Photoluminescence of BaCd₂P₂ as a New Defect-Insensitive Solar Material

Searching for earth-abundant, defect-insensitive solar materials can play a significant role in future photovoltaics. Here we experimentally verify BaCd₂P₂ as a promising candidate for low-cost, high-efficiency photovoltaic material, as predicted by high-throughput screening of nearly 40,000 semiconductors. Its optimal band gap of 1.45 eV and properties such as temperature stability and long carrier lifetime make BaCd₂P₂ a strong photovoltaic candidate. We were able to experimentally demonstrate the optical performance of BaCd₂P₂ by comparing its photoluminescence (PL) intensity with that of GaAs. The relatively low-purity (99.9%) BaCd₂P₂powder had direct gap PL intensity in the same order as GaAs powder obtained from a prime single crystalline wafer, suggesting that it is insensitive to impurities and defects. We further present an in-depth investigation of the PL spectrum of this promising material by studying the impact of temperature, excitation laser power, and composition variations. Nearly all our experimental data have agreed with our theoretically calculated properties of BaCd₂P₂. The calculated band structure suggested a direct gap of 1.45 eV and an indirect gap of 1.48 eV. The PL spectrum of BaCd₂P₂ at 298 K, had a peak at 847 nm (1.46 eV), aligning well with the predicted band structure, and another, broader, peak at 980 nm. At 78 K we observed a separate peak on the lower wavelength side of the 847 nm peak resulting from the indirect gap transition, further validating our computed band structure. This material also showed great temperature stability, increasing the temperature to 368 K had only a slight decrease in the PL emission by ~5%, which fully recovered when going back to 298 K. Our work on excitation power dependence of PL intensity has enabled us to confirm the nature of 847 nm peak and gain a better understanding of the 980 nm peak. We did this by modifying the well-known equation I_PL= C * (P_laser)^α (meant to model the change in PL peak intensity with power) to account for competing nonradiative processes such as Auger recombination at higher excitation powers and consider the potential saturation of available states. We developed a more generalized equation that treats PL intensity as a function of thermal excitation, laser excitation, and nonradiative recombination. For the 847 nm peak, we repeatedly got a value of the α parameter that was strongly indicative of a band-to-band transition. We saw that the 980 nm peak does not blueshift with decreasing temperature and that it had a strikingly similar response to PL laser excitation power variations as the radiative defect of Si-doped GaAs, appearing at 950 nm, leading us to hypothesize that the 980 nm peak of BaCd₂P₂ is due to a shallow radiative defect. We conducted an evaluation of the change in the 950 nm defect PL peak when using off-stoichiometric BaCd₂P₂ samples. We observed that the peak intensifies and is slightly redshifted in Ba-poor samples. We further used the data from our studies of PL intensity change with temperature to calculate the activation energy of the prominent nonradiative defects.