Hyperaromatic stabilization of arenium ions

Article abstract of DOI:10.1021/ol1014027

Benzene-cis- and trans-1,2-dihydrodiols undergo acid-catalyzed dehydration at remarkably different rates: k_cis/k_trans = 4500. This is explained by formation of a β-hydroxycarbocation intermediate in different initial conformations, one of which is stabilized by hyperconjugation amplified by an aromatic no-bond resonance structure (HOC₆H₆ ⁺ HOC₆H₅ H⁺). MP2 calculations and an unfavorable effect of benzoannelation on benzenium ion stability, implied by pK_R measurements of -2.3, -8.0, and -11.9 for benzenium, 1-naphthalenium, and 9-phenanthrenium ions, respectively, support the explanation.

Full text of DOI:10.1021/ol1014027

ORGANIC
LETTERS
2
010
Vol. 12, No. 23
550-5553
Hyperaromatic Stabilization of Arenium Ions
§
5
†
Jaya S. Kudavalli, Derek R. Boyd, Dara Coyne, James R. Keeffe,
‡
†
†
David A. Lawlor, Aoife C. MacCormac, Rory A. More O’Ferrall,*
†
,†
†
S. Nagaraja Rao, and Narain D. Sharma
‡
School of Chemistry and Chemical Biology, UniVersity College Dublin,
Belfield, Dublin 4, Ireland, School of Chemistry and Chemical Engineering, Queen’s
UniVersity of Belfast, Belfast BT9 5AG, Northern Ireland, and Department of
Chemistry and Biochemistry, San Fancisco State UniVersity, 1600 Holloway AVenue,
San Francisco, California 91432, United States
rmof@ucd.ie
Received October 9, 2010
ABSTRACT
Benzene-cis- and trans-1,2-dihydrodiols undergo acid-catalyzed dehydration at remarkably different rates: kcis/ktrans ) 4500. This is explained
by formation of a ꢀ-hydroxycarbocation intermediate in different initial conformations, one of which is stabilized by hyperconjugation amplified
+
+
by an aromatic no-bond resonance structure (HOC
benzenium ion stability, implied by pK measurements of -2.3, -8.0, and -11.9 for benzenium, 1-naphthalenium, and 9-phenanthrenium ions,
respectively, support the explanation.
6
H
6
6 5
T HOC H H ). MP2 calculations and an unfavorable effect of benzoannelation on
R
Arene-cis-1,2-dihydrodiols are products of bacterial oxidation
of aromatic molecules by microorganisms such as Pseudomonas
putida containing dioxygenase enzymes. A characteristic
1
Scheme 1. Dehydration of cis-Benzene-1,2-dihydrodiol
reaction they undergo is acid-catalyzed dehydration to form
phenol as illustrated for benzene-cis-1,2-dihydrodiol 1 in
2
Scheme 1. The mechanism of this reaction is similar to that
of the dehydration of alcohols, with initial formation of a
carbocation (arenium ion) intermediate 2. However, in contrast
to most alcohols, the rate-determining step is formation rather
than reaction of the carbocation. This is because loss of a proton
to form an aromatic product or the alternative hydride rear-
The trans-isomers of arene dihydrodiols are not accessible
by bacterial oxidation, but they are accessible by synthesis.
4
Comparison of rate constants for the 1,2-dihydrodiols of
benzene shows that the cis-isomer is 4500 times as reactive
as the trans. This is a surprising result because one expects
the two isomers to react via a common carbocation inter-
mediate (2) and to do so at similar rates. Thus, it is unlikely
3
rangement (NIH shift) are both rapid reactions.
†
University College Dublin.
Queen’s University of Belfast.
San Francisco State University.
‡
§
(
(
1) Boyd, D. R.; Bugg, T. D. Org. Biomol. Chem. 2006, 4, 181–192.
2) Boyd, D. R.; Hand, M. V.; Kelly, S. C.; More O’Ferrall, R. A.;
(3) Nashed, N. