SPECTROSCOPIC STUDIES OF, IONIC INTERACTIONS AND COMPLEXATION ’ ~ 0F ALKALI METAL IONS IN V " ' VARIGUSSOLVENTS‘ , f * DIssertatm Io’rtIIeVDegreeofPIID [T ' 7 I ‘ mcmem STATE UNIVERSITY was: It. CAHEN: ‘ *1975 ’ ‘ ‘w :qig I! '59:?“ ._ 1‘ I ‘ 1 . 7 This is to certify that the id, thesis entitled Spectroscopic Studies of Ionic "- ‘2' Interactions and Complexation I '31 of Alkali Metal Ions in Various Solvents presented by Yves M. Cahen has been accepted towards fulfillment of the requirements for Ph . D . (19mm, in Chemistry ///W/ saggy Major profesgor/ Date ngruagx'ZS, 1975 0-7639 fj ABSTRACT ’17 SPECTROSCOPIC STUDIES OF IONIC INTERACTIONS AND {I @W COMPLEXATION OF ALKALI METAL IONS IN VARIOUS SOLVENTS BY Yves M. Cahen Chemical shifts of lithium-7 nucleus were measured in eleven nonaqueous solvents against 4.0 fl aqueous lithium perchlorate solution. Lithium perchlorate, chloride, bromide, iodide, triiodide and tetraphenylborate were used. The shifts ranged from +2.80 ppm for acetonitrile, down to —2.54 ppm for pyridine. In dimethylsulfoxide and dimethyl- formamide no evidence for contact ion pairing was observed. Formation of contact ion pairs was particularly evident in tetrahydrofuran, nitromethane and tetramethylguanidine. 35Cl resonance of the per— The large broadening of the chlorate ion in these solvents is in agreement with the above explanation. In contrast to sodium-23 NMR, no cor— relation was found between limiting chemical shifts in different solvents and the Gutmann donor numbers of these solvents. Formation constants of lithium ion complexes with 1,5- polymethylenetetrazoles and 3,3-disubstituted glutarimides have been determined in nitromethane solutions by lithium—7 Yves M. Cahen NMR. Concentration dependence of the obtained values are explained by the competitive ion pair formation. Glutarimide complexes were found to be somewhat more stable than the pentamethylenetetrazole complexes. Lithium-7 NMR studies were performed on lithium ion complexes with cryptands C222, C221 and C211 in water and in several nonaqueous solvents. In the case of the first two cryptands the exchange between the free and complexed lithium ion was fast by the NMR time scale and only one population—average resonance was observed. Cryptand 211 forms much more stable lithium complexes and two 7Li reson— ances (corresponding to the free and the bound Li+) were observed for solutions containing excess of the Li+ ion. The limiting chemical shifts of the complex were found to be independent of the solvent indicating that the lithium ion is completely shielded by the cryptand. Formation constants of lithium-C222 complexes were determined in water and pyridine solutions. The values obtained were: log K = 0.99 i 0.15 and log K = 2.94 i 0.10. PY The kinetics of complexation reactions of the lithium H20 ion with cryptand C211 in pyridine, water, dimethylsulfoxide, dimethylformamide and formamide and with cryptand C221 in pyridine were investigated by temperature dependent 7Li NMR. The energies of activation for the release of Li+ from Li+-C211 complexes increase with the increasing donicity of the solvent as expressed by the Gutmann donor number. The transition state of the complexation reaction E 7 Yves M. Cahen must involve substantial ionic solvation. Using the formation constant of the Li+-C211 cryptate in water, the rate constant for the forward reaction was found to be kf = 0.98 x 103 sec. Far infrared spectra of sodium and lithium cryptates were observed in several nonaqueous solvents. The spectra are characterized by a broad band whose frequency is in- dependent of the solvent or of the anion and which is assigned to the vibration of the cation in the cryptand cavity. The band frequencies were 234 i 2, 218 i l, 1 for Na+-C222, Na+—C221, Li+— 234 i 3 and 348 i 1 cm” C221 and Li+—C211 cryptates respectively. These bands were found to be Raman-inactive,indicating that the cation— ligand interaction is very largely electrostatic in nature. SPECTROSCOPIC STUDIES OF IONIC INTERACTIONS AND COMPLEXATION OF ALKALI METAL IONS IN VARIOUS SOLVENTS BY Yves MI Cahen A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1975 ACKNOWLEDGMENTS The author wishes to express his sincere gratitude to Professor Alexander I. Popov for his guidance, encourage- ment and friendship thrOughout this study, Professor Bernard Tremillon of "Ecole Nationale Superieure de Chimie de Paris" and Dr. Richard Combes for their introduction to research and their friendship. Professor James L. Dye is acknowledged for his helpful suggestions and discussions as second reader. Gratitude is also extended to the Department of Chemistry, Michigan State University, and the National Science Foundation for financial aid. Special thanks are given to Dr. Richard Bodner and Dr. Mark S. Greenberg for numerous enlightening discussions and their help during the early part of this work, to the former and present members of the group for their constant moral support, to Joe Ceraso for his constant help through- out the kinetic part of this work, to Wayne DeWitte for his expenditure of time in proofreading this thesis, to Messrs.Eric T. Roach and Frank Bennis, without whose co— operation the NMR investigations would have been much more difficult, to the friends in Michigan and elsewhere who contributed to render our stay in America a wonderful and unique experience. Deep gratitude to our families for their encouragement, abundant mail and numerous Visits. ii Deep appreciation to my wife, Yvette, for her love, patience and encouragement throughout the years of graduate s tudy . To her and to our families, I dedicate this thesis. iii TABLE OF CONTENTS Chapter LIST OF TABLES. . . . . . . . . . . . . . . . LIST OF FIGURES O O O O O O O O O O O O O O 0 LIST OF NOMENCLATURE, ABBREVIATIONS AND SYMBOLS . . I HISTORICAL A. SPECTROSCOPIC STUDY OF SOLVATION AND ASSOCIATION. . . . . . . . . . . . . 1. Introduction . . . . . . . . . . 2. Vibrational Spectroscopy . . . . 3. Nuclear Magnetic Resonance . . . COMPLEXATION OF ALKALI METAL IONS BY ORGANIC LIGANDS. . . . . . . . . . . 1. Tetrazoles and Glutarimides. . 2. Crowns and Cryptands . . . . . II EXPERIMENTAL PART III A. SALTS. . . . . . . . . . . . . . . LIGANDS. . . . . . . . . . . . . . . SOLVENTS . . . . . . . . . . . . . SAMPLE PREPARATION . . . . . . . . . INSTRUMENTAL MEASUREMENTS. . . . l. Lithium-7 NMR. . . . . . . . . 2. Chlorine-35 NMR. . . . . . . . . 3. Infrared Spectra . . . . . . . . 4. Laser Raman Spectra. . . . . . . 5. Data Handling. . . . . . . . . SPECTROSCOPIC STUDIES OF IONIC SOLVATION BY LITHIUM-7 NMR, CHLORINE-35 NMR AND RAMAN SPECTROSCOPY A. B. INTRODUCTION . . . . . . . . . . . RESULTS AND DISCUSSION . . . . . . . iv Page .vii ix xiv u3Id hard . 24 . 24 . 25 . 26 . 26 Chapter IV SPECTROSCOPIC STUDIES OF COMPLEXATION OF ALKALI METAL IONS I. LITHIUM-7 NMR STUDY OF THE Li+ COMPLEXES WITH CONVULSANT POLYMETHYLENE TETRAZOLES AND GLUTARIMIDES IN NITROMETHANE SOLUTIONS. . A. INTRODUCTION. . . . . . . . . . . . . . . . B. RESULTS AND DISCUSSION. . . . . . . . . . . II. STUDY OF THE COMPLEXATION OF ALKALI METAL IONS BY CRYPTAND LIGANDS IN VARIOUS SOLVENTS O O O O O O O O O C O O O O O O O O O A. LITHIUM-7 NMR STUDY . . . . . . . . . . . Lithium-7 Chemical Shift of Lithium-222, l. 2. 221, 211 Cryptates in Various Solvents. Lithium NMR Study of the Lithium Ion- Lithium Cryptate Exchange in Various Solvents. . . . . . . . . . . . . . . B. FAR INFRARED AND RAMAN STUDY OF LITHIUM AND SODIUM CRYPTATES IN NONAQUEOUS SOLVENTS. . . . . . . . . . . . . . . . . . 1. Introduction. . . . . . . . . . . . . . 2. Results and Discussion. . . . . . . . . C. CONCLUSIONS AND SUGGESTIONS FOR FURTHER STUDIES . . . . . . . . . . . . . . . . . . APPENDICES I LITHIUM-7 CHEMICAL SHIFTS YE 4.0 g AQUEOUS LiClO4 OF VARIOUS LITHIUM SALTS IN VARIOUS SOLVENTS O O O O O O O O O O O O O O O O O O O 0 II DETERMINATION OF COMPLEX FORMATION CONSTANTS BY THE NMR TECHNIQUE, DESCRIPTION OF COMPUTER PROGRAM KINFIT AND SUBROUTINE EQN . . . . . . . III NMR LINE SHAPE ANALYSIS FOR TWO SITE EXCHANGE. DESCRIPTION OF COMPUTER PROGRAM AND SUB- ROUTINE EQN . . . . . . . . . . . . . . . . . . A. NMR LINE SHAPE ANALYSIS FOR TWO SITE EXCHANGE. . . . . . . . . . . . . . . . . . Page‘ 52 52 53 60 60 60 70 88 88 88 101 103 113 118 118 Chapter B. Page DETERMINATION OF T VALUES FROM LINE SHAPE ANALYSIS OF AN NMR SPECTRUM AT A GIVEN TEMPERATURE. DESCRIPTION OF COMPUTER PROGRAM AND SUBROUTINE EQN. . . . . . . . . . . . . . . . . . . . . 128 C. DETERMINATION OF THE ACTIVATION ENERGY (Ea). DESCRIPTION OF THE COMPUTER PROGRAM AND SUBROUTINE . . . . . . . . . . . . . . . 129 LITERATURE CITED. . . . . . . . . . . . . . . . . . . 133 Vi Table II III IV’ VI VII VIII IX LIST OF TABLES Donor Number (DN) and Dielectric Constant (e) of Some Important Solvents (38): . . . Key Solvent Properties and Correction for Magnetic Susceptibility on DA~60 . . . 7 Limiting Chemical Shifts of Li (0.02 M LiI3) in Various Solvents. . . . . . . . . 35 Line Broadening of Cl Resonance in LiClO4 Solutions . . . . . . . . . . . . . Formation Constants of Li+-Tetrazole and Li+-Glutarimide Complexes . . . . . . 7Li-NMR Study of c222, c221, c211, Lithium Complexes in Various Solvents at Room Temperature. . . . . . . . . . . . 7Li Chemical Shifts as a Function of Ligand/Li+ Mole Ratio for the Determina- tion of the Formation Constants of Li+- C222 Complexes in Water and Pyridine . . Description of A(6) as a Function of Temperature. . . . . . . . . . . . . . . . Exchange Rates and Thermodynamic Parameters of Lithium Cryptate Exchange in Various Solvents. . . . . . . . . . . . Far Infrared Bands of Sodium and Lithium Salts in Various Solvents. . . . . vii Page 23 34 40 57 61 68 81 84 92 Table XI XII Page Frequencies (cm-l) of the Major Far Infrared Bands for Cryptates in Various Solvents. . . . . . . . . . . . . . . . . . 94 Raman Bands of Li+-C211 Complexes in Pyridine and Nitromethane Solutions . . . . 100 viii LIST OF FIGURES Figure Page 1 Structure of (A) 1,5—cyclopoly— methylenetetrazoles, and (B) 3,3— disubstituted glutarimides . . . . . . . . 10 2 Dibenzo-lS-crown-G. The number 6 refers to the total number of oxygens and 18 to the total number of atoms in the poly- ether ring . . . . . . . . . . . . . . . . ll 3 Cryptand 222, 221, 211 and their cavity diameter (2) , . . . . . . . . . . . . . . 15 4 Exo—exo, endo—endo and exo—endo con— formations of 222 cryptate . . . . . . . . l6 5 Structure of (A) "proton" cryptate and (B) "silver" cryptate. . . . . . . . . . . 18 6 7L1 chemical shifts of lithium salts in propylene carbonate and dimethyl- formamide. . . . . . . . . . . . . . . . . 30 7 7Li chemical shifts of lithium salts in dimethylsulfoxide and methanol. . . . . 31 8 7Li chemical shifts of lithium salts in tetrahydrofuran . . . . . . . . . . . . 32 9 7Li Chemical shifts of lithium salts in acetonitrile and nitromethane . . . . . 33 10 7Li chemical shift of lithium salts in acetic acid and tetramethyl- guanidine. . . . . . . . . . . . . . . . . 38 ix Figure 11 12 13 14 15 16 7 35 C1 NMR resonance line width of LiClO4 in Li chemical shift and nitromethane as a function of lithium concentration . . . . . . Raman spectra of C10; 935 cm—1 band as a function of LiClO4 con— centration in nitromethane. . . . Raman spectra of the C10; 935 cm- band as a function of LiClO4 con- centration in, (A) methanol, (B) tetramethylguanidine, (C) aceto- nitrile . . . . . . . . . . . . . Plot of 7Li chemical shift with reference to 4.0 M aqueous LiClO4 Kg mole ratio of 3,3-dimethyl glutarimide to Li+ at constant [Li+] = 0.100 131. Solid line is computer-generated curve and dots are the experimental points . Plot of formation constants XE lithium ion concentration. A - Glutarimide. B - 3-ethyl-3- methyl glutarimide. C — Penta— methylenetetrazole. . . . . . . . Lithium-7 NMR spectra of lithium- C211 cryptate in various solvents; [c211] = 0.25 H, [Li+] = 0.50 M- Chemical shift of Li+-C211 is at 0.41 ppm gs aqueous LiClO4 solution at infinite dilution. . . . . . . l Page 46 49 50 56 59 64 Figure l7 18 19 20 21 22 23 Plot of 7Li chemical shift with reference to 4.0 M aqueous LiClO4 yg mole ratio of cryptand 222 to Li+ at constant [Ii+] = 0.25 M. Solid line is computer-generated curve and dots are experimental points . . . . . . . . . . Lithium-7 NMR spectra of 0.50 M LiClO4, 0.25 M C211 solution in pyridine at various temperatures Lithium-7 NMR spectra of 0.50 M LiClO4, 0.25 M_C211 solution in formamide at various tempera- tures. O O O O O O O O O O O O O Lithium-7 NMR spectra of 0.50 M LiClO4, 0.25 M C211 solution in dimethylformamide at various temperatures . . . . . . . . . Lithium-7 NMR spectra 0.50 M_LiClO4, 0.25 M C211 solution in dimethyl- sulfoxide at various temperatures. Lithium-7 NMR spectra of 0.50 M LiI, 0.25 M C211 solution in water at various temperatures. . . . . . Lithium-7 NMR spectra of 0.50 M LiClO4, 0.25 M c221 solution in pyridine at various temperatures xi Page . . 69 . 73 . 74 . . 75 . . 76 . . . 77 . 78 Figure 24 25 26 27 28 Page Computer fit of spectra obtained with 0.50 M LiClO4, 0.25 M C211 in formamide at 143.30C. X means an experimental point, 0 means a calculated point, = means an experimental and calculated point are the same within the resolu- tion of the plot . . . . . . . . . . . . . 80 Log kb gs l/T plot for 0.50 M LiClO 0.25 M C211 in (A) pyridine, (B) formamide, (C) dimethylformamide, (D) dimethylsulfoxide, (E) 0.50 M LiI, 0.25 M C211 in water, and (F) 0.50 M_LiClO4, 0.25 M C221 in pyridine. . . . . . . . . . . . . . . . 83 4’ Schematic representation of the complexation of lithium ion by cryptands 211 and 221. S1 repre- sents a good donor solvent, 8 2 a poor one . . . . . . . . . . . . . . . . 87 Far infrared spectra of nitro- methane, pyridine and DMSO solutions of Na+x' salts (x‘ = BPhZ and I") (solid lines) and of analogous 222 cryptates (dashed lines) . . . . . . . . . 90 Far infrared spectra of nitromethane, pyridine and acetonitrile solutions of Li+x’ salts (x' = C102 and 1’) (solid lines) and of analogous 211 cryptates (dashed lines) . . . . . . . . . . . . . . 91 xii Figure 29 30 Page Raman and far infrared spectra of 6Li—czll and 7 nitromethane solution. S-solvent band. . . . . . . . . . . . . . . . . . . . 96 Li-Cle cryptates in Comparison of ion motion band frequencies for sodium and lithium salts and their 222 and 211 cryptates respectively in pyridine, aceto- nitrile, nitromethane, dimethyl- sulfoxide solutions . . . . . . . . . . . . 98 xiii LIST Contact Solvent OF NOMENCLATURE, ABBREVIATIONS AND SYMBOLS Ion Pairs. Pairs of ions, linked electrostatically, but with no covalent bonding between them. Shared Ion Pairs. Pairs of ions, linked electro- Solvent statically by a single, oriented solvent molecule. Separated Ion Pairs. Pairs of ions, linked electro- CH CN DMF DMSO PC THF TMG CH NO MeOH statically but separated by more than one solvent molecule. Acetonitrile N,N-Dimethv1formamide Dimethylsulfoxide Propylene Carbonate Tetrahydrofuran 1,1,3,3-Tetramethylguanidine Nitromethane Methanol C222 : Cryptand 222 (C18H36N206) MW = 376.5 g C221 : Cryptand 221 (C16H32N205) MW = 332.4 9 C211 : Cryptand 211 (C14H28N204) MW = 288.4 g xiv CHAPTER I HISTORICAL PART A. SPECTROSCOPIC STUDY OF SOLVATION AND IONIC ASSOCIATION 1. Introduction Alkali metal ions and their salts play an important role in chemistry as well as in biological processes. However, their solution chemistry, especially in nonaqueous solvents, still remains largely unknown. For example, we have only a very imperfect knowledge of the exact nature of the chemical species present in such solutions, of the equi- libria between them, and of the role of the solvent in these equilibria. Classical techniques such as conductance are still used for characterization of ionic equilibria in solutions, either alone (1-4) or in combination with ultrasonic relaxation (5). Recently spectroscopic techniques have been shown to be very useful tools for such investigations. They are now extensively used to obtain qualitative information about the kinds of interactions which are predominant in solutions (solvent-solvent, solvent—solute, solute-solute interactions) as well as quantitative data about equilibria in solution (ion pair formation, complexation, kinetics of complexation). 2. Vibrational Spectroscopy Evans and Lo (6) studied far infrared spectra of tetra- Pentyl- and tetrabutylammonium chlorides and bromides in benzene solutions, and observed bands which could not be attributed to a vibrational mode of either the solute or the solvent. They concluded that these bands were due to cation—anion ion pair vibration in solution. At the same time Edgell E: 3i- (7-8) likewise reported ion pair vibra— tion bands in cobalt tetracarbonyl and manganese pentacar— bonyl solutions in tetrahydrofuran. Popov and coworkers (9-17) extended these far infrared studies to a wide range of nonaqueous media. They found that in solvents of high solvating ability,such as dimethylsulfoxide or l-methyl- 2—pyrrolidone new far infrared bands were observed whosefre- quencies were dependent on the solvent and on the cation but were not affected by a change in the counter ion, thus representing vibration of a cation in a solvent cage ("solvation bands"). In solvents of medium and low donor ability the frequencies of such bands were occasionally anion-dependent. In these cases the anion penetrates the first solvation shell of the cation forming contact ion pairs. Tsatsas and Risen (18) observed far infrared bands for lithium and calcium ions in ethylene methacrylate ionic polymers. Recently, Edgell and coworkers (19) examined the infrared spectrum of thallium tetracarbonylcobaltate in several solvents as a function of temperature. Only a single ion site was found in dimethylformamide, dichloro— ethane and dimethylsulfoxide solutions. Several kinds of sites were found in tetrahydrofuran, acetonitrile and rdtromethane solution, including free ions, solvent separated ion pairs, contact ion pairs and triple ions. Barriol g3 al., (20) tried to calculate the frequency shift attributed to the solvent effect on an infrared band using the classical model of an anharmonic oscillator centered in the Onsager cavity. Corset 22.2i- (21—22) and Popov and coworkers (13) studied preferential solvation of the lithium ion in binary solvent mixtures (acetone, nitromethane). Recently Baum and Popov (23), extended the earlier work and used Raman spectroscopy, infrared spectroscopy, lithiumr7 NMR, and chlor- ine-35 NMR to study Li+ ion solvation in acetone-nitro— methane mixtures. They found that Li+ ion is solvated by four acetone molecules and calculated equilibrium constant values for the stepwise solvation reaction. 3. Nuclear Magnetic Resonance Nuclear magnetic resonance (NMR) has become a powerful tool for the investigation of electrolyte solutions and of complexation reactions. The chemical shifts and line widths of the nuclear resonances of various nuclei can yield information about ion-ion, ion-solvent and ion- ligand interaction. There have been many studies of solvent molecules or the solvated species by proton NMR, some typical examples can be found in References (24) to (27). However, relatively few studies of the magnetic resonances of nuclei other than protons have been reported. All the alkali metal and halide ions possess at least 7 23 39 one isotope with a magnetic nucleus, i.e., Li, Na, K, 87 133 19 35 79,81 127 Rb, Cs, F, Cl, Br and I. Deverell and Richards (28) studied the chemical shifts of 23Na, 39K, 87Rb and 133 Cs nuclei in aqueous solutions of alkali halides and nitrates as a function of salt concentration. The concentration dependences of the chemical shifts were attributed to interactions between the cations and anions in solutions. Recent use of high resolution NMR pulse Fourier transform techniques (29-30) has made possible the investigation of nuclei with low magnetic moment or low natural abundance. The exact nature of the chemical shift is not com- pletely understood at this time. The chemical shift arises from the fact that a magnetic nucleus may experience local magnetic fields. Such magnetic fields are due to the sur- rounding electronic motion as modified by chemical bonding and molecular association. 'Atomic chemical shifts may be considered to be a sum of diamagnetic term (0d) arising from the inner symmetrical electrons which set up a small magnetic field opposed to the large, externally applied field, and a paramagnetic term (op) arising from non- spherically symmetrical valence electron orbitals which restrict electronic motions. Jameson and Gutowsky (31) have discussed the apparent increase in observed chemical Shift with increasing atomic number and pointed out that the diamagnetic contribution to the chemical shift can be calculated and that the often predominant paramagnetic contribution is difficult to estimate. The chemical shift of 23 Na is dominated by the para- magnetic contribution. The range of chemical shifts ob- served is rather large, and the large quadrupole moment of the nucleus renders this nucleus a sensitive probe of the neighboring electronic environment. Sodium—23 NMR has been recently used to study nonaqueous electrolyte solutions. Popov and coworkers observed the chemical shifts of various sodium salts in neat and mixed nonaqueous solvents (32-36). They were able to identify contact ion 23Na pair formation and found an upfield change of the chemical shift with increasing concentration of sodium perchlorate solutions in various solvents. Similar results were obtained by Van Geet and Templeman for aqueous perchlorate solutions (37). Popov 32 El' also observed a linear relationship between Gutmann donor numbers* (38) 23 of various nonaqueous solvents and the Na chemical shifts (at infinite dilution) in those solvents (39). 23Na chemical shift as a function By monitoring the of the amount of water added, Van Geet (40) determined that the hydration number of the sodium ion is between three and four. * Gutmann's donor number (38) is the enthalpy of complex formation between the given solvent and antimony penta- chloride in 1,2—dichloroethane solution: 5 + SbClS $432953 s.snc15 Gutmann used the term "donicity" when referring to the donor ability of a solvent. Donor numbers of some im— portant solvents are given in Table I. Table I. Donor number (DN) and dielectric constant (e) of some important solvents (38). Solvent 1,2-dichloroethane Nitromethane Nitrobenzene Acetic anhydride Benzonitrile Acetonitrile Sulpholane Propionitrile Ethylene carbonate Acetone Ethyl acetate Tetrahydrofuran (THF) IfiJnethylformamide (DMF) Dimethylsulfoxide (DMSO) Pyridine Hexamethylphosphoramide Water" a Predicted by 23 Na NMR (39). DN 10. 11. 14. 14. 16. l6. 17. 17. 20. 26. 29. 33. 38. 18. l 2 0 (33.03) 10.1 10.0 34.8 20.7 25.2 38.0 42.0 27.7 89.1 81.0 Unlike sodium, the paramagnetic contribution to the 7Li chemical shift is small enough to cause the diamagnetic contribution to be equally important. Akitt and Downs (41) pointed out that the lithium nucleus should be highly suitable as a nuclear magnetic resonance probe because of its high sensitivity, enhanced by an exceptionally narrow line width for ionic solutions. Consequently, very ac- curate measurements of chemical shifts are possible. Lithium-7, and lithium-6 NMR techniques were used by Attalla and Eckstein to determine the isotopic ratios in isotopic mixtures (42). In an early work Graig and Richards (43) did not observe any significant differences in 7Li chendcal shifts of lithium chloride in different solvents’ (n: at varying salt concentration. The negative results arts probably due to the inaccuracy of their measurements. Marxiel g; 3;. (44) and Akitt and Downs (41) measured the -Hxi chemical shifts of solutions of lithium bromide and perchlorate in water and in eleven organic solvents and observed that the frequency of the 7Li resonance is indeed sensitive to the environment. Recently Cox 2E.El° (45-46) studied the 7L1 nuclear magnetic resonance of some aromatic ion pairs in various solvents. They discussed their results in terms of the type of ion pairs formed in solution and the possible structures of these ion pairs (46). Both 23Na NMR and 7 Li NMR have been found useful for the determination of formation constants of weak complexes (47‘48). It is clear that 7Li NMR may provide useful information concerning the presence and types of inter— actions in electrolyte solutions and can also be a tool for the investigation of complexation reactions. A more extensive historical and theoretical discussion on 7Li NMR can be found in the Ph.D. thesis of P. R. Handy (49). To complement alkali metal cation NMR, nuclear mag- netic resonance of anions can also provide information on electrolyte solutions. Dodgen 33 El- (50) reported that 35 the Cl chemical shift of HClO was -946i6 ppm from con— 4 centrated aqueous HCl solution, with a linewidth of 62.5 Hz. The same chemical shift was observed by Saito (51). In this case, however, the linewidth was found to be only 42 Hz. Richards 23 al., (52) studied chemical shifts and transverse nuclear relaxation times of 35C1 and 81Br in aqueous-methanol solutions of LiCl and LiBr. They observed that most of the line broadening is due to changes in the viscosity of the solutions with concentration. Langford and Stengle (53) studied the solvation of the chloride ion by 35 Cl NMR in mixtures of acetonitrile and dimethyl- sulfoxide with water. They observed that a nearly equivalent competition by the two solvents for the Cl' ion solvation sites occurs in both mixtures and that the immediate en— vironment of Cl— is related to the long—range structural aspects of the solvent mixtures. Deverell and Richards (54) surveyed the chemical 35 81 127 shifts of Cl, Br and I nuclei in aqueous solutions of alkali halides and postulateithat.direct cation-anion collisions were the predominant cause of the concentration dependence of the observed chemical shifts of potassium, rubidium and cesium halide salts. 'For lithium and sodium halide solutions they found that the ionic interactions were of prime importance for the determination of the chemical shift of the halogen. Recently, Hall (55) pointed out that halogen NMR can be a useful tool for the investigation of biochemical and biophysical processes and systems. Langford and coworkers (56) reported results on solvent and counterion 19F- chemical shifts. The collision processes dependence of concept allowed them to separate the large effect arising from collisions and the primary solvation sphere from the smaller effects of outer sphere interactions. B. COMPLEXATION OF ALKALI METAL IONS BY ORGANIC LIGANDS l. Tetrazoles and Glutarimides Compounds such as l,S-cyclopolymethylenetetrazoles (Figure 1A) and 3,3-disubstituted glutarimides (Figure 1B) are characterized by their strong stimulating action on the central nervous system. For example, the convulsant activity of polymethylenetetrazoles increases with increas— ing length of the polymethylene chain (57-58). Popov g3 §l° (59) investigated donor properties of tetrazoles in 1,2- dichloroethane by spectrophotometric measurements. They did not observe a correlation between the length of the hydrocarbon chain and the stability of the iodine complex. 10 Complexing ability with alkali metal ions could be an im- portant factor in the physiological activity of these com- pounds, if such molecules act as ionic carriers through the membranes of neural synapses. Alkali metal NMR has been used for the determination of formation constants of com- plexes of pentamethylenetetrazole with sodium (48) and lithium (47) in nitromethane, however more work is needed to determine whether or not convulsant activity of these drugs is due to their ability to complex alkali metal ions. Figure 1. Structure of (A) l,S-cyclopolymethylenetetrazoles, and (B) 3,3-disubstituted glutarimides. 2. Crowns and Cryptands In recent years, macrocyclic polyethers have been synthesized which have a remarkable ability to form very stable complexes with alkali metal cations. Not only do these compounds present considerable interest from a purely chemical point of View, but they can serve as models in simulating the processes which govern ion—transport through membranes 11 in biological systems (60). Cyclic polyethers, or "crown" ethers, developed by Pedersen (61L‘were the first such complexing agents to appear. A typical "crown" is shown in Figure 2 Figure 2. Dibenzo-lB-crown—6. The number 6 refers to the total number of oxygens and 18 to the total number of atoms in the polyether ring. Truter g; 21., determined the crystal structures of Sodium dibenzo-lB-crown-6 (62), potassium dibenzo-BO-crown-IO (62), and potassium benzo-lS-crown-S (63). They showed that in the solid state, the alkali metal ion is located in the Middle of the polyether ring. In solution, crowns are also likely to form a bidimensional complex where the cation lies in the center of the polyether ring. Pedersen (61) Studied crown complexes in solution by proton NMR and vibra- tional spectroscopy. He measured the solubility of dibenzo- + . l8~crown-6-K salts in seventeen nonaqueous solvents and 12 pointed out that the saturated cyclic polyethers have the useful property of solubilizing salts in aprotic solvents. Dye 32 El° (64) using this property, reported a new tech~ nique for dissolving alkali metals in solvents, such as ethers, in which they are ordinarily either insoluble or only slightly soluble. Pedersen (65) found that crowns do not necessarily form only one to one complexes with metal ions. Thus with dicyclo-lB-crown-6, in addition to the 1:1 complexes, he also obtained 2:1 complexes, Eiflir dibenzo-lB-crown-6-K+, and 3:2 complexes ELSL’ dibenzo-lB-crown-G-Cs+. The 2:1 and 3:2 complexes may have a "sandwich" structure. Further work by Pedersen (66) showed that dicyclohexyl- lS-crown-S is the best complexing agent of the crown type for Na+ ion; on the other hand, dicyclohexyl-lG-crown-S, is the most specific crown for Na+ as compared with other alkali cations. Frensdorff obtained formation constants Of crown complexes in aqueous and methanolic solution by Potentiometry (67), and in chloroform by picrate extrac- tions (68). Risen 23 El- (69) made an interesting far infrared study of dibenzo-lB—crown—6 complexes of sodium and potassium. They found a cation dependent band at 167 l for sodium. This vibration mm‘1 for potassium and 213 cm— iS due to the crown—encaged cation and is solvent and cation independent. Shchori 3E 3;. (70) monitored the complexation of Sodium ion by dibenzo—l8—crown—6 in dimethylformamide by 13 23Na NMR measurements. They were unable to observe using two peaks because the line width of the complexed Na+ ion was very broad, although the line shape analysis indicated that exchange was slow. They reported an activation energy of 12.5 kcal. mole—l. Wong, Konizer and Smid (71) studied these systems by using proton NMR with several etheral solvents and pyridine. They observed two sets of ligand protons, one set corresponding to complexed crown and the other to uncomplexed crown. The spectra were analyzed at the coalescence temperature, and results were obtained consistent with those of Shchori eE El' (70). Pedersen and Frensdorff (72) noted that very few data are available on complexation reactions in solvents less polar than methanol, where ion-pair formation becomes sig- nificant so that anion effects would be appreciable. Smid 33 El. (73) investigated the interactions of alkali metal ions and their fluorenyl ion pairs with crown ethers in THF. They concluded that crown ether complexes represent convenient systems to study ion or ion—pair interactions With neutral molecules in both aqueous and nonaqueous media. The mechanism of complexation reaction of the crown ethers was investigated by Chock (74). Enthalpy of hydra- tion involved in the desolvation step and the binding energy fOr a given ligand were found to be the two most important faCtors in the mechanism of complexation and the stability 0f the resulting complex. These results are in agreement 14 with the work of Cussler E£.2l° (75» who precisely measured conductances of sodium, potassium and cesium salts in methanol and acetonitrile solutions containing cyclic poly- ethers (dibenzo-lS-crown-6), and obtained the association constants of the respective complexes. They also observed that dibenzo-lB-crown-6 complexes selectively sodium as compared to potassium in methanol but in acetonitrile it complexes both cations evenly; thus the observed selectivity is not only a function of the ionic size of the complexed ligand but also of the solvent used. Recently Simmons gE §l° (76) published an interesting paper about the applica- tion of crowns in substitution reactions using potassium hydroxide complexes of dicyclo-lB-crown-G. The complexa- tion of the cation enhances the reactivity of the anion (OH’). At the height of the interest in crown ethers, Lehn and coworkers (77-78) introduced a new class of complexing agents: diaza-polyoxamacrocycles and macrobicycles called crthands. Cryptands are bicyclic molecules in which the length 0f the ether bridges may be changed to vary the size of the cavity inside the cryptand to accommodate different Cations. The word cryptand refers to the ligand and cryptate t0 the complex. Cryptand 222, 221, 211 are shown in Figure 3. Like crowns, cryptands have aroused the interest of biolo- giSts as carriers in ion selective membranes (79). Their 15 .Amv nooosoao sus>to hams“ out HAN .Hmm .NNN osooosuo .m ousoaa $395+: Cass—80 z/\/o/\/z Ao<§S+oz A2381 Ca 2:20 risomwmo .omldo so mpflaflnflpmwomsm oaumcmmz How cofluomuwou paw mmfluwodowm ucm>Hom SUM .HH Tahoe 24 volumetric flask (1 ml, 2 ml or 5 ml) and then introduced into a dry—box for subsequent manipulation. E. INSTRUMENTAL MEASUREMENTS l. Lithium-7 NMR Lithium—7 nuclear magnetic resonance measurements were obtained using a Varian Associates DA-6O spectrometer operating at a field of 1.4092T and a frequency of 23287 MHz. The spectrometer was frequency locked to an approp— riate reference solution (4.0M LiClO4 in water, 3.5M LiClO4 in nitromethane, 3.5M LiClO4 in acetone or 5.0M LiClO4 in methanol) contained in a 1 mm melting point capillary and centered in the 5 mm NMR tube by Delrin spacers. All the chemical shifts reported in this thesis are with respect to 4.0M LiClO4 aqueous solution or aqueous LiClO4 at infinite dilution, as specified. A positive shift from the reference is upfield. The chemical shifts reported are corrected for dif—. ferences in bulk diamagnetic susceptibility between sample and reference according to the following equation: 6 = 6 + 21(xref _ Xsample corr obs v v I (1) ref sample V where XV and X are the volume susceptibility of the reference and sample solutions respectively and Gobs and 6corr are the observed and the corrected chemical 25 shifts. Values of acorr were calculated on the basis of published magnetic susceptibilities of various solvents (106). Several of these values were confirmed by using the method of Live and Chan (107) which involves the mea- surement of the chemical shift of lithium salt solutions at two different field strengths. The magnitude of the correction for various solvents is shown in Table II. When the contribution of the salt to the magnetic suscept- ibility of the solution could not be assumed negligible, volume susceptibility of each solution was measured using a Guoy balance. Temperatures were measured with a cali- brated thermocouple. Pressurized NMR tubes (30 to 60 psi of N2) were used when it was necessary to record a spectrum above the boiling point of the solution. 2.Chlorine-35 NMR Chlorine-35 spectra were obtained with the DA-6O spectrometer operating at a field of 1.0378T and a frequency of 4.33 MHz. Modulation frequencies in the ranges 20-30 Hz and 800-1000 Hz were used, depending on the linewidth to be observed. Care was taken to avoid modulation broad— ening; the modulation amplitude was progressively reduced until the width of the resonance being observed showed no further narrowing. The radio frequency power and sweep rate were also optimized. All experiments were done at r00m temperature (25°C). Cylindrical nonspinning sample tubes of about 15 mm diameter were used. Spectra were 26 calibrated using the high frequency sidebands in the case of relatively narrow lines. Line widths were determined with an estimated accuracy of :10% as an average of two to four measurements. Viscosities of lithium perchlorate solutions were measured with an Ostwald viscometer at 250C in a constant temperature bath. 3. Infrared Spectra Far infrared measurements (600—50 cm_l) were per- formed on a Digilab FTS-16 Fourier transform spectrometer. The theory and operation of this instrument have been described by P. Handy (49). Most of the spectra were obtained at a nominal resolution of either 2 or 4 cm_l. A standard demountable Barnes liquid cell with polyethylene windows and a 0.1 mm path length was used. 4. Laser Raman Spectra Raman spectra were obtained on the Spex Ramalog 4 Laser-Raman system equipped with the Spectra-Physics model 164 argon-ion laser. The 5145 A line was employed for excitation and data were obtained in the pulse counting mode with a nominal resolution of 2-4 cm—l. Samples were injected into 1.6—1.8 x 90 mm melting point capillary tUbes and sealed. 27 5. Data Handling Extensive use of the CDC-6500 computer was made to evaluate data. Program KINFIT (108) was employed to determine complexation constants (Appendix II), exchange rates and activation energies (AppendiszII). CHAPTER III SPECTROSCOPIC STUDIES OF IONIC INTERACTIONS BY LITHIUM-7 NMR, CHLORINE-35 NMR AND RAMAN SPECTROSCOPY A. INTRODUCTION Previous studies in this laboratory (35) and elsewhere (28, 40, 109) have shown that sodium-23 NMR offers a very sensitive probe of the environment of sodium ions in various solvents and solvent mixtures. The purpose of this study is to extend such investigations to salts of other alkali metal ions in order to determine the influence of the cation on the ionic equilibria and ionic species present in various nonaqueous solvents. The exchange of ions between different environments is usually rapid with respect to the NMR time scale, result- ing in only one resonance signal at an average frequency determined by the magnetic shielding and lifetime of the nucleus in each of the sites. Alteration of parameters such as concentration, counter ions and solvent produces changes in the relative proportion and type of environment which may be reflected by a change in chemical shift and/or line shape and/or line width of the observed resonance. 7Li nucleus are quite favorable for The properties of NMR studies. The resonance lines of Li+ ion in solutions are exceptionally narrow and chemical shiftscxnlbe measured with considerable accuracy (41). B. RESULTS AND DISCUSSION 7 Variation of the Li chemical shifts with concentration for various salts in various solvents is shown in Appendix 28 29 7Li chemical shifts I. In dimethylformamide (DMF) the for lithium perchlorate, chloride, bromide and iodide are essentially independent of the counter ion and of the concen— tration (Figure 6). Somewhat similar behavior is found in propylene carbonate, methanol, and dinethylsulfoxide (DMSO) (Figures 6—7). In tetrahydrofuran (THF), there is little concentration dependence but a significant difference in the chemical shifts of various lithium salts (Figure 8). Acetonitrile (CH3CN) and nitromethane (CH3N02) solutions 7Li chemical shift on show considerable dependence of the the concentration and on the counter ion (Figure 9). Since the limit of detection of 7Li resonance with our instrument is ~0.01M, it was difficult to establish the limiting chemical shifts of Li+ ions in such solvents as acetonitrile, THF, nitromethane and acetone where ion pair formation was especially evident. It has been shown previously by electrical conductance measurements that in general, the triiodide salts behave as strong electrolytes in nonaqueous solutions with high and medium dielectric constants (110). It was found that the chemical shifts Of lithium triiodide were essentially independent of concentration in all solvents tried and, therefore, the Values of the chemical shifts in 0.02 M LiI3 solutions are reasonably close to the limiting chemical shifts of the Li+ ion in the same solvents (Table III). It should be noted that in cases where extrapolations of chemical 30 1.0 - PROPYLENE CARBONATE ‘\\\‘ ::45; CL\-€:::::::::;5!::——__ x. __43_ 0.0-J DMF I -—LO-« 0) (:“()4— x ll- 0 Br: El CI O 13— <> BPh4 —20 1 1 r 1 1 0.0 0.1 0.2 0.3 0.4 0.5 CONC M Figure 6. 7Li chemical shifts of lithium salts in propylene carbonate and dimethylformamide. 31 1.5 - DMSO ’O_0 O - O —o— k _ * 4— (z 1’0 69% x— ~0— E O. 2 MeOH Q—o___ oath—W 0- 0.5 . fl? .0 0'04— 0 Br" x I- D CI- (9 13- 0'0 i r l (>1 Bph4- l 0.0 0.1 0.2 0.3 0.4 0.5 0.6 CONCM Figure 7. 7Li chemical Shifts of lithium salts in dimethyl- sulfoxide and methanol. 32 THF 1.0- W 01)u E W #3. a x. . .. < o CIO4 0 Br— 1.0- x [- El Cl— C) _ 3 _ <> BPh 4 2'0 I I I T I I 01) 0.1 (12 0J3 Ov4 OAS (16 CONC M FiSUre 8. 7Li chemical shifts of lithium salts in tetra— hydrofuran. 33 3.0. E D. 0- 10- CH NO 3 2 Q 4 001%):(0- 0.0- M «:10. x I _ )x\ 0 BL \x (913 _ <>BPh4 -—LO - - - r I 0.0 0.1 0.2 0.3 0.4 0.5 CO NC M Figure 9. 7Li Chemical shifts of lithium salts in aceto— nitrile and nitromethane. 34 Table III. Limiting Chemical shifts of 7L1 (0.02M LiI3) in various solvents. Dielectri Solvent 6(ppm)a Donor Numberb Constant (25%C) Acetonitrile +2.80 14.1 37.5 Dimethylsulfoxide +1.01 29.8 46.68 Propylene Carbonate +0.61 15.1 65.0 Tetrahydrofuran +0.60 20.0 7.58 Methanol +0.54 25.7C 32.7 Nitromethane +0.36 2.7 35.87 Acetic Acid +0.03 ---- 6.2 Dimethylformamide -0.45 30.9 36.71 Tetramethylguanidine -0.63 --—- ll Acetone -l.34 17.0 20.7 Pyridine -2.54 33.1 12.4 aZ§.4.OM aqueous LiClO4 as external standard. Corrected for bulk susceptibility. bRef. (38). C Ref. (35). 