THE PREPARATION AND _ f OPTICAL AND ELECTRON SPIN RESONANCE SPECTRA 0F somE HEXACHLORO AND PENTACHLOROALKDXD ‘ ’ ' VANADATES A III) I PARTII : THE INVESTIGATION OF THE APPARENT THERMOCHROMISM" ' OF SOME VANADIUM (III) COMPLEXES I Thesis for the Degree of Ph D MICHIGAN STATE UNIVERSITY ROBERT DEANE BEREMAN ' 195917” fi,ILII£RH4Ill? Michigan SEate _ ‘ University 3. 515 This is to certify that the thesis entitled . PART I. THE PREPARATION AND OPTICAL AND ELECTRON SPIN RESONANCE SPECTRA OF SOME HEXACHLORO AND PENTACHLOROALHOXO VANADATES(IV) PART II. THE INVESTIGATION OF THE-APPARENT THERMOCHROMISM OF SOME VANADIUAMIII) COMPLEXES presented by ROBERT DEANE BEREMAN has been accepted towards fulfillment of the requirements for Ph.D. degreein CHEMISTRY tad. gawfifio. Major professor ' ABSTRACT PART I THE PREPARATION AND OPTICAL AND ELECTRON SPIN RESONANCE SPECTRA OF SOME HEXACHLORO AND PENTACHLOROALKOXO VANADATES(IV) BY Robert Deane Bereman A new class of crystalline vanadium(IV) compounds, the pentachloroalkoxovanadates(IV), has been prepared and characterized. The alkoxo group was methoxo, ethoxo, nf propoxo, or nfbutoxo, while the cation was tetramethyl— ammonium, tetraethylammonium, or pyridinium. The relative stabilities of the complexes are dependent on the cations. The color of all the complexes is golden. Two new salts of the hexachlorovanadate(IV) ion were also prepared. The tetramethylammonium and tetraethylam— monium hexachlorovanadates(IV) were prepared by adding a solution of VCl4 in thionyl chloride to a solution of the tetraalkylammonium chloride in thionYl chloride. The pentachloroalkoxovanadates(IV) were prepared by the addition of one equivalent of the appropriate alcohol to a slurry of the tetraalkylammonium or pyridinium hexachloro- vanadate(IV) in an acetonitrile - ethyl ether mixture. The alkoxo complexes could be converted to the appropriate hexachloro complex by the addition of HCl or thionyl chloride and thus indicated the absence of a vanadyl species. *7 Robert Deane Bereman Magnetic studies indicated Curie-Weiss paramagnetism and verified the presence of the vanadium(IV) ion with 3d1 configuration. The infrared spectra of the complexes were in agreement with the formulation V(OR)C15-2, showing characteristic alkoxide C-O absorptions in the region 1000-1100 cm—l. The ultraviolet—visible spectra of the alkoxides consisted of two peaks around 14,000 cm-1 while the Spectra of the hexachlorovanadates(IV) had an asymmetric peak around 15,500 cm-l. The reflectance spectrum of each of the complexes agreed with the solution spectrum. Attempts to prepare a tetrachlorodialkoxovanadate(IV) species as well as other examples of the pentachloroalkoxo— vanadate(IV) species were unsuccessful. A complete investigation of the electron spin resonance spectra of the pure solids, solutions and frozen solutions (glasses) of the pentachloroalkoxovanadates(IV) and hexa- chlorovanadates(IV) was carried out. Trends in the hyperfine splitting constants as well as in some 9 values were observed in going from the methoxide to the n—butoxide. The observed g and A values were used to calculate the coefficients in a simple molecular orbital scheme. The two molecular orbitals made up of the dxy and dX2_ 2 metal Y orbitals became more covalent in going from the methoxo to the nfbutoxo complex. The molecular orbital made up of the metal dXz or dyz metal orbitals became more ionic. The unpaired electron density in each of the four equatorial Robert Deane Bereman ligand 3p? orbitals was calculated from the molecular orbital coefficients. The density increased along the series (methoxide to nfbutoxide) as was expected. PART II THE INVESTIGATION OF THE APPARENT THERMOCHROMISM OF SOME VANADIUM(III) COMPLEXES Several crystalline octahedral vanadium(III) complexes of the general type [(C2H5)4N]VBr ClX'ZCH3CN where x = 4-x 0—4 were prepared and investigated. The color of the com— plexes changes gradually from yellow for the [(C2H5)4N]VC14°2CH3CN complex to red—brown for [(C2H5)4N]VBI4‘2CH3CN complex. All the complexes have the unusual property of being yellow at 779K. The ultraviolet—visible spectra of the solid complexes were investigated at several temperatures in an attempt to explain the color changes. A large charge transfer band at approximately 20,000 cm"1 for the [(C2H5)4N]VBr4°2CH3CN complex shifts to a shorter wavelenth and narrows slightly to produce the color change. Spectra at 3000K, 1950K, and 770K show the shift is gradual. Magnetic studies indicated temperature dependent para— magnetism and verified the presence of the vanadium(III) ion with 3d2 configuration. No abrupt changes were found in magnetic behavior as the color of the complexes changed. Robert Deane Bereman The C1 nuclear quadrupole resonance spectra of the [(C2H5)4N]VC14°2CH3CN complex gave two peaks at 770K. The position of these peaks shift only slightly when the sample is warmed to room temperature. No Cl nuclear quadrupole resonance signal could be obtained on any of the other complexes. The infrared spectra were also investigated at room temperature and 770K. Only minor differences in the spectra were noted. Most of the peaks due to metal—ligand vibrations as well as those due to the acetonitrile group could be assigned. PART I THE PREPARATION AND OPTICAL AND ELECTRON SPIN RESONANCE SPECTRA OF SOME HEXACHLORO AND PENTACHLOROALKOXO VANADATES(IV) PART II THE INVESTIGATION OF THE APPARENT THERMOCHROMISM OF SOME VANADIUM(III) COMPLEXES BY Robert Deane Bereman A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1969 /0~1 3—70 To Barbara ii ACKNOWLEDGMENT I would like to extend my appreciation to Professor Carl H. Brubaker, Jr. for his interest, patience, and encouragement during this investigation. I am deeply grateful to my wife, Barbara, for her inspiration and unrelenting encouragement. -I wish to thank my parents, Mr. and Mrs. Howard L. Bereman of Lawrence, Indiana for their assistance, guidance, and encouragement during my educational pursuits. Appreciation is also extended to the National Science Foundation, Lubrizol Chemical Company, and Phillips Petro— leum Oil Company for financial assistance during the period 1966-1968. iii TABLE OF CONTENTS PART I INTRODUCTION . . . . . . . . . . . . . . . . . . . HISTORICAL . . . . . . . . . . . . . . . . . . . . A. Vanadium(IV) Chemistry . . . . . . . B. Anionic d1 Transition Element Alkoxides C. The Application of Electron Spin Resonance to Transition Metal Complexes . . . . . D. -Electron Spin Resonance Studies of d1 Transition Element Alkoxides . . . . . THEORETICAL . . . . . . . . . . . . . . A. The Spin Hamiltonian . . . . . . . . . B. Molecular Orbital Theory . . . . . . . C. The Theory of Obtaining Molecular Orbital Parameters from Esr g and A Values EXPERIMENTAL . . . . . . . . . . . . . . . . . . . A. Materials . . . . . . . . . . . . . . . B. Analytical Methods . . . . . . . C. General Experimental Procedure . . . . . .D. Preparation of Compounds . . . . . . . E. Magnetic Moment Measurements . . . F. Spectroscopic Measurements . . . . . . . G. ~Determination of g and A Values from Esr Spectra . . . . . . . . . . . . . . . iv PAGE 030300 coco 12 12 13 23 TABLE OF CONTENTS (Cont.) PAGE RESULTS AND DISCUSSION . . . . . . . . . . . . . . . 46 A. Preparation of Complexes . . . . . . . . . 46 B. Magnetic Moments . . . . . . . . . . . . . 47 C. Optical Spectra . . . . . . . . . . . . . . 50 D. Infrared Spectra . . . . . . . . . . . . . 57 E. Electron Spin Resonance Spectra . . . 64 IF. Calculation of Molecular Orbital Parameters 73 SUMMARY . . . . . . . . . . . . . . . . . . . . . . 78 PART II INTRODUCTION AND HISTORICAL . . . . . . . . . . . . 80 A. Thermochromism . . . . . . . . . . . . . . 80 B. Vanadium(III) Complexes . . . . . . . . . . 83 XPERIMENTAL . . . . . . . . . . . . . . . . . . . . 86 A. Materials . . . . . . . . . . . . . . . 86 B. Analytical Methods. . . . . . . . . . . . . 87 C. General Experimental Procedure . . . . . . 87 D. Preparation of Compounds . . . . . . . . . 87 E. Magnetic Moment Measurements . . . . . . . 90 F. Spectroscopic Measurements . . . . . . . . 90 ESULTS AND DISCUSSION . . . . . . . . . . . . . . . 93 A. Preparation of Compounds . . . . . . . . . 93 B. Magnetic Moments . . . . . . . . . . . . . 94 C. Infrared Spectra . . . . . . . . . . . . . 95 D. Nuclear Quadrupole Resonance Spectra . . . 99 E. Optical Spectra . . . . . . . . . . . . . . 105 )NCLUSIONS . . . . . . . . . . . . . . . . . . . . 115 IBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . 117 LIST OF TABLES ABLE PAGE PART I I. Magnetic parameters for V(OR)4 . . . . . . . 9 II. -Esr and molecular orbital parameters for Nb(OCH3)C15=................10 III. Estimated errors in esr parameters . . . . . 45 IV. Magnetic moments of vanadium(IV) complexes . 48 V. .Magnetic properties of vanadium(IV) complexes 50 VI. Optical spectroscopic features of the tetra— methylammonium complexes . . . . . . . . 51 VII. Optical spectroscopic features of the tetra— ethylammonium complexes . . . . . . . . . 52' [II. Infrared spectroscopic features of vanadium(IV) 5 complexes . . . . . . . . . . . . . . . IX. Esr parameters for tetramethylammonium complexes . . . . . . . . . . . . . 65 X. Esr parameters for tetraethylammonium complexes . . . . . . . . . . . . . . 66 XI. Molecular orbital parameters and spin densities for V(OR)Cl5 complexes . . . . . 77 II. Molecular orbital coefficients for MoOX5= . 77 PART II I. Vanadium(III) nitrile adducts . . . . . . . 84 II. Magnetic moments of vanadium(III) complexes. 95 [1. Features of the far infrared spectra of the vanadiumf interest because of the large amount of work which has l done on the vanadyl systems. The pentachloroalkoxo— .date(IV) ion should be six—coordinate and the metal to 1 2 xygen bond would formally be a single bond. A study of a eries of these complexes in which the alkoxide group is hanged would show the effect on the bonding parameters of hanging the donating strength of the oxygen. The effect n the molecular orbital parameters of_changing the donating trength of the halide by changing from F- to C1- to r- has been shown in the case of the pentahalo(oxo)molyb— ate(IV) ion.1 HISTORICAL Vanadium(IV) Chemistry ,Of those d1 transition elements which have been estigated, vanadium in the tetravalent state has cer- nly been the most thoroughly studied. Vanadium(IV) normally bonded to oxygen to give the familiar vanadyl ++> I unit. The vanadium—oxygen bond has been shown to essentially a double bond, V=0.15§HMTVO(H20 four water molecules in a plane with a VSO distance 2.3 8; ‘While perpendicular to this plane is the V=O d of length 1.67 .17 The multiplicity of the bond see from the flow of electron density 0(pW)I—-> VVPT). ause of the strong V-O N-bonding in oxovanadium(IV) plexes, the interpretation of the electronic spectra is as simple as it would be for anordinary octahedral Ilex. There are presently unresolved differences of .ion as to the exact ordering Of the orbitals.18 Sidgwick19 provides a suitable review Of vanadium co- nation complexes through 1949. -Excellent and compre- ive reviews of oxovanadium(IV) compounds have recently ared.3°'21 ASomewhat fewer examples exist Of vanadium(IV) Dn-vanadyl systems. The reaction of vanadium(IV) 3 FIIIIIIIIll----::___________' 4 slides with nitrogen and oxygen bases has been surveyed 1 Fowles.22 Bridgland,-§5_§l,23caused vanadium tetra- Tloride and vanadium tetrafluoride to react with a variety 5 N—containing and O—containing ligands which did not con— Iin protonic hydrogens and obtained complexes of the general 'pe VX4L and VX4L2. All of the complexes are readily 'drolyzed and oxidized in air. Vanadium tetrachloride also rms 1:1 or 1:2 adducts with sulfur24:25, selenium26. osphorous27 and arseniczsv29 donor molecules. ‘Nicholls30 fers a thorough review of non-vanadyl complexes since 49. Solvolytic reaction of vanadium(IV) chloride with :Ohols and phenols causes cleavage of two V—Cl bonds, :ming the dichloroalkoxides. The general properties of ass compounds have been reviewed by Nicholls.30 .The alkoxides formed by the aliphatic alcohols are dark green solids, dimeric in boiling benzene. These dimers probably contain hexa-coordinate vanadium with a structure consisting of two octahedra sharing a com- mon edge through alkoxide bridges. They cannot be sub- limed, but on heating at 1500 C/0.1 mm they yield the oxychloride alkoxides, [V20C13(OR)3]. The tert—butyl and tert-amyl alkoxides are most conveniently prepared by alkoxide ion exchange with isopropoxides. Bradley andMenta31 and Thomas32.have prepared a iety of tetraalkoxovanadates(IV). These compounds are nally prepared by causing tetrakis(dimethylamide)vana- 2(IV) or tetrakis(diethylamide)vanadate(IV) to react with >hols. Primary,secondary and tertiary alkoxides of this a have been prepared. *The tertiary and secondary alkoxides 5 re predominantly monomeric in benzene. However,.tetra- ethoxovanadate(IV) is apparently a trimer in benzene which 5 the limiting degree of association for an octahedrally- oordinated metal alkoxide. Chamberlain, gg_al.33 prepared etrakis(triphenylsiloxy)vanadium(IV) from vanadium tetra— hloride and sodiumtetraphenylsilanolate. The liquid, thyl analogue was prepared by Thomas34 from vanadium(IV) iethylamide and triethylsilanol in benzene. I Only two previous examples of anionic,octahedral,non- nadyl vanadium(IV) complexes exist. Potassium, rubidium. ad cesium.hexafluorovanadate(IV) were prepared from vana- ium(IV) fluoride and the appropriate fluoride salt in alenium tetrafluoride or by fluorination of the appropriate entafluorovanadate(III) salt.35” Six saIESNOf the hexa- Ilorovanadate(IV) ion have been prepared. Gutmann35first .entified the hexachlorovanadate(IV) ion by the conducti— tric titration of potassium chloride with vanadium tetra- loride in iodine monOchloride. .Subsequent conductimetric trations of vanadium(IV) chloride with various.bases have own the presence of hexachlorovanadates(IV) in a range of Lorinated solvents.37'39 Other titratiOns gave evidence : the