Skip to main content
v2026.11,986 entries · CC-BY 4.0

QM/MM: Quantum Mechanics/Molecular Mechanics Explained

QM/MM treats a reactive region quantum-mechanically and its surroundings with a force field; learn the schemes, embedding options and what to report.

Written and maintained by CASRAI Editorial Board

Published

On this page

QM/MM (quantum mechanics/molecular mechanics) is a hybrid simulation method that treats the small part of a molecular system where bonds form or break with quantum mechanics and treats the rest, such as the surrounding protein and solvent, with a classical molecular-mechanics force field. It exists because quantum calculations are accurate for chemistry but too costly for large systems, while force fields are cheap but cannot describe reactions.

Short answer: what QM/MM means

QM/MM combines two descriptions of one system in a single calculation. A chemically active region, such as a substrate and cofactors in an enzyme active site, is described by a quantum-mechanical (QM) method. The remainder is described by molecular mechanics (MM). Senn and Thiel’s 2009 review in Angewandte Chemie summarises the reason: QM methods are restricted to systems of up to a few hundred atoms, while biopolymers call for methods that handle up to several 100,000 atoms and simulations over tens of nanoseconds, which force-field MM methods provide.

This page is for researchers and lab staff who meet the abbreviation in a paper, a software manual or a collaborator’s methods section and need to know what was actually done, which choices define a QM/MM calculation, and what a reader needs in order to reproduce it.

Definition and scope

The two halves of the name map onto two older families of method. Quantum-mechanical methods compute the electronic structure of a set of atoms, which is what allows bond breaking, charge transfer and electronic excitation to be described. Molecular mechanics represents atoms as particles interacting through a parameterised force field, which is what makes large, flexible systems tractable. The molecular dynamics simulation guide covers the MM side, including why classical force fields cannot, on their own, describe reactive chemistry. QM/MM is the bridge between that method and the quantum chemistry covered in the physical chemistry overview and in software guides such as ORCA.

The Royal Swedish Academy of Sciences announced the 2013 Nobel Prize in Chemistry on 9 October 2013 to Martin Karplus, Michael Levitt and Arieh Warshel “for the development of multiscale models for complex chemical systems”. The press release states that their work, whose foundation was laid in the 1970s, made Newton’s classical physics work side by side with the fundamentally different quantum physics, where chemists previously had to choose one or the other. It describes the trade-off plainly: classical programs handle large molecules with simple calculations but offer no way to simulate chemical reactions, while quantum calculations need enormous computing power and are therefore used only for small molecules. An early landmark from that decade is Warshel and Levitt’s 1976 paper in the Journal of Molecular Biology on enzymic reactions, using the lysozyme reaction as its case.

In scope, QM/MM is a method for modelling reactions and other electronic processes inside large environments. Cao and Ryde, in a 2018 paper on additive and subtractive calculations in Frontiers in Chemistry, call it a popular method for studying reactions in biochemical macromolecules. It is not a separate theory from QM or MM. It is a way of coupling them, and every published QM/MM result is therefore only as defined as its coupling choices are stated.

How QM/MM is used in practice

The Nobel Prize press release gives the practical picture with a drug-binding example: quantum calculations are run on those atoms in the target protein that interact with the drug, while the rest of the large protein is simulated with classical physics. Each QM/MM project involves the same sequence of decisions, taken from the structure of the method as Cao and Ryde describe it.

Step 1: choose the QM region and the MM region

The QM region is the small region of central interest. Cao and Ryde give a typical size of 20 to 300 atoms. The MM region is the rest of the macromolecule plus explicit solvent. Which atoms are in the QM region is a modelling decision, and it is the first thing a reader of a methods section needs.

Step 2: handle the boundary

When the QM and MM regions are separated by covalent bonds, the QM system has to be truncated. Cao and Ryde note that this is typically done with hydrogen link atoms, and that in subtractive schemes MM parameters for the link atoms are needed. Reports that cut a covalent bond between regions should say how the cut was capped.

Step 3: choose how energies are combined (the scheme)

Two schemes are described in the review, and they differ in which calculations are run.

  • Additive scheme: one QM calculation on the QM region plus a single MM calculation on the whole system, using an internal MM program. The MM term is often split into the MM energy of the MM part and a QM/MM interface energy, and care is needed to avoid omitted or double-counted interactions.
  • Subtractive scheme (ONIOM-type): the QM energy of the truncated QM system plus two MM calculations, one on the QM region alone and one on the full system, using an external MM program.

Step 4: choose how the regions interact electrostatically (the embedding)

Embedding describes how the MM environment influences the QM region. It is a separate choice from the scheme, as the next section explains.

Types: scheme and embedding are two separate choices

Cao and Ryde argue that the literature often confuses additive and subtractive schemes with mechanical and electrostatic embedding, and stress that the two choices are separate. A calculation therefore needs both labels, for example “additive scheme with electrostatic embedding”. The data table below sets out the five options with what each one computes and what it requires.

Mechanical embedding treats the QM-MM electrostatic interaction at the MM level, so the QM region is not polarised by its surroundings. Electrostatic embedding includes the MM atoms as point charges inside the QM calculation, which polarises the QM region, but only one way. Polarised embedding is mutual and self-consistent, so the MM region responds to the QM region as well; it needs a polarisable force field and software that supports it.

Where it matters for labs and research offices: reproducibility and reporting

Because the scheme, the embedding, the QM region and the boundary treatment each change the energy that comes out, they are part of the method, not implementation detail. This is an editorial reading of the sourced choices above rather than a published standard: a result stated only as “computed with QM/MM” cannot be reproduced or compared. The checklist below is built directly from the decisions the sources identify. For the general principle, see the computational reproducibility entry.

