Trapped-electron centers in pure and doped glassy silica: A review and synthesis

Abstract

Abstract This paper reviews half a century of research on radiation-induced point defects in pure and doped glassy silica and its crystalline polymorph α quartz, placing emphasis on trapped-electron centers because the vast majority of all presently known point defects in various forms of SiO2 are of the trapped-hole variety. The experimental technique most discussed here is electron spin resonance (ESR) because it provides the best means of identifying the point defects responsible for the otherwise difficult-to-attribute optical bands. It is emphasized that defects in α quartz have been unambiguously identified by exacting analyses of the angular dependencies of their ESR spectra in terms of the g matrix of the unpaired electron spin and the matrices of this spin's hyperfine interactions with non-zero-nuclear-spin 29Si and 17O nuclides in pure α quartz and/or with substitutional 27Al, 31P, or 73Ge in quartz crystals respectively doped with Al, P, or Ge. Many defects in pure and doped glassy silica can be unambiguously identified by noting the virtual identities of their spin Hamiltonian parameters with those of their far better characterized doppelgangers in α quartz. In fact, the Ge(1) trapped-electron center in irradiated Ge-doped silica glass is shown here to have a crystal-like nature(!), being virtually indistinguishable from the Ge(II) center in Ge-doped α quartz [R.J. McEachern, J.A. Weil, Phys. Rev. B 49 (1994) 6698]. Still, there are other defects occurring in glassy silica that are not found in quartz, and vice versa. Nevertheless, those defects in glasses without quartz analogues can be identified by analogies with chemically similar defects found in one or both polymorphs and/or by comparison with the vast literature of ESR of paramagnetic atoms and small molecules. Oxygen “pseudo vacancies” comprising trigonally coordinated borons paired with trigonally coordinated silicons were proposed to exist in unirradiated B2O3–3SiO2 glasses in order to account for observations of γ-ray-induced trapped-electron-type B- and Si-E′ centers [D.L. Griscom et al., J. Appl. Phys. 47 (1976) 960]. Analogous Al-E′ trapped-electron centers have been elucidated in silica glasses with Al impurities [K.L. Brower, Phys. Rev. B 20 (1979) 1799]). And it has been proposed [D.L. Griscom et al., J. Appl. Phys. 47 (1976) 960] that trapping of a second electron on such oxygen pseudo vacancies accounts for the predominant ESR-silent trapped-electron centers in irradiated silica glasses containing B or Al. The present paper additionally attempts to divine the identities of some of the ESR-silent radiation-induced trapped-electron centers in silica glasses free of foreign network-forming cations. This quest led to the doorstep of the most famous ESR-silent defect of all, the twofold-coordinated silicon, which is found only in silica glasses (not in quartz) and is responsible for the UV/visible optical properties of the oxygen-deficiency center known as ODC(II). The oxygen-deficiency center called ODC(I) is associated with an absorption band at 7.6 eV and, though commonly believed to be a simple oxygen mono-vacancy, is clearly more complicated than that [e.g., A.N. Trukhin, J. Non-Cryst. Solids 352 (2006) 3002]. Certain well documented but persistently enigmatic ODC(I)↔ODC(II) “interconversions” [reviewed by L. Skuja, J. Non-Cryst. Solids 239 (1998) 16] have never been explained to universal satisfaction. An innovative combined ESR/thermo-stimulated-luminescence (TSL) study of a series of pure low-OH silica glasses with oxygen deficiencies of 0.000, ~ 0.015, and ~ 0.030 vol.% [A.N. Trukhin et al., J. Non-Cryst. Solids, 353 (2007) 1560] places new constraints on all future models for ODC(II). Taking this history into account, specific redefinitions of both ODC(I) and ODC(II) are proposed here. The present review also revisits a study of X-ray-induced point defects in an ultra-low-OH, high-chlorine but otherwise ultra-high-purity silica glass [D.L. Griscom, E.J. Friebele, Phys. Rev. B34 (1986) 7524], arguing that (1) most of the reported E′γ and E′δ centers were created via the mechanism of dissociative electron capture at chlorine-decorated oxygen vacancies, (2) the concomitantly created interstitial chloride ions serve as ESR-silent trapped-electron traps, (3) the ESR-detected “Cl0” centers arise from hole trapping on O3≡ Si–Cl units without detachment of the resulting Cl atom, and (4) those chlorine atoms that are detached by homolytic bond fission are ESR-silent. Finally, in chlorine-free, low-OH, high-purity silica glasses, up to 100% of the trapped-electron centers appear to be ESR silent and are tentatively ascribed to electron trapping in pairs below the mobility edge of the conduction band. In such cases, the sum of all trapped-hole centers has been found to decay exponentially with increasing isochronal annealing temperature in the range 100 to ~ 300 K [D.L. Griscom, Nucl. Inst. & Methods B46 (1990) 12]. Overall, this review consolidates a large amount of long-existing but often underappreciated knowledge bearing on the natures of trapped-electron centers in pure and doped glassy silica, proposes new models for some of these, and raises a number of questions that cannot be fully answered without future performance of new experiments and/or ab initio calculations.