iOpenShell » Q-Chem specific questions » scf_guess read for AO-XSAPT(KS)-D3

scf_guess read for AO-XSAPT(KS)-D3

Page: 1

Author Post
Member
Registered: Oct 2016
Posts: 5
Hi,

I am trying using AO-XSAPT(KS)-D3 with fragment wavefunctions inherited from previous scf jods using scf_guess fragmo.
But I encountered problems.
I simplify my model into a unrestricted Li2 AO-XSAPT-D3 model with opposite spin on two atoms, given the error information is the same.
The fragmo step in this system seems redundant but just for demonstrating the error.

final out:
"
**************************************
* *
* XPOLMAN *
* Leif D. Jacobson *
* and *
* John M. Herbert *
* *
**************************************

XPOL setup time 0.07 (cpu) 0.07 (wall)
Requested basis set is aug-cc-pVTZ
There are 28 shells and 92 basis functions
A cutoff of 1.0D-14 yielded 406 shell pairs
There are 4446 function pairs ( 6399 Cartesian)
Smallest overlap matrix eigenvalue = 2.62E-05
Smallest overlap matrix eigenvalue = 2.62E-05
Guess MOs from SCF MO coefficient file
Reading MOs from coefficient file
Reading MOs from coefficient file

Q-Chem fatal error occurred in module /home/scratch/svnadmin/10478_zgan/qchem/libmdc/newfileman.C, line 293:

Error reading in TMP file 869/0 (135424)

p0_67826: p4_error: interrupt SIGx: 6

"
What's the content of 869.0 file? -save of previous job didn't creat this file.
The decomposition without MO read can be completed.


I submitted jobs through specifying the savedir and below is the input:

#1
fragmo guess:
"
$rem
JOBTYPE SP
SCF_GUESS FRAGMO
SCF_PRINT_FRGM TRUE
scf_print true
sym_ignore true
basis aug-cc-pvtz
GUI 2
UNRESTRICTED TRUE
method hf
$end

$molecule
0 1
--
0 2
Li 0.0 0.0 1.0
--
0 -2
Li 0.0 0.0 -1.0
$end
"

#2
read and re-scf: ! just for ensurence and this step can produce a .fchk with the whole-system wavefuctions suitable for third-party visualization software,
otherwise, it will be visualized only containing the last fragment.
"
$rem
JOBTYPE SP
SCF_GUESS read
sym_ignore true
basis aug-cc-pvtz
GUI 2
UNRESTRICTED TRUE
method hf
SCF_FINAL_PRINT 4
print_orbitals true
$end

$molecule
read
$end
"

#3
Final AO-XSAPT-D3:
"
$rem
gui 2
scf_guess read
unrestricted true
sym_ignore true
exchange hf
basis aug-cc-pvtz
xpol true
xpol_mpol_order gas
xpol_print 3
sapt_print 3
sapt true
sapt_ao true
sapt_order 2
sapt_basis dimer
sapt_exchange s_inverse
sapt_disp_corr true
sapt_disp_version 3
sapt_dscf true
sapt_cphf true
mem_total 62000
mem_static 4000
ao2mo_disk 50000
thresh 14
scf_convergence 8
$end

$molecule
0 1
--
0 2
Li 0.0 0.0 1.0
--
0 -2
Li 0.0 0.0 -1.0
$end
"


Besides, the model of H2 can not complete even without MO read
"
**************************************************************
SAPT

Leif D. Jacobson, Ka Un Lao, John M. Herbert
**************************************************************
...
...
===========================================================
CPHF solver time: 0.16 (cpu) 0.15 (wall)
===========================================================

K construction time: CPU 0.01 s wall 0.01 s
K construction time: CPU 0.01 s wall 0.01 s
K[O] Non-systematic Exchange Matrices Time: 0.01 (cpu) 0.01 (wall)

MKL ERROR: Parameter 4 was incorrect on entry to DGETRF
Info = -4

Q-Chem fatal error occurred in module /home/scratch/svnadmin/10478_zgan/qchem/liblas/MatInvert.C, line 54:

