Re: [ESPResSo-users] directional lennard jones

[Tristan Bereau]

Dear Mohammad,

That's a standard limitation of the potential: you can't avoid infrequent highly energetic configurations. That, as you say, comes from nearby particles that are at first very close but at specific angles such that the interaction is turned off. If the system reorients itself such that it turns on the interaction, you may hit the repulsive part of the potential with unreasonably large forces. The way I deal with it is to apply a high force capping to that interaction throughout the simulation. The capping is high enough that it mostly doesn't affect the dynamics--except during these unlikely events when it prevents the system from blowing up. Use the command "inter ljangleforcecap" (see the manual). Note that this command *only* applies to the LJANGLE interaction.

Best,
Tristan

On Sun, Jul 15, 2012 at 12:29 PM, mehdi ezatabadi <address@hidden> wrote:

Hi

Sorry to interrupting you again
I am using the directional lennard jones potential to simulate  H-bond. I have some problem in implementation of this potential. Using this problem leads to break of my bonded interactions like Fene bond. First to rectify this problem I tried to use smaller time step but after some steps My bonded interactions that is in neighborhood of none bonded directional lennard jones will break. This problem does not exist for other non-bonded potentials like usual lennard jones. Changing the sigma of lj-angle or reduction the number of particles would not solve this problem.

This potential lets the particles that their orientations are different from the sample particle to go so near ..... and the little changement of the orientation lead to
enormous force at the time and more than bonding potential. I dont know how to rectify this problem. I will be thankful to know your experience and suggestions

Is there any one be familiar with this potential ?

Would you have some problems like mine in implementation of this potential ?

Thanks,Mohammad