We are aware of an issue with certificate availability and are working diligently with the vendor to resolve. The vendor has indicated that, while users are unable to directly access their certificates, results are still being stored. Certificates will be available once the issue is resolved. Thank you for your patience.

(706a) Mining and Screening of Ab Initio Data for Catalyst Design

Botu, V., University of Connecticut
Ramprasad, R., University of Connecticut
Suib, S., University of Connecticut

Dopants have often been used to optimize, enhance, or tailor the behavior of a parent material in many situations ranging from material strengthening to electronics to electrochemistry to catalysis. The search and identification of suitable dopant candidates has been laborious though, and dominated either by lengthy trial-and-error strategies (guided by intuition) or plain serendipity. As we enter the data-driven era, such traditional Edisonian approaches are being augmented (and sometimes, replaced) by rational strategies based on advanced computational screening. In this contribution, we offer such a paradigm for the selection of suitable dopants in metal oxides for catalytic reactions involving oxygen. As an example, we explore dopants across the periodic table for enhanced thermochemical splitting of water on doped ceria surfaces. Using ab initio methods the oxygen vacancy formation energy was identified as the suitable descriptor, and used to develop screening criteria based on Sabatier’s principle. Sc, Cr, Y, Zr, Pd and La were identified as the best candidates, and in good agreement with past experimental data. However, owing to slow predictive capability of ab initio methods, machine-learning methods (specifically feature selection techniques) were employed to mine and discover patterns amongst the spectrum of dopants, to quickly classify a dopants impact on activity. By using methods such as principal component analysis and random forest methods, we identified the key properties of dopants that make them most attractive. A dopant’s oxidation state, ionic radius and electron affinity were found to correlate strongly with enhanced activity. Thus, by combining the power of ab initio with speed of machine learning methods, we can therefore quickly pre-screen dopants simply based on their properties, and further expand the dopant chemical space by searching for those elements with the desired properties. Though, the framework was developed for thermochemical dissociation of water, we believe it can be applied to several other processes, e.g. CO oxidation, chemical looping and solid oxide fuel cells, etc.