l LIBRARY THESE; Michigan State University L -. e 4) This is to certify that the thesis entitled Potassium l8-Crown-6 Sodium: a Study of Optical and Electrical PrOperties presented by Mary L. Tinkham has been accepted towards fulfillment of the requirements for M.S. degree mineralogy Major professor Mew/7} x982... 0-7639 MS U is an Mmau've Action/Equal Opportunity Institution 'bvisSIyJ RETURNING MATERIALS: Place in book drop to LJBRARJES remove this checkout from .J-lltfll-IL your record. FINES will be charged if book is returned after the date stamped below. ___- POTASSIUM lB-CROWN-G SODIUM A STUDY OF OPTICAL AND ELECTRICAL PROPERTIES by Mary L. Tinkham A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Chemistry 1982 ABSTRACT Potassium lB-Crown-G Sodium: a Study of Optical and Electrical Properties by Mary L. Tinkham Potassium lB-crown-G sodium aIkalide crystals of stoi- chiometry K-18C6-Na were prepared from equimolar solutions of potassium metal, sodium metal and complexing agent, lB-crown-é. Optical and EPR spectra suggest that K-18C6—Na is a simple alkalide with Na_ as the anion, however, the possibility of Na+18C6 K"cannot be ruled out. Three separate preparations of K-18C6-Na resulted in two solvent free batches, and one batch of crystals which contained ~l4 mole percent isopropylamine. The solvent-free crystals were stable up to +48°C, whereas the solvent- containing crystals melted at room temperature. Optical spectra of thin films formed from solutions of either type of crystal in methylamine by fast solvent evaporation had a single narrow peak at 14,0fl0 cm"1. The absorption maximum Mary L. Tinkham lies between the maxima observed for Na' and K- in ethylene- diamine and, therefore, does not clearly indicate the nature of the anion. The observed absorption maximum could either be a red-shifted Na' peak, or a blue-shifted K'peak, but the narrowness of the peak suggests a simple alkalide rather than a mixed system. EPR spectra show a weak signal line, presumably from trapped electrons (F-centers), with a 9 value of 2.06259. Hyperfine interaction of the electron with either the Kq'or Na+ nucleus (but not both) is observed at higher temperatures. DC conductivity measurements indicate that K-18C6-Na is an extrinsic semiconductor with an apparent band gap of 1.03 eV. Photoconductivity studies showed no photoresponse in the region BEG-325 nm but the method for these measurements is still in the developmental stage. A recrystallization procedure was developed to obtain better crystals for future single crystal studies. to my mother 11 ACKNOWLEGEMENTS I would like to express my gratitude to Professor James L. Dye for his constant guidance and support during this work. Of the scores of other people who have helped me along the way and to whom I am indebted. I would like to particularly thank Brad Van Eck, Drs. Long D. Le and Dheeb Issa fer their tutelage and help. Also. my colleagues in the research group, Steve Dawes, Ahmed Ellaboudy, Margie Faber. Dr. Stephan Jaenicke and John Papaioannou, have my heartfelt appreciation for their aid. discussions and morale support. The technical and clerical staff deserve thanks also. especially glassblowers Keki Mistry, Andy Seer and Manfred Langer for their excellent and speedy service, and to Naomi Hack for her clerical help and general concern. A special "thank you" goes toward Kermit Johnson for his help and patience in the qualitative NMR work. Research support from NSF Grant DMR-79-21979 is gratefully acknowledged. iii TABLE or CONTENTS LIST OF FIGURES LIST OF TABLES I. II. INTRODUCTION A. The Nature of Semiconductors B. Alkali Metal Anions in Solution C. Stability of Alkali Metal Anions in Condensed Phases D. Properties of Alkalides EXPERIMENTAL A. Reagents l. Metals 2. Solvents 3. Complexing Agent- B. Glassware Cleaning C. Sample Preparation and Handling D. Analysis Techniques 1. Hydrogen Evolution 2. pH Titration iv page vi viii 12 16 17 21 21 21 21 22 22 23 26 26 28 III. Iv. V. 3. Flame Analysis 4. Proton NMR E. Instrumental Techniques 1. 2. 3. 4. 5. Optical Spectra DC Conductivity Recrystallization Photoconductivity EPR POTASSIUM 18-CROWN-6 SODIUM A. Elemental Analysis B. Solubility Studies and Recrystallization Attempts ’1 C. Optical Spectra D. 'Powder DC Conductivity E. Photoconductivity F. Electron Paramagnetic Resonance G. Melting Point CONCLUSIONS AND SUGGESTIONS FOR FURTHER WORK A. Conclusions B. Suggestions for Further Work REFERENCES 29 29 30 3O 33 34 37 41 42 43 48 49 S3 56 57 62 63 63 64 65 Figure 10. 11. 12. TABLE OF FIGURES Range of electrical conductivity at room temperature. Band diagram for metals, insulators, and semiconductors. Energy diagram for extrinsic semi- conductors. Temperature dependence of resistivity for doped germanium. Optical spectra of alkali metals in ethylenediamine. Apparatus for the synthesis of K-lBCG-Na. Optical spectra apparatus for thin films Recrystallization apparatus. Photoconductivity apparatus designed by M.R. Yemen. Photoconductivity apparatus designed by L.D. Le. Optical spectrum of K-18C6-Na from equi- molar methylamine solutions. Optical.spectrum of K-18C6-Na films from solutions of crystals in methylamine. page 14 24 31 35 38 40 50 52 13. 14. 15. 16. Current vs. voltage for K-18C6-Na powders. Dependence of log R on l/T for K-18C6-Na. EPR spectra of K-18C6-Na powder at +38.3°C. EPR spectra of K-18C6-Na powder at —101.l°C. V11 54 55 59 61 TABLE OF TABLES Table page 1. Results of elemental analysis of K-18C6-Na. 44 2. Comparison of 18C6 to sodium acetate proton ratios determined by mass and by NMR. 47 viii I. INTRODUCTION The successful synthesis of a crystalline salt which contained alkali metal anions (i.e.,a1kalide) opened a new field of alkali metal chemistry. The first "alkalide", Na+C222 Na': was prepared in 1974 and is the most fully characterized (1-4). Since that time, numerous other "alkalide" systems have been made by using cation complexing agents such as crown ethers, (called n-crown-m where n is the ring size and m is the number of (-CH2CHZO-) in the ring). and cryptands, (called Cjkl where j,k,and l are the number of (CHZCHZO-) in each of the three Strands connecting two tertiary amine nitrogens) (5-16). One of these systems, potassium lB-crown-6 sodium, had been synthesized but the analysis and characterization were incomplete (11). Due to the relative availability and low cost of the starting materials, this salt is the most likely candidate observed so far which might be practical for use in synthesis. This thesis describes the synthesis and characterization of the K 18-crown-6 Na system. I.A. The Nature of Semiconductors When an investigator encounters a new substance, she will probably try to classify it according to her own field of interest. If the interest lies in the electrical properties of matter, the major categories are conductor, insulator and semiconductor. Conductors in the solid state are often metals or semimetals which can conduct appreciably at all temperatures and have negative temperature coefficients of conductivity. On the other end of the spectrum, insulators have high resistances over large temperature ranges. Semiconductors have intermediate conductivities. which are highly temperature dependent: the conductivities increase exponentially with temperature. Figure 1 shows the range of conductivities observed for these categories at room temperature (12). To describe the differences in conductivity of insulators. semiconductors and metals, a simplified presentation of band theory may be used. The hydrogen molecule is an example of two equal ground state atomic energy levels which combine to form two molecular energy levels. Similarly,a molecule with n atoms will form n molecular energy levels for each atomic level. In sOlids, the density of atoms is so great that the discrete energy levels of molecules give way to energy continua; i.e., bands in the bulk state which can accept 2n electrons. There may remain energy separations between the bands corresponding roughly to differences in the atomic energy levels, but the differences between discrete levels within a given band are extremely small (13). The highest completely occupied energy level is the valence band. The band immediately higher in energy than the highest completely occupied band 1010 h " Copper Metals and I- Iron . 6- Bismuth Semi-metals 10 *- ---_---_-—_—-_--------—fi . Intrinsic indium antimonide‘ - Intrinsic tellurium 102 - Transistor—grade silicon 0' (ohm"meter") Intrinsic germanium , t Semiconductors 10"2 " . . .. Intrmsrc Silicon l- Crystalline selenium IO‘6 - : Pure cadmium sulphide E Pure iodine IO“10 *- Room Temperature Bulk Electrical Conductivity .. Polyethylene Insulators and _ Plate glass . insulating .