35 shifts to infinite dilution are possible, the value obtained for LiI3 agrees well with the extrapolated value (for example, in dimethylformamide, methanol and propylene carbonate solu- tions). The chemical shift values obtained in this investiga— tion agree reasonably well with those reported by Maciel 3: 21. (44) and by Akitt and Downs (41). Our values, however, seem to be uniformly displaced by m0.2 ppm towards higher field. This difference may be due to the fact that we have included corrections for the bulk susceptibility of the solvents. It should be noted that in most solvents with low or 7Li chemical shifts intermediate solvating ability, the are strongly influenced by the presence of even small amounts of water. It has been shown by Akitt and Downs (41) that the addition of small amounts of water to lithium perchlorate solutions in pyridine results in a sharp upfield shift of the 7Li resonance. It was found that this effect is even more drastic in non—solvating solvents such as nitromethane. Since most organic solvents cannot be obtained completely anhydrous, it is obvious that in SUCh cases meaningful chemical shifts for Li+ ion can only be obtained if the concentration of water is much smaller than the concentration of the salt. It has been previously observed (35, 111) that the cOntact ion pair equilibrium strongly depends on the donor ability of the solvent molecule as well as on the bulk _¥—_ 36 dielectric constant of the medium. Although nitromethane has a high dielectric constant of 36, its donor ability is very low and on Gutmann's scale (38), its donor number is 2.7. We see from Figure 9 that the chemical shifts of lithium perchlorate and lithium iodide are concentration dependent and, therefore, that there is contact ion pair formation. There is also some evidence for contact ion pair formation in lithium iodide and bromide solutions in propylene carbonate (Figure 6), (dielectric constant 65, donor number 15.1). These results are in agreement with the data obtained with 23Na NMR in propylene carbonate solutions where the chemical shifts of sodium bromide and of sodium thiocyanate are strongly concentration-dependent (17). On the other hand, in dimethylformamide, with a high dielectric constant of 36.7 and a donor number of 30.9, there is very little influence on the chemical shift by the counter ion or by concentration (Figure 6). In fact, the chemical shifts for the perchlorate, bromide and iodide are essentially superimposable and only the chloride shows some evidence of contact ion pair formation. Tetrahydrofuran is an interesting solvent in that it has a low dielectric constant of 7.58 but a respectable donor number of 20.0. The similarity of the chemical shifts for the chloride, bromide and iodide (Figure 8» and the fact that they are downfield from the perchlorate, may indicate contact ion pair formation. The low dielectric constant would preclude any ion pair dissociation in the concentration 37 range used (0.01—0.6M). In fact, conductance measurements reported in a previous paper (35) showed that sodium— anion ion pairs do not dissociate in the same concentration range. In the case of acetic acid solutions (Figure 10) we see very little concentration or counter ion dependence of the 7Li chemical shift. In fact, the greatest difference we observe is 010.2 ppm between 0.5M solutions of the per— chlorate and the triiodide. Acetic acid, however, is a solvesnt of low dielectric constant (6.3 at 25°C) and it is naitural to expect that there will be a considerable amourit of ionic association in this medium. It seems reascanable to assume that in this case we have largely solveant—separated ion pairs and that a very slight concen- tratuion dependence indicates an equilibrium between solvent- sepalrated ion pairs and a small amount of contact ion pairs. At tile limiting concentration of 0.01M essentially only SolV’EEnt-separated ion pairs exist in solution. This asstunption is strongly supported by a previously reported Obsfierrvation from this laboratory that in acetic acid solu— ti0r1 the frequency of the lithium ion vibration in a solvent Cage is independent of the nature of the counter ion and 6 Conuess at 390 cm-1 for 7Li salts and at 407 cm-1 for Li SaJVtJS (14). On the other hand it has been shown that the freflqllency of the solvation band is anion—dependent when IthEE éanion penetrates the inner solvation shell to form a C=01'ltact ion pair (7,17,69). A PPM Figure 38 05- ACETIC ACID . Q A OIL .‘.—WEr——“‘———CI_______________-__—_:V Ch . TMG- -05 " V .' if W _]0_ 0 CK); 0 Br— x [—— 0 Cl 9 13‘ _ <> BPh .15 - 4' I - - 00 OJ 02 Q3 Q4 05 CONC M 10. 7Li chemical shift of lithium salts in acetic acid and tetramethylguanidine. 39 We mentioned above that we attribute the concentration dependence of the 7Li chemical shifts to the formation of contact ion pairs, iiiir to cases where the anion directly replaces a solvent molecule or molecules in the inner solvation shell of the cation. It seems reasonable to assume that gross variations in the chemical shift of an alkali metal ion is a direct influence of the change in its immediate chemical environment, irgir it reflects the influence of its nearest neighbors. It has been shown, for example, that NMR techniques for the determination of solvation numbers of ions invariably yield numbers indica— tive of the inner solvation shell. Thus the hydration number of magnesium (II) ion was found to be six by the NMR measurements (112) while electrical conductance technique, which also reflects the contribution of the outer solvation (113) shell, yields a solvation number of 14-15. 35 The results of the Cl NMR study are shOWn in Table IV. In general, in dilute solutions the width at half height 35 of the Cl resonance was 10—20 Hz as compared with 45 or 65 Hz reported previously for aqueous solutions. It is 35Cl seen that there is a considerable broadening of the resonance with increasing concentration of the salt in nitromethane and to some extent in tetrahydrofuran and tetramethylguanidine. On the other hand, very little con— centration dependence is evident in acetone, methanol or acetonitrile solutions. It seems that the concentration—dependent broadening 40 Table IV. Line broadening solutions. Solvent Acetone Methanol Nitromethane Tetrahydrofuran Tetramethylguanidine Conc (M) 0.26 0.52 1.00 1.49 2.03 2.52 3.02 3.50 4.00 0.49 1.00 1.99 4.00 5.98 0.26 0.51 1.01 1.50 2.00 2.50 3.02 0.26 0.49 1.44 2.01 1.00 of 35Cl resonance in LiClO4 Wl/2a n(cp)sample Wl/2a (obs,Hz) n(cp)solvent (corr,Hz) 11 1.12 10 16 1.21 13 27 2.07 13 38 2.63 14 55 3.80 14 84 6.22 13 142 10.82 13 257 19.29 13 485 32.16 14 19 1.30 15 19 1.37 14 26 1.90 14 43 3.76 11 121 9.14 13 85 1.04 82 120 1.16 103 178 1.40 127 220 1.67 132 260 1.97 132 304 2.46 124 325 2.66 122 53 1.19 44 63 1.39 45 202 3.14 64 349 5.19 67 280 5.52 51 41 Table IV - continued a Conc w1/2 n(cp)sample 0(cp)solvent a Wl/2 (corr,Hz) Solvent (M) (obs,Hz) Acetonitrile 0.26 15 0.50 27 1.00 36 aWidth at half height. 1.10 1.26 1.67 14 21 22 42 3f the 35 C1 resonance is indicative of contact ion pair formation for the reasons given below. In theory, spin-lattice relaxation mechanisms may be iivided into five categories (114% dipole—dipole relaxa- tion, chemical shift anisotropy, scalar coupling, spin- rotation relaxation, and quadrupole relaxation. Since the abserved relaxation time is assumed to be in the motionally larrowed limit so that Tl equals T2, the observed results nay be expressed as: 1 _ _ TI _ R1 exp — R1 dip-dip + Rl scalar + R1 chem shift anis . + +R1 spln rot Rl quad (2) In general, the relaxation mechanisms have a Hamiltonian Jperator HC = -hI.TC.0 (3) Ihere I is the spin of interest, 0 is a physical quantity Ihich interacts with I to provide the relaxation mechanism, Ind Tc is the coupling interaction tensor. For the case >f the dipole-dipole relaxation mechanism we have H = hI.T.S (4) C 43 vhere S is the nuclear spin of the species coupled to the spin of the nucleus of interest. Clearly, this relaxation mechanism is negligible when compared to smallest observed 35C1 in the 35 3 {l for C104- anion. Since the interaction falls off as r_ , where r is the distance between spins, 35C1 and nuclei other lipole—dipole interactions between :han nearest neighbors need not be considered, at least :0 a first approximation. For chemical shift anisotropy to be a significant 'elaxation mechanism, the chemical shift of the species lust vary radically with orientation in the magnetic field. 'he C1047 anion cannot satisfy this criterion, since even L fairly large deviation from T symmetry does not seem d :o affect chemical shift significantly. Scalar coupling of the first kind is due to the ef— ‘ects of chemical exchange of the species of interest, »r of nuclei directly coupled to it. This is of no relevance 0 35Cl resonance of the C104— anion, since the nonexchang— ng oxygens shield the chlorine effectively from chemical xchange effects. In contrast, scalar coupling of the ecxond kind would be due to magnetic field fluctuations at 17 he (chlorine nucleus due to motion of the 0 nuclear spins, llich should be negligible compared to the observed R1 5 a consequence of the extreme magnetic dilution of 17O. In the isotropic molecular reorientation limit, the ontribution from the spin rotation mechanism may be ex- ressed as 44 ZUIkT h2 2 = ( )ceff Te (5) Rl spin rot vhere I is moment of inertia of the C104_ anion, Ceff is the spin rotation coupling constant and T is the angular 0 :orrelation time. A calculation using the largest reason- is able estimates of 1 spin rot T6 and Ceff shows that R several orders of magnitude smaller than the smallest )bserved R . 1 exp The term involving quadrupole relaxation may be ex- >ressed as 2 2 2 + R :R2=43—0—2—-2L(1+L)(-e_Q£-) Tc (5) I (21-1) 3 h Where I is the nuclear spin of the observed species, n is :he assymmetry parameter, eZQq/h is the quadrupole coupling :onstant, and TC is the translational correlation time. ‘or 35C 104—, the assymmetry parameter is zero, or at most Very small for small distortions from Td symmetry, and the :erms involving the nuclear spin are of course constant. 'herefore concentration dependence of R must depend l quad In the influence of concentration on either the correlation .ime, TC, or the quadrupole coupling constant, or both. or pure Td symmetry, eZQq/h is zero, since this term may 150 be expressed as (eQ/h)(32V/3Z2) (115), and the electric ield gradient at the chlorine nucleus is obviously zero or pure tetrahedral symmetry. Therefore, for the quad— upole relaxation mechanism to make any contribution to R1, some distortion from Td symmetry must be involved. Changes in this distortion with concentration and, of course, any change in R would be due to an ion pair phenomenon 1 quad rather than classical solvation or bulk viscosity effects. Also it should be noted that changes in the electric field gradient at the chlorine nucleus affect the R as the 1 quad square of the perturbation, while TC is influential only to the first power. Several authors have assumed that this effect on the translational correlation time might be ninimized by using a correction term linear in bulk vis- cosity, but recently (116,117) this correction term has been questioned on the basis that bulk viscosity does not accurately reflect changes in TC, especially for relatively 2 high concentrations (>10_ M). Despite this, the correction is still useful in predicting the sign and order of magnitude of the change in R due to the concentration effects 1 quad on TC. 35 In conclusion, it seems that the Cl NMR data support 7Li chemical our assumption that the concentration-dependent shifts are indicative of the contact ion pair formation. Phis is particularly evident in nitromethane as shown in ?igure 11. It is, of course, possible that the second learest neighbors (solvent shared ion pairs) may slightly influence the chemical shift of the alkali nucleus, however, the predominant effect must be due to the nearest neighbors. Similar results were recently obtained by Stengle and co— vorkers (118) who have studied sodium, lithium and 46 I!) In. Appm 0. o' 0. °. . 'o E fiF—‘Fé O) D: “E E o —D_ 2E '06 :9 NC”) W o—{j— d-Cl— A El —o- .3; z O U —Cl- _C? '\ _{j<3\\\$>clr I .TJk—e I I fig—Lg. o o o o o o o s 2 r: N ° °° 0' (2H)H10|M aNII Figure 11. 7Li chemical shift and 35C1 NMR resonance line width of LiClO4 in nitromethane as a function of lithium concentration. 47 particularly magnesium perchlorates in nonaqueous solvents 35 by C1 NMR. An interesting difference was observed in the relation 23Na and 7Li chemical shifts in different between the solvents and the Gutmann donor numbers (38). It was pointed out in previous publications that a plot of the limiting 23Na chemical shifts gs donor numbers yieldsa respectable straight line. It is interesting to note that no such cor— relation is observable with the 7Li chemical shifts in different solvents (Table 3, page 34). For example, the limiting chemical shift of a poor donor, nitromethane, is between the limiting chemical shifts of two excellent donors, dimethylsulfoxide and pyridine. Several workers (109) have pointed out that in the case of sodium, the paramagnetic screening constant is dominant over the diamagnetic screening constant. For lithium, however, this is not the case; the diamagnetic and paramagnetic terms are of the same order of magnitude, and tend to cancel one another (31). For this reason, affects such as ring currents and neighbor—anisotropy affects become more important for lithium chemical shifts, is pointed out by Maciel s3 sl. (44). For example, in :he case of pyridine, the deshielding of the lithium lucleus may be accounted for by the effect of the circulat— _ng electrons, if it is assumed that the lithium nucleus .5 coordinated with the nitrogen in the plane of the ring. Conversely, the upfield shift of the lithium when 48 coordinated to acetonitrile may be attributed to a strong neighbor-anisotropy effect analogous to the extraordinary shielding of acetylene protons. Evidence for contact ion pair formation, in the case of lithium perchlorate—acetone system, was also ob— tained by Popov and coworkers (13) from the behavior of 1 tflue 935 cm_ Raman band corresponding to the symmetrical stretch of the C104- ion (01, A1). They observed that the 9355 cm_1 perchlorate Raman band splits upon formation of ccnitact ion pairs. Greenberg (39), monitoring the three eaisily observable perchlorate bands at 935, 460 and 626 cut—1 respectively, in NaClO4 solutions, observed a split 1 band in solvents which gave indication cxf contact ion pairs formation by 23Na NMR; acetonitrile, of the 460 cm" tertrahydrofuran and pyridine. He observed only a marked brrbadening of the 935 cm_1 band. It is clear then, that thee formation of contact ion pairs, which perturbs the Td synunetry of the C104— ion, can be observed by Raman spec- troscopy. We monitored the 935 cm—1 perchlorate Raman band as a function of LiClO4 concentration in nitromethane, methanol, acetonitrile and tetramethylguanidine (Figures 12 and 13). At high concentration of LiClO4 the 935 cm-1 band splits and a new band appears at 938, 940 and 946 cm-1 in tetramethylguanidine, methanol and nitromethane, 1 respectively. In acetonitrile the 935 cm_ band becomes unsymmetrical when the LiClO4 concentration reaches 1.0 M. ¥ 49 NEAT SOLVENT 3.0 M 0.50 M 0.25 M IKE: l l l l I 920 960 930 A FREQUENCY (cm—I) FiSure 12. Raman spectra of the C102 935 cm—1 band as a function of LiClO4 concentration in nitro— methane. .O_ neat solvent neat solvent 0 ::;\/ 980 900950 'fi 950 950' ' THO 0 as w >\ 0 % 5 0. 25M - Q 1 I; I p U! % . A FREQUENCY (cm-l) Figure 13. Raman spectra of the C10; 935 cm—1 band as a function of LiClOA concentration in, (A) methanol, (B) tetramethylguanidine, (C) acetonitrile. 51 These observations confirm the results obtained by 7Li NMR in nitromethane and tetramethylguanidine where indica— tions of contact ion pairing have been found. In methanol the appearance of the new band occurs only at concentra- tions higher than 4.0 M, where the deficiency of solvent molecules causes the formation of contact ion pairs. CHAPTER IV SPECTROSCOPIC STUDIES OF COMPLEXATION OF ALKALI METAL IONS I. LITHIUM-7 NMR STUDY OF THE Li+ COMPLEX WITH CONVULSANT POLYMETHYLENE TETRAZOLES AND GLUTARIMIDES IN NITROMETHANE SOLUTIONS A. INTRODUCTION Both l,5-polymethylenetetrazoles and glutarimides (E>age 10) are convulsant agents. Small changes in the srflostituent groups, however, may change drastically the convulsant activity of the compounds. For example, the convulsant activity of polymethylenetetrazoles increases with increasing length of the polymethylene chain. The minimum convulsant dose varies from 1000 mg/kg for tri— methylenetetrazole to 30 mg/kg for heptamethylenetetrazole (57). (Tetrazoles with larger methylene chains are in- soluble in water.) While the mechanism of convulsant activity of the above compounds remains unknown, there is a possibility that their interaction with the alkali metal ions of the cerLtral nervous system may be an important factor in their physiological activity. Consequently we investigated the interactions of alkali metal ions with convulsant drugs. Initial studies on alkali ion-tetrazole systems (47) showed that while tetrazoles do form complexes with alkali Inetal ions, such complexes are very weak and cannot be studied by conventional potentiometric or spectroscopic tecflnaiques. On the other hand, alkali metal NMR Spectra 52 53 are very sensitive probes of the immediate environment of tflie alkali metal ion and the addition of a tetrazole to a Li. or Na+ salt solution results in a definite shift of tIle 7Li or 23Na resonance. In order to maximize the interaction between the drugs arui the lithium ion, measurements were carried out in nitro— nuethane,which is a poorly solvating solvent (although with a luigh dielectric constant of 35.87) and thus offers a Ininimum of competition for the complexation reaction with the drug molecules. B. RESULTS AND DISCUSSION We assume that in all solutions containing the drug molecule and Li+ ion, the latter is found in two environ— ments: free solvated lithium ion and the complexed lithium ion. The exchange between the two environments is fast as conqpared to the NMR time scale, therefore, only the mass— average chemical shift is observed Gobs ‘ M M MLXML (7) r . . . whe e aobs IS the observed chemical shift, XM and XML are respectively the mole fractions of the free and complexed metal ion while 6M and 5ML are the respective chemical shifts for the two species. Assuming a 1:1 complex, we have the equilibrium 54 where L is the ligand. The formation constant of the complex, in concentration units, becomes K = ML (9) where CM and CML are the equilibrium concentrations of free ligand and of the complex respectively. Equation (7) can be written as: 2 2 2 2 L _ t_ t_ 2t t_ tt t t 2 Gobs - (KCM KCL l) i (K CL +K CM 2K CMCL+2KCL+2KCM+1) l—L + 5 (10) Since 6M can be easily determined from measurements on solutions of lithium salts without the ligand, and C5 and CE, respectively, the total concentration of metal ion and ligand, are known, Eq. (10) contains two unknowns K and 6ML° For a fairly strong complex GML can be determined experimentally by the addition of such excess of L that essentially all of the metal is in the complexed form. For a weak complex, however, either the limiting shift value is unobtainable or such large excess of L is needed that the solution loses even a semblance of ideality. The procedure we use for solving Eq. (10) is to sub— ct Ct M’ L and 0M stitute the experimental parameters dobs’ 55 and vary K and 6ML until the calculated chemical shifts correspond to the experimental values within the error limits on concentration (0.01 M) and shifts (0.05 ppm). The data were analyzed on a CDC-6500 computer using the Fortran IV program KINFIT (108). A more detailed derivation and the description of the subroutine EQN used are given in Appendix II. A typical plot of experimental points and of the com— puter-generated curve for Li+-3,3-dimethyl glutarimide system is shown in Figure 14. It is seen that a satisfactory agreement is obtained between the calculated and the ex— perimental values. Similar plots were obtained for other systems. The results are given in Table V. Both glutarimides and polymethylenetetrazoles form complexes with the lithium ion in a non—solvating (or poor donor) solvent such as nitromethane. In aqueous solutions, where alkali cations are much more strongly solvated, the competition between the solvent molecules and the rela- tively weak ligand, glutarimides or tetrazoles, may be heavily weighted in favor of the solvent and the complexation constant would be much smaller. As seen from Table V, the formation constants of the lithium—drug complex are dependent on the total concentra— tion of the lithium salt, CS. The inconstancy of the Kf values may be due to (a) activity effects, (b) formation of more than one complex in solution and (c) competing equi— libria involving Li+C104— ion pairs. Since the 060 56 T010 9.'o 20 530 do M LIGAND/M Li+ 40 3'0 2D 10 050- Figure 14. 040- 20- 010d 00 030- -010- -020 de V Plot of 7Li chemical shift with reference to 4.0 M aqueous LiClO4 gs mole ratio of 3,3— dimethyl glutarimide to Li+ at constant [Li+]= 0.100 M. Solid line is computer—generated curve and dofs are the experimental points. 57 Table V. Formation Constants of Li+-Tetrazoles and Li+— Glutarimides Complexes. Formation Constants of Li+-Tetrazoles Conc of Ligand LiClO4(M) Thrimethylenetetrazole 0.000 0.050 0.104 Pentamethylenetetrazole 0.000 0.011 0.050 0.100 Hexamethylenetetrazole 0.000 0.011 0.104 Ligand Conc Range (M) b 0-2.3 O-3.0 0—0.9 0~2.5 O—3.2 b 0-1.0 0-1.5 Formation Constants of Li+-Glutarimides Glutarimide 0.000 0.010 0.050 0.099 3-Ethyl-3—methyl glutarimide o . 000 0.010 0.049 0.100 3:3-Dimethyl glutarimide 0.012a b 0—1.0 0-1.0 0-1.4 0-0.8 0-0.8 0-1.0 0-1.0 Kf (M 6.25:0. 4.97:0. 3.73:0. : 4.85:0. 4.54:0. 3.19:0. 1.72:0. 5.05:0. 4.91:0. 3.85:0. -1 ) 10 10 05 10 10 09 10 10 07 06 a Complex precipitated out at higher lithium concentrations. Extrapolated value. 58 complexation reaction does not involve separation of charges and since the solutions are relatively dilute, the activity effects should be minimal. The formation (of 2:1 complex seems to be excluded by the very good agree- rnent between the experimental and calculated chemical sllifts, especially at high ligand/Li+ mole ratios where frarmation of 2:1 complex should be more apparent. In addi— tion, results from 7 Li NMR in various solvents presented in Chapter I of this thesis, strongly indicate ion pair formation in nitromethane solutions of lithium perchlorate. Consequently, we assume that the competing equilibrium with Li+ClO4_ ion pairs is the principal cause for the variation of Kf values with salt concentration. Plots of Kf gs. lithium ion concentration are essen— tially linear (Figure 15). Extrapolation to infinite dilu— tion should yield quasi-thermodynamic constants for the complexation reactions. The lithium complexes of polymethylenetetrazoles appear to be more stable than the corresponding complexes of the sodium ion. For example, the formation constant of the pentamethylenetetrazole—Na+ complex in nitromethane is 0.76 (48) as compared with 4.85 for lithium. 59 10.0 concentration. 3—methyl glutarimide. zole. 1 O '0' 2| < m 0 0 In 2 -Q O O 0 f- .J -5. I ’ O / ’ I I I I, I ,’ , O ,1 - I ,L 1* I Iés I I I E; SEQ Q0 0 o 009 o o_ 0 Lo 8:0 not N o to}; v m N .— I (m) x Figure 15. Plot of formation constants vs lithium ion A — Glutarimide. B - 3—ethyl- C — Pentamethylenetetra— 60 II. STUDY OF THE COMPLEXATION OF ALKALI METAL IONS BY CRYPTAND LIGANDS IN VARIOUS SOLVENTS A. LITHIUM-7 NMR STUDY 1. Lithium—7 Chemical Shift of Lithium 222, 221, 211 Cryptates in Various Solvents The 7Li chemical shifts were determined as a function of cryptand/Li+ mole ratios with the results shown in Table VI. Typical spectra obtained with cryptand C211 are shown in Figure 16. The stability of a cryptate complex is largely de— termined by the size of the cryptand cavity and the sol- vating ability of the solvent. If the rate of exchange of the lithium ion between the two sites, free ion in the bulk solution and the complex, is greater than Z/FAV, whereAv is the difference between the characteristic resonance (in Hz) of each site, only one population—average resonance is observed. This is the case with C222 which has a much larger cavity (2.8 A) than the bare lithium ion (1.56 2). Only one 7Li resonance is observed in nitro— methane, dimethylsulfoxide, pyridine, and water solutions. In dimethylsulfoxide and aqueous solutions, the sol— vent molecules have a strong solvating ability and compete quite successfully with the ligand. Consequently, only a weak lithium complex is formed and a large excess of 61 Table VI. 7Li- NMR Study of C222, C221, C211, Lithium Complexes in Various Solvents at Room Tempera— ture. Cryptand 222 + + 7Li Chemical Solvent Salt [Li J (M) [CryptandJ/[Li J Shift (Ppm)a CH3NO2 LiClO4 0.025 0.0 0.35 0.5 0.75 1.0 1.02 2.0 1.03 DMSO LiClO4 0.025 0.0 0.97 0.5 0.97 1.0 0.96 2.0 0.96 Pyridine LiClO4 0.025 0.0 —1.52 0.7 0.30 1.0 1.04 2.5 1.61 00(1)) 1.73 H20 LiI 0.010 0.0 0.00 1.0 0.005 10.0 0.095 20.0 0.11 mm 0.18 Cryptand 221 CH3NO2 LiClO4 0.05 0.0 0.38 0.81 1.04 1.03 62 Table VI - Continued Cryptand 221 — Continued + 7Li Chemical Solvent Salt [Li J (M) [CryptandJ/[Li'j Shift (ppm)a DDdSO LiClO4 0.05 0.0 0.94 0.5 0.96 1.0 0.97 2.0 0.98 Pyridine LiClO4 0.05 0.0 -2.16 0.5 -2.21, 1.87 Cryptand 211 CH3NO2 LiClO4 0.15 0.0 0.61 1.0 0.41 LiI 0.14 0.0 0.49 1.0 0.42 LiI3 0.14 0.0 0.11 0.4 0.01 and 0.37 0.9 0.37 LiCl 0.13 1.0 0.41 DMSO LiClO4 0.20 0.0 0.97 0.8 0.95 and 0 39 1.0 0.39 LiI 0.14 0.0 0.95 1.0 0 39 63" Table VI - Continued Caryptand 211 - Continued 7 1 Chemical L. S ion pair formation. On the other hand, the limiting chemical shifts of I¢i+-C222 and especially Li+—C221 complexes are definitely Sc>lvent-dependent indicating that the looser structure Of the complex permits the solvent molecules to approach sufficiently close to the metal ion to affect its resonance frequency. Lithium NMR has been shown to be a useful technique for the determination of the formation constants of weak and medium strength complexes (Chapter IV-I). This approach 67 vvas used in this work to determine the formation constants (3f Li+—C222 complexes in water and pyridine. The technique involves the measurement of 7Li chemical shifts as a func— ‘tion of ligand/Li+ mole ratio (Table VII), followed by a (:omputer fit of the data as described in Appendix II. The Iglot of experimental points of the computer—generated curve for Li+—C222 in pyridine is shown in Figure 17. The values obtained were: log Kf = 0.99i0.15 and log I/(Li+) 6(ppm)a (222)/uso owumuwcomluwpsmeoo mH mafia oflaom ME mm.o u h+fiqw ucmumcoo um wcfiofluhm :fl +HA Op NNN ocmumxuo mo oflumu wHoE m> voHoflq msowswm 2 o.v ou wocwuwwwu aufls uMHnm HmoHEwno figs mo uon .ha wusmflm +_._ .3285. ohm 0..“ o: 0.0 o.mu no._I o V Au .0 ..o.o 0 yo.— o.w 70 2. Lithium NMR Study of the Lithium Ion-Lithium Cryp— tate Exchange in Various Solvents The drastic effect of the solvent on the complexation reaction has been illustrated (Chapter IV—II—A—l) by the difference in the formation constants of the Li+-C222 complex in water and in pyridine. Thus, the strong solvating ability of water drastically depresses the value of the Li+-C222 formation constant. We have seen also that the lithium ion forms stable complexes with the cryptand 211, and the ex— change between the free and the bound lithium ion is slow on the NMR time scale. Thus,solutions containing lithium in excess can be examined by measuring the 7Li resonances. The exchange kinetics can be deduced from changes in line shapes as a function of temperature. The complexation process between a ligand and a cation M+ in a solvent S can be represented by the following general equation, L + M+ K; LM++solvent (12) solv £- which assumes a first order process for the backward re— action, ngL in our case, the dissociation reaction or the r81ease of the lithium ion from the cryptate cavity. Such a mechanism was found to be predominant for the complexa— tion of sodium ions by dicyclohexy1—18-crown-6 (121). The general case of exchange between two sites A and B with 71 different relaxation times is described by the following modified Bloch equations (98,122) G = u + iv (13) SU + TV V = “YH M ——~———- (14) 1 0(82 + T2) UT - SV u = -yH M ——————- (15) l 0(82 + T2) where G is the complex moment of magnetization, u and v are the pure absorption and pure dispersion line shapes,respec— tively, and P P A B T S = —-— + -—— + -———— - (w - w)(m - w) (16) TZA T23 TZATZB A B U = l + (pB/T2A + pA/TZB) (17) (mA - w) (wB - w) T = (pAwA + PBwB - w) + T + (18) T T 2B 2A V = (p80)A + pAwB - w) (19) Where pA and pB are the relative populations atsites A and B,reSpectively,and T is the mean lifetime of the interaction defined by, Pi +3fi a w (20) 3’ CU 72 The quantities w and wB are the resonance frequencies A in radians per second at the two sites at a given tempera— ture in the absence of exchange and T2A and T2B are the respective relaxation times at each site at a given tempera- ture in the absence of exchange. If at a given temperature the lifetime T is greater than /2/(0Aw), where Aw = IwA - wBl, two separate resonances are observed for the two respective sites; if T is less than /2/(nAw), only one population-averaged resonance is observed. Since it was experimentally difficult and inconvenient in the case of coalescing lines to obtain a pure absorption mode signal, a phase correction was made and the observed line shape was fitted by the following equation: V = usine + vcose + c (21) where 0 is the phase correction parameter and c the base line adjustment parameter. Spectra with C211, obtained at selected temperatures in pyridine, formamide, dimethylformamide, dimethylsulfoxide and water are shown respectively in Figures 18 to 22. Spectra with C221 obtained at selected temperatures in pyridine are shown in Figure 23. Spectra were analyzed by using the Fortran IV KINFIT program (108) based on a generalized weighted non-linear least—squares analysis. A more detailed derivation and subroutine EQN are given in Appendix III. Each spectrum was fitted with four parameters, 73 :1 ll __ ______ W181J°C _H__. ].()ppm Figure 18. Lithium-7 NMR spectra of 0.50 M LiClO , 0.25 M C211 solution in pyridine at various tempera- tures. 74 ~fl'fl#J\\U-FN*dWM'J/\\-Mflfl 105.5OC 120.5 °c M 131.5 0C 143.3 °c [I r--—--fi 0.5;nom Figure 19. Lithium-7 NMR spectra of 0.50 g LiClO4, 0.25 94 C211 solution in formamide at various tem— peratures. : 25.5°c M 105.5°c 12o.5°c M 137.5% A Lil-4°C (0 H —-——i 0.5 ppm Figure 20. Lithium-7 NMR spectra of 0.50 g LiClO , 0.25 g C211 solution in dimethylformamide at various temperatures. e 28.5°C M 86'9OC M 124.50C A 139.50C 157.0°c 168.5% m H ——————+ f——I 0.5 ppm Figure 21. Lithium—7 NMR spectra 0.50 La LiClO4, 0.25 r_4_ C211 solution in dimethylsulfoxide at various temperatures. 77 I ll } \ 25.80C M 62_3°C M 80.3°C A 109.2°c A 119.2°c A 128-5°C A H 02ppm Figure 22. Lithium-7 NMR spectra of 0.50 g LiI, 0.25 g C211 solution in water at various temperatures. Figure 23. l. _____ l. A... l ________ l 70.1°C WA...“ 126.6 °C 142.7 °C 153.4 °c 2.0 ppm 1. Lithium—7 NMR spectra of 0.50 g LiClO4, 0.25 M C221 solution in pyridine at various tem— peratures. 79 the lifetime T, the phase correction 0, the base line adjustment c and a normalization factor. A typical com— puter output (Figure 24) shows the fit of a spectrum (LiClO c211) in formamide at 143.3°c. 4, No exchange is observed for the Li+—C211 systems at room temperature. Measurable exchange, detected by the onset of broadening of both resonance lines, begins at higher temperatures, tE which were found to be about 75°, 80°, 105°, 145°, and 85°C in dimethylsulfoxide, water, formamide, pyridine and dimethylformamide respectively. It was noted that wA and wB varied linearly with tempera- ture relative to the lock frequency. The difference between the chemical shift (ppm) of the solvated ion A and that of the complexed ion B is a linear function of temperature as expressed by 6 - 53 E A(6) = 4(60) - S(t-25) (22) in Vvhich A(60) is 5A — GBat 25°C. The values of A(6O) and.:3 are given in Table VIII at temperatures higher than tEr “A and “B were obtained by extrapolation. The validity Of this extrapolation was verified in the case of the DMSO SYStem by separate experiments on solutions which contained Orlly A or B. The T and T values given in Table VIII 2A 28 were determined by measurement of the full width at half height of each resonance line and were found to be tempera— ture independent. It should be noted that the cryptate 80 .pon on“ mo COHMSHOmmH wnu Canvas wEmm one mum ucflom pmumaso Iamo ocm Hmucwfiflummxw cm mcme u .ucflom omumasoamo m mcme o .ucflom Hmucmfiflummxm cm mcme x .00m.mva um mpHEmEHOM :H HHNu m mN.o .eoHoeq a om.o rues emcempno manomem no new nousesoo I .em museum I CY. X Ox X 0 Ema to o x amplitude 0< OX 0 OX O>< OX 81 sme.o eee.o wesee.e- ewm.m1 Ham eoHUeg meeceuse esm.o mom.o osmoo.o- OHO.H Ham eoHoeq meesmauom mme.o Heo.H emmoo.o- mam.o HHN eoHqu moesmsuOMHsrumeHo esm.o mom.o Hwaoo.o mmm.o- HHN eoHoeu weexoWHsmesruwaea emm.o eme.o Nmmoo.o- emm.o HHN Heq umumz sme.o ems.o emeoo.o- mme.m Ham eoHoea mceeeusm Aoomv Roomy Ema mme awe 00\Eee m Aceva ecmuesno “Ham ua0>aom m a o Asofl owxmamfioo n m .cofl omum>HOm u do 00mm us © 1 m n A mv< Aoov u musumumemu co>fim m pm me I do u va< 1mm - now + leecs u lees .mugmuomEmB mo coflpoqsm m mm S: mo 833.8me .3: 0369 82 resonance line is 2 to 3 times broader than that for the solvated ion. Therefore, the width of the cryptate resonance line cannot be entirely caused by field inhomo- geneities. In the case of the Li+-C221 system in pyridine, T and T2B were measured by separate experiments “’A’ ”B' 2A since some exchange occurs at room temperature. Activation energy plots, log kb vs 1/L are shown in Figure 25. Activation energies (Ea), rate constants (kb), and values of AH:, As: and AG: for the release of Li+ from the cryptate are given in Table IX. A complete error analysis (123), including cross correlation terms which ac- count for the coupling of parameters, (particularly evident loetween AH: and 15:) was performed in all cases. Not sur— prisingly, AGi, which is directly determined from kb, has the smallest standard deviation. The accuracy of the determination of the activation energy depends upon the range of temperatures over which the exchange can be measured. For example, with the Li+— C211 system in pyridine (Figure 18) the difference between the chemical shift of the solvated and the complexed lithium ion is 2.7 ppm, which is considerably larger than the shifts found in other solvents (0.5—l ppm). Thus, for this system the exchange is observable over only a limited temperature range and coalescence could not be observed. Therefore, the activation energy could only be determined from the line broadening below the coalescence, which accounts for 35 30 25- b(sec—1) M (0 Loqk Figure 25. 83 23 2Q 23 2b 27 2b 2b 3b 3 4 1/Tx10 (°K ) Log k vs l/T plot for 0.50 g LiClO4, 0.25 g C211 in (A) pyridine, (B) formamide, (C) di— methylformamide, (D) dimethylsulfoxide; (E) 0.50 M LiI, 0.25 g C211 in water, and (F) 0.50 E LiClO4, 0.25 5 C221 in pyridine. 84 le.ec m.sH lm~.oo m.o~ .me.e. e.o~ lee.e. h.aH lm~.eo o.e~ AH.H. e.- Asommmo H1H9: Hung O me + Am.ov m.va1 Am.Hv m.~m1 Av.HV m.mHI Av.Hv m.ma1 Ao.mv v.o + Am.mv m.NH| Av.ov m.NH Ah.ov m.ma Am.ov v.mH Aw.ov m.mH AH.HV b.0N Av.mv o.mH HOE Hmox H: o m< + Amway OMNH Am.NV v.5 Am.mv o.ma Av.mv «.mm Ao.NV m.v Avm.ov NH.o lxowmmee-oom nee x ex Av.ov m.ma Ah.ov H.va aw.ov o.wH Am.ov H.©H AN.HV m.HN Avoim.mv m.mH Hoe Hmox H1 8 m .coflumfl>wc oumvcmuwo .ucmumcoo ofiuuowawwon .Ammo “096:2 nocoo :cmEusom .mu:w>aom muoHuw> :H omcmgoxm oumumxuo Edenuflq mo mumuwsmumm owEmnzcoaumna cam H.mm mcflpfluhm HNN ecmumxuo o.vm woflsmsuom w.w~ mpfifimsuowaxnumafio m.m~ weexouflsmesrumEMQ o.mm Hmumz H.mm wcflnfluam HHN cumumwmu Amvzo ucm>How mwumm wwcmgoxm .xH anwe 85 the relatively large standard deviation. On the other hand, for the Li+—C221 system in the same solvent (Figure 23), the exchange is observable over a large temperature range, 62.60C to 159.4OC, and the activation energy can be determined with a higher accuracy. The energies of activation for the release of Li+ from the Li+-C211 complexes in pyridine, water, dimethyl- sulfoxide, dimethylformamide and formamide, seem to be determined by the donicity of the solvent as expressed by the Gutmann donor number (38), rather than the dielectric constant. The activation energies vary from 14.1 kcal -1 1 in water mol in formamide (DN = 24) to 21.3 kcal mol— (DN = 33.0). By contrast, Shchori, gt al. (121), found that the (smaller) activation energies for the release of Na+ from several l8-crown-6 complexing agents were indepen- dent of the solvent used. However, two of the three sol- vents used, methanol and dimethylformamide have the same donicity,while that of the third solvent, dimethylethane, is not known. We expect that the net-energy required to transfer Li+ from the cryptate to the solvent should decrease with increasing donicity of the solvent since this scale is a good measure of the primary solvation energy. The solva— tion energy of the cryptate and the secondary solvation energy of the lithium ion. both depend primarily upon the dielectric constant of the solvent and change in the same direction from solvent to solvent. 86 Since the activation energy increases with increasing donicity, opposite to the overall energy change, the transi— tion state must involve substantial ionic solvation. The energy profile is illustrated schematically in Figure 26. The solid line represents the complexation path of Li+- C211 in a poor donor solvent (82) and the dashed line in a good donor solvent (S1) with reverse activation energies of E l and Ea a , respectively, (E > E 2). In the same sol— 2 a1 a vent, for example 52, the energy level of the 221 cryptate will be higher than for the cryptate 211 because of the better fit of the lithium ion in the C211 cavity. On the otherhand if the transition state is similar to the sol- vated lithium ion, its energy should not depend strongly upon the cryptand used. The energy of activation for Li+-C221 (Ea3) is lower than that for Li+-C211 (Eaz) in pyridine, 13.5 and 19.6 kcal mol—l, respectively. Although Ea’ and hence AHt, are very sensitive to the solvent used, the values of AG: (2980K) are nearly inde- pendent of solvent. Changes in AH: are compensated for by corresponding changes in As:, a not uncommon occurrence (124). Using the formation constant of the Li+-C211 cryptate in water determined by Lehn and coworkers (80), log K = 5.3, we can calculate the rate constant for the forward 3 -1 reaction, kf = Kk = 0.98 x 10 sec for Li+-C211 in b water. 87 ‘~ Li ($2) ———+ reaction coordinate Figure 26. .\ \ \ \ l \ A A '- N O") . m m m Lu Lu LLl \ , \ . \ L1 C22](S2) \ \\ x J. Li c211(52) \ \\ \ J LEJC2H(S]) Schematic representation of the complexation S 1 of lithium ion by cryptands 211 and 221. represents a good donor solvent, 82 a poor one . 88 B. FAR INFRARED AND RAMAN STUDY OF LITHIUM AND SODIUM CRYPTATES IN NONAQUEOUS SOLVENTS 1. Introduction Far infrared spectroscopy has been used extensively (6-17) for the investigations of the motion of alkali metal cations relative to their immediate environment. In general, variations of the cation's motional frequencies in solution are expected to occur with variations in the immediate en- vironment of the cation, and are indicative of interactions occurring in solutions. For example,cation-dependent fre— quencies can be either dependent or independent of the anion,indicating the presence or absence of the anion in the first solvation sphere of the cation, iLEL, formation of the contact ion pair. When an alkali ion is complexed by a macrocyclic poly— ether, the ligand insulates it from the medium and a cation— ligand, rather than cation-solvent vibration is observed. Risen and coworkers (69) investigated the far infrared spectra of the sodium and potassium—dibenzo-18-crown—6 systemsixlseveral solvents and found a band whose frequency was solvent independent. 2. Results and Discussion a. Salts in solution. Far infrared spectra (100—500 cm‘l) of sodium salts (NaBPh4 and NaI) in pyridine, DMSO and nitromethane and of lithium salts in nitromethane, 89 pyridine and acetonitrile are represented in Figures 27 and 28 respectively. Frequencies of the major far infra- red bands, observed in the 100-500 cm-1 region, are reported in Table X. For sodium ion solutions in pyridine and DMSO, anion independent and solvent dependent bands are observed at l in pyridine and at 205 cm.1 in DMSO. These bands 180 cm— represent the vibrations of the sodium cation in a solvent cage (9—17). Pyridine and DMSO are solvents of high don— icity and no significant concentration of contact ion pairs is present in these solutions. In nitromethane, a poor donor solvent, an anion dependent band is observed at 135 4 and at 148 cm—1 for NaPF6. The anion dependence of the band frequency indicates significant cm_1 for NaBPh anion participation in the near-neighbor environment of the cation. For lithium salts in acetonitrile and pyridine, cation 1 1 dependent bands are observed at 381 cm- and 387 cm— , respectively, due to the solvated cation. In nitromethane solutions the frequency of the lithium ion vibration is strongly anion-dependent. The bands are found at 358 cm—1 1 band in for LiClO4 and 330 cm-1 for LiI. The 420 cm— pyridine is interpreted to be an activated complexed pyri— dine band (16). The other bands, reported in Table X, are due to solvent or internal vibration modes. 90 PYRIDINE 200 cm III \\'II III-II I‘lul'nl I 100 300 200 cm4 100 pyridine BPh4 +X' salts (X' and of analogous 222 cryptates (dashed lines). Far infrared spectra of nitromethane, (solid lines) and DMSO solutions of Na and I‘) Figure 27. PYRIDINE s / \ 91 300 Figure 28. I *1 I I ‘fi 400 500 300 400 500 ch cmq Far infrared spectra of nitromethane, pyridine and acetonitrile solutions of Li+X’ salts (X- = C103 and 1‘) (solid lines) and of analogous 211 cryptates (dashed lines). 92 lmc+sae lmc+maa AmVva Am>vromv Am>vrmmm Am>v¥Nmm Am>vrnov Am>vromv Am>vtowv Am>vshov Am>vshov Am>vmov Am>vtowv Am>vrhov Am>vmov A3V+wwm Am>vsvmm Am>vahov Am>vawmm ABmeN ABVHmN ABVHwN mwflocwdwoum ocmm ocouum mum> 1 m> lme+emm lme+eam Am>vrawm Am>vomm A3v+hmm lm>e.amm Am>v+mmm lm>cmem Am>vwha lmaessm Am>vmma lmscmma Am>vmva .mcouum wcfipflumm wcflofiuhm zommo 02mmu ocfipfluwm 20mmo NOmeo OmSQ wcflcfluhm N omzo mafipflumm NOZMmo NOmeo ucm>H0m ow>H0me >HHMfluumm+ .EdflomE om.o om.o om.o om.o om.o om.o om.o om.o om.o om.o om.o om.o mN.o m 0:00 pawn ucm>H0wr 1 E .xmmB 1 3 Hmflq sodoaq Hmz asemmz mmmmz perm .mpcw>aom msoflum> as muamm Esflnpflq one Esfitom mo modem owumnmcH Hum .x wanna 93 b. Cryptates in Solution. The far infrared spectra (100-500 cm-l) of Na+—C222 complexes in pyridine, DMSO, and nitromethane and of Li+—C21l complexes in pyridine, acetonitrile and nitromethane are shown in Figures 27 and 28, respectively. The frequencies of major far infrared bands observed in this region for the sodium and lithium cryptates are given in Table XI. The addition of cryptand to a sodium or lithium ion solution results in the disappearance of the cation-solvent vibrational band and the appearance of a new band whose frequency is both anion and solvent independent. For example, the Na+-C222 system in pyridine, DMSO and nitromethane has 1 a new band at 234:2 cm_ in all three solvents. Similarly, the Na+-C221 complex has a solvent and anion independent 1, the Li+-C221 complex, at 243:3 cm“1 and that of the Li+-C211 complex, at 348:1 cm_1. band at 218:1 cm— The substitution of 6Li for 7Li in pyridine solutions changes the frequency of the Li+—C211 band from 348 cm—1 1 confirming that the vibration involves the to 369 cm— lithium ion (Figure 29 and Table XI). These results are analogous to those of Tsatsas gt 31. (69) with the crown ethers. They indicate that the alkali metal ions are completely enclosed in the cryptand cavity and the new bands indicate the vibrations of the cations inside the cavity. It is interesting to note that the frequencies of sodium ion motion in the C222 and C221 cryp— tands are at higher frequency than those for the simple 94 Am>vrmmm Am>vrmvm mdoflum> CH Am>vrmmm Am>vabov Am>vremm Am>vtbov Am>vroam nm>vrmwm Am>vthov Am>v¥owv AH: mwumummuu How mpcmm owumumcH new “one: wnu mo AH Eov Am>crcoe Am>vsomv Am>vaomv Am>vrhoe Am>vaowv Am>vsvmm A3Vomm vamhm vaomm szomm Am>vrvmm sznmm ABVHmN Amv+mmm Amvramm Amvamm vamow Am>vthov Amvmmm Amvramm Amvmmm Am>vnHN Am>vmam Amvmmm Amvvmm “mommm Amvsmm vawmm Amvemm Eco mwflocoswoum ocmm Am>vomm Am>vomm Am>vmem Am>vmmm Am>vonm flm>vmvm Am>omvm Am>vmvm Am>vmha Am>voma Am>vena Am>voma Am>anH Am>vwba Am>vmoa Am>omea weapausm acmmo NOmeo NOZMmo masseuse masseuse zommo NOZMmo Omza masseuse OmZD masseuse wGIOmZQ Omzo wcflofluwm NOmeo pcm>aom cm.o mH.c mm.o 1H Hamuu+ag . 4 cm 0 -oao HHN01+an . e 1 cm s -oHo Heme +ags cm.o oc.o om.o moHo Hamoi+aq cc.o 1H ammou+mz me.o wcam ammoi+cz mN.c mH.o 1H NNNU1+cz mm.o mm.c om.o om.c wsdm mmmoi+mz Afiv mpmum>uo 0:00 . mpgwwaom mwflocmsgoum .Hx magma 95 machum >Hm> 1 m> Am>vrmmm Am>vawmm Ame/V show Am>v¥bov va+mwm AHIEOV moflocwsgoum ocwm AEvmvN AEVovN Am>vomm .mcouum 1 m omzo wcfiofluhm masseuse pcw>aom po>a0mwu waawfluumm+ pawn ucw>H0mr sgflUwE I E .Mww3 l 3 cm.c 1H amm61+aq cs.o onu ammoi+aq cm.c lam Hamoi+ae Ame oumummuu ocoo owsqflucoo .Hx manta 96 6Li“'—C211 7LiJF—C211 RAMAN 250 ' 3éo_1 450 cm ’ 6Li+—C211 1110211 FAR IR | i1 '1 1 s 300 ' 400_1 ' 550 cm Figure 29. Raman and far infrared spectra of 6Li-C21l and Li—Cle cryptates in nitromethane solu- tion. S — solvent band. 97 sodium salt solutions in the same solvents while the fre— quencies of lithium motion in C211 and C221 are generally at a lower frequency than those for the simple lithium salt solutions in the solvent (Figure 30). The only ex- ception is the frequency of the lithium iodide vibration in nitromethane where the salt exists largely as a contact ion pair. The fact that these frequencies reflect the strength of interaction of the alkali ion with the cryptand is i1- lustrated by the difference in the vibrational frequency between Li+—C221 and Li+-C2ll complexes which are observed at 243 and 348 cm-1, respectively. In the latter case, the Li+ ions fit exactly into the cryptand cavity and the more rigid structure results in a higher vibrational frequency. The other interesting feature of cryptate far infrared spectra is the appearance of another band whose frequency is solvent dependent but anion independent. Such bands are found at 161 cm—1, 176 cm-1 and 145 cm"1 for Na+—C222 in pyridine, dimethylsulfoxide and nitromethane, respectively, 1 l and at 388 cm_ and 385 cm_ for Li+—C211 in nitromethane and pyridine respectively. Substitution of DMSO by d6— DMSO shifts the 176:1 cm_1 band to 172:1 cm—l clearly indicating the participation of the solvent in the observed vibration. Such a band is not observed for Li+-C21l complex in acetonitrile solution; it is possibly masked by the strong 381 cm"1 solvent band of acetonitrile. It seems reasonable to assume that the observed band can be assigned 420—1 -—Ci~————Cl'—-\ — \ __A__A_\ \\ 380- \\ \\\\\ — —o——~--_- \ ““““ 9HLI—C 211 340— : / 348+1cm — ‘3 300 9 _ 9 a: 1— _ _J _J E o 260— ‘ HNa— —C222I 220_ /:I/:l” 2:34I-.-2C|’Y1_.l I’I’ll _ C C ,’ I I l ’ I ’I 130- __A—A—J 1’ I . . _ I A PyrIdIne I, I 140— II D CH3CN _ _g o Nitromethane O. 100— cc ._. 0 DMSO <5 to Z 2 Figure 30. Comparison of ion motion band frequencies for sodium and lithium salts and their 222 and 211 cryptates respectively in pyridine, aceto- nitrile, nitromethane, dimethylsulfoxide solu- tions. 99 to the vibration of solvent molecules which are in the first solvation shell of the cryptate moiety. Raman spectra were obtained for the Li+—C2ll cryptates in pyridine and nitromethane solutions and it was found 1 that the 348 cm_ infrared band is Raman—inactive. The data are shown in Figure 29 and Table XII. Two strong bands, 1 1 however, are observed at 367 cm- and 312 cm_ in both solvents. These bands cannot be assigned to a vibration of the cryptate or to a displaced solvent band. Their frequencies are unaffected by isotopic substitution of 6 7 Li for Li (Figure 29), and, therefore, they cannot be a displaced cryptate vibration. We feel that they represent activated cryptand vibrations. The fact that the lithium—cryptand vibration is Raman- inactive allows some speculation on the nature of the complexation interaction. It was pointed out by Edgell and coworkers (8) that the far infrared cation-solvent vibrational bands are Raman—inactive, indicating the electrostatic nature of the interaction. Since the same selection rules apply to the cation-cryptand vibration, it seems reasonable to conclude that in this case the cation-ligand interaction is predominantly electrostatic in nature. 100 Table XII. Raman Bands of Li+—C21l Complexes in Pyridine and Nitromethane Solutions Pyridine Conc Observed Bands(cm—l)in the Solute (g) 150 to 450 cm" region c211 0.50 406a:1 378a:1 325b:3 LiBr 0.55 406a:1 3783:2 Li+-C211 Br" 0.50 406a:1 378a'°:2 366 :2 311:1 298a'c:3 Nitromethane c211 0.45 320b:5 Li+—c211c1o; 0.55 367 :1 312:1 298°:2 7Li+—C21l Clo—0.60 367 :1 313:1— 2990:2 4 aSolvent band. bBroad band. CPartially resolved. .101 C. CONCLUSIONS AND SUGGESTIONS FOR FURTHER STUDIES Lithium—7 NMR has been applied to the study of ionic interactions of lithium in various solvents, to the deter— mination of complex formation constants of weak and rela- tively strong complexes, and, furthermore, to the investiga- tion of the kinetics of the complexation. Other techniques, such as 35C1 NMR, laser Raman spectroscopy and far infrared spectroscopy have been used to obtain complementary and additional information in order to more fully understand the role of the solvent in chemical processes. In light of the studies accomplished, the following suggestions can be made for further investigations: 1. Study of the variation of the 7Li chemical shift as a function of concentration in the low concentration range; (<0.01 M). The limiting chemical shift could then be reached even in solvents where ion pairs are formed. Pulse technique will have to be used, and the water content of each solution will have to be kept at an extremely low level so the 7Li chemical shift will not be affected. 2. Investigation of the kinetics of the formation of Li+—C211 complexes in low donor solvents to see if the same trend is observed between the activation energies and the donicity of the solvent in which the complexation is carried out. 3. Investigation of the kinetics of the complexation of Na+-C221 in various solvents for which the cation- 102 . . . + cryptand cavity fit 18 analogous to the flt between L1 and C211. 4. Determination of complex formation constants of cryptates in various solvents. Nuclear Magnetic Resonance technique could be used by either the direct method in the cases of a fast exchange and weak to relatively strong complexes, or by stepwise method, carrying out competitive complexation reactions. APPENDICES APPENDIX I LITHIUM—7 CHEMICAL SHIFTS ZS 4.0 M AQUEOUS LiClO4 of VARIOUS LITHIUM SALTS IN VARIOUS SOLVENTS. ACETONITRILE LiClO4 Conc (M) _EPE LiBr Conc (L4) ppm 0.500 2.17 0.503 1.52 0.250 2.29 0.376 1.62 0.123 2.33 0.251 1.59 0.062 2.49 0.126 1.67 0.031 2.55 0.075 1.72 0.016 2.60 0.037 1.77 0.025 1.87 E31114 0.509 1.75 [£13 0.501 2.55 0.254 1.99 0.251 2.59 0.127 2.19 0.125 2.71 0.064 2.35 0.063 2.72 0.032 2.46 0.031 2.75 0.016 2.49 0.016 2.76 Lg; 0.500 2.31 0.250 2.44 0.125 2.53 0.062 2.60 0.031 2.64 0.016 2.67 104 NITROMETHANE LiClO4 Conc (M) (5me LiI3 Conc (L4) Em 3.50* 0.90 0.506 0.07 3.00 0.93 0.253 0.20 2.50 0.91 0.126 0.30 2.00 0.89 0.063 0.34 1.50 0.88 0.032 0.36 1.00 0.85 0.016 0.38 0.510 0.82 0.300 0.78 Ll; 0.250* -0.63 0.200 0.74 0.125 -0.58 0.100 0.67 0.062 -0.48 0.082 0.65 0.031 -0.40 0.061 0.62 0.016 -0.34 0.030 0.52 0.020 0.47 0.014 0.41 £12324 0.250* 0.40 0.125 0.39 0.062 0.37 0.031 0.35 0.016 0.34 *Solubility limit 105 DIMETHYLSULFOXIDE LiClO4 Conc (M) :EPE LiBr Conc (M) EEPE 0.526 1.07 0.680 1.10 0.394 1.05 0.511 1.08 0.263 1.03 0.341 1.12 0.131 1.05 0.171 1.10 0.078 1.03 0.102 1.12 0.058 1.03 0.068 1.07 0.026 1.07 0.034 1.12 giggg4 0.497 1.05 Lil 0.500 1.00 0.248 1.03 0.250 1.03 0.124 1.06 0.125 1.04 0.062 1.06 0.062 1.04 0.031 1.05 0.031 1.05 0.016 1.06 0.016 0.96 1,2-DICHLOROETHANE 9191 0.500 0.90 0.250 0.93 LiBPh4 Conc (M) :m 0.125 0.97 0.496 1.18 0.062 1.02 0.248 0.82 0.031 1.03 0.124 0.80 0.016 1.04 METHANOL L1C104 6.50* 6.00 0.062 0.031 0.016 LiBPh4 0.250 0.125 0.062 0.016 *Solubility limit Conc (M) 106 LiCl LiBr Conc (M) 0.500 0.250 0.125 0.062 0.031 0.016 0.500 0.250 0.125 0.032 0.500 0.250 0.125 0.062 0.031 0.016 107 ACETIC ACID PROPYLENE CARBONATE LiClO4 Conc (M) 6 m LiClO4 Conc (M) 6 m 0.490 0.15 0.500 0.47 0.245 0.11 0.250 0.54 0.125 0.57 0.122 0.08 0.062 0.60 0.061 0.05 0.031 0.62 0.031 0.04 0.016 0.63 LiBPh4 0.505 0.01 LiBPh4 0.500 0.48 0.252 0.02 0.250 0.51 0.125 0.55 0.126 0.02 0.062 0.57 0.016 0.60 LiI 0.500 -0.02 LiBr 0.500 0.21 0.250 +0.00(4) 0.250 0.30 0.125 +0.03 0.125 0.37 0.062 +0.04 0.062 0.44 0.031 +0.06 0.031 0.51 0.016 0.55 LiI3 0.500 -0.12 LiI 0.500 0.26 0.250 -0.05 0.125 0.44 0.125 —0.03 0.062 0.51 0.062 —0.01 0.031 0.56 0.031 +0.01 0.016 0.61 0.016 +0.02 108 DIMETHYLFORMAMI DE LiClO4 Conc (M) SEER LiBr Conc (M) 022$ 0.500 -0.41 0.250 -0.41 0.500 -0.45 0.125 -0.42 0.250 -0.43 0.062 ~0.42 0.125 -0.43 0.031 -0.42 0.062 -0.42 0.016 -0.42 0.031 -0.41 0.016 —0.41 @114 0.503 -0.46 0.251 -0.44 all 0.500 -0.42 0.125 -0.42 0.250 -0.41 0.062 -0.41 0.125 —0.40 0.031 -0.39 0.062 -O.40 0.016 ~0.37 0.031 -0.40 Li_c1 0.500 -0.60 0.250 —0.57 0.125 -0.55 0.062 —0.52 0.031 —0.50 0.016 —0.49 109 TETRAHYDROFURAN LiClO4 Conc (M) 6 m LiBr Conc (M) EEET 2.00* 0.64 1.124 -0.76 1.50 0.66 0.562 -0.71 l.00 0.66 0.281 -0.69 0.44 0.64 0.140 -0.69 0.22 0.63 0.070 -0.63 0.110 0.63 0.035 -0.60 0.055 0.61 0.018 -0.60 0.028 0.63 0.014 0.61 Li: Conc (M) _EEE LiBPh4 0.510 0.62 0.500 -0.66 0.255 0.60 0.250 ~0.67 0.128 0.58 0.125 -0.69 0.062 -0.69 1.31 0.500* -0.53 0.031 -0.67 0.250 -0.53 0.016 -0.58 0.125 -0.52 0.062 —0.52 Li£3 0.500 0.50 0.031 -0.50 0.250 0.55 0.016 -0.48 0.125 0.60 0.062 0.61 0.031 0.60 0.016 0.58 *Solubility limit 110 ACETONE LiClO4 Conc (M) 5223 LiBr Conc (M) 522$ 0.418 -1.47 3.50 —0.76 0.314 -1.51 4.00 -0.85 0.209 -1.51 3.00 -0.89 0.105 -1.54 2.50 —0.90 0.063 -1.53 2.00 —1.00 0.042 —1.53 1.00 —1.05 0.021 -1.51 0.50 —l.06 9;; 0.500 -l.46 0.25 —1.07 0.250 -1.43 0.125 -1.14 0.125 —1.43 0.062 —1.14 0.062 -1.43 0.031 —l.18 0.031 —1.43 0.016 -1.24 0.016 -1.44 My: giggg4 0.504 -1.11 0.252 —1.18 £11 9929 (3) E239 0.126 -1.23 4.00 0.000 0.062 —1.25 1.00 0.064 0.031 -1.31 0.25 0.060 0'016 “1'31 0.06 0.081 0.015 0.086 111 TETRAMETHYLGUANIDINE LiClO4 Conc (M) 022$ Bi: Conc (M) 322$ 0.500 -0.44 0.500 -0.78 0.250 -0.44 0.250 -0.74 0.125 —0.43 0.125 -0.70 0.062 -0.41 0.062 -0.69 0.031 -0.41 0.031 —0.68 £i§3h4 0.500 -0.54 Lil3 0.500 —0.76 0.250 —0.55 0.250 -0.69 0.125 —0.55 0.125 -0.64 0.062 -0.58 0.062 -0.64 0.031 —0.58 0.031 -0.63 0.016 —0.65 L_i_§_1 0.500 —0.80 0.250 -0.77 0.125 -0.74 0.062 —0.71 0.031 —0.71 0.016 -0.71 112 PYRIDINE LiClO4 Conc (M) 32m LiBr Conc (M) SEPT 0.500 -2.01 0.500 -2.88 0.250 -2.07 0.250 -2.93 0.125 -2.12 0.125 -2.96 0.062 -2.13 0.062 -2.98 0.031 -2.17 0.031 -3.02 0.016 -2.22 0.016 -3.06 §i§3g4 0.498 —2.41 all 0.500 —2.82 0.249 -2.43 0.250 -2.78 0.124 -2.43 0.062 -2.74 0.062 -2.47 0.031 -2.70 0.031 -2.47 0.016 -2.67 0.016 -2.47 gig; 0.500 -2.68 0.250 -2.75 0.125 -2.79 0.062 -2.84 0.031 -2.84 0.016 -2.86 APPENDIX II DETERMINATION OF COMPLEX FORMATION CONSTANTS BY THE NMR TECHNIQUE, DESCRIPTION OF COMPUTER PROGRAM KINFIT AND SUB- ROUTINE EQUATION. Let's consider the following equilibrium for a one to one complex, k M + L 2? ML (23) kb with the concentration formation constant K K = cML / CM°CL (24) C, stands for concentration. The observed chemical shift of M (éobs) is a mass average of the characteristic chemical shift of M at each site (M in the bulk solution, and M complexed), assuming that a fast exchange occurs between these two sites with respect to the NMR time scale. = X 6 + X 6 (25) Where: 6M is the characteristic chemical shift for M in the bulk solution, 6ML is the characteristic chemical shift for M COmplexed (ML), XM is the fraction of M (CM/(CM + CML))’ 113 114 XML is the fraction of ML (CML/(CM + CML)), then 6Obs = XM6M + (l-XM)6ML ) + 6 (26) 6obs = XM(6M-6ML ML t _ . . CM — CM + CML (the analytical concentration of M)(27) CM 6obs = _E (GM—GML) + 6ML (28) C M Ct = C + C (the analytical concentration of L)(29) L ML L _ t _ CL ' CL CML . _ t t u51ng (27) and (29), CL — CL - (CM - CM) t C - C K = M f (30) t (CM)(CL — CM + c CM is solved in (30) t t _ t _ cM(cL cM + CM)K ~ CM cM 2 t t t _ KCM + (KcL KCM + 1)cm CM — o t t t t 2 t c 2 (KcL KCM + 1) : J(KCL KcM + 1) + 4KcM 14 2K the positive root is 115 t _ t _ 2 t2 2 t2 _ 2 t t t t c (KCM KCL 1) +J K cL + K CM 2K CLCM + 2KcL + 2KCM+1 M 2K (31) Substitution of CM from (31)’in Equation (28) _ t _ t _ 6dE-[m%4 Kg: 1)+ 2 t2 2 t2 _ 2 t t t t ] J K cL + K cM 2K CLCM + 2KcL + 2KcM +1 [(6 - 6 )/2CtK + 6 M ML M ML (32) We assume a constant value for 6M and that 6ML and K are unknown. In order to fit the calculated shift (the right hand side of Equation (32)) to the observed chemical shift, the program may vary the values of 5ML and K. Hence, the number of unknowns, NOUNK, equals two as does the number of variables, NOVAR. The first card contains the number of experimental points (columns 1—5 (F.15)), the maximum number of iterations allow— ed (columns 10—15 (F.15)), the number of constants (columns 36-40 (F.15)) and the maximum value of (A parameter/parameter) for convergence to be assumed (0.0001 works well) in columns 41-50 (F10.6). The second data card contains any title the user desires. The third data card contains the value of CONST(1) (CE) columns 1—10 (F10.6) in M, CONST(2) (6M) columns 11-20 (F10.6) other constants can be listed on columns 21—30, 31-40, etc. The fourth data card contains 116 the initial estimates of the unknowns U(l) = 6ML and U(2) = K, in columns 1-10 and 11—20 (F10.6), respectively. The fifth through N data cards contains XX(1) = c; in columns l-10 (F 10.6) variances on XX(1) in columns 11—20, XX(2) = the chemical shift at XX(1) in columns 31-40 (F 10.6) followed by the same parameters for the next data point. Each card may contain two data points. If no further data are to be analyzed the next card after the last data point(s) should be a blank card followed by a 6789 card. If more data sets are to be analyzed, the next card after the last data point(s) is the first data card of the next set. Generally the most common error using this program is an error MODE 4 in an address of the SQRTE subroutine. Usually this implies that the initial guess for K is quite inaccurate. If this becomes a problem, the argument of the square—root expres— sion could be tested to make certain that it does not be- come negative. Before using this program, the user should see Reference (108), or the materials of CEM 883, Chemical Kinetics, to become familiar with the mode of operation and further application of program KINFIT. 117 123 567 60 888 888 90 227 7?? 23 111 111 11 NNN NNN NN n 000 00 F FFF FF OVARvNOUNKOX0U~ITMAX9FQ ToFOpOFOQFUvDVZLnTOvF-FO F‘ 0 FQN 1332 ) n R M N 7 I N1 90 L Ox )7. TDT4( Y ) 909(0 R I UDNXT )( NVOX T )H OTInl, 0 Q. )0 PXwn) N ?I C9T)0 O (C NnSOO C ”C IXCO‘ (an XXN‘( N fi) OQIQL 0 ) )) PDTQ7 I 1 )- AYC( T ( ll LTFX. A T (6 IIVTI M S TF TvoWl) R N 50 WSY ab 0 0 ND IPUool F C 09 OE.)0( )0 CC FQY0?T F n!) (9 PDQ?(SP n n) I bH AAJl‘TNY )QI. ) 00 TIJHCOT N 2)( 1 2A 1011. EC! 0 fill)l\) (l. vnOQV I 4(X)I) /T (VIM) I 0 TI \Ivll‘IDQ‘I \ID pSOOo)) A 1%8(N( )n o AELOIOI N (N)XOT)ls ) 0 rD‘Qlll' )1 XDQVNCQIllO T 0H1.To?(5 6M XPO((N(H) I 00.V04¢Yc oR ((4960X..) F OETAO(OUQ HE #9)))CX)16 N 0 NnTXZ u DY ))2))(u?.( I OEU19 (93 AE ??(E?§)(HX K 0N0.TND)v TD 4 U(())TQX o CIKTSU 0?E01Jfl Ffl#(”d))§6. T 0T 9X101(lh6((72 U))fi((?(MA§ 71.! OHMFQQ.)( N:: VI::II\Vn/POFG(HH,A: FT OOOTLNOYOIEEEAKQPI((.00H(C(U Avl on.“ QAEn TTpn—TNKBV TV“? 0 Iii—((1 1F ORMXVMlv NAAIDHVTN((.??:=:(Q KN 0H0TGY()OOTTDOOOFO:==::AurzF (1 ORCHID” OGCIJ WFNNDCA RCQFARFSP HK 0 12 12 Cal 0 1! 6 7 30.0 NTAOO TTOGO prtnpu mm 0000000000009000033000 ?.0001 ’0 “9 oeoooonoococootoonoooooo lQQ6900011g731R?11026702 6050107511801010040.?610? 011?a§7l46715°4726094406 000000011112273344555788 ICOOOIUIOOOOOOOOOCOOOOOO 00.00.000.00...CIOOOCIOI. 35555:599559555555§§q995§ 27227222?22??2222222?722? E n n T 07978R26fi3309775749444011 M 097:1§918210q20763107h?P3 I .99996887777666555554444603 I.” I.can...contonooocauoucccepurv F. 33 M 91 luau h nv.v U 33 L“ n, C3 Cu 1000000 0033 I...odeco-0..canoe-cocooooooon.u 14 R 9., A n L 9 O 93 M an 5 94?433?14491?77?49817fi \ l3 28279012?190037944FFL¢(If 85701.:(1. 45(t.(..(2 .I). (1 .040000000111)]??23344567TR?9 nno.09.09.00...-cocoon-0.0.000 APPENDIX III NMR LINE SHAPE ANALYSIS FOR TWO SITE EXCHANGE. DESCRIPTION OF COMPUTER PROGRAM AND SUBROUTINE EQN. A. NMR LINE SHAPE ANALYSIS FOR TWO SITE EXCHANGE. Definitions G 2 complex moment of magnetization. A,B E Respective sites. pA'pB 5 Fractional population at site A and B. (pA + pB = l) TZA’ TZB Transverse relaxation times of nuclei at the two sites, in the absence of exchange. TA,TB Life time on site A and B at a given temperature. wA,wB Chemical shift at site A and B in the absence of exchange. y Gyromagnetic ratio. Hl Radio frequency magnetic field Mo(A B) Magnetization in the Z direction at both sites. I For two sites in the absence of exchange one can write (122): dGA IR? + O‘AGA = —1YHlMoA (33) dGB Tfi? + dBGB = -1yH1MOB (34) where 0A and GB are complex quantities defined by a =T‘1—i(o - ) (35) A 2A A w =T‘1-i( - ) (36) 0‘B 23 “B 9 118 119 Let's assume an exchange between site A and B following the two basic assumptions. 1. A11 nuclei remain in one site until they make a sudden rapid jump to another (nuclear precession during jump being neglected). Under these circum- stances, it is clear that a nuclear exchange between positions of the same type will have no effect. 2. It will be assumed that while a nucleus is in A position, there is a constant probability TA 1 per unit time of its making a jump to a B position. Fractional populations pA and pB are related to TA and TB by A TA + TB B TA + TB Modified Bloch equations proposed by McConnell (125) are A _ . —l _ -1 dt + dAGA — lYHlMoA + TB GB TA GA (37) dG B _ _ . —1 _ —l -aE-+ aBGB — 1YH1M0B + TA GA TB GB (38) Assuming a small radio frequency field H1, for slow passage conditions we have: dG dG A_ B= 71t— dt 0 (39) 120 Also M0A = pAMo’ MoB = pBMo Solving (33) for GA 0 G = - in M + T_1G - T_lG (40) A A 1 0A B B A A A GB 1 G = (~1leM0A + ;;)/(0A + ¥;) (41) substitute equation (41) into (38) and solve for GB 0 _iYHlMoA + —E _ _ - _ __________Ji__ ‘1 aBGB — lYHlMoB 1 TB GB (42) (“A + ?_)/TA A inlMoA (‘lYHlMoB - W) (TB) (OLA-[A + 1) GB = (43) (dBTB + l)(01AIA + l) - l and then from inspection in M . 1 0B (‘1YH1MoA ' aBTB+l )(TA)(aBTB + 1) GA = (44) (OIATA + 1)(01BTB + 1) -1 G = G + G = u + iv 121 in M - —fl) (TA) (aBrB + 1) + + 1 aBTB G = (-1yHlM0A in M - —1-LA) (TB) (aA—rA + 1)] / + 1 0‘ATA (—1YH1M0B (OLATA + l)(dBTB + l) - l (45) using pA = TA/TA + T3' p3 = TB/TA + TB pAMo = MoA pBMo = MoB p —_' B c — 1yHlMo (pA + ————————)(TA)(GBTB + l) + d + 1 P (pB + ————A——I)(TB)(GATA + lfl / “GATA + l)(dBTB + l) -1 aATA + (46) G = -inlMO (pAdBTB + pB + pAhA + pBaATA + pA + pB)TB]/[(GATA + l)(aBTB + l) -l (47) Vvith pA + pB = 1 G = —inlMo “pAaBTBTA + TA + pBaATATB + TB) / flaArA + l)(aBTB + 1) — 1 (48) 122 or [(nArA + 1) (6313 + 1) — 1] (49) as obtained first by Gutowsky, McCall and Slichter (126). We must now separate the imaginary part from the real in Equation (49) Define T = TATE/TA + TB Therefore TA = I/pB and TB = T/pA let y = G = TA + TB + TATB(aApB + “BPA) (50) -in M _ 1 o aAaBTATB + 1 + 0318 + dATA 1 dividing numerator and denominator by TA + TB T T T T A B A B TA + TB + 1A + TB + TA + TB(GAPB + 0‘BPA) TATB a a + TB + TA a + a TA TB A B TA + TB B TA + TB A 1 1 let k = ——— k = ——— A I TZA B T23 123 Y = 1 + TpB(kA — i(wA - w)) + pA(kB - i (wB -ww))/ [pA(kA - i(wA - on] + [(kB - imB - wHPB] + EAkB — i(kB(wA — w) + (6B — w)kA) — (6A — 00) (0B - 6)] r (52) Y = 1 + TkaA + pAkB — iT(PB(wA - w) + pA(0.)B - w))/pAkA + kaB + T [kAkB - (0)A - 0)) (01B - 01)] '- i[TkB(u)A - 00) + kA(0)B - 0)) + PA(wA - w) + pB(wB - wfl (53) Let S = pAkA + kaB + T[kAkB - (01A - 01) (0)8 - (0)] T = (PA + k3) (01A - w) + T(pB + kA) (01B - 0)) U = 1 + rka k A + pA B V = T[PB(UJA - w) + pA(wB - 01)] U - iv = (U — iv)(s + iT) Y: s — iT 52 + T2 (54) 124 SU + TV + i(UT - SV) $2 + T2 G = — inlMoY (56) G = - yHlMo i 53 + T; + 03 — SE (57) S + T S + T Since G = u + iv (58) UT - SV u = - yH M ——————— (59) l o [52 + T2] SU + TV v = - yH M ———-——— (60) 1 o [52 + T2] where pA pB r pA ’ PAPB(wA * M)(wB - w) (61) pA pB r S = ——— + ——— + - T(w - w)(w - w) (62) T2A T23 TZATZB A B P p U = 1 + T(—-—TB + T—A) (63) 2A 28 (“A - w) (wB - w) T = p (w - w) + p (w - w) +T + T (64) A A B B TZB T2A T ( ) “A - w ”B - w (6 ) = p w + p w - w p + p + T —————— + ————-— 5 A A B B A B TZB TZA T ( ) (wA - w wB - w) (6 ) = p w + p w - w + T -———-— + -————- 6 A A B B TZB TZA 125 V = (pBwA + pAwB - w) (67) G = u + iv (68) UT '- SV u = _ Ya M (___) <69) 1 0 S2 + T2 SU + TV V = - yH M <——-) (70) l 0 S2 + T2 P P A B T S = ——- + ——- + —-—-—- - T(w - w)(w - w) (71) TZA T23 TZATZB A B m - w w - w A B T=pw +pw -w+T(———+—-——) (72) A A B B TZB TZA P P U=1+T(T—B+T—‘3—) (73) 2A ZB V = 1(pBwA + pAwB - w) (74) Relation to a Complexation Equilibrium Let's consider the following equilibrium k M + L :f ML (75) kb where M is the observable nucleus which can be found, under certain circumstances in both sites M and ML the complex formation constant K is 126 In general the relaxation time 1 = rate of removal of molecule from i H! (76) H1 is given by (127a) th state, by exchange PI P- then ;L = §%%L§E TB ML L aM/at TA CM aML _ at ‘ kb CML aM _ _ 3E ‘ kf CM CL Therefore _£.= kb TB 1 —— — -k c TA f L T = TATE/TA + TB pA = TA/TA + TB pB = TB/TA + TB then number of molecules in the 1 'th state (77) (78) (79) (80) (81) (82) (83) 127 T = pATB _ pBTA with (83) p r = FA (84) b For equal population case pA = pB = 0.5 (84) becomes 1 T = _ (85) Zkb Therefore, from the value of T at a given temperature, kb can be obtained with (84). Thermodynamic parameters can then be calculated from the following equations (127b) aznk _ 2 TIT- — Ea/RT (86) V or P AHO* = Ea — RT (87) i _ _ 52 + ASO — Rln kb R fink) + AHO (88) i = + _ * AGO AHO TASo (89) where AGO+, AHO+ and ASO+ are the standard free energy of activation, the standard enthalpy of activation and the standard entropy of activation respectively. 128 Ea is the Arrhenius activation energy T the absolute temperature h the Plank's constant K the Bolzman's constant B. Determination of T Values from Line Shape Analysis of an NMR Spectrum at a Given Temperature. Description of Computer Program and Subroutine Equation. The lineshape is a mixture of the absorption (v) and dispersion (u) components V = usine + vcose + c (90) u and V are defined in Equations (69) and (70), respectively. The KINFIT program is used as described in Appendix II with the following parameters NCST = 6 (number of constants) CONST (l) = pA CONST (2) = pB CONST (3) = T2A CONST (4) — T2B CONST (5) = wA/Zfi CONST (6) = wB/Zn NOVAR = 2 (number of variables) XX (1) = w/ZW xx (2) = v 129 Each spectrum is fitted to four parameters (NOUNK = 4) U (l) = normalization factor U (2) = T U (3) = C(base line adjustment) U (4) = 6 C. Determination of the Activation Energy (EH). Descrip— tion of the Computer Program and Subroutine Equation. The Arrhenius activation energy (Ea) is given by 3£nk _ 2 —SE— - Ea/RT (91) v or p or -E_a l- 1 k = kref e R T Tref (92) where Tref is a reference temperature to which kref cor- responds, and l o R = 1.987 cal. mole- K. KINFIT program is used with NCST = 2 (number of constants) CONST (1) = R CONST (2) = T 130 ToKINFIY) l?3 .567 0 888 R88 0 ??? ??? 1 1‘1 III)". 1 NNN NNN N 000 000 0 FFE FF? F 9E F0 ) ( X 9 l \I 1| An 9( A ( MT )T l a T 1 as T ‘I. IL 0X 5 I 07 l N D H. (9 n # OP 0) C ) X9 F0 D 2 9U 2 ) ( KF o( 3 U G No )L ( ()) U0 0A T .)) OF 0V Q ))1 Nm 16 N ))3( 9D. (I 0 )IleX Q0 PE C )(TX AF ( 41$— Vv F0 /)(XN) Owl. \I )\lv|_\}) N1 90 215)C6 9X )2 ((N)/( TUT4( HX06)T) pcS(0 ) .XC()S) 0P XT G )./T1N3 NV )))<(0( 0T 1 F bfi‘HXCH RX 0) C (((OX§( CQ )0 N TTTC.)0 N 0.500 F. ngd)‘.) IKCO] n NNN)6() Y‘XNI.‘ N. 0002(T)) IvovL E CCC(TST) DDT42 P /((T§N¢T AY E )§ OgN‘UTé LTCLX o n P))N\(;C(T) OleT) ())“C(Ql‘1z T VQWI) E T13C(9)9( WSY 26 R S(((()S) ‘90..) H NXToo)u§o 1L.“ ‘)nu( T nuXQ.))C)cyfl ) F 0Y0?T A r _N))I\(Q,) PD:?(§D P ()HQhT((4 AAJ‘TNV F. Qrfi(((n\/(( TI ”COT P )(/TTh)/l J9 ECI M )T)§§n))( -n -V F 79)MMFU)§ FT -. T (r(qnaavn 09.0.)) T0TCF)T¢P AEL0)UI. p SCSfi /2(S¢ TROII). )U H(H))(o(C NI.T¢?(§ 6M 0v01)T).o novnhoYo u 9C)C(157))) .LrlA 1(0n4 E1 .5/)(T1\N#.l)). N nTIXut. 9 I. l)7(§X.US§ all EHI '. l.\ I}: \H I411!” NIAPI(Z(X NU QTNP) v T7 1(H)I (X IKTgfl 07FOIJu F.T§?C)(a((. 7 €XI.I(H66((A? U3