Base-catalyzed dehydration of 3-substituted benzene cis -1,2-dihydrodiols: Stabilization of a cyclohexadienide anion intermediate by negative aromatic hyperconjugation

Article abstract of DOI:10.1021/ja304366j

Evidence that a 1,2-dihydroxycyclohexadienide anion is stabilized by aromatic "negative hyperconjugation" is described. It complements an earlier inference of "positive" hyperconjugative aromaticity for the cyclohexadienyl cation. The anion is a reactive intermediate in the dehydration of benzene cis-1,2-dihydrodiol to phenol. Rate constants for 3-substituted benzene cis-dihydrodiols are correlated by σ^- values with = 3.2. Solvent isotope effects for the reactions are kH₂O/kD ₂O = 1.2-1.8. These measurements are consistent with reaction via a carbanion intermediate or a concerted reaction with a "carbanion-like" transition state. These and other experimental results confirm that the reaction proceeds by a stepwise mechanism, with a change in rate-determining step from proton transfer to the loss of hydroxide ion from the intermediate. Hydrogen isotope exchange accompanying dehydration of the parent benzene cis-1,2-dihydrodiol was not found, and thus, the proton transfer step is subject to internal return. A rate constant of ～10¹¹ s^-1, corresponding to rotational relaxation of the aqueous solvent, is assigned to loss of hydroxide ion from the intermediate. The rate constant for internal return therefore falls in the range 10¹¹-10¹² s ^-1. From these limiting values and the measured rate constant for hydroxide-catalyzed dehydration, a pK_a of 30.8 ± 0.5 was determined for formation of the anion. Although loss of hydroxide ion is hugely exothermic, a concerted reaction is not enforced by the instability of the intermediate. Stabilization by negative hyperconjugation is proposed for 1,2-dihydroxycyclohexadienide and similar anions, and this proposal is supported by additional experimental evidence and by computational results, including evidence for a diatropic ("aromatic") ring current in 3,3-difluorocyclohexadienyl anion.

Full text of DOI:10.1021/ja304366j

Article
pubs.acs.org/JACS
Base-Catalyzed Dehydration of 3‑Substituted Benzene cis-1,2-
Dihydrodiols: Stabilization of a Cyclohexadienide Anion
Intermediate by Negative Aromatic Hyperconjugation
Jaya Satyanarayana Kudavalli,† S. Nagaraja Rao,† David E. Bean,‡ Narain D. Sharma,§ Derek R. Boyd,§
Patrick W. Fowler,‡ Scott Gronert,# Shina Caroline Lynn Kamerlin,∥ James R. Keeffe,*,⊥
and Rory A. More O’Ferrall†
†School of Chemistry and Chemical Biology, University College Dublin, Belfield, Dublin 4, Ireland
‡Department of Chemistry, University of Sheffield, Sheffield S3 7HF, U.K.
§School of Chemistry and Chemical Engineering, Queen’s University of Belfast, Belfast BT9 5AG, Northern Ireland
∥Department of Cell and Molecular Biology, Uppsala University, S-751 24 Uppsala, Sweden
⊥Department of Chemistry and Biochemistry, San Francisco State University, San Francisco, California 91432, United States
#Department of Chemistry, Virginia Commonwealth University, Richmond, Virginia 23284, United States
S
* Supporting Information
ABSTRACT: Evidence that a 1,2-dihydroxycyclohexadienide anion is
stabilized by aromatic “negative hyperconjugation” is described. It
complements an earlier inference of “positive” hyperconjugative aromaticity
for the cyclohexadienyl cation. The anion is a reactive intermediate in the
dehydration of benzene cis-1,2-dihydrodiol to phenol. Rate constants for 3-
substituted benzene cis-dihydrodiols are correlated by σ− values with ρ = 3.2.
Solvent isotope effects for the reactions are kH O/kD O = 1.2−1.8. These
2
2
measurements are consistent with reaction via a carbanion intermediate or a
concerted reaction with a “carbanion-like” transition state. These and other experimental results confirm that the reaction
proceeds by a stepwise mechanism, with a change in rate-determining step from proton transfer to the loss of hydroxide ion from
the intermediate. Hydrogen isotope exchange accompanying dehydration of the parent benzene cis-1,2-dihydrodiol was not
found, and thus, the proton transfer step is subject to internal return. A rate constant of ∼1011 s−1, corresponding to rotational
relaxation of the aqueous solvent, is assigned to loss of hydroxide ion from the intermediate. The rate constant for internal return
therefore falls in the range 1011−1012 s−1. From these limiting values and the measured rate constant for hydroxide-catalyzed
dehydration, a pKa of 30.8 0.5 was determined for formation of the anion. Although loss of hydroxide ion is hugely exothermic,
a concerted reaction is not enforced by the instability of the intermediate. Stabilization by negative hyperconjugation is proposed
for 1,2-dihydroxycyclohexadienide and similar anions, and this proposal is supported by additional experimental evidence and by
computational results, including evidence for a diatropic (“aromatic”) ring current in 3,3-difluorocyclohexadienyl anion.
through a carbocation intermediate 2. The occurrence and
mechanism of the reaction are understandable by analogy with
the dehydration of alcohols and vicinal diols of nonaromatic
carbon−carbon double bonds.4 More surprising is that the
dehydration also occurs with base catalysis under the mild
conditions of aqueous sodium hydroxide at room temperature.
This reaction is shown in the lower pathway of Scheme 1.
The reaction in basic media is surprising because base-
catalyzed dehydration to form a carbon−carbon double bond
normally requires activation of the carbon β to the reacting
hydroxyl group by a substituent capable of stabilizing a
carbanion intermediate, such as a keto, cyano, or fluorenyl
group.5 It is quite unexpected that a cyclohexadienide anion
INTRODUCTION
■
The acid-catalyzed dehydration of arene dihydrodiols is now a
well-studied reaction.1−3 As shown in Scheme 1 for the example
of benzene cis-1,2-dihydrodiol (1), it is believed to proceed
Scheme 1
Received: May 5, 2012
Published: July 25, 2012
© 2012 American Chemical Society
14056
dx.doi.org/10.1021/ja304366j | J. Am. Chem. Soc. 2012, 134, 14056−14069

Journal of the American Chemical Society
Article
such as 3 would be a viable reactive intermediate under such
mildly basic conditions. In principle, the reaction could proceed
in a concerted manner, and this possibility is considered below,
but to our knowledge, there is no example of a dehydration
reaction in solution that proceeds by such a mechanism.
In this paper, we report a study of the base-catalyzed
dehydration of a series of 3-substituted benzene cis-1,2-
dihydrodiols. Measurements of hydrogen isotope effects and
isotope exchange point to a mechanism in which a carbanion
intermediate is subject to internal return. Assignment of a rate
constant for the rotational relaxation of water (1011 s−1) to the
loss of hydroxide ion from the carbanion intermediate has
allowed the derivation of a pKa value for the carbanion, leading
to an evaluation of the pKR of the phenolic product of the
reaction. It is proposed that the carbanion is stabilized by
negative hyperconjugation6−8 from the presence of a β-OH
group, and this is supported by MP2 calculations on carbanion
3, the structure of which corresponds to a Meisenheimer
intermediate of nucleophilic aromatic substitution. Calculations
also confirmed that the corresponding β-difluoro-substituted
carbanion supports a diatropic (“aromatic”) ring current.
the 3-carboxylate derivative was taken to be identical to that for
the 3-phenyl derivative (25% ortho) on the basis of their similar
σ values.
Rate constants for the dehydration reactions were measured
spectrophotometrically. Plots of first-order rate constants
against the concentration of sodium hydroxide showed
characteristic saturation with increasing base concentration, as
illustrated in Figure 1 for the 3-cyano-substituted reactant. As
RESULTS
■
Kinetic Measurements. Rate constants for the reaction in
aqueous sodium hydroxide at 25 °C were measured for 3-
substituted benzene cis-1,2-dihydrodiols 4 with X = Br, Cl, I,
CN, H, Ph, CF3, MeSO, and EtSO2 (Scheme 2). The reactants
Figure 1. Plots of rate constants (in s−1 × 104) for dehydration of 3-
○
cyanobenzene cis-1,2-dihydrodiol ( , H) and its pentadeuterated
analogue ( , D) in aqueous sodium hydroxide at 25 °C. The lines
represent the best fits of the H and D rate constants to eq 1.
●
Scheme 2
Scheme 3
were obtained as the products of oxidative biotransformations
of the parent aromatic molecules by the UV4 mutant of
Pseudomonas putida.9 As shown in Scheme 2, the reaction leads
to o- or m-phenolic products, depending on which of the
hydroxyl groups of the reactant is lost. Product ratios were
established for selected cis-diols by HPLC, GC−MS, or NMR
spectroscopy. For benzene cis-1,2-dihydrodiol itself, phenol was
the only detectable product. Phenol ratios for other substituted
cis-dihydrodiols are given in the Experimental Section. In
general, for electronegative substituents such as CN and CF3,
the sole product was the m-phenol. Only for Br (11%), Me
(13%), Ph (25%), and But (15%) was an appreciable amount of
o-phenol detected. This behavior is consistent with the
expectation that the formation of a carbanion intermediate at
the 1-position of the ring should be favored by electron-
withdrawing substituents at the 3-position. On the basis of the
measured product analyses, one can with some confidence
predict product ratios for substituents for which measurements
were not made. This can be done with sufficient accuracy to
derive rate constants for the reaction of the benzene cis-
dihydrodiols to form the major meta isomer as the product.
Thus Cl and I substituents should have product ratios similar to
that for Br, while electron-withdrawing EtSO2 and MeSO
should yield the meta product exclusively. The product ratio for
shown in Scheme 3, this behavior was attributed to the
ionization of one of the hydroxyl groups, and the measured rate
constants (kobs) were fitted to the kinetic expression given by eq
1, in which k2 is the rate constant for dehydration of the un-
ionized diol and Kb = Kw/Ka is the base dissociation constant of
the hydroxyl group.
k2[HO−]
1 + [HO−]/Kb
kobs
=
(1)
The ionization was assumed to yield a species that would be
unreactive because of the presence of a negative charge close to
the site of attack of the hydroxide ion (and formation of a
carbanion) on a hydrogen of the carbocyclic ring. The rate
constants and apparent pKa values for the two hydroxyl groups
are listed in Table 1. These are apparent pKa values for
ionization of one of the OH groups, and each is actually a
function of the microscopic pKa values of both OH groups. The
14057
dx.doi.org/10.1021/ja304366j | J. Am. Chem. Soc. 2012, 134, 14056−14069

