摘要:
1976 999Paramagnetic Anisotropy and Electronic Structure of [NN'-Ethylenebis-(salicylideneiminato)]cobalt(~i), its Pyridine Adduct, and [NN'-Ethylene-bis(thiosalicylideneiminato)]cobalt( 11)By Keith S. Murray and Robert M. Sheahan, Chemistry Department, Monash University, Clayton, VictoriaSingle-crystal magnetic-anisotropy measurements have been made over the temperature range 80-300 K onthe oxygen-inactive title chelates [(Co(salen)),], [Co(salen)]*py, and [(Co(tsalen)),]. These five-co-ordinatespecies all show rhombic magnetic ellipsoids with [{Co(tsalen)),] being the least anisotropic. Both salen com-plexes display an unusually rapid increase in magnetic moment at higher temperatures. The theoretical modeldeveloped to interpret the results incorporates spin-orbit coupling between the ground doublet stateand theexciteddoublet and quartet states, the latter being low-lying and important in the case of the salen complexes.Smallchanges in ligand field with temperature have been invoked to explain the rapid increases in p4 at higher tempera-tures. The relative ordering of the cobalt d-orbital levels has been deduced, and correlations between the positionsof the out-of-plane x orbitals, d, and dvz, and the bonding of the axial donor ligands (e.g. pyridine) and of 0, arebriefly discussed.31 68, AustraliaSINCE the early work of Tsumaki, and Calvin andtheir co-workers, there has been continued interest inthe properties of low-spin cobalt (11) Schiff-base chelatecomplexes. This is chiefly due to the unusual reactivitydisplayed by this important class of compounds.Thecomplexes are best known for their ability, under certainconditions, to reversibly co-ordinate oxygen both in thecrystalline state and in solution, and this has led to theiruse as models for biological oxygen carriers3-6 and totheir application as catalysts for certain organic oxi-dation~.~~' They also form very stable cobalt-carbona-bonded organometallic compounds which have servedas important model systems for vitamin B,, coenzyme.*It seemed to us important, therefore, to obtain a de-tailed picture of the electronic structures of such com-plexes and to this end we have investigated the aniso-f salen = NN'-Ethylenebis(salicy1ideneiminato) , tsalen =NN'-ethylenebis(thiosalicylideneiminato), and py = pyridine.T.Tsumaki, Bull. Chevrt. SOC. Japan. 1938, 13, 252.A. E. Martell and M. Calvin, ' Chemistry of Metal ChelateCompounds,' Prentice Hall, New York, 1962.E. Bayer and P. Schretzmann, Structure and Bonding, 1967,2, 181. * G. Henrici-Olive and S. Olive. Aweew. Chenz. Internat. Edn.. , " 1974, 13, 1.55.A. L. Crumbliss and F. Basolo, J . Amev. Chem. SOC., 1970,92,G. A. Rodley and W. T. Robinson, Nature, 1972,235, 438.M. Fieser and L. F. Fieser, ' Reagents for Organic Synthesis,'* D. Dodd and M. D. Johnson, J . Organometallic Chem., 1973,Wiley, 1972, vol. 3, p. 246.52, 1.tropy in the magnetic susceptibility of single crystals of arange of salicylaldimine and acetylacetoneimine deriv-atives. The theoretical analysis of the principal sus-ceptibility measurements, obtained over a wide temper-ature range, has yielded the relative ordering of the&orbital energy levels. We discuss here the resultsobtained for three examples of the salicylaldimine type,[(Co(salen))J, [Co(salen)]*py, and [{Co(tsalen))J,t allof which are inactive towards oxygen in the crystallinestate, but active to varying extents in solution.Aver-age susceptibility measurements on polycrystalliiiesamples of [{Co(salen))& and [Co(salen)]*py have beenmade previo~sly.~ Recently there have been numerouse.s.r. studies on complexes of this type in powder form orin frozen solvents, which have yielded some informationon the ground state and orbital occ~pancy.l~-~~ SomeA.Earnshaw, P. C. Hewlett, E. -4. Icing, and L. F. Lark-worthy, J . Chem. SOC. ( A ) , 1968,241.lo B. M. Hoffman, D. L. Diemente, and I?. Basolo, J . Avzev.Chem. Soc., 1970, 92, 61.C. Busetto, F. Cariati, A. Fusi, M. Gullotti, F. Morazzoni, A.Pasini, R. Ugo, and V. Valenti, J.C.S. Dalton, 1973, 754.l2 C. Busetto, F. Cariati, P. Fantucci, D. Galizzioli, and F.hlorazonni, J.C.S. Dalton, 1973, 1712.l3 E. I. Ochiai, J . Inorg. Nuclear Chem., 1973, 35, 1727.l4 Y. Nishida and S. Kida, Bull. Chem. SOC. Japan, 1972, 45,l5 A. von Zelewsky and H. Fierz, Helv. Chim. Acta, 1973, 56,l6 L. M. Engelhardt, J. D. Duncan, and M. Green, Iizovg.l7 R. Karlsson, L. &I. Engelhardt, and &I. Green, J.C.S.461.977.Nuclear Chem. Letters, 1972, 8, 726.Dalton, 1972, 2463J.C.S.Daltonof our results have been presented in a preliminary and [{Co(tsalen)),] the modified approach of Gerlochcommunication .u and Quested was employed to obtain the best set of directioncosines.a0 Assumption of rhombic symmetry gave anEXPERIMENTAL excellent fit a t all temperatures. The direction cosines forthese and orthorhombic [Co(salen)]*py are given in Tablegrown as described previously;l* [{ Co(tsa1en) )J was 2. The derived molecular-susceptibility ellipsoids for theLarge crystals of [{Co(salen) )J and [Co(salen)]*py were-I I &I I (J \synthesized as described lo and crystals were grown fromNN-dimethylformamide (dmf) solution using the samesealed-tube techniques as for the other crystals. Thecrystals of [{Co(tsalen)),] were much smaller than theothers, ca.0.2 mg, but were sufficiently large to obtainreproducible anisotropy data.The anisotropy measurements over the temperaturerange 90-300 K were made using a null-deflection torsionbalance, described earlier. The wheel device of Gerlochand Quested was employed for the monoclinic crystals of[{Co(salen)}J.20 The smaller crystals of [{Co(tsalen)},]were mounted along the a*, b, and c axes respectively.Crystal axes were located by X-ray photography. Powdersusceptibilities were determined by the Gouy method in theliquid-nitrogen range, and on a Foner vibrating-samplemagnetometer in the range A 7 7 K for [{Co(salen))J.Crystal Structures.-The complexes [{ Co(sa1en) },I and[{Co(tsalen) )J are both monoclinic, with four centrosym-metric dimer groups in the unit ~ell.21-2~ They have similarmolecular structures, (I) and (11) , in which ligand-bridgeddimerization occurs to give square-pyramidal stereo-chemistry around each cobalt.The complex [Co(salen)]*pyis orthorhombic and has a monomeric square-pyramidalstructure, (111), in which the plane of the pyridine ringbisects the ligand N,O direction.a*RESULTSThe crystal anisotropies, X i , and average susceptibilities, x, over the range 90-300 K are given in Table 1. Theprincipal molecular susceptibilities, Kg, were obtained fromthese by the tensor-transformation methods of Krishnanand Lonsdale.26 For the monoclinic crystals of [{Co(salen) }Jl8 K. S. Murray and R. M. Sheahan, Chew.Phys. Letters, 1973,22, 406.l9 A. van den Bergen, M. F. Corrigan, K. S. Murray, R. M.Slade, and B. 0. West, Inorg. Nuclear Chem. Letters, 1974.10, 869.2o M. Gerloch and P. N. Quested, J . Chm. SOC. ( A ) , 1971,2307.21 S. Brtickner, M. Calligaris, G. Nardin, and L. Randaccio,Acta Cryst., 1969, B25, 1671.22 R. de Iasi, B. Post, and S. L. Holt, Inorg. Chem., 1971, 10,1498.23 M. F. Corrigan, K. S. Murray, R. M. Sheahan, B. 0. West,G. D. Fallon, and B. M. Gatehouse, Inorg. Nuclear Chem. Letters,1976, 11,626.m M. Calligaris, D. Minichelli, G. Nardin, and L. Randaccio,J . Chem. SOC. ( A ) , 1970, 2411.25 K. S. Krishnan and K. Lonsdale, Proc. Roy. SOC. , 1936, A156,597.TABLE 1Experimental? crystal anisotropies and average sus-ceptibilities ( lo8 cmS mol-l)(a) [{Co(salen)hlb axis (xa) Xa verticalTIK320300280260240220200180160140120100vertical40839238338439441 0432462495638696670(xz - XI)(4 CCo(salen)l*PYa axisvertical(xc - xb)320 776300 725280 687260 664240 665220 666200 692180 735160 797140 884120 999100 1166(4 [{Co(tsalen)121b axis (XJvertical(xa - XI)300 76280 70260 67240 62220 60200 60180 60160 60140 68120 68100 68(xs - XI)1140114611601198133314191 6291665I 8332 0902 4181268c axisvertical476423380360336320330343371412466636(x8 - xb)t axisvertical(Xa* - x1001001001001001009796969494x1 9431 9631 9962 0822 1932 3202 4712 6602 8823 1673 6494 0322 8792 2002 1782 1942 2602 3762 6362 7362 9863 3003 7104 245a* axisverticalx n - x s ) x’150 1660148 1 676144 1 696139 1 780134 1 870127 1976127 2 116127 2 300126 2 630125 2 866126 3 300t The quoted values are smoothed data obtained for severalcrystals of each complex1976TABLE 2Direction cosines[(Co(sawIal I{Co(tsalen)> 21 LCo(salen)l 'PYr I I > I \K, -0.1303 -0.9671 0.2688 0.3762 0.4844 -0.7903 -0.9999 0.0016 0.00570.9896 -0.1414 -0.0267 -0.1062 0.8695 0.4825 -0.0069 -0.2678 -0.9636-0.061t -0.2677 -0.9666 -0.9208 0.0971 -0.3776 0 0.9636 - 0.2678a b G * a b C * a b GKUK,1001TABLE 3Observed and calculated molecular susceptibilities (108 cm3 mol-l) and magnetic moments (1 B.M. % 9.27 x A m2)(4 [~Co(salen)hlTIK 320 300 280 260K , (obs.) 2669 2687 2 746 28601813 1810 1842 19161347 1369 1400 1471Ka (calc.) 2678 2701 2728 28641808 1798 1853 19211349 1396 1452 1623I.r, (obs.) 2.61 2.64 2.48 2.44c(v 2.16 2.08 2.03 2.001.86 1.81 1.77 1.762.23 2.17 2.11 2.082.62 2.65 2.47 2.44PY 2.16 2.08 2.04 2.00I4 1.86 1.83 1.80 1.78 rz 2.23 2.17 2.12 2.09= 600 crn-l, J = -40 cm-l.K Y K,KVKar Pi% (talc.)Fit (4 (b) (c) (d)Best-fit parameters:2403 0142 0111 6642 96619991 6072.411.961.732.062.391.961.762.06(4220 2003192 34012122 22581646 17643 162 33962119 22611699 18092.37 2.331.93 1.901.70 1.682.02 1.992.36 2.331.93 1.901.73 1.702.02 2.00(4 (4A1crn-l1803 6642 42718883 6732 4311 9062.301.871.661.962.301.871.671.97(41603 9782 6242 0444 0122 6382 0992.261.831.621.922.271.841.641.93(el, - 14T1 aT2(141 f >XnY z 6 r)10 000 1 600 1 0002 000 1 300 10 000 2 6004 000 1 400 3 000 12 0004 000 1 600 3 000 12 0004 000 1 600 3 300 14 0002 300 (8)(b) (4(4(4A(2T1)(aAJA(2T1)(1E)x = 2 000, y = 1 200,~ = 6 000x = 2 000, y = 16 000, z = 20 0003202 7022 3382 7972 7002 3691 8042.632.452.142.422.632.462.152.42Lmeters: t: =(43002 6032 24117672 6022 2501 7662.502.322.053.302.502.322.062.30(b)600 cm-l2802 6672 2031 7652 6562 19617762.402.221.992.212.392.221.992.21(42602 6762 2061 8002 5562 2121 8232.312.141.932.142.312.151.952.14(dl2402 6402 26618762 6672 2961 9082.262.091.902.082.262.101.912.10(el2202 7692 36819912 7912 3882 0162.212.041.872.042.222.061.882.05(f)2002 9462 5262 1372 9362 5232 1462.172.011.862.012.172.011.862.01(g)PT, - Y z(4 760 500900 7501100 1 0001 500 1 3001 800 1 5002 000 1 8002 000 2 000(b)(4(d)k)A('TJ('Aa)A('TI)('E)1))X,Y = 760, z = 4500X,Y = 2 750, z = 8 7501803 1722 7192 3143 1612 7122 3012.141.981.831.982.131.981.821.98(g)1603 4592 9672 6293 4422 9492 4982.101.961.801.962.101.941.791.94(g)I4 0004 2004 6005 0005 0005 0006 200-417 60017 50017 50017 60017 60017 60017 5001404 3762 8772 2474 4332 8942 2972.211.801.691.882.231.801.601.90( 41403 8263 2812 7943 8033 2632 7612.071.921.771.922.061.911.761.91(g)1204 9343 2062 6064 9663 2202 6482.181.761.661.862.181.761.661.86(41204 3042 6883 1384 2843 6583 0892.031.881.741.892.031.871.721.88(g)1005 6413 6232 8325 6593 6422 8742.121.701.511.802.131.711.621.80(41004 9414 2143 5804 9574 2243 5641.991.841.691.841.991.841.691.84(g1002 J.C.S.DaltonTABLE 3 (Continzced)3001 7021 66613931 6841 52013662.021.931.831.932.011.911.801.9328017731 6321 47017611 59214271.991.911.811.911.981.891.791.9126018391 703164218271 6761 5081.961.881.791.881.961.871.771.892401919179016311914177016011.921.861.771.861.921.841.751.872202 005188017242 016188117111.881.821.741.811.881.821.741.842002 108198418332 1372 0121 8401.841.781.711.781.861.791.721.821802 2472 12319752 2812 16819961.801.761.691.751.811.771.691.791602 4312 3082 1602 4662 3692 1831.761.721.661.721.771.741.671.76Best-fit parameters:A(*Tg)(lE) 4 = 10 000, q = 16 000.