Geometry optimizations that won't converge
Clicked A Few Times
10:18:21 AM PDT - Thu, Aug 21st 2014
This is a question that isn't really specific to NWChem, since I think all codes struggle with this. I am working with a benzaldehyde dimer (28 atoms total, 14 in each molecule) that is failing to converge in the geometry optimization; it is not going wild, it is just getting to a reasonable place and then making tiny changes and never totally converging. This is a DFT run with the BLYP functional. Are there any thoughts (with this case, and in general) on how to handle these failures to converge with a geometry optimization? Thanks in advance for any advice.

Forum Vet
4:18:41 PM PDT - Thu, Aug 21st 2014
Could you post the input file?
Cheers, Edo
PS By the way, BLYP might not be the best XC functional for a system that is likely to have hydrogen bonds.

Clicked A Few Times
8:19:04 AM PDT - Fri, Aug 22nd 2014
Quote:Edoapra Aug 21st 4:18 pm
Could you post the input file?
Cheers, Edo
PS By the way, BLYP might not be the best XC functional for a system that is likely to have hydrogen bonds.



Could you please tell me more about picking a good functional for a system with hydrogen bonds?

Here is the input file:

title "benz"
echo
memory total 8000 stack 2000 heap 2000 global 4000 mb
scratch_dir ./scratch
geometry noautosym units angstroms noautoz
C -0.0104264112060349 1.2217277514897262 -0.0067215915170059
C 1.3532236922262595 1.0376207989405550 -0.0100473424117131
C 1.8827338023761457 -0.2427426217894013 -0.0047394994210278
C 1.0450793476024702 -1.3417295269712477 0.0038063008011855
C -0.3226230587407987 -1.1576391608095158 0.0071742984298133
C -0.8588428428176198 0.1217738325711968 0.0019385807161303
C -2.3141267455224108 0.2958852986137319 0.0059091006174388
O -2.8806249352935782 1.3636886481992183 0.0030452114213137
H 2.0212406382283934 1.8993024513319567 -0.0170051706955214
H 2.9650290473933638 -0.3837195354887875 -0.0074984890651671
H 1.4654738778600864 -2.3478515703138321 0.0080308956664231
H -0.9941868126164205 -2.0194959957310901 0.0141343124104505
H -2.9034663695358489 -0.6674590749110183 0.0128044859411187
H -0.4484831321651365 2.2206487109982476 -0.0108410920382612
C -0.0104264112060349 1.2217277514897262 2.4932784084829942
C 1.3532236922262595 1.0376207989405550 2.4899526575882871
C 1.8827338023761457 -0.2427426217894013 2.4952605005789721
C 1.0450793476024702 -1.3417295269712477 2.5038063008011857
C -0.3226230587407987 -1.1576391608095158 2.5071742984298133
C -0.8588428428176198 0.1217738325711968 2.5019385807161303
C -2.3141267455224108 0.2958852986137319 2.5059091006174388
O -2.8806249352935782 1.3636886481992183 2.5030452114213135
H 2.0212406382283934 1.8993024513319567 2.4829948293044786
H 2.9650290473933638 -0.3837195354887875 2.4925015109348330
H 1.4654738778600864 -2.3478515703138321 2.5080308956664230
H -0.9941868126164205 -2.0194959957310901 2.5141343124104507
H -2.9034663695358489 -0.6674590749110183 2.5128044859411189
H -0.4484831321651365 2.2206487109982476 2.4891589079617389
end
basis
* library 6-311G*
end
driver
 maxiter 400
end
dft
#BLYP functional

xc becke88 lyp

direct

iterations 400
end
task dft optimize

Forum Vet
9:39:46 AM PDT - Fri, Aug 22nd 2014
I have found no problem running your input file.
What version of NWChem have you been using?
Could you post your output file to a website?
There might be issue with the NWChem installation you are using

Clicked A Few Times
12:29:34 PM PDT - Fri, Aug 22nd 2014
Quote:Edoapra Aug 22nd 9:39 am
I have found no problem running your input file.
What version of NWChem have you been using?
Could you post your output file to a website?
There might be issue with the NWChem installation you are using


I am using version 6.1.1. Do you have any suggestions on a website that I could post my output file to? I am not sure how to do this.

Thanks very much for your help.

Clicked A Few Times
10:24:16 AM PDT - Mon, Aug 25th 2014
Quote:Edoapra Aug 22nd 9:39 am
I have found no problem running your input file.
What version of NWChem have you been using?
Could you post your output file to a website?
There might be issue with the NWChem installation you are using


I tried running this on another computer and had no trouble, so it seems that you are right that it is an installation issue. Thank you very much for your help.


Forum >> NWChem's corner >> General Topics