[gmx-users] Model N2 molecules

[David van der Spoel]

On 2012-01-20 14:04, Kiwoong Kim wrote:
> Hi,
>
> I don’t know how to model N2 molecule yet.
>
> If I use virtual site for N2 model, then, this can be done like below.
>
> Hence, where do I have to apply plus charge for making N2 neutral??
>
> Can it be D1 and D2 for plus charge??
>
> Thx.
>
> [ atomtypes ]
>
> ; type mass charge ptype sigma epsilon
>
> D1 14.00674 0.0000 A 0.0000 0.00000000
>
> D2 14.00674 0.0000 A 0.0000 0.00000000
>
> N 0.0000000 -0.4048 V 0.332 0.3026
>
>
>
>   [ moleculetype ]
>
>   ; name  nrexcl
>
>   N2         2
>
>
>
>   [ atoms ]
>
>   ;   nr  type    resnr   residu  atom    cgnr    charge        mass
>
>   1       D1       1       N2    D1       1        0.0000  14.00674
>
>   2       D2       1       N2    D2       1        0.0000  14.00674
>
>   3       N        1       N2    N        1       -0.4048  0.0000000
>
>   4       N        1       N2    N        1       -0.4048  0.0000000

What is the physical meaning of a N2 molecule with net charge -0.8 ?
For a physical model with three particles I would suggest to make the 
atoms +0.1 or so and the central virtual particle -0.2. This is 
realistic since the triple bond between the atoms yields a significant 
electron density. In order to tune the charges you should compute the 
quadrupole moment and compare that to high level AI calculations or 
experimental data.

You might want to check this reference for a more complex N2 model:

Jordan et al. J. Chem. Phys. 103 (1995) p. 2272-2285.

>
>
>
> [ constraints ]
>
>   ;  ai  aj funct           c0           c1
>
>   1       2     1          0.1098
>
>
>
>   [ virtualsites2]
>
>   ;  ai    aj    ak       funct   c0      c1
>
>        3     1     2       1       -0.5
>
>        4     1     2       1        0.5
>
>
>
>   [ exclusions ]
>
> 3 4
>
> 4 3
>
> Below is the previous post message
>
> On 20/01/2012 5:51 PM, Kiwoong Kim wrote:
>
>     Hi,
>
>     I asked this question before.
>     Thanks to Justin, he did reply as follows.
>
>     Kiwoong Kim wrote:
>
>          Hi,
>
>          I'm getting better to use Gromacs owing to many posts on this sites :) .
>
>          I have several questions about adding the counter-ions.
>
>
>          My system has a number of N2 molecules which has charge,
>          -0.40484(for single N) X 2.
>
>          Thus, I have to add some counter-ions to make the system neutral.
>
>
>          However, because I'm newbie on Gromacs, I thought of several
>          clumsy ways myself.
>
>          #1. Add virtual sites (virtual atoms) which has counter-ions like below.
>
>          [ atomtypes ]
>          ; name  mass   charge  ptype  sigma  epsilon
>          DUM     0        0.80968 V    0.0    0.0
>
>
>          I set the coordinates of each virtual DUM atoms to the center of
>          N2 molecules.
>
>          #2. using genion in Gromacs.
>
>
>          But, I have no idea on this. What molecules do I have to designate
>          to charge plus ion using genion ??
>
>          I typed below line.
>
>          genion -s md_0_1.tpr -n index.ndx -o ton_genion.pdb -g
>          md_0_1_genion.log -p topol_genion.top -np 322 -pname dum -pq 0.80968
>
>          and selected N2 molecules which is diffusing particle.
>
>
>          It results that the name of half of N2 molecules is changed as
>          DUM. (maybe the system become neutral)
>
>          The number of N2 molecules should be fixed.
>
>          Do I have additional N2 molecule for charging using genion?? Then,
>          what are the initial coordinates???
>
>          please help me....
>
>          any advises would be helpful.
>
>
>     How can I do that??? I see no reason why you should do either. For
>     dinitrogen, which has no net dipole, it seems intuitive to me that
>     both N atoms should have zero charge.
>
>     -Justin
>
>     I made linear rigid model (CO2) by introducing virtual interaction. However, I
>     have troubles in modelling N2 molecules.
>     I want to model N2 with quadrupole moment according to the paper (Adsorption
>     (2007) 13: 469-476).
>     In the paper, the charge N2 model (originally developed by Murthy et al. (1980)
>     can be described as follows.
>     N (-0.40484), Dummy (N2, +0.80968).
>
> Right... that's totally different from N2 with charged N atoms :)
>
>
>     Actually, I don't know how to model N2 molecules with dummy. Please let me know
>     any advises.
>
> Sounds like virtual site type 2 is perfect - and didn't you already use
> that for CO2? See section 4.7 and somewhere in chapter 5.
>
> Mark
>
>
>     Thx. Have a great day. :)
>
>
>
>
>
>
>


-- 
David van der Spoel, Ph.D., Professor of Biology
Dept. of Cell & Molec. Biol., Uppsala University.
Box 596, 75124 Uppsala, Sweden. Phone:	+46184714205.
spoel at xray.bmc.uu.se    http://folding.bmc.uu.se