Degree of polarization of photoluminescence from facets of InP as a function of strain: some experimental evidence

Previous work demonstrated a good fit to the degree of polarization (DOP) of luminescence measurements on {110} facets of InP using a simple dependence of DOP of luminescence on strain: ${-}{K_e} ({e_1} - {e_3})$, where ${K_e}$ is a positive calibration constant, and ${e_1}$ and ${e_3}$ are normal components of strain in the plane of the facet and along $\langle 1\bar 10\rangle$ and $\langle 001\rangle$ directions [Appl. Opt.43, 1811 (2004)APOPAI0003-693510.1364/AO.43.001811]. Recent analytic modeling, which by necessity to be analytic must make simplifying assumptions, has suggested that unless the measurements are along crystallographic axes, the dependence of the DOP of luminescence on strain is more complicated: ${-}{K_e} (1.315 {e_1} - 0.7987 {e_3})$ for measurements from an InP facet, with a similar "excess" ${e_1}$ for GaAs [Appl. Opt.59, 5506 (2020)APOPAI0003-693510.1364/AO.394624]. In this work, we fit finite element simulations (FEM) to DOP measurements of the photoluminescence from facets of InP bars with ${\{111\} _B}$ v-grooves that have been placed in a cylindrical bending moment. We find that the more complicated dependence of DOP on strain, as derived by the analytic model, fits the data better than the previously assumed simple dependence. This finding thus corroborates the analytical model and should have an impact on understanding the strain-dependent operation of optoelectronic devices.