MSU LIBRARIES m \' RETURNING MATERIALS: PIace in book drop to remove this checkout from your record. FINES wiII be charged if book is returned after the date stamped below. SYNTHESIS AND GIARACTERIZATION OF NEW ALKALIDES AND WES VIA 'IHE TERTIARY AMINE WING AGENTS: STEPS IMRDS THERMAL STABILITY by Mark E . Kuchenmeister A Dissertation Submitted to Michigan State University in partial fulfillment of the requirements for the degree of WOFPI-IILOSOPHY Department of Chanistry 1989 .7) 4}? :lkr ‘35 fig. SYNTHESIS AND WTION OF NEW ALKALIDES AND WIDE VIA THE TWIARY AMINE MIEXING mm: STEPS ms m STABILITY Mark E . Kuchenmeister Altrnaghammber of salts with the alkali metal anions (alloalides) or trapped previously all electrons (electrides) are lawn, were thermally unstable to irreversible decomposition of the cation complexant (crown ether or cryptand) . By using the fully methylated nitrogen analogs of the crown ethers, five crystalline sodides and one crystalline electride have been synthesized. The ranarlable feature of the aza—crowns is their extrane resistance to reductive deoanposition. Differmtial scanning calorimetry (080) of the three sodides that cautain PMICY (heiamethyl hexacyclen) shows that each compound melts witl'nut decomposition at temperatures ranging from 7 to 40 ‘C and decanplexes to yield liquid MO! and the alkali metals at tenaperatures about 30 degrees above the meltirg point. For all three compounds. irreversible decomposition does not occur until ~14O ‘C. The crystal structures for the three sodides with I-MiCY have been determined, and they are isostructural in the orthorhanbic space group P212121. In each case, the structure shows that Na- is very nearly in cmmtact with one alkali metal cation. 'Ihe solid my be viewed as containing closest-packed contact ion-pairs in michthecatimisanbeddedinthecanplacantmt is exposed on one face to the sodide anion. The optical spectra of thin films and the 23Na MAS-MR spectrum confirmed the presence of Na.. M additional compounds have been synthesized with PMPCY (pentamethyl pentacyclen) , one sodide and one electride. Both oanpounds are ranarlably stable at roan tanperature for over 10 days, but eventually decanplex into PMPCY and the alkali metals. The DSC spectrum of each canpound indicates a phase transition at ~-54 and ~-74 ‘C for the sodide and electride respectively. ‘mis transition is also detected for the electride in magnetic susceptibility and conductivity mamts with very different behaviors above and below the transition. All of these canpamds with the aza—croms are very soluble in dimethyl ether and sanewhat soluble in diethyl ether and trimethylamine. Solutions in dimethyl ether are renarkably stable. These studies point to a strategy for synthesizing thermally stable allalides and electrides in which tertiary amine cyclic or bicyclic oanplexants are used to enhance stability. The limiting factor is weak canpleiation that tends to yield the free mine and alkali metals at elevated tenperatures. to my Parents and Claudia iv I would like to express my gratitude to Professor Dye for his constant support and guidance throughout this work. I would also would like to thank Lauren Hill and Jinem Kim for helping with the MR spectra, and armed Ellaboudy for helping with the EPR spectra. Special thanks goes to Dr. ward and Rui Huang for their assistance in obtaining the crystal structures and for the informative discussions I had with Dr. Ward. I would also like to thank Professor Far-mm and his group for the an—complexants they provided. The technical and clerical staff deserve thanks also, especially the glassblowers Keki Mistry, Manfred Langer, and Scott Bancroft for their outstanding service. Research sumort from NSF Grants DIR 87-14751 is gratefully acerawleged . WWW Page List of Tables ...................................... ix List of Figures ...................................... x I INTRODUCTION ........................................... 1 I. A. History of Canplexing Agents ...................... 1 I. B. Alkalides and Electrides .......................... 4 I. C. Current Work ...................................... 8 II. SYNTHESIS AND ANALYSIS ................................. 10 II. A. Vacuum Lines and Inert Atmosphere Techniques ..... . 10 II. B. Reagents .......................................... 11 II. B. 1. Complexing Agents ........................... 11 II. B. 2. mtals ...................................... 13 II. C. Synthesis ......................................... 13 II. D. Analysis ..................................... 15 II. D. 1. Hydrogen Evolution .......................... 16 II. D. 2. pH Titration ................................ 17 II. D. 3. Plasmaanission ......... 18 II. D. 4. 1am ...... . ........ 18 II. E. Results.............. .............. .. ....... 19 II. E. 1. PMCY Compounds ............................. 19 II. E. 2. PMPCY Compounds ............................. 20 vi II. E. 3. TRIMJY Canpourxis ............................ 20 II. E. 4. m Compounds ............................ 21 II. F. Summary ........................................... 21 III OPTICALSPmSGJPY ................................... 23 III. A. Introduction ...................................... 23 III. B. Dcperimental ...................................... 25 III. B. 1. Compamds with HMHCY ........................ 27 III. B. 2. Compounds with my ........................ 32 III. B. 3. L1+(m)2.Na' ............................ 36 III. c. Sumnary ........................................... 36 IV. WDYNAMIC PROPERTIES ............................... 39 IV. A. Introduction ...................................... 39 IV. B. Experimental Methods .............................. 40 IV. C. Thermodynamic Properties .......................... 42 IV. c. 1. may Compounds .............................. 42 IV. c. 2. PMPCY Compounds .............................. 56 IV. c. 2. Li+(m)2.ua’ ............................. so IV. D. Sunmary ........................................... 60 V. MAGNETIC PROPERTIES .................................... 62 V. A MAS-MGR ........................................... 62 v. A. 1. I'M-{CY Compounds . ............................. 63 v. A. 2. m Compounds ................. . ............ 64 V B EPR ......... .5” . ........................ 64 V. B. 1. Alkalides and Electrides . .................... 66 V. B. 2. Results .................................. 67 V. C. mgnetic susceptibility ........................... 70 vii v c 1 Li (maybe ................................ 72 v. c Sumnary ........................................... 76 VI. x-RAY s'mwmms .............. . ....... . ................ 73 VI. A. Introduction ...................................... 78 VI. B. Experimental ......... . .................. . ........ . 79 VI. C. X—Fay Data Collections ............................ 80 VI. c. 1. K+(HMI-ICY).Na- ................................ so VI. 0. 2. Cs+(I-1MHCY)-Na- ............................... 83 VI. 0. 3. Rb+(deCY)oNa- ............................... 89 VI. D. Discussion ........................................ 86 VI. E. Canpounds with PMPCY .............................. 93 VII. CONCLUSIONS ............................................ 95 VII A Designing new Complexants ......................... 96 APPENDIX A ............................................. 98 REFERENCES ............................................. 107 viii LISIWTABLES TABLE Page 1. Smmary of Purification Conditions For the Various Aza-Crowns ............................................ 12 2. Summary of Alkalides and Electrides Synthesized and Characterized in This Study ....................... 22 3. Absorption Maxima Position For Na- in Various Sodide Films ......................................... 24 4. Smunary of Optical Absorption Peak Positions ......... 38 5. Values Used to Obtain the Enthalpy of Formation of the Metal Alloys .................................. 49 6. Sunlnary of Enthalpy Charges for Various Steps in the Thermodynamic Cycle Shown in Figure 14 ........... 50 7. Stmmaryofthe'l'hermalProcesses'matmcurinthe DSC Traces With a Heating Rate of 5 deg/min in ‘C 61 + _ 8. Summary of Cell Constants f r K (I-M-ICY).Na (I), Rb (I-MHCYbNa (II), and 039(IMHCY).Na (III) ........ 88 9. ammry of Atanic Distances With Respect to the Plane in Figure 26 ................................... 90 A1. Positional Parameters and Their Esrimated Standard Deviations for K (WbNa (I) ..................... 99 A2. Positional Parameters and Their Estimated Standard Deviations for Rb (mph (II) ................... 100 A3. Positional Parame ers ard Their Estimated Standard Deviations for Cs (3010‘!th (III) .................. 101 A4. Bqd Distances (in Angstans) for {(1811}??th (I), Rb (WhNa (II), and Cs (momma (III) ........ 102 A5. Bond Angles (in Degrees) for K+(I-m-ICY):_Na- (I). Rb (HIGH-Na (II), and Cs (momma (III) ........ 104 LISTG’FI“ FIGURE Page 10. 11. 12. 13. A) 1,4,7,10,13,16,-He)aoocacyclooctadecane [18-Crown—6]; B) 1,4,13,16,21,24-Hexaoxa-1,10-Diazabicyclo [8.8.8] Hexacosane [Cryptand 2,2,2]; C) 1,4,7,10,13,16—Hexaaza- 1,4,7,10,13,16—Hexamethylcyclooctadecane [HMICY] ....... 2 Mechanism for Crown Ether Decanposition. A) By hydrogen atcm renoval. B) By proton ranoval ............ 6 Vacuum apparatus used in the synthesis of alkalides and electrides [K—Cell] ................................ 14 Apparatus used to obtain optical spectra ............... 26 Absorption Spectrum of KI(HMHCV).Ns‘ at -120 '0 ........ 28 Absorption Spectrum of Rb+(l-WCY)oNa- at -100 “C ........ 29 Absorption Spectrum of Cs+(HMHCY)-Na- at -107 '0 ........ 30 Absorption Spectrum of Ba2+(I-M-ICY)o (Naxe2_x)2— at -117 °C ............................................. 33 Absorption Spectra of A) Li+(PIPCY)-Na- at -80 ‘C; and B) Li (PMPCY)oe at -90 ’C ......................... 34 Absorption Spectrum of Li+(TRIMZY)2-Na- at -114 'C ..... 37 Differential scanning calorimetry trace for a 3.3 mg sample of K (EM-ICYhNa at 10 deg/min. Endothermic values are negative .................................... 43 Differential +scanning calorimetry trace for a 2.5 ng sample of Cs (MOSH-Na at 5 deg/min. Endothermic values are negative .................................... 46 Differential +scanning calorimetry trace for a 2. 2 119 sample of Rh (WhNa at 5 deg/min. Endothermic values are negative .................................... 47 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. Thermodynamic le used to calculate enthalpies of formation of I_( (WbNa , Rb (BIRCH-Na , and Cs (Wbfla from the metals and complexant at 225 K from the DSC measurenents ........................ Thennodynanic Cycles for Rb+(M!CY).Na-; A) Processes occuring during decanpleation over a large tenperature range. B) Decanplexation process at 300 K. Enthalpy values are in kJ mol ................................. Thermodynamic cyCles for 05+(MICY).Na—; A) Processes occuring during deoanplexation over a large tanperature range. B) Deoanplexation process at 300 K. Enthalpy values are in kJ mol ................................. nsc spectrum for A) L1+(I=MPCY).Na'; and B) Li (PMPCY)oe ; at 5 deg/min. Endothermic values are negative .................................... DSC trace over the phase transition in L1 (PMPCY)-Na . A) Heating; B) Cooling ................. DSC spectrum of Li+(TRIM3Y) oNa- at 5 deg/min. Endothermic values are nega ive ........................ EPR spectra for the three sodides K+(1MHCY).Na_; Rb (WhNa ; C) Cs (PMCYhNa . Taken with an X-band spectrometer .................................... Plot of the molar susceptibility of Li+(PIVPCY).e- as a function of temperature in a 3 k6 magnetic field ..... Plot of 1/x as a function of tanperature. There are two noticeably different behaviors above and below the transition tenperature at ~190 K ................... The molecular structure and the ring of_the stars; and B) stereopacking diagram of K (EM-ICYhNa .......... The molecular structure and the bering of the atoms; and B) stereopacking diagram of Cs (HMHCYbNa ......... The molecular structure and the mimbering of the atoms; arri B) stereopacking diagram of Rb (W)oNa ......... Diagram of plane through the Nitrogens 1,7,10,16. The axis are only approximate ........... . .............. Canputer simulation of one "molecule" of Cs+(IMHCY)oNa-. Generated on an Evans +and Sutherland P3350 system with the program FRODO. [K (HM-ICY).Na and Rb (HMICYhNa would appear similar ................................... xi 50 57 59 59 68 73 75 82 87 90 91 28. 'mo cutaway views of the cation-anion packing in K+ (HM-1C3!)- Na .A) View down the y-axis; B) View down the x-axis. Distances are in Angstans. The Na is not drawn to scale. The structures obe Rb-(IM'ICY)0N3 and Csu-IMHCY) Na are similar ............................. xii 94 1mm Alkalides and electrides are a novel class of canpounds first synthesized in 1974.1'2 An alkalide is an ionic solid in which the cation is an alkali metal encased in 1 or 2 organic macrocyclic or macrobicyclic ligani(s). The anion is than an alkali metal in the -1 oxidation state. An electride is also an imic solid with the same or a similar cation, but the anion is a solvent-free "trapped" electron. Electrides may be thought of as "stoichianetric F-centers". Alkali metal anions in solution have been hum since 1969 and solvated electrons since the work of Kraus in 1908. Alkali metal solutions have been studied since Weyl's work with sodium and potassium in amnonia in 1864. The solid state counterparts were not possible until the developnent of suitable cyclic or bicyclic polyether complexants. The role of the ligand is to stabilize the alkalide or electride by complexing the alkali cation to replace the salvation sheath present in solutions. I. A. Histogy g; M19293 m: In 1967, Pederson first reported his revolutionary synthesis of 33 cyclic polyethers ["crovm ethers", Fig. 1A] and the observation that many of those containing five to ten oxygai atone formed surprisimly stable complexes with the alkali and alkaline 3,4 earth metal cations He also stowed that there was a certain amamt of selectivity between the cram compounds and the cations m f° W (Cw K/O OJ \ao OJ L_/ \_/ 18 Crown~6 Crypund-2.2.2 'A B Me Ma \r-\ / Figure 1. A) 1,4,7,10,13,16-I-Iexaoxacyclooctadecane [is-Cram-GJ; B) 1,4,13,16,21,24-Hexama-1,10—Diazabicyclo [8.8.8] Beacosane [Cryptand 2,2,2]; C) 1,4,7,10,13,16-Bexaaza-1,4,7,10,13,16- Hamlethylcyclooctadem [W]. based (:1 the relative sizes of the holes and tta cation diameters. It was later observed that when tla sizes did not match, 2:1 and even 3:1 canplexes of cram to metal could be formeds. With the danonstrated sucwss of the macrocyclic polyettars as canplexants, Lehn went one step further and synthesized nacrobicyclic canpounds ["cryptarris", Fig. 18], which contairad two bridging N(01{2)3 groups. Time "crypts" fornad a cage for tta cation; thereby increasing stability and selectivityG—s. Rigid ligands, generally those with small cavities and short nonflexible chains, displayed peak selectivities while the more flexible ligands had plateau selectivities]. Cram used the concept of rigid ligands to furtl'ar improve on the selectivity by introducing additional binding sites or steric barriers at strategic positions on the crown etrars ["sptarands'flg-n. These compounds are macrocyclic or nacrobicyclic systems that contain preformed cavities fully organized for complexation during synthesis, rather than durirg the canplexation process. Since tra initial work by Pederson, many researchers rave studied the synthesis, complexation, kinetics, reactivity and interactions of nacrocyclic ligands with alkali and alkaline earth natal cations. The interest from tra syntratic commity Ias broadened the range of available nacrocyclic