Chemistry & ADMET · isayevlab
AIMNet2-Pd
A neural network interatomic potential predicting energies, forces, and charges for palladium-containing systems.
What AIMNet2-Pd is
AIMNet2-Pd is a neural network interatomic potential developed by the Isayev Lab, designed for molecular simulations involving palladium-containing organometallic compounds. It targets homogeneous catalysis and coordination chemistry, providing near-DFT accuracy at a fraction of the computational expense required by conventional electronic structure methods.
By predicting molecular properties orders of magnitude faster than quantum mechanical calculations, the model enables researchers to perform extended molecular dynamics, structural optimisations, and conformational sampling on transition metal complexes that would otherwise be computationally prohibitive.
How it works
The model builds on the AIMNet2 architecture, which combines Atomic Environment Vectors with message-passing multilayer perceptrons to model multi-body atomic interactions within a 5.0 Å cutoff radius. It operates without embedded short-range Coulomb interactions, while D3BJ dispersion corrections are handled externally by the calculator.
Trained to reproduce the wB97M-D3 density functional level of theory, the system evaluates potential energy, atomic forces, partial atomic charges, Hessians, and stress tensors. The distribution includes four ensemble members to assist with prediction uncertainty quantification.
Key facts
- Architecture
- AIMNet2 (Atomic Environment Vectors with message-passing MLPs)
- Cutoff radius
- 5.0 Å
- Reference functional
- wB97M-D3 with D3BJ dispersion
- Supported elements
- H, B, C, N, O, F, Si, P, S, Cl, Se, Br, Pd, I
- Ensemble size
- 4 members
- Primary outputs
- Energy, forces, partial charges, Hessian, stress tensor
What you can use it for
- Homogeneous Catalysis Modelling
- Evaluating reaction pathways, intermediates, and transition states in organopalladium-catalysed cross-coupling and related catalytic cycles.
- Organometallic Molecular Dynamics
- Executing rapid dynamic simulations and conformational searches of flexible palladium coordination complexes in gas or periodic environments.
- Uncertainty-Aware Property Calculation
- Computing energies, atomic forces, and partial charges across the four ensemble members to quantify predictive uncertainty in novel chemical spaces.
Run it here
No installation and no GPU of your own: fill in the inputs and it runs on our servers. Results stay in your browser unless you sign in.
Runs on On-demand GPU
Runs on our servers with on-demand compute. A first run needs time to load the model; active capacity can be reused and scales down when idle.Run this model
AIMNet2-Pd
Strengths and limits
Good at
- Delivers near-DFT accuracy for energies, atomic forces, and atomic partial charges at orders of magnitude lower computational cost.
- Includes four distinct ensemble members to support straightforward uncertainty quantification.
- Supports periodic boundary conditions, stress tensors, and Hessian calculations for bulk or condensed-phase workflows.
- Integrates with the Atomic Simulation Environment (ASE) for straightforward molecular dynamics and structure relaxation.
Watch out for
- Element coverage is restricted strictly to 14 elements (H, B, C, N, O, F, Si, P, S, Cl, Se, Br, Pd, I), and the presence of any other element will trigger an error.
- Arsenic (As) is not supported and requires alternative models such as the general aimnet2-wb97m-d3.
Inputs and outputs
You provide
Molecule (SMILES)
We generate and relax a 3D conformer before the model sees it.
You get back
Energy, maximum force and per-atom partial charges.
Licence
Check the licence before commercial use · See model card
The See model card licence carries terms worth reading before you depend on it.
Questions
Can I use AIMNet2-Pd commercially?
Check the licence before commercial use. The See model card licence carries terms worth reading before you depend on it.
How do I run AIMNet2-Pd without installing anything?
Use the form under “Run it here” on this page: fill in the inputs and press Run model. It runs on our servers on CPU only, typically in 12s, and a run costs about 28 credits. Visitors without an account get 1,000 credits a month.
Which elements does AIMNet2-Pd support?
It supports 14 elements: hydrogen, boron, carbon, nitrogen, oxygen, fluorine, silicon, phosphorus, sulfur, chlorine, selenium, bromine, palladium, and iodine.
What reference quantum chemistry method does the model replicate?
The model is parameterised against the wB97M-D3 density functional, with D3BJ dispersion corrections managed externally.
How is prediction uncertainty calculated?
Uncertainty can be estimated by running predictions across all four provided ensemble parameter sets and assessing the variance between their outputs.
Can the model evaluate periodic systems?
Yes, periodic boundary conditions and stress tensors are supported through the Python API and the ASE interface.
How to cite it
Anstine, Dylan and Zubatyuk, Roman and Isayev, Olexandr. AIMNet2: A Neural Network Potential to Meet your Neutral, Charged, Organic, and Elemental-Organic Needs. Chemical Science, 2025. doi:10.1039/D4SC08572H
Written from AIMNet2-Pd's own documentation (model card). Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.