 by the action of chloroform and alcoholic potash on primary amines : — 
 
 CHg-NHa + CHCI3 = CH3-N=C + 3 HCl. 
 These two reactions are still the most important for their preparation. They are 
 also got as by-products in the preparation of the nitriles. 
 
 They are volatile liquids with a very powerful and extraordinarily repulsive 
 smell. They are not basic, but nevertheless combine with hydrochloric acid in 
 ethereal solution to form compounds such as 2CH3NC, 3HC1 : these, however, 
 are not stable to water, as aqueous hydrochloric acid readily saponifies the 
 isoeyanides to form an amine and formic acid : — 
 
 Et-N=C + 2 HgO = Et-NHg + HCOOH. 
 This is the most typical reaction of the isoeyanides, and proves their constitu- 
 tion. It is to be noticed that whereas the nitriles are saponified with equal 
 ease by acids and alkalies, but only on warming, the isoeyanides are stable 
 towards alkali, but are saponified with great readiness — almost explosively in 
 some cases — by acids in the cold. 
 
 As the nitriles by taking up one molecule of water are converted into the 
 amides of the corresponding acids, so the isoeyanides take up a molecule of 
 water to give the corresponding alkyl-formamides : — 
 
 Et-N=C + H2O = Et-NH-C<2. 
 
 This change may be brought about by means of glacial acetic acid. The 
 
 ' Diels, Stein, Ber. 40. 1655 (1907). " Diels, Pillow, Ser. 41. 1893 (1908). 
 
208 Cyanogen Compounds 
 
 necessary water is supplied by the acid, which is thereby converted into its 
 anhydride. 
 
 On reduction the isocyanides yield secondary amines, one of the groups being 
 of course methyl : — 
 
 0.N=C -^ 0-NH-CH3. 
 
 The hydrolysis of the isocyanides to an amine and formic acid proves that 
 the CN group is linked to the hydrocarbon radical by the nitrogen and not by 
 the carbon. But this grouping may be written in two ways, either as E-N=0 
 or as E-N=C. In the first ease the carbon is represented as tetrad, in the 
 second as dyad. The first formula was that originally adopted, but subsequent 
 investigations, especially those of Nef, have shown that the second, with the 
 nitrogen doubly linked to dyad carbon, is much more probable, and in fact 
 practically certain. 
 
 To begin with, there are great difficulties in giving a physical interpretation 
 to the first formula. If we hold, as we must, that the carbon atom is of the 
 form of a tetrahedron, we should have to suppose that the nitrogen atom is of 
 such a shape that it is able to surroimd it entirely, touching all the four points 
 at once, and this does not seem probable ; moreover, apart from these compounds, 
 there is no established case in the whole of chemistry of a quadruple bond. 
 Further, the behaviour of the isocyanides strongly supports the dyad carbon 
 structure. A body of the type E-N=C,in forming an addition-compound, must, 
 like a nitrile, break one of the bonds between the carbon and the nitrogen, so 
 
 that the product would have the structure ^ . A body of the type E-N=C 
 
 E-N— C'-Y 
 might indeed behave in the same way, giving a product ^ ; but as the 
 
 dyad condition for carbon is very abnormal, we should rather expect that the 
 carbon would first assume its normal valency of four by attaching both of the 
 
 two added groups to itself, forming E-N=:C/-jt^ : a reaction which could not take 
 
 place with a derivative of tetravalent carbon, E-N=C. Thus the difference 
 
 between the two alternative formulae for the isocyanides is not merely formal, 
 
 but is real, and conditions a difference in behaviour ; and this enables us to 
 
 decide which is correct. Nef has shown that in the addition-reactions of the 
 
 isocyanides the primary product possesses, certainly in many cases, and probably 
 
 in all, the second of the two possible constitutions, having both of the new 
 
 groups attached to the carbon. For example, the isocyanides reduce mercuric 
 
 oxide to metallic mercury, being converted into isocyanates : — 
 
 Et-N=:C + HgO = EtN=C=0 + Hg. 
 
 Again, phenyl isocyanide in chloroform solution adds on two atoms of chlorine 
 
 /CI 
 to the carbon, forming isocyanphenyl chloride, (p-'N=G\p-, . In the same way, 
 
 when the isocyanides are treated with phenyl magnesium bromide, addition 
 
 takes place to the carbon only, giving a compound Et-N=C<^?L -p . the struc- 
 
 ' Ann. 270. 267 (1892) ; 287. 265 (1895). 
 
Isocyanides : Constitution 209 
 
 ture of which is proved by its giving with minerals first an imino-com- 
 pound Et-N=CH-0, and then benzaldehyde. It is thus evident that whereas 
 a nitrile adds on both to the carbon and to the nitrogen, forming compounds 
 
 of the type • , an isocyanide does not form the analogous bodies 
 
 R'N— C-Y X 
 
 ^ , but adds on to the carbon only, giving E-N=0\y • This can only be 
 
 explained if the carbon in the isocyanides is dyad, and in the formation of the 
 addition-compounds assumes its higher valency of four. 
 
 Constitution of Prussic Acid and its salts 
 
 This is a question of great complexity, on which every possible theory has 
 been advanced ; and it cannot even now be said that any one of them is estab- 
 lished. Prom the relation between prussic acid and cyanogen, in which it can 
 be proved that the two ON groups are united through carbon, and also from the 
 fact that the nitriles were the first class of organic derivatives to be discovered, 
 it was for a long time assumed that free prussic acid possessed the formula 
 H-C=N. This view became more firmly established when Gautier, the discoverer 
 of the isocyanides, published in 1869 an elaborate comparison of the reactions 
 of prussic acid with those of the nitriles on the one hand and the isocyanides on 
 the other, from which he concluded that the nitriles were the true analogues of 
 prussic acid. Of recent years this view has been challenged, especially by Nef, 
 who brought forward much evidence in favour of the opposite theory that 
 prussic acid is H-N— C, and its salts M-N=C. Nefs theory has since been 
 attacked by Wade,^ who thinks that the acid is H-C=N, while he agrees with 
 Nef that the salts are M-N=C. Others maintain that the alkaline salts are 
 nitriles, M-C=N, but that the salts of the heavy metals, such as silver and 
 mercury, are isocyanides, M-N^C. 
 
 Ostwald points out, in favour of the isocyanide structure for the free acid, 
 that the influence of the true nitrile group, -ON, in raising the acidity of an 
 organic acid is enormous, as is shown by the following series of dissociation 
 constants : — 
 
 Acetic acid K = 0-0018 
 
 Monochloracetic 
 
 Monobromacetic 
 
 Cyanacetic 
 
 Thiocyanacetic 
 
 0-155 
 0-138 
 0-370 
 0-265 
 
 Thus the negative influence of cyanogen is more than twice as great as that of 
 chlorine or bromine ; in fact it is among the most strongly acidic groups known. 
 We should therefore expect a compound H-C-N to be among the strongest acids, 
 as is thiocyanic acid, which on chemical grounds we know to be H-S-C=N. But 
 as a fact prussic acid is among the weakest. It therefore cannot be H-C-N, but 
 must be H-N-C. 
 
 To this argument Wade ' answers that the substitution of the group CCI3 for 
 
 1 J. C. S. 1902. 1596. 2 1. c. 1616. 
 
210 Cyanogen Compounds 
 
 hydrogen enormously raises the strength of an acid, as is shown by the com 
 parison of formic with trichloracetic acid : — 
 
 Formic acid K= 0-0214 
 
 Trichloracetic .... 120-0 
 
 Hence, if Ostwald's analogy holds, chloroform ought to be an acid of immense 
 strength ; whereas it is not an acid at all. This objection is certainly difficult 
 to answer. 
 
 Of the other arguments which have been advanced, some may be dismissed 
 at once : such as the formation of prussic acid (as a nitrile) by heating form- 
 amide, since a tautomeric change is obviously probable in a reaction of this 
 kind. The same may be said of Chattaway's argument for the isocyanide 
 formula from the undoubted fact that the cyanogen halides have the halogen 
 attached to nitrogen. The whole question is whether the two classes of bodies 
 are analogously constituted or not. 
 
 The problem really divides itself into two parts : the constitution of prussic 
 acid, and that of its salts. The free acid, in the liquid, dissolved, or gaseous 
 state, must, as a tautomeric substance, be a mixture of the two forms ; and 
 hence when we ask what its structure is, we are really asking which of the two 
 predominates. If their—praportions are not very unequal, we cannot expect 
 a definite answer. But the salts in the solid state, or at least any given salt, 
 must have one form only ; and even in solution we should expect to find that 
 one form greatly predominated. The facts agree with this. The arguments as 
 to the structure of the metallic cyanides point fairly definitely to the isocyanide 
 formula; but with regard to the free acid there is very much more doubt. 
 Wade points out that the hydrolysis of a prussic acid resembles that of a nitrile 
 and not that of an isocyanide ; it is effected only slowly by strong acids, which 
 hydrolyse an isocyanide almost explosively, but it is also effected by boiling 
 alkalies, which have no action on an isocyanide. Again, aU organic isocyanides 
 dissolve silver cyanide at the ordinary temperature, while nitrites do not. The 
 alkaline cyanides have this power, but not prussic acid, which thus appears 
 as a nitrile. Further, prussic acid forms additive compounds with stannous 
 chloride, antimony pentachloride, and cuprous chloride, herein resembling the 
 nitriles but not the isocyanides. Its reaction with diazomethane has already 
 been referred to. Now that it has been shown that this produces methyl 
 isocyanide as well as methyl cyanide, this reaction cannot be regarded as 
 throwing any light on the question. Nef ^ finds that prussic acid is not acted 
 on by ethyl hypochlorite at — 15°, whereas ethyl isocyanide acts violently at this 
 temperature. BrilhP has shown that its molecular refraction is much lower 
 than that required for an isocyanide, and agrees with that of a nitrUe. On the 
 other hand Lemoult ' finds that its heat of combustion is exactly that calculated 
 for H-NC from the values obtained for isocyanides, and is lower than that 
 required for a nitrile. Its high dielectric constant * and its great ionizing power ^ 
 both show an analogy to the nitriles ; but the behaviour of the isocyanides in 
 these respects is unknown. 
 
 1 Ann. 287. 274 (1895). « Z. Ph. Ch. 16. 512 (1895). 
 
 ' C. R. 143. 902 (1907); 148. 1602(0. 09. ii. 272). 
 
 * Schlundt, J. Fhys. Chem. 5. 157 (1901). ^ Centnerszwer, Z. Ph. Ch. 39. 218 (1897). 
 
Constitution of Prussic Acid 211 
 
 The physiological effects of prussic acid are stated by Nef to resemble those 
 of an isocyanide ; and by Wade, with equal confidence, to be almost identical 
 with those of a nitrile, so that this evidence must be considered to be ambiguous. 
 Great stress has been laid on the smell, but all that can really be said is that 
 while the smell of the nitriles is rather pleasant, those of prussic acid and the 
 isocyanides are very repulsive ; they are not very similar. Nef shows that the 
 smell of prussic acid can scarcely be distinguished from that of fulminic acid, 
 which he has proved to be H-O-N=0'''. But if the smell of isocyanogen is so 
 powerful as it appears to be, that of prussic acid may be due only to a small 
 proportion of H-N-C, the acid being mainly HC-N. 
 
 It would seem that the balance of evidence is, on the whole, in favour of the 
 nitrile formula for free prussic acid, but it is very unsatisfactory, which is, 
 perhaps, inevitable. When we come to the salts, the ground is rather firmer. 
 Take first those of the alkalies. Nef has shown that most of their reactions can 
 only be explained on the supposition that the primary product is of the type 
 
 K-N=C<^Y • ^°^ example, with ethyl hypochlorite potassium cyanide gives ethyl 
 
 cyanimido-carbonate. This is inexplicable on the nitrile formula, but it can 
 easily be explained on the other theory by the following series of reactions : — 
 
 EtO-Cl T^^_p/OEt ^-^'f ^^_p/OEt ^ „^_p/OEt + KOH 
 
 + K-N=C = ^-^-^VCI * ^-^-^ -* HN-t\c-N + KCl • 
 
 C1-C=N-K 
 
 So, too, it is probable that potassium cyanide combines with chlorine to give 
 
 /CI 
 KN=C\pi, , as it certainly does with oxygen to form K-N=C=0 ; and there are 
 
 many other cases. 
 
 Again, the alkaline cyanides form stable double salts with many salts of the 
 heavy metals. The isocyanides do the same, while the corresponding compounds 
 of the nitriles, where they exist at all, are much less stable.^ 
 
 The cyanides of the heavy metals, and especially those of silver and mercury, 
 have in many respects a difi'erent behaviour from the alkaline salts. For 
 example, they cannot be oxidized by potassium permanganate, as those of the 
 alkalies can ; and on treatment with alkyl halides they yield mainly iso- 
 cyanides, whereas the alkaline salts give mainly nitriles. This last reaction will 
 be discussed later ; but the other differences, though they may be due to a 
 diffference in structure between the two classes of salts, are, in all probability, 
 sufficiently accounted for by the fact that the cyanides of the heavy metals are 
 scarcely ionized at all, and hence their reactions are those of the undissociated 
 compound, while those of the alkaline salts are those of their ions. We may 
 therefore conclude that the metallic cyanides all have the metal attached to 
 the nitrogen. 
 
 We have still to consider the reaction of the metallic cyanides with alkyl and 
 acyl halides.^ The alkaline cyanides with the alkyl halides give mainly nitriles, 
 
 ' Hofmann, Bugge, Ber. 40. 1772 ; Eamberg, ib. 2578 (1907). 
 
 ' Cf. GuiUemard, C. S. 143. 1158 ; 144. 141, 326 (1907) ; Am. Chim. Phys. [8] 14. 311 
 (C. 08. ii. 583). 
 
212 Cyanogen Compounds 
 
 but especially in the presence of solvent small quantities (up to 5 per cent.) of 
 isocyanides are formed. Alkyl sulphuric or sulphonic acids, or alkyl sulphates, 
 produce the same result.' Silver cyanide, and the other heavy metal cyanides, 
 so far as they have been investigated, give with alkyl hahdes practically only 
 isocyanides ^ ; but with acyl halides, such as acetyl and benzoyl chlorides (and 
 also apparently with certain substituted alkyl halides, as the chloro-methyl 
 ether Cl-CHa-OAlk"), they form nitriles. 
 
 Gautier supposed that the salts were all M-C-N, and that a nitrUe was always 
 the primary product, but that this changed over into an isocyanide (where that 
 was the ultimate product) in the course of the reaction. This view has been 
 shown to be untenable, as the change does not occur ; and so has a later view, 
 that the primary product is always an isocyanide, which in some cases changes 
 over into a nitrile : because, although this change does occur, it only does so very 
 slowly under the conditions of the experiment. 
 
 Again, it has been suggested that the potassium salt is K-C-N, and the silver 
 salt Ag-N-C. But against this we have all the evidence that the potassium salt 
 is K-N-C, and, moreover, it leaves us with the same difficulty of explaining how 
 Ag-N-C with acyl chlorides gives nitriles. 
 
 Nef, regarding all the salts as of the type M-N-C, considers that the iso- 
 cyanides are produced from silver cyanide by direct replacement : — 
 
 Ag-N=C + Alk-I = Alk-N=C + Agl, 
 
 and that where nitriles are produced, this is due to the intermediate formation 
 of addition-compounds : — 
 
 K-N=C -h Alk-I = K-N=C<^^^ = KI -f N=C-Alk. 
 
 It can be shown that the metallic cyanides do in many cases combine with the 
 halides ; we know that dyad carbon forms additive compounds of this type ; and 
 in similar cases Nef has actually prepared additive compounds containing both 
 metal and halogen. Thus his theory seems satisfactory as far as the formation 
 of nitriles is concerned. But there is evidence that additive compounds are 
 formed also in the cases where an isocyanide is the ultimate product. Dry 
 silver cyanide absorbs methyl iodide at its boiling-point, giving a viscous liquid 
 which solidifies on cooling, and evolves methyl isocyanide on further heating. 
 What structure is to be attributed to these compounds? Wade, who accepts 
 Nef s view as to the formation of nitriles, suggests that the intermediate com- 
 pounds produced in those cases in which an isocyanide is obtained have the 
 alkyl and the halogen attached to the nitrogen thus : — 
 
 Ag-N=C -f CHs-I = Ag-N=C = ^^^ "^ 
 
 CH3 CH3-N=C. 
 
 The objection to this view is that there is no evidence for the structure of these 
 additive compounds, beyond the fact that their supposed analogues, the com- 
 
 1 Auger, C. R. 145. 1287 (1908). 
 
 ' The nitrile is, however, formed in inereasing quantity as the alkyl group geta larger. 
 
 s Sommelet, C. Ji. 143. 827 (1907). 
 
Constitution of Prussia Add 213 
 
 pounds of the alkyl halides with the isocyanic esters, yield on hydrolysis vei-y 
 small quantities of secondary amines ; and this is not a proof of the structure of 
 these compounds, much less of those obtained from the metallic cyanides. 
 Moreover, the structure assigned to them is intrinsically improbable ; for we 
 must suppose that the halogen attaches itself to the nitrogen, for which it has 
 very little afBnity, rather than to the excessively active dyad carbon atom. 
 
 It seems more natural to suppose that in both eases— whether the final 
 product is a nitrile or an isocyanide — the addition takes place to the dyad 
 carbon. And it is not difficult to see how this might lead to the formation of an 
 ester of either type.^ On this hypothesis, the body formed by the addition 
 of any metallic cyanide to any alkyl or acyl halide will have the structure 
 
 M-N=C^T > or, as it may also be written, II • This latter formulation 
 
 ^I M-N 
 
 shows that there is a formal analogy between the structure of the compound and 
 
 that of the unsymmetrical oximes. It has a carbon atom linked to two different 
 
 groups, and doubly linked to trivalent nitrogen. It obviously can occur in two 
 
 stereoisomeric forms : — 
 
 Alk-C-I AlkC-I 
 
 II and II . 
 
 M-N N-M 
 
 In the second of these, which may be called the synhaloid form, the metal 
 and the iodine, being near to one another, will split off easily as metallic iodide, 
 leaving a nitrile, just as a synaldoxime loses water : — 
 
 Alk-CI Alk-C AlkCH AlkC 
 
 II = III + MI : 11 = 411 + H„0. 
 
 N-M N NOH N ' 
 
 This is essentially Nef's formulation of the reaction of potassium cyanide, the 
 formation of nitrUes. On the other hand the anti-compound cannot react in this 
 way, because the metal and the halogen are on opposite sides of the C=N 
 group. It therefore undergoes the Beckmann reaction, which brings the metal 
 and the halogen nearer together, so that they can split off, giving an isocyanide : — 
 
 Alk-C-I M-CI C 
 
 II -> II -^ W + MI. 
 
 M-N Alk-N Alk-N 
 
 This is analogous to the formation of an isocyanate from the potassium salt of an 
 amide in the Hofmann reaction : — 
 
 Alk-C-OK Br-C-OK C=0 
 
 II -♦ II _> II + KBr. 
 
 Br-N Alk-N Alk-N 
 
 On this hypothesis everything depends on which stereoisomer is produced ; the 
 synhaloid gives a nitrile, and the antihaloid an isocyanide. We must there- 
 fore suppose that with an alkyl iodide and potassium cyanide, or with an alkyl 
 acyl halide and sUver cyanide, the synhaloid form is obtained ; while the anti- 
 haloid is formed from an alkyl halide and silver cyanide : — 
 
 Alk-C-I AcC-Cl Alk-CI 
 
 II : II : II ^ 
 
 N-K N-Ag Ag-N ' 
 
 1 Sidgwick, Pi-oe. O. S. 21. 120 (1905). 
 1175 P 
 
214 Cyanogen Compounds 
 
 CYANIC ACID 
 
 Cyanic acid, HCNO, like hydrocyanic acid, gives two series of derivatives ; 
 and it may therefore have either or hoth of the formulae H-0-C=N and 
 H-N=C=:0. It is by no means certain which of these should be adopted, but 
 the imide formula, M-N=C=0, is at any rate the more probable as far as the salts 
 are concerned. This is most nearly related to the formula of the metallic cyanides, 
 M-N=:C", and it explains the ease with which the latter are converted into 
 cyanates, the dyad carbon atom taking up an atom of oxygen to become tetrad. 
 
 Cyanic acid is prepared from its polymer cyanuric acid C3N3O3H3, which is 
 obtained by the action of water on cyanuric bromide, CgNgBrg, or by heating 
 urea* : — 
 
 3CO(NH2)2 = C3N3O3H3 + 3NH3. 
 
 The cyanuric acid is heated in a stream of carbon dioxide, and the vapours of 
 cyanic acid which are evolved are condensed to a liquid in a freezing mixture. 
 Cyanic acid is a very volatile liquid, giving an acrid and corrosive vapour. It is 
 excessively unstable. In about an hour, if it is kept below 0°, it is completely 
 converted into a white porcelain-like mass : this is generally said to consist of 
 cyamelide, an insoluble polymer of cyanic acid which regenerates cyanic acid on 
 heating. It has been shown,* however, that this mass is composed of about 
 30 per cent, cyamelide and 70 per cent, cyanuric acid. If the liquid cyanic acid 
 is taken out of the freezing mixture it changes into this white mass with 
 explosive violence in the course of a few minutes. (Formaldehyde undergoes 
 a similar explosive polymerization to the triple polymer trioxymethylene.') If 
 it is dissolved in ether, or if a few drops of triethyl-phosphine, PEtg, are added to 
 it, it is converted into cyanuric acid. 
 
 In aqueous solution it is scarcely more stable, and rapidly breaks up, at any 
 temperature above 0°, into carbonic acid and ammonia. The potassium salt 
 breaks up in the same way in solution, but is quite stable in presence of a 
 slight excess of potash.* 
 
 Potassium cyanate is obtained by heating potassium cyanide or ferrocyanide 
 with an oxidizing agent such as manganese dioxide, litharge, or potassium 
 bichromate ; or by the electrolytic oxidation of potassium cyanide in aqueous 
 solution.'' It is of course on its ready oxidation to the cyanate that the use of 
 potassium cyanide as a reducing agent in inorganic chemistry depends. The 
 cyanate is also got by passing cyanogen gas or cyanogen chloride into potash 
 solution, in the first case together with potassium cyanide, in the second with 
 potassium chloride. 
 
 It is to be noticed that silver cyanate, unlike the cyanide, reacts with acyl 
 chlorides to give acyl isocyanates, and not the normal esters." Also, free cyanic 
 acid with diazomethane gives only methyl isocyanate.' 
 
 > Cf. V. "Walther, J. pr. Ch. [2] 79. 126 (C. 09. i. 841). 
 
 2 Senier, Walsh, J. C. S. 1902. 290. " Kekul^, £er. 25. 2435 (1892). 
 
 * Siepermann, 0. 07. i. 1569. " Paterno, C. 04. ii. 982. 
 
 « BUIeter, Ser. 36. 3213 (1903) ; 38. 2018, 2015 (1905). 
 
 ■' Palazzo, Carapelle, C. OB. ii. 1723. 
 
Cyanic Add 215 
 
 Cyanic acid gives rise to the following series of derivatives : — 
 
 1. The halides of cyanogen, or acid halides of cyanic acid, such as CN-Cl. 
 
 2. The amide, cyanamide, C^-^j ^ (or carbo-diimide, HN=C=NH), and its 
 derivatives. 
 
 3. The cyanic and isocyanic esters. 
 
 4. The thiocyanic and isothiocyanic esters. 
 
 All these classes of compounds can form triple polymers, which will be 
 considered later. 
 
 Cyanogen Halides 
 
 These bodies are obtained by the action of the halogens on metallic cyanides 
 or on aqueous hydrocyanic acid. On the large scale cyanogen bromide is 
 generally made by the action of sulphuric acid on a mixture of the bromide, 
 bromate, and cyanide of sodium. The velocity of the change has been investi- 
 gated by Ewan,' who stops the reaction at a given point by adding soda, and 
 determines the amount of unchanged sodium cyanide by titration with silver 
 nitrate and potassium iodide. His results indicate that the reaction on which 
 the change depends is : — 
 
 HBrOa + 5 HBr + 3 HON = 3 BrCN + 3 HBr + 8 H^O. 
 
 Cyanogen chloride is a colourless, very poisonous liquid, boiling at + 15-5° 
 and melting at — 6°. It is somewhat soluble in water. On keeping it changes 
 into cyanuric chloride C3N3CI3. Its formula was originally assumed to be that 
 of a normal cyanogen derivative, N=C-C1. But Chattaway " has shown that this 
 is not the case. When a halogen atom is attached to carbon, it yields fairly 
 stable compounds ; or if the negative character of the body loosens the bond 
 between the two (as in the acid chlorides), then on treatment vnth water the 
 chlorine is removed as hydrochloric acid, and hydroxyl takes its place : — 
 
 X-CCl + HOH = XCOH + HCl. 
 Where the halogen is joined to triad nitrogen, as in the halogen-substituted 
 amines and amides, Chattaway finds that it behaves quite differently. It can 
 easily be removed, but is replaced not by hydroxyl but by hydrogen, the chlorine 
 forming hypochlorous instead of hydrochloric acid. The typical reactions of 
 chlorine attached to trivalent nitrogen are thus reactions of hypochlorous acid, 
 that is, oxidations. Such bodies oxidize sulphurous acid to sulphuric, hydriodic 
 acid to iodine, and hydrogen sulphide to sulphur. All these reactions are given 
 quantitatively by the cyanogen halides : — 
 
 C^lSTBr + 2 HI = HCN + HBr + I^, 
 C=N-Br -I- H2SO3 H- H2O = HCN + HBr + H2SO4, 
 C=N-Br + H2S = HCN + HBr + S. 
 This is strong evidence that these halides are really isocyanogen compounds. 
 It also supports Nef's view that the isocyanides contain divalent carbon. For 
 these characteristic reactions are those of chlorine attached to triad nitrogen. 
 If the nitrogen is pentad, the chlorine behaves quite differently, as we see in 
 
 1 J. Soc. Chem. Ind. 25. 1130 (C. 07. i. 691). 
 
 2 Chattaway, Wadmore, J. 0. 8. 1902. 191. 
 
 p2 
 
216 Cpanogen Compounds 
 
 ammonium chloride. The structure C=N-C1 is therefore excluded, and we must 
 fall back on the dyad carbon. 
 
 When cyanogen chloride is treated with ammonia, it gives cyanamide, 
 H2N-C=N. This was regarded as a proof of the normal formula, from which it 
 follows by direct substitution. But it can easily be derived from the isocyanogen 
 formula by Nef s addition method : — 
 
 NH3 + C=N-C1 = ■^2g>C=N-Cl = H^NCHN + HCl. 
 
 Cyanogen iodide gives a curious reaction when its ethereal solution is treated 
 with zinc. It yields zinc cyanide, with separation of iodine. Eegarding the 
 zinc salt as an isocyanide, this is a case of direct substitution : — 
 
 C=N-I , „ C=N\,, T 
 
 C!=N-I + Zn = c^jj>Zn + I^. 
 
 The cyanogen halides, and especially the bromide, have been the subject of 
 numerous investigations. Some of the more important of their reactions are 
 described in the papers referred to below.' 
 
 Oyanamide 
 
 Cyanamide, C^j^ ^ ^ ^j. possibly, though less probably, HN=C=NH, is best 
 
 prepared by the action of freshly precipitated mercuric oxide on thio-urea, in the 
 presence of a little ammonium thiocyanate, which dissolves some of the mercuric 
 oxide as the double thiocyanate, and so renders it more active : — 
 
 It may also be obtained in many other ways, as by the action of ammonia on 
 cyanogen chloride, or by passing carbon dioxide over heated sodamide : — 
 
 2 NaNHa + CO2 = ISTHaCN + 2 NaOH. 
 
 The cyanamide derivatives have recently assumed a considerable practical 
 importance as affording a method for obtaining atmospheric nitrogen in a 
 combined form for agricultural purposes.'' In 1904 Caro and Prank showed 
 that when calcium carbide ' is heated with nitrogen in the electric furnace, carbon 
 separates and calcium cyanamide is formed : — 
 
 CaCa + N2 = CaN-CN + C. 
 The crude product, which contains about 20 per cent, of nitrogen, is known as 
 Kalkstickstoff, and is used as a manure. It is hydrolysed in the soil to form 
 cyanamide, urea, and ammonia. It has to be used with care, as some of the 
 earlier products of hydrolysis, probably the cyanamide, have a poisonous action ; 
 
 With cyclic amines : v. Braun, Ber. 33. 1483 (1900) ; 40. 3914 (1907); with tertiary amines : 
 V. Braun, Ber. 36. 1196 (1903) ; 40. 8933 (1907) ; with hydrazine : Pellizari, Cantoni, Gaz. 35. 
 i. 291 (C. 05. ii. 122) ; Pellizari, Koncagliolo, ib. 3». i. 434 (C. 07. ii. 685) ; Pellizari, ib. 37. i. 
 611 (C. 07. ii. 801) ; with hydroxylamine : Wieland, Ber. 38. 1445 (1905). 
 
 » Kfihling, Ber. 40. 810 (1907) ; P. F. Frankland, J. Soc. Chem. Ind. 26. 175 (1907). 
 
 ' Barium carbide can also be used. Kuhling, Berkold, Z. angew. Ch. 22. 193 (C. 09. i. 629). 
 
Cyanamide 217 
 
 but this clanger can be avoided by putting on the manure some time before the 
 crop is sown.^ 
 
 Cyanamide forms a colourless deliquescent crystalline mass, melting at 40°, 
 and is easily soluble in water, alcohol, and ether. Its hydrogen atoms are 
 readily replaced by metals. If treated with an ammoniacal silver solution, it 
 gives a yellow precipitate of silver cyanamide, CN^Aga, which on treatment 
 with ethyl iodide forms an ester whose formula is N=C-NEt2. Like all 
 cyanogen derivatives, cyanamide readily forms addition-compounds. It poly- 
 merizes spontaneously to the so-called dicyandiamide,^ which is probably 
 
 cyanguanidine, C— N-CN. When treated in ethereal solution with hydrochloric 
 
 \NH2 
 acid it gives a compound CN2H2-2HC1. It also undergoes the typical cyanide 
 reactions. It takes up water to form the amide, urea, it forms with hydrogen 
 sulphide the thioamide, thio-urea, and with ammonia the amidine, guanidine : — 
 
 N^C-NHa + NH3 = HN=C<^^2. 
 
 The mono-alkyl cyanamides are obtained from cyanogen chloride and 
 
 primary amines : — 
 
 ^•NHa + Cl-CN = ^-NH-CN or ^•N=C=NH : 
 
 or by the action of mercuric oxide on the mono-alkyl thio-ureas : — 
 
 a_p,/NHCH3 „« n/NH-CHs n^N-CH, 
 
 ®-^\NH2 H^S = C4jj 3 or C<j^jj \ 
 
 There is the same doubt about their formulae as about that of cyanamide itself. 
 
 Among the di-substitution products, on the other hand, the derivatives 
 of both forms are known. The true cyanamide derivatives, N=C-NR2> are 
 formed, as has been mentioned, from silver cyanamide and alkyl iodide, and 
 their constitution is shown by the fact that on decomposition they yield dialkyl- 
 amines. They are also formed from the dialkyl chloramines and potassium 
 cyanide : — 
 
 EtgN-Cl + KNC = KCl + EtaN-CSN, 
 
 which is a further proof of their structure. 
 
 The derivatives of the other form have already been discussed as carbo- 
 diimide compounds. They are formed by the action of mercuric oxide on certain 
 symmetrical di-substituted thio-ureas : — 
 
 «=<?:§!; = <c:i: ^ ^^^- 
 
 Esters of Cyanic Acid 
 
 The normal esters of cyanic acid,* E'0-C=N, are said to have been obtained 
 by the action of cyanogen chloride on sodium ethylate. But later work has 
 shown that this is a mistake, and as a fact they have not yet been isolated by 
 
 1 PoUacoi, Z. EleUrochem. 14. 565 ; Ulpiani, Gaz. 38. ii. 358 (C. 08. ii. 1342, 1627) ; Lohnis, 
 Moll, C. 09. i. 310; Kappen, ib. ; Loew, ib. 785. 
 
 2 Cf. Pohl, /. pr. Oh. [2] 11. 633 (C. 08. ii. 151). 
 
 3 Cf. Nef, Ann. 287. 310 (1895). 
 
218 Cyanogen Compounds 
 
 any means whatever. It is, however, clear that they must be the primary 
 product of this reaction, since the substance actually obtained is the normal 
 
 cyanuric ester : — 
 
 N 
 
 EtO-C^C-OEt 
 
 N N 
 
 C-OEt 
 It would seem, therefore, that the normal cyanic esters have in such a high 
 degree the tendency to polymerization which characterizes the whole group, 
 that they condense to cyanurates as soon as they are formed. 
 
 On the other hand, the isocyanic esters, E-N=:C=0, are well known. They 
 were first prepared by Wurtz in 1864 by distilling the salts of the alkyl sulphuric 
 acids with potassium cyanate, and led him to the discovery of the amines. A 
 better method is to act with alkyl iodide on silver cyanate. 
 
 They are also got by oxidizing the isocyanides with mercuric oxide ; and by 
 the action of bromine on amides, being intermediate products in the Hofmann 
 reaction for preparing amines. They may be obtained by the action of carbonyl 
 chloride on amines. If carbonyl chloride is passed over the heated amine hydro- 
 chloride, a carbamic chloride is formed, which on treatment with lime loses 
 hydrochloric acid to give an isocyanate : — 
 
 ClCOCl + HaNEt = ClCONHEt + HCl 
 = 0=C=]SrEt + 2 HCl. 
 
 Another method of making them is by a modification of the Hofmann 
 mustard oil reaction, using carbon oxysulphide instead of carbon bisulphide.' 
 This reacts with a primary amine to give a thiocarbamic acid, which, on treat- 
 ment with mercuric oxide, loses hydrogen sulphide and yields the isocyanate : — 
 
 ^SH 
 COS + EtNHa = C=0 = H^S + 0=C=NEt. 
 
 ^NH-Et 
 
 The isocyanic esters are volatile liquids of a powerful, unpleasant smell, 
 which, on keeping, change rather rapidly into the polymeric isocyanuric esters. 
 They are, of course, derivatives of the imide of carbonic acid ; compare : — 
 
 /OH /NHa ^,^ 
 
 C=0 : C=0 : C=ONH. 
 
 \0H \nH2 ^-' 
 
 p,.C<gjj CH,.C<g CH,-C<0 
 
 : : I NH. 
 
 CH,.C<gH CH,.C<gH2 CH,-C<o 
 
 Dibasic Acid. Diamide. Imide. 
 
 The great strain in the molecule causes them to react readily with the breaking 
 of a double link. Now there are in the molecule two double links, one between 
 the carbon and the nitrogen, and the other between the carbon and the oxygen. 
 Of these two the latter is comparatively stable. The carbonyl compounds — 
 
 1 Anschutz, Ann. 359. 202 (1908). 
 
Esters of Cyanic Add 219 
 
 aldehydes, ketones, acids — though they can behave as unsaturated bodies, are 
 yet capable in a great number of reactions of maintaining the carbonyl group 
 intact. On the other hand the imide group, ^0=NH, is an extremely unstable 
 one. Accordingly we find that in the reactions of the isocyanates it is always 
 the imide group that is attacked, and in consequence these reactions are very 
 similar to those of the isocyanides, in which the double link between the carbon 
 and the nitrogen is almost always broken, while the carbon becomes tetrad ; and 
 in this case also it is nearly always a hydrogen atom which goes to the nitrogen, 
 while the residue of the added molecule attaches itself to the carbon. 
 
 The most important reaction of the isocyanates is one in which the carbon 
 and nitrogen are completely separated. If they are warmed with alkali they 
 split up into carbon dioxide and an amine, which proves their constitution : — 
 
 ^ CH3.N=C=0 _ cHg-NHa + 0=0=0. 
 
 But often this link is not completely broken ; for example, with alcohols they 
 give urethanes : — 
 
 0=C=N-CH3 + C^Hj-OH = O^C<^f^^K 
 With ammonia or primary or secondary amines they give substituted ureas : — 
 0=C=NCH3 + HNH-C^Hg = 0=C<^g:^^^. 
 
 By a combination of the first and last of these reactions, they give, when boiled 
 with water alone, symmetrical di-substituted ureas. One molecule breaks up 
 into carbon dioxide and an amine, and then the latter reacts with a second 
 molecule of isocyanate to form the urea. They combine directly with hydro- 
 chloric acid to form substituted carbamic chlorides : — 
 
 0=C=NGH3 + HCl = 0=C<^j^'-'^3. 
 
 This reaction is to be contrasted with that of an isocyanide with hydrochloric 
 acid, which is due to the dyad carbon : — 
 
 KN=G + HCl = KN=C<^^. 
 
 Phenyl isocyanate, 0-N=C=O, is a substance which' has been much used for 
 determining questions of constitution. Although it may be obtained by any of 
 the ordinary reactions for preparing isocyanates, and also by one of the 
 Sandmeyer reactions, by treating diazobenzene with potassiiun cyanate and 
 copper powder, it is unfortunately difficult to prepare in any quantity, as the 
 yield by any of these methods is small. 
 
 The use of phenyl isocyanate as a reagent for determining constitution is 
 due to Goldschmidt,^ who first applied it in the case of succinosuccinic ester and 
 analogous compounds, and shortly afterwards to the more important case of the 
 oximes. It enables us to detect the presence of hydroxyl and imide or amide 
 groups, as with hydroxyl it gives urethanes : — 
 
 -^. /NH-0 
 
 X-COH + Cl^^ = C=0 , 
 " NO-CX 
 
 1 Ber. 23. 2179 (1890) ; Dieckmann, Ber. 33. 2002 (1900), 
 
the two isomeric benzaldoximes had the formulae Ji and ' \ / 
 
 220 Cyanogen Compounds 
 
 while with amines or imines it gives urea derivatives : — 
 
 x-c=]srH + cC^ = 0=0 
 
 '^ \n=c-x 
 
 It is especially useful from the fact that it can be made to react at a low 
 temperature and in the absence of a solvent. There is thus less danger that the 
 reagent itself will bring about tautomeric change, which is particularly liable to 
 happen where a solvent of some ionizing power, such as alcohol, has to be used. 
 Also it breaks up the substance with which it reacts, and adds on the two parts 
 to itself in such a way that they can be easily identified. 
 
 The cases where it has proved most useful are those of the oximes, and of 
 bodies with keto-enolic tautomerism. In the oximes the question was whether 
 
 d)-C-H d>CHNH 
 
 f II ___. r , , 
 
 N-OH 
 
 (of which the first would give a urethane and the second a urea), or whether 
 they both had the first formula ; and Goldschmidt showed that in both cases 
 a urethane was formed. 
 
 In the case of the keto-enolic esters, such as the /3-keto-esters, the problem 
 is a difierent one. We have here to decide between the groupings -CH=C-OH 
 and -CH2-C=0. The former gives a urethane, while the latter does not react 
 at all. It has, however, recently been shown ' that this test must be used with 
 caution, as some ketonic compounds are able to react with phenyl isocyanate, 
 changing over into their isomers in the course of the reaction. But Goldschmidt ^ 
 finds that this danger is to a great extent avoided if care is taken to exclude 
 traces of alkali, which have a great power of promoting the tautomeric change, 
 and will often enable the isocyanate to react with the ketonic form when it 
 cannot do so in their absence. * Even then, however, this reagent is not always 
 trustworthy. There are cases of hydroxyl compounds (especially those with 
 negative substituents) with which it will not react ; while there are also cases 
 of undoubted ketones with which it combines by direct addition.' 
 
 Tidocycmic Acid and its Derivatives 
 
 The salts of thiocyanic acid are obtained from the cyanides by direct 
 combination with sulphur, just as those of cyanic acid are got by direct combina- 
 tion with oxygen. But whereas it has been shown that the metallic cyanates 
 have the formula M-N=C=0, that is, are really isocyanates, free thiocyanic acid 
 and its salts are derived from the normal formula H-S-CSN, M-S-C=N. Isothio- 
 cyanic acid, H-N=C=S, and its salts are unknown. This is in accordance vrith the 
 general tendency of the C=S group to pass into C-S-H, which is far greater than 
 that of 0=0 to go over into 0-OH, as we have already had occasion to notice in 
 comparing the amides with the thioamides and the ureas with the thio-ureas. 
 
 Free thiocyanic acid is generally described as a liquid. But this is because 
 it has never until recently been obtained in the pure state. The pure acid = can 
 
 ' Michael, £er. 38. 22 (1905). 2 Ber. 38. 1096 (1905). 
 
 ^ Cf. Dieokmann, Hoppe, Stein, Ber. 37. 4627 (1904). 
 
 * Michael, Cobb, Ann. 383. 64 (1908). ^ Koaenheim, Levy, Ber. 40. 2166 (1907). 
 
Thiocyanic Compounds 221 
 
 be prepared by treating a mixture of dry potassium thiocyanate and phosphorus 
 pentoxide with sulphuric acid in an atmosphere of hydrogen, avoiding any rise 
 of temperature and collecting the vapours evolved in a freezing mixture. It is 
 then found to be a white crystalline mass, which is stable at 0°, but at + 5° 
 melts to a liquid which soon begins to polymerize, first turning yellow and then 
 suddenly resolidifying to a yellow amorphous mass. The molecular weight was 
 shown by the boiling-point in ethereal solution to be that required by the 
 formula HCNS. The dilute aqueous solution is stable at 0°, but on warming 
 polymers separate out. The concentrated solution breaks up into hydrocyanic 
 acid and persulphocyanic acid, H2C2N2S3. If a fairly concentrated solution is 
 treated with a large excess of sulphuric acid, it decomposes mainly into ammonia 
 and carbon oxysulphide, COS. When treated with diazomethane at — 5°, thio- 
 cyanic acid is wholly converted into methyl thiocyanate, CH3-S-C=N.^ 
 
 Ammonium thiocyanate may be obtained by the action of carbon bisulphide 
 on alcoholic ammonia : — 
 
 /NH2 T.T 
 
 CS2 + 4NH3 = q=S + 2NH3 = C^aTjTT + (NHJ^S. 
 
 On heating to 160° it gives thio-urea, but if heated for twenty hours to 180° it is 
 converted into guanidine thiocyanate, as we have already seen. , 
 
 The behaviour of potassium thiocyanate with organic halides '^ is remarkable. 
 With alkyl halides and many of their derivatives, such as chloracetic ester, it 
 gives normal thiocyanic esters, E-S-C^N. But with bodies of the type 
 Arg-CHBr, where Ar is an aromatic radical— for example, with CH^jBr — it 
 gives isothiocyanates, E-N=C=S. No explanation has been offered of this. 
 
 The brilliant red colour which is produced with ferric salts on addition of 
 potassium thiocyanate solution is due to the formation of a double thiocyanate, 
 9 KCNS-Fe(CNS)3. It may be distinguished from all other substances which 
 could possibly be mistaken for it by the fact that on shaking with ether it 
 goes over entirely into the ethereal solution. This test is very delicate. A 
 thousandth of a milligram of ferric iron produces a quite visible pink colour. 
 
 The esters of thiocyanic acid belong to two series : (1) the normal thio- 
 cyanates E-S-C=N, and (2) the isothiocyanates or mustard oils K-N=C=S. 
 
 The normal esters (S-esters) are obtained by the action of potassium thio- 
 cyanate on alkyl iodides or the salts of alkyl sulphuric acids, or from cyanogen 
 chloride and the mercaptides, which shows that the alkyl is attached to sulphur : — 
 
 (EtS)2Pb + 2 C<g^ = 2 C<|-^* -f PbCl^. 
 
 They are liquids of a leek-like smell. Their structure is also shown by their 
 being reduced by nascent hydrogen to mercaptan and prussic acid, which is then 
 further reduced to methylamine. Another proof is that when oxidized with 
 nitric acid they give alkyl sulphonic acids. With alcoholic potash they form 
 potassium thiocyanate, which the isomeric isothiocyanates do not. They have 
 
 1 Palazzo, Scelsi, Gas. 38. ). 659 (C, 08. ii. 774). 
 
 2 Wheeler, C. 02. i. 1400, ii. 577, 788. Cf. Johnson, Guest, Am. Ch. J. 41. 337 (C. 09. i. 1547), 
 who failed to observe any regularities. 
 
222 Cyanogen Compounds 
 
 a remarkable tendency to change over into the mustard oils. Methyl thio- 
 cyanate is very largely converted into the iso-compound by heating to 180° for 
 several hours, and with allyl thiocyanate the change takes place so easily that 
 if it is once distilled the conversion is complete. 
 
 Compounds containing the thiocyanate group in the a-position to carbonyl 
 or carboxyl are capable of intramolecular condensation to heterocyclic bodies. 
 Thus thiocyanacetone (from chloracetone) if heated with sodium carbonate forms 
 methyl-oxy-thiazol : — 
 
 CH=C<2§3 r!H=C^^^3 
 
 The isothiocyanic esters or mustard oils may be obtained from the primary 
 amines, by first combining them with carbon bisulphide to dithiocarbamates : — 
 
 CS2 + 2 HgN-Et = S=C<(g_-j^jj^j,j.. 
 
 This is converted by silver nitrate into the silver dithiocarbamate, which on 
 boiling with water breaks up into silver sulphide, hydrogen sulphide, and the 
 mustard oil: — 
 
 2 S=C<^.5^* = 2 C<^^* + Ag,S + H,S. 
 
 This is Hofmann's mustard oil test for primary amines, the product being 
 recognized by its smell. 
 
 They are also formed by heating the normal thiocyanates, and from the 
 isocyanates by heating them with phosphorus pentasulphide. 
 
 The aromatic mustard oils are also got from the condensation product of 
 the primary aromatic amines with carbon bisulphide. This is of course not 
 a dithiocarbamate but a thio-urea, and to convert it into the mustard oil merely 
 requires boiling with concentrated hydrochloric acid : — 
 
 S=C<^g^ + HCl = C<^^ + <^-NH3Cl. 
 
 The mustard oils are liquids of an unpleasant odour, which are insoluble in 
 water and boil undecomposed. Their reactions are almost all strictly analogous 
 to those of the isocyanates. Thus with dilute acid they are hydrolysed to the 
 amine, carbon dioxide, and hydrogen sulphide ; vnth concentrated sulphuric 
 acid they give amine and carbon oxysulphide. They have a large number of 
 additive reactions : for example, they combine with ammonia and amines to 
 form thio-ureas, with alcohols to form thio-urethanes, &c. 
 
 Nascent hydrogen converts them into amine and thioformaldehyde, which 
 appears in the form of its trimolecular polymer : — 
 
 EtN=C=S + 2H2 = Et-NHa + H-C<|-. 
 
 Isothiocyan-acetic acid,' like its normal isomer, goes over readily or even 
 spontaneously into a ring compound, dioxy-thiazol : — 
 HaC-S-C^N HaC-Sx 
 
 t = I /C=0. 
 
 0=COH 0=C-NH 
 
 1 Frerichs, Beckurts, C. 00. i. 589, 02. ii. 931 ; Hantzsoh, Vogelen, Ber. 35. 1007 (1902). 
 
Tkiocyanic Compounds 223 
 
 AUyl isothiocyanate is the body from which the name of the whole class is 
 derived. It is obtained by grinding the seeds of the black mustard in water, 
 and distilling the liquid after it has stood for a certain time. The seeds contain 
 a glucoside which is of a rather unusual kind, since it is a potassium salt, 
 potassium myronate. This glucoside breaks up under the influence of a ferment 
 known as myrosin, which is also contained in the seeds, into allyl isothiocyanate, 
 grape sugar, and potassium hydrogen sulphate. 
 
 The bodies obtained by the interaction of potassium thiocyanate with acyl 
 halides have been shown to be isothiocyanates. The normal thiocyanates of 
 the acyl radicals do not appear to exist.' 
 
 rULMINIC ACID 
 
 Mercury fulminate was discovered by Howard in 1800. In 1823 Liebig 
 showed that the silver salt has the same composition as silver cyanate. This 
 was the first case of the discovery of two different substances having undoubtedly 
 the same composition ; and it was to designate this phenomenon that Berzelius 
 invented the word isomerism. 
 
 Mercury fulminate is formed by the action of alcohol on a solution of 
 
 mercury in excess of nitric acid. The constitution of the salt has long been 
 
 a matter of great uncertainty. It is extraordinarily reactive, and very explosive, 
 
 which made the investigation more difficult. Moreover it has not been found 
 
 possible, even up to the present time, to obtain a stable volatile derivative of 
 
 fulminic acid,'' so that its molecular weight can only be inferred from indirect 
 
 evidence. For a long time the accepted view was that of Divers and others, 
 
 that the molecule of the acid contains two atoms of carbon, so that it must be 
 
 written HgCaNgOj. This view was based on its formation from ethyl alcohol, 
 
 and on the fact that when treated with bromine or iodine it yields compounds 
 
 containing two carbon atoms in the molecule. On this assumption the most 
 
 various structures have been erected for it. Kekule suggested that it was 
 
 nitro-acetonitrile, CHaNOj-CN. This was soon abandoned, and as a matter of 
 
 fact this body has recently been prepared,' and is found to be a fairly stable oil 
 
 of rather high boiling-point. Another suggestion is that it is glyoxime 
 
 HC CH HC CH 
 
 peroxide, II ^11 , or more probably II „ 1^0 • ^ut this has also been 
 NO -ON •' N-O-N^ ' 
 
 made,* and though it resembles fulminic acid in some respects, it has been proved 
 
 to be different from it. The fact that fulminic acid, like the oximes, yields 
 
 hydroxylamine when treated with hydrochloric acid, showed that it must contain 
 
 C=:NOH 
 the group :NOH, and hence Steiner suggested the formula J|_t.t„„ , which was 
 
 generally accepted some years ago. 
 
 The overthrow of this and all other two-carbon formulae, and the final 
 
 ■ Dixon, Taylor, /. C. S. 1908. 684. 
 
 ' Biddle, Am. Gh. J. 33. 60 (C OS. i. 590), believes that be has obtained methyl fulminate, 
 CHsO'N = C, as a liquid boiling at 50-60°, with a strong isonitrile smell ; but it was too unstable 
 to admit of further inveBtigation. 
 
 5 Steinkopf, Bohrmann, Ber. 41. 1044 (1908). * Jovitschitsch, Ann. 347. 233 (1906). 
 
224 Cyanogen Compounds 
 
 establishment of our knowledge of fulminic acid on a firm basis, is due to the 
 work of Nef.' He showed in the first place that all the evidence was in 
 favour of its containing only one carbon atom in the molecule. Its decom- 
 position-products are nearly always one-carbon bodies. It is formed, as we 
 have seen, by treating sodium nitromethane with mercuric chloride, and if 
 
 NO 
 
 treated with nitrous acid it is converted " into methyl-nitrolic acid, H- C'^jtqjj" 
 
 If we take it as proved that there is only one carbon atom in the molecule, 
 and also consider the fact that on treatment with hydrochloric acid it gives 
 hydroxylamine, only one formula is possible, namely H-O-N^C". The real 
 reason which prevented the earlier investigators from adopting such a one- 
 carbon structure as this was their unwillingness to admit the existence of 
 a dyad carbon atom. But this objection has been removed by Nef 's other work 
 on dyad carbon, and he has been able to show that his formula is alone capable 
 of explaining the very remarkable reactions of fulminic acid ; while all subsequent 
 work on these bodies has only served to confirm his views. 
 
 The formula explains the production of the salt from mercury nitromethane 
 by ' intramolecular oxidation ' : — 
 
 CH,=N<gj^g = C=N.Ohg + H^O. (hg = iHg.) 
 
 It represents it as an oxime, and hence accounts for the formation of 
 hydroxylamine on hydrolysis. Further, it represents it as the oxime of 
 carbon monoxide, containing therefore, like carbon monoxide itself, a dyad 
 carbon atom. Now the dyad carbon atom, wherever it occurs, is always 
 extraordinarily active, as we have seen in the case of the isocyanogen com- 
 pounds. It has the strongest tendency to pass into the tetrad condition, 
 and will do so if it possibly can, even at the cost of forming very unstable 
 com^pounds. The formation of methyl-nitrolic acid is the simplest case of 
 this : — 
 
 HON=C + HNO2 = HON=C<^Q . 
 
 A similar reaction takes place with hydrochloric acid. If sodium fulminate 
 is treated with this acid at 0°, the fulminic acid first liberated forms an 
 addition-compound with the hydrochloric acid, the carbon becoming tetra- 
 valent : — 
 
 HON=C" + HCl = HON=:C<^j. 
 
 This body is the oxime of formyl chloride, 0=C\pj, and is therefore called 
 
 formyl chloride oxime. Its properties are remarkable. It forms large colourless 
 crystals, which are rapidly and completely volatile at the ordinarj^ temperature. 
 If it is allowed to stand in the cold it soon begins to decompose, and the decom- 
 position when once it has set in proceeds with almost explosive violence, and 
 with the evolution of enormous quantities of heat. In this reaction carbon 
 monoxide is evolved, and a residue left behind consisting mainly of hydroxyl- 
 amine hydrochloride. That is, the oxime dissociates into fulminic acid and 
 
 ' Ann. 280. 303 (1894). ■'■ Palazzo, C. 07. ii. 135. 
 
Fulmink Add 225 
 
 hydrochloric acid, and the former is hydrolysed by the latter in the normal 
 manner. 
 
 The structure of formyl chloride oxime is proved by its behaviour on 
 treatment with aniline in ethereal solution, when it gives a precipitate of 
 phenyl isuretin : — 
 
 The formula of the product was shown by synthesizing it from isuretin and 
 aniline : — 
 
 HONHC<Jjj + H2N0 = HO.NH-C<J^ + NH3. 
 
 If silver fulminate is suspended in water and one equivalent of hydro- 
 chloric acid added, part of the silver is precipitated as silver chloride, but 
 a clear solution remains which contains both silver and chlorine. The first 
 effect of the hydrochloric acid is to add on to the dyad carbon, forming the 
 
 XT 
 
 silver salt of the oxime AgO-N=C<^Qj. This soluble compound is partly decom- 
 posed by more of the acid to form silver chloride and the free oxime, but not 
 wholly. This is a remarkable proof of the activity of the dyad carbon ; it holds 
 the chlorine so firmly that it cannot react with the silver. It is also a strong 
 argument in favour of Nef 's view that when a metallic cyanide reacts with an 
 alkyl halide to form a nitrile, an intermediate additive compound is formed with 
 the carbon tetrad ; for we have here an example of an additive compound of this 
 type which is capable of more or less permanent existence. 
 
 It has recently^ been discovered that fulminic acid can exist in the free 
 state, though it is too unstable to be isolated. It was shown by Scholvien ^ 
 that if a solution of potassium fulminate is treated with excess of sulphuric 
 acid, ether extracts from it a substance which with silver nitrate forms silver 
 fulminate. This he assumed to be free fulminic acid. Nef,'' on the other 
 
 hand, argued that it was formyl sulphate oxime, Tr^CtNOH. Wieland and 
 
 Hess have repeated the work, and have shown that the ethereal solution contains 
 no trace of sulphuric acid, so that Nef 's view must be wrong, and the solution 
 must really contain free fulminic acid. They also showed that if the ether is 
 distilled at 0°, the acid goes over with the vapour, and will form silver fulminate 
 in the receiver. Hence the free acid exists in the vapour as well as in the 
 solution. If the solution is allowed to stand, it rapidly polymerizes to meta- 
 fulminuric acid. 
 
 Fulminic acid is easily formed from methyl-nitrolic acid * : — 
 
 HC<^2^ = HNO2 -f^ C=NOH, 
 
 a reaction which, as we have already seen, is reversible. It is also produced in 
 three analogous reactions, from amino-methyl-nitrosolic acid " : — 
 
 H2N-C<^.^^ = H2O + N2 -1- C=NOH, 
 
 ' Wieland, Hess, Ber. 42. 1346 (1909). * / ^^_ cj^_ 32. 461 (1885). 
 
 2 Ann, 280. 315 (1894). ♦ Wieland, Ber. 40. 418 (1907). ° Wieland, Ser. 42. 820 (1909). 
 
226 Cyanogen Compounds 
 
 from the hydrochloride of form-oxy-amidoxime ' : — 
 
 H^\NH0HHC1 = NH2OHHCI + C=NOH, 
 and from the oxidation product of this amidoxime, methyl-nitrosolic acid : — 
 2 HC<^^^ = HPN=N0H + 2 C^NOH. 
 
 Nef has shown that there is an extraordinary resemblance between fulminic 
 and hydrocyanic acids, which is explained if we suppose that one is the 
 hydride and the other the hydroxide of isocyanogen : H-N=C, HO-N=C. The 
 older writers state that when a fulminate is treated with hydrochloric acid, 
 prussic acid is formed, though they were not able to isolate it. They were 
 misled by the smell, which is very like that of prussic acid. This smell is 
 not that of formyl chloride oxime (the actual product), but is evidently due 
 to traces of fulminic acid, arising from a slight dissociation of the oxime. It 
 can scarcely be distinguished from the smell of prussic acid. 
 
 But the resemblance goes further than this. The fulminates are nearly if 
 not quite as poisonous as the cyanides, and their physiological effects are 
 identical. Again, Nef has succeeded in preparing sodium ferrofulminate, 
 Na4re(ONC)5, by treating an alkaline solution of sodium fulminate with 
 ferrous sulphate. This is strictly analogous to sodium ferrocyanide, which 
 should really be written Na4Fe (NC)^. The solution of the ferrofulminate gives 
 none of the reactions of iron, and on evaporation the salt crystallizes out in 
 large yellow needles. It resembles the ferrocyanide in giving a colour with 
 ferric chloride, which in this case is purple ; and Nef states that this test is 
 very delicate. It differs, however, from the ferrocyanide in being very unstable, 
 breaking up into its constituents sodium fulminate and ferrous fulminate at 
 once on losing its water of crystallization, and refusing to form the corre- 
 sponding ferrofuhninic acid on treatment with mineral acids. Also it will not 
 form the ferrifuhninate when oxidized with bromine water. 
 
 Nef 's views on the structure of the fulminates have received strong confirma- 
 tion from the work of SchoU and his pupils '" on the syntheses effected by means 
 of mercury fulminate and aluminium chloride. They find that if benzene is 
 treated with mercury fulminate and a mixture of anhydrous and hydrated 
 aluminium chloride, it is converted into benzaldoxime. This reaction is easily 
 explained by Nef 's formula of the fulminates, but by no other. The water of 
 the hydrated aluminium chloride liberates a certain quantity of hydrochloric 
 acid, and this combines with the fulminic acid which it sets free from the 
 mercury salt to give formyl chloride oxime. This then reacts with the benzene 
 in the normal manner of the Friedel-Crafts synthesis, forming an oxime : — 
 
 HON=C<^j + CeHg = HON=C<^jj + HCl. 
 
 The yield under favourable conditions amounts to 70 per cent. The presence of 
 the hydrated aluminium chloride is necessary to the reaction. If it is not used 
 the nitrile is formed instead (up to 80 per cent.). This is not due to a dehydra- 
 
 1 Wieland, Hess, Ber. 42. 4175 (1909). 
 
 2 Ber. 32. 3492 (1899); 36. 10, 322, 648 (1903). 
 
Fulminic Add 227 
 
 tion of the oxime by the aluminium chloride, since it was shown that the oxime 
 is practically not converted into the nitrile at all by aluminium chloride. It 
 seems probable that in the absence of water the fulminic acid reacts with the 
 aluminium chloride to give cyanogen chloride, which then changes over into its 
 tautomeric form ; and this by the Friedel and Crafts reaction forms the 
 nitrile : — 
 
 C=NOH -^ C=N.C1 -^ C1-C=N + G^B.^ -* C0H5CN. 
 
 The conditions of the reaction, which must be very carefully observed, are 
 curious. With benzene the best yield is got by using a mixture of anhydrous 
 and hydrated aluminium chloride corresponding to the formula A120C14. But 
 with the homologues of benzene, which favour the formation of the nitrile, the 
 amount of hydrated salt must be increased, up to the composition AlOCl. It is 
 also better to add a certain quantity of aluminium hydrate, but this seems to 
 have only a mechanical action, preventing the solid salt from caking together. 
 A further point is that in order to get the oxime the fulminate must be added first 
 to the benzene, and then the aluminium compounds : while the reverse order 
 must be adopted if the nitrile is wanted. These syntheses can also be carried 
 out with the phenol ethers. 
 
 The analogy between fulminic and hydrocyanic acids holds here as well, as 
 is shown by Gattermann's synthesis of the aldimines from phenol ethers on 
 treatment with anhydrous prussic acid and hydrochloric acid in the presence of 
 aluminium chloride (see above, p. 199). 
 
 The one thing needful to establish Nef's formula beyond all doubt is a 
 determination of the molecular weight. If it can be shown that fulminic acid 
 contains only one carbon atom in the molecule, we have no alternative but to 
 accept Nefs view as to its structure. A direct determination of this magnitude 
 is impossible, but indirect evidence of great force has recently been adduced,^ 
 from the electrolytic behaviour of the sodium salt in aqueous solution. The 
 salt can be prepared by treating mercury fulminate in alcohol with sodium 
 amalgam, and is fairly stable. On Nef's theory it is the salt of a monobasic 
 acid NaONC ; on any other the acid is dibasic, and the salt Na2(ONC)2 . Now 
 van 't Hoff has shown that the alkaline salts of monobasic acids in fifth to tenth 
 normal solution give a value of the dissociation factor i of about 1-85, and those 
 of dibasic acids about 2-5. From this it follows that if sodium fulminate is 
 NaONC, its apparent molecular weight in solution should be 35, but if it is 
 Na2(ONC)2, it should be about 52. From the depression of the freezing-point 
 in fifth normal aqueous solution, it was found to be 34-9, agreeing with that 
 required for the monomolecular formula. Again, Ostwald found that the 
 increase in molecular conductivity in passing from N/32 *•' ^/io24 solution, for 
 the salt of a monobasic acid was from 4 to 8 units, and for that of a dibasic acid 
 about 11. The observed increase for sodium fulminate was 5 units. These 
 results make it certain that the acid is monobasic, and has the formula HONC ; 
 and hence Nef's view as to its structure must be adopted. 
 
 The reactions which take place in the ordinary method of preparing mercury 
 fulminate by the violent oxidation of ethyl alcohol by a solution of mercury in 
 
 ' L. Wohler, Ber. 38. 1351 (1905). 
 
228 Cyanogen Compounds 
 
 nitric acid are obscure and difficult to explain. It has been shown' that 
 oxides of nitrogen (fuming nitric acid) must be present ; that acetaldehyde gives 
 a much larger yield of fulminate than alcohol ; and that one-carbon bodies 
 (methyl alcohol, formaldehyde) and also 3- and 4-carbon compounds give none 
 at all. The explanation has been given by Wieland.^ He finds that methyl- 
 
 yNOH 
 nitrolic acid, CH , can be converted into fulminic acid, as indeed we should 
 
 expect, since it is the analogue of formyl chloride oxime : — 
 
 HO-N=C<^j : HON=C<gQ^. 
 
 It is evident that the alcohol is first oxidized to aldehyde (hence the latter gives 
 the best yield): this is converted by the nitrous acid present into isonitroso- 
 acetaldehyde, and this further oxidized to isonitroso-acetic acid. This last body 
 has been shovni ' to be converted by nitrous fumes (by nitration followed by loss 
 of carbon dioxide) into methyl-nitrolic acid, which reacts with mercuric nitrate 
 to give mercury fulminate : — 
 
 CH3CH2OH -^ CH3CHO -♦ HON:CHCHO -» HON:CHCOOH -> 
 H0N:C(N62)C00H -> H0-N:C(N02)H -^ HO-N:C. 
 
 The one-carbon compounds cannot give fulminate, since it is only when the 
 methyl group is united to a strongly negative grouping such as carbonyl that 
 the oxides of nitrogen are able to introduce the isonitroso-grouping into it. 
 
 The peculiar reaction of the fulminates with bromine is discussed below 
 (p. 230). 
 
 Polymers of Fulminic Acid 
 
 Fulminic acid gives rise to a series of polymerization products. If mercury 
 fulminate is boiled with potassium chloride solution, it gives the salt of ful- 
 minuric acid, as was discovered by Liebig. This acid * has been shown to be 
 
 nitrocyan-acetamide, -^^^ J^CH-CO-NHj ; but the mechanism of its formation is 
 
 obscure, and even its structure is disputed. 
 
 The other polymerization products form a continuous series. If the ethereal 
 solution of free fulminic acid, obtained by extracting the acidified solution of 
 the potassium salt with ether, is allowed to stand, a triple polymer, originally 
 known as isocyanuric acid, but now more conveniently termed metafulminuric 
 acid, is produced. This has been shown ^ to be isonitroso-isoxazolone : — 
 3 C=NOH -» rC C C -I H-C C C=NOH 
 
 rC C C -I H-0 
 
 Lnoh noh nohJ ^ 
 
 NOH 
 
 K O 
 
 Metafulminuric acid. 
 
 This polymerization brings out the resemblance of fulminic and hydrocyanic 
 
 ' L. Wohler, Theodorovits, Ber. 38. 1351 (1905). = Ber. 40. 418 (1907). 
 
 » Ponzio, Gas. 33. i. 510 (1903). 
 
 * Steiner, Ber. 9. 781 (1876) ; Conrad, Schultee, Ber. 4S. 735 ; Steinkopf, ib. 2026 (1909). 
 
 « Wieland, Hess, Ber. 42. 1346 (1909). 
 
Poly^ners of Fulminic Add 229 
 
 acids as compounds of dyad carbon. It is exactly analogous to the polymerization 
 of hydrocyanic acid in presence of alkali to amino-malonitrile :— 
 3 C=NH -► rC C C "i -* C— CH— C 
 
 The corresponding compounds of tetrad carbon, the nitriles, cyanates, iso- 
 cyanates, and the nitrile-oxides, all polymerize to form a symmetrical 6-ring, as 
 in cyamelide and cyanurio acid. 
 
 When metafulminuric acid is warmed with water, it goes over into cyau- 
 isonitroso-acethydroxamic acid, whose constitution was established by Nef ; — 
 NOH NOH 
 
 H-C-C-C=NOH -* C-C — cS^^ 
 
 II T III ^OH • 
 
 N N 
 
 This is a not unusual rearrangement of the isoxazole ring, and occurs, for example, 
 
 with isoxazole itself, which is converted by alkalies into cyan-acetaldehyde ^ : — 
 
 HC-CH=CH 
 
 II T -* NCCHa-CHO. 
 
 N O 
 
 Finally, the hydroxamic acid derivative can be converted into the so-called 
 isofulminuric acid of Ehrenberg," whose constitution is still unknown. 
 
 NITEILE-OXIDES 
 
 The nitrile-oxides, of which until recently very little was known, are deriva- 
 tives of the third isomer of cyanic acid, which probably has the structure 
 
 H-C=:iN 
 
 When silver nitrite acts on bromacetic ester, a small quantity of the so-called 
 oxalic ester nitrile-oxide is formed,' by loss of water from the primary product, 
 nitro-acetic ester : — 
 
 EtOCOCHa-NOj = EtO-COC=N -I- H2O. 
 
 The analogous benzonitrile-oxide is obtained by the action of soda on 
 benzhydroxamic chloride * : — 
 
 ^<OH = ^-^/-^H^^ = 
 
 or by the spontaneous decomposition of benzonitroUc acid ^ : — 
 
 Their reactions have been further examined by Wieland.* They polymerize, 
 with extraordinary ease, in three different ways. By themselves, or in neutral 
 solution, they form double polymers of the class previously known as glyoxime- 
 
 ' Claieen, Ber. 36. 3666 (1903). " J. pr. Ch. [2] 30. 38 (1884). 
 
 ^ Soholl, Schofer, Ber. 34. 862, 870 (1901). ' Werner, Buss, Ber. 27. 2199 (1894). 
 
 '■ Wieland, Semper, Ber. 39. 2522 (1906). « Ber. 40. 1667 (1907) ; 42. 803 (1909). 
 
 1175 Q 
 
230 Cyanogen Compounds 
 
 peroxides, which, however, have now been shown ' to be 5-ring compounds, and 
 are called furoxanes : — 
 
 0-C Cr^ 
 
 2 0-C=N _ II |>0. 
 \o/ - N N 
 ^(/ 
 Diphenyl-f uroxane . 
 
 This polymerization enables us to explain the singular action of bromine on 
 silver or mercury fulminate, whereby dibromo-furoxane (dibromo-glyoxime- 
 peroxide) is formed.'* The bromine first attaches itself to the dyad carbon, as 
 hydrochloric acid does, and then the product loses mercuric bromide : — 
 
 Br2 + C-NOhg = |j;>C=NOhg = hgBr + Br-q==N. 
 
 This gives bromo-formonitrile oxide, which will at once polymerize to the 
 dibromo-furoxane which is actually obtained: — 
 
 ^T, ^ ,x Br-C C-Br 
 
 2Br-C=N = II |>o. 
 
 ^0^ N-O-N 
 
 In presence of hydrochloric acid two molecules of benzonitrile-oxide combine 
 
 in a different way, giving dibenz-oxo-azoxime : — 
 
 O-N=C-0 
 Finally, in alkaline solution they give another series of polymers, the 
 trifulmines, which differ from the first two in readily going back to the mono- 
 molecular nitrile-oxides, with which they are identical in most of their reactions. 
 These bodies are probably triple polymers, and their formulae are written by 
 Wieland, on the analogy of cyanuric acid, thus : — 
 
 C-0 
 
 0| I 
 
 The reactions of the nitrile-oxides fall into two classes, depending on a 
 preliminary isomeric change either into a fulminate or into an isocyanate : — 
 K.q==N ^ (1) C=:N-OE 
 ^0'''^ -^ (2) EN=C=0. 
 The production of fulminic acid from methyl nitrolic acid evidently^takes place 
 through the unstable formonitrile-oxide, which can be isolated as||the triple 
 polymer : — 
 
 ■rr^^NOH _ H-C==N r.-wnTj 
 
 ' Wieland, Semper, Ann. 358. 36 (1908). 
 
 • Holleman, Eee. Trav. 10. 77 (1891) ; ■Wieland, Ber. 42. 4198 (1909). 
 
Nitrile-oooides 231 
 
 If benzonitrile-oxide is treated with hydrochloric acid, it is to some extent 
 hydrolysed to benzoic acid and hydroxylamine, which again indicates its 
 character as a hydroxylamine derivative : — 
 
 0-C=N -» 0-CO.OH + HjNOH. 
 The second type of reaction, leading to an isocyanate, is commoner. Thus the 
 triple polymer of benzonitrile-oxide is converted quantitatively into phenyl 
 isocyanate by heating with toluene : — 
 
 ^■^ - «^-N=C=0. 
 
 In the same way carbon dioxide and substituted ureas are always found among 
 the products of hydrolysis of these compounds. 
 
 This reaction, involving the migration of a hydrocarbon radical from the 
 carbon to a nitrogen doubly linked to it, is obviously analogous to the reactions 
 of Beckmann, Hofmann, and Curtius. In the case of the last two of these, 
 where an isocyanate is known to be formed, it is not an improbable suggestion ' 
 that the nitrile-oxide is an intermediate product. Thus in the Hofmann reaction 
 we should have : — 
 
 ONa /(X^ 
 
 EC=N-Br -♦ NaBr + EC^N -♦ EN=C=0. 
 and in the Curtius reaction : — 
 
 E-C-N<ll -* Ng + EC^N -* E-N=C=0. 
 
