fix langevin command

Accelerator Variants: langevin/kk


fix ID group-ID langevin Tstart Tstop damp seed keyword values ...
  • ID, group-ID are documented in fix command

  • langevin = style name of this fix command

  • Tstart,Tstop = desired temperature at start/end of run (temperature units)

  • Tstart can be a variable (see below)

  • damp = damping parameter (time units)

  • seed = random number seed to use for white noise (positive integer)

  • zero or more keyword/value pairs may be appended

  • keyword = angmom or gjf or omega or scale or tally or zero

    angmom value = no or factor
      no = do not thermostat rotational degrees of freedom via the angular momentum
      factor = do thermostat rotational degrees of freedom via the angular momentum and apply numeric scale factor as discussed below
    gjf value = no or vfull or vhalf
      no = use standard formulation
      vfull = use Gronbech-Jensen/Farago formulation
      vhalf = use 2GJ formulation
    omega value = no or yes
      no = do not thermostat rotational degrees of freedom via the angular velocity
      yes = do thermostat rotational degrees of freedom via the angular velocity
    scale values = type ratio
      type = atom type (1-N)
      ratio = factor by which to scale the damping coefficient
    tally value = no or yes
      no = do not tally the energy added/subtracted to atoms
      yes = do tally the energy added/subtracted to atoms
    zero value = no or yes
      no = do not set total random force to zero
      yes = set total random force to zero


fix 3 boundary langevin 1.0 1.0 1000.0 699483
fix 1 all langevin 1.0 1.1 100.0 48279 scale 3 1.5
fix 1 all langevin 1.0 1.1 100.0 48279 angmom 3.333


Apply a Langevin thermostat as described in (Schneider) to a group of atoms which models an interaction with a background implicit solvent. Used with fix nve, this command performs Brownian dynamics (BD), since the total force on each atom will have the form:

\[\begin{split}F = & F_c + F_f + F_r \\ F_f = & - \frac{m}{\mathrm{damp}} v \\ F_r \propto & \sqrt{\frac{k_B T m}{dt~\mathrm{damp}}}\end{split}\]

\(F_c\) is the conservative force computed via the usual inter-particle interactions (pair_style, bond_style, etc). The \(F_f\) and \(F_r\) terms are added by this fix on a per-particle basis. See the pair_style dpd/tstat command for a thermostatting option that adds similar terms on a pairwise basis to pairs of interacting particles.

\(F_f\) is a frictional drag or viscous damping term proportional to the particle’s velocity. The proportionality constant for each atom is computed as \(\frac{m}{\mathrm{damp}}\), where m is the mass of the particle and damp is the damping factor specified by the user.

\(F_r\) is a force due to solvent atoms at a temperature \(T\) randomly bumping into the particle. As derived from the fluctuation/dissipation theorem, its magnitude as shown above is proportional to \(\sqrt{\frac{k_B T m}{dt~\mathrm{damp}}}\), where \(k_B\) is the Boltzmann constant, \(T\) is the desired temperature, m is the mass of the particle, dt is the timestep size, and damp is the damping factor. Random numbers are used to randomize the direction and magnitude of this force as described in (Dunweg), where a uniform random number is used (instead of a Gaussian random number) for speed.

Note that unless you use the omega or angmom keywords, the thermostat effect of this fix is applied to only the translational degrees of freedom for the particles, which is an important consideration for finite-size particles, which have rotational degrees of freedom, are being thermostatted. The translational degrees of freedom can also have a bias velocity removed from them before thermostatting takes place; see the description below.


Unlike the fix nvt command which performs Nose/Hoover thermostatting AND time integration, this fix does NOT perform time integration. It only modifies forces to effect thermostatting. Thus you must use a separate time integration fix, like fix nve to actually update the velocities and positions of atoms using the modified forces. Likewise, this fix should not normally be used on atoms that also have their temperature controlled by another fix - e.g. by fix nvt or fix temp/rescale commands.

See the Howto thermostat page for a discussion of different ways to compute temperature and perform thermostatting.

The desired temperature at each timestep is a ramped value during the run from Tstart to Tstop.

Tstart can be specified as an equal-style or atom-style variable. In this case, the Tstop setting is ignored. If the value is a variable, it should be specified as v_name, where name is the variable name. In this case, the variable will be evaluated each timestep, and its value used to determine the target temperature.

