Hello!
I am trying to restrain the rotation of an SO4 cluster using a harmonic bias around the center (1, 0, 0, 0). I want to restrict the cluster about its initial orientation, but no matter what I do, it rotates first and then stops. The first quaternion is (0.605, 0.000, 0.563, 0.563), and it eventually approaches (1, 0, 0, 0). For my reference file, I just use the initial coordinates of the atom in the same order that they appear in my LAMMPS data input file. Is that wrong? If not, do you have any other ideas for what the issue could be?
I have also tried just implementing the restraint for the Oxygens, and it exhibits similar behavior. Any help is much appreciated, and I have pasted my files below!
here is my colvar.in file:
colvar {
name Orient0
orientation {
atoms {
atomNumbers 1 2 3 4 5
}
refPositionsFile ref.xyz
}
}
harmonic {
colvars Orient0
centers (1.0, 0.0, 0.0, 0.0)
forceConstant 10000000
}
my ref.xyz file:
5
ā# File I madeā
S 0.37406 3.58503 3.58503
O 0.13104 2.43206 2.43206
O 1.97427 3.86856 3.86856
O 6.75545 3.08470 4.98643
O 6.75545 4.98643 3.08470
and a clip of my lammps data input file:
Atoms # full
1 1 1 1.2 0.37406322117761903 3.585033905119269 3.5850339051192863 0 0 0
2 1 2 -0.8 0.13104212276656693 2.4320558626769717 2.4320558626767848 0 0 0
3 1 2 -0.8 1.9742741542602702 3.8685611624642804 3.86856116246994 0 0 0
4 1 2 -0.8 6.755453565927884 3.084701450834368 4.986434315070869 -1 0 0
5 1 2 -0.8 6.755453565934712 4.9864343150732084 3.0847014508313886 -1 0 0
1 Like
Hi @morgannguyen3917 what are the physical interactions present? Are there angles and torsions defined or are all the atoms free to move?
The algorithm used by Colvars to define rotations assumes that the group of atoms has a stable internal structure. If it doesnāt, then the algorithm (see the ref cited in the Colvars doc) is no longer numerically stable.
Giacomo
Hello Giacomo,
I appreciate your help and directing me to the reference. I do define my angles with the angle style Fourier. Would you recommend increasing my angle coefficients so the angles deviate less? Or is there something else?
Also by āstable internal structureā, do you only mean the angles and torsions or does bond length matter too?
Thank you so so much!
Hi Morgan, the basic idea is that there should be no swaps in position between the oxygen atoms.
E.g. if you look down the direction of the bond from O1 to the S atom, O2 O3 and O4 should remain in the same order. (The exact order, clockwise or counter-clockwise, depends on your system, but I believe you get the idea.)
Giacomo
That makes sense! Thank you for that explanation.
Without the COLVAR restriction, my oxygens donāt swap in position, so I feel like it has a stable internal structure? But when I apply the COLVAR restraint, the bond lengths increase, and then the Oxygens swap in position, indicating that I am using COLVARS incorrectly. Do you think it is an issue with my reference coordinates or something else?
Thank you!
Hi Morgan, without explicit potentials on the angles and torsions from the force field, the atoms will eventually swap. This may occur more quickly if you add restraints. Plus, as you pointed out, it is worth checking that the order of atoms in the simulated and reference coordinates match.
Just a check: is there a reason why using fix rigid or rigid/small wouldnāt work for this āmoleculeā?
Giacomo