MSU LIBRARIES n \— RETURNING MATERIALS: P1ace in book drop to remove this checkout from your record. FINES w111 be charged if book is returned after the date stamped be1ow. SYNTHESIS AND PROPERTIES OF ALKALI METAL lB-CROWN-6 ALKALIDES AND ELECTRIDES BY Dheeb Issa A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1982 ABSTRACT SYNTHESIS AND PROPERTIES OF ALKALI METAL lB-CROWN-6 ALKALIDES AND ELECTRIDES Dheeb Issa Alkalides and electrides of alkali metals with 18- crown-6 as a complexing agent were synthesized by various procedures and the properties of these salts were studied by a number of different techniques. Two crystalline sodide salts, Cs+18C6-Na- and K+18C-Na-, were prepared. Analyses of these salts verified the stoichiometry, CslBC6Na with no solvent present and K18C6Na (with, how- ever, about 16% of amine solvent in the one sample analyzed). These are the first alkalide salts which utilize 18-crown- 6. Optical spectra of C518C6Na films showed one sharp peak at 16,500 cm-l attributed to the Na' anion. Films of K18C6Na showed a broad peak at 13,000 cm”1 and a shoul- der at 10,000 cm'l. Whether the peak at 13,000 cm”1 is caused by overlap of Na- and K- peaks or a red-shifted Na- Dheeb Issa peak is not known at this time. Powder d.c. conductivity measurements showed that these salts are semiconductors with average band gaps of 0.93 eV and 1.70 eV for K+18C6-Na- and Cs+18C6'Na-, respectively. Films and powders of electrides were prepared from methylamine solutions of cesium and 18-crown-6 in various proportions. The properties of the electrides were dependent on the ratio of the metal to crown used. Optical spectra of films with R=l showed a localized t "plasma type" absorption. EPR and magnetic suscepti- peak due to e while the films with R=2 showed a bility studies showed that the behavior of electrides with R=l can be explained by the presence of two trapping sites, while for electrides with R=2 the results indicated a metallic behavior. Films of electrides and alkalides were also prepared by direct reaction of the alkali metals and 18C6 deposited from the vapor phase. Optical spectra of these films showed the trapped electron and/or alkali metal anions can be obtained depending on the ratio of the metal to crown ether used. A crystalline electride, Cs+18C6°e- was prepared from equimolar solutions of cesium metal, lithium metal and lB-crown-G in a mixed isopropylamine-diethylether solvent. The presence of lithium greatly stabilizes the solution but the crystals are essentially free of lithium metal. The lithium probably acts as a scavanger Dheeb Issa for radicals produced by solvent and/or crown ether decomposition. Analyses of these crystals confirmed the stoichiometry C518C6. Optical spectra of annealed films showed a peak at 7,500 em“l attributed to trapped electrons. Powder d.c. conductivities of a polycrystalline sample showed a semiconductor behavior with an average band gap of 0.60 eV in the temperature range -54°C to +10°C. The magnetic susceptibility of crystals of C518C6 measured from 1.7 K to 300 K showed that the compound is essentially diamagnetic, with only about 1% unpaired spins. Single crystal EPR spectra showed the presence of two overlapping lines. One has g = 2.0023 and a linewidth of %O.75 G, independent of temperature from —135°C to +51°C. The other line has a smaller g-value (W2.0016) and shows g—anisotropy upon rotation. It has a linewidth which broadens from m1.25 G at -135°c to m3.5 G at +31°c. This represents the first synthesis of a crystalline electride. To My Family ii ACKNOWLEDGMENTS I am grateful to Professor James Dye for his encouragement, enthusiasm and guidance during this work. I would like to thank other members of the Group for their help. In particular, thanks go to Dr. Long Dinh Le, Michael Yemen, Brad VanEck and John Papaioannou. The MSU glassblowers, Keki Mistry, Andrew Seer and Manfred Langer, have given excellent service and cheerful cooperation. Special thanks go to the Yarmouk University in Irbid-Jordan for supporting financially this educational opportunity. Research support from NSF under Grant Number DMR-79-21979 is gratefully acknowledged. iii TABLE OF CONTENTS Chapter Page LIST OF TABLES. . . . . . . . . . . . . . . . . . . viii LIST OF FIGURES . . . . . . . . . . . . . . . . . . ix CHAPTER I - INTRODUCTION. . . . . . . . . . . . . . l I.A. Metal Ammonia Solutions . . . . . . . . . 1 I.E. Metal Ammonia Compounds . . . . . . . . . 5 I.C. Electrons in Low Temperature Glasses . . . . . . . . . . . . . . . . . 7 I.D. F-Centers . . . . . . . . . . . . . . . . 13 I.E. The Metal-Nonmetal Transition . . . . . . 15 I.F. Alkali Metal Anions . . . . . . . . . . . 16 I.G. Nature of Alkalides and Electrides. . . . . . . . . . . . . . . . 19 I.H. Alkali Metal Crown Ether Systems . . . . . . . . . . . . . . . . . 25 CHAPTER II - EXPERIMENTAL . . . . . . . . . . . . . 30 II.A. Materials. . . . . . . . . . . . . . . . 30 II.A.l. Complexing Agent . . . . . . . . . 30 II.A.2. Solvents . . . . . . . . . . . . . 30 II.A.3. Metals . . . . . . . . . . . . . . 31 II.B. Glassware Cleaning . . . . . . . . . . . 34 II.C. Sample Preparation and Instrumental Techniques. . . . . . . . . 34 II.C.1. General Preparative Methods. . . . . . . . . . . . . . 34 iv Chapter II.C.2. Optical Spectra. . . . . . II.C.3. EPR Spectra. . . . . . . . . II.C.4. Magnetic Susceptibility. . . II.C.S. Pressed Powder Conductivity. II.D. Analysis . . . . . . . . . . . . . II.D.l. Hydrogen Evolution . . . . . II.D.2. pH Titration . . . . . . . . II.D.3. Flame Emission . . . . . . . l II.D.4. H NMR 0 O O O O O O O O O 0 CHAPTER III - CESIUM lB-CROWN-G ELECTRIDES. III.A. Optical Spectroscopy. . . . . . . III.A.1. Cs-18C6 Films from Ammonia. III.A.2. Films From Methylamine. . . III.B. EPR O O O O O O O O O O O O O O III.B.l. Results and Discussion. . . III.C. III.B.1.a. Solids From Cs—18C6 Ammonia Solution. . . III.B.1.b. Solids from Cs-18C6 Methylamine Solutions Magnetic Susceptibility . . . . . III.C.l. Results and Discussion. . . III.D. III.C.l.a. Cs(18C6)2 Solid From Methylamine Solution. III.C.1.b. (CS)-(18C6)1. . . . . III.C.l.c. (Cs)2(18C6)1. . . . . Microwave Conductivity - Results and Discussion. . . . . . . . . . Page 40 41 42 44 45 45 46 47 47 49 50 50 50 57 59 59 59 64 76 79 79 81 84 84 Chapter CHAPTER IV - ALKALI METAL 18-CROWN-6 ALKALIDES. . . . . . . . IV.A. Optical Spectroscopy . . . IV.A.l. Films of CsNa18C6. IV.A.2. Films of KNa18C6 . . IV.A.3. Films of Cst18C6. . IV.B. Sample Analyses. . . . . . IV.C. Powder dc Conductivity . IV.D. Magnetic Susceptibility of Cs+18C6-Na-. . . . . . . . IV.B. X-ray Study of Cs+18C6-Na- IV.B.l. Single Crystal Isolation IV.B.Z. Results of X-ray Study . CHAPTER V - PREPARATION OF ALKALIDES AND ELECTRIDES BY DIRECT VAPOR DEPOSITION. . . . . . . . V.A. Experimental. . . . . . . . V.B. Results and Discussion. . . CHAPTER VI - PREPARATION AND CHARACTERIZATION OF A CRYSTALLINE ELECTRIDE Cs+18cs-e' . . . . . . . Introduction. . . . . . . . . . . VI.A. Preparation Procedure. . VI.B. Analyses . . . . . . . . . VI.C. Optical Spectra. . . . . . VI.D. Powder dc Conductivity . . VI.E. EPR Study. . . . . . . . vi Page 89 90 90 90 93 95 97 97 100 100 101 104 104 106 110 110 112 112 114 117 120 Chapter Page VI.F. Magnetic Susceptibility Study. . . . . . 123 CHAPTER VII - CONCLUSIONS AND SUGGESTIONS FOR FUTURE WORK . . . . . . . . . . . . . 126 VII.A. Conclusions . . . . . . . . . . . . . . 126 VII.B. Suggestion for Future Work. . . . . . . 127 BIBLIOGRAPHY. . . . . . . . . . . . . . . . . . . . 129 vii Table LIST OF TABLES Page Values for the Parameters in the Fit of the Static Susceptibility for Three Different Samples with Two Non-Interacting Centers . . . . . . . . . . 83 Microwave Conductivity Measurements of Cs+18C6-e- Compared with the Standards . . . . . . . . . . . . . . . . . 87 Microwave Conductivity Results of Cs+l8C6'e- Electrides Compared with Standards . . . . . . . . . . . . . . . . . 88 Results of the Analyses of Cs+18C6°Na-. The Values in the Parenthesis are the Percent Deviation from the Predicted Stoichiometry . . . . . . . . . . . . . . . 96 Results of the Analyses of K+18C6'Na-. The Numbers in the Parentheses are the Percent Deviation from Predicted Stoichiometry . . . . . . . . . . . . . . . 96 Results of the Analyses of Crystal- line Cs+18C6'e-. The Numbers in Parentheses are the Percent Deviation From Predicted Stoichiometry. . . . . . . . 113 viii Figure LIST OF FIGURES Reflectance spectra of Na-NH3 solutions . . . . . . . . . . . . . Optical Spectra of alkali metal-EDA solutions . . . . . . . . . . . . . Apparatus for distribution of alkali metal under vacuum. . . . . . . . . . Apparatus for preparation of powders and films of Cs+18C6'e- . . . . . . . Apparatus for the preparation of crystalline alkalides . . . . . . . Optical spectrum of a dry film of C518C6 (R = 1) from ammonia solu- tion. 0 O O 0 O O O O O O O O O O O 0 Optical spectra of dry films of C518C6 from methylamine solutions ( R = 1), (---- R = 0.5). . . . . . . . Optical spectra of Csl8C6 (R = 1.5) from methylamine solution (---- film made at -42°C) ( film made at -55°C) . . . . . . . . . . . . . . ix Page 18 33 35 38 51 52 54 Figure 10 11 12 13 14 Page Optical Spectra of films of C518C6 R = 2 from methylamine solution ( temperature of the film -52°C), (---- temperature of the film -45°C) . . . . . . . . . . . . . 56 A/B ratios of Cs+18C6°e- with R = 1 prepared from ammonia solution. . . . 61 Linewidths of Cs+18C6°e- (R = 1) from ammonia solution. Solid symbols data collected with k-He cryostat; open symbols i-Nz cryostat. . . . . . 62 Number of spins vs. temperature for Cs+18C6'e- (R = 1) from ammonia solution. . . . . . . . . . . . . . . 63 Linewidth of Cs+18C6-e- with R = 0.5 from methylamine solution. Solid symbols - data collected with R-He cryostat; Open symbols - R-Nz cryostat . . . . . . . . . . . . 65 A/B ratios of Cs+18C6-e- R = 0.5 from methylamine solution. Solid symbols - data collected with the k-He cryostat; open symbols - i-Nz cryostat . . . . . . . . . . . . 66 Figure 15 16 17 18 19 20 Page Number of spins against reciprocal temperature for Cs+18C6'e- with (R = 0.5) from MeNH solution . . . . . 68 2 Linewidth of Cs+18C6-e- with R = 1 from methylamine solution. Solid symbols - data collected with Q-He cryostat, Open symbols - t—NZ cryostat, solid squares - data col- lected on another sample from dif- ferent preparation. . . . . . . . . . 69 EPR spectra of Cs+18C6'e- (R = 1) from methylamine at two different temperatures. . . . . . . . . . . . . 70 A/B ratios for Cs+18cs-e' (R = 1) from methylamine. Solid symbols - data collected with z-He cryostat; open symbols - l-Nz cryostat; squares - data collected on another sample from different preparation . . 71 Number of spins vs reciprocal tem- perature for Cs+18C6‘e- (R = 1) from methylamine solution . . . . . . 72 EPR spectra of C518C6'e- with R = 2 from methylamine solution . . . . . . 75 xi Figure 21 22 23 24 25 26 27 Page Plot of reciprocal molar suscep— tibility vs. temperature of Cs+18C6-e- (R = 0.5) from methylamine solutions. . 80 Plot of reciprocal molar sus- ceptibility vs. temperature for Cs+18C6°e- (R = 1) from methyl- amine solutions - A data taken as temperature increases; 0 — data taken as temperature decreases. . . . . 82 Plot of molar susceptibility vs. temperature for Cs+18C6'e- (R = 2) from methylamine solution. Circles resemble data taken as temperature increases; squares resemble data taken as temperature decreases. . . . . 85 Optical spectrum of dry film of Cs+18C6'Na- from methylamine solu- tion. . . . . . . . . . . . . . . . . . 91 Optical spectrum of a dry film of KNa18C6 from methylamine solution . . . 92 Optical spectrum of a dry film of Cst18C6 from methylamine solu- tion. . . . . . . . . . . . . . . . . . 94 Plot of log resistivity vs. recipro- cal temperature for polycrystalline Cs+18C6'Na-.. . . . . . . . . . . . . . 98 xii Figure 28 29 30 31 32 33 34 35 Page Plot of log resistance vs. recipro- cal temperature for polycrystalline K+18C6-Na-. . . . . . . . . . . . . . . 99 Apparatus for preparation of films of electrides and alkalides from vapor deposition. . . . . . . . . . . . 105 Optical transmission spectra of 1) C518C6; 2) --- Rb18C6; 3) --- K18C6; 4)--- Na18C6 by direct vapor deposition. . . . . . . . . . . . 107 Optical transmission spectra of Cs+18ce-e' film made at -so°c . . . . . 115 Optical transmission spectra of Cs+18C6°e— film at temperature -37°C. 1) --— unannealed film at -50°C; 2) annealed film at +12°C . . . . . . . . . . . . . . . . . 116 Ohm's Law plot of polycrystalline powder of Cs+18C6'e-. . . . . . . . . . 118 Plot of log resistance vs. reciprocal temperature for a polycrystalline Cs+18C6'e-. . . . . . . . . . . . . . . 119 EPR spectra of a single crystal of Cs+18C6'e- at different (arbitrary) orientations. . . . . . . . . . . . . . 121 xiii Figure 36 37 Page Effect of temperature on the EPR spectra of single crystal (Cs+18C6°e-). . . . . . . . . . . . . . 122 Plot of molar susceptibility against temperature for poly- crystalline Cs+18C6°e—. . . . . . . . . 124 xiv CHAPTER I INTRODUCTION Solutions of the alkali and alkaline earth metals in ammonia, some other amines and ethers have been studied for well over a century, and an extensive literature exists in this field (1-6). The work described here is based on that research. In the early 1970's Dye and coworkers used crown ethers and cryptands as complexing agents to enhance the solubility of alkali metals in amines, which opened the door to a new field in alkali metal solu- tion chemistry. In this section the background for the research will be presented, beginning with metal ammonia solutions (MAS). This includes the nature of the species present in metal solutions, the solution properties, the metal-nonmetal transition and solid metal ammonia and methyl- amine compounds. The role of the complexing agents in en- hancing metal solubilities is also described. I. A. Metal Ammonia Solutions Alkali metals dissolve in ammonia to a considerable extent giving blue solutions when dilute and metallic bronze solutions when concentrated. Sodium, potassium and rubidium form saturated solutions at about 16 mole percent metal (MPM). Lithium is soluble up to 20 MPM and cesium to about 65 MPM (7). The solution properties range from electrolytic in the dilute and intermediate concen- tration ranges (<0.5 MPM) to metallic in concentrated solu- tions (>8 MPM) with a nonmetal to metal transition occur- ing between 2 and 9 mole percent metal. Very dilute metal ammonia solutions (<10-3 M), contain solvated cations and solvated electrons. Although the structure of the latter is open to question, the most commonly accepted structure for the solvated electron (based upon theoretical calcula— tions) is that of an electron trapped in a cavity from which one or more ammonia molecules have been excluded (8-11). This model was proposed first by Ogg and developed semi—quantitatively by Jortner. The cavity radius is 3.2 - 3.4 A, and the electron is stabilized by orientation of nearest neighbor dipoles and by long-range polarization. As the concentration increases cation-electron and electron- electron interactions become important. However, the changes in such properties of these solutions as optical spectra, magnetic susceptibility, conductivity, density and activity coefficient are nearly independent of the cation. This suggests that if any new Species are formed, they retain the essential characteristics of the solvated cation and the solvated electron. An ionic aggregation model has been proposed which explains these properties (12,13). Based on this model, ion-pairing occurs between M+ and e which forms a non-conducting ion-pair. The mag- Solv netic interaction between the electron spin and the metal nucleus in this ion-pair is weak as Shown by the small paramagnetic Shifts of the alkali metal cations in the nmr Spectra (14,15). Static susceptibilities (16,17) and EPR spin susceptibilities (18-20) Show that in addition . . . + - to ion-pairing between M and e solv’ Spin-pairing occurs to give diamagnetic states which become more predominant as the concentration increases. This has been attributed to the ion-triple, e_°M+-e-, although higher aggregates of stoichiometry M2 and M2 might also form. AS the metal concentration is increased (3 < MPM < 9), the prOperties change from non-metallic to metallic. Many theories have been developed to explain the nature of this transition (1). The Simple qualitative picture for this transition is that, although a second electron may be added to a 2 trations eventually there is not enough ammonia to provide cavity to form Spin paired Species e , at higher concen- Such Sites. Therefore, the electrons are forced to enter the conduction band at high concentrations resulting in the non-metal to metal transition. Concentrated metal am- monia solutions exhibit metallic characteristics. The absorption spectra Show plasma edges, predicted by the Drude model for conduction electrons in a metal (21). Figure 1 shows the reflectance spectra of Na-NH3 solution .mcoflusHOm mmZIwz mo mupoomm mocmuowammm .H musmfim 31 use: 55 outages 35821.9. 23.330. IOO O O m m 4 2 A A d % q q u _ . II . 12 m \A 13 5 5 2 8. l m. 6. 8. M W W. 4 oz 29: one 2 mIz 8.2: So 28.. 4 8 6. 7 5 5 n m 5 3. 2 .29: 28.3 206 . h . h b p . _ b in the intermediate and concentrated regions (22). As the concentration of the metal increases above 5.6 MPM, the reflectance Spectra Show a very Sharp drop at the plasma frequency which is a collective resonance of the conduction electrons. In contrast, the absorption spectrum of the solvated electron consists of a distinct peak in the in- frared which drOpS to low absorbance as the wavelength is increased. The conductivity of concentrated metal am- monia solutions is high (comparable to that of liquid mercury) (23). The EPR Spectra also Show that concentrated metal ammonia solutions are metallic (24) aS indicated by the asymmetric first derivative absorption curve predicted by Dyson for highly conducting materials (25). I. B. Metal Ammonia Compounds In contrast to the well-investigated metal ammonia solutions, relatively few studies have been made on metal ammonia compounds. Lithium is known to form the compound Li(NH calcium, barium and strontium form M(NH3)6 as 3)4. do eurOpium and ytterbium (l). The work that has been done to date suggests that these "expanded metals" have many novel and interesting structures as a result of their low conduction electron density. Recent re-examination of the crystal structure of Li(NH3)4 by Sienko and co- workers (26) Shows the presence of three solid phases at different temperatures. All are indexed as body-centered cubic, but with different space groups. Solid phase I is stable in the range 89-82 K, solid phase II in the range 82-25 K and the third phase below 25 K. Glaunsinger and Sienko (27) studied the EPR Spectra of Li(NH3)4 in phase I. The experimental results were fitted to the theoretical lineshape equation derived by Dyson (25). In 1955 Dyson formulated a theory to account for the diffusion of conduc- tion electrons into and out of the region penetrated by the EPR microwave-frequency field. For a metallic sample which is thick compared to the skin depth (6) and with a relaxa— tion time long compared to the diffusion time, the first derivative lineshape is asymmetric about its center with the low-field amplitude A greater than the high field amplitude B. This is now known as the Dysonian lineshape A and results in 2.7 < E < 20. The EPR spectra of Li(CH3NH2)4 system have also been studied (28-30). The spectra indicate A an asymmetrical lineshape with 1 < E < 2.7. The magnetic susceptibility of Li(NH3)4 and Li(CH3NH have been measured 2)4 over the temperature range 4.2 - 300 K. The results for Li(NH3)4 (31) showed that the susceptibility is small and positive and that it is temperature independent above 89 K. In solid phase I (between 89 and 82 K) a smaller temperature independent paramagnetism was observed, while in solid phase II (from 82 to about 25 K), Curie-Weiss paramagnetism was found which changed to antiferromagnetism below 25 K. However, the magnetic behavior of Li(CH3NH2)4 is quite dif— ferent (25). The susceptibility was found to decrease with decreasing temperature showing a large drop at 155 K which is believed to be the melting point Of the compound Li(CH3NH2)4. It then became constant and diamagnetic in the solid phase. The hexammine compounds M(ND3)6 where (M is Ca, Ba, Sr) Show interesting structural features which are worth mentioning. Powder neutron diffraction and x-ray diffraction studies (32—34) Show that within the complexes the six nitrogens are arranged in an exact octahedron with a rather long M-N distance. The neutron diffraction data indicate that the ND3 molecules are nearly planar and have two inequivalent sets Of deuterons. One N-D bond is short (No.9 A), while the other two are extremely long (m1.4 A). A second feature Of the M-ND3 structure is its built-in disorder. Although each ammonia molecule is coordinated to the metal by its nitrogen, its "pseudo—trigonal" axis is not coincident with the M'-N bond but makes an angle Of N13° with it. Based on the properties of these compounds such as electrical transport, magnetic susceptibility and magnetic resonance, these compounds constitute model ex- amples Of metals with low electron densities. I. C. Electrons in Low Temperature Glasses In 1962 the transient Optical Spectrum Of the solvated electron in water was discovered by pulse radiolysis methods (35). Since that time transient solvated elec- trons have been Observed in a variety Of other polar and nonpolar liquids (36). One offshoot of this work was the study of trapped electrons in glassy disordered media at liquid nitrogen temperatures, and recently in the liquid helium range on the picosecond and nanosecond time scales. Electrons are usually generated by ionizing radiation, and the formation Of solvated electrons can be visualized kin- etically as occurring in two stages (37,38). The first stage is localization which depends upon the relative ener- gies of the conduction electron level Of the medium and the energy Of the localized electron state in the medium. The second stage, after localization occurs is solvation, in which the electron induces rearrangements in the surround- ing solvent—Shell molecules to create a geometry charac- terized by the forces between the electron and the solvent molecules. Electrons trapped in organic and alkaline ice glasses all give broad absorption spectra in the visible or near-infrared with long tails on the high energy side Of xmax and shorter tails on the low energy Side (39). These spectra have linewidths at half-heights ranging from m 0.5 eV for hydrocarbons to >1 eV for some alcohols, and trap depths which increase with increasing polarity Of the matrix molecules. The initial spectrum Of the trapped electron at :200 nsec in CZHSOD glass at 77 K has Amax ml300 nm and shifts with time to a stable position near 540 nm. If 10-20 mOl % Of water is added tO the ethanol, the absorption band remains structureless but its maximum shifts slightly toward the red (40). These time-dependent spectral changes reflect the orientation Of surrounding molecular or bond dipoles around the electron in the pro- cess Of solvation. In 1975 Willard and co-workers (41) conducted a series Of bleaching experiments (on the spectra Of produced in 2-methy1tetrahydrofuran (MTHF) glass) etrap' in which the Optical density in the near infrared is reduced at the bleaching wavelength and at longer wavelengths, but the remainder Of the spectrum is unaffected. They found that the entire spectrum bleaches nearly uniformly when 1064-nm, rather than l338-nm light is used. This indicates that all electrons apparently have high energy tails. Trap- ed electrons contribute differently to the Optical density according to their Amax' Consistent with the trapping model suggested by the Optical Spectra, the linewidths Of the EPR singlet of the trapped electron in glasses at 77 K does not change indicating a symmetrical equilibrium environ- ment, but as the polarity Of the matrix molecules increases the linewidths increases indicating increasing hyperfine interaction with atoms along the walls Of the trap as the traps become deeper (42). The structure Of the solvated electron in different glassy media has been reported (43). Most Of the information about the structure came from Op- tical spectra and spin-echo studies. In aqueous systems, Spin-echo studies Of 10 M sodium hydroxide aqueous glasses Showed that the electron has only water molecules in its 10 first solvation shell. The electrons are solvated by Six water molecules, each with one Of its OH bonds arranged ap- proximately octahedrally around what is presumably the center Of the Odd electron density. The distance to the closest proton Of the solvent water molecules is 0.21 nm while that to the more distant proton is 0.36 nm. These two distances, with an apprOpriate OH bond orientation are consistent with the known structure Of the water molecule. The structure indicated that there are relatively strong interactions between the solvated electron and the first solvation shell. Another system that has been studied in great structural detail is electron solvation in MTHF (43). A quantitative model Of the organization Of molecules around trapped electrons at 77 K based on electron spin-echo studies indicates that three MTHF molecules are oriented around each electron with the plane Of each molecule per- pendicular to a line from the center Of the ring to the electron at a distance Of m3.7 A, and that the CH3 group is on the side away from the center Of the electron Spin density. The molecules Of MTHF are oriented statistically causing multiple environments for the electron. In 1975 Willard (41) prOposed that, because the Spectrum Of sol- vated electrons in MTHF showed three different peaks, the structures ascribed to different discrete preferred orien- tations Of MTHF molecules would give different trap depths. The structure Of the solvated electron in glassy ethanol 11 was also reported, which shows that the first solvation shell consists Of four molecules with the CD3 deutrons at 0.38 nm, the CD2 deutrons at 0.33 nm and the OD deutron at 0.22 nm from the electron. In methanol glasses the solvation shell is 4 i 2 molecules with an electron to hydroxyl proton distance Of 0.23 i 0.02 nm and the OH bond points towards the electron. In mixed matrix glasses, there is a single absorption peak due tO the electron and Amax Shifts ac- cording tO the composition Of the matrix provided the glass is made from a mixture Of similar types Of molecules. By contrast when a glassy matrix is made from different mole- cules, two distinct absorption peaks characteristic Of the two components are found. The electron is initially trapped in the less polar solvent and then moves to more polar regions as it anneals. This could also be due tO solvent orientation effects. Electrons generated in single carbo- hydrate crystals show similar properties tO those in glassy media (44,45). The electron g-value is nearly the same as that Of the free electron. In both media the EPR absorption can be bleached by using visible light and electrons in single crystals show strong anisotropic hyperfine inter- actions with the hydroxyl groups. However, the traps in single crystals are shallower. It is Of interest tO note that in single crystals only a Single trapping site is utilized even though many Of the crystal structures Offer more than one site. There are several theoretical models 12 tO explain the structure of solvated and trapped electrons. They have been reviewed by Feng and Kevan (46) and Only two Of them will be mentioned here, since the others have failed tO explain satisfactorily the properties Of solvated electrons. Even the two models described here are con- troversial. The continuum model was first formulated mathematically by Jortner (47). In this model the medium is assumed to be a polarizable continuous dielectric, and the electron is assumed to be located physically in a spherical symmetric cavity. At distances outside the cavity, the electron interacts with the continuous di- electric through a % Coulombic potential, while within and at the cavity boundary R, the potential remains constant. The semi-continuum (48,49) model considers the electron to be at the center Of the cavity surrounded by a fixed number of solvent molecules which are arranged symmetrically. These solvent molecules interact with the electron tO give Short-range interactions which account for the values Of electronic energy levels. Outside Of the solvation shell is the dielectric continuum. However, there are still several arguments about these models. In 1980 T. Ichikaw and H. Yoshida (50) prOposed a cavity model for localized electrons in nonpolar media. In this model, the depth Of the potential well was derived from the Hartree—Fock molecu- lar potential and the polarization potential. It was found deep enough to account for electron localization. They 13 found that the energy of quasifree electrons, the total energy for electron localization and the cavity radius (estimated from the simulation Of the absorption spectra) were in reasonable agreement with the experimental data available for the 3-methylpentane glassy matrix used in the study as a typical example. Another theoretical formalism was developed by Banerjee and Simon (51). In their formula, the Hamiltonian was written as the sum Of electronic and vibrational contribution Of the system which includes a trapped electron plus solvent. The absorption band shape results from two major contributions, the localized (bound- bound) transition which makes the largest contribution and the nonlocalized (bound-free transitions). The latter, with the effect of rotational-translational contributions ac- counts for the asymmetry Of the absorption band. The model was found tO explain successfully the absorption spectra Of excess electrons in ethanol and anthracene glasses, sug- gesting that this model might be a general model for excess electrons in condensed media. I.D. F-Centers F—centers are defined as anion vacancies in an ionic lattice occupied by electrons. Their characteristic color (from which they got the name "color centers") is due tO an interaction, or series Of interactions, between excess electrons and negative vacancies. F—centers have 14 been studied most completely in alkali halide crystals. The Optical spectrum of an F—center consists of a broad absorption band, described as a bell-shaped curve. The wavelength at which the absorption peak occurs varies from 250 nm for LiF which has the shotest interionic distance to 785 nm for CSI with the longest interionic distance (52). MOllow first reported a relation between um (the frequency Of the absorption maximum) and the inverse square Of the 2 for the NaCl interionic distance a. Vm (in eV) = 20.7 a- structure. Later this equation was modified to give a better fit. The new relation known as the MOllow-Ivey a-l'84. The temperature ef- relation is um (in eV) = 17.6 fect was studied by MOllow and others. It was found that the absorption (A(v)) drops Off faster towards the red-end Of the peak than towards the violet, and that the shape is more symmetric at higher temperatures. Pressure also af- fects the peak; as the pressure increases, the position Of the peak shifts toward higher energy and the peak gets broader. Studies Of the EPR spectra Of F-centers Show that the g-values vary Slightly depending on the system, but that all the g-values are close tO 2.0 (53). The line- widths are very broad because Of hyperfine interactions Of the electron with its nearest neighbors. In 1975 Schin- dewolf (54) studied the Optical absorption spectra Of ex- cess electrons produced by electrolysis in melts of alkali halide salts in the temperature range 600-900°C. As expected, 15 the position Of the peak varied with temperature. The rela- tive half-width (half-width/peak position) Of the Spectra is almost 1, and a quadratic relation similar to that given by MOllow was Obtained but with a wider spread Of data and a steeper slope than for solids. This was explained by assuming that the cavity radius of the electrons in melts is larger than in the corresponding solids. I.E. The Metal-Nonmetal Transition A metal-nonmetal transition occurs in many crystalline and disordered systems. There are several theories which explain the transition and which will be reviewed here in a qualitative manner. Mott (55) first showed that a metal-nonmetal transition can occur if the lattice spacing is varied. When the electrostatic repulsion between elec- trons equals the band width a transition tO the conduction state occurs. Using a screening constant, Mott calculated the criterion on the atomic density for the metal—nonmetal transition as nl/3 a = 0.25 where n is the electron density and aH is the Bohr radius Of the isolated species. This relation was found tO be empirically applicable to several systems (56) when an "effective" Bohr radius is used together with a critical 16 electron density nc: nl/3 a = 0.26 In his approach, Hubbard (57) viewed the problem in terms Of band theory. In isolated atoms the bound energy levels Of the electrons are discrete. When the energy levels overlap they smear out into bands with a bandwidth w. If u is the electrostatic repulsion energy at a particular site, then when g i 1.15 a transition from nonmetal to metal occurs. Later Mott found the condition for the transition is E m 1, hence the theory has been advanced as the Mott-Hubbard model. Anderson's approach (58) was based on the assumption that the lattice is composed Of potential energy wells with variable depths with VO being the Spread in energy. Using w, the band width, as in the Hubbard model, a metal-nonmetal transition occurs at w VSW 0.5 for Z = 6, where Z is the number of nearest neigh- bors. A combination Of the Mott-Hubbard and Anderson's models has been used to explain the metal-nonmetal transi- tion in systems such as Li(CHBNHZ)4 solutions (30). I.F. Alkali Metal Anions It has been known for some time that the alkali metals dissolve in a number Of amines and ethers, giving blue solutions with one or two broad absorption bands. One 17 Of these, in the infrared region at 1200-2000 nm corresponds tO the metal-independent band Of the solvated electron. The position Of the other band depends upon the metal, solvent and temperature and the absorption peak is strongly asym- metric On the high-energy Side (59). These features are characteristic Of the SO called charge-transfer to solvent (CTTS) bands Of other anions such as the halides. Figure 2 shows the spectra Of Na-, K-, Rb , Cs- and e- 501v in ethylene- diamine (60). The infrared shoulders Observed for solutions Of potassium, rubidium and cesium are attributed to the sol- vated electron and the ratio Of this absorbance tO that of the anion depends strongly on concentration as indicated by the following equation. M- Z M+ + 2e;01V Matalon, Golden and Ottolenghi showed that the wavenumber Of maximum absorbance Of the alkali metal anions varies inversely with the estimated anion radius, in accord with the predictions Of CTTS theory (60). The Observed tem- perature dependence Of the peak position is also predicted by this theory. The oscillator strength (61), Faraday effect (62) and formation Of Na- in pulse-radiolysis studies (63)a11 supported the assignment Of the metal-dependent band to M-. The strongest evidence for "genuine" alkali metal anions in solution came from the 23Na NMR spectrum 18 33 x .mGOflusHOm <0muawuoa HHmMHm mo muuoomm HMOfiuQO .m musmfim TO. A 758m ON 9 O. m q _ r and XOE< ' . x . . “OZ _ _ am. m0 _ _ 10. 