THE UNIVERSITY OF MICHIGA N COLLEGE OF LITERATURE, SCIENCE, AND THE ARTS Department of Chemistry Progress Report A STUDY OF "INDUCED" ABSORPTION IN URANYL SOLUTIONS AND GLASS Thomas M. Dunn Lee A. CrIs, ORA Poe 1.t 4 under contract with: U. S. AIR FORCE AIR FORCE OFFICE OF SCIENTIFIC RESEARCH OFFICE OF AEROSPACE RESEARCH GRANT NO. AF-AFOSR-865.-65 administered through: OFFICE OF RESEARCH ADMINISTRATION ANN ARBORugh: February 1966

. 2f\,,, L IAPf I

ABSTRACT An analysis of the absorption induced by an intense xenon light flash in several media containing fluorescent uranyl ions has been accomplished. Spectrographic studies on uranyl glass (Canary glass) and a solution of uranyl sulfate in concentrated sulphuric acid have shown the induced absorption to be a broad band extending from 4600 A to 7000 A with a maximum in the neighborhood of 5800 A. Spectrophotometric studies on the same systems have shown that the lifetime for this absorption is the same as for the normal fluorescence and that the transition being observed via the absorption measurements is from the initial state of the fluorescence to a still more highly excited state. The failure of uranyl salts to exhibit the induced absorption in solutio.ns in which they are only faintly fluorescent (for example, in water or alcohol) confirms the assignment of the initial state of this transition. The absorption induced in a solution of uranyl phosphate in 80% phosphoric acid has been shown by spectrophotometric means. to be nearly indistinguishable from that induced in the sulfate solution. Ultraviolet absorption measurements of the spectra of the sulfate and phosphate solutions have shown that the absorption which normally appears as a high frequency continuum beginning at 3500 A in samples which are concentrated enough to show the visible absorption system in the region from 4900 A to 3500 A has its peak at 2200 A (or perhaps to shorter wavelengths in the phosphate) and is detectable at 3500 A only because of its moderately high molar extinction coefficient which is much larger than that of the visible system. In the sulfate solution, the band at 2200 A has two longer wavelength shoulders at 2600 A and 2830 A and it is possible to show, by assuming that the initial state of the induced absorption is also the initial state of the ordinary fluorescence, that the final state of the induced absorption must be the same as the terminal level of the 2600 A shoulder. The oscillator strengths for the three transitions considered are: (1) from the vibrationless ground state to the vibrationless first excited electronic state (the resonance line): 1.28 x 10-6, (2) from the vibrationless ground state to the central UV level: 2.29 x 10-3, and (3) from the vibrationless first excited electronic state to the central UV level (the induced absorption): 6.54 x 10-5. iii

ACKNOWLEDGMENTS I would like to acknowledge Professor Thomas M. Dunn for his continuing interest and support of this work, for many invaluable discussions regarding the interpretation of molecular spectra, and for his assistance in making the following pages more intelligible. To Professors Chihiro Kikuchi and Edgar F. Westrum goes my appreciation for their valuable suggestions concerning this manuscript. The cooperation and skill of Mr. Norman G. Johnston and Mr. Ernest Metzner of the chemistry machine shop who constructed the flash tube heads is gratefully recognized. The construction of the power supplies and mechanical timing equipment by Mr. Victor Podolak of the chemistry electronics shop was of considerable assistance. The support of this research by the United States Air Force under Grant No. AF-AFOSR-865-65 is thankfully acknowledged.

TABLE OF CONTENTS Page ABSTRACT iii ACKNOWLEDGMENTS v LIST OF TABLES ix LIST OF FIGURES xi LIST OF APPENDICES xv I. STATEMENT OF THE PROBLEM 1 II. EXPERIMENTAL 2 A. Ordinary Absorption Spectra of Uranyl Glass and Uranyl Sulfate and Phosphate Solutions 2 B. Ordinary Fluorescence Spectra of Uranyl Glass, Uranyl Sulfate Solution, and Polycrystalline Uranyl Sulfate Trihydrate 9 C. Induced Absorption Spectra 19 D. Time-Resolved Induced Absorption in Uranyl Glass and Uranyl Sulfate and Phosphate Solutions 27 E. Time-Resolved Fluorescence in Uranyl Glass and Uranyl Sulfate Solution; Comparison of DC- and Flash-Excited Fluorescence Spectra of Uranyl Sulfate Trihydrate 42 F. Ultraviolet Absorption Spectra of Uranyl Sulfate Solution and Uranyl Phosphate Solution 50 III. DISCUSSION OF RESULTS AND SUMMARY 57 BIBLIOGRAPHY 91 vii

LIST OF TABLES Table Page I Absorption Bands of Uranyl Glass 2 II Absorption Bands of Uranyl Sulfate Solution 6 III Absorption Bands of Uranyl Phosphate Solution 7 IV Molar Extinction Coefficients of Uranyl Sulfate Solution 7 V Fluorescence Bands of Uranyl Glass 9 VI Fluorescence Bands of Uranyl Sulfate Solution 16 VII Fluorescence Bands of Uranyl Sulfate, Polycrystalline, 3000K 17 VIII Changes in Ordinary Absorption Peaks of Sulfate Solution Upon Optical Pumping 25 IX Ultraviolet Absorption Bands of Uranyl Sulfate and Phosphate Solutions 55 X The Character Table for DIh 62 XI Assignment of Terms for a Molecule of Symmetry DIIh 6h XII Assignment of the Uranium and Oxygen Orbitals to the Representations of the DTh Group 64 XIII Assignment of the Oxygen Linear Combinations to the Representations of the Doh Group 65 XIV The Character Table for the Doh Double Group 70 T1 XV Assignment of Terms for a Molecule of the D-h Double Group 71 XVI Resolution, Dispersion, and Wavelength-.Ranges for the 31,000 A Blaze Side of 3.4 Meter Ebert Spectrograph 87 XVII Resolution, Dispersion, and Wavelength Ranges for the 59,000 A Blaze Side of the 3.4 Meter Ebert Spectrograph 87 ix

LIST OF FIGURES Figure Page 1. Visible Absorption Spectrum of Uranyl Glass. 3 2. Visible Absorption Spectrum of Uranyl Sulfate in Sulphuric Acid Solution. 4 3. Visible Absorption Spectrum of Uranyl Phosphate in Phosphoric Acid Solution. 5 4. Fluorescence Spectrum of Uranyl Glass, 3000K. 10 5. Fluorescence Spectrum of Uranyl Glass, 770K. 11 6. Fluorescence Spectrum of Uranyl Sulfate in Sulphuric Acid Solution, 3000K. 12 7. Fluorescence Spectrum of Uranyl Sulfate in Sulphuric Acid Solution, 770K. 13 8. Fluorescence Spectrum of U02S043H20, 300~K. 14 9. Fluorescence Spectrum of U02S04.3H20, 770K. 15 10. Energy Level Diagram for Uranyl Solutions, the Glass, and Uranyl Sulfate. 18 11. Photographically Recorded Transmission Spectra of Optically Pumped and Unpumped Uranyl Glass. 20 12. Induced Absorption Cross-Section of Uranyl Glass. 21 13. Photographically Recorded Transmission Spectra of Optically Pumped and Unpumped Uranyl Sulfate Solution. 22 14. Induced Absorption Cross-Section of Uranyl Sulfate Solution. 23 15. Time-Resolved Induced Absorption in Uranyl Glass. 28 16. Induced Absorption Cross-Section for Uranyl Glass as Deduced from Photoelectric Measurements (versus wavelength). 29 xi

LIST OF FIGURES (Continued) Figure Page 17o Induced Absorption Cross-Section for Uranyl Glass as Deduced from Photoelectric Measurements (versus energy), 30 18. Time-Resolved Induced Absorption in Uranyl Sulfate Solution. 31 19o Induced Absorption Cross-Section for Uranyl Sulfate Solution as Deduced from Photoelectric Measurements (versus wavelength. 33 20. Induced Absorption Cross-Section for Uranyl Phosphate Solution as Deduced from Photoelectric Measurements (versus wavelength). 34 21. Theoretical Time-Profile for the Excited State Population Assuming Square-Wave Pumping. 37 22. Induced Absorption Decay Measurement for Uranyl Glass. 38 23. Induced Absorption Decay Measurement for Uranyl Sulfate Solution. 39 24. Induced Absorption Decay for Uranyl Glass. 41 25~ Time-Resolved Fluorescence for Uranyl Glass. 43 26, Time-Resolved Fluorescence for U02S04 in H2S04 Solution. 44 27. Fluorescence Decay Measurements for U02S04 in H2S04 Solution. 45 28. Fluorescence Decay for Uranyl Glass. 46 29. Fluorescence Decay for U02S04 in H2S04 Solution. 47 30. DC- and Flash-Excited Fluorescence of U02SO043H20 at 77~K. 49 31. Ultraviolet Absorption Spectrum of U02SO04 in H2S04 Solution (versus wavelength). 51 32. Ultraviolet Absorption Spectrum of U02S04 in H2S04 Solution (versus energy). 52 335 Resolution of the Ultraviolet Absorption Spectrum of U02S04 in H2S04 Solution by Subtraction of the Induced Absorption Contour. 53 xii

LIST OF FIGURES (Concluded) Figure Page 34. Ultraviolet Absorption Spectrum of UO2PO04 in H3P04 Solution Showing the Relationship Between the Long Wavelength Shoulder and the Induced Absorption Maximum. 54 35. Principle Transitions Observed for the Uranyl Ion. 60 36. Flash Head and Arc Lamp Arrangement. 73 37. Optical Bench Arrangement for the Recording of Induced Absorption Spectra. 75 38. Camera Shutter, Mirror, and Trigger Photocell Arrangement. 76 39. Oscilloscope and Delay Network Arrangement. 78 40. Induced Absorption Tracings for U02S04 in H2S04 Solution Illustrating Film Error. 82 41. Intensity "Amplification" by Film H & D Curve. 83 42. H & D Curve for Kodak Pan-X Film. 85 43. Order-Sorter for 3.4 Meter Jarrell-Ash Spectrograph. 88 44. An Attempt to Observe Quenching of the Induced Absorption in Uranyl Sulfate Solution. 90 xiii

LIST OF APPENDICES Appendix Page I. DISCUSSION OF THE MOLECULAR ORBITAL REPRESENTATIONS FOR THE URANYL ION AND THE ASSIGNMENT OF TERM LEVEL DESIGNATIONS TO THE LOWER URANYL ION LEVELS 62 II. THE EXPERIMENTAL ARRANGEMENT AND ITS OPERATION 72 III. ATTEMPTS AT CRYSTAL GROWTH AND REMARKS ON SIZE AND CLARITY REQUIREMENTS 79 IV. FILM ERRORS, THEIR EFFECTS ON THE SPECTRA, AND THEIR CORRECTION 81 V. A DESCRIPTION OF THE SPECTROGRAPHS 86 VI. AN ATTEMPT TO OBSERVE "BLEACHING" OF THE INDUCED ABSORPTION 89 XV

SECTION I STATEMENT OF THE PROBLEM In connection with work concerning the construction of high power lasers, this author and a co-worker1 reported the observation of an absorption "induced" in Corning No. 3-79 filter glass (a glass whose coloring constituent is the uranyl ion) by the light flash from a high-intensity xenon flash lamp. These authors noted that this absorption was quite broad, extending from 5460 A to at least 7000 A and decayed at the same rate as the fluorescence of the glass. By the construction of a laser cavity incorporating this glass as a saturable filter, they were able to demonstrate that the absorption could be wiped out by the application of an intense "depumping" light provided by the laser,2 Because of experimental limitations, this induced absorption could only be monitored photoelectrically and then at only a few wavelengths so that the shape of the absorption remained unknown, as did the exact nature of the levels taking part in the absorption. As it seemed likely that the induced absorption was in fact a transition from the initial state of the normally observed fluorescence to some higher excited state, it was decided that a series of experiments could be set up to establish whether or not this was the case, One alternate proposal for the appearance of this absorption would be that some sort of photodecomposition product is formed in the glass which happens to have the same recombination rate as the normal fluorescence decay rate and that it is this product which gives rise to the absorption. If the previous hypothesis concerning the nature of the initial level is correct, then those systems which contain uranyl ions but the mechanics of which are such that no fluorescence occurs, the induced absorption should not be observable and all uranyl systems which are fluorescent should exhibit it. The terminal level of the transition also must be established, i.e., the higher excited state which is the upper state for the induced absorption pxocess., Assuming that the transition being observed is in fact between states of the uranyl ion, the possibility exists that the transition from the ground state to this excited state is forbidden and therefore the presence of this state could not be detected by ordinary absorption spectroscopy. 1

SECTION II EXPERIMENTAL A. ORDINARY ABSORPTION SPECTRA OF URANYL GLASS AND URANJYL SULFATE AND PHOSPHATE SOLUTIONS Using a Cary Model 11 spectrophotometer, whose resolution is approximately 10 cm1, the absorption spectra in the region from 3500 to 5500 A of the following materials were run at 300 ~K: (1) a 5 mm thick sample of Corning No. 3 —79 filter glass (uranyL glass); (2) a 10 mm path length sample of a solution of uranyl sulfate (UO2SO4o3H2O) in concentrated sulphuric acid (4~7 x 10-2gm/ml); and (3) a 10 mm path length sample of a solution of uranyl carbonate* in 80% phosphoric acid (7.0 x 10-2gm/ml). These spectra are shown in Figures 1-3, respectivelyo Table I lists the transitions which are observed for the uranyl glass, TABLE I ABSORPTION BANDS OF URANYL GLASS Wavelength (X) Frequency (cm-l) Av5130 19500 1070 4862 20570 4601 21740 1980 4216 23720 400 4147 24120 2510 3756 26630 650 3668 27260 The glass spectrum is quite diffuse but is much more detailed than the Corning *Prepared by precipitation from a concentrated solution of uranyl nitrate with ammonium carbonate. 2

2.0 Sample = One 5 mm. thick section of Corning 43-79 filter glass 1.5 >l-, s0ac1.0 0.5 0.0 1. I 3500 4000 4500 5000 5500 WAVE LENGTH (A) Figure 1. Visible Absorption Spectrum of Uranyl Glass, 3000K.

