Computational screening of new inorganic materials for highly efficient solar energy conversion

Abstract

The world’s energy consumption is rising, and the demand for energy is projected to increase even more in the future. At the moment, we rely on energy from fossil fuels, which are harmful to the environment, and the future increases in consumption will be strongly harmful to both the environment and human well being. Additionally, fossil fuels are non-renewable, and there are only limited reserves on the planet. An alternative to using fossil fuels is to harvest energy from renewable and environmentally benign sources, such as the sun, which theoretically provides the largest source of energy that we have access to. Despite the vast amounts of energy at our disposal, we are not able to harvest this solar energy efficiently. Currently, there are a few ways of converting solar power into usable energy, such as photovoltaics (PV) or photoelectrochemical generation of fuels (PC). PV processes in solar cells convert solar energy into electricity, and PC uses harvested energy to conduct chemical reactions, such as splitting water into oxygen and, more importantly, hydrogen, also known as the fuel of the future. Further progress in both PV and PC fields is mostly limited by the flaws in materials used as photoabsorbers. Silicon as an absorber dominantes the PV community and all other semiconductors face significant challenges, for various different reasons, when suggested for large scale deployment. Advances in PC, on the other hand, are mostly held back due to the fact that no single material, which can both absorb light and catalyse the relevant chemical reactions, has been found. A proposed alternative, using two materials in tandem instead of one, has the potential to successfully perform the task. However, progress in this field is inhibited by the lack of high band gap photoabsorbing materials. In this work a high-throughput computational search for suitable absorbers for PV and PC applications is presented. A set of descriptors has been developed, such that each descriptor targets an important property or issue of a good solar energy conversion material. The screening study was performed step-wise, so that in each step a new descriptor and associated criterion were introduced and all the materials failing to satisfy the criterion are removed from the study. The corresponding descriptors were obtained within the scope of quantum mechanics using Density Functional Theory. This method of materials design is first applied to materials found by substituting atomic cations in crystal structures of the ABS3 stoichiometry, resulting in several candidates which we believe have the potential to work in a PV and PC device. One of these candidates has been successfully synthesized by our collaborators, and the measured band gap is in accordance with the theoretically calculated one. Furthermore, a study on previously synthesized semiconductors yielded a list of stable materials which have not yet been explored for PV or PC. A similar study has been performed on II-IV-V2 compounds, and has revealed some interesting trends within the class, resulting in several interesting candidate materials. A few of these have already been extensively investigated by others.