an! nacrobicyclic compounds. “areas the first complexants had only oxygen donors, there lave since been synthesized naterials with nitrogen, sulfur and phosphorous substituted for sane or all of the oxygen atoms. The revelation that cram etl'ars and related canpounds form stable complexes with nany different cations, and also demonstrate selectivity, has pranpted mearcl'ars to explore tta possible applications of these new: nterials. Sane of tta processes investigated include: tra transport of has thragh 12,13 14,15 menbranes , tra construction of ion selective electrodes 16'”, and the increase in solubility of alkali 18,19 isotope separation natals in a wide variety of solvents I. B. Alkalides & Electrides: It was while working with the concentrated solutions (~O.4 M) of sodium an! tra cryptand 0222 in ethylamira, trat Dye and canorkers were able to form a solid gold—colored precipitate by 1'2. Based on cooling tta solution to Dry Ice tenperatures stoichianetric analysis, the crystallira material was identified as having the formula Na+Lo Na', where Na+L refers to the sodium- cryptand canplex. This identification was verified by X—ray crystallography giving tIa first structure of a salt of an alkali metal anion; called an Alkalide. Ellaboudy and Dye further surprised tl'a scientific caununity in 1983 by reporting tla first crystalline salt with tl'a formula M+L2 e. [Cs4'(18-cram-6)2 e-] wtare all the anialic sites were occupied by solvent-free trapped electrons”. Inis naterial was called an Electride. Firal conclusive evidence for tlne assigned 1 who determined the stoichianetry was provided by Danes et a1.2 structure of (:(s+'(18-cr'av~rrn-6)2 e. by using single crystal x—ray diffraction methods. Since tl'a first allelide was prepared in 1974, there lave been ' about 30 alkalides and 7 electrides syntl'asized and thoroughly ctaracterized by various methods in these laboratories. The optical and magratic properties as well as a number of crystal structures of these naterials have provided a great deal of iraight into the nature of trapped electrons arr! alkali metal anicns.22'23 The applicability of solutions that contain these species to organic reduction prooesoesz“29 and to the doping of organic semiconductor-83° l'as already been denonstrated. The utility of alkalides and electrides has been limited, hanever, by trair tendency to decanpose trar‘mally. Their inherently strong reducing craracter requires that they be landled under an irart atmosptare or in vacuo, problems trat can readily be overcane. T‘l‘a irreversible decanposition of these canpounds at roan temperature (intimes ranging franafewsecondstoseveral days, dependingon the system) poses far more severe handling problems. Prior to the present work, all reported alkalides and electrides have utilized either of two classes of canplexing agents, crown etl'ars or cryptands. T‘l'a trarnal instability of ionic alkalides and electrides results fran reductive attack on tl'a etl'ar canpleant by tl'a unbound or weakly bound electrons. Tl'a cleavage of etl'ars by alkali natals is a canplex but mien-snidied pm”. Two possible mecl‘anisms of crown ettar decanposition are sham in Figure 2. With sonata B, tta radical fornad can calms further decanposition of the ether and foniaticn of a variety of products. With alkalides and electrides, it is believed that proton abstraction (sclneme B) is tra prevalent form of decomposition. T'Inis belief is based on tla fact tlat when divalent cations such .Hu>QSou beyond am An .Ha>06cu soon concave: an A< .coauamoc60uoo hogan cwouo you «Emacunuoz .~ ouswam u: «\p so: + as Ba“. or 0a” are used in conjmrction with creme or crypts, even thin films and solutions at -60 ’C are extremely susceptible to decanposition as a result of the even greater 0-}! dipoles induced in the molecule by the divalent cations.32 The fact that decanposition is autocatalytic, also favors a radical mechanism. The thermal instability of the crown ethers and cryptands led us to search for more robust canplexing agents that would permit crystallization of allcalides and electrides, yet not be as easily destroyed by reduction. As mentioned earlier, ligands have been synthesized with nitrogen, sulfur and ghoszhorous replacing sane or all of the oxygen donors, as well as various canbinatiors of nitrogen, sulfur, and drosphorous. Substituting only a fraction. oftheoxygenswouldmtbeofanybenefitastheether decanposition would still be prwent. Therefore, one must consider only the fully am, thia or phospha ligands. The thia crowns have been studied, and there are indications that they are good complexing agents for alkali and alkaline earth ,33,34,35 metals however, the high reactivity of the sulfur would make them unlikely candidates for alkalides and electrides. The phospha ligands have not been extensively investigated, but most 36.37 of the reported work has been with transition metals. Work with the allali and alkaline earth metals is sparse,38 but they do offer sane prunise for the future. The aza-crovms, cm the other hand, have been studied to saneextentarfihavebemstnmto 34.39.40 canplex allali and alkaline earth metal cations. Further, 25. 26 Barrett et al. andPezet a1.28 haveshownthat mixtures of potassium, sodium and hexamethyl hexacyclen (PMCY, the fully methylated nitrogen analog of ia-crowm-G: Figure 10) form stable chrk blue aoluticrns [the color one would expect with a solvated electron]. With the alkalides and electrides, the fully methylated canpounds are needed as the amino protons would be reduced in the enviranent of the weakly bound or unbound electrons . I. C. Current Work: ItwasthesuccessofBarrettardPezthatprunptedustouse the axe—cram may to synthesize the alkalide [x+(n»n{cy).Na']‘1. The unprecedented thermal stability of this canpound led to the investigation of other metals with this canplexant, and to the, exploration of the possibilities of using other amine-based ligands. Several new alkalides and one new electride which employ aza- crowns have been synthesized and characterized. The Differential Scanning Calorimetry (DSC) spectra show that some of these new materials melt and decomplex before any signs of ligand decomposition occur. Other compounds appear to be stable to melting arr! decomplexaticm during a DSC scan, yet decomplex if left at room temperature for an extended period of time (over two weeks). The structures of threeofthesecouponmiswithm imply a near overlap of the cation and anion electron densities; yet the optical, “IR and EPR spectra verify the retention of ationic and anionic characters respectively. An electride with pentamethyl pentacyclen [PMPCY, the fully methylated nitrogen analog of 15-crowm-5] , Li+(PIPCY)oe-, is extremely resistant to decanposition. bhgnetic studies show it to have localized electrons that display antiferromagnetic behavior. These studies, along with others, will be presented for the new mine-based alkalides and electrides synthesized, and the steps taken towards the synthesis of alkalides and electrides with true thermal stability will be described. II. WIS AND ANALYSIS II- A- mmfigmmwm = As a result of the highly reductive nature of alkalides and electrides, high vacuum or an inert atmosphere was required for the synthesis and all menipulatiors of the alkalides and electrides. The vacuum lines cesieted of an all-glass manifold withalowvacmnmandahighvacunmport. Onthehighvacmlmside a liquid nitrogen trap was installed just before the diffusion pump to prevent contamination of the pump oil. A more complete description of the construction of the lines is presented by Van. Eek“ . Manifold pressures (the pressure measured at the manifold arnd reported as the working pressure throughout this dissertation, although the actual pressure of the systen might be greater) of less than 2 x 10—5 Torr are routinely achieved. A majority of the inert atmosphere work was performed with the ample in a helium-filled glove box [Vac Atmospheres 00.; Model DIX-001-S.G Dri lab] with an wed nitrogen scrubbing system [Vac Atmospheres Go. Model NI-20-3]. The contamination of the abnospterewasmaintainedat lessthano.5ppn02 arnd less than 0.01 ppn H20 and was contimnously monitored by both moisture and oxygen analyzers [Vac Atmospheres Co. inn—2032 and AO-316-H respectively]. In order to introduce samples and materials into the glove box, an evacuable port was used. The time this required (~25 minutes to prevent contamination of the glove box atmosphere) could allav sane samples to decanpose. As a result, most samples 10 11 were loaded for physical measurements, such as optical studies, trermal analysis and magnetic studies, by use of a polyethylene glove bag (nan Go.) with a liquid nitrogen bath. The atnnospl'ere in the glove bag was maintained by a constant flow of boil-off nitrogen gas. 11- 3- E99133? II. B. l. W m: The complexing agents obtained from Professor Fannmn and his group members needed to be furtl'er purified. Vacuum distillation or vacuum sublimation (depending on the melting point of the. connponmd)wasusedtoentractttecompomxifremtlecrnxiemixture. All purifications were performed at less than 2 x 10'"5 Torr with the conditions listed in Table 1. The first distillation usually yielded a slightly yellow material. Furtl'er purification was obtained by washing the compound over a film of sodium netal with dimethyl ether, then redistilling or resubliming the resulting mixture. After tl'e initial distillation or sublimation, all handlim of the canpound was doneinthegloveboxtoprevent absorption of water or oxygen. 12 Table 1. Sumary of Purification Oddities For the Various Aza- Crowns. Oanpleant Formula Tanperature Method (‘0) Hecanethyl Heacyclen (HMHCY) c18H42N6 125 dist. Pentanetthyl Pentacyclen (PMPCY) 015%st 75 dist. Tetramethyl Wolen (TM) 612K28N4 130’ dist. Tetramethyl Cyclam (TE'IW) C1 41132N 4 85 subl. Trinethyl Tricyclen (TRINEY) 09H21N3 25" dist. 'It was later discovered that this canpound was extreely impure. ”Could not obtain 2 x 10—5 Torr due to its high vapor pressure. 13 II. B. 2. m The metals sodium, potassium, and rubidium (Alfa-Ventron Products, 99.95% purity) were obtained underargoninsealed glass ennpoules with breaksealsmoneeri. Cesiun (a gift from the Dow Chenical Co.) was in 50 gramampoulesanrihadtobe distributed to no gram ampoules with breeneeeels“. The contents of each tube could then be redistributed into smaller tubes following a procedure described elsewhere“. The required quantities of metal were obtained by preeasuring the inner diameter of the tube and isolating legtts to give the appropriate volumes. Lithium, on the other hand, was cut from a large piece stored in the glove box and weigred on an analytical balance. II. C. §ynthesis All of the alkalides and electrides were prepared in an apparatus called a K-Cell [Figure 3]. Glass ampoules of tlne appropriate netals were broken and loaded through side arm A while under the helium atmospl'ere of the glove box. The stoichiometric amount of complexant was tien added through side arm B. Both side arnns were then capped with cajon Ultra-Torr couplings and a sealed glass tube to create a vacuum tight system. The K—Cell could tlen be renoved fran the glove boxanndevacuated to 2 x 10"5 Torr. Both of tie Ultra-Torr couplings were rennoved by flame seal off, the metal was distilled, annd the renaming part of side arm A was also sealed off. The apparatus was then cooled to below -20’C (care must be taken nnot to let the stopcock get cold) arnd the appropriate solvent, eitl'er dimethyl ether or methylamine, was Figure 3. 14 Sample Fingers / // L, Vacnnlmapparatususedinthesynthesisofalkalidesand electrides [K—Gell] . 15 introducedastrevaporthrough the stopcock and condensed in chamer C. The canplenant was dissolved and poured onto the metal in chamber M. Onnce all the metal had dissolved, the solutian was poured through the sintered glass frit back into chamber C. A co- solvent, eitrer dimethyl etl'er or trimethylaminne, was tren added and the solutionn allowed to stand at ~-78 ‘C for 3-4 hours to pranote crystal growth. To enhance the yield, solvent could then be removed by vacuum distillation unntil only a very small amount of solvent was left. The remaining solvent was poured back through the glass frit and also distilled off. A washing solvent was then added to rinse the crystals tlnat nowwere in chamber C. Once the crystals appeared clean, the washing solvent was also‘ removed. The crystals were dried on tlne vacuum line overnight, tren poured into tie glass fingers and sealed off for future use. when lithium was used, it was dropped through the stopcock into chamber M. Methylamine was then required as the initial solvent to pranote tie solubility of the lithium so that it might canplex. Once all the metal had dissolved, the solution was allowed to stand at ~-40’C for 3-4 hours to insure canplete canplenation. The nethylamine was then canpletely removed and dimethyl ether was added. 'I‘heprooedurewasthenthesaneaswithtreotrernetals. II. D. Analygis: When an alkalide or electride is first syntl'esized, a specific end product is always expected, but nnot always realized. So. after making a new canpound, a stoichiometric analysis must be canpleted to help determine tlne identity of the end product. As 16 an example of the need for analysis, an attenpted syntresis of a "sandwiched" mound of os+[mncy-iz-cmn-41.Na’ resulted in the first successful synttesis of Gs+moNa-. This result would not have been recognized by optical studies or MAS-MR, and could have caused errors in interpretation at a later time. The analysis of the alkalides is based on Equation 1“. M+Lx.N- + 21120—-~ M+ + N+ + x1. + 321 + 203' 1 M-netali L=connplexant N=meta12 x=ior2 In the case of an electride, the appropriate balanncing is employed. The products of the reaction are then evaluated by, hydrogen evolutionn (Hzt), pa titration (OH- and xL in the case of amines), plasma enission (M+ annd N+), annd proton MGR (xL) . II. D. 1. m Evolution: Reaction 1 was carried out using a vacuum systen that had been specially designed for hydrogen evolution“. The sample was kept at liquid nitrogen temperature annd evacuated to ~2 x 10-5 Torr. Water that had been previously degassed by repeated freeze-pump- thaw cycles was then distilled into the sannple vessel. The liquid nitrogen was renoved annd as the water nnelted, it would begin to decanpose the material. (hoe the reaction was canplete, a Toepler pnmlpwasusedtocollect thehydrogengas inacapillary tube of known volume. All the hydrogen had been collected when successive pumpings no longer charged tre pressure in the capillary. By measuring tl'e total pressure change annd temperature of the 17 capillary, the ideal gas law could be used to calculate the moles of Hz evolved. Gare must be taken when analyzing samples that contain ligands with highvaporpressures suchasTRDCYnot to lose all or part of the ligand to tie liquid nitrogen trap. In this situation, the hydrogen was not collected. II. n. 2. pg Titration: A pH titration of the residue from the hydrogen evolution offers a second check on the total amount of metal in the sample. when the canplexants are amines, which are also tases, information on the amounnt of oanplencant present can also be obtained. Using tlne information fran hydrogen evolution to estimate the number of~ equivalents of base present, the residue was dimlved in a known amount of HCl and conductance water to give a solution of pflv2.75. This solution was then divided into five parts, one for plasma enission, one for in MR, and three for pH titrationn. A fresh solution of NaOH was standardized with potassium hydrogen phthalate, then used to titrate the residue. A digital pH meter (ORIEN Research, Inbdel 701A) and a Corning electrode were used to determine the endpoints for all titrations. To prevent 002 absorption by the base, the titrationwas done in a sheathed bnlret with a constant flow of dry nitrogen. It wm observed that us titration curves of the solutions were identical to the curves of the free acidified amines, but shifted to the left by an annncnnnt proportional to the number of equivalents of OH- present. The titration curve for PMPCY is similar to that of mm“, with endpoints at a pH of ~3.5, 6.5, 18 and 10.0. The first endpoint is due to a neutralization of the encessacidwhilethe renaining two are due to one and two equivalents of ligand respectively, for a total of three equivalents. The curve for m has only two endpoints, ans at ~3.5 for tie excess acid, annd one at ~8.0 for are equivalent of the ligand. By accurately measuring tress endpoints, the quantity of ligand in tl'e residue can be determined. II. D. 3. M Emissicm: Tne metal content of the solution could be determined directly by using atanic emission on a Beclonnan Spectraspan IV Emission Spectraneter. Solutions of ~30 ppn for each of the metals present~ (solutions of >100 ppn were needed for Rb+ and 08+) were naie frann a portion of the residue and analyzed by canparison with a curve obtained fran stanndard solutions. II. D. 4. in m: Tie remining amount of residue was analyzed by proton 1MB for the compth content. Tl'e solution was first neutralized by adding an equimolar amount of NaOH, as determined by the titrationn curve, then allowed to dry by slow evaporation. caution must also betakenat this point teprevent the less of ligandwhenithasa high vapor pressure. A weighed amount of sodium acetate was added to the remaining solid and diluted to ~4 ml with D20. Tie integrated intesities of the resulting spectrnnn could then be compared to determine tie mnmber of moles of complainant present. 