2..O 1.5 z 1.0 0.5 0.0 3700 4000 4300 4600 4900 5200 0 WAVE LENGTH (A) Figure 2. Visible Absorption Spectrum of Uranyl Sulfate in Sulphuric Acid Solution, 300~K. 4

2.0 1.0 n 0.5 0.0 3400 3700 4000 4300 4600 4900 5200 WAVE LENGTH (A) Figure 3. Visible Absorption Spectrum of Uranyl Phosphate in Phosphoric Acid Solution, 3000K.

Company's published transmission curve* and it does show some trace of the different electronic transitions of the uranyl ion to be expected in this region3 as well as some finer structure which may either be attributed to vibrational intervals or to the presence of different uranyl ion sites in the glass. The solution spectra, whose bands are listed in Tables II and III, reTABLE II ABSORPTION BANDS OF URANYL SULFATE SOLUTION In Sulphuric Acid In Water (This Study) (Morton and Bolton4) x(A), v((cm1l) Av,() v(cm-1) Av 4922 20320 750 4955 20160 850 4751 21050 940 4750 21050 850 4548 21990 800 4570 21900 650 4589 22790 740 4434 22550 620 4250 25550 700 4315 23120 4127 24230 700 4200 23810 700 4010 24930 750 4080 24510 3895 25680 710 3789 26390 74 3686 27130 3539 28260 spectively, have a much better resolved vibrational structure than the glass. The vibrational interval observed here is the symmetric stretching frequency. The average value of this interval is 720 cm-1 for the sulfate and 640 cm1l for the phosphate but some large deviations are observed in the high- and low-wavelength portions of this region. These are probably due to the fact that there are at least three electronic transitions in this region and that the intervals which deviate represent the transition from the vibrational structure of one electronic state to that of another, Table II also lists for comparison the absorption bands of an aqueous solution of uranyl sulfate and Table III the absorption bands of an aqueous solution of uranyl monophosphate as measured by Morton and Bolton4 and corrected by Kayser.5 *"Glass Color Filters," Corning Glass Works, Corning, New York 6

TABLE III ABSORPTION BANDS OF URANYL PHOSPHATE SOLUTION In Phosphoric Acid In Water (This Study) (Morton and Bolton4) x(X) v(cm-l) Al} -v (a) v(cm-l) Av 5036 19860 710 5060 19760 1180 4862 20570 680 4775 20940 730 4706 21250 620 4615 21670 650 4572 21870 4480 22320 650 700 4444 22520 600 4345 23020 570 4325 23120 600 4240 23590 4216 23720 670 4100 24390 610 4000 25000 890 3862 25890 490 3753 26380 930 3662 27310 The absorption spectrum of the sulfate is sufficiently well resolved so that the oscillator strength of the resonance line at 20,320 cm1l may be calculated. Table IV lists the molar extinction coefficients for the 11 absorption peaks of TABLE IV MOLAR EXTINCTION COEFFICIENTS OF URANYL SULFATE SOLUTION Band No. v(cm-l) c 1 20320 1 1 2 21050 2.2 3 21990 4. 0 4 22790 9.3 5 23530 11.8 6 24230 10.0 7 24930 6.8 8 25680 4. 2 9 26390 3.7 10 27130 3.6 11 28260 6.8

the uranyl sulfate solution spectrum. If a Gaussian line shape of the form E(V) = Eo exp (-k(v-v)e2) is assumed for the resonance line then the oscillator strength may be directly obtained. The oscillator strength can be calculated via a formula given by Mulliken6: f = 4.20 x l0-8fkvdV where V is the frequency in cm-1 and kv is the absorption coefficient at the frequency v measured in (cm-atmospheres)-l at O0C. This form is clearly most useful for gasses but it can be normalized to the measurement of the absorption coefficient as a molar extinction coefficient, e(v), and in this form is given by Rabinowitch and Belford14 as: f = 4.32 x lo-9fEc()dv For a Gaussian line, this reduces to f = 4.32 x 10-9eo Av The half width of the resonance line is 271 cmnl and from Table IV the molar extinction coefficient at the center of the line is 1.1. This yields an oscillator strength of 1.23 x 10-6. The absorption cross-section at line center is 3.24 x 10-23 cm2. The oscillator strength for the entire first electronic system, assuming that the absorption profile is symmetric about the 21,990 cm-1 band (the 3rd vibrational quantum of the first electronic state) is 1.25 x 10-5. 8

B. ORDINARY FLUORESCENCE SPECTRA OF URANYL GLASS, URANYL SULFATE SOLUTION, AND POLYCRYSTALLINE URANYL SULFATE TRIHYDRATE The fluorescence spectra of uranyl glass, a solution of uranyl sulfate in concentrated sulphuric acid (4.7 x 10-2 gm/ml), and polycrystalline uranyl sulfate trihydrate (UO2S04K3H20) were recorded using a 1-meter Jarrell-Ash scanning spectrometer at 300~K and 77~K. This instrument had a grating ruled with 590 lines/mm and a total ruled area of 102 x 102 mm. In the first order, the dispersion is 16.4 I/mm. These spectra are shown in Figures 4 through 9. The room temperature spectra of the two solid samples were taken by supporting the sample directly in front of the spectrometer input slit and illuminating it from the side. The solution spectrum and all the spectra taken at 770K were made by placing the sample in the square cross-sectioned tip of a Dewar flask and supporting this section of the flask directly in front of the slit. The exciting light was provided by a low pressure mercury-neon discharge lamp* and the fluorescent output was detected by an RCA 1P21 photomultiplier operated at 1200 VDC. The slit width employed varied from 200 to 25 microns depending on the intensity of the fluorescent output which implies a resolution of from 9 to 1.6 cm1l. Fortunately, the bandwidth of the fluorescent bands was always much larger than the necessary bandwidth which was resolvable because of intensity considerations. The 770K solution spectrum was obtained by dropping the solution into the Dewar flask filled with liquid nitrogen; the frozen globules then fell through the nitrogen to the bottom of the flask. The polycrystalline uranyl sulfate showed a considerable amount of short-lived triboluminescence as it warmed up. TABLE V FLUORESCENCE BANDS OF URANYL GLASS Temperature Band 3000K 770K N(A) v(cm-l) Av X( A) v(cml-1) Av 1 5181 19300 5138 19460 780 2 5327 18770 700 5352 18680 890 3 5534 18070 5558 17790 *"Pen-Ray" type obtained from Ultraviolet Products Co., San Gabriel, California. Operated with their Type SCT-3 power supply. 9

7 X - Mercury Emission Lines 6A I.2 x 1 5675 5550 5425 5300 5175 5050 4925 z3 10

6 x X -?Mercury Emission Lines 5 x LU Ln 0 2 O. I I I I 5675 5550 5425 5300 5175 5050 4925 WAVE LENGTH (A) Figure 5. Fluorescence Spectrum of Uranyl Glass, 770K. 11

x x X-Llercury Emission Lines UEMISSION 5550 5425 5300 5175 5050 4925 4800 WAVE LENGTH (A) Figure 6. Fluorescence Spectrum of Uranyl Sulfate in Sulphuric Acid Solution, 3000K.

X -iercury Elmrission Lines EM I SSI ON X 5675 5550 5425 5300 5175 5050 4925 4800 WAVE LENGTH (A) Figure 7. Fluorescence Spectrum of Uranyl Sulfate in Sulphuric Acid Solution, 770K. 13

X - Mercury Emission Lines X XXX~~~~~x fEMISSION x 5800 5625 5550 5425 5300 5175 5050 4925 0 WAVE LENGTH (A) Figure 8. Fluorescence Spectrum of U02S04-3H20, 300~K.

X-Ilercury Emission Lines I__ _1.x In I I 1 I 1X 5675 5550 5450 5350 5150 5050 4950 4850 0 WAVELENGTH A Figure 9. Fluorescence Spectrum of U02S04.3H20, 770K. 15

The spectrum of uranyl glass shows a considerable amount of sharpening as it is cooled from 300'K to 770K as well as shifts in the positions and relative intensities of the bands. Table VI lists the positions of the bands at 300 and 770k along with the data of Gordon7 of the spectrum of a 3 molar frozen aqueous solution at 770K. The sulfate solution and polycrystalline uranyl sulfate also exhibit shifts in the positions and relative intensities of the bands and the spectra of both become sharper at lower temperatures. Only in the case of the polycrystalline uranyl sulfate, however, is there any more structure to the fluorescence at 77~K than is observable at room temperature. The vibrational quantum observed in the fluorescence spectrum of the sulfate solution is that of the symmetric stretch and the bands are given in Table VI. The bands observed for crystalline uranyl sulfate at 300~K are listed in Table VII along with the same data obtained by Pant.8 The variations between the positions of the bands as measured here and those of the other authors are not significant since they are within the experimental error. It is interesting to note that in the case of the frozen sulfate solution none of the other vibrations of the molecule are seen (as in the case of the crystals even at room temperature) even through the bandwidths of the fluorescence bands which are seen are narrow enough to prevent their masking the presence of other bands. TABLE VI FLUORESCENCE BANDS OF URANYL SULFATE SOLUTION In Sulphuric Acid In Water Band (This Study) (Gordon7) No. T = 300~K T = 770K T =77K K(a) v( cm-l) Av x(A) v(cm-l) AV ( cm-l) 1 4915 20346 926 4871 20530 950 20310 2 5150 19420 917 5106 19583 19460 917 10690 3 5404 18503 5362 18649 870 18610 4 5625 17780 17750 5 16900 Figure 10 shows a resume of the energy levels of the uranyl ion deduced from analysis of the absorption and fluorescence spectra in these various hosts. 16

TABLE VII FLUORESCENCE BANDS OF URANYL SULFATE, POLYCRYSTALLINE, 3000K Band (This Study) (Pant') No.?(a) v(cm-l) Av v(cm-l) 1 4888 20460 230 20460 2 4944 20230 210 20244 3 4994 20020 410 19976 4 5099 19610 240 19615 5 5163 19370 630 19392 6 missing 19134 7 5337 18740 210 18755 8 5398 18530 350 18560 9 5500 18180 310 18304 10 5595 17870 200 17899 11 5660 17670 17695 17

UO2SO4 IN H2SO4 CANARY GLASS UO2SO4'3H20 UO2HPO4 IN H3P04 1129.8 i 750 27.0 738.6 760 26.0 - 715.8 960 743.4 25.0 670 701.6 620 24.0 - 702.3 640 23.0 752.7 FLUORESCENCE FLUORESCENCE FLUORESCENCE 23" 0 — 680 x 797.4 640 22.0 -77 / [300~K] [77~K] 2 937.6 [300~K] [7707K] [77K] 730 21.0 660 [ 731.1 (20346) (20228.2) (20318) 20.0 (19495) (19461) (20270) (19302) 2.5 -2066 ABSORPTION ABSORPTION 870.1 3000 K 2.0 - [300~K] 30KABSO 3R08 8 [300*K] 215. _ 1.5 - 934.1 934.1 ________ 692.7 630.4 700.7.5 957.0 888.0 776.8 532.0 615.3 Figure 10. Energy Level Diagrams for Uranyl Solutions, the Glass, and Uranyl Sulfate. 18

