Semiempirical and Composite Methods

Between cheap empirical models and the most rigorous wavefunction theories lie semiempirical and composite methods, which trade selected approximations or clever combinations for accuracy at reduced cost.

Definition

A grouping of quantum-chemical strategies that either introduce empirical parameters, combine several calculations, or extend the formalism to handle strong correlation and relativity, balancing accuracy against computational cost.

Scope

Covers semiempirical molecular-orbital methods that parameterize or neglect expensive integrals, multireference and multiconfigurational approaches for systems where a single determinant fails, composite 'model chemistry' protocols that combine calculations to reach high thermochemical accuracy affordably, and relativistic methods for heavy elements. These complement the standard single-reference electronic structure and density functional methods.

Sub-topics

• Composite Thermochemistry Methods
• Multireference and Multiconfigurational Methods
• Relativistic Quantum Chemistry
• Semiempirical Quantum Chemistry Methods

Core questions

• How do semiempirical methods achieve speed by approximating or parameterizing integrals?
• When does a single reference determinant break down, requiring multireference methods?
• How do composite protocols reach high accuracy at moderate cost?
• When must relativistic effects be included for heavy elements?

Key theories

Integral approximation and parameterization

Semiempirical methods neglect or approximate the most expensive electron-repulsion integrals and replace them with parameters fitted to data, gaining orders of magnitude in speed.

Composite model chemistries

Additive schemes combine results from different methods and basis sets to estimate a high-level result that would be too costly to compute directly, targeting chemical accuracy in thermochemistry.

Clinical relevance

These methods extend computational chemistry to regimes that standard approaches cannot handle affordably: very large molecules and screening with semiempirical methods, strongly correlated and excited systems with multireference methods, benchmark thermochemistry with composite schemes, and heavy-element chemistry with relativistic methods.

History

Semiempirical methods grew from Hückel and Pariser-Parr-Pople theory through Dewar's MNDO, AM1, and PM3 models; multiconfigurational methods such as CASSCF were developed by Roos and others; composite schemes like the Gaussian-n and Weizmann methods emerged in the 1990s to deliver benchmark accuracy.

Key figures

• Michael Dewar
• Walter Thiel
• Björn Roos
• Larry Curtiss

Seminal works

• thiel2014
• cramer2004

Frequently asked questions

Are semiempirical methods still relevant given fast computers?

Yes; they remain valuable for very large systems, high-throughput screening, and as fast components in multiscale and machine-learning workflows where full ab initio treatment is impractical.

What does a composite method actually compute?

It combines several smaller calculations, typically a high-correlation result in a modest basis with basis-set and other corrections, to approximate an accurate energy at far lower cost than the direct high-level calculation.