19 II. 8. Results: Asunnnaryof thecanponmdssynthesizedandanalyzedane given below with the determined stoichionnetry. Also inncluded are outpounds that were attempted but could not be synthesized. II. E. 1. ENJOY M: With new as the canplexing agent, three sodides were 41 synthesized and fully characterized; x+(mm).ua' (I) , Rb+(I-MiCY)oNa_ (II), and os+(m).Na’ (III). The crystal structures have been determined and they verify the identifications. A fourth compound with barium has also been synthesized, but is not completely understood at this time. line e )2— 1.5 0.5 excess amount of barium netal in the product could account for the initial analyses indicate Ba2+(m).(fla Mr, an apparent electride, or an excess amount of sodium metal in the product could account for the excess sodide. A more careful study of this systen is necessary to accurately determine tne stoichiometry of the material [more will be said about this compoundinlater chapters]. From the success with the above coupounnds, 'l'l'e syntleses of two electrides were attempted, x+(m).e'andos+(mncv).e“. Itwaspossible to form stable solutions in dimethyl ether, but all attempts to crystallize led to deccnpleationn to the metal annd mm. A synthesis of Rb+(mm).nh‘ was also attempted with similar results. With the rubidide, films could be obtained, but nnot harvested. 'n'ese results will also be discussed in more detail later. 20 II. E. 2. WM: With HPCY, two compounds have been synthesized and characterized, Li+(HPCY)oNa- (IV) and Li+(PbPCY)oe- (V). An attempt to synttesize Na+(HPCY)-Na-wasmade, but there was no evidence that the my complexes sodium. 11. E. 3. WM: The canplexant TEDDY has so far enabled us to synthesize only one canpound. Li+(m)2.Na’ the crystal structure has nnot been determined. but we believe that the lithium cation is "sandwiched" between the two Tam molecules. Ann attenpt to synthesize the corresponding electride was unsuccessful. The material decompleued back to the metal and complainant during the crystallization sme. Rb+(‘IRnCY)ona- was also sought, but there was only a slight indication of complexation in dinethyl ether. At this time, the lithium sodide is the only salt synthesized with TRIMZY as the canplexing agent. Cannplexes of TRIMJY with the other alkali metalsmaystill bepossible, and warrant further investigation. 21 II. E. 4. mm: In order to test the stability of compounds that contain 8— hydrogens in the presence of the solvated electron, the synthesis of m+(nm)x.ua’ was investigated. Dark blue solutions (the color expected for the solvated electron) could be obtained in dimethyl etl'er after first dissolving the materials in methylamine and gaming dry. However, the solution was found to be tenperature depedent. At ~-20'C, the color would fade and metal flakes could be seen in the solution. If the temperature was then lowered back to below -50'C, the netal redissolved and the solution would again amear dark blue. Even with the dark blue solution at -78‘C, attempts to crystallize resulted in decomplexation. The conclusions are that this ccmpound with B- hydrogens is stable to decomposition, but that the complexation constant of m and lithium. and probably the other alkali netals, is too low for the synthesis of alkalides and electrides. II. F. m: A list of the alkalides and electrides synthesized to date with the awe-cm is given in Table 2. The physical properties of ttese compounds will be discussed in the pages to follow. 22 Table 2. Summary of Alkalides and Electrides Synthesized and Qaracterized in This Study. Compound Stoichiometric Analysis M:L:N K+(WCY)oNa" 1:1:1 Rb+(IiMiCY)-»Na_' 1:1:1 05+(W)°Na-' 1:1:1 Ba2+(mcn1).Na’ 1:1:1.5” Li+(PB‘PCY)oNa- 1:1:1 L1+(PMPCY)oe- 1:1:0 Li+(m)2.Na‘ 1:2:1" 'Structures known. ”Not positively identified. III. MICAL WY III. A. Introduction: The optical spectra of alkalides and electrides offer clues about the interaction of the most weakly bound electron with its surroundings and about how tightly it is held. In general. absorption bands occur at 620-760 m for Na-, ~800 m for K‘, ~900 nm for Rb-, ~1000 m for Ga“, and 1100-1700 m for e- (trapped)45'48. convenient method of identifying the the anionic species in an Thus, the optical spectrum also offers a alkalide or an electride. (are must be taken, however, in using such an assignment for new materials, since the absorptionn band has been shown to bedepedentontl'enatureofthecanpleting agent, the metals innvolved, and the ratio of metal to canplexant. As an eample of the variation in the position of the absorption maximum, a list of the wavelegtlns of maximum absorption is given in Table 3 for various sodide films. For the alkalides, tlne major absorptionn peak is believed to result of an rn_s_to ngtrannsition of M‘ (where M can be Na, K, Rb, or Us). A shoulder, which is quite distinct in most sodides and slightly apparent in some potassides, could be due to a bound to continnuum trannsition. Since a pronounced shoulder was present in films of na*c222.ua5, but not in. some other sodides such.as Ha+1806-Na-, it was suggested that a charge transfer frann Na- to Na" was involved”. Since the shoulder has since been observed in sodides where the cation is not Na+, the bound to continuum 23 24 Table 3. Morption Maxims Position for Na- in Various Sodide Films.(a) Film Peak Position (nm) cs+0222.ua7 685 Rb+C222.Na- 715 KI0222.Na7 660 na+0222.Na' ~ 550 x+lBC6.Na’ 715 Rbfiacs.Na’ 725 Cs+(18C6)2.Na— 715 os+(isc5)2.Na' 720 Rb+(1505)2.ua’ 720 K+(1505)2.Na’ 700 a) Reference 5. 25 transition scene to offer the better explanation. Fluorescence stndies are currently underway in these laboratories and others to better identify the origin of these transitions. The observed spectra for electride films fall in the near infrared as would be expected for a solvent—free trapped electron. These spectra can be divided into two categories, those resulting from localized electres (trapped in a cavity) and those that appear to involve more delocalized electrens (which can nave through the lattice more freely). We refer to the latter systns as those with "netallic" behavior although it now appears that they are nnot true metals. Localized electrides have electrons trapped in cavities with absorption maxima that correspond to trap depths of 0.5-0.9 eV. Those electrides tend to give broad bands that "tail" into the visible. This may result from either electron-electron interactions or a delocalized excited state. The "metallic" electrides have either free or much more shallowly trapped electrons and lnave absorption spectra that continue to rise from the blue end of the visible spectrum into the infrared. The spectra are renarlably similar to those omerved for concentrated (metallic) netal-aumonia solutiorsso. III. B. Egrinental: Althongh there are nanny ways to determine the optical properties of materials, all of the compounds prepared in this work were studied by absorption spectroscopy of thin dry filns, prepared by solvent evaporation in a specially designed optical cell [Figure 4]. A previously syntlnesized sample was loaded into 26 Graded Seal h—d \_a \_/ / Quartz ‘ 80” Optics: ‘1' Chamber CeH Figure 6. Apparatus used to obtain optical spectra. 27 tte bulk chamber under an inert atmosrinere at liquid nitrogen tenperature to prevent decanposition. The opening was capped with a Gajon Ultra-Torr coupling, the cell was renoved from the glove bag, evacuated to less than 2 x 10-5 Torr, and the arm was flame sealed off. The sample was kept at or below -40 ’C during this process and throughout tte ecperiment. A solvent, eitrer dimethyl etl'er or metnnylannine, was added and a portien of the solution was poured into the quartz optical cell. 'n'en, while splashing the solution onnto the sides of tie cell, the solvent was relieved by imnersing the bulk chamber in liquid nitrogen. The dry polycrystalline film was then placed in the optical path of a Beckman DK-2 spectrophotanneter. The tenperature of the sample, controlled by a cold dry nitrogen stream that impinged on the cell, was measured with a thermocouple. To prevent solvent frann distilling back onto the film, the bulk chamber was kept in a liquid nitrogen bath. A lead sulfide detector with a range of 400-2850 nm, was then used to obtain the absorption spectra. Efforts to study the region below 400 nm were nnot made, as light scattering by the polycrystalline films, which increases as v4 , would ted to ml: any absorption bands present. III. B. 1. M with W: The absorption spectra for tl'e three alkalides with HMICY, xflmrcymva' (I), Rb+(l-MiCY)oNa- (II), and os+(amcnr).Na' (III), shown in Figures 5, 6, and 7 respectively, were all obtained from dinethyl ether. In each case an intese peak can be observed, at 660 m («45,100 cm-l) for I, 670 m («44,900 can-1) for II, and 710 28 1.0 A L 3‘0 run it '2' {I o s-J ( 0.0 " “r I I T I 1 400 500 700 1000 1400 2200 Wavelength (nm) Figure 5. Absorption spectrum of K+(HMHCY)o Na' at -120 'c. 29 1.0 [- 670 um I E 0,5. x“ < 055 nfn 0.01 V 7 1 7 V r 400 500 700 1000 1400 2200 Wavelength (nm) Figure 6. Absorption spectrum of Rb+(MCY)'Na' at -100 ‘C. 3O 1.. 710 "m n E 0.5 ' \( ( 0-0 " 1 1 I W T V ‘00 50° 70° 1000 1‘00 2200 Wavelength (nm) Figure 7. Absorption spectrum of Cs+(HMHCY)°Na' at -107 'C. 31 m («44,100 can-1) for III. each ounpourd also had a broad higher energy shoulder centered at about 530 nm. These absorption profiles are characteristic of other sodide films an! the mximum positions of the more intense peaks fall within the range of acceptable values for the sodide anion. The red shift with increasing cation size is also expected as the transition energy dependsmtheawiromlentoffla: andascan be seen from the structures [see chapter 5 of this Dissertation] , the larger cations have a greater overlap of densities with the sodium anion. In the spectrum for II, there is also a distinct lower energy band at 855 nm (~11,7OO cut-'1 ). It is believed that this absorption is due to a smedl ammmt of contamination from Rb+(I~M-ICY)-Rb-. The "mrmal" position for the rubidide peak is ~900 nm, but as stated earlier, this depends on the compound involved. For example, Rb+(C222)-Rb- shows an absorption peak at 860 m (~11,600 and)”. The contamination could either result fran nonstoichianetric ratios in the initial synthesis, or fran the method used to produce the film. The salt Rh”(rmcsr).Na' is the least stable to decanplexation of the three discussed here (see chapter 4 of this Dissertaticm), and part of the sample may have decanplexed to form rubidium W and sodium in the local environnent. Rb+(l-MICY)oRb- might then form, leaving excess my and sodium metal behind. A similar effect has been observed with other canpounds in vapor deposition experiments in this lab“. ‘Iheabsenceofanylowerenergypeaksand the sharpness of the sodide peak in the spectra for K+(HMHCY)oNa- and cs+(mmY).Na’ indicates that there has no similar contamination by K. or Cs- in 32 I or III respectively. Although it is possible that a small peak of K“ might be hidden beneath the Na’ peak, it is not likely. netmudprobablybroadmunem‘peakorappearasa shoulder. Based m the above observations, the coupounds were identified as sodides, with the understardim that II had some contaminatien fran rubidide. The coupound synthesized with barium, sodium, and PEI-{CY has also studied by optical absorptim to help determine the nature of the species present. The absorption spectrum is slam in Figure 8. Although the analysis indicates that there is a mixture of sodide and electride present, the optical spectrum shores only the presence of sodide with a maximum absorption at 700 m (14,300 em“1 ). The small apparent peakat ~14OO nm is a result of light scattering by the optical cell, and is also present in blank runs. III. B. 2. Mammal: The absorption profiles of Li+(PM?CY)oNa- (IV), and Li+(PMPCY)-e- (V) were also studied bytheoptical absorption technique, with dimethyl ether as the working solvent . Their spectra are slam: in Figure 9. Li+(HPCY).Na- has the characteristic spectrum of other sodides with the major absorption peakat 680 m (14,700 cal-1) and a slightly apparent shoulder centered at about 535 nun. Li+(m)oe-, on the other hand, has a broad absorption band centered at about 1400 m (7,100 can-1) with 48.23.50 a profile similar to that of a localized electride [the apparent shoulder at ~2200 m is an artifact of the optical cell]. 33 1.0 700 m X . as (,5.‘ vfirh \ ( W 0.0.. 1‘ —i i 1‘ ? s too 500 woo 1000 1‘00 ’7” Wavelength (nm) Figure 8. Absorption spectrum of Bs2+(mfilCY)-(Nexe2_x)2' at -117 'C. 34 1.0 tr— 5.0 nm (A) .. i . E 0.5 d l‘ooflm ( \ B < ‘ ) D-Od 1 V 1 T I —T T‘ 400 50° 700 ‘000 "00 2200 2700 Wavelength (nm) Figure 9. Absorption spectrum of A) LiWPMPCY) Ne- at -30 'c; and a) Li+(PMPCY)oe' at -90 'c. 35 The low background for IV in the lower energy region and a similar low background for V in the higher energy region verify that these materials the are sodide and electride for Iv and V respectively. An important observation illustrated by this Figure is that there are two distinct spectra for the two very similar canpounds. 'nne mly difference in the preparatian of these two materials was the presence of sodium in the synthesis of IV, which gives rise to a Na' absorption in thin films of the resulting coupound. Since one would expect lithium and MC? to behave consistently under similar conditions, any naterial formed in the awence of sodium should still contain Li+(PMPCY) as the cation. As can be seen in Figure 9, a material with an optical absorption is obtained, even when there is no sodium present. This leaves only e. as a possible counter ion since Li- has yet to be seen in an alkalide, and in any event would absorb at higher energies. These results provide added confirmation for the formation of the electride, Li+(PMPCY)-e-. 36 III. B. a. gifimncwzma': The optical spectrum of Li+('IRIlCY)2oNa- was similar to that observed with other sodides. The major absorptian peak was at 700 m (14,300 can—1 ) (Figure 10). In the higher wavelength (lower energy) region out to 2200 m, there were no significant features that might indicate a mixture of sodide and electride. It is interesting that the higher energy shoulder ascribed to the bound- to—cantimmm transition is not apparent. It could be hidden under the high energy "tail" of the main peak, but cauparison with Figure 5, for eample, suggests that it is qualitatively different inthis case. III. C. m: A summary of the electronic absorption naxima for the canpalnds studied is given in Table 4. Based on the optical spectra and the results of the stoichionnetric analyses, it is believed that six new sodides ard one new electride have been synthesized. 37 1.0 ——700 nm i.‘ 0.5- E 6 < 0.0— f 1 '1 ' V 400 500 700 1000 1400 2200 We velength (nm) Figure 10. Absorption spectrum of Li+(TRIMCY)2oNa' at -lllb 'C. 38 Table 4. Snmary of Cptical Absorption Peak Positions. Wavelength of the Compound Maximum (m) anmncwma’ 660 Rb+(m)oNa- 670' Cs+(I~M-ICY)-Na- 710 Ba+(amcy).ua‘ 700 Li+(pMpc:Y).Na’ 680 Li+(Pl\PCY).e_ 1400 Li+(mIM:Y).Na’ 700 "me Rb. absorption has been omitted. 39 NWCW IV. A. Introggction: fineprimaryreasonforecpandirgtheresearchenallalidssand electrides into the area of an—crowns was to find materials that were thermally stable. the method of testing the thernal stability is by heating a sample at a constant rate arri measuring the enthalpy changes that occur. 'Ihis method doesnottake kinetic effects into consideration, as most of the compounds decanplex or deoanpose with time at tenperatures much lever than indicated by heating at a constant rate. This results fran processes that are kinetically slow at the lower tenperature although thermodynamically favored. beaver, this method does provide an accurate measure of the heat of reaction for a transition and offers a means of comparing the relative stabilities of various alkalides and electrides. The enthalpy (AH) is determined by converting the temperature axis to a time axis. then measuring the area under the curve of heat flow (W/g) vs. time arr! dividing by the mass of the sample. The transition temperature is then obtained by extrapolating the steepest part of the onset curve to the baseline. The temperature at this intersection is the reported temperature of the transition. Many of the alkalides (especially sodides) and a few electrides are relatively stable up to their melting point, above which they decompose rapidly and irreversibly, presumably as a 40 result of attack on the ligand by the strongly reductive enviromnent. The unique property obtained by the use of amine based ounpleants such as them-crowns, is thattheytendto deoanplexratherthandeoanpose. Insanecases, thefreeamineis then stable in the presence of the highly reducing alkali metals (even the alkali metal alloys) to tenperatures above 120 ‘C. IV. B. grimental Methods: The thermal behavior of these oanpounds was explored by differential scanning calorimetry (DSC) using an 2.1. duPont de Nanours 9900 Series thermal analysis systen. Microcrystalline samples were loaded into anodized aluminum DSC sample pans under nitrogen atmosphere at liquid nitrogen tenperatures. The pans were then hermatically sealed to permit handling outside of the glove bag. The studies were made by introducing the samples to the instrument at between -50 and -80 'C and scanning at between 2 and 10 deg/min until deoanposition occured. As previonsly mentioned, the peak temperature is dependent on the heating 8.51'52 rat A relationship between the heating rate and peak temperature for an n'nth order reaction is given by Equation 253; L9 3 A e—E/R'nn Rl‘m whereRisthegasconstant, 'l‘misthepeaktenperature, Bis the activation energy for the transition, A is the pre—exponential factor, and o is the heating rate. By determining the peak temperature at various lneating rates, the activation energy and 41 frequency factor may be determined.51 In this stuiy, only one heating rate was anployed to determine mly the enthalpy of reactim (AH) and the approximate temperature of the axiothermic and exothermic transiticns of various oanpounds. The rate of 5 deg/min was crash to give the best results on determining the enthalpy for most oanpcunds. A slower rate would provide a better estinate of the transitim temperature; however, the slower rates resulted in a baseline drift which made enthalpy determination more difficult. When the observed tramitions in the DSC trace could not be unequivocally assigned, a visual observation of the heating was also obtained by using a Thomas Hoover capillary melting point amaratus. In this case, capillary tubes were loaded at room temperature under a helium atmosphere, then evacuated and flame sealed while keeping the sample in liquid nitrogen. The observations were started at at ~O °C, by allowing the temperature to come to equilibrium first, then heating at ~20 deg/min until the sample noticeably decomposed. 42 IV. C. Thermodynamic Properties IV. c. 1. m mzs‘ As mentioned before, the met noteworthy feature of these materials is their resistance to thermal decomposition. when a microcrystalline sample of x+(mm).ua’ (I) is heated, as mitored by DSC (Figure 11) , it begins to melt at ~42 ‘C and decanplexes into the allali metals (presunably NaK55 ) and free amine at ~74 ‘C. Flutter heating yields no sign of decanposition of the canplezant at tanperatures up to 120 ‘C at which point the MICY decanposes andevolvesagas. Toverifythatthe transitions indicated by DSC were melting and decanplexation respectively, a visual observation of the heating process was also made. At ~40 ’C, the red-orange metallic-appearing crystals melted to form a deep blue liquid. At this point, if the sample was rancved fromthehot bathandinmersedinliquidnitrogen, a red—orange film could be obtained. The melting arri freezing were repeatable over more than one cycle. Continued heating of the sanple to ~60 '0 resulted in the appearance of a silvery metallic substance and a light blue liquid which eventually became clear. The metal and ccmplexant remained stable until the temperature reached ~14O 'C, at which point the solution became dark and a gas was evolved (decomposition of the m; see Figure 2). In a separate amerinent, the crystals were heated until they decomplezed to give a slightly pale blue solution, then cooled to -78 '0. Upon addition of dimethyl ether, the soluticm became dark blue, the color expected for a solvated alkali metal anion, and HEAT FLOW (w/g) 43 0.5 L 0.0- ’ 50°C -05P ‘ - LO - -l.5~ -2.0~ _ .5 1 l 1 J 1 l r l . J 1 l 1 l 1 L 1 -I50 -IOO -50 O 50 IOO ISO 200 250 300 TEMPERATURE (°C) Figure 11. Differential scanning calorimetry trace for a 3.3 ng sample of x+(mm).ua' at 10 deg/min. Endothermic values are negative. The peaks are shifted to higher tenperatures than reported; possibly as a result of the faster scan rate . 44 red-orange films could be seen on the walls of the glass cell. These observations indicate that the endothermic process that occurs at ~74 ’C in the [BC Lattern is decanplexation. Further, since the exothermic decanposition does not occur until the tauperature is above 120 'C, m is resistant to reductive attack in the presence of alkali metals at tenperatures well above 100 ‘C After measuring the melting point in the ISO, the sample was cooled to below -20 ‘C and reheated to check the reproducibility of the melting. Thus was repeated several times, but the endothermic transition at the melting point*was absent except for a small dip in the baseline. If, however, on the second heating the temperature was allowed to continue to increase, decomplexation occured at the normal temperature of 74 ‘C. If the sample was held just above the melting point for several hours, it would remain stable until the temperature was increased to the decomplexation.point. It is our belief that the crystalline compound melts at 42 ‘C to give a liquid, but when it is recooled it remains in.a glassy state so that no melting transition can. be detected. Crystals of xflnmcnnma‘ that are left at roan temperature decomplexuafter about 4 days. This indicates that the solid compound is thermodynamically unstable to deconpleation at roan tanperature, even though it must be stable at the tenperature of formation (-50 ‘C). The remarkable stability of the melt below 74 ‘C suggests that the decomplexation.prccess must be very slow, even in the liquid state. Polycrystalline samples of Cs+(M{CY)oNa- (III) had a behavior 45 similar to those of K+(m)-Na-. The former canpound appeared to melt at ~8 'C and decanplex into alkali metals [presunably Na(s)+Gs(l)] and free anine at ~37 ’C, (Figure 12). Race forms a liquid alloy at ~45 °c,"’5 near the end of the decanplenatioln process. As with compound I, a visual nelting point was obtained for this sodide which agreed with trees assignments. This canpound also decanplelnes if left at roan temperature after about 4 days. Thus, decanplexation of the melt is slow in this case also. The third canpound in this series, Rh+(mncn').ua’ (II), has a thermal profile similar to tie otter two (Figure 13). This canpound has an apparent melting transition at ~7 ‘C and decanplelcation into the alkali netals [presumably Na(s)+Rb(s)] and free amine at ~30 'C. Rb would nelt at 39 ’C under the decanplexation curve and RbNa would then fornn at ~53 .055 near the end of the decanplexation process. Due to the similarity in behavior of this canpound to x+(nmcnr).ua' (I) and os+(nmm).ua’ (III) and to tie similarity in tleir structures (chapter 6 of this Dissertation), a visual melting point was not obtained. This canpounnd also decanplexes if left at roan tanperature as with canpouncbIandIII, butinlday or less. This is not very surprising, as the decanplexation tanperature is‘anly slightly above roan tanperature. However, the decanplexation is still slaw for the melt as with r and III. The fact that Rb+(nmcnr).Na’ does decanplex much more readily than eitl'er x+(m).ua' or mflnmcnnma‘ and that its enthalpy of formation is so much lower, pranpted the proposed nechanism of Rh- formation found in 46 9‘ O T .N on 37 °C N 0 'UI 'o .0 U I HEAT FLOW (w/g) O r C ,6 ‘3' r C -I.O -I.5- 1 L 1 J 4 1 . 1 . 1 1 1 i -80 -60 -40 -2o 0 20 40 so TEMPERATURE (°C) Figure 12. Differential scanning calorimetry trace for a 2.5 no sample of Gs+(I-M{CY)-Na- at 5 deg/min. andotlermic values are negtive. 47 1.5 7 0c 30°C 1.0J I I 9 E g 0.54 9 U. *— 3 U I O J -005 ' ' ' I f T e n v— 1 v 1 v e, v ‘40 0 40 30 120 TEMPERATURE (°C) Figure 13. Differential sannning nalorinetry trace for a 2.2 mg sample of Rb+(HMHCY)oNa- at 5 deg/min. Endothernnnic values are negative. 48 the studies of the optical spectra. To estimate the enthalpy of formation of tlnese three alkalides fran the DSC innformation, the tternnodynamic cycle shown in Figure 14 was need. The enthalpy of formation of tie liquid alloy was obtained by sunning the enthalpy of mixing of the liquid metal56 and tlne enthalpy of fusion” of the metals involved (is. AHfNa + Afo + Ali ‘ I =Anformationnax‘ Table 5). The value obtained for NaK in this way was comparable to tie value reported by McKim and Branley (5.8 to 6.3 )0 nnntal'-1).58 The enthalpy of fusian of m was also obtained fran DSC measurenents. The heat capacities for tie various canponents were not included, as terms that contain them should nearly cancel out in the cycle. The results are given in Table 6. By examining the neasured enthalpy of decomplenation at various temperatures annd tlne entropise of melting of xflmncnnma’ , 1mm, and NaK, we can understand wlny the compound is stable to decanplexation at 225 K but decomplemes (slowly) at 300 K and above. Since the sodide is stable at tl'e tennperature of formation (225 K) and the value for tlne enthalpy of decanplenation of the sodide at 225 K, directly fran Figure 12, is AB = 52.5 It.) -1 -1 mol , it would imply AS < 233 J mol"1 K at this tenperature. At 300 K, the sodide decanplexes slowly with tine. This indicates flat, at this tenperature, the canpound is not stable, or A6 2 0. By calculating tie enthalpy of decompleation at 300 K, we can obtain a lower limit to the enthalpy of stabilization. Since the the products of decanplenatian at 300 K are NaK arnd I-M-ICY (1) (1)' 49 Table 5. Values Used to Obtain the Enthalpy of Formation of tie natal Alloys. Metal (infusim AH’ i 1 (no Incl-1) (no Incl-1) Na 2.7 -— x 2.3 0.3 Rb 2.2 1.2 Cs 2.1 0.9 'for tie reaction M(l) + Na(l) .. MNa(l) 50 5 M“) + Nam + HMHCYm ——>— MYHMHCY) Nais) l 2 4 3 MNaa) + HMHCYG) —-—>M+(HMHCY) N37!) Figure 14. Tnermcdynamic cycle med to calculate enthalpy of fornation of x+(nnnncn').Na', Rb+(l-M-iCY)-Na-, arnd Cs+(I-M~ICY)oNa— fran tie netals and canplexant at 225 K fran the DSC measurenents. Table 6. Sunrnary of Enthalpy Changes for Various Steps in tl‘e Thermodynamic Cycle Shown in Figure 14. Step K+(nn~nncy>.Na" Rb+(HMiCY)oNa-' eflmcwma’ kJ/mol kJ/mol kJ/mol 1" 5.8 6.0 5.6 2” 5.9 5.9 5.9 3" —29.0 -28.0 -31 .0 4” -35.2 -12.2 -35. 1 5 -52.5 -28.4 -54.6 "The value for steps 3and4aresuspect: see text. I'Calculated values. ”Fran DSC neasurenents. 51 enthalpychangeisa sum of the values for the melting and decanplexation steps [steps 3 and 4 in Figure 14 and Table 6] for the sodide. This yields AH - 64 kJ mol"1 213 J mol-1 K-l. In order to determine a lower limit to AS at 225 andimplies that AS > K, the entropies of crystallization for flex and m (AH/Tm; - 22.6 and -21.8 J mol-1 K-1 respectively) must be added to the entropy calculated for decouplenation at 300 K. This correction isnecessar'yasNaKammaresolidatzzfax. This requires -1 that the lower limit be AS > 170 J EDI-1 K for the decouplexationprocess at 225 K. Thus, at the temperature of fornation of the crystalline canpound, 170 J mol"1 K-1 < AS < 233 J mol'1 (1. Decoupleation above the melting points of the products would occur for any value of AS between these limits. Application of the same analogy to ah+(mcy).Na' (II) and mflmcmmi (III), which are stable at 225 x with AH = +23.4 and AB = +54.6 for decomplexation, respectively [calculated from Figure 14 and Table 6], implies that AS < 126 J moi”1 K-1 1 for II and AS < 243 J mol' x’1 for III at 225 x. In order to calculate thelowerlimit forAS, thereneed to be some charges in the thermodynamic cycles used. At 300 x, bothcompmmdsIIarriIIIaswellasI-MiCYarein the liquid state. Also, in an ideal case the "solid metals precipitate without formation of the alloy when the allalides deocnplex. Meyer, since the enthalpic values reported are under a large temperature envelope (Figures 12 & 13), alloy formation must be taken into account in the calculations. From the thermodynamic cycles shown in Figure 15 A and 16 A for 52 Rb+(IMICY)oNa- am 03+(1-MiCY)oNa— respectively, the energy required to decouple): the liquid allalides into M Na and (s) ' (8) ’ m can be calculated (22.0 and 27.2 kJ mol"1 x’1 for n and (1) III respectively). Then, Figures 15 B and 16 8 allow one to calculate the energy required for the solid alkalides to decomplex intoM Na arriI-M-ICY at300K.'1‘heenthalpy of (8y (8y (1) decanplexation world then be AH = 34.2 kJ llDl-l and AB = 60.2 kJ -1 mol for II and III respectively. This implies that AS > 114 kJ "1 K’1 for II and as > 200 J mol"1 :61 for III. mol In order to determine the lower limit for the entropy charge at the temperature of formation, 225 K, we must add the entropy of crystallization for the W, —21.8 J mol’1 {1, for both II and III (since both M and Na are already in the solid state, their crystallization entropies need not be added). This requires AS > 1 +92 J molm1 K"1 for II and AS > 178 J mol- l(-1 for III. Thus, at the tanperature of formatim, 225 K, for Rb+(I-M{CY).Na- +92 < AS < 126 J mol'1 1('1 and for es+(m~m<:Y).Na' 178 < AS < 243 J mol"1 {1. Decomplexation above the melting points of the two alkalides would occur for any value of AS between the respective limits. It is inocnsistent that Rb+(I~M-ICY)oNa- has such a different thermal behavior than the other tm allalides, Mm their structures and other mysical properties are so similar. By exanining Figures 15 and 16, it can be seenthatthemajor difference is in the enthalpy of meltirg of the two ooupounds. cs+(m).ua‘ has an enthalpy of melting similar to that of x+(m).ua', while that of Rb+(MICY).Na- is much lower. All other values, including the enthalpy of decouplecation, are 53 28 Rb+(HMHCY) N311) -+—__*RbNa(l) + HMHCYU) (303-325 K) + 22 + 3.8 (303 K) (325 K) + 2.2 Rb(s) + Nam + HMHCYU) (312 K) Rba) + Nam + HMHCYU) A + _ + 34.2 Rb (HMHCY) Na (8) > Rb(s) + Na(s) + HMHCY0) +22 +12 Rb+(HMHCY) Na'a) B Figure 15. 