reforming Gaussian user
Clicked A Few Times
9:09:15 AM PST - Wed, Jan 13th 2016
looking for a little help with Syntax- interested in selenium-cesium salts; been running trial cesium methylselenide;
trying to run carbon and hydrogen with 6-31G* and Cesium and Selenium with LANL08; from error messages, looks like i need to set charges- can someone suggest how to fix the syntax to get this to run with MP2 optimiziation

  1. ================================================================
  2. NWChem input file made in Gabedit
  3. ================================================================

charge 0


geometry
C  -3.844301 4.526917 -12.172369

H  -4.171634 3.870637 -12.978749

Se  -5.045227 5.281263 -11.611687

H  -3.114974 5.239638 -12.557353

H  -3.387928 3.931306 -11.381754

Cs  -7.241826 6.632592 -10.568067
end

basis "ao basis" spherical
carbon library 6-31G* 

hydrogen library 6-31G*

se library LANL08

cs library LANL08
end

  2. SCF on the full system #

unset scf:* # This restores the defaults
charge 0
set atomscf:tags_z Se Cs
set atomscf:z -1 1
scf
sextet

vectors fragment ring.mo imid.mo fe.mo water.mo

maxiter 50
end
scf
Singlet

maxiter 150

thresh 1e-6

rhf
end
mp2
freeze core atomic
  1. freeze virtual 5
  2. tight
  3. print
  4. noprint
  5. riapprox V
  6. riapprox SVS
end

task MP2 optimize

Forum Regular
11:51:22 AM PST - Thu, Jan 14th 2016
First, you need to specify an ECP input block for Se and Cs, i.e.

ECP
Se library LANL08

Cs library LANL08
END

Second, change "set atomscf:z -1 1" to "set:atomscf:z -1.0 1.0"

Third, I don't understand why you have two SCF input blocks. My guess is that the first input block is from some other calculation because it does not appear consistent with the rest of the input you show above. If that is the case, you would want to remove that from your input file, otherwise the program will try to perform a fragment initial guess using those files, which based on their names seem to describe atoms/molecules not present in this input.

Fourth, only the DIRECT_MP2 and RIMP2 modules are compatible with frozen virtual orbitals. However, these two modules lack analytical gradients, so you should be aware that the program will automatically switch to numerical gradients, which will probably result in a significantly longer calculation than you may expect for optimization.

With the above mentioned changes, I was able to run this input without trouble (note I did not do the full optimization so I don't know if it will converge in the default number of steps, but the calculation does proceed without issue).


Forum >> NWChem's corner >> Running NWChem