Thermostats and Statistical Ensembles

Bare Newtonian molecular dynamics conserves energy and samples the microcanonical ensemble, but real experiments hold temperature and pressure fixed, so thermostats and barostats are added to make a simulation sample the desired statistical ensemble.

Definition

A thermostat is an algorithm coupled to molecular dynamics that controls the system temperature so that time averages sample a chosen statistical ensemble; a barostat does the same for pressure.

Scope

This topic covers the methods that control temperature and pressure in molecular dynamics: velocity rescaling and stochastic thermostats, the deterministic Nose-Hoover thermostat and its chains, and barostats for constant-pressure simulation, together with the ensembles, microcanonical, canonical and isothermal-isobaric, that they realize.

Core questions

• How does adding a thermostat change microcanonical dynamics into canonical sampling?
• Why is the Nose-Hoover thermostat preferred over simple velocity rescaling for correct ensembles?
• How do barostats allow the simulation box to fluctuate at constant pressure?
• How can a thermostat distort dynamical properties if applied too strongly?

Key theories

Canonical sampling and thermostats

Coupling the system to a heat bath, by stochastic collisions or rescaling, drives the time-averaged kinetic energy to the target temperature so the trajectory samples the canonical ensemble rather than fixed energy.

Nose-Hoover dynamics

The Nose-Hoover thermostat introduces an extra dynamical variable representing the heat bath, giving deterministic, time-reversible equations whose trajectory provably samples the canonical distribution.

Barostats and the isothermal-isobaric ensemble

Barostats let the simulation volume fluctuate by coupling to a pressure bath, so that, combined with a thermostat, the dynamics sample the constant-temperature, constant-pressure ensemble of typical experiments.

Clinical relevance

Correct ensemble control is essential for computing free energies, phase behavior and response properties under experimentally relevant conditions, and is standard practice in materials, soft-matter and biomolecular simulations.

History

Constant-temperature molecular dynamics developed through the 1980s, with Andersen's stochastic thermostat and barostat, Nose's extended-system formulation in 1984 and Hoover's reformulation in 1985 providing the now-standard deterministic route to canonical sampling.

Debates

Ergodicity of deterministic thermostats

Single Nose-Hoover thermostats can fail to be ergodic for small or stiff systems, sampling the wrong distribution; thermostat chains and stochastic alternatives were introduced to address this, and the best choice remains system-dependent.

Key figures

• Shuichi Nose
• William G. Hoover
• Hans Andersen

Seminal works

• nose1984
• hoover1985

Frequently asked questions

Why not just rescale velocities to fix the temperature?

Simple velocity rescaling holds the kinetic energy fixed but does not reproduce the proper canonical fluctuations, so it samples the wrong ensemble. Methods like Nose-Hoover or stochastic thermostats allow the correct temperature fluctuations while keeping the average on target.

Can a thermostat affect the dynamics being studied?

Yes. A strongly coupled thermostat perturbs the natural motion and can bias transport properties, so weak coupling or a thermostat applied only to control equilibration is used when accurate dynamics are needed.