 The peculiar position of these bodies on the line between the hydroxylamine 
 derivatives of carboxylic acids (as shown by their giving hydroxylamine and 
 fulminic acid derivatives) and amine derivatives of carbonic acid (giving iso- 
 cyanates and their hydrolytic products) is well explained by the oxygen bridge. 
 This has the further advantage over the alternate formula E'C=N=0, that it 
 accounts for their comparatively saturated character, addition-reactions (apart 
 from polymerization) not taking place readily. 
 
 TEICYANOGEN COMPOUNDS 
 Nearly all cyanogen compounds polymerize with great ease to form tri- 
 molecular polymers, which in most cases can readily be reconverted into the 
 simple bodies. The most natural way of accounting for this is to suppose that 
 the three CN groups combine to form a tricyanogen ring, of the form 
 
 /C\ 
 
 N N 
 
 c ' 
 
 \J 
 
 A strict proof of the existence of this ring is still wanting ; but it has been 
 shown that the tricyanogen complex has a symmetrical structure. If cyanuric 
 chloride is treated with ammonia or amines, the chlorine atoms are replaced by 
 NHj or substituted NHg groups. By regulating the quantities, tl^e replacement 
 
 ' See however Wieland, Ber. 42. 4207. (1909). 
 q2 
 
H-N N-H N N 
 
 I I and II 1 „ 
 
 «< .. 9 ^-l,^/^ 
 
 232 Cyanogen Compounds 
 
 may be made to occur in three stages, and by employing successively ammonia, 
 • methylamine, and ethylamine, we can replace the three chlorine atoms by 
 amino-, methylamino-, and ethylamino-groups respectively, giving a substance 
 known as methyl-ethyl-melamine (C3N3) (NHg) (NH-CHg) (NH-C^Hs). Now it is 
 found that whatever be the order in which the three groups are introduced, the 
 same product is obtained, proving that the central C3N3 nucleus is symmetrical ; 
 and this can only be accounted for by assuming the existence of a ring of the 
 form suggested above, which is known as the triazine ring. 
 
 The mother substance of the group, triazine or tricyanogen hydride, has not 
 yet been prepared. It may have two tautomeric formulae : — 
 
 H 
 
 N-H 
 I and 
 
 \n/ '' Xn/ 
 
 H 
 The simple derivatives — those which contain no oxygen and only have the 
 hydrogen of the triazine replaced by hydrocarbon groups — must all be derived 
 from the first or second formula, since they are formed by the polymerization 
 of nitriles and not of isocyanides. But, as has been pointed out, the usual 
 product of the polymerization of a nitrile is a cyanalkine, an amino-pyrimidine, 
 such as 
 
 CH3.C Y 
 H.N-C ^.C^Hs 
 
 That is to say, instead of the three CN groups forming the ring, one nitrogen is 
 driven outside, while a carbon atom from one of the alkyls takes its place. 
 This is obviously only possible if the nitrile employed has, next to the CN, 
 a carbon atom united to hydrogen. Hence, if a nitrile is used which has no 
 hydrogen on this carbon, the reaction takes a different course, and a triazine 
 (tricyanogen) compound is formed. Tor instance, benzonitrile polymerizes in 
 presence of sulphuric acid to form triphenyl triazine. In this case a pyrimi- 
 dine could not be formed without breaking one of the benzene rings. Again , 
 a,a-dichloro-propionitrile condenses to a tricyanogen body which, on reduction, 
 yields triethyl-triazine : — 
 
 CCl2'CH3 CH2*CH3 
 
 /Cv (H) /C^ 
 
 N N U N N 
 
 CHaCCl^-n C-CClaCHa CH^CH^-C CCHa-CHg 
 
 \n/ \]Sr/ 
 
 This is a colourless compound of peculiar smell, melting at 29° and boiling at 
 193-195°, which, when heated with hydrochloric acid, gives ammonia and 
 propionic acid, i.e. is at once depolymerized and saponified. 
 
 Among the cyanic acid derivatives the formation of tricyanogen compoimds 
 
Cyanuric Add 233 
 
 is quite general. Cyanuric acid, CgNaOgHg, the mother substance of this group, 
 is formed by the spontaneous polymerization of cyanic acid, and also by heating 
 urea and many of its derivatives. It can obviously occur in two tautomeric 
 forms, one derived from normal cyanic acid HO-CN, and the other from isocyanic 
 aeidHN=C=0:— 
 
 N N-H 
 
 HOC-^COH 0:0-^^0:0 
 
 normal ll^]^ iso H-A^ij-H " 
 
 C-OH C:0 
 
 The balance of evidence is on the whole in favour of the iso (imide) structure 
 for the free acid. The tri-N-ester, the derivative of this form, is the product of 
 the action of diazomethane on the acid.^ Cyanuric acid is a pseudo-acid, and 
 the sodium salt, which is derived from the true acid form, shows by its behaviour 
 that it has the metal attached to oxygen : from which it follows that the free 
 acid must be the imide. The same conclusion has been arrived at by Hartley 
 from the absorption spectra of cyanuric acid and its derivatives. 
 
 The acid can also be obtained by heating biuret with cyanic acid : — 
 
 /NHa /NHg /NHa NH 
 
 ^C=0 C:=0 \c=o 
 
 It is best prepared by treating cyanuric bromide with water. On heating it 
 breaks up, regenerating cyanic acid. 
 
 The chloride and bromide of cyanuric acid may be got by heating the acid 
 with phosphorus pentahalide ; but they are more readily obtained by allowing 
 cyanogen halide to stand in the presence of a little halogen or halogen acid. 
 Chattaway has shown that in these bodies the halogen does not behave as if it 
 were attached to nitrogen. Hence they must have the normal structure 
 
 C1-C=N-, with the halogen on the carbon. They yield cyanuric acid on heating 
 with water. 
 
 The esters of cyanuric acid exist in four isomeric forms, as each alkyl group 
 
 may be attached either to oxygen or to nitrogen, AlkO-C=N- or Alk-N-C:0. 
 The tri-0-ester and the tri-N-ester have long been known ; and Hantzsch " has 
 recently prepared the two mixed compounds. The tri-0-ester is obtained from 
 cyanuric bromide and sodium methylate ; the di-0-mono-N-ester from silver 
 cyanate and methyl iodide below 0° ; the mono-0-di-N-ester from silver 
 cyanurate and methyl iodide at low temperatures : while these bodies if warm 
 give the tri-N-compound. The constitution of these esters is proved by heating 
 them with hydrochloric acid, when only the methyls which are attached to 
 oxygen are saponified. The tri-N-ester can also be obtained by boiling the tri- 
 0-ester with water, when it slowly changes over. 
 
 The isomerism of the esters is repeated, at any rate as regards two of them, 
 
 1 Palazzo, Scelsi, Oaz. 38. i. 659 (0. 08. ii. 774). 
 
 2 Hantzaoh, Bauer, Ber. 38. 1005 (1905). 
 
234 Cyanogen Compounds 
 
 in the mercuric salts, as Hantzsch^ has shown. If sodium cyanurate is treated 
 with mercuric chloride at 0°, it precipitates the oxygen salt ; but at 100° the 
 N-salt is formed. These salts are unfortunately very inactive and quite insoluble, 
 so that their molecular weights cannot be determined, nor the products of their 
 interaction with alkyl iodides ; but they are isomeric, and their formulae are 
 sufficiently established by the fact that one of them is decomposed by alkali to 
 form alkaline cyanurate and mercuric oxide, while the other is not affected. 
 This stability to alkali is a general character of compounds containing mercury 
 attached to nitrogen, as for example mercury acetamide. It is remarkable that 
 though by precipitating them at temperatures between 0° and 100° a mixture 
 of the two salts can be obtained, in which there is more of the N-salt the higher 
 the temperature, it is not found possible by any means to convert one salt into 
 the other. The isomeric salts are thus more stable than the isomeric esters. 
 This may be due to the high atomic weight of mercury, which renders it difficult 
 to move. 
 
 Such cases of isomeric salts are rare ; but the silver salt of benzamide ^ has 
 
 been obtained in two forms, which are no doubt 0"C^>tttao. ^nd 'P'^'\j^ri '• 
 
 while the mercury salt of nitroform offers the still more remarkable phenomenon 
 of an actually tautomeric salt, having one structure in ionizing and another in 
 non-ionizing solvents. 
 
 Cyamelide is an obscure polymer of cyanic acid, which may be mentioned 
 here, though it does not belong to the tricyanogen compounds in the strict 
 sense. It is formed by the spontaneous polymerization of anhydrous cyanic 
 acid, and also by the action of carbonyl chloride on ammonia. Its slow pro- 
 duction by the polymerization of cyanic acid in the gaseous state has been 
 investigated in great detail by van 't Hoff,' and it affords a classical example of 
 the influence of disturbing factors on the rate of gaseous reaction. The cyanic 
 acid vapour was enclosed over mercury, and the rate of reaction followed by 
 means of the change of pressure. It was found that the velocity was affected 
 by the area of the glass surface, being nearly half as great again in a long 
 narrow tube as in a globe. It was also much slower at first than after a certain 
 time had elapsed, and this was shown to be due to the accelerating influence 
 exerted by the layer of cyamelide on the surface of the glass : in fact the change 
 was more than three times as rapid in a vessel whose walls were covered with 
 a layer of cyamelide as in a clean vessel. Both these influences were much 
 diminished if the gas was diluted with air. If these disturbing influences are 
 eliminated, either by observing the initial velocities at various pressures in clean 
 vessels, or by using vessels already coated with cyamelide, it is found that the 
 reaction is approximately trimolecular. The structure of cyamelide was until 
 recently quite unknown, but Hantzseh * has been able to throw a certain amount 
 of light upon it. As it is non-volatile and insoluble in the ordinary solvents, 
 its molecular weight cannot be determined directly. But its effect on the 
 freezing-point of pure sulphuric acid (in which it dissolves) shows that its 
 
 ' Ber. 35. 2717 (1902). 2 Titherley, /. C. S. ISSV. 468. 
 
 ° See Studies in Chemical Dynamics. * Bet. 38. 1013 (1905). 
 
Cyamelide 235 
 
 moleculav weight cannot be very high, while the fact that its formation from 
 cyanic acid is a trimolecular reaction indicates that its formula must be 
 (CaNaHgOg)^. If w was greater than one, we should expect to find that it was 
 very much less active than cyanuric acid, which is not the case : for example, 
 it is much more easily decomposed by sulphuric acid. We may therefore 
 conclude that its formula is probably O3N3H3O3. It is feebly acidic, but gives 
 no salts, and is converted by alkalies into a salt of cyanuric acid. The relations 
 of these bodies are shown in the following scheme : — 
 
 Cyamelide 
 ■(CN0H)3 
 
 Cyanic Acid 
 HOCN ■ 
 
 OH' 
 
 + NH,. 
 
 Cyanuric Acid 
 (CNOH), 
 
 Since cyanuric acid cannot be converted into cyamelide directly, but only 
 
 through cyanic acid, cyamelide cannot have any of the possible pseudo-cyanuric 
 
 formulae, nor can it be a stereoisomer of cyanuric acid. It cannot in fact have 
 
 a tricyanogen (triazine) ring at all ; nor, since it forms no salts, can it have any 
 
 I I I 
 
 group capable of forming a salt directly (as HO-C-N-) or indirectly (as HN-C=0). 
 
 Since it gives a mercury derivative in which the metal is attached to nitrogen, 
 
 it must contain imide groups. Hence it must be a polymer not of HO-C=N but 
 
 of H-N— C=0, and the three molecules of this must be joined up through oxygen. 
 
 We thus arrive at the structure : — 
 
 Q 
 
 HN=C^ ^C=NH 
 
 V 
 
 C=NH 
 
 This accounts for all the relations described above ; in fact cyamelide is related 
 to cyanic acid in the same way as paraldehyde to aldehyde ; and hence it is 
 more easily converted than is cyanuric acid into cyanic acid (or carbon dioxide 
 and ammonia) under the influence, for example, of acids. 
 
 TMocyanuric Compoimds 
 
 Corresponding to the cyanuric compounds are a group of the thiocyanuric 
 derivatives, in which the oxygen is replaced by sulphur. They resemble the 
 cyanuric compounds so closely that they need not be described in detail. But 
 they afford another example of the rule which holds generally among bodies 
 exhibiting amide-imide tautomerism : that while in the oxy-derivatives the imide 
 (NH) form is the most stable, both in the free hydrogen compound and in the 
 salts and esters, the reverse is the case among the thio-compounds. Thus both 
 
236 Cyanogen Compounds 
 
 the N- and the 0-esters of cyanuric acid are stable, but the former are the more 
 stable, since the normal esters are slowly converted into them on boiling. 
 Among the esters of thiocyanuric acid these relations are reversed. Normal (8-) 
 thiocyanuric esters are easily prepared, while it is doubtful whether the iso (N-) 
 esters exist at all. 
 
 Melamines, 4'C. 
 As a tribasic acid, cyanuric acid can give a mono-, di-, and triamide ; and 
 each of these, like cyanuric acid itself, has two tautomeric forms. The most 
 important of these bodies is the triamide, melamine 
 
 N N-H 
 
 HaN-C'^ C-NHj HN=C C^NH 
 
 %/N °^ H-N N-H 
 
 CNH2 C=NH 
 
 It is formed by the action of ammonia on the normal esters of cyanuric or thio- 
 cyanuric acid, or better on cyanuric chloride. The symmetrical trialkyl 
 melamines may have three formulae:— 
 
 N N-H l^CHj 
 
 CH3.NH-(^/ C-NH-CH3 CHgN^CjT ^(]=NCH3 HN=C'^\=NH 
 N^/N H-N^^N-H CHj-N. N.CH3 ' 
 
 CNH-CHg C=NCH3 C=NH 
 
 The first and second of these differ only in the position of a hydrogen atom, 
 and are therefore tautomeric, while the third is quite distinct. Accordingly, we 
 find that two series of trialkyl melamines are known. The members of one 
 series are got by acting on thiocyanuric esters or cyanuric chloride with primary 
 amines : on decomposition they give cyanuric acid and primary amines. They 
 therefore possess the first or the second formula. To distinguish them from the 
 compounds of the third type they are called ea;o-trialkyl melamines, since the 
 nitrogen of the amine group is outside the ring. Derivatives of the third form, 
 with the amine nitrogen forming part of the ring, are known as eso-melamines. 
 They are obtained by the polymerization of the alkyl cyanamides, and their 
 constitution is proved by their giving on saponification ammonia and isocyanuric 
 esters. 
 
 Summary 
 
 The following summary of the polymerization products of the different 
 classes of cyanogen compounds may be of use. 
 
 Cyanogen forms paracyanogen, of unknown molecular weight and structure. 
 
 Hydrocyanic acid on standing in aqueous solution in the presence of a little 
 potassium carbonate gives a triple polymer, which is probably amino-malonio 
 nitrile 
 
 CN 
 
 H-C-NHa. 
 
 CN 
 
 The ordinary nitriles of the fatty series, if treated with sodium in ethereal 
 
Polymerization of Cyanogen Compounds 237 
 
 solution, yield double polymers which are imino-nitriles : for example, methyl 
 cyanide gives 
 
 CHg-fcNH 
 
 CHa-CN ' 
 and when treated with sodium in the absence of a solvent they give cyanal- 
 kines, i.e. pyrimidines. 
 
 Those fatty nitriles which have no hydrogen attached to the same carbon as 
 the CN, and all aromatic nitriles, give derivatives of the tricyanogen ring : thus 
 benzonitrile gives 
 
 N 
 
 Cyanic acid changes spontaneously into a mixture of cyanuric acid, a tri- 
 cyanogen or triazine compound, and cyamelide, a body containing a ring of 
 three carbons and three oxygens. 
 
 The cyanogen halides give the halides of cyanuric acid on standing. The 
 isocyanic esters give the isocyanuric esters. 
 
 Cyanamide on heating is converted into the corresponding cyanuric com- 
 pound, melamine or cyanuric triamide, and also into a double polymer dicyan- 
 diamide, which is cyan-guanidine 
 
 Fulminic acid polymerizes spontaneously to an isoxazolone derivative, which 
 changes to cyanisonitroso-acethydroxamie acid: while its salts on boiling in 
 aqueous solution form the salts of fulminuric acid or nitrocyanacetamide : — 
 
 NO2 
 H-C-C<jjjj . 
 CN 
 Finally, the nitrile-oxides change spontaneously into two classes of double 
 polymers, furoxanes (glyoxime-peroxide) and oxo-azoximes : and also into triple 
 polymers (trifulmines) probably containing the tricyanogen ring. 
 
DIVISION III 
 
 COMPOUNDS CONTAINING AN OPEN CHAIN OF TWO 
 OE MOKE NITROGEN ATOMS 
 
 These compounds fall into two subgroups : — 
 
 1. Those containing two, and 
 
 2. Those containing more than two nitrogen atoms in the chain. 
 
 The second of these subgroups is comparatively unimportant. The first 
 consists of: — 
 
 1. The derivatives of hydrazine, HgN-NHa, including the hydrazo-com- 
 pounds. 
 
 2. The diazo- and azo-compounds, derived from the hypothetical diimide, 
 HN=NH. 
 
 3. The oxyazo-compoimds, which, though strictly speaking ring bodies, 
 
 -N— N- 
 containing the group \/ , are more conveniently dealt with here. 
 
 O 
 
 -'2 
 
 XT TT 
 
 4. The nitroso- and nitro-amines, E-N<(j^q and E-N^^^j^q 
 
CHAPTER X 
 
 HYDRAZINE DERIVATIVES 
 
 These may be divided into three classes : — 
 
 1. Purely aliphatic derivatives. 
 
 2. Purely aromatic and mixed derivatives, which have either only one 
 aromatic group, or two attached to the same nitrogen. 
 
 3. Compounds containing aromatic groups attached to both nitrogen 
 atoms. These are the hydrazo-compounds, such as hydrazobenzene, 0-NH-NH-0. 
 Their behaviour is so different from that of the other members of the group 
 that they must be dealt with separately. 
 
 FATTY HYDRAZINES 
 
 The primary alkyl-hydrazines may be obtained by the action of alkyl 
 halides on free hydrazine, or from the mono- or symmetrical di-alkyl ureas by 
 conversion into nitroso-compounds, and reduction with zinc and acetic acid to 
 hydrazine-ureas or semicarbazides, which are then hydrolysed with fuming 
 hydrochloric acid : — 
 
 /NH.Et /N<?J> ^N<JH^ ^H,H-CO, 
 
 CO -» CO ^* -» CO ^* -* NHEt 
 
 \NH.Et ^NHEt "^NH-Et +NH2-Et 
 
 They are very hygroscopic, strongly basic liquids. They smell like ammonia, 
 and are corrosive, rapidly destroying cork. When warmed with potassium 
 pyrosulphate, KaSgO^, they give salts of the hydrazine sulphonic acids, such as 
 
 Et-NH-N<^gQ r^. If this body, the potassium salt of monethyl-hydrazine 
 
 sulphonic acid, is oxidized with mercuric oxide, it is converted into diazo- 
 ethane-sulphonic acid, CgHg-N^N-SOgK, one of the very few known instances of 
 a fatty diazo-compound of the aromatic or true diazo type. 
 
 The secondary fatty hydrazines were obtained by Emil Fischer by the 
 reduction of secondary nitrosamines : — 
 
 EtaN-N:© -* EtgN-NHj. 
 They may also be obtained by the direct action of alkyl iodides on hydrazine 
 hydrate. In this reaction it has been shown that the alkyl groups attach 
 themselves only to one of the two nitrogen atoms : the products, for example, 
 from hydrazine and methyl iodide being CHg-NH-NHa, (CH3)2N-NH2, and 
 finally (CH3)3NI-NH2.^ 
 
 The secondary alkyl-hydrazines, like the primary, are strongly basic hygro- 
 
 1 Harries, Haga, Ber. 31. 56 (1898). 
 
242 Hydrazine Derivatives 
 
 scopic liquids, which are easily soluble in water. Nitrous acid converts them 
 into secondary amines and nitrous oxide : — 
 
 Et,N H(p ^ ^ ^ ^ 
 
 NHa 0:N ' 
 
 Like all hydrazines they are easily oxidized, and hence are strong reducing 
 agents. Mercuric oxide converts them into tetrazones, such as Et2N-N=N-NEt2 , 
 which are strongly basic liquids. 
 
 AEOMATIC HYDRAZINES 
 
 The aromatic hydrazine derivatives were the first to be discovered, and are 
 by far the most important. Our knowledge of them, as of the fatty compounds, 
 is mainly due to EmU Fischer. The first member of the group was prepared 
 by Strecker and Eomer in 1871, by the action of potassium hydrogen sulphite 
 on diazobenzene nitrate : — 
 
 ^-Nj-NOg + KHSO3 -* 0NH-NHSO3K. 
 
 In 1875 Fischer showed that this body on boiling with hydrochloric acid 
 gave phenyl-hydrazine hydrochloride, and by treating this with alkali he prepared 
 phenyl-hydrazine itself, one of the most important reagents in the whole of 
 organic chemistry. 
 
 Formation 
 
 The primary and secondary compounds are both obtained from the corre- 
 sponding amines by replacing one hydrogen attached to nitrogen by NHj, and in 
 both cases by treatment with nitrous acid and subsequent reduction. With 
 the primary amines the diazo-compounds are first produced, and these can then 
 be reduced in three ways : — 
 
 1. With stannous chloride in hydrochloric acid solution, giving the hydrazine 
 hydrochloride directly. 
 
 2. They can be coupled with potassium hydrogen sulphite, forming the 
 diazo-sulphonate 0-N=N-SO3K, and then reduced with excess of sulphite, or 
 better with zinc dust and acetic acid, to the hydrazine-sulphonate 
 
 0NH-NHSO3K, 
 which is subsequently hydrolysed by boiling with acids. 
 
 3. The diazo-compound is coupled with an amine to form a diazoamino- 
 body, which is then split by reduction : — 
 
 0-N=N-NH-0 + 4 H = 0-NH-NH2 + HgN-^. 
 
 4. They may also be obtained' directly from the phenols (and still more easily 
 from the naphthols) by heating them in aqueous solution with hydrazine sulphite 
 and excess of hydrazine hydrate : — 
 
 Ar-OH -f H2N-NH2 = Ar-NH-NH^ -t- H2O. 
 This reaction corresponds exactly to the conversion of phenols (and more easily 
 naphthols) into amines. 
 
 5. The less important secondary hydrazines are prepared by reducing the 
 nitrosamines with sodium amalgam, or better with zinc and acetic acid in cold 
 alcoholic solution. 
 
 » Franzen, J. pr. Ch. [2] 76. 205 (C. 07. ii. 1337). 
 
Aromatic Hydrazines: Properties 243 
 
 Properties of the primary and secondary aromatic hydrazines 
 
 They are colourless liquids or solids, of a feebly aromatic smell, which are 
 easily soluble in alcohol and ether, but only slightly in water, and scarcely at 
 all in concentrated alkali. They are decided monacid bases, forming salts with 
 mineral and some organic acids. Unlike the primary fatty hydrazines they will 
 not form salts with two equivalents of acid, but only with one. Secondary 
 aromatic hydrazines, such as ^a^-NHj, will form salts with one equivalent of 
 acid, but these are partially decomposed by water. This is like the behaviour 
 of the secondary aromatic amines, and is a sign of the negative character of the 
 phenyl group. It is practically certain that it is only the NHg group which 
 takes part in the formation of these salts. On the other hand, the hydrogen of 
 the NH group in phenyl-hydrazine can be replaced by an alkali metal. Sodimn 
 dissolves in phenyl-hydrazine, evolving ammonia and producing aniline. That 
 is, the hydrogen driven out from the first molecule of phenyl-hydrazine reduces 
 a second molecule. On evaporating oflf the anUine, the sodium compound, 
 ^-NNa-NHg, remains as a yellow-red transparent mass, absorbing moisture from 
 the air to form sodium hydrate and phenyl-hydrazine, and giving with ethyl 
 iodide unsymmetrical ethyl-phenyl-hydrazine, ^EtN-NHg. 
 
 The aromatic hydrazines are very easily oxidized. It is probable that the 
 first oxidation product of a primary hydrazine is a diazo-compound. Phenyl- 
 hydrazine siJphate can be partially converted by mercuric oxide into benzene 
 diazonium sulphate, ^-N^N-SO^H. Chattaway has shown' that all primary 
 aromatic hydrazines are oxidized by chlorine at -15°, or by bromine in acetic 
 acid solution, to give diazo-compounds ; and he suggests the following series of 
 reactions to account for the change : — 
 
 ENH E-N-Gl KN-Cl 
 
 H^.H -* H-A-Cl -* i^ ' ^ ^^^- 
 
 Further oxidation '^ by means of copper sulphate or ferric chloride, or by bromine 
 in presence of alkali, or even by atmospheric air, especially in the presence of 
 alkali, removes all the nitrogen as nitrogen gas, while the aromatic radical is 
 reduced to the hydrocarbon :— 
 
 Ar-NH-NHa -}- O = Ar-H -h N^ -H HgO. 
 This reaction may be used for determining a hydrazine quantitatively; and it also 
 affords an indirect means of replacing the diazo-group by hydrogen. Chattaway' 
 explains it in a similar way to the formation of a diazo-compound : — 
 
 ENH EN-H ENH 
 
 I -» I -* ! -^ 111 -f- I • 
 
 HN-H HNOH H N OH 
 
 The intermediate compounds assumed in these reactions have not, however, been 
 
 isolated. '' 
 
 Secondary aromatic hydrazines behave differently. Gentle oxidation, as in 
 
 ' J. C. S. 1808, 852. ' Chattaway, J. C. S. 1907. 1323 ; 1909. 1065. 
 
 ' /. C. S. 1908. 270. 
 
>N-NH2 +0 = "; T '^ + Na + H^O. 
 
 ^2 
 
 244 Hydrazine Derivatives 
 
 the case of the fatty compounds, gives the tetrazones Ar2N-N=N-NAr2 : whUe 
 a stronger oxidizing agent converts them into nitrogen and secondary amines : — 
 
 2Ar2N-NH2 + O = 2Ar2NH + N^ + H2O. 
 An exception to this rule, which otherwise holds both for fatty and for aromatic 
 derivatives, is presented by dibenzal-hydrazine, (ip-GR^^-'S'H.^} When this 
 is oxidized by mercuric oxide, it does not give a tetrazone but loses its nitrogen 
 and forms dibenzyl : — 
 
 0.CH2^^ NH, + O - ^^^^ 
 
 In the elimination of the nitrogen, and the union of the two groups attached to 
 it, this reaction is exactly similar to the oxidation of phenyl-hydrazine to 
 nitrogen and benzene. 
 
 The hydrazines can be reduced, though not easily, and then always split 
 between the two nitrogen atoms, giving ammonia and an amine. 
 
 Nascent nitrous acid acts on the primary hydrazines at low temperatures to 
 
 NH 
 
 form very unstable nitroso-derivatives,^ such as 'P'^'C-Kn'^t which on gentle 
 
 ^N 
 warming with alkali lose water and pass into azides, ^-N l| . 
 
 When heated with fuming hydrochloric acid to 200°, phenyl-hydrazine is 
 converted into ^-phenylene diamine: — 
 
 CeH^-NH-NH^ = C6H,(NH2)2, 
 the usual migration of a substituent from the NH2 of an aniline to the ring. 
 
 1,2,4-dinitro-phenyl-hydrazine is remarkable for the extraordinary ease with 
 which it is acetylated.' It is sufficient to warm it with even highly diluted 
 aqueous acetic acid to obtain the acetyl derivative CeH3(N02)2-NH-NH-CO-CH3. 
 
 The mono-alkyl derivatives of phenyl-hydrazine are of two kinds, the sym- 
 metrical (/3) and the unsymmetrical («). The a or unsymmetrical are got from 
 the mono-alkyl-anilines by conversion into nitrosamines and reduction ; or from 
 sodium phenyl-hydrazine and alkyl iodide. If phenyl-hydrazine itself is treated 
 with alkyl bromide, a mixture of the a- and ^-alkyl derivatives is obtained. To 
 separate the /3, the mixed product is oxidized with mercuric oxide : the a-body is 
 thus converted into the unstable basic tetrazone, while the /3- (symmetrical) 
 gives a mixed azo-compound, as ^-N^N-Et, which, from its volatility and 
 indifference to acids, can easUy be separated and re-reduced. 
 
 The dialkyl derivatives are got from the a-mono-alkyls in the following 
 manner. The mono-derivative is converted into the formyl-hydrazide by 
 heating with formic acid : the sodium derivative of this is then treated with 
 alkyl iodide, which gives the dialkyl compound, and this on saponification gives 
 the dialkyl-aryl-hydrazine :— 
 
 ' Busoh, WeisB, Ber. 33. 2701 (1900). ^ Cf. Thiele, Ber. 41. 2806 (1908). 
 
 5 Curtius, Mayer, J. pr. Ch. [2] '76. 369 (C. 08. i. 125). 
 
Aromatic Hydrazines: Properties 245 
 
 These bodies, when boiled with alkyl halide, are mainly converted into the 
 quaternary azonium compounds, as 
 
 CH-NI-NH-OoH., 
 but also to some extent into the tri-alkyl-aryl-hydrazines, such as 
 
 Tetra-atyl-hydraeines 
 Tetra-phenyl-hydrazine was first obtained ^ by the action of iodine on the 
 sodium derivative of diphenylamine, ^j^-Na, but it and its analogues can be 
 prepared more simply by the oxidation of the diarylamines with lead dioxide or 
 potassium permanganate.' 
 
 These bodies are remarkable for giving brilliant colours with mineral acids ; 
 in fact it is to the production of tetra-phenyl-hydrazine that the blue colour 
 formed in the diphenylamine test for nitric acid is due. These colours do not 
 depend, as was at first supposed, on the splitting of the molecule between the 
 two nitrogen atoms, since under proper conditions the hydrazine can be recovered 
 unchanged. The coloured bodies are therefore coloured salts of the colourless 
 hydrazines. Similar coloured compounds (not double salts) are formed by the 
 addition of the halides of phosphorus, tin, iron, aluminium, and zinc. To account 
 for their colour Wieland suggests * that they contain a quinoid ring, and are in 
 fact quinol derivatives : for example, the body obtained from tetra-tolyl-hydra- 
 zine and hydrochloric acid may have the formula : — 
 
 CI 
 CHa-CsH^xij — .^/CHg 
 CH3-CoH/Y~W\H . 
 
 CgH^CHg 
 At first this seems impossible, since the quinols have no colour. But quinone 
 diimine, HN=C5H4=NH, is itself colourless, while its derivatives of the type 
 
 in which the nitrogen has become pentad, and has no hydrogen attached to it, 
 are brilliantly coloured ; and we may well suppose that a similar change would 
 produce a coloured compound from a colourless quinol. 
 
 These coloured derivatives are especially stable in the case of tetra-tolyl- 
 hydrazine, and it is to be noticed that the stability of the quinols themselves is 
 greatly increased by the presence of a para-methyl group. In the case of tetra- 
 phenyl-hydrazine they soon decompose and lose their colour. The compound 
 splits between the two nitrogen atoms, no doubt primarily thus : — 
 
 /CI 
 0,N-N=CeHa.H = ^^NH + Cl-N^^. 
 
 ' For a remarkable case of stereo-hindrance in the reaction of the hydrazines with the chloro- 
 fatty acids see Busch and MeussdorfBer, J. pr. Ch. [2] 75. 1035 (C. 07. i. 121). 
 
 " Chattavfay, Ingle, J. C. S. 1895. 1090. * Wieland, Gambarjan, Ber. 39. 1499 (1906). 
 
 * Ber. 40. 4260 (1907). 
 1175 E 
 
246 Hydrazine Derivatives 
 
 The diphenylamine can actually be isolated ; the second half of the molecule 
 decomposes further,' giving mainly a diphenylamine derivative : — 
 
 ^^N.Cl + H-C>-NCl-0 -> HCl + [02N-<3-NC10 -»] 
 
 02N-C3-N-OO1. 
 
 ACID HYDEAZIDES 
 These bodies are derived from the acids by replacing the hydroxyl by the 
 hydrazine residue, as in C!H3'C'^^tt_-^ttt j . They thus correspond to the amides. 
 
 They can be obtained from inorganic as well as organic acids. 
 
 Thionyl-phenyl-hydrazone, 0NH-N=S=O, is a kind of dihydrazide of sul- 
 phurous acid. If sulphur dioxide is passed into a solution of phenyl-hydrazine 
 in benzene, the additive compound (0NH-NH2)2-SO2 is precipitated ; and this on 
 heating loses water to give the thionyl-hydrazone, a very stable substance which 
 is volatile without decomposition in steam, but is easily broken up by 
 alkalies, re-forming phenyl-hydrazine and potassium sulphite. The hydrazido- 
 sulphonic acids, as "^NH-NH-SOgH, which are intermediate products in the 
 reduction of diazobenzene-sulphonic acid, are half-hydrazides of sulphuric acid. 
 
 The hydrazides of the fatty acids are of two kinds, a and ^. In the a the 
 acyl is attached to the same nitrogen as the phenyl ; in the ^ or symmetrical it 
 is on the other nitrogen atom. 
 
 The a (unsymmetrical) hydrazides may be obtained by the action of an acid 
 
 chloride or anhydride on sodium phenyl-hydrazine, or by treating yS-acet- 
 
 phenyl-hydrazine with an acid chloride, which gives an a-;3-di-acyl compound, 
 
 ^■N-NHCOCHa 
 
 I , and then boiling with dilute sulphuric acid, which splits off 
 
 CO-R 
 
 the /3-acyl group, and leaves the a- untouched. The a-compounds are distinguished 
 
 from the yS- by not giving the so-called Bulow reaction — a colour with a trace of 
 
 ferric chloride in concentrated sulphuric acid. 
 
 The /S- or symmetrical hydrazides are the product of the direct action of 
 phenyl-hydrazine on the acid chlorides, anhydrides, or amides. With one or 
 two exceptions, such as ^-tolyl-hydrazide, they give a red or violet colour in the 
 Bulow reaction. 
 
 Mercuric oxide oxidizes them in chloroform solution to give red unstable 
 bodies, which evolve nitrogen and are apparently diazo-compounds, such as 
 0-N=N-COCH3. 
 
 Many acids react so readily that even on warming with a solution of phenyl- 
 hydrazine in acetic acid they precipitate the hydrazide. This is particularly the 
 case with the highly oxidized acids of the sugars, where this reaction affords 
 a valuable method of separation, as the hydrazides are easUy purified, being only 
 slightly soluble, and can then be broken up again into their components by 
 boiling with bai-yta. 
 
 J Wieland, Ber. 41. 3478 (1908). 
 
Hydrazones 247 
 
 HYDEAZONES 
 
 These are among the most important derivatives of phenyl-hydrazine and its 
 analogues. They were discovered by Emil Fischer in 1883. They are of great 
 value for characterizing compounds containing the carbonyl group. They are 
 produced, as a rule, very easily — generally by warming the carbonyl compound 
 with free phenyl-hydrazine in acetic acid solution — when the two condense with 
 the elimination of a molecule of water : — 
 
 ^ >C=0 + H^N-NH^ = ^ >C=N-NH0 + H^O. 
 
 The rate of formation of the phenyl-hydrazones of a variety of ketones in 
 a variety of solvents has been investigated by Petrenko-Kritschenko.* The 
 results are not easy to reduce to a system, but the general conclusions are (1) that 
 the formation is most rapid in acetic acid solution, the other solvents coming in 
 the order ligroin (hexane) — nitrobenzene — benzene, the last being the slowest ; 
 and (2) that the ring ketones, such as hexamethylene ketone, are, except in 
 acetic acid solution, much more rapidly acted on than the open-chain compounds. 
 
 The phenyl-hydrazones may conceivably have any one of the three formulae : — 
 
 I. |>C=N-NH0. II. ^ >C<r . III. \yO<^^^^ . 
 
 Formula III is excluded by the fact that ethane-azo-benzene, ^N^N-CHg-CHa, 
 which has this structure, and is obtained by oxidizing ethyl-phenyl-hydrazine 
 with mercuric oxide, is different from the phenyl-hydrazone of acetaldehyde. 
 Formula II is impossible because the unsymmetrical secondary hydrazines, 
 EEiN-NHj, also form hydrazones, the body obtained from benzaldehyde and 
 unsymmetrical phenyl-ethyl-hydrazine, ^EtN-NHg, which must have the struc- 
 ture I, being identical with that got by ethylating benzaldehyde-phenyl-hydrazone. 
 We are therefore left with the first as the only possible structure for the 
 hydrazones. 
 
 The hydrazones, when treated with Fehling's solution, evolve no nitrogen, 
 whereas, as we have seen, hydrazines evolve all their nitrogen under these con- 
 ditions. This fact may be made use of to determine the carbonyl group quanti- 
 tatively. A known weight of phenyl-hydrazine, in excess of that required, is 
 added to the carbonyl compound, and the excess determined by measuring the 
 nitrogen evolved on treatment with Fehling's solution. 
 
 On warming with acids the hydrazones are broken up back again to a greater 
 or less extent, but seldom quantitatively, into ketone or aldehyde and phenyl- 
 hydrazine. The best way of obtaining a carbonyl compound from a hydrazone, 
 which is sometimes important when it is used for a separation, is to treat it with 
 an aqueous solution of pyroracemie acid, when the hydrazone group migrates 
 from the one compound to the other : — 
 
 • ^>C=N-NH^ + CHgCO-COOH = ^ >C=0 + CH3C<^qJ§*^. 
 
 ' C. 03. i. 1129 ; ii. 491. 
 
 k2 
 
248 Hydrazine Derivatives 
 
 The hydrazones of fatty (not aromatic) carbonyl compounds add on hydro- 
 cyanic acid to give hydrazido-nitrUes : — 
 
 0.NH-N=CH.CH3 _ 0-NH-NH.CH.CH3 
 + H.CN - CN 
 
 By careful reduction they can sometimes be converted into hydrazido- 
 
 compounds ' : — 
 
 0.NH-N=CH.COOH -» 0.NH-NHCH2.COOH. 
 
 Stronger reducing agents convert them at once into a mixture of amines : — 
 (CH3)2C=N.NH.0 + 4 H = (CH3)2CH.NHis + NHj^. 
 This is Tafel's method of preparing amines. The reducing agent used is either 
 sodium amalgam and acetic acid in alcoholic solution, or, more recently, 
 electrolysis in sulphuric acid.' 
 
 Gentle oxidation with amyl nitrite converts the phenyl-hydrazones into 
 
 hydrotetrazones : e. g. 
 
 6.CH=N N=CH.0 
 2fCH=N-NH0-.^ 4_li.^ • 
 
 These bodies give a brilliant red solution in sulphuric acid. This is probably the 
 origin of the Btilow reaction, a bright red or yellow colour which all hydrazones 
 give when treated with a crystal of ferric chloride or potassium bichromate in 
 concentrated sulphuric acid. 
 
 When they are heated with zinc chloride they give off ammonia, with the 
 production of indol derivatives, a reaction which has not yet been satisfactorily 
 explained : — 
 
 0.NH-N=C<gg3 _, Q:gg>C.CH3. 
 
 (./CH3 
 
 ^.NH-N^CH.CHa-CHg -» [Y \m . 
 
 "^NH 
 A large number of phenyl-hydrazones have been found to occur in two more 
 or less stable modifications.' In all these cases the bodies are of the type 
 Ar.NH-N=CEKi, while no such isomerism is observed in the derivatives of 
 symmetrical ketones of the type Ar.NH-N^CEg. It is therefore probable that 
 the two forms are stereoisomers, corresponding to the stereoisomeric oximes : — 
 
 II \ II : II \ II 
 
 HO-N ' N.OH Ar-NH-N ' N-NH.Ar 
 
 The other possible isomeric formula, that of the azo-compounds, ArN^N-CHERj , 
 is excluded by the fact that in many cases these bodies are known as well as the 
 two forms of the hydrazone, and that they are coloured. 
 
 In the case of the derivative of cyanacetic ester, Ar.NH-N=C\pj-.^Tj , 
 
 Hantzsch and Thompson ' find that in the a-modification the hydrogen attached 
 to the nitrogen is more acidic than in the /8-, and more easily replaced by the 
 
 » Cf. Schlenk, J. pr. Ch. [2] 78. 49 (1908). 
 
 ' Tafel, Pfeffermann, C. 02. i. 1207. Cf. Ber. 33. 2209 (1900). 
 
 ' Cf. Lookemann, Liesohe, Ann. 342. li (1905). •• Ber. 38. 2266 (1905). 
 
Hydrazones 249 
 
 acetyl group. They hence assume that in the a- this hydrogen is nearer to 
 
 the acidifying ON group than in the other :— 
 
 NC-C-COOR NC-0-COOR 
 
 a II : j3 II 
 
 Ar-NH-N '^ N-NH-Ar 
 
 This explanation is ingenious and probable ; but it is to be noticed that the 
 
 assumption is made here for the first time that the chemical influence of one 
 
 group on another (as distinguished from the reaction of one group with another) 
 
 is determined by their stereo-chemical distance apart. 
 
 The hydrazone of acetaldehyde, 0-NH-N=C\qtt , occurs in two forms,* 
 
 melting at 98° and 57° respectively. Here the azo-com pound, ^•N=N-CH2-CH3, 
 is known, and is quite distinct. The isomerism of the hydrazones, as described, 
 cannot be reconciled with any theory ; for each is supposed to be the stable form, 
 and one passes over into the other when treated with a trace of ammonia, while 
 in presence of a trace of sulphur dioxide the reverse change takes place. It is 
 evident that the two differ but little in stability, and that they are near their 
 transition point, but the facts require further investigation. 
 
 Acetaldehyde-phenyl-hydrazone is formed from the isomeric ethane-azo- 
 benzene by the action of strong acids '' ; and there is some reason to think that 
 alkalies effect the reverse change. The azo-compound is red ; and it has been 
 suggested ' that the red colour which many hydrazones assume on standing, and 
 especially on exposure to light, is due to the change into the azo-compound. The 
 ultra-violet absorption of the product supports this view.* 
 
 Para-nitro-phenyl-hydrazine ' combines with acetone with such readiness, 
 giving an almost insoluble hydrazone, that it can be used for the quantitative 
 determination of acetone in methylated spirit, the spirit being treated with the 
 hydrazine in acetic acid solution, and the precipitate filtered off and weighed. 
 
 The compound obtained from acetoacetic ester and phenyl-hydrazine is not 
 a true hydrazone but a derivative of the enolic form of the ester. It must be 
 phenyl-hydrazo-crotonic ester, since on oxidation it gives benzene-azo-crotonic 
 ester. When warmed in vacuo to 200° it loses alcohol and forms phenyl- 
 methyl-pyrazolone : — 
 
 (JH3 CH3 (.jj^ 
 
 C.N=N.0 C-NH-NH^ \^_^^ 
 
 * 
 
 CH CH I >N-0' 
 
 COOEt COOEt HC— CO 
 
 If the pyrazolone is treated with methyl iodide, the hydrogen atom marked with 
 a * is replaced by methyl, giving dimethyl-phenyl-pyrazolone, which is antipyrine. 
 This is the method used for the commercial preparation of this substance. 
 
 Compounds containing two carbonyl groups in the molecule can give both 
 mono- and di-hydrazones. a-dihydrazones are known as osazones. These bodies 
 are of great importance in the chemistry of the sugars, and are there formed by 
 a peculiar reaction. The sugars have a hydroxyl group attached to every carbon 
 
 « Lockemann, Liesche, Ann. 342. 14 (1905). ' E. Tischer, Bei: 29. 793 (1896). 
 
 ' Chattaway, J. C. S. 1906. 462. * Baly, Tuck, J. C. S. 1906. 982. 
 
 ' Blanksma, van Ekenstein, Bee. Trav. 22. 434 (C 04. i. 14). 
 
250 Hydrazine Derivatives 
 
 atom but one, that one having a oarbonyl group in the form of ketone or 
 aldehyde. On treatment with phenyl-hydrazine the carbonyl forms a hydrazone 
 in the normal manner. But if excess of phenyl-hydrazine is used and the 
 mixture warmed in acetic acid solution, the reaction goes further. Another 
 molecule of the hydrazine removes two hydrogen atoms from the next CHOH, 
 converting it into CO, which condenses with a third molecule, while the 
 hydrogen removed reduces the phenyl-hydrazine into aniline and ammonia. 
 For example : — 
 
 CH2OH CH2OH 
 
 (CHOH), (CHOH), 
 
 N ^' -> N '^ + 6NH-NH, 
 
 CHOH CHOH ^ ^ 
 
 HC=0 HC=N-NH0 
 
 CH,OH CH,OH 
 
 "^3 
 
 f 
 
 (CHOH) 3 -I- 0NH2 (CHOH) 3 
 
 C=0 + NH3 ""* 0=N-NH^' 
 
 HC=N-NH^ HC=N-NH^ 
 
 The corresponding ketose, CH20H-(CHOH)3-CO-CH20H, gives the same 
 osazone. The importance of these osazones in the chemistry of the sugars is due 
 to the fact that they are almost insoluble in water, whereas the hydrazones are 
 very soluble. They can therefore be easily separated and identified by means of 
 their melting-points, &c., whereas the sugars themselves are often difficult to 
 crystallize, and so can neither be purified nor identified directly. 
 
 Many osazones, when treated with ferric chloride in alcoholic solution, give 
 a red colour, by which they can be distinguished. This is due to their oxidation 
 to osotetrazones : — 
 
 CH3-C=N-NH-0 CH3-C=N-N.0 
 
 CH3-C=N-NH-^ ~* CH3-C=N-N-0' 
 
 Besides the usual method of forming hydrazones — by the action of a carbonyl 
 compound on phenyl-hydrazine — there is another which is commonly assumed 
 to lead to the production of hydrazones (though this is not certain), and which 
 is in any case of great theoretical interest. This consists in the action of diazo- 
 compounds on open-chain bodies containing an acidic methylene group : that is, 
 a CH2 attached to negative groups such as NO2, CO, or COOEt. Thus the 
 primary nitroparaffins, y8-keto-esters, malonic ester, &c., combine with diazo- 
 bodies to form derivatives which were at first supposed by their discoverer 
 V. Meyer, to be azo-compounds : — 
 
 COOEt COOEt 
 
 CH2 + HON=N-0 = CH-N^N-^ -f H^O. 
 COOEt COOEt 
 
 But V. Meyer found that this compound gave on saponification an acid 
 
 identical with that obtained from mesoxalic acid and phenyl-hydrazine, which 
 
 one would assume to be a hydrazone : — 
 
Hydrazones from Diazo-compounds 251 
 
 COOH COOH 
 
 0=0 + H2N-NH0 = C=N-NH^ + H2O. 
 COOH COOH 
 
 It was therefore evident that in one or other of these reactions an intramolecular 
 change occurred. The question of the structure of these bodies was for a long 
 time disputed. It is certainly a strong argument in favour of the azo-formula 
 that they are all brightly coloured, as are all azo-compounds but no hydrazones. 
 On the other hand, if the acetoacetic ester derivative, for example, is an azo- 
 body, 
 
 an analogous compound, 
 
 CO-CHg 
 
 ^•N=N.CH 
 
 COOEt 
 
 COCH3 
 
 0-N=N-C-CH3 , 
 
 COOEt 
 
 should be formed from methyl-acetoacetic ester. But as a fact methyl-acetoacetic 
 
 ester reacts with diazobenzene in quite a different way. The acetyl group splits 
 
 /OH 
 off and the phenyl-hydrazone of pyroracemie acid, 0-NH-N==C\p j-, j^-p,, , is produced. 
 
 The breaking of the carbon chain is most easily explained if we suppose that 
 the divalent hydrazo-residue has to make a place for itself on the carbon by 
 turning out not only the hydrogen but also the acetyl. It will be shown, 
 however, that in bodies of this type the acyl groups have unusual mobility. It 
 has been advanced in favour of the hydrazone structure that these bodies give 
 the Billow reaction for hydrazones ; but in strong sulphuric acid, which is used 
 in this reaction, tautomeric change is obviously probable. 
 
 The mechanism of these reactions, and the nature of the products, have been 
 elucidated by Dimroth. The bodies which react with diazobenzene in this way 
 are all such as exhibit keto-enolic tautomerism. Dimroth ' selected cases (such 
 as his own triazol derivatives, Claisen's mesityl-oxide-oxalic ester and triacyl- 
 methanes, and Knorr's diaceto-succinic ester) where the two tautomers can be 
 separated, and where it is known that under the conditions of the reaction one 
 tautomer does not go over into the other ; and he showed that in every instance 
 the enol form reacts with diazobenzene and the ketone does not, and further 
 
 that the enol reacts even in bodies of the type t>/C=C<(tj , where there is no 
 
 hydrogen attached to the carbon next to the COH group. The combination 
 with the diazo-compound may thus take place in either of two ways: — 
 
 (1) 11" + HON=N-R -^ l^OH _^ ~Y ; 
 
 ^ ' -C- -'C-N=:NE -C-N=NR ' 
 
 I I 
 
 -COH -CON=N.E -C=0 
 
 (2) 11 + HON=N-R 
 
 Further investigation'' showed that the reaction takes the second course. 
 For example, we may take the case of tribenzoyl-methane. The ketp-form of 
 
 > Ber. 40. 2404, 4460 (1907). " Dimroth, Hartmann, Ber. 41. 4012 (1908). 
 
252 
 
 Hydrazine Derivatives 
 
 this will not react. The enol form reacts easily, forming a yellow body, which 
 should be an 0-azo-compound with the structure 
 
 0-C-O-N=N-0 
 
 0-CO^II 
 
 0-CO^^ 
 For if it is boiled with alcohol it breaks up into tribenzoyl-methane, nitrogen, 
 and benzene, while the alcohol is oxidized to aldehyde. This is the behaviour 
 of a diazo-ether, as are also its other properties : it couples with a-naphthol, it 
 gives a diazonium salt with hydrochloric acid, and it reduces to phenyl-hydrazine, 
 in all cases with regeneration of tribenzoyl-methane. It should, however, be 
 noticed ^ that all these reactions only show that the body is not a C-azo-com- 
 pound : they do not prove it to be an 0-azo-compound. They are quite compatible 
 
 . , . , . ,. -C-ONE 
 
 with its bemg a diazomum salt, 111 . 
 
 N 
 
 If the dry substance is heated, it first turns red, and then colourless. The 
 red substance is unstable, but can be isolated. It must be an N-azo-derivative ; 
 the N2 group is firmly attached, and the body will neither couple with naphthol 
 nor form a diazonium salt with hydrochloric acid. On reduction it gives 
 a leuco-compound, which is the corresponding hydrazo-body. The white sub- 
 stance, made by heating the red or by acting on it with acid, is a hydrazone, 
 and on reduction gives benzanilide. The whole reaction is therefore : — 
 
 C-O-'N-'N-cp 6=0 
 
 CO CO CO CO 
 
 
 
 Azo- (red). Hydrazone (colourless). 
 
 On treatment with sodium ethylate all these three compounds split off 
 
 benzoic acid and form the phenyl-hydrazone of diphenyl-triketone. We may 
 
 suppose that the reaction takes place thus: — 
 
 yo 
 
 
 
 CO 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 C=0 
 
 
 c=o 
 
 C=N-N.0 
 
 C-0-N= 
 
 11 
 
 =N-0 
 
 -> 
 
 C-N=N-0 
 
 -» 
 
 CO^^O 
 
 
 
 co^^o 
 
 
 
 
 CO CO 
 
 
 
 
 
 
 
 
 
 
 i 
 
 
 
 i 
 
 
 i 
 
 
 
 
 
 
 
 
 
 
 
 
 
 f^ 
 
 
 ^=0 
 
 C-O.N= 
 
 ::N-0 
 
 -* 
 
 H-C-N=:N-0 
 
 -^ 
 
 C=N-NH0 
 CO 
 
 CH 
 
 
 
 T 
 
 
 T 
 
 
 
 
 
 
 
 
 _ 
 
 
 
 
 
 
 Cf. Auwers, Ser. 41. 4304 (1908). 
 
Hydrazones from IHazo-compounds 253 
 
 the analogous series of changes taking place after the loss of one benzoyl group, 
 with the difference that the reactions which require a measurable time in the 
 former instance now occur instantaneously. 
 
 If we start with dibenzoyl-methane, we must suppose that the lower series 
 of compounds are formed, and that the only product which can be isolated is 
 the hydrazone. 
 
 These experiments show that as far as the acylated compounds are concerned, 
 the hydrazone is the stable form. All attempts to convert the hydrazone back 
 into the azo-body were unsuccessful. We may, therefore, fairly assume that 
 in those compounds which have a hydrogen atom instead of the aeyl group the 
 hydrazone is still the stable form, and that the only difference will be that in 
 this case the change will be more rapid. Hence the compounds formed by the 
 action of diazo-bodies on substances with an acidic methylene group must be 
 regarded as hydrazones. 
 
 This result, which seems beyond dispute, is particularly remarkable in view 
 of the fact that, as we shall see later, in the aromatic series (oxyazo-compounds) 
 the reverse is the case, and the stable form is the azo. 
 
 The readiness with which azo-compounds and hydrazones go over into one 
 another (of which we have already had an example in ethane-azo-benzene and 
 acetaldehyde-phenyl-hydrazone) is shown in the derivatives obtained from diazo- 
 benzene and triacyl-methane.^ If they are allowed to stand in indifferent 
 solvents or heated above their melting-points, they lose their colour and pass, 
 with the migration of an acyl group, into substances which are undoubtedly 
 acyl-hydrazones : — 
 
 0-N=N-C(CO.CH3)3 -> (,jj^ ^,g>N-N=C(CO.CH3)2. 
 
 This mobility of the acyl group weakens the argument from the behaviour of 
 methyl-acetoacetic acid. 
 
 The formulation of these bodies as hydrazones brings out the resemblance 
 which exists between hydroxylamine and phenyl-hydrazine on the one hand, 
 and nitrous acid and diazobenzene on the other. We have seen that there are 
 two ways of making oximes : the normal method, by the action of hydroxyl- 
 amine on a compound containing a carbonyl group: — 
 
 >C=0 H- H2NOH = >C=NOH -1- H2O, 
 and a second method, by the action of nitrous acid on bodies containing an 
 acidic methylene: — 
 
 >C=H2 + 0=NOH = >C!=NOH + H2O. 
 The two are so to speak reciprocal ; in the first the doubly linked oxygen is on 
 the carbon and the Hg on the nitrogen ; in the second the oxygen is attached to 
 nitrogen and the hydrogen to carbon. 
 
 Exactly the same relations hold in the case of the hydrazones, which differ 
 from the oximes only in having the OH of the :NOH replaced by NH^. Corre- 
 sponding to the normal formation of oximes from ketone and hydroxylamine is 
 the formation of hydrazones from ketone and phenyl-hydrazine : — 
 >C=0 + HgN-NH^ = >C=N-NH^ + HjO, 
 
 ' Dimroth, Hartmann, Ber. 40. 4460 (1907). 
 
254 
 
 Hydrazine Derivatives 
 
 and corresponding to the action of nitrous acid on the methylene group that of 
 diazobenzene, which also only occurs when this group has acidic properties, and 
 which may be written : — 
 
 )CH2 + O:N-NH0 = >C:N-NH0 + H2O, 
 so as to bring out the analogy. A further point of resemblance is that in both 
 classes of reactions — with nitrous acid and with diazobenzene — we get under 
 similar conditions a breakage of the carbon chain, as is shown in the acetoacetic 
 ester group : — 
 
 With acetoacetic ester : 
 
 
 
 CHa-C^O CHg-C^O 
 
 
 CH3C=0 
 
 C=NOH «- CH2 
 
 -» 
 
 C=N-NH0 . 
 
 COOEt COOEt 
 
 
 COOEt 
 
 With acetoacetic acid : 
 
 
 
 CHs-C^O CH3-C=0 
 
 
 CH3-C=0 
 
 H.C=NOH <- CH2 
 
 — > 
 
 H.C=N-NH^ . 
 
 + CO2 COOH 
 
 
 + CO2 
 
 With methyl-acetoaeetic ester : 
 
 
 
 CH3.COOH OHa-C^O 
 
 
 CH3COOH 
 
 CH3-C=N0H *- CHg-C-H 
 
 — > 
 
 CH3-C=N-NH0 . 
 
 COOEt COOEt 
 
 
 COOEt 
 
 HYDEAZO-COMPOUNDS 
 
 The hydrazo-compounds are those hydrazine derivatives in which one 
 hydrogen of each NHg is replaced by an aromatic radical. Their behaviour 
 differs in many respects from that of the other hydrazine derivatives. 
 
 They were discovered by Hofmann in 1863. They are obtained by the 
 reduction of the nitro-compounds in alkaline solution, for example by the action 
 of zinc dust, or by electrolysis. A remarkable method of forming them is by 
 the action of chloro-dinitro- or chloro-trinitro-benzene (in which the chlorine is 
 mobile) on phenyl-hydrazine : — 
 
 (N02)2C6H3.C1 + H2N-NH0 = (N02)2C6H3-NH-NH^ + HCl. 
 They are colourless crystalline substances, and are in fact the leuco-eompounds 
 of the azo-series. They are soluble in alcohol, but not in water. They become 
 coloured in air, especially in the presence of moisture, and still more in that of 
 alkali, going over partially into the azo-compounds. They are neither basic nor 
 acidic, the amino-group being neutralized by the negative phenyl. 
 
 On energetic reduction they give two molecules of amine. In many cases 
 this occurs with especial ease at the moment of their formation, so that it is 
 often difficult to stop the reduction of the azo-compounds at the right point. 
 
 The hydrazo-compounds cannot be distilled, as on heating one molecule gives 
 up its two hydrogen atoms to another, forming the azo-derivative and two mole- 
 cules of amine : — 
 
 2 ^NH-NH0 = 0-N=N-^ 4- 2 ^NHj. 
 
 The hydrogen attached to the nitrogen is easily replaced, acetic anhydride, 
 
Hydrazo-compounds 255 
 
 for example, forming a di-acetyl derivative, phenyl isocyanate a urea, and nitrous 
 
 acid at low temperatures a very unstable yellow crystalline compound, which 
 
 0-N N-0 
 
 probably has the formula I I 
 
 NO NO 
 
 The most remarkable reaction of the hydrazo-compounds is their intra- 
 molecular change to benzidine and similar bodies. This occurs with great ease 
 on treatment with acids, so that if azobenzene is reduced in acid solution the 
 hydrazo is not obtained at all, but only the products of its rearrangement. The 
 reaction has already been sufficiently discussed. It is enough to repeat that it 
 goes in two stages, giving first a semidine (^-amino-diphenyl-amine) and then 
 a diphenyl derivative, either benzidine (di-p-diamino-diphenyl) or, if the para 
 position is occupied, a diphenyline, which is the corresponding ortho-par*- 
 diamino-compound. 
 
 Hydrazo-triphenyl-methane, ^gC-NH-NH-C^s, like so many other compounds 
 containing this radical, has a very peculiar behaviour.^ It is a comparatively 
 stable substance, and is not oxidized at all by the air, or by silver oxide. 
 Stronger oxidizing agents, such as potassium permanganate or chromic acid, 
 remove the hydrogen of the hydrazo-group, but the azo-compound, ^gC-N^N-C^s, 
 which we must suppose to be found, at once breaks up into nitrogen and tri- 
 phenyl-methyl, (f>^G, which appears as its peroxide, ^gC-O-O-C^g. 
 
 Thus the relations which hold with the simple aromatic derivatives (such as 
 hydrazobenzene) are here reversed. The hydrazo-compound is much more stable, 
 and the azo-body almost infinitely less so. Wieland expresses this by saying 
 that the weak affinity of the triphenyl-methyl group makes the attachment of 
 the hydrogen to the nitrogen in the hydrazo-compound much weaker than in 
 hydrazobenzene, while it is unable to hold the azo-group at all. 
 
 ' Wieland, Ber. 42. 1902 (1909). 
 
CHAPTEE XI 
 DIAZO-COMPOUNDS ^ 
 
 The next group of compounds consists of the derivatives of the hypothetical 
 diimide HN^NH. Here, as in the hydrazine compounds, we have to distinguish 
 two classes of bodies : those in which one of the two nitrogen atoms is attached 
 to a hydrocarbon group, and those in which both are so attached. The difference 
 here is, however, for whatever reason, far more striking than among the 
 hydrazines. 
 
 The diazo-compounds are those in which the Ng group is attached only on 
 one side to a hydrocarbon radical. 
 
 It is to be noticed that the so-called fatty diazo-compounds do not belong to 
 this group, since they have the Ng joined at both ends to the same carbon atom, 
 
 as indiazo-methane, CH2II. They will therefore be discussed among the ring 
 
 compounds. The very small number of true diazo-compounds of the fatty series 
 which are known are more conveniently dealt with along with the diazomethane 
 derivatives. 
 
 The study of the diazo-compounds has contributed to the development of 
 organic chemistry in an extraordinary degree and in a variety of ways. The 
 detection and isolation of bodies of so unstable and explosive a nature required 
 the highest experimental skill. The compounds were of the utmost service to 
 synthetical chemistry, owing to the great variety of gi'oups which they made it 
 possible to introduce into the benzene nucleus ; they are of immense technical 
 importance as the foundation of the vast group of azo-dyes ; and in recent times 
 the study of their constitution has thrown great light on some of the obscurest 
 questions of tautomeric change. 
 
 The diazo-compounds were discovered by Peter Griess '' in 1858. Their 
 importance for synthetic and technical purposes was soon realized, and their 
 more important reactions were established. Then, more than thirty years later, 
 the question of their constitution, which seemed to have been fairly satisfactorily 
 settled, was again raised, and became the subject of a prolonged controversy 
 between Hantzsch and Bamberger, in which both sides exhibited the highest 
 degree of skill, and Hantzsch in particular developed a new and most fruitful 
 method — the electrochemical — of attacking problems of this kind. The con- 
 troversy has finally been decided in Hantzsch's favour. It is clear that we 
 
 ' Cf. Hantzsch, Die Diazoverlindungen, Sammlung Ahrens, Bd. viii (Enke, Stuttgart, 1903) ; 
 Eibner, Zur GescJiichte der aromatischen Diazoverbindungen (Oldenbourg, Munich, 1903) ; Cain, 
 Chemiatry of the diazo-compounds (Arnold, 1908). See also Morgan, Brit. Ass. Beporfs, 1902, 
 p. 181, and Ch. News, 86. 189, 204, 213, 225 (1902). 
 
 ' Ann. 106. 123. 
 
Diazo-compounds 257 
 
 have to deal with an extremely complicated case of combined tautomerism and 
 stereoisomerism, the diazo-group being capable of assuming no less than four 
 different configurations. By his elucidation of these extraordinarily intricate 
 relationships Hantzsch has won a place in the first rank of chemists, and has 
 advanced our knowledge of the intimate nature of chemical reactions to a 
 remarkable degree. 
 
 Before discussing these questions, it wiU be well to consider briefly the 
 methods of formation and the more important reactions of the diazo-compounds. 
 In order not to prejudge the question of constitution, this will be represented 
 by the symbol Ng. 
 
 Methods of formation 
 
 1. By the action of nitrous acid on primary aromatic amines : either 
 
 (a) By passing the gaseous oxides of nitrogen evolved from starch or arsenic 
 trioxide and nitric acid into a solution of the amine salt. This is the usual 
 method for isolating the solid diazo-salts in the laboratory, and was the way in 
 which Griess originally prepared them. The gas is passed into a paste of the 
 amine salt and a little water, cooled with ice. The much more soluble diazo- 
 salt goes into solution, and is precipitated with alcohol and ether. A modificar 
 tion^ of this method is to diazotize the sulphuric acid solution with barium 
 nitrite (which can now be easily obtained), to filter off the barium sulphate, and 
 to add a mixture of alcohol and ether to the filtrate. 
 
 (6) By the action of alkyl (ethyl or amyl) nitrite on an acid solution of the 
 amine.'' This depends on the extraordinary rapidity with which the alkyl 
 nitrites break up in the presence of acid. It is sometimes used to make the 
 solid diazo-salt; the amine salt is dissolved in alcohol, and the alkyl nitrite 
 added. An almost quantitative yield of the solid diazo-salt may be got by 
 dissolving or suspending the amine salt in glacial acetic acid, cooling below 10°, 
 adding a slight excess of amyl nitrite, and precipitating with ether.' 
 
 (c) By the action of nascent nitrous acid on the amine salt, i. e. by adding 
 sodium nitrite to the acid solution. This is by far the commonest method 
 when, as is usually the case, an aqueous solution only is wanted, and the 
 precautions necessary in preparing the solid salt (which are due to its great 
 solubility in water and its insolubility in ether, making it impossible to separate 
 it from a soluble inorganic salt) are no longer required. But other precautions 
 have to be observed. It is important to use the exact equivalent of sodium 
 nitrite, and allowance must be made for the impurity which this salt always 
 contains. The nitrite solution may be titrated, or as a rough guide the salt 
 assumed to contain 98 per cent, of nitrite ; the older books recommended taking 
 the molecular weight of sodium nitrite as 72 instead of 69, but it is now 
 generally better than this. A common way is to run in the nitrite solution 
 until the liquid will darken starch-potassium iodide paper, showing the presence 
 of free nitrous acid. 
 
 The amount of acid used is also an important factor. Theoi-y requires two 
 
 ' Witt, Ludwig, Ber. 36. 4384 (1908). » Knoevenagel, Ser. 23. 2995 (1899). 
 
 ? HantzBoh, Jochem, Ber. 34. 3337 (1901). 
 
258 Diazo-covipounds 
 
 molecules of hydrochloric acid to one NHg. Generally more than this must be 
 used : commonly two and a half molecules, but in exceptional cases a stiU larger 
 excess. 
 
 The temperature must be carefully regulated. The ordinary rule is to cool 
 the amine solution beforehand to 5° or below, and not to allow it to rise aboye 
 10° during the reaction. It is then still kept cold, and used within twenty 
 minutes. But with some amines the diazotization goes slowly, and it is 
 necessary to work at the ordinary temperature, or even slightly above it. 
 
 If the amine salt is only sparingly soluble in water it can be suspended in 
 a state of fine division, as the diazo-salt is always much more soluble. When 
 the amine is so feebly basic that its salt is broken up by water, it is sometimes 
 suspended in the sodium nitrite solution and the acid run in ; or the base may 
 be dissolved in strong nitric acid, and the nitrous acid formed in the solution 
 by adding potassium pyrosulphite, K2S2O5, which reduces some of the nitric 
 acid : potassium nitrite cannot be used in this case, as it reacts too violently 
 with the strong acid.^ 
 
 Under ordinary conditions the diazotization of a base by nitrous acid takes 
 place with great rapidity. But Hantzsch and Schumann '' have been able, by 
 using very dilute solutions, to measure its velocity. They determined the 
 extent to which the reaction had proceeded by measuring the amount of un- 
 decomposed nitrous acid remaining by means of the colour which it gave with 
 zinc iodide and starch. They showed that all amines examined (both positively 
 and negatively substituted anilines) were diazotized with the same velocity ; 
 that this velocity was proportional to the product of the concentrations of the 
 nitrous acid and the base ; that it was somewhat increased by an excess of 
 hydrochloric acid, but that a greater excess than one equivalent of acid had no 
 further effect. This indicates that the reaction takes place between the undis- 
 sociated nitrous acid and the cation of the base. In the presence of hydrochloric 
 acid the nitrous acid, being a weak acid (about twice as strong as formic acid), 
 will be practically undissociated ; and the increase of velocity in presence of 
 excess of acid must be due to the disappearance of the hydrolysis of the aniline 
 salt, and the consequent increase of its cations. Similar results were obtained 
 by Schumann,' by determining the velocity from the fall of conductivity of the 
 solution. 
 
 Other methods for preparing the diazo-compounds are : 
 
 2. In some cases by treating the amine with another diazo-compound. Thus 
 nitro-diazobenzene and toluidine give nitranUine and diazotoluene. 
 
 3. By the oxidation of aryl-hydrazines with mercuric oxide. 
 
 4. By the action of a diazonium perbromide on an aryl-hydrazine : — 
 
 2 Ar-Na-Bra + Ar-N^Ha = 3 Ar-N^Br + 3 HBr. 
 This affords a convenient method for preparing the solid diazonium salt. The 
 perbromide is suspended in alcohol, an alcoholic solution of the hydrazine added, 
 and the precipitation completed with ether.' 
 
 5. By the reduction of the amine nitrate with zinc and hydrochloric acid. 
 This is really a modification of 1. 
 
 ■ Witt, Ber. 42. 2953 (1909). 2 Ber. 32. 1691 (1899). 
 
 ' Ber. 33. 527 (1900). < Chattaway, /. C. S. 1908. 958. 
 
Diazo-compounds : the Formation 259 
 
 6. By the action of hydroxylamine on a nitroso-derivative : — 
 
 ArNO + H^N-OH = Ar-Nj-OH + H^O. 
 In this reaction the normaP and not, as was previously supposed,'' the iso- 
 diazotate is formed. The importance of this result will be seen later, 
 
 7. A peculiar method ' is by treating certain nitroso-compounds with nitrous 
 acid, some of which is oxidized to nitric acid : — 
 
 Ar-NO + 3 HNO^ = ArNgONOg + HNO3 + HjO. 
 
 Properties 
 
 As the diazo-compounds do not, as a rule, require to be isolated, there is 
 comparatively little known about their individual properties, considering that 
 almost every primary aromatic amine which has been described has been 
 diazotized. But thanks to the work of Griess, and especially, in more recent 
 times, to that of Hantzsch and his pupils, we have considerable knowledge of 
 some of the simpler derivatives. 
 
 The mineral acid salts of diazobenzene are colourless crystalline solids, very 
 easily soluble in water, less in alcohol, and scarcely at all in ether. They have 
 the full character of salts, being highly ionized and not hydrolysed, as is shown 
 by their having, when pure, a neutral reaction. They are endothermic bodies,'' 
 and in the dry state explode when heated, and sometimes when struck. The 
 nitrate in particular is violently explosive, far more so than nitrogen iodide or 
 mercury fulminate ; the sulphate much less so. If the aqueous solution is 
 shaken with nitrobenzene, benzene, chloroform, &c., not a trace of the diazo- 
 compound is taken up ; but it can be easily and completely removed by phenol, 
 which turns a deep brown, possibly from the formation of diazobenzene-phenyl 
 ester, ^-NjO-*^.^ The aqueous solutions of diazo-salts are usually stable if they 
 are kept cold ; but on heating the diazo-nitrogen is evolved quantitatively and 
 a phenol is produced. 
 
 The haloid salts have a characteristic power of taking up two more atoms of 
 halogen to form perhalides, such as ^-Ng-Brg, which are analogous to the alkaline 
 perhalides, like KI3 . They are crystalline feebly explosive compounds, which 
 on treatment with ammonia give azides, e.g. ^-Ng. 
 
 The N2 group, though normally a strong base, is also capable of behaving as 
 a weak acid. If a concentrated aqueous solution of diazobenzene chloride is 
 poured into a large excess of very concentrated potash, potassium benzene 
 diazotate, ^NjOK, separates out. But this and similar bodies will be considered 
 in dealing with the question of constitution. 
 
