[CP2K-user] [CP2K:18281] Difference between CP2K and Gaussian on 4th period transition metals

[Jürg Hutter]

Hi

if you are using the same method (functional, TDDFT vs. Tamm-Dancoff) with the same
numerical settings (pseudpotentials and basis sets), the two codes give the same answer.
However, if you start to make changes in CP2K results will differ.
It is up to you to verify, which of the two settings in the codes are more reliable.
The first question will be what is the target result?

regards
JH

#These are the coefficients used for B3LYP using VWN3, this is the default Gaussian definition, but not recommended
    &XC
      &XC_FUNCTIONAL
       &LYP
         SCALE_C 0.81
       &END
       &BECKE88
         SCALE_X 0.72
       &END
       &VWN
         SCALE_C 0.19
         FUNCTIONAL_TYPE VWN3
       &END
       &XALPHA
         SCALE_X 0.08
       &END
      &END XC_FUNCTIONAL
      &HF
        &SCREENING
          EPS_SCHWARZ 1.0E-10
        &END
        &MEMORY
          MAX_MEMORY  100x
        &END
        FRACTION 0.20
      &END
    &END XC

________________________________________
From: cp2k at googlegroups.com <cp2k at googlegroups.com> on behalf of Elvis <elvisnwh at gmail.com>
Sent: Wednesday, December 28, 2022 5:13 PM
To: cp2k
Subject: [CP2K:18265] Difference between CP2K and Gaussian on 4th period transition metals

Hi,

I am seeing very dramatic difference between TDDFT spectra obtained by Gaussian and CP2K. I am not sure if I should just trust the built in basis set / pseudopotential.

First of all, for some systems, e.g. H2O, pyridine or pyZnCO3, the spectra are very similar,  when using large cell or with wavelet solver, and using large enough cutoff. One example is shown below. Even though CP2K and Gaussian uses very different pseudopotentials and also very different algorithms in implementations, we can see qualitative match between the two software.

(e.g. pyZnCO3)
[Zn.png]

However, when the Zn is replaced by Cd, or other 4th period transition metals. The spectra are no longer similar. One example is shown below.

(e.g. pyCdCO3, where the spectra are very different. )

[pyCdCO3.png]

I have tried different cutoff values, different cell sizes, non-periodic wavelet solver, no or larger cutoff for truncated 1/r potential for HFX, finer grid etc. and none of them resulted in better match between the two softwares. It should be pointed out that the good match for H2O, py or even pyZnCO3 (where the only difference is Zn replacing Cd) showed that the software was set up correctly and the mismatch should not be due to wrong software setting. This mismatch also happens for test molecule, CdO and PdO, and therefore not an isolated example.

I can understand that this difference is due to the different pseudopotential used with CP2K and Gaussian. However, in Gaussian, even different ECPs (Lanl2dz or SDD) gave qualitatively similar results.

I am a bit concerned about the big difference between the two softwares when Cd or other 4th period transition metal (e.g. Pd) is involved. Can anyone comment on this? Any suggestion or comment for me will be greatly appreciated. I am not sure if I should just trust the built in basis set / pseudopotential and just proceed with my complex system.

Thanks again.




--
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+unsubscribe at googlegroups.com<mailto:cp2k+unsubscribe at googlegroups.com>.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/7693fb34-974b-475c-bd44-7f09f48f0ddfn%40googlegroups.com<https://groups.google.com/d/msgid/cp2k/7693fb34-974b-475c-bd44-7f09f48f0ddfn%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+unsubscribe at googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/ZR0P278MB0759296688AF9D3F1E26A2009FF49%40ZR0P278MB0759.CHEP278.PROD.OUTLOOK.COM.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Zn.png
Type: image/png
Size: 152224 bytes
Desc: Zn.png
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20230103/83ea5398/attachment-0002.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: pyCdCO3.png
Type: image/png
Size: 194636 bytes
Desc: pyCdCO3.png
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20230103/83ea5398/attachment-0003.png>