espressomd-users
[Top][All Lists]

## [ESPResSo-users] WARNING: Major Change in Langevin Thermostat

 From: Joost de Graaf Subject: [ESPResSo-users] WARNING: Major Change in Langevin Thermostat Date: Sat, 18 Jul 2015 21:01:41 +0200

Dear all,

As of commit

Merged #358.

there are major changes in the ESPResSo Langevin thermostat. I know that some of you are wondering: "Will this effect my simulations if I pull a newer version of ESPResSo, some time in the future?"

The good news is, probably not: if you had not enabled the feature MASS and used non-unit-masses, the behavior should be exactly the same. The same goes for ROTATION and ROTATIONAL_INERTIA, provided you used a unit interia tensor.

So what has changed? We have repaired several bugs in the code and modified the behavior such that the 'friction coefficient' which you specify, is an actual friction coefficient and is mass and inertia independent. Before, if you changed the mass, the diffusion rate would also change, which is inappropriate if you consider that number to represent something like $6 \pi \eta R$ for the Stokes' friction of a sphere, for instance.

If you did have masses or rotational inertia, with a non-unit inertia tensor, you should verify if this version changes your results (which it should) and make modifications to the input accordingly.

There are also several neat new features in the update. You can now set the rotational and translational friction coefficient independently. Which means you can match to a sphere solution for which you get

$6 \pi \eta R$ (translation)
$8 \pi \eta R^{3}$ (rotation)

respectively. Also, by specifying two instances of 'on/off' keys, you can thermalize only the translational or the rotational modes, or both. This has applications in high-Peclet number active suspensions, for instance, where the self-propulsion dominates the translational diffusion, but the rotational diffusion might still be quite crucial.

We hope that the previous behavior of ESPResSo's thermostat has not severely inconvenienced you, and that the transition to the new behavior will be smooth. This change will be encorporated in the next major release of ESPResSo. Please let us know if there is anything you need help with, if the new behavior causes you problems.

Sincerely Yours,

Joost de Graaf
on behalf of the ESPResSo developer's team