UCCSD(T)-F12 with UHF reference

[Matthew Graneri]

HI there,

I just wanted to check something...

I'd like to run a CCSD(T)-F12 calculation using a UHF reference. Using the input:

memory,500,m
gthresh,energy=1.d-12,zero=1.d-22,oneint=1.d-22,twoint=1.d-22;
gthresh,optgrad=1.d-8,optstep=1.d-8;

angstrom
geometry={
O 0.000000000000 0.000000000000 -0.554771997800
F 0.000000000000 1.068922596900 0.233599084700
F 0.000000000000 -1.068922596900 0.233599084700

}

basis={
default,cc-pCVTZ-f12
}
set,charge=1
set,spin=1
uhf,accuracy=16,energy=1.0d-10
{CCSD(T)-F12,thrden=1.0d-12,thrvar=1.0d-10}
{optg,mppx=0,grms=1.d-8,srms=1.d-8}

the calculation starts without any error messages, and gives me these details:

 PROGRAM * CCSD (Unrestricted open-shell coupled cluster)     Authors: C. Hampel, H.-J. Werner, 1991, M. Deegan, P.J. Knowles, 1992

                           UCCSD-F12 implementation by G. Knizia and H.-J. Werner, 2008

                   Density fitting integral evaluation by F.R. Manby, 2003,2007, G. Knizia, 2010


 Generating basis VTZ-F12-JKFIT
  Atom=O        basis=aug-cc-pVTZ/JKFIT
  Atom=F        basis=aug-cc-pVTZ/JKFIT
 Number of basis functions:     312

 Generating basis CC-PCVTZ-F12_OPT
  Atom=O        basis=cc-pCVTZ-F12_OPT
  Atom=F        basis=cc-pCVTZ-F12_OPT
 Number of basis functions:     285

 Generating basis CC-PCVTZ-F12-MP2FIT
  Atom=O        basis=aug-cc-pwCVTZ/MP2FIT
  Atom=F        basis=aug-cc-pwCVTZ/MP2FIT
 Number of basis functions:     408


 Convergence thresholds:  THRVAR = 1.00D-10  THRDEN = 1.00D-12

 CCSD(T)     terms to be evaluated (factor= 1.000)

But, when I take a look at the results, the only energies I get are:

 RESULTS
 =======

  Reference energy                   -273.132598260137
  UCCSD singles energy                 -0.000000014085
  UCCSD pair energy                    -0.773131651354
  UCCSD correlation energy             -0.773131665438
  Triples (T) contribution             -0.035208033385
  Total correlation energy             -0.808339698823

  UHF-UCCSD energy                   -273.905729925575
  UHF-UCCSD[T] energy                -273.946735356472
  UHF-UCCSD-T energy                 -273.940046283858
 !UHF-UCCSD(T) energy                -273.940937958960

Is Molpro just calculating the UCCSD(T) energies without the F12 correction? Is CCSD(T)-F12 with UHF actually implemented in Molpro, or can you only use ROHF orbitals? If it's not implemented, I'm a little confused about why I didn't get any error message, and why the calculation went ahead in the first place...

Thanks!

Matthew