[ase-users] NPTBerendsen()

[Vitaly Vitalievich]

Adam, thank you for your insights.

Here is how the stress looks in MOPAC22. If you happen to know the
internals of ASE deeply, unlike me, it must be straightforward to code.
Also, the cell vectors-related procedure does not work with the ASE MD
functions.


          GRADIENT NORM           =       3903.37562          =
400.47787 PER ATOM
          IONIZATION POTENTIAL    =          5.403926 EV
          HOMO LUMO ENERGIES (EV) =         -5.404 -1.421
          NO. OF FILLED LEVELS    =         94
          MOLECULAR WEIGHT        =        436.8466         POINT GROUP:
 C1

          VOLUME OF UNIT CELL     =       1000.000 CUBIC ANGSTROMS

          DENSITY                 =          0.725 GRAMS/CC
                              A   =         10.000 ANGSTROMS
                              B   =         10.000 ANGSTROMS
                              C   =         10.000 ANGSTROMS
                            ALPHA =         90.000 DEGREES
                            BETA  =         90.000 DEGREES
                            GAMMA =         90.000 DEGREES

           Pressure required to constrain translation vectors
           Tv(  96)  Pressure:  16.32 GPa
           Tv(  97)  Pressure:  52.10 GPa
           Tv(  98)  Pressure: 136.90 GPa



*           Stress tensor in GPa using Voigt notation (xx, yy, zz, yz, xz,
xy):              -26.397   -22.259   -23.777     5.216    -4.175    -1.922*


On Mon, Oct 28, 2024 at 10:43 AM Adam Jackson <a.j.jackson at physics.org>
wrote:

> No, the temperature is as I've set it. I guess the Mopac calculator simply
> does not have its "get_stress()" function.
>
> This is correct, the ASE MOPAC calculator does not read stress from the
> output file. Perhaps it was not present (or difficult to read reliably)
> when it was written; I’ve just run a quick calculation on diamond and in
> that case MOPAC 22 did write a stress tensor in Voigt form that we might be
> able to work with. Perhaps somebody can create a Pull Request including a
> unit test; we do have MOPAC 22 on the test runners so once set up it can be
> checked frequently.
>
>
>
> I also found it necessary to set ‘GEOM-OK’ in the MOPAC parameters or the
> calculation tends to complain about the unit cell and quit early.
>
>
>
> All the best,
>
> Adam
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.fysik.dtu.dk/pipermail/ase-users/attachments/20241028/aba77367/attachment-0001.htm>