Question on using MOPAC API

[rajgopals]

Greetings! I am trying to use the MOPAC API to run some computations. Since some of the systems I am trying to model are charged, I specified the charge using system%charge. However, while the program compiles, the reported heat of formation is 0.0, irrespective of the charge. If I omit the charge then the computation runs and the system is treated as having unpaired electrons. I have illustrated this in the attached 2 f90 files - one where the system is modeled as hydroxide (charge = -1) and the other as hydroxyl radical (no charge specified). The results from running the computation are as follows

$ compile_mopac hydroxyl mopac_test_pm7_oh_radical.f90
$ compile_mopac hydroxide mopac_test_pm7_oh.f90
$ ./hydroxyl
Using PM7
7.2626149968318714
$ ./hydroxide
Using PM7
0.0000000000000000

Is charge not supported by the API?

Thanks,
raj
files.zip

[godotalgorithm]

I certainly intend for the API to support charge, so this is definitely a bug. The API has better testing coverage than the rest of MOPAC, but it is still a new feature that has yet to be used extensively.

Fixing any API bugs that are discovered is a high priority for me, so I'll try to fix this and release a patch in the next few days.

[rajgopals]

Thanks a lot for the insanely quick response. While, we are at it, may I also add that the API does not support PM6-D3H4, while other models - namely PM7, PM6-ORG, PM3, AM1 and RM1 work as expected. I have illustrated this issue in the following 2 examples.

$ compile_mopac pm7 mopac_test_pm7.f90
$ ./pm7
Using PM7
-57.769745741145925
$ compile_mopac pm6d3h4 mopac_test_pm6d3h4.f90
$ ./pm6d3h4
Using PM6-D3H4
0.0000000000000000

Sorry to overload the discussion with these issues. If you prefer, I can raise these as issues.

Thanks,
raj
pm6d3h4issue.zip

[godotalgorithm]

It's fine to keep these bugs together, as they are probably related, and I'll transfer this from Discussions to Issues.

[rajgopals]

Thanks yet again.

raj

[godotalgorithm]

If you encounter anymore related API bugs before I fix these two, then post them in the Issue (#296).