ELECTRON PARAMAGNETIC RESONANCE ABSGRPTEON STUDEES ON SOLUTIONS OF METALS IN AMMONEA AND AMINES That: for Hm Degree of DH. D. MECHIGAN STATE UNIVERSETY Kenneth Dean V05 1962 TH 5315 0.7... FJA Lr6§1"" Michigan State University W” uA ' . \ ' ABSTRACT ELECTRON PARAMNENETIC RESONANCE ABSORPTION STUDIES ON SOLUTIONS OF METALS IN AMMONIA AND.AMINES by Kenneth Dean VCs Solutions of metals in ammonia, methylamine and ethylenediamine 'were studied by means of electron paramagnetic resonance absorption. Two instruments, a homebuilt apparatus and a varian Modei hSOO Spec— trometer with 100 KC. field modulation.were used in these studies. Line widths and g values were measured for alkali and alkaline earth metals in ammonia and alkali metals in methylamine and ethylenediamine. Samples of sodium, potassium, and rubidium in ammonia which sep- arated into two phases gave two epr absorptions. The less dense bronze phase showed a broad asymmetric absorption curve. The blue phase gave a narrow Lorentzian absorption. The line widths of the bronze and blue phases were different for different metals. Saturated solutions of cesium in ammonia and solutions of potassium in ammonia that contained excess metal also gave asymmetric curves. For all of the asymmetric absorptions, the low field side of the first derivative curve was about 2.5 times farther from the base line than the high—field side. This result is in accord with the theory of Dyson for the epr absorption of highly conducting samples. Some of the two-phase potassium samples showed more than two absorption lines with a maximum of four being observed. Solutions of lithium, sodium, and potassium in methylamine and lithium, sodium, potassium, rubidium, and cesium in ethylenediamine gave Lorentzian curves. These solutions were measured at various tem- peratures and the direction and extent of change of the line width during decomposition were noted. Kenneth Dean Vos Solutions of rubidium and cesium in methylamine exhibited nuclear hyperfine structure, consisting of six and eight hyperfine lines re- Spectively. The total hyperfine separations were compared to those determined from molecular beam experiments. The electron densities at the nuclei were computed and compared to the values for the free atoms. .An extra epr absorption line was also evident in both the rubidium and cesium solutions. Solutions of cesium in methylamine were studied as a function of temperature. The total hyperfine separation for the ces- ium solution at -lOOOC was a factor of four smaller than the room temperature value. The extra absorption line narrowed by a factor of 20 and its intensity rapidly increased as the temperature of the cesium solution decreased to -lOO°C. The results obtained indicate the complex nature of these solutions. INo simple model can be used to explain all of the results, but some of the observations can be correlated with existing theories. ELECTRON PARAMAGNETIC RESONANCE ABSORPTION STUDIES ON SOLUTIONS OF METALS IN AMMONIA.AND.AMINES By Kenneth Dean V05 A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1962 To my wife Irene ii ACKNOWLEDGMENTS The author wishes to express his sincere appreciation to Professor James L. Dye for his guidance, assistance, and encouragement during the course of this investigation and the preparation of this thesis, to Pro- fessors Max T. Rogers and Richard H. Schwendeman for their assistance in the investigation. Financial aid from the Atomic Energy Commission is gratefully acknowledged. iii TABLE OF CONTENTS Page 1.:mnmmmnON ... ... ... ... ... ... ... 1 II. HISTORICAL . . . . . . . . . . . . . . . .‘. . . . . . S III. THEORY OF‘ELECTRON PARAMNSNETIC RESONANCE . . . . . . . 13 IV. THE HOMEBUILT EPR SPECTROMETER . . . . . . . . . . . . 21 v. DETERMINATION OF‘PARAMAGNETIC SUSCEPTIBILITY . . . . . 2h VI. EXPERIMENTAL . . . . . . . . . . . . . . . . . . . . . 28 Ethylenediamine Purification . . . . . . . hl Sample Preparation . . . . . . . . . . . . b6 A. Measurement Of Magnetic Field . . . . . . . 28 B. Glassware Cleaning . . . . . . . . . . . . 32 C. Metal Purification . . . . . . . . . . . . 33 D. Ammonia Purification . . . . . . . . . . . 35 E. Methylamine Purification . . . . . . . . . 3? F. G. VII. RESULTS . . . . . . . . . . . . . . . . . . . . . . . . N9 A. Solutions in Ammonia . . . . . . . . . . . A9 B. Solutions in Methylamine . . . . . . . . . 56 C. Solutions in Ethylenediamine . . . . . . 60 D. Rubidium and Cesium in Methylamine . . . . 67 E. Measurements of 9 value . . . . . . . . . . 7h VIII. DISCUSSION . . . . . . . . . . . . . . . . . . . . . . 76 Solutions in Ammonia . . . . . . . . . . . 78 Solutions in Methylamine . . . . . . . . . 80 Solutions in Ethylenediamine . . . . . . . 81 Solutions of Rubidium and Cesium in ' Methylamine . . . . . . . . . . . . . . 82 . Comparisons . . . . . . . . . . . . . . . . 85 Future Work . . . . . . . . . . . . . . . . 86 'RN PPR? IXREFERENCES 87 iv LIST OF TABLES TABLE Page I. Lithium in Methylamine . . . . . . . . . . . . . . . . . 57 II. Sodium in Methylamine . . . . . . . . . . . . . . . . . 59 III. Potassium in Methylamine . . . . . . . . . . . . . . . . 61 IV. Lithium in Ethylenediamine . . . . . . . . . . . . . . . 63 V. Sodium in Ethylenediamine . . . . . . . . . . . . . . . 65 VI. Potassium in Ethylenediamine . . . . . . . . . . . . . . 65 VII. Rubidium in Ethylenediamine . . . . . . . . . . . . . . 66 VIII. Cesium in Ethylenediamine . . . . . . . . . . . . . . . 67 IX. Measurements of 9 Value . . . . . . . . . . . . . . . . 7S Figure amt—u.» \] 10. ll. 12. 13. 1h. 15. 16. 17. 18. 19. 20. 21. 22. LIST OF FIGURES Schematic of the NMR Oscillator . . . . . Line Width and 9 Value of DPPH . . . . . . . . . . Metal Degassing Tube and Metal Make-up Vessel . Purification Train . . . . . . . . . . . . . . . . Fractionation System . . . . . . . . . . . . . . . . Evers' Purification System . . . . . . . . . . . . . Nitrogen Bubbling System . . . . . . . . . . . . . Long Purification Train . . . . . . . . . . . . . . Final Storage Vessel; Ethylenediamine . Sample Make-up Vessel . . . . . . . . . . . . . . Sodium - Ammonia Spectrum . . . . . . . . . . . . Potassium - Ammonia Spectrum .. . . . Potassium - Ammonia Spectrum . . . . . . . . . . Rubidium — Ammonia Spectrum . . . . . . . . . . . . . Line Width vs Temperature 5 Lithium Methylamine . Line Width vs Temperature ; Lithium ~ Ethylenediamine . Sodium - Ethylenediamine Spectrum . . Rubidium - Methylamine Spectrum . . . . Rubidium - Methylamine Spectrum . . . Cesium - Methylamine Spectrum . . . . . . . . . . . Cesium - Methylamine Spectrum . . . . . . . . . . . Change of the total Hyperfine Separation and the Line Width of the Extra Line with Temperature; Cesium — Methylamine . . . . . . . . . . . . . . . . vi Page 29 3 1 3h 36 38 ho h2 MI AS A? 51 52 53 SS 58 62 6h 68 69 7O 71 73 I. INTRODUCTION The physical properties of solutions of alkali and alkaline earth metals in ammonia have been the subject of study for almost one hundred years. Since the early work of Weyl (1) many chemists and physicists have studied these systems because of their unusual physical and chemical properties. When metals dissolve in ammonia a metastable blue solution is formed which becomes a copper-bronze color as the solution becomes more concentrated. Much of the early work on metal ammonia solutions was carried out by C. A. Kraus. Kraus (2) postulated that the dissolved metal ionized according to the mass action expression _) _ M 4' X <_ M... +2 (NH3)X where M is the metal atom, M+ is the positive metal ion, and e' is the solvated electron. It was also postulated that the equilibrium shifted to the right as the concentration decreased giving at infinite dilution only metal ions and solvated electrons. The solutions should be paramagnetic if they contain unpaired electrons. This was verified by Freed and Sugarman (3), and Huster (h,S). These workers measured the static susceptibility of sodium, potassium, cesium, calcium, and barium in ammonia as a function of temperature and concentration. At high dilutions the static suscepti- bility approaches the value expected for free Spins. Dilute barium solutions have susceptibilities greater than expected for one electron per barium atom indicating that both valence electrons are ionized. Quantitative electron paramagnetic resonance experiments have shown the paramagnetic susceptibility of sodium and potassium in ammonia (6) and potassium in deutero-ammonia (7) to increase but the molar susceptibility to decrease with increasing concentrations. These results are in agree- ment with the static measurements and show that a concentration depend— ent equilibrium exists between paramagnetic and diamagnetic species. Transference number experiments (8,9,10) have shown the primary current carrier to be anionic. The negative species carries about seven times more current at low concentrations than does the positive carrier. The percentage of the current carried by the negative Species increases with increasing concentration until it is greater than 99 % at high concentrations. Nuclear magnetic resonance experiments on sodium ammonia solutions (11,12) have shown a chemical shift for the nitrogen and sodium nuclei (usually called a Knight shift for metals) but no detectable chemical shift for the protons. These results indicated considerable electron density at the nitrogen and sodium nuclei but very little electron density at the proton nucleus. In contrast to solutions of inorganic salts in ammonia, which show an increase in density and viscosity with increasing concentration, solutions of metals in ammonia Show a decrease in both density and vis— cosity with increasing concentration (13,1h). A number of theories have been proposed to explain the observed physical prOperties of which two have recently gained prominence (15,16). Both of the models involve the ionization of the metal in dilute solu- tion to give a metal cation and a solvated electron, as first postulated by Kraus. The main difference between the theories relates to the posi- tion of the electron in the intermediate concentration range (0.005 to 0.1 molar). The "cavity" model (15) considers the electron to be com— pletely removed from the metal and located in a "cavity" within the solvent. The Becker Lindquist Alder model (16) considers the electron to be in an orbit about the metal ion but outside of its solvation sheath. Each of the theories is able to explain some of the experimental data better than the other but neither can explain all of the observed facts. The original purpose of this research was to quantitatively deter- mine the paramagnetic susceptibility of metal-amminé*solutions using the electron paramagnetic resonance spectrometer. Hutchison and his students have made the only quantitative paramagnetic susceptibility measurements by electron paramagnetic resonance of sodium and potassium in ammonia and deutero-ammonia. The present research was to have ex- tended the sodium and potassium measurements to lower concentrations and lower temperatunxsand to make measurements on other metals in am- monia, methylamine and other solvents. By using higher magnetic fields than those employed by Hutchison and his students (3300 gauss 22° 5 gauss) it would be possible to obtain data at much lower concentrations and lower temperatures. Extending the measurements to lower tempera— tures would give a better measure of the temperature dependence of the equilibrium constant for the equilibrium between the paramagnetic and diamagnetic species. By making measurements at lower concentrations it would be possible to determine more precisely whether the molar sus- Ceptibility approaches the value for free spins or a lower value. The % . In order to avoid repeated use of the phrase, "ammonia and amines", true word "ammine" will be used in this thesis to designate these solvents. Curie law for the variation of magnetic susceptibility with temperature is not followed for the concentrated solutions, but it is not known whether the extremely dilute solutions follow the Curie law. Extending the measurements to other metals and solvents would give a clearer in- dication of the role played by the metal ion and the solvent in the susceptibility. After considerable effort, the original plan for quantitative measurements was abandoned and a number of qualitative measurements were made instead. Instrumental problems with the homebuilt electron paramagnetic resonance spectrometer that could not be adequately solved made quantitative measurements impossible. The Varian electron paramag- netic resonance spectrometer, which arrived after qualitative measure- ments had commenced, should, however, be satisfactory for making the quantitative measurements as described here. The line widths and g values for alkali and alkaline earth metals in ammonia and the alkali metals in methylamine and ethylenediamine were measured as functions of temperature at various concentrations. The concentrations were not determined so that the correlation with concentration is only quali- tative. II. HISTORICAL The static magnetic susceptibilities of solutions in ammonia have been measured as a function of temperature and concentration for sodium, potassium, cesium, calcium, and barium (3,h,5). At low concentrations the molar susceptibility approaches that of a free electron gas but at higher concentration the molar paramagnetic susceptibility falls off rapidly. A 0.5N potassium solution at -53°C is diamagnetic but at -300C it is paramagnetic. This is a dramatic demonstration that at high con- centrations the solutions do not obey the Curie law. Hutchison has theorized that in extremely dilute solutions the susceptibilities would obey the Curie law (6). Quantitative measurements of the paramagnetic susceptibility of potassium in ammonia over the concentration range from E X 10-3 to 1.6 molar have been carried out by Pastor (6). Pastor has also measured the paramagnetic susceptibility of sodium over the range 5 x 10-2 to 0.9 molar. These measurements covered the temperature range of -33°C to 270C. O'Reilly has measured the susceptibility of potassium in ammonia and deutero-ammonia as a function of temperature from -33°C to 25°C over the concentration range 2 x 10-3 molar to 0.6 molar (7). The re- sults of both the static and electron paramagnetic resonance measurements show that the molar susceptibility is larger at low concentrations than at high concentrations. The molar susceptibility for the concentration range studied by epr increased with increasing temperature, in agree- ment with the static measurements. Potassium and sodium appear to have the same molar susceptibilities at -33°C but at 25°C the value for sodium is 13%.kugen The molar susceptibility of potassium in deutero-ammonia and ammonia are the same at -33°C but at 25°C the sus- ceptibility in deutero-ammonia is lh% larger than in ammonia. The metal and the solvent are both important in determining the molar sus- ceptibility, eSpecially at the higher temperature. The spin—lattice (T1) and spin—spin (T2) relaxation times for metal- ammonia solutions have been measured. Pollak has measured both T1 and T2 for sodium and potassium in ammonia over the concentration range 1 x 10-2 to 0.h molar and at temperatures from -50°C to 30°C (17). The pulsed magnetic resonance technique developed by Hahn (18) was employed by Pollak using 30 mcruf. The results showed T1 = T2 for both the sodium solutions and the potassium solutions with the relaxation times rang- ing from 3.3 u sec. to 1.0 u sec. At temperatures below -30°C the re- laxation times increase as the concentration increases while at temper- atures above 0°C the relaxation times decrease as the concentration in- creases. In dilute solutions the relaxation times increase with in- creasing temperature. For solutions of higher concentration, the times pass through a maximum and then decrease as the temperature increases. The spin-lattice and spin—spin relaxation times of sodium in N15 ammonia have the same value but this value is somewhat larger than in N14 ammonia. Blume has measured T1 and T2 for sodium in ammonia using the pulse method at l7.h mc. He found T1 = T2 for those samples on which he was able to measure T1 (19). The T2 measurements of Pollak agree with those of Blume at concentrations below 0.15 M but at higher concentrations, Blunm's T2 measurements are longer than those of Pollak. 7 The continuous wave technique of measuring T1 and T2 has been used by Hutchison and O'Reilly with solutions of potassium in ammonia and deutero-ammonia (20). It is difficult to measure T1 and T2 as accur— ately with the continuous wave technique as with the pulse method. Never- the less, O'Reilly has made very careful measurements and at 25°C and con- centratiOns above 0.1 molar the T1 and T2 measurements of O'Reilly and of Pollak are in reasonable agreement. At concentrations below 0.1 molar, O’Reilly finds T1 different from T2 with T1 as much as 25% larger than T2. At the lowest concentrations studied, T1 decreases rapidly with decreasing concentration and it would appear that T1 and T2 might become equal at infinite dilution, as required by the theory of Bloembergen, Purcell, and Pound (21). The relaxation times measured by O'Reilly for ammonia and deutero-ammonia solutions are nearly in- distinguishable when a correction is made for the difference in viscos- ity of ND3 and NHS. Cutler and Powles have measured T1 and T2 for lithium, sodium, potassium, calcium and barium solutions in ammonia at 23°C using the pulse method at 93h0 mc (22,23). The measurements on lithium, sodium, calcium, and barium solutions showed T1 = T2 with a limiting value at low concentrations of approximately 3 u sec. As the concentration in- creased the times decreased for these solutions until the times reached the detection limit of the instrument, 1 u sec. Potassium showed T1 = T2 only for the concentrations greater than 1 molar. At lower concen- trations T1 was as much as 25% larger than T2. The pulse method has also been used by Kaplan, Browne, and Cowen to measure T1 and T2 for potassium in ammonia at 9300 mc (2h). The re- sults of these workers indicated T1 = T2 = 3 u sec at 0°C. The Spin—Spin relaxation time for potassium and sodium in ammonia may be calculated from the line width measurements of Pastor (6) because the curves are Lorentzian. Pastor's measurements on line width gave T2 as a function of temperature and concentration that agreed with the measurements of Pollak, Blume, and O'Reilly. The fact that the relaxation times in N15 ammonia were longer than in N14 ammonia is a clear indication that the nitrogen is important in the relaxation mechanism. The metal atom is also involved in the relaxa- tion as evidenced by the difference between sodium and potassium. The role played by the protons in the relaxation process is not as large as that of the nitrogen or the metal nucleus as evidenced by the negligible difference between H1 ammonia and H2 ammonia. Levy has investigated the electron spin resonance of frozen metal ammonia solutions at 30 mc in the temperature range of L° K to 200° K (25). The line widths and g values of lithium, sodium, potassium, and cesium were studied. The line widths of frozen sodium and potassium solutions appear to increase linearly with temperature. The dependence of the line width on temperature for lithium is anamolous in that it shows a sudden break at approximately 80° K, this may be due to a eutectic. Frozen cesium-ammonia solutions gave Spectra only in the temperature range of ho K to 770 K. Other electron spin resonance experiments of a qualitative nature have been carried out by several workers. Gastens and Ryan (26); Beeler and Roux (27,28); and Wysocyanski and Cowen (29) have studied epr in sodium-ammonia solutions. Hutchison and Pastor have observed epr resonance in potassium-ammonia solutions (30,31,32). Electron spin resonance experiments by Levinthal, Rogers, and Ogg have shown lithium in methylamine to have a line width of approximately 0.h gauss (33). The line width and amplitude of lithium in methylamine is reported not to show as large a temperature dependence as in the ammonia solutions. Levinthal gt al.also report that lithium in ethylenediamine has a line width of 0.6 gauss and that it undergoes rapid decomposition. Fbwles, McGregon,and.Symons have observed epr resonance in metal ammine solutions (3h), They have shown that lithium in methylamine gives a strong res- onance, potassium in methylamine and ethylenediamine gives a weak reso- nance but no resonance in propylenediamine, and sodium did not give a resonance in ethylenediamine or propylenediamine. Windwer and.Sundheim have measured the line width and g values for sodium, potassium and rubidium in ethylenediamine (35). The Q values for the metals are the same, within experimental error, but the line widths are much different. The