why symmetry contamination removed under C1 symm

[jxzou]

Hi,

Recently I performed a CASSCF(6,6) computation for the N2 molecule at d(N-N) at 1.0A, using basis set cc-pVTZ. I noticed that in the output, there are several remarks like

*** IN SYMMETRY 1 ORBITAL  3 SYMMETRY CONTAMINATION OF 0.821D-04 HAS BEEN REMOVED

But I used 'symmetry,nosym,noorient'. I read in the Molpro manual that by default symmetry contaminations are not removed. So what does the above remark mean?

Besides, I compared CASCI and CASSCF energies with those from OpenMolcas (using the same orbitals as initial guess in two software):

                           Molpro                            OpenMolcas

CASCI(6,6)    -109.07947055          -109.07922934

CASSCF(6,6) -109.08301400          -109.08301400

The CASSCF energy matches excellent. But CASCI does not. Is this due to removal of symmetry contaminations?

Thank you for any suggestions or answers. Related files are attached.

[Hans-Joachim Werner]

By default, the casscf in Molpro removes orbital rotations between orbitals that have different symmetries, for example sigma and delta orbitals. This is
seen by the SYMMETRY CONTAMINATION output lines. This feature can be disabled using the „noextra“ directive. However, this should have no effect in casci.

Reading in orbital coefficients from some source is dangerous. This will only work if (i) the order of the basis functions is identical; (ii) the normalization of the basis functions is identical. The normalization is not the same in Molcas and Molpro, and perhaps this is the reason for the difference.

Joachim Werner

> --
> You received this message because you are subscribed to the Google Groups "molpro-user" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to molpro-user...@googlegroups.com.
> To view this discussion on the web, visit https://groups.google.com/d/msgid/molpro-user/8d45845e-8a20-47b9-a0de-55a41c6e2e81n%40googlegroups.com.
> <N2_cc-pVTZ_rhf_gvb3_2CASSCF.zip>

[jxzou]

Prof. Werner, thank you very much!

I just tried 'NoExtra', it works. The CASCI energy now turns into -109.07922946 a.u., which is much more close to that of OpenMolcas. In fact, I found that 'NoExtra'  affects the CASCI energy not only in CASCI computation (DONT,ORBITAL) but also in the 0-th step of CASSCF computation.

I certainly agree with you on  'Reading in orbital coefficients from some source is dangerous', any tiny mistake will result in non-sense energy. So I had written several utilities (Jingxiang Zou / MOKIT · GitLab) to read and transfer MOs among software packages. These utilities carefully take care of the order of basis functions (up to H) and normalizations. I've tested the R(O)HF, UHF and CASSCF orbitals, they all works.

Again I want to thank you for your quick reply.