Magnetic Circular Dichroism Studies of Ions in Solutions and Crystals BY P. N. SCHATZ R. B. SHIFLETT J. A. SPENCER A. J. MCCAFFERY' S. B. PIEPHO, J. R. DICKINSON AND T. E. LESTER~ Department of Chemistry University of Virginia Charlottesville, Virginia 22901 U.S.A. Received 2nd September 1969 Magnetic circular dichroism (MCD) studies are reported on several metal d1OsZ systems both in solutions at room temperature and as dopants in crystals at room temperature and liquid helium temperature. Magnetic moments of the excited T1,(3P1) state obtained from the data are con-siderably lower than theoretically expected for octahedral complexes and suggest substantial quenching of angular momentum probably at least in part through the participation of the metal p orbitals in covalent bonding.The TlU(lPl) region clearly indicates distortion from high symmetry in most cases with apparent complete quenching of the angular momentum in the excited state. In addition, high resolution MCD data are presented on CsaZrC16 Os4+ at liquid helium temperature. A vast amount of fine structure is revealed often exceeding in clarity the previously studied absorption spectrum. A preliminary analysis of the results is discussed. The utility of magnetic circular dichroism (MCD) spectroscopy in clarifying spectroscopic assignments characterizing the symmetry of transitions and extracting magnetic moments of excited states has now been widely demon~trated.~ The technique involves the measurement as a function of frequency of the difference induced in the absorption coefficients for left and right circularly polarized light by a longitudinal magnetic field.If the energy levels involved in the transition possess degeneracy they will generally be split by the magnetic field and the selection rules involving the various Zeeman sublevels will differ for left and right circularly polarized light. It is often convenient to distinguish three different effects which give rise in turn to A C and B terms. The first arises if the ground and/or excited state involved in the transition undergoes a Zeeman splitting (and hence possesses degeneracy). The absorption of left and right circularly polarized light will then occur at slightly different energies and a sigmoid dispersion curve ( A term) will result which has its zero at the zero-field absorption ma~imum.~ The C term arises if thermally accessible levels undergo a Zeeman splitting so that the absorption of left and right circularly polarized light occurs from levels with differing populations.This gives rise to a term peaking at the absorption maximum and varying inversely with absolute tem-perat~re.~ Finally the B term in general is always present and arises from field induced mixing of the unperturbed states of the system. It is temperature independent and has the same dispersion form as the C term.3 All of the systems we shall discuss have nondegenerate ground states and hence only A and B terms can occur. In this paper we present some recent applications of the MCD technique to in-organic ions of high symmetry both in solutions and crystals.We start with a room temperature solution study of a number of dlos2 systems which involve broad absorp-tion bands ; we then consider some of these same systems as dopants in alkali halide 1 SCHATZ SHIFLETT SPENCER MCCAFFERY PIEPHO DICKINSON LESTER 15 crystals at low temperature. Finally we discuss a high resolution study of the d4 system Cs,ZrCl Os4+ at liquid helium temperature where extensive fine structure is resolved in both the absorption and MCD spectra. These latter data have been obtained recently and we restrict ourselves at this time to only a few general remarks. dlos2 SYSTEMS The ultra-violet absorption spectra of dl0s2 metal complexes generally show two fairly strong absorption bands interpreted as arising from transitions of the type s2+sp.In the atomic case the configuration sp gives rise to the states 3P2 3P1 3P0, and lP1 and the only allowed transition from the ground state is lS0-+lP1. How-ever lP1 and 3P1 mix under spin-orbit coupling and the lower frequency less intense band is assigned as 1S0+3P1 (made allowed through the admixture of lP1) and the higher frequency band is assigned as lSO-+ lP1. The spectra of these systems has been known for a long time and a large number of studies have been made both in solution and in the solid state. Jarrgensen et aZ.4-6 have given discussions of the spectra with additional literature references and good summaries of the solid state work have been given by McClure and Honma.8 T2u -NO ATOM . s - 0 s-0 0; COUP LING COUPLING FIG. 1 .-Energy levels arising from the excitation s2 +sp for a metal atom in spherical and octahedral symmetry.< is the p-electron spin-orbit coupling constant and the star indicates double group. Throughout this paper term symbols without a left superscript designate double group states. Our MCD study was undertaken to measure excited state magnetic moments and to see if we could throw any additional light on the nature and symmetry of the species in solution and in the crystal. Since the transitions are presumed to be to degenerate excited states A terms are expected. Fig. 1 shows a schematic energy level diagram for the central atom in spherical and octahedral environments. The MCD of [Bu2NH2I3BiBr6 in CH3CN shown in fig. 2 is typical of the solution MCD of all of these systems except Sn(I1).Band 1 is certainly the Alg-+T1,,(3P1) (‘* triplet ”) transition and in all cases we have studied (Table 1) shows a distinct, positive A term (though in a few cases there is clear evidence of some splitting of thi 16 MAGNETIC CIRCULAR DICHROISM STUDIES OF IONS band). There is a small negative MCD at about 31,000 cm-1 which does not corres-pond to any obvious absorption band in solution but corresponds to a clear shoulder in crystalline KBr Bi3+. This is likely due to the vibronically allowed 1S0-+3P2 I -so,ooo -40,000 -30,000 -20,000 - I 0,000 - 0 30,000 40,000 50,000 frequency (cm- l) FIG. 2.-Absorptionand MCD spectrum of [Bu2NH213BiBr6 in CH3CN. [O] M is the molar ellipticity (defined as in natural optical activity in deg. dl dm-' mol-') per gauss in the direction of the light beam.E is the molar extinction coefficient. The numbering of the bands is indicated. TABLE 1 Experimental magnetic moment (AID) in Bohr magnetons for the T1,(3P1) excited state of some dlos2 complexes. A-values determined by the method of l o D-values obtained by numerical integration except as noted. T = 300 K except as noted. metal Sn(1I) Sb(II1) SbClg- SbBrz- TKI) solvent HC1 KCl(s) HC1 HBr CH3CN CH3CN HCl HBr KBr(s) KI(s)e A/Da > O b >Ob >Ob 0.60 >Ob 0.21d 0.86 0.76 0.97 0.84d metal Pb(I1) BI(II1) Bi& BiBrz- * solvent HCl HBr KI(s) KI(s)f HCl HBr KBr(s) CH3CN CH3CN A/Da 0.79 0.91 0.44 0.63d 0.95 0.97 0.68 0.76 0.74 {(c/v)dv; bD could not be reliably measured; a D = 9.1834~ =run as an alkyl ammonium salt; d D obtained by gaussian fit; e 1 2 K ; f16K.transition. Bands 2 and 3 are almost certainly associated with the Al,-+Tl,(lPl) (" singlet ") transition and the much more intense band 4 probably corresponds to a charge-transfer or intra-ligand type transition. In the 'P1 region the MCD first shows a positive peak ( N 36 000 cin-I) which possibly corresponds to a shoulde SCHATZ SHIFLETT SPENCER MCCAFFERY PIEPHO DICKINSON LESTER 17 in absorption and it then shows a negative peak (-40 000 cm-l) which is quickly overlapped by what appears to be the first half of a large positive A term associated with band 4. In contrast to all the other ions studied in solution Sn(1I) in con-centrated HCl (fig. 3) shows an apparently simple MCD throughout the triplet-singlet region.There are distinct positive A terms corresponding to Alg+T1u(3PJ N 36 000 cm-,) and Alg-+Tlu(lPl) (-46 000 cm-l) and the characteristic negative MCD peak in the region of the lS-+ 3P2 shoulder ( - 38 000 cm-l). I I I I I 30,000 4 0,000 50,000 frequency (cm-') FIG. 3.-Absorption and MCD spectrum of Sn2+ in concentrated HCI; symbols and units are as in fig. 2. If we assume initially that we are dealing with octahedral MX6 ions it is not difficult to derive for the A,,+T, transition a general group theoretical expression for the ratio of Faraday A parameter to dipole strength (AID) which in this case is simply the magnetic moment of the excited state. If one assumes that the transition is simply s2-+sp (i.e. that the tFu functions are pure metal p functions) the result is (1) where /3 is the Bohr magneton I c1 l2 and I c2 l2 are respectively the relative weights of lP and 3P1 in the T, excited state (I c c2 l2 = l) and p and p/2 in the parentheses are the spin and orbital contributions respectively of 3P1.If we con-sider the triplet region first (roughly I c2 l2 - 1 I c1 l2 -O) we would predict an A/D value approaching 1.5 P. Inspection of table 1 shows in fact that the experimental AID values are always less than one often considerably so. Quenching of the orbital angular momentum due to mixing of the metal p functions with ligand orbitals (covalency) does not seem able to account fully for the observed reduction in AID, certainly not for Sb(III) since even complete quenching (AID = I c2 I2p) gives an AID = I c1 12P+ I c2 12W+8/2) 18 MAGNETIC CIRCULAR DICHROISM STUDIES OF IONS AID value of around one Bohr magneton due to the spin angular momentum of the triplet state.MCD studies in the triplet region by Yoshikawa and Mabuchi l2 (In+), Onaka et aZ.13 (TI+ and Pb2+) and Topa et a1.I4 (Ag-) also indicate substantial reductions of the magnetic moment from the limiting value of 1.5 p. The singlet region is much more complex and only with Sn2+/HC1 is a simple result obtained. There a clear A term is observed and AID is found to be -0.33 p. Since I cl l2 - I and I c2 l2 -0 a reduction in orbital angular momentum by about a factor of 3 is indicated. This could be due to covalency effects lowering of symmetry or a combination of both. In none of the other systems studied in solution was there any clear indication of an A term in the singlet region.frequency (cm- l) FIG. 4.-ROOm temperature absorption and MCD spectrum of Sn2+ doped into crystalline KCl; symbols and units are as in fig. 2. The concentration of dopant is not known but a nominal value is obtained by comparing the crystal absorption spectrum at room temperature with that of a cor-responding solution spectrum of known concentration. In order to define these systems more clearly and to permit low temperature work, we have extended our studies to alkali halide crystals containing the d10s2 ions as dopants. Fig. 4-6 show some typical results. We note from table 1 that the AID values for the triplet region in the crystals agree reasonably well with the corres-ponding ions in solution suggesting that no drastic lowering of the symmetry occurs in solution.This is of interest because there is not complete agreement that MX6 is always the limiting species in solution though we have always chosen our conditions to favour this. On the other hand it is generally assumed that the metal ion is octahedrally coordinated in the alkali halide host crystal. KC1 Sn2+ (fig. 4) shows the typical complex singlet region. There are three distinct bands (starting with the shoulde SCHATZ SHIFLETT SPENCER MCCAFFERY PIEPHO DICKINSON LESTER 19 35.000 40,000 45,000 frequency (cm-') FIG. 5.-12 K absorption and MCD spectrum of T1+ doped into crystalline KI; symbols and units are as in fig. 2. Dopant concentration determined as in fig. 4 caption.30 2 20 SL 10 0 -10 -20 -30 W 60000 40000 20000 0 2 5. 30.630 35,000 4 0 b O frequency (cm-') FIG. 6.-.16 K (solid line) and room temperature (dashed line) absorption and MCD spectrum of PbZ+ doped into crystalline KI; symbols and units are as in fig. 2. Dopant concentration determined as in fig. 4 caption 20 MAGNETIC CIRCULAR DICHROISM STUDIES OF IONS at -41 700 cm-l) and a correspondingly complex MCD. A great reduction in symmetry of the ion has occurred possibly through a Jahn-Teller effect in the excited state. Unfortunately we have not yet succeeded in making MCD measurements on this system at low temperature. The TI(I)/KI MCD spectrum (fig. 5) in the singlet region appears relatively more simple with no indication of any A term character though there may be significant overlap from the so-called 2240A band which is thought to be due to an intra-ligand (iodide) transition.’ The Pb(II)/KI spectrum (fig.6) in the singlet region is especially complex particularly at low temperature. We have as yet made no attempt at a detailed analysis of these crystal results. In any case with the possible exception of Sn2+/HC1 it seems clear for all of the systems studied both in solution and crystal that there is substantial distortion from octahedral symmetry in the TJP1) excited state apparently with complete quenching of the orbital angular moment urn. I I -__ 1- 1 . I I I I I I I I I I I I I I I I I I I I I I ;,’ I,,,’ I frequency (cm-’) FIG.7.-The liquid helium (solid line) and room temperature (dashed line) absorption and MCD spectrum of crystalline Cs2ZrCls Os4+ over the region 19-24 000 cm-l; symbols and units are as in fig. 2. It has been suggested that some of the ions studied here may be tetrahedrally coordinated. Our MCD studies do not distinguish in any simple way between octa-hedral and tetrahedral species since one finds that eqn. (1) applies in both cases. CszZrCls Os4+ High resolution MCD measurements on crystals at low temperatures has great potential because of the possibility of studying individual vibronic lines rather than the broad bands generally present in solution spectra at room temperature. We have recently succeeded in interfacing a Spex 1400-11 2 m double monochromator with our Durrum-Jasco CD spectrophotometer.One of the first systems we studied was Cs2ZrCl Os4+ at liquid helium temperature and we were able to obtain excellen SCHATZ SHIFLETT SPENCER MCCAFFERY PIEPHO DICKINSON LESTER 21 quality high resolution MCD data. Most of our results all from one small dilute crystal are summarized in fig. 7-1 1 the absorption spectra having been recorded on a Cary 14. f I I ! I I -13000 - ROO0 - 6000 -;OW -2WO 24,000 2 5,000 2 6,000 2 7,000 frequency (cm-') FIG. 8.-The liquid helium (solid line) and room temperature (dashed line) absorption and MCD spectrum of crystalline Cs2ZrC16 Os4f over the region 24-27 000 cm-' ; symbols and units are as in fig. 2. The labelled points are discussed in the text. In regions containing sharp features we were able to make MCD measurements at spectral slit widths in the approximate range 2-6cm-l.Except in the 19000-22 000 crn-I region where the MCD is very small we had an excellent signal-to-noise ratio and were able to obtain clear reproducible data. (We have obtained a nominal value for the product of concentration and path length (cl) in the crystal by setting ( N 26 000 cm-') for the crystal at room temperature equal to 9 600 the value observed 24 in [Bu4V20sCl6/C2H4Cl2 at room temperature. In quantitative con-siderations ratios are always considered which are independent of cl.) The absorption spectrum of Os4+ in single cubic crystals Of K2PtCl6 and Cs2ZrC16 at 4.2 K have been studied in detail by Dorain Patterson and Jordan (hereafter abbreviated DPJ).They found that many of the transitions are characterized by series of very narrow absorption lines. All the bands below 35 OOO cm-l were interpreted as d-+d transitions on the basis of a crystal field calculation. Previously, those transitions occurring in solution between 23 900 and 30 000 cm-I had been assigned to charge transfer transitions since their extinction coefficients in solution range up to 9 000. Intensity calculations were made for the vibronically allowed d+d transitions and it was determined that the more intense transitions would be those from the A1J3T1,) ground state to the Tlg or T2g excited states with one quantum of the vq(tlu) v3(tlu) v7(tlu) or Vg(t2u) odd vibrational modes excited. Experimentally 22 MAGNETIC CIRCULAR DICHROISM STUDIES OF IONS vibronic transitions of this sort are found to occur as progressions in the totally symmetric vibration superimposed on one quantum of an odd vibration.The authors found that an excellent fit of experimental energies with those from their crystal field calculation resulted if it was asszrrned that transitions to T2 +~4(t,,) +n v,(a,,) excited states were much the most intense transitions. l 2 frequency (cm-') FIG. 9.-The liquid helium (solid line) and room temperature (dashed line) absorption and MCD spectrum of crystalline CszZrCls Os4+ over the region 27-28 500 cm-' ; symbols and units are as in fig. 2. Our absorption spectra agree well with DPJ and though the crystal is rather dilute so that we do not observe many of the weaker lines observed by them we can clearly make some preliminary tests of their detailed assignments.Os4+ has a 5d4 configuration. In a strong cubic electrostatic field the ground state is the 3T1g state of the tiQ configuration. But since the spin-orbit interaction is large (lsd-2 000-3 000 cm-l) the 3T1g state will be split into states of Alg TI, Es and T2g symmetry with the A state lying lowest by several thousand cm-l. Thus the MCD should consist of A and B terms only. When the ground state is nondegenerate the experimental parameter AID can be calculated as the product of a group theoretic factor and the magnetic moment of the excited state. If the transition is vibronically allowed and standard vibronic theory1** l9 is used AID will depend on the symmetry of the excited state (Tlg or T,,) the symmetry of the odd vibration (tl or t2,) and the sign and magnitude of the excited state magnetic moment.For example for an A1g-+T2g transition made allowed by vibronic coupling to a v4(tlu) v3(tlU) or v7(tlu) vibration (2) AID = *(T2,1 I IL? I T2,l) SCHATZ SHIFLETT SPENCER MCCAFFERY PICPIIO DICKINSON LESTER 23 10 2: 22 -2 c -2 -6 6 030 450C 3 On@ 1mo C frequency (cm-l) FIG. 10.