Firefly and PC GAMESS-related discussion club

 
Learn how to ask questions correctly  
 
 

 

Alex Granovsky
gran@classic.chem.msu.su

>Thanks for the guide, I will put them into consideration but the only thing I am afraid of is the time cost of using higher basis set and the cost of implementing $force nvib=2 $end as it requires 6N+1 evaluations  because some of my molecules are up to 70 atoms with Ru metal.

I see. By the way, you can reuse already computed information
to reduce the cost of $force nvib=2 $end computations.
Please let me know if you are interested and I'll post here a
couple of useful tricks.

On the other hand, polarization functions are very often a
prerequisite to get the qualitatively correct picture and so  
you may consider to add at least one shell of polarization functions
per heavy atom.

>The imaginary frequency is the first frequency of each molecule and it is within the range of 21.70I and 25.00I.

This can be indeed caused by the nvib=1 numerical inaccuracies.
Are there exact six additional modes corresponding rotations +
translations? Did you check all these modes using e.g. ChemCraft
(the trial version is free)?

Another typical reason causing small imaginary frequencies to appear
is the internal rotation of, say, methyl groups if there are any of
them in the molecule. Some people bravely ignores these floppy modes
while computing hessians.  

Kind regards,
Alex Granovsky

>Kind regards,
>Adebayo
>
>
>On Thu Apr 26 '12 7:32pm, Alex Granovsky wrote
>----------------------------------------------
>>Dear Adebayo,

>>how large are these imaginary frequencies? If they are small,
>>this could be an artifact from numerical hessian calculations.
>>You can use $force nvib=2 $end to improve precision of the computed
>>hessian and frequencies. I'd also strongly suggest you to visualize
>>these "bad" frequencies to understand their origin. Another thing
>>to try is use larger and better basis set.  

>>Kind regards,
>>Alex Granovsky
>>
>>
>>
>>On Tue Apr 24 '12 12:58pm, Adebayo A. Adeniyi wrote
>>---------------------------------------------------
>>>Dear all Firefly Users,

>>>I optimized my metal organic compounds of Ru metals (derivatives of anticancer called RAPTA) using external basis set with PBE0 functional. The basis set for the Ru, Cl, P is SBKTC with ECP while other atoms have 6-31g. But after obtainig the optimized geometry, I extracted the optimized geometry using Open Babel and proceed to frequency calculation using RUNTYP=HESSIAN, GUESS=READ and HESS=CALC, but against my expectation, almost all the supposed stable moleculed were having their first frequency as Imaginary. I want to know what is wrong since the initial geometry of some is from crstal structure.