AdsNet: A Graph Neural Network for Direct Adsorption Energy Prediction
Organizations: Department of Chemical Engineering, Stanford University, Stanford, California 94305, United States · SUNCAT Center for Interface Science and Catalysis, SLAC National Accelerator Laboratory, Menlo Park, California 94025, United States
Abstract
Fast and accurate prediction of adsorption energies is critical for high-throughput computational catalyst screening, existing methods compromise either on evaluation speed (full-graph machine learning potentials) or geometric sensitivity (coordination-based models like the alpha-scheme). Here, we present AdsNet, an equivariant graph neural network that directly predicts adsorption energies in a single forward pass, which operates on a local graph restricted only to adsorbate atoms and its immediate surface neighbors within a cutoff radius. This local graph formulation decouples computational complexity from total system size while encoding continuous geometric features (interatomic distances and directional vectors) enabling sensitivity to subtle variations in local bonding environments. AdsNet adopts a message-passing architecture with scalar and vector node features over multiple interaction layers. Trained on AdsTraj, a dataset of 33,920 density functional theory (DFT) calculated structures spanning 26 transition metals and 6 adsorbates (H*, C*, O*, CO*, OH*, HOCO*), the model achieves an adsorption energy root-mean-square error (RMSE) of 0.032 eV and a force RMSE of 0.047 eV/A. Analytic forces obtained via automatic differentiation enable structure relaxation through a projected gradient descent algorithm with adaptive line search, eliminating entire geometry optimizations. Additionally, AdsNet uniquely decomposes predicted adsorption energies into contributions from individual atoms within the local adsorbate graph - an interpretability feature unattainable with conventional machine-learning potentials. AdsNet thus establishes an accurate, highly efficient, and interpretable framework for large-scale adsorption energy prediction and local interface relaxation across various material surfaces.
Figures & tables
| Metal | Structure | Facet | H* | C* | O* | CO* | OH* | HOCO* | Total |
|---|---|---|---|---|---|---|---|---|---|
| Sc | hcp | (0001) | 84 | 164 | 143 | 225 | 213 | 663 | 1492 |
| Ti | hcp | (0001) | 78 | 83 | 51 | 313 | 237 | 469 | 1231 |
| V | bcc | (110) | 51 | 76 | 92 | 212 | 169 | 528 | 1128 |
| Cr | bcc | (110) | 43 | 67 | 68 | 224 | 182 | 526 | 1110 |
| Mn | fcc | (111) | 99 | 64 | 96 | 229 | 221 | 576 | 1285 |
| Fe | bcc | (110) | 96 | 110 | 48 | 244 | 194 | 672 | 1364 |
Appendix figures & tables2 assets
Supplementary material from the paper’s appendix.
Appendix
| Train | Test | |||||
|---|---|---|---|---|---|---|
| Metal | MAE | RMSE | MAE | RMSE | ||
| Sc | 1355 | 17.4 | 23.8 | 137 | 19.4 | 26.5 |
| Ti | 1103 | 23.4 | 30.5 | 128 | 25.3 | 33.4 |
| V | 1004 | 19.0 | 31.0 | 124 | 19.1 | 30.0 |
| Cr | 1001 | 19.6 | 33.3 | 109 | 18.4 | 32.5 |
| Mn | 1162 | 15.4 | 24.0 | 123 | 18.1 | 29.5 |
| Train | Test | |||||
|---|---|---|---|---|---|---|
| Adsorbate | MAE | RMSE | MAE | RMSE | ||
| H* | 1798 | 19.8 | 31.1 | 186 | 22.2 | 36.6 |
| C* | 2368 | 25.0 | 35.1 | 259 | 31.2 | 50.2 |
| O* | 2174 | 27.1 | 39.5 | 233 | 30.1 | 44.6 |
| CO* | 6383 | 18.5 | 27.8 | 661 | 19.0 | 29.3 |
| OH* | 4435 | 17.5 | 27.0 | 526 | 18.0 | 27.1 |
Materials Property Prediction
Materials property prediction uses data and physical models to estimate how composition, structure, and processing determine a material’s behavior.
See all 204 papers