Sept. 20, lDGl
ELECTRON SPINRESONANCE OF VANADIUM AND TITANIUM COMPOUNDS [ CONlRIBUTION FROM
TIIE
3767
HERCULES RESEARCH CENTER, WILMINCTON, DELAWARE]
Electron Spin Resonance of Some Vanadium and Titanium Compounds BY JAMES C. W. CHIENAND CHARLES K. Boss RECEIVED MARCH1. 1961 The electron spin resoiiance, e.s.r., of V(1V) in heptane solutions of VCh (8 X lo-‘ t o 8 X 10-*111 concentrations) and of VCI,(O-n-CJio)+-., where x = 0, 1, 2 and 3, cannot be observed from 90 to 330’K. VCI~(OK)?(KOH).where R is nand I-butyl, give well-resolved c.s.r. spectra. The e.s.r. spectra of ( a-CsH6)?Ti(CsHs), ( ~ - C ~ H ~ ) ~ V(a-CsHs),V( CIT, Cab) and (a-C6H6)lV(CsH4CH:)are also presented. The spectra of the last two compounds constitute the hrst observed roomtemperature spin resonance of V( 111). These results are interpreted with the paramagnetic relaxation mechanism based on interaction between the ground state and a nearby Stark excited state. The variation of line width and h y p e r h e constant on the nuclear orientation and on the solvent used are attributed to the “microcrystal” mechanism.
Introduction
of these vanadium esters were various hues of brown. The vanadium and its oxidation state were determined spectro-
Electron spin resonance, e.s.r., has been used photometrically. The Vohdrd method was used in the widely to elucidate the symmetry of crystals of chlorine determinations and the Zeisel method was used in paramagnetic salts, to study the structures of the analysis of the butoxy groups. The V:CI:O-n-C4Ho mole ratios found were: VCII(O-~-C,H~);1.0:2.8:0.8; organic free radicals, to measure the magnetic VC12(O-n-C,He)z; 1.0: 1.8:2.1; VCI(O-n-CrHo)a; 1.0: 1.312.9; properties of color centers and to determine the and V(O-n-C4Ho),; 1.0:0.05:3.9. The reaction products from VC&and n- and I-butyl alcohols spin densities in aromatic anions and cations. Triplet state molecules, biradicals and carbonaceous also were investigated. In tliese reactions, a heptane soluof the alcohol was added t o a heptane solution of VClr materials also have been investigated using this tion under nitrogen. At alcohol to VCI4 ratios of 2:1 t o 6:1, new technique. However, only a few organo- brown to black precipitates were obtained. These precipimetallic compounds have been studied. Singer’ tates were insoluble in the parent alcohols and contained reported the e.s.r. of tris-(acety1acetonato)-Cr- two chlorines per vanadium. Excess I-butanol reacted VClr to give a clear, green solution with an absorption (111). Jarrett2 extended the measurements to with maximum a t 14,300 cm.-l. Drying this solution gave the acetylacetone complexes of Ti(III), Mn(III), solid, green VClz(O-n-C~Hp)2.(n-C~HQOH) (I). Excess t-butyl Fe(III), Mo(II1) and Rh(II1). These authors alcohol reacted with VCI, to give a clear, blue solution found a strong axial field in these complexes. with an absorption maximum a t 15,400 cm.-l. Drying solution gave solid, dark green VC12(O-I-C4Ha)~.(tJarrett suggested that the three de degenerate this CrHoOH) (11). orbitals of the metal overlap with the T orbitals The brown heptane solution of VCII(O-~-C,H~)~ turned of the oxygen. The result of such “r-conjuga- green upon addition of excess I-butanol. This solution tion” is to lift the degeneracy of the de orbitals gave a visible spectrum identical with that of the solution of I. t o give an A, and an E, state and is equivalent The compounds (a-CsH&Ti( C2Hs) and (a-C&&VCb t o superposing on the cubic field a rhombic crystal were prepared by Dr. W. P. Long of this Laboratory. (r-CsH&V(GHs) and (~-CIH&V(GH,CH:) were gifts field of Da symmetry. Several chelates of Cu(I1) have been studied from Prof. van der Kerk.