High Resolution Magnetic Circular Dichroism and Absorption Spectra of Cs2ZrBr6:Ir4+

High Resolution Magnetic Circular Dichroism and Absorption Spectra of Cs2ZrBr6:Ir4+ J. R. Dickinson, S. B. Piepho, J. A. Spencer, and P. N. Schatz Citation: The Journal of Chemical Physics 56, 2668 (1972); doi: 10.1063/1.1677595 View online: http://dx.doi.org/10.1063/1.1677595 View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/56/6?ver=pdfcov Published by the AIP Publishing Articles you may be interested in Communication: Fullerene resolution by the magnetic circular dichroism J. Chem. Phys. 138, 151103 (2013); 10.1063/1.4802763 Photovoltage detection of x-ray absorption and magnetic circular dichroism spectra of magnetic films grown on semiconductors J. Appl. Phys. 93, 2028 (2003); 10.1063/1.1537453 A combination spectrophotometer for measuring electronic absorption, natural circular dichroism, and magnetic circular dichroism spectra Rev. Sci. Instrum. 61, 2073 (1990); 10.1063/1.1141420 Confirmation of a Ham effect in the 22 900 cm−1 band of Cs2ZrCl6:Ir4+ J. Chem. Phys. 61, 4868 (1974); 10.1063/1.1681815 High Resolution Magnetic Circular Dichroism Spectrum of Cs2ZrCl6:Ir4+ J. Chem. Phys. 57, 982 (1972); 10.1063/1.1678349 This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 155.33.120.209 On: Sat, 22 Nov 2014 02:12:48 http://scitation.aip.org/content/aip/journal/jcp?ver=pdfcov http://oasc12039.247realmedia.com/RealMedia/ads/click_lx.ads/www.aip.org/pt/adcenter/pdfcover_test/L-37/327320036/x01/AIP-PT/JCP_ArticleDL_101514/PT_SubscriptionAd_1640x440.jpg/47344656396c504a5a37344142416b75?x http://scitation.aip.org/search?value1=J.+R.+Dickinson&option1=author http://scitation.aip.org/search?value1=S.+B.+Piepho&option1=author http://scitation.aip.org/search?value1=J.+A.+Spencer&option1=author http://scitation.aip.org/search?value1=P.+N.+Schatz&option1=author http://scitation.aip.org/content/aip/journal/jcp?ver=pdfcov http://dx.doi.org/10.1063/1.1677595 http://scitation.aip.org/content/aip/journal/jcp/56/6?ver=pdfcov http://scitation.aip.org/content/aip?ver=pdfcov http://scitation.aip.org/content/aip/journal/jcp/138/15/10.1063/1.4802763?ver=pdfcov http://scitation.aip.org/content/aip/journal/jap/93/4/10.1063/1.1537453?ver=pdfcov http://scitation.aip.org/content/aip/journal/jap/93/4/10.1063/1.1537453?ver=pdfcov http://scitation.aip.org/content/aip/journal/rsi/61/8/10.1063/1.1141420?ver=pdfcov http://scitation.aip.org/content/aip/journal/rsi/61/8/10.1063/1.1141420?ver=pdfcov http://scitation.aip.org/content/aip/journal/jcp/61/11/10.1063/1.1681815?ver=pdfcov http://scitation.aip.org/content/aip/journal/jcp/57/2/10.1063/1.1678349?ver=pdfcov THE JOURNAL OF CHEMICAL PHYSICS VOLUME 56, NUMBER 6 15 MARCH 1972 High Resolution Magnetic Circular Dichroism and Absorption Spectra of CS2ZrBr6:IrH* J. R. DrCKINSoN,t S. B. PrEPHo,t J. A. SPENCER,§ AND P. N. SCHATZ Department of Chemistry, University of Virginia, Charlottesville, Virginia 22901 (Received 13 October 1971) The high resolution absorption and MCD spectra of IrH doped into the cubic host Cs,ZrBr. are reported over the range "'11 000-20000 cm-I at liquid helium temperature and as a function of temperature. This host has not been used previously for optical studies and yields an extraordinarily well resolved IrH spectrum with individual vibronic linewidths in the range "'3-15 em-I. A study of the detailed vibronic structure of three of the bands including the associated "hot" absorption proves very helpful in making definitive as- signments. It is shown clearly that the strong bands at "'17000 and 19000 em-I correspond to the allowed ligand-to-metal charge-transfer transitions, Eo" ('T,o)----> Eu" ('T,u) and Eo" (2T,o) ---->U u' ('T,u) , and that in each case the extensive fine structure is associated with a single allowed electronic transition. The third band which shows sharp vibronic structure, at ",14500 em-I, is very likely the parity-forbidden charge- transfer transition Eo"---->Eo' ('Tlo ) whose lower spin-orbit component, Eo"---->Uo' ('Tlo ), corresponds to a pair of weak absorptions ",12 000 em-I. The present results are consistent with our previous assignment of the intense band system ",13000 em-I to the ligand-to-metal charge-transfer transition Eo"---->Uu'('Tlu ) split by Jahn-Teller interaction. The other component of this transition, Eo"---->Eu' ('Tlu ) , is orbitally for- bidden but is nominally assigned to an intense band overlapping the high energy side of the Eo"---->Eu" ('T'n) transition with which it is strongly mixed. Aside from some very weak absorption in the 15000 em-I region, there seems to be no basis for assigning any of the spectral features observed to ligand field transitions. I. INTRODUCTION We have previously studied the magnetic circular dichroism (MCD) spectra of a variety of metal ions, usually in environments which give broad-band or at will withstand detailed analysis and to assess the general utility of high resolution MCD spectroscopy. Recently, sharp-line optical spectra and accompany- ing detailed interpretations have been published for ReH ,1 OSH,2 and Ir4+ 3 doped into Cs2ZrCl6 and related hosts. High resolution MCD results now also exist on 80 l 0~'---------L~-----4----~~--~~--~ 8000- € 4000-j 11000 11800 12600 13400 14200 15000 FREQUENCY (em-') FIG. I. Absorption and MCD spectrum of Cs,ZrBr6:lrH. [O]M is the MCD in molar ellipticity units (defined as in natural optical activity in degrees· deciliter decimetecI·mole- l ) per gauss in the direction of the light beam. [O]M=3.30X1()3 (EL- ER) III. E is the molar extinction coefficient. The MCD and sharpest absorption spectrum (solid line) were run at ~8°K, the dashed absorption spectrum was run at liquid nitrogen temperature, and the broad solid one at room temperature. The bands are numbered in accordance with our discussion in the text. 13500 10500 € 7500 4500 1500 -c:-=------T 16000 o 5 8 ----,-----1 ------r-.~ T--- 17000 18000 19000 20000 21000 FREQUENCY (em-I) FIG. 2. Absorption spectrum of CS2ZrBr6:Ir4+. Units, notation, and temperatures are as in Fig. I. OSH,4.5 IrH ,5 and ReH S in CS2ZrCls, and preliminary analysis leads to markedly different spectral interpreta- tions from thosel-3 based on the optical spectra alone. In this paper, we shall analyze in some detail the best partially resolved spectra, and have proposed absorption and MCD spectra of IrH in the closely assignments on the basis of such data. It is of consider- related host lattice, Cs2ZrBrs. This lattice has not able interest to apply these techniques to sharp, well- previously been used in optical studies and provides an resolved vibronic lines of the same ions at low tempera- environment which yields an extraordinarily well- ture both to ascertain whether our previous assignments resolved Ir4+ spectrum at low temperatures which is 2668 This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 155.33.120.209 On: Sat, 22 Nov 2014 02:12:48 MAG NET I C D I C H R 0 ISM AND S P E C T R A 0 F C S 2 Z r B r s : I r 4 + 2669 FIG. 3. Absorption and MCD spectrum of CS2ZrBrs: Ir4+ including hot bands. Units and notation are as in Fig. 1. The cold spectrum was run at ~8°K. The hot bands have been blown up along the ordinate for clarity the factors being 10 and 2 for [OJM and E, respectively. € -20 -80 3200 1600 o 14150 probably the easiest to interpret of the systems men- tioned. In subsequent papers, we plan to discuss the high resolution MCD spectra of several second and third transition series ions both in C&.lZrCI6 and Cs2ZrBrs. Ir4+ has been of particular interest to us because its d5(t2g5) configuration gives rise to relatively few low- lying ligand-to-metal charge-transfer excited states. With fewer plausible alternatives, allowed charge- transfer transitions can be assigned with considerable confidence on the basis of MCD data. The Cs2ZrBr6:IrH spectrum (Figs. 1-5) shows incredible detail with linewidths generally in the 3-15 cm-1 range-far narrower than are usually observed through charge-transfer bands in transition-metal com- pounds. The spectra are much superior to the best of those we reported recently1 in various tin hexahalide hosts. The narrow linewidths have allowed us to study 120 [elM -40 25.3"K II'K 1\ ~ f\ " 1\ I'- "V 1J '.J ~ v 25.6°K \. J v HOT BAND 5 BANDS 2 ~ 5 25.So K 1 3h 2h Ih .JJ .A 1 v-Z .A ..A ./'--.. 14250 14350 14450 14550 14650 14750 FREQUENCY (em-I) the hot bands associated with three different electronic transitions thereby furnishing additional insight into the origin of fine structure observed, both here and, by analogy, in related compounds. Recently we measured the MCD of several groups of vibronic lines in Cs2ZrCls:UH, a system whose vibronic structure is relatively well-understood, and obtained encouraging agreement between experimental Faraday parameters and values calculated using first-order vibronic theory.s In the present paper we extend these calculations to the octahedral double group states and use them in interpreting our results. II. EXPERIMENTAL Our low temperature experimental techniques have been described previously.9 The crystals of C&.lZrBr6 FIG. 4. Absorption and MCD spectrum of CS2ZrBrs:Ir4+ including hot bands. Units and notation are as in Fig. 1. The cold spectrum was run at ~8°K. The hot bands have been blown up along the ordinate for clarity the factors being 40 and 3.75 for [OJM and E, respectively. -200 ~ HOT 600 BANDS 0 € 16900 17100 17300 17500 17700 17900 18100 18300 FREQUENCY (em -I) This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 155.33.120.209 On: Sat, 22 Nov 2014 02:12:48 2670 DICKINSON, PIEPHO, SPENCER, AND SCHATZ -40 -200 HOT BAND 8 10000 BANDS 4 E ~~ 3 I 14.6°K i 25.9' K • 5000l 43' K·" 1 o L--"""r""""'1P" .