rubidium line width is 8 times that of sodium which is approximately 7 times wider than that of potassium. Cafasso and Sundheim have noted the presence and absence of epr resonance in metal solutions in some organic solvents (36). Paramagnetic resonance was observed in solutions Of rubidium in bis (2-(2methoxyethoxy)ethyl) ether with a line width of approximately 5 gauss. Frozen solutions of potassium in 1,2—dimethoxy— ethane also gave a resonance but none was observable in the liquid. IParamagnetic resonance could not be found for solutions of potassium in 1,2-dimethoxyethane, sodiumspotassium alloy (70 wt % K) in 1,2-dimethoxy- ethane, rubidium in 1,2-dimethoxyethane, potassium in 1,2 bis(2—methoxy- ethoxy)ethane, or potassium in bis(2-(2-methoxyethoxy)ethyl)ether. The Knight shift measurements on sodium-ammonia solutions have in- dicated considerable electron density at the sodium and nitrogen \ 10 nuclei (11,12) but no detectible density at the protons (11). The electron density at the sodium and nitrogen nuclei are temperature and concentration dependent. As the concentration of the solution decreases, the electron density at the nitrogen nucleus increases but the density at the sodium nucleus decreases. When the temperature of the solution is lowered, the electron density at the sodium and nitrogen nuclei decreases. The Overhauser effect has been measured for sodium-ammonia solutions (37). In an Overhauser experiment, the nuclear resonance is observed, witheuxiwithout equal populations in the electron-Spin energy levels. If there is an interaction between the unpaired electron and the nucleus and if the electron—Spin energy levels have equal populations, there can be an enhancement of the nuclear resonance signal compared with the sig- nal obtained when the electron energy levels are in thermal equilibrium. This enhancement arises because the electrons can provide an additional relaxation mechanism for the nucleus. The Overhauser experiment is per- formed by applying a large radiation field to the electron Spin system such that the electron Spin resonance condition is met and observing the nuclear resonance while the electron resonance is saturated. The measurements of Carver and Slichter on the proton enhancement in sodium- ammonia solutions Show considerable electron density at the nucleus. They observed an enhancement of 150 for the proton resonance in a 0.hN solution and an enhancement of 500 in a 0.9N solution. While signifi- cant.e1ectron density at the proton was not indicated by the Knight shift or the epr relaxation times in deutero-ammonia and ammonia, the results of Carver and Slichter leave little doubt that there is consid- erable electron density at the protons. 11 0f the several models that have been proposed to fit the experimental data two have gained prominence, the Becker Lindquist Alder (16) and the cavity model (15). The Becker Lindquist Alder model considers the solute to consist of four species in equilibrium. The first of these, a monomer unit, which is a neutral species, consists of a metal ion surrounded by approx- imately six solvent moleculaywfiheuielectron circulation on the protons of the solvent molecules "coordinated" to the metal ion. In addition, a solvent metal ion and an electron are produced by dissociation of the monomer unit. Finally, a dimer unit is formed consisting of two monomer units bound together by the quantum mechanical exchange forces of the electrons and van der Waals' attractions. The model prOposes the fol- lowing equilibria: + - M<:> M++€— K1=[M [e] —> ‘where M is a monomer, M+ a solvated cation, e- a solvated electron, and M2 a dimer. Schmidt has carried out small angle x-ray scattering experiments in approximately 1 molar metal ammonia solutions (38). The data on these experiments indicated the presence of scattering centers of approximately 15 A diameter. The size of the scattering center depends on the metal ion, in agreement with the Becker Lindquist Adler model. The model can qualitatively account for the magnetic susceptibility, the density, the conductance in dilute solutions, and the Knight shift. l2 Blumberg and Das (39), and Pitzer (h0) have applied, with some success, quantum mechanical calculations to Show that the Knight shift is consis- tent with the Becker Lindquist Adler model. The cavity model was first proposed by Ogg in l9h6 (bl) and improved by Kaplan and Kittel (15). The cavity model considers the metal to be completely ionized into metal ions and electrons. The electrons are presumed to be in cavities within the solvent and to be in equilibrium according to the following reaction _ —+> = 2e > > Wm. >m m : a HHH a . tare , as p... H .8: C608 >. . i x. . 0 H430.” I]. ..I .lllJ i. 2 8. o 4 . + olll H. n .nmmo. H. a 1 an s. _ swag < m: aao coma .. as Haemam P Sam 30 higher oscillator powers are employed. The nmr sample was placed be- tween the absorption cavity and one of the magnet's pole faces with the nmr sample parallel to the paramagnetic sample. The oscillator output was impressed across the vertical plates of a Du Mont Type 30E A Oscilloscope as well as across the recorder input. Before entering the recorder, the oscillator output passed through a full wave diode rectifier to convert the a-c voltage to a d-c voltage. In this way, the proton absorption was visible on the oscilloSCOpe and was simultaneously recorded on the chart. The frequency of the proton oscillator was measured with a Hewlett Packard 52E Electronic Counter containing a 525 A Frequency Converter Unit. The magnetic field at the epr absorption maximum and the line width of the epr absorption were determined from two proton markers, one on either side of the epr absorption, Fig. 2. The method used to determine the line width and g value is illustrated in Fig. 2. To determine the specroscopic splitting factor, 9, it is necessary to know the klystron frequency as well as the magnetic field at the absorption maximum. A Waveline 698 Wave Meter that had been calibrated to an accuracy of at least i1 mc was used to measure the klystron fre- quency. The frequency measuring apparatus in Dr. R. Ht Schwendeman's microwave laboratory was used to calibrate the wave meter. The .klystron frequency was determined by measuring the difference between the accurately known frequency, a harmonic of an oven-controlled 1 mo czgrstal, and the klystron frequency. The micrometer reading on the watna meter was recorded for each 10.0 mc change of the klystron. The acetrracy of the calibration was limited by the precision attainable in .mdmn no aaama a ear ream: mead .N .aaa mmamm H.mmmm mm, III So m4.wa IIIIIIIIwJ _ . . _ _ _ emoo.m u a . " mafia Baez. u a . _ was 4.82 a Team + as nil Rama m9 _ - mowwmm I — _ . . i ‘v r _ m sea Rama ea 85;: woman m.mmmm us NH04.:H magma aa.a u aaeaz eeau_w . Eom.H oE .momm zoconoopm coupmxax w.©~ Hmmqug u nova: mama . 32 reading the micrometer. The spectroscopic splitting factor can be found by using the following formula with the appropriate constants (59): h‘u s 98H g = 0.7111115 x 10"5 'V/H where 1);; klystron frequency in cycles per second. The nmr markers and the epr absorption were recorded on the chart by switching the recorder input alternately between the full wave diode rectifier and the 100 KC field modulation and control unit, V-E560, while the magnetic field was being swept linearly. After the first nmr marker had been recorded and the oscillator frequency noted, the recorder input was switched to the 100 KC Unit and the epr absorption recorded. The nmr oscillator frequency was changed so that the nmr marker could again be recorded after the epr absorption. The magnetic field was modulated with E0 cycle per second field modulation during the proton absorption by adjusting the V-E250A sweep unit to give a field modulation of two to six times the proton line width. The E0 cycle per second field modulation was employed during the proton absorption only. B. Glassware Cleaning The solvent make-up train and the metal make-up vessels were cleaned with either hot chromic acid cleaner or a hydrofluoric acid cleaner. The hydrofluoric acid cleaner was prepared according to the following formula: 12% acid stable detergent 33% concentrated nitric acid 5% hydrofluoric acid 60% water. 33 After the acid soak, the glassware was rinsed approximately twelve times with distilled water and demineralized water and finally dried in an oven operating at lEOOC. The sample make-up vessel was cleaned by filling the vessel with fuming nitric acid and allowing it to stand at least two hours. The vessel was rinsed six times with distilled water and at least six times with demineralized water and finally dried in a 1E0°C oven. C. Metal Purification The sodium was obtained in a hermetically sealed can as a large dry piece from J. T. Baker Chemical Company. The potassium, from Mallinckrodt Chemical Works, was in the form of 1/2 inch rods stored under petroleum hydrocarbons. The metal was cut from a large piece with a stainless steel knife and forceps. After the metal was trans- ferred to a degassing tube, Fig. 3, the tube was connected to a mechan- ical roughing pump through cap A sealed with Apiezon W wax. When the tube was evacuated, the metal was gently heated until it flowed through the constrictions into chamber B leaving the oxide at the con- strictions. Chamber B was sealed off under vacuum, when the metal transfer was completed. The metal make-up vessel, Fig. 3, was attached to the high vacuum line using a 12/5 ball joint sealed with Apiezon W wax. Chamber B from the degassing tube was broken and quickly transferred upside down into the metal make-up vessel. .After capping and sealing the 2E/E0 taper with Apiezon W wax, the system was evacuated to 5 x 10'6 mm pressure at which time the metal make-up vessel was flamed with a gas-oxygen 3E m mesh. 3 Bamm NH on o IlilIlL m\ma a r annom> cmwm oh EEOH .Hemma>,aa-exmz Haas: ear ears aeammmaaa Haas: .m .aae \/ nubsmeu use seem og\am m uoeamcu mac Hmam . a sac gang ssoom> ado Emacs: OH. gfiam 3S torch to drive off volatiles. The metal was heated until it flowed through the constrictions into Chamber C. The upper constrictions were then.removed by sealing Off the lower constriction under vacuum. The metal was repeatedly degassed until the pressure stabilized at less than 2 x 10'5mm at which time the metal was distilled into the sample tubes. The sample tubes were removed by sealing under vacuum. The rubidium from Fairmont Chemical Company and the cesium from varlacoid Chemical Company came as sealed ampoules. The ampoules were .cooled in liquid air, broken, and immediately transferred to the metal .make-up vessel. It was necessary to cool the samples to avoid exces— sive oxidation of the metal during transfer. The metal make-up then proceeded as described above for sodium and potassium. D. .Ammonia Purification Ammonia, obtained from 01in Matheson Company Inc., was distilled from the commercial cylinder into a four liter evacuated tank that had been charged with sodium metal. The condensation was effected by cooling the small tank with a dry-ice and alcohol mixture. Sodium.metal was cut from the center of a large piece and placed into vessel A of the purification train shown in Fig. E. The metal was outgassed by heating under high vacuum until the pressure stabilized at»2 x 10'5mm. The vacuum rubber tubing connecting the small ammonia tank and the purification train was evacuated, flushed.with ammonia, and evacuated several times before the ammonia was condensed into ves- sel A. A dry-ice and trichloroethylene bath was used to cool the ves- setls. The pressure in the line connecting the ammonia tank and vessel A knas maintained above atmospheric pressure during distillation. The 36 Eooom> nmwm oh ii) iii. 0 (n om}: canoe eoabnoauanan .4 .aaa mmmao acne dead mwcoafid ESSom> oh awowcmcooz DH nupoeocmz ow \ a 37 sodium-ammonia solution was allowed to stand in vessel A for several hours and stopcock B was periodically opened to a vacuum pump to remove any volatile gases. Vessel B was cooled and the bath on.A removed to effect the transfer of the solution into vessel B through the glass frit. When-the liquid level was just above the glass frit, the cold trap was replaced on vessel A to stop the transfer and stopcock C was closed. The ammonia stood in vessel B for several hours before it was used. Just prior to distillation the ammonia was pumped on again to remove gaseous impurities. E. Methylamine Purification The monometbylamine obtained from Matheson Companx,Inc. was puri- fied by two different methods. The first method was fractionation through a two-foot column packed with 1/E inch glass helicies, Fig. 5. After the system was evacuated, methylamine was condensed into flask #1 from the commercial cylinder by cooling the flask with a dry-ice alcohol bath. When 600 ml. had been condensed, flask #1 was allowed to warm up and the column's temperature was set at -15° to —25°C by ad— justing the temperature of the bath from which the alcohol for the column was obtained. The cold finger was cooled with dry-ice and alco- hol, and a dry-ice alcohol bath was raised around collecting flask #2. After-about six hours, 60 m1. of solution had been collected in flask #2. The solvent purification train, Fig. E, was evacuated to 2 x 10‘5mm of pressure after vessel A had been charged with potassium and lithium. Approximately 200 ml. of methylamine was condensed from flask #1, Fig. 5 into vessel A of the purification train. The manipulation of the 38 Cold Finger To Mechanical Vacuum.Pump I A gnnmmnus monumental #1 Fig. 5. Fractionation System 39 solution in the purification train then proceeded as described for ammonia. The second purification method, Fig. 6, was similar to that employ— ed in Evers' laboratory (60). After the system was evacuated, 600 m1. of methylamine was condensed from the commercial cylinder into flask #1. Flasks #1 and #2 of the system contained clean lithium. The cold finger was charged with dry—ice and alcohol and the dewar was removed from flask #1; the methylamine was allowed to reflux four hours. The heat from a 250 watt infrared heat lamp was impressed on the solution during the final E5 minutes of reflux to accelerate the process. After reflux, the cold finger was allowed to warm up and 100 m1. of solution was blown off through #E of Fig. 6. Approximately E00 ml. of methylamine was distilled into flask #2 with the last 100 ml. in flask #1 being discarded. The methylamine remained in flask #2 at -78°C for 68 hours after which it was allowed to warm up and 100 m1. of solution was blown off through #5 of Fig. 6. The solvent was then distilled into vessel #3 until 200 m1. had been transferred, the last 100 ml. being again discarded. Methylamine was distilled from vessel #3 into vessel A of the puri- fication train Fig. E. The manipulation of the methylamine in the puri- fication train then proceeded as described for ammonia. The infrared spectrum of the methylamine vapor was recorded with a Perkin Elmer Model 21 Recording Infrared Spectrometer and a Beckman Model I R 7 Infrared Spectrometer with sodium chloride prisms. A ten centimeter gas cell filled to 55 mm pressure was used for the infrared determination. The infrared Spectra indicated that the two;purification E0 .Eopmtam cowbmoamusd .muo>m .0 .mE me. Q mm 4% uoHCH >3 133m >3130Hm unnsouog unhdouoz all. 1 -...o , 4 I a a SSMMM P . a". home: Hmoflcmeoofi . J . r) Boo OH. )1.\ TIM . g . El methods gave methylamine of comparable purity but that neither method completely removed the ammonia. Ammonia has strong absorptions at 932cm“1 and 968cm‘1 which make it easy to detect small amounts of am- monia (61). The methylamine samples from the purification procedures were com- bined by distilling them into an evacuated nitrogen bubbling system, Fig. 7. Dry-ice and alcohol was added to the cold finger and the dewar was removed from the flask when the nitrogen bubbling commenced. During three hours of bubbling prepurified nitrogen through the methylamine under reflux conditions, one third of the solution was lost. One third of the solution was then distilled into an evacuated transfer flask and the final third was discarded. The methylamine was distilled from the ‘transfer flask to an evacuated vessel containing degassed potassium. This methylamine was used for the lithium, potassium, rubidium and cesium solutions. The infrared spectrum of the methylamine obtained after the nitrogen bubbling did not show any detectible absorption due to ammonia. F. Ethylenediamine Purification Ethylenedfimfineobtained from Matheson Chemical Company Inc. was listed as 98-100% purity. A 500 ml. flask was filled with ethylene- diamine and the flask was attached to the nitrogen bubbling system ; Fig. 7. To effect the distillation of ethylenediamine into the evacet.fl uated system, a 50°C bath was placed on the distilling flask and an ice water bath on the collecting flask with a heating tape operating at 55-6000 wrapped around the connection between the two flasks. When E00 ml. had been condensed in the flask of the nitrogen bubbling system, E2 Annaofim annouoz soonam menaaam coach; .s are poHCH comoupwz mesa sonom> acumcmcooz oh m\ma C I A homomm fllllk taco E3 the nitrogen bubbling was commenced and the solution was heated with a 250 watt infrared heat lamp. After twelve hours of bubbling with pre- purified nitrogen, 150 ml. of ethylenediamine had been evaporated, then 150 ml. of ethylenediamine was distilled into a transfer flask. The transfer flask was attached to the long purification train Fig. 8, with vessels A and C containing degassed potassium. When the pressure had stabilized at less than 2 x 10'6 mm, 100 ml. of ethylene- diamine was distilled from the transfer flask into vessel A using a hot water bath, heating tape, and an ice-water bath. The ethylenediamine remained over potassium metal in vessel A for eight hours at room temper- ature and eight hours at 50°C. The solution was pumped on periodically to remove volatile gases. A liquid transfer of the potassium-ethylene- diamine solution was made through the glass frit into vessel B. The potassium-ethylenediamine solution stood at room temperature for two hours and at 50° for four hours before distillation into vessel C. After 70 ml. had been condensed, vessel C was removed and attached to the final storage vessel. The potassium used invvessels A and C of the long purification train had been degassed but not distilled. The potassium in the final storage vessel Fig. 9, was distilled into the vessel from a metal sample tube (see metal purification). Ethylenediamine was distilled from ves- sel C into the evacuated final storage vessel. The metal solutions for study were made from this ethylenediamine. The infrared spectrum of ethylenediamine vapor in the ten centimeter' gas cell showed no absorption attributable to ammonia. Eh I 11m. flea nnflo .538 ceapmoamuow 954 U ./ .w a: E5 m\fi Hommo> ommuoum Hug...“ .m .mwm use Heb 0:38 1 \ Honnwwommmoxnz cabs f Baa m SENS , 5 B roam ca. )\ r1 (1 E6 G. Sample Preparation The sample make-up vessel Fig. 10, was attached to the high vacuum system by means of a 12/5 ball joint using Apiezon W wax. In the prep- aration of sodium and potassium solutions a metal sample tube was broken and quickly inserted upside down into the make—up vessel. Rubidium and cesium samples were first cooled in liquid air, broken, and inserted into the sample make-up vessel. The 2E/EO taper was then capped and Sealed with Apiezon W wax and the system evacuated. Lithium, obtained from the Lithium Corporation of America, was in the form of l/E inch rods stored under mineral oil. Small pieces of lithium were cut with a stainless steel knife and forceps from the cen- ter of a large piece under dry purified benzene. The benzene had pre- viously been fractionally distilled and stored over sodium and lithium. The metal was transferred to the sample make-up vessel while the vessel was being swept out with argon. Calcium from the Fisher Chemical Company and strontium and barium from the Fairmont Chemical Company were cut from large pieces with diagonal cutters under dry benzene. When the pressure reached 2 x 10'6 mm, the sample make-up vessel was flamed with a gas-oxygen torch to drive out volatiles. The metals that could be distilled in Pyrex glass (sodium, potassium, rubidium, and cesium) were distilled into the sample tubes by gently heating the metal. The solvent was then distilled into the sample tubes by cooling the tubes with a dry-ice alcohol bath for the ammonia and methylamine distillation and an iceéwater bath for the ethylenediamine distillation. When the prOper amount of solvent had been distilled into the sample E7 moQSH oaqemm OH ca 4 / 33a Eoumhm Hommo> nouoxmz oaaawm .2 are E8 tubes, the tubes were cooled in liquid air and.sealed off under vacuum. In the preparation.of lithium, calcium, strontium, and barium solution, which