CHAPTER 1
Electronic Spectra
BY P. DAY AND E. R. KRAUSZ
1 Introduction
The basic format of this Report is little changed from that of Volume 3. As in all the previous volumes in this series we have selected those papers which we consider to contain particularly significant new observations for more detailed coverage under individual subject headings; papers which report electronic spectral data largely in the context of preparative and characterization work on new compounds are dealt with according to the central metal atom in the molecule. We also include a Table of references to other spectra which appear either as peripheral to some other aspect of the characterization of a molecule or complex or in journals which we have not had an opportunity to examine.
One or two other general comments are in order. The Report is somewhat briefer than last year's, partly as a result of a decrease in the number of papers coming to our notice, but partly also because we have decided not to present some of our material in as much detail as in the past, so as to keep the volume's cost down to a level which individual users can still afford. It is to be hoped that this terser style of presentation does not reduce the readability of the text too much.
Two reviews contain material of interest to inorganic spectroscopists. Electronic Raman spectroscopy is now being used on a widening variety of inorganic materials, and recent work on transition-metal and rare-earth compounds has been assembled by Koningstein.' High-valent first-row transition-metal oxides and halides are among the model compounds most favoured by spectroscopists, so a comprehensive review of their preparations and properties which has appeared is welcome.
2 Polarized and Low-temperature Crystal Spectra
In Volume 3 we noted the increasing proportion of work on the electronic spectra of complex ions which is taking advantage of the extra dimension of information brought in by making polarized measurements on single crystals. This year we have noticed a continuation of the same trend with, in addition, emphasis on higher resolution yielding site-group splittings and vibronic information. On the other hand, since only relatively simple molecules yield spectra that sharpen sufficiently at low temperatures to justify increasing resolution beyond, say, 1 A, preoccupation with detail of this kind may serve to focus attention on an artificially limited range of examples within the huge available field of inorganic materials.
We shall use again the sub-divisions for this section introduced in Volume 3, including separate accounts of spectroscopy whose main emphasis is on vibronic interactions and spectra in the far and extreme ultraviolet.
No reviews devoted to polarized or low-temperature crystal spectra were noted this year.
Discrete Complexes in Crystals. — Monoatomic Ligands. In contrast to the large number of papers reported in Volume 3 on tetroxo-complexes of the 3d elements, in 1973 no such papers fall within this section.
In the field of halide complexes, a number of papers have reported fine detail in the spectra of tetrahalogeno-complexes of 3d elements, as well as hexahalides of all three transition series.
The tetragonal crystals Cs3MX5 (M = 3d ion; X = Cl or Br) are popular with electronic spectroscopists as high-symmetry hosts for [MX4]2- ions, the spin-forbidden ligand-field transitions in the cobalt(II) examples providing a particularly rich field for study. In a very elaborate series of experiments, Tsujikawa and his co-workers have measured Zeeman splittings of many of the band origins in Cs3CoCl5 and Cs3CoBr, as a function of the angle between the applied magnetic field and the crystal axes. Deriving the selection rules, they show that the anisotropy of the Zeeman splitting allows one to determine whether the excited states are Γ6 or Γ7 in D*2d, and whether g[perpendicular to] is positive or negative.
Tetrahedral nickel(II) complexes have been of interest for a number of years because of the possibility that their ground and excited states might exhibit Jahn-Teller distortions. A further contribution to this field is a study of the polarized spectra of [NiX4]2- (X = Cl or Br) in the tetragonal crystals [N(Et)4]2NiX4,4 with a special emphasis on the components of 1G. The resolved fine structure on one hand was assigned as a progression in totally symmetric stretching modes, and origins were assigned to representations of D*2d double group as follows (tetrahedral parent states in brackets):
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]
Part of the spectrum is shown in Figure 1.
It has been known for some time that the 3d3 ion [MnF6]2- has an elaborate sharp-line spectrum, parts of which have now been definitely assigned, using Cs2GeF6 and K2GeF6, in which the site symmetries are respectively Oh and D3d, as host lattices. Transitions gain electric dipole intensity by vibronic coupling with all three of the odd-parity intramolecular modes of the octahedral complex, and many also appear in combination with the totally symmetric modes (Figure 2). The 4A2g ->2Eg and 2T1g transitions can be assigned quite unambiguously, but for 4T2g, it is necessary to assume that the spin-orbit coupling factor is reduced about tenfold, presumably by a Ham effect. Unfortunately the higher lying 4T1g is only poorly resolved. In contrast to [MnF6]2-, the spectrum of [MnCl6]2- contains only broad transitions, and in addition there are chemical difficulties because of disproportioaation. After sorting out the MnIV band from others due to MnII and MnIII, [MnCl6]2- is assigned the parameters B, 584 cm-1 and Δ, 18 240 cm-1 in a K2SnCl6 host.
Last year we remarked on the extensive activity in the spectroscopy of 4d and 5d hexahalide complexes, referring tg work from the groups of Schatz and Patterson. The latter now contributes a study of [OsBr6]2- in the cubic host Cs2ZrBr6 at 20 K, including both absorption and luminescence. Again, vibronic absorption bands result from coupling to the odd-parity modes of the octahedral complex, the much weaker zerophonon lines being magnetic dipole in character. Jahn-Teller effects may be important in this spectrum, the major origins of which are assigned as follows:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]
The salts [EtNH3]2SnX6 (X = Cl or Br) are hexagonal, and provide trigonally distorted sites for [MX6]2- ions. Values of Δ and the trigonal distortion parameter have been calculated from the 4A2g ->2T2g transitions of [TcX6]2- and [ReX6]2- in this host at 77 and 298 K. Optical electronegativities derived from the charge-transfer spectra are 2.3 and 2.1 for TcIV and ReIV respectively.
The lowest-energy absorption band of the dimeric anion [Re2Cl8]2- was suspected some time ago as being the δ -> δ* transition within the quadruple Re — Re bond. This idea receives support from the fact that at low temperature it resolves into a prominent progression in the totally symmetric Re — Re stretch at 248 cm-1, each member of which is combined with one quantum of the alg ReReCl band (115 cm-l). The integrated intensity of the transition does not decrease down to 4 K, so it is definitely allowed.
Polyatomic Ligands. Work on the polarized and low-temperature crystal spectra of complexes of polyatomic ligands is arranged in order of increasing atomic number of the central metal atom.
In bis(methoxyacetato)diaquocobalt(II) the two organic ligands occupy a trans-planar arrangement, the two water molecules completing a distorted octahedron for which the approximate ligand-field symmetry is C2h. Magnetic dipole and vibronic electric dipole contributions to the intensity of the ligand-field transistions werc separated by measuring the temperature dependence of the band envelopes.
Over the past 10 years there has been a great deal of controversy about the assignment of the visible and near-u.v. spectrum of the planar [Ni(CN4)2- ion. In most crystals the spectrum of this ion is complicated by intermolecular interaction effects, because the planar units frequently occur in closely spaced stacks. However, this complication may be avoided by examining salts of bulky organic cations, such as [Bu4N]2Ni(CN)4. Two ligand-field bands with opposite polarizations appearing in the 5 K absorption spectrum of this crystal, at 31 000 and 31 650 cm-1, were assigned as 1A1g ->1B2 and 1E respectively in a highly distorted D2d excited state. Intense sharp bands appearing in the 34 000-37 000 cm-1 region are the spin-allowed and spin-forbidden metal-to-ligand charge-transfer transitions, of the type d ->a2u (4pz, and CN π*). To be precise, the two charge-transfer transitions are both to A2u (D*4h) states, one (34 360 cm-1) coming from 3Eu and the other (35 840 cm-1) from 1A2u Russell-Saunders parent states. The crystal spectrum of tris-(1,lO-phenanthroline)nickel(Ir) nitrate at 77 K has been analysed to yield a set of ligand-field parameters, but the authors' assumption of a tetrahedral ligand field in this salt seems rather questionable.