0.0 m0 . NO m6 0.. N. m._ m._ 19 Of Na- in three solvents (64,65). The absence Of a solvent paramagnetic Shift (compared with the gaseous anion) shows that the solvent cannot interact strongly with the p electrons of the alkali metal as it does in the case of solvated cations. In addition, the narrowness Of the lines suggests that M- is large and spherically symmetric. Later Observations Of the NMR spectra Of Rb- and Cs- were also in accord with this picture (65). Finally, in 1974 the proof was completed when Dye and co-workers (66,67) iso- lated and characterized a crystalline salt Of the sodium anion which was free of solvent. All of these Observations provide proof of the existence Of alkali metal anions both in solution and in ionic solids. I.G. Nature Of Alkalides and Electrides After the synthesis Of crown ethers by Pedersen (68) in 1967 and cryptands by Lehn (69) in 1969, Dye and co- workers used these macrocyclic compounds as cation complex- ing agents. They found that 18-crown-6 greatly enhanced the solubility Of metals in amines and ethers, and that the cryptands were even better complexing agents. This work also expanded the range Of solvents in which the alkali metals could be dissolved (70-72). The equilibria in metal solution can be described as 20 2M(s) Z M + M’ (1) M Z M + e (2) M Z M + e (3) solv Depending on the metal and solvent used, the Species esolv’ M- and M can be detected. When a complexing agent is added tO the solution, the complexation equilibrium (4) occurs M+C+MC (4) This will shift the equilibria in 1 and 3 to the right. Depending on the solvent used, either e- of M— or both can be produced. Another characteristic Of the cryptand complex- ing agents is that M+C consists Of the bare cation trapped in the cryptand with no solvent included. Other important properties Of these complexing agents are their high stability, high selectivity towards various cations and ability tO shield the cation from the solvent environment. By controlling the ratio Of the complexing agent to the metal it was possible to differentiate among various models for the solution structure Of M- in amines and ethers. This was done by Dye, Ceraso and Andrews by using alkali 23Na NMR studies for solutions Of metal NMR (64,73). sodium bromide and C222 cryptand at several mole ratios in various solvents showed separate peaks for Na+ and Na+C222 21 which merged into a single peak as the temperature was increased. Therefore, because Of the slow release Of Na+ from the complex it was possible to Obtain separate signals due to Na+C and Na- for solutions which contained C222 and sodium in a 1:2 mole ratio. The position and linewidth of the signal from Na+C coincided with those Ob- tained for solutions Of simple sodium salts in the pres- ence Of the cryptand. The Signal from Na- is extremely narrow and is shifted diamagnetically from that Of the cation in the same solvent. The absolute chemical shift Of the Na- resonance is practically the same as that Of gaseous Na- and is independent Of the solvent. These results showed that the sodium p—electrons in Na- are well-shielded from interactions with the solvent molecules. The narrowness Of the line indicated that Na- is highly symmetric, and the solvent-independence strengthened the overall conclusion that Na- is a spherically symmetric anion with two electrons in the 3s-orbitals. It was also pos- sible to Obtain the NMR spectra Of Rb- and Cs-. Solvent- free solids have been Obtained from concentrated alkali metal amine and ether solutions which contain complexing agents. This has been done in two ways: first by Slow cooling Of relatively concentrated solutions to form crystals Of M+C-N-, ("alkalides"), and second, by evaporating the solvent to dryness leaving a precipitate with the composi- tion M+C-e- ("electrides"). Early attempts to grow crystals 22 from solutions of sodium and C222 resulted in the crystal- lization Of the compound Na+C222-Na- (66,67,74). This com- pound has been fully characterized. The crystals are stable for long periods Of time in vacuum or in an inert atmosphere at about 0°C (at least several years). The crystals are, however, sensitive to air, water, high tem- peratures and light. The stability Of the crystals made it possible to determine the crystal structure. The x-ray structure showed that the twO sodium moieties are in dif- ferent environments in the crystal. One is inside Of the cryptand and the other is outside Of the cryptand. The sodium anions form parallel planes separated by the cryptated sodium cations. The cryptand has three-fold symmetry and an antiprismatic oxygen atom arrangement about Na+. The Na--N closest distance is 8.83 A, the Na-N closest distance is 5.55 A and the Na--O distance is 5.76 A. Although the compound appears metallic, conductivity measurements showed that it is a semiconductor. After the synthesis Of Na+C222-Na- a number Of different kinds Of crystals which used different cryptands and metals were prepared. How- ever, the instability and/or poor quality Of these crystals made them unsuitable for x-ray structural studies. At the beginning Of the present research none Of these crystals had been completely analyzed or characterized. The second type Of compound, called "electrides", were studied by Optical spectroscopy, electron paramagnetic 23 resonance, microwave conductivity and magnetic suscep- tibility, and the properties Of these electrides will be reviewed here (75—78). Dry films made from ammonia solu- tions Of potassium and C222 with a metal to cryptand ratio R = 1 yielded time-dependent spectra with two peaks; one at 11,000 cm-1 1 (900 nm) attributed to K-, and the other at 6500 cm- (1800 nm) corresponding tO the trapped electron. With time, the K- peak decayed tO a Shoulder and the absorp- tion Of e; increased. The final spectrum showed a "plasma— type" absorption spectrum similar to that Of conduction electrons in concentrated metal ammonia solutions. The EPR spectra Of solid K-C222 with R = 1 showed that the ratio Of the low-field derivative peak to the high-field peak 5 is larger than 1. This result is one characteristic B' Of a metallic system as shown by Dyson. The spectra showed that multiple lines appeared below 77 K and that the system may be anisotropic. X-band microwave conductivity studies Of a cryptand-rich K+C-e- sample indicated that the solid is highly conducting in the microwave region. Another type Of electride which was investigated more thoroughly is that formed with lithium and cryptand C211 (79). The properties Of LiC21l electride depend on the mole ratio Of metal to cryptand, R. Films from solutions with R < 1.15 have an absorption spectrum with low far-infrared absorbance and peaks at 5000 cm.1 and 7000 cm-1, and a high energy shoulder at m12000 cm-1 suggesting that the electrons are 24 trapped in several different non-equivalent environments. On the other hand, dry films with R = 2 showed a "plasma" edge presumably due to conduction electrons. A metal non— metal transition apparently occurs between R = 1.5 and R==2. The EPR spectra of Li+C211-e- showed that the g-values are at or near the free electron value, 2.0023, the linewidth AHp_p is only about 0.5-0.6 Gauss, but below 60 K the linewidth increases gradually to 1.5 Gauss at 3 K. The % ratios are less than 1.25 even for samples which Showed metallic character. Except for a sample with R = 1.5, all samples showed at least one additional EPR peak at low temperatures. The EPR spectra showed that there is a tendency for the spins to pair as the temperature is de- creased. There was apparently only a single spin-pairing process in the sample for which R = 1.57. This sample also showed only one peak in the Optical spectrum. Systems which showed more than one peak in the Optical spectra ex- hibited two spin-pairing processes. Static magnetic sus— ceptibility measurements (80) verified the EPR results, which showed that the number Of unpaired spins tends to decrease as the temperature is decreased. The temperature Of maxi- mum susceptibility ranged from 70 K for the system with R = 0.6 to 20 K for the system with R = 1.57. The sus- ceptibility approached zero at low temperatures and the number Of unpaired spin (based upon a Curie-Law approach) varied exponentially with % down tO 3 K. Except for the 25 most homogeneous sample (R = 1.57), it was difficult to describe the behavior below the temperature Of maximum susceptibility. Above this temperature the susceptibility was described by the Spin-pairing equation 2 2 f Navg g pB Xm = kBT[3 + exp(-J/kBT)] (5) in which f is the fraction Of electrons which participate an J is the Heisenberg exchange integral. At high tempera- tures the above formula approaches the Curie-Weiss Law. The fit to a spin-pairing equation does not necessarily imply only pair-wise interaction Of the electrons. How- ever, the presence Of several trapping sites and the lack of information about the structure made it impossible to interpret the data quantitatively. I.H. Alkali Metal Crown Ether Systems The macrocyclic polyethers first reported by Pedersen (68), the so-called crown ethers, are known to be effective in increasing the solubility Of inorganic salts in non- polar media (81-87). The common factor governing all these processes is the ability Of crown ethers to form stable complexes with cations. These complexes are mostly Of 1:1 stoichiometry where the cations are held in the crown 26 ether ring by ion-dipole forces. The stability Of crown- ether complexes depends on several factors; these include cavity size Of the ligand, cation diameter, spatial dis- tribution of ring binding sites and the type Of solvent used. Very stable complexes can be formed when there is a good match between cation and crown ether with respect to size. However, the results Of Izatt et 31. (88) con- firmed that correspondence Of ligand ring cavity "size" to cation size is an important factor in determining the stability of complexes formed with cations small enough to enter the ligand cavity. Specifically, those cations which exactly fit the ligand cavity are bound more strongly than those which are too small. However, among cations which are tOO large to enter the ligand cavity, size correspon- dence is not predominant in determining cation selectivity, which is then governed by other factors. The cavity size of 18-crown-6 is large enough to allow entry Of all alkali metal cations except Rb+ and Cs+. The selectivity of 18- crown-6 towards K+ has been demonstrated. On the other hand, for cations whose radii are larger than the hole Size, sandwich complexes can be formed in which the cation is located between two molecules Of crown ether (68,89,90). The formation Of such complexes was detected by alkali metal NMR techniques (91). Crown ethers were used in the early 1970's by Dye and co-workers (70) to enhance the solu- bility Of alkali metals in amine and ether solvents. It was found that the crown ethers form stable complexes with 27 alkali metal cations. These complexes were detected by alkali metal NMR (93). Evaporation Of alkali metal solutions in methylamine which also contain 18—crown-6 produces dark blue film or powders which apparently contain complexed alkali metal cations and either trapped electrons, "elec- trides", or alkali metal anions, "alkalides", depending On the ratio Of the metal to crown ether. Previous studies on "electrides" and "alkalides" Of alkali metal 18-crown-6 systems have concentrated on the optical spectra Of thin films made from solutions containing Na, K, Rb and Cs with 18-crown-6 (75,76,78). The Optical spectra Of films made from sodium 18-crown-6 solutions with R = I%%%%T = 2, Showed a strong band at 16,000 cm"l characteristic of Na-, as well as, a small broad peak at 25000 cm-1 and a shoulder at 19000 cm-l. Potassium films with R = 2 have a strong 1 absorption band at 12,200 cm- and a pronounced shoulder at 9000 cm-1, while films made with R = 1 have bands at 1 and 6500 cm'l. The peak at 12000 cm'1 is 1 12,000 cm“ 1 probably that Of K‘, while those at 9000 cm" and 6500 cm' may be caused by trapped electrons. Films Of Rb with 18- 1 crown-6 with R = 2, show absorptions at 12000 cm- and 9100 czm-1 which are assigned to Rb- and e;, respectively, with the band Of e; shifted tO higher energies probably by interaction with Rb-. However, the behavior Of Cs lB-crown-6 films is different. Films formed from solu— l tions showed a single band at 6400-6700 cm- which prob- ably originates from a trapped electron. The EPR spectra 28 Of Cs 18-crown-6 with R = 0.5 consisted of a single line with g-value 2.0022 i 0.0001. The linewidth changed from 0.36 G at 214 K tO 0.54 G at 106 K. Several attempts to produce crystalline alkalides by using 18-crown-6 as the complexing agent failed, perhaps this was due to the use Of unsuitable crystallization solvents. In addition, characterization of the cesium 18-crown—6 electrides was fragmentary. Therefore, a thorough investigation was re- quired to prepare and characterize "alkalides" and "elec- trides" which utilize 18-crown-6 as a complexing agent. The Objectives Of the present work can be summarized as follows. 1. Preparation and characterization Of powder and/or film electrides which are formed from cesium and 18C6. Several Cs-18C6 systems were prepared by using different solvents and different metal to crown ratios. The prOp- erties Of these systems were studied by optical and EPR spectroscopy, microwave conductivity and magnetic susceptibility. These properties will be discussed in Chapter III. 2. Attempts to prepare and characterize alkalides by using 18C6 as a complexing agent resulted in the prepara- tion Of two crystalline sodide salts Cs+18C6-Na- and K+18C6°Na_. The properties of these new alkalides will be discussed in Chapter IV. 3. Explore the possibility Of preparing alkalides 29 and electrides by direct vapor phase deposition. Pre— liminary experiments showed this to be possible and the films prepared were characterized by Optical spectroscopy. The experimental procedure and the Optical spectra will be presented in Chapter V. 4. Finally, the major breakthrough in this research was the first successful preparation Of a crystalline elec— tride, Cs+18C6-e-. This experiment Opens the door for the preparation Of other new crystalline electrides. The syn- thesis and properties Of this crystalline electride will be discussed in Chapter VI. CHAPTER II EXPERIMENTAL II.A. Materials II.A.l. Complexing Agents l8-crown-6. IUPAC: l, 4, 7, 10, 13, l6-hexaoxacyclo- octadecane (18C6). (Purchased from PCR, Inc. or Parish) was recrystallized from warm acetonitrile tO give the crown- acetonitrile complex (93). The crown ether Obtained by vacuum decomposition Of the acetonitrile complex was then sublimed under high vacuum at 60°C and stored for further use. It is recommended that the purified 18C6 be stored in the dark 13 vacuO or under an inert atmosphere. II.A.2. Solvents After purification, all solvents were stored in vacuum storage bottles and were degassed by freeze-pump-thaw cycles before use. Ammonia: Ammonia (anhydrous, 99.99% Matheson) was treated twice with Na-K alloy, ratio 1:3 then transferred to a vacuum storage bottle. 30 31 Methylamine (MA): Methylamine (98% Matheson) was stirred over calcium hydride for about 48 hours. It was then treated with Na-K alloy until the characteristic deep blue color remained for at least 48 hours at room temperature. The solvent was then vacuum transferred to a vacuum storage bottle. Ethylamine (EA). Ethylamine (anhydrous, Eastman Kodak) was treated in the same way as methylamine. Isopropylamine (IPA). (Anhydrous, Eastman Kodak) was first dired over calcium hydride, then degassed by freeze— pump-thaw cycles. It was then treated with excess Na-K alloy over benzophenone until the characteristic violet color remained for at least 72 hours at room temperature. The solvent was then vacuum transferred immediately to a vacuum storage bottle. The violet benzophenone ketyl served both as a drying agent and a dryness indicator. Diethyl Ether (DEE). Ether (Fisher) was treated in the same way as iSOprOpylamine. II.A.3. Metals Cesium. This metal (a gift from Dow Chemical CO.) had been previously transferred into sealed glass ampoules with breakseals. The metal was stored under vacuum in measured small diameter glass tubes (2-8 mm O-D) with sealed 32 ends. Vacuum transfer Of the metal to these tubes was ac- complished as follows. The ampoule with breakseal was attached tO the distribution apparatus shown in Figure 3. After evacuating the apparatus (<1 x 10"5 torr) and breaking the breakseal, the metal was heated and allowed tO run into the reservoirs after which vacuum seal-Offs were made at the constrictions. The metal was heated and allowed to run down into each of the tubes which had known internal diameters. Due tO the possible contamination Of cesium with sodium from sodium borosilicate glass, it is recom- mended that the final distillation Of cesium should be made in a fused silica apparatus (94). Sodium, Potassium, and Rubidium. These metals (Alfa— Ventron) were supplied under argon in sealed glass am- poules with breakseals. Total purities were 99.95%, 99.45%, and 99.93% reSpectively according to the supplier. These metals were treated in the same way as cesium. Lithium: (Automergic Chemical CO. 99.99%). Lithium was treated in an argon-filled dry box (95). A clean knife was used tO out small pieces Of lithium which were then weighed on a model RTL Cahn Electrobalance inside the dry box. Then each Of the pieces was loaded into a 5 mm tube sealed at one end. A cap tO which pOlyOlefin heat-shrink- able tubing (Alpha-wire Corporation) had been previously attached, was sealed onto the sample tube, providing a gas— tight seal. After removal from the dry box, the end Of the 33 1 4’3 heat ' l ’\G _ . I . /, LASS COVERED L -1 IRON ROD (.