The optical properties of Be,Mg and Zn-diffused gallium phosphide

Interest in the Ga-site acceptors Be and Mg was stimulated by the possibility that they might produce efficient luminescence on association with O, analogous to the well-known red Zn-O luminescence in GaP but at higher transition energy. Attention was directed to diffusion doping by Be and Mg of GaP O-doped during growth because the reactivity of Be and Mg with O renders double doping during crystal growth very difficult. Structured green donor-acceptor pair spectra were observed at 1.6°K from many Be-diffused crystals, yielding an accurate measure of (E_A)_BE, 50 ± 1 meV. Moderately efficient orange-red luminescence was also observed below ∼ 100 °K from these crystals, but the intensity of this luminescence decreased rapidly to negligible levels by ∼ 200°K. This luminescence also contains sharp structure at 1.6°K, of a form characteristic of the decay of excitons bound to complex centres. Many sharp phonon replicas occur, involving local modes as well as characteristic GaP modes. One set of no-phonon lines, at least, near 2.19 eV, shows zero-field splitting, luminescence decay times and behaviour in magnetic and external strain fields characteristic of exciton decay at a centre with <100>; or <111>-type symmetry axes, containing no extra electronic particles. The exciton state is split by 2.4 meV by J-J coupling, and the axial field of the centre splits the hole states by ∼ 1.0 meV. These bound excitons are specifically characteristics of diffused GaP and appear analogous to bound excitons observed below 2.12 eV in Zn-diffused GaP. It is probable that the relevant centres contain diffusion components such as Be or Zn interstitials and improbable that O_P is involved. By contrast, weak orange bound exciton luminescence observed in Mg-diffused GaP does involve O, presumably as O_P. No analysis of the magneto-optical behaviour of this Mg-related bound exciton was possible in our crystals, so its symmetry axis was not established. It is possible that this is the Mg_Ga-O_P bound exciton. If so, the two-fold reductions in the exciton localisation energy from ∼ 0.32 eV to ∼ 0.15 eV and in the mass of the Ga-site substituent has produced dramatic changes in the form of the phonon cooperation between the Zn-O and “Mg-O” excitons. The “Mg-O” exciton luminescence is not dominant in our crystals, even at low temperature. The exciton state is again split by a local crystal field as well as by J-J coupling, but here the former splitting is predominant; 2_∈0 = 3.9 meV, Δ = 0.60 meV.