F-SAPT analysis on transition states

I am currently familiarising myself with F-SAPT analysis on transition states, for which I am trying to reproduce results of a paper from Prof. Sherrill’s group (DOI: 10.1039/c5cp07281f).

I am working with the RR_ENT-SYN_1 transition state.

However, I am getting an error due to creation of ‘link bonds’. I am not able to understand what they are and how I should correct my input file.

I have attached an image of the output file

Any help regarding this would be appreciated!

Looks like you may be partitioning your molecule such that not all pieces are closed shell (hence the warning of charge incompatible with singlet).

Barring that, you’ll have to post the input file. Maybe @BrandonB can consult.

The partitioning of the molecule is closed shell because the SCF calculations don’t show any error. If my partition was not closed shell even the SCF calculations wouldn’t be done.

This error pops up only when the F-SAPT calculations start to run.

I have attached my input file below:

memory 1 GB

molecule mol {
0 1
O -2.09080353 -2.32271788 -0.13680149

H -0.98362848 -1.81070860 -0.74427008

O -3.44688855 -2.08959932 1.62287150

H 1.64508302 0.11815505 2.52135970

H -2.33364003 0.20000484 -2.31733416

C 1.00249418 0.90539504 2.11929114

H -4.48849883 -0.70001211 -0.86808663

H -1.86134690 1.86968967 -2.01677616

C -2.40233474 1.02610210 -1.60417397

C -2.88460909 -0.10399470 0.48054963

H -3.91635717 2.29278052 -0.74723701

C -2.77659945 -1.62998698 0.69163491

C -0.65811838 0.96648889 0.19542889

C -0.08603926 0.28990263 1.27339373

C -4.19960661 0.18322038 -0.29509699

N -1.83415048 0.57925989 -0.31902023

H -0.71458720 -0.41431133 1.80456187

H -2.90542298 0.36208143 1.46500599

H -4.49841719 1.35106295 -2.12994534

C -3.85325040 1.32652894 -1.25246634

H -5.01101186 0.41733955 0.39042567

H 0.52608713 1.36857422 2.99244469

C 0.11993320 2.07819516 -0.46195920

C 1.84345901 1.95332385 1.38594177

H 0.77380970 1.64729325 -1.22733846

C 0.95399462 2.87089169 0.55410428

H -0.55389409 2.76063238 -0.97569374

H 2.56474524 1.46378908 0.73041106

H 0.27732414 3.42711851 1.21056857

H 1.55392075 3.60832100 0.01681598

H 2.41838333 2.53378718 2.11093514

0 1
C 0.76605487 -1.35176510 -0.04638422

O -0.02116915 -1.45848109 -1.07766692

C 3.88978666 -0.00990930 -1.64791414

C 4.43866294 -0.87926637 0.53729422

H 5.86117432 0.03879013 -0.78770611

C 2.15909060 -0.97905251 -0.26651793

H 2.81310660 -1.74427871 1.63846621

H 1.83472208 -0.20323412 -2.24522330

H 0.55681420 -1.98869340 0.80933209

H 4.19610883 0.46251175 -2.57273917

H 5.16901259 -1.08465593 1.30953263

C 4.82785157 -0.25108394 -0.64377757

C 3.11248867 -1.24601846 0.72487957

C 2.56437813 -0.37551024 -1.46574611
symmetry c1

set {
basis jun-cc-pVDZ
scf_type df
guess sad
freeze_core true


Hi @shreesowndarya, you will have to lower your charge completeness threshold from the default 0.8 value. I lowered this threshold to 0.6 to converge this result with the expected charge/multiplicity. Otherwise, your input looks fine!

I guess that is because the code checks whether or not partial charges are compatible with the charge that was given in the input. The default threshold is usually fine but there is more charge moving around in transition states.

Yup I guess that’s the reason. Anyways I have fixed the problem! Thanks :slight_smile:


I would like to know if this F-SAPT analysis could be done on transition states involving metal such as Rh, Pd etc.,

Kindly let me know about this.

Thank you.


As far as I know, F-SAPT can in principle be done on transition states involving metals. However, if you use ECPs you should know that the SAPT code is not well-tested for use with ECP (as noted in another thread).


Thanks for letting me know that F-SAPT can be used on transition states involving metals.

I have been working with systems that dont involve metals and I was able to obtain the results i wanted till now.

However, I started working on a system involving Rh with the LANL2DZ basis set. I am able to pass the basis set externally by creating a .gbs file for the Rh atom.

I am getting the following error:

Could you please let me know how to go about fixing this.

Thank you!


if I remember correctly, the density fitting approximation has to be used in the SAPT code at some point, which means that you need an appropriate fitting basis set in addition to a regular basis set for Rh. If you can find a reasonable fitting basis set somewhere (for example on the EMSL basis set exchange website), you should be able to use it.

1 Like


I found this density fitted basis set for metals avaliable on PSI4 itself. Would this solve the error I described above?

The Path is as follows on GitHub : psi4/psi4/share/psi4/basis/def2-qzvpp-jkfit.gbs

It contained basis set for certain metals like Rh which I need to use.

I think that is what you need. I also just noticed that in the error you copied, the path to that folder seems to contain some very unusual characters, so maybe that could also be a problem. Try to include the jkfit basis first, i.e. make sure it is in your Psi4 basis path and add the appropriate keyword to your input file (I can’t remember what it is but it should be in the manual).

If that still gives you the above error, consider changing your path so that it only contains standard ASCII characters to make sure Psi4 can recognize it properly.

The path to the folder contains those characters because when we installed the program the folder was named with ‘.’ ‘/’ and ‘?’. However we are able to access the files properly. Hence I guess that is not the source of any error. I was able to successfully complete the calculations after I added those respective basis sets.

Thank you for the help!

Sorry to bring up this topic again!

However, I have been trying to quantify certain interactions in a transition state, the following issues pop up

The position of the methyl in the optimised geometry, are such that they are not expected to interact with the phenyl ring in the transition state. However, the F-SAPT analysis shows interactions i.e., a negative total energy. This is counter-intuitive on geometric grounds. Could you please let me know if there is some basis for such an interaction identified by F-SAPT analysis?

Similarly, all other interactions which seem to exist in the transition state are turns out to be positive values. Could you please comment?

I hope this could be addressed I can provide specific details to the people who would help me in solving this problem.