ELECTRON SMN RESONANCE 0F TRANSITION METAL COMPLEXES "fizesis {‘03 the Dogma of Ph. D. #eiCHEGAN $1‘AYE UNWERSE'FY Henry Anmn Kuska W625 This is to certify that the thesis entitled ELECTRON SPIN RESONANCE OF TRANSITION METAL COMPLEXES presented by HEN RY ANTON KUSKA has been accepted towards fulfillment of the requirements for M.— degree mm A “ 7, 4 7’ /4j”4}/I #07 6/14 Major talessor Aug. 19 1965 Date ’ 0-169 LIBRARY Michigan State University ABSTRACT ELECTRON SPIN RESONANCE OF TRANSITION METAL COMPLEXES by Henry Anton Kuska Central metal and ligand nuclear hyperfine splittings and g values were obtained for a number of transition metal complexes by electron paramagnetic resonance spectrosc0py. The hyperfine interactions studied include: (1) Carbon—13 ligand hyperfine splittings in Cr(CN)6-3, Cr(CN)5NO'3, Fe(CN)5NO-3, and VO(CN)5-3. (2) c0pper nuclear hyperfine splittings in a series of substituted acetylacetonate complexes, a series of amino acid complexes, and a series of aliphatic amine complexes, and (3) vanadium nuclear hyperfine splittings in a large number of vanadyl complexes. The metal-cyanide sigma bonds in Cr(CN)5NO-3, Cr(CN)g3 and VO(CN)5_3 were found to have a large amount of covalent character, but the amount of metal-cyanide covalent n bonding appears to be small. The data for the substituted acetylacetonate complexes of copper (II) indicate that the isotropic hyperfine splittings and g values give the Opposite trend of covalent character from that obtained by other methods. This discrepancy was explained by postulating that the isotrOpic hyperfine splittings are de- pendent on two contributions, a polarization of filled s orbi- tals and a fractional occupancy of the empty As orbital. The anisotrOpic hyperfine splittings and g values, which are not dependent on fractional As character, did give the expected trend in molecular orbital coefficients. The isotrOpic g values and hyperfine splittings were found to vary in the same manner Henry Anton Kuska when the complexes were dissolved in more basic solvents as they did when more electronegative groups were substituted in the acetylacetonate structure. The nitrogen ligand hyperfine splittings in copper (II) glycinate indicate appreciably less covalency in the c0pper—nitrogen bond then found for copper (II) phthalocyanine. For the series of copper (II) complexes with amino acids and aliphatic amines a larger isotropic copper nuclear hyperfine splitting correlated with a higher stability constant; however, there were many exceptions which appear to be due to steric effects changing the local symmetry at the copper ion. For the vanadyl complexes it was possible to correlate the isotropic hyperfine splittings with the ratio of axial to equatorial crystal field. The nuclear hyperfine splitting was found to decrease with increasing covalency and with a lowering of symmetry. The decrease with lowering of Symmetry is thought to be due to an increase in the positive contribution of Us electron density which decreases the net spin density since the polarization of filled s orbitals gives a negative Spin density. ELECTRON SPIN RESONANCE OF TRANSITION METAL COMPLEXES by Henry Anton Kuska A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1965 To my Wife and Parents ACKNOWLEDGMENTS The author wishes to eXpress his sincere appreciation to Professor M. T. Rogers for his continued interest, counsel, and encouragement during the course of this investigation. I would also like to thank Professor B. R. McGarvey, Polytechnic Institute of Brooklyn, and Professor R. G. Hayes, University of Notre Dame, for helpful discussions and pre- prints of related papers. The experimental assistance of Mr. Lowell Kispert, Mr. John Wreede, Mr. Robert Drullinger, Miss Sandra Bartlett, Mr. Robert Wegener, and Mr. Ward Collins is gratefully acknowledged. Also I would like to thank the Atomic Energy Commission, the Dow Chemical Company, and the United States Army for fi- nancial support during the course of this investigation. ii TABLE OF CONTENTS Page INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . l HISTORICAL . . . . . . . . . . . . . . . . . . . . . . . 3 Transition Metal Chemistry . . . . . . . . . . . . . 3 Review of the Application of Electron Spin Resonance to Transition Metal Complexes. . . . . . . . . . . A Development of the Theory of Obtaining Covalency Parameters from the ESR g Values . . . . Theory of Owen . . . . . . . Theory of Owen plus Screening Effects. Alternate Theory to Owen's Covalent Theory- The Free Ion Theory. . . . . . . . . . . . . . lO Present g Value Theory. Inclusion of Charge Transfer and Ligand Spin Orbit Coupling Con- stant Contributions. . . . . . . . . . . . . ll Development of the Theory of Obtaining Covalency Parameters from the ESR Hyperfine Splittings . . . ll Restricted Hartree-Fock Model-Higher Orbital 3 Character. . . . . . . . . . . . . . . . . . . ll Unrestricted Hartree-Fock Model-Spin Polarization ll THEORETICAL . . . . . . . . . . . . . . . . . . . . . . 16 Covalent Molecular Orbital Theory of Ligand Hyper- fine Splittings for Cr(CN)6'3 and CrF6-3 . . . . . 16 Energy Leve§s and Molecular Orbitals of Cr(CN)6 -3 and ch6- o o o o o o o o o o o o o 16 Experimental ESR Parameters . . . . . . . . . . 17 Relationships Between the Experimentally Deter— mined A'Values and the Molecular Orbitals. . . 19 Higher Order Considerations and Alternate Theory 21 Covalent Molecular Orbital Theory for the Determina- tion gf M. O. Coeffigients from the ESR g Values of CrF6' and Cr(CN)6' . . . . . . . . . . . . . 22 iii Molecular Orbital Theory of d1 and Low-Spin d 5 Systems in Tetragonal Field . . . . . . . . . . Theory of ESR g Values . . . . . . . . . . . Metal Ion Hyperfine Splitting Theory Ligand Hyperfine Splittings . . . . . . Molecular Orbital Theory of Ligand Hyperfine sp1§t- tings for Copper (II) Complexes and Fe(CN) SNO- Crystal Field Theory for Tetragonal Symmetry5 EXPERIMENTAL . . . . . . . . . . . . . . . . . . . . ESR System . . . . . . . . Second Derivative Presentation Measurement of Magnetic Field Measurement of Klystron Frequency . . . . . . . Linear, Precalibrated Magnetic Field Sweeps Single Crystal Holder Standard Samples . . . . . . . . . . . Preparation of Samples. . . . . . . . . . . . . . Crystal Growing . . . . . . . . . K Cr(CN)5NO in Alkali Halide Lattices 3 Preparation of Other Single Crystals . . . . Equations and Conversion Factors. . . . . . . . . . RESULTS 0 O O O O O O O O O O O O K3Cr(CN)5NO . . . . . . . . . . . . . . . . . Electronic Configuration and Optical Spectra. . Solution ESR Spectra. . . . . . . . . . . . Single Crystal Studies. . Comparisons of Powder Spectra Data with Single Crystal Data . . . . . . . . . . . . . . Infrared Spectra . . . . . . . . . . . . . . . K3cr(CN)6 o o o o o o o o o o o o o o o o o 0 Crystal Orientation . . . . . . . Vanadyl Complexes . . . . . . . . . . . . . . . . . Ligand Hyperfine Splittings in Vanadyl Com- plexes. O O O O I O O O O O O O O O Covalency in Vanadyl Complexes. . . . . . . iv Page 22 22 25 25 27 28 32 32 32 32 33 3A 3A 37 38 38 38 39 39 Al Al Al Al 42 52 55 55 55 62 6U 71 Copper Complexes . . . Single Crystal Study of COpper Glycinate in Cadmium Glycinate . . . . . . . . . . Solution Spectra of Copper Amino Acid Complexes . . . . . . . . . Solution Spectra of Substituted Copper Ace- tylacetonates . . . . . . . . . . . . . ESR Study ofl§r(NO)(NH3)§C13 DISCUSSION . . . . . . . . . . . . . . . . . K3Cr(CN)5NO . . . . . Solution ESR Spectra . . . . . . . . . Electronic Configuration and Optical Spectra Molecular Orbital Coefficients From Single Crystal Data . . . . . . . . . . . . . . . . Anisotropic l3C Splittings . . . . Variation of ESR and Infra-Red Data with the Various Alkali Halide Matrices . . . . . K3Cr(CN)6 Covalent Bonding Determination of gy, D, and E . . . . . 53Cr Hyperfine Splittings ESR Studies of Copper Glycinate . . . . . . . . . Solution Spectra of Substituted COpper Acetylace- tonates . . . . . . . . . . . . . . . . . . Solution Spectra of Substituted Vanadyl Acetylace- tonates . . . . . . . . . . . . . . . . . . Correlation of Optical Spectra With The ESR A values 0 O O O O O O O I O O O O 0 C ESR Spectra of Vanadyl Complexes . Solution Spectra of Copper (II) Amino Acid Com- plexes . . . . . . . . . . . . . . SUMMARY . . . . . . . . . . . . . . APPENDIX I-GLOSSARY o o o o o o 0 REFERENCES 0 O O O O O O O O O Page 73 73 78 78 88 90 9O 90 91 98 98 99 99 100 102 102 10A 110 112 117 118 126 127 130 Table 1. 9b. 10. 11. l2. 13. 1A. 15. 16. LIST OF TABLES Electron distribution in the metal d and ligand p orbitals due to metal-ligand overlap Reduction of spin—orbit coupling constants for transition metal hydrates Spin density in KNiF3 Ox and "x values for Co III ions . . Experimental ESR data for K3 Cr(CN)5NO in H O 2 0 Data for Cr(CN)5NO-3 in various lattices . . Comparison of powder and crystal ESR data , Infrared data for [M(CN)5NO]X compounds Direction cosines between the principle axes (x,y,z) and the crystallographic (a,b,c) for K3Co(CN)6 . . . Direction cosines between the molecular (a, B,t) and the crystallographic axes for }{3CO(CN)6 o o o o o o o 130 ESR data for Cr(CN)6-3 in K 3CO(CN)6 magnetic axes axes (a,b,c) Solution ESR spectra of substituted vanadyl acetylacetonates . . . . . . . . . ESR and Optical data for vanadyl oxalate ESR data for vanadyl complexes . 130 hyperfine splittings in VO(CN)5-3 Molecular orbital coefficient for vanadyl xy orbital O O O O I O O O O 0 I O O ESR data for bis-(glycino)-c0pper (II)-xH O vi 2 O t Page 15 31 A3 53 5A 56 59 59 62 63 6A 65 71 73 7A Table Page 17. ESR hyperfine splittings for copper (II) amino acid complexes . . . . . . . . . . . . . . . 78 18. Isotropic and anisotrOpic A values for substi- tuted copper (II) acetylacetonates . . . . . . 79 19. s orbital Spin density (fS ) in sigma ligand or- bitals o o o o o o o o o o o 92 20a. Possible energy level assignment for Cr(CN)5NO-3 93 20b. Alternate energy level assignment for Cr(CN)5NO_3 9A 21. K3CO(CN)6 vibration frequencies . . . . . . . . 99 22. Values of g , D, and E for K30r(CN)6 in KCO(CN)y..................101 3 6 23. Comparison of ligand spin densities for copper complexes . . . . . . . . . . . . . . . . . . 103 24. ESR hyperfine splitting constants and other data for c0pper acetylacetonates . . . . . . . . . 105 25. Comparison of copper (II) phthalocyanine and c0pper (II) glycinate ESR data . . . . . . . . 109 26. ESR A values for Cu+2 and VO+2 adsorbed on ion exchange resins . . . . . . . . . . . . . . . 111 27. Energy levels, hyperfine splittings, and p values for vanadyl complexes . . . . . . . . 113 28. o and n electrostatic interactions for VYX5 compounds . . . . . . . . . . . . . . . . . . 116 29. Comparison of the ESR data for VO(CN)5 and [VS6C6(CN)6 J o o o o c o o o o o o o o o o o o 118 +2 30. Literature ESR A_va1ues for V0 complexes . . . 119 31. Group overlap integrals for some vanadyl com- plexes O 0 O O O I O O O O I O O 0 0 C O O o 0 121 32. ESR data and literature data for cOpper amino acid complexes . . . . . . . . . . . . . . . . 122 33. ESR and stability constants for copper (II) amine complexes . . . . . . . . . . . . . . . . . 125 vii 10. ll. 12. 13. 1n. 15. LIST OF FIGURES Page Energy level diagram for Cr+3 ions , , , , 18 Energy level diagram for a d1 transition metal ion in an octahedral and in a tetragonal field . . . . . . . . , . . 2A Crystal field splittings for v+u . , , , 29 Hall probe schematic . . . . . , . , , , 35 Adjustable single crystal holder , , , , 36 Energy level diagram for K3Cr(CN)5N0 , , AA Orientation dependence of Cr(CN) 5N0"3 ESR data U6 Saturation of the g position of K3Cr(CN)5 NO in KBr single crystal at 770K . U7 Saturation of g, spectrum but not the g, spec— trum for K3Cr(CN)5 NO in KBr single crystal at 770K 0 O O O O O O O O I O O O O O O O O “7 Collapse of nitrogen lines by overmodulation of the 100 Kc unit for g" position of 15.5% 13C enriched K3Cr(CN)5 NO in KCl . . . , , , A9 Superposition of the g ESR spectrum of 15.5% enriched 13 3c K 3Cr(CN‘§ N0 in KBr single crystal and in 3KCl siRgle crystal . . . . . 50 130 splittings of 15.5% enriched l3c K3Cr(CN)5NO in 8 plane 0 O O O O O O O O O O O O O O O 51 .1 Infrared spectra, from 5,000 cm.1 to 1,100 cm- of K3Cr(CN)5N0 in various hosts . . . 57 Infrared spectra, from 5,000 cm-1 to 1,100 cm- of Fe(CN) 5N0‘ in various hosts . , , , , 58 ESR spectrum of 15. 5% 130 enriched Cr(CN)63 K 3CO(CN)6 o o o o o o o o o o o o o 0 viii 60 Figure 16. 17. 18. 19. 20. 21. 22. 23. 2A. 25. 26. 27. 28. 29. 30. 31. Single crystal ESR spectrum of VO(CN)5-3 in KBr O O O O 0 O O O O O O O O ESR spectrum Of VO(NCS)5-3 in CHCl Aqueous solution Spectrum of 15.5% VO(CN)5"3......... solution 3 l3 ESR Spectra of copper (II) glycinate . . C enriched Second derivative ESR spectrum at g position (lines two and three) of a copper (II) glycinate-cadmium (II) glycinate single crystal . . . . . . . . . Second derivative ESR spectrum at a copper (II) glycinate-cadmium single crystal . . . . . . . . Frozen CHCl solution ESR spectra hexafluor8acetylacetonate . . Frozen CHCl solution ESR spectra trifluorogcetylacetonate . . . Frozen CHCl solution ESR spectra benzoylacétonate. Frozen CHCl solution ESR spectra 3—methy1a8ety1acetonate . . . . Frozen CHCl solution ESR spectra acetylaceéonate . . . . . . . . Frozen CHCl solution ESR spectra benzoyltrIfluoroacetone . . . Frozen CHCl solution ESR spectra g, position of (II) glycinate of c0pper of copper of copper of copper of copper of copper of copper (II) (II) (II) (II) (II) (II) dibenzoylémethane . . . . . . . . . . Hyperfine structure observed between g" lines two and three in a frozen CHCl ESR spectrum of copper (II) dibenzoylmethané . . . . . . . ESR spectra of Cr(NO)(NH3)5+3 in NHuCl single crystal . C . . . C O C . . . C O C C C . Solvent and substituent effects on copper (II) acetylacetonates . . . . . . . . . . . . . ix (II) Page 0\ \O 80 81 82 83 8A 85 86 87 89 108 Figure Page 32. Correlation of ratio of axial to equatorial crystal field,(2, and ESR A values for a number of vanadyl complexes . . . . . . . . . 115 33. Plot of A" versus g” for a number of vanadyl complexes 0 O O O O O O O O O O 0 O O O O O O 120 INTRODUCTION During the past ten years our knowledge of transition metal complexes has grown considerably. Experimentally, vi- sible, infrared, and electron spin resonance (ESR) Spectro— scopy have provided the data required to test old theories and formulate new ones. Theoretically, a combination of crystal field and molecular orbital theory have been deveIOped to explain the experimental data. This new theory called ligand field theory, is now generally accepted and is at pre— sent being refined tO the point that quantitative calculations of the various observables are being made with reasonable success.1 In this thesis the ESR spectra of a number of transition metal systems have been studied, the data have been interpreted in terms of present theory, and correlations of the ESR data with other types of measurements have been made. At the beginning of this investigation there was a sizeable literature of ESR work on transition metal fluoro- complexes which indicated that fluorine-metal bonds were not ionic as had been assumed, but had a definite fraction of covalent character.l’2 It was decided that a comparison of ESR data for compounds normally considered to be covalent, such as the cyanides, with data for the fluoro-complexes would provide a basis for interpreting the significance of the re- ported covalent character in the fluoro-complexes. A relative comparison was considered desirable since there were indica— tions that the approximations involved in the ESR theory were such that the amount of covalent character was being over- estimated.3 However, in making such a comparison an additional factor had to be considered. The theory of hyperfine inter-- actions had been developed for the nearly ionic case; in the 1 2 extension of this theory to highly covalent cases, additional terms may be significant. Because of this uncertainty, it was desirable to have ESR data from an addition series of compounds which had been well characterized by other physical methods and were known to be intermediate in covalency be- tween the fluorides and the cyanides. For the extremely covalent series, the cyanides and nitrosyls were chosen. The series consisted of the t2 1 compound V0(CN)5-3, the t2g3 compounds Cr(CN)6_3 and Cr(N0) (NH3)5‘3, the t285 compounds fliréCN)5N0_3 and Fe(CN)6-3 and the t2g eg compound Fe(CN)5NO' . ESR spectra of several series were studied for the intermediate covalency situation. Substituted acetylacetonate +2 and V0"-2 On ESR parameters of having the electron in an orbital of 0 complexes of Cu were chosen to Observe the effect +2 +2 (for Cu ) or n (for V0 ) symmetry. The acetylacetonates were chosen because there exists a large amount of other types of physical data to correlate with the ESR data. Several 2 were investigated, and the ESR amino-acid complexes of Cu+ data were compared with other physical data from the liter— ature. Several vanadyl complexes with bonding intermediate in character between