= 600 cm-1, J = -40 cm-1, A in cm-1.A(2T1)(3AJ x = 1 260, y = 2 000, z = 12 600.26 0001402 6592 6382 3922 6742 6952 4171.731.691.631.691.731.711.661.721202 9942 8732 7282 9602 8962 7131.691.661.621.661.681.671.611.681003 4283 3083 1633 3083 2843 0981.661.631.591.621.631.621.671.62A(aTJ('E) x = 10000, y = 15000, Z=three complexes have K2 and Kv in the ligand plane andmidway between the 0,O (or S,S) and 0,N (or S,N) donoratoms, respectively; K, is a t right angles to the plane,along the axial direction. In all cases K, > K , > Kz,with [(Co(tsalen) )J being less anisotropic than the othertwo complexes.The Ki values (per cobalt atom) are given in Table 3.They have been given earlier '8 in graphical form whereit can be seen that plots of 1/Ki against temperature for[{Co(salen)),] and [Co(salen) ]*py exhibit maxima a t ca.300 and 270 K, respectively.In contrast, [(Co(tsalen)),]exhibits Curie-Weiss behaviour. The related plots ofeffective magnetic moment (per cobalt atom) showed apronounced increase with temperature a t higher temper-atures for the salicylaldimine complexes, although not forthe thiosalicylaldimine derivative (Figure 1). Theseincreases were thought, at an early stage, to be indicative ofexcited quartet-state interactions. Measurements of Rin the range 4-77 K were obtained for one of the com-plexes, [{ Co(sa1en) 12], and a susceptibility maximum wasobserved at ca.46 K.The Magnetic IWodeZ.-Low-spin d7 cobalt(I1) Complexeshave a 2E ground state which is derived from the t26e1orbital configuration (using octahedral notation for con-venience). The theoretical model usually used to deter-mine the magnetic and e.s.r. properties of such a system isthat developed by Griffith 26 and (later) by Maki et ~ 1 . ~ 'We have recently outlined the inadequacy of this simplifiedtreatment,28 which in essence considers the electronicstates as single orbital wavefunctions, i.e. analogous tothose of CuII, d9. The present model, which is itselfsimplified to some extent, and which has many moreadditional parameters, does however lend itself to easyutility in the reproduction of principal magnetic momentsand g values, something which could not always be obtainedeven qualitatively with the early model.It consists ofthe crystal-field-spin-orbit matrix V,.f. + <l.S, whichuses states based on the electronic configurations t , V(ground) and f26e2 (excited). The states, which consist ofvarious linear combinations of orbital^,,^ include spin26 J. S. Griffith, Discuss. Faraday Soc., 1968, 26, 81.27 A. H. Maki, N. Edelstein, A. Davidson, and R. H. Holm,28 K. S. Murray and R. 3%. Sheahan, J.C.S. Chem. Comnt., 1975,29 J . S. Griffith, ' Theory of Transition Metal Ions,' CambridgeJ . Amer. Chena. Soc., 1964, 86, 4680.475.University Press, 1964.doublets and quartets: 2E; 4T1(3A2); 2T1(3A2); 2T1(1E);,T2(IE) ; and 2T2(1A,).The strong-field limit was assumedwith the parametrized crystal-field elements diagonal.r21-51 ' I I Ic1 I I 1100 150 200 250 300T l KFIGURE 1 Temperature dependence of principal magneticSolid lines are calculated curves using parameters moments.given in Table 3. (a), [(Co(salen)),] ; (b) [C'o(salen)]*py;(4, [(co(tsalen)}JMatrix elements between *T1(3A 2) and 2T1(3A2) wereignored. Further excited states arising from t,*e3 an1976 1003tZ3e4 were excluded as these do not mix to first order byspin-orbit coupling with the ground state. Eigenvaluesand eigenvectors were obtained on diagonalization of thematrix (see Table 4). The principal susceptibilities wereTABLE 4Off-diagonal non-zero matrix elements under spin-orbitcoupling. Multiply each matrix element A (u,b) by ec,the spin-orbit coupling constant.A (a,b) = A(b,u).The diagonal elements are the parametrized crystal-field energiesStatenumber(a, b)12345678910111213141516171819202122232425262728State *2E 402TZ -34 ('E)A ( 3 , l ) = -*A ( 5 , l ) = 8 . 