— Figure 3. Apparatus for distribution Of alkali metal under vacuum. 34 tube containing the lithium was placed in liquid nitrogen and a seal was made to remove the heat-shrinkable tubing. II.B. Glassware Cleaning A general method was used for all glassware cleaning. The apparatus was first filled with an HF cleaner. This cleaner contains 33% HNO 5% HP, 2% acid-stable detergent, 3’ and 60% H20 by volume (96). After thorough rinsing, the apparatus was filled with aqua regia and allowed tO stand overnight. The apparatus was then rinsed thoroughly with doubly distilled water and dried overnight in an oven at mlSOOC. II.C. Sample Preparation and Instrumental Techniques II.C.1. General Preparative Methods (a) Electrides (CS-18C6) - The apparatus shown in Figure 4 was used tO prepare samples for the study Of Op- tical and EPR spectra and magnetic susceptibilities. This procedure permitted the same preparation to be used for all studies in order tO eliminate differences in behavior which might arise from different preparations (for example, dif- ferent crown-to-metal ratios). A known mass Of 18C6 equivalent to the amount Of Cs metal used was introduced into the apparatus. The small glass tube which contained cesium metal was scored around the middle and introduced 35 Graded Seal Heat . Shrinkable Tubing Reservoir Sidearm Sample Sidearm Consrricfion I [U Quartz OpticalCeH Apparatus for preparation Of powders and films Of Cs+18C6-e-. Figure 4. 36 into the apparatus sidearm. A piece Of Teflon shrinkable tubing was heated gently and sealed at the end Of the side- arm. The other end Of the shrinkable tubing was then sealed onto a short piece of glass tubing which had been sealed at one end. Then the apparatus was connected tO the vacuum line. To avoid contamination Of the manifold, the connection was made through a tee and the tee was con- nected to the manifold gig an intervening liquid nitrogen 5 torr and trap. The apparatus was evacuated to %1 x 10- then removed from the vacuum tee with the valve closed. The cesium ampoule was broken. (It was necessary to cool the cesium prior to breaking since the melting point Of cesium is low and molten cesium reacts with the shrinkable tubing). After breaking the ampoule, the two pieces were shaken down into the sidearm and a glass seal-Off was made behind the pieces in order tO remove the shrinkable tubing. The cesium was distilled, and another seal was made at the constric- tion in order to remove the pieces Of glass. The solvent bottle was then connected to the tee and the connecting tubes were evacuated tO m1 x 10"5 torr. The solvent was distilled into the apparatus by keeping the temperature Of the apparatus lower than the temperature Of the solvent bottle. After introducing a reasonable volume of solvent, the valves were closed and the apparatus was removed from the vacuum line to an iSOprOpanol bath at a temperature Of -40°C. The apparatus was allowed to stand for a period Of time with occasional shaking tO dissolve the metal and 37 the crown and to ensure complexation. Once the solution had been prepared, it was poured gently into the sample tubes for EPR and magnetic susceptibility measurements. The solvent was evaporated slowly tO avoid "bumping" by keeping the temperature Of the bulk solution in the main stem at -78°C and that in the sample tubes at m-60°C. For some unstable electrides such as (CS)2(18C6), the main stem were kept at liquid nitrogen temperatures while the sample tubes was kept at a temperature of -78°C. It was necessary to wash the sample tubes several times to bring the sample down to the bottom Of the tube. After removing essentially all Of the solvent, the bulk solution was frozen with liquid nitrogen, then the samples were dynamically pumped for about 45 minutes to insure that the samples were solvent-free. The sample tubes were then sealed-Off and stored in liquid nitrogen. The remaining solution was used for studies Of optical Spectra. For systems, such as CSl8C6Li in which two metals were used an additional side arm with a frit was added to the appara- tus. (b) Alkalides - The same general procedure used to prepare electride solutions was also used to prepare solu- tions of the alkalides. The apparatus shown in Figure 5 was used to prepare crystalline alkalides. After dis- tilling the metals and making the seals, methylamine was distilled into the apparatus. The solution was maintained 38 Km"! ‘ . “90". To l-i ~Voc - W —O- leo M‘ C-N‘ mm: SYNTHESIS Sample APPARATUS . /C Nm-Distlllabll ' Metal Ampaulo \ °fifi;h‘f E J that m Tub'no )j/A Figure 5. Apparatus for the preparation Of crystalline alkalides. 39 at temperatures Of about -25°C, the methylamine was poured back and forth between the main chamber A and chamber B in order to dissolve the metals and the crown. After that, the whole solution was poured into the main chamber A. The methylamine was distilled out Of the apparatus into an evacuated empty bottle. IsoprOpylamine, which served as a crystallization solvent was then distilled into the apparatus. The solution was kept initially at m—15°C to dissolve all Of the alkalide, and then cooled slowly to -78°C and allowed tO stand overnight during which time the crystals precipitated from the solution. The supernatant iSOprOpyl- amine solution was poured into chamber B, frozen with liquid nitrogen and a seal was made at the constriction. Then di- ethylether was distilled into the apparatus to wash the crystals and transfer them to part C. The diethylether solution was poured back into the main chamber A, redis- tilled onto the crystals several times to wash them and finally frozen in the main chamber A with liquid nitrogen and the final seal-Off was made at the constriction. The crystals were then distributed into the sample tubes which were sealed—Off and stored at dry-ice temperatures. In the preparation Of K+18C6°Na- a co-solvent Of iSOprOpyl- amine and diethylether in a ratio Of about 2:1 was used for the crystallization rather than pure iSOprOpylamine. 40 II.C.2. Optical §pectra Solutions were prepared in the apparatus shown in Figure 4. However, films for Optical Spectra were formed in the apparatus which consisted only Of the main stem and the reservoir side arm, all other arms having been sealed- Off. With the exception Of the graded seal and the boro- silicate glass vacuum valve, the apparatus was constructed Of fused silica in order tO avoid sodium exchange contamina- tion Of the alkali metal solutions by contact with sodium borosilicate glass (97). Films were formed in the Optical cell by leaving 0.2 ml of the solution in the cell, while the bulk solution in the reservoir arm was frozen in liquid nitrogen. During this process, the Optical cell was kept in a cold iSOprOpanol bath and was shaken rapidly from side to side about the axis through the side arm. This rapid Shaking splashed the solution onto the cell walls where the film formed during the flash evaporation Of the solvent. It was necessary sometimes tO repeat this process tO Obtain films Of prOper thickness and uniformity. Because Of the nature Of this method Of film preparation, films were Often non-uniform, especially when a solvent Of high vapor pressure such as ammonia was used. It was noticed also that slow evaporation Of the solvent resulted in more uniform films. In general, all films from methylamine solutions were more uniform than those from ammonia 41 solutions. Some Of the films, classified as damp and wet films were prepared by bringing the temperature Of the reservoir up to temperatures about 5-20°C below the Optical cell temperature. A11 Optical spectra were recorded on a double beam recording spectrophotometer (Beckman DK-2), modified to permit control of the sample compartment tempera- ture between -65°C and 0°C. An ethanol cooling bath (Neslab Model LTE-9) provided rough temperature control for the compartment and was augmented by nitrogen gas which flowed through a coil immersed in liquid nitrogen. A copper- constantan thermocouple placed near the cell provided the signal for temperature readout (Doric Model DS-350). Spectra l l were recorded from 4000 om’ (2500 nm) to 25,000 cm" (400 nm) for standard fused silica cells and from 3125 cm”1 (3200 nm) to 25000 cm"1 (400 nm) for infrasil cells. The reference beam passed through air. The Spectra were normalized to a scale Of 0.0 tO 1.0 by subtracting a base- line correction, setting the lowest absorbance tO zero and the maximum to 1.0 and then scaling the measured absorbance 1 intervals. The baseline was Obtained by at 500 cm” using the spectrum Of an empty cell. It was necessary to use a new baseline correction for each new run. II.C.3. EPR Spectra EPR spectra were recorded On an x-band spectrometer (Varian Model E-4 or Bruker Model 200) over the temperature 42 range 3.6-225 K. Above 100 K, a variable temperature con- troller (Varian) was used which utilized a stream Of tem- perature-regulated cold nitron gas. The temperature set- ings were calibrated with a copper-constantan thermocouple with digital readout (Doric Model DS-350). Temperatures between 3.6 and 77 K were provided by a continuous flow liquid helium system (Oxford Instrument CO., Ltd., Model ESR 9). Digital temperature readout was based on an (Au + 0.03% Fe/chromel) thermocouple just below the sample. II.C.4. Magnetic Susceptibility Samples for magnetic susceptibility were prepared in 4-6 mm O.d. "Spectrosil" fused silica (Thermal American fused quartz). The sample was located in the bottom Of the tube. A seal was made about 1 cm above the sample while the sample was immersed in liquid nitrogen. A small fused silica hook was attached to the top of the sample tube ("bucket"). Care was taken to avoid contamination of the bucket and the hook during seal-Off and handling. For example, handling the hot quartz with metallic forceps can lead to contamination. Magnetic susceptibility measure- ments were made with a SQUID (Superconducting Quantum Inter- ference Device) magnetometer purchased from S.H.E. Company. Since most Of the electrides decompose at temperatures of -30°C and above, it was necessary to load the samples into the SQUID while keeping them cold. This was achieved 43 by using a copper block that fit into the airlock of the SQUID. This block had a central hole Of 7 mm diameter so that the sample could pass through it. TO avoid SQUID contamination with air and moisture, a glove bag was placed around the airlock and purged with a continuous helium flow. Prior to loading the sample, a thread 15—20 cm long was attached tO the hOOk Of the bucket. While the sample was kept in liquid nitrogen, the copper block was cooled tO liquid nitrogen temperatures, then placed in the airlock and the thread was hooked to the tape Of the drive motor. The sample was placed into the hole in the copper block, the airlock cover was put back and the airlock was rapidly evacuated and pressurized with helium gas three times. Then the sample was loaded into the SQUID. It is essential that the loading process be completed as rapidly as pos- sible so that the copper block does not warm up with result- ing sample decomposition. Measurements were made over the temperature range 1.7 - 280 K. To correct for the dia- magnetism Of the bucket, it was ejected from the SQUID and allowed to remain in the airlock at room temperature long enough to decompose the sample. Then the bucket and de- composed sample were loaded back intO the SQUID for further measurement. The reading given by the SQUID is the total magnetic moment in e.m.u. The magnetic susceptibility is calculated by using the formula 44 magnetic moment Xgm = (SF) x field in Gauss x mass in gm. where SF is the total scale factor which equals the scale Of the SQUID control multiplied by the scale Of sample measurement control. The magnetic susceptibility Of the sample was calculated by subtracting the diamagnetism Of the bucket and decomposed sample from the total magnetic susceptibility. II.C.5. Pressed Powder Conductivity Conductivity measurements were made in an apparatus designed by J. L. Dye and Michael R. Yemen (98). A powder sample was placed between two stainless steel electrodes inside a 2 mm I.D. heavy wall fused silica tube. A steel spring whose force constant had been measured was used to compress the sample. A variable temperature controller (Varian Model V-4540) was used to control the sample tem- perature. Powdered samples under vacuum which had been stored at dry-ice temperatures were broken in an inert atmosphere glove bag. The sample was transferred to the pre-cooled conductivity sample chamber. After the sample had been loaded, the current at various voltages was measured to determine whether Ohm's law was obeyed. Then the current through the sample was measured at a constant voltage at 45 a number Of different temperatures. II.D. Analysis TO begin the analysis Of samples they were first de— composed with doubly distilled water in a closed vacuum system (95). The hydrogen evolved was collected and measured and the unreduced water which had been distilled into a trap was analyzed for amine solvent content in the crystalline sample. An aqueous solution Of the decomposed material was made and divided into two parts, one for titration with a standard acid, the other for metal determination by flame emission spectrometry. Finally, the solution was evaporated tO dryness and the residue was dissolved in D 0 so that 2 quantitative proton NMR could be used to determine the amount Of complexant present. Potassium hydrogen phthalate (KHP) was used as an internal reference for the NMR integra- tion. II.D.l. Hydrogen Evolution The sample tube was scored and sealed into a glass ap— paratus with Teflon heat-shrinkable tubing. This apparatus was then connected tO the hydrogen collection apparatus. The entire system was evacuated tO W3 x 10"5 torr and the conductance water used to decompose the sample was de- gassed several times through freeze-pump-thaw cycles until 46 no detectable gas remained. The sample tube was then broken and the conductance water was condensed onto the sample very slowly and at a low temperature. The sample was decomposed by reacting with water according to M+C°N- + ZHZO + M+ + c + H2 + N+ + 20H". The hydrogen evolved was pumped through two liquid nitro- gen traps with a manually Operated Toepler pump and col- lected in a calibrated pipet. After repeated cycles Of a mercury leveling bulb, the height Of mercury, the atmos- pheric pressure and the temperature at the pipet were measured and the number Of moles Of evolved hydrogen was calculated by using the ideal gas law. II.D.2. pH Titration The decomposed solution was evaporated during hydrogen collection by condensing the water in a trap at liquid nitro- gen temperature. The water was analyzed for amine solvent by measuring its pH. The residue vessel was removed, and under a nitrogen atmosphere in a glove bag a known amount Of conductance water was added and a measured portion was then titrated with standard HCl solution by using a pH electrode (Corning, catalog number 476050) and digital pH meter (Orion Research Model 701A) which had been calibrated with pH buffer solutions. However, it was not possible to 47 calculate the exact hydroxide ion concentration in the presence Of 18C6 because Of the instability Of the pH meter, especially in the vicinity Of the end point. This behavior was verified by titration Of a solution Of cesium hydroxide and 18C6. II.D.3. Flame Emission Part Of the solution was used for flame emission. The flame emission instrument (Jarrell Ash) was adjusted for the estimated parts per million (ppm) concentration Of the unknown sample. Standard solutions Of the apprOpriate metal(s) were run followed by the unknown solution. The emission value was read from a digital averager. The reading from conductance water was determined between all measurements which yielded a background emission or noise level. Calibration curves for the standards were prepared by plotting the emission output versus concentration in ppm. The concentration Of the unknown was then calculated from the calibration curve. 11.0.4. 