■ 7 ->«*>■' >. ^ "^>-^^' - ^c^N ^:, ■ >fe,?w^i:- ...l^' •^^ <• . ..^^^'^-t^ ■^'*^ ...i.a.i^. A. Cw» 'Si QfatneU UttiwctHita ffiibtarg THE GIFT OF Cornell University Library QC 477.N61 Fluorescence of the uranyl salts, 3 1924 012 330 746 Date Due 1 1 1 /"^^i^^^^^ ^ii^=2^ tf%«>n ,„£ iM?-'°Mr^tr1 D SIP***** " *-" nil "^A D ^jtf^MJ ^^^^^^^^^^^^Bt^~^^ i "TM^ m ^^^^^^ATJ J- « r-'! 3 ur ■^ df f Z3233G Cornell University Library The original of this book 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/cu31924012330746 FLUOEESCENCE OF THE URANYL SALTS 5 ^^■ BY EDWARD L. NICHOLS and HORACE L. HOWES IN COLLABORATION WITH ERNEST MERRITT, D. T. WILBER, and FRANCES G. WICK Published by the Caknegie Institution op Washington Washington, 1919 FLUORESCENCE OF THE URANYL SALTS BY EDWARD L. NICHOLS and HORACE L. HOWES IN COLLABOKATION WITH ERNEST MERRITT, D. T. WILBER, and FRANCES G. WICK Published by the Carnegie Institution oe^ Washington Washington, 1919 \ r /\4^034\ CARNEGIE INSTITUTION OF WASHINGTON Publication No. 298 PRESS OP GIBBON BROTHERS, INC. WASHINGTON, D. C. ■m:^hhoo ■ CONTENTS. PAGE I. Historical Introduction 5 II. The Structure of Fluorescence Spectra 10 III. Prelimiaary Observations on Certain Uranyl Salts 15 IV. Phosphorescence of the Uranyl Salts 38 V. Intimate Structure on Cooling to —185° C 61 VI. Polarized Spectra of Double Chlorides 102 VII. The Nitrates and Phosphates; Water of Crystallization; Crystal Form 122 VIII. The Acetates 146 IX. The Sulphates 169 X. Frozen Solutions 180 Appendix 1. Chemistry and Crystallography of the Uranyl Salts 207 Appendix 2. On Phosphoroscopes 231 3 PREFACE. This volume, the completion of which has been much delayed by the participation of America in the World War, contains the results of an investigation covering a period of eight years. The discovery by Becquerel and Onnes, that the fluorescence of certain uranyl compounds is resolved into groups of narrow line-like bands when these substances are excited to luminescence at very low temperatures, suggested to the present authors the desirability of a thorough and systematic study of this subject. The spectra of numerous uranyl salts, many of which were espe- cially prepared for this pmpose, have now been mapped. Owing to the extraordinarily complex character of the phenomena, no satis- factory theory has as yet been evolved, but the mass of facts here recorded and the general principles established will, it is hoped, afford a basis for the successful theoretical development of this important and httle understood branch of the science of radiation. Physical Laboratory of Cornell University, May H, 1919. FLUORESCENCE OF THE URANYL SALTS. By Edward L. Nichols and Horace L. Howes. I. HISTORICAL INTRODUCTION. The beginnings of precise knowledge concerning the luminescence of the compounds of uranium are to be found in the classical memoirs of George Gabriel Stokes and of Alexandre Edmond Becquerel. It is true that Brewster^, who observed the fluorescence of chlorophyl and other substances in 1833 and gave the phenomenon the name of internal dispersion, mentioned a yellow glass, doubtless the "canary glass" of commerce, which exhibited the same property, but it remained for Stokes,^ by means of the beautiful experiments described in his papers entitled ''The Change in the RefrangibiUty of Light," to really eluci- date the phenomena and to lay the foundation for all subsequent work on fluorescence. Having observed, by the use of suitable Ught-filters and by his ingenious and elegant method of transverse dispersion, the unusual character of the fluorescence and absorption of this glass, Stokes pro- ceeded to the investigation of such compounds of uranium as he was able to procure. From the nitrate he made the acetate, oxalate, and phosphate; also uranates of potassium and calcium and the oxides. He also obtained specimens of autunite (uranyl calcium phosphate) and chalcolite (uranyl copper phosphate). After observations of these minerals he writes (Sec. 145) : "The intervals between the absorption bands of green uranite were nearly equal to the intervals between the bright bands of which the derived spectrum (i. e.j the fluorescence spectrum) consisted in the case of yellow uranite. After having seen both systems I could not fail to he impressed with the conviction of a most intimate connection between the causes of the two phenomena, unconnected as at first sight they might appear. The more I examined the compounds of uranium, the more this conviction was strengthened in my mind.*' Upon reading Stokes's memoir one can not but feel that had he had at his command a modern spectroscope he would infalUbly have antici- pated by more than half a century much of the recent work on fluores- cence. He used light-filters to prevent the exciting beam from sub- merging the fluorescence on the one hand and to exclude the exciting ^ Sir David Brewster, Trans. Roy. Soc. Edin., vol. xii. 1833. 2 stokes, Phil. Trans., 1852, p. 463; 1853, p. 385. 6 FLUORESCENCE OF THE URANYL SALTS. rays from entering the eye on the other, and thus by means of a prism held to the eye was able to observe the spectra of both fluorescence and absorption with surprising accuracy. Paragraph 148 of his paper describes his observations on uranyl nitrate. In the following quotation of that paragraph certain pas- sages forecast in an extraordinary manner some of the conclusions reached in subsequent chapters of the present monograph: "The sun's light was reflected horizontally by a mirror and condensed by passing through a large lens. It was then transmitted through a vessel with parallel sides containing a moderately strong axomoniacal solution of a salt of copper. The strength of the solution and the length of the path of the hght within it were such as to allow of the transmission of a little green besides the blue and violet. *'A crystal of nitrate of uraniiun was then attached to a narrow slit and placed in the blue beam which had been transmitted through the solution, the crystal being turned toward the incident light. The hght coming from the crystal through the slit was then viewed from behind and analyzed by a prism. A most remarkable spectrum was then exhibited, consisting from end to end of nothing but bands arranged at regular intervals. The interval between consecutive bands appeared to increase gradually from the red to the violet, just as is the case with bands of interference. Although this interval appeared to alter continuously from one end of the spectrum to the other, the entire system of bands was made up of two distinct systems, different in appearance and very different in nature. The less refrangible part of the spectrum, where, only for the crystal, there would have been nothing but darkness, was filled with narrow bright bands due to the Ught that had changed its refrangibihty. The more refrangible part of the spectrum was occupied by the system of bands of absorption. The interval between the most refrang- ible hght band and the least refrangible dark band of absorption appeared to be a very little greater than one band interval, so that had there been one more band of either kind the least refrangible absorption band would have been situated immediately above the most refrangible bright band. With strong hght I think I have seen an additional band of this natiu'e." Becquerel, in the course of his work on phosphorescence, notes the fact that most of the compounds of uranium show a strong emission of light when examined with the phosphoroscope. He determined the duration as three to four thousandths of a second; and to test his empirical formulae made measurements of the rate of decay which, as wiU be seen, are in good agreement with the results described in Chapter IV of the present treatise. With a prism of carbon bisulphide he observed 8 bright bands in the spectrum of the phosphorescent Ught of uranyl nitrate; he determined the approximate range in the violet and ultra-violet of the exciting rays; noted that the bands in the spectra of various uranyl salts, such as the chloride, fluoride, and uranyl potassium sulphate, occupy different places. He also esti- mated their relative displacements when compared with the bands of the nitrate. By comparing the spectrum of the nitrate during excita- HISTORICAL. 7 tion with that of the afterglow, Becquerel reached the very important conclusion that the fluorescence and phosphorescence are identical. This point finds ample confirmation in the present work. In 1872, E. Becquerel returned to the study of the uranyl salts. The following are the conclusions reached in this investigation:^ (1) The salts of the protoxide of uranium are inactive. (2) Many, but not all, salts of the sesquioxide (uranyl salts) are active. (3) Five, six, and sometimes seven bright bands, or groups, are visible; lying between the Fraunhofer lines C and F. (4) The positions of the bands vary for different salts, but are always the same for a given salt. (5) The acid of composition determines the disposition of both bright and dark bands. (6) In double salts of the same acid the composition of individual groups is the same, but their position is not the same for the different salts. (7) In a given substance the distance between bands, as viewed in the spectroscope, increases from red to violet; but the differences of wave-length decrease. The ratio of the above distances to the square of the mean wave-length is nearly constant throughout the spectrum, and this ratio (d/2) is the same for the various salts. (8) No simple relation is apparent between the location of homologous bands in different compounds and the chemical properties of the compounds. (9) The absorption spectra also differ for the various compounds and the absorption bands seem to form a continuation of the fluorescence series. (10) The location and character of these spectra being fixed and definite for each compound, we have the basis for an analytical method similar to but less general than ordinary spectrum analysis. In 1873, Henry Morton and H. Carrington Bolton published an account of extended studies of the fluorescence and absorption of the uranyl salts.^ Their list contains 85 substances, chiefly of their own preparation, including 17 double acetates; but not all of these com- pounds were found to be fluorescent. Readings were made on the Bunsen scale in vogue at that time, and some of these, for comparison with our own determinations, will be found, reduced to approximate wave-lengths, in Chapter III. Figure 1, which is reproduced from the paper of Morton and Bolton, gives an excellent general view of some of the most interesting of their observations. The unshaded portions are fluorescence bands, the shaded regions are the bands of absorption. The partial resolution of the bands in several cases is clearly shown and the breaking-up into distinct groups of the uranyl ammonium chloride; also the coincidence in certain cases of absorption and fluorescence in what in this mono- graph we shall term the reversing region. ^ E. Becquerel, Compteg Rendus, lxxv, p. 296. 1872. 2 Morton and Bolton, Chem. News, pp. 47, 113, 164. 273, 244, 257, 268. 1873. 8 FLUORESCENCE OF THE URANYL SALTS. The authors note specifically the following further important char- acteristics of the spectra of the uranyl salts: (1) The steeper gradation of light on the side toward the violet in the case of bands showing a single crest. (2) The weakness of the outer bands, both toward red and violet, com- pared with the central bands of the spectrum. (3) The overlap of fluorescence and absorption. (4) The systematic shift of bands when a salt is dissolved in water and other solvents. (5) The remarkable changes due to the dehydration of salts containing water of crystallization. (6) The effects of heating. 2.3.4.5 8 . 9 , 10 , 1.1 , 12 , 13 , 14 , 1.5 , 16 iihMilili 3 imliiTi 4 Mm fAJrJitolbWljfiy 6 jukw Ip , 11 . 1.2 . 13 , 14 , 15 , 1.6 ij iiiiiliiiiliiiiliiiiliiir 1.1 , 12 , 1,3 . 14 . 1.5 . 16 2 3 4 IP . 1.1 . 1.2 . 13 . 14 . 15 . 1.6 FlQ. 1. — 1. Uranic nitrate. 2. Uranic acetate. 3. Sodio-uranic acetate. 4. Uranic oxychlo- ride (acid), mixed hydrates. 5. Potassio-uranic oxychloride. 6. Uranic oxyfluoride. 7. Bario- uranic oxyfluoride. 8. Uranic phosphate, mixed hydrates. 9. Calcio-uranio phosphate 10. Ammonio-iiranic sulphate. HISTORICAL, y Morton and Bolton, like Becquerel, refer to the possibility of deter- mining the composition of uranyl compounds from the observation of their fluorescence spectra and state that even minute quantities, present as impurities, may be detected by means of their characteristic bands. Hagenbacy likewise published a considerable list of fluorescence bands for the uranyl salts, but his paper adds little to the data of Becquerel and of Morton and Bolton. In 1903, J. Becquerel and Onnes, working in the cryogenic laboratory at Leyden, excited various uranyl salts to fluorescence at the tempera- tures of liquid air and of liquid hydrogen respectively. At —185° C. each band of the spectrum was found to be resolved into a group of much narrower bands. The spectra of a number of compounds were photographed, using a grating spectrograph, and the most prominent bands were mapped. It was shown in the course of this investigation that the resolved spectra are made up of series of bands, the frequency interval varying slightly for different compounds; also that each group in a given spec- trum is similar to all the other groups as regards the arrangement and the relative intensities of its components. In the reversing region, where fluorescence goes over into absorption, the coincidence in posi- tion of bright and dark bands was pointed out. Further cooling to the temperature of liquid hydrogen rendered the individual bands sharper and more line-like, but there was no further resolution. This resolution of the fluorescence spectra by cooling constitutes the most important advance subsequent to the discoveries of Stokes and of E. Becquerel, since it affords a means of studying the more intimate structure of these remarkable spectra. It forms, indeed, the starting- point for the present investigation. 1 Hagenbach, Annalen der Physik., v. 146, p. 395. 1872. 11. THE STRUCTURE OF FLUORESCENCE SPECTRA. A fluorescence spectrum consists of one or more bright bands, and these may greatly vary in width, from the very broad bands, filling a great part of the visible spectrum, characteristic of the fluorescent dyestuffs and the phosphorescent sulphides, to the line-hke bands of the ruby. Such a spectrum is either a homogeneous complex of systematically related components or a heterogeneous complex of unrelated compo- nents. In either case the components frequently overlap, giving the appearance of a single band, which may be described as a mixed band (an unresolved heterogeneous complex) or a homogeneous band, respec- tively. Where the components overlap less completely or not at all the appearance is that of a group of bands. It is probable that a heterogeneous complex is always the result of a mixture of two or more compounds the fluorescence of each of which by itself gives a homogeneous complex. The phosphorescent sul- phides afford spectra which may serve to illustrate the above classification. A stron- tium sulphide with bismuth as the active metal and a flux of sodium sulphate, for example, has a fluorescence spectrum which appears to the eye to consist of a single band with its crest at 0.480 ju. A recent spectrophotometric explora- tion by Dr. H. L. Howes,^ however, shows a group of closely over-lapping compo- nents (see fig. 2). The crests of these are located as shown in table 1; and as they are systematically related, form- ing members of a series having a uniform interval of frequency difference, this is to be regarded as a homogeneous band or homogeneous complex. Similarly, the fiuorescence of a barium sulphide with copper as the active metal and a flux of sodium borate, when viewed through an I L. &K. No. 13. SR. BI.NAaSO*. ■80 rt \ J -60 \ ^ ^0 \ I , » -2D 1 / 1 \ 1 Al .5/4 Fia. 2. * Proceedings American Fhiloaophical Society, lvi, p. 258. 1917. 10 GENERAL DISCUSSION OP FLUORESCENCE SPECTRA. 11 ordinary spectroscope, has a spectrum which seems to consist of a single very broad band. A spectrophotometric study reveals, how- ever, two neighboring and overlapping bands. These have their crests in the red and green respectively and are complex. (See fig. 3.) Table 1. — Approximate wave-lengths of visible crests in the spectrum of a phosphorescent strontium sulphide (Sr; Bi; Na2S0), No. 13. fX, Visible crests 1^X103. Intervals. /i. Visible crests 1/mX103. Intervals. 0.4430 .4547 .4670 .4801 .4938 2257 2199 2141 2083 2025 58 68 58 58 0.5238 1909 2X58 .5562 1793 2X58 .5921 1677 2X58 The components of the band in the green are members of a series having a constant frequency interval of 70 (see table 2), while the components of the band in the red form a series with an interval of I \ L&K. No. 33. BA. cu. NA^ B4 07 -15 ^ ! 1 >i 1 -10 ^J i V ; t v t H 1 / 1 1 \ Ai^ rS/i .6m Fig. 3. 26.6. The two series overlap, as may be seen from figure 3. In this example the spectrum as a whole forms a heterogeneous complex made up of two homogeneous complex bands which are partially super- imposed. The fluorescence spectrum of commercial anthracene affords an example of a heterogeneous complex easily resolved into a group of 12 FLUORESCENCE OF THE URANYL SALTS. bands. There are at least 7 such bands, 4 of which are seen in the spectroscope with a region in the violet not readily resolved by visual observations. This violet fluorescence has, however, been determmed photographically by Miss McDowell.^ The approximate location of the bands in this spectrum is shown in figure 4. Table 2. — Approximate wave-lengths and frequencies of visible crests in the spectrum of a phosphorescent barium sulphide (Ba; Cu; NaaBiO?). Green complex. Red complex. Visible crests. Intervals. l//xX10^ Visible crests. M- Intervals. 0.4255 .4386 70 0.5000 26.6X2 70X2 70 70 .5136 "'"!5283' .4673 .4831 .5000 26.6X2 26.6X7 70X2 .5861 .5376 26.6X2 .6049 26.6X2 .6250 26.6X2 26.6 .6465 .6578 .6695 26.6 While all of these bands are present in the fluorescence of the impure commercial product, they are not aU due to any one constituent. By solution and subsequent fractional sublimation, as is weU known, it is possible to partially separate the substance into pure anthracene, which has a violet fluorescence and a residue containing chrysogen, the fluorescence of which is green. Fig. 4. Miss McDowell has shown that the bands 6, 7, and 8 belong to the anthracene thus obtained, while band 4 is also present in its spectrum. Bands 1, 2, 3, and 4 are characteristic of the green residue. 1 Miss L. S. McDowell, Physical Review (1), xxvi, p. 155. 1908. GENERAL DISCUSSION OF FLUORESCENCE SPECTRA. 13 The presence of band 4 in both spectra may be due to the imperfect separation of the two substances. That it is much stronger in spectra obtained from mixtures showing green fluorescence than from those the fluorescence of which is blue-violet would seem to warrant ascribing it to the chrysogen component, a conclusion strengthened by the con- sideration of the placing of the bands. The most probable positions of the crests are given in table 3. The positions of the three bands assigned to anthracene are from photographic measurements by Miss McDoweU;^ those due to chrysogen are from spectrophotometric read- ings made in 1910,^ combined with more recent observations. Table 3. — Wave-lengths and frequencies of the bands of commercial anthracene. Band. /i. l/fxXlOK Difference. 1 0.6235 .5790 .5362 .5005 .4750 .4490 .4260 1603 1733 1860 1996 2105 2227 2347 130 127 136 2 3. Chrysogen . . . 4 5. . . , 122 120 6. Anthracene. . . 7 It will be noted that the three bands in the blue-violet are members of a series having a frequency interval of about 121; also that the 4 bands of greater wave-length form a series with a somewhat greater interval, i. e., about 131. Band 4 is too near to band 5 to belong to the anthracence series, but may, within the rather large errors due to the breadth and vagueness of these bands, be regarded as one of the chrysogen series. There are several criteria based on experimentally estabhshed facts by which the homogeneity or heterogeneity of a fluorescence band or complex may be determined. CRITERIA OF HOMOGENEITY. (1) The position and distribution of intensities in a homogeneous band is independent of the mode of excitation. This was established by various observations published several years ago,^ and subsequent experience strengthens our conviction that it is a general principle and that shifts in position and change of form are to be regarded as indica- tions of heterogeneity due to the presence of more than one lumines- cent substance. (2) The distribution of intensities in a homogeneous band is such that the curve has a single well-marked maximum. The slope toward the violet is steeper than that toward the red, like that in the corre- sponding curve of intensities of an incandescent black body. 1 Miss L. S. McDowell, 1. u. ^ Nichols, E. L., Proc. Am. Philos. Soc, xux, p. 277. ^ See Nichols and Merritt, Studies in Luminescence, Carnegie [Inst. Wash. Pub. No. 152, pp. 24, 38, 144. 14 FLUORESCENCE OF THE TJRANYL SALTS. In the case of a partially resolved or wholly resolved homogeneous complex the envelope obtained by drawing a curve through the crests of the group of bands has the above form, as has the curve of intensities of each of the component bands. A departure from this type indicates heterogeneity. Thus, for example, the curve in figure 2 suggests a partially resolved homogeneous complex, while that in j&gure 3 indicates a heterogeneous complex or mixed band. The best examples thus far are found among the uranyl salts. Figure 5 shows a typical case in which the envelope of the 7 bands of a uranyl salt is shown and with an enlarged scale of wave-lengths the distribu- tion of intensities of a single band of the same spectrum. Other illustrations will be found in subsequent chapters of this monograph. (3) In a homogeneous com- plex, the fluorescence spectrum is identical with that observed during phosphorescence as regards the position, relative intensity, and structure of its component bands. Nor is there any change in these respects during the process of decay. Change of color in passing from the fluorescent to the phos- phorescent stage or during phosphorescence is therefore a criterion of heterogeneity, since such changes are due to the presence of bands having different rates of decay. Such subjective changes of color as are due to the loss of intensity during decay are excluded from the above statement. Most of the phosphorescent sulphides afford examples of heterogene- ity clearly indicated by the above criterion and confirmed in other ways, while the spectra of the uranyl salts, in spite of their great complexity, are found, from this criterion, too, strictly homogeneous. (4) Persistence of color and of structure when excited to fluorescence at different temperatures, the different components of the spectrum suffering the same relative changes of intensity, may be regarded as a criterion of homogeneity, but the complex changes of structure revealed by the resolution of spectra in the process of cooling to the temperature of liquid air do not necessarily indicate heterogeneity. As will be shown in subsequent chapters, the fluorescent spectra of the uranyl salts, for example, are profoundly modified by the cooling of the substance, and yet these spectra conform to all other known criteria of homogeneity. Lo. ■*** 4io/* .5^ (\ .s6o/A ■' 1 5-0 / \ 1 \ -■' / \ ^ A 1 \ 20.00 . , 1 2^° . 1^0 . 18,00 a^oo , 1 Fig. 5. III. PRELIMINARY OBSERVATIONS ON CERTAIN URANYL SALTS.i Because of their brilliant luminescence and the interesting character of their spectra of fluorescence and absorption, the uranyl compounds have been the subject of extended study. A brief account of the work of previous observers in this field has been given in Chapter I. Our original purpose in taking up the study of these substances was to determine whether the different bands of the fluorescence spectrum are to be regarded as independent, each with its own region of excita- tion, or whether they form a homogeneous complex, such that the excitation of one necessarily involves the excitation of all. In this inquiry we have been led to the investigation of many other questions. Since, as was first shown by C>0 n L.B. Becquerel and Onnes^ in the paper cited in Chapter I, the bands of the uranyl salts are resolved into groups of narrow components by cooling, it is at the temperature of liquid air and chiefly by photographic methods that the intimate structure of the fluo- rescence and absorption spectra is to be determined. The study of the spectra at ordinary tem- peratures, however, is not without significance. In this work, where the width of the bands is from 50 to 100 A. u., the spectrophotometer is indispensable. Many of the measurements to be described were made with a special in- strument which combines the features of the constant deviation spectrometer and the Lummer- Brodhun spectrophotometer. It is essentially a spectrometer of the Hilger type, with two collimators C and C (fig. 6), a Lummer-Brodhun cube L. 5., and a constant-deviation prism P with carefully calibrated drum. ^_^ 1 Certain of the observations contained in this chapter have been published in the Physical Review (1), xxxiii, p. 355, but many of the data there given have been replaced by more complete investigations kindly done at our request by Dr. Frances G.Wick (Physical Review (2) , v. 11, p. 121. Feb. 1918. 2 BecQuerel and Onnes, Leiden Communications, 110. 1909. 15 Fig. 6. 16 FLUORESCENCE OF THE URANYL SALTS. For the detenniaation of wave-lengths the eyepiece is provided with a pointer in the focal plane and also with the usual slides for isolating the region under observation. The collimator slits have micrometric adjustment, and to provide for convenient comparison through the very great range of intensities occurring in the study of fluorescence, the illumination of the com- parison slit can be varied by moving the comparison light along a photometer bar to any desired distance from the slit. The observing telescope can be replaced by a camera whenever photographs of the spectra are desired. With this instrument the wave-lengths of the bands could be determined by setting the pointer to the region of greatest brightness as estimated by the eye and the relative intensities could be measured spectrophometrically. Tables 4 to 10 contain the resulting data for several salts; also the frequencies corresponding to the wave-lengths and frequency intervals. The measurements and computations were kindly made by Miss Wick, who likewise determined the relative brightness of the bands in several of the spectra. From the data in these tables some of the salient features of the uranyl spectra may be deduced, viz : (1) The weakest bands are at the ends of the spectrum, i, e., in the red and the blue. (2) The brightest band is not in the center, being third from the violet end and sixth from the red end when all 8 bands of the spectrum are visible, (3) Taking the frequency intervals, instead of the differences of wave-length, the bands, with the exception of the band of shortest wave-length (band 8), are equidistant, at least within the rather large Table 4. — Fluorescence hands of the nitrates. Uranyl (N03)2) 70 A. u nitrate (Anhydrous); (UO2 . Width of bands about Uranyl nitrate (tri-hydrate) (UO2 (N03)2+3H20). Observations on a single large crystal; width of bands about 100 1. U. Uranyl nitrate (hexahy- drate) (U02(N03)2+6 H2O). Position of crest of band. 1/mX103. Inter- val. Inten- sity. Position of crest of band. l/iuX103. Inter- val. Inten- sity, Position of crest of band. l/fiXW. Inter- val. 4720.0 2118.6 65.8 4871.0 2052.9 84.3 49.2 4720.8 2118.2 69.8 4871.2 2052.8 88.2 50.2 5079.7 1968.6 86.8 100.0 4881.8 2048.4 86.4 5090.0 1964.6 87.6 100.0 5314.0 1881.8 87.6 64.5 5096.3 1962.0 86.1 6327.6 1877.0 87.4 62.7 5573.4 1794.2 87.4 20.0 5330.6 1875.9 87.3 6587.6 1789.6 87.2 25.1 5859.0 1706.8 84.6 5591.2 1788.6 87.2 6874.0 1702.4 84.7 6164.3 1622.2 5877.5 1701.4 85.0 6181.2 1617.7 6186.5 1616.4 FLUORESCENCE AND ABSORPTION OF THE URANYL SALTS. 17 Table 5. — Fluorescence hands of the double niiTOtes. Ammonium uranyl nitrate (/3), U02(N03)2-NH4 NO3. (Accuracy of setting not to be expected; the double crest (a and 6) of each band is flat of nearly equal intensity, rather broad, and not well separated.) Potassium uranyl nitrate, U02(N03)2-KN03. (Bands broad and fuzzy; over 100 units wide; crests distinctly double, with a third crest toward the violet.) Position of crest of band. i/m+io^ Interval. Position of crest of band. 1//XX103. Interval. a-a 6-6 a-a. 6-6. a-4639.6 6-4705.3 O-4800.0 6-4841.6 0-4992.6 6-5048.3 a-5224.0 6-5280.6 a-5473 . 6-6535.6 a-5748.7 6-5825.6 a-6026. 6 6-6130.8 2155.3 2125.0 2083.3 2065.4 2003.0 1980.8 1914.2 1893.6 1827.1 1806.5 1739.5 1716.5 1669.3 1631.0 a-4696.0 a-4860.0 6-^904.0 a-6077.5 6-5108.3 a-6303.5 6-5347.0 a-5556.0 6-5607.3 a-5836. 5 6-5891.5 a-6146.3 6-6209.0 2129.5 2057.6 2039.0 1969.4 1957.5 1885.5 1870.2 1799.8 1783.5 1713.3 1697.3 1627.0 1610.6 71.9 72.0 59.6 88.2 81.5 80.3 84.6 83.7 87.3 88.8 87.2 85.7 86.7 87.1 87.1 86.6 86.2 87.6 90.0 8ff.3 86.7 80.2 85.5 Table 5. — continued. Rubidium uranyl nitrate, U02(N03)2'RbN03. (Bands very broad with two poorly defined crests. With less excitation the bands appear single. Settings are only roughly approximate.) Position of crest of band. 1//XX103. Internal. Intensity. a-a. 6-6. 6-6. 6-4828.0 a-4994.5 6-5046.6 ar-5210.7 6-5265.3 a-5464.6 6-6525.0 a-5738.8 6-6820.5 ar-6012.0 6-6119.0 2071.0 2002.2 1982.0 1919.2 1899.0 1833.3 1810.0 1742.6 1718.0 1663.0 1634.0 89.0 68.3 83.0 83.0 100.0 85.9 89.0 39.4 90.7 92.0 7.6 79.6 84.0 2.1 Table 6. — Fluorescence hands of the sulphate. Uranyl sulphate (UO2SO4-I-3H2O). Position of crest of band. 1/tiXlOK Interval. Intensity- a-a. 6-6. a. 6. a- 6-0.4763 a- .4886 b- .4941 a- .5099 6- .5158 a- .6335 6- .5397 a- .5692 6- .6662 a- .6877 6- .5962 a- .6183 6- .6284 2103.9 2041.7 2023.9 1961.0 1938.7 1874.4 1852.9 1788.3 1766.2 1701.5 1677.3 1617.3 1590.2 80.0 85.7 86.2 27.97 38.72 86.6 85.8 91.79 100.00 86.1 86.7 50.04 53.79 86.8 88.9 19.27 20.14 84.2 87.1 18 FLUORESCENCE OF THE URANYL SALTS, uncertainties inevitable in the attempt to locate the crests of such broad bands. Of this band, which occupies the region Ijdng, roughly, between 0.4650 m and 0.4750 m, only the less-refrangible edge is seen, the other side being more or less cut off by absorption. Its apparent distance from band 7 is thus reduced. (4) In some cases there is sufficient evidence of resolution to enable the location of two or more crests. Further evidences of complexity will be found in the spectrophotometric study of these spectra, to be considered in a subsequent paragraph. Table 7. — Fluorescence hands of the double sulphates. Sodium uranyl sulphate (U02S04.Na2S04). Potassium uranyl sulphate (UO2SO4.K2SO4). Position o f crest of band. 1/mX103. Interval. Intensity. Position of crest of band. l//iXl03. Interval. Intensity. 4744.0 4910.0 5125.0 5354.2 5608.4 5890.5 6200.2 2107.9 2036.6 1951.0 • 1867.6 1783.0 1697.6 1612.8 71.3 85.6 83.4 84.6 85.4 84.8 4778.0 4935.5 5133.6 5365.6 5619.2 5902.9 6201.2 2093.0 2026.1 1948.0 1863.7 1779.5 1694.3 1612.6 66.9 78.1 84.3 84.2 85.2 81.7 27.29 100.00 46.77 15.13 5.88 35.33 100.00 41.22 12.85 Ammonium uranyl sulphate (UO2SO4. (NH4)2S04). (The two crests of each band very close and narrow.) Rubidium uranyl sulphate (UOiSOi-RbaSG^). Position of crest of band. 1/mX103. Interval. Intensity. Position of crest of band. VmXIO^. Interval. Int,ensity. a-a. b~b. 0-4929.0 6-4950.0 a-5140.8 6-5164.8 a-5374.5 6-5399.5 a-5627.0 6-5657.8 a-5906.5 6-5935.5 a-6213.7 2028.8 2020.2 1945.2 1936.2 1860.6 1852.0 1777.0 1767.5 1693.0 1684.8 1609.3 83.6 84.6 83.6 84.0 83.7 84.0 84.2 84.5 82.7 85.2 25.5 4757.0 4930.0 5136.0 5368.7 5619.8 5894.0 6195.5 2102.2 2028.3 1947.4 1862.6 1779.4 1696.6 1614.1 73.9 80.9 84.8 83.2 82.8 82.5 38.52 100.00 49.35 19.77 100.00 21.43 11.86 6-6251.4 1599.6 1 1 FLUORESCENCE AND ABSORPTION OF THE URANYL SALTS. 19 Table 7. — Fluorescence hands of the double sulphates- continued. Uranyl cesium sulphate (UO2SO4.CBSO4). (Bands over 100 units wide with sharp maxima near the end toward red and secondary maxima (brack- eted) much less bright and sharp toward the violet.) Wave-length of crest of 1/mX103. Interval. Intensity. band. 4702.0 2126.7 75.8 4876.0 2050.9 50.49 (4933.0) 85.7 5088.5 1965.2 100.00 (5151.5) 84.4 5317.0 1880.8 53.09 (5338.0) 86.6 5573.5 1794.2 16.90 (5670.0) 85.4 5852.0 1708.8 86.8 6.36 6165.0 1622.0 RELATIVE INTENSITIES OF THE BANDS. To indicate graphically the relative intensities of the bands, we may plot their strength, ex- pressed in terms of energy, as ordinates and wave-lengths of the crests as abscissse. The resulting curve (see figs. 7, 8, 9, and 10) is a sort of enve- lope for the entire spectrum cor- responding to the curve of distri- bution of energy. It resembles in type the curve of energy found in the case of the broad, single- banded fluorescence described in earUer communications ,^ being single-crested and steeper toward the violet. As has been pointed out in Chapter II, these curves are very similar to the energy curve for temperature radiation. Figure 7 contains the envelopes thus plotted of 5 uranyl double sulphates. Of these, 4 have their crests at approximately the same wave-length (0.515 m). Curve E (csesium uranyl sulphate) is shifted slightly, an effect due to the presence of a strong component of each band on the violet side of the main crest which influences the estimates ^ Nichols and Merritt, Physical Review (1), xviii, p. 403; xix, p. 18. Table 8. — Fluorescence hands of two acetates. Uranyl acetate. Position of crest of band. l/juXl03. Interval. Intensity. 4710.0 4878.0 5094.0 532S.0 5586.0 5869.2 6182.3 2123.0 2050.0 1963.1 1876.9 1790.2 1703.8 1617.5 73.0 86.9 86.2 86.7 86.4 86.3 48.16 100.00 48.86 22.36 Ammoniima uranyl acetate. Position of crest of band. 1/mX103. Interval. Intensity. a— a. 6-6. 6-4680.0 a-^804.5 &-4884.3 a-5016.3 fe-5094.6 0-5242.2 6-5330.8 a-5487.8 6-5581.0 a- 6-5870.6 a— 6-6207.5 2136.0 2081.3 2047.4 1993.5 1962.9 1907.5 1875.9 1822.2 1791.7 87.8 86.0 85.3 88.6 84.5 87.0 84.2 88.3 92.5 61.42 100.00 46.12 18.40 1703.4 1610.9 20 FLUORESCENCE OF THE URANTL SALTS. Table 9. — FliioresceTice hands of two phosphates. Uranyl phosphate (H-U02-P04). (Bands narrow and distinct.) Ammonium uranyl phosphate (H,(NH4)2U02(PO*)2). (Bands very distinct with narrow crests.) Position of Position of crest of 1/mX103. Interval. crest of l//zXlO^ Interval. band. band. 4847.0 2063.0 71.7 4845.0 2063.9 69.7 5020.3 1991.9 83.8 5014.6 1994.2 82.7 5240.7 1908.1 83.7 5231.3 1911.5 83.2 5481.1 1824.4 84.9 5469.6 1828.3 83.7 5748.6 1739.5 82.9 5730.8 1744.6 83.8 6036.5 1656.6 6021.0 6336.0 1660.8 1578.3 82.5 Table 10. — Fluorescence bands of a nitrate, oxalatej and fltioride. Uranyl nitrate Uranyl oxalate Potassium uranyl fluoride (U02(N03)r6H20). U02C204'3H20. (UO2 F2-2KF). Position of Position of Position of crest of i/mxio^ Interval. crest of X/fiXlO\ Interval. crest of l//iX10=». Interval, band. band. band. 4700.0 2127.6 72.1 4715.0 2120.9 75.1 " 4865.0 2055.5 89.3 4888.0 2045.8 92.3 4803.2 2081.9 79.8 5085.8 1966.2 86.6 5119.0 1953.5 86.3 4994.8 2002.1 86.8 5320.0 1879.6 88.3 5355.0 1867.2 86.9 5219.5 1915.3 85.6 5582.3 1791.3 87.0 5'617.0 1780.3 88.2 5465.2 1829.7 89.1 5867.0 1704.3 86.2 5910.0 1692.1 88.2 5745.0 1740.6 89.1 6179.8 1618.1 6235.0 1603.9 6055.0 1651.5 of the location of the latter. The envelopes of the 4 nitrates in figure 8 have the same characteristics. The crests of 3 agree (at 0.510 /x), while the envelope D of rubidium uranyl nitrate, in the spectrum of which the bands are vaguely double-crested, is displaced. In the two uranyl acetates (fig. 9) the same identity of type and position shows itself. To reduce these spectrophotometric measurements to relative energy units the distribution curve of the comparison Ught must be known. This curve for the acetylene flame, which was the source employed, has been carefully determined, and data pubhshed by Coblentz were used in the computation. FLUORESCENCE AND ABSORPTION OF THE URANYL SALTS. 21 In certain cases the resolution of the bands is such that the brightness of two crests can be determined and two overlapping envelopes drawn, as in figure 10, which pertains to the spectrum of UO2SO4+3H2O. The relative brightness of the two crests is seen to vary sUghtly from band to band. -T .50 /^ ibIoo lOlOO 2olqo Fig. 7. Fig. 8. The changes in the position of the crests in the case of the sulphates, nitrates, and acetates is illustrated in figure 11, in which a typical curve from each family of salts is given. While these measurements by Dr. Wick do not include all the spectra for which such determinations are possible, they suffice to demonstrate the essential uniformity of type of the envelopes and to show that within a given family, such as the double sulphates or the nitrates, the crests occupy the same region in the spectrum. It will be seen, when we come to the consideration of the detailed structure of these fluores- cence spectra, that there is a slight but definite shift of all the bands with molecular weight. Spectrophotometric measurements of single unresolved bands, prac- ticable with accuracy only in the case of some of the brightest, show the 22 FLUORESCENCE OF THE URANYL SALTS. curve of distribution to be of the same type as that obtained when the envelope is drawn for the entire spectrum, t. 6., the type associated with what we have termed a simple band. (See A in fig. 5, Chapter II, which is the energy curve for the brightest band of uranyl potassium sul- FiG. 9. Fig. 10. phate with the scale of wave-lengths, adjusted so as to make the width nearly the same as that of the envelope iB) for the same substance.) The most striking feature ' distinguishing these spectra ri I T" from one another to the eye, exceptiQg where partial resolu- tion occurs, is the varying width and sharpness of the bands. With the spectrophotometer it is possible to obtain a more definite expression of this fea- ture, as may be seen from fig- ures 12 and 13, in which are depicted, from such measure- ments, the three brightest bands of uranyl nitrate (crystalUzed) and uranyl potassimn sulphate. It wiU be noted that the bands overlap at the base, but to a greater extent in the nitrate than in the potassium sulphate, where the bands are narrower and more sharply defined. Fig. 11. FLUORESCENCE AND ABSORPTION OF THE URANYL SALTS. 23 A more detailed use of the instrument, working with narrow shts and making settings at closer intervals, will often bring out the complexity of single bands, where the overlapping of the components is such as to conceal the structure Figures 14, 15, 16, and 17 give the results of such a study by Miss Wick. The existence of numerous partly sub- FiG. 12. Fig. 13. Showing the relative intensities of the brightest fluorescence bands of uranyl nitrate (fig. 13) and uranyl potassium sulphate (fig. 14). merged crests is apparent in the curves, corresponding to the com- plexity of structure which these bands show when the substance is excited at the temperature of liquid air. The vertical lines indicate the position of the bands, as observed at low temperatures by methods to be described in subsequent chapters. That these hnes in general do not co- incide with the positions of the crests might seem to indicate that there is no definite relation between the spectra at the two temperatures or that the accu- racy of the curves is in doubt; but the discrepancies are quite in accordance with the results obtained by the detailed study of the spectra of the double chlorides (Chapter \), where the spectra are sufficiently resolved at +20° C. to enable us to trace the changes on cool- ing, measure the definite shifts, and dis- cover the remarkable mechanism of the process of resolution. p^^ ^^ I zo|oo l»|50 —20 • 1 — IS \ — 10 _ J 1 1 1 1 1 49 .50 .51 A 24 FLUORESCENCE OF THE URANYL SALTS. EXCITATION BY LIGHT OF DIFFERENT WAVE-LENGTHS. If all the bands of the luminescence spectnim are due to the vibra- tions of a single connected system it would be natural to expect that an agency which excited one would also excite the rest, especially if luminescence is due to the recombination of ions dissociated by the exciting hght, or to the return of an electron set free by the exciting I 20|00 1 1 A! l»|00 — r ^vfk-L J — 20 f \ V 10 1 1 1 1 1 1 1 Fig. 15. sT/T >9 ^0 Fig. 16. -TTjr agency. On the other hand, if each band is dne to some process going on in one particular compound or molecular aggre- gation, wave-lengths might be found which would excite one band and not the rest, or which would at any rate excite the bands in different degree. To test this matter we have measured the distribution of intensity in the bands for excitation by different hues in the ultra-violet spectrum of the quartz mercury lamp. The intensity of fluo- rescence with this excitation is not suffi- cient to permit the measurement of all the bands, so that the three brightest bands only have been measured. In table 11 the intensities for excitation by the different lines in the mer- cury spectrum are given for five different uranyl salts. Curves show- ing the variation of the relative intensity with the wave-length of the exciting hght are shown for uranyl-nitrate cr^'stals in figure 18, and for the double sulphate in figure 19. In each case the intensity of the most intense band has been put equal to 10. The variation was greater id the case of the double sulphate than in the case of any other salt studied. The observations were repeated in the case of this sub- stance on two different days and a comparison of the fuU and dotted I 20|00 L I9|50 '■ / M —20 / \ — 15 / y \ — 10 \ 1 1 1 1 1 ^ .51^ Fig. 17. FLUORESCENCE AND ABSORPTION OF THE URANYL SALTS. 25 curves indicates the extent to which the results agree. In the case of the other salts studied, curves very similar to that of figure 18 were obtained. Table 11. — Relative intensity of excitation of the three brightest hands by five different lines in the spectrum of the mercury arc. Wave- Intensity^ of luminescence length of at crest. Ratio a/b. Ratio c/h. exciting light. Band a. Band h. Band c. Uranyl-potassium sulphate; 0.436m 12.5 22.7 7.45 0.55 0.33 .407 12.6 26.4 8.45 .48 .32 Band a 4,920 .366 10.5 23.2 6.65 .45 .29 b 5,130 .313 16.25 22.9 8.27 .71 .36 V 5,360 .254 8.83 13.9 4.25 .64 .31 Uranyl phosphate: .436m 10.0 10.2 3.07 .98 .30 .407 6.04 6.06 1.90 1.00 .31 Band a 5,015 .366 .313 9.7 11.1 9.7 10.4 2.47 3.04 1.00 1.07 .25 .29 h 5,?39 c 5,483 .254 6.85 5.35 1.28 Uranyl nitrate (anhydrous) : .436m 22.9 38.7 15.5 .59 .41 .407 15.3 22.7 9.30 .67 .41 Bando 4,849 .366 15.3 20.7 8.87 .47 .43 b 5,071 .313 21.8 31.6 11.7 .69 .37 V 5,311 .254 6.0 7.7 3.2 .78 .42 Uranyl nitrate (crystals) : .436m 18.6 27.2 10.0 .69 .37 .407 17.5 24.5 8.6 .73 .35 Bando 4,869 .366 12.2 18.3 7.5 .67 .41 b 5,086 .313 17.3 25.6 9.6 .67 .38 c 5,329 .254 8.3 10.5 4.0 .79 .38 Uranyl fluorid fluor- ammonium: .436m 10.3 11.2 3.7 .92 .33 .407 23.3 25.0 7.7 ,93 .37 Band a 5,008 .366 20.4 25.7 7.6 .79 .30 b 5,237 .313 37.5 43.4 11.5 .87 .27 ^ 5,460 .254 15.8 17.6 5.7 .90 .32 'The intensities given in table J 1 are not corrected for energy distribution in the acetylene flame. It will be noticed that the lower curve in figures 18 and 19 indicates a very nearly constant ratio between the intensity of the brightest band and that of the band lying next in the direction of the red. But if we compare the brightest band with the band lying next to the violet side we find a considerable variation intheratio of intensities, especially in the case of the double sulphate. It appears tons probable that this variation is the result of a partial absorption of the luminescence by the substance. The absorbing power of a given salt differs for the different mercury lines used, so that in some cases the exciting light may pene- trate much further into the substance than in others. It is clear that those bands for which the absorption is greatest will appear relatively weaker when the exciting light penetrates a considerable distance into the substance, even if the relative intensity of the excitation of the different bands is really the same for all wave-lengths of the exciting 26 FLUORESCENCE OF THE URANYI. SALTS. light. The observed distribution of energy would correspond with the actual distribution only in case an excessively thin layer of the sub- stance is excited — so thin that the absorption of the Ught emitted is neg- Ugible. As a matter of fact, the band lying to the ^dolet side of the maximum is in a region where the absorption is considerable, while the brightest band and those lying to the red are in the region where the absorption is small. The constancy of the ratio in the case of the lower curves, and the small variation of the ratio shown by the upper curves, are therefore entirely consistent with the view that the observed variations are the result of absorption, and that the first effect of excitation, whatever may be the wave-length of the exciting hght, is to produce aU of the bands with a definite and constant intensity distribution. Fig. 18. Fig. 19. Relative intensdtaes of the brightest fluorescence bands of uranyl nitrate (fig. 18) and uranyl- potassium sulphate (fig. 19). The intensity of the brightest band is put equal to 10. The upper curve in each figure refers to the band Ij'ing next to the brightest toward the violet. The lower curve refers to the band toward the red. Abscissse give the wave-length of the exciting light. (See table 11.) The observations recorded in the foregoing paragraphs all tend to indicate that the fluorescence spectrum of a uranyl salt is a homo- geneous complex. The envelope is single-crested and has the form typical of a simple band. Neither its position nor form is modified by changing the mode of excitation. To test this conclusion we have made many experiments under widely varying conditions, especially in the way of selective and monochromatic excitation of the resolved spectra, where it should be possible to observe critically the disappearance or enhancement of single narrow components of groups of series. The remarkable effects of selective excitation recorded by Wood in the case of fluorescent vapors might lead to the expectation of similar or analogous changes in the uranyl spectra. All these attempts have thus far been without result, and we are inclined, therefore, to regard the spectrum as a unit and to consider it as a broad, simple band, which unlike the other bands of this type as yet discovered, consists of resolved instead of completely overlapping components. Studies to be described in Chapter IV are in confirmation of this "^^ew in that the criterion for a simple band, based upon the phenomena of phosphorescence, is fulfilled. FLUORESCENCE AND ABSORPTION OF THE URANYL SALTS. 27 THE ABSORPTION SPECTRUM AT ORDINARY TEMPERATURES. The resemblance of the absorption spectra of the uranyl salts to their fluorescence spectra, which is so striking as to have led both E. and H. Becquerel to regard the absorption series as a continuation of the series of fluorescence bands, can be fully investigated only by observations at low temperatures. Since the absorption extends into the ultra- violet, moreover, photographic methods are necessary. The study of the absorption at ordinary temperatures is, however, not without sig- nificance, and the use of the spectrophotometer in this work brings out certain features not easily discernible in the photographic plates. The salts thus studied by us were in powdered form and the location, relative intensity,and character of the bands lying within the visible spectrum were determined by measur- ing the intensity of the light transmitted by an extremely thin layer between glass plates, or in some instances by ob- serving the spectrum of white light reflected from the surface of the powder. Recourse to the latter method is, indeed, frequently necessary because of the great and rapidly in- creasing opacity of these sub- stances in the blue and violet. The nature of the results of such measurements is suffi- ciently shown in figure 20, which is plotted from deter- minations of the light transmitted by a thin layer of uranyl potassium sulphate. The source of light was an acetylene flame. The measurements cover not only a considerable portion of the absorbing region, but also a part of the region containing the fluores- cence bands. Three of these bands show very clearly, even when superposed upon the brilliant continuous spectrum of the acetylene flame. The absorption begins a little on the violet side of the brightest luminescence band and extends into the ultra-violet. It will be noticed that there are several definite and narrow absorption bands, which appear to be superposed upon a broad band, or region, of general absorption. This appearance of a broad band might result from the overlapping of the group of narrow absorption bands, only the crests of which can be observed. In estimating the relative intensity of the Fig. 20, — Transmission of a thin layer of uranyl- potassixim sulphate, showing absorption bands and three of the fluorescence bands. Curves F and A show the relative intensities of the bands of fluorescence and absorption respectively. 