Counterpoise-Corrected Ab Initio Potential Energy Curves and Analytic Pair Potentials for Ag, Au, and Ni Interacting with He and Ar: A CCSD(T) Benchmark with Machine-Learning Comparison
Abstract
Accurate pairwise potentials for coinage-and transition-metal atoms interacting with noble gases are required for molecular dynamics simulations of metal clusters and matrix-isolation environments, yet such potentials are rarely benchmarked with counterpoise-corrected coupledcluster theory. In this work, potential energy curves of Ag, Au, and Ni interacting with He and Ar were computed at the CCSD(T) level using the aug-cc-pVTZ and aug-cc-pVTZ-PP basis sets, with full counterpoise correction and a wavefunction-stability-verified, orbitalcontinuity-chained self-consistent-field protocol. An initial dataset computed with Gaussian 16 was found to contain a spin-multiplicity error for nickel and uncorrected electronic-state discontinuities and was discarded in favor of a complete recomputation with ORCA 6.0.1. Basis-set superposition error was found to account for 59–166% of the well depth even at the aug-cc-pVTZ(-PP) level, and a genuine avoided electronic curve crossing was identified for the Ni–Ar pair near 2.6–3.0 Å, corroborated by both an energy discontinuity and a T1-diagnostic spike. The curves were fitted to the Lennard-Jones, Morse, Buckingham, and Improved Lennard-Jones (ILJ) forms, and to a Gaussian process regression model. The Improved Lennard-Jones and Morse forms were found most reliable classically, while the Lennard-Jones form was poorly constrained for every pair. The Gaussian process regression model outperformed every classical form on every pair. Well depths were found to follow the trend Au > Ag > Ni for both gases, and Ar wells were 11–16 times deeper than He wells, consistent with the larger polarizability of argon.
Related
- Engineering entanglement and transport in interacting quantum walks with tailored potentials
- Diagnosing Latent Energy Decomposition in Machine-Learning Interatomic Potentials via Interacting Quantum Atoms
- Mass-Scaling of Quantum Tunnelling in Hydrogen Bonds: Analytical Model and Comparison with Multidimensional Potentials
- Reconstructing sign-switching dark energy histories: Scalar-field regularity, conditional potential comparison, and representative dynamics
- Encrypted cloning of qubits is able to serve as a quantum primitive, already on non-error-corrected quantum hardware
BibTeX
@misc{mohammadi2026counterpoisecorrected,
title = {{Counterpoise-Corrected Ab Initio Potential Energy Curves and Analytic Pair Potentials for Ag, Au, and Ni Interacting with He and Ar: A CCSD(T) Benchmark with Machine-Learning Comparison}},
author = {Mohsen Doust Mohammadi},
howpublished = {ChemRxiv},
year = {2026},
month = aug,
doi = {10.26434/chemrxiv.15008051/v1},
url = {https://chemrxiv.org/doi/full/10.26434/chemrxiv.15008051/v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - Counterpoise-Corrected Ab Initio Potential Energy Curves and Analytic Pair Potentials for Ag, Au, and Ni Interacting with He and Ar: A CCSD(T) Benchmark with Machine-Learning Comparison
AU - Mohsen Doust Mohammadi
PY - 2026
DA - 2026/08/30
PB - ChemRxiv
DO - 10.26434/chemrxiv.15008051/v1
UR - https://chemrxiv.org/doi/full/10.26434/chemrxiv.15008051/v1
AB - Accurate pairwise potentials for coinage-and transition-metal atoms interacting with noble gases are required for molecular dynamics simulations of metal clusters and matrix-isolation environments, yet such potentials are rarely benchmarked with counterpoise-corrected coupledcluster theory. In this work, potential energy curves of Ag, Au, and Ni interacting with He and Ar were computed at the CCSD(T) level using the aug-cc-pVTZ and aug-cc-pVTZ-PP basis sets, with full counterpoise correction and a wavefunction-stability-verified, orbitalcontinuity-chained self-consistent-field protocol. An initial dataset computed with Gaussian 16 was found to contain a spin-multiplicity error for nickel and uncorrected electronic-state discontinuities and was discarded in favor of a complete recomputation with ORCA 6.0.1. Basis-set superposition error was found to account for 59–166% of the well depth even at the aug-cc-pVTZ(-PP) level, and a genuine avoided electronic curve crossing was identified for the Ni–Ar pair near 2.6–3.0 Å, corroborated by both an energy discontinuity and a T1-diagnostic spike. The curves were fitted to the Lennard-Jones, Morse, Buckingham, and Improved Lennard-Jones (ILJ) forms, and to a Gaussian process regression model. The Improved Lennard-Jones and Morse forms were found most reliable classically, while the Lennard-Jones form was poorly constrained for every pair. The Gaussian process regression model outperformed every classical form on every pair. Well depths were found to follow the trend Au > Ag > Ni for both gases, and Ar wells were 11–16 times deeper than He wells, consistent with the larger polarizability of argon.
ER -