ionic and covalent were investigated, and the ESR data were compared with recently published infra- red and Optical data. For ionic complexes published data were utilized for all of the fluoro-complexes with the exception of VOF5'3 for which measurements are reported here. HISTORICAL Transition Metal Chemistry The general field is covered by several excellent books and many recent review articles. Ballhausen's book“ gives an excellent introduction and review of ligand field theory in general with an excellent balance of theory and chemical applications. A pair of complementary books, one by Orgel5 and the other by Griffith6 provide an exhaustive pre-l96l coverage of ligand field theory. Orgel's book is a concise review of the qualitative correlations of the theory with experimental data,while Griffith's book is a complete presentation of the mathematical and physical development of ligand field theory. Jorgensen has two recent books, one discusses the pre—l960 field of Optical spectroscopy from a ligand field view-point,7 while the second is a pre—lgou survey of the general scientific literature of transition metal complexes.8 Jorgensen also has three comprehensive review articles. The two most interesting from the standpoint of the work of this thesis are a9pre-1963 review and deveIOp- ment of the nephelauxetic series and a review and further deveIOpment of the application of ligand field theory to Opti- cal spectrosc0py.10 The third is a more general review of 11 The applications of group theory to ligand field theory are well reviewed and explained by Cotton.12 The theory and pre-l963 the application of spectrOSCOpy to chemical bonding. applications of infrared spectroscopy to transition metal chemistry are comprehensively reviewed by Nakamoto.l3 ‘Optical and infrared spectrOSCOpy and ligand field theory have received comprehensive and authoritative reviews so they will not be generally reviewed in this thesis and 3 only the basic ideas needed to correlate with the present work will be presented. Since the application of electron spin resonance to transition metal complexes does not have the benefit of a recent comprehensive review article, the following section will be primarily devoted to this area. Review of theprplication of Electron Spin Resonance to Transition Metal Complexes A review of the early development of electron Spin resonance, ESR, has been given by previous theseslu"l7 from this department and will not be duplicated. In addition to the early reviews and books, see Faber,1u a number of recent reviews and books have appeared which ade— quately cover the basic fundamentals and the pre-l960 liter- 18 is a textbook covering the theory of ature. Pake's book ESR from the vieWpoint of a physicist. It does not discuss eXperimental details or, in general, chemical applications. It is especially useful as an abstract of, and a reference source to, the early theoretical papers. A slightly Older but extremely useful book is that of Low.19 This book is more restricted than that of Fake in that it only covers the ESR of solids, but it is extremely thorough in its coverage of this area and therefore is also useful as a source-book. A recent book by Slichter2O is useful as a textbook presenting the theoretical develOpment of ESR of the solid state. The book that is the nearest comparison to the comprehensive Optical and infrared books discussed earlier is that by Al'tshuler and Kozyrev.21 It is a translation of a 1961 book; and, therefore, its usefulness is limited to the literature and theory of that period. To partially fill the gap between the limitations of the above books and the general literature there are several excel- lent reviews. Anderson22 has written a very good general review of the experimental develOpments with well chosen exam— 23 ples of experimental applications. Stevens has reviewed the use of the spin Hamiltonian (the mathematical shorthand used to describe the ESR experimental observables). This review provides detailed mathematical examples of the application of the Spin Hamiltonian. A useful introductory article covering the application of ESR theory to transition metal ggmpounds is that of Carrington and Longuet-Higgins.27 Jarrett has writ- ten a detailed review of the theoretical treatment of ESR which serves as an excellent recent review and textbook of advanced ESR. Robertson26 has written an excellent review article in which he discusses in detail representative exam- ples of the application of ESR to transition metal complexes with organic ligands. There are also several concise reviews‘27’28’29 in the annual review literature which, in general, abstract the important ESR literature of that year with a limited amount of critical comment. As a concentrated source of recent eXperimental papers, there are several published proceedings of conferences which dealt with ESR3O‘36. A source of comprehensive reviews of particular aspects of ESR is the recent Ph.D. thesis litera— ture. Of particular interest are ESR studies of those transi— tion metal ions which give narrow lines at room temperature or at liquid nitrogen temperature, 770K. These ions normally are of the outer electronic configuration 3dl, 3d5, 3d9, OP 5fl. In general the theory of each of these cases is dif- ferent. Fortunately there is available a Ph.D. thesis which covers each of these configurations. Faber's thesislu con- cerns the random orientation spectra of dl, d5, and d9 ions. Feltham's highly diverse investigation37 Spectra, optical spectra, and ligand field theory of d1 and -'d9 d5 ions in single crystals is the thesis of Drumheller. Neiman's thesis39 is a very useful thesis covering the theory of powder Spectra and molecular orbital analysis of d9 Cu com— plexes. The theory of ligand hyperfine splitting in d3 covered solution ESR 'complexes. A good source of the present state of ES§80f fluoride complexes is presented in Guzzo's thesis“0 and sub- ui,u2 sequent publications, but in Guzzo's publications the analysis of the data has been changed from that given in the thesis. Two theses from the University of Californiau3’uu report an investigation of the optical and ESR prOperties of protactinium in single crystals. Hayes' thesisl45 gives a complete review of the theory of line widths of paramagnetic ions in solution. Development of the Theory of Obtaining Covalency Parameters from the ESR g Values. Theory of Owen Owenu6 found by utilizing optical and magnetic data that the spin-orbit coupling constant A in the equation g = 2.0023 — %} (1) was smaller by 20 to 30% than the free ion 1 value. In equa— tion 1 A is the energy separation between the ground state and the first excited state and g is the electron gyromagnetic ratio. He interpreted this reduction by saying that partial covalent bonding forces some of the unpaired electron out on the ligands. Theory of Owen_plus_Screening Effects Murao)47 noted that the decrease in A is greater for metal ions of smaller atomic number and larger valency. He attributed this to screening by the additional 3d electron density produced by the admixture of 3d wave function into the bonding orbitals. The bonding orbitals are of the form ¢b = Mo pcmwfla lampoe on one mamuwnmo a pcmwfla one U Hopoe or» CH :ofipsnwmpmfip coupooamll.a mqm m+so m+Hz m+oo m+om m+cz m+ho m+> m+fle .moumspms HeuoE.COHpHmcmmp pom mpcmpmcoo wcwadsoo OHOLOICHQm mo COHpozcmmtl.N mqm<8 10 An alternate procedure for considering screening effects appears to be currently more pOpular. This method consists of using the tabulated spin-orbit coupling constants of Dunn“8 for the various possible oxidation states of the metal to calculate the covalency parameters for that parti- cular oxidation state and then choosing the set of covalency parameters which gives a spin density on the metal in closest agreement with the corresponding assumed oxidation state. Alternate Theory to Owen's Covalent Theory - The Free Ion Theory Although Murao was able to obtain good agreement be- tween the theoretical and experimental reductions in the spin— orbit coupling constants, Marshall and Stuart“9 felt that the amount of covalency required was too large. They pro- posed a theory based on an ionic model. In an ionic model there is no covalent character to the wave functions and c in equation 3 is Just equal to the metal-ligand overlap inte- gral. They postulated that the decrease in A is due to an over- all shift of the d wave function outward due to a repulsion between the d electron density and the 2p electron density which overlaps the metal ion. They supported this theory by citing a neutron diffraction study which reported that the 3d wave function was expanded by 10% over the free ion wave func- tion. Using 10% eXpanded wave functions they were able to obtain good agreement between the experimental fluorine hyper- fine splittings for MnF and their calculated values. 2 Marshall and Stuart's theory is known as the free ion theory since covalent bonding is not considered in their model. 50 Their theory has been reviewed by Anderson and by Shulman and Sugano.51 Shulman and Sugano point out that the agreement of Marshall and Stuart's theory with the experimental MnF6_u data was fortuitous and could not be extended to other ions. An- derson discusses the theoretical Justifications given by Mar- shall and Stuart in prOposing their model and concludes that their interpretation is not the only possible one. 11 Present g Value Theory. Inclusion of ChargeITransfer and Ligand Spin Orbit Coupling Constant Contributions Several investigatorsSZ-Su have recently corrected the theory of Owen to include charge transfer and ligand spin— orbit coupling constant contributions. A published comparison of the estimates of covalency obtained with these contribu- tions and the estimates of covalency obtained from Optical data indicates55 that the former g value theory without charge transfer terms overestimated the degree of covalency. Development of the Theory of Obtaining Covalency Parameters from the ESR Hyperfine Splittings Restricted Hartree-Fock Model- Higher Orbital 8 Character In 332 3dn transition metal ions one would not predict an isotrOpic nuclear hyperfine splitting since the 3d orbitals have a node at the metal ion nucleus. The earliest eXplana- tion of the observed Splitting is that in the ground state there is a small admixture of 3sl3dnllsl character. Since an s electron has a high electron density at the nucleus, only a small percentage of 3sl3dnl—lsl character is required to account for the observed Splittings. This theory was introduced by 56 67 Abragam and Pryce and utilized by Van Wieringenl to explain the observed variations of the Mn+2 nuclear hyperfine split- tings in a series of Mn+2 complexes. Unrestricted Hartree-Fock Model- Spin Polarization For Metal Ion Nuclear Splittings Although the theory of Abragam and Pryce was able to quantitatively account for the magnitude of the splitting, subsequent investigations showed that it predicted the wrong 57, Sign for the splittings. 58’59 This discrepancy was accounted 12 for by theoretical unrestricted Hartree-Fock calculations of the exchange polarization mechanism which showed that an un- paired 3d electron will polarize the core 8 electrons to give a net Spin at the nucleus of opposite Sign to the unpaired electron. For example, Freeman and Watson57 reported for vanadium metal that the contributions of the various orbitals were +0.05 (ls), —u.85(2s), +1.6.(3s) and +2.7“(Us) to give and overall effective field at the nucleus of -0.A5 atomic units. 61 +1 for Fe+3 and Cr , were able to interpret the A values using only spin polariza— Matamura6O for Mn+2, and Title tion and covalency. They plotted the experimental A values versus estimated percentage ionicity62 and extrapolated back to 100% ionicity. The A values at 100% were in close agree- ment with the calculated values of Freeman and Watson.57 +1 +3 The agreement for Mn+2 is impressive, but for Cr additional data are needed.62 and Fe Konig63 has interpreted the isotrOpic A values of Cr+1 the spin polarization theory and the higher s-state contribu— and V(0), both d5 low spin cases using a combination of tion theory. He calculated the exchange polarization con- tributions to be +26.A and —13A gauss for Cr+1 and V(O) respectively. Since the experimental values of A were +21.8 and -83.5 gauss, he calculated the As contribution to be -“.6 gauss and +50.5 gauss by the following relationship: A(experimental)=(A exchange polarization)+ A(Us character.) He calculated that One electron in a As orbital would cause a splitting of -u22.5 and 923.1 gauss for Cr+1 and V(0), respectively. He was then able to calculate the As spin density as 0.0109 and 0.05U7, respectively. Because of the uncertainties in the calculation of exchange polarization and Us character and the lack of consideration of the effects of covalent bonding on the calculated parameters, the As spin densities cannot be considered as quantitatively significant. Additional experimental data are needed. For example chromium metal is 3d5usl; since there is an unpaired electron in the 13 As orbital, a large As contribution is expected. Childs et al.67 report that the experimental value for chromium metal is 29.A7 gauss. This value is only slightly larger than the Cr Splittings when there are only 3d unpaired electrons. This result casts considerable doubt on Konig's estimation of -A22.5 gauss for a As Cr+1 electron. Childs et a1. calculate ' -51 gauss for the As splitting. As a further example of the qualitative nature of this type of treatment, Davison, et a1. estimate the As splitting by Cr to be -ll70 gauss. Another illustration of the use of-both the As contri— bution and the spin polarization is the published interpreta- A 65 cm-l) of the vana— / tion of the unusually low value (25.7 x 10- f dium hyperfine splitting found in bis—cyclopentadienyl vanadiumTU Since proton NMR spectra of the complex rule out the possi- bility that the electron is predominately localized on the ligand, the small vanadium splitting cannot be attributed to a large covalency. The small splitting was-explained by par- tial cancellation of the negative-spin polarization contribu- tion by a positive As contribution. Evidence which appears to support the combined polar— ization and As character theory is the large difference in 2 in two non- hyperfine splitting constants reported for Cu+ equivalent sites in an NHuCI single crystal.67 Since the g values for both sites are near 2, the difference cannot be due tO‘a large orbital contribution. Also, the observation of an additional 13-line superhyperfine structure due to A equiva- lent Cl- ions rules out a large difference in covalency be- tween the two sites. The hyperfine splittings for one of the sites have a pronounced temperature dependence. All of these observations can be explained by assuming that the amount of As cfldaracter is dependent on the local crystal field symmetry. FRJP Ligand Hyperfine Splittings Two early theories were proposed to account for the <fl3£3erved ligand hyperfine splittings. Mukherji and Das,68 M<‘ii-II."shall and Stewart,9 and Freeman and Watson69 utilized ionic 1A models and were able to quantitatively account for the mag- nitudes of the splittings by polarization of the ligand or- bitals due to overlap of the metal-ligand wave functions. 51 However, subsequent experiments which determined the Sign of the hyperfine Splittings and the dependence of the split- tings on whether the unpaired electron was in an orbital of 0 or W symmetry indicated that covalent contributions must be considered. The covalent contribution at first was considered to arise from covalency of the antibonging unpaired electron. Using this model Shulman and Sugano 'were able to obtain good agreement between the Observed and theoretical fluorine hyperfine splittings in KNiF3. However, Watson and Freeman70 have recently claimed that the model used by Shulman and Sugano is incorrect. They propose that the covalent contri- bution arises from spin unpairing in the bonding molecular orbitals rather than the antibonding molecular orbitals. The numerical parameters calculated by them did not give as good agreement with the experimental parameters as the theo- retical values calculated by Shulman and Sugano; however, a more recent paper by Ellis71 has reconsidered Watson and Freeman's theory and found additional contributions so that the theoretical and experimental values are now in better agreement; as illustrated in Table 3. An additional mechanism which can account for observed ligang hyperfine splittings has been prOposed by Kivelson and Lee. This mechanism is a configuration interaction whereby a bonding electron is promoted from its spin paired orbital to the corresponding antibonding orbital leaving a net negative Spin density in the original bonding orbital. This mechanism predicts a ligand hyperfine splitting even when the unpaired electron is in an ionic orbital which does not overlap the ligand in question. l5 w>m0.0 5000.0 00H0.0 Hmm0.0 0% HmpOB .IIIIII .IIIIII 111111. a an maao-o oooo.o om:o.o oeoHs>oo sadm L b .. fisoo.o smoo.o Haoo.o doaso>o each a m amoo.o mooo.o Hmoo.o maoo.o L Heooe 11:11:. .1111]: .IIIII. m a: Haoo.o oaoo.o omoo.o ocoan>oo cadm o m in. mmoo.o Hmoo.o mmoo.o dofito>o each a mpcmeflnooxm mmHHHm ,owcmEoonm Hmocmwsm Hm zpflmcom sfiom H H pco comumz paw Chafisnm .m maze ea moaneoo caamnu.m mamae THEORETICAL COVALENT MOLECULAR ORBITAL THEORY OF LIGAND HYPERFINE SPLITTINGS FOR Cr(CN)6-3 AND CrF6-3 For a more general discussion of this theory the 50 review article and Shul- reader is referred to Anderson's man and Sugano's51 paper.- The Cr(CN)6'3 and CrF6-3 ions will be used to illustrate the application of the theory to complexes in which only molecular orbitals (M.O.'S) of n symmetry contain unpaired metal electrons in the ground state. Energy Levels and Molecular Orbitals of Cr(CN)6‘3 and Cr F6-3 The energy levels of Cr(CN)6-3 and CrF6—3 are given in Figure 1. The cubic field splits the five degenerate chro— mium 3M1 orbitals into~a higher energy eg doublet and a lower energy t2g.triplet. The separatio§3between the eg and t2g levels is called A. For the CrF6 complex the chromium t2g levels form molecular orbitals with the occupied. 2p1T orbitals of fluorine. The bonding t2g by the paired fluorine electrons forcing the unpaired metal t28 thereby decreasing'A.‘ See Figure 1. For Cr(CN)6-3 the low molecular orbitals are occupied * electrons into the antibonding t2g molecular orbitals energy p1T carbon orbitals are utilized in carbOn to nitrogen n bonding leaving only an empty high energy antibonding carbon orbital to form n molecular orbitals with the metal d orbitals. Since the carbon pTr orbital is empty and less stable than the metal d orbitals, the metal electrons are stabilized by the interaction and form a t molecular orbital which is slightly 2s 16 l7 3 bonding. This causes an increase in. A (A Cr(CN)6- > A in cubic field > A in CrF6-3). Therefore, for CrF6-3 the n molecular orbitals con— taining the three unpaired electrons are antibonding, while for Cr(CN)6"3 the n molecular orbitals are bonding. The molecular orbitals are: * ¢ = N (v - A w - A w ) o o 3dZ2 2s 28 p0 pO (8) ¢ = N (w i A w ) (9) T'1 Trl 3dzx Tr1 "1 ¢ = N (w i A w ) (10) Tr2 1T2 3dyz 1T2 1T2 The negative signs in equations (9) and (10) are for CrF6-3; the positive signs are for Cr(CN)6-3. For Cr(CN)6'3 there is an additional negative (antibonding) term in equations (9) and (10) due to a contribution from the bonding occupied carbon p1T orbitals.73 there would not be sufficient experimental data to solve for the M.O. coefficients. Since the stability of transition metal If this term was kept cyanides increases as the number of d1T electrons increases the important term must be the positive (bonding) term and as an approximation the antibonding term was set equal to zero and not included in equations (9) and (10). Experimental ESR Parameters If the ESR absorption lines are narrow and if the ligand nuclei have nuclear magnetic moments, it is sometimes possible to observe a splitting of the main lines. This split— ting is due to unpaired electron density reaching the ligand nucleus. The spectra can be fitted to the equation: 2 2 2- . 2 9 H—H = (An cos 9 + A,- Sln 6)‘Tfi; (ll) 0 .maow m+no new ampmmHe Ho>od mwnonm .H unamwm .waflenon : vcmmwa ou Hence we ebflumuaouonmon ow dam m mAzovao now mcmupwamm any aw <+< I . .wawvaon : on nu“: «+90 now wawuuwamm may aw < .mdavaon n dupes Op enmmwa mo u>mu~ucomonmea mm enumnmmno pom mnwpuwamm on“ ma (15) ° 5(2311 ° " r3 n 2 11n 8n 1 A = (f — f ) —— (16) 7‘ 5 (23) In Trl Tr2 A1T is zero for cylindrical symmetry around the metal- ligand bond since f1T = f" . Cylindrical symmetry is found in CrX6-3 ligand-metél bongs because each of the three t2g orbitals contains one unpaired electron. AD = —— <17) 20 C is a correction factor due to the fact that the magnetic electrons cannot be treated as point dipoles.72 2 2 8nun 8n NO A28 A = t<0)2 <18) 3 23 In 3 8n is the nuclear magneton, pm is the magnetic moment of the ligand nucleus, 8 is the total electron spin of the complex and AD is the point dipole term. The molecular orbital coefficients are related to 3' where U is a measure of the covalency in the bond formed by the central metal ion 3d orbital and the x orbital of the ligand by the followingg' A2s = Us + J§.S3d,2s (19) Ap = K“ + 583d’po (20) A“ = A; + 2 S3d’pn (21) where S is the metal 3d orbital—ligand x orbital atomic d,x overlap integral. N0 and NTr are the normalization coefficients and are given by 2 ll 2 -% (l-2J3Aosd,pO + AU) (22) 2 ll 2 -% fl (1 :A Aflsd’pTr + A“) (23) where the plus sign in equation (23) is for bonding mole- 3 9 cular orbitals, as in Cr(CN)6' and the negative Sign is for antibonding molecular orbitals, as in CrF6-3' 21 The unpaired spin density in the ligand 0 and n orbitals, f0 and f“, respectively, are related to the mol- ecular orbital coefficients and to the-experimental hyperfine Splittings by the fOllOWing'equations:51 (2 N2 2s A O O S f = = -—r— (2“) U 3 A s 2 2 Afl~N1T 28 AG f,” = = ' (25) A A 0 28 As f8 = (26) ' A s ' 8 un 8n 2 As = __ [¢(O)2S] (in gauss) (27) 3 I n 2 ' un 8n 1 O 5 In r3 2p HigherAOrder Considerations and Alternate Theory The above theory is only a first order theory; there are-several higher order effects which have been examined in the literature. The ESR theory for COF3-l to include the effects of unquenched orbital momentum and 7A has been extended mixing of excited states. Marshall75 has investigated higher order contribu- tions for Cr+3 and Ni+2 and finds that terms from spin-orbit interactions which produce unquenched orbital moment on the ligand and also modify the spin distribution are sometimes important. 22 Helmholz, Guzzo, and Sandersl42 give equations which take into consideration that the magnetic field at the ligand has components perpendicular to the external field of the unpaired electrons. 76 70 Simanek and Sroubek and also Watson and Freeman have considered an alternate theory to the theory used in this thesis. They consider that the unpaired spin density reaches the ligands through the bonding electrons and is of opposite Sign to the unpaired metal electrons. This theory has been developed, at present, only for the d8 case so that +3 a comparison of its use for Cr with the theory used in this thesis is not yet possible. COVALENT MOLECULAR ORBITAL THEORY FOR THE DETERMINATION OF M.O. COEFFICIENTS FROM THE ESR g VALUES OF CrF6‘3 AND Cr(CN)6-3 The theory for the determination of the M.O. coef- ficients from the g values has been presented by Lacroix and 53 Emch52 and by McGarvey. It will not be presented here since it was not used in this research. MOLECULA ORBITAL THEORY FOR (11 AND LOW- SPIN Ci SYSTEMS IN A TETRAGONAL FIELD Theorycfl‘ESR g Values In a tetragonal field for the d1 and low-Spin d5 cases the experimental g values can be-related to the mole- cular orbital coefficients by:72’77 8AN§2N§ S = 2.0023 — ——————__ [A _ %(y" A ) T(n) " Ab2+bl(I> 2 02 8A Ni (1 - N: ) (29) 2- 2 —A 31 s — 2A s — O2 C1"’2 T'2 d’"2} Ab2 + bl (II) 23 2A N2 N2 1T2 Til ;’ e e E = 2.0023 - [l - (%)2A A - 28 A i Ab2-+e (I) Tr1 1T2 d’"2 7T2 2 _ 2A N (1 - N2 ) (30) g e a 1T2 T'1 — (5)228d " A1T - Sd A ] - ’ 2 l ’"1 1'1 A02 + e (II) where T(n) is defined by Kivelson and Lee,72 Sd is the ’TT metal d, ligand pTr atomic overlap integral,)$ 3ndJfiALare the xz,yz molecular orbital coefficients for the equational and axial ligands, respectively, and A is the spin-orbit coupling constant. The spin-orbit coupling constant is posi- tive if the transition is one where an unpaired electron goes from a half-filled orbital to an unoccupied orbital and is negative if the transition is from a filled orbital to a half-filled orbital. The latter type of transition is com- monly called a hole transition. Transitions I and II are the d-d and the charge transfer transitions, respectively. Ab2 + b1 and Ab2 + e are determined from optical spectroscopy with the aid of the energy level diagram given in Figure 2. The g values are Obtained from the experimental spec- trum with the use of the following equations:78’38 by For g isotropic g = 8H (31) 0 A2 2 A3 HO = H( ) + A-m + [I(I + 1) — m 1 + (32) m 2 H A H 2 (m) (m) A12 [I(I + 1) - m2] For g" HO = H(m) + Au.m + 2 H (33) (m) (A 2 + A %)[I (I + 1) - m2] ll 1 + A 'm + (3A) For g1 H = H(m) 1 A H (m) 2A d (all (bl) (e) octahedral (b2) field tetragonal field 1 for d d (a1) d 2_ 2 (bl) d 2_ 2 (bl) d 2 (31) or d (b2) d (b2) dxz’ d z (e) dxz’ d z (e) 5 Figure 2. Energy level diagram for a d1 transition metal ion in an octahedral and in a tetragonal field. 25 Here H(m) is the magnetic field position of the ESR line due to the component (m) of the nuclear magnetic moment (I), v is the klystron frequency in megacycles and A, A", and AL are the nuclear hyperfine splitting constants. Metal Ion Hyperfine Splitting Theory The experimental metal ion nuclear hyperfine split- ting is related to the molecular orbitals by the equations:72 A 8). N2 N2 AO g AN2 N2 A0 2 1'2 02 7 1'2 Trl u: -—N1rA0-A- - (35) 7 2 Ab2 + bl (I) ab2 + e (I) 2 11 AN2 N2 A . 2 1T2 171 0 AL = —-NH AO - A - —— (36) 7 2 7 A b2+ e (I) 2 l where A = —-A‘ + “Au and A0 = 3 Estimation of 3 r <: 3) values. value in ln(< —) term. of a known similar ion. Take the calculated (r 3)putZ it back in the right side of Equation (37) and repeat the cal- culation; continue until ’ ”M (x2 - 53)“) (x2 - 372)”) (mm) f (22) Rb) (nym y (22)“) (22mm (mun (£45) 27 The superscript b indicates a bonding orbital and * indicates an antibonding orbital. For the equatorial ligands the first promotion is the important one. The isotropic split- ting is given by: 2 2 ' N (l - N ) (d ,d ) A O2 O2 xy X2_ y2 s AS = - (A6) * 2A E (x2 - Y2)b + (X2 - Y2) where (dxy’d 2 ) is an exchange integral and the % comes X _ from the assumption that the sigma ligand orbital is an s-p hybrid. A similar equation applies for the isotrOpic nitrogen Splittings except that the promotions involve the 22 orbitals. The approximations involved in equation (“6) are such that molecular orbital coefficients cannot be calculated from them; however,the spin density at the ligand can be calculated with the use of equation (26L An alternate mechanism also considered by Kivelson 72 and Lee is that the AS values are due to polarization of the ligand by unpaired spin density in the ligand n orbitals. MOLECULAR ORBITAL THEORY OF LIGAND HYPERFINE SPLITTINGS FOR COPPER (II) COMPLEXES AND Fe(CN)5NO-3 For the equatorial ligands in copper(II)complexes and the axial ligands in Fe(CN)5NO-3 the unpaired electron is in an orbital of 0 symmetry. Equations (11) through (28) are applicable with the assumption that 2U A0x__5(23)1nf°:_<3> in equation (15) and that AnxiO in equation 16. 28 CRYSTAL FIELD THEORY FOR TETRAGONAL SYMMETRY A crystal field of tetragonal symmetry is obtained by compression or elongation of an octahedron along a four- fold axis. An octahedral field causes the free ion degenerate d orbital energy levels to be split into a triplet and a doublet as shown in Figure 3. The addition of a tetragonal . distortion splits the doublet into two singlets and the triplet into a doublet and a singlet. The energies of the orbitals can be represented in terms of three crystal field parameters Ds, Dq and Dt by the following equations:80 E(z2) = 6 Dq - 208 - 6 Dt (A7) E(x2 - y2) = 6 Dq + 2Ds - Dt (u8) E(xy) = - A Dq + 2Ds - Dt (A9) E(xz,yz) = - u Dq - D8 + u Dt (50) where Dq, D3, and Dt;are defined as: l Dq = _ puxy (51) 6 1 DS = ; (202xy ‘ p2+2 ‘ 02-2) (52) l at = -— (2%xy - ou+z - pu'z) (53) 21 n P o = eq( ) (5U) Rn+l 29 2 2 2 z, x -y ‘— 2 / gaseous +1} V xz, yz XZ. ya. xy /\ octahedral tetragonal field compression Figure 3. Crystal field splittings for V+u. 30 e and r are the electronic charge and radius, and q and R are the effective ligand charge and the internuclear distance. In this thesis we are interested in the ratio of the axial field to the equatorial field, 0:81 +2 -Z on + “A p = (55) 2 puxy For dl vanadyl complexes 0 is obtained from the experimental Optical spectrum by use of the equation:81 3 (E xy + Z2) — U (E xy + xz,yz) l o = — — (56) 2 (E xy + x2 — y2) 2 In addition to the ratio of the axial to equatorial field strength it is also possible to determine the field strength of the various ligands using the expressions of Yamatera.82 Defining Ox as the o electrostatic effect due to the 0- bond between ligand x and M and fix as the electro- static effect due to the n bond between ligand x and M, he obtains the following expressions for the d orbital energy levels for MX Y type ions: 5 3 l EXZ "-' Eyz = ‘14- TTX + 1 fly (57) E 2 2 = Ox (58) X - y 2 l E = - o + —. 0 z2 3 x 3 y (59) Some of the Ox and n x values for Co+2 complexes are listed in Table “.10 TABLE u.—-o X and 1T X 31 values for Co III ions. Example OX "x 0+2 ON‘ 2u,000 cm‘l -8,000 cm"1 N+3 NOE 20,000 —6,000 N0 NH3 22,000 0 N‘ N3 2u,000 10,000 00 H2O 25,000 8,000 0‘ OH‘ 32,000 16,000 Oxalate 23,000 6,000 0'2 0’2 39,000 24,000 E“ F' 26,000 10,000 01‘ 01' 24,000 10,000 Br“ Br“ 23,000 10,000 1' 1' 21,000 10,000 EXPERIMENTAL ESR SYSTEM Second Derivative Presentation A Varian83 model #500 spectrometer with 100 kc field modulation was used in this research. It was modified so that the sample could be modulated at both 100 he and “00 cycles, and the resulting signal doubly detected by the two lock-in amplifiers. This procedure gives a second deriva- tive presentation of the signal. Second derivative presen- tation is useful when there are overlapping lines. A dis- advantage is that the sensitivity is lowered by a factor of approximately 16 from first derivative 100 he operation. Measurement of Magnetic Field A small proton marginal oscillator8u was used to measure the magnetic field. The proton frequency was mea- sured with a Hewlett-Packard85 Electronic Counter, Model 52U C, which has an accuracy of i 20 cps. When Operating the Spectrometer in the second deri— vative mode, the proton signal was passed through a variable attenuator into a Hewlett-Packard Model M60 AR Wide Band Amplifier. The output from this amplifier was fed to the Y axis input of the ESR console oscilloscope. With the func- tion switch in position 2 the signal is fed to the U00 cycle lock-in amplifier where it is detected and amplified. The second derivative of the proton signal then appears in the recorded spectrum superimposed on the second derivative of the ESR signal, while the first derivative of the proton signal appears on the scope. The frequency is recorded by 32 33 the H.P. Counter by setting the display time on infinity and pressing the recount button when the chart and scope signal appear. The proton oscillator is then changed-to another frequency down field (for down field sweeps) and the process repeated. The magnetic field in gauss is calculated from the following equation: H in gauss = 2.3A86855 x (01) x 10.4 (50) 01 is the frequency of the proton resonance at the magnet field in question. Measurement of Klystron Frequency Three methods of measuring the klystron frequency are available. (a) A Waveline 698 Wave Metergé-The normal accuracy is i 1 mo, however wave meters are temperature dependent and it has been in error by as much as t 2 me. Other disadvantages are that it is insensitive at low klystron powers and that it is not calibrated directly in megacycles. The frequency is obtained by interpolation from a table of actual fre- quencies versus wave meter readings. (b) A TS-IUB/UP U.S. Navy Spectrum Analyzer.—-The spec- trum analyzer also operates on the wavemeter principle but is preferable to method (a) since the dial is calibrated directly in megacycles and the instrument does not lose sensitivity at low klystron powers. Although the accuracy of calibration as checked against method (c) is only i 2 mc., the precision of measurement ist0.5 me so that an initial calibration before a series of runs by method (c) followed by the use for each spectrum of the spectrum analyzer has proven satisfactory. (c) A Micro-Now Model 101 frequency calibrator.