28 FLUORESCENCE OF THE URANYL SALTS. absorption bands we have adopted the first view and have assumed a general absorption^ such as is indicated by the dotted line of figure 20. The deviations from this dotted curve have been ascribed to the effect of the narrow bands. The intensity of each band is determined by taking the ratio of the diminution of the transmission which it produces to the transmission which would be expected if the general absorption only were present. Both the absorption bands and the fluorescence bands have been indicated in figure 20 by lines whose lengths are proportional to the intensities of the bands. If a Une is drawn through the ends of the lines that give the intensity of the absorption bands a curve (A) is obtained which is very similar in form to the absorption curve for a substance having a single broad band. This curve also has the same position with reference to the envelope of the luminescence bands (F) that the absorption curve in such cases has to the luminescence curve. It appears highly probable that just as a broad luminescence band may result from the overlapping of a group of bands, so the absorption of the same substance may result from the overlapping of a similar group of absorption bands. The transmission curve for a thin layer of powdered uranyl sulphate is shown in figure 21, the source of hght being an acet3^ene flame. In its general features this curve is similar to that for the double sulphate of uranyl and potassium. The fluorescence of the sulphate is not so brilliant and the fluorescence bands therefore show less prominently. The sulphate, as has been shown in a preceding paragraph, has the peculiarity of possessing two series of fluorescence bands lying close together, one set of bands being much more intense than the other. It will be noticed that the absorbed bands are also double. If we think of the more intense luminescence bands as constituting the principal series and the less intense bands forming a secondary series, a curious reversal is noticeable as we pass from the region of fluorescence to the region of absorption. Each band of the principal series in the lumines- cence region lies a little to the right of the corresponding band of the secondary series. The positions of the bands are indicated by short vertical lines in the lower part of figure 21, the bands of the secondary series being represented by dotted lines. When we pass to the absorp- tion series, however, the more intense band lies to the left in each case. For example, the absorption band at 4,925 corresponds in position with a fluorescence band of the principal series; but the absorption band at 4,880, which probably corresponds to the band 4,890 of the secondary fluorescence series, is by far the more intense of the two, ^ The fact that all the uranyl salts, so far as known, increase rapidly in opacity as the wave- length of the transmitted light decreases, even when the bands are greatly reduced in width by cooling, seems conclusive as to this assumption. FLUORESCENCE AND ABSORPTION OF THE URANYL SALTS. 29 It will be observed that the absorption bands of uranyl potassium sulphate occurring at 4,760 and 4,920 (fig. 20) appear to coincide in position with two of the luminescence bands of the same substance. In other words, these two bands are ''reversible" and may appear either as absorption bands or as luminescence bands, according to the con- ditions under which they are observed. The double sulphate thus shows the same phenomenon that was first described by H. Becquerel^ in 1885 in the case of uranyl nitrate. f\ /^ r \ /^AT V 10 f v >. / t\ J (\ 5 /\ / f" J J 1 1 1 1 1 1 f t t \ 1 1 1 1 .45 Fig. 21.- -so .55, -Transmission of a thin layer of uranyl sulphate. It was, however, of interest to study these relations in the case of the uranyl spectra at ordinary temperatures also. Special precautions were necessary, for when a luminescence band occurs in a region where there is appreciable absorption it is clear that the apparent position of the crest of the band may be influenced by absorption in case the latter is not uniform. Where measurements of absorption are made with light containing rays that are capable of exciting fluorescence there may also be a displacement of the crest of the absorption band, owing to the presence of luminescence. There could be no displacement of 1 Comptes Rendus, vol. 101. p. 1252. 1885. 30 FLUORESCENCE OF THE URANTL SALTS. this sort in case the light emitted were strictly proportional to the coefficient of absorption; but if the fluorescence band and the absorp- tion band do not exactly coincide in position or in form, such a dis- placement is to be expected. In order to avoid the necessity of changing the adjustment of the spectrophotometer, or the position of the substance, between measure- ments a thia layer of the uranyl potassium sulphate was in some cases mounted permanently in front of the sUt. To locate the absorption bands the sUt was illuminated, through the specimen, with Ught from an acetylene flame. To observe the luminescence bands a piece of blue glass was placed in front of the flame, so as to cut off the rays having the same wave-length as the bands, while permitting the exciting rays to pass ; or in some cases the acetylene flame was replaced by a mercury arc. To guard against the presence of fluorescence in measurements of absorption a green glass was sometimes used. With the relatively thick specimen first used the absorption was so great that the band at 4,760 could not be observed. The band at 4,920 was well defined, howe\'"er, and could be accurately located. If the eyepiece pointer was set at the crest of the absorption band and the source of Ught then changed so as to bring out the fluorescence band, the latter was seen to be very obviously displaced toward the red. Photographs of the absorption and fluorescence spectra taken on the same plate also showed the relative displacement of the two bands very clearly. The wave-length of the fluorescence band as measured imder these conditions was not the same, however, as that previously deter- mined, and the whole appearance of the band was different from what had been observed when looking at the front surface of the luminescent substance. More definite conditions for observing the absorption band were obtained by using nearly monochromatic light for transmission meas- urements. The spectrum of a Nemst glower was formed by a large spectrometer and a small region of the spectrum was isolated by means of a suitable screen containing a sUt. The fight coming through this slit, after passing through the specimen to be studied, fell upon the sfit of the spectrophotometer. By suitable adjustment the center of the band of transmitted fight could be made practically coincident with the center of the absorption band and the latter could be located with considerable accuracy. Under these circiimstances the transmitted fight contained no rays capable of exciting any observable fluorescence, so that we may look upon the determinations of absorption by this method as uninfluenced by errors due to the presence of luminescence. Using a relatively thick layer, the absorption band was located at 4,919, while the crest of the fluorescence band (observed by transmis- sion) lay at 4,974. An excessively thin layer, formed by depositing the salt from a solution, or suspension, in alcohol, gave a fluorescence FLUORESCENCE AND ABSORPTION OF THE URANYL SALTS. 31 band whose crest was at 4,925, while the wave-length of the very faint absorption band was 4,922. Our previous determination of the wave- length of the luminescence band, when looking at the surface exposed to the exciting rays, was 4,920. These results appear to us to warrant the conclusion that if disturbances due to absorption could be entirely eliminated the two bands would be found to have exactly the same wave-length. It must not be forgotten, however, that it is nearly impossible to observe the fluorescence spectrum under conditions which entirely eliminate effects due to absorption. The exciting light always pene- trates to some extent beneath the surface, so that some of the emitted light must pass through the fluorescent material before it can reach the eye. It is natural, therefore, to expect a slight displacement in all cases. Although our most reliable measurement of the wave-length of the absorption band, 4,919, and our best determination of the crest of the luminescence band, 4,920, differ by less than the probable errors of measurement, we feel that it is not unlikely that the difference is a real one, due to the cause just cited. The absorption band at 4,760 in the double sulphate differs in posi- tion by 5 units from the fluorescence band at 4,765. A portion of this difference may also be explained by absorption. But it is probably chiefly due to the difficulty in accurately locating the crests of these bands. The fluorescence band is extremely faint, while the absorption band is not very sharp, because of the large general absorption in this region. Using a thick layer, formed by grinding down a translucent mass of adhering crystals until a piece about 0.5 mm. thick was obtained, a faint absorption band was observed at 5,127. This corresponds to the brilliant fluorescence band at 5,130. In all likelihood the coincidence here is complete, since measurements of the fluorescence band made at the same time and with the same specimen as that used for absorption measurements gave the same wave-length, 5,127, for both bands. EXCITATION BY LIGHT CORRESPONDING TO DIFFERENT PARTS OF THE ABSORPTION REGION. It seemed a matter of some interest to determine the relative effec- tiveness of light of different wave-lengths in producing fluorescence, and experiments having this end in view have been made in the case of the double sulphate. We were particularly interested in determining whether wave-lengths falhng within the sharp absorption bands at 4,918, 4,760, 4,615, etc., were especially effective in exciting lumines- cence. The source of the exciting light used in these experiments was a Nernst glower which was mounted in place of the slit of a spectrometer. The spectrum was focussed upon an opaque screen containing a narrow slit, and the light passing through this slit was used in exciting the speci- 32 FLUORESCENCE OP THE URANTL SALTS. men tested. The fluorescence spectrum was observed in a spectro- photometer, the specimen being set up at an angle of approximately 45^ with th^ path of the exciting light, so that the collimator of the spectrophotometer could be pointed at the illuminated surface without interfering with the exciting light. Enough of the exciting rays were reflected into the spectrophotometer to enable the range of wave-lengths used in each case to be determined. The spectrophotometer was then set at the crest of the principal fluorescence band and the intensity measured by com- parison with an acetylene standard. Observations of this sort were repeated throughout the absorbing region. The results are shown in figm-e 22. It will be noticed that the regions of strong excita- tion at 4,910 and 4,775 correspond very closely to the two absorption bands at 4,920 and 4,766. Some slight indication is also present of the other absorption bands. It is clear, however, that the abihty to excite luminescence is not con- fined to rays falling within the narrow absorption bands, but extends to the region of general absorption lying be- tween. It is not possible to determine the specific exciting power of different rays, as has been done in the case of eosin and resorufin,^ because of our ignorance of the absorbing power of the salt for different wave-lengths.^ The results indicate, however, that the specific exciting power varies only slightly with the wave-length, as in the case of resorufin and eosin. THE RELATION BETWEEN ABSORPTION AND FLUORESCENCE AS IT APPEARS AT ORDINARY TEMPERATURES. In 1885 H. BecquereP made measurements of the spectrum of uranyl nitrate from which it would appear that the frequency interval remains constant in passing from the fluorescence to the absorption spectrum and that the suggestion of E. Becquerel in his classical memoir of 1872, that the emission bands and absorption bands belong to the same series, is in accordance with the facts. H. Becquerel also showed that two of the bands are reversible, ap- pearing as emission bands when suitably excited, whereas if light free * Physical Review, xxxi, p. 381. * The distribution of energy in the spectrum of the Nernst glower also has not been determined. ' H. Becquerel, Comptes Rendus, 101, p. 1252. \ A \ 5 / \ / f / / / .42 .46 50/6 Fig. 22. — Intensity of fluorescence (ordinates) produced by exciting light of different wave-length (ab- Bcissse). FLUORESCENCE AND ABSORPTION OF THE URANYL SALTS. 33 from exciting rays be passed through the substance, absorption bands in the same location are observed. In our own work upon uranyl nitrate and potassium uranyl sulphate we have confirmed the results of H. Becquerel so far as the existence of reversible bands is concerned and have found for these substances 3 such bands instead of 2. The frequency interval between absorption bands, like the fluores- cence interval, is approximately constant, but, as may be seen from tables 12 and 13, it is much smaller. Additional evidence on this point will be found in the chapters dealing with the double chlorides and with the spectra at low tempera- tures, where it will be established as a relation conomon to all uranyl spectra. The study of the absorption spectra at +20° C. is uncertain and unsatisfactory, because we have to do with unresolved groups of bands, and these two examples will suffice to illustrate the remarkable way in which the two frequencies interlock where fluorescence goes over into absorption. Table 12. — Absorption and fluorescence hands of potassium uranyl sulphate at -\-20° C. Absorption. Fluorescence. JU. l/)tiX103. Interval. M- 1//XX102. Interval. 0.4350 2298.9 62.8 .4472 2236.1 68.8 .4614 2167.3 66.5 .4760 2100.8 68.3 0.4765 2098.6 66.1 .4920 2032.5 82.0 .4920 2032.5 83.2 .5127 1950.5 .5130 .5360 .5606 .5881 .6190 1949.3 1865.7 1783.8 1700.4 1615.5 83.6 81.9 83.4 84.9 It will be seen from tables 12 and 13 that the last 3 fluorescence bands, counting from the red, are nearly or quite coincident with the first three absorption bands. Whether or not these coincidences are to be regarded as exact can not be determined from observations on unresolved spectra. It will be demonstrated later that reversals are exact between the ultimate components of bands, but not, in general, between unresolved aggregates. That the fluorescence interval changes to conform to the absorption interval at the last step appears not only from the data in tables 12 34 FLUORESCENCE OF THE URANYL SALTS. and 13, but also in the determinations for other salts (tables 4 to 10) wherever the final fluorescence band (8) has been observed. The corresponding change in the absorption interval to conform with the fluorescence interval is much more difficult to estabhsh, because the last absorption band toward the red is entirely invisible under ordinary conditions. Table 13. — Absorption and fluorescence hands of uranyl nitrate at -\-W° C. Absorption bands.^ Fluorescence bands. M- l/juX103. Interval. jU. 1/mX103. Interval. 0.3830 2610.9 69.6 .3935 2541.3 72.2 .4050 2469.1 71.0 .4170 2398.1 58.9 .4275 2339.2 69.1 .4405 2270.1 72.3 .4550 2197.8 71.9 ,4705 2125.4 72.0 0.4708 2124.0 70.2 .4870 2053.4 87.2 .4869 2053.8 86.6 .5086 1966.2 .5086 .5329 .5585 .5866 .6188 1966.2 1876.5 1790.5 1704.7 1616.0 89.7 86.0 85.8 88.7 * Absorption bands, excepting that at 0.5086 are from measurements by Jones and Strong (Am. Chem. Journal, 1910). EFFECT OF WATER OF CRYSTALLIZATION— BEHAVIOR OF SOLUTIONS. The effects of water of crystallization and the comparison of the spectra of the solid uranyl compounds with those of their solutions are to be treated at some length in subsequent chapters. A few points which have been brought out in the course of our work on the spectra at +20° C. are, however, recorded here. The effect of water of crystallization in the case of uranyl nitrate is to shift the luminescence bands shghtly in the direction of the longer waves. (Compare the hexahydrate with the anhydrous form in table 1.) This is the effect which it would seem most natural to expect, since the mass of the vibrating system is increased by the addition of water of crystallization without any increase, so far as we know, in the elastic FLUORESCENCE AND ABSORPTION OF THE URANYL SALTS. 35 /3'a' /3'« Position of fluorescence and absorption bands of uranyl sulphate. forces of the system. In fact, the presence of water so intimately associated with the salt molecule would probably increase the effective dielectric constant of the region in which the vibrations occur, and would thus cause a decrease in frequency quite independent of any effect due to increase in mass. It has been shown by Deusen^ and by Jones and Strong^ that the absorption spectrum of the crystalhzed nitrate is nearly coincident with the absorption spectrum of the aqueous solution. In many cases no difference can be detected in the wave-length of the band in solu- tion and in the solid crystal. In the case of other bands, however, the difference appears to be too great to be accidental. It seems not unlikely that the absorption spectrum contains several series of bands, some of which occupy almost identically the same po- sitions for the solution as for the solid salt. We must assume, therefore, that at least a part of the dissolved salt has the same molecular structure as the soUd crystals. In the case of the uranyl sulphate studied by us the phenomena are more complicated. As has already been shown, the luminescence spectrum of this salt, even at ordinary temperatures, contains two series of bands, which for convenience we shall designate the a and j8 series respectively. The a bands are by far the stronger and 6 of these could be observed. Of the relatively weak /3 bands only 3 could be seen. In the absorption spectrum of the solid salt 2 series of bands were also found (see fig. 21) which we shall call the a' and j8' bands. Two of the oJ bands corresponded in position with two of the a bands of luminescence, while one band of the /3' series corresponded with one of the jS bands. The wave-lengths are given in table 14 and are shown graphically in figure 23. It is a remarkable fact that while the a bands Table 14. — Uranyl sulphate fluorescence and absorption hands. Fluorescence: Crystals — Principal series (a) 4763 Crystals — Secondary series 03) Dehydrated salt (7) Concentrated solution Absorption; Crystals — a' series 4595 Crystals — ^' series 4555 Concentrated solution ^ Annalen der Physik, 43, p. 1128. 1898. ^ American Chemical Journal, vol. xliii, p. 37, 1889. 4929 5148 5395 4894 5098 5340 4843 5049 5285 4928 5145 ' 6387 4755 4925 4720 4880 4718 4887 6095 5659 6538 5925 1910. See also Vogel, Spectral analyse, p. 270, 36 FLUORESCENCE OF THE URANYL SALTS. are much the brighter in the luminescence spectrum, the ol bands in the absorption spectrum are much weaker than the )3' bands. The sulphate used in this experiment was in the form of small crystals, When the salt was dehydrated by being kept for about an hour in a stream of warm, dry air its luminescence spectrum was found to be absolutely different, each band being shifted toward the violet by about 100 A. u. Brief exposure to the air apparently permitted a portion of the salt to return to the original condition, so that the original a and /3 bands could be seen as well as the 7 bands charac- teristic of the dehydrated salt. In the case of a thin layer of the sulphate which had been dehydrated and then exposed for a short time to the air, each of the lumi- nescence bands was found to consist of three overlapping bands, the components corresponding in position to the a, jS, and 7 bands respectively. Spectrophotometric measurements (with a rather wide sUt) of the brightest luminescence band and of a portion of the absorption spectrum of the same layer are shown in figure 24. In the luminescence spectrum the /3 bands are by far the most prominent,^ while in the ab- sorption spectrum the ol bands are strong- est and no 7' bands can be detected. The results point to the existence of two dif- ferent hydrated salts corresponding to the a and j3 bands respectively, but further study would be necessary to make possible an entirely satisfactory explanation of the observed phenomena. The concentrated aqueous solution of the sulphate showed weak fluor- escence, and the three brightest bands, which could be located with reasonable accuracy, were found to agree in position with three of the a bands of the solid crystallized salt. In the absorption spectrum of the concentrated solution it was possible to locate three well-defined bands, two of which corresponded with two of the /3 bands of the soUd salt (see fig. 25). The solution showed no trace of any fluorescence corresponding to the j8 series, nor did it show any trace of absorption corresponding to the ol series. ^ The a' band appears in fig. 24 to be shifted by about 15 Angstrom units toward the violet; whether this is a real shift, or whether it is due to disturbances caused by simultaneous absorp- tion and luminescence we are unable to say. / 10 / / 1 I J 1 7 A / 1 ./ ^ [ \ / \ / \ / \ Y Q X 1 s (X. .48 .50 .52>y Fig. 24. — TJranyl sulphate (solid), showing the brightest fluo- rescence band at about 0.51 ju and a group of absorption bands at about 0.49 /i. FLUORESCENCE AND ABSORPTION OP THE URANYL SALTS. 37 In a concentrated solution of potassium uranyl sulphate (see fig. 26) three absorption bands were found at 4,910, 4,730, and 4,570. These r 10 Jt / N f i-'V- ./ r 1 5 1 f\ J ; ml i I 1 /3 ' 4 .48 .50/^ ZZZZpf EEgE Fig. 25. — Uranyl' sulphate (solution), show- ing at the left a portion of the trans- mission spectrum for a thin layer and at the right for a thick layer. .44 j4e ^e .50 Fig. 26. — Transmission of a concentrated solution of tiranyl potassium sulphate. do not agree in position with the corresponding bands of the solid salt, which occur at 4,920, 4,760, and 4,472. The solution of the double sulphate shows no trace of fluorescence. IV. PHOSPHORESCENCE OF THE URANYL SALTS. Concerning the phosphorescence of the uranyl compounds, we find little on record beyond the early observations of E. Becquerel/ who, in his classic paper of 1861, noted the brilliant and very short-lived after-glow and made some obser\''ations on the law of decay. For the study of the phenomena of phosphorescence in these sub- stances and in other cases having a duration of glow of a few thou- sandths of a second, we devised a new instrument, the synchrono- phosphoroscope. Indeed, for the experiments to be described in this chapter, and which involved the use of surfaces of considerable size, the cooling of the substance during excitation, simultaneous observa- tions during fluorescence and phosphorescence, etc., none of the exist- ing forms are easily adapted. The original phosphoroscope of Bec- querel,^ later modified by E. Wiedemann,^ and also the revolving drum type used successively in various forms by Becquerel,^ Tyndall,^ Kester,® and Waggoner,^ afford sufficient speed, as does Merritt's^ phosphoro- scope of 1908; but none of these could be used without modification. The new apparatus® consists of a small synchronous, alternating- current motor A. C, figure 27, and a small direct-current motor D. C upon a common shaft. To one end of the shaft is attached a sectored disk, WW, figures 27 and 28, with four equal open and four closed sec- tors, corresponding to the four poles of the A. C. motor. On the cir- cuit of 60 cycles this machine, when brought to speed by the D. C. motor and released, runs steadily at 30 revolutions per second. A "step-up" transformer TT, in the same alternating-current circuit, produces discharges at the spark-gap, or series of gaps {E), at each alternation, i. e., 120 times a second. This discharge may be reduced to a single spark by proper adjustment of the resistance and capacity of the circuit, or more conveniently for many purposes the discharge may be confined to the peak of the wave by means of the four-pointed star-wheel SS (figs. 27 and 28), which is mounted on the shaft and carefuUy adjusted as to phase. The direct-current motor may also be used to drive the sectored disk at other speeds, in which case the circuit of the motor A. C. is broken and the discharge is derived from any convenient source capable of producing a proper spark at each quarter revolution. ^ E. Becquerel. Annales de Chimie et de Phyaque (3), t.xtt, p. 1. 1861. 'Ibid., Lv, p. 5. 1859. ' E. Wiedemann, Wiedmann a Annalen, xxxrx, p. 446, 1888. * E. Becquerel, 1. u. ^ Tyndall. See I^wis Wright's volume on light, p. 152. London, 1882. • Kester, Physical Review (1), rx, p. 164. ' Waggoner, Carnegie Inst. Wa^h. Pub. No. 152. ^ Nichols and Merritt, Carnegie Inst. Wash. Pub. No. 152. *E. L. Nichols: Proc. Nat. Acad, of Sciences, v. 2, p. 328. 1916. Also Science, xun, p. 937. 1916. 38 PHOSPHORESCENCE SPECTRA. 39 When the sectored disk WW is so adjusted on the shaft that the closed sectors conceal the phosphorescent surface during excitation by the spark, an observer, looking through the open sectors as they pass, sees the phosphorescence as it appears a few ten thousandths of a second after. The apparatus is thus suitable for the study of phosphorescence of very short duration or of the earliest stages in cases of slow decay. By shifting the sector on the shaft it is possible without variation in D Fig. 27, Fig. 28. the rate of rotation to make observations at the very beginnings of phosphorescence and to compare, by simultaneous vision, the appear- ances just before and immediately after the close of excitation, or, on the other hand, the earher with the later stages, up to about 0.004 second. The photometer, spectroscope, spectrophotometer, camera, etc., may all readily be used with this form of phosphoroscope and studies of the most varied character become possible. Phosphorescence is commonly regarded simply as the after-effect of fluorescence, the emission spectrum immediately after the close of exci- tation being identical with that immediately before excitation ceases. This has hitherto been only an assumption, since it is thinkable that the process which prepares a substance for phosphorescence might pro- duce emission during excitation differing from that which consti- tutes phosphorescence and which together with the latter would be present during fluorescence. It is also thinkable, although unlikely, that the phosphorescence might contain some components requiring a measurable time for development and observable only after an appre- ciable interval. This is a matter which it would be very difiicult to settle in the cases of phosphorescence hitherto studied, because the spectrum of fluorescence and phosphorescence consists of broad bands or complexes of overlap- ping bands, and almost the only criterion of identity is that of color. 40 FLUORESCENCE OP THE URANYL SALTS. The uranyl salts, because of their remarkable spectra, afford an unusual opportunity to estabUsh the exact relation between the emis- sion of light during excitation and at various times after excitation has ceased, and it was for this purpose that the first experiments with the new phosphoroscope were undertaken. The method, briefly outlined, is as follows: The substance, inclosed in a flat tube of glass BA about 8 cm. long and 2 cm. wide, is viewed through the rapidly revolving sectored disk of the synchrono-phos- phoroscope. It is mounted vertically, with its axis at right angles to the radius of the disk, as shown in figure 29. FiQ. 29. Fig. 30. It is uniformly excited by zinc sparks 120 times a second while hidden by the closed sectors and is visible for 1/240 of a second during the passage of each of the intervening open sectors. A phosphorescent substance of slow decay appears under these cir- cumstances to be equally bright from top to bottom, but if one of the uranyl salts, such as the double uranyl-anomonio sulphate, which was the substance selected for detailed study, be used, it appears a very bright green at the bottom of the tube, shading off to bare visibihty at the top. The rate of decay of this substance and of the other uranyl salts is so rapid that the upper end of the tube, which is seen at the intensity which corresponds approximately to the instant 0.003 second after excitation, has only a small fraction of the brightness of the lower end, which is viewed about 0.0005 second after excitation. The particular salt mentioned above was selected because at low temperatxires its spectrum is unusually well resolved in groups of com- plexes of narrow, line-like bands, making it possible to detect changes in the individual components. To obtain simultaneous observations a pair of right-angled prisms was mounted before the sUt of a Hilger spectroscope, as shown in figure 30. Light from the lower end of the tube A enters the lower half of the slit. That from the upper end £, after two total reflections, enters the upper half of the slit, and we have two spectra one above the other, PHOSPHORESCENCE SPECTRA. 41 Fig. 31a. Fig. 316. coinciding throughout as to wave-length, but separated by a dark line formed by the lower edge of the second prism (R^). To compare fluorescence with phosphorescence, the sectored disk was shifted upon its shaft until the lower end of the tube was viewed during excitation, the upper end immediately after (fig, 31a). To compare the phosphorescence spectrum at an earlier and later stage, the disk was so set that its position at the moment of excitation was as shown in figure 31b. By means of the reflecting prisms at the sUt of the spec- troscope, already described, the spectrum of the light emitted from region A was compared with that at B in each case. At 4-20° C. the banded spectra were found to be identical in every respect, except in brightness ; and the same was true at low temperatures, where it was possible to in- spect each of the numerous line-like bands individually. Of the seven homologous series distinguishable in the fluorescence spectrum, all were present in phosphorescent Ught, unshifted as to posi- tion and not perceptibly enhanced or diminished in relative brightness. The comparison was less satisfactory as regards minor details in the case of the early and late stages of phosphorescence, some of the fainter bands being invisible, but changes such as Doight be looked for, i. e., those due to the greater persistence of certain series, could scarcely have escaped notice. The significance of these observations is two-fold: On the one hand we find that for the only examples of luminescence which admit of such detailed inspection the spectrum of phosphorescence is identical with that of fluorescence, and since there are no indications to the contrary in the case of other classes of substances thus far studied, it is probable that the above statement will apply to all phosphorescent materials. On the other hand, we find that, in spite of its great complexity, the lumi- nescence spectrum of a uranyl salt is to be regarded as a unit, all its components decaying at the same rate after the cessation of excitation. Thus this class of substances (i. e., the uranyl salts) not only conform to the first three criteria of homogeneity discussed in Chapter II but likewise to that based upon the phenomena of phosphorescence. CURVES OF DECAY. To determine the change of intensity of phosphorescence with the time a simple form of photometer previously used in a study of the phosphorescence of kunzite^ was mounted in front of the sectored disk. ^ Nichols and Howes, Physical Review (2), iv, p. 19. 1914. 42 FLUORESCENCE OF THE URANYL SALTS. A lateral strip of the phosphorescent salt 1 cm. wide was excited by sparks from a single spark-gap between zinc terminals and measure- ments of the brightness were made at various times after the close of excitation. The necessary conditions were attained by shifting the disk successively through small angles, so as to vary the interval between excitation and observation. The time could be estimated with sufficient accuracy by noting the instantaneous positions of the disk for each adjustment, as given by the strictly synchronous illumina- tion due to the spark. X .L>iL 1_ L.B. FiQ. 32. The arrangement of the apparatus is shown in figmre 32, in which P is the phosphorescent surface, DD the sectored disk, L. B, the Lununer- Brodhun cube of the photometer, E the eyepiece, S a color-screen and matte translucent plate, C the comparison lamp which traveled along the track of an optical bench. The cross at Z indicates the position of the spark-gap. In table 15 relative intensities /, the reciprocals l/V/^and times T after excitation are given. Figure 33 shows the relations between I and T, and l/VFand T respectively in the usual manner. Table 15. T. /. i/Vi. T. I. 1/Vj. 0.000479 59.49 0.130 0.00170 2.03 0.702 .000637 27.78 .190 .00193 .971 1.014 .000856 16.02 .250 .00212 .610 1.280 .000949 12.62 .281 .00247 .296 1.836 .00110 9.80 .319 .00287 .159 2.524 .00146 5.03 .446 As appears from the table and curve ABC, figure 33, this substance exhibits a remarkably rapid decay, falhng in the interval between 0.0005 second after close of excitation and 0.003 second to less than three-thousandths of its intensity at the beginning of that interval. To show the degree of accuracy with which the lower intensities were observed, the portion of the curve BC is reproduced with ordinates magnified ten times B^C\ _The results are hkewise plotted in the cus- tomary manner with 1/ V/ as ordinates (curve DEF), and this brings PHOSPHORESCENCE SPECTRA. 43 out an unusual characteristic. It is usual to find two processes o phosphorescence succeeding one another and represented by the two straight arms of the curve DE and FGj but in all the numerous cases hitherto described, excepting two to be discussed in a subsequent para- graph, the later process (FO) is indicated by a curve of lesser slope. In the case of this uranyl salt, however, FG trends very sharply up- ward, showing a greatly accelerated decay. .003 SEC Fig. 33. Fig, 34. By means of these preliminary observations certain facts may be regarded as established.^ These may be summarized as follows: (1) There is no appreciable change of color during decay. (2) The decay of phosphorescence is exceedingly rapid, the intensity falling to one-thousandth of its initial value within 0.0035 second. (3) The very complex fluorescence spectrum at — 180° C. is identical in structure and relative distribution of intensities with that observed during the earUer and later stages of phosphorescence. ^ Nichols, Proceedings National Academy of Sciences, vol. ii, p. 328. 1916. 44 FLUORESCENCE OF THE URANYL SALTS. Table 16. (4) The curve of decay of phosphorescence differs from the prevailing type in that although as usual two successive processes are distinguish- able, the second process is more rapid instead of being slower than the first. The study of these phenomena has since been extended to several other typical uranyl salts, the curves of decay of which were deter- mined by the method just described and under conditions of excita- tion, etc., as nearly constant as possible.-^ These curves of decay are of the same new type originally found in the uranyl ammonium sul- phate. The two processes, as determined by the customary method of plotting / -1/2 as a function of the time are indicated by straight Unes differing from one another in slope and the second process has in all cases the steeper gradient. Later experiments, in which the intensity of excitation was increased, revealed the presence of a third process not included within the interval of time covered by our earlier experiments. STUDIES INVOLVING THE FIRST AND SECOND PROCESSES. The curves shown in figures 34 and 35 are typical of the results obtained with all the salts under observation. They represent the decay of the phosphorescence of the compounds shown in table 16. The initial intensity, under like excitation, varies greatly in the different salts, as also, to some ex- tent, does the rate of decay. It will be noted that the initial intensities of the ammonium and potassium sulphates, for example, are several times greater than those of the nitrate, the sulphate, and the am- monium chloride. This is, however, a question of previous history as well as of chemical and physical constitution, as was determined in the following manner: Uranyl potassium sulphate was dissolved in hot water and a mass of the minute crystals which were thrown down on cooling the solution were immediately sealed up in a glass tube. Care was taken through- out these manipulations to protect the precipitate from the action of light. This sample, still in darkness, was mounted in the synchrono-phos- phoroscope and a curve of decay was taken, the first exposure to excit- ing light being that at the beginning of the run. The substance then showed, temporarily, a brilliancy of phosphorescence much above that to be obtained under ordinary circumstances, but was soon reduced to its normal and semi-permanent condition, after which the usual curve of decay was obtained. Ciirve. Substance. 1 2 3 4 5 Uranyl ammonium sulphate. Uranyl potassixmi sulpl^ate. Uranyl nitrate + 6H2O. Uranyl sulphate. Uranyl ammonium chloride. 1 Nichols and Howes, Physical Review (2), ix. p. 292. 1917. PHOSPHORESCENCE SPECTRA. 45 EXCITATION IN THE PRESENCE OF RED AND INFRA-RED RAYS. To determine whether red or infra-red rays have an effect on these substances similar to that observed in the case of the phosphorescent sulphides, a modification of the apparatus was made such that the surface under examination could be subjected to the intense illumina- tion obtained by focusing the crater of an electric arc upon it. A screen of excellent ruby glass was interposed to cut off all but the longer waves and observations were made through a screen quite impervious to red. Exposure to this source was found to affect measurably neither the brightness of fluorescence nor of phosphorescence. Curves taken after exposure to this source, those taken with the substance subjected to it interruptedly throughout the run, and curves in the determination of which readings were taken alternately with and without red light were all identical with those taken in entire absence from such expos- ures. The striking contrast between this negative result and the well-known effects of infra-red radiation upon .the phosphorescence of the sul- phides is notable. The observations already cited, showing the complete identity of the spectrum of fluorescence with that of phos- phorescence seemed to indicate that the intensity would go over from that of fluorescence to that of phosphorescence without discontinuity. This conclusion was confirmed, within the errors of observa- tion, by measurements just before and after the close of excitation. The only previous instances where this relation has been experimentally established, so far as we know, are to be found in Waggoner's^ studies of phosphorescence of short duration and in recent observations on the luminescence of kunzite.^ In view of the unexpected character of the decay curves for the phos- phorescence of the uranyl compounds, the question arises whether the rather unusual mode of excitation employed, i.e., periodically repeated exposures, 120 times a second, to groups of sparks of high frequency, might produce such a result, or whether the decay curves are character- ^ Waggoner, Physical Review, xxvii, p. 209. * Nichols and Howes, Physical Review (2), iv, p. 26. I-i- 5. ao 4. / 2./ / / 1. 60 60 i / / /.. 40 // 40 20 y. ly^ • 20 ^ ^ .001 1 .002 1 SEC. Fig. 35. 46 FLUORESCENCE OF THE URANYL SALTS. istic of this class of compounds, whatever the mode of excitation. It is true that both Waggoner^ and Zeller,^ using a Merritt phosphoro- scope, found in their studies of phosphorescence of short duration that excitation by means of a spark discharge very similar to our own gave decay curves of the usual type. It is also obvious from the measurements already described that the interval between excitation, i. e., 1/120 second, is sufficient for the com- plete discharge of the phosphorescent glow, and since the absence of any effect of red and infra-red indicates that there is no storage of undeveloped energy to be carried over, such as occurs in the phosphor- escent sulphides, it seems probable that the decay curves do not vary greatly from that which might be obtained, were it possible to make the experiment, from a single exposure. To test this a run was made upon the sample of uranyl anmionium sulphate previously used, but with the Merritt phosphoroscope. By driving the disk of this instrument 3,000 revolutions a minute, much the same range of time intervals was available as with the synchrono-phosphoroscope. To further vary the conditions, a quartz mercury arc was substituted for the spark-gap of Waggoner and Zeller. The arrangement of appar- atus was as shown in figure 36, in which DD is the revolving disk, H the mercury lamp, P the phosphorescent substance, LB the Lummer-Brodhun cube of the photometer, SS a color- filter and milk-glass screen. The device for shifting the oblique mirror M with reference to the aperture A in the disk is not shown. Although the decay was some- what more rapid in this determination on account of the less intense excitation, the curve was of precisely the type obtained by the prev- ious method. Measurements upon some of the bands of brief duration in the spec- trum of the phosphorescent sulphides, recently made with the syn- chrono-phosphoroscope under experimental conditions identical with those here described,^ yield curves of the usual type associated with these sulphides, so that the question of the change of form being due to the phosphoroscope employed is effectually eliminated. S.&E / 3 p / ' D 1 / y L.B. FiQ. 36. ^ Waggoner, Physical Review, xxvii, p. 209. 2 Zeller, Physical Review, xxxi, p. 367. 3 Nichols, Proc. Am. Philosophical Society, lv, p. 494. 1916. PHOSPHORESCENCE SPECTRA. 47 SOLID SOLUTIONS AND SEMI-FLUIDS. The uranyl salts differ from nearly all if not all phosphorescent sub- stances hitherto studied. We do not have, as in the phosphorescent sulphides, the preparations of Waggoner, the ruby, etc., to deal with a trace of active material in solid solution, but with compounds that are in themselves brilliantly phosphorescent. If the peculiar character of the curve of decay is due to that fact it might be expected that uranium UFUNYL AMMONIUM SULPHATE Fig. 37 Fig. 38. glass, in which the active material is considered to be in a state of solid solution, would have a law of decay corresponding to the prevaiUng type for such solutions, i. e,, with the first process as indicated by the curve for /-i/^^ and time, represented by a hne of steeper slope thaij the Une for the second process. A piece of uranium glass gave, however, a decay curve similar to those of the uranyl salts (see fig. 37). Another 48 FLUORESCENCE OF THE URANYL SALTS. preparation which differs from most of the uranyl salts is the uranyl sodium phosphate, a sample of which was made by D. T. Wilber for certain studies in fluorescence recently published.^ This substance is a very viscous liquid with the characteristic green fluorescence of the uranyl compounds. One might expect, in accordance with the findings of Becquerel for liquids in general,^ that there would be no observable after-glow. It EFFECT OF TEMPERATURE ON URANYL AMMONIUM NITRATE SEC. FiQ. 39. is true that Becquerel expressed the belief that with a phosphoroscope of sufficient speed, phosphorescence would probably be detected in fluorescent liquids, but no one, so far as we know, save Dewar in an unconfirmed statement concerning a supposed phosphorescence of liquid air, has since that time (1859) recorded an instance of phos- phorescence excepting in solids and gases. When a tube containing the phosphate was tested with the syn- chrono-phosphoroscope no phosphorescence was found of duration 1 Howes and WUber, Physical Review (2), vol. 7, p. 394. Mar. 1916. ^ See E. Becquerel, La Lumiere, vol. i, chapter on Phosphorescence. PHOSPHORESCENCE SPECTRA. 49 sufficient to be detected. Another sample so prepared as to reduce the amount of water to a minimum did, however, exhibit phosphorescence of measurable duration. This preparation, so slow was its rate of flow, might be regarded as a plastic solid rather than a viscous liquid. A bead of microcosmic salt, colored in the usual manner with uranium oxide, was comparable in its phosphorescence with canary glass. Fig. 40, It appears that the persistence of luminescence is due to the con- sistency of the substance and disappears as the fluidity increases; also that the peculiar type of decay here described is common, not only to the crystalline uranyl salts in general, but also to the gelatinous forms, as in this double salt, and to substances in which uranium appears in solid solution, as in the case of the canary glass. THE THIRD PROCESS. ' E. Becquerel,^ in the course of his great pioneer work on phosphores- cence, made a number of observations on the uranyl salts and on ^ E. Becquerel, Annales de Chimie et de Physique (3), lxii, p. 1. 1861. 50 FLUORESCENCE OF THE URANYL SALTS. uranium glass. He noted the brilliant initial intensity and very rapid decay, and to test the independence of the constant in his equation of decay when the illumination varied he made many measurements. If from his data we compute Z-1/2, as a function of the time, we obtain curves of the same general form as those in figure 32. Becquerel's observations are not numerous enough, taken by them- selves, to determine completely the type of curve. His measurements, however, cover a larger time interval than ours and the values for the longest times indicate an even more rapid decay following the second process. We had, indeed, found some indications of a similar tendency which had been omitted from our curves as lying almost beyond the range of definite determination. To investigate the further trend of the curves of decay, the intensity of excitation was increased by readustment of the sparking circuit, by which means it was found possible to extend the time interval for more than 0.006 second beyond the cessation of excitation. Careful, often repeated measurements, of the various salts showed in fact a third linear process beginning where our previous determina- tions had ceased and having a steeper slope, indicative of still more rapid decay. Typical results are indicated in figures 38, 39, 40, etc. These processes may be numbered for convenience 1, 2, and 3 in the order in which they occur. Processes 1 and 2 are in general of about equal duration for a given salt. The abruptness of transition, however, varies greatly, and in some instances the change of slope is so gradual as to encroach seriously on process 2 at both ends. THE INFLUENCE OF TEMPERATURE. The only previous instances of decay of phosphorescence in which the later stages are more rapid than those preceding are noted by Ives and Luckiesh^ in their study of the influence of temperature on phosphores- cence, and by E. H. Kennard^ in a more recent paper. Ives and Luckiesh measured the phosphorescence of one of Lenard and Klatt's sulphides (BaBiK from Leppin and Masche). This sub- stance was found to be very sensitive to change of temperature and the results at 0°, 22"^, and 35°, C. when plotted for /-1/2 and time in the usual manner, gave curves varying greatly in slope. The curve for 0** is concave toward the time axis, that for 22° linear, and that for 35° strongly convex. They show that a Unear relation may be obtained for each of these curves by varying the exponent of /. The effect of temperature in the case of the phosphorescent sul- phides, where one has to do with a composite of many overlapping bands of varjdng duration, is undoubtedly different from that to be * Ives and Luckieah, Astrophyaical Journal, xxxvi, p. 330 (1912). *Kennard, Physical Review (2), iv, p. 278 (1914). PHOSPHOKESCENCE SPECTRA. 51 expected with the uranyl salts, where the spectrum, in spite of its complexity of structure, is a unit. It was deemed of interest, however, to determine the effect of temperature upon the latter. For this purpose a specimen of the uranyl ammonium nitrate was mounted within a cylindrical Dewar flask with unsilvered walls and its decay of phosphorescence was determined with a synchrono-phos- phoroscope at a temperature a few degrees above that of liquid air (about -180°) at +20° and at +60°. The last-named temperature was maintained during the run by means of an electrical heating-coil. Fig. 41. The principal change consists in a marked retardation of decay with lowering temperature (see fig. 39), but this is not a universal charac- teristic of the uranyl compounds. Uranyl anamonium sulphate, for example (fig. 40), is but slightly influenced in its rate of decay by change of temperature and the curve for — 180° is intermediate between those for +20° and +60^. 52 FLUORESCENCE OF THE URANYL SALTS. THE EFFECT OF VARYING THE INTENSITY OF EXCITATION. To determine the effect of the intensity of excitation, a series of measurements were made with the spark-gap at various distances from the phosphorescent surface. The substance observed in these experi- ments was uranyl rubidium nitrate. It was found possible to make observations of the decay of phosphorescence with the excitation reduced to a two-hundredth of that usually employed. From the curves obtained, of which four are given in j&gure 41, it will be noted that all three processes are present, whatever be the intensity of the exciting hght; also that, taken roughly, processes 1 and 2 are of nearly equal duration, and that with decreasing intensity of excitation the duration of each of these processes diminishes. DURATION OF PROCESSES WITH EXCITATION E. 1. I&2. a .40 / ^0 ^x / .M 7 ^^^^ i^ M x^ 1 y — -'*'*""T> ' ' r 1 2 .002 Fig. 42. .003 SEC. These relations are better shown in figure 42, in which the duration of process 1 and the sum of the duration of processes 1 and 2, counting from the close of excitation, are plotted, with the intensity of the ex- citing light as ordinates. Approximately in both cases the duration is proportional to the natural logarithm of the excitation. (See table 17.) This decrease in the duration of the two processes with falling excitation affords an obvious explanation of the varying character of PHOSPHORESCENCE SPECTRA. 