Firefly and PC GAMESS-related discussion club

 
Learn how to ask questions correctly  
 
 

 

Thomas Pijper
thomaspijper@hotmail.com

For the geometry optimization and Hessian calculation it is indeed necessary to use the exact same methodology. However, for the subsequent Raman calculation it is allowed to change the method, such has modifying the basis set.


Kind regards,
Thom


On Wed Feb 15 '17 10:58am, Siddheshwar Chopra wrote
---------------------------------------------------
>Dear Thomas,
>Thank you for the valuable information that "Raman calculations can give better results when diffuse functions are used". Yes I agree about 18+1 gradient calculations needed.
>However about the diffuse functions use for raman, shouldn't we use the same SET for both preceding and raman calculations? This question was asked to me too.

>Regards,

>On Tue Feb 14 '17 11:49pm, Thomas Pijper wrote
>----------------------------------------------
>>Dear Siddheshwar,

>>As stated in the manual, the calculation requires 18+1 gradient evaluations. The Hessian should be calculated in advance and needs to be provided in the input.

>>Note though that Raman calculations can give better results when diffuse functions are used (whereby these functions do not have to be used on the preceding geometry optimization and Hessian calculation). This increases computational cost.
>>
>>
>>Kind regards,
>>Thom
>>
>>
>>
>>
>>On Tue Feb 14 '17 11:35am, Siddheshwar Chopra wrote
>>---------------------------------------------------
>>>Dear All,
>>>I have never worked with Raman spectrum calculations using FIREFLY. Can anybody having experience with these calculations tell about their intensiveness of computations? It needs $HESS as noted from the manual.
>>>
>>>
>>>Regards,