-The liquid helium (solid line) and room temperatwe (dashed line) absorption and MCD spectrum of crystalline CsrZrC16 Os4+ over the region 28 500-32 000 cm-l ; symbols and units are as in fig. 2. The labelled points are discussed in the text. I I I I I I I I I frequency (cm-') FIG. 11. The liquid helium (solid line) and room temperature (dashed line) absorption and MCD spectrum of crystalline CszZrC16 Os4+ over the region 32-34 OOO cm-l; symbols and units are as in fig.2 24 MAGNETIC CIRCULAR DICHROISM STUDIES OF IONS where p = - p(Lz + 2Sz) while with a v,(t2,) vibration AID = -3(Tzg1 I p:B I T*gl). (3) The notation and symbols are those of Griffith.zO As an initial step in interpreting our MCD data we made a complete crystal field calculation using the DPJ parameters and with the eigen-vectors obtained calculated the entire magnetic moment matrix. All matrix elements for the spin-orbit and crystal field interactions were machine computed in the weak field basis using input from the d4 matrices of Nielson and Koster.21 The magnetic moment matrix was also calculated by computer using standard formulasz2 and orbital reduction factors of one.Agreement with the eigenvalues published 1 5 3 l7 was to within 4 1 cm-l in all cases and the g factors published15 were reproduced with a maximum deviation Using our magnetic moments we are thus able to calculate the AID value pre-dicted by the DPJ assignment for each line. Our MCD data indicate a number of difficulties with their assignments. Even with our relatively dilute crystal there are two spectral regions in which we clearly observe an MCD though the corresponding absorption spectrum is almost undetectable. The first such region (19-22 000 cm-l) shows only some broad weak MCD but the second (23-23 900 cm-l) exhibits a totally symmetric progression in at least six distinct features. We hope to be able to identify the absorption bands to which this MCD corresponds when we run more concentrated crystals.It is possible that these bands correspond to transitions between states primarily of different spin multiplicity. The MCD of such transitions is frequently an order of magnitude more intense than the optical spectrum of the corresponding absorption^.^ MCD in this spectral region might indicate that the 5Eg state lies considerably higher in energy than predicted by DPJ who assign none of their transitions to the 23 000 cm-' region. With the dilute crystal used for our studies no MCD or absorption spectra were observed for the band at 17 000 cm-l. We hope to examine the spectral region below 25 000 cm-l with a more concentrated crystal. The first intense absorption band (26-27 000 cm-l) shows a clear progression of negative A terms in MCD.Using the DPJ parameters and assignment (Alg-+TZg (3Tzg)) the observed sign of the A terms is obtained only if V g ( t 2 ) is chosen as the intensity-producing vibration. This clearly contradicts the DPJ assumption that the most intense vibronic lines arise from transitions to T2g + ~ 4 ( t 1 ) +nvl(alg) excited states. The second intense band (28 500-32 000 cm-l) assigned by DPJ as Alg+T2g(3A2g) seems from the MCD to consist of two different transitions (fig. lo) although in the absorption spectrum the dominant peaks appear to be members of progressions in the totally symmetric vibration which proceed through the entire band. The overall shape of the band is however asymmetrical. The two transitions might correspond to the peak and shoulder observed at 29 200 and 29 900 cm-' respectively in the solu-tion spectrum of (NH4)20sC16.z4 The MCD in fig.10 shows a totally symmetric progression in a complex unit containing five distinct features for what appears to be the first transition of the band. Lines A and B give negative B terms lines C and C' negative A terms while line D gives a positive B term. These features are repeated several times but at line L seem to disappear. The spectrum of the second transition of this absorption band is less clear but the MCD appears to be dominated by a positive A term together with a positive B term for each of the lines P R S and T. The DPJ assignments are consistent with our MCD for lines C G J and L but we do not observe any angular momentum for lines A E and I lines B and F or lines D, H and K which by the DPJ assignment can show A terms.of 50.01 p SCHATZ SHIFLETT