53 the curves of decay of phosphorescent substances. Where the excita- tion is chiefly superficial, as in the case of some powders, the excitation may be nearly of one intensity and the curve inade up of well-defined Unear processes with sharp inflection-points. We have found this to be the case in many instances. Where, on the other hand, fluorescence is excited within the crystalline mass by rays that have suffered con- siderable loss by absorption, etc., there will be a wide range of intensi- ties of excitation and a curve results with distributed knees and linear processes shortened and sometimes almost obliterated. We observed this particularly where a clear crystal was mounted with faces per- pendicular to the photometer and was excited from behind so that the light emitted by the surface nearest the exciting source passed through Table 17. — Variation of length of processes with excttaiion {phosphorescence of uranyl rubidium nitrate). Intensity of Duration, Nat. log. E. excitation Process 1. Process 2. Process 1+2. sec. sec. sec. 41.70 6.033 0.0022 0.0018 0.00400 13.70 4.919 .00170 .0014 .00310 2.52 3.220 .00147 .000993 .00240 1.35 2.590 .00118 .00080 .00198 .900 2.190 .00080 .00090 .00170 .476 1.560 .00070 .00076 .00146 .201 .698 .00050 .00060 .00110 the crystal and was partially absorbed. Excitation occurred within the crystal in diminishing amount with increasing depth and the com- posite phosphorescence reaching the eye under such conditions showed this blending effect to a marked degree. The same crystal when excited from in front gave a curve in which the angles between pro- cesses were made more sharply defined. The effect in question is probably a general one and may well account for the perplexing differ- ences in the curves of decay obtained under slightly varying circum- stances. Thus, one observer will obtain an angular curve, where another, studying the same material, can detect no linear processes. The same observer, indeed, in attempting to repeat his measurements, will often find the above-mentioned change of type under conditions which seem to be identical but in which the same relations as regards superficial and internal excitation are not preserved. We found in the study of this effect a crystal one smooth face of which gave the blended curve, while the opposite face, which was rough, gave the angular curve, a change produced and reproducible by merely rotating the crystal through 180*^. 54 FLUORESCENCE OF THE TJRANYL SALTS. THE PHOSPHORESCENCE OF VARIOUS NITRATES. Observations were made on a series of nitrates previously prepared for the detailed comparison of the fluorescence spectra of that salt> These consist of crystals with 6 H2O (rhombic), 3 H2O (triclinic), and 2 H2O (system undetermined) as water of crystallization and a speci- men sealed in glass which had been rendered as nearly anhydrous as was possible without decomposing the nitrate. The curves of decay indicate a much slower rate of decay for the crystalline forms than for the anhydrous nitrate. Whatever effect the amount of water of crystallization may have is doubtless masked by the far greater influence of the crystalline form. This is perhaps to be expected, since, as will be shown in Chapter VII, these specimens exhibit as great differences in the structure and appearance of their fluores- cence and absorption spectra as commonly exist between entirely distinct uranyl salts. Similar differences in the case of salts similar in composition but differing in crystalline form will Ukewise be described in a subsequent chapter. Fig. 43. OBSERVATIONS ON POLARIZED PHOSPHORESCENCE. Certain crystals of the double chlorides of uranyl exhibit fluorescence spectra consisting of sets of bands polarized at right angles to one another. To determine whether these components after the close of excitation decay independently or without change in their relative intensities, the following experiment was made: A crystal of the rubidium uranyl chloride that exhibited the phe- nomenon of polarized fluorescence was mounted behind the disk of the synchrono-phosphoroscope and was observed with a spectroscope. ^Nichols and Howea; Physical Review (2), ix, p. 292. 1917. PHOSPHOKESCENCE SPECTRA. 55 The slit of the latter instrument was divided into two parts by means of an opaque strip across the middle {S, fig. 43). Within the colhmator a doubly refracting rhomb R and Nicol prism N were mounted. The rhomb gave two slit-images vertically displaced and the adjustment was such that the lower part (A) of one image was contiguous with the upper part (B) of the other. Thus two spectra of the phosphorescent field were obtained corre- sponding to the two polarized components. These presented the usual distinctive structures at whatever stage of the phosphorescent decay they were observed. By rotation of the Nicol prism the two fields could be brought to equal brightness for any given part of the spec- trum, and this balance, if made with the sector of the phosphoroscope set so as to give observations at 0.0005 second after extinction, was found equally correct up to 0.005 second or as long as phosphorescence was observable. The two components therefore decay at the same rate. SUMMARY OF PHOSPHORESCENCE OF SHORT DURATION. (1) All uranyl salts thus far examined possess the same type of phosphorescence; i. e., with increasing instead of diminishing rates of decay. (2) This is true not only of the crystalline forms, but also of uranyl compounds in solid solution or in the plastic state characteristic of the double phosphates. (3) The initial brightness of phosphorescence under like excitation varies greatly with the different salts, as does also to some extent the rate of decay. (4) The brightness of a salt newly prepared in darkness is greater when first excited than subsequently, but it soon reaches a nearly stable condition. (5) Exposure to red and infra-red rays is without effect as regards the rate of decay. (6) The phosphorescence, Hke the fluorescence, of the uranyl salts appears to be independent of the mode of excitation, and the structure of the intricate spectrum is the same during excitation and throughout the observable phosphorescent interval. (7) Changes in the rate of decay are not continuous, but occur in definite steps, there being at least three successive processes within the interval covered by observations, i. e., about 0.006 second. These processes follow a law such that /-1/2 is in Unear relation to the time. (8) The first and second processes, counting from the close of excitation, are of nearly equal duration, increasing in duration with the intensity of excitation in such a manner that the duration of the process is approximately proportional to the natural logarithm of the excitation. 56 FLUORESCENCE OF THE URANYL SALTS. (9) In certain salts, such as uranyl ammonium nitrate, decay is retarded by cooling; in other cases the temperature effect is slight. (10) Uranyl nitrates with 2, 3, and 6 molecules of water of crystal- lization vary greatly in the rate of decay, but the changes in crystalline form appear to be more important in this respect than the amount of water. (11*) In the case of the polarized spectra of the double chlorides, both components decay at the same rate and no change in relative bright- ness can be detected throughout the range covered by observation. PHOSPHORESCENCE OF LONG DURATION. While comparing the spectra of uranyl salts under excitation by kathode rays and under photo-excitation, in 1917, Misses Wick and McDowell discovered that certain salts continued to glow for several minutes after bombardment in the vacuum tube, at the temperature of liquid air. Many uranyl compounds are unstable in vacuo, and of those which are not decomposed rapidly, some, notably the chlorides, are prac- tically inactive under the kathode rays. The following salts, which were prepared by Mr. Wilber in the form of fairly large, well-formed crystals, gave bright fluorescence and were fairly stable: Uranyl potassium nitrate, K2U02(N08)4 (crystallized from 10 to 30 per cent nitric acid). Uranyl potassium nitrate, K2U02(N03)4 (long crystals from 2 to 3 per cent nitric acid). Uranyl potassium nitrate, KU02(N03)s (water form). Uranyl potassium nitrate, KU02(N03)8 (anhydrous). Uranyl potassium sulphate. Uranyl potassium sulphate (with 2 molecules of water). An examination was made of all of this group. They were found to exhibit phosphorescence in varying degree. Some showed no phos- phorescence of noticeable duration. The following, which were among the brightest, were selected for study: (1 and 2) K2U02(N03)4. The first form, A, was crystallized from a 10 to 30 per cent solution of nitric acid, and the second form, 5, from a 2 to 3 per cent solution. Although the crystallographic form is identical, form A crystallizes in short, thick crystals and form B in long, slender crystals. There appeared to be a slight difference in the phosphorescence of the two forms. It is possible, however, that the difference observed might have been due to some variation in the conditions under which the phosphorescence was produced. (3) K2U02(S04)y. To ascertain whether, as the result of exposure to the kathode rays, the surface layer of the crystals had undergone some change which rendered them capable of persistent phosphores- cence under photo-excitation, they were alternately illuminated by the light of a carbon arc and bombarded by the kathode rays. To accom- PHOSPHORESCENCE SPECTRA. 57 plish this without changing any conditions except the mode of excita- tion the tube containing the crystal under observation was mounted within an unsilvered cyhndrical Dewar flask and cooled to the tempera- ture of liquid air. Light from a carbon arc was focussed upon the crystal through the walls of the Dewar flask and of the vacuum-tube, producing intense fluorescence, but there was no after-glow of duration sufiicient to be detected. The kathode discharge, however, caused the persistent phosphorescence already described and the effect appeared to be distinctly cumulative, requiring excitation for several seconds. After the phosphorescence had died away, photo-excitation was re- sumed, and this process was repeated many times without observable change in the effect of the light. IDENTITY OF THE SPECTRA DURING FLUORESCENCE AND KATHODE- PHOSPHORESCENCE. To determine whether the spectrum, during this persistent phos- phorescence, corresponded with the fluorescence spectrum, settings on several of the brightest bands were made with the Hilger spec- troscope. The result was the same as the observations upon the brief phosphorescence following photo-excitation, described in an earlier paragraph of this chapter; i. e., the spectra were found to be identical during and after excitation and remained unchanged in character as long as they were visible. CURVES OF DECAY FOR THE KATHODE-PHOSPHORESCENCE. Misses Wick and McDowell also determined the law of decay for the three salts (1, 2, and 3) selected for investigation. Since the effect lasted for several minutes, it was possible to use the method commonly employed in such measurements. The arrangement of the apparatus is shown in figure 44. ® Fig. 44. A Lummer-Brodhun cube A was placed at one end of a track XY, about 3.5 meters long. The crystal B was placed opposite one face of the cube. The comparison source L was a 5-volt tungsten lamp placed in parallel with a suitable rheostat upon a 55-volt circuit. The lamp was mounted in a carriage C, running on the track XF, on which, at intervals of about 25 cm., stops were placed. Green, blue, and ground glass absorption plates P and P' were inserted to obtain a comparison source of the proper color and intensity. A chronograph was used to record the time. The zero of time was in every instance recorded when the primary circuit of the induction coil was broken. When the 58 FLUORESCENCE OF THE URANYL SALTS. intensity of phosphorescence matched that of the source in the fibrst possible position, the time was again recorded and the carriage moved to the next stop, and allowed to remain until a match was made as before. This procedure was continued until the phosphorescence was too faint to observe or until the end of the track was reached. The interpretation of the results was somewhat difficult, since the instability of the crystals rendered uncertain both the control of the vacuum and the maintenance of the crystal surface unchanged during prolonged bombardment. The general shape of the decay curve after long excitation is shown in figure 45. The curves are of the type usual with phosphorescence of long duration, consisting of two linear processes, of which the first is the more rapid, whereas, as has been shown in the previous portions of this chapter, the decay following photo-excitation is of a new and entirely different type. ri > / 1 / / / y / / 1/ / / J / / / / ^ y / / ./ / / 9A ^ ^- / ^ y ^ WO / / / / 7 / ^ ^ I + v^ (/ / // r 1 Jl // f 1 ^ il ■■ "4 M " IKC( 4 )HOS Qd u SfCO Z( HDS K M SI ««s= M FiQ. 45. Fia. 46. Fig. 47. Under different conditions, phosphorescence was observed to last from less than a minute to 10 or 15 minutes. The exact form of the curve varied with the time of excitation. The time of decay was found to increase with the time of excitation, as shown in figure 46, but the initial brightness changed relatively httle. There was some evidence to indicate that under similar conditions of vacuum the rate of the first process remained practically unchanged for varying times of excita- tion, but that the second process began sooner for longer excitation, as shown in figures 47 and 48. In figure 48, curves 1 and 2, obtained by a short-time excitation, show only the first process, whereas curves 45 and 46, obtained by excitations of 40 and 80 seconds respectively, indicate that a state of saturation had been reached such that added excitation produced no change in the phosphorescence. As has been stated, the initial brightness and rate of decay were found also to depend upon the strength of the bombardment, as varied by the pressure in the tube and by the voltage applied to the induction PHOSPHORESCENCE SPECTRA. 59 coil. The curves of figure 45, for example, were obtained with a rela- tively high vacuum, whereas those of figure 49 were obtained with a very low vacuiun, so that the decay was comparatively rapid and there was only a suggestion of the beginning of the second process in the position of the last point observed. Slight changes in temperature, such as were produced when the liquid air fell below the line of the crystal, were found also to produce changes in the initial brightness and rate of decay. ,-i 1 ? 3. • 200 / ->^ • J •^ '/ > ^ -^T 100 .0 ^- / f 90 79 SECONDS Fig. 48. 100 To determine whether the excitation produced any secondary change in the crystal, which persisted after the phosphorescence had dis- appeared, so that there would be a progressive building up of the phosphorescence, excitations were made of equal length, repeated at as nearly equal intervals as decay observations permitted. Figure 49 shows that, at a fairly low cathode vacuum, an excitation of 20 seconds, repeated at approximately 1-minute intervals, produced identical 1-^ /+o, AOO y r 100 X. y y y ■ 1 \ % 6 3 6 ■ l-i ^ A A ^ r wu o^ ^ 1 4 % »d aecQNDS Fig. 49. SECONDS Fig. 50. decay curves. The same effect is shown in figure 50 for a much longer period of decay. When the time between excitations was short as compared to the time and strength of excitation, there appeared to be a progressive change, as indicated in figure 51. 60 FLUORESCENCE OF THE URANYL SALTS. From this investigation by Misses Wick and McDowell, two definite conclusions may be drawn: (1) The spectrum of the long-time phosphorescence produced by cathode-ray excitation at Uquid-air temperatures is identical with the fluorescence spectrum. rs i so 100 secoNos Fig. 51. (2) The decay curve of the cathode phosphorescence differs in the most striking manner from that of the brief photo-phosphorescence. It corresponds in type with that usually found in cases of phosphores- cence of long dm-ation. V. THE MORE INTIMATE STRUCTURE OF URANYL SPECTRA AS REVEALED BY COOLING, It was first shown by J. and H. Becquerel and Onnes/ who studied the spectra of several of the uranyl salts when excited to fluorescence at the temperature of liquid air and ultimately at that of liquid hydro- gen, that each band of the spectrum as we know it at +20° is resolved into a group of much narrower bands. It was further shown by these investigators that all of the various groups of bands in a given spectrum were resolved in precisely the same manner, the homologous com- ponents forming series. This more intimate structure, which is revealed by cooUng, may be studied to great advantage in the case of the double chlorides, which salts, as has been noted in Chapter III, have spectra sufficiently resolved at +20° so that the origin of the components observed at — 185° can be traced and the relation of the two spectra to one another much more definitely determined than is the case where the spectrum at +20° consists of unresolved bands. Four of these chlorides have the following composition: UO2CI2. 2KCI+2H2O. U02Cl2.2RbCl+2H20. UO2CI2. 2NH4CI+2H2O. UO2CI2.2CSCI. They crystallize in tricUnic plates which are strongly fluorescent and their spectra, which are almost identical in structure, are resolved at room temperature into 8 groups of narrow bands. Each group, which corresponds to a single band of the ordinary uranyl fluorescence spectrum, consists of 5 nearly equidistant bands. The symmetry of these spectra, as they appear to the eye when viewed with a spectro- scope of moderate dispersion, is most striking. The bands are well separated from their neighbors and are about one-tenth as wide as the bands of the ordinary type of uranyl spectra. The distribution of intensities within the group has been determined for the visually brightest group in the spectrum of the ammonium uranyl chloride by means of the spectrophotometer. The results of such a determination are given in table 18, and are shown graphically in figure 52. The curve (fig. 52) which forms an envelope of the group of bands is of the same type as that for the distribution of intensities in a single band of the ordinary uranyl fluorescence spectrum and of the envelope of the set of bands in such a spectrum and is also similar to curves of distribution of the fluorescent spectra having a single broad band.^ The effect of cooling is likewise analogous, the envelope for —185° being narrower on account of the great relative reduction in brightness of the outlying members of the group. All the bands are shifted in ^ Becquerel and Onnes, Leiden Communications, 110. 1909. 2 See Chapters II and III. 61 62 FLUORESCENCE OF THE URANYL SALTS. Table 18. — Intensities of hands in group 6 {excited at +20° C). Band. Intensity. 5306m 18 5259 33 5208 24 5159 11 5119 (0 ^ Visible, but too dim for spectrophotometric measurement. position as well as changed in intensity in a manner to be described in a subsequent paragraph. To determine as closely as possible the wave-lengths of the bands, photographs of the spectra of the four double chlorides were taken and many visual settings were made. Fluorescence was excited by means of the carbon arc, the Ught from which passed through a screen opaque to rays of wave-length greater than about 0.45 fi and was f ocussed upon the crystal. Various exposures were employed on account of the great d fferences in the intensity of the bands and special plates were used for the red end of the spectrum. The exciting Ught was excluded from the camera by the use of suitable screens opaque to the blue and violet , except where the absorption spectrum was to be recorded. The various negatives were measured by mount- ing them on a micrometer stage in the field of the lantern. The micrometer-screw was carefully cali- brated, so that wave-lengths could be determined by measuring the distance of the crests of the bands from certain reference lines of the mercury spectrum, which was photo- graphed on each negative so as to overlap the fluorescence spectrum. This method of projection was found better than the use of the comparator commonly employed in the measurement of line spectra, because of the hazy character of the bands and because bands that are so weak and vague as to be invis- ible even under a low-power micro- scope could be seen and located by means of the lantern. Many measurements of the stronger bands were made with the comparator as a check on the determinations with the lantern. These measurements confirmed to a remarkable degree the apparent symmetry of the spectrum. When all the bands are plotted on a large scale, in a diagram with the recip- rocal of wave-lengths as abscissse, the spectrum is seen to consist of 8 groups of 5 bands each, as already described. The nearly uniform arrangement of the bands of each group repeats itself precisely from group to group, so that corresponding members of the groups form an .53/6 FiQ. 52. INTIMATE STRUCTURE ON COOLING. 63 homologous series of equidistant bands. This interval, moreover, is very nearly the same for all five of these homologous series; but although the departures from equahty are of the same order as the errors of measurement, there is reason, as will be seen later, to regard them as real. The general arrangement of the bands in these spectra is roughly depicted in figiu-e 53, which is based upon measurements of the fluores- cence spectrum of the ammonium uranyl chloride. Horizontal dis- tances are plotted on the scale of frequencies, the corresponding wave- / \ / \ / ' / / \ \ \ / / \ \ / \ / n ^ / ^ A r \ \ / /\ -. \ .6'4 /IaJ ,ul[/i.1 j \ i-Llif i,.Il/uA FiQ. 53. lengths being indicated for convenience. Vertical heights indicate relative intensities, but with some pretence of accuracy. The first and eighth groups at the extreme left and right, for example, if drawn to scale, would be scarcely visible. They are, in fact, so feeble that they can be observed only with the greatest difficulty. The location of the various bands of the 4 double chlorides, in wave-lengths and in frequencies (1/m X 10^) is given in table 29 at the end of this chapter. The values given are the averages of several readings from the photo- graphs and from visual settings. The bands in each group from the red toward the violet are designated by the letters 5, C, Z>, E^ and A, and bands having the same letter thus form homologous series. To determine the intervals between groups, the position of what may be called the center of each group was found by averaging the frequencies of all 5 bands. The intervals between these centers for groups 2, 3, 4, 5, 6, and 7 are given in table 19. Groups 1 and 8, for which insufficient data were available, were omitted, except in the case of the ammonium chloride. The only indication of a systematic departure from uniformity of interval for a single salt appears in the case of the caesium chloride, the average group-interval for which is smaller than that of the other salts by nearly 0.5 per cent. 64 FLUORESCENCE OF THE URANYL SALTS. As will appear in the course of the subsequent consideration of individual series, the tendency of the group intervals of the csesium salt to diminish toward the violet is not, as might seem at first sight, an indication that the groups are made up of series having a diminish- ing interval. As regards the other salts, it will be noted that the dis- tance between groups is essentially constant. Table 19. — Distances between fluorescence groups. Group. Potassium uranyl chloride. Ammonium uranyl chloride. Rubidium uranyl chloride. Caesium uranyl chloride. Center of group. Inter- val. Center of group. Inter- val. Center of group. Inter- val. Center of group. Inter- val. 1. 2. 3. 4. 5. 6. 7. 1505.3 1588.6 1671.9 1755.0 1838.4 1922.3 2005.6 1587.2 1670.8 1754.0 1836.6 1919 6 83.6 83.2 82.6 83.0 83.7 83.3 83.3 83.1 83.4 83.9 83.3 1591.5 1674.9 1758.3 1840.7 1924.2 2007.8 83.4 83.4 82.4 83.5 83.6 1592.6 1675.9 1759.3 1841.8 1924.7 2006.6 83.3 83.4 82.5 82.9 81.9 Average dist. 2003.3 mcea .... 83.22 83.38 83.26 82.80 DISTRIBUTION OF BANDS WITHIN THE GROUPS. While to the eye the fluorescence spectra under consideration present the appearance of evenly spaced bands varying periodically in intensity so as to form the groups, this is not strictly the case, as may readily be shown by subtracting neighboring values in table 29. The average distances thus obtained are given, for convenience, in table 20. The greatest departures from uniformity of distribution occur in the spectra of the rubidium chloride and the csesium chloride. Table 20. — Average distances between neighboring bands in the fluorescence spectrum at -{-20° C. Fluorescing substances. Distances between bands. C to B. D to C. E to D. A toE. B to A. UO2CI2.2KCI UO2CI2.2NH4CI.... U02Cl2.2RbCl UO2CI2.2C3CI Averages 15.97 17.56 16.20 18.25 18.66 17.74 18.43 12.85 17.96 17.86 18.50 18.63 14.70 15.67 12.75 14.52 15.58 15.67 17.12 18.10 16.99 16.92 18.24 14.41 16.62 In the rubidium spectrum, bands A and E are crowded together, the average interval being 12.75, and in the csesium spectrum D and C are similarly crowded. It will be noted that the average distance between A and E is less for the four chlorides than any of the corre- sponding distances between other bands. The arrangement of the bands within the group in the four salts is conveniently compared by means of the diagram in figure 54, in which INTIMATE STRUCTURE ON COOLING. 65 the geometrical centers of the groups are in the same vertical line. It will be seen from the diagram: (1) That the group center is in all four cases almost coincident with the crest of the D band. (2) That the distance between D and E is approximately the same in all. (3) That the arrangement of bands within the group is essentially the same in all, except for the displacement of band A in the spectrum of the rubidium and of B and C in that of the caesium salt, as mentioned above. Further discussion of these discrepancies will be found in a later paragraph of this chapter. INTERVALS OF THE INDIVIDUAL SERIES. For the consideration of the frequency intervals of the individual series, the values from table 29 have been arranged by series in table 30. Distances between the observed positions of neighboring members of each series are given in the column marked ''Intervals." To facilitate the detection of systematic departures from uniformity of interval, a column of values calculated by the following method is likewise given: A ''center" for each group was found in the manner already employed for determining the group centers. Around this the calculated positions were arranged under the assumption of a con- stant frequency interval equal to the average of the observed intervals for each series sepa- rately. The column marked ' ' Differences ' ' indicates the de- parture of the observed values from those thus calculated. The departures from uniform- ity of interval are unsyste- matic, indicating, for all the salts, that the series may be regarded as having a constant frequency interval. This interval has been computed for each series by subtracting the observed frequency of each band from the frequencies of all the other bands of the series and dividing the sum by the total number of inter- vals in question. The results are presented in table 21. These data indicate no progressive change of interval with the molecular weight, except that the interval is definitely smaller for the caesium uranyl chloride. The other three salts, so far as this deter- mination goes, must be regarded as having the same average interval. It is likewise difficult to distinguish with certainty differences in the intervals of different bands in a given salt, except that the C band has in general a smaller interval than the other series, or of a given band in the different salts, excepting in the case of the caesium chloride. [ K 3 C 1 1 ) E I i \ NH4 1 j Rb Cs . 40 2.0 20 4p 1 Fig. 54. 66 FLUORESCENCE OF THE ITRANYL SALTS. At the same time, while not obviously systematic, the variations in these values are considerably larger than those resulting from the measurement of the interval of any given series, taken by itself, which should not exceed, at most, 0.2. Table 21. — Average intervals at -\-W° C. for the four double chlorides. Series. K. NH4. Rb. Cs. Av. by- series. B C D E A Av... 83.42 83.11 83.30 83.00 83.23 83.34 82.99 83.21 83.81 83.65 83.49 82.97 83.17 82.97 83.77 82.32 82.50 82.85 83.45 82.85 83.14 82.89 83.13 83.31 83.37 83.21 83.40 83.27 82.80 83.17 These seeming discrepancies are not to be considered as wholly accidental, but as being due to the fact that the bands are complex, and variously so, as will appear from the study of these spectra at low temperatures. While he determinations thus ' far described may be regarded as indecisive as to small differences of interval between the various series and salts, excepting as noted above, the iufluence of molecular weight upon the posi- tion of bands in the spectrum is unmistakable. In tables 29 and 30 (at end of chapter) the almost universal and fairly regular increase in the frequency of each band as we pass from potassium to caesium is sufficiently evident. In figure 55 this general shift, which is present in all the groups and affects all series, can be seen at a glance. Almost the only re- versed shifts occur in the case of those bands of the csesium spec- trum which show anomolous plac- ing in the spectral groups. In table 19, where the groups are units, the accidental errors pertaining to individual bands are submerged in the processes of aver- aging and the shift with molecular weight appears as a still more — — 1 [■■■-■ 1 1 1 1 1 "X 15 16 17 18 19 20 21 SERIES B K 1 1 1 1 1 1 1 NH,1 1 1 i 1 1 1 Rb 1 1 1 1 1 1 1 Cs 1 1 1 1 1 1 1 C K 1 1 1 1 1 1 1 NH, 1 1 1 1 1 1 1 Rb 1 1 1 1 1 1 1 Cs 1 1 1 1 1 I 1 D K 1 1 1 1 1 1 NH, 1 1 1 1 1 1 Rb 1 1 1 1 1 1 Cs 1 1 1 1 1 1 E K 1 1 1 1 1 1 NH4 1 1 1 1 1 1 Rb 1 1 1 1 1 1 Cs 1 1 1 i 1 1 A K 1 1 1 1 1 1 NH4 1 1 1 1 1 1 Rb 1 1 1 1 1 i Cs 1 1 1 1 1 1 .J ^^4/* .5,6a' .4,8;.| INTIMATE STRUCTURE ON COOLING. 67 systematic phenomenon. Ignoring group 7, in which the bands are displaced by absorption in a manner to be discussed later, we find the following shifts to exist. Table 22. — Shift of the groups. Group 2 3 4 6 6 Shift 5.4 5.1 5.3 5.2 5.1 Average shift from K to Cs» 5.2 Ths shift is therefore to be regarded as approximately uniform throughout the spectrum. The shift is much greater between NH4 and Rb than in the other cases, the averages being as shown in table 23. Table 23. — Average shift of groups. K-NH4 1.6 NH4-Rb 2.7 Rb-Cs 9 It will be noticed that in this discussion the order of molecular weights used is K, NH4, Rb, Cs — NH4 being placed between K and Rb instead of in its proper position. This is in accordance with the results of Tutton/ who has shown that in various optical properties of crystals which depend on the molecular weights, NH4 always lies between K and Rb, as though its effective molecular weight were larger instead of being smaller than K, THE EFFECTS OF TEMPERATURE. The narrow, line-like bands into which the ordinary uranyl spec- trum is resolved at low temperatures^ form a rather complex aggre- gation separable into a series of identically arranged groups corre- sponding to the unresolved bands at +20°, but related to the over- lapping components of the latter in a manner not easily capable of direct determination. It was deemed of especial interest, therefore, to observe the effect of cooling on the double chlorides, where the relation, owing to the partial resolution at +20°, should be more obvious. For this purpose a crystal, C, of the salt to be examined was mounted within a long cyUndrical Dewar flask, D, with unsilvered walls (fig. 56). The carbon arc A was f ocussed on the crystal by the lens L. A water- cell W was inserted between the arc and the condenser. The light- filter F was opaque to all but the violet and ultra-violet rays used for excitation. Observations with the Hilger spectroscope fl", a portion of the collimator of which is shown, were made through a second filter E opaque to the exciting light but transmitting the fluorescence. The arc and specimen were well screened by an opaque box BB, When it was desired to photograph the spectrum a camera was substituted for the observing telescope of the spectrometer. The control and adjustment of temperature were effected by attach- ing the crystal at the upper end of a vertical copper rod which could be ^ Tutton, A. E., Crystalline Structure and Chemical Constitution. (London, 1916.) ^ See Becquerel and Onnes, 1. c. 68 FLUORESCENCE OF THE URANYL SALTS. immersed more or less deeply in the liquid air by raising or lowering the Dewar flask. To preclude the gathering of frost or moisture on the surface of the crystal, it was kept during the entire experiment at a sufficient distance below the lip of the flask, where it was surrounded with the dry atmosphere above the slowly evaporating mass of hquid air. Measurements of the temperature were by means of a small coil of fine copper wire mounted at the same level as the crystal, so as to have always, as nearly as possible, the temperature of the latter. Changes in the resistance of the coil were indicated on the sheet of a Callender recorder, carefully cahbrated to read directly in degrees centigrade and adjusted for a range from +20° to —200°. B C D E A 85' / K /n J J A /^ j/l 6 B' C C < ti E E* A /f FiQ. 56. Fig. 57. The crystal was mounted so as to cover a transverse slot in the copper rod. It could thus be illuminated either from the front, as shown above, or from behind by light transmitted through the slot. The latter arrangement was employed especially in the study of the absorption spectrum. When the substance, excited to fluorescence in the manner already described, was gradually cooled to the temperature of liquid air and the spectrum was observed through the Hilger spectrometer, the fol- lowing changes were noted: (1) The bands become narrower and better defined until at the temperature of liquid air they correspond in appearance to the usual line-like bands characteristic of the fluorescence spectra of the uranyl salts at low temperatures. (2) As the temperature falls the bands are gradually resolved into doublets. One component of each doublet becomes rapidly brighter, while the other frequently becomes more indistinct and sometimes disappears. The general effect is that of a shift toward the violet INTIMATE STRUCTURE ON COOLING. 69 amounting to about a third of the distance between the original bands. The nature of this apparent shift is as follows: Each band at +20° may be regarded as an unresolved doublet, of which in general the member of longer wave-length is relatively so much the stronger that its position determines approximately the location of the crest of the composite band (see fig. 57). The effect of cooUng is to resolve this doublet into separately distinguishable bands and at the same time to cause a subsidence of the stronger and an increase of the weaker member. The member of the shorter wave- length usually becomes dominant at low temperatures, and in so far as this occurs the arrangement of the spectrum appears to be undis- turbed but shifted toward the violet by an amount representing the width of the doublet. There are, however, certain exceptions to this rule, so that the relation of the resolved spectnmi to that at +20*^ is not so simple as th^ above description would imply. The appearance of the group, if this be its real structure {i. e., a set of neariy equi- distant doublets, the distance between the members of all the doublets being nearly the same), would then be as shown in figure 57. At +20°, £', C", D\ E\ and A' are entirely concealed by the over- lapping of the bands. At —185°, 5, C, D\ E, and A may or may not be visible, according to their intensity or the completeness of the resolu- tion, which in fact varies greatly in different parts of the spectrimi. It will be noticed that in the lower diagram in figure 57, D and not D' is the dominant component. This is a condition which obtains in the ammonium chloride, with the result that C and D, which appear to have replaced the strong C and D bands of the spectrum of +20°, are near together, D and E^ far apart, and the symmetry of the group is impaired. Similar compUcations occur likewise in the spectra of the other double chlorides. To illustrate the application of this assumption, the spectrum of the ammonium uranyl chloride has been mapped in the manner shown in figure 58, in which the fluorescence bands of the 8 groups as they occur at —185° are shown in their relation to a hypothetical grouping given at the head of the diagram. This grouping consists of the set of imagined doublets of which, as in a previous paragraph, the spectrum at +20° is supposed to be made up. The spacing for each doublet is that determined from the observed average shift on cooling and the relative divergence from this arrangement is shown for all the bands of each group. A scrutiny of the fluorescence spectnmi at —185°, group by group, by means of this diagram, affords very satisfactory confirmation of this hypothesis concerning the apparent shift. It is obvious: (1) That not all the components B,C,D, £', and A will necessarily be visible in every group of the resolved spectrum. 70 FLUORESCENCE OP THE URANYL SALTS. (2) That lack of resolution in any region may give the appearance of a single band with intermediate crest in place of the doublet. (3) That the position of crests of the xmresolved doublets at +20° will not necessarily coincide exactly with that of either component. Bearing these points in mind, it will be seen that were resolution complete all the observed bands of the spectrum at —185° would probably fall into the system proposed above. We may imagine that the difference between the resolution of the bands C and D, for example, as seen in figure 54, is produced by changes in the unresolved doublets at +20° when the temperature is reduced to —185°, of the kind indicated in figure 59. The doublet CC" forms a single band with crest nearly coincident with C at +20°, and this owing to the sub- sidence of C and growth of C takes the resolved form shown at — 185°. In the case of D, however, the unresolved . c Jk ( B 0' ^ e' A ^' ! ! 1 1 1 ! 1 1 ! 1 I . 1 . 1 1 r 1 1 1 1 1 1 1 1 . 1 1 II 1 1 1 1 1 1 1 -2 0" Fig. 58. -I8S Fig. 59. -20 -185 band has an intermediate crest at D, but is really composed of over- lapping components D' and 1>" which are separately visible at —185°. The wave-lengths and frequencies of the bands in the resolved spectra of the four double chlorides, as observed when excited at the temperature of liquid air, are given in table 31 at the end of the chapter. The nomenclature used in this and subsequent tables is chosen to indicate as far as possible the relation of the bands at —185° to those. at +20°. Thus Bi, B^^ etc., denote components of 5, etc., which have been rendered visible by the resolution effected by cooling. The explanation offered above to account for the relation between the spectra at +20° and at —185°, and which was illustrated in the case of the anunonium uranyl chloride (see fig. 58), was confirmed by observations upon the spectrum of that salt at intermediate tempera- tures. It was thus possible to watch the gradual appearance of the INTIMATE STRUCTURE ON COOLING. 71 'gsj' I 1+20' B +20** : B Rfr UlB5* ^ iin C ^ D 1+20'' Cs UBS' iti ^ fe. components characteristic of the spectrum at low temperatures and the simultaneous fading away of those dominant at +20°. The same explanation apphes equally well to the potassium and rubidium double chlorides. In the case of caesium uranyl chloride the relations are comphcated by the further resolution of these components, so that the connection with the original complexes is less easily traced. To indicate the general character of these resolutions and the apparent temperature shift which results therefrom, the positions of the bands of group 6 at —185° are plotted for all four chlorides (see fig. 60). Intensities of the —185° bands are indicated roughly by the height of the lines. The corresponding crests of the bands at +20° are represented by dotted lines. Group 6 was selected because it offers better examples of the further breaking-up of the components and of other phases of the process of resolution than do groups toward the red in which resolution is progressively less complete. Two questions which were left undetermined in the study of the spectra at +20° may be regarded as settled by these measurements of the bands at — 185°. (1) That the intervals are not the same for aU series in a given spectrum is clearly estabhshed. For example, the components Ci, C2, which take the place of the C bands in all four spectra, have distinctly different in- tervals, i. 6., 84.00 for Ci and 82.75 for C2. It is noteworthy that C2, which becomes the crest of the group in place of C also, has the small interval. It might be questioned whether these so-caUed components are not merely accidental neighbors rather than products of the same vibrating system, but for the fact that they are present in aU the spectra and have very nearly if not precisely the same relative positions to each other in aU. (2) The average interval of aU series in the spectnun of the csesium chloride (82.80) at +20°, which causes the notable displacement of the bands of that substance, becomes 83.44 when we take the average of the intervals of the bands at —185°. That is to say, it is, within the errors of observation, the same as the general average for the other salts. On the basis of the measurements at low temperatures (see table 24), we must conclude that the four double chlorides have approx- imately the same average frequency interval. The averages given in table 24 are obtained from the data of table 32, which contains the frequencies of all the fluorescence bands observed in the spectra of the four double chlorides when excited at the tempera- ^ ^ +20° II h . t 1900 Fig. 60. I I 1950 72 FLUORESCENCE OF THE URANYL SALTS. ture of liquid air. As in the corresponding table for +20*^ (table 30) the arrangement is by series. Table 24. — Average intervals of the fluorescence series at ~18S° C. Series. K. NH4. Rb. Ca. Average. Bi.... B2.... Ba.... Ci.... C2.... Di.... Di' 83.9 83.1 83.5 84.9 82.7 83.1 83.0 83.2 84.2 83,6 83.0 83.4 83.53 83.33 84.1 82.7 83.8 84.0 82.9 83.6 83.7 82.8 83.1 84.5 83.6 83.6 83.2 83.5 84.18 82.78 83.40 D2.... D2' . 84.1 84.2 84.0 83.98 E2". 83.6 82.5 83.3 83.1 83.10 83.40 82.83 83.50 82.1 83.6 Ai.... A2 ... 83.3 83.4 83.58 83.32 83.50 83.44 THE ABSORPTION SPECTRA. A glance at the absorption spectra of the double chlorides, obtained by viewing through a spectroscope the light transmitted by the crystals at room temperature, shows the same higher degree of resolu- tion that characterizes the fluorescence spectra of these salts. The saUent feature is a series of strong, rather narrow bands, equally spaced as to frequency, like the broader bands of the other uranyl compounds. The interval, as in all uranyl absorption spectra, is distinctly smaller than the fluorescence interval. Between these are several series of weaker bands. The complete mapping of the absorption spectra is difficult. It can not be done visually, since the bands extend out into the darkness of the ultra-violet. Photography adds considerable detail, but does not greatly extend the range toward the shorter wave-lengths on account of the rapidly increasing opacity. In the brighter regions of the spectrum, on the other hand, more can be seen with the eye than can be found on the photographic plate. The data which we have obtained and which are presented in the tables at the end of this chapter have been procured by supplementing the photographic method, wherever desirable, by visual observations. A great variety of light-filters and combinations of light-filters have been employed in different parts of the spectrum, with widely different exposures for the strong and weak bands. The thickness of the trans- mitting layer has likewise been varied as far as the available material would permit. We are convinced, however, that the extreme limits of the absorption, in both directions, have not as yet been reached. By using crystals of unusual thickness, especially prepared for this work and sometimes by mounting several crystals one behind the INTIMATE STRUCTURE ON COOLING. 73 other, so as to greatly increase the depth of the transmitting substance, it has been found possible^ to greatly extend the absorption spectrum toward the red. Since the crystals are of a greenish-yellow color, they become rapidly transparent as the light admitted is changed from blue to yellow; hence the use of increasingly thicker layers to bring out the absorption bands. To a certain extent the crystal acts as a screen to absorb the blue light which would cause fluorescence; nevertheless it was found necessary to interpose orange or yellow screens of different densities to eliminate fluorescence in a region where ordinarily it is at a maximum. At first colored glasses obtained from Dr. H. P. Gage, of the Corning Glass Company, were used as filters; later, solutions of potassium bichromate of varying concentration. It is evident that the screening must be constantly changed when hght from the arc is used as a background for bands of increasingly longer wave-length. It was thought that a beam of monochromatic hght could be used as a background and thus obviate exciting the crystal to fluorescence, but a preliminary study indicated that such a beam of dispersed light could not be made of sufficient intensity to bring out the dimmer bands. "I — ~l ■52 Xt 4T+ltTf •48 yU WAVI-Lf HOTH. I ihIj -*--<-. ' ' ' ■ ■ I I l. hil 1 1 U! !i'i ! -H- -fi-l- III I ' ' u i 'i . i 1 1 1 h u i 'l i ' u ! ^ _ ie|oo ibIoo 20|00 2tl00 Fig. 61. — Fluorescence bands are indicated by lines above the horizontal. Old absorption bands are indicated by dotted bands below the line; new absorption bands by solid bands below the horizontal. The plot shows only a portion of the complete spectra of the following salts at +30° C: (1) potassium uranyl chloride; (2) ammonium uranyl chloride;. (3) rubidium uranyl chloride; (4) CEesium uranyl chloride. In figure 61 is pictured a portion of the fluorescence and absorption spectrum of each of the double chlorides studied. Fluorescence bands are designated by lines above the horizontal line. The older, well- estabHshed absorption bands are designated by dotted lines below the horizontal and the new bands by sohd lines below the horizontal. The relative positions of the fluorescence and absorption bands are readily seen. These bands appear to be of two distinct classes: 1 Howes, H. L., Physical Reviuw (2), xi, p. 66. 1918. 74 FLUORESCENCE OF THE URANYL SALTS. (1) Most of them at +20°, as may be seen from the diagram and from table 25, in which they are listed together with the corresponding fluorescence, are reversals of fluorescence. These do not form a con- tinuation of the absorption series lying farther toward the violet, nor can they be grouped in series having the absorption interval of 71=»=. In all four species every fluorescence band of groups 5 and 6 has its Table 25. — New absorption hands at +20 °c. Potassium uranyl chloride. Ammonium uranyl chloride. Absorp- tion. Fluores- cence. Fluores- cence. series. Absorp- tion series. Absorp- tion. Fluores- cence, Fluores- cence series. Absorp- tion series. 1802.1 1801.4 B 1802.5 1803.1 B 1820.2 1819.3 C 1820.8 1820.7 C 1836.5 1837.6 D 1838.9 1839.7 D 1846.0 1855.3 c 1848.8 1857.8 c 1855.3 E 1856.9 E 1865.0 1869.4 d 1869.2 1871.8 d" 1869.6 A 1871.8 A 1879.0 1885.1 e 1886.5 1906.2 1886.8 1904.6 B C 1884.7 B 1902.2 1901.5 C 1924.2 1923.2 D 1920.9 1920.1 D 1942.3 1940.5 E 1937.6 1938.3 E 1957.4 1956.3 A 1954.7 1953.5 A Rubidium uranyl chloride. Caesium uranyl chloride. Absorp- tion. Fluores- cence. Fluores- cence series. Absorp- tion series. Absorp- tion. Fluores- cence. Fluores- cence series. Absorp- tion series. 1740.0 1741.6 C 1791.5 1789.7 A 1778.7 1777.8 E 1808.0 1808.6 B 1789.5 1789.4 A 1829.2 1827.5 C 1806.1 1806.1 B 1843.0 1840.5 D 1823.2 1822.8 C 1846.4 P 1834.9 1841.6 6 1861.2 1873.0 1859.1 1873.1 E A 1841.5 D 1859.8 1859.8 E 1890.7 1891.1 B 1872.0 1873.1 A 1911.1 1910.4 C 1889.0 1890.0 B 1923.8 1923.6 D 1907.2 1905.5 C 1944.4 1942.7 E 1926.7 1925.0 D 1957.8 1955.7 A 1941.7? d 1944.0 1943.5 E 1952.0? e? 1958.7 1957.1 A corresponding absorption band, and this relation extends to some of the bands of group 4. Indeed, the suspicion would seem warranted that were the proper experimental conditions attainable throughout the spectrum, every fluorescence band would be found to have its related absorption band and to be reversible in the sense in which that term is defined in a subsequent paragraph. INTIMATE STRUCTURE ON COOLING. 75 (2) The remaining bands listed in table 25 are not reversals of fluorescence. They belong to existent absorption series, of which they are the members of greatest wave-length as yet observed. It should be noted that special precautions were taken to avoid bias. They were not sought for by locating the fluorescence bands and looking for reversals, but found under conditions of illumination which Table 26. — New absorption hands at —185° C. Potassium uranyl chloride. Ammonium uranyl chloride. Absorp- tion. Fluores- cence, Fluores- cence series. Absorp- tion series. Absorp- tion. Fluores- cence. Fluores- cence series. Absorp- tion series. 