[gmx-users] oxydiffusion

[Sunil Thapa]

Respected experts
Thank you a lot for your valuable suggestions for my previous mail.
 
Taking your advice I used SPCE water model with ffg43a1 forcefield. I modified the non bonded parameters for oxygen-water interaction taking the parameters from CRC as a source in the nonbonded parameter file. 
 
In the non-bonded parameter section I entered the values for oxygen and water using Lorentz-Bertholet rule but forgot to enter the values for water and water oxygen interaction. With that change I prepared a system of one oxygen in 508 molecules of water inside a 2.5 nm cubic box. I saw that the density of the system was 997.8 kg/m3 close to what I want for 298 K. The system was energy minimized and I got a straight line having negative slope in the negative energy region with an average negative value.
 
Is my system not in local equilibrium because I suspect I should get a horizontal line in the negative region?
 
Thinking that i would give a try, I simulated the system directly in NVT ensemble at equilibration temperature 298 K (Berendsen). I ran the system for 100 ns. I fitted the data for first 50 ns and got the diffusion coefficient of 0.38 vs 2.4 unit. I have attached graph in this mail.
 
Is it due to insufficient statistics ? Would you please suggest me some reference that clarifies this insufficiencey. If then, should I extend this run to some 50 ns more?
 
Or is it due to the fact that I forgot to enter the water-water interaction parameter. The self diffusion coefficient of water was 2.6 unit versus 2.3 unit.
 
Also, the density calculated by g_energy is different from that we calculate manually (i.e., total mass divided by box volume) why?
 
Thank you for your consideration and patience.
 
Neal
 


      
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20100316/1461010b/attachment.html>