ERROR : charge solution failed in setreaxffQiter

Hi,

I am new to GULP and trying to use ReaxFF to perform MD simulation using Material Studio.

In previous step, I perform a NPT dynamics to get the proper density of the system. Then I perform NVT dynamics to equilibrate the system without assigning velocity, but the initial stage shows extremely high temperature and the trajectory shows unreasonable movement.

So I added the final velocities from the previous step of density adjustment, but by doing so the program report ERROR : charge solution failed in setreaxffQiter and I have no clue what it suggest to be.

It would be more than appreciated if anyone could provide any kind help or instruction!

Best,
Tingxuan

p.s. Sorry I can only use this method to upload the input file, as I am new user that cannot upload a file and the full file is to long which there is a length limit here :frowning:

md conv conse qok nomod pres spat pot efg qiter 
ensemble nvt
tau_thermostat  0.1 ps

temperature 298.000000 K

equilibration 0.001 ps
timestep 0.1 fs
production 1 ps
write 100
sample 100


title
GULP calculation from Materials Studio for Cellubiose
end


cell
21.094300 24.041900 25.890000 92.351300 96.460200 97.279200

fractional
C1 core 0.186141 0.335796 0.120541 0.000000 1.000000 0.0 1 1 1
H1 core 0.175245 0.292751 0.132285 0.000000 1.000000 0.0 1 1 1
O1 core 0.160211 0.339541 0.066440 0.000000 1.000000 0.0 1 1 1
...

VELOCITIES
1 3.604728 -1.250909 3.044476
2 -0.595171 -3.385330 4.164299
3 3.177406 -0.598277 0.179269
...

Species
C1 core C
H1 core H
O1 core O

spacegroup
P 1

library reaxff_oxidative

dump Cellubiose.grs
output movie arc Cellubiose
mdarchive Cellubiose.arc
output trajectory Cellubiose.trg

This sounds like a standard MD issue of starting from a high energy structure. As the MD proceeds, a lot of excess energy gets dumped into the system and so unless the thermostat can remove it fast enough you end up with problems. It’s always worth running a bit of optimisation (even with conjugate gradients) before starting MD to try to remove any excess internal energy. This make the MD much more stable when it starts and hopefully keep the temperature in the region where it should be. The failure to converge the charges in ReaxFF is likely to be just a consequence of the geometry going crazy due to the excess energy.
Hope that helps,
Julian