Electronic Spectra

BY P. DAY

1 Introduction

The overall outline of this Report follows quite closely from last year's. Papers which are felt to represent particularly significant developments in inorganic spectroscopy, or which exemplify fields at present in a particularly lively state of development, are dealt with in the earlier sections, while papers whose main emphasis is on the preparation and characterization of compounds, and in which electronic spectroscopy is used as a structural tool, are referred to in the sections classified according to the metal atom in the molecule. Finally, papers in which electronic spectroscopy plays only an incidental part or which, for one reason or another, we have not been able to scan in detail, are collected in the Table of Section 14, again classified by metal atom.

A few words about the principles defining our range of coverage of inorganic materials, and classes of spectra, are relevant here. In general, no transitions of entirely delocalized (i.e. band-to-band) character are included, although one could certainly make a case that compounds such as GaP are as worthy of attention by inorganic chemical spectroscopists as KMnO4. Nevertheless, bowing to the traditional subject demarcation between inorganic chemistry and solid-state physics, almost all the materials we refer to are either discrete molecules or molecular ions or, if continuous lattice solids, are predominantly of ionic character. Also, we have refrained from including any references to emission spectra, although emission data are of course quite complementary to absorption and in many cases serve to confirm the absorption assignments. Our reason is quite simply the volume of luminescence work, to treat which properly would necessitate a separate chapter.

There have not been any books entirely devoted to inorganic electronic spectroscopy during 1972, but an important introductory reference is an account of the electronic spectra of co-ordination compounds, both ligand field and charge transfer, which appears in the American Chemical Society Monograph on 'Coordination Chemistry'. A review on stereochemical and electronic structural aspects of five-co-ordination also includes material on the electronic spectra of this important class of compound. Two new volumes have appeared in a useful compilation of literature references to electronic spectra, and reviews of nuclear hyperfine structure in the spectra of diatomics and of molecular geometries of excited states' are included in a new book on aspects of molecular spectroscopy.

2 Polarized and Low-temperature Crystal Spectra

A steadily increasing proportion of all the work on electronic spectra of inorganic molecules and complex ions is being carried out on crystals rather than solutions, thus providing the extra dimension of information concerning polarization of the transition, and removing ambiguities often found in older assignments, even of quite simple chromophores. Because of the great variety of chemical systems, and types of spectra, now being studied in the crystalline state, this section of the Report is for the first time subdivided. For this purpose we have not chosen the conventional classification of inorganic spectra into "ligand field', 'charge transfer', etc., partly because many papers report results on both kinds of spectra for the same compound and partly because we feel that the inorganic interest of the work under review is best served by concentrating on particular problems or classes of compound.

Among general trends which may be discerned in the year's activity are the continuation of definitive high-resolution work on simple octahedral and tetrahedral chromophores, mainly with oxide and halide ligands, and a marked increase in the amount of work on more complicated complex ions containing multidentate chelates, often producing unusual stereochemistries. The former activity has now produced sets of unambiguous assignments of both ligand-field and charge-transfer energies against which theoretical calculations can be subjected to stringent verification.

Much detailed work on Jahn-Teller and other vibronic interaction effects continues to appear, and warrants treatment as a separate section. We also noted in last year's Report the important resurgence of interest in the old established technique of soft-X-ray absorption and emission spectroscopy, as well as in the previously inaccessible extreme ultraviolet region, brought about by the use of synchrotron radiation from particle accelerators as a light source. Although this field still lies somewhat nearer the realm of solid-state physics than of inorganic chemistry, we again include a section devoted to new work in it because it is clear that the range of compounds being examined is widening rapidly and the results will soon warrant careful attention by inorganic chemists. Worth noting too is the way in which the new technique complements X-ray and ultraviolet photoelectron spectroscopy. This section of last year's Report also included an account of spectra measured at cryogenic temperatures on evaporated films of compounds which exist as gases at room temperature, as well as of molecules isolated in rare-gas matrices. No relevant spectra of evaporated films were noted this year, and references to matrix-isolation spectra are to be found in Section 8.

The only review entirely devoted to crystal spectra this year is by Martin, on square-planar PtII complexes. It emphasizes the work, particularly on exciton effects, carried out recently in his own laboratory.

