Stress splitting of theEL2 zero-phonon line: Need for reinterpretation of the main optical transition

The T₂ final state of the main optical transition of EL2 can couple to lattice distortions of a, {ε}, and {τ} symmetry. The stress splitting of the zero-phonon line (ZPL) observed by Kaminska et al. gives information about the strength of the coupling between the lattice and the T₂ final state. We have obtained that information and have shown that the lattice relaxation associated with the {τ} mode used by Kaminska et al. to account for their results cannot be larger than about 70 meV, and is probably much closer to 45 meV. We find that the lattice relaxation associated with the {ε} modes is negligible, and we estimate, from analysis of hydrostatic-pressure experiments, that the lattice relaxation associated with the a modes is about 90 meV. Thus, the separation of 140 meV between the peak in the optical absorption at 1.18 eV and the ZPL at 1.04 eV (a separation commonly taken to be due to lattice relaxation) can be accounted for. If this agreement is interpreted as strengthening the case that the ZPL and the peak in the optical absorption are related to each other by lattice relaxation, it exacerbates well-known problems in understanding the optical absorption. Among these problems are the intensity of the ZPL, the fact that it moves in a direction opposite to that of the main peak under hydrostatic pressure, and the fact that it rides up over the beginning of the main peak when hydrostatic pressure is applied. Our calculation does make it clear that there must be at least two local modes, one of a symmetry and the other of {τ} symmetry, in addition to any of the perfect-crystal phonon modes usually considered. Although this potentially adds a new feature to the analysis, it does not by itself resolve the problems with the absorption. We discuss the role that ionization and recapture may play in resolving these problems and suggest an experiment that might clarify this issue.