Beilstein J. Org. Chem. 2013, 9, 1083–1092.
SOMOs for spiro-cyclisations of 5a and 11a. For both radicals converged in indicating that the “default” ring closure mode for
the iminyl units approach the acceptor rings from vertically iminyl radicals is ortho-cyclisation onto suitably situated
above; the angle between the ArC(R)=N unit and the plane of aromatic rings. The exception to this rule was spiro-cyclisation
the acceptor ring is ~90° for both TSs. This ensures that overlap onto benzofuran and benzothiophene rings at low temperatures.
between the incoming iminyl orbital and the acceptor ring Even for these acceptors, however, thermodynamic control
orbitals is optimum. The bottom line shows the TS structures ensured that the only isolable products were from ortho-clos-
and SOMOs for the ortho-cyclisations of 5a and 11a. For both ures. Rate parameters for the spiro-cyclisation of 11a, estim-
radicals the architectures oblige the iminyl units to approach the ated by the steady-state kinetic EPR method, showed the
acceptor rings at much more oblique angles. Thus, overlap of process to be about an order of magnitude slower than for
the in-plane iminyl orbital with the acceptor ring orbitals is archetype C-centred radicals. This was in harmony with the
poorer than for spiro-cyclisation. However, this is offset by the generally slower reaction rates of iminyl radicals compared with
greater resonance delocalisation in the TS frontier orbitals for alkyl radicals. DFT-computed energetics were in good agree-
ortho-cyclisation (see Figure 4). There is a trade-off between ment with experiment and supported the idea of thermody-
these two factors and the default is that the resonance delocal- namic control for the cyclisation of 11a,b.
isation factor in ortho-cyclisation prevails. Evidently 11a,b, are
exceptional in that the iminyl/acceptor overlap factor outweighs Experimental
the resonance delocalisation factor.
EPR spectra were obtained at X-Band on Bruker EMX 10/12
spectrometers at St. Andrews and Manchester. Oxime carbon-
ates (2 to 15 mg) and MAP (1 equiv wt/wt) in t-BuPh or
benzene (0.5 mL) were prepared in quartz tubes and deaerated
by bubbling with nitrogen for 15 min. Photolysis in the resonant
cavity was by unfiltered light from a 500 W super-pressure
mercury arc lamp or, in the Manchester experiments, from a
Luxtel CL300BUV lamp. EPR signals were digitally filtered
and double integrated by using the Bruker WinEPR software,
and radical concentrations were calculated by reference to the
double integral of the signal from a known concentration of the
stable radical DPPH run under identical conditions. The
majority of EPR spectra were recorded with 2.0 mW power,
0
.8 Gpp modulation intensity and gain of 106.
QM calculations were carried out by using the Gaussian 09
program package. Geometries were fully optimised for all
model compounds. Optimised structures were characterised as
minima or saddle points by frequency calculations. The experi-
mental kinetic and spectroscopic data was all obtained in the
nonpolar hydrocarbon solvents t-BuPh or cyclopropane. Solvent
effects, particularly differences in solvation between the neutral
reactants and neutral transition states, were therefore expected
to be minimal. In view of this, no attempt was made to model
the effect of the solvent computationally.
Figure 4: DFT computed structures for 5a, 11a and their cyclisation
transition states (TS). Top line: spin density distribution maps for
iminyls 5a and 11a. Middle line: spiro TSs, r(N···C) for 5a = 1.928 and
for 11a = 2.064 Å. Bottom line: ortho TSs, r(N···C) for 5a = 1.954 and
for 11a = 2.027 Å.
Supporting Information
Supporting Information File 1
Conclusion
General procedures. Preparation and characterization data
of oxime carbonates. Sample EPR spectra and kinetic data.
Ethoxycarbonyl oximes of aromatic and heteroaromatic ketones
are easily made and handled and have long shelf lives. We have
found them to be outstanding precursors for the photochemical
generation of aromatic iminyl radicals. Product analyses,
EPR spectroscopic observations and DFT computations all
1090