Wrong elastic tensors

Hi dear LAMMPS Community,

I wish you all a great day. I’m having a bit of trouble with my simulations and could use some advice.

I’ve been working on simulating and calculating the elastic tensors of hcp titanium using the stable_2Aug2023 version of LAMMPS on Windows 11. The problem is that I’m consistently getting an elastic modulus of 140 GPa, while the literature suggests it should be around 170 GPa.

Initially, I used NPT and fix deform, but when I saw that the example code in the LAMMPS directory was working fine, I tried to replicate it and still got the same 140 GPa result. I even combined different simulation techniques, using NPT for stabilization and trying NVE, Langevin, and change_box for the tensile test, but the issue persists.

I’ve changed my potential file and simulation box multiple times, consulted with my instructor, and searched online for solutions, but no luck so far.

My code is attached. Does anyone have an idea what causes that issue?
in.elastic.lmp (3.7 KB)

Potential files:
ti.meam (248 Bytes)
library.meam (2.9 KB)

The structure:
Ti_hcp_0.5-3.lmp (20.2 KB)

And the log file:
log.lammps (160.6 KB)

Thanks in advance.

Were these results from computations with the exact same potential parameters?

Yes, they are. I got the potentials and the article from nist.gov.

Then you need to carefully compare your computations with what is is described in the article (and potentially its supporting information).
I would also look around for other publications on the same material. Not always are the results consistent and not all publications are done with sufficient care these days.

It is next to impossible to comment and correct calculations like this from the outside.
The only caveat we have with the example scripts provided with LAMMPS is that they are usually set up to run fast but do not always provide converged results, so you have make certain that you are getting properly converged results (also with respect to system size!).

If you cannot resolve the discrepancy, it is probably more helpful to try an contact the authors.

1 Like

Thank you for your insights Axel!

Hello dear community, I fixed the problem by changing the last “fix npt” to “fix langevin” and “fix nve.” I’m not exactly sure what was wrong, but this adjustment worked well and got the results pretty close to what’s described in the article. Also just “fix nvt” works fine. I hope that helps for whom is interested.