Re: Open-Shell Total Energies

Posted by Per-Olof Widmark on October 29, 2001 at 15:56:31:

In Reply to: Open-Shell Total Energies posted by Yannis G. Lazarou on October 29, 2001 at 13:50:51:

: Hello everybody!
: Can anyone tell me if it is possible with MOLCAS 5 to
: perform calculations on Open-Shell Species and get
: meaningful results? I have recently tried to use MOLCAS
: and I took some strange results:
: The Hartree-Fock energies of open-shell species such as
: H, C (triplet), F atoms and CH2F radical deviate by as
: large as 0.14 a.u. (max. dev. for H and CH2F) from the
: corresponding values by using Gaussian (with the
: cc-pVTZ basis set in all cases and C1 symmetry).
: In the MOLCAS manual (SCF program)it is explicitly
: written that MOLCAS is not able to perform ROHF
: calculations (neither UHF, I suppose...)
: Since the HF step is required for the higher electron-
: correlation level calculations (e.g. CCSD(T), MCPF),
: are the resultant total correlated energies of
: Open-Shell species meaningful?

: Thanks in advance for any replies.

For restricted open shell calculations you have to use the rasscf program and
define the active space such that you have one configuration only. With this
wavefunction you can proceed to make, for axample, CCSD calculations.

Follow Ups:

Post a Followup

Name:
E-Mail:

Subject:

Comments:
: : Hello everybody! : : Can anyone tell me if it is possible with MOLCAS 5 to : : perform calculations on Open-Shell Species and get : : meaningful results? I have recently tried to use MOLCAS : : and I took some strange results: : : The Hartree-Fock energies of open-shell species such as : : H, C (triplet), F atoms and CH2F radical deviate by as : : large as 0.14 a.u. (max. dev. for H and CH2F) from the : : corresponding values by using Gaussian (with the : : cc-pVTZ basis set in all cases and C1 symmetry). : : In the MOLCAS manual (SCF program)it is explicitly : : written that MOLCAS is not able to perform ROHF : : calculations (neither UHF, I suppose...) : : Since the HF step is required for the higher electron- : : correlation level calculations (e.g. CCSD(T), MCPF), : : are the resultant total correlated energies of : : Open-Shell species meaningful? : : Thanks in advance for any replies. : For restricted open shell calculations you have to use the rasscf program and : define the active space such that you have one configuration only. With this : wavefunction you can proceed to make, for axample, CCSD calculations.