 The next reactions to be discussed are those in which the diazo-bodies pass 
 into compounds of other types, and of these first those in which the diazo- 
 nitrogen is eliminated. It is these reactions which have given the diazo-com- 
 pounds their enormous importance for synthetical work. They all consist in 
 the evolution of the whole of the nitrogen of the Ng group as gas, while its place 
 
 ' Hantzsch, £er. 38. 2066 (1905). ' Bamberger, Ber. as. 1218 (1895). 
 
 ' 0. Fischer, Hepp, Ann. 255. 144 (1889) ; Hantzsch, Ber. 35. 894 (1902). 
 
 * Berthelot, VieUe, C. B. 92. 1076 (1881). 
 
 ' Hirsch, Ber. 23. 8705 (1890) ; Norris, Maointire, Corse, Am. Ch. J. 29. 120 (1903). 
 
260 Diazo-compounds 
 
 on the ring is taken by some other radical present, commonly that which was 
 previously attached to the diazo-group as an acid radical. There are scarcely 
 any of the simpler groupings which cannot in this way be introduced into the 
 ring under suitable conditions. 
 
 Thus we can replace the Nj by the following groups : 
 
 1. Hydroxyl. This usually occurs very readily, on keeping or heating the 
 aqueous solution of the diazo-salt. It is best to use the chloride or sulphate ; 
 if the nitrate is employed the phenol is liable to be attacked by the liberated 
 nitric acid. In some cases it is not suflScient even to boil with water, but the 
 salt must be heated with moderately concentrated sulphuric acid of boiling- 
 point 140-150°, or with copper sulphate solution. The presence of a chlorine 
 atom, or still more of a methoxy- or ethoxy-group in the ortho position to the 
 diazo, greatly hinders this reaction. 
 
 2. Alkoxyl, with the formation of phenol ethers. This change often occurs 
 readily on boiling with absolute alcohol, but the reaction in many cases goes in 
 a different way, the diazo-group being replaced by 
 
 3. Hydrogen, with the production of the hydrocarbon, the alcohol being 
 oxidized to aldehyde : — 
 
 ^•Na-Cl + CHs-CHjiOH = 0-H + HCl + N^ + CHs-CHO. 
 
 Which of these two reactions takes place depends entirely on the precise 
 conditions. Thus if diazobenzene-sulphonic acid ' (vrith the sulphur on the 
 nucleus) is treated with methyl alcohol under diminished pressure, only benzene- 
 sulphonic acid is formed ; but the same reagents under thirty atmospheres 
 pressure give only anisol-sulphonic acid, CH30-C6H4-S03H ;' while at the ordinary 
 pressure a mixture of the two is obtained. The formation of phenol ethers is, 
 however, the normal reaction. The presence of negative groups, especially in 
 the ortho position, promotes the replacement by hydrogen.'' 
 
 The production of the hydrocarbon is often effected by treating the alcoholic 
 solution or suspension of the amine salt with amyl nitrite, and then boiling. 
 
 The diazo-group can also be replaced by hydrogen in other ways. The body 
 can be reduced to the hydrazine, and then this converted into the hydrocarbon 
 by oxidation with copper sulphate or ferric chloride as already described. Or 
 the diazo-group may be replaced by iodine, and the product reduced, for example, 
 by distillation over zinc dust. A reaction which is often used for this purpose 
 is Friedlander's. This consists in treating the diazo-salt with an alkaline 
 stannous solution. It has been shown' that in this reaction part of the salt 
 is reduced to the hydrazine, which is then oxidized by the rest of the diazo- 
 compound to the hydrocarbon. 
 
 4. Halogens. This is effected in various ways : sometimes by heating with 
 excess of hydrochloric or hydrobromic acid, or by heating the platinichloride 
 with soda. 
 
 The best method is that of Sandmeyer,* which consists in heating the chloride 
 or bromide in presence of cuprous chloride or bromide : — 
 
 Ar-Na-Br = Ar-Br + Ng. 
 
 ' Eemsen, Palmer, Am. Ch. J. 8. 243 (1886) ; Keinsen, Dasliiell, ib. 15. 105 (1893). 
 " Eemsen, Palmer, 1. c. ; Cameron, ib. 20. 229 (1898). = Eibner, Ber. 36. 813. 
 
 * Ber. 17. 1633, 2650 (1884). 
 
Reactions of Diazo-compounds 261 
 
 The diazo-solution is usually added to the cuprous solution, which has been 
 warmed to a suitable temperature. Sometimes the amine is diazotized in 
 presence of the cuprous salt. In many cases it is better to adopt Gattermann's 
 modification,' and to treat the diazo-solution with copper powder, which de- 
 composes it in the cold. The powder is prepared by adding zinc dust through 
 a fine sieve to a cold saturated copper sulphate solution, and then washing the 
 precipitated copper with water and hydrochloric acid. 
 
 Iodine requires none of these elaborations. It is enough to pour the 
 diazonium sulphate solution into potassium iodide solution, or vice versa. 
 
 In order to replace the diazo-group by fluorine the diazo-salt is coupled with 
 piperidine to form the diazo-piperidide 
 
 Ar.N=N.N<gH.-CH,\(^^^ 
 
 and when concentrated hydrofluoric acid is poured over this it breaks up into 
 fluorbenzene, nitrogen, and piperidine hydrofluoride : — 
 
 ^•N^N-NCsHio + 2 HF = ^-F + CsHnN-HF + N^. 
 
 5. Cyanogen. This is a most important discovery of Sandmeyer's, enabling 
 us to introduce the carboxyl group into an aromatic compound. The diazo- 
 solution is poured into a hot solution of potassium cuprous cyanide, made by 
 the action of potassium cyanide on copper sulphate. 
 
 6. Various groups containing sulphur may be introduced by decomposing 
 the diazo-salt with potassium xanthogenate (potassium ethyl-dithiocarbonate 
 
 Q "IT" 
 
 S=C<^^-t;,,). If they are heated alone they give carbon oxysulphide and the 
 
 thio-ether Ar-S-Et; and if they are heated with alkali they give carbon oxy- 
 sulphide, alcohol, and the thio-phenol Ar-SH. 
 
 7. The nitro-group. This replacement is not usually required ; but it is 
 useful in certain cases. For example, /3-nitro-naphthalene cannot be obtained 
 by direct nitration, whereas ^-naphthylamine is easily got from /3-naphthol. In 
 order to introduce the nitro-group the diazo-solution is treated with an equivalent 
 of sodium nitrite, and the diazonium nitrite so formed is decomposed by cuprous 
 oxide. 
 
 8. Aromatic hydrocarbon groups, such as phenyl, can be introduced by 
 acting on the dry diazo-salt with the requisite hydrocarbon : if necessary alu- 
 minium chloride can be used as a catalyst, or in other cases an acyl chloride 
 Such as acetyl chloride '^ : — 
 
 ^•Na-CI + H-0 = 0-0 H- N2 -f HCl. 
 Again, if the brown solution obtained by extracting a diazo-solution with 
 phenol is heated,' oxy-diphenyls are produced : — 
 
 0-N2-C1 + H-CaHiOH = ^-CeH.OH + Ng -j- HCl. 
 This by no means exhausts the list of radicals which can be introduced in 
 place of the diazo-group, but it includes all those of any considerable importance. 
 Of the diazo-reactions in which the Nj group is not removed, some have 
 
 ' Ber. 23. 1218 (1890). " See refereneea in Meyer and Jaoobson, Lehrbueh, ii. 2. 17. 
 
 ' Norris, C. 03. i. 705. 
 1175 S 
 
262 Diazo-compounds 
 
 already been mentioned, such as the combination with certain methylene com- 
 pounds to form hydrazones. There are also others of great importance. Thus 
 on reduction the diazo-compounds give primary aryl-hydrazines : — 
 
 ArNa-Cl 4- 4 H = Ar-NHNHg + HCl. 
 On oxidation in alkaline solution they give nitroso-compounds and then nitra- 
 mines, such as diazobenzenic acid or phenyl nitramine, ^-NH-NOg. 
 
 The so-called coupling reactions are of immense technical importance. They 
 take place in two ways : — 
 
 1. With amines (fatty or aromatic) the first product is a diazoamino-com- 
 pound ^ : — 
 
 ^-N^-Cl H- 0-NH2 = ^•N=N-NH0 + HCl. 
 
 2. Directly in some cases (e.g. with some tertiary aromatic amines') and 
 always easily by a rearrangement of the first product an aminoazo-compound 
 is formed. This is the typical 'coi^pling', the Ng group attaching itself to 
 the nucleus to give an azo-derivative. Thus an amine gives an aminoazo- 
 derivative : — 
 
 ^•NjCl -t- H<3]SrH2 = HCl -f 0N=N-C>-NH2 ; 
 and in the same way phenols form oxy-azo-compounds : — 
 
 ^Na-Cl + HC^OH = HCl + 0N=N-C>-NH3. 
 These products will be further discussed under the azo-derivatives. 
 
 Constitution of the Biaeo-compoimds 
 
 The original formula proposed by Griess ' for the diazo-compounds regarded 
 them as derived from the hydrocarbons by the replacement of two atoms of 
 hydrogen by two of nitrogen, whence the name diazo. The salts were supposed 
 to be formed by the direct addition of acid : thus 
 
 Diazobenzene C5H4N2 : chloride C5H4N2-HC1. 
 
 Kekule* had no difficulty in overthrowing this view. He showed that in 
 all reactions in which the diazo nitrogen is removed a mono- substitution product 
 remains. Hence the diazo-salt must be written Ar-N2-Cl. This he expanded 
 into Ar-N=N-Cl for the salt, and Ar-N=N-OH for the free base. 
 
 In 1869 Blomstrand suggested another formula.^ He argued that the diazo- 
 aalts were strictly analogous to the ammonium salts, and therefore must contain 
 
 pentad nitrogen. Hence he wrote them Ar-N<^p|, , corresponding to Ar-N^p,^ . 
 
 The same view was put forward independently and almost at the same time by 
 Strecker,' but on a different ground — that the very unstable diazo-compounds 
 were so unlike the comparatively very stable azo-bodies, which were undoubtedly 
 Ar-N=N-Ar, that they could not be built up on the same type. Erlenmeyer' 
 suggested the same structure, also independently, in 1874. 
 
 ' For the occurrence of this reaction among the fatty amines seeDimroth, Ber. 38,2328 (1905), 
 tnd among the benzylamine bases, Goldschmidt, Holm, Ber. 21, 1016 (1888), 
 
 2 Haeussermann, Ber. 39. 2762 (1906), ^ Ann. 121. 257 (1862); 137. 39, (1866), 
 
 * Lehrl. d. org. Ohem., ii. 717 (1866), ' Chemie der Jetztzeit, 4, 272; Ber. 8. 51 (1875). 
 
 ' Ber. 4, 786 (1871), ' Ber. 7. 1110. 
 
Diazo-compounds : Constitution 263 
 
 Blomstrand and Strecker's views were, however, little regarded, and Kekule's 
 formula was generally accepted, especially after Emil Fischer's work' on the 
 hydrazines. Fischer showed that diazobenzene could easily be converted by 
 reduction into phenyl-hydrazine, whose formula he proved to be ^-NH-NHg. 
 This is to be expected from Kekul6's diazo-formula : — 
 
 ^■N=N-OH + 4 H -» ^-NHNHa, 
 ■while Blomstrand's formula (tautomerism being of course at this time excluded) 
 required that the product should be ^■NH2=NH. 
 
 The question remained in this position for over ten years. Much attention 
 ■was bestowed upon the diazo-compounds, but it was devoted entirely to working 
 out the practical details of the reactions in which the diazo-group was replaced 
 hy other groups, and it threw little or no light on the constitution of these 
 bodies. 
 
 In 1892 the question of constitution was again raised by v. Pechmann.^ On 
 the formation of hydrazones by the interaction of diazo-salts with methylene 
 •compounds, and on the general similarity of behaviour of the diazo-compounds 
 and nitrous acid, he based a new formula, that of a nitrosamine, Ar-NH-NO. 
 This view received additional support from the discovery of v. Pechmann and 
 Wohl ° that diazobenzene reacts with benzoyl chloride to give nitrosobenzanilide, 
 which is easily explained on the nitrosamine formula : — 
 
 0-N-NO 
 ^ ^ CO-0 
 
 but which on Kekule's theory requires us to suppose an elaborate addition process. 
 Two years later (1894) the problem entered on a new stage with the dis- 
 covery of isomers of the diazo-compounds, which was made simultaneously and 
 independently by Schraube and Schmidt,* by v. Pechmann and Frobenius,' and 
 by Bamberger.'' They found that potassium diazobenzene possesses entirely the 
 character of a diazo-compound, and in particular the power of coupling, for 
 example with /3-naphthoI, to give an azo-dye. But if it is heated with its 
 mother liquor (excess of concentrated potash) the precipitate no longer couples, 
 although after treatment with excess of acid it will again do so. This new body 
 was called the iso-diazo-salt, and as Schraube and Schmidt found that it reacted 
 with methyl iodide to give nitroso-methyl-aniline, ^-NCHg-NO, they assumed 
 that it had the structure 0-NK-NO, derived from v. Pechmann's nitrosamine 
 formula. In view of v. Pechmann's arguments for this structure, it is remark- 
 able that the iso-diazo-compound will not give hydrazones with aceto-acetic ester 
 and similar bodies. The argument from the reactions of the potassium salt is 
 one of a class which we know are not to be trusted ; and v. Pechmann showed 
 its weakness by proving that the silver salt of iso-diazobenzene gave with 
 methyl iodide the 0-ether of normal diazobenzene, ^-N^N-OOHg. 
 
 All this work on the iso-diazo-compounds was done in 1894. In the same 
 year Hantzsch ' produced a paper in which he reviewed the whole question, and 
 
 1 Ser. 8. 589, 1005 (1875) ; 9. 881 (1876). ^ Ber. 25. 3505 (1892) ; 27. 651 (1894). 
 
 » ■Wohl, JBer. 25. 8631 (1892). * Ber. 27. 514. = Ber. 27. 672. ' Ber. 27. 679. 
 
 ' Ber. 27. 1702 (1894)". 
 
 s2 
 
264 Diazo-compounds 
 
 offered a novel solution. He pointed out that there was a remarkable analogy 
 between the history of the diazo-compounds and that of the oximes. The 
 explanation oifered by Schraube and Schmidt of the isomerism of the former 
 corresponded exactly to that which Beckmann had given of the isomeric oximes ; 
 and it was possible to give a stereo-chemical explanation of the isomerism of the 
 diazo-compounds similar to that by which he and Werner had solved the oxime 
 problem. There are three analogous series of compounds which all admit of the 
 same stereoisomerism : — 
 
 ECH KCH EN 
 
 II : II : II • 
 
 ECH NOH NOH 
 
 Fumaric, maleic. Oxime. Diazo. 
 
 He further argued that the evidence offered in support of the nitrosamine 
 
 formula for the diazo-compounds was worthless, and that the normal and iso- 
 
 alkaline salts must have the same structural formula, and must therefore be 
 
 stereoisomers. The more easily decomposed normal compounds were assumed 
 
 to have the syn-formula, II , and the more stable iso-salts the anti-, H . 
 
 To sum up, the state of affairs in 1894, after the appearance of Hantzsch's 
 paper, was this. Two different series of isomeric diazo-compounds were known, 
 the normal and the iso ; and there were four different formulae in the field : — 
 
 1. Kekule's, Ar-N=N-OH (Diazo). 
 
 2. Blomstrand's, Ar.N<^^-rT (Diazonium). 
 
 3. v. Peohmann's, Ar-NH-NO (Nitrosamine). 
 
 Ar-N Ar-N 
 
 4. Hantzsch's stereoisomers of Kekule's formula, II and II (Syn- 
 
 and anti-diazo). 
 
 As will appear later, all these formulae are under some conditions correct. 
 
 It is not necessary to follow the historical development of the question any 
 further, or to describe the long controversy which ensued between Hantzsch and 
 Bamberger. Hantzsch modified his original views on several important points ; 
 but he has established his theory, with these modifications, on so firm an experi- 
 mental basis, that we need not consider the rival theories which have been 
 proposed. The views as to the constitution of the diazo-compounds which are 
 now accepted are in all essentials those of Hantzsch, and their more important 
 points are summarized in the following pages. 
 
 We will consider first the salts of the diazo-compounds with the mineral 
 acids. Here the evidence is conclusive in favour of Blomstrand's pentad 
 nitrogen formula, which Hantzsch has called the diazonium formula, to indicate 
 its analogy to that of the ammonium salts : — 
 
 Ar.N=N H-NBH, Ar-N=H3 
 
 I i ^ I ^ 
 
 X XX 
 
 Diazonium. Ammonium. 
 
 This was revived by Bamberger in 1895 in opposition to Hantzsch's stereo- 
 chemical views ; it was soon adopted by Hantzsch himself, to whom the most 
 important arguments in its favour are due. These consist in an elaborate com- 
 
Diazonium Salts 265 
 
 parison of the chemical and physico-chemical properties of these salts with those 
 «f the corresponding salts of the alkalies and the quaternary ammonium bases. 
 
 In the first place, it is obvious that the diazo-salts do not behave like com- 
 pounds of triad nitrogen. We have an example of triad nitrogen with a hydroxyl 
 group attached in hydroxylamine HgN-OH. This is not only a very weak base, 
 but when it forms salts with acids it does so by addition, giving, for example, a 
 hydrochloride, H2NOHHCI. The diazo-salts, on the other hand, are derived from 
 & very strong base, which in some of the substituted compounds is far stronger 
 than ammonia, and almost if not quite as strong as the alkalies. Also when it 
 forms salts it does not do so by addition, but by replacing the hydroxyl by an 
 acid radical. In both of these points it closely resembles a quaternary ammonium 
 hydroxide, such as ^N(CH3)30H. 
 
 Again, diazonium chloride and nitrate have a neutral reaction ; they are not 
 hydrolysed in aqueous solution, while they are ionized to the same high degree 
 as potassium chloride or nitrate, not only in aqueous but also in alcoholic 
 solution ^ ; and the cation has almost the same velocity as that of an alkali metal 
 or ammonium.'' The carbonates, like those of the alkalies, are soluble in water 
 and of a strong alkaline reaction. The double salts" resemble those of the 
 alkalies. The platinichlorides, (ArN2)2PtClc , have long been known ; and Hantzsch 
 has shown that the diazonium salts give cobaltinitrites and double chlorides with 
 mercuric chloride analogous to those of potassium and ammonium. Further, as 
 will be shown later, the cyanide forms a double salt with silver cyanide, 
 Ar-Ng-AgCyg, closely resembling in behaviour potassium silver cyanide, KAgCyj . 
 
 A direct determination of the afiinity constant of phenyl-diazonium hydrate 
 showed that while it is a decidedly weaker base than tetramethyl-ammonium 
 hydrate it is still nearly seventy times as strong as ammonium hydrate.' The 
 strength is greatly diminished by the introduction of negative groups like 
 bromine, but on the other hand it is enormously increased by positive groups 
 such as methyl or methoxyl.^ Thus anisol (CH30'C8H4-) and pseudocumene 
 {(CH3)3-05H2-) diazonium hydrates are, like the alkalies, such strong bases that 
 their strength cannot be exactly measured. 
 
 Moreover, it has been shown that the only ions into which the hydrate breaks 
 up are Ar-Nj and OH ; that is to say, that it is a true hydroxide. This excludes 
 the nitrosamine formula, and leaves us only the choice between Ar-N=N-OH and 
 
 Ar-N^^TT ; and the highly basic character of the compounds makes it certain that 
 
 they must contain pentad nitrogen. 
 
 A further point of resemblance between the diazonium and the ammonium 
 
 compounds is found in the perhalides. These are formed by the action of the 
 
 halogens on the diazonium halides, and have the composition Ar-N2-Hal3. They 
 
 Ar-N— N-Br 
 were originally written on the type I I • But it has been shown by 
 
 Br Br 
 
 » Hantzsch, Davidson, Ber. 31. 1612 (1898). 
 
 ' Hantzsch, Ber. 28. 1740 (1896) ; Hantzsch, Davidson, 1. c. 
 
 3 Hantzsch, Danziger, Ber. 30. 2629 (1897). * Hantzsch, Davidson, 1. c. 
 
 = Hantzsch, Engler, Ber. 33. 2147 (1900). 
 
266 Diazo-compounds 
 
 Hantzsch ' that they are strictly analogous to the alkaline perhalides, such as KI3 . 
 The diazonium perhalides, like those of the alkalies, are coloured, are not very 
 soluble in most solvents, and readily break up again into the simple halide and 
 free halogen. The resemblance is most marked between the diazonium com- 
 pounds and those of caesium. Taking the three halogens, chlorine, bromine, and 
 iodine, and neglecting structural or stereoisomerism, ten compounds X-Halg are 
 possible. Of these in the diazonium series nine are known — all but -CI3, and in 
 the caesium series eight— all but -CI3 and -IgCl. The similarity in colour is most 
 striking. In both series you start with a blue-black almost opaque tri-iodide, and 
 the colour gets lighter as the atomic weight of the halogen diminishes, passing 
 through red, orange (-Brg), and yellow, and ending with the pale yellow -ClaBr. 
 la colour and in stability the corresponding members of the two series are almost 
 identical. A further argument for their being diazonium compounds and not 
 tri-haloid hydrazines is that on the latter hypothesis they should give structural 
 isomers according to the order ia which the halogens are introduced : — 
 
 Ar-N=N-Cl _ Ar-N-N-Cl 
 + Br-I ^ BrI ' 
 
 ArN=NBr ArN-NBr 
 
 and = 11.; 
 
 + CM ClI 
 
 but no such isomers have ever been obtained. 
 
 It is clear then that a solution of a diazonium salt closely resembles that of 
 an alkali or compound ammonium salt ; and that just as the latter contains- 
 potassium ions or X-N^Hg ions, so the former must contain diazonium ions, 
 Ar-N=N. But it does not necessarily follow from this that the undissoeiated 
 
 part consists wholly of the compound Ar-N<^-j^ (where X is hydroxyl or the acid 
 
 radical). An aqueous solution of ammonia contains the ions NH4 and OH j 
 but we know that the undissoeiated NH4OH breaks up in the solution, to the 
 extent of about two-thirds, into ammonia and water, so that we have an equili- 
 brium : — 
 
 NH'^ + OH' ;?i NH4OH ^ NH„ + K^O. 
 
 Now in the case of the undissoeiated diazonium hydroxide we cannot have 
 this precise change with loss of water, but there is every probability of a change 
 of another kind. It has been shown that all quaternary ammonium hydroxides 
 in which the nitrogen is either doubly linked to carbon, or forms part of a ring, 
 are liable to a peculiar transformation in the presence of a base, the hydroxyl 
 migrating from the nitrogen to another part of the molecule, forming a substance 
 which is no longer a base. In other words they are pseudo-bases. Of the first 
 kind we have already had an example in the triphenyl-methane dyes, where the 
 highly ionized salt, X2-C=CeH4=NH2-Cl, on treatment with alkali first gives the 
 corresponding hydroxide Xa-C^CjH^^NHg-OH, which then changes spon- 
 taneously into the undissoeiated carbinol X2C(OH)-C|3Hj-NH2. That is, the 
 nitrogen becomes triad through the migration of the hydroxyl to another atom, 
 
 1 Ber. 28. 2764 (1895). 
 
Diazotates 267 
 
 and a non-basic substance is formed. Another such case is that of the pyridonium 
 salts, where the base changes over into the non-basic oxy-pyridine : — 
 
 H^ (in 
 
 
 HC CHOH 
 
 N 
 
 -^ 
 
 \/ 
 
 N 
 
 R OH 
 
 
 i 
 
 Now diazonium hydroxide is a substance peculiarly adapted for this kind of 
 ionization-isomerism. The basic nitrogen is triply linked to another nitrogen 
 atom, and it is quite probable that in the presence of a base the hydroxyl (of 
 the undissociated portion) may, at any rate partially, migrate to this : — 
 
 ArN^Qjj -> Ar-N=N-OH, 
 
 the body changing, to use Hantzsch's nomenclature, from a diazonium to a diazo- 
 hydrate. This change is not hypothetical. It is absolutely required to explain 
 the extraordinary behaviour of diazobenzene hydrate. It has been shown that 
 the strongest argument for the diazonium structure of the dissociated part of 
 the molecule is its powerfully basic character. A diazo-hydroxide would be 
 a weak base, like hydroxylamine. But though it is so strong a base, yet it is 
 capable of forming a salt with potash, the normal diazotate. If the solution of 
 the base is treated with a solution of potash, there is a considerable evolution of 
 heat, and measurements of conductivity show that with a not veiy large excess 
 of potash it is completely converted into such a salt. That is to say, this body, 
 which behaves as a base in some derivatives as strong as potash, is also capable 
 of acting as a rather weak acid. Now we do know of some very weak bases which 
 are able to act also as weak acids, X-OH giving either cations X or anions X-0' ; 
 these are the so-called amphoteric electrolytes, such as the hydroxides of 
 aluminium, zinc, lead, and tin. But these bodies differ from the diazo- 
 compounds in being only weak bases and excessively weak acids, their alkaline 
 salts being often so unstable that they cannot be obtained in the solid form. 
 
 A compound, Ar-N^j-.T7-, would not be analogous to such compounds as these, but 
 
 rather to an alkaline oxide (anhydride) like KgO. It could only exist in the 
 absence of water, whereas the diazotates are distinct salts, derived from a weak 
 but not excessively weak acid, Ar-NgOH. This unique behaviour of diazobenzene 
 hydrate must be due to isomerism. The substance which reacts as a strong 
 base must have a different structure from that which reacts as a weak acid ; 
 
 and if so, it is clear that the base must be the diazonium compound Ar-N<^Qjj > 
 
 analogous to ammonium hydrate, and the weak acid the diazo-body Ar-N=N-OH, 
 corresponding to hydroxylamine. This hypothesis will also account for the 
 unusual behaviour of sodium normal diazotate. This body, as the salt of a 
 weak acid, is hydrolysed in water to an extent which can be determined by 
 conductivity measurements. But the hydrolysis increases with the dilution 
 enormously faster than in any other known case. It is hard to see how this 
 can be explained except by supposing that the equilibrium is destroyed by the 
 weak acid first liberated going over into the ions of the strong diazonium base. 
 
268 Diazo-compounds 
 
 So far, then, we may conclude that the mineral acid salts of diazobenzene 
 have the diazonium structure, but that on addition of a base the hydrate which 
 is first formed changes over to give the alkaline salts of the weakly acidic diazo- 
 hydrate Ar-N=N-OH, which are the normal diazotates. 
 
 But, as Hantzsch pointed out in his original paper, a substance of this latter 
 
 Ar-N , Ar-N 
 structure should exist in two stereoisomeric forms, II and jJ^Qg-- ^^"^ 
 
 in this case of the alkaline diazotates we have two series of compounds — the 
 normal and the iso — for which two types of formulae must be found. We have 
 seen that the normal salts cannot be true diazonium compounds, because so 
 strong a base could not form salts with another strong base. This is even more 
 obviously true of the iso-diazotates, since the acid from which they are derived 
 must be stronger than that which gives the normal compounds, as they are less 
 hydrolysed. There are therefore three formulae possible for the normal and 
 iso-diazotates : — 
 
 ArN ArN Ar-NK-N:0. 
 
 II II 
 
 KON NOK 
 
 Syn-diazo. Anti-diazo. Nitrosamine. 
 
 The third of these formulae is on various grounds very improbable. The 
 isomerism of the alkaline salts is repeated in other classes of derivatives, such as 
 the cyanides and sulphonates ; and any explanation must apply to these as well. 
 But it is evident that the cyanides and sulphonates cannot be given a structure 
 analogous to that of the nitrosamines. Moreover, as wUl be shown later, the 
 nitrosamines can actually be prepared, and they have quite different properties. 
 Further, a more detailed examination of the chemistry of the diazo-compounds 
 proper — as opposed, that is, to the diazonium derivatives — shows that their 
 isomerism resembles that of stereo- rather than structural isomers. The normal 
 and iso-diazotates (the alkaline salts) are colourless salts, which both form in 
 aqueous solution the isomeric ions Ar-NjO', but are also to some extent hydrolysed, 
 the normal more than the iso. Their reactions are practically all similar in kind, 
 but the iso react much less readily in most cases. Both classes are easily 
 reduced to hydrazines, and are converted by benzoyl chloride into nitroso- 
 anilides, Ar-N(CO-0)-NO. Both give on oxidation the salts of a nitramic acid, 
 as diazobenzenic acid, Ar-NjO-OM. The same resemblances are shown by the 
 isomeric cyanides and sulphonates. This indicates that the same differences 
 condition the isomerism in all three cases (of the diazotates, cyanides, and 
 sulphonates) ; and the fact that these are differences rather in degree than in 
 kind points to stereoisomerism. A further consideration shows that they can 
 easily be explained on stereo-chemical grounds. 
 
 If we draw the structure of the syn- and anti-diazo-compounds on the basis 
 of the usual tetrahedral models, we get figures corresponding to those for fumaric 
 and maleic acids. But there is a remarkable difference between the two cases. 
 In the latter case the two doubly linked carbon atoms are in both modifications 
 more or less symmetrically related to the rest of the molecule. But in the diazo- 
 compounds this is not so. In the syn-compound the two nitrogen atoms must 
 lie wholly on one side and the other two groups on the other. Moreover, if we 
 
Syn- and Anti-diazo 269 
 
 suppose, as is generally done, that the three valencies of the doubly linked 
 trivalent nitrogen act from one point of a tetrahedron towards the other three, 
 it is clear from the model that the position of least strain for the syn-compound 
 is that in which the two nitrogen atoms are inclined towards one another, and 
 the two groups attached to them are brought close together ; and that even here 
 there will be considerable strain in Baeyer's sense ; while in the anti-compound 
 no such strain will exist : — 
 
 There is thus an essential difference between the two classes of isomers in 
 the diazo-compounds which does not occur either in the case of fumaric and 
 maleic isomerism or in that of the oximes : namely, that the syn-compounds are 
 necessarily in a state of much greater strain than the anti-. They will therefore 
 be the less stable class, since the strain will endeavour to relieve itself. This it 
 can do in two ways : either by changing into the more stable anti-isomer, or by 
 splitting off the nitrogen entirely : — 
 
 Ar-Na-X = Ar-X + Nj. 
 We should further expect to find that the syn-form was only capable of existence 
 ■when the second group X attached to it was a small one ; as otherwise the two 
 nitrogen atoms would not be able to turn over towards one another sufficiently. 
 And this agrees with experience : the only comparatively stable syn-compounds 
 are the diazotates and the cyanides (X = OM or CN). The syn-sulphonates 
 {X = 8O3M) are much less stable, while compounds like the syn-diazoamino 
 bodies (X = NH-Ar) appear scarcely to exist at all. 
 
 There are several methods for determining whether a given diazo- compound 
 belongs to the syn- or the anti-series. If both isomers are known, we can say 
 that the syn-, as it has the largest energy content, must be the most reactive of 
 the two. It will be the more explosive, and will be the more readily reduced to 
 hydrazines and oxidized to nitramines and so forth. All these properties are 
 those of the normal as opposed to the iso-diazotates ; so that it is clear that the 
 normal are the syn- and the iso the anti-isomers. It is also found that the syn- 
 are the more soluble and have the lower melting-points. 
 
 A reaction which is much used for distinguishing between the two classes is 
 that of coupling, i.e. of combining with amines or phenols to give azo-dyes. 
 This always occurs more easily with the syn-compound. Indeed, it was at first 
 supposed that the anti-compounds did not couple at all, but Hantzsch has shown 
 that they always do so, though often very slowly. In other cases, however, 
 they couple quite rapidly, so that this test is not absolutely trustworthy except 
 
270 Diazo-compounds 
 
 where both isomers are known ; but then it is invariably found that the syn- 
 compound couples with the greatest rapidity. 
 
 Again, the typical diazo decomposition into Ar-X and nitrogen is, as has- 
 been pointed out, peculiar to the syn-series ; and accordingly we find that it is 
 only the normal diazotates which break up easily in this way; the anti- or iso- 
 compounds decompose much more slowly, and require a higher temperature, 
 which probably converts them first into the syn-isomers. 
 
 It is evident, then, that the alkaline diazotates are all true diazo-compounds, 
 and that the so-called normal series, which are the first product of the action of 
 alkali on a diazonium salt, are the syn-compounds, while the iso-diazotates, into- 
 which these can be converted, belong to the anti-series. 
 
 There remains to consider the constitution of the normal and iso-diazo- 
 hydrates, which are got by the action of acid on the corresponding diazotates. 
 We have already seen that the diazonium hydrate, which may be obtained in 
 solution by treating a solution of a diazonium salt with silver oxide, is a strong 
 base. But we should expect that we might be able to get two isomers of this, 
 the syn-diazo-hydrate and the anti-diazo-hydrate, by the action of acid on the- 
 two diazotates. It was formerly supposed that this was the case ; but Hantzsch 
 has shown that the true syn-diazo-hydrates cannot be isolated at all, partly 
 because of the ease with which they break up into phenol and nitrogen, and 
 partly because they change with great readiness into their isomers, the diazonium 
 hydrates and the anti-diazo-hydrates. We may, however, infer from the hydro- 
 lysis of the syn-diazotates in solution that they are weak acids. The anti-diazo- 
 hydrates are obtained by precipitating the solution of an anti-diazotate with the 
 calculated quantity of acetic acid at a low temperature. They are colourless- 
 crystalline compounds, which are very unstable, and are converted even by 
 dissolving in water into the nitrosamines : — 
 
 ArN Ar-NH 
 
 NOH "^ N=0' 
 
 It is, however, possible, by taking proper precautions, to isolate and examine 
 these anti-diazo-hydrates. They behave as true acids. They react as hydroxyl 
 compounds with phosphorus pentachloride, with acid chlorides, and with phenyl 
 isocyanate ; and they give the ammonia reaction. They also couple fairly 
 rapidly with phenols. 
 
 If an anti-diazotate is treated with acid without special precautions, or if an 
 anti-diazo-hydrate is allowed to stand, the nitrosamine is produced. Owing to- 
 this fact, the diazotate behaves at the ordinary temperature as the salt of a pseudo- 
 acid. Its solution, when treated with an equivalent of hydrochloric acid, has the 
 conductivity of the potassium chloride which it contains. The nitrosamines 
 are yellow crystalline compounds, rather sparingly soluble in water, and easily 
 in organic solvents. As pseudo-acids they have a neutral reaction, but neutralize 
 an alkali ; and they do not give the ammonia reaction. They do not react with 
 phosphorus pentachloride in the cold, and hence do not contain a hydroxyl 
 group. They couple very slowly with phenols. Mineral acids in the presence 
 of water convert them into diazonium salts, but dry hydrochloric acid combines- 
 
Diazo-hydrates 
 
 271 
 
 with them in ethereal solution without isomeric change to give the nitrosamine 
 salt Ar-NH-NOHCl. 
 
 It is thus evident that the mineral acid salts of the diazo-oompounds have 
 the diazonium structure, and when treated with an equivalent of alkali yield 
 
 a solution of the diazonium hydrate, which contains the ions Ar-N*^ and OH', 
 
 while the undissociated part consists partly of the true diazonium hydrate 
 
 /^N Ar-N 
 
 Ar-N^Qjj and partly of the feebly acidic syn-diazo-hydrate II . This explains 
 
 why the same normal diazo-hydrate solution is completely converted by a mineral 
 
 acid into a diazonium salt, and by excess of alkali into a normal (syn-) diazotate. 
 
 If the solution is heated with excess of alkali, the syn-diazotate is converted into 
 
 the anti-(iso)-diazotate : — 
 
 ArN Ar-N 
 
 II -> II , 
 
 KON NOK 
 
 and when this is treated with an equivalent of acid it gives the very unstable 
 anti-diazo-hydrate, which rapidly goes over into the non-acidic-nitrosamine : — 
 
 Ar-N 
 II 
 NOH 
 
 It is to be noticed that whereas the syn-diazo-hydrate is converted by a mineral 
 acid into the diazonium salt too quickly for the progress of the reaction to be 
 observed, the iso-diazo- hydrate only changes slowly, as may be seen by measuring 
 the conductivity. 
 
 Hantzsch has expressed these rather complicated relationships in the form of 
 a diagram, which renders them somewhat clearer : — 
 
 Ar-NHNO. 
 
 
 
 Diazonium salt. 
 Ar-N-Cl 
 
 vf ill t 
 
 
 
 Ar-ir -^ r-Ar-N-OH- 
 HO-N «- LhO-NH . 
 
 Na 
 
 Diazo 
 hyd 
 
 -^ Ar-1^ 
 
 1 
 «- I' 
 
 1 ^V 
 
 •OH 
 
 nium 
 rate. 
 
 roH 
 
 ■ 
 
 Ar^N Ar^N-H 
 NOH ~* N=0 
 
 Syn-diazo- 
 hydrate. Co-'"' si- 
 lt I'oSlaUi'cal . 
 
 
 Anti- 
 
 diazo. Nitros- 
 , Hv^"'^"'- amine. 
 
 Na 
 
 1 
 
 •OH 
 
 \ 
 
 
 
 HCl 
 
 Ai 
 
 NaO 
 
 Syn-di 
 
 (unsi 
 
 ■N Stereo change 
 
 A 
 
 r-N 
 II 
 N^ONa 
 
 i-diazotate 
 stable). 
 
 •N (heating, &c.) 
 
 azotate 
 
 able). 
 
 
 Ant 
 ( 
 
 These are the main outlines of Hantzsch's theory of the diazo-compounds. 
 
272 Diazo-compounds 
 
 There are a large number of smaller points, some of which deserve further 
 consideration. 
 
 In the first place, there is the question of the structure of the solid salts of 
 the mineral acids. What has hitherto been discussed is their structure in 
 solution. Now they differ very greatly both in colour and in stability, and these 
 two properties always go together, the darkest salts being the most readily 
 exploded. Thus, to take a striking example, 2,4,6-tribrom-phenyl-diazonium 
 bromide, 
 
 Br 
 
 Br 
 
 is deep orange-red and excessively explosive, while the corresponding nitrate 
 is colourless and can scarcely be made to explode at all. If we tabulate the 
 compounds in the following way : — 
 
 Trimethyl-benzene diazonium Chloride Bromide Thiocyanate Iodide 
 
 Methoxy- ,, ,, 
 
 Benzene ., 
 Monobromo- ,, ., 
 
 Dibromo- ,, ,, 
 
 Tribromo- „ ,, ~] j 
 
 we find that travelling in the direction of the arrows, either downwards or 
 from left to right, we always get an increase of colour and of explosiveness.' 
 The less basic the diazonium compound is, the less explosive and coloured its 
 salts are ; and as the atomic weight of the halogen increases, these qualities 
 increase also. This indicates a possible difference of constitution, and Hantzsch 
 ingeniously suggests that the solid is a solid solution of the diazonium salt and 
 the syn-diazo-compound. We know that the diazonium salt changes very easily 
 into the diazo ; we know that in the case of the cyanides the diazonium is less 
 reactive and is colourless, while the syn-diazo is more reactive and coloured ; 
 and we, know that the more basic the group is the stronger is the tendency 
 under all conditions to form the diazonium compound. Further, it is obvious 
 that a diazo-halide, having the halogen attached to trivalent nitrogen, would 
 share the explosiveness of the nitrogen halides. It is remarkable that the equi- 
 librium between the two forms in the solid solution seems to be displaced in 
 the diazonium direction by a lowering of temperature. Many of the halides 
 are almost colourless at —60°, and their colour deepens continuously as the 
 temperature rises.' 
 
 This conception of a solid tautomeric substance existing as a solid solution 
 of the two forms has subsequently been extended by Hantzsch to other cases, 
 and in particular to the nitrophenols (p. 174) ; and he has accumulated 
 a considerable amount of evidence in its favour. 
 
 The diazonium salts have a remarkable power of exchanging the acid radical 
 for a substituent in the ortho or para position on the nucleus. This was first 
 
 ' Hantzsch, Ber. 33. 2179 (1900). ^ Hantzsch, Euler, Ber. 34. 4166 (1901). 
 
Solid Diazoniam Salts 273 
 
 discovered by Hantzsch and Hirsch ' in the case of the thiocyanate of j9-chloro- 
 diazobenzene : in alcoholic solution the chlorine and the thiocyanogen groups 
 change places : — 
 
 C1-<I>-N-S-C=N -* N=C-S-0-N-Cl. 
 
 N N 
 
 The chlorides^ behave in the same way, though less readily, the change only 
 occurring when there are at least two halogen atoms in the ortho or para 
 position : halogens in the meta position have no effect. The velocity of this 
 change has been measured by Hantzsch and Smythe,' by removing portions of the 
 solution and precipitating with silver nitrate : the proportion of bromide and 
 chloride in the precipitate was determined by converting a weighed quantity 
 into the chloride in a stream of chlorine. They find that the reaction is roughly 
 of the first order, the constant falling off slightly with dilution. It is slowest in 
 water, more rapid in methyl alcohol, still more so in acetic acid, and quickest in 
 ethyl alcohol, being thus greater the less the dissociating power of the medium. 
 From these facts Hantzsch infers that it is not the ions which change, but the 
 undissociated salt : — 
 
 Br CI 
 
 Br-O-N-Cl -* Br-O-N-Br. 
 
 1^ 111 ^-f III' 
 
 Br N Br N 
 
 We may assume that, as in practically all cases of the tautomeric change of 
 an ionizable substance, the amount of change is proportional to the concentration 
 of the undissociated part, or to the product of the concentrations of the two ions, 
 which both come to the same thing.' The mono-halogen derivatives, whether 
 ortho or para, will not react at all ; the di- do so much less easily than the tri-. 
 The chlorine changes places with an ortho-bromine- atom more easily than with 
 a para. It is remarkable that the presence of a methyl group on the nucleus in 
 any position, even meta, delays the reaction. The temperature coefficient of the 
 velocity, as in other cases of tautomeric change, is unusually high, being about 
 five for ten degrees. 
 
 Cyanogen compownds 
 
 Perhaps the most striking confirmation of Hantzsch's views on the diazo- 
 compounds is to be found in their application to the cyanides, which he has 
 investigated in great detail. These cyanides exist in three distinct series. If 
 a diazonium chloride, particularly one with negative substituents in the ring, 
 such as the dibromo-compound, is dissolved in water and an equivalent of 
 potassium cyanide added, a precipitate is obtained of what is known as a normal 
 cyanide. The normal cyanides are more stable than the diazonium salts, but 
 they decompose easily, form the aryl nitriles with copper powder, couple with 
 ;8-naphthol to azo-dyes, &c. On standing in the solid state, or in alcoholic solu- 
 tion, or on heating in alcoholic solution (according to the precise compound 
 employed), they change over into the isomeric iso-diazo-cyanides — much more 
 stable bodies, which do not give nitriles or azo-dyes, or only do so with difficulty. 
 
 These two classes resemble one another in many ways. Besides the easy 
 conversion of one into the other, their reactions are in many respects identical, 
 
 » Ber. 29. 947 (1896) ; 31. 1253 (1898). ^ Hantzsch, Bei: 30. 2334 (1897). 
 
 ' Ber. 33. 505 (1900), * See p. 184. 
 
274 Diazo-compounds 
 
 and so are their molecular solution volumes, as determined by Traube's method. 
 They always behave as cyanogen and never as isocyanogen derivatives. But 
 their behaviour is totally at variance with what we should expect from a diazo- 
 
 nium compound. Such a body, Ar-N^Qjj, like Ar-N^^p should closely resemble 
 
 the cyanides of potassium and tetramethyl-ammonium. But these last are 
 
 highly ionized and hydrolysed salts, colourless, easUy soluble in water, sparingly 
 
 in organic solvents, and decomposed with evolution of prussic acid by acetic 
 
 and even by carbonic acid. The diazo-cyanides, both normal and iso, are quite 
 
 different. They are coloured, volatile in steam, scarcely soluble in water but 
 
 easily in organic solvents, not directly decomposed {i.e. with evolution of 
 
 prussic acid) except by the strongest acids. They are not ionized at all, and in 
 
 fact are not in any respect to be compared to salts. On the contrary their 
 
 behaviour is rather that of organic cyanides (nitriles), as is shown in their 
 
 saponification to amides and acids ; and of azo-compounds, as in their colour and 
 
 volatility in steam. Considering the close resemblance between the diazonium 
 
 salts and those of the quaternary ammoniums and the alkalies, we cannot doubt 
 
 that normal and iso-diazo-cyanides are not diazonium compounds at aU, but are 
 
 diazo-compounds, with the formula Ar-N=N-CN ; and as the differences between 
 
 the two series correspond closely to those between the normal and iso-diazotates 
 
 {stability, reactivity, coupling power), we must conclude that the normal low 
 
 melting unstable reactive compounds are the syn-, and the iso-, which are 
 
 comparatively stable and inactive, are the anti-isomers :— 
 
 Ar-N normal, Ar-N iso, 
 
 II II 
 
 NO-N syn-diazo. N-CN anti-diazo. 
 
 A particularly strong proof of the formulae is the fact that it is only the normal 
 
 which lose nitrogen to form the nitrile : — 
 
 Ar-N Ar N 
 
 II -» 1 -I- III • 
 NC-N NC N 
 
 Besides these two series of syn- and anti-diazo-cyanides there exists a third, 
 
 which is of a totally different character. The diazo-cyanides of anisol and 
 
 pseudocumene (i. e. of the most basic compounds), when solid or when dissolved 
 
 in indifferent organic solvents or in alcohol, behave entirely as syn-diazo- 
 
 cyanides. They are coloured, they are not dissociated in aleohoUc solution, and 
 
 they are not salts. But when they are dissolved in water they give a colourless 
 
 solution which is an excellent conductor — is highly ionized — and is decomposed 
 
 by acetic or carbonic acid. (It is to be noticed that this is a curious inversion 
 
 of the usual order, where the ionized tautomer is coloured, and the non-ionized 
 
 colourless.) The conductivity of these bodies is of the same order as that of 
 
 potassium cyanide, which they resemble in every respect. These are clearly the 
 
 true diazonium cyanides, Ar-N^Q-xx. The investigation of the ultra-violet 
 
 absorption of these bodies confirms this conclusion, the spectra of the syn- and 
 anti-diazo-cyanides being almost identical, while those of the diazonium 
 compounds are quite different.' 
 
 » Dobbie, Tinkler, J. C. S. 1905. 273. 
 
Diazo-cyanides 275 
 
 With other diazo-compounds which give a less positive diazonium- group, the 
 tendency to form a diazonium cyanide is less. Ordinary diazobenzene cyanide 
 in aqueous solution gives an equilibrium between the undissociated syn-diazo 
 and the dissociated diazonium ; whereas compounds containing negative substi- 
 tuents, such as the di- and tri-bromo-bodies, are present in aqueous solution 
 almost if not quite entirely as syn-diazo-cyanides. In fact we can compare 
 a whole series of diazo-compounds, starting with the anisol and pseudocumene 
 ■derivatives, and passing through diazobenzene itself to the mono-, di-, and tri- 
 bromo-bodies : and as we go down the series we find that the basicity of the 
 diazonium hydrate diminishes, and with it the dissociation of the cyanide in 
 aqueous solution, and also its solubility in and decomposition by acids. The 
 same influences which cause the normal diazo-hydrate — the syn-diazo-hydrate — 
 to change into the dissociated diazonium hydrate also cause the syn-diazo-cyanide 
 to change into the dissociated diazonium cyanide. 
 
 The diazonium cyanides are very difficult to isolate. Like potassium cyanide 
 they are largely hydrolysed in solution, i. e. the solution contains a considerable 
 amount of hydroxyl ions ; and like diazonium hydrate, only to a still greater 
 •extent, they are isomerized by hydroxyl ions to form syn-diazo-cyanides. 
 Moreover, these latter are much less soluble in water, and so crystallize out first. 
 
 They can, however, be obtained as double cyanides with silver cyanide, 
 
 Ar-N\p,1ig-, AgCN. These are colourless soluble neutral salts, decomposed by 
 
 acetic acid to form nitrogen, hydrocyanic acid, and a phenol, with precipitation 
 ■of silver cyanide : in other words, strictly analogous to potassium silver cyanide, 
 
 KAgCy^. 
 
 By treating pure anisol-diazonium hydrate at 0° with excess of a very concen- 
 trated prussic acid solution, and evaporating the solution in vacuo over phos- 
 phorus pentoxide at 0°, Hantzsch succeeded in obtaining crystals of the 
 composition ArNj-CN, HON, HjO, which appear to be the soUd diazonium 
 •cyanide. It forms a highly dissociated solution in water, with a conductivity 
 nearly as high as that of potassium cyanide, which has all the reactions of 
 a diazonium cyanide solution, and which, when treatedwith a little alkali, gives 
 a yellow precipitate of the syn-diazo-cyanide. 
 
 Theory of the typical diazo-reactions 
 
 It remains to consider what explanation this theory has to offer of the 
 mechanism of those reactions in which the diazo complex is eliminated and 
 replaced by a different group. There is an obvious peculiarity about them, in 
 that they scarcely ever occur directly, according to the equation : — 
 
 Ar-Nij-X = Ar-X -i- Nj, 
 but almost always require the presence of another substance (water, copper 
 powder, &c.), and hence proceed thus : — 
 
 ArNj-X -1- HE = Ar-E + H-X + Ng. 
 Now it is found by experiment that the only bodies which decompose directly 
 are the syn-diazo-compounds, and that the diazonium compounds always do so 
 indirectly, in accordance with the second equation. This is illustrated by the 
 
276 Diazo-compounds 
 
 case of the sulphonates and the cyanides, and also, as we have seen, of the 
 halides, where the most easily decomposed are the most deeply coloured, i. e. are 
 those which contain the largest proportion of the syn-diazo-compound. 
 
 The indirect decomposition of the diazonium compounds is shown in the 
 ordinary diazo-reactions, such as the formation of phenols with water, and of 
 phenol ethers with alcohol. The first of these reactions, the formation of 
 phenol and nitrogen in aqueous solution, has been the subject of several investi- 
 gations.^ The method used in all cases was to determine the amount of change 
 by measuring the volume of nitrogen given off. It was found that the reaction 
 was monomolecular, that is, the amount of change was proportional to the 
 quantity of diazonium salt present. It was the same for the chloride, bromide, 
 sulphate, nitrate, and oxalate : it was unaffected by excess of mineral acid, 
 unless this was very large. SubstitUents on the nucleus, especially in the 
 ortho position, whether positive or negative, diminish the velocity. The case of 
 the iodide is peculiar. Its decomposition is much more rapid, is greatly 
 increased by the presence of light (as indeed are all diazo-reactions)," and owing 
 to the occurrence of side reactions no definite constants can be obtained. In 
 the case of the chloride, Hantzsch and Thompson' have shown that the 
 decomposition is slower if the salt is dissolved immediately after its preparation, 
 than if it is kept for some time in a desiccator, although no visible change and 
 no loss of weight take place. This must be due to the disappearance of some 
 negative catalyst, whose nature is, however, unknown. 
 
 As regards the mechanism of the reaction, the simplest hypothesis is that it 
 is due to hydrolysis : — 
 
 ArN-Cl H2O ArNOH Ar-OH N 
 
 -» "I — > + 
 
 N N ^ ^ N 
 
 But if this were so, the velocity would be greater the weaker the diazonium, 
 whereas it is independent of this. Further, the hydrochloric acid liberated in 
 the reaction would diminish the hydrolysis, and therefore the velocity, and the 
 monomolecular constant would not be maintained ; moreover, the addition of 
 excess of acid at the beginning of the reaction would greatly diminish the 
 velocity, which it does not. Hence the reaction cannot be due to hydrolysis, 
 since its rate depends on the concentration not of the hydrolysed but of the 
 dissociated part of the compound, i.e. of the diazonium ion. But it is quite 
 possible that some addition reaction takes place, for example, with the water, 
 whose concentration remains constant during the reaction. And the fact that 
 in all these reactions the presence of another substance is required points clearly 
 to the occurrence of some such addition reaction. It is easy to see how this 
 would happen. We may suppose that water is first added to the diazonium 
 molecule, and then hydrochloric acid split off: — 
 
 Ar OH Ar OH Ar OH ArOH 
 
 N=N + -^ N=:N -♦ N=N -^ N=N . vi^y^'^ 
 
 01 H cm CLH QVR ^'"jji^j^'' 
 
 1 Hantzsch, Oswald, Ber. 33. 2528 (1900) ; Euler, Ann. 325. 295 (1902) ; Cain, J. C. S. 1902. 
 1412 ; 1903. 470 ; Schwalbe, Ber. 38. 2196, 3071 (1905) ; Cain, ib. 2511 ; Hantzsch, Thompson, 
 Ber. 41. 3519 (1908). ' Orton, Coates, J. C. S. 1907. 35. s i^^_ 
 
Theory of Diazo-reactions 277 
 
 Addition-compounds such as are here assumed have been found to exist in some 
 cases ; and we know that the syn-diazo-compound which would be formed 
 breaks up in this way into nitrogen and phenol. Since it is found that the 
 velocity is in most cases independent of the acid radical present (provided that 
 it is the radical of a strong acid), it is more correct to formulate the reaction as 
 being due to the diazonium ion, so that the anion does not appear in it at all : — 
 
 Ar OH Ar OH Ar OH ArOH 
 
 I- I. I II + 
 
 N=N + -* N=N -» N=N -» N=N • 
 
 H H + H" + H" 
 
 A peculiarly complicated case is presented by the reaction with alcohol, 
 because this may go, as we have seen, in two directions, giving either the hydro- 
 carbon or the phenol ether. Both these results are easily explained on Hantzch's 
 theory. It has been shown that the normal course of the reaction is to give the 
 phenol ether : this is the main product with monovalent alcohols, and the only 
 product with the polyvalent, such as glycerine. The formation of hydrocarbons 
 is only the main reaction when there are negative substituents present on the 
 nucleus. But if these are numerous, as in the tri-bromo-compound, the phenol 
 ether is not formed at all, even with dilute alcohol. Now it is clear that on the 
 addition hypothesis the alcohol can attach itself in two ways, and one of these 
 will lead to the formation of a phenol ether, the other to that of a hydro- 
 carbon : — 
 
 I. 
 
 
 Ar 
 
 1. 
 
 
 
 1 
 
 NE 
 
 1 
 
 N 
 
 Lr 
 
 1 
 
 k 
 
 H 
 
 1 
 
 EN 4 
 
 
 Cl 
 
 
 
 
 O-C^Hg Ar O-C^Hs Ar^OC^Hs 
 
 _♦ N=N -» N=N . 
 
 I I; + 
 
 H Cl H CIH 
 
 Ar H ArH 
 
 II. N=N + -* N=N -* N=N 
 
 _ .^ Cl O-OaHs cfn + CH3.CHO. 
 
 Which of these two reactions will occur will depend on the relative attraction 
 of the groups for one another, the positive hydrogen going on the side of the 
 negative chlorine, except when it is attracted away by a strongly negative aryl 
 radical. 
 
 By means of similar additive compounds it is possible to explain the other 
 diazo-reactions, such as those of Sandmeyer and Gattermann. 
 
 Another reaction which supports this view is the formation of diazo-com- 
 pounds by the action of hydroxylamine on a nitroso-derivative. Bamberger' 
 stated that this produced the iso- or anti-diazotate ; but Hantzsch ' has been 
 able to show that it is really the syn-diazotate which is formed. This is to be 
 expected if we suppose that an intermediate addition-compound is formed, 
 which then loses a hydrogen and a hydroxyl which are near to one another : — 
 
 Ar-N=0 Ar-N-OH Ar-N 
 
 -* I -* W + H„0. 
 
 + HONH2 HONH HON ^ 
 
 It is evident from the foregoing that the constitution of the diazo-complex 
 in a compound ArNg-X is mainly determined by the nature of the radical X. 
 
 » Ber. 28. 1218 (1895). ' Ber. 38. 2056 (1905). 
 
 1175 T 
 
278 Diazo-compounds 
 
 But it is also affected by the nature of the aryl radical, or in other words by the 
 substituents on the benzene ring. The main directions in which this latter 
 influence acts have been determined by Hantzsch, by the investigation of a large 
 series of compounds. They are as follows. The more important substituents 
 are on the one hand alkyl groups and methoxyl, which render the nucleus more 
 basic, and on the other halogen atoms, which make it more negative or acidic. 
 As regards the relation between diazonium and syn-diazo bodies, it is found that 
 the more basic the nucleus is, the more the diazonium form is favoured ; and 
 the more negative, the more the syn-diazo. Thus the trimethyl-diazonium 
 bromide is almost colourless and hardly explosive ; the tribromo-diazonium 
 bromide is very explosive and deeply coloured ; hence, the former consists 
 mainly of the diazonium, the latter mainly of the syn-diazo. So, too, the 
 trimethyl-diazonium cyanide in aqueous solution is almost wholly in the form 
 of diazonium cyanide ions, of which the tribromo-derivative scarcely forms any. 
 Again, the trimethyl-diazonium hydrate is nearly as strong as potassiimi hydrate, 
 while the tribromo-diazonium hydrate is weaker than ammonia ; that is, it 
 contains very little dissociated diazonium hydrate. 
 
 Between these two extreme limits come the other diazo-compounds, for 
 which the same general rule holds : the fewer alkyls or the more halogen atoms 
 there are on the ring, the less the tendency to produce the diazonium form. It 
 is to be observed that the methoxyl group has the same powerful basic influence 
 as the alkyl ; a conclusion to which Baeyer has also come from a study of the 
 aromatic oxonium salts. 
 
 The relations among the true diazo-compounds between the syn- and anti- 
 series are less regularly affected. In the diazotates the change of syn- into anti- 
 is hindered by methyl groups and promoted by halogens ; while in the sul- 
 phonates and the cyanides exactly the reverse effect is produced, the more 
 positive compounds changing the most rapidly. The presence of nitro-groups 
 increases the velocity of the change in all cases. 
 
 The conversion of the anti-diazo-hydrates into the nitrosamines seems to be 
 hastened by the presence of halogen atoms. 
 
 The following summary of the more important results of these investigations 
 may be of service. 
 
 We find among the diazo-compounds representatives of all the four possible 
 formulae : — 
 
 ArNX 
 
 1. in . Diazonium. (Blomstrand.) 
 
 Ar-N Ar-N 
 
 2, 3. II : II . Syn- and anti-diazo. (Kekule-Hantzsch.) 
 
 4. Ar-NH-N:0. Nitrosamine. (v. Pechmann.) 
 
 The normal salts with strong (mineral) acids are in the solid state mainly 
 diazonium, but partly (especially the halides) solid solutions of diazonium and 
 syn-diazo. In solution they give almost entirely the diazonium ions. The 
 normal hydrates, obtained by treating the salt with an equivalent of acid, are 
 
Diazo-compounds : Constitution 279 
 
 in solution equilibrium mixtures of the ions of the strongly basic diazoniura 
 hydrate, undissociated diazonium hydrate, and undissociated syn-diazo-hydrate. 
 The normal diazotates which this solution forms with alkalies are salts of the 
 weakly acidic syn-diazo-compounds ; but on heating, or sometimes on standing 
 at the ordinary temperature, they undergo a stereoisomeric change, giving the 
 iso-diazotates, which are anti-diazo-compounds, derived from the less reactive 
 form, which is more acidic than the syn-. If the iso (anti-) diazo-hydrate is set 
 free, it changes with great ease into the nitrosamine. 
 
 The ordinary diazo-reaetions which result in the elimination of nitrogen are 
 not reactions of the diazonium salts at all, but of the syn-diazo-bodies formed 
 from them through an additive compound with some other substance present. 
 
 Finally, among the cyanides there are three distinct series, corresponding 
 to all the three possible formulae : — 
 
 1. Syn-diazo: generally the most easily formed, especially when there are 
 negative substituents on the ring. They are coloured, insoluble in water, easily 
 soluble in organic solvents, not dissociated, behaving like organic nitriles, and 
 not decomposed except by strong acids. They break up into nitrogen and 
 nitriles directly. 
 
 2. Anti-diazo : formed by spontaneous change from 1, which they closely 
 resemble. They are coloured, insoluble in water, soluble in organic solvents, 
 not ionized, not true salts, not decomposed by weak acids. But they are more 
 stable than the syn-, and do not give the nitrUe directly. 
 
 3. Diazonium : formed from the syn-diazo-compounds with positive sub- 
 stituents on the ring, by dissolving them in water. They are colourless, easily 
 soluble in water, going back in organic solvents into the coloured syn-diazo. 
 They are ionized and hydrolysed in water, are decomposed by weak acids, and 
 are converted into the insoluble syn-diazo-bodies by hydroxy! ions. 
 
 With regard to the tautomerism of the anti-diazo-hydrates and the nitros- 
 amines, it is to be noticed that it really belongs to the keto-enolic type, and 
 completes the series of the possible replacements of the carbon in that type by 
 nitrogen. We have the following four groups, all formed on the same plan, and 
 all characterized by the same tendency to tautomerism : — 
 
 1. CH, 
 
 2. NH 
 
 1 
 
 3. CH, 
 
 4. NH 
 
 1 
 
 c=o 
 
 h 
 
 N=0 
 
 N=( 
 
 CH 
 ^.OH 
 
 i 
 
 COH 
 
 CH 
 loH 
 
 1 
 N 
 
 II 
 NO 
 
 Of these, the first type provides the classical examples of tautomerism, as in the 
 /3-keto-esters and diketones. The second is exemplified in the amides and iso- 
 amides, and in such ring-compounds as isatin. The third type occurs in the 
 nitroparaflSns, and in the tautomerism of the secondary or primary nitroso- 
 compounds and the oximes. 
 
CHAPTER XII 
 AZO-COMPOUNDS, AZOXY-COMPOUNDS, NITRAMINES 
 
 The azo-eompounds differ from the diazo in having the -N=N- group 
 attached to carbon on both sides. According to this definition the diazo- 
 cyanides are, strictly speaking, azo-compounds, but it is more convenient to treat 
 them as belonging to the diazo-group. The azo-bodies with which we are 
 practically concerned are almost entirely those in which the Ng group connects 
 two aromatic nuclei. This conditions much greater stability, and excludes the 
 possibility of the complicated isomerism and tautomerism of the diazo-group. 
 
 The nomenclature usually adopted for the simpler azo-bodies is to put the 
 word azo between the names of the hydrocarbons from which the two radicals 
 are derived. Thus : — 
 
 CH3-C5H4-N=N-^ = Toluene-azo-benzene, 
 0-N=N-C5H4-N(CH3)2 = Benzene-azo-dimethyl-aniline. 
 
 Other systems have been suggested, which are perhaps more convenient in 
 dealing with the more complicated derivatives, and may be used by the specialist 
 who wishes to distinguish the enormous number of azo-dyes. But for ordinary 
 purposes the simpler method will suffice. 
 
 The azo-compounds are usually prepared by the moderated reduction of the 
 nitro-compounds in alkaline solution : generally with zinc dust in alcoholic 
 solution in presence of potash, or with sodium amalgam, or more recently by 
 electrolysis. This, of course, yields only the symmetrical bodies. It is worth 
 noticing that the yield is worse the more side chains are present. They may 
 also be obtained by oxidizing the amines in alkaline solution with potassium 
 permanganate or ferricyanide. In this case the yield is better the more side 
 chains there are. 
 
 The unsymmetrical compounds are prepared by heating a mixture of an 
 amine and a nitro-derivative with powdered soda to 180-200°, a method devised 
 by Sandmeyer. The first product is the azoxy-compound : — 
 
 CH3-C6H4-N02 + H^N-^ = CHsCgH^-N — -N-^ + HgO. 
 
 Then one part of the azoxy-body reduces the rest to azo, being itself burnt up 
 in the process. 
 
 A formation of theoretical interest is by the action of a nitroso-compound on 
 an amine : — 
 
 CHgCeH^-NHa + 0=N-^ = CHgCoH^-N^N-^ -t- H^O. 
 This corresponds to the formation of a diazo-compound from a nitroso-body and 
 hydroxylamine. 
 
 The mixed azo-compounds (containing a fatty as well as an aromatic group) 
 
Azo-compounds : Formation 281 
 
 are got by oxidizing the mixed hydrazo-compounds or symmetrical secondary 
 fatty-aromatic hydrazines with mericuric oxide : — 
 
 0-NH-NHCH3 + O = 0-N=N-CH3 + H2O. 
 The aromatic azo-bodies are crystalline substances (some of the mixed deriva- 
 tives are liquids) of an intense red or orange colour. They distil unchanged at 
 rather high temperatures (azobenzene, for example, at 293°), and they are volatile 
 in steam. They are insoluble in water, alkali, and acid, but dissolve in organic 
 solvents. They are remarkable for their extreme stability ; they are not affected 
 by distillation or by boiling with acid or alkali, a striking contrast to the diazo- 
 compounds. As regards their structure, it is obvious that, like the diazo-bodies, 
 they might be either syn- or anti-compounds ; from their great stability it is 
 clear that they must be anti-, 
 
 Ar-N 
 II 
 N-Atj 
 
 No cases of stereoisomerism have yet been observed among them for certain, 
 although there is an instance of isomerism among the aminoazo-compounds, 
 which is very probably to be referred to stereo-causes. 
 
 Though they are so stable under ordinary conditions, they are very sensitive 
 to reducing agents, which, if alkaline, convert them iato hydrazo-compounds, or, 
 on more energetic reduction, into amines : acid reducing agents generally trans- 
 form them directly into benzidines. They are even reduced, especially in 
 methyl alcohol solution, by hydrochloric acid, with the formation chiefly of 
 chlorination products of hydrazobenzene and anUine.^ 
 
 Azobenzene, ^•N=N-^, itself was discovered by Mitscherlich in 1834. It 
 forms fine orange-red crystals, melting at 68° and boiling at 293°. It may be 
 prepared in any of the ways that have been mentioned, and is generally made 
 by reducing azoxybenzene (obtained by the reduction of nitrobenzene in alkaline 
 solution) by distillation with iron filings. 
 
 The mixed azo-bodies, in which the N2 is attached on one side to an aromatic 
 and on the other to a fatty hydrocarbon, are liquids. They resemble the purely 
 fatty azo-bodies, and differ from the purely aromatic, in being able to split off 
 the nitrogen on heating, the two residues combining. This reaction has been 
 made use of by Gomberg for the preparation of tetraphenyl-methane. He made 
 triphenyl-methane-hydrazo-benzene, ^gC-NH-NH-^, oxidized this to the corre- 
 sponding azo-body 03C-N=N-^, and by heating a mixture of this with sand 
 obtained, though only in very small quantity, the previously unknown tetra- 
 phenyl-methane 0^0. 
 
 Another peculiarity of these mixed azo-compounds is that sulphuric acid 
 converts them into the isomeric hydrazones : — 
 
 0-N=NCH2CH3 -* 0.NH-N=CHCH3. 
 This has already been mentioned in connexion with the supposed formation of 
 hydrazones by the action of diazo-compounds on the acidic raethylene group. 
 
 The simplest azo-compound, azomethane, CH3-N=N-CH3, has been prepared 
 quite recently by Thiele,' who obtained it by oxidizing hydrazomethane, 
 
 1 Jacobson, Ann. 367. 304 (1909). = Ber. 42. 2575 (1909). 
 
282 Azo-compounds 
 
 CHg-NH-NHCHg (s-dimethyl-hydrazine), with potassium chromate. It is a 
 
 colourless gas with no alkaline reaction, which condenses to a liquid boiling at 
 
 + 1-5°. The liquid has only a pale yellow colour, in which respect it resembles 
 
 , (CH3)2C-N=N-C(CH3)2 
 another purely fatty azo-compound, azoisobutyric ester,- I I • 
 
 C02Et C02i!jt 
 
 It explodes on heating, but if mixed with excess of carbon dioxide decomposes 
 quietly, mainly into ethane and nitrogen : — 
 
 CH3N=NCH3 = CH3-CH3 + N^. 
 
 The alternative formula, that of a hydrazone, CH3-NH-N=CH2, is excluded 
 by its distinct though pale colour in the liquid state, by its decomposing to give 
 ethane (which is exactly analogous to the formation of tetramethyl succinic 
 ester from azoisobutyric ester, an undoubted azo-body), and by the fact that 
 it is easily reduced to the hydrazo-compound. On the other hand, it is easily 
 broken up by acids into methyl hydrazine and formaldehyde, so that it can 
 readily assume the hydrazone structure. The same thing has been observed 
 with the mixed azo-body, phenyl-azo-ethane, ^-N^N-CHa-CHs.^ 
 
 The remarkable instability of azo-triphenyl -methane, <f>jiu-l^—^-G<p^, has 
 already been mentioned (p. 255). 
 
 By far the most important azo-derivatives are the aminoazo- and oxy-azo- 
 compounds which are used as dyes. Their formation has already been referred 
 to. Practically any primary aromatic amine can be diazotized, and the body so 
 formed can be coupled with any aromatic amine (primary or otherwise) or 
 phenol ; so that the number of azo-dyes which can theoretically be prepared 
 is enormous. Already the components described in patents for azo-dyes are 
 sufficient to yield several millions ; and the number which could be obtained 
 from all the known amines and phenols amounts to several hundred millions. 
 In comparison with these figures the 75,000 compounds in Beilstein and the 
 100,000 in Richter's Lexicon appear insignificant. 
 
 As the azo-dyes are perhaps the most important and certainly the most 
 numerous class of organic dyes, a few remarks on the process of dyeing may be 
 useful. 
 
 The cloth is dyed by bringing it into a solution of the dye in water : this is 
 the dyeing bath. It is therefore necessary to render the dye soluble in water, 
 if it is not so already. This is frequently done by sulphonation, i. e. by intro- 
 ducing an SO3H group into the nucleus, since all sulphonic acids are soluble in 
 water ; the substitution in most cases has little or no effect on the colour. The 
 dyes.are divided into two classes, adjective and substantive dyes. A substantive 
 dye is one which is capable of dyeing the fibre as it is, that is, without any 
 previous treatment ; an adjective dye requires the use of a mordant. Many 
 dyes are either substantive or adjective according to the fibre used. A great 
 number are substantive to animal fibre (silk or wool) and adjective to 
 vegetable. 
 
 As mordants, tannic acid and compounds of it with preparations of antimony 
 are often used. Acid dyes, such as oxy-azo-compounds, are often fixed by first 
 soaking the cloth in a solution of an easily decomposed aluminium, ferric, or 
 
 » Thiele, Heuser, Ann. 290. 5. (1896). « E. Fischer, Ser. 29. 793 (1896). 
 
Azo-dyes 288 
 
 chromic salt (especially the acetate), and then passing it through a feebly basic 
 bath or steaming it, so as to separate the oxide of the metal. The dyeing then 
 consists in the formation of an insoluble compound of the dye with the oxide, 
 known as a lack. The colour with many dyes depends on the base used ; such 
 dyes are called polychromatic. Thus alizarin dyes red with aluminium salts, 
 violet with ferric, and brown with chromic. 
 
 Sometimes the dye is produced on the fibre itself, either by soaking it in 
 a solution of the leuco-base (e. g. indigo-white, which is probably di-indoxyl) and 
 then oxidizing it by exposure to the air ; or, in the case of azo-dyes, by passing 
 it successively through solutions of the two components. 
 
 The value of a dye depends, of course, on its beauty, on its dyeing power, 
 and on its fastness — its power to resist washing, soap, and light, and in special 
 cases acids, &c. 
 
