Computational high throughput screening methods and data driven materials design

点击翻译按钮获取中文摘要

Climate change is one of the greatest challenges of our generation and the need to mitigate CO2 emissions is urgent. In the October 2018 Intergovernmental Panel on Climate Change (IPCC) report it is stated that limiting global warming to 1.5ºC would require CO2 emissions to fall by an astonishing 45% from 2010 levels by 2030. With such short timelines "Carbon Capture is the only hope for mankind" as dramatically put by Sir David King, the former UK government chief scientist. Since ~35% of the world's anthropogenic CO2 emissions arise from electrical power generating plants that burn fossil fuels, there has been substantial interest in technology to capture CO2 from the combustion flue gases of such point sources. Carbon capture and storage (CCS) involves scrubbing CO2 from the combustion flue gas and permanently storing the CO2 in relatively pure form. Although several large scale CCS projects exist that capture and store more than a million tons of CO2 per year, the solvent-based CO2 scrubbing technologies currently used in these projects are too energetically costly for wide-scale deployment. At the forefront of alternative technologies for low energy CCS is the use of solid sorbents (instead of liquids) to capture the CO2 within what are called pressure swing adsorption (PSA) systems. This should offer low energy and low cost CO2 capture as long as the right materials can be found and to date such materials have not been identified. The short and long term goals of the proposed research are to develop new computational tools to accelerate the discovery of new materials, with a particular focus on materials that will enable low cost CCS. We have two distinct avenues towards this goal. The first is to integrate industrial process simulations into the rational design of materials at the atomistic level. That is, to bridge the gap between materials design and process engineering. The second avenue is to develop cutting edge high-throughput screening, data-mining and machine learning methods for materials discovery. In these so-called data driven methods, thousands to millions of materials are evaluated within the virtual space of the computer to look for `hidden' patterns that can be used to identify potential high performing materials to target for synthesis. Trainees in this research program will receive multi-disciplinary training that will provide them with highly sought after skills, particularly in the data sciences and machine learning. Although the research program focuses developing materials related to green house gas mitigation, the novel methods proposed are expected to be more broadly relevant to other classes of materials and other applications.