[molpro-user] Unreasonable T2 norm for CCSD(T) calculation with cc-pCVTZ but not for cc-pCVDZ

[Paul Jerabek]

Hi Tatiana,

thank you very much for your answer!

If I use the CC_NORM_MAX=XYZ option and play around with the numbers I
get arbitrarily large t2 vector norms (the higher XYZ, the higher the
norm). No convergence behaviour can be observed.

I very curious why this happens for the VTZ basis but not in VDZ.

Cheers,

	Paul

Am 05.02.2015 um 16:18 schrieb Tatiana Korona:
> Hi,
> 
> If you put
> 
> CC_NORM_MAX=100.0 ! some large number
> 
> somewhere in your input, it will force CCSD to assume that your T2 norm
> is "reasonable", but of course your CCSD should converge in order to get
> sensible results.
> 
> Best wishes,
> 
> Tatiana
> 
> On Thu, 5 Feb 2015, Paul Jerabek wrote:
> 
>> Dear all,
>>
>> I'm trying to calculate the core electron ionization energy for CO with
>> CCSD(T) including core-valence correlation. The problem arises with the
>> singly 1s(C) core-electron ionized CO molecule.
>>
>> With a cc-pCVDT basis I get a nicely converged CCSD(T) result.  When
>> using the larger basis cc-pCVTZ, the calculation stops with the
>> "unreasonable norm" error and a large value for the t2 norm vector.
>>
>> Also: The MP2 part of the CCSD(T) calculation won't converge with the
>> cc-pCVTZ basis as well, although I overcame this obstacle by playing
>> around with the DIIS options: "diis, 1,1,6,1" worked for me.
>>
>> Do you have any idea what exactly causes the problem at the cc-pCVTZ
>> level and how to deal with it this? I'd really like to get the energies
>> at the  CCSD(T)/cc-pCVTZ level of theory.
>>
>> Thank you very much for your help!!
>>
>>     Paul

-- 
Dr. Paul Jerabek
Fachbereich Chemie, AK Berger, Theoretische Chemie
Philipps-Universität Marburg
Hans-Meerwein-Strasse
D-35032 Marburg
Germany

Phone: +49-(0)6421-28-27001