Basis set incompleteness errors in fixed-node diffusion Monte Carlo calculations on non-covalent interactions
Kousuke Nakano, Benjamin X. Shi, Dario Alfè, Andrea Zen · 2024
Read Basis set incompleteness errors in fixed-node diffusion Monte Carlo calculations on non-covalent interactions by Kousuke Nakano, Benjamin X. Shi, Dario Alfè, Andrea Zen (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2412-00368.txt.
Introduction
Kousuke Nakano and co-authors measure a quiet error source in quantum chemistry in this 2024 paper. The method, fixed-node diffusion Monte Carlo, is often thought immune to basis set bias. Basis sets are the limited alphabets used to write electron states, and small ones leave gaps.
Total energies barely move with a small set, the team found. Binding energies of weakly linked molecule pairs are another story: the error can grow large there. They tested a standard set of 24 linked pairs to map the effect. The study warns that weak bonds, common in drug and materials work, need bigger basis sets than assumed.
Catalog record
| CREATOR | Kousuke Nakano, Benjamin X. Shi, Dario Alfè, Andrea Zen |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 43 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2412-00368
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Basis set incompleteness errors in fixed-node diffusion Monte Carlo calculations on non-covalent interactions Kousuke Nakano, Benjamin X. Shi, Dario Alfè, Andrea Zen Abstract Basis set incompleteness error (BSIE) is a common source of error in quantum chemistry (QC) calculations, but it has not been comprehensively studied in fixed-node Diffusion Monte Carlo (FN-DMC) calculations. FN-DMC, being a projection method, is often considered minimally affected by basis set biases. Here, we show that this assumption is not always valid. While the relative error introduced by a small basis set in the total FN-DMC energy is minor, it can become significant in binding energy ( E b E_{\rm b} ) evaluations of weakly interacting systems. We systematically investigated BSIEs in FN-DMC-based binding energy ( E b E_{\rm b} ) evaluations using the A24 dataset, a well-known benchmark set of 24 non-covalently bound dimers. Contrary to common expectations, we found that BSIEs in FN-DMC evaluations of E b E_{\rm b} are indeed significant when small localized basis sets, such as cc-pVDZ, are employed. We observed that BSIEs are larger in dimers with hydrogen-bonding interactions and smaller in dispersion-dominated interactions. We also found that augmenting the basis sets with diffuse orbitals, using counterpoise (CP) correction, or both, effectively mitigates BSIEs. † † email: [email protected] † † affiliation: Center for Basic Research on Materials, National Institute for Materials Science (NIMS), Tsukuba, Ibaraki 305-0047, Japan † † affiliation: Yusuf Hamied Department of Chemistry, University of Cambridge, Cambridge CB2 1EW, United Kingdom † † affiliation: Dipartimento di Fisica Ettore Pancini, Università di Napoli Federico II, Monte S. Angelo, I-80126 Napoli, Italy † † affiliation: Department of Earth Sciences, University College London, Gower Street, London WC1E 6BT, United Kingdom † † affiliation: Thomas Young Centre and London Centre for Nanotechnology, 17-19 Gordon Street, London WC1H 0AH, United Kingdom † † email: [email protected] † † affiliation: Dipartimento di Fisica Ettore Pancini, Università di Napoli Federico II, Monte S. Angelo, I-80126 Napoli, Italy † † affiliation: Department of Earth Sciences, University College London, Gower Street, London WC1E 6BT, United Kingdom 1 Introduction Diffusion Monte Carlo (DMC) Ceperley (1986) ; Foulkes et al. (2001) is a state-of-the-art electronic structure method used for predicting and understanding phenomena in materials science, chemistry, and physics. In particular, DMC can achieve highly accurate quantitative predictions, typically surpassing those of mean-field approaches like density functional theory (DFT). This level of accuracy has proven essential for studying systems challenging for DFT, such as high-pressure hydrogen Drummond et al. (2015) ; Mazzola et al. (2018) ; Tirelli et al. (2022) ; Ly and Ceperley (2022) ; Niu et al. (2023) ; Monacelli et al. (2023) ; Tenti et al. (2024) , layered materials Krogel et al. (2020) ; Ichibha et al. (2021) ; Nikaido et al. (2022) ; Wines et al. (2023) ; Wines et al. (2023) , molecular crystals