- Amorphous selenium photoconductors ‘0.” : Paraffin wax . Pure diamond p lor'lfl Figure 1. Range of electrical conductivity observed at room temp- erature, with some typical members of the three classes of conductors (with permission from J.S. Blakemore, "Solid State Physics", W.B. Saunders, Co., 1974). is called the conduction band. It is partially filled in conductors and empty in semiconductors and insulators at very low temperatures. The energy difference between the filled valence band and the empty conduction band is known as the energy gap (E9). Electrons in a metal can easily move in response to an electric field to carry current. because available empty states lie at energies only infinitesimally above the filled states. However, for semiconductors and insulators the ability to conduct electricity depends on the ability for electrons to be excited from the valence band to the conduc- tance band. Electrons in a completely filled band have no empty states within the band to which they can be excited. One method of excitation from a filled band to an empty one is through the addition of thermal energy. The temperature dependence of the conductivity,a , is given by a = A exp(-Eg/2kB T) where A is the conductivity at infinite temperature and k3 is the Boltzmann constant (13). For insulators, the energy gap is very much greater than kBT and appreciable conduction does not occur at any accessible temperature. In semiconductors. the conduction band contains only a few electrons as long as Eg>>kBT. As the temperature increases and k3; becomes an appreciable fraction of E more 9 I electrons can be thermally excited to the conduction band. Figure 2 shows typical band diagrams illustrating the differences between metals, semiconductors and insulators. Conductivity in semiconductors arises from electron- hole pair formation where the electron in the conduction _ band and the hole it leaves in the valence band can both contribute to the conductivity. The concentrations of holes and conduction electrons are equal. This type of semi- conductor is called an intrinsic semiconductor. Intrinsic semiconductivity exists only in extremely pure substances. Much more common are extrinsic semiconductors which have contributions to the conductivity from impurities. Extrinsic semiconductors are sometimes referred to as doped semiconductors, since the conductive properties can be controlled by the type and level of deliberately added impurities. The model for the impurity contribution is given in Figure 3. The impurity has an energy level which lies somewhere in the energy gap. If this level is empty and is close to the valence band, electrons can be excited from the valence band into the acceptor impurity band leaving holes. The observed conductivity at thermal energies much less than the intrinsic value of E arises from the mobility of the holes in the valence band. This type of semiconductor is known as a p-type semiconductor. On the other hand, if the impurity level contains electrons and lies close to the conduction band. electrons can be donated from the impurity band to the conduction band. This is called an n-type semiconductor and the conductivity is 3 \ \\\X\V \\\\\\\\\ m \\< Aws (a) metal (b) semiconductor (C) insulator Figure 2. Band diagram illustrating the difference between (a) metals, (b) semiconductors and (c) insulators. E (7/1/72. 7/////// Figure‘3. (a) (b) Energy diagram for extrinsic semiconductors: (a) p-type semiconductors where electrons can be excited from the host lattice valence band to the accepter center, EA' (b) n-type semiconductors where electrons D can be excited from the donor energy level, ED to the conduction band. mostly due to electrons rather than positive holes. In either case, the contributions to the conductivity from the electrons in the conduction band and holes in the valence band are vastly different. If the temperature is increased a state is reached where nearly all impurity states con- tribute to the conductivity. This maximum current is refer- red to as the saturation current. Further temperature in- crease can still create new electron-hole pairs in the‘host lattice and the conduction returns to instrinsic behavior. Figure 4 shows the temperature dependence of electrical resistivity (l/o ) fer germanium doped with different concentrations of antimony (12). The change in lepe occurs in the temperature region at which the impurity band has donated all its mobile carriers and intrinsic behavior predominates. Therefore, most of the low temperature conduction is from antimony "impurities" in the germanium. There are several types of impurities or imperfections that can contribute to the conduction of semiconductors, or create conductivity in insulators. Donor levels can be created by anion vacancies in the crystal lattice while cation (or metal) vacancies produce acceptor centers. A common imperfection in ionic solids such as alkali halides is an F-center or color center. F-centers are crystal defects where an anion site is occupied by an electron. These electrons often lie at energy levels less than that of the conduction band but higher than the valence band (14) and can be thermally excited to the conduction band. Resistivity p = (1/0) [ohm-meter] Figure 4. 106 I I I r r __ £3 _‘ £2 10" — '3 d ' F “" L: -a i 102 L- - . F2 1 I—' L211 , ? - ‘ - c : 3 : ..... 3 : ‘4 CI / '7 yr 5'": Ho 0 2 4 6 (lo/T) (K'l) Temperature dependence of resistivity for germanium doped primarily with antimony, in amounts ranging from 7x1021 m“3 for sample E3 to 2x1023 m'3 for sample Fl (with permission reference 12). 10 Therefore, F-centers are donor type impurities. Metals and insulators differ also in their optical properties. The absorption of light by metals, insulators and semiconductors can be explained qualitatively by band theory. For metals, free electrons in the conduction band absorb incident radiation by making transitions to higher energy levels. These levels form essentially a continuum so that absorption occurs from the ultraviolet (the so called plasma edge) through the infrared region. The high reflection at essentially all visible wavelengths gives the characteristic metallic luster. Most semiconductors do not absorb radiation until the critical incident energy is reached. The absorption is abrupt and the absorption coefficient increases rapidly with increasing energy until the solid is essentially Opaque to radiation of higher energy. The wavelength at which absorption begins is often called the absorption edge and corresponds roughly to the energy gap of the semiconductor (13). In the presence of F-centers and other impurities, absorption maxima are seen at longer wavelengths than the absorption edge. Generally, the spectra of F-centers are recorded at low temperatures (<26 K) to prevent thermal ionization of these impurity states which increases the spectral bandwidths. The most noted contribution of imperfections is in insulators where the absorption edges caused by trapped electrons produce intense colors which give this defect its name of F-center (Farbe means color in German). 11 The conduction of electricity in semiconductors can be enhanced by excitation with light at temperatures below those where the current is due to thermally created carriers. This phenomenon is called photoconductivity and is observed in insulators as well as semiconductors (14). The photoresponse spectrum often closely resembles the optical spectrum with maxima at the same wavelengths. The presence of imperfections such as trapped electrons can create bands in the photoresponse spectra just as in the optical spectra. Photoconductivity in semiconductors makes these materials suitable for radiation detection (gamma rays, electrons and alpha particles). Photoionization creates electron-hole pairs which are easily detected electronically and give information not only about the amount but also the energy of the radiation. The final property of solids discussed in this section is the electron paramagnetic resonance (EPR) observed for unpaired electrons in a magnetic field. EPR can provide information about the basic structure of defects and can also indicate the distribution of the charge over the surrounding nuclei (14). The EPR spectra of F-centers in most alkali halides consist of single broad Gaussian lines that have linewidths ranging from 47 to several hundred Gauss, without further structure. However, in some cases (LiF,NaF, RbCl and CsCl) the broad EPR signal is distinctly structured with a discernible number of components (15). The electronic configuration for each ion in an alkali halide is 12 that of the nearest group VIII element and all electrons are paired. Therefore, no EPR signal is expected in a pure salt. The signal observed is due to the F-centers and other impurities within the crystals. The additional fine structure is hyperfine interaction between the electron and the surrounding nuclei (16). Doped crystals such as donor silicon (containing phosphorus which contributes electrons to the lattice) or acceptor silicon (with boron which removes electrons from the lattice) have been used to study the EPR spectra of transition metals dissolved in the lattice. EPR spectra of the conduction electrons of some pure metals have also been observed but not for heavy metals, presumably due to the large width of the line. The intensity of the EPR signals for metals are temperature independent as expected, but semiconductors show a temperature dependence of the signal intensity (l6). I.B. Alkali Metal Anions in Solutions Since the discovery of potassium ammonia solutions by Weyl in 1864 (17) ,metal ammonia solutions have been widely studied. Alkali metals dissolve in ammonia and certain amines to give blue solutions with one or two Optical ab- sorption bands. One peak is metal, solvent and temperature dependent. This band is assigned to solvated metal anions, M‘. The second band, which appears at 1200 to 2000 nm, is 15 metal independent and is attributed to solvated electrons, eSOhr Figure 5 gives the spectra of Li,Na,K,Rb and Cs in ethylenediamine (EDA) (18). The peak for lithium is actually that of egolvas confirmed by pulse radiolysis studies (7). The infrared shoulders on the low energy side of the peaks of K‘, Rb. and Cs-Fare attributed to egdhrwhich arise from the following equilibrium: M"—--M+ + 2e‘golv (1) Evidence in support of the assignment of the metal dependent peak to M"is the formation of Na- by pulse radiolysis (3,19), the Faraday effect (20), the oscillator strength of f"2 (21) and 23Na NMR spectra (22,23). The study of alkali metal solutions before 1970 was re- stricted by the low solubilities of the metals in most sol- vents. Ammonia is "too good" as a solvent in that the dis- sociation reaction, equation (1), lies far to the right (7). To expand the number of solvents, complexing agents such as crown ethers, first synthesized by Pederson in 1967 (24), and cryptands, developed by Lehn and co-workers in 1969 (25) were used. The solubility of ordinary alkali salts had been observed to be greatly enhanced by the presence of these complexing agents. Crown ethers are monocyclic ethers which form a complex with the alkali cation in the center of the ring. Cryptands are three dimensional complexants in which the cation is surrounded by three strands which contain 14 .mcoausaom m:«6mwp ocoaanOIHmuoE «meac mo muuumdm HmOfiudo .m muzwfim MIO_ .A TEUVNM ON 0. O_ . m _ m, l _ 4.. _ t am. 8 1o._ oz 3 _ _ _ mo mo no mo 9 . m... 6.. 