Configuration interaction

Configuration interaction
Name	Configuration interaction
Type	electronic structure method
Introduced	1920s–1950s
Primary authors	Walter Heitler, Fritz London, Douglas Hartree, Vladimir Fock
Related methods	Hartree–Fock method, Møller–Plesset perturbation theory, Coupled cluster, Density functional theory

Configuration interaction. Configuration interaction is a quantum chemistry electronic structure approach that constructs correlated many-electron wavefunctions by combining multiple Slater determinants derived from independent-particle references such as Hartree–Fock method, Hartree method, or self-consistent field solutions. It provides a systematically improvable framework to capture electron correlation by including excited determinants (single, double, triple, etc.) and is widely used alongside methods developed by Walter Heitler, Fritz London, Douglas Hartree, and Vladimir Fock in computational studies of atoms, molecules, and solids.

Introduction

Configuration interaction builds correlated wavefunctions as linear combinations of determinants or configuration state functions sourced from a reference like the Hartree–Fock method or multiconfigurational references such as Complete active space self-consistent field and Multi-configuration self-consistent field. The CI expansion can be truncated at various excitation levels — CIS, CISD, CISDT, CISDTQ — to balance accuracy and cost, with benchmark comparisons often made to Coupled cluster and high-level correlated methods validated against experimental data from facilities like the National Institute of Standards and Technology and spectroscopic measurements cataloged by institutions such as the Royal Society of Chemistry.

Theory and formalism

The CI formalism expresses the exact non-relativistic electronic wavefunction as a linear combination of orthonormal determinants formed from one-electron spinorbitals, typically derived from a Hartree–Fock method or generalized Kohn–Sham reference influenced by work at institutes like Max Planck Society and University of Cambridge. The CI secular problem leads to a configuration interaction matrix whose diagonalization yields energies and eigenvectors, analogous to procedures in matrix methods used at centers such as Los Alamos National Laboratory and Lawrence Livermore National Laboratory. Spin adaptation and point-group symmetry from groups like International Union of Pure and Applied Chemistry reduce dimensionality, while second-quantized operator algebra connects CI to field-theoretic formulations used by researchers at CERN and Princeton University.

Computational methods and approximations

Practical CI calculations rely on truncated expansions (CIS, CISD) and selected CI algorithms such as configuration selection algorithms pioneered by groups at University of California, Berkeley and Massachusetts Institute of Technology. Deterministic full configuration interaction (FCI) is feasible only for small basis sets or few electrons; stochastic approaches like full configuration interaction quantum Monte Carlo developed by teams at University of Cambridge and University of Toronto extend reach, as do selected CI strategies like CIPSI used in collaborations involving CNRS and Université Paris-Saclay. Basis sets from authors such as John Pople and families like correlation-consistent sets by David Sherrill or Dunning interplay with orbital optimization schemes influenced by John Roos and Boys and Handy to manage convergence. Implementation advances at software projects such as Gaussian (software), MOLPRO, Psi4, and ORCA leverage parallel computing architectures found at national supercomputing centers including Argonne National Laboratory and Oak Ridge National Laboratory.

Applications and examples

CI methods are applied to compute excitation energies, potential energy surfaces, and ionization potentials for molecules studied by research groups at Caltech, Harvard University, and ETH Zurich. Examples include spectroscopy of small molecules benchmarked against experiments at National Institute of Standards and Technology and photochemistry investigations in collaborations with Lawrence Berkeley National Laboratory and Max Planck Institute for Chemical Physics of Solids. Multireference problems such as bond breaking in diatomics, transition metal complexes characterized by experimental teams at Stanford University and Imperial College London, and electronic states of organic chromophores examined by groups at University of Oxford showcase CI’s utility. In materials science, CI-derived insights complement many-body approaches used at Argonne National Laboratory and Brookhaven National Laboratory to interpret correlated electron behavior.

Limitations and improvements

CI suffers from combinatorial growth of the determinant space, making full CI impractical beyond small systems; this bottleneck has motivated reduced-scaling methods developed by researchers at ETH Zurich and University of Waterloo. Truncated CI (e.g., CISD) lacks size-consistency, prompting corrections and the adoption of size-extensive alternatives like Coupled cluster methods championed by scientists at University of Florida and Weizmann Institute of Science. Advances such as internally contracted CI, density matrix renormalization group inspired schemes from groups at University of Stuttgart and adaptive sampling CI methods from teams at University of Toronto aim to mitigate scaling. Hybrid strategies combining CI with Density functional theory embedding and quantum embedding theories explored at Argonne National Laboratory and Flatiron Institute address strong correlation in extended systems.

Historical development

The conceptual roots of CI trace to early quantum theory and determinant expansions following the foundational works of Walter Heitler and Fritz London on chemical bonding and the independent-particle formulations by Douglas Hartree and Vladimir Fock. Semi-empirical and ab initio CI techniques matured through mid-20th century efforts at institutions such as University of Cambridge and Bell Labs, with algorithmic and computational progress driven by supercomputing initiatives at Los Alamos National Laboratory and code development from groups at IBM and University of California, Los Angeles. Later innovations in selected CI, stochastic FCI, and embedding approaches have emerged from collaborations spanning Université Paris-Sud, Columbia University, and California Institute of Technology.

Category:Quantum chemistry