Using COSMO with UHF

Peter Walton

unread,

Nov 4, 2020, 11:07:58 AM11/4/20

to NWChem Forum

Hi there,

I am trying to calculate the solvent-corrected energy for systems using the Hartree-Fock method. I need to use UHF and the documentation states that COSMO can be used with UHF by invoking the DFT module with specific keywords. However I am unsure which of these keywords I need to use and so I was hoping someone may be able to shed some light on the matter.

Kind Regards

Peter

Edoardo Aprà

unread,

Nov 4, 2020, 3:17:31 PM11/4/20

to NWChem Forum

You are not required to use the DFT module to perform UHF calculations with COSMO.

From

https://nwchemgit.github.io/COSMO-Solvation-Model.html

... Note that COSMO for unrestricted Hartree-Fock (UHF) method can also be performed by invoking the DFT module with appropriate keywords ...

Anyhow, this is how to do a UHF calculation with the DFT module

dft

odft

xc hfexch

end

task dft

Peter Walton

unread,

Nov 5, 2020, 9:29:29 AM11/5/20

to NWChem Forum

Many thanks for the reply Edoardo, I am perhaps slightly confused. Do I have to first perform a single point calculation in DFT to allow me to calculate the solvent correction in UHF? Currently with the input file:

echo

START

memory 30000 mb

charge 1

geometry

C 2.2426050700 1.5250285200 0.1230799900

C 2.0744508700 0.1890610100 -0.2232484600

H 2.8655380700 -0.3598444700 -0.6945083000

H 3.1535870400 2.0643483900 -0.0478092100

C 0.8617940400 -0.4444204900 0.0935986600

C 0.7415571600 -1.9100992100 -0.2280555700

C -0.1627851600 0.2693830700 0.6983030400

N 1.2536143400 2.1855975200 0.7075861800

H 1.4027477300 3.1453312000 0.9589375900

C -0.0046199700 1.6624597100 0.9163758600

H -0.6045703800 2.1884823300 1.6338463800

C -1.1157060500 2.6978411300 -0.6011771600

H -0.9314487400 2.2671251200 -1.5688877700

F -2.3538065100 2.6043605400 -0.2015273800

F -0.6656402300 3.9275709000 -0.4773257700

O 1.6166141300 -2.4877361700 -0.7817330200

O -0.3971217000 -2.4056079900 0.1630917000

H -1.0809723100 -0.2055985600 0.9795574300

C -0.6364745800 -3.8147562100 -0.0522553200

C -2.0948676500 -4.0594088400 0.2542368400

H 0.0241107500 -4.3614021600 0.6050566600

H -0.3816797400 -4.0483479400 -1.0742687000

H -2.3321182600 -3.7467098800 1.2641831100

H -2.3153063300 -5.1160619000 0.1615700300

H -2.7268764300 -3.5149960600 -0.4375155800

end

basis

* library 6-31G*

end

scf

doublet ; uhf

end

driver

xyz geo-opt

end

stepper

TS

maxiter 1000

end

cosmo

solvent dcm

end

task scf saddle

I am getting the error seen below:

NWChem Gradients Module

-----------------------

wavefunction = UHF

------------------------------------------------------------------------

grad_hnd_cos: rtdb load failed for Bq 916

------------------------------------------------------------------------

current input line :

48: task scf saddle

------------------------------------------------------------------------

An error occured in the Runtime Database

------------------------------------------------------------------------

For more information see the NWChem manual at

http://www.nwchem-sw.org/index.php/NWChem_Documentation

For further details see manual section:

0:0:grad_hnd_cos: rtdb load failed for Bq:: 916

What is it that I need to do so I can perform a solvent corrected calculation with the UHF wavefunction?

Kind Regards

Peter

Edoardo Aprà

unread,

Nov 5, 2020, 12:50:21 PM11/5/20

to NWChem Forum

It turns out that the documentation is misleading and your original intution was correct.

The UHF part of the SCF module is not compatible with COSMO (see bug report I have just filed in https://github.com/nwchemgit/nwchem/issues/267) while the ROHF method is compatible, instead.

Therefore, you can either switch from UHF to ROHF and use the SCF task or use the UHF methods in the DFT task.

Here are the two input snippets.

ROHF from SCF task

...

scf

rohf

doublet

end

task scf

UHF from DFT task

...

dft

xc hfexch

mult 2

end

task dft

Reply all

Reply to author

Forward