Hello !
I have performed AIM analysis by means Gaussian 09 program and using the following keywords in the roote section:
# b3lyp/6-311g(d,p) geom=connectivity density=current aim=all, but the job stops and gives this message:
FATAL ERROR: DUPLICATE CRITICAL POINTS FOUND: 42 4
Error termination via Lnk1e in /home/kader/g09/l609.exe at Mon Feb 22 17:32:34 2016.
Thanks for any help!
Best Regards,
Without the knoweldge of your full input (and the molecule you have analyzed) it is hard to give precise answer. But maybe you could check the analysis using the DGrid program (link to download the program is in web, respectively on my ResGate account -- the download is free of charge). You just need to write out the formated CheckPoint file (using fchk=filename in the G09 input) and then proceed with this file as described in the DGrid manual. Maybe this will help you to find out the origin of the problem with the Gaussian09 density.
Thanks a lot Dr Miroslav Kohout for you answer. Actually, I am studying a kappa-carrabiose molecule which is a disaccharide, you can find in attached document the input file.
Thanks again,
For some molecules the density has such distribution that AIM analysis does not make sense anymore. Tune your basis and method a bit to see if the change of density associated with those would make the extra critical points on the density distribution go a way. However, it basically means that for this particular molecule, AIM analysis, even if you can get it done without error, does not have much physical meaning. AIM is theoretically clean and good looking, and the quantities are well defined, but its benefits stops there. AIM charges and bond order usually have very weak correlation with the actual physical picture and provide little intuition for the nature of your system, and in this case it basically cannot tell you anything useful at all.
The error "DUPLICATE CRITICAL POINTS FOUND" means that two critical points are closer than about 1e-6 bohr to each other. Unfortunately, Gaussian does not output the coordinates of those points (because it stops without further output). It still makes sense to analyze the density, because the error can signal a tiny small nonnuclear attractor in the electron density. There is principally nothing wrong about it, but Gaussian will not let you proceed with the analysis -- use another program. The AIM analysis for your molecule is meaningful -- more about the nonnuclear attractors see paper of A.M. Pendas.
Thank you all for your help, I will take your recommendations in consideration.
Hallo, Rachida!
Gaussian is not suitable for the AIM analysis of more complex molecules. For this purpose I propose to generate a .wfn file (keyword Output=WFN with a subsequent .wfn file name definition behind coordinates and an empty line) and this can be treated using AIMALL, AIM2000 or other similar software.
Best wishes
mb
Hi!
Thank you very much Martin Breza . I have already tried with the free AIMAALL but it is limited to 12 atoms or fewer and my systems are composed by 46 and 49 atoms, the same for the aim 2000, so I wonder if there is another free software that allows me to perform this analysis.
Thank you in advance for your answer.
Another alternative software to do AIM analysis is opensource Multiwfn, which is really easy to learn and can provide really good results. Also no restriction as AIMALL. It takes gaussian .fchk files as input.
Link of Multiwfn : https://multiwfn.codeplex.com/
I've been using Multiwfn for a year now, and it has been excellent. The only draw back I find is that it is predominately a Windows distribution, with all my efforts at compiling a Mac/Linux distribution have failed. I login to a virtual Windows machine to use it.
If think it is a real shame that the Gaussian AIM functionality isn't forgiving, otherwise it would make an excellent full integrated QM-to-AIM software package.
For Linux one can use the DGrid program (which will be this year distributed in the 5.0 version). DGrid works with molecular wavefunctions from Gaussian, Gamess, ADF, Molpro, Molcas and Turbomole now already for around 25 years. The 5.0 version is able to analyze solid state wavefunction from the FHI-aims program (the 4.6 version was able to analyze the results of the solid state program Elk). All the bonding analyses with DGrid go far beyond the QTAIM approach.
- Badges
- Science topic