cannot be distilled in Pyrex glass, the solvent was con- densed Onto the metal by cooling part D of the sample make-up vessel. When adequate solvent had been condensed on the metal, the sample tubes were cooled and the dewar removed from.part D. The vapor pressure of the solvent pushed the solution into the sample tubes. The sample tubes were cooled in liquid air and the excess solvent was evaporated from the vessel. The sample tubes were then sealed off under vacuum. VII. RESULTS The concentration dependence of the results reported here is quali- -tative in nature because the concentration of the solutions was not de- termined. The change of line width with concentration was followed as the sample decomposed. The samples had been at room temperature for 20 to 30 minutes before the first Spectrum.was measured. The initial con- centrations varied from one sample to another but the final Observations were for samples with comparable weak absorptions. Attention should be called to the fact that the decomposition times varied considerably from one metal to another in a particular solvent. In spite of the varia- tion from sample to sample, the measurement of line width for any given sample should be accurate to about i10%. The temperatures reported here were measured by a thermocouple near the bottom of the dewar that passed through the microwave cavity. A temperature gradient existed in the dewar so that the measured temperatures were lower than those of the sample. .A. Solutions in Ammonia Calcium, barium, and strontium are soluble in ammonia. Samples of these metals were prepared and their line widths and g values measured at room temperature on the homebuilt epr spectrometer. The line width for these metals decreased as the concentration decreased. The minimum line width for these metals is approximately 0.06 gauss. Studies were made on concentrated sodium, potassium, rubidium, and cesium solutions in ammonia. These metals, with the exception of cesium, form two phases at high concentrations and low temperatures in ammonia; the two phase region was investigated. Large samples were prepared so E9 50 that it was possible to investigate the blue and the bronze phases simultaneously or singly in the microwave cavity. The results showed the bronze phase to give a much broader epr absorption than the blue phase. The absorption from the bronze phase is asymmetric with the.lowsfield side of the first derivative spectrum w larger than the high-field; i.e., the low-field side is farther from the base line than the high-field side. This result was independent of the direction of magnetic field sweep. The Spectrum of the blue phase was a symmetric Lorentzian curve. The peak-to-peak distances reported for the blue phase were measured from spectra such as those in Fig. 11 to 1E, However, the rate of magnetic field scan was much slower for the peak-to-peak measurements of the blue phase than in the Spectra of Figs. 11 to 1E. The sodium spectrum in Fig. 11 was obtained with the bronze and blue phases together in the cavity at -55°C. The broad asymmetric curve from the bronze phase has the low-field side approximately 2.5 times farther from the base-line than the high-field side. The peak~to-peak distance of the asymmetric curve is 3.3 gauss. The spectrum of the blue phase is the sharp absorption on the broad curve and has a peak-to-peak dis- tance of 0.15 gauss. The spectra in Figs. 12 and 13 are those for potassium in ammonia. Both the bronze and the blue phases were in the cavity when the spectra were taken. The spectrum in Fig. 12 shows a broad asymmetric curve with the low-field side approximately 2.5 times larger than the high- field side. The peak-to-peak distance of the asymmetric curve is 5.0 gauss. The absorption due to the blue phase has a peak-to-peak distance of O.E gauss. The temperature of the solution was -75°C. 51 adupooam m~CoEE¢ : fiasco O m A .HH .mmm 52 A Edupooam mwcoead n aswmmwuom .3 a: mafia ommm 53 £3.36on 3:052. u sommmwpom .9 are u: Magma-i SE The spectrum of Fig. 13 shows potassium in ammonia at -75°C. The appearance of the two lines, one on either side of the absorption due to the blue phase, was strongly dependent on temperature, a change of 3°C would result in their disappearance. The extra lines were not present in all samples but were present in some. The peak-to-peak dis- tances are; for the broad asymmetric curve due to the bronze phase, E.3 gauss; the curve due to the blue phase, 0.3 gauss; the extra curve on the low-field side, 0.1 gauss; the extra curve on the high-field side, 0.05 gauss. Solutions of potassium in ammonia that contain excess metal give an asymmetric epr absorption. The low-field side of the first deriva- tive curve is about 2.5 times farther from the base line than the high- field side. The peak-tO-peak distance decreases from E2 gauss for a sample containing a COnSiderable excess of metal, to 10. gauss for a sample containing less excess metal. The epr spectrum of the two-phase rubidium solution at -75°C is shown in Fig. 1E. The broad asymmetric curve is due to the bronze phase and has a peak-to-peak distance of 72 gauss. The narrow absorp- tion of the blue phase has a peak-to-peak distance of 1.1 gauss. The dilute solutions gave a symmetric Lorentzian curve with a line width of 0.03 gauss. Cesium does not form two phases, and only one absorption curve was observed. At high concentration the cesium curve is asymmetric with the low-field side approximately 2.5 times larger than the high- field side. The peak-to-peak distance decreased from 1.1 gauss as the concentration decreased because of decomposition. 55 £336on mmcossd .. Signom .aa .maa . 56 B. Solutions in Methylamine The infrared spectrum of the methylamine used for the lithium, potassium, rubidium, and cesium epr spectra did not Show any absorp— tions due to ammonia while the methylamine used as a solvent for sodium did show small absorptions in the infrared.which were assigned to ammonia. Lithium is quite soluble in methylamine and gives a large epr signal in the more concentrated solutions. The epr spectra of both the liquid solution and the frozen samples were Lorentzian. Data for typical samples are given in Table I and Fig. 15. The line width of the samples increased approximately 20% as the samples decomposed at room temperature. The line width of the frozen samples increased as the temperature increased until the samples approached the melting point, the line width then decreased as the samples passed through the melting point. A very pronounced difference exists between room tem- perature line widths of samples immediately after they had been frozen and remelted, and the line width after they had been at room temperature for some time. This indicated the existence of complex changes in the solution which do not reach equilibrium rapidly. The solubility of sodium in methylamine is very low. The results reported here were obtained with the methylamine purified by the frac- tionation procedure described earlier; this methylamine was not cem- pletely free of ammonia. The line width of the frozen samples increased as the temperature increased until the temperature approached the melt- ing point. Near the melting point, no Spectrum was detected but the Table I: LithiumeMethylamine 57 Decomposition 5.21mi: m l 1.32 gauss 3 1.05 Sample 1 Temperature Line Width -150°C E.E gauss -1E8 E.2 -l33 5.2 -112 5.2 -105 E.8 -1oo 2.E - 98 1.0 - 88 0.76 - 70 0.63 - 59 0.53 - 52 0.51 - 51 0.50 - 38 0.52 - 19 O.E9 - 1 0.50 + 19 0.57 Time Interval Between First and Last Spectra Final During Decomposition 1.67 gauss 85 minutes. 1.56 70 Sample E Temperature Line Width -133°C E.7 gauss -118 5.3 -109 5.5 -102 E.S - 88 2.E - 81 1.E - 62 1.1 - 38 0.95 - 1E 0.89 + 17 0.87 58 ceaseless: - sass: “awesomeness as 52: 83 .3 am Us 1. ouopwpodaoe 03 u 03.. 03 a 00.7 om: om- 0;: Own. 0 Om _ _ _ _ _ _ _ _ o iuAJ/ko « T II. o; , IL, on IL, o.m . ....._ .. . I o.m . a a £95m a II 68 a 295m 0 ssmeB - lupin 9an 59 Spectrum was detected again at slightly higher temperatures. The line width of the solution broadened as the temperature increased. Lorents Zhn.epr absorption curves were Obtained for the saturated solutions, but it was difficult to determine the shape of the curves for frozen samples. To obtain a signal for the frozen material, it was necessary to over-modulate the signal so that the true shape could not be deter- mined. [ata.for sodium-metbylamine samples are given in Table II. Table II: Sodium-Methylamine Sample 1 Samle 2 1.E8 gauss Temperature 'Line Width 1.EE -lE0°C 1.72 gauss Sample 3 -11E 1.92 1.51 gauss - 93 no signal 1.36 - 88 no Signal - 73 no Signal - 60 1.23 - E3 1.211 - 22 1.3E + 1E 1.E3 + 22 1.E8 Solutions of potassium in methylamine gave a large epr Signal. The line width increased as the solution became less concentrated going from 0.73 gauss to 2.69 gauss. The absorption curves obtained were Lorentzian. The epr absorption obtained in frozen samples depended on 60 of how the sample was cooled. If a sample at room temperature was thrust into liquid air, and the Spectrum examined at -150°C, no absorption was obtained. However, if the solution was cooled slowly, a Spectrum was observed at —150°C. A sample which had been cooled very suddenly gave a spectrum when it had warmed up to -l35°C. and was then recooled to -150°C. Near the melting point, the Spectrum became very weak and could not be detected in some cases. When the sample was liquid, a narrow resonance was observed. The line width for the liquid increased as the temperature increased. Line width data for potassium-methylamine samples are given in Table III. C. Solutions in Ethylenediamine Ethylenediamine has a very limited temperature range, it freezes at 8.5°C. The measurements with temperatures below 5°C deal with solid samples. The solubility of lithium in ethylenediamineiS<1311(61) so that concentrated samples give a large epr signal. The line width of the Lorentzian absorption at room temperature decreased approximately 20% for 0.18 gauss as the concentration decreased because of decomposition. The line width for frozen samples decreases as the temperature increases from -l30°C, Fig. 16. However, at -79°C the Spectrum was so weak that it was just detectible; the spectrum of solutions more than five de- grees on either side of —79°C is again easily detectible. Table IV gives line width data for lithium. Table III: 61 Potassium in Methylamine Sample Concentrated l 0.73 gauss 2 0.90 5 0.81 Sample E,Rapid Cooling Temperature Line Width Dilute 2.69 2.13 Sample 2, Slow Cooling Sample 2,Rapid Cooling. Temperature Line Width Temperature Line Width -155°c -1E8 no Spectra no Spectra warmed to -130°C -l60 2E gauss -1Eo 25 -138 23 -lE7 23 -1E2 23 -126 27 -119 no spectra -100 narrow line -116 no spectra -107 1.17 - 96 0.51 - 90 0.5E - 79 O-SE - E1 0.60 - 26 0.68 - 18 0.72 - l 0.88 + 16 0.89 -152 -1E0 -138 -126 -118 -122 -138 -11E -10E 87 6E 58 - 22 + 20 °C 27 gauss 33 22 25 very weak 32 22 very weak 1.0 0.70 0.70 0.88 1.17 1.E6 -1E8°C -1E7 ~1EE 7133 -1E1 -1E3 Repeat Rapid —l50 —1E6 -1E6 no no no 19 17 18 no no no Spectra Spectra Spectra gauss Cooling Spectra Spectra spectra Warmed to -1300C -1EE -1E8 29 25 gauss 62 Opal ocHEmeocufizzpm n Esweumq monopmquth m> can“: scam Do u unopmuodspw Oman OOH: Om- Op: 0:. Own .2 are _ _ _. _ _ Om OH NH ssmefi - lupin aUI'I 63 Table IV: Lithium in Ethylamine Samplg Initial H [Fipgl Time interval between initial 1 0.18 gauss. 0.12 gauss and final -- 80 minutes Sample 2 Temperature Line Width + 10°C 7 0.22 gauss — 2 0.31 gauss - 35 E.O - 79 very weak - 85 9.3 -120 12.5 Y The line width of sodium in ethylenediamine increases from .33 to 0.6 gauss as the concentration decreases because of decomposition. The line width decreased as the temperature was raised from -130°C to -90°C, at which point a sharp decrease occurred in the line width of the sample. The change in line width is shown in the spectrum of Fig. 17. This Spectrum was taken while the change in the sample was occurring and thus does not represent a steady state situation. The absorption of the broad symmetric curve, of width 9.8 gauss, decreased in intensity as the intensity of the narrow absorption of width, O.E2 gauss, increased. This sample, when melted, showed an increase in line width to 0.60 gauss. The data for sodium in ethylenediamine are given in Table V. Potassium is.about three times as soluble in ethylenediamine as sodium, the epr absorption is also stronger for potassium. The Lorent- zian epr absorption obtained for potassium solutions broadened from 6E Subcomm o: mammcocofixcpm .. Eamon/... mAIIIIII .: .aE 65 Table V: Sodium in Ethylenediamine Time Interval Between Sample Initial Final First and Last Spectra l 0.29 gauss 0.5 gauss E 1/2 hours. 2 0.36 0.5 > E Sample 2 Temperature Line Width + 22°C 0.60 gauss + 10 0.62 O O.EE - 20 0.37 - 59 0.33 — 9o ' O.E2 - 90 O.E2 9.8 -100 13.7 -130 ' 23.3 Table VI: Potassium in Ethylenediamine Time Interval Between Sample Initial Final First and Last Spectra l 0.3E gauss 0.55 gauss >120 hours Sar_np 1e 2 Temperature Line Width + 230C O.EE gauss + 23 O.E3 + 10 O.El - 2 O.E3 - 2O O.EE - 36 no Spectra 66 0.33 to 0.6 gauss as the concentration of the solution decreased. Potas- sium in ethylenediamine gave no Spectrum below -20°C. The line width appeared not to change as the temperature increased from ~20°C. The. data for potassium solutions are given in Table VI. The line width of rubidium in ethylenediamine decreased from 8.7 to 3.9 gauss as the concentration decreased during decomposition. The absorption curves are Lorentzian. Rubidium gave no absorption below —29°C. The line width decreased from 8.0 gauss to 2.0 gauss as the temperature decreased from room temperature to -29°C. The rubidium data are given in Table VII. Table VII: Rubidium in Ethylenediamine Sample Initial Final Time Interval Between First and Last Spectra 3 8.7 gauss 3.9 gauss , 15 hours Sample 1 Temperature Line Width + 22°C 8.0 gauss + 22 6.E + 22 6.E + 10 E.1 + 10 E.1 - 2 3.2 - 13 2.6 -.29 2.0 67 The line widths obtained from solutions of cesium in ethylenediamine were much broader than for any of the other metals in ethylenediamine. The Lorentzian line width was 50 gauss at high concentrations and'inarr creased to 77 gauss as the sample decomposed. The cesium samples gave no epr Spectra below 0°C. The line width decreased from 85 to 52 gauss as the temperature increased from 0°C to room temperature. Table VIII gives the cesium data. Table VIII: Cesium in Ethylenediamine Sample Initial Final ‘ 2 50 gauss 77 gauss Sample 2 Temperature Line Width + 22°C 52 gauss + 22 59 + 22 63 + 10 85 0 85 - 20 no spectra D. Rubidium and Cesium in Methylamine Rubidium and cesium in methylamine gave unusual epr spectra. Room temperature spectra for rubidium are shown in Figs. 18 and 19 and for cesium in Figs. 20 and 21. The Six-line spectrum of rubidium and the eight-line Spectrum of cesium are asymmetric about the base line. The low-field side of the spectrum is farther from the base line than the 68 sateen. ceaseless: .. Sagas .2 am 69 m - 583: am 83338: soupoo .2 are 70 A 8393on 6583553: .. £3.30 .8 6.E 71 A esnbooam ocwsmfixrpoz I Sawmoo .HN .aaa UoOOHI onsBmEOanh Eoom 72 high-field side. This is the first instance in which metals in ammine solutions have shown hyperfine interaction. Rubidium has two isotopes, Rb3§ and Rb°7 with nuclear spin of 5/2 and 3/2 respectively. The Rb°5 would be expected to have a six-line spectrum and the Rb87 a four~line spectrum. Six lines are observed. Cesium has one isotope, 03133, with a nuclear spin of 7/2 which should give eight lines, and, indeed, eight lines are observed. The rubidium and cesium Spectra in Figs. 19 and 21 Show the presence of an extra absorption line. There are two peaks in the rubidium and cesium spectra that are usually much larger than the other peaks, and this enhancement must be due to an extra line. The effect of temperature on the spectrum of cesium in methyl- amine was investigated. As the Sample was cooled from room temperature, the extra line increased rapidly in intensity. At #8°C, the center hyperfine peak.was not detected, being completely masked by the extra line, leaving seven peaks. .As the sample cooled, the extra line con- tinued to increase in intensity until at -50°C the hyperfine Splitting was just detectible over the strong extra line. At -70°C the hyperfine Splitting was again visible although with a much reduced line width. The extra line was very intense at -70°C and was again masking the center hyperfine peak. The peak-to-peak distance of the extra line at -100°C had been reduced to 2.7 gauss. The tOtal hyperfine separation (the distance between the extreme lines of the Spectrum at —lOO°C) was only 25% of the value at room temperature. The two Spectra in Fig. 21 are for room temperature and —100°C. The two spectra were obtained us- ing the same magnetic field sweep rate. The gain of the amplifier used for the spectrum at 7100°C was 1/E of the gain used for the room temper- ature spectrum. Returning the sample to room temperature from ~100°C 73 again showed the spectrum characteristic of room temperature. The g value of the extra line at -100°C was 2.0017. The frozen sample did not give an epr spectrum. The effect of temperature on the total hyperfine separation and the extra line is illustrated in Fig. 22. The hyperfine separation de— creases by E as the temperature decreases to -lOO°C and the line width of the extra line decreases approximately 20 as the temperature de- creases to -100°C. E. Measurements of g Value The 9 value measurements are given in Table IX. The uncertain- ties given are the experimental uncertainties in the measurements. The values given fOr sodium, potassium, calcium, barium, and strontium in ammonia were determined on the homebuilt epr Spectrometer. No attempt was made to measure the g values of the symmetrical Lorentzian curves for sodium and potassium in ammonia on the Varian spectrometer. The Q: values given for rubidium and cesium in methylamine are the values for the center of the spectrum. Because of the asymmetric curves, these values are less accurate than the other values given. 7E 658.3553; .. E538 mousbmuodsg 5H3 65H muuxo can Ho v.53: on: o5 cam compmumaom season? 133 on» E omcmzo .Nw .mE ... Do u undamuoaaoh OOH 1 Om 1 00.. 041 ON 1 0 ON _ _ _ _ m CH _.||l AA 0 . o mH I S b S u . A nnawooN ill _ m U. mm ill m t on i A - a . .. f O ..m mm l m w A m 04 .III a .1 L ma ,lllcombmuwaom cannon»: .238. < we: ..mupfim o . - A _ ONH 3.3 8 0.1 Total Hyperfine Separation - gauss Cam 0 co N O N m 0 \o m 75 Table IX: Measurements of 9 Value Ammonia Methylamine Na* 2.001 t 0.001 Li 2.0019 x 0.0002 K* 2.001 r 0.001 Naw 2.0017 i.0.0002 Rb 2.0003 i 0.0002 K 2.0017 i 0.0002 CS ---- Rb 1.998 t 0.003 ca* 2.001 t 0.001 Cs 2.00 r 0.01 Ba* 2.001 r 0.001 *- Sr 2.001 t 0.001 Ethylenediamine 2.0021 : 0.0002 2.0017 i 0.0002 2.0018 : 0.O@O2 2.0017 : 0.0005 1.998 x 0.003 * Homebuilt epr Spectrometer. VIII. DISCUSSION The metastable blue solutions of metals in ammines have many char— acteristics in common but few quantitative similarities. An important Similarity is the decomposition of the sample by reaction of the metal with the solvent to produce hydrogen. The decomposition rate varies tremendously; samples of sodium and potassium in ammonia have remained blue for up to five months at room temperature while samples of lithium in ethylenediamine remained blue at room temperature for only 90 min— utes. The frozen samples appear to last indfifinitely. Rapid decomposi- tion can make qualitative measurements difficult and quantitative measurements almost impossible. A considerable body of information has been amassed on alkali metal—ammonia solutions, but the available information on other metal- anmine systems is limited. Many of the published papers dealing with electron paramagnetic resonance in metal—ammine solutions are discussed in the historical section of this thesis. The information reported here for the methylamine and ethylenediamine systems is the result of introductory work and many of the observations noted are qualitative in nature. These results require verification and extension and sug- gest many new problems for future research, a few of which are given at the end of this discussion. Some of the complicating factors, which are possible avenues for future work, cannot be sorted out at this time and make a complete interpretation of the research impossible. Some of the results reported need further experimentation to definitely establish the trends noted. 