From owner-chemistry@ccl.net Mon Oct 29 00:17:00 2012
From: "Raman Kumar Singh singh,yamanashi.ac.jp" <owner-chemistry::server.ccl.net>
To: CCL
Subject: CCL: SCF convergance failure
Message-Id: <-47805-121028221457-28626-7PsYfBPBEsHUvUKgHqKZRw::server.ccl.net>
X-Original-From: "Raman Kumar Singh" <singh-$-yamanashi.ac.jp>
Date: Sun, 28 Oct 2012 22:14:56 -0400


Sent to CCL by: "Raman Kumar Singh" [singh a yamanashi.ac.jp]
Hello Everyone,

I want to do single point energy calculation using CCSDT (DFT)/aug-cc-pVTZ//B3LYP/cc-pVTZ for the molecules like linear anthracene and Tetracene .... but end up with convergence failure. For Benzene and Naphthalene, no problem. I have tried following things:

1) increased the number of cycles
2) scf=qc
3) scf=xqc
4) scf(vshift=150)
5) scf(xqc, vshift=150, maxcyc=512)
6) using ultrafine grid
7) scf=noincfock
8) scf=fermi
9) increased the convergence criteria from 10^(-8) to 10^(-9/-10/-12)

and possible combinations of above-mentioned keywords.

Besides above, I also tried using lower basis sets but the problem remained the same.

If somebody could help me out, I shall deeply appreciate. 

Regards
Raman


From owner-chemistry@ccl.net Mon Oct 29 08:33:01 2012
From: "Mariusz Radon mariusz.radon^_^gmail.com" <owner-chemistry- -server.ccl.net>
To: CCL
Subject: CCL:G: SCF convergance failure
Message-Id: <-47806-121029041358-5317-lf9JTXZ1wkgJpwvfis/z6Q- -server.ccl.net>
X-Original-From: Mariusz Radon <mariusz.radon,+,gmail.com>
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=ISO-8859-1
Date: Mon, 29 Oct 2012 09:13:50 +0100
MIME-Version: 1.0


Sent to CCL by: Mariusz Radon [mariusz.radon~~gmail.com]
Hi Raman:

The first thing to make sure is what exactly means your "convergence
problem": that SCF does not converge or that CCSD does not converge? All
tricks (1-9) you mentioned affect only the SCF convergence...

A good idea to improve convergence of SCF (beside the tricks you tried)
is to supply a better initial guess through ``scf=read''. You can read
it from the checkpoint file obtained in the previous calculations with a
smaller basis set (e.g., cc-pVTZ you used).

In contrast, if your problem is CCSD not converging, I see that Gaussian
has a ``MaxCyc'' keyword as an option to ``CCSD''. But this is
applicable only if your SCF does converge.


Regards,
Mariusz Radon



On 10/29/2012 03:14 AM, Raman Kumar Singh singh,yamanashi.ac.jp wrote:
> Sent to CCL by: "Raman Kumar Singh" [singh a yamanashi.ac.jp]
> Hello Everyone,
> 
> I want to do single point energy calculation using CCSDT (DFT)/aug-cc-pVTZ//B3LYP/cc-pVTZ for the molecules like linear anthracene and Tetracene .... but end up with convergence failure. For Benzene and Naphthalene, no problem. I have tried following things:
> 
> 1) increased the number of cycles
> 2) scf=qc
> 3) scf=xqc
> 4) scf(vshift=150)
> 5) scf(xqc, vshift=150, maxcyc=512)
> 6) using ultrafine grid
> 7) scf=noincfock
> 8) scf=fermi
> 9) increased the convergence criteria from 10^(-8) to 10^(-9/-10/-12)
> 
> and possible combinations of above-mentioned keywords.
> 
> Besides above, I also tried using lower basis sets but the problem remained the same.
> 
> If somebody could help me out, I shall deeply appreciate. 
> 
> Regards
> Raman> 
> 


-- 
Dr Mariusz Radon, Ph.D.
Coordination Chemistry Group
Faculty of Chemistry
Jagiellonian University
ul. Ingardena 3, 30-060 Krakow, Poland
http://www.chemia.uj.edu.pl/~mradon