Author Topic: ridft+dsp problem with Curium/HOH  (Read 4838 times)

corey.taylor

  • Newbie
  • *
  • Posts: 6
  • Karma: +0/-0
ridft+dsp problem with Curium/HOH
« on: November 15, 2019, 02:21:01 PM »
I set up a bond scan between a single HOH and Cm and am getting some errors. I ran define and allocated basis set def-TZVP (no def2 for Cm in our basen) for all atoms, ecp-60-SV added for Cm. Dispersion correction and RIwere both turned on.

Upon running ridft, it appears the DFTD3 routine crashes with the following outputted in job.last:

Code: [Select]
|                                 |
 |         DFTD3 V3.1 Rev 0        |
 | S.Grimme, University Bonn       |
 |            June  2014           |
 |   see standalone version        |
 |       dftd3 -h for options      |
 |_________________________________|
 
 Please cite DFT-D3 work done with this code as:
 S. Grimme, J. Antony, S. Ehrlich and H. Krieg,
 J. Chem. Phys, 132 (2010), 154104.
 If used with BJ-damping cite also
 S. Grimme, S. Ehrlich and L. Goerigk,
 J. Comput. Chem. 32 (2011), 1456-1465
 For DFT-D2 the reference is
 S. Grimme, J. Comput. Chem., 27 (2006), 1787-1799
 
C6 coefficients used:
                     2  C6 for element                      1
Z=  1 CN= 0.912     C6(AA)=    3.03
Z=  1 CN= 0.000     C6(AA)=    7.59
                     3  C6 for element                      8
Z=  8 CN= 0.000     C6(AA)=   15.51
Z=  8 CN= 0.993     C6(AA)=   12.82
Z=  8 CN= 1.989     C6(AA)=   10.37

#           XYZ [au]              R0(AA) [Ang.]  CN          C6(AA)     C8(AA)   C10(AA) [au]
  1  -0.27766   0.00012   0.00051   ^@^@  0.000  0.000************         0.0         0.0
  2   3.69128  -0.00873   0.01775   o   1.241  1.985        10.4       210.2      5195.4
  3   5.03685   0.06287   1.26045   h   1.091  0.993         3.1        37.4       553.8
  4   4.41182   0.04607  -1.66684   h   1.091  0.993         3.1        37.4       553.8

molecular C6(AA) [au] = ************
                     1                     2                     8
                    96
  --- ERROR in DFT-D dispersion routines ---
 program stopped due to: radius missing

Followed by MPI crash messages.

The original paper (https://aip.scitation.org/doi/full/10.1063/1.3382344) states coefficients were calculated H-Pu which would appear to suggest I'm out of luck. I have a much larger system (protein binding pocket) to calculate with Curium so dispersion effects will be important.  Is anyone able to confirm that Cm for sure isn't covered by the correction and that there are no plans to change this?

Cheers all.

corey.taylor

  • Newbie
  • *
  • Posts: 6
  • Karma: +0/-0
Re: ridft+dsp problem with Curium/HOH
« Reply #1 on: November 15, 2019, 02:37:11 PM »
Forgot to mention, I know there's an extension to the D3 model available (D4) but as far as I'm aware, it doesn't extend the element range beyond Pu.

uwe

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 558
  • Karma: +0/-0
Re: ridft+dsp problem with Curium/HOH
« Reply #2 on: November 15, 2019, 04:38:24 PM »
Hi,

you are right, DFT-D3 is parametrized from Z=1 to 94, so from H to Pu.

But DFT-D4 contains parameters up to Z=118, so I'd simply try it. DFT-D4 is ready for usage in Turbomole versions 7.3 and newer.

Regards,

Uwe

corey.taylor

  • Newbie
  • *
  • Posts: 6
  • Karma: +0/-0
Re: ridft+dsp problem with Curium/HOH
« Reply #3 on: November 15, 2019, 04:53:45 PM »
Cheers Uwe. However...

Code: [Select]
^@^@^@
   |            DFT-D4            |
   |          ==========          |
   |  E. Caldeweyher,  S. Grimme  |
   |   University of Bonn, MCTC   |
   |     02/2018, Version 1.0     |
   '------------------------------'
   '------------------------------'
  This code comes without any warranty
  This code comes without any warranty

 Found param_gfn2.xtb file !
========================================================================
    G F N  parametrization, version 3
   Geometries Frequencies Non-covalent interactions
 /bigdata/chemistry/opt/TURBOMOLE/parameter/param_gfn2.xtb
========================================================================
 method parameters:
------------------------------------------------------------------------
k(s)        :               1.8500
k(p)        :               2.2300
k(d)        :               2.2300
k(f)        :               2.0000
kEN (H0ij)  :              -2.0000
D3 a1       :               0.5200
D3 a2       :               5.0000
D3 s8       :               2.7000
D3atm       :               5.0000
alphaj      :               2.0000
AESdamp     :               4.0000
AESrad      :               3.0000
expscalexp  :               5.0000
------------------------------------------------------------------------
------------------------------------------------------------------------
 xTB basis
------------------------------------------------------------------------
#ERROR! f-functions not implemented

I didn't find an option in define so just changed $disp3 to $disp4. Did I miss something in the setup?

corey.taylor

  • Newbie
  • *
  • Posts: 6
  • Karma: +0/-0
Re: ridft+dsp problem with Curium/HOH
« Reply #4 on: November 26, 2019, 11:21:07 AM »
Just wanted to bump this in case it got missed. Is there an obvious thing I missed in the setup? Perhaps fixed in TM 7.4 (fairly sure we're using 7.3)? Or perhaps do I need to contact Grimme and get the data directly? Appreciate any insight.