6.1 Mössbauer Spectroscopy
This extremely sensitive technique relies on the recoil‐free resonant absorption and emission of gamma rays in solids (Mössbauer effect). It probes small changes in energy levels associated with the nucleus. However, only a relatively limited number of elements are suitable for study. This is because Mössbauer spectroscopy relies on the radioactive decay of a parent isotope to that of interest, which has a sufficiently long half‐life to make real‐time experiments realistic. In glasses, the most common elements exhibiting this effect are iron (57Fe) and tin (119Sn), the radioactive sources being 57Co and usually 119mSn, respectively (cf. [23]). The samples are usually powders (mg) that are mixed with some sort of inert matrix such as sucrose in order to dilute the concentration of the Fe or Sn. If the Fe or Sn concentration is too high, a useful Mössbauer spectrum will not be obtained. One can determine the oxidation state and coordination of Fe in glasses based on the analysis of the isomer shift (IS) and quadrupole splitting (QS) values extracted from Fe Mössbauer spectra. The spectra of iron‐containing glasses generally exhibit an asymmetric doublet whose full width at half maximum are larger than those of crystalline materials. A number of different models are usually fitted to the doublet (Figure 11). In early studies sets of symmetric Lorentzian curves were used to represent the different potential Fe sites but in more recent years fits of the quadrupolar splitting or hyperfine field distributions are usually preferred. Typical IS (δ) values for silicate glasses fall between 0.20 and 0.32 mm/s for Fe3+ in tetrahedral coordination, 0.35–0.55 mm/s for Fe3+ in octahedral coordination, 0.80–0.95 mm/s for Fe2+ in tetrahedral coordination, and 1.05–1.55 mm/s for Fe2+ in octahedral coordination. The presence of fivefold Fe has been observed in glasses with IS and QS values lying between those of four‐ and sixfold Fe. However, it is difficult to distinguish Fe in fivefold coordination from a mixture of Fe in four‐ and sixfold coordination.
Tin spectra are usually broad (FWHM ~1 mm/s) symmetric doublets. For Sn4+, QS values are small and the IS is close to 0 mm/s whereas Sn2+ has IS and QS values of 1–2 and 2–3 mm/s, respectively. For Sn2+, the IS is more negative for tetrahedral than for octahedral sites.
6.2 X‐ray Photoelectron Spectroscopy
Although it has been widely used to study the surfaces of materials (see [24]), X‐ray photoelectron spectroscopy (XPS) can also provide information on the bulk structure. Its application to glasses has been relatively limited to date but it can determine quantitatively the NBO and BO concentrations as well as the concentration of free oxygen (O2−), the oxygens not bound to a network former. The technique measures the kinetic energy of photoelectrons that are ejected from the sample (a mm‐glass chip with a freshly exposed surface) as a result of ionization of an element by an incident X‐ray beam. The kinetic energy of the photoelectron can be simply related to the binding energy (BE) of the electron to the nucleus so that it is characteristic of both the element from which it has been ejected and the environment around the element. One can scan for the presence of specific elements and obtain high‐resolution spectra of well‐resolved energy ranges. The concentration of the different oxygen species are determined by curve fitting of the O 1s XPS spectrum [25]. The BO BE is at higher energy (Δ ~ 1.4 eV) than the NBO peak but their FWHM are comparable (average ~ 1.22 eV). The free oxygen peak is incorporated into that of the NBO to the low‐energy side and is observed as an increase in the FWHM of the NBO peak. This is shown in Figure 12 for a PbO–SiO2 glass. The NBO peak has two contributions, one from the NBOs and one from the presence of free oxygen (O2−), also called non‐network oxygen.
Figure 11 Contributions of Fe2+ (light gray) and Fe3+ (dark gray) to the Mössbauer spectrum of a Fe‐containing borosilicate glass as determined from fits made with 2‐d Gaussian distributions [15].
6.3 Ultra Violet/Visible Spectroscopy
The absorption, reflection, or emission of light in the near UV to near IR (~250–3000 nm) is also a source of structural information. In UV/Vis spectroscopy, one thus measures the absorption of light by a material caused by a number of factors: electronic transitions of transition metal d electrons, intervalence charge transfer (IVCT) and anion–cation charge transfer, electronic transitions between the conduction and valence bands, vibrational overtones, electronic transitions between f‐orbitals, defects, and electron holes [26]. The technique is primarily used to investigate the causes of color in glasses through the oxidation and coordination environment of coloring transition metal elements such as Ni within glasses (Chapter 6.2). In general, specific absorption bands are observed that are characteristic of the transition metal, its oxidation state, and coordination. The samples are usually polished glass chips (mm) or slabs (mm–cm). In addition, sample thickness may need to be adjusted for specific experimental conditions.
Figure 12 The different oxygen species identified from the XPS O 1s spectrum of a lead silicate glass. BO peak is constrained to a linewidth of 1.22 eV and fitted with two peaks in the NBO envelope, one due to NBOs and the other to the presence of free oxygen. Glass made and spectrum obtained by Ryan Sawyer, University Western Ontario.
7 Perspectives
In this chapter I have by no means covered all the possible methods that can be used to investigate the structure of glasses. In addition to variations in many of the specific techniques covered in this chapter, many more techniques including imaging methods such as atomic force microscopy (AFM) and high‐resolution transmission electron microscopy (HRTEM) can be used (Chapter 2.3). Whereas the ability to “solve” the structure as done in crystal‐structure analysis is not possible, our ability to probe the structure of glasses has greatly improved since the first studies in the early decades of the twentieth century. Progress in X‐ray and neutron sources, lasers, detector sensitivity and resolution, and computers have all contributed to improve greatly the resolution and length scales observable in glasses and other amorphous materials.
Unlike for crystalline materials, however, any structural study of a glass must involve a multi‐technique approach since no single method can supply information on all aspects of the structure. In addition, it is advantageous to use numerical techniques such as simulations of spectra, first‐principles molecular dynamics (MD) simulations, and reverse Monte Carlo (RMC) calculations to aid interpretation of the experimental data. While our understanding of the SRO continues to improve, that of the IRO and LRO remains murky, as is the link between structure and, physical properties and behavior, as well as, structural changes that occur with increasing pressure and temperature. With continued advances in experimental instrumentation and computational methodologies,