--The kly- stron frequency is allowed to beat with a harmonic of a Micro- Now87 Model 101 Frequency Calibrator and the difference 3A frequency is tuned with a National88 NC-l73 communications receiver. The difference frequency can be read from the calibrated dial to t 0.05 me on the low frequency band and to i 0.5 mc on the highest frequency band. An output is pro- vided after the first R.F. amplification stage of the commu- nications receiver so that the signal can be fed to the Hewlett-Packard wide band-amplifier and then to.the Hewlett- Packard Electronic Counter, Model 52“ C. The frequency read on the counter, minus 955 t 2 kc (The I.F. frequency of the communications receiver) gives the difference between the klystron frequency and the harmonic of the Micro-Now to an accuracy greater than presently needed for ESR experiments. Linear,_Precalibrated Magnetic Field Sweeps A Hall probe89 was used in constant current mode to drive the x - axis of an x-y recorder. The experimental set- up is given in Figure A. In constant current operation the output of the Hall probe is directly prOportional to the applied magnetic field. At 3000 gauss the output is approxi- mately 0.16 volts. Most of the 0.16 volts are bucked out by the feedback network shown in Figure 3 so that a change of twenty gauss at a magnetic field of 3000 gauss will cover the full x axis of the recorder on the 7.5 mv. scale. The method allowed the use Of precalibrated standards to determine the A values of a series of samples by running the known before the first and after the last sample of the series. The reproduci- bility has consistently been at least one part in two hundred for narrow lines. Single Crystal Holder The single crystal holder is pictured in Figure 5. It permits rotation of the crystal in two mutually perpendi— cular directions in the cavity. Its design and Operating principles are as follows: 35 .owpmaonoe mocha Ham: .: onsmwm vomuoo 4 fl- anchm Adam >com coca v8 zom 4'0 x: XmH .HGPOE - + 9.333 >om on on o 36 Finely threaded Brass washer cemented t brass screw Teflon washers to top of glass rod I Brass nut soldered e; to Brass screw I I 241 Brass nut cemented to top of plastic tube aflaltic tube Cork cemented to J glass tube fl 7 Glass tube Vertical slit Adjustable cork —-—) in plastic tube Copper wire through cork Glass tube Thin plastic or quartz rod Stiff thin plas- tic (Duco cement covered Scotch tape) Rubber cement-—] al Pliable bend (flexed Scotch Plastic cryst 4 7H holder Figure 5. Adjustable single crystal holder. tape) 37 The holder body consists of two concentric pieces of tubing, the outer one of glass and the inner one of plastic. The inner tube has a nut cemented to its top and a vertical slit in its body. The outer has a screw secured to its top such that the screw can be rotated but not raised or lowered. This outer tube consists of two pieces which are held together by the fixed cork. The division is necessary so that the heavy c0pper wire can be drawn through the cork and the ver- tical slit in the inner plastic tube. The copper wire pre- vents the inner tube from rotating when the screw is turned. Since the inner tube cannot rotate, it will move vertically as the screw is turned. The left end of the plastic crystal holder (see diagram) will be raised or lowered while the right side acts as a hinge thereby giving a vertical rotation. The adjustable cork is used to position the holder so that the crystal will be in the center of the microwave cavity. Standard Samples Several standard samples were used to check the cali- bration of the x— axis of the xy recorder and the klystron frequency readings. They are: Aqueous peroxylamine disulphonate g = 2.00550 t 0.00005, A(high field) = 13.110 gauss . _ 9O A(low field) - 13.031 gauss Varian standard sample of pitch in KCL g = 2.0028, 1013 Spins for the 0.00033% sample per cm. of length and 3 x 1015 spins for the 0.1% sample t 25% accuracy for the number of spins. Line width (separation of peaks of the derivative):: 1.7 gauss.91 Aqueous K3Cr(CN)5N0 e = 1.99u5u . 0.00005, AN = 5.265 . 0.0590.92 38 Aqueous VOSOu A = 116.13 1 0.2 gauss (separation between the fourth and fifth 11ne)93 DPPH in KCL g = 2.003622 PREPARATION OF SAMPLES Substituted copper acetylacetonates. The preparation of the substituted acetylacetonates is described by Burdett.9u The copper complexes were prepared by adding cOpper acetate to a solution of the ligand in CHCl3. K3Cr(CN)6. The method of Cruser and Miner95 to prepare the potassium hexacyanochromate. was used K3Cr(CN)5N0. This compound was prepared by the method of Griffith et. a1. 96 VO(CN)§3. This complex ion was prepared in water solu- tion by adding excess KCN to a solution of VOSOMP7 K3V0(NCS)u. This complex was prepared by adding KNCS to a water solution and extracting with ethyl acetate.98 Cr(N0)(NH3)SCl3. This complex was prepared by the method of Mori et. al.99 The vanadyl acetylacetonates were prepared by adding the appropriate substituted acetylacetonate to vanadyl sulphate in dimethylformamide (DMF) or by the method of Feltham.37 The copper salts of the amino acids were prepared by ad- 100 ding CuCO3 to an aqueous solution of the amino acid and boiling. The other vanadyl complexes were prepared by the methods 97 . of Holmes. CRYSTAL GROWING K3Cr(CN)5N0 in Alkali Halide Lattices Single crystals of alkali halides containing about 0.1% Cr(CN)5N0-3 were grown from saturated aqueous solutions. 39 The KBr crystals were very easy to grow. KCl crystals were moderately easy to grow. KI crystals were easy to grow, but they often were inhomogenous and contained visible areas of trapped water and visible areas of high concentrations of K30r(CN)5NO. NaCl crystals were very hard to grow and only small (2—3 mm on a side)crystals were obtained. The NaCl crystals were not cubes but contained a combination of the cubic and octahedron faces. The crystals were grown in a desiccator kept at 23 t 200. Small seed crystals of the respective alkali halide were grown from solutions left standing in Open beakers for K01 and NaCl. For KBr and KI small crystals from the reagent bottle were used as seeds. The seed crystals were suspended in the mother liquor by a fine piece of nylon thread. In order to prevent capillary action in the thread and the depo- sit of rings of the alkali halide on the side of the beaker, both the thread and the entire inside of the beaker were coated with Dow Corning high-vacuum silicone grease. Preparation of the Other Single Crystals The other single crystals used in this investigation were also prepared by the desiccator evaporation method described above. The K3Cr(CN)6,K3Co(CN)6 crystals were rela- tively easy to grow as were the OOpper glycinate, cadmium- glycinate crystals. The Cr(NO)(NH3)5+3, NHuCl single crystals had to be grown very slowly. VO(CN)5—3 could be grown in KBr only if there was an excess of KCN in the solution to pre- vent decomposition of the VO(CN)5-3 ion. EQUATIONS AND CONVERSION FACTORS The following equations and conversion factors are included for convenient reference.lOl V g = 7.1uu89 x 10‘7 — (51) H HO v is in megacycles and H is in gauss. V B [A(-)] = (—-) g AH(in gauSS) (52) c in cm-1 ho a = 0.92732 x 10‘20 h = 6.6256 x 10‘27 c = 2.997925 x 1010 " 5 A(-) = (4.668567 x 10‘ ) g AH(in gauss) (53) c in cm-1 1 V AH in gauss = (2.1Ul98u x 10“) — [A (—)] (5“) g c in cm-1 v(megacyc1es) = 1.3996 g H(in gauss) (55) -1 _ -5 0 (cm ) — 3.3356 x 10 (MO) (56) l gauss = 1 oersted (for air) (57) l a.u. = 5.29167 x 10"9 cm (58) a.u.-3 x 6.7u872 x 102“ = cm-3 (59) RESULTS K3Cr(CN)5NO Electronic ConfigUration and Optical Spectra 5 2s of the energy levels are given in Figure 6. At present it This complex is t low spin. The possible orderings is not possible to assign the relative positions of the d and the d levels. x2 _ 2 OpticaI spectra of KBr single crystals and of the 2 2 water solution were obtained with a Beckman Model DB Spec— trometer. Visible peaks were observed at 13,700 cm-l, 22,000 cm'l, and a shoulder at about 30,u00 cm-l. These are in good agreement with the published solution data.102’103 In addition a faint peak was observed which varied from l6,U00 to 17,100 cm-lin the two KBr crystals used. U.V. peaks were observed at 2: 37,000 cm”1 and 2: uu,000 cm’l. Solution ESR spectra The ESR data for K3Cr(CN)5NO in water solution are given in Table 5. The nitrogen hyperfine splittings were measured from samples containing natural abundance 13C while the carbon hyperfine splittings were measured from samples enriched to 15.5% 13C. The 53Cr hyperfine splittings were measured from samples containing natural abundance 53Cr. For the measurements of the relative intensity ratios of the axial and equatorial 130 splittings the experimental Spectra were taken on the low field side of the natural abundance (10%) 53Cr lines to avoid overlapping of the 53Cr and 52Cr sets. No splittings of the 13C lines due to the cyanide ni- trogen was observed even in D20 solution. In concentrated U1 A2 H2SOLl only a single line of width four gauss was present. In DMF the Splittings were about the same as in H 0 except 2 that over several months a second peak (or peaks) appeared. Single Crystal Studies Angular Dependence of ESR Spectra The spectra obtained using KBr, KCl, and NaCl as host lattices are similar and the triplet separation, due to hyperfine splitting by the 1“N (of the NO group), varies with angular setting as shown in Figure 7. When the host lattice was KI no angular variation was observed at room temperature and the spectrum was similar to the aqueous solution spectrum. Some KI samples showed an angular depen- dence at lower temperatures (-10 to —2000) while others gave the same Spectrum as a powdered crystal. The KI crystals chosen were desiccator dried for several months, were uni- form in appearance and did not contain visible areas of trapped water; nor did they Show, by infrared analysis, any greater percentage of water than the KBr crystals. However, the low temperature Spectra indicate that the complex ion has not completely entered the lattice in all the crystals. In describing the crystal Spectra it is assumed that 5 group of the MX alkali the complex ion has replaced an MX6- halide host lattice and that the complex ions are arranged with the Cr-N-O direction randomly distributed among the Six possible positions (which are also the cubic axes of the cry— stal). The vertical axis, about which our rotations were made, is chosen as the z-axis and was always a cubic axis of the crystal; the x and y axes are then the Cr-C-N directions in the plane.lz. The applied field H is along x initially O and e is the angle between H and x at any given setting. 0 The crystals were oriented by manually adjusting the crystal holder in two mutually perpendicular directions until the gH Spectra observed on the oscilIOSOOpe were at highest magnetic field; this was seen visually to place HO along a cubic axis. “3 moooo. oh w mom.e ommm.fi m:mm.e ememm.fi m.o. om.o. m:.m om.w me.m Assess omeve m.oH oa.oh U m.HH m.HH me.mH eo.mfi me.ma “Hoesooos o omeva mods .. m.e mm.m e.m om.m sm.m me.e . om.m AZeHve mo.o. mo.o. Aso Va mm.sfi m.ms me.we mm.ms me.ws mm soa.oom moa.oom moaaeom eoe.eom ‘-mofi.eom.. . . wflmm m om . he Jere A 4 se oZmAzovsomx sou some mmm HoosoeHsooxmun.m momee Au antibondi-: 0 levels e (s*) ‘ * * a1 (a ) (a 2) \ u (no) * b1 (0 ) (dx2_y2) A a (CH) a (NO) bonding 0 levels _$igand atomic orbitals with. Figure 6. Energy level diagram for K3Cr(CN)5NO. “5 Since visual orientation could not be used for the NaCl cry- stals, which were small and irregular, and since the oscillo- scope signal was insensitive over about a 50 range about the maximum, values of A(Nn' (Table 6) for NaCl have a larger uncertainty than normal. AS shown in Figure 7 the g, lines did not change with angular setting Since,when the unique axis of the octahedral ion is perpendicular to H one third of the octahedral ions 2 are always oriented withgheir unique axis..‘L.HO and give rise to a triplet centered at g, which does not Shift on ro— tation about the z-axis. One third of the octahedra are oriented at the initial setting (HO H cubic crystal axis) with the unique axis parallel to H0; these give rise to a triplet centered at gl'which moves monotonically on rotation until at 6 = 900 it coincides with the g1 lines. The third triplet coincides with g, at e = 0 and moves on rotation until it is at g, at e = 90°. The line widths are about 1.5 gauss at g and 3 gauss at g1 and neither the widths nor the Splittingsuwere altered appreciably from the room tem- perature values when the KBr crystal was heated to 3000 0, although decomposition began. A precise study by Hayes“5 indicates a small linear increase in line width with tem- ‘ perature over the range 2A0 to 10000 for K3Cr(CN)5NO in water solution. Spin Lattice Relaxation The gI'lines saturated at about 300 milliwatts, see Figures 8 and 9, but the g1 lines could not be saturated with our instrument. At present there is no theory for this anisotropic relaxation phenomenon. Dr. CowenlOBOf our Physics Department is studying the dependence of the relaxation on temperature and alkali halide lattice. Collapse of Nitrogen Splittings As illustrated in Figure 10 it is possible with over- modulation by the 100 Kc unit-(setting of 800 where modulation 146 \ \ \IAI / .mumv Mmm m OZmAzovno mo oosovcomoc cowumucowno .5 onawwm / \ / to 147 Solid line - 1% db attenuation Dotted line - 0 db attenuation The spectra were obtained at 77°K with the same instrument ‘ settings for both. Figure 8. Saturation of the 3" position of K3Cr(CN)5NO in KBr single crystal at 770K. Figure 9. Saturation of g" spectrum but not the 8L spectrum for K3Cr(CN)5NO in KBr single crystal at 77°x. u8 broadening starts at 125) and second derivative presentation to collapse the nitrogen triplet and leave the 130 splittings undisturbed. This technique iS‘Of limited use since the two sets of splittings must be sufficiently different in width so that both sets will not be overmodulated. With this limi- tation it can be used to resolve overlapping Spectra and remove closely spaced fine structure from larger splittings. l3c Splittings The 13C hyperfine Splittings were obtained from samples prepared with KCN enriched to 15.5% in 13C. The Splitting in the g" position (see Figure 11) is 12.9M gauss and results from four equatorial carbon atoms for which the Cr—C bonds are perpendicular to H0. One axial carbon for which Cr—C is parallel to H0 should also produce a splitting but this is not seen either because it falls at about the same place as the stronger peak or because it lies beneath a carbon peak from the other nitrogens (the observed line widths are 1.5 to 3 gauss).v We symbolize this Splitting as A(13C)L since HO In the g; position in KCl two 130 doublets with in- tensity ratios 3:2 are seen (Figure 12.) If the axial and is perpendicular to the Cr-C bonds. equatorial l3C Splittings were the Same, this intensity ratio could be interpreted as the splittings due to the three 130 nuclei of the three Cr-C bonds perpendicular to H0 and the Splittings of the two 130 nuclei oiofihe two Cr-C bonds parallel to H0. However, Foigman and Hayes have been able to dis- tinguish the axial C splittings from the equatorial for Cr(ON)5N0‘3 in Na2[Mn(CN)5NO] and find that they are not equi- valent. Since the behavior of the axial 13C splitting is, no doubt, Similar in the two systems, the actual intensity ratio is probably 1 to 1; the observed ratio of 3 to 2 may be due to the overlap of the'neighboring more intense nitrogen line which would have the effect of shifting the baseline of the recorded Spectrumt* The outer peak gives A(13Ch. = 12.97 gauss, as at the g”*position; and the inner peak gives 149 .Hox as ozmazovsoox eoeoesao onH «m.ma mo sowuwmom :m now was: ex ooa one mo oo«uma560lho>o he nomad somonuws mo oomeaaou .oa enamam 50 .Heuuhno oaudw» aux ow one deuuhno mamswm sax aw OZmAZUVLOmx omH nonofinno wm.ma mo sapwoomm «mm =u any mo sofipauonpomsm .HH enough ~ ._ . . fl _ l u . fl _ _ . . . . . a — _ .. . . a . c a _ . . . . a s a — . a — — c . . . . a _ . . . . a . . . u . . . u a . .n . . _ n . a a. . . . n . .. _ . l _ . . . a . . _ __ — . ~ . ~ _ ~ — a x — - _ ~ — . __ ._ ._ ._ _ i. - - — - ~ 1 u . _ . .. u n u N . _ semen mo . . _ n .. _ . 63» t3 mmo " a u a n u oowa and Hesowuaoem — . _ . _ .. . . s mafia somonuwa I a a _ . u x a and u . ,/ nmx am 02 Azovno x a oawa weapon ,L aux ow OZmAzovaomx u osma vfiaow a a r _ .g '8 .3 '8 'I '8 51 .osmam 4 m ea ozmxzousomx 0 ma ooeoasao rm.ma mo omsapuaaoo Una .ua enemas eoaa una a a osaa somonuas u a a aepahnu eawsae nmx ea 4w so next owoso scam one - mafia eaaom amuuhnu manna» aux ca soauanoa 4w um mane cameo macaw a oaaa weapon 52 A(lBCb' = 9.35 gauss. In KBr,the 130 lines were not Split in the gL' position. At e‘= H50 the four equatorial cya- nides Should be equivalent. The Splitting at this position is 11.35 gauss (labeled A(l3Ce ) A50 in Table 5) and was used along with the A(13C)‘L value to calculate A:3C l Ap3C in KBr. and NO—Cr-CN Angle For the KBr and KCl crystals the g1 data were taken at the same crystal orientation as the g”. X-ray Spectra of the Similar compound Na 2Fe(CN) 5NO indicate that the NO- Fe-CN bonds make an angle of 960 with each other. An attempt was made to check the gH to g; angle; however, 109 the g, spectrum is made up of two mutually perpendicular orientations which behave differently on rotation. Any ro— tation of the crystal to see if g, would move farther down- field resulted in a broadening of the line which masked any shift which might have been present. The regular structure of the alkali halide lattice would tend to minimize the non— perpendicularity of the NO-Cr-CN bonds. As a check on the assumption that the g” orientation also gave the g“L orien- tation the powder Spectra of crushed KBr and NaCl crystals were also run. Comparisgps of Powder Spectra Data With Single gpystal Data Second derivative spectra gave excellent resolution of the Si lines in the crushed single crystals but did not resolve the g” Splittings. A comparison of the powder data with the Single crystal data for KBr is given in Table 6. Within experimental error the data are in agreement.- The g; and AL data for KBr and NaCl listed in Table 5 were obtained from the powder spectra. 53 TABLE 6.--Data for Cr(CN)5NO-3 in various lattices. Quantity KI KBr KCl NaCl K3Mn(ON)5NOP ——. 1.9722 1.9722 1.9750 1.97u5 8” —— 10.0005 10.0005 10.001 —— —— 2.00uu7 2.001u 2.00u5 2.0052 5; -- 10.0002 10.0005 10.0005 -— 1 99975 1.9937 1.9936 1.99M? 1.9950 g 10 0001 10.0008 10.0008 10.001 -- AN -- 2.89 2.11 2.1 2 06 on -- 10.1 10.1 10.2 —— AN -" 701 701 7.12 6090 l -- 10.2 1O 1 10.2 -— AN 5027 i 0.05 -5071 _5053 .50)4 -5029 s -- 10.2 10.1 10.2 -- An -— +l.A1 +1.57 +1.70 1.63 p -- 10.1 10.1 10.2 -— A130 —- 13.05 12.95 -— 15.0a l. —- 10.1 10.05 -- 8.79b aequatorial baxial CRef. 104 54 TABLE 6.--(Continued) Quantity KI KBr KCl NaCl K3Mn(CN)5NOC A13C -- -- 9.35 —- 10.78a " -- -- 10.5 -- 7.59b l3c —- 11.35 -- —— —- A o -- 10 5 —- -— —— 45 ' A130 -- -11.92 —11.75 —- -12.16 s —- 10.3 10.2 -- -- A130 _- +1.13 +1.19 -— 1 15 p -— 10.3 10.2 -_ __ A53Cr -- 33.“ 33.4 -— 32.5 In —- 10.8 10.1 —- —- A53Cr -- 11.9 11.7 -- 11.5 .L -- 11.0 10.5 -- -- 530r -- 19.1 18.9 —— 18.u6 A -- 11.0 10.5 -- -- aequatorial baxial cRef. 104 TABLE 7.--Comparison of powder crystal ESR data. and KBr Powder KBr Single Crystal t0.0002 i0.0010 N _ N _ AJ. - 7013 J- — 7005 t0.0S t0.20 55 Infrared Spectra Infrared spectra of K3[Cr(CN)5NO] and K2Fe(CN)5NO in KBr powder in mulls, of KBr single crystals,and of KBr and KCl pellets made from crushed KBr and KCl single crystals were made using Beckman IR 5 and IR 7 IR Spectrometers. Since the NO stretching frequency and a water peak fall in the same region (1600 cm-1) single crystals were also grown from D20 where the water peak is shifted to 1200 cm'l. The IR data, along with some data from the literature for com- parison, are given in Table 8. The IR spectra are given in Figures 13 and 1n. K3Cr(CN)6 Crystal Orientation Single crystals of K Co(CN)6 containing less than 3 1% K30r(CN)6 were grown by slow evaporation of the saturated aqueous solution in a desiccator kept at 23 t 200. By visual inspection the crystals resembled the crystal illustrated in Figure 256 of Chemische Kristallographie.llO The crystal angles of the apparent'myand a,faces were measured with a reflecting goniometer. The three principle magnetic axes x,y,z and the crystallographic axes a,b,c are related by the direction cosines in Table 9a.111 There are two nonequivalent octahedra which transform into each other by reflection in the ac plane. Each of the two cyanide octahedra has a two-fold axis along the c axis. The four-fold axis normal to the c axis makes an angle of 27° with the b axis. The direction cosines of the molecular axes (a,8,6) of the two octahedra are given in Table 9b. The crystals were cemented, in the approximate orien- tation desired, to a cavity insert which was adjustable in two perpendicular directions as shown in Figure 5. The accurate positioning of the crystal was then accomplished by observing the ESR spectrum on an oscilloscope, adjusting the two probe directions and rotating the electromagnet; also the above 56 TABLE 8.-—Infrared data for [M(CN)5NO]X compounds. l Host v(NO) in cm- V(CN) 0(M-C) [Fe(CN)5NO]-2 Mu11a 1938 —— -- KBra’b 1938 -- __ KCla’b 1925 _- __ NaCla 1938 -- -- [Cr(CN)5NO]-3 Mu11a 1616(3) 2116(3) -— 163M(m) -- —- Grounda in '1637(8) 2123(8) —— KBr Pellet froma 1634(5) 2123(s) 350 KBr Single 1656(s) 2101(wtos) u01 Crystals __ __ 93A a figilggngggm 1707(5) 2103 3u7 Cr stals 1685(5) 2112(m) 351 y 1654(8) -- 397 __ __ ”28 a Pellet from 1696(8) 2118(3) __ NaCl Single Crystals iggggzg aThis work. bReference 169. c s = strong, m = medium, w = weak. .muuon usofitgp aw oZmAzovnomx mo .Huau ooaa op, <1; Haau ooo.m Baum .muuoona vununmcH .ma «human an: ad aux ad 57 Hon: 64 An uo:v xauan aux and: unaonu ll\\\lll\\\\lllllu oaaan» 58 .opoon osoaum> cw mcozmhzovom mo . H .38 ooaa op a uau ooo.m 309m .mnpuomm vvnmpmcH L .34 ansmfiu ca 59 TABLE 9a.--Direction cosines between the principle magnetic axes (x,y,z) and the crystallographic axes (a,b,c) for x y, z a 0.10u 0 0.998 b 10.994 0 10.109 c 0 l 0 TABLE 9b.--Direction cosines between the molecular axes (a,8,5’) and the crystallographic axes (a,b,c) for K300(0N)6.l a B 1U a 0.U5 —0.63 0.63 b 10.89 10.32 1:0.32 c 0 0.71 0.71 mentioned relationships between the magnetic and octahedra axes were utilized along with the fact that there are two non-equivalent octahedra that transform by reflection.111 The spectra were taken at room temperature and at approxi— mately 1260K. The line width decreased as the temperature decreased,but the 130 splitting remained constant.‘ The spec- tra were taken in five crystal orientations - the magnetic x, y, and z axes, along a Cr;C bond axis, and at 610 to each of two of the Cr—C axes. Second derivative presentation was used. A typical spectrum is given in Figure 15. 60 .mAzovoomx chImAzovno 60:3".qu Una wmda mo savouao «mm .3. ogmwm 61 ESR Spectra The spectra were fitted to equation (11) where I is the projection on H of the 130 nuclear spin of the complexes containing only oneoisotrOpically substituted 13C (which consti— tute A0% of the total number of complex ions in our case), and is equal to i %. H0 is the center of the spectrum for the complexes (33% abundance) which contain no 130. In each orientation chosen two of the Cr-C bond di- 0' The hyperfine splitting observed is attributed to the cyanide rections are equivalent and make an angle 6 with H carbon atoms lying along these axes. ~Any splitting due to the remaining two cyanide carbon atoms would be half the in- tensity and is neglected. The splittings due to the 9.5% abundant 53Cr isotope and to the complexes containing two or more 13C atoms are neglected; no correction has been applied for distortion of the measured peaks by these or by the cen- tral line of the spectrum. Off-diagonal terms in the hyper- fine splitting tensorA2 are also neglected. The values Ap =1 0.69 gauss and AS = -9.80 gauss account for the data from four orientations (resolution along the fifth direction, H0 parallel to the x - axis, was so poor that quantitative measurements could not be made). At least seven spectra were taken for each orientation and the average deviations were:: 0.03 gauss. In view of the approximations involved, it appears more realistic to report Ap = 0.7 i 0.2 gauss and As = -9.8 i 0.1 gauss. The experimental data are given in Table 10. There is another possible set of solutions to the ESR data. It is AS = 13.73 gauss and Ap = i 6.76 gauss. This set gives a carbon n spin density ofgg 60% on each of the six carbon atoms. Since the total spin density in the complex is only three electrons, the value of the total carbon spin den— sity obtained with this set of As and Ap values,:: 3.6 elec— trons is physically unacceptable. 62 TABLE 10.--13C ESR data for Cr(CN)g3 in K3Co(CN) O‘\ m I .‘I: ll Along the C-Cr bond axis (8 = 90 5.26 gauss Along the y magnetic axis (8 = M5 H - H = “.79 gauss Along the 500 rotation o position (8 = 76 ) H - H 5.99 gauss I CE ll ' Along the magnetic z axis (8 = 56Ou5 )H 9.93 gauss VANADYL COMPLEXES The isotropic g and A values were obtained from the liquid solution spectra. In solution vanadyl complexes have an eight-line ESR spectra due to the hyperfine splitting by the vanadium nucleus with nuclear spin of g. The A value is measured as the magnetic field separation betweefi lines A \) and 5. To first order the g value is equal to 8H where v is the klystron frequency in megacycles and H iésthe mag- 5,5 netic field in gauss halfway between lines ”sand 5. The second order correction is significant so that the actual magnetic field that is used is: 31 A2 H = H + ——————- (60) U,5 u H ( “’5) The anisotropic terms were determined from dilute powder spectra, frozen non-aqueous solutions, or dilute sin— gle crystals. A" and AL were measured from the separation between their respective*1ines H and 5 when resolution per- mitted. Otherwise they were determined from the other line separations with the use of equations (26) and (27) The ESR A_values for substituted vanadyl acetylace- tonates are given in Table 11. Vanadyl acetylacetonate, o u mo u no a.mouo ma ecmwfifi mesa .o.HH mmsww CH mzam>.mo .m.ow mmzmm CH odam>.mm 63 mo.mHH mm.mHH mm.mHH am.:HH mmo mmo we.m0H no.moH mo.moH 8H.HHH Hwy mmo am.moH no.00H am.mOH 83.0HH mmo mmo am.NOH ee.moa mo.m0H as.soa e e «H.50H no.m0H mo.mOH 88.50H mmo a aw.eoa no.00a mm.m0H mm.moH mzo mmo use :H weepmee cH mfiomo 8H. mcmNcmm :H m m 2 opmcopoomampmom szwcw> condufimeSm mo mppomdm mmm coaudaomll.afi mqm H. moo.o. m. moo.o. m.o. moo.o. mam :H m.mo ssm.H omH mam.H w.moH mom.H mHo m.o. mooo.o. m. moo.o. m. mee.o. amazom hmm :H H.Hm magma.H emH mmam.H e.mm Hma.H m-mAzevo> . . . Hopmzho HH mooo oH NH mooo OH NH mooo OH mammam hmm :H Hm 22mm H mHH HHsm H m mm comm H mumAzovo> In I: a: u- m.o. mooo.o. 0mm 8H u- u- u- .. a.aa mamwm.H mnmfizavo> I: u: nu a: m.o. moo.o. mam :H u: u- an n. H Hm mmsm H mumfizovo> 4. A. z 2 I m< -.w dd 1 w m¢ w ocuoafioo mmonQEoo Hmpmzm> pom dump mmmll.mH mqm In I: m. moo.o. m.o. In eoHom :H u- u- mom ommm.H o.omH u- AeoHoo> H. moo.o. m. moo.o. m.o. moo.o. mHomo :H a.ee eeaa.H m.me mama.H a.eeH aea.H mkeeeevo> II II we moo.oH Q.HH moo.oH maomo Cfi .. -- mmH eeea.H o.aeH weem.H mmxmozvo> H. moo.o. m. moo.o. m.o. moo.o. mam cH m.ee aaa.H omH maea.H m.HOH amea.H mmxmozvo> a- nu u: u. m.o. moo.o. accumoa :H u: nu u: I: m.MHH mwm.H mHo .4 4 : .. m< w m< w m< .w UCSOQEOQ omscfipcoovll.ma mqm nu nu u- I- H. .1 12a c rmeHHoaepce nu I- -u u: HOH nu u: figopgepoao> H. moo.o. nu u- m.o. nu mzo eH m.mo sam.H nu u- m.BOH nu meHeHasa+Homo> HH moo.o« II II m.oH II mzm cfl m.om mmam.H a: nu m.moH uuNmeeHeHasaHmua.avo> II II NH moo.oH m.oa II omza CH u- nu mmH mamm.H m.mHH nu Homo> HH moo.oH NH moo.oH. m.oH II mzo :H 9:. SSH HaH 23H :52 I mxaaaemaemw v8 4. d. 2 2 . m¢ w m< w m< w UCSOQEOO Aomscfipcoovll.ma mqm¢9 68 mmsmw CH 0 I. u- I- :u 0.0. 000.0. temhemm 0H -u u- u- I- 0.00H 000.H 016000000 :1 u- nu u- 0.0. 000.0. 020 0H u: u: u- a- 0.:0H 0000.H NH000<00> nu nu I: u- 0.0. u: 0mm.mm 2H II II II II o.@HH II mmmo> II II II II m. 0“ II HOS CH u- u- I- nu 0.0HH nu NH00> I- 000.0. I: u- 0.0. u: 020 0H 0.Ha 0ea0.H -- -- 0.0HH .. a0000. m.o. moo.o. II II II II mus pcomso 0.01 00.0.H .. -- -- .. eH.mAeeoevo> 4. #w : :m w 0900 d¢ dd. dd. USS U Avmscfiucoovll.ma mqm<8 69 nmcwa 4< mmpmoavaa \/ moswa =< moumowvaw < .nmx ca mumazuvo> mo asnuooam «mm dupmuuu mawcwm .ma onswwm 70 39.3 4< «33209.0 > was: :< moumuwvcw < .eoHuaHounHomo acuoam 0H 0-0H002000 mo 55100000 «mm .sH ohamHn 71 TABLE 14.-- 13C hyperfine splittings in VO(CN)g3 A A...‘ Solvent or~- A_ A(at gl) A(at g") Matrix H2O 11.3 Gauss -- —- 10.2 __ -_ DMF Not Observed 10.65' 9.95 i 0.05 KBr (Single Crystal) -- 1M 11.7 i 0.5 Covalency in Vanadyl Complexes The equations used for the determination of the mole- cular orbital coefficients from the ESR data are equations (29), (30>, <35), and (36). Using {-l3->= 3.67 a.u.,ll2 A0 (see equation (35) and (36) is equalrto’l8u gauss. Approxi- mating equation (35) by A" = — $.Ng AO - A one obtains the N82 values given in Table 15. This2set of Ni coefficients indicates considerable xy n covalency for all of the complexes studied. (An Ni value of 1 indicates a completely ionic com— plex and an Ni value of 0.5 indicates a completely covalent complex.) Perhaps a closer approximation to the actual cova- lency can be obtained by setting A0 at the value which gives N32 = 1 for VO(H2O)5+2 and calculating the molecular orbital coefficient for the remaining complexes on this basis. These values are listed as Ni . For comparison the molecular orbital coefficients obtained from the rati22of Aisotropic for the isotropic"for VO(H20)5 are included. These values are listed as Ng2. complex over A 72 .0-0H2000> nonoHnno mafia Una hexane nowumnnwamo vaowm owuouwma 00H oeHH eaHemam> and mo sapwood» scapsaou uaoo=a< .mH onswwm mafia 0 ma noxnma newumnnwamo,. vauwm owuoawua TABLE 15.--Mo1ecu1ar orbital coefficient for vanadyl xy orbital. Ni2a Ni; b N22" c VO(H2O);2 0.79 1.00 1.00 V0(DM30)5(010u)2 in DMSO 0.81 1.02 0.93 vosou in DMSO 0.79 0.995 0.95 VOClOu in H010“ 0.79 0.99 1.01 v06? CF3AcAc)2 in DMF 0.775 0.975 0.92 VO(CF3AcAc)2 in DMF 0.775 0.97- 0.91 VO(CF3CF3AcAc)2 in DMF 0.765 0.96 0.905 V0(NCS)5-3 in 011013 0.71 0.93 0.90 VO (porphyrin) in THF 0.7“ 0.93 0.805 v001g3 0.735 0.925 0.92 VO(AcAc)2 in 0H013 0.72 0.91 0.895 V0(NCS)-31n DMF 0.715 0.90 . 0.88 V0(0N)§ in KBr 0.625 0.79 0.70 aBased on A0 189 gauss in Equation (35). bBased on A0 145.1u gauss in Equation (35). Chosen 2 to make N = 1 for VO(H 0)+2. n2 2 5 CBased on ratio of A over A . . of +2 isotropic VO(H2O)5 (119 gauss). isotropic COPPER COMPLEXES Single Crystal Study of Cgpper Glycinate in Cadmium Glycinate 8 H mutually perpendicular directions and observing the ESR signal and g_L were found by rotating the crystal in two on the oscilliscope. g" is at the lowest possible magnetic field and g; is at the highest possible magnetic field. Spectra of crushed dilute COpper glycinate in cadmium glycinate single crystals were also taken as a check on the data obtained from the oriented single crystals. The data are given in Table 16. The spectra are given in Figures 19-21. 7H .mmsmw CH mum moCHm> «a II II II II II II II II II COHpsHom II II II m.mo II II II II II 0mm CH II II II II II II II II II N Coozom II II II II II mMH II II oom.m Amevoo CH ApmpmHCOHmov ApmpmHsonov . AoopmHCOHmov Hmpmmkw m.oh m.OHm.OH H.0H O.HH m.o« Noo.oH moo.oH HOO.OH meCHm o.w w.© :.OH N.mw H.:: MMH mmNH.N w:mo.m :Nwm.m NAmevUO CH m 4. a 4. 2 .1 .. Z? 2% 2% 30¢ 50< 30< m . w w Ommx . AHHV CmQQoo I AOCHoszVImHQ Com mmpmo mmMII.oH mqmHum>Hnov vacuum .ou 013mm 77 .Hmumhhu onCHm onwaHohHw “HHvauHavmo IoumCHOHHm HHanommoo m mo aoHuHmon 4w um asnpomam «mm o>pr>Hnuv vacuom .Hw whamHm 78 TABLE 17.--ESR hyperfine splittings for copper(II) amino acid complexes . Ligand A(gauss) Ligand A(gauss) Isoleucine 72.8 a-Alanine 69.6 2-Aminobutyric acid 71.8 Glycine 68.8 Valine 71.8 * NePhenylglycine 68.6 2-Aminoisobutyric acid 71.2 B-Alanine 51.0 Serine 71.0 Solution Spectra of COpper Amino Acid Complexes The isotropic COpper nuclear hyperfine splitting constants of some water soluble copper amino acid complexes were determined from their solutipn spectra, see Table 17 above. Solution Spectra of Substituted Copper Acetylacetonates The complexes were dissolved in CHCl3. The high pre— cision of measurement of the ESR A values was made possible by recording the spectra on an X-Y recorder whose x axis dis- placement is proportional to the magnetic field. This was accomplished by using the Hall probe circuit described in Figure 4. The A values are given in Table 18. The frozen solution ESR Spectra are given in Figures 22 through 29. The ESR results were all obtained in chloroform solu— tion since the effect of changing solvent can be larger than the substituent effects; thus, the hyperfine splitting A for copper acetylacetonate in chloroform is 77.2 x lo-ucm-l (g = 2.123) whereas it is 66.1 in dimethylformamide (g = 2.138) and 56.7 in pyridine (g = 2.198). .moCHm> HmHHMCma UCm CoHpCHom 0C» Eopm UmpmHConoo o 79 o 0 .HI50 000.0 I a 00H: 000.0 + A0 I400 mm I H0 I .mv_m + I« I 0&0 m." NZ de0: 0000H00H000 O . M pCmmeoo m Csz Aowv COHpmsvo wCHm: pmpmHCOHmom AVG 000. 000. 000.0 000.0 0HH.0 0.0H 00H 0.00 HspHssdHn 000. 000. 000.0 000.0 H0H.0 00 0.00H 0.00 000 000 000 000. 000. 000.0 000.0 00H.0 0.00 00H 0.00 000 0 000 000. 000. 000.0 000.0 00H.0 0.00 00H 0.00 000 m Hsspna 000. 000. 000.0 000.0 00H.0. 0.00 00H 0.00 000 0 000 000. 000. 000.0 0Hm.0 HmH.0 0.00 00H 0.Hs Hsocmse m 000 000. 00H.0 HIE0 0.00 000 m 000 b o 4 ._ 4 _. p 02 a 02 00 m w 00 a 0 .sm :0 .0 .mmpr0p ImomHmpoom AHHV Cmddoo ompCpHumCCm Com mmCHm> fl oHQoppomHCm UCm oHQOCpOmHII.wH mqmg8H is discussed in detail by Schulz-Du Bois.113 The effective g value along the mole- cular g_axis is still g, but the effective g value in the xy plane is approximately M. This situation is not unusual for Cr+3 complexes. In order to obtain precise data one must go to higher klystron frequencies than x-band and higher magnet fields than the 3000 gauss region. 'A limited amount of in- formation can be obtained by utilizing Shulz-Du Bois' work and looking at the angular dependence at x-band frequencies. However, our main interest in this compound is in the nitrogen hyperfine splittings. Since they were not observed, the spec- tra were not analyzed further. 