pentachloro—, heptachloro—, and octacthrovanadate(1v) 1s which have not yet been isolated.40 Fowles andWalton41 :st prepared the diethylammonium and triethylammonium salts the-hexachlorovanadate(IV) ion by the reaction of either ,adium(IV) chloride or its ethyl cyanide adduct with the ylammonium salt in chloroform. .Kilty and.Nicholls42 6 'epared the pyridinium, quinolinium, isoquinolinium, and :sium salts of the hexachlorovanadate(IV) ion by the re- tion of thionyl chloride with the corresponding tetra— loroToxo)vanadate(IV) salt at room temperature. M2VOC14 + SOC12 > M2VC16 + $02 ,All the salts except the cesium hexachlorovanadate(IV) Fsolve in acetonitrile to give red solutions. .Two peaks Ie been observed in the solution and reflectance spectra. ...1 . a peak at 15,000 cm is believed to be the 2T2g —v ZEg Insition. vThis peak is somewhat broad and asymmetric. I more intense peak around 21,000 cm"1 is a charge trans- ‘ band and accounts for the dark red color. Anionic d1 Transition Element Alkoxides Recently several examples of d1 anionic alkoxides e been prepared. »Wentworth and Brubaker9l10 first pre— ad the pentachloroalkoxoniobate(IV) complex; the alkoxide methoxo, ethoxo, or iSOpropoxo, while the cation was Elly a large protonated organic base. IThe complexes 2 prepared by the addition of the appropriate base to a Ition of niobium(v) chloride in alcohol which had been .ced electrolytically to niobium(IV). The color of the lexes was found to depend onthe cation used. -Funk, gE_al.43 first prepared pyridinium tetrachloro- thoxomolybdate(V) by the treatment of Mo(OCH3)2Cl3°3CH30H pyridinium chloride in methanol. .McClung, et al.12 7 epared and investigated various other salts of the methoxide mplex as well as the ethoxide and isopropoxide complexes. ese complexes were prepared by a method somewhat different om.Funk's. -MOlybdenum(V) chloride was dissolved in alcohol 1959K and allowed to warm to room temperature. An excess the cation was then added to the green solutions to give e crystalline products. Funk and Naumann44 first prepared the pyridinium and iethylammoniumsalts of the tetrachlorodimethoxotungstate(V) n. .Rillema,§t_al,14 have recently prepared other examples the tetrachlorodialkoxotungstate(IV) complex as well as ier salts of the dimethoxo complex. ~They also reported a first examples of the pentachloroalkoxotungstate(V) ion. ese complexes decompose in the solid state to give the :rachloro(oxo)tungstate(V) ion. The decomposition is (R4N)W(OR')C15 > (R4N)WOC14 + R'Cl test when R and R' are methyl. The decomposition is er as either R' or R becomes larger. An apparently n coordinate species, tetramethylammonium hexachloro- xotungstate(v), was also isolated from an ROI saturated nol solution in which tungsten(v) chloride was dissolved. Giggenbach and Brubaker13 repOrted preparing a series itanium(III) mixed chloride-alcoholates. .More recently same authors reported the preparation of two other nium(III) chloride—methanolatesle, [C5H6N12T1Cl5(CH30H) [C5H6N]Ticl4(CH3OH)2, which are similar to the tungsten(v) 8 and molybdenum(V) alkoxides discussed earlier. .NO anionic alkoxide complexes have been reported in the case of zirconium(III), hafnium(III), tantalum(IV),or chromium(V). C. The Application of Electron.Spin Resonance to Transition ‘Metal Complexes _Several comprehensive reviews on the early development of electron spin resonance-are given in previous theses from this department and will not be duplicated.45'43 Various recent books and reviews adequately cover the >asic fundamentals of electron spin resonance. .Low‘s49_book .s extremely useful although it covers only the esr of solids. :lichter50 presents the theoretical development of esr in :he solid state andPake's51 book is also a good reference. nderson's52 review gives various experimental applications f esr. .Carrington and Lonquet-Higgens53 have a thorough rticle covering the applications of esr to transition metal mplexes and Robertson‘s54 review also covers the applica— 'ons of esr to transition metal complexes. ISeveral other more recent sources are available. .KuSka d Rogers5 certainly offer the most complete review of ectron spin resonance studies of first row transition ement complexes available at this time. Their chapter om "Radical Ions" is an extremely useful source as a complete View of the literature as well as a concise statement of e basic theories of covalent bonding parameters. Several 9 other chapters in "Radical Ions" also deal with inorganic systems.55:56 PD. .Electron.Spin Resonance Studies of d1 .Transition ,Element Alkoxides ,Only one example exists of an esr study of vanadium(IV) alkoxides. .Kokoszka, §E_al,57 investigated the tetrakis; (Efbutoxo)vanadate(IV) complex. .The Spectra were measured in the temperature range of 295 to 770K. The measurements were made on pure V(OR)4, 1-2% V(OR)4 in Ti(OR)4, and 1-2% V(OR)4 in C82 and the spectra were approximately the same in all samples. 'The electron spin resonance parameters for this compound are given in Table I. * Table I. .Magnetic parameters for V(OR)4 l‘ '— Liquid Spectrum (2950K) = 1.964 r 0.005 = 64.0 s 2.0 Polycrystalline Spectra (770K) g = 1.940 r 0.005 A = 125 i 5.0 g = 1.984 1 0.005 A = 36 i 4.0 Hyperfinesplittings are given in 10—4 cm_1. .Rasmussen, et al.11 investigated the solution and rozen solution (glasses) spectra of Nb(OCH3)C15=. Both e isotrOpic g values and hyperfine splittings due to e 93Nb (100% I = 9/2) nucleus were observed. The results 10 were interpreted in terms of an approximate molecular orbit— al approach to the bonding which takes into account the ef- fect of charge—transfer states. The esr parameters and molecular orbital parameters obtained for this complex are listed in Table II. Because of the broad lines and lack of sufficient optical data, an accurate calculation of the molecular orbital parameters could not be made. Table II. Esr and Molgcular Orbital Parameters for Nb(OCH3)Cl5 . Liquid.Spectrum 1.869 r .002 = 178 i 3 gauss Frozen Solution Spectrum gu = 1.923 r .008 A“ = 248 i 6 gauss 91. = 1.842 r .010 A1. = 144 i 10 gauss Molecular Orbital Parameters 2 = 2 ss.45 N2 21.6 NW} .62 N02 72 The electron spin resonance data which were reported for alcohol solutions of tetrachlorodialkoxomolybdates(V) now appear to be in error. The species being studied were probably the oxyion species and other hydrolysis products. Phe tetrachlorodialkoxomolybdates(V) are stabilized in Ilcoholic solutions which contain small amounts of HCl. The decomposition probably involves mechanisms such as: 11 Mo(0R)2c14' + ROH > Mo(OR)3Cl3— + HCl fast Mo(OR)Cl3- > MoO(OR)Cl3_ + ROR HCl +MoO(OR)Cl3— < > Mooc14‘ + ROH + .MoOCl4— + HCl > MoOCl5_2 + H This would explain the stabilizing effect of HCl as well as the esr being essentially the same as molybdenum oxochlorides which have been studied. Rillema and Brubaker58 have re— investigated the esr spectra of solutions of tetrachloro— dialkoxomolybdates(v). Rillema59 studied the solution and frozen solution spectra of various pentachloroalkoxotungstates(V) as well as various tetrachlorodialkoxotungstates(V). The broad lines and lack of known constants ($42. spin orbit coupling con- stants, overlap terms, etc.) for tungsten(V) made the calcu— lation of accurate molecular orbital parameters impossible. Giggenbach and Brubaker16 have recently reported the esr spectra for a number of titanium(III) chloride-methanolates. Although very accurate g values could be obtained, no metal hyperfine structure was observed in any of the spectra and hence molecular orbital calculations could not be made for these complexes. THEORETICAL A. The Spin Hamiltonian An unpaired electron in a transition metal complex interacts with its environment in several ways which are sensitive to study by electron spin resonance. These inter— actions are normally described in the form of spin Hamilton- ian of the general form: X = “925sz + ngXHX + gySyH ] + [AZSZIZ + AXSXIX + AySny] L Y L L + A S I 1 X yy] H L + AXSXI Y + Z[ALSIL L ZZZ where gX, gy, and gZ are the spectroscopic splitting factors, 6 is the Bohr magneton (0.92731 x 10_20 erg/Gauss), x’ Hy’ and H2 are the components of the magnetic field along the x, y, and z direction, and Sx’ Sy’ and Sz H are the components of the electronic spin operator along the X, y, and z magnetic field axis respectively. .I:, Iy’ and L . . . . I are the ligand nuclear spins in the x, y, and z direc- z t' . , . . ions Ax AY and AZ 2, A?, and A: are the ligand hyperfine are the metal hyperfine interaction constants and A interaction constants. 'Ix"Iy’ and IZ are the components of the metal nuclear spin along the respective axes. 12 13 .For those cases in which the complex possesses axial symmetry, x = y = ‘i and z = I] to give the more familiar form of the spin Hamiltonian. 3? = B[CEI'ISZHZ + gJ-(SXHX + SYHYH + [Allszlz + AJ-KSXIX + SYIYN + [z [Atszii' + AL (5X1: + sy1§)]] (2) L . The first term in brackets in equation 2 is commonly referred to as the zeeman portion of the spin Hamiltonian. The second term is the metal hyperfine portion of the Spin ‘Hamiltonian. The thud camlflsthaligand hyperfine, or super- hyperfine, portion of the spin Hamiltonian and it may be disregarded if no ligand hyperfine splittings are The observed in the electron spin resonance spectrum. principal information gained from the esr Spectrum is the evaluation of the various g and A values of the spin Hamiltonian. B. -Molecular Orbital Theory Various theories have been proposed and used to ex- plain bonding in inorganic complexes. The first such theory was called Valence Bond Theory, VBT, and assumed that the ligands were groups which in some way donated electron pairs to metals, thus forming the so-called coordinate link. This theory enjoyed great and almost exclusive popularity in the 1930's and 1940's but was supplemented during the 1950’s by Ligand Field Theory, LFT, LFT was developed between 1930 14 and 1940 by physicists, mainly, J. H. Van Vleck. The Ligand Field Theory that we employ today evolved from a purely electrostatic theory called Crystal-Field Theory, CFT, which treats the metal and the ligands as pure point charges or point dipoles. At the opposite extreme, Molecu— lar Orbital Theory, MOT, treats the metal—ligand interac— tion in terms of molecular orbitals formed by overlap of metal and ligand atomic orbitals. These various theories are covered in several books and reviews.6°I64 Molecular Orbital Theory has been extremely valuable in interpreting cases where there is a strong metal—ligand interaction (covalency). Molecular Orbital Theory starts with the premise that metal and ligand orbitals will overlap to some degree when- ever symmetry permits. -It thus includes the electrostatic situation (no overlap) as one extreme, maximal overlap as the other extreme, and all intermediate degrees of overlap in its scope. The first task in working out the molecular orbital treatment for a particular type of complex is to find out which orbital overlaps are and are not possible because of the inherent symmetry requirements of the problem. This can be done elegantly and systematically by the use of group theory principles. If one considers octahedral complexes, it is quite easy to obtain the pertinent molecular orbitals involved in 15 sigma (o) and pi (w)-bonding. It is easy then to extend this treatment to complexes with lower symmetry. In complexes involving transition metals one needs to only consider those metal orbitals which are valence orbitals, 3dzz, 3dx2—y2’ 3dxy’ dez’ 3dyz’ 4s, 4px, and 4py, and 4pz. It can be shown easily that the 3dZ2 and 3dx2-y2 orbitals transform as the E9 representation in Oh symmetry. Also the 3dxy’ 3dyz and 3dXZ orbitals transform as the ng representation, the 4s orbital transforms as the totally symmetric Alg representation and the 4px, 4py, and 4pZ orbitals transform as the T1u representation. Our task then is to construct ligand orbitals which have symmetries such that they can overlap with these metal orbitals to form either sigma or pi bonds. These ligand symmetry orbitals will be a linear combination of atomic ligand orbitals, LCAO, so that the bonding theory we are using is commonly called, LCAO-MO theory. One first needs to obtain six ligand symmetry—orbitals to form six sigma—bonds with the metal orbitals. It is known from symmetry considerations that the 3dx2—y2’ 3dzz,4s, 4g{,4p&, and 4pz metal orbitals lie along the x, y, z axes toward the ligand so the ligand sigma orbitals will be constructed to transform as Eg, A19, and T1u representa— tions. -In this case, the six ligand symmetry—orbitals denoted 1 through 6, will be made from the six spz hybrid orbitals directed along the x, y, and z coordinates toward the metal. (A local right—handed coordinate system 16 where the a-bond is the z axis has been taken on each on each ligand.) These six LCAO orbitals are given below. ¢Eo = %'(01 + 02 + 03 + 04 + C5 + 06) (A19) (3) $20 - ‘1' (01 - 02 + 03 _ O4) (Eg) (4) L _ 1 $30 E (‘01 " 02 — 03 " 04 + 205 + 206)(Eg) (5) ¢£o fJ%'(O5 _ 06) (Tiu) (6) 65:0 7% (a. - a.) (arm) (7) «>20 7;- (a. - a.) (Tm) (8) Where 01 and 03 lie along the + and — x axis, 02 and 04 along the + and — y axis, and 05 and 06 along the + and — z axis respectively. Now these six symmetry orbitals can overlap with the respective metal orbitals to form a sigma- bonding orbital (positive overlap) or a sigma—antibonding orbital (negative overlap). These molecular orbitals are commonly written as wb = N<¢m + ML) (9) m V = N(¢> — NDL) (10) . . L . where ¢ is the metal atomic orbital, ® is the ligand symmetry orbital, N is the normalization constant, and A is the coefficient for the symmetry orbital. The separation between these two orbitals can be calculated. However, the separation is proportional to the amount of overlap and a 17 rough calculation of the overlap will give a good estimation of the separation. A molecular orbital diagram for an octa- hedral complex where no w—bonding has been considered is shown in Figure 1. The t2g (dxz, dyz’ and dxy of the metal) orbitals are nonbonding. The asterick denotes an antibonding orbital which always lies higher in energy than its corresponding bonding orbital. If the ligands have v—orbitals, filled or unfilled, it is necessary to consider their interaction with t2g d orbitals on the metal. The simplest case occurs in the event that each ligand has a pair of p orbitals mutually perpendicular. Thus one has twelve linear combinations of atomic orbitals to consider. It can be shown that these twelve atomic orbitals transform as T19, T29, Tiu’ and T2u representations. Those atomic orbitals in the classes T19 and T2u will remain rigorously nonbonding. This is for the simple reason that the metal does not possess any orbitals of these symmetries with which they could inter— act. The T set could interact with the metal p orbitals, iu which are themselves a set with T1u symmetry and in a quantitative discussion it would be necessary to make al— lowances for this. ‘However, in a qualitative treatment it may be assumed that since the metal p orbitals are already required for sigma-bonding and no 7—bonding will take place involving the T1u orbitals. This leaves only the tzzg set of symmetry orbitals to overlap With the metal tzg d orbitals. "" .» .5... —--‘~._........ 