@ The e.s.r. spectra were obtained with a Varian Associates b y e.s.r., namely: bis-(dimethylg1yoximato)-CuX-band spectrometer and a 12-in. electromagnet. In these (11), 3 bis-(acet l a c e t o n a t o ) - C ~ ( I Iand ) ~ bis-(salicyl- expcriments, the concentration of unpaired spins in a specialdehydeiminerCu(I1). Spin resonance spectra men was determined by comparing the e.s.r. spectral inof vanadyl acetylacetonates and vanadyl etio- tensity (first moment) with that of a standard solution of vanadyl acetylacetonate ina medium identical with that of the porphyrin 1” have been reported. unknown. Both spectra were obtained with the same moduT h e e.s.r. spectrum of dibenzenechromium cat- lation amplitude and microwave power. This procedure ion’ shows a g-factor of 1.98, a 15-gauss line width is necessary because solvents with different values of loss and barely resolved proton hyperfine structures. tangent affect the “loaded” Q of the microwave cavity Therefore, the relative e.s.r. spectral intensiT h e e.s.r. spectrum of bis- (cyclopentadienyl) - differently. ties of vanadyl acetylacetonate a t the same concentrations vanadium8 shows a g-factor of 2.00 and a V61 in benzene, toluene, methylene chloride, chloroforni, chlorobenzene, o-dichlorobenzene and 1: 1 benzene-chlorohyperfine splitting of only 26 gauss. Recently, research activity on the chemistry of benzene were 10.0, 7.0, 6.1, 2.9, 2.7, 1.1 and 4.05, respectively. The cavity resonance disappeared when a dimethylorganometallic compounds of transition metals formamide solution of vanadyl acetylacetonate was inhas been greatly accelerating. I n this note, the serted into the cavity. The relative e.s.r. spectral intensifirst e.s.r. results for several such compounds are ties of a,a-diphenyfpicrylhydrazyl at the same concentrations in benzene, chloroform, chlorobenzene and odichloropresented.
Experimental VCI, was obtained from Stauffer Chemical Co., N. Y. VClZ(0-n-C4Hs)e. was prepared by the slow addition of a heptane solution of VC4 t o a stirred suspension of NaOn-CrHp in heptane under nitrogen. The heptane solutions L. S. Singer, J . Chem. Phy5.. 23, 379 (1955). H. S. Jarrett, i b i d . , 27, 1298 (1957). H. S. Jarrett, ibid., 28, 1260 (1058). A. H. Maki and B. R. McGarvey, i b i d . , 29, 31, 35 (1958). W. A. AndPrvon and L. H. Piette. ibid., SO, 501 (1959). D. E.O’Reilly, i b i d . , 29, 1188 (1950). (7) R . D. Feltham, P.Sogo and M. Calvin, i b i d . . 26, 1354 (1957). ( 8 ) H. M .McConnell. W. W . Porterfield aud R . E. Robertson, ibid., 80, 442 (1959). (1) (2) (3) (4) (5) (6)
benzene were 10,3.1, 2.8 and 1.2, respectively.
Results The e.s.r. of V(1V) in heptane solutions of VC14 (8 X lo-’ t o 8 X 10-2M concentrations) cannot be observed from 90 to 300’K. and from 500 to 14,000 gauss. At 90”K., VC14 is brilliant green. Heptane solutions of VCl,(O-n-C~H~)~-,, where x = 0, 1, 2 and 3, also failed to give observable e.s.r. absorption at both 90 and 300’K. (Table I). Addition of excess 1-butanol to the brown heptane solution of VC12(0-n-C4H& changed the color (9) H. J. de Liefde Meijer, M. J Janseen and G . J. M. van der Kerk, Chon. 6 ’ I n d . (London), 119 (1960).
TABLE 1
E.S.R.SPECTRAL PARAMETERS oi? SOME VANADIUM A N D TITANIUM ComPouNm Solvent
Heptane Heptane Heptane Heptane Heptane Heptane Heptane Heptanc Heptane
Heptane Heptane Heptane Heptane Heptane Rctizcne Chloroform Toluene
No observable e.s.r. absorption. width cannot be accurately measured.
