Electronic Spectra

BY P. DAY

1 Introduction

The most obvious difference between this Report and the one which appeared in Volume 4 of this series is its slightly greater length. Partly this is due to the fact that the period under review this time is eighteen months instead of one year, but it also reflects a genuine increase in the volume of work containing at least reference to electronic spectroscopy as a tool for characterizing new compounds, and perhaps also an increased level of activity in those laboratories which use the more elaborate refinements of low temperatures, high resolution, or unusual sample conditions such as high pressure or high magnetic fields to probe, often with great subtlety, the bonding characteristics of inorganic molecules. Such activity, and the understanding which it gives, is quite as central to the progress of inorganic chemistry as synthesizing new compounds or investigating reactivity. Indeed, the three march together.

In format this Report is like last year's. In style too it aims at conciseness, in an effort to keep the total bulk of the volume down. What we consider the most significant advances appear in the subject sections; spectra measured as part of a broader study of a group of compounds are dealt with according to the central metal atom of the complex, while papers making only passing reference to spectra, or in journals to which we have not had access, appear in the final Table.

2 Polarized and Low-temperature Crystal Spectra

Comparing the contents of this section with that in Volume 4 it is at once apparent that polarized single-crystal spectroscopy is now established as a relatively routine technique in many laboratories, and that it has become an integral part of the portfolio of physical techniques to be tried out on any newly prepared or specially interesting substance. Even the use of liquid helium, which a few years ago would have been confined to physics departments and the most avant garde physical chemistry laboratories, is now taken completely for granted in inorganic chemistry as a routine method of simplifying and improving the resolution of a wide range of spectra. In large measure this is the result of a new generation of cryogenic devices, such as continuous-flow cryostats and closed-cycle refrigerators, which are simple to set up and more or less trouble-free in operation. In part too, though, it reflects a greater readiness by inorganic chemists to get involved with more complex instrumentation in order to obtain more subtle information about the bonding in the compounds they make.

Reviews on aspects of crystal spectroscopy published in the period surveyed here include a most elegant synthesis of the intra- and inter-subshell transitions of metal impurities in ionic crystals by McClure, a useful survey of the vibronic spectra of co-ordination compounds, showing what a wealth of information is contained in such spectra, even of large molecules, and an account of work (much of it from the author's own group) on polyatomic impurities as guests in alkali halide crystal.

Discrete Complexes in Crystals. — Monoatomic Ligands. Oxide. After a gap last year the oxide ion once again figures as one of the simplest ligands in these pages, with work at low temperatures both on tetraoxo-ions and, perhaps for the first time, on substituted 0x0-species. Far less studied than the inorganic spectroscopists' favourite molecule the permanganate ion is the next member of the series, manganate. Polarized spectra of this ion doped in K2S04, Rb2S04, and Cs2S04 have now been reported over the range 10 000 — 40 000 cm-1. In tetrahedral symmetry the ground state of [MnO4]2- is 2E, and both ligand-field and charge-transfer states are well resolved (Figure 1). The Cs local symmetry of the host lattice splits the 2T2 ligand-field band into three zero-phonon components, whose polarization behaviour is quite different from that observed in the lowest-energy charge-transfer band having the same cubic symmetry. This is because spin-orbit coupling makes a contribution to the splitting of the former which is comparable to the low-symmetry field, while the orbital degeneracy in the charge-transfer state is derived from a hole localized on the oxygen, which is therefore subject to a much smaller spin-orbit interaction. This appears to be the first example of such a phenomenon. Also much less studied than permanganate, though for different reasons, is the isoelectronic but radioactive ion pertechnetate. Its spectrum in CsCIO4 between 20 000 and 47 000 cm-1 contains two band systems, each with partially resolved vibronic fine structure, though no discrete zero-phonon lines can be seen.

The substituted d0 chromate ions [CrO3X]- (X = For Cl) are attractive objects for crystal spectroscopy, since they provide a substantial trigonal perturbation on the parent tetraoxo-species, and reports on their charge-transfer spectra, in addition to that of the 5d0 ion [OsO3N]-, have come from two groups. The Copenhagen group explains the sharp line structure of the lowest-frequency band system on the assumption that 3E and 3A2 states originating from the tetrahedral parent 3T lie a few hundred cm-1 above the zero-phonon line of the lowest 1E state, originating from 1T2. The Oxford group examined [CrO3X]- and [OsO3N]- in KClO4 at 4 K and found, rather surprisingly, that the dipolar guest ions were oriented by the dipole of the Cs site in the perchlorate lattice. Assignments of the various band systems to 1E or 1A1 components of the tetrahedral 1T2 then followed from the observed dichroic ratios. The lowest-energy charge-transfer transitions result from donation of an electron from a2 and e (C3v) orbitals localized on the oxygen atoms, which correlate with the t1 shell in the parent tetraoxo-ions. Similar conclusions about nitrido-osmate were reached by the Copenhagen group, who used LiClO4,3H2O as a host. Although the highest filled levels are oxygen-localized, the π-bonding in the ion is dominated by the nitrogen. In the polarized single-crystal spectra of salts such as Ph4As[MoOCl4(H2O)], containing the MoO3+ moiety, the lowest-energy transitions are O(2pπ) -> Mo(xy) charge-transfer and Mo(xy) + Mo(x2- y2) ligand-field types.

Halide. Some of the richest ligand-field spectra to be found anywhere are those of the tetrahedral ions [MX4]2- (M = 3d ion, X = halide). The wealth of fine structure revealed at 4K is interesting, not only for the very detailed information about vibronic interactions which can be extracted from it but, more generally, for unambiguous assignments of the electronic states themselves, which then provide the starting point for searching tests of theoretical models for describing ligand-field states. An extremely detailed study of the Co" ions, in the lattices Cs3CoX5(X = Cl or Br), has been used in this way, the site-group and spin-orbit splittings of the vibronic origins giving the independently assigned input for a least-squares fit to a new molecular orbital-based model of ligand-field spectra. From a theoretical point of view, general expressions for the electron-repulsion matrix elements were derived using a molecular orbital basis and values of the LCAO coefficients in the ligand-field manifold follow from fitting the experimental spectra. One of the many sharp band-origin lines in the Cs3CoX5 spectrum, that of 2E(D), has been the subject of an elegant series of uniaxial stress experiments, from which it has been found that the tetragonal splitting of the 4A2 ground state should disappear at 7.57 and 8.78 kbar for the chloride and bromide respectively. Jahn-Teller interaction makes only a small contribution (ca. 3%) to the tetragonal distortion of the ground state at ambient pressure, but is important in the excited states. Zeeman measurements on the same excited states yield spin-Hamiltonian parameters in agreement with those found from e.p.r., which in turn agree with crystal-field calculations.

Detail comparable to that found in the Cs3CoX5 crystals also appears in the ligand-field bands of [NiX4]2-, both in the tetraethylammonium salt and doped into Cs3ZnCl5, (Figure 2). With the exception of 1A1(S), band systems corresponding to all the tetrahedral d8 terms can be identified, as follows:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

Because of a combination of spin-orbit coupling and static tetragonal distortion the components of 3T1(p) near 16 000 cm-1 span an unusually wide energy range.

So sharp are the ligand-field bands in some chloride complexes that it is even possible to see fine structure due to the different vibrational frequencies of M — Cl and M—Cl when, as would normally be the case, the two chlorine isotopes are present in their natural abundances. An example is shown in Figure 3.

Like its tetrahedral analogue [CoX4]2- the d3 ion [ReCl6]2- has highly resolved intrasubshell transitions 4A2g (Γ8) ->2T2g(Γ7) whose Zeeman splitting can be measured directly. The right and left circularly polarized spectra of the zero-phonon line agree nicely with expectations but the higher vibronic components cannot be rationalized unless one assumes that the [ReCl6]2- impurity ion couples to the vibrations of the entire K2PtC16 host lattice, and not simply to the zone-centre phonons. When the spherical Group I cation in this type of lattice is replaced by a 'cylindrical' group such as CH3NH+3 dichroism is observed and thus, for example, transitions appearing in the [IrCl6]2- spectrum at 19 760 and 20 000 cm-1 can be assigned as Γ6g -> Γ7u and Γ6g -> Γ7u respectively in the double group D*3d.

Crystals of K2PtBr4 have the K2PtCl4 structure in which square-planar complexes are stacked plane-to-plane along the c-axis. The polarized ligand-field spectra measured at 4 K are quite similar to those of the chloride, though the ratio of xy : z intensity is a little higher, probably because the first electric-dipole-allowed transition in the U.V. is now an xy-polarized ligand-to-metal charge transfer, and the d ->p transitio - is much higher, at about 48 000 cm-1 (Figure 4). The ion [Re2Cl8]2- is famous for having an Re — Re bond of order 4. In the polarized crystal spectrum of its tetrabutylammonium salt at 4 K the principal band maxima at 14 180,30 870, and 39 215 cm-1 were assigned to [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (Figure 5).