'The ESR spectrum of IVIn+2 in NHuCl has been reportedllu and it also has D)> gBH. 89 .Hmumhuo onCHm H0332 CH m+mAmCZVAozvao mo manned» mmm .om unamHm mason coo an“ 0>IJHIIIIII(I()I( vHon uHuunuma on comm Huumhno OHACO O I 7.31 pHon oHuonmma on 00 00mm Hmummno 0H 0 mmsmw o DISCUSSION K3Cr(CN)5NO Solution-ESR Spectra There are several possible explanations for the dif- ference in A values of the equatorial and"axia1 carbons. One is axial destabilization of thetcomplex so that the axial cyanide is not held as strongly as the equatorial cyanides. Neither the internuclear distanceslog in Fe(CN)5NO-2 nor the rates of cyanide exchange106 in Cr(CN)5NO"3 support this suggestion. Another possibility is that the unpaired spin density reaches the cyanide n orbitals through the xy -'n molecular orbital and that the n spin densit 'polarizes the paired s elggtrons to give the equatorial As C value while the axial AS value can only arise from a configuration interaction mechanism of the type given by equation (42). Single crystal work rules out this possibility since the n spin density would only produce an AS value of 2 gauss (or less) if the ratio of Ap to AS is similar to that found in organic systems.115 In addition, spin polarization would give AS a positive value115 while the observed AS values are negative. It appears that configuration interaction as formulated in equation (U6) is the principle mechanism for pppp the axial and equatoriall3C hyperfine splittings. ‘The larger value of the equatorial Splitting can be explained by putting the x2 - y2 energy level below the 22 energy level since A: C in this mechanism is inversely preportional to the energy difference E - E (zzor x2-y2)b. * (z or x2-y2) 9O 91 One cannot estimate the coefficients of the molecular orbitals from equation (M6) since the excitation energy for (x2—y2)b + (x2-y2)* is not known. 'The net spin density at the ligand can be calculated from equation (26) (see Table 19.) The nitrogen As hyperfine splitting of -5.56 gauss can be accounted for by an explanation similar to that employed for the cyanide hyperfine splittings. Electronic Configuration and Optical Spectra Assuming that the x2 - y2 level can be assigned to a lower energy than the Z2 level on the basis of the isotropic A values two possible energy level assign- ments are given in Table 20. The assignment in Table 20a is supported by the observed charge transfer transitions at 37,000 cm"1 and uu,000 cm'l assigned as xy + «*(NO) and xz,yz + n*(NO), respectively, which can consistently be since the difference in energy between'these two bands is approximately the observed xz or yz + xy energy. The assign- ment in Table 20b draws support from the ESR gL value which can only be reasonably explained by placing the xy + 0 *(NO) charge transfer transition at low energy. Gray et. al.162 have recently calculated the energy levels for Fe(CN) 5NO 2 and find that the 0*(NO) is at low energy. However, 5their extension to Fe(CN) 5NO 3 appears incorrect since the ESR data indicate that 5the unpaired * electron istin the Z2 orbital rather than the 0 (NO) as they 2 propose.77 The z orbital assignment is supported by the similarity of the anisotrOpic nitrogen hyperfine splittings163 with those found in copper‘phthalocyanine.126 The 13C A value of the equatorial cyanides is 10.0 i 2 gauss.77 This is consistent with a 22 ground state but is too large for * an 0 (NO) ground state. 92 HmpHCCo an C0 wCHECmmm oopmHCOHmoo .HMHHCCO m mCCQ m wCHECmmm ooHMHConom II o I o I o I o I o I O 00 m 00H 0 00 H 000 0 000 0 e 00 000 0I 00 0I 000 HI 00 0I 000 HI 00H HI HmHHOHpsam II II II II II 000.HI copsmo II II II II II 000.0I H0HH< II II II o.m.m+ II 9mm.HI Cowomsz II II II 000.0+ II 0000.0I HmeH 000000 0Hzovoomx :H 00m :0 020 0H 000 dH 000 0H 5000 0H20Vsomx 0Hzovpmmx 020H2000000z 0H2000>mx 020H20000 x m . II I I II V I III I __ -..w -.. m I mHmthsp unstH ast0 0H H 00 spHmcmp eHdm Hsthso pII.0H mHm one 00H.ComCMCHHmm oCm mmpo go mCOHumzdo 0C» UCm HIEo meH H mm oCm HIEo H.H0 u 2m wCHms oopmHCOHmo whopoEowd COHmHCQoH OHCoHpooHoCopCH. HO H:. II 000.H0 "H020H 0zVH0H00V + 000H I 00 I 000H HO H: 000.00 000.00 "H020H 020H0H00v + 000H I 00 I 000H 00 I 00 so 00 HO NF II 000.00 I H 020H 0200000 + 000 I 000 I 000H 00 I 0x 000.0H 000.0H n 000H 00 I 0x I 00 II oom.w n Pam I mom ax I 0% Ho 0x Nb Hz. II 000.H0 u A 020H 0200000 I 000 I 000 + 000H I mo d: HIed 000.00. HIsp 000.00 I H 020H 0zv0H0H + 000 I 000 + 000H 00 I 0x I 00 so 00 HwCoCm oo>pomno zwpoCm HmoHpoCooCe. COHpHmCmCB m IOZmAZQVHo HOH pCmECmemm Hm>oH meoCo oHnHmmomII.mOm mqm¢e 99 TABLE 20b.--A1ternate energy level assignment for Cr(CN)5NO-3 Transition ~Theoretical Energy - Observed Energy xz or yz + x2 - y2 -- 30,900 cm-l xz or yz + xy -- 13,700 2 2 xy + x — y 16,900 16,900 2 xy + z _- -- 2 xz or yz + z -- —— * xy + 0 (NO) -- 22,000 Dq = 1,690 cm'l, Ds and Dt cannot be determined. Molecular Orbital Coefficients From Single Crystal Data The experimentally determined hyperfine splittings, Ap, are related to the spin density by equations (19) through (23). Note that A1T = 0 for nitrogen since fTr = f1T . The ligand molecular orbitals chosen are: XZ yz 1* -N [I (I II) (I I») I (61> 02 02 x2 _ y2 2s 2s l3C pO pG 13C 9 ' N [w or w + (A w ) + (A w ) ] (62) T‘1 Tr1 3dxz 3dyz "1 pn N 1T1 pt 130 = N + 6 9,2 “2tw3dxy (Ifl2wpfl)l3cl ( 3) The overlap integrals were estimated by graphical interpolation from the tables of Jaffe and Doak116 using the Z6 values of Hartree117 for carbon and nitrogen atoms. ff 95_ The values obtained are 82p 3d = 0.15 (Cr-N) and 0.10 (Cr-C); , s = 0.20 (Cr-N) and 8.17 (Cr-C); S = 0.10 (Cr-N) 2po93d0 and 0.12 C9-C). For d5 low—spin complexes the g values are related to 2s,3dO the molecular orbitals by equations (29) and (30). These equa- 1 case but can be used for d5 low tions were derived for the d spin complexes by changing the sign of the spin-orbit coupling constant when the transitions are of the hole type. The value of the spin-orbit coupling constant,A , for Cr(I) is 212 cm-1.6 Since the chromium is already in the low valence state of +1 no correction of the type proposed by Murao“7 (to account for a reduction in 1 due to screening by the partial transfer of the bonding ligand electrons to the metal) was applied. Estimation of N 1'2 A value of N1T = 0.91 is obtained from the observed chromium hyperfine splittings (Table 6) by use of equations (29) and (25). Where A0 = E$32- = 29 gauss was computed using r"3 = 2.73 i 0.9 A.U. egtimated by the semi-empirical 79 equation of Korol’kov. From the values of NH and S 1r(Cr—C) estimated above 2 d, 2 —1 we obtain 1.2 = 0.32, since N"2 - [l + 9xfl2Sd,fl(Cr-C) + Afl2] . The equatorial Cr-C n— covalent bonding is 2 (I — 23d 1T(Cr—0) )2 = 0.012 (69) Ni 3.2 = N. I. 2 2 2 2 ’ and so is quite small. Carbon—Chromium o—Bonding *— 2+ bl ._ + b2 (charge transfer) = 50,00 cm_l (approxi- The observed value of g , along with values of b 13,700 cm‘l, b2 mated). T(n) = 0.273 for an s-p hybrid carbon orbital, S = {£3 (S + S ) 1 2 d,s d,pO ’ 0 estimated constants leads to the value NC2 = 0.81 if the = 1.0 (by iteration), and the previously 96 charge transfer correction, equation (22), is used. This is reduced to N02 = 0.78 if the X2A2 term of Lacroix and Emch52 is used which takes into consideration the spinborbit coupling constant of the ligand. The amount of covalent bonding may now be estimated as N 2 yo2 = N02 [10 — Sd’O(Cr-C)] = 0.39 (65) * The alternate energy level assignment b2 + bl = -l _ 2 _ 16,900 cm leads to No2 ~ 0.86 and N02 02 - 0.26. The Cr—C 0- bonds thus have considerable covalent character. n- Character in the Axial Bonds to Chromium If it is assumed that the metal dxz and dyz orbitals are involved in n-bonding with nitrogen only, and that charge transfer contributions can be neglected, the amount of elec- tron-transfer2 to the ligand, the n-covalent bonding, is found to be N1T H591 ”v 0.66, for the optical assignment b2 e = 8 ,300 cm-1; this value is reduced to.~,0. 63 if the assignment b2 + e = 13,700 cm 1is usid. If only the xy + n (NO) charge transfer correction is considered, Niffil'2:.50 for the optical assignments b2 + e = _ * - 8,300 cm 1 and b2 + e(n ) = 37,000 cm 1; but anfnlnz.35* for the Optical assignments b2 + e = 13,700 cm 1and b2 22,000 cm’l. These values qualitatively indicate a highly +e(1r*)= covalent xz, yz n bond to the nitrosyl nitrogen. Estimation of Values of AD The internuclear distanceslogFe - N = 1.63AO and Fe — c = 1. 91A0 found for Fe(CN) 5N0 2 should be close to the Cr - N and Cr — 0 distances in Cr(CN) 5N0'3. These lead, with 13 equation (17), to the values AN = 0.97 gauss and AD C = 1.06 D gauss. 97 UnpairedEflectron Density on Nitrogen The observed A: value of + 1.91 is too large to be accounted for by the-AD term alone in equation (19). The anisotropic splitting in excess of'the1dipolar (AD) splitting is about + l gauss. This can be accounted for by positive Spin density in the sigma orbital or negative spin density in the w orbital or the difference between~+~(-)>spin density in the sigma orbital and + (0) Spin density in the n orbital if both are appreciable. The npgative spin density in the nitrogen s orbital indicates that the sigma p Spin density will also be negative assuming that the sigma orbital is an sp hybrid. With this hybridization we can estimate the contribution of the p part of the sigma orbital to be -0.16 gauss.' This value combined with the AD value of +0.97 gauss and N2 *= 0.83 leaves a splitting of +1.2 gauss to be accounted for by'n spin density This is a n spin density, f", of -0.07. Assuming that the spin density arises from configuration interaction Fortman and Hayes102 have derived an equation similar to equation (96) which accounts for the n spin density. 9 n 2 u (dxysdxz)Nfll(Al) f" = - — , T (66) 3 xz + fl (NO) x - With the assignment xz + n (N0) = 95,000 cm 1 (Table —1 20a), and Fortman and Hayes estimate of (dXy dxz) = 3,868 cm , one lobtains Nfi fi<1 )2 = 0. 61 while with xz + n*(NO) = 36,000 1(Table 200) one obtains N:(11N)2 = 0.99. Calculations of this type are not expected t8l give quantitative estimates of the molecular orbital coefficients. In this light the qualitative agreement between the N: (1N)2 value (230.35) obtained from gL by the reduction of the spin-orbit coupling constant (equation (30) and the value obtained with equation (66), Ngl(xi)2:: 0.55 can be interpreted to indicate signifi- cant xz and yz chromium - nitrogen n bonding. 98 Anisotropic l3C Splittings The equations for the angular-dependence of the 13C hyperfine splittings (equations (38), (39), and (90))indicate that if there is appreciable covalency in the xy w bond such that I f 2Ap |>> | N2 2~AD| the symmetry will be AE‘“ AG 7‘: AC but if the xy 1! bond 2is essentially ionic, AC’VAC ¢ AC since in this case the main contribution to the angular1T dependence should be NgeAD. In KBr we were not able to determine the axis of symmetry, but in K01 it appears.that the AO axis is the symmetry axis. The A: value found with this.assignment, A; = +1.13 D,~l.06 gauss. This interpretation is in.agreement with our work on gauss,is consistent with the value calculated for A K3Cr(CN)6 where the anistrOpic l3C.splitting is of the same magnitude as AD. Fortman and Hayes,102 in a study of Cr(CN) SNO—3 in K3 Mn(CN) 5,N0 have found that A C3 AC 0;: A: Their analysis indicates that this is due to f 2Ap beingZ of 2 comparable magnitude to N "22D Using equations similar to (38), (39), and (90) with the assumption that ffllAp = fG G2Ap = 0 and Marshall's 75 higher order equations for N22AD, Fortman and Hayes obtained a pX spin density (f1T 2) of’V'O..03. Equations (38), (39), and (90) allow three separate determinations of fn 2Ap. From Fortman and Hayes' analysis one obtains fngAp = 0. 69, 0.96, and 1.20 gauss. The agreement is satisfactory considering the approxi- mations made and the experimental errors; however, if one attempts to further improve the treatment by estimating the fO2Aé contribution from the experimental fSA; and a model of an s-p hybrid orbital, the agreement is lost; and one obtains f"222 = 1.55, O, and 1.57 gauss. Variation of ESR and Infra-Red Data with the Various Alkali Halide Matrices The hyperfine splitting constants and the infra-red frequencies are found to depend on the host lattice. The [Cr(CN)5NO]-3ASN values and the NO stretching frequencies for 99 the complex in KBr and KCl matrices areteonsistent with the NO group being forced outward(i.e., the SEN-distance increased) as the smaller lattice forces the CN-CreNO angle from the ex- pected obtuse angle tO‘a more perpendicular one.--However, the AN values and the NaCl nitrosyl stretching frequency are incon- sistent with this model.- The effects of the water molecules of hydration, pellet preparation, and“of*local charge compen— sating defects are other possible'causes for the variations. Jones122 found that for K3Co(CN)6 the.CN vibration frequencies in various lattices had the values given in Table 21. A stronger metal—carbon 0 bond is found to lead to a stronger C-N bond.r Therefore, the strongest metal-carbon 0 bond is expected in NaCl. The K30r(CN)5N®.data cannot be ex- plained by such a simple model. Further experimentation in this laboratory on other cyanides in alkalithalide matrices is planned. TABLE 21.--K3Co(CN)6 vibration frequencies. -Host Lattice NaCl ' KCl H20 Solid K3Co(CN)6 2145 cm.1 2127 cm"1 2127 cm.1 2150 cm-1 21U0 2119 -- 2137 2128 2110 -- 2129 K3Cr(CN)6 Covalent Bonding The general theory for 130 hyperfine splitting in Cr(CN)g3 was presented on pages 16-17. The value-of the overlap integral was estimated by graphical interpolation from the tables of Jaffee and Desk116 and is 83d 2 = 0.065. The internuclear distance used was that of the 6026 lattice lOO K3Co(CN)6.118 A is calculated to be 0.8Ni gauss using C = D 0.75. Using the experimental value'of‘Ap and the calculated value of A , we obtain A - A = A - A = 0.7 — 0.8N2 2:0. 0 w p n D D The possible error in AD is such that“a quantitative estimate of A1r cannot be madea- Qualitatively the results~indicate little w bonding in agreement with the analysis by Jones119 of the infrared spectra. McGarvey53 estimates that Nflfiz 0.89 and that N: 2: 0.68 from the ESR chromium nuclear hyperfine Splitting and g value, respectively.- From the nephelauxetic 55,120 one obtains Nix 0.83 ahle2 ratios C2:20.72. Determination of gy, D, and E The energy level equations for Cr+3 ’along the y mag- netic axis (see Table 9 for the relationships between y and the crystallographic axes) are: 3(D+3E)(E-D)2 3/2 to 1/2 transition g 8H = hv + D + 3E - 2 (67) y 8(hv) 2 3(E-D) (68) 1/2 to —l/2 transition g 8H = hv - —————-— y Hhv 3:H macaw Hoo.o. :u nu mooo.o. oaoo.o. omeoo.o oemoo.o omomoo.o mowoo.o Hugo moao.o . . m Hoo.o. In nu . mooo.o. oaoo.o. mmmo.o oomeo.o mmweo.o Huse ammo.o Hugo ammo.o o Aampmzmo OHQmpmHoa Hoo.o. eooo mocha ozov mooo.o. Hoo.o. a mwm.a mmmm.a mamm.H. icy: ommm.a Hmm.a m «mommm u a Hmfimommm u e mmamommm 1.9 Haamoom u a A ¥. , a QAZOVoomm ea oAzQVaOmx too.m oco;.o . w mo mosao>un.mm mamas 102 bond sharing increases as the volume is reduced and that the Spin-orbit coupling eventually dominates as the“volume becomes smaller. Artman et. al.121 found evidence‘for“polytypism and Raoultlzu found that aged crystals gave extra ESR lines which were not present in new crystals. 53Cr Hyperfine Splittings 53 The in K Cr(CN) . Recently McGarvey 3 6 53Cr 3 Cr A value varied from 15 to 17 gauss 125 observed an anistrOpic A_va1ue in Cr+ acetylacetonate and was able to theoreti- cally account for the anisotropic behavior. 'He found that the six oxygen atoms surrounding the chromium are distorted from the octahedral configuration by compression along the trigonal axis. Assuming an ionic model he calculated that a distortion of 0.60 in the angle between the trigonal axis and the Cr-O bond was sufficient to account for the anisotropic 53Cr_A_ value. A distortion of this small magnitude would not pro- duce a detectable difference in the resulting non—equivalent 130 A values. ESR STUDIES OF COPPER'GLYCINATE The single crystal, solution, and powder data are given in Table 16. The powder data and single crystal data are in agreement. Solution Spectra are not expected to agree since copper glycinate is only soluble in highly polar sol- vents which can coordinate along the z axis and change the hyperfine splitting constant by as much asfi3 “0%. Using equations (11) through (28) with AD = 0.25 gauss one obtains the values of the s and'p*spin density fS and fp, respectively, given in Table 23. The literature data for several other copper complexes are included in Table 23 for comparison purposes. The f; and f; valugswere calculated with the assumption that the a bond is an sp hybrid. If this assumption is correct, f; Should equal fg. .oCHUthQ "kmnm .COHcm poem oo:0ma:moco3H0pmhwq n .oma monopomomo .omH oocopomomo .xhoz hogan .mmsmw ch we .ocfipflhma one mo coapfimomlz on» CH musozpfimeSm map ohm poxomph esp so mm wsfizoaaom maonEmm one ‘11 «41 b7 ooo. Moo..omo. Hmo. mo. mm. N.H e.HH oo.oH oo.oa mflooaooofizoaooso mooo. meo. omo. omo. o.H om. m.H o.mH oe.HH oe.oa mfloeooAmAmmoVZooooo m. oomo. ooo. omo. mmo. oo. :H. o.o m.mH oe.aa oa.oa mAmeoeAmmoaoooo 1 ooeo. moo. moo. Hmo. o.o mm. H.H o.HH om.oa oe.ma mfloeosamoooso memo. oeo. moo. oomo. H.H Hm. o.H o.mH oH.mH oo.oH mflmozvofimoovso ooH. ooo. Foo. oomo. 3H.H mm. so.a e.ma om.oa oo.oa oeaooaooaooooo so zoo. zoo. omo. oefio. mm. mm. m.H o.o oo.o oo.oa ooooaoaam so we we oo oo mo mo mo mo mo M$ .moonQEoo hoaooo,&0h-mowvwmcop swam pcwwfifi mo somHLmQEOOII.mm mqmde 10” SOLUTION SPECTRA OF SUBSTITUTED COPPER ACETYLACETONATES For the spectrum of a copper (II) complex taken in solution the isotropic nuclear hyperfine Splitting A_has been related to the covalency of the unpaired electron by the equation: A = AO(-N§KO + g-2.0023 + smaller terms) (70) where N3 is the fractiggagf time the unpaired electron is on the copper atom, A0 = T’ and K0 is the Fermi contact term due~to spin unpairing in the copper s orbitals. In this equation K0 is assumed constant but if this assumption is not Justified, the value of A_wi11 also be influenced by any change in position of the c0pper d electron energy levels such that the amount of unpaired s electron character at the nucleus changes. For copper complexes K0 is usually considered to be constant and the above equation is considered to be ap— plicable.l‘28’l‘29’130 The ESR A values for the substituted copper acetylace- tonates decrease with change of substituent in the same man— ner that the hydroxyl proton NMR chemical shift decreases for the corresponding enol tautomar,131 with the exception of the phenyl and thenoyl substituents (see Table 2“). Here ring current effects might be expected to alter the NMR values. Both sets of data would give the order of electronegativity CF3>C1>H>CH3. and thenoyl groups with the Cu-O bond is not important the ESR data would give these groups an electronegativity greater than that of methyl, in agreement with the conclusions from 132 Ifit is assumed that conjugation of the phenyl NMR studies. However, Nakamoto et. al. reported on the basis of some infrared data that phenyl substitution increased the Cu-O bond strength relative to methyl and suggested that release of electrons to oxygen by phenyl would increase the Cu—O n—bonding. They also cited some stability constant data133 .Hma cocopowom 105 a .omax woo. .wH canoe mom .ox pcwpmcoo m QpHB Aoov coopmsvo Song pocfiahopoon .o.