The DX Center: How Complicated can a Point Defect Be?
- PDF / 2,207,310 Bytes
- 12 Pages / 420.48 x 639 pts Page_size
- 44 Downloads / 225 Views
THE DX CENTER: HOW COMPLICATED CAN A POINT DEFECT BE? Thomas N. Theis IBM Research, T.J. Watson Research Center, P.O. Box 218, Yorktown Heights, NY 10598 ABSTRACT I summarize a large body of experimental and theoretical work, especially in Sidoped GaAs and AlxGalxAs, which has led to our present understanding of the DX center. There is good evidence that the DX center is just the simple donor, but each donor atom can exist in either of two distinct lattice configurations, each with its own spectrum of bound electronic states. At present the model which appears to best explain this unexpected complexity is that of Chadi and Chang, which predicts a large bond-breaking distortion of the lattice, accompanied by capture of two electrons by the donor. I argue that the best evidence to date for the predicted distortion is provided by observations of alloy perturbations of the DX level. Furthermore, there is now very convincing evidence that the DX level is a two-electron state. I briefly summarize what is known about the bound state spectra of the substitutional and relaxed (DX) configurations, and then discuss the very interesting question of how the donor captures two electrons from the conduction band to the DX level. There is good, although indirect, evidence that an excited one-electron state acts as an intermediate in the thermal emission and capture process. A different excited state appears to 15e involved in photoemission. Although much of this complexity was unforeseen by Chadi and Chang, it nevertheless seems to be consistent with their model. However, I point out a number of issues which still require experimental or theoretical resolution. INTRODUCTION It is well known that the impurity potential of a substitutional donor in an otherwise perfect semiconductor crystal can be written as the sum of a long-range coulombic potential and a short-range central cell potential [I]. The resulting spectrum of electronic states can be (somewhat arbitrarily) divided into shallow or hydrogenic states associated with the coulomb potential and "deep" or localized states associated with the central cell potential. Either sort of state may exist as a bound state in the fundamental gap, or as a resonance in the conduction band. Deep or shallow, bound or resonant, such states are expected to equilibrate rapidly with the states of the conduction band by radiative processes. However, in many compound semiconductors, another donor-related deep state, the DX state, is observed, which at low temperatures, equilibrates exceedingly slowly. In order to explain the thermally activated capture rate, the absence of an observable radiative capture rate, and the very large photoionization energy, Lang et al. [2] argued that this state was separated from the conduction states by an energetic barrier corresponding to a large distortion of the lattice around the donor (large lattice relaxation). There is abundant evidence for large lattice relaxation upon electron capture to the DX level. Photoionization studies yield an optical ionization ener
Data Loading...
