THE PREPARATION AND REACTIONS OF PARACHLOROPHENYLARSINE BY WALLACE HUME CAROTHERS B. S. Tarkio College 1920 THESIS Submitted in Partial Fulfillment of the requirements for the Degree of MASTER OF SCIENCE IN CHEMISTRY IN THE GRADUATE SCHOOL OF THE UNIVERSITY OF ILLINOIS 1921 Digitized by the Internet Archive in 2016 https://archive.org/details/preparationreactOOcaro UNIVERSITY OF ILLINOIS THE GRADUATE SCHOOL x r 192 % I HEREBY RECOMMEND THAT THE THESIS PREPARED UNDER MY SUPERVISION BY Wal lace H. Caro the re ENTITLED .The Pre paration a nd Reactions o f p-Ch lorophenylarsine BE ACCEPTED AS FULFILLING THIS PART OF THE REQUIREMENTS FOR THE DEGREE OF Master of Scienc e IaC J)- In Charge of Thesis Head of Department Recommendation concurred in* Committee on Final Examination* ^Required for doctor’s degree but not for master’s ...V, -• • • . . 'll ASIA OF COiS'i'iiil'Ji’S THJSORRTICAL AM) HISTORICAL I DESCRIPTIVE 4 SIGHIFICAHCE OF RESULTS 12 MPRRIMiSHTAL 16 Preparation & properties of p-ehlorophenylarsonic acid 16 Preparation and properties of p-chlorophenylarsine £0 Analysis of p-chlorophenylarsine-benjzaldehyde product £6 Analysis of p-chlorophenylarsine -paraldehyde product £7 Preparation of arsanilic acid and its reduction £b Preparation of p-phenoiarsonic acid and its reduction £1 Preparation of p- phene tylarsonic acid and its reduction £4 SUMMARY 35 ACKHOWnCilXJisiMJiRT 36 -- . X . . V » — - . r - ■w* . ... ~ - tit • - • J - * - J : , v. • ,x ^ ■>' ‘ ■' U - • . •- • o . 1 THE PREPARATION AND REACTIONS OF P AR A CH LOR OPIIENYL ARSINE. I. THEORETICAL AND HISTORICAL. Arsenic occurs in the periodic table in the fifth vertical column below phosphorous which stands just under nitrogen, a fnct which would lead to the prediction that arsenic and nitrogen would show some similarities. Physically they are, of course, quite dif- ferent, but chemically their properties bear out this prediction in many respects. Thus, both elements form stable hydrides in which the element is trivalent; both form stable oxides in which 1 2 the element is either tri- or pentavalent; both form halides \ Similar analogies exist in the formation of organic deriva- tives and in the chemical properties of these derivatives. The hydrides of both elements yield derivatives in which one hydrogen is replaced by an alkyl or aryl group, the primary amines being thus derived from ammonia, and the primary organic arsines from arsine. Amines were predicted to be capable of existence by 3 4 Liebig in 1842 , were first prepared by Wurtz , and since then have been extensively studied and have been found to be of enormous importance in organic chemistry, the aromatic amine compounds be- ing especially important in the synthesis of other compounds of many diverse types. The existence of primary organic arsines would of course, be predicted by analogy, but attempts to prepare compounds of this 5 type failed until A. W. Palmer succeeded in reducing methyl and 1 Wm. H. Dehn Amer. Chem. Journ. 33, 101 (1905) 2 Gilbert T. Morgan, Organic compounds of Arsenic and antimony, xi. 3 Handworterbuch, 1, 689. 4 Ann. Cham. (Liebig) 71, 330, 76, 318. 5 Ann. Chem., 107, 285. . ■ . .. * . ‘ 2 . 0 phenyl arsenic acids and isolating the corresponding arsines.' It then appeared that the failure of previous investigators had been due to the absence of precautions to protect the product from the air; primary arsines being oxidized instantaneously. Since then several other primary arsines, both aromatic and aliphatic 7 have been prepared , and there is no apparent theoretical reason why the list should not be extended indefinitely. The properties of primary arsines have been studied most ex- Q tensive ly by Dehn . He found that like the primary amines they react readily with alkyl halides with the formation of secondary and tertiary arsines and quarternary arsonium compounds, that, they are formed by the reduction in acid solution of any monoarylated or monoalkylated arsenic compound, that in certain cases they com- bine with sulphuric acid with the formation of unstable salts. On the other hand the primary arsines are oxidized much more read- ily than are the amines, halogens readily replace the hydrogen attached to the arsenic, etc., showing the effect of the more pos- itive character of the arsenic atom. Phenyl arsine is the structural analogue of aniline, and as may be seen from a comparison of the formulas of the two compounds the formal resemblance is very close, especially when the proximity of nitrogen and arsenic in the periodic table is taken into con- sideration. 6 Ber. d. Chem. Ges. 34, 3594. 7 Ber. (1901) 34, 3594. Amer. Chem. Tourn. (1908) 40, 113. D. R. P. 269, 843; 269, 74^7 251, 571; 275, 216; Chem. Abs. 11, 3256. 8 Am. Chem. Journ., 33, 101; 40, 88; 35, 1 (1906) . . ' 3 To what extent this structural analogy finds its outward ex- pression in chemical properties has been indicated above for cer- tain reactions. But there are other reactions in which these similarities might be expected to manifest themselves. Thus, aniline condenses with certain aldehydes according to the follow- ing scheme^: Phenyl arsine might be predicted to react in a similar manner: This reaction has been studied by Adams and Palmer who found that no reaction takes place until a drop of Hcl is added. A vigorous reaction then sets in. The reaction instead of follow- ing the above course, went as follows: The product is thus a new type of organic arsenic compound. This result is of considerable importance for two reasons. In the first place, the existence of such a type is of theoretical int- erest; in the second place, the most promising field for researches in which the purpose to discover superior compounds to replace arsphenamine and neo-ars phenamine, is in the study of new types of organic arsenic compounds 11 . Old types have been pretty carefully studied, and the chief objection to those which have so far proved most successful in the treatment of syphilis and virulent skin diseases is instability, a defect which seems to be inherent in the type (the arseno grouping) and which is fatally retained thru 9 Vide L. Rugheimer, Ber. d. Chera. Ges., 39, 1653 (1906) 10 Journ. Am. Chem. Soc., 42, 2375 (1920). The structure of these compounds has never been definitely proved. 