Interest in the polarized spectra of CuII complexes of chelating ligands hinges on attempts to assign effective point symmetries in distorted environments and to derive one-electron orbital orders without the complications of electron repulsion. Dichlorobis( tripheny1phosphine oxide)copper(II), a CuO2Cl2 chromophore, has a compressed tetrahedral geometry. In the ground state the single hole is in xy, and the polarized ligand-field spectrum suggests the following assignments (in the C2 point group) :
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]
The orbital ordering, and even separation, is surprisingly similar to that in [CuCl4]2-. Hathaway's group continue their long series of papers on CuII complexes with a study of three crystals containing co-ordinated nitrate groups, two forms of bis(nitrato)bis(α-picoline)copper(II) and bis(nitrato)mono(pyrazine)copper(II). The spectra were assigned in C2v and D2h point symmetries, the differences between them being due partly to the differences in the molecular structures and partly to the variable positions of the nitrate oxygen atoms, which lie off the z-axis. The dodecahedra1 crystal field in calcium copper(II) acetate hexahydrate has attracted attention in the past, but in a new study of its single-crystal spectrum bands are assigned as follows: 16 500 cm-1 to xz, yz ->xy and 16 950 cm-1 to z2 ->xy. The single-crystal electronic spectrum of cis-bis(hexafluoroacetylacetonato)bis(pyridine)ZnII, doped with CuII, supports the conclusion from the e.p.r. spectrum that there are three types of copper site in the crystal, each with approximately axial symmetry about one of the three orthogonal axes of an octahedron.
A further contribution to the long-standing discussion about the d-levels in square-planar PtII complexes is the polarized spectrum of the orthorhombic crystal KPtCl3(NH3), measured at 10 K. Starting from the ligand-field parameters which fit the [PtCl4]2- spectrum, the effect of substituting one chlorine by NH3 is brought in through the angular-overlap method. Unfortunately it is not possible to observe a high-energy band corresponding to the a1g ->b1g transition in K2PtCl4 at 37 000 cm-1. Thus the relative ordering of alg(z2) and eg(xz, yz) remains obscure.
A most unusual example of an inorganic molecular crystal is that of uranium borohydride, U(BH4)4, for which the isomorphous Hf(BH4)4 acts as a convenient host. The electronic spectrum of this molecule has been measured over the entire energy range from the charge-transfer or f-d cut-off at 2900 A down to the limit placed by the vibrations of the host lattice at 20 000A (Figure 3). From the temperature variation (300, 77, and 4 K), deuteriation, and intensity patterns, over twenty possible forced electric dipole origins, due to transitions within the 5f2 manifold, were identified. Most of the vibronic sidebands were also assigned to specific internal modes of the M(BH4)4 molecules. A valuable aid in assigning the spectra was a Zeeman effect study up to 95 kOe. When least-squares fitted to fifteen of the band systems, the following parameters gave a mean-square error of 160 cm-1 : F2 187.8, F4 35.9, F6 3.94, ζ5f 1908.0, and A -371.9, B -261.6cm-1. Within each band system the spectra fit a crystal field whose two parameters A60 (r6) and A40 (r4) have signs compatible with the simple expectation for twelve negative point charges (the H atoms) placed around the metal.
Continuous Lattices. — Doped Crystals. Rather less work falls in this section than in previous years, probably because most of the simpler high-symmetry host lattices have now been quite thoroughly examined.
