error in nuclear spin-spin coupling calculation

Find answers or ask questions regarding Dalton calculations.
Please upload an output file showing the problem, if applicable.
(It is not necessary to upload input files, they can be found in the output file.)

Post Reply
p_petkov17
Posts: 4
Joined: 11 Feb 2015, 14:20
First name(s): Petko
Last name(s): Petkov
Affiliation: Jacobs University Bremen
Country: Germany

error in nuclear spin-spin coupling calculation

Post by p_petkov17 » 11 Feb 2015, 17:19

Dear all,
I am trying to calculate nuclear spin-spin coupling in 3He dimer. My input looks like :

test.mol:
BASIS
6-31G**
He2 spin-spin coupling
6-31G** basis set
Atomtypes=1 Nosymmetry
Charge=2.0 Atoms=2
He 0.0 0.0 2.8
He 0.0 0.0 -2.8


test.dal:
**DALTON INPUT
.RUN PROPERTIES
**WAVE FUNCTIONS
.HF
.CC
*CC INPUT
.SOPPA(CCSD)
**PROPERTIES
.SOPPA(CCSD)
.SPIN-SPIN
*SPIN-S
.ISOTOP
2 2
**END OF DALTON INPUT

When I run the program it gives me an error message:
No symmetry -> DSO by numerical integration.
Electrons: 4.0000005( 4.63e-07): DFTDSO time: 0.1 s

TRACTL_1: Integral transformation abandoned,
the required MO integrals are already available.
RSPORT: 1 out of 1 new trial vectors linear dependent

--- SEVERE ERROR, PROGRAM WILL BE ABORTED ---
Date and time (Linux) : Wed Feb 11 17:02:10 2015
Host name : animal

Reason: LRST, START VECTORS NOT LINEAR INDEPENDENT.

>>>> Total CPU time used in DALTON: 0.35 seconds
>>>> Total wall time used in DALTON: 0.36 seconds

I have done calculation for D2 dimer and it works with the same input.
Any suggestions are very welcome. Thak you in advance!
Petko

frj
Posts: 11
Joined: 09 May 2014, 08:37
First name(s): Frank
Last name(s): Jensen
Affiliation: Aarhus University
Country: Denmark

Re: error in nuclear spin-spin coupling calculation

Post by frj » 11 Feb 2015, 21:30

I don't have an answer for your problem, but if you get it working, I strongly suggest using a different basis set than 6-31G**. Spin-spin coupling constants require significant uncontraction and additions of several tight functions to get any sensible results. Of course the pcJ-n would be my suggestion...

Frank

taylor
Posts: 545
Joined: 15 Oct 2013, 05:37
First name(s): Peter
Middle name(s): Robert
Last name(s): Taylor
Affiliation: Tianjin University
Country: China

Re: error in nuclear spin-spin coupling calculation

Post by taylor » 12 Feb 2015, 01:09

I admit to some confusion here now. You are using He_2 as a further test, but surely the He nuclei have no nuclear magnetic moment, and no spin? In that case it would appear that there is nothing to couple? I wonder whether as a result the program is trying to solve response equations with a right-hand side that is a zero vector, in which case the solution would also be a zero vector. Could you perhaps try Li_2 instead? I had thought also of suggesting the Isotope keyword in the mol file in order to select the second isotope of He, which is of course 3He and is magnetic. But I am not sure that the program is designed to do isotopic substitution for NMR calculations, I am only sure for vibrational calculations. Hence I think Li_2 is a more reliable test.

I strongly endorse what Frank Jensen said in his posting, both that the unmodified 6-311G** type basis sets are totally inadequate for describing properties at or near the nucleus, and that the pc basis sets are considerably to be preferred.

