Constant-temperature molecular dynamics

Abstract

A review is given of how molecular dynamics methods have been modified to perform simulations in the constant-temperature condition. We usually consider a system which is thermally connected with a huge external system (a heat reservoir) to describe a canonical ensemble in statistical mechanics. The way in which this situation is reflected is a key factor for simulations under isothermal conditions. The total kinetic energy is kept to a constant value in constraint methods. In stochastic methods, interactions with a heat bath are treatedasrandomcollisionswith hypotheticalatomsor randomforcesactingonparticles. In the extended-system method, a degree of freedom which mimics a heat bath is introduced, and the total energy of a physical system is allowed to fluctuate.

On the Combination of TDDFT with Molecular Dynamics: New Developments

“…(1.40). This property, which is also satisfied by other dynamics like the one proposed by Nosé [Nosé 1984, Nosé 1991 if the H eff in Eq. (1.44) is used, comes out in a very natural way from the present formalism while it is yet unclear of other ab initio MD candidates for going beyond gsBOMD [Mauri 1993, Parandekar 2005, Schmidt 2008, Bastida 2007.…”

Nonequilibrium adiabatic molecular dynamics simulations of methane clathrate hydrate decomposition

“…Simulations of ten methane molecules in 1242 water molecules were performed for seven temperatures in the range of 273-310 K. The solutions were first equilibrated with a NpT simulation for 100 ps ͑with 50 ps equilibration͒ and the resulting configuration was simulated for 200 ps ͑with 50 ps equilibration͒ in a NVE simulation. In the NpT simulations, the Nosé-Hoover thermostat-barostat algorithms 32 were used with 0.5 and 2.0 fs relaxation times, respectively. The temperature dependence of the methane diffusion coefficients in water are plotted as an Arrhenius plot in Fig.…”

Reactivity and Dynamics at Liquid Interfaces

“…These particles are usually identified as the individual atoms in the molecules, but in many applications a particle can represent a group of atoms ("united" atom) or a fictitious mass for modeling coupling to an external bath 66 or for describing fluctuating charges in a molecule (see below).…”

“…These terms are obtained from a fit to the ab-initio calculations of Tucker and Truhlar 637 and to experimental data. 622 The generalization to non-linear geometry is accomplished by making some of the parameters θ-dependent and adding a bending energy term with parameters determined by a best fit to the gas-phase ab-initio values of the energy, to the location of the transition state, and to the ion-dipole well-depth as a function of θ. , [66] where S(r) is the overlap integral for the sigma orbital formed from the carbon 2p and chlorine 3p atomic orbitals. S(r) is determined using Slater-type orbitals and the approximation of Mulliken et al, 636,640 and Q = 678.0 kcal/mol is a parameter that is fitted to obtain the correct gas-phase activation energy.…”