3-1A (4,2) = - B . 3 4A (6,2) = 8A (7, 1) = -4, A (7.3) = 6A (8, 2) = -+ . 3-4, A (8,4) = -6A (9, 1) = -+. 3-4, A (9,6) = -QA (10, 2) = A (lo, 6) = --* A (11, 1) = - 2 . 3-*,A(11J5) = -*,A(11, 3) = A (11, 7) = +. 3-*, ( AA (12, 2) = - 2 . 3-4, A (12, 6) =- 4 . 3-4, A (12, 4) = A (12, 8) =(11.9) = 4+, A (12, 10) = Q .3 4A (13,2) = -i$. 6-*A (14, 1) = -+. 6-4A (15, 2) = - & . 6-4, A (15, 13) =zA (16, 1) = 4 . 6-4, A(16, 14) = -QA (17, 1) = 6-*, A (17, 14) = i,QA (17,16) = -QA (18, 15) = -QA (18, 2) = -6-4, A (18, 13) = -*,A (19,2) = Q . 2-4A ( 2 0 , l ) = Q . 2-4A (21, 2) = 4 . 2+*, A (21, 17) = -4A (22, 1) = - 4 . 2-4, A (22,18) =A ( 2 3 , l ) = -+ . 2-+, A (23, 18) = +,A (23, 20) =A (24, 2) = Q .A (24, 21) = -&A (24,19) = -$,34 A (25,2) = - 2.243a A (26, 1) = - 2.2k- 3, A (27, 2) = mi, A (27, 25) = -*3 i A (28, 1) = -mi, A (28, 26) = 4* The wavefunctions are labelled according to Table A24 ofref. 29 i.e. (term; m,; orbital basis; (e2 term component)).Orbitalsymbolism: 8 = @, ( 4 ) = dyr, r) = &.then calculated in the usual way, using the Van Vleckequation ; some pertinent aspects of the calculations aredescribed below.No particular symmetry was assumed inthe model, the energies of each excited state being para-metrized relative to the ground state. The actual sym-metry around the cobalt(r1) atoms in the present chelates isC, but is often approximated to Caw.Catcutations. The calculations of the magnetic sus-ceptibilities of the complexes exhibiting quartet interactions,[(Co(saien) )J and [Co(salen)]*py, required the diagonaliza-tion of a 28 x 28 matrix, while those which did not,[(Co(tsalen) j2] and other (unpublished) cobalt(I1) chelates,required an 18 x 18 matrix. Only those excited statesfrom the t,W and t26e2 configurations which coupled directlyby spin-orbit coupling to the ground state were included.The states derived from 2T2(1A1) were excluded as they areexpected to be rather high in energy, and the influence onthe magnetic behaviour is mimicked by the states derivedfrom 2T2(1E), although it must be borne in mind that thecalculated energies for these latter states will be ratherlower than is actually the case because of the neglect of2T2(lA To further minimize parameter variation, theone-electron spin-orbit coupling constant, c, was set a t500 cm-1, slightly reduced from the free-ion value.Anysymmetry mixing of states was neglected. A summary isgiven below of the effect of the energies of the excited stateson the calculated susceptibilities. It was quickly found thatthe ground state for all complexes was zE6, i.e.the unpairedelectron is in the d , ~ orbital. To first order, we havealready shown the effect of the excited states on the gvalues for a d,a ground configuration.28The spin-quartet states onlyaffect the susceptibilities to second order via spin-orbitcoupling. The 4T1(3A2) (x, y, z) states increase the first-order Zeeman term on approaching the ground state for allthree directions. The 'T1(3A2) (x, y ) states also increasethe second-order Zeeman terms in the x and y directions,but the *T1(3A2) z term reduces it in the z direction. The2T1(3A2) (x, y , z) states decrease the first-order Zeemanterms on approaching the ground state, but increase thesecond-order terms, whilst the 2Tl (1E) (x, y, z) statesincrease both first- and second-order Zeeman terms whenthe energy separation from the ground state decreases.Thus the origin of a large high-frequency (or second-orderZeeman) susceptibility is easily explained ; the 2Tl termshave a small effect on the low-frequency susceptibilitybecause of the opposite sign of mixing into the groundstate, but have an additive effect on the second-orderZeeman term.The 2T2(1E) and 2T2(1A1) states behavein the same way, mixing into the ground state, and raisingthe susceptibilities in the x and y directions [2T2 (5) and2T2 (7) respectively] for both first- and second-order Zeemanterms.We have alreadymentioned briefly the rapid increase in moment a t highertemperatures for [{ Co(sa1en) j2] and [Co(salen)]*py, whichsuggests the possible importance of quartet-state inter-actions in these two complexes.The low-temperatureregion of the susceptibility plots (80-240 K) for thesecomplexes and the whole of the temperature range for[{ Co(tsa1en) },I showed Curie-Weiss behaviour . These linearregions could be well fitted to the theoretical model de-scribed above, without the need of including quartet-stateinteractions in the case of [{ Co(tsa1en) j2]. The substantialtemperature variation in this Curie-Weiss region arisesfrom a large second-order Zeeman term. The best-fitparameter values are given in Table 3, and a discussion ofthese is given below. The model was not, however, able topredict the rapid increase in p at higher temperatures forthe salen complexes.A number of approaches were examined to tryto explain the behaviour in this region.Mixing byspin-orbit coupling of the doublet ground state withexcited spin-quartet states or a thermal population ofthese states could not reproduce the observed behavio~r.~~, 3130 C. G. Barraclough, Trans. Faraduy Soc., 1966, 1033.s1 C. M. Harris, T. N. Lockyer, R. L. Martin, H. R. H. Patil,(a) 2Tl and 4T1 States.(b) 2T2 States.Comparison of theory a%d observed data.E. Sinn, and I. M. Stewart, Austral. J . Clzetw., 1969, 22. 21061004 J.C.S. DaltonThe existence together of doublet and quartet-state mole-cules, with a thermodynamic equilibrium between the two,was rejected both because of the lack of any direct structuralevidence and the large number of rather ad ~ O G assumptionsinherent in the model.The existence of such spin isomersI 1 I I I I100 150 200 250 300'T / KFIGURE 2 Effect on of varying the crystal-field parameterswith temperature. Solid lines are calculated values for thebest-fit parameters a, at 320 K and a,at ~ 2 0 0 Kfor [Co(salen)J.-py, and b, a t 320 K and b, at ~ 2 4 0 K for [{Co(saJen)}J (seeTable 3). Broken lines are the experimental curves forCwsawl*PY (.I and [{~(salen)),I (0)is reasonably well established in some tris(monothio-acetylacetonato)iron(m) chelates 31 and have been postu-lated to explain the anomalous behaviour of some terpyridyl-cobalt(I1) 33 and tris(dithiocarbamato)iron(IIr) complexes,84although recent low-temperature X-ray studies castdoubts on the applicability of the method to the dithio-carbamate derivati~es.~s We prefer the present quantum-mechanical approach based on the electronic configurationof a single species and find that the only way to satisfactorilyreproduce the higher-temperature data for [{ Co (salen) ),Iand [Co(salen)]*py is to assume that small changes in theligand field occur as the temperature varies. This isperhaps not surprising when excited quartet states are soclose to the ground state.Furthermore, it has been shownthat slight movement of the cobalt(I1) ion out of the ligandplane tends to stabilize quartet states energeti~ally,~~e.g. [Co(msalen)]*H,O * is high spin and the cobalt(I1) ionis 0.2 further out of the plane37 than in [{Co(salen)),]and [Co(salen)].py (each 0.2 A out of plane).z1-23 Thechanges in the ligand field were assumed not to significantlyaffect the molecular orientations and hence not the directioncosines in Table 2.The assumption of changes in ligand-field parameters with temperature has also been recentlyinvoked in the case of some iron(II1) octaethylporphyrinderivatives s* and in earlier e.s.r. studies of some magneticallyanomalous cobalt ~helates.