'mennodynamic cycles for Rb+(m)oNa-; A) Processes occur'ing during decouple-anon over a large teupera- ture rame. B) Deoomplexation process at 300 K. mthalpy values are in 10 1101-1. 54 +3l.0 Cs+(HMHCY) Na.(I)——>CSNam + HMHCYU) (305-320 K) +27.2 305 K +33 312 K CS“) 4- Nam + HMHCY0) A + . + 60.2 CS (HMHCY) Na (S) > CS“) + Na“) + HMHCY(|) + 35.1 - 2.1 + _ + 27.2 C8 (HMHCY) Na (I) > CS“) '1' Nam HMHCY0) B Figure 16. Thermdynamic cycles for ceflmcwma“; A) Processes oocuring during deoompleaation over a large tapere- ture rarge. B) Decomplexation process at 300 K. Enthalpy values are in kJ tool—1. 55 similar for the three alkalides. We believe that the enthalpy of melting for II is inaccurate. As mentioned earlier, these canpcnmds glassify if melted and recooled. In the process of either loading the sample into the DSC pan or collecting the sanple tron the initial synthesis, the sample may have partially melted and then glassified upon recooling. Since theglassy product does not contribute to the enthalpy of meltirg, the associated value valid be lower. Ibsever, the molten canpound would decouple: at the expected tenperature. This would account for the low values for the enthalpy of melting yet "normal" values for decanplecation. The thermodynamic calculations for Rb+(m1CY)oNa-shouldberepeated on a sample which has been treated more carefully and not allowed to melt, in order to more accurately determine the enthalpy of melting. The thermdynamic properties of the barium compound with MIC? were not obtained. However, thermdynamic measurements might provide a useful method of determining whether a mixture of two canpounds exists, or if it is one material with a unique stoichionetry. IV. C. 2. m M When microcrystalline samples of Li+(PbPCY).Na— (IV) and Li+(m).e’ (V) are heated at 5 deg/min, as monitored by use (Figure 17 A & B), there are no melting or decompleation transltiors seen. The spectrum for m+(HPCY)oNa- does have a enall (~1.1 kJ Incl-1) etiothermic transitions at -s4 °c and a larger (~40 kJ moi-1) emthermic transition at ~120 ’C. The spectrlml for Li+(BPCY)oe- has a similar behavior with a small (~2.3 kJ mol-l) endothermic transition at ~-74 °c with a corresponding larger (~32 kJ moi-1) emthermic transition at ~80 'C. The e:othermic peaks are a result of sample decanposition for each canpound. This assignment is based on the behavior of V under visual observation of the heating process. Eben this sample was heated at ~20 deg/min, it appeared to renain stable until the tenperature reached ~85 ‘C. At that point, the sample started to melt then inmediately turned yellow and evolved a gas. The similar DSC pattern of canpound IV would lead one to believe that it is undergoing a similar procees at the higher temperature. The lower tenperature endotherm is more difficult to explain. Unlike the melting and decanplecatial steps in the study of the three sodides with may, this transition was reversible (Figure 18) and reproducible. By alternately raising and lavering the tenperature thragh this range (but renaming well below the decanposition tenperature), the transition was repeated several times, and the positicn and intensity vere consistent. Since there is no visually noticeable change in the materials and the process is present in both the sodide and the electride, there 57 1.6 4 ‘3. 1 g‘ 1.2.. 3 < 3 . “- 0.8« a ‘ o d I k 0.4“ J ' 1 ' I ‘ 1 ' ‘ ' .100 -50 o 50 100 150 Temperature (°C) A 2.5 1 A 2‘ 3‘ . 3 V1.5« 3 3 1 LI. ‘d ‘5 . 0 Iosl 0 v V 1 Y t V v -1OO '50 0 50 100 150 Temperature ('C) Figure 17. DSC spectrum for A) Li+(PbPCY).Na-; and B) Li+(m).e'; at 5 deg/min. Endothermic values are negative. 58 must be sane internal charge in the molecule, possibly the orientation of the my ring around the Li+, or a partial melting of the material. This transition will be discussed in rare detail later, as it was also detected in the electride by magnetic susceptibility measurenents and conductivity.59 It would be an interesting experiment to test whether it is prwent or not in model salts of Li+(PtPCY) with calventicnal anions such as I’ or scu’. Notwithstandirg their novel properties, these two compounds do behave in a similar fashion to the allalides synthesized with W in one respect. They decouple: if left at room temperature for an ertended period of time (over 60 dew: and over 10 days for IV and V respectively). In order to verify that the compounds did truly decanple:, samples of each canpound, Li+(P!\PCY).Na- and Li+(PMPCY)oe-, were left at room temperature, under vacuum, until they noticeably "died" (they lost their color and crystallinity so all that remained were a clear liquid and metallic flakes). In the case of the sodide, dimethyl-ether was then added to produce a dark blue solution, the color erpected with allaalide and electride solutions, and orange metallic films. With the electride, methylamine was added to give a dark blue solution [Li4'(CI-13Nl-I)4 + e(s)"; this is proof that the metal had not been destroyed], then M to give a black of material which appeared to be the electride. If either sample had decomposed rather than decompleed, the dark blue solutions would not have been seen. 59 1.6 J 314‘ J E A 2 3 1.2- U. .6 '1 d) I 1.o~ n 0.8 v I v 1 ‘ I ‘- ‘100 -80 -60 _40 -20 Temperature (°C) . . + - Figure 18. DSC trace over the phase tranSltlon ln Li (PMPCYhNa . A) Heating; B) Cooling. 9’ o l '9 on 1 1 I" o 1 1 .5 U! l 1 Heat Flow (W/g) and 'o l A .° 01 t A oo~ '1'}. ‘ ' U -150 -1OO 50 O 510 100 Temperature (°C) Figure 19. DSC spectrum of Li+(mrmY)2.Na' at 5 deg/min. Endothermic values are negative. IV. c. a. Li+jTRDCY)2.Na-: The final canpound in this study, Li+(TRDCY)2oNa-, was also investigated by DSC (Figure 19) . This material presented no indication of melting, decanpleration, or any other type of transition other than a sharp enothermic process at ~80 ’C which has been identified as a decanposition. Preliminary studies have indicated that this canpound is stable at roan tenperature for up to three malths. Although there is no quantitative proof, it appeared that decanpleatial rather than decanpositiai was the governing process in its "death". This canpound shcmld be studied in greater detail to determine the ertent of its thermal stability. IV. 1?. my: Asumnaryof thethermalprocessesthatoccurinthefisodides and 1 electride are given in Table 7. It seems that these ccmpounds have provided a solution to the problen of autocatalytic decanposition at roan temperature, but they no» suffer fran lav canpleation strength. The fact that sane of then decanpose at temperatures near or above 100 'C is not a najor caicern to the goal of finding alkalides and electrides that are stable at roan tanperature. The neat canplerants also tend to decanpose if heated (~140 'C for W). However, the problen of decanpleratial when the materials are left at roan tenperature for periods ranging fran 1 toGOchysrenains; orennlst tryto find ligands with a greater affinity for the alkali metals. 61 Table‘l. Samar-yoftheThermalProcessesThatOccurinthe DSC Traces With a Heating Rate of 5 deg/min in ‘C. Canpound M.P. Decaupleation Decanposition Phase Temperature Temperature (marge K+(MCY)-Na' 42* 74 140" __-- Rb+(MICY).Na' 7 30 140" ---.. Cs+(1-M-ICY).Na- 8 37 140" ---- Li+(mPCY).Na' --- -- 120 -54 Li+(PMPCY).e- -—- --— so -74 Li+(T‘RIlVCY) Zara" —- -— 30 __.. "Heating rate at 10 deg/min. ”Decanposition of the MCY ligand. 62 VWCWIES The nagnetic properties of alkalides and electrides offer a method of identifying the cationic and anialic species in these unusual solid state materials as well as a means of understanding the nature of the cation-anion interactions. _l~_t_agic Angle §pinning yuclear gagnetic W (MAS—M) spectroscopy is a powerful technique for probing the local structure around the nucleus, while Electron Paramagnetic W (EPR) spectroscopy can reveal information about the trapping sites of the electron, the electron-electron interactions, and electron-nuclear interactions for electrides and contaminant electrons in alkalides. The magnetic susceptibility is another technique used to probe the interactions between the unpaired electrons in the electrides. All these techniques are enployed, when possible, to help understand the behavior and properties of these unusual materials. v. A. MAS—NMR The chemical shift in the MAS-M spectrum of alkalides and electrides can be used as a "fingerprint" to identify the animic or catialic species presentz3 Also, there are several al-going projects in this laboratory to better understand the interactions of the canpleued cations and the solvent-free anions with their surroundings. Plavever, inthisstudy, MAS-MRwasusedallyasa means of identifying the anialic and cationic species. All samples were studied with a Bruker 180 m2 (proton frequency) spectraneter using polycrystalline samples.60 63 v. A. 1. mm: The regials of both ”X and 2”Na were mined for K+(I-MiCY).Na- (I). Sodium-23 MAS—m gave a sirgle sharp resalance at -60.8 ppm [relative to Na+ (aq.)], a chemical shift 61,62 that is characteristic of all sodide salts. No peaks of Na+ orx'wereobserved. As e:pected on the basis of its large quadrupolar broadenirg,63 the signal of 1:“ was not seen. For Cs+(HM-ICY)oNa- (III), the regions of was and 2=Na were earnined. Again, sodium-23 MAS-MGR gave a single resonance at -60 ppn [relative to Na+ (aq.)]. Surprisingly, MAS-M gave no discernible peaks of 13308 even though 05+ is normally detectable in other crystalline alkalides and electrides.61'64 No peaks were observed for Na+ or 08-. The MAS-M spectrum of Rb+(I~MiCY).Na-' has not yet been determined. It is expected that the 23Na spectrum will have a peak at ~-60 ppn [relative to Na+ (aq.)] as with the two similar conpounds. The ”Rb spectrum would probably not be observed for Rb+ as a result of its large quadrupolar broadening.‘55 tamer, there might be a peak due to Rb- if the contamination observed with the optical sample is present. An interesting stuiy would be to carefully load a sample (not allow it to 'warm up to decaupleation temperatures), and then study the possible formation of Rb- upal tanperature cycling. V. A. 2. MM. The regions of both 7Li and “Na were examined for samples of Li+(PMPCY)oNa- (IV). As with the other sodides, sodium-23 MAS-M2 gave a sirule sharp rescnance at -60.7 ppm [relative to Na+ (aq.)]. Lithium-7 mus-m gave a single sharp resonance at ~ -O.1 pm [relative to Li+ (aq.)]. No peaks were observed for either Na+ or 1.1-. For Li+(PMPCY)oe- (V), ally the region of 7Li was examined. miscanpmmigaveasingle very sharp resonance at 421.0 ppm [relative to Li+ (aq.)]. 'lhesharpless of this peak (Av = 160 Hz; 3.8 kHz spinning rate) indicates that the 7L1 1/2 nucleus is surrounded by a very symnetric environment. Also, it an be seen that the Li+ signal is paramagnetically shifted in this compound, as would be expected for an electride. For both compounds, the temperaturedependence of the peak position and line shape was alanined. The MAS-MR signal has not noticeably affected by the transitions that occur at -54 ’C for IV and -74 °C for V. This indicates that the transition found by DSC and, for canpound V, by conductivity and susceptibility measuranents (vide infra) is probably due to lattice changes rather than changes in the local structure around the lithium cation. V. B. a: The electron has a spin angular muentum with two degenerate energy levels. In the presence of a magnetic field, the degeneracy is removed coming the levels to split into two states characterized by the spin quantum numbers 111 = .4: 1/2. Each level has an energy Em = 9811111, where g is the electronic g-value 65 (similar to the chenical shift in MR). 8 is the Bohr magneton (B 21 erg games-'1),66 and H is the applied nagnetic = 9.2741 x 10- field. 'lhe separation between these energy levels (Zeeman levels) is AB = hv -= mu. In this equation, h is Planck's constant and v is the frequency. The latter is determined by the type of EPR spectrmeter used. The most comm spectrometer is called an x- bard spectraleter with v = 9 6112. For a free noninteracting electron, the EPR spectrum would have a sirgle sharp resonance at g = 2.0023.66 In practice, the electron interacts with its surroundings causing a shift in the parameter g, and hence the resonance field Ho, as veil as a change in the line shape. Electron-electron interactions can produce either a strong g—shift or the appearance of a triplet state. If the electron couples (for times long canpared with the EPR time scale) to a nucleus of spin I, the signal will split into (21 + 1) lines. As an eiemple, an electron coupled to a potassium nucleus (I = 3/2) will produce 4 lines in the EPR spectrum. This results from a nonvanishing electron density at the nucleus because of an overlap of electronic mvefunctions. Qualitatively, the species formed Mild be the maneric pair Kte -; where (x) is a mmber (8) less than one and is directly related to the hyperfine coupling constant (A) . By measuring the separation between the resonance lines, the hyperfine coupling constant (A) may be determined, «here A is defined by Equation 3.66 8" 2 A = T- gNBN S(S+1) (vol 3 In this equation. 9N and 8N are the nuclear g—factor ard the nuclear Bohr magneton. BycanparingthemeasuredvalueofAto that for the atanic species, the percent atanic character of the nucleus can be determined. 'misisameamlreoftheamountof electron density at the nucleus, ard can be thought of as the value of x in the manomeric pair. An atomic species world have the value x = 1. With polycrystalline solids, anisotropic interactions are also detected in the EPR spectrum. These interactions can reveal valuable information about the local structure of the species in the vicinity of the resonant electron. Such information is lost in liquid samples due to molecular and ionic motion. V. B. 1. Alkalides and Electrides: In a pure polycrystalline alkalide, all of the electrons are spin paired (nsz). As a result, one veuld not expect to detect an EPRsignal. Hovever.asaresultofthemethodswedto synthesize alkalides, there is a low caloentration of electron impurities in most alkalides. Ihis results from the equilibria shown in Equation 4. 67 solvent _ _ M°+xL <====> MLX +e(s) 4 No + 9(5)- (E) N During the crystallization process, electrons may becane trapped in the crystal in a fashian similar to "color centers". As an emple, electron doping can be produced in in allali halides by heating in the pmence of alkali metal vapors or by irradiation with r-rays.67 The samples used in this stuiy vere obtained from a "normal" synthesis. A polycrystalline sample was loaded into a quartz tube (3 or 4 run diameter) and sealed under vacuum at liquid nitrogen temperature. The spectra were recorded on a Bruker model 200 X- band spectrcmeter . 