Mössbauer Spectroscopy and Transition Metal Chemistry

Some 50 years ago, Rudolf L. Mössbauer, while working on his doctoral thesis under Professor Maier-Leibnitz at Heidelberg/Munich, discovered the recoilless nuclear resonance absorption (fluorescence) of γ rays, which subsequently became known as the Mössbauer effect [1–3]. Some three decades before this successful discovery, Kuhn had speculated on the possible observation of the nuclear resonance absorption of γ-rays [4] similar to the analogous optical resonance absorption which had been known since the middle of the nineteenth century. The reason that nuclear resonance absorption (or fluorescence) was so difficult to observe was clear. The relatively high nuclear transition energies on the order of 100 keV impart an enormous recoil effect on the emitting and absorbing nuclei, such recoil energies being up to five orders of magnitude larger than the γ-ray line width.

Mössbauer spectroscopy is based on recoilless emission and resonant absorption of γ-radiation by atomic nuclei. The aim of this chapter is to familiarize the reader with the concepts of nuclear γ-resonance and the Mössbauer effect, before we describe the experiments and relevant electric and magnetic hyperfine interactions in Chaps. 3 and 4. We prefer doing this by collecting formulae without deriving them; comprehensive and instructive descriptions have already been given at length in a number of introductory books ([7–39] in Chap. 1). Readers who are primarily interested in understanding their Mössbauer spectra without too much physical ballast may skip this chapter at first reading and proceed directly to Chap. 4. However, for the understanding of some aspects of line broadening and the preparation of optimized samples discussed in Chap. 3, the principles described here might be necessary.

In this chapter, we present the principles of conventional Mössbauer spectrometers with radioactive isotopes as the light source; Mössbauer experiments with synchrotron radiation are discussed in Chap. 9 including technical principles. Since complete spectrometers, suitable for virtually all the common isotopes, have been commercially available for many years, we refrain from presenting technical details like electronic circuits. We are concerned here with the functional components of a spectrometer, their interaction and synchronization, the different operation modes and proper tuning of the instrument. We discuss the properties of radioactive γ-sources to understand the requirements of an efficient γ-counting system, and finally we deal with sample preparation and the optimization of Mössbauer absorbers. For further reading on spectrometers and their technical details, we refer to the review articles [1–3].

In Chap. 2, for the sake of simplicity, we dealt with transitions between unperturbed energy levels of “bare” nuclei with the mean transition energy E ₀. In reality, however, nuclei are exposed to electric and magnetic fields created by the electrons of the Mössbauer atom itself and by other atoms in the neighborhood. These fields generally interact with the electric charge distribution and the magnetic dipole moment of the Mössbauer nucleus and perturb its nuclear energy states. The perturbation, called nuclear hyperfine interaction, may be such that it shifts the nuclear energy levels, as is the case in the electric monopole interaction (e0), or such that it splits degenerate states, as afforded by the electric quadrupole interaction (e2) and the magnetic dipole interaction (m1). Only these three kinds of interaction must be considered in practical Mössbauer spectroscopy.

The underlying physics and analysis of Mössbauer spectra have been explained in detail in Chap. 4. In that chapter, the principles of how a spectrum is parameterized in terms of spin-Hamiltonian (SH) parameters and the physical origin of these SH parameters have been clarified. Many Mössbauer studies, mainly for ⁵⁷Fe, have been performed and there is a large body of experimental data concerning electric- and magnetic-hyperfine interactions that is accessible through the Mössbauer Effect Database.

Studies of magnetic relaxation phenomena are important for understanding the properties of magnetic materials, and Mössbauer spectroscopy has contributed much to the elucidation of the magnetic dynamics in solids. In experimental studies of relaxation phenomena, the time scale of the experimental technique is a crucial parameter, because it determines the range of relaxation times that can be measured. DC magnetization measurements have a time scale on the order of seconds, and in AC magnetization measurements the time scale can be varied from ∼1 to ∼10⁻⁷ s by varying the frequency.

The previous chapters are exclusively devoted to the measurements and interpretation of ⁵⁷Fe spectra of various iron-containing systems. Iron is, by far, the most extensively explored element in the field of chemistry compared with all other Mössbauer-active elements because the Mössbauer effect of ⁵⁷Fe is very easy to observe and the spectra are, in general, well resolved and they reflect important information about bonding and structural properties. Besides iron, there are a good number of other transition metals suitable for Mössbauer spectroscopy which is, however, less extensively studied because of technical and/or spectral resolution problems. In recent years, many of these difficulties have been overcome, and we shall see in the following sections a good deal of successful Mössbauer spectroscopy that has been performed on compounds of nickel (⁶¹Ni), zinc (⁶⁷Zn), ruthenium (mainly ⁹⁹Ru), tantalum (¹⁸¹Ta), tungsten (mainly ¹⁸²W, ¹⁸³W), osmium (mainly ¹⁸⁹Os), iridium (¹⁹¹Ir, ¹⁹³Ir), platinum (¹⁹⁵Pt), and gold (¹⁹⁷Au). The nuclear γ-resonance effect in the case of technetium (⁹⁹Tc), silver (¹⁰⁷Ag), hafnium (¹⁷⁶Hf, ¹⁷⁷Hf, ¹⁷⁸Hf, ¹⁸⁰Hf), rhenium (¹⁸⁷Re), and mercury (¹⁹⁹Hg, ²⁰¹Hg) has been of relatively little use to the chemists, so far. There are various reasons responsible for this, viz., (1) extraordinary difficulties in measuring the resonance effect because of the long lifetime of the excited Mössbauer level and hence the extremely small transition line width (e.g., in ⁶⁷Zn), (2) poor resolution of the resonance lines due to either very small nuclear moments or the very short lifetime of the excited Mössbauer level resulting in very broad resonance lines, (3) insufficient resonance effects due to unusually high transition energies between the excited and the ground nuclear levels, which in turn increase the recoil energy and thus reduces the recoilless fraction of emitted and observed γ-rays.

We have learned from the preceding chapters that the chemical and physical state of a Mössbauer atom in any kind of solid material can be characterized by way of the hyperfine interactions which manifest themselves in the Mössbauer spectrum by the isomer shift and, where relevant, electric quadrupole and/or magnetic dipole splitting of the resonance lines. On the basis of all the parameters obtainable from a Mössbauer spectrum, it is, in most cases, possible to identify unambiguously one or more chemical species of a given Mössbauer atom occurring in the same material.

Conventional Mössbauer spectroscopy (MS) can be considered as “spectroscopy in the energy domain.” It has been widely used since its discovery in 1958 [1]. Nuclear resonant forward scattering (NFS) of synchrotron radiation has been successfully employed as a time-differential technique since 1991 [2]. Another related technique, nuclear inelastic scattering (NIS) of synchrotron radiation [3], can be regarded as an extension of conventional, energy-resolved MS (in the range 10⁻⁹ to 10⁻⁷ eV) to energies on the order of molecular vibrations (in the range 10⁻³ to 10⁻¹ eV). So far only a few “Mössbauer” stations for NFS and NIS measurements have become available in synchrotron laboratories, i.e., in Germany, France, Japan, and the USA.

This paragraph presents a summary of the most relevant expressions provided by Long et al. [1] for the optimization of “thin” absorbers (effective thickness t ≪ 1) with high mass absorption. The result of this work is used in Sect. 3.3.2 of the book. Following the approach of [1], we adopt for the signal-to-noise ratio: