Condensed Matter > Materials Science
[Submitted on 1 Aug 2024 (v1), revised 23 Aug 2024 (this version, v2), latest version 12 Sep 2024 (v4)]
Title:Thermal Conductivity Predictions with Foundation Atomistic Models
View PDF HTML (experimental)Abstract:Recent advances in machine learning have led to foundation models for atomistic materials chemistry, enabling quantum-accurate descriptions of interatomic forces across diverse materials at reduced computational cost. These models are currently benchmarked on harmonic phonons and related quantities, leaving their reliability for complex, technologically relevant anharmonic heat-conduction properties uncertain. Here, we introduce a framework that leverages foundation models and the Wigner formulation of heat transport to automatically predict anharmonic vibrational properties and thermal conductivity in solids with arbitrary compositions and structures. We employ this framework with the foundation models M3GNet, CHGNet, MACE-MP-0, and SevenNET to analyze 103 diverse compounds, comparing predictions against first-principles references. Thus, we introduce a benchmark metric based on conductivity, and discuss data-efficient fine-tuning protocols to achieve first-principles accuracy. Our framework enables physics-aware, accurate predictions of vibrational and thermal properties, paving the way to discover materials that violate semiclassical Boltzmann transport and exhibit record-high heat-shielding or thermoelectric performance.
Submission history
From: Michele Simoncelli [view email][v1] Thu, 1 Aug 2024 17:57:24 UTC (1,371 KB)
[v2] Fri, 23 Aug 2024 17:58:01 UTC (3,057 KB)
[v3] Thu, 29 Aug 2024 16:32:58 UTC (3,130 KB)
[v4] Thu, 12 Sep 2024 05:13:52 UTC (3,218 KB)
Current browse context:
cond-mat.mtrl-sci
Change to browse by:
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
IArxiv Recommender
(What is IArxiv?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.