Re: zeroMomentum for water droplet

From: David Hardy (
Date: Wed Dec 04 2013 - 09:32:41 CST

Dear Mihaela,

The zero momentum option, which subtracts the mass-weighted averaged force at every step, is intended to fix the lack of conservation of linear momentum introduced by grid-based long-range electrostatic solvers, like PME. This approach is not compatible with simulation techniques that do not already conserve linear momentum, such as the use of fixed atom restraints.

Best regards,

On Dec 3, 2013, at 10:38 PM, Axel Kohlmeyer wrote:

> On Wed, Dec 4, 2013 at 12:38 AM, Mihaela Drenscko <> wrote:
>> Hi,
>> I found that if I use zeroMomentum for water droplet on graphite, I can
>> prevent drifting of water molecules towards adjacent cells. I keep the
> this would be a *very* bad idea. zeroing the total momentum only
> really is useful for a bulk liquid system to remove a drift which can
> lead to the so-called "flying ice cube syndrome". if you would enable
> this for a droplet you get very unphysical behavior. if a water
> molecule will (partially) leave the droplet, it will be subject to
> huge momentum fluctuation because of leverage imposed by this
> algorithm.
> if you want to confine water molecules you will be best of to define a
> reflective boundary wall using tclBC.
> axel.
>> Carbon atoms fixed, that why when I run simulation with namd, I get the
>> following message:
>> ------------- Processor 0 Exiting: Called CmiAbort ------------
>> Reason: FATAL ERROR: Cannot zero momentum when fixed atoms are present.
>> Charm++ fatal error:
>> FATAL ERROR: Cannot zero momentum when fixed atoms are present.
>> Aborted (core dumped)
>> How do I fix it?
>> Thank you.
>> Mihaela Drenscko
> --
> Dr. Axel Kohlmeyer
> International Centre for Theoretical Physics, Trieste. Italy.

David J. Hardy, Ph.D.
Theoretical and Computational Biophysics
Beckman Institute, University of Illinois

This archive was generated by hypermail 2.1.6 : Wed Dec 31 2014 - 23:21:58 CST