Multiplicity problem


Just Got Here
Hi,
I want to calculate an electron transfer reaction, NaCl -> Na+Cl-, with follow input file. But this calculation stops with a error, “scf: no. of closed-shell electrons is not even! 0”.


TITLE "NaCl_et"
start NaCl_et

geometry Na units angstroms print xyz noautoz noautosym
 Na         0.0  0.0  0.0
end

geometry Cl units angstroms print xyz noautoz noautosym
 Cl         5.0 0.0 0.0
end

geometry Na_Cl units angstroms print xyz noautoz noautosym
 Na        0.0  0.0  0.0
Cl 5.0 0.0 0.0
end

basis
* library 6-31G*
end

scf
  doublet; uhf; vectors input atomic output Na.movecs
end
charge 0
set geometry Na
task scf

scf
  doublet; uhf; vectors input atomic output Cl.movecs
end
charge 0
set geometry Cl
task scf

scf
  singlet; uhf; vectors input atomic output Nap.movecs
end
charge 1.00
set geometry Na
task scf

scf
  singlet; uhf; vectors input atomic output Clm.movecs
end
charge -1.00
set geometry Cl
task scf

  1. ET reactants:
scf
 singlet; uhf; vectors input fragment Na.movecs Cl.movecs output NaCl.movecs
end
set geometry Na_Cl
task scf

  1. ET products:
scf
 singlet; uhf; vectors input fragment Nap.movecs Clm.movecs output NapClm.movecs
end
task scf

et
vectors reactants NaCl.movecs
vectors products NapClm.movecs
end
task scf et

I think this error is caused by inappropriate multiplicity. In this file, Cl has more than one alpha spins to beta spins and Na does too. So if Cl and Na are combined, they would contain more than two alpha spins to beta spins and it would be not singlet anymore.

Then, I want to calculate to make Cl have more than one beta spins to alpha spins.

How can I assign as it?

Forum Vet
Both the reactants and products molecule calculations are run with a charge -1, which you used in the Cl- calculation. Remember, the data is persistent. So, if you specified a charge -1 before, it will continue to use it until you change it. Simply do a "charge 0" before "task scf" of the reactant.

Bert

Just Got Here
Bert,

Thank you for your fast reply.
I am ashamed that I did not notice such a simple mistake.
The error, “scf: no. of closed-shell electrons is not even! 0”, was resolved if I followed your advice.
But I got a new error, ” movecs_fragment: open shell mismatch 14”.
If I calculated NaCl.movecs with triplet option (not singlet option), then I got the output file without any error.
So I still need to get a Cl’s data which has more than one beta spins to alpha spins to calculate singlet.
If you know, could you tell me what I must do?

Rui

Forum Vet
In the SCF you will always have more alpha than beta spins when you specify a doublet and higher. However, there is a way for you to get this, but you will have to use the DFT module. Here you can specify a "mult -2" etc to tell the code to put more electrons in the beta orbitals. And, by setting "xc HFexch" the DFT code will be effectively doing an SCF calculation.

Just Got Here
I calculated with HFexch and I reached successful termination!

Thank you very much, Bert!


Forum >> NWChem's corner >> Running NWChem