 A dye is necessarily a coloured substance capable of adhering to the fibre ; 
 and it, therefore, must have in the molecule two kinds of groups, one giving it 
 colour, and the other giving it this power of adherence. The first of these 
 groupings is known as a chromophor, such as the azo-group, -N=N-, which of 
 itself gives colour, but does not yield a dye ; the second is called an auxochrome, 
 but about this there is a certain amount of confusion. The terminology is 
 originally due to Witt, who supposed that certain groups, which he called the 
 chromophors, were essentially colour-producing, but in some cases required the 
 presence of a second group, not in itself coloured, to call out or at least increase 
 the colour. Groups of this second kind he called auxochromes. Subsequently, 
 as the interest of chemists was diverted from the purely scientific question of 
 the cause of colour to the more practical problem of producing dyes, the word 
 auxochrome changed its meaning, and was applied to those groups which give 
 a coloured compound its power of adhering to the fibre. This confusion was 
 increased by the fact that the two most important auxochrome groups in the 
 dyeing sense are hydroxyl and the amine or substituted amine group ; and 
 these are also auxochromes in the original Witt sense, of increasing the colour 
 of a compound without reference to its dyeing power. These two groups are 
 both salt-forming, one being basic and the other acidic ; but other salt-forming 
 groups, such as COgH and SO3H, will not act in this way. 
 
 Another remarkable point about dyes, and, indeed, about coloured organic 
 compounds in general, is that they can be converted by reduction into colourless 
 bodies, the so-called leuco-compounds, which generally go back readily, often 
 even on exposure to the air, into the original dye. The leuco-derivatives of the 
 azo-dyes are the hydrazo-compounds, which themselves are colourless, but are 
 reconverted by the air into the coloured azo-compounds. 
 
 The azo-dyes may be divided into two classes, according to the auxochrome 
 (in the dyeing sense) which they contain : — 
 
 1. The aminoazo-compounds, with NHj. 
 
 2. The oxy-azo-compounds, with OH. 
 
 The first azo-dye put on the market was chrysoidine (from diazobenzene and 
 meta-phenylene-diamine), which was prepared in January, 1876, by an English 
 firm, by the method discovered, of course, by Griess, but developed by 0. N. Witt. 
 
284 Azo-compounds 
 
 AMINOAZO-COMPOUNDS 
 
 In practice the aminoazo-compounds are always prepared by the action of 
 an amine on a diazo-solution. But in many cases this gives first, as an inter- 
 mediate product, the diazo-amino-body. Direct formation of the aminoazo- 
 compound occurs with a tertiary amine, as it obviously must, and with 
 »t-diamines in neutral or feebly acid solution. It is also observed with some 
 secondary monamines in presence of mineral acids, while in neutral or acetic 
 acid solution they usually give the diazo-amino. The conversion of the diazo- 
 amino into the aminoazo-body occurs slowly even on standing in alcoholic 
 solution ; but it is hastened by the addition of small quantities of the amine 
 salt. If the diazo-amino-compound is treated with an equivalent of the salt of 
 another amine, it is possible to introduce this second amine, e. g. : — 
 CH3.C6H4-N=N-NH.C6H4CH3 + CeHs-NHgCl 
 
 = CHg-CeH^-N^N-CeHi-NHa + CHa-CgH^-NHsCI. 
 The aminoazo can be prepared directly, under suitable conditions, by acting on 
 excess of the amine with nitrous acid ; but this implies the intermediate formation 
 of the diazo and diazo-amino-compounds. 
 
 The position taken up by the azo-group on the nucleus of the amine is 
 always para, if this place is open ; and it is only in this case that the reaction 
 goes easily. If this place is occupied, the reaction can be brought about, though 
 much less readily, and the azo-group takes up the ortho position : — 
 
 i-N=:N-lj 
 
 ^•N=N-NH-<3CH3 -». f 
 
 NH, 
 
 It is remarkable that dialkyl-amines, in which the para position is filled up, 
 usually refuse to give ortho products.^ Either no reaction takes place, or the 
 para substituent is expelled, and a J^aminoazo-compound produced. The most 
 striking case of this is afforded by p,jp-tetramethyl-diamino-diphenyl-methane, 
 where both the nuclei separate from the methane carbon : — 
 2 HSO3.CeH4.N2.OH + (CH3)2NC3-CH2.0.N(CH3)2 
 
 -^ 2 HS03.C6H4-N=N.CI>N(CH3)2 + CHjO + H^O. 
 The structure of the product is determined by energetic reduction, which con- 
 verts it into a mixture of a monamine and an 0- or ^-diamine. 
 
 We have thus two classes of aminoazo-compounds, the ortho and the para. 
 They differ so remarkably in behaviour that it has been suggested that they 
 have different constitutions. 
 
 The para compounds have the normal behaviour of primary amines, and 
 hence must certainly have the normal aminoazo formula. But the behaviour 
 of the ortho-compounds makes it possible that they contain a closed nitrogen 
 ring, e.g. :— 
 
 0.NH.l[r-<3-CH3 
 
 ' Scharwin, Kaljanov, Ser. 41. 2056 (1908). 
 
Aminoazo-compounds 285 
 
 For instance, on oxidation they give colourless azimido-compounds such as 
 
 CHa-CeH^.N-N-O-CHs 
 
 Again, whereas the para react with aldehydes like primary amines, giving 
 indifferent compounds which readily decompose again into their components, 
 the ortho give colourless basic substances, stable on heatiag with acids, which 
 belong to the triazine group . — 
 
 H 
 
 It is to be noticed, however, that these differences are scarcely greater than 
 those between the para and ortho diamines, where no difference of constitution 
 is to be suspected. The peculiarities of the ortho bodies are probably only 
 illustrations of the general tendency of ortho di-substitution-products to form 
 closed rings ('ortho-condensation'). 
 
 Properties 
 
 They are yellow or brown substances, scarcely soluble in water : hence the 
 importance of sulphonation when they are to be used as dyes. They are basic, 
 and easily dissolve in acids to form brilliantly coloured salts, which usually 
 have a different colour from the base ; these salts are discussed below. They 
 have the behaviour both of amines and of azo-compounds. As amines they give 
 acyl and alkyl derivatives and salts ; and they can be diazotized and used as 
 components for new azo-dyes, producing the dis-azo-compounds. As azo-bodies 
 they can be reduced to hydrazo-compounds and ultimately split up into a mona- 
 mine and a diamine. If they are boiled for a long time with strong hydro- 
 chloric acid the same decomposition occurs, the acid acting like a mixture of 
 free hydrogen and free chlorine (compare the action of hydrochloric acid on 
 quinone chlorimines and tetraphenyl-hydrazine). Thus aminoazo-benzene yields 
 mainly aniline and ^phenylene diamine, together with chlorinated hydroquinones 
 such as CeHCl3(OH)2. 
 
 Another peculiar method of splitting the aminoazo (and also the oxy-azo) 
 compounds into their components is by the action of fuming nitric acid.' This 
 is sometimes of use for determining their constitution where the reduction 
 method fails. The azo-group does not split, as it usually does, between the two 
 nitrogen atoms, but it breaks off entirely from one nucleus, appearing as the 
 diazonium salt, while its place is taken by a nitro-group : — 
 
 HS03<3-N=NC3-N(CH3)2 -» S03<3N=N + N0,<3-N(CH3),. 
 
 It has been shown by Bamberger'' that the azoxy-compounds are oxidized to 
 diazo by potassium permanganate, and it is therefore probable that in this case 
 the fuming nitric acid first oxidizes the azo-group to azoxy and then further to 
 the diazo ; indeed, simple oxidizing agents such as chromic acid and potassium 
 permanganate will break up azo-compounds in the same way. 
 
 1 0. Schmidt, Ser. 38. 8201, 4022 (1905). ' Ber. 33. 1957 (1900). 
 
286 Azo-compounds 
 
 The mineral acid salts of the aminoazo-compounds have recently been 
 investigated by Hantzsch and Hilscher/ who have shown that they occur regu- 
 larly in two series, one yeUow and the other violet. An isolated case of this 
 was, indeed, discovered by Thiele some years ago. Which of these two is 
 obtained in any particular case depends both on the aminoazo-compound used 
 and on the acid ; nor could any regularities in their influence be discovered. 
 In many cases, however, both the isomeric salts could be isolated, the yellow 
 unstable salt usually crystallizing out first, and the violet stable form later. In 
 the solid state the unstable form does not change over as long as it is kept dry 
 and cold ; but on warming, or in presence of traces of free acid, or sometimes 
 even on rubbing, it goes over into the stable form. In solution equilibrium 
 between the two forms is as a rule established at once, the proportions, and 
 hence the colour of the solution, depending on the base and the acid, but very 
 largely on the solvent, indifferent solvents such as chloroform favouring the 
 violet form, and solvents of a higher dielectric constant, like ether and alcohol, 
 the yellow. 
 
 As regards the question of structure, that of the orange-yellow series is fairly 
 certain . Their colour is very similar to that of azobenzene itself and the salts 
 of azobenzene-trimethyl-ammonium, 0-N=N-C6H4-N(CH3)3X, which, as it contains 
 no mobile hydrogen atom, must have the true azo-structure, and whose salts 
 are accordingly found to be in all cases orange-yellow. It is therefore certain 
 that these orange-yellow salts of the aminoazo-compounds are also true azo-salts 
 of the type 0-N=N.C6H4.NK2HX. 
 
 The violet salts are not polymers of the yellow, as was shown by a determina- 
 tion of their molecular weight in solution, which was found to be normal. 
 They are, therefore, true isomers of the yellow form. The hydrazone formula 
 
 ArNHN=C6H4=NR 
 suggested by Auwers,'' /\ , is excluded by the fact that similar 
 
 salts are given by the dialkyl-aminoazo-bodies. Hantzsch and Hilscher therefore 
 propose the quinoid formula Ar-NH-N=C6H4=NR2X for the violet modification. 
 While this view is quite probable, it is to be noticed that the relations of these 
 two series of salts are almost precisely the same as those observed in the yeUow 
 and red salts of the nitrophenols, the dinitroparaflSns, and the a-nitroketones. 
 In these three last-mentioned cases Hantzsch attributes the isomerism with 
 great probability to stereo-chemical causes ; and the same explanation will 
 account for the aminoazo-salt if we suppose that the more stable orange-yeUow 
 salts belong, like azobenzene, to the anti-series, and the unstable violet salts to 
 the syn- : — 
 
 Ar-N violet, Ar-N orange-yellow, 
 
 HXE^N-CeHi-N unstable " N-CoH^-NEaHX stable. 
 
 Or perhaps an even closer analogy to these other cases is expressed by a formula 
 such as Ar-NH-N-CgH^-NEaX, which might also admit of stereoisomerism. 
 
 • Ber. 41. 1171 (1908); Hantzsch, Ber. 43. 2129 (1909). = Ber. 33. 1314 (1900). 
 
Oxy-azo-compounds 287 
 
 OXY-AZO-COMPOUNDS OR AZO-PHENOLS 
 
 The symmetrical oxy-azo-compounds may be obtained by the reduction of 
 nitrophenols in alkaline solution ; but these are of small practical importance in 
 comparison with the unsymmetrical, which are made by coupling a diazo-solution 
 with a phenol. The reaction takes place practically only in a weakly alkaline 
 solution. In neutral solution it often leads to the formation of nitrogen and 
 a phenol ether ; while a great excess of alkali may prevent the reaction altogether, 
 probably from the formation of the isodiazotate. Goldschmidt has shown that 
 the velocity of formation of the dye is proportional to the concentration both of 
 hydrogen and of hydroxyl ions. It is also, of course, aifected by the nature 
 both of the phenol and of the diazo-compound. As regards the phenol, it is 
 found that the presence of an azo-group in it hinders the coupling ; oxy-azobenzene 
 will couple again to give bis-azo-phenol, HO-CjHsC-N^N-^) 2, but this body 
 is not capable of taking on another azo-group. It is remarkable that the 
 presence of a doubly linked carbon attached to the ring of the phenol has the 
 same effect as an azo-group, so that oxy-cinnamic acids will only couple once 
 and not twice.' As regards the influence of the structure of the diazo-body, the 
 reaction velocity is diminished by the introduction of halogens, and increased 
 by that of alkyls ; that is, it is greater the more basic the diazo-compounds. 
 
 The similarity of the reactions of diazo-compounds with phenols and with 
 amines makes it probable that there may be an intermediate stage in the former 
 case analogous to the diazoamino-bodies in the latter : — 
 
 0N2OH + HOC:>-H -> 0-N=NO<OH -* H0<ON=N-^, 
 as was indeed originally suggested by Kekule. It has been proved by Dimroth ' 
 that in many cases an intermediate product is formed, to which he attributes 
 a structure of this kind. For instance, ^-bromophenyl diazonium chloride com- 
 bines with jp-nitrophenol to form an unstable substance which soon decomposes 
 with loss of nitrogen, and is at once broken up by acids to give the diazonium 
 salt. If this is heated in the dry state to 80° it goes over into the stable oxy-azo- 
 compound 
 
 Br<3-N2-0-<3-N02 -> H0-<p-N02 
 
 N=:N-<3-Br " 
 
 Other similar cases are recorded by Bucherer ' and by Auwers,'' who, however, 
 
 points out, as we have seen already, that the behaviour of these bodies is quite 
 
 compatible with their being not 0-azo-compounds, Ar-N=N-OAr, as Dimroth 
 
 Ar-N-0-Ar 
 suppjoses, but merely diazonium salts, III' , which would explain why they 
 
 are formed with peculiar ease from negatively substituted phenols. Whichever 
 view may be right, it is evident that in coupling with phenols the Ng group first 
 attaches itself to the oxygen, as it does with amines to the nitrogen, and then 
 subsequently passes over to the ring. 
 
 As we have seen, the power of coupling is one of the best tests for dis- 
 
 ' Borache, Streitborger, Ber. 37. 4116 (1904). ^ Dimroth, Hartmann, Ber. 41. 4012 (1908). 
 s Ber. 42. 47 (1909). * Ber. 41. 4304 (1908). 
 
288 Azo-compounds 
 
 tinguishing between the syn- and anti-diazo-compounds. An anti-compound 
 will couple, but always much more slowly than its syn-isomer. 
 
 The following rules govern the position of the azo with respect to the 
 hydroxyl group. If the para position is open it is always taken. If it is 
 occupied and the ortho is open, then that is taken. There are only one or two 
 examples of the ortho position being occupied when the para is open, and a few 
 where one or both of these positions are unoccupied, and where yet no coupling 
 occurs at all. The same applies to the case where a second azo-group is intro- 
 duced, except that the two will not both take up the ortho position. Coupling 
 in the meta position never occurs at all, and the meta-oxy-azo-compounds can 
 only be got by indirect methods. Para-oxy-benzoic acid couples with diazo- 
 benzene to give para-oxy-azobenzene, the carboxyl being eliminated, as also 
 happens in certain benzidine reactions. 
 
 There is another method of preparing the oxy-azo-compounds which is of 
 some theoretical interest, and that is by the action of amines on the quinone 
 oximes or nitrosophenols. This can be written in two ways, which illustrate 
 the two possible formulae of the oxy-azo-bodies : — 
 
 ^•NHa + H0N=O=0 = 0NHN=O=O + HgO; 
 
 ^■NHa + 0=N-O>-0H = (;6.N=N-CI>-0H + H^O. 
 This raises the question of the constitution of the oxy-azo-compounds, which 
 have so far been assumed to be true phenols. This was the original and is the 
 most probable view, as their behaviour is in many respects that of phenols. 
 The question has, however, been much disputed,^ and a great deal of work has 
 been devoted to it, especially in recent years, without any very satisfactory 
 results being arrived at. The first doubt as to the correctness of the phenol 
 structure was raised by Liebermann,^ on the ground of certain differences in 
 behaviour between the ortho and the para series, and by Zincke and Bindewald,' 
 who showed that a-naphthoquinone and phenyl-hydrazine gave the same product 
 as that obtained from diazobenzene and a-naphthol : they therefore suggested 
 that the body might really be a hydrazone : — 
 
 ^•NH-NH, O 0-NH-N 
 
 *C0- iCO- 
 
 II II 
 
 o o 
 
 Goldschmidt and his pupils * then found that the ortho compounds would not 
 react with phenyl isocyanate, and therefore attributed to them the hydrazone 
 (quinoid) structure. Meldola^ investigated the products of the reduction of the 
 acyl derivatives, and found that in all cases they behaved as if they were 
 a mixture of the two forms ; for example, acetyl^-oxy-azobenzene as if it were 
 partly 0N=NC6H4-O-COCH3 and partly <p-^{GO-GR^)Yi-C^B.i-0. It has 
 since been shown that the acyl groups in these compounds are unexpectedly 
 
 ' A brief and very clear Bummary of the question is given by Auwers, Ann. 360. 11 (1908). 
 » Ber. 16. 2858 (1883). 3 Ber. 17. 3032 (1884). 
 
 * Ber. 23. 487 (1890); 24. 2300 (1891); 25. 1324 (1892). 
 
 " Meldola, Morgan, J. C. S. 1889. 114 ; Meldola, Hawkins, ib. 1893. 923 ; Meldola, Burls, 
 ib. 930; Meldola, Hanes, ib. 1894. 834. 
 
Occy-azo-compounds : Structure 289 
 
 mobile, and much of the confusion which has arisen is due to the fact that this 
 mobility was not recognized.' 
 
 Hewitt ^ then investigated the action of substituting agents, especially 
 bromine and nitric acid, on the oxy-azo-compounds, and proved that the para and 
 the ortho derivatives behaved in the same way, and as if they were true phenols. 
 
 Jacobson and Honigsberger ° attacked the problem by preparing the 
 previously unknown meta-oxy-azobenzene, which might be assumed not to 
 have a quinoid structure, and instituted an elaborate comparison of its properties 
 with those of the other two isomers. The results are as follows. The meta 
 and para compounds dissolve in the equivalent quantity of soda, whereas the 
 ortho will only dissolve in a very large excess. In toluene solution the meta 
 and para take up dry ammonia readily, but the ortho not at all. All three bodies 
 form a salt with hydrochloric acid ; but the meta and para salts are stable, 
 whereas the ortho salt loses its hydrochloric acid in the air very readily. Again, 
 the molecular weight of the ortho compound, as determined by the freezing- 
 point method in benzene solution, is normal, and independent of the concentra- 
 tion ; whereas that of the meta and para rises rapidly with the concentration. 
 In all these cases the meta and para derivatives behave alike, so that as the 
 meta compoimd is a true phenol we must conclude that the para is so also. But 
 the peculiar behaviour of the ortho compound led Jacobson and Honigsberger to 
 the view that it was the quinone-hydrazone 
 
 at the same time its alkaline salts, which exactly resemble those of the para 
 body, they assumed to be true phenol salts with the structure 0-N=N-< >-0M. 
 If this view is correct, it is remarkable that it should be the ortho compound 
 which has the strongest tendency to assume the quinoid form, since among the 
 quinones themselves it is the para which is the most easily produced and the 
 most stable. It is, however, doubtful whether the differences which have been 
 enumerated are sui&cient to establish the difference of structure. The peculiarities 
 of the ortho compound, some of which have since appeared to be differences in 
 degree rather than in kind, may be merely due to the influence of the hydroxyl 
 in immediate proximity to the azo-group. There is no doubt that in this 
 position the hydroxyl is much less acidic, and in general (perhaps through 
 stereo-hindrance) much less active. This is shown not only in the insolubility 
 of these compounds in alkali (which, however, is less marked in some cases than 
 it is in that of o-oxy-azobenzene), but also, for example, in the ease of methyla- 
 tion by methyl sulphate, a reaction which goes quantitatively in the para series, 
 but with much less ease in the ortho.' In the same way, while the para bodies 
 react with diazomethane to give their methyl ethers, the ortho do not react at 
 all ^ ; and this last fact cannot be explained by giving them a quinoid structure, 
 since a hydrazone should react with diazomethane through the NH group. 
 
 > Cf. McPherson, JBer. 28. 2414 (1895) ; Am. Ch. J. 22. 364 (1899). 
 
 2 Hewitt, /. C. 8. 1900. 99 ; Hewitt, Aston, ib. 712, 810 ; Hewitt, Pox, ib. 1901. 49 ; Hewitt, 
 Lindfield, ib. 155 ; Hewitt, Phillips, ib. 160; Hewitt, Auld, ib. 1902. 1202 ; Hewitt, Walker, ib. 
 1906. 182. ' Ber. 36. 4093 (1903). 
 
 * Colombano, C. 08. i. 23. " C. Smith, MitoheU, J. O.S. 1908. 843. 
 
290 Azo-compounds 
 
 The argument of Goldschmidt, that the ortho compounds are quinoid because 
 they do not react with phenyl isocyanate, he has himself shown to be untenable/ 
 as the reaction can be' brought about under suitable conditions. Goldschmidt 
 has therefore abandoned his former view, and now considers that all oxy-azo- 
 compounds are true phenols. 
 
 An investigation of the ethers and esters (alkyl and acyl derivatives) of the 
 osy-azo-compounds by Auwers, Goldschmidt, Hewitt, Willstatter, and others, 
 has shown that these bodies are all true phenol derivatives. In those cases 
 where the acyl hydrazone compounds can be prepared, as by the action of asym- 
 metrical phenyl-benzoyl-hydrazine on quinone, they change over with unexpected 
 ease into the true azo-compounds ^ : — 
 
 ^.cS>N-NH, + 0=0=0 -* ^.c§>N-N=O=0 -* 0-N=N-C>-OCO.^. 
 
 This suggests that the free hydrogen compounds, which can change much 
 more easily, will also assume the phenol structure. 
 
 Borsche ' has examined the case of bodies containing the group 
 
 HOCeH^-N^N- 
 
 attached to other than aryl radicals, and his results indicate that here also the 
 azophenol structure is more stable than the isomeric quinone-hydrazone. 
 
 The evidence from the physical behaviour of the free oxy-azo-compounds, 
 such as it is, points to the same conclusion, or at least does not contradict it. 
 Auwers * showed that the cryoscopic behaviour of the para oxy-azo-compounds 
 was that of phenols, while the ortho bodies gave ambiguous results. The 
 evidence on which Hantzsch and Farmer" based the view that these bodies 
 were pseudo-acids was weak, and they have subsequently abandoned it. But 
 more recently Hantzsch and Glover" have shown, from the consideration of 
 their colour, that the free para compounds must have the same constitution as 
 their ethers and esters, i.e. that they must be phenols. Lemoult^ concludes 
 from the heat of combustion of a series of azo-dyes that they are all true azo- 
 compounds. Tuck * claims to have proved from the ultra-violet absorption that 
 all the para derivatives are of the phenol type, and also the ether of the ortho 
 series ; but that the free ortho compounds and their acyl derivatives are hydra- 
 zones. Not only, however, is the whole basis of his method open to some 
 doubt, but the actual results point just as much in the opposite direction. 
 
 We may conclude, therefore, that the free oxy-azo-compounds, whether they 
 belong to the ortho or to the para series, have the true phenol structure 
 
 Ar-N^N-CeH^-OH. 
 This is an example of the general tendency of the quinoid derivatives (and the 
 hydroaromatic bodies in general) to pass back if possible into aromatic com- 
 
 ' GoIdBchmidt, Liiw-Beer, Ser. 38. 1098 (1905). 
 ^ Willstatter, Veraguth, Ber. 40. 1432 (1907). 
 
 3 Ann. 334. 148 (1904) ; Borsche, Ookinga, ib. 340. 85 (1906) j Borsche, ib. 343. 176 (1905) ; 
 357. 171 (1907). 
 
 * Auwers, Orton, Z. Ph. Ch. 21. 356 (1896) ; Auwers, Ber. 33. 1802 (1900). 
 » Hantzsch, Ber. 32.690 ; Hantzsch, Fanner, ib. 3089 (1899). 
 
 • Ber. 39. 4153 (1906). ' C. M. 143. 603 (1906). • J. C. S. 1907. 449. 
 
Oxy-azo-compounds : Structure 291 
 
 pounds. (Compare the case of the quinols.) When the passage of the azo into 
 the hydrazone type does not involve the production of a quinoid ring (as in the 
 mixed azo-compounds), the hydrazone structure is the more stable, as is shown 
 by the ease with which benzene-azo-methane goes over into acetaldehyde 
 phenyl-hydrazone. 
 
 The alkaline salts may be taken to have the same constitution as the mother 
 substances, for it is scarcely likely that the replacement of the hydrogen by 
 a metal will favour the hydrazone as opposed to the phenol form. On the other 
 hand the oxy-azo-compounds can also act as very weak bases and form salts with 
 acids. This salt formation produces a difference in colour, but Tuck' has 
 shown that the same difference is produced in the free oxy-azo-compound as in 
 its ethyl ether. This excludes the possibility of its going over into a hydrazone. 
 Fox and Hewitt "^ have found that while j3-oxy-azobenzene gives with dilute 
 nitric acid (which does not produce the salt to any great extent) the ortho-nitro 
 derivative, if it is nitrated in concentrated sulphuric acid solution, where it is 
 present practically entirely as the salt, the nitro-group goes into the para position. 
 This shows that the formation of the salt involves some change of structure, as 
 might be expected from the change of colour, while it must be a change which 
 can occur with the alkyl derivative. They therefore suggest that the salt is an 
 oxonium compound with a quinoid ring : — 
 
 Hewitt and Mitchell^ point out that both in the oxy- and in the aminoazo- 
 compounds, the introduction of a nitro-group in the para position to the azo 
 results in a compound which changes colour and becomes bluer on treatment 
 with alkali. This indicates a change of constitution, due to the presence of the 
 nitro-group, and on the analogy of the nitrophenols we may assume in this case 
 the production of the quinoid form : — 
 
 02N<3]sr=]src>0H -». koon=o=n-n=o=0- 
 
 The di-oxy-azo-compounds, such as azo-phenol, HO-C8H4-N=N.CgH4'OH, 
 present certain remarkable peculiarities.' Para-azo-phenol (but not the meta or 
 ortho compounds, nor the corresponding amino-derivatives) can be oxidized by 
 silver oxide or lead dioxide to a substance which must have the formula 
 
 and is known as quinone-azine. The failure of this reaction in the ortho series 
 is a strong argument against the ortho oxy-azo-compounds having a quinoid 
 structure, as in that case they ought to oxidize easily. Quinone-azine occurs in 
 two different solid forms, which are possibly stereoisomers. On reduction it 
 takes up two atoms of hydrogen again, but the product is not the original 
 azo-phenol but a new modification. The differences between the two forms of 
 azo-phenol are small, but quite distinct (in colour, solubility, &c.). The isomerism 
 is maintained in solution and in the salts. The fact that neither form is reduced 
 further by such agents as phenyl-hydrazine indicates that neither of them 
 
 M. c. ' J. C. 8. 1908. 333. 5 J. C. 8. 1907. 1251. 
 
 * Willstatter, Benz, Ser. 39. 3482, 3492 (1906) ; 40. 1578 (1907). 
 
292 Azo-compounds 
 
 possesses a quinoid structure. This is further proved by the fact that they give 
 two di-ammonium salts (which cannot be derived from a quinoid form) which 
 regenerate the original a- and /3-azo-phenol with acids. This seems to exclude 
 structural isomerism, and to make it certain that they are stereoisomers. 
 There are other cases in which stereoisomerism among azo-compounds is 
 suspected/ though in none has it been definitely established. A more detailed 
 investigation of the two para-azo-phenols shows that each of the two forms is 
 capable of existing in two modifications, each pair changing more readily into 
 one another than into the other pair ; and the difference persists even in the 
 acetyl derivatives. No explanation has been offered of this. 
 
 AZOXY-COMPOUNDS 
 These bodies, which have the general structure 
 
 Ar-N N-Ar, 
 
 are the first bimolecular reduction-products of the aromatic nitro-compounds. 
 They are made by boiling the nitro-bodies with alcoholic potash. The real 
 reducing agent is the alcohol, which is oxidized to the corresponding acid. In 
 this reaction it is found that ethyl is more powerful than methyl alcohol, and 
 potash than soda. 
 
 This method cannot be used if there is a methyl group in the para position 
 to the NO2, as in that case some of the hydrogen required for the reduction is 
 taken from the methyl, and a dibenzyl or stilbene derivative is formed. It is 
 also possible to reduce the alcoholic solution of the nitro-body with sodium 
 amalgam, or to use zinc dust in the presence of alcoholic ammonia. 
 
 The azoxy-compounds are likewise formed by the oxidation of azo- and 
 hydrazo-compounds, and, as we have seen, by the action of nitroso-compounds 
 on /3-aryl-hydroxylamines. This last method of formation is of the utmost 
 importance, since it is practically the sole source of the bimolecular reduction- 
 products of the aromatic nitro-compounds. 
 
 The azoxy-compounds are crystalline bodies of low melting-point, and of 
 a rather pale yellow colour. They cannot be distilled unchanged, and are con- 
 verted into the azo-bodies by distillation with iron filings. One of their most 
 interesting reactions is that when gently warmed with sulphuric acid many of 
 them are converted into the isomeric oxy-azo-compounds (Wallach). 
 
 They have not been investigated in very great detail, and there is even some 
 doubt as to their structure. But the only alternative to the usually accepted 
 
 Ar-N=N-Ar 
 formula is II , which represents them as to some extent similar to the 
 
 A.r*N*Ar 
 nitrosamines, I ; and it has been shown by Lachman that they differ 
 
 N:0 
 
 from the nitrosamines in every possible way, being very inactive, while the 
 
 nitrosamines are excessively active. So we may take the usual formula as 
 
 sufficiently established. 
 
 ' Cf. Jaeobson, Honigsberger, Ber. 36. 4093, 4123, note 1 (1903). 
 
Azoxy-compounds 293 
 
 Beissert * has obtained isomeric forms of azoxybenzene and of o-azoxytoluene. 
 
 They are nearly colourless, even in solution, and have higher melting-points 
 
 than the normal forms, into which they readily pass on warming or in presence 
 
 of catalytic agents, such as bromine. They show no definite differences &om 
 
 the normal forms in chemical behaviour. Keissert suggests that they may have 
 
 Ar-N=N-Ar 
 the structure II 
 O 
 
 The most remarkable point about the azoxy-compounds is that in a certain 
 
 number of them the mysterious phenomenon of liquid crystals has been 
 
 observed. If 2J-azoxy-anisol, ^' ^^ — ''\^ /^ — ' " ^, is heated, it melts at 
 
 116° to a turbid liquid, which at 184° suddenly becomes clear. The clear liquid 
 is perfectly normal in its behaviour, and like other liquids. The turbid liquid 
 is found to possess strong double refracting power. This would seem to imply 
 that the orientation of the molecules which exists in the (solid) crystal is main- 
 tained in the liquid. Various theories have been advanced to explain away the 
 phenomenon, as that the turbid liquid is an emulsion of two liquid phases, or 
 a suspension of small solid crystals in the liquid. But these have been 
 disproved, and there can be no doubt that we are dealing with a real liquid, in 
 which, nevertheless, the molecules are maintained in a certain fixed arrangement. 
 The turbidity is merely due to the fact that the various crystal drops have their 
 axes in different directions, and so resembles that of a mass of powdered crystals. 
 The phenomenon is widely spread among certain classes of azoxy-derivatives, 
 and is also observed in bodies of quite a different type, such as the cholesteryl 
 esters. With some bodies the clearing temperature is below the melting, so 
 that the liquid crystal phase can only be observed in the supercooled liquid. 
 In certain compounds Vorlander has noticed the occurrence of two different 
 liquid crystalline phases, one dark and the other Ught in colour, with 
 a perfectly definite transition point. The most inexplicable point about the 
 whole phenomenon is perhaps the fact that the viscosity of the crystalline 
 liquid is in nearly all cases less (and sometimes much less) than that of the 
 isotropic form. 
 
 NITEOSAMINES 
 
 The nitrosamines are derived from the amines by replacing a hydrogen 
 attached to nitrogen by NO. Hence two classes are theoretically possible, the 
 primary E-NH-NO, and the secondary KaN-NO. Among the fatty compounds 
 (other than the derivatives of carbonic acid) the primary nitrosamines do not 
 exist at all. There are several reactions which might be expected to produce 
 them, but they yield only their decomposition products. Thus a primary 
 amine with nitrous acid might give either a primary nitrosamine or a diazo- 
 
 compound : — 
 
 CH3NH2 -f HON:0 = CHgNHNiO -h H^O, 
 or CHj-NHjj + 0:NOH = CH3-N=N-0H ^- HjO. 
 
 As a fact, of course, neither is obtained, but only nitrogen and the alcohol. 
 
 1 Ser. 42. 1364 (1909). 
 1175 U 
 
294 Nitrosamines 
 
 A reaction which gives a good illustration of the degree of stability of these 
 primary nitrosamines is the saponification of nitroso-methyl-urethane. This 
 would naturally lead to the formation of methyl nitrosamine : — 
 
 (^0 ■ + 2H2(> = CO2 + C2H5.OH + HN<gj3. 
 
 / 
 
 \OC2H5 
 
 Hantzsch has shown that in the presence of excessively concentrated aqueous 
 potash this will give a true diazotate, CHg-N^N-OK, a tautomer of the expected 
 nitrosamine, but this loses potash with the utmost ease, in fact almost 
 explosively, to form diazomethane : 
 
 CHo-N=NOK = KOH + CH, 
 
 'N 
 
 .11 > 
 
 N 
 
 and the latter reacts with water to give nitrogen and methyl alcohol. 
 
 On the other hand, in the aromatic series, as we have seen, the primary 
 nitrosamines can be isolated. They are formed by spontaneous change from 
 the anti-diazo-hydrates. But even here they are very unstable, and can exist 
 only in the solid state or in neutral solution. The presence of hydrogen or 
 hydroxyl ions converts them either into a diazonium or into a diazo-compound. 
 In fact these aromatic nitrosamines are at once pseudo-acids and pseudo-bases. 
 
 The secondary nitrosamines exist both in the fatty and in the aromatic series: 
 the instability of the primary being obviously due to the hydrogen which still 
 remains attached to the nitrogen, and which has the same tendency to go over 
 to the oxygen which we observe in the simple primary or secondary nitroso- 
 compounds : — 
 
 E2C<J.Q -> E2C=N0H : EN<^.q -^ RN=N-OH. 
 
 Secondary nitrosamines are formed by the action of nitrous acid on secondary 
 amines : — 
 
 ^j^3\NH + HO-N:0 = ^^3^N-N:0 + H2O. 
 
 In one case, that of di-isopropylamine, the nitrite, /pTT^{^pTT/'NH-HN02, can be 
 
 isolated ; and a salt of this type may be assumed to be an intermediate product 
 in all cases. This nitrite is a crystalline substance, stable in cold aqueous 
 solution, and only goes slowly on boiling into the nitrosamine. In all other 
 cases, even with normal dipropylamine, the nitrosamine is formed at once in the 
 cold. 
 
 The fatty nitrosamines are also formed in a curious way, by heating the 
 nitrates of the secondary amines to 150°, when they give off oxygen : — 
 
 (CH3)2NHHN03 = (CH3)2N-]Sr:0 + HjO + O. 
 The secondary nitrosamines are liquids or solids, volatile in steam. The fatty 
 distil without change, but the aromatic do not. They are much used for 
 isolating the secondary amines. They separate out as oils on treating the 
 mixed bases with potassium nitrite and acid. They can be made to re-form the 
 secondary amines, the fatty by heating with concentrated hydrochloric acid, 
 the aromatic by reduction with tin and hydrochloric acid. 
 
Nitrosamines 295 
 
 Weaker reducing agents, such as zinc and acetic acid in alcoholic solution, 
 
 convert them into the secondary hydrazines such as ?)>N-NH2. They give 
 
 the Liebermann reaction for nitroso-compounds. This is effected by wai-ming 
 them with strong sulphuric acid and phenol, when a red colour is produced, 
 which, on pouring the liquid into water and adding excess of alkali, turns to 
 a beautiful blue. 
 
 When the mixed (fatty-aromatic) secondary nitrosamines are fused with 
 potash, a part is decomposed into nitrous acid and the secondary amine, while 
 another part splits between the nitrogen and the fatty radical, and gives 
 phenyl-nitrosamine, which forms its potassium salt, the anti-diazotate. 
 
 If they are treated with alcoholic hydrochloric acid the nitroso-group 
 
 migrates as usual to the ring, giving p-nitroso-methyl-aniline, 0:N< ^^\tt ^• 
 
 The same effect is produced by alcoholic hydrobromic acid, but curiously not 
 by alcoholic sulphuric acid. 
 
 NITRAMINES 
 
 The nitro-amines or nitramines are bodies in which one hydrogen atom of 
 an NHj group is replaced by NO2. They are more stable than the nitrosamines, 
 as NO2 is usually more stable than NO, and both the primary and the secondary 
 compounds are known. 
 
 The primary fatty nitramines are formed from the monalkyl-urethanes or the 
 monalkyl-oxamides. These bodies on treatment with anhydrous nitric acid are 
 nitrated in the NH group thus : — 
 
 CH3NH CH3NNO2 ^ 
 
 CO-OEt ~* COOEt' 
 
 0=C.NH.Et 0=C-N<gJ^2 
 
 0=C.NH.Et "* 0=C.N<g*oJ 
 
 When the products are treated with ammonia, the nitramine group is 
 replaced by the NHj of the ammonia, giving acetamide or oxamide and the 
 ammonium salt of the nitramine, which we may write CHg-NHNOg-NHg . This, 
 when boiled with alkali, loses the ammonia and gives the free nitramine. 
 
 The primary alkyl nitramines are acid substances, which react with alkyl 
 
 C H \ 
 
 iodide and potash to form the dialkyl nitramines, such as ^TjV^'^Oa. 
 
 The aromatic nitramines are the so-called diazobenzenic acid and its substitu- 
 tion-products. This body is got by oxidizing potassium diazobenzene, or better 
 potassium isodiazobenzene, with potassium ferricyanide.' It is also produced 
 by the action of nitrogen pentoxide on an ethereal solution of aniline at — 10", 
 which is evidence for its being the true nitro-amine ^■NH'N02 . In the same way 
 its alkyl derivatives can be made by the action of fuming nitric acid on the 
 monalkyl- or (with the loss of an alkyl group) on the dialkyl-anilines. 
 
 ' Bamberger, Ber. 21. 915 (1894). 
 
 u2 
 
296 Nitramines 
 
 Phenyl nitramine melts at 46° and explodes at 100°. It is slightly soluble 
 in water, giving an acid solution. It is very sensitive to acids, being converted 
 mainly into ortho but partly into para nitraniUne, a further proof of its constitu- 
 tion. This tendency of the nitro-group to migrate to the ring is of course the 
 normal behaviour of a substituent attached to the nitrogen atom of aniline ; 
 and as usual the ortho and para positions are taken up, but under no conditions 
 the meta, and the reaction is promoted by the presence of hydrogen ions. If 
 the para and both the ortho positions are already occupied by halogen (in s-tri- 
 chloro- and s-tribromo-phenyl nitramine) the nitro-group is able to turn out the 
 para halogen atom, either expelling it from the compound altogether, or moving 
 it on to the next carbon of the ring.' 
 
 Phenyl nitramine is very stable to alkali, and is not decomposed by it unless 
 it is fused with potash at 230-260°, when it is converted into aniline and 
 potassium nitrite and nitrate. On reduction in alkaline solution it gives first 
 phenyl nitrosamine and then phenyl-hydrazine. 
 
 Its sodium salt reacts with methyl iodide to give phenyl-methyl nitramine, 
 
 J^NNOj, which is converted by acids into ortho and para nitro-methylaniline, 
 
 and reduces, like the mother substance, to phenyl-methyl nitrosamine and then 
 to unsymmetrical phenyl-methyl-hydrazine : whereby its formula is established. 
 
 Its silver salt, when treated with methyl iodide, yields a very unstable isomer 
 of this, which decomposes on standing, and is probably an oxygen ester. 
 
 As regards the constitution of these bodies, it is fairly certain that free 
 phenyl nitramine has the structure ^-NH-NOa. This is supported by its forma- 
 tion from aniline and nitrogen pentoxide, and by its conversion into ortho and 
 para nitraniline. It is, however, tautomeric in behaviour, and gives two series 
 
 of ethers, one of which undoubtedly has the structure .7jr>N-N02, the bodies 
 
 being true secondary nitramines. These must therefore be the derivatives of the 
 normal form. On the other hand Hantzsch and Dollfus' have shown that it 
 gives the ammonia reaction for pseudo-acids : that is, in dry benzene solution it 
 combines with dry ammonia much more slowly than a true acid (like benzoic 
 acid) does, which indicates that it requires to go over into the tautomeric form 
 in order to form the salt. In aqueous solution, since it is a fairly strong acid 
 (about as strong as acetic acid), it must be largely present in the isomeric hydroxyl 
 form, from which the alkaline salts and no doubt the second series of ethers — 
 the 0-ethers— are derived. This form may probably be, on the analogy of the 
 
 isonitro-compounds, Ar.N=N^Qg ; but there is no direct proof of this, and it 
 
 Ar-N N-OH 
 
 may conceivably be \rv/ ' * formula 'which is in fact adopted by 
 
 Scholl.' 
 
 * Orton, Smith, J. C. S. 1905, 889 ; 1907. 146. 
 
 » Ber. 36. 259 (1902) ; Hanteich, ib. 39. 2103 (1906). 
 
 = Ann. 338. 11 (1904). 
 
Nitrimines 297 
 
 NITKIMINES > 
 
 Certain ketoximes, especially those of the camphor series, when they are 
 treated with nitrogen peroxide, instead of forming pseudonitrols in the normal 
 manner, convert the oxime group into the group NjOa- This reaction seems to 
 be confined to those compounds which have the C=NOH group attached on one 
 or both sides to a tertiary or a doubly linked or a quaternary carbon atom. The 
 products are known as nitrimines, and it is evident from their behaviour that 
 they belong to the class of nitramines. They give the Liebermann reaction, the 
 Thiele-Lachman reaction for nitramines (the formation with strong sulphuric 
 acid of nitrous oxide and nitric acid), they can in some cases be reduced to 
 hydrazines, and their behaviour with alkalies is that of the alkyl nitramines 
 and other pseudo-acids. There can thus be no doubt that the characteristic 
 
 group, C2HN2O2, which they contain must be written either >C=C-NH-N02 or 
 
 >CH-Ct=N-N02 . It is a strong argument in favour of the second of these two 
 structures that unlike the primary nitramines they will react neither with 
 diazomethane nor with phenyl isocyanate. 
 
 They give two series of alkyl derivatives, one having the alkyl attached to 
 nitrogen and the other to oxygen. The salts are obviously derived from the 
 same form as the second series of ethers, and in many cases this second form, 
 
 which must be X'=C-N=N<(qjj or >C=CN NOH, can actually be isolated. 
 
 The passage of the pseudo-acid form into the true acid form (e. g. in presence of 
 an alkali) involves the formation of an intermediate compound : — 
 
 >CHC=NN02 -^ >C=CNH.N02 -» >C!=CN=N<2jj . 
 The N-ethers are obviously derived from this intermediate compound. \ 
 
 In the nitrimine derived from chlorocamphor a remarkable case of intra- 
 molecular migration has been observed." The nitro-group changes over spon- 
 taneously from the nitrogen to the next carbon atom, giving the free imine, 
 which can easily be hydrolysed to the chloronitro-camphor : — 
 
 /C=N.NOo /Ct:NH /C=0 
 
 * "^CHCl * "^C-Cl-NOa * "^CCl-NOa 
 
 Chlorocamphor- Chloronitro- Chloronitro- 
 
 nitrimine. camphor-imine. camphor. 
 
 This change is exactly analogous to the transformation of phenyl nitramLne into 
 ortho-nitraniline ; but among other than aromatic bodies such changes are rare. 
 
 ISONITEAMINES' 
 Isomeric with the nitramines are the so-called isonitramines, which are 
 probably nitroso-hydroxylamines, EN\jt.q. The fatty isonitramines are ob- 
 
 ' SchoU, Ann. 338. 1. (1905) ; 345. 363 (1906). 
 
 " Angeli, Angelico, Castellana, Atti Lino. 12. i. 428 (C. 03. ii. 373). 
 
 s Traube, Ber. 28. 1785 (1895) ; 29. 667 (1896) ; Gomberg, Awn. 300. 59 (1898). 
 
298 Isonitramines 
 
 tained by treating sodium acetoacetic ester with nitric oxide in the presence of 
 sodium ethylate : — 
 
 EtO-CO/^^^2 + ^ JNO - EtOCO/^\Na 
 When the product is hydrolysed by alkali, it gives the salt of isonitramine 
 
 CH2-N{0H)N0 
 acetic acid, I 
 
 ' COOH 
 
 With dUute mineral acids these bodies change over into the so-called 
 
 CH,NHOH , . 
 
 amidoxyl-fatty acids, such as 1 , which are yS-hydroxylamme deriva- 
 
 CO'OH 
 
 tives. On reduction in alkaline solution they give diazoacetic acid, no doubt 
 
 through the intermediate production of the nitrosamine and the true diazo- 
 
 compound : — 
 
 CH2-N(0H)N0 CHa-NHNO CHa-N^N-OH _^ CH<^ , 
 
 COOH "^ COOH ~* COOH ~* (Iq-OH 
 
 The aromatic isonitramines are obtained directly from the ^-aiyl-hydroxyl- 
 amines by treatment with nitrous acid : — 
 
 <P-K^ + Ho-No = <t>Km + ^^^' 
 
 The isonitramines are not pseudo-acids, and even in benzene solution give 
 salts at once with ammonia. This shows that under all circumstances the 
 molecule contains a hydroxyl group. They are, however, tautomeric, giving two 
 series of esters ; but both of these have the alkyl attached to oxygen. One of 
 these series is probably derived directly from the nitroso-hydroxylamine formula, 
 which may therefore be assumed to be one of the tautomeric structures of the 
 free isonitramines : — 
 
 g.j./0CH3 . R.jT/OH 
 
 The structure of the other form is unknown ; it may possibly be either 
 ^NOCH ^^ \n/ ^ ' ^^i*'^®^®'^ o^ these two formulae is not to be 
 
 assigned to the 0-esters of the normal nitramines. 
 
 KN<S.f 
 
CHAPTER XIII 
 DERIVATIVES OF CARBONIC ACID 
 
 Amino-urethane, or hydrazine carboxylic ester, (^O , can be obtained 
 
 XOEt 
 by the reduction of nitro-urethane with zinc and acetic acid. A similar but 
 
 ^NH-NHa 
 more important body is amino-urea, CO , generally known as semi- 
 
 carbaeide. This is obtained in the same way, by the reduction of nitro-urea, 
 
 CO 
 
 \NH2 
 It may also be got &om hydrazine and cyanic acid : — 
 
 CC + \ = CO 
 
 ^0 HN.NH, \NH.NH, 
 
 It is a crystalline substance melting at 96°, and has basic properties, forming 
 salts with one equivalent of acid. Its importance lies in the fact that it reacts 
 very readily with aldehydes and ketones after the manner of phenyl-hydrazine, 
 to give semicarbazones : — 
 
 /NHNHa ^p„ /NH.N=C<^g3 
 
 CO + 0=CC„3 = CO ^^3 + HjO. 
 
 These compounds usually crystallize well and are not over soluble, so that in 
 many cases semicarbazide is better than phenyl-hydrazine for characterizing 
 such carbonyl compounds. 
 
 It is remarkable that the monosemicarbazones of the a-diketones form deep 
 yellow solutions in alkali, in which respect they resemble the monoximes of the 
 a-diketones. It has been suggested^ that this is due to their going over as 
 pseudo-acids into an acidic enoUc form : — 
 
 CHgC^O CHa^COH 
 
 CH3-C=N-NH-CONH2 ~* CHj-C^N-NH-CO-NH; 
 This view is supported by the fact that the corresponding derivative of benzil, 
 
 0C=O 
 0.C=N-NH.CO-NH2' 
 
 which could not form such an end, is the only body of this type which is 
 insoluble in dilute alkali. 
 
 ' Diels, van Dorp, Ber. 36. 3183 (1903). 
 
300 Carbonic Add Derivatives 
 
 ^NHNH2 
 
 Thiosemicjtrbazide, CS , is interesting from the fact that it forms 
 
 two isomeric diphenyl derivatives. They are prepared by acting on phenyl- 
 hydrazine with phenyl miistard oil, a reaction which one would expect to take 
 place in this way : — 
 
 If the reagents are heated, the stable or /3-form is obtained. But at a low 
 temperature an unstable isomer (the «-form) is produced. This changes very 
 readily, even on recrystallization from alcohol, into the /3 modification. It 
 might be supposed that the isomerism was due to one form having the normal 
 
 structure (given above) and the other being the iso-compound, &SH 
 
 But it has been shown by Marckwald' that both forms contain an SH group, 
 since they both react with methyl iodide to give an ether which has the methyl 
 group attached to sulphur. There thus remain only two possible explanations : 
 either that they are structural isomers with the formulae : — 
 
 ^N.NH.<^ H,N^.^ 
 
 C-SH and 
 
 C-SH 
 
 or that they are stereoisomers : — 
 
 HSCNH0 HS-CNH-d) 
 
 II ^ and ^ II ^ • 
 
 NNH-(^ 0-NHN 
 
 Marckwald decided against the first of these theories, mainly on the ground that 
 
 so great a change of structure could not occur so easily. He therefore concluded 
 
 that the bodies were stereoisomers. Eecently, however, it has been shown' 
 
 that the first explanation (that of structural isomerism) is undoubtedly the true 
 
 one. The substance formed by the action of phenyl-hydrazine on phenyl 
 
 mustard oil at a low temperature (the unstable form) has the structure 
 
 HSC^N.(^ V,-, .u .^, ^ . HS(;^N-0 .,, HS-CNH-0 
 
 , I ^^^^ , while the stable form is I ,^ , ,^ , or possibly II 
 
 ^•N-NHa NH.NH.0 ^ ' N-NH-^ 
 
 The case is of considerable interest from the remarkable ease with which the 
 mustard oil residue migrates from the a- to the ^-nitrogen atom of the phenyl- 
 hydrazine. 
 
 Both amino-urethane, which has already been mentioned, and amino-guani- 
 
 dine, (^NH , which is obtained by the reduction of nitro-guanidine, and 
 
 is used for preparing hydrazine, react with carbonyl compounds in the same way 
 as semicarbazide. In fact this reaction is common to all bodies of the type 
 HjN-E, where E is a more or less negative group (OH, NH^, or a modified urea 
 residue). 
 
 • Ber. as. 8098 (1892) ; of. Hantzsch, Ser. 26. 16 ; V. Meyer, ib. 17 (1893). 
 ' EuBoh, Holzmann, Ber. 34. 820 (1901). 
 
Semicarhazide 301 
 
 All the bodies so far considered are of course hydrazine derivatives. We 
 might expect to get a corresponding series of derivatives of diimide, NH=NH. 
 This substance is itself unknown : indeed it seems to be incapable of existence, 
 since several reactions are known which ought to produce it, but in none of 
 them is it formed. It appears to break up as soon as it is produced, usually 
 into hydrazine and nitrogen : — 
 
 HN=NH H2N-NH2 
 
 + = + . 
 
 HN=NH N=N 
 
 In the same way the mono-derivatives of diimide cannot be prepared. It is 
 necessary that both hydrogen atoms should be replaced before a compound is 
 formed sufficiently stable to be isolated. 
 
 Derivatives in which the N=N group is attached only on one side to carbon 
 we have already met with in the true aromatic diazo-compounds. Another 
 class of bodies of this kind is found among the nitroso-derivatives of the amides 
 of carbonic acid, the desmotropic form of-NH-NO being -N=N-OH. But there 
 are also some true azo-bodies among the carbonic acid derivatives, which have 
 the N=N group linked to carbon on both sides. These are the derivatives of 
 azo-dicarboxylic acid, COaHN^NCOgH. They are made in two ways. 
 
 1. If hydrazine is treated with a solution of a cyanate, the hydrazine 
 cyanate first produced changes into hydrazo-dicarboxyl-amide, just as ammonium 
 cyanate changes into urea : — 
 
 NH3-N=C=:0 NH-CO-NH, 
 I =1 
 
 NH3N=C=0 NH-CO-NHa ' 
 
 as: — 
 
 NH4-N=C=0 = NHg-CO-NHa, 
 
 and when this is treated with chromic acid, the two hydrogen atoms of the 
 hydrazo-group are oxidized off, and azo-dicarboxylamide, ^ „« t^ti^> ^^ formed. 
 
 2. The second method is to oxidize amino-guanidine with nitric acid. 
 The hydrazine NH2 groups are oxidized off along with the hydrogen atom from 
 the next NH, and the two residues unite : — 
 
 H,N-C-NH-NH, H,N-NHCNH, 
 
 '+30 = 
 
 NH NH 
 
 H„N-C-N=N-0-NH2 
 ^11 II ^ + SHoO + N„. 
 
 NH NH 22 
 
 This gives the amidine of azo-dicarboxylic acid, which like all amidines exchanges 
 its NH for oxygen on saponification (boiling with water), to form azo-dicarboxyl- 
 amide. This is an orange-red powder (the usual azo colour), and on treatment 
 with concentrated potash gives the potassium salt of the corresponding azo- 
 
 NCOOK 
 dicarboxylic acid, II , a very unstable salt, which readily splits up into 
 
 carbon dioxide, nitrogen, and hydrazine. Many attempts have been made to 
 isolate diimide from this compound, a reaction which one would expect from 
 analogy to go easily, but they have all failed. 
 
302 Carbonic Acid Derivatives 
 
 Mtroso-derwatives of the amides of carbonic mid 
 
 In the case of the secondary compounds (the alkyl derivatives) these can be 
 formed by the direct action of nitrous acid ; in the case of the primary amides, 
 such as urethane itself, the nitroso-derivatives are formed by the reduction of the 
 nitro-compounds. 
 
 The behaviour of the two classes of derivatives, the primary and the 
 secondary, is very different. 
 
 Of the primary, nitrosourethane is almost the only representative. It is 
 obtained by reducing the ammonium salt of nitrourethane with zinc and acetic 
 acid, a reaction which may be written : — 
 
 ^NHNOj ^NHNO 
 
 CO + 2H = CO + HjO. 
 
 \0-Et \OEt 
 
 This body, ethyl nitrosocarbamate (not to be confused with nitroso-ethyl- 
 urethane), is a crystalline compound melting at 51°, which readily decomposes, 
 the course of its decomposition being somewhat remarkable. 
 
 With alkalies it breaks up into carbon dioxide, nitrogen, and alcohol : — 
 
 ^NHNO 
 
 CO = CO2 + N2 + CjHs-OH, 
 
 \OC2H5 
 
 whereas with acids only half the alcohol appears as such, the other half being 
 broken up into ethylene and water : — 
 
 ^NH-NO 
 2 CO = 2 COg + 2 N2 + C2H5OH + C2H4 + H2O . 
 
 XO-CaHj 
 
 ^NHNO 
 The methyl compound, CO , is similar, but even more unstable. On de- 
 
 \OCH3 
 composition with sulphuric acid it does not form any hydrocarbon, but only 
 methyl alcohol and traces of methyl sulphuric acid. 
 
 The structure ascribed to these bodies above is that which was originally 
 assigned to them. But it has been shown by Hantzsch* that this does not 
 correctly represent their constitution. There are two classes of compounds 
 containing the group — NgOH, the nitrosourethanes and the diazo-hydrates. 
 Both these classes give highly ionized salts, which must be derived from the 
 most negative of the two possible desmotropic formulae, the potassium salt, for 
 instance, being EN=N-OK and not ENK-NO. The possibility of their having 
 this form is shown by the fact that the silver salt of nitrosourethane fornjs an 
 O-ether with ethyl iodide. 
 
 In the case of the diazo-hydrates, as we have seen, the hydrogen compounds 
 of the strongly acidic form Ar-N=N-OH can be isolated, but are very unstable ; 
 and the only stable form is the tautomeric nitrosamine Ar-NH-NO. 
 
 But, curiously, the behaviour of the nitrosourethanes is the reverse of this. 
 Here the free hydrogen compound, instead of being a neutral non-electrolj^, as 
 
 ' Ber. 32. 1703 (1899). 
 
Nitrosourethanes 303 
 
 a nitrosamine would be, is always an electrolyte of strongly acid reaction, 
 whose affinity constant can be measured, and which combines with dry ammonia 
 even in benzene solution. These bodies must therefore have a constitution 
 corresponding to that of their salts, and must really be diazo-urethane hydrates. 
 
 Op . This view accords with their decomposition, in which, as in the 
 
 \0-Alk 
 
 normal decomposition of the diazo-compounds, nitrogen is evolved. 
 
 Hantzsch's results are in striking contradiction to the views previously held 
 as to the constitution of these bodies. We have been accustomed to regard the 
 free diazo-hydrates as true diazo-compounds, and the urethane derivatives as 
 nitrosamines. But his work shows conclusively that the reverse is the case. 
 The free iso-diazo-hydrates have at best only a temporary existence, the stable 
 form being the nitrosamine, while the nitrosourethanes are really diazo-hydrates. 
 
 The secondary nitrosourethanes are obtained by the action of nitrous acid on 
 the alkyl-urethanes : — 
 
 (3^Q^H + HONO = ^0 -"^^ ■*■ ^^O- 
 \0-Et \0-Et 
 
 The desmotropic diazo-formula is here excluded, since the mobile hydrogen 
 is replaced by alkyl. 
 
 This body, nitroso-methyl-urethane, is chiefly remarkable as the source from 
 which V. Pechmann first prepared diazomethane, a reaction which has already 
 been discussed. 
 
 Of the nitroso-derivatives of urea only the secondary are known. The 
 reduction of nitro-urea itself cannot be stopped at the stage of nitrosourea, but 
 proceeds further to amino-urea or semicarbazide. 
 
 The secondary compounds are obtained in the usual manner from alkyl 
 ureas and nitrous acid. On reduction they yield the amino-ureas or alkyl- 
 semicarbazides, which are hydrolysed by alkalies to give the alkyl-hydrazines : — 
 J./CH3 JJ/CH3 j,/0H3 CH3.NH.NH2 
 
 \NH, ' xSh, "^NH, - ^^3 
 
 ^^--NHNO 
 
 Nitroso-guanidine, Q=NH , is prepared by reducing nitro-guanidine with 
 
 zinc and sulphuric acid. It is a yellow powder which explodes without melting 
 at 160°. It forms metallic salts of the usual anti-diazo type, but seems in the 
 free state to be a true nitrosamine and not a diazo-hydrate. Thus it is neutral 
 to litmus and is not an electrolyte. This view is supported by the fact that on 
 decomposition it gives almost a quantitative yield of nitrous acid, and scarcely 
 any nitrogen. 
 
 Nitro-derivatives 
 
 The nitro-derivatives of the amides of carbonic acid are obtained by 
 direct nitration. For this purpose it is necessary, as in the aromatic series, to 
 
304 Carbonic Add Derivatives 
 
 employ concentrated nitric acid in the presence of excess of strong sulphuric 
 acid. In some cases the nitro-compounds are decomposed further by excess 
 of nitric acid, and so only the theoretical quantity must be employed. 
 This is most conveniently done by using the nitrate of the amide, which 
 necessarily contains one equivalent of nitric acid, and adding this to 
 excess of strong sulphuric acid, or, in some cases, of acetic anhydride, Nitro- 
 
 /.NH-NOa ^N=N<(*^ 
 
 urethane, CO or CO OH, is made in this way. It is a crystal- 
 
 \OEt \0-Et 
 
 line body melting at 64°, which has an acid reaction and forms salts. From its 
 acid character it is fairly certain that it must in the free state be an iso-nitro- 
 compound. On reduction it gives first nitroso- and then amino-urethane. 
 
 Nitro-urea, obtained by dissolving urea nitrate in strong sulphuric acid at 
 a low temperature, is a crystalline compound which, when dry, is remarkably 
 stable, and does not melt or decompose unless strongly heated ; but in water it 
 breaks up if heated above 60°. It is a strong acid, expelling acetic acid from its 
 salts, and forms salts of neutral reaction. Hence it must be the iso-nitro- 
 compound : — 
 
 (^0 "^ • 
 
 \NH2 
 
 Nitro-guanidine, which is prepared in the same way, is a substance of similar 
 properties. 
 
CHAPTER XIV 
 
 COMPOUNDS CONTAINING A CHAIN OF THREE OR MORE 
 
 NITROGEN ATOMS 
 
 These bodies have been classified by Curtius, on the analogy of the carbon 
 compounds, as derivatives of (mainly hj^othetical) compounds of nitrogen and 
 hydrogen corresponding to the hydrocarbons. Thus we have :— 
 
 NH3, Ammonia, analogous to GH4. 
 
 NH2NH2, Hydrazine, analogous to CHj-CHg. 
 
 NHj-NH-NHg, Prozane, analogous to CHg-CHg-CHg, propane. 
 
 NHj-NHNH-NHa, Buzane, analogous to CHs-CHg-CHa-CHg, butane, 
 and so forth. This nomenclature, though it is systematic, has not been much 
 used, as nitrogen chains are much less stable than carbon chains, and therefore 
 form a less satisfactory basis of classification. 
 
 DIAZOAMINO-COMPOUNDS 
 
 Of the three-nitrogen compounds the most important are the derivatives of 
 
 prozene, NH2-N=:NH, which may be regarded as the amidine of nitrous acid. 
 
 To this class belong the diazoamino-compounds, which, in the aromatic series, 
 
 have long been known. They are readily formed by the action of aromatic 
 
 diazo-compounds on aromatic (and in some cases fatty) amines in neutral or 
 
 acetic acid solution, and are of course intermediate products in the formation of 
 
 the aminoazo-dyes : — 
 
 Ar.N=N-OH + H2N.Ar = Ar-N=N-NH-Ar + H2O. 
 
 This method of formation can be modified in various ways. Solid diazobenzene 
 
 chloride and normal potassium benzene-diazotate will combine with amines in 
 
 the absence of water. The symmetrical derivatives can also be got directly by 
 
 the action of one molecule of nitrous acid on two molecules of an aromatic 
 
 amine. Another reaction which leads to them is that of the nitroso-anilides on 
 
 the primary amines : — 
 
 0-N-N=O 
 
 ^ I + H2N-(^ = ^-NH-N^N-^ + CHs-CO-OH. 
 
 C0*CH3 
 
 The method of preparation through the diazo -compounds can of course only give 
 
 derivatives containing at least one aromatic nucleus. But a general method 
 
 of preparation is afforded by the action of an azide on an organo-magnesium 
 
 compound ' : — 
 
 MgBr 
 
 E-N<]f + EjMgBr = R.lI-N=N-Bi -> R-NH-N=N-Bi + MgBrOH. 
 
 N 
 
 1 Dimrotb, Ber. 36. 909 (1903). 
 
306 Diazoamino-compounds 
 
 By means of this reaction Dimroth ' has succeeded in preparing the previously 
 unknown diazoainino-para£Bns, including the mother substance of the group, 
 diazoamino-methane CHa-NHN^N-CHa, or, as it is more conveniently named, 
 dimethyl-triazene. It is made by the action of methyl azide CHj-Ng on 
 methyl magnesium iodide. It is very difficult to isolate, being decomposed at 
 once by all acids, even carbonic, and being also miscible with water and volatile 
 even with ether vapour. It was obtained from the ethereal solution in the form 
 of its cuprous compound CHj-NCu-N^N-CHg. Acids cannot be used to de- 
 compose this compound, as they destroy the triazene, and it was therefore 
 mixed with an equivalent of diazoamino-benzene, which is more acidic, and 
 therefore takes up the copper, and the dimethyl-triazene was distilled over 
 under reduced pressure. It is a colourless liquid, which solidifies on cooling to 
 crystals which melt at - 12°. It boils with slight decomposition at 92°, and 
 explodes when suddenly heated. In spite of its power of forming salts with 
 metals, its aqueous solution has an alkaline reaction, but it does not form salts 
 with acids, as they decompose it at once ; even water containing carbon dioxide 
 immediately breaks it up, two-thirds of the nitrogen being evolved, and methyl 
 alcohol and methylamine produced : — 
 
 CHgNH-N^NCHa + H2O = CHg-NHj -1- N2 -f CH3OH. 
 In pure water it is fairly stable, but on addition of colloidal platinum nitrogen 
 is evolved. 
 
 In the same way Dimroth ^ has prepared benzyl-methyl-triazene 
 
 ^-CHa-NgH-CHs, 
 a liquid which is very unstable to acids,. and a series of mixed derivatives, such 
 as methyl-phenyl-triazene (M.Pt. 37°), and others, which are intermediate in 
 properties between the purely fatty and the purely aromatic compounds. As 
 the alkyl groups are replaced successively by aromatic radicals, the stability of 
 the compound, especially to acids, increases, the aryl-triazenes being very fairly 
 stable. At the same time the colour, both of the free triazenes and of their 
 metallic derivatives, increases, as is shown by the following table : — 
 
 Triaeene. 
 Dimethyl- .... Colourless 
 Methyl-benzyl- . . Colourless 
 Phenyl-methyl- . . Colourless 
 Diphenyl- .... Yellow-brown 
 Dimroth ' has shown that the mixed derivatives can be made by treating the 
 diazo-salts with the alkylamines ; but only imder special conditions, as even 
 traces of acid break up the triazene, while if the amine is not in excess the bis- 
 diazoamino-compound (Ar-N=N-)2N-Alk is formed. 
 
 The aromatic triazenes or diazoamino-compounds, which are much better 
 known, are crystalline substances, insoluble in water, acids, and alkalies. They 
 are generally yellow or brown, but in some cases the colour has been found to 
 be due to an impurity : thus diazoamino-toluene, if carefuUy purified, is colourless. 
 
 » Ber. 39. 3905 (1906). ' Ber. 38. 670 (1905). 
 
 s Ber. 38. 2328 (1905). 
 
 Silver salt. 
 
 Gwgrous salt. 
 
 Colourless 
 
 Pale yellow 
 
 Colourless 
 
 Pale yellow 
 
 Yellow 
 
 Orange 
 
 Eed 
 
 Brick red 
 
Diazoamino-compcmnds 307 
 
 They are feebly basic, forming unstable compounds with platinic chloride and 
 unstable salts with strong acids. As in the fatty compounds the imide hydrogen 
 can be replaced by metals, giving salts of the type Ar-N=N-NM-Ar. The 
 mercury, copper, and silver salts are much more stable than those of the alkalies, 
 as we should expect from the metal being attached to nitrogen. 
 
 Their behaviour in many respects is that of diazo-anilides. They easily split 
 up into an amine and a diazo-compound or its decomposition-products. Thus 
 in ethereal solution they are broken up by hydrobromic acid into the diazonium 
 bromide and aniline hydrobromide. Cold concentrated hydrochloric acid con- 
 verts diazoamino-benzene into aniline hydrochloride and chlorobenzene. 
 
 On reduction they give aniline and phenyl-hydrazine. All attempts to obtain 
 as an intermediate reduction-product the prozane or triazane derivative 
 
 Ar-NH-NH-NHAr, 
 hydrazoamino-benzene, have failed ; although, as we shall see later, they have 
 to a certain extent succeeded with some of the fatty derivatives. 
 
 When boiled with dilute hydrochloric acid they give phenol, nitrogen, and 
 aniline, though they are often only slowly decomposed. 
 
 In all these reactions they behave as diazo-compounds, from which, however, 
 they are distinguished by their much greater stability. Thus on heating they 
 usually melt without decomposition at a fairly high temperature (diazoamino- 
 benzene at 98°), and though on further heating they ultimately explode, yet 
 they do so with much less violence than the diazo-compounds. Indeed by 
 mixing them with sand, or by dissolving them in solvents of high boiling-point, 
 such as aniline or liquid paraffin, this decomposition may be made to go quite 
 quietly. It is then found that two-thirds of the total nitrogen is evolved as 
 such, while the residue mainly consists (in the case of diazoamino-benzene) of 
 ortho and para-amino-diphenyl. 
 
 The most important reaction of the diazoamino-bodies is their conversion 
 into the aminoazo-compounds, which has already been discussed. This reaction 
 is greatly assisted by the presence of an amine salt. Its velocity has been 
 investigated by Goldschmidt and Eeinders,^ who dissolved in the amine known 
 quantities of the salt and of the diazoamino-compound, and determined the 
 amount of the latter which remained unchanged after a given time by measuring 
 the volume of nitrogen evolved on heating it with acid. They found that the 
 rate of formation of the aminoazo-compound was proportional to the quantity of 
 diazoamino-body present (monomolecular), and also to the amount of amine salt. 
 If different acids were used, the velocity was proportional to the strength of the 
 acid. Where the ortho-aminoazo-compound is formed (owing to the para 
 position being occupied) the reaction is much slower. 
 
 The structure of the diazoamino-compounds obviously admits of the same 
 
 stereoisomerism as is observed among the simpler diazo-compounds. They may 
 
 Ar-N Ar-N 
 either be syn- or anti-bodies, II or II . Their comparatively 
 
 Ar-NH-N N-NH-Ar 
 
 great stability, the absence of explosiveness, and the slowness with which they 
 
 1 Ber. 29. 1369, 1899 (1896) ; Goldschmidt, Merz, Ber. 30. 670 (1897). For the influence of 
 substituents on the formation of aminoazo-compounda see Morgan, Wootton, J. C. S. 1905. 9S5. 
 
308 Diazoamino-compounds 
 
 couple, all show that they must have the anti-formula. Orlow ' claims to have 
 prepared an unstable isomer of diazoamino-benzene, which is more reactive than 
 the normal form. This may possibly be the syn-compound, but its existence is 
 doubtful. 
 
 There is another interesting question as to the structure of these bodies. 
 The unsymmetrical diazoamino-compounds, i.e. those containing two different 
 aromatic groups, exhibit a peculiar tautomerism. By condensing diazobenzene 
 with toluidine we should expect to get the reaction : — 
 
 0.N=N-OH + HjN-CtH, = ^-N^N-NH-CtH, + HjO, 
 and from diazo-toluene and aniline : — 
 
 C7H7-N=NOH + H2N.0 = C7H,-N=N-NH-^ + H^O. 
 The two compounds with these formulae would be different, but they would be 
 desmotropic, that is, they would differ only in the position of a hydrogen atom, 
 and in the consequent position of the double bond. 
 
 As a fact the product obtained in both these reactions is the same. Moreover, 
 the investigation of its behaviour for a long time led to no results capable of 
 determining which of the two formulae it possesses. For example, on heating 
 it with dilute hydrochloric acid we should expect that (as with the symmetrical 
 compounds) the radical next to the NH would separate as amine, and the other 
 as phenol. But the actual products are all the four possible substances : aniline 
 and cresol, and toluidine and phenol. 
 
 It has, however, been shown by Dimroth that if the decomposition by acids 
 is carried out at 0°, the reaction only goes in one direction, and only a single 
 amine is formed. Now the structure of these bodies can be investigated in 
 another way, which is due to Groldschmidt. They combine with phenyl iso- 
 cyanate to form a diazo-urea: — 
 
 A..N=H.HH.Ar. ^^''t^' .NH-Ar. . Ar-OH 
 
 and this decomposes, splitting off the diazo-group, and leaving a simple di- 
 substituted urea. The constitution of this last body may fairly be assumed to 
 indicate the position of the hydrogen in the original diazoamino-compound. 
 The aryl group which was joined to the imide nitrogen in the original compound 
 will be that which remains in the resulting urea, while the one which was next 
 to the diazo-group will be split off. Goldschmidt has applied this method to 
 a large number of diazoamino-compounds with various substituents in the 
 benzene ring, and finds that the imide group is always next to the most positive 
 radical. Dimroth ' has extended it to his mixed fatty-aromatic derivatives with 
 a similar result, the imide group being always next to the alkyl radical. Thus, 
 for example, from their behaviour with phenyl isocyanate we should conclude 
 that benzene-diazoamino-toluene is ^-N^N-NH-C,!!,, and the phenyl-methyl 
 compound 0-N=N-NH-CH3. Acetyl chloride behaves in the same way, giving 
 a diazo-urea of the same type. 
 
