Advancements in thermochemical predictions: a multi-output thermodynamics-informed neural network approach
Authors:
Open Access
- 16 June 2025
- journal article
- research article
- Published by Springer Nature in Journal of Cheminformatics
- Vol. 17 (1) , 1-8
Abstract
The Gibbs free energy of an inorganic material represents its maximum reversible work potential under constant temperature and pressure. Its calculation is crucial for understanding material stability, phase transitions, and chemical reactions, thus guiding optimization for diverse applications like catalysis and energy storage. In this study, we have developed a Physics-Informed Neural Network model that leverages the Gibbs free energy equation. The overall loss function is adjusted to allow the model to simultaneously predict all three thermodynamic quantities, including Gibbs free energy, total energy, and entropy, thus transforming it into a multi-output model. In recent literature, there is a growing emphasis on evaluating machine learning models under challenging conditions, such as small datasets and out-of-distribution predictions. Reflecting this trend, we have rigorously benchmarked our model across these scenarios, demonstrating its robustness and adaptability. It turns out that our model demonstrates a 43% improvement for normal scenario and even more in out-of-distribution regime compared to the next-best model. Scientific Contribution This study introduces the application of a Physics-Informed Neural Network to simultaneously compute multiple thermodynamic properties, including Gibbs free energy, total energy, and entropy. By integrating the Gibbs free energy equation into the loss function, the model achieves superior accuracy in low data regimes and enhances robustness in the out-of-distribution scenarios.Keywords
All Related Versions
Funding Information
- Tata Institute of Fundamental Research
This publication has 36 references indexed in Scilit:
- First-principles thermodynamics of precipitation in aluminum-containing refractory alloysActa Materialia, 2024
- Global reaction neural networks with embedded stoichiometry and thermodynamics for learning kinetics from reactor dataChemical Engineering Journal, 2024
- Towards physics-informed explainable machine learning and causal models for materials researchComputational Materials Science, 2023
- A critical review on transition metal phosphide based catalyst for electrochemical hydrogen evolution reaction: Gibbs free energy, composition, stability, and true identity of active siteCoordination Chemistry Reviews, 2022
- A Neural Network Approach to Predict Gibbs Free Energy of Ternary Solid SolutionsJournal of Phase Equilibria and Diffusion, 2022
- Data‐Driven Materials Innovation and ApplicationsAdvanced Materials, 2022
- Computational thermodynamics and its applicationsActa Materialia, 2020
- A First‐Principles‐Based Sub‐Lattice Formalism for Predicting Off‐Stoichiometry in Materials for Solar Thermochemical Applications: The Example of CeriaAdvanced Theory and Simulations, 2020
- FactSage thermochemical software and databases, 2010–2016Calphad, 2016
- First principles phonon calculations in materials scienceScripta Materialia, 2015