1941.7 1947.6 1954.7 1960.4 1965.8 1972.4 1977.6 1984.9 1989.3 1998.0 2008.8 1940.0 E2' d2 ei Ol &2 &3 C2' C2" di d2 1945.9 1953.5 1956.6 1963.5 1967.7 1973.6 1977.1 1981.0 1984.9 1992.0 1996.8 2002.8 2006.8 2014.1 1945.0 1953.7 E2" Ai d2" ei 62" bi &2' &2 C2 1963.9 1972.3 1977.8 Bi B2 B3 1968.7 Bi 1977.9 B2 1997.2 2007.4 Di D2 1992.7 C2 di' dl" d2" Rubidium uranyl chloride. Cfesium uranyl chloride. Absorp- tion. Fluores- cence. Fluores- cence series. Absorp- tion series. Absorp- tion. Fluores- cence. Fluores- cence series. Absorp- tion series. 1944,4 1952.4 1954.7 1958.1 1963.9 1973.9 1981.0 1985.7 1995.6 2005.2 2010.1 2016.1 ^2" 1953.9 1956.6 1958.9 1967.0 1970.8 1974.3 1978.2 1982.9 1987,7 1991.3 1997.6 2005.6 2009.6 2016.1 2022.2 ds" 62' ai' 1957.9 Ai C2' 62 fei' bi" h' 62" bz Cl C2 di da' d2" C2 da' d2" 2003.7 Di 1997.6 Cl 2008.5 2014.9 Di rigorously excluded fluorescence, and in many instances their existence and place was checked by two observers working independently. The fact that practically the entire group was in approximate coin- cidence with fluorescence was an unlooked for result of which we became aware only after the measurements had subsequently been plotted. The expectation was that these bands would prove to be members of the absorption series lying farther toward the violet. 76 FLUORESCENCE OP THE URANYL SALTS. A search by similar methods failed to reveal any bands of class (1), mentioned above, in the spectra of the crystals when cooled to the temperature of liquid air. No selective absorption could be detected beyond the violet end of group 6, 1/X 1940, and while a considerable number of new absorption bands were detected, nearly all of these (see table 26) were found to be members of series already recognized. The exceptions, two each in the spectra of the ammonium, rubidium, and csesium double chlorides, do not appear to be related to the fluores- cence. Coincidences between fluorescence and absorption are of the sort already established as characteristic of the reversing region. Table 27. — Average intervals of absorption series at -\-W° C. Series. K. NH4. Rb. Cs. Av. h 70.5 71.6 70.4 70.9 70.6 70.9 70.3 70.9 70.0 70.5 70.9 70.5 70.3 70.3 69.7 B c 70.8 70.4 68.8 70.7 70.3 70.4 y d' d d" e 71.1 70.6 70.0 70.4 71.0 70.3 69.7 70.8 71.2 70.8 69.6 70.6 70.8 P." a a" Av.... 70.0 69.3 70.5 70.4 70.3 70.6 The failure to find the bands in groups 5 and 6 is not surprising. They are sufficiently difficult objects at +20°, where two or more components are blended into a broader band. The existence of these components at —185° may be regarded as probable, but they were invisible under the conditions which we have thus far been able to obtain. The absorption spectra of the double chlorides do not exhibit the same remarkable approach to identity of structure and regularity of arrangement manifested in the fluorescence spectra. Upon analysis, however, they are all found to consist of series having intervals of approximately 70 frequency units. As may be seen from table 27, this interval for a given series is very nearly the same for all four salts. The average interval for all the series of a given salt is constant within the errors of observation. These averages are based on the values in table 33 at the end of this chapter. The absorption bands, unUke those of the fluorescence spectrum, do not appear to fall into a succession of strictly homologous groups, but this is because some series disappear, while others increase in strength INTIMATE STBUCTUBE OX COOLXN'G. toward the violet. A group near the ^uoreset-nce region. Therefore. differs notably in aspeci from one in tlie extreme violet, and it is difficult to h-3>se conclusions on the location of the ceniei^ of the groups. as vrzs done in the study :: the fluorescence sr-ec:r:-. A5 may be obser.ei in nguxe o2. where the ninth group for the : :ur spe^r-ira at —20" is plottedj the dis:ance5 between the consecutive bands 5-re less nearly equal than the distances between fluorescence bands. It is als^D e'ddent from tnis iig^ure that with increasing molecuLir weight there is a general shift t : v ,":;rd the violet. The shift is apparently less systematic than with the duorescence bands and several reverse shifts seem to occur. In general, however, 'he total (iisnlacenient is approximately :hr T>^-.T :'Serv^ra lor ^uoresCfrnce. i. e,, o frequency units from T)otassiiiin to cc^siurn. EriTCT OF TEMPERATItRE OX THE a3Sj?.? TIOX SPECTRA- In the srudy of the absorption c: the double chlorides at — 1S5". a m^dihcation of the method described in a predus paragr:iph in the investigation of the nu:rescen:e at low temneratures was made. 'See ng. o'o. The cr;.*stal under observaticn was niotnited within a Dewar :5a sh and submerge^! in :ne cr^-: air. Light ^as tra:tsm::ted tai msteau ci c^mg reiiecte^n irom its s^urface and a nitrogen-nlled tungsten lamp w:^s . in general . substitute-d for the carbon arc. Both p h : ^ographic and visual methods were tried, and in the re.ersing region, especiahy. where £uorescence and absorption overlap, much attention was given to uie selection of color-screens to exclude ^uorescence frcni the portion under consideration. A com^plete list of the abs irption bands observedat — ISo'willbefoundintable-Sl. The three mc^t obdous resulis of cool- ing to the temperature of Ucuid air are: 'Ij a general shift to'^'ard the ^dolet : 2 , a great iacrease in themumber of bands: 3) a very decided narrowing and sharp-ening of the ban«ds. These changes are readily accounted for by the ass^umpti- -n already made, in this chapter, that the ban 'ds at — 20"C.arec:n- cealed doublets andtha" thee^ect ofcool- ms IS tores-: ■~ r- ^"ry em wnue sunuit aneou slv Fzc-. e: reducing the strength of the str-mger and increasing the strength of the weaker component. The apparent shift thus prcduce-i will var^" from zero to o or more units, according to the distance between the com- ponents. 78 FLUORESCENCE OF THE URANYL SALTS. A few bands at — 185° are so located with regard to the +20° bands that to explain them by this theory we must suppose them to be too feeble at +20° for detection and greatly increased in intensity by cooUng. There is also evidence in places of further resolution into closer narrow doublets and as the degree of resolution is not always the same with fluorescence and the corresponding absorption, this is a soxirce of trouble in the attempt to find the fluorescence series which belongs to each series in the absorption spectrum. Every low-temperature band, however, falls into a series of constant frequency, whatever its position or degree of resolution. The effect of temperatm-e on the average intervals can be studied by comparing tables 27 and 28. Although the intervals range from 69 to 71, there is little that can be termed systematic in the variations. At liquid air temperature, where two or more components are present, we have used subscripts. Thus di, corresponds to Di, ^2 to D2, etc. Where the reversal is doubled in the manner shown in flgure 63, we have designated this doublet as di and di', etc. The average interval of each salt is approximately the same at both temperatures. It will be noticed in table 27 that 70.28, the average of the c components is smaller than the h, d, c, or a averages. This is of interest because the strong C series, which join these series, are also the shortest of the fluorescence series. Since the — 185° bands are very sharp and easy to locate, no doubt the differences found in table 28 are indicative of real variations in 'the constant-frequency intervals. It does not follow that the smaller intervals are con- fined to one salt or one set of bands, however, since, as has been noted in the case of series Ci and C2 of the fluorescence series, the maximum difference in interval may be associated with two series which are nearly co- incident. The comparison of table 27 with table 28 shows that the effect of changing temperature on the ^' average interval of a salt is almost negligible, but that the two com- ponents of one series of the +20° spectrum may vary by as much as 1.9 units in frequency interval. The character of the change in the absorption spectra when we pass from +20° to —185° can best be seen in detail by plotting a single group in the spectrum of each salt, as has been done for group 9 in figure 62. A better idea of the phenomena of cooUng, as a whole, is obtained by means of maps like those in figures 64, 65, 66, and 67, in which all the bands of fluorescence and absorption are given at both temperatures, first in a single line as they occur in the spectrum of each salt. Fluorescence is indicated by vertical lines above the horizontal and absorption below. Length of line indicates roughly the strength d' ►, iip M If i"iffii' i| i'***'|iii III 1 i II II II II 11 II II 1 1 1 1 II 1 11 1 II 1 II II i 1 II li [l II 1 1 1 B 1 1 1 II II II : II II 1 , 1 1 i( 1 I ' 1,111. tl 1 I t 1 +";i,i,,, ,, ) Il I J 1 1 1 1 . 1 1 . . 1 i 1 i 1 , iri 1 I 1 1 1 Ml i'jIm |i I j |i| ' jl i <| 1 T-l j 1 '1 I II II 1 1 II 1 1 1 1 1 1 I : ' 1- 1 I 1 : j II 1 1 1 1 1 1 1 1 1 ■ ' 1 1 1 1 1 t 1 1 1 ■ -.85-1 ., 1 i,ii 11L...1 1 Li. *.c UJ. Xl A^ URANYL RUBIDAIH CHLORIDE C, V D GREEK WHFTE URANYL AMHONIUM CHLORIDE CREEK £'^ nrr URANYL CAESIUM CHLORIDE C^ Fig. 71. From this figure it will be seen that bands C, E, and A of xu-anyl potassium chloride appear as doublets, polarized at right angles. Band B has no green component visible, but, as will be shown in a subsequent paragraph, at —185° a green component is present, which lies nearer the red than B^. Bands Cg and C^ are the two components of band C, while no component of band D has been found on the white side. Bands E and A are also well resolved; the white component of band E is of longer wave-length than the green component, while the white components of C and A, and probably of B, are of shorter wave- length than their respective green components. 106 FLUORESCENCE OF THE URANYL SALTS. The uranyl amraonium chloride group shows a strong similarity to the preceding group. All except band B appear as polarized doublets. Components D^ and A^ were discerned only with the greatest diffi- culty. The uranyl rubidium chloride group is very similar to the uranyl potassium chloride group. Band Bg is missing, but as in the potassium chloride, there is a —185° component to the red of B^. Component Cw has a position nearer Dg than has Cw in the preceding spectra. This is also the condition existing in the csesium-chloride spectrum, and it is possible, since no D^ component is visible in either spectrum, that D^ is very dim, and hidden in C^. Uranyl csesium chloride gives the most satisfactory set of fluores- cence bands, since both components of band B are present, and the C, E, and A components are very well separated. It is interesting to note that Bg is of longer wave-length than B„, as is the —185° com- ponent of Bg in the preceding salts. It has been previously stated that the absorption spectra, like the fluorescence spectra, are composed of series, which begin with the bands which terminate the fluorescence series. The absorption bands which lie nearest the fluorescence region can also be arranged in recm*- ring groups. The absorption series will be designated fe, c, d, e, a; since they join the B, C, D, E, and A fluorescence series, respectively. The e and A series are the strongest in the reversing region, but grad- ually vanish, while the D series becomes stronger toward the ultra- violet. Figure 72 gives a typical absorption group for each of the four salts. A^ before, the components above the line belong to the green; those below to the white polarization. By comparing the uranyl potassium chloride absorption group in figure 72 with the fluorescence group of the same salt in figure 71, it will be seen that there is no bg component present, as there was no Bg component present, but that Cp, dg, e^, and ag, corresponding to series Cg, Dg, Eg, and Ag are present and that there are no other series represented. Although the relative intensities of the absorption com- ponents are almost reversed when compared with the relative intensi- ties of the fluorescence bands, the same spacing exists between the green components of both fluorescence and absorption. In the white polarization group, c^, corresponds, in position, to C^, and e„ to E^, while b^ a„ serves both B'^ and A^ series in the following way: B^ is the first member of each fluorescence group, while A^ is the band of the preceding group which is nearest to B^. As the fluoresence intervals of both the A and B series are approximately 83 frequency units, and A^ is 12 units distant from B^, the reversing band of the a„ series must coincide with the second member of the by, absorption series, since it is 71 units from the reversing band B^ or &„,. The d„ component is absent, as is D^, and there are no superfluous series. POLARIZED SPECTRA OF DOUBLE CHLORIDES. 107 The absorption group of the uranyl ammonium chloride is very similar to that of the potassium chloride. Again, the bg component, like the Bg component, is absent, but the other fluorescence series are represented by absorption series, save that no component of d was found to join the very weak D^ fluorescence band. Uranyl rubidium chloride shows a grouping analogous to that of the potassium and ammonium chloride, while the uranyl caesium chloride group is only slightly different. A bg series is present, which is properly related to the Bg series, so that bg and b^ are the same relative positions as are Bg and B^. POLARIZED ABSORPTION GROUPS +80" URAHYL POTASSIUM CHLORIDE ORE EN WHITE &>f btv B 9 > AB I II II II II II ^-20°| II I I I I I c e da c e ha. WHITE-IBSHI I I II I I II I I I I I I WHITE URANYL RUBIDIUM CHLORIDE C D EA C D EA TTTT rr GREEN f 20 C D EA C D EA I I II I I II GREEN -IBS _LL U c d, e a. c d- ecL c I I II M II I WHITE +20' B P E A B C E A B I 1 I I I I M I I III I lljc e aJ} c e ah WHITEM85- II III M I 1 1 1 I I I I I I I I URANYL CAESIUM CHLORIDE aCDEABCDElA eREEN-4-20°l lllllllll^'Cf^ eab c d e ab c SREEW-ieS WHITE 4-20 BCEABCEA •I I I I I I I I II 1 1 1 I I 1 1 1 111 I I m ^ I I III M I inn bee cbb c e a WHITE-ieS I I I I 1 1 I I II I III leloo laloo I III III 20|00 I 2ll00 FiQ. 75. nearly of the same interval, not only in the same salt but in all the salts. When, however, we make further averages of the average intervals from table 37, taking the nlean of all green components of fluorescence, then of all white components, for each salt separately, and do the same for the absorption intervals, we find an approach to systematic arrange- ment which is suggestive if not altogether conclusive. (See table 37.) Both components of the fluorescence spectrum show an average interval in the inverse order of the molecular weights, and while the absorption series do not give so decisive an indication the salts of lesser POLARIZED SPECTRA OF DOUBLE CHLORIDES. 117 molecular weight, NH4 and K show again a longer interval than do Rb and Cs. Averaging by series affords no such direct indication* as to differences of interval, as will appear from table 38. It will be noted that while the averages for the green and white com- ponents of fluorescence are in very close agreement at +20° and also at —185°, there is a difference of about 0.5 between the averages for +20° and those for —185°; also that the interval is greater for each individual series at —185° than at +20°, with the single exception of ew. This difference does not appear, however, in the case of the absorp- tion intervals. Table 37. — Average freqiiency intervals, -\-20° C. and —185° C. Fluorescence series. Green component. White component. Series. K NH4 Rb Cs Series. K NH4 Rb Cs Ba 82.9 82.6 83.3 83.3 82.9 0(0. . . 83.0 82.1 83.4 82.9 83.6 83.3 82.9 82.8 c,.... D,.... E,.... A,.... 81.9 83.0 83.2 83.4 82.8 83.5 83.3 83.8 82.2 83.1 83.5 83.2 A„... 83.5 83.1 83.5 82.9 83.1 82.4 Absorption series. Series. K NH4 Rb Ca Series. K NH* Rb Cs be 70.5 70.6 70.5 71.0 70.8 71.5 71.3 71.3 70.3 70.4 70.0 70.8 70.5 Cff H 71.3 71.1 70.8 70.7 70.7 70.5 70.8 71.5 71.2 70.2 71.7 71.1 70.4 71.6 71.7 71.3 70.8 70.4 70.4 71.4 Average frequency intervals, —186 °C. Fluorescence series. Green component. 1 White component. Series. K NH* Rb Cs Series. K NH4 Rb Cs C,.... Do.... E,.... A,.... 84.0 82.5 83.7 83.3 84.1 83^4 84.4 83.2 84,4 83.4 82.6 83.9 84.1 82!2 83.3 83.6 84.1 c„... 84.1 84.8 84.1 84.4 83.2 83.9 83.6 83.3 83.0 84.0 83.1 84.1 83.1 Absorption series. Series. K NH* Rb Cs Series. K NH4 Rb Cs ha 70.9 70.7 71.3 70.6 70.5 70.8 71.4 71.0 71.1 c-„... 71.3 69.8 71.3 70.6 70.8 70.3 70.6 70.7 cp dg H «ff 71.6 70.3 71.8 71.0 70.9 71.1 71.6 70.9 71.4 70.4 70,7 71.1 70.9 70.6 118 FLUORESCENCE OF THE URANYL SALTS. Table 37. — Average frequency intervals, -\-W° C. and 185° C. — continued. General averages of intervals (by salts). Fluorescence. NH4. K. Kb. Cs. Green +20° and -185°... White +20° and -185°... All fluorescence 83.60 83.78 83.25 83.43 83.25 83.42 83.19 83.16 83.69 83.34 83.33 83.17 Absorption. Green +20° and -185°... White +20° and -185°... All absorption 70.99 71.02 71.07 70.96 70.96 70.55 70.74 70.70 71.00 71.01 70.75 70.72 Table 38. — General averages of intervals (by series). Fluoiescence. Series. Green. Av. Series. White. Av. +20° -185° +20° -185° Eg.... Be 83.33 83.-37 83.70 84.18 83.83 82.68 83.35 83.70 83.72 83.53 82.53 Ew . . . Bw... Aw... 83.37 83.23 82.80 83.05 83.75 83.73 83.17 83.49 83.74 Ag.... Dg.... Cg.... Av.... 83.26 83.23 82.38 Cw... Av.... 82.78 84.10 83.44 83.05 83.54 83.37 83.05 83.66 83.36 Absorption. +20° -185° Av. +20° -185° Av. 6(7 ha 71.07 71.40 71.23 ay,.... d,n 70.82 71.00 70.90 71.02 71.00 70.70 70.90 71.00 70.80 «ff d, Cff Av.... 71.02 70.58 70.95 70.90 71.02 70.98 70.96 70.75 70.96 Av. . . . 70.50 70.35 70.42 70.90 71.07 70.97 70.81 70.77 70.78 On the other hand, differences so large are not to be regarded as errors of observation, it being possible to determine the average inter- val of any series, excepting possibly Ag and A^, which are very weak ahd rather vague, within about 0.2. It does not follow, however, that the bands are really thus irregularly placed. The discrepancies are due rather to the fact that resolution is not equally complete in all portions of the spectrum and that on cooling the crystal structure was more or less disturbed and the polarization always much less complete. The anomalous values above 84 frequently observed at —185° (see table 37) are probably due to varying components of the opposite polar- ization superimposed on the bands in question and producing a false POLARIZED SPECTRA OF DOUBLE CHLORIDES. 119 shift. Thus, for example, the position of C^ would be modified by the presence of the overlapping of Dg or Cgj Dg by Cg, etc. In short, it is probable that if observ^atipns could be had on crystals which at —185"^ preserved their structure, the difference in interv^al between +20° and — 185° would disappear. THE INFLUENCE OF MOLECULAR WEIGHT LTON THE POSITION OF BANDS. While some doubt may be felt as to the validity of the suggestion, based upon the averages presented in the foregoing paragraphs, that there is a relation between frequency interv^als and the molecular weight, there can be no question as regards the influence of molecular weight upon the position of the bands. Ill 1 1 NH, 11, = ;: ^•^ Rt, III!!;; Cs 1 ! 1 1 I K 1 1 1 ; i i ^^ NH, 1 1 1 r i ! " Rt> 1 1 1 1 ! 1 Cs III ^ 1 1 1 ! 1 ^ NH, 1 1 1 1 1 ; '- Rb , , , , 1 ; ; Cs I 1 1 1 1 t ! ^ 1 1 1 1 ; ; 1 ■ ^ NH. , , !■;::;!;;: ■' Rb III;:; ; ; ! i Cs III;;:;;;;: '^ 1 1 1 ■ i ■ NH. 1 1 1 ; ; : : : ^-^ Rb 111;;; Cs III;;;; « 1 1 1 i ; ■ e. NH, t 1 I ; ; ; ; ; ; Rb 111;; Cs III;; ^^ III;: ^:(. NH^ III-;;; Rb 1 1 : : Cs 1 1 1 ; : • K I : : : ««- NH^ 1 : ; Rb 1 1 1 1 ; ! ■ Cs III;;;; ■ V \« '? . ^? . V . =? . -? . ^f . V Fig. 76. If we select a typical region in the spectrum and arrange the bands belonging to a single group as in table 39, we find a general drift of the various bands toward the violet as we pass from, salt to salt in the order K, XH4, Rb, Cs. 120 PLirORESCENCE OF THE URANYL SALTS. The same drift occurs quite systematically throughout the entire fluorescence and absorption spectrum, as may be seen from figure 76. In this chart such of the fluorescence and absorption series as are present in all four salts at +20^ are plotted on the frequency scale. The solid lines represent observed fluorescence bands; the dotted hues represent observed absorption bands; no hypothetical values are indicated. The order of the salts is the same as in table 39 and follows Table 39. Green polarization, 185°. White polarization, 185°. Cg. Dg. Eg. Ag. Bw. Cw. Ew. Aw. K NH4... Rb.... Cs.... 1903.3 1911.7 1912.0 1916.8 1842.3 1843.5 1845.7 1852.9 1940.8 1941.0 1950.3 1868.8 1868.1 1874.8 1878.5 1891.4 1894.1 1895.9 1899.8 1911.7 1919.7 1920.9 1924.6 1854.9 1870.9 1862.5 1879.2 1879.9 that given by A. E. Tutton in his Treatise on Crystalline Structure and Chemical Constitution (London, 1916). He found for both single and double salts of the alkaU metals that several of their optical properties, such as refractive index, etc., follow the order of the molecular weights, but that in the ammonium salts the NH4 radical often acts as if it were much heavier than the combined weights of its components would indicate, so that its position is quite close to rubidium and sometimes on the side toward caesium. It will be observed that there are several examples of this in figure 76, particularly in the case of the Cg series. SUMMARY. (1) The four double chlorides, uranyl ammonium chloride, uranyl potassium chloride, uranyl rubidium chloride, and uranyl caesium chlo- ride, crystalhze in the triclinic system. The crystals are pleochroic and their fluorescence spectra and absorption spectra are polarized. (2) The spectra differ from those of other m-anyl compounds thus far examined in that both in the fluorescence and absorption regions each band is resolved at +20° C. into a group of five bands forming homologous series of constant frequency interval, (3) The structure of the fluorescence spectrum is essentially the same in the different salts, the spacing of the bands of each group repeating itself in the successive groups, excepting in the reversing region, the appearance of which is modified by the overlapping of fluorescence and absorption. (4) Each of the five bands which constitute a group is a doublet, the two components of which are polarized at right angles to one another. (5) The frequency interval is the same or nearly the same for each series in a given salt. POLARIZED SPECTRA OF DOUBLE CHLORIDES. 121 (6) Variations in the average interval for the four salts are scarcely greater than the errors of observation, but there are indications of a very shght decrease of interval with increase of molecular weight, and this apphes alike to fluorescence and absorption series. (7) The position, in the spectrum, of a given band varies slightly but systematically with the molecular weight of the salt. The order of diminishing wave-lengths is K, NH4, Rb, Cs; the shift from K to Cs being of the order of 5 A. u. This shift is in the same direction — from red toward violet — for all the homologous series and of the same size within the errors of observation. (8) Cooling to the temperature of Uquid air produces the usual narrowing of bands, apparent shifts of position, and apparent changes of interval, all of which changes are explained by the relative enhance- ment or diminutioii of components of the bands. VII. THE NITRATES AND PHOSPHATES; INFLUENCE OF WATER OF CRYSTALLIZATION AND OF CRYSTAL FORM. I. URANYL NITRATE AND EFFECT OF WATER OF CRYSTALLIZATION. The spectra of the different uranyl salts are so similar in their general characteristics that we can scarcely doubt that the nature of these spectra is chiefly determined by the radical UO2. Apparently the uranyl radical contains a group of electrons whose arrangement is such as to permit of vibrations that give this type of spectrum; and although UO2 is not stable in the chemical sense and must be com- bined with some acid in order to form a stable compound, yet the effect of the acid radical is merely to modify the constants of this vibrating system in the UO2 radical without changing the type of vibration. It is natural to expect that the addition of water of crystallization would produce a similar effect, and it is our intention to present in this section of Chapter VII the results of a study of the influence of water of crystallization upon the fluorescence and absorption spectrum in the case of uranyl nitrate. The nitrate is particularly suited for such an investigation because of the fact that several different hydrates are formed. The crystals grown from a water solution contain 6 molecules of water. In an acid solution crystals are formed with 3 molecules of water. In both cases crystals may be obtained which are large enough to permit of observations being made with a single crystal. By methods described later, small crystals containing only 2 molecules of water are readily obtained. It is a matter of some difficulty to push the dehy- dration further, but specimens have been prepared for us by Mr. D. T. Wilber which we have reason to believe are either anhydrous or formed of a mixture of the anhydrous salt and the monohydrate. The fluorescence of the nitrate, like that of the other uranyl salts, with the exception of the double chlorides, the resolution of the bands of whose spectra into groups of five at +20° has been described in Chapter VI, is unresolved at ordinary temperatures. Careful spectre- photometric measurements of what appear to be unresolved bands reveal, however, indications of overlapping components, as has already been shown in Chapter III. At the temperature of liquid air the resolution into narrow bands characteristic of the uranyl spectra in general takes place, and it is to these resolved spectra that the following discussion refers. In the case of the hexahydrate, wave-lengths were in most cases determined photographically, \isual observations, however, were also made, although these could not be extended throughout the whole spectrum. The agreement between measurements made by the two methods was surprisingly good. In the case of weak bands lying near 122 THE NITRATES AND PHOSPHATES. 123 to bands of great intensity the visual observations were found to be best. The results given for the fluorescence spectra of other hydrates and for the anhydrous salt are based upon visual observations exclu- sively. The Hexahydrate: U02(N03)2+6H20. The hexahydrate crystallizes in the rhombic system with the axial ratio a: b: c = 0.6837 : 1 : 0.6088. The crystals were grown in the form of plates by using a water solution whose depth was equal to the thick- ness of the plate desired. Single crystals as large as 15 mm. in diameter were obtained with relatively little difficulty. All of the results here discussed are based upon observations made with single crystals. In selecting the data to be used in taking a final average, each nega- tive was carefully studied and measurements that seemed for any reason doubtful were discarded. The eUmination of doubtful observa- tions was made without reference* to the agreement or lack of agree- ment between the different measurements, and was, in fact, completed before the measurements of the different negatives were compared. About 40 negatives were used, although the number for any one line was rarely more than 10. The errors of calibration of the spectrograph and spectrometer can hardly exceed 1 1. u., except perhaps in the extreme red end of the spectrum. The uncertainties due to the faintness of certain bands, to their finite width, and to photographic broadening are more difficult to estimate and undoubtedly differ greatly with the character of the band and its position in the spectrum. In the case of the sharper bands of moderate intensity we feel that the averages that are here tabulated are reliable within 1 A. u. In other words, the reciprocal wave-lengths are accurate to within about 0.02 per cent. For the faint or hazy bands the possible error is undoubtedly much greater. Of the 55 fluorescence bands observed, 46 can be arranged in 9 series, as tabulated below, the frequency interval being nearly constant in each series. Two of the remaining bands have the same interval, and apparently form part of a series whose other members were too weak to detect. The 7 bands that do not fall in any series arrangement are all extremely weak, and since in most cases they are recorded only once, their existence is subject to considerable doubt. Estimates are given in table 40 of the intensities of the different bands and of the reliability of the measurements. In some cases the series seem to extend into the region of absorption, and in such cases the absorption bands that seem to form part of the series are also given. The data for series B, D, E, and F, which are made up of the stronger bands and those of medium intensity, are undoubtedly the most reli- able. The values of the average interval between bands in these series are 86.0, 85.8, 85.9, and 86.1 respectively. In taking these averages, 124 FLUORESCENCE OF THE URANYL SALTS. the first band in the case of series D and E has been left out of con- sideration on account of its relative uncertainty. For the other series the interval, although less certain, has nearly the same value. It will be noticed that there is nothing to indicate any change in the interval as we pass from the longer to the shorter waves. Table 40. — Series in the fluorescence spectrum of uranyl nitrate hexahydraie lU02{NOz)2+6H20]. Inten- sity.i Relia- biUty.2 X 1 AX Inten- sity.i Relia- biUty.2 X 1 AX A B C V. d. V. d. d. m. V. d. d. m. ni. str. d. d. d. d. V. d. m. 2 1 2 4 I 4l 6 5 3 3 2 2 I 2 F 1760.1 1846.0 1930.1 F 2018.2 F 1689.5 1776.0 1861.1 1947.1 F 2034.5 A 2207.3 F 1699.0 1785.1 1869.0 1956.2 2041.5 85.9 84.1 88.1 F G m. d. m. m. m. d. m. m. m. str. V. d. d. d. V. d. d. 1? 3 4 5 6 4 4 4 4 3 1 3 4 2 1 F 1631.3 1718.4 1803.6 1889.6 1976.4 F 2061.9 A 2061.7 A 2148.7 A 2234.1 A 2321.0 A 2491.9 F 1810.4 1897.3 1983.1 A 2241.6 87.1 85.2 86.0 86.8 85.5 86.6 86.1 86.0 87.4 86.4X2 87.0 86.4 86.9 85.6X2 86.1 83.9 87.2 85.3 86.9 86.8 86.2X3 D E V. d. d. m. str. 8tr. str. d. V. d. d. m. str. str. str. str. m. 1 3 4 5 5 5 3 1 2 3 5 5 5 6 5 F 1534.9 1621.0 1706.8 1792.5 1877.8 1963.6 2050.0 F 1540.1 1629.2 1715.0 1800.0 1886.1 1972.6 F 2058.5 A 2058.6 H ■{ J V. d. V. d. d. m. m. V. d. V. d. d. m. d. m. d. d. 3 2 1 4 4 2 2 3 3 3 4 4 1 F 1649.7 1737.4 1822.0 1906.7 1993.4 1826.0 1911.7 F 1666.6 1751.8 1837.8 F 1923.1 A 2268.3 A 2440.2 86.1 85.8 85.7 85.3 85.8 86.4 87.7 84.6 84.7 86.7 85.7 89.1 85.8 85.0 86.1 86.4 86.0 86.3 86.0 85.3 86.3X4 86.0X2 ^ Estimated, Str., strong; m., mediTim; d., dim; v. d., very dim. * The most reliable results (as indicated by the number and consistency of the individual measurements, the appearance of the negatives, etc.) are marked 6; the least reliable by 1. ' The tmit in which 1/X is expressed is such that for X=5,000 A., v. 1/X is written 2,000, In a spectrum consisting of so many bands, the occasional repetition of any given interval between bands is to be expected, even if the bands are distributed at random. It is proper to inquire, therefore, whether this interval of about 86.0 really occurs more frequently than would be expected for a random distribution. Data bearing on this point are plotted in the upper curve of figure 77. In this curve horizontal dis- tances indicate the lengths of different possible intervals between THE NITRATES AND PHOSPHATES. 125 U02(N03)2 + 6H2 U02(N03)2 + 3H U02(N03)2-h2H2 bands, while ordinates give the number of times each interval occurred. The range of possible error in the location of each band is arbitrarily- assumed to be 2 units. Thus for a frequency interval 32 (abscissa) the ordinate is 10. This means that 10 pairs of bands were found for which the interval lay between 31 and 33. It is evident from the chart that certain intervals occur with much greater frequency than would be expected if the bands were distributed at random, and this is most conspicuously true of the interval 86. It will be noted that the curve also shows lesser maxima for several other frequency intervals: e. g., 8, 16, 70, 78, and 94. These intervals corre- spond to the spacing of the bands in the successive groups which make up the spectrum. On account of the fact that large, clear crystals could be obtained, the hexahydrate offered an especially favorable case for the study of the absorption spectrum. Obser- vations were made with a number of dif- ferent crystals ranging in thickness from a few tenths of a millimeter to 3 or 4 mm. The averages given in table 41 are in many cases based upon 15 or more independent measurements. In the case of the band at 2,148.7, for example, 17 measurements of wave-length were made, of which 3 were discarded because of the unsatisfactory character of the negatives. In the 14 measurements used in forming the average, the reciprocal wave-length ranged from 2,147.8 to 2,149.6, most of the values lying close to the average. In other cases the wave-length is much more uncertain. The extremely faint bandat 2,536.4, for example, was observed on only two negatives, while the dim, broad band at 2,720.3 was observed only once. The reliability of the recorded average has been estimated in each case and is indicated in the table. A study of the absorption spectrum shows that an interval of about 71 between bands is of relatively frequent occurrence. (See the lower curve, fig. 77.) In several instances definite series exist with this con- stant interval. The values of 1/X for the bands forming these series are given in table 41. The two series e and / begin with reversible bands. Thus, the first band of series e, at 1/X = 2,058.6, can not be distinguished in position from the last band, 1/X = 2,058.5, of the fluorescence series E, while the first band, 1 /X = 2,061 .9, of series/ is coincident with the band 2,061 .7 of U0.2CN03)2 + 6HO ABSORPTIQN Fia. 77. — Frequency of occurrence of different intervals between bands. AbscissEB show the in- tervals (1 division=10); ordi- nates show the number of timea the interval occurs (1 division «10). 126 FLUORESCENCE OF THE URANYL SALTS. Table 41. — Series in the absorption spectrum [UO^{NO,)2-\-6H20 of uranyl nitrate hexahydrate Series. Inten- sity. Relia- biUty. 1 X 1 AX Series. Inten- sity. Relia- bility. 1 X 1 AX c d' e f V. d. str. d. V. d. d. d. str. m. str. str. str. m. str. m. d. d. m. str. m. 2 4 5 4 2 4 3 3 3 3 6 5 4 4 4 4 4 5 4 2127.2 2200.4 2272.3 [2053.4] 2125.0 2196.8 2268.3 2340.4 2412.0 2484.1 2555.3 [2058.5] 2058.6 2131.2 2203.8 2277.8 [2061.9] 2061.7 2134.1 2207.3 73.2 71.9 71.6 71.8 71.5 72.1 71.6 72.1 71.2 72.6 72.6 74.0 72.4 73.2 k- •{ d 5'- m. str. str. m. m. d. d. d. str. m. str. str. str. str. d. d. m. 5 5 4 4 4 5 4 4 3 2 4 4 4 2 3 4 2148.7 2219.2 2290.2 2359.4 2430.1 2500.0 2164.0 2235.4 2321.0 2390.0 2464.3 2533.2 2552.8 2623.3 2695.0 2559.5 2630.5 70.5 71.0 69.2 70.7 69.9 ilA 69.6* 68.9 '70.5' 71.7 *7i!o" fluorescence series F. There is some indication that several other absorption series may be looked upon as associated with fluorescence series in the same way. Thus the series a may perhaps be associated with a very weak fluorescence series falling between D and E. Two bands of such a series were occasionally observed at 1,968.0, and 2,053.4 (interval 85.4). The interval between the line at 2,053.4 and the first line of series a is 70.6, which is almost exactly the average interval for the absorption series. Again, in the case of series h we might expect an absorption band to fall at 2,078.7, while series H might have a fluorescence band at nearly the same point, viz, 2,079.3. Neither band was observed; but it must be remembered that the detection of reversible bands is only possible when the conditions of excitation are suitable. Any trace of fluorescence tends to mask an absorption band, and vice versa. The scarcity of bands, either of fluorescence or absorp- tion, in the ''reversal region'' lying between 2,060 and 2,120 is perhaps due to this cause. There are other cases which suggest the same rela- tionship between fluorescence and absorption series, although less definitely. While there are thus strong reasons for believing that certain fluores- cence series are to be looked upon as associated, in the manner indi- cated above, with absorption series, yet there are several series in the fluorescence spectrum for which no related absorption series have been observed; and, on the other hand, there are several absorption series which do not appear to be related with the observed fluorescence. In THE NITRATES AND PHOSPHATES. 127 this respect the systematic relation of fluorescence and absorption is not so completely brought out as in the spectra of the chlorides. The observed intervals between bands are not so nearly constant in the case of the absorption series as in the fluorescence series. It seems to us probable that this is due to the greater uncertainty in the wave- length determinations; for on account of the lack of sharpness of the absorption bands and their greater width, as compared with fluorescence bands, the accuracy that is attainable in determining their location is considerably less than in the fluorescence spectrum. It will be noted also that the interval between bands is different for different series. For series e the average interval is 73.1 ; for series d! it is 71.7; for series /i, 70.3; while for the two pairs of lines in the ultra- violet, which have been designated as series /3 and 7, the interval is in one case 69.0 and in the other 68.9. This change in interval as we pass from one series to another, which is too great to be accounted for by experimental errors, appears of especial significance when it is remem- bered that in the fluorescence spectrum the interval is the same for all the series. One of the most puzzling points brought out by the detailed study of the observed spectra is the fact that a considerable number of the absorption bands are spaced with the interval corresponding to the fluorescence series, and in some cases appear to form a continuation of these series. In such cases the reciprocal wave-lengths for these bands are included in table 40, but are preceded by the letter A. Thus, there are four absorption bands, in addition to the reversible band, which apparently belong to series F. If it is assumed that they do form a part of this series, the average interval for the whole series comes out exactly the same as for the fluorescence bands alone. Series J appears to include two absorption bands, and series B, E, and G each show one band. On the whole, however, we are inclined to look upon these cases as the result of accidental coincidences and to believe that the fluorescence series do not extend into the absorption region beyond the reversible band. The interval of about 70, which appears to be characteristic of the absorption spectrum, is also found in the fluorescence spectrum. Thus the bands of series C are displaced from those of series E by intervals ranging from 69.0 to 70.2. A sinailar relation appears to exist between series J and series H, the average displacement being 70.1. Between series B and A the average shift is 70.6, between series F and H, 68.1, between series D and B, 69.1. With the exception of series I the series thus seem to be grouped in pairs, the interval between pairs being in the neighborhood of 70. It seems not unlikely that a companion series exists for I also, since faint bands were occasionally observed at 1,826.0 and 1,911.7 (interval 85.7) which are displaced by the intervals 71.3 and 71.4 from the corresponding bands of series I. 128 FLUOEESCENCE OP THE URANYL SALTS. The Trihydrate: U02(N03)2+3H20. The trihydrate crystallizes in the triclinic system with the axial ratios a :b : c = 1.2542 : 1 : 0.70053.^ The crystals were grown by evaporating the nitric-acid solution in a desiccator over caustic potash and sulphuric acid or over calcium chloride. To obtain large crystal plates the bottom of a dish 6 cm. in diameter was covered with the solution to a depth of about 2 mm. and the solution was "seeded" near the center. A cover was then placed over the solution with a small opening at the center, so that evaporation took place directly over the crystal. The trihydrate was also obtained in the form of a fine powder by efflorescence of the hexahydrate crystals in dry air. Although most of the observations recorded below were made with single crystals, a few measurements were made, with concordant results, on the powder. In the fluorescence spectrum visual observations only were made. In the absorption spectrum the measurements were in most cases photographic. 6 H^O J_ J A ' DC D E F q" H i J A 3 H,0 ■ I I I L 'A B C E J L_L a HO I ■ I I M ""a BCD ANHYDROUS F G ni i J K L J I I I F 6 'H i J K LM I8<50 19*00 Fig. 78. — One group of bands fiom the fluorescence spectrum of each of the salts studied. The spectrum of the hexahydrate contains 7 such groups; that of the trihydrate, 6; the dihy- drate, 5; and the anhydrous salt, 3. The fluorescence spectrum of the trihydrate was found to consist of 63 bands, of which 55 fell into 12 constant-interval series of from 3 to 6 bands each. The intervals for the different series ranged from 86.5 to 87.5, the average being 86.8. The reciprocal wave-lengths are given in table 42. The numbers in parentheses refer to absorption bands which seem to fall into the fluorescence series. The relative intensities ^Wyrouboff. Sur quelques composes de Turanium, Bull. Soc. Frangaise Mineral, 32, 349-350. 1909. THE NITRATES AND PHOSPHATES. 129 of the bands are shown roughly in figure 78, where a typical group, I. e.j one band from each series, has been plotted for each of the salts studied. Table 42. — Series in the fluorescence spectrum ofuranyl nitrate trihydrate, U02{NOs)2-\-3H20. A. 1677.0, 1765.0, 1851.5, 1938.8, 2025.0, (2112.7, 2200.2). B. 1686.1, 1772.0, 1858.7, 1945.7, 2033.2, (2120.7). C. 1778.7, 1865.1, 1952.9, 2041.1, (2041.1). D. 1699.8, 1785.9, 1873.0, 1959.1, 2046.8, (2134.0, 2220.6, 2307.9). E. 1704.2, 1791.8, 1877.4, 1965.3, 2051.0. F. 1629.2, 1715.2, 1802.1, 1889.0, 1976.4, 2064.3? G. 1637.2, 1722.8, 1808.7, 1895.3, 1982.3, 2070.7. H. 1643.9, 1729.2, 1816.1, 1903.0, 1989.9, 2076.3, (2076.5, 2251.2). I. 1821.3, 1908.9, 1995.9, (2083.8). J. 1741.8, 1828.2, 1915.9, 2002.1, (2089.7). K. 1748.6, 1835.3, 1923.3, 2009.8. L. 1667.2, 1755.0, 1842.4, 1930.1, 2017.4, (2103.5. 2189.2, 2277.9). The second curve of figure 77, which shows the frequency of occur- rence of the different possible intervals between the bands of the fluores- cence spectrum, indicates a remarkably regular grouping of the bands. Besides the principal interval 87 which is characteristic of the series in this spectrum, a number of other intervals are almost equally promi- nent, e. g.j 7, 14, 30, 36, 43.5, 50, 80, 94. In the case of each of these intervals the frequency of occurrence is far above the average. These intervals, of course, correspond to the spacing of the bands in the groups. Thus the interval 43.5, which is just half the principal interval 87.0, occurs twice in each group, the bands of series K lying half-way between the bands of series F and the bands of series L half-way between those of series G. In each case the two series might be com- bined to form a single series with half the interval. Since, however, the bands of the combined series would be alternately strong and weak, it does not appear that such a combination is justified. In the absorption spectrum of the trihydrate, 48 bands were observed, several of which, however, were so faint and indistinct as to make their existence doubtful. As in the case of the hexahydrate, an interval between bands of a little more than 70 is of frequent occurrence, and 37 of the bands (including all that are strong and well defined) can be arranged in 9 constant-interval series. The interval does not appear to be the same for all of these series, however. In one case the interval is as high as 73.8, while in another case its value is 71.0. For most of the series the interval lies near 72.0. In many instances the absorption series start with reversed fluores- cence bands. Reversals are especially sharp and definite in the case of the final bands of series C, G, H, and I (1/X = 2,041.1, 2,070.7, 2,076.3, 2,104.9). In other cases, the absorption series begins at a point where a fluorescence band might be expected, but where none was actually observed. Thus, we should expect the final bands of series I, J, and K to lie at 2,082.7, 2,009.0, and2,096.6 respectively. These bandswerenot observed, probably because of the fact that the three series in question 130 FLUORESCENCE OP THE URANYL SALTS. are made up of very faint lines. But the first bands of the absorption series %, j, and k fall at 2,083.8, 2,089,7, and 2,095.6, and it would seem, therefore, that they might properly be looked upon as resulting from the reversal of the final bands of the corresponding fluorescence series, even though these bands escaped observation. The reciprocal wave-lengths for the principal absorption series are given in table 43. Each series is lettered in such a way as to indicate Table 43. — Series in the absorption spectrum ofuranyl nitrate trihydrate [U02{NOs)2-\-3H20]. h B = 2033.2 b =(2033.2) 2107.0 2180.8. c C= 2041.1 c= 2041.1 2112.7 2186.7. e C'=? c'= 2043.3 2116.0 2189.7 2261.0. r F'= 2058.5 /'= 2057.6 2129.0 2200.2 2271.7. a G = 2070.7 ff= 2071.7 2142.7 2213.2 2285.7 2357.9. h H = 2076.3 A =2076.5 2148.6 2220.6 2368.0. i I =(2082.7) 1=2083.8 2154.5 2226.9 2297!8 2368.0. J J =(2090.0) i= 2089.7 2162.0 2235.4 2307.9. k K ==(2096.6) k = 2095.6 2166.7. V L'=? V= 2103.0 2173.4 2245.7 2317.5 the fluorescence series with which it appears to be related, e. g,, the first band of series c is the reversal of the last band of series C. In each case also the reciprocal wave-length is given for the last band of the fluorescence series. Thenumbers in parentheses indicate bands that were not observed, but would be expected to occur with the indicated value of 1/X. It is to be observed that no absorption series were found which corre- sponded to the fluorescence series A, D, E, and L. On the other hand, no fluorescence series were observed to correspond with the absorption series c' and V, The Dihydbate: U02(N03)2+2H20. Although the dihydrate has been made by several observers, the crystalline form does not appear to have been studied. The crystals used in this investigation were in most cases made by heating a tube containing the crystallized trihydrate to a temperature somewhat above 100° C. and passing through it a current of air which had been run through a mixture of sulphuric acid and nitric acid and over phosphorus pentoxide. The melted trihydrate slowly evaporated and recrystallized as dihydrate without losing nitric acid. The tube was then sealed up to prevent the entrance of moisture. The dihydrate was also prepared synthetically by treating dry H2UO4 with the ^'monohydrate" nitric acid, the two being sealed in a glass tube and allowed to react. The crystals obtained were in each case small, so that the observations were made on a mass of crystals having no systematic arrangement of the axes. Seventy-four bands were observed in the fluorescence spectrum, 61 of which fell into 12 constant-interval series of from 4 to 6 bands each. The interval ranged from 87.6 in series F to 88.3 in series A, but in most cases lay near the average of all, viz, 88.1.^ The values of \/\ ^ For series C, which contained two bands only, the interval was 87.2. THE NITRATES AND PHOSPHATES. 131 are given in table 44, and a typical group, consisting of one band from each series, is shown in figure 78, Many of the bands that do not fall into these 12 series nevertheless seem to belong to similar series of which only 2 or 3 bands could be detected. Two of these suspected series have been included in figure 78, where they are indicated by dotted lines. Table 44. — Series in ; the fluorescence spectrum of ura'> nyl nitrate c lihydrate [ UO^iNOz] A. 1676.7 1765.9 1853.0 1942,7 2030.7 B. 1684.9 1773.7 1861.6 1951.0 2037.9 C. 1780.6 1867.8 D. 1609.0 1695.5 1783.5 1871.8 1959.7 2047.4 E. 1618.9 1705.9 1794.4 1880.3 1967.5 F. 1625.3 1714.3 1801.5 1889.5 1977.1 2063.1 G. 1632.7 1721.2 1808.5 1896.8 1985.2 2072.5 (2072.7) H. 1637.7 1725.3 1814.6 1900.8 1989.7 2077.3 (2078.2) I. 1641.5 1729.8 1905.1 1993.8 J. 1645.3 1734.6 1821.2 1909.0 1998.4 K, 1830.6 1918.3. L. 1658.4 1745.5 1834.9 1923.8 2009.3 (2095.9) M. 1751.0 1838.2 1926.8 2015.6 (2102.7 2191,3 2278.0) The absorption bands were located by observing the spectrum of Ught from a coi^tinuous source after diffuse reflection from a mass of small crystals. Although the bands observed in this way are sur- prisingly sharp, the method is not so satisfactory as that in which the light is analyzed after direct transmission through a single crystal. It is doubtful whether the reflection method gives as great accuracy in the location of the bands, and many of the weaker bands, which would have been easily detected if large crystals had been available, were probably not observed at all. For this reason, perhaps, the absorption spectrum of the dihydrate shows only four well-defined series. The reciprocal wave-lengths for these series are given in table 45. The interval between bands is 70.0 for series w, 70.4 for series j, and 71.3 for series g and h. In each case the first band in the absorp- tion series occupies nearly the same position as the last band in one of the fluorescence series. Table 45. — Series in the ahsorption spectrum of uranyl nitrate dihydrate [U02{NOi)%-\-2H^O]. G= 2072.4 (;= 2072.7 2144.7 2215.3. H= 2077.3 A= 2078.2 2149.4 2218.9 2290.4 2362.9 J= (2085.7) ;■= (2086. 