 The remarkable point is that the results obtained by this method are 
 ' C. 06. ii. 1569. " Dimroth, Eble, Gruhl, Ber. 40. 2S90 (1907). 
 
Diazoamino-compounds : Structure 309 
 
 diametrically opposed to those obtained by decomposing the bodies with acids. 
 If methyl-phenyl-triazene is 0-N=N-NH-CH3, we should expect that with acids 
 at 0° it would give diazobenzene and methylamine. But it forms methyl alcohol 
 (or methyl chloride), nitrogen, and aniline, not even a trace of a diazo-compound 
 being produced. A certain amount of light is thrown on this question by the 
 behaviour of the bis-diazoamino-compounds. These bodies are formed by the 
 action of two molecules of the diazo-salt on one of a primary amine, and, as wiU 
 be shown later, their structure must be of the type 0'N=N-N(CH3)-N=N-^. 
 They are decomposed by alcoholic hydrochloric or sulphuric acid below 0°, and 
 instead of giving two molecules of diazo-compound and one of amine, they give 
 only one of diazo (together with one of alkylamine) and evolve one molecule of 
 nitrogen. This formation of nitrogen must be due to a preliminary decomposi- 
 tion into a diazo-body and the triazene, which then breaks up in the normal 
 manner ; and the triazene so formed must, at least in the first instance, have 
 its hydrogen atom attached to the nitrogen which carried the diazo-group : — 
 (0-N=N-)2N-CH3 -^ ^-NaCl + ^-N^N-NH-CHs -♦ 0-OH + N2 + HaN-CHg, 
 We must, therefore, admit that the decomposition with acids does take this 
 abnormal course, and that the reaction with phenyl isocyanate, supported as it 
 is by those with acid chloride and with diazo-salts themselves, gives a true 
 indication of the formulae of these bodies. 
 
 To explain why it is that the same compound is obtained by diazotizing 
 either of two aromatic amines and coupling it with the other, Goldschmidt has 
 advanced a rather elaborate theory, involving the formation of addition-products 
 with the acid present. This has been disproved by Dimroth, who showed that 
 the same body was obtained from the azide of one aryl radical and the organo- 
 magnesium compound of another, in whichever way the reaction was carried 
 out. The theory is, indeed, obviously superfluous, the tautomerism being strictly 
 analogous to that of the anti-diazo-hydrates and the nitrosamines : — 
 
 ArN _^ ArN-H ArN _^ ArNH 
 
 N-OH *^ N:0 ' il-NH-Ari *~ N=N-Ari 
 
 The only important difference is that whereas in the first case the two tautomers 
 belong to quite different chemical types, and so can sometimes be separated, in 
 the case of the diazoamino-compounds they afe so similar that no such separation 
 is possible ; this being what v. Pechmann has distinguished as ' virtual tauto- 
 merism '. The readiness with which the tautomeric change occurs is a further 
 argument for the anti-formula. As we see in the case of the diazo-hydrates, 
 such a change is characteristic of the anti- and not of the syn-compounds ; and 
 this is natural, since the change in position is less in the anti-series. 
 
 Incidentally this mobiUty of the imide hydrogen furnishes a conclusive proof 
 of the correctness of the general structural formula (apart from stereo-chemical 
 considerations) adopted for the diazoamino-compounds. Their method of forma- 
 tion only admits of three possible structures. First, they might be symmetrical 
 ring compounds of the type 
 
 NH 
 
 Ar-N N-Ar' 
 
 1175 X 
 
310 Diazoammo-compounds 
 
 This is at once ruled out by the fact that it is impossible to give an analogous 
 
 formula to the condensation-product of diazobenzene hydrate with secondary 
 
 amines like ethyl-aniline ; and yet these bodies behave exactly like the ordinary 
 
 diazoamino-bodies. Also it is impossible to explain the phenyl isocyanate 
 
 reaction on this formula. Secondly, we might suppose them to be diazonium 
 
 Ar-N-NH-Ar 
 compounds, 111 ; and thirdly, there is the usual diazoamino-foi'mula 
 
 ArN=NNH-Ar. Now the identity of the two desmotropic unsymmetrical 
 bodies which have just been discussed is conclusive in favour of the third 
 formula. It is evident that the two pass into one another with great ease. On 
 the third or diazo-formula this only requires the migration of the imide hydrogen, 
 which, as we know from many parallel instances (such as that of the anti-diazo- 
 hydrates), can occur very easily. But on the second or diazonium structure, it 
 would require not only that the imide hydrogen should go from one nitrogen 
 to the other, but also that the triply linked nitrogen should make a similar 
 migration in the opposite direction: — 
 
 Ar-N-NH-Ar, Ar-NH-N-Ar, 
 
 111 ^ -^ III ; 
 
 N N 
 
 and it is practically impossible that so great a change of structure should occur 
 so easily. 
 
 The first diazoamino-compounds of the aliphatic series to be prepared belong 
 to a quite different class from those mentioned above, and were discovered by 
 Thiele. The starting-point of the investigation is the so-called diazoguanidine, 
 which was discovered by Thiele' in 1892, and was re-examined in 1901 by 
 Hantzsch and Vagt,'' who showed that it is not a diazo-compound at all, but an 
 azide. Thiele found that when amino-g^anidine is treated with potassium 
 nitrite in acid solution, a compoimd of the composition CNjHg-HX is produced, 
 which he supposed to be diazoguanidine : — 
 
 /NH-NHa, HNO3 /NHNa-ONOg 
 
 C=NH + ONOH = C=:NH + 2 H,0 . 
 
 XNH^ XnHj 
 
 This body, if it existed, would be a most remarkable substance. It would be 
 a unique case of the diazotizing of an amino-group attached to nitrogen. Its 
 properties are also very unusual for a diazonium compound. It is not decom- 
 posed even by warming the solution ; it does not give off nitrogen when boiled ; 
 and on treatment with alkali it forms metallic azides. Its salts have an acid 
 reaction, whereas a diazonium salt, with so positive a group as the guanidine 
 residue, should be neutral. Moreover, on treatment with alkali a diazonium salt 
 would first form the basic diazonium hydrate and then the alkaline diazotate. 
 But this body first gives an indifferent unstable compound, and then breaks up 
 into hydrazoic acid and cyanamide. 
 
 For all these reasons it is clear that the body is not a diazonium salt at all. 
 On the contrary, all its reactions indicate that it is, as one would expect from its 
 formation, an azide. The fact that it forms a salt is merely due to the presence 
 
 1 Ann. 270. 1. ' Am. 314. 339. 
 
Fatty Diazoamino-compounds 311 
 
 of the unmodified NH^ group in the guanidine, and the equation which 
 represents its formation is: — 
 
 /NHNH2 
 
 /4 
 
 C=NH + HONO = q=NH + H^O . 
 
 \NH2HNO3 XNHa-HNOs 
 
 It is the nitrate of carbamide-imide-azide. The first effect of treating it with 
 alkali is to liberate the &ee azide, which, if left to itself, rapidly changes into 
 a ring compound, aminotetrazole : — 
 
 
 -<^* - 
 
 .NHN 
 
 
 
 
 If the free azide is 
 and cyanamide : — 
 
 treated with more 
 
 alkali it breaks up 
 HN3 
 
 into 
 
 hydrazoic 
 
 acid 
 
 One remarkable reaction, which appears to favour Thiele's original view, is 
 that when treated with potassium cyanide it is converted, like a diazonium 
 salt, into a true diazocyanide : — 
 
 /NH-N=N-CN 
 (^NH ; 
 
 but Hantzsch has shown that other similar compounds, and especially the azide 
 
 /-N3 
 of urea, CO , behave in the same way. 
 
 It is through these diazo-cyanides that Thiele' was able to prepare the 
 diazoamino-compounds. In fact diazoguanidine cyanide is itself a diazoamino- 
 compound, as is obvious from the formula. It is the amidine-nitrUe of triazene 
 dicarboxylio acid, COaH-NH-N^N-COgH. Like aU the diazo-cyanides it has the 
 full nitrile character, and thus the CN group can be converted into an amide 
 or an ester : — 
 
 C=NH -^^2 or c=:NH "■^*. 
 
 \NH2 NNHg 
 
 These bodies resemble the other diazoamino-compounds in giving off two-thirds 
 of the triazene nitrogen as such on heating with dilute acids ; but they differ 
 from them in being decomposed by alkalies, to which the others are stable, and 
 in being unaffected by cold concentrated acids. 
 
 Thiele has also succeeded in reducing these compounds to triazane derivatives 
 or hydrazoamino-compounds. Like so many bodies with doubly linked nitrogen, 
 they add on sulphurous acid to form the sulphonic acid of the reduced derivative. 
 Since the product no longer behaves as a diazoamino-compound the reduction 
 
 > Ann. 305. 64 (1899). 
 x2 
 
312 Diazoamino-compounds 
 
 must take place in the triazene group, and the only question is whether the 
 SO3H goes to the middle or the end of the nitrogen chain : — 
 
 (1)C=NH SO3H ■ (2)q=NHS03H 
 
 If the resulting substance had the first formula, it ought to be easily oxidized 
 back to a diazoamino-derivative ; but it is stable to oxidizing agents, and 
 therefore must have the second formula. 
 
 All attempts to remove the SO3H and replace it by hydrogen have failed. 
 We should expect the reaction to take place as it does with phenyl-hydrazine : — 
 
 /NH.]|f.NH.C0.NH3 /NH-NH-NH-CO-NH, 
 
 Q=NHS03H + HOH = (^NH + HOSO3H. 
 
 If it is warmed with acids sulphuric acid is split off, but the triazane which 
 must be produced breaks up at once in a complicated manner, giving nitrogen, 
 carbon monoxide, guanidine, and other products. 
 
 Thiele has also attempted to reduce these diazoamino-compounds directly. 
 Under ordinary circumstances this gives, as it does in the aromatic series, 
 a mixture of an amine and a hydrazine derivative. But if the body is carefully 
 reduced with zinc dust and ammonium chloride, a colourless solution is obtained, 
 which gives no reactions of the triazene. The product must have the N3 chain 
 intact, since on oxidation the diazoamino-compound is re-formed. It has strong 
 reducing properties, and on warming readily decomposes into the same sub- 
 stances as are obtained by splitting off sulphuric acid from the sulphonic acid. 
 It is evident, therefore, that in both cases the true triazane derivative, 
 
 /NHNH-NH-CO-NHa 
 q=NH 
 \NH2 
 is formed, but that it is too unstable to be isolated. This great instability 
 makes it improbable that any attempts to prepare triazane itself, NH2-NH-NH2, 
 will succeed. 
 
 Darapsky ' has endeavoured to form triazane derivatives by an extension of 
 Schestakow's application of the Hofmann reaction. Schestakow obtained 
 hydrazine compounds in this way from derivatives of urea (p. 187), and Darapsky 
 applied the reaction to hydrazine derivatives of urea, which should produce 
 triazanes, phenyl-semicarbazide, for example, giving phenyl-triazane thus : — 
 
 0-NH-NH-CO-NH2 -* 0-NHNH-N=C=O -» 0-NH-NH-NH2. 
 But in no case was such a triazane obtained. The oxidation always proceeded 
 further, with the formation either of an azide, 
 
 HaN-CONH-NHNHa -* H2NC0.N<^ -* H-Ng, 
 
 or of nitrogen itself. This affords a further illustration of the instability of the 
 triazane derivatives. 
 
 J. pr. Ch. [2] 76. 433 (C. 08. i. 452). 
 
Tetrazane Derivatives 313 
 
 BUZANE OK TETEAZANE DEBIVATIVES 
 
 Derivatives of bu2ane, NHa-NH-NH-NHg, are v. Pechmann's hydrotetrazones; 
 which are formed by oxidizing hydrazones with amyl nitrite : — 
 
 0-NH "^ HN-0 ~ ^ ~ ^-N— N-0 
 
 They all dissolve in concentrated sulphuric acid to give brilliantly coloured 
 solutions ; the compound whose formula has just been given (yellow needles, 
 M.Pt. 180°) forming a deep blue solution. This is probably the origin of the 
 BUlow reaction, the brilliant colour which hydrazones give with ferric chloride 
 or potassium bichromate in concentrated sulphuric acid. 
 
 The hydrotetrazones are readily reduced back to hydrazones. When heated 
 with alcoholic potash they undergo a remarkable change — ^a sort of reversal of 
 the Beckmann reaction — to form benzil osazone :— 
 
 0-CH HC-A <t>-C C-^ 
 
 ^11 II ^ ^11 I! ^ 
 
 The buzylene compounds or tetrazenes, as they are properly called (though 
 they are generally known as tetrazones), are of two kinds, the derivatives of 
 symmetrical tetrazene, NHa-N^NNHj, and of unsymmetrical, NH=N-NHNH2. 
 The former are obtained by oxidizing the unsymmetrical secondary hydrazines 
 with mercuric oxide * : — 
 
 (CH3)aN-NH2 + 0^ + H2N-N(CH3)2 = (CH3)2N.N=N-N(CH3)2 + 2H2O. 
 
 They are strongly basic colourless oily substances, easily volatile with steam, 
 which explode when heated to a somewhat high temperature. They reduce 
 silver nitrate in the cold, and form very soluble and unstable salts. When the 
 solution is boiled, half the nitrogen comes off as such, with the formation of 
 an amine and an aldehyde. 
 
 The aromatic tetrazones are very similar.^ 
 
 The derivatives of tmsymmetiical tetrazene, NH^N-NH-NHj, are obtained 
 by the action of diazo-compounds on hydrazines in acetic acid solution,^ and 
 hence are known as diazo-hydrazides. Their structure has been determined in 
 the following way.* Benzal-benzyl-hydrazone, 0-CH2-NH-N=CH-0, combines 
 with nitrophenyl-diazo-hydrate to form a diazo-hydrazide which can only have 
 the formula : — 
 
 N02-CcH,.N=N-N<g=JJ-'^. 
 This same compound is formed from the diazo-hydrazide got by treating with 
 
 ' Cf. Franchimont, van Erp, Sec. Trav. 14. 317 (1896). 
 
 " Cf. Wieland, Ber. 41. 3498 (1908). ' Wohl, £er. 26. 1687 (1893). 
 
 * Wohl, Sohiff, Ber. 33. 2741 (1900). 
 
314 Tetrazane Derivatives 
 
 benzaldehyde the diazo-hydrazide obtained from nitrophenyl-diazo-hydrate and 
 benzyl-hydrazine. Hence this last-named body must have the structure 
 
 and in general the diazo-hydrazides must be of the type Ar-N=N-N<( • ^ . 
 
 They are unstable basic substances, which break up on reduction into phenyl- 
 hydrazine and two molecules of hydrazine. 
 
 BIS-DIAZOAMINO-COMPOUNDS 
 
 These bodies, containing a chain of five nitrogen atoms, are formed by the 
 
 condensation of two molecules of a diazo-compound with one molecule of 
 
 ammonia or a primary amine ' : — 
 
 ^■N=NOH + NH3 + HO-N=:N-0 = 0-N=N-NH-N=N-^ + 2 H2O. 
 
 It is conceivable that the reaction might go in a diiferent way, diazobenzene 
 
 0-N-N=N-CH3 
 combining, for example, with methylamine to give a body „_!, 
 
 Now these bodies are also formed by the action of a diazo-compound on an aryl- 
 alkyl-triazene, and if they had the unsymmetrical formula one would obtain 
 a different product by treating phenyl-methyl-triazene with diazotoluene and 
 tolyl-methyl-triazene with diazobenzene. Experiment '^ showed, however, that 
 the two products were identical, so that the bis-diazoamino-compound must 
 have the sjonmetrical structure Ar-N=N-N(Alk)N=N-Ar. These bodies are 
 coloured, generally very explosive substances. When boiled with alkali they 
 break up in the normal manner into two molecules of phenol, one of amine, and 
 two of nitrogen. Their singular decomposition with acids has already been 
 mentioned. Those which are derived from alkylamines give, even at 0°, one 
 molecule of diazo, one of nitrogen, one of phenol, and one of alkylamine. This 
 is no doubt due to the preliminary formation of the diazoamino-compound : — 
 (Ar-N=N-)2NCH3 -^ Ar.N=NOH + ArN=N-NH-CH3 
 
 -^ Ar-OH + N2 + H2N-CH3, 
 Bis-diazoamino-benzene itself, instead of giving, as we should expect, phenol, 
 nitrogen, and ammonia, forms phenol, aniline, and nitrogen. This may be ex- 
 plained in a similar way, by the intermediate production of phenyl-triazene : — 
 (0.N=N-)2NH -> 0.N=N-OH + ^-N^N-NHg -♦ 0-OH + Ng -i- ^-NHa + Nj. 
 
 OCTAZONES 
 The longest known chain of nitrogen atoms consists of 8, and occurs in the 
 octazones. These curious bodies were obtained by Wohl and Schiff ' by oxidizing 
 the diazo-hydrazides with potassium permanganate : — 
 
 2 0.N=N.N0-NH2 -)- 20 = 0-N=N-N0.N=N.N0-N=N-0 + 2H2O, 
 the reaction being exactly analogous to that by which a hydrazine is converted 
 into a tetrazone. 
 
 The octazones are yellow bodies which are, as one might expect, very 
 unstable and highly explosive. 
 
 ' V. Peohmann, Frobenius, Ber. 27. 703, 898 (1894) ; 28. 170 (1895). 
 '^ Dimroth.Eble, Gruhl, Ber. 40. 2390 (1907). ' Ber. 33. 2741 (1900). 
 
CHAPTER XV 
 
 URIC ACID DERIVATIVES 
 
 These bodies are, strictly speaking, heterocyclic compounds, but from their 
 close relationship to the open-chain compounds, and the ease with which they 
 pass into them, they occupy a somewhat anomalous position, and are most con- 
 veniently treated as a class by themselves, before we come to deal with the 
 various types of nitrogenous rings. 
 
 The central point of the whole group is uric acid itself ; but this cannot be 
 discussed until after the consideration of some of the simpler compounds from 
 which it is built up. 
 
 Urea, as a substituted ammonia, can combine with acids with loss of water 
 to form a series of compounds analogous to the amides. These are known as 
 ureides : e. g. : — 
 
 /NH2 /NH-COCHs 
 
 CO + HOCOCH3 = CO + H„0: 
 
 Acetyl urea, 
 as NH3 -f HOCOCH3 = NH2COCH3 + H2O. 
 
 If this condensation occurs twice we get di-ureides, such as 
 
 /NHCOCH; 
 
 QO^ 
 
 3 
 
 NHCO-CH 
 
 3 
 
 But if, instead of two molecules of a monobasic acid, one molecule of a dibasic 
 acid combines with one of urea, the resulting compound contains a closed ring. 
 Such compounds are called cyclic ureides : an example is oxalyl-urea or parabanic 
 
 /NH-CO 
 acid, CO I . 
 
 \nh-co 
 
 It is unfortunate that as many members of this group were known long 
 before their formulae were made out, they have received a series of irrational 
 names ; but these names are now so weU established and so constantly employed 
 that they cannot be altogether neglected. 
 
 In behaviour the cyclic ureides show a certain resemblance to the imides of 
 the dibasic acids, such as succinimide. There is the same tendency to break the 
 ring in the presence of alkali, with the formation of mono-ureides which are at 
 the same time acids : the ultimate effect of alkali being of course to split up the 
 compound completely into the dibasic acid and urea, or ammonia and carbonic 
 acid. Compare 
 
316 Uric Acid Group 
 
 CHo-CO CHo-COOH CH2-COOH 
 
 I >NH ^1 -> I +NH3: 
 
 CH2-CO CHa-CONH^ CH^COOH 
 
 Succinimide. Succinamic acid. 
 
 CO-NH COOH COOH NHj 
 
 and I >C0 -♦ I ^CO-NHg -> | + >C0 . 
 
 CO-NH CONH COOH NHg 
 
 Oxalyl-urea. Oxaluric acid. 
 Parabanic acid. 
 
 Oxalyl-urea is obtained from oxalic acid and urea by the action of phosphorus 
 pentachloride. Bromine converts it into oxaluric acid, which is changed back 
 into oxalyl-urea by phosphorus oxychloride, and is broken up by alkali into urea 
 and oxalic acid. 
 
 Of greater importance is malonyl-urea or barbituric acid, 
 
 CO-NH 
 
 CH2 CO, 
 
 ^CO-N'H 
 
 which may be got from malonic acid, ui'ea, and phosphorus oxychloride. It is 
 a crystalline substance which is broken up into its constituents on boUing with 
 alkali. The hydrogen atoms of the methylene group, as in malonic acid itself, 
 are very reactive. They can be replaced by bromine, -NO2, =NOH, &c., with 
 great ease, and thus indirectly by other groups as well. Malonyl-urea acts as an 
 acid, forming salts in which these hydrogens are replaced by metal ; and these 
 salts, when treated with alkyl iodide, give C-alkyl derivatives. In these ways 
 a series of derivatives of malonyl-urea have been prepared which have been of 
 the greatest value in determining the constitution of the members of the uric 
 acid group. 
 
 For example, if dibromo-barbituric acid is reduced with hydrogen sulphide, 
 it gives hydroxy-barbituric acid, dialuric acid, or tartronyl-urea, 
 
 NH- 
 
 <f« ^<0H- 
 NH-CO 
 
 >- tn 
 
 Again, the dihydroxy-derivative, mesoxalyl-urea, 
 
 NH— GO NH— CO 
 
 (jO (|<^g or CO CO + H^O, 
 
 NH— CO NH— CO 
 
 is alloxan. This is one of the oxidation-products of uric acid, and its constitu- 
 tion is shown by the fact that its oxime is identical with the isonitroso-compound 
 obtained from barbituric acid and nitrous acid, which must be 
 
 NH-CO 
 
 CO C=NOH. 
 
 NH-CO 
 
Barbituric Add Derivatives 317 
 
 This last body is known as violuric acid.' It is broken up by alkali into urea 
 and isonitroso-malonic acid. 
 
 Fuming nitric acid converts barbituric acid into the nitro-derivative, dilituric 
 acid (whose formula is given below), which is also formed by the oxidation of 
 violuric acid. 
 
 Finally, by the reduction of violuric or of dilituric acid is formed amino- 
 barbituric acid or uramil. 
 
 The formulae and common names of the more important of these compounds 
 are given in the following table. They are all derived from barbituric acid or 
 malonyl-urea by modifying the methylene group. The ring represents the group 
 
 NH— CO 
 
 V« , ■ 
 NH— CO 
 
 which they all contain. 
 
 CHg. Malonyl-urea, barbituric acid. 
 
 0\qjj . Hydroxy-malonyl-urea, dialuric acid. 
 
 VOH 
 
 ^ 
 
 OH ""^ 
 
 Cfc=0. Dihydroxy-malonyl-urea, mesoxalyl-urea, alloxan. 
 
 C=N0H. Isonitroso-malonyl-urea, violuric acid. 
 
 0\jjQ . Nitro-malonyl-urea, dilituric acid. 
 
 ~k/^ 
 
 S'^NH ■ -Ajoino-barbituric acid, uramil. 
 
 Uric Acid 
 
 Uric acid was discovered in urinary calculi by Scheele in 1776, and simul- 
 taneously by Bergmann. It occurs in various parts of the animal organism — 
 muscles, blood, urine, &c. — especially in the carnivora ; also in guano, up to 
 25 per cent. , and to a still greater extent — 90 per cent. — in the excrement of 
 serpents. It forms small crystals, without smell or taste. It is insoluble in 
 alcohol and ether, but dissolves in 10,000 parts of water at 18-5°, and in 1,800 
 at 100°. 
 
 ' Violuric acid is reiiui.rkable for the changes of colour which it undergoes. In the solid state it 
 is colourless : in solution, especially in the presence of small quantities of bases, it is blue (see 
 Donnan, Schneider, J. G. 8., 1909. 956). Its esters are colourless, and its salts in the solid state 
 are colourless, yellow, red, and blue. The colourless salts and the esters are evidently of the type 
 
 _C=0 
 
 I • The coloured salts must have three other structures, but it is at present quite uncertain 
 
 -C=NOR 
 
 what these may be. Cf. Hantzgch, Ser. 42. 966 ; Hantzscb, Isherwood, ib. 986 ; Hantzsch, Issaias, 
 ib. 1000 (1909). 
 
318 Uric Add Group 
 
 It is a feeble dibasic acid. Its alkaline salts are mostly very insoluble, and 
 are highly hydrolysed in solution. The potassium salt requires 800 parts of 
 water to dissolve it, and the sodium and ammonium salts a still larger quantity. 
 The lithium salt is more soluble, dissolving in 368 parts of cold water, and the 
 
 salt of piperazine, NH /NH, still more so, requiring only fifty parts at 
 
 ^^CH2~Cxl2 
 
 17°. Hence lithium, and more recently piperazine, have been used in medicine 
 to remove deposits of uric acid in the body in gout and rheumatism ; though it 
 is doubtful whether their use, or at any rate this explanation of it, does not rest 
 on a physico-chemical error. 
 
 The first systematic investigation ^ of uric acid was that of Liebig and Wohler 
 (1826-1838), who discovered an immense number of its reactions, and converted 
 it into a series of other substances, many of which were already known to exist 
 in nature. But their work threw little or no light on its constitution. Our 
 knowledge of this is due in the first place to Baeyer, who began his researches 
 on it in 1863.° He showed that the whole class of substances included in the 
 group could be regarded as derivatives of barbituric acid, which he proved to be 
 malonyl-urea from its hydrolysis to ammonia, carbon dioxide, and malonic acid. 
 He also prepared from it the hydroxy- (dialuric acid), the keto- (alloxan), the 
 isonitroso- (violuric acid), and the nitro-derivative (dilituric acid). He proved 
 the constitution of the last two by reducing them to uramil, which was shown 
 to be amino-malonyl-urea by its giving the hydroxy-compound (dialuric acid) 
 when treated with nitrous acid. This uramil combines with cyanic acid to give 
 a body which Baeyer called pseudo-uric acid. The reaction must go in this way : — 
 NH— CO NH— CO 
 
 CO CHNHj + C<Q^ = CO CHNH-CONHa. 
 NH— CO NH— CO 
 
 Pseudo-uric acid differs in composition from uric acid only by containing the 
 elements of one molecule of water more ; and Baeyer hoped to be able to convert 
 it into uric acid by the action of dehydrating agents. But he eould not bring 
 about this change, though E. Fischer subsequently showed that it is possible. 
 
 In 1875 Medicus ' published a most remarkable paper on the constitution of 
 this group of compounds. He produced hardly any new facts ; but sometimes 
 on the basis of facts already known, and sometimes, apparently, on no basis at 
 all, he suggested formulae for nearly every known compound of the group — 
 uric acid, xanthine, caffeine, theobromine, guanine, and hypoxanthine. The 
 singular point is that with the exception of the position of one of the methyl 
 groups in theobromine, all the formulae which Medicus proposed are absolutely 
 correct ; although it was not until the most recent work of Fischer, in which he 
 revised and modified many of his own previous formulae, that they were 
 recognized as being so. 
 
 We are at present concerned only with uric acid. For this Medicus proposed 
 the now accepted formula : — 
 
 ' Tor a summary of the histoi-y of this subject see Lachman, Spirit of Organic Chemistry. 
 2 Arm. 127. 1, 199. ^ ^„„ i^g^ 236. 
 
Uric Acid: Structure 319 
 
 NH— CO 
 
 CO C-NHv„^ 
 
 I II >co, 
 
 NH— 0-NH^ 
 on the following grounds : — 
 
 1. On heating with acids it gives glycocoll, ammonia, and carbon dioxide 
 (Strecker). 
 
 NH— CO 
 
 2. On oxidation it gives alloxan, CO CO, and urea (Liebig and Wshler). 
 
 NH— CO 
 
 3. On treatment with alkaline permanganate it gives allantoin, which breaks 
 up into glyoxylic acid, CHOCOOH, and urea. 
 
 It is obvious that these arguments are quite insufficient to establish his 
 formula, and it is not easy to see how they led him to adopt it. 
 
 In 1882 E. Fischer ' published a detailed examination of uric acid and its 
 derivatives, in which the constitution of uric acid was finally determined, in 
 accordance with Medicus's suggestion, but on much more satisfactory evidence. 
 Baeyer had shown that uric acid must be formed from malonyl-urea by the 
 addition of another molecule of urea and the loss of four atoms of hydrogen and 
 two of oxygen. The whole question was how this loss occurred. Baeyer had 
 proposed a formula in which no second ring was formed ; Medicus had suggested 
 the one which has just been given ; while Fittig had put forward yet a third 
 formula, also assuming the production of a second ring, and attractive on 
 account of its symmetry : — 
 
 
 
 NH 6 NH 
 
 NH — CO 
 
 NH— CO 
 
 
 \ 1 
 
 
 1 
 
 CO 
 
 CO CO 
 
 CO HC-NHC=N 
 
 CO CNH 
 
 
 / 1 
 
 1 
 
 >C0 
 
 NH C NH 
 
 NH CO 
 
 NH— CNH 
 
 
 Baeyer. 
 
 Medicus. 
 
 Fit 
 
 tig. 
 
 Before the publication of E. Fischer's paper, Fittig's view was the one 
 generally adopted. It is to be noticed that his formula is symmetrical about 
 both of the dotted lines, so that all the four NH groups are similarly related to 
 the rest of the molecule. 
 
 Fischer showed in the first place that there are four imide (NH) groups in 
 uric acid ; for a tetramethyl derivative can be prepared which on saponification 
 yields all its nitrogen in the form of methylamine, and gives no ammonia. 
 Hence all the four methyls are attached to the four nitrogen atoms. This makes 
 Baeyer's formula impossible, as it only contains three NH groups, the other 
 hydrogen being attached to carbon ; but it does not decide between Medicus and 
 Fittig. 
 
 Secondly, Fischer obtained two different monomethyl-uric acids. Hence 
 the compound is not symmetrical, and Fittig's formula must be rejected. More- 
 
 1 Ann. 215. 253. 
 
320 Uric Add Group 
 
 over, one of these monomethyl derivatives on oxidation gives alloxan and methyl- 
 urea, while the other gives methyl-alloxan and urea ; showing that one of the 
 NH groups is not contained in the alloxan or barbituric acid ring. The isomerism 
 and behaviour of the numerous other methyl derivatives prepared by Fischer are 
 found to be in full accordance with Medicus's formula ; but the two reactions 
 which have been mentioned are the most important, and are in themselves quite 
 a suflScient proo£ 
 
 The three most important syntheses of uric acid are :— 
 
 1. From acetoacetic ester (Behrend, 1888). 
 
 2. Prom malonic acid, begun by Baeyer in 1863, and completed thirty-two 
 years later (1895) by E. Fischer. 
 
 3. From cyanacetic acid (W. Traube, 1900). 
 
 1. Behrend's Synthesis 
 
 Acetoacetic ester condenses in the enolic fonn with urea to give j8-uramino- 
 crotonic ester ; this on saponification gives the corresponding acid, which loses 
 water to form a cyclic ureide, methyl-uracil : — 
 
 NHg EtOCO NH2 EtOCO NH— CO 
 
 CO + CH -» CO CH -* CO CH . 
 
 NH2 HOCCH3 NH CCH3 NH— CCH3 
 
 Acetoacetic ^-uramino-crotonic Methyl- 
 
 ester, ester. uracU. 
 
 Nitric acid nitrates methyl-uracil in the CH group, while the methyl is oxidized 
 to carboxyl, the product being nitro-uracil carboxylic acid. When this is boiled 
 with water it loses carbon dioxide and forms nitro-uracil : — 
 
 NH— CO NH— CO 
 
 (^0 C.NO2 -^ (^0 C-NO^. 
 
 NH— CCOOH NH— CH 
 
 Nitro-uracil Nitro-uracil. 
 
 carboxylic acid. 
 
 Nitro-uracil, on treatment with tin and hydrochloric acid, is converted partly into 
 amino-uracil and partly also into oxy-uracil: the former dissolves in the acid 
 liquid, while the latter remains undissolved, and so can be filtered off. The oxy- 
 uracil is oxidized by bromine water to dioxy-uracil (isodialuric acid), which 
 condenses with urea in the presence of sulphuric acid to give uric acid : — 
 
 NH— CO 
 
 NH— CO 
 -* CO COH H2N. 
 
 1 1 + >co 
 
 NH— COH n^w 
 
 NH 
 
 1 
 
 CO COH 
 NH— CH 
 
 -J. CO 
 NH 
 
 Oxy-uracil. 
 
 Dioxy-uracil. 
 
 U 
 
 r 
 
 C-NH 
 
 II >co. 
 
 -C-NH 
 Uric acid. 
 
 2. Baeyer and Fischer's Synthesis 
 The first part of this synthesis has already been described. Baeyer showed 
 that malonic acid condenses with urea in the presence of phosphorus oxychloride 
 
Syntheses of Uric Add 321 
 
 to give malonyl-urea or barbituric acid : that this is converted by nitrous acid 
 into the isonitroso-compound, violuric acid : — 
 
 NHa HOCO NH— CO NH— CO 
 
 CO + CHa -> CO CH2 -♦ CO C=NOH : 
 
 NH2 HOCO NH— CO NH— CO 
 
 and that this on reduction gives amino-malonyl-urea or uramil, which condenses 
 with cyanic acid to form pseudo-uric acid : — 
 
 NH— CO NH— CO 
 
 y^ y<^NHa + ^^NH -* 9^ y^^NH-CO-NHa" 
 NH— CO NH-CO 
 
 Beyond this point Baeyer was unable to go. All that was wanted was 
 that the pseudo-uric acid should lose one molecule of water to form uric acid ; 
 but he could not make it do this. It was not until thirty-two years later that 
 E. Fischer showed that the reaction can be brought about by treatment with 
 fused oxalic acid, and even by boiling with mineral acids, uric acid being 
 produced. This change is most simply expressed if we suppose the pseudo- 
 uric acid to react in the enolic form : — 
 
 NH-CO NH— CO 
 
 CO CNHCONHo -♦ CO CNH . 
 
 I II I ll>co 
 
 NH— COH NH— CNH 
 
 Pseudo-uric acid. Uric acid. 
 
 3. Traube's Synthesis 
 
 This work,' which is subsequent to Fischer's investigations, starts with 
 cyanacetic acid. Traube finds that cyanacetic acid readily condenses with urea 
 in presence of phosphorus oxychloride to give cyanacetyl-urea. This body 
 when treated with dilute alkalies is converted into the isomeric cyclic pyri- 
 midine derivative, which is imino-barbituric acid. This yields with nitrous 
 acid an isonitroso-derivative, which on reduction gives the corresponding 
 
 NH-CO 
 CO CHa 
 NHa CN 
 
 NH— CO 
 
 -^ <^0 C^B, 
 NH— C=NH 
 
 amino-compound : — 
 
 NH— CO NH— CO 
 
 . CO C=NOH -• CO CH-NHa. 
 
 NH— C=NH NH— C=NH 
 
 This last body, which may also be written in a tautomeric form as a di- 
 amino-compound, when shaken with chlorocarbonic ester in alkaline solution, 
 gives a urethane ; and the sodium salt of this urethane, when heated to 180- 
 190°, loses water and forms the sodium salt of uric acid :— 
 
 1 Ber. 33. 1371, 3035 (1900). 
 
322 Uric Acid Group 
 
 NH-CO NH-CO ^H— CO -^K-^O 
 
 CO CHNH„ or CO ONH» -^ CO CNHCOOEt -* CO CNH . 
 
 I I I II I II I ll>co 
 
 NH— C=NH NH— CNH2 NH— CNH2 NH— CNH 
 
 The other derivatives must be dealt with more briefly, begimiing with the 
 mother substance of the whole group, which historically was the last to be 
 prepared. Uric acid contains three carbonyls adjacent to imide groups. Each 
 of these carbonyls is capable of passing over, in the presence of a suitable 
 reagent, into an enol (hydroxyl) group. Thus on treatment with phosphorus 
 oxychloride these three hydroxyls are replaced by chlorine : — 
 
 NH— CO N=COH N=C-C1 
 
 CO CNH -> COH CNH -> CCl CNH 
 I II >C0 II II >COH II II >C.C1 
 NH— C-NH N C-N^ N C-N^ 
 
 Now all the complicated uric acid derivatives, such as caffeine and xanthine, 
 contain this carbon and nitrogen skeleton ; and as it is necessary to have 
 a rational nomenclature for them, Fischer proposed in 1884 to call the then 
 hypothetical hydrogen compound, of which the last-mentioned body is the 
 trichloro-substitution product, purine (from purum and uricum), and to name 
 the other substances from it, the atoms being numbered thus : — 
 
 iN==CH 
 
 I ^1 
 
 „CH sC-^NH . 
 
 II II '>H 
 
 3N — jc— ^nA 
 
 The preparation of purine itself presented extraordinary difficulties, as all 
 the usual reducing agents when applied to trichloro-purine either failed to 
 reduce it or broke up the molecule altogether. It was not vmtil 1898 that 
 Fischer succeeded in obtaining it, which he did as follows.' 
 
 Trichloro-purine, from uric acid and phosphorus oxychloride, was treated 
 with hydriodic acid and phosphonium iodide at 0°. This replaces one chlorine 
 by hydrogen and the other two by iodine. The resulting diiodo-purine can 
 be shown to have the iodine in the positions 2 and 6. When it is reduced 
 with zinc dust alone in boiling aqueous solution, the two iodine atoms are 
 replaced by hydrogen, and purine is produced : — 
 
 N^C-Cl T^^^?^ V"^^^ 
 
 CCl C-NH -♦ C-I C-NH -> CH C-NH 
 II II >C.C1 i II >CH II II >CH 
 
 N C-N^ N C-N^ N C-N^ 
 
 Trichloro- Diiodo- Puiine. 
 
 Purine is a colourless crystalline substance melting at 216°. It volatilizes 
 partly undecomposed. It is remarkably soluble in water and alcohol, but 
 only slightly in ether and chloroform. It combines with one equivalent of 
 
 ^ Ber. 31. 2550. 
 
Purine 
 
 323 
 
 acids to form salts such as CgH^N^-HNOs. It also forms salts with, bases, 
 the alkaline salts being excessively soluble in water. It is fairly stable to 
 oxidizing agents, not being attacked by chromic acid even on boiling, nor 
 oxidized by potassium permanganate at once in the cold. 
 
 The numerous naturally occurring derivatives of purine, whose constitu- 
 tion has been established by E. Fischer, may be divided into three classes, 
 according as they are derived from mono-, di-, or trihydroxy-purine. The last 
 of these is of course uric acid itself, written on the trienolic formula: while 
 the mono-hydroxy-compound is hypoxanthine, and the di-hydroxy- is xanthine, 
 both of which substances occur in the animal organism :^ 
 
 N=C.OH 
 
 CH C-NH 
 
 II II >CH 
 N C-IT 
 
 N= 
 
 I 
 
 =?■ 
 
 OH 
 
 COH C-NH 
 
 or 
 NH— 
 CH 
 
 N- 
 
 or 
 
 -C-1 
 
 N=C!.OH 
 
 COH C-NH 
 
 >C-OH 
 
 -C-N^ 
 
 r 
 
 C-NH 
 
 II II >CH 
 N C-N^ 
 
 6-Oxy-purine. 
 Hypoxanthine. 
 
 NH- 
 CO 
 
 C-NH 
 
 I II >CH 
 NH-C-N'^ 
 
 2, 6-Dioxy-purine. 
 Xanthine. 
 
 NH 
 I 
 CO 
 
 T 
 
 O-NH 
 
 I II >co 
 
 NH— C-NH 
 
 2, 6, 8-Trioxy-pm-ine. 
 Uric acid. 
 
 The compounds of this group may be prepared from uric acid.^ This is 
 first converted into trichloro-purine. Of the three chlorine atoms which this 
 body contains, the most mobile is in position 6 ; while that at 2 can also be 
 readily though less easily removed. The other chlorine atom (at 8) is more 
 firmly attached. It is on these differences that the whole of these syntheses 
 depend. 
 
 If trichloro-purine is treated with aqueous ammonia, the chlorine at 6 is 
 replaced by NHg, giving diehlor-adenine. If it is treated with aqueous 
 potash, the same chlorine atom is replaced by hydroxyl, giving dichloro- 
 hypoxanthine. These two bodies can have the remaining chlorine atoms re- 
 placed by hydrogen, by treatment with hydriodic acid, being thereby converted 
 respectively into adenine and hypoxanthine. Further, alcoholic ammonia acts 
 on dichloro-hypoxanthine to replace the next chlorine atom (at 2) by NHg, 
 giving 2-amino-6-oxy-8-chloro-purine, which is reduced by hydriodic acid to 
 guanine. The constitution of this body as a guanidine derivative is shown 
 by its splitting off guanidine on oxidation. 
 
 Finally, if trichloro-purine is treated with sodium ethylate, the two chlorine 
 atoms at 6 and 2 are replaced by ethoxy-groups, forming diethoxy-ohloro- 
 purine, and when this is reduced the chlorine is replaced by hydrogen at the 
 same time that the ethoxy-groups are saponified, and 2,6-dioxy-purine or 
 xanthine is formed. These relationships are shown in the following table : — 
 
 ' E. Fischer, Ber. 30. 2220, 2226 (1897). A brief but clear account of the relationships and 
 syntheses of these bodies is given in Bichter's Organisers Chenie, vol. i, pp. 609-14 (10th ed., 
 1903). 
 
324 
 
 Uric Add Group 
 
 =(J!NH. 
 C-NH 
 
 1,^ 
 
 CCl 
 
 II 
 N— 
 
 Dichlor-adenine, 
 
 i 
 N=CNH2 
 
 CH C-NH 
 
 II II > 
 
 Adenine. 
 
 N=C-C1 
 ,CC1 C-NH 
 
 II >c-ci 
 
 C-N^ 
 
 Trichloro-purine. 
 
 i 
 COH 
 
 CCl C-NH 
 
 II II >c-ci 
 
 N C-N^ 
 
 Dichloro-hypoxanthine. 
 
 1 
 =COH 
 
 C-NH 
 
 ■ II II >c.ci 
 
 N C-N^ 
 
 Amino-6-oxy-8-chlorO' 
 purine. 
 
 I 
 =COH 
 
 C-NH 
 
 II >CH 
 
 N=COH 
 
 CH C-NH 
 
 II II >CH 
 N C-N^ 
 
 Hypoxanthine. 
 
 N= 
 I 
 H,NC 
 
 N C-N 
 
 Guanine. 
 
 !-OEt 
 C-NH 
 
 >cci 
 
 N C-N^ 
 
 Diethoxy-chloro-purine. 
 
 N==COH 
 
 COH C-NH 
 II II >CH 
 N C-N^ 
 
 Xanthine. 
 
 These four bodies all occur naturally in various animal substances, and 
 adenine and xanthine also in tea. Their constitution is established by various 
 reactions. Thus nitrous acid converts adenine into hypoxanthine, and guanine 
 into xanthine : while xanthine itself is converted by the introduction of three 
 methyl groups into caffeine, which belongs to the next class of uric acid 
 derivatives. 
 
 This class consists of the methyl derivatives of xanthine : heteroxanthine 
 (7-methyl-xanthine), theobromine (3,7-dimethyl-xanthine), paraxanthine (1,7-di- 
 methyl-xanthine) theophylline (1,3-dimethyl-xanthine), and caffeine (1,3,7- 
 trimethyl-xan thine). 
 
 They all occur in nature. Caffeine, the most important of them, is found 
 to the extent of 0-5-2 per cent, in coffee, and 2-4 per cent, in tea, whence 
 it has also been called theine. The formulae of these bodies are established 
 by their syntheses. 3,7-Dimethyl-uric acid is converted by phosphorus oxy- 
 chloride into a mono-chloro-derivative, which is chloro-theobromine, and gives 
 theobromine on reduction. Theobromine when treated with phosphorus oxy- 
 
Syntheses of Purine Derivatives 
 
 325 
 
 chloride and pentachloride, has one of its methyls removed, and the two 
 hydroxyls replaced by chlorine. When the body so formed is treated with 
 ammonia and the product oxidized with chlorine, it gives guanidine. This 
 shows that the remaining methyl group cannot be at 8 (or it would give 
 methyl-guanidine), and so must be at 7. When this body is treated with 
 fuming hydrochloric acid the chlorine atoms are removed, with the production 
 of 7-methyl-xan thine, which is heteroxanthine. 
 
 On the other hand, if this 7-methyl-2,6-dichloro-purine is treated with 
 potash, only the chlorine at 6 (the most mobile) is saponified. The product 
 on methylation takes up a methyl group at 1, and the body so formed is 
 hydrolysed by hydrochloric acid to give paraxanthine. 
 
 NH— CO 
 
 CO C-N 
 
 I II >co 
 
 CH3N C-NH 
 
 3,7-Dimethyl-uric acid. 
 
 NH— CO 
 
 I I /CH3 
 
 CO C-N 
 
 I II >c-ci 
 
 CHg-N C-N*^ 
 
 Chloro-theobromine. 
 
 NH- 
 
 I 
 CO 
 
 CO 
 
 I /CH3 
 
 C-N 
 
 II >CH 
 
 =CC1 
 
 CCl 
 
 / 
 
 CH, 
 
 C-N 
 II II >CH 
 
 N C-N 
 
 7-Methyl-2,6-dichloro- 
 purine. 
 
 NH- 
 
 CO 
 
 NH- 
 
 CHg-N C-N 
 
 Theobromine. 
 
 CO 
 
 I /CH3 
 
 C-N 
 
 II >CH 
 
 -C-N 
 
 Heteroxanthine. 
 
 NH-CO 
 
 CCl C-N 
 
 /CH3 
 
 -X? 
 
 CH 
 
 N 
 
 7-Methyl-6-oxy-2- 
 chloro-purine. 
 
 CHg-N CO 
 
 I I /CH3 
 CCl C-N 
 
 II II >CH 
 N O-N^ 
 
 l,7-Dimethyl-6-oxy-2- 
 chloro-purine. 
 
 CH3N 
 
 CO 
 
 I I /CH3 
 
 CO C-N 
 
 I II >CH 
 
 NH— C-ir 
 
 Paraxanthine. 
 
 Caffeine can be obtained by the methylation of xanthine, and therefore is 
 trimethyl-xanthine. Its formula is proved by its synthesis from dimethyl- 
 alloxan, which can only have one constitution. This body combines with 
 methylamine in presence of sulphur dioxide to give trimethyl-uramil : — 
 
 CH3N CO CH3N— ^0 
 
 CO CO + H,N.CH3 -» (^0 (^<fH.CH3- 
 
 CH3-N — CO CH3.N— CO 
 
 This, like uramil itself, condenses with cyanic acid, forming trimethyl- 
 pseudo-uric acid : and this loses water, when treated with mineral acids, to give 
 trimethyl-uric acid : — 
 
 CH,-N — CO 
 
 CHoN — CO 
 
 ^CH, 
 
 <fO C-N<^O.kH, 
 CHoN — COH 
 
 I I /' 
 CO C-N 
 
 CH, 
 
 >C0 
 
 CH,N — C-NH 
 
 1175 
 
CHg 
 
 CH3 
 
 ,.N— 
 CO 
 
 ;-N— 
 
 -CO 
 
 
 Caffeine. 
 
 326 Uric Add Group 
 
 This is a substance already known as hydroxy -caffeine, and it yields caffeine on 
 reduction. This reduction obviously consists in the replacement of the hydroxy! 
 of the enol form by hydrogen, and so can only take place in one way : — 
 
 CHs-N — CO CHs-N — CO 
 
 I I /CH3 I I /CH3 
 
 CO C-N -» CO C-N 
 
 I II >C0 I II >C.OH 
 
 CHg-N — C-NH CHg-TSr G-W 
 
 Hydroxy-caffeine. 
 
 Any of the other syntheses of uric acid may be modified so as to produce 
 the other members of the group. Traube, for example, as has already been 
 mentioned, converted cyanacetyl urea into a pyrimidine, which, on treatment 
 with nitrous acid and subsequent reduction, gave a diamino-derivative. 
 
 NH2 HO-GO NH— CO NH— CO NH— CO 
 
 CO + CH2 -* CO CH2 -^ CO CH2 -» CO CNHj- 
 NH2 CN NH2 CN NH— C=NH NH— CNHg 
 
 The product condenses with formic acid to give a formyl derivative, whose 
 sodium salt loses water to form the sodium salt of xanthine : — 
 
 NH— CO NH— CO 
 
 CO CNHCHO = H„0 + CO C-NH 
 
 I II I II >CH 
 
 NH— CNH2 NH— C-N^ 
 
 If we start with symmetrical dimethyl urea instead of urea, and take it 
 through the same series of reactions, we arrive at a dimethyl-formyl compound, 
 which can then be further treated in two ways. If it is heated, it readily 
 loses water to give 1,3-dimethyl-xanthine, which is theophylline. On the 
 other hand, if it is treated with sodium ethylate and methyl iodide, the 
 acidic imide hydrogen next to the formyl group is replaced by methyl, and 
 the product condenses to caffeine : — 
 
 CHg-^ — CO CH3-N — CO CHg-N — CO 
 
 CO CH2 -^ CO CNHCHO -9. CO C-NH 
 
 II II I II >CH 
 CH3-NH CN CH3N — CNH2 CHj-N C-N^ 
 
 Cyanacetyl-di- 1 Theophylline, 
 
 methyl urea. * 
 
 CH3N — CO CHa-N — CO 
 
 CO C-N<^553 _^ CO C-N 
 
 ^CHO 
 
 CHa-N — C-NH2 CH3-N — C-N^ 
 
 Caffeine. 
 
 r<^H 
 
 So also guanidine gives cyanacetyl-guanidine, which can be transformed in an 
 analogous way into guanine : — 
 
Syntheses of Purine Derivatives 327 
 
 NH— CO NH— CO NH— CO 
 
 HN=C GHg -♦ HN=C C-NH„ -^ HN=C C-NH 
 11 I II I II >CH 
 
 NHg ON NH— CNHa NH— C-IT 
 
 Cyanacetyl- Guanine, 
 
 guanidine. 
 
 Further details as to these syntheses, and further evidence as to the 
 structure of the products, may be found in Fischer's original papers. But it 
 is to be observed that while Fischer's proof of the formula of uric acid itself 
 has never been disputed, in his first and very important paper on caffeine," 
 though many of his proofs are of great value and still hold good, yet he adopts 
 the view that all the xanthine derivatives, such as caffeine and guanine, contain 
 a different nucleus from uric acid ; and this false theory was not corrected 
 until his last series of papers in Ber. 30, 31, 32 (1897-8-9), of which the most 
 important is in Ber. 32. 435, which contains a summary of the whole group, 
 and a history of the steps by which the respective formulae were established. 
 
 ' Ber. 17. 1776 (1884). " Arm. 215. 253 (1882). 
 
 y2 
 
DIVISION IV 
 
 RING COMPOUNDS 
 
 Of the large number of known compounds containing nitrogen in a closed 
 ring, only a few of the most important can be discussed. 
 
 The simplest method of classification is to divide them into groups according 
 to the total number of atoms in the ring, subdividing each group according to 
 the number of nitrogen atoms which the ring contains. 
 
 The first group thus consists of the 3-rings ; and of this there are three 
 subdivisions : — 
 
 l>N : Y>N : 1>N. 
 
 cr w w 
 
CHAPTEE XVI 
 
 3-RINGS AND 4-EINGS 
 
 I. 3-RINGS. CaN 
 
 Of this subgroup very few representatives are known, and only one is of 
 
 CH 
 any importance. This is ethylene imine, I ^^NH. When bromethylamine, 
 
 CHaBr-CHa-NHa, is treated with potash it loses hydrobromic acid, forming 
 
 a base, C2H5N. This may have either of two formulae, CHa^CH-NHa, vinyl- 
 
 CHav 
 amine, or I /NH, ethylene imine, according as the hydrogen of the hydro- 
 CHa 
 
 bromic acid comes from the CHg or the NHa- The compound behaves in many 
 
 respects like an unsaturated body. It readily adds on hydrochloric acid to 
 
 form chlorethylamine, CHaCl-CHa-NHa, and sulphurous acid to give taurine, 
 
 CHa-NH, 
 
 I . Hence it was at first supposed to be vinylamine. 
 
 CHa-SOgH ^ 
 
 The resemblance of its mode of formation, and to a certain extent of its 
 
 CHa-NH 
 properties, to those of trimethylene imine, I I , led Marckwald ' to suspect 
 
 CHa-CHa 
 
 that it might have the ring formula. The ordinary reagents for the NHa group, 
 
 such as nitrous acid, cannot be applied to this substance on account of its 
 
 instability. Marckwald therefore used the Hinsberg reaction with benzene 
 
 sulphonic chloride, which has already been described (p. 19), and showed that 
 
 the body gave a sulphonamide insoluble in alkali. It follows that it must be 
 
 a secondary amine, i. e. ethylene imine. 
 
 This substance is a colourless liquid soluble in water, which attacks the skin 
 
 and smells strongly of ammonia. Its unsaturated character is clearly due to the 
 
 great strain in the ring. It is of interest as the first member of the group of 
 
 polymethylene imines, including trimethylene imine, pyrrolidine (tetrahydro- 
 
 pyrrol), and piperidine. 
 
 II. CNa GEOUP. DEEIVATIVES OF DIAZOMETHANE 
 
 The bodies contained in this group are the so-called fatty diazo-compounds. 
 The name is too well established to be given up, but it is unfortunate and mis- 
 leading, as it suggests a resemblance to the aromatic diazo-compounds, whereas 
 the fatty diazo-compounds, though allied to them, are, as ring compounds, very 
 different both in structure and in behaviour. 
 
 Tho first member of the group to be discovered was diazoacetic ester, 
 
 ' Howard, Marckwald, Ser. 32. 2036 (1899) 
 
932 Diazomethane Derivatives 
 
 obtained by Curtius in 1883 by the action of nitrous acid on glycocoU ester 
 
 hydrochloride. The mother substance, diazomethane, CH^ II , was not discovered 
 
 by v. Pechmann till 1894/ It is formed by the action of alkalies on various 
 nitroso-derivatives of methylamine, containing the group CHs-N-NO, the one 
 most commonly employed being nitrosomethyl-urethane. This reaction, as 
 has already been explained, must proceed in several stages. The first is no 
 doubt the formation of an unstable nitroso primary amine : this then changes 
 into the isomeric open-chain diazo-compound, which can actually be isolated, 
 and this again loses water to form the diazomethane : — 
 
 CO ^•^ -^ CH3-N<^" -» CH3-N=N0H -> CHJ + H^O. 
 \OEt ^ ^N 
 
 / Pechmann's method '^ was to dissolve nitroso-methyl-urethane in ether and 
 add a 25 per cent, solution of potash in methyl alcohol. The mixture is warmed 
 on a water bath, when a yellow liquid distUs over, which is a solution of the 
 gaseous diazomethane in ether. The yield is about 50 per cent./ 
 
 The true open-chain diazo-compound, which is an intermediate product in 
 this reaction, may be described here, though it is strictly analogous to 
 the aromatic diazo-compounds. It was obtained by Hantzsch' by treating 
 the urethane with excessively concentrated aqueous potash, or with ethereal 
 potassium ethylate. The urethane, without any ether, is dropped into the 
 very concentrated potash at 0°. In a few moments the salt, CH3-N=NOK,H20, 
 crystallizes out. It is extraordinarily sensitive to water, a few drops of which 
 change it explosively into diazomethane and potash. This is because, as the 
 salt of a very weak acid (even weaker than diazobenzene hydrate), it is very 
 readily hydrolysed ; and the excessively unstable hydrate instantly loses water 
 to form diazomethane. The same fact explains why this body is not formed 
 in V. Pechmann's method of preparation. The methyl alcohol, which must 
 be present in considerable excess, owing to its comparatively slight power of 
 dissolving potash, causes sufficient hydrolysis to bring about this change. 
 
 Diazomethane can also be made,* though only in very small quantity, by 
 the reduction of methyl nitramine, CHgNHNOj ; and in rather better yield 
 by the action of hydroxy lamine on dichloro-methylamine ° : — 
 
 CH3NCI2 + H2NOH = CH3N=N.0H + 2 HCl = CHJ h- H^O -I- 2 HCl. 
 
 The first stage of this reaction is analogous to the formation of diazobenzene 
 hydrate from hydroxylamine and nitrosobenzene : — 
 
 0-N=O ■+- H2NOH = 0-N=N-OH + H2O. 
 Diazomethane is an intensely yellow gas. It can be condensed ^ in a freezing 
 mixture of snow and calcium chloride to a yellow liquid boiling below 0°. It 
 
 > Ber. 27. 1888. = Ber. 28. 855 (1895). ' Ber. 35. 897 (1902). 
 
 « Thiele, Meyer, Ber. 29. 961 (1896). ' Bamberger, Renauld, Ber. 28. 1682 (3895). 
 
 « V. Pechmann, Ber. 28. 855 (1895). 
 
Diazotnethane : Properties 333 
 
 has a violently poisonous* action on most people (though some are little 
 affected by it), attacking the eyes and lungs. Chemically it is extremely active. 
 There is obviously a great strain in the ring, which is broken in nearly all 
 its reactions. The normal type of reaction consists in the elimination of the 
 nitrogen and the introduction in its place of two monad groups. 
 
 In particular, it reacts with all bodies containing hydroxyl, converting this 
 group into methoxyl : thus with water it gives methyl alcohol : — 
 
 CHJ +1 = l"^ + N,. 
 \N OH OH ' 
 
 In the same way acids are converted into their methyl esters, and phenols 
 into their methyl ethers (anisols). 
 
 The reactions of some isomeric oximes with diazomethane are peculiar and 
 difiScult to explain. We should expect both forms to be converted, and as far as 
 we can see with equal ease, into the two stereoisomeric 0-ethers : — 
 
 E-CH=NOH + CH2N2 = RCH^NOCHs + N^. 
 Forster and Dunn ' have examined the behaviour of the oximes of benzaldehyde, 
 of its three nitro-derivatives, and of its ^J-triazo-derivative, and find it to be quite 
 different from this. The syn-compounds are not acted on at all, except that of 
 jw-nitro-benzaldehyde, which gives a yield of the 0-ether of the anti-compound 
 much larger than that obtained from the anti-compound itself. The anti-oximes 
 are converted to a greater or less extent into their 0-ethers. Camphoroxime and 
 benzophenone oxime do not react. 
 
 This reaction is of some practical importance for methylation where other 
 methods fail. It has also been used ^ for determining the structure of tautomeric 
 bodies. It has the advantage that it proceeds in the absence of any third body 
 (except ether), generally quantitatively, and at the ordinary temperature ; and 
 the only products are the ester or ether and nitrogen. It has, however, been 
 less used for methylation since the discovery of the methyl sulphate method ; 
 and recent work has shown that the results obtained in the case of tautomeric 
 substances must be received with caution. 
 
 An analogous reaction occurs with aldehydes, which are converted into 
 the corresponding methyl-ketones,' e.g. : — 
 
 CCl3-C<J + CH2N2 = CCl3.C<g^3 + N^. 
 
 If an acetyl derivative of a phenol is treated with diazomethane, the acetyl 
 group is expelled and the methyl takes its place : for example : — 
 
 CaH3(OEt)2-OCO-CH3 -> C6H3(OEt)20CH3. 
 (This reaction does not occur in solvents quite free from water, nor if there 
 are other groups present in the ortho position to exert stereo-hindrance.) On 
 the other hand, if the acetyl group is attached to nitrogen (as in acetanilide 
 
 » J. C. S. 1909. 425. 
 
 ' Cf. H. Meyer, Mon. 26. 1311 (C. 06. i. 557) (pyridones) ; Peratoner, Azzarello, C. 06. i. 1489 
 (pyridoneB) ; Forster, Holmes, J. C. S. 1908. 242 (ieonitroBO-comphor) ; Aoree, Johnson, Brunei, 
 Shadinger, Nirdlinger, Ber. 41. 3199 (1908) (an elaborate investigation of certain nrazoles). 
 
 3 Schlotterbeck. Ber. 40. 479 (1907) j 42. 2559 (1909). 
 
334 Diazomethane Derivatives 
 
 derivatives), it is not expelled under these conditions. Hence the reaction can 
 be used to determine, in the case of acetyl derivatives of aminophenols, 
 whether the acetyl group is attached to oxygen or nitrogen.' 
 
 With primary and secondary amines, diazomethane forms the methyl 
 derivatives: e.g. with toluidine, methyl-toluidine. 
 
 When it is treated with an ethereal solution of iodine the colour disappears 
 
 at once, with the production of methylene iodide : — 
 
 ^N 
 CH2II + I2 = CH2I2 + N^. 
 \N 
 
 This reaction is useful for determining the strength of the solution of 
 diazomethane obtained in the preparation. It is titrated with an ethereal 
 solution of iodine of known strength as long as the colour disappears.^ 
 
 There is also a class of reactions in which the nitrogen is not eliminated. 
 Thus with unsaturated bodies it is capable in many cases of forming addition- 
 compounds. The simplest instance is that if it is left to stand for some time in 
 contact with acetylene, pyrazole is formed, to the extent of 50 per cent, of the 
 theory " : — 
 
 CH N^ CH-NH. 
 
 HI + I >N = I >N. 
 
 CH CH^ CH-CH^ 
 
 The analogous reaction takes place, but much less easily, with ethylene, 
 giving pyrazoline or dihydropyrazole.'' 
 
 In the same way, but even more readily, diazomethane will combine with 
 unsaturated acids or their esters, having a double or triple bond next to the 
 carboxyl: for example, with fumaric ester it forms the ester of pyrazoline 
 dicarboxyUc acid : — 
 
 COoKCH N%^ COjECH-NH. 
 
 COaKtJH CHj CO2ECH-CH 
 
 The acids so obtained have the remarkable property of losing their nitrogen 
 when heated alone, giving trimethylene dicarboxyUc acids : — 
 
 COoHCH — K COoHCH. 
 
 '1 >N = ' T >CH2 + N.,. 
 
 CO2HCH-CH/ COaHCH^ ^ 
 
 A similar intermediate formation of a nitrogenous ring may probably explain 
 the singular reaction of diazomethane with aldehydes, which have the aldehydic 
 hydrogen replaced by methyl, with the production of ketones" : — 
 
 0CH=O ^'9^~^\-M ^CO-CHg 
 
 + CH,N, -* H,C— r ^ + N, • 
 
 Of the homologues of diazomethane, Pechmann has prepared diazoethane, 
 
 1 Hertzig, Tiohatsohek, Ber. 39. 268, 1657 (1906). 
 
 2 The strength may also be determined by shaking with decinormal hydrochloric acid, and then, 
 after all the diazomethane has been converted into methyl chloride, titrating the excess of acid. 
 H. Meyer, Mon. 26. 1296 (C. 06. i. 555). 
 
 = V. Pechmann, Ber. 31. 2950 (1898). * Azzarello, C. 05. ii. 1236. 
 
 ^ Schlotterbeck, Ber. 40. 479, 3000 (1907) ; 42. 2665 (1909). 
 
Diazomethane : Properties 335 
 
 C2H4N2, which closely resembles it. He failed to obtain phenyl-diazomethane 
 by acting on nitroso-benzyl-urethane with alcoholic potash, but Hantzsch,' by 
 employing saturated aqueous potash, obtained first the true diazo-hydrate, which, 
 on treatment with water, gave phenyl-diazomethane : — 
 
 ^ \N0 _> 0.OH2-N=NOK -♦ ^-CH || . 
 \OEt ^N 
 
 It is a dark red oil, only slightly volatile; it does not boil without great 
 decomposition, evolving nitrogen and leaving stilbene, 0-CH=CH-0. 
 
 Diasoacetio ester, the first member of the group to be discovered, was 
 obtained by Curtius, in 1883, by the action of potassium nitrite on glycocoU 
 ester hydrochloride : — 
 
 H<f<NH, + o=NOH = ^^<N=NOH + h,0 = ^f N + 2H,0. 
 CO2R CO2R CO2K 
 
 It is to be noticed that free glycocoU does not give any diazo-compound with 
 nitrous acid, but goes directly to glyeoUie acid, with elimination of nitrogen. 
 The whole question of the conditions which determine the reaction of an NHj 
 group with nitrous acid is very obscure. The main facts ,as far as they are known , 
 are these. Apart from the simple formation of a nitrite, which occurs in some 
 cases, there are three possible reactions. First, the NHg group may be replaced 
 by hydroxyl, which is the most frequent case, occurring with all ordinary alkyl- 
 amines and with aU amides. Secondly, a true open-chain diazo-compound may 
 be formed, as in the aromatic amines. Thirdly, a ring compound may be pro- 
 duced, a fatty diazo-compound. This last reaction obviously requires that there 
 should be a hydrogen atom on the same carbon atom as the NHj ; if this is 
 absent, as in the aromatic amines, only the first or the second reaction can take 
 place. In fact there are two questions concerned, first, whether an alcohol or 
 an open-chain diazo-compound is the primary product, and secondly, whether 
 the latter, if produced, goes over at once into the ring compound. 
 
 As regards the first question, the normal reaction is the production of 
 a hydroxyl compound. The formation of a diazo-compound of any sort seems 
 to depend on the presence of acidic groups in the neighbourhood of the NH2, 
 though it is by no means all acidic groups which will produce this effect, and in 
 particular it is to be noticed that it does not occur with amides, which are 
 always converted into the acids. 
 
 But at any rate those amines which give diazo-compounds are always of 
 a more or less negative character. They comprise in the first place the 
 aromatic amines, in which the NH2 is attached to the negative benzene nucleus. 
 The only fatty amines which give this reaction are those which contain a car- 
 bonyl group, and it would seem that this must be in the a-position to the NHg , 
 i. e. that the compound must contain the grouping COCNHj. Thus among 
 the amino-acids, Curtius^ finds that the free acids never form diazo-compounds, 
 
 » JBer. 35. 897 (1902). 
 
 2 Curtius, Mftller, JSer. 37. 1261 (1904). Cf. Angeli, Ber. 26. 1715 (1893). 
 
336 Diazomethane Derivatives 
 
 and of the esters, only those which have the NHg in the a-position to the 
 
 carboxethyl. For example, a-^-diamino-propionic ester gives a-diazo-/3-oxy- 
 
 propionic ester. He has recently shown that the esters of the polypeptides, 
 
 which have the NHj in the same relative position to the carbonyl, can likewise 
 
 be converted into diazo-compounds : — 
 
 N 
 NH2CH200NH-CH2.C02Et -»• II >CH-CO-NHCH2-C02Et. 
 
 Again, the a-amino-ketones, such as amino-acetophenone, ^•CO-CH2-NH2, will 
 give diazo-compounds (Angeli), these also containing the grouping C0-C-NH2. 
 The same rule holds with certain uric acid derivatives such as amino-methyl- 
 
 uracil, 
 
 NH— CO 
 
 (j!0 (^•NH2, 
 
 NH— CCH3 
 though this body can only give an open-chain compound, as there is no 
 hydrogen on the carbon carrying the NHg. On the other hand, the presence of 
 hydroxyl on the carbonyl carbon prevents the formation of diazo-derivatives, as 
 is shown by the behaviour of the free amino-acids. Further, it is found that 
 while the presence of a hydrogen atom on the same carbon as the NH2 is 
 necessary to the production of a ring diazo-derivative, the stability of this body 
 is greatly increased if there is a second hydrogen atom, i. e. if we start with 
 a body -CH2NH2 and form a diazo-compound -CH:N2. It is possible that this 
 may be due to the body being then able to assume the tautomeric structure 
 
 JH 
 , although, as we shall see later, the supposed evidence for the existence 
 
 of compounds of this type has been shown to be incorrect. 
 
 Diazoacetic ester, N2-CH-C02Et, is a yellow liquid of peculiar smell, boiling 
 at 143°. It does not explode when struck, but does so violently when brought 
 into contact with sulphuric acid. It begins to decompose near its boiling-point 
 into nitrogen and fumaric ester, as phenyl-diazomethane does into nitrogen and 
 stilbene : — 
 
 COoECH'lf 
 
 ^^ = 2N2."^^^-^^. 
 
 C02E.C^^ ^^^^-^^ 
 
 \N 
 
 Its reactions closely resemble those of diazomethane ; thus when boiled with 
 water or dilute acids it liberates nitrogen quantitatively, giving the ester of an 
 oxy-acid : — 
 
 HC<II I HC<^„ 
 
 pN + OH = pOH + N^. 
 
 CO2E CO2E 
 
 This reaction has been shown by Bredig and Fraenkel ' to afford a valuable 
 
 ^ Bredig, Fraenkel, Z. f. Mektrochem. 11. 525 (1905); Ber. 39. 1756 (1906); Fraenkel, 
 Z. Ph. Ch. 60. 202 (1907); Bredig, Kipley, Sei: 40. 4015 (1907); Mumm, Z. Ph. Ch. 62. 589 
 (1908) ; Holmberg, Ber. 41. 1341 (1908). 
 
 -4 
 
Diazoacetic Ester: Properties 337 
 
 means of determining the concentration of hydrogen ion, especially at high 
 dilutions. The rate of change was measured by observing the volume of 
 nitrogen evolved, the solution being shaken in a flask connected with a 
 measuring vessel. The reaction is monomolecular, that is, it is proportional to 
 the amount of diazoacetic ester present. It is hastened by the presence of acids, 
 the monomolecular constant being proportional to the concentration of hydrogen 
 ion (with a limit of error of about 3 per cent.) down to N/5,000 concentration 
 of this ion. With weak acids the addition of salts of the acid greatly diminishes 
 the velocity, as we should expect. The addition of alcohol also greatlj' 
 diminishes it, no doubt by depressing the ionization. On the other hand, when 
 the alcohol is nearly absolute, this effect is reversed, and the addition of traces 
 of water lowers the rate of reaction. Thus with picric acid the addition of 
 0'18 per cent, of water to absolute alcohol lowers the velocity constant by 
 22 per cent. ; the reaction still continuing to be monomolecular. This anti- 
 catalytic action of water is unexplained ' ; but Goldschmidt and Sunde ' observed 
 a similar effect on the rate of esterification. 
 
 If hydrochloric or sulphuric acid is used to promote the decomposition of the 
 diazoacetic ester, the change gets rapidly slower and finally stops long before the 
 ester has all disappeared. If more acid is added more change takes place, but 
 this again soon comes to an end. This is due to the fact that under these 
 circumstances another reaction takes place, with the production of chloracetic 
 
 ester:— NgCHCOaEt + HCl = CLCHa-COaEt + Ng, 
 
 and this removes the catalyst. By determining the amount of chloracetic 
 ester produced, it is possible to calculate the velocities of the two reactions ; 
 and it has been shown that the formation of the haloid ester is promoted by 
 the presence of a neutral salt of the acid (e.g. sodium chloride), and that its 
 rate of production is proportional to the concentration of the diazo-ester, to 
 that of the hydrogen ion, and to a function of the concentration of the chlorine 
 ion which could not be exactly determined, but which approximates to the 
 cube root. 
 
 The same reaction occurs with nitric, hydrobromic, and hydriodic acids. 
 As regards the three haloid acids, it is found that the proportion of haloid 
 ester produced is greatest ceteris paribus with the iodide and least with the 
 chloride, this being the reverse order to that of their tendency to ionization. 
 
 When diazoacetic ester is treated with a halogen, it gives the di-halogen- 
 acetic acid : — • f^ 
 
 HC<11 HCI, 
 
 pN + I2 = I + N^. 
 CO2K CO2 
 
 With aldehydes ketonic esters are formed: e.g. with benzaldehyde, 
 benzoyl-acetic ester: — 
 
 pN + HCO-^ = HC-00-0. 
 CO2R io^R 
 
 ' See, however, Lapworth, J. C. S. 1908. 2187; Lapworth, Partington, ib. 1910. 19. 
 2 Ber. 39. 711 (1906). 
 
338 Diazomethane Derivatives 
 
 With bodies containing double or triple links it combines in the same way 
 as diazomethane. Thus it reacts with phenyl-acetylene at 100° to give phenyl 
 pyrazole-carboxylic ester,' 
 
 COgEt 
 
 and with unsaturated esters, such as acrylic, cinnamic, and fumaric, it forms 
 various pyrazoline carboxylic esters : e. g. with fumaric ester : — 
 
 N. CO2ECH— NH 
 
 CO,RCH ^ \ ^ CO,R.CH-C 
 
 ^^2^ CO2E 
 
 This last body is also produced by the spontaneous decomposition of diazo- 
 acetic ester, if it is allowed to stand for several years exposed to the light, 
 or if it is warmed for a few days ; or more rapidly (explosively at 80°) on 
 warming with copper powder.'' It is obvious that a part of the ester loses 
 nitrogen to form fumaric ester, and this combines with the rest. 
 
 If this product is heated, it loses nitrogen in the normal manner to give 
 trimethylene tricarboxylic ester, 
 
 COoECH. 
 
 ^ I >CHC02E. 
 COaECH-^ ' 
 
 An attempt has been made' to use this reaction of diazoacetic ester with 
 unsaturated bodies to test Thiele's theory of double and conjugate links, by 
 treating it with phenyl-butadiene, 0-CH=CH-CH=CH2. If addition occurred in 
 the 1,4 positions, as Thiele supposes, the product after the elimination of 
 nitrogen would be a pentamethylene derivative with the formula 
 
 0CH-CH=CH-CH2 
 
 H-CCOaEt 
 
 whereas it was found to be a trimethylene compound, 
 
 i^CH=CHCH-CH„ 
 
 HCCOjEt- 
 
 This shows that addition takes place in the 3,4 positions ; but as regards its 
 bearing on Thiele's general theory it is to be observed that the intermediate 
 nitrogen-ring compound, in the reaction which actually occurs, contains a 5-ring, 
 whereas 1,4 addition would involve the improbable formation of a 7-ring. 
 
 On treatment with ammonia, diazoacetic ester gives under ordinary condi- 
 tions its amide: but this will be dealt with more fully in discussing the 
 complicated question of its behaviour with alkalies. 
 
 Baohner, J3er. 35. 35 (1902). » silberrad, Eoy, /. C. S. 1906. 179. 
 
 = V. der Heide, Ber. 37. 2101 (1904). 
 
IHazoacetic Ester: Properties 339 
 
 On reduction it first yields hydrazi-acetic acid, I NH» which, when 
 
 COOH 
 treated with acids, gives hydrazine and glyoxylic acid, CHO-COOH ; further 
 reduction converts it into ammonia and glycollic acid. 
 
 One of the most remarkable reactions of diazoacetic acid is its behaviour with 
 benzene. This was first discovered by Curtius and Buehner in 1885, and the 
 product has since been investigated by Buehner and his pupils.' If diazoacetic 
 ester is heated with a very large excess of benzene, the whole of the nitrogen is 
 evolved, and a body is produced which is isomeric with phenyl-acetic ester, 
 ^•CHg-COjEt. On saponification the corresponding acid is obtained, which is 
 known as pseudo-phenyl-acetic acid. When the amide of this singular body is 
 boiled with soda it is converted into a heptamethylene derivative, cyclohepta- 
 triene carboxylic acid : — 
 
 CO2H 
 
 A 
 
 CHoCH 
 
 / ' \ ■ 
 CH CH 
 
 CH— CH 
 
 On careful treatment with bromine at 0°, pseudo-phenyl-acetic acid yields 
 a saturated tetrabromide, showing that it contains only two double bonds: 
 hence from its composition it must have two carbon rings. More energetic 
 treatment converts the tetrabromide into an unsaturated cyeloheptene acid, 
 which, in order to become saturated, requires to take up two more atoms of 
 bromine or one molecule of hydrobromic acid. It follows from this that pseudo- 
 phenyl-acetic acid must contain a 7-ring, with a bridge link easily broken. On 
 the other hand, it easily passes into a benzene derivative. Its amide is con- 
 verted by concentrated sulphuric acid into phenyl-acetamide, ^-CH^-CO-NHg, 
 and by oxidation with acid permanganate into benzoic acid. Hence it must 
 contain a 6-ring as well as a 7-ring. This fixes the position of the bridge link. 
 It shows that the COgEt-CH residue, which is added on to the benzene by the 
 diazoacetic ester, must attach itself in the ortho position : — 
 
 CH CH 
 
 H'^ ^N HC^JJB. H/ >a ^CH ' 
 
 CH ^ CH 
 
 As Buehner has shown, the decompositions of this body afford a very com- 
 plete proof of its structure. It can theoretically break up, as the formula 
 indicates, in three ways, to give three different forms of ring : a 7-ring (hepta- 
 methylene), a benzene ring, or a trimethylene ring. All these three transforma- 
 tions actually occur. Two of them we have seen already: soda converts the 
 body into a 7-ring compound, the isomeric cycloheptatriene carboxylic acid ; and 
 
 • Ber. 29. 106 ; 30. 632, 1949 ; 31. 399, 402, 2004, 2241, 2247 i 32. 705 ; 33. 684, 3453 ; 
 34. 982(1896-1901). 
 
340 Diazomethane Derivatives 
 
 sulphuric acid breaks the trimethylene ring to give a benzene derivative, also 
 isomeric, namely, ordinary phenyl-acetic acid. Finally, oxidation with acid 
 permanganate breaks the 6-ring, and gives trimethylene tricarboxylic acid : — 
 
 CH . 
 
 H'' ^CH CH H/ \CHC0,,H 
 
 This last reaction enables us to fix with fair certainty the position of the two 
 double bonds in pseudo-phenyl -acetic acid. They always form the point of 
 attack when an unsaturated body is oxidized ; and the fact that in this reaction 
 two carboxyls — that is to say, two carbon atoms — remain attached to the tri- 
 methylene ring, shows fairly conclusively that these two carbon atoms must be 
 attached only by single bonds in the original acid. That is to say, the formula 
 of pseudo-phenyl-acetic acid is what it is assumed to be in the above equation. 
 
 In a similar manner diazoacetic ester combines with toluene' and with 
 meta-xylene,^ giving compounds whose formulae have been shown to be 
 CH CH 
 
 ^^^SXVM-^^^' and ^«^g>C<?^l-^«3. 
 H-^ ^CH CH H ^CH CH 
 
 CH (>CH3 
 
 The formation of these substances is of interest from the point of view of the 
 benzene theory. The reaction is really precisely similar to that which occurs 
 between diazoacetic ester and the esters of unsaturated acids. Thus, as we have 
 seen, it reacts with fumaric ester to give an addition-product — pyrazoline tricar- 
 boxylic ester — and this on heating loses nitrogen to form trimethylene 
 tricarboxylic ester: — 
 CO,Ex(./N ^ CH.CO,K ^ ^q^^^^^-^^^-QH-QO,^ 
 
 H-^^N CH-CO,E H/ ^--^^H-CO^K 
 
 (Isoform). 
 
 ^ I^ ^ CO,Rx^./CH.CO,R 
 ' H/ \CHCO2E 
 
 In exactly the same way we can write the reaction with benzene, using 
 Kekule's formula : — 
 
 CO^JX^ + H^ N^ = CO,Rs^^/N=N-CH^CH 
 
 ^ N Hc. y:iH H/ ^--. I 1 
 
 CH 
 
 CH CH 
 ^CH 
 
 CH 
 
 ' H/ ^CH CH 
 
 ' Buchner, Feldmann, Ber. 36. 3509 (1903). 
 
 ' Buchner, Delbrflok, Ann. 358. i. (1908). This paper affords a remarkably ingenious and 
 complete example of the way in which the structure of such compounds can be established. 
 
Diazoacetic Ester : Polymerizations 341 
 
 That is, the benzene (and similarly toluene and xylene) behaves as though it 
 
 had three ordinary double links. The reaction occurs at a higher temperature 
 
 than with the unsaturated esters, and the (hypothetical) intermediate pyrazoline 
 
 derivative must be supposed to be unstable at this temperature and to break up 
 
 as fast as it is formed. 
 
 A still closer analogy is the condensation of diazoacetic ester with A'-tetrar 
 
 hydro-benzoic ester, which undoubtedly contains an ordinary double link, the 
 
 product being the carboxylic ester of the saturated tetrahydro-pseudo-phenylr 
 
 acetic ester : — 
 
 CH9 CHo 
 
 COjKx ^\^ COgEv ^\^ 
 
 H-^ ^N CH CH, ^ H/^OH CH, 
 
 CII2 CH2 
 
 Of course this reaction is no argument against Thiele's benzene theory, which in 
 fact does ascribe three double links to benzene, though it goes further and 
 explains the peculiarities of their behaviour. 
 
 It may be pointed out here that the occurrence of a cycle of reactions of this 
 kind — the formation of a nitrogenous ring followed by the elimination of the 
 nitrogen — throws doubt in some cases on the validity of the arguments as to the 
 structure of tautomeric bodies derived from their reactions with fatty diazo- 
 eompounds. For example, the formation of acetonitrile from diazomethane and 
 prussic acid, which was assumed to be evidence of the nitrile formula, can 
 equally well be explained on the imide formula thus: — 
 
 CH^il + C=NH -^ CH, I -» N2 + CH2=C=NH -» CH3CN. 
 
 This particular reaction has lost its importance since it has been found that 
 methyl isocyanide is actually produced along with acetonitrile ; but the example 
 is sufficient to show that a certain amount of caution is required in interpreting 
 the evidence in such cases. 
 
 Another important series of reactions of diazoacetic ester are its polymeriza- 
 tions with alkalies. This subject has had a singular history. The phenomenon 
 was first observed by Curtius,' and led him to the discovery of hydrazine and so 
 of hydrazoic acid. He regarded the products as triple polymers of the original 
 ester, and called them triazo-compounds. They were then reinvestigated by 
 Hantzsch and SUberrad,^ who showed that their molecular weight was not 
 three times but twice that of the diazo-ester ; they characterized many of the 
 more important products, and suggested formulae for them. The work has 
 recently been revised by Curtius, Darapsky, and Mtiller,^ who have greatly 
 
 '■ Ber. 18. 1283 ; J. pr. Ch. [2] 38. 408 (1888). Cf. Traube, Ber. 29. 669 (1896). 
 
 = Ber. 33. 58 (1900); Silberrad, /. 0. S. 1902. 598. 
 
 ' Curtius, Thompson, Ber. 39. 1383, 3398, 4140; Curtius, Darapsky, MuUer, Ber. 39. 3410, 
 3776 (1906); 40. 84, 815, 1176, 1194, 1470 (1907); 41. 3140, 8161 (1908). The last of these 
 papers contains a summary of the whole subject. Cf. E. Miiller, Ber. 41. 3116 ; Billow, Ber. 39. 
 2618, 4106. 
 
 1175 z 
 
342 Diazomethane Derivatives 
 
 extended our knowledge of these compounds, and have shown that the conclu- 
 sions of Hantzsch and Silberrad are in many points incorrect. The present state 
 of the subject is as follows : — 
 
 If diazoacetic ester is treated in the cold with dilute alkali or dilute ammonia, 
 it is converted in the usual way into the potassium salt or the amide of diazo- 
 
 acetic acid, II /CHCOOK or ll/CHCO-NHg. Free diazoacetic acid is too 
 
 unstable to exist, and decomposes as soon as it is liberated. If diazoacetamide 
 is warmed with alkali, it undergoes a remarkable isomeric change into a 
 triazolone : — 
 
 N=N j^jj N K 
 
 \/ f^2 _^ "l >NH. 
 
 CH-CO CH,-CO^ 
 
 On hydrolysis this body breaks up, as does diazoacetamide itself, into nitrogen, 
 ammonia, and glycollic acid. 
 
 If diazoacetic ester is treated in the cold with concentrated potash or liquid 
 ammonia, it gives the amide or the potassium salt of the so-called pseudo- 
 diazoacetic acid, a double polymer (incapable of existence in the free state) which 
 can be shown to be a dihydro-tetrazine of the formula : — 
 
 .N— NH 
 COOHC<f >CHCOOH. 
 
 The same reaction occurs with primary amines ; but secondary amines react 
 more slowly and form derivatives of another series of tetrazine derivatives, the 
 bis-diazo-compounds, which will be mentioned below. 
 
 The alkaline alcoholates act on diazoacetic ester to give yellow very unstable 
 ester-salts, which Hantzsch and Lehmann ' regarded as isodiazo-compounds, e. g. 
 
 t-c<X 
 
 •COaEt-C^j I . Curtius '^ and his collaborators have, however, proved from their 
 
 decomposition-products that they are tetrazine derivatives, probably belonging 
 to the pseudo-diazo series. The only true salt of diazoacetic ester known is that 
 of mercury,' which almost certainly has the metal attached to carbon : — 
 
 COaEt-C^g . 
 
 On the other hand, if diazoacetic ester is heated with concentrated potash or 
 strong ammonia, the salt or amide of an isomeric dihydro-tetrazine derivative is 
 formed, known as bis-diazoacetic acid (the original triazo-compound of Curtius) : 
 it may have either of two structures : — 
 
 N ^N. .N NH 
 
 COaH-Cr >C-C02H or CO^IL-Gf Sc-CO,H. 
 
 ^NH-NH'^ ^NH-IT 
 
 ' Ber. 34. 2506 (1901). = Ber. 41. 3140 (1908). 
 
 s Buohner, Ber. 28. 216 (1895). 
 
Diazoacetic Ester: Polymerizations 343 
 
 The structure of these 6-rrDg compounds is determined mainly by their 
 behaviour on hydrolysis. Under these circumstances the grouping =:N-N= 
 always breaks off as hydrazine, and -N=N- as nitrogen. Thus bis-diazoacetic 
 acid is hydrolysed to two molecules each of oxalic acid and hydrazine : — 
 
 + HO 
 
 + O^Ha H2=0 
 
 >N N.! ' + NHg.NHa 
 
 C02HC<^ JcCOaH = COjjH.CO.H + CO^TLCO^B., 
 
 NH-NH 
 H H 
 
 + NH,-NH, 
 
 i-OH 
 
 while pseudo-diazoacetamide ultimately gives two molecules of ammonia, two of 
 glyoxylic acid, one of nitrogen, and one of hydrazine : — 
 
 + OiH, H;OH 
 
 N-NH : + NH„-NH, 
 
 CO-NHn-Cf" ^ "])CH.CONH, -» 2 NH, + OHOCO2H'"'' *+'cH0-C02H. 
 
 >N=N ; + Na 
 
 The bis-diazo series may be obtained from the pseudo-diazo-compounds by further 
 treatment with potash. The change is exactly analogous to that of an azo- 
 compound into the isomeric hydrazone. 
 
 If the bis-diazo-compounds are further heated with excessively concentrated 
 potash, one of the nitrogen atoms is extruded from the ring, and a triazole 
 derivative is formed : — 
 
 COaH-CC >CC02H -* COaH-C CCO2H. 
 
 ^NH-NH \/ 
 
 N.NH2 
 
 On oxidation, both the pseudo-diazo- and the bis-diazo-series are converted 
 into derivatives of tetrazine dicarboxyUc acid : — 
 
 N=N 
 COaHCi, >CC02H. 
 
 "We should expect to find that these 6-ring derivatives could be made to lose 
 the carboxyl groups and yield the simple ring compounds. Hantzsch and 
 Silberrad considered that they had done this, but the bodies they obtained were 
 really triazoles. Thus bis-diazoacetic acid loses two carboxyls on heating, giving 
 l-N-amino-3,4-triazole.' Pseudo-diazoacetic acid cannot be isolated, so in this 
 case the reaction is impossible. It is, however, possible in an indirect way to 
 obtain the dihydro-tetrazine which is the real bis-diazomethane." If the tetrazine 
 dicarboxylic acid (the oxidation-product whose formula is given above) is heated, 
 it loses its carboxyls and gives the free tetrazine, a deep red solid : and this on 
 reduction is converted into the dihydro-tetrazine which is the mother substance 
 of the bis-diazo series : — 
 
 COaH-Ci >C-C02H -> HOC >CH -» HC< >CH. 
 
 N— IT ^N— N NH-NH 
 
 1 BiUow, Ber. 39. 2618, 4106 (1906). ^ Ber.iO. 84, 836 (1907). 
 
 z 2 
 
344 Diazomethane Derivatives 
 
 Its structure is shown by its giving formic acid and hydrazine (and no nitrogen) 
 on hydrolysis. It exhibits the characteristic tendency of these 6-rings to 
 ' crumple ', with the formation of a 5-ring, being converted by gentle warming 
 into an amino-triazole : — 
 
 HCO„H +'HC02H ^ HC< >CH -^ HC< >CH. 
 
 + NHg-NH, ^NH-NH N 
 
 We have thus seen how the 3-ring of diazoacetic ester can be converted into 
 a 5-ring (triazole), and into two different forms of a 6-ring (dihydro-tetrazine). 
 There is yet another change possible. If diazoacetic ester is treated with 
 hydrazine, aji Ng-ring is formed, the product being the hydrazide of azido-acetic 
 acid.^ This singular reaction is most easily explained by the intermediate 
 production of a ' buzylene ' derivative : — 
 
 NH2-NH2 + II >CH-C02Et = ]SrH2-NH-N=N-CH2-C02Et 
 
 -> II >N-CH2-C02Et + NH, -» II >N-CH2-CO-NH-NH2. 
 
 There is another method of preparing diazomethane derivatives due to 
 V. Pechmann, which may be mentioned on account of its simplicity. Sulphurous 
 acid condenses with hydrocyanic acid (one molecule of the latter to two of the 
 former) to give a disulphonic acid, which can be shown to be a primaiy amine : 
 for example, by its giving Hofmann's test with chloroform and potash. It must, 
 
 therefore, be aminomethane disulphonic acid, go^H/ ^H ^" WI1611 this is 
 treated with nitrous acid it forms diazomethane disulphonic acid : — 
 
 S03H/^<H + HO-N ~ So'h/^^I^ + ^^'^- 
 
 This body shows all the reactions of fatty diazo-compounds : for example, with 
 
 iodine it gives nitrogen and diiodo-methane disulphonic acid, so that there can 
 
 be no doubt about its constitution. It is to be noticed that we have here the 
 
 case of a fatty compound which is capable of being diazotized although it does 
 
 not contain the group CO-C-NH,. But it has a very similar structure. It 
 
 contains an NH2 in the a-position to a strongly negative group, namely, the 
 
 SO3H. It is also to be observed that in this case the free acid can be diazotized, 
 
 whereas the amino-carboxylic acids must first be converted into their esters. 
 
 The body is further interesting as providing a practical method of making 
 
 hydrazine. It forms a direct addition-compound with another molecule of 
 
 SO3K .N-H 
 sulphurous acid, of the formula „/C< I ^^ „ , and this when boiled with 
 
 acids splits up with the formation of hydrazine, 
 
 1 Ber. 41. Zii (1908). 
 
Azides: Preparation 345 
 
 III. N3 GEOUP. AZIDES 
 
 The azides, otherwise known as azimido- or triazo-compounds, or as diazo- 
 
 N 
 imides, are the esters of hydrazoic acid or azoimide, H-N<^ II . The number 
 
 which is theoretically possible is limited, from the fact that the ring is only 
 monovalent; but considerable attention has been directed to them in recent 
 years. 
 
 They were first prepared by Griess, by the action of ammonia on the diazo- 
 perbromides : — 
 
 0.N.Br3 + H3N = 0.N<1I + 3HBr. 
 N 
 
 They may also be formed by treating the diazo-compounds with hydrazoic 
 acid, a reaction resembling Sandmeyer's : — 
 
 0-N2-OH + HNs = ^-Ng + N2 + H2O, 
 and further, in an unexpected way, by the action of the diazo-compounds on 
 hydroxylamine,' possibly through the nitrosamine : — 
 
 0-NH-NO + H^N-OH -* 0-NH-N=N.OH -* 0N<'if + HgO. 
 
 Another method, which is most convenient for preparing them on a large 
 scale, is by the action of nitrous acid on hydrazine derivatives. An intermediate 
 product, a nitrosohydrazine, is formed first, which can in many cases be isolated 
 by working at a low temperature. This method is particularly used in preparing 
 the azides of acids, and is part of the cycle of changes by which Curtius 
 prepares hydrazoic acid from hydrazine : — 
 
 t'''''S)":N2^ = H^O -^ t>-00.1^<Z' = 2H,0 4- 0.GO.N<|- 
 
 Curtius's method of making hydrazoic acid consists in starting with an amide 
 insoluble in alcohol — he generally uses hippuramide — and treating it with 
 hydrazine hydrate, whereby it is converted into the acid hydrazide. This is 
 then transformed into the acid azide as above, and the azide dissolved in alcohol 
 and treated with ammonia, which reproduces the original amide, and, as it is 
 insoluble in alcohol, precipitates it ; while the hydrazoic acid remains in 
 solution : — 
 
 E-CONa + NH3 = ECONH2 + HN3. 
 
 In the same way phenyl azide, ^-Ng, is best made by acting on phenyl 
 hydrazine with nitrous acid.^ 
 
 The simplest organic azide, methyl azide, CHg-Ng, was discovered by Dim- 
 roth and W. Wislicenus,' who made it by dissolving sodium azide, formed by 
 the action of nitrous oxide on sodamide, in water, and treating it with methyl 
 sulphate : — 
 
 2 NaNa + (CH3)2S04 = NajSO^ + 2 CH3N3. 
 
 ' E. Fischer, Ann. 190. 96 (1877); Mai, Ber. 25.372 (1892). 
 ' Dimroth, Ber. 35. 1029 (1902). = Ber. 38. 1573 (1905). 
 
346 Azides 
 
 The methyl azide is evolved as a gas and condensed in a freezing mixture, 
 a yield of over 80 per cent, being obtained. It is a colourless liquid, boiling at 
 20-21°, and has an ethereal but unpleasant odour, resembling that of azoimide 
 itself. The authors say that in great dilution the vapour smells like the air of 
 a mouldy cellar. It explodes violently when brought into a flame, but its 
 temperature of explosion is high, over 500°, so that it vras found possible to 
 determine its vapour density, which is normal, and even to analyse it completely 
 by combustion, by mixing its vapour with a large excess of air or carbon 
 dioxide. 
 
 Several azido-derivatives of the simpler fatty compounds, such as azido- 
 acetic acid, Ng-CHj'COgH, and azido-acetone, Na-CHj-CO-CHg, have been prepared 
 by acting with sodium azide on the corresponding halogen derivatives. They are 
 colourless liqiiids or solids of low melting-point, which, though they explode on 
 heating, are fairly stable. 
 
 The aromatic azides, like diazobenzene-imide or phenyl azide, ^-Ng, are oUy 
 or crystalline substances of a strong peculiar smell, which explode at a somewhat 
 high temperature. On boiling with dilute sulphuric acid they evolve nitrogen 
 and form aminophenols. This is no doubt due to the primary production of 
 phenyl-hydroxylamine : — 
 
 so that the reaction is exactly similar to that of the fatty diazo-compounds with 
 dilute acids : — 
 
 The evolution of nitrogen on treatment with either acid or alkali is a general 
 reaction of the azides ; but in the case of many of the fatty derivatives the 
 ultimate product is a ketone-imine, which easily loses ammonia to form the 
 ketone or aldehyde : for example, camphoryl azide '■ or azido-acetone : — 
 CHgCO-CHa-Ng -^ CHa-COCH^NH -> CH3-C0-CH=0. 
 
 On the other hand, many azides when treated with alkali undergo a different 
 reaction, the whole azide group being eliminated as hydi-azoic acid. In the 
 aromatic compounds this reaction is promoted by the presence of negative 
 groups on the ring: thus when dinitrophenyl azide, (N02)2CgH3-N3 , is boiled 
 with alcoholic potash it is converted into potassium azide and dinitrophenol. 
 This may be compared to the weakening of the attachment of chlorine to 
 the nucleus by negative groups; indeed this is only one of many points of 
 resemblance between the Nj group and the halogens. Among the fatty com- 
 pounds the relations are more complicated.'' Camphoryl azoimide is hydrolysed 
 by potash to nitrogen and the amide ; but the simpler fatty compounds, such 
 as azidoacetic ester and azido-acetone and its oxime, all of which contain the 
 group Ng-OH^-CO, give at the same time a considerable amount of free azoimide. 
 
 ' Forster, Fierz, /. C.S. 190S. 826. 
 
 2 Forster, Fierz, J. C. S. 1905. 722, 826 ; 1908. 1174, 1859 ; Forster, ib. 1909. 184, 191. 
 
structure O— <^=NsC1.tt ; l^ut there is no evidence of this. 
 
 Azides: Properties 347 
 
 The behaviour of the two isomeric derivatives of propionic ester is peculiar. 
 The a-compound, CHj-CHaNg-COaEt, is very stable, and when the Ng ring is 
 attacked it gives nitrogen and no azoimide ; while the i3-compound, 
 
 CHaNs-CHa-CO^Et, 
 splits off azoimide with great ease. It is possible that these differences are 
 connected with the readiness with which the bodies are able to go over into 
 a tautomeric enolic form. A similar change is indicated by the behaviour of 
 para-azido-phenol/ Na-< >-0H. Whereas the corresponding ortho and meta 
 compounds are colourless substances, existing only in one form, the para body 
 occurs in two forms, one colourless, the other blue ; and these differences are 
 maintained in solution. The blue form may be a quinoid derivative of the 
 
 '^ 
 
 Various attempts have been made to reduce the azides, but in all cases these 
 resulted in breaking up the molecule, until recently Dimroth '^ was successful in 
 obtaining a direct reduction-product of phenyl azide, with very remarkable 
 properties. He used a solution of stannous chloride in ether containing hydro- 
 chloric acid, which has the advantage that the product is precipitated in the 
 form of its tin double salt. If phenyl azide is reduced in this way at - 18°, it 
 takes up two atoms of hydrogen, and yields a very unstable substance, to which 
 we may provisionally give the formula of phenyl-triazene, ^-N^N-NHj . Like 
 the other triazenes this forms metallic derivatives, the copper salt being fairly 
 stable, and the silver salt spontaneously explosive. Phenyl-triazene forms 
 colourless plates which melt at 50°, but rapidly decompose, even in the cold, in 
 contact with a trace of any solvent, giving aniline and nitrogen : — 
 
 ^■N=N-NH2 = 0-NH2 + Ng. 
 If these crystals are filtered off and placed on a porous tile under the micro- 
 scope, they begin in a very few minutes to break up into small fragments, 
 many of which are thrown up into the air. In about five minutes the crystal- 
 line plates are converted into a fine powder, which soon begins to decompose 
 into aniUne and nitrogen. This powder is an isomer, melting at 40°, which is 
 reconverted into the original plates (M. Pt. 50°) by recrystallization from ether. 
 It must be an isomer, because its formation is practically unaccompanied by 
 loss of weight ; hence its production cannot be due to the loss of ether of 
 crystaUization. It cannot be a mere case of physical dimorphism, because the 
 more stable form has the lower melting-point. The two must therefore be 
 chemical isomers, and there are various possible formulae : ^-N^N-NHg (syn- 
 
 ,NH 
 and anti-), 0-NH-N=NH, and also a ring structure, 0-N<^ I , phenyl cyclo- 
 
 triazaue. No evidence of differences in chemical behaviour between the two 
 forms has yet been obtained, but the general reactions of the body point clearly 
 to two structures, that of a triazene and that of a cyclo-triazane. It condenses 
 with phenyl isocyanate to form a urea, 0-N=N-NH-CO-NH-^, this being the 
 normal behaviour of the triazenes, as we have seen, so that we should infer that 
 
 1 rorster, Fierz, J. 0. 8. 1907. 855, 1350. " Ser. 40. 2376 (1907). 
 
348 Azides 
 
 it had one of the first two formulae given above; while on oxidation it is 
 reconverted quantitatively into phenyl azide, which points to its being the ring 
 
 compound <i>-'B'(\ . It is at least probable that these two formulae represent 
 
 the two isomeric forms. 
 
 This tautomerism of the derivatives of triazene and those of cyclo- 
 triazane : — 
 
 HN^N-NH, :;^ HN< 1 , 
 
 is of great interest. It serves to explain, for example, the fact^ that phenyl 
 semicarbazide cannot be made in the Hofmann reaction to go into phenyl- 
 triazene, but phenyl azide is obtained instead. The triazene which is no doubt 
 first produced changes into the ring compound, and this is then oxidized by 
 the hjrpobromite : — 
 
 /NH-NH.^ HN. N. 
 
 CO -> HgN-NH-NH-^ -» H^l^-T^-lS-^ -» l>N-0 -^ ll>N-0. 
 
 \NH2 hn n 
 
 Dimroth has shown ^ that phenyl azide can be used for several important 
 syntheses, some of which have already been mentioned. It condenses with 
 bodies containing an acidic methylene group, such as acetoacetic and malonic 
 esters, in presence of sodium ethylate, in somewhat the same way as diazomethane 
 does with doubly linked compounds, except that in this case water or alcohol is 
 eliminated. The N3 ring is broken unsymmetrically, and the residue joins on to 
 two contiguous carbons of the ester to give a derivative of a 1,2,3-triazole, 
 a ring containing three consecutive nitrogen atoms and two carbons. It is very 
 probable that a diazoamino-compound (triazene) is formed as an intermediate 
 product ' : for example, with malonic ester : — 
 
 ,N=N CH^COjEt ^N=N-CH-C0-OEt 
 
 / ^ + \ ^ ^ = 0-NH I 
 
 0N-^ COOEt ^ EtOCO 
 
 = 0-N< I + EtOH. 
 
 ^CO-CHCO-OEt 
 
 The product in this case is of especial interest from the fact that it occurs in 
 two distinct desmotropic forms.* The keto-form, whose structure is given above, 
 is a neutral substance, while the enol : — 
 
 ^ ^C=CCOOEt, 
 OH 
 
 is an acid strong enough to be titrated with phenol phthalein. The two can be 
 converted into one another, and their desmotropy is repeated throughout the whole 
 
 ^ Darapsky, Ber. 40. 3033 (1907). 
 
 2 Ber. 35. 1029, 4041 (1902) ; 36. 909 (1903) ; 38. 670 (1905). 
 
 ^ Dimi-oth, Frisoni, Marshall, Ber. 39. 3920 (1906). 
 
 * Dimroth, Ann. 335. 1 (1904) ; 338. 148 (1905). 
 
Phenyl Azide 349 
 
 series of their derivatives, including the free acids (of which the keto- is 
 monobasic and the enol dibasic) and their salts. 
 
 It is remarkable that this condensation with the azides is not confined to 
 compounds with an acidic methylene group, but also occurs, for example, with 
 acetic and propionic esters, in which the methylene is only attached on one side 
 to a negative group, and is not generally regarded as having acidic properties. 
 Thus propionic ester forms a compound : — 
 
 \C=:C-CH3. 
 OH 
 
 Another synthesis which Dimroth has carried out by means of the azides is 
 that of the fatty and the mixed diazoamino-compounds, as has already been 
 described, by condensation with Grignard's reagent :— 
 
 ^•N<| + BrMg.CH3 -^ 0-N<g^'cH3 "* '^•NH-N=N.CH, . 
 
 The same opening of the Ng ring, with the production of a diazoamino-com- 
 pound, occurs under the influence of hydrocyanic acid, as we have already seen 
 in dealing with the so-called diazoguanidine (carbamide-imide-azide). Phenyl 
 azide undergoes the same change when treated with hydrocyanic acid ' : — 
 
 N 
 0.N<ll + HCN = 0.NH-N=N-CN or 0-N=]Sr-NH-CN. 
 
 The second formula is the more -probable, but of course the two are desmo- 
 tropic. 
 
 It is to be noticed ° that one at least of the fatty diazo-compounds, diazo- 
 acetophenone, 0-CO-CH || , behaves in the same way, and gives with prussic acid 
 
 an open-chain diazo-cyanide, ^•C0-CH2-N=N'CN. 
 
 The diazonium azides, Ar-N-N\ II , are worth noticing from the very singular 
 
 N 
 group of five nitrogen atoms which they contain. They have been prepared by 
 Hantzsch * by treating the anti-diazo-hydrates with certain azide derivatives, such 
 
 as azido-carbonic ester, CO . They are crystalline substances, soluble in 
 
 ^CO^Et 
 water, insoluble in ether, which are violently explosive, and are decomposed by 
 soda to a diazotate and sodium azide. 
 
 1 Wolff, Lindenhayn, Ber. 37. 2374 (1904). « Wolff, Ann. 325. 129 (1902). 
 
 3 £e>-. 36.2056(1903). 
 
350 Polymethylene Imines 
 
 IV. 4-EINGS 
 
 Of the 4-rings which contain nitrogen very little is known. We have, 
 however, an example of a C3N ring in trimefhylene imine, which is made in the 
 following way. The amide of toluene-2>sulphonic acid condenses with tri- 
 methylene bromide to give the trimethylene-imide : — 
 
 BrCHg CH2 
 
 CHgCeH^SOaNHa + >CH2 = CHaCeH^-SOg-N^ >CH2 + 2HBr. 
 BrCHg CH2 
 
 The decomposition of the imide is not easy. The imine required would be at 
 
 once decomposed by acids, and it is necessary to treat the compound with sodiimi 
 
 in boiling amyl alcohol solution. The nascent hydrogen splits off the imine, on 
 
 which it has no further action, and reduces the sulphonic acid to toluene and 
 
 sulphurous acid. 
 
 Trimethylene imine ' is a liquid boiling at 63°, which fumes in the air and 
 
 smells strongly of ammonia. In fact it closely resembles ethylene imine, and 
 
 not least in the remarkable ease with which the ring is opened.^ Acids break it 
 
 at once. Even on evaporating the solution of the hydrochloride considerable 
 
 decomposition occurs. Hydrochloric acid merely adds itself on to form chloro- 
 
 propylamine, CHgCl-CHg-CHg-NHj. With sulphuric acid water is taken up, 
 
 giving oxypropylamine, CHaOH-CH^-CHa-NHa. The fact that the body reaUy 
 
 is trimethylene imine, and not the isomeric propylene-amine, is shown by its 
 
 CH2— CH2 
 giving with nitrous acid a nitroso-derivative, I I , which proves it to be 
 
 OXX2 — ^ *JN \J 
 a secondary base. 
 
 We may consider here the question of the stability of the polymethylene 
 imines, bodies of the general formula (CH2)„NII. They are commonly formed 
 either by loss of ammonium chloride from the hydrochloride of the corresponding 
 diamine : — 
 
 (CH2)„<J5g^.jj(.l -* (CH2)„>NH + NH.Cl ; 
 
 or by loss of hydrochloric acid from the chloramine : — 
 
 (CH2)Xnh^ -> (CH2)„>NH + HCl. 
 
 The readiness with which they are formed increases ° from the ethylene com- 
 pound up to the tetramethylene compound (tetrahydro-pyrrol or pyrrolidine). 
 The pentamethylene derivative is easily produced, but less so than the tetra-*, in 
 accordance with the usual rule that the 6-ring is the most stable. As regards 
 the higher members of the series, though these have been described up to 
 deeamethylene imine, it is probable that they have not the structures assigned 
 to them. Hexamethylene-diamine (and its chloramine) seems to form a small 
 quantity of the corresponding imine, derivatives of which undoubtedly exist, as 
 wUl be shown later, but the heptamethylene compound goes entirely to other 
 
 1 Howard, Marckwald, Ber. 32. 2036 (1899). ' Cf. Gabriel, Colman, Ber. 39. 2889 (1906). 
 
 = Cf. V. Braun, Steindorff, Ber. 38. 3083 (1905). « Cf. WiUstatter, Ber. 33. 365 (1900). 
 
Polymethylene Imines 351 
 
 products,' and the same is the case with the 8- and lO-carbon analogues. These 
 products are in some cases probably polymers of the type of piperazine, 
 
 HN'(,Qjj2,")>NH (in which the larger rings seem to be more stable),^ while in 
 
 other cases the ring closes up through a carbon atom which is not at the other 
 end of the chain. Thus Blaise and Houillon' have shown that the product 
 obtained on heating octomethylene-diamine, NH2-(CH2)8-NH2, is not ootomethy- 
 lene imine but mainly a-w-butyl-pyrrolidine, 
 
 CHjCHaCHa-OHa-CH CH^, 
 NH 
 and that the supposed decamethylene imine of Krafft* is really hexyl- 
 pyrroUdine. 
 
 The derivatives of hexamethylene imine have been obtained in a different 
 way, and identified beyond doubt, by Gabriel." Methyl-e-aminoamyl ketone 
 loses water to form a base, C7H13N, presumably in this way : — 
 (^Ha (J3H3 
 
 H2NCH2-CH2-CH2 ~^ ^CHa-CHa-CHa' 
 
 If this body is reduced (or if the ketone is reduced, when water is split off at 
 the same time) it gives the saturated 7-ring imine, methyl-hexamethylene 
 imine :- - 
 
 CH3 
 
 CH2~Cxl3— CH2 
 
 The only other formulae that this body might possess are those of ethyl- 
 piperidine, propyl -pyrrolidine, or of the corresponding derivatives of trimethylene 
 or ethylene imine. The last two are excluded, since hydrochloric acid, which 
 would break their rings, has no action on this body. The ethyl-piperidine and 
 the propyl-pyrrolidine in question are known, and are quite different substances. 
 There can thus be no doubt that the substance is really a 7-ring imine, of the 
 constitution which Gabriel assigns to it. 
 
 In connexion with the formation of these imine rings a remarkable case of 
 stereo-hindrance has been observed." The 1,4- and 1 ,5-dibromides, such as 
 o-xylylene bromide and 1,4- and 1,5-dibromopentane, condense with primary 
 aromatic amines to form phenyl-imines : — 
 
 CH4i::c?:lr + H2N.^ = €H2<^H2-CH2>j^.^ ^ 2 H3^ 
 
 Substituted amines behave in the same way, if their substituents are in the 
 
 ' V. Braun, C. MuUer, Ber. 38. 2203 (1905) ; 39. 4110 (1906). 
 ' Howard, Marckwald, Ber. 32. 2038 (1899). Cf. v. Braun, Ber. 39. 4347 (1906). 
 ' C. R. 142. 1541 (C. 06. ij. 527). 
 
 * Krafft, Phookan, Ber. 25. 2252 (1892) ; Krafft, Ber. 39. 2193 (1906). 
 5 Ber. 42. 1259 (1909). 
 
 Scholtz, Ber. 31. 414, 627, 1154, 1707 (1898) ; Scholtz, Friemehlt, Ber. 32. 848 (1899) 
 Scholtz, Wassermann, Ber. 40. 852 (1907). 
 
352 Trimethylene Imine 
 
 meta or para position. But if they have a substituent in the ortho position to 
 the NH2, the reaction takes a different course, and an open-chain compound is 
 formed : for example, with ortho-toluidine : — 
 
 p„ /CH2.CH2.NHC5H,CH3 
 ^^aNCHaCHaNH-CeHt-CHj ' 
 
 In this reaction a-naphthylamine behaves as an ortho-substituted compound, 
 while y3-naphthylamine does not. In the case of one di-ortho-substituted com- 
 pound which was investigated (mesidine) no reaction was obtained at all. 
 
 A compound containing a CgNg ring is Cxirtms's dimethi/l-asiethane,^ obtained 
 by the condensation of hydrazine with diacetyl in molecular proportions : — 
 
 CH3CO H2N CHg-C^N ^ 
 
 It is a crystalline powder, melting at 270°. 
 
 1 Curtius, Thun, J. pr. Ch. [2] 44. 175 (1891). 
 
CHAPTER XVII 
 5-RINGS 
 
 Op the great variety of 5-ring compounds which have been investigated, only- 
 one class will be dealt with : those containing a ring of four carbon atoms and 
 one nitrogen. This includes the pyrrol and the indigo derivatives. 
 
 PYEEOL GEOUP' 
 
 The pyrrol group of compounds has assumed of recent years a greatly 
 increased importance from many points of view. It has been found, for example, 
 that they, and especially the reduced pyrrol derivatives, are very widely dis- 
 tributed in nature. The pyrrol ring has been found by Pinner in nicotine, by 
 Liebermanu in hygrine, and by Willstatter in atropine and cocaine. Even more 
 important natural substances have been shown to belong to this group. Thus 
 the researches of Kiister and others have proved that both haemoglobin and 
 chlorophyll are to be regarded as derivatives of methyl-propyl pyrrol, and, as 
 we have seen, Emil Fischer has proved that a-pyrrolidine carboxylic acid is 
 a constituent of nearly all proteid substances. 
 
 If we add to this the great practical and theoretical importance of indigo 
 and its allies, it is evident that the pyrrol group is one which deserves careful 
 consideration. 
 
 The formula 
 
 CH— CH 
 
 II II 
 CH CH 
 
 \,..-^ 
 NH 
 
 was proposed for pyrrol by Baeyer in 1870, and has maintained its position 
 (with some reservation in respect to the distribution of the double bonds) 
 up to the present day. It is supported by the relation between pyrrol and 
 thiophene and furfurane, as well as by a whole series of syntheses in which 
 derivatives of these three classes result from y-dicarbonyl compounds, y-dike- 
 tones, y-ketonic acids, &c. This method of synthesis was discovered by Knorr ; 
 perhaps the simplest example is the preparation of dimethyl -pyrrol from 
 acetonyl-acetone, CH3-CO-CH2-CH2-CO-CH3. The ketone itself is obtained by 
 treating sodium acetoacetic ester with iodine, when two molecules condense 
 to give diaceto-succinic ester : — 
 
 CH3 <^H3 (|3H3 (^H3 
 
 GO CO CO CO 
 
 CH-Na -f I2 H- NaCH = CH CH -t- 2 Nal. 
 
 COaEt COaEt COgEt COaEt 
 
 ' See Ciamician, ' On the development of the chemistry of pyrrol iu the last twenty-five years ' 
 [_Ser. 37. 4200 (1904)], a very full and clear account on which the following is largely based. 
 
354 Pyrrol Group 
 
 This diacetosuceinic ester (a body remarkable for the enormous number of tauto- 
 meric forms in which it can occur) is itself a y-diketone, and therefore it will 
 condense with ammonia to a pyrrol derivative, the ester of dimethyl-pyrrol 
 dicarboxylic acid. Knorr and Kabe'have investigated this reaction in detail, 
 and find that an intermediate compound is formed. It is probable that the 
 ammonia first forms an addition-compound with the ester, and that this then 
 loses first one molecule of water, and then a second, to give the pyrrol : — 
 
 EtOaC-C^COHCHg _ EtOaC-C^COHCHg 
 
 EtOgC-CH-COCHs "^ ^ ~ Et02C-CHCOH{NH2)CH3 
 
 Et02C-Cl=COH-CH3 2 V-. 
 
 "* Et02CC=CCNH2)-CH3 
 
 )NH, 
 EtO,C-C=CCH, 
 
 The second of these bodies can actually be obtained by working in ethereal 
 solution at 0° ; at the ordinary temperature the third, the pyrrol, is produced. 
 The same intermediate stages probably occur in all syntheses of this kind." 
 
 In the ordinary Knorr synthesis of dimethyl-pyrrol the diacetosuceinic ester 
 is left to stand with cold soda solution, which saponifies it and at the same time 
 splits off the very unstable /3-carboxyls, leaving acetonyl-acetone. The interest 
 of this body lies in the fact that it is capable of giving all the three heterocyclic 
 5-rings. With a dehydrating agent it gives dimethyl-furfurane ; with phosphorus 
 pentasulphide, dimethyl-thiophene ; and finally, with ammonia, dimethyl-pyrrol. 
 If we disregard the intermediate compounds already referred to, these reactions 
 are best written with the ketone in the dienolic form, when the products appear 
 as its anhydride, its thio-anhydride, and its imine respectively : — 
 
 CH=C<^53 CH=C/^^3 CH=C/*-'^3 CH=C/^^3 
 
 I _ OH = I _ )0 = I _ )S ^ i _ >H- 
 
 CH-C<^jj^ CH-C\(,jj^ CH-C\(.jj^ CH-C\^g^ 
 
 Ace tony 1- Dimethyl- Dimethyl- Dimethyl- 
 
 acetone, furfurane. thiophene. pyrrol. 
 
 The methyl groups in dimethyl-pyrrol can then be oxidized to carboxyl and 
 so eliminated, giving pyrrol : a fairly conclusive proof of its formula. 
 
 Knorr's synthesis is capable of great extension. Practically any y-ketone, 
 ■y-ketonic acid, or y-diketonic acid can be used, while the ammonia can be 
 replaced by amines, hydrazines, or hydroxylamine. It is therefore of the utmost 
 value for the production of compounds of this type. 
 
 Pyrrol can also be got by distilling ammonium mucate or saccharate, or by 
 heating them to 200° with glycerine. The mucic acid first forms pyromucic 
 acid (furfurane a-carboxylic acid), which then reacts with the ammonia, and at 
 the same time loses another carboxyl : — 
 
 .COOH 
 CHOHCHOHCOOH CH=C<^ CH=CH. 
 
 CHOHCHOHCOOH CH=CH CH=CH^ '" 
 
 • Ber. 33. 3801 (1900). » Cf. Borsche, Fels, Ber. 39. 3877 (.1906). 
 
Pyrrol: Syntheses 355 
 
 Another general reaction, also discovered by Knorr,^ for the preparation of 
 pyrrol derivatives, is that of ketones with amino-ketones, including ketonic 
 acids. The method is to start with the easily obtained isonitroso-ketone, and 
 reduce this to amino-ketone in the presence of the other ketone. This reaction, 
 like the former, has been shown to go in two stages, an open-chain compound 
 being first formed ; for example, in the simplest case of acetone and isonitroso- 
 acetone : — 
 
 CH3 C0.CH3 <rH3(^o.cH3 CH-C.CH3 
 
 CH,(i0 ^ NH,(^H, - ^^3-C OH, -* CH3.C^^CH. 
 
 N NH 
 
 Pyrrol can also be formed (though only in traces) by the reduction of 
 auccinimide with zinc dust or sodium, which is an important proof of its 
 formula : — 
 
 1' "^NH + 211, = .1 >NH -1- 2H2O. 
 
 CH2-CSQ CH=CH'^ ' 
 
 and by passing acetylene and ammonia through a red-hot tube : — 
 
 r-a—cTf + NH3 = I >NH -f H,. 
 
 CH=CH 3 CH=Cff 
 
 The positions of the substituents on the ring are indicated in three ways, 
 which tends to confusion. The commonest and simplest way is 
 
 /3 a 
 
 /3' of 
 
 'but Beilstein uses 
 
 3 2 
 
 4 5 
 
 and there is a third method sometimes employed, 
 
 2 1 
 
 3 4 
 
 The formula of pyrrol is sufficiently proved by these syntheses, and is sup- 
 ported by the fact that the required number of isomers (e.g. three {n, a, j3) 
 mono-derivatives) are found to occur among the substitution-products. The 
 position of the substituents is often known already from the synthesis ; but an 
 important method of determining it is by breaking the ring to form an open- 
 chain compound. There are various ways in which the ring can be broken, and 
 they are all really reversals of different syntheses. The method which is of most 
 importance for our present purpose — that of orientation — is an indirect hydro- 
 lysis. Unlike the furfurane bodies, the pyrrols cannot be hydrolysed directly, 
 
 ■ Ser. 17. 1635 (1884) ; Ann. 236. 290 (1886). 
 
356 Pyrrol Group 
 
 except in the case of a few complicated derivatives, the reason apparently being- 
 that the reverse action, that of closing the ring, occurs much more easily. But 
 the hydrolysis is possible if a substance is present which can prevent this 
 reversal of the action by combining with the product of hydrolysis. Such a 
 substance is hydroxylamine. If a pyrrol is treated with an alkaline solution of 
 hydroxylamine, it is at first hydrolysed to a dialdehyde, and this at once forms 
 its dioxime : — 
 
 CH=CHv CH„-CHO CHa-CH^NOH 
 
 I >NH + 2H„0 = I ^ + NH, -> I ^ ; 
 
 CH=CH^ ' CH2-CHO ^ CH2-CH=N0H' 
 
 the body obtained from pyrrol itself being the dioxime of succinic aldehyde. 
 This reaction enables us to determine the position of any number of sub- 
 stituents. If the substituent is on the nitrogen it is, of course, removed, giving 
 the simple dioxime which, on hydrolysis with concentrated potash, is converted 
 into succinic acid. If the alkyl groups are in the /3- or ^'-positions, they give 
 by a similar reaction an alkyl succinic acid. If one a-hydrogen is substituted, 
 the product is the dioxime not of a dialdehyde, but of a ketone-aldehyde, and on 
 treatment with potash it is oxidized to a keto-acid. Thus a-/3'-dimethyl-pyrrol 
 gives ;8-aceto-isobutyric acid : — 
 
 CH3.C=CH^j.„ CH3-CH.CH=N0H CHa-C-COOH 
 
 I ^^sMi. _^ I /NOH — > I 
 
 HO-C^CH, CH^-^XCH, CHa-CO-CHg" 
 
 ±3 VJJ.±3 
 
 If both the a- and a.'- are occupied, a diketone-dioxime is formed, which on 
 hydrolysis gives only the diketone itself. 
 
 Another method of breaking the ring, which is also the reversal of a 
 
 synthesis, is by oxidation. As pyrrol can be got by reducing succinimide, so it 
 
 has been found' that pyrrol can be oxidized by chromic acid, not to succinimide, 
 
 CH-COs 
 but to the previously unknown maleinimide, II )>NH; and in the same 
 
 CH-CO 
 
 way a-/3'-dimethyl-pyrrol gives the imide of citraconic acid : — 
 
 zQ(^^i CH-CO. 
 
 , >NH -* ^„ ;i r.n>^H' 
 CHa-C^Cff CH3C CO 
 
 (^H= 
 
 the a-methyl being oxidized to carboxyl and eliminated. It is to be noticed 
 that these oxidations are in strict accordance with Thiele's rule ; we are dealing 
 with a conjugated system, and so the points of attack for the oxygen are the 
 1,4-positions, that is, the two a-carbon atoms. We may suppose that as usual 
 the oxidizing agent adds on two hydroxyls at these points, while the double 
 bond migrates to the 2,3-position ; and the hydrogen attached to the a-carbons 
 is also oxidized to hydroxyl. We thus get : — 
 
 HC CH H(j3=CH HC=CH HC=CH 
 
 HC^^CH -^ HOHC CHOH -* (H0)2C CCOH)^ ^00 CO. 
 
 NH ^NH ^NH ^NH 
 
 '■ Planolier, Cattadori, Gazz. 33. i. 402 (1903). 
 
Pyrrol: Properties . 357 
 
 Pyrrol itself is so called from Trvpp6% fiery red, because it gives this colour to 
 a pine shaving. It was discovered by Eunge in 1834, in coal tar, and in 1858 
 by Anderson in bone oil, from which it is now practically always obtained. 
 The bone oil is freed from its strongly basic constituents, mainly the pyridine 
 compounds (pyrrol and its homologues are very weak bases), and then contains 
 the nitriles of the fatty acids, the benzene hydrocarbons, and pyrrol and its 
 homologues. The nitriles are removed by saponification with potash, and the 
 liquid fractionated. The pyrrol is contained in the fraction boiling at 115-130°, 
 and is isolated by conversion into the solid potassium pyrrol. 
 
 Pyrrol is a colourless liquid boiling at 131°, which smells like chloroform. 
 It appears from its formula as a secondary amine, but its basic properties are 
 extraordinarily weak. They are to some extent concealed by the fact that pyrrol 
 and most of its homologues are very easily converted by strong acids into 
 complicated polymers ; but even where these bodies are not formed, and the 
 salts can be obtained, they are at once hydrolysed by water. In dilute acids 
 pyrrol only dissolves slowly ; and the tendency of the nitrogen to pass into the 
 pentad condition is so slight that it will not combine with alkyl iodide. 
 
 The weak basicity of the pyrrols is obviously related to their peculiar aromatic 
 character. The remarkable point about this is that whereas thiophene and its 
 homologues resemble the aromatic hydrocarbons, as has often been emphasized, 
 the striking analogy of pyrrol is not to the hydrocarbons but to the phenols. 
 Of this analogy many instances might be given. The imide hydrogen, like the 
 hydroxyl hydrogen of the phenols, can be replaced by potassium, forming potas- 
 sium pyrrol, C^H^NK. The hydrogen of the CH group, as in phenol, is 
 extraordinarily easily replaced : for example, by the halogens. Indeed, the 
 resemblance in the action of the halogens goes deeper than this. Tetrachloro- 
 pyrrol, C4CI4NH, can be further chlorinated to pentachloropyrrol, a derivative 
 of a pseudo-form, analogous to the pseudophenol compounds, such as hexa- 
 chlorophenol : — 
 
 C1-C=C-C1 C1-C!=C-C1 
 
 Cla-C C-Cl : Cla-C ^C-Ch 
 
 IT 0^-C^l 
 
 The analogy is also shown in the formation of nitroso- and nitro-compounds ; in 
 
 the power of direct coupling with diazonium salts to give azo- and dis-azo-dyes ; 
 
 and in their behaviour on alkylation. 
 
 It would seem as if we ought to be able to give some reason for so striking 
 
 a resemblance ; but we can do no more than guess. "We may take Baeyer's 
 
 CH=CH. 
 formula for pyrrol, I _ /NH, as established ; and the only question is as to 
 CH— CH 
 
 the exact interpretation which is to be given, in the light of Thiele's discoveries, 
 
 to the double bonds. The analogy to benzene is indicated by the undoubted 
 
 aromatic character, though of a peculiar kind ; and stiU more by the fact that 
 
 on reduction this entirely disappears, the dihydro-pyrrols being much more 
 
 unsaturated than pyrrol itself. From this point of view Bamberger, in 1891, 
 
 before the publication of Thiele's theory, suggested that the so-called hexacentric 
 
 1175 A a 
 
358 , Pyrrol Group 
 
 equilibrium was maintained in pyrrol by means of the two extra valencies of 
 the nitrogen : — 
 
 and that this would explain the feeble basicity of pyrrol, since its nitrogen atom 
 was pentavalent. But as regards the last argument, on which Bamberger 
 chiefly insists, Marckwald' has pointed out that the fact that theNH is attached 
 to two -C=C- groups is itself suiBcient to explain its feeble basicity, as is 
 illustrated by the cases of diphenylamine and dihydro-acridine, 
 
 which contain the same grouping, and exhibit the same suppression of basic 
 properties. The highly negative influence of this structure is further shown 
 in indene, 
 
 the carbon-analogue of indol, where the two -C=C- groups communicate to the 
 methylene a distinctly acidic character. 
 
 Thiele has now introduced a greater flexibility into our conception of link- 
 ages, and we may fairly adopt a modified form of Bamberger's suggestion. It 
 must be modified, because pyrrol is much less completely saturated than benzene, 
 as is shown by its behaviour on oxidation. But we may suppose that in pyrrol, 
 and similarly in thiophene and furfurane, the higher valency of the nitrogen (or 
 the sulphur or the oxygen) is partially exerted, and that thus there are two 
 bonds which can go some way towards saturating the residual valencies of the 
 carbons in the two a-positions. This would be expressed in Thiele's symbols 
 by the formula 
 
 --■HC^^^JbH -* H,0 CH,. 
 
 "NH NH 
 
 On reduction this passes as benzene does into an ordinary unsaturated system. 
 
 In this way we get an explanation of the peculiar, so to speak, semi-aromatic 
 character of pyrrol.^ As regards its striking resemblance to phenol, we can see 
 some sort of reason for this also. The peculiarities of phenol are certainly due 
 to a great extent to its power of assuming a second tautomeric form, from which 
 the pseudophenols are derived : and pyrrol possesses the same power, the three 
 forms corresponding exactly in the two cases : — 
 
 1 Ann. 279. 8 (1894) ; Ser. 28. 114, 1601 (1895). 
 ' Cf. Ciamieian, Gazz, 35. ii. 384 (C 05, ii. 1797). 
 
Pyi-rol: Character 359 
 
 H H H H2 H_H 
 
 HO-H O H a^i 
 
 H H H H2 H_H 
 
 NH N f 
 
 Pyrrol is characterized by a strong tendency to polymerization, shown, as it 
 so often is, by a readiness to turn into a resin, which is no doubt a high polymer. 
 All strong acids resinify it rapidly. If its ethereal solution is treated with 
 hydrochloric acid, the salt of tripyrrol, (C4HgN)3-HCl, is precipitated ; and free 
 tripyrrol may be obtained by neutralizing a solution of pyrrol in dilute hydro- 
 chloric acid with ammonia. It decomposes on heating into ammonia, pyrrol, 
 and indol, whence its formula probably is : — 
 
 CH-CH-OH— CH-CH— CH CH=CH-C — CH CH— CH 
 
 CH CH-CH CH-CH CH = CH=:CH-CI CH + OH CH + NH3. 
 
 NH NH ^11 ^Im ^H 
 
 A very remarkable characteristic of the pyrrol ring is the ease with which it 
 goes into the pyridine ring — one of the small class of cases of the expansion of 
 rings which are to be noticed from their bearing on the question of the stability 
 of ring structures. If potassium pyrrol is heated with chloroform it is converted 
 into ^-chloro-pyridine : — 
 
 CH=CH. CH=CC1-CH 
 
 I >N-K + CHCL =1 II + KCl + HCl. 
 
 CH=CH'^ ^ CH=CH-N 
 
 Many other halogen derivatives act in the same way, the methane carbon 
 always going between the a- and the /3-carbon atoms of the pyrrol, i. e. taking 
 up the meta position to the nitrogen, and becoming the /3-carbon atom of the 
 pyridine.' Thus with bromoform, potassium pyrrol gives /3-bromo-pyridine, 
 with methylene iodide, pyridine itself, and with benzal chloride, /3-phenyl- 
 pyridine, and so forth. The reaction takes place with many of the homologues 
 of pjrrrol and with the indols more readily than with pyrrol itself. From 
 experiments made with the indols and carbazols it is probable that the reaction 
 goes in two stages. A pseudo-pyrrol derivative is iirst formed, and then the ring 
 opens and admits the new carbon atom : — 
 
 HC-CH.CHC1, /^^ 
 
 The conversion of the pyrrol into the pyridine ring can also take place in other 
 ways. If a derivative of pyrrol with the side chain either on the nitrogen or on 
 
 ' Emerson Keynolds (/. C. S, 1909. 505, 508) has tried to obtain compounds with silicon in the 
 ring, by the action of silicon tetrachloride or silico-chloroform on potassium pyrrol, but without 
 success. 
 
 A a 2 
 
360 Pyrrol Group 
 
 the next (a-) carbon is passed through a tube at a low red heat, the carbon of the 
 side chain enters the ring. Thus N-benzyl-pyrrol is converted first into a-benzyl- 
 pyrrol, and then into /3-phenyl-pyridine : — 
 
 CH— CH CH-CH 
 
 X 
 
 CH CH -* ciijc-m,-(i> 
 
 CH C-(f> 
 
 11 1 
 
 ^^H^i!) "nH 
 
 CH CH 
 
 In the same way a-methyl-pyrrol gives pyridine.^ 
 
 Derivatives of Pyrrol 
 
 Potassium pyrrol, C4H4NK, is formed with evolution of hydrogen by dis- 
 solving potassium in pyrrol : and also by boiling pyrrol with solid potash, which 
 is a proof of the acidic character of the imine hydrogen. Sodium goes in much 
 less readily. Soda has no action on pyrrol at all ; and metallic sodium only 
 turns out the hydrogen at a very high temperature. This may be compared 
 with the much slower reaction of benzene, which acts on potassium at a high 
 temperature to give potassium benzene, C^Hj-K, but does not act on sodium 
 at all. 
 
 Potassium pyrrol is a solid crystalline substance, decomposed by water into 
 potash and pyrrol. It is the source of a large number of pyrrol derivatives, as 
 on treatment with alkyl and acyl halides, chlorocarbonic ester, &c., it gives the 
 corresponding N-substituted pyrrols. These products are remarkable for the 
 ease with which on heating the substituent passes from the nitrogen to the carbon, 
 as so often happens in the case of aniline. For example : — 
 
 CH=CH\ CH=CH. CH=C^^2^5 
 
 1 >N-K + C2H5I = 1 y^-C^m + KI -> V^ "^NH . 
 
 CH=CH^ ' ' CH=CH^ ' " CH=:CH 
 
 It appears that in these cases the a- and /3-positions are occupied indifferently, 
 a mixture of the two C-derivatives being generally obtained. 
 
 The further behaviour of the pyrrols on alkylation is extraordinarily com- 
 plicated, and has not yet been fully made out. By a succession of reactions 
 a series of pseudo-pyrrol, pyrroline, and pyrrolidine derivatives are produced, 
 the alkyl groups being added on to the ring one after another. 
 
 The alkyl pyrrols are of two kinds, the N- and the C-compounds. The first 
 class are obtained from potassium pyrrol, or by any of the syntheses, if an amine 
 is substituted for ammonia. They closely resemble pyrrol itself. 
 
 The C-alkyl derivatives may be got from the N- by heating, or they may be 
 formed by passing the mixed vapours of pyrrol and an alcohol over heated zinc 
 dust. This is essentially a dehydrating action : — 
 
 CH=CH. f^TT-p^CHg 
 
 ^„_p„>NH + HO.CH3 = ^H-C<j^jj ^ jj^O_ 
 CH-CH CH=:CH 
 
 ' Pictet, Ber. 38. 1961 (1905). 
 
Pyrrol Derivatives 361 
 
 On further heating they are converted into pyridine derivatives. They can also 
 be prepared by modifications of the various pyrrol syntheses ; thus we have seen 
 that acetonyl-acetone gives «,a'-dimethyl pyrrol. 
 
 They also resemble pyrrol, but both the imine and the methine hydrogens 
 are often even more reactive. They are easily resinified by acids. Fusion with 
 potash oxidizes the alkyls to carboxyls, giving rise to the pyrrol carboxylic 
 acids. 
 
 The a-mono-alkyl and a,/3-dialkyl compounds— i. e. those which have sub- 
 stituents only on one side of the molecule — when treated with hydrochloric acid 
 in ethereal solution give double (not, like pyrrol itself, triple) polymers ; and 
 these when warmed with dilute sulphuric acid break up into ammonia and 
 alkyl indols : — 
 
 CH3C — CH-CH-C-CHg CH.5C=CH-C — G-GR^ 
 
 ^„ II T I II " M II II 
 
 CH3C CH-CH CCH3 = CH3C=CH-C CCH3 + NH3. 
 
 NH NH ^m 
 
 The halogens act on pyrrol with extreme violence, even more so than on phenol 
 
 and aniline. The halogen must be very dilute, or the pyrrol resinifies ; and 
 
 even when it is dilute, it is not easy to stop short of a complete replacement of 
 
 all the four hydrogen atoms attached to carbon. 
 
 Tetraiodopyrrol, obtained from tetrachloropyrrol and potassium iodide, is 
 
 known as iodol, and is sometimes used as an antiseptic, having the same action 
 
 as iodoform, but being free from the smell which generally accompanies the latter. 
 
 An alkaline solution of chlorine or bromine both substitutes and oxidizes pyrrol, 
 
 CCl— CO 
 giving the imide of di-chloro-maleic acid, II ^NH. 
 
 CCl-CO^ 
 
 The production by further chlorination of pseudo-pyrrol derivatives analogous 
 to the pseudo-phenols, such as pentachloro-pyrrol, C4CI5N, has already been 
 mentioned. 
 
 The lower halogen derivatives have recently been prepared by the action 
 of sulphuryl chloride in ethereal solution.^ It is found that the halogen first 
 occupies the a-positions and then the /3-, and it is only when these places are 
 filled up that the imine hydrogen is attacked. The conversion of these products 
 by oxidation into substituted maleinimides gives a means of determining the 
 position which the halogen has taken up. 
 
 The nitrosopyrrols can be prepared by the action of amyl nitrite and sodium 
 ethylate. This gives the sodium salt, which is really derived from the isonitroso- 
 compound or oxime, e. g. 
 
 HC— C=NONa 
 
 HC^ CH 
 
 \^ 
 
 In the case of pyrrol itself only this sodium salt is stable, but in some of the 
 derivatives the free compound can also be prepared, 
 
 '■ Mazzara, Borgo, Gazz. 35. i. 477; ii. 100 (C. 05. ii. 488, 829). 
 
362 Pyrrol Group 
 
 The nitropyrrols can be made by direct nitration in the case of the more 
 stable compounds, such as the carboxylic acids. Otherwise it is necessary to 
 use amyl nitrate and sodium ethylate, which give the sodium salt of the 
 isonitro-compound : — 
 
 HC- 
 
 -C^NOONa 
 
 ^ 
 
 \ 
 
 HC 
 
 CH 
 
 \/ 
 
 This is a very unstable compound, which explodes like gunpowder when touched 
 with a hot wire. The dinitropyrrols are stable, though their salts also are 
 explosive. 
 
 The aminopyrrols' can be prepared from the nitroso- or the nitro-derivatives 
 by reduction with zinc and acetic acid. They resemble the aromatic amines in 
 being converted into diazo-compounds by nitrous acid. They are somewhat 
 unstable bodies, as are the aminophenols. 
 
 The azopyrrols are formed, like the azophenols or oxy-azo-compounds, by the 
 direct coupling of the pyrrols with diazonium salts. The diazo-group takes up 
 the a-position, and only goes to the fi- when the a- is occupied. The products 
 closely resemble the oxy-azo-bodies, and like them are dyes. They also show 
 the same tendency as the oxy-azo-bodies to go over into a tautomeric hydrazone 
 form : — 
 
 ^•N^NCoH^OH -» 0NH-N=CeH4=O 
 
 HC— CH HC=CH 
 
 ^ \ / \ 
 
 0-N=NC CH -> 0]SrH-N=C OH. 
 
 The /3-azo-derivatives, obtained from compounds in which the two cc-positions 
 are already occupied, undergo a remarkable change when heated with excess of 
 hydroxy lamine, being converted into pyrazoles ' : — 
 
 0N=N-C CH CH 
 
 II II J\ /CH 
 
 CHg-C CI-CHa — > CHg-C ^■^'^NOH' 
 
 NH 0-N N 
 
 This implies that the ring first opens, and then closes again through the 
 azo-group. 
 
 The pyrrol carboxylic acids can be made by introducing carboxyl into pyrrol 
 by any of the methods used in the case of phenol : for example, by heating 
 potassium pyrrol in a stream of carbon dioxide, or by acting on p3n'rol with 
 chloroform in the presence of alcoholic potash (Tiemann and Eeimer's reaction). 
 In either case an a-acid is obtained. The alkyl-pyrrols cannot be oxidized to 
 acids by potassium permanganate, as in the benzene series, but only by fusion 
 with potash ; in this they resemble the phenols. But if the ring contains 
 an acyl group, as CHg-CO, as well as an alkyl, then the alkyl can be oxidized to 
 
 ' Angelico, C. 05. ii. 900 ; Angeli, Mai-chetti, C. 08. i. 739. 
 * Castellana, Gazz. 36. ii. 48 (C. 06. ii. 1126). 
 
Pyrrol Derivatives 363 
 
 carboxyl by permanganate, though in this case the acyl only yields to potash 
 fusion. 
 
 The carboxylic acids of pyrrol resemble those of phenol in many respects. 
 They lose carbon dioxide even more easily than the latter, often merely by 
 boiling with water, or by heating alone above their melting-points. This, 
 however, is a common property of carboxyl attached to a nitrogenous ring. 
 
 It is remarkable that the /3-acids, both of pyrrol and of indol, are much 
 weaker than the corresponding o-acids, as is shown by the following values of 
 the dissociation constant K ^ : — 
 
 a-pyrrol-carboxylic acid 0-00403 
 
 0-00012 
 0-0002 (about) 
 0-000075 
 00177 
 0-00056 
 
 a,/3-dimethyl-pyrrol a'-carb. acid 
 
 a-indol-carboxylic acid 
 P" )» J) » • • 
 
 This is the more strange since the opposite is the case with the acids of 
 pyridine : — 
 
 a- 0-0003 : /3- 0-00137 : y- 0-00109, 
 
 The reduced pyrrol derivatives are of less importance. Their nomenclature 
 should be observed, as it is typical of that adopted in all these heterocyclic 
 compounds. By treating pyrrol with zinc and acetic acid, two hydrogen atoms 
 are introduced, and dihydropyrrol is formed, which on treatment with hydriodic 
 acid and phosphorus takes up two more hydrogen atoms to give tetrahydropyrrol. 
 Dihydropyrrol is known as pyrroline, tetrahydro- as pyrrolidine : one syllable 
 being added for every pair of hydrogen atoms : — 
 
 CH— CH 
 
 II II 
 CH CH 
 
 Pyrrol. 
 
 This system is always used : 
 
 CH— CH 
 
 II II 
 CH N 
 
 ^H 
 Pyrazole. 
 
 CH 
 
 CH2-CH2 
 CH2 CH2 
 
 Pyrrolidine. 
 
 NH 
 Pyrroline. 
 
 thus we have in the pyrazole series : 
 
 CH,-CH (j^H, 
 
 CH, N 
 
 -CH, 
 
 I ^ 
 NH 
 
 NH NH 
 
 Pyrazoline. PyrazoUdine. 
 
 The reduced pyrrols have entirely lost the peculiar aromatic character of 
 pyrrol, and behave as unsaturated or saturated fatty compounds. Thus pyrro- 
 line is a strong base, forming stable salts with acids, and giving a quaternary 
 iodide with methyl iodide. The further reduction of the pyrrolines to pyrroli- 
 dines is not easy. It requires the action of hydriodic acid and phosphorus, and 
 it is difficult to prevent this from going too far, and giving an alkylamine or 
 even a hydrocarbon. 
 
 Pyrrolidine is tetramethylene imine, and can be made, as has already been 
 
 Angeli, Gazz. 22. ii. 1 (1892). 
 
364 Pyrrol Group 
 
 mentioned, by heating the hydrochloride of tetramethylene diamine. In its 
 physical and chemical properties it closely resembles the next member of the 
 series, pentamethylene imine or piperidine. When rapidly heated or distilled 
 over zinc dust it is converted into pyrrol. It shows the characteristic behaviour 
 discovered by Hofmann for piperidine on what is known as exhaustive methy- 
 lation. This is the effect of continued treatment with methyl iodide, and is 
 a remarkable series of reactions, which has often proved of great value in 
 elucidating the structure of the nitrogenous rings of the alkaloids. In the case 
 of pyrrolidine the successive stages are as follows : — 
 
 ?H-^H <?H-CH ^H,-CH /CH 
 
 ch^-ch/ ch,-ch/ ' ch,-ch/ \i 
 
 -* CH2=CH-CH2CH2N(CH3)3l -» CH2=CH-CH=CH2 + N(CH3)J. 
 The main effect of these reactions is to split off the nitrogen from the ring first 
 on one side and then on the other, in each case with the production of an 
 ethylene group, the final product being trimethylamine and a doubly unsatu- 
 rated hydrocarbon. In the same way /3-methyl-pyrrolidine can be converted 
 into isoprene, CH2:::C(CH3)CH:=CH2. 
 