~@It follows from the above discussion that a single set ofbest-fit parameters was obtained for [{ Co(tsa1en) },I and forthe Curie-Weiss regions of [{Co(salen) 12] and [Co(salen) ]*py.For the last complexes some of the ligand-field energy para-meters were then varied as a function of temperature at* msalen = NN'-Ethylenebis( 3-methoxysalicylideneiminato) .a2 M.Cox and J. Darken, Co-ordination Cltem. Rev., 1971, 7 , 29.33 R. C. Stoufer, D. W. Smith, E. A. Clevenger, and T. E. Morris,sa R. L. Martin and A. H. White, Transition Metal Chem.. 1968,35 J. G. Leipoldt and P. Coppens, Inorg. Chem., 1973, 12, 2269.36 J. A. Varga and C . A. L. Becker, Canad. J . Chem., 1974, 52,Inorg. Chem., 1966, 5, 1167.4, 114.79.higher temperature (Table 3). The effect of this is shownin Figure 2. The values of energies of the excited statesdeduced for best-fit should not be taken too literally, sincesmall changes in one parameter can often be absorbed bychanges in another parameter, and the value of c, the spin-orbit coupling constant, was fixed arbitrarily (at a reasonablevalue).However, the deduced energy parameters are in aphysically reasonable range, and their variation is restrictedconsiderably by the necessity of calculating all threeprincipal susceptibilities with one set of parameters.Unfortunately it would be extremely difficult to obtainpolarized spectral data for these complexes to confirm thepositions and assignments of the excited states. Thesolution and solid-reflectance optical data which have beenreported show considerable inconsistencies, although thegeneral spread of transition energies lies within the valuesestimated here.11-13,40 In the case of [{Co(salen) )dl anantiferromagnetic exchange constant, J , of - 40 cm-lwas included in the calculations, and experimental con-firmation of this value was obtained from observation of theN8el temperature a t ca.46 K. A similar value of J wasalso incorporated into the [{Co(tsalen)}.J calculations, nodirect observation of To being available in this case. Itshould be stressed that the curvature in K at high temper-atures for [(Co(salen)),J and [Co(salen)]*py is not due tomagnetic exchange, since J is zero in the pyridine adduct.z2-YZ -YZ-YZ -XZ -X I -XZ -XFIGURE 3 Ordering of out-of-plane d-orbital levels (not toscale)DISCUSSIONIn Figure 3 we show the relative ordering of the d-orbital levels, deduced from the energies of the excited37 M. Calligaris, G. Nardin, and L. Randaccio, J.C.S. Dalton,88 A. K. Gregson, H. A. 0. Hill, and P.Skyte, personal com-J. G. Schmidt, W. S. Brey, and R. C . Stoufer, Inorg. Chem.,40 F. L. Urbach, R. D. Bereman, J. A. Topich, M. Hariharan,1974,1903.munication.1967, 8, 268.and B. J . Kalbacher, J . Amer. Chem. Soc., 1974,96, 60631976 1005states. The anisotropy in the susceptibility of thesemolecules is dominated by the close approach of the dvzorbital to the tiz% ground-state orbital. The complex[(Co(tsalen))J is only weakly anisotropic because of therather strong ligand field exerted by the sulphur atoms,which causes large orbital separations. The dyz orbrtalconsequently lies at a much lower energy in this caserelative to that in [(Co(salen)),] and [Co(salen)]*py, therelative ordering of orbitals being the same for all threecomplexes.It