Machine Learning for Novel Thermal-Materials Discovery: Early Successes, Opportunities, and Challenges

Abstract

High-throughput computational and experimental design of materials aided by machine learning have become an increasingly important field in material science. This area of research has emerged in leaps and bounds in the thermal sciences, in part due to the advances in computational and experimental methods in obtaining thermal properties of materials. In this paper, we provide a current overview of some of the recent work and highlight the challenges and opportunities that are ahead of us in this field. In particular, we focus on the use of machine learning and high-throughput methods for screening of thermal conductivity for compounds, composites and alloys as well as interfacial thermal conductance. These new tools have brought about a feedback mechanism for understanding new correlations and identifying new descriptors, speeding up the discovery of novel thermal functional materials.

Nanostructured and Heterostructured 2D Materials for Thermoelectrics

“…For the materials genome, a highthroughput calculation method is required and this can be achieved with machine learning. Successful examples for machine learning in materials search and design can be found for interfacial thermal conductance, [175], [255], [256] bandgap, [257] and interatomic force constants [258], [259], [260] used in MD simulations. Machine learning driven by experimental data is desired for thermoelectric studies but is still lacking due to the challenge of high-throughput measurements at the nanoscale.…”

Autonomous Discovery in the Chemical Sciences Part I: Progress

“…[61,461,464] Thep redominant use of machine learning in computational materials discovery has been to fit surrogate models to existing (often, experimental) data and screen al arge design space. [465][466][467][468][469][470][471] To the extent that performance can be correlated to structure,t hese models can reveal opportunities for the design of new catalysts/ligands for organic synthesis [223,[225][226][227]472] (Figure 11), metallic catalysts, [473,474] Heusler compounds, [475] metal-organic frameworks (MOFs), [476] hybrid organic-inorganic perovskites, [477] superhard materials, [478] thermal materials, [479] organic electronic materials, [480][481][482][483][484] polymers for electronic applications, [485,486] porous crystalline materials for gas storage, [487,488] and reductive additives for battery electrolyte formulations. [42] Computational models have also been used to determine when calculations are likely to fail [489] and to identify associations between materials and specific property keywords through text mining.…”

Autonome Entdeckung in den chemischen Wissenschaften, Teil I: Fortschritt

“…Die vorherrschende Anwendung von maschinellem Lernen bei der rechnergestützten Entdeckung von Materie besteht bislang in der Anpassung von Surrogatmodellen an vorhandene (oft experimentelle) Daten und der Untersuchung von großen Designräumen. [465][466][467][468][469][470][471] In dem Maße,wie die Leistung einer Verbindung mit der Struktur korreliert werden kann, bieten die Modelle neue Designoptionen fürK atalysatoren/Liganden fürd ie organische Synthese [223,[225][226][227]472] (Abbildung 11), Metallkatalysatoren, [473,474] Heusler-Verbindungen, [475] Metall-organische Gerüste (MOFs), [476] hybride organisch-anorganische Perowskite, [477] superharte Werkstoffe, [478] thermische Materialien, [479] organische elektronische Materialien, [480][481][482][483][484] Polymere fürd ie Elektronik, [485,486] porçse kristalline Materialien fürd ie Gasspeicherung [487,488] und reduzierende Additive fürB atterie-Elektrolyt-Formulierungen. [42] Mit rechnergestützten Modellen wurde bestimmt, wann Rechnungen wahrscheinlich versagen [489] und wie sich durch Te xt-Mining Assoziationen zwischen Materialien und bestimmten Eigenschafts-Stichwçrtern finden lassen.…”