Discrete Complexes in Crystals. — In this section we review work on both pure and doped crystals in which discrete molecules or complex ions may be distinguished. In contrast to organic molecular crystals, almost all the spectroscopic work on inorganic molecular crystals is interpreted implicitly using the 'oriented gas' model. There is no reason in principle why interactions between transition dipoles, on neighbouring molecules, giving, for example, Davydov splittings, should not be found in the inorganic crystal spectra, but scarcely anyone has ever seriously probed the question. The very occasional instances where such effects have been postulated are dealt with in Section 3.

Monatomic Ligands. Oxides. New data continue to appear on the tetra-oxo-anions which have been used so much in the past as models for testing assignments of charge-transfer spectra. Ballhausen gives some arguments based on a simplified treatment of electron repulsion for believing, as has in fact been subsequently confirmed experimentally, that the weak near infrared band of MnO4 at 14 450 cm-1 is the forbidden 1T1 state coming from the t512e1 configuration. The low-temperature spectrum of permanganate has of course been measured many times, but an interesting new host lattice for this ion is KBr, in which it substitutes for Br-. The site symmetry is therefore cubic, and the familiar rich vibronic structure of the first charge-transfer transition contains bands due not only to coupling to the molecular modes of the complex, but to lattice modes of KBr. Perhaps this type of substitution could be used more widely as a probe to lattice phonon spectra.

The congenors of permanganate, TcO4- and ReO4-, do not form ideal solid solutions in KCIO4, probably because they are too big to fit properly. Consequently the charge-transfer bands remain relatively broad even at 4 K, though in each case the major progression in the totally symmetric internal mode is clearly apparent. Compared with the ground states the frequencies of these modes are reduced in the first two charge-transfer states by the following amounts:

I II

TcO4- 12.0% 14.0%
ReO4- 12.7% 15.8%

The band origins lie much further into the ultraviolet than those of MnO4-(TcO4- : 32 600, 38 600; ReO4-: 40 066, 45 934 cm-1), though the assignments are analogous, but no forbidden 1T1 transitions could be detected.

Because it contains a single d-electron in the ground state, the charge-transfer spectrum of MnO42- should be the next in order of increasing difficulty to interpret after the closed-shell a0 ions. There have been two reports about this ion in the past year, one using K2SO4 as host lattice (Figures 1 and 2) and the other BaSO4. Both agree that the weak band with extensive vibrational structure in the near infrared is the 2E ->2T2 ligand-field transition, thus defining Δ for the ion. For example, in K2SO4 the average frequency of the three components split by the Cs site perturbation is 11 177 cm-1, with a total site splitting of 667 cm-1. (In BaSO4 the site group is C2u.) The two studies also agree that the tetrahedral parentage of the first charge-transfer band, near 18 000 cm-1, is 2T2. If, by analogy with MnO4-, this state arises from transferring an electron from the non-bonding t1 to the partly occupied 2e orbital, the fact that it lies below 2T1, which also comes from the same configuration, can be rationalized by examining the effect of electron repulsion using a simplified model which only takes into account one-centre metal (i.e. 'ligand field') contributions. In this way the sequence of the first four charge transfer transitions (2T2< 2T2< 2T1< 2T2 from site group splittings) are accounted for.

The ions MO43- are rather more difficult to stabilize in host lattices than the corresponding mononegative and dinegative anions, but phosphate and vanadate lattices related to chlorapatite are convenient. Hypomanganite, having a d2 configuration, has a spectrum even more elaborate than that of manganate, since several ligand-field transitions are now expected, as well as the charge-transfer and, according to Holt's assignment based on the 4 K polarized spectrum, the two types of transition are interspersed among each other. As with other tetroxo-ions, a lot of vibronic structure is found at 4 K (see, e.g. Figure 3). Based on site splittings and arguments based on analogy with related ions, Holt proposes assignments of the main absorption regions in CrO43- and MnO43- as follows:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

Halides. We deal first with tetrahedral halogeno-anions of the first transition series, for which the reports this year concern exclusively ligand-field spectra, followed by octahedral 4d and 5d hexahalide complexes, whose ligand-field and charge-transfer spectra have received attention.

The Copenhagen group have made a careful examination of the MnX4-2(X = Cl or Br) ions, as the orthorhombic tetramethylammonium salts. At 4 K many of the ligand-field transitions are accompanied by long progressions in vl, whose frequency varies somewhat from state to state. Gross assignments of the band groups are accomplished using a strong field scheme with Trees correction (Table 1). Although the metal site in these salts is formally Cs, it approximates to a tetragonal distortion of the tetrahedron which, nevertheless, is much smaller than the spin-orbit splitting of most of the states. Perhaps the most interesting feature, however, is that the splitting of the 4A1, 4E origin (Figure 4) cannot be accounted for either by spin-orbit coupling alone, or by any combination of spin-orbit coupling with the static tetragonal perturbation. Some combination of spin-orbit with vibronic interaction remains a possibility though.

