| correlation (quantum chemistry) | |
|---|---|
| Name | Electron correlation |
| Caption | Schematic of correlated electron motion |
| Field | Quantum chemistry |
| Introduced | 20th century |
| Notable practitioners | D. R. Hartree, V. Fock, John A. Pople, C. C. J. Roothaan |
correlation (quantum chemistry)
Correlation (quantum chemistry) refers to the parts of the electronic energy of atoms and molecules that are not captured by single-determinant mean-field theories such as the Hartree–Fock approximation. It quantifies the instantaneous interactions and quantum entanglement among electrons and is crucial for accurately predicting observables in Quantum Physics-informed chemistry, including binding energies, spectra, and reaction dynamics. Accounting for electron correlation links theoretical models to experimental results and underpins advances in computational methods across academia and industry.
In quantum-many body terms, electron correlation arises because the exact electronic wavefunction of an N-electron system cannot generally be represented as a single antisymmetrized product (one Slater determinant). Correlation embodies both Coulombic and exchange effects beyond mean-field, and it is formally defined relative to the Hartree–Fock reference energy as the difference between the exact nonrelativistic energy and the Hartree–Fock energy. Accurate treatment of correlation is essential in fields ranging from molecular spectroscopy studied at NIST to condensed-matter problems at facilities like Argonne and CERN where many-body quantum effects are central. The concept also connects to quantum information notions such as quantum entanglement and reduced density matrices used in reduced-density-matrix functional theory.
Electron correlation is often partitioned into two physically distinct components. Static correlation (also called nondynamic) appears when multiple electronic configurations have near-degenerate energies — typical in bond breaking, diradicals, or transition metal complexes (studied at Max Planck institutes and university research groups). Dynamic correlation describes the short-range avoidance between electrons due to Coulomb repulsion and is required even in single-reference systems like closed-shell organic molecules. Multireference approaches (e.g., CASSCF) capture static correlation, while perturbative corrections (e.g., CASPT2) or coupled cluster recover dynamic effects. Distinguishing both is important for interpreting spectra measured by groups at LBNL and for materials modeling at ORNL.
A range of theoretical frameworks exist to include correlation. Configuration interaction (CI) builds many-electron wavefunctions from linear combinations of Slater determinants but suffers from exponential scaling; full CI is exact within a finite basis and used as a benchmark in studies by theoretical chemists such as Roald Hoffmann and John Pople. Coupled cluster (CC) theory, notably CC with singles, doubles, and perturbative triples (CCSD(T)), is considered a gold standard for dynamic correlation. Møller–Plesset (MP) methods (MP2, MP3...) provide perturbative corrections to Hartree–Fock. Density functional theory (DFT) offers an efficient alternative by approximating the exchange–correlation functional; efforts like the PBE and B3LYP functionals trade accuracy for tractability and are widely used in industry and academia (e.g., IBM Research, Dow). Hybrid approaches—range-separated functionals, double-hybrid DFT, and methods combining DFT with many-body perturbation theory such as GW—seek improved treatment of correlation. Specialized methods for strongly correlated solids include DMFT used in studies at Princeton and Rutgers.
Accurate correlation methods typically scale steeply with system size: full CI scales factorially, CCSD(T) scales approximately N^7 with basis size, and high-level multireference methods can be even more demanding. This cost creates disparities: well-funded groups at national labs and top universities can access supercomputers (e.g., NERSC, XSEDE) and proprietary software (Gaussian, Molpro), while researchers in low-income institutions face barriers to participation. These inequities affect who can validate new materials or catalysts, influencing whose priorities are advanced. Open-source projects like Psi4 and PySCF and community initiatives (training workshops sponsored by IUPAC or regional consortia) aim to democratize access, but persistent resource gaps and software licensing practices continue to skew research opportunities.
Correctly accounting for correlation is central to predicting spectroscopic observables (UV–vis, X-ray, IR) and interpreting experiments from facilities like ESRF and Diamond. Reaction barrier heights and thermochemistry—critical for industrial chemistry at firms like BASF and in atmospheric chemistry modeled by NOAA—depend on accurate correlation treatment. In materials science and catalysis, correlation affects band gaps, magnetic ordering, and active-site energetics; methods combining DFT with beyond-DFT corrections inform design in battery and photocatalyst research at institutions such as MIT and Stanford.
Major challenges include balancing accuracy and cost, treating strong correlation in transition-metal systems, and developing functionals and wavefunction methods that scale to large systems. Ongoing research explores machine learning corrections trained on high-level correlated data (efforts at DeepMind and academic labs), tensor-decomposition techniques, quantum algorithms for chemistry (e.g., variational quantum eigensolver researched by IBM Quantum and Google Quantum AI), and reduced-density-matrix methods promoted by theoretical groups at Cambridge and ETH Zurich. Ethical and social considerations—such as prioritizing open access, capacity building in underrepresented regions, and aligning computational chemistry research with public-interest goals like climate mitigation—are increasingly discussed at conferences like the ACS national meetings and in policy dialogues.