is pertinent to see if there is any relation between theelectronic properties deduced here and the bonding ofthis type of cobalt(@ chelate to molecules such aspyridine and 0,.Pyridine is a weak Q donor (Kb =1.7 x and hence will only react to any significantextent with the cobalt complex if the latter is a suffi-ciently strong Lewis acid or, if this is not the case, ifthere is a significant x bonding to ensure the stability ofthe resulting complex. Both Q and x bonding willgenerally therefore be present in such a cobalt-pyridinelinkage. The complex [(Co(salen))J reacts readily withpyridine to form stable [Co(salen)]*py. The latterreacts further, to some extent, with pyridine. Thecomplex [(Co(tsalen) )J reacts only weakly with pyri-dine.19*41 X-Ray structures on five-co-ordinate pyi-dine 6~24~42-44 or phenyl 46 derivatives of these chelatesshow that the axial group invariably lies almost exactlyin the yz plane.Because of the steric constraintsimposed by the ethylene bridge of the chelate backbonding though x orbitals on the metal will play animportant r6le only if the orbitals are suitably oriented.The appropriate orbital in this situation is the out-of-plane:d,, orbital, which must of course have a favourableenergy. The present results correlate well with this lineof argument, the d,, orbital in [(Co(salen))J being of* 5C1-amben = NN'-ethylene(bis-2-adno-Ei-chlorobenzyl-41 31. F. Corrigan, unpublished work.42 M. Calligaris, G. Nardin, and L. Randaccio, Inorg.Nuckar43 M. Calligaris, G. Nardin, L. Randaccio, and G. Tauzher,ideneiminato) .Glaem. Letters, 1972, 8, 477.Inorg. hTz.rclrnr Chern. Letters, 1973, 9, 419.suitable energy relative to pyridine p, orbitals for strongbonding, whilst that of [(Co(tsalen))d, at much lowerenergy, will lead only to weak bonding. The argumentis further substantiated in the case of [Co(EiCl-amben)],* arelated complex which does not react with pyridine;the d,, and duz orbitals are in reverse order relative to thepresent situation.l69%The bonding of 0, to form 1 : 1 adducts of the type[CoL(py)]*O, (L = chelate ligand) can be rationalizedin a related manner. The 0, molecule has the bentconfiguration and lies in the xz plane perpendicular to theaxial base which is in the yz plane.6*43 This suggeststhat the out-of-plane duz orbital on cobalt may be in-volved to some extent in x bonding to the oxygenmolecule, there being no steric restraints imposed by thechelate ligand. x Bonding to dyd is necessitated by theprior involvement of dz# with the axial base. In arecent bonding scheme deduced for such complexes(L = porphyrin), the x bonding between Co and 0,was considered to be reduced relative to o bonding,because of the bent nature of the Co-0-0 moiety.&The inactivity of cobalt(r1) complexes of the presenttype towards oxygen in the solid state would appear tobe chiefly a function of the crystal packing, since e.s.r.data on active solutions of [Co(salen)] in pyridine suggestthat the electronic structure is very similar to thatdeduced here for the solid. It would be most useful tohave the X-ray structure and crystal anisotropy of theactive solid form.We thank the Australian Research Grants Committeefor support, Professor B. 0. West for advice and encourage-ment, Mr. A. van den Bergen and Mr. M. F. Corrigan forexperimental assistance, and Drs. A. K. Gregson and J. E.Davies for help with computer programs.[6/1484 Received, 28th Judy, 197614* M. Cesari, G. Neri, G. Perego, E. Perrotti, and A. Zazzetta,*6 S. Bruckner, M. Calligaris, G. Nardin, and L. Randaccio,46 B. B. Wayland, J. V. Minkiewicz, and M. E. Abd-Elmageed,Chem. Comm., 1970, 276.Chem. Comm., 1970,162.J . Amer. Chem. Soc., 1974,96, 2796
ISSN:1477-9226
DOI:10.1039/DT9760000999
出版商:RSC
年代:1976
数据来源: RSC