Basis sets for NMR

[Ronald Cohen]

I am reposting this with a proper subject from conversation"REFTRAJ for LINRES".

When I use the EMSL 6-31G** basis sets I get reasonable nmr shifts, but with basis sets that are supposed to be optimal for nmr from basissetexchange.org like pcSseg and x2c-tavpall-s.1 I get crazy huge NMR shifts (printed as ****). It is not a cut off issue because results change little doubling the cutoff from 100 to 200 Ryd.

Is there an issue with what basis sets are supported, or is there a problem with the cp2k formats from www.basissetexchange.org?

There are no error messages.

Thanks for any help.

Sincerely,

Ron

[Ronald Cohen]

Here is the example for iceX with x2c-tavpall-s.

Sincerely,

Ron

[Krack Matthias (PSI)]

Hi Ron

 

The exponent 5.11 of the O x2c-TZVPall-2c basis set is not included in the GAPW soft basis set. I suggest to reduce EPSFIT to 1.0E-6 (or less) to achieve that. This is just one possible issue and I do not claim that it will solve all your problems.

 

HTH

 

Matthias

--
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/4de62505-93ed-48f5-8c79-a23bd8de7062n%40googlegroups.com.

[Marcella Iannuzzi]

Dear Ron

Could you check the ground state electronic structure with the different basis sets? 

For instance PDOS, population etc. Are they different? 

You could also check and compare the localisation and the position of the Wannier centres with the different basis sets. 

Best 

Marcella

[ma455...@gmail.com]

Hi Matthias,

I'm a newbie in cp2k. In the basis set file, there is a function row with the exponent of 5.2. However, in the output file, exponent of 5.2 is not seen both in the orbital basis set and the GAPW soft basis set sections. I'm wondering is this reasonable? This phenomenon is observed for both O and H.

Additionally, is there a rule that determines if the exponent should be included in the GAPW soft basis set?

Best,

Hongyang

[Jürg Hutter]

Hi

I cannot reproduce your findings. Can you send input and output files and the Basis set file
for testing?

regards

JH

________________________________________
From: cp...@googlegroups.com <cp...@googlegroups.com> on behalf of ma455...@gmail.com <ma455...@gmail.com>
Sent: Saturday, August 13, 2022 2:33 AM
To: cp2k
Subject: Re: [CP2K:17474] Re: Basis sets for NMR

Hi Matthias,

I'm a newbie in cp2k. In the basis set file, there is a function row with the exponent of 5.2. However, in the output file, exponent of 5.2 is not seen both in the orbital basis set and the GAPW soft basis set sections. I'm wondering is this reasonable? This phenomenon is observed for both O and H.
Additionally, is there a rule that determines if the exponent should be included in the GAPW soft basis set?

[微信截图_20220813101839.png]

Best,
Hongyang
在2022年8月12日星期五 UTC+10 18:04:19<Matthias Krack> 写道：
Hi Ron

The exponent 5.11 of the O x2c-TZVPall-2c basis set is not included in the GAPW soft basis set. I suggest to reduce EPSFIT<https://manual.cp2k.org/cp2k-2022_1-branch/CP2K_INPUT/FORCE_EVAL/DFT/QS.html#EPSFIT> to 1.0E-6 (or less) to achieve that. This is just one possible issue and I do not claim that it will solve all your problems.

HTH

Matthias

From: "cp...@googlegroups.com" <cp...@googlegroups.com> on behalf of Ronald Cohen <reco...@gmail.com>
Reply to: "cp...@googlegroups.com" <cp...@googlegroups.com>
Date: Thursday, 11 August 2022 at 21:35
To: "cp...@googlegroups.com" <cp...@googlegroups.com>
Subject: [CP2K:17466] Re: Basis sets for NMR

Here is the example for iceX with x2c-tavpall-s.

Sincerely,

Ron

On Thursday, August 11, 2022 at 3:27:43 PM UTC-4 Ronald Cohen wrote:
I am reposting this with a proper subject from conversation"REFTRAJ for LINRES".

When I use the EMSL 6-31G** basis sets I get reasonable nmr shifts, but with basis sets that are supposed to be optimal for nmr from basissetexchange.org<http://basissetexchange.org> like pcSseg and x2c-tavpall-s.1 I get crazy huge NMR shifts (printed as ****). It is not a cut off issue because results change little doubling the cutoff from 100 to 200 Ryd.

Is there an issue with what basis sets are supported, or is there a problem with the cp2k formats from www.basissetexchange.org<http://www.basissetexchange.org>?