The sharp transitions and long lifetimes of the excited states of CrIII continue to make them a favourite subject for spectroscopists. Transitions from the long-lived 2E up to higher doublets have been examined by a Russian group, using yttrium aluminium garnet (YAG) as host. The no-phonon line of the spin-allowed 4A2g ->4T2g transition of CrIII doped in TiO2 is sufficiently sharp for Zeeman spectroscopy, the results of which were analysed in terms of a D2h ligand field, including spin-orbit coupling. Other work on the spectra of chromium impurities has stressed the question of charge conversion between CrIII and CrII. In CdF2, CrIII is converted quantitatively into CrII when the crystal is heated in Cd metal vapour, and back to CrIII again on heating in the vapour of a halogen. In CdSe, chromium substitutes both as CrII and CrIII, and the spectroscopic consequences of interactions between the two oxidation states are seen in highly doped samples.
Stress-induced splitting of sharp absorption lines is sometimes a useful alternative to Zeeman spectroscopy, but what would appear at first sight to be a favourable case in which to attempt such an experiment, namely MnII in ZnSe, turns out disappointingly negative: stress applied to the crystal parallel to 110 caused no observable splitting of the 4E state.
An unusual example of excited-state spectroscopy (of which, of course, CrIII is the favourite subject) is provided by a paper on Cu1 in KBr. The low-temperature absorption spectrum of an optically pumped crystal contains bands due to transitions from the 3d94s excited state of Cu1 to the continuum states of the conduction band.
Pure Crystals. Apart from papers primarily concerned with exchange interactions, which are dealt with in Section 3, only a single paper falls in this section this year.
Low-temperature measurements in the i.r. up to ca. l000 cm-1 display the complete spin-orbit and low-symmetry splittings of the ground 4T1g state for the trigonally distorted octahedral CoII ions in CoCl2 (Figure 4). On the other hand the exchange splittings remain unresolved. Zeeman spectroscopy in fields up to 43 kOe again proves a valuable assignment tool for the zero-phonon magnetic dipole lines, whose calculated positions, polarizations, and intensities agree quite well with observation.
Vibronic Effects. — In Volume 3 we commented on the very precise information about vibrational-electronic interactions now emerging from high-resolution spectroscopy of d- and f-block ions. 1973 saw more examples of this development, but, in one case at least, some care is needed in interpreting the very high-resolution data.
The fine structure accompanying the 4A2 ->4T2 and 4A2 ->4T1 transitions of CoII in both cubic and hexagonal ZnS in the near i.r. cannot be explained by invoking a static crystal-field model (Figure 5). Inclusion of Jahn-Teller coupling which is weak compared with the static low-symmetry field, however, is sufficient to give a good account of the spectrum.
Previous Reports in this series have mentioned the highly resolved spectra given by octahedral 4d and 5d hexahalide ions when doped into cubic hosts such as Cs2ZrCl6. The wealth of spin-orbit and vibronic detail has made them excellent subjects for theoretical 'analysis, although with such elaborate data it is necessary to ensure that all the factors which might lead to band splittings are taken into account. For example, it has been proposed that the splittings of individual members of the progression in the totally symmetric Ir-Cl stretching mode which accompanies one of the U' states of [IrCl6]2- doped in Cs2ZrCl6 is due to higher-order coupling between U' and either a t1 or a t2 vibrational mode. However, one can rationalize the band envelopes equally well as vibrational isotope splittings, from the natural abundances of Cl and Cl in the sample.
As well as resolving the vibronic fine structure itself, information about coupling of electronic states to the lattice phonon spectrum can also be had from the temperature broadening and shifts of the zero-phonon lines themselves. An example is the spectrum of the uranyl ion in the two crystals [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]. At 77 K the shapes of the zero-phonon lines in these two compounds were respectively asymmetric Lorentzian and Gaussian, but at 120-130 K they were both Lorentzian. The effective Debye temperatures of the two lattices were estimated as 100 [plus or minus] 50 and 500 [plus or minus] 50 K. The lines broaden through interaction with the intramolecular modes of the UO2 groups but they shift through interaction with the lattice phonon spectrum.