18 19 / . / . \\ / t1u \ tlu ’// \\ 4p I / \\ \ \ \ \l a / \\ 45““ }g \ ‘\\ \ \ e; \\ \‘\ // , \ \ \\\ \ \ I, \~.\ ‘ \\ I \\ a1u tiu 69(0) \(( I] ‘ t29 eg / \\ t /// \. \ 4’/I \\ /// \\ /,/ \\ / e /./ \ \\ g / / / \ \\ / 1 \ \ tiu / / \ / \ / t 19 / Metal Ligand Figure 1. Molecular orbital diagram for an octahedral complex with no w—bonding. 19 The ligand symmetry orbitals which overlap with the metal dxz orbital can be constructed as at =% £qu (1) + px (s) '+ px (3) + py (6)“) (11> where the numbering system is the same as used in the sigma— bonding caSe. Two similar symmetry orbitals can be written which overlap with the metal dxy and dyz orbitals. Again a set of bonding and antibonding orbitals are formed to give the MO diagram in Figure 2. Now that the molecular orbital diagram has been developed for an octahedral complex, it is easy to obtain the molec— ular orbitals necessary for the discussion of other complexes of lower symmetry derived from octahedral symmetry. In the case of the vanadium.alkoxide complexes discussed in this work, the symmetry is C4v' From a correlation table such as that given by Cotton55, the degenerate sets of sym- metry orbitals in Oh symmetry can become nondegenerate in the lower C4v symmetry. For example, the tzg anti- bonding w-orbitals in Oh symmetry become b2 and £3 anti— bonding w—orbitals in C4V symmetry. Therefore the molecu- lar orbital diagram for a complex with C4V symmetry ap— pears to be somewhat more complicated than the basic Oh case but is easy to understand if this approach is used. (Figure 3). It cannot be shown from group theory which of the two orbitals b2 or e is lower in energy. Similarly, it cannot be shown which of the two orbitals, a1 or b1 20 19 /' V 1 / .. \w // ’ tm \\ / \ tlu ’,/f/ \ \ 4P / / \ \ \/ \ \ a \ 4s lg X * \\ \ \_ eg \\ \ \ / \\ \\ \ \ / \ \ a t e ((7) / * \ 19 111 g \) tzg (F) I] t e /« \ "" ‘\“\. [VI . 3d 29 9 bflfi . \ tzg (V) {A tzg W [717,” \ \ ‘\\ / // \ ‘m e9. / // \\\ t / / \ ‘ 1u I/l/ '\ / \ a1. / Metal Ligand Figure 13. Molecular orbital diagram for an octahedral complex with w—bonding. / \ / \ / \ / \ / \ l a: \ I /$'—a§ _\\ \\ /// Q\\ 4p L’ l/ \\ \ / .1, \§ \ / I \ §\ 4s _1 / \\ §|\ W / b1 \ \\ \\\ ’ \ \ \ \l 9* (7T) \\\\\ a, e (W) \ \ ///// \‘l \\ NW \\ \\.—.—a-L—/ /// Mt.e // \‘\ 'bi //'l V\ a1 / / Metal \‘ / Ligand ‘\ a1 / Figure -3. Molecular orbital diagram for a complex with C v symmetry and v—bonding. 22 arising from the <39 orbitals is lower in energy. However, either intuitive arguments or crystal field theory will show that the b2 orbitals lies at a lower energy than the e orbital and the b1 orbital lies at a lower energy than the a1 orbital. The determination of the separation of the levels must come from the ultraviolet-visible spectrum of the complex. Now if one fills in the orbitals with the available electrons, there are 12 electrons from the 6 ligand spz hybrid orbitals and 6 electrons from the 3 5) orbitals which form bonds with the t2g metal orbitals for a total of 18 electrons. The molecular orbitals are filled up to the lowest lying w-antibonding orbital so that in a d1 case, the unpaired electron is in the b2 antibonding molecular orbital. The first two excited states are e and b1 so that the molecular orbitals necessary for the discussion of the bonding in a d1 case are the b2, e, and b1. These molecular orbitals may be written similarly to those given above (Equation 10). L * = _ IB2> NV2(dXy xvz ¢W2) (12) _ _ L IB1>* — N02(dX?‘Y2 102 $02) ‘13) * = _ e e _ a a ]E > N71(dxz or dyz xvi $71 XVI ¢F1 (14) where <1>L = -1—(p (.1) + p (2) + p (3) + p (4)) (15) 23 ¢gz = %(0(1) + 0(2) + 0(3) + 0(4)) (16) Q -J% + px<3>> (17> ¢31 = éépgm + py(6>) (18) and A31 and A31 are the ]E>* molecular orbital coef- ficients for the equatorial and axial ligands respectively. The ligand orbital of the chlorine and oxygen involved in W-bonds are pure p orbitals while the ligand orbitals involved in o—bonding are spz hybrids. -In the actual calculation of the molecular orbitals, it is convenient to e a assume ha = he so that ¢ + ¢ 3 ¢ .( 7T1 7T1 7T1 T7'1 7T1 C. -The Theory of Obtaininngolecular Orbital Parameters from ESR g and A Values For a free electron the value is 2.0023; however, in transition metal complexes, spin—orbit interaction mixes some excited state into the ground state. The actual (9) value is given by the expression g = 2.0023 (éij — c1\ij) (19) c where 5ij is the Kronecker delta, Q is the spin orbit coupling constant, and-[iij is defined by 0 Aij = z 0. Iii 0,201,112]. 00/02,, - E >66 (20) n#o ‘ where i is the angular momentum operator, and (En - E0) is the energy between the ground and nth excited state. 24 Owen67 found that by using optical and magnetic data the value for the spin—orbit coupling constant in equation 12 was smaller by 20—30% than the free ion value. This de- crease was interpreted in terms of covalent bonding between the metal and the ligand which forces some unpaired electron density onto the ligand. Murao68 attributed the lowering of the spin—orbit coupling constant to screening by the additional. 3d elec— tron density produced by the mixing of 3d chlorine wave functions into bonding orbitals. -Several alternate procedures for considering screening effects have also been proposed.3: 69:70 Several investigatorse’s’“71 have recently amended Owen‘s early theory to include charge transfer and ligand spin—orbit coupling contributions. The complexes considered in this investigation have C4v symmetry (Figure 4) and form coplanar bonds between the metal and each of four equatorial chlorine ligands. The alkoxide ligand is attached axially along the positive 2 axis. The fifth chlorine lies along the negativer z axis. The V(OR)Cl5= ion has one unpaired electron (d1) so the ground state is 2B2. (See molecular orbital section.) The molecular orbitals which are necessary for this discus- sion are the ground state and the first two excited states and are given by equations 12, 13, and 14. 25 CW c1 / C1 C1 Cl Figure 4. Symmetry of pentachloroalkoxovanadate(IV) complexes. The experimental g and A values can~be related to the molecular orbital coefficients by: ‘8CN2 N2 1 -2.0023 = ”2 02 1-— T —2 s -2 g“ AKBZ - b1) [ 2()\'IT27\O'2) (n) 7\0'2 b1 .A'Wzsbz] (21) -2CN2 N2 .. . : 77-2 EL -1 e _ , 9L 2 0023 Asz _ e) [1 f2(?\7’_1?\w2) ZAWZsz ( ) . 22 e a . — J2 iwisbz — xvlse] _ 2 2 2 2 2 _ 4NF2P SCNFZNOZP 6CNW2NF1P A - — __.____ — — (23) " 7 “("“A b2 --"'—b1) 72092 - e) where S S and .8e are the metal-ligand overlap b1' b2' 26 integrals. T(n) is defined by Kivelson and Lee72 as: T(n) = n -(%)%(1 — n2)-;- R éwr2R31(r)-&%[R3o(r)]dr (24) where R is the vanadium—ligand distance and R31(r) and R30(r) are the normalized radial 3p and 38 functions respectively. P = 2.0023 gNBefiN avg., where Be and SN are the Bohr and nuclear magnetons, respectively, and g is the nuclear g factor. N and A are related N TTZ 1T2 by the normalization requirement. 2 = _ 2 ‘1 NW2 [1 “Was +i7T2] (25) b2 This procedure can be extended to permit calculation of the spin density in the p” orbitals of the equatorial chlorides and involves use of Mulliken's populational analysis which assumes the electron density is proportional to the square of the molecular orbital coefficient.73 Several authorsll2 have found excellent agreement between spin densities calculated in this manner and those calcu— lated from ligand hyperfind splitting values. 2 N2 _ N2 Unpaired electron density = #2 02 vzxwzsbz (26) in each ligand 3pTr orbital 4 . EXPERIMENTAL A. Materials Vanadium Tetrachloride.-— Vanadium tetrachloride was prepared by elemental synthesis. Vanadium metal powder was obtained from Alfa Inorganics, Inc. (m2N8). The re- action took place in a tube furnace at 390° afl.the gaseous VCl4 was condensed in a water condenser and subsequently collected in an ice bath. The use of the water condenser allowed the reaction to proceed much more rapidly than if a direct connection had been made to the ice bath.‘ The crude vanadium tetrachloride was stored under chlorine in the absence of light until just before use. The VCl4 was then distilled at atmospheric pressure under chlorine and degassed. (Calcd: V, 26.43; Found: V, 26.14.) Tetraalkylammonium Chlorides.—— Tetramethylammonium chloride was obtained from Eastman Organic Chemicals and dried at 1100 before use. -Tetraethylammonium chloride was also obtained from. Eastman Organic Chemicals and dried at 800 before use. Pyridine.—— Reagent grade pyridine was stored over sodium hydroxide. Prior to use, the pyridine was allowed 27 r___—f 28 to reflux over finely crushed barium oxide and was distilled. Phenol.-- Reagent grade phenol was sublimed twice at room temperature. Hydrogen Chloride, Chlorine, and Nitrogen.-— Anhydrous hydrogen chloride was obtained from.Matheson Chemical Company and passed through concentrated sulfuric acid before use. Chlorine was obtained from Matheson Chemical Company and was used without additional purification. Pure nitrogen was obtained by passing Matheson prepuri- fied nitrogen through a three foot BTS74 column and sub— sequently through calcium chloride and barium oxide drying towers. Solvents.-- Methanol, ethanol, and isopropanol were dried by reaction with magnesium and were subsequently dis— tilled. Cyclohexanol, nfpropanol, nfbutanol, and n7 octanol were reagent grade and were used without additional purification. Acetonitrile was caused to reflux over phosphorus pentoxide and distilled under a nitrogen stream. This process was repeated at least three times and the solvent was then stored over Linde—4A molecular sieves. Methylene chloride and carbon tetrachloride were allowed to reflux continuously over calcium hydride and distilled as needed. 29 Chloroform was distilled twice from phosphorus pent— oxide and stored over Linde-4A molecular sieves in the absence of light. Nitromethane was distilled from calcium chloride and passed over a two foot column of Dowex 1—X8 resin in the acid form, which had been previously dried with anhydrous ethanol. Ethyl ether was distilled from solid sodium and stored over fresh sodium wire. Thionyl cthride was reagent grade and used without additional purification. B. Analytical Methods Vanadium Analysis and Oxidation State Determination75.—- Weighed samples of a vanadium complex were dissolved in dflute sulfuric acid and the solution was heated to boiling and allowed to cool to 60-800. The solution was then titrated with standard KMnO4. Sulfur dioxide gas was then passed through the cooled solution for five minutes followed by nitrogen for 20 minutes. The solution was then retitrated with standard.KMnO4. If the results of the two titrations agreed, the oxidation state was confirmed as +4. The second titration proved to be more reproducible and those values are reported. Chloride Analy§i8.—~ Weighed samples of the vanadium complexes in Schlenk—tubes were cooled to 770K. Dilute ‘ 30 sulfuric acid was then added to a sample and the mixture was warmed slowly to room temperature. This procedure prevents the loss of hydrogen chloride. Such losses may occur if the samples are dissolved directly in water or dilute sulfuric acid at room temperature. Chloride was determined by potentiometric titration of aliquots of the resulting solution with standard silver nitrate. Gravimetric chloride determinations were made on several of the less stable complexes. The silver nitrate solution was added directly to the sample in a cooled Schlenk-tube. This procedure eliminated the possibility of loss of hydrogen chloride but was very time consuming. Carbon, Hydrogen, and Nitrogen Analyses.-— The analyses were performed by Spang Microanalytical Laboratory, Ann Arbor, Michigan which reported, “The compounds hydrolyze very easily." Pyridine Analysis.—— Solutions of pyridine up to 1.7 x 10-4 M'in 0.05 N sulfuric acid have been found to obey Beer's Law at 255 mu.76 Samples were dissolved in aqueous solutions of sulfuric acid of known concentration. The solutions were diluted to one liter with sufficient sulfuric acid and water so that the final concentration was 0.05 N in H2804. The pyridine concentration was determined spectrophotometrically at 255 mu. (6 = 5.32 x 103 l. mole~1 cm~1.)13 31 C. General Experimental Procedure All reactions, transfers, weighings, etc. were carried out in either an inert atmOSphere (dry N2) or in a vacuum, mainly by use of Schlenk-tube methods. Filtra— tions and washings were all performed by the application of pressure or by suction. All drying was in vacuo and compounds were stored under nitrogen in the absence of light. D. Preparation of Compounds Tetramethylammonium HexachlorovanadateLIV).