[ase-users] Optimization with constraints using VASP

[Massimo Bocus]

Dear ASE users,

I am performing an optimization with constraints using vasp. In particular, I am constraining the bond lengths between three pairs of atoms to keep my system close to a transition state region using:

fix2 = FixBondLengths([[0,204],[204,209],[209,66]])

At the same time, I want to perform a scan of a dihedral angle, so I start by fixing it at a certain value 'i' using:

system.set_dihedral(66,196,197,206,i,indices=[206,207,208,215,216,217,218,219,220,221,222,223])
dihedral = [system.get_dihedral(66,196,197,206), [66,196,197,206]]
fix = FixInternals(dihedrals_deg=[dihedral], epsilon=1e-03)

Then I optimize using the BFGS optimizer. 

system.set_constraint([fix, fix2])
system.set_calculator(calc)
opt = BFGS(system, trajectory='pos.traj') 
opt.run(fmax=5.0E-3, steps=1000)

However, since one of the atoms involved in the transition state is a H, it experiences quite high forces (due to the fact that the ideal TS distance can change based on the movements of the molecule) that tend to increase during the optimization while the energy of the system decreases (see attached figures). This makes the optimization basically impossible to converge based on a force criteria only.
I was therefore wondering how the problem could be solved, maybe using some other convergence criteria or taking into account the constraint in the force calculation.

Many thanks in advance, 
Massimo
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Figure_2.png
Type: image/png
Size: 17928 bytes
Desc: Figure_2.png
URL: <http://listserv.fysik.dtu.dk/pipermail/ase-users/attachments/20210430/fb485307/attachment-0002.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Figure_1.png
Type: image/png
Size: 25561 bytes
Desc: Figure_1.png
URL: <http://listserv.fysik.dtu.dk/pipermail/ase-users/attachments/20210430/fb485307/attachment-0003.png>