 A somewhat similar reaction is that of phosphorus pentachloride on benzoyl- 
 pyrrolidine (and also on benzoyl-piperidine).' According to the conditions the 
 ring can be split off on both sides from the nitrogen, or only on one, its place 
 being taken by chlorine : — 
 
 CH„-CH,. CH.,-CHov 
 
 1 ' '>N.CO-<i -> 1 ' '>N-CCl2-0 
 
 CHa-CH/ ^ CH^-CH/ '^ 
 
 ^ C1-(CH2)4-N=CC1-^ -* Cl-(CH2)4-NH-CO>. 
 ^ Cl-(CH2)i-Cl -t- NC-^. 
 
 INDOL GEOUP 
 
 If a benzene nucleus is imagined to condense -with a pyrrol nucleus so that 
 the resulting compound has two ortho-carbon atoms of the benzene identical with 
 the a- and ji- of the pyrrol, we get a body bearing the same relation to pyrrol 
 that naphthalene does to benzene, whose systematic name is benzopyrrol : — 
 
 CH CH 
 
 ^\ CH— CH --"\ 
 
 HC CH IT HC C CH 
 
 T I + CH CH : I T T ■ 
 
 HC CH \^ HC C CH 
 
 \/- NH \/^\/- 
 
 CH CH NH 
 
 This body is indol, the mother substance of indigo and its derivatives. This 
 group of compounds is for various reasons of the highest importance. The 
 physical and chemical properties of indigo early attracted the attention of 
 chemists. Its volatility and its bronzy lustre seemed particularly mysterious, 
 and the latter was at one time supposed to indicate something of a metallic 
 
 ' V. Braun, Beschke, Ber. 39. 4119 (1906). 
 
Indol Derivatives 
 
 365 
 
 nature. In a more scientific age, the desire to elucidate the constitution of the 
 substance was naturally stimulated by the hope of discovering a method of pre- 
 paring artificially the most valuable and most beautiful of natural dyes. The 
 history of the steps by which these two problems, so closely interwoven — the 
 scientific and the technical — were ultimately solved by the labours of Baeyer 
 and others, extending over the last forty years of the nineteenth century, forms 
 one of the most interesting chapters in the history of chemistry.* 
 
 The main credit of these great achievements belongs of course to Baeyer, who 
 seems to have been destined by nature from a very early age for the investiga- 
 tion of the subject. But the scientific examination of indigo began many years 
 before his time, and in the first half of the nineteenth century several valuable 
 benzene derivatives were obtained from it. Thus in 1826 Unverdorben prepared 
 aniline from it by dry distillation ; and in 1841 Fritsche obtained by oxidation 
 anthranilic acid (o-amino-benzoic acid), a body which has since become important 
 in connexion with the synthesis of indigo. 
 
 The first work which was really of value for the elucidation of the formula 
 was Laurent and Erdmann's discovery that indigo when oxidized with nitric 
 acid yields isatin, CsHgNOg. They were not able to throw much light on the 
 structure of this body, though it was clearly a benzene derivative ; and for some 
 years the subject was not further investigated. Then Baeyer took it up, and 
 showed that isatin, C6H4-C2HN02, could be converted by reduction first into 
 dioxindol, CeH^-CaHgNOa , and then into oxindol, CeH^-CgHsNO. Baeyer had 
 just been working on barbituric acid and its derivatives, and had proved bar- 
 bituric acid to be malonyl-urea : and he was struck by the remarkable resem- 
 blance of the series isatin, dioxindol, oxindol, to the series alloxan, dialurie acid, 
 barbituric acid, a resemblance fully borne out by their modern formulae :— 
 
 CO 
 CO 
 Isatin C\'^CC\ Alloxan 
 
 CeH^-CaHNOa Un/ 
 NH 
 
 CoHNO„C,HNO„ 
 
 NH 
 
 CO 
 
 I 
 NH 
 
 Dioxindol 
 CeH^-CaHsNO^ 
 
 Oxindol 
 CeH^-CiiHsNO 
 
 
 
 CHOH 
 CO 
 NH 
 
 CH, 
 
 '^GO 
 
 0. 
 
 NH 
 
 Dialurie acid 
 C,HNO,C,HoNOo 
 
 Barbituric acid 
 C2HNO2C2H3NO 
 
 CO 
 
 CHOH 
 CO CO 
 NH NH 
 ^CO 
 
 ^^ 
 CO CO 
 
 NH Nh" 
 
 CO 
 
 He regarded oxindol as a phenol, arid considered that on reduction it should 
 give a body in which the hydroxyl was replaced by hydrogen, which he caUed 
 
 1 This subject was summarized from the scientific side by Baeyer, and from the technical aide by 
 Brunck, in two lectures delirered at the opening of the Hofmannhaus in Berlin in 1900 {Ber. 33. 
 Sonderhefi, li, lxxi> The following account is largely based on these papers. 
 
366 
 
 Indol Group 
 
 indol. This lie actually prepared in 1866 by passing oxindol vapour over heated 
 zinc dust — a method introduced here for the first time, and adopted immediately 
 afterwards by Graebe and Liebermann for the conversion of alizarine into 
 anthracene, by which they were led to the synthesis of alizarine. 
 
 In 1869 Kekule suggested that isatic acid, which is formed from isatin by 
 the addition of a molecule of water, and readily passes back into isatin again, 
 was amino-benzoyl-formic acid, and isatin its anhydride : — 
 
 Isatic 
 acid 
 
 CeH^- 
 
 COCOOH 
 NH, 
 
 Isatin 
 
 /COCO 
 
 He suggested that it might be synthesized from o-nitro-phenyl-acetic acid, 
 but he was unable to prepare this acid, and so could not test the synthesis. 
 Baeyer saw that if Kekule's view was right, the reduction-products of isatin, 
 dioxindol and oxindol, might be similarly formed anhydrides of the reduction- 
 products of o-amino-benzoyl-formic acid, i. e. of o-amino-mandelic acid and of 
 o-amino-phenyl-acetic acid : — 
 
 Isatic 
 acid 
 
 o-aniino- 
 
 mandehc 
 
 acid 
 
 o-ammo- 
 
 phenyl-acetic 
 
 acid 
 
 I^COCOOH 
 
 U-NH^ 
 
 f>,-CHOHCOOH 
 
 U-NH, 
 
 fvCH^-COOH 
 
 Isatin 
 
 Dioxindol 
 
 Oxindol 
 
 03 
 
 CO 
 CO 
 NH 
 
 
 0: 
 
 CHOH 
 
 "co 
 
 NH 
 
 CH, 
 CO 
 NH 
 
 He proved that isatic acid gave on reduction o-amino-mandelic acid, which on 
 further reduction gave oxindol: and on June 6, 1878, he for the first time 
 synthesized isatin from o-nitro-phenyl-acetic acid by reduction to oxindol, followed 
 by oxidation : — 
 
 I^CH^-COOH 
 
 U-NO2 
 
 tt 
 
 CH2COOH 
 NH, 
 
 CHj 
 J^O 
 NH 
 
 Oj^-O 
 
 CO 
 CO. 
 NH 
 
 As he had ah-eady, in 1870, prepared indigo from isatin by treatment with 
 phosphorus pentachloride and reduction, this was the first true synthesis of 
 indigo. 
 
 The formulae of isatin, dioxindol, oxindol, and indol were thus determined 
 by the year 1878. The question of the formula of indigo still remained. It was 
 certainly closely related to the indol derivatives, as was shown by its formation 
 from isatin. Its synoptic formula was proved by analysis and vapour density to 
 be CijHioNgOa, while that of isatin is CgHjNOa. Hence indigo is formed by 
 the condensation of two molecules of isatin with the loss of two atoms of oxygen. 
 It was therefore to be expected that it would be obtained by the reduction of 
 isatin. But Baeyer had shown that isatin could be reduced by successive stages 
 
(^-CH=CHCOOH f^-CHBrCHBrCOOH 
 
 IJ-NO2 -* U-NO2 
 
 Indigo : Constitution 367 
 
 through dioxindol and oxindol to indol itself, which contains no oxygen at all, 
 without indigo being formed in the process. He had further shown that in 
 dioxindol and oxindol it was the ^-carbonyl which was reduced, and he there- 
 fore suggested that the presence of the carbonyl group in the /3-position may be 
 necessary to the formation of indigo. 
 
 We are practically sure, then, from the relation of indigo to the indol group, 
 
 C 
 that it contains the same nucleus as indol, M^. We know that it consists 
 
 N 
 of two such nuclei joined together, no doubt by means of the pyrrol ring. The 
 next question is how this junction is effected, Avhether through the carbon or 
 the nitrogen. If it is through the carbon, it is probably the a- and not the /3-, 
 since the presence of a /3-carbonyl seems to be required for the production of 
 indigo. 
 
 This question was settled by Baeyer in 1882, by the synthesis of indigo 
 from di-o-dinitro-diphenyl-diacetylene. o-nitro-cinnamic acid forms a dibromide 
 which when treated with alkali loses two molecules of hydrobromic acid, and 
 gives o-nitro-phenyl-propiolic acid : — 
 
 0: . 
 
 f^-C=CCOOH |^-C=CH 
 
 -* U-NO2 ~* U-NO2 • 
 
 This acid loses carbon dioxide to form o-nitro-phenyl-acetylene, and when the 
 copper derivative of this is oxidized with potassium ferricyanide, two molecules 
 
 condense to give di-o-dinitro-diphenyl-diacetylene, [ J^-hT) ~ NOJl J ' which on 
 reduction yields indigo. 
 
 This proves that the two phenyl nuclei in indigo are joined by an unbranched 
 chain of carbon atoms : or, in other words, that the two indol nuclei are joined 
 together through the two a-carbon atoms ; and this gives, as the most probable 
 formula for indigo, 
 
 NH NH 
 
 The only other possible formula would have the hydrogen attached to the 
 oxygen instead of the nitrogen. This, however, was shown by Baeyer to be 
 
 xs,_CO 
 impossible, for the following reason. The N-ethyl ether of isatin, kJ\^CO» 
 
 * N-Et 
 
 condenses like isatin itself to diethyl-indigo, which must therefore have two 
 ethyl-imino-groups, and this substance exactly resembles indigo in its properties. 
 It follows that indigo must contain two imine groups, and so have the formula 
 which Baeyer assigned to it. 
 
 This is in outline the history of the discovery of the constitution of indigo 
 and the indol compounds. Before considering the individual substances in 
 detail, it will be well to give a list of the more important of them. 
 
368 Indol Group 
 
 They are all derived from indol or benzopyrrol, and from dihydro-indol or 
 benzopyrroline, a body which has recently been isolated : most of them from 
 either, according as we regard them as enols or ketones. By replacing the 
 a- or ^-hydrogen, or both, in either of these bodies, by hydroxyl or ketonic 
 oxygen, we arrive at the following compounds : — 
 
 -CH, 
 
 _CH p 
 
 II 
 'OH a 
 NH 
 Indol, benzopyrrol. 
 
 a 
 
 a 
 
 I 
 
 'CH2 
 
 NH 
 
 Dihydro-indol, indoline, 
 
 benzopyrroline. 
 
 _COH 
 
 II 
 'CH 
 NH 
 /3-oxy-indol, indoxyl. 
 
 0^ 
 
 0: 
 
 'CH^ 
 
 NH 
 
 /3-indolinone, pseudo-form 
 of indoxyl. 
 
 0; 
 
 .CH 
 'COH 
 
 a 
 
 CO 
 
 NH 
 
 a-oxy-indol, oxindol, end form. 
 
 NH 
 Oxindol, keto-form. 
 
 _CHOH 
 
 
 
 a 
 
 _C0 
 
 I 
 'CO 
 
 
 
 NH 
 
 CO 
 NH 
 Dioxindol. 
 
 ,_co 
 
 \^C-OH 
 
 N 
 
 Isatin. 
 
 NH NH 
 
 Indigo. 
 
 Of these compounds, those which have the a-carbon oxidized (oxindol, 
 dioxindol, and isatin) are the lactames or lactimes of o-amino-benzoyl-formic 
 acid and its reduction-products. These names are intended to indicate the 
 analogy of the products to the lactones on the one hand and the amides or 
 imides on the other. They are formed by loss of water between the NHj group 
 (instead of the OH, as in the lactones) and the carboxyl : and this may occur in 
 two ways. Either the OH of the carboxyl goes out with one hydrogen of the 
 NHj, giving a kind of amide, or the :0 of the carboxyl with both hydrogens of 
 the NH2, giving an intramolecular imide : — 
 
 <, 
 
 ^OH 
 -NH, 
 
 -NH 
 
 p/OH 
 -NH, 
 
 — O 
 
 — N 
 
 ,/OH 
 
Indol: Derivatives 369 
 
 It is to be noticed that the two formulae are desmotropic, and are related in the 
 same way as ketone and enol : — 
 
 acocooH r\ — 90 r\-^^ 
 NH, -* U\^(lo U^C.O 
 
 Compare : — 
 
 NH N 
 Isatin. 
 
 Laetame. Lactime. 
 
 CO COH 
 
 -CH2 -CH 
 
 Ketone. Enol. 
 
 Indol or benzopyrrol is a crystalline solid, melting at 52° and boiling at 
 245°. It is obtained from its oxy-derivatives such as indigo and indoxyl by 
 treatment with zinc dust (Baeyer, 1866-8) or sodium amalgam. A large 
 number of syntheses have subsequently been discovered by Baeyer and others, 
 of which the following are the most important. 
 
 1. From o-amino-(o-chloro-styrol by treatment with sodium ethylate (removal 
 of hydrochloric acid) : — 
 
 _CH y^ CH 
 
 C>CH = HCl + U\/C 
 NH2 NH 
 
 CH 
 
 2. From o-nitro-cinnamic acid by reduction, the NO2 being first converted 
 into a hydroxylamine group : — 
 
 0_cH n— ^? C\ ^^ 
 
 s CH-COOH -* U\ dboOH -* U^^/CCOOH 
 NO2 Nli\0'fll NH 
 
 _CH 
 
 'CH + CO2. 
 NH 
 
 a 
 
 Indol is also formed to a small extent by passing the vapour of methyl-o- 
 toluidine together with hydrogen over reduced nickel at 300°.' This reaction 
 is reversible, and indol may in the same way be partially reduced to methyl-o- 
 toluidine. 
 
 The alkyl-indols may be prepared in various ways. For example, from 
 aniline and chloro-aeetone, the latter reacting in the enolic form : — 
 
 ^_H CICH _ y^ CH 
 
 U-NH2 + HO-C-CH., ~ U^^CCH, + HCl + H^O. 
 
 NH 
 
 Another way is from acetone and phenyl-hydrazine in presence of sulphuric 
 acid or zinc chloride.' The hydrazone which is first formed eliminates an atom 
 of nitrogen as ammonia from the middle of the chain in a peculiar manner, so 
 
 1 Carrasco, Padoa, C. 06. ii. 683 ; 07. i. 571. ^ Plancher, Caravaggi, C. 05. i. 1154. 
 
370 Indol Group 
 
 as to form the stable 5-ring ; but the reaction proceeds quite smoothly and 
 gives an excellent yield : — 
 
 A-H H,CH ^^„ , r^ CH 
 
 •VH H,CH ^^„ <v 
 
 I r "I = NH, + If 
 
 Indol itself resembles pyrrol in many respects. Its solution dyes a pine 
 shaving moistened with alcohol and hydrochloric acid a cherry red. It is only 
 feebly basic, and is easily resinified. 
 
 The homologues of indol resemble indol in properties, but are more stable 
 towards acids. They all, except the a,/3-dialkyl compounds, give the pine- 
 shaving reaction. On fusion with potash the side chains are oxidized to 
 carboxyl, giving the indol carboxylic acids. 
 
 Dihydro-indol is made by reducing N-methyl-indol with zinc and hydro- 
 chloric acid to N-methyl-indoline, and heating this with hydriodic acid and 
 phosphorus, whereby the methyl group is split off.' It is a liquid boiling at 
 220-221°. 
 
 p^ COH 
 
 Indoxyl, or /3-oxy-indol, 's/'--v^^ CH ' occurs in the indigo plant as the 
 
 NH 
 glucoside indican, and may be obtained by fusing indigo with potash in the 
 absence of air. It is an unstable oil which is rapidly oxidized by the air 
 to indigo. It is one of the intermediate products in the commercial manufac- 
 ture of indigo. 
 
 
 Indoxyl is tautomeric. It can react also as a ketone, v'^^^CH j ^^^ form 
 
 NH 
 known as pseudo-indoxyl, which, however, is not capable of separate existence. 
 In favour of the enol form there ai-e many reactions. Indoxylic ester behaves 
 as a phenol, dissolving in alkali, and being reprecipitated by carbon dioxide. 
 
 |/\ COH 
 
 It must therefore be k/^^-C-COOEt- ^^ gives with ethyl iodide an ethyl ether, 
 
 NH 
 and this saponifies to ethyl-indoxylic acid (the phenol ether), which loses carbon 
 dioxide to form ethyl indoxyl, 
 
 _COEt 
 
 0; 
 
 ^^CH 
 NH 
 
 On the other hand, indoxyl and indoxylic acid react with aldehydes and 
 ketones in the pseudo-form : thus with benzaldehyde they give : — 
 
 Q 
 
 _C0 
 
 ^-^C=CH-0 • 
 NH 
 
 ^ Plaucher, Eavenna, C. OS. ii. 335. 
 
Indoxyl 371 
 
 It is to be noticed that the oxidation of indoxyl to indigo is most simply 
 represented as a reaction of the keto-form : — 
 
 NH NH NH NH 
 
 ..a 
 
 -CH, 
 
 Oxindol, or a-indolinone, IJ\/'CO > melts at 120°. It is the lactame of 
 
 NH 
 o-amino-phenyl-acetic acid, into the barium salt of which it is converted by 
 heating with baryta to 150°. So great is the tendency to close the ring that the 
 free acid is incapable of existence : if the barium salt is treated with acid the 
 oxindol is formed at once. (Compare the behaviour of the y-oxy-acids.) It is 
 also produced by the reduction of dioxindol, and hence of isatin. 
 
 |/N CHOH 
 
 Dioxindol, W-^^^QQ > ^^ t^® lactame of o-amino-mandelic acid, which also 
 NH 
 can only exist in the form of its salts, and passes over as soon as it is set free 
 into the lactame. 
 
 CO CrS^ 
 
 -v,^^^QO > '^^ lactime, '^s^G-OH.t ^^ o-amino-benzoyl- 
 
 NH N 
 
 aCO-COOH 
 j^T^ , generally known as isatic acid, which can be isolated, 
 
 but if warmed in aqueous solution goes over into isatin. Isatin forms orange 
 red prisms, melting at 201°. It was first obtained by the oxidation of indigo, 
 and was first synthesized by Baeyer from o-nitro-phenyl-acetic acid. Baeyer 
 also synthesized it from o-nitro-phenyl-propiolic acid, which on treatment with 
 alkali is transformed into the so-called isatogenic acid, which then loses carbon 
 dioxide to give isatin : — 
 
 aC=CCOOH _^ f^ CO y^ CO 
 
 NO2 ~* IJ\ /CCOOH -> CO2 + U\^CO 
 
 o-nitro-phenyl- Isatogenic acid. Isatin. 
 
 propiolic acid. 
 
 It is also formed by the oxidation of oxindol and of dioxindol. If it is 
 reduced with zinc and hydrochloric acid it goes back into dioxindol.^ 
 
 Isatin was one of the earliest observed cases of tautomerism. It forms an 
 acetyl derivative, which must have the acetyl attached to nitrogen, since it 
 yields acetyl-isatic acid. It is therefore derived from the lactame form : — 
 
 oz:i:^(xt -^m 
 
 -COCOOH 
 . -NHCOCH, 
 NH N-CO-CHs 
 
 ' Hydro-isatin, the supposed intermediate reduction-product of isatin, has been shown not to 
 exist. Heller, Ber. 37. 938 (1904). 
 
372 Indol Group 
 
 On the other hand, silver isatin gives vrith ethyl iodide an ether which must 
 have the alkyl attached to oxygen, as it is readily saponified to isatin. This 
 would not in itself prove that the silver in the salt is on the oxygen ; but it is 
 fairly good evidence of this ' that the silver is removed as silver oxide when the 
 salt is treated with potash. The mercury compound, on the other hand, must 
 be an N-salt, for on treatment with potash the metal is not removed, though 
 the ring is broken and a potassium salt of the mercury compound is formed : — 
 
 0— QO ^-COCOOK 
 
 vCO -^ U-NH-hg . 
 N-hg 
 
 We may therefore assume that the silver salt is derived from the lactime form : — 
 
 rr^^ ^_co ^_co 
 
 vk^C-OH v^vC-OAg v'x^C-OEt 
 
 Free isatin probably has the lactame structure, both in the solid state and 
 in acid solution. Goldschmidt and Meissler - have shown that its reaction with 
 phenyl isocyanate is in accordance with this view, and Hartley and Dobbie ' 
 have been led to the same conclusion by a study of the absorption spectrum. 
 On the other hand, the alkaline solution and the alkahne salts appear to possess 
 the tautomeric enol structure. The change from one to the other can be shown 
 by means of a characteristic colour change.* Isatin dissolves in soda to form 
 a bluish-red solution of the N-salt. On standing at the ordinai-y temperature 
 the colour soon changes to pale yellow, owing to the formation of the 0-salt. 
 The salt of isatic acid is not formed unless the solution is boiled. If the yellow 
 solution of the 0-salt is acidified, the colour only gradually goes back to reddish, 
 and then the isatin crystallizes out. 
 
 When treated with phosphorus pentachloride, isatin reacts, as we should 
 expect, according to the enol formula (compare the case of uric acid), and gives 
 
 a 
 
 ,— CO 
 
 I 
 
 isatin chloride, V\^C-C1 > which on reduction with zinc and acetic acid, or with 
 
 N 
 
 hydriodic acid, gives indigo. It is probable that the first effect of the reduction 
 is to add on two atoms of hydrogen, with the formation of chloro-indoxyl, which 
 then simply loses hydrochloric acid : — 
 
 /^_C0 t^ CO ^ CO OC XV, 
 
 Ox/(^.ci - CC(i<^, -- 0^^h=h^S}- 
 
 N NH '^^ NH NH 
 
 Indigo, sometimes known as indigotin, occurs in the form of indican 
 (probably a glucoside of indoxyl) in the indigo plant (Indigofera Tinctoria) of 
 China and India, and in smaller quantities in the European woad {Isatis 
 Tinctoria). To obtain it fi-om the indigo plant, this is crushed in water and 
 allowed to stand exposed to the air. An enzyme contained in the cells of the 
 
 ' Peters, Ber. 40. 235 (1907). ' Ser. 23. 253 (1890). 
 
 8 .7. C. S. 1899. 640. * C.^ Heller, Ber. 37. 933 Anm. (1901). 
 
Indigo 373 
 
 plant breaks up the indican into glucose and indoxyl, and the latter is oxidized 
 by the air to indigo.^ In dyeing with indigo two processes are employed. 
 Either the dye is dissolved in sulphuric acid, and so converted into a disulphonic 
 acid which is soluble in water and can be used for dyeing directly (Saxon 
 process), or it is reduced to the leuco-base, indigo-white, by treatment with 
 grape sugar in alkaline solution. The cloth is then dipped in this and the 
 leuco-base oxidized on the fibre by the air. This indigo-white is probably 
 a di-indoxyl: — • 
 
 ^_-C .OH HO. C__^ 
 
 NH NH 
 
 Indigo is a dark blue powder with a coppery-red reflection when rubbed. It 
 is without smell or taste. It is insoluble in water, alkalies, acids, alcohol, and 
 ether ; but dissolves in aniline, molten paraffin, and some other organic sub- 
 stances, crystallizing out again on cooling. It has been found ^ that if indigo- 
 white is oxidized in a solution containing certain colloidal substances, the 
 indigo, instead of being precipitated, remains in solution in a soluble colloidal 
 form. The best substances to employ as colloids are the higher hydrolytic 
 products of the proteids, such as lysalbic acid. The solution is quite stable, and 
 on evaporation leaves the indigo behind as a blue friable amorphous mass, 
 easily soluble in water. It would seem, however, that this soluble form is 
 a compound of indigo with lysalbic acid. Prom its aqueous solution the indigo 
 is precipitated in the ordinary insoluble form by organic acids. 
 
 Indigo is further remarkable for its volatility. It forms a dark red vapour 
 on heating, and under diminished pressure can be sublimed without decom- 
 position. 
 
 The molecular weight of indigo in solution has been examined by Vaubel,* 
 with rather unsatisfactory results. His work has been repeated by Beckmann 
 and Gabel,* who find that indigo gives the normal molecular weight by the 
 boiling-point method in aU the solvents investigated (quinoline, aniline, phenol, 
 and jp-toluidine). The same value was obtained by the freezing-point method 
 in aniline and phenol. But the lowering of the freezing-point of ^-toluidine 
 indicates a double molecular weight, so that it appears that in this solvent 
 the molecules of indigo are double at the freezing-point (45°) but single at the 
 boiKng-point (198°). The crystals of j)-toluidine which separate out are quite 
 colourless, so that there can be no question of the results being vitiated by the 
 formation of a solid solution. 
 
 The constitution of indigo has already been discussed. The main points 
 are: — 
 
 1. Its relation to indoxyl and isatin shows it to contain two benzopyrrol 
 
 nuclei. 
 
 2. Its formation from di-(o-nitro-phenyl-)-diacetylene shows that these two 
 nuclei are joined through the two a-carbon atoms. 
 
 » Cf. Bergtheil, /. C. S. 1904. 870. '^ Mohlau, Ziramermann, C. 03. i. 640. 
 
 3 C. 01. ii. 779 ; 02. i. 936. * Ber. 39. 2611.(1906). 
 
 1175 B b 
 
374 Indol Group 
 
 3. The production of N-diethyl-indigo with properties simDar to those of 
 indigo shows that it contains two NH groups. 
 
 The syntheses of indigo are very numerous, and some of them have been 
 mentioned already. The more important are the following: — 
 
 The first true synthesis was completed by Baeyer in 1878, when he synthe- 
 sized isatin from o-nitro-phenyl-acetic acid, as he had shown in 1870 that isatin 
 can be converted into indigo. 
 
 There are also a whole series of syntheses which Baeyer has worked out, 
 starting with o-nitro-cinnamic acid. 
 
 The acid may be oxidized with permanganate to o-nitro-benzaldehyde : this 
 condenses with acetone to the methyl ketone of o-nitro-phenyl-lactic acid, which, 
 on treatment with alkali, gives acetic acid, water, and indigo : — 
 
 f^-CH=CHCOOH f^CHO + CH3COCH3 _^ f^ CHOH 
 
 U-NO, -* U-NO2 "* U\ CH,.C0.CH3- 
 
 Or it may be converted by bromine into the dibromide, and this with alcoholic 
 potash into the propiolic acid : which, on reduction in alkaline solution, gives 
 first isatogenic acid, and then loses carbon dioxide to form indigo : — 
 
 r^-CHBrCHBrCOOH fV-C=CCOOH 
 
 -^ iJ-NOa -^ U-NO2 
 
 
 
 ^"cOOH <'f;fr' 00. . indigo. 
 
 N— O 
 
 Lastly, the propiolic acid may be converted by loss of carbon dioxide into 
 o-nitro-phenyl-acetylene, and this condensed by the action of potassium ferri- 
 cyanide to the diacetylene derivative, which, on treatment with alkali and reduc- 
 tion, yields indigo. 
 
 Since indigo can be obtained from indoxyl and from isatin, every new 
 synthesis of either of these bodies is a synthesis of indigo. Indoxyl is con- 
 verted into indigo by oxidation, while isatin (acting in the pseudo-form) is 
 converted by phosphorus trichloride into isatin chloride, which is reduced by 
 zinc dust through a-chloro-indoxyl to indigo. 
 
 A comparatively simple synthesis is that of Camps. ^ o-Nitro-acetophenone is 
 converted by a variety of reducing agents (stannous chloride and hydrochloric 
 acid, aluminium amalgam, zinc dust, &c.) into the corresponding hydrazo- 
 compound, di-o-diaceto-hydrazobenzene : — 
 
 f^CO-CHs CHg-CO-r^ 
 U— NH NH— U • 
 
 When this is heated alone it is converted largely into indigo, together with 
 various reduced decomposition-products. 
 
 Another remarkable method'' starts with dibromo-maleic acid. If this is 
 heated with aniline it gives dianilido-maleic anhydride. Sodium methylate 
 
 ■ C. 02. ji. 939. ^ Salmony, Simonis, Ber. 38. 2680 (1905). 
 
Syntheses of Indigo 375 
 
 converts this into its sodium salt, which, when fused with potash and sodamide, 
 gives indigo : — 
 
 COOH COOH ^ NaOCO COONa^"*'^ CO CO ^' 
 
 Commercial Synthesis of Indigo 
 
 The most important synthesis of indigo still remains to be described. It is 
 that by which indigo is now prepared on the large scale, and is known as 
 Heumann's synthesis. In its original form aniline was condensed with chloracetic 
 acid to give phenyl-glycine. This, when treated with alkali or sulphuric acid, 
 loses water and goes into indoxyl, which on oxidation yields indigo : — 
 
 IJ-NH2 + Cl-CHa-COOH ~* (j- 
 
 NH.CH„COOH 
 
 
 
 _co 
 
 '^^CHa 
 NH 
 
 CO 
 
 (X:X 
 
 NH 
 
 An important improvement was the substitution of anthranilic acid (o-amino- 
 benzoic acid) for aniline. This is converted by the same series of reactions 
 through pheuyl-glycine-o-carboxylic acid into indoxyl-carboxylic acid. This 
 body loses carbon dioxide to form indoxyl, which is oxidized as before to 
 indigo : — 
 
 a COOH f^COOH 
 
 NH2 + Cl-CHjCOOH -^ U-NHCH2COOH 
 Anthranilic Phenyl-glycine- 
 
 acid. o-carboxylic acid. 
 
 a 
 
 '\ _» Pi T -» Indigo. 
 'CH-COOH U\^CH2 * 
 
 NH NH 
 
 Indoxyl- Indoxyl. 
 
 a-carboxylic acid. 
 
 The following account of synthetic indigo is mainly taken from Brunck's 
 paper already referred to, and so represents the state of the industry in 1900. 
 
 The commercial preparation of synthetic indigo was, from the beginning of 
 the aniline dye industry, an object of the researches of chemists. But the earlier 
 syntheses of indigo were useless from a commercial point of view ; and it was 
 not till 1880, when Baeyer discovered the synthesis from o-nitro-phenyl-propiolic 
 acid, that the question became one of practical importance. The first patent 
 was taken out by Baeyer on the 19th of March, 1880. But it was only after this 
 had been followed by nearly twenty years of continuous research that the present 
 process, covered by 152 patents in Germany alone, was arrived at, and the 
 synthetic product obtained at a price which could compete with that of the 
 natural indigo. 
 
 Bb2 
 
376 Indol Group 
 
 By 1881 the synthesis from o-nitro-phenyl-propiolic acid was so far improved 
 that indigo could be made at a cost not much greater than that of the natural 
 product, and it was actually employed to a limited extent in cotton-printing. 
 
 In 1882 Baeyer's method of preparation from o-nitro-benzaldehyde appeared. 
 This was an improvement on the older method, but only a slight one. Indeed, 
 there was a fundamental objection to both of these methods which must, in 
 spite of an improvement in detail, have prevented them from becoming serious 
 rivals of the natural process. They both start with toluene, and the supply of 
 toluene is necessarDy very limited, and so must have limited the supply of 
 artificial indigo. Toluene is of course obtained entirely from the distillation 
 of coal tar, of which the main product is benzene. The annual production of ben- 
 zene and toluene together amounts to 25,000-30,000 tons, which is almost wholly 
 used for dyeing, and consists of only about one part of toluene to four of benzene. 
 Thus only 5,000-6,000 tons are produced annually, which no more than suffice 
 for the present demand. Any increase in this production must be accompanied by 
 four times as great an increase in the production of benzene, and therefore by an 
 enormous rise in the price of toluene, unless a demand could be created for this 
 additional quantity of benzene. 
 
 Now, by the most recent and improved methods, 1 kilogram of indigo cannot 
 be obtained from less than 4 kilograms of toluene ; and the annual consamption 
 of indigo amounted, before the introduction of the synthetic compound, to 
 5,000 tons. Thus the whole of the present production of toluene would only 
 produce a quarter of the indigo consumed ; and in order to replace entirely the 
 natural by the artificial indigo it would be necessary to produce five times as 
 much toluene as is now prepared, and hence to increase the annual production of 
 benzene by at least 80,000 tons, for which there is no demand. 
 
 It was therefore evidently necessary to find some other source for indigo ; 
 and so the problem entered on a new stage when, in 1890, Heumann discovered 
 the synthesis from phenyl-glycine. The raw materials required for this were 
 aniline, acetic acid, chlorine, and alkali, which could all be obtained in any 
 quantity. But the yield by this method was small and the cost great ; and it 
 was not until Heumann discovered that the substitution of anthranilic acid for 
 aniline gave a much better yield, that the method could be used on the large 
 scale. Then, indeed, the problem was in outline solved. But it stUl required 
 seven years of unremitting labour, taxing the utmost resources both scientific 
 and technical, in inorganic as well as organic chemistry, of the two greatest dye 
 factories in the world, before the details of the process could be sufficiently 
 improved to admit of competition with the natural product. It is extraordinary 
 to see what a wide range so apparently simple a process covers, and how the 
 latest refinements in the most remote fields of chemical industry have to be 
 called in in order to perfect it, 
 
 Heumann's improved synthesis of indigo begins with the condensation of 
 anthranilic with chloracetic acid to phenyl-glycine-o-carboxylic acid. The first 
 condition of success was the cheap production of anthranilic acid from some 
 substance which could be obtained in sufficient quantity. We need only 
 consider the process which was ultimately adopted. This starts with naphtha- 
 lene. This in itself was an enormous improvement on the older methods. 
 
Commercial Synthesis of Indigo 377 
 
 Naphthalene may be obtained in any quantity at a very low price. The amount 
 of coal tar annually worked up for hydrocarbons (about two-thirds of the total 
 production of coal tar) contains from 40,000-50,000 tons of naphthalene. It had 
 long been a problem to find some use for the large quantities which are formed 
 as a by-product in making benzene and toluene. It had been used to a small 
 extent for increasing the illuminating power of coal gas (as ' albocarbon '), and 
 there had been in recent years an increasing demand for it as a source of the new 
 naphthalene azo-dyes. But the amount so employed was extremely small, and 
 the greater part was used for purposes which returned only a minute profit, 
 such as burning to make lampblack. The total quantity of naphthalene which 
 was isolated annually was not more than 15,000 tons. There remained, there- 
 fore, the difference of 25,000 tons, at the lowest estimate, which could be isolated 
 at the same price as that already on the market ; and this was more than enough 
 to supply the whole of the world's consumption of indigo. 
 
 Anthranilic acid is obtained from naphthalene through phthalic acid. This 
 is converted into phthalimide, which, on treatment with chlorine and potash, 
 goes, by a modification of Hofmann's reaction, into anthranilic acid. The 
 reaction is most easily expressed if we suppose the phthalimide to act as the 
 monamide of phthalic acid : — 
 
 fvCO-NHj, _^ f^-CO-NHCl ^ f^-N=C^O _^ r^-NH^ , ^^ 
 U-CO-OH ~* U-COOH "^ U-COOH ~* U-COOH "^ ^^2- 
 
 The first problem was the oxidation of naphthalene to phthalic acid. The 
 older method of using chromic acid was too expensive, and it was discovered that 
 the oxidation could be performed by heating the naphthalene with sulphuric 
 acid containing a high percentage of sulphur trioxide. But this would only be 
 economical if some cheap method were found for preparing this acid, and for 
 converting the sulphur dioxide which is formed in the reaction back into the 
 trioxide. Such a method was the (then) new contact process of the Baden 
 works, which is probably destined to supersede entirely the old lead-chamber 
 process. This consists, of course, in mixing the gases from the pyrites burners 
 with air, and passing them (after due purification) over platinized asbestos at 
 a suitable temperature, whereby the sulphur dioxide combines with the oxygen 
 of the air to form sulphur trioxide, which is absorbed in strong sulphuric acid. 
 By this means it was easy, not only to prepare the original acid, but also to 
 reoxidize the dioxide formed in the oxidation of the naphthalene. The process 
 is in fact reduced to a cycle. The sulphur dioxide takes up the oxygen of the 
 air and passes it on to the naphthalene. The importance of carrying on this 
 cycle economically is shown by the fact that the annual production of phthalic 
 acid leads to the formation of 35,000-40,000 tons of sulphur dioxide. If no sub- 
 stitute for the lead-chamber process had been discovered, it would have been 
 impossible to produce indigo at a sufficiently low price to compete with the 
 natural dye. 
 
 It is interesting to notice that the discovery that the addition of mercuric 
 sulphate greatly facilitates the oxidation of the naphthalene by the sulphur tri- 
 oxide was due to an accident. In one experiment the vessel in which the 
 oxidation was being carried on contained an iron cup filled with mercury (no 
 
378 Indol Group 
 
 doubt holding a thermometer). This was eaten through by the acid mixture, 
 which came in contact with the mercury, and the yield was found to be much 
 increased. This catalytic influence of the mercuiy, which occurs in other cases 
 of sulphonation as well," is remarkable, and a very probable explanation of it has 
 been given by Dimroth.' He has shown that if a mercury salt is heated with 
 an aromatic derivative, a compound is formed in which one hydrogen on the 
 ring is replaced by mercury. This substitution is quite peculiar in that the 
 position taken up by the mercury is not determined by the substituent already 
 present : it is always ortho. Now it is found that if benzoic acid is sulphonated 
 in the presence of mercury, the amount of meta and para acids produced is the 
 same as if the mercury were not there ; but in addition a considerable quantity 
 of the o-sulphonic acid is formed, which is not obtained at all in the absence of 
 mercury. This shows that the catalytic influence of the mercury is due to the 
 formation and decomposition of a mercury derivative, as it introduces the 
 sulphonic group at the position which the mercury (but no other substituent) 
 would occupy. 
 
 The anthranilic acid having been obtained, it has to be combined with 
 chloracetic acid : it was therefore necessary to improve the methods for pre- 
 paring the latter. Acetic acid is, of course, easily obtained by the distillation 
 of wood. Every year two million kilograms ' of acetic acid are used for making 
 indigo, and this quantity is got from more than 100,000 cubic metres of wood. 
 In order to chlorinate this a large and cheap supply of fairly pure chlorine 
 is required ; and chlorine is also used for an earlier stage of the synthesis, the 
 conversion of phthalimide into anthranilic acid. The ordinary processes for 
 preparing chlorine were not found to be satisfactory ; Weldon's was too expen- 
 sive, and Deacon's gave too dilute a gas. The electrolytic method of preparation 
 was finally adopted, but even here the gas required to be further purified by 
 liquefaction. 
 
 In this way the process of manufacture of synthetic indigo was ultimately 
 perfected, and the cost of production reduced to (it is believed) less than a 
 quarter of that of natural indigo. 
 
 A brief recapitulation of the method may be useful. 
 
 Naphthalene, obtained from coal tar, is oxidized to phthalic acid by heating 
 with fuming sulphuric acid prepared by the contact process. The sulphur 
 dioxide which is given oflf in this reaction is reconverted into sulphur trioxide 
 by the same process. The phthalic acid is converted into phthalimide, and this 
 by treatment with chlorine and soda (Hofmann's reaction) into anthranilic acid 
 (o-amino-benzoic acid). The anthranUic acid is then condensed with chloracetic 
 acid, which is made by chlorinating acetic acid obtained by the distillation 
 of wood with chlorine produced by the electrolysis of sodium chloride and 
 purified by liquefaction. The phenyl-glycine-o-carboxylic acid so obtained is 
 fused with potash, which gives first indoxyl-carboxylic acid and then indoxyl. 
 On treating the fused mass with water and exposing it to the air, the indoxyl 
 
 ' e. g. anthraquinone, C. 04. ii. 751 ; 05. ii. 864. 
 " Dimrotli, v. Schmaedel, Ber. 40. 2411 (1907). 
 ' In 1900 : the amount must now be far greater. 
 
Commercial Synthesis of Indigo 379 
 
 is oxidized to indigo. The more important stages are given in the following 
 formulae : — 
 
 Cf\ -* rV^^NNH -» n-COOH fvCOOH 
 
 UJ ^ U-CO/^^ ^ U-NH^ ^ U-NH-CH^-COOH 
 
 Q^_(^0 ^. (JO ^_(^0 0(f__^ 
 
 NH NH NH NH 
 
 The advances which have been made in the preparation of synthetic indigo 
 since the publication of Brunek's paper (1900) are not of a fundamental character." 
 Heumann's synthesis still holds its own, though it has been considerably 
 modified, and there are now several rival processes in the field. 
 
 The earlier stages of Heumann's synthesis have not been altered, and are in 
 fact very satisfactory as they stand. In the conversion of phthalimide into 
 anthranilic acid two intermediate compounds have been isolated.^ The first 
 
 action of the chlorine is to form the imido-chloride, [J p^)>NCl, and this is 
 converted by the alkali into [ jZtjtt.qq/O ; but nothing appears to be known as 
 
 to how this change takes place. 
 
 The conversion of the anthranilic acid into phenyl-glycine-o-carboxylic acid 
 by means of chloracetic acid gives unsatisfactory results, partly owing to the loss 
 of carbon dioxide, and partly owing to the formation of the diacetic acid deri- 
 
 aOOOT-T 
 ■NTfTT roOTT'* " ^^^^ i® ^•'^ avoided by not condensing the free 
 
 acids but their alkaline salts, which are found to react completely in aqueous 
 solution at 40° ; and the product, the acid salt of phenyl-glycine-carboxylic acid, 
 being only slightly soluble, crystallizes out, and so is removed from the further 
 action of the chloracetic acid. Although the yields obtained in this way are 
 very good, an even better method has been worked out by Bender, which is 
 based on Miller and PlOchl's ' reaction of the anhydro-bases with formaldehyde 
 and prussic acid. If the hydrochloride of anthranilic acid is treated in aqueous 
 solution first with potassium cyanide and then with formaldehyde, the nitrile 
 of phenyl-glycine-carboxylic acid separates out : — 
 
 C.h4«^H . CH.0 - C.H.<S ^ ^^ _ c.H.<««>H 
 
 \CN 
 The simple nitrile of phenyl-glycine, ^-NH-CHg-CN, can be made in the same 
 way by treating aniline with prussic acid and formaldehyde. In both cases the 
 yield is very good, and the nitriles are easily saponified. 
 
 Of recent years, an increasing use has been made of the original form 
 of Heumann's synthesis ; starting with aniline instead of anthranilic acid, and 
 going through phenyl-glycine itself, instead of its carboxylic acid. The aniline 
 and chloracetic acid are allowed to react in the presence of ferric oxide,* the 
 
 1 For an account of the development of this industry up to 1904 see Eeissert, Zeit. f. angew. 
 Chem. 17. 482. 
 
 ' Bredt, Hof, Ber. 33. 24, 27 (1900). » Ber. 25. 2020 (1892). ' C. 06. ii. 1746. 
 
380 Indol Group 
 
 glycine being thus removed as the insoluble ferric salt, and so protected from 
 further action. The glycine can even be prepared by heating nitrobenzene with 
 iron and chloracetic acid.' 
 
 For the condensation of the glycine to indoxyl derivatives Heumann uses 
 potash fusion. The conversion of phenyl-glycine-carboxylic acid into indoxyl- 
 carboxylie acid requires a temperature of 240-280°, and the yield, though fair, 
 is not nearly quantitative. The analogous conversion of phenyl-glycine into 
 indoxyl requires a higher temperature (300-350°), and the yield under ordinary 
 circumstances is very small. It was this fact which prevented the method 
 from being adopted at first. A large number of suggestions have been made 
 for the improvement of the process, of which perhaps the most important 
 is the addition of sodamide. In the presence of sodamide phenyl-glycine can 
 be converted into indigo at a temperature of 180-240°, and the yields are much 
 better. The sodamide is prepared by passing ammonia through liquid sodium. 
 Its action is so energetic that it must be diluted with potash. The water 
 evolved in the reaction liberates ammonia, and so the necessity of employing 
 special means to exclude atmospheric oxygen is avoided. In this way it is 
 possible to start with benzene, which, like naphthalene, can be obtained cheaply 
 in any desii-ed quantity. The sodamide is, of course, expensive ; but in spite 
 of this the method is used on the large scale by the Hochst dye works, while 
 the Badische at Mannheim still keeps to Heumann's method, employing at 
 the same time to some extent Bender's synthesis with formaldehyde and 
 prussic acid. 
 
 Other suggestions for improving the yield in the potash fusion ai-e the 
 addition of alkaline earths,'' of sodium oxide,' of magnesium nitride,'' or of 
 calcium carbide.^ In all these cases the real point appears to be to secure 
 the absence of water throughout the reaction. Instead of potash it is of 
 advantage to use a mixture of potash and soda in molecular proportions, 
 on account of its lower melting-point. It is also found that the yield is 
 increased if ammonia is passed over or through the fused mixture ^ ; other gases 
 such as hydrogen, nitrogen, coal gas, or the vapour of hydrocarbons may be 
 used instead with the same results.' 
 
 An essentially new and practicable synthesis has been published by Sand- 
 meyer." He starts with thiocarbanilide, which when warmed in aqueous 
 alcohol to 50-60° with potassium cyanide and white lead (basic lead carbonate) 
 loses hydrogen sulphide to form carbo-diphenyl-imide, which then adds on 
 prussic acid : — 
 
 If the product is treated with yellow ammonium sulphide for two days at 
 25-35°, it is converted into the thio-amide, ^[^^^C-CS-NHg . This thio-amide 
 
 ' C. 06. ii. 1700. ' C. 07. i. 383. ^ C. 06. i. 514. 
 
 * C. 06. i. 617. = Ibid. 6 C. 06. i. 618. 
 
 ' C. 07. i. 435, 519. » Zeif. Farb. u. Text. 2. 129 (1903). 
 
Commercial Synthesis of Indigo 381 
 
 is then heated with concentrated sulphuric acid at 95-110° till no more sulphur 
 dioxide is evolved, whereby «-isatin-anilide is produced : — 
 
 C6H5\?^2?^C=N-.^ + H,SO, = OC^tiN.^ + SO2 + NH3 + H,0 + S. 
 
 NH 
 The a-isatin-anilide, when heated with dilute mineral acids, splits up into aniline 
 and isatin, while on warming with yellow ammonium sulphide it is readily 
 reduced to indigo. In practice the solution of the anilide in concentrated 
 sulphuric acid is allowed to flow directly into water, simultaneously with a 
 
 dilute solution of sodium hydrogen sulphide. a-Thio-isatin, [JIjth/^CS, is pre- 
 cipitated ; and this when treated with alkali or alkaline carbonate gives a 
 mixture of indigo and sulphur. 
 
 This synthesis of indigo is somewhat complicated ; but the yields at each 
 stage are so good, and the original materials so cheap, that it is quite possible 
 that it may become a serious rival to the Heumann process. 
 
 The introduction of artificial indigo at first met with various difficulties, the 
 chief of which was, as Brunck points out, that the general public are quite 
 incapable of understanding what is meant by a chemical individual. They could 
 not realize that the synthetic and the natural products were (except for the 
 impurities of the latter) identical, and insisted on regarding the one as an inferior 
 substitute for the other. But these objections are rapidly disappearing. The 
 synthetic indigo is purer than the natural, and of far more constant composition, 
 which makes it much easier to use ; and above all its cost price is far lower. The 
 annual value of the indigo produced in India before 1897, when the first artificial 
 indigo was placed on the market, was about three millions sterling. By 1902 it 
 had fallen to one and a quarter millions. This was partly due to a decrease in 
 the amount produced, but largely to the remarkable fall in the price. The price 
 of indigo before 1896 was about £600 per metric ton. By 1900 it had fallen to 
 £250, and by 1905 to £115. These latter figures refer to the synthetic product: 
 the natural is still somewhat more expensive. This great fall in price has 
 naturally led to a great increase in consumption ; and the decrease in the produc- 
 tion of the natural dye has been far more than counterbalanced by the enormous 
 increase in the manufacture of the artificial. In 1900 the world's production of 
 indigo was about 5,000 tons ; in 1905 the amount exported from Germany was 
 over 11,000 tons, to which the still considerable Indian production has to be 
 added, as well as the consumption within Germany itself. The whole amount 
 cannot fall far short of 20,000 tons. 
 
 Before we leave the subject of indigo, it is interesting to notice the scientific 
 advances, both theoretical and practical, which arose more or less directly through 
 the investigation of the constitution and syntheses of the compounds of this group. 
 
 The method of reduction by passing the vapour of the substance over heated 
 zinc dust was discovered by Baeyer in 1866 in his endeavours to reduce oxindol 
 to indol ; it immediately afterwards led Graebe and Liebermann to the synthesis 
 of alizarine, and has subsequently been most widely extended. 
 
382 Indol Group 
 
 Another very important method of reduction — the use of sodium and boiling 
 amyl alcohol — was invented by Baeyer in 1879 for the reduction of dichlorindol 
 to indol at a lower temperature than that required by the zine method, in order 
 more completely to exclude the possibility of intramolecular change. 
 
 Again, to take theoretical questions, it was from the study of the ring- 
 formation in the syntheses of the indol derivatives that Baeyer came to the con- 
 clusion, in 1880, that such closing of the ring only occurs when by its means 
 a ring of either five or six atoms is produced ; and it was in the development of 
 the views thus originated that he was led to formulate his theory of strain 
 in 1885. 
 
 Finally, his investigation of isatin revealed, in 1882, three years before the 
 first of Laar's classical papers appeared, one of the earliest clearly recognized 
 cases of tautomerism ; and it was in connexion with this that he gave, in 1883, 
 the first distinct expression of his view that the peculiar behaviour of a tautomeric 
 substance is due to the mobility of a hydrogen atom which can migrate, according 
 to circimastances, from one position to another ; whereas in its two stable series 
 of derivatives this hydrogen atom is replaced by other and heavier groups, which 
 do not possess the same mobility. 
 
CHAPTER XVIII 
 6-E,INGS: C^N 
 
 Of these bodies, as of the 5-rings, only a single class will be discussed, the 
 derivatives of a ring of five carbon atoms and one nitrogen, comprising the 
 pyridine, quinoline, and isoquinoline groups. 
 
 1. PYRIDINE 
 
 The discovery of pyridine is due to Anderson, who in 1845-51 succeeded in 
 isolating it from bone oil, together with picoline (methyl-pyridine) and lutidine 
 (dimethyl-pyridine) : the three names were invented by him. The pyridine bases 
 have since been found to occur in the products of distillation of almost all kinds 
 of nitrogenous matter, and the source from which they are now usually prepared 
 is coal tar. It is to be noticed that their formation from animal matter requires 
 the presence of unsaponified fats. If the fats are removed, pyrrol compounds are 
 formed, but no pyridines ; while if free fatty acids are present, they are converted 
 into the corresponding nitriles. It would seem therefore that the production of 
 pyridine derivatives only occurs in the presence of glycerine. This is converted 
 at the high temperature into acrolein and similar aldehydes, which combine 
 with the ammonia formed from the nitrogenous matter (gelatine, &c.) to give 
 pyridine bases. 
 
 As regards the formula of p3rridine, we have as usual two questions to con- 
 sider. It is easy to show that it contains a ring of five carbons and one nitrogen, 
 with one hydrogen on every carbon ; but this leaves six valencies to be accounted 
 for, about which there has been much dispute. 
 
 In 1869 Korner proposed for pyridine a formula, 
 
 CH 
 CH^H 
 
 II T » 
 
 CH CH 
 
 which is that of benzene with one CH replaced by N. Whether this is correct 
 as to the six extra valencies or not, it certainly is so as to the arrangement of the 
 atoms. The various pyridine syntheses aiford many proofs of this : the simplest • 
 and most important is given by the preparation from pentamethylene-diamine. 
 The hydrochloride of this base on heating loses ammonium chloride to give 
 piperidine, whose formula is thus established : — 
 
 /CHj-CHax-^TT pi /CHg-CHay 
 
 CH, ^g^^^ = CH, >NH + NH,C1. 
 
 \CH2-CH/^-"2 \CH2-CH2 
 
384 
 
 Pyridine Group 
 
 Piperidine is readily oxidized (by heating with concentrated sulphuric acid, 
 nitrobenzene, or silver acetate) to pyridine, which thus must have the arrange- 
 ment of atoms which Korner supposes. 
 
 Korner's formula obviously corresponds to Kekule's benzene formula ; and, 
 as with benzene, a variety of other arrangements of the six internal valencies 
 have been suggested. Biedel's, in which the nitrogen is united to three different 
 
 /\ 
 
 , is supported by a certain amount of evidence, especially 
 
 carbon atoms, 
 
 ~^N- 
 
 ■/ 
 
 by the various reactions and syntheses of acridine (which is undoubtedly dibenzo- 
 pyridine), which point to para-linkage : in particular by its formation from 
 formyl-diphenylamine : — 
 
 o=gH 
 
 h/\ /\/CH\/\ 
 
 \/\ 
 
 •N- 
 
 /\/ 
 
 ^N- 
 
 /\/ 
 
 Formyl-diphenylamine. Acridine. 
 
 This peculiarity of acridine is exactly parallel to that of anthracene, where the 
 same question of para-linkage arises. And therefore as with benzene we accept 
 Kekul6's structure, so with pyridine we may take Korner's as the most probable : 
 in both cases with Thiele's interpretation. 
 
 The equivalence of the a- and a'-positions in pyridine has been proved (if the 
 proof is of any value) by synthesizing a-phenyl-a'-tolyl and a-tolyl-a'-phenyl 
 pyridine, which were found to be identical.' 
 
 The syntheses of pyridine derivatives are numerous. The following are 
 among the most important : — 
 
 Aldehyde ammonias condense with themselves or with other aldehydes or 
 ketones thus : — 
 
 CH3CHO + Ho 
 
 CH 
 CH3 
 
 HCO 
 
 H,N 
 
 ^H 
 
 CHCH, 
 / 
 
 4H„0 -*- 
 
 CIi3'CHii-C5 
 
 H 
 
 C 
 
 """^n/ 
 
 CH 
 
 II 
 
 CCH, 
 
 the product being a-methyl-/3'-ethyl pyridine, known as aldehyde collidine. 
 
 A similar reaction is the formation of picoline by the condensation of 
 glycerine with ammonia, to which its production in the dry distillation of 
 animal substances is no doubt due. The glycerine first gives acrolein : — 
 CHO CH 
 
 CH CH=CH, ^ cA3.CH3^2„0. 
 
 CH, 
 
 CHO 
 NH, 
 
 CH CH 
 
 ' Soholtz, Wiedemann, Ber. 36. 845 (1903). 
 
Pyridine: Syntheses 385 
 
 Secondly, there is Hantzseh's synthesis from ^-diketo-compounds or ketonic 
 
 esters with aldehydes and ammonia. This gives dihydro-pyridine derivatives : 
 
 e. g. dihydro-collidine dicarboxylic ester from acetaldehyde and acetoacetic ester. 
 
 The reaction has been further investigated by Knoevenagel,^ who finds that it 
 
 takes the following course. The aldehyde and the ammonia combine to form 
 
 aldehyde ammonia ; this reacts with a molecule of acetoacetic ester to form 
 
 CH •CO'C-CO Et 
 ammonia and ethylidine acetoacetic ester, ^ II ^ . The ammonia then 
 
 CH-CHg" 
 
 acts on another molecule of acetoacetic ester, forming /3-amino-crotonic ester, 
 CHa-CtiCHCOaEt 
 
 1 TT • Finally, one molecule of each of these products condenses 
 
 to give the dihydro-collidine derivative : — 
 
 + H,0. 
 
 This is the reaction which occurs when the acetoacetic ester is in excess ; but 
 a different, though similar, product is obtained if the aldehyde is in excess. In 
 this case an unreduced pyridine derivative is formed. The simplest explanation 
 is that two molecules of aldehyde condense with one of acetoacetic ester and one 
 of ammonia to a dihydro-compound, and then this is oxidized by excess of 
 aldehyde to a simple pyridine : — 
 
 CHo 
 
 CH3 
 
 CH 
 
 hB. 
 
 COaEt-Cr H-C-COaEt 
 CH3-C=0 /C-CHo 
 
 COaEt-C C-COaEt 
 
 Oxig'C G'Cxlg 
 
 H^N 
 
 NH 
 
 /CO2R ^pxT CO^R 
 
 HC— C<Vfi3 J. 
 
 2-' 
 
 OH C==CCH3 
 
 CHa-CHO + + NH3 -> CH3CH ^H 
 
 .OH CH=CH 
 
 HCH==CH 
 
 oR 
 
 <j!02 
 
 C CCH3 
 
 -* CH3C N . 
 
 CH=CH 
 a-y-dimethyl-pyridine- 
 /3-carboxylic ester. 
 
 A remarkable synthesis has been discovered by E. v. Meyer,^ which starts 
 from the so-called benzo-aceto-dinitrile. This is an unsaturated amino-nitrUe 
 formed by the addition of benzonitrile to acetonitrile under the influence of 
 sodium (by a sort of polymerization). It will condense with acetoacetic ester in 
 
 1 Ber. 31. 745 (1898); of. Beyer, Ber. 24. 1662 (1891). 
 
 2 C. 05. i. 262. 
 
386 Pyridine Group 
 
 presence of hydrochloric acid to give a cyan-hydroxy-pyridine (or, in the tauto- 
 meric form, pyridone) : — 
 
 EtOCO CO C^H 
 
 CNCH H^CH CN-C^^H CNC CH 
 
 II I = II II or II I . 
 
 0Cv O^CCHg 0C C-CHj ^.n c-CHs 
 
 The synthesis from pentamethylene-diamine through piperidine has already 
 been given ; and also that from potassium pyrrol and chloroform or methylene 
 iodide. 
 
 The pyridine bases are colourless liquids of peculiar smell. Pyridine itself is 
 miscible with water, but the higher homologues rapidly become less soluble, and 
 are generally more soluble in cold water than hot. 
 
 Pyridine boils at 115-2°. It is a weak base of about the strength of aniline ' ; 
 but it is much stronger than pyrrol, and forms fairly stable salts. It is not 
 unfrequently used in order to bring about reactions which involve the separa- 
 tion of hydrochloric or hydrobromic acid, the reaction being sometimes carried 
 out in pjrridine solution. It would seem that this action does not merely depend 
 on the basic character of pyridine ; but it is not properly understood.' Thus it 
 has been used for benzoylation ' ; the benzoyl chloride is dissolved in pyridine 
 (which causes great evolution of heat, suggesting the formation of some com- 
 pound), and the amine, amide, or alcohol is added. Again, pyridine (or quinoline) 
 will remove hydrobromic acid, for example, from bromo-succinic ester, giving 
 f umaric ester * ; and this has the advantage over the use of alkali that the ester 
 group is not saponified. Pyridine forms an efficient carrier in the chlorination 
 and bromination of aromatic hydrocarbons," and it also acts as a catalyst (as 
 does quinoline) in promoting both the Claisen reaction and the formation of 
 Grignard's reagent.* 
 
 As we should expect from the formula, the derivatives of pyridine exhibit the 
 aromatic character in a very high degree. This is shown even in their physical 
 properties — the small molecular volume and the high refiactive power — and also 
 in their very remarkable stability, Pjrridine is even more stable than benzene. 
 It resembles benzene in not being oxidized by chromic acid or fuming nitric 
 acid ; and that it is still more stable than benzene is shown by the fact that 
 when phenyl-pyridine is oxidized, it is the benzene and not the pyridine nucleus 
 that is destroyed, the product being pyridine carboxylic acid. 
 
 It has recently been found, however, that the pyridine can be broken between 
 the nitrogen and the carbon by several somewhat unexpected reagents: for 
 example, by dichloro-nitrobenzene (Zincke), by cyanogen bromide (Konigs), and, 
 
 ' Constam, White, Am. C'k. J. 29. 1 (C. 03. i. 523). 
 
 2 Cf. Eckstein, Ber. 39. 2135 (1906). 
 
 '■' Freundler, C. R. 136. 1553 ; 137. 712 ; Bull. Soc. [3] 31. 616 (1908, 1904). 
 
 < Dubreuil, C. B. 139. 870 (C. 05. i. 25). ^ Cross, Cohen, Proc. C. 8. 24. 15 (1908). 
 
 » Tingle, Gorsline, Am. Ch. J. 37. 483 (C. 07. li. 30). 
 
Pyridine : Properties 
 
 387 
 
 •what is more remarkable, by sodium bisulphite, an intermediate addition- 
 compound being formed in this case.^ 
 
 As in the benzene series, the homologues of pyridine are easily oxidized by 
 strong oxidizing agents (especially potassium permanganate), the side chains, 
 whatever their length, being converted into carboxyl. 
 
 Almost the only reactions of which pyridine is capable are substitutions, and 
 these occur less readily than with benzene. The halogens substitute more or 
 less easily, sulphuric acid only at a high temperature, and nitric acid usually not 
 at all. Marckwald " has advanced a theory to account for this latter phenomenon, 
 as well as certain other peculiarities of pyridine. He points out that a nitrogen 
 atom in a ring is able to behave under certain circumstances as a strongly 
 negative group. For example, tetrazol, 
 
 HN N 
 
 HC N, 
 
 is an acidic compound which reddens blue litmus. This is not necessarily in- 
 compatible with the nitrogen also possessing basic properties, i. e. the power of 
 becoming pentad. In the compounds in which it remains triad and exhibits its 
 negative character, it behaves like a carbon atom in the benzene ring attached 
 to a strongly negative substituent, e. g. C-NOg. Thus, if pyridine is substituted 
 the substituent takes up the /3-position, as nitrobenzene gives meta-di-derivatives. 
 Again, in the chloro-pyridines we find the «- and y-chlorine extremely mobile — 
 easily replaced by NHg — as in o- and j)-chloro-nitrobenzene, whereas in /3-chloro- 
 pyridine, as in wi-chloro-nitrobenzene, it is very firmly attached : — 
 CI CI 
 
 /\ /\ /\ /\ 
 
 V 
 NO 
 
 -CI 
 
 V 
 
 NO. 
 
 \^/ 
 
 -CI 
 
 \^/ 
 
 ortho 
 
 para a- 
 
 Labile. 
 
 /\ 
 
 CI 
 
 A-Cl 
 
 \n/ 
 
 NO, 
 meta ^- 
 
 Stable. 
 
 Again, the repellent influence of this negative nitrogen explains why it is so 
 difficult to introduce into pyridine negative substituents like SO3H and NOg. 
 But if the acidifying influence of the nitrogen is diminished by the introduction 
 of a basic group, nitration becomes possible; amino-pyridine can be nitrated 
 directly. This, however, is a general characteristic of aromatic NH,. 
 
 Although pyridine is so stable a compound it is comparatively easily reduced. 
 When treated with sodium and alcohol in the cold, it takes up six hydrogen 
 atoms and forms piperidine. Here, again, we may compare its behaviour with 
 that of certain negatively substituted benzenes, namely, the carboxylic acids, 
 which can easily be made to take up two or four hydrogen atoms in the cold. 
 On the other hand, it is remarkable that if the vapour of pyridine mixed with 
 hydrogen is passed over heated nickel, no piperidine is formed at any tempera- 
 
 Buoherer, Schenkel, Bei: 41. 1316 (1908). 
 
 Ber. 26. 2187 (1893); 27. 1317 (1894). 
 
388 
 
 Pyridine Group 
 
 ture.' Below 220° normal amylamine is produced ; while above 220° the 
 nitrogen is split off from the carbon in both sides, with the formation of 
 ammonia and pentane. The absence of piperidine is due to the fact that the 
 reverse action takes place much more easily : if piperidine vapour is passed over 
 heated nickel it breaks up into pyridine and hydrogen. 
 
 Pyridine is also capable of addition to the nitrogen, which becomes pentad. 
 The pyridine compounds are, as has been pointed out, moderately strong bases, 
 though weaker than the reduced compounds, such as piperidine. They form 
 fairly stable salts with one equivalent of acid, and also characteristic gold and 
 platinum double salts. They further combine directly with any metallic salts 
 in the same way as ammonia. 
 
 They are also able to add on alkyl iodide to form the pyridinium derivatives, 
 which undergo three remarkable changes. On heating, the alkyl group migrates 
 to the a- or y-position on the ring, as in Hofmann's aniline reaction, where the 
 0- and jp-positions are taken up : — 
 
 CHg 
 
 / \ 
 CH, I 
 
 / \ 
 
 H I 
 
 X 
 
 H I 
 
 On treatment with silver hydrate the pyridinium iodides yield strongly basic 
 hydroxides (resembling the quaternary ammonium bases), and when these are 
 warmed the oxygen migrates to the carbon. The reaction is complicated, but 
 apparently gives a dihydro-pyridine and a pyridone : — 
 
 H 
 
 H- 
 
 H 
 
 H 
 
 Hr^H 
 
 s]sr^ 
 
 i-H 
 
 H 
 
 / \ 
 CH, OH 
 
 CH. 
 
 H, 
 
 H 
 
 H 
 
 ./\h 
 
 U 
 
 CH3 
 
 This behaviour is common with hydroxyl attached to doubly linked pentavalent 
 nitrogen ; as we have seen in the triphenyl-methane dyes, and the diazonium 
 hydroxides. The quaternary bases formed by the addition of methyl iodide to 
 «- or y-alkyl pyridines, which, like the others, are colourless, when treated with 
 soda lose hydriodic acid and are converted into bright yellow so-called pyridane 
 compounds.^ On dilution, or on dissolving the compound in pure water, the 
 colour disappears again ; water is added on and the quaternary hydroxide 
 formed : — 
 
 CH3 
 
 /\ 
 
 / \ 
 CH, I 
 
 -CH,.0 ^ L/'^CH.^ 
 
 V 
 
 -CH,.<^. 
 
 Colourless. 
 
 Yellow. 
 
 CH3 OH 
 
 Colourless. 
 
 1 Sabatier, Mailhe, G. II. 144. 784 {C. 07. ii. 73). 
 
 2 Decker, Ber. 38. 2493 (1905). 
 
Pyridine : Properties 
 
 389 
 
 This curious reaction, with its characteristic change of colour, also occurs in the 
 quinoline and isoquinoline series, and in compounds in which the nitrogen is 
 replaced by oxygen or sulphur, such as the xanthones and thio-xanthones. 
 
 The determination of position among the pyridine derivatives is carried out 
 as follows : The mono-derivatives (omitting the N-compounds, which obviously 
 belong to a different class) are of three kinds, a-, ^-, y-, as required by the 
 formula. The position is determined by reference to the three monocarboxylie 
 acids, which are known respectively as picolinic, nicotinic, and isonicotinic acids, 
 and whose constitution has been made out by Skraup. The starting-point is the 
 naphthoquinolines, which are obtained by a modification of the well-known Skraup 
 synthesis of quinoline — from a- and /3-naphthylamine respectively. This reaction , 
 the details of which will be discussed later, consists in condensing an aromatic 
 amine with glycerine, the 3-carbon chain joining up with the nitrogen of the 
 NHg to form a 6-ring. Since both a- and /3-naphthylamine give in this reaction 
 compounds undoubtedly containing a pyridine nucleus, there can be no doubt 
 about their formulae : — 
 
 ^\/\ /\/\ 
 
 NH, I 
 
 a-naphthyl- a-naphtho- 
 amine. quinoline. 
 
 /\/COOH 
 
 COOH 
 
 \/\/\^\AA ■ 
 
 N 
 
 \y'Y\ 
 
 N 
 
 CO,H- 
 
 V 
 
 Picolinic 
 acid, (a-) 
 
 \/\Ah, 
 
 /3-naphthyl- 
 amine. 
 
 /\/\ 
 
 N 
 
 /3-naphtho- 
 quinoline. 
 
 
 CO,H-/'\ 
 
 Nicotinic 
 acid. (/3-) 
 
 On oxidizing these two naphthoquinolines, they both split the middle ring 
 to give phenyl-pyridine dicarboxylic acids, which lose two molecules of carbon 
 dioxide to form phenyl-pyridines. Now it is obvious from the formulae that 
 «-naphthoquinoline must give a-phenyl-pyridine, and the /3-compound /3-phenyl- 
 pyridine. Finally on oxidation the phenyl group is destroyed, only carboxyl 
 remaining. Hence we know that the resulting pyridine monocarboxylie acids 
 must be a- and ^- respectively. And it is found that the acid obtained from 
 a-naphthylamine is picolinic, which must therefore be the a-, and that from 
 /Q-naphthylamine, which must be the /3-, is nicotinic. The third acid, isonico- 
 1175 c c 
 
390 Pyridine Group 
 
 tinic, which is got from the third isomeric phenyl-pyridine, must therefore 
 be the y-carboxylic acid : — 
 
 
 
 
 
 COOH 
 
 
 a-. 
 ■COOH Pi«oli°i«- ' 
 
 
 •COOH ^. 
 
 Nicotinic. 
 
 A 
 
 Isonicotimc 
 
 A rather neater method of orienting these acids depends on the following 
 reactions. Quinoline and isoquinoline, whose formulae are proved by their 
 syntheses, both give on oxidation dicarboxylic acids of pyridine. Quinoline 
 gives quinolinic acid, which thus has the carboxyls in the a- and ;3-positions, 
 and isoquinoline gives cinchomeronic acid, where the positions must be 
 ;3- and y- : — 
 
 /\/\ 
 
 ^N^X/ 
 
 ^"^COOH 
 
 \ /-COOH 
 
 Quinolinic 
 acid (a,/3). 
 
 Vv 
 
 |/^-COOH 
 ^^COOH ' 
 
 Cinchomeronic 
 acid (/3,y). 
 
 Both these acids can be made by heating to lose one molecule of carbon dioxide, 
 leaving monocarboxylic acids. Quinolinic gives only one such acid, nicotinic, 
 and cinchomeronic a mixture of nicotinic and isonicotiuic. Hence nicotinic 
 acid, which is given by both, must have the carboxyl in the position common 
 to both, that is, the /3-position. Isonicotiuic acid, on the other hand, must retain 
 the other carboxyl of cinchomeronic acid, that is, it must be the y-acid. 
 
 The orientation of the di-substitution-products of pyridine is fixed by refer- 
 ence to the six dicarboxylic acids, whose constitution has been determined by 
 Ladenburg by a similar method. 
 