=mo:=mouoo.m ma oooooooaaoooo ooosoaooooo ooeo ooo. oe.m ooe me.a mas oeo.o+ ooa.m o.a. e.sm moo o moo mom. oo.o ooo -- .. ooa.o- Hma.m e.o. o.ae aaoeoee o moo moa. om.o ego -- one mea.o- oma.m e.o. o.me moo o moo .oee. mo.o omoH Hm.m moo oem.ou omfi.m o.H. m.oe Haooeo m Hacooo 5N5. w©.m owm ll mm: :wm.ol :NH.N m.0u :.©b mmo m Hhflmnm II II mmm II II II II ow mmo HO mmo ope. me.o :mo om.m H-Eo mos eos.o- mma.m o.o. m.ee moo o moo so». .. ooo -- .. ooo.o- Hma.m m.o. o.oe moo moo moo D 2 *pQMchoo pmahm oomph mocosv. Hoop ooowao aoHHHoEWSSEooo sooo mmwao moo mmmcouoo ooao> Hugo.wwoa .=m em .m -Hooo,umxwoooa ooooao ooaom neoooaom o>o2 w ooao> «.mmm \. mooosoaooosm loo OZ lmflo SO mzz OISU OISU IMHmm , .m .mopwcouoomazpoom pooooo pom mpmp honpo new mpcmpmcoo mooooaaoo oofiohooao mmmnu.om mqmoe 106. to support this order.of electronegativities but these appear inconclusive since some of the data for diphenylacetylace- 13H tonates could not be repeated. Also steric interactions between phenyl and the a-hydrogen would be expected to reduce the importance of "-bonding. The N: values given in Table 24 indicate that the more electronegative the substituent the~greater~the~covalency This conclusion is not iggported by'eithcr the polarographic or the correlation between the stability constants and the basicity of the pure ligands}37 132 reduction investigation The polarographic reduction data were interpreted to indi- cate that a more negative value of the half-wave potential is evidence for a more covalent bond.; This interpretation was supported by the similarity between the half-wave poten— tials for cOpper hexafluoroacctylacetonate and copper nitrate. The correlation between the stability constants and the basicity of the pure ligands-indicates that electron with- drawing substituents tend to withdraw the bonding oxygen electrons and so decrease the amount of ligand—to—metal elec- tron donation. It therefore appearsathat equation (70) cannot be used with a constant KO to calculate N§.ashas.been«previous1y assumed. This conclusion is.not completely startling as car- licr papers on the ESR of copper complexes have indicated that, although the g values were following the expected trend (a g value nearer 2.0023 qualitatively indicates a greater covalency) tfluaA_values seemed to be increasing-with greater covalency rather than decreasing as expeCted.l38’129’139 The consistent trend in the substituted acetylacetonate A values can be explained by considering that K consists of two contribu- 63,66,26_ ° tions: a negative contribution which arises from the exchange polarization of the electron spin in filled s-orbitals and a positive contribution which arises from unpaired s elec— tron density in the HS copper orbital.‘ Since the experimental ' A value sis negative, an increase in IA]' can be ex- plained by an increase in the ‘relative importance 107 of the negative exchange polarization term over the positive us population contribution. There is strong evidence61 for +2 Mn the exchange polarization contribution is dependent upon the , in which the As contribution appears negligible, that covalency of the complex. The MS contribution should have a dependence on both the covalency and.the separation between the ground state and the excited AS level. Because of this dependency of the Us term”on an additional parameter besides covalency it is possible for the relative contribution of the two terms to change, as appears to have occured in this series, such that the absolute magnitude of the A value actually in- creases with covalency. A surprising correlation can be made between solvent effects and substituent effects. A plot of g versus A values is given in Figure 31 for a series of substituted copper acetylacetonates in chloroform solution (circles) and the relationship is seen to be nearly linear. The points for copper acetylacetonate in-various solvents (triangles) appear to obey the same relationship and‘it therefore appears that increasing the basicity of the solvent has the same effect on the d—electron energy levels and on the unpaired electron density on copper as does increasing the electronegativity of the substituents on the ligand.- The similarity in ESR parameters for COpper hexafluoroacetylacetonate in chloroform solution and for copper acetylacetonate in pyridine‘ solution is also consistent with the similarity between their low- energy optical absorption Spectra.lLlO’lul A Further evidence that reliableNi values cannot be obtained from equation (70) when K6 is constant comes from an ESR study of nitrogen and copper splittings in c0pper com- plexes in which there are c0pper-nitrogen bonds. Thus we may compare copper phthalocyanine and copper glycinate. The former is considered to be highly covalent while the latter Should be reasonably ionic. In addition the copper—nitrogen 1263142 The ESR data and bond lengths are given in Table.25. A smaller cop- bond lengths for both have been reported. per A value, interpretedeitha~constantKO in equation (70) 108 .monocouoomahuuomAHHvuuamoo no muoommo ucvauwuunnu can uco>aom .Hm onswwm AHn80V< oo me on oo oo oo oo _ _ _ _ _ «H.« o 0 \mV 3 mo o no uoH.~ «no osam> w o o o .mw Home no oo o no \o -3H.N, one no moo x omo \u oouoaoso ca one x moo \H o (o o moo mo mo o MAN e m e . 3 moo a one o 1085 .1 Jo a moo m moo a roa.~ m m m o g a. - 109 .mma oocohomoh omam com .mma oocohomoh .w coapmzvo anm popMHSOHmom mono moa oo Hmo.m ooo zoo.H :Ioa x m o x . :IOH H w :IQH x m w :IOH x m.o « 0.0H EU on . n. a: onoH o H Hod moo.o a mmo.m moo.o u wwm.m ooo.o oo mo.H x .H :IoH H :IOH x ©.mH x . :IQH m :H a . ouoa o ea x ouoa oH Huse onoa x mom omo.m moa.m b m2 oocmpmfim ZISQ m z< oHQOLpomH z< a. z< opmcfiozfim so ocficmmooaospnm so oma .Mpwc mmm opocflozaw AHHV hooqoo cam ocflcmzooamhuhd AHHV Lodaoo eo COmflthEooll.mm mqm . . a a a m m N omao ooH ooo ooaooo mm ooaoos oH ooaoom ea m mooo mo oh Aoooooo> ooam.ooa mas. *ooo.mm *ooo.oH *oom.oa o zommo ea onooaoo> o o Q n m mom sod oofi oooo mm ooom ea oooo.mfi m m: Amozoo> moo.:oa moa. *ooa.om *ooo.sa *ooo.mH m ago as maoooooo> ooam.moa on. sauso ooo.om *HnEo omo.sa *H15o ooo.mH a mommo so onooooo> magma o.os om.a I: n: In a: moOmm ca mumfizooo> < a mm + ax mm Imx + ax Na.nx + ax mm.wwm ocsoofioo cooufimcmpe HmOHon no pcoecwwmm< CH .02 .moxOHQEoo Hzomcm>.now mosao> a cam .mwcfipufiaam MCHMhoamn .mao>oa mwhocmll.wm mqm<9 11H .HoH .oom .flomm. .mo mom .o m m HOCHQ mAzzzv CH mmm mooH .wom .0 mm. HOHB M CH ESLpoon HGOHQQO .Hm oocohomom * m .am as mmm .HoH .oom .oofiaoo too so some Hooaoooo m m m n mmo CH Euppooom HQOfipoom .mzo to mom ”so .oom .zo oHH n: oom.oH A omm.ma V I: oommm oosaao ea oomo> o.oHH ooo.u . .oom.oa ooH.oH oom.ma ma omao> o.oHH mom.u ooo.oH oom.ma ooo.ma ma moazo ca oomo> o.mHH oom.u oos.om oom.oH .ooo.mH HH , momaomo ea onooomaomaooo> o.oHH moa.n oom.mm ooo.oa oos.oa oa mooooooom so on mom.ooa omo.u *oos.om *ooo.oH *oom.oH o omoo oHvo .:.ooa ooo.n ooo.om oom.oa oom.ma o omaoo> o.ooH ooo. ooo.mm ooo.oa ooo.ma s moomzo ca mflouoaoxooo> < a mm + %x mzlmx + ax stnx + ax mm.me pszoqeoo coapfimcohe HmOHon mo pcoecwfimm< CH .02 floozcflpcODv-1.Nm maooe D 115 0.25 0.20” 0.15 0.10” 0.05 0.00- -0005 -0010“, -0. 20¢» .0e 25 -o. 30, -0.35 “Ce ”0+ ’0.“5 b The numbers correspond to the numbered complexes in Table 27. 12‘) 102 Figure 32. L I 104 106 108 110 ESR A Value qt 112 db 11k 116 Correlation of ratio of axial to equatorial crystal field,p, and ESR'A_ values for a number of vanadyl complexes. 116 TABLE 28.-—C and n electrostatic interactions for VYX5 Compounds Compound Cx "x Cy ny VO(NCS)'5'3 29,u8o cm"l 12,280 cm‘1 u3,580 cm-l 66,280 om“l VOF§3 22,920 8,820 35,220 62,020 VOSOu 23,540a 7,5AOa 107,5u0a 59,5140a aUsing Optical assignment of reference lUS. The correlation cannot be extendedeto include com- plexes of lower symmetry than C It appears that as in the copper (II) case, there are twogzontributions to the hyper— fine Splitting, a negative contribution from the polarization of filled E orbitals and a positive contribution due to some Us character. In order for~the-Hs contribution to be appre- ciable the symmetry must be lower than C n~since the only 2v vanadium A orbital which can admix with the As in*C or Cuv is the d22 orbital.65 This theory is consistenivwith the available solution data.‘ For vanadyl oxalate in H20, with no excess ligand present, a second set of ESR peaks slowly grows with an A value of 90 gauss while-the original peaks have an A value of 106.6 gauss.~ Since there is nothing in the solution which can form a more covalent bond than the oxalate, a higher covalency cannot be used-as the-explanation for the lower A values -AS a-check of this interpretation the ESR spectra-of several known low-symmetry cOmplexeslu? were obtained. ~Thc A values-for the tartrate, lactate, malate, and mandelate vanadyl complexes in basic solutiOn Were all within the.range of 88 to-90 gauss. Also, vanadyl glycinate (in which the glycinate is a chelate-ligand with one vanadium- nitrogen-and one vanadiumaoxygen bond) has an ESR A value of 99.2 1'1 gauss.93 This explanation of a positive As-contribution and a negative polarization contribution has-been used-to account 117 for the unusually low ESR A value, 27.H gauss, found for " 1u8 V(CSH5)2 (bis-cyclopentadienyl vanadium) in benzene. The largely ionic complex V(NHA)2(SOA)2 - 6H20 has an A value of 9A gauss.lu9 The V(CSH5)2 case is particularly interesting because proton nuclear magnetic resonance Spectra have shown that the electron spin density on each carbon atom is 0.009. This gives a total.Spin density on the ligands of 0.9. Since the total spin density in the complex is 3 elec- trons the maximum value of the covalency is 30%. Taking into account the large metal-ligand orbital overlap would decrease this quantity so that it appears that this compound does not have an abnormally high covalency as would be predicted from the small value of the vanadium ESR A value.’ In Table 29 the publishedESR data for [VS6C6(CN)6]- are compared with the data for V0(CN)g3. IThe g values are 2 the same within experimental error, and A" is greater than A, for both compounds. On the-basis of the low A value, gL> g , II and A" > AL, and with the assumption of trigonal symmetry, Davison et. a1.65 suggested that the unpaired.electron is in a molecular orbital which is mainly ligand in character since for trigonal symmetry one predicts that AL> A” and gn> gL. The similarity of their data to the data for V0(CN)'5'3 suggests that their symmetry is actually distorted tetragonal. The low A value can then be explained by the admixture of Us character rather than an abnormal amount of covalency. ESR SPECTRA 0F VANADYL COMPLEXES The ESR data for a series of vanadyl complexes, inclu— ding the anisotropic hyperfine splitting constants, are shown in Table 13. These values can be compared with the previously published ESR data-collected in Table 30. A plot of the gH values versus-the A” values is given in Figure 33; within the average experimental error of 10.005 for the g” values, there appears to be a linear correlation between the A and gH values N for complexes with tetragonal symmetry.- With this assumption, 118 TABLE 29.--Comparison of the ESR data for V0(CN)g3 and [VS6C6(CN)6]-2 g gH gL A Al! A1 VS6C6(CN)6"2 1.980 1.97A (1.983) " 63.3 ' 100 A5 3 ‘ ‘ gauss V0(CN);. 1.983, 1.972 1.98A 79.9 1A9 51 the A” values could.be-used to estimate-the gH values for addi- tional complexes-and~possib1y give a more accurate value of g” for the complexes in Table 13. This correlation is surprising since the principle dependence of A” is on N32 (see equation (35))while the principle dependence of gH is on N2 N2 1T2 02 -—————— (see equation (29)) Ab+ b 2 1 The overlap integrals for some vanadyl complexes are given in Table 31. The relative magnitudes of these overlap integrals Should give an approximate indication of the relative magnitudes of the covalency in the bonds involved. SOLUTION SPECTRA OF COPPER (II) AMINO ACID COMPLEXES The isotropic ESR A values and some literature data for comparison are given for a series of copper (11) amino acid complexes in Table 32. Using the results of the substi- tuted c0pper acetylacetonate series as a model, one would pre- dict that the more basic the amino acid, the larger the ESR A value of the corresponding complex, the more stable the com- plex, and the greater the covalency. There does appear to be a general overall correlation between the basicity of the ligands as measured by the ligand proton dissociation constants and the A values but the correlation does not hold for all of the TABLE 30.--Literature ESR A values for VO+2 119 AA complexes. Compound A(in gauss) A”(in gauss) A; (in gauss) VO(porphyrin)72 96 to 98 173 to 176 56 to 60 V0 2 150 (acetylacetonateg2’ 102 to 108 181 to 191 6A to 69 VO(H20)5+2 116, 118 —- —- 120, 116.6 -_ __ v0(0204)‘2 112 (103) 180 t 65 . 2 VO+u in amorphous 0e0278 (11u.2) 19u.8 10.1 73.9 . 0.1 +u V0 7%“ tetragonal —— 1A6.66 u0.90 : 0.01 GeO2 -- i0.02 (x axis) and A1.85A 10.01 (y axis) v0+u in IR 100138 —— 200 81 Dowex 50 -- 210 80 Charcoal -— 190 76 IR NB —- 175 66 V0“4 in 153 Zn(NHu)2(SOu)2 —- 202.54 71.15(X axis) __ -_ 78.A9(y axis) -- 202.63 76.81(x axis) __ -_ 78.A9(y axis) v0F2152 116 t 3 __ _- 152 VOCl -- 202 i 10 76 i 5 VOC15-3 151 (109.u) 173.0 77.6 190 120 lemt 170+. 160-_ H votiougln Hc10u V0(DMSO)5(ClOu)2 in DMSO VOSOu in DMSO V0(AcAc)2 in DuPont glue VO(hexaFAcAc)2 in DMF V0“? CFaAcAc ) 2 in DMF VO(CF3AcAc)2 in DMF v0(NCS)5'3 in CHCl -3 3 V0(NCS)S in DMF omqmmsww 0 co . ‘| -1 in cm 150 L luod 130 10 11 V0(AcAc)2 in CHCl v0(CN)5'31n KBr 3 1.93 1.9“ 1.95 1.96 1.97 Figure 33. Plot of A" versus 3" for a number of vanadyl complexes. 121 TABLE 31.--Group overlap integrals for some vanadyl complexes. +2 -3 -3 VO(H2O)5 VO(NCS)5 VO(ACAC)2 VO(CN)5 V = 0 Distance (A0) 1.67a 1.62a 1.59a 1.75b V-ligand Distance (A0) 2.3a 2.0Ma 1.97a ::1.91 302 0.168 0.22 0.2u9 0.28u 8,2 0.056 0.132 0.12M 0.26 "l 0.12“ 0.139 0.198 0.105 aSee Ref. 15M. bEstimated from infrared data, Ref. 97. compounds in the series. There appears to be no-correlation between the stability constants and the_A values; however, there is a correlation between the A values and the separation betweenSShe COO stretching frequencies. According to Nakamoto's model, increased covalent character leads to a more asym- metrical carboxyl group and hence to an increase in the fre- quency separation of the two bands. From the available data it appears that thelfifil A_va1ue increases with increasing covalency, but that the stability of the complexes is dependent upon an additional factor which does not proportionally affect theififll A value. Irving and Pettit155 noted the discrepancy between the basicity of the ligands and the stability constants of the complexes and suggested that the discrepancy is probably due to steric effects. This explanation is consistent with the ESR behavior of the bipyridyl complex, see Table 18, which does not follow the smooth correlation which is.obtained for the substituted acetylacetonates. TABLE 32.-ESR data and literature data for COpper amino acid complexes. Kouenb -eag Butqoqeaqs ueBoaqtN-IBQGN gsatouenb -8Jfi Butqoqeaqg 000 991 USGMQ -eq uotqeaedeg quads -uoo eoaog N-no guaxtx 301 ggIQU’BQSUOO [101:3 -etooss:q uoqoad uaBoaqtN pueBIq Iqueqsuoo uotq -etoosstq uoqoaa {Axoqaeo pueB:q g I ( a"II/IX Bot queqs -u00 thttqeqs 8019A '17 ass (I: N I :I.‘ m-o-z \ :3 I O 03-0-0" 122 m) U\ H H I I I E I o 3' O\ «1 I I I I I I a) ux H 0 mx I I . I I 0 r1 . I I U\ I I Hi N \o I I - l l O\ m N) I I . I I m :r I oo \o a) m) cm H .H H b— b— >- W) W) m m ? o -(i)~:: I W) m WI m m m o «—O L) a“ m I: C) 14.6 71.0 2.21 9.15 CH2OH H 9.62 2.26 70.0 CH - - H H TABLE 32.