68 v. 8. 2. Results: The spa spectra for K+(PM-!CY).Na- (I), Rb+(I~MHC)-Na- (II), and cs+(mncv).Na’ (III) are shown in Figures 20 A, 20 s, and 20 c 68 All three alkalides show a clear hyperfine respectively. couplirg to the metal cation, where I = 3/2, 5/2, and 7/2 for K, in), and Cs respectively. The measured hyperfine coupling ccnstants are 3.5 G for I, 38 G for II, an! 100 G for III. The corresponding hyperfine coupling constants for atmic ”K. "Rb. and mcsare 82.38 G, 361.06 a, and 819.94 6 respectivelysg. These values correspond to approximate percent atanic characters of 4.2 for I, 10.5 for II, and 12.2 for III. 68 106 K*(HMHcv)-Na‘ ‘_\\\M ‘A ”WH Rb"(HMHCY)-Na‘ B H) Hr Cs"(HMHcv)-Na' C Figure 20. EPR spectra for the three sodides; A) K+(HMHCY).Na'; a) Rb*(mcr)-Na‘; C) cflwmcnmi. Taken with an X-band spectrometer. 69 FrantheEPRspectrait isapparent thatelectronsaretrapped in the crystal, but the trapping site is still unknown. The strong hyperfirne interaction indicates that the electrun is in clwe proximity to a sirngle alkali metal cation. The crystal structuresof thesethreesodidesshanaclosecantact betweenthe sodium anim and the capleed metal caticrns [Figure 27; Chapter 6 of this Dissertation] . The EPR spectra provide strong evidence for electron trapping at the aniornic site. This is also in accord with charge considerations that require e- to replace Na-, unless excess catiors are present. The EPR patterrns for all three canpounds are characteristic of ewiremnents with axial synmnetry (showing parallel arr} perpendicular componernts) . For compound I, the anisotropy is not obvious because of the small overall breadth of the spectrum. However, the line shapes for all three were fitted by assuming that g and A have parallel arnd perpendicular canponents7o, end the crystal structures shove that there is axial synnnetry around the sodimnaniorsinthethreecanpounds(chapter6 of this Dissertation) . The most surprisirg result is the absence of any apparent interactions with the protons on the my ring. A power study (the line intensity as a function of microwave paler) shows a canpletely harngeneous hyperfine couplirng. This indicates that the electron is not strongly coupled with all the protons on the "veils" of the cavity. Also, sirnoe the protons from the methyl groups are protrudirg into the cavity between the arnionic site and the cation [Figure 27], it would seen that the electronic vevefunction is "bending" itself around the protons to interact 70 with the cation. A similar behavior has been noted for solvated electrons in amines, where there is a positive electron desity at the nitrogen atan arnd a "negative" density at the proteas.85 I-bever, it is still possible that the electrons may interact through space (via dipolar coupling) to rnearby protons as well as to other mlclei. This behavior can be investigated by _E_lectron fiuclear D_ouble Besonance spectroscopy (ENDOR) . These studies are currently urnder investigation in our laboratory in cooperation with Professor Babcock end his group.71 V. C. Wtic Sinsceptibility: The magnetic susceptibility (x) of a sample is an inportant piece of information in understanding the interactions between urnpaired electrons. Since the alkalides all have an n52 electronic ground state, . they would be expected to have a diamagnetic behavior”. The small dopirng of electrons seen by EPR methods would probably give rise to a small Curie tail, but not much information could gairned. Therefore, only the electride, Li+(PMPCY)oe-, has been investigated by using uagnetic susceptibility. In order to obtain the susceptibility of the "pure" electride, the contributions fran the bucket arnd other impurities (possibly fran partial decanposition) must be renoved. Landers et al.73 devised a method to subtract the bulk diamagnetic contributions by allowing the sample to decompose after the susceptibility measurements on the "live" sample had been made. The susceptibility of the decomposition products could then be measured arnd subtracted from the original data. This removes all 71 cantributions to the susceptibility except those due to the unpaired electrons. The susceptibility of a pure paramagrnetic material has been found to be inversely proportienal to the temperature and can be related by a proportionality constant [0; the Curie constant]. finenthenaterialhas 100%of its spirsunpairedinthestates= 1/2and in the absence of orbital arngular unnentun of the electrons, C = 0.3760. This "lav" was modified to incllde mean field effects generated by ferrcnnagnetic and antiferranagnetic materials above their transition tennperatures. The resultant relationship is given by Equation 5; where C is tlne Curie constant and 6 is the Weiss constant. To reflect the posibility that a mole of compournd may not contain a mole of urnpaired electrons, C in equation 5 should be replaced by fC in which f is the fraction of electras that are unpaired. This law is then called the Curie—Weiss law. The high tenperature susceptibility may then be used to predict whether a material has a tendency to be ferromagnetic (e > 0), paramagnetic ( e :- 0), or antiferromagnetic (e < 0). In a true antiferromagnetic sanple, a characteristic feature of a plot of x vs T‘ is a cusp at a tenperature T = TN, where TN is called the Nee'l or ordering 72 tenperature. Below the Nee’l temperature, the electronic spirns tend to be aligned in an antiparallel arrangemnt ard are canpletely ordered at T = 0. This ordering temperature may be apprmtinated by TN - -e. Below the Neél tenperature, the conponents of the spin axis which are parallel arnd perpendicular to the applied nagnetic field will be different. The perpendicular coupenent (in the limit '1' - 0))(1 = i/p, where u is thenagnetic mnent, arnd the parallel component X" = 0. The susceptibility of a random powder will therefore level off at 2/3 of the maximum value at TN. v. c.1. Li+(m).e’: The temperature dependent susceptibility was treasured for Li+(pMPCY).e' up to 300 x with an applied magnetic field of 3 k6 (Figure 21) . This componmd exhibits antiferromagrnetic behavior with a Neél temperature of 35 K. The susceptibility levels off at about 75 96 of the maximum value rather than 67 96; however, this could be due to the rourndirng of x at the maximmn. One interesting feature of the spectrum is a sharp drop in susceptibility at ~ 190 K. This is the same tenperature region as the unusual endothermic trarnsition observed in the DSC trace (195 K). As with the DSC spectrum, this trarnsition is reversible and reproducible. In order that we might understand the interactiorns more clearly, we will ermine the spectrum from a different perspective. If we invert arnd rearrange Equation 5, we obtain Equation 6. 73 .eamwm uaumcwes ox m a as musuoumdsmu mo coauucsw a we ne.isoezsv+es so essenceseoomem cases as» so does .Hs seamen C: ossuoswan... see can see as. as. as ea _ L _ _ s _ _ _ e p D D D D U D D D me D U U D D v.c a.o o.P N.w v.w m.w e.w UN))‘: (801x) (aguonnoala) 74 1 1 —=—T- X 6 15C fC Byplotting 1/xvsT, thevalueoff can be obtained from the slopeandtheWeiss cestantcanbeobtainedfmmtheintercept of the line. This plot is shown in Figure 22. An interestirng feature of this curve is that 2 separate straight line regions are obtained, one between the Neel temperature arnd the transition tenperature, R arnd a second above the transition temperature, 1! R2. Clearly, these two regiorns must be treated irndividually. From Fqnetion 6, values of fC = 0.237 arnd f0 = 0.120 are obtained for R1 arnd R2 respectively. Time correspond to a value of ~65 X unpaired spirns for R1 arnd ~32 X urnpaired spins for R Similarly, 2. the Weies constant may be obtained fran the intercept of each line. Region 1 has an intercept en the y—axis of 290 giving a Weiss constant of e = -68. Region 2 has an intercept on the y- axis of ~ 0 to give a corresponding Weiss constant of 0. The values obtairned for both regions have sane uncertainty since there were few points to fit with each line. For Region 1, Curie lav behavior is not generally considered valid until the temperature reaches about 4 TN or 140 K, arnd would cease at the transition. Region 2 was limited by the upper limit in temperature. A more careful temperature study should be performed with a larger number of points taken, which might change the values slightly. However, using the values obtairned as rough 75 .u sea a so coarseness on» soaoo ecu u>ooe wuoa>ocmo acoumuuae haoaoofiuoc o3» one muons .uusueuoasuu mo coauucsu a we x\H uo used .- cheese 33 03333.3... one cow use of. as... a s s, r. _ u _ . s . s .1 _ .Jv _ a nzunfiuuu zflflw D D on DO U D on U D D D D D Q..- N.— c.—. e.« (9-!” X) (oguonnoon) (w) ’9. 76 estimates, some general conclusions can be reached. Recalling the previous "rules" relatirng e to the expected magrnetic behavior,in region 1 the sample behaves as an antiferrunagnet while in region Zitbehavesasaparamagnet. Frmthe observed charge in the "Curie constant" and the Weiss constant, there appears to be an interestirng reorganization of the crystal lattice at tlne trarnsition tenperature which allows a greater pairirng of the electrons . Interest irngly , pressed powder conductivity measnlrenents indicate an incrwed cornductivity and a smaller band 59 This showsthatthe gap above tlne transition temperature. electrons can move more freely through the lattice. The structure of Li+(PMPCY)oe— has not yet been determined, so the erect nature of the lattice transformation is not known. A remaining puzzle is the low value of f in each region and its constancy at various temperatures. Clearly, more work rneeds to be done to understand this systen. V. D. M: The magnetic properties have helped to confirm the identities of the species present inthecompoundsstudied. Further, the MAS-MR arnd nagnetic susceptibility studies on Li+(pnpcv).e’ irndicate that the trarsitian occurrirng at ~ 190 K is a result of a lattice transformation. The interactions of the unbound electrons appear to be affected, rather than the environment of the lithimn with respect to the PMPCY. The initial assumption that the PMPCY ring was sonnel'nw "twisting" around the lithium cation is not valid. 77 TheMAS-MresultsfcrthethreeMiCYsodidesstmthatthe eations and anions retain their respective characters, ever though theyappeartobeincontact. TheEPRspectragiveussane clues as to the location of the trapped electrons. The evidence suggests that e. replaces Na- in an anionic site, coupling to the catim but not directly to the protons surrcurriing the cavity. “WWW VI. A. Introduction: Single crystal x-ray diffraction is the ultimte method of determining the identity and stoichiometry of a compound. As an sample, a canpound thought to be the electride C}s+(0222).e-,75 was proven to be a csside, Cs+(0222)oGs-,74 by single crystal structure determination. Although the electron density due to the electrou'e in the electride is so diffuse that it is hidden in the background, the absence of significant electron density in sane regions can be informative. It is believed that these "holes" are sites at which the electron is "trapped". With alkalides and electrides, the structure also helps one understand the electronic, magnetic, optical, and thermal properties of these unique solid state materials. It may also help to predict certain properties, such as how to increase thermal stability, by understanding the nature of the canplexation of new ligand types. Fran an understanding of cation-ligand and cation-anion interactions, we may be able to "modify" the ligand to increase canplexation of the cation. This might then overcane the decanplaation problems which currently plague the amine-based allcalides . 78 79 VI . B. Egrimental: The factor that limits crystal structure determinations of alkalides and electrides is the difficulty of growing suitable single crystals. Simle crystals can sanetimes be obtained fran the initial synthesis, but usually the sample rust be recrystallized. The two most successful recrystallisation methods used in the study of alkalides and electrides are slow temperature scanning and slow solvent evaporation. The crystals used in this study were all grown by the second method. Polycrystalline samples, synthesized by the methods described in Chapter 2, were loaded into one chamber of a two—chamber apparatus used for crystal growing. Dimethyl ether was then added to dissolve the material and the solution was filtered through a sintered glass frit, after which the less polar solvent, trimethylamine, was aided. The solution was then cooled to —78 'C and all the solvent was renoved by slow evaporation (~6 hrs.). In order to select a good single crystal, the solvent-free sample was transferred onto a cold copper block kept at —45 ’C in a nitrogen-filled glove bag. The crystals were then covered with cold, purified n—octane to protect then fran the atmosphere of the glove bag long ernigh for a well formed crystal to be selected by microscopic examination. (nice a suitable crystal had been found, it was mamted on the end of a glass fiber with Celvacene (medium) high vacuum grease and transferred under a strewn of cold nitrogen gas (-60 ’C) to a Nicolet P3F counter-controlled 4-circle diffractometer that med graphite-monochronatized MoKo radiation (A = 0.71073 A). The crystal washept underacold nitrogen 80 strem throughout the entire data collection with a locally mdified Nicolet LT-l low temperature system to prevent thermal decomposition or midation. A more couplete description of the procedureused to select andmount acrystal isgiveribynawes.76 v1. 0. may; Collection v1. 0. 1. K+LHMHCY).Na-:41'54'77 Preliminary examination of the bronze, irregularly-shaped crystal as well as the subsequent data collection were performed with the crystal kept at or below --67 ‘0. Due to its irregular shape, the dimersions of the crystal were not obtained. The unit cell constants and an orientation matrix for the data collection were obtained by least squares refinement fran the setting angles of 16 reflections in the range 7.5 < e < 10° . The orthorhanbic primitive cell constants are a = 11.091(3), _c_= 11.172(4), and _c_= 22.531(7) for a calculated cell volume of 2791.8(14) A3 . With a fornmla weight of 404.67 ard Z = 4, the ealculated density is then 0.96 g/cm3 . From the systematic absences of reflections and fran least squares refinerient, the spacegroupwasdetemined tobeP2 rabies") . 12121 (#19 in the International The intensity data were collected by using the 6-26 scanning technique at 4’/min (in 29) up to sine/A = 0.5385 Ad, (29 = 45'). A total of 6405 reflections (2094 unique) were recorded. Three maritar reflections (measured every 93 reflections) indicated that there was negligible decay of the crystal during the data 81 collection. Therefore, a decay correction was not anployed. With a data cut-off of I > 30(1), there were 1222 observed and 872 unomerved (mmeasurably weak but not systenatic absences) data points. The structure was solved by using the Patterson method. Full-matrix least squares refinenent was on F (Ii/2) . The non-H- at‘s were refined anisotropieally while the H—atans were constrained to ride on their bonded C-at- with fixed isotropic thermal parameters. After the final cycle of refinement, the maximum shift/error ratio [the shift in a parameter divided by the standard deviation in the measurement] (6/0) = 0.03, and R = 0.045; where R is defined by Equation 7.,29 >2 I IFOI - IFCI l >2 |Fo| F0 is the observed structure factor and Po is the structure factor calculated from the refined structure. R then compares how well the calculated model fits the measured data. The final difference nap peak heights ranged fran -o.20(4) to 0.22m e A’s. The molecular structure and numbering of the atans are shown in Figure 23 A; andthestereOpackingdiagramisshowninFigure 233. The positional parameters, bond distances, and bond angles are given in Appendix A; Tables A1, A4, and A5 respectively. 82 Figure 23. A) The molecular structure and the numbering of the atoms; and B) stereographic packing diagram of K+(HMHCY).N3'. 83 VI. 0. 2. os+(m0Y).Na’:5"77 Preliminary sanitation of the copper colored crystal and data collection were performed with the sanple kept at ~-75 ‘C. This crystalappearedtobeprisnshapedandhaiapprcndmatedimensions of 0.20 x 0.28 x 0.45 cm. The unit cell and an orientation natrix for the data collection were determined by least squares refinement from the setting angles of 12 reflections in the range 7.5 < e < 10'. The orthorhombic primitive cell constants are a = 11.021(4), g = 11.411(4), and g = 22.886(6) A. Theseyielda calculated cell volume of 2878.2(14) A3 . With a formula weight of 498.47 g/mole and Z = 4, the calculated density of this sodide is 1.16 g/cm3. From the systematic absences of reflections and least squares refinement, the space group was determined to be P212121, isostructural with K+(HMHCY).Na-. The intensity data were also collected by using 6-26 scanning techniques at 4°/min (in 29), to sine/A = 0.7035 A“ . (29 = 60’ ) - A total of 4726 reflections (4696 unique) were recorded. In this case, the three monitor reflections decreased by an average of 0.6% in intensity. It was therefore neccessary to anploy a linear decay correction. With a data cut—off of I > 30(1), there were 2996 observed and 1700 unobserved data points. This structure was solved by using the coordinates of the xflmhua‘ structure as initial values for full-matrix least squares refinenent on F. The non-H-atans were refined anisotrepically while the H—atans ware castrained to ride on their bonded C-atans with fixed isotropic thermal parameters . 84 After the final cycle of refinanent, the maxim shift/error ratio (6/0’) = 0.07 and R = 0.033. The final difference map peak heights ranged fran —0.37(7) to O.61(7) e A's. The mlecular stnictureandmmberingoftheatcuareslminFigure 24A; the stereo—graphic packing diagram is shown in Figure 24 B. The positional parameters, bard distances, and bond argles are given in Appendix A in Tables A3, A4, and A5 respectively. v1. 0. 3. unfit-momma?“ Preliminary examination of the bronze colored crystal and data collection were performed with the sample kept at ~—87 °C. This crystal appeared rod-shaped and had approximate dimersions of 0.18 x 0.25 x 0.40 m. The unit cell and orientation matrix for the data collection were determined by least squares refinenent from the setting angles of 13 reflections in the range of 6.5 < 8 < 10'. The orthorhombic primitive cell constants are _a_ = 11.075(4), g = 11.227(6), and _c_:_ = 22.781(10) A for a calculated cell volume of 2832.6(3.8) A3 . With a formula weight of 451.12 g/mole and Z = 4, the calculated density of this sodide is 1.06 g/cm’. From the systenatic absences of reflections and least squares refinenent, the space group was determined to be P212121; isostructural with xflmncwma’ and 03+(mm).Na’. The intensity data were collected by using an Omega scanning technique at 4 ’/min (in 28), to sire/A = 0.5385 A-:1 , (28= 45'). The 0-28 method could not be used inthiscase,asthepoor quality of the crystal caused a broadening of the scattering Figure 24. A) The molecular structure and the numbering of the atoms; and B) Stereographic packing diagram of “Vanuatu Na". 86 profile. A total of 4037 reflections (2121 unique) were recorded. The three mitor reflectias indicated that there was negligible decay of the crystal durim the data collection. Therefore, it was not neccessary to employ a decay correction. With a data cut- off of I > 30(1), there are only 984 observed and 1137 unobserved data points. An attenpt was made to solve the structure by least squares refinanent on F using the coordinates fran the K+(MCY)-Na- structure as initial values, but the coordinates did not refine. The Patterson method was then anployed to solve the structure. Full-matrix least squares refinanent was enployed on F with non-H-atans refined anisotropically, while the H-atans were constrained to ride on their bonded C-atans with fixed isotropic thermal parameters. After the final cycle of refinement, the maidmum shift/error ratio (6/0) = 0.09 and R = 0.037. The final difference map peak heightsrangedfran -0.27(6) to 0.2(1) e A.-3 The molecular structure and mnnbering of the atoms are shown in Figure 25 A; and the stereo-packing diagram is shown in Figure 25 B. The positional parameters, bond distances, and bond angles are given in Appendix A; Tables A2, A4, and A5 respectively. VI. D. Discussion: The cell constants for the three sodides are sunmarized in Table 8. Qualitatively, the “CY ring [1,4,7,10,13,16—hexaam— 1,4,7,10,13,16-hexamethyl-cyclooctadecane] forms a "cup" around the alkali cation. Nitrogens 1,7,10,16 are in the ring, coplanar within 3 0.09 A for it"(nmcmma' (I), z 0.06 A for Rb+(HM-ICY).Na- Figure 25. A) The molecular structure and the numbering of the atoms; and B) Stereogrephic packing diagram of Rb+(HMHCY) Na“. 88 Table 8. Summary of Cell Constants For x+(m).ua' (I), Rb+(I-MiCY)oNa- (II), and cs+(mm)ma' (III). x‘mncmmi Rb+(mm).Na‘ Gs+(I-MiCY).Na- (I) (II) (III) a = 11.091(3) A 11.07am A 11.021(4) A b = 11.172(4) A 11.227(6) A 11.4mm A c = 22.5310) A 22.781(10) A 22.886(6) A v = 2791.8(14) A3 2832.6(38) A3 2878.2(14) A3 d = 0.96 g/cm3 1.06 g/cm3 1.15 g/cm3 R = 4.5 X 3.7 X 3.3 X 89 (II), and t 0.03 A for 08+(MCY).Na— (1m with the corresporriing methyl groups forming a cage arourd the cation. Nitrogens 4 and 13 are below the plane 1.27 and 1.63 A for I, 1.12 and 1.51 A for II, and 1.14 and 1.43 A for III, with the methyl groups effectively closing off the bottan of the ring. The cation is located 0.29 A, 0.59 A, and 0.91 A above the carter of the plane for I, II, and III respectively. The caticns are coordinated to all six nitrogas with K—N distances ranging from 2.903(6) to 3.014(6) A for I, Rb—N distances ranging from 3.014(10) to 3.170(11) A for II, and Gs-N distances ranging from 3.157(6) to 3.333(6) A for III. The sodide anion is 4.280(3) A from the potassium cation in I, 4.202(10) A from the rubidium cation in II, and 4.254(3) A from the cesium cation in III. The planar arrangement is slam in Figure 26 with the associated values given in Table 9. The example of a "molecule" shown in Figure 27, which was obtained from a computer generation of the structure of Cs+(I~lMHCY)-Na- shave the nature of the Cat-Na. ion pair. The structures of K+(HMHCY)-Na- and Rb+(HMHCY).Na' would be similar. At first glance, it might seem umlsual that the acell edge gets smaller with increasing cation size. However, the concept of the plane with its associated atomic distances allows one to understand the slight collapse. The larger cation must be farther above the center of the I-M-ICY molecule as a result of size constraints. This requires N4 and N13 to move closer to the plane in order that they might still interact strongly with the cation. As they move up, thering ispinchedinonedirection (the;— axis) and thereby expanded along the perpendicular axis (the g — xi, 3 Figure 26. Diagram of plane through the nitrogens 1,7,10,16. The axes are only approximate. Table 9 . in Figure 26 . Sunnar'y of Atomic Distances With Respect to the Plane xflrgmmmi Rb+(mflCY).Na- 08+(MCY1oNa- N—1,7,10,16 plarar t N-4 below plane N—13 below plane cation above plane sodide from cation 0.09 A 1.27 A 1.63 A 0.29 A 4.28 A $0.06A 1.12 A 1.51 A 0.59 A 4.20 A i 0.03 A 1.14 A 1.43 A 0.91 A 4.25 A 91 Figure 27. Oanputer sinulation of one "bblecule" of Gs+(I-MiCY).Na-. Generated on an Evens & Sutherland P5350 systan with the program FRDDO. [K+(MICY).Na- and Rb+(m'lCY)-Na- appear similar]. 92 axis). The nest unusual feature of these molecular structures is the close cation-anian distances, especially in 08+(W).Na-. Taking the effective radii of K+, not and Gs+ to be 1.33 A, 1.49 80 A, and 1.69 A respectively, and the effective radius of Na- to be between 2.50 and 2.70 A (determined fran the anionic cavity size in other sodidesz'al )the separationbetweenthevander Paals surfaces of the cation and anion is between 0.25 and 0.45 A for I, between 0.02 and 0.22 A for II, and between -0.13 and 0.07 A for III. This is the first evidence for contact ion pairs between alkali metal cations and alkali metal anions in allcalides. It is interesting that similar close contacts are observed in the rubidide Rb+(18C6).Rb-.82 It is remarkable that, in spite of the close proximity of the opositely charged ions, the optical and ma properties of Na’ show no evidence of appreciable charge transfer to the cation; that is, the properties are the same as for other sodides in which Na- is well isolated from the cation. The absence of an m signal for 13303, however, may be due to a strong perturbation of the p- electrons in 08+ as a result of this close proximity to Na-. In the unit cell, the molecules are ordered in a staggered arrangenent to give a distrorted octahedral coordination between the canplexed cation and the sodium anions. The distances between K“ and the six nearest Na‘ neighbors are 4.28 A for the "ion-pair" and fran 6.99 to 8.31 A for the next five closest neighbors in I. The distances between Rb+ and the nearest six Na’ neighbors are 4.21 A for the "ion-pair" and from 7.12 to 8.49 A for the next 93 five closest neighbors in II. The distances between cs” and the nearest six Na- neighbors are 4.26 A for the "ion—pair" and from 7.23 to 8.63 A for the next five closest neighbors in III. Also, the sodium anias are effectively shielded fran one another by the large canplexed catians. The distances between Na" arnd the six nearest neighboring anions range frau 8.66 to 9.94 A for I, fran 8.88 to 9.88 A for II, and fran 9.94 to 10.00 A for III. The views of the packing are shown in Figure 28 [these views are for K+(mm0Y).Na'; however, Rb+(m).nva' and 03+(MICY).Na- would appear the same except for small dimensional changes]. VI. E. Lego; m: We also wished to obtain the crystal structures of Li+(PMPCY)-Na- (IV) and Li+(PMPCY)-e- (V) in order that we might better understand the peculiar magnetic properties of the electride arnd the nature of the apparent mass change which occurs in each species. However; several attempts to grow single crystals of each canpound, by both slow solvent evaporation and slaw cooling, were unsuccessful. Oannpound IV tended to form golden flakes similar to metal filings, while canpound V always ended up as a black powder. It is possible thattheuseof different solvents or a canbination of solvents might yield suitable crystals, and this should be explored in the future. ‘g’li ' 0".. -‘.:' .. - (‘- 0' o l ‘ -‘;rn‘ - z..- 0' '.0 O”... .- fl -.. .0, . .... ‘. '9. n. - at all .- .- u 3 o‘- a. mo .? "2):... No”. .¢!=’=z'.‘" ., _., ".09 O . - .u . ' .... ,;- '4: we. 4;..- ‘ 3...}: 3i i .‘a " "., e K.l .rk..:. fg‘géxi ‘1‘. ' D- .2... ‘ I . ‘.. Y. C ... ||.l7 1,. “.K.::": ,- {i \—." \ . . I K”; .—) 4— , .1. . .-|':I.:- .. . 0’ L’;3::.o" ‘ :.-,‘ ‘:‘:.:. 2...“; 428 '3 o I”. ' J". 1..) , .- z- - I -. 3". .- |o ', .3. ' N° ---wr"" ;' ‘2. : v -.No ~— L" :5 'E:.::."'. 1‘ : ‘31:. (.2330. H ‘.....‘ a ’0 :. '0 '. "}?a." " ‘..'..Q.I a “‘ 'g. -.:'o l v' t .3. 4 l' I" .~’ ,_K:r n: ' a . :‘I I Figure 28. Two cutaway views of the cation-anion packing in K+(HMHCY) Na'. A) View down the y-exis; B) View down the x-axis. Distances are in Angstroms. The Na- is not drawn to scale. The structures of Rb+(HMHCY) Na- and Cs+(HMHCY) Na- are similar. VII WIRE The pirsuit of thermally stable alkalides and electrides via the ans-crown canplexants has not only yielded pranise of success, but has also provided acne interesting new canpounds. The study of these new alkalides and electrides has provided new insights into the nature and properties of solvent-free trapped electrans and alkali metal anions. These componmds have shown that the complexed cation can be in contact with an alkali metal anion yet not destroy the essential character of either species. Through EPR and ENDOR we have also been able to better understand the nature of the electron trapping as well as how the electron interacts with its surroundings. Probably the most important resultof thisworkis thatwehave proven that aza-crams are stable to decanposition in the strong reducing environment of alkalidse and electrides. The only drawback to the current am- canplexants is their weak affinity for the alkali metal cations, which causes decanplexation to occur slowly when the tenperature is raised. In many cases, decanplexation occurs at roan tenperatures. The fact that time connpourris are less stable to decompleation than their onto-analogs is not too surprising. Studios have shown that the effect of substitutirng N—CI-Ia for 0 in cram ethers decreases the stability of tine cannplex by a factor of 83 10 for each substitution. This results from the lower dipole momentordonicityoftheaminegroup. Sonnegooddoescomeoutof 95 96 this "problen" wearer. This is probably what enables us to synthesize a canpound with barium when all other attenpts that usedethershadfailed. We must now try to find amine-based canpleents which can canpensate for the lower affinity, yet still fornnn stable alkalides and electrides. Studies are currently in progress in our laboratories, in cooperation with Professor Farnum, to acquire "hinged" sandwich complexes or fully methylated aca-cryptands. This slnould yield a more stable canplex by encapsulating the cation. It has been shown that macrobicyclic ligands (cryptands) increase the stability of metal canpleoes by 4 enthalpic stabilization3 . This process is designated as the cryptate effect. The sane stability considerations should apply to the "hinged" complexants. VII. A. Desigg'gg _Ne_w_ (mm-ants: In the design of new completing agents, certain factors which influence the stability and selectivity must be considered. The first consideration is the donor atom type or types. There is a limitation in this respect when working with alkalides and electrides. For the alkali and alkaline earth metals, the oxygen donnors are the most favorable, while the amine donors are better suited for metals such as Cu2+, 0o2+, N12+, etc.83' Fbever, as mentianed in Chapter One of this Diwertation, the etlner links are more susceptible to reducticxn. Therefore, we must settle for the less favorable amirne coupounds and try to maximize the other variables. A second consideration is the legth of the bridging groups. Izatt et.al.84 have shown that by alternating ethane and 97 propane or by using ally proparno groups, strong stability losses occurcannparedwiththe ethano-crowns. This is presumbably a result of partial diversion of the dipole eds may fran the center of the cavity. As an example, replacing one etlnano group in 18-Crownn-6 with a propano group (19-Cram-6) reduces the complexation constant by 50 8.84 The failure of all attempts to synthesize alkalides with tetramethyl cyclam (N4-14-Crewn—4) is probably due to this effect. A third casideration is the over- all size of the ligand. The most stable complexes are formed when the ligand cavity closely matches the cation size. This effect is not as critical, since macrocycles are not completely rigid. Fist of then are quite fletible and capable of orienting their donor groups in space. Larger macrocycles are free to fold, resulting in variable 3-dimensional cavities. However, this factor should not be ignored since we already have "one strike" against us in using tlne amines. In conclusion, the ability to synthesize thermally stable alkalides and, more importantly, electrides is just around the corner. The limiting factor is in maximizing the canplexation strength of the ligand by adjusting several pareneters. We require a canplecant that will canplet and stabilize the cation, yet not be easily reduced by the strongest loswn reducing agents. The wide variety of optical, uagnetic, and electron enission properties of electrides might make than useful in device applicatias if the problem of stability can be overcanne. This work appears to be a step in the riglnt direction. APPENDIX APPENDIXA Positional parameters, bond distances, and bond angles for K+(I-MiCY)-Na— (I), Rb+(mncn).ua’ (II), and cs+(nnmcnr).Na' Table A1. A4 Positionnal parameters and their estimated standard deviations for K+(HMHCY)-Na— (I). Positional parameters and their estimated standard deviations for Rb+(mncy).Na‘ (II). Positional parameters and their estimated standard deviations for Cs+(mncnr).ua’ (III). Bond distances for x+(mmcy).Na’ (I), Rb+(I-M-ICY).Na- (II), and Cs+(l-M-ICY).Na— (III). Bond angles for K+(PM-ICY)eNa- (I), Rb+(mncnr).Na' (II), ard 0s+(mncn).Na' (III). The parameters for tine hydrogen atoms have nnot been included, as they were constrained to ride an their C—bonded hosts in all threecases. 