--''-r--,~~- 18500 18700 18900 FREQUENCY (em -') were grown by slowly heating to the melting point a sealed, evacuated ampoule containing an intimate mixture of CsBr and ZrBr4. (CAUTION: Explosions often occur in this procedure.) Preliminary room tem- perature x-ray photographs taken and analyzed by Dr. Peter T. Greene indicate that CS:!ZrBr6 belongs to space group Fm3m with a~1O.83 A. The crystals are extremely hygroscopic and deteriorate rapidly if exposed to the atmosphere. The doped crystal, Cs2ZrBr6:Ir4+, was prepared by adding a small amount of Cs2IrBr6 to the Cs.2ZrBr6 and passing the solid slowly through a Bridgeman furnace at about 7S0°C. The re- sulting red crystals were dissected and small pieces were extracted and polished (all in a dry atmosphere). These were of remarkable optical quality (in contrast to the tin hosts) showing no strain birefringence at low tem- perature. The MCD data obtained thus are of excellent quality with no appreciable baseline uncertainties. The data (Figs. I-S) throughout should be reliable to ±20% or better. We have not been successful to date in quantitatively analyzing our small crystal samples for Ir4+. Therefore, our [O]M and E values (Figs. I-S) are nominal and have been obtained by assuming that Emax for band 3 at room temperature is the same as the corresponding value for [(C4H9)4N]2IrBr6 in CH2Ch solution at 300oK. The important quantities in our theoretical discussion are always ratios which are independent of assumptions about concentration. It is conceivable that some Ir3+ may be present as an impurity in our crystals. However, IrBr63- in solution at room temperature shows1o only very weak d-d ab- sorption (Emax"-'30) in the region we have studied. Though our CS:!ZrBr6:Ir4+ E values are nominal, there is no reason to doubt their order of magnitude validity. Thus, the absorption associated with any Ir3+ present should be several orders of magnitude less than for 19700 19900 FIG. 5. Absorption and MeD spectrum of CS2ZrBr6:IrH including hot bands. Units and notation are as in Fig. 1. The cold spectrum was run at ,...,goK. The hot bands have been blown up along the ordinate for clarity the factors being 20 and 4.17 for [OJM and E, respectively. Ir4+, and it is extremely unlikely that any important features of our spectra can be attributed to this im- purity. III. RESULTS AND DISCUSSION The general outlines of the Ir4+ hexachloride and hexabromide spectra are reasonably well understood as a result of our previous MCD studies.7 •9 •11 The pre- dominant bands below 27 000 and 20 000 cm-I, re- spectively, are assigned to the allowed ligand-to-metal charge-transfer transitions arising from the tlu (1r+u)---? t2g and t2,,(1r)---?t2g excitations. The strong-field d5(t2g5) configuration of the Ir4+ hexahalide ions is ideal for the study of these transitions since excitations from the ligand orbitals fill the t205 hole and each of the single- hole excited state configurations (-y5t2g6) gives rise to only one state, neglecting spin-orbit coupling. When spin-orbit coupling is introduced, each of these states splits into two (Fig. 6), the magnitude of the splitting being directly proportional to the halide ion (but independent of the metal) spin-orbit coupling constant. Since .\Br"-'2460 cm-I and '\cl,,-,S90 cm-I, the bromide splittings are considerably larger and this constitutes an important simplifying feature of the spectra since the U ,,' eT2,,) and E,," eT2,,) states are cleanly sep- arated (Fig. 2). Arguments given previously support assignment of the 13 00(}-14 000 cm- I band (bands 3 and 4, Fig. 1) and the 17300 cm- I band (band 7, Figs. 2 and 4) in IrBr62- to the Eo" eT2g)---?U ,,'eTI,,) and Eo"---?Eu ' eT1u) transitions, respectively, arising from the tl .. (-11"+U)---?i2g excitation and the 16800 cm- I band (band 6, Figs. 2 and 4) and 18300 cm-I band (band 8, Figs. 2 and S) to the Eu"---?E,," eT2,,) and Eu"---?U .. ' eT2U ) transi- tions, respectively, both arising from the t2,,(-rr )---?t2g excitation; all but the Eo"---?E./ transition are allowed in 0,,* symmetry (see Fig. 6). Several other forbidden This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 155.33.120.209 On: Sat, 22 Nov 2014 02:12:48 MAG NET leD I C H R 0 ISM AND S PEe T R A 0 F C S 2 Z r B r 6: I r 4 + 2671 bands are present in the 11 000-20000 cm-1 region. We give arguments in Ref. 7 supporting assignment of bands 1, 2, and 5 (Figs. 1 and 3) to vibronically allowed, electronically forbidden charge-transfer transitions. We have suggested on the basis of our data for 11""+ in the tin hexabromide hosts that the shape of bands 3 and 4 arises from a Jahn-Teller splitting of the U ",'eTl ",) state.7 Figure 1 illustrates that the temperature de- pendence of these bands and the increase in splitting from 380 cm-l at 5.2°K to 555 cm-l at room temperature follows the pattern of our earlier data, although band 4 seems to be somewhat less intense at the lower tempera- tures than in some of the tin hexabromide hosts. Bands 5, 6, and 8, at liquid helium temperature, exhibit structure in Cs2ZrBrs:I1""+ even more extensive than that found for OSH and 11""+ in Cs2ZrCIs or for 11""+ in K2SnBrs. The origin of the structure through the strong bands in the latter compounds proved puzzling and the sharp lines have been assigned to d-d or other electronically-forbidden transitions2.3 •7 ; these assign- ments were based either on the assumption that the band(s) are completely d-d,2 or that the structure arises from forbidden transitions which overlie allowed charge-transfer transitions.3 •7 Our new data show that both assumptions are incorrect. In order to clarify the origin of the fine structure, we have studied the hot band patterns obtained for the crystals as the temperature is raised. Such experiments give comparable information to that obtained from emission studies yet have the advantage that data may be obtained for any band with sufficiently sharp struc- ture whose initial lines are not appreciably overlapped by lower energy bands. CS2ZrBrs:I1""+ is particularly well-suited for these studies as the spectrum contains three well-isolated bands and sharp-line structure for two of the three persists even at liquid nitrogen tem- perature. Hot band data is presented in Table I and Figs. 3-5. Turning first to the U ",' (2T2",) band, Fig. 5 shows the absorption spectrum at temperatures between '"'-'8 and 43°K. The pattern of hot and cold lines clearly indicates that the initial strong line is a no-phonon line and that the additional lines are vibronic structure associated with the same electronic transition (see detailed dis- cussion below). The strongest series of lines is a pro- gression in the totally symmetric mode built upon the no-phonon line as expected for an allowed transition. Hot and cold absorption associated with the Eo"----+ E"," (2T2",) transition (band 6) is presented in Fig. 4. Here the lines broaden much more quickly as the tem- perature is raised. Once again the data strongly support assignment of the initial line as a no-phonon line. We note that the temperature dependence of both bands 6 and 8 also indicates that they arise from electric-dipole allowed transitions (Fig. 2). Finally, Fig. 3 gives the hot and cold absorption associated with the 14 500 cm- l band (band 5). The absence of a no-phonon line, as U~ (-.15) [1/3] 2 " T2U " --=--< , '" Eu (0) [0] ',i 1 1''--,,--- 1 Eu (.60) [213] 1 1 I 2 I TIU I E~ (0)[0] ---- 2672 DICKINSON, PIEPHO, SPENCER, AND SCHATZ TABLE 1. Cs.ZrBr6: IrH hot band data near liquid helium temperature compared with ground state vibrational energies. Band 5 Figure 3 Hot bands Band 6 Figure 4 Band 8 Figure 5 Ground state vibrational frequencies at room temperature Energy' (em-I) Energy Energy Hot band Line Line (em-I) Line (em-I) assignment Mode lit 111 36 liz 46 38 47 Lattice modes (46)- (49)8 26 37 51 (~70)b ~(7o-75)b "S(tl.) "9 (t •• ) "7 (tIu) 60 61 21z 89 "6 21z 108 21z 108-112 "" (75)- 109 3h 116 31z ~151 4lt ~21O "4 "2 "I '" (tlu) "2(e.) "I (al.) "3 (tlu) 82 235 118 138 185 222 ",258 "I + lattice • Measured from the virtual origin estimated at 14 401 cm~l. d M. Debeau and H. Poulet, Spectrochim. Acta 25A. 1553 (1969). b Very weak lines. e Parentheses designate calculated values. eM. Debeau, Spectrochim. Acta 25A. 1311 (1969). an absorption band is then expanded in terms of the zero-order functions as a sum over vibronic components where Do= (ell m I ea)(v/l va), D 1 = L [L (ell (aHjaQi)O I ek)(ek I m I ea)(v/l Qi I va) + L (ell m I ek)(ek I (aHjaQi)O I ea)(v/l Qi I Va)] (2) i 'k""! E(ej)-E(ek) 'k"'e. E(ea)-E(ek) etc. Here e=ifi.(Qo)' v=ifiv(Q) and m is the electric dipole moment operator. ifie(Qo) and ifiv( Q) are electronic and vibrational wave functions respectively, g and f label ground and excited state functions, and summa- tion over components of degenerate representations is understood. Do will be nonzero for electronically allowed transi- tions (e/lmlea)~O) for which vaxvi is totally symmetric. Thus for allowed bands, progressions (~nk= 0, ± 1, ±2,' .. ) occur in totally symmetric modes (nk is the number of quanta of the kth vibra- tional mode excited). For nontotally symmetric vibra- tions, ~nk = 0 transitions are by for the most intense as long as the symmetry in the upper and lower states is the same. This is the case because the potential mini- mum must occur at the same value of any antisymmetric coordinate (at Q=Qo) irrespective of any expansion or contraction of the molecule that conserves the sym- metry; ~nk~O transitions then occur appreciably only when I'k varies enormously from ground to excited state. 