11 Vide ibid. 0 -4- oc*/C 0 ^ H n.0 10 . ■ . ■ * . . all th6 enormous aeries of derivatives of this type which have been studied. The purpose of the present investigation was to stuuy tne properties of certain analogues of phenyl arsine and their re- actions, especially the reaction with aldehydes. The character of the reaction product between primary amines and aldehydes de- pends upon the character of the aldehyde, and aiso upon the na- y ture of the substituent in the ring of the amino compound. It should be of interest to determine if similar effects are opera- tive in the case oi the arsines. i 8 . % arsenic Caloul ted for CH^O gJI/! A sH- { OgHgCHO )2 19 « 71 .150? gran of aubatanoe red.oed 8*03 o.o. of iodine solution (1 c .o.»#005515 . arsenic ) 18.7.3 .1391 gram of eubatance reduced 7.45 c.c. o* io- dine solution 18.8.: The above product evidently Wa3 a condensation of one mole of o-fcolylaraine and two moles of onzaldohyde. The following for- mula: ia a possible and probable one. This condensation product is hytoo- lyced by cold, concentrated, hydrochloric acid and slightly by boil- ing dilute hydrochloric. It is not hydrolyse by boiling five min- utes with 30% so&iur hydroxide solution. o-Tolyl Arsine and ar aldehyde. A similar experiment was carried out to determine the action of paraldehyde on o-tol larsine. To about 25 grams of the arsine a few drops of concentrated hydrochloric acid and about 35 prams of par al- dehyde were added. The mixture became slightly r.rr;. but not so warm A as in the previous experiment • When the reaction had ceased, the flash was sealed and set aside. After a few days the r ixture con- taine a white solid and w,-.s rather turbid. The excess paraldehyde , about 18 *rmm >w;...a distilled off at 180 nr. pressure, ieavi p a clear yellov7 solution. An after .pt was made to distill this at the same pressure, but it decompose a, turning red. At 23 mm* it boiled at 165°. n second distillation, about 18 ;rars of product, all rhtly u 3 — ■ T-r ===== turbi J , was obtained. In neither case .v is .11 ox the product dis- tilled, as it charred slightly on gettiri" to a low volur.c. m ol.uB- i nr a few days a white solid settled out. The 11 uid was accented from the solid and or, distillation, a clear product, boiling at 165°.21rnnu , was obtained* The index of refraction la 30 (n) 1.5573 $ aroenic Calculated for C%C6H4AaHg( CH?CH0) 2 29.25 .0974 -ran substance reduced 7.67 c.c. of iodine solution ( lo .c .-*003528 *rai ar- senic) 27.78 .1527 or am substance reduced 12.04 c.o. Iodine solution 27.82 This substance po-tcibly • as t o structure*. H JH 3ir.il ar to the one suggested for the condensation product of bensal- dohyde and o-tol; l.roine. The p .raldohyde condensation gro¬ mentioned above was hy- drolyzed by cold, eoncentraod hydrochloric acid, giving a strong odor of paraldehyde -nd yellow and red solids. These li. ely were oxidation products of the o-tolylsraine which was foxed by the hy- drolysis. Boiling dilute hydrochloric hydrolyses the condensation product slightly. It is not hydrolysed with boiling dilute or cold 30 .alkali, but on boiling with 30f sodium hydros *de, it is hydro- lysed# II. DifiSCKIPTIVii; a . p-Uhlorophenyiarsine does not appear in tne literature. J?'or the present investigation it was prepared by a method similar to that described in 1). K. P. £5i,o7x ±C '~ J . Its pxeparatxon ana prop- erties are aesenbed in aetail in the experimental part oi this paper . i'ne condensation of p-chiorophenyiarsme with benzaldehyde was carried out as ionows : ‘i‘o one more oi' the pure redistilled arsine a little over two mores oi‘ tne aldehyde were added, tre n a lew drops oi cone. liCi solution. The mixture immediately became hot, an a completely solidified within a half' an hour to a whife , amorphous solid. The reactioxi mixture was, oi course, kept in an atmospnere of CO . On standing for an nour xii an atmospnere of CO , the mass gradually began to turn yeliow and at one ena ui i,nx«& «ji lour nours iu was xound to nave changed completely to a canary yeixow. This material was partiaiiy soluble in hot chloro- benzene. It was extracted with hot chlorobenzene, and the hot extract filtered. On cooling the riitrate, long silky needles separated which gradually aggregated to cottony masses which seem- ed to render the solution almost a soxid mass. These cr^s-ais , v n filtering off the liquid were found to oe very light, the yield from £0 grams of the arsine never amounting to over three grams and often failing beiow O.n & ram. They were recrystarlized se verai times from a mixture of chlorobenzene and alcohol. The results were the same when dry, gaseous instead of qqueous HCl wao used as a cataiysu. x£. Friedlander jlx, iO£4. 13. Vide Adams and Palmer ibid. . . - ' ■ The product thus obtained wua an extremexy ligu*, cottony tass, melting at 2l8-2lb°." C. (corr. ) { 21i .0-211° . ‘c , unc . ) and gradually changing to a hard, brown, caramel-like mass wnen be ated for some time doiow its T.eiting puxnl (e. g. at 140°) . It was in- soluble in water, dilute acias or aixalies, siightxy soluble in alcohol, ether, benzene, and ligroin, ana very soluble in ‘Chioro- ocnzene. It thus resenbled quite ciosely in all its pnysicax properties the a^axu^ous coxpound of phenyl arsine and oenzaidehyae obtained by Adams and Palmer. Tne analytical data, nov/ever, aia not utar out this conclusion very cios6iy. if this substances has it is evident that its formation does not represent the chief proauct when ultimate ana most stable conditions of equilibrium have been reached, in an cases tn.6 chlorobenzene insoluble residue constituted at least of the reaction product, ana apparently the relative proportion of the latter proauct increased as the reaction mixture was anowed to stand before extraction with chlorobenzene. This residue was at first a bright yellow, sticky mass, smeiiing strongly of benzaiaehyde . On standing in th6 air it gradually became lighter in color ana ary. After a coupxe of weeks it was found to be mostly soluble in hot iO% NaQH solution although a dirty, sticky residue remained insoluble, un dilution, this solution Decame turbid, but no precipitate appeared until the solution was made acid with HC1. The precipitate thus formed was .light yellow-grey, and of the consistency of molasses. On stand- ing for some time, it gradually became solid and granular. This , t N I precipitate was filtered oil aim dissolved up xu y bfi aioohol, al- most completely covered and anowed to evaporate very siowiy . Alter three weeks it was round that the amorphous yellow mass which remained contained may