In contrast to the tetramethylammonium salts, in the tetraethylammonium salts of the MX42- ions the site group is strictly tetragonal. Smith and his colleagues have looked for the lowest ligand-field excited states of NiX42-(X = Cl or Br), both in this lattice and in the caesium salts, which are again orthorhombic. The ground state 3T1(F) is split in first order by spin-rbit coupling so the ligand-field transitions occurring in the infrared are from the A1 spin or component of 3T1(F) to the other components, or from 3T1(F) to the other crystal-field components of 3F, i.e. 3T2 and 3A2. In the Cs salts Smith claims to have identified one of the former, while in the tetraethylammonium salts, bands centred at 7000 cm-1 (chloride) and 7400 cm-1 (bromide) are of the latter type. The higher-energy infrared transitions are accompanied by progressions in v1 with average frequencies of 268 (Cl) and 167 cm-1 (Br), which may be compared with 271 and 168 cm-1 for the ground state, from the Raman spectrum. Subsidiary excitation of v2 is also observed.

Caesium hexahalogenozirconates have proved most versatile host lattices for examining the spectra of 4d and 5d hexahalide complexes, since their relatively simple phonon spectra mean that both ligand-field and charge-transfer bands become highly resolved at low temperatures. Schatz's group, in particular, have made good use of these lattices, which have the further advantage of being cubic, in their m.c.d. studies of charge transfer spectra, referred to in Chapter 2. They are equally suitable, of course, for examining ligand-field spectra, and a number of examples have appeared in the past year. At 4 K, in Cs2ZrCl6, the d1 ion MoCl6- has a single transition, assigned at [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII], accompanied by much sharp vibronic detail between 23 830 and 25 697 cm-1. There appears to be no evidence for a static Jahn-Teller effect in this cubic host: the major progressions, of seven members, are in the totally symmetric mode (292 cm-1), built as would be expected on false origins, owing to excitation of odd parity modes, v7 (t1u, lattice mode, 49cm-1) and v4 (t1u, Mo — Cl stretching mode, 170 cm-1). The quartet-doublet transitions within the t32g configurations are of course renowned for their sharpness, and the transitions from Γ8(4A2g) to the Γ7 and Γ8 components of 2T2g in ReBr62- are no exception. The origins, which lie at 13 144 and 14 917 cm-1 are accompanied by vibrational structure, including not only the usual progressions in v1 but also short ones in eg and t2g modes, revealing a degree of vibronic interaction (Figure 5). There are also differences between the frequencies of the odd-parity enabling modes in the ground and excited states, as follows (in cm-1):

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

Octahedral hexahalogeno-anions are not the only kind to substitute into the Cs2ZrX6 lattices, as some experiments by Patterson on PtC142- demonstrate. This anion has of course already been thoroughly examined in K2PtCl4 and other hosts, but in the zirconate the [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] transition, which lies between 24 800 and 28 000 cm-1, is particularly well resolved (Figure 6). The major superimposed vibrational mode is again the totally symmetric one: its frequency is 293 cm-1 compared with 329 cm-1 in the ground state. A crystalfield model taking the orbital sequence as x2 - y2 >xy >xz,yz >z2 fits the entire set of bands to a standard deviation of about 100cm-1. The resulting orbital energy differences electron repulsion parameters and spin-orbit coupling constant are: [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]; and ζ, 1013 cm-1.

As noted above, most of Schatz's recent work on the high-resolution chargetransfer spectra of 4d and 5d hexahalides in Cs2ZrX6 is referred to in detail in Chapter 2, since it is principally concerned with m.c.d. measurements. However, the papers also report full absorption data and one, in particular, deserves mention here since it clears up an old controversy about the nature of the absorption bands in OsCl62-. Dorain originally assigned the entire spectrum as ligand field in origin, but it is now clear, both from the m.c.d. and vibronic fine structure, and from a comparison of the gross spectral features with those of related molecules, that all the absorption features between 23000 and 34 000 cm-1 are due to charge transfer, although weaker features below 18 OOO cm-1 are certainly d-d transitions. The m.c.d., in particular, leads to a set of assignments which order the ligand-to-metal charge-transfer excitations as follows (in order of increasing energy): [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII].