Hi Lammps Users:
I’m a freshman to MD, now I want to do some simulation about Cu nanowires. The first problem for me is how to relax a circular nanowire with fcc lattice. In my scripts, r and l represents the radius and length respectively, T represents the temperature. I’d like to construct a nanowire in two different ways.
1.The first method is that the nanowire is the whole box, like:
boundary s s p
region wire cylinder z $r $r $r 0 $l
create_box 1 wire
create_atoms 1 region wire
(1)nvt ensemble
I tried nvt relaxation, the commands is:
fix nvt_relax2 all nvt temp $T 1000 0.005 drag 0.2
fix nvt_relax4 all nvt temp 1000 $T 0.005 drag 0.2
It seems there is large tensile force or something cause the nanowire to warp. This happened with both Cu Mishin eam potential or Cu Zhou eam potential, and the range of r is from 5 lattice constants to 30 lattice constants. I don’t know why this happened and whether this nvt relaxation method is proper.
(2)npt ensemble
I also tried npt relaxation with commands ‘fix npt_relax all npt temp $T 1000 0.1 z 0 0 1’, but the value of temperature, energy and press will be ‘Nan’ when fix npt starts.
When relax at constant temperature like ‘fix npt_relax all npt temp $T $T 0.1 z 0 0 1’ , Nan appears too if T equals 0.001K.
But command ‘fix npt_relax all npt $T $T 0.1 z 0 0 1’ works if T equals 300K.
2.The second method is that the nanowire is in a bigger box, like:
boundary p p p
region whole block -50 50 -50 50 0 $l
create_box 1 whole
region rwire cylinder z $r $r $r 0 $l
create_atoms 1 region rwire
(1)nvt ensmble
The relaxation command is same to the first method, and the results are similar too. Still, warp and distortion of the nanowire’s cross section seems this is not a good way to relax nanowire.
(2)npt ensemble
Similar to the first methond, commands ‘fix npt_relax all npt temp $T $T 0.1 z 0 0 1’ works if T equals 300K. But lammps shows error ‘Domain too large for neighbor bins (neighbor.cpp:1479)’ when T equals 0.001K.
And npt relaxation with increasing the temperature like command ‘fix npt_relax all npt temp $T 1000 0.1 z 0 0 1’ doesn’t work too.
I also tried minimize command before or after fix nvt or fix npt. But about 100 steps the minimization was finished. The difference of total energy is very small. I am not sure whether it is enough that only use minimize command to relax the system. Referring to some refs, I want to add fix npt or fix nvt command.
My question is which ensemble should I use to relax the nanowire, nvt or npt? How should I use them? How to avoid warp when using nvt or errors when using npt?
Any help would be much appreciated. Thanks very much for your help!
My Best Regards
Hi Lammps Users:
I’m a freshman to MD, now I want to do some simulation about Cu nanowires. The first problem for me is how to relax a circular nanowire with fcc lattice. In my scripts, r and l represents the radius and length
Three things:
It is not a good idea to try and learn MD from a mailing list. You need to work under the direct personal supervision of somebody that has experience in MD.
It is also not a good idea to start with a complex nanostructure. Learn how to reproduce some simple bulk simulations and work your way up to more complex setups.
For an isolated system you have neither an nvt or an npt ensemble. This topic has been discussed many times on this list. Please see my previous comments on the subject in the mailing list archives.
respectively, T represents the temperature. I’d like to construct a nanowire in two different ways.
Dear Axel:
Thanks very much for your reply!
Actually, I have tried several bulk simulations and learned some basic knowledge and skills. I also reproduce a work of nanowires, my result is similar to the publication's result. The relaxation method is that relax the nanowire using npt ensemble at constant temperature (300K), which is described in my previous email. But I am wondering whether the system can gain enough energy to get over the potential barrier to reach the most stable state. So I'd like to try nvt or npt ensemble with increasing temperature, but there is distortion of cross section or errors happened. I can not solve this problem now and post my questions for help.
Would you please give me more details?
I have searched on the mailing list, you have thousands of replies on the list. Would you please give me some keywords or guide to search for my problem?
Thanks a bunch!
With Best Regards
Dear Axel:
Thanks very much for your reply!
Actually, I have tried several bulk simulations and learned some basic
knowledge and skills. I also reproduce a work of nanowires, my result is
similar to the publication's result. The relaxation method is that relax the
nanowire using npt ensemble at constant temperature (300K), which is
described in my previous email. But I am wondering whether the system can
your previous description is unusable. there is far too little context
to work with. ...and as i already alluded to, there is no meaning to
use a (periodic) variable cell integrator (for bulk systems) on an
isolated object.
gain enough energy to get over the potential barrier to reach the most
stable state. So I'd like to try nvt or npt ensemble with increasing
temperature, but there is distortion of cross section or errors happened. I
i have no idea what you mean by this.
can not solve this problem now and post my questions for help.
Would you please give me more details?
as i mentioned before. this not really a LAMMPS question, but a
question about the science that you are doing and your adviser and
your (senior, more experienced) colleagues that know about your
project are the best people to discuss this with.
I have searched on the mailing list, you have thousands of replies on
the list. Would you please give me some keywords or guide to search for my
problem?
<alert>
the comment to this question has been censored by the mailing list decency board
</alert>
axel.