Call to dgetrf failed in MatInvert

p0_67346: p4_error: interrupt SIGx: 6
"


And sometimes I found this error :shock: , I don't know what's the reason.
What is GH standing for? It it a input format error? The input looks the same with above steps.
"
**************************************
* *
* XPOLMAN *
* Leif D. Jacobson *
* and *
* John M. Herbert *
* *
**************************************

XPOL setup time 0.08 (cpu) 0.08 (wall)
Requested basis set is aug-cc-pVTZ
There are 28 shells and 92 basis functions
A cutoff of 1.0D-14 yielded 406 shell pairs
There are 4446 function pairs ( 6399 Cartesian)
Smallest overlap matrix eigenvalue = 2.62E-05
Smallest overlap matrix eigenvalue = 2.62E-05

Unexpected End of File for file

/home/wangzg/Q-Chem4.4/Build/qcaux/basis/aug-cc-pVTZ.bas

was found while looking for Atomic Symbol GH.

Check to see if this file has a basis for that atomic symbol.
The input format searched for here should be:
<atomic symbol> <0>
Q-Chem fatal error occurred in module /home/scratch/svnadmin/10478_zgan/qchem/forms1/AtomicBasis.C, line 833:

Basis not supported for the above atom.

p0_66392: p4_error: interrupt SIGx: 6

"

Thanks for your patience and looking for your help.

Best regards.
Member
Registered: Oct 2016
Posts: 5
I found the last error, namely, "atomic symbol GH", occurs when using "GUI 2" together with AO-XSAPT(KS)-D.

The main problem above it is still needing help.

Best regards.
Member
Registered: Sep 2007
Posts: 58
Hi warnJ,

XSAPT uses FRAGMO as initial guess. You don't have to run a separate FRAGMO calculation. Do you have any reason for such an inheriting setup?

Wesley
Member
Registered: Oct 2016
Posts: 5
Hi Wesley,

Thanks for your reply.

In my practical calculation, each fragment possesses a complex broken-symmetry singlet state, which can not be obtained in a one-step XSAPT job using the default FRAGMO based on a two-fragment molecule block.

I actually constructed the wavefuction for the XSAPT dimer through a separate FRAGMO jobs, where ten fragments are defined, five for each monomer.

Is it possible to complete such a two-order FRAGMO combining with the XSAPT? I just tried using MO READ as described.

wanrJ
« Last edit by wanrJ on Fri Oct 14, 2016 3:58 am. »
Member
Registered: Sep 2007
Posts: 58
Hi WanrJ,

I understand what you are doing. Unfortunately, combined FRAGMO results can not be directly used by XSAPT code because XSAPT saves FRAGMO results into a different scratch file. I can contact the XSAPT developer and see if they can help to add this feature, but you might need to wait till the coming release.

If you are in a hurry, the workaround is to manually copy those combined results and save them as xSAPT density file. If you want to do this, I can help you on this matter.

Thank you,
Wesley
Member
Registered: Oct 2016
Posts: 5
Dear Welsley,

Thank you a lot.

I'm actually trying the demo of Q-chem and not in a hurry.

I'll keep following about the feature and may all the best with Q-chem.

Best Regards,
WanrJ
Member
Registered: Sep 2007
Posts: 58
Hi WanrJ,

Thank you for being interested in Q-Chem. For testing this feature, would you mind to provide me with some example for test?

Thank you,
Wesley
Member
Registered: Oct 2016
Posts: 5
Hi Wesley,

I'm glad to provide some examples.

