Prediction of seebeck coefficient for compounds without restriction to fixed stoichiometry: A machine learning approach

• Published in: Journal of computational chemistry 2018-02, Vol.39 (4), p.191-202
• Main Authors: Furmanchuk, Al'ona, Saal, James E., Doak, Jeff W., Olson, Gregory B., Choudhary, Alok, Agrawal, Ankit
• Format: Article
• Language: English
• Subjects: Coefficients; Crystal structure; Crystallinity; data mining; Machine learning; nonstoichiometric materials; prediction; Predictions; Regression models; Seebeck coefficient; Stoichiometry; Thermoelectric materials; thermoelectric properties; web application
• ISSN: 0192-8651; 1096-987X
• DOI: 10.1002/jcc.25067

The regression model‐based tool is developed for predicting the Seebeck coefficient of crystalline materials in the temperature range from 300 K to 1000 K. The tool accounts for the single crystal versus polycrystalline nature of the compound, the production method, and properties of the constituent elements in the chemical formula. We introduce new descriptive features of crystalline materials relevant for the prediction the Seebeck coefficient. To address off‐stoichiometry in materials, the predictive tool is trained on a mix of stoichiometric and nonstoichiometric materials. The tool is implemented into a web application (http://info.eecs.northwestern.edu/SeebeckCoefficientPredictor) to assist field scientists in the discovery of novel thermoelectric materials. © 2017 Wiley Periodicals, Inc. Machine learning is employed for prediction of the Seebeck coefficient of crystalline materials in the temperature range from 300 K to 1000 K. To address off‐stoichiometry in experimental samples, new descriptive features are introduced. The tool is implemented into a web application (http://info.eecs.northwestern.edu/SeebeckCoefficientPredictor) to assist field scientists in the discovery of novel thermoelectric materials.