The design of materials for increased turbine operating temperature is a critical goal from energy to aerospace applications. Proper material design includes protective native oxide scales or applied coatings with high-melting temperatures, thermodynamic stability, and low ionic diffusivity. To expedite the fabrication process the above properties would be known for all oxides from experiments or first principles calculations with quantified uncertainties. While some properties of interest are known for many oxides (e.g. elastic constants exist for over 1,000 oxides), melting temperature is known for a relatively small subset. Furthermore, melting temperature measurements can be computationally or physically prohibitive, and the limited amount of centralized data precludes the use of standard machine learning models. To address this gap, we use a multi-step approach based on sequential learning where information from data-driven models is leveraged on smaller datasets. We extend our models to nearly 11,000 oxides, and quantify uncertainties in the space.

This effort was supported by the US National Science Foundation DMREF program under contract number 1922316-DMR.

Citrine Informatics, Citrination Database, citrination.com, (2020).

S. P. Ong, S. Cholia, A. Jain, M. Brafman, D. Gunter, G. Ceder, and K. A. Persson, Computational Materials Science 97, 209 (2015).

Wolfram Research, Inc., Wolfram|Alpha Notebook Edition, Champaign, IL (2019).

J. Ling, M. Hutchinson, E. Antono, S. Paradiso, and B. Meredig, Integrating Materials and Manufacturing Innovation 6, 207 (2017).

M. Hutchinson, Citrine informatics lolo, (2016).

Researchers should cite this work as follows:

• Zachary D McClure, Alejandro Strachan (2023), "High Temperature Oxide Property Explorer," https://nanohub.org/resources/htoxideprop. (DOI: 10.21981/6EVM-0137).

  BibTex | EndNote

1. machine learning
2. materials databases
3. oxides