[gmx-users] shell water and epsilon

David van der Spoel spoel at xray.bmc.uu.se
Fri May 13 10:56:53 CEST 2005


On Fri, 2005-05-13 at 03:53 -0400, David L. Bostick wrote:
> This will most likely (effectively) be answered by David vdS,
> 
> I have been interested in using a polarizable water model via gmx, and have
> been looking at similar models to the shell model in gmx.  I was looking at
> the Drude model of Roux
> and co-workers in J. Chem. Phys. (2003) 119(10) p. 5185.  I noticed their
> dielectric constant appears to have converged very quickly compared to the
> rigid shell model in the publication.  Is there any rationale to explain
> why? Have there been any improvements/changes on the shell model? In the
> publication, it is mentioned that the model may not be sticky enough ... is
> there any way to improve it, or is the model tweaked as far as it can go?
I don't understand the question I'm afraid...
Do you mean Fig. 2?

One thing that is different from our SW models is that they do not
minimize the shell positions at each time step, but rather give it a
mass of 0.4 amu. With a timestep of 1 fs it seems that the shells will
not be very well optimized, i.e. the model violates the Born-Oppenheimer
approximation. 

I'm attaching the itp file, with a real shell that is minimized at each
time step. If you want to reproduce the original work you can change the
particle type to A and give it a small mass.

You will need yesterdays CVS code to run it...
-- 
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
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

-------------- next part --------------
; 
; Topology file for SSWM4-DP
;
; G. Lamoureux, A. D. MacKerell, Jr., B. Roux
; A simple polarizable model of water based on classical Drude oscillators
; J. Chem. Phys.,  119 (5185-5197), 2003
;
; Possible defines that you can put in your topol.top:
; -DRIGID       Rigid model (flexible is default)
; -DPOSRES      Position restrain oxygen atoms
;

[ defaults ]
LJ	Geometric

[ atomtypes ]
;name        mass      charge   ptype   c6	c12
   WO    15.99940       0.0     A   	0.0	0.0
   WH     1.00800       0.0     A	0.0	0.0
   WS     0.0           0.0     S   	0.0	0.0
   WD	  0.0		0.0	D   	0.0	0.0

[ nonbond_params ]
WO      WO      1       3.561658e-03	3.685198e-06

;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
;; This is a the 'classical YAW' model, in which we do have the dummy.
;; The shell is attached to the dummy, in this case the gas-phase
;; quadrupole is correct. Water_pol routine can be used for this
;; model. This has four interaction sites.
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
[ moleculetype ]
; molname	nrexcl
SW		2

[ atoms ]
; id	at type	res nr 	residu name	at name		cg nr	charge
1	WO	1	SM2		OW1		1      -1.77185
2	WH	1	SM2		HW2		1	0.55370
3	WH	1	SM2		HW3		1	0.55370
4	WD	1	SM2		DW		1      -1.10740
5	WS	1	SM2		SW		1       1.77185

[ polarization ]
; See notes above.	alpha (nm^3)
1	5	1 	0.00104252

;[ settles ]
; i	funct	dOH	dHH
;1	1	0.1	0.16333

[ constraints ]
; i     funct   doh     dhh
1       2       1       0.09572
1       3       1       0.09572
3       2       1       0.15139

[ dummies3 ]
; The position of the dummies is computed as follows:
;
;		O
;  	      
;	    	D
;	  
;	H		H
;
; 2 * b = distance (OD) / [ cos (angle(DOH)) 	* distance (OH) ]
;	  0.0137408 nm	/ [ cos (104.52 / 2 deg) * 0.09572 nm	]
;         0.01557 nm 
; Dummy pos x4 = x1 + a*(x2-x1) + b*(x3-X1)
;
; Dummy from			funct	a		b
4	1	2	3	1	0.2031807494	0.2031807494

[ exclusions ]
; iatom excluded from interaction with i
1	2	3	4	5
2	1	3	4	5
3	1	2	4	5
4	1	2	3	5
5	1	2	3	4

#ifdef POSRES
; Restrain the oxygen...
[ position_restraints ]
; iatom type	fx	fy	fz
1	1	100	100	100
#endif



More information about the gromacs.org_gmx-users mailing list