I have practically no experience with Car-Parinello, BO, or path integral Molecular dynamics. In principle I want to perform dynamics of a 10-atom organic system in a small box of water with periodic boundary conditions.
The main problem is how to start and solvate my molecule in a small box. Is it possible in CPMD package, or do I have to use another program to do so? Maybe I should first equilibrate my system with molecular mechanics and then use the box for CPMD calculations? My main problem is the technical side of such calculations and all the procedures/tools that are necessary.
I searched the internet but I haven't found anything that could help me. Maybe you have some hints about how to start or what to use? If such a tutorial does not exist, maybe someone could briefly describe what to do?
I will be grateful for any help. Thanks in advance!
Dear Rafal, CPMD is the appropriate program to use for such task, you need to have an initial atomic positions of your 10 atoms inside the box, then optimize the wave functions and finally you can perform an annealing of your ssystem at a given temperature or any kind of MD schemes. The attached pdf may help you to begin. Best
Thanks a lot! The tutorial looks pretty clear. What about preparing the box with the solute and the surrounding water molecules and finally how to properly choose the size of the box?
I have just found a nice paper on glycine in water solution:
Chem. Phys. 133, 114508 (2010); http://dx.doi.org/10.1063/1.3481576
They use a rather small cubic box of 9.85 A x 9.85 A x 9.85 A. The question is how to find out that such a box is sufficiently large, so that the solute does not interact with its own image? I think this has to be an optimal choice due to the costs of ab initio MD...
Sure, it has to be a compromise. In general, the size of the box has to be as small as possible because of the cost and large enough to avoid interaction of the molecule with its "copy" in the neighbour cell. To adjust the box size, I would choose possibly small one initially, optimize, extend the box, calculate energy, repeat the procedure until no energy change is observed (the neighbouring molecues do not interact).
Dear Rafał,
Adding a bit to the previous, very good suggestions: I/we often use the software 'packmol' for solvating a molecule (or mixtures of liquids), then classical molecular dynamics if there is just a reasonable force field available, and then the Car-Parrinello- or Born-Oppenheimer dynamics.
In my (uneducated) opinion the boxes of 10 A are historical, nowadays one should be able to reach larger cells (even if not much; we use 128 water molecules as "standard", 50 ps trajectories; in more complicated solutes and/or solvents possibly larger/longer). Usually it is not only the solute itself, but the solvent might have one or two clear solvation shells, and then it is tough to disentangle the periodic images.
Just my opinions; Good Luck. :)
Greetings from Zurich,
apsi
- Badges
- Science topic
- Similar topics
- Chemistry
- Theoretical Chemistry
- Computational Chemistry
- Molecular Dynamics