Concn.,
molar
8X 8X 1x 1x 1x 1x 7 x 6 X
lO-'to 10-2 10-2 10-2
'Temp.,
OK.
p-value
A , gauss
L i n e width. gauss
90 and 00 and 90 and 90 and
Remark
Negative"
90 to 300 300 300 :300 :N)O
Negative Negative 10-2 Negative 10Negative 10-2 3O(J I .M 360 No hP c d 10-2 300 98, 131 Two 8-line spectrum c d 5 x 10-2 300 98, 131 Two %line spectrum 7 x 10-2 :300 1.95 112 :32 60% of theoretical intensity 6 X 10-2 300 1.95 112 25 70% of theoretical intensity 5 x 10-2 :100 I .95 111 70% of theoretical 25 intensity 7 x 10-2 3H1 1.95 45y0of theoretical 112.5 21 in tensity 6 x 10-2 8670 of theoretical 112.5 21 intensity 300 1 .!$5 5 x 10-2 100% of theoretical 112.5 21 intensity 71.2 0 . 9 x 10-3 7 2 lOO7', of thenretic:rl intensity 300 1.99 74.1 4.6 x 10-3 10.4 98% of theoretical intensity -1 x 10-1 300 2.13 IO00 Nohf 300 2.13 -1 x 10-2 1000 No hf :30( ) 1.94 -1 x 10-3 20 No hf Single line spectrum Line Center of spectrum cannot be accurately located.
to green. This solution gave a well-resolved e.s.r. spectrum with an average V1 hyperfine splitting, A , of 111 gauss, and a g-factor of 1.95. The spectral intensity was 70'g of the theoretical intensity. The insoluble reaction product of an equimolar mixture of VCl, and 1-butanol gave a very weak and broad spectrum with :I line width of about 360 gauss. The I : 3 and 1 :4 mixtures had very weak e.s.r. spectra which appeared to consist of 16 lines. Green solutions were ohtained when excess 1-butanol was added to VC14. The spectral intensities for the 1 :20 and 1 : 30 solutions were ii0 and 70% of the theoretical intensity, respectively. The spectral pnrameters were: A = 112 gauss and g = 1.95. Keactionsof VC14and t-butyl alcohol gave products which had e.s.r. spectra similar to the correspondiiig products from I-butanol. 'The blue solutions obtained at VC14:t-BuOH ratios of 1: 10, 1:20 and 1:30 had spectral intensities 45, S6 and 100% of the theoretical intensity, respectively. The spcc-tral p;irztnctcrs were: A =z 113.5 gauss :tnd g = 1.95. 'rhc e.s.r. spectra of ( T - C ~ H ~ ) ~ both V C ~in. ~beii~ e n eaiid in chloroform solutions have a near free electron g-factor of 1.99. T h e spectral parameters in the 0.92 X l0-3M benzene solution were: A = 71.2 gauss and line width = 7.2 gauss. The corresponding parameters in thc 4.6 X 10+ chloroform solution were: R = 71.1 gauss and line width = 10.4 gauss.
The toluene solutions of ( T - C ~ H ~ ) ~ V ( C Eand ") ( T - C & , ) ~ V ( C ~ H ~ Cgave H ~ ) identical room-temperature e.s.r. spectra. The line width was about 1000 gauss, the complete spectrum had a span of about 5000 gauss. The value for the g-factor w;ts 2.13. A toluene solution of (T-C~H~)~T~(C~HF,) gave an unsymmetrical room-temperature e.s.r. spectrum The line width was 20 gauss and the g-factor was 1.94. Discussion of Results VC1, is pararnagnetic1° and has a moment close to that of one unpaired electron spin both in the pure compound and in CCIJ solution. The magnetic susceptibility of chlorovanadic ester has not been measured. T h e failure t o observe the V(1V) spin resonance in VC14 and VCl,(O-n-C4H9)4-x is attributed to very short relaxation times in these systems. The line width is related approximately to the relaxation tinies by Line width =
1
1
T T ~+ ' ;Ti
where T I is the spin-lattice relaxation tinic aiid 1'2' is the transverse relaxation time. Both spin lattice and spin-spin relaxation processes contribute to Tz'. The failure to observe e.s.r. absorption in these vanadium compounds means that their line widths are. in excess of 2000 gauss. Van Vleckll showed (10) A. G. Whittaker and D. AT. Yost, J. C h e ~ i .Piiyr., 17, 188 (1940,. (11) J . I T . V a n Vleck, P h y s . R e i ) . 67, 426 (1940).
