Molcas Forum

maheshg

Member

Registered: 2018-09-25

Posts: 9

[SOLVED] Problem in using 6-31G(d,p) with AMFI card

Dear All,

I am trying to calculate the triplet state energies relative to the singlet ground state at CASSCF/6-31G(d,p) level of theory. According to the manual, I have added AMFI keyword to the SEWARD input. The calculation exits with the following error in the log file.

     Basis set specifications :
      Symmetry species         a
      Basis functions           80

  Only real spherical harmonics allowed
  for AMFI.
--- Stop Module: seward at Mon Oct  8 10:55:20 2018 /rc=_RC_INPUT_ERROR_ ---

It seems AMFI only works with spherical Gaussian functions. By default, the s, p and d functions of 6-31G(d,p) basis set are in Cartesian, and this can be changed by using "spherical"  card in Basis set keyword. However, I could not find the correct usage of it. Any help regarding to solve the issue would be highly appreciated.

Here is the input which gives the above error.

>export MOLCAS_MOLDEN=ON
&GATEWAY
Coord
8
acrolein
O  1.5686705444  -0.1354553340  3.1977912036
C  -0.1641585340  0.2420235062  -0.0459895824
C  0.1137722023  -0.1389623714  1.3481527296
C  1.3218729238  0.1965728073  1.9959513294
H  2.0526602523  0.7568282320  1.4351034056
H  -0.6138178851  -0.6941171027  1.9113821810
H  -0.8171509745  1.0643342316  -0.2648232855
H  0.1260134708  -0.4020589690  -0.8535699812

BASIS=6-31G**
Group=Nosym
AMFI
&SEWARD
&SCF
&RASSCF
Nactel = 4 0 0
Inactive = 13
RAS2 = 4
Ciroot = 1 1 1

&CASPT2
Multistate= 1 1

>>COPY $Project.JobMix JOB001

&RASSCF
Spin = 3
Nactel = 4 0 0
Inactive = 13
RAS2 = 4
Ciroot = 4 4 1

&CASPT2
Multistate= 4 1 2 3 4

>>COPY  $Project.JobMix  JOB002

&RASSI
Nr of JobIph= 2 1 4; 1; 1 2 3 4
Spin
EJob

Thank You
Mahesh

Ignacio

Administrator

From: Uppsala

Registered: 2015-11-03

Posts: 1,085

Re: [SOLVED] Problem in using 6-31G(d,p) with AMFI card

Cartesian d functions for the 6-31G are hard-coded (based on the basis set name), and the "Spherical" keyword is only available for "native input". You have two options I can think of:

1. Input the basis set and geometry with the native format, instead of "xyz" (note that in the native format the default is no symmetry and bohrs):

&GATEWAY
Basis
  O.6-31G**
  O1 1.5686705444  -0.1354553340  3.1977912036 angstrom
  Spherical d
End of Basis
Basis
  C.6-31G**
  C2 -0.1641585340  0.2420235062  -0.0459895824 angstrom
  C3 0.1137722023  -0.1389623714  1.3481527296 angstrom
  C4 1.3218729238  0.1965728073  1.9959513294 angstrom
  Spherical d
End of Basis
Basis
  H.6-31G**
  H5 2.0526602523  0.7568282320  1.4351034056 angstrom
  H6 -0.6138178851  -0.6941171027  1.9113821810 angstrom
  H7 -0.8171509745  1.0643342316  -0.2648232855 angstrom
  H8 0.1260134708  -0.4020589690  -0.8535699812 angstrom
End of Basis
AMFI
&SEWARD

2. Make a copy of basis_library/6-31Gpp with a different name. You can make it in your input directory, for instance: SIX-THREEONEGPP (use only uppercase to make it simpler, otherwise you need a trans.tbl file too). Inside this new file, change the 6-31G** in the H, C, and O basis names (and any other atom you intend to use) to SIX-THREEONEGPP too. Then just add to &GATEWAY:

BasDir = $CurrDir
BASIS=SIX-THREEONEGPP

maheshg

Member

Registered: 2018-09-25

Posts: 9

Re: [SOLVED] Problem in using 6-31G(d,p) with AMFI card

Dear Ignacio,

Thank you very much for the helpful reply. Giving the geometry and basis set in the native format solved the issue.

Mahesh