2 2.12 15 ether linkages between two tertiary amine nitrogen atoms. These complexed cations are readily solvated by polar solvents such as amines and ethers. When added to alkali metal solutions in the presence of excess metal in such solvents as ethylamine and tetrahydrofuran, the solubilities were increased by many orders of magnitude (26-28). The equilibria describing alkali metals in solutions (3) are: 214(8) , M++ m“ (1) M’- v -—I- M(S)+ eSolv (2) M(s) :" Mt+ ensolv (3) in which M(S)is a contact ion pair or "monomer" present in low concentrations. The addition of a complexing agent, C, introduces a fourth equilibrium: \ M+ + c ____.. M+C (4) which drives equations (1) and (3) to the right. Cryptands complex the cations more strongly than do crown ethers and therefore have a greater effect on the equilibrium. The range of solvents available for dissolving metals was thereby greatly expanded (27,28). Due to.the cation selec— tivity brought about by variations in the cavity size of the complexing agent, as well as differences in thermodynamic stability of the various alkali metal anions, selectivity of anions in mixed metal solutions is possible. For example, 16 solutions which contain equimolar amounts of Na,K and C222 give only the Na' absorption band indicating that all the potassium is complexed (9). Also, the 23Na NMR spectrum of the solution indicated only Na- present with no peak corre- sponding to Na+C (6). By using X-ray crystallography it was determined that 18-crown-6 has a cavity that will accommo- date both Na+ and K+ (29) but is selective towards K+ (30). The larger cations, Rb4'and Cs+, are unable to enter the cavity but can form both 1:1 complexes in which the cation lies outside the plane of the ring and sandwich complexes in which the cation lies between two molecules of crown ether. The formation of such complexes was detected by alkali metal NMR studies (31). I.C Stability of Alkali Metal Anions in Condensed Phases The increased concentration of the alkali metal anion and/or solvated electrons in solution brought about by the use of complexants for the cations made it conceivable that crystalline salts of these unusual anions could be prepared. In fact, when a ~0.2 M solution of Na in ethylamine with C222 present was cooled from +10 to -20°C, Na+C222 Na- crystals appeared spontaneously (32). This spontaneous and reversible process would occur isothermally when C222 was added to a saturated Na solution over solid sodium. Therefore, Na+C222 Na- is thermodynamically stable with respect to Na(s)and C222solv . By examining the enthalpy of l7 formation of the hypothetical ionic solid,M+M?s), from the metallic solid, it is easy to see why M"in the solid phase can be thermodynamically stable. The estimated energy required for: 2M —’ M+M- (S) (s) is less than +15 Kcal/mol because of the positive electron affinity relatively low ionization potentials of the gaseous alkali atoms, and the relatively low lattice energies of the solid metals. Lattice energies were estimated by assuming that the structure of the M+M- salt would be the same as the corresponding alkali halide and by using the interatomic distance of the metal as the interionic distance in the salt. By means of Born-Haber cycle calculations, the enthalpy of formation of M+M(;)was estimated to range from 0.4 Kcal/mol for Li to 13.5 Kcal/mol for Cs (32). Since these values are only slightly endothermic, stabilization of M+ by a complexant could thermodynamically allow the forma- tion of an alkali metal anion salt. The complexation of a1- kali metal cations by crown ethers is thermodynamically fea- sible (29). Therefore, a variety of alkalides are thermo- dynamically stable (7) and have subsequently been made (10). I.D. Properties of Alkalides Solids which contain alkali metal anions have been prepared by three methods. The first is Slow solvent evap- 18 oration to yield powders with the same stoichiometry as the solution. By decreasing the ratio of metal to complexing agent, electrides (salts that have an electron as the anion rather than the alkali metal anion) could also be made in this manner. The second method is direct vapor deposition of the metal and complexant (33). The third method (which was historically the first method used) is gradual cooling with or without a change in the solvent composition to pre— cipitate crystals (1,2,11). These crystals were not neces— sarily of the same stoichiometry as the solution (11,34). Rapid evaporation of primary mono-alkylamine or ether solutions which contain M+C M"where C can be a crown ether or cryptand, yield blue or gold colored films (4,6,35-38). Solutions of Na+C222 Na‘ give films with a strong absorption band at 15,000 cm‘l and a pronounced shoulder at 18,900 cm"1 Also present is a small peak at 24,500 cm-j. Decomposition of the films resulted in a loss of all spectral features. Films from solutions which contained K', Rb' and Cs' were also prepared but are considerably less stable. These films lacked the high energy shoulder and the distinct peak at 24,500 cm'd. The absorption bands are at 11,900, 11,600 and 10,500 cm'4 for K',Rb' and Cs" respectively. These locations correspond to the M" ‘bands in ethylenediamine at 12,000, 11,200 and 9800 cm'”. Na" has a band at 15,400 cm"1 in EDA but no other bands or shoulders (18). Films prepared from solutions containing excess potassium and sodium metals with c222 showed only the major peak of Na“ at 15.1% cm" (6). 19 Other mixed alkali systems show little change in the absorption band locations for the M“ positon in the M+C M’ films. Cs+18C6 Na- has a sharp peak at 16500 cm'1compared to that of 16,000 cm“1 observed for Na+18C6 Na-in the presence of excess methylamine. In the case of Cs-Rb-18C6 where neither cation can enter the crown ether cavity, two peaks appear with one at 11,500 cm-‘which extending over a 1000 cm" range. The center is at 12,000 cm 'and corresponds to the Rb- peak observed in the Rb+18C6 Rb' spectra. The second band occurs at 9000 cm'"1 and was assigned to trapped electrons. These bands were temperature dependent in that the band width increased at lower temperatures. The presence of Cs- could not be ruled out due to the broadness of the band (11). Films prepared from a solution containing equi- molar amounts of K, Na and 18C6 were observed to have a strong band at 13,300 cm"with an infrared shoulder at 9000 cm'“. This was thought to be either a Na' peak with a 2700 cmT'red shift or a blue shift of 1100 cmf'for K“ (35). The films of K-l8C6-Na showed no change with time or temperature and because Of the broadness of the peaks were first thought to contain a mixture of K'and Na‘. Alkalide systems have electrical conductivities which correspond to semiconductors. \Reported band gaps from powder conductivity measurements are 1.7 and .93 eV for Cs+18C6 Na- and K-18C6-Na respectively (11). Na+C222 Na' has the largest band gap of ~2.4 eV (9). Rb+18C6 Na- has also been prepared and has a band gap of ".89 eV (39). In 20 most cases the solids behave like doped semiconductors with the exception of Na+C222 Na‘ which appears to be intrinsic. True alkalides are expected to have all electrons paired since the outer orbital has an 82 configuration. Samples of NaIC222 Na’ are indeed diamagnetic and show no appreciable EPR signals. Crystals of Cs+18C6 Na- are also diamagnetic. Rb+18C6 Na- shows a weak EPR signal with the presence of fine structure (39). Magnetic studies had not been performed on K-18C6-Na before this investigation. A new study of the K-18C6-Na system has been made in this thesis research in an attempt to resolve the nature of the anionic species (K- or Na“). The synthesis has been modified from that previously used (11) and is described in detail in Chapter II. The conductivity, optical spectra and magnetic properties have also been examined and are described in Chapter III. II. EXPERIMENTAL II.A. Reagents II.A.l. Metals Potassium and sodium (Alpha Ventron Products, total purity for each 99.95%) were obtained under argon in breakseal ampules. The metal was distributed into smaller tubing following the procedure outlined by Issa (l0). Desired quantities of metal were obtained by premeasuring the inner diameters of the tubing and isolating lengths to give the corresponding volumes. II.A.2. Solvents Methylamine (98% pure, Matheson)(MA) was stirred over calcium hydride at -20°C for 1 hour, frozen in liquid nitrogen and pumped to < 2x10'fi torr. This stir-freeze-pump cycle was repeated until no gas evolution was observed during stirring. The MA was transferred to bottles containing NaK3 alloy, frozen, pumped and stored overnight before transferring to another NaK3 bottle. After the solution maintained the characteristic