Hi Dongbo,
Thank you for your reply.
I used the crystal hydroxylated silica. I would like to know how you generate the bond connection for this system?
Thank you
Liyi
王冬博 <bo.wang94@…1193…> 于2019年5月9日周四 下午4:54写道：
Moltemplate can generate data file with everything you want, please try it.
Jun Zhou <jun.zhou@…3633…> 于2019年5月9日周四 下午5:33写道：
Moltemplate can generate data file with everything you want, please try it.
Thank you very much. I will try it.
Thanks for recommending moltemplate. However I should warn you that (as of 2019-5-10), moltemplate does not have the ability to automatically connect pairs of nearby atoms with bonds. This is fine if you are content to immobilize the atoms in the bulk (interior) of the SiO2 crystal (by omitting them from the group of atoms sent to “fix nvt” or “fix npt”). In that case, you don’t need to connect the Si atoms to the O atoms with bonds because they are not going to move.
However if you want to vibrations (phonons, movement of the atoms deep within of the SiO2 crystal), then you will either need to 1) add explicit covalent bonds to keep these atoms close together (using the ClayFF or CWCA (?) force fields or something similar), OR 2) use a “many-body” force field such as ReaxFF (“pair_style reax”) or Tersoff (“pair_style tersoff”) which does not require explicit bond constraints to hold the atoms in place. This is a choice you will have to make regardless of whether or not you want to use “moltemplate”. Tools like moltemplate, topotools (or even the simple commands that come with LAMMPS) can be used to build these kinds of crystals, assuming you do not need to automatically connect nearby atoms with bonds.
Similarly, once you have built the interior of the SiO2 crystal, it is not difficult to use moltemplate, topotools, or other lammps builder tools to cover the surface with a layer of silicon bonded to hydroxyl groups. If you are willing to immobilize these hydroxyl groups as well, then you do not have to worry about choosing realistic force-field parameters for this part of your simulation. You are done. (Or presumably, if you have water molecules, and salt ions, or other solvent molecules, then add those and you are done.)
On the other hand, if you want to allow the hydroxyl group atoms on the surface to move, you will also need to add bonds that connect the silicon atoms in the outermost layer with the nearby oxygen atoms in the crystal. This way, realistic bond-angle forces can be applied to prevent these hydroxyl-group atoms from flopping back and fourth loosely. Once you have decided what kind of bond-angle force constraint you want to use for these atoms, tools like moltemplate, or topotools can generate these bond-angle interactions automatically for all of the hydroxyl atoms on the surface of your SiO2 crysta. (Other tools likel EMC, or materials studio can do this as well, but I suspect they will probably choose the force field parameters for you instead of allowing you to choose what these parameters yourself.)
Alternatively, if you are using ReaxFF, then perhaps ReaxFF can be used to describe the forces acting on these hydroxyl group atoms (as well as the interior atoms)? In that case, you don’t need to add any bonds or bond-angle forces to your system at all. (In that case, I suppose the ReaxFF will effectively compute forces on these atoms which are somewhat similar to the forces that the bonds and bond-angle interactions would have applied. Forgive me, I have no experience using ReaxFF.)
In summary:
If you plan to use use ClayFF, or a similar force-field that requires explicit bonds between all adjacent atoms in the crystal, then it may be difficult to build your system with moltemplate. But it should be relatively easy if you are using ReaxFF, Tersoff, or are simply immobilizing the atoms in the crystal.
Incidentally, a discussion of how to build graphene-oxide came up several times recently. I realize some of the issues discussed there are not relevant to your question, but I thought I’d post a link to these discussions anyway.
https://sourceforge.net/p/lammps/mailman/message/36580756/
https://sourceforge.net/p/lammps/mailman/message/36612046/
(Apparently several people have tried using moltemplate to build graphene oxide surfaces.)
I hate to admit that moltemplate might not be the optimal tool to build these kinds of covalently bonded crystalline materials. This ended up being a long and confusing email. But I hope something I wrote was helpful at pointing you in a useful direction. If you have access to materials studio, then by all means use that.
Cheers
Andrew
Thank you very much for your detailed explanation. The system consists water and SiO4H wall .I only apply LJ potential for Si and O to describe the pair interaction between wall and water. Si and O atoms are located in fixed positions but the H atoms on the surface are able to move with fixed bond lengths (eg, 0.1nm) and bond angles,each H atom can reorient in the circle. Such circular motion occurs in a plane parallel to the plate at a distance 0.033 nm away from the O-atom plane of the Si-O-H group. So I need obtain the information: the nearby atom oxygen ID of H.
Andrew Jewett <jewett.aij@…24…> 于2019年5月11日周六 上午11:53写道：