Polyatomic Ligands. In 1974 and 1975 the volume of work in this field expanded far beyond the modest number of references referred to in our previous volumes. These embrace complexes from simple cyano-species up to elaborate conjugated ligands, and emphasize once again how powerful polarized optical spectroscopy is in establishing, often quite unambiguously, the symmetries and orbital characteristics of excited states, even in quite complicated molecules. Our examples are set out in order of increasing atomic number of the central metal atom in the complex.

In K3Cr(CN)6 the zero-phonon line of the 4A2 ->2E transition is sharp enough for Zeeman splittings to be measured, so the g-value for this excited state can be determined. It is nearly isotropic but the symmetry of the complex is lowered slightly by the neighbouring cations. By comparing the vibronic structure in the 2E, 2T1 region of the optical spectrum with frequencies obtained from Raman spectroscopy it has also proved possible to assign nearly all the vibronic sidebands for the related salt K3Cr(NCS)6,4H2O. Lowering the symmetry from octahedral to tetragonal of course increases the number of ligand-field parameters necessary to fit the spectra, and polarized spectra are almost obligatory to give the degree of confidence in the assignments needed before the band energies are fitted to a theoretical model. Quadratic CrIII complexes are favourite subjects in this area and assignments and parameters for trans-[Cr(en)2XY]n+ based on crystal spectra at 77 K are listed in Table 1. They form an excellent self-consistent series.

In an experiment paralleling the one on K3Cr(CN)6 described above, the 77 K spectrum of a crystal of K3Mn(CN)6 has also been reported, including a previously unobserved band system with its attendant hot bands. With the ligand tetraphenylthioimidophosphinate MnII forms the unusual tetrahedral complex [Mn(SPPh2NPPh2S)2], the bands in which are nicely fitted by the parameters Δ = -4685, B = 559, and C = 31 19 cm-1. Built on the spin-orbit components of 4A, 4E are vibrational progressions in quanta of 254 cm-l, corresonding to the Mn — S totally symmetric stretching frequency.

The ligand-field spectra of hydrated Co" chlorides, reported independently by Australian and American groups, contain a number of unusual features, showing that even such apparently straightforward substances may spring surprises when looked at in polarized light at low temperatures. The gross features of the hexahydrate spectrum may be fitted to ligand-field parameters for an axially elongated octahedral site, though the actual values differ considerably:

[FORMULA NOT REPRODUCIBLE IN ASCII]

Since it was performed at a lower temperature (4 K instead of 77 K) and used data from three different faces of the crystal we consider the Australian work to be the more reliable. Separated some thousands of cm-1 from several of the major ligand-field bands, some extra, so-called 'anomalous' small bands appear, which do not fit easily into the ligand-field analysis (Figure 6). They are strongly polarized along the x-axis of the CoCl2(H2O), chromophore and shift strongly on deuteration.

The suggestion is that they are vibronic sidebands involving OH stretching overtones, which break the centre of inversion. Similar 'anomalous' bands appear in the spectrum of the dihydrate but in this compound the very large orthorhombic component in the crystal field makes assignment rather difficult.

A point which has been laboured several times in these pages is that quite precise information on bonding parameters can often be extracted from the polarized spectra even of quite complicated complexes. Examples of this observation could be taken from three sets of data on CoII complexes from the Florence group. Bis-(N-butylpyrrole-2-carbaldimino)CoII contains an elongated tetrahedral CoN4 chromophore, whose ligand-field bands are assigned in D2d using polarization data to obtain tetragonal field parameters as follows:

[ILLUSTRATION OMITTED]

Δ = 13 000 cm-1, B2/B4 = 0.20 and β = 0.80. The corresponding angular-overlap parameters would be e'sigma 2750 cm-1 and e'n/e'σ = 0.43. The second example is 2-picoline N-oxide, which forms a pentakis five-co-ordinate complex with CoII perchlorate, whose crystal spectrum is similarly assigned using a C2v point symmetry. Finally, the four-co-ordinate tribromo(quino1ine)CiII anion, in its tetrabutylammonium salt, has approximately C3v symmetry, though a closer look at the polarized spectra along three independent crystal directions shows that the true symmetry is C1.