blue color of solvated electrons for two consecutive cycles, the MA was tranferred to a heavy walled storage bottle. The MA was then subjected to several freeze-pump-thaw cycles until the intial pressure at pumping was less than lxl0—‘ torr. Isopropylamine and n-Pentane: The procedure for 21 22 isopropylamine (Eastman Organic Chemicals) and n-pentane (J.T. Baker Chemical Co.) was the same as for methylamine with the addition of benzophenone as an indicator of , dryness. ' Diethyl Ether (Ethyl ether, anhydrous, Mallinckrodt,Inc) was treated as isopropylamine. Final storage was over NaK with no benzophenone. Trimethylamine (Matheson) was purified following the procedure of methylamine. II.A.3. Complexing agent 18-Crown-6 (18C6 or IUPAC:1,4,7,10,13,16 hexaoxacyclo- octadecane, purchased from PCR,Inc. and Parish Chemicals) was first recrystallized from acetonitrile then vacuum sublimed at 60°C. The 18C6 was stored in vacuo in the dark until use. II.B. Glassware Cleaning All glassware was first rinsed with an HF-cleaner solution (5% HF(28M), 2% detergent, 33% HN03(16 M) and 60% deionized H20 by volume) then immediately rinsed six times with deionized water. The glassware was filled with freshly prepared aqua regia (3HC1:1HNO3) and left overnight. The glassware was emptied, rinsed six times with deionized water followed by six rinses with conductance water (house deion- ized water which has been further deionized with a Crystalab Deeminizer and distilled through a high reflux ratio column to less than 1ppm impurity) and dried in a 450°F oven for 23 several hours. Each apparatus used in synthesis was rinsed an additional time with boiling conductance water before drying. II.C. Sample Preparation and Handling Since the deSired crystals contain sodium, Pyrex apparatus was suitable for synthesis. Crystals were prepared in a vessel such as shown in Figure 6. Two glass tubes which contained measured amounts of K and Na respectively, were scribed with a glass knife and placed in side arm A. The side arm was capped with a sealed Pyrex tube connected with flexible heat-shrink Teflon tubing at B. 18C6 was weighed under flowing nitrogen and introduced into the apparatus through side arm C which was then quickly sealed. The apparatus was immediately evacuated to 2x10"5 torr to prevent water absorption by the crown. At this point the system was left under static vacuum overnight with the crOwn chamber,D, covered with aluminum foil. After a quick check on the vacuum, the vessel was removed from the line and the metal-containing vials were shaken into the heat-shrink tubing at B and broken. The metals were then distilled into the metal chamber,G, making seals at E and H as they were distilled past these points. The system was brought to less than 10-5 torr and cooled to -40°C in an isopropanol-dry ice bath. Methylamine was distilled into chamber D and the crown was dissolved. The solution was 24 .mZIoUw~ Ix LO mamosucam can now maumumdd< n— .c ouswfim 25 poured through frit J, into G where the metal mirror was dissolved. Care was taken that the bath temperature did not exceed -40°C.1 The solution was poured back into D and the MA was distilled into a waste bottle to leave a damp film. Isopropylamine was then introduced to make a concentrated solution in which the film from the methylamine evaporation could dissolve without the appearance of crystals at -40°. The system was then cooled slowly to -78°C and left for several hours. Attempts to make better and larger crystals were made by raising the temperature to -50°C where only a few crystals were present and again lowering the temperature slowly to -78°C for several hours. However, these attempts did not Succeed in producing larger crystals. After the formation of crystals had stopped, the mother liquor was poured to chamber G, frozen in liquid nitrogen and a seal under dynamic vacuum was made at K. Diethyl ether was then introduced as a washing solvent and the crystals suspended in ether were transferred to chamber L. By distilling ether from D to L and decanting liquid to D, the crystals were washed six times. The ether was then frozen in liquid nitrogen and the system was pumped to l0_5torr or less. During the pumping, the crystals were allowed to warm to 0°C for several minutes to insure complete solvent removal. A vacuum seal-off was made at M and the apparatus cooled. Thei crystals were distributed among the storage tubes N by in- verting and rotating the apparatus. After sealing the in- dividual tubes, the samples were stored in dry ice. 26 II.D. Analysis Techniques To determine the stoichiometry of alkalide and electride systems, a thorough analysis scheme has been developed. The crystals were first reacted with water to yield a basic solution and hydrogen gas. The volume of the gas is directly proportional to the amount of reductant. The decomposed material was then divided into two parts: one for pH titration and the other for alkali metal flame absorption spectroscopy. The solution from titration was further analyzed for crown content by first drying it, dissolving the residue in D20 and using quantitative 1H NMR. In addition, crystals were directly dissolved in D20 to check for solvent content. This procedure gives three checks for metals and one for crown. II.D.l. Hydrogen Evolution Care was taken that the sample decomposed in this analysis step did so only according to the reaction: K-18C6-Na + 2H,o —— K+ + Na+ + 1scs + on‘ + H2 To prevent thermal decomposition, the apparatus used had a large surface area so that the crystals could be spread out. A scribed sample tube was introduced into a large side arm which was sealed by a cap connected by flexible heat-shrink Teflon tubing. After evacuation to "10'"5 torr, the sample tube was moved to the heat-shrink 27 tubing. To insure that all the crystals fell into the working part of the apparatus, the heat-shrink tubing was partially compressed to hold the sample tube and the system was shaken until all the crystals were present in the bottom of the tube. The sample was then broken and both parts of the tube were allowed to fall into the reaction chamber. The crystals were spread over the bottom of the apparatus. This part of the vessel was kept in dry ice to prevent thermal decomposition. A vacuum seal-off below the Teflon portion of the tube was then made and the apparatus was attached to a vacuum system for hydrogen collection. The entire system was evacuated to 10"5 torr. This system consists of a gas burst and a modified Toeppler pump and is described elsewhere (38). The conductance water used for the decomposition was first degassed by three freeze-pump-thaw cycles and then allowed to be distilled into the crystal containing vessel. The water immediately froze upon condensation on the walls above the crystals. The vessel was warmed‘spot by spot to melt the ice and allow only a small amount of water to contact the crystals at any time. When the reaction began, a dry ice-isoprOpanol bath was brought up around the chamber to slow or halt any further reaction. By repeating this thawing-quenching procedure, the crystals were slowly destroyed without thermal decomposition to yield a white residue. At this time, more of the water was rapidly distilled onto the decomposed crystals to form a solution and to insure complete reaction. 28 The hydrogen evolved was pumped through a double liquid nitrogen trap into a tube of known volume by means of a manual Toeppler pump. After several cycles, the pressure of the H2 was measured by noting the height of a column of mercury opened on one side to the H2 gas and on the other to the atmosphere. The endpoint was determined to be the pressure at which five additional pump cycles did not change the mercury level. By noting the temperature at the H2 chamber and the atmospheric pressure, the ideal gas law could be employed to calculate the number of moles of H2 evolved. II.D.2. pH Titration The residue from the H2 evolution was dissolved in a known amount of HCl and conductance water to give a solution of pH ~3. The amount of H2 evolved was used to estimate the number of equivalents of base present. To prevent absorption of CO2 by the residue, the vessel was opened under a nitrogen atmosphere. This solution was divided into two portions: one for flame analysis and another for pH titration. The pH titration portion was then titrated with NaOH solution freshly standardized with potassium hydrogen phthalate. To prevent COz absorption by the base, the titration buret has a glass sheath allowing dry nitrogen gas to flow over the solution. The end point was determined by using a digital pH meter (Orien Research, model 701A) and a Corning electrode (catalog No.476050). The solution was titrated to the endpoint, then past the endpoint, and back- 29 titrated with standardized HCl solution several times to determine the correct endpoint. II.D.3. Flame Analysis The determination of the amount of potassium and sodium present in the sample was made with a Jarrell-Ash Atomic Absorption/Flame Emission spectrophotometer. Only the flame emission feature was used. The second portion of the solution from the H2 evolution step was divided and diluted to give solutions of ~50 ppm fOr K+’and Na+ . It is extremely important that the solution used for the flame emission analysis is not the same as that used in the pH titration, since the calomel reference electrode adds K+ to the system (not to mention the Na+ from the NaOH). Standard solutions of each metal in the 20-100 ppm range were used to obtain the calibration curve. The emission values were read from a digital averager. The reading obtained with conductance water was noted between each standard in order to give the baseline correction. Calibration curves of ppm vs. output were used to find the concentration of the unknown in ppm. II.D.4. Proton NMR The 18-crown-6 content of the crystals was determined by drying the neutral solution remaining after the pH titra— tion. The solution was placed in a desiccator with Drierite as a drying agent to permit slow evaporation of the water. The residue was further dried in a vacuum desiccator before dissolving in D20. A known amount of sodium acetate was 30 added to the solution as an internal standard. A 1:1 ratio of acetate protons to crown protons was approximated by using the amount of H2 evolved as an estimate of the number of moles of K-18C6-Na. Spectra were taken on a Bruker 250 MHz Fourier Transform NMR instrument. Individual peaks were isolated and fit to either Lorentzian or Gaussian curves by a line fitting program provided by Bruker. This program gives the amplitude, full width at half height and the standard deviation of the curve. A ratio of the area under each curve could be made to give the ratio of acetate to crown protons. Proton NMR was also used to determine the solvent content of the crystals, if any. Crystals were dissolved in D20 and spectra were taken of the resulting solution. The l8-crown-6 peak was used as the internal standard. II.E. Instrumental Techniques II.E.l. Optical Spectra Optical spectra had previously been taken of films from solutions of known stoichiometry (11,35). Because of the red shift of the optical peak of Na observed with prior films of K-l8C6-Na, it was desired to make films directly from solutions prepared by dissolving the crystals. This technique used an apparatus such as that shown in Figure 7. The scribed sample tube was placed in side arm A which was connected to a Pyrex cap with a piece of flexible heat— 31 D Figure 7. Apparatus for optical spectra of thin films of K-18C6-Na from crystals dissolved in methylamine. 32 shrink Teflon tubing at B. After evacuation to 10-5torr, the vial was moved to the heat-shrink region and broken. As in the hydrogen evolution procedure, care was taken so that all of the crystals were in the lower portion of the tube. By keeping the heat-shrink tubing compressed, it was possible to make the seal at C with only the lower half of the sample tube in the apparatus. This reduced the chance for decomposition due to contamination by the glass tubing. MA was distilled into B and the solution was poured through the frit into chamber D, the optical cell. The cell was then kept in dry ice to reduce the vapor pressure of the methylamine, while a dynamic vacuum seal was made at E. Film making was a trying process since the majority of the films were not of proper thickness or uniformity.t To make films, most of the solution was poured to chamber F leaving between one and two ml in D. While F was submerged in liquid nitrogen and the cell was kept in a dry ice-isoprOpanol bath, the apparatus was vigorously shaken to spread the solution over the walls of the cell. The films left by the evaporation of methylamine in D were made at different annealing temperatures by varying the bath temperature. Unsatisfactory films were remade by thawing the solution in F and distilling methylamine back into the cell. Spectra were recorded on a Beckman DK-2 double beam recording spectrophotometer that had been modified to permit temperature control of the sample compartment between 0 and -65“C. Rough temperature control was achieved by cold 33 ethanol circulated through the cell compartment from a controlled temperature cooling bath (Neslab LTE-9), and the temperature was further controlled by cooled N2 gas. A copper-constantan thermocouple was placed in the cell compartment to indicate the temperature. Spectra were recorded in the 1500 to 400 nm range with the reference beam passing through air. The spectra were normalized by subtracting a base line correction of an empty cell and then rescaled from the lowest absorbence at zero to the maximum at 1.0. The dependence on annealing temperature, time and film temperature were studied. II.E.2. Powder D.C. COnductivity Powder conductivity measurements under pressure were made in an apparatus designed by J.L.Dye and M.R.Yemen. Under a dry nitrogen atmosphere, a sample was loaded into a precooled 2mm inner diameter heavy-walled fused silica tube which rested on a stainless steel electrode. A second stainless steel electrode was inserted into the top of the tube and was used to compress the sample by means of a spring with a measured force constant. The sample cell was loaded into a cryostat which was cooled by controlled liquid nitrogen boil off. Ohm's law obedience was checked by measuring the current as a function of voltage. At constant voltage, the crystals were slowly cooled to less than 50°C and warmed again to 0°C. The current was read at approximately two degree intervals during the temperature cycle. The apparatus was then disassembled to check the 34 powder for decomposition. If the crystals maintained the dark blue color, the system was reassembled and taken through another temperature cycle. This second cycle was used to determine the conductive properties of the K-l8C6-Na. At the completion of the second run, the height of the pressed powder column was measured so that the specific resistivity could be estimated. II.E.3. Recrystallization Crystals grown during synthesis are Often of unsatis— factory quality for X-ray crystallography and single crystal conductivity experiments. Rapid growth of the crystals can lead to twinning, clumping and other deformities. Therefore, the following recrystallization scheme was used in attempts to grow more perfect crystals. The apparatus is shown in Figure 8. The top of chamber B was closed by a Pyrex cap connected by a 9mm Fisher Porter joint. In an Argon atmosphere dry box, crystals were placed through the Kontes valve into chamber A. The apparatus was closed quickly, removed from the dry box to a vacuum line and evacuated to 2x10.”:5 torr. The crystallization solvent (either isopropylamine or an isoPropylamine-diethyl ether mixture,3:l) was distilled onto the crystals to make a saturated solution at -40°C. The solution was poured through a frit to the crystal growing chamber,B. The apparatus was then placed in a precooled temperature- controlled ethanol bath (Neslab LTE-9) which had been modified for slow cooling. This modification consisted of a 35 .__° \ \ / , Z/ A Figure 8. Apparatus for recrystallization. 56 clock motor which drove a rotating digital counter. The counter in turn, was connected by a belt to a ten turn wire- wound potentiometer (Helipot) which replaced the 3/4 turn potentiometer in the temperature controller of the LTE-9. By choosing the proper wheel of the clock, the potentiometer could be given one turn every hundred minutes. This corresponded to a lowering of the bath temperature of " 4°C per hour. The solution was first subjected to two or three ten minute temperature cycles between -36 and -44°C to reduce the number of crystal seeds before cooling slowly to -70°C. Removing the apparatus from the bath to examine the crystal growth often resulted in the precipitaion of many small crystals. Therefore, curiosity was held in check. In the event of this rapid precipitation of small crystals, the bath temperature was raised to -400 and the slow cooling process was once again begun. During the cooling process, the clock motor was stopped for 12 hours at 50 intervals to overcome diffusion problems. This waiting period permitted maximum growth of most of the crystals at a given tempera- ture before cooling resumed. Also, in order to increase the yield, some of the solvent was removed by the following procedure. A beaker was placed inside the bath around chamber B so that the rim of the beaker was at the fluid level of the bath. An immersion heater controlled by a Variac was used to raise the temperature of the ethanol in the beaker to about 5°C above the temperature Of the rest of the bath, so that the solvent would distill from B into A. 37 This procedure raised the entire bath temperature so that the distillation described above was usually between the temperatures of -55 and -60°C. After half of the solvent had been distilled, the entire apparatus was cooled to —65'C and tranSferred to a dry ice-isoproPanol bath. The re- maining solution was poured from B to A and the solvent was distilled into a waste bottle. The crystals were washed with ~5 ml of n-pentane which was then distilled into another waste bottle. The lower part of the apparatus was packed in dry ice until use. The crystals were removed in an Argon atmosphere dry box through the Fisher Porter opening on chamber B. II.E.4. Photoconductivity Attempts were made to determine the effect of light on the conductive properties of thin films of the K18C6Na system. This was done by making a dry film from a solution of the crystals in methylamine over two silver strip electrodes. The current between the two strips was measured as the film was irradiated over the 800-325 nm range. Two types of cells were used in this study. The first was designed by M.R.Yemen and is Shown in Figure 9. The main chamber has four silver strips (Decal Craft,Penn.) initially bonded to the inner wall by decal type adhesive. The silver was permanently fixed to the wall via a silicate adhesive by baking in a 550°F oven for four to twelve hours. Two of these strips were soldered to leads connected to tungsten wires which were sealed vacuum tight in a Pyrex 38 Figure 9. Apparatus for photoconductivity measurements. Cell designed by M.R. Yemen. 