Convergence criteria for geometry optimization with B3LYP

Clicked A Few Times

4:28:53 PM PST - Fri, Nov 19th 2010
Dear all, I have optimized geometry at B3LYP/6-31++G(d,p) using G09. I used default optimization criteria ( Max Force = 0.00045, RMS Force = 0.0003, Max Displacement = 0.0018 and RMS Displacement = 0.0012). When I compare this geometry with (default) optimized geometry given by NWChem, the structure turned out to be different. I used the following input setting for NWChem optimization: ...... geometry end basis library 6-31++G** end DFT GRID FINE (I tried with medium grid too) XC B3LYP ITERATIONS 100 END task DFT optimize Since I am new to NWChem I could not understand the convergence criteria for NWChem which is eqivalent to the G09 default convergence as listed above. Can you please suggest me the input setting in order to have equivalent convergence criteria to that of gaussian? Thank you very much in advance for your valuable suggestion ram

Hi Ram,

The code will reorient the geometry by default.

For your geometry block try the following

geometry nocenter noautoz

...

end

as far as the dft defaults just use

driver

 maxiter 50

end

dft

 xc b3lyp

 iterations 100

end
task dft optimize

Clicked A Few Times

11:39:02 AM PST - Sun, Nov 21st 2010
Hi Guest; Thank you very much for your suggestion. I still have the problem with the optimized geometry. When I use dft defaults the optimized geometry from NWChem is different for that of default DFT in G09. Could you please suggest me the DFT convergence setting which is equivalent to that of default DFT criteria for G09 ? Thank you very much Ram

Forum Regular

6:34:13 PM PST - Sun, Nov 21st 2010
Hi Ram, How different is your converged geometry ? I don't use G09 so I cannot suggest an equivalent setting. If you don't wish to post your geometry or input, then just send me your input file and output so that I can take a closer look. My email and webpage information are listed in the Developers page. Best, -Niri