C. INDUCED ABSORPTION SPECTRA The spectrum of a 1600 watt compact arc lamp* transmitted through a 5-mm thick section of uranyl glass and a solution of U02S04 in H2S04 (0.0466 gm/ml) contained in a 10-mm path length Beckman liquid sample cell were recorded on Kodak Pan-X film in the region from 3750 J to 6500 A using the first order of Al.5 meter Bausch and Lomb grating spectrograph both with and without excitation with a xenon flash lamp.** The resolution of this spectrograph is about 50,000 in the first order and the input energy to the flash tube was nominally 500 joules (2000 VDC at 250 mfd). These spectra for uranyl glass and uranyl sulfate are given in Figures 11 and 13 respectively. The details of the timing and triggering networks used to obtain the spectra only when the flash lamp is firing are discussed in Appendix II. The difficulties encountered due to the nonlinearity of the film blackening with respect to log (exposure) are discussed in Appendix IV. If the optical density of the film versus log (exposure) curve for the film is assumed to be linear, then the difference in optical density between the pumped and unpumped tracings should be directly proportional to the induced absorption coefficient. This does, of course, ignore the spectral response of the film which will modify the constant of proportionality between the change in optical density and the absorption coefficient. The correction for this variation can be made by reference to the published spectral response curves for a given film but since it was known that spectrophotometric measurements of the absorption were also to be made, this correction was not applied as the photoelectric measurements can give a much better measure of the magnitude of the induced absorption as well as its period of induction and decay. The induced absorption cross-sections for uranyl glass and sulfate solution are given in Figures 12 and 14, respectively. The induced absorption in the uranyl glass is quite diffuse and shows only a vague structure. The sulfate spectrum is equally diffuse and shows no sign of any separation which might be taken to be a vibrational separation. (The dips around 5275 A were subsequently shown by photoelectric measurements to be fallacious. Their exact cause is uncertain.) Since the sulfate solution absorption and fluorescence speptra at 300'K both show considerable vibrational structure, it was concluded that only a crystal cooled to 77~K or lower could yield more information than the room temperature solution. *Osram type XBO-1600, operated with a model MHXM 2500-2S power supply (Christie, Los Angeles 43, California). **EG and G type FX-42. 19

Unpumped. Xenon Emission Lines o Pamped 0~~~~~~~~~~~~~~~~~~~~~ 6450 6150 5850 5550 5250 4950 4650 4350 4050 3750 WAVE LENGTH (A) Figure 11. Photographically Recorded Transmission Spectra of Optically Pumped and TJnpunped Uranyl Glass.

mn 4 Z n 0~_} O ta1ae X O. LL _ c~Ln 4350 4650 4450 5250 5850 6150 6450 6750 o WAVE LENGTH (A) 12. Induced Absorption Cross-Section of Uranyl Glass. (Note: absorption cross-section is not corrected for film response.)

Xenon Emission Lines V) <I r ~~~~~~~~Unpump ed o. Region of Ordinary Uranyl Absorption Pumped 6750 6450 6150 5850 5550 5250 4950 4650 4350 4050 3750 WAVE LENGTH (A) Figure 13. Photographically Recorded Transmission Spectra of Optically Pumped and Unpumped Uranyl Sulfate Solution.

0 z 7 0 I (Po 6 o0 UIQ I0 0 0 0 0 L0. 0 oL I I X I I I I I A, 3750 4050 4350 4650 4950 5250 5550 5850 6150 6450 6750 WAVE ENGTH (A) Figure 14. Induced Absorption Cross-Section of Uranyl Sulfate Solution. O 35 0 45 6 45 25 — 5 c80 65 40 65 a_~~~~~~WV EDOTH(~ ~iue 14XIducdAsrto rs-eto fUay uft oui

An estimation of the number of ions giving rise to this absorption can be made by noticing the difference between the ordinary absorption peaks and dips in the pumped and unpumped spectra. If the average number of ions giving rise to absorptions at frequencies v and v' is N and the absorption cross sections at these frequencies are a and a', then the difference in optical density of the film at these two frequencies will simply be N(a-')2, where 2 is the path length of the sample. If the number of ions decreases, then the difference in optical density at v and v' will also decrease so that the new difference will be N*(a-&')o, Thus, the ratio between the original population and the new (in the pumped condition) population for the ground state can be calculated by comparing the ratio between the differences in optical density between two predetermined points in the pumped and umpumped spectra: N*/N — = x (ODp(v) - ODp(v'))/(ODu(v) - ODu(v')) (1) where OD (v) means the optical density of the film at frequency v for the pumped spectra and ODU(v) means the optical density of the umpumped spectra at the same frequency, Equation (1) holds only under the condition that the number of ions depleted from the ground state is small compared to the number originally there since the actual amount of absorption is proportional to the difference between the population of the excited state and the ground state, not just the population of the ground state alone. If the transition being considered can have an appreciable lifetime, a sufficient concentration of ions could be built up so as to invalidate this result. It is necessary to derive a expression for the excited state population which better fits the conditions of this experiment. There are several assumptions to be made and experimental conditions which must be taken into account: (1) Due to the fact that the flash lamp does not illuminate the sample during the entire time that the camera shutter is open, there is a period of time during the recording of the pumped spectrum in which the ion populations are the same as in the umpumped spectra. This situation arises both before and after the flash lamp fires since the shutter is open longer than the duration of the discharge and the decay of the induced absorptiQn in the uranyl sulfate is very rapid. (See the next section.) (2) A great deal depends on the nature of the upper level involved in the absorption process, If the level has a long half-life, then the disappearance of the ordinary absorption will be proportional to the difference between the upper state and ground state populations, but if the upper state may always be regarded as being empty, the disappearance of the ordinary absorption can be ascribed to the depletion of the ground state population alone, The relaxation time from the other states to the resonance state has been measured as less than 3 x 10-6 seconds by Levshin and SheremetJev9 by noting the delay between the application of exciting light and the onset of fluorescence. This does not mean, however, that no population 24

density in any of these states can be observed because, if the exciting light is sufficiently intense, the pumping rate may be an appreciable fraction of the normal relaxation rate. (3) In accordance with the foregoing considerations, the following assumptions are made: (a) The exposures of both the pumped and unpumped spectra are made with a light pulse which has a square time profile of length At1 + At2, (b) The difference between the ground and excited state populations during the unpumped spectrum is AN and this persists throughout the duration of the exposure, (c) The difference between the ground and excited state populations in the pumped spectrum is broken into two time regions, each of which is square and during the time Atl the difference is AN (the same as in the;unpumped ~case) and during the time At2 the difference is AN*, The ratio of the differences in optical densities between the pumped and unpumped spectra is then given by: AOD' = Atl AN*(a-a'),+At2AN( N -' ), (2) AOD (At +At2)AN( C- ) where AOD' is the difference is the pumped spectrum and hOD is the difference in the unpumped spectrum, Rearrangement of this equation yields: AN*/AN = (l/At)((AOD*/AOD) (Atl+At2)-At2) (3) Table VIII shows the results of the measurements of several of the absorption TABLE VIII CHANGES IN ORDINARY ABSORPTION PEAKS OF SULFATE SOLUTION UPON OPTICAL PUMPING Change in Optical Density (Arb, Units) Wavelengths where Optical Densities Determined Unpumped Pumped AOD1/AOD2 Max, of 4389 A abs. band-Min, between 4389 and 4250 A band. 3.8 3.0.79 Mino between 4389 and 4250 A band-Max. of 4250 A band, 709 5.8.74 Max, of 4250 A band —Min, between 4250 and 4127 A band, 7.6 5.3.70 Min, between 4250 and 4127 A band-Max, of 4127 A band, 4,2 2,4 ~57 Max, of 4127 A band-Max, of 4010 A band, 10.0 9.0.90 Max, of 4010 A band —Max. of 3895 A band. 11o5 10.4.90 25

peaks and dips in the pumped and unpumped spectra. We shall assume the following values for the variables appearing in Equation (3): Atl = 0.5 msec. At2 = 0.5 msec. AOD*/AOD = 0.77 which yields the result: AN* = 0.54 AN. If we further assume that there is no population in the upper states, then the population of the ground state has therefore been decreased by 46% since AN* equals the ground state population. The population of excited ions (all of which have been assumed to reside in the lowest vibrational level of the lowest excited electronic state) is then in the neighborhood of 0.0565 molar. The experimental evidence for the exact population is admittedly open to question since the range of ratios in Table VIII is quite large. Complications also arise because of the nonlinearity of the film which may be the cause of the large variation seen there. The mean deviation for the optical density ratios is 0.13 which would give the limits of the ground state depletion as 20% to 72%. This latter figure is quite high but it is certainly reasonable to suppose that the population in the resonance level is within 50% of the previously calculated value. 26

D. TIME-RESOLVED INDUCED ABSORPTION IN URANYL GLASS AND URANYL SULFATE AND PHOSPHATE SOLUTIONS Using the experimental setup described in detail in Appendix II, the induced absorption spectra of a 10-mm thick section of uranyl glass, a 10-mm path length solution of uranyl sulfate in concentrated sulphuric acid (4.7 x 10-2 gm/ml) and a 10-mm path length solution of uranyl phosphate in 80% phosphoric acid (7.7 x 10-2 gm/ml) were recorded on Polaroid type 107 film using an RCA lP21 photomultiplier as a detector and the Jarrell-Ash 3.4 meter Ebert spectrograph as a monochromator. Some of these time-resolved traces are given in Figures 15 and 18 for the uranyl glass and the sulfate solution respectively. (The results for the uranyl phosphate are very similar to those for the sulfate, so none of these are included.) The photographs in Figure 15 were taken in the following manner. (1) The scale was put on the film with a rather low level of illumination for the grid and exposure settings of 1/10 sec and f/l.9. The grid illumination was then turned down and the rest of the photograph taken with the aperture stopped down to f/22 and the trace intensity adjusted to give a good record. The purpose of this process was to insure that the grid intensity would always be the same, regardless of the manner in which the rest of the photograph was exposed. (2) The record of the induced absorption was then made by firing the flash tube with the cammera shutter held open in the "bulb" position and the oscilloscope sweep set to trigger off an input signal taken directly from the flash tube trigger. Only one beam of the oscilloscope was used and the signal from the photomultiplier monitoring the xenon arc light and the signal from an RCA 933 photodiode monitoring the output from the flash lamp was fed into the two inputs of the dual trace plug-in amplifier which was operated in the chopped mode. The lowest trace on the photograph is the record of the light emitted by the flash lamp; an increase in light signal causing an upward deflection of the beam. The middle trace on the photograph is the xenon arc light as monitored by the photomultiplier. The sense of the beam deflection is reversed from the normal, however, since a downward deflection of the beam corresponds to an increase in the photomultiplier output signal. Thus, in Figure 15, all of the photographs from 4800 A to 6400 A show a decrease in the xenon arc light when the flash lamp is fired since the beam deflects upward. The appearance of the "increased" xenon arc light at 4400 A is dealt with below. (3) The record of the light scattered from the sample and detected by the photomultiplier was then made by blocking the xenon arc light from passing througn the sample with an opaque shutter and firing the flash lamp a second time under the same conditions as before. The flash tube capacitor bank was not always fully charged the second time so the flash light trace is sometimes slightly different in this second exposure than in the first giving the appearance of a doubling in the lowest trace. Were there no flash lamp light scattered from the sample, the uppermost trace would have been a straight horizontal line since no lignt would have been present. The fact that there is a downward deflection of the beam indicates that a small amount of light is being scattered from the sample. 27

4400 A 4800 A 5200 A 5600 A 6000 A 6400 A Figure 15. Timne- soie1d Induced Absorption in Uranyl Glass. (Time scale is 2 msec./div.) 28

0.7 0.6 ow, 0.5 v 0.4 uV 0 _) 0.2 < O 0.1 0 64oo00 6200 6000 5800 5600 5400 5200 5000 480 4600 0o WAVE LENGTH (A) Figure 16. Induced Absorption Cross-Section for Uranyl Glass as Deduced from Photoelectric Measurements (versus wavelength).

0.6 0o 0.5.OL L <0. I 0.1 l l0 20..000 20,00 18,000 19, 000 1,7,0000 ENERGY (CM-I) Figure 17. Induced Absorption Cross-Section for Uranyl Glass (versus energy).

5200-5600 A 5700-6100 A 6200-6600 A Figure 18. Time-Resolved Induced Absorption in Uranyl Sulfate Solution. (Time scale is 10 msec./div.) 31

The photographs for the uranyl glass were taken every 200 A since no structure was expected to appear of a spacing closer than this and the minimum percent of transmission at each wavelength determined by measurement of the difference between the uppermost and middle tracings at the maximum of the middle trace and dividing by the spacing between them when no induced absorption or scattered light was present, i.e., at the far right of the photograph. (The gain of the oscilloscope and the aperture of the sample were adjusted to give a nearly constant signal at each of the different wavelengths.) These results were then converted to optical density and plotted versus wavelength in Figure 16 and versus energy in Figure 17. The appearance of the plateau in the decay of the induced absorption is attributable to the presence of a second hump in the flash tube light output —the circuit was "ringing." This disappeared when a second inductor which quenched the ringing was substituted. Because of the high ordinary absorption in the sample, the photograph at 4400. in Figure 15 had to be taken with a high gain on the oscilloscope so that the scattered light appears to be larger at this wavelength than at the others. Such is not the case, however, and this scattered light signal is more than enough to make up for the diminution of the xenon arc signal due to the induced absorption so the middle trace of this photograph deflects downwards instead of upwards as in the rest of the photographs. There is just a barely detectable induced absorption at this wavelength as can be observed by subtracting the uppermost trace from the middle one and observing that the middle trace does deflect upwards once this source of noise has been accounted for. The photographs of the uranyl sulfate solution were taken in exactly the same manner as those of the uranyl glass with but one exception: in order to conserve film, the sweep time of the oscilloscope trace was lengthened to 10 msec/cm and the time base operated in the "strobe" mode which brightens any particular preselected portion of the trace. In this manner, all of the sweep can be eliminated from the photograph except this brightened portion since the beam intensity can be adjusted so that all the rest is too weak to record on the film. The beam was then positioned so as to start in the middle of the oscilloscope face and the camera moved between exposures. In this manner, five separate exposures can be made on each film. A survey every 100 A was first made and then a series every 10 A in the region from 4800 to 5270 a and every 20 A in the region from 5350 to 5530 A to check for finer structure. None was detected so it was concluded that its appearance is only an artifact of the photographic film. The profiles of the induced absorption in the sulfate and phosphate solutions were determined in the same manner as was that of the glass and these are plotted in Figures 19 and 20, respectively. The general outline of the band in all three media is basically the same but the oscilloscope photographs give more detail than just the intensity of the absorption. One immediate discernible difference between the glass and sulfate traces is the prominence of the second hump in the induced absorption decay. As stated before, this is due to the presence of a second hump in the flash tube output (more prominent because of the condensation of the time axis in Figure 18 but of equal magnitude). The decay of the absorbpion induced in the glass is so slow that this 32

0.26 0.24 00222: L 0.20 I L/) v LA 0.18 t 0.16 o 0.14,, 0.12 O 0.08 0.06 0.04 o 0.02- 0.00 6600 6400 6200 6000 5800 5600 5400 5200 5000 4800 WAVE LENGTH (A) Figure 19. Induced Absorption Cross-Section for Uranyl Sulfate Solution as Deduced from Photoelectric Measurements (versus wavelength).

0.7 O _ 0.6'r) 0,>- 0.4 Zu,. —,,- 0.32 6200 6000 5800 5600 5400 5200 5000 0.1-_ / WAVE LENGTH (A) Figure 20. Induced Absorption Cross-Section for Uranyl Phosphate Solution as Deduced from Photoelectric Measurements (versus wavelength). 54

second pumping appears only as a plateau in the induced absorption curve while in the sulfate the relaxation is so much faster that the second hump is very prominent in the induced absorption curve. This difference in decay time evidences itself in other ways. The effect of the integration of the pumping light in the glass can be seen in the fact that the maximum of the induced absorption occurs slightly after the peak of the flash lamp while the sulfate, which apparently has a much faster relaxation time has its peak absorption at the same time as the flash lamp. The differential equation describing this process is dN*/dt = N a I(t) - a N* (4) where N* is the concentration of excited ions, N is the concentration of ground state ions, P is a factor dependent on pumping efficiency, ioe., geometry and size of sample, flash tube, and reflector and intensity distribution of the pump lamp and absorption cross-section of the sample as a function of wavelength, etc., and a is the decay time constant. Assuming that the ground state population does not change during this process, the solution for this equation is t N*(t) = N a edt I(t')e dt (5) 0o where it is assumed that N*(t = O) = O. The assumption that the ground state population does not change is not a particularly good one at these high pumping intensities. The solution obtained when taking this into account is -t pt At" N*(t) = N +t+ (t)e dt,at x/ I(t')e e o Cet 1. dt' e[at+pJ I(t')e t dt] 0dt o (6) where No is the total number of ions. The right hand side of (6) is seen to converge to that of (5) if the integral fI(t')e dt 0o is small. As an example of the behavior of the function in (6), suppose the pumping intensity is assumed to be a square wave extending in time from O to T and is of amplitude I. Then (6) reduces to: 35

(a+BI)t NODI e -1 N*(t) = (2+I)t O < t <T (7) and NON* = (e+PI)TA )-CZ(t-T) N*(t) +I e (cz+PI)T e, t > T (8) Figure 21 illustrates the behavior of this solution. The characteristics of these curves which are of interest is that the faster the decay time of the excited state, the sooner the maximum population is reached and the smaller it is. These are exactly the phenomena observed by contrasting the induced absorption traces of the glass and sulfate solutions: (1) the maximum of the induced absorption in the sulfate solution occurs before that of the glass, and (2) the induced absorption in the sulfate is smaller (for roughly the same number of ions). In order to determine the decay constant of the induced absorption, the same experiment was run but with an inductor which did not cause the flash tube to ring. This was an air core inductor having a calculated inductance of 700 t-henries. (No attempt was made to optimize the discharge circuit for critical damping or to measure the inductance of the circuit.) The induced absorption traces for uranyl glass and uranyl sulfate are shown in the normal perspective in Figures 22(a) and 23(a), respectively. In order to examine that portion of the induced absorption trace which occurs after the pump lamp has cut off (and thus to see the pure exponential decay of the absorption), the same photograph was then repeated using thle "strobe" mode of triggering so that only a portion of the trace has sufficient intensity to be visible on the film. Part (b) of each of these figures shows that portion of the curve ) which is to be expanded for better time resolution. Part (c) then shows the decay curve recorded at a higher sweep speed. The sweep speed of the fastest traces of the sulfate absorption is fast enough to show the chopping between the beams. The intensity of the light transmitted through the sample is assumed to follow Beer's law I= Ii e-N c (9) where I is the transmitted intensity, Ii is the incident intensity, N is the number of active centers/cm3, a is the absorption cross section at a particular frequency, and 2 is the path length. The number of active centers, however, varies as a function of time and this is assumed to be a pure exponential decay: -Dt N = No e (10) Substitution from (10) into (9) yields:

50 45 | ---, r.o-. _.,... 40. ~35 - 30. 25 Un, 10 0 1 _ 3 45 67 TIME (ARBITRARY UNITS) Figure 21. Theoretical Time-Profile for the Excited State Population Assuming Square-Wave Pumping. 37

(a) (b) Figure 22. Induced Absorption Decay Measurement for Uranyl Glass. (a) Normal sweep, 0.5 msec./div. (b) Strobe sweep showing region to be expanded. (c) Delayed sweep, 0.2 msec./div.

(a) (b) (C) Figure 23. Induced Absorption Decay Measurement for Uranyl Sulfate Solution. (a) Normal sweep, 0.2 msec./div. (b) Strobe sweep showing region to be expanded. (c) Delayed sweep, 0.05 msec./div..=_~~3

I = Ii exp (-ai No exp (-nt)) (11) At time t = 0, let I = Io so that (11) becomes Io = Ii exp (-ca No) or: ac No = -2.303 logo ((12) Equation (11) may be rearranged to: I -2.303 log10o = a2 No exp (-nt) (13) and substitution from (12) yields: Exp (-nt) loglo( I) log10 (I (14) Equation (14) can also be written as: Pt = -2-303 logo (log ( log1o (15) i.e., the rather complicated expression on the right hand side should be a linear function of the time. Equation (15) is plotted in Figure 24 for the data obtained from Figure 22 for the uranyl glass. The agreement is very good indicating that the decay of the induced absorption is purely exponential. The time constant is 429 >sec. The data for the sulfate solution is much poorer, however, but data from Figure 23 indicates a decay time of 52 ~isec. (This figure is probably larger than the true decay time because measurement of the decay of the absorption induced in the sulfate solution must be made while the flash tube is still pumping the sample and this will give an apparent lengthening of the decay time. As can be seen from Figure 23, the induced absorption is practically zero by the time the flash tube has finished firing.) Coupled with the result from the photographically recorded spectrum that approximately 50% of the uranyl ions participate in the induced absorption, it is possible to calculate the molar extinction coefficient and the oscillator strength for this transition in the case of the sulfate where the concentrations are known. Based on the previous calculation of the population of the resonance level, the molar extinction coefficient at the maximum of the induced absorption curve at 5700 ~ is 5.65 and the calculated oscillator strength is 6.54 x 10-. As remarked before, these values may be as much as 50% in error. 40

0.8- o 0.6 / (.),~ 0.4 Wavelength = 5800 i 0.2 - 0.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 - 1oglo(log10 ( Il i) 1og0o(1 I!i )) Figure 24. Induced Absorption Decay for Uranyl Glass. 41

E. TIME-RESOLVED FLUORESCENCE IN URANYL GLASS AND URANYL SULFATE SOLUTION; COMPARISON OF DC- AND FLASH-EXCITED FLUORESCENCE SPECTRA OF URANYL SULFATE TRIHYDRATE The induction and decay of the fluorescence of a 5-mm thick section of uranyl glass and a 10-mm path length solution of uranyl sulfate solution in concentrated sulfuric acid (4.7 x 10-2 gm/ml) were recorded using the same apparatus as in the case of the time-resolved induced absorption measurements. The physical setup, however, was different because the flash head was positioned differently on the optical bench so that the sample was at a focus point. (Appendix II describes the optical path and how this can be accomplished.) The samples were excited with the xenon flash lamp as in the case of the induced absorption measurements but it was necessary to impose a 3-mm thick section of deep blue Corning No. 5-57 filter glass between the sample and the flash lamp to avoid picking up the flash lamp light which scatters off the sample and which is at the same frequency as the fluorescence. The photographs of the fluorescence decay were made at the wavelengths of the peaks of the emission bands at room temperature, at 5180, 5327, and 5534 A for the glass and at 4915, 5150, and 5405 A for the sulfate solution. These are shown in Figures 25 and 26. These photographs are mounted upside down from the position in which they were taken so that time proceeds from right to left. The upper trace is a record of the light emitted by the flash tube with an increase in output corresponding to a downward deflection and is the result of the superposition of several traces. Several exposures were made on each photograph, the only change being made from one exposure to the next being the vertical gain of the lower beam which displays the fluorescent light signal. The gains employed for the lower beam are listed at the bottoms of the figures. Comparing the two sets of traces makes the faster decay time of the fluorescence in the sulfate solution very evident and, in fact, while it was possible to determine the decay rate directly from these photographs for the glass, the sulfate solution required the use of the technique employed in measuring the induced absorption decay: delaying the sweep and expanding a portion of the decay curve after the pump lamp has cut off. Figure 27 illustrates this process. Since the decay rate measurements were performed an appreciable time after the pump lamp had cut off, no short-lived second component of the fluorescence decay such as that observed by BillingtonlO is expected to be detectable. The decay should be a pure exponential and this was verified by plotting log (intensity) versus time. This is shown in Figure 28 for the glass and Figure 29 for the sulfate. The decay constant for the glass is 458 tsec which is in good agreement with the induced absorption decay measurement of 429 isec. The agreement between the decay constants of the fluorescence and induced absorption in the sulfate solution is much poorer, being 13.8 and 52 isec, respectively, but this is due to the inaccuracy of the induced absorption measurement. 42

(a) (C) (b) Figure 25. Time-Resolved Fluorescence for Uranyl Glass. (a) 5180 A, lower scales are 5, 2, 1 V/div. (b) 5527 A, lower scales are 5, 2, 1 V/div. (c) 5534 A, lower scales are 5, 2, 2, 1, 1/2 V/div.

(a) (b) Figure 26. Time-Resolved Fluorescence for U02S04 in H2S04 Solution. (a) 4915 A, lower scales are 5, 2, 1 V/div. (b) 5150 A, lower scales are 5, 2, 1 V/div. (c) 5405 A, lower scales are 5, 2, 1, 1/2 V/div. 44

(a) Figure 27. Fluorescence Decay Measur'ement for U02S04 in H2S04 Solution. (a) Strobe operation showirlg r,.gon to be expanded. (b) Expanded sweep,.0j mse. /div. _ _ e4

0.8 0 0.6 LAJ 0.4A 3- 5180 i s / IA- 5327 i I 0 - 5534 i 0.2 0.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.7 - 1ogo( I / Io ) Figure 28. Fluorescence Decay for Uranyl Glass.

250 225 200 175 g) 150 = 125 L 100 t - 75 S v Wavelength = 5150 i 50 - 25- 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 - log 10 ( / )io Figure 29. Fluorescence Decay for U02SO4 in H2S04 Solution.

In order to establish whether or not there could be any appreciable population of the uranyl ions in the excited vibrational states of the first excited electronic state, the fluorescence spectrum of polycrystalline uranyl sulfate trihydrate was recorded at 770K on Kodak Pan-X film using the 1.5 meter B & L spectrograph. The first fluorescence spectrum was recorded using the DC xenon arc as the exciting source and the second using the xenon flash lamp. In both cases the exciting light was filtered using the Corning No. 5-57 filter. The microdensitometer tracings of these spectra are shown in Figure 30. There were, of course, no differences between the spectra which indicates that, despite the higher transition probability from the ground state to the v + 1 level of the first excited electronic level (as evidenced by the absorption spectrum), the population of this state is so low during the time the sample is pumped with the flash Lamp that its fluorescence is negligible. 48

A I \O ~ ~ ~ ~~~~~; 0 Flash-excited Excitirn?- Linght.F- X Figure 30. DC- and Flash-Excited Fluorescence Of U02SO4~3H20) 770K.

