Subroutine NAOANL could not find a p-type valence orbital on

Subroutine NAOANL could not find a p-type valence orbital on

Postby seebol » Tue Aug 04, 2015 10:57 am

I'm trying to run NBO on an old Gaussian09 calculation, using the gaussian "pop=nbo" and "guess=(read,only)" commands. When the program gets to the NBO section, I get an error that I don't understand. Does anyone have any insight into this?

Code: Select all
******************************Gaussian NBO Version 3.1******************************
             N A T U R A L   A T O M I C   O R B I T A L   A N D
          N A T U R A L   B O N D   O R B I T A L   A N A L Y S I S
 ******************************Gaussian NBO Version 3.1******************************
       /RESON  / : Allow strongly delocalized NBO set

 Analyzing the SCF density

 Job title: SL1074_NBO                                                     

 Storage needed:   1902604 in NPA,   2527629 in NBO (  33544252 available)
 Remove ICRem=  455 ICRemI=  348 L=0 PNAO number  455 atom   25 type  51 NPNAO=  794
   due to symmetric orthogonalization 1, value= 0.00D+00.
 Remove ICRem=  597 ICRemI=  348 L=0 PNAO number  597 atom   35 type  51 NPNAO=  793
   due to symmetric orthogonalization 1, value= 0.00D+00.

[...]

Remove ICRem=   72 ICRemI=   64 L=1 PNAO number   72 atom    4 type 251 NPNAO=  727
   due to symmetric orthogonalization 1, value= 0.00D+00.

 Subroutine NAOANL could not find a p-type valence orbital on atom Ni    2.
 IVAL :  1  1  1  0     M :  1     LA :  3
 Error
 Error termination via Lnk1e in C:\G09W\l607.exe at Tue Aug 04 11:50:53 2015.
 Job cpu time:  0 days  0 hours  1 minutes 11.0 seconds.
 File lengths (MBytes):  RWF=     62 Int=      0 D2E=      0 Chk=      9 Scr=      1
seebol
 
Posts: 4
Joined: Tue Aug 04, 2015 10:52 am

Re: Subroutine NAOANL could not find a p-type valence orbita

Postby seebol » Tue Aug 04, 2015 3:51 pm

To further narrow down the problem, if I just run the Gaussian calculation as a single-point calculation (without reading or restarting from previous calculation files), I still get this problem with NBO. Here's my Gaussian input:

Code: Select all
%chk=C:\G09W\calcs\SL1074_NBO\SL1074_NBO.chk
# genecp rtpssh pop=nboread

SL1074_TPSSH

0 1
 Fe                -0.00000000   -0.00000000    2.30182878
 Ni                -0.00000000    1.17459200    0.21102269
 Ni                 0.00000000   -1.17459200    0.21102269
 C                 -0.11313124   -1.30092705    3.46494117
 C                  0.11313124    1.30092705    3.46494117
 C                 -1.63205039    0.18007981    1.60456178
 C                  1.63205039   -0.18007981    1.60456178
 P                  0.25693204    3.32233999    0.46752130
 C                  0.38607739    1.40438766   -1.79153741
 P                 -0.25693204   -3.32233999    0.46752130
 C                 -0.38607739   -1.40438766   -1.79153741
 O                 -0.18248458   -2.19181382    4.21863463
 O                  0.18248458    2.19181382    4.21863463
 O                 -2.76523290    0.28849625    1.29995509
 O                  2.76523290   -0.28849625    1.29995509
 C                  0.66074842    3.85890620   -1.24602451
 C                 -0.99303865    1.00243852   -1.73076190
 C                  1.35642086    0.36009324   -1.71979075
 C                  0.75049798    2.80780841   -2.18210205
 C                 -0.66074842   -3.85890620   -1.24602451
 C                  0.99303865   -1.00243852   -1.73076190
 C                 -1.35642086   -0.36009324   -1.71979075
 C                 -0.75049798   -2.80780841   -2.18210205
 C                  0.94165682    5.17530018   -1.64401909
 C                  1.14920438    3.09202933   -3.49905640
 C                 -0.94165682   -5.17530018   -1.64401909
 C                 -1.14920438   -3.09202933   -3.49905640
 C                  1.32679245    5.44950811   -2.95854412
 C                  1.43801136    4.40396106   -3.88321303
 C                 -1.32679245   -5.44950811   -2.95854412
 C                 -1.43801136   -4.40396106   -3.88321303
 H                  0.87443339    5.98796792   -0.92464804
 H                  1.54839207    6.46951419   -3.25905648
 H                  1.74452425    4.61219556   -4.90429600
 H                  1.22216019    2.28314946   -4.22099571
 H                 -1.77095183    1.75594969   -1.81126384
 H                  2.40986493    0.62266028   -1.69910579
 H                  1.77095183   -1.75594969   -1.81126384
 H                 -2.40986493   -0.62266028   -1.69910579
 H                 -0.87443339   -5.98796792   -0.92464804
 H                 -1.22216019   -2.28314946   -4.22099571
 H                 -1.54839207   -6.46951419   -3.25905648
 H                 -1.74452425   -4.61219556   -4.90429600
 C                 -1.11965514    4.41951221    1.02529043
 H                 -0.85572046    5.47852158    0.93325685
 H                 -2.01485565    4.21577832    0.43269651
 H                 -1.33700651    4.19356380    2.07316496
 C                  1.11965514   -4.41951221    1.02529043
 H                  0.85572046   -5.47852158    0.93325685
 H                  2.01485565   -4.21577832    0.43269651
 H                  1.33700651   -4.19356380    2.07316496
 C                 -1.68353904   -3.94710142    1.45993339
 H                 -1.82505333   -5.02348515    1.31467341
 H                 -1.49664278   -3.74331634    2.51706657
 H                 -2.58990343   -3.41769869    1.15649579
 C                  1.68353904    3.94710142    1.45993339
 H                  1.82505333    5.02348515    1.31467341
 H                  1.49664278    3.74331634    2.51706657
 H                  2.58990343    3.41769869    1.15649579

C H O P 0
6-31++G(d,p)
****
Ni Fe 0
lanl2dz
****

Ni Fe 0
lanl2dz

$nbo bndidx $end


seebol
 
Posts: 4
Joined: Tue Aug 04, 2015 10:52 am

Re: Subroutine NAOANL could not find a p-type valence orbita

Postby seebol » Wed Aug 05, 2015 9:29 am

I've also tried running a singlepoint and NBO using 6-31++g(d,p) basis from the atomic coordinates from a previous optimization that used 6-31++g(d,p) for the main group elements and lanl2dz for the metal centers. This gives me a different error during the NBO run:

Code: Select all
 ******************************Gaussian NBO Version 3.1******************************
             N A T U R A L   A T O M I C   O R B I T A L   A N D
          N A T U R A L   B O N D   O R B I T A L   A N A L Y S I S
 ******************************Gaussian NBO Version 3.1******************************

       /RESON  / : Allow strongly delocalized NBO set
       /BNDIDX / : Print bond indices based on the NAO density matrix

 Analyzing the SCF density

 Job title: SL1074_TPSSH                                                   

 Storage needed:   2479356 in NPA,   3296847 in NBO (  33544252 available)
 Remove ICRem=  105 ICRemI=  508 L=0 PNAO number  105 atom    3 type  51 NPNAO=  907
   due to symmetric orthogonalization 1, value= 0.00D+00.
[...]
 Remove ICRem=   53 ICRemI=  501 L=0 PNAO number   53 atom    2 type  51 NPNAO=  887
   due to symmetric orthogonalization 1, value= 0.00D+00.
LAPACK or BLAS routine DGESVD set Info=          -3.
 Bad arguments to LAPack or BLAS routine.
 Error termination via Lnk1e in C:\G09W\l607.exe at Tue Aug 04 20:48:38 2015.
 Job cpu time:  0 days  3 hours 41 minutes 42.0 seconds.
 File lengths (MBytes):  RWF=    430 Int=      0 D2E=      0 Chk=     15 Scr=      1
seebol
 
Posts: 4
Joined: Tue Aug 04, 2015 10:52 am

Re: Subroutine NAOANL could not find a p-type valence orbita

Postby seebol » Wed Aug 05, 2015 1:18 pm

Update: if I rerun the singlepoint and NBO calculations using a 6-31g basis set, NBO finishes without an error! So somewhere between using 6-31g and using 6-31++g(d,p), something is causing NBO to crash. Anyone have any insight on which parts of the basis set are causing the problems?
seebol
 
Posts: 4
Joined: Tue Aug 04, 2015 10:52 am

Re: Subroutine NAOANL could not find a p-type valence orbita

Postby ericg » Sun Aug 09, 2015 2:35 pm

The diffuse functions in the 6-31++G(d,p) basis set give rise to linear dependence issues that NBO 3.1 attempts to resolve. We didn't write this part of the code (the Gaussian folks did), so the error messages you get are a bit of a mystery. No such problems occur in NBO 6.0 (see attached).

The original NBO 3.0 code [E. D. Glendening, A. E. Reed, J. E. Carpenter, and F. Weinhold, QCPE Bull. 10, 58 (1990)] on which "Gaussian NBO Version 3.1" is based is now more than a quarter-century out of date with respect to the type of heavy-metal and highly extended basis applications now in common usage. Differences between older and newer NBO versions are likely to be particularly dramatic in cases of incipient linear dependence pathologies, as illustrated by your example. Except for routine organic applications in modest basis sets, we recommend that "NBO 3.1" results not be trusted to be of publishable reliability unless compared with results of current NBO 6.0.

Eric
Attachments
seebol.log
(1.15 MiB) Downloaded 888 times
ericg
 
Posts: 265
Joined: Sat Dec 29, 2012 9:31 am


Return to Bug Reports

Who is online

Users browsing this forum: No registered users and 1 guest

cron