39 cap. The other two strips were held in reserve. The chamber and cap were joined by a 9mm Fisher Porter connection. Crystals were loaded through the Kontes valve in a nitrogen atmosphere glove bag. The Kontes valve was then replaced and the apparatus was immmediately evacuated to 2x10"6 torr. Approximately 7 mls of methylamine were introduced to dissolve the crystals. Films were made by vigorously shaking the apparatus to form a film over the entire surface. The wall opposite to the silver strips was washed by distilling solvent from the bottom of the chamber. The solvent was frozen in liquid nitrogen. The film was illuminated by a beam from a 75W DC Xenon high pressure lamp which passed through a Bausch and Lomb high intensity grating monochromator followed by a chopper. Filters were used to suppress higher order radiation and the beam was focused onto the film with lenses. Current was measured with a Keithley 610B electrometer. The side chamber of this apparatus had been used in studies which preceded this in- vestigation (37) and was not used here. The second apparatus, Figure 10, was designed by Dr.L. D.Le. It consists of two concentric tubes joined by an Ace 15mm connector (Cat. No. 7644-15) with a plastic coupler. The inner tube has a long finger with two silver strips sealed to the glass in the manner previously described. Shielded cable was soldered to the strips and passed through the insert at A. Torr-Seal epoxy (Varian) was used to seal the entry holes. The upper portion of the insert held an 40 (a) (b) Figure 10. Cell designed by L.D. Le for photoconductivity measurements. Insert (a) fits into outer tube (b) aligned by arrows as indicated. 41 extension of the hollow finger. The outer tube had a large lower bulb with a quartz Optical cell insert. A graded seal joined this to the Ace connector. The inner tube fit exactly into the plastic coupler and the seal was made with a Viton O-ring at the bottom of the coupler. The inner tube could then be moved up and down by a slow rotation pull motion without loss of vacuum. Crystals were loaded in a nitrogen glove bag into the outer tube. The apparatus was assembled and evacuated to 2x10"5 torr. Films were made by lowering the inner tube to slightly above the solution level and shaking vigorously as the solution was frozen in liquid nitrogen. The finger was raised so that the film was centered in the optical cell. Cooled nitrogen gas was used to cool the finger to retard thermal decomposition. Unfortunately, the Torr-Seal seals were not stable under changing temperatures. Flowing gas also led to mechanical vibrations which made sensitive measurements impossible. II.E.S. Electron Paramagnetic Resonance The sample for EPR spectra was loaded into an evacuable 3mm outer diameter quartz tube in a argon atmosphere dry box. The tube was then removed to the vacuum line, evacu- ated to 3x10‘5 torr and sealed. Spectra were recorded on an X-band spectrometer (Bruker model 200) over the tempera- ture range l72.l-311.5 K. Temperature regulated N2 gas was used to cool the sample. A c0pper-constantan thermocouple with digital readout (Doric model DS-350) placed directly under the sample was used for temperature readings. III. POTASSIUM 18-CROWN-6 SODIUM The major goal of this thesis investigation was the synthesis and analysis of the K-18C6-Na system. Prior to this study, the preparations of K-18C6-Na had given various analytical results. Crystals were found to have up to 16 mole percent solvent (10) and the exact endpoint was difficult to determine in the pH titrations (11). The properties of K-18C6-Na had not been explored and only optical spectra and DC conductivity measurements had been made. The optical spectra suggested that the compound was either a mixed potasside-sodide salt or a Simple alkalide (either K’ or Na") with a large spectral shift from those normally observed (35). Three separate preparations of K-l8C6-Na were made in the present work. In addition to the basic elemental analysis, the optical spectra and DC conductivity measure- ments have been repeated. To further characterize the system, photoconductivity measurements, EPR spectra and thermal decomposition studies have been done. A recrystallization procedure has been developed and imple- mented to produce better crystals for single crystal studies. The experimental techniques for these procedures have been described in the previous chapter. The results of these studies are reviewed in the following sections. 42 43 III.A. Elemental Analysis The three preparations of K-18C6-Na differed only in the amount and ratio of starting materials. The initial ratio of K:18C6:Na was essentially 1:1:1 in the first two preparations (synthesis I and II), but an excess of Na was used in the third preparation (synthesis III). Another dis- tinguishing feature in synthesis II was that the metal did not completely dissolve during preparation. The crystals from II and III were identical in appearance and formed as small dark bronze crystals of irregular shape. Crystals from synthesis I on the Other hand, were large flat rhombic crystals which were gold in color. They were temperature sensitive, in that they would adhere to the wall of the sample tube if warmed to 0°C or above. Crystals of synthesis II and III remained free-falling in the sample tubes over this temperature range. A sample from each preparation was analyzed according to the reaction: K-18C6-Na + 2H20 —-—— K“ + Na+ + ZOH“ + 18C6 + H2 The number of moles of hydrogen gas was calculated by using the ideal gas law and was used as the reference for the total number of moles of K-18C6-Na throughout the rest of the analysis. Table 1 gives the results of the analysis scheme described in chapter II. As an additional check, the mass of the sample tube with and without crystals was .memccuu ca umoH m~mum>uo AU .oocoummou ~scumu:« mzu om oum~ uses: :22 :— >n Odessa mumumeom O £uw3 cocaEuouoc . Ac .oumanmzux MO Ofiumu Auauu O meassmoo Ono coduquQEOooc Looms USO daemon Once Odessa UcfizmwOB >9 poc«Euouoo moo: Am .~.m~+. .m.o+. .o. A..~+. Am.m+. .u. 3.5 .mh.m mm.m m~.n so.> O¢.m m~.m HHH -4 Am.a+. .o.+. .m.~+. Aa.n+. .o.o+. Ase t. u mm.~ ~m.~ mm.a oe.~ .ee.~ m~.~ HS .m.e+. .m.e+. .o.aa. . .m.e+t .u. ea O&.~ oe.~ Q~.~ mu.m I 0mm. H seaflEOA>QCuQOma cad x «Ga x wad x cad u some: «OH x mfinO5uc>m m Odo: wood OOAOZ +oz oodoz .4: node: IIO mode: a: node: .oocou0mou on» on com: ma po>~o>o new evacupas no dado: .ouoozucouoc cw co>wm ow >uu050a50aouu couOucoua one EOuu ecuuca>oc because one .mznmomth wo nfio>~cco on» no nausoom .H Ounce 45 measured, and the corresponding number of moles of K-18C6-Na in the crystals was calculated from the difference. The ratio of K:Na:18C6 is essentially 1:1:1 except in the case of synthesis III where there is an excess of sodium and crown. This sample could be a mixture of an electride and an alkalide or could have excess sodium metal. In any event, the analysis suggests that it is not apprOpriate to use excess sodium in the preparation. A new technique of 18C6 determination was used in this investigation. Previously, determination of the number of moles of complexing agent had been done by integration of the 1H NMR signal of the complexing agent and an internal standard (11,38). For the earlier studies, potassium hydrogen phthlalate (KHP) had been used as the internal standard. Because of its ease of dissolution and its simple spectrum (one singlet), sodium acetate (NaAc) was used as the internal standard in this work. The Bruker 250 MHz Fourier Transform NMR instrument has a program which will fit isolated peaks to either a Gaussian or Lorentzian function. The respective forms of these functions are: f(6 )=2Aexp(-25w at.” 2.2)” f(%;=A/E1-2 )] where A is the maximum amplitude and Gage”) is the change in and 46 frequency (40). Integration of these functions shows that the area under each curve is proportional to the amplitude times the full width at half height (fwhh). The program provided by Bruker gives these values as well as the standard deviation of the calculated curve from the transformed signal. Several spectra of D20 solutions with known ratios of 18C6 to NaAc protons were measured to determine the accuracy of this method. Most of these signals were essentially Lorentzian in shape as indicated by the quality of the fit to this function compared to a Gaussian. The ratio of 18C6 to NaAc was determined by the ratio of the products of the- height and fwhh. We found poor lineshape resolution when using the multi-nuclear probe. Therefore, a proton-only probe was used. High quality signals are absolutely essential in obtaining good results. To further improve the form of the signal, an exponential multiplier was applied to the free induction decay (FID) in the time domain to broaden the Fourier transformed spectral lines and to provide more symmetry. Table 2 gives the results of these studies by comparing the mass ratio to the ratio given by this method. It is clear from examination of the standard deviation of the individual peaks, as well as the relative accuracy of the ratios, that the exponential multiplier enhanced the overall accuracy of this method. The determination of solvent content could not be made by the curve-fitting method due to the resulting spectra. 47 Table 2. Comparison of ratios of lB-Crown-6 protons to sodium acetate protons by mass and by NMR measurements. 18C6 : NaAc Standard deviation of peaks Percent Standard by mass by NMR high field 19! field diffexenee l - 1.077 1.08838 .009 .006 1.1 1.100853 .07 .02 2.23 2 2.166 2.20209 .08 .14 1.6 2.32535 7.3 2.83628 .36 .46 30.7a 3 0.488 0.50124 .19 .03 ‘ 2.6 a) Without exponential multiplication of the FID. E8 Two samples of K-18C6-Na were directly reacted with D20 and the spectra were immediately taken. The sample from synthesis I gave a spectrum with a septet and a high field doublet in addition to the 18C6 peak. This spectrum corresponds to the presence of isopropylamine in addition to K-18C6-Na. By integration of the doublet and the 18C6 peak, the sample was found to contain ~l4 mole percent isopropyl- amine. Crystals from synthesis III showed no evidence of isopropylamine, even at high gain. The presence of the solvent in crystals from synthesis I is probably the origin of the difficulty in handling these crystals at higher temperatures. III.B. Solubility Studies and Recrystallization Attempts To obtain crystals adequate for x-ray crystallography, the recrystallization scheme described in chapter II was developed. Before this investigation, crystals for X-ray structure determination had to be selected from those grown during synthesis. The crystals remaining at the end of the K-18C6-Na preparations were either too small or existed as clumps of several crystals. Crystals grown by recrystal- lization were generally larger, but would redissolve if the temperature exceeded -50°C. Also, observations made during synthesis indicated that diethyl ether was unsuitable as a wash solvent since it severely etched the crystal surface. To find a suitable wash solvent, crystals from the syn- 49 thesis were loaded into a recrystallization apparatus and various solvents were distilled onto the crystals. N-pentane had no effect on the already etched crystals. The pentane remained colorless even after standing over the crystals for several hours. However, trimethylamine immediately reduced the crystals to a fine dark powder that could easily pass through the coarse frit of the apparatus. N-pentane was therefore choosen as the wash solvent but unfortunately, the recrystallized crystals changed from a gold color to plum purple in its presence. Reduction of the contact time with pentane lessened its effect on the crystals. The color change is probably due to removal of isopropylamine or 18C6 from the crystals. Thus, although the successful growth of larger crystals has been achieved, their isolation for X-ray crystallographic studies has not been accomplished. III.C. Optical Spectra Spectra of films from methylamine solutions which contain equimolar amounts of K, Na and 18C6 have already been reported (35). Figure 11 gives a spectrum of a dry film from one of these solutions (11). Note the broadness of the peak at 13,000 cm‘1 and the shoulders at 10,000 and 14,000 cm”. The width of the peak may be attributed to the non-uniformity of the film (41). The location of the 1 maximum was shifted 2700 cm" to the red from that of Na‘ in Na+18C6 Na' films formed by vapor deposition, and was blue 50 .OOOfiusaom mafiamaanuoe umaoawaao Bonn mzucomfitx mo Edam app c «O enuuooam HmOfiuao .da ouswfim WIO_ .A-tEUvm mm on o. - . o. k 1 0.. no 6.0 No mo 0.. m. e... 3.1: 51 shifted 800 cm"-1 from that of K- in Kf18C6 K"films from methylamine solutions (35). The Na peak shift observed from films from methylamine solutions between the corresponding pair Of salts, K+C222 Na" and Na1C222 Na' , is only 300 err1 (35). Figure 12 is a typical film spectrum from methylamine solutions of K-18C6-Na crystals. The maximum is at 14000 cm‘1 and the shoulder observed in previous spectra is conspicu- ously absent. The peak is also considerably narrower than in the previously recorded spectra. As in previous studies of this system, the spectra of dry films from crystals in methylamine were temperature- and time-independent. The difference in the locations of the maximum in this study and that of prior work may be due to the nature of the solutions from which the films were made. (The dry film spectra of the equimolar solutions were taken shortly after the solutions were prepared. An exchange reaction such as,. M‘C + N‘——.M' + N+C may not have had time to go to completion if the exchange rate is slow. Thus, films made from freshly prepared equi- molar solutions could contain a mixture of Na’ and K’ . A more likely explanation is the possibility that these solu- tions did not possess a 1:1:1 stoichiometry. The K"absorption from vapor deposition of potassium; metal and 18C6 in a 2:1 ratio, shows a maximum at 13,000 cm‘1 52 .osaacahnuoe a“ mamumauo azuoomulx mo chquHom Eouw Edam cfinu mo Esuuooam HwOfiudo .- ouswfim A 75on loooé ooo.m_ ooo.o~ ooo.m~ _ a _ . _ . .to.o ,. and XOE d. < to Ad.’ P _ Pl _ . L . 3 o ._ o. o. 4.. o. m. e. 55 and has a very broad peak (33).. Films of Na*18C6 Na, prepared by the same method have a very broad maximum at 15,700 cm‘d. Superposition of these peaks with a strong emphasis on K’ could give the previously observed 13,000 cm‘1 peak. The single absorption peak at 14,000 cm'"1 for films from solutions of the crystals indicates a change to a simple sodide with a decrease in the energy gap compared to that of Na+C222 Na", possibly due to the interaction with K*18C6. In addition, films made from methylamine solutions of Rb-18C6-Na have a maximum at 16,400 cm”1 (38). This blue shift from the Na“ peak in Na’C222 Na‘ films further in- dicates the strong dependence of the peak location on counter ion and complexant. The sharpneSs of the absOrption band also seems to indicate that the crystals are a single alkalide rather than a mixed alkalide. III.D. Powder DC Conductivity All conductivity measurements were preceded by a check of adherence to Ohm's law in the assembled system to make sure that contact polarization did not affect the resistance. Figure 13 is a plot of current versus voltage at constant temperature. The current at a set voltage of 5 volts was measured with changing temperatures through two temperature cycles. Figure 14 is a plot of log R vs l/T for the K-l8C6-Na system. Data from the first temperature cycle deviated severely from the calculated curve at low 54 .muopsoa mzuoum_tx so“ owcuao> .m> ucouuso Op 0 O .ML enemas ad ad. (endure) 010‘ x | QON 55 Afin ATV: no. x K. 00 ad” @— .mzuoomalx pow ouaucuoaSoO Hmooudwoou co moamumfimou woa mo mococcmaoa .q~ muzwfim (a) 901 56 temperatures. These deviations were not present in the second temperature cycle and are most likely due to changes in packing density of the powders. A fit of the equation: I=V/(R°. exp(-E9/2k'T)) (l) (where I is the observed current at temperature T, V is the set voltage,R, is the resistance at infinite temperature, E9 is the energy gap and kl is Boltzmann's constant), to the data with the aid of the nonlinear least squares program KINFIT (42) gave the average band gap as 1.02 + .04 eV. Estimates of the specific resistivity at infinite tempera— ture,nw, ranged from 4.37xl0'3 to 6.32x10"2 Ohm-cm. The resistivity is a measure of the ability of a material to resist a current flow and ranges from 1.7xl0"6 Ohm-cm for copper to about 101'ohm-cm for fused silica at room temp- erature (43). The moderate band gap and the exponential behavior indicates that K-l8C6-Na is a semiconductor. The value of’q3° greatly exceeds that of instrinsic semiconduc- 4ohm-cm (12). We therefore tors which are on the order of 10' conclude that K-l8C6-Na is an extrinsic semiconductor, where the impurities may be F-centers, free metal atoms or anions in a perturbed environment. III.E. Photoconductivity The technique for photoconductivity measurements is still under development. Preliminary results of the 57 photoresponse in Na+C222 Na“ films indicate that surface properties play a more important role than the bulk prOperties (44). Although photoconductivity is observed in most NaIC222 Na' films, the wavelength and time dependence suggest that it is due to the trapped electron rather than 'Na’ . PhotocOnductance "spectra" decay with time, indicating thermal or optical bleaching of traps. Films of K718C6-Na showed only a slight photoresponse in the 500-325 nm range, at much higher energies than the optical transition in both cells. Unfortunately, changes in the dark current also occurred in the same region and decomposed films gave similar responses. Therefore, these intial photocon- ductivity studies suggest that there is little or no photo effect on the conductivity of such films. It should be noted, however, that many films of Na+C222 Na' also shOwed no photoconductivity (37). III.F. Electron Paramagnetic Resonance Electron paramagnetic resonance (EPR) spectra give information about the enviroment of unpaired electrons. True alkalides would be diamagnetic with the outer electrons having an 52 configuration, and therefore would have no EPR signal. However, paramagnetic impurities,such as F-centers or metal atoms, would give EPR signals. The spectrum of an F-center is usually inhomogeneously broadened due to hyperfine interactions between the electron and the 58 surrounding nuclei (16). Hyperfine splitting (hfs) of EPR signals have been observed for alkali metals in amines and ethers (45-50) and in frozen organic glasses (51,52). The hyperfine splitting in solution was found to be solvent. metal and temperature dependent. The nuclear spins of Na, K, Rb and Cs are 3/2, 3/2, 5/2 and 7/2 respectively and give hfs quartets for Na and K, sextets for Rb and octets for Cs. The species reponsible for the hfs spectra has been proposed to be the solvated electron (52) ion-paired to the alkali cation or a combination of solvated metal atoms and ion- pairs (47). The K-18C6-Na system has an easily observable EPR signal although the free spins constitute only a small fraction of the electrons: the majority of the valence electrons are spin paired in the anion's s’ outer shell and are EPR inactive. The observed spectrum for the K-l8C6-Na powder is quite different from the previously reported spectra of alkalides and electrides (11,36,42). In the previous cases (except Li+l8C6 e’), the signal was due to a small fraction of trapped electrons and consisted of a single line or several overlapping lines. Figure 15 shows the EPR spectrum at +38.3°C of K-18C6-Na. Five peaks are evident: a strong central peak and a quartet, pressumably due to the hfs by either a Kl'or Na+ nucleus. The 9 value of the central peak, determined by calibration with a standard sample of diphenylpicrylhydrazyl is 2.00259 while the quartet is symmetrically situated about g=2.00236. The S9 0 6.0.. .oom.mm+ um Henson mznoomfiux mo ESHOOOOm mam .m~ ouswam 60 hyperfine coupling constant (hfcc) observed is 11.5 Gauss and is compatible with that observed in amine and ether solutions (49,50) but is temperature independent. In solu- tion, the hfcc decreased as the temperature increased (48). The hfcc for atomic potassium is 165.1 G (16) indicating that there is 7.0% atomic character in the sample if the cation is K+'. On the other hand if the cation is Na+, there is 1.8 % atomic character when compared to the free atom hfcc of 632.3 (16). At lower temperatures, the hfs spectrum loses intensity although the coupling constant remains the same. The spectum at -10l.l°C, Figure 16, shows only the typical asymmetric shape found in other alkalide and electride systems. The signal is still modulated by the unresolved hyperfine structure at lower temperatures. The "solvation shell" of the cation in K-18C6-Na is the crown ether, therefore the resolution of the hfs at higher temperatures indicates a partial melting of the crystal lattice, leading to a motional narrowing of the signal. However, exchange is not fast enough to eliminate the hyperfine splitting. Due to the rigid structure of the macrocyclic crown ether, compared to a solvation shell made from small molecules, we expect relatively little change in the nearest neighbor distances. This is consistent with the temperature independence of the hfcc observed in this study. _o ode— .o.~._o_a he hooeoe azuoowaue to eeheooem mam .OH shaman 62 111.6. Melting Point To determine the melting range of the K-18C6-Na crystals, a simple melting point determination was per- formed. A few solvent free crystals from synthesis II were sealed under vacuum in a Pyrex tube. Observations were made through a microscope as the sample tube was heated from -25 to +579C in an isopropanOl bath. No detectable changes were observed until the temperature reached +48.2°C. At this point, the crystals began to clump together and adhere to the sample tube walls. As the temperature was further increased to 52°C, the crystals became quite shiny and bronze in color. The crystals Spread to form a viscous film in the 52.4 - 53.20 range, but a true liquid state was never reached. Upon further heating, the films darkened and became black. When cooled, the color did not revert to either the dull bronze of the crystals or to the shiny bronze of the films. These black films were then cooled to room temperature and reheated to 56°C. They became clear and colorless as the temperature increased. This observed melting range is significantly higher than the melting point of the NaK alloy which falls between -3.5 and +5.0"C (54). IV. Conclusions and Suggestions for Further Work IV.A. Conclusions An alkalide salt of K, Na and lB-crown-G has been prepared. Analysis indicates a 1:1:1 ratio of K,Na and l8-crown-6. The form, temperature sensitivity and color of the resulting crystals are dependent on solvent content. The sharpness of the optical absorption peak indicates that this compound is a simple alkalide with either K' or Na’ as the anionic species, although the peak location does not allow us to decide unambiguously which is the anionic species. EPR spectra show a weak signal due to a small number of para- magnetic trapped electrons. At high temperatures, inter- action of the electron with surrounding K+ or Na+ nuclei is present. Powder conductivity measurements indicate that this system is an extrinsic semiconductor, having conductivity due to "impurities" such as trapped electrons. There was no photoresponse Observed from films prepared from methylamine solutions. However, the method for photoconductivity measurements is still in the developmental stages. Crystals of K-l8C6-Na can be recrystallized from appropriate solvent mixtures and are relatively stable at room temperature, decomposing only at elevated temperatures. 63 IV.B. Suggestions for Further Work 1) The question of whether K’ or Na" is the anionic species has not been fully answered. To affirm K-18C6-Na is a sodide, solid state 23Na NMR should be. perfOrmed. Solid state spectra have been taken of the Na*C222 Na' system and separate signals with distinct line shapes are present for Na+ and Na" (38). The ionic nature of the sodium in the crystals can therefore be determined by this method. Suc; cessful recrystallization may lead to X-ray structure determination as well as single crystal conductivity and single crystal EPR. The photoconductivity of this system needs to be pursued to further elucidate the nature of the alkalide. 2) Synthesis of other similar alkalide and electride systems is needed. The stabilization of Cs*18C6 e” by lithium in solution (11,34) can perhaps be mimicked to attempt the synthesis of K*18C6 e' and Na+18C6 e“. Nitrogen analogs of the complexant where the oxygen atoms are replaced by -NH- or -NCH2- should also be used in attempts to make the corresponding alkalides and electrides. 9. 10. 11. 12. 13. 14. 15. 16. V. REFERENCES J.L.Dye, J.M.Ceraso, M.TaLok, B.L.Barnett and F.J.Tehan, J. Amer. Chem. Soc., 96, 608 (1974). F.J. Tehan, B.L.Barnett and J.L.Dye, J. Amer. Chem. Soc., 96, 7203 (1974). J.L.Dye, J. Chem. Ed., 54, 332 (1977). Frederick John Tehan, Ph.D. Dissertation, Michigan State University, 1973. J.L.Dye, Scientific American, 237, 92 (1977). J.L.Dye, M.R.Yemen and M.G.DaGue, J. Phys. Chem., 68, 1669 (1978). ’ J.L.Dye, Angew. Chem. Int. Ed., Eng., 18, 587 (1979). M.G.DaGue, J.S.Landers, H.L.Lewis and J.L.Dye, Chem. Phys. Lett., 66, 169, (1979). J.L.Dye, J. Phys. Chem., 84, 1084 (1980). B.Van Eck, L.D.Le, D.Issa and J.L.Dye, Inorg. Chem., 21, 1966 (1982). Dheeb Issa, Ph.D. Dissertation, Michigan State University, 1982. J.S.Blakemore, "Solid State Physics", W.B.Saunders, Co.(Philade1phia), 1974. P.F.Weller, "Solid State Chemistry and Physics", vol. 1, Academic Press (New York), 1968. P.D.Townsend and J.C.Kelly, "Colour Centres and Imperfections in Insulators and Semiconductors", Crane, Russak and Co. (New York), 1973. W.Beall Fowler, "Physics of Color Centers", Academic Press (New York), 1968. M.Bersohn and J.C.Baird, "An Introduction to Electron Paramagnetic Resonance", W.A.Benjamin, Inc. (New York), 1966. 65 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. W.Weyl, Am.Phys, 197, 601 (1863): Poggendorffs Annln, 121, 601 (1864). S.Matalon, S.Golden and M.Ottolenghi, J. Phys. Chem., 73, 3098 (1969). J.L.Dye, M.G.DeBacker, J.A.Eyre and L.M.Dorfman, J. Phys. Chem., 76, 839 (1972). T.R.Tuttle,Jr., Chem. Phys. Lett., 20, 371 (1973). M.G.DeBacker and J.L.Dye, J. Phys. Chem., 75, 3092 (1971). J~M.Ceraso and J.L.Dye, J. Chem. Phys., 61, 1585 (1974). J.L.Dye, C.W.Andrews and J.M.Ceraso, J. Phys. Chem., 79, 3076 (1975). C.J.Pendersen, J. Amer. Chem. Soc., 89, 7017 (1967). B.Dietrich, J-M.Lehn and J.P.Sauvage, Tetrehedron Lett., 34, 2885 (1969). J.L.Dye, M.G.DeBacker and V.A.Nicely, J. Amer. Chem. Soc., 92, 5226 (1970). M.T.Lok, F.T.Tehan and J.L.Dye, J. Phys. Chem., 76, 2975 (1972). J.Lacoste and F.Schue, J. Organometallic Chem., 231, 279 (1982). . J.D.Lamb, R.M.Izatt, C.S.Swain and J.J.Christensen, J. Amer. Chem. Soc., 102, 475 (1980). G.Wipff, P.Weiner and P.Kollman, J. Amer. Chem. Soc., 104, 3249 (1982). E.Mei, J.L.Dye and A.I.Popov, J. Phys. Chem., 81, 1677 (1977). J.L.Dye, C.W.Andrews and S.E.Mathews, J. Phys. Chem., 79, 3065 (1975). L.D.Le, D.Issa, B.Van Eck and J.L.Dye, J. Phys. Chem., 86. 7 (1982). D.Issa and J.L.Dye, J. Amer. Chem. Soc., 104, 3781 (1982). J.L.Dye, M.G.DaGue, M.R.Yemen, J.S.Landers and H.L. Lewis, J. Phys. Chem, 84, 1096 (1980). 36. 370 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 67 Michael G. DaGue, Ph.D. Dissertation, Michigan State University, 1979. Michael R. Yemen, Ph.D. Dissertation, Michigan State University, 1982. Bradley Van Eck, Ph.D. Dissertation, to be submitted to Michigan State University, 1983. Unpublished results by Ahmed Ellaboudy in this laboratory. DISNMR, Disk Interactive SpectrosOOpy on the Aspect 2000 minicomputer, version 81091, Bruker Instrument, Inc. 1981. J. Steven Landers, Ph.D. Dissertation, Michigan State University, 1981. J.L.Dye and V.A.Nicely, J. Chem. Ed., 48, 433 (1971). David Halliday and Robert Resnick, "Fundamentals of Physics", John Wiley and Sons (New York), 1974. S.Jaenicke and J.L.Dye; paper in progress: “Photo- conductivity in Alkalide Films". L.R.Dalton, J.D.Rynbrant, E.M.Hansen and J.L.Dye, J. Chem. Phys., 44,3969 (1966). R.L.Harris and J.J.Lagowski, J. Chem. Phys., 84, 1091 (1980). K.Bar-E1i and T.R.Tuttle, J. Chem. Phys., 40, 2508 (1969). R.Catterall, J.Slater and M.C.R.Symons, J. Chem. Phys., 52, 1003 (1970). R.Catterall, M.C.R.Symons and J.W.Tipping, J. Chem. Soc., (A), 1530 (1966). J.L.Dye and L.R.Da1ton, J. Phys. Chem., 71, 184 (1967). R.Cattera11 and P.P.Edwards, Chem. Phys. Lett., 42, 540 (1976). R.Catterall and P.P.Edwards, Mol. Relaxation, 13, 123 (1978). J.L.Dye, Pure and Appl. Chem., 49, 3 (1977). th Handbook of Chemistry and Physics, 58- edition, The Chemical Rubber Company, 1978.