Properties of the As-related shallow acceptor level in heteroepitaxial ZnSe grown by molecular-beam epitaxy

The As-related shallow acceptor level in ZnSe is characterized in detail by low- and variable-temperature photoluminescence (PL), selective-pair luminescence, and magnetospectroscopy measurements, using intentionally As-doped material grown by molecular-beam epitaxy on GaAs. The shallow-acceptor-related PL features grow progressively stronger with Zn3As2 flux, while deep-level peaks at 2.20 and 1.7 eV are observed only at the highest doping levels. In addition to the previously reported donor-to-acceptor peak at about 2.693-2.697 eV, we observe a corresponding band-to-acceptor peak at temperatures of ∼25 K and above. The temperature dependence of the band-acceptor peak position yields a light-hole-acceptor binding energy of 114.6±0.7 meV for thermally strained material, which compares to values of 114.1±0.4 and 113.0±0.6 meV we recently obtained in the same way for Li and N acceptors, respectively. The corresponding value for unstrained material is calculated to be EAAs=114.8±0.7 meV. Selective pair luminescence is used to study the excited states. Four states are observed and assigned to the 2p3/2, 2s3/2, 2p5/2 (Γ7), and 3s3/2 levels, respectively. These levels lie about 73.6, 83.9, 93.5, and 98.4 meV above the ground state, respectively, which agrees with calculations based on effective-mass theory. In the excitonic region, the previously reported As acceptor-bound exciton peak at 2.7888 eV is found to exhibit a high energy component at 2.7903 eV. The observed splitting, which is similar to that of other shallow acceptor-bound excitons in relaxed heteroepitaxial layers, is modeled by a calculation of the effects due to thermal mismatch strain. The splitting of the acceptor-bound exciton is studied in magnetic fields up to 12 T as a function of orientation. The results suggest that the As-related acceptor has the point symmetry of the lattice, implying that it may involve a simple substitutional acceptor involving As on the Se site. The results imply that the failure to obtain p-type conductivity to date with this dopant is not due to the nonexistence of a suitable shallow level. Further work is necessary to investigate improved incorporation methods and to determine if the shallow level is metastable with respect to lattice distortion.