X-ray creation and activation of electron spin resonance in vitreous silica.

Abstract

We report strong dependencies of electron-spin-resonance (ESR) signals in fused silica on x-ray dose, sample hydroxyl concentration [OH], and preirradiation sample-annealing temperature ${\mathit{T}}_{\mathit{F}}$. In general, the strengths of the E'-center ESR signal and the two oxgyen-hole-center signals increase nonlinearly with Cu-target x-ray dose; we ascribe this behavior to the coexistence of two processes: (1) The x-ray activation of the spins of pre-existing structural defects, and (2) the x-ray creation of new similar defects. From this model and the observed nonlinear dependence on dose, we obtain relative values for the number of pre-existing defects, their rate of activation by x rays, and the rate of creation of new defects. These relative values are then examined for their trends with respect to [OH] and ${\mathit{T}}_{\mathit{F}}$: For example, there are far more pre-existing E' precursor defects in ``dry'' Suprasil-W1 ([OH]2 ppm) than in ``wet'' Suprasil-1 ([OH]\ensuremath{\approxeq}1200 ppm) fused silica, yet new E' defects are about twice as easy to create with x rays in the ``wet'' material. In general, we find that the relative x-radiation ``hardness'' of different fused silica samples is strongly dependent on dose, hydroxyl content, and thermal history. From the [OH] and ${\mathit{T}}_{\mathit{F}}$ dependencies of spin concentrations, we conclude that none of the three spin-active defects can be identified with either of the two previously reported Raman-active defects, traditionally labeled ${\mathit{D}}_{1}$ and ${\mathit{D}}_{2}$. Moreover, we conclude that ${\mathit{D}}_{2}$ is not a strongly preferential site for formation of the spin-active defects under x irradiation, and that ${\mathit{D}}_{1}$ is unlikely to be such a site.