Performant automatic differentiation of local coupled cluster theories: Response properties and ab initio molecular dynamics

Abstract

In this work, we introduce a differentiable implementation of the local natural orbital coupled cluster (LNO-CC) method within the automatic differentiation framework of the PySCFAD package. The implementation is comprehensively tuned for enhanced performance, which enables the calculation of first-order static response properties on medium-sized molecular systems using coupled cluster theory with single, double, and perturbative triple excitations [CCSD(T)]. We evaluate the accuracy of our method by benchmarking it against the canonical CCSD(T) reference for nuclear gradients, dipole moments, and geometry optimizations. In addition, we demonstrate the possibility of property calculations for chemically interesting systems through the computation of bond orders and Mössbauer spectroscopy parameters for a [NiFe]-hydrogenase active site model, along with the simulation of infrared spectra via ab initio LNO-CC molecular dynamics for a protonated water hexamer.

Copyright and License

Acknowledgement

This work was primarily supported by the United States Department of Energy, Office of Science, Basic Energy Sciences, Chemical Sciences, Geosciences, and Biosciences Division, FWP LANLE3F2 awarded to Los Alamos National Laboratory under Triad National Security, LLC (“Triad”) contract Grant No. 89233218CNA000001, subaward C2448 to the California Institute of Technology. Additional support for G.K.C. was provided by the Camille and Henry Dreyfus Foundation via a grant from the program “Machine Learning in the Chemical Sciences and Engineering.” Some of the calculations were performed at the National Energy Research Scientific Computing Center (NERSC), a U.S. Department of Energy, Office of Science User Facility located at the Lawrence Berkeley National Laboratory.

Contributions

X.Z. and C.L. contributed equally to this work. X.Z., C.L., and G.K.-L.C. designed this project. H.-Z.Y. and T.C.B. contributed to the original implementation of the LNO-CC method, based on which X.Z. developed the differentiable version used here. H.-Z.Y. and T.C.B. also provided help and advice with the methodology development in the early part of the project. C.L. developed the AIMD simulation workflow. X.Z. and C.L. performed the calculations and wrote the original draft. X.Z., C.L., and G.K.-L.C. validated the data, and all contributed to the reviewing and editing of the manuscript.

Xing Zhang: Conceptualization (equal); Methodology (equal); Software (equal); Validation (equal); Writing – original draft (equal); Writing – review & editing (equal). Chenghan Li: Conceptualization (equal); Methodology (equal); Software (equal); Validation (equal); Writing – original draft (equal); Writing – review & editing (equal). Hong-Zhou Ye: Methodology (supporting); Software (equal); Writing – review & editing (equal). Timothy C. Berkelbach: Methodology (supporting); Writing – review & editing (equal). Garnet Kin-Lic Chan: Conceptualization (equal); Funding acquisition (lead); Supervision (lead); Writing – review & editing (equal).

Data Availability

The data that supports the findings of this study are available within the article, and/or from the corresponding author upon reasonable request.

Additional implementation details and computational results supporting the findings in this work are provided in the supplementary material.

Code Availability

The PySCFAD source code can be found at https://github.com/fishjojo/pyscfad.

Files

Embargoed

The files will be made publicly available on July 7, 2025.

	All versions	This version
Views	0	0
Downloads	0	0
Data volume	0 Bytes	0 Bytes