2
$\begingroup$

I've been getting the following error whenever I use CCSD(T) and the ORCA Forums don't have an answer to this one:

[file orca_numfreq/numfreq_utils.cpp, line 421]: Error (ORCA_NUMCALC): NCalcs not set

The input file is as follows: 

!CCSD(T) DKH DEF2-TZVP FREQ VeryTightSCF

* xyz 0 1
Zn 0.000 0.000 0.000
*

Is there something missing here? Sorry, I am still relatively new to ORCA and DFT calculations. I am trying to calculate the ionization enthalpy of Zn, and I'm basing this on a paper.

$\endgroup$
3
  • 1
    $\begingroup$ +1 and welcome to our new community! Thank you for your contribution and we hope to see much more of you in the future!!! Just to let you know, CCSD(T) is not a DFT calculation, so I removed the DFT tag from your question. Also, hopefully you're able to make a more recognizable username in the future. Hopefully this community can help you! $\endgroup$ Sep 27, 2023 at 0:15
  • $\begingroup$ Thanks Nike, that helps to clarify a lot. I'm still dipping my toes into this field. The software seems to easy to use and I'm getting hang of the syntax. Are there any references or a direction you can point me to? I eventually wanna use this in my everyday work. $\endgroup$
    – znc204
    Sep 28, 2023 at 13:54
  • $\begingroup$ If you want a reference for something specific, you can ask a question and use the reference-request tag. If you just say "are there any references you can point me to, for nothing in particular", then you will get no answers in particular. $\endgroup$ Sep 28, 2023 at 21:03

1 Answer 1

3
$\begingroup$

There is no point of doing frequency calculations for a single atom, since atoms do not have vibrations, which means that they do not have vibrational frequencies. The enthalpies of single atoms can be trivially calculated by hand, from its definition: $$ H = E_{ele} + ZPE + (U(T)-U(0)) + pV $$ For atoms, and under the standard approximations, ZPE is zero (because there are no vibrations), U(T)-U(0) (the difference of internal energies at the target temperature and at 0 K) is completely contributed by translational degrees of freedom and amounts to (3/2)RT, and pV=RT. Therefore, to get the enthalpy of a single atom, you just need to add (5/2)RT to the electronic energy of the atom, $E_{ele}$. Note that this is also why the isobaric molar heat capacities of monoatomic gases are (5/2)R. Of course, if you are solely interested in calculating the ionization enthalpies, then things are even simpler: as the enthalpy corrections of both the neutral atoms and the atomic cations are (5/2)RT, they cancel out when calculating the ionization enthalpies, so that you don't even need to calculate the (5/2)RT correction.

Furthermore, as the paper you mentioned is quite old, it may be worthwhile to consider using methods developed (or gained popularity) in the past two decades instead, which gives you much more accurate energies at affordable cost. For example:

  1. The electronic energies may be calculated using a selected CI, a DMRG, or a FCIQMC method, which give results that are much closer to full CI than CCSD(T) gives. Within ORCA, the most convenient choice may be the ICE method, which belongs to the category of selected CI methods, and has the extra advantage that its wavefunction has no spin contamination.
  2. Explicitly correlated methods (e.g. F12 or transcorrelated methods) or complete basis set (CBS) extrapolations can be used to minimize the basis set error. ORCA supports both F12 and CBS extrapolations.
  3. Spin-orbit coupling should be considered, since single atoms have unquenched orbital angular momenta, which may give a non-negligible energy contribution when they couple with the spin angular momenta.
  4. If possible, and particularly if you want to study heavier elements, use X2C instead of DKH. X2C is not much more expensive than DKH, but provides essentially exact relativistic energies compared to solving the Dirac equation. For atoms you can even remove the word "essentially" - X2C is an exact method, and it acquires minuscule errors for molecules only because the inter-atomic blocks of certain matrices are typically neglected for efficiency reasons.
$\endgroup$
3
  • $\begingroup$ Wow! Thank you so much. I took the advice to heart and found major differences in computational time and much better agreement with experiment. I'm still relatively new to the field, but would you mind clarifying what the abbreviations in point #1 are and their differences? Thank you for your time and help. $\endgroup$
    – znc204
    Sep 28, 2023 at 14:00
  • $\begingroup$ @znc204 have you looked up DMRG and FCIQMC and "selected CI" online? I strongly recommend to do that before asking other people to spend time clarifying "obvious" abbreviations. ICE might be harder to find because it means more than one thing, but it's basically Frank Nesse's method for trying to solve "full" CI problems, and I suggest to look at the ORCA manual or Frank's papers before asking about it here. If you do have to ask something because you can't find the answer online, I recommend to ask a new question to the whole community rather than asking one person in a comment. $\endgroup$ Sep 28, 2023 at 14:58
  • 1
    $\begingroup$ @NikeDattani, Got it. Thanks for the tip. $\endgroup$
    – znc204
    Sep 28, 2023 at 17:22

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .