Hi Marawan,
There are several issues.
1) Dipole moment magnitude
The ORCA module has a bug printing the dipole moment. Multiply with the conversion factor for Debye to au (2.5417) to get the correct number. Sorry for that. I will release a bugfix for this in the near future.
2) Dynamics
Your input file requests a calculation that differs from Gaussian. You did not provide the orca and gaussian outputs but I notice the SCF energy in the first MD step differs by 11 kcal/mol. Orca by default uses density fitting (RI approximation) for non-hybrid functionals. But it won't know which fit basis to use with 6-31G*. There might also be a difference in the use of Cartesian / spherical harmonic functions for d orbitals and higher. Finally, gradients will differ unless you increase XC quadrature grid accuracy. If you send me all input files you have been using I can have a look at this to clarify. Thanks.
As a note aside, for debugging it helps to try to generate identical trajectories - use identical restart files and MD settings that do not depend on random numbers (e.g. specify the random number seed if you are using Langevin dynamics). I did a short test with Gaussian / Orca for NMA in a water box and get comparable trajectories.
All the best,
Andy
On Dec 13, 2012, at 7:47 PM, marawan hussain wrote:
> Dear Andy,
> Attached is some outputs i got from two different runs with identical parameteres in QM/MM simulation, one with Gaussian-AMBER and the other one is with ORCA-AMBER. The dipole moment steadily decreases with time in ORCA-AMBER but almost very stable in Gaussian-AMBER. As if the QM system is steadily desolvated in ORCA-AMBER. Also, as an example i have included the H-bonds between the nucleoside and two representative water molecules hydrogens. starting from the ORCA run (from step 177) you will notice that the distance increases steeply with time, the plots are also telling the same story.
> My system is a nucleoside in a water box of (20 Angstrome). Only the nucleoside is treated as the QM system. I'm using the same QM level of theory, functional(Blyp) and convergence criteria, input, etc.
> If i run longer simulations i got the same behaviour.
>
> King regrads
> Marawan<Comp_QM-MM_Gaussian_ORCA.rar>_______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
--
Dr. Andreas W. Goetz
Assistant Project Scientist
San Diego Supercomputer Center
Tel : +1-858-822-4771
Email: agoetz.sdsc.edu
Web : www.awgoetz.de
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Dec 14 2012 - 02:00:02 PST