clusters of tiny, needlo-iike crystals. These were isolated and identified as p-chioropnenyiarsonic acid. Tne inference was tnat the main product oi the reaction was not the condensation product expected, but p-dichioroarsenobenzene : This accounts for tne changes above mentioned, ior the arseno compounds are usually yeliow, and they are all oxidized slowly by the air to arsonic acids. Attempts to isolate tne arseno compound failed, as would be expected because these compounds are difficult- ly soluble in most reagents, and usuany amorphous, ana some ben- zaidehyde was present to complicate matters. It was also attempted to condense p-chiorophenylarsine with paraldehyde. The reaction was carried out in the same way as with benzaldehyde . On the addition of HC1, the mixture became warm, but remained liquid. After standing ior i£ hours it was dissolved in Denzene , dried with fused calcium chloride , filtered, and fractionated under diminished pressure, most of the condensation product was found to distill over, out at the end decomposition took pxACe rather suddenly, xeavmg in the flask, a red, sticky mass. The distiixed product thus obtained was a colorless liquid boiling at i83° under £3 mm. oi mexcury. It was redistilled. The yield in this case was much better, amounting to about SOfc of theory. The product nad a refractive index of l.t>7£6 at £5°, and a specific gravity of . 74ol. it was insoluble in water and dilute ■ ” — — 77 HCl, Dut soluble m benzene ana chlorobenzene . Analysis indicated the formula ^iqH^Q^asCI which corresponds to O ^sfcHOKCfa/vj* Inis compound was very unstable, un standing it gradually under- went change with th6 separation of a soiid white materia. me nature of this product was not determined, cne amount obtained Do- ing msuiiicient lor analysis. it was insoluble in ZQ°/o HaOH, even after standing ior a xong time m tne air, and hence could scarcely have represented an oxidation product oi the arsine. It was soxubie in chior obenzene anu Ufjon evaporating oil the solvent under dimin- ished pressure at ordinary temperatures , no soxiu separated out, but a viscous xiquid remained very like the material from which it originally separated. This investigation was extended to a study oi the properties of certain other arsines. The preparation oi p-aminophenyiarsme i4 is described in i). n. P. 25i,t>7i . a or the present investigation it was pxeparea d/ the same metnou as tnat used m the preparation of p-chioropnenyiarsme . The pure aminophenylarsine was axso iso- lated and condensed in the same was, using three moles of ben- J zaidehyde instead oi two. it was supposed that the amino group would iirst condense m the absence oi HCi, and that the arsine woura tnen react upon the addition oi HCi. This predict tion was borne out by the generar uehavior of the mixture, un adding the oenzarden^ae tile mixture became quite warm, and alter cooling, when a few drops oi cone. nCi had oeen added, the mixture neated up again. The reaction mixture now gradually solidified to a trans- 14 Priediander Al, 1024 . . ' Q. xucent, almost coioness raad. It was Kept in an atmosphere of CO^ throughout. After standing m carbon dioxide for twelve hours xv, was found that rea spots nad appeared ana, upon expoexng tne mass to tne axr, tiie red ooior gradua^y spread tnroughout the whole mass, henzaidenyde was regenerated. The mass u.o'fi was founa to be soluble m cnxui uucnzene , from which solution it was precip- itated oy tne addition oi ailute aqueous HC1. After filtering and drying, a rather heterogeneous and very hara redaisn mass remained whicn was apparently compxetery xixSuxuble in any common organic solvents. It was suspected of being: tempt to hyarolyze the compound with dilute HOI and thus to obtain p-diammoarsenobenzene resulted in a decomposition 01 tne molecule with the formation of black tarry material. Analysis of tne red- disn material showed 22.26?fe arsenic. The theoretical value for a compound oi formula I is 2^.3b?b. The disparity is probably due to the impurity of the matenax. its hard, non-crystaiiine cnaracter and insoxubixity made it impossible to purify. XJ-Phenetyiarsine was prepared in the same way as p-emuxo- ana p-aminophenyiarsine . Apparently it underwent very considerable decomposition eitner during tne xeduction or during distillation, for the product showed i4.G4 jo of arsenic instead of the theoi tutox value oi 32. Vo;*. it smelled strongly of phene toie . About half of ih K. H. P. 206,057 Since the compound .LID resembles it in color and general solubility behavior . An at- ' ■ ■ a gram oi the axaiut) was therefore exposed to the air. A white precipitate formed at once , and al ter standing lor several hours an oil still remained. i'his was separated oy extracting tne suxid material with a little ether, filtering, and pulling off tne otnex from tne filtrate au ordinary temperatures under diminished pres- sure. i’he resraual on amounting to aoout u.o e . uuc. identified as ph6netole by its odor and boiling point. ihe residual white soiid amounting to about 0.2 g. was analyzed after drying at 60° for two nours. it showed Sl.oo^b Ox arsexiic . ihe theoretical value for p-pnene tyiarsonic acid is 30.4y> and for the correspond- ing arsine oxide, 55.2>8^fa. ihe wnite powder was tnerefore chieny pnene tyxarsomc aciu, containing some phene tyiarsine oxide as an o impurity, it melted at 174-6 in a capnxaiy m an oil Datn xu whicn the temperature was rapidly rising, if, nowever, tne temper- o ature rose slowly, especially within 10 of this melting point, it did not melt under 2Vo°. If a sample. of the material was melted and frozen repeatedly it became infusible, rhis behavior is char- lo acteristic oi the aryl arsonic acids , and is due to tne xus» oi The observed melting point is of course of no special signu icance except bo indicate tnat tne sample was not pure p-phene tyiarsonic acid, a conclusion which is ooxne out, by, the results of the analysis The product obtained by the reduction or p- pnene tyiarsonic a moiecul6 oi water. — > 16 Morgan, Organic compounds oi Arsenic and Antimony, p. 74 17 Ber. (1908) 41, 18b4 ■ . 