Research on selenium solar cells is regaining momentum due to the exciting prospect of integrating a single-element, wide-bandgap (≈1.95 eV) photoabsorber in tandem with a lower bandgap photovoltaic device. Low temperature processing of selenium makes it a potentially inexpensive candidate for a top cell absorber that can easily be integrated with existing photovoltaic technology. However, single-junction selenium devices still exhibit a significant voltage deficit when compared to the Shockley–Queisser limit, despite recent efficiency improvements through device engineering. This calls for more detailed characterization of the selenium material properties linked to photovoltaic performance. In this work, we first fabricate large-area selenium solar cells with a record open-circuit voltage of 0.99 V and a pseudo fill factor of 80%. Then, we characterize the mobility-lifetime product, built-in voltage and low-temperature radiative recombination in these state-of-the-art devices. We attribute most of the Voc-deficit to non-radiative recombination in the selenium absorber. Thus, improving the bulk optoelectronic quality of selenium appears to be a more urgent need than the optimization of its device structure.
Elemental selenium (Se) is experiencing a renaissance as a p-type direct wide bandgap (1.95 eV) photoabsorber, appropriate for integration with lower bandgap materials in tandem photovoltaic devices. However, single-junction selenium devices are typically in the superstrate configuration with the charge-separating p–n junction located very close to the substrate. For tandem devices, the p–n junction should ideally lie near the surface of the cell to maximize photocarrier collection, implying that the n-type heterojunction partner should be deposited on top of Se. Since Se is a soft, low-melting-point material, a low-damage deposition technique should then be identified for the partner material. Here, we investigate the suitability of CdS grown by chemical bath deposition as an n-type partner for Se cells. We demonstrate the first functioning CdS/Se solar cell in the standard superstrate configuration. Although it may be possible to realize a substrate configuration, we find that the bandgap and electron affinity of CdS are not optimal for efficient carrier absorption and transport in selenium devices, suggesting that CdS is a poor partner for selenium.