F. ULTRAVIOLET ABSORPTION SPECTRA OF URANYL SULF'ATE SOLUTION AND URANYL P HOSPHATE SOLUTION The absorption spectra of a solution of uranyl sulfate in concentrated sulphuric acid (2-77 x 10-4 gm/ml) and uranyl phosphate in 80% phosphoric acid (4*66 x l0-4 gm U02C03/ml) were taken in a 10-mm path length cell on a Cary Model 15 spectrophotometer in the region from 3000 to 1900 X using the pure acids as references, The acids themselves were also run using air as a reference to establish how good their transmission was in this region and therefore how good the resolution of the spectra would be. The sulphuric acid did not show any appreciable absorption above 1900 A so the resolution should be near that of the manufacturers ratings (0.3 cm-l). The phosphoric acid, however, is not well suited for use in cells with such a long path length since by 2300 A it has an optical density of 0.5 and this rises to about 0.75 at 1900 A. Unfortunately, shorter path length cells were not available so that the resolution in the phosphate solution is poor below 2400 A. The results of these measurements are presented in Figures 31 and 32 for the sulfate and Figure 34 for the phosphate. The results for the sulfate are different from those of previous authors, notably Henri and Landau,ll Ahrland 12 and McGlynn and Smith,13 in that this ultraviolet absorption band is shown to have a definite upper limit and its peak lies at 2200 A. There are also two longer wavelength shoulders at 2600 A and 2830 A which are barely discernible on the tracings and are very apparent only if the wavelength scale is greatly compressed. The phosphate spectrum shows no such maximum and this is probably due to the poor transmission of the phosphoric acid in this region. In fact, the same behavior, ioe., the indicated absorbance of the sample monotonically becoming greater above 2200 A was also observed in the case of the sulfate when the same experiment was run on a Cary Model 11 spectrophotometer which had a range of usefullness limited to 2300 A. This behavior seems to be linked with the performance of the machine when the source intensity drops to too low a value so that scattered light becomes important. Instead of two longer wavelength bands as in the sulfate, only one is present at about 2500 A but the relative intensities of the bands may be shifted sufficiently so that the longer wavelength band, which was marginal to begin with, is no longer discernible. Table IX shows the molar extinction coefficients and absorption crosssections of these bands for the sulfate and the phosphate. The oscillator strength of this higher system (treated as a single band) is much larger than for the visible system but is not as large as that derived by Rabinowitch and Belford, The oscillator strength calculated using the data obtained here for the entire UV system obtained by Simpson's rule integration of the absorption curve is 1.77 x 10-2 in contrast to the previous authors' value of 7.89 x 10-2. Assuming that the degeneracy of the upper state may be regarded as the

0.30 0.25 (Arrows indicate the positions of the long wavelength shoulders) c 0.20 V) 0,1 C-) 0.10 S 0.05 0.00 2000 2100 2200 2300 2400 2500 2600 2700 2800 2900 3000 WAVE LENGTH (A) Figure 31. Ultraviolet Absorption Spectrun of U02S04 in H2S04 Solution (versus wavelength),

0.30 0.25 0.20 < 0.15 C) I- -- 0.10 0.05 0.00 49 47 45 43,41 39 37 35 33 ENERGY ( CM-' x 103) Figure 32. Ultraviolet Absorption Spectrum of U02S04 in H2S04 Solution (versus energy).

0.30 (The arrows indicate the positions of the shoulders in the UV absorption spectrum) 0.25 0.20 (V) C~ 0.15 \J1 00.10 0.05 1 0.00 49 47 45 43 41 39 37 35 33 ENERGY ( CM-'x 103) Figure 33. Resolution of the Ultraviolet Absorption Spectrum of U02S04 in H2S04 Solution by Subtraction of the Induced Absorption Contour.

0.35 0.30 0.25 0.20 LLU Shoulder o 0.15 I XL |MT.aximum of Induced I Absorption Curve 0.10 0.05 0.00 2200 2300 2400 25000 2600 2700 2800 2900 3000 WAVE LENGTH (A) Figure 34. Ultraviolet Absorption Spectrum of U02HP04 in H3P04 Solution Showing the Relationship Between the Long Wavelength Shoulder and the Induced Absorption Maximum.

TABLE IX ULTRAVIOLET ABSORPTION BANDS OF URANYL SULFATE AND PHOSPHATE SOLUTIONS Uranyl Sulfate Uranyl Phosphate %(AO.) c a(cm2 x 10-23) E V(m2 x 10-23) 2200 500 82.5 2600 146 24.1 2830 48 7.9 2500 117 18.8 same as the ground state, the mean intrinsic lifetime for relaxation from the UV system to the ground state is 4.2 x 10-8 seconds, in contrast with the previous authors' value of 1.63 x 10-8 seconds. If the initial state of the fluorescence is the initial state of the induced absorption, as the close correspondence between the decay rates of the fluorescence in uranyl glass and uranyl sulfate solution and the decay rates of the induced absorption in the corresponding media seems to indicate, then the sum of the resonance fluorescence energy and the energy of the peak of the induced absorption band should give the energy separation of the higher excited state from the ground state of the uranyl ion. In the case of the uranyl sulfate, the resonance energy is 20,320 cm-1 and the peak of the induced absorption is at 17,240 cm-1l (5800 A) so the energy sum is 37,560 cm-1, ioe,, there may be an absorption at 2660 A. In the case of the uranyl phosphate, the resonance energy is 19,860 cm-l and the peak of the induced absorption is at about 17,540 cm-1 (5700 A) so the energy sum is 37,400 cm-1 and the expected absorption should occur at 2670 Ao The agreement between the calculated positions of the previous bands and the positions of the shoulders in the ultraviolet absorption spectrum is quite satisfactory. It indicates that the transition probabilities for transitions from the first excited electronic state to the UV states must be quite different from those for transitions from the ground state to the UV states since the profile of the induced absorption does not all follow the profile of the upper band as given by the ordinary absorption spectrum. The induced absorption "eats a hole" in the UV band profile in the case of the sulfate and shows that the UV band must indeed be resolved into three bands as the shoulders would indicate. This is illustrated in Figure 33 which was obtained by adding 20, 230 cm-1 to all the frequencies of the induced absorption profile, normalizing the curve thus obtained to agree with that of the ultraviolet absorption curve at 37,470 cm-l (the maximum of the curve), and then subtracting the 55

resulting curve from the ultraviolet absorption curve. This method of analysis assumes, of course, that the Frank-Condon profile of the UV level appears the same whether the transition to it occurs from the ground state or from the resonance level, i.e., that there is no great shift in the internuclear distance between the minimum of the potential curve in the first excited state and the minimum of the potential curve in the ground state and that these two have the same shape. This assumption is not too ill-founded on experimental evidence. The ordinary absorption spectrum shows that there must be some expansion of the uranyl molecule in the first excited electronic state since the 0 ~ 0 band is less intense than the 1 - 0 or 2 ~ 0 bands. The 2 + 0 band is more intense than the following bands (of the same electronic state) indicating that the minima of the two potential curves cannot be too different, It is therefore reasonable to assume that the maxima of the absorption curves obtained may be slightly different but not enough to cause confusion as on which of the three UV levels the induced transition terminates. The remaining portion of the UV bands of wavelengths longer than 2700 A is resolved into a band whose peak agrees with the shoulder seen in the ultraviolet absorption curve at 2830 X while the profile of the band to shorter wavelengths remains nearly the same but assumes a more symmetric shape. A few photoelectric measurements were conducted on uranyl sulfate solution at wavelengths shorter than 3700 A to see if there was any indication of a higher induced absorption band but the results were entirely negative. Experiments conducted elsewhere15 have shown that no absorption exists at the neodymium laser wavelength in uranyl glass which would correspond to a ground state to excited state separation 27,360 cm-l or an ultraviolet absorption at 3650 A which adds additional confirmation to this assignment. 56

Thus it seems well established that, unless by some coincidence there is another level at 38,000 cm-l which is not detectable from the ground state, the induced absorption has as its initial level the resonance level and as its terminal level the middle one of the three ultraviolet levels. Figure 35 shows the relative positions and oscillator strengths of the transitions discussed here. The Oscillator strengths of the three main transitions shown in Figure 35, coupled with previous results on the absorption and fluorescence spectra of uranyl compounds, permit some conclusions to be drawn concerning the nature of the states involved. The ground state of the uranyl ion is almost certainly 1Z, (see Appendix I) and the agreement between the fluorescence and absorption resonance line frequency measurements (Dieke and Duncan3) precludes the possibility of the transition from the ground state to the first excited electronic state being vibronically induced so that the transition is Laporte-allowed and the first electronic state must be ungerade. The intensity of this transition, however, is so low that it must be forbidden on some grounds, the most likely possibility being that the state is a triplet and the transition is spin-forbidden. Accordingly, the first excited electronic state may be designated 3x, where X = T -, A, ~, j as no low-lying 3L states are to be expected. (See Appendix I.) The oscillator strength of the transition from the ground state to the center UV level is three orders of magnitude higher than that of the first electronic transition so it is unlikely that it is forbidden overall, i.e., electronic x vibrational, on either spin or Laporte grounds. However, the oscillator strength is still not high enough to correspond to a fully allowed electronic transition. One explanation for this situation would be that the transition is vibronically allowed, and "borrows" intensity from a |j Iu lying some 7000 cm-1 above it. The electronic state is presumed to be a singlet and gerade, the only low-lying possibilities for these being the 1Z, ]Tg, and 1/\ derived from the configuration (br) (l3 g)l Each of these states, coupled with the proper vibration of the uranyl ion, can give rise to a vibronic level whose symmetry is luu all of them coupled with a iu vibration and only the lg coupled with the au vibration. Since the first excited electronic state is assumed to be an ungerade triplet (or a component of one), a transition from it to a l'F,, vibronic state would be forbiaden on uisn basis oi ooTn the Laporte rule and spin selection rule. Accordingly, it would be extremely weak. However, if the induced absorption transition terminates on the presumed pure electronic level, then the transition is 3xu (or one component of it)+Yg and is forbidden only on spin grounds and would be of comparable intensity to the transition from the ground state to the first excited electronic state. This agrees with the experimental result. There should, accordingly, be a shift in the apparent position of this center level depending on whether the transition to it occurs from the ground state or from the resonance level. The measurements made in this study are not 59

ENERGY IN CM-(x 103) 46 46 ______ ----- ('Tru) 44 42 ULTRAVIOLET SYSTEM 38 (Xgrug,'g) 34 32 f = 6.54 x105 28 26 24 22 | / I FIRST 1 I - A d ~EXCITED 20 L ELECTRONIC 1 (19)R8 STATE 16.-1.77xo-~3 ( 3 3 3 Fu, Au Pu, I 14 12 f = 2.29x10'3 f: I.28 x I0-6 ~~~10 f=l. 25x 10-5 8 6 4 2 0 ('Z+)GROUND STATE Figure 35. Principle Transitions Observed for the Uranyl Ion. 60

accurate enough to establish this difference but it should be possible to test these assignments with crystal spectra. As mentioned by McGlyan and Smith13 (see Appendix I), the lowest excited electronic level shows no Zeeman splitting and probably corresponds to a 3A1 state.