12 Thus in Oh symmetry the progression in the I'I(ala) totally symmetric mode is usually the only set of "allowed" lines observed. Line intensities are governed by the Franck-Condon overlap factors, (VII va). Dl may be nonzero if v I x Va contains a representation in common with el xm x ea. Thus both electronically forbidden and allowed transitions contribute to D1• Here, for the non totally symmetric vibrations, ~nk = ± 1 lines make the largest contributions for reasons anal- ogous to those in the previous paragraph [unless J ahn- Teller interactions are important, in which case Eq. (2) is not applicable]. Superimposed upon these lines are progressions in the totally symmetric mode(s), where the line intensities within a progression will be governed by overlap factors as in Do. Quadratic terms in the Hamiltonian [Eq. (1) ] allow combinations of antisymmetric modes to gain intensity. Combination lines, and progressions in the totally symmetric modes built upon them, are generally assumed to be less intense than those allowed through lower degree terms in the Hamiltonian. All of the above discussion applies to the J ahn-Teller susceptible U' states when J ahn-Teller effects are negligible. J ahn-Teller interactions will split the ea and t2a vibronic levels into the number of states contained in e x v and thus both splittings of these lines and pro- gressions in these modes may be observed. It follows, however, that no-phonon lines cannot be split by Jahn- This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 155.33.120.209 On: Sat, 22 Nov 2014 02:12:48 MAG NET leD I C H R 0 ISM AND S PEe T R A 0 F C S 2 Z r B r 6 : I r 4 + 2673 Teller interactions. For Eu"-+U ... ' transitions, (unsplit) t2v and eg vibronic lines are also permitted by the Herz- berg-Teller mechanism [Eq. (2) ] with selection rule, Ank= ±1. Lines may also be split by low symmetry perturbations and anharmonicity. The discussion so far has been limited to molecules or molecular ions and has neglected all solid state effects. The "internal" modes (Ill through 116) of octahedral complex ions doped into hosts with the K2PtCl6 struc- ture might be expected to have energies close to those found for the ions in solution. Experimentally, this has been demonstrated for several of the ligand-field bands of ReH 1 and UH 13 doped into Cs2ZrCI6. In the crystal, however, additional modes representing the motion of the complex ion with respect to the cations ("lattice" modes) may also couple with electronic motions. "Lattice" modes are expected at lower frequencies than the "internal" modes and we assume that the very low energy vibrations « 70-80 cm-1) arise from the former though we do not attempt detailed assign- ments. In general, the Kramers doublet ground state should have less complex vibronic structure than the excited states. Furthermore, the ground state symmetry is probably very close to Ok since the hot band data indi- cates parity is a good quantum number, none of the no-phonon lines appear split, and progressions occur in only one mode. Thus, the hot band energies associated with parity-allowed electronic transitions should reflect even vibrational mode frequencies while those asso- ciated with parity-forbidden transitions should reflect the odd ones. Therefore, hot band data in favorable cases allows one to distinguish between orbitally for- bidden (but parity-allowed) and parity-forbidden electronic transitions. Vibrational energies from hot band data are given in Table I together with tentative assignments. The fre- quencies of the vibrational modes of CS2IrBr6 have not been reported but may be inferred from infrared and Raman spectra of related hexabromide ions (Table I). The even-mode assignments for the lines of bands 6 and 8 are consistent with our assignment of the bands to allowed transitions, while the hot band data for band 5 suggests it arises from a parity-rather than an orbi- tally-forbidden transition. Certainly the infrared and Raman spectrum of C~IrBr6 at liquid helium tempera- ture is needed to confirm this choice, but the distinctive hot band pattern-two strong lines to the red of the low-lying hot lattice modes-is suggestive of odd rather than even modes since there are two of the former and only one of the latter in this energy range (Table I). Analysis of the cold (or liquid helium temperature) vibronic lines is more difficult; vibronic patterns may be complicated by excited state force constants and/or geometries different from those in the ground state and by Jahn-Teller coupling in U' states. Surprisingly, vibronic assignments seem least ambiguous for the lines of the U v' (2T2v ) band (band 8). There are no indications of strong Jahn-Teller coupling since no obvious progressions in the t29 or eg vibrations are ob- served and no large splittings of t2g or ev vibronic lines occur. Furthermore, force constants are close to those in the ground state, since the hot and cold lines mirror each other closely. Long progressions in the II1(a1g) totally symmetric mode are superimposed upon an initial pattern of vibronic lines. This is in accord with Ok excited state symmetry. Additional information may be obtained by analysis of our MCD data which is conveniently discussed using Serber's A, B, C notation; the terms arise respectively from Zeeman splittings of ground andlor excited state, field-induced mixing of unperturbed states, and popula- tion differences in a Zeeman-split ground state. The A term dispersion is sigmoid with a zero at the absorption maximum whereas Band C terms peak there.14 Since the Eu"(2T2g) ground state of Ir4+ has a magnetic moment, as do (in general) all excited states in an odd electron system, all three Faraday parameters may have nonzero values. C terms vary as liT while A and B terms are temperature independent; thus at liquid helium temperature the experimentally measureable quantity, (B+C/kT) XkT, is approximately equal to C. A values may be extracted independently of Band C since they have a different line shape. Since AIC::::::! kT I (bandwidth), A terms are generally swamped by C terms for broad-band transitions; thus, previous MCD data for IrBr62- in the tin hexabromide hosts was dominated by C terms.7 Clearly, in CS2ZrBr6:Ir4+ the situation is reversed; bands are now resolved into series of sharp lines with helium temperature widths of the order of 3-15 cm-I, and A terms are prominent in the M CD. Thus our focus is shifted from bands to individual vibronic lines and from C terms to A terms. Experi- mental A/D values (D is the dipole strength) for in- dividual lines may be extracted from the data using Gaussian fits, or the method of moments assuming a rigid shift model for the line.8 If an isolated vibronic line is split by small perturbations, as is quite likely the case for many of the lines in the C~ZrBr6:Ir4+ spectrum, experimental parameters for the unperturbed vibronic state may still be extracted using the method of moments, provided the transitions can still be described by states within a single vibronic manifold. While theoretical A/D values for no-phonon transi- tions depend only on the ground and excited state symmetries and magnetic moments, those for other vibronic lines in systems where the ground state is degenerate reflect in addition the symmetry of both the vibration and the perturbing state(s). We find, using first-order vibronic theory, that for low-lying iridium hexahalide states, A / D is negative for vibronic lines which gain intensity by mixing with Evil states, but may be of either sign when intensity is borrowed pri- marily from U v' states. Mixing with all other states is forbidden by symmetry. These results are summarized in Table III. We do not calculate A/D for the case This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 155.33.120.209 On: Sat, 22 Nov 2014 02:12:48 2674 DICKINSON, PIEPHO, SPENCER, AND SCHATZ TABLE II. Energies of initial vibronie lines and MCD data for Cs.ZrBr6:IrH at liquid helium temperature. A/D (Bohr magnetons) Band Line Energy (em-I) Assignment Experimental Theoretical 5 14 401" Eu'(2T,g) 5 27 + lattice 0.35b 37 51 5 2 "'(88)0 +V6(t2u) , V4(t,u) 0.35 b,d 0.5°,1,& 105 111 "'( 122) 0 5 3 166 >0 5 4 208 >0 5 5 231 +"3 (t,U) 0.35b 0.5°,& 5 6 ",245 +lattice+vI (alo) 5 7 291 +"6(~u) +VI (al g) , 0.25b ,d 0.5°,1,& 304 v4(h,,) +"1 (al g) 312 327 6 0 17 131 em-1 li,," (2T2,,) +no'phonon -0.6h -5/3& 6 38 +alo lattice MAG NET leD I C H R 0 ISM AND S PEe T R A 0 F C S 2 Z r B r 6: I r' + 2675 when perturbing states of both E" and V' symmetries simultaneously contribute significantly to the vibronic intensity since such a calculation requires a determina- tion of the difference in magnitude of vibronic interac- tion matrix elements, a calculation for which an accurate method has yet to be developed. Furthermore, we ne- glect mixing with the ground state in all cases; i.e., we drop the second term in D1, Eq. (2). For transitions to V' states repeated representations occur in the V' xv f direct product when v f is threefold degenerate; the re- duced matrix elements for the repeated representations may, however, be related to one another for a particular V' (2S+lr) mixing state. (The parentage of the V' mixing states need not be specified in the analogous C I D calculation if the above approximations are made since the same sum of reduced electric dipole matrix elements factors out of both C and D and thus cancels.) For band 8 we assume 2SHr is 2T2" and for band 5, 2T1u• In contrast to Zeeman spliUings,15 AID values are iso- tropic for all V' states.16 This has the important prac- tical consequence that we need not worry about the orientation of our crystal. Experimental and theoretical AID values for the better resolved vibronic lines are presented in Table II. Agreement is quite good for the lines of Band 8. The lines of band 6 overlap considerably so only signs are given for lines other than the no-phonon line. The negative sign of AID for the no-phonon line is in accord with our earlier E,," assignment based on C term data, although the magnitude is low (see Table II). The next four lines (labeled 1-4 on Fig. 4) which show approxi- mately equal spacings of about 40 cm-I all clearly correspond to (overlapping) negative A terms. Any attempt to assign these lines to non-alo vibrations in Oh* fails because positive or zero A terms are cal- culated for t20 or eo vibronic lines of the Eo'l~E,.." (2 T2 ,,) transition (Table III). We believe it likely that these lines are members of a progression in a "localized" alo lattice mode which arises with the disappearance of translational symmetry in the vicinity of the impurity ion. This hypothesis accounts for the regular spacing of these four lines and their observed negative A terms. The entire pattern (lines 0-4) is clearly repeated once with a spacing corresponding to the internal alo vibra- tion of the complex (,.....,194 cm-I ), and the ~2 and ~3 members of the 194 cm- I progression can be clearly discerned at higher energies as a sharp line on the red side of each of the first two (much broader) lines of band 7. Band 7, which we nominally assign to the forbidden Eo"~E,/ eTI ,,) transition, is characterized by five broad lines which are clearly members of a progression in the alo vibration. We are convinced that the Eo'l~E,/' transition (assigned nominally to band 6) is the domi- nant intensity source for both bands 6 and 7 since no other allowed transition gives a positive C term, the principal MCD feature of these bands in all hosts (and in solution). An obvious mechanism for intensity borrowing is vibronic coupling through a g vibration. Using Eq. (2), one finds that a t20 vibration will permit the E,,' state to borrow intensity from E,," yielding the observed positive C term (see Table III). Furthermore, the very small energy denominator in this case should make the effect large. The possibility of interchanging the assignments of bands 6 and 7 is excluded as long as the excited states are assumed to maintain Oh* sym- metry since a vibrationally induced transition could not produce the observed no-phonon line of band 6. How- ever, it is also possible that a lowering of the symmetry of the IrBr62- ion in the excited states occurs permitting a mixing of the E,,' and E,," states. Conceivably such a distortion could also playa role in the coupling re- quired for the appearance of the ,.....,40 cm-I lattice progression already discussed for band 6. There is another puzzling observation that can be accounted for by postulating excited state distortion. The ,.....,108 cm-I hot band (line 2h) observed to the red of both bands 6 and 8 almost certainly corresponds to the JJ5(t2g) vibration. For band 6, however, the MCD of line 2h shows a negative A term in contradiction to the theoretically predicted positive (or zero) value (Table III). An appropriate lowering of the symmetry of the excited state could resolve this contradiction. A lowering of the symmetry to allow strong mixing of the E,,' and E,/' electronic states would render meaning- less the assignments of bands 6 and 7 respectively to the individual "allowed" E,/' and "forbidden" E,,' transitions. Both excited states would contain sub- stantial E,," character and thus transitions to either could give rise to a strong no-phonon line and an accompanying totally symmetric progression. Clearly, when two states interact as strongly as those giving rise to bands 6 and 7, perturbation theory is not really applicable. A proper treatment, however, ap- pears to be a difficult undertaking which we do not attempt here. Our hot band data together with arguments pre- sented in Ref. 7 indicate that band 5 arises from a parity-forbidden charge-transfer transition; assign- ments considered were transitions to states arising from the tlo err), eo ( 2676 DICKINSON, PIEPHO, SPENCER, AND SCHATZ TABLE III. Theoretical values of A/D and C/D for some vibronic transitions." A/D (Bohr magnetons) C/D (Bohr magnetons) Symmetry of coupling Symmetry of coupling vibration vibration Transition [from Eo" (2T2o) ground state] Symmetry of Symmetry of P (llu) p(lzu) mixing state P(llu) P (l2u) mixing state 110 (7r) ->1.0 Uo'(2Tla) -1/20 +1/12 Uu'(2TIU ) -1/2 -1/2 Uu' -3/2 -1/6 HI! ~" E,/' 1/2 1/2 Uu' -1/2 -1/2 Uu' forbidden -1 Eu" forbidden E~" 0 0.48 Uu'(2Tlu) -1/2 -1/2 Uu' -4/3 -4/3 Eu" Eo" -1/18 5/6 U1t ' -1/2 -1/2 Uu' forbidden -7/9 Eu" forbidden Eu" 13/36 19/60 Uu'(2Tlu) -1/2 -1/2 Uu' Notationb Pure 7r a=Oja=1/V3 b= 1 b=V2/vJ I." (7r) ->1,(/ Uu'(2T,,,) -13/18 1/2/7/9 -1/-10/9 7/12 -7/6 -7/6 1/2/1/3 forbidden 7/120 -7/6 1t'1/ "U Uu' 1/ " -u j MAG NET I C D I C H R 0 ISM AND S P E C T R A 0 F C S 2 Z r B r 6 : I r 4 + 2677 in the same spectral region as the charge-transfer bands and may be badly overlapped. The situation is further complicated by the possibility of configura- tion interaction mixing the even parity charge-transfer and ligand-field states. IV. CONCLUSIONS The crystal, CS:lZrBrs, shows great promise as a new host for the study of the optical spectra of a variety of metal ions in the +4 oxidation state. For Ir4+, the spectra, if anything, exceed in quality and sharpness the excellent results previously obtained in Cs2ZrCIs,a and the site symmetry of the Ir4+ ion ap- pears to be octahedral or very close to it at low tem- perature. Thus a means has become available to study at high resolution and in an octahedral environment the sharp-line absorption and MCD spectra of ions such as Ir4+, OSH, and ReH as a function of ligand spin-orbit coupling parameter. Through such studies we believe it will be possible to assign these spectra in considerable detail. Our present study of the detailed vibronic structure of Cs2ZrBrs: Ir4+ has confirmed beyond any reason- able doubt the charge-transfer assignments of the prominent bands of Ir4+ previously proposed on the basis of broad band MCD data. Hot band studies of the sharp lines at low temperatures allows us to conclude that the Eo"----+Eu" (2T2u) (band 6) and Eo"----+U,/(2T2u) (band 8) allowed charge-transfer transitions are built on progressions in the g vibra- tions. Furthermore, in each case it is clear that all (or virtually all) of the fine structure is associated with a single allowed electronic transition. No fine structure is revealed for bands 3 and 4 and the MCD shows a negative C term as in all other environments, consistent with our previous assignment of the ex- cited state as U u' (2TIu ) split into two J ahn-Teller components. The hot band data show clearly that the transition corresponding to band 5 is forbidden- a fact deduced previously from the temperature de- pendence of the intensity. The detailed vibrational pattern strongly suggests u activating vibrations and hence a g-g transition. Furthermore, a detailed analy- sis of the vibronic MCD strongly suggests the Eo"----+ Eo'(2TIg ) assignment. Band 7 which we nominally assign to the orbitally forbidden transition Eo"----+ Eu' (2Th ,,) has an anomalously large intensity which we attribute to strong mixing of the overlapping E,/ and E u"(2T2u ) excited states, both of which may be distorted. Finally, we should like to reemphasize the fact that there is no evidence whatever for assigning any part of the intense bands (bands 3, 4, 6, 7, and 8) to anything but charge-transfer transitions. Douglas' sug- gestion3 that lines of an analogous second strong band in Cs2ZrCIs: Ir4+ (at ",23 000-24000 cm-I ) be assigned to d-d transitions seems to us to be without founda- tion. We shall discuss this spectrum in a future pub- lication. Furthermore, we shall present detailed analyses in future publications arguing that the intense transi- tions in OSH and ReH are also due to charge-transfer transitions contrary to previous suggestions that some or all of such transitions are ligand field in nature.1.2 ACKNOWLEDGMENTS We are much indebted to Dr. Peter T. Greene for the preliminary x-ray analysis of the crystal structure of Cs2ZrBrs and we thank P. J. Stephens for several helpful discussions. This work was supported by a grant from the National Science Foundation. * A preliminary account of this work was presented at the American Physical Society Meeting, Cleveland, 1 April 1971: Bull. Am. Phys. Soc. 16, 448 (1971). t Present Address: Department of Inorganic Chemistry, University of Leeds, Leeds LS2 9JT, England. t Present Address: Department of Chemistry, Sweet Briar College, Sweet Briar, Virginia. § Present Address: Department of Chemistry, Southern Illinois University, Edwardsville, Illinois. 1 P. B. Dorain and R. G. Wheeler, J. Chem. Phys. 45, 1172 (1966). 2 P. B. Dorain, H. H. Patterson, and P. C. Jordan, J. Chem. Phys. 49, 3845 (1968); P. C. Jordan, H. H. Patterson, and P. B. Dorain, J. Chem. Phys. 49, 3858 (1968). 3 I. N. Douglas, J. Chem. Phys. 51, 3066 (1969). 4 Symposia of the Faraday Society No.3, 1969, pages 14, 92. 5 S. B. Piepho, Ph.D. dissertation, University of Virginia, August, 1970. 6 P. A. Dobosh, J. R. Dickinson, J. A. Spencer, and P. N. Schatz, Bull. Am. Phys. Soc. 16,448 (1971). 7 S. B. Piepho, T. E. Lester, A. J. McCaffery, J. R. Dickinson, and P. N. Schatz, Mol. Phys. 19, 781 (1970). 8 S. B. Piepho, J. R. Dickinson, and P. N. Schatz, Phys. Status Solidi 47,225 (1971). 9 A. J. McCaffery, P. N. Schatz, and T. E. Lester, J. Chem. Phys. SO, 379 (1969). 10 C. K. J prgensen, Acta Chem. Scand. 10, 500 (1956). 11 G. N. Henning, A. J. McCaffery, P. N. Schatz, and P. J. Stephens, J. Chem. Phys. 48, 5656 (1968). 12 See, for example, G. Herzberg, Molecular Spectra and Molecular Structure III. Ekctronic Spectra and Electronic Struc- ture of Polyatomic Molecules (Van Nostrand, New York, 1966), Chap. II. 13 D. Johnston, R. A. Satten, and E. Y. Wong, Optical Properties of Ions in Crystals, edited by H. M. Crosswhite and H. W. Moos (Interscience, New York, 1967), p. 429. 14 See, for example, P. N. Schatz and A. J. McCaffery, Quart. Rev. Chem. Soc. 23, 552 (1969); Erratum 24,324 (1970). IS See, for example, R. A. Satten, Phys. Rev. A 3, 1246 (1971). 16 S. B. Piepho (unpublished calculation). This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 155.33.120.209 On: Sat, 22 Nov 2014 02:12:48