Fix freeze requires atom attribute torque (../fix_freeze.cpp:34).

LAMMPS LAMMPS Mailing List Mirror
1

Dear All users,

I have a problem with freeze fix. this is my input file:

#-------------------------- This input is related to 300K run ------------------------------------

– Init Section –

units metal
dimension 3
boundary p p p
atom_style hybrid full
pair_style hybrid buck/coul/long 10.1 11 lj/cut/coul/long 11 11
bond_style hybrid morse harmonic
angle_style harmonic
dihedral_style opls
kspace_style ewald 1.0e-6

kspace_modify mesh 40 10 10 order 7

pair_modify mix geometric tail yes

– Atom Definition Section –

Here I just read the raw coordinate data generated by moltemplate.sh.

(If I were more careful, I would minimize the system before the run command.)

####reading my system data

– Settings Section –

include system.in.settings

– Run Section –

1 C3

2 Ca

3 Cc

4 H3

5 Oc

neighbor 1 bin
neigh_modify check yes every 1 delay 0 exclude type 3 5
region boundary block -19.443871 -5 -23.882012 22.023986 -26.467619 25.029379 side in units box
group A region boundary
timestep 0.001 # <- can be increased to 5.0 or 10.0 if careful
dump 1 all custom 4000 traj_nvt.lammpstrj id mol type x y z ix iy iz
compute 1 mobile temp
compute 2 all pressure 16

fix shake all shake 0.0001 20000 0 b 1

fix nose all temp/rescale 1 300 300 0.05 1

fix beren all press/berendsen iso 1 1 .01

fix timintg all nve

fix fxnvt all nvt temp 10.0 10.0 0.1
fix frozen A freeze

thermo 100 # time interval for printing out “thermo” data
thermo_style custom step temp pe etotal press vol

run_style verlet
run 10000

write_restart system_after_nvt.rst
print “All done”

2

No, because this fix zero out the force and torque "on a granular
particle", which you don't have in your system. Why not just use fix
rigid, not even not integrating the group at all?

Ray

3

Dear Ray,

Thank you very much for your reply.

I have used fix rigid.

4

If by “freeze” you mean hold
motionless, than fix rigid is
overkill. Fix setforce 0.0 or simply
not integrating those atoms will

do the same.

Steve

5

Dear Steve,

Thank you very much for your reply.

I want to hold them motionless and as a united surface while they have the temperature similar to other movable atoms and groups.

Besides I want to perform an NVT ensemble. Is it possible?

Thanks in advance.

6

Dear Steve,

Thank you very much for your reply.

I want to hold them motionless and as a united surface while they have the
temperature similar to other movable atoms and groups.

this is plain impossible. motionless atoms have a temperature of 0K.

Besides I want to perform an NVT ensemble. Is it possible?

it *cannot* be an NVT ensemble. that would require a bulk system in
the first place.

axel.

7

Dear Axel,

Thank you very much for your reply.

8

Dear Axel,

I have a question based on your reply. What are the differences between rigid, freeze, and ignoring atoms for integrating?

All of them make the atoms motionless. We can specify the movable atoms by specifying groups.

Thanks in advance.

9

Dear Axel,

I have a question based on your reply. What are the differences between
rigid, freeze, and ignoring atoms for integrating?

you forgot in your list the option of setting the velocity to zero and
using fix setforce 0.0 0.0 0.0.

see below for fix rigid. fix freeze is needed for extended particles,
e.g. in granular media simulations, since such particles can have
rotations as well, hile typical atomistic simulations have point
particles, which cannot have rotations by there very nature. with this
aside the difference is obviously whether you do integrate the
equations of motion for those immobile particles or not. the visible
result on the trajectory would be the same (atoms don't move), but the
circumstances of the model may require you to prefer one option over
the other.

also, you didn't consider in your list the option to have atoms not
move *on average* (through using fix spring/self), which is very
different as well and - judging from the trajectory of this discussion
- may be the option that you *really* want.

All of them make the atoms motionless. We can specify the movable atoms by
specifying groups.

it would help a lot, if you simply properly read the documentation and
not guess what something does by its name.
fix rigid does *not* make atoms motionless. it moves groups of atoms
as rigid particles, that is not at all motionless. you can ask all the
people around here that use fix rigid for the simulation of CO2
molecules in the gas phase, for example. they are quite mobile...

axel.

10

Dear Axel,

Thank you for your complete reply.