I am attempting to locate the transition state (TS) for the reaction of 4-chloropyridine with the azide anion, which corresponds to the Meisenheimer complex model. However, OptKing (opt_type="ts") fails to identify a valid transition state, as checking the vibrational frequencies does not return the expected one imaginary frequency criterion.
psi4.set_options({
“reference”: “UHF”,
“geom_maxiter”: 2000,
“maxiter”: 2000,
“basis”: “Def2-SVPD”,
“e_convergence”: 1e-5,
“d_convergence”: 1e-5,
“ints_tolerance”: 1e-8,
“opt_coordinates”: “internal”,
“full_hess_every”: 0,
“dynamic_level”: 0.05, # Tested from 0.05 to 0.3
“MAX_FORCE_G_CONVERGENCE”: 3e-3 # Also tested 1e-4
})
I have tested different functionals, including B3LYP and M06-2X, but without success.
Approach Taken:
- Preliminary SCF scanning – I performed an iterative scan across 3000 coordinate points along the nucleophile and leaving group axis to probe the potential energy surface. However, no clear local saddle point was identified.
- Exploring different orientations of the azide anion, considering both the LUMO symmetry of 4-chloropyridine and the HOMO of azide to optimize orbital overlap.
- Comparative analysis – I have successfully located TS structures for other SNAr reactions involving nucleophiles such as the fluoride anion attacking chlorobenzene or dinitrochlorobenzene. In those cases, **OptKing reproduced geometries reported in the literature.
Nucleophilic aromatic substitution (SNAr) reactions with highly polarizable nucleophiles such as the azide anion tend to proceed via a concerted mechanism, where no distinct Meisenheimer complex is observed. This could mean that the transition state is more subtle, making it harder to locate.
I am reaching out to the community to ask:
Have I overlooked any critical Psi4 settings that might improve TS search reliability?
Are there any methodological refinements I could apply to better locate the TS for this system?
-best regards