[gmx-users] Re: Re: David, doubt about the definition of polarization.

[David van der Spoel]

Yinghong wrote:
> David:
>  
>  >/ Yinghong wrote:
> />/
> />>/ David:
> />>/
> />>/  >/ Dear Dr. David:
> />>/ />/
> />>/ />/ According to the shell water model, I used this method to another
> />>/ kind of molecule, which is composed of 6 atoms (e.g. benzene).
> />>/ Initially, I put a dummy and shell particle (a small mass is given to
> />>/ shell, and doing a normal dynamcis) in the center of this hexagon, in
> />>/ which shell particle is connected to dummy through your defined
> />>/ isotropic polarization method.
> />>/ />/
> />>/ />/ Theoretically, polarization can be looked as a spring-like
> />>/ connection with constant Kr = sqr(qS)/4*PHI*Epsilon*Alpha, and the
> />>/ distance between dummy and shell particle can be decided by rsd =
> />>/ 4*PHI*Epsilon*Alpha * E0 / qS. Is it right?
> />>/ />/
> />>/ />/ Now, in my simulation, I applied an external electric field along
> />>/ Z direction, and the interactions (vdws + coulomb) between shell
> />>/ particle and all the other atoms are exclued. (Of course, here, 
> What I /
>  >>/ did is only to make a test instead of a real case). Obviously, for
> />>/ dummy and shell particles, E0 is currently only referred to the
> />>/ external field, because local field is excluded.
> />>/ />/
> />>/ />/ Quantitively,  I set alpha = 0.3 nm^3, qS = 3.0e and E0 = 1.5
> />>/ V/nm, through "mdrun -debug", alpha and qS can be correctly output,
> />>/ and the calculated value for Kr = 4168 KJ/mol/nm^2 is also in the
> />>/ right way. After simulation, I used "g_dist" to check the distance
> />>/ between dummy and shell particle (rsd) under such electric field. But
> />>/ the calculted value for rsd is only 10 percent of the theoretical
> />>/ value although I have tried for many times.
> />>/ />/
> />>/ />/ So, Could you tell me some possible errors in my defined model,
> />>/ and why rsd can not approach to the theoretical value? What is the
> />>/ principle for GMX to calculate this rsd?
> />>/ />/
> />>/ /
> />>/  > Isn't the problem nm vs. Ångström?
> />>/ I am very sure it is not that problem. Upon the parameters mentioned
> />>/ above, rsd should be ~0.1nm theoretically, but my calculation gave a
> />>/ value of only 0.01nm. So, any other suggestion?
> />>/ 
> />>/ 
> />/
> />/
> />/ We have
> />/
> />/ F = q E = k r or
> />/ r = q E / k
> />/ r = 0.00108 (eV/kj/mol) nm
> />/   = 0.1 nm
> /
>  > Just realized that I repeated your calculation and got the same result.
>  > How about exclusions? Have you checked the tpr file for that?
>  
> Firstly, thanks for your calculating in person. Which result did you 
> get, 0.1nm or 0.01nm?
See above.

>  
> In my simulation, I did not define the exclusions in top file. Instead, 
> I defined two energy groups in mdp file: SHELL & Others. Here, I 
> wanna check whether the movement of shell particle is only related to 
> external field in the absence of any other non-bonded interations 
> between shell and other atoms. So, I defined "energygrp_excl = SHELL  
> Others" in mdp files. Is that right?
Maybe, but I'm not sure.


>  
> Besides, I checked the tpr file, which seems ok.
Does this mean that all exclusions were there?
Is there interaction energy between shell and others in the output edr file?

>  
>  
> Xie Yinghong
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> gmx-users mailing list
> gmx-users at gromacs.org
> http://www.gromacs.org/mailman/listinfo/gmx-users
> Please don't post (un)subscribe requests to the list. Use the 
> www interface or send it to gmx-users-request at gromacs.org.


-- 
David.
________________________________________________________________________
David van der Spoel, PhD, Assoc. Prof., Molecular Biophysics group,
Dept. of Cell and Molecular Biology, Uppsala University.
Husargatan 3, Box 596,  	75124 Uppsala, Sweden
phone:	46 18 471 4205		fax: 46 18 511 755
spoel at xray.bmc.uu.se	spoel at gromacs.org   http://xray.bmc.uu.se/~spoel
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++