Dear QChem developers and users,
I am running an SF-TDDFT (spin-flip TDDFT) calculation on an organic molecule (gas phase). As stated in the manual, SF-TDDFT does not support formatted checkpoint or cube file generation, and the only available orbital output appears to be via Molden / .mo files.
The calculation runs smoothly and produces:
- sf_no.mo, sfnto.mo, sf*_ndo.mo files;
- a .plots directory containing the sa_nto_nto.mo file.
However, I am unable to visualize or post-process these orbitals with common programs (IQmol, Avogadro, Multiwfn).
In particular, when loading the Molden file generated by Q-Chem into Multiwfn, I obtain the following error:
Loading various information of the wavefunction
Loading basis set definition…
The highest angular moment basis functions is D
All basis functions are Cartesian type
Loading orbitals…
forrtl: severe (24): end-of-file during read, unit -5, file Internal List-Directed Read
The $rem section used is:
$rem
method lrc-wpbe
omega 200
lrc_dft true
basis def2-SVP
unrestricted true
cis_n_roots 15
spin_flip true
cis_singlets true
cis_triplets true
state_analysis true
rpa false
cis_ampl_print 1
print_orbitals 5000
molden_format true
$end
The Molden file is partially read correctly (geometry and basis are detected) if I rename the files from .mo to .molden, but orbital data cannot be parsed by Multiwfn or other viewers.
My questions are:
- Is visualization or external post-processing of SF-TDDFT orbitals
(NOs / NTOs / response orbitals) officially supported in Q-Chem?
- If so, is there a recommended workflow or keyword combination?
Thank you very much for your time and help.
Best regards,
Arianna
University of Parma
What version of Q-Chem are you running? With the latest (v.6.4), I was able to successfully generate a Molden-formatted file (and load it into Molden to look at orbitals) using your $rem section.
Hi, the Molden files produced can be read by a variety of programs (like Molden itself or jmol). Not quite sure about multiwfn. The problem is that Molden format is not fully standardised. So, sometimes problems pop up with various codes.
We have been using Molden files from Q-Chem and fed them into Molden2AIM. Then you can get wfn/wfx files. Maybe those work better with multiwfn.
BTW, for cube files you can use something like
STATE_ANALYSIS true
WFA_REF_STATE 1
NTO_PAIRS 3
MAKE_CUBE_FILES true
WFA_LEVEL 4
WFA_ORB_THRESH 3
$end
$plots
Write cube files for all 4 states
35 -3.5 3.5
35 -3.5 3.5
15 -1.5 1.5
0 4 0 0
1 2 3 4
$end
See Example 10.9 here
https://manual.q-chem.com/latest/ex_StateAnal_MolDen.html
And you should also make sure to specify the appropriate WFA_REF_STATE
FCheck files it’s GUI=2
Or to be more specific regarding the recommended workflow. With your original input for Molden files, I would go via jmol. Optionally you can use TheoDORE as an interface layer that automates the plotting of the orbitals
https://theodore-qc.sourceforge.io/docs/orbitals_and_densities.html#density-plotting-jmol-molden
theodore jmol_mos *.mo
If you use my second workflow to produce cube files, then I would go via VMD. In this case there is also a utility script theodore vmd_plots that allws you to automate things.
Thanks to both of you!
I am using the Q-Chem 6.1.1 version and I was actually having some troubles to visualize the orbitals with Molden, hence I switched to Multiwfn to produce cubefiles.
I’ll try to use the latest suggestions and will come back if anything goes wrong.
Thanks again!
It’s also possible to get Q-Chem to generate cube files for the NTOs directly, although I think you have to do it one state at a time. Here’s the relevant section of the sample job:
$rem
method pbe0
basis 6-31+G*
cis_n_roots 3
cis_triplets false
cis_ampl_print 10 ! set print thresh to 0.10 (default=0.15)
cis_ampl_anal true ! calculate NTOs and other amplitude analysis
nbo 2 ! NBO analysis for ground & excited states
make_cube_files ntos ! generate cube files for the NTOs...
cubefile_state 1 ! ...associated with the first excited state
$end