-—(Continued). Aouenb -eag Butqoqeaqg. UBSOJQTN-IBQGWi Bsetouenb -eag Butqoqeaqs 000 991 USGMQ -eq uotqeaedeg‘ queqs —uoo eoaog N-no guaxr‘x 301 ggtqueqsuoo uotq -etooss:q uoqoaa ueBoaqtN pueBIq ggtqueqsuoo uotq -etoosstq uoqoaa {Axoqaeo pueBIq Z 9§I( TWX Bot) queqs -u00 Katttqens enteA'? ass 220 6.78 2.3“ 9.69 15.1 69.6 CH3 2.3M 9.60 15.u 0.62157 206 6.91 68.8 161 9.13 10.19 1.83 3.60 68. 51.0 0 92 I H H N- \ 2| 2 O / --CH2 ¢sN/C_u\o I 2/ H-C-C=O H2C In hydrated crystal.159 bWhen nitrogen are cis to each other.157 a b When nitrogens are trans.157 1211 For comparison the solution EESR A_values and sta- bility constants for some copper (II) amine complexes are given in Table 33. For the bidentate ligands a higher A value corresponds to a higher stability constant. 125 TABLE 33.--ESR and stability constants for copper (II) amine complexes. Ligand in Water Solution A (in gauss) Log K1K2c H2NCH2CH2NH2 b Ethylenediamine 89 90 19.72 H2NCH2CH2CH2NH2 b Trimethylenediamine -- 87 16.9 H2NCH2CH20H Ethanolamine 87,5a 87 __ HN(CH2CH2OH)2 Diethanolamine _- 83 __ HN2CHZCH2HH CH CH l HN2CH2CH2NH Triethylenetetramine 82a —- 20.5 2 2 NH 3 ' 81a 8b 12 (K K K K ) Ammonia 7 '7 1 2 3 A N(CHZCH2OH)3 Triethanolamine 77 7A -— H 2N- -CH2—CH2 N(CH 2CH3) 2 N,N- -Diethylenediamine 77.5 -- 13.7 (CH3)2CH2 NH2 ISOpropylamine -- 61 -- aThis work. bReference 139. CReference l36, page 286 and 518-523. SUMMARY The electron spin resonance spectra of a number of transition metal complexes have been obtained. The interpre- tation of the spectra has been shown to be more complex than expected. Effects which previously were ascribed to covalency may also have appreciable contributions from excited states so that it now appears that additional information about the sym— metry, energy levels, and effects of spin polarization are required in order to accurately describe the covalency of a transition metal complex from its electron Spin resonance spectrum. 126 10. 11. 12. 13. AcAc— CF3AcAc- D- APPENDIX I GLOSSARY Isotropic hyperfine coupling coeffi- cient, page 19, 20, 25, 26. Ligand hyperfine coupling constant due to direct dipole interaction from the electrons on the metal ion, page 20,21. Theoretical anisotropic nuclear hyper— fine splitting value, page 25, 80. Theoretical nuclear hyperfine splitting value for an electron which spent 100% of its time in the orbital indicated by the subscript, page 21. AnisotrOpic hyperfine coupling coeffi- cient along the chosen axial symmetry axis, page l8, 19, 2A, 25, 26. Anisotropic hyperfine coupling coeffi- cient perpendicular to the chosen axial symmetry axis, page 18, 19, 2A, 25, 26. Acetylacetonate, page 65. Trifluoroacetylacetonate, page 65, 73. Zero-field splitting coefficient along the chosen symmetry axis, page 102. Diphenylacetylacetonate- DMF— DMSO- Ds,Dq,Dt— Dibenzoylmethane, page 89. Dimethylformamide, page 38, 63, 65, 73. Dimethylsulphoxide, page 6“, 66, 73. Crystal field splitting parameters, page 28. 127 14. 15. l6. 17. 18. 19. 20. 21. 22. 23. 2A. 25. 26. 27. 28. 29. HexaFAcAc,CF K _ o m— 128 Zero-field splitting coefficient per— pendicular to the chosen symmetry axis, page 100. Transition metal d orbitals of correct symmetry to form sigma molecular orbi- tals in an octahedral complex. Electron spin density in the 0, n, and s orbital respectively, page 21, 26, 98. Spectroscopic splitting factor, page 6. Spectrosc0pic splitting factor parallel to chosen axial symmetry axis. page 22, 23, 2A. Spectrosc0pic splitting factor perpen- dicular to chosen axial symmetry axis, 22, 23, 2A. CF AcAc- Hexafluoroacetylacetonate, page 65, 73. Nuclear magnetic moment, page 24, 25. Fermi contact term contribution to the nuclear hyperfine splitting constant, page 105. Component of nuclear magnetic moment, 1, page 2A, 25. Molecular orbital coefficient of metal atomic d orbital, page 17, 19, 20, 21, 22. Pyridine, page 105. Metal 3d orbital—ligand x orbital atomic overlap integral, page 20. Transition metal d orbitals of correct symmetry to form 0 molecular orbitals in an octahedral complex. Paratoluenesulfonic acid anion, page 103. Nuclear magneton, page 20. 30. 33. 3A. 35. s CF3AcAc- E3 - 129 Molecular orbital coefficient which indicates covalent character in excess of that caused by metal—ligand overlap, page 20.‘ Spin—orbit coupling constant, page 6, 2“. Molecular orbital coefficient of the ligand orbital which includes both co- valency and overlap, page 20. Magnetic moment of the nucleus, page 20. Klystron frequency, page 2A, 61. Ratio of axial crystal field to equa- torial field, page 30. Thenoyltrifluoroacetonate, page 66, 73. Thenoyl group, page 63. 10. 11. 12. 13. 14. 15. REFERENCES U) Sugano, and R. G. Shulman, Phys. Rev.‘ 130, 517 (1963). T. P. P. Hall, W. Hayes, R. W. H. Stevenson, and J. Wilkens, J Chem. Phys.. 8. 1977 (1963); 32, 35 (1963). R. Lacroix and G. Emch, Helv. Phys. Acta. 35, 592 (1962). C. J. Ballhausen, "Introduction to Ligand Field Theory," McGraw-Hill, New York, 1962. L. E. Orgel, "An Introduction to Transition-Metal Chemistry, Wiley, New York, 1960. J. S. Griffith, "The Theory of Transition-Metal Ions," University Press, Cambridge, England, 1961. C. K. Jorgensen, "Absorption Spectra and Chemical Bonding in Complexes," Pergamon Press, New York, 1962. C. K. Jorgensen, "Inorganic Complexes," Academic Press, New York, 1963. C. K. Jorgensen, "Progress in Inorganic Chemistry," 3, Interscience, New York, 1962. C. K. Jorgensen, "Advances in Chemical Physics," V, Inter— science, New York, 1963. C. K. Jorgensen, "Solid State Physics," 13, Academic Press, New York, 1962. F. A. Cotton, "Chemical Applications of Group Theory," Interscience, New York, 1963. K. Nakamoto, "Infrared Spectra of Inorganic and Coordina- tion Compounds," Wiley, New York, 1963. R. J. Faber, Ph.D. Thesis, Michigan State University, 1957. Title: "Paramagnetic Resonance of Some Adsorbed Transi- tion Metal Ions." K. D. Vos, Ph.D. Thesis, Michigan State University, 1962. Title: "Electron Paramagnetic Resonance Absorption Studies on Solutions of Metals in Ammonia and Amines." 130 l6. 17. 18. 19. 20. 21. 22. 23. 2A. 25. 26. 27. 28. 29. 30. 31. 32. 33. 3A. 131 S. K. Bolte, Ph.D. Thesis, Michigan State University, 1963. Title: "A Study of the Free Radicals in Irradiated Amides by Electron Spin Resonance Spectroscopy." J. E. Wreede, M. S. Thesis, Michigan State University, 1963. Title: "Hyperfine Splitting in Electron Paramag- netic Resonance Studies of Metal-Amine Solutions." G. E. Pake, "Paramagnetic Resonance," Benjamin, New York, 1962. W. Low, "Solid State Physics," Suppl. 2, Academic Press, New York, 1960. C. P. Slichter, "Principles of Magnetic Resonance," Harper and Row, New York, 1963. S. A. Al'tshuler and B. M. Kozyrev, "Electron Paramagnetic Resonance," Academic Press, New York, 196“. R. S. Anderson, "Methods of Experimental Physics," 3, Aca- demic Press, New York, 1962. K. W. H. Stevens, "Magnetism," 1, Academic Press, New York, 1963. A. Carrington and H. C. Longuett-Higgins, Quart. Rev. 13, A27 (1960). H. S. Jarrett, "Solid State Physics," 13, Academic Press, New York, 1963. R. E. Robertson, "Determination of Organic Structures by Physical Methods," 2, Academic Press, New York, 1962. S. I. Weissman, Ann. Rev. Phys. Chem.. 12, 151 (1961) R. G. Shulman, Ann. Rev. Phys. Chem. 13, 325 (1962). M. C. R. Symons, Ann. Repts. Progr. Chem., 59, A5 (1962). W. Low, Ed., "Paramagnetic Resonance," 1, 2, Academic Press, New York, 1963. Colloque Ampere, 9, University of Geneva, Switzerland, 1960. Colloque Ampere, 19, North-Holland, Amsterdam, 1961. Colloque Ampere, 11, North-Holland, Amsterdam, 1962. R. Servant and A. Charru, Ed. "Electronic Magnetic Reso- nance and Solid Dielectrics," North-Holland, Amsterdam, 196A. 35. 36. 37. 38. 39. A0. A1. A2. A3. AA. A5. A6. A7. u8. A9. 50. 51. 52. 53. 5A. 55. 56. 57. 132 J. Phys. Soc. Japan. 31, Suppl. B-l, 1962. J. Appl. Phys. 33, Suppl. 3, 1962. R. D. Feltham, Ph.D. Thesis, U. of Calif., Berkfley,UCRL 3867, 1957. J. E. Drumheller, Ph.D. Thesis, U. of Colorado, 1962. R. Neiman, Ph.D. Thesis, U. of Calif., Los Angeles, 1960. A. V. Guzzo, Ph.D. Thesis, Washington University,St. Louis, 1960. L. Helmholz and A. V. Guzzo, J. Chem. Phys. 33, 302 (1960). L. Helmholz, A. V. Guzzo, and R. N. Sanders, J. Chem. Phys. 35, 13A9 (1961). g R. Kyi, Ph.D. Thesis, U. of Calif., Berkfley,UCRL 9109, 1960. J. D. Axe, Ph.D. Thesis, U. of Calif., Berkfiey, UCRL 9293, 1960. R. G. Hayes, Ph.D. Thesis, Univ. of Calif., Berkfley, UCRL 9873, 1961. J. Owen, Proc. Roy. Soc. (London). £_331, 183(1955). T Murao, Progr. Theoret. Phys. (Kyoto) 33, 657 (1959). T. M. Dunn, Trans. Faraday Soc. 31, 1AA1 (1961). W . Marshall and R. Stuart, Phys. Rev. 123, 20A8 (1961). P. W. Anderson, "Solid State Physics," 33, Academic Press, New York, 1963, p. 99. R. G. Shulman and S. Sugano, Phys. Rev. 130, 506, (1963). R. Lacroix and G. Emch, Helv. Phys. Acta 5, 592 (1962). B. R. McGarvey, J. Chem. Phys. 33, 37A3 (196A). 2 H. Kon and N. E. Sharpless, J. Chem. Phys. 906 (1965). H. A. Kuska and M. T. Rogers, J. Chem. Phys. 33, 3802 (196A). A. Abragam and M. H. L. Pryce, Proc. Roy. Soc. (London A 205, 135 (1951). A. J. Freeman and R. E. Watson, Phys. Rev. 123, 2027 (1961). 58. 59. 60. 61. 62. 63. 6A. 65. 66. 67. 68. 69. 70. 71. 72. 73. 7A. 75. 76. 77. 78. 79. 80. 133 J. H. Wood and G. w. Pratt, Jr.,Phys. Rev. 107, 995 (1957). v. Heine, Phys. Rev. 107, 1002 (1957). On.Matamura. J. Phys. Soc. Japan 39, 108 (1959)- R. S. Title, Phys. Rev. 131, 623 (1963). B. Welber, Phys. Rev. 138, A1A81 (1965). E. Kdnig, Z. Naturforsch. 195, 1139 (196A). W. J. Childs, L. S. Goodman, and D. von Ehrenstein, Phys. Rev. 132, 2128 (1963). A. Davison, N. Edelstein, R. H. Holm, and A. H. Maki, J. Am. Chem. Soc. §§, 2799 (196A). R. E. Robertson and H. M. McConnell, J. Phys. Chem. 6A, 70 (1960). —_ N. J. Trappeniers and S. H. Hagen, Physica, 33, 122 (1965) and 3;, 251 (1965). A. Mukherji and T. P. Das, Phys. Rev. 111, 1A79 (1958). A. J. Freeman and R. E. Watson, Bull. Am. Phys. Soc..§, 23A (1961). R. E. Watson and A. J. Freeman, Phys. Rev. 13A, A 1526 (196“). D. E. Ellis, Solid-State and Molecular Theory Group, M.I.T. Quarterly Progress Report No. 55, January 15, 1965, p. 58. D. Kivelson and S. Lee, J. Chem. Phys. 93, 1896 (196A). G. Shulman and S. Sugano, J. Chem. Phys. 93, 39, (1965). Tsang, J. Chem. Phys. 39, 729 (196A). . Marshall, "Paramagnetic Resonance," 1, Academic Press, ew York, 1963, p. 3A7. Simanek and Z. Sroubek, Phys. Stat. Solidi 9, 291 (196A). .‘IEFJZSI-JSU A. Kuska and M. T. Rogers, J. Chem. Phys. 93, 303A (1965). H Siegel, Phys. Rev. 13A, A 193 (196A). S. Korol'Kov, Opt. Spectry. ll, 73 (1961). V. R. A. D. Wentworth and T. S. Piper, Inorg. Chem. A, 709 81. 82. 83. 8A. 85. 86. 87. 88. 89. 90. 91. 92. 93. 9A. 95. 96. 97. 98. 99. 100. 101. 102. 103. 13A J. Selbin and T. R. Ortolano, J. Inorg. Nucl. Chem..g§, 37 (196A). H. Yamatera, Bull. Chem. Soc. Japan, 33, 95 (1958). Varian Associates, Palo Alto, California. L. Buss and L. Bogart, Rev. Sci. Instr. 3;, 20A (1960). Hewlett-Packard Co., Palo Alto, California. Waveline Inc.,Ca1dwe11, New Jersey. Micro-Now Instrument Company, Chicago, 111. National Company, Malden, Mass. Beckman Model 350, Beckman Helipot Division, Fullerton, California. H. A. Kuska and M. T. Rogers, J. Chem. Phys._flg, 910 (196A). Instruction Manual V-A503EPR Spectrometer Systems, Varian Associates, Palo Alto, Calif., p. 5-11. J. B. Spencer and R. J. Myers, J. Am. Chem. Soc.§§,522 (196A). This work. J. L. Burdett, Ph.D. Thesis, Michigan State University, 1963. F. Cruser and E. H. Miller, J. Am. Chem. Soc. 33, 1132 (1906). . W. P. Griffith, J. Lewis, and G. N. Wilkinson, J. Chem. Soc. 1959, 872. L. H. Holmes Jr., Ph.D. Thesis, Louisiana State University, 1961. A. M. Golub and R. A. Kostrova, Russ. J. Inorg. Chem. 3, 3A9 (1960). M. Mori, S. Ueshiba, and S. Kawaguchi, Bull. Chem. Soc. Japan_3§, 796, (1963). J. P. Greenstein and M. Winitz, "Chemistry of the Amino Acids, 3, Wiley, New York, 1961, p. 569. Physics Today, Feb., A9 (196A). 1. Bernal and S. E. Harrison, J. Chem. Phys. 39, 102 (1961). W. P. Griffith, J. Chem. SOC. 1963, 3286. 135 10A. J. J. Fortman and R. G. Hayes, J. Chem. Phys..A;, 15 (1965). 105. R. G. Hayes, J. Chem. Phys. 3g, 2580 (1963). 106. J. B. Spencer and R. J. Myers, J. Am. Chem. Soc. 86, 522 (196A). ' —- 107. B. Elschner and R. Neubert, "Hochfrequenzspektroskopie" Akadamie-Verlag, Berlin, 1961, p. 185. 108. J. Cowen, Physics Department, Michigan State Universtiy, Personal Communication. 109. P. T. Manoharan and W. C. Hamilton, Inorg.Chem. 2, 10A3 (1963). ‘ — 110. P. Groth, "Chemische Kristallographie," I, Verlag von Wilhelm Engelmann, Leipzig, 1906. p. A18. 111. J. M. Baker, B. Bleaney, and K. D. Bowers, Proc. Phys. Soc. (London) B69, 1205 (1956). 112. R. M. Golding, M01. Phys. 5, 369 (1962). 113. E. o. Schulz-Du Bois, Bell System Tech. J. ;§, 271 (1959). 11A. H. Abe and H. Shirai, J. Phys. Soc. Japan 15, 1711 (1960). 115. J. R. Morton, Chem. Rev. 63, A53 (196A). 116. H. H. Jaffe and G. o. Doak, J. Chem. Phys. 21, 196, 258 (1953). —_ 117. D. R. Hartree, "The Calculation of Atomic Structures," Wiley, New York, 1957, p. 67, 118. N. A. Curry and W. A. Runciman, Acta Cryst. 12, 67A (1959). 119. L. H. Jones, Inorg. Chem. 2, 777 (1963). 120. R. Krishnamurthy and W. B. Schaap, Inorg. Chem. 2, (1963). 121. J. 0. Artman, J. C. Murphy, J. A. Kohn, and W. D. Townes, Phys. Rev. Letters. A, 607 (1960). 122. L. H. Jones, J. Chem. Phys. 16, 1A00 (1962). 123. W. M. Walsh, PhyS. Rev. 11A, 1A85 (1959). 12A. G. Raoult, J. C. Parouty, M. Lacombat, and A. M. Duclaux, "Colloque Ampere," 10, North-Holland, Amsterdam, 1961, p. 165. "’ 125. B. R. McGarvey, J. Chem. Phys. 59, 809 (196A). 126. 127. 128. 129. 130. 131. 132. 133. 13A. 135. 136. 137. 138. 139. 1A0. 1A1. 1A2. 1A3. 136 S. E. Harrison and J. M. Assour, J. Chem. Phys. A0, 365 (196A). 3 A. V. Zelewsky and W. Schneider, "Proc. 8th International Conference on Coordination Chemistry," Springer-Verlag, Wien, 196A, p. A3. H. R. Gersmann and J. D. Swalen, J. Chem. Phys..36, 3221 (1962). ' D. Kivelson and R. Neiman, J. Chem. Phys. 35, 1A9 (1961). A. H. Maki and B. R. McGarvey, J. Chem. Phys. 29, 31 (1958). J. L. Burdett and M. T. Rogers, J. Am. Chem. Soc. 96, 2105 (196A). ' K. Nakamoto, Y. Morimoto, and A. E. Martell, J. Phys. Chem. 66, 3A6 (1962). ‘ L. G. Van Uitert, W. C. Fernelius and B. E. Douglas, J. Am. Chem. Soc._15, A57, 2736 (1953). G. S. Hammond, A. W. Moschel, and W. G. Bordium, J. Phys. Chem. 6A, 1782 (1960) H. F. Holzclaw, Jr., A. J. Carlson, and J. P. Collman, J. Am. Chem. Soc. 18, 1838 (1956). A. E. Martell and M. Calvin, "Chemistry of the Metal Chelate Compounds," Prentice-Hall Inc., New York, 1952, Appendix 1. R. L. Belford, A. E. Martell, and M. Calvin, J. Inorg. and Nucl. Chem. 2, 11 (1956). R. J. Faber and M. T. Rogers, J. Am. Chem..Soc. 8;, 18A9 N. N. Tithomirova and K. I. Zamaraev, Zhurnal Strukturnoi Khimii. A, 22A (1963). R. L. Belford, M. Calvin, and G. Belford, J. Chem. Phys. 36, 1165 (1957). J. P. Fackler, Jr., F. A. Cotton, and D. W. Barnum, Inorg. Chem. a, 97 (1963). H. C. Freeman, M. R. Snow, I. Nitta, and K. Tomita, Acta Cryst. ll, 1A63 (196A). K. Nakamoto, Y. Morimoto, and A. E. Martell, J. Am. Chem. Soc. 83, A533 (1961). 1AA. 1A5. 1A6. 1A7. 1A8. 1A9. 150. 151. 152. 153. 15A. 155. 156. 157. 158. 159. 160. 161. 162. 137 S. Ahrland and B. Noren, Acta Chem. Scand. A2, 1595 (1958). C. J. Ballhausen and H. B. Gray, Inorg. Chem..1, 111 (1962). "— "Stability Constants of Metal—Ion Complexes," Publication number 17, Chemical Society, London, Burlington House, London, 196A. J. Selbin and L. Morpurgo, J. Inorg. Nuc1.Chem. 27, 673 (1965). ‘__ H. M. McConnell, W. W. Porterfield, and R. E. Robertson, J. Chem. Phys. 30, AA2 (1959). C. Kikuchi, M. S. Sirvetz, and V. W. Cohen, Phys. Rev. 92, 109 (1953). I. Bernal and P. H. Rieger, Inorg. Chem? a, 256 (1963). K. DeArmond, B. B. Garrett, and H. S. Gutowsky, J. Chem. Phys. A2, 1019 (1965). N. S. Garif'yanov and S. E. Kamenev, Soviet Physics, JETP, 19, 3A0 (196A). R. H. Borcherts and C. Kikuchi, J. Chem. Phys. A0, 2270 (196“). _— J. Selbin, Chem. Rev. 65, 153 (1965). H. Irving and L. D. Pettit, J. Chem. Soc.,1963, 15A6. J. P. Greenstein and M. Winitz, "Chemistry of the Amino Acids, A, Wiley, New York, 1961, p. A86. T. J. Lane, J. A. Durkin, and R. J. Hooper, Spectrochim. Acta .gg, 1013 (196A). 1. Nakagawa, R. J. Hooper, J. L. Walter, and T. J. Lane, Spectrochim. Acta AA, (1965). K. Nakamoto, Y. Morimoto, and A. E. Martell, J. Am. Chem. Soc. 83, A528 (1961). R. A. D. Wentworth and T. S. Piper, J. Chem. Phys. Al, 388A (196A). T. R. Ortolano, J. Selbin, and S. P. McGlynn, J. Chem. Phys. 31, 262 (196A). H. B. Gray, P. T. Manoharan, J. Pearlman, and R. F. Riley, Chem. Communications, 1965, 62. 163. 16A. 165. 166. 167. 168. 169. 138 D. A. C. McNeil, J. B. Raynor, and M. C. R. Symons, J. Chem. Soc. 1965, A10. H. B. Gray and C. J. Ballhausen, J. Chem. Phys. (1962). 6, 1151 J A. K. Wiersema and J. J. Windle, J. Phys. Chem. 68, 2316 (196“). ' —_ C. A. Bates, Proc. Phys. Soc.(London) B2, 465 (1964). J. S. Van Wieringen, Disc. Faraday Soc.l9, 118 (1955). C. J. Ballhausen and H. B. Gray, Inorg. Chem. g, A26 (1963). J. Lewis, R. J. Irving, and G. Wilkinson, J. Inorg. Nucl. Chem. 1, 32 (1958). '