98 99 Table A1. Positional Parameters and Their Estimated Standard Deviations for K+(m)oNa- (I). Atan x y z B(A2) K1 0.7677(1) 0.5798(1) 0.87093(6) 2.93(3) Nai 0.7203(3) 0.9527(3) 0.8917(1) 6.90(9) N1 0.5287(6) 0.5065(6) 0.9129(3) 6.8(2) N4 0.6097(5) 0.4297(6) 0.7892(3) 5.8(2) N7 0.8190(5) 0.5798(6) 0.7444(2) 5.3(2) N10 1.0219(5) 0.6063(5) 0.8335(3) 4.7(2) N13 0.9497(5) 0.4166(6) 0.9208(2) 5.1(1) N16 0.7444(7) 0.5057(7) 0.9946(2) 6.5(2) C2 0.4665(7) 0.4186(8) 0.8736(4) 8.6(3) C3 0.4859(7) 0.4447(8) 0.8098(4) 8.0(3) C5 0.6214(8) 0.4774(9) 0.7292(3) 8.2(3) C6 0.7529(9) 0.4907(8) 0.7099(3) 7.5(2) C8 0.9458(7) 0.5672(8) 0.7307(3) 6.4(2) 09 1.0260(8) 0.6402(7) 0.7702(3) 7.1(2) C11 1.0877(7) 0.4952(7) 0.8431(4) 6.4(2) C12 1.0741(7) 0.4406(8) 0.9051(3) 6.9(2) C14 0.9390(8) 0.3963(8) 0.9849(3) 7.4(3) C15 0.8104(9) 0.3947(8) 1.0068(3) 8.1(3) 017 0.6179(9) 0.498(1) 1.0109(4) 12.3(4) C18 0.5380(8) 0.453(1) 0.9734(4) 12.0(4) C19 0.4632(8) 0.6206(8) 0.9119(4) 9.6(3) C20 0.6425(9) 0.3025(8) 0.7906(5) 10.1(3) C21 0.7734(9) 0.7009(7) 0.7304(4) 8.7(3) C22 1.0722(8) 0.7015(7) 0.8699(4) 7.7(3) C23 0 9044(9) 0.3155(8) 0.8872(4) 8.8(3) C24 0 8010(8) 0.6076(8) 1.0245(3) 8.0(3) 100 Table A2. Positional Parameters and Their Estimated Standard Deviations for Rb+(l-M-ICY).Na- (II). Atom x y z B (A2 ) Rbl 0.7621(1) 0.6028(1) 0.87207(5) 3.31(2) N31 0.7174(6) 0.9721(5) 0.8939(3) 6.6(2) N1 0.525(1) 0.498(1) 0.9132(6) 7.1(4) N4 0.607(1) 0.443(1) 0.7857(6) 6.3(4) N7 0.8214(9) 0.594(1) 0.7421(5) 4.3(3) N10 1.0269(9) 0.615(1) 0.8314(5) 4.7(3) N13 0.952(1) 0.4319(9) 0.9174(4) 4.6(3) N16 0.740(2) 0.504(1) 0.9945(4) 5.7(3) C2 0.464(1) 0.420(1) 0.8714(8) 7.6(5) 03 0.482(1) 0.456(1) 0.8072(7) 6.6(5) 05 0.621(1) 0.497(2) 0.7281(6) 7.0(5) 06 0.751(2) 0.510(1) 0.7079(6) 6.2(4) 08 0.950(1) 0.576(1) 0.7299(6) 5.4(4) C9 1.031(1) 0.648(1) 0.7689(6) 6.5(5) 011 1.086(1) 0.503(1) 0.8401(7) 6.3(4) 012 1.080(1) 0.453(1) 0.9013(6) 5.4(4) 014 0.939(1) 0.401(2) 0.9797(6) 6.1(4) C15 0.809(1) 0.390(1) 1.0038(6) 6.5(4) 017 0.617(2) 0.483(2) 1.0112(8) 10.0(6) C18 0.536(1) 0.436(2) 0.9698(7) 9.3(5) 019 0.457(1) 0.609(2) 0.9195(7) 8.1(5) 020 0.637(2) 0.314(1) 0.7831(8) 8.9(6) 021 0.781(2) 0.717(1) 0.7295(7) 7.6(5) 022 1.075(2) 0.712(1) 0.8666(8) 8.2(5) 023 0.898(1) 0.338(1) 0.8813(7) 6.3(4) C24 0.793(2) 0.603(2) 1.0262(5) 8.2(5) 101 Table A3. Positionnal Parameters and Their Estimated Standard Deviations for 0s+(nMiCY).Na' (III). Atom x y z B(AZ ) 081 0.75388(4) 0.63250(3) 0.87396(1) 2.807(4) Na1 0.7166(3) 1.0010(3) 0.8958(1) 5.40(9) N1 0.5218(6) 0.4904(6) 0.9134(3) 4.5(1) N4 0.6059(5) 0.4527(5) 0.7839(3) 3.9(1) N7 0.8189(5) 0.6101(5) 0.7397(2) 3.3(1) N10 1.0292(5) 0.6256(5) 0.8277(2) 3.4(1) N13 0.9470(5) 0.4446(5) 0.9140(2) 3.5(1) N16 0.7418(8) 0.4993(5) 0.9955(2) 4.2(1) C2 0.4626(6) 0.4192(7) 0.8687(4) 5.4(2) 03 0.4798(7) 0.4651(7) 0.8056(4) 4.6(2) C5 0.6160(8) 0.5144(8) 0.7278(3) 4.8(2) 06 0.745(1) 0.5284(6) 0.7054(3) 4.7(1) 08 0.9479(6) 0.5899(7) 0.7272(3) 4.1(2) 09 1.0352(6) 0.6579(7) 0.7654(3) 4.4(2) 011 1.0929(7) 0.5146(7) 0.8372(3) 4.1(2) 012 1.0739(7) 0.4622(7) 0.8973(3) 4.1(2) 014 0.9395(7) 0.4004(8) 0.9752(3) 4.8(2) 015 0.8120(7) 0.3908(8) 0.9983(3) 5.2(2) 017 0.6133(8) 0.4791(9) 1.0132(4) 5.7(2) 018 0.5369(7) 0.4232(7) 0.9668(3) 4.8(2) 019 0.4505(8) 0.5975(8) 0.9243(4) 6.0(2) 020 0.6346(8) 0.3287(6) 0.7761(4) 5.3(2) 021 0.7845(8) 0.7310(6) 0.7267(3) 5.3(2) 022 1.0793(7) 0.7184(7) 0.8634(4) 5.4(2) 023 0.8877(7) 0.3614(6) 0.8745(4) 4.7(1) 024 0.7938(9) 0.5947(8) 1 0296(3) 6.2(2) 102 Table A4. Bond Distances (in Angstrans) for x+(mm).ua‘ (I), Rb+(mncnn).Na' (II), and Cs+(mcy).Na" (III). (I) (II) (III) Atan 1 Atan 2 Distance Distance Distance K,Rb,Cs Nal 4.280(3) 4.205(10) 4.254(3) K,Rb,Cs N1 2.917(6) 3.025(14) 3.160(6) K,Rb,Cs N4 3.014(6) 3.170(11) 3.333(6) K,Rb,Cs N7 2.907(5) 3.035(10) 3.165(5) K,Rb,Cs N10 2.964(6) 3.080(10) 3.215(5) K,Rb,0s N13 2.909(6) 3.026(11) 3.157(6) K,Rb,Cs N16 2.903(6) 3.014(10) 3.173(5) N1 02 1.491(11) 1.46(2) 1.461(11) N1 018 1.493(12) 1.47(2) 1.452(10) N1 019 1.468(11) 1.47(2) 1.474(11) N4 03 1.459(10) 1.47(2) 1.483(9) N4 05 1.459(10) 1.45(2) 1.468(10) N4 020 1.466(11) 1.49(2) 1.461(9) N7 06 1.460(11) 1.45(2) 1.465(10) N7 08 1.447(9) 1.46(2) 1.468(9) N7 021 1.479(10) 1.48(2) 1.461(9) N10 09 1.478(10) 1.47(2) 1.473(9) N10 011 1.456(10) 1.44(2) 1.465(10) N10 022 1.454(10) 1.45(2) 1.447(10) N13 012 1.450(10) 1.49(2) 1.463(9) N13 014 1.466(9) 1.47(2) 1.491(9) N13 023 1.451(10) 1.46(2) 1.466(9) N16 015 1.466(11) 1.50(2) 1.461(11) N16 017 1.452(12) 1.43(2) 1.491(12) N16 024 1.465(11) 1.45(2) 1.459(11) 103 Table A4. Bond distances cont. (1) (II) (III) Atan 1 Atan 2 Distarnce Distance Distance 02 03 1.482(12) 1.53(2) 1.549(12) 05 06 1.530(13) 1.52(3) 1.522(14) 08 09 1.498(11) 1.51(2) 1.514(10) 011 C12 1.532(11) 1.50(2) 1.513(11) 014 015 1.510(13) 1.55(2) 1.506(11) 017 018 1.326(14) 1.41(3) 1.497(12) 104 Table A5. Bond Angles (in Degrees) for K+(MTCY)-Na- (I), Rb+(I-MICY)oNa- (II), and 0s+(mm).Na’ (III). (I) (II) (III) £3991 areas. ere-us side. are e291; N31 K,Rb,0$ N1 96.4(2) 104.3( ) 113.3(1) Nal K,Rb.Cb N4 121.7(1) 124.6( ) 129.3(1) N31 K,Rb,08 N7 96.5(1) 100.0( ) 102.5(1) Nal K,Rb,Cb N10 91.8(1) 95.5( ) 98.9(1) Na1 K,Rb,08 N13 129.9(1) 131.8( ) 134.6(1) Nal K,Rb,Cs N16 98.4(2) 104.2( ) 111.5(1) N1 K,Rb,Cs N4 61.9(2) 60.1( ) 57.7(2) N1 K,Rb,Cs N7 120.1(2) 118.5(3) 114.8(2) N1 K,Rb,Cs N10 171.1(2) 159.6(4) 147.7(2) N1 K,Rb,Cs N13 110.3(2) 104.4(3) 96.6(2) N1 K,Rb,0$ N16 62.5(2) 60.0(4) 58.0(2) N4 K,Rb,CS N7 61.7(2) 59.5( ) 57.4(1) N4 K,Rb,0$ N10 116.3(2) 110.9( ) 104.1(1) N4 K,Rb,® N13 108.2(2) 103.2( ) 95.2(1) N4 K,Rb,Cs N16 113.0(2) 108.8( ) 103.1(2) N7 K,Rb,® N10 62.2(2) 60.1(3) 57.6(1) N7 K,Rb,Cs N13 104.8(2) 99.3(3) 94.2(1) N7 K,Rb,Cs N16 164.5(2) 155.2(3) 145.3(2) N10 K,Rb,0$ N13 61.4(2) 58.1(3) 56.1(2) N10 K,Rb,08 N16 112.9(2) 111.8(4) 108.5(2) N13 K,Rb,08 N16 61.9(2) 60.3(4) 56.5(2) K,Rb,08 N1 02 113.5(5) 115.5(9) 116.5(5) K,RB,08 N1 018 109.8(5) 112.6(9) 114.8(4) K,Rb,0$ N1 019 102.6(5) 98.3(9) 93.1(4) 02 N1 018 108.1(7) 109. (1) 110.3(6) 02 N1 019 109.5(6) 110. (1) 109.9(6) Table A5. iflsafl; one K,Rb,Cs K,Rb,Cs K,Rb,Cs c3 05 K,Rb,Cs K,Rb,Cs 1K,Rb,Cs 06 06 08 K,Rb,Cs K,Rb,Cs KgRb,Cs 09 C9 011 KgRb,Cs KfiRbVCS KgRb,Cs C12 012 C14 K;Rb,Cs KgRb»Cs KQRb»Cs 015 Bond.Angles cont. Atoma 2 H 0 0 0 N10 N10 N10 N13 N13 N13 N13 N13 N13 N16 N16 N16 N16 5.3553353153555535 m 019 ca 05 020 cs 020 020 ca 021 021 021 011 022 011 022 022 012 014 023 014 023 023 015 017 024 017 105 (I) 50912 113.4(7) 106.9(5) 108.5(5) 111.4(5) 109.6(6) 109.8(7) 110.6(7) 114.3(4) 113.4(4) 99.2(4) 107.9(6) 109.8(6) 112.1(7) 109.5(4) 109.5(4) 107.0(4) 110.2(6) 110.2(6) 110.4(6) 117.0(5) 114.8(5) 91.8(5) 110.3(6) 110.2(6) 111.4(6) 111.1(4) 110.1(5) 101.3(4) 112.8(8) (11) 13.19.13 111. (1) 104.2( ) 105.6( ) 116.9( ) 111. (1) 109. (1) 110. (1) 115.3(8) 113.6(7) 95.3(8) 110. (1) 110. (1) 113. (1) 109.6(8) 110.6(9) 102.5(8) 110. (1) 110. (1) 114. (1) 118.5(8) 114.3(8) 88.8(8) 112. (1) 112. (1) 110. (1) 114.0(8) 112. (1) 98.3(7) 108. (1) (III) £19.13 110.9(6) 101.2(4) 102.0(4) 124.5(5) 108.5(6) 109.6(6) 109.9(6) 116.5(4) 114.9(4) 93.6(4) 109.4(6) 110.4(6) 111.1(6) 110.8(4) 115.1(4) 99.0(4) 109.8(6) 110.3(6) 111.5(5) 118.4(4) 117.7(4) 87.7(4) 110.2(6) 110.7(6) 109.6(6) 115.0(4) 110.6(5) 95.4(4) 111.1(6) Table A5. 9399.11 015 017 N1 NM N4 lNlO N10 1N13 N43 1N16 1N16 1N1 Bond Angles cont. Atan2 N16 N16 02 888888 1 012 014 015 017 018 ers-.23. 024 024 03 02 06 05 09 08 012 011 015 014 018 017 (I) (II) (III) eagle A_ngl_e sale 110.9(7) 113. (1) 113.7(7) 110.0(7) 112. (1) 110.1(6) 112.3(7) 114. (1) 114.2(6) 115.0(7) 114.1(1) 113.2(6) 112.5(7) 115.1(1) 114.6(7) 113.1(7) 114. (1) 113.9(6) 113.4(6) 114. (1) 115.0(6) 114.5(7) 114. (1) 113.7(6) 115.1(7) 116. (1) 114.2(6) 113.0(6) 110. (1) 115.1(6) 113.6(7) 117. (1) 114.0(6) 113.6(7) 110. (1) 114.6(7) 120.5(8) 119. (2) 114.1(6) 118. (1) 118. (2) 115.9(7) LISTOFREFERENCES 10 11 12 13 14 15 16 17 18 .L. Dye, J.M. Ceraso, M.T. Ink, 8.1’... Barnett, and F.J. Tehan; . 9133. Chen, Soc. 96, 608, (1974). ILOLI F.J. Tehan, 8.1.. Barrett, and J.L. Dye; :1. Am, Chen. Soc. 96, 7203. (1974). 0.3. Pederson; g. 511. Chem Soc. 89, 2495, (1967). 0.3. Pederson; g Am Chem Soc. 89, 7017, (1967). 0.3. Pederson; g. A_m. Chem. Soc. 92, 386, (1970). J.M. Lehnn; Struture andgedim, 16, 1, (1973) J.M. Lehn and J.P. Sauvage; 1. fl- Chan. Soc. 97, 6700, (1975). 3.14. Lehn and 3.9. Sauvage; gm. m 1971, 440. D.J. Cram and 3.14. Cram; Science, 183, 803, (1974). D.J. Cram and J.M. Cram; Accounts Chem. Res. 11, 8, (1978). D.J. Cram; m gimp _I_n_t_. g. m. 25, 1039, (1988). D.C. Tostesonn; E- grpg. 27,1289, (1968). 8.14. Izatt, 0.0. Lindh, R.L. Bruening, J.S. Bradshaw, J.D. Lamb, and 3.3. Christensen; Pure All. Chem. 58, 1453, (1986). 0. Ryba and J. Petranek; 9;. Electroanal. (men. Interfacial glgtmdm. 44, 425, (1973). T. Ogata; mgaimtsu, 25, 598, (1987). 8.8. Jepson and R. DeWitt; l. Inogg. Nucl. Chan. 38, 1175, (1976). 8.8. Jepson and W.F. Evans; _S_ep. Sci. Teclnnol. 22, 1029, (1987). J.L. Dye, M.G. De8acker, and V.A. Nicely; _._I_. _A_m_. Chem. Soc. 92, 5226. (1970). 107 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 35 36 108 J.L.Dye, M.T. Lok, F.J. Tehan, R.B. Coolen, N. Papadakis, J.M. Ceraso, and 14.0. De8acker; Ber. W: P_hy§. 011311. 75, 659, (1971). A. Ellaboudy, J.L. Dye, and P.B. Snith; i. 55. Clem. Soc. 105. 6490, (1983). 8.8. Dues, D.L. thrd, R.H. lining, and J.L. Dye; _J_. .A_m_. Chem. _Sgg. 108, 3534, (1986). J.L. Dye; Egg. 1mm. Chem. 32, 327, (1984). J.L. Dye and 14.6. De8acker; Ann. Rev. EYE.- Chem. 38, 271, (1987). D.J. Mathre and W.0. Guida; Tetraledron Lett. 21, 4773, (1980). A.G.M. Barrett, C.R.A. Godfrey, D.M. Hollinshead, P.A. Prokopiou, D.H.R. Barton, R.B. Boar, L. Joukhadar, J.F. McGhie, and 8.0. Misra; l. _0_. §_. Perkin Trans. _I_, 1981, 1501. A.G.M. Barrett, P.A. Prokopiou, and D.H.R. Barton; _q_. _0__. g. Perkin Trans. _L 1981, 1510. T. Ohsawa and T. Oishi; ,1. Inncl. Phen. 2, 185, (1984). G.P. Pez, LL. Fador, J.E. Galle, R.K. Crissey, and 0.8. Forbes; g. Am. Chem. Soc. 107, 4098, (1985). T. Olsawa, T. Kobayashi, Y. Mizuguchi, T. Saitoh, and T. Oishi; Tetraredron Lett. 26, 6103, (1985). R. Liepins and M. Aldissi; Mol. grygg. Lig. _CEEL 105. 151. (1984). A. Maercker; Egg. Chem. Int. _E_d_. £131, 26, 972, (1987). R.H. Huang, M.K. Faber, B. VanEck,ard J.L. Dye; unpublisled results, this laboratory. R.M. Izatt, L. Eblerhardt, G.A. Clark, R.L. Bruening, J.S. Bradshaw, M.H. Cho, and J.J. Christensen; Maxim g Technolggy, 22. 701. (1987). RM. Izatt, J.S. Bradshaw, S.A. Nielsen, J.D. Lamb, and J.J. Christensen; Chan. Rev. 85, 271, (1985). M.N. Bell, A.J. Blake, M. Schroder, and T.A. Stephenson; :1. Clnan. Soc., Chan. Caumm. 1986, 471. T.A. Del Donnno and W. Roach; ._T_. 533;. Chan. Soc. 99, 8051, (1977). 37 40 41 42 43 44 45 46 47 48 49 51 52 53 109 8.N. Diel, R.0. Haltararger, and A.D. Nornan; l. _An_nn_. Glenn. g. 104, 4700. (1982). .I. Sinyavskaya; Koord. Khim. 12, 1155, (1988). M. Kodama, E. Kim, and S. Yanaguchi;_._1_. _0_. _S_. Dalton, 1980, 2536. M. Kodama and E. Kimura; l. g. g. Dalton, 1978, 1081. M.E. Mister; M.S. Thesis, Michigan State University, (1986). B. VanEck; Ph.D. Dissertation, Michigan State University, (1981). J.L. Dye; g. Phys. Chan. 84, 1084, (1980). B. VanEck, L.D. Le, D. Issa, and J.L. DYE: Inogg. Gran. 21, 1966. (1982). M.T. Lok, F.S. Tehan, and J.L. Dye; 1- Phys. Chem. 78, 2975, (1972). J.L. Dye, M.R. Yenen, M.G. Dague, and 3.14. Lehn; _.1_. Gem. Egg. 68, 1665, (1978). L.D. Le, D. Issa, B. VanEck, and J.L. Dye; ;. Phys. Gnem. 86, 7, (1982). J.L. Dye, M.G. Dague, M.R. Yenen, J.S. Landers, and R.L. Lewis; 1. P_hE. Gem. 84, 1096, (1980). J. Skowyra and J.L. Dye; unpublished results this Laboratory. M.G. Dague, J.S. Lenders, R.L. Iewis, and J.L. Dye; Gem. Phys. Lett. 66, 169. (1979). R.G. Patel and M.M. Gnaudhri; Thermochimica Acta. 25, 247, (1978). Z. Osawa; Bull. Gem. Soc. M 38, 1881, (1965). H.E. Kissinger; Anal. Gen. 29, 1702, (1957). M.E. Mister and J.L. Dye; 9;. _A_m;. Gan. Soc. in press. 55 57 58 59 61 62 63 65 66 67 68 69 7O 71 72 73 11.0. Borgstedt am C.K. mm; Amlied Chemistry githe 110 Alkali Metals, Plenum: New York, p. 166 1987. '1'. m and 0.J. Kleppa; ,1. Clan. Phys. 40, 46, (1964). Handbook QLClanistgy and _P_hEics-, Weast, R.C., £21.; 62 nd. editian, CRC Press Inc. (1981). R.L. McKisson am L.A. Branley; ._7_. AIL. Clan. Soc. 73, 314, (1951) . K. Moeggenborg and J. Papaiouannou; unpublished results this laboratory. Spectra obtained by L.E. Hill and J. Kim A. Ellaboudy, M.L. Tinkhaxn, B. VanEck, J.L. DYE. and P.B. Smith; .1. PhYS. Chan. 88, 3852, (1984). A. Ellaboudy and J.L. Dye; _J_. m. Reson. 66, 491, (1986). M.L. Tinkham and J.L. Dye; J. (1985). 8.8. Dawes, A.S. Ellaboudy, and J.L. Dye; l. 511: Clan. Soc. 109, 3508, (1987) . M.L. Tinkham, A. Ellaboudy, J.L. DYE. and P.B. Snith; 9;. _P_lnys_. Clan. 90. 14. (1986) . in; Clan. Soc. 107, 6129, J.E. Hertz and J,R. Bolton; "Electron §pin Resonance: Elanentaty 'llneory and Mlications", MoGraw Hill (New York) 1972 . W. Bell Flower, "P_hy§ics 5g. Color Centers", Academic Press, (New York) (1986) . Spectra obtained by A. Ellaboudy and DH. Shin. R. Catterall and P.P. wards; Adv. _igMolecular Pelaxation ant} Interaction Process, 13, 123, (1978). Fitting was done by J. Kim and D.H. Slnin. A.S. Ellaboudy, C. Bender, 9.7. Baboock, M.E. Kuclanneister, an! J.L. DYE; to be published. D. Issa, A. Ellaboudy, R. Janakiraman, and J.L. Dye; _J_. m Clan. J.S. Ianders, J.L. Dye, A. Stacey, and M.J. Sienko; l. m. Clan. 88. 3847. (1984) . 85, 1096, (1981) . 74 75 76 77 78 79 80 81 82 83 84 85 111 A. Ellaboudy; Ph. D. Dissertation, Michigan State University, (1984). R.H. Huang, D.L. Ward, M.E. Imclanmeister, and J.L. Dye; ,'_I_. M. Clan. Soc. 109, 5561, (1987). 8.3. Dares; Ph. D. Dissertation, Michigan State University, ( 1986). D.L. Ward, M.E. Kuclanneister, and J.L. DYE; Suhnitted for Publication. D.T. Craner arxi J .T. Faber; International Tables for X-ray W, Vol IV, (1974), Table 2.2 8. Birmingham: Kynoch Press. G.H. Stout and L.H. Jenson; x-ray Structure Determination: _A_ Pratical Guide, manillan Publishing Co.; New York, (1968). R.D. Shannon and C.'I'. Prewitt; Acta. Crystallggg. 1969, 825, 925. 12.11. Huang, S.B. Dawes, D.L. Ward, and J.L. Dye; Unpublisle results this laboratory. R.H. Huang and D.L. Ward; unpublished results this laboratory. F. chtle and E. Weber, m. Clan. Int. 33. 13, 814, (1974). R.M. Izatt and J .J . Christensen; Synthetic Miltidentate m (1978). J.C. Tlanson; Electrons _igLifld Amnonia, Claredon Press, Oxford, (1976).