Journal of the American Chemical Society
Article
Table 1. Second-Order Rate Constants (k2, in M−1 s−1) and
Acid Dissociation Constants of Hydroxyl Groups (pKa) for
the Dehydration of 3-Substituted Benzene cis-1,2-
ments, including first-order rate constants and wavelengths at
which the reactions were monitored, are contained in the
Supporting Information.
In attempts to extend the study of the dehydration reactions
to arene 1,2-dihydrodiols other than substituted benzene cis-
1,2-dihydrodiols, it was found that no reaction of the trans-
dihydrodiol of bromobenzene could be detected in 1 M
aqueous sodium hydroxide over a period of 3 weeks. This
implies a strong dependence of the rate of reaction on the
stereochemistry, with kcis/ktrans > 50. Likewise, no reaction was
detected for the cis-1,2-dihydrodiol of naphthalene. In an earlier
study it was concluded that benzene hydrate (2,4-cyclo-
hexadien-1-ol, i.e., benzene-1,2-dihydrodiol lacking one hydrox-
yl group) did not undergo dehydration in aqueous sodium
hydroxide, although reaction occurred readily in acidic solution.
Isotope Effects. A number of measurements of isotope
effects were undertaken utilizing biotransformation products of
deuterated aromatic substrates. Hexadeuterated benzene,
pentadeuterated chlorobenzene, and iodobenzene were com-
mercially available. Pentadeutero 3-iodobenzene cis-1,2-
dihydrodiol diacetate was converted to pentadeutero cyano-
benzene cis-dihydrodiol diacetate by reaction with tributyltin
cyanide. Deprotection then afforded the required deuterated
dihydrodiol. Rate constants and pKa values for the deuterated
1,2-dihydrodiols are also listed in Table 1. Figures 1 and 2 show
Dihydrodiols 4 in Aqueous Sodium Hydroxide at 25 °C
pKa
% m-
base/
3-
phenol in
a
b
c
d
solvent substituent
HO−/ EtSO2
k2
product
e
kinetic
equilib.
1.61 × 10−2
9.35 × 10−3
1.16 × 10−3
2.64 × 10−4
5.48 × 10−5
4.77 × 10−5
3.96 × 10−5
2.54 × 10−5
9.90 × 10−6
5.25 × 10−6
4.52 × 10−6
13.86
13.21
13.51
13.92
13.46
13.25
13.48
14.94
13.76
13.83
14.12
13.45
in 1 M
H2O
CN
100
e
MeSO
CF3
I
100
f
Br
89
f
Cl
13.9
14.0
COO−
Ph
75
H
D (4-d5)
CN (4-d5) 5.61 × 10−4
Cl (4-d5)
100
f
g
1.22 × 10−6
NaOH
g
f
I (4-d5)
1.82 × 10−6
in 1 M
NaOH
h
f
i
hj
,
DO−/ Cl
D2O
(2.8 × 10−5
1.70 × 10−2
)
(14.0)
13.5
CN
100
13.05
a
Calculated as the measured k2 value multiplied by the precentage of
b
m-phenol product. Product analyses were also carried out for Me
(89% meta) and But (87% meta), for which no kinetic measurements
c
d
were made. Based on kinetic measurements and eq 1. From initial
absorbances of kinetic measurements at different [HO−] (see Table S2
e
in the Supporting Information). Formation of 100% m-phenolic
product was assumed. Formation of 89% m-phenolic product was
assumed. The rate constant was extrapolated from a single
measurement as explained in the text. The limited amount of data
f
g
h
added to uncertainty of this measurement (see Figure S3 in the
i
Supporting Information). The data were unsuitable for obtaining a
j
measurement of pKb (see the discussion of Figure 3). Taking pKw =
14.85 for D2O.
experimental rate constants have been multiplied by the
fraction of the dominant m-phenol product formed to give
rate constants k2 specifically for this reaction path. The pKa
values are only “apparent” because normally the ionization of
two hydroxyls with different acidities is occurring. There is little
correlation between the acidity and the electron-withdrawing
character of the 3-substituent. Presumably this is partly because
of uncertainty in separating the two parameters in eq 1 (k2 and
Kb). Measurements in some cases were also based on fairly
small but consistent variations in the initial absorbances of
kinetic measurements with [HO−], as reported in Table S2 in
the Supporting Information. Of the two sets of measurements,
these probably yield the more reliable pKa values.
In the case of the carboxylate substituent, the reactant was
the corresponding methyl ester of carbomethoxybenzene cis-
1,2-dihydrodiol. This substrate showed a rapid hydroxide ion-
catalyzed decrease in absorbance at 290 nm with a rate constant
of 7.0 × 10−2 M−1 s−1, which was attributed to hydrolysis of the
ester group; this was followed by a much slower hydroxide ion-
catalyzed reaction with a rate constant of 2.5 × 10−5 M−1 s−1,
consistent with dehydration to form predominantly the m-
hydroxybenzoate anion and probably a lower concentration of
its ortho isomer. Details of this and other kinetic measure-
Figure 2. Plots of rate constants (in s−1 × 106) for dehydration of
○
benzene cis-1,2-dihydrodiol ( , H) and its hexadeuterated analogue
●
( , D) in aqueous sodium hydroxide at 25 °C. The line represents the
best fit of the H rate constants to eq 1.
plots of first-order rate constants against hydroxide ion
concentration for deuterated and undeuterated cyanobenzene
cis-1,2-dihydrodiol and unsubstituted benzene cis-1,2-dihydro-
diol, respectively.
It is apparent from Figure 1 that deuteration of
cyanobenzene cis-1,2-dihydrodiol leads to a large isotope effect.
The deuterated substrate shows the same characteristic
saturation of the rate at high base concentrations attributed
to ionization of the hydroxyl groups of the reactant as the
protio substrate. However, the rate constants are substantially
smaller than for the undeuterated substrate at all base
concentrations, with a limiting kH/kD value of 16.7 at low
concentrations of hydroxide ion. In contrast, in the case of
benzene cis-1,2-dihydrodiol, deuteration causes little change in
the measured rate constants and yields a correspondingly small
14058
dx.doi.org/10.1021/ja304366j | J. Am. Chem. Soc. 2012, 134, 14056−14069

Journal of the American Chemical Society
Article
Figure 3. Partially reacted benzene cis-1,2-dihydrodiol in NaOD/D2O at 25°. The peaks at 7.2 and 6.5 ppm correspond to the phenol product, and
the peaks near 5.9 and at 4.15 ppm represent unreacted (and unexchanged) benzene cis-dihydrodiol. The peak at 4.8 ppm is assigned to an OH
group.
isotope effect of kH/kD = 1.16. The isotope effect is >1.0, even
though at higher base concentrations the deuterated substrate
reacts more rapidly than the undeuterated substrate, with values
of kH/kD decreasing to 0.8 at high concentrations of sodium
hydroxide. This is because kH/kD is based on its limiting value
at low base concentration and the hydroxyl groups of
deuterated benzene cis-1,2-dihydrodiol are less acidic than
those of the undeuterated substrate (with KHa /KaD = 1.9). We
infer from Figure 2 that the value of kH will level off at a lower
base concentration, leading to a lower limiting value for kH than
for kD at high base concentrations.
The small isotope effect in the case of the perdeutero
benzene cis-dihydrodiol was confirmed by a similar observation
in the case of toluene cis-1,2-dihydrodiol and the corresponding
perdeuterated substrate 3-CD3C6D5 cis-1,2-dihydrodiol. These
substrates reacted too slowly for conventional kinetic measure-
ments, but repetitive scans of spectra for the reaction in 1 M
NaOH over 30 h (Table S1 and Figure S5 in the Supporting
Information) indicated the similarity of their rates. A crude
estimate of the isotope effect gave kH/kD ≈ 1.5.
In addition to these substrate isotope effects, a solvent
isotope effect of kD O/kH O = 1.2 was measured for the chloro-
2
2
substituted benzene cis-1,2-dihydrodiol. In principle, this is
consistent with a solvent isotope effect on rate-determining
proton transfer, for which kD O/kH O normally falls in the range
2
2
1.2−1.5.10 However, although the measurement was unchanged
for hydroxide (or deuteroxide) ion concentrations between 0.3
and 0.9 M, it was not possible to separate accurately the effects
on the kinetic and equilibrium steps. Thus, the greater rate in
D2O is consistent with the expectation that ionization of the
hydroxyl groups of the cis-1,2-dihydrodiol reactant occurs more
readily in H2O than D2O, and correspondingly, that the reverse
of this equilibrium (which is fully expressed at higher base
concentrations) is more favorable in D2O. This makes it likely
that the limiting rate constant for the dehydration reaction at
low base concentrations is greater in H2O than in D2O,
although the difference cannot be large (say, <1.5).
On the other hand, more extensive measurements for the 3-
cyanobenzene cis-1,2-dihydrodiol could be extrapolated to
dilute solution and yielded kD O/kH O = 1.8. At higher base
2
2
The isotopic behavior of the cis-1,2-dihydrodiols of
chlorobenzene and iodobenzene is intermediate between that
of the cyano-substituted and unsubstituted (or methyl-
substituted) benzene cis-1,2-dihydrodiols. The reactivities of
pentadeutero chloro- and iodobenzene cis-1,2-benzene dihydro-
diols are very low, and only single measurements of rate
constants based on UV measurements at 1.0 M NaOH
extending over 500 h were undertaken. Details of these
measurements are contained in the Supporting Information,
but at the indicated base concentration, the values of kH/kD
were 8.5 for the chloro-substituted diol and 6.6 for the iodo-
substituted compound. When the rate constant at 1 M was
crudely corrected to its limiting value in dilute base using eq 1
and pKa = 14.0, we obtained values of kD = 6.3 × 10−7 M−1 s−1
and kH/kD = 6.3 for the chloro compound.
concentrations, the solvent isotope effects for cyano- and
chloro-substituted substrates are similar, as may be seen by
comparing Figures S3 and S4 in the Supporting Information. As
discussed below, there is good evidence that the cyano-
substituted substrate reacts by rate-determining proton transfer
to form a carbanion intermediate, and the solvent isotope effect
is fully consistent with this. The fact that the isotope effect is
nearly independent of the concentration of hydroxide ion
suggests that the isotope effect on the equilibrium ionization of
the hydroxyl groups of the benzene cis-1,2-dihydrodiols is small,
and this is probably reasonable even though in the equilibrium
a hydroxide ion is pitted against an alkoxide ion. For the chloro-
substituted substrate, there is probably a partial change in th
rate-determining step (see below), but the similar isotope effect
is not unreasonable.
14059
dx.doi.org/10.1021/ja304366j | J. Am. Chem. Soc. 2012, 134, 14056−14069