that T I decreases greatly when there is interaction vated dimeric species and the non-solvated tribetween the ground state and a nearby excited meric species. state. Therefore, transition metal compounds with The symmetry of ( B - C ~ H ~ ) ~ V is Criot I ~ known. degenerate ground states or with snitill separation I t is likely that the symmetry is lower than Czv. between the ground state and the next higher state The e.s.r. spectruni of ( a-CaHs)r\~C12 consisted have efficient paramagnetic relaxation. The e.s.r. of eight well-resolved lines having about the same line widths are broad. For compounds of low width, Fig. 1. Rogers and Pake’’ found that for symmetry, the ligand field splitting is large, leading to large separation between the ground state and the next excited state. Paramagnetic reP laxation is slow and the e s r . line width is narrow. Ti(IT.1) has the W-configuration. In the aquated Ti(II1) ions, the octahedral ligand field splits the five-fold orbital level into a r3and a rh state, the latter being lower. The ground sextet is further split by spin-orbit coupling into three Kramers’ doublets. ‘The separation between them is small. The “turned over” spin is relaxed by the spin-orbit and orbit-lattice perturbation. McGarvey’* reported that the e.s.r. of aquated Ti(II1) ion was not observed. The resonance of the -m m-• Ti(II1) in C ~ T i ( S 0 ~ ) ~ . 1 2 Hwas ~ 0 observed13 m @ below 8’K. The dominant crystal field was probably cubic. lo/ The intrinsic symmetry of (n-CsHb)zTi(CzHs) is very low. The strong ligand field of low symmetry increases the separation between the 5 I I I Kramers’ doublets. Consequently, the transition 7/2 5‘7 3.’2 112 -112 4 2 -5/2 -7/2 probability of paramagnetic relaxation is reduced, mi. leading to narrow line width e.s.r. absorption. V(1V) also has the 3d1-configuration. VC14 Fig. 1.-Plot of line width versus ;If, at X-band frequency: was found t o be a tetrahedral molecule by electron 0 , 0.92 X M (a-CsHs)z\’Clz in benzeiie; 0 , 4.6 X in chloroform; 0 , lo-+iM and 1 0 - 3 M diffraction measurements. l 4 The tetrahedral field 10- a M ( T-C~HF,)~VCI* splits the five-fold orbital level into a r b and a I’nvnriadyl acetylacetonate in heiizene; A , 6 X 1 0 2 df VO++.’g state, the latter being lower. The ground quartet is further split to a pair of Kramers’ doublets by aqueous solutions of V 0 + + , the electron line width spin--orbit coupling and by the Jahn-Teller effect. varies with the vanadium nuclear orientation acThe separation of the two KrdlnerS’ doublets must cording to the “microcrystalline” mechanism probe very small because the e.s.r. absorption of posed by McConnell.18 In this mechanism, a relatively stable short-range order between the VCI, was not observable a t 90’K. The fact that the e.s.r. in VClr(O-n-C4H9)4--s transition metal ions and the nearest solvent was not observed indicates a symmetrical struc- neighbors was postulated to form ‘microcrystals” ture for these esters. A probable one is the cyclic giving rise to anisotropies in the g-tensor and in the trimeric structure of a trigonal prism. This struc- hyperfine interaction tensor. Therefore, widths ture also was postulated15 for the trimeric Ti(0- of lines in aquated VO++ with different values n-C4H9).< and TiCl(O-n-CaHe)s. The insoluble of nuclear spin quantum number, M I , differ by as products from the reactions of VC14 and butyl much as 10 gausstR (Fig. 1). Similar line width alcohols probably are polymeric octahedra bridged dependence on MI was found in benzene solutions of vanadyl acetylacetonate. The lines in the by alkoxy groups. Bradley, et aZ.,16 found that VC12(OR),.(ROH) heptane solutions of VC12(OII)2.(ROH) were not is dimeric in boiling benzene and postulated a sufficiently resolved to be included in these discusstructure comprising two octahedra bridged by sions. The electron line width showed only slight dealkoxy groups. Strong axial fields are possible in these dimers. Narrow e.s.r. absorption lines pendence on nuclear orientation in both the benwere resolved in I and 11. The experimental zene solution and the chloroforni solution of (aspectral intensity of I1 agreed with the theoretical C&&VC12, (Fig. 1). On the other hand, the intensity, indicating that VCI?