 It is remarkable that while the carboxylic acids of pyridine and quinoline, 
 and their esters, amides, and nitriles, resemble the corresponding derivatives 
 of benzene, the acid chlorides are quite dififei-ent. They do not dissolve in 
 organic solvents except when they react with them, their melting-points are 
 very high, lying near those of the hydrochlorides of the acids, and they ai-e 
 marked by great stability and the absence of smell. These peculiarities suggest 
 that they are really polymers of the acid chlorides ; but as no solvent can be 
 found in which to determine their molecular weight, this question cannot 
 be settled.^ 
 
 The amino-derivatives of pyridine have a somewhat singular behaviour. 
 The a- and y- may be obtained, as we have seen, by treating the corresponding 
 chlorine compounds with ammonia. The /3-derivative cannot be prepared in 
 this way, as the chlorine is too firmly attached. It is therefore necessary to use 
 an indirect method, that by which the amino-pyridines were first obtained. 
 This is by means of the Hofmann reaction, the amide of nicotinic acid being 
 converted by bromine and soda into /3-amino-pyridine : — 
 
 H. Meyer, Ber. 38. 2488 (1905). 
 
A mino-pyridines 
 
 391 
 
 /\ 
 
 -CO-NHa <^ VNH, 
 
 Xn/ 
 
 
 The /3-amino-compounds, in which, as in the meta- derivatives of nitroben- 
 zene, the negative influence of the nitrogen on the NHg is least, behave, as is to 
 be expected, like aromatic amines ; they give with nitrous acid diazo-compounds, 
 which form diazo-amino-bodies and azo-dyes. But the behaviour of the a- and 
 ■y-amines is peculiar. In dilute acid solution nitrous acid has no action on them 
 at all. In concentrated sulphuric acid they can be diaaotized, but the diazo- 
 compounds cannot be isolated ; if the sulphuric acid solution is poured on ice, 
 the diazo-nitrogen is evolved quantitatively, and the oxy-pyridine formed : — 
 
 \^/ 
 
 NH, 
 
 .\n/ 
 
 -Na-OH 
 
 \n/ 
 
 -OH 
 
 Of the reduction-products of pyridine the most important is piperidine or 
 hexahydropyridine, 
 
 NH 
 
 It is to be noticed, however, that in this series the tendency to form the hexa- 
 hydro-eompound on reduction is less marked than usual. Thus it is much 
 easier to obtain the di- and tetrahydro-compounds directly than with the benzene 
 derivatives, where it is scarcely possible to stop short of complete reduction. 
 
 Piperidine is a colourless liquid boiling at 106°, which is miscible with water. 
 It can be obtained from pepper, whence the name. Its formation by heating 
 the hydrochloride of pentamethylene diamine (or t-chloramylamine) has already 
 been mentioned, and also the ease with which it can be oxidized to pyridine. 
 It is a strong base. The ring is weakened by the reduction," and can be broken 
 by a variety of reagents. Thus oxidation with hydrogen peroxide converts 
 it partly into glutarimide (to which it bears the same relation as tetrahydro- 
 pyrrol to succinimide) and partly to 6-amino-valerianic aldehyde : — 
 
 CH2 
 
 QH^ CHO 
 
 S-amino- 
 valerianic 
 aldehyde. 
 
 CH, 
 
 CO 
 
 ?■ 
 
 H, 
 
 NH 
 Piperidine. 
 
 CO 
 
 NH 
 
 Glutarimide. 
 
 ' The same is the ease with the tetrahydro-derivatives, 
 Wallach, ibid. 2803 (1905). 
 
 Cf. Lipp, Widnmann, Ber. 38. 2471; 
 
 cc2 
 
392 Pyridine Alkaloids 
 
 ALKALOIDS 
 
 Among the compounds of the pyridine group which are found in nature are 
 several important members of the class of alkaloids. The alkaloids are defined 
 by Ladenburg' as 'those naturally occurring vegetable substances of basic 
 character which contain at least one nitrogen atom forming part of a hetero- 
 cyclic ring'. These bodies^ are of great interest from many points of view, and 
 especially as providing examples of the methods used in determining the for- 
 mulae of bodies of natural origin, that is, bodies whose origin gives us no clue 
 to their structure. The most important alkaloids of the pyridine group are 
 coniine, piperine, and nicotine. 
 
 Coniine, dextro-a-w-propyl-piperidine, 
 
 CHj OH2 
 
 CH2 CH-CH2'CH2-CIl3 
 
 is of interest as the first alkaloid whose synthesis was successfully carried out, 
 owing no doubt to its comparatively simple structure. Coniine occurs in 
 hemlock (Conium maculatum, whence the name), especially in the seeds, 
 together with several other alkaloids. It is a colourless liquid boiling at 167°, 
 which has a strong smell and is a violent poison. It was discovered in hemlock 
 seeds by Giesecke in 1827 ; its correct molecular formula, CgHj^N, was estab- 
 lished in 1881 by Hofmann, who also succeeded in throwing much light on its 
 structure ; and its synthesis was carried out by Ladenburg in 1886. In investi- 
 gating its formula Hofmann made use for the first time of his method of 
 exhaustive methylation.' This reaction has already been mentioned ; but it is 
 so important for determining the structure of alkaloids that it is worth considering 
 more in detail. 
 
 If a quaternary ammonium halide is heated, it splits ofi' the halogen together 
 with the smallest of the alkyl groups. The same occurs with the halide of 
 a tertiary base ; and this fact has been made use of in the case of the alkaloids.'' 
 When the hydriodide of a base with a methyl on the nitrogen is heated, it forms 
 methyl iodide : — 
 
 /CH3 
 'R=^-1I = K=NH + CH3L 
 
 The methyl iodide can then be determined, as in Zeisel's method, with alcoholic 
 silver nitrate. 
 
 ^ Ann. 301. 117 Anm. 
 
 ^ See J. Schmidt, Konstitution der Pflanzenalkaloide, 1904 ; Die Alltaloidchemie in. d.Jalirea 
 1900-1904 (Enke, Stuttgart). 
 
 ^ Ann. 78. 268 ; Ber. 14. 494, 659 (1881). 
 
 ' Herzig, H. Meyer, Mon. 15. 613 (1894) ; 16. 599 (1895) ; 18. 379 (1897). 
 
Coniine 393 
 
 But if the halogen of a quaternary halide is replaced by hydroxyl, and this 
 body is heated, then it is the largest of the alkyl groups which splits off from 
 the nitrogen. This is the foundation of the method of exhaustive methylation, 
 which was discovered by Hofmann, and fully explained some time later by 
 Ladenburg. 
 
 In the case of an alkaloid, where the nitrogen forms part of a ring, the 
 largest hydrocarbon radical is of course the ring itself. When the quaternaiy 
 hydroxide is heated, this breaks off from the nitrogen on one side, but still 
 remains attached on the other. The result is that only a molecule of water 
 is lost, and a base with the same number of carbon atoms remains. For example, 
 with piperidine : — 
 
 CH, (j)H, 
 
 CH2 CH2 = CH2 CH, + H,0. 
 
 ^CH, 
 
 N 
 
 V 
 
 cnf^ 
 
 "CH3 
 
 N-OH 
 
 The product adds on another molecule of methyl iodide, and the corresponding 
 hydroxide when heated separates the nitrogen from the long chain altogether, 
 giving a hydrocarbon piperylene, CgHg. The reaction is most naturally repre- 
 sented thus : — 
 
 CH2 CH, CH ^CH 
 
 I II ^ II II 
 
 CH2 CHg _ CH2 CH2 + H2O. 
 
 N-OH 
 
 N 
 
 CH3CH3 CH3 CH3CH3 CH3 
 
 This must undoubtedly be the formula of the primary product, and was for 
 a long time supposed to be that of piperylene itself. But certain peculiarities 
 in its behaviour led Thiele ^ to doubt this, and he has been able to show that its 
 true formula is CH2=CH-CH=CH-CH3 . One of the strongest proofs is its 
 reaction with potassium permanganate. This must break the chain at the 
 double bonds ; and di-vinyl-methane would give formic and malonic acids : 
 whereas the actual products from piperylene are formic and acetic acids. It is 
 evident that the di-vinyl-methane first formed undergoes a rearrangement, no 
 doubt under the influence of the simultaneously produced trimethylamine, which 
 exerts a similar effect on the unsaturated lactones.'' 
 
 This intramolecular change, however, does not affect the value of the reaction, 
 which essentially consists in splitting out the nitrogen from the ring of which 
 it forms part, and hence enables us to determine what the nature of that ring 
 was : a point of the utmost importance in investigating a class of bodies like the 
 alkaloids, whose origin leaves us quite in the dark as to the type of compounds 
 to which they may belong. 
 
 It has recently been discovered that if the compound contains methoxyl 
 groups, as many alkaloids do, these are converted by the action of the methyl 
 1 Ann. 319. 226 (1901). » lb. 129. 
 
394 Pyridine Alkaloids 
 
 iodide into hydroxyls, so that a certain caution is necessary in employing the 
 method in these cases.' 
 
 In the case of coniine, which contains no oxygen, we do not meet with any 
 such difficulties. When coniine is subjected to this treatment, Hofmann showed 
 that it passes through the following stages : — 
 
 CgHi.N -> C8Hie(CH3)NI -> C3Hie(CH3)N.OH -^ C,-a^,{C,Tl^)^Ti 
 
 Coniine. 
 
 -^ C8Hig(CH3)3N-OH -^ CsHi, + N(CH3)3 + H^O. 
 
 Conylene. 
 
 The resulting hydrocarbon differs from piperylene, as coniine does from piper- 
 idine, by CjH,;. Hofmann therefore suggested that coniine was a homologue 
 of piperidine, and soon after Konigs proposed for it the formula of a propyl- 
 piperidine, though there was as yet no evidence as to the disposition of the three 
 extra carbon atoms. 
 
 Konigs' view was completely confirmed by the very unexpected results 
 which Hofmann obtained when he distilled coniine hydrochloride with zinc 
 dust.'' His object was to reduce it ; but instead of this, he found that the 
 coniine lost six atoms of hydrogen, and gave a base of the formula CgHjiN, 
 which he called conyrine. This was easily shown to be a pyridine derivative, 
 and on oxidation gave picolinic acid (pyridine a-carboxylic acid). It therefore 
 only contained one side chain, and that in the a-position. It followed that 
 conyrine was a-propyl- or a-isopropyl-pyridine, and coniine, containing six 
 hydrogen atoms more, was a-propyl- or a-isopropyl-piperidine. That it was a 
 propyl and not an isopropyl compound was rendered veiy probable by Hofmann's 
 further discovery ' that coniine on reduction with hydriodic acid is converted 
 into ammonia and normal octane, which indicates that it contains an unbranched 
 chain of carbon atoms : — 
 
 CHj CH2 
 
 CH2 CIH2 -f 4H CH^^CHa 
 
 CH2 CHCH2CH2CH;; ^ CH3 CH2CH2CH2-CH3. 
 NH + NH3 
 
 The question was finally set at rest by Ladenburg's synthesis of coniine 
 in 1886.' 
 
 The first point was to prepare a-propyl-pyridine. This he endeavoured to 
 do by heating propyl-pyridinium iodide. As we have seen, the alkyl-pyridinium 
 iodides when heated change into C-alkyl-pyridines, the alkyl taking the a- or 
 y-position ; and so it should have been possible to obtain normal propyl-pyridine 
 from normal propyl-pyridinium iodide : — 
 
 /\ y\ 
 
 N]Sf/ \j^/-C!H2-CH2-CH3. 
 
 / \ / \ 
 
 I CHa-CHa-CHa I H 
 
 ' Freund, Becker, Ber. 36. 1528 (1903). ' Ber. 17. 825 (1884). 
 
 ' Ber. 18. 13 (1885). * Ber. 22. 1404. 
 
Coniine 395 
 
 It was found, however, that whether you start with normal or isopropyl- 
 pyridinium iodide, the same product is formed,^ and this will not yield coniine, 
 because, as was afterwards discovered, it is the isopropyl compound : the alkyl 
 undergoing a rearrangement in the course of its migration. 
 
 The method by which Ladenberg finally succeeded in preparing the normal 
 propyl compound was to condense a-picoline (a-methyl-pyridine) with paralde- 
 hyde, by heating under pressure. This gives a-allyl-pyridine ; and when this 
 is reduced with sodium in alcoholic solution, the allyl group is converted into 
 normal propyl, and at the same time the pyridine ring is reduced to piperidine : — 
 
 H 
 
 h/\h 
 
 V /CHs + 0=CHCH3 ^ h'^ JJ-CH=CHCH3 
 
 H. 
 
 H,,/ ^H, 
 
 
 H 
 
 In this way he obtained a-w-propyl-piperidine. But as this contains an asym- 
 metric carbon atom, the product is of course a mixture of the dextro- and 
 laevo-forms. If its salt with dextro-tartaric acid is recrystallized from water, 
 the salt of the dextro-base separates out first, and when this is treated with 
 alkali it gives the dextro-«-M-propyl-piperidine, which was shown to be identical 
 with the natural coniine. As this was the first synthesis of a natural alkaloid, 
 it is worth pointing out that it is a complete one ; that is to say, it makes 
 it possible to prepare coniine from its elements. The preparation of aldehyde, 
 (^-tartaric acid, and glycerine from their elements is sufficiently obvious ; from 
 glycerine to a-picoline the stages are : allyl bromide, trimethylene bromide, 
 pentamethylene-diamine, piperidine, pyridine, pyridine methyl iodide, a-picoline. 
 A different synthesis of coniine has since been carried out by Engler.^ He 
 distilled the calcium salt of picolinic acid with calcium propionate, and so got 
 a-ethyl-pyridyl ketone. On energetic reduction of this with sodium and alcohol, 
 the ketone group is converted into CHg, and also the ring reduced to piperidine, 
 and thus a-w-propyl-piperidine, that is, inactive coniine, is obtained : — 
 
 H H^ 
 
 /\ h/\h h,/^P, 
 
 -COOca H 
 
 I 
 
 COCH,-CH, Hal.'HCHaCHa-CH, 
 
 H- CHs-CHgCOOca H 
 
 ' Ladenburg, Ber. 18. 1587 (1885). " Ber. 24. 2530 (1891). 
 
396 Pyridine Alkaloids 
 
 Piperme 
 
 Piperine occurs in the seeds of various kinds of pepper. Its formula is 
 C17H19NO3. When it is boiled with alcoholic potash it takes up a molecule 
 of water, and is converted into piperidine and a monobasic acid known as 
 piperinic acid ' : — 
 
 Ci,HigN03 + H2O = CsHnN + Ci^HioN,. 
 Piperine. Piperidine. Piperinic acid. 
 
 The reverse change occurs when the acid chloride of piperinic acid is heated 
 with piperidine, piperine being formed." It follows that piperine must be a 
 kind of amide of piperidine with piperinic acid, and must contain the group 
 
 ^,^CH2— Cllgv 
 CH2 >NCOC,iH902. 
 
 \CH2-CH2 
 
 Hence it only remains to determine the formula of piperinic acid. 
 
 This was done by Fittig. He showed in the first place that when piperinic 
 acid is oxidized with potassium permanganate, it gives first an aldehyde, pipero- 
 nal, and then the corresponding acid, piperonylic acid, C^HgO^. This last acid 
 when heated with hydrochloric acid is converted with separation of carbon into 
 protocatechuic acid, which was shown to be 3,4-dihydroxy-benzoic acid: — 
 
 H0./\ 
 
 HO 
 
 + C. 
 COOH 
 
 Since this acid can be reconverted into piperonylic acid by heating with methylene 
 iodide and alkali, the latter must be the methylene ether of protocatechuic acid, 
 and piperonal the corresponding aldehyde, the methylene ether of protocatechuic 
 aldehyde : — 
 
 H0-/\ ^0- 
 
 CH2 
 
 COOH HO-I 
 
 /\ 
 
 CH2 
 
 COOH \0- 
 
 \/ 
 
 CHO 
 
 Piperonylic acid. Protocatechuic Piperonal. 
 
 acid. 
 
 Now piperonylic acid is got from piperinic acid by splitting off C^H^ ; and as 
 the product contains only one carboxyl group, the C4H4 must have been between 
 the carboxyl of the piperinic acid and the benzene ring. Thus we get for 
 piperinic acid the structure 
 
 CH.] 
 
 ^»-\/ 
 
 -C4H4COOH 
 
 and it only remains to determine the arrangement of the C4H4 group. The 
 behaviour of piperinic acid, and especially of its reduction-products, shows that 
 
 ' Rochleder, Werthelm, Ann. 54. 255 (1845) ; Babo, Keller, J. pr. Ch. 72. 53 (1857). 
 ' Kugheimer, Ber. 15. 1390 (1882) ; Fittig, Remsen, Ann. 159. 142 (1871)'. 
 
Piperine 397 
 
 it contains a normal side chain with two double links, and so the only probable 
 formula appears to be this ' : — 
 
 ^0-. .-CH=CHCH=CHCOOH 
 
 This view is completely supported by Ladenburg and Scholtz's synthesis of 
 piperinic acid." They started with protocatechuic aldehyde, which can be 
 synthesized, and which when treated with methylene iodide gives piperonal.' 
 Piperonal condenses with acetaldehyde, when heated with it in very dilute soda 
 solution, to give piperonyl-acrolein. Finally, this product, when heated with 
 acetic anhydride and sodium acetate, condenses after the manner of the Perkin 
 I'eaction, giving piperinic acid : — 
 
 ^0 
 
 CH2 
 
 -\/ 
 
 CH3CHO Q-g^' 
 -CHO *■ ^0- 
 
 \X 
 
 -CH=CHCHO 
 
 Piperonal. Piperonyl-acrolein. 
 
 CH,COOH ^^jj 
 
 "\/ 
 
 -CH=CHCH=CHCOOH 
 
 Piperinic acid. 
 
 The acid can then be converted into its chloride, and this, as we have seen, 
 combines with piperidine to form piperine, whose complete synthesis is thus 
 possible. 
 
 Nicotme 
 
 Nicotine occurs in the tobacco leaf, as the salts of citric and malic acids, to 
 the extent of 0-5-8 per cent. Its formula is C10H14N2 . It is a di-tertiary base ; 
 that is, both its nitrogen atoms are joined to carbon by three bonds. This is 
 shown by the fact that it gives two different quaternary iodomethylates. 
 
 On oxidation with nitric acid, chromic acid, or potassium permanganate, it 
 yields nicotinic acid (pyridine /3-carboxylic acid). It therefore contains the 
 pyridine nucleus connected to carbon in the /3-position. 
 
 With bromine it gives two different dibromo-compoimds — dibromo-cotinine 
 CjoHioBrgNaO, and dibromo-ticonine CioHgBrgNaO^. The first of these gives, on 
 treatment with alkali, methylamine, oxalic acid, and ^-methyl-pyridyl ketone, 
 
 /N-CO-CHg 
 
 7 
 
 the second gives with alkali methylamine, malonic acid, and nicotinic acid. 
 The formation of methylamine from both of these substances proves that 
 
 ' Fittig, Mielck, Ann. 172. 134 (1874). ' Ber. 27. 2958 (1894). 
 
 = Wegscheider, Mon. 14. 382 (1893). 
 
398 
 
 Pyridine Alkaloids 
 
 nicotine contains a methyl group attached to nitrogen, and therefore cannot be 
 a dipyridyl compound, as might otherwise be supposed. Further, if we put 
 together the fragments of chains obtained in these two reactions we get from 
 the first : 
 
 /^C-C C-C -NCH. 
 
 ^N^ 
 
 and from the second ; 
 
 /\. 
 
 \^/ 
 
 C-C-C 
 
 -N-CH... 
 
 From the fact that we can either split off a 2-carbon chain, and leave two 
 carbons attached to the nucleus, or split off three, and leave one, it follows that 
 these four carbon atoms must form a continuous chain, and the nicotine molecule 
 must contain 
 
 /Vc-C-C-C 
 
 + -N-CH.. 
 
 V 
 
 Now we have seen that both the nitrogen atoms are tertiary ; hence the nitrogen 
 in the pyridine nucleus must be unreduced, and the other nitrogen, which carries 
 the methyl, must be attached to two other carbon atoms as well. This condition 
 is most easily fulfilled if we suppose that this second nitrogen joins up the four 
 carbon atoms of the side chain into a pyrrol nucleus, which would give us 
 for the formula of nicotine : — 
 
 H 
 
 h/\ 
 
 H, 
 
 H, 
 
 H 
 
 V 
 
 H 
 
 ?\-nt/ 
 
 .^z 
 
 H, 
 
 Vt/ 
 
 H CH. 
 
 This is confirmed by the fact that if nicotine is gently oxidized by potassium 
 ferricyanide or silver oxide, it is converted into a new base, nicotyrine, with the 
 loss of four hydrogen atoms. This new body must be the corresponding pyridyl- 
 methyl-pyrrol : — 
 
 H 
 
 H 
 
 H 
 
 H CH, 
 
 ^N^ 
 
 The final confirmation of these views was furnished by the synthesis of 
 nicotine, which was carried out by Am6 Pictet. 
 
Nicotine 
 
 399 
 
 The first stage was the synthesis of nicotyrine. When /3-amino-pyridine is 
 heated with pyromucic acid it condenses with it to form N-/3-pyridyl-pyrrol :— 
 
 /V 
 
 ■NH, + 
 
 \n/ 
 
 COOH 
 
 ^CH=CH + CO2 + H2O. 
 
 ^n/ 
 
 This body when passed through a tube at a low red heat undergoes the same 
 change as the corresponding acetyl- and phenyl-pyrrols, the substituent going 
 from the nitrogen to the /3-carbon atom of the pyrrol ring, and /3-pyridyl-a-pyrrol 
 is produced : — 
 
 CH— CH 
 
 .CH=CH 
 
 ^CH=CH 
 
 \^/ 
 
 CH 
 ^H ■ 
 
 \^/ 
 
 On treatment of the product with methyl iodide the imine hydrogen of the 
 pyrrol is replaced by methyl, while the pyridine nitrogen takes up methyl and 
 iodine, becoming pentad. The resulting compound, 
 
 OH— CH 
 
 /\ II II 
 ^ N-C CH 
 
 / \ 
 CH, I 
 
 N-CH» 
 
 is identical with the iodo-methylate of nicotyrine, and is converted into nico- 
 tyrine on treatment with potash. This settles the structure of nicotjrrine, and 
 the only stage left was to reduce it to nicotine. For this purpose it was neces- 
 sary to reduce the pyrrol ring while leaving the pyridine ring unaffected. This 
 was found to be extremely difficult, as both rings behaved alike ; a weak reducing 
 agent reduced neither, and a strong both. It was finally discovered that the 
 end could be attained by first converting the nicotyrine into a mono-iodo- 
 derivative, and then reducing this. In this way a dihydro-nicotyrine was 
 formed, with the two hydrogens on the pyrrol ring ; and this, when treated 
 with bromine and again reduced with tin and hydrochloric acid, gave a base 
 identical with nicotine in composition and in all its properties with the exception 
 of the optical activity. After resolution by means of rf-tartaric acid, the laevo- 
 base was obtained, which agreed in eveiy respect with the natural nicotine, 
 whose formula was thus proved to be : — 
 
 CH2-CH2 
 /\-CH CH2. 
 
 \/ CH3 
 
400 
 
 Quinolifie 
 
 II. QU INCLINE 
 
 By the fusion of a pyridine with a benzene nucleus we arrive at benzopyri- 
 dine, which bears the same relation to pyridine that naphthalene does to benzene, 
 or indol to pyrrol. Owing to the presence of the nitrogen, this may lead to the 
 formation of either of two compounds, according as the two carbon atoms 
 common to both rings occupy on the pyridine ring the a,/3- or the ^,y-positions. 
 In the former case we get quinoline, in the latter isoquinoline : — 
 
 x/^N^ Vx/ 
 
 Quinoline. Isoquinoline. 
 
 The formula of quinoline was first established by its giiring on oxidation a 
 pyridine dicarboxylic acid (quinolinJc acid), and by its synthesis from allyl- 
 aniline by passing its vapour over heated lead oxide (the first synthesis of 
 quinoline, KOnigs, 1879): — 
 
 CH 
 
 C6H4^ 
 
 + 02= CgHi CH + ^ ^^O- 
 
 NH 
 
 This does not show where the new ring attaches itself to the benzene nucleus, 
 but it shows that the nitrogen is directly attached to the benzene, and so, taken 
 in conjunction with the previous evidence of the existence of a pyridine ring in 
 quinoline, is a sufiicient proof of the formula. 
 
 A further proof is given by Baeyer's synthesis from o-nitro-hydrocinnamic 
 acid, discovered in 1879, the same year as KOnigs'. This body on reduction 
 gives o-amino-hydrocinnamic acid, which however is unstable, and at once 
 forms its lactame hydrocarbostyril, the ketone of dihydroquinoline. This re- 
 action is exactly parallel to the formation of oxindol from o-amino-phenyl-acetic 
 acid ; indeed, it was in connexion with his work on the indigo compounds that 
 Baeyer discovered it : — 
 
 CHg 
 
 \COOH 
 
 /\ 
 
 CO 
 
 /\/CH,^ 
 
 NH 
 
 Oxindol. 
 
 \/\ 
 
 COOH 
 
 /VCH,. 
 
 NH, 
 
 \y 
 
 CO 
 
 NH 
 
 Hydrocarbostyril. 
 
 On treatment with phosphorus pentachloride and hydriodic acid hydrocarbo- 
 styril loses one oxygen atom and four hydrogens and gives quinoline. 
 
 The usual types of formulae have been proposed to express the internal 
 structure of pjnidine ; the most satisfactory is the analogue of Thiele's naphtha- 
 lene formula : — 
 
 ./--W 
 
Quinoline Syntheses 
 
 401 
 
 The substitution-products of quinoline are naturally very numerous. As the 
 formula shows, all the seven hydrogen atoms are differently related to the 
 nucleus, and thus seven monosubstitution-produots are possible. The number- 
 ing of the ring is confusing, as no less than three systems are in use :— 
 
 /3 
 
 Richter. 
 
 4 3 
 
 Bz- Py- 
 
 Beilstein. 
 
 m 
 
 /A 
 
 
 
 a = ana. 
 
 A very large number of syntheses of quinoline and its derivatives are known. 
 
 They may be obtained by intramolecular condensation from o-amino-deri- 
 vatives of benzene having a side chain of at least three carbon atoms, and 
 an oxygen atom on the third carbon ; e. g. from o-amino-cinnamic aldehyde : — 
 
 C^H _ 
 CHO 
 
 or from o-amino-cinnamyl-methyl ketone : — 
 
 CH 
 
 \/\^/ 
 
 CH 
 
 \/ NH, 
 
 \/\J 
 
 H 
 
 CCH, 
 
 Quinaldine 
 (a-methyl-quinoline). 
 
 A similar reaction is Friedlander's synthesis from o-amino-benzaldehyde and 
 any compound containing a CHg-CO group, such as aldehydes, ketones, aceto- 
 acetic ester, &c. This occurs in the presence of soda, and depends on the 
 intermediate formation of a benzylidine compound ; e. g. with acetoacetic ester : — 
 
 /\/CHO 
 
 \/^NH, 
 
 ?^ 
 
 H^CO^R ^ C-CO,R 
 
 0C-CH3 ^11 C-CHa " 
 
 a-methyl-quinoUne- 
 /3-carboxylic ester. 
 
 By substituting anthranilic acid for o-amino-benzaldehyde we get y-oxy-quino- 
 lines; e.g. with acetaldehyde, y-oxy-quinoline itself: — 
 
 OH 
 
 CH3 
 
 COH 
 CH 
 
 The best known synthesis of quinoline is Skraup's. It consists in heating a 
 primary aromatic amine with concentrated sulphuric acid, glycerine, and an 
 
402 
 
 Quinoline 
 
 oxidizing agent. The latter is usually the nitro-compound corresponding to the 
 amine employed, which is reduced in the reaction to that amine. Thus, quinoline 
 itself is got from aniline, nitrobenzene, glycerine, and sulphuric acid. We may 
 assume that the glycerine is converted by the sulphuric acid into acrolein, which 
 combines with the aniline to give acrolein-aniline ; and then the nitrobenzene 
 oxidizes oif the terminal hydrogen of the side chain together with the ortho- 
 hydrogen of the ring : — 
 
 OH, 
 
 .H 
 
 \/^NH, 
 
 CH. 
 
 2\ 
 
 CHO 
 
 /\^\ 
 
 n/ 
 
 CH 
 CH 
 
 H 
 
 
 There is some reason, however, to think ' that the reaction in its last stage is 
 not quite so simple as this. If crotonic aldehyde is condensed in this way with 
 aniline, we should expect it to give y-methyl-quinoline : — 
 
 /CH3 
 
 \rnT Y ^cjj 
 
 \/\n^ 
 
 CH 
 CH 
 
 \/\N^ 
 
 CH 
 
 but as a fact the product is a-methyl-quinoline. This can be explained if we 
 suppose that the aldehyde condenses with two molecules of aniline, one of which 
 splits off again in the closing of the ring : — 
 
 0NH2 + CH3CH=CH-CH0 
 
 /\ 
 
 ?^ 
 
 H, 
 
 NH 
 
 /\ 
 
 CHO 
 
 ^CH 
 
 ^^ NH 
 
 2 
 CH-CHo 
 
 > 1 ?^ . 
 
 The Skraup reaction is very violent, and requii-es great care ; the purity of 
 the glycerine seems to be of importance. It has been found ^ that if arsenic 
 acid is used as the oxidizing agent instead of nitrobenzene, the reaction proceeds 
 quite quietly. 
 
 In the place of aniline, any other aromatic amine having at least one 
 ortho position free can be used. Diamines react twice, giving the so-called 
 phenanthrolines ; e. g. : — 
 
 H^Nw^/NH, 
 
 \/ 
 
 /^\/\/ 
 
 \y\/ 
 
 > Blaise, Maire, C. R. 144. 33 (C. 07. i. 974) ; Ball. Soc. [4] 3. 667 (C. 08. ii. 174) ; cf. Simon. 
 C. R. 144. 138 {0. 07. i. 973). » Knfippel, Ser. 29. 704 (1896), 
 
Quinoline Syntheses 
 
 403 
 
 Another general method of great importance for the synthesis of quinoline 
 compounds is that of Dobner and v. Miller, the action of sulphuric or hydro- 
 chloric acid on a mixture of an amine and an aldehyde. The reaction proceeds 
 in three stages : (1) the formation of alkylidene-aniline ; (2) the formation of a 
 bimolecular condensation-product of this ; (3) the loss of aniline and hydrogen 
 to give a quinoline compound, for example with acetaldehyde : — 
 
 0-N=CH.CH3 
 + CH3-CH=N-^ 
 
 0-N=CH 
 
 vHo 
 
 ?■ 
 
 NH 
 
 CHCH, 
 
 
 \/\/ 
 
 CCHg 
 
 + ^■'Rllr, + H2 
 
 The hydrogen which is set free in this reaction sometimes reduces part of the 
 quinoline compound formed to a tetrahydro-derivative. 
 
 Quinoline itself is usually obtained by Skraup's synthesis ; it may also be 
 prepared by an extremely simple method due to Kulisch, by warming a mixture 
 of o-toluidine and glyoxal with soda : — 
 
 + 
 
 0=0H 
 0=CH 
 
 /\/CH^ 
 I CH 
 
 \/\^/ 
 
 CH 
 
 2H2O. 
 
 Quinoline is a colourless oil, boiling at 239°. It is very hygroscopic, forming 
 a hydrate ( + 1| HjO) in moist air. It closely resembles pyridine in behaviour. 
 It is a tertiary base, and forms quaternary quinolinium iodides with the alkyl 
 iodides. The corresponding hydroxides are strong bases; but, as with the 
 pyiidinium bases, the oxygen very readily migrates from the nitrogen to 
 the carbon. 
 
 Of the homologues of quinoline, the a- and y-methyl compounds are remark- 
 able for the mobility of the methyl hydrogen. (The same is the case with the 
 picolines.) Thus they condense with aldehydes, either to give compounds of 
 the aldol type, e. g. with chloral, 
 
 ■CHa-CHOH-CCIa ; 
 or with loss of water to form alkylidene derivatives, e.g. with benzaldehyde, 
 
 ■CH=CH-^. 
 
 Since the benzene nucleus is more easily substituted than the pyridine, the 
 direct substitution-products of quinoline (for example, with the halogens, sul- 
 phuric or nitric acid ') contain the substituent in the benzene ring. Those which 
 are substituted in the pyridine ring must be made by indirect methods, as by the 
 
 ' Kaufmann, Hiisey, Ber. 41. 1735 (1908). 
 
404 
 
 Quinoline 
 
 action of phosphorus pentachloride on the oxy-quinolines. These latter resemble 
 the substitution-products of pyridine itself. Thus the a- and y-chloro-quinalines 
 have the chlorine very mobile and easily replaced. 
 
 Quinoline is easily reduced, the hydrogen atoms going almost exclugiV'ely to 
 the pyridine ring, as we should expect, since pyridine is more easily reduced than 
 benzene. We thus get first di- and then tetrahydro-quinoline, 
 
 H, 
 
 
 H 
 
 \/\n^ \An/^^ 
 
 H 
 
 H 
 
 which latter behaves as far as the pyridine ring is concerned as a secondary 
 fatty amine, like piperidine: while the benzene ring has the full aromatic 
 character. Precisely the same phenomena occur of course in the case of 
 naphthalene. 
 
 The oxy-quinolines behave at once as acids and as bases. The o- and y- may 
 be obtained from the chlorine compound with potash. The /3- is unknown. 
 They react both as phenols and as ketones. The Bz-oxy-derivatives are pre- 
 pared by modifications of the syntheses ; e. g. from glycerine and aminophenol, 
 with nitrophenol as the oxidizing agent. The most important of the oxy- 
 quinolines is the a-, which is carbostyril, 
 
 \ Jv COH 
 
 .CH; 
 
 \ 
 
 
 CH 
 
 ho' 
 
 whose formation by the reduction of o-nitro-cinnamio acid has already been 
 mentioned. 
 
 It forms salts both with acids and with bases, without opening the ring, but 
 they are very unstable and are decomposed by water. It forms both O- and 
 N-ethers. 
 
 The nitrogenous ring of quinoline can be broken by various reagents ; thus, 
 benzoyl chloride and soda convert it into o-benzoylamino-cinnamic aldehyde ' : — 
 
 "H . ^„^„, . „^ f T CH 
 
 \/\J 
 
 CH 
 
 H^O = 
 
 CHO 
 
 + HCl. 
 
 + 0-CO-Cl -I- 
 
 ^'^NHCO-^ 
 
 It is to be observed that isoquinoline is not decomposed in this reaction. 
 
 In most cases this breaking of the ring is preceded by reduction. Thus if 
 the vapour of quinoline mixed with hydrogen is passed over reduced nickel at 
 260-280°, «-methyl-indol is formed, together with small quantities of methyl-o- 
 toluidine and toluidine.^ This is the unusual case of the conversion of a 6-ring 
 
 ' Eeissert, Ber. 38. 3415 (1905). 
 
 Padoa, Carughi, C. 06. il. 1011. 
 
Quinoline Syntheses 405 
 
 into a 5-riiig. It is probably due to the intermediate formation of an open- 
 chain body : — 
 
 CH 
 
 ? 
 ^^J 
 
 !H 
 
 CH 
 
 /CH, 
 
 /\ 
 
 NH 
 
 ,CH„CH, 
 
 -CH 
 
 ^ NH 
 
 C-CH, 
 
 In the same way tetrahydroquinoline when heated with nickel gives seatol 
 (^■methyl-indol) together with methyl- and ethyl-aniline, which are no doubt 
 intermediate products.^ 
 
 III. ISOQUINOLINE 
 
 Isoquinoline is isomeric with quinoline, the two carbon atoms common to 
 both rings being in the ^,y-positions. The positions in the ring are usually 
 numbered thus: — 
 
 This group is far less numerous than that of quinoline, but it is interesting 
 from its relation to various natural alkaloids, such as papaverine and berberine. 
 
 Isoquinoline can be synthesized in various ways. If the ammonium salt of 
 homophthalic acid is heated, it is converted into its imide, which is diketo- 
 tetrahydro-isoquinoline. This, on treatment with phosphorus pentachloride, 
 gives dichloro-isoquinoline, which on reduction yields isoquinoline itself. 
 
 CH 
 
 /\- 
 
 CH2COOH 
 COOH 
 
 Homophthalic 
 acid. 
 
 CH, 
 
 I 
 NH 
 
 
 
 Homophthal- 
 imide. 
 
 C-Cl 
 
 \^/ 
 
 CCl 
 Dichloro- 
 isoquinoline. 
 
 CH 
 
 CH 
 
 Isoquinoline. 
 
 The reaction can also be carried out in one operation, by heating homophthaUmide 
 with zinc dust in a stream of hydrogen. 
 
 Another general method of obtaining isoquinoline derivatives is to start from 
 benzene derivatives containing the side chain C-N-C-C ; thus from benzylidene- 
 amino-acetal (the condensation-product of benzaldehyde and the acetal of acetal- 
 dehyde) by heating with sulphuric acid : — 
 
 (EO),CH CH 
 
 (EO),C^H /\ ^CH„ /\/Y„ 
 
 \CH, _ I i ■■ _ V^ + 2 EOH. 
 
 ■^CHO 
 
 H^N 
 
 1175 
 
 CH 
 
 Padoa, Scagliarini, C 08. ii. 614. 
 
 Dd 
 
 N 
 
 CH 
 
406 
 
 Isoquinoline 
 
 A still simpler synthesis of the same kind is from benzylidene-ethylamine, 
 by passing it through a red-hot tube : — 
 
 CH 
 
 /V CH3 
 
 N 
 CH 
 
 /\/\ 
 
 CH 
 
 f..H. 
 
 A more unusual method was discovered by Gabriel and Colman,' who treated 
 the ester of phthalyl-glycine with sodium ethylate, whefeby the carboxylic ester 
 of oxy-isocarbostyril was formed :— 
 
 >•"="■ COH 
 
 A/'""'" /Vccooa 
 
 NCHg-COOEt -* CHj-COOEt -♦ | 
 
 X/^CO' \/\/NH XA^^ 
 
 CO COH 
 
 Another remarkable synthesis is by the distillation with phosphorus pentoxide 
 of the oxime of einnamic aldehyde ; we should expect this to give quinoline : — 
 
 CH 
 /X/X 
 
 X/ 
 
 ;CH 
 
 CH 
 /X/\ 
 
 NOH 
 
 H 
 
 
 As a fact, however, quinoline is not formed, but isoquinoline is, though only in 
 small quantity. This can be explained by supposing that the oxime (as an 
 anti-aldoxime) first undergoes the Beckmann reaction, and then the product 
 condenses : — 
 
 0-CH-CH-CH HOCH 
 
 HON 
 CH 
 
 HOCH 
 
 II 
 0.CH=CH-N 
 
 CH 
 
 /\/^. 
 
 X, 
 
 CH 
 
 N 
 /- 
 CH 
 
 The unreduced isoquinoline compounds are tertiary bases, giving stable salts 
 with acids and quaternary compounds with alkyl iodides. The stability of the 
 pyridine ring is less in isoquinoline than in pyridine itself and in quinoline, as 
 is shown by the fact that when isoquinoline is oxidized in acid solution both the 
 benzene and the pyridine rings are attacked, a mixture of phthalic and cincho- 
 meronic acids being produced : — 
 
 /\-COOH /\/\ CO,h/\ 
 
 X/ 
 
 COOH 
 
 N^\/ 
 
 CO,H.^ 
 
 1 Ber. 33. 980 (1900) ; cf. Findsklee, Ber. 38. 3542 (1905). 
 
Isoquinoline : Properties 407 
 
 If the nitrogen is made to assume the pentad condition, the pyridine ring is 
 still further weakened ; so that if the halogen alkylates (isoquinolinium salts) 
 are oxidized, the pyridine ring is broken and not the benzene ring. This reaction 
 takes a peculiar course. One carbon is eliminated from the ring, giving an 
 N-substituted phthalimide : — 
 
 Xvco 
 
 Isoquinoline occurs among the products of the distillation of coal tar, from 
 which it is separated along with the quinoline. It is isolated by dissolving the 
 raw quinoline in alcohol and adding sulphuric acid, when the less soluble 
 isoquinoline sulphate crystallizes out. 
 
 It boils at 240-5° (quinoline at 239°). It is more basic than quinoline, and 
 attracts carbon dioxide from the air. 
 
 The derivatives of isoquinoline closely resemble those of quinoline. 
 
 Dd2 
 
INDEX 
 
 The more important references are printed in thick type. 
 
 Acetanilide, 86. 
 
 Aeetic anhydride, 92, 154. 
 
 Aceto-acetic ester, 169, 254. 
 
 Acetonyl-acetone, 353. 
 
 Acetyl chloride, 114. 
 
 Acetyl nitrate, 12. 
 
 Acetylation, 86, 244. 
 
 Acid anhydrides, 103. 
 
 Acid hydrazides, 246. 
 
 Acidic groups, 19, 44, 46, 51, 82, 88, 89, 91, 
 
 104, 107, 129, 131, 140, 141, 151, 169, 183, 
 
 217, 248, 295, 304, 363, 387. 
 Acidic methylene group, 8, 9, 107, 140, 
 
 141, 169, 196, 250, 254, 316, 848. 
 Adenine, 323. 
 Adjective dye, 282. 
 Aldehyde ammonias, 32. 
 Alkaline stannous solution, 161. 
 Alkaloids, 392. 
 Alkyl pyrrols, 360. 
 Alkylamines, hydrates of, 19. 
 Alkylamines, properties of, 18, 20. 
 Alkylamines, substituted, 31. 
 Alkylation (of amines), 49. 
 Alkylene diamines, 68. 
 Allophanic acid, 183. 
 Allophanic ester, 180. 
 Alloxan, 316. 
 Allyl isothiocyanate, 223. 
 Aluminium chloride, 107, 180, 199, 226,261. 
 Amic acids, 84. 
 Amides, 76. 
 Amides as solvents, 82. 
 Amides of dibasic acids, 84. 
 Amides, properties of, 79. 
 Amidines, 95. 
 Amido-chlorides, 92. 
 Amidolysis, 82. 
 Amidoximes, 105. 
 Amidoxyl-fatty acids, 298. 
 Amine group : reaction with nitrous acid, 
 
 335, 344. 
 Amine hydrates, 19, 69. 
 Amine oxides, 101. 
 Amine oxide, active, 30. 
 Amines, 15. 
 Amines, aromatic, 45. 
 Amines, basicity of, 24. 
 Amines, mixed, 48. 
 Amines, mixed tertiary, 49. 
 
 1175 
 
 Amines, oxidation of, 162. 
 
 Amines, separation of, 19. 
 
 Amines, solubility of, 27. 
 
 Amino-aoids, 34. 
 
 Amino-acids of proteins, 40. 
 
 Amino-alcohols, 32. 
 
 Aminoazo-compounds, 284. 
 
 Amino-diphenyl-methanes, 59. 
 
 Amino-guanidine, 300. 
 
 Amino-ketones, 34. 
 
 Aminophenols, 56. 
 
 Aminopyrrols, 362. 
 
 Amino-triphenyl-methanes, 59, 124. 
 
 Amino-urea, 299. 
 
 Amino-urethane, 299. 
 
 Ammonia reaction, 296. 
 
 Amphoteric electrolytes, 107. 
 
 Amygdalin, 197. 
 
 Amyl nitrite, 248, 257. 
 
 Anhydrides, acid, 103. 
 
 Anhydro-bases, 70. 
 
 Anhydro-formaldehyde aniline, 49. 
 
 Anilides, 86. 
 
 Aniline, 45, 46. 
 
 Aniline derivatives substituted in the 
 nucleus, 51. 
 
 Aniline, halogen derivatives of, 54. 
 
 Aniline, sulphonic acids of, 55. 
 
 Anilines, substituted, 48. 
 
 Aromatic amines, 45. 
 
 Aromatic character, 183. 
 
 Aromatic and fatty compounds, distinc- 
 tions between, 71. 
 
 Aromatic nitro-compounds, 153. 
 
 Aromatic poly-nitro-derivatives, 164, 
 
 Artificial musk, 155. 
 
 Aspartic acid, 85. 
 
 Auramines, 59. 
 
 Autoxidation, 98. 
 
 Auxochrome, 172, 283. 
 
 Azides, 345, 346. 
 
 Azido-acetic acid, 344. 
 
 Azimides, 71. 
 
 Azines, 71. 
 Azobenzene, 281. 
 
 Azo-compounds, 280. 
 Azo-dicarboxylic acid, 301. 
 Azo-phenols, 287, 291. 
 
 Azo-pyrrols, 362. 
 Azoxy-compounds, 292. 
 Azulmic acid, 196, 199. 
 
410 
 
 Index 
 
 B 
 
 Baeyer's strain theory, 70, 71, 84, 90, 218, 
 
 331,333,406. 
 Barbituric acid, 816. 
 Basic gi-oups, 18, 50, 54, 55, 57,71, 73, 80, 
 
 87, 101, 105, 107, 133, 185, 191, 193, 241, 
 
 243, 265, 299, 307, 313, 357, 386, 407. 
 Basicity of amines, 24. 
 Beckmann reaction, 82, 89, 104, 105, 114, 
 
 119. 213, ai3. 
 Benzaldoximes, 118. 
 Benzamide, tautomeric salts of, 81. 
 Benzene sulphonic chloride, 19, 120. 
 Benzidine derivatives, 58. 
 Benzil, oximes of, 118. 
 Benznitrosolic acid, 133. 
 Benzo-aceto-dinitrile, 385. 
 Benzoin synthesis, 200. 
 Benzoyl nitrate, 12. 
 Benzylamine bases, 43. 
 Biguanide, 193. 
 Bis-diazoacetic acid, 342. 
 Bis-diazoamino-compounds, 314. 
 Bis-diazomethane, 343. 
 Biuret, 183, 188. 
 Biuret test, 48, 188. 
 Bromamides, 82. 
 Bijlow reaction, 246, 248, 313. 
 Buzane derivatives, 313. 
 
 Caffeine, 824, 325. 
 
 Carbamic acid, 179. 
 
 Carbamio chloride, 180. 
 
 Carbamide-imide-azide, 311. 
 
 Carbo-diimide, 192. 
 
 Carbo-diphenyl-imide, 192. 
 
 Carbon bisulphide, 47, 183. 
 
 Carbonic acid derivatives, 177, 299. 
 
 Carbylamines, 207. 
 
 Caro's acid, 102, 103, 122, 162. 
 
 Catalysis, 22, 45, 46, 48, 86, 94, 137, 182, 
 
 191, 200, 214, 261, 293, 838, 386. 
 Catalysis, negative, 123, 199. 
 Catalytic action of acids, 52, 125, 133, 154, 
 
 295, 296, 306, 807. 
 Catalytic action of hydrogen ion, 76, 77, 
 
 106,337. 
 Catalytic action of hydroxyl ion, 77, 106, 
 
 158. 
 Catalytic action of light, 52, 70, 123, 138, 
 
 146, 276, 838. 
 Catalytic action of mercury, 878. 
 Catalytic action of neutral salts, 97, 156. 
 Celluloid, 10. 
 Chloramines, 21, 52. 
 Choline, 32. 
 
 Chromo-ethers, 170, 178. 
 Chromophor, 172, 283. 
 Colloidal platinum, 806. 
 Collodion, 10. 
 Colour and constitution, 60, 64, 67, 73, 110, 
 
 123, 145, 146, 149, 166, 170, 171, 174, 245, 
 
 286, 299, 306. 
 
 Colour and ionization, 171, 174. 
 Condensation products of ortho-derivatives, 
 
 70. 
 Coniine, 892. 
 
 Coniine, syntheses of, 394. 
 Conyrine, 394. 
 Copper cyanide, 196. 
 'Coupling' reactions, 262, 287. 
 Creatin, 194. 
 Creatinin, 194. 
 
 Crum Brown's rule, 51, 53, 378. 
 'Crumpling' of ring, 848, 344, 404, 407. 
 Cuprous chloride, 185. 
 Curtius reaction, 18. 
 Cyamelide, 234. 
 Cyanalkines, 206. 
 Cyanamide, 216. 
 Cyanic acid, 214. 
 Cyanic esters, 217. 
 Cyanides, double, 201. 
 Cyan-isonitroso-acet-hydroxamic acid, 229. 
 Cyanogen, 195. 
 Cyanogen compounds, 195. 
 Cyanogen halides, 215. 
 Cyan-urea, 188. 
 Cyanuric acid, 233. 
 Cyanuric bromide, 283. 
 Cyanuric chloride, 233. 
 Cyanuric esters, 233. 
 Cyclic diamides, 85. 
 Cyclic ureides, 315. 
 
 D 
 
 Di-acetamide, 83. 
 
 Bi-acetyl-ortho-nitric acid, 12. 
 
 Di-acyl-anilines, 87. 
 
 Dial^l-hydx-oxylamines, 101. 
 
 Dialuric acid, 316. 
 
 Diamides, cyclic, 85. 
 
 Diamines, 68. 
 
 Diamines, aromatic, 70. 
 
 Diazoacetic ester, 385. 
 
 Diazoaoetic ester, polymerization of, 341. 
 
 Diazoamino-compounds, 305. 
 
 Diazoamino-compounds, structure of, 309. 
 
 Diazoamino-cyanides, 311. 
 
 Diazoamino-paraffins, 306. 
 
 Diazobenzenic acid, 295. 
 
 Diazo-compounds, 256. 
 
 Diazo-compounds, constitution of, 262, 278. 
 
 Diazo-compounds, formation of, 257. 
 
 Diazo-cyanides, 273. 
 
 Diazo-ethane, 334. 
 
 ' Diazo-guanidine,' 310. 
 
 Diazo-hydrates, 270. 
 
 Diazo-hydrazides, 313. 
 
 Diazomethane, 332. 
 
 Diazomethane disulphonic acid, 344. 
 
 Diazomethane group, 331. 
 
 Diazo-reactions, 260. 
 
 Diazo-reactions, theory of, 275. 
 
 Diazonium perhalides, 265. 
 
 Diazonium salts, 265. 
 
 Diazonium salts, solid, 272. 
 
Index 
 
 411 
 
 Diazotates, 267. 
 
 Dicyan-diamide, 217. 
 
 Dihydro-tetrazine, 342. 
 
 Diimide, 301. 
 
 Diketones, oximes of, 109. 
 
 Dilituric acid, 317. 
 
 Dimethyl-aniline, 50. 
 
 Dimetliyl-triaaene, 306. 
 
 Dinitro-paraffins, salts of, 149. 
 
 'Dinitrure', 128. 
 
 Di-ortho-dinitro-diphenyl-diacetylene, 367. 
 
 Dioxindol, 366, 371. 
 
 Diphenyl-carbo-diimide, 192. 
 
 Diplienyl-furoxane, 230. 
 
 Diphenyl-glyoxime peroxide, 134. 
 
 Diphenylamine, 50. 
 
 Diphenylamine test for nitric acid, 245. 
 
 Distillation in vacuo, 39. 
 
 Dithiocarbamic acid, 182. 
 
 Diurea, 187. 
 
 Dyad carbon, 208, 211, 224. 
 
 Dyeing, 282. 
 
 Dyes, 59, 282. 
 
 Dynamite, 11. 
 
 E 
 
 Electrolytic reduction, 156. 
 
 Electro-synthesis, 69. 
 
 Ethanolamine, 32. 
 
 Ethionic acid, 33. 
 
 Ethyl hydroperoxide, 7, 9. 
 
 Ethyl oxalate, 19. 
 
 Ethylene imine, 331. 
 
 Exhaustive methylation, 364, 392. 
 
 Expansion of rings, 359. 
 
 P 
 
 Ferments, natural, 81. 
 Ferric chloride, 143, 183. 
 Ferricyanides, 201. 
 Ferrocyanides, 201. 
 Ferrofulminates, 226. 
 Formaldehyde, 49, 102. 
 Fonnaldoxime, 109. 
 Formamidoxime, 106. 
 Form-hydroxamic acid, 104. 
 Formic acid as solvent, 168. 
 Formyl chloride oxime, 104, 224. 
 Friedel and Craft's reaction, 180. 
 Fulminic acid, 223. 
 Fulminic acid, polymers of, 228. 
 Fulminuric acid, 228. 
 Purazane derivatives, 130. 
 
 G 
 
 Gabriel's phthalimide reaction, 16, 32, 34, 
 
 35, 69. 
 Gatterman's reaction, 199, 261. 
 Glucosamine, 36. 
 Glucosides, 197- 
 Glycine, 37. 
 GlycocoU, 37. 
 
 Glyoxime peroxide, 130. 
 Guanidine, 193. 
 Guanine, 323, 326. 
 Gun cotton, 10. 
 
 H 
 
 'Halochromie,' 153. 
 
 Halogen carriers, 182, 386. 
 
 Heteroxanthine, 325. 
 
 Heumann's synthesis of indigo, 375. 
 
 Hinsberg reaction, 19, 331. 
 
 Hofmann's amide reaction, 17, 82, 187, 204, 
 
 377, 379. 
 Hofmann's amine reaction, 52, 57. 
 Hofmann's isonitrile reaction, 20. 
 Hofmann's mustard oil test, 20, 222. 
 Homolka base, 68, 74. 
 Hydrates of alkylamines, 19. 
 Hydrates of diamines, 69. 
 Hydrazi-acetic acid, 339. 
 Hydrazides, acid, 246. 
 Hydrazido-nitriles, 248. 
 Hydrazine derivatives, 241. 
 Hydrazines, aromatic, 242. 
 Hydrazines, fatty, 241. 
 Hydrazoamino-compounds, 311. 
 Hydrazo-compounds, 254. 
 Hydrazones, 247. 
 
 Hydrazones from diazo-compounds, 251. 
 Hydrocarbostyril, 400. 
 Hydrochloric acid as reducing agent, 22, 73, 
 
 285. 
 Hydrocyanic acid, 197. 
 Hydrogen peroxide, 7, 98, 101, 162. 
 Hydrotetrazones, 248, 313. 
 Hydroxamic acids, 103. 
 Hydroxamoxiraes, 106. 
 Hydroxy-anilines, 56. 
 Hydroxy-caffeine, 326. 
 Hydroxylamine, 101, 106, 253, 259. 
 Hydroxylamine derivatives, 06. 
 a-Hydroxylamines, 96. 
 /3-Hydroxylamines, 96. 
 Hypoxanthine, 323. 
 
 Imidazoles, 70. 
 Imides, 90. 
 Imido-chlorides, 92. 
 Imido-compounds, 16. 
 Imido-ethers, 93. 
 Imidohydrine, 92. 
 Imino-dicarboxylic acid, 183. 
 Indicators, 171. 
 Indigo, constitution of, 366. 
 Indigo, synthesis of, 374. 
 Indigo -white, 373. 
 Indol, 369. 
 Indol group, 364. 
 Indophenols, 133. 
 Indoxyl, 370. 
 Internal salts, 33, 36, 55. 
 Intramolecular change, 21, 38, 50, 51, o6, 
 60, 87, 91, 93, 97, 98, 103, 126, 144, 222, 
 
412 
 
 Index 
 
 231, 244, 253, 266, 273, 292, 295, 296, 297, 
 300, 301, 313, 360, 388, 393, 395. 
 
 Ionization and colour, 171, 174. 
 
 Ionization-isomerism, 267. 
 
 Isatin, 366, 371. 
 
 Iso-amides, 80. 
 
 Isocyanic esters, 218. 
 
 Isocyanides, 207. 
 
 Isocyanides, constitution of, 208. 
 
 Isocyanphenyl-chloride, 208. 
 
 Isocyanuric acid, 228. 
 
 'Iso-diazo-compounds,' 342. 
 
 IsoMminuric acid, 229. 
 
 Isomerism, 223. 
 
 Isonitramines, 297. 
 
 Isonitrile reaction, Hofmann's, 20. 
 
 Isonitriles, 207. 
 
 Isonitroso-compounds, 106. 
 
 Isoquinoline, 405. 
 
 Isothiocyan-acetio acid, 222. 
 
 Isothiocyanic-esters, 222. 
 
 Iso-urea, 188. 
 
 Isuretin, 106. 
 
 Lecithins, 33. 
 
 Leuco-bases, 60. 
 
 Leuco-ethers, 170. 
 
 Liebermann reaction, 295. 
 
 Light, catalytic action of, 52, 70, 123, 133, 
 
 146, 276, 338. 
 Liquid crystals, 293. 
 
 M 
 
 Magnesium alkyl halide, 5, 101, 144. 
 
 Malonyl-urea, 316. 
 
 Melamines, 236. 
 
 Melinite, 177. 
 
 Melting-point and constitution, 69. 
 
 Mendius reaction, 17, 35, 205. 
 
 Mercuric cyanide, 201. 
 
 Mercury, catalytic action of, 378. 
 
 Mercury fulminate, 107, 223. 
 
 Mercury salt of nitroform, 151. 
 
 Mercury salts of tautomeric compounds, 
 
 234. 
 Meri-quinoid compounds, 74. 
 Metafulminuric acid, 228. 
 Meta-quinones, 175. 
 Methyl azide, 345. 
 Methyl cyanide as solvent, 168. 
 Methyl-nitrosamine, 181. 
 Methyl-nitrosolic acid, 184. 
 Methyl sulphate as alkylating agent, 16, 
 
 140. 
 Methylation, 333, 364, 392. 
 Methylene blue, 133. 
 Methylene group, acidic, 8, 9, 107, 140, 141, 
 
 196, 250, 254, 316, 348. 
 Michler's ketone, 59. 
 Migration, see Intramolecular change. 
 ' Mobility,' 94, 403. 
 Mordants, 282. 
 
 Musk, artificial, 155. 
 
 ' Mustard oil ' reaction, Hofmann's, 20, 222. 
 
 Mustard oils, 222. 
 
 N 
 
 Negative groups, 140, 188, 243, 272, 278, 
 
 335, 358. 
 Nicotine, 397. 
 Nicotyrine, 398. 
 Nitramines, 295. 
 Nitranilines, 55. 
 Nitrates, acyl, 12. 
 Nitration, 153. 
 
 Nitration of fatty compounds, 139. 
 Nitration of phenol, 169. 
 Nitric esters, 8. 
 Nitrile-oxides, 229. 
 Nitriles, 203. 
 Nitrimines, 297. 
 Nitrites, metallic, 140. 
 Nitrites, organic, 21, 37. 
 Nitro-celluloses, 10. 
 Nitro-compounds, aromatic, 153. 
 Nitro-compounds, fatty, 130. 
 Nitro-compounds, reduction of, 155. 
 Nitrocyan-acetamide, 228. 
 Nitro-derivatives of amides of carbonic acid, 
 
 303. 
 Nitro-diphenylamines, 178. 
 Nitro-explosives, 10. 
 Nitro-glycerine, 9. 
 Nitro-guanidine, 304. 
 Nitro-hydroquinone methyl ether, 177. 
 a-Nitro-ketones, 145. 
 Nitro-methane, 140. 
 Nitro-paraffins, 139. 
 Nitro-paraffins, structure, 141. 
 Nitro-phenols, 169. 
 Nitro-pyrrols, 362. 
 Nitro-urea, 804. 
 
 Nitro-urethane, 304. 
 Nitroform, 150. 
 
 Nitroform, mercury salt of, 151. 
 Nitrogen, oxides of, 6, 124,128. 
 Nitrogen, pentavalent, 23. 
 
 Nitrogen, pentavalent, stereochemistry of, 
 27. 
 
 Nitrogen tricarboxylic acid, 183. 
 
 Nitrolic acids, 146. 
 
 Nitronic acids, 165. 
 
 Nitrosamines, 270,293. 
 
 Nitrosates, 128. 
 
 Nitrosites. 128. 
 
 Nitrosoanilines, 58. 
 
 Nitrosobenzene, 182. 
 
 Nitrosobutane, 122. 
 
 Nitroso-compounds, aromatic, 131. 
 
 Nitroso-compounds, fatty, 122. 
 
 Nitroso-derivatives of amides of carbonic 
 acid, 302. 
 
 Nitroso-hydroxylamines, 297. 
 
 Nitrosolio acids, 183. 
 
 Nitroso-methyl-urethane, 181, 294, 332. 
 
 Nitroso-orcin, 136. 
 
Index 
 
 413 
 
 Nitrosophenols, 135. 
 
 Nitroso-pyrrols, 361. 
 
 Nitrosyl chloride, 182. 
 
 Nitrous acid, 20, 21, 36, 37, 47, 69, 71, 72, 
 
 107, 109, 133, 146, 253, 255, 259. 
 Nitrous acid, analogy with triphenyl 
 
 carbinol, 8. 
 Nitrous acid, estimation of, 47. 
 Nitrous acid, reaction with NHj group, 335, 
 
 o44. 
 Nitrous esters, 6. 
 Nitrous oxide, 105. 
 
 
 Octazones, 314. 
 
 Optically active nitrogen compounds, 28. 
 Optically active oximes, 116. 
 Organo-magnesium compounds, 305. 
 Ortho-condensation products, 70. 
 Ortho-esters, 95. 
 Ortho-, meta-, and para-derivatives, distinc- 
 
 tionsbetween, 72. 
 Osazones, 249. 
 Osotetrazones, 250. 
 Oxamic acid, 84. 
 Oxamide, 84. 
 
 Oxidation, 70, 72, 73, 97, 108. 
 Oxidation of amines, 162. 
 Oximes, 106. 
 Oximes of benzil, 118. 
 Oximes of diketones, 109. 
 Oximes, stereoisomerism of, 110. 
 Oximes, velocity of formation of, 106. 
 Oxindol, 366, 371. 
 Oxy-amidoximes, 106. 
 Oxyamines, 101. 
 Oxy-azo-compounds, 287. 
 Oxy-azo-compounds, constitution, 289. 
 
 Para-hydrogen atom, reactivity of, 49. 
 
 Paracyanogen, 197. 
 
 Para-urazine, 187. 
 
 Para-xanthine, 325. 
 
 Pentavalent nitrogen, 23. 
 
 Pentavalent nitrogen compounds, 74. 
 
 Pentavalent nitrogen compounds, stereo- 
 chemistry of, 27. 
 
 Peptones, 41. 
 
 Persulpho-cyanic acid, 221. 
 
 Phenyl-aoetamide, 83, 339. 
 
 Phenyl-azide, 345. 
 
 Phenyl-cyclo-triazane, 347. 
 
 Phenyl-diazo-me thane, 335. 
 
 Phenyl-hydrazones, 247. 
 
 0-Phenyl-hydroxylamine, 96. 
 
 Phenyl-isocyanate, 111, 143, 219. 
 
 Phenyl magnesium bromide, 132, 208. 
 
 Phenyl nitromethane, 143. 
 
 Phosphorus oxychloride ascondenbing agent , 
 90, 316. 
 
 Phosphorus pentachloride, 92, 114, 167, 168. 
 
 Phosphorus pentasulphide, 88, 89. 
 
 Phosphorus pentoxide, 81, 84. 
 
 Phosphorus, red, as reducing agent, 156. 
 Phthalimide reaction, Gabriel's, 16, 32, 34, 
 
 35,69. 
 Picric acid, 177. 
 Picryl chloride, 169. 
 Piperidine, 391. 
 Piperine, 396. 
 Piperylene, 393. 
 Polymerisation, 32, 109, 123, 126, 134, 197, 
 
 201, 205, 214, 218, 229, 236, 341. 
 Polymethylene imines, 351. 
 Poly-nitro-derivatives, aromatic, 164. 
 Poly-nitroparaffins, 148. 
 Polypeptides, 38, 42, 366. 
 Position of substituents, influence of, 54, 55, 
 
 56, 58, 278, 280. 
 Potassium nitrite, 148, 155. 
 
 Potassium permanganate, 155, 162, 163. 
 
 Potassium pyrrol, 360. 
 
 ' Protection ' of amines for nitration, 53. 
 
 Proteins, 38. 
 
 Prussian blue, 195. 
 
 Prussic acid, 197. 
 
 Prussia acid, constitution of, 209, 341. 
 
 Prussic acid, salts of, 201. 
 
 Pseudo-acids, 82, 144, 170. 
 
 Pseudo-diazo-acetic acid, 342. 
 
 Pseudo-nitrols, 147. 
 
 Pseudo-nitrosites, 128. 
 
 Pseudo-phenyl-acetic acid, 339. 
 
 Pseudo-salt, 152. 
 
 Pseudo-uric acid, 318. 
 
 Purine, 322. 
 
 Pyridane, 388. 
 
 Pyridine, 120, 124, 386. 
 
 Pyridine, amino-derivatives of, 390. 
 
 Pyridine as catalyst, 386. 
 
 Pyridine as condensing agent, 203, 386. 
 
 Pyridine group, 383. 
 
 Pyridine substitutions, 387. 
 
 Pyridine, synthesis of, 384. 
 
 Pyridinium derivatives, 388. 
 
 Pyrrol, 357. 
 
 Pyrrol carboxylic acids, 362. 
 
 Pyrrol group, 353. 
 
 Pyrrol, synthesis of, 353. 
 
 Pyrrolidine, 363. 
 
 Q 
 Quaternary ammonium compounds, 22. 
 Quaternary hydroxides, 22. 
 Quinhydrones, 74. 
 Quinoline, 400. 
 Quinoline, synthesis, 401. 
 Quinone-azine, 291. 
 Quinone diimines, 72. 
 Quinone imines, 72. 
 Quinone oximes, 135. 
 ' Quinoid' constitution, 73, 110. 
 Quinoxalines, 71. 
 
 R 
 
 Racemisation, velocity of, 29. 
 Raoult's cryoscopic method, 111. 
 
414 
 
 Index 
 
 Rate of reaction, see Velocity. 
 Reduction, 17, 37, 45, 311. 
 Reduction of nitro-compounds, 155. 
 Reduction with nickel in hydrogen, 17. 
 Reduction with red phosphorus, 156. 
 Reduction with sulphurous acid, 311. 
 Ring formation, 34, 36, 56, 69, 70, 86, 90, 
 
 350. 
 Rings, change of size, 21, 343, 344, 359, 404, 
 
 407. 
 Rosaniline dyes, constitution of, 61. 
 Rosaniline hydrocyanide, 64. 
 
 S 
 
 Salicylamide, 57. 
 
 Sandmeyer's diazo-reaction, 260, 261. 
 
 Sandmeyer's synthesis of indigo, 380. 
 
 Schiff's test for aldehydes, 67. 
 
 SohoU's reaction, 107. 
 
 Schotten-Baumann reaction, 69. 
 
 Semicarhazide, 299. 
 
 Serine, 36. 
 
 Silver cyanate, 16, 214. 
 
 Silver nitrite, 140, 229. 
 
 Slow neutralization, 149. 
 
 Smell, 155, 198, 211, 226. 
 
 Sodium ethylate as condensing agent, 8, 9, 
 145, 196. 
 
 Sodium hypobromite, 187. 
 
 Sodium methylate as reducing agent, 156. 
 
 Sodium nitrite, 105, 187. 
 
 Sodium peroxide, 155. 
 
 Solid solutions of tautomers, 173, 174, 272. 
 
 Solubility of amines, 27. 
 
 Solvents, influence on tautomeric equilib- 
 rium, 173. 
 
 Skraup's synthesis of quinoUne, 401. 
 
 Stability of rings, 406. 
 
 Stannous halide, 161. 
 
 Stannous solution, alkaline, 161. 
 
 Stereo-chemistry of pentavalent nitrogen, 
 27. 
 
 Stereo-hindrance, 50, 78, 86, 117, 132, 245, 
 260, 284, 289, 333, 351. 
 
 Stereo-isomerism, 105, 248, 307. 
 
 Stereo-isomerism of diazo-compounds, 264. 
 
 Stereo-isomerism of oximes, 110. 
 
 Strain, 70, 71, 84, 90, 218, 331, 333, 406. 
 
 Strecker reaction, 36. 
 
 Substantive dyes, 282. 
 
 Substituents, influence of position, 54, 55, 
 56, 58, 278, 280. 
 
 Succinimide, 84. 
 
 Sulpho-carbamic acids, 20. 
 
 Sulphurous acid, 48. 
 
 Sulphurous acid as reducing agent, 311. 
 
 Syn- and anti-diazo compounds, 269. 
 
 T 
 Tartronyl-urea, 316. 
 Taste, 36, 118. 
 Taurine, 33. 
 Tautomeric equilibrium, 173, 189. 
 
 Tautomeric groups, 279. 
 
 Tautomeric salts, 80, 81, 87, 135, 234. 
 
 Tautomerism, 85, 97, 110, 115, 140, 251, 
 
 308, 348, 362. 
 Tautomerism, amide-imide, 80, 87, 88, 94, 
 
 103, 182, 189, 190, 233, 235, 279. 
 Tautomerism, cyanide-isocyanide, 197, 214. 
 Tautomerism, keto-enol, 91, 279, 347, 348, 
 
 370. 
 Tautomerism of nitro-compounds, 142, 144, 
 
 145, 147, 149, 151, 165, 170, 178, 296. 
 Tautomerism, nitroso-diazo, 270, 302. 
 Tautomerism, quinone-oxime, 135, 136. 
 Tautomers, separable, 66, 136, 143. 
 Temperature coefficient of velocity, 273. 
 ' Tertiary ' hydrogen, 139. 
 Tetra-aryl-hydrazines, 245. 
 Tetra-iodo-pyrrol, 361. 
 Tetramethyl-ammonium, 23. 
 Tetranitro-methane, 153. 
 Tetrazane derivatives, 313. 
 Tetrazenes, 313. 
 
 Tetrazine dicarboxylic acid, 343. 
 Tetrazones, 313. 
 Theobromine, 324. 
 Theophylline, 326. 
 Thiazol, 88, 89. 
 Thiazoline, 88. 
 Thiele's theory, 338, 356. 
 Thioacetic acid, 87. 
 Thioamides, 88. 
 Thioanilides, 89. 
 Thiocarbamic acid, 182. 
 Thiocarbanilide, 190. 
 Thiocyanacetone, 222. 
 Thiocyanic acid, 220, 221. 
 Thiocyanuric compounds, 235. 
 Thioformamide, 88. 
 Thionyl-phenyl-hydrazone, 246. 
 Thiosemicarbazide, 300. 
 Thio-ureas, 189. 
 
 Three-nitrogen compounds, 305. 
 Triazane derivatives, 311. 
 Triazene dicarboxylic acid, 311. 
 Triazenes, 308. 
 Triazine, 232. 
 ' Triazo-compounds,' 342. 
 Triazoles, 343. 
 Triazolones, 342. 
 Tribenzoyl-methane, 252. 
 Tricyanogen compounds, 231. 
 Triethyl-triazine, 232. 
 Trifulmines, 230. 
 Trimethylene imine, 350. 
 Trinitrobenzene, 165. 
 s-Trinitro-benzoic acid, 167, 178. 
 Triphenylamine, 51. 
 s-Triphenyl-guanidine, 194. 
 Triphenyl-methyl compounds, 8, 124, 152, 
 
 255 
 Tripy^ol, 359. 
 
 U 
 
 Uramil, 317. 
 Urea, 184. 
 
Index 
 
 415 
 
 Ureides, 315. 
 
 Ureides, cyclic, 315. 
 
 Uretbanes, 181. 
 
 Uric acid, 317. 
 
 Uric acid, constitution of, 318. 
 
 Uric acid group, 315. 
 
 Uric acid, syntheses of, 320. 
 
 Velocity of reaction, 7, 22, 29, 52, 56, 76 , 77, 
 86, 106, 108, 121, 154, 158, 160, 163, 184, 
 186, 188, 189, 190, 200, 202, 204, 215, 284, 
 247, 258, 278, 287, 307, 337. 
 
 Violuric acid, 317. 
 
 ' Virtual ' tautomerism, 309. 
 
 W 
 
 'Wandering,' 94,382. 
 Wurater's salts, 74. 
 
 Xanthine, 323, 326. 
 
 X 
 
 Zinc alkyl halide, 101, 144. 
 
 Zinc chloride as condensing agent, 199, 248. 
 
 Zinc ethyl, 142.