Hello all,
I am trying to perform a simulation of CNT in an epoxy polymer system. I primarily simulate the bond formation between the resin and the hardener (using create bond command). The bond formation is performed under periodic BC. I then have an energy minimized, well equilibrated system and I write a data file at this point (with image flags) and a dump file with unwrapped coordinates (using dump xu yu zu).
At the next step (using the data file generated from the previous step), I want to perform a simulation with a non periodic boundary in x direction (s p p). When I change the boundary, I get a warning “Inconsistent image flags”. I loaded the resulting data file from the “s p p” simulation on VMD to see the configuration, and it looks like the atoms are mapped wrong - some atoms in the CNT chain appear broken and way outside the box. I checked the image flags in the new data file and they appear to be the same as the image flags in the previous data file.
I think I can fix the problem if I just write data file with purely unwrapped atom coordinates. Does write_data have the option to write unwrapped atoms coordinates (and avoid using image flags altogether)? If not, then is there an alternative way to fix the inconsistency in atom mapping?
Thank you for your time. Please let me know if more information is required to better understand this problem
Hello all,
I am trying to perform a simulation of CNT in an epoxy polymer system. I
primarily simulate the bond formation between the resin and the hardener
(using create bond command). The bond formation is performed under periodic
BC. I then have an energy minimized, well equilibrated system and I write
a data file at this point (with image flags) and a dump file with unwrapped
coordinates (using dump xu yu zu).
At the next step (using the data file generated from the previous step), I
want to perform a simulation with a non periodic boundary in x direction (s
p p). When I change the boundary, I get a warning "Inconsistent image
flags". I loaded the resulting data file from the "s p p" simulation on VMD
to see the configuration, and it looks like the atoms are mapped wrong -
some atoms in the CNT chain appear broken and way outside the box. I
checked the image flags in the new data file and they appear to be the same
as the image flags in the previous data file.
I think I can fix the problem if I just write data file with purely
unwrapped atom coordinates. Does write_data have the option to write
unwrapped atoms coordinates (and avoid using image flags altogether)? If
not, then is there an alternative way to fix the inconsistency in atom
mapping?
Thank you for your time. Please let me know if more information is required
to better understand this problem
-Nithya Subramanian
Arizona State University
your simulation protocol is flawed. why not set up a system with fixed
boundaries for the non-periodic dimension from the beginning and two
walls, do your bonding and equilibration and then remove the walls and
convert from fixed to shrinkwrapped? seems more straightforward and
cleaner than ripping a bulk system apart (which is what you are
doing).
axel.