There are no error messages.

Thanks for any help.

Sincerely,

Ron

--
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/4de62505-93ed-48f5-8c79-a23bd8de7062n%40googlegroups.com<https://groups.google.com/d/msgid/cp2k/4de62505-93ed-48f5-8c79-a23bd8de7062n%40googlegroups.com?utm_medium=email&utm_source=footer>.

--
You received this message because you are subscribed to the Google Groups "cp2k" group.

To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com<mailto:cp2k+uns...@googlegroups.com>.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/8ba5ac86-414e-4c0a-9666-676bf3167152n%40googlegroups.com<https://groups.google.com/d/msgid/cp2k/8ba5ac86-414e-4c0a-9666-676bf3167152n%40googlegroups.com?utm_medium=email&utm_source=footer>.

[ma455...@gmail.com]

Dear Prof. Hutter, 

The files were uploaded by Ron in previous replies. I have attached again here. Please have a look.

Regards,

Hongyang

[Krack Matthias (PSI)]

Hi Hongyang

 

That is strange, I get the exponent 5.2 printed in the full and the soft basis printout with the current CP2K release version 2022.1.

 

Best

 

Matthias

 

From: "cp...@googlegroups.com" <cp...@googlegroups.com> on behalf of "ma455...@gmail.com" <ma455...@gmail.com>
Reply to: "cp...@googlegroups.com" <cp...@googlegroups.com>
Date: Saturday, 13 August 2022 at 02:34
To: "cp...@googlegroups.com" <cp...@googlegroups.com>
Subject: Re: [CP2K:17474] Re: Basis sets for NMR

 

Hi Matthias,

 

I'm a newbie in cp2k. In the basis set file, there is a function row with the exponent of 5.2. However, in the output file, exponent of 5.2 is not seen both in the orbital basis set and the GAPW soft basis set sections. I'm wondering is this reasonable? This phenomenon is observed for both O and H.

Additionally, is there a rule that determines if the exponent should be included in the GAPW soft basis set?

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/8ba5ac86-414e-4c0a-9666-676bf3167152n%40googlegroups.com.

[Ronald Cohen]

Thank you so much for the suggestions. I have not been able to find a difference that explains why one basis set works for nmr and another does not. They all seem to work for molecules like TMS. Again, I have been trying EMSL 6-31G** which seems to give reasonable results for everything, and pcSseg-3, pcSseg-2, and x2c-tavpall-s.1 which do not. However, they all give similar PDOS and energies. They do differ in which states are included in the Soft GAPW set, but that seems is not unexpected.

Some sample results of the tests for a 6 atom cell of ice X (H2O) at high pressure. It has a cubic cell.

O 

x2c-tavpall-s.1>

Total energy: -153.2775234508810

 cat *pdos

# Projected DOS for atomic kind O at iteration step i = 0, E(Fermi) =    0.248490 a.u.

#    MO Eigenvalue [a.u.]      Occupation                s                p                d                f

      1        -18.366065        2.000000        0.99890162        0.00000737        0.00000010        0.00000310

      2        -18.259142        2.000000        0.99916313        0.00000579        0.00000098        0.00000170

      3        -0.627219        2.000000        0.43501495        0.00064038        0.00014477        0.00055218

      4        -0.335410        2.000000        0.78411097        0.00655622        0.00016477        0.00049770

      5        -0.063359        2.000000        0.00569524        0.71262321        0.00810277        0.00102167

      6        -0.047443        2.000000        0.00000000        0.71521699        0.00653835        0.00082817

      7        -0.047443        2.000000        0.00000000        0.71521699        0.00653835        0.00082817

      8          0.235229        2.000000        0.00097929        0.82612193        0.00387274        0.00042107

      9          0.248490        2.000000        0.00000000        0.87600716        0.00171687        0.00032413

      10          0.248490        2.000000        0.00000000        0.87600716        0.00171687        0.00032413