Journal of the American Chemical Society
Article
Isotope Exchange. The reaction of undeuterated benzene
cis-1,2-dihydrodiol in D2O was also monitored by NMR
spectroscopy, and the occurrence of H for D isotope exchange
in the reactant and phenolic product was examined.
Remarkably, no exchange was detected. Figure 3 shows NMR
spectra of the reactant and product after 60% reaction. The two
low-field multiplets in the spectrum correspond to the
phenoxide ion, and the peak at highest field corresponds to
the potentially exchangeable methine hydrogen atoms adjacent
to the hydroxyl groups of the dihydrodiol. It is apparent that
there was no decrease in the 2:1 ratio of vinylic to methine
hydrogens (4.3 ppm) in the unexchanged reactant, nor was
there any sign of exchange in the product. The hydroxyl peak at
4.8 ppm decreased a little more than expected from the loss of
hydrogen from the reactant. This is readily attributable to
exchange of D2O with water from the atmosphere over the 14 h
of the reaction. If we suppose that a maximum of 5% of the
exchange remained undetected, the implication is that exchange
is at least 10 times slower than the dehydration.
Equilibrium Constant for Dehydration. Of interest in
connection with the mechanism of dehydration is the
thermodynamic driving force for the reaction. As shown in
Scheme 1, this is represented by the energy difference between
the phenolic product and the cis-1,2-dihydrodiol reactant
(augmented by ionization of the product in aqueous sodium
hydroxide). For phenol, a free energy of formation, ΔG°f (aq),
may be obtained from the convenient compilation of free
energies of formation in aqueous solution at 25 °C by
Guthrie.11 For the cis-1,2-dihydrodiol, ΔGf°(aq) may be
estimated by analogy with that for benzene hydrate (2,4-
cyclohexadien-1-ol) . As shown in Scheme 4, ΔG°f (aq) for this
Scheme 5
substituents are at the 6-position of the cyclohexadienide
anion or the 7-position of the cycloheptadienide anion and
protonation yields the corresponding neutral X-substituted
cyclohexadiene or cycloheptadiene. Also shown are ΔPA
values, which are differences in proton affinities with respect
to the unsubstituted cyclohexadienide anion, C6H7 . The
−
effectiveness of the G3 method was demonstrated by showing
that it gives proton affinities that correlate well with
experimental values for 18 anionic bases: PAG3 = 1.10PAexptl
+ 5.2 (r2 = 0.999).13 This correlation is shown in Figure S8 in
the Supporting Information on the basis of data summarized in
Table S3 in the Supporting Information. More details of
calculated energies, zero-point energies, and geometries of the
cyclohexadienide and cycloheptadienide ions and their isomeric
complexes and transition states are shown in Tables S4
(G3MP2) and S5 in the Supporting Information. Included in
Table S4 (entry 8a) is a 1,6-dihydroxycyclohexadienide anion
hydrogen-bonded to a water molecule. This is the immediate
(stable) product of proton transfer from the corresponding 1,6-
dihydroxycyclohexadiene to hydroxide ion in the gas phase.
Enthalpies and free energies (G3MP2) of the conjugate acids of
the calculated anions are given in Table S8 in the Supporting
Information.
Scheme 4
The X substituents correspond to leaving groups in an
elimination reaction in which the cyclohexadienide anion is a
reactive intermediate. The products of this elimination are
shown in Table 2. The anions also correspond to Meisenheimer
intermediates in nucleophilic substitution reactions, for which
the reverse of the loss of leaving group represents the initial
attack of a nucleophile on the aromatic ring. Table 2 lists
energies and activation energies for these reactions. For the
molecule was obtained from that for 1,3-cyclohexadiene by
adding an increment representing the energy of replacing a
hydrogen atom by a hydroxyl group.12 In principle, the
conversion of cyclohexadienol to benzene-1,2-dihydrodiol 1
(Scheme 4) should involve the same increment adjusted for any
repulsion between the vicinal hydroxyl groups. This repulsion
may be estimated from ΔG° for the bond separation
equilibrium between ethylene glycol and ethanol (eq 2).8
−
reaction of C6H5X with X−, the intermediate C6H5X2 is the
product of reaction at the ipso carbon of the aromatic ring.
G3MP2 enthalpies and free energies of the nucleophiles and
substrates in the nucleophilic substitutions are given in Table
S3 in the Supporting Information.
When X is bound to the aromatic ring by a second-row
element, as with F, OH, NH2, and CH3 (and CN or CCH),
the cyclohexadienide anion exists as an energy minimum.
Indeed, this Meisenheimer complex is more stable than the
separated nucleophile (X−) and aromatic molecule. However,
for third-row elements, only SiR3 gives a stable monosubstituted
anion with benzene, and the diphosphino anion is also stable.
For the other third-row elements (SR and Cl), there is no
energy minimum for mono- or disubstituted ions, and it
appears that these are too reactive as leaving groups to maintain
attachment to the benzene ring. The same was shown to be
true for X = Br or NO2. Previous work indicated that for these
groups, the corresponding nucleophilic substitution reaction
proceeds by a concerted mechanism without the formation of a
Meisenheimer intermediate.14, As also shown previously, the
The derived value of ΔG°(aq) = −33.2 kcal mol−1 for benzene
cis-dihydrodiol may be combined with the free energies of
formation for water and phenol to give ΔG°(aq) = −36.1 kcal
mol−1 for the dehydration reaction in Scheme 5. This translates
to an equilibrium constant of log K = 36.1/1.364 = 26.5.
Computed Energies and Geometries of Cyclohexa-
dienide and Cycloheptadienide Anions. Energies of
substituted cyclohexadienide and cycloheptadienide anions in
the gas phase were obtained from G3MP2 and MP2
calculations. In Table 2, these are expressed as proton affinities
(PAs) of the mono- and disubstituted anions C6H6X− (5) and
−
−
C6H5X2 (6) or C7H8X− and C7H7X2 , where the X
14060
dx.doi.org/10.1021/ja304366j | J. Am. Chem. Soc. 2012, 134, 14056−14069

