NBO 6.0 NRT calculation problems

NBO 6.0 NRT calculation problems

Postby ktpob » Tue Nov 04, 2014 6:58 am

Hi,

There was a post from 2012 of a user using NBO 5.9 and getting the error message similar to this one that I'm getting:

NATURAL RESONANCE THEORY ANALYSIS:

Maximum reference structures : 20
Maximum resonance structures : 300
Memory requirements : 11444808 words of 99275923 available

TOPSTR: too many structures; consider setting NRTRES > 300 or NRTCHG < 1
Maximum scratch memory used by GENNBO was 957851 words

I've tried changing the number of NRTRES and NRTCHG but it isn't making any difference. I still get the same error message. I've tried adding NRTTHR=20 in the .47 file too, but again, I get the same error message.

Could someone help please?

Thanks

ktpob
ktpob
 
Posts: 8
Joined: Fri Oct 31, 2014 5:52 am

Re: NBO 6.0 NRT calculation problems

Postby ericg » Wed Nov 05, 2014 10:51 pm

What error message is reported when you use, say, NRTRES=1000 and NRTCHG=0? How large is the system that you are attempting to analyze?

Eric
ericg
 
Posts: 277
Joined: Sat Dec 29, 2012 9:31 am

Re: NBO 6.0 NRT calculation problems

Postby ktpob » Thu Nov 06, 2014 5:01 am

Hi Eric.

Whatever number I put for NRTRES=x, even for 1000, it says "TOPSTR: too many structures; consider setting NRTRES > 1000 or NRTCHG < 1"

If I put NRTCHG=0 it says the same error message except it says "or NRTCHG < 0" instead.

My system is bis(benzene)chromium, so 25 atoms and 108 electrons. This is just a "test" to see how I can do NRT analysis on eta-6 bonded metal-benzene bonds. My main system I want to study if bis(benzene)chromium works is an organometallic with up to 59 atoms and 316 electrons, depending on which ligand I'm changing. But each complex has this eta-6 bonding mode.

Thanks

ktpb
ktpob
 
Posts: 8
Joined: Fri Oct 31, 2014 5:52 am

Re: NBO 6.0 NRT calculation problems

Postby ericg » Sun Nov 09, 2014 11:18 pm

I tried bis(benzene)chromium myself, and show a sample G09 input file below that yields reasonable NRT output.

The algorithm that generates candidate resonance structures for NRT can have difficulties with transition metal centers. In these cases it is often better to propose a set of candidate structures using the $NRTSTR keylist. I do this below, specifying the four Kekule forms. These four structures are sufficient to initiate the NRT search for additional structures, including those that reveal bonding interactions between the central metal center and the benzene rings.

Although not necessary, I use the NRTTHR keyword to limit the number of delocalizing interactions that will be considered to generate new resonance forms, and the NRTCML keyword to write a CML file that can be used with applications like ChemAxon's MarvinView (highly recommended!) to display the calculated resonance structures.

Eric

Code: Select all
%mem=1gb
%nproc=2
#b3lyp/sdd pop=nbo6read symm=veryloose

bis(benzene)chromium...b3lyp/sdd

0  1
Cr    0.000000    0.000000    0.000000
C    -1.662634    1.432048    0.000030
C    -1.662851    0.716023   -1.240053
C    -1.662851   -0.715971   -1.240082
C    -1.662634   -1.432048   -0.000030
C    -1.662851   -0.716023    1.240053
C    -1.662851    0.715971    1.240082
C     1.662633    1.432048    0.000023
C     1.662851    0.715977    1.240079
C     1.662851   -0.716017    1.240056
C     1.662633   -1.432048   -0.000023
C     1.662851   -0.715977   -1.240079
C     1.662851    0.716017   -1.240056
H    -1.636845    2.517251    0.000052
H    -1.638173    1.258511   -2.179990
H    -1.638173   -1.258421   -2.180042
H    -1.636845   -2.517251   -0.000052
H    -1.638173   -1.258511    2.179990
H    -1.638173    1.258421    2.180042
H     1.636845    2.517251    0.000040
H     1.638173    1.258431    2.180036
H     1.638173   -1.258501    2.179996
H     1.636845   -2.517251   -0.000040
H     1.638173   -1.258431   -2.180036
H     1.638173    1.258501   -2.179996

$nbo nrt nrtthr=10 nrtcml $end
$nrtstr
  str
    lone 1 3 end
    bond d 2 3 s 2 7 s 2 14 s 3 4 s 3 15 d 4 5 s 4 16 s 5 6 s 5 17 d 6 7
         s 6 18 s 7 19 s 8 9 d 8 13 s 8 20 d 9 10 s 9 21 s 10 11 s 10 22
         d 11 12 s 11 23 s 12 13 s 12 24 s 13 25 end
  end
  str
    lone 1 3 end
    bond s 2 3 d 2 7 s 2 14 d 3 4 s 3 15 s 4 5 s 4 16 d 5 6 s 5 17 s 6 7
         s 6 18 s 7 19 s 8 9 d 8 13 s 8 20 d 9 10 s 9 21 s 10 11 s 10 22
         d 11 12 s 11 23 s 12 13 s 12 24 s 13 25 end
  end
  str
    lone 1 3 end
    bond s 2 3 d 2 7 s 2 14 d 3 4 s 3 15 s 4 5 s 4 16 d 5 6 s 5 17 s 6 7
         s 6 18 s 7 19 d 8 9 s 8 13 s 8 20 s 9 10 s 9 21 d 10 11 s 10 22
         s 11 12 s 11 23 d 12 13 s 12 24 s 13 25 end
  end
  str
    lone 1 3 end
    bond d 2 3 s 2 7 s 2 14 s 3 4 s 3 15 d 4 5 s 4 16 s 5 6 s 5 17 d 6 7
         s 6 18 s 7 19 d 8 9 s 8 13 s 8 20 s 9 10 s 9 21 d 10 11 s 10 22
         s 11 12 s 11 23 d 12 13 s 12 24 s 13 25 end
  end
$end
ericg
 
Posts: 277
Joined: Sat Dec 29, 2012 9:31 am

Re: NBO 6.0 NRT calculation problems

Postby ktpob » Mon Nov 10, 2014 4:24 am

Hi Eric,

Thanks for this reply. I copied the NRTSTR keyword and Kekule forms into my .47 file and the .out file has the message:

NATURAL RESONANCE THEORY ANALYSIS:

Maximum reference structures : 20
Maximum resonance structures : 300
Memory requirements : 8985603 words of 99275923 available

4 reference structure(s) read from $NRTSTR keylist

SELECT: no reasonable reference structures
Maximum scratch memory used by GENNBO was 957851 words

The only difference is I used VTZ quality basis sets to generate the .47 file. Could this have something to do with it? Also for my main structures I want to analyse once I get chromium to work has pseudopotentials as well. Not sure if that's relevant or not.

Thanks

ktpob
ktpob
 
Posts: 8
Joined: Fri Oct 31, 2014 5:52 am

Re: NBO 6.0 NRT calculation problems

Postby ericg » Mon Nov 10, 2014 7:33 am

VTZ and pseudopotentials shouldn't matter; NBO can work with both. Note that atom ordering is important here. The structures listed in $NRTSTR of my Gaussian input are appropriate for the order in which the atoms are listed in the geometry section.

Eric
ericg
 
Posts: 277
Joined: Sat Dec 29, 2012 9:31 am

Re: NBO 6.0 NRT calculation problems

Postby ktpob » Tue Nov 11, 2014 6:21 am

Hi Eric,

Thanks a lot. This has worked, and I've also used it to generate a .47 file.

If I'm running the .47 file in a later version of NBO, do I need to input these NRT keywords into the .47 as well?

Thanks again

Kieran
ktpob
 
Posts: 8
Joined: Fri Oct 31, 2014 5:52 am

Re: NBO 6.0 NRT calculation problems

Postby ericg » Tue Nov 11, 2014 10:47 am

Kieran,

Yes, I'd recommend the keywords as shown in the G09 input above. (Note that NRTCML in only available in NBO6, not in early versions of the program.)

Eric
ericg
 
Posts: 277
Joined: Sat Dec 29, 2012 9:31 am

Re: NBO 6.0 NRT calculation problems

Postby ktpob » Wed Nov 12, 2014 6:28 am

Excellent, many thanks Eric.

Kieran
ktpob
 
Posts: 8
Joined: Fri Oct 31, 2014 5:52 am

Re: NBO 6.0 NRT calculation problems

Postby ktpob » Thu Nov 13, 2014 6:31 am

Hi Eric,

Sorry to ask more questions, I just have one more thing. Like I said, the molecule I ideally want to analyse has a metal centre sandwiched between two arene rings, however when I run $NRTSTR it says there's an error in the structures. I've checked the kekule structures and they seem to be correct.

In addition to the benzene rings the ligand also contains 2 pyrrole rings, however these are bonded with a hapticity of just 1, so normal NBO deals with these bonds. Having said that, do I still need to include all the kekule structures for both pyrrole rings in the NRTSTR list as well? Is the fact I'm not including these the reason why there's an error? This will be a lot of when added to the 4 different benzene kekule structures as well, as just one pyrrole ring has 5 resonance structures.

Thanks

Kieran
ktpob
 
Posts: 8
Joined: Fri Oct 31, 2014 5:52 am

Next

Return to General NBO Discussion

Who is online

Users browsing this forum: Bing [Bot] and 2 guests

cron