SPENCER MCCAFFERY PIEPHO DICKINSON LESTER contradicts. For example they have assigned lines G and K (fig. 8) as 25 There are also other details in the exhaustive DPJ assignments which our MCD while our MCD gives a positive A term for each of these lines and thus demonstrates that the excited state must be degenerate. Certainly DPJ's most interesting assertion is that all transitions in the Os4+ spectrum below 35 000 cm-1 are ligand field (d-+d) transitions. We have measured the dipole strength of the two intense absorption bands-at room temperature by a gaussian fit (which was quite convincing) and at liquid helium temperature by numeri-cal integration.The room temperature dipole strength of the first band (-25-27 000 cm-I) is about twice the liquid helium temperature value while for the second (-28 500-32 000 cm-l) it is about three times that at liquid helium temperature. Simple theory1 * predicts a temperature dependence for oscillator strengths of approxi-mately this magnitude for vibronically allowed transitions and thus the intensity ratios support the DPJ contention. It is also possible that these transitions are partly or entirely parity forbidden charge-transfer bands. Such transitions would be expected to have greater oscillator strengths than d-+d transitions. It is clear from this preliminary discussion that the high resolution MCD spectrum contributes a great deal of additional information regarding the interpretation of the absorption spectrum.It is relatively easy to check suggested assignments and in spectra of the complexity observed for Cs,ZrCl Os4+ it is not surprising that contra-dictions are found in assignments based on the absorption spectrum alone. A more challenging task is to propose detailed assignments which fit both sets of data. In a forthcoming paper we argue that much of our observed spectrum arises from charge-transfer transitions contrary to the DPJ interpretation. We are greatly indebted to Prof. Paul B. Dorain for supplying the Cs,ZrCI Os4+ crystal used in this work. We thank Mr. R. L. Mowery for much help with the data processing. One of us (S. B. P.) acknowledges support under an N.D.E.A.fellowship. This work was supported by a grant from the National Science Foundation. permanent address Department of Chemistry University of Sussex Brighton Sussex. permanent address British Petroleum Co. Ltd. Sunbury Research Laboratories Sunbury, Surrey. For a review of the theory and recent experimental work in this field see P. N. Schatz and A. J. McCaffery Quart. Rev. 1969 552; A. J. McCafTery P. N. Schatz and T. E. Lester J. Chem. Phys. 1969 50 379 and references therein. C. K. Jerrgensen Absorption Spectra and Chemical Bonding in Complexes (Pergamon Press, Oxford 1962); chap. 10. C. K. Jlzrrgensen Halogen and the Noble Gas Complexes in Halogen Chemistry Vol. 1 ed. V. Gutmann (Academic Press New York 1967). R. A. Walton R. W. Matthews and C .K. Jerrgensen Inorg. Chim. Acta 1967 1 355. York 1959) p. 162. C. H. Henry S. E. Schnatterly and C. P. Slichter Phys. Rev. 1965 137 A583. ' D. S. McClure Electronic Spetrca of A/iolecules and Ions in Crystals (Academic Press New * A. Honma Sci. Light 1967 16 229. l o P. J. Stephens Chem. Phys. Letters 1968 2 241. I ' See for example Schatz et al. J. Chem. Phys. 1966 45 722. l 2 A. Yoshikawa and T. Mabuchi J. Phys. SOC. Japan 1968,24 1405. l3 R. Onaka T. Mabuchi and A. Yoshikawa J. Phys. SOC. Japan 1967,23 1036. l4 V. Topa L. Taurel J. C. Rivoal and B. Briat Phys. Stat. Sol. 1969 33 K17. l 5 P. B. Dorain H. H. Patterson and P. C. Jordan J. Chem. Phys. 1968 49 3845. l6 C. K. Jerrgensen Disc. Faraday Soc. 1958 26 175; C. K. Jnrrgensen Mol. Phys. 1959,2,309 ; C. K. Jrargensen 2. Naturforsc. 1967 22 945; R. B. Johannesen and G. A. Candela Inorg. Chem. 1963 2 67 26 MAGNETIC CTRCULAR DTCMROTSM STUDIES OF IONS P. C. Jordan H. H. Patterson and P. B. Dorain J. Chem. Phys. 1968,49 3858. C. J. Ballhausen Introduction to Ligand Field Theory (McGraw-Hill Book Co. Inc. New York 1962) chap. 8. l9 P. J. Stephens J. Chem. Phys. 1966,44,4060. *O J. S. Griffith The Theory of Transition-Metal Ions (Cambridge University Press CAmbridge, England 136J) table A16. C. W. Nielson and G. W. Koster Spectroscopic Coefficients for thepn dn and f n Configurations (M. I. T. Press Cambridge Mass. 1963). 22 ref. (20) chap. 5. 23 A. J. McCaffery P. J. Stephens and P. N. Schatz Inorg. Chem. 1967 6 1614. 24 G. N. Henning Dim (University of Virginia June 1968) p. 84-86