Equal-style variables can specify formulas with various mathematical functions, and include thermo_style command keywords for the simulation box parameters and timestep and elapsed time. Thus it is easy to specify a time-dependent temperature.

Atom-style variables can specify the same formulas as equal-style variables but can also include per-atom values, such as atom coordinates. Thus it is easy to specify a spatially-dependent temperature with optional time-dependence as well.

Like other fixes that perform thermostatting, this fix can be used with compute commands that remove a “bias” from the atom velocities. E.g. to apply the thermostat only to atoms within a spatial region, or to remove the center-of-mass velocity from a group of atoms, or to remove the x-component of velocity from the calculation.

This is not done by default, but only if the fix_modify command is used to assign a temperature compute to this fix that includes such a bias term. See the doc pages for individual compute temp commands to determine which ones include a bias. In this case, the thermostat works in the following manner: bias is removed from each atom, thermostatting is performed on the remaining thermal degrees of freedom, and the bias is added back in.

The damp parameter is specified in time units and determines how rapidly the temperature is relaxed. For example, a value of 100.0 means to relax the temperature in a timespan of (roughly) 100 time units (\(\tau\) or fs or ps - see the units command). The damp factor can be thought of as inversely related to the viscosity of the solvent. I.e. a small relaxation time implies a high-viscosity solvent and vice versa. See the discussion about \(\gamma\) and viscosity in the documentation for the fix viscous command for more details.

The random # seed must be a positive integer. A Marsaglia random number generator is used. Each processor uses the input seed to generate its own unique seed and its own stream of random numbers. Thus the dynamics of the system will not be identical on two runs on different numbers of processors.

The keyword/value option pairs are used in the following ways.

The keyword angmom and omega keywords enable thermostatting of rotational degrees of freedom in addition to the usual translational degrees of freedom. This can only be done for finite-size particles.

A simulation using atom_style sphere defines an omega for finite-size spheres. A simulation using atom_style ellipsoid defines a finite size and shape for aspherical particles and an angular momentum. The Langevin formulas for thermostatting the rotational degrees of freedom are the same as those above, where force is replaced by torque, m is replaced by the moment of inertia I, and v is replaced by omega (which is derived from the angular momentum in the case of aspherical particles).

The rotational temperature of the particles can be monitored by the compute temp/sphere and compute temp/asphere commands with their rotate options.

For the omega keyword there is also a scale factor of \(\frac{10.0}{3.0}\) that is applied as a multiplier on the \(F_f\) (damping) term in the equation above and of \(\sqrt{\frac{10.0}{3.0}}\) as a multiplier on the \(F_r\) term. This does not affect the thermostatting behavior of the Langevin formalism but ensures that the randomized rotational diffusivity of spherical particles is correct.

For the angmom keyword a similar scale factor is needed which is \(\frac{10.0}{3.0}\) for spherical particles, but is anisotropic for aspherical particles (e.g. ellipsoids). Currently LAMMPS only applies an isotropic scale factor, and you can choose its magnitude as the specified value of the angmom keyword. If your aspherical particles are (nearly) spherical than a value of \(\frac{10.0}{3.0} = 3.\overline{3}\) is a good choice. If they are highly aspherical, a value of 1.0 is as good a choice as any, since the effects on rotational diffusivity of the particles will be incorrect regardless. Note that for any reasonable scale factor, the thermostatting effect of the angmom keyword on the rotational temperature of the aspherical particles should still be valid.

The keyword scale allows the damp factor to be scaled up or down by the specified factor for atoms of that type. This can be useful when different atom types have different sizes or masses. It can be used multiple times to adjust damp for several atom types. Note that specifying a ratio of 2 increases the relaxation time which is equivalent to the solvent’s viscosity acting on particles with \(\frac{1}{2}\) the diameter. This is the opposite effect of scale factors used by the fix viscous command, since the damp factor in fix langevin is inversely related to the \(\gamma\) factor in fix viscous. Also note that the damping factor in fix langevin includes the particle mass in Ff, unlike fix viscous. Thus the mass and size of different atom types should be accounted for in the choice of ratio values.

The keyword tally enables the calculation of the cumulative energy added/subtracted to the atoms as they are thermostatted. Effectively it is the energy exchanged between the infinite thermal reservoir and the particles. As described below, this energy can then be printed out or added to the potential energy of the system to monitor energy conservation.


This accumulated energy does NOT include kinetic energy removed by the zero flag. LAMMPS will print a warning when both options are active.

The keyword zero can be used to eliminate drift due to the thermostat. Because the random forces on different atoms are independent, they do not sum exactly to zero. As a result, this fix applies a small random force to the entire system, and the center-of-mass of the system undergoes a slow random walk. If the keyword zero is set to yes, the total random force is set exactly to zero by subtracting off an equal part of it from each atom in the group. As a result, the center-of-mass of a system with zero initial momentum will not drift over time.

The keyword gjf can be used to run the Gronbech-Jensen/Farago time-discretization of the Langevin model. As described in the papers cited below, the purpose of this method is to enable longer timesteps to be used (up to the numerical stability limit of the integrator), while still producing the correct Boltzmann distribution of atom positions.

The current implementation provides the user with the option to output the velocity in one of two forms: vfull or vhalf, which replaces the outdated option yes. The gjf option vfull outputs the on-site velocity given in Gronbech-Jensen/Farago; this velocity is shown to be systematically lower than the target temperature by a small amount, which grows quadratically with the timestep. The gjf option vhalf outputs the 2GJ half-step velocity given in Gronbech Jensen/Gronbech-Jensen; for linear systems, this velocity is shown to not have any statistical errors for any stable time step. An overview of statistically correct Boltzmann and Maxwell-Boltzmann sampling of true on-site and true half-step velocities is given in Gronbech-Jensen. Regardless of the choice of output velocity, the sampling of the configurational distribution of atom positions is the same, and linearly consistent with the target temperature.

Styles with a gpu, intel, kk, omp, or opt suffix are functionally the same as the corresponding style without the suffix. They have been optimized to run faster, depending on your available hardware, as discussed on the Accelerator packages page. The accelerated styles take the same arguments and should produce the same results, except for round-off and precision issues.

These accelerated styles are part of the GPU, INTEL, KOKKOS, OPENMP, and OPT packages, respectively. They are only enabled if LAMMPS was built with those packages. See the Build package page for more info.

You can specify the accelerated styles explicitly in your input script by including their suffix, or you can use the -suffix command-line switch when you invoke LAMMPS, or you can use the suffix command in your input script.

See the Accelerator packages page for more instructions on how to use the accelerated styles effectively.

Restart, fix_modify, output, run start/stop, minimize info

No information about this fix is written to binary restart files. Because the state of the random number generator is not saved in restart files, this means you cannot do “exact” restarts with this fix, where the simulation continues on the same as if no restart had taken place. However, in a statistical sense, a restarted simulation should produce the same behavior.

The fix_modify temp option is supported by this fix. You can use it to assign a temperature compute you have defined to this fix which will be used in its thermostatting procedure, as described above. For consistency, the group used by this fix and by the compute should be the same.

The cumulative energy change in the system imposed by this fix is included in the thermodynamic output keywords ecouple and econserve, but only if the tally keyword to set to yes. See the thermo_style page for details.

This fix computes a global scalar which can be accessed by various output commands. The scalar is the same cumulative energy change due to this fix described in the previous paragraph. The scalar value calculated by this fix is “extensive”. Note that calculation of this quantity also requires setting the tally keyword to yes.

This fix can ramp its target temperature over multiple runs, using the start and stop keywords of the run command. See the run command for details of how to do this.

This fix is not invoked during energy minimization.


For gjf do not choose damp=dt/2. gjf is not compatible with run_style respa.


The option defaults are angmom = no, omega = no, scale = 1.0 for all types, tally = no, zero = no, gjf = no.

(Dunweg) Dunweg and Paul, Int J of Modern Physics C, 2, 817-27 (1991).

(Schneider) Schneider and Stoll, Phys Rev B, 17, 1302 (1978).

(Gronbech-Jensen) Gronbech-Jensen and Farago, Mol Phys, 111, 983 (2013); Gronbech-Jensen, Hayre, and Farago, Comp Phys Comm, 185, 524 (2014)

(Gronbech-Jensen) Gronbech Jensen and Gronbech-Jensen, Mol Phys, 117, 2511 (2019)

(Gronbech-Jensen) Gronbech-Jensen, Mol Phys (2019); https://doi.org/10.1080/00268976.2019.1662506