Zen et al. (2018) ; Della Pia et al. (2024) , and molecular adsorption on surfaces Beaudet et al. (2008) ; Zen et al. (2016) ; Al-Hamdani et al. (2017) ; Hsing et al. (2019) ; Shi et al. (2023) . In theory, DMC is an exact technique to project the ground state (GS) of a Hamiltonian. However, in practical applications to fermionic systems (e.g., atoms, molecules, and materials), it relies on the fixed-node (FN) approximation to maintain the antisymmetry of the wavefunction. The FN approximation constrains the nodal surface of the projected state to that of a trial wavefunction, which can be generated by methods such as DFT, Hartree-Fock (HF), or correlated quantum chemistry (QC) methods, including the complete active space self consistent field (CASSCF) method. The approaches used to generate the trial wavefunction are not exact, so its nodal surface is not exact either, yielding an error on the FN-DMC evaluations called the FN error. The closer the nodal surface of the trial wavefunction to the nodal surface of the exact GS, the smaller the FN error. There are other approximations in FN-DMC, but typically the major source of error is the FN error. The FN error depends on the accuracy of the trial wavefunction, which in turn depends on the level of theory employed to generate it (e.g., we expect a CASSCF wavefunction to have a better nodal surface than a DFT or an HF wavefunction) and on the completeness of the employed basis set representation (the larger the basis set, the better is the wavefunction and usually its nodal surface). While previous work has explored the influence of trial wavefunction accuracy, Korth et al. (2011) ; Nemec et al. (2010) ; Petruzielo et al. (2012) ; Zen et al. (2013) ; Scemama et al. (2016) ; Caffarel et al. (2016) ; Scemama et al. (2020) the impact of the basis set incompleteness errors (BSIEs) in FN-DMC has yet to be comprehensively and systematically explored in the context of non-covalent interaction evaluations, which is one of the most prominent applications of FN-DMC Korth et al. (2008) ; Nemec et al. (2010) ; Dubecky et al. (2013) ; Dubeckỳ et al. (2014) ; Mostaani et al. (2015) ; Nakano et al. (2019) ; Zen et al. (2018) ; Benali et al. (2020) ; Al-Hamdani et al. (2021) ; Krogel et al. (2020) ; Ichibha et al. (2021) ; Nikaido et al. (2022) ; Wines et al. (2023) ; Wines et al. (2023) ; Raghav et al. (2023) ; Shi et al. (2023) . In QC and DFT methods, BSIEs are a dominant error source that requires careful control, yet it has been often assumed that FN-DMC is relatively immune to BSIEs from the trial wavefunction Dubecký et al. (2016) because it depends only on the nodal surface, not on the full wavefunction amplitude. In this work, we systematically investigate how these assumptions hold up by analyzing BSIEs in FN-DMC calculations. BSIEs are especially pronounced in QC and DFT methods when describing non-covalent interactions. In this context, the quantity of interest is typically the binding energy of a dimer complex (AB), defined as: {E_{\rm b}}={E^{\textrm{AB}}}-E^{\rm{A}}-E^{\rm{B}}, (1) where E^{\rm{A}} , E^{\rm{B}} , and {E^{\rm{AB}}} are the total energies of monomer A, monomer B, and the AB dimer complex, respectively. This study focuses on the propagation of BSIEs from the trial wavefunction in FN-DMC calculations of E_{\rm b} , a particularly relevant area of investigation given the high sensitivity of non-covalent interactions to basis set quality Al-Hamdani et al. (2021) ; Nakano et al. (2024) ; Schäfer et al. (2024) . A basis set consists of a number of basis functions that are used to represent the electronic wave function, with the complete basis set (CBS) limit achieved when expanded towards an (infinite) set of functions. The BSIE is the deviation from the CBS limit Van Duijneveldt et al. (1994) ; Dunning (2000) and for a binding energy E_{\textrm{b}} , it is defined as Van Duijneveldt et al. (1994) ; Dunning (2000) : E_{\rm b}^{\rm BSIE}(M^{\textrm{A}},M^{\textrm{B}},M^{\textrm{AB}})={{E_{\rm b}}(M^{\textrm{A}},M^{\textrm{B}},M^{\textrm{AB}})}-E_{\rm b}^{\rm CBS}, (2) where M^{\textrm{A}} , M^{\textrm{B}} and M^{\textrm{AB}} denote the number and type of basis functions employed in the calculation of E^{\textrm{A}} , E^{\textrm{B}} and E^{\textrm{AB}} respectively within Eq. 1 , and E_{\rm b}^{\rm CBS} denotes the binding energy in the CBS limit. Two common choices of basis function types are plane