76 77 The number and nature of the paramagnetic species present in the solutions are unknown, but such information would be most useful in understanding the systems. There is no doubt that at least two para- magnetic species exist in solutions of rubidium and cesium in methyl- amine. One of the paramagnetic Species in solutions of rubidium and cesium in methylamine ultimately involves the metal nucleus. There appear to be Species in the lithium-methylamine solutions which are not in rapid equilibrium. It is not uncommon for several workers to report different results when studying metal-ammine Systems. Cafasso and Sundheim (36) were unable to observe infrared absorption of metals in dimethoxyethane but Ihinton Q a}: (63‘) were able to detect absorption in the infrared. The . Optical absorption spectra of metal-ethylenediamine solutions reported by Windwer andESundheim (35) are different from those found by Dewald and Dye (62). The line width measurements reported here are different, in some cases, than those reported by Levinthal gt §£:(33), and Windwer and Sundheim (3S). Fowles (3b) and Windwer (35) were unable to prepare lithium-ethylenediamine solutions stable enough to permit measurement of the epr aspectra, but our solutions were blue for 90 minutes. These differences in optical Spectra and other prOpertieS may be due in part to the differences in stability of the solutions. The Species in the solutions are not in rapid equilibrium with each other so that decom- posing solutions may not have the same composition from sample to sample. Thus, it becomes understandable that different workers obtain different results for systems as unstable and as complex as these. 78 A. Solutions in Ammonia The epr results for calcium, barium, and strontium in ammonia are consistent with the results obtained by other investigators for dilute solutions of the alkali metals in ammonia. (The epr measurements for the two phase region of metal-ammonia solutions Show a distinct differ- ence in the absorption of the two phases. The less dense bronze phase has a higher concentration of metal than does the denser blue phase. For all metal-ammonia solutions Showing two phases, the lowafield Side of the first derivative curves for the bronze phase is about 2.5 times farther from the base line than the high-field Side. The same type of asymmetric curve was also observed for saturated potassiumeammonia solutions in the presence of excess metal and for concentrated cesium- ammonia solutions. Dyson has developed a theory which deals with the diffusion of the conduction electrons in a metal into and out of the thin Skin into which the microwave field can penetrate (6b). This theory was applied by Feher and Kip to diSperSionS of lithium, sodium, beryllium, and cesium in mineral oil, and to Sheets of the alkali metals (65). They Show that for a thick plate, as TD/T2-—eq>xn, the low-field Side of the first derivative curve, Should be 2.5 times farther from the base line than the high—field Side. (TD is the time it takes an electron to dif- fuse through the Skin depth, and T2 is the Spinéspin relaxation time.) The largest value of TD/TZ experimentally measured by Feher and Kip was about 0.5 for lithium sheets at 29601:. For this value of TD/TZ the low- field side of the derivative curve was about four times as far from the base line as the high—field Side, in agreement with theory. 79 Parkas reports that the conductance of saturated sodium-ammonia solutions is only a factor of 50 lower than that of pure sodium (66). For a sample diameter of 1 mm, the Skin depth (approximately 0.02 mm) is much less than the diameter of the sample. The theory of Dyson (6h) for TD/T2 large thus appears to agree exactly with our observations on con- centrated metal—ammonia solutions since the predicted factor of 2.5 is observed in these systems. The large value of TD in the solutions is not unexpected since the conductivity is lower than that of the bulk metal by a factor of 50. By using smaller sample diameters, it would be possible to definitely establish the validity of this explanation. These observations furnish a test of the Dyson theory for a different type of sample than that previously employed. The peak-to-peak distance for sodium is 3.3 gauss; potassium, approximately h.5 gauss; rubidium, 72 gauss; and cesium, 1.1 gauss. It should be noted that for solutions of cesium in ammonia, no attempt was made to scan the magnetic field over a large area to detect a broad absorption. Therefore, it is possible that a very broad absorption, in addition to the narrow one observed, is present in concentrated solu- tions of cesium in ammonia. The two extra lines that appeared in some of the two phase potas- sium—ammonia solutions are not easily explained. The Spectrum does not indicate the presence of a mixture of absorption and dispersion modes. The field modulation was below the line width of all the absorption lines so thatcnmx=m3dulation was not a factor. Potassium has a nuclear Spin of 3/2 which could give a four-line hyperfine spectra. If the blue phase absorption occurred at the center of the four-line spectra, 80 it could mask out the center two of the four-line Spectra. If potassium gives a hyperfine Splitting then other metals might be expected to Show similar phenomena, but none was.observed. There is the possibility that the extra lines were caused by a concentration gradient in the samples. However, the Spectra were reproducible and did not change as long as the temperature remained constant. B. Solutions in Methylamine A.large difference existed between the line width of lithium-methyl- amine samples at room temperature after they had been frozen and re- melted, and the line width after the samples had been at room tempera- ture for some time. This change in line width suggests the existence of Several Species in the solution which do not reach equilibrium rapidly. These could be two paramagnetic Species with different line widths or one which is paramagnetic and another which affects the line width. The difference between the areas under the narrow and broad absorptions is small, indicating that the change is probably not due to decompoSition of the paramagnetic Specie. Levinthal, Rogers, and Ogg (33) report that lithium in methylamine has a line width of approximately O.h gauss. This line width is below the value reported here.. The differences between frozen potassium—methylamine solutions which have been cooled at different rates indicate the presence of processes that are not in rapid equilibrium. Sudden cooling of the sample could result in so much line broadening that the absorption is undetectible. Alternately, a diamagnetic Species could form.which is 81 "frozen in" at low temperatures. Warming the sample to ~135°C might then allow the metal to precipitate giving an epr Spectrum character- istic of the metal. C. Solutions in Ethylenediamine The line width reported by Levinthal gt al.(33) for lithium in ehtylenediamine is 0.6 gauss. This value is much larger than the line width reported here, 0.12 to 0.18 gauss. The differences between the line width measurements reported here and those of Levinthal (33) may be due to differences in sample preparation or in the method of measur- ing the line width; information is not available on Levinthal's procedure. Windwer (35) and Fowles (3h) were unable to prepare solutions of lithium in ethylenediamine which were stable enough to allow measurement. This is understandable since it is difficult to get lithium as free from im- purities (oxide and nitride) as the other alkali metals which can be distilled in Pyrex. FOW1€5.EE.§l°(3h) reported that sodium in ethylenediamine did not give an epr Signal. Windwer and Sundheim (35) were able to get an epr signal for sodium—ethylenediamine Solutions and reported that the line width, 0.75 gauss, did not change as the sample decomposed over a three hour period. We found the line width, 0.29 gauss, not to change for the first two hours. However, after our sample had been at room temperature for h 1/2 hours, it was light blue and the line width had increased to 0.5 gauss. Because of the difficulty of preparing reasonably stable solutions the differences between results obtained by various workers is not as large as it might at first appear. Small amounts of impurities can catalyze the decomposition and could affect the measurements. 82 The change that occurred in the spectrum of frozen sodium-ethylene- diamine solutions at -90°C is unique for the ethylenediamine Systems. The broad symmetric curve at temperatures below -90°C appears to be due to the precipitated metal. The sample does not melt at ~90°C so a solid—liquid transition cannot be occurring unless a low-melting eutectic forms. However, sodium is only soluble in ethylenediamine to the extent of 3 x 10'3 molar (62) and could not show a eutectic at such a low tem- perature and concentration. The phase diagram for the system is not known, but a reaction in the solid state might be occurring. The line width for potassium-ethylenediamine solutions increases as the concentration of the sample decreases, varying from 0.33 to 0.6 gauss. We found the line width to be much broader than that reported by Windwer (35), 0.15 gauss. Frozen potassium—ethylenediamine samples did not give a detectible epr absorption. The samples were cooled both Slewly and rapidly but neither procedure gave samples that exhibited epr absorption. The line width reported by Windwer (35) for rubidium-ethylene- diamine, 7.5 gauss, is in agreement with the value reported here, 8.7 to 3.9 gauss, as the sample decomposed. Rubidium samples did not give a detectible epr absorption below ~29OC. Cesium-ethylenediamine solutions gave a very broad epr absorption. The line width increased as the sample decomposed and increased as the sample was cooled. D. Solutions of Rubidium and Cesium in Methylamine The nuclear hyperfine Splitting of rubidium and cesium in methylamine 83 is unique for metal~ammine systems; no other metal—ammine solution has been reported to Show this phenomenon. The.existence of nuclear hyper- fine Splitting shows that one of the paramagnetic Species in these solu- tions involves contact of the electron with the metal nucleus. The re- sults of the paramagnetic susceptibility (6) and the Knight shift measure- ments (11,12) at relatively high concentrations in ammonia have also in- dicated that the metal nucleus is involved with a paramagnetic Specie. A comparison may be made between the hyperfine Separation determined by molecular beam experiments for the atom, and the total hyperfine sep- aration of the epr experiments (distance between the extreme lines of the Spectrum). The electron density at the nucleus for the atom, deter- mined by molecular beams, and the electron density determined by epr are compared to find the percentageof time the electron iS at the nucleus relative to the free atom. The hyperfine separation in wave numbers is given by Fermi's formula (67,68): 3h. 50" = (37%;) uomflI/Zbflg-I-Ifl] where no is the Bohr magneton, u is the nuclear magnetic moment, I is the nuclear spin, andlI! 2(0) is the value of the normalized wave func- tion. The hyperfine separations determined from molecular beams by ..I Millman and Fox for rubidium and cesium are (70): Rb85 0.1018 cm‘1 Rba? 0.229 cm‘1 C6133 = 0.3067 cm-1 By using the approximate constants (60,69) in equation 3h one obtains for 472(0) for the free atom: 8h 2 - 1]} Rbas(e) = 2.63 x 1025 cm 3 Z _ ll} R1387(o) = 2.21; x 1025 cm 3 2 _ WC5133(0) = thé X 1025 CHI 3 The value of the normalized wave function at the nucleus may be calculated from epr results (71) by: _ 8n 2 35. (AH)total — (3411’ (0)11 where.(AH)total is the total hyperfine separation. Using equation 35 and the room temperature rubidium and cesium values for (AH)total gives 87 7 = » Z ' = 25 -3 Rb (AH)total 128 gauss 11811387“) 0.06 x 10 cm 133 = 2 = 25 » ‘3 CS (AH)total 39h gauss WITC8133(0) 0.18 x 10 _ cm For cesium, this gives an electron density at the nucleus of h.h% of the value determined by molecular beams. Calculating the electron density for rubidium is complicated by the presence of the two isotopes. The percentage and nuclear moments of Rb35 and Rb87 are 72.8% and 27.2%; and 1.35 and 2.7h respectively. This gives the four—line Rb“7 Spectrum a separation of 2.0 times the Rb35 separation and an intensity of 0.76 times the Rb85 intensity (72). This means that the outside peaks in the rubidium Spectrum must be due to the Rb37. Thus, the electron density at the nucleus for Rb57 is 2.7% of the value determined by molecular beams. The large change in total hyperfine separation for cesium with temperature, Fig. 21, gives an electron density at the nucleus of only 1.1% of the free atom value, at -100°C. This decrease of a factor of 85 four in the line width coupled with the increased absorption of the extra line is a dramatic demonstration of the temperature dependence of the relative concentrations of various Species in the solution. The line width of the extra line decreased by approximately a factor of 20 upon cooling to -1000C. These results leave no doubt but that at least two paramagnetic Species exist in solutions of cesium in methylamine. The electron density at the sodium nucleus calculated from the Knight shift experiments (12) for fairly concentrated sodium-ammonia solutions changes from 9% to(ll9% as the concentration decreases from 10 to 0.2 molal. However, epr eXperiments on sodium-ammonia solutions Show no Sign of hyperfine Splitting. The value obtained by epr measurements for rubidium and cesium are within the values calculated from Knight Shift measurements. E. Comparisons Some of the differences between the solvents are evident when the results obtained for a given metalina different solvent are examined. The line width of lithium is nearly a factor of 10 larger in ethylene- diamine than in methylamine. Potassium has a larger line width in methylamine than in ethylenediamine. The line width of lithium broadens as the solution decomposes in methylamine and narrows in ethylenediamine. The line width of potassium increases upon decomposition in both methyl- amine and ethylenediamine. Sodium, potassium, and cesium exhibit in- creasing line width as the samples decompose in ethylenediamine while lithium and potassium Show decreasing line widths. The line widths of the metals decrease with decreasing concentration in ammonia. It is 86 difficult to make any generalizations about the systems. Dr. J. A. Cowen had this to say about the systems, "There appear to be more dis- similarities than Similarities". The effect of the decomposition products on the epr absorption is not.known; it is possible that theSe products might affect the line width. Impurities such as oxide and nitride have an effect on the stability of the solutions but their role in determining the line width is unknown. A model for the system must recognize the possibility that more_, than one paramagnetic species must exist at least in some of the solu- tions. The model must allow for unpaired electron density at the metal nucleus, and, at least in ammonia, at the nitrogen nucleus and at the proton nucleus. This information must be combined with other data, such as absorption Spectroscopy and conductivity, to provide a basis for a general model. F. Future Work Some avenues for future epr work on metal-ammine systems are as follows: The effect of temperature on the line width and the extra absorption line in rubidium-methylamine solution would allow a compari- son with the cesium results. Quantitative or semi-quantitative measure- ments of the Spin concentration in the rubidium and cesium-methylamine solutions as a function of temperature would give an estimate of the effect of temperature on the equilibrium constant between the Specie giving the hyperfine Spectrum and that giving the extra line. 87 Quantitative measurements of'the paramagnetic susceptibility as a function of concentration and temperature-would give a value for the equilibirum constant between paramagnetic and diamagnetic species. Measurements in the very dilute region would tell if the system followed the Curie law. The Spin-lattice and Spin-Spin relaxation times would be of value in generalizing a theory for the metalrammine system. 10. ll. 12. 13. 1h. 15. 16. 17. 18. 19. 21. IX. REFEREIVICES Weyl, Ann. Physik, 1%, 601 (18611). A. Kraus, J. Am. Chem. Soc., g9, 1323 (1908). Freed and N. J-. Sugarman, J. Chem. Phys., fl, 3514 (19113). . Huster, Ann. Physik, _3_3_, 1177 (1938). . Huster, Ann. Physik, h, 183 (19118). A. Hutchison, Jr. and R. C. Pastor, J. Chem. Phys., 21, 1959 (1953). E. O'Reilly, Ph.D. Dissertation, University of Chicago, Chicago (1955). A. Kraus, J. Am. Chem. Soc., 36, 8611 (1,911.1). . Cw. EverS, A. E. Young II, and A. T. Panson, J. Am. Chem. Soc., 12, 5118 (1957). 1... Bye, R. F. Sankuer, and G. E. Smith, J. Am. Chem. Soc., . M. McConnell and C. H. Holm, J. Chem. Phys., 26, 1517 (1957). v. Acrivos and K. s. Pitzer, J. Phys. Chem. 66 1693 (1962). ’6...) . A. Kraus, E. S. Carney, and W. C. Johnson, J. Am. Chem. Soc., 59, 2206 (1927). . C. Johnson and A. W. Meyer, J. Am. Chem. Soc., ill: 3621 (1932). Kaplan and C. Kittel, J. Chem. Phys., 2;, 1&29 (1953). BeCker, R. H. Lindquist, and B. J. Alder, J. Chem. Phys., 2_5, 97'1 (1956)- ' L. Pollak, J. Chem. Phys., 2b 861; (1961). L. Hahn, Phys. Rev., E, 580 (1950). T. Blume, Phys. Rev., 102, 1867 (1958). A. Hutchison, Jr. and D. E. O'Reilly, J. Chem. Phys., .3_11', 1279 (1961). Bloembergen E. M. Purcell, and R. V. Pound, Phys. Rev., 3 679 (19u85. —’ 88 '(6561) 6E9 ‘6? "5693 'waqo 'r ‘JaU3Jor '(8761) W6: ‘9T"'SKQd 'wauo 'r ‘IIIH '1 ‘(ES6I) 35 ‘T3"'sfiud 'wauo 'r ‘qwoosdii 'N “(9861) 899 ‘39 "A88 'sfiUd "Jr ‘560 'v '(8561) {SW ‘63 "9693 'mauo 'r ‘122313 '3 '(6561) 153 ‘0? "sfiud 'wavo 'r ‘SEC 'd '1 Due Baaqwnia '3 '(1561) E3 ‘ZE"'sfiud 'mauo 'r ‘ipiwuos 'M a '(9561) 516 ‘30? "Aaa 'Sfiud ‘JaquIIs 'd '0 DUB IaAJeo ' '(656I) 609 ‘TE "8693 'wauo 'r ‘WIaupuns 'H '8 Due 0888380 ' '(3961) WSar ‘99"'wauo 'sfiUJ 'r ‘WIaupuns 'H '8 DUB JaMDUIM '6ZEE (LS6I) "OOS 'man ‘f ‘onmflg 'H '3 'W pue ‘Jofialgow"a 'M ‘sotnog “V 'M __. '(156i) 391 (E8 "A33 'SKI-{d “.If‘ (550 'V 'H pue (8.18503 'H .3 (IBQQUIAa’I ,3 '(Esoi) 995 ‘2? "wauo 'sfiud 'r "Jr ‘u081991nH 'v __ '(ES6I) S93 ‘53 "sflud 'pow 'SAaH ‘JOQSBd 'o '3 DUB 'Jr ‘UOSiuvinH 'v '(IS6I) 393 ‘T9"'Aaa 'sfiUd ‘JOised '0 '3 DUB 'Jr ‘u081901nH 'v °(1561) 91g ‘E"'oog 'sfiqd 'Jomv 'IIng ‘uonog 'V 'p pue insuefioosfim .JQZ. (Z 'v '0 ‘(156I) 913 ‘0? "830v 'sfiud 'Aiah ‘Xnou '0 DUB Jaiaaa ' _ _ 'JEW93I ‘05 'v '0 ‘(9561) 99 ‘6 (BAauao) IDS 'qoiv ‘xnoa '0 Due J9I398 M 3d 1 J 'f 'f hi cé :s 'S '0 '3 '3 'O '3 'M 'H '(IS6I) 999 ‘T9"Aaa 'sKQd ‘uefiu 'H 'v DUB SU31$I£>‘V'°W '(9561) If ‘36? "A28 'sflud ‘fiAai 'v 'a ___ ‘(T96I) 39II ‘39 "qsu1 'iog 'AZH ‘uonog 'V 'p pue ‘ounosg '3 'w ‘ ueIdey 'g '(396i) {6 ‘UaAOQDUIH ap anbOIIoo ‘saIMOd '0 'r Due 121390 '(396I) 0E1 ‘09 "Oos 'sfiud 'OOJd ‘saIMOd '9 'r Due 121390 68 '0 '0 '0 'WW 'fW ‘ZW 'TW 'OW '6E 'QE °L€ '99 'SE OWE ‘€€ 'ZE ‘TE '06 ‘68 '98 “L3 '93 ‘53 .178 ‘E8 '38 '(5561) 695 ‘96 "A28 'Sfiud ‘uosfih 'r '(0961) 6931 ‘:Dos 'wau0 'r '1WBEJM 'N 'v Due ‘821IM 'w '0 ‘u01UIEU 's '(3961) 81-6 'idas "r 'N ‘6110 oiquetqv ‘Aqoioos {e01mou3 ue0112mv ‘AJqsiwauQ {eoisflud JO uoisiniq euq oq paquason Jaded ‘oAfi '1 'p pue ptenaq '3 '563 'd ‘(6W6I) 810A mam "0uI Anedmog pueJqSJoN den '0 'II aanoans JeIn0oIoM pue BJqoodS JEIUOBION ‘BJequoH '(LS6T) eiuenrfisuuad JO KaisJaniun ‘uo1qeqsoss1q 'Q'ud ‘uioty [aeuoim '(E561) 169 ‘53 "sfiud 'pow 'Aau ‘uauo0 '3 pus puowna 'w ‘M __ '(0561) 191 91 "A83 'sfiud ‘31dd1H 'v 'r PUB ‘110081J0 '1 'H 'SBMOQI 'v C '(0561) 995 ‘103 v ‘(UOPUO1) '00s 'fiou 'DOJd‘sq1IJJIJo '9 'H -p pue Aatnfifiea o5 'w . '(1961) £51 ‘5? "sfiud 'wauo 'r ‘Jnapuowwox 'r PUB JaBUIs 's '(0961) 9051 ‘1? "sfiUd 'Iddv 'r ‘JOQSEd '0 '3 Due SUIHSOH 'H ' '(6561) £191 ‘0? "sfiud 'Iddv 'r ‘J35UES 's '(1561) 61W ‘9? "r '9021 '1sfis '1128 ‘Jauaa ' '(3961) ‘6-5 'AON ‘doqsxdom HdH-HWN IBnuuv UQXIS saqeioossv ueiseA 'quH U} sonbiuuool {equamisadxgu ‘apAH 'g ' 'qg-o 'd (1561) HJOA mom "0uI Anedmog wooa II1H-nesgow ‘nooqpueH SJoufiisaq 01u0110313 ‘qu0anIV 'd 'v pue ‘sineq '3 'Q ‘oepuej 'M '(1561) BUISUET 1593 ‘AliSJafliufl 91918 UEBIQQIW ‘uqueQJoss1q 'G'Qd 'JZQEJ 'f ' '(ES6T) 831 ‘35 "A83 'sflud ‘qoota 'g pue ssausfiuem 'H ' '(3961) '0uI u1wefuag 'V 'M ‘a0ueuosoa oiiaufiewEJBd ‘awed '3 C HJOA noN '(9W6T) 09W ‘0: “A83 'Sfiqd ‘qootg '(056T) 8E8 ‘8T"'Sfiud 'r 'wv ‘axed '3 '911-911 'd ‘(0561) 'ssew Snipeaa "OUI Anedmog Buiusitqnd flotsam uosippv ‘Soiueu0ow {BOISSBIQ ‘uioqsprog 06 'J 'J 'H '0 'H 'f 'H 'H '9 'J '0 'W9 '19 '65 '85 ‘15 '95 ‘55 'WS '85 '35 '15 '05 .617 'BW .LW '9? '57 '(1561) 199 ‘5_6 "A88 'Sflqd ‘qiasow '3 '3 pue ‘uossead '1 '9 'JQBBA 'V"M ‘Jauoqarg '3 '(5561) £89 ‘26 "A88 'sfiud ‘JaBUI13n1 'w 'r DUE“UQOH '(9861) 088 ‘05 "A88 'SAQd (X03 'W DUB UBWIIIW 'H 'M 'S '11120 ‘011v 0123 '931913033V’UBIJBA '3IQBI BMW '(3961) {15 C '(1561) 035 ‘69“'Hisfihd 'z ‘1w123 '(3E61) 555 ‘191 v "w290 'Hisfind '2 ‘SBHJBJ ' '(556T) ZEE ‘83 "A38 'Sfiud ‘dix 'x 'v pue Jaqad ' I6 55? "A83 'Sfiud ‘anequouitg '3 'Q pue ‘uemaoug 'H '3 ‘wassng ' V '3 I 9 '81 'IL '01 .69 '89 ‘19 '99 '59 cums TRY UERARY 1r1111)111101111911111