Striking the right balance of encoding electron correlation in the Hamiltonian and the wavefunction ansatz

Kalman Szenes, Maximilian Mörchen, Paul Fischill, Markus Reiher

Published Nov 1, 2024• 1 citation

Abstract

Multi-configurational electronic structure theory delivers the most versatile approximations to many-electron wavefunctions, flexible enough to deal with all sorts of transformations, ranging from electronic excitations, to open-shell molecules and chemical reactions. Multi-configurational models are therefore essential to establish universally applicable, predictive ab initio methods for chemistry. Here, we present a discussion of explicit correlation approaches which address the nagging problem of dealing with static and dynamic electron correlation in multi-configurational active-space approaches. We review the latest developments and then point to their key obstacles. Our discussion is supported by new data obtained with tensor network methods. We argue in favor of simple electron-only correlator expressions that may allow one to define transcorrelated models in which the correlator does not bear a dependence on molecular structure.

DOI: 10.3929/ethz-b-000688070

Related papers

Striking the right balance of encoding electron correlation in the Hamiltonian and the wavefunction ansatz

Kalman Szenes, Maximilian Mörchen, Paul Fischill +1 more

methods for chemistry. Here, we present a discussion of explicit correlation approaches which address the nagging problem of dealing with static and dynamic electron correlation in multi-configurational active-space approaches. We review the latest developments and then point to their key obstacles. Our discussion is supported by new data obtained with tensor network methods. We argue in favor of simple electron-only correlator expressions that may allow one to define transcorrelated models in which the correlator does not bear a dependence on molecular structure.

arXiv:2404.19172Faraday DiscussionsJan 1, 20247 citations

New electron correlation theories for transition metal chemistry

Konrad Heinrich Marti, Markus Reiher

Electronic structure theory faces many computational challenges in transition metal chemistry. Usually, density functional theory is the method of choice for theoretical studies on transition metal complexes and clusters mostly because it is the only feasible one, although its results are not systematically improvable. By contrast, multireference ab initio methods could provide a correct description of the electronic structure, but are limited to small molecules because of the tremendous computational resources required. In recent years, conceptually new ab initio methods emerged that turned out to be promising for theoretical coordination chemistry. We review and discuss two efficient parametrization schemes for the electronic wave function, the matrix product states and the complete-graph tensor network states. Their advantages are demonstrated at example transition metal complexes. Especially, tensor network states might provide the key to accurately describe strongly correlated and magnetic molecular systems in transition metal chemistry.

Physical Chemistry Chemical PhysicsJan 1, 2011130 citations

Ab initio Density Matrix Renormalization Group Theory for Multireference Quantum Chemistry

Takeshi Yanai

A major challenge in molecular electronic structure theory arises from chemical systems where the mean-field picture based upon a single electronic configuration fails to properly describe mechanics of many electrons. Such complex electronic states should be accounted for using a correlated superposition of multiple determinants and are referred to as multireference in quantum chemistry. Application domain of traditional multireference methods is severely limited to systems containing a small number of correlated electrons because of their exponential complexity. In recent years, much attention has been drawn to the density matrix renormalization group (DMRG) theory as a promising alternative. It has been introduced in ab initio quantum chemistry calculations and shown to be a highly-scalable multireference method that can handle full quantum degrees of freedom for the strongly-correlated wavefunction with unprecedented large configuration space. In this article, we provide an explicative introduction to the underlying theory and formalism built into the DMRG method and give a brief overview of our methodological extensions that combine DMRG and active-space correlation models.

Molecular ScienceJan 1, 20140 citations

Automated Identification of Relevant Frontier Orbitals for Chemical Compounds and Processes

Christopher Johannes Stein, Markus Reiher

Quantum-chemical multi-configurational methods are required for a proper description of static electron correlation, a phenomenon inherent to the electronic structure of molecules with multiple (near-)degenerate frontier orbitals. Here, we review how a property of these frontier orbitals, namely the entanglement entropy is related to static electron correlation. A subset of orbitals, the so-called active orbital space is an essential ingredient for all multi-configurational methods. We proposed an automated selection of this active orbital space, that would otherwise be a tedious and error prone manual procedure, based on entanglement measures. Here, we extend this scheme to demonstrate its capability for the selection of consistent active spaces for several excited states and along reaction coordinates.

arXiv:1702.00450CHIMIA International Journal for ChemistryApr 26, 201781 citations

Making quantum chemistry compressive and expressive: Toward practical ab‐initio simulation

Jun Yang

Abstract Ab‐initio quantum chemistry simulations are essential for understanding electronic structure of molecules and materials in almost all areas of chemistry. A broad variety of electronic structure theories and implementations has been developed in the past decades to hopefully solve the many‐body Schrödinger equation in an approximate manner on modern computers. In this review, we present recent progress in advancing low‐rank electronic structure methodologies that rely on the wavefunction sparsity and compressibility to select the important subset of electronic configurations for both weakly and strongly correlated molecules. Representative chemistry applications that require the many‐body treatment beyond traditional density functional approximations are discussed. The low‐rank electronic structure theories have further prompted us to highlight compressive and expressive principles that are useful to catalyze idea of quantum learning models. The intersection of the low‐rank correlated feature design and the modern deep neural network learning provides new feasibilities to predict chemically accurate correlation energies of unknown molecules that are not represented in the training dataset. The results by others and us are discussed to reveal that the electronic feature sets from an extremely low‐rank correlation representation, which is very poor for explicit energy computation, are however sufficiently expressive for capturing and transferring electron correlation patterns across distinct molecular compositions, bond types and geometries. This article is categorized under: Electronic Structure Theory > Ab Initio Electronic Structure Methods Software > Quantum Chemistry Software > Simulation Methods

Wiley Interdisciplinary Reviews Computational Molecular ScienceMar 1, 20244 citations

强关联分子体系电子结构方法研究进展

Wei Liang Huang, Yifan Cheng, Haibo Ma

In recent years, numerous novel materials have emerged in fields such as high-temperature superconductivity and quantum communication, including unconventional superconductors, topological insulators, and spin liquids. These materials exhibit significant quantum phenomena such as quantum interference, tunneling, fluctuations, and entanglement, often arising from strong correlation effects among electrons within molecules. In the community of quantum chemistry, strong correlation effects are typically classified into dynamic electron correlation and static electron correlation. Dynamic electron correlation, caused by the instantaneous movement of electrons leading to charge density fluctuations, presents computational challenges due to the necessity of introducing a multi-particle form of the wave function. This departure from the mean-field single-particle perspective results in a substantial increase in the computational cost for calculating electron correlation levels. On the other hand, static electron correlation arises from the near-degeneracy of numerous frontier molecular orbitals, resulting in a quantum superposition of electron configurations. It requires the use of multi-configuration methods and faces the well-known “exponential wall” challenge, where computational complexity exponentially grows with the system size. Therefore, due to the involvement of both dynamic and static electron correlations in strongly correlated systems, achieving precise electronic structure calculations is exceptionally difficult, representing a significant challenge in theoretical chemistry and computational physics. Traditionally, to describe the static correlation, quantum chemists often define crucial correlated orbitals and their electrons within the complete active space (CAS). Full configuration interaction (FCI) calculations are performed within the CAS, while the region outside the CAS is described by mean-field methods. While these multi-configuration methods allow for arbitrary electron arrangement within the CAS, effectively capturing the static correlation of the system, they struggle to handle the dynamic correlations beyond the CAS. Therefore, combining them with multi-reference configuration interaction (MRCI), multi-reference perturbation theory (MRPT), multi-reference coupled cluster theory (MRCC), and other multi-reference methods becomes essential to accurately handle strong correlation systems. Constructing multi-configuration wave functions to obtain static correlation by multi-configuration methods, and then generating excitation configurations based on these multi-configuration wave functions to obtain dynamic correlation has become the mainstream of the development of current wave function methods. Challenges in this direction include the exponential growth in computational complexity with an increase in CAS size and the precision of multi-reference methods affecting the accuracy of capturing dynamic correlations beyond the active space. Over the last decade, various multi-configuration methods, such as selected configuration interaction (sCI), Density matrix renormalization group (DMRG), full configuration interaction quantum Monte Carlo (FCIQMC), and neural-network quantum states (NQS), have successfully expanded the computable active space by explicitly or implicitly extracting and compressing complete configuration space information. Furthermore, combining these multi-configuration methods with traditional multi-reference approaches, such as MRCI, MRPT and MRCC, enables capturing dynamic correlations caused by electronic instantaneous interactions beyond the active space. This achieves higher precision and quantitative descriptions for mid-sized strongly correlated molecular systems (around 50 active orbitals). Additionally, novel methods like Block-Correlated Coupled Cluster Theory (BCCC) have made progress in attempting to discuss static and dynamic correlations within a unified framework. This article primarily introduces the advancements in wave function theory electronic structure methods for strongly correlated molecular systems in recent years. The first section details multi-configuration methods that have expanded the CAS size, including sCI, DMRG, FCIQMC, NQS, and outlines their developmental trajectories. The second section delves into the combination of these multi-configuration methods with traditional multi-reference approaches, leveraging larger active spaces to address dynamic correlations and achieve higher precision results. Finally, the article discusses the prospects of these methods in three directions: Handling static and dynamic correlations within a unified framework, integrating with relativistic effects, and combining with machine learning.

Chinese Science Bulletin (Chinese Version)Apr 1, 20241 citation