APPENDIX I DISCUSSION OF THE MOLECULAR ORBITAL REPRESENTATIONS FOR THE URANYL ION AND THE ASSIGNMENT OF TERM LEVEL DESIGNATIONS TO THE LOWER URANYL ION LEVELS The symmetry of the uranyl ion, ignoring the presence of the secondary ligands, is Emh with the two oxygen nuclei and the uranium nucleus colinear and the oxygen nuclei equally displaced from the uranium nucleus. This configuration has been verified by x-ray diffraction in several salts.l6,l7 The symmetry operations of the molecule are E (the identity), C(p (rotation around the internuclear axis through an angle pp), av (reflection through any plane containing the internuclear axis), i (inversion through the uranium nucleus), iCCp (any improper rotation around the internuclear axis), and iav = C2 (rotation through the angle Ec about any axis perpendicular to the internuclear axis and intersecting it at the uranium nucleus). The character table for this symmetry group is given in Table X. Because of the symmetry about the internuclear axis, the projection of the angular momentum along this axis is a good quantum number, designated Q, and, in the absence of spin-orbit coupling, so are the projections of the orbital angular momentum, designated A, and the spin, designated Z. Such a state is specified in the spectroscopic notation as an+d Q,,u or g and it is useful to find the correspondence between these states and the representations listed in Table X. Such a correspondence has been given by Rabinowitch and Belfordl4 but there is some confusion in their table regarding the assignments of the 3Z states and the 3V0 and 3T2, 3 1, and 3/3, etc., which they list in the form 3TTO 2 under one representation and as 720 under another. TABLE X THE CHARACTER TABLE FOR DIh (After 14) D0h E Cp Cov i iC p C2 Al + g+)2 2 2 A.g(oa) 1 1 1 1 1 1 z,x + y Alu(Oa) 1 1 1 -1 -1 -1 A2g( cg) 1 1 -1 1 1 -1 Iz A2u(l+) 1 1 -1 -1 -1 1 z Elg(ng) 2 2coscp 0 2 2coscp 0 xz,yz,Ix,Iy Elu(tu) 2 2cosp 0 -2 -2coscp 0 x,y E2g(Sg) 2cos2cp 0 2cos2cp 0 x2-y2, xy E2u(6u) 2 2cos2cp 0 -2 -2cos2cp 0 E3g(Cpg) 2 2cos3cp 0 2 2cos3cp 0 E3u((Pu) 2 2cos3p 0 -2 -2cos3cp 0 62

The assignment of spectroscopic terms to the representations is very easy since the representations are already characterized by Q. For instance, the representations Alg...A2u which have constant characters under Ccp must correspond to terms with Q = O and these can only be 1Z, 370 or 3Zo- Similarly, the El representations must have Q = 1 and therefore correspond to -Z,, 371, or 3, 1. Table XI gives the representations and the spectroscopic terms which are appropriate to them. The assignment of the triplets to the representations can be made more rigorous by observing that the direct product space of a singlet with that representation which transforms like a rotation about the z-axis is reducible to the representations of the corresponding triplet state and the same value of Q. Similarly, the product space of the direct product of the singlet representation with the representation which transforms like a rotation about the x- or y-axis is reducible into the representations of the triplet state whose Q value is 1 removed from the Q value of the singlet state. From Table X, these representations are A2g and E-g, respectively. For example, the Q = 0 component is corresponding to the l., obtained by the direct product A2gX Alg which is, of course, A2g so that the term 3Og is assigned to A2g. TABLE XI ASSIGNMENT OF TERMS FOR A MOLECULE OF SYMMETRY ID0h Representation Singlets Triplets A'.g g 7 O Og A2g ZU 3W- 3ZE Aog 1 - 3T- 3-7+ It. 3 f+ 3ulu 37 u Ou E2g 1' 3TT2g A2g E2u Ln, 32 3Au 32 2u E3 3 g g 3 E3u lu 3R 3u u 63

A complete discussion of the attempts to explain the visible and ultraviolet absorption and the visible fluorescence spectra for the uranyl ion is given by Rabinowitch and Belford (op. cit.) and only a brief review of their work will be given here. The molecular orbitals are to be constructed from linear combinations of the unoccupied 5f, 6d, and 7s orbitals of the uranium and the occupied 2s and 2p orbitals of the oxygens. The representations to which these orbitals belong are listed in Table XII where it is assumed that the z-axis of the atomic wave functions is coincident with the internuclear axis. TABLE XII ASSIGNMENT OF THE URANIUM AND OXYGEN ORBITALS TO THE REPRESENTATIONS OF THE Dooh GROUP Representation Uranium Orbitals Oxygen Orbitals A^g g(_g) 7s, 6do 2s A.ju(a ) 5fo 2pO A2g(ag) A2U( u) Elg (1g) 6dl, 6d_1 Elzu(tu) 5fl, 5f_1 2pl, 2P_1 E2g (g) 6dl, 6d_, E2u(6u) 5f1, 5f-_ E3g( CPg) E3u(CPu) 5f3, 5f-3 The symmetry for the oxygen nuclei alone is also DIoh so the only linear combinations of these which preserve this symmetry are permissable. These are listed in Table XIII. The designation 2po + 2p0o means a 2po orbital centered on oxygen nucleus (1) and one centered on oxygen nucleus (2) and the sense of the z-axis of these two wave functions being the same along the internuclear axis. The designation 2pO - 2pO means that the senses of the two z-axis are reversed (or, in the case where the wave functions are gerade, the sign of one of the wave functions has been reversed). 64

TABLE XIII ASSIGNMENT OF THE OXYGEN LINEAR COMBINATIONS TO THE REPRESENTATIONS OF THE Dooh GROUP Representation Oxygen Linear Combinations Ajgg(og) 2s + 2s, 2po - 2po Aiu(a) 2s - 2s, 2p0 + 2po A2g( ag) A2U( a+) Elg(~g) 2pl - 2pl, 2p.1-2p_1 Elu(Iu) 2pl + 2pl, 2P-1+2P_1 There are, accordingly, 4d+ states possible, constructed of linear combinations of 7s, 6do, 2s + 2s, and 2pO - 2Po, 3au states constructed of linear combinations of 5fo, 2s - 2s, 2po + 2po, 2Ag states constructed of linear combinations of 6dl and 2pl - 2pl or 6d-1 and 2p_1-2pl, and 2,u states constructed of linear combinations of 5fl and 2pl + 2pl or 5fl and 2pl1+2pl. To construct the molecular ion (OUO), 16 electrons must be put into the above one-electron orbitals and Rabinowitch and Belford give the ground state configuration as: (. )2(dg) Oru) (gg) This does, of course, account for only 12 electrons, not 16, but they have implicitely assumed by the use of hybrid sp orbitals that there are two occupied a orbitals below all of the ones above. (These can be obtained by taking the symmetric and antisymmetric linear combinations of the orbitals which these authors designate as sp(I) and sp(II).) The complete configuration is: ( )~2 ( 1: )2 U ( 2-aU)2 ( 2-g ) u(g This configuration gives a 1 state as the ground state which agrees with the apparent diamagnetism of the ground state. The excited configurations are assumed by these authors to rank in energy as: 65

The spectroscopic terms which can arise from each of these configurations are: (lg)4: g Ugg) 3(1 pu): fn(5,4,3,)Uj I4U 3& 3, 2.9ly Utjn6 (,gg)3(16 )1 3 331(4 3 2)TUs1E,120)U} Flu It is assumed by McGlynn and Smith13 that the states giving rise to the absorption in the blue are either the components of the 3/U or 3TU since these are spin-forbidden transitions and would most likely have the low transition probabilities observed for these transitions from a ground state. The level which gives rise to the absorption in the ultraviolet is assumed to be 1TU since a fully allowed transition must be present to give the large transition probability observed for this transition. There is, however, conflicting evidence for the nature of the lower levels since the Zeeman splittings do not agree with the lower levels being the three components of a single state. The splitting due to a magnetic field is given by 212A + 2IP[ io Hz and the 0, 1, and 2 components of a s1 state should split in the ration 1: 1: 3 while the 1, 2, and 3 components of a 3L state should exhibit no splitting for the 1 component, and a ratio of 1: 2 for the 2 and 3 components. It has been experimentally observed that only the middle electronic state of the three which are assumed to make up the visible electronic absorption system splits. It is therefore clear that the assignments of these levels is far from complete. It is of interest to examine the Dch double group, not because it applies to the uranyl ion, but because the next actinyl ion, neptunyl, and the second one after that, americyl, must belong to this group. The irreducible representations of this group, because it is infinite, cannot be derived in as simple a manner as those of the finite groups and it is necessary to return to first principles and the full rotation group. (A complete description of the method is given by Heinel8.) The symmetry operations of this double group will be the same as those of the single group with the addition of: 9 (rotation through 21), Cq, av6, i~, and iCcq6. (Since C2 is single-valued, C26 and C2 are in the same class.) Consider first the representations having a = 1/2. Heine gives as the matrix representing a rotation about the z-axis through an angle ( for a representation having J = 1/2 as: D(1/2) (pz) = (I-1) 66e 66

and for a rotation about the y axis through an angle G as: /cosl/2Q sinl/21 D(l/2) (,y) = sinl/ (I-2) L-sinl/29 cosl/2j Thus, according to (I-1), the representative for j = 1/2 of the rotation Cp is: e(i/2)cp 0 0 e(-i/2)cp and for the rotation CCp) is: (i/2)cp + it O Lo0 e(-i/2)q - i The characters of these representatives are then: X(CC) = e(i/2)p + e(-i/2)c = 2 cos (cp/2) and x(Cp) = 2 cos(p/2 + i) = -2cos(p/2) The character for j = 1/2 for C2 can be found from (I-2) by observing that the character of C2 must be independent of the axis about which the rotation is performed as long as it is in the x-y plane so that the y axis is perfectly general, not for the rotation matrix, but for the character. Accordingly, the representative is: cos (i/2) sin(xi/2) Lsin(X/2) cos (i/2) and X(C2) = o 67

For the representation D(3/2), the step-up and step-down matrices are: I+ = I_ = so the Ix and the Iy matrices are: Ix- 3/2 Iy = 3/2i and Iz is Iz = 3/2 Using Heine's equation for the rotation operator for a rotation through an angle a about an axis ~: (c,6), = L + iac It + (iO I )2/2' + the representative for D(3/2) for CFp may be derived from: D(3/2) = E + (ip zn/n (I-3) n=l Since I = (9/4) 1, where 1 is the unit matrix, (I-3) may be broken into a sum of even powers of Iz and a sum of odd powers: D(3/2) =1 + 1 I ((3i/2) )2n/(2n): n=l + Iz ((3i/2)cp) /(2n + +)' n=O which reduces to: (3i/2)cp (3~/2)((pOe(-3i/2)2j 68'

whose character is 2cos(3/2)cp. The character for Di(5/2) (Ccp ) can similarly be shown to be -2cos(3/2)cp and that for D(3/2)(C2) to be zero. The complete double group character table is given in Table XIV and the spectroscopic terms belonging to each of the representations in Table XV.

TA.BLE XIV THE CHARACTER TABLE FOR THE Dooh DOUBLE GROUP!~oh E R Ccp RCcp 9V Rvv i RI ic RiC.C C. Alg 1 1 1 1 1 1 1 1 1 1 Alu 1 1 1 1 1 1 -1 -1 -1 -1 -1 A2g 1 1 11 1 1 1 1 1 -1 A2u 1 1 1 1 -1 -1 -1 -1 -1 -1 -1 Eig 2 -2 2cos p/2 -2cos p/2 0 0 2 2cos cp/2 -2cos p/2 0 Elu 2 -2 2cos cp/2 -cos os /2 0 0 -2 2 -2cos cp/2 2cos p/2 0 Elg 2 2 2cos cp 2cos cp 0 0 2 2 2cos p 2cos c O Eju 2 2 2cos qp 2cos f 0 0 -2 -2 -2cos q -2cos p O Etg 2 -2 2cos3cp/2 -2cos'3p/2 0 0 2 -2 2cos 35p/2 -2cos 3cp/2 O Elu 2 -2 2cos 3qp/2 -2cos 3qp/2 0 0 -2 2 -2cos 3q)/2 2cos 3p/2 0 E2g 2 2 2cos.2p 2cos' 2p 0 0 2 2 2cos 2cp 2cos 2p O E2u 2 2 2cos: 2cp 2cos 2p 0 0 -2 -2 -2cos ap -2cos 2p o E2g 2 -2 2cscsos /2 0 0 2 -2 2cos /2 -2cos'5/2 0 Etu 2 -2 2cos'5p/2 -2cos 5Pp/2 0 0 -2 2 -2cos 5cp/2 2cos'5/2 0 E3g 2 2 2cos 3q 2cos 3cp 0 0 2 2cos3cp 2coscp E3 2 2 2cos'3p 2cos 35p 0 0 -2 -2 -2cos 35p -2cos 3cp O

TABLE XV ASSIGNMENT OF TERMS FOR A MOLECULE OF THE Dh DOUBLE GROUP Rep. Singlets Doublets Triplets Quartets Ajg 3TTg Alu ZuTu A2g ogTTg A2U'Z8Tou Elg 2]T/2g 2Z1/2g 21T1/2g 1/2g H Eu VFU!21 2Z1/2u 411i/2u 4A1/2u Eig I11g 3frlg 3Zig 3Ajg Elu "1Tu 3rTiu 3Zu 3A, E2g 2113/2g 2A 3/2g; 1Tr3/2g An3/2g 43/2g Eiu 21T3/2u _3/2U lT3/2u Ai/2u 4Q1/2u E~g Ag +3T g 3A,4 E2u A7u 31T2u 3A 3.2U Ei~~~T/u ~A~/~ TL/ ~z/u ~ E2g 2A5/2g 2Z1s/2g 4TT/2g AL3/2 3/2 5/2 Eu /2u /2U4l1/2u 1A5/2u 15/2u75/2u EE~g 1A 9 7g 3Z23A 33g2r3g ~E3gU 1~~~~~U 3a33u 3au 3 E2tl l Au 32U3L 2 32U F1 I 2Z~~~ 21 g 4(~~~5

