In order to clarify the nature of the neutral Mn acceptor in GaP, we have carried out optical-absorption and electron-paramagnetic-resonance ͑EPR͒ experiments using both conventional and thermally detected EPR on semi-insulating GaP:Mn. In thermal equilibrium at low temperatures, all the manganese occurs in the charged acceptor state Mn Ga 2ϩ ͑A Ϫ ͒. By illumination with photon energies greater than 1.2 eV, it can be partially converted into the neutral charge state. The arising photostimulated EPR spectrum shows the characteristic of a tetragonally distorted center with an integer spin. The resonance lines are detectable only at temperatures below 7 K, and their linewidth of about 50 mT is due to the unresolved Mn-hyperfine splitting. We interpret the experimental data in terms of Mn Ga 3ϩ ions on strain-stabilized sites of tetragonal symmetry due to a strong T ⑀ Jahn-Teller coupling within the 5 T 2 ground state. Such a behavior is expected for a 3d 4 defect, as observed for the isoelectronic impurity Cr 2ϩ in GaAs, and other tetrahedrally coordinated semiconductors. The analysis of the EPR spectra thus verifies that, in GaP, the neutral charge state of the Mn acceptor is Mn Ga 3ϩ ͑A 0 ͒ in contrast to its behavior in GaAs and InP. ͓S0163-1829͑96͒06139-5͔

Estimates of the first-order Jahn - Teller reduction factor () of the state of in III - V materials are made from an examination of FTIR studies. Values of between 0.01 and 0.2 are deduced. .

No abstract

The propagation equation, written in a curvilinear coordinate system, is solved by using a perturbation method inspired by quantum physics. The parameter of perturbation is the groove depth which is small compared to the period. The method is expanded up to second order for the non-degenerate problem. In this way the solutions have analytical form compared to a numerical method. They present the advantage of providing evidence for the evolution of the energy distribution for different diffraction orders as a function of the magnitude of the perturbation. The results that are deduced from these analytical solutions are within less than 5 × 10 − 3 of those obtained by the curvilinear coordinate method. Thus, this new approach opens a new range of applications for inverse problems.