|
What if we could compute a ‘Materials Genome' -- the calculated properties of all known, reported materials whose structures have been discovered? While the number of atom combinations is almost limitless -- just like the permutations of DNA patterns -- there is a great deal we can learn about materials, new and old, by looking carefully at the properties and trends existing in such a vast database. This knowledge could then be used in the design and screening of advanced materials in a multitude of applications.
My project deals with the development of such a large-scale database towards the discovery of a new high-rate, high energy density lithium ion battery cathode. We are running ab initio density functional theory calculations in a high-throughput mode, which has already generated data on tens of thousands of materials. Using this information, we are currently able to evaluate many key aspects of a potential battery cathode, such as its energy density and thermodynamic stability, and are able to screen thousands of battery cathode candidates for potentially interesting properties.
In the future, we expect to go beyond data generation and screening to materials property prediction using statistical analysis and machine learning. As one example, we will data mine the existing voltages in our database to predict new materials that are likely to be in our desired voltage range. Therefore, we can start to solve the inverse problem of "what materials will give us property X?" rather than asking "what is property X of this material?". Going further, we can recycle the information in the database for batteries towards the guided search for new materials in many other applications. In this way, we hope to develop a new model of materials discovery in which we largely describe, design, and discover new materials in silico.
Publications:
A. Jain & V. Stojanoff (2007). Are you centered? An automatic crystal-centering method for high-throughput macromolecular crystallography. Journal of Synchrotron Radiation, 14, 4, pp. 355-360.
M. Allaire, A. Berntson, A. Jain et al. (2005). A Modular Approach to Beam Line Automation: The NIGMS Facility at the NSLS. Synchrotron Radiation News, 18, 2, pp. 23-27.
Presentations:
"High-Throughput Density Functional Theory as a Search Tool for Novel Li-ion Battery Cathodes"
Poster at the Materials Research Society Fall Meeting, Boston, MA 2009
"A High-Throughput Computational Search for Novel Li-ion Battery Cathode Materials"
Presentation at the Electrochemical Society Fall Meeting, Vienna, Austria 2009
"Screening of Sorbents for Mercury Capture Using Ab initio Computations"
Presentation at the American Institute of Chemical Engineers Spring Meeting, Tampa, Florida 2009
Cryscent: Automatic crystal-centering for high-throughput macromolecular crystallography. 2007 Annual Meeting of the ACA. [poster presentation]
Are You Centered? Automatic sample alignment at the X6A beamline. National Synchrotron Light Source Lunch Talks, August 2006.
|
