Phonon scattering by impurities in semiconductors
Thermal conductivity measurements have been used to study the low lying energy levels of Cr ions in GaAs. Strong resonant phonon scattering was observed in semi-insulating (SI) and p-type samples, which is attributed to Cr 2+ or Cr 3+ ions, while the scattering in the n-type samples additional to that in undoped material was very small. From the computer fits of the thermal conductivity, zero-field ground state splittings have been deduced. A splitting at ~ 23 cm -1 is attributed to Cr 3+ ions, others at ~ 0.7 cm and 4.9 cm are due to Cr 2+. The phonon scattering in the n-type samples did not show any magnetic field dependence while big effects were observed in SI and p-type ones. This seems consistent with the results of the zero-field work. The effect of uniaxial stress on the phonon scattering has been measured in the temperature range ~2-15 K. Again no effects were seen in n-type material. The results for SI and p-type material are interpreted in terms of a static Jahn-Tellereffect of Cr 2+ ions. A preliminary investigation was made of the effect on the phonon scattering of sub-band-gap illumination. In an-type sample, the decay in the increase in the thermal resistivity produced by photoexcitation showed two parts. The first part with a characteristic time of ~ 1 hour is attributed to electron-capture at Cr 2+ ions. The second decay was very slow (persistent) with a characteristic time T » 10 5 s. This effect has tentatively been attributed to the occurrence of large lattice relaxation. The phonon scattering by the Cr ions is found to be consistent with the 'double acceptor' model for Cr in GaAs. Another model where er can act as a hole trap is discussed. Finally the effect of high magnetic fields on the thermal conductivity of acceptors in Ge was measured. From this and previous results, the g-values describing the ra8 ground state were found to be much lower than the predicted ones. A Thermally Detected EPR apparatus was designed and constructed in an attempt to check on the ground state structure of p-Ge and also p-Si but no results were obtained. This is believed to be due to the very large line widths resulting from strain splitting of the ra8 ground state.