10 acid thus undoubtedly consisted oi p-phenetyj.ars±xic mixed with phenetoie, the latter being fiomed by some secondary reaction tak- ing place during the reduction or the steam, distillation, and in- volving the splitting of f oi arsinic from, the ring. The boiling point of the highest boiling fraction of the mixture was 162-b 0 under 18 mm. and this undoubtedly represents the boiling point of pure phene tyiarsine . A rather remark ble iact about this mixture is that on oxidation it immediately yields a white powder. how the arseno compound represents the first pos- sible stage in the oxidation of an arsine, and p-die thoxyarseno- IQ benzene is known to be yellow. Apparently the arseno compound either was not intermediate, or it was under these conditions ox- idized more rapidly than it was formed. This behavior is rather unusual, and combined with the lability of the AsH 2 group in phene tyiarsine , indicates the remarkable and unexpected influence which the OC 2 H 5 group has on the properties of the phenylarsine rest . The mixture of phene tyiarsine and phenetoie was condensed with benzaldehyde as follows: To 14 g. of the mixture, 16 g. of benzaldehyde was added (2 moles assuming that the phenetylarsine is 100$ pure) and then a few drops of aqueous HC1. Heat was evolved, and after standing for 12 hours in an atmosphere of COg it was observed that crystals had separated. Without admitting any trace of air, the mixture was now steam, distilled until nothing further was carried over. About two grams of a yellow, sticky solid re- mained in the distilling liask, while in the receiver a considerable quantity of yellow oil had collected under the water. It was 18 Michaelis., Ann., 320, 300 (1902) f ; ' 11 . separated off and exposed to the air. .benzaldehyde and ohenetoie were obviously present, and later some crystals separated which proved to be phene tylarsonic acid, flow phenetylarsonic acid is not volatile, nor are any of the other possible oxidation products of phene tylarsine . it was .therefore , necessary to conclude that th6 arsine itself had distilled over. Som.6 of the ar3ine must, therefore, have remained unchanged in spite of the tact that the benzaldehyde was present in large excess, or the arsine must have been regenerated trom. its combination during the distillation. The residue from the steam distillation was now examined. it was found to be soluble in chlorobenzene, and ethyl alcohol, irom which sol- ution on cooling there separated long, silky needles, which were supposed to do the condensation product; OC x h 6 ' After recrystallization and drying they melted at 220-223° (corr). This product amounted to less than 0.1 g., and an attempt to deter- mine the arsenic content did not yield a result of any significance. The filtrates from both crystallizations were combined and the solvent evaporated off at ordinary temperatures . The residue re- gaining was a sticky viscous, yellow liquid. it amounted to about 1.5 g. It was supposed to contain some arseno compound, and som.6 benzaldehyde which nad been mechanically held back during the dis- tillation. 12 III. SIGNIFICANCE OF RESULTS It is evident that the reactions of these arsines with ben- zaldehyde are very complex, and that in any case, when equilibrium is reached, the product of the type RAs (CH0HCgH 5 ) 2 represents but a small fraction of the products formed. It is possible to explain the phenomena observed on the addition of p-chlorophenyl arsine to benzaldehyde by the reactions indicated below. The reaction goes practically quantitatively at first to the form- ation of IV which is a white solid. At the same time reaction B sets in, but is much slower than A. However, it is more complete, so that ultimately most of the arsine present is changed into V. This theory is borne out by several facts, the most important of which is the color change above mentioned. Benzyl alchol was never isolated from the reaction mixture, but Mr. 0» S. Palmer of this laboratory has identified it as one of the products in the reaction between phenylarsine and benzaldehyde. In one case alcohol was added to the arsine before the addition of benzaldehyde. Heating and the formation of a solid took place as usual on the addition of benzaldyhyde, but neither reaction A nor B could have been com- plete ; since most of the arsine was unchanged at the end of an hour and on exposing the mixture to the air, much heat was evolved, and p-dichloroarsenobenzene formed. The reversibility of reaction A must be questioned. Products . ■ ' X . . . 13 of the type IV are fairly stable although, at least in certain cases, they decompose much below their melting points, this de- composition being almost complete in two hours at 140° in the case of the product IV. The behaviour of the other two arsines studied furnished certain additional evidence. Thus, in the case of p-aminopheny- larsine care was taken to add just exactly the theoretical amount (3moles) of benzaldehyde. All the benzaldehyde disappeared after the addition of the HC1, but on exposing to the air, it was at least partially regenerated. The most rational explanation of this behaviour necessitates the assumption of a reversible reaction similar to the above: («) JU(chOHCuHs)^. 0 ■f* C> ■ ch syv O'"-*' 0—0 is less in evidence here, since in the course of 12 hours only a few reddish spots had appeared in the mixture in the absence of air; but these were sufficiently obvious to justify the assumption that some aldehyde had been reduced, providing F is the proper explanation for the formation of the reddish product. 14 Similar explanations suffice to account for the observations made in the case of p-phenetylarsine In this case also reaction H was not so much in evidence, but that it actually took place to some extent is indicated by the yellow color of the product obtained after the completion of the steam distillation. That is to say, reaction H was very slow, just as was reaction F; but that reaction G was reversible was indicated by the fact that most of the arsine originally present appeared in the distillate from the steam distillation, in spite of the fact that a large excess of benzaldehyde was present. Thus it is concluded that compounds of the type IV, VIII and X were formed, but that the reaction by which they formed was in each case reversible so that they were each in equilibrium with the arsine from which they are derived. In each case the equilib- rium was displaced by the removal of the arsine mom che field o^. the reaction; the p— chlorophenylansine was oxidj.zeo. by ohe benzal— dehyde to the corresponding arseno compound; the p-aminophenylar- sine was oxidized on exposure to the air to the arseno compound; the p-phenetylarsine was removed by the steam. A 15 The chief objection to this explanation is that there is no direct evidence for the evolution of arsine from these condensa- tion products, and while it is pointed out that those compounds with which this study is concerned showed no very high degree of stability, and in certain cases (notably the compound of paralde- hyde with p-chlorophenylarsine ) were decidedly unstable; yet it must be admitted that the above explanations cannot be regarded as completely consistent and satisfactory unless they are fortified with some additional assumptions, such as: that the reverse reac- tion is catalyzed by some substance which is absent after the puri- fication, or that in some manner, the purified product is more stable than that from which it was derived. The basis of these speculations is made additionally insecure through the fact that the structore of these condensation products has never been defi- nitely established. 16 IV . EX PER IMENTAL P ART . 1 . Preparation and properties of p-chlorophenylarsonic acid : This substance has been prepared by Bertheim 18 from arsanilic acid by the Gattemiann diazo reaction. For the present investigation it was prepared from p-chlorooniline by a modification of the Bart’s 19 reaction , the coupling being carried out in the absence of free 20 alkali . A typical run Is described below: 126 g. of p-chloroaniline was stirred up by means of an ef- ficient mechanical stirrer with 600 cc. HgO and 180 cc. cone. HC1 (s. g. 1.19), cooled to 0°, and diazotized by the addition of 68 g. of sodium nitrate in 250 cc. of water, the temperature being kept below 7°. 350 g. of crude ASgO^ (theoretical plus 80$) was dis- solved in 1.5 liters of water to which 565 g. of sodium carbonate had been added by heating on the water bath for two hours with oc- casional shaking. Some of this material usually failed to go into solution. The mixture was, therefore, filtered with suction, and to the filtrate 10 g. of anhydrous copper sulphate was added with stirring. The solution was then cooled to approximately room tem- perature and the diazo solution siphoned into it very slowly. The solution was stirred constantly with a motor stirrer during this addition. A tendency to foam limits the speed of the addition of the diazo solution. It Is, therefore, necessary to use as large a vessel as possible for this reaction, not smaller than 5 liters preferably much larger. Foaming can often be greatly mitigated by the addition of a few cc. of benzene occasionally. Stirring is 18. Ber (1908) 41, 1854. 19. D. R. P. 2517 092, Frdl., 11, 1030 (1913) 20. J. Ind. Eng. Chem. 11, 825"Tl919) ■ . , t * • 17 continued for four hours after the addition of the diazo solution is complete. The mixture is then allowed to stand for 12 hours and filtered with suction from the tar. The filtrate is clear and ciUXU slightly greenish in color. It is made/ -with "glacial acetic. The addition of the acetic acid must be carried out very carefully, since the mixture may foam violently. If the approximate quantity of acid required is known, the solution may be added to the acid. Pi The foaming under these conditions is much less . The purpose of the addition of the acetic acid is to precipitate out the excess of ASgO^, it having been found experimentally that arsenious acid is 2 p precipitated almost quantitatively ” from solutions of its salts by the addition of acetic acid, while p-chlorophenylarsonic acid is not so precipitated. If the As^O,, thus precipitated is now filter- ed off, the filtrate will be found to be perfectly clear, and on the addition of cone. HC1 (about one-fourth the total volume) the 21. J. Am. Chem. Soc. 43, 161 (1912) 22. That is, for the purposes under consideration. The solubility of AsgOg in water approaches 20 g. per liter at ordinary tempera- tures. HAsOp is, however, ionized in 0.1N solution very slightly between 0.002 and 0.008% according to A. A. Noyes, Qualitative Chem- ical Analysis, p. 125. Acetic acid in similar concentrations is ionized to the extent of 1«2%. Hence, in a solution of AsgO* con- taining acetic acid, the concentration of AsOg will be very low, — - probably of the order of .0001 equivalents per liter (mass action effect). Any further repression of the ionization of HAsOg by an in- crease of the hydrogen ion concentration (addition of HC1) will, therefore, of necessity be extremely slight; and hence will have very little effect on the total concentration of HAsOg. H-AsO^ is a much stronger acid than HAsOp (Ionization 20-45%) and p-chlorophenylar- sonic acid would be predicted to be of the same order, and its solu- bility in water should be quite appreciable (cf. phenylarsonic acid). Acetic acid in the presence of large quantities of sodium acetate which are present does not furnish a sufficiently high concentration of hydrogen ions to exceed the solubility product of p-chlorophenyl- arsonic acid, but HC1 does; and, moreover, the salting out effect of a moderately concentrated solution of HC1 on the neutral molecules should be much greater than that of HC1 on the neutral molecules of HAsOp, a very weak electrolyte. See Washburn, Principles of Physical Chemistry 1915, pp. 227, 228. * , . . . . . ■ * » in' 18 p-chlorophenylarsonic acid will be precipitated as a perfectly white, granular precipitate. It is filtered off with suction, and thoroughly washed with cold water to remove any traces of salt, and dried for several days by exposure to the air. The product is a fine, white tasteless, dusty powder. The yield amounts to about 200 g. or 85$ of theory. Its solubility behavior is indicated below: Solvent Hot Cold Water Dilute alkalies Cone HC1 Alcohol Ethyl acetate Glacial acetic acid Ether Somewhat Very soluble Very soluble Very soluble Soluble Quite soluble Somewhat soluble Sparingly Very soluble Insoluble Sparingly Somewhat soluble Soluble Somewhat soluble In a capillary tube a sample of the crude acid did not melt, but decomposed at about 320°. Similarly a sample recrystallized from alcohol decomposed at 348°. It crystallized in needles from alcohol, glacial, acetic, or hot concentrated HCl. The latter is an ideal solvent so far as solubility behavior is concerned, and moreover, boiling with HCl would probably distill off as AsClg any traces of AsgOg which might be present; but analysis of samples re- crystallized from this solvent always showed too high a chlorine content, hov/soever thoroughly they were washed with boiling water. The sample used in obtaining the following analytical data was crys- tallized once from glacial acetic and twice from 95$ ethyl alcohol and dried to constant weight at 90° under diminished pressure. .1612 g. sample used 26.38 cc. I 2 sol. =-31.58$ As .1782 g. sample used 28.77 cc. Ig sol. - 31.89$ .1966 g. sample used 31.60 cc. Ip sol. = 31.74$ 1 cc. sol = .001975 g. As Average 31.74$ .4190 g. sample gave .2539 g. AgCl z 14.99$ Cl .5190 g. sample gave .3176 g. AgCl =-15.14$ Cl Average = 15.06 Calculated for C 6 H 6 0 3 C1 As: As, 31.71$; Cl, 15.01$. 19 9 ^ The arsenic was determined by the method of Ewins and the chlor ine by the method of Carius. i i 23 Chem. Soc. Trans. (1916) 109, 1356. ' . 20 2. Preparation and properties of p-chlorophenylarsine • Consider- able difficulty was experienced in working out the proper conditions for the reduction of p-chlorophenylar sonic acid. The method finally adopted was as follows: 70 g. of p-chlorophenylar sonic acid and 350 g. of thoroughly amalgamated zinc dust were placed in a 3 liter round bottom flask, and 250 cc. of methyl alcohol added. The flask was provided with a two hole rubber stopper. Through one of these holes extended the lower end of a long glass condenser, the upper end being provided with a mercury trap consisting of a bent glass tube dipping into a tube of mercury. The other hole of the stopper carried a 3mm. glass tube about two decimeters long, and connected at the upper end with a 500 cc, dropping funnel. All joints were sealed with paraffin. The dropping funnel was filled with cone. HC1, the long tube filled with the acid and the stop-cock so reg- ulated that the acid dropped in at the rate of about 3 or 4 drops to the minute. The combined effects of capillarity and the pressure inside the flask prevented the long tube from emptying even if the acid should all run out of the funnel during the night. About a liter of cone. HC1 was thus added, and the run was considered com- plete when the zinc had all or practically all disappeared. This usually required from three days to one week. The stopper was now removed from the flask and quickly replaced by another bearing three glass tubes, one connected to a source of steam, another to a source of COg, and another to a condenser. The whole set-up is illustrated in the following diagram: ' . . ■ . . 21 Before admitting steam the apparatus was permitted to fill complete- ly with COg. Steam was then passed through for about four hours. The arsine solidified from time to time in the condenser, and was pushed out by shutting off the condenser water and blowing out with COg. When the distillation was complete, the stopper bearing the adaptor was removed and immediately replaced by one bearing a right- angled tube attached to a source of COg. The suction flask was now removed from its ice bath. The p-chlorophenylarsine was found col- lected as dark solid at the bottom of the flask. A little glass wool was stuffed into the tubulure of the suction flask, the flask tipped, and the water pushed through the tubulure by the entrant steam of CO p. The stopper was partially removed without, however, removing the end of the COg tube and 200 cc. of ether added. The tubulure of the suction flask was inserted into the mouth of a 500 cc. separatory funnel filled with CO 2 and the etheral solution poured out. The flask was washed with a little ether and the pro- cess repeated. Water was drawn off from the bottom of the funnel • ' - . . and then solid KOH and fused CaC12 were added to the solution, and it was allowed to dry for two hours. The method of transferring the ethereal solution to the distilling flask was as follows: The Claissen flask was provided with a stopper hearing two tubes, one it a capillary for CO (as in a Bruhl apparatus), and the other bent twice at right angles and closed at the lower end with a piece of stoppered gum tubing. Dry COg was admitted through the capillary from mercury trap so arranged as to provide the gas at a constant pressure slightly greater than atmospheric. The Claissen flask was connected with a receiver consisting of an ordinary distilling bulb, and the latter with an air pump and a manometer as in an or- dinary vacuum distillation. After the system had been alternately evacuated to the limits of the pump and filled four times with COg at the pressure of the source, the stopper was pulled off from the right-angled tube and the latter dipped into the ethereal solution, CO 2 being continuously admitted. The COg was then shut off and the suction turned on, and the flask filled with ether. The suction was so regulated as not to draw in any air, and as soon as the de- sired amount of ether was drawn in the suction was shut off and the COg turned on. The separatory funnel was then withdrawn and stop- pered, the right-angled tube closed, the ether pulled off in vacuo, and the process repeated until all the solution had been transferred to the flask. The solution was then fractionated. No special re- ceiving flask was necessary, providing that the whole system was filled with COg before changing receivers. Considerable experience taught that the best method of ensuring complete exclusion of air from the final product was as follows: . . 23 When distillation was complete, the suction was cut off and the system filled with COg the stop-cock to the source being left open. When the rubber suction tube was now pulled from the glass delivery tube of the receiving flask, a stream of COg rushed out of that tube and effectively prevented the ingress of any air. The rubber suction tube was then replaced by a rubber tube through which a stream of COg poured, and the receiver drawn off from the delivery tube of the distilling flask. No air could enter the top of this flask because COg was passing continuously out; and hence it was not necessary to stopper the flask until the COg was cut off. This arrangement made it possible to draw out samples, etc. without any danger of contamin- ation from the air. The optimum yield of twice redistilled product was about 26 g. Several other methods for reducing the arsonic acid were tried without success. The presence of the methyl alcohol seems to be ab- solutely essential, probably because the solubility of the arsonic acid in water is not great. Some traces of arsine as indicated by the odor were always formed in reductions carried out without the use of methyl alcohol, but no arsine could be isolated. Attempted re- ductions using zinc dust and solid NaOH also failed. Ether extrac- tion methods also were unsuccessful, supposedly because no consider- able reduction took place unless methyl alcohol was present, and in its presence the ether extraction method is complicated by the con- siderable solubility of ether in the alcohol water mixture. p-chlorophenylarsine which has been twice redistilled is a solid crystallizing in very thin, flat, perfectly transparent leaves. These crystals often attain a size of 1.5 cm. square, and in the re- duction of the arsonic acid where some of the arsine vapor is carried 24 . up out of the surface of the liquid, they condense against the sides of the flasks, standing out at right angles to its surface. The odor is characteristic and persistent. It is not unpleas- and, and resembles benzene to some extent. The observed boiling points at various pressures are recorded below. Pressure Temperature 18-20 mm. 98-101° 33 116 38 119 67 •143-6 200 159 It melts at 30.5-30.7°. It is soluble in ether and alcohol, but insoluble in water. On exposure to the air it is immediately oxidized to the yellow arseno compound with the evolution of heat. Analysis of the arsine was carried out as follows: Small bulbs resembling Victor Meyer bulbs were blown on long, thin capillaries. The bulb was weighed, heated in a flame, and the tip quickly dropped into the molten arsine. On cooling, a drop of the arsine sucked back into the bulb. This was carefully heated to boil- ing and the bulb again cooled. The bulb no?/ filled with the arsine. The capillary was drawn rapidly through a flame to expell the arsine in it, and the tip sealed off. It was then weighed. For the deter- mination of arsenic the capillary was broken off and the broken peices together with the bulb wrapped up in a peice of filter paper, the bulb smashed with a hammer, and the package dropped into a 500 cc Kjeldahl. The proceedure from here on was simply the method of Ewins, the addition of starch being omitted. Chlorine was determined by the method of Carius. The bulb was placed in the digestion tube and broken by dropping in a heavy glass slug. The arsine was then . - , ; v 25 allowed to oxidize in the air, and the nitric acid added very care- fully a drop at a time and with careful cooling, since nitric acid sometimes reacts with explosive with the arseno compound. After the tube was opened, the silver chloride was dissolved out with ammonia, filtered from the glass, and reprecipitated with HNO3 . It was, of course, necessary to digest the reprecipitated AgCl for several hours in the dark to obtain a filterable precipitate. Some of the silver chloride usually remained stuck in the peices of capillary and did not dissolve in ammonia. It was necessary to shake the tube until these capillaries were completely broken up against the glass slug. .5147 g. sample required 103.64 cc. Ig sol. 39.75$ As .3622 g. 72.20 cc. 39.37 1 cc. ig sol. = .001975 g. as Average 39.51 .4259 g. sample gave .3184 g. AgCl 18.50$ Cl .4592 g. .3520 g. 18.97 .2539 .1914 18.65 Average 18.71 Calculated for CgH4AsCl: Cl, 18.81$; As, 39.79$ ' , • « * . • • * • • • - 26 3 . p-chlorophenylarsine benzaldehyde condensation product ; The preparation of this compound and its properties is in the descriptive part of this paper. Analytical data: .1406 g. sample gave .0473 g. AgCl = 8.32$ Cl .1676 g. .0558 g. " = 8.23$ .1082 g. required 8.95 cc. Ig sol. 16.34$ .1144 g. required 9.21 cc. Ig sol. 15.90$ 1 cc. i .2 sol = .001975 g. As Calculated for CgOHqQOgClAs : As, 18.63$; Cl, 8.74$ analysis . described . . . 27 4. p-chlorophenylarsine paraddehyde condensation product; Analysis. The preparation and properties of this compound are described in the descriptive part of this paper. Analytical data: .3215 g. sample gave .1620 g. AgCl = 12.47$ Cl .3601 g. .1820 g. , 12.50$ Average 12.49$ .3674 g. sample required 50.01 cc. Iq sol. 26.88$ As .3771 g. 52.18 cc. 27.33$ 1 cc Ig sol = .001975 g. As Average 27.11$ Calculated for C 10 H14 02 As Cl: Cl, 12.82$; As, 27.12$. . . ' . . * . - 28 5. Preparation of arsanilic acid and its reduction: Arsanilic acid was first prepared by Bechamp in 1863 2 ^ by heating aniline arsenate to 190-200°. This method i3 apparently the only one which has been used in preparing this important arsonic acid. The latest modifica- tion of this method is that of Cheetham and Schmidt '^ ' J , who claim 20$ yields when working with small quantities. The modified Bart's reaction had been found in this laboratory to work so smoothly in the preparation of many arsonic acids that it was thought worth while to work out the conditions for the preparation of p-acet7/larsanilic acid by this reaction and for its subsequent hydrolysis to arsanilic acid. The preparation of acetylarsanilic acid by Bart's reaction is O described in D. R. P. 250,264 AvU . The proceedure used in the present investigations was as follows: 50 g. of pure (Eastman) p-aminoacete.nilide in 200 cc. HgO and 100 cc. EC1 (£.g. 1.19) was diazotized with 23 g. of solution of NaN02 in 500 cc, HgO. 120 g. AsgO^ and 200 g. NagCOg were dissolved in 350 cc. H2O with heating, filtered, and to the filtrate 3 g. GuSO^ was added. To this cooled solution the diazo solution was added with stirring, the mixture made acid with acetic acid, filter- ed, and the filtrate treated ?;ith an excess of HC1. 65 g. of p-acetylaminoarsonic acid was obtained 75$ of theory. HCl> NaOH, and HgSO^ in various concentrations were tried as hydrolytic agents. NaOH was found to give the best results and to permit the isolation of the arsanilic acid most readily. The proceedure was as follows: 24 Compt. rend. 56, I, 1172 (1865) Abs. in Jahrsbericht (L. & K. ) (I860) 414. 25 J. Am. Chem. Soc. 42, 828, (1920) 26 Friedlander VIII, T£25 * ■ ’ , • * * . • • . . - . 29 6,5 g. |>-acetylaminoarsonic acid was boiled under reflux for 2 hours with 35 cc. of 20$ NaOH solution. The mixture was then boiled ten minutes with a little animal charcoal, and filtered into 50 cc, of 95$ ethyl alcohol. The resulting solution was cooled with ice and 10 cc. of 40$ acetic acid added. Crystals separated out after a short time until the mass was nearly solid. 30 cc more of alcohol was then added and the mixture allowed to stand over night. The sodium salt was then filtered off, and washed with alcohol and ether. Yield, 3.6 g. or 60$ of theory. With a little experimentation the yields of this product (calculated on the acetylarsanilic acid) should undoubtedly be run up to very nearly 100$. The amount of sodium hydroxide solution used in this experiment was probably ex- cessive. Using smaller relative amounts of sodium hydroxide solu- tion and working with larger quantities this would undoubtedly be an entirely practicable method for the preparation of arsanilic acid. The process would also be much simplified if the free acid were pre- fab nyarogen ion conctentx'tit-Lon instead of isolating p r . cipitated out by properly adjusting the A sodium salt with alcohol. In the hydrolysis it is quite essential that the acid or alkaline hydrolytic agent be not too dilute, since arsanilic acid is decom- posed by hot water rather rapidly. In the presence of fairly con- 2ft eentrated acids or alkalies, it is, however, perfectly stable. Reduction of arsanilic acid: This reduction was carried out precisely as in the case of p-chlorophenylarsonic acid, the propor- tions being the same. It was, of course, necessary to make the re- action mixture alkaline before steam distilling. It distills much more slowly than does the chlorophenylarsine. The yield of twice re- distilled product from 100 g. of arsanilic acid was 24 g. 27 Sec J. Am. Chem. Soc. 42, 828, (1920) 28 E. Schmitz, Ber., (191T7 47, 365; ibid 996 . . . . . . * «* . , , " • : , . . . . • 30 The product was a colorless liquid boiling at 145-7° sure of 41 mm. This agrees with the description of oq arsine in D. R. P. 251, 571. under a pres- p-aminophenyl- 29 Friedlander, XI, 1024. . ♦ • . *v 31 6. Preparation of p-phenolarsonic acid and its reduction: The method used was that of Jacobs and Heidelberger . Yields of crude product were approximately 20 %, but considerably less than this when calculated against the recrystallized product. Contrary to the statements of Jacobs and Heidelberger, tar was always formed, and the crude sodium salt was always colored. This color was removed by a single precipitation from a concentrated aqueous solution by alcohol. Syrupy arsenic acid was prepared from crude arsenic trioxide according to the method of Vannino (3± ). In dealing with quantities of over 100 g. it is quite important to carry the reaction out in a large dish (not in a flask, or as Vannino recommends, in a retort) and to add the ASgO^ to the acid. The method used by Jacobs and Heidelberger in isolating the product is very tedious and attempts were made to simplify it. The magnesium salt was isolated in two runs instead of the sodium salt. This was carried out as follows: The reaction was run as usual, using 450 g. arsenic acid and 200 g. phenol, heating at 150° for 6 hours, and extracting the pro- duct with two liter of HgO. This filtered extract was divided into two equal fractions. The sodium salt was isolated from the first by neutralizing with Ba(0H)g, exactly removing the excess barium ions with H 2 S0 A , filtering, exaporating in vacuo, exactly neutralizing with NaOH, evaporating in vacuo until crystallization began, adding a large excess of alcohol, cooling for 12 hours, and filtering off the sodium salt. The second fraction was treated as follows: 250 g. ice, 300 g. MgCl 2 , 200 g. NH^Cl, and 100 cc. excess cone. NH^OH, and were added and the mixture allowed to stand for several hours. 30 J. Am. Ghem. Soc. (1914) 41, 1446. 31 Handbuck der ?rapMrativen Chemie I, 184 - ' . . . . [tout n . . . . * 32 . It was then filtered. The filtrate was heated to boiling, whereupon the magnesium salt of the p-phenolarsonic acid separated. A compar- ison of the yields from the two fractions is shown below: Run III Run IV Mg salt 30 g. 49 g. Na salt 62 g. 57 g. The isolation of the magnesium salt is extremely simple compar- ed with the isolation of the sodium salt, but the magnesium salt 13 not so clean a product and the yields are not so good. It is proba-- Difi that further experimentation would improve both factors and make this process entirely practicable. The effect of the use of an excess of phenol, instead of an ex- cess of the arsenic acid was tried in one experiment. The excess phenol was removed by distillation in vacuo. A dirty gum was formed from which no crystalline product could be isolated. Preparation of p-hydroxyphenylarsine : p-phenolarsonic acid was reduced by the same method used in preparing p-chlorophenylar sonic acid excepting that the methyl alcohol was omitted, p-phenolarsonic acid being extremely soluble in water. The reaction mixture was steam distilled. The distillate contained only water. The contents of the reduction flask were then extracted with ether. The extract was black, the black material supposedly coming from the zinc. It was extracted with 10$ NaOH. The NaOH extract was black. It was saturated with COg, and extracted with ether. The ethereal solution was now colorless. It was again extracted with 10$ NaOH and the ex- tract saturated with COg. A buff colored precipitate was formed, apparently the p-dihydroxyarsenobenzene. If HCl was used instead of COg as a precipitant, the precipitate was red. All the reactions were carried out in an atmosphere of CO 2 . p-hydroxyphenylarsine is . - . . , 53 2 . 32 described in D. R. P. 251, 571 as a white powder which readily decomposes on heating or exposure to the air. Apparently it was not formed in the present case; or else decomposed during the attempt to isolate it. I 32 Friedlander XI, 1042 ii « . . I « 34 . 7. Preparation of p-phenetylarsonic acid an d its reduction; This acid was prepared in excellent yields from p-phenetidine by the mod- ified Bart’s reaction • It was reduced in the same way as the other arsonic acids described above, and the arsine isolated in the same way. The properties of the arsine are described in the descriptive part of this paper. 33 This preparation was carried out by Mr. (J. H. Cheney of this laboratory. 4 " 35 IV. SUMMARY . 1. The preparation of p-chlorophenylarsine, p-aminophenylarsine, and p-phenetylarsine is described. Of these p-chloropnenyl- and p-phenetyl -arsine are new. 2. The reactions of these arsines with certain aldehydes have been studied. 3. The preparation of p-chlorophenylarsonic acid and of arsanilic acid by the modified Baht’s reaction is described. . - , * 36 V. ACKNOWLEDGEMENT. This problem was suggested by Professor Roger Adams and the experimental work was done under his direction. The writer ex- presses his appreciation of the kindly interest and guidance which Professor Adams offered, and without which this investigation would have been impossible.