(O-t-C4Hs)z,( t - c d - e.s.r. spectrum of (a-CbH&\’Cl* in benzene has H90H) is completely dimeric in heptane. On the an average line width of 7.2 gauss and an average other hand, the spectral intensity of I was only hyperfine constant of 71.2 gauss. The spectrum 60-70% of the calculated intensity. It appears of the chloroform solution of (T-C&)zVC1z has that in I there is an equilibrium between the sol- an average line width of 10.4 gauss and an average hyperfine constant of 71.1 gauss. The variation (12) B. R . hlcOarvey, J . Phys. Chrm., 61, 1232 (1957). of the magnitude of the hyperfine constant and of (13) D. Bijl. Pror. P h y s . SOC.( L o n d o n ) , 8 6 3 , 405 (1950). (14) W. N. Lipscomb and A. G. Whittaker, J . Ani. Chenr. S O P 67. , the line width with the solvent used indicate the 2019 (19.15). presence of “niicrocrystallites” even though the 11.5) K . 1,. M a r t i n and G. Winfer. N o l i r r e , 188,3 1 3 (1060) (16) L). e. Bradley, R . K . Mnltaiii and W . Wardlaw. 1.C h e i r i .Cor., (17) K . K. Rogers and G. E. Pakc, J . CAcm. P l t y s . , 33, 1107 (1060). I
46,47 (19.58)
I
I
(18) H. hf, McConnell. ihid., ’26, 700 ( 1 9 3 3
I
I
3770 J. V. QU.\GLIANO, J. FUJITA, G. PRANZ, D. J. PHILLIPS, J. A. WALMSLEY AND S. E'. TYREEVol. 83 line width vari ,ition with A l l is not significant. gauss, the line widths in the spectra of (a-CaH&Vwere about 1000 Inspection of tile equation derived by Kivelson1° (GH6) and (~-CF,H&V(C~H~~CH~) showed that t liese observations are consistent gauss. The difference is partly attributable to with small anis kropies in the g-tensor and in the the difference in the spin-orbit coupling coefficient hyperfine intera tion tensor. It will be interesting for Ti(II1) and V(III), the valuesz1 are 154 and to make h e \,idth measurements on (T-CF,HF,)L-217, respectively. Van Vleck" showed that the VCl2 a t the K-band frequency to bring out signifi- paramagnetic relaxation time is inversely procant line width dependence on nuclear orientation. portional to the square of the spin-orbit coupling The e.s.r. spectra of (a-C6H&V(C&) and (r- coefficient. The small coupling coefficient in CSH6)2V(C&C&) have line-widths of 1000 gauss. Ti(II1) leads to long relaxation time and narrow The ground state of V(II1) is 3F2. The seven-fold e.s.r. line width. orbital levcl is split by a cubic field to I'4, ri and Others have also noticed the dependence of e.s.r. r2 states, the first being the lowest. No room line widths upon field symmetry. Pake and Sandsz2 temperature V(II1) spin resonance has been re- obtained a narrow e.s.r. spectrum in aqueous soluported before. JarrettZ was not able to observe tion of VOSOI. The resonance disappeared when the spin resonance in tris-(acety1acetonato)-V(II1) the solution was made alkaline. These authors a t liquid nitrogen temperature. The e.s.r. spec- attributed this disappearance to the formation of trum of tris-(acety1acetonato)-Ti(II1) was readily complex ions in which the field about the V(1V) observed under similar experimental condit.ions. ion is more symmetrical. He concluded that in V(I1I) the orbital degeneracy This work points out the importance of symof the r4 state was not lifted by the axial field. metry considerations in e s r . studies. Lack of Lambe arid Kikuchi2" reported that in cu-Ald&, success in observing e.s.r. absorption in a transithe spin resonances of V(I1) and V(1V) were ob- tion metal compound may be remedied by lowering tained a t room temperature but the spin resonance the symmetry of the molecule by changing the of V(II1) was obtained only a t 4'K. Here, the number or the type of the ligands or both. crystalline field was primarily trigonal. Acknowledgments.-The authors appreciate The symmetries in ( n-CbH6)zV(CeH~)and in ( n-C6&)2V(CeIhC&) are lower than Czv. These helpful discussions with Dr. H. S. Jarrett of the compounds gave the first observed room tempera- du Pont Company and Dr. A. D. Liehr of the Bell ture e.s.r. spectra of V(II1). While the line width Telephone Laboratories. in the spectrum of ( T - C ~ H ~ ) ~ T ~ ( C Z isHonly F , ) 20 (21) B. N. Figgis and J . Lewis, "Modern Coordination Chemistry," (19) D. Kivelson, J. Chcm. Phys., 33, 1094 (1960). (20) J Lambe and C. Kikuchi, Phys. Rev., 113, 71 (1960).
