Date/Time Date(s) - 27/02/20262:30 pm - 3:30 pm
Categories
Prof. Holger Kleinke
Department of Chemistry, University of Waterloo
Artificial intelligence (AI) studies to predict or suggest efficient thermoelectric materials have become increasingly important. Previous machine learning (ML) studies have used different literature sources or density functional theory calculations as input. As quality and consistency of data play a crucial role in the success of the model’s outputs, we originally developed an ML model using data exclusively from our own lab experiments with various doped SnSe materials to predict their total thermal conductivity.[1]
Subsequently, we developed a machine learning pipeline trained inputs on a massive public dataset to predict total thermal conductivity as well as Seebeck coefficient using four different test sets: three publicly available datasets and one dataset created using all our own previous results.[2,3] With the aid of feature selection and importance analysis, useful chemical features were chosen that ultimately led to higher accuracy in the test sets. Finally, we used a similar procedure to develop a model for the prediction of the thermoelectric figure-of-merit, which brings us into the exciting situation to predict the performance simply based on the chemical formula.[4]
[1] N. K. Barua, A. Golabek, A. O. Oliynyk, H. Kleinke, J. Mater. Chem. C 2023, 11, 11643–11652.
[2] N. K. Barua, E. Hall, Y. Cheng, A. O. Oliynyk, H. Kleinke, Chem. Mater. 2024, 36, 7089–7100.
[3] N. K. Barua, H. Kleinke, ACS Appl. Energy Mater. 2025, 8, 16110–16121.
[4] N. K. Barua, S. Lee, A. O. Oliynyk, H. Kleinke, ACS Appl. Mater. Interfaces 2025, 17, 1662–1673.
Brief Professional History
Prof. Kleinke, Full Professor at the University of Waterloo since 2006, obtained his PhD at the University of Mainz in Germany in 1994, before his postdoctoral stint at the Ames Laboratory, US Department of Energy, in Iowa from 1995 to 1997. Thereafter he moved to the University of Marburg in Germany to begin investigating thermoelectric antimonides for his Habilitation from 1997 to 2001, which overlapped with his career at the University of Waterloo since the year 2000. There, Prof. Kleinke worked as Tier II Canada Research Chair from 2002 to 2011, and served as Interim Executive Director of the Waterloo Institute for Nanotechnology from 2016 to 2017 and as Director of the Nanotechnology Graduate Program from 2018 to 2019. Since 2018, he contributes to the community as Associate Editor for the journal ACS Applied Materials & Interfaces. Prof. Kleinke is also a member of the Waterloo Data and Artificial Intelligence Institute, the Waterloo Institute for Nanotechnology, and the Waterloo Institute for Sustainable Energy.
Research Focus & Background
Prof. Kleinke has over 25 years of experience of working on materials for thermoelectric refrigeration and power generation applications, with well over 200 peer-reviewed journal articles on that topic. His primary research expertise lies in exploratory chemistry to identify and characterize new thermoelectric materials. He continues to investigate a variety of different materials, from silicides to antimonides, to sulfides, selenides and tellurides as well as half-Heusler materials, including nanostructured materials and nanocomposites. Since several years now, he is using artificial intelligence employing different machine learning algorithms to successfully predict thermoelectric properties of hitherto unexplored materials.