Elastic constant for C-S-H in cement

Hi Aidan
I used the elastic example for C-S-H phase in cement. the box is triclinic.
All the minimization were converged but the Pressure increase from 30 to 2000 atm during “minimization” and “fix deform_box”.

Another thing is that Cij for negative and positive deformations are negative and positive receptively. and also Cij are very large. for example:
C11 pos = 153331Gpa
C11 neg = -152895Gpa

It is not normal.

Do have any idea what is my mistake?

Mehdi Eftekhari

C11pos and C11neg should both be positive. The fact that C11neg is negative suggests that Pxx is decreasing in both directions. This is opposite to what you state below about the pressure. I am guessing that in reality, you observed C11pos to be negative. Either way, the deformed state may be jumping to a new energy minumum, in which case the PE vs. strain function is effectively discontinuous, inconsistent with the formal definition of elastic constants. I would try two things:

  1. Do a better job of relaxing the reference state, including the cell dimensions.
  2. Use a smaller deformation

Aidan

There is the a coul/long pair style in my input script:

pair_style hybrid/overlay buck 3.0 sw coul/long 3.0
kspace_style ewald 0.0001
pair_coeff 1 1 buck 4369.01 0.29 0.0
pair_coeff 2 2 buck 1171.52 0.29 0.0
pair_coeff 3 3 buck 452.5051 0.29 0.0
pair_coeff 4 4 buck 21.228 0.35 0.0
pair_coeff 1 2 buck 1382.4811 0.29 0.0
pair_coeff 1 3 buck 3557.62 0.29 0.0
pair_coeff 2 3 buck 1848.7174 0.29 0.0
pair_coeff 3 4 buck 248.6324 0.29 0.0
pair_coeff 2 4 buck 430.6566 0.29 0.0
pair_coeff * * coul/long
pair_coeff * * sw Si.sw NULL Si O NULL

I read user manual and find that long/range couloumbic style may cause the pressure not to converge to zero.

do you have any comment how can I fix this?

There is the a coul/long pair style in my input script:

pair_style hybrid/overlay buck 3.0 sw coul/long 3.0
kspace_style ewald 0.0001
pair_coeff 1 1 buck 4369.01 0.29 0.0
pair_coeff 2 2 buck 1171.52 0.29 0.0
pair_coeff 3 3 buck 452.5051 0.29 0.0
pair_coeff 4 4 buck 21.228 0.35 0.0
pair_coeff 1 2 buck 1382.4811 0.29 0.0
pair_coeff 1 3 buck 3557.62 0.29 0.0
pair_coeff 2 3 buck 1848.7174 0.29 0.0
pair_coeff 3 4 buck 248.6324 0.29 0.0
pair_coeff 2 4 buck 430.6566 0.29 0.0
pair_coeff * * coul/long
pair_coeff * * sw Si.sw NULL Si O NULL

I read user manual and find that long/range couloumbic style may cause the
pressure not to converge to zero.
do you have any comment how can I fix this?

this looks quite strange. how did you test that his force field works
in the first place?

a 3 angstrom cutoff is quite short. have you checked how much of a
discontinuity in the forces/virial you have because of that?

axel.

I dont know how to do that?
would you please tell me which command should I use to calculate the discontinuity of the force/virial?
I searched the doc but couldn’t find it.

I dont know how to do that?
would you please tell me which command should I use to calculate the
discontinuity of the force/virial?
I searched the doc but couldn't find it.

you have to be a bit more creative. don't expect the documentation to
contain an immediate solution with a detailed description for all
problems.

it is easy to check this for the pairwise additive interactions. so if
you disable the stillinger-weber term in your compound potential, you
could just use pair_write and then look at the force and energy vs.
distance graphs in the generated tables.

another option is to pick some 2-3 atoms and generate a trajectory
moving them around in straight motions with a script in a supported
format and then use the rerun option to compute the forces and
energies and then post process the data into plots as well.

as i said, a 3 angstrom cutoff is highly suspicious. how did you come
up with these parameters?

axel.