1H NMR The solution was evaporated to dryness in a partially evacuated desiccator with "Drierite" as a drying agent followed by dynamic pumping tO be sure that nO water was left. Then the residue was dissolved in 2 ml Of a known 48 concentration Of KHP solution in D20, which was used as a reference. Measurements were done on a Bruker 250 MHz NMR instrument. The integrated areas Of the KHP and the unknown were compared and the number Of moles Of the un- known sample was calculated. CHAPTER III CESIUM lB-CROWN-G ELECTRIDES The Optical spectra Of films of composition Cs+l8C6-e- have been reported previously. The results indicated non- metallic behavior since the Optical spectrum was quite similar tO that Of dilute metal ammonia solutions. How- ever, since the earlier investigations Of the nature Of these electrides were only fragmentary, further investif gation, using different techniques, was undertaken to better understand the nature Of these electrides. The abbreviation Cs+18C6-e— indicates the materials from which the films or powders are made (Cs and 18C6). The symbol Cs+l8C6-e- does not represent the nominal stoichiometry or the metallic or insulator character Of the electrons. The ratio Of the moles of metal tO moles Of 18C6, R, will identify the stoi- chiometry Of a particular preparation. The characteriza- tion Of the nature Of these electrides was done by Optical spectra, EPR, microwave conductivity and magnetic sus- ceptibility. In some cases the characteristics are similar tO those Of metal-ammonia compounds, though there are many differences. The various methods Of study and their re- sults are detailed below. 49 50 III.A. Optical Spectroscopy III.A.1. Cs-18C6 Films From Ammonia The spectra Of dry films made from ammonia solution containing Cs and 18C6 with R = 0.5 and 1 are similar. In the spectrum shown in Figure 6 with R = l, the main peak at 7000 cm.1 is presumably due to trapped electrons (et due to interaction Of e; with the cation Cs+18C6 or else to the presence Of different traps for the electron be- ). A shoulder at 8500 cm.1 is also Observed, probably cause Of inhomogenities in the solid. III.A.2. Films From Methylamine Due to the inhomogenity and instability Of dry films produced from ammonia solutions, it was difficult tO characterize the nature Of the electrides. However, using methylamine as a solvent produced more homogenity and en- hanced the stability Of the films. Figure 7 shows the spectrum Of a dry film from methylamine solution with R = The main peak at 7000 cm—1 is due to the trapped electron e;, in the far-infrared region. The damp film shows the same the peak is somewhat broad and there is no absorption features except that the peak is less broad. Dry films from methylamine solution with R = 1 give a Single peak at 6500 cm-1 (shown in Figure 7) due to trapped electrons. Interaction Of these films with methylamine vapor causes 51 .COHu IDHOm mac0EEw EOHMAH n mv momamu mo EHHM who m mo Eouuooom Hmowumo .m anomam n..O_. 3%:th ON 9 O_ m _ q . . 0.0 t .. no SEN no mo are 0... om o._ 2.1:. 52 v .Aaum .lm.e u a ----c mSOHusHOm ocHEMHSnuoE Eoum wOmHmO mo mEHHm Sup mo muuoomm HMOAHQO .5 musmflm MIO_ . ATEUVNN ON 0. o. m . . . _ 00 III § . // _ / _ x _ / x _ . x2: I // 7400 4 / ..I..__ / x A‘ / l / / s .l/ x / \ x \ , . o._ md 53 a shift in the peak position towards shorter wavelength and produces shoulders on both sides of the peak. This might be due to more interaction between the electron and the cation and the presence Of more than one site for the elec- tron. These two systems show almost the same Optical prop— erties and are similar to dilute metal ammonia solutions in- dicating the presence Of localized electrons. However, making the film richer in crown ether destabilizes the films and produces more heterogenity. In order tO per- mit comparison with the lithium-cryptand (C211) systems which always give absorptions in the infrared region even up to R = 2 (79), a series Of methylamine solutions which contained cesium metal and 18C6 in different ratios, R = 1.25, 1.5, 1.75 and 2 were prepared. The optical spectra Of dry films formed from these solutions were studied. Figure 8 shows the spectrum Of a dry film made at a tem— l.5. There are two peaks, one perature Of -42°C with R at 10500 cm"1 probably due to Cs- and a second peak at 1 8500 cm- due to trapped electrons. However, the Cs- t t peak is shifted (compared with the et peak from Cs18C6, R = 1) towards shorter wavelength, peak is a little more intense than the e peak. The position Of the e probably due to the interaction Of the trapped electron with Cs-. Another characteristic Of the spectrum is the high absorption in the far-infrared region compared with the spectrum Of a film Of Cs 18C6 R = l. The Optical spectra for this system were also recorded when the films 54 ON .AUomml um OOOE EHHM nThV_.A-.%PRMYR 9 O_ V Avomvl um acme EHHM unnnv cofiusHOm anaemamnuoe Eoum Am.H u my mOmHmO mo muuomam Acapumo .m wusmflm - u 0.0 0.0 x08 <1 0.0 O._ 55 were made at lower temperatures. Figure 8 shows the Optical spectrum Of a dry film made at a temperature Of -55°C. As the spectrum indicates, there is a major peak 1 at 11000 cm- (probably caused by Cs-) and an ill-defined broad shoulder at 8000 cm-1. From these spectra we can conclude that in the range 1 < R < 2 there are two Species e; and CS- and that they are temperature dependent. High temperatures favor the formation Of the trapped electron over Cs-. However, once the film has been formed, tem- perature does not affect the relative concentrations Of e; and Cs-. When these films are made wet with methyl- amine vapor they shift to give a localized peak due to the solvated electron. However, these systems are difficult to characterize since there is more than one species. This was not the case with C518C6 when R = 2. Figure 9 shows the spectrum Of a dry film made at a temperature of -52°C Of C318C6 with R = 2. The main feature Of the spectrum is the high absorbance in the infrared region and the ap- parent metallic "plasma" character similar to that Of con- centrated metal ammonia solutions. When the film was made at a temperature Of about -45°C or higher the spectra Show some changes. Figure 9 shows the spectrum Of such a dry film with R = 2. The major characteristic Of this spectrum is the high absorption in the far-infrared region which lOOks "plasma" like and two peaks at 7000 cm-1 and 8500 cm-1. These multiple peaks are probably due tO different 56 .AOomv I Edam onu mo ououmuomamu IIIIV .AOonI Edam asp mo ousumuomfiou V coHusHOm ocflsmamnuofi Eoum m u m mumamu mo mEHHm mo muuoomm HMOflumo nuO_ 87.58%. .m anomwm 57 environments Of the electron. Making these films "damp" with methylamine vapor tends to localize the electron peak. In spite of these complexeties, all Of the above mentioned spectra were reproducible. It was also noticed that the intensities Of the spectra increase dramatically at a cer- tain temperature which depends on the ratio R. As R in- creases that temperature decreases. This is probably due to melting or annealing Of the film which then produces more homogeneous and uniform films. III.B. EPR Electron paramagnetic resonance spectrOSCOpy (EPR), in contrast tO other techniques such as magnetic sus- ceptibility detects the local environment Of the unpaired spins. When a magnetic field is applied to a spin system, the energy levels of an unpaired electron are separated due tO the Zeeman effect AB = gBHO (6) where B is the Bohr magneton, g is the Landé g-factor and H0 is the applied field. Transitions between the Zeeman levels can be achieved by radiation of appropriate fre- quency according to the equation AB = hv. 58 For an electron with a spin angular momentum % and L = 0, so that there is no spin-orbit interaction, g = 2.0023. The g-value is used as a measure Of the electron's inter- action with its environment. The lineshape is determined by the type Of coupling between the spins and their en- vironment. However, there are two theoretical expressions commonly used to describe the EPR spectra which have sym- metrical 1ineshapes (53). For systems where the spin- lattice and exchange effects predominate, the lineshape is Lorentzian. The Gaussian shape usually applies to systems where spin-Spin couplings predominate. The linewidth of the EPR line is determined by the degree Of coup- ling Of the unpaired electrons tO their environment. If such coupling is strong, i.e., the relaxation times are short, then the electrons spend only a small time At in the upper energy level. This time At corresponds to an uncertainty AB in the energy given by AEAt W h. There- fore, there will be a broadening Of the lines Observed in the EPR spectrum. In addition, there are other sources of line broadening which can be divided into two major groups. The inhomogeneous broadening, in which the unpaired electron is subjected to slightly different effective fields. Therefore, the observed line is a superposition Of a number of individual components referred tO as "spin packets". _The reasons for inhomogeneous broadening are, inhomogeneous field, anisotropic interaction in randomly oriented systems and unresolved hyperfine structure. The 59 other kind is homogeneous broadening caused by non-uni- formity Of the instantaneous magnetic field at each di— pole. There are several mechanisms which account for this homogeneous broadening. These include (a) electron Spin- electron spin dipolar interaction which depends on the concentration Of the paramagnetic centers; (b) electron spin-nuclear spin interaction, caused by the random local field produced by magnetic nuclei in the vicinity Of the electron; and (c) electron exchange where the electrons in different Sites exchange their Spin states. This ex- change-narrowing is important in bimolecular reactions. For dilute samples the spectra could have more than one line and as the concentration increases, these lines coalesce to a Single line which becomes narrower at even higher concentrations. In solids this exchange-narrowing arises from the overlap Of molecular wave functions (99). III.B.l. Results and Discussion III.B.l.a. Solids From Cs-18C6 Ammonia Solution The EPR spectra Of films and/or powders Obtained by evaporating ammonia from solutions which contained Cs and lB-crown-6 with a ratio Of R = 0.5 or 1 where (R = moles of metal moles Of crown ) were studied in the temperature range 4.2 to 230 K. In general both systems showed the same line- shape and the value Of g (the ratio of the low-to high-field 60 peak amplitudes in the EPR spectrum) was always greater than 1. The variation Of g-with temperature is shown in Figure 10 for the solid Cs-18C6 with R = 1. This high value of %~might be due to inhomogenities of the system which give more than one trapping site for the electron, resulting in closely-spaced peaks. The g-value determined by cali- bration with a standard diphenylpicrylhydrazyl (DPPH) (99) sample is 2.0019:0.0001ikn:the system with R = 0.5 and 2.0028i0.0001 for the system with R = 1. The g-values determined in the temperature range 230 to 140 K showed no temperature dependence. The g-values for these systems suggest that there is some interaction between the electron and its environment (presumably the Cs+18C6 cation). The possibility Of the presence Of other phases and excess crown and metal atoms cannot be ruled out. The linewidth was 0.6 Gauss at 230 K and increased tO 2.2 Gauss at 4.2 K for Csl8C6 with R = 0.5, and 0.5 Gauss at 230 K which in- creased tO 1.05 Gauss at 4.2 K for Csl8C6 with R = 1. The variation Of linewidth with temperature for Csl8C6 with R = l is shown in Figure 11. The general behavior is that the linewidth increases with decreasing tempera- ture. The discrepancies in the values Of the linewidths in the region Of temperature where the values determined in the nitrogen temperature range Should overlap with those determined in the liquid helium range might be due to the temperature cycle and tO the conditions Of the instrument 61 .cOHuSHOm mac086m Eoum pmummmnm H I 225.: l m rows lo.GOmH+mo mo mowumm mxd .OH apnoea om ON CNN Om. OW. 00. - - — - q q o O 0000 n;— AU .0 O o 83v nu no .I-. .v_1 Mw_ . mm. 0 O H... O O o o . im._ . L p L _ . 62 .umumowuo NZIa maonamm ammo upcomomuo wmla cufl3 couomHHOO mump mHOQESm OHHOm .cofluoaom wficoaam Eoum AH n my Im.mUmH+mO mo mnupflzwcfiq .HH ousmflm A v: .anP OWN om. 9w. co. om ON a _ ssnog) d-de 63 MHCOEEM Eouw AH n mv OdN .COHusHOw no.mUmH+mO now ousumuomfiau .m> mcflmm mo nooaoz .NH ouomwm Av: .95... CON Om. OW_ 1 In. N N . .0 .mmo Ila O S m. U . . the . ..o.. m 0 l m. N 64 since the two runs were done separately. The number Of spins calculated by using a spin standard Ruby sample 3+ (A1203 Cr creases with decreasing temperature. Figure 12 Shows ) (100) indicated that the number Of spins in- the plot Of the number Of spins with temperature for C518C6 with R = 1. However, the inhomogenity Of the films and/or powders Obtained from ammonia solutions made it difficult to further characterize these systems. III.B.1.b. Solids from Cs-18C6 Methylamine Solu- 31.9% A series Of methylamine solutions which contain cesium and 18-crown-6 were prepared with different ratios Of metal to crown. The EPR spectra Of the solid with R = metal crown g-value was found tO be 2.0023:0.0001 which is close to the = 0.5 were recorded in the range 4.2 tO 230 K. The free electron value and suggests little interaction between the electron and its environment. The variation Of line- width with temperature is shown in Figure 13, which shows that the linewidth increases gradually from 0.45 Gauss at 230 K tO 2.0 Gauss at 4.2 K. The % ratio (shown in Figure 14) was always greater than 1, probably due tO the presence Of several peaks originating from different trapping sites Of the electron. This was also indicated by the broadness Of the infrared peak in the Optical Spectra. The number Of spins calculated using a spin standard Ruby as a function 65 .umpm Iowuo NZIa I maonemm some “umumomwo mmla nufl3 wouooaaoo dump I maonfimm UHHOm .cOHuSHOm acflsmahauoe Eoum m.o n m nuflz Io.mOmH+mO mo nuofl3mcwa .ma whomflm 3:. deep ONN om. o: oo. oo om . . . . q 1. 7 q . 100000 0000 coo IVO o o oo o . I ImOv o H 4 a -m... 9 D n M» 66 .umumowuo NzIa I mHOQESm ammo “unpmowuo omla on» nufl3 pwuomaaoo camp I maon Ifiwm pflaom .coflunaom onwEmenuoE Eoum m.o n m Io.mUwH+mU mo mowumu m\< A v: asap 0mm am. o: oo. oo om _ . _ .wa musmfim - u - J_._ 8:v 0008 9m.— 67 Of temperature is shown in Figure 15. The figure Shows that the number of Spins increases with increasing temperature. We found the system Cs-18C6 with R = l to be more attractive in terms Of homogenity and stability than with R = 0.5. Films and/or powder samples Of Cs18C6 with R = l were pre- pared and studied by EPR spectroscopy. The g-value cal- culated for this system is 2.0024 0.0001 which is close to the free electron value and might suggest very little inter- action between the trapped electron and the Cs+18C6 cation. The linewidth, as shown in Figure 16, changes with tem- perature. It is 0.6 Gauss at 230 K and increases to about 1.5 Gauss at 4.6 K. This increase in linewidth might be explained by the presence of more than one peak. The fact that the lineshape has a structure at about 4 K (Figure 17) supports this, while at higher temperatures these peaks overlap to form a single line. The % ratio shown in Figure 18 in the liquid nitrogen temperature range is about 1.15, while in the liquid helium temperature range it is about 1 at 4 K which decreases gradually to about 0.71. A possible explanation for values which are less than 1 is the pres- ence Of more than one peak. As shown from the figures, the two samples showed essentially the same behavior. Figure 19 shows the variation Of the number Of spins calculated by using a spin standard Ruby. It is difficult tO con- clude any Specific behavior, but the general behavior is that the number Of spins is less at higher temperatures. 68 ruflz Id.oOmH+mo pom _.v_ N. .COHuoHOm mmzoz Eoum Am.o u my ououmummamu HMOOHQAOOH umcflmmm mcflmm mo HOQEDZ .ma wusmflm - fi 2 9‘. : 2 show suIds )0 on: 5". 69 .cowumnmmoum u:muoumap Eoum onEwm “anyone so pouooaaoo numb moumsqm OHHOm .uwum Iomwo sza I mHOQE>m ammo .ucumowuo mmIa Suez Oouoaaaoo camp I maonshm pflaom .cOHuoHOm acflfiwawnuwe Eonm H u m nuw3 Io.wUmH+wO mo nupfizocaq A v: .anp 03 00. ON 00 .oH ousoam ONN Om. d d — .d v ..N.O I I I I I I I I . v I fly Av AU .. C. O. C. I. |.mw.nv “H” O 000 O o O OI Aw . d fi 010; ma nu 33 ea 1. I: .l. .- . h .lb - . — 70 ucaummmap ozu um mcHEmawnuoE Eonw AH x t? «F xmm u... 0N .mmusumnmmeu ml I0.eOmH+mo mo mwuooam mam .SH magmas 71 .cOHumummaum ucouomwap Eoum mamewm wmnuocm co oouooaaoo pump I moumoqm “umumomwo Nqu I maoofixm coma «umumomno omIa nufl3 pmuowaaoo dump I maonfixm Uwaom .wsHEmamsuaE Eoum AH n my Io.mUmH+mU you mowumn m\< .mH ousmflm 3: deep CNN 09 ov. 00_ cm ON q q — _ Ifi - O I I I I II I I I . .mo 0. II Qu—V . no III.C— w. mm. O oo OO o o . .N. p P - _ _ _ 72 .cofluoaom anaemawnuofi Eonm AH I my Io.mOmH+mU you muopmnmemu Hmooumfloou .m> mcflmm mo HmoEoz .ma anomflm I. ..v. no. .I_- t J C)! 00 1 .m.m N .0 o m. 1 a 40m A.» w. s o . I.I r Iva. .u nv II 9 nWU 1 l a). 