Best regards
Pete

p_petkov17
Posts: 4
Joined: 11 Feb 2015, 14:20
First name(s): Petko
Last name(s): Petkov
Affiliation: Jacobs University Bremen
Country: Germany

Re: error in nuclear spin-spin coupling calculation

Post by p_petkov17 » 12 Feb 2015, 12:25

Dear Pete and Frank, thank you for the advices. The basis set is not an issue at this stage because I just want to get a working input.
According to the isotope specification, in the manual is written:
.ISOTOP
READ (LUCMD,*) (ISOTPS(IS), IS=1, NATOMS)
Calculate the indirect spin–spin coupling constants for a given isotopic constitution
of the molecule. The next line reads the isotope number for each of the atoms in
the molecule (including also symmetry-dependent molecules). The isotopic number
for each atom is given in terms of the occurrence in the list of natural abundance
of the isotopes for the given atom, i.e. the most abundant isotope is number 1, the
second-most abundant is number 2 and so on.

That is why I used command line like :
.ISOTOP
2 2
because the most abundant isotope is number 1 (He_4 99.999863%, has no magnetic moment) the second one i.e. isotope number 2 is H_3 only 0.000137%.
I have tried also a command lines for isotope specification like :
.ISOTOP
3 3
but I get the same error.
If I do not specify any isotopes the program ending without error and without results for spin-spin coupling because He_4 isotope has no magnetic moment. It is like that, because the program by default takes all isotopes in the molecule with a natural abundance above 1.0%.
I have tried with the same type of input to calculate the spin-spin coupling for C_13 and it works, just for He I got this error.
So I am stacked at the moment and appreciate all further suggestions.
Cheers,
Petko

sauer
Posts: 45
Joined: 27 Aug 2013, 16:37
First name(s): Stephan P. A.
Last name(s): Sauer
Affiliation: Department of Chemistry, University of Copenhagen
Country: Denmark
Contact:

Re: error in nuclear spin-spin coupling calculation

Post by sauer » 12 Feb 2015, 13:43

Hi Petko,

I have not final solution to that yet, but could you try a coupled of things, which you should try independently?
a) could you remove the
.ISOTOP
2 2
part of the input?
This would tell us, whether the problem is related to using different isotopes, although I doubt it.
And if that is the case, there is an easy fix for it: Multiply by hand the results by the ratio of the g-factors of the two isotopes. That is the only difference between the outputs with and without specially chosen isotopes. There would be a more significant difference, if you would ask for vibrational averaging, but not for a single point calculation.
b) could you remove the
Nosymmetry
in the *mol file?
There is absolutely no reason for running without symmetry, if you are not trying to debug the program. And there are all reasons for using symmetry - it will go much faster.
c) could you ask for
.NODSO
in the
*SPIN-S
section.
This would skip the calculation of the DSO term, where you get the
"No symmetry -> DSO by numerical integration.
Electrons: 4.0000005( 4.63e-07): DFTDSO time: 0.1 s"
message from.
But I believe that this is just a message and not related to your error. Putting the .NODSO would tell me, whether I am right.

Best wishes
Stephan

p_petkov17
Posts: 4
Joined: 11 Feb 2015, 14:20
First name(s): Petko
Last name(s): Petkov
Affiliation: Jacobs University Bremen
Country: Germany

Re: error in nuclear spin-spin coupling calculation

Post by p_petkov17 » 12 Feb 2015, 15:11

Hi Stephan,
I did the test calculations, which you suggested.
for your suggestion a) I removed the .ISOTOP part of the input.
**DALTON INPUT
.RUN PROPERTIES
**WAVE FUNCTIONS
.HF
.CC
*CC INPUT
.SOPPA(CCSD)
**PROPERTIES
.SOPPA(CCSD)
.SPIN-SPIN
**END OF DALTON INPUT

The program finished without error, but JJ coupling constants are missing because it uses by default He_4 isotope , which nuclear has no magnetic moment.
*******************************************************************************
**************** ABACUS - INDIRECT NUCLEAR SPIN-SPIN COUPLINGS ****************
*******************************************************************************

Definitions from Smith, Palke, and Grieg, Concepts in Mag. Res. 4 (1992) 107



+------------------------------------+
! ABACUS - Final spin-spin couplings !
+------------------------------------+



Interatomic separations (in Angstrom):
--------------------------------------

He He
------ ------
He : 0.000000
He : 2.963392 0.000000


Max interatomic separation is 2.9634 Angstrom ( 5.6000 Bohr)
between atoms 2 and 1, "He " and "He ".

Min YX interatomic separation is 2.9634 Angstrom ( 5.6000 Bohr)


Bond distances (Angstrom):
--------------------------

atom 1 atom 2 distance
------ ------ --------
bond distance: He He 2.963392

CPU time statistics for ABACUS
------------------------------

>>>> Total CPU time used in ABACUS: 0.48 seconds
>>>> Total wall time used in ABACUS: 0.48 seconds

For your b) and c) suggestions i got absolutely the same error as before:

Changes of defaults for SPIN-S:
-------------------------------

Spin-spin couplings printed for atoms with abundance greater than : 0.000 %

WARNING SIR_INTOPEN, old integral transformation not found as expected, ICASE = 1
- ICASE = 1: MO integral file MOTWOINT does not exist
- ICASE = 2: no MO integrals on MOTWOINT
- ICASE = 3: transformation level on MOTWOINT not sufficient

NSETUP: MO transformation level too low or no MO integral file found.
NSETUP: generating MO 2-el. integral file.


Integral transformation: Total CPU and WALL times (sec) 0.152 0.156

Center of mass dipole origin : 0.000000 0.000000 0.000000

Center of mass gauge origin : 0.000000 0.000000 0.000000


.------------------------------------------------.
| Starting in Static Property Section (ABACUS) - |
`------------------------------------------------'



Date and time (Linux) : Thu Feb 12 14:56:21 2015
Host name : animal

Symmetry -> DSO by analytical integration.

TRACTL_1: Integral transformation abandoned,
the required MO integrals are already available.
RSPORT: 1 out of 1 new trial vectors linear dependent

--- SEVERE ERROR, PROGRAM WILL BE ABORTED ---
Date and time (Linux) : Thu Feb 12 14:56:23 2015
Host name : animal

Reason: LRST, START VECTORS NOT LINEAR INDEPENDENT.

>>>> Total CPU time used in DALTON: 2.66 seconds
>>>> Total wall time used in DALTON: 2.66 seconds


QTRACE dump of internal trace stack

========================
level module
========================
8 RSPCTL
7 RSPLLE
6 ABARSP
5 LRSDRV
4 ABACTL
3 ABACUS
2 DALTON
1 DALTON main
========================


I did additional tests and calculated He3-He3 spin spin coupling constants but instead of SOPA(CCSD) I used DFT. With DFT my *.dal file works with .ISOTOP keyword:
**DALTON INPUT
.RUN PROPERTIES
**WAVE FUNCTIONS
.DFT
B3LYP
**PROPERTIES
.SPIN-S
*SPIN-S
.ISOTOP
2 2
**END OF DALTON INPUT

And the result in the output is :
*******************************************************************************
**************** ABACUS - INDIRECT NUCLEAR SPIN-SPIN COUPLINGS ****************
*******************************************************************************

Definitions from Smith, Palke, and Grieg, Concepts in Mag. Res. 4 (1992) 107



+------------------------------------+
! ABACUS - Final spin-spin couplings !
+------------------------------------+



Indirect spin-spin coupling between He and He :
======================================================


Mass number atom 1: 3; Abundance: 0.000 %; g factor: -4.255250
Mass number atom 2: 3; Abundance: 0.000 %; g factor: -4.255250

Isotropic coupling : 4.5693 Hz
Anisotropic coupling : 0.8996 Hz
Asymmetry : 0.0000
S parameter : 0.8996 Hz
A parameter : 0.0000 Hz
Isotropic DSO contribution: 0.0000 Hz
Isotropic PSO contribution: -0.0001 Hz
Isotropic SD contribution : -0.0003 Hz
Isotropic FC contribution : 4.5697 Hz




Total Coupling Tensor in Hz
---------------------------

x y z

x 4.269457 -9.3041210E-19 6.4756996E-16
y -9.3853819E-19 4.269457 1.0417028E-15
z -8.1557867E-16 1.2065007E-16 5.169052

It looks like SOPPA(CCSD) has some problems in this case.

Cheers,
Petko

sauer
Posts: 45
Joined: 27 Aug 2013, 16:37
First name(s): Stephan P. A.
Last name(s): Sauer
Affiliation: Department of Chemistry, University of Copenhagen
Country: Denmark
Contact:

Re: error in nuclear spin-spin coupling calculation

Post by sauer » 12 Feb 2015, 16:26

Thanks.

I have run some test calculations myself now and I can see that is applies to SOPPA and SOPPA(CCSD) and the PSO and SD terms. Apparently the program tries to solve linear equations, which is should not, because the property gradient is zero, which it notices in the case of HF but not for SOPPA. I wonder, what would happen for an MCSCF calculation with this system, because it is the same routines.

So, to summarize, this is a bug, which I do not have a solution right now.
I am sorry
Stephan

p_petkov17
Posts: 4
Joined: 11 Feb 2015, 14:20
First name(s): Petko
Last name(s): Petkov
Affiliation: Jacobs University Bremen
Country: Germany

Re: error in nuclear spin-spin coupling calculation

Post by p_petkov17 » 12 Feb 2015, 16:33

Thank you for the efforts Stefan! I will try an MCSCF calculation with this system. The good point is that we figured it out where the problem comes from.
Regards!
Petko

andrey
Posts: 1
Joined: 17 Nov 2015, 01:56
First name(s): Andrey
Last name(s): Yachmenev
Affiliation: UCL
Country: United Kingdom

Re: error in nuclear spin-spin coupling calculation

Post by andrey » 17 Nov 2015, 12:34

Hi Stephan,

is there a chance this bug is fixed in the near future? I have the same problem with SOPPA complex frequency-dependent polarizabilities.

With Best Regards,
Andrey

xiongyan21
Posts: 183
Joined: 24 Sep 2014, 08:36
First name(s): yan
Last name(s): xiong
Affiliation: CENTRAL CHINA NORMAL UNIVERSITY
Country: China

Re: error in nuclear spin-spin coupling calculation

Post by xiongyan21 » 10 Dec 2015, 10:25

I used the input prepared by Dr. Pektokv (SOPPA(CCSD)) to run Dalton 2015.1 sequentially, and got the following result with 711: WARNING SIR_INTOPEN, old integral transformation not found as expected, ICASE = 1 in the bash,

ABACUS - Final spin-spin couplings !
+------------------------------------+



Indirect spin-spin coupling between He and He :
======================================================


Mass number atom 1: 3; Abundance: 0.000 %; g factor: -4.255250
Mass number atom 2: 3; Abundance: 0.000 %; g factor: -4.255250 ,

Isotropic coupling : 0.6971 Hz
Anisotropic coupling : 0.9607 Hz
Asymmetry : 0.0000
S parameter : 0.9607 Hz
A parameter : 0.0000 Hz
Isotropic DSO contribution: 0.0000 Hz
Isotropic PSO contribution: -0.0000 Hz
Isotropic SD contribution : -0.0000 Hz
Isotropic FC contribution : 0.6971 Hz




Total Coupling Tensor in Hz
---------------------------

x y z

x 0.3768500 -2.0172046E-18 1.5697769E-16
y -1.8898870E-18 0.3768500 2.0081125E-17
z 8.2345985E-17 1.4282135E-17 1.337538


Interatomic separations (in Angstrom):
--------------------------------------

He He
------ ------
He : 0.000000
He : 2.963392 0.000000


Max interatomic separation is 2.9634 Angstrom ( 5.6000 Bohr)
between atoms 2 and 1, "He " and "He ".

Min YX interatomic separation is 2.9634 Angstrom ( 5.6000 Bohr)


Bond distances (Angstrom):
--------------------------

atom 1 atom 2 distance
------ ------ --------
bond distance: He He 2.963392




CPU time statistics for ABACUS
------------------------------




>>>> Total CPU time used in ABACUS: 0.19 seconds
>>>> Total wall time used in ABACUS: 0.21 seconds


.-------------------------------------------.
| End of Static Property Section (ABACUS) - |
`-------------------------------------------'

>>>> Total CPU time used in DALTON: 0.23 seconds
>>>> Total wall time used in DALTON: 0.31 seconds ,
and a parallel run finished with the same result and the following there

711: WARNING SIR_INTOPEN, old integral transformation not found as expected, ICASE = 1


===================================================================================
= BAD TERMINATION OF ONE OF YOUR APPLICATION PROCESSES
= PID 528 RUNNING AT
= EXIT CODE: 136
= CLEANING UP REMAINING PROCESSES
= YOU CAN IGNORE THE BELOW CLEANUP MESSAGES
===================================================================================
Last edited by xiongyan21 on 27 Jan 2016, 11:13, edited 6 times in total.

xiongyan21
Posts: 183
Joined: 24 Sep 2014, 08:36
First name(s): yan
Last name(s): xiong
Affiliation: CENTRAL CHINA NORMAL UNIVERSITY
Country: China

Re: error in nuclear spin-spin coupling calculation

Post by xiongyan21 » 10 Dec 2015, 10:55

xiongyan21 wrote:I used the input prepared by Dr. Pektokv (SOPPA(CCSD)) to run Dalton 2015.1 sequentially, and got the following result but with 711: WARNING SIR_INTOPEN, old integral transformation not found as expected, ICASE = 1 in the bash,

ABACUS - Final spin-spin couplings !
+------------------------------------+



Indirect spin-spin coupling between He and He :
======================================================


Mass number atom 1: 3; Abundance: 0.000 %; g factor: -4.255250
Mass number atom 2: 3; Abundance: 0.000 %; g factor: -4.255250 ,

Isotropic coupling : 0.6971 Hz
Anisotropic coupling : 0.9607 Hz
Asymmetry : 0.0000
S parameter : 0.9607 Hz
A parameter : 0.0000 Hz
Isotropic DSO contribution: 0.0000 Hz
Isotropic PSO contribution: -0.0000 Hz
Isotropic SD contribution : -0.0000 Hz
Isotropic FC contribution : 0.6971 Hz




Total Coupling Tensor in Hz
---------------------------

x y z

x 0.3768500 -2.0172046E-18 1.5697769E-16
y -1.8898870E-18 0.3768500 2.0081125E-17
z 8.2345985E-17 1.4282135E-17 1.337538


Interatomic separations (in Angstrom):
--------------------------------------

He He
------ ------
He : 0.000000
He : 2.963392 0.000000


Max interatomic separation is 2.9634 Angstrom ( 5.6000 Bohr)
between atoms 2 and 1, "He " and "He ".

Min YX interatomic separation is 2.9634 Angstrom ( 5.6000 Bohr)


Bond distances (Angstrom):
--------------------------

atom 1 atom 2 distance
------ ------ --------
bond distance: He He 2.963392




CPU time statistics for ABACUS
------------------------------




>>>> Total CPU time used in ABACUS: 0.19 seconds
>>>> Total wall time used in ABACUS: 0.21 seconds


.-------------------------------------------.
| End of Static Property Section (ABACUS) - |
`-------------------------------------------'

>>>> Total CPU time used in DALTON: 0.23 seconds
>>>> Total wall time used in DALTON: 0.31 seconds ,

and a parallel run finished with the same result and the following there

711: WARNING SIR_INTOPEN, old integral transformation not found as expected, ICASE = 1


===================================================================================
= BAD TERMINATION OF ONE OF YOUR APPLICATION PROCESSES
= PID 528 RUNNING AT
= EXIT CODE: 136
= CLEANING UP REMAINING PROCESSES
= YOU CAN IGNORE THE BELOW CLEANUP MESSAGES
===================================================================================

Last edited by xiongyan21 on 27 Jan 2016, 11:12, edited 2 times in total.

taylor
Posts: 545
Joined: 15 Oct 2013, 05:37
First name(s): Peter
Middle name(s): Robert
Last name(s): Taylor
Affiliation: Tianjin University
Country: China

Re: error in nuclear spin-spin coupling calculation

Post by taylor » 15 Dec 2015, 14:48

You appear to have posted the same thing twice: was that intentional? And if the two postings are different, exactly how are they different? Finally, it is not clear to me at least whether you are reporting a bug, or saying the results are wrong, or that the program runs to completion but does not shut itself down correctly? Exactly what problem are you drawing attention to.

Best regards
Pete
P.S. It would be much more helpful to post files than to include output fragments in the posting itself, at least unless the latter are formatted in fixed spacing. It is not straightforward for an old guy with crap eyesight (like me...) to figure out some of the numbers in output that is re-rendered in non-fixed format!

xiongyan21
Posts: 183
Joined: 24 Sep 2014, 08:36
First name(s): yan
Last name(s): xiong
Affiliation: CENTRAL CHINA NORMAL UNIVERSITY
Country: China

Re: error in nuclear spin-spin coupling calculation

Post by xiongyan21 » 16 Dec 2015, 02:48

Dear Dr. Taylor
It is not intentional. This maybe caused during the modification by clicking on the mouse for the second time when the submission seems to be unsuccessful, especially when the network is busy, of course with identical content, and I just want you to help me to delete the latter one.
Do the repeatedly incorrect shutdowns when a parallel run is carried out mean a bug or the wrong configuration of Dalton on my MAC?
Thanks a lot for your helps.
Last edited by xiongyan21 on 03 Feb 2016, 16:44, edited 6 times in total.

taylor
Posts: 545
Joined: 15 Oct 2013, 05:37
First name(s): Peter
Middle name(s): Robert
Last name(s): Taylor
Affiliation: Tianjin University
Country: China

Re: error in nuclear spin-spin coupling calculation

Post by taylor » 17 Dec 2015, 16:45

I would defer to someone who is more expert in the parallel code than myself here, but it is difficult at this point to know whether the problem is in the code or in the system setup. Do you only get this error for this particular property, or for other properties run in parallel as well? If you get this error for other properties I would suspect the setup, rather than the code. I do not run the code on my Mac because I have other more convenient resources available, so it is some time since I even built it for the Mac, and anyway I don't recall ever having tried to run in parallel on the Mac, only under Linux.

It is not straightforward to connect this to the original posting, either. That seemed to reveal a bug in that the program was trying to calculate the solution to a linear equation system where the right-hand side was completely zero by definition, but I do not see quite how that could be connected to an error opening a file.

Many people are approaching holidays now, for the Xmas/New Year break, and in addition the release of Dalton2016 is imminent. I fear not much progress will be made on your posting until after New Year.

Best regards
Pete

xiongyan21
Posts: 183
Joined: 24 Sep 2014, 08:36
First name(s): yan
Last name(s): xiong
Affiliation: CENTRAL CHINA NORMAL UNIVERSITY
Country: China

Re: error in nuclear spin-spin coupling calculation

Post by xiongyan21 » 18 Dec 2015, 02:50

On OS X EI Capitan 10.11.2, Xcode7.2, GFortran 5.2.0, MPICH2 3.1.4_1, Cmake 3.3.2, etc., the sequential originally intended calculation(SOPPA(CCSD))using DALTON 2015.1 gets a normal termination, but with information like solving SOPPA linear response skipped because norm of p-h property vector SD 002 x is only although ...

Best Regards!
Last edited by xiongyan21 on 28 Jan 2016, 03:30, edited 5 times in total.

xiongyan21
Posts: 183
Joined: 24 Sep 2014, 08:36
First name(s): yan
Last name(s): xiong
Affiliation: CENTRAL CHINA NORMAL UNIVERSITY
Country: China

Re: error in nuclear spin-spin coupling calculation

Post by xiongyan21 » 20 Dec 2015, 03:45

The incorrect shutdown maybe caused by the compilation without mpi included.

I can confirm the same results through a sequential run terminating normally can be repeated by Dalton2016.0.

Post Reply

Who is online

Users browsing this forum: No registered users and 5 guests