Journal of the American Chemical Society
Article
Table 2. G3MP2-Calculated Proton Affinities of 6-Substituted Cyclohexadienide Anions C6H5XY−, Enthalpies ΔHadd and
a
Activation Enthalpies ΔH⧧add for Addition Reactions of the Elimination Products (C6H5X + Y− → C6H5XY−), and Calculated
Proton Affinities of Selected 7-Substituted Cyclohexadienide Anions
b
c
d
entry
X, Y (point group)
PA (kcal/mol)
ΔPA (kcal/mol)
elimination products
ΔHadd (kcal/mol)
ΔH⧧add (kcal/mol)
Cyclohexadienide Anions
e
1
H, H (Cs)
374.3
0
benzene + H−
−17.0
−28.3
−30.5
4.3
+6.6
−1.7
−3.7
−
−3.7
−6.8
−6.6
−
2
CH3, H (Cs)
CH3, CH3 (Cs)
371.5
370.2
369.7
364.2
361.0
357.75
364.0
356.9
350.9
345.8
369.2
363.8
371.8
−
2.8
benzene + CH3
−
3
4.1
toluene + CH3
f
−
4
−CH2−CH2− (Cs)
NH2, H (Cs)
4.6
C6H5−CH2−CH2
−
5
10.1
13.3
16.5
10.3
17.4
23.4
28.5
5.1
benzene + NH2
−15.2
−16.6
−6.9
−0.6
−15.1
+3.7
−
6
NH2, NH2 (C1)
aniline + NH2
g
7
OHax, H (Cs)
benzene + OH−
benzene + OH−
phenol + OH−
benzene + F−
g
8
OHeq, H (Cs)
9
OH, OH (C1)
−12.9
−3.7
−4.9
+12.1
+4.2
−
h
10
11
12
13
14
15
16
17
18
F, H (Cs)
F, F (C2v)
fluorobenzene + F−
−10.7
+5.8
−
SiH3, H (Cs)
SiH3, SiH3 (Cs)
(CH3)3Si, H (Cs)
benzene + SiH3
−
9.5
silylbenzene + SiH3
−0.9
−7.2
−5.1
−6.2
+3.1
2.5
benzene + (CH3)3Si−
i
−
PH2, H (Cs)
−
14.9
benzene + PH2
−
j
−
PH2, PH2 (Cs)
359.5
368.8
359.5
phosphinobenzene + PH2
benzene + HCC−
benzene + CN−
+5.6
+13.1
+23.5
HCC, H (Cs)
5.5
CN, H (Cs)
14.8
+18.4
Cycloheptadienide Anions
19
20
21
22
23
H, H (C2)
F, H (C1)
371.6
362.4
355.8
360.5
360.3
2.7
11.9
18.5
13.7
14.0
F, F (C1)
OH, H (C1)
OH, OH (C1)
a
b
c
Initial step of nucleophilic aromatic substitution. The pseudoaxial group is indicated in bold type. Geometries are outlined in the main text. PA =
d
−ΔHrxn for protonation of the anion to give the corresponding 1,3-cyclohexadiene. ΔPA = PAHH −PAXY, where PAHH is the PA for
cyclohexadienide anion, C6H7 . Thus, a positive ΔPA value indicates that the anion is less basic (more stable) than C6H7− relative to their respective
−
e
f
g
−
protonation products. The experimental PA for C6H7 is 373.3 4.1 kcal/mol. Spiro[2.5]octa-4,6-dienide−. The OH bond is endo to the ring.
The exo form is a transition state [iν = 306 cm−1 (C−O torsion)] lying 4.4 kcal/mol above the endo form. In entry 8, the OH group was made
h
equatorial by constraining dihedral angles to ensure Cs symmetry. A covalent addition complex with a bond-length extension of 0.193 Å relative to
its conjugate acid was found at the G3MP2 level. However, at the MP2/6-311+G** level, this structure optimized to a H-bonded benzene−fluoride−
i
j
complex. Optimization at the G3MP2 or MP2/6-311+G** level gave a H-bonded benzene−PH2− complex. At the G3MP2 level with MP2/6-31G*
frequencies, a stable covalent adduct was formed. At the MP2/6-311+G** level, the structure became a transition state (iν = 117 cm−1).
stability of the intermediate is increased by inclusion of one or
more nitro or other electron-withdrawing groups in the
aromatic ring.16
carbon is 15−18° out of plane. All of the X substituents occupy
pseudoaxial positions with C−X bonds that are longer than the
axial C−X bonds in their neutral protonation products. A
pseudoaxial−pseudoequatorial disposition of the X substituents
is maintained for the disubstituted ions, except for the difluoro
and diamino ions, both of which have C2v heavy-atom
frameworks. Further geometric details are contained in Tables
S4 and S5 in the Supporting Information.
It should be mentioned that for the fluoride anion, the most
reactive of the second-row leaving groups, although a stable
Meisenheimer complex with a highly extended fluoride bond
was found at the G3 level, it was not found at the MP2/6-
311+G** level. Rather, a hydrogen-bonded benzene−fluoride
complex is formed. For the difluoro anion, on the other hand,
the complex is stable, probably as a result of the stabilizing
influence of geminal fluorine atoms. In this case also, the
enthalpy of activation for the identity aromatic substitution of
fluoride by fluoride (−4.9 kcal mol−1) is in satisfactory
agreement with that found by Glukhotsev et al. [−2.2 and
−3.7 kcal mol−1 by MP2 and B3LYP respectively, both at
MP2/6-31+G(d)].15
A previously calculated geometry of the cyclohexadienide
anion gave a planar structure.17 Our calculations for the
monosubstituted ions correspond to shallow boat conforma-
tions in which carbons 1, 2, 4, and 5 define a plane and the X
substituent is at C6, leading to overall Cs symmetry. The carbon
at position 6 has an out-of-plane tilt that varies from 45° for
groups with low electronegativity (e.g., H and CH3) to smaller
values for electronegative groups (e.g., 6° for OH). The 3-
DISCUSSION
■
The dehydration of benzene cis-1,2-dihydrodiol shows a first-
order dependence on the hydroxide ion concentration, albeit
modified by partial ionization of a hydroxyl group as shown in
Scheme 3. This observation, coupled with a positive ρ value of
3.2 and the correlation of the effect of 3-substituents by the σ−
values (Figure 4), is clearly consistent with reaction that occurs
via a carbanion intermediate. However, these criteria do not
rule out a concerted mechanism involving a carbanion-like
transition state.6,18 Indeed, the following facts and observations
appear to favor a concerted mechanism.
The very high exothermicity of the reaction suggests that the
stability of the double bond of the product, which reflects its
aromatic character, should strongly stabilize the partially
formed double bond in the transition state of a concerted
14061
dx.doi.org/10.1021/ja304366j | J. Am. Chem. Soc. 2012, 134, 14056−14069

Journal of the American Chemical Society
Article
unsubstituted cis-1,2-benzene dihydrodiol, within an overall
E1cB mechanism for the dehydration reaction.
An obvious test for a mechanistic change is to look for rapid
exchange of hydrogen for deuterium isotopes in the pre-
equilibrium proton transfer step in the case of benzene cis-1,2-
dihydrodiol. As shown in Scheme 7, reaction of the protio
Scheme 7
benzene cis-1,2-dihydrodiol with NaOD in D2O should lead to
exchange of a hydrogen for deuterium in two C−H bonds in
the dihydrodiol reactant and one in the phenolic product.
Remarkably, when the reaction was monitored by NMR
spectroscopy, no exchange was observed in the reactant or
product over a period that implied that the rate of exchange was
at least 10 times slower than the overall dehydration reaction.
At first sight, this appears to exclude the possibility of a pre-
equilibrium proton transfer. However, a number of examples of
related behavior are known for carbanion reactions.19,20 They
have been interpreted as implying the occurrence of “internal
return” of the initial proton transfer reaction, i.e., the formation
of a carbanion that is hydrogen-bonded to the hydrogen atom
to which it was originally attached. In the present instance, the
carbanion would be bound to a water molecule formed by
proton transfer between the dihydrodiol and hydroxide ion
reactants. Provided that further reaction of the carbanion (to
form phenol) occurs within this reversibly formed ion−
molecule hydrogen-bonded pair and that dissociation of the
pair leading to isotope exchange is slower than product
formation, the observed behavior becomes explicable.
A schematic representation of the dehydration reaction
embodying these mechanistic features is shown in Scheme 8, in
which the benzene cis-1,2-dihydrodiol is denoted as H−Ar−
OH or H−Ar−OD (in D2O) and phenol as Ar. The H−Ar
bond of the reactant is the carbon−hydrogen bond that
undergoes ionization to form the carbanion intermediate.
Perhaps the simplest interpretation of the slow isotope
exchange reaction is that it involves dissociation of the
hydrogen-bonded carbanion to give an unsolvated species of
significantly higher energy.20 This feature is incorporated into
Scheme 8, in which the relative rates of the processes
controlling dehydration, exchange, and internal return are
indicated. The reaction is shown as occurring in D2O in order
to include the exchange feature. Richard has suggested
rearrangement of the hydrogen-bonded water molecule as an
alternative to dissociation,21 but as noted below, this fails to
account for the slowness of the reaction.
Figure 4. Plot of log k (M−1 s−1) vs σ− for the dehydration of 3-
substituted benzene cis-1,2-dihydrodiols in aqueous NaOH at 25 °C:
log k = −5.16 + 3.21σ−.
reaction. The much greater reactivity of the cis-1,2-dihydrodiol
of benzene relative to that of naphthalene is consistent with this
possibility. Also consistent with a concerted mechanism is the
much greater reactivity of a cis-dihydrodiol compared with a
trans-dihydrodiol. In the case of the cis-dihydrodiol, the OH
leaving group and the β-carbon−hydrogen bond can achieve an
antiperiplanar arrangement, which is normally considered
optimum for a concerted reaction. This is not possible for
the trans-dihydrodiol.
However, the measurements of hydrogen isotope effects for
deuterated reactants appear to be decisive in indicating reaction
via a carbanion intermediate. For pentadeutero cyanobenzene
cis-dihydrodiol (7), the measured value of kH/kD is 16.7, which
must represent a large primary isotope effect. This isotope
effect is still consistent with operation of either a concerted
mechanism or a stepwise (E1cB) mechanism in which the step
forming the carbanion intermediate is rate-determining, as
shown in Scheme 6. However, when the substituent is changed
Scheme 6
from cyano to chloro, the primary isotope effect is reduced to
∼6.3, and when the substituent is changed to deuterium (i.e.,
for the cis-1,2-dihydrodiol of hexadeuterobenzene), the isotope
effect disappears altogether. For benzene cis-1,2-dihydrodiol,
kH/kD = 1.2, which is a normal magnitude for a secondary
isotope effect.
Such a large change in the magnitude of the isotope effect for
a relatively small change in reactivity between cyano, chloro,
and deutero substituents would be unprecedented for a proton
transfer reaction not subject to a change in mechanism or rate-
determining step. What the behavior suggests is that there is a
change from rate-determining proton transfer for the cyano
substituent to pre-equilibrium proton transfer for the
Scheme 8 is complicated in that it includes the initial low-
energy step representing diffusion together and resolvation of
the reactants to form the ion−molecule pair within which
proton transfer occurs. The reaction is more clearly represented
by the free energy diagram in Figure 5, in which the initial
diffusion step has been omitted, and further simplified in
Scheme 9, in which only the preequilibium and rate-
determining steps for the dehydration reaction are shown. As
usual, k1 in Scheme 9 represents the composite of an
equilibrium constant for formation of the encounter complex
between the substrate and hydroxide ion (i.e., in the initial
14062
dx.doi.org/10.1021/ja304366j | J. Am. Chem. Soc. 2012, 134, 14056−14069

