complex we generally find that as the carbene-TMB
interaction distance decreases, the binding enthalpy in-
creases. In type A complexes, the carbenic Cl atom points
toward the open back-side of an adjacent MeO group and
does not appear to be sterically encumbered. In type B
complexes, the Cl atom is situated above an adjacent
oxygen atom and is sterically encumbered by the MeO
group. As illustrated in Figure 2, the Cl-MeO steric
interaction forces the aromatic rings in Type B complexes
to rotate relative to each other, and consequently, there is
more stabilizing face-to-face ring overlap in complexes of
type A. For a particular para substituent (X), the A - B
energy difference is 1-2 kcal/mol (Table 3).
The entropy changes associated withcomplex formation
(ΔSo ≈ -35 eu, Table 3) are overestimated by the electro-
nic structure calculations, which refer to idealized gas
phase rather than solution phase conditions. Thus, despite
the very favorable complexation enthalpies obtained at the
B97D/6-311þG(d) level, the computed Gibbs free energies
of complexation are too large, and the computed equili-
brium constants are consequently smaller than the experi-
mentally derived quantities. Unquestionably, electron-
withdrawing groups (NO2, CF3, and Cl) in the p-position
lead to morefavorable free energiesof complexation (larger
K), and electron-donating groups (Me and MeO) disfavor
complex formation. The relative magnitudes of the sub-
stituent effects appear to be reproduced much better for the
electron-withdrawing groups than for the electron-donat-
ing groups; in particular, the destabilizing effect of MeO is
seriously underestimated by the calculations.17 Neverthe-
less, a plot of the Gibbs free energy changes, or equiva-
lently, a logarithmic plot of the equilibrium constants
versus the Hammett substituent parameters, displays a re-
asonable straight line. For the B complexes, the slope F =
1.95 (r = 0.982) is in satisfactory agreement with the
experimentally defined value of 2.48 (Figure 4); for the A
complexes F = 2.49 (r = 0.979).13 These Hammett correla-
tion diagrams appear in the Supporting Information.
p-X-PhCCl/TMB complex formation is accompanied
by the appearance of an electronic absorption band in
the 400-500 nm region. The new band embodies consider-
able π(HOMO, TMB) f p(LUMO, carbene) charge trans-
fer character, although neither experimental (Table 1) or
computed (Table S-1 in Supporting Information) band
maxima shift in a systematic manner as the para sub-
stituent (X) changes. For a particular substituent, the B
complex absorbs at lower energy with a smaller oscilla-
tor strength than the A complex (Table S-1), qualita-
tively in accord with the lesser overlap and interaction
between the aryl rings present in the B complexes
(Figure 2). In the experimental spectra, we therefore
associate the higher-energy band with the A-type com-
plexes and the lower-energy band with the B-type com-
plexes. Flanking these signature bands are mixed
absorption bands reflecting the coexistence of free car-
benes and carbene complexes in solution.
The experimental spectra appear to show larger in-
tensities in the low-energy bands, which we ascribe to B-
type complex formation, than in the high-energy bands
(A-type complexes). In contrast, the calculations predict
that a carbene/TMB solution mixture at equilibrium
should contain more complexes of type A than of type
B (Table 3); furthermore, the signature transition of
complex formation is predicted to carry larger intensity
in the A-type complexes (Table S-1). The electronic
structure calculations locate the single most stable car-
bene/TMB configuration on the potential energy surface
in an idealized gas phase. It is plausible that on the
“solution free energy surface” complexes of the less
geometrically restricted B-type actually dominate, so
that the absorption toward the red appears more intense
in the experimental spectra as a result of a larger
equilibrium concentration of complexes.
In summary, using a combination of LFP and DFT
techniques, we have determined equilibrium constants for
a family of p-substituted phenylchlorocarbene/TMB com-
plexes. The substituent effects on the experimental and
computational equilibrium constants are well correlated
by the Hammett equation.
Acknowledgment. We are grateful to the National
Science Foundation for financial support.
(16) All calculations made use of the Gaussian 09, Revision A02
program package: Frisch, M. J. et al. . Gaussian, Inc.: Wallingford, CT,
2009; see Supporting Information for additional computational details and
the complete reference for Gaussian 09.
(17) It was necessary to tighten the geometry optimization criteria
and to increase the size of the integration grid relative to their default
values to reach even this level of agreement and consistency. Test
calculations with larger basis sets and/or different dispersion-corrected
functionals did not show any improvement for X = MeO.
Supporting Information Available. Figures S-1-S-51,
Table S-1, computational details, and optimized geometries,
absolute energies, electronic excitation energies and oscilla-
tor strengths of all relevant species. This material is available
Org. Lett., Vol. 13, No. 5, 2011
1201