1) 2156.6 2227.2 2297.3. M=(2103.6) m= 2102.7 2172.7 2242.7. The Anhydrous Nitrate. Specimens of uranyl nitrate that were in all likelihood anhydrous, and which certainly contained less water than the dihydrate, were prepared by allowing nitric anhydride, N2O2, to react with uranic oxide. The nitric anhydride was distilled from a mixture of nitric acid and phosphorus pentoxide, while the uranic oxide was prepared by heating uranic acid, H2UO4. In preparing the oxide, the heating was not con- tinued so long as to completely drive off the water from the uranic 132 FLUORESCENCE OF THE URANYL SALTS. acid, since when this was done no reaction occurred between the oxide and the nitric anhydride. At temperatures above 30° C. the N2O2 reacts with the mixture of UOs and H2UO4 to form uranyl nitrate, presumably anhydrous, and a considerable amount of HNO3. To free the specimen from acid, the tube containing it was placed in a freezing mixture until the N2O5 was frozen, when the HNO3, which still remained a hquid, was poured off. This process was repeated several times. While the amount of water remaining after this treatment must have been extremely small, we cannot feel certain that all traces were removed, and it is possible, therefore, that the nitrate formed may have consisted in part of the monohydrate. No fluorescence bands belonging to the other hydrates could be observed. The method of preparation was varied by changing the temperature at which the reaction was allowed to occur, and by heating the salt, after it had been formed, to different temperatures and for different periods. The fluorescence spectra of the different preparations differed widely. In one case the spectrum was found to consist of 3 narrow bands only, but in all other cases bands were observed which remained broad (about 100 A. u.) even at the temperature of liquid air. These bands were spaced with a constant-frequency interval of approximately 88 to 89. It seems probable that these broad bands were due to the solution of the anhydrous salt in nitric acid. To test this point a specimen was prepared under conditions which made certain the presence of a considerable excess of nitric acid. The fluorescence spectrum contained two series of broad bands, the central band of the stronger series lying at 1/X = 1,939.0 and that of a weaker series at 1/X = 1,920.0. The specimen was then gently heated and the nitric acid driven off was condensed in a connecting- tube. After this process had continued for a short time, several series of narrow bands or lines appeared in the fluorescence spectrum (at —186° C). As this pro- cedure was repeated, the line spectrum became more prominent and the broad bands fainter. In one instance the specimen was heated nearly to decomposition — ^in fact, part of the salt was undoubtedly decom- posed — and in this case the very faint fluorescence spectrum consisted of three narrow bands only. The same three bands were observed in the case of most of the specimens that were prepared in the attempt to remove the water of crystallization, although in other cases they were accompanied by other lines or broad bands. It seems probable that these three bands constitute the brightest part of the fluorescence spectrum of the anhydrous salt, and that the additional lines and bands that were observed in some specimens are due to traces of the mono- hydrate or to a solution of the nitrate in HNO3. The three bands formed a series with the interval 88.5, the central band lying at 1/X = 1,902,0. THE NITRATES AND PHOSPHATES. 133 Summary of Section I. (1) In the case of each of the nitrates, the fluorescence spectrum is made up of series in which the intervals between bands are constant and the same for all of the series. The interval increases slightly, but unmistakably, as the amount of water of crystallization decreases. For the hexahydrate the interval is 86.0, for the trihydrate 86.8, for the dihydrate 88.1, and for the anhydrous salt 88.5. (2) Numerous constant-interval series occur in the absorption spec- trum, the interval being approximately 71. But the interval does not appear to be the same for different series, even when these occur in the spectrum of the same salt. No systematic variation with the amount of water of crystallization could be detected. (3) Nearly all of the series in the absorption spectrum have their origin in the ''reversing region," the first member of the absorption series being in coincidence with the last member of a fluorescence series and constituting a ''reversible'' band. (4) There is some slight resemblance between the different hydrates as regards the grouping of the bands (see fig. 77). In each case, for example, a certain short interval appears with a frequency consider- ably above the average. In the case of the hexahydrate and the dihydrate this interval is about 8; in the case of the trihydrate is almost exactly 7. The interval 14, in the case of the trihydrate, and 16, in the case of the other two salts, is also of unusually frequent occurrence. (5) It is clear from inspection of figure 78, in which a characteristic group of bands is shown for each of the hydrates studied, that the different spectra are not in the least similar in their general appear- ance. It might at first appear that the three hydrates have one series in common, viz, that designated as series F. But while the central band of this series does occupy practically the same position in each of the three spectra, the fact that the interval between bands is different for the different salts causes the bands to fall more and more out of step as we proceed in either direction from the center of the spectrum. On the whole, the spectra of the different hydrates differ from one another fully as much as do the spectra of two different uranyl salts. This result is surprising, since it is customary to think of water of crystallization as rather loosely attached and therefore incapable of exerting a great influence upon properties which depend upon the internal structure of the molecule. Indeed, were the amount of water of crystallization the only difference between the forms of uranyl nitrate under consideration, we should look upon the attachment as more intimate than has generally been supposed. There is, however, another distinction, that of crystalline structure, and, as will appear from the subsequent sections of this chapter, the crystal form has a profound influence upon the character of the fluores- cence and absorption spectra of the uranyl salts. 134 FLUORESCENCE OP THE URANYL SALTS. II. THE DOUBLE NITRATES: INFLUENCE OF CRYSTAL FORM.i There is good reason to think that crystal form has an important bearing upon the structure and arrangement of fluorescence spectra. The polarized spectra of the four double uranyl chlorides, described in Chapter VI, are almost identical in arrangement and in the absolute position, relative intensity, and resolution of their bands. These sub- stances all crystallize in the triclinic system. Further evidence bearing upon this subject will be found in section iii of this chapter. Our object in the present section is to describe the fluorescence and absorption of four double uranyl nitrates and to throw further Ught on the r61e played by crystal structure. The two pairs of double nitrates studied are mono-ammonium uranyl nitrate, NH4UO2 (N03)3 ; di-ammoniumuranylnitrate, (NH4)2U02(N03)4 2H2O; the mono-potassium uranyl nitrate, KUO2 (N03)3; and the di-potassium uranyl nitrate, K2U02(N03)4. The crystallographic features of these four compounds may be briefly specified as follows : (1) The mono-ammonium salt, which crystallizes from a solution of the two component salts in concentrated nitric acid, was described by Meyer and WendeP and crystallographically by Steinmetz.^ The crystals are of the trigonal system, with an axial ratio of a : c = 1 : 1.0027. (2) The di-ammonium salt crystallizes from a slightly acid water solu- tion of the two salts in which the ammonium nitrate is in excess of that required for the mono-ammonium salt. This salt was at fitTst thought to be the a modification of ammonium uranyl nitrate made by Rim- bach^ and measured by Sachs,^ but an examination of the spectrum of the a modification so-called proved that it was simply lu-anyl nitrate hexahydrate. The crystals analyzed by Rimbach were probably the mono-ammonium form, as this sometimes forms in the same solution. The crystals of the di-ammonium salt belong to the monoclinic system. (3) The mono-potassium salt crystallizes from nitric-acid solution in the rhombic system, as described by Steinmetz, with axial ratio a :b :c = 0.8541 : 1 : 0.6792. (4) The di-potassium salt crystallizes with reluctance; but when seeded from an acid aqueous solution, it forms in beautiful, fluorescent crystals of the monoclinic system. The axial angle is j3 = 90°=*= and the axial ratio a :b :c — 0.6394 : 1 : 0.6190. The composition is different from that of the di-ammonium salt, since it lacks the water of crystallization. Both visual and photographic measurements of the spectra were taken, and, since they agreed well, were averaged together. When ^ Howes, H. L., and D. T. WUber: Physical Review (2), ix, p. 125 (1917). 2 Meyei and Wendel, Ber. d. d. Ch. Gea., vol. 36, 4055. 1903. ' Steinmetz, Groth'a Chem. Krya., ii, p. 150. * Rimbach, Ber. d. d. Ch. Ges., vol. 37. 472. 1904. ^ Sachs, Zeitschr. f. Krya., vol. 38, 497. 1904. THE NITRATES AND PHOSPHATES, 135 possible the absorption spectrum was obtained by transmitted light. The crystals from an acid solution were of a deeper green color than those from a water solution, which necessitated grinding to about 0.4 mm. thickness to make them sufficiently transparent. Since the immersion in liquid air spoiled a crystal, many crystals of each form had to be prepared. Since, as is usual with the uranyl salts, we have in these spectra series of constant-frequency intervals, the location of the bands (tables 46, 47, 48, 49) is indicated in frequency units. As elsewhere in this treatise, fluorescence series are denoted by capital letters (A, B, C, etc.) and the related absorption series by a, 6, c, etc., or where the relation is not obvious, by Greek letters. The four spectra are mapped in the usual manner in figure 79. '.6^ 1. FLUORESCENCE K5fi II ■ .III mill kllil jiLliLl ABSORPTION '.4 /A I' ■ Ml' .l.l Mil.Mlilhi.Ul..liM.Hiill.lll[. TWTFFF li|ll||l'l|l'i{l'|l''>PII|ltlllll ' ' ' I" il'ii ' ililiii il'll'i 'liln llU liilrn I liillih 1 liJiilili I lillHiii III 3^1600 ll|l|l| I 2000 2400 Fig. 79. — 1. Fluorescence and absorption spectra of mono-ammonium uranyl nitrate, NH4UO (N03)a. 2. Diammonium uranyl nitrate, (NH4)2U02(N03)i.2H20. 3. Mono-potaasium uranyl nitrate, KU02(N03)3. 4. Di-potassium uranyl nitrate, K2U02(NOa)4. In the spectra of these double nitrates, the relation of absorption to fluorescence is somewhat simpler than is the case with the uranyl nitrates described in section i of this chapter, but it is less systematic and complete than in the spectra of the double chlorides. Thus, in mono-ammonium and mono-potassium salts (table 46 and 48), there are not sufficient absorption series to match all the fluorescence series, but there are no absorption series that do not join. In the di-ammonium spectrum (table 47) there is a related absorption series for each fluores- cence series and three extra absorption series that are not obviously related to fluorescence. (Seeplatel,6.) In the spectrum of the di-potas- sium salt, fluorescence series B, D, G, I, and J have no corresponding absorption of measurable intensity, while there are two absorption series, V and g, apparently with a direct reversing hnkage with the fluorescence. As to the completeness of classification of bands, it can 136 FLUORESCENCE OF THE URANYL SALIS. be said that not a fluorescence or absorption band of any of the salts fails to fit into one of the constant-frequency series. Table 46. — Series in the fluorescence spectrum of mono-ammonium uranyl nitrate. l/\ Ad/X) l/X A(l/X) l/X A(l/X) B D ' 1797.6 1885.9 1972.1 2058.3 1629.4 1738.] 1805.9 1893.4 1984.1 1555.5 1645.1 1734.0 1821.0 1909.1 1996.9 2086.1 88.3 86.2 86.2 88.7 87.8 87.5 90.7 89.6 88.9 87.0 88.1 87.8 89.2 G I ]573.9 1663.0 1751.0 1838.2 1925.9 2013.8 2101.3 1670.0 1758.0 1845.7 1934.2 2022.3 2110.2 1852.5 1941.4 89.1 88.0 87.2 87.7 87.9 87.5 88.0 87.7 88.5 88.1 87.9 88.9 K L M 1859.1 1948.8 2035.8 1602.1 1692.7 1780.9 1869.0 1953.5 2041.0 1704.2 1790.4 1878.8 1964.9 2050.6 89.7 87.0 90.6 88.2 88.1 84.5? 87.5 86.2 88.4 86.1 85.7 Series in the absorption spectrum of mono-ammonium uranyl nitrate. l/X A(l/X) l/X A(l/X) l/X A(l/X) a' 2132.8 2207.1 2280.5 2356.3 2430.1 2502.5 74.3 73.4 75.8 73.8 72.4 2163.1 2237.3 2313.2 2386.9 2469.1 2621.9 2693.9 74.2 75.9 73.7 76.4X2 72.0 V 2111.0 2187.7 2412.5 2562.1 76.7 74.9 74.8 In section i it was shown that in the spectrum of the uranyl nitrate the intervals of the fluorescence series are the same for all series within the errors of observation ; in the case of the absorption series, however, the interval is not the same for all series. In the double nitrates we find an unmistakable variation in the fluorescence intervals as well as in the absorption. In the mono-ammonium nitrate the interval varies from 86.4 for series M to 89.0 for series B. In the di-ammonium nitrate the interval varies from 83.7 for G to 85.0 for E. In the mono-potas- sium nitrate spectrum the interval varies between 86.4 for G and 87.9 for J. In the di-potassium nitrate the interval varies between 86.2 for E and 87.6 for J. The variation of the interval in the absorption series is of the same order of magnitude; 6. gr., an extreme variation of 1.4 in the mono- ammonium, 3.5 in the di-ammonium, 4.3 in the mono-potassium, and 4.4 in the di-potassium nitrate. In this connection it was thought to be of interest to compare the ratios of related fluorescence and absorp- THE NITRATES AND PHOSPHATES. 137 tion intervals. In table 50 these ratios are given. The ratios are nearly constant for the mono-ammonium and mono-potassium uranyl nitrates, but differ in the case of the other two salts. Table 47. — Series in the fluorescence spectrum of di-ammonium uranyl nitrate. lA Ad/X) 1/X A(l/X) 1/X A(l/X) B C D 1773.6 1857.4 1941.8 2026.4 1695.6 1779.5 1864.4 1949.3 1786.4 1871.6 1955.2 2039.6 1628.8 1713.1 1796.3 1880.8 1965.2 2050.2 83.8 84.4 84.1 83.9 84.9 84.9 85.2 83.6 84.4 84.3 83.2 84.5 84.4 85.0 G I 1637.7 1722.5 1806.8 1891.1 1976.5 2061.9 1564.0 1650.0 1733.3 1816.8 1900.8 1985.3 2068.5 1572.9 1657.8 1741.6 1824.5 1908.5 1993.2 2076.9 84.8 84.3 84.3 85.4 85.4 86.0 83.3 83.5 84.0 84.5 83.2 84.9 83.8 82.9 84.0 84.7 83.7 J' L 1664.5 1748.7 1832.2 1918.2 2002.6 2084.6 1754.1 1838.0 1923.1 2008.0 1595.4 1681.0 1764.4 1847.9 1931.2 2015.9 84.2 83.5 86.0 86.3 82.1 83.9 85.1 84.9 85.6 83.4 83.6 83.3 84.7 Series in the absorption spectrum of di-ammonium uranyl nitrate. 1/X Ad/X) 1/X A(l/X) 1/X ACl/X) a c- d 2114.8 2185.4 2254.7 2531.0 2669.5 2178.1 2248.7 2321.3 2392.6 2124.5 2332.5 2401.3 2472.2 2344.9 2414.9 2484.7 2552.8 2621.2 70.6 69.3 69.1X4 69.3X2 70.6 72.6 71.3 69.3X3 68.8 70.9 70.0 69.8 68.1 68.4 Q i f 2131,0 2201.0 2268.8 2338.4 2408.3 2477.4 2545.6 ^ 2683.8 2140.8 2210.9 2279.1 2077.3 2148.2 2218.0 2291.0 2358.5 2429.2 2497.7 2567.0 2154.2 2224.3 2365.4 2436.1 2604.2 70.0 67.8 69.6 69.9 69.1 68.2 69.1X2 70.1 68.2 70.9 69.8 73.0 67.5 70.7 68.5 69.3 70.1 70.6X2 70.7 68.1 k l\ a 2092.9 2163.9 2233.2 2304.0 2373.8 2443.6 2511.1 2584.7 2657.5 2102.3 2173.9 2245.1 2316.8 2384.8 2454.2 2523.3 2592.7 2422 . 3 2490.9 2559.7 2628.1 2538.1 2602.2 71.0 69.3 70.8 69.8 69.8 67.5 73.6 72.8 71.6 71.2 71.7 69.4 69.1 69.4 68.6 68.8 68.4 68.1 138 FLUORESCENCE OF THE URANYL SALTS. Table 48. — Series in the fluorescence spectrum of mono-potassium uranyl nitraU. lA A(l/X) lA A(lA) lA AdA) \ D 1725.3 1813.8 1900.4 1988.1 1569.1 1655.9 1742.7 1830.2 1916.3 2003.8 2090.5 1754.1 1842.3 1928.5 2015.7 88.5 86.6 87.7 86.8 86.8. 87.5 86.1 87.5 86.7 88.2 86.2 87.2 G J 1589.8 1674.8 1762.0 1848.5 1934.9 2021.4 2107.8 1867.6 1955.1 2043.2 85.0 87.2 86.5 86.4 86.5 86.4 87.5 88.1 K I 1615.8 1700.4? 1790.2 1877.5 1964.5 2050.3 2136.3 1683.2 1769.4 1856.9 1943.7 2030.5 2118.2 84.6 89.8 87,3 87.0 85.8 86.0 86.2 87.5 86.8 86.8 87.7 Series in the absorption spectrum of mono-potassium uranyl nitrate. 1/X A(l/X ) 1/X A(lA) lA A(lA) d 2167.0 2238.3 2313.8 2386.1 71.3 75.5 72.3 '{ 2117.7 2189.6 71.9 h 2140.3 2213.2 2287.9? 72.9 74.7 1900 I 1. B A c i B A t c i 1 1 1 1 \ [ 2. A^9 c E 3 I L D A?C C E I 1 ] L 1 JM c 3. D I C D B 1 F 1 1 1 i I ( 4. D 1 \v, i Hi jKl II 1 ill C D B IeF 1 1 1 3 Hi jKl II t ill Fig. 80, — ^A single group from each of the four spectra. 1, Mono-potassium uranyl nitrate — trigonal. 2. Di-potassium uranyl nitrate — monoolinic. 3. Mono-ammonium uranyl nitrate — ^rhoqibic. 4. Di-anmionium uranyl nitrate — ^mono- clinic. The bands occupy their natxiral positions in the left-hand panel, but have their strongest bands in vertical alignment in the right-hand panel. THE NITRATES AND PHOSPHATES, 139 That the crystal system to which a salt belongs is an important factor in determining the position of the bands can be seen in figure 80. In the left-hand panel a single group is shown in its natural position; in the right-hand panel the strongest bands of each group are placed in the same vertical line, to show the resemblance in grouping. This similarity is probably due to the fact that all four belong to the same chemical family. If we compare this grouping with that of the uranyl nitrate spectra in section i we find little resemblance, hence the grouping is probably characteristic of the double uranyl-nitrate family. In the left-hand panel it will be seen that the second and fourth groups occupy almost identical positions, while the first and third occupy positions which differ from one another and from the second or fourth. As has previously been stated, the second and fourth groups belong Table 49. — Series of the fluorescence spectrum of dir-potassium uranyl nitrate. l/X Ad/X) lA AdA) l/X AdA) B D E 1775.2 1861.9 1948.9 2034.6 1621.3 1708.0 1793.7 1880.0 1966.9 2053.7 1631.6 1717.9 1802.7 1889.3 1975.8 2062.2 86.7 87.0 85.7 87.7 85.7 86.3 86.9 86.8 86.3 84.8 86.6 86.5 86.4 F G H 1723.8? 1808.6 1894.1 1980.3 2068.8 1554.0 1640.4 1727.6 1813.2 1899.0 1986.2 /1903.7 \ 1906. 8 /1989.7 \1993.8 r2075.8 \2080.7 84.8 85.5 86.2 88.5 86.4 87.2 85.6 85.8 87.2 86.5 86.6 I f J K L 1651.5 1737.6 1823.8 1911.2 1998.5 2085.5 1831.8 1919.5 2007.0 1663.3 1751.8 1837.8 1925 . 1 2011.9 1670.8 1757.8 1844.6 1931.4 2018.3 86.1 86.2 87.4 87.3 87.0 87.7 87.5 88.5 86.0 87.3 86.8 87.0 86.8 86.8 86.9 Series in the absorption spectrum of di-potassium uranyl nitrate. lA Ad/X) lA AdA) lA A(l/X) i 2196.6 2269.3 2210.9 2285.7 2369.4 2444,4 72.7 74.8 75.0 k 2152.6 2224.3 2169.2 2240.1 2310.9 2382.6 71.7 70.9 70.8 71.7 8 2105.9 2179.9 2253.4 2325.1 2396.9 2361.6 2437.8 2513.2 74.9 71.7 71.8 76.2 75.4 140 FLUORESCENCE OF THE URANYL SALTS. to the monoclinic crystal systems, the first to the trigonal and the third to the rhombic system. Since all four spectra vary slightly in their frequency intervals, the relative positions would change sUghtly if we compared homologous groups in the other end of the spectrum, but this gradual and slight shifting would not change the general condition, which indicates that the absolute position of a group is largely deter- mined by the crystal system. This is not entirely new, as the four triclinic crystals of the double uranyl chlorides exhibit spectra which are as nearly coincident as could be expected of salts which vary in molecular weight. Table 50. — Average intervals. Mono-ammonium uranyl nitrate. Fluorescence series. Absorption seiies Ratio of fluorescence to absorption . A 86.6 D 88.3 G 87.7 a 73.7 d 74.5 g 74.2 1.18 1,19 1.18 I 88.1 75.1 1.18 Di-ammonium uranyl nitrate. Fluorescence series. Absorption series . Ratio of fluorescence to absorp- tion A 84.4 B 84.4 C 84.3 D 84.5 E 85.0 G 83.7 I 83.9 J 83.8 K 84.8 a 69.2 b 71.6 c 69.8 d 68.9 e 68.7 Q 68.8 i 69.6 3 69.7 h 70.9 1.22 1.18 1.21 1.23 1.22 1.22 1.21 1.20 1.20 L 84.0 I 71.5 1.18 Mono-potassium iiranyl nitrate. Fluorescence series . Absorption series , Ratio of fluorescence to absorption . D 86.9 I 87.2 d 73.2 i 71.9 1.18 1.16 K 86.6 h 74.1 1.17 Di-potassi\m3 uranyl nitrate. Fluorescence series . Absorption series Ratio of fluorescence to absorption . D 86.6 E 86.2 F 87.2 H 86.5 K 86.9 d 72.7 74.8 / 76.0 h 71.7 k 71.3 1.19 1.16 1.16 1.21 1.22 L 86.9 I 74.9 1.16 THE NITKATES AND PHOSPHATES. 141 Again, in the case of the uranyl nitrate, the crystals of the hexa- hydrate are of the rhombic system, while those of the trihydrate and dihydrate are of the triclinic system. In spite of slight shifts due to changing molecular weight, the strong bands of the two spectra pro- duced by the crystals of the triclinic system agree fairly well, while the strong bands of the spectrum produced by the rhombic crystal reside in entirely different positions. There is one more bit of evidence which adds weight to the above view. The chemical formulae of the two potassium salts are more nearly alike than those of the two anmionium salts, since the di- ammonium salt has 2 molecules of water of crystallization, while the other salts have none, yet there is a greater difference between the first and second spectra than there is between the second and fourth spectra. Summary of Section IL (1) The spectra of the double uranyl nitrates resemble those of the previously studied uranyl salts in that the bands can be arranged in series having constant frequency intervals. (2) These intervals, while constant for any given series, are different for different series. (3) In the mono-ammonium uranyl nitrate and the mono-potassium uranyl nitrate the ratio of the interval of a fluorescence series to the interval of the absorption series which joins that fluorescence series is approximately a constant. (4) Although the grouping of the bands shows a strong family resemblance in the four spectra, yet the absolute position of a group is largely determined by the crystal system. III. RESOLUTION ON COOLING AND ITS DEPENDENCE ON CRYSTALLINE STRUCTURE. The crystal system of any uranyl compound is an important factor in determining the character of its fluorescence and absorption spectra, as we have endeavored to show in the foregoing section. There is equally good evidence that resolution is dependent on the existence of a crystalline condition. Table 51. — Bands of fluorescence in canary glass. '^ At 4-20° C. At -185° C. X 1 Ax 103. X 1/XX103. 5280 5180 1894 1931 5330 5140 1876 1946 1 R. C. Gibbs, Physical Review (1), vol. 30, p. 382. 142 FLUORESCENCE OF THE URANYL SALTS. Not all uranyl fluorescence spectra are well resolved on cooling. In the case of a piece of canary glass, for example, the rather unusually broad, vague doublet occurs at +20° (see table 51) . At the temperature of liquid air the doublet is partially resolved, but no narrow components appear. The solid solution of uranyl phosphate in microcosmic salt, the phosphorescence of which has already been described in Chapter IV, yields a narrowing of the bands on cooling and a shift, but no resolution. (See table 52.) Table 52. — Bands of uranyl phosphate in microcosmic salt.^ At +20° C. At -185° C. X 1/\X103. X 1/XX103. 5670 5421 5183 4970 1764 1845 1929 2012 5680 5430 5190 4980 1761 1842 1927 2008 ^ The bands are 160 a- u- in width. The inference that the failure to obtain resolution of the bands is due to the non-crystalline structure of the substance is confirmed by the observations described below. EXPERIMENTS ON THE SPECTRA OF SODIUM URANYL PHOSPHATES.^ Stokes,^ in an early paper on the ultra-violet spark spectra of the metals, described a fluorescent screen prepared by treating the ordi- nary uranyl phosphate with a solution of phosphoric acid and sodium or ammonium phosphate. While the uranyl phosphate is only feebly fluorescent, the double salts thus produced were very brilliant. To investigate the fluorescence spectra of these double phosphates, the following preparations were made: (1) A mixture of uranyl phosphate and sodium phosphate in the ratio of 4 molecular weights of HUO2PO4 . 3 JH2O to 1 molecular weight of HU02PO4. (2) A similar mixture in proportions 2 to 1. (3) A similar mixture in proportions 1 to 1. These three specimens, when cooled to —180° C. and excited by radiation from the carbon arc, yielded precisely similar and well- resolved spectra. (See flg. 81, i, ;§, and S.) In addition to the above, four further specimens were made by mixing increasing amounts of phosphoric acid with sodium uranyl phosphate, i. e.: (4) One molecule of phosphoric acid to 2 molecules of uranyl phosphate and 1 molecule of sodium phosphate, giving the composition H3NaU02(P04)2 This was a powder, similar to preparations 1, 2, and 3. * Howea and Wilber, Physical Review (2), vii, p. 394. 2 Stokes, Philos. Trans., 152, p. 599, 1862. 1916. THE NITRATES AND PHOSPHATES. 143 (5) One molecule of phosphoric acid to 1 molecule of uranyl phosphate and 2 molecules of sodium phosphate. When dried, this contained much free sodium phosphate. (6) Two moleciiles of phosphoric acid to 1 molecule of uranyl phosphate and 1 molecule of sodium phosphate. This specimen did not dry, but remained syrupy at room temperature and appeared to be vitreous at — 180°. (7) A solution of uranyl phosphate in a considerable excess of syrupy phosphoric acid. This gave a glass-like mass even at +20°. The fluorescence spectra of these 7 substances are plotted to the usual frequency scale in figure 81. Table 53 gives the location of the narrow bands, and approximately of the crests of the broad, unresolved groups; also the frequencies and frequency intervals. It will be seen that the spectra of 1, 2, and 3 consist of recurring groups of narrow bands and that homologous members of these groups Table 53.— Wave-lengths and freqiiencies of the line series of the flitorescence of the sodium uranyl phosphates. X lA Al/X X l/X AlA Series A 5640 5396 5171 1773.1 1853.1 1933.8 80.0 80.7 Series F .... 6015 5739 5484 5248 5034 1662.4 1742.6 1823.4 1905.3 1986.3 80.2 80.8 81.9 81.0 (very dim) Average (dim) Average 80.4 Series B 6153 5862 5599 5359 5136 1625.2 1705.8 1785.9 1865.9 1947.0 80.6 80.1 80.0 81.1 81.0 (Him) Average Series G 5460 5227 5014 1831.6 1913.0 1994.6 81.4 81.6 (dim) Average 80.5 81.5 Series C 6114 5827 5568 5327 1635.7 1716.0 1796.0 1877.3 80.3 80.0 81.3 Series H 5429 5199 4989 1841.8 1923.4 2004.6 81.6 81.2 (very dim) (very dim) 81.4 80.5 ♦Series B' 6169 5877 5607 5363 5136 1621.0 1701.5 1783.5 1864.6 1947.0 80.5 82.0 81.1 82.4 Series D 6075 5794 5538 5299 1646.0 1726.0 1805.6 1887.1 80.0 79.6 81.5 (medium) Average 1 (medium) Average. . . . 80.4 Series E 6350 6040 5760 5506 5270 5057 1574.9 1655.7 1736.2 1816.2 1897.4 1977.6 80.8 80.5 80.0 81.2 80.2 (very strong) Average 80.5 ♦ Series B' is found in spectrum No. 5 only. 144 FLUOKESCENCE OF THE URANYL SALTS. Table 53. — Wave-lengths and frequencies of the broad-band series of the fltiorescence of the sodium uranyl phosphates. X l/X Al/\ X l/X Al/X Spectrum No. 4 . . . . Average 6219 5909 5641 5390 5158 4948 1608.0 1692.6 1772.7 1855.3 1938.7 2021.0 83.6 80.1 82.6 83.4 82.3 Spectnun No. 6 Average 5932 5644 5388 5147 1685.8 1771.8 1856.0 1942.9 86.0 84.2 86.9 85.3 82.6 Spectrum No. 7 . . . . Average 5958 5661 5400 5157 4935 1678.4 1766.5 1851.9 1939.1 2026.4 88.1 85.4 87.2 87.3 Spectrum No. 5 Average 5927 5647 5398 5174 4956 1687.2 1770.9 1852.5 1932.7 2017.8 83.7 81.6 80.2 85.1 87.0 82.7 ill _Lii _lU II 1 I ll I I h nil _b Luul Hi I 111 II . I Wi 4. form the usual constant-interval series. The intervalj which is the same for all within the errors of observation, is the shortest yet observed in the study of the fluorescence of the uranyl salts. Position as well as the arrangement of the bands is identical, and it is highly probable that we have to do with the same crystalline fluorescent compounds in these three preparations. The broad bands of specimens 4 and 5 form series with a constant interval of 82.5 units. Evidently the increase in the proportion of phosphoric acid tends to suppress the strongest line series and merge the dimmer series into broad bands. With the increas- ing predominance of the broad bands, caused by the increasingly larger pro- portion of acid present, there is a si- multaneous increase in interval from 82.5 units to 85.1 units for specimen No. 6, and 87.0 units for specimen No. 7. Experiments similar to the above were made in which ammonium phos- phate was substituted for sodium phos- phate. The results were in all respects analogous to those above described. That in general the resolution of the uranyl spectra by cooling occurs only when the fluorescing substance is in crystalline form is further substantiated by numerous experiments on frozen solutions to be described in detail in Chapter X. JXn ./^i ./li.ii../:^.. 5. -ZV /\ M/) x^ AAA 'V\AAA »7l00 ipIoq -lilfifi- FiG. 81. THE NITRATES AND PHOSPHATES. 145 Summary of Section III. (1) Where uranyl compounds occur in solid solution, as in canary- glass, or in a bead of microcosmic salt, the banded fluorescence spec- trum with constant frequency intervals, as observed at +20° C, is not further resolved into groups of narrow, line-like bands by cooling to the temperature of liquid air. (2) Sodium uranyl phosphate or ammoniimi uranyl phosphate, when prepared in the form of crystalline powder, gives fluorescence spectra which are fully resolved at low temperatures. (3) In the presence of an excess of phosphoric acid, where the above compounds, or uranyl phosphate, form solid solutions of Adtreous structure, resolution does not occur on cooling, (4) There is reason to think that the dependence of resolution by cooling upon the existence of crystalhne structure apphes in general to the fluorescence of the uranyl salts. VIII. THE ACETATES. The uranyl acetates afford a broader field for investigation than the chlorides or nitrates, the spectra of which have been considered in previous chapters. In addition to two forms of the single acetate U02(C2H302)2, we have the double salts of all the alkali metals except caesium; the double salts of calcium, barium, strontium, magnesium, zinc, lead, silver, and manganese; the triple salt NaMg U02(C2H302)5. In the fluorescence spectra of the acetates, as in the case of all uranyl salts thus far studied, the broader bands observed at room temperature are resolved into groups when the substance is excited at the temperature of liquid air, and the constitution of these groups, which repeat themselves at regular intervals from the red to the region in the blue, where absorption begins to replace fluorescence, is very similar in the acetates to that of the groups in the spectra of the com- pounds already discussed. THE SINGLE ACETATE. Two distinct varieties of this salt were available for observation — the finely powdered anhydrous form, U02(C2H302)2, and the crystalline form,U02(C2H302)2.H20. The spectra of the two are very similar in appearance; each being characterized by two strong, well-defined series forming a set of doub- lets. They are easily distinguished, however, by the widely different location of the doublets. In the spectrum of the anhydrous variety these occur near the group centers of the alkaUne double salts, whereas in the crystalline form they fall nearly midway between these groups. The strong series of the crystalline salt, which we have denoted as E and F, frequently appear in greatly reduced intensity in the spectra of the double salts, due doubtless to the presence of traces of the single acetate. The strong doublets C and D of the anhydrous acetate, if they ever appear in the spectra of the double salts, would be more difficult to detect, as they would overlap bands in the groups of the latter. Wave-lengths and frequencies of these two forms of uranyl acetate are given in tables 54 and 55. Intensities are designated as very strong (vs), strong (s), medium (m), dim (d), very dim (vd), and very very dim (wd). Studies of a Single Group. Since the acetates, like the chlorides and nitrates discussed in pre- vious chapters, have spectra consisting of similar recurring groups, it is convenient and sufficient, in the study of the structure of the ensemble of the fluorescence, to consider a single group. For this 146 THE ACETATES. 147 Table 54. — Fluorescence hands in spectrum of uranyl acetate (anhydroits), —185°. Group. Series. M 1/mX103 Int. Group. Series. f* 1//1X103 Int. C" 0.6037 1656.4 m. A 0.5295 1888.5 d. 3 D .6011 1663.5 m. B .5273 1896.6 d. E .5975 1673.6 vd. C .5229 1912.5 d. F .5950 1680.6 vd. G' .5223 1914.4 8. * 6 D .5202 1922.4 S. A .5822 1717.5 d. E .5183 1929.4 vd. B .5799 1724.3 d. F .5161 1937.5 m. C .5749 1739.4 d. G .5117 1954.4 vd. 4' C D .5739 .5713 1742.6 1750.5 m. m. H .5088 1965.5 vd. E .5687 1758.5 vd. A .5062 1975.4 vd. F .5663 1765.7 vd. B .5041 1983.6 vd. 7' C' .5006 1997.4 d. A .5548 1802.6 d. D .4979 2008.6 s. B .5523 1810.5 d. E .4961 2015.6 vd. C .5481 1824.6 d. F .4940 2034.3 d. C .5469 1828.5 s. 5 D E F G H .5445 .5424 .5401 .5352 .5318 1836.4 1843.6 1851.5 1868.5 1880.5 a. d. d. vd. vd. Table 55. — Fliiorescence bands in spectrum of uranyl acetate [UOi{C2HzOt)2'\-2HiO], at -{-185'' C. Group. Series. M l/juX103 Int. Group. Series. M 1/mX10^ Int. '{ E 0.6158 1623.9 d. A 0.5166 1935.7 vd. F .6133 1630.5 m. B .5149 1942.3 d. C .5130 1949.2 d. E .5860 1706.6 m. C' .5122 1952.0 vd. 4 E' .5849 1709.8 m. D .5107 1958.0 vd. F .5825 1716.8 s. 7' El .5096 1962.3 d. E .5090 1964.6 m. B .5648 1770.5 vd. Ft .5075 1970.4 . 8. C' .5630 1776.1 vd. F .5067 1973.5 8. El .5583 1791.2 vd. F' .5059 1976.7 vd. E .5575 1793.6 m. G .5027 1989.1 vd. 6 F .5550 .5529 1801.9 1808.8 s. vd. H .5005 1998.2 vd. G .5504 1816.9 vd. A .4962 2015.3 d. H .5473 1827.0 vd. B .4930 2028.5 m. I .5442 1837.6 vd. C C' .4913 .4904 2035.5 2039.2 m. d. Ai .5416 1846.4 vd. 8- D .4892 2044.1 d. B .5385 1857.0 d. Fi .4863 2056.3 m. C .5367 ]863.3 d. F .4857 2058.7 m. C' .5357 1866.7 vd. F' .4848 2062.7 8. D .5342 1872.1 vd. G .4823 2073.2 vd. 6 El E' F F' F' G H .5329 .5322 .5305 .5300 .5289 .5258 .5231 1876.5 1878.9 1885.0 1886.7 1890.7 1901.9 1911.7 d. m. vd. 3. vd. vd. ■ vd. 148 FLUORESCENCE OF THE URANYL SALTS. Anhijdrous J L I 1 I I Cri^stalline purpose group 7, which is in the brightest part of the spectrum and is free from the complications due to the overlapping of fluorescence and absorption in the reversing region, is most favorable. In figure 82 the spectral region of this group is plotted for the anhydrous and crystalline forms of the single acetate to depict the remarkable dis- placements brought about by the presence of water of crystallization and the consequent modifi- cation of crystal structure. vr^^v^. acetate The effect is very similar, both as regards the direction of the shift of the groups and the amount of shift, to that already described in the case of the nitrates. (Compare fig. 78 in Chapter VII.) In both instances it is not the transfer of the groups toward the blue without change in their structure that occurs, but something much less obvious. In fact, it is not possible to identify any of the bands iii the spectra of the hydrated form with those belonging to the an- hydrous salt- To produce this change in the spectrum it is only necessary to add a drop of water to a small portion of the anhydrous powder and to compare the fluorescence of the dry and moistened substance when excited in the usual way at — 185°. Frequency Intervals of the Single Acetates. The frequencies and frequency intervals of the series occurring in the spectra of the two forms of the single acetates are given in tables 56 and 57, from which it will be seen that the two forms of the acetate Table 56. — Uranyl acetate {anhydrous). s c J L 1900 Fig. 82. Series. Group 3. Group 4. Group 5. Group 6. Group 7. Average interval. A 1717.5 1724.3 1739.4 1742.6 1750.5 1758.5 1765.7 1802.6 1810.5 1824.6 1828.5 1836.4 1843.6 1851.5 1868.5 1880.5 1888.5 1896.6 1912.5 1914.4 1922.4 1929.4 1937.5 1954.4 1965.5 1975.4 1983.6 1997.4 2000.6 2008.6 2015.6 2024.3 85.97 86.40 86.00 86.00 86.27 85.50 85.91 85.90 85.00 B Ci C 1656.4 1663.5 1673.6 1680.6 D E F G H General average . . 85.96 THE ACETATES. 149 appear to have the same interval. The difference between the weighted averages is much less than the uncertainties in the determination of the intervals of the dim bands of the weaker series. Table 57.~ Uranyl acetate (crystalline; 2H2O). Series. Group 3. Group 4. Group 5. Group 6. Group 7. Group 8. Average interval. A2 2015.3 Ai 1846.4 A 1935.7 1942.3 1949.2 1952.5 1958.0 1962.3 1964.6 1970.4 1973.5 1976.7 B 1770.5 1776.1 1857.0 1863.3 1866.7 1872.1 1876.5 1878.9 1885.0 1886.7 1890.7 2028.5 2035.5 2039.2 2044.1 2056^3 2058.8 2062.7 86.0 86.45 86.25 86.0 85.23 85.18 85.65 85.66 86.00 c C D El 1706.6 1709.8 1791.2 1793.6 E 1623.9 Fi F 1630.5 1716.8 1801.9 F' F" 1808.8 1816.9 1827.0 1837.6 G 1901.9 1911.7 1989.1 1998.2 2073.2 85.42 85.60 H I General average 85.72 THE DOUBLE ACETATES. The fluorescence spectra of these salts have as a rule lower frequency intervals than the two forms of single acetate. The average interval is below 85, as compared with 85.7 for U02(C2H302)22H20 and 85.9 for the anhydrous single acetate. The group structure is in general less symmetrical than that of the double chlorides or the double nitrates and precise comparisons are therefore more difficult. Corresponding groups in the majority of cases, however, occupy very nearly the same position in the spectrum, and the system of designating the various bands employed in the discussion of the chlorides and nitrates has been used. If we neglect some of the weaker outlying bands, the group structure of several of the double acetates is found to consist of 4 nearly equi- distant bands the wave-length of which is almost if not quite inde- pendent of the metal which enters into the composition of the double salt. The substances which most nearly conform to this type are the double acetates containing lithium, potassium, calcium, and strontium. Manganese uranyl acetate differs from these only in the absence of band B in some groups. (See fig. 83.) In the spectrum of the barium double acetate the groups are shifted bodily toward the red about 5 frequency units. In the spectra of the ammonium and rubidium salts band D is doubled. 150 FLUORESCENCE OF THE URANYL SALTS. The double acetates of sodium, magnesium, zinc, silver, and lead (fig. 84, a) axe made up of groups which, while they overlap, are by no means identical, either as to the location or arrangement of their bands. The spectra of these 5 salts agree, however, in this: They contain in each group 5 bands which correspond so closely with the bands B, C, D, E, and F of the double acetates depicted in figure 83 that by a bodily shift of the group as a whole they may be made to conform to the uniform arrangement, so far as those bands are concerned, quite as well as do the latter. This may be seen from figure 84 6, in which the dotted vertical lines indicate the positions of the bands in the uni- form type, while the group in each case has been shifted so as to register approximately. Doublx AecTATca OouskC AccTATca u C 3 ' ' r H 1 K 1 Ctt 1 1 Mn , 1 Sr 1 1 Na J_ ,11 II 1 1 Na II II in — ' C JD t If 1 1 Me 1 1 1 1 1 1 i 1 Zn 1 II 1 1 Zn 1 1 1 ' 1 1 1 1 1 1 As 1 1 1 A, 1 i 1 1 i 1 Pb ll 1 Pb 1 1 1 1 1 i] Fig. 83. Fig. 84a. Fig. 846. The distinction between the spectra under discussion and those previously considered, which were described as having group spectra conforming to an essentially uniform type both as to location of bands and group-structure, is twofold : (a) there is a shift of the groups as a whole; (b) there are additional series, varying in intensity, some of which are among the strongest in the spectra and which are charac- teristic of the individual salt. It should be reiterated in this connection that neither the bands B, C, D, E, and F, which, although sometimes uniformly shifted, are common to the spectra of the double acetates, nor the additional bands, are found in the spectra of the single acetates. The spectrum of neither the anhydrous acetate nor the crystalline form can be made to conform to the uniform type by a general shift. THE ACETATES. 151 A Possible Relation to the Metallic Spark Spectra. It appears from the foregoing that any metal capable of forming a double uranyl acetate modifies the constitution of the fluorescence spectrum both as to the composition of the groups and their location. Certain metals, such as lithium, potas- sium, calcium, manganese, and strontium, produce one and the same modification, irrespective of the metal which is present. Other metals shift the group slightly (e. g., barium) or vary slightly the relative distances between neighboring bands without otherwise changing the struc- ture of the group. The presence of still other metals, such as sodium, magnesium, zinc, silver, and lead, results in a considerable general shift and the introduction of new series into the spectrum characteristic of the partic- ular metal in question and existing only in the doublet salt of which it forms a part. Some of these groups are much more complex than the uniform type depicted in figure 83. The others are accompanied by strong bands or minor groups of bands lying outside the usual boundaries. One might imagine, to account for this type of spec- trum, that in addition to the metal in combination as a part of the double salt, there are in solution certain other radiators. If these are uncombined particles of the metal existing in a condition akin to the gaseous state, one might expect a type of radiation, under exci- tation, similar to that discovered by Wood^ in sodium- vapor; i. e., series of constant frequency made up of bands instead of lines because of damping. One member of each such series should coincide or nearly coincide with some Une in the arc or spark spectrum of the metal. Now, there are in fact various coincidences or approxi- mations thereto close enough to bring hues of the emission spectrum well within the brighter portion of one of the fluorescence bands in question. In the spec- trum of silver uranyl acetate, for example, there is a strong series which does not coincide with any series in the fluorescence spectra of the other acetates thus far observed. One member of this series coincides with the brightest visible line in the spark spectrum of silver (Haschek 0.54655 ^t; frequency number 1,829.6). Our reading of the corresponding band, made before we had any suspicion of the possible relation here suggested, was 1,829.8. The rather bright hne (0.51838) and the neigh- boring doublet (0.51729-0.51675) in the spark spectrum of magnesium correspond similarly to bands 1,928.9 and 1,934.4 of the fluorescence spectrum. Fig. 85. 1 R. W. Wood. Physical Review (2), xi, p. 76. 152 FLUOEESCENCE OF THE URANYL SALTS. In the spark spectnun of lead, of the 9 lines hsted by Haschek which lie in the fluorescence region, 7 are within one frequency unit of our readings of the corresponding bands; 4 of these are in practically perfect coincidence, the departures from the crests of the bands being only one or two tenths of a unit. Of the 25 spark lines of zinc within the fluorescence region, 15 are certainly not related to fluorescence in the manner here under con- sideration, 4 in somewhat doubtful coincidence, and 6 are in close approximation. Of these last, 5 are consecutive lines of the spark spectrum, all of which are in group 7 of our fluorescence system. The evidence of any significant relation based upon these coincidences is obviously far from conclusive. The matter is mentioned here solely in view of possible developments in the further study of the connection between fluorescence and temperature radiation. The search for possible coincidences in the case of sodium led to the discovery of a striking arrangement, which seems to be peculiar to that element. The doublets and triplets of the spark spectrum, while they do not form constant-frequency series, are so located that they could be excited to radiation of the type described by Wood, with a common interval equal to the fluorescence interval of the acetates; L e., about 85, the result would be a well-defined group spectrum of the type of the fluorescence spectrum of the uranyl salts. (See fig. 85.) There are, however, only two individual coincidences with bands of the sodium uranyl acetate. In the figure, the actual arc-lines of sodium are elongated. The shorter lines are derived from them by assuming constant-frequency series having the interval 85, as described above. Fluokescence Series in the Spectra of the Double Acetates. In tables 58 to 70 the fluorescence bands in the spectnun of the double salts are arranged in the order of their wave-length. In tables 71 to 83 the frequencies and average intervals of each series in the various salts are given. It will be seen by comparison with tables 56 and 57 that the average interval for the double acetates is less by more than one frequency unit than for the single acetates; also that the av- erage for the various double salts differ from the general average of all (table 84) by an amount no greater than the difference between the intervals of the various series present in the spectrum of a given salt. In brief, whatever real differences may exist are too small to be deter- mined from our data. THE ACETATES. Table 58. — Lithium uranyl acetate. 153 Group Group Group and M 1/mX103 Int. and M 1/fxXW Int. and M l/fiXlO^ Int. senes. senes. senes. A 0.6105 1638.0 d. C 0.5481 1824.5 m. ^C 0.5016 1993.7 m. Q G .6033 1657.4 vd. D .5451 1834.4 d. D .4987 2005.1 c ^)E .5963 1677.0 m. 5 E .6423 1844.0 s. 7 E .4967 2013.1 nas. H .5858 1707.0 vd. F .5399 1852.0 m. F .4948 2021 . 1 ma. G .5363 1864.7 vd. H .4889 2045.6 d. 'C .6740 1742.1 d. H .5336 1874.1 vd. D , .5711 1751.0 d. C .4808 2080.0 ms. 4 E .5680 1760.5 m. G .5238 1909.0 m. 8 D .4784 2080.6 s. F .5653 1769.0 d. D .5209 1919.8 a. F' .5751 2104.8 ms. H .5590 1789.0 vd. 6 E .5184 1928.9 8. F .5166 1935.8 m. G .5128 1950.2 vd. [h .5102 1960.2 d. Table 59. — Sodium uranyl acetate. ■if 0.6262 1597.0 d. 'Bi 0.5510 1815.0 vd. B 0.6027 1989.2 .6107 1637.5 B. B .5494 1820.2 vd. C .4998 2001.0 C .5468 1828.8 m. C' .4988 2004.9 Bi .6085 1643.5 m. C' .5452 1834.1 m. D .4972 2011.3 B .6068 1648.0 d. D .5432 1840.9 m. D' .4965 2014.2 C .6028 1659.0 d. D' .5422 1844.3 m. 7- E .4949 2020.8 0' .6007 1664.8 d. E .5404 1850.6 s. F ,4931 2028.0 D .5978 1672.8 d. 5 F .5383 1857.8 m. F' .4917 2033.6 3 E .5948 1681.1 8. F' .5371 1862.0 wd. G .4900 2040 . 9 F .5924 1688.0 m. Gi .5359 1866.0 vvd. G' .4891 2044.5 G .5898 1695.5 vd. G .5345 1871.0 vd. H J .4874 2051.5 G' .6875 1702.0 vvd. G' .5332 1875.5 vd. .4844 2064.3 H .5846 1710.5 d. H .5313 1882.0 m. J .5805 1722.5 m. I .5283 1893.0 vd. ^Bi .4821 2074.1 I' .5269 1897.8 vd. C .4796 2085.1 B .5782 1729.5 d. C' .4787 2089.2 B' .5764 1735.0 d. B .5249 1905.0 vd. 8- D .4769 2097.0 .5731 1744.9 m. B' .5242 1907.5 vvd. D' .4762 2100.0 C' .5717 1749.0 m. C .5221 1915.3 s. E .4734 2112.5 D .6693 1756.5 na. C' .5210 1919.3 m. G .4707 2124.5 D' .5687 1758.5 m. D .5191 1926.3 s. 4' E .5662 1766.1 ' s. 6< D' .5182 1929.9 m. F .6642 1772.5 m. E .5166 1935.6 vs. F' .5625 1777.8 wd. F .6148 1942.5 vs. Gi .5614 1781.4 wd. G .5115 1955.0 vd. G .6599 1785.9 vd. G' .5099 1961,0 vd. G' .5590 1789.0 vd. H .5083 1967.5 m. H .5565 1796.9 m. I .5048 1980.8 d. I .5529 1808.6 d. 154 FLUORESCENCE OF THE URANYL SALTS, Table 60. — Magnesium uranyl acetate. Group Group Group and A* 1/mX103 Int. and M l/iuX103 Int. and M l/jLtX103 Int. senes. senes. senes. 2 I 0.6147 1627.0 m. ^A 0.5537 1806.1 d. rA 0.5063 1975.1 wd. B .5507 1815.9 d. B .5036 1985.8 m. [A .6115 1635.4 m. C .5475 1826.5 m. C .5007 1997.1 m. B .6072 1647.0 vd. 5- D .5447 1836.0 vs. 7' D .4983 2006.8 zn. D .6002 1666.1 m. E .5423 1844.0 s. E .4966 2013.8 8. 3 E .5971 1674.8 d. F .5408 1879.0 vvd. F .4951 2019.6 Vd. G .5926 1687.6 vd. G .5382 1858.0 d. G .4929 2028.7 d. H .5897 1695.7 vd. H .5358 1866.5 d. H .4908 2037.4 d. I .5831 1715.0 m. A .5289 1890.8 vd. B .4828 2071.4 m. fA .5810 1721.2 d. B .5261 1900.9 d. C .4802 2082.6 m. B .5774 1731.8 vd. C .5230 1912.0 m. Di .4791 2087.1 m. D .5708 1751.8 u. 6' D .5206 1920.9 s. 8i E .4764 2099.0 m. 4 E .6684 1759.2 m. E .5184 1928.9 d. F .4754 2103.6 d. ' F .5672 1763.2 vd. F .5170 1934.4 vd. G .4733 2112.6 d. G .5640 1773.1 vd. G .5145 1943.8 d. H .4707 2124.5 d. H .5613 1781.6 d. H .5128 1950.1 d. [l .5556 1799.8 d. Table 61. —Ammonium uranyl acetate. A 0.6097 1640.2 m. 'B 0.5520 1811.5 d. [B 0.5048 1981.0 d. C .6044 1654.5 vd. c .5484 1823.5 m. c .5020 1992.0 m. 3 E .5970 1675.5 m. D' .5453 1834.0 m. D' .4994 2002.5 B. F .5945 1682.0 d. 5 E .5424 1843.5 s. 7 D" .4983 2007.0 S. I .5867 1704.5 F .5404 1850.5 m. E .4970 2012.0 o. G .5379 1859.2 vd. F .4953 2019.0 a. [B .5789 1727.5 vd. H .5360 1865.5 vd. G .4929 2029.0 d. C .5749 1739.3 d. I .5338 1873.5 d. I .4897 2042.0 d. D' .5709 1751.5 d. 4 E .5681 1760.3 d. B .5271 1897.0 vd. C .4816 2076.5 m. F .5661 1766.5 m. 0' .5242 1907.5 m. 8 D' .4787 2089.0 vs. H .5613 1781.5 d. D' .5212 1918.5 m. D" .4778 2093.0 vs. I .5589 1789.2 vd. D" .5200 1923.0 m. F .4756 2102.5 m. 6 E F G H I .5189 .5171 .5142 .5127 .5105 1927.0 1934.0 1944.8 1950.5 1958.9 s. m. vd. vd. d. Table 62. — Potassium uranyl acetate. 0.6009 .5975 .5958 .5936 .5858 .5738 .5701 .5677 .5652 .5614 .5582 1641.2 m. 'B 0.5512 1673.7 d. .5484 1673.5 m. D .5444 1684.7 d. 5 E .5420 1707.8 vd. F H .5399 .5330 1742.8 d. 1754.0 d. fB .5262 1761.5 m. C .5237 1769.1 d. D .5202 1781.3 vd. 6' E .5182 1791.3 d. F .5163 G .5138 G' .5121 [H .5101 1814.2 vd. fB 0.5041 1983.9 wd. 1823.4 m. C .5016 1993.8 a. 1837.0 m. D .4986 2005.7 a. 1845.0 s. 7' E .4967 2013.3 s. 1852.2 m. F .4948 2021.0 vs. 1876.0 d. G G' .4923 2031.3 2036.7 d. d. 1900.3 vd. H .4891 2044.7 d. 1909.3 m. 1922.3 m. C .4811 2078.9 B. 1929.6 ». ^ F .4781 2091.5 a. 1937.0 s. .4749 2105.9 8. 1946.3 vd. G .4724 2116.8 vd. 1952.7 vd. 1960.4 d. THE ACETATES. Table 63. — Calcium uranyl acetate. 155 Group Group Group and M 1/mX103 Int. and M 1/mX103 Int. and M 1/mX103 Int. series. series. series. C 0.6017 1662.0 d. fc 0.5483 1823.8 d. fC' 0.5017 1993.2 m. D .6987 1670.3 wd. D .5461 1834.5 na. D .4990 2004.0 m. 3 E .5968 1675.6 d. 5- E .5427 1842 . 6 s. E .4970 2012.1 a. F .5939 1683.8 vd. F .5403 1850.8 m. 7 F .4953 2019.0 8. H .5869 1703.9 vd. G" .5361 1865.3 d. G" .4918 2033.3 d. J .5821 1717.9 d. H .5339 1873.0 d. H .4895 2042.9 d. I .4852 2051.0 wd. C .6740 1742.2 d. C .5240 1908.4 m. D .5703 1753.5 d. D .5212 1918.6 m. fC .4875 2076.8 m. 4- E .5682 1759.9 m. E .5190 1926.8 s. 8 D .4790 2087.7 m. F .5659 1767.1 d. 6 F .5169 1934.6 m. ]F .4755 2103.0 s. G" .5612 1781.9 vd. G" .5130 1949 . 3 d. G .4729 2114.6 vd. H .6590 1788,9 vd. H .5107 1958.1 d. [K .5059 1976.7 wd. Table 64. — Manganese uranyl acetate. < 0.5718 1748.9 vd. \o 0.5240 1908.4 d. fc 0.4812 2078.1 m. .5692 1757.0 vd. D .5213 1918.4 m. 8-^D .4786 2089.4 m. 6 E .5187 1927.9 Cf. F .4751 2104.8 m. C .5491 1821.0 vd. F .5168 1935.0 S. D .5454 1833.5 vd. H .5108 1957.9 d. 5 E .5428 1842.3 m. F .5408 1849.1 m. C .5018 1992.8 m. H .5339 1873.0 vd. D .4994 2002 . 6 s. 7- E .4971 2011.7 s. F .4952 2019.4 s. G .4936 2025.9 vd. .H .4895 2042 . 9 vd. Table 65. — Zinc uranyl acetate. 2 E 0.6288 1590.3 d. fD 0.5460 1831.6 vd. fA 0.5050 1980.3 wd. D' .5449 1835.0 d. .5026 1989.7 m. A .6234 1640.0 vd. E .5430 1841.6 s. C' .5011 1995.8 m. C .6045 1654.3 vd. E' .5420 1845.0 m. D .4998 2000.9 d. W .5967 1675.8 m. 5> F .5411 1848.0 m. D' .4988 2004.9 s. 3 F .5942 1682.9 d. F' .5400 1851.9 in. E .4977 2009.1 s. G .5921 1689.0 vd. G .5382 1858.0 vd. 7i E' .4965 2014.0 m. H .6893 1697.0 wd. H .5362 1865.0 vd. F .4957 2017.2 s. 1 .5870 1703.6 vd. I .5342 1872.0 m. F' .4946 2021,9 s. J .5331 1875.9 d. G .4935 2026.3 vd. A .6800 1724.0 d. H .4918 2033.4 vd. B .5780 1730.1 vd. ^A .5275 1895.7 vd. I .4899 2041.4 d. C .5749 1739.3 d. C .5249 1905.0 d. J .4888 2045.9 d. E .5701 1754.0 d. C' .5238 1909.1 d.' 4 E' .6685 1759.0 8. D .5219 1916.1 d. 'A .4843 2064.8 vd. F .5661 1766.6 m. D' .5209 1919.9 m. C .4819 2075.1 m. G .5640 1773.1 vd. 6 E .5193 1925.6 a. C' .4809 2079.6 m. H .5624 1780.6 wd. E' .5182 1929.9 m. & D .4795 2085.5 m. 1 .5618 1788.0 d. F .5171 1934.0 m. D' .4786 2089.3 s. F' .5160 1938.0 m. E .4776 2094.0 m. A .6528 1809.0 d. G .5149 1942.0 wd. F .4758 2101.9 e. b B .6612 1814.1 vd. H .5131 1949.1 wd. Gi .4746 2107.0 d. C .5482 1824.0 m. I .5110 .5099 1956.9 1961.3 d. d. 156 FLUORESCENCE OF THE URANYL SALTS. Table 66. — Rubidium uranyl acetate. Group Group Group and M l//iXl08 Int. and M I/mXIO' Int. and M i/mxio» Int. senes. senes. senes. fB 0.6093 1641.2 vd. 'C 0.5489 1821.8 B. ofH 0.6129 1949.7 vd. 3^ 3W .5977 1673.1 8. C .6476 1826.2 vd. "ll .6109 1967.3 m. .5955 1679.3 d. D .5458 1832.3 m. H .5879 1701.0 wd. D' .6445 1836.5 d. C .6022 1991.2 B. 5 E .5428 1842.3 vs. D .4995 2002.0 a. 'A' .5811 1720.9 vd. F .6406 1850.0 s. D' .4984 2008.4 xn. C .5753 1738.2 m. G .6377 1859.8 vd. 7' E .4971 2011.7 B. c .5739 1742.6 wd. H .6360 1865.7 vd. F .4953 2019.0 B. D .5717 1749.2 d. I .5338 1873.4 s. G .4927 2029.6 wd. 4- D' .6703 1753.3 d. H .4916 2034.2 wd. E .5685 1759.0 va. fc .5241 1908.0 8. I .4893 2043.7 na. F .5662 1766.2 s. C .5229 1912.4 wd. G .5639 1773.4 wd. D .5214 1917.9 Q. C .4813 2077.7 t)t H .5618 1780.0 wd. 6 D' .5202 1922.0 m. D .4790 2087.7 D. I .6592 1788.3 m. E .5188 1927.5 vs. 8J D' .4780 2092.1 m. F .6170 1934.2 8. F .4761 2104.8 B. a .6554 1800.6 d. G .6144 1944.0 vd. G .4730 2114.2 va. .6625 1810.0 d. Table 67 — Strontium uranyl acetate 2 I 0.6112 1636.1 d. C 0.5483 1823.9 m. 'C 0.5017 1993.2 m. D .5453 1834.0 vd. D .4990 2004.0 zn. [E .5974 1674.0 m. E .6427 1842.6 m. 7' E .4969 2012.6 m. 3P .6951 1680.5 m. 5 F .6407 1849.4 m. F .4950 2020.4 B. .5872 1703.0 d. G .5363 1864.6 d. G .4916 2034.0 vd. I .5812 1720.7 d. H .5334 1874.9 d. H .4888 2045.9 vd. I .6300 1886.8 vd. 'C .5752 1738.6 d. C .4812 2078.2 a. D .6721 1747.9 d. ^C .5239 1908.6 m. 8 D .4789 2088.1 m. 4 E .6687 1758.6 m. D .6211 1919.0 d. F .4761 2105.0 s. F .5663 1765.8 m. E .6187 1928.0 B. G .5615 1780.8 vd. 6 F .5169 1934.5 s. H .6589 ,1789.1 vd. G .6131 1949.0 d. H .6104 1959.1 d. [l .5078 1969.4 d. ' Table 68. — Silver uranyl acetate. 0.6979 .5961 .5878 .6806 .6764 .5730 .6699 .6678 .6625 .5600 1672.6 1677.6 1701.2 1722.5 1736.0 1746.1 1754.7 1761 . 1 1777.7 1786.7 d. vd. vd. vd. d. d. m. d. vd. vd. 0.5500 .5465 .6437 .5417 .5379 .5345 .6250 .6227 .6198 .6183 .6141 .6113 1818.4 1829.7 1839.4 1846.0 1859.0 1871.0 1904.9 1913.2 1923.9 1931.0 1945.1 1955.7 d. m. a. m. vd. vd. d. m. vd. vd. 0.6027 .6000 .4979 .4960 .4930 .4902 .4824 .4796 .4759 1989.2 2000.2 2008.6 2016.1 2028.6 2040.0 2073.0 2085.1 2101.5 THE ACETATES. Table 69. — Barium uranyl acetate. 157 Group Group Group and /* l//iX103 Int. and -M I/mXIO^ Int. and M 1//XX103 Int. senes. senea. senes. '0 0.6055 1651.5 vd. 'C 0.5498 1818.8 d. fc 0.5025 1990.0 d. E .5993 1668.6 d. D .5471 1827.8 d. D .5000 2000.0 m. 3 F .5963 1677.0 d. 5 E .5441 1837.9 s. 7 E .4978 2008.8 m. G .5923 1688.3 vd. F .5415 1846.7 s. F .4959 2016.5 s. H .5888 1698.4 d. G .5387 1856.3 d. G .4933 2027.2 vd. H .5349 1869.5 d. H .4901 2024.4 d. C .5764 1734.9 vd. D .5733 1744.3 d. C .5250 1904.8 d. (C .4878 2075.5 m. 4 E .5700 1754.4 m. D .5226 1913.5 m. ^ F .4795 2085.5 B. F .5672 1763.0 m. 6 E .5201 1922.7 a. .4769 2101.3 m. G - .5638 1773.7 d. F .5180 1930.5 S. G .4731 2113.7 vd. H .5602 1785.1 d. G .5150 1941.7 vd. ■ .H .5116 1954.7 d. Table 70. — Lead uranyl acetate. 2 K 0.6202 1621.6 m. C 0.5489 1821.9 m. ^1l 0.5100 1960.8 vd. C' .5472 1827.5 m. .5089 1965.0 vd. B .6086 1643.1 vd. D .5459 1831.7 a. C .6048 1653.4 vd. E .5439 1838.7 s. B .5042 1983.5 m. D .6015 1662.5 m. E' .5423 1844.0 vd. Ci .5028 1989.0 vd. E .5991 1669.1 m. 5 F .5410 1848.3 vd. .5019 1992.4 a. 3 G .5938 1684.1 vd. G .5390 1855.1 d. c .5008 1997.0 a. I .5909 1692.4 d. H .5369 1862.4 m. D .4996 2001.8 fl. H' .5891 1697.5 vd. H' .5352 1868.5 d. 7- El .4983 2006.8 wd. I .5859 1706.8 m. I .5333 1875.1 vd. E .4976 2009.6 vst. L .5841 1712.1 d. L .5320 1879.6 vd. F .4948 2021 . 1 vd. : ' G .4936 2025.8 vd. B .5788 1727.7 d. [B .5269 1897.8 m. H .4916 2034.0 m. C .5751 1738.7 d. Ci .5252 1904.0 vd. J .4886 2046.7 vd. C' .5741 1741.8 d. c .5242 1907.6 s. D .5722 1747.7 a. C' .5230 1911.9 s. B .4837 2067.4 m. 4 E .5700 1754.3 S. D .5219 1916.2 vs. Ci .4824 2073.0 vd. G .5649 1770.1 d. 6< El .5201 1922.7 vd. C .4814 2077.4 a. H .5621 1778.9 m. E .5195 1925.0 vs. 8' D .4790 2087.5 a. H' .5611 1782.3 d. E' ,5186 1928.3 vd. E .4772 2095.7 a. I .5591 1788.7 vd. F .5172 1935.0 vd. E' .4766 2098.2 wd. L .5567 1796.3 d. G .5149 1942.1 d. F .4749 2105.9 d. H .5130 1949.1 m. G .4737 2111.0 vd. .5518 1812.1 m. H' .5119 1953.4 vd. .5500 1818.2 vd. 158 FLUORESCENCE OF THE URANYL SALTS. Table 71. — Lithium uranyl acetate. Series. Group 2. Group 3. Group 4. Group 5. Group 6. Group 7. Group 8. Average interval. A 1638.0 1657.4 c 1742.1 1751.0 1760.5 1769.0 1824.6 1834.4 1844.0 1852.0 1909.0 1919.8 1928.9 1935.8 1993.7 2005.1 2013,1 2021.1 2080.0 2090.5 2104.8 84.52 84.89 84.03 84.05 D .. . . E 1677.0 F F' G 1864.7 1874.1 1950.2 1960.2 85.50 84.65 H 1707.0 1789.0 2045.6 Gen. av . 84.50 Table 72. — Sodium uranyl acetate. Bi 1643.5 1648. C 1729.5 1735.0 1815.0 1820.2 2074.1 86.12 86.43 B 1905.0 1907.5 1915.3 1919.3 1926.3 1929.9 1935.6 1942.5 1989.2 B' c 1659.0 1664.8 1672.8 1744.9 1749.0 1756.5 1758.5 1766.1 1772.5 1777.8 1781.4 1785.9 1789.0 1796.9 1808.6 1828.8 1834.1 1840.9 1844.3 1850.6 1857.8 1862.0 1866.0 1871.0 1876.5 1882.0 1898.0 1897.8 2001.0 2004.9 2011.3 2014.2 2020.8 2028.0 2033.6 2085.1 2089.2 2097.0 2100.0 2112.5 85.22 84.88 84.84 85.37 84.76 84.90 85.27 85.25 84.50 85.17 85.20 85.36 C' D D' E 1597.0 1681.1 1688.0 F F' Gi 1695.5 1702.0 G 1965.0 1961.0 1967.5 1980.8 2040.9 2044.5 2051.5 2064.3 2124.5 G' H 1710.5 1722.5 I 1637.5 I' Gen. av 85.22 Table 73. — Magnesium uranyl acetate. A 1635.4 1647.0 1721.2 1731.8 1806.1 1815.9 1826.5 1890.8 1900.9 1912 1975.1 1985,8 1997.1 2071.4 2082.6 2087.1 '2699!6' 2103.6 2112.6 2124.6 84.93 84.88 85.37 B C C' D 1666.1 1674.8 1751.8 1759.2 1763.2 1773.1 1781.6 1799.8 1836.0 1844.0 1849.0 1858.0 1866.5 1920.9 1928.9" 1934.4 1943.8 1950.1 2006.8 2013.8 2019.6 2028.7 2037.4 85.18 84.84 85.10 85.00 86.76 86.40 E F G 1687.6 1695.7 1715.0 H I 1627.0 Gen. av 85.19 Table 74. — Ammonium uranyl acetate. A 1640.2 B 1727.5 1739.3 1751.5 1811.5 1823.5 1834.0 1897.0 1907.5 1918.5 1923.0 1927.0 1934.0 1944.8 1950.5 1958.9 1981.0 1992.0 2002.5 2007.0 2012.0 2019.0 2029.0 2076.5 2089.0 2093.0 *2i62!5* 84.50 84.40 84.38 85.00 84.13 84.10 84.90 84.50 84.38 C 1654.5 D D" E 1676.5 1682.0 1760.3 1766.5 1843.5 1850.5 1859.2 3865.6 1873.5 F G H 1781.5 1789.2 I 1704.5 2042.0 Gen. av . . . . 84.40 THE ACETATES. Table 75. — Potassium uranyl acetate. 159 Series. Group 3. Group 4. Group 6. Group 6. Group 7. Group 8. Average interval. A 1641.2 B 1814.2 1823.4 1837.0 1845.0 1852.2 1900.3 1909.3 1922.3 1929.6 1937.0 1946.3 1952.7 1960.4 1983.9 1993.8 2005.7 2013.3 2021.0 2031.3 2036.7 2044.7 2678!9 2091.5 2105.9 2116.8 84.83 84.56 84.63 84.95 84.24 85-25 85.13 84.26 c 1656.1 1742.8 1754.0 1761.5 1769.1 D E 1673.5 1684.7 F G G' 1781.3 1791.3 1845.0 1876.0 H 1707.8 Gen. av. 84.57 Table 76. — Calcium urany ' acetate. Series. Group 3. Group 4. Group 6. Group 6. Group 7. Group 8. Average interval. c 1742.2 1823.8 1908.4 1993.2 2076.8 83.87 C' D 1662.0 1670.3 1675.6 1683.8 1753.5 1759.9 1767.1 1834.5 1842.6 1850.8 1918.6 1926.8 1934.6 2004.0 2012.1 2019.0 2087.7 2l63!6 2114.6 83.91 84.12 83.85 ssiso 84.66 E F G' G" 1781.9 1788.9 1865.3 1873.0 1949.3 1958.1 2033.3 2042.9 2051.0 H 1703.9 I J 1717.9 Ki 1976.7 Gen. av . 83.88 Table 77.— 'Manganese urany] acetate. Series. Group 3. Group 4. Group 6, Group 6. Group 7. Group 8. Average interval. C 1821.0 1833.5 1842.3 1849.1 1908.4 1918.4 1927.9 1935.0 1992.8 2002.6 2011.7 2019.4 2025.9 2042.9 2078.1 2089.4 2i04!8 85.70 85.12 84.90 85.23 D 1748.9 1757.0 E F Gi H 1873.0 1957.9 84.95 Gen. av. . 85.19 160 FLUORESCENCE OF THE URANYL SALTS. Table 78. — Zinc uranyl acetate. Series. Group 2. Group 3. Group 4. Group 5. Group 6. Group 7. Group 8. Average interval. A 1640.0 1724.0 1730.1 1809.0 1814.1 1895.7 1980.3 2064.8 84.96 84.00 84.16 85.03 84.48 84.77 83.95 85.22 83.80 85.00 B c 1654.3 1905.0 1909.1 1916.1 1919.9 1925.6 1929.9 1934.0 1938.0 1989.7 1995.8 2000.9 2004.9 2009.1 2014.0 2017.2 2021.9 2075.1 2079.6 2085.6 2089.3 2094.0 2101.9 2107!o c 1739.6 1824.0 1831.6 1835.0 1841.6 1845.0 1848.0 1851.9 D D' E 1590.3 1675.8 1682.9 1754.0 1759.0 1766.6 E' F F' Gi G 1689.0 1697.0 1703.6 1773.1 1780.6 1788.0 1858.0 1865.0 1872.0 1875.9 1942.0 1949.1 1956.9 1961.3 2026.3 2033.4 2041.4 2045.9 84.33 84.20 84.45 86.00 H J Gen. av 84.61 Table 79, — Rubidium uranyl acetate. Series. Group 2. Group 3. Group 4. Group 5. Group 6. Group 7. Group 8. Average interval. A.. .. 1800.6 A' 1720.9 1738 ! 2 1742.5 1749.2 1753.3 1759.0 1766.2 1773.4 1780.0 B 1641.2 1810.0 1821.8 1826.2 1832.3 1836.5 1842.3 1850.0 1859.8 1865.7 84.40 84.88 84.45 84.62 84.70 84.65 85.10 85.20 84.73 G 1908.0 1912.4 1917.9 1922.0 1927.6 1934.2 1944.0 1949.7 1991.2 2077.7 C' D.;;.;:::;::: 2002.0 2006.4 2011.7 2019.0 2029.6 2034.2 2087.7 2092.1 '2i64!8* 2114.2 D' E 1673.1 1679.3 F G H H' 1701.0 I 'i788!3* 'i873!4 1967.3 2043.7 85.13 Gen. av 84,86 Table 80. — Strontium uranyl acetate. Series. Group 2. Group 3. Group 4. Group 5. Group 6. Group 7. Group 8. Average interval. C 1738.6 1749.9 1758.6 1765.8 1780.8 1823.9 1834.0 1842.6 1849.4 1864.6 1908.6 1919.0 1928.0 1934.5 1949.0 1993.2 2004.0 2012.5 2020.4 2034.0 2078.2 2088.1 "2i65!6' 84.90 85.05 84.63 84.90 84.40 D E 1674.0 1680.5 F G Hi 1703.0 H 1789.1 1874.9 1886.8 1959.1 1669.4 2045.9 85.60 83.32 I 1636.1 1720.7 Gen. av. 84.74 THE ACETATES. Table 81. — Silver uranyl acetate. 161 Series. Group 2. Group 3. Group 4. Group 5. Group 6. Group 7. Group 8. Average interval. A 1722.5 1735.0 1745.1 1754.7 c 1818.4 1829.7 1839.4 1904.9 1913.2 1923.9 1989.2 2002.2 2008.6 2073.0 2085.1 84.50 85.00 84.63 D E E' 1672.5 1677.5 F 1761.1 1846.0 1859.0 1931.0 1945.1 2016.1 2028.6 2101.5 84.80 84.80 G G' 1777.7 1785.7 H 1701.2 1871.0 1955.7 2040.0 84.70 Gen. av. . . . 84.74 Table 82. — Barium uranyl acetate Series. Group 2. Group 3. Group 4. Group 5, Group 6. Group 7. Group 8. Average interval- c 1651.5 1734.9 1744.3 1754.4 1763.0 1773.7 1785.1 1818.8 1827.8 1837.9 1846.7 1856.3 1869.5 1904.8 1913.5 1922.7 1930.5 1941.7 1954.7 1990.0 2000.0 2008.8 2016.5 2027.2 2040.4 2075.5 2085.5 2101.3 2113.7 84.80 85.30 85.05 84.86 85.08 85.50 D E 1668.6 1677.0 1688.3 1698.4 F G H Gen. av . . . . 85.08 Table 83. — Lead uranyl acetate. Series. Group 2. Group 3. Group 4. Group 5. Group 6. Group 7. Group 8. Average interval. B 1643.1 1727.7 1812.1 1818.2 1821.9 1827.5 1831.7 1897.8 1904.0 1907.6 1911.9 1916.2 1922.7 1925.0 1928.3 1935.0 1942.1' 1949.1 1953.4 1960.8 1983.5 1989.0 1992.4 1997.0 2001.8 2006.8 2009.6 2021 !l 2025.8 2034.0 2067.4 2073.0 2077.4 2087!5 "2695;7' 2098.2 2105.9 2111.0 84.86 84.93 84.75 85.07 85.00 84.10 85.32 84.30 84.40 85.38 85.40 85.30 85.02 Ci c 1653.4 1738.7 1741.8 1747.7 C' D 1662.5 El E 1669.1 1754.3 1838.7 1844.0 1848.3 1855.1 1862.4 1868.5 1875.1 E' F G 1684.1 1692.4 1697.5 1706.8 1770.1 1778.9 1782.3 1788.7 1796.3 H H' I 1621.6 2046.7 L Gen. av 85.12 162 FLUORESCENCE OP THE URANYL SALTS. Table 84. — Summary of average intervals of the dovhle acetates. Substance. Interval. Substance. Interval. Li(U02)(C2H802)3.3H20 Na(U02)(C2H802)s 84.50 85.22 85.19 84.40 84.57 83.88 85.19 Zn(U02)2(atH802)(..7H20 Rb(U02)(C2Hs02)s 84.61 84.86 84.74 84.74 85.08 85.12 Mg(U02)2(C2H802)6.7H20.... NH4(U02)(C2H802)8 Sr(U02)2(C2H802)6.6H20 Ag(U05)(C2H302)« K(U02) (C2H302)S BaCUOj) (C2H802)8 . 6H2O Pb(U02)(CaH802)4.4H,0 General average Ca(U02)2(C2H802)6.8H20 Mn(U02)(C2H302)4.6H20 84.76 Absorption Spectra of the Acetates. The fluorescence and absorption of the acetates are related to each other in a manner entirely similar to that already established in the case of the other uranyl compounds. The absorption bands occur in series of constant interval and this interval is much shorter than that of the fluorescence series. Fluores- cence and absorption overlap in the reversing region, with numerous coincidences and an interlocking of the fluorescence and absorption intervals. Reversals, both exact and of the well-known displaced type, are more frequent, perhaps, than in any family of uranyl salts as yet studied. A notable example occurs in the spectrum of lead uranyl acetate (see fig. 86). PR. Rrvt RSA J . 21 00 Fig. 86. The absorption spectra fall into two fairly well defined classes: (1) Double acetates of Li, NH4, Na, K, Ca, Zn, Rb, Sr., Ag., Ba. In this class the system of bands having a series interval of 70+ ends at about 2,180, where is located the head of the strongest series. Band E of the fluorescence series is usually missing in group 8 and is supplanted by a strong absorption band (1) located 85=t= frequency THE ACETATES. 163 units from the terminating absorption band mentioned above, des- ignated as e in accordance with convention used in previous papers. An excellent example of this type of change from fluorescence to absorption is afforded by the spectrum of barium uranyl acetate (fig, 87). Here an exact reversal of band Eg^ occurs and the strong absorption band 6, which takes the place of E in that group, is 85 frequency units from the first member of the strong e series which extends toward the ultra-violet with the usual absorption interval of 70 units. Displaced reversals F, G, and H also occur — an indica- tion of the probably complex structure of these bands. The corre- sponding absorption bands are likewise 85 units from the first members of the/', g'y and h' series of the absorption spectrum. There is almost as notable a resemblance between the absorption spectra of this class as between their fluorescence spectra. The resolu- tion is, however, not so good, and all the members of the various series are not so easily located. Almost without exception the bands which can be observed are definitely related, in the manner just described, to the fluorescence series. (2) The single acetates U02(C2H302)2 and U02(C2H302)2+2H20; double acetates of Mg, Mn, Pb. Here the absorption system (interval 70 =t ) distinctly overlaps the fluorescence system extending into the region of groups 8 and 7 beyond, without change of interval. flROUP 7 r I flROUP o QROUP 9 _ji5. IF -85. _70_ _70_ 20 00 21100 22100 Fig. 87. The various series of absorption bands located in our visual and photographic studies of the acetates are contained in tables 85 to 97 inclusive. Frequencies and average intervals are given for each salt, the series being designated as usual by small letters, which indicate their relation to fluorescence series denoted by the corresponding capital letter. The three examples of bands or series not thus related to visible fluorescence are indicated by means of the Greek letter 7. * This band Eg may appear either as fluorescence or as absorption according to the conditions of illumination, etc. It is commonly seen as fluorescence in the spectrum of the zinc uranyl acetate and as absorption in the spectra of other salts of this class. 164 FLUORESCENCE OF THE URANYL SALTS. Table SS.—LHUOiXdHzOih.SH^O. Series. Frequencies. Interval. c c' e €' /' /" 01 h A' 2234.5 2375.0 2446.5 2514.8 2584.8 2085.0 2097.8 2185.0 2256.6 2325.0 2396.2 2100.0 70.06 *76!46" 2026,6 2108.8 2110.8 2113.0 2117.8 2201.5 2623.8 2124.0 ' 70.39*' 2129.0 Weighted average 70.27 Table m.—NHAiJJO'diCiHzO'd^ Series. Frequencies. Interval. A 7 c e /' a h i' 2145.6 1996.5 2082.4 2163.4 2299.2 2511.6 2232.4 2374.0 2446.3 2096.2 2183.2 2253.8 2322.8 2394.2 2107 8 '69!62'* 71.73 70.12 2114.8 2119.8 2207.0 2278.8 2348.6 2422.8 2487.5 2047.8 2130.0 '76!i3 Weighted average 70.19 Table S7.—Na{U02){C2Hs02h. Series. Frequencies. Interval. b 6' c di d Cl e e' / /' h i 2229 .2 2375 3 2239.6 2311.1 2382.1 2454.0 2524.0 2591.8 2093.4 2395.8 2328.3 2472.8 2102.6 2190.1 2259.4 2332.1 2403.8 2542.4 2265.0 2336.4 2409.1 2475.9 21110 2478.9 69.95 70.46 70.30 2117.7 2277.4 2343.0 2415.3 2126.8 2211.4 2285.7 2350.7 2423.7 2569.1 2137.7 2217.8 2291.0 2358.5 2430.1 2499.4 2336.3 70.40 71.54 70.15 Weighted average 70.46 THE ACETATES, Table ^%.—Mg{U02)2.{C2HzOi)f,.7H20, Table m.—K{V02){C2Hz02), Table ^0.—Ca{UO2h{C2HzO2)^.8H2O, Table 9h—Mn{UOMC2ff 302)5. OH^O. 165 Series. Frequencies. Interval. Cl c c' d' e / ffi g 0' 2161.0 2168.3 2239.2 2172,8 2315.0 2092.8 70.90 71.10 2182.8 2102.2 2108.4 2266.0 2336.2 2115.3 2199.3 2269.0 70.20 69.70 Weighted average 70.60 Series. Frequencies. Interval. c c' y d e cl /i g g' h A' 2166.0 2234.0 2084.0 68.00 2000.0 2176 1 2011.5 2017.0 2100.0 2186.0 2256.3 2326.5 2025.0 2109.3 76^25 ' 2116.0 2203.0 2124.0 2211.6 2292.2 2349.0 2131.8 68.70 2061 8 Weighted average 69.18 Series. Frequencies. Interval. y d e / /i 01 g h' h i 2023.1 2092.5 2166.8 2173.0 2245.2 2312.7 2384.9 2180.1 2249,2 2317.5 2390.1 2104 8 71.85 70.63 70.00 2185.0 2254.3 2325.6 2395.8 2198.8 2340.3 2409.6 2032.1 2115.1 2203.6 2274.3 2415.5 2123.6 2278.4 2349.1 2126.3 2436 6 70.27 70.27 70.63 70.70 Weighted average 70.54 Series. Frequencies. Interval. c e cl 2425 2011.5 2094.0 2178.4 2236.2 2307.5 2311 3 71.62 2110.6 2181.2 2251.5 70.50 Weighted average 71.25 166 FLUORESCENCE OF THE URANYL SALTS. Table ^2.~Zn{TJOMC2H^02\.7H20, Series. Frequencies. Interval. c ;'•;::::: Q h V i 2372.5 2442.6 2096.4 2181.5 2251.7 2322.9 2392.5 2188.2 70.10 70.10 2263.0 2178.6 2343.6 2415.0 2205.6 2277.9 2350.7 2129.9 71.40 72.80 Weighted average 70.77 Table m.—Rh{U02){C2Hz02)z- Series. Frequencies. Interval. a &i c c' d' d" c ffl ff' »1 i 2059 3 . 2070.8 2158.9 2292.2 2299.4 2373.6 2440.8 2081.6 2093.8 2246.7 2317.0 2392 3 . 70.27 "76!35" 2186.7 2259.4 2326.1 2399.8 2027 2 71.03 211.'; 1 2041 7 . 2129.0 2209.5 2279.6 2350.2 2423,1 71.20 Weiehted averacre 70.78 Table M,—Sr{U02)2{C2H,02)s^6H20. Series. Frequencies of absorption bands at —185** (groups 8 to 12). Interval. ci c «i e €' c" h V i o•V7^ 2 ^V?^ fi ... 2390 ... 2094!5R 2i8i!4 225i!9 2322.5 2185.7 2254.1 O^QQ Q ... "76!65'" 68.40 2273.8 2412.7 2206.0 2347.8 2209.0 2279.1 2350.0 2420.0 2490.0 69.45 70.90 70.25 frPTiPTal TppiffKtfid averaze ..•• 70.11 Table 95.—AgU02{C2Hz02). Series. Frequencies. Interval. c e /' 9 h i 2231.4 2369.7 2180.0 2251.3 2321.2 2390.1 2107.0 2191.4 2116.0 2125.0 2133.0 2204.1 2276.3 2344.7 69.15 70.03 "76!39 " Weiehted averaee 70.01 THE ACETATES. Table QG.—Ba{UOMCAO^),i.6H20. 167 Series, Frequenciea. Interval. b c e /' g' gh 2294.1 2227.2 2370.8 2094.7 2180.1 2250.2 2322.5 2393.5 2105.3 2187.2 2260.0 2115.1 71.82 71.13 72.80 2204.1 2275.8 2419.0 2486.9 70.70 Weighted average 71.36 Table ^l.—PhiUO^W^HzO^W^O, Series. Frequencies. Interval. h 2065.0 Cl 2072.0 2142.5 2355.0 2426.5 2495.0 70.5 c'd 2365.0 2437.5 2506.0 71.5 e 2094.5 2236.5 2309.8 2380.0 2450.0 2521.0 2594.5 71.43 e' 2098.0 2168.5 2240.0 2313.5 71.83 e" 2101.5 2389.0 71.87 / 2105.5 2176:5 2248.4 2321.5 2392.5 71.75 g 2110.5 k 2119.5 V 2125.0 2194.5 2264.0 2332.5 71.00 k 2201.5 2272.0 2341.0 2413.5 2482.0 2552.5 71.00 V V^eighted average 71.21 Table 98. — General weighted averages of intervals of absorption series in the spectra of the acetates at —185° C. Substance. Interval. Substance. Interval. U02(C2H302)2 71.04 72.35 70.27 70.19 70.46 70.60 69.18 70.54 71.25 Zn(U02)2(C2H302)6.7H20 Rb U02(C2H302)3 70.77 70.78 70.11 70.01 71.36 71.21 UOaCCaHaOa)? 2H2O LiU02(C2H802)3.3H20 NHiU09fO>Ha09^« Sr(U02)2(C2H302)6 . 6H2O Ac 1X02(6211302)3 Na U02(C2H302)3 Ba(U02)2(C2H302)6.6H20 Pb(U02)(C2H302)4.4H20 Av. interval for all acetates . . Mg(U02).(C2H302)fl.7H20.... Ca(U02)2(C2H302)8.8H20 Mn(U02)2(C2H802)«.6H20. . . . 70.68 From the list of general, weighted averages of the intervals for the various salts (tables 97 and 98) it appears that the frequency interval of the single acetates is larger than the general average, corresponding in this respect with the larger interval of their fluorescence spectra, as has been previously noted. The determination of intervals is, however, somewhat less accurate than in the case of the fluorescence bands, and, as in that instance, no difference between various series, or various salts, can be considered as positively established. 168 FLUORESCENCE OF THE URANYL SALTS. Summary. (1) The spectra of the uranyl acetates consist of the usual equi- distant fluorescence bands. (2) When excitation occurs at the temperature of liquid air, these bands are resolved into groups the homologous members of which form series of constant-frequency intervals. (3) There are two single acetates — Si finely powdered, anhydrous variety and a crystalline form with 2 molecules of water of crystalliza- tion, whose spectra differ widely, particularly as to the groups of fluorescence bands. (4) Of the double acetates, those containing hthium, potassium, calcium, manganese, and strontium have spectra which may be regarded as essentially identical both as regards the location of the principal bands and the structure of the fluorescence groups. The only distinctions between their spectra are in the sharpness of reso- lution and relative brightness of the various components. (5) The spectrum of barium uranyl acetate differs from the above in that the groups are shifted, as a whole, about 5 frequency units toward the red. (6) In the spectra of the double acetates of ammonium and rubidiiun, band D in each group is doubled, but there is no shift of the groups. (7) The presence of sodium, magnesium, zinc, silver, and lead modifies the group structure by the addition of bands characteristic of the metal and causes slight relative displacements of the group system as a whole. Otherwise the spectra resemble those mentioned under (4). (8) The frequency interval for the fluorescence series of the double acetates is probably the same for all series and for all salts, the depar- tures of the general averages for the various salts being less than one frequency unit from the average for all, i. c, 84.76. The same is probably true of the absorption series, the general average for which is 70.68. (9) The frequency intervals, both in the fluorescence and absorp- tion spectra, are larger by more than one frequency unit for the single acetates than for the double acetates. IX. THE SULPHATES. Uranyl sulphate (UO2SO4.3H2O) and the double uranyl sulphates of the alkaline metals are among the most brilliant of known fluores- cent substances. Their spectra are characterized by an unusual com- plexity of narrow bands brought out by cooling to the temperature of liquid air. The group structure is by no means so obviously uniform as in the case of the compounds already considered, nor is there the marked similarity between the spectra of the double sulphates which has been noted in the discussion of the fluorescence and absorption of the chlorides, nitrates, and acetates. There are, however, certain characteristics common to all the sulplmtes thus far examined; L e.: (1) Fluorescence at —186° vanishes with the group 7 (frequency 2000 to 2070), which is the reversing region for this family of salts, and the eighth group lies entirely within the absorption region. 3UUPMATCS u 1 1 1 II 1 u 1 1 1 II 1 NH4. Ill 1 1 1 1 1 NH4 1 1 1 1 1 1 1 Na III 1 1 II Na II 1 1 II K III 1 111 1 1 K 1 1 1 1 III 1 t Rb 1 1 1 II M Rb 1 III II Ca III III 1 Ca 1 i M 1 1 Fig. 88. Fig. • (2) Absorption of the type having the usual 70 =^ frequency interval extends without change of interval into group 7. In discussing the acetates, what we have called the heads of the prominent absorption series lie in the region between 2040 and 2060 instead of at about 2170, as in the spectra of the acetates. 169 170 FLUORESCENCE OF THE URANYL SALTS. (3) The fluorescence groups are distinguished by a strong pair of bands, fairly dominant in all the spectra excepting that of the sodium salt. The series formed by the members of shorter wave-length of these pairs terminates toward the violet, where it meets the head of the corresponding absorption series mentioned above. (4) The location in the spectrum of the fluorescence groups in the spectnmi of the sulphates is not approximately the same for the different salts, as is the case with the corresponding double acetates. On the contrary, there is in general a shift toward the violet with increasing molecular weight, as may be seen from figure 88, in which group 5 of the 6 sulphates under consideration are depicted. This shift is larger than that observed in the double nitrates, but not quite so systematic. Table Q9.~Uranyl sulphate: UOlSOSH^O. Fluorescence at -185° C. Prepared by extracting an excess of uranium oxide (UaOg) with sulphuric acid and oxidiz- ing the solution to UO2 . SO4 by means of H2O2. This neutral solution was evaporated to crys- tallization. The crystals were needles, some being 1 by 2 by 5 mm. in size, apparently ortho- rhombic, with three good pinacoidal cleavages. The angle pf the optical axes is very ne^ly 90° and the double refraction is positive. Group Group and /* l/tiXW Int. and M 1/mX103 Int. series. series. i^ 0.6254 1599.0 d. A 0.5256 1902.6 d. .6223 1606.9 d. B .5237 1909.5 vd. B' .5228 1912.7 vd. B .6046 1654.0 vd. .5210 1919.4 d. C .6009 1664.2 vd. C' .5203 1921.9 d. D .5976 1673.4 d. 6 D .5182 1929.8 d. 3 E .5942 1682.9 vd. E .5157 1939.2 m. F .5914 1691.0 d. F .5135 1947.5 s. H .5857 1707.4 vd. F' .5123 1951.8 m. I .5827 1716.1 vd. H .5091 1964.4 m. I .5071 1972.0 d. 'B .5740 1742.2 vd. J .5054 1978.6 d. C .5716 1749.4 d. ■ D .5686 1758.8 d. ^A' .5027 1989.2 d. E .5657 1767.7 m. B .5014 1994.4 vd. 4 F .3630 1776.1 m. B' .5004 1998.5 m. G .5600 1785.7 vd. C .4990 2004.0 m. H .5574 1794.1 d. C' .4981 2007.6 vd. I .6551 1801.3 vd. C" .4978 2008.8 vd. J .5534 1807.0 vd. C'" .4974 2010.5 vd. D .4964 2014.3 vd. A .5506 1816.5 d. 7< D' .4955 2018.2 vd. B .5478 1825.6 vd. El .4941 2023.9 vd. C .5450 1834.9 d. E .4938 2025.3 m. C .5441 1837.9 d. E' .4933 2027.2 d. D .5423 1843.9 vd. Fi .4926 2030.0 vd. 5' E .5394 1853.8 m. F .4917 2033.9 s. F .5369 1862.4 s. F' .4912 2035.8 vd. F' .5357 1866.7 vd. F" .4905 2038.7 m. G .5346 1870.7 vd. H .4878 2049.9 m. H .5321 1879.2 d. J .4843 2064.8 vd. I .5301 1886.4 d. J .5280 1893.9 vd. THE SULPHATES. 171 If the above groups are aligned by bringing band F into vertical registration, as in figure 89, it will be seen that the apparent dissimi- larity in the composition of the group in the various salts is due rather to the occurrence of various weak bands than to the arrangement of the stronger bands, which, while not identical, approximates to identity almost as closely as in the acetates or the nitrates. As in previous diagrams (see the chapters dealing with the spectra of the chlorides, nitrates, and acetates), the vertical lines indicate the position of the crests of the bands and, qualitatively only, their relative intensities. They are estimated in making observations merely as very strong (vs), Table 100.- -Uranyl ammonium sulphate: {NH\)2 ■ UO2 . {804)2 ■ 2HiO. Fluorescence at —185° C. Prepared by crystallizing a solution of the two component salts in the proportions of the double salt. The composition has been determined by Rimbach (Ber. d. d. Chem. Ges., 37, 479 (1904) ; the crystallization by de la Provastaye (Ann. Chem. Phys. (3), 5, 51 (1842), who described it as being monoclinic. The preparation consisted of square and rounded plates of diameter from 0.025 to 0.050 mm. The needle-like crystals showed distinct pleochroism from colorless to yellow, the greatest absorption being in the direction of greatest index. Group Group and M l//iXl03 Int. and M 1/mX103 Int. series. series. fE 0.6214 1609.2 d. 'Hi 0.6300 1886.8 vd. 2 P .6185 1616.7 d. H .5295 1888.7 d. [h .6100 1639.3 d. 5 H' I .5290 .5280 1890.4 1839.3 d. vd. B .6007 1664.7 d. J .5265 1899.3 vd. C .6977 1673.1 d. 3- D .5941 1683.2 d. A .5249 1905.2 d. E .5911 1691.8 B. Bi .5230 1912.0 vd. F .5883 1699.7 B. B .5214 1917.8 d. .H .6809 1721.6 vd. Ci .5198 1923.8 m. C .5194 1926.4 m. ^A .5766 1737.6 vd. C .5190 1926.8 m. B .5717 1749.2 d. Di .5177 1931.6 d. C .6697 1755.3 d. D .5169 1934.6 d. C .5689 1757.8 d. 6- El .5152 1941.0 vd. D .6663 1765.8 m. E .5147 1943.0 8. 4- E .5632 1775.5 B. Fi .5128 1950.1 vd. F .5608 1783.3 S. F .5123 1961.9 VS. G .5579 1792.3 vd. F' .5116 1954.7 vd. Hi .5549 1802.1 vd. Gi .5106 1958.6 vd. H .5539 1805.5 m. H .5073 1971.2 vd. J .5508 1816.5 vd. H' .5066 1974.0 d. J .5038 1984.9 vd. A .6491 1821.2 d. Bi .5472 1827.5 vd. Ai .5025 1990.0 d. B .5456 1833.0 m. A .5015 1994.0 vd. Ci .6440 1838.2 m. B .4998 2000.9 d. C .5435 1840.1 m. C .4976 2010.0 m. C' .5430 1841.6 m. Dx .4960 2016.1 d. 5 Di .5415 1846.7 vd. 7- D .4965 2018.3 d. D .5406 1849.8 d. E .4933 2027.1 s. E .5380 1858.8 s. Fi .4916 2034.2 vd. Fi .6361 1865.3 vd. F .4912 2036.0 vs. F .5354 1867.6 s. F' .4905 2038.7 vd. F' .6346 1870.6 d. Gi .4893 2043.7 vd. Gi .5337 1873.7 vd. 172 FLUOEESCENCE OP THE URANYL SALTS. Strong (s), medium (m), dim (d), very dim (vd), and very very dim (wd) respectively. No attempt is made in the diagram to indicate the width of the bands. The spectrum of the single sulphate resembles those of the double sulphates much more nearly than is the case with the single and double salts of the other acids. Wave-lengths, frequencies, and relative intensities of the bands observed in the fluorescence spectra of uranyl sulphate and the double sulphates of ammonium, sodium, potassium, rubidium, and caesium are given in tables 99 to 104. Similar measurements of the bands in the absorption spectra are given in table 10. The determination of wave-lengths were made by the visual and photographic methods de- scribed in the foregoing chapters. Table 101. — Uranyl sodium sulphate: Na^. U02-{SOi)2-2H20. Flitorescence at —185° C. Prepared by crystallizing a solution containing the two component salts in the proportions of the double salt. (See O. de Coninck, Chem. Centralblatt, ix, I, 919, 1905.) The preparation consisted of crystalline grains about 0.5 mm. in diameter, with much mother liquor or deliques- cence. The crystals are apparently monoclinic, with positive double refraction. Group Group and M 1/mX103 Int. and M 1/mX103 Int. series. series. 'B 0.6296 1588.3 vd. D 0.5374 1860.8 d. C .6255 1598.7 vd. E .5355 1867.3 d. 2- E .6182 1617.6 d. F .5330 1876.0 m. F .6151 1625.8 m. 5 G .5311 1882,7 vd. G .6122 1633.5 vd. G' .5301 1886.4 vd. H .6093 1641.2 vd. H .5287 1891.4 d. H' .5278 1894.7 d. B .5976 1673.3 d. I .5260 1901.1 d. B' .6963 1677.0 vd. C .5945 1682.1 d. ^A .5226 1913.5 d. 3- D .5908 1692.6 d. B .5197 1924.2 vd. E .5880 1700.7 d. Ci .5181 1930.1 vd. F .5851 1709.1 8. c .6167 1935.4 m. G .5828 1715.9 d. Dt .5152 1941.0 d. H .5797 1725.0 d. D .5141 1946.1 vd. 6 E .5123 1951.8 d. "A .5729 1745.5 vd. Fi .6110 1956.9 vd. B .5698 1755.1 m. F .5101 1960.4 8. Ci .5677 1761.5 vd. Gi .5087 1965.8 vd. C .5665 1765.1 m. G .5077 1969.5 vd. Di .5642 1772.4 d. Hi .6061 1976.9 d. D .5629 1776.5 d. H .6050 1980.2 m. 4- E .5607 1784.1 d. I ,5038 1984.9 vd. F .5579 1792.3 m. G .5560 1798.6 vd. ^A .5006 1997.6 vd. H .5532 1807.7 d. Bi .4977 2009.2 vd. H' .5522 1810.9 vd. B .4965 2013.9 8. Ii .5509 1815.9 vd. Ci .4955 2018.2 m. I .5501 1817.9 vd. 7 C .4943 2022.9 m. E .4910 2036.6 m. A .5468 1828.8 d. F .4890 2045.0 d. B .5439 1838.4 vd. G .4873 2052.3 vd. 5- Ci .5418 1845.7 vi. H .4857 2058.9 d. C .5406 1849.9 m. [H' .4847 2063.1 vd. Di .5388 1856.0 d. THE SULPHATES. 173 Table 102. — Uranyl potassium sulphate: Ki.UOiiSO^i.^HiO. Fluorescence at —185° C. Prepared by crystallizing u solution of the two component salts in the proportions of the double salt. The composition has been determined by Rimbach (Ber. d. d. Chem. Ges., 37, 478 (1904). The crystals obtained in this laboratory were orthorhombic. The preparation con- sisted of 6-sided plates and rounded grains about 0.045 mm. in diameter, the plane of the optical axis being a (100) and h the acute bisectrix. Double refraction positive. Group Group and M 1/mX10» Int. and /* l//iXlO» Int. series. series. 0.6267 1695.7 vd. 'F' 0.5332 1875.5 d. D .6229 1605.9 vd. Gi .5324 1878.3 vd. 2 E .6188 1616.0 d. 5- G .6319 1879.9 vd. F .6164 1622.3 m. G' .5314 1881.8 vd. F' .6150 1626.9 vd. H .5295 1888.6 vd. ,G .6129 1631.6 vd. J .5276 1895.2 vd. 'B .6010 1663.9 vd. 'A' .5240 1908.4 vd. Ci .5981 1672.0 d. Bi .5235 1910.2 vd. C .5957 1678.8 d. B .5226 1913.4 d. Di .5941 1683.2 vd. Ci .5199 1923.4 vd. 3 D .5921 1688.0 vd. C .5189 1927.2 m. E .5886 1698.9 vd. Di .5173 1933.1 vd. F .5859 1706.7 m. D .5164 1936.5 d. F' .5851 1709.1 d. D' .5157 1939.1 vd. G .5830 1715.3 vd. 6 El .5144 1944.0 vd. H .5804 1723.0 vd. E .5137 1946.7 m. F .6115 1955.4 s. ^B .5718 1748.8 vd. F' .5107 1968.1 d. Ci .5697 1755.3 d. Gi .6097 1961.9 vd. C .5680 1760.7 d. G .5093 1963.4 m. Di .5662 1766.3 vd. G' .5088 1965.4 vd. D .5644 1771.9 d. H .6069 1972.6 d. 4- E .5616 1780.5 m. r .5054 1978.6 d. F .5589 1789.1 s. F' .5683 1791.3 d. A' .5023 1991.0 m. G .5563 1797.5 vd. Bi .5015 1994.0 vd. H .5539 1805.5 vd. B .5007 1997.3 d. I .5519 1811.9 vd. Ci .4988 2004.7 vd. C .4973 2010.9 s. A' .5481 1824.5 vd. Di .4959 2016.5 vd. B .5461 1831.0 d. 7^ D .4951 2019.7 m. Ci .5438 1838.9 d. El .4935 2026 . 5 m. C .5424 1843.8 m. E .4923 2031.3 d. 5 Di .5405 1850.1 vd. F .4906 2038.5 s. D .5391 1854.9 d. Gi .4894 2043.3 vd. El .5374 1860.8 vd. G .4889 2045.2 d. E .6366 1863.5 m. G' .4884 2047.5 vd. iF .5342 1871.8 vs. 174 FLUORESCENCE OP THE URANYL SALTS. Table 103. — Uranyl rubidium svlphaie: -K62. UO2. {SO^z.2H20. Fluorescence, at—185'' C. Prepared by crystallizing a solution containing the two component salts in the proportions of the double salt. The composition has been determined by Rimbach (Ber. d. d. Chem. Ges., 37, 479, 1904). The crystallization ia in every way like the potassium salt, although the solubil- ity is less and the crystals smaller. The preparation consisted of 6-sided plates about 0.02 by 0.04 mm. in size. Group Group and /* 1/mX108 Int. and /* 1/mX103 Int. series. series. 1 F 0.6485 1542.0 vd. F' Gi 0.5332 .5320 1875.8 1879.7 vd. vd. C .6269 1595.1 d. 5 G .5310 1883.2 vd. D .6225 1606.4 d. H .5292 1889.6 vd. 2e .6187 1616.4 vd. I .5276 1895.4 vd. [f .6157 1624.1 d. A' .5240 1908.4 vd. B .6004 1665.5 vd. B .5223 1914.5 d. C .5975 1673.7 d. Ci .5198 1923.8 d. C .5954 1699.5 d. C .6187 1927.8 m. 3 D .5919 1689.5 d. D .5163 1937.0 d. E .5887 1698.7 m. 6 E .5136 1946.9 m. F .5860. 1706.5 s. F .5115 1955.2 vs. F' .5105 1958.7 vd. 'B .5719 1748.5 vd. Gi .5096 1962.3 vd. Ci .5692 1757.0 vd. G .6087 1965.8 vd. C .5679 1760.9 vd. H .5068 1973.3 vd. D .5643 1772.2 d. I .5053 1979.0 vd. E .5616 1780.6 m. 4' F .5591 1788.7 s. A' .6021 1991.8 vd. F' .5580 1792.0 vd. B .5008 1996.7 d. G .5560 1798.6 vd. Ci .4987 2005.2 d. H .5538 1805.7 vd. C .4973 2010.9 m. I .5520 1811.6 vd. 7 D .4963 2018.9 d. El .4930 2028.4 d. fB .5461 1831.2 d. E .4922 2031.8 d. Ci .5438 1838.8 d. F .4906 2038.4 s. 5^ C .5424 1843.7 d. Gi .4890 2045.1 vd. D .5392 1854.6 d. G .4880 2049.2 vd. E .5364 1864.3 m. F .5342 1871.8 vs. THE SULPHATES. 175 Table 104. — Uranyl ccesium sulphate: Cs'iU02{SO\)2 2H2O. Fluorescence at —185° C. Prepared by precipitating uranyl sulphate by adding caesium sulphate in calculated amount to form the double salt, which is very insoluble. The composition of the crystals is given as above by O. de Coninck (Chem. Centralblat, ix, 1, 1306, 1095). The preparation consisted of very small square plates about 0.01 mm. on a side, the largest of which showed an apparently uniaxial negative figure. The crystals are therefore presumably tetragonal. Group Group aud M 1/mX103 Int. and M I/mXIO^ Int. series. series. 2 F 0.6129 1631.8 d. [El E .5336 0.5321 1874.2 1879.3 d. m. A .5989 1669.7 vd. F .5299 1887.0 s. B .5964 1676.7 vd. 5' G .5276 1895.4 vd. Ci .5939 1683.8 d. G' .5261 1900.8 vd. 3^ C .5916 1690.3 d. H' .5239 1908.6 d. D .5894 1696.6 vd. I .5228 1912.9 vd. El .5869 1703.9 vd. E .5850 1709.3 d. ^A .5189 1927.2 d. F .5825 1716.6 m. B .5168 1935.0 vd. Ci .5148 1942.5 vd. \^ .5695 1755.9 vd. .5140 1945.7 m. B .5671 1763.4 vd. D .5118 1953.7 vd. Ci .5652 1769.2 d. D' .5110 1956.9 vd. C .5636 1774.3 d. 6 El .5099 1961.0 d. D .5612 1782.0 d. E .5088 1965.5 m. 4 El .5593 1788.0 d. F .5067 1973.6 vs. E .5574 1794.0 m. G .5047 1981.4 vd. F .5550 1801.7 s. G' .5035 1986.1 vd. G .5517 1812.5 vd. H .5019 1992.6 d. H .5490 1821.4 vd. I .5003 1999.0 m. [I .5472 1827.4 vd. ■ : 'A .4970 2012.1 d. A .5434 1840.3 d. C .4920 2032.6 m. B .5410 1848.3 vd. 7 D .4902 2040.0 vd. 5 Ci .5390 1855.2 d. El .4888 2045.8 d. C .5374 1860.8 m. E .4874 2051.6 m. ,D .5353 1868.0 d. .F .4858 2058.6 s. FREQUENCY INTERVALS OF THE FLUORESCENCE SERIES. The average frequency intervals of the various series, as derived from the foregoing tables, are given in table 105, together with the weighted average for each salt. It will be noted that the intervals of the single sulphate and the double salt of csesium are distinctly greater than the intervals of the other four sulphates. There is nothing fortuitous about these differences, for, as will be seen from the table, the different series for each salt have intervals within one frequency unit of the general average for that salt, with three exceptions. These exceptions are series Ci in the ammonium and sodium double sulphates and Gi in the ammonium salt. Such occasional apparent discrepancies are not uncommon in the fluorescence spectra of the uranyl salts. They are not due to accidental errors, but are probably ascribable to the complexity of bands having overlapping components the relative intensity of which in different portions of the spectrum varies progressively. Many such cases are known. A doublet ill- resolved and appearing as a single hazy band, the component of longer 176 FLUORESCENCE OP THE URANYL SALTS. Table 105 — Average frequency intervals in the fluorescence spectra of the sulphates. Series. Intervals. Weighted averages. UO2SO4 +3H2O. (NH4)2U02(S04)2 +2H2O. Na2U02(S04)3 +2H2O. +2H2O. RbjUOiCSOOz +2HaO. CsaUOaCSO*), +2H2O. A A' Bi B B' Ct C C Di D El E Fi F F' Gi G G' Hi H H' I Ji J Average 84.1 84.0 85.6 83.3 85.8 84!9 86!2 84.5 83.5 84.2 84.2 85.6 84.8 83.3 82.8 86.1 85.6 84.2 84.5 83.2 83.8 83.2 83.1 82.7 83.2 86.2 85.6 84.3 84.2 83.3 82.8 83.2 82.1 86.8 82.5 86.7 86.6 85.0 "s5a" 84.6 85.2 83.6 84.4 83.9 84.0 85.0 83.8 83.1 83.8 83.3 83.0 82.5 82.7 82.9 82.7 83.3 86.4 84.4 83.5 86.0 85.3 "sh.k" "S5'.2" 85.2 84.6 83.0 83.6 84.4 84.1 84.9 83.2 83.8 85.6 83.3 83.7 86.8 84.7 85.2 83.7 84.3 83.0 83.2 86.7 wave-length being much stronger in the bands toward the red and d3dng away gradually in subsequent bands as we approach the blue, while the other component steadily increases, will give the effect of an increased frequency interval for the series. The increase might easily be of the general order observed in this case. There is also always the possibility of the presence of a trace of another uranyl compound which would yield additional series. Such cases, for example, are not uncommon in the study of the acetates, where an admixture of the single acetate occurs. ABSORPTION SPECTRA. The difl&culties in obtaining a complete record of the absorption bands of the uranyl sulphates are similar to those described m the preceding chapters. The transmission, like that of the other uranyl salts, ranges progressively from almost complete transparency in the red, yellow, and green to a high degree of opacity in the ultra-violet. Large, clear crystals of the sulphates are not obtainable and there- fore it is not possible to use very thick layers and thus to follow the selective absorption far beyond the reversing region toward the red, as has been done in the case of the chlorides.^ The bands which we were able to locate lie approximately between 2,000 and 2,600 fre- quency units. They belong almost exclusively to the system having 1 H. L. Howes. Physical Review (2). xi, p. 66. 1918. THE SULPHATES. 177 the shorter frequency interval of 70=*=. A few end members of the reversing system, which presumably extends throughout the fluores- cence region, were discernible. Determinations were made in part by photographing the spectrum of the light transmitted by thin layers, in part by the method of reflection. In table 106 the frequencies of the bands in the spectra of the 6 sulphates are arranged by series. Each series, as usual, is designated by a small letter corresponding to the capital letter which denotes the fluorescence series to which it is related. Table 106. — Absorption spectra of uranyl sulphates at —180° C. Salt. Series. Frequenciea. Average interval. 'a, 2056.8 2128.1 2399.8 68.6 a 2060.6 2202.2 70.8 y 2068.7 2140.5 2209.2 70.1 dx 2218.3 d 2016.1 2081.2 2152.9 68.4 U02S04+3H20. ■ e 2093.6 2160.3 2229.5 68.0 /i 2029.9 2170 70.1 / 2102.6 2236.9 2373.0 70.1 /' 2035.8 2102.6 hi 2045.8 2U5.9 2186.7 2327.2 70.3 '» 2260.4 General average 69.6 2061.2 2135.3 2205.8 2275.3 a 71.3 h 2072.5 2212.4 69.9 c 2082.8 2155.1 2226.2 2295.2 2510.0 71.2 di 2016.5 (R)2199.8 2263.0 2409.6 69.9 e 2096.0 2236.2 2305.7 2444.4 69.6 (NH4)2U02(S04)8 +2H2O. fx 2031.8 2448.6 2244.4 2315.6 2383.3 2524.6 2595.4 69.5 70.2 f 2107.7 2178.7 2250.0 2454.3 69.3 r 2039.6 2253,8 2323.8 2392.0 2464.0 2532.8 70.4 G\ 2044.0 2116.0 2187.1 71.5 Q 2330.5 2401.0 2541.0 70.2 h 2127.7 2198.8 2236.0 2409.6 2477.1 70.3 General average 2549.4 70,3 2069.5 'a 0' 2144.6 2214.7 70.2 b 2080.8 2153.3 72.5 c 2093.7 2306.8 2375.9 70.5 dx 2167.3 d 2172.5 2243.8 2314.8 2385.5 71.6 Na2U02(S04)2 +2H20. e e' f 2035.8 2107.5 2250.1 2039.2 2043.9 2114.3 2184.5 71.4 ffi 2050.0 2120.3 2260.4 70.1 (t 2190.1 Q' 2054.7 2055.2 h 2063.4 2128.6 2199.9 71.3 h' 2135.2 2207.7 General average 72.1 71.3 178 FLUORESCENCE OF THE URANYL SALTS. Table 106. — Absorption spectra of uranyl sulphates at —180° C — continued. Salt. Series. Frequencies. Average interval. 'fei 2063.1 2412.3 69.8 b 2064.4 2136.4 2204.6 2276.9 70.8 5' 2065.7 2341.0 68.8 Cl 2071.9 2353.5 70.4 c 2078.6 2289.4 70.2 di 2017.3 2444.1 71.1 d 2375.3 d' 2158.2 K2U02(S04)2 ei 2174.0 +2H2O. e 2095.5 2307.0 2379.8 71.1 e' 2241.7 2312.1 2385.0 71.6 fx 2035.0 2106.8 2246.7 70.6 f 2039.6 2109.3 2179.4 2248.2 2396.6 70.2 r 2250.7 (71 2293.5 g 2116.4 2188.5 2256.7 2323.7 69.1 a' 2047.9 [h 2056.8 2266.3 General aver 2065.6 2136.6 ace • .r> • • • 69.8 70.3 2205.7 2276.1 ffei 69.8 * h 2071.2 ...:.. ci 2218.8 c 2434.3 dx 2014.9 d 2230.2 2301.0 2443.2 71.0 d' 2370.5 Rb2U02(S04)2 +2H2O. ■ e 2028.4 2096.2 2241.2 2307.3 2312.2 2375.1 2378.7 2449.8 76.2 68.8 /i 2106.6 2174.5 2249.2 71.3 / 2037.9 2107.6 69.7 /' 2040.3 2109.7 2179.4 2385.7 69.1 ffi 2045.4 2322.0 2391.3 69.2 a 2048.5 2117.7 2187.7 2256.7 2325.8 69.3 h 2193.9 [h 2085.5 2267.3 General aver 2096.0 2165.4 2340.8 aee 2408.5 70.0 69.8 2236.9 2375.9 2445.0 a 70.3 2517.7 a' 2311.9 c 2103.5 c' 2035.8 2106.4 76.6 d^ 2115.1 CS2U02(S04)2 4-2H20. e f f 2120.0 2261.0 2061.2 2130;8 2133.4 2329.6 2266.6 2399.8 2542.8 2471.6 2613.4 76.5 **69!7 " 2065.5 2206.5 2279.5 71.3 g' 2071.9 2143.6 2215.8 2353.1 2498.3 2426.9 2567.2 70.8 h 2077.2 i 2084.6 2294.1 General average 69.8 70.4 THE SULPHATES. 179 Table 107. — Average frequency intervals for the six sulphates. UO2SO4 69.6 (NH4)2U02(S04)2 70.3 Na2U02(S04)2 71.3 K2U02CS04)2 70.3 RB2U02(S04)2 69.8 C82U02(S04)2 70.4 As is frequently the case in these absorption spectra, one or more bands of a given series are commonly missing or at least not discerni- ble in the negatives. On the other hand, nearly all the bands are found to be members of a series which is definitely related to a fluores- cence series and has the proper frequency interval. The occasional isolated bands, moreover, are so located that they may be definitely associated with a fluorescence series and may reasonably be classed as the sole visible member of an absorption series the remainder of which fails to appear in our photographs. These show no systematic departure from the general average (70.3) for the entire group. Lying as they do within one frequency unit of the average, we may fairly conclude that within the errors of observa- tion, which are rather large on account of the lack of definition and incomplete resolution of these absorption groups, the various sulphates have a common frequency interval. The frequency intervals of the various series of a given salt depart somewhat more widely from the average for that salt, but again there is no systematic variation, and it is probable that all the series would be found to have the same interval, were it possible to locate the bands with greater certainty. Summary. (1) The fluorescence spectrum of the uranyl sulphates consists of 8 equidistant bands, the first and eighth of which disappear at the temperature of liquid air. (2) The remaining bands are resolved into groups of narrow line- like bands, the homologous members of which form series having constant frequency intervals, ranging from 85.7 in caesium uranyl sulphate to 83.0 in potassium uranyl sulphate. (3) The fluorescence groups are distinguished by a strong pair of bands about 8 frequency units apart and 7 weak bands, some of which are doublets. (4) There is a shift of all bands toward the violet, with increasing molecular weights, of about 15 frequency units in passing from the spectrum of uranyl sulphate to that of caesium uranyl sulphate. (5) The absorption spectra of the sulphates are made up of series of bands having a frequency interval of 70.3 (general average). (6) These absorption series extend into group 7 of the fluorescence without break of interval. There are many reversals where fluorescence and absorption overlap. The reversing region is therefore one group farther toward the red than in most spectra of the uranyl compounds. X. THE FLUORESCENCE OF FROZEN SOLUTIONS. The fluorescence spectra of solutions generally consist of only one or two very broad bands. Such bands undoubtedly possess component bands in considerable number, but spectrum analysis often fails to reveal them because of extensive overlapping. Chlorophyl in alcohol possesses a series of absorption bands^ which resemble the absorption bands of the uranyl solutions. The very broad fluorescence band in the orange and red probably consists of several components which form a similar series. Anthracene in solution^ presents a fluorescence series of at least 4 bands which strongly resembles the series foimd in the fluorescence spectra of uranyl solutions. Probably the first observer to note the fact that uranyl solutions yield fluorescence spectra consisting of several bands was G. C. Stokes.^ He states that ''a solution of nitrate of uranium is decidedly sensi- tive," i. e., fluorescent. Later, in the same paper, he writes "I have observed seven of these bands arranged at regular intervals." E. Becquerel,^ in his monumental work on the uranyl salts, makes this observation: " Certain solutions of the salts of uranium give, in the violet rays, a luminous emission scarcely less brilliant than the crystals themselves .... several [bands] appear to correspond to the bands given by the solid salts; the sul- phate and the double sulphate of potassium and uranium are in this class." In the same year H^genbach,^ who was studying many fluorescence materials, observed that the uranyl oxide in nitric acid shows 8 very sharply outlined maxima in the fluorescence spectnmi. Morton and Bolton^ studied the absorption of the uranyl solutions and noted the fluorescence. These investigators were the first to recognize the possibility of the existence of more than one hydrate of the same salt, which, they state, "enables us to explain some discrepancies of authori- ties on this point." Our present work has brought out the necessity, of such a hypothesis. Jones and Strong,^ following the same method as Morton and Bolton, have published the most extensive data on the absorption spectra, but their work does not include temperatures below the freezing-point. This chapter contains the results of experiments which were de- scribed in two papers, together with some additional data heretofore unpubhshed. The first, a preliminary paper, ^ showed that the bands of 1 Nichols and Merritt. Carnegie Inst. Wash. Pub. No. 130, p. 85. 1910. 2 Louise MacDowell. Physical Review (1), 26, p. 155. 1908. ' Stokes. Philosophical Transactions, p. 463. 1852. * E. Becquerel. Comptes Rendus, 75. p. 296. 1872. 5 Hagenbach. Poggendorf Annalen, 146, p. 582. 