0’) 73 However, since the possibility Of the existence Of two kinds Of material is demonstrated by the linewidth and % ratio. The calculated number Of spins could be distorted by changes in lineshape. The system Cs-18C6 with R = 1.5 was also studied by EPR in the liquid nitrogen tem- perature range. The % ratio was less than 1, which sug- gests the presence Of more than one peak. The g-value is 2.0022:0.0001, which is close to the free electron value and suggests little interaction between the trapped elec- tron and its environment. The linewidth in the liquid nitrogen range is about 0.5 to 0.6 Gauss. The number Of spins calculated indicates that only about 4% Of the avail- able electrons contribute tO the EPR signal. However, due to the presence Of at least two kinds Of species, the trapped electron, e; and Cs- as indicated by the Optical spectra it is very difficult to characterize this system. On the other extreme, the system Cs-18C6 with R = 2 showed com- pletely different properties as indicated by its Optical and EPR spectra and other techniques. The average g- value in the temperature range 118-200 K is 2.0026i0.0001. The g-value Showed some dependence on temperatures. While the g-value is 2.00247 at 118 K it shifted to 2.00253 at 171 K and to 2.0027 at 201 K. This indicates that the interaction between the electrons and their environment is greater at high temperatures. The 5 ratio is about 1.3, B a possible indication Of Dysonian lineshape which represents 74 metallic systems (25). The Optical Spectra Of films made for this system also showed this metallic behavior. The linewidth in the liquid nitrogen temperature range is about 0.5 Gauss. The percentage Of the free Spins (calculated by using a spin standard Ruby) which participate in the EPR absorption is only about 0.05%. This calculation is only approximate because it is valid only for systems which have Lorentzian lineshape. The EPR spectra in the liquid helium range were measured on two samples from two different preparations. The first sample which appeared tO be a powder gave % z 1.1 and a linewidth of about 37.5 Gauss at 4.3 K which decreased tO 28.5 Gauss at 120 K. As the temperature increased the linewidth decreased gradually until the temperature reached about 140-150 K, at which point there was a sudden change in the spectrum. The linewidth drOpped tO about 5 Gauss. This change was re- versible with reSpect to temperature. Reducing the tem- perature once again yielded large linewidths. The line- shape below 140 K indicates the presence Of one peak and two shoulders at the ends Of the peak (Figure 20), while above 140 K there is only one single peak (Figure 20). The other sample, which appeared to be a film on the walls Of the EPR tube, showed qualitatively the same EPR spec- tral behavior. g-was about 1 and the linewidth was about 25 Gauss at 4.2 K. It remained almost constant at this value up to N140 K. At 140 K the EPR spectrum changed to 75 .cowuoHOm OGHEma>nuoE Eouw m n m nuw3 Io.mOmHmU mo mwuommm mom .om ousmflm xméu ._. YES; \I 1 com 76 give a more symmetrical lineshape and a smaller linewidth. This behavior might be due to an electron exchange pro- cess (99) which participates in the broadening Of the peak and there might be several unresolved peaks which overlap at higher temperatures to give a single symmetrical line. The dependence Of the g-value on temperature supports this explanation. However, possible phase change cannot be ruled out. This system looked crystalline under the micro- SCOpe. However, the instability and the reactivity Of this system made it impossible at this point to prepare tem— perature stable powders or crystals. III.C. Magnetic Susceptibility Magnetic susceptibility is determined by the response Of the material tO an applied external magnetic field. When a substance is placed in a magnetic field of strength H, the magnetic induction B in the substance is given by the equation B = H + 4NM (7) where M is the intensity of magnetization or the magnetic moment per unit volume. For an isotropic material M is proportional to H and the proportionality constant is the volume magnetic susceptibility (8) EH: Xv 77 Quite frequently the susceptibility is referred tO unit mass or to a mole of the substance. The gram susceptibility is given by x = -— (9) where p is the density. If M is the molecular weight, the molar susceptibility is given by = MX (10) XM gm All materials have diamagnetic contributions to their susceptibility which results from the orbital motion Of the electrons. If the net susceptibility is negative, the substance is diamagnetic. Substances with a positive sus— ceptibility are paramagnetic. Paramaqnetism occurs only in materials in which individual atoms or molecules have permanent magnetic moments. The paramagnetism arises from several sources. The free spin paramagnetism which is due to non-interacting unpaired electrons can be described by the Curie-Law x:—=————=— 11 where N is Avogardo's number, k is Boltzmann constant, “B is the Bohr magneton (eh/4 Me) and P is the effective 78 number Of Bohr magnetons and is defined as t P = g[J(J+l)] (12) where J is the total angular moment quantum number and g is the Landé g-factor. For a system with single spin and no spin-orbit interaction P = g[S(S+l)]%, where S is the spin quantum number. For the case S = t, the Curie constant, 3 k C, calculated from the above equation is 0.37604 cm mole"l for a mole Of free spins. Equation 12 must be modified when internal interactions occur between the mag- netic atoms, which tend to align the spins. Weiss first postulated such interactions and it was shown later by Heisenberg that this interaction can be described as a result Of quantum mechanical exchange interactions (101). The exchange energy Of a two electron system is written as —2J12(Sl-SZ), where J is the exchange integral and is related to the overlap Of the charge distribution Of the atoms 1 and 2. For a ferromagnetic system in which the magnetic moments tend to align parallel to each other the exchange integral J is positive while for antiferromagnetic systems it is negative. In the temperature range where antiferromagnetic and ferromagnetic interactions are too weak to lead to spontaneous ordering, the susceptibility can be described by the Curie-Weiss Law _ C X — _T-0 (13) 79 where 0 is a temperature characteristic Of the material generally called the Weiss constant. 0 is positive for ferromagnets and usually, but not always, negative for antiferromagnets. Another type Of paramagnetism is due to conduction electrons in a metal. Pauli explained this phenomenon by the application Of the Fermi-Dirac distribu- tion (102). In this model, which is called the Fermi-gas model, the net magnetization of the conduction electrons is given by Nqu/kBT and only the electrons within the range kBT Of the Fermi energy are likely to change spin in an applied field. Therefore, only a fraction T/TF (where T is the Fermi temperature) of the total number Of F electrons contribute to the susceptibility, hence N112 = B (14) This result is the net magnetic susceptibility for con- duction electrons after making the Landau's correction for diamagnetism. III.C.l. Results and Discussion III.C.l.a. Cs(18C6)2 Solid From Methylamine Solution Magnetic susceptibility measurements were done on solid powder, produced by evaporation of the solvent from methyl- amine solution with the composition Cs and 18C6 in a ratio 80 .mcofiusHOm onwEmHmnumE Eonm Am.o u my Io.muma+mu mo amouMHOQEau .m> muflaflnflumoomom HMHOE Hmooumflowu mo uOHm .HN musmflm germs mm. am. no. oo 9. _ _ . _ X/I . w I pp. 0.0 I... x O \ a \s 90 t . 1 .n IN_O O (O. o \\ o (\O. . T. \\.\\.\\ liq—O 81 1:2. The measurements in the temperature range 1.7 to 200 K showed a temperature dependent paramagnetism. The plot of A; vs. T (Figure 21) does not follow the Curie-Weiss Law.m This behavior was also observed in another more stable system, Cs-18C6 in a ratio 1:1 (see below). Because Of this behavior and other reasons such as the possible presence of different phases in the solid, this system was not studied further. III.C.1.b. (CS)-(18C6)l: Several samples from methylamine solution of composi— tion Cs-l8C6 in a ratio 1:1 were prepared and their mag- netic susceptibilities were measured in the temperature range 1.7 to 250 K. All the samples showed the same quali- tative behavior, but quantitatively there were some dif- ferences which might be due to partial decomposition which is likely to occur to different extents in different sam- ples because Of difficulties in handling these reactive com- pounds. The results reported here are those for the most stable samples determined by its blue color and the absence of white color due to decomposition after ejecting the sample from the SQUID when the run had been completed. The reciprocal molar susceptibility i: was plotted yg temperature as shown in Figure 22. The plot does not follow the Curie-Weiss law in any of the samples over the whole temperature range. A possible explanation is the 82 .mommouoop ousumuomfiau mm :oxmu pump I o «mommouo IcH ousumuoofiou mm coxmu pump < I mSOAuSHOm mcHEmamnuoE Baum AH u my Ia.wuma+mo Ham ousumuomfiou .m> aufiaanwumoomom HMHOE Hmooumfloou mo uoam .mm ouooflm A v. v .an... mm. on. mo. 00 m. - . _ 4 _ \ Q 0 83 presence of two non-interacting magnetic sites or two types of materials. In this case the susceptibility would be the sum of two individual susceptibilities. The results were fitted to the following equation by using the non- linear least squares program KINFIT (103). _1_ Xm T-6 T-0 The fit of the measurements on the three samples gave al- most the same qualitative behavior, but they differed quantitatively. The values of the parameters from the fit are tabulated below: Table 1. Values for the Parameters in the Fit of the Static Susceptibility for Three Different Sam- ples with Two Non-Interacting Centers. Sample Cl 01 C2 02 l 0.033:0.003 -2.06i0.35 0.275i0.06l -262:85 2 0.012i0.001 -l.44:0.l6 0.095r0.260 -7981247 3 0.023i0.002 -l.82i0.25 l.330:0.480 -725:309 The values of 0 and 62 suggest that the sample exhibit l antiferromagnetic behavior. 84 III.C.1.c. (Cs)2(l8C6)l This system Showed very different behavior. Figure 23 shows a plot of the molar magnetic susceptibility ys tem- perature. The sample has almost a temperature-independent paramagnetism above 30 K and a temperature dependent para- magnetism below 30 K. The behavior above 30 K can be interpreted as the Pauli paramagnetism for conduction elec- trons in a metallic system. The optical and EPR spectra indicate such metallic behavior. The behavior at low temperatures might be due to some magnetic impurities. To be certain about this behavior, reproducibility Of these results should be demonstrated, but the high re- activity and instability of this system made it very dif- ficult to prepare. III.D. Microwave Conductivity - Results and Discussion The microwave conductivity measurements Of several Cs- 18C6 systems were conducted in the microwave region (x- band) following the method of Lok (104). In this method the relative power absorption of the sample was compared with known standards ranging from metals to insulators. The samples were prepared in 3 mm O.D. fused Silica tubes, placed in a microwave cavity (TE ) and the power trans- 103 mitted was measured with a power meter (Hewlett-Packard Model 432 A) attached to a 10 db coupler. All standard samples were commercially available and were used without 85 .momMOHOOO ousumuoofiou mm coxmu pump oHQEomou mmumsqm “mammouocfi who» ImuomEou mm coxmu pump oaofiomou moaouflu .cofiuoaom ocflEmHmnuoE Eoum AN 0 mv Io.mOmH+mO Mom onsumuomamu,.m> wuflaflnwumoomom HMHOE mo poam .mm onsmwm 3: deep com mm. on. no om . m. u — - 86 further purification. Because of the unknown skin depth of the Cs-18C6 samples and because the cavity filling factor varied between samples and standards, this method gives only qualitative results, but it can certainly distinguish qualitatively between metals, low-gap semiconductors and insulators. It should be noted that for band gaps much larger than kBT the results are indistinguishable from those Obtained with insulators. Tables 2 and 3 summarize the results of the microwave conductivity measurements of Cs-18C6 systems compared with the standards. The measure- ments showed that solids prepared from ammonia and methyl- amine solutions with R = 0.5 and l are poor conductors. The solids from methylamine solutions with R = 1.25 and 1.5 also showed similar behavior in Spite of the fact that there is some Optical absorption in the I.R. region at 4000 cm-1. The solid Cs-18C6 with R = 2 which shows a "plasma" type absorption in the optical spectra did not Show truly metallic character by the microwave conductivity measurement, but it certainly was different from the rest of the systems. The power absorbed by this sample was of comparable value to that of a low band-gap semiconductor. However, by the method we used we were not able to dis- tinguish between the other electrides prepared from Cs-18C6 with different proportions. 87 ma.a sea mmmao modem AHHSMV ma.H boo oomruoeHSdordwue mm.H moo mmzozmAGOmvao meo.o ooa Amm.oc eaaoflfiamd em.H mom mmzozxmxeo@flcmo «no.0 ema lmo.oc esHoAHHma ma.H med mmZNIooSHcoo o Nmo laasac adaoaafimd ma.H boa mm2\HAOmHvHAmOV amo.o sea 1 av Hoodoo <8 ammuao> oHoEmm me ammuao> oaofimm pouuflamcmue pouuflEmcmue Ho3om mason .mpnmp locum asp npw3 poumofiou Io.mOmH+mO mo mpcofiouommoz hufl>flpospcou o>m3ouow£ .N magma 88 em.o s.boo mmoao Rodeo ma.o e.ooo armamficmloov em.o o.oeo ozmo oa.o mm.eeo Amm.ov asaoaaaoa mm.o a.ooo bamam.axmoc emo.o mm.ooo Amo.oc sshoaaaod mm.o mm.ooe albumacmm.axmoc mo.o m.ooo laasov soaoaaaoa mm.o m.eoe mmzoz\aloOmHVHlmoc meo.e mm.oeo a do poddoo <5 ammuao> onEmm «E ammuao> mamsmm cmggflngMHB muwnwu. HEWCMHB Ho3om II III :I lIlliIIIIIIII. HOBON Il .mcumpcmum any Spas OOHMQEOO mopflwuooam Io.mOmH+mO mo moaomom wpfl>wpoopcoo o>m3OHOHZ .m OHQMB CHAPTER IV ALKALI METAL lB-CROWN-6 ALKALIDES Since the preparation and characterization Of the first crystalline salt (Na+C222-Na-) of an alkali metal anion in 1974 (66,67) the formation Of other alkali metal anions was Observed in the film spectra (75,78). The first syn- thesis (105) and characterization of two sodide salts which use 18-crown—6 is reported here. The synthesis of alkalide salts has been greatly aided by the ability to measure the transmission spectra Of thin solvent-free films produced by rapid solvent evaporation. These spectra permit the identification of the alkali metal anion and also indicate when trapped electrons are likely to be present. Besides that, they provide informa- tion about the stability of alkalides in the presence of the complexing agents. The preparation procedure Of the alkalides has been discussed in the experimental section. The prOperties Of these alkalides will be reviewed in the following sections. 89 90 IV.A. Optical Spectrosquy IV.A.l. Films of CsNa18C6 Figure 24 shows the optical spectrum of a dry film made from a methylamine solution which contained equimolar amounts of cesium, sodium and 18-crown-6. As the Spectrum shows, there is a sharp peak at 16500 cm—1 which corres- ponds to the Na- anion. This peak is slightly shifted compared to the Na- peak (16000 cm-l) produced from the Na 18C6 system with R = metal/crown = 2 (78). The spectrum showed no changes with time or temperature. IV.A.2. Films of K Na 18C6 The spectrum of a dry film made from methylamine solu- tion which contains equimolar amounts of potassium, sodium and 18C6 is shown in Figure 25. There is a broad peak 1 l at 13000 cm- and a Shoulder at 10000 cm- . If the major peak is due to Na- then the presence of K+18C6 causes a red-shift Of 3000 cm-1. On the other hand, if it is due to K-, then the presence Of Na+18C6 causes a blue—shift of 800 cm-1. This contrasts with the behavior of K+C222°Na- films in which the peak shifted only 300 cm”1 from that of Na+C222°Na- (78). The shoulder is attributed to a trapped electron shifted to shorter wavelength probably due to interaction with the cation and anion. The spectrum of K-Na 18C6 films showed no change with time or temperature. 91 .GOAu Inflow ocHEmamsuoE Eouw Imz.00ma+m0 mo SHAH who no Eonuooom HMOflumo .vm anomflm nIO_ ATEUK 0N ON m. 0. m . 0.0 0.. v0 0 .0 N0 0.. ON .53.. 92 MN .cofluoaom ocflsmamnuofi Eoum mOmHmzx mo EHHM >up a mo Esuuoomm HMOflpoo .mm anomfim MIO_ .ATEUYR 0N m. 0. k _ . _ 0.0 I 1 .0 m onc. a. t .. . . . _. . 0.. no we no mo 9 m_ s. .53.: . 93 We expect the major peak to be due to Na- and not K-. This is because the complexation constant of K+18C6 is greater than Na+l8C6 in solution (88) and Na- is generally much more stable than K-. However, the Observed peak might be due to a mixture of K- and Na-. IV.A.3. Films of Cs Rb 18C6 The Spectrum of a dry film made from a methylamine solu- tion which contained equimolar amounts of cesium, rubidium and 18C6 is Shown in Figure 26. There are two peaks, the 1 major peak is at 11500 cm- and it is very broad, extending over a 1000 cm.1 range. The center of the peak is at 1 12000 cm- which corresponds to Rb- observed in the spectrum of Rb 18C6 with R = metal/18C6 = 2 (78). The second peak at 9000 cm’1 assigned to the trapped electron is shifted to higher energies than expected, probably because of inter- action with the anion Rb-. However, the presence of the Cs— anion cannot be excluded since these films showed dif- ferent behavior at different temperatures. When the films are made at lower temperatures the absorption is very broad and probably includes Cs-, Rb- and trapped elec- tron. This system was difficult to characterize and at- tempts to prepare crystals from this system have failed. 94 .coHuSHOm OCHEMH>£uoE Eoum oumanmmu mo Eaflm who M NO Eowuoaom HMOHuQO .mm musmwm nIO_ . A .IE0vh mm 0m 9 o_ b . . a a . 