APPENDIX II THE EXPERIMENTAL ARRANGEMENT AND ITS OPERATION The problem of obtaining the absorption spectrum of a sample while it is exposed to a bright secondary illumination poses a major difficulty which is not encountered in ordinary absorption spectroscopy: that of isolating the signal light from that of the pumping light which excites the absorption. If the sample exhibits absorption in a region far different from that where it is pumped then this isolation is accomplished by the frequency difference between them. However, in the case of the uranyl ion, the pumping band extends from 3700 to 5000 A. and the induced absorption from 4600 to 7000 A so there is a significant region where the two overlap. Having the pump and signal frequencies different is not sufficient, moreover, because if the sample fluoresces in the same region in which it absorbs, this also causes an error signal to be present. The uranyl ion fluorescence spectrum almost completely covers its induced absorption spectrum. Thus, it is not possible to set up such an experiment in the same manner as the conventional absorption spectrum would be taken: to image the source light on the surface of the sample and then to focus the transmitted light onto the entrance slit of the dispersive device. Since the exciting light is, in general, so much brighter than the source lamp, illuminating a sample set up in the preceeding manner would simply yield a spectrum consisting of the light scattered from the sample and its fluorescence. The isolation between the signal and pump (or fluorescence) may be obtained in another manner: that of spatial (or optical) separation. This is accomplished in the following manner: (1) the source light is collimated and passed through the sample as a parallel beam; (2) the sample is contained in a flash head arranged so that the sample is illuminated in a direction perpendicular to the direction of the source beam; and (3) the source light is refocused and may be passed through a series of beam stops to eliminate as much scattered flash light as desired. This arrangement then has the following characteristics: (1) the flash light cannot enter directly into the path of the signal light but must be scattered through 900 to even be propagating in the same direction; and (2) as viewed from the lens following the flash head, the two sources of light have an infinite separation: the source is at infinity and the flash lamp scatter and sample fluorescence is at some small finite distance. Figure 36 illustrates the flash lamp and arc lamp setup. There is still another method of obtaining isolation between the two signals: that of a difference in time. Suppose the exciting light is fired in such a manner that it cuts off very rapidly and that immediately following 72

. I\\ ~ XENON ARC LAMP I \ I \ FLASH TUBE I POWER FROM COLLIMATING LENS CAPACITOR BANK (200 mmf.I.) (00 SAMPLE f.I.) Fge6FsHdI I I ADJUSTABLE - ] I II ~i I I I i I /(500 mm f.'.) Figure 36. Flash Head and Arc Lamp Arrangement. 73

this, a signal flash lamp is fired that is of so short a duration that no appreciable decay of the absorption can occur in the time interval. Then the difference in time of these two signals will provide the isolation required. There are, however, disadvantages to this procedure which make it unsuitable under certain circumstances: the signal light must be much more intense than in the previous case since the same amount of light must get through in a much shorter time. Ordinarily this would make no difference but in the case of induced absorption it would at best be questionable to assume that no changes in energy level populations (and hence in absorption spectrum) would occur under the application of a high intensity signal light. If the fluorescence of the sample is located in the same region as the induced absorption, it can still give an error signal under these civ:-cumstances. Figure 37 shows the arrangement of the xenon arc light source, -hle Ilash head, and the spectrographs. The photographically recorded spectra were taken with the 1.5 meter Bausch & Lomb spectrograph in the following manner: (1) the mirror which reflects the light beam onto the slit of the spectrograph had a thin plane glass (a photographic plate with the emulsion removed) attached to the front of it forming an angle of about 5O with the first surface of the mirror. This glass provided a secondary image of the arc source on the Exakta camera shutter about 1/2 inch to the right of the primary image from the mirror as viewed from the front of the camera. (Many secondary images were thrown to the left as well.) This is illustrated in Figure 38. When the camera shutter is triggered, the shutter sweeps open and permits the secondary image to pass through first, leading the main image by about 10 msec. A. small 45~ mirror mounted beside the slit reflected the secondary beam onto an RCA 922 photodiode. The purpose of this arrangement was to provide a precise reliable trigger by which the exact time the arc light would pass through the slit of the spectrograph could be established. By mounting an RCA 921 photodiode at the film holder of the spectrograph, measurements were made to check both the reproducibility of this trigger and the shape of the light pulse transmitted through the slit. It was found that the width of the light pulse was a strong function of the distance between the slit and the focal plane shutter, and when the shutter is placed as close to the slit as possible, the transmitted light pulse has the appearance of a Gaussian wave form with a half-width of 1.0 msec. The time variation between the trigger provided by the 922 photodiode and the light transmitted through the shutter was 50 Lsec or less, a small fraction of the pulse duration. The camera shutter was wound and triggered by a mechanical timing network which cycled automatically. This process was limited by the time that was required to rewind the shutter and was ordinarily repeated every 15 seconds. (2) The trigger pulse from the RCA 922 photodiode was fed into the input of a Fairchild Type 76-OlI\ Single Trace amplifier using a Type 74-03A Time 74

PHOTO DIODE FOR MONITORING FOCUS 8S"X" MIRROR"M" XeARC LIGHT PASSING / THROUGH SPECTROGRAPH SLIT ~ MIRROR/fiD == i_=~cr --— 0:0 —>-I |ORDER-SORTER I I CAMERA SHUTTERiI - 500 mm f.l. I TRIGGER<O i LENS m? PHOTODIODE tl | BAUSCH tI I'& LOMB l-. BEAM STOP & LOMB BEAM STOP 1.5 METER FLASH TUBE | |FLASH HEAD SPECTROMONITORING: GRAPH'IR JARREL-ASH PHOTODIODE' 200 mm f.I. LENS 3.4 METER EBERT SPECTROGRAPH XENON ARC LAMP Figure 37. Optical Bench Arrangement for the Recording of Induced Absorption Spectra.

FIRST SURFACE MIRROR LENS ~~-~~ ~ ~~~~ i\! SPACER THIN GLASS PLATE CAMERA SHUTTER, i I (IN COCKED POSITION) I I TRCAIGGE92R i'j! I DIRECTION OF PTRIGGER I SHUTTER SWEEP PHOTODIODE( — - " \ I FIRST SURFACE MIRROR PRIMARY BEAM - - SECONDARY BEAM - Figure 38. Camera Shutter, Mirror, and Trigger Photodiode Arrangement.

Base on the 777 Main Frame (dual beam). The time base was set to trigger from this input signal. (3) The + gate from the Type 74-03A. Time Base was then used to trigger a Tektronix Type 162 Waveform Generator, i.e., the oscilloscope in this application acts as an amplifier for the trigger pulse, and the correct delay such that the flash tube would trigger when the camera shutter was open was selected from a Tektronix Type 161 Pulse generator fed by the waveform generator. Both of the Tektronix units were powered by the same Type 160A. power supply. Figure 39 shows the arrangement of this timing network. The pulse duration of the flash tube was, unfortunately, only about 0.5 msec so the sample could not be irradiated throughout the time that the shutter was open. In the case of the uranyl sulfate solution, the decay of the induced absorption was known to be rapid so, in order to achieve a maximum absorption signal, the maximum of the flash tube discharge was adjusted to the maximum of the transmission through the spectrograph slit. In uranyl glass, however, the induced absorption decays more slowly and a larger signal on the film could be recorded if the flash tube maximum occurred 50-100 ULsec before the transmission maximum. The solution to this problem, which unfortunately was not available at that time, would have been to lengthen the discharge of the flash lamp. No camera shutter was employed in the recording of the fluorescence spectra on the 1.5 meter Bausch & Lomb spectrograph and the only change in the setup was to move the flash head from its position shown in Figure 37 to the focus point marked "X" in the same figure. This permitted the sample to be imaged on the slit of the spectrograph and the fluorescence spectrum to be taken. The photoelectrically recorded fluorescence and induced absorption spectra were taken in a manner analogous to that above except, of course, the mirror "I'M" (Figure 37) was removed and there was no necessity for the use of a camera shutter. The output from the photodiode monitoring the flash tube light output (see Figure 37) and the output from the RCA 1P28 photomultiplier were fed into the inputs of the Fairchild Type 76-02A Dual Trace amplifier. The sweep was provided by a Fairchild Type 74-llA. Delaying Sweep whose trigger was obtained by simply wrapping the innter conductor of a RG-58A./U coaxial cable several times around the trigger lead to the flash tube and leading the other end to the external trigger input of the sweep amplifier. The whole system was then activated by firing the flash tube. 77

FAIRCHILD 777 MAIN FRAME TYPE 76-02A DUAL TRACE TP76 AMPLIPIER DELAYING TYPE 160A O1INPUT I SWEEP POWER SUPPLY EXTERNAL TRIGGER QINPUT 2\........ - _ TRIGGER TO FLASH TUBE TYPE 76-01 TYPE7403A INPUT INPUT AMPLIFIER TIMEBASE ~ + GATE OUTPUT TYPE 162 SC WAVE FORM I, TYPE 161 FLASH TUBE TRIGGE PULSE GENERATOR GENERATOR c INPUT INPUT TEKTRONIX RCA 922 RCA 921 PHOTODIODE PHOTODIODE MONITORING NOTE: FIGURE ILLUSTRATES ARRANGEMENT > TRIGGER I I MONITORI G ARC LIGHT FOR SYNCRONIZATION OF FLASHTUBE PHOTODIODE )FLASH TUBE ARC LIGHT FIRING AND TRANSMISSION OF LIGHT BY PASSING THROUGH THE SPECTROGRAPH SLIT | OUTP1U T SPECTROGRAPH SLIT Figure 39. Oscilloscope and Delay Network Arrangement.

APPENDIX III ATTEMPTS AT CRYSTAL GROWTH AND REMARKS ON SIZE AND CLARITY REQUIREMENTS Following the procedure given by Dieke and Duncan,3 many attempts to grow single crystals of cesium uranyl nitrate, Cs2U02(N03)4, from a concentrated nitric acid solution were made. This crystal was chosen because it, unlike the majority of uranyl single and double salts, has no water of hydration which must be avoided if the crystal is to withstand the high intensity light flash from the xenon flash lamp and because the spectrum has been analysed in some detail by previous workers (cf. Dieke and Duncan). (To test the stability of a crystal containing water of hydration, a large single crystal of manganous chloride, MnCl2'2H20, was placed in the flash head and the flash lamp fired. The crystal was very clear (but not without imperfections) and after two firings of the flash lamp had become frosted throughout. It is probably the momentary heating which causes the fracture.) Despite all precautions, which finally included placing the nitric acid solution in a dessicator suspended in a 30-gallon water tank to provide shock and thermal insulation and sweeping the solvent out with dry nitrogen, the crystals failed to grow large and clear enough for use in these experiments. The cesium uranyl nitrate, which is only moderately soluble in concentrated nitric acid, crystallizes in flat square plates having a width (and length) to thickness ratio of about 5:1. Although many seed crystals of about 2 mm cross-section were obtained which were without imperfections, these could not be grown to a larger size without developing a very characteristic pair of flaws: two lines of imperfections which run diagonally across the large square face of the crystal. As it is imperative to have samples which are extremely clear and of at least 5 mm diameter, attempts at the growth of crystals had to be abandoned. Clarity, more than size, is of extreme importance in these experiments because of the need to isolate the xenon arc lamp signal, which provides the information concerning the induced absorption, from the light signal from the xenon flash lamp which excites the induced absorption. The light intensity from the flash lamp is so much higher than that of the arc lamp that if the sample scatters the flash lamp light, some of it, albeit a very small fraction, can register as the arc signal. Since the arc signal is expected to go down during the time of the flash but the flash signal would increase at the same time, the two effects oppose one another. The glass and solutions employed were without imperfection but even these gave a small scattering of the xenon light. It does not require much degradation of the optical quality from this high standard before the scattered light overwhelms the relatively small arc signal. For example, a crystal of fusion-grown 79

sodium fluoride containing a small amount of uranium trioxide* which showed no imperfections of any kind other than a slight amount of schlieren was tried but produced a signal-to-noise ratio at least five times worse than the uranyl glass. This is due to two causes: (1) a scattering of the flash lamp light into the optical path of the arc light, and (2) a loss in the arc light signal due to imperfections. It might be possible, by the use of repeated exposures and a subsequent mathematical analysis of the photoelectric signal, to analyse the spectrum of a material which scatters badly but it is unlikely that such a solution is available in the case of the spectruii recorded on photographic film. There are two areas in which the size of the sample is important: (1) The number of exposures which would be required for the photographic record will increase as the sample is decreased. This is not of too much concern however, because with the existing setup only 50 exposures on Pan-X film are required to give a readable spectrum. This film is quite slow and with the faster films available a considerable reduction in sample size could probably be tolerated. (2) The aperture of the system is limited by the size of the sample since it is in a parallel beam of light and not at a point of focus. It was just possible to fill the aperture of the 3.4 meter Ebert spectrograph with a 5-mm diameter sample with the setup described here so any smaller sample would have vignetted the grating and spoiled its resolution. This problem was not encountered with the 1.5 meter B & L spectrograph which has approximately the same aperture as the Jarrell-Ash. (This was due to the difference in lens arrangement.) However, it would be possible, with a different physical arrangement, to use a smaller aperture. *Provided by Ojars Risgin, WRRL, University of Michigan. 80

APPENDIX IV FILM ERRORS, THEIR EFFECTS ON THE SPECTRA, AND THEIR CORRECTION Figure 40 shows the tracings obtained when the induced absorption spectrum of uranyl sulfate solution was recorded for the first time. There are two features which are particularly evident and whose cause needed to be investigated: (1) the pumped sample apparently transmitted more light in the region with wavelength longer than 5800 A; and (2) the strong emission lines of the xenon lamp appear to be preferentially absorbed more than the continuum around them, i.e., the heights of the peaks are greatly reduced. The first of these anamolies was discovered to have been caused by a malfunction in the camera shutter. Being a focal plane shutter, the two "leaves" of the shutter are drawn back together when the camera is recocked but a failure in the shutter permitted a small gap between the leaves to form during this process so that light could leak through during the recocking process. Thus, there was more light passing through the slit of the spectrograph when the pumped exposure was made and the first of these errors was disposed of. (The pumped exposure was made after the unpumped one and the failure in the shutter must have occurred at the end of one run or the beginning of the other.) The second one, however, required more analysis of the behavior of the photographic film. Ideally, the optical density of an exposed film is a linear function of the logarithm of the exposure the film has received. In practice, this condition is met for a large number of films over a large range of optical density. There are three main deviations from this ideal behavior, the first two of which do not appear in the results here. These are: (1) reciprocity failure, (2) the intermittancy effect, and (3) the curvature of the optical density versus log (exposure) curve at low log (exposure) values. The real optical density versus log (exposure) curve has an "S" shape, the slope of the curve being smaller close to the origin than further out. The apparent overabsorption of the xenon lines can then be interpreted according to this curve. In the pumped tracing, the general light level is lower due to absorption by the sample so the portion of the curve whose shape is reflected in the tracing is close to the origin. There, the slope of the optical density versus log (exposure) curve is small so that differences in light level are compacted into small differences in optical density. In the unpumped tracing, however, the small increase in light level has caused a greater amplification of the differences in light intensity. This process is illustrated in Figure 41.