1872. * Morton and Bolton. Chemical News, pp. 113, 164. 1873. ^ Carnegie Inst. Wash. Pub. No. 130. » Nichols and Merritt. Physical Review (2), 3, p. 457. 1914. 180 FROZEN SOLUTIONS. 181 the solutions, even at — 180**, resembled in breadth and regular spacing those of the solid salts at room temperature. The uranyl acetate in alcohol proved to be the exception, since at —180"^ it resolved into faint lines, which did not, however, coincide in position with those of the solid acetate at that temperature. The variety of shifts with systematic dilution and temperature change led to the second investigation,^ in the hope that some general law of shift might be deduced. It was also planned to study the funda- mental relations between concentration and frequency interval, temperature, and state of resolution, etc. With these relations in view much work was done which led to the discovery of many beautiful and unique spectra. EXPERIMJINTAL METHOD. For the study of the spectra, except where otherwise specified, a Hilger constant-deviation spectrometer was used. The apparatus for the cool- ing and excitation of the sub- stances under observation was designed to enable the observer to hold the tempera- ture of the specimen con- stant at any temperature between 0^ and -180° C. The mounting consisted of a cylindrical copper block M (fig. 90), the top of which was bored to receive a small test-tube Fj which contained the fluorescent solution. The side of the block was channeled to let the exciting light fall on the specimen and to , let the fluorescent light out. To the bottom of this copper block was soldered a cylinder of sheet copper, which could be partially or completely covered by the liquid air in the unsilvered Dewar bulb D, thus produc- ing different temperatures in the specimen. This mount- ing was rigidly suspended from above by partially non- KiG. 90. 1 Howes. Physical Review (2), vol. 6. p. 193. 1915. 182 FLTJORESCENCE OF THE URANYL SALTS. conducting material. The Dewar bulb was fastened to an adjustable support E, and the mounting could be submerged in the hquid air to various depths by raising or lowering the bulb by means of a cord R. ^ Fluorescence was excited by the rays from a carbon arc A. The light passed successively through a water-cell W, a large short-focused condenser C, and a solution of ammonio-copper sulphate B, This solution absorbed all of the exciting Hght of a wave-length greater than 0.4780 ii, so that the fluorescent light, which entered the colli- mator sUt S of the spectrometer, could be viewed on a black back- ground. A small resistance coil T was inserted in a glass tube, which was always placed in the middle of the solution. The temperatures were recorded on a Callender recorder. The massive copper block M served to reduce the vertical temperature gradient in the frozen solu- tion to less than 1"^ per centimeter. It made the apparatus rather slow in responding to changes, but afforded an excellent control of tempera- ture. The salts were carefully weighed and "normal" solutions were prepared.^ Acid solutions were made by adding a definite volume of the commercial concentrated acid to a definite volume of a water solution of known concentration. Although readings were taken and tables made in units of wave- length, the diagrams of spectra are plotted on an arbitrary scale of frequencies, L e., l/ji X 10^ URANYL SULPHATE IN AQUEOUS SOLUTION, The uranyl sulphate in water, upon excitation with the carbon arc, yields 4 bands at +20° C; but when cooled 6 bands are visible, the new bands being of longer wave-lengths. This phenomenon — increase of intensity with cooling — is a very fortunate one, for otherwise the study of the more dilute solutions would be limited to the lowest temperatures. In table 108 will be found the bands of the 1/1, 1/10, 1/100, and 1/1000 normal aqueous solutions. In the spectrum of the 1/1 normal solution band 7 is at 0.4927 ix at +20°, which is of interest because the crystalline salt was found to give a fluorescence band at 0.4925.2 If a reasonable error is assumed, these bands may be considered to be coincident. In this region they are approximately 75 a.u. in width; hence measurements were taken on the crest rather than on the middle of the band, the crests being located slightly nearer the violet edge. The absorption spectrum^ of the normal solution presents a band in this region at 0.4910, which is 17 a.u. nearer the violet than the fluores- ^ The term "normal" solution, as used in this paper, means one which contains the same num- ber of grams of solute to the liter of solvent as the number which represents the molecular weight of the particular salt dissolved. 2 Nichols and Merritt. Physical Review (1), 33, p. 354. 1911. « Jones and Strong. Carnegie Inst. Wash. Pub. No. 130, p. 109. FROZEN SOLUTIONS. 183 cence band. Jones and Strong employed the continuous spectrum of the Nernst lamp, which produces fluorescence in this region. Such luminescence, although masked by the more intense background, tends to shift the crest of the absorption band toward the violet. A comparison of the wave-lengths of the bands of the solid with those of the solution indicates a progressive difference. When the spectra of the solution and of the solid are plotted in frequency units, either is found to include only one series of bands, those of the solution being of a slightly smaller interval than those of the solid. Table 108. — Urany sulphate in water. 1 Solution and temp. Band 2. Band 3. Band 4. Band 6. Band 6. Band 7. Nonnal solution. 1/10 normal solution. 1/100 normal solution. 1/1000 normal solution. '+ 20 - 35 - 60 - 90 -120 -150 -180 ■+ 20 - 30 - 60 - 90 -120 -150 -180 - 35 - 60 - 90 -120 -150 -180 - 90 -120 -160 0.5641 .6641 .6641 .5637 .5635 .5634 .5634 .5636 .5636 .6631 .6629 .5621 .5624 .5629 .5633 .5633 .5634 .6631 .6629 .6631 .6638 .5583 .5574 .5574 0.5383 .5383 .5378 .5376 .6370 .5370 .5373 .6375 .5379 .5373 .5368 .5364 .5365 .5367 .6371 .5373 .5373 .5368 .5365 .5368 .6370 .6330 .6324 .6346 0.5143 .6141 .5141 .5137 .6131 .5133 .5136 .5140 .5139 .5136 .6130 .6125 .6127 .5130 .5133 .5135 .5129 .5125 .6126 .5129 .5132 .5102 .5097 .5105 0.4927 .4926 .4927 .4924 .4919 .4919 .4919 .4921 .4918 .4913 .4911 .4910 .4909 .4913 .4916 .4913 .4911 .4907 .4907 .4909 .4911 .4904 .4904 .4909 0.6229 .6227 .6230 .6236 .6241 0.6934 .5928 .5923 .5923 .5924 .5924 .6228 .6231 ,5911 .5904 .5904 .5917 .5921 .5924 .5928 * The numbers by which unresolved bands are designated in this and the following tables correspond to the group numbers used in previous chapters, since each band corresponds to a group in the resolved spectra of the crystallized salts. In figure 91, band 5 is seen to shift with falling temperature toward the violet, the shift amounting to 13 A.u. when a temperature of — 100° is reached. With further cooling to — 180° the band shifts back toward the red. It would be interesting to ascertain whether this shift toward the red would continue with further decrease in temperature. The other bands of the normal solution behave similarly with falling temperature, i. e., the entire spectrum undergoes a shift to the violet, followed by a reverse shift to the red. The wave-lengths of the bands at —180° are approximately the same as the wave-lengths at —60°. 184 FLUORESCENCE OF THE URANYL SALTS. Jones and Strong discovered a shift of 15 A.u. toward the red for the absorption band of wave-length 0.4910 when the temperature of their solution was raised from +5'' to +84°. Our fluorescence band at 0.4927 shifts in the same direction, with a rise in temperature from -90^0 +20^ H. Becquerel beUeved that any modification of the absorption spectrum is accompanied by a similar change in the fluorescence spec- trum, and these shifts lend strength to his generahzation. A brief study of the fluorescence spectra of the 1/10 and the 1/100 normal solutions at different temperatures indicates that a similar temperature shift occurs. -120 -1 80 Fig. 91. — Uranyl sulphate — Temperature shift. fl) 1/1 normal, aqueous; (2) 1/10 normal, aqueous: (3) 1/100 normal, aqueous; f4J 40 o.o. 1/10 normal, aqueous to 1 o.o. sulphuric acid; (5) 1 o.c. I/IO normal, aqueous to 1 c.c. sulphuric acid. The increase in the amount of solvent produces a shift of the spec- trum toward the violet. For example, band 7 at —90° shifts as fol- lows: in the 1/1 normal solution the wave-length is 0.4919, in the 1/10 normal 0.4910, in the 1/100 normal 0.4907, in the 1/1000 normal 0.4904 m. The 1/1000 normal solution shows a spectrum which is very strongly shifted toward the violet. The above comparison of the wave-lengths of band 7 fails to indicate the shift of the spectrum, because it is accom- panied by a marked decrease in interval, while of all the bands, number 7 is the least shifted. For example, band 4 of the 1/1000 normal at a temperature of —90"^ is of wave-length 0.5583, while band 4 of the 1/100 normal at —90° is of wave-length 0.5631 units. Bands 5, 6, and 7 show progressively less variance in wave-length with the correspond- ing bands of the 1/100 normal solution because of the shorter frequency of the 1/1000 normal interval. Measurements of these spectra plotted on a frequency scale indicate that while the bands are spaced by about 85.7 units in the 1/1, 1/10, and 1/100 normal solutions, the 1/1000 FROZEN SOLUTIONS. 185 normal bands are of only 82 units interval. (See table 109.) This may- be due to a change in the ionization with dilution. To ascertain whether the rate of cooling caused a change in the spectra, a solution was suddenly plunged into hquid air and excited to fluorescence. Measurements were then taken on the bands, but no change in wave-lengths was observed. Table 109.— J Uranyl sulphate in water. Frequencies and average intervals of fluorescence bands. Band. +20° 0° -35° -60° -90° -120° -150° -180° Frequencies in normal solution. [2 3 4 5 6 7 1605.4 1686.9 1773.9 1860.1 1946.7 2030.9 1605.9 1688.3 1774.6 1862.2 1948.9 2032.9 1605.1 1688.3 1774.9 1862.2 1948.2 2032.9 1603.8 1688.0 1774.9 1861.2 1947.0 2032.1 1602.3 1688.0 1774.3 1860.5 1945.5 2030.0 1685.2 1772.7 1857.7 1945.1 2029.6 1772.7 1857.7 1944.4 2029.5 1772.7 1857.7 1945.1 2030.0 Average int . . 85.6 85.8 86.1 85.1 85.4 85.6 85.7 85.5 Frequencies in 1/10 normal solution. 2 3 4 5 6 7 1605.6 1690.0 1776.5 1863.3 1949.3 2035.4 1604.9 1688.9 1775.8 1861.9 1948.2 2034.2 1691.8 1776.5 1862.9 1949.3 2036.2 1693.8 1779.0 1864.3 1951.2 2036.7 1693.8 1778.1 1863.9 1950.5 2037,1 1774.3 1859.1 1945.9 2033.3 1775.9 1861.2 1947.0 2035.4 Average int , . 86.3 86.5 86.1 85.7 85.8 86.0 85.9 Frequencies in 1/100 normal solution. 3 4 5 6 [7 1688.0 1775.9 1862.9 1949.7 2037.1 1686.9 1773.7 1862.3 1948.6 2036.2 1775.3 1861.2 1947.4 2035.4 1774.9 1861.2 1949.7 2036.2 1775.9 1862.9 1951.2 2037.9 1776.5 1863.9 1950.8 2037.9 Average int 86.7 87.1 87.3 87.1 87.3 87.3 Frequencies f4 in 1/1000 Is normal 1 6 solution. [7 1791.2 1876.2 1960.0 2039.4 1794.0 1878.3 1961.9 2039.4 1794.0 1870.9 1958.9 2037.1 1794.0 1870.9 1958.9 Average in< 82.7 81.8 81.0 82.4 URANYL SULPHATE MIXED WITH SULPHURIC ACID. It has been observed that the bands of the aqueous solutions move toward the violet with progressive dilution with water; hence it was of considerable interest to ascertain the effect of dilution with sul- phuric acid. The addition of one volume of acid to 40 volumes of the 1/10 normal aqueous solution (table 110) produces a negligible effect, but a mixture of equal volumes shifts the bands back toward the red, in fact, the wave-lengths of the bands at 4-20° are longer than those of the normal solution — ^aqueous. This can be discovered from a comparison of the wave-lengths of the 1/10 normal solutions with those of the normal solutions in tables 108 and 110. The effect is not evident at low tern- 186 FLUORESCENCE OF THE URANYL SALTS. peratures, because the spectrum of the 1 to 1 acid solution persistently shifts toward the red instead of reverse shifting at —100° (see fig. 91). The frequency interval remains unchanged, with a proportionately large acid dilution. Table 110. — Uranyl sulphate in sulphuric acid. Temp., etc. Band 2. Band 3. Band 4. Band 5. Band 6. Band 7. + 10** 0.5634 0.5376 0.5140 0.4931 - 30° 0.5929 .5635 .5376 .5139 .4928 40 c.c. of 0.1 normal aque- - 60° 0.6238 .5924 .5631 .5373 .5135 .4919 ous solution with 1 c.c. - 90° .6224 .5910 5621 .5365 .5129 .4919 of acid. -120° .6226 .5914 .5625 .5362 .5126 .4920 -150° .6234 .5918 .5627 .5365 .5128 .4925 ,-180° .6242 .5919 .5630 .5367 .5130 .4925 f-h 20° .5945 .5657 .5388 .5149 .4921 - 30° .6246 .5935 .5640 .5376 .5139 .4916 1 c.c. of 0.1 normal aque- - 60° .6231 .5921 .5630 .5368 .5133 .4911 ous solution with 1 c.c.-* - 90° .6215 .5906 .6616 .5359 .5124 .4905 of acid. -120° .6199 .5893 .5608 .5348 .5114 .4898 -150° .6203 .5893 .5607 .5347 .5113 .4897 [-172° .6203 .5890 .5602 .5345 .5109 .4893 Table 111. — Uranyl sulphate in sulphuric acid. — Frequencies and average intervals of fluorescence hands. Band. +10° -30° -60° -90° -120° -150° -180° Frequencies of 40 c.c. of 0.1 normal aqueous with 1.0 c.c. of acid. ^2 3 4 5 6 7 1603.1 1688.0 1775.9 1861.2 1947.4 2032.9 1606.7 1692.0 1779.0 1863.9 1949.7 2032.9 1606.2 1690.9 1777. S 1865.0 1950.8 2032.5 1604.1 1689.8 1777.1 1863.9 1950.1 2030.5 1602.1 1689.5 1776.1 1863.2 1949.3 2030.5 i7U.9 1860.1 1945.5 2028 . 1686.7 1774.6 1860.1 1945.9 2029.2 Average interval. . . 84.4 85.6 86.0 85.2 85.3 85.3 85.7 Frequencies of 1 c.c. of 0.1 normal aqueous- with 1 c.c. of acid. 2 3 4 5 6 7 'i682!i' 1767.7 1856.0 1942.1 2032.1 1601.0 1684.9 1773.1 1860.0 1945.9 2034.2 1604 . 9 1688.9 1776.2 1862.9 1948.2 2036.2 1609.0 1693.4 1780.6 1866.0 1951.6 2038.7 1613.2 1696.9 1783.2 1869.9 1955.4 2041.7 1612.1 1696.9 1783.5 1870.2 1955.8 2042,1 1612.1 1697.8 1785.1 1870.9 1957.3 2043.7 Average interval ... 87.5 86.6 86.3 86.0 85.7 86.0 86.3 URANYL POTASSIUM SULPHATE IN WATER. The spectra of the aqueous solutions of uranyl potassium sulphate, like those of uranyl sulphate, consist of a single series of bands. The temperature shift of the more concentrated solutions, e, g., the 1/15 and 1/150 normal, is at first toward the violet, followed by a reverse shift toward the red. The wave-lengths of the bands of the aqueous solutions are recorded in table 112. It will be seen that the shift toward the red is more marked than in the uranyl sulphate solu- tions. (See also fig. 92.) FROZEN SOLUTIONS. 187 The 1/1500 and 1/15,000 normal solutions yield bands which present a hazy appearance, lacking the pronounced crests of the more con- centrated solutions. For this reason the readings of such wave- lengths are more Ukely to be in error. The same tendency to first shift toward the violet and then shift toward the red is evident. Table 112. — Uranyl potassium sulphate in water. Solution and temp. 1/15 normal solution. 1/150 normal solution. 1/1500 normal ■ solution . 1/15000 normal solution. + 20° 0° - 30° - 60° - 90° -120° -150° -180° + 20° - 30° - 60° - 90° -120° -150° -180° - 30° - 60° - 90° -120° -150° -180° - 60° - 90° -120° -150° -180° Band 3. 0.5942 .6945 .5945 .6946 .5953 .5966 .5956 .5966 Band 4. 0.5656 .6651 .5660 .6652 .5663 .5653 .5653 .5669 .5659 .6657 .5667 .6656 .5667 .5679 .5612 .5612 .5624 .6630 .5643 .5660 .5483 .5464 .5456 .5465 .5476 Band 5. Band 6. Band 7 0.5388 .6387 .5386 .5386 .5386 .5387 .5393 .5406 .5398 .5391 .5388 .5393 .5394 .5403 .5411 .5378 .6368 .5373 .5382 .5382 .5385 .5257 .5338 .5233 .6234 .5235 0.5154 .6152 .6148 .5148 .5148 .5149 .5154 .5166 .6155 .6153 .5152 .5153 .6162 .5166 .5171 .5147 .6149 .5149 .6162 .5152 .5153 .5038 .6027 .6021 .5020 .5021 0.4932 .4928 .4926 .4927 .4925 .4928 .4933 .4945 .4931 .4928 .4931 .4932 .4941 .4951 .4938 .4938 .4936 .4931 .4931 .4938 .4831 .4829 .4829 .4828 The frequency interval, as nciay be seen from table 113, suffers a marked change with dilution. The 1/15 and 1/150 normal solutions show spectral series of 86.8 and 85.5 average interval respectively, the 1/1500 series of 83.4 units, the 1/15,000 series of only 80.2 units. The 1/1500 series undergo marked increase of interval on cooling. The bands of the 1/15,000 normal are so greatly shifted that they lie approximately in the middle of the intervals between the bands of the 1/1500 normal solution. Such a "shiff of the entire spectrum must be due to a marked change in the molecular arrangement ; hence it can hardly be designated as a shift of the 1/1500 normal spectrum. Pre- sumably a new hydrate has been formed by the freezing of the 1/15,000 normal solution. The 1/150,000 normal solution gave a spectrum which was too dim to permit of measurement, except at the lowest temperatures. From the three bands which are visible, it appears that the frequency interval 188 FLUOKESCENCE OF THE URANYL SALTS. is approximately 77 units. If so, it is the shortest frequency interval yet discovered in the fluorescence of the uranyl salts. A comparison of the location of the bands of the sohd potassixmi sulphate with those of the 1/15 normal solution shows that the bands of the solutions are from 12 to 40 A.u. nearer the red, according to their wave-lengths. URANYL POTASSIUM SULPHATE IN SULPHURIC ACID. The addition of sulphuric acid in moderate proportions to the aqueous solutions of the potassium sulphate increases the intensity and improves the resolution of the bands. A solution of the 6 c.c. of 4/15 normal aqueous solution to 1 c.c. of acid was subjected to the coohng process. In figures 92 and 93 and table 114 the shifts will be observed. Band 5, which is typical of the other bands, shifted toward the violet by 21 A.u. at shifted 6 A.u. toward the red at —180*^. ■120^ and then + 20 . -60 . -180 Fig. 92. — Uranyl iwtassium sulphate — ^Temi>erature shift. (1) 1/15 normal, aqueous; (2) 1/150 normal, aqueous; (3) 1/1500 normal, aqueous; (4) 5 p.o. of 4/15 normal, aqueous to 1 c.o. aulphurio acid; (5) 1 c.c. of 4/15 normal, aqueous to 1 c.o. sulphurio acid. The addition of acid in larger proportion — 1 c.c. of acid to 1 c.c. of 4/15 normal solution — ^resulted in a spectrum which shifted toward the violet in a peculiar fashion with each decrement of temperature. (See curve No. 5 of fig. 92.) The total shift of band 5 amounted to 42 A.u. Further dilution with acid, e. g., 20 c.c. of acid to 1 c.c. of solution, resulted in another broad-banded spectrum at +20°. With cooling, however, partial resolution occurred. This is best observed in figure 94, where several narrow bands appear in the regions formerly occupied by the broad bands. Homologous components are lettered a, 6, and c. Still greater dilution — 50 c.c. of acid to 1 c.c. of aqueous solution — produced a spectrum which passed through the same development; FROZEN SOLUTIONS. 189 +2 * ±un. POTASSIUM SULPHATE IN SULPHURIC ACID. 5:1 0- .^^ ^T\ /i\ /i\ -30^ XfA / I \ ^ 1 \ /i\ A\ -eo' yT> ^-T\ y]\ A\ /l\ -»()• yr\ XiA A\ /W /r\ -no* -/Tv A\ ylA / ' \ As /]\ -ISO" ^T\ y\ x^ /^ /W r\ -"">' ^T\ ^T\ /iA / 1 \ A\ /i\ I6l00 1 leloo 1 20IOO I .?oi-« .S6|>^ .321^ .4fll>^ Fig. 93. Table 113. — Uranyl potassium sulphate in water. — Frequencies and average intervals of fluorescence hands. Band. +20'' 0'' -30'' -60° -90° -120° -150° -180° Frequencies in 1/15 normal solution. ^3 4 5 6 7 1682.9 1769.9 1856.7 1942.5 2030.0 1682.1 1769.3 1856.7 1942.6 2029.6 1682.1 1769.0 1856.6 1942.5 2030.0 1681.8 1769.0 1856.3 1942.1 2029.2 1679.8 1769.0 1854.3 1940.2 2027.2 1676.2 1764.0 1849.8 1935.7 2022.2 1768.0 1856.0 1940.2 2027.6 1769.6 1856.3 1941.0 2029.2 Average int . . 86.5 86.6 86.8 86.9 87.0 86.9 86.9 86.5 Frequencies in 1/150 normal solution. 3 4 5 6 7 1679.0 1764.6 1850.8 1935.7 2023.9 1676.1 1760.9 1848.1 1933.9 2019.8 1767.1 1854.9 1940.6 2028.0 1767.7 1856.0 1941.0 2029.2 1767.7 1864.6 1940.0 2028.0 1768.0 1853.9 1941.0 2027.6 1852.5 1939.9 Average int . . Frequencies (4 in 1/1500 15 normal 16 solution. [7 87.4 87.0 87.2 86.8 86.6 86.2 85.9 1781.9 1860.1 1942.9 2026.1 1781.9 1862.9 1942.1 2026.1 1778.1 1861.2 1942.1 2025.9 1776.1 1858.0 1941.0 2028.0 1772.1 1858.0 1941.0 2028.0 1769.9 1867.0 1940.6 2025.1 Average im 81,1 81.1 82.6 84.0 85.3 86.1 Frequencies (4 in 1/15000 15 normal ] 6 solution. [7 1823.8 1902.2 1984.9 1830.2 1909.1 1989.3 2070.8 1832.8 1911.0 1991.6 2070.8 1829.8 1910.6 1992.0 2070.8 1826.2 1910.2 1991.6 2071.3 Average im 80.6 79.3 79.3 80.3 81.7 190 FLUORESCENCE OF THE URANYL SALTS. +?• jl UR. POTASSIUM SULPHATE IN SULPHURIC ACID. 1:20 ---TV -38' /T\ /TN X^ /T\ -90'* yvK A /R /A /l\ -izo-* yry yn /i^ /A Afi\ -150" /!> /A /l\ m -180° A °i/f cJl ^_ 16100 isloo 20|00 .641/* .€0\^ .56l>«- ,52|^ .481^ , Fia. 94. hence it is clear that the presence of sulphuric acid in excess is essential to resolution of this type. The homologous components formed frequency intervals which were constant and of the same length as those of the parent bands; i. e., the homologous components form separate series. Table 114. — Uranyl potassium sulphate in sulphuric acid. Temp., etc. Band 2. Band 3. Band 4. Band 5. Band 6. Band 7. 6 c.c. of 4/15 normal aque- ous solution with 1 c.c. of acid. 1 c.c. of 4/15 normal aque- ous solution with 1 c.c. of acid. H- 20° 0° - 30° - 60° - 90° -120° -150° -180° -1- 20° - 2° - 35° - 63° - 90° -120° -150° -180° ' 0.5392 .5388 .6382 .5379 .5377 .5371 .5373 .5377 .6385 .5378 .5374 .5364 .5362 .5350 .5347 .5343 0.5150 .5150 .6145 .6142 .5142 .5136 .5139 .5141 .5141 .5142 .6136 .5131 .5127 .5117 .5113 .5107 0.4931 .4933 .4927 .4928 .4926 .4921 .4926 .4928 .4921 .4918 .4916 .4911 .4908 .4898 .4896 .4892 0.5650 .5642 .5638 .5635 .6625 .5631 .5633 .6650 .6647 .5635 .5629 .5621 .5611 .6603 .5600 0.5917 .5917 .5911 .5914 .5917 0.6219 .6220 .6226 .5946 .5938 .5921 .5915 .5903 .5887 .5885 .6230 .6227 .6211 .6203 .6194 Temp., etc. Band 3. Band 4o. Band 5a. Band 56. Band 6a. Band Band 7a. Band 7b. 1 c.c. of 4/15 normal aque- ous solution with 50 c.c. acid. 0° - 30° - 60° ( - 90° -120° —150° -180° )!5966 .5966 .5968 .6979 .5942 0.6662 .5666 .5662 .5661 .5659 6fifi3 0.5388 .5394 .5388 .5387 .5385 .6380 .6376 0.5168 .5151 0.5138 .6146 .6141 .5140 .5139 5119 0.4914 .4916 .4911 .4905 .4902 .4897 .4895 4930 .5 634 0.5401 5118 .49 28 FROZEN SOLUTIONS. 191 10 :o .60U* 10 :i /A ^ /A A\ y1\ 5:1 ^^ /T\ /i\ /^ y^ 1:1 yT\ /[\ y^ A\ y?i 1:5 ^^ y1\ A\ /A y1\ i:20 >^ aAA MA y/ViU /^A 1:50 >rv A aAA J\f\ AA 1 1 1 1 1 1 1600 1800 Fig. 95. 2000 Table 115.- —Uranyl potassium sulphate in sulphuric acid- intervals of fluorescent^ hands. —Frequencies and average Band, etc. +20° 0° -30° -60° -90° -120° -160° -180° Frequencies of 5 c.c. 4/15 normal - aqueous with 1 c.c. acid. 2 3 4 5 6 [7 1608.0 1691.8 1777.8 1861.9 1947.0 2032 . 1 1607.7 1690.9 1775.9 1861.2 1945.9 2030.0 1606.2 1690.0 1775.3 1859.8 1945.1 2029.2 1690.0 1773.7 1859.1 1944.8 2029.2 1690.0 1774.6 1859.8 1944.8 2030.0 1941.7 2028.0 1769.9 1856.0 1941.7 2027.2 1772.4 1858.0 1943.6 2029.6 Av. int 86.7 85.8 85.7 84.8 84.8 84.8 84.5 84.6 Band, etc. +20° -2° -35° -63° -90° -120° -150° -183° Frequencies of 1 c.c. 4/15 normal - aqueous with 1 c.c. acid. r2 3 4 6 .7 1605.1 1688.9 1776.6 1864.3 1948.9 2036.2 1605.9 1690.6 1779.0 1865.0 1950.5 2037.5 1610.0 1694.1 1782.2 1869.2 1954.3 2041.7 1612.1 1698.7 1784.8 1870.2 1955.8 2042.5 1614.5 1699.2 1785.7 1871.6 1958.1 2044.2 1769!9 1857.0 1945.1 2032.1 1681.8 1770.9 1859.4 1944.8 2033.3 1684.1 1774.6 1860.8 1947.0 2034.2 Av. int 87.4 87.9 87.5 86.2 86.3 86.3 86.1 85.9 Band, etc. +20° 0° -30° -60° -90° -120° -150° -180° Frequencies of 1 c.c. 4/15 normal ■ aqueous with 50 c.c. acid. f3 4a 5a 5b 6a 66 7a [7b 1676.2 1766.2 1676.2 1766.5 1675.6 1767.1 1672.5 1765.8 1682.9 1774.9 1851.6 1860.1 1941.4 1953.9 2029.2 2042.9 1766.2 1764.9 1856.0 1853.9 1856.0 1856.3 1857.0 1858.7 1938.7 1953.5 2028.4 2042.1 1946.3 1943.3 1945.1 1945.5 1945.9 2035.0 2034.2 2036.2 2038.7 2040.0 Av. int 89.6 89.8 90.0 90.6 91.1 92.4 90.0 192 FLUORESCENCE OF THE URANYL SALTS. The effect of dilution with acid at one temperature is given in table 115. The —180" spectra of the 4/15 normal aqueous solution with varying proportions of acid are shown in figure 95. With the addition of acid the bands at first move toward the violet without resolving, then become stationary in position, and finally resolve. The ratio by volume of aqueous solution to sulphuric acid is given for each spectrum. The shift is not the same for the different bands, because the frequency interval, beginning with about 85 units for the aqueous solution, increases with increase of acid component to about 90 units in the 50 parts acid to 1 part water solution. With the exception of the two resolved spectra, the bands are too diffuse to permit of satisfactory intermediate measurements on the frequency intervals. Dilution with acid has undoubtedly increased the interval by 5 units, whereas dilu- tion with water decreased the interval by 8 units. URANYL CHLORIDE IN WATER. The absorption spectrum of the chloride is of particular interest, since Jones and Strong first located absorption bands^ in the fluores- cence region in an aqueous solution of this salt. Observations by the authors on the transmission spectrum of several crystals of the uranyl double chlorides of potassium, ammonium, rubidium, and caesium have Table 116. — Uranyl chloride in water. Solution and temp. Band 2. Band 3, Band 4 0.05 normal solution. 0.6247 .6250 .6254 .6255 .6245 .6246 .6252 .6252 .6236 .6249 .6251 .6254 .6253 .6254 .6252 0.5935 .5940 .5939 .5939 .5931 .5935 .5936 .5934 ,5938 .6936 .5933 .5936 .5941 .5942 .5939 .5935 0.6639 .5644 .5644 .6645 .5643 .5643 .5643 .5642 .5648 .5646 .5646 .5645 .5645 .5647 .5649 .5646 Band 6. Band 6, 0.6383 .6379 .6379 .6386 .5382 .6382 .6382 .6382 .5385 .6386 .5385 .6382 .6382 .6381 .6381 .6382 0.5142 .5140 .5139 .6143 .5141 .5142 .5141 .5141 .5146 .5144 .5143 .6144 .5144 .5141 .5142 .5144 Band 7. 0.4927 .4926 .4925 .4926 .4926 .4929 .4926 .4926 .4926 .4923 .4924 .4926 .4923 .4923 .4923 .4924 resulted in the discovery of absorption bands in the same region. The view held by Jones that the fluorescence spectrum is a continuation of the absorption spectrum is to be gravely doubted, for while the chloride solution shows a fluorescence band at 0.4926 and Jones has established the position of an absorption band at 0.4920, none of the other bands ^ Jonea and Strong. Carnegie Inst. Wash. Pub. No. 130, p. 90. FROZEN SOLUTIONS. 193 located by him at 0.6070, 0.6040, 0.6020, 0.6000. 0.5200, or 0.5185 coincide with a band of the fluorescence spectrum. Furthermore, it has previously been indicated that often the last band of a fluorescence spectrum coincides fairly well with a strong band in the absorption. It has also been shown in our study of the fluorescence and absorption spectra of the crystalline salts (see Chapters III to IX) that the interval between the absorption bands, although constant, is much smaller than that between fluorescence bands. The bands of the chloride in solution are separated by a very black background, but are so dim that cooling to —90° is necessary before measurements can be made. The bands continue to increase in brightness as the temperature is further decreased. The temperature ^ shift between —90° and —180° is toward the red in the spectrum of the 3.0 normal solution. The measurements on the chloride, to be found in tables 116 and 117, indicate that difficulty is experienced in locating the positions of the bands. The remarkable Table 117. — Uranyl chloride in water — Freqiiendes and average intervals, flitorescence hands. Band. -97° -120° -150° -180° \2 1600.8 1600.0 1599.0 1598.7 Frequencies 3 1684.9 1683.5 1683.8 1683.8 in 3.0 4 1773.4 1771.8 1771.8 1771.5 normal 5 1857.7 1859.1 1859.1 1856.7 solution. 6 1944.8 1945.5 1945.9 1944.4 Av. int . . . 7 2029.6 2030.5 2030.5 2030.0 85.8 86.1 86.3 86.3 1601.3 1601.0 1599.5 1599.5 Frequencies 3 1686.1 1684.9 1684.6 1685.2 in 1.5 4 1772.1 1772.1 1772.1 1772.4 normal i 5 1858.0 1858.0 1858.0 1858.0 solution. 6 1945.1 1944.8 1945.1 1945.1 Av. int . . . [7 2030.0 2028.8 2030.0 2030.0 85.7 85.6 86.1 86.1 f2 1603.6 1600.3 1599.7 1599.0 Frequencies 3 1684.1 1681.8 1685.4 1584.6 in 0.5 4 1770.5 1771.2 1771.2 1771.5 normal 5 1857.0 1856.7 1857.0 1858.0 solution. 6 1943.3 1944.0 1944.4 1944.0 Av. int . . . L7 2030.0 2031.3 2030.9 2030.0 86.3 86.2 86.2 86.2 ^2 1599.2 1599.0 1599.5 Frequencies 3 1683.2 1682.9 1683.8 1684.9 in 0.05 4 1771.5 1770.9 1770.2 1771.2 normal 5 1858.0 1858.4 1858.4 1858.0 solution. 6 1944.0 1945.1 1944.8 1944.0 Av. int . . . J 2031.3 2031.3 2031.3 2030.9 87.0 86.4 86.5 86.3 194 FLUORESCENCE OF THE URANYL SALTS. fact is that the bands of the 1.5, 0.5, and 0.05 normal solutions are not shifted by temperature, and that dilution from 3.0 normal to 0.05 normal produces a negligible shift. There is no tendency toward reso- lution. Clearly, the uranyl chloride in aqueous solution furnishes spectra of great stability, especially in view of the behavior of the bands of the uranyl nitrate. URANYL NITRATE IN WATER. The solutions of uranyl nitrate present at once the most interesting and most complicated spectra. In our first investigations the solutions were studied at —185° after suddenly plunging them into liquid air. Later it became of interest to study them at several intermediate NITRATE IN WATER. _^Ds ^^TX -180* m ^ /h^T\M A\ I >^T^ ^ I I y'T\y,. I 1 >^Tv/K J_L ^^r-\ LI_^^MLdL^iJj_^^a/Lj I I f^NlTRATE IN WATER. -180° .-^IL_ Jh. ^^rs. JJL -.4L JL y^ y^V^ JL NITRATE IN NITRIC ACID, -leo" i:io .ZH >^ y i \ .^T\ ^TV /lA /^ ^ aA 1:3.5 VI\ I8|00 2000 .sel^^ .n\^ Fig. 96. temperatures and the freezing and subsequent cooling was of necessity done slowly. To our surprise, the normal solution of the nitrate yielded an entirely different type of spectrum. Spectrum No. 1 at the top of figure 96 represents the old type and spectrum No. 5 the new type. It was found possible to produce intermediate degrees of resolu- tion somewhat similar to Nos. 2, 3, and 4 by intermediate rates of cooling. The pertinent fact is that the identical solution could, by FROZEN SOLUTIONS. 195 manipulation of the cooling process, be made to yield either an unre- solved or a highly resolved spectrum. The intermediate forms were not easy to reproduce at will. A comparison of the wave-lengths of the strongly resolved bands of the solution at —180° with those of the crystalline salt at the same temperature showed that they were identical. The uranyl ammonium nitrate and uranyl potassium nitrate in aqueous solution were similarly cooled and showed resolution of the same type. Resolution of this type has not been discovered in any other aqueous solutions, but in our first investigation uranyl acetate in alcohol was found to give highly resolved but quite dim bands superimposed on a continuous background. The spectra of the 1/100 normal solution were similarly affected by retarding the rate of cooling. Spectra L, M, and N of figure 96 show the results of successively slower rates of cooling. EFFECT OF TEMPERATURE ON SLOWLY COOLED SOLUTIONS OF URANYL NITRATE. Since the changes in the spectrum of the normal solution are very striking and are typical of the changes in many other more dilute solutions, a detailed account of the changes in this spectrum is given. Figure 97 gives a plot of the spectra. Some attempt at indicating the NITRATC IN WATER. + 20" -60' /\ /^ A\ A A_ _yi^ /\ , A ^ n ^ /; !l ^ A ■""T^- U -180 Art 16100 I laiOO I 20100 ^ .64t^ .M\^ ,56\Ji .S2I^ .481^ Fig. 97. form of the bands is made, but the changes in intensity are too great to be represented on such a plot. The wave-lengths are tabulated in table 118, and frequencies in table 119. At +20"" only two broad bands located at 0.5323 and 0.5088 were of sufficient intensity to be measur- 196 FLUORESCENCE OF THE URANYL SALTS. able. With falling temperature these increased in brightness and two more bands came up to the threshold of vision. Bands 0.4890 at 0° corresponds with an absorption band at 0.4870 discovered by Jones and Strong. The crystalline nitrate, with 6 H2O, also has an absorp- tion band at 0.4870. While continuing the cooling process at a slow rate, a sharp rise in temperature from -25° to -18° was invariably noticed, probably due to undercooling or change in hydration. Inamediately following this stage portions of the background increased greatly in brightness so as to broaden each band on the violet side. These very broad bands, which exist at temperatures between —25° and —40°, were found on subsequent cooling to be the parents of groups of resolved bands. The —40° bands were five times the intensity of the +20° bands. Table 118. — Uranyl nitrate in water — normal solution. Temperat lire. +20** 0« -40° -60° -90° -120° -160<» -180° Strong. . 0.6175 0.6174 .6039 .5855 0.6161 .6030 .5857 .5829 .5722 .5579 .5553 .6457 0.6183 .6022 .5861 .5827 .5713 .5577 .5554 .6454 0.6165 .6006 .5857 .6823 .5712 .5576 .5553 ,5455 ,5368 .5322 .5299 .5217 .5132 .5089 .5068 .6008 .4991 .4951 .4914 .4855 Weak. Strong. . 0.5932-0.5696 .5864 Weak. Dim. , . . .5723 .5584 Strong 0.5622 .5582 Dim. . . . Dim.. . .5461 Dim. Strong. . Dim. 0.5323 .5360 .5375-0.5182 .5323 .5331 .5299 .5219 .5322 .5300 .5216 .5322 .6299 .5214 .5129 .5090 .5068 Dim. . . . Dim. strong. . Dim. .5088 .5110 .5129-0.4951 .5082 .5093 .5065 .5093 .5069 Dim Dim. . . . .4924^6.4831 .4999 .4997 .4997 Dim. Dim .4914 .4857 .4916 .4856 Strong. . .4890 .4862 .4868 At —46° the portion of each band toward the violet decreased in intensity as the part of longer wave-length became stronger, thereby tending to both narrow the band and produce a decided crest. It was found, with the aid of the spectro-photometer, that the intensity of the stronger crest at — 60° was 85 times that of the homologous band at +20°. Further cooling resolved the stronger band into doublets without a real shift, but the dimmer component was not so easily resolved. At temperatures between — 120° and — 180° the strongly resolved doublets formed two series, both of a constant frequency interval of 88 units, the single band at 0.4885 being a member of one series. There was FROZEN SOLUTIONS. 197 ro NITRATE IN WATER. -30 -60" /^ A A\ A Jl .^ffl- -^01. JL H ^ l\^ /^ -120" .Zi- JL .--^ /il ^ /il ^ ^ -ISO* _-db ^^ ^^ -^ ^di_ A -185' _-dki ^3i- .-i:lL y^ ^ ^'^T\ ^ idoo J_ laloo I aoloo -»ol>^ .sg|>^ ■52l^ ■4gl^ Fig. 98. no shift here, but increasingly better resolution. The very dim inter- mediate bands resolved to form series of approximately the same interval. It was thought that by reversing the coohng process the spectra might go through the same forms at the same temperatures, which Table 119. — Uranyl nitrate in water — frequencies and average intervals of fluorescence hands. Band. -20° 0« -40° -60° -90° -120° -160° -180° n 3 4 6 I 7- Av.int 1619.4 1619.7 1655.9 1705.0 1623.1 1658.4 1707.4 1715.6 1747.6 1792.4 1800.8 1832.5 1622.6 1660.6 1706.2 1716.1 1750.4 1793.1 1800.6 1833.5 1616.8 1665.0 1707.4 1717.3 1750.7 1793.4 1800.8 1833.2 1862.9 1879.0 1887.1 1916.8 1948.6 1965.0 1973.2 1996.8 2003.6 2019.9 2035.0 2059.7 1685.8-1755.6 1705.3 1747.3 1790.8 1778.7 1791.5 1831.2 1878.6 1869.2 1860.5-1929.8 1878.6 1875.8 1887.1 1916.1 1879.0 1886.8 1917.2 1879.0 1887.1 1917.9 1949.7 1964.6 1973.2 1965.4 1966.9 1949.7-2019.8 1967.7 1963.5 1974.3 1963.5 1972.8 2030.9-2070.0 2000.4 2001.2 2001.2 2035.0 2058.9 2034.1 2059.3 2045.0 2056.8 2058.5 86.8 88.8 89.8 87.5 87.8 87.2 87.3 88.6 198 FLUORESCENCE OF THE URANYL SALTS. proved to be the case when the temperature was raised from — 180° to —60°, but on further heating to —30° the —60° spectrum failed to change over to the very broad banded form. Finally, the temperature was raised to —18°, the cryohydrate point. The original spectrum of the unfrozen solution then reappeared. 100 NITRATE IN WATER. -«0" ^^ ^T-\ ^T\ -120' -^ — ^:i^^ /iX yr\ /il .^r^ y\ -150' A y^r\ -185" /n ^ ^^ m ^T\ 7L wloo leloo 20|00 .60Ia ■56U .52|>^ ■48l>«' Fig. 99. The slow cooling of the 1/10 normal solution produced a series of spectra similar to the above, but dimmer and not quite as well defined. (See fig. 98.) The 1/100 normal, although too dim to be measured 2o:i J ^Ei_ NITRATE IN NITRIC ACID. -160 I A A A A A I A A I A U 10 :i 1 -iEi ^ I J /\ y\ ^ \ \ A A AaI \Aj\ a A/\\\ A r\ I A 2:1 \j\ A^ /T\ ^ Atl/ t /^ -A AA I A A A AA\\aj\ I /1\ \/\ _I " I A 1^2.5 /^ ^ y\ ± 1:2.75 /^ A _ZIV- A /DA JL ^r\ A xr\ /I / ! \ Jh^ JL -ZIL. ^^ ^^ . ^\ ^ A /^ /i\ A 1:75 isloo A isloo _Zii ^^n^ ^i:^ rx /I\ >^rv _-/n_ ^TV 20IOO ,^t\xt ■321.^ .491^ Fig. 100. FROZEN SOLUTIONS. 199 until a temperature of —60° was reached, behaved similarly. (See fig. 99.) The more dilute aqueous solutions, e. gr., the 1/200 and 1/500 normal, gave broad bands with no important shifts. The very dim, broad bands of the 1/1000 and 1/10,000 normal are probably due to the production of different hydrates. URANYL NITRATE IN NITRIC ACID. The spectra of the normal aqueous solution diluted wifch nitric acid in varying proportion are shown in figure 100. Data for 5 c.c. of acid are given in tables 120 and 121. Table 120. Uranyl nitrate in nitric acid (1 c.c. of normal aqueous solution with 5 c.c. of acid). -30° -60° -90° -120° -150° -180° 0.5958 .5818 .5673 .5527 .5399 .5280 .5163 .5045 .4935 .4825 0.5965 .5806 .5654 .5531 .5393 .6279 .5154 .5043 .4930 .4823 0.5817 0.5810 0.5808 .5681 .5528 .5412 .5274 .5165 .5042 .4941 .4824 0.5540 .5540 .5531 .5289 .5283 .5274 .5064 .5047 .5045 .4832 .4823 Uranyl nitrate in methyl alcohol (0.1 normal solution). Uranyl nitrate in ethyl alcohol (0.1 normal solution). -120° -135° -150° -180° -90° -120° -150° -180° 0.6045 .5882 .5780 .5600 .5503 ,5341 .5252 .5104 .5025 .4887 .4815 6!5537" .5287 .5069 0.5834 .5557 .5301 .5072 0.5836 .5552 .5300 .5072 .05788 .5521 .5270 .5040 .4889 .4828 0-.5841 ".5569 .5691 0.5894 .5804 .5605 .5525 .5345 .5272 .5106 .5041 .4887 0.5891 .5780 .5602 .5510 .5344 .5263 .5108 .5031 .4888 .4819 The uppermost spectrum, denoted at the left by 20 : 1, was pro- duced by slowly cooUng to —180° and exciting to luminescence a solu- tion of 20 c.c. of the normal aqueous solution mixed with 1 c.c. of acid. The first effect of the acid was to bring out more distinctly the dimmest bands of the aqueous solution. There was no marked shift or change in resolution as the acid component was increased until the solution contained 1 c.c. of normal aqueous solution to 2 c.c. of acid. With further dilution, e. g., 1 c.c. of solution to 2.75 c.c. of acid, a marked change in the spectrum occurred, for only a broad-banded series of 200 FLUORESCENCE OF THE URANYL SALTS. Table 121. — Uranyl nitrate in nitric acid — Frequencies and average intervals of fluorescence bands. Frequencies in 1 c.c. of normal aqueous with 5 c.c. of acid. Band. -30° -60° -90° -120° -150° -180° 2 = { 4 I 6 { ' { 7 Av. int . 1678.4 1718.8 1762.7 1809.3 1852.2 1893.9 1936.9 1982.2 2026.3 2072.5 1676.4 1722.4 1768.6 1808*0 1854.3 1894.3 1940.4 1982.2 2028.4 2073.4 1719.1 1721.2 1721.8 1760.3 1809.0 1847.7 1896.1 1936.1 1983.3 2023.9 2073.0 1805.1 1805.1 1808.0 1890.7 1892.9 1896.1 1974.7 1981.4 1982.2 2069.5 2073.4 84.8 87.6 88.1 87.8 88.4 87.8 Uranyl nitrate in methyl alcohol — Frequencies in 0.1 c.c. normal solution in alcohol. Band. -30« -60° -90° -120° -135° -150° -180° 2 ' { '{ » { 6 1 7 ' I Av.int. 1654.3 1700.1 1730.1 1785.7 1817.2 1872.3 1904.0 1959.2 1990.1 2046.2 2076.8 1712.0 1966.6 1723.0 1784.1 1810.0 1870.9 1896.8 1958.5 1983.7 2046.2 1697.5 1730.1 1785.1 1814.9 1871.3 1900.1 1957.7 1987.7 2045.8 2075.1 1795.7 1880.4 1964.3 2046.2 83.6 87.4 87.1 86.8 Uranyl nitrate in ethyl alcohol — Frequencies in 0.1 c.c. normal solution in alcohol. Band. -90° -120° -150° -180° 3 4 5 6 ' { Av. int. 1714.1 1799,5 1886.4 1971.7 1713.5 1801.2 1886.8 1971.6 1727.7 1811.3 1897.5 1984.1 2045.5 2071.3 1806.0 1891.4 1972.8 83.4 85.8 86.1 85.9 FROZEN SOLUTIONS. 201 doublets is present. This is probably caused by a change in the hydrate at this dilution. This type of spectrum persisted through five more dilute solutions, even when the solution contained only 1 part aqueous solution to 1,000 parts acid. The effect of slowly cooling five of the acid solutions is seen in figures 101, 102, 103, 104, and 105. Very often combinations of acid and aqueous solution proved to be unstable on freezing; consequently it was difficult to reproduce at + 20" NITRATE IN NITRIC ACII >• io:i ^T-X .--T-N ^-TN -JO' -60* yn /r\ /i\ A A /\ -90* /TV A A M M ./\ -120' A\ yT\ ^A Ik 1 1 yK^TV wiiil A A -150* A /1\ .1^^ /TV/I'll A\ A .k A .1 1 1 1 1 . 1 1 1 i 1 III II 1 1 1 1 loloo I leloo 1 20t00 1 M^ .S6|y** .521^ .481 ,« Fig. 101. -o' NITRATE IN NITRIC ACID. r.2 -30' ^^ ^,,^ ^^ -45" ,,- -60- y\ yi\ A. A A A -90' y\ y\ A A A A . -120" y?t n A ^ 1 A -150° jTK y> Jh /i n 1 . h -180' 1 1 1 1 1 1 1 1 1 1 1,, , II til icIoo I loloo 1 20I00 1 .60]^ M^ .52 ^ •A9\^ FiQ. 102. 202 FLUORESCENCE OP THE URANYL SALTS. -30^ NITRATI : IN NITRIC ACID r.3.5 -60' ^-n-N ^^^ /T\ -90* /T\ /i\ A\ /]) /n\ -120" yw /7-yi\ /ivA A/l\ /^ /[Vi\ -I50' /T\ yVK /^A\ J^ /5/I /n/i\ -180° /T\ /^/^ ^li\ ^^ ^^ /^/i\ leloo 1 l9lOQ 1 20|00 1 .«!^ 56l^ sSai^* .481^. Fig. 103. -30" NITRATE IN NITRIC ACtO. I:S -eo« ^ yri /^ /\ -90'' /\ Til /[I 1^ -120" ^ ^ /!> /. A\A\ m fW /\\ -150" /^^ ^ /^ /A XTy /i^ /r\ /^ /i\ A -180" /r\ ^ /T\ y^ y i \ /^ / i \ ^ /i\ A leloo 1 isloo 1 20|00 1 .60|x, .561^. .SZl^' .481^ Fig. 104. I NITRATE IN NITRIC ACID. lOM 10 + 10* -30" -60'' -90* -150" -180" ^TN /n/n /n icIoo 20IOO -*oU .56[> Fig. 105. •S2U _,48L£_ FROZEN SOLUTIONS. 203 will the spectra of such solutions. This appeared to be somewhat independent of the rate of cooling; thus, it was discovered that a solu- tion might yield a spectrum consisting of a set of broad-banded doublets at one time and a narrower set of doublets differently spaced at other times. The three spectra shown at the bottom of figure 96 illustrate this phenomenon. The first two spectra, although entirely different, were produced from a solution of 1 part normal aqueous solution with 10 parts nitric acid. The first was obtained after very slow cooling, the second after moderately slow cooling to liquid air. The second spectrum is identical with the third, obtained by slowly cooling a solution of 1 c.c. of normal solution to 3.5 c.c. of nitric acid. Such experiments lead to the view that the luminescence spectrum is determined by the particular hydrate which is formed on freezing. The frequency interval of an acid or aqueous solution was always constant. The change in interval through a wide range of dilutions was slight. The largest interval was of 87+, the smallest of 84+ units. NrTRATE IN WATER £ ETHYL ALCOHOL. -»0* -120° -150' -i ^n "^ ^ I ~\ r~ i ^ /-r\r^-^ ^ NITRATI IN ALCOHOLfETHYL. -120* ■ISO' -zn zn i6loo I laloo I goloo .60l^ 56l^«- .52! >^ .A^VM -120' To ^ NITRATE IN ALC0H0L:METHYL. -J^ -185' j^^ ^=^ Aya y\J\ A -ISO" JL A^ A_ A^ A^ -IBS' I ■ leloo I 20I00 .60|^« .Sel^^ .52!^^ .481^ Fig. 106. 204 FLUORESCENCE OF THE URANYL SALTS. THE URANYL NITRATE IN ALCOHOL. The luminescence spectrum of the normal aqueous solution diluted with ethyl alcohol is distinctly different from that of the aqeous solu- tion. (See fig, 106.) The sharply resolved spectrum is quenched and the new bands are not in the same positions. The unfrozen solutions in a mixture of alcohol and water are not so opaque as the aqueous solutions; hence it is necessary to freeze them to produce sufficient absorption to bring out the luminescence. The first readings were taken when the temperature was —90°, and a consistent shift to the violet was effected by further reduction in temperature. The spectrum of a solution of uranyl-nitrate crystals in ethyl alcohol will also be found in figure 106. At -90°, -120°, and -150° slight change in form or wave-length occurs, but at — 185° fairly well resolved, crests protrude above the crests of the broad bands, still existent. It is probable that one series is due to the water of crystallization, the other to the alcohol. Jones and Strong^ have attributed the presence of two sets of absorption bands in the water and alcohol solutions to the presence of both a hydrate and an alcoholate, and the two lumines- cence spectra are undoubtedly caused by such a combination. CNCRGY IN CRESTS OF BANDS. NITRATE AT -90. 6 180 A 5870 k 5586 A 5324 A 5088 A 4883 k Fig. 107. A solution of uranyl nitrate in methyl alcohol (fig. 106) presented bands which in the manner of development with temperature resem- bled the aqueous bands. The doublets fall into two series of constant intervals. It will be observed in figure 106 that the bands of the alco- holic solutions are in approximately the same positions. ^ Loc, cit., p. 104. FROZEN SOLUTIONS. 205 ENERGY DISTRIBUTION IN THE BANDS OF URANYL NITRATE. The normal aqueous solution at —90° was studied with the aid of the spectrophotometer and bar, the intensity of the crests of the bands being matched by the intensity of the acetylene flame at the same wave-length. These values were multiplied by the ordinates of the corresponding wave-lengths of the energy curve for acetylene.^ Figure 107 shows the manner in which the bands differed in intensity. The envelope is of the same form as that determined by Nichols and Merritt^ for the individual bands of the crystalline salts. SUMMARY OF CHANGES. The changes produced by slowly changing the temperature from +20° to -180° include: (1) An increase in intensity of the entire spectrum. (2) A shift which is more often toward the violet than toward the red, although both shifts may occur between the above temperatures. (3) A narrowing of the bands and in some solutions a resolu- tion of the bands. (4) A slight change in the frequency interval. (5) The formation of one or more definite hydrates. (6) A change in the form of the bands. The clianges produced by dilution include: (1) A shift of the entire spectrum. (2) A change of interval. (3) A change in the hydrate. (4) A decrease in the resolution, excepting when small amounts of acid are added to an aqueous solution. (5) A decrease in intensity. CONCLUSIONS. (1) The constant-frequency intervals are due to the uranium oxide. (2) The small shifts are due to a change in the relative intensity of two or more components of a band. (3) The more remarkable changes in position are caused by the presence of a new hydrate. (4) The change in hydrate is probably often associated with a change in the crystal system, and when this phenomenon occurs a change in the grouping of the component bands occurs. The work on four double nitrates^ (Chapter VII) indicates that the crystal system is an important factor in the determination of the positions of the bands. (5) The invariable production of broad bands with extensive aqueous dilution is due to complete ionization. ^ Coblentz. Bureau of Standards, v. 7, No. 2, p. 259. 2 Nichols and Merritt. Physical Review (1), 32, p. 358. 3 Howes and Wilber. Physical Review (2), xi, p. 66. 1918. NICHOLS PLATE 1. !1 1 III B (A) A double reversal in uranyl sulphate at 185° C. {B) The fluorescence of uranyl ammonium nitrate at 185°jC. (C) The polarized fluorescence and absorption of uranyl ccesiura chloride at 185° C. Photographic reproductions from the original spectrographs by Dr. R. C. Rodgers APPENDIX 1. CHEMISTRY OF FLUORESCING URANYL SALTS. The compounds studied in this work were those uranyl compounds which showed a bright fluorescence. These in general were salts of the stronger acids and usually double salts with the alkah metals. The further general characteristics were high solubility and much water of crystallization, i. e., the more water of crystalhzation the more intense the fluorescence, as in the case of the hthium manganese acetates. The nonfluorescing compounds of lower valence or those without the "uranyl" oxygen, as well as the sodium carbonate and zinconium oxide solutions of uranic oxide, which, though having pecuUar and characteristic absorption, do not fluoresce, were not taken up. The particular groups taken up largely were the nitrates, chlorides, sulphates, and acetates, with potassium, rubidium, caesium, ammonium, and sodium in double salts. The phosphates, fluorides, oxalates, and tartrates and some double salts with the bivalent elements were studied in some cases. The material for use in this investigation was obtained at first from Kahl- baum. Later a number (25) of compounds were prepared by G. 0. Cragwall in the Chemical Laboratory of Cornell University. The remainder were pre- pared by the authors. CragwalFs material was a large quantity of uranium residue originally from Kahlbaum, during the purification of which by con- version to ammonium diuranate and hen to the chloride the first ammo- nium uranyl chloride crystals with the resolved spectrum were observed. The material used by the author was chiefly uranyl nitrate hexahydrate purchased as chemically pure, but which was found to contain noticeable quantities of sodium nitrate crystals. This was dissolved in water, precipi- tated with ammonium hydroxide, washed by decantation to incipient suspen- sion, to which HCl was added until nearly all the precipitate was dissolved. This leaves most of the iron in suspension if present in small quantities and was used to separate out iron in reworking material contaminated from spatulas, etc. On boiUng this solution, if much iron is present it further coagulates and can be filtered, but if the acid concentration is low enough, quite a portion of the uranium separates as H2UO4. The chloride solution was precipitated again with ammonia, washed, and redissolved in nitric acid. The nitrate was evaporated until the salt (trihydrate) began to crystallize out, and was then carefully heated until decomposition took place, with the formation of the red uranic oxide. Care was taken not to form the black uranous uranic oxide UsOs by overheating. The red oxide containing some undecomposed nitrate was digested with water, which converted the oxide into the hydroxide, or acid H2tJ04, a bright yellow powder. This was washed free from nitrate by decantation and air-dried and formed the major part of the material used. Some stock uranyl acetate was used, but as this usually contains sodium acetate also, it is not advisable when preparing sodium-free salts to be com- pared with sodium triple salts. Some material was also precipitated as the oxalate, but this does not give complete precipitation, and as it gives the black UsOs on ignition, which is not as readily soluble, it is not of much value except for preparing oxalates. 