What Is Computational Chemistry?

Computational chemistry is a branch of chemistry that uses computer simulation and numerical approximation to predict the structure, energetics, and reactivity of molecules and materials. It applies quantum mechanics and classical mechanics to chemical systems that are too large, too short-lived, or too hazardous to characterize directly in a laboratory. The field sits between theoretical chemistry, which derives the underlying models, and experimental chemistry, which supplies the measurements those models are calibrated against.

Its practical beginnings date to the 1960s and 1970s, when programmable computers first made it feasible to solve approximate forms of the electronic Schrödinger equation for polyatomic molecules. The 1998 Nobel Prize in Chemistry recognized Walter Kohn for density functional theory and John Pople for the computational methods and software, beginning with the GAUSSIAN program released in 1970, that turned quantum chemistry into a routine working tool. The discipline now overlaps heavily with computational physics, materials science, and structural biology.

Electronic Structure Methods

Electronic structure methods solve for the distribution of electrons in a molecule and derive observable properties from it. Ab initio approaches begin from the Schrödinger equation with no empirical parameters: Hartree-Fock theory treats each electron in the average field of the others, and post-Hartree-Fock methods such as Møller-Plesset perturbation theory and coupled cluster recover the electron correlation energy that Hartree-Fock omits. Coupled cluster with single, double, and perturbative triple excitations, written CCSD(T), is widely treated as the reference standard for small molecules, though its cost scales roughly as the seventh power of system size. Density functional theory replaces the many-electron wavefunction with the electron density, which reduces cost enough to handle hundreds of atoms and accounts for the majority of published calculations. Every method depends on a basis set, a finite collection of functions used to represent molecular orbitals, and the choice of basis set is often the dominant source of error.

Molecular Mechanics and Molecular Dynamics

Where electronic detail is unnecessary, molecular mechanics models atoms as classical particles linked by springs, with bonded and nonbonded interactions defined by an empirical force field such as AMBER, CHARMM, or OPLS. Molecular dynamics then integrates Newton's equations of motion, typically with a one or two femtosecond timestep, to follow a system through nanoseconds to milliseconds of simulated time. This is the standard route to protein folding pathways, membrane behavior, and solvation free energies. Hybrid quantum mechanics and molecular mechanics schemes treat a reactive site quantum mechanically while the surrounding protein or solvent remains classical, an approach recognized by the 2013 Nobel Prize in Chemistry awarded to Martin Karplus, Michael Levitt, and Arieh Warshel.

Benchmarking and Reference Data

Because every method is an approximation, the field depends on curated reference data to establish how far a given approach can be trusted. The NIST Computational Chemistry Comparison and Benchmark Database collects experimental and calculated thermochemical properties, geometries, vibrational frequencies, and reaction barriers for several hundred gas-phase species, letting practitioners compare method and basis set combinations against measured values. Benchmarking of this kind established the informal target of chemical accuracy, roughly 1 kcal/mol in computed reaction energies, and it supplies the scaling factors applied to calculated vibrational frequencies. Benchmark sets have more recently become training data for machine-learned interatomic potentials that aim to reproduce quantum-level accuracy at classical cost.

Applications

Computational chemistry has applications in a wide range of fields, including:

• Pharmaceutical research, for virtual screening, docking, and binding affinity prediction
• Catalysis, for mapping reaction mechanisms and transition state energies
• Materials design, including battery electrolytes, photovoltaics, and metal-organic frameworks
• Atmospheric and combustion chemistry, where rate constants are difficult to measure directly
• Spectroscopy, for assigning experimental infrared, NMR, and UV-visible spectra