Unpumped?_-.fI7 Note "preferential" absorption z7 |/ %\t\0 Jl 1Z; 11of Xenon lines LC. C-) 6750 6450 6150 5850 5550 5250 4950 4650 4350 4050 3750 WAVE LENGTH (A) Figure 40. Induced Absorption Tracings for U02S04 Solution Illustrating Film Error.

i 7 r | I I HIGH LEVEL BACKGROUND VARIATION IN OPTICAL DENSITY I: I.Z/I: ~u BAC/ I] LOW LEVEL BACKGROUND:_ ~;_..tVARIATION IN OPTICAL DENSITY log (INTENSITY) II LOW LEVEL | I HIGH LEVEL I BACKGROUND Figure 41. Intensity "Amplification" by Film H & D Curve. I:F _RIloNI~oPI85DEST

To test the validity of the above conclusion, and optical density versus log10 (exposure) curve was constructed for Kodak Pan-X film, beginning with a background blackening of 0.3 optical density. The background was used to see if this would provide a sufficient pre-exposure to get the film over the "toe" of its response curve. The data points were obtained by the use of a tungsten lamp and five 0.5 optical density neutral filters to provide calibrated changes in exposure with the 1.5 meter B & L spectrograph to provide a wavelength dispersion. All of the data points thus obtained could be fit to a single curve regardless of wavelength indicating that the same background pre-exposure could be used to lift the film over the toe regardless of wavelength. This curve is shown in Figure 42. It will be noted that there is still some of the toe of the curve left even at such a pre-exposure and it is not until a total optical density of 0.6 is reached that the curve becomes linear for about 1.0 optical density. Accordingly, the spectra which were recorded on the 1.5 meter spectrograph were all given a pre-exposure of green light sufficient to blacken them to 0.6 optical density. The spectra thus taken are quite poor for viewing with the eye because of the lack of contrast but quite suitable for obtaining a microdensitometer tracing. Small deviations from linearity probably account for the unreal structure seen in the induced absorption spectrum of the uranyl sulfate solution (Figure 14). 84

1.9 1.7 1.5 7 1.3 < 1.1 CIa — 0.9 0.7 0.5 o.3I 0 1 2 3 4 5 6 LOG (EXPOSURE) (ARBITRARY UNITS) Figure 42. H & D Curve for Kodak Pan-X Film.

APPENDIX V A DESCRIPTION OF THE SPECTROGRAPHS The following is a brief description of the spectrographs employed in the recording of the induced absorption photographically and photoelectrically. BAUSCH & LOMB 1.5 METER SPECTROGRAPH This instrument, which was used to record the induced absorption spectrum photographically and also to record the fluorescence spectrum of uranyl sulfate trihydrate with DC and flash excitation, has an end-on Eagle mounting using a concave grating of 1.5 meter focal length with 450 lines/mm and a ruled area of 40 by 80 mm. In the first order the wavelength range covered is from 3750 to 7500 A at a dispersion of 15 s/mm and a theoretical resolution of 35,000 in the first order. This instrument is equipped with a fixed slit having widths of 10, 30, and 60 microns and was always run with the 10micron slit. The instrument is rendered stigmatic by a cylindrical lens located between the slit and grating. JARRELL-ASH 3.4 METER EBERT SPECTROGRAPH This instrument, which was employed as a monochromator to record the timeresolved induced absorption and fluorescence spectra, is and over-and-under Ebert mount with a 4 inch x 8 inch interferometrically ruled plane grating. The grating has 300 rulings/mm and is blazed on one side at 59,000 A in the first order and at 31,000 A in the first order on the reverse side. Since high resolution was not required, the 31,000 A blaze side was always employed because fewer orders have to be used to cover the visible region on this side. For the 39,000 A blaze side, the orders for the visible region are fifth to eighth. On the 59,000 A blaze side the same region is covered by the eighth to seventeenth. Table XVI gives the resolution, approximate dispersion, and wavelength region for each of the orders used to cover the visible region on the 39,000 A blaze side. Had it been necessary, the 59,000 A blaze side of the grating would have been employed. Table XVII lists the resolution, approximate dispersion, and wavelength region for this side of the grating. The resolution has been experimentally verified by the observation or the Iodine (12) absorption spectrum which showed that at least 90% of theoretical resolving power is being obtained. Because of the large dispersion of the spectrograph and the large bandwidth of the spectrum being observed, the photomultiplier was operated with 86

TABLE XVI RESOLUTION, DISPERSION, AND WAVELENGTH RANGE FOR THE 31,000 A BLAZE SIDE OF 3.4 METER EBERT SPECTROGRAPH Order >D —~.ngth Range (A) Resolution Approximate Dispersion (A/mm) 5 5700-7000 305,o000 1.68 6 4970-5690 366,000 1.40 7 4l40-4960 427,000 1.20 8 365o-4130 488,000 1.05 TABLE XVII RESOLUTION, DISPERSION, AND WAVELENGTH RANGES FOR THE 59,000 A BLAZE SIDE OF THE 3.4 METER EBERT SPECTROGRAPH Order Wavelength Range (A.) Resolution Approximate Dispersion (i/mm) 7 6940-7000 427,000.721 8 62o10-6940 488,ooo.631 9 5920-6210 549,000.561 10 5130-5920 610,000.505 11 4720-5130 671,000.459 12 4370-4720 732,000.421 13 4068-4370 793,000.388 14 3806-4068 854,000.361 15 3575-3806 915,000 337 16 3371-3575 976,ooo.316 17 -3371 1,037,000.297 slits from 1 to 5 mm wide giving a bandwith of from 2 to 15 ~ which is much below that of the vibrational spacings which were to be observed. The entrance slit of the spectrograph was opened to 400 microns. Because of the small free spectral range of the spectrograph, a predispersion of the light before it enters the spectrograph is required. This was accomplished by the use of a small prism apparatus having a focal Length of 23 cm constructed with a Pellin-Broca constant deviation prism. Figure 43 shows the arrangement for this predisperser (order-sorter). 87

0 - 1/2 INCH MICROMETER HEAD CENTER OF ROTATION TURNTABLE LENS'I )~~~~~~~.... II APERATURE-STOP -_SPRI__ Z./ DASHED LINES SHOW OPTICAL PATH THROUGH ORDER-SORTER \/ NOTE: PRISM-TO-SPECTROGRAPH DISTANCE IS CONSIDERABLY FORESHORTENED i SPECTROGRAPH SLIT Figure 43. Order Sorter for 3.4 Meter Jarrell-Ash Spectrograph.

APPENDIX VI AN ATTEMPT TO OBSERVE "BLEACHING' OF THE INDUCED ABSORPTION As it was known that the light from a ruby Q-switched.laser is capable of "bleaching" the uranyl glass induced absorption and to cause it to become transparent again, even though a flash lamp may be exciting the sample, an experiment was tried in which an attempt to observe this bleaching directly was made0 To accomplish this end, two sources of pumping light are required, one to induce the absorption and the second to bleach the absorption in the same manner that the laser light does, by ptu-i.-ing the sample with light of a frequency which will not pump the uranyl:ons from the ground state but which can be absorbed by the sample once the induced absorption is present. To acco:mplish this double pvimp?ing, the sample of uranyl glass was placed in a flacs11 head so constructed that it would hold two linear flash tubes each at the focus of a different ellipse but whose other foci were coincident at which the sample was placed. The flash tube used to induce the absorption was an'FX-iO4 dissipating only 67.5 joules (1500 VDC at 60 mfd). The other flash tube, an FX-42*, had a 2-nmu thick section of Jena OG-2 filter glass located between it and the sample. This glass is a sharp-cut red filter which transmits about 90% of the light of wavelength longer than 5800 A. and whose transmissioz has dropped to zero before the uranyl glass absorption bands have begun. Although the input energy to this second flash tube was 700 joules (2000 VDC at 350 mfd) it was impossible for it to provide any radiation which would induce the absorption but could only cause transitions between the resonance level and the middle level of the ultraviolet system. The small flash tube was fired by itself and the induced absorption recorded. The large flash tube was then fired by itself and it was verified that no induced absorption was observed~ Both the flash tubes were then, fired at the same time (the larger one lagging because of the longer time constant of its circuit) with the hope of observing that the induced absorption would diminish due to the depumping provided by the larger flash lamp. The oscilloscope traces for this experiment are shown in Figure 44~ The failure of the secondary pump lamp to produce any change in the absorption induced by the primary one can probably be ascribed to the fact that the relaxation time from the middle level of the ultraviolet system to the resonance level must be considerably shorter than 10-7 seconds, a process with which the flash tube illumination cannot compete, only a light source as intense as the laser can show any changes in the population of these levels, *Edgerton, Germeshausen, & Grier, Boston, Mass. 89

*Ei!1111 (a) (b) (c) Figure 44. An Attempt to Observe Quenching of the Induced Absorption in Uranyl Sulfate Solution. (a) Small pumping flash tube only firing. (b) Large "depumping" flash tube only firing. (c) Both flash tubes firing. 90

BIBLI OGRAPHY 1. L. A. Cross, and L. G. Cross, IEEE Conference, Cincinnati, Ohio, April 7, 1964 (1st paper). 2. L. G. Cross, and L. A. Cross, IEEE Conference, Cincinnati, Ohio, April 7, 1964 (2nd paper). 30 Go Ho Dieke, and A.B.F. Duncan, "Spectroscopic Properties of Uranium Compounds," National Nuclear Energy Series, Div III, 2, McGraw-Hill, New York, 1949. 4. H. Morton, and H. C. Bolton, Chemo News, 28: 47, 113, 164, 233, 244, 257, 268 (1873)o 5. H. Kayser, "Handbuch der Spectroskopie," S. Herzel, Leipzig, 1905. 6. R. S. Mulliken, J. Chem. Physo, 7: 14 (1939)o 7. Bo E. Gordon, Dokl, Akad. Nauk SSSR, 74: 913 (1950). 8. D. D. Pant, JO Sci. Research Benares Hindu Univ., 3: 27 (1952). 9o V. Lo Levshin, and Go Do Sheremetjev, Zhuro Eksptl. Theoret. Fiz., 17: 209 (1947). 10o Co Billington, Phys. Rev., 120: 710 (1960), 11, V. Henri, and Mo Landau, Compto Rend., 158: 1511, 1688 (1914)o 12. S. Ahrland, Acta. Chem. Scand., 5: 1151 (1951), 135 S. P. McGlynn, and J. K. Smith, J. Mol. Spec., 6: 164, 188 (1961)o 14, E. Rabinowitch, and R. Lo Belford, "Spectroscopy and Photochemistry of Uranyl Compounds," MacMillan Co., New York, 1964. 15. L. G. Cross (private communication). 16. W. H. Zachariasen, Acta. Cryst o, 1: 288 (1948). 17. W. H. Zachariasen, Actao Cryst,, 7: 783 (1954), 18. V. Heine, "Group Theory in Quantum Mechanics," Pergamon Press, New York, 1960. 91

UNIVERSITY OF MICHIGAN I t3Ililt II I III Il9 0ll 02 Ilt8 6Ill0 7 3 9015 02086 6607