207 208 FLUORESCENCE OF THE URANYL SALTS. For making triple sodium acetates, some material was precipitated as the sodium uranate, dissolved in sodium carbonate, and the solution treated with acetic acid, from which the sodium uranyl acetate crystalHzes, leaving the sodiimi acetate in solution with very Uttle waste uranyl salt. NITRATES. URANYL NITRATE. This is the commonest and best known of the uranyl salts, crystalUzing ordinarily as the hexahydrate, which readily forms large, clear crystals by cooUng or evaporation. It is prepared by dissolving either the uranic oxide or hydroxide H2UO4 or the uranous xu-anic oxide UaOs in nitric acid and crystal- Uzing. This salt shows strikingly the property of most of the uranyl salts of strong acid, of dissolving noticeable amounts of the oxide in the neutral solution, so that a clear solution may be strongly basic. This oxide pre- cipitates on heating or evaporation. Uranyl Nitrate Hexahydrate. U02(N03)26H20. The complete description^ of the crystal properties of this hydrate are given in Groth's Chemische Krystallographie, II, page 142. System rhombic; axial ratio a:b:c = 0.8737: 1: 0.6088. Forms b (010), a (100), making short rectangular prisms with pyramidal ends formed by h (111), usually cut also by q (Oil), making a a six-sided face and b eight- sided. The prism (110) was observed on one crystal which was deformed by growing near another. Specific gravity, according to Boedeker (1860), is 2.807. No statement is made as to cleavage, but it was found that very sHght temperature changes produce a spontaneous cleavage, generally along q (Oil), so that the crystals can not be handled on a cold day and immersion in Uquid air completely powders them. The optical properties are double refraction +, plane of axes b (010), acute bisectrix the c axis, apparent angle of optic axes 67° to 69°, mean index j8 1.495 to 1.502. The pleochroism, according to Schabus, gives bright yel- low-green parallel to a, b greenish yeUow, c deep citron yellow. The fluorescence and absorption were investigated by Stokes^, E. Becquerel, Hagenback,^ and H. Becquerel.* The tribo-luminescence was noticed by Herschel.^ Wasiljew^ gives the melting-point of the hexahydrate as 60.2° C. and gives the solubiUty curve for the hexahydrate in water. SiUiman^ gave 1 de la Provostage, Ann. der Chim. Phys. (3), 5, 48. 1842. Schabus, Preiachr. Wien (1855), 40. ' Sitz. berichte d. A. d. W. Wien, 27, 41. 1857. Rammelsberg, Neuat. Fortsch. in de Kryat. Chem. Leipzig, 58. 1857. Lang, Sitz. ber. Wien, 31, 120. 1858. des Cloiseau, Annales dea Mines (5), 14, 348. 1858. Quercigh, Riv. Min. crist. Ital., 4, 6-14. 1915. 2 Stokes, Phil. Trana., 142, 517, 520. 1852. 3 Hagenback, Poggendorff's Annalen, 146, 395. 1872. * H. Becquerel, Ann. Chim. Phys. (6), 14, 230. 1888. 5 Herschel, Nature, 60, 29. 1899. ^Wasiljew, Chem. Zentralblatt 14, 2, ii, 1527. 1910. Jour. Russ. Phys. Chem. Ges. 42, 577. 1910. 'Silliman, Amer. Jour. Science (2), 27, 14. 1859. CHEMISTRY OF FLUORESCING URANYL SALTS. 209 59.5° C. which is probably not as accurate. Lowenstein^ gives the vapor- pressxire of the saturated solution as 18 mm. approximately at 25° and the pressure of the equilibrium between hexahydrate and trihydrate as over 4 mm. The author found the two hydrates to be stable together at 5 mm. at 20°. The result is that the crystals always effloresce and fall to a yellow powder if left in the air, in the winter especially if the sun falls on them, and may dehquesce in the summer. Lescouer^ gives the vapor-pressure of the solution at 6° as 12 mm. and for the trihydrate below 3 nun. The author found that the hexahydrate dissolved in various concentrations of nitric acid at 20° C. in the following ratio: 1.6 grams of hexahydrate in 1 gram of 10 per cent HNO3, 1.15 grams in 20 per cent, 0.8 gram in 30 per cent, 0.65 gram in 40 per cent to 70 per cent HNO3. The values have not been determined accurately above 40 per cent on account of the compHcations due to the occasional formation of the trihydrate. These crystals were usually grown by evaporation in the room. For work on the polarization they were grown in thin plates tabular on a or b by putting small seed crystal in a solution of the depth desired for the thickness of the crystal, in the position desired. Uranyl Nitrate Trihydrate. U02(N03)23H20. This hydrate is mentioned by Lescouer^ and by Ditte^ as being formed when the hexahydrate is heated to boihng. Drenkman^ and Schultz-Sellack^ found that on adding the hexahydrate to strong nitric acid and crystalhzing by cool- ing or evaporation the trihydrate was obtained. Lebeau^ also obtained it by heating the hexahydrate on the water-bath or by evaporating the nitric acid solution in a dessicator over H2SO4 or KOH. Marketos^ mentions it as formed directly from the hexahydrate over sulphuric acid in a dessicator, as does also Forcrand.^ As can be deduced from the vapor-pressure data of Lowenstein and the author, this air-drying takes place as soon as the vapor- pressure of the water in the atmosphere goes below 5 mm. The best crystals are obtained by slow evaporation of the solution of the hexahydrate, dried on the water-bath in concentrated nitric acid in a dissicator over sulphuric acid and caustic potash or quicklime. The crystalUne form was measured by G. Wyrouboff,^*' who obtained his crystals by evaporating the neutral solution at 65°. System triclinic; axial ratio, a :h :c= 1.7753: 1 : 1. 4104. a 85°35'; /3_94°12'; 7 81°44'. Forms p (001), making plates with h' (100), a' (101), and a* (201) on the edges, and c* (111) and 6* (111) obhque-angled ends. The specific gravity was found to be 3.345, No cleavage has been noticed, although the crystals are likely to form with irregular cracks across or radiat- 1 Lowenstein, Zeit. Anorg. Chem. 63, 105-107. 1909. ^Lescouer, Ann. Chim. Phys. (7), 7, 429. 1896. * Ditte, Ann. Chim. Phya. (5), 18, 337. 1879. Compt. Rend. 89, 643. 1879. * Drenkman, Jahrsber der Fortschritt Chem., 256. 1861. « Schultz-SeUack, Jahrsber. Fort. Chem., 365. 1870; Zeit. fur Chem., 646. 1870. ' Lebeau, BuU. Soc. Chim. (4), 9, 299. 1911. 8 Marketos, Comptes Rendus, 155, 210. 1912. * Forcrand, Comptea Rendus, 156, 1044, 1207, 1954. 1913. ^° Wyrouboff. Bull. Soc. fran. Mineral, 32, 340. 1909. 210 FLUORESC3CNCE OF THE URANYL SALTS. ing from the seed. Schultz-Sellack gives the melting-point as 120° and Wasiljew as 121.5° C. It is really only a partial melting-point, as the dihy- drate is not completely soluble in the resulting solution. The solubility of the trihydrate in water above 60° or in nitric acid has not been determined, although Ditte gives a solubility of 14.39 parts of the trihydrate in mono- hydrated (91 per cent) nitric acid. Uranyl Nitrate Dihydrate. U02(N03)22H20. Ordway describes the dihydrate as resulting by boiling off the fused hexahydrate, which Lowenstein confirms. The latter finds it as the product of 6 days' dihydration over sulphuric acid of over 80 per cent strength, al- though Fourand finds 6 days required in a vacuum over strong sulphuric. Lebeau finds powdered hexahydrate converted to dihydrate in a vacuunj desic- cator with concentrated sulphuric acid in 72 hours. Lebeau finds that on treating the hexahydrate with ether, two layers are formed, of which the ethereal layer can be dried with anhydrous calcium nitrate, which leaves the dihydrate on evaporation . It is to he noted in this connection that the ethereal solution, which is also used for separating uranium X, can* not he boiled off, as it decomposes with explosive violence after some heating, liberating copious nitrous fumes.^ Lebeau also obtains crystals with ether of crystallization at 10° and —70°. The dihydrate may also be formed by adding dry UO3 to fuming nitric acid (92 per cent), from which solution it is readily recrystal- lized. Wasiljew, crystallizing the dihydrate from fuming nitric acid (s. g. 1.502), finds quadratic tables of the rhombic system with strong fluorescence. The author found yellow plates with marked fluorescence at lower tfirrjpera- tures of probably rhombical pinaeoid and pyramid, with some other forms. The crystals weather so rapidly, having a vapor-pressure of 0.2 mm., accord- ing to Lowenstein, that changing from one closfid vessel to another usually tarnishes them so that little can be done in the way of measuring, liandling for cleavage, etc. Wasiljew gives the melting-point as 179. 'j''. Thf; mixture of dihydrate and solution obtained by rrjelting the trihydrate goe-s over to solu- tion at about 160° and then goes unchanged except for slight }>oiling to 240°. Uranyl Niitiate Anhydrous. V02(K(hh or UOa.NzOr,. Marketos produced anhydrous uranyl nitrate by heating the nitrate to 170° to 180° C, since total decomposition took place at 200", and passing over it dry carbon dioxide saturated with nitrie-aeid vapors by huhfjling through concentrated nitric and sulphuric acids. This produced a yellow amorphous salt soluble in water which decomposed ether, with the liberation of nitrous vapors. Forerand found thiat long heating above 120"-' C. in a current of dry carbon dioxide produced basic anhydrous nitrate and below 100° only a monohydrate. Twelve hours at lOo In a current of carbon dioxide cliarged with nitrif^acid vapors gave VO'JlsO-.'jz + l/31H2L'04. The method evolved for producing anhydrous uranyl nitrate was to place in a train of U-tubes a tube containing uranic oxide made by fieating the hydroxide or acid H2UO4 until it began to turn red and di>.tilling over it nitric » Mullcr. Chcrrj. ZtK., 41, 4-39, 1917; 40, 30, 1916. CHEMISTRY OF FLUORESCING URANYL SALTS. 211 anhydride, N2O6. This was accomplished by having a reaction flask fitted to the system by a ground-glass joint, in which were placed phosphorus pentoxide and fiuning (92 per cent) nitric acid in calculated amounts. From this, on heating to 50°, the NaOe distilled out and was condensed by freezing mixture in the first U-tube, which served as a reservoir. When this was filled with solid N2O6 and a two-hquid layer of N2O5 and HNO3, the flask was removed, the joint covered by a cap, and the anhydride distilled over on the uranic oxide by placing the reservoir tube in a bath of hot oil. No re- action took place between the oxide and the acid until the oxide tube was in turn put in the oil-bath and the anhydride boiled off into the last tube, which served as a second reservoir for the acid. As soon as the acid began to boil the reaction took place, producing a vivid green fluorescence and a fight yel- low color instead of the reddish oxide. The anhydride could be distilled off and run back over while holding the tube with the uranyl nitrate at any temperature. Also, any acid which did not sofidify could be poured off and the remaining N2O5 run back over the nitrate, insuring absolute freedom from water. The resulting compound was found to be stable up to 180°, at which temperature it broke up into N2O5 and UO3, which could be recombined if the temperature was lowered. Distilfing the acid on and off was performed several times with one specimen, examining the spectra each time, which showed first the anhydrous salt fluorescence and then none for the oxide. DOUBLE NITRATES. Meyer and WendeP prepared double salts of ammoniimi, potassiimi, rubi- dium, caesimn nitrates with uranyl nitrate. These crystals were described by Steinmetz.^ They were grown from a solution in nitric acid and were of the type KU02(N03)3- Rimbach^ endeavored to determine the solubility of these salts in water at various temperatures. He found large crystals in the am- monium and potassium solutions unhke those of Meyer and Wendel which were measured by Sachs^ and assigned formulae like those of Meyer and Wen- del, but since they were alike were called isomorphous and the a forms of NH4 and KU02(N03)3. Examination of the spectra of these forms in the labora- tory indicated and Sachs's data itself proves that the a form is simply uranyl nitrate hexahydrate. In attempting to grow crystals according to Rimbach^s method which would not be uranyl nitrate, however, two new forms were discovered containing two molecules of alkali nitrate to one of uranyl nitrate. In the process of growing the potassium salt for experimental purposes, still a third was found, but so rarely that it was not studied. In order to find out the conditions under which the various salts were formed, a series of solubiUty determinations were undertaken, being run at constant temperature of 20° C. in a thermostat, with varying percentages of aqueous nitric acid as a solvent. From these incomplete results it will be seen that from solutions of less than 30 per cent nitric acid and less than 1 molecule of uranyl nitrate to 1 of potassium nitrate, potassium nitrate only will crystalhze; that in a 1 to 1 1 Meyer and Wendel, Ber. d. d. Ch. Ges., 36, 4055. 1903. 2 Groth's Chem. Kryst, 11, 150. SRimbach. Ber. d. d. Ch. Gea., 37, 472. 1904. * Sachs, Zeit. f. Krys., 38, 497. 1904. 212 FLUORESCENCE OP THE URANYL SALTS. solution above 40 per cent the so-called 7 form or monopotassium salt will appear. In the 1 of uranyl nitrate to 2 of potassium nitrate, the double nitrate crystaUizes only above 50 per cent of nitric acid, and as the 8 phase or the dipotassium salt and the metastable phase at this concentration is the 7 form. Presumably at higher concentrations of potassium nitrate the last- found and imdetermined form would appear. Grams of solute in 100 grams of solvent. Solvent (p. ct. HNO3). KN08.U02CN03)2. 2KN03.U02(N08)2. KNOj. 85.5 82.0 99.8 89.0 81.3 54.0 33.9 Solid phase. KNOs KNOa KNOs Hex y y y ■Solid pAose. 62.9 52.2 45.7 51.8 67.2 52.3 Solid phase. KNOa KNO3 KNOs KNO3 KNOs 5 Solid phase. 31.4 19.1 14.5 11.4 15.2 18.6 19.6 29,6 34.2 48.8 10 20 30 89.5 y 40 50 57.5 y 60 70 80 .... 90 Solvent (p. ct. HNOa). NH4N03.U02(N08)2. 2NH4N03.U02(NOa)2. NH4.NOS 165 128.6 80.3 68.2 60.5 61.4 38.6 Solid phase. Hex. Hex. Hex. Hex. Hex. Solid phase. 251 201 150 144 98.2 58.0 35.6 Solid phase. Hex. Hex. Hex. Hex. /3 380 380 215 150 Solid phase. /3 191 151 127 104.8 86.4 76.5 10 20 144.5 144 95.6 /3 30 40 50 60 Solid phases appearing are: Potassium nitrate, KNO3. Ammonium nitrate, NH4NO3. Uranyl nitrate hexahydrate, U02(N03)2 6H2O, Hex. Monopotassium uranyl nitrate, KU02(N03)3, y. Dipotassium uranyl nitrate, K2U02(N03)4. S. Monoammonium uranyl nitrate, NH4U02(N03)3» ^. From solutions 1 molecule of uranyl nitrate to 1 of ammonium nitrate, uranyl nitrate hexahydrate crystaUizes xmless the per cent of nitric acid is at least 50, above which the j8 or monoammonium form crystaUizes, which is metastable practically to water solution. From 2 molecules of ammonium nitrate to 1 of uranyl the hexahydrate crystaUizes up to 40 per cent, above which the j8 form appears, which is also metastable to pure aqueous solution. It wUl be noted that while potasisum nitrate is about as soluble as uranyl nitrate and the phase in equilibrium with the more acid solutions corresponds to the composition of the solution, the ammonium nitrate is much more solu- ble and not only does not form the soUd phase in case uranium is present, but does not form the diammonium salt from solutions of that composition. Laboratory experience showed that a large excess of ammonium nitrate and CHEMISTRY OF FLUORESCING URANYL SALTS. 213 rather low acid concentration was necessary to produce this form. These higher ratios of ammonium nitrate should be investigated. These results do not check well with those of Rimbach, who presumably- crystallized considerable portions of the salt, instead of determining the phase with which the solution was in equihbrium by the addition of seeds of known phases. They do not, however, materially conflict with those of EngeP in the case of the solubiUty of potassium nitrate in nitric acid, where the solubihty is found greater at 20° than at 0° in dilute solutions and less in highly acid solution. The mono or acid forms of the double potassium ammoniimi salts and the corresponding rubidium and caesium salts were found to be as described by Sacks from the preparation of Meyer and Wendel. There is no indication that the corresponding double salts containing 2 atoms of rubidium or caesium could not be produced by using solutions similar to that used for the dipotas- siimi salt. Other double uranyl nitrates with other bases than the four alkalies dis- cussed do not seem to form, with the exception of thaUimn, which is reported by Meyer and Wendel as forming but being non-fluorescent, as the double thallous sulphate is. Sodium nitrate crystalhzes side by side with the hexa- hydrate or trihydrate, according to the acidity of the solution, but no con- ditions were found under which the two salts would crystallize together. Silver, cadmium, zinc, calcium, bariiun, and magnesium were also tried with- out success, although a modification of the hexahydrate spectrum was pro- duced by the calcixmi and magnesium. Meyer and Wendel also tried hthium, sodiiun, and the bivalent metals without formation of double salts. MoNOPOTASsiuM Urantl Nitrate. (y form) KUOaCNOa)^. These crystals were prepared by Meyer and Wendel by crystaUizing potas- sium nitrate and uranyl nitrate in equal proportions from a nitric-acid solu- tion. The crystals were examined by Steinmetz. System rhombic; axial ratio 0.8541 : 1 : 0.6792. Thick tabular combinations of c (001), m (110), with subordinate forms of b (010), s (102), (111), sometimes a (100), and rarely a (Oil) and 122. Steinmetz and Sykes report good cleavage on &, and good cleavage was also observed on a. The specific gravity was found to be 3.503. Crystals of this form are stable at 20° if the partial pressure of the water-vapor is not over 9 mm. Hg, but at that point begin to dehquesce, changing to a whitish yel- low chalky mass. On heating the crystals, yellow crusts begin to form on the crystals at 150°, violent decrepitation begins at 200°, and decomposition with hberation of nitrous fumes at 270° C. According to Steinmetz, the plane of the optical axes is c (001), acute bisectric a. Axes visible through (110). The best crystals were obtained by cooHng of hot solutions supersaturated 2 grams in 50 c.c. in glass-stoppered bottles. The composition as determined by Meyer and Wendel was KU02(N03)3. An ignition run to check this gave 65.13 and 64.82 per cent, the theoretical form K2U2O7 being 67.29, the low values being due to loss by decrepitation. 1 Engel, Comptes Rendus, 104, 913. 1887. 214 FLUORESCENCE OF THE URANYL SALTS. DiPOTASsiuM Urantl Nitrate. (5 form) K2U02(N03)4. These crystals appeared, after a year of effort to obtain crystals from neutral solution which were not hexahydrate, in a slightly acid solution containing an excess of potassium nitrate. It shows marked reluctance to appear and does not grow well if the room temperature is below 20° C, the hexahydrate forming instead, but above that temperature gives fine crystals, especially if seeded, although it does not grow as rapidly as the other members of the group. System monocUnic; axial ratio a :h: c = 0.6394 : 1 : 0.6190; i3 = 90=t=. calc. ohs. calc. ob8. c:a =001 : 100 = 90° 0' O'.ir =133 : 232 =19° 19' 20° 44' p:p'=331 : 331 = 63° 30' o:q =133 131=52° 3' 61° 50' c:p =001 :331 = 77° 38' o:d =133 101 =45° 38' 47° 36' c:d =001 : 101 =50° 1' 52° 62' piir =331 : 232 =23° 39' 22° 49' c:o =001 133 =42° 43' 43° 10' p:d =331 : 101 =38° 42' 38° 26' c:q =001 : 131 =59° 22' 60° 39' p:q =331 : 131 =43° 37' 42° 12' o:o =133 : 133 =73° 58' 73° 15' d:q =101 : 131 =56° 31' 66° 17' o:p =133 : 331 =42° 58' 42° 48' d'.ir =101 : 232 =37° 6' 38° 66' o:p' = 133 : 331 =84° 19' 84° 0' d:m=101 : 230 =61° 7' 62° 48' These axes are probably not those of the space lattice, being taken from the first habitus observed, which formed in a solution having barely enough potassium nitrate to produce this phase, and consisted of e (001), o (133), and p (331) meeting in a point in front which was sometimes cut off by a (100), usually accompanied by d (101). tt sometimes occurred between o (133) and p (331) and q (131) between o 133 and p 311; 6 (010) and m (230) were found in measurement. One crystal showed c, o, p, d, and d' (101), a and probably q and (lU). Most of the crystals grown later had a prismatic or needle habitus in which p (331) was the predominant form, with small e faces on the ends and occasionally some of the other forms. All faces on these crystals gave reflections which appeared in the goniometer as a flattened figure 8, and best agreements were found in the angles taken from the outside of every pair of readings. In case the whole figure did not appear, results were unsatisfactory. On dissolving a large crystal in the mother-liquor by heat, c (001) was un- touched, d (101) was left even and a little pitted, and the edge between p (331) and (133) was rapidly dissolved, leaving the deepest etching where these met at d (101). No conspicuous cleavage was noticed. The specific gravity was found to be 3.359. On heating crystals of this phase, they first decrepitate to an opaque yellow powder at 200° C., which, at 260° C, flows together. Above this temperature decomposition sets in, with the evolution of nitric fumes. Various colored masses result, deep red UaOs, bright red UO3, bright yellow K2UO4, and on cooling a beautiful rose pink pervades the mass. This phase does not change over concentrated sulphuric acid, i, e., has zero vapor-pressure at 20° C, but over acid corresponding to 11 mm. of mercury of partial pressure of water-vapor it turns whitish without becoming moist, probably due to the formation of KNO3 and U02(N03)2 6H2O. This is the same pressure at which the diammonium salt deliquesces. The refractive index of dipotassium uranyl nitrate given by the faces (001) and d (101) were found to be 1.5422 for light vibrating parallel to the b axis CHEMISTRY OF FLUORESCING URANYL SALTS. 215 and 1.5349 for light vibrating in the ac plane 26° 34 J' from a toward c in the *' acute" angle j8. The composition was investigated by igniting to K2UO4 and by washing out the K2SO4 from sulphuric-acid solution precipitated with NH4OH and weighing the resulting U3O8 and K2SO4. Theoretical. First. Second. K2U04 . . K2S04. . . U308.... p. ct. 63.78 47.09 29.19 p. ct. 64.61 46.85 30.36 V-ct. 63.52 47.03 29.01 The best crystals were obtained by cooling solutions supersaturated 1 gram in 100 c.c. MONOAMMONTUM UrANYL NiTRATE. /3 fonn NH4U02(N03)3. This salt forms from solutions containing uranyl nitrate and ammonium nitrate at room temperature if the acid-content is high and from water at higher temperatures. System trigonal; axial ratio a :c = 1: 1.0027 (a 97°6'). The forms are prismatic combinations of prism a (1120) with the rhombohedron r (1101) on the end, on the edges of which occur s (1012). Theory. Steimnetz. W. r:r = 1101 TlOll ... s:s=0il2 :il02 ... 81° 47' 51° 19' 81° 54' 51° 30' 82° 8' 51° 36' The column headed Theory gives the values for a substance having an axial ratio a : c = 1 : 1, the close approximation to which makes this a remarkable case and probably indicates something concerning the structure. Twins were observed in which the contact plane was s 1 102 and the twin- ning axis, the axis of reference to which s was parallel, making the angle be- tween the two unique axes 119°54' and giving the crystal the appearance of a flat hemimorphic orthorhombic crystal. The angle between the two r (iTOl) faces was calculated to be 21°32' and found to be 21°=^. Crystals of this form are stable in dry air, but begin to deliquesce and change to a light yeUow chalky mass if the vapor-pressure of water is above 9 mm. Hg. Intense pleochroism was observed in crystals about 0.01 mm. thick on a microscope sHde, when they occurred lying on a prism face, the hght vibrat- ing parallel to the unique axis appearing white, f. e., less yellow than the mother-liquor in which the crystal lay, while that ordinary ray appeared a deep yellow. The monopotassium and monoammonium crystals separate from the same solution, there being no tendency to form mixed crystals. The composition of the crystals was checked as being the same as that given by Meyer and Wendel by igniting to the oxide. The'ory, 59.22 per cent found, 58.97, and 58.86. 216 FLUORESCENCE OF THE URANYL SALTS. DiAMMONIUM UbANTL NiTBATB. a form (NH4)2U02(N08)42H20. This phase crystallizes from slightly acid solutions of uranyl nitrate con- taining a large excess of ammonium nitrate. Its solubility increases very rapidly with rising temperatxire, until at about 60° C. the uranyl nitrate dis- solves out, leaving a residue of ammonimn nitrate. This salt crystallizes very readily in large sulphur-yellow perfect crystals of 5 or 10 grams from a volimie of solution that only gives 1 or 2 grams of monoammonium salt and crystallizes best if the room temperature is 10° to 15°. The fluorescence is very faint at 20° C, but increases rapidly below 0°, becoming stronger than that of the monoammonium salt at hquid-air temperatures. System monocUnic axial; ratio a:h:c =^ 0.8419: 1 : 0.5594; /3 94°55', The habitus resembles that of a cube with octahedron, the forms being c (001), a (100), & (010), and o (111), The faces p (110) were also occasionally observed, and possibly (211). No cleavage was observed, thus increasing the resemblance to sulphur crystals. However, on heating rapidly, the crystals fill with cracks, and con- sequently seeded crystals often have a large single crack more or less parallel to a (100). Etch figures produced by resolution in a crystal in the mother-liquor were found once, the distinct forms being on b (010), with two rounded faces meet- ing in a line in the bottom as though a lens had been pressed in. The bottom edges were all parallel and about haKway between the edges of h (010) and o (111), z. e., parallel to (102). The specific gravity was found to be 2.777. When these crystals are heated slowly they break down at about 140°, giving a pasty mass full of bubbles, which clears up shghtly at 220°, but does not give a clear solution below 240°. On cooling, bright green crystals of the monoammoniimi salt form in a background of white ammoniimi nitrate. Crystals of this phase placed over sulphuric acid with a pressure of water- vapor of 4 mm. started to lose water, turning to a whitish powder, and con- tinued to do so slowly at 5 mm., although then they do not start. Placed over sulphuric acid with a vapor-tension of 11 mm., they deliquesce rapidly, soon going completely into solution. The refractive index was observed in two different directions, using natural faces. First, between 111 and 111 giving the light vibrating in the ac plane nearly parallel to the edge olllrolll, with an index of 1.546, and that at right angles to the ac plane as 1.639; between the faces h 010 and o 111, giving for_Ught appearing on b to vibrate nearly parallel to edge between b (010) and o' 111 as 1.508, and at right angles to that 1.619. The composition was determined by ignition to U3O8, which gave 47.68 and 47.62 per cent against theoretical 47.58. URANYL CHLORIDE. Neither the monohydrate UO2CI2H2O described by de Coninck^ as being formed by evaporating the solution prepared by precipitating the sulphate solution nor the trihydrate, which, according to Myhus and Dietz,^ forms from 1 de Coninck. Comptes Rendus, 148, 1769. 1909. 2 Mylius and Dietz, Ber. d. d. Ch. Ges., 34, 2774. 1901. CHEMISTRY OF FLUORESCING URANYL SALTS. 217 the evaporation of the solution of the oxide in hydrochloric acid, were suc- cessfully prepared and freed from the sirupy mother-liquor so as to give good fluorescence spectra. Double Chlorides. The alkah double chlorides, as was discovered early in this investigation in the case of the ammonium salt, give resolved spectra at room temperature. This makes the group quite important. The four double salts of ammonium, potassium, rubidium, and caesium were prepared. Attempts were made to prepare the double salts with silver, cadmium, zinc, and calcium, and also hydrazine and hydroxylamine, but resulted in each case in the formation of the crystals of the chloride added, in a sirup or mat of the uranyl chloride. The silver was sealed in a tube with a strong HCl solution as solvent, but although remaining white did not dissolve and recrystallize. The tetramethyl and tetraethyl ammonium chlorides described by Rimbach were not made. The alkali double chlorides in general were grown by evaporation in a des- iccator in presence of an excess of HCl, which is necessary to prevent hydrol- ysis of the uranyl chloride. This forces back the solubility of the alkali salt, so that the solutions usually contain an excess of uranyl chloride. The crystals, if allowed to stand in the open air, readily give off acid, turning the color of any indicator paper on which they are placed and in moist atmosphere deliquesce readily, the sirupy uranyl chloride running away from a skeleton of alkali chloride. Potassium Uranyl Chloride. K2-U02C14-2H20. This salt was described by de la Provostage^ as occurring in hexagonal tables on c (001) bounded by o (111), w (111), m (110), m (110), h (010), q (021). The crystals measured by Rammelsberg were mostly prismatic along the a axis. The first type were those used in this work, although crystals tabular on h (010) were fairly frequent. System triclinic ; axial ratio, a:h\c\ = 0.607 : 1 : 560. a = 80°41' ; /? = 77''42' ; 7 = 91°18'. K2UO2CI4.2H2O. (NH4)2U02Cl4.2H20. Calculated. de la Provostage. Rammelsberg. Grailich. W. m :6=(110):(010)... H :6=(1T0):(0T0)... M :c = (110):(001)... 6 :c = (010):(001)... q :c = (012):(001)... : c = (111): (001) .. 60° 28' 61° 33' 83° 48' 99° 15' 55° 21' 60° 30' 61° 10' 83° 55' 60° 52' 60°30'± 61° 0' 60° 48' 85° 26' 99° 34' 56° 22' 59° 40' 47° 4' 83° 0' 98° 53' 55° 45' 55° 30' 60° 15' 46° 55' 81° 0' 75° 20' 46° 5' 66° 45' w':c = (ni):(001)... :b= ail): (010)... 80° 0' 75° 30' w': 6= (Til) :(010) ,. . «j':g= (111): (012)... o:^'=(lll):aiO)... m: c = (110) :(001)... 46° 17' 66° 33' 76° 8' The properties of the crystals were similar to those of the ammonium salt, although the crystals seemed to grow larger more readily. 1 de la Provostage, Ann. Chim. Phys. (3), 6, 165. 1842. 218 FLUORESCENCE OF THE URANYL SALTS. In the work of Jones and Strong/ it was found that the absorption bands persist far out into the red, only the intensity decreases with such rapidity that great depths of solution were required to show them. To try this out in the case of the resolved spectra of the chlorides, thick layers were built up of several crystals and found out to hold. A very deep crystal was grown in a glass tube ground into the bottom of an inverted bottle-neck which held the solution. This crystal, which was 3 cm. thick, was never tried. Another investigation which was never finished was that of the char- acter of the spectra of mixed ^ crystals, of which potassium ammonium salt K'NH4-U02*Cl4-2H20 was prepared as an example. Ammonium Urantl Chloride. (NH4)2-U02-Cl4-2H20. The crystal forms of this salt are practically identical with that of the potassium salt, as shown by the table of angles under that salt. Intense pleochroism is noticed in this crystal when viewed through the c (001) face if the crystal is about 1 mm. thick. The hght vibrating parallel to the b (010) edge of the face, i. e., parallel to the axis, is so httle absorbed as to appear whitQ and is also least refracted. The hght vibrating nearly parallel to the h axis is strongly absorbed in the blue-violet and appears deep yellow even in these crystals. The refractive indices parallel to a and nearly parallel to h and c were measured on prisms cut so that hght traveled parallel to the c face and at right angles to it. It happens that the letters of the refractive indices correspond to the axes to which they are nearest. u 6 c X720. X580. X500. 1.564-1.566 1.566-1.574 1.576-1.581 1.619 1.633 1.622 1.637 1.650 The absorption is so great parallel to h that the value for X 500 could not be obtained. Rubidium Urantl Chloride. Rb2-U02-Cl4-2H20. This is similar to the potassium and ammonium salts; although no measure- ments were taken, the crystals could not be distinguished, except by knowing the individual crystals. CESIUM Urai^yl Chloride. CS2UO2CI4. The salt was crystalhzed as above from a solution containing csesium chlo- ride and uranyl chloride and presented a distinctly different appearance from the other members of the group. This is accounted for by the composition, which, according to Rimback, Wells and Boltwood,^ is the anhydrous chlo- ride instead of containing 2 molecules of water, as with the other salts. The crystals were elongated rhombs of yellow color, showing less fluorescence than the other salts. Under the polarizing microscope they showed a striking ^ Jones and Strong, Carnegie Inst, Wash. Pub. No. 130, 90. 2 WeUs and Boltwood, Zeit. Anorg. Chem., 10, 181. 1895. CHEMISTRY OF FLUORESCING URANYL SALTS. 219 resemblance to gypsum, possessing the "fish-tail" twins and approximately the same angles and appearing different only in the yellow absorption. The crystals were so universally twinned that the interfacial angles could not be determined certainly. The largest face was b (010), with the two prism faces m (110) and m (110), and, as determined by measurement, practically all the end faces were d (Oil), r (031), s (031), although these usually appeared twice on a crystal, and other faces indicated by the measurement were q 111, x (111). System triclinic. 6:m = 010:110 = 49° 7' 6: /i = 010: 110 = 50° 50' b:d = 010:011=40°31' fc:r = 010:031=67°48' 6:8 = 010:021 = 57° 41' m:g = 110:m=43°41' m:r = 110:031 = 86°49' m:7- = 110:031=44° 0' /i:a: = 110:lll = 75°24' The refractive index was determined through the faces h (010) and ix (110), the more deviated ray vibrating at an angle of about 15° from the prism edge toward the a axis. X720 1.618 1.692 X580 1.625 1.695 X500 1.634 1.714 The index was also determined through the faces h (010) and d (Oil), the less-deviated ray vibrating parallel to this prism edge. X580 X580 1.614 1.622 1.691 1.698 Urantl Sulphate. UOaSOi-BHaO. This salt was prepared by Cragwall, presumably as the trihydrate. On recrystallization of Kahlbaum material two forms appeared, a yellow opaque needle-mass tending to replace the bright green fluorescent grains. The yellow needles became more numerous on adding sulphuric acid, so uranic oxide was added to saturation, which proved to be in excess on evaporation. The bright green fluorescent crystals were difficult to keep, as they dried out readily in the air. Microscopic examination of the Cragwall preparation showed needles with parallel extinction and greater absorption and greater index the long way of the crystals, while the angle of the optical axes was very large and the sign positive. The acid sulphate, H2U02(S04)25420, was reported by Wyrouboff.^ DOUBLE URANYL SULPHATES. The double sulphates differ from the previous double salts in the occurrence of the sodium salt. The potassimn, ammonium, rubidium, caesium, and thallous salts were also prepared by Cragwall in the form of powders result- ing from rapid precipitation by cooling. The potassium and rubidium salts especially showed very strong fluorescence, the sodium and ammonium good fluorescence, the caesium less, and the thalHum practically none, and that was not resolved. Rimbach^ describes a dipotassium salt which was not pre- pared and one of hydroxylamine also. ^Wyrouboff, Bull. Soc. fran. min. No. 32,351, 1909. 2Rimbach, I. c, 479. 220 FLUORESCENCE OF THE URANYL SALTS. These salts were all prepared by crystallization from water of the calculated quantities of the two single salts. Rimbach^ describes the potassium ammo- nium and rubidium salts as having each 2 molecules of water of crystaUization. On the other hand, de Coninck^ describes the sodium potassium and caesium salts as having 3 molecules of water, while the ammonium commonly has 2, but can be made by special conditions with 3 molecules. Potassium Uranyl Sulphate. K2U02(S04)22H20. This salt separates readily on cooling a hot saturated solution of the two salts in equimolecular proportion. This, according to Rimbach, gives the dihydrate, according to de Coninck^ the trihydrate. The salt prepared in this way is a fine crystalline powder, larger masses being clusters of crystals. A good crystal was discovered in an old solution of known strength. They were then obtained by supersaturating 0.1 to 0.5 gram of salt in 50 to 200 c.c. of solution, seeding, and allowing the tightly stoppered solution to stand from 3 to 6 months, especially in the fall, when the room temperature gradually decreases. The crystals tend to form rosettes, clusters of crystals arising from the middle of the basal pinacoid. The smaller crystals are tabular on the base; the larger ones have large prism faces made up really of repeated pyramids. These are capped by the unit pyramid with brachydome and basal pinacoid. System rhombic; axial ratio a:b: c== 0.5889: 1: 0.6253. calc. oha. calc. oba. b: m = 010:110 = 32° 1' 32° 13' c: p = 001:lll=48° 2' 48" 6' c. n = 001:101=43°18' 43° 0' c r = 001: 332 = 69° 3' 69° 30' c: = 001:201= 62° 3' c « = 001: 221 = 66° 47' 64° 46' c: fc = 001:011=30°30' 30° 19' c « = 001:331 = 73° 18' 71° 52' c: 2 = 001:021= 49° 20' c: w = 001: 441 = 77° 20' 77° 25' c: 5 = 001:112 = 29° 4' 30° 0' II to b. II to a. C II toe. X720... X580. . . X600... 1.6610-1.6627 1.6670-1.5705 1.6786-1.6847 1.6220 1.6266 1.6360 1.6096 1.6144 1.6202 No pleochroism was observed. Plane of axes a (100), obtuse bisectrix normal to base. Double refraction +. Cleavage was observed on the base. Rubidium Uranyl Sulphate. Rb2U02(S04)42H20. The rubidium salt was more fluorescent than the potassium salt and mor difficult to crystallize, so that measurements of it were not obtained. It is, however, completely isomorphous with the potassium salt. The composition, according to Rimbach,^ is as above, with 2 molecules of water. 1 Rimbach, I. c, 478. 2de Coninck, BuU. Acad. Roy. Belg., 1904, 1171; 1905, 50, 94. « Ibid., 1905, 50. * Rimbach, I. c, 479. CHEMISTRY OF FLUORESCING URANYL SALTS. 221 C^siTTM Urantl Sulphate. CS2U02(SO4)23H2O. This salt is so insoluble that no crystals could be produced. The Cragwall product recrystalUzed showed on the microscope-sHde square plates about 10 ju in length, which showed a negative uniaxial figure. This salt, according to de Coninck/ has the same composition which he finds for the potassium salt, that is, 3H2O. Ammonium Urantl Sulphate. (NH4)2U02(S04)22H20. This salt was not recrystalhzed, but showed similar characteristics to the potassimn salt. When recrystalhzed it gave bundles of needles, the vibra- tions across which were most absorbed and most refracted. Rimbach de- scribes the salt as having 2 molecules of water. The crystals were found to be monoclinic by de la Provostaye.^ Sodium Urantl Sulphate. Na2U02(S04)23H20. This salt is described by de Coninck as having 3 molecules of water, which he finds also in the potassium salt. This salt as prepared by Cragwall was not recrystalhzed, but showed under the microscope one optical axis and the acute bisectrix with positive double refraction. It is, therefore, presumably triclinic, as the acute bisectrix was off normal in both directions. Thallous Urantl Sulphate. Tl2U02(S04)23H20. This salt was prepared by Cragwall from weighed amounts of the two salts according to Kohn,^ who found the salt to be of the above composition with probably 3H2O. The crystal description by Himmelbauer in the same article gives the system as rhombic, the symmetry from etch figures pyramidal. The forms are the three pinacoids with pyramid faces at the corners which were too small to measure. He observed through (100) in converged polarized light the plane of the axes for red and blue, at right angles the plane of the blue being that of the a and c axes; for green nearly uniaxial; for red fight a is the acute bisectrix, the pleochroism on (100) distinct, parallel to c, deep yellow; parallel h yellowish white, but no noticeable pleochroism on (010). Crystals up to 3 mm. in diameter and 1 mm. thick, produced by slow cooling of the Cragwall salt, showed the axial figure, but it could not be surely seen to agree with the description, due to the intense absorption in the blue and green. The figure might be explained by anomalous dispersion due to the absorption band. PHOSPHATES. Uranyl phosphate (HUO2PO4.3IH2O), which precipitates from uranyl solu- tions on adding phosphates, possesses no fluorescence. If it is dissolved in an excess of acid it gives a glass or sirup with a brilhant fluorescence which can not be resolved beyond the bands. The sodium double salt was made by adding sodium phosphate to produce H2Na2C02(P04)2 to the uranyl phosphate with an excess of water, which on standing and evaporating gave 1 de Coninck, Bull. Acad. Roy. Belg., 1905, 94. 2 de la Provostaye, Ann. Chim. Phys. (3), 6, 51. 1842. 3 Kohn, Z. Anorg. Chem., 59, 111. 1908. 222 FLUORESCENCE OF THE URANYL SALTS. a fine crystalline mass which was very fluorescent. The spectrum of this was studied. The potassium, ammonium, lithium, and calcium salts were also prepared and seen to have characteristic line spectra, but were not studied further. The mineral autunite is a basic calcium uranyl phosphate which Stokes^ says shows briUiant fluorescence, while chalcolite, the analagous copper compound, has none, but shows the same absorption bands characteristic of uranyl compounds. CHROMATES. An attempt was made to prepare the sodium uranyl chromate described by E-imbach, which resulted in a brown mass. The uranyl chromate U02Cr04- 3H2O (Orloff)2 from UOs-and CrOa gave yellow needles with no fluorescence. The potassium salt from K2Cr207 and UO3 was also without fluoresecence. FLUORIDES. Cragwall prepared the uranous and uranyl fluoride from UaOs and HF, which showed practically no fluorescence. He also prepared the double potassium salt K3UO2F6 by adding KF to uranyl nitrate and (NH4)3 UO2F5 by dissolving (NH4)2 U2O7 in HF.^ The double salts showed characteristic spectra, but the fluorescence was very weak. URANYL lODATE. This salt was prepared from sodium iodate and uranyl nitrate by Cragwall by a method which, according to Artmann,^ would result in U02(I03)2H20 of the rhombic form. This showed Uttle fluorescence. MISCELLANEOUS INORGANIC COMPOUNDS. An attempt was made to produce bromides and iodides analogous to the chloride salts without results, due to decomposition with the Hberation of bromine and iodine. An attempt was also made to produce molybdyl and timgstyl ammonium chloride double salts analogous to the uranyl salts by heating the oxides with ammonium chloride and hydrochloric acid in sealed tubes, which in some cases resulted in crystals, which, however, showed no fluorescence. Uranic acid or H2UO4 was also sealed up in tubes with anhy- drous Uquid NH3, CO2, SO2, and HCl. None of the resulting compounds were soluble or fluorescent, although changes took place, the carbonate being nearly white, the sulphur-dioxide tube greenish, due to reduction, and the ammonia tube reddish Hke the diuranate. Of the uranates, the sodium potassium calcium and barium were made, none of which showed fluorescence, the first two being golden yellow plates, the latter two an amorphous greenish mass. URANYL ACETATES. The anhydrous uranyl acetate U02(C2H302)2 was. prepared by Cragwall according to Spath^ by adding acetic anhydride to uranic oxide. This latter took up some water and became partially the dih3^drate. On recrystalUzing some of the material from acetic-acid solution, small clear cubes were obtained which appeared to contain acetic acid of crystalUzation. 1 stokes, Phil. Trans. Roy. Soc. London, 142, 518. 1852. 2 Orloff, Chem, Ztg., 31, 375. 1907. "" H. F. Baker, Chem. Soc. Jour., 35, 763-769. 1879, * Artmann, Z. Anorg. Chem., 79, 327, 1913. B Spath. Monatsh. J. Ch. 33, 248. 1912. CHEMISTRY OF FLUORESCING URANYL SALTS. 223 Urantl Acetate Dihtdrate. U02(C2H302)22H20. This salt was prepared by Cragwall by recrystallizing the anhydride from water solution. There was also a stock of material from Kahlbaum and un- known sources. In an attempt to recrystaUize this salt in large, clear crystals much difficulty was met, as it usually fills with cracks as it grows. The best material, having as much as 4 mm. cube of clear material, was obtained by supersaturating 2 grams in 200 c.c. and allowing a month or two to crystalhze. The crystal properties were found to be similar to those described by Schabus,^ the system being rhombic, with an axial ratio of 0.7817: 1:0.3551, with the forms m (110), a (100), r (101), n (120), and h (010). The prism zone is very much striated, affording a continuous procession of reflections in the gonio- meter. Bchabus finds cleavage on m, a, h, c, which accounts for their extreme friability. Zehenter^ finds the specific gravity to be 2.893. The refractive index was determined through the dome as being 1.490 for hght vibrating parallel to the c axis and 1.521 parallel to h. The uranyl acetate trihydrate, which, according to Schabus, forms below 10° C, was not prepared. It crystaUizes in the tetragonal system with an axial ratio of a : c = 1 : 1.4054. DOUBLE URANYL ACETATES. The sodium salt occurs in the acetates as well as the sulphates and is the only m^anyl salt crystaUizing in the regular system. The ammoniimi and potassium salts seem to be isomorphous, in spite of the fact that the potassium salt is said to contain 1 molecule of water and the ammonium salt to be anhy- drous. The silver salt, which contains 1 molecule of water, is apparently isomorphous and the rubidiiun salt was f oirnd to have a similar axial ratio, as usual very near that of the ammonium salt. The similarity of axial ratio to that of the uranyl nitrate trihydrate suggests that these salts form a group that might weU be studied further. The csesiimi salt could not be obtained, the uranyl dihydrate crystaUizing out and leaving the caesiiun acetate in solution. The two hydrates of Hthium uranyl acetate fuUy described by Wyrouboff^ as being monochnic were attempted, but only the room-tempera- ture form was obtained. The double salts of uranyl acetate with bivalent acetates were in general prepared by dissolving the oxide or carbonate of the second metal in acetic acid in excess, adding lu^anyl acetate in calculated amount, with water enough for complete solution, and allowing to crystallize by slow evaporation. Difficulties were encountered in the preparation of some of the salts, such as the calcium salt, which Rammelsberg also could not obtain, as described by Weselsky,* which was finally prepared by Weselsky method of precipitation with calcixmi carbonate and solution of the precipitate in acetic acid. The barixun salt was finally prepared by this method. The cadmium was never prepared at all; at least, no specimen that gave anything but the uranyl-acetate spectrima. An attempt to produce a mercuric acetate also failed. 1 Schabus, Prieschr. Wien, 207. 1855. 2 Zehenter, Monats f. Ch., 21, 235. 1900. " Wyrouboff, Bull. Soc. fran. min., 8, 115-122. 1885. * Weselsky. J. Prakt. Chem., 75, 55. 1858. 224 FLUORESCENCE OF THE URANYL SALTS. The double acetates as a group were studied by "Wertheim/ Schabus,^ Grailich/ Weselsky,^ and Rammelsberg.^ The acetate group in general show much less intense fluorescence, tending to be of a dull yellow color. The uranyl double acetates with bivalent metals may be divided into two classes— the normal and the abnonnal. The group HU02(C2H302)33H20 appears to act as a imit in forming crystals. In the alkali double salts the water of crj^staUization seems to be lacking, at least in the well-confirmed cases of the sodium and ammonium salts. The case of the manganese, cad- mium, and lead double salts seem also to be an exception, but with the other double acetates the ratio of uranium acetate to bivalent acetate seems to be 2 to 1. The water of crystalHzation is variously given from 7, which was found uniformly by Rammelsberg, to 8 by Wertheim and 10 by Graihch. Since the water is likelj' to nm high, due to occluded mother-liquor, and is such a small per cent of the total weight, it is not unreasonable to assume that these reaUy are all hexahydrates. The manganese, when satisfjdng this valence ratio, and magnesiimi salts seem also to have 2 molecules of water to each uranyl radical. In the case of the triple salts this requirement is exactly fulfilled, each valence of base having a U02(C2H302)3H20 group attached to it. In the case of the manganese, cadmiimi, and lead salts, this radical does not seem to act, but simply the two acetates are present in a 1 to 1 ratio. The spectra of the manganese salt was like that of the other double acetates; the cadmium was not formed or else gave a spectrum Uke that of the single ace- tate, and the lead was one of the salts which showed fluorescence lines coinci- dent with spark lines, so that no generalization can be made. Sodium Ur-ajntij Acetate. NaUOaCCsHaOj),. This well-known salt described by Graihch^ crystaUizes in the regular system with the least or pentagonal dodecahedral symmetry. It is usually in the form of tetrahedra, yellow, with hght green fluorescence. Johnsen^ gives the specific gravity as 2.562 and the refractive index as 1.5014. Marback^ and Traube^ give the optical rotation as 1.48°. The best crystals, up to 3 mm. in thickness by 8 mm. diameter, were obtained on long standing of sKghtly supersaturated solutions. Dr. Xishikawa tried to obtain X-ray diffraction patterns with these crj'stals, but could obtain nothing. Potassium URA>TrL Acetate. KU02(C2H302)3H20. This salt was described by Wertheim^** as having 1 molecule of water, which was also fotmd by Schabus" and Rammelsberg.^ A recent determination by 1 Wertheim, Jour, of Prakt. Chem. 29. 207-231. 1843. * Schabus, Best. d. KystAll gest. i. chem. Lab. Erz Prod. Preischr. Wien. 1855. * Grailich, KJo'st Opt. Untersuchung. Preisch. Wien, pp. 151-17o. 1S5S.