(CONTRII4ITION FROM
Interscience Publishers, Inc., N e w York, N. Y.,1960, Chap. 0. (22) G . E. Pake and R. H Sands, Phys. Rev., 98, 266 (1955).
DEPARTMENTS O F CHEMISTRY O F THE UNIVERSITY O F NORTHCAROLINA, CHAPEL HILL, NORTH CAROLINA, AND THE FLORIDA STATE UNIVERSITY, TALLAHASSEE, FLORIDA]
TlIB
The Donor Properties of Pyridine N-Oxide BY J. V. QUAGLIANO, J. FUJITA, 6. FRANZ, D. J. PHILLIPS, J. A. WALMSLEY AND S. Y. TYREE RECEIVED MARCH15, 1961 The coordination chetnistry of pyridine N-oxide with a variety of acceptors has been investigated. The new substances Co(C1O4)?6L,CO(NO~)~.OI,, CoC12.3L, CoC12.L.H20, Ni(C104)2.6L, NiCI*.L.H10, NiBr2.6L, Ni12.6L, NiBrZ.LaH*O,Cu(C104)z.4L, C I I ( N O ~ ) ~ CuCI*.L, ~~L, CuCI2.2L, Zn(C10,)~6L,ZnC12.2L, .Zn(N0&.6L, Fe(C104)2.6L, Fe(Cl04)~.6L,and SnRrr,ZL (where I, = pyridine N-oxide) have been isolated and characterlzed by molecular conductance measurements in non-aqueous solvents, magnetic susceptibility measurements and infrared spectra. All of the perchlorates and some of the nitrates and halides have only the pyridine N-oxide in the h s t coordination sphere. However, in some cases, both nitrates and halides are coordinated t o the central ion.
Introduction As a logical extension of previous work on the donor properties of substituted phosphine' and arsine2 oxides, we wish to report now the results of a study of the ability of pyridine N-oxide to behave as a ligand. We find only three prior pyridine N-oxide adducts, the hydrochloride,a a 1 : 1 adduct with SOs,' and the hexakis-(pyridine N-oxide)-cobalt(11) tetracarbonylcobaltate( - 1).6 However, i t (1) J. C. Sheldon and S. Y. Tyree, J . Am. Chcm. Soc., 80, 4775 (1958). (2) D. J. Phillips and S . Y.Tyree, ibid., 83, 1808 (1961). (3) J. Meisenheimer, B e t . , 69, 1848 (1920). (4) P. Baumgarten and H. Erbe, i b i d . , 71. 2603 (1938). (6) C. Franr, Dissertation, Technixbe Hochffhule Munich, 1969.
is our understanding that two adducts with zinc chloride have also been prepared.8 Experimental Reagents.-Reagent or analytical grade chemicals were used without further purification except in the cases noted. Pyridine N-oxide was prepared once by the method of Ochiai.7 Other samples were obtained from Reilly Tar and Chemical, and K and K Chemical laboratories. All samples were purified by vacuum distillation, collecting the fraction distilling between 125 and 130" a t 8 mm. N,N-Dimethylfonnamide (DMF) was purified by a modification of Method I of Rochow.8 (6) Sister M. Ellen Dolores Lynch, C.S.C., Dunbarton College of Holy Cross, Washington, D. C.,private communication. (7) E. Orhid. J . 011. Chcm., 18, 648 (1953). (8) A. B. Thomas and E. C. Rochow, J . A m . C h m . .FQc.. 79, 1843 (1867).