Item to report Why it is needed (source statement)
Atoms in the QM region, and its size The QM region is the small region of central interest, typically 20-300 atoms (Cao and Ryde 2018).
What the MM region contains The rest of the macromolecule plus explicit solvent (Cao and Ryde 2018).
Scheme: additive or subtractive The two schemes use different calculations and, for subtractive, an external MM program (Cao and Ryde 2018).
Embedding: mechanical, electrostatic or polarised Determines whether, and in which direction, the QM region is polarised by the MM environment (Cao and Ryde 2018).
Treatment of covalent boundaries The QM system is typically truncated with hydrogen link atoms; subtractive schemes need MM parameters for them (Cao and Ryde 2018).
MM force field and QM method The two halves of the method; polarised embedding additionally needs a polarisable force field (Cao and Ryde 2018).

For lab managers and research offices, the practical consequence concerns the computing and the people. QM calculations are the expensive part of the method, so the size of the QM region and the system’s total atom count feed directly into the computing time a project needs and into the data a project should archive. Reviewers and data-management plans benefit when input structures, region definitions and software settings are deposited alongside the paper.

Common misunderstandings

“QM/MM is just a faster quantum calculation”

It is a different model. Only the QM region receives the quantum treatment. Everything outside it is described by a force field, so the result carries the approximations of both halves and of the coupling between them.

“Additive and subtractive are the same as mechanical and electrostatic”

They are not. Cao and Ryde say the literature often confuses them, and that the scheme and the embedding are two independent choices. A methods section should therefore name both.

“The QM region can be any size”

QM methods are restricted to systems of up to a few hundred atoms according to Senn and Thiel, and Cao and Ryde give a typical QM region of 20 to 300 atoms. The region should include the atoms of central chemical interest, but it cannot grow without limit.

“The method is new”

The foundations date to the 1970s. Jorgensen, writing in Cell in 2013 about the work honoured by that year’s Nobel Prize, says it provided a foundation for the widespread activities today in modelling organic and biomolecular systems.

Frequently asked questions

What does QM/MM stand for?

QM/MM stands for quantum mechanics/molecular mechanics. It is a hybrid method that treats a small, chemically active region quantum-mechanically and the surrounding system with a classical force field (Senn and Thiel 2009).

Who developed QM/MM?

The 2013 Nobel Prize in Chemistry was awarded to Martin Karplus, Michael Levitt and Arieh Warshel for the development of multiscale models for complex chemical systems, with the foundation laid in the 1970s (Nobel Prize press release, 9 October 2013).

What is the difference between additive and subtractive QM/MM?

The additive scheme runs one QM calculation on the QM region and one MM calculation on the whole system. The subtractive scheme runs the QM calculation on the truncated QM system and two MM calculations, on the QM region alone and on the full system (Cao and Ryde 2018).

What is electrostatic embedding?

Electrostatic embedding places the MM atoms as point charges inside the QM calculation, which polarises the QM region. The polarisation runs one way: the MM region is not polarised by the QM region unless polarised embedding is used (Cao and Ryde 2018).

Sources

Data: QM/MM scheme and embedding options compared

Compiled from Cao and Ryde (2018): the two QM/MM energy schemes and the three embedding treatments, with the calculations each requires and its key condition.

Period covered: 2018.

QM/MM scheme and embedding options compared
OptionChoice typeWhat it involvesKey condition
AdditiveSchemeOne QM calculation on the QM region plus one MM calculation of the whole system, using an internal MM programAvoid omitted or double-counted interactions
Subtractive (ONIOM-type)SchemeQM energy of the truncated QM system plus two MM calculations (QM region alone, full system), using an external MM programMM parameters needed for link atoms
Mechanical embeddingEmbeddingQM-MM electrostatics treated at the MM levelQM region is not polarised by the surroundings
Electrostatic embeddingEmbeddingMM atoms included as point charges in the QM calculationPolarises the QM region, one way only
Polarised embeddingEmbeddingMutual, self-consistent polarisation of QM and MM regionsNeeds a polarisable force field and supporting software

Sources

Compiled by CASRAI from the sources above. Free to reuse under CC BY 4.0 with attribution.

Follow CASRAI

Research-administration guidance, standards updates and independent tool reviews.

Ask CASRAI · full access

Ask about QM/MM: Quantum Mechanics/Molecular Mechanics Explained

Full access: large monthly usage, up to 10 source passages behind each answer, and MCP and API access for Claude Code, Codex and other CLIs. Cancel any time.

An AI assistant for research: funder rules, compliance, methods and publishing. It links the sources behind its answers and labels answers that come from a web search.

Answers draw on CASRAI's guides and dictionary plus the federal and funder documents we index: Federal Register, Grants.gov, Regulations.gov and UKRI.

Works on this site and inside Claude, Cursor and the AI tools you already use.

Everything CASRAI publishes — this page, the dictionary, the guides and the news — stays open to read, with no account and no card.

Referenced across the research world

University of Cambridge logoColumbia University logoCrossref logoUniversity of Edinburgh logoHarvard University logoUniversity of Oxford logoPrinceton University logoStanford School of Medicine logoUniversity College London logoORCID logoUniversity of Cambridge logoColumbia University logoCrossref logoUniversity of Edinburgh logoHarvard University logoUniversity of Oxford logoPrinceton University logoStanford School of Medicine logoUniversity College London logoORCID logo
  • University of Cambridge logo
  • Columbia University logo
  • Crossref logo
  • University of Edinburgh logo
  • Harvard University logo
  • University of Oxford logo
  • Princeton University logo
  • Stanford School of Medicine logo
  • University College London logo
  • ORCID logo

View CASRAI adoption →