# Projected DOS for atomic kind H at iteration step i = 0, E(Fermi) =    0.248490 a.u.

#    MO Eigenvalue [a.u.]      Occupation                s                p

      1        -18.366065        2.000000        0.00095690        0.00013090

      2        -18.259142        2.000000        0.00079274        0.00003566

      3        -0.627219        2.000000        0.54707105        0.01657667

      4        -0.335410        2.000000        0.03624861        0.17242172

      5        -0.063359        2.000000        0.26420001        0.00835710

      6        -0.047443        2.000000        0.26557542        0.01184108

      7        -0.047443        2.000000        0.26557542        0.01184108

      8          0.235229        2.000000        0.00750535        0.16109963

      9          0.248490        2.000000        0.00314739        0.11880444

      10          0.248490        2.000000        0.00314739        0.11880444

EMSL 6-31G**

Total energy: -153.17163103454254

# Projected DOS for atomic kind O at iteration step i = 0, E(Fermi) =    0.258425 a.u.

#    MO Eigenvalue [a.u.]      Occupation                s                p                d

      1        -18.391769        2.000000        0.99911933        0.00000095        0.00000011

      2        -18.287838        2.000000        0.99931300        0.00000016        0.00000091

      3        -0.621948        2.000000        0.46425916        0.00050230        0.00011727

      4        -0.330907        2.000000        0.85874910        0.00580413        0.00013641

      5        -0.055743        2.000000        0.00664550        0.71507112        0.00712147

      6        -0.039314        2.000000        0.00000000        0.72193723        0.00526740

      7        -0.039314        2.000000        0.00000000        0.72193723        0.00526740

      8          0.243824        2.000000        0.00090864        0.86122143        0.00339013

      9          0.258425        2.000000        0.00000000        0.90746434        0.00061255

      10          0.258425        2.000000        0.00000000        0.90746434        0.00061255

# Projected DOS for atomic kind H at iteration step i = 0, E(Fermi) =    0.258425 a.u.

#    MO Eigenvalue [a.u.]      Occupation                s                p

      1        -18.391769        2.000000        0.00086680        0.00001282

      2        -18.287838        2.000000        0.00065225        0.00003368

      3        -0.621948        2.000000        0.52448247        0.01063880

      4        -0.330907        2.000000        0.03064083        0.10466953

      5        -0.055743        2.000000        0.26540735        0.00575456

      6        -0.039314        2.000000        0.26249598        0.01029939

      7        -0.039314        2.000000        0.26249598        0.01029939

      8          0.243824        2.000000        0.00792147        0.12655833

      9          0.258425        2.000000        0.00392651        0.08799661

      10          0.258425        2.000000        0.00392651        0.08799661

So those are similar, but for the NMR shifts:

x2c-tavpall-s.1

Shielding atom at atomic positions. # tensors printed      6                    
     1O  O          0.000000       0.000000       0.000000
 SIGMA from SOFT J
  XX = 25520.4169  XY = -9631.0091  XZ = -9479.0517
  YX = -9639.7596  YY = 25609.1113  YZ = -9614.6107
  ZX = -9671.1728  ZY = -9507.2257  ZZ = 25776.7155
 SIGMA from LOCAL J
  XX = -1651.2810  XY =   747.3953  XZ =   737.5936
  YX =   410.9983  YY = -1647.5841  YZ =   710.1340
  ZX =   412.4832  ZY =   729.2310  ZZ = -1702.0936
 SIGMA TOTAL
  XX = 23858.3300  XY = -8905.9700  XZ = -8764.3298
  YX = -9252.3112  YY = 23949.7063  YZ = -8927.6288
  ZX = -9282.5714  ZY = -8801.8203  ZZ = 24064.7122
  ISOTROPY =   23957.5828  ANISOTROPY =   13597.1974

     2H  H          1.221236       1.221236       1.221236
 SIGMA from SOFT J
  XX =  7106.0609  XY = -2398.1745  XZ = -2394.6978
  YX = -2400.3592  YY =  7110.2761  YZ = -2395.4492
  ZX = -2406.3554  ZY = -2410.7282  ZZ =  7097.8572
 SIGMA from LOCAL J
  XX =     7.7628  XY =    -6.4796  XZ =    -6.3556
  YX =    -8.0445  YY =     8.8412  YZ =    -5.1384
  ZX =    -8.9021  ZY =    -6.2532  ZZ =     7.3640
 SIGMA TOTAL
  XX =  7103.0178  XY = -2427.0102  XZ = -2423.9251
  YX = -2431.9536  YY =  7107.2965  YZ = -2423.7397
  ZX = -2439.1393  ZY = -2440.8070  ZZ =  7095.3116
  ISOTROPY =    7101.8753  ANISOTROPY =    3650.1793

EMSL 6-31G**

Shielding atom at atomic positions. # tensors printed      6                    
     1O  O          0.000000       0.000000       0.000000
 SIGMA from SOFT J
  XX =   255.4334  XY =    31.8292  XZ =    31.8290
  YX =    31.8292  YY =   255.4334  YZ =    31.8288
  ZX =    31.8290  ZY =    31.8293  ZZ =   255.4335
 SIGMA from LOCAL J
  XX =   198.7708  XY =    -5.8487  XZ =    -5.8488
  YX =    -3.9238  YY =   198.7708  YZ =    -5.8489
  ZX =    -3.9238  ZY =    -5.8486  ZZ =   198.7709
 SIGMA TOTAL
  XX =   461.5226  XY =    24.0795  XZ =    24.0793
  YX =    26.0044  YY =   461.5227  YZ =    24.0789
  ZX =    26.0042  ZY =    24.0797  ZZ =   461.5227
  ISOTROPY =     461.5227  ANISOTROPY =      74.1672
     2H  H          1.221236       1.221236       1.221236
 SIGMA from SOFT J
  XX =     1.0927  XY =    -6.7369  XZ =    -6.7370
  YX =    -6.7370  YY =     1.0927  YZ =    -6.7370
  ZX =    -6.7369  ZY =    -6.7370  ZZ =     1.0927
 SIGMA from LOCAL J
  XX =     2.5894  XY =    -1.0185  XZ =    -1.0185
  YX =    -2.3805  YY =     2.5894  YZ =    -1.0185
  ZX =    -2.3805  ZY =    -1.0185  ZZ =     2.5894
 SIGMA TOTAL
  XX =    11.0005  XY =    -9.6564  XZ =    -9.6565
  YX =   -11.0185  YY =    11.0005  YZ =    -9.6565
  ZX =   -11.0185  ZY =    -9.6565  ZZ =    11.0005
  ISOTROPY =      11.0005  ANISOTROPY =      15.8518

Any suggestions are appreciated!

Sincerely,

Ron

[Ronald Cohen]

Thank you very much for the suggestion. I tried EPSFIT=1e-07 and x2c-tavpall-s.1 is still terrible or even worse:

Shielding atom at atomic positions. # tensors printed      6                    
     1O  O          0.000000       0.000000       0.000000
 SIGMA from SOFT J

  XX = 31638.5971  XY = **********  XZ = **********
  YX = **********  YY = 31748.2856  YZ = **********
  ZX = **********  ZY = **********  ZZ = 31930.4031
 SIGMA from LOCAL J
  XX =  -818.8182  XY =   385.1581  XZ =   378.9823
  YX =   206.3453  YY =  -816.8239  YZ =   366.3063
  ZX =   206.6750  ZY =   375.1724  ZZ =  -845.2912
 SIGMA TOTAL
  XX = 30805.1110  XY = **********  XZ = **********
  YX = **********  YY = 30915.7381  YZ = **********
  ZX = **********  ZY = **********  ZZ = 31071.2597
  ISOTROPY =   30930.7029  ANISOTROPY =   16782.2884

     2H  H          1.221236       1.221236       1.221236
 SIGMA from SOFT J

  XX =  7207.3967  XY = -2385.0046  XZ = -2381.5027
  YX = -2387.5623  YY =  7210.8500  YZ = -2381.6744
  ZX = -2393.5466  ZY = -2397.6103  ZZ =  7197.9781
 SIGMA from LOCAL J
  XX =    -3.7427  XY =     0.6390  XZ =     0.5561
  YX =     0.0340  YY =    -3.0295  YZ =     1.9120
  ZX =    -0.7421  ZY =     1.0471  ZZ =    -3.8553
 SIGMA TOTAL
  XX =  7188.9861  XY = -2406.0057  XZ = -2403.0887
  YX = -2410.4407  YY =  7192.0970  YZ = -2402.2695
  ZX = -2417.4847  ZY = -2419.7320  ZZ =  7180.2706
  ISOTROPY =    7187.1179  ANISOTROPY =    3618.1086

with EPSFIT=1e-7 the soft basis shows:

                  Set   Shell   Orbital            Exponent    Coefficient

                          1       1    1s                5.119805       0.000000
                                                         1.176403       0.000000
                                                         0.465461       0.000000
                                                         0.185009       0.000000

but it does not seem to have helped.

Sincerely,

Ron

[Ronald Cohen]

So what is it about periodic boundary conditions that can made an all-electron basis set fail for nmr with GAPW? The parameters discussed in the forum for GAPW do not seem specific for periodic boundary conditions. If this is a convergence issue rather than a bug, I would appreciate any guidance on what I should tune. Thank you!

Sincerely,

Ron