—- Tetra— methylammonium chloride (5.0 g) was dissolved in approxi— mately 50 ml of thionyl chloride. A solution consisting of 3 ml of freshly distilled vanadium tetrachloride in 20 ml of thionyl chloride was added and the solution stirred vigorously during the addition. A red-black precipitate formed immediately. The precipitate was extremely fine and difficult to filter. The filtration took several days. The product was washed twice with 100 ml of carbon tetra- chloride and once with ethyl ether and was dried under vacuum. Axial. Calcd for VC16C3H14N2: v, 12.3: cl. 51.65: C, 23.31; H, 5.87; N, 6.80. .Found: V, 12.34; Cl, 51.65. (gravimetric), 51.20 (potentiometric); Cl, 23.01; H, 5.68; N, 6.68. Oxidation Number: 4.07. 32 Tetraethylammonium Hexachlorovanadgte(lv).—— Tetra— ethylammonium chloride (9.0 g) was dissolved in approxi— mately 40 ml of thionyl chloride. A solution consisting of 3 ml of freshly—distilled vanadium tetrachloride in 20 ml of thionyl chloride was added and the solution was stirred vigorously during the addition. A red—black oil formed in the reaction flask. Dry carbon tetrachloride was added slowly until a crystalline precipitate was observed. The precipitate was filtered under nitrogen, washed twice with carbon tetrachloride, and with dry ether, then dried under vacuum. AEEL- Calcd for VC16C16H40N2: V, 9.72; Cl, 40.60; C, 36.64; H, 7.69. Found: V, 9.51; Cl, 40.47; C, 36.31; H, 7.83. Oxidation Number: '4.11. Pyridinium Hexachlorovanadate(IV).—— Pyridinium chlor— ide (5.0 g) was dissolved in 50 ml of chloroform. A solu— tion consisting of 3 ml of freshly—distilled vanadium tetrachloride in 20 ml of a 5% solution of thionyl chloride in chloroform was added and the solution was stirred vigor- ously during the addition. A dark red precipitate formed immediately. The mixture was stirred for one hour, filtered under nitrogen, washed with 100 ml of chloroform and dried under vacuum. This complex had been prepared by Kilty and Nicholls.42 Anal. Calcd for VCl6C10H12N2: Cl, 50.19. Found: Cl, 49.92 (gravimetric). 33 Tetramethylammonium.Pentachloromethoxovanadate(IV).-- Tetramethylammonium hexachlorovanadate(IV) (2.7 g) was dis— persed in a solution of 20% acetonitrile in‘ether° The slurry was stirred for two hours to break up any lumps in the starting material so that the reaction could proceed smoothly. (The reaction also takes place in methylene One equivalent of methanol was chloride or nitromethane.) -The reaction proceeded added while the mixture was stirred. immediately with the formation of a golden product. Hydro— gen chloride could be detected above the reaction flask. The product was stirred one hour, filtered under nitrogen, and dried under a washed with warm acetonitrile and ether, vacuum. 12.50; Cl, 43.50; _Anal. Calcd for VCl5C9H27N20: V, 6.68. 'Found: V, 12.60; Cl, 43.56: c, 26.44; C, 26.50; H, 4.07. H, 6.78. Oxidation Number: Tetramethylammonium Pentachloroethoxovanadate(IV).-— This complex was prepared in the same manner as the above methoxo complex. 12.09; Cl, 42.06; Anal. Calcd for VC15C10H29N20: VI C, 28.47; H, 6.94. AFound: V, 12.20; Cl, 42.03; c, 28.56; H, 6.98. Oxidation Number: 4.00. Tetramethyiammonium.Pentachlorogg—propoxolyanadate IV ,__ This complex was prepared in the same manner as the above methoxo complex except the tetramethylammonium hexachloro— vanadate(IV) was diSpersed in a 50% acetonitrile—ether solu— tion. 34 .AnEL.Cahxifor\K35CLfih1Nflhz V,]J.707Cfl,‘“L71. Found: V, 11.69; Cl, 40.67° Tetramethylammonium Pentachloro(n:butoxo)vandate(IV).-— This complex was prepared in the same manner as the above methoxo-complex except the tetramethylammonium hexachloro- vanadate(IV) was dispersed in acetonitrile. Afléln Calcd for VC15C12H33N203 V, 11.33; Cl, 39.44. Found: v, 11.22; c1, 39.21. Tetraethylammonium Pentachloromethoxovanadate(IV).-- This complex and subsequent complexes were prepared in the same manner as the corresponding tetramethylammonium com- plexes. -Ag§l. Calcd for VCl5C17H43N20: V, 9.81; cl, 34.12. Found: V, 9.81; Cl, 34.10. Oxidation Number: 4.17. Tetraethylammonium Pentachloroethoxovanadate(IV).—- Found: V, 9.70; Cl, 33.41. Tetraethylammonium Pentachloro(nepropoxo)vanadate(Iv).-— Anal. Calcd for VC15C19H47N202 V, 9.30; Cl, 32.26. Found: V, 9.20; Cl, 32.17. Pyridinium PentachloromethoxovanadatejIV .-- Anal. Calcd for VC15C11H15N20: V, 12.15; Cl, 42.26; C5H6Nj3 38.19. Found: v, 12.16; c1, 42.30; C5H6N+, 38.61. 35. -Attempts to Prepare Other Monoalkoxides.—— Attempts to prepare the tetraethylammonium pentachloro(nfbutoxo)- vanadate(IV) as well as various other monoalkoxides were unsuccessfifli Evenin refluxing acetonitrile, cyclohexanol, isopropanol, tfbutanol, and nfoctanol do not react with the hexachlorovanadate(IV) ion to any appreciable extent. One assumes that steric factors play an important role here. -No attempt? WGSE made to prepare other monoalkoxides starting with sodium or potassium alkoxides. There is no obvious method available to separate the product from the NaCl or KCl. Phenol reacts extremely rapidly with the hexachloro- vanadate(IV) ion in acetonitrile. However, no stoichio- metric compound could be isolated. Attempts to Prepare Complexes Containing the Tetra- chlorodialkoxovanadate(IV) Ion.—- Considerable time was spent trying to prepare dialkoxo compounds. Two basic methods were tried and will be discussed. Green solutions were obtained if the hexachlorovana— date(IV) salts were dissolved in alcohols. When the solu— tions were evaporated to dryness, only salts containing the tetrachloro(oxo)vanadate(IV) ion could be obtained. Variations of temperature of the reaction and in the con— centration of the reactants had no effect on the product. The addition of HCl gas during the reaction did not elimin— ate or slow down the decomposition. Although the 36 hexachlorovanadate(IV) salts dissolved slowly in nfpropyl and nfbutyl alcohol, the decomposition seemed to be as rapid. If two or more equivalents of alcohol were added to a slurry of the hexachlorovanadate(IV) salts in ether or ether—acetonitrile mixtures, the isolated product appears to be the monoalkoxo species. However, a series of prepara— tions that used 1, 2, 3, or 4 equivalents of methanol per vanadium was carried out. An infrared spectrum was obtained for each product and these are shown in part in Figure 5. The spectra show a decrease in the intensity of the peak at 1050 cm~1 normally assigned to the presence of a meth— oxide ligand in going from 1 to 4 equivalents/‘7'6 There is also an increase in the intensity of the peak at 1000 cm-1 normally associated with the presence of a vanadyl unit. Figure 6 shows the infrared spectrum of pure [(CH3)4N]2- V(OCH3)Cl5 and a completely hydrolyzed sample.) Experiments with various other solvents such as nitromethane, chloro— form, and ether gave similar results. Attempts to prepare the dialkoxide complexes by the reaction of sodium or lithium alkoxide with the hexachloro— vanadate(IV) salts were not made. There is no reason to believe these Would be any more successful than the two methods described. The decomposition must involve a reaction of a higher alkoxide species formed in solution and not a decomposition of the monoalkoxide° The mechanism is probably similar to .thUm mOmmU mo mpcmam>flfivm 9 .Q .Umpvm mommo mo mucwam>flsvm m .0 I’m aIEo oomzoofifiv MHuommm @mHmHmQH .tmotm mOMmo mo muqmam>flsam m .m .omotm mOmmo mo hamam>asvm H "huaz 6Ho>NHZaAmmovL How A .m musmflm EU ,, a- com oooH ooHHoom coca ooHH oom oooH OOHH 00m OOOH OOHH Ci 37 38 gJM‘“x a... :00. . I. ‘0... o: :. 3 3 3 £ .5 v 35 0.; a: l I l 1200 1009 800 cm Figure 6 . Infrared spectrum of pure [(CH3)4N]2V(OCH3)C15 —1 (800-1200 cm ) (~—----) . Infrared spectrum of hydrolyzed [(CH3)4N]2V(OCH3)C15 (800—1200 cm_1) ( ~~~~~ ). 39 that proposed for the decomposition of the molybdenum di- alkoxides. vc16‘ + CH3OH > V(OR)C15= + HCl < > V(OR)2C14= + HCl \ V(OR)C15= + CH30H <—__ V(OR)2C14_ £335——> VOCl4= + ROR. -E. .Magnetic Moment Measurements Magnetic susceptibilities were measured by the Gouy method by use of methods similar to those described by Vander Vennen.77 The major difference lwasg that the apparatus was constructed in order to allow a constant stream of helium to pass over the sample tube. This pre— vented water from condensing on the sample tube at low temperatures and also protected the sample from hydrolysis. The calculation of the magnetic moment was made by the use of the equation:78 106 X = f' x 5e- (27) S where X is the gram—susceptibility of the sample; f' is the force exerted on the sample alone, i.e., the measured force corrected for the force experienced by the tube alone; WS is the weight of the sample in grams; and B is the tube constant. In practice the constant 6 must be determined for a particular tube by use of a material of known susceptibility 40 In this work, Hg[Co(SCN)4] was used: its susceptibility is 16.44 x 106 cgs units.79 ' The molar susceptibility, Xm’ of the sample is ob— tained by multiplying the gram—susceptibility by the molecu— lar weight. The susceptibility of the metal ion, x$, is obtained by correcting the molar susceptibility for any dia— magnetic species present. Pascal's constants74 were used to estimate the diamagnetism of the ligands and cations. In normal paramagnetic substances XQ is related to the absolute temperature as (Curie Law) (28) x B I aka or I C Xm = my (Curie—Weiss Law) (29) For the latter case a plot of 1/X$ against T allows evalu— ation of 9 from the intercept. The magnetic moment u of the sample may be calcu— lated from the molar susceptibility by (1 = 2.84 (T x Xr'n)1/2 (30) The low temperature studies were performed by using a Specially constructed Dewar flask similar to that described by Vander Vennen.77 The magnetic moment was measured at room temperature and at 770K. In those cases where a large Change occurred, the moment at 1950K was also determined. 41 F. .Spectroscopic Measurements The infrared spectra of the complexes were obtained by use of Nujol mulls and a Unicam Model SP—200 spectrophotometer (5000 cm-1 to 650 cm-1) and a Beckman IR—7 spectrophotometer (700 cm"1 to 200 cm_1). The visible and ultraviolet spectra were obtained by use of a Unicam Model SP—800 spectro— photometer and a Cary Model 14 spectrophotometer. The visible and ultraviolet spectra were determined in either methylene chloride or acetonitrile by preparing the complexes in solu— tion. Great care was taken in handling the solutions and in filling the cells to prevent hydrolysis. »Reflectance spectra were determined by means of a BauSCh and Lomb Spectronic 600 spectrometer with diffuse reflectance attachment. ‘X-Band electron spin resonance spectra were determined for the pure solids and solutions at 1000K and room tempera— ture. The spectra were determined at frequencies from 9.2 to: 9.5 KHz by using a Varian—4502—O4 spectrometer and were recorded on an x—Y recorder with the X-axis proportional to the magnetic field strength. A Hall probe was used as a field sensor» First derivative curves were recorded. The hyperfine splittings were measured by means of a Hewlett— Packard-524C Frequency counter which was checked against aqueous V0804 (A = 116.13 i 0.2 gauss between fourth and N fifth lines)48 and aqueous K3[Cr(CN)5NO] (A = 5.265 r .05 gauss).48 42 Peak separations were measured by markers, correspond— ing to measured proton frequencies, placed as near each peak as conveniently possible. The magnetic field in gauss was calculated from the following equation: H in gauss = 2.3487465 x v1 x 10“ (31) where v1 is the frequency of the proton resonance at the magnetic field in question. Peak separations, in gauss, were converted to frequencies by the following equation: A(cm—l) 4.668597 x 10‘7 x g x A(gauss) (32) where g is the g value for the set of lines being con— sidered. Values of g were determined from the measured klystron frequency and the field strength. The klystron frequency was determined by means of a TS—148/UP U.S. Navy Spectrum Analyzer which operates on the wave meter principle. The analyzer is calibrated directly in megacycles and the instrument does not lose accuracy at low klystron powers. Although the accuracy is reported to be i2 mc, the pre— cision of the measurements are reported to be :0.5 mc.48 Thus for a series of measurements this method proves to be quite satisfactory. Pitch in KCl (g = 2.0028) and aqueous K3[Cr(CN)5N] (g = 1.9945) were used as standards to check the accuracy of the measured 9 values. Most of the low temperature experiments were conducted employing a Varian V-4547 variable temperature Dewar with liquid nitrogen as coolant. 43 G. .Determination of g and A values from.ESR Spectra ‘During this research, it was necessary to determine 9 and .A values as accurately as possible. Any differences in the spectra of successive complexes was expected to be small. Thus random errors in g and A bad to be eliminated. ~The isotropic g and .A values were determined from the solution spectra at room temperature. The anisotropic terms 9" , gu_, A" , and Al. were determined from the frozen solution (glasses) spectra at either 779K or 1009K. .Since the hyperfine splittings were on the order of 100 gauss, the high field approximation could not be ap- plied. -The perturbation of the Zeeman transitions resulting from the hyperfine interactions was corrected by means of the following equations: hv = gBHo (33) for isotropic 9 Ho = Hm + mI +-§Eie-[I(I + 1)— mi] (34) fOr g" Afi 2 = ___. + — ' 35 for gi- -2 A2 + A II J. _ 2 36 Ho = Hm +Al-mI + ‘ 4Hm ”(I + 1) m1] _( ) where 'Hm is the magnetic field position of the esr line due to the component mI of the nuclear spin I, v is the 44 klystron frequency and (a), A”, and AL are the hyperfine aplitting constants. The corrections are necessarily re- iterative and were carried out by desk calculator. Normally three iterations were suEicient. The hyperfine splitting h . . . t constants were determined from the pOSitions of the 4 and 5th, 3rd and 6th, 2nd and 7th, and 1St and 8th lines where resolution permitted. For example, if the 4th and 5th lines are considered for g isotropic, one obtains: 31A2 Ho = H4,5 + 4H4’5 (37) where H4,5 is the midpoint between the 4th and 5th lines. Similar equations can be written for the other 3 pairs of lines. -The values obtained for H0 from each spectrum (3 or 4 separate values depending on the resolution) were averaged to obtain the final value of H0 that was used to calculate g in the following equation: k g = 0.714489 v /H0 (38) where Vk is the klystron frequency. .The A values reported are the averaged A values obtained by considering either three or four pairs of lines. The estimated errors in the measurements of g, A, g“, A”, ql and AL are given in Table III. 45 Table III. Estimated errors in esr parameters* g i 0.0004 A .i 0.2 g“ 1‘ 0.0002 A” i 0.1 gl 4 0.0005 AJ- i 0.3 * _ _ Hyperfine values given in 10 4 cm 1. -RESULTS AND DISCUSSION A. Preparation of Complexes During the course of this research, two new salts of the hexachlorovanadate(IV) ion as well as various salts of a new class of complexes, the pentachloroalkoxovanadates(IV), were prepared. . AThe tetramethylammonium and tetraethylammonium hexa- chlorovanadate(IV) complexes were prepared by a method some— what different from.that used previously for hexachloro- vanadate(IV) salts. Kilty andNicholls42 prepared four such salts by the reaction of thionyl chloride on various oxotetrachlorovanadate(IV) salts. AFowles and/Walton41 pre— pared several salts in chloroform starting with bis(aceto- nitrile)tetrachlorovanadate(IV). The method employed here was a combination of these two methods. »Thionyl chloride is a very good chlorinating reagent and has the added ad- vantage of eliminating hydrolysis of the complexes. Thus the hexachlorovanadate(IV) salts which were to be used as starting materials for the preparation of the alkoxo com— Plexes could be prepared in high purity. The pentachloroalkoxovanadates(IV) were prepared by a method not yet used in the preparation of alkoxo complexes. 46 47 Almost all examples of alkoxo complexes or compounds were prepared in anhydrous alcohols. As discussed above, this method gave only vanadyl species when employed during this research. The hexachlorovanadate(IV) complexes were known to be soluble and stable in acetonitrile. It was hoped that the hexachlorovanadate(IV) salts would be more stable toward decomposition than the species which forms in alcohols (probably V(OR)2C14=) and that the acetonitrile would also help to stabilize any product (acetonitrile is a less polar solvent than alcohols and it was felt that the polar alcohol solvents probably contributed to the lability of the vanadium species). This method proved to be successful as well as similar preparations employing nitromethane and methylene chloride as solvents. The tetramethylammonium complexes seemed to be more stable than the tetraethylammonium complexes. The pyridinium salts of the hexachlorovanadate(IV) ion and the pentachloro- methoxovanadate(IV) ion were extremely labile toward hydrolysis and could not be completely characterized. B. Magnetic Moments The magnetic moments of the new complexes were deter— mined at room temperature and in some cases at 77°K and 1950K. The results are given in Table IV. Both the hexa— chlorovanadate(IV) and pentachloroalkoxovanadate(IV) complexes exhibit Curie—Weiss paramagnetism. The magnetic moments of the alkoxo complexes show little temperature 48 .haso Hmnfioc %Q Op Uwuuwmmu on HHHB mossomsoo wmmflu meQ0p msfiommoosm CH * BB.H Hx maoflmmoov>mflzommol ms.H x 6Ho>NH26mmoL m>.H mn.H 8H maoflhmmoov>NH24AmmaoVL NB.H 88.H HHH> AHoAnmmoov>8226AmmNovL m>.H w>.H HH> mHoAmmoov>ufl23Ammuov mm.H sm.H ¢>.H H> 6Ho>NH24AmmNUVL 98.H mh.H > mHoAamvoov>Nfl24Ammovl NB.H 88.H >H maoflpmmoov>mfizaflmmoVL ms.H m8.H HHH maofimmmoov>ufl24flmmovl o>.H m>.H HH mHoAnmoov>NH24AmmovL mm.H mm.H m8.H *H 6HU>NH24Ammovl Mosh momma momma .mmxwamsoo A>HVE5HUme> mo mquEoE UHpocmmE .>H CHQMB 49 dependence and their Weiss constants are therefore small. The magnetic moments of the hexachlorovanadate(IV) com— plexes show a strong temperature dependence. In the deter— mination of the Weiss constant, 0 , for the hexachloro— vanadates(IV), the three points do not define a curve ac— curately enough to estimate the value of 9. However, by analogy to previously studied vanadyl complexes, only two points for the alkoxo complexes allow: an estimation of the Weiss constant.80 This is the case if the magnetic moments change only slightly with temperature. The shoulder on the szg ——> 2Eg transition observed in the ultraviolet spectra of the hexachlorovanadate(IV) complexes corresponds to removal of the degeneracy of the ground state by axial distortion to give tetragonal sym— metry. Thus it is possible to employ the calculations out— lined by Figgis82 to obtain k (a measure of the delocaliza— tion of the 3d1 electron onto the ligand), A (the separa— tion between the orbital levels of the szg created by axial distortion of the ligand field), and v (which is A/E where C is the spin orbit coupling constant). The values obtained by this treatment are given in Table V and are quite similar to those obtained by Machin and Murray.81 The value of 160—170 cm'1 for the spin orbit coupling constant indicates that the formal charge on the vanadium in the hexachlorovanadate(IV) complexes is approximately +2.83 50 Table V. Magnetic properties of vanadium(IV) complexes (9°K) C(cm‘l) K 2' v A(cm_1) I 160—170 0.8-0.85 2 320—340 II —6 III -6 IV -4 V —7 VI 160—170 0.8—0.85 2 320—340 VII -3 VIII -4 IX -2 C. Optical Spectra The optical spectra in solution and the reflectance spectra of the complexes were determined and the results are given in Tables VI and VII. The solution spectra were determined in either acetonitrile or methylene chloride. Traces of representative solution spectra as well as plots of representative reflectance spectra are given in Figures 7 to 10. -1 The distorted peak at approximately 15.5 x 103 cm for the hexachlorovanadate(IV) complexes is the 2T2g _) 2E9 transition. The other two maxima are charge transfer bands. -1 The two maxima at approximately 13.5 x 103 cm and 51 Table VI. Optical spectroscopic feabnes of the tetra— methylammonium complexes. (CH3CN) Electronic absorptions Electronic absorptions Complex maxima x 10 cm 1 reflectance (e in parentheses) maxima x 10_3 cm—1 I 13.8 (sh)a 15.2 15.3 ) 19.4 21.2 (330) 22.5 37.5 22.5 II 13.5 (3.0) 13.3 a 14.5 (3.4) 15.5 (vw) 26.1 (1250) 28.6 ~ 41.0 III 13.5 (2.3) 13.4 14.5 (2.6) 15-5 (VW) 26.1 (1070) 26-0 ~ 41.0 IV 13.4 (2.1) 13-6 14.8 (2.3) 15-4 (VW) 26.1 (1070) 26-0 ~ 40.0 v 13.5 (2.0) 13-7 14_4 (2.0) 15.5 (vw) 26.1 (1030) 26-0 ~ 40.0 ash = shoulder; vw = very weak. 52 Table VII. Optical spectroscopic features of tetraethyl— ammonium complexes. CH3CN cnzcl2 Electronic absorptions Electronic absorptions Complex maxima x 10 3 cm 1 reflectance e in parentheses) maxima x 10 3 cm"1 V: 15.4 15.2, 15.5 21-2 17.8 37.3 20.6 ~ 41.0 22.8 VII 13.7 14.3 (8.0) 13.8 14.9 26.0 (2000) 15.5 (vw) 26.0 ~ 37.9 ~ 40.0’ 27.0 VIII 13.3 14.4 (9.7) 14.1 15.0 26.2 (1420) 15.5 (vw) 25.9 ~ 40.0 28.5 (vw) 37.8 IX 13.5 14.5 (9.5) 14.3 14.9 26.3 (1490) 15.5 (vw) 25.9 ~ 40.0 28.5 ~ 37.8 I I I g I ' j I 200 250 300 350 400 mu I I I l l I I A 450 550 650 750 8.30 mu Figure 7 . Solution spectrum of [(CH3 )4N] 2VC15~ I I I 1 j n I \ 200 250 300 350 400 mu m I I I I l I I I 450 550 650 750 850 Figure 8 . Solution spectrum of [ (CH3 )4N12V(OCH3 )C15 . i l.‘ 1 I , l . 1 .. g. g. I 14000 16000 18000 20000 22000 cm Figure 9. A. Reflectance spectrum of [(C2H5)4N]2VC16. B. Reflectance spectrum of [(CH3)4N]2VC16 ——__7 + ________ /\/ M | a | 1 1 11000 15000 _119000 23000 cm 1 27000 Figure 10. A. Reflectance spectrum of [(CH3)4N]2V(OCH3)C15. B. Reflectance spectrum of [(CH3(4N]2V(OC2H5)015, 57 14.5 x 103 cm"1 for the alkoxo complexes are probably the 2 —-> 2E and the 2B2 -> 2B1 transitions respectively B2 (see the molecular orbital diagram for C4v complexes, 1 page 21), The two maxima,approximately 26.0 x 103 cm— and 40,0 x 103 cm_£ are charge transfer bands. The reflec- tance spectra and solution spectra agree well in most cases and thus indicate that the species in Solution is the same species as in the solid. The tetraethylammonium pentachloroalkoxovanadates(Iv) were too unstable in acetonitrile to permit determination of molar absorptivities. The difference between the solu— tion Spectra in acetonitrile and methylene chloride is probably due to solvent effects. In methylene chloride the 2B2 ——> 2E and 2B2 ——> 2B1 transitions may be superposed. D. Infrared Spectra The infrared Spectrum for each of the complexes was determined in Nujol mulls and the results are given in Table VIII. Traces of representative spectra as well as plots of representative far infrared spectra are shown in Figures 11 through 19. The peaks marked with an asterisk in Table VIII are those normally associated with the c—o stretch in alkoxo complexes. The peaks around 250 cm—1 are metal—chlorine vibrations. The peaks at approximately 600 cm—1 observed in the tetramethylammonium alkoxo complexes as in the correct region to be the metal—oxygen stretch but *omOH .omHH mmHH ovNH mMMH onwfi mNmH oowH OM®H XH omfifi mGHH OVNH mmMH nva mNmH oomH OM®H N owm Ohm.omm mmv mow OHOH *omOH mmHH xH 0mm 0mm mmv mow OHOH *ovOH.*owOH mQHH HHH> omm omv mow OHOH *owOH mmHH HH> Ohm.owm.mwm mov.NH¢ mow mHOH mmHH H> omw 00> Ohm *OHOH.*OQOH oomH omvH > % owm.omm ovm mmw mbm.mo> ONQ *omoH OOMH ome >H O¢N www com omm *omOH.*mon omNH omwd HHH OVN wmw mom omm.mmm *OQOH omNH omvfi HH mom mwv omm.®mm omNH omwfi H aIEo .mEmeE COHuQMOQO UmumumcH xmamfioo .mwxwamsoo A>Hvfisflwmsm> mo wwusumww UHQOUmOHumem @meanH .HHH> CHASE 59 ) I l I I I I I I l l 5000 3000 2000 1800 1400 1000 800 2000 cm"1 Figure 11. Infrared spectrum of [(CH3)4N]2VC16. I I l I I l I L . 5000 4000 2000 1800 _1 1400 1000 800 2000 cm I Figure 12. Infrared spectrum of [(CH3)4N]2V(OCH3)C15- L l I I I I 1 l J 4000 2000 1800 _ 1400 1000 2000 cm 1 Figure 13. Infrared spectrum of [(CH3)4N]2V(OC2H5)C15. I I 3 I . l 4000 2000 1800 2000 cm I I 1 1400 1000 Figure 14. Infrared spectrum of [(CH3)4N]2V(O—nC3H7)C15. 61 l l | I l l I I I. I L 4000 2000 1800 _1 1400 1000 2000 cm Figure 15. Infrared spectrum of [(CH3)4N]2V(O-EC4H9)C15. M I I 4000 2000 1800 _1 1400 1000 2000 cm Figure 16. Infrared spectrum of [(C2H5)4N]2VC16. 62 4W K; 4‘ I ‘ I 4000 2000 1800 2000 L A + I 1400 1000 -J. cm Figure 17. Infrared spectrum of (C5H6N»2VC16. «WW I I I I I J I I I I 4000 2000 1800 _1 1400 1000 2000 cm Figure 18. Infrared spectrum of (C5H6N)2V(OCH3)C15. 63 .AOTDZ Ca u:O\mmUOV>mHZwA ”mo H - EUHMUmmm wwmwumcfl Hmm .U oaorSZ GA wHU>mHZw mmU H we Eflurommm meMHwQfl Hum .HOwHZ mo 51.Huooau wwmmmwia Ham .4 .mH musmflm +ILU OON Com 00¢ ocm com 00% N - a a n q 64 no corresponding peaks were observed in the tetraethyl- ammonium complexes. E. Electron §pin Resonance Spectra ,The esr spectra were determined for solutions and solids at room temperature and at 1009K (glasses) for each of the complexes. Acetonitrile was used as a solvent for both the tetramethylammonium and the tetraethylammonium complexes while nitromethane,was also used as a solvent for the tetraethylammonium complexes. The results are shown in Tables IX and X and traces of representative spectra are given in Figures 20 through 24. Figure 25 shows a superposition of the esr spectra of [(CH3)4N]2V(OCH3)C15 and [(CH3)4N]2V(O—QC3H7)C15 where the differences in A” and AL are small but very evident. No esr signal could be detected for the hexachloro— vanadate(IV) complexes in either the solid state of in thionyl chloride solutions at 2950K or 770K. In fact the metal hyperfine structure of the alkoxo complexes could be observed in diluted powders of the pentachloroalkoxovana- dates(IV) in the hexachlorovanadates(IV) at room tempera— ture. Thus the hexachlorovanadates(IV) acted as a dia— magnetic host but these spectra were not investigated further. Vanadium metal hyperfine was observed in both the solution and frozen solution spectra but no chlorine super— hyperfine strucume was observed. The solution spectra con— sisted of 8 lines (I = 1) while the frozen solution spectra 2 could be resolved into parallel and perpendicular components. 65 «a EU OH an cm>flm mum mmcfluwflamm wcflwuwmmm I * «ima NONQH méuw manqfi N.©®H manqfi w.©m Hwom.H mwm.H > miwm mmom.H méum omnm.fi m.®mH mfimmé 0.th mamm.H owm.H >H . N45 OHbmJ“ 9mm mwbm.fi 0.5m: 269a dbm omwm.H oma.H HHH N203” mwmm.H ©.®® vmba.fl 99: whvm.fi v.56 wwwm.H mvm.H HH oamofl Damom |1< .fiw __< :m 4 Amv UflHOmm memEoo *mmwaQEOU Esflqo=:emam£uwammump How muwuwfimumm Hmm .xH manma 66 ®.OOH mmwm.H HI EU v OH GH Qw>Hm mum mmcfluuflamm mcflwnwmhm * fi.mo «msm.fi m.HsH nova.fi N.nm hwa.H mvm.H xH o.HOH mmom.a ¢.mo nmum.a m.msH ofivm.fi w.>m mwmm.H mvm.a HHH> m.fi0H wfimm.fl o.om fivbm.fi >.N>H sawm.fi m.>m mwmm.H ovm.H HH> «02mmo qH m.wm «woo.fi w.mm nunm.a w.smfi wmvm.H p.mm mama.H mvm.H xH m.ooa ssom.a u.wo umnm.fi m.mmH wmwm.fi w.mm omom.H mvm.H HHH> H.HOH ommm.H H.mw pupa.” o.mua mmwm.a o.um umwm.fi mvm.H HH> Zommo qH oamod oamom .4¢ 4w =< :m < Amv UHHowm xwamfioo *mmxmamfioo Esflsofifimamnuwmnump How mequMHmm Hum .x wanna ml 6 u 1!. nflo.nmoov>wHZ$fimmUVH Uflaom wc Eonuommm G .MOOOM ocmcommu Seam nonpomam .ON muswem tavfim carmcmmfi mdflwmmnucfl Manon um 20mg cu maoflmmoovskmzflmmovr mm Euwuemmm moamcowwu cflmm GOHuomam .HN h wampw oeumqmmfi mCemmmHUGH fl, d 69 .xooo ,, «u H ,.r um 20mmo cu .Hoflmmoov>NH2whmmov A, H.mmrmomvm« m0 Ermuowmm wCHH _ m mofiMQOmGH cam mumoemmfl » x: U gonuowam .NN ouzmflm \ Raga pamflw Usumuvmfi mcflmmwuoca be“ cuflH.fl meMOHUCH « wsfla ~_ wwwmoflpsfl A .MoooH mm ZUmmo CH mHUAnmmUOV>m~ZwAmmUVH mo monmdomwu Qflmm conuomam .mN musmflm eeeeeehe IIIIHIIIII .xocoH um ZUmEU as mHUA. QCAH fi.r¢ ammomuovsrm Eu; «mo: Hr uHUHUCH w mGHH __ nmumwflpflfl 9 O Ezuuoram occmcommh cflmm couuomam .VN CHSWHm 71 eeeee pamam ospoqmmfi mcemmmuoce bee 72 increasing magnetic field 3 I .2 x.- .. .I DO ~I :. b 4 ‘ . . OJ :1 ‘. :- :." .1. I. .0 I :a'. " I':. :3; 0". 5': E? :33 .0 :0 .0 90“ 5’." I...~jl.L—u A B Figure 25. Electron spin resonance spectrum of 5(CH3)4N]2V(OCH )Cl5 (A) at 1000K and [(CH3(4N]2V(O-£03H7 (:15 (B at 1003K superposed to Show differences. ....,... ..... I..'........OQ'ODOO.0000O... Iii-ta": I II II 73 Various trends can be observed in the esr parameters in going from the tetramethylammonium pentachloromethoxo- vanadate(IV) to the pentachloro(nfbutoxo)vanadate(IV). A" . Al , and A all decrease through the series while 9” increases and QL seems to decrease. No trends are ap— parent in the isotropic (9) value. The tetraethylam— monium salts show similar trends. The same trend is seen in CH3CN and CH3N02. Kuska“8 studied a series of vanadyl complexes and found an inverse relationship between g” and and A” similar to that observed here. Figure 26 shows a plot of gll gs A” for the tetramethylammonium alkoxo complexes. F. Calculation of Molecular Orbital Parameters The observed g and A values were used to calcu— late the coefficients in a simple molecular orbital scheme. The procedure was similar to that of Kuska and Rogers ex- cept contributions from charge transfer bonds were not included.5 2 2 . In order to solve for N , N , and N2 in equa— W2 02 7T1 tions 21, 22, and 23 above, values must be obtained for S C. P, T(n), S , and Se' The values of C and P b2' b1 depend on the formal charge assigned to the vanadium atom. In calculations involving the V0++ unit, most authors feel the charge is reduced to approximately +2. It would seem reasonable then to assume the charge in the similar n m .o .mmuo01 u m .m .mMUO: n m .« .Afiusu wuofi E s muoxmoruflzimmo: H8 a 3 a no 83¢ .2 was: mumm.H mamm.fi momm.H mmwm.fi mmwm.fi mswm.a . a a . u a a a 4 1 av . sea a 1 sea “0 4 7 I sea 4. I mo“ 75 ++ alkoxo group V(OR) + formal charge does not have to be an integer, +3 was used. Thus4 P taken as 210 cm-1 the ligand sigma orbital hybridization. the chlorine sigma bond is an T(n) was estimated as 0.25. and Se is somewhat larger. is 1.50 x 10‘ from Dunn's77 tables. sp hybrid, The overlap terms S While the the value of 2 —1 CH1 and C was T(n) depends on In this case where the value of b2 I Sbll were not determined but should be very near those values calculated by Gutowsky3 for the tetrachloro(oxo)vana— date(IV) ion. Thus S S b2, b1 0.165, and 0.139 respectively. , and Se were taken as 0.099, The assumption was made that little change occurs in these values in going from the methoxo to the n—butoxo complex. a ~ e XVI = XVI. It was also assumed that The values obtained by an iterative treatment of equa- tions 21, 22, and 23 are given in Table XI. are strongly dependent on the trends A decrease is observed in N: and 2 group becomes larger and an increase increase in N2 7T2 ”lB1>* 2 . . and No indicates 2 The calculations in All' gll’ and gl_ . 2 N as the alkoxo 02 occurs in N2 . An 7T1 that the |B2>* and molecular orbitals become more covalent through the . . . 2 . . series. The increase in NW indicates the |E>* molecu— 1 lar orbital becomes more ionic from the methoxo to the n: butoxo complex. Similar trends would result from a treat— ment of the data obtained from CH3N02 solutions. 76 Thus it would appear that as the alkoxyl group becomes longer, electron density flows toward the metal in the bond involving the alkoxo group, the metal, and the axial chlorine. Also electron density flows from the metal to the equatorial chlorines in both sigma and .pi bonds in the same series. The unpaired electron density in the four equatorial 3p orbitals of chlorine is given by equation 25. The F values of the spin densities are given in Table XI. The values vary as would be expected. The spin densities in— crease in going from the methoxo to the n—butoxo complex. Monoharan and Rogers2 and Dalton, gt_3l.1 have found excel— lent agreement between the spin densities calculated in this manner and those calculated from ligand hyperfine splittings. This work can be compared to the trends found in the molecular orbital parameters for MoOX5= as X = F-, C1—, or Br_.1l2 Table XII gives the coefficients found in a similar molecular orbital approach for these complexes. As the ligand X is changed from F— to C1_ to Br—, the electron density on the halide becomes larger in those molecular orbitals involving the metal dx2_y2 and the d or dyx orbitals. No trend is observed in the coef— XZ ficients of the molecular orbital involving the dx2_y2 metal orbital. As would be expected these changes are much larger than similar changes in the alkoxo series. 77 Table XI. Molecular orbital coefficients and spin densities. for V(OR)C15 2 complexes. 2 2 2 Complex sz NV1 N02 AFZ %fw II 0.947 0.882 0.794 0.506 4.87 III 0.934 0.910 0.767 0.529 5.31 IV 0.921 0.924 0.751 0.552 5.75 V 0.919 0.924 0.739 0.555 5.81 VII 0.982 0.914 0.811 0.438 3.64 VIII 0.961 0.949 0.767 0.481 4.41 IX 0.944 0.955 0.765 0.512 4.99 _ 2 Table XII. Molecular orbital coefficients for MoOX5'.’ X N N N 7T2 7T1 02 F"1 0.956 0.960 0.891 01'1 0.905 0.901 0.754 Br"1 0.932 0.844 0.596 SUMMARY The object of this research was to prepare and char— acterize a series of pentachloroalkoxovanadates(IV) and to study their esr spectra in detail. It was hoped that trends would be found in the esr parameters which could be related to the molecular orbital parameters. Little was known about trends in the molecular orbital coefficients until recently.1 With the large number of electron spin resonance studies which have been made on transition metal complexes, it was necessary to understand these trends. 78 PART II THE INVESTIGATION OF THE APPARENT THERMOCHROMISM OF SOME VANADIUM(III) COMPLEXES 79 INTRODUCTION AND HISTORICAL A. Thermochromism During the investigation of the vanadium(IV) complexes described in Part I of this thesis, two vanadium(III) complexes were prepared. One of these complexes, [(C2H5)4N]VBr4-2CH3CN, possessed the unusual property of changing color with temperature change. The second van— adium(III) complex which was prepared, [(C2H5)4N]VCl4-2CH3CN, did not apparently change color with temperature change. It was interesting that the tetraethylammonium bis(aceto— nitrile)tetrabromovanadate(III) complex appeared to be the same color at 77°K as the tetraethylammonium bis(aceto— nitrile)tetrachlorovanadate(III) complex at room temperature. Thus the effect of warming the tetrabromo complex was equiv— alent to substituting Cl— for Br_ in the equatorial posi- tion in the coordination shell. For this reason, an at- tempt was made to prepare all the intermediate complexes containing Cl- and Br_ with the hope that their properties might elucidate the thermochromic mechanism. Thermochromism is defined as the reversible change in color of a compound when it is heated or cooled.1r2 The thermochromic color change is distinguished by being 80 81 noticeable and sometimes occurring over a narrow temperature range. For inorganic complexes this transition in color may result from a change in the crystalline phase, change in ligand geometry, or to a change in the number of solvent molecules in the coordination sphere. For those complexes with a sharp transition, the thermochromic temperature may be changed by dispersing the compound into solid matrices such as paraffin waxes or oils. Houston3 first examined the property of thermochromism by heating solids on copper strips over a bunsen burner. Day's1 review article adequately reviews most inorganic compounds which are known to be thermochromic. Thermochromism is believed to be a general property of chromium(III) compounds. The continuous thermochromic transition is from red to violet to green as the temperature is raised. Poole4 found that the color change which oc— curred in mixed oxides of chromium(III) with aluminum, lanthanum-gallium, lanthanum—gallium-aluminum, and yittrium— aluminum was a consequence of the lattice expansion which occurs when the sample is heated. Theoretical and electron spin resonance studies have shown that the chromium ion occupies octahedral or quasi—octahedral sites exclusively and that the color changes depend on the distance to the neighboring central ions. .The structure of the thermochromic complex AgZHgI4 has been studied as a function of temperature by several 82 physical techniques.5r7 It was found that variations in various physical properties could be correlated to thermo— chromic temperature changes. No thermochromic complexes of vanadium(III) are known. A striking apparent thermochromism of liquid VOClé is due to traces of water or even hydroxyl groups on the glass containers.8 Dry VOC13 which contains traces of water is yellow,but at -700 is bright red. A solution of the com— pound in methylene chloride is orange, again due to traces of water, and is thermochromic due to the narrowing of a strong absorption band which extends into the visible region. There are no distinct bands in the ultraviolet but continu— ously increasing absorption. The absorbing species has not been identified, but it is possibly a mixture of partially hydrolyzed products. Thermochromism in liquids is well known.1 Most of these cases either involve simple equilibria between sol— vent molecules and ligand groups or a change in coordina— tion number from 4 to 6.9I10 There are in general two broad classes of thermo— chromism. The first is a gradual deepening in color with rising temperature and occurs for a great many substances. The complexes investigated in this thesis would belong to this class. However, Day notes that this property is seldom reported.1 The second is a drastic color change over a very limited temperature range. In the case of solids the temperature range may be dependent on the rate of heating. 83 Thermochromism in many inorganic complexes depends on the position of charge transfer bands. In many transition element complexes the wave—lengths of the first strong (charge transfer) bands vary in an interesting way with the nature of the ligands. For example, in the Series of halide complexes, [Co(NH'3')5X]+2 the fluoride begins to absorb strongly in the far—ultraviolet at about the same place as the [Co(NH3)6]+3 ion itself (50,000 cm—l), but the chloride (45,000 cm—l), bromide (37,000 cm_1), and iodide (32,000 cm-l) have strong bands at progressively longer wave—lengths and finally in the iodide these largely obscure the weaker d—d transitions.11 Thus the strong absorption bands move to longer wave—lengths as the ligand becomes more easily oxidized. Many colorimetric analytical reagents used for detecting transition element ions are in fact ligands which form complexes having strong charge transfer bands. B. Vanadiumglllz Complexes Vanadium(III) chemistry has been investigated quite extensively and vanadium(III) is known to form cationic, neutral, and anionic complexes. Nicholl's12 review of vanadium chemistry covers these areas and only complexes of interest to this research will be discussed here. Vanadium(III) halides react with donor molecules to form six—coordinate adducts. Examples of such complexes are given in Table 1. Little is known concerning most of 84 these complexes except those compounds with nitriles as ligands, which may be prepared by causing the halide to reflux in the solvent. The alkyl cyanide complexes can also be prepared by the direct reaction of the nitriles with vanadium(IV) chloride. Table I. Vanadium(III) nitrile adducts VCls '3C2H5OH VC13 '3RCN R = CH3_I C2H5_l C3H7" VBr3-3RCN R = CH3—, csz- VCla ‘3C5H5N The diffuse reflectance spectra of [VC13°3CH3CN] and the absorption spectrum ofla solution of vanadium(III) chloride in acetonitrile are almost identical.13/14 These compounds are non-«electrolytes in the parent nitriles and are formulated, 24g., as [VC13'CH3CN]°. Two peaks are observed in the visible region for this complex. The maxima at 14,400 cm—l has been assigned to the 3Tlg(F)-—> 1 3T2g(F) transition and the shoulder around 21,000 cm_ to the 3 (F) —> 3T1g(P) transition. The corresponding T1g bromo—compound shows only the first ligand field peak while a strong charge transfer band obscures the second peak. Both the tetraphenylarsonium and the tetraethylammon- ium salts of the octahedral bis(acetonitrile)tetrachloro- vanadate(III) complex have been prepared.15r16 The 85 tetraphenylarsonium complex shows two peaks in the visible spectrum at 13,500 cm—1 and 20,400 cm—l. Several salts of the octahedral bis(acetonitrile)tetrabromovanadate(III) complex have also been isolated.16 One mixed octahedral complex, tetraethylammonium bis(acetonitrile)bromotri— chlorovanadate(III) has been prepared.16 All these com— plexes lose acetonitrile when heated to 1000 under vacuum to give tetrahedral vanadium(III) complexes. While the tetrahedral complexes have received some attention, little is known about the octahedral complexes. EXPERIMENTAL A. Materials Tetraethylammonium hexachlorovanadate(IV).— The preparation of tetraethylammonium hexachlorovanadate(IV) is explained in Part I of this thesis. Vanadium(III) chloride.17—— Vanadium(III) chloride was prepared by the reaction of sulfur monochloride (Szclz) with vanadium pentoxide. The vanadium pentoxide was ob— tained from K and K Laboratories and the sulfur monochlor— ide was obtained from Eastman Organic Chemicals. Fine, pure V205 powder (18 g) and 40 ml of Szclz were caused to reflux under anhydrous conditions for 8 hours (constant stirring). The excess Szclz, containing dissolved S, was decanted and the VC13 was washed 10 times with CS2 to re- move adhering sulfur. .The product was then dried at 120° under vacuum. The yield was about 27 g. Tetraethylammonium halides.- Tetraethylammonium chlor— ide and tetraethylammonium bromide were obtained from Eastman Organic Chemicals and were dried at 80° before using. Ethanethiol.— Ethanethiol was obtained from Eastman Organic Chemicals. 86 87 Hydrogen Bromide and Nitrogen.- Anhydrous hydrogen bromide was obtained from Matheson Chemical Company. Pure nitrogen was obtained in the same manner as described in Part I of this thesis. Solvents.— 'Acetonitrile and ethyl ether were puri— fied in the same manner as described in Part I of this thesis. B. Analytical Methods All analyses were performed in the same manner as described in Part I of this thesis. C. General Experimental Procedure All experimental procedures were the same as described in Part I of this thesis. D. Preparation of Compounds Tetraethylammonium bis(acetonitrile)tetrachlorovana- date(III).- This complex could be prepared by two methods. Method 1. To a known amount of tetraethylammonium hexachloro— vanadate(IV) in acetonitrile, one equivalent of ethanethiol was added. The color changed from red brown to yellow im— mediately. The mixture was stirred for 1 hour and filtered. The product was recrystallized several times from boiling acetonitrile and dried under vacuum. 88 Anal. Calcd for VC14C12H26N3: v, 12.58;.01, 35.01; c, 35.58; H, 6.47; N, 10.37. Found: v, 12.32; c1, 35.41; c, 35.51; H, 6.47; N, 10.17. M- A solution of vanadium trichloride in acetonitrile was treated with one equivalent of tetraethylammonium chloride. The green solution became bright blue on the addition of the tetraethylammonium chloride, and when cooled became violet. The complex precipitated as a yellow powder and was recrystallized twice from boiling acetonitrile. This complex has been prepared by Clark et al.16 by the same Method 2 described. Anal. Found: V, 12.38; Cl, 35.37. Tetraethylammonium bis(acetonitrile)bromotrichloro— vanadate(III).- One equivalent of tetraethylammonium bromide was added to a slurry of vanadium(III) chloride in warm acetonitrile. The mixture was heated and stirred for one hour and cooled. The yellow—orange product was recrys- tallized twice from boiling acetonitrile and dried under vacuum. This complex was previously prepared by Clark gt al.16 in the same manner as described here. Anal. Calcd for VBrC13C12H26N3: V, 11.33; Br, 17.78; Cl, 23.66; C, 32.05; H, 5.83; N, 9.35. Found: V, 11.17; Br, 17.99; Cl, 24.03; C, 31.91; H, 5.887 N, 9.30. 89 Tetraethylammonium bis(acetonitrile)dibrgmodichloro— vanadate(III).- No pure material with this composition could be obtained but products whose composition were near this could be obtained by two methods. M_et_hp_d_1- If the bright red product whose composition was near that of tetraethylammonium bis(acetonitrile)tribromochloro— vanadate(III) was washed well with warm acetonitrile, an orange product was obtained which gave an analysis near that expected for the dibromodichloro product. Anal. Calcd for VBr2C12C12H26N3: V) 1031; Br, 32.35; Cl, 14.35. Found: V, 10.14; 10.27; Br, 32.08, 31.74; Cl, 13.71, 13.83. regress- If one equivalent of tetramethylammonium bis(aceto— nitrile)tetrabromovanadate(III) was mixed with one equiv— alent of tetraethylammonium bis(acetonitrile)tetrachloro- vanadate(III) in a small amount of acetonitrile, an orange product could be obtained. This product gave analyses similar to those above. Tetraethylammonium bis(acetonitrfle)tribromochloro— vanadate(III).— No pure product with this composition could be obtained but a product whose composition was near this could be produced by bubbling anhydrous HBr into a slurry of tetraethylammonium hexachlorovanadate(IV) in acetonitrile. 90 After 10 minutes, the HBr gas flow was removed and ethyl ether added. -A bright red precipitate was obtained which was filtered, washed with ether, and dried under vacuum. Anal. Calcd for VBr3ClC12H26N3: v, 9.96,- Br, 44.52,- Cl, 6.53. Found: V, 9.37; Br, 43.97; Cl, 6.44. Tetraethylammonium bis(acetonitrile)tetrabromovana- date(III).— If the red solid obtained above was dissolved in hot HBr saturated acetonitrile, large red—brown crys— tals of the product could be obtained. This complex has been prepared by Clark, et al.16 by the addition of tetra— ethylammonium bromide to vanadium(III) bromide in aceto— nitrile. Anal. Calcd for VBr4C12H24N3: V, 8.74; Br, 54.83; C, 24.72; H, 4.50; N, 7.21. Found: V, 8.75; Br, 55.01; C, 24.45; H, 4.46; N, 7.18. E. -Magnetic Moment Measurements The magnetic moments of the complexes were determined in the same manner as described in Part I of this thesis. F. Spectroscopic Measurements The infrared spectra were obtained by use of Nujol mulls and a Perkin—Elmer Model 457 Spectrophotometer (4000 cm—1 to 250 cm_1). The ultraviolet—visible spectra were obtained by use of Nujol mulls and a Unicam Model SP—800 spectrophotometer and a Cary Model 14 spectrophoto- meter. Both infrared and uv—visible spectra on the Cary 91 Model 14 spectrophotometer were obtained at 770K and 1959K by means of a specially constructed cell (See Figure 1). The cell was constructed so that it could be used for either low temperature uv—visible absorption spectra, low tempera— ture reflectance spectra, or low temperature infrared spectra depending on the cell windows employed. The sample was ground in Nujol and pressed between two plates (quartz for uv—visible or reflectance spectra and cesium iodide for infrared spectra). The plates were tightened between the two metal plates of the sample holder (A). The sample holder (A) is connected by a kovar seal to the Dewar. The lower half of the cell could be evacuated (B) to prevent condensation of moisture onto the cell windows or sample plates. The external cell windows (C) could be varied again according to the area of the spectrum being investi- gated. The cell base and Dewar portion of the cell are connected by an O—ring seal so that the cell may be dis— assembled easily. The nuclear quadrupole resonance spectra were deter; mined with the aid of Dr. E. Carlson, Department of Physics, Michigan State University, on a superregenerative spectrometer similar to that of Dean.18 All frequencies were measured with a Hewlet-Packard frequency counter. A Princeton Ap— plied Research lock—in amplifier was used for derivative detection. Figure 1. Low temperature cell for infrared and uv—visible spectra. RESULTS AND DISCUSSION A. Preparation of Compounds During the preparation of the complexes described in this thesis, it became quite evident that the mixed halo complexes tend to disproportionate. For example, if one starts with vanadium(III) bromide and tetraethylammonium chloride in acetonitrile, the products isolated are tetra- ethylammonium bis(acetonitrile)tetrabromovanadate(III) and impure tetraethylammonium bis(acetonitrile)dibromodichloro- vanadate(III). The reactions which occur in solution are probably: < VBr3'3CH3CN + (C2H5)4Ncl :—-> [(C2H5)4N]VBr3Cl'2CH3CN + CH3CN 2[(C2H5)4N]VBr3Cl'2CH3CN <_ > [(C2H5)4N]VBr4'2CH3CN + [(C2H5)4N]VBr2C12'2CH3CN Either the disproportionation is very rapid or the equié librium lies far to the right for the complex with 3 Br's_ and 1 C1—. The disproportionation is less apparent for the complex with 2Cl's- and 2Br's_ and almost nonexistent for the complex with 3Cl's— and 1 Br—. The disproportionation 93 94 process can become an advantage in preparing the mixed complexes. 'For example, if HBr is bubbled into a slurry of the tetraethylammonium bis(acetonitrile)tetrachlorovanadate(III) complex in acetonitrile at room temperature, apparently only three chlorides are replaced. If ethyl ether is added, the total solvent becomes much less polar and the dispropor- tionation is stopped so that the complex with 3Brfi's and 1 Cl_ can be isolated. Of course, this complex cannot be recrystallized in acetonitrile because it would then dis— proportionate. However, if small amounts of warm aceto— nitrile are added, it is possible to obtain the complex with 2Cl_‘s and ZBr-'s because the disproportionation pro— ducts differ greatly in solubility. The tetrabromo complex is much more soluble and is washed out while the dibromo— dichloro complex is less soluble and is left behind. The fact that the uv—visible spectrum of each of these "impure" complexes is different and consists of one narrow band in the visible region tends to support the belief that the products are fairly pure. B. Magnetic Moments The magnetic moments for each of the complexes were determined at room temperature. The magnetic moments for the tetraethylammonium bis(acetonitrile)tetrachlorovanadate- (III) and tetraethylammonium bis(acetonitrile)tetrabromo— vanadate(III) were also determined at 1950K and 77°K. The results are listed in Table II. 95 Table II. Magnetic moments of vanadium(III) complexes Complex 3009K 1950K 77°K [(CH3CH2 )4N]VBr4 -2CH3CN 2 .72 2 .67 2 .58 [(CH3CH2 )4N]VBr3Cl '2CH3CN 2 .73 [(CH3CH2 )4N] VBr2C12 -2CH3CN 2.73 [(CH3CH2 )4N1VBrCl3 '2CH3CN 2 .75 [(CH3CH2)4N]VC14 '2CH3CN 2.75 2 .71 2 .66 It was hoped that the magnetic behavior of the tetra— bromo complex would be quite different from that of the tetrachloro complex——that is not the case. All the com— plexes exhibit temperature dependent paramagnetism with the moments being near that for a spin only d2 system. The three points for the tetrabromo and tetrachloro com— plexes do not define a line well enough to allow an esti— mation of the Weiss Constants. C. Infrared Spectra Little or no differences were observed in the infra— red spectra obtained at room temperature and the spectra obtained at 77°K for any of the complexes. Figures 2—7 Show the spectra obtained for the complexes. The features of the infrared spectra are listed in Tables III and IV. along with possible aSsignments of the peaks. Tables V and VI give the vibrational modes and the description of the 96 1400 1200 em“‘1000 800 600 400 Figure 2. infrared spectrum: of (C2H5)4NVCl4-2CH3CN (77°K). I I I I I 1400 1200 1000 cm-1 800 600 400 Figure 3. Infrared spectrum of (02H5)4NVBrC13-2CHSCN (77°K). 97 I I l 41 J; 1400 1200 1000 Tm~l 800 600 400 Figure 4. Infrared spectrum of [(C2H5)4N]VClBr3'2CH3CN (77°K). ((0 m I I I I I I 1400 1200 1000 cm—1 800 600 400 Figure 5. Infrared spectrum [(C2H5)4N]VCIZBr2-2CH3CN (77°K). 98 I 1400 Figure 6 1 I 1200 1000 ID frared spectrum 0 i‘ I I m\—i 800 600 L 400 [(212115 )4N] VBI.‘4 '2CH3CN (3000K) o I 1400 Figure 7. ,1_Jr l ‘L I 1200 1000 cm_1 800’ 600 Infrared spectrum of [(C2H5)4N]VBr4-2CH3CN (77°K). I 400 99 Table III. Features of the far infrared Spectra of the vanadium(III) complexes Complex M—Cl M-Br M—N [(C2H5 )4N] v13r4 '2CH3CN I . 275-285 420 [(C2H5)4N]VBr3Cl°2CH3CN II . 320 260-280(sh) 418 [(C2H5 )4N] VBr2C12 -2CH3CN,III 330(broad) 270 418 [(C2H5 )4N] VBrCl3 ~2CH3CN IV 330(broad) 270(weak) 420 [(C2H5)4N]VC14 20113qu v 335(broad) 420 modes expected for tetraethylammonium chloride and aceto- nitrile respectively. In a complex with D symmetry, there should be two 4h metal—ligand vibrations due to equatorial ligands. In this case the metal-bromine vibrations should lie lower in energy than the metal—chlorine vibrations. With this in- formation and the changes in the far—infrared spectra from complex to complex, it is possible to assign most of the metal-ligand vibrations. D. Nuclear Quadrupole Resonance Spectra Nuclear quadrupole resonance has been used extensively to study phase transitions in inorganic solids. The transi— tion points of R2MX6 type complexes, located by the tempera— ture dependence of the NQR signal, are given in Table VII. At a transition point the resonance frequency may move to 100 Table IV. Features of infrared spectra of the vanadium(III) complexes I II III IV va 2740(vw) —— 2740(w ) 2740(vw) 2740(vw) v3 + v7 2405(w) —- 2400(w) 2415(vw) 2410(vw) v3 + v4 2320(s) 2320 2320(m) 2325(m) 2330(w) v2 2290(5) 2290( ) 2290(s ) 2290(s) 2290(s) 2V4 + v8 2245(vw) —— —— —— 2245(vw) 1490(vw) —— —— 1490(sh) 1490(sh) 1470(vw) —- 1470(sh) 1470(sh) —— .v6 1390(w) —— 1390(m) 1390(m) 1395(m) v3 1360(w ) —- 1360(m) 1360(m) 1365(m) 1190(m) 1190(sh) 1190(sh) 1190(sh) 1190(Sh) (C2H5)4N+ 1175(s) 1170(5) 1170(s) 1170(s ) 1170(s) 1090(w ) 1090(w) 1090(m) 1090(w) 1090(w) 1065(w) 1070(w) 1070(m ) 1070(w) 1070(m) v7 1050(s) 1050(m ) 1050(m) 1050(m) 1050(m) 1030(w ) 1030(vw) -- —— -- (C2H5)4N+ 995(5) 990(w) 1000(s) 1000(s) 1000(s) v4 950(m) 950(5) 950(m) 9500.) 950(m) 900(w) 900(w) 900(vw) 900(Vw) 900(w) 885(w) 885(w) 885(vw) —— —— 2V3 820(m) 830(m ) 840(m) 820(m) 820(m) (C2H5)4N+ 790(s) 780(s) 790(s) 790(5) 790(5) 370(s) —— —— —— —— 3200.) a . Numbers refer to complexes in Table III. 101 Table V.18 Infrared peaks of (C2H5)4NC1. 1195 1019 805 467 425 392 350 Table VI.19 Infrared peaks of CH3CN. F6223? 32:32; vs (E) C-CEN bend. 380(m) 380-420 2v3 (A1) overtone 750(m) 750-825 v4 (A1) C-C str. 920(m) 924—980 v7 (E) CH3 rock. 1040(8)‘ 1038-1025 v3 (A1) CH3 def. 1376 (s) 1374-1355 v6 (E) CH3 def . 1442 (s) 2V4 + v3(A1) comb. 2208(vw) 2215-2285 v2 (A1) CEN str. 2257(m) 2266-2325 v3 + v4(A1) comb. 2297(m) 2300-2355 V3 + v7(E) comb. 2412(vw) v2 + v8(E) comb. v1 (A1) C-H str . v5 (E) C—H str. v2 + V4(A1) comb. abend. = bending mode; str. = stretching mode; rock. = rock— ing mode; def. = deformation mode; comb. = combination band. b s = strong; m = medium; w = weak; v = very. C . . '71 FrequenCies in cm 102 much higher or lower frequency or the resonance may disap— pear altogether. Table VII. Transition points observed by NQR spectroscopy. Compound Transition Point Reference (°C> KZSeBrG -64,-52,-33 21 (NH4)2TeBr6 —52 22 szTeIG -40,—16, 55 23 K28n16 —8.5 24 (NH4)2PtBr6 0.5 23 K2ReBr6 -27,—16,-4 25,26 K2Re16 166 26 A chlorine nuclear quadrupole resonance has been ob— served for VCl3 at 9.40 MHz at room temperature.27 In this investigation only the tetraethylammonium bis(aceto— nitrile)tetrachlorovanadate(III) complex gave a resonance signal. Two signals were observed at 9.45 and 9.60 MHz which indicated two chlorine environments in the solid state. The observed spectrum is given in Figure 8. If the signals which were observed at 77°K were followed as the sample was warmed to room temperature only a decrease in intensity and a gradual shift of the two Signals to (9.49 and 9.62 MHz) was observed. Since no major shift in the signal took place, there is probably no phase transition for 103 || 0 .musmfisoufl>so usmuomwflp OBD How Hmcmsm HUum .m usmfisouw>so mcflaamummuo sH mEoum HUmmN How Hmcmflm H0 n m .m Damfisonfl>sm mcflaamummuo CH mEOum HUmmN How Hmcmflm H0 H < AME: CH moflosmsqmnhv .Mobb um ZUmEUN.wHU>ZvAmmmov mo Eduuowmm OUCMGOmmH waomsnomsv HmmHoss HO .w ousmflm 104 07' 669'6 897's 619'6 IIL'L :0 fl 709'6 Figure 8. \ 5 .~ In .;:f‘fl.',k.. "any” . - . .r ' -.. . ....‘ . .. 2"... ., -\ w . -..~ ~ ...v(~.. ., 105 tetraethylammonium bis(acetonitrile)tetrachlorovanadate(III) between 77°K and room temperature. E. Optical Spectra The uv—visible spectrum of each of the complexes was determined as an absorption spectrum of the Nujol mull of the solid. The results are given in Table VIII and the Spectra are shown in Figures 9—13. There is a trend in 3T1 (F) ——> 3ng(F) transition from a lower to a higher 9 energy in going from the tetraethylammonium bis(acetonitrile)— tetrabromovanadate(III) complex to the tetraethylammonium bis(acetonitrileketrachlorovanadate(III) complex. The 3Tlg(F) ——> 3Tig(P) was observed only in the tetraethyl— ammonium bis(acetonitrile)tetrachlorovanadate(III) complex because the broad shoulder in all the other complexes due to the low lying charge transfer band obscures the transi— tion. The smooth variation in color from the light yellow tetrachloro complex to the red—brown tetrabromo complex is due to the position of the first charge transfer band. A peak near 20,000 cm"1 will produce a red—bIOWn color while no peaks in this area result in a light color depending on the position of the d—d transition. This color can be compared to the red—black hexachlorovanadate(IV) and golden pentachloroalkoxovanadate(IV) species investigated in Part I of this thesis. 106 Table VIII. Optical spectroscopic features of vanadium(III) complexes. (bands in cm 1 Complex 3Tlg(F) —> 3T29(F) 3Tlg(F) —> 3T1g(P) (C2H5)4NVBr4-2CH3CN 12,950 (C2H5)4NVBr3Cl-2CH3CN 13,200 (C2H5)4NVBr2Clz-2CH3CN 13,420 (C2H5)4NVBrC13-2CH3CN 13,700 (C2H5)4NVCl4-2CH3CN 13,900 20,800 Position of center of first charge transfer band Color 20,200 rédtfiEBGh 22,000 red 22'600 orange 23,800 yellow—orange 25,600 yellow The charge transfer peak which produces the color change was investigated in some detail for tetrabromo and tetradflbro—complexes. The results of these spectra are shown in Table IX and representative spectra are given in Figure 14. As the temperature is lowered, the charge trans- fer band present in the tetrabromo complex shifts slightly to a higher energy and becomes narrower. This behavior is observed to a lesser extent for the tetrachloro complex. Thus while it would appear that the tetrabromo complex is 107 zoamou.amo>~zafl.maoVw cm_ 00.. 0.. MO E7. #00.an CO,.m.t.Q:OuQm waofimflm -Iq+mao.m..mmuab 18 com .rsooomv com . a 0:3qu CA:V com I .IIIIIIIIIIIIIIIIIIIIrI .Asooomv zowmom, 36852.1 emaov , Uswo: we Ell. -..Umm. ..toeyioflrosmm wanflnsn.-wmao,n .m ...”.HD 18 com com com 108 — a - .1 d 109 IMooomv zommow.uH0uum>HZeimmmovL p-40. wo Eflunccmm coaumnomnm OHQHmeluwH a muuHD .HH mnsqem 18 com cos com com com 110 Mooomv 203.5“ «HO ...m> :23. ammo: tween mo Er grown coflum (not mamswu,,rweon.ngqb NH mesmam 1E 00w CON. com Com DOV w a u a u . I q - 111 Umao mo Eu rvmmm zoewmnomgm wanemabluwaombmnsab .MH mmnqam 1E com com com com 00¢ 4 a a a q 1- n 1 ~ 112 I l L l 350 400 450 500 550 600 ‘ mu Figure 14. Ult.aviole:—visible ab orption spectrum of .,;c.‘tid. [ (c2115 )4N] VBr4 -2CH3CN. 113 thermochromic and the tetrachloro complex is not, the quality which produces the thermochromism is present in both of the complexes studied. The charge transfer band is located such that all the complexes with bromide ligands appear to be thermochromic while the tetrachloro complex is not. Table IX. Spectroscopic features of first charge transfer band in [(C2H5)4N]VC14°2CH3CN and [(C2H5)4N]VBr4-2CH3CN Compound Temperature Peaks (103 cm-1)* [(C2H5)4N]VC14'2CH3CN 3000K 25.6 1950K 25.7 77°K 25.8 [(C2H5)4N]VBr4°2CH3CN 3000K 18.3,19.2,21.5,23.8 1950K 18.8,19.4,21.8,24.1 779K 19.2,19.8,22.0,24.2 9(- Averaged from 4—6 spectra at each temperature. The structure on the charge transfer band could result from the superposition of the 3T1(F) ——> 3A2(F) ligand field transition as well as from the two spin forbidden ligand field transitions 3T1(F)'—-> 1A1(G) or 3T1(F) _-> 1T2(G) band onto the charge transfer band. No shift in the position of the 3T1(F) -—> 2T1(F) transition was observed as the temperature was changed. 114 However these peaks are less intense and small changes might go undetected. CONCLUSIONS Table X lists all of those compounds given in Day‘s review1 with thermochromic properties similar to the com— plexes investigated in this research. It would seem then that all thermochromic transitions actually should be listed in three classes: (1) Those transitions which are gradual and in— volve a mechanism similar to that found for the vanadium(III) complexes investigated here. (2) Those transitions which involve a change in the phase of the solid. (3) Those transitions which take place in solution and involve equilibria. It is felt that this investigation more clearly defines the phenomenon of thermochromism, and will encourage future thorough investigations of the theoretical implications of the color changes. 115 116 Table X. Compounds with thermochromic properties .Compound Color change on heating1 CuzFe(CN)6 red -—> dark red or black szsa red -—> dark red or black FeO red -—> dark red or black Cu12 red -—> dark red or black HgS red ——> dark red or black PbCrO4 red ——> dark red or black PbO red -—> dark red or black K2Cr207 red ——> dark red or black A5283 yellow —-> orange red HgSO4 yellow ——> orange red BaCrO4 yellow ——> orange red Snsz yellow ——> orange red H92I2 green ——> yellow ——> orange ——> red SnOz white —-> green -—> yellow ——> orange AgI yellow ——> red brown AnggI4 yellow ——> orange CqugI4 red —-> dark red TlZHgI4 orange ——> red Color change on coolipgl HgS red —-> light red SnSz yellow ——> light yellow HgSO4 yellow -—> green PbI2 orange ——> yellow orange PbCrO4 yellow orange -—> yellow green BIBLIOGRAPHY 10. 11. 12. 13. BIBLIOGRAPHY PART I Dalton, L. A., R. D. Bereman, and c. H. Brubaker, Jr., Inorg. Chem., 8 (1969? in press. Monoharan, P. T. and Max T. Rogers, J. Chem. Phys., 49, 5510 (1968). .DeArmond, K., B. B. Garrett, and H. S. Gutowsky, ibid., 42, 1019 (1965). McGarvey, B. R., J. Phys. Chem., 11, 51 (1967) Kuska, H. A. and M. T. Rogers, "Radical Ions", E. T. Kaiser and L. Devans, Eds. New York: Interscience Publishers, Inc., 1968. Chapter 13. Monoharan, P. . and M. T. Rogers, J. Chem. Phys., 49, 3912 (1968). Kon, H. and N. Sharpless, J. Phys. Chem., 12' 105 (1966) Ballhausen, C. J. and H. B. Gray, Inorg. Chem., 1, 111 (1962). Wentworth, R. A. D. and C. H. Brubaker, Jr., ibid., 2, 551 (1963) Wentworth, R. A. D. and C. H. Brubaker, Jr., ibid., 3, 47 (1964). Rasmussen, P. G. , H. A. Kuska, and C. H. Brubaker, Jr., ibid., 4, 343 (1965). McClung, D. A., L. R. Dalton, and C. H. Brubaker, Jr., ibid., 5, 1985 (1966) Giggenbach, W. and C. H. Brubaker, Jr., ibid., 1, (1968) 117 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 118 -Reagan, W. . P. Rillema, and C. H. Brubaker, Jr., ibid., 8, 581 (1969). Rillema, D. P. and C. H. Brubaker, Jr., ibid., 8, 1645 (1969) Giggenbach, W. and C. H. Brubaker, Jr., ibid., 8, 1141 (1969 Cotton, F. A. and G. Wikinson, "Advanced Inorganic Chemistry". New York: John Wiley and Sons, Inc., 1966. Selbin, J. and L. Morpurgo, J. Inorg. Nucl. Chem., 81, 673 (1965) Sidgwick, N. V., "The Chemical Elements and Their Compounds". London: Oxford University Press, 1950. Selbin, J., Chem. Rev., 88, 153 (1965). Selbin, J., Coord. Chem. Rev., 1, 293 (1966). Fowles, G. W. A., Progr. Inorg. Chem., 8, 25 (1965) Bridgland, B. E., G. W. A. Fowles, and R. A. Walton, J. Inorg. Nucl. Chem., 81, 383 (1965). Cozzi, D. and s. Cecconi, Ric. Sci., 88, 609 (1953) Neubauer, D. and J. Weiss, Z. Anorg. Allg. Chem., 303, 28 (1960). Aynsley, E. E., N. N. Greenwood, and J. B. Leach, Chem. Ind., 379 (1966). Markov, B. F., B. A. Voitovich, and A. S. Barabanova, Urk. Khim. Zh., 81, 580 (1961) Clark, R. J. H., J. Chem. Soc., 5699 (1965). Clark, R. J. H., J. Lewis, and R. S. Nyholm, ibid. 2460 (1962) Nicholls, D., Coord. Chem. Rev., 1, 379 (1966) Bradley, D. C. and M. L. Menta, Cand. J. chem., 88, 1183 (1962). Thomas, I. M., ibid., pg, 1386 (1961) Chamberlain, M. M., G. A. Jabs., and B. B. Wayland, J. Org. Chem., 81, 3321 (1962) 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 119 Thomas, I..M., Can. J. Chem., 88, 1386 (1961). Liebe, W., E. Weise and W..Klemm, Z. Anorg. Allg. Chem., 311, 281 (1961). Gutmann, V., 1888., ggg, 151 (1951). Gurmann, V., Monatsh., 88, 286 (1954). Gutmann, V., 1888., 88, 404 (1954). Gutmann, v., 1219-: §§, 1077 (1954). Gutmann, V., Z. Anorg. Allg. Chem., 888, 331 (1951). Fowles, G. W. A. and R. A. Walton, J. Inorg. Nucl. Chem., 21, 735 (1965). »Kilty, P. A. and D. Nicholls, J. chem. Soc., 1173 (1965). Funk, H., F. Schmeil, and H. Scholz, Z. Anorg. Allg. Chem., 0, 86 (1961). Funk, H. and H. Naumann, ibid., 343, 294 (1966). Faber, R. J., Ph.D. Thesis, Michigan State University, 1957. Title: "Paramagnetic Resonance of Some absorbed Transition Metal Ions". Vos, K. D., Ph.D. Thesis, Michigan State University, 1962. Title: "Electron Paramagnetic Resonance Absorp— tion Studies of Solutions of Metals in Ammonia and Amines". Wreede, J. E., M-S. Thesis, Michigan State University, 1963. Title: "Hyperfine Splitting in Electron Para— magnetic Resonance Studies of MetaleAmine Solutions”. Kuska, H. A., Ph.D. Thesis, Michigan State University, 1965. Title: "Electron Spin Resonance of Transition Metal Complexes". Low, W., “Solid State Physics", Suppl. 2, New York: Academic Press, 1960. Slichter, C. P., ”Principals of Magnetic Resonance”, New York: Harper and Row, 1963. Pake, G. E., "Paramagnetic Resonance", New York: ; Benjamin, 1962. / 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 120 Anderson, R. 8., "Methods of Experimental Physics", 8, New York: Academic Press, 1962. Carrington, A. and H. C. Lonquet—Higgins, Quart. Rev., 11, 427 (1960). Robertson, R. E., "Determination of Organic Structure by Physical Methods", 8, New York: Academic Press, 1962. Urry, Grant, "Radical Ions", E. T. Kaiser and L. Devans, Eds., New York: Interscience Publishers, Inc., 1968. Chapter 7. Treinin, A., "Radical Ions", E. T. Kaiser and L. Devans, Eds. New York: Interscience Publishers, Inc., 1968 Chapter 12. Kokoszka, G. E., H. C. Allen, Jr., and G. Gordon, Inorg. Chem., 8, 91 (1966). Rillema, D. P. and C. H. Brubaker, Jr., Inorg. Chem., 8 (1969) in press. Rillema, D. P., Ph.D. Thesis, Michigan State University, 1969. Title: "The Preparation and Properties of Some d1 Molybdenum and Tungsten Compounds". Ballhausen, C. J., "Introduction to Ligand Field Theory”. New York: McGraw-Hill, 1962. g Figgis, B. N., "Introduction to Ligand Fields", London: Wiley, 1966. Gray, H. B., J. Chem. Ed., 11, 2 (1964). Moffitt, W., and C. J. Ballhausen, Ann. Rev. Phys. Chem., 1, 107 (1956). Orgel, L. E., "An Introduction to Transition Metal Chemistry, Ligand Field Theory". New York: John Wiley and Sons, Inc., 1960. Cotton, F. A., ”Chemical Applications of Group Theory", New York: John Wiley and Sons, Inc., 1963. Abragam, A. and M. H. L. Pryce, Proc. Roy. Soc. (London), , A205, 135 (1951). / Owen, J., ibid., A 227, 183 (1955). , Murao, T., Progr. Theoret. Phys. (Kyoto), 81, 657 (1959). / 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 121 -Kon, H. and N. E. Sharpless, J. chem. Phys., 88, 906 (1965). .McGarvey, B. R., ibid., 31, 3743 (1964) Lacroix, R. and G. Emch, Helv. Phys. Acta., 88, 592 (1962). .Kivelson, D. and S. Lee, J. Chem. Phys., 81, 1896 (1964). Mulliken, R. s., ibid.,_88, 1833 (1955). Broadbent, A. D., J. Chem. Ed., 88, 145 (1967). Hillebrand, W. F., G. E. F. LUndell, H. A. Bright, and J. I. Hoffman, "Applied Inorganic Analysis", 2nd Ed. New York: Wiley and Sons, Inc., 1953. P. 458. Wentworth, R. A. D., Ph.D. Thesis, Michigan State University, 1963. Title: ”The Preparation and char- acterization of some Alkoxo'Niobium(IV) Complexes". Vander Vennen, R. E., Ph.D. Thesis, Michigan State University, 1954. Title: "Study of the Surface Phases of Charcoal and Copper". Figgis, B. N. and J. Lewis, ”Modern Coordination -Chemistry", J. Lewis and R. G. Wilkens, Ed. New York: Interscience Publishers, Inc., 1960. P. 412. Figgis, B. and R. S. Nyholm, J. Chem. Soc., 4190 (1958). Figgis, B. N. and J. Lewis, "Modern Coordination Chem- istry", J. Lewis and R. G. Wilkens, Ed. {New York: Interscience Publishers, Inc., 1960. P. 406. Machin, D. J. and K. S. Murray, J. chem. Soc., 1330 (1967). Figgies, B. N., Trans. Faraday Soc., 157, 198 (1961). Dunn, T. M., ibid., 157, 1441 (1961). __—__. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. .122 PART II Day, Jesse E., Chem. Rev., 88, 649 (1963). Day, Jesse E., 1818., 88, 65 (1963). Houston, E. J., Chem. News, 81, 177 (1871). Poole, C. P., J. Phys. Chem. Solids, 88, 1169 (1964). Wendlandt, W., and T. W. George, Chemist—Analyst, 88, 100 (1964). Asmussen, R.-W. and P. Anderson, Acta. Chem. Scand., 18, 939 (1958). Suchow, L. and G. R. Pond, J. Am. Chem. Soc., 18, 518 (1953). Miller, F. A. and W. B. White, Spectrochim. Acta, 2, 98 (1957). Gill, N. s. and R. s. Nyholm, J. Chem. Soc., 3997 (1059). Scaife, D. E. and K. P. Wood, Inorg. Chem., 8, 358 (1967). Linhard, v. M. and M. Weigel, Z.-Anorg. Allg. Chem., 266, 49 (1951). Nicholls, D., Coord. Chem. Revs., 1, 379 (1966). Duckworth, M., G. W. A. Fowles, and R. A.-Hoodless, J. Chem. Soc., 5665 (1963). Clark, R. J. H., J. Lewis, D. J. Machin, and R. S. Nyholm, ibid., 379 (1963). rNyholm, R. S., Croat. Chem. Acta, 157 (1961). Clark, R. J. H., R. S. Nyholm, and D. E. Scaife, J. Chem. Soc., 1296 (1966). Brauer, G., "Handbook of Preparative Inorganic Chemist I. ll. London: Academic Press, 1963. Vol. II, p. 1256. y Dean, C., Rev. Sci. Intru., 81, 934 (1960). -Reedijk, J. and W. L. Groeneveld, Rec. Trav. Chim., 88, 663 (1969). 20. 21. 22. 23. 24. 25. 26. 27. 123 Reedijk, J., A. P. Zuur, and W. L. Groeneveld, Rec. Trav. Chim., 88, 1127 (1967). Nakamura, D., K. Ito, and M. Kubo, Inorg. Chem., 8, 61 (1063). Nakamura, D., K. Ito, and M. Kubo, J. Am. Chem. Soc., 88, 1636 (1962). Nakamura, D. and M. Kubo, J. Phys. Chem., 88, 2986 (1964). ' Nakamura, D., Bull. Chem. Soc., Japan, 88, 1662 (1963). Ikeda, R., D. Nakamura, and M. Kube, ibid., 88, 1056 (1963). Ikeda, R., D. Nakamura, and M. Kube, J. Phys. Chem., 88, 2101 (1965). Barns, R. G., S. L. Segel, and W. H. Jones, J. Appl. Phys., 88, 296 (1962). “1111111()11111111“