There are some typical systems usually used for demonstrations of constructing a broken-symmetry singlet wavefunction
through the fragment guess.
You can build dimers taking them as monomers.(flip the spin of one monomer for testing an antiferromagnetically coupled dimer)

Below are the structures in the fragment guess:

# *CH2-CH2-CH2-CH2*

$molecule
0 1
--
0 3
C 0.09727146 1.92239482 0.00000000
H 0.47392697 2.13428940 0.97885116
H 0.47392697 2.13428940 -0.97885116
C -0.55878611 0.59240328 0.00000000
H -1.16892856 0.49566725 0.87365029
H -1.16892856 0.49566725 -0.87365029
--
0 -3
C 0.55878611 -0.59240328 0.00000000
H 1.16892856 -0.49566725 -0.87365029
H 1.16892856 -0.49566725 0.87365029
C -0.09727146 -1.92239482 0.00000000
H -0.47392697 -2.13428940 -0.97885116
H -0.47392697 -2.13428940 0.97885116
$end



# Fe4S4, two Fe spin up and the other two Fe spin down,
# S atoms are included in the fragment just for conveinet.

$molecule
0 1
--
0 5
Fe -0.000000000000 -1.256142548900 0.888226914500
Fe 0.000000000000 1.256142548900 0.888226914500
S -1.845393493800 0.000000000000 1.304890253400
S 1.845393493800 -0.000000000000 1.304890253400
--
0 -5
Fe -1.256142548900 0.000000000000 -0.888226914500
Fe 1.256142548900 -0.000000000000 -0.888226914500
S -0.000000000000 -1.845393493800 -1.304890253400
S 0.000000000000 1.845393493800 -1.304890253400
$end

# Fe and S system with four ligands (sorry for the ignorance about its specific chemical compound name),
# a more practical monomer for the inermolecular interaction but also more complicate.

$molecule
0 1
--
3 6
Fe 15.263 -1.0091 7.0068
--
3 -6
Fe 17.043 1.0091 7.0068
--
-2 1
S 14.8495 1.149 7.0431
--
-2 1
S 17.4565 -1.149 7.0431
--
-1 1
S 14.3762 -2.1581 8.7983
C 12.5993 -2.1848 8.6878
H 12.3743 -3.6513 10.1678
C 10.4994 -3.1122 9.4309
H 9.9929 -3.7579 10.0022
C 9.8049 -2.2791 8.5639
H 8.805 -2.2873 8.5744
C 10.4833 -1.4146 7.6615
H 9.973 -0.8525 7.0106
C 11.9048 -1.3675 7.7057
H 12.4158 -0.7843 7.0743
C 11.8886 -3.065 9.5194
--
-1 1
C 17.2999 3.4265 4.6624
C 16.6376 4.1967 5.609
H 16.5022 3.8494 6.5369
C 16.153 5.4856 5.2463
H 15.6665 6.0472 5.9155
C 16.3468 5.9257 4.0431
H 16.0236 6.8408 3.802
C 17.0091 5.1398 3.0522
H 17.133 5.4944 2.1254
C 17.4775 3.8823 3.3884
H 17.94 3.3149 2.7071
S 17.9266 1.8091 5.0446
--
-1 1
H 23.501 2.2873 8.5744
S 17.9298 2.1581 8.7983
C 19.7067 2.1848 8.6878
C 20.4174 3.065 9.5194
H 19.9317 3.6513 10.1678
C 21.8066 3.1122 9.4309
H 22.3132 3.7579 10.0022
C 22.5011 2.2791 8.5639
C 21.8227 1.4146 7.6615
H 22.333 0.8525 7.0106
C 20.4012 1.3675 7.7057
H 19.8902 0.7843 7.0743
--
-1 1
S 14.3794 -1.8091 5.0446
C 15.0061 -3.4265 4.6624
C 15.6684 -4.1967 5.609
H 15.8038 -3.8494 6.5369
C 16.153 -5.4856 5.2463
H 16.6395 -6.0472 5.9155
C 15.9592 -5.9257 4.0431
H 16.2824 -6.8408 3.802
C 15.2969 -5.1398 3.0522
H 15.173 -5.4944 2.1254
C 14.8285 -3.8823 3.3884
H 14.366 -3.3149 2.7071
$end

Hope this could be helpful.

All the best,
wanrJ

Page: 1

iOpenShell » Q-Chem specific questions » scf_guess read for AO-XSAPT(KS)-D3