Journal of the American Chemical Society
Article
Scheme 8
its relationship to k1 and k−1 implied by Scheme 9 and shown in
eq 3:
k1k2
k−1 + k2
kOH
=
(3)
It follows that k1 = kOH(1 + k−1/k2), and with k2 = 1011, this
expression gives k1 = 1.1 × 10−5 M−1 s−1 for k−1 = 1012 s−1 and
k1 = 2.0 × 10−6 M−1 s−1 for k−1 = 1011 s−1. The corresponding
values of Kb are 1.1 × 10−17 and 2.0 × 10−17. From the average
of these values, pKb for the 1,2-dihydroxycyclohexadienide
anion becomes 16.8, which yields pKa = 14.0 + pKb ≈ 31 with
an estimated uncertainty of 0.5. We consider it possible that
k2 is somewhat less than 10−11, as HOD hydrogen bonded to
the carbanion could impose a barrier to reorganization. If so, kex
would be even smaller.
Figure 5. Schematic plot of free energy against reaction coordinate for
base-catalyzed dehydration and isotope exchange of benzene cis-1,2-
dihydrodiol. The deuteron transfer step completing the exchange
reaction has been omitted.
A more detailed argument regarding the limiting rate
constant values focuses on the possible relative magnitudes of
k−1 and k2 in Scheme 9 and kr, which for purposes of this
analysis we will call the rate constant for solvent relaxation (or
solvent reorganization) within the hydrogen-bonded
DOH···−Ar−OD complex (Scheme 8 and Figure 5). These
are the three rate constants leading from the complex. The
absence of exchange implies that kr is less than k2 and therefore
that k−1 > k2 > kr. All three rate constants must be very large.
We can also state that the observed rate constant kOH is given
by kOH = k1k2/(k−1 + k2 + kr) ≈ k1k2/(k−1 + k2) and that the
rate constant for the unobserved exchange process, kex, is given
by kex = k1kr/(k−1 + k2 + kr) ≈ k1kr/(k−1 + k2). We can conclude
from experiment that kex/kOH < 0.1 and that kex/kOH = [k1kr/
(k−1 + k2)]/[k1k2/(k−1 + k2)] = kr/k2. The hydrogen bond in
DOH···−Ar−OD, even though it is between C and O, will make
solvent reorganization slower than in the solvent itself, by
perhaps as much as 2 orders of magnitude; thus, 109 s−1 < kr <
1011 s−1. This implies that 1010 s−1 < k2 < 1012 s−1. We argue
above that k−1 lies between 1011 and 1012 s−1. Thus, the
constraint that k−1 > k2 > kr leaves very little freedom. Assuming
that “greater than” means a factor of 10 or more, the set of rate
constants is then 1011 < k−1 < 1012, 1010 < k2 < 1011, and 109 <
kr <1010 (all in s−1). This is sufficient to put narrow limits on k1
and thus on the proposed pKa for benzenediol. We thank a
reviewer for this analysis.
Thermodynamic Driving Force. The thermodynamic
driving force (exothermicity) required to ensure that the
chemical barrier to reaction is less than the limiting value of the
barrier for solvent relaxation depends on the intrinsic barrier to
the reaction. Without entering into a full discussion of the point
here, we note that for proton transfer to a hydroxide ion from
carbon activated by a series of carbonyl, carbethoxy, amido, and
carboxylate groups, Richard has shown that the intrinsic barrier
is 10 kcal mol−1 and that the reaction becomes diffusion-
controlled as the pKa of the carbon acid approaches a value of
30.26 Our assignment of a limiting rate constant to the reaction
between water and a 1,2-dihydroxycyclohexadienyl anion for
Scheme 9
diffusion step) and a rate constant for the reaction of this
complex.
A clue to the interpretation of the observed behavior is
provided by the large exothermicity of the dehydration reaction.
This must apply a fortiori to the reaction step leading to
formation of the phenolic product from a high-energy
carbanion intermediate such as 8. As discussed in more detail
below, this must in turn imply that loss of hydroxide ion from
the carbanion occurs at a rate limited only by relaxation of
water to and from the solvation shell of the ion, with k2 ≈ 1011
s−1.22−24 If the internal return of the proton to reform the
benzene cis-1,2-dihydrodiol is faster than this step [i.e., loss of
the hydroxide ion to form the product (phenol)], then the
dehydration reaction would indeed occur without a significant
primary isotope effect.
An upper limit to proton transfer within a hydrogen bond is
probably given by the rate constant of 1012 s−1 for proton
transfer between H3O+ and H2O in ice.25 It is reasonable that
k−1 should be as large as 1011−1012 s−1 if the thermodynamic
driving force for the protonation of the carbanion is sufficiently
great. The size of this driving force can be determined by
deriving the equilibrium constant Kb for protonation of the
carbanion intermediate. This is given by k1/k−1, the ratio of rate
constants for formation of the carbanion from the benzene cis-
1,2-dihydrodiol and hydroxide ion reactants and for reversion
to these reactants. A value of k1 may be obtained from the
experimentally measured rate constant for the hydroxide-
catalyzed dehydration reaction (kOH = 1.0 × 10−6 M−1s−1) and
14063
dx.doi.org/10.1021/ja304366j | J. Am. Chem. Soc. 2012, 134, 14056−14069

Journal of the American Chemical Society
Article
which the pKa of its conjugate acid is 30.8 is therefore
reasonable.
A difference between the study by Richard and Williams21
and the present work is that they used hydrogen isotope
exchange of nitriles to measure their rates of ionization to form
anions. When reprotonation of the anion is sufficiently
exothermic, the isotope exchange ceases to be catalyzed by
buffer base, and Richard supposed that the rate-determining
step for the exchange is rearrangement of a HDO solvent
molecule, with a rate constant kreorg. This is shown in Scheme
10, in which AH denotes a reactant and A− its conjugate-base
anion.
+ H+) is a favorable reaction; for example, pKR = −16 for the
tert-butyl cation.31 Hydrolysis becomes unfavorable and pKR
positive for weaker electrophiles such as aldehydes or ketones.
In the case of benzaldehyde, for which the resonance form
PhCH+−O− may be considered to represent a highly stabilized
carbocation, pKR = 14.7. The much larger positive value for
phenol reflects both the stability of the double bond of the
weakly electrophilic reactant and the instability of the
carbanionic adduct. These are also reflected in the large
exothermicity associated with expulsion of hydroxide ion from
the adduct.
Scheme 10
The two equilibrium constants KR and Ka may be related to
each other through the equilibrium constant KH O for addition
2
of water to phenol, which corresponds to the reciprocal of the
equilibrium constant for the overall dehydration reaction
calculated above. When expressed in their logarithmic forms,
the three equilibrium constants can be related through a Hess-
like thermodynamic cycle, as shown in Scheme 11. Single rather
than double arrows are used to represent the equilibria in order
to indicate the direction of the reaction to which each
equilibrium constant refers.
Richard assumed a kreorg value of ∼1011 s−1, which
corresponds to rotational relaxation of the solvent.16−19 In
our case, however, a rate constant corresponding to rotational
relaxation is assigned to the loss of hydroxide ion from a
dihydroxycyclohexadienide anion intermediate. Loss of hydrox-
ide occurs at least 2 orders of magnitude more rapidly than the
isotope exchange reaction. This implies that the isotope
exchange cannot be controlled by solvent relaxation.
Similarly, Koch showed that loss of a leaving group, typically
F− or Cl−, from a carbon β to a carbanion center can occur
more rapidly than hydrogen isotope exchange in deuterated
methanol as the solvent.27 Formation of a desolvated higher-
energy carbanion intermediate with a rate constant of <1011 s−1
seems to offer the simplest explanation of the behavior that he
observed and we have observed here.
Scheme 11
Is the thermodynamic driving force for the expulsion of
hydroxide ion from the dihydroxycyclohexadienide anion
sufficient to ensure that this occurs at the limiting rate of
solvent relaxation? Bernasconi showed that the intrinsic barrier
to loss of a β-hydroxide ion to form a double bond is
significantly greater than that for protonation of the same
carbanion.28 The magnitude of the exothermicity (strictly, the
reaction free energy) can again be assessed from the
equilibrium constant for the reaction. It was shown above
that log K for the overall reaction of benzene cis-1,2-dihydrodiol
to form phenol is 26.5. To arrive at the value of log K for the
conversion of the carbanion intermediate to product, we must
add to this the pKb value of 16.8; this gives an overall value of
log K = 43.3; which corresponds to ΔG = −59 kcal mol−1 at 25
°C. This is an overwhelmingly large thermodynamic driving
force, which leaves little doubt that the rate of expulsion of
hydroxide ion must indeed occur at the limit of solvent
relaxation.
Isotope Effects. The appearance of isotope effects on the
dehydration reaction when the 3-chloro or 3-cyano substituents
are present in the benzene cis-1,2-dihydrodiol reactant is readily
understandable. The electronegative substituents stabilize the
carbanion and increase the barrier for reprotonation. This
renders the proton transfer rate-determining, insofar as the
much greater exothermicity of the loss of hydroxide ion ensures
that the rate constant for this reaction step remains close to its
limiting value of 1011 s−1. This is illustrated schematically in
Figure 6, in which the upper free energy profile corresponds to
Cycle of Equilibrium Constants. The value of log K
corresponds to log Kc, the logarithm of the equilibrium
constant for addition of hydroxide ion to an ortho position of
phenol. Values of log Kc for the addition of hydroxide ion to
various electrophiles have been discussed by Hine.29 They may
be converted to values of pKR by recognizing that KR and Kc are
related in a manner comparable to that KR/Kc = Kw, where Kw
= 10−14 is the autoprotolysis constant of water.30 The value of
pKR for phenol, which refers to the equilibrium shown in eq 4,
is 43.3 + 14 = 57.3.
Figure 6. Schematic free energy profiles for hydroxide ion-catalyzed
dehydration of unsubstituted (upper profile, from Figure 4) and 3-
cyano-substituted (lower profile) benzene cis-1,2-dihydrodiol.
Normally, pKR refers to cabocations and has a negative value,
implying that hydrolysis of the carbocation (R+ + H2O → ROH
14064
dx.doi.org/10.1021/ja304366j | J. Am. Chem. Soc. 2012, 134, 14056−14069

Journal of the American Chemical Society
Article
the unsubstituted benzene cis-1,2-dihydrodiol and the lower
(dashed) profile to its 3-cyano-substituted derivative. The sharp
change in isotope effect in going from the chloro (kH/kD ≈ 6.3)
to the cyano compound (kH/kD = 16.7) suggests that the
isotope effect may be reduced in the case of the chloro
substituent because the proton transfer step is only partially
rate-determining.
pseudoequatorial β-C−H bond, a geometry that is much less
favorable for hyperconjugation. The large magnitude of the
effect was attributed to the aromatic character of the no-bond
resonance structure associated with the hyperconjugation, as
illustrated in structure 13; it is further magnified by good
electron-donating groups (e.g., silyl) at the β-carbon.
A similar dependence of the isotope effect on the substituent
was observed by Eaborn in the cleavage of benzyltrimethyl-
silanes 9 in methanolic sodium methoxide (Scheme 12).32 The
Scheme 12
It is natural to hypothesize that the difference in reactivities
in the base-catalyzed reaction is also due to hyperconjugation.
In this case, however, stabilization of a cyclohexadienide anion
would be by negative hyperconjugation, with a pseudoaxial β-
C−OH bond being more favorable than a C−H bond. The
preferential formation of carbanions with C−OH and C−H
bonds in β-pseudoaxial positions from the reactions of cis and
trans reactants, respectively, is illustrated in Scheme 13
(hydrogen bonding between the newly formed water and the
carbanion has been omitted). An important assumption is that
the proton transferred from the reactant benzene dihydrodiol
departs from a pseudoaxial position. This should optimize the
overlap between the developing negative charge and the π
orbitals of the cyclohexadienyl ring.
Calculated Stabilities of Cyclohexadienide Anions.
Support for the concept of negative hyperaromaticity was
provided by MP2 calculations on substituted cyclohexadienide
anions in the gas phase. As noted above, these ions have the
structure of Meisenheimer intermediates in nucleophilic
aromatic substitution reactions. Thus the 2-hydroxycyclohexa-
dienide anion 3 corresponds to a Meisenheimer complex
formed by nucleophilic attack of a hydroxide ion on benzene.
Previous computational studies of Meisenheimer complexes
have focused on their formation from different combinations of
the aromatic substrate and the nucleophile in the gas phase.14,15
They have emphasized ipso reactions, especially of halide ions
with the corresponding halobenzenes. Depending on the
aromatic molecule and nucleophile, the Meisenheimer structure
may correspond to a stable intermediate or a transition state.
A difficulty with these calculations is that the energy of the
intermediate is computed relative to the strongly variable
energies of the nucleophile and aromatic molecule. In this
study, we assessed the stabilization of cyclohexadienide ions
(14) through comparisons of gas-phase proton affinities for X-
and X2-substituted ions, where the X substituents are located at
the 6-position with respect to the carbanionic charge. The
approach is similar to that used to investigate the stabilization
of cyclohexadienyl cations through comparisons of hydride ion
affinities.2,3 As shown for these ions, it is particularly effective
for investigating stabilization by hyperconjugation. Thus, it
seems reasonable to suppose that geminal (anomeric)
interactions between electronegative atoms, which can be
expected to stabilize X2-substituted cyclohexadienide anions,
will be largely canceled with the cyclohexadiene reactant.
isotope effect was measured as the ratio of H to D in the
toluene product, [R−H]/[R−D], following the reaction of the
benzylic anion intermediate in a mixed MeOH/MeOD solvent.
Table 3 shows values of [R−H]/[R−D] for various
substituents along with values of log krel, where krel is the
relative rate constant for generating the anion from the
benzyltrimethylsilane precursor.
One question that remains is why does the reaction proceed
so readily for so apparently unstable an intermediate as a
cyclohexadienide anion? The following may provide at least a
partial answer.
Negative Hyperaromatic Stabilization of the
Hydroxycyclohexadienide Anion. A number of clues
suggest that the cis-1,2-dihydroxycyclohexadienide anion may
be unusually stable. We have seen that no reaction is observed
when the benzene cis-1,2-dihydrodiol reactant is replaced by
benzene hydrate [2,4-cyclohexadien-1-ol (10), i.e., the dihy-
drodiol lacking a 2-hydroxyl group], naphthalene-cis-1,2-
dihydrodiol (11), or benzene trans-1,2-dihydrodiol (12). The
structures of these potential reactants are compared below.
The more than 50-fold greater reactivity of the benzene cis-
dihydrodiol relative to the trans isomer in particular recalls the
behavior observed for the acid-catalyzed dehydration, a reaction
for which the reactivity difference amounted to kcis/ktrans
=
4500.2 This difference was attributed to hyperconjugative
stabilization of a carbocation intermediate in the case that the
carbocation was formed with a pseudoaxial C−H bond β to the
center of charge and the β-C−OH bond in a pseudoequatorial
position. It was argued that although this occurred for the
reaction of the cis-dihydrodiol, the carbocation of the trans
isomer was formed with a pseudoaxial β-C−OH bond and a
Table 3. Isotope Effects for Protonation of Substituted Benzylic Anions by Methanol
substituent
H
m-CN
m-NO2
p-CN
p-SO2Ph
p-COPh
o-NO2
p-NO2
[R−H]/[R−D]
log krel
1.2
0
1.2
1.3
2.0
2.9
7.0
10.0
6.67
10.0
7.36
3.45
3.72
5.80
5.97
6.11
14065
dx.doi.org/10.1021/ja304366j | J. Am. Chem. Soc. 2012, 134, 14056−14069

Journal of the American Chemical Society
Article
Scheme 13
a
Table 4. Proton Affinities (PA) and Relative Proton Affinities (ΔPA) of 6-Monosubstituted Cyclohexadienide Ions 11
substituent
H
CH3
NH2
OH
F
SiH3
CCH
CN
PA
374.3
0
371.5
2.8
364.2
10.1
357.8
16.5
350.9
23.4
369.2
5.1
368.8
5.5
359.5
14.8
ΔPA
a
Abstracted from Table 2 in the Results; all values are in kcal mol−1.
Calculated PAs of monosubstituted cyclohexadienide ions
computed proton affinities of the p-phenyl anions33 against the
proton affinities of the correspondingly substituted stable
cyclohexadienide anions is given in Figure 7. Groups that are
and ΔPA values relative to the unsubstituted cyclohexadienide
ion are summarized in Table 4. Additional substituents and
energies of nucleophilic substitution reactions are given in
Table 2 in the Results. A further difficulty with the
computations is that the reaction of the anion to form the
product aromatic molecule and nucleophile, i.e., the second
step of an elimination reaction of a substituted cyclohexadiene
such as benzene cis-1,2-dihydrodiol (or of a two-step
nucleophilic substitution), has a very low energy barrier, and
the structures of anions for which X is a good leaving group
correspond to transition states for this reaction rather than true
energy minima. Nevertheless, the electronegative substituents
F, HO, and NH2 all give minima, as do less electronegative
groups such as SiH3 and CH3. All substituents are stabilizing,
but the effect is much larger for electronegative groups than for
CH3 or SiH3.
For the stable cyclohexadienide anions, the X substituents
consistently occupy pseudoaxial positions (14), and there
appear to be no minima for the corresponding isomers with X
in a pseudoequatorial position. If a conformation with X
Figure 7. Plot of computed cyclohexadienide PAs vs p-phenyl anion
PAs.
not expected to have a strong contribution from hyper-
conjugation give a reasonable correlation line with a slope of
0.8. These include H, CH3, SiH3, CCH, and CN. The groups
where hyperconjugation is expected (F, OH, NH2) fall below
this line and may be considered to provide extra stabilization of
the cyclohexadienide relative to the phenyl anion. The
magnitude of the implied stabilization is large, amounting to
19 kcal/mol for F, 15 kcal/mol for OH, and 11 kcal/mol for
NH2. The relative magnitudes of the stabilizations fall in the
order expected for negative hyperconjugation.
These comparisons do not demonstrate that the hyper-
conjugative effect confers aromaticity. Evidence regarding this
point comes from a comparison between para- and ortho-
substituted phenyl anions. As illustrated by 15o and 16,
hyperconjugative stabilization but not aromaticity can be
envisaged for the ortho anion. Moving a cyano group from a
para to an ortho position reduces the proton affinity of the ion
by only ∼2 kcal/mol. In contrast, moving a fluorine atom
reduces the affinity by over 7 kcal/mol.32 This suggests that
there is some contribution from hyperconjugative stabilization.
However, the effect is much smaller than the stabilization
estimated above for the 6-fluorocyclohexadienide anion (14, X
= F). Hyperconjugative interactions in 16 may be limited to
equatorial is constrained, the difference in energy between this
and the equilibrium geometry is greatest for electronegative
substituents, with an energy difference of 6.3 kcal/mol for an
OH group when the dihedral angle formed by the axial
hydrogen, C6, C1, and the H atom at C1 is 80°. This difference
compares with 8.8 kcal mol−1 for the corresponding 6-
hydroxycyclohexadienyl cation in which C−OH rather than
C−H is constrained to be pseudoaxial.2 This result provides the
first computational argument for the importance of hyper-
conjugation. The stabilizing effect of electronegative substitu-
ents is apparent in the relative proton affinities (ΔPA) of the
cyclohexadienide anions in Tables 2 and 4. The geometrical
dependence of this stabilization confirms that it is not solely an
inductive effect.
Another approach for assessing the stabilization derived from
hyperconjugation is to compare the proton affinities of 6-
substituted cyclohexadienide ions with those of carbanions that
cannot profit from hyperconjugation. For this comparison, we
chose para-substituted phenyl anions 15p. A plot of the
14066
dx.doi.org/10.1021/ja304366j | J. Am. Chem. Soc. 2012, 134, 14056−14069

Journal of the American Chemical Society
Article
some extent by the implied contribution from a high-energy
benzyne resonance structure, but the large difference (19 vs 7
kcal mol−1) suggests that the cyclohexadienide anions
experience added stabilization of up to 12 kcal mol−1 due to
hyperaromaticity in 14.
A further way of estimating the aromatic contribution to the
hyperconjugation is to compare the proton affinities of the F-
and OH-substituted cyclohexadienide ions in Table 4 and
Table 2 with those of the corresponding unsubstituted and
substituted cycloheptadienide anions 17−19. Values of ΔPA
conventional π system of the ring and any contributions that
may arise from the antisymmetric combinations of CF bonds
and of the F “lone pairs” of the CF2 group. The main
observation is that there is a clear indication of a global π ring
current passing above and below all of the carbon centers, with
the same sense of circulation as in benzene itself. This is fully
consistent with an attribution of aromatic character deriving
from hyperconjugative participation of the fluorine atoms.
Incidentally, the map also shows a pronounced but tightly
localized diamagnetic circulation associated with the hyper-
conjugating CF2 group, which is a reflection of the presence of
a closed-octet valence shell on the fluorine atom and is not
connected with the global aromaticity of the ring. There is no
equivalent in the benzenium ion 21, where of course the H
centers do not carry these extra valence electrons. Estimates of
the intensities of the ring currents in these systems were
available from the maximum magnitudes of the current density
per unit inducing field taken over the 1a0 plotting plane. For
benzene itself, in a calculation of the present type, the value of
this quantity is 0.079 au, and in 20 and 21, the values are 0.056
au and 0.046 au, respectively. Hence, the π ring current in the
benzenium cation has a little under 60% of the strength of the
full π current of benzene, and this increases to ∼70% in the
difluoro anionic species, confirming that both ions, and in
particular the gem-difluoro-substituted cyclohexadienide anion
20, are aromatic on the magnetic criterion.
A final puzzle concerns the nature of the potential energy
surface for the dehydration reaction. Despite the experimental
evidence that the reaction proceeds in a stepwise manner via
formation of a carbanion intermediate, the exceptional stability
of the aromatic double bond of the product makes it somewhat
surprising that the reaction is not concerted, with simultaneous
loss of a proton and HO− leaving group. Indeed, the
combination of a stable double bond with close to zero
activation energy for loss of both the proton and the leaving
group from the intermediate suggests an unusual potential
energy surface for the reaction. Ab initio quantum chemical
calculations40,41 are currently underway to explore this surface
and the balance between the concerted and stepwise pathways.
The results will be reported in a later paper.
relative to the unsubstituted cyclohexadienide ion are shown
under the structures. From Table 2 we see that the
unsubstituted cycloheptadienide ion and the 6-methylcyclo-
hexadienide ion have practically the same PA. The difference
between these values and that for unsubstituted cyclohexa-
dienide ion is ∼2.8 kcal/mol, an increment that can be assigned
to the polarizability effect of having an additional carbon in the
seven-carbon-ring anions. The difference between the PAs of
the monofluorocyclohexadienide and monofluorocyclohepta-
dienide ions is 11.5 kcal/mol, which, allowing for the effect of
the extra carbon in the latter, gives a corrected difference of
∼14 kcal/mol. The same analysis using 7-hydroxycyclohepta-
dienide (19) shows that OH confers an extra stabilization of ∼6
kcal mol−1. It is likely that the fluorinated and hydroxylated
cycloheptadienides are somewhat stabilized by hyperconjuga-
tion, as indicated by the extended bond lengths of their
pseudoaxial C−F and C−OH bonds relative to those in their
conjugate acids. To that extent, the increased stabilities of the
fluorinated and hydroxylated cyclohexadienides reveal the
additional “aromatic” stabilizing effect over and above the
“nonaromatic” hyperconjugative stabilization of the seven-
membered-ring model ions. In the case of the fluoro
substituent, the magnitude of the stabilization (14 kcal
mol−1) agrees satisfactorily with that estimated above on the
basis of the o-fluorophenyl anion model (12 kcal mol−1).
Finally we mention that this aromaticity differs from that
inferred for transition states of nucleophilic aromatic
substitution reactions proceeding in a concerted manner
without formation of a Meisenheimer complex, as is character-
istic of most nth-row nucleophiles with n ≥ 3. Uggerud and
Frenkel attribute the relatively low activation energies of these
reactions to loose binding of a nucleophile and leaving group,
which allows a fraction of the aromaticity of the reactants to be
retained in the transition state.14
Orbital Interactions and Ring Current Calculations. A
calculation of the “ipsocentric”34 type confirmed that there is a
ring current associated with the gem-difluoro-substituted
cyclohexadienide anion 20. A current density map at a height
of 1a0 above the median plane was calculated for this system
with the distributed-gauge CTOCD-DZ method35,36 at the
RHF/6-31G** level of theory using the SYSMO37 and
GAMESS38 packages and a structure optimized at the
B3LYP/6-31G** level with GAUSSIAN 09.39 This map is
shown as structure 20. It displays the summed contributions of
all occupied molecular orbitals of π symmetry, which in
qualitative terms captures both the contributions from the
EXPERIMENTAL SECTION
■
General. NMR spectra were recorded on 300 or 500 MHz
instruments. Kinetic measurements utilized double-beam or single-
beam UV−vis spectrophotometers. For measurements over extended
time periods, a multiple cell changer was used and a reference
absorbance recorded as a function of time for subtraction from the
measured values for the sample(s). HPLC measurements normally
employed a dual-pump instrument, usually under isocratic conditions.
Standard reagents such as sodium hydroxide, tributyltin cyanide, and
14067
dx.doi.org/10.1021/ja304366j | J. Am. Chem. Soc. 2012, 134, 14056−14069

Journal of the American Chemical Society
Article
ethyl phenyl sulfide were purchased and used without further
purification. Reagents for kinetics were of AR or ARISTAR quality;
the water used as a solvent was distilled. Concentrated solutions of
NaOD in D2O were purchased and diluted with 99.8% D2O.
AUTHOR INFORMATION
■
Corresponding Author
keeffe@sfsu.edu
Reagents for Kinetic Measurements. The substituted benzene
cis-1,2-dihydrodiols studied kinetically in this work were prepared
earlier by bacterial oxidation of the corresponding substituted benzene
substrates by dioxygenase enzymes present in cultures of P. putida
UV4.42a−c The structures and spectra of most of the dihydrodiol
metabolites, including the unsubstituted benzene cis-dihydrodiol and
benzene cis-dihydrodiols with 3-substituents chloro, bromo, iodo,
phenyl, carbethoxy, trifluoromethyl, and methylsulfoxy were reported
previously.42 Proton NMR spectra are included in the Supporting
Information. Deuterated cis-1,2-dihydrodiols were prepared by similar
biotransformations of hexadeuterobenzene, pentadeutero bromoben-
zene, pentadeutero chlorobenzene and perdeuterated toluene. They
were identified from their Rf values on silica preparative layer
chromatography plates, which were identical to those of their
undeuterated analogues, and similarly identical reactant and product
UV spectra before and after hydroxide ion-catalyzed dehydration. The
perdeutero 3-cyanobenzene cis-dihydrodiol was prepared from the
corresponding 3-iodo-substituted cis-dihydrodiol by reaction with
tributyltin cyanide and Pd(Ph3P)4 for 4 h in THF solvent, as described
by Boyd et al.42a for the corresponding undeuterated compound.
Product Analyses. Products of the reactions of substituted
benzene cis-1,2-dihydrodiols in sodium hydroxide were analyzed by
NMR spectroscopy, HPLC, or GC−MS to identify and determine the
ratio of ortho- to meta-substituted phenols formed. The following
analysis for cis-1,2-dihydroxy-3-bromo-1,2-cyclohexa-3,5-diene is typ-
ical. The cis-dihydrodiol reactant (30 mg) in acetonitrile (5 mL) was
treated with a solution of 1.0 M sodium hydroxide for 15 h at 25 °C.
The pH of the reaction mixture was adjusted to 6−7 with saturated
aqueous sodium bicarbonate, and this was followed by extraction with
ethyl acetate and evaporation of the solvent under reduced pressure.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This paper is dedicated to the memory of Rory More O’Ferrall,
our colleague, collaborator, and friend. This work was
supported by the Science Foundation Ireland (Grant 04/
IN3/B581). Financial support for D.E.B. from the EPSRC and
the University of Sheffield (U.K.) is also gratefully acknowl-
edged, as is a Royal Society Conference Travel Grant for P.W.F.
Support for S.G. was provided by the National Science
Foundation (CHE-0716147). Support for S.C.L.K. was
provided by the Swedish Research Council, Vetenskapradet
(Grant No. 2012-5026).
̊
REFERENCES
■
(1) Boyd, D. R.; Blacker, J.; Byrne, B.; Dalton, H.; Hand, M. V.;
Kelly, S. C.; More O’Ferrall, R. A.; Rao, S. N.; Sharma, N. D.;
Sheldrake, G. N. J. Chem. Soc., Chem. Commun. 1994, 313−314.
(2) Boyd, D. R.; Coyne, D.; Keeffe, J. R.; Lawlor, D. A.; MacCormac,
A. C.; More O’Ferrall, R. A.; Rao, S. N.; Sharma, N. D. Org. Lett. 2010,
12, 5550−5553.
(3) Lawlor, D. A.; Bean, D. E.; Fowler, P. W.; Keeffe, J. R.; Kudavalli,
J. S.; More O’Ferrall, R. A.; Rao, S. N. J. Am. Chem. Soc. 2011, 133,
19729−19742.
(4) (a) Schubert, W. M.; Keeffe, J . R. J. Am. Chem. Soc. 1972, 94,
4048−4050. (b) Richard, J. P.; Jencks, W. P. J. Am. Chem. Soc. 1984,
106, 1373−1383. (c) Modena, G.; Rivetti, F.; Scorrano, G.; Tonellato,
U. J. Am. Chem. Soc. 1977, 99, 3392−3395.
(5) (a) Fishbein, J. C.; Jencks, W. P. J. Am. Chem. Soc. 1988, 110,
5074−5086. (b) Fedor, L. R. J. Am. Chem. Soc. 1969, 91, 908−913.
(c) More O’Ferrall, R. A.; Slae, S. J. Chem Soc. B 1970, 260−268.
(6) Saunders, W. H., Jr. J. Org. Chem. 1999, 64, 861−865.
(7) Reed, A. E.; Schleyer, P. v. R. J. Am. Chem. Soc. 1990, 112, 1434−
1445.
1
The H NMR spectrum was recorded, and the ratio of products was
determined by HPLC. Comparison of the HPLC peaks with those of
authentic samples of the isomeric products confirmed their identity
and allowed correction of the measured peak intensities for a small
difference in response factors (m-bromo/o-bromo = 0.9), to give 11%
as the fraction of the ortho isomer in the mixture. Analyses of the
products from the reaction of the 3-methylbenzene- and 3-
trifluoromethylbenzene cis-dihydrodiols were carried out in the same
way as for the bromo compound.
Kinetic Measurements. Kinetic measurements in aqueous sodium
hydroxide and NaOD/D2O were based on UV spectra. The cell
compartment of the spectrophotometer was thermostatted at 25 °C.
Initially, for each substrate, repetitive scans of spectra were recorded
throughout the dehydration reaction. Samples of the spectral changes
are shown in the Supporting Information. Measurements of rate
constants were based on kinetic measurements at a single wavelength,
usually at an absorption maximum of the substituted benzene cis-
dihydrodiol reactant or phenoxide ion product. Measurement
wavelengths and first-order rate constants measured at different
concentrations of sodium hydroxide or deuteroxide are reported in the
Supporting Information.
(8) Keeffe, J. R.; Gronert, S.; Colvin, M. E.; Tran, N. L. J. Am. Chem.
Soc. 2003, 125, 11730−11745.
(9) Boyd, D. R.; Bugg, T. H. Org. Biomol. Chem. 2006, 4, 181−192.
(10) Kresge, A. J.; More O’Ferrall, R. A.; Powell, M. F. In Isotopes in
Organic Chemistry; Buncel, E., Lee, C. C., Eds.; Elsevier: Amsterdam,
1986; Vol. 5.
(11) Guthrie, J. P. Can. J. Chem. 1992, 70, 1042−1054.
(12) Dey, J.; O’Donoghue, A. C.; More O’Ferrall, R. A. J. Am. Chem.
Soc. 2002, 124, 8561−8574.
(13) Bartmess, J. E. In NIST Standard Reference Database Number 69;
Mallard, W. G., Linstrom, P. J., Eds.; National Institute of Standards
and Technology: Gaithersburg, MD, 2003; http://webbook.nist.gov.
(14) Fernan
2971−2980.
́
dez, I.; Frenking, G.; Uggerud, E. J. Org. Chem. 2010, 75,
ASSOCIATED CONTENT
■
(15) Glukhotsev, M. N.; Bach, R. D.; Laiter, S. J. Org. Chem. 1997,
62, 4036−4046.
S
* Supporting Information
(16) Terrier, F. Nucleophilic Aromatic Displacement: The Influence of
the Nitro Group; Organic Nitro Chemistry Series; VCH: New York,
1991; Chapter 1.2.
Complete ref 39; details of kinetic measurements, including
UV−vis spectra of reactants and products (Figures S1 and S2);
isotope effects (Table S1 and Figures S3−S5); product analyses
(Figure S7); measurements of ionization constants (Table S2);
computational results (Tables S3−S7); observed and calculated
absorbances as a function of [HO−] for cyanobenzene 1,2-
dihydrodiol (Figure S6); plot of computed versus experimental
proton affinities for various anions (Figure S8); and NMR
spectra. This material is available free of charge via the Internet
at http://pubs.acs.org.
(17) Birch, A. J.; Hinde, A. L.; Radom, L. J. Am. Chem. Soc. 1980, 102,
6430−6437.
(18) Kim, Y.; Mohrig, J. R.; Truhlar, D .G. J. Am. Chem. Soc. 2012,
132, 11071−11082.
(19) Buncel, E.; Dust, J. M. Carbanion Chemistry: Stuctures and
Mechanisms; Elsevier: New York, 2003.
(20) Koch, H. F.; Biffinger, J. C.; Mishima, M.; Mustanir; Lodder, G.
J. Phys. Org. Chem. 1998, 11, 614−617.
14068
dx.doi.org/10.1021/ja304366j | J. Am. Chem. Soc. 2012, 134, 14056−14069

Journal of the American Chemical Society
Article
(21) Richard, J. P.; Williams, G.; Gao, J. J. Am. Chem. Soc. 1999, 121,
715−726.
(22) Giese, K.; Kaatze, U.; Pottel, R. J. Phys. Chem. 1970, 74, 3718−
3725.
(23) Kaatze, U. J. Chem. Eng. Data 1989, 34, 371−374.
(24) Kaatze, U.; Pottel, R.; Schumacher, A. J. Phys. Chem. 1992, 96,
6017−6020.
(25) Eigen, M. Angew. Chem., Int. Ed. Engl. 1964, 3, 1−19.
(26) Richard, J. P.; Williams, G.; O’Donoghue, A. C.; Amyes, T. L. J.
Am. Chem. Soc. 2002, 124, 2957−2968.
(27) Koch, H. F.; Dahlberg, D. G.; Lodder, G.; Root, K. S.;
Touchette, N. A.; Solaky, R. L.; Zuck, R. M.; Wagner, L. J.; Koch, N.
H.; Kuzemko, M. A. J. Am. Chem. Soc. 1983, 103, 2394−2398.
(28) Bernasconi, C. F. Adv. Phys. Org. Chem. 2010, 44, 223−324.
(29) Hine, J. J. Am. Chem. Soc. 1971, 93, 3701−3708.
(30) Bunting, W. J. Adv. Heterocycl. Chem. 1979, 25, 1−82.
(31) Toteva, M. M.; Richard, J. P. J. Am. Chem. Soc. 1996, 118,
11434−11445.
(32) Alexander, R.; Asomaning, R. A.; Eaborn, C.; Jenkins, D.;
Walton, D. R. M. J. Chem. Soc. Perkin Trans. 2 1974, 490−494.
(33) Bouchoux, G. Chem. Phys. Lett. 2011, 506, 167−174.
(34) Steiner, E.; Fowler, P. W. J. Chem. Soc., Chem. Commun. 2001,
2220−2221.
(35) (a) Keith, T. A.; Bader, R. F. W. Chem. Phys. Lett. 1993, 210,
223−231. (b) Coriani, S.; Lazzeretti, P.; Malagoli, M.; Zanasi, R. Theor.
Chim. Acta 1994, 89, 181−192. (c) Steiner, E.; Fowler, P. W. J. Phys.
Chem. 2001, 105, 9553−9562.
(36) Steiner, E.; Fowler, P. W.; Jenneskens, L. W. Angew. Chem., Int.
Ed. 2001, 40, 362−366.
(37) Lazzeretti, P.; Zanasi, R. SYSMO; University of Modena:
Modena, Italy, 1980. With extra mapping and orbital contribution
routines by E. Steiner, P. W. Fowler, R. W. A. Havenith, and A. Soncini
and an interface to GAMESS by R. W. A. Havenith.
(38) Guest, M. F.; Bush, I. J.; van Dam, H. J. J.; Sherwood, P.;
Thomas, J. M. H.; van Lenthe, J. H.; Havenith, R. W. A.; Kendrick, J.
Mol. Phys. 2005, 103, 719−747.
(39) Frisch, M. J.; et al. Gaussian 09, revision A.1; Gaussian, Inc.:
Wallingford, CT, 2009.
(40) Kamerlin, S. C. L. J. Org. Chem. 2011, 76, 9228−9237.
(41) Kamerlin, S. C. L.; Florian, J.; Warshel, A. ChemPhysChem 2008,
9, 1767−1773.
(42) (a) Boyd, D. R.; Hand, M. V.; Sharma, N. D.; Chima, J.; Dalton,
H.; Sheldrake, G. N. J. Chem. Soc., Chem. Commun. 1991, 1630−1632.
(b) Boyd, D. R.; Sharma, N. D.; Byrne, B.; Hand, M. V.; Malone, J. F.;
Sheldrake, G. N.; Blacker, J.; Dalton, H. J. Chem. Soc., Perkin Trans. 1
1998, 1935−1943. (c) Boyd, D. R.; Sharma, N. D.; Byrne, B. E.;
Shepherd, S. D.; Ljubez, V.; Allen, C. C. R.; Kulakov, L. A.; Larkin, M.
L.; Dalton, H. Chem. Commun. 2002, 1914−1915.
14069
dx.doi.org/10.1021/ja304366j | J. Am. Chem. Soc. 2012, 134, 14056−14069