00 0.. no .oo oo 9 od .512 95 IV.B. Sample Analyses The procedure and the analysis scheme have been dis- cussed in the experimental section. Several different samples of the alkalides Cs+18C6-Na- and K+18C6-Na- were analyzed. The samples were analyzed according to the re- action M+18C6 N” + 2H20 + M+ + N+ + 18C6 + on’ + Hzt The number of moles of hydrogen was calculated by using the ideal gas Law and the number of moles of hydroxide ion was determined by titration with standard HCl solution. The number of moles of M+ and N+ were determined by flame emission. The amount of crown ether was determined by 1H NMR signal compared with a known integration of the concentration of potassium hydrogen phthalate (KHP). Table 4 gives the results of analysis Of Cs+18C6-Na- and the percentage deviation from the amount of sample present as determined by hydrogen evolution. As shown from Table 4, the stoichiometry of the compound is C518C6-Na. Since the optical spectrum shows only the peak of Na-, the formula of the compound must be Cs+18C6-Na-. The 5% deviation for OH- is higher than for the metals. This is due to the difficulty in the titration of OH- in the presence of crown ether which causes instability of the pH meter es- pecially near the end point. The pH Of the water distilled 96 Table 4. Results of the Analyses of Cs+18C6oNa-. The Values in the Parenthesis are the Percent Devia- tion From the Predicted Stoichiometry. Sample Moles of Moles of Moles of Moles Of Moles of NO' H2-104 0H’-104 Cs+-104 Na+-104 18C6-104 1 0.471 0.889 0.473 0.483 (-) (-5) (0) (+3) 2 0.197 0.198 (-) (0.5) from the sample after the hydrogen evolution step showed that there was no solvent associated with the crystals. Table 5 shows the results Of the analysis Of K+18C6-Na-. + .— Table 5. Results of the Analyses of K 18C6-Na . The Num- bers in the Parentheses are the Percent Deviation from Predicted Stoichiometry. Sample Moles of Moles of Moles of Moles Of Moles Of No. 142-104 0H'-104 K-lo4 Na-lo4 18C6-104 1 0.858 1.645 0.900 0.813 (-) (—4) (+4.9) (-5) 2 0.171 0.186 (-) (8.8) 97 The analysis Of the water after hydrogen evolution showed that the crystals contained about 16% amine even though they had been vacuum-dried before decomposition. IV.C. Powder dc Conductivity Conductivity measurements were performed on alkalide compounds. This was done first by checking Ohm's Law by reading the current versus voltage at different tempera— tures. Then at a certain voltage the current was measured with changing temperature. The resistance was calculated and a plot of log R y§,% shown in Figures 27 and 28 for Cs+18C6-Na_ and K+18C6-Na- gave a straight line from which the band gap was calculated. A fit of the data to a linear relationship using the non-linear least-squares KINFIT program (103) gave the relation log R = -5.579:0.637 + (4.54l:0.l72) x 103(%), the average band gap calculated is 1.7 eV. The conductivity measurements of K+18C6-Na- yielded the relation log R = -l.397iO.169 + 2.287:0.042 x 103(%) which gives an average band gap Of %.93 eV. As the values of the band gaps indicate, these compounds are semi- conductors. IV.D. Magnetic Susceptibility of Cs+18C6-Na- Crystals of Cs+18C6-Na- are diamagnetic in the tempera- ture range 1.6 and 300 K and there is no evidence for any 98 l2.9 - ' 0.;“(0 ' An 0 l3 0 O O a O O 2.0 - 8° 0° -’ o O o 0: (3O (9 00/ ‘ 3 ”.0" (9° '— I0.0 - ‘ .J J 1 3.5 40 4.4 +(K")-Io3 Figure 27. Plot of log resistivity vs. reciprocal tempera- ture for polycrystalline Cs+18C6-Na'. 99 T I ‘ _ 9. 8.8 y a,“ 07 a 'O 8.5— ’/°D Oo "‘ :3 O D O 8.2- 0° 0 .. D o D O 0: O 79- //0 0° - 0 O O 3 . .° 76- ” o/ - O o 0 o O 7.3- .3 0° 1 O o 0° 3.5 4.0 +(K")-Io~” Figure 28. Plot of log resistance vs. reciprocal tempera- ture for polycrystalline K+18C6-Na'. 100 magnetic ordering. The observed molar magnetic suscep- tibility can be compared to that calculated using Pascal's constants for diamagnetic susceptibility. The summation of the atomic susceptibilities (Pascal's constants) give the molar susceptibility (106). Therefore, the calculated molar susceptibility for Cs+18C6-Na- is XCS+18C6-Na' = XCS+ + X18C6 + XNa' (15) The diamagnetic susceptibility Of Na- can be approximated to be equal to that of Cl- (since the radius Of Cl- is close to that of Na-) according to Langevin equation for diamagnetism which states that the diamagnetic suscep- tibility is prOportional to the mean square distance of the electrons from the nucleus. Then the calculated molar susceptibility value 2.27 x 10-4 emu/mole. This value is in the same range of the experimental value 1.72 x 10-4 emu/mole within the experimental error. IV.B. X-ray Study of Cs+18C6-Na- IV.E.1. Single Crystal Isolation The crystalline compound Cs+18C6°Na- is stable at room temperature in an inert atmosphere. Single crystals were loaded in small thin wall capillary tubes (107) in the argon atmosphere box (dry box). The crystals were loaded with the help of a microscope mounted outside the 101 box. Due to difficulties in adjusting the micrOSCOpe, the crystals which were loaded were not all Of good quality. After loading the crystals in the capillary tubes (0.3- 0.2 mm D), they were sealed inside the box by using resistively heated nichrome wires. After that the crystals were taken out of the box and examined under the micrOSCOpe to assess their quality. However, this technique has its faults. The compound is extremely reactive and the crystals are degraded by any impurities in the inert atmosphere box. Temperature changes may also alter the quality Of the crystals. Problems also have occurred because of the move- ment of the crystal in the capillary. Furthermore, there were many problems with twinned crystals, unfavorable crystal dimensions and irregular shapes. Because Of all these difficulties only a few crystals satisfied all of the necessary conditions for X-ray structure determination and even these gave fewer reflections than would have been desired. IV.E.2. Results of x-ray Study Different procedures were used by Dr. Donald Ward for data collection from three different crystals. All of them gave essentially the same results, but the structure could not be solved. A summary of the information Obtained about the crystalline compound Cs+18C6'Na- will be given. Crystal symmetry: monoclinic. 102 Unit Cell parameters: 13.895 (11) A a = b = 15.501 (24) c = 8.932 (7) B = 93.25 (5) 3 II .15 V (volume) = 1920.7 A , and assuming Z the number of molecules per unit cell, the density D = 1.453 g cm-3. Possible space groups: C2, Cm, C2/m. Patterson map peaks are (no. U, V, W, height) 1. 0.000 0.000 0.000 999 2. .500 .000 .500 147 .361 .500 .652 122 298 .339 .155 105 .420 .332 .697 86 .395 .500 .467 71 .000 .183 .726 67 8. .000 .000 .500 45 9. .260 .040 .000 43 10. .260 .000 .000 41 Remaining peaks are of heights less than 9. The Cs-Cs peaks should be of the type (depending on space group): 0 0 0 0 2y 0 2x 0 22 2x 2y 22 103 and a single-weighted Cs-Cs peak (scaled to the origin = 999) should have a height about 190. The structure making the most "sense" places Cs at: 0, 0, 0; k, k, 0; k, 0, S; and 0, k, A. This is a face-centered arrangement and gives an R-factor of about 55%. (Placing Cs only in the first two positions gives an R-factor of about 25% but is not face-centered and corresponds to only two cesiums per unit cell. The structure could not be extended by repetitions of structure-factor calculations and electron density maps. Direct methods were not successful in solving the structure. We conclude that the microstructure must be more complex than a simple face-centered monoclinic system. CHAPTER V PREPARATION OF ALKALIDES AND ELECTRIDES BY DIRECT VAPOR DEPOSITION The films from solution used to study the optical spectra were always prepared at low temperatures (<—30°C) in order to minimize decomposition. The Spectra of some dry films showed changes with time which suggested "anneal- ing" in the solid state. In addition, dark blue films Often formed when cesium metal was distilled into a vessel which contained 18—crown-6 at room temperature. These observa- tions lead us to investigate the possibility of preparing alkalides and electrides by direct vapor deposition (108). V.A. Experimental The apparatus shown in Figure 29 was used to prepare films of alkalides and electrides by vapor deposition. The crown ether (18C6) and the metal were introduced into the side arms by using Teflon heat-shrinkable tubing. 5 torr, the glass After evacuation to about 1 x 10- ampoule containing the metal was broken, moved to the constriction and a seal was made behind it to remove the heat—shrinkable tubing. The side arms were then 104 105 VAPOR SYNTHESIS APPARATUS . O-Ring Seal To High Vacuum Crown or Cryptand Small Quartz Optical Cell 1 (I .4 <—v Quartz Optical Cell .5117) Figure 29. Apparatus for preparation of films of electrides and alkalides from vapor deposition. 106 electrically heated to the appropriate temperature by adjusting the voltage across the Nichrome wire coils. While the inner cell was above the Side arms, a film was first deposited on the inside of the outer vessel. When an appropriate film was formed (judged by the appearance of the film), the inner cell was lowered so that the Optical part was facing the two side arms. After the film had been formed, the inner cell was lowered down into the Optical portion of the outer cell, the apparatus was closed, removed from the vacuum line and the spectra were recorded with a Beckman DK-2 Spectrophotometer. Except for the films formed from cesium and 18C6, the films were rather unstable and tended to decompose when formed at room temperature. Therefore, it was necessary to cool the inner cell by inserting powdered dry-ice into the cell while forming the film. After that, the dry-ice was removed and the film was allowed to come to room tem- perature and the spectra were recorded. V.B. Results and Discussion The Optical Spectra Of films made from the alkali metals (Na, K, Rb and Cs) with 18C6 are shown in Figure 30. The Spectrum of Cs-18C6 shows a sharp peak at 5250 cm_1 in the IR region attributed to the trapped electron and a very intense peak which was out of scale in the visible region probably due to the Cs- anion. In the spectrum of 107 .ooaowmodoo aodm> uoowao so bamamz II. .a “mamas I.I Am “mumaom .II AN “mumamo AH 00 muuoomm coflmmHEmcmHu Havaumo .om anomflm m.O_ . $.5on _ _ . Lug \\\ .\ \ \ // \\\ \ / \ \ /. \ \ /. \ . \\ t\ / / \ \ / / / / \ \ I.I ll. \\ \ / ll. \\ .\ /. l/ . /. .I/ x .\ DO /. ./l\ \ l .I ./ \ ’./\ \4 “DE . x r 4‘ / \ \. .. / . / \ T’I/ . \ \ // q // /. \ \. / I.I / \ ,,/ / /. \ \ / ~ I / \ . ‘ I .A I . I, \ / .lr.\ b o . . O. #0 m0 108 Rb-18C6 film there is only a major peak at 6500 cm.1 due to the trapped electron. However, this peak is broad with extensive tailing in the visible region. For the films Of Na-18C6 and K-18C6, the peaks are at 17500 cm-1 and 13250 cm.1 probably corresponding to Na- and K-, res- pectively. These peaks were very broad. The spectra of all films made from vapor deposition were compared with the corresponding known spectra Of films made from solution. Blank experiments were also done, in which only films of metals were formed. Their spectra were recorded and com- pared with spectra which have been previously reported (109). The inability to control the stoichiometry and thickness often caused interferences from light scattering, absorption by excess metal and absorbances which were out of range. With the apparatus we used it was impossible to prepare a uniform homogeneous film. Therefore, some films (for example that with Cs-18C6) showed two kinds of peaks, one due to the trapped electron and the second caused by Cs-. The reason for this behavior is the nature of the film formed. The part Of the film which faces the metal side arm is richer in metal and therefore produced Cs-, while the other part which faces the crown ether is richer in the crown and gives the trapped electron. Depending on the metal to crown ratio, it is possible to prepare either electrides or alkalides and this can be accomplished by increasing the rate of evaporation of either the crown or 109 the metal. Although it was not possible by this method to prepare films with controlled thicknesses and stoichiom- etries, these preliminary results clearly show that alka- lides and electrides can be prepared by direct vapor deposi- tion. CHAPTER VI PREPARATION AND CHARACTERIZATION OF A CRYSTALLINE ELECTRIDE Cs+18C6-e- Introduction The major thrust in the research in this laboratory since 1970 has been to prepare and characterize crystalline "electrides" and "alkalides". Successful preparation of the first crystalline alkalide, Na+C222-Na-, came in 1974. Continuous attempts over almost twelve years to prepare crystalline electrides have failed. The extreme reac- tivity and instability of the electrides has made the synthesis of electrides a very difficult and Often frus- trating task. The easiest method for the preparation of electrides is the solvent evaporation discussed in Chapter II. It was clear from the beginning that our major prob- lems with this work were due to decomposition. Even with extreme care in cleaning glassware, purification of com- plexants, metals and solvents, we were able to minimize decomposition but not entirely prevent it. The second major problem with the crystallization procedure is that changing the initial solvent from methylamine to a less polar 110 111 crystallization solvent always resulted in decomposition. Even a gradual change of the polarity Of the solvent did not help. Recently, after I had prepared powders and films of the extremely reactive electride Cs+18C6-e- in a mole ratio of up to two males of metal per mole of crown ether and had shown that it displayed different prOperties com- pared with the other systems discussed in Chapter III, a preparation was tried in which I dissolved cesium, lithium and 18-crown-6 in equimolar amounts in methylamine. Sur- prisingly, the solutions were very stable even at high temperatures (about m-10°C). The stability of films made from this solution was remarkable. It was clear that the lithium played a major role in stabilizing the solution and preventing decomposition. The lithium probably serves as a scavanger for radicals (such as ~CH2NH2) which may enhance decomposition as chain carriers. Cooling the methyl- amine solution to dry ice temperatures did not produce crystals. This was expected because the materials are very soluble in methylamine. Finally, crystallization was made possible by using a co-solvent of isopropylamine and diethylether. (It is possible to Obtain crystals from iso- propylamine solutions alone at low temperatures but the yield is low). However, it is quite possible that the composition Of the solvent mixture might be very important because of problems that might arise from precipitation of lithium metal which would contaminate the crystals and give false 112 analyses. Because of all these problems, a very careful procedure has to be used in order to get metal-free crystals (110). The preparation procedure and characterization and prOperties of this crystalline electride are discussed in the following sections. VI.A. Preparation Procedure The synthesis procedure used here is similar to that used in alkalide synthesis as discussed in Chapter II. The apparatus in this case was constructed Of fused silica instead of borosilicate glass in order to avoid sodium contamination of the lithium solutions by contact with sodium borosilicate glass. In the final and most success- ful preparation the solution was filtered through a coarse frit prior to the crystallization step in order to exclude precipitated lithium. IV.B. Analyses The analysis scheme used for alkalides was also used here. This is based on the reaction of the electride with water according to 1 Cs+18C6-e- + H20 + Cs+ + 18C6 + OH- + EHZI The results of the analyses of three samples taken from 113 .v+. Am+v .m+. .o.o. .HI. «IOmemhm.m oIOHxvmmm.m oIOHxOmNmN «IOonon.m vIOmeN¢N.H vIOmeHmvN m .ou. .I. mIOmemmm.m mIOmemmo.o mIOHxOHom.m mIOwaNom.m m a: 3+. .3 T. mIOHthbm.HN mIOwammm.n mIOonmmm.om mIOonmvo.om mIOHxvao.OH III H mOmH HA m0 ImO mm po>Ho>m mmmz Op .02 mo mmHoz mo mOHoz mo moHoz mo mOHoz mo mOHoz deouoood OHmEmm moHoz .