Genomic or data-driven materials research is emerging as a new paradigm of materials discovery. By integrating experiments, computation, data mining and informatics, genomic approach is expected to help shorten the time it takes to move a newly discovered advanced material from the laboratory to the commercial market place, at a reduced cost. Due to strong support and intensive activities in research on 2D materials in local research centres and labs, we are developing a database for 2D materials. Various methods are used to generate 2D structures and high throughput first-principles calculations based on density functional theory are carried out to study their stability and physical properties.
As one example of materials prediction based on such an approach, we have carried out high-throughput computational screening to identify van der Waals heterostructures for efficient excitonic solar cell application. Band alignments of 1540 vertical heterostructures formed by 56 two-dimensional semiconducting/insulating materials were investigated. More than 90 heterostructures with estimated power conversion efficiency (PCE) higher than 15% have been identified, of which 17 heterostructures are predicted to have PCE higher than the best value (20%) reported or proposed in the literature. More details can be found in Linghu, J. et al., “High-Throughput Computational Screening of Vertical 2D van der Waals Heterostructures for High-efficiency Excitonic Solar Cells“, ACS Applied Materials & Interfaces, 10, 32142-32150 (2018).
On-going projects include: (i) further expansion of the 2D materials collection by both top-down and bottom-up approaches; (ii) collection of 2D materials data from literature; (iii) analysing 2D materials data to identify trends and descriptors; (iv) adapting machine learning methods to design and prediction of 2D materials; (v) applications of 2D materials, such as development of sensors for Internet of Things.