‘lll‘llllll'i. 11.1.10)-.. .II I. [I.I IVl I! .ll Icoumm OH .. illa]. .xuuoEoH£OHoum OOUOHOOMQ Eoum coHumH>oo pcoowmm mop mum womonp muoofioz one .Io.00mH+m0 OOHHHOHOSHU mo mommHmc< any 00 muHOmom .0 OHQOB 114 three different preparations are tabulated in Table 6. The crystals were free of solvent as indicated by the following results: In two of the analyses the water left after the decomposition step was collected by distillation and its pH was checked. The pH value indicated no amine solvent. In the third analysis the crystalline sample was decomposed with D20 and the solution was taken for 1H NMR measurement. The Spectrum showed no evidence of contamina- tion of the crystals by solvent. From the results in Table 6 it is clear that lithium metal is not an essential constituent of the crystals. The percent deviations of the values from those predicted are comparable to the estimated experimental errors of each method. Thus, these results prove that the stoichiometric formula is C518C6. VI. C. Optical Spectra The films studied here were formed as discussed in the Experimental section but in this case the solutions were prepared by distilling methylamine into the optical appara- tus which contained some crystals of C518C6. The Optical spectrum of a solvent-free film made at a temperature of -50°C is shown in Figure 31. As the spectrum indicates there is high absorption in the IR region which peaks at 6000 - l 1 7250 em“ but there is also a shoulder at 10250 cm" . These features are probably due to trapped electrons and 115 0N t0 .0oomI um acme EHHm Io.mOmH+m0 mo muuooom conmHEmcmuu HOOHDQO .Hm ouoon n.0_ ATP—8h om b. o_ b . H _ A I 00 no xosurm no no 0. Que. .51: 116 .OomH+ um EHHm OOHaocca Am uOoomI um EHHM OOHmoccmco III .H .OohmI onoumuooEou um EHHM IO.wOmH+m0 mo mnuooom conmHEmcmuu HOOHuoo MIO_ . ATP—0K . .mm owsowm 117 Cs-. However, this Spectrum does not represent the most stable state because it changes upon annealing. Figure 32 shows the spectra of a dry film made at temperature of -37°C. The spectrum (recorded at a temperature of -50°C) shows a high absorption in the IR region and a peak at 10000 cm-l. As the temperature increases, the films "anneal" and the absorption in the IR region drops and the peak tends to localize to give a maximum at 7250 cm-1. Once the film annealed at this temperature its spectrum stays the same and is not affected by time or temperature change. There- fore, this spectrum represents the most stable state pro— duced in the experiment. The absence of a peak due to Cs- in the annealed film is the strongest evidence for the assignment Cs+18C6-e- rather than Cs+(18C6)2-Cs- for the crystalline material. VI.D. Powder dc Conductivity Powder dc conductivity measurements were made with two samples from different preparations. The Ohm's Law behavior was checked at different temperatures and it was found that Cs+18C6-e- nearly obeys Ohm's Law as shown in Figure 33 although some polarization is evident. The current at various temperatures in the temperature range +10°C to -54°C was measured at constant voltage and con- verted to resistance. Figure 34 is a plot of log R against reciprocal temperature. A non-linear least-squares fit 118 I I I 1 1 2.2- - l.8- - b _ n I.4- - .9 - - .'_. Io- . 0.6- - 0.2- - 1 I J l I 2 4 6 8 IO V Figure 33. Ohm's Law plot of polycrystalline powder of + .. Cs 18C6-e . 119 A A A 9" A0 A0 ° £:,(D A 86 006 a: 8 A 01 " 0° 0 A .J 38 l 7 31.6 470 4.4 4.6 7}: '03 K’1 Figure 34. Plot of log resistance vs. reciprocal tempera- ture for a polycrystalline Cs+l8C6-e-. 120 of the line yields log R = l.850:0.076 + (l.611r0.019) x 103 %. The average value of the band gap calculated from this result is m0.60 eV in this temperature range. These powder conductivity measurements suggest that the crystal- line Cs+18C6-e- is a semiconductor. VI.E. EPR Study Both a polycrystalline sample and a Single crystal were used in the EPR study. It was found that the spectra show two peaks and the resolution of these peaks depends on the orientation of the crystal with respect to the magnetic field. Figure 35 shows a series of spectra taken at (ar- bitrary) different orientations. It is clear from these spectra that there are two peaks. Measurements Of the g- values showed that one of them is essentially isotrOpic with g = 2.0023 while the other is anisotropic. The effect of temperature on the peaks was also studied. Figure 36 shows the spectra for a single crystal where the two peaks almost resolved at a temperature of m-100°C. As shown in the figure, the isotropic peak was not affected by tem- perature, while the second anisotropic peak was greatly affected. At lower temperatures the anisotropic peak tends to shift downfield (higher g-values) and the linewidth be- comes smaller, while at high temperatures this peak gets broader, decreases in amplitude and vanishes at 51°C. It is not possible experimentally to separate the peaks 121 O .l- Figure 35. EPR spectra of a single crystal Of Cs+18C6-e- at different (arbitrary) orientations. 122 mHmch mo muuoomm mom any so ououmuomfiou mo uoommm .mm ousmHm .AIw.waH+m0. Hmum>uo v- .I- 0 . Uomum+. 123 completely. It was possible, however, to measure the g- values by using DPPH as calibrating agent. The g-values indicate that the isotropic peak has a g-value of 2.0023, the free electron value. This suggests no appreciable spin- Orbit interaction with the cesium cation. The second an- isotropic peak has a smaller g-value indicative of some interaction. This value is temperature-dependent, it varied from 2.0015 to 2.0020 and also depends upon the orientation of the crystal. It was difficult to measure the g-value at high temperatures because this peak is very broad suggesting a short relaxation time. An approximate calculation of the number of spins was done by using a Spin standard Ruby crystal. This was accomplished by orienting the crystals so that the two peaks almost overlap which could only be done at low temperatures. The calculations show the presence of less than 1% free Spins (in the tem- perature range 128 K to 240 K). VI.F. Magnetic Susceptibility Study The magnetic susceptibility of crystalline Cs+18C6-e- was measured in the temperature range 1.7 K to 300 K. Figure 37 shows the magnetic behavior yg. temperature for the crystalline Cs+l8C6-e- electride. This is a plot of the susceptibility of the electride minus that of the de- composed sample. It is seen that the electronic suscep- tibility is diamagnetic in the range 130 K to 300 K. Below 124 .Io.00mH+mU OOHHHmumwho ImHoo How auouauomfiou umchom muHHHQHuooomOm HOHOE mo uon .bm anomHm 3...? 00m 00m 00m 0..». 0m ON 4 .4l _ _ _ _ AV I O o o o IONI 0 O I O O .IAH_I .I lllllllllllllll mull IIOIQIIO llllllllll 0.0 O O o I o o 1 N O X I O 4.? W I O .me r 10 . o 1.9 T 0 MON 0 I 0.? p — p p h — 125 130 K the magnetic susceptibility is paramagnetic and is temperature dependent down to 1.7 K. The fact that Cs+18C6-e- is nearly diamagnetic rather than strongly paramagnetic is attributed to antiferromagnetic ordering of the electrons in the lower Hubbard band. Cyrot (111) and others pointed out that when electron correlations are taken into account, the singlet ground state is favored over the paramagnetic state. CHAPTER VII CONCLUSIONS AND SUGGESTIONS FOR FUTURE WORK VII.A. Conclusions Films and/or powders Of Cs-18C6 electrides prepared by solvent evaporation showed properties which are dependent on the metal to crown ratio. The prOperties range from nonmetallic when R = 1 to metallic when R = 2. Two stable crystalline alkalides, Cs+18C6-Na_ and K+18C6°Na- were prepared and their properties were in- vestigated. Analyses and the properties Of these compounds confirmed the proposed formula. Films of alkalides and/or electrides were prepared by direct reaction Of the alkali metals and l8-crown-6 deposit- ed from the vapor phase. The optical spectra of these films guide their classification. The first crystalline electride, Cs+18C6°e-, was syn- thesized and its properties were investigated. These crystals were stable at room temperature, and have a melting point Of m65°C. Analyses of different samples confirmed the stoichiometry C518C6. Such properties as Optical spectra and powder d.c. conductivities support the formula CS+18C6'e-. Single crystal EPR spectra showed 126 127 two lines, one of which is isotropic and the other an- isotropic. Magnetic susceptibilities and EPR spectra in- dicated a substantial spin-pairing. VII.B. Suggestion for Future Work (1) After the success in preparing the first crystalline electride, Cs+18C6'e-, further investigation of its prOper- ties such as quantitative single crystal EPR spectra and single crystal conductivity should be made. X-ray struc- ture determination is absolutely necessary for a better understanding of the properties of electrides. Therefore, systematic procedure for growing high-quality single crystals should be develOped. The synthesis Of other electrides which use 18C6 and cryptant C222 might be possible if the stabilization by lithium is a general phenomenon and should be investigated. (2) Work should continue to solve the crystal struc- ture Of the stable alkalide Cs+18C6'Na—. This will probably require crystals of better quality. Other properties such as the single crystal conductivity and photoconductivity should also be investigated. Quantitative work should be done to understand the nature Of the system K+18C6-Na-. Other alkalides such as Rb+18C6'Na-, Cs+(18C6)2-Na-, etc., might be feasible to synthesize. If so, their properties should be investigated in order to establish trends in the properties of such sodides. 128 (3) Quantitative optical and electrical studies should be done on films prepared by vapor deposition using Bell- Jar high-vacuum techniques. (4) The synthesis of mixed alkalides and electrides M+C-N;-e1_X would provide an ideal opportunity to study the influence of electron density on electron-electron inter- actions and should also be explored. BIBLIOGRAPHY 10. ll. 12. 13. 14. 15. 16. BIBLIOGRAPHY J. C. Thompson, "Electrons in Liquid Ammonia", Oxford University Press, Oxford, 1976. G. LePoutre and M. J. Sienko (eds.), "Metal-Ammonia Solutions, Colloque Wey 1", W. A. Benjamin (New York) 1964. J. J. Lagowski and M. J. Sienko (eds.), "Metal-Am- monia Solutions", Butterworth (London) 1970. J. Jortner and N. R. Kestner (eds.), "Electrons in Fluids", Springer-Verlag (New York) 1973. J. Phys. Chem. 12 (26) 1975. J. Phys. Chem. 84 (10) 1980. G. LePoutre and J. P. LeLieur, Reference 3, p. 247. R; A. Ogg, Jr., J. Am. Chem. Soc., 68, 155 (1946), J. Chem. Phys. 14, 114, 295 (1946); Phys. Rev., 69, 243, 668 (1946). J. Jortner, J. Chem. Phys., 30, 839 (1959). D. A. Copeland, N. R. Kestner and J. Jortner, J. Chem. Phys., 53, 1189 (1970). N. R. Kestner and J. Jortner, J. Chem. Phys., 11, 1040 (1973). M. Gold, W. L. Jolly and K. S. Pitzer, J. Am. Chem. Soc., 84, 2264 (1962). J. L. Dye, Reference 3, p. 1. J. V. Acrovis and K. S. Pitzer, J. Phys. Chem., 66, 1693 (1962). J. P. LeLieur, P. DaMay, and G. LePoutre, J. Phys. Chem., 19, 26 (1975). E. Huster, Ann. Phys., 33, 477 (1938). 129 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 130 S. Freed and N. Sugarman, J. Chem. Phys., 11, 354 (1943). C. A. Hutchison, Jr. and R. C. Pastor, Rev. Mod. Phys., 23, 285 (1953); J. Chem. Phys., 31, 7959 (1953). A. Demortier and G. LePoutre, C. R. Acad. Sci. Paris, 2 8, 453 (1969). A. Demortier, M. DeBacker, and G. LePoutre, J. Chem. Phys., 62, 380 (1972). F. Wooten, "Optical Properties of Solids", Academic Press (New York) 1972. T. A. Beckman and K. S. Pitzer, J. Phys. Chem., 66, 1527 (1961). J. A. Vanderhoff, E. W. LeMaster, W. H. McKnight, J. C. Thompson and P. R. Antoniewicz, Phys. Rev. 66, 427 (1971). R. Catterall, J. Chem. Phys. 66, 2262 (1965). F. J. Dyson, Phys. Rev. 26, 349 (1955). A. M. Stacy and M. J. Sienko, J. Chem. Phys, in press. W. S. Glaunsinger and M. J. Sienko, J. Chem. Phys., 62, 1883 (1975). P. P. Edwards, A. R. Lusis and M. J. Sienko, J. Chem. Phys., 12, 3103 (1980). J. R. Buntaine, M. J. Sienko and P. P. Edwards, J. Phys. Chem., 66, 1230 (1980). P. P. Edwards, J.R. Buntaine and M. J. Sienko, Phys. Rev., B19, 5835 (1979). W. S. Glaunsinger, S. Zolotov and M. J. Sienko, J. Chem. Phys. 66, 4756 (1972). W. S. Glaunsinger, T. R. White, R. B. VonDreele, D. A. Gordon, R. F. Marzke, A. L. Bowman, and J. L. Yarnell, Nature, 271, 414 (1978). H. Oesterricher, N. Mammano and M. J. Sienko, J. Solid State Chem. 1, 10 (1969). R. B. VonDreele, W. S. Glaunsinger, A. L. Bownman and J. L. Yarnell, J. Phys. Chem., 79, 2992 (1975). 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 131 J. E. Hart and W. J. Boag, J. Am. Chem. Soc., 66, 4090 (1962). L. Kevan and B. C. Webser (eds.), "Electron-Solvent and Anion-Solvent Interactions", Elsevier, Amsterdam (1976). L. Kevan, J. Phys. Chem., 66, 838 (1972). J. H. Baxendale and P. H. G. Sharpe, Chem. Phys. Lett., 66, 401 (1976). T. Shida, S. Iwata and T. Watanabe, J. Phys. Chem., 76, 3683 (1972). L. Kevan, J. Phys. Chem., 66, 1232 (1980). J. E. Willard, J. Phys. Chem., 26, 2966 (1975). A. Ekstrom and J. E. Willard, J. Phys. Chem., 26, 4599 (1968). L. Kevan, Acc. Chem. Res., 14, 138 (1981). E. E. Budzinski, W. R. Potter, G. Potienko and H. C. Box, J. Chem. Phys., 16, 5040 (1979). H. C. Box and H. G. Freund, Appl. Spectroscopy, 66, 293 (1980). D. F. Feng and L. Kevan, Chem. Rev. 66, l (1980). J. Jortner, Mol. Phys., 6, 257 (1962). D. Copeland, N. R. Kestner and J. Jortner, J. Chem. Phys., 66, 1189 (1970). K. Fueki, D. F. Feng and L. Kevan, J. Phys. Chem., 16, 1976 (1970); J. Am. Chem. Soc., 66, 1398 (1973). T. Ichikawa and H. Yoshida, J. Chem. Phys., 16, 1540 (1980). A. Banerjee and J. Simons, J. Chem. Phys., 66, 415 (1978). J. J. Markham, "F-Centers in Alkali Halides", Academic Press (New York) (1966). R. S. Alger, "Electron Paramagnetic Resonance; Tech- niques and Applications", John Wiley and Sons, Inc. (New York) (1968). 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 132 W. Schmitt and U. Schindewolf, Ber. Bunsenges, Phys. Chem., 66, 584 (1977). N. F. Mott, Proc. Phys. Soc., 666, 416 (1949); Phil. Mag. 6, 287 (1961); Phil. Mag. 19, 835 (1969); and "Metal-Insulator Transitions", TEylor and Francis (London) (1974). P. P. Edwards and M. J. Sienko, Phys. Rev. B17, 2575 (1978). J. Hubbard, Proc. R. Soc., A276, 238 (1963); A277, 237 (1964); and A281, 401 (1964). P. W. Anderson, Phys. Rev., 109, 1492 (1958) and Comments Sol. St. Phys. 6, 190 (1970). R. R. DeWald and J. L. Dye, J. Phys. Chem., 66, 121 (1964). S. Matalon, S. Golden and M. Ottolenghi, J. Phys. Chem., 16, 3098 (1969). M. G. DeBacker and J. L. Dye, J. Phys. Chem., 16, 3092 (1971). T. R. Tuttle, Jr., Chem. Phys. Lett., 66, 371 (1973). J. L. Dye, M. G. DeBacker, J. A. Eyre and L. M. Dorfman, J. Phys. Chem., 16, 839 (1972). J. M. Ceraso and J. L. Dye, J. Chem. Phys., 66, 1585 (1974). J. L. Dye, C. W. Andrews and J. M. Ceraso, J. Phys. Chem., 16, 3076 (1975). J. L. Dye, J. M. Ceraso, M. T. Lok, B. L. Barnett and F. J. Tehan, J. Am. Chem. Soc., 66, 608 (1974). F. J. Tehan, B. L. Barnett and J. L. Dye, J. Am. Chem. Soc., 66, 7203 (1974). C. J. Pedersen, J. Am. Chem. Soc., 66, 7017 (1967); 66, 3299 (1968). B. Dietrich, J.-M. Lehn and J.-P. Sauvage, Tetrahedron Lett., 2885, 2889 (1969). J. L. Dye, M. G. DeBacker and V. A. Nicely, J. Am. Chem. Soc., 66, 5226 (1970). 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 133 M. T. Lok, F. J. Tehan, and J. L. Dye, J. Phys. Chem., 66, 2975 (1972). J. L. Dye, M.-T. Lok, F. J. Tehan, R. B. Coolen, N. Papadakis, J. M. Ceraso and M. G. DeBacker, Ber. Bunsengens. Phys. Chem., 16, 659 (1971). J. L. Dye, C. W. Andrews and S. E. Matthews, J. Phys. Chem., 16, 3076 (1975). J. M. Ceraso, Ph.D. Dissertation, Michigan State Uni— versity, 1975. J. L. Dye, M. R. Yemen, M. G. DaGue and J.-M. Lehn, J. Chem. Phys., 66, 1665 (1978). M. G. DeGue, Ph.D.Dissertation, Michigan State Uni- versity, 1979. M. G. DaGue, J. S. Landers, H. L. Lewis and J. L. Dye, Chem. Phys. Lett., 66, 169 (1979). J. L. Dye, M. G. DaGue, M. R. Yemen, J. S. Landers and H. L. Lewis, J. Phys. Chem., 66, 1096 (1980). J. S. Landers, Ph.D. Dissertation, Michigan State University, 1981. J. S. Landers, J. L. Dye, A. Stacy and M. J. Sienko, J. Phys. Chem., 66, 1096 (1981). C. J. Pedersen and H. K. Frensdorff, Angew. Chem. Int. Ed. Engl., 6, 16 (1972). R. M. Izatt, D. P. Nelson, J. H. Rytting, B. L. Haymore and J. J. Christensen, J. Am. Chem. Soc., 66, 1619 (1971). . Christensen, D. J. Eatough and R. M. Izatt, . Rev., 16, 351 (1974). . M. Izatt, R. E. Terry, B. L. Haymore, L. D. Hansen, . Dalley, A. G. Avonded and J. J. Christensen, . Am. Chem. Soc., 66, 7620 (1976). . M. Izatt, R. E. Terry, D. P. Nelson, Y. Chan, D. J. Eatough, J. S. Bradshaw, L. D. Hansen and J. J. Christensen, J. Am. Chem. Soc., 66, 7626 (1976). E. Mei, J. L. Dye and A. I. Popov, J. Am. Chem. Soc., 66, 1619 (1976). 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 134 E. Mei, J. L. Dye and A. I. Popov, J. Am. Chem. Soc., 66, 6532 (1977). J. D. Lamb, R. M. Izatt, C. S. Swain and J. J. Chris- tensen, J. Am. Chem. Soc., 102, 475 (1980). C. J. Pedersen, Fed. Proc., 27, 1305 (1968). H. K. Frensdorff, J. Am. Chem. Soc., 66, 600 (1971). E. Mei, A. I. Popov and J. L. Dye, J. Phys. Chem., 66, 1677 (1977). Sadegh Khazalei, Ph.D. Dissertation, Michigan State University, 1982. G. W. Gokel, D. J. Cram, C. L. Liotta, H. P. Harris and F. L. Cook, J. Org. Chem., 66, 2445 (1974). M. G. DeBacker and J. L. Dye, J. Phys. Chem., 16, 3092 (1971). Bradley VanEck, Ph.D. Dissertation, Michigan State University, 1982. L. H. Feldman, R. R. DeWald and J. L. Dye, Adv. Chem. Ser., 66, 163 (1965). I. Hurley, T. R. Tuttle, Jr. and S. Golden, J. Chem. Phys. 66, 2818 (1968). M. R. Yemen and J. L. Dye, private communication. J. E. Wertz and J. R. Bolton "Electron Spin Reson- ance: Elementary Theory and Practical Applications", McGraw-Hill (New York) 1972. T. Chang and A. H. Kahn, National Bureau of Standards Special Publication 260-59: Standard Reference Ma- terials: Electron Paramagnetic Resonance Intensity Standard: SRM-2601; Description and Use," U.S. De- partment of Commerce (Washington, D.C.) 1978. A. H. Morrish "The Physical Principles Of Magnetism" John Wiley and Sons, Inc. (New York) 1965. C. Kittle, "Introduction to Solid State Physics", 5th ed., John Wiley & Sons, Inc. (New York) 1976. J. L. Dye and V. A. Nicely, J. Chem. Educ., 66, 443 (1971). 104. 105. 106. 107. 108. 109. 110. 111. 135 M. T. Lok, Ph.D. Dissertation, Michigan State Uni- versity, 1973. B. VanEck, L. D. Le, D. Issa and J. L. Dye, Inorg. Chem. to be published (1982). L. N. Mulay, Ch. 7 in A. Weissberger and B. W. Rossi- ter, "Physical Methods of Chemistry", Vol. 1, John Wiley & Sons (New York) 1972. Thin Wall Capillary Tubes, from "Charles Supper CO. Inc. (15 Tech. Circle, Natick, Mass. 01760). L. D. Lee, D. Issa, B. VanEck and J. L. Dye, J. Phys. Chem., 66, 7 (1982). R. Pyan, Ann. Phys. (Paris), 6, 543 (1969). D. Issa and J. L. Dye, in press. M. Cyrot, J. de physique (Paris), 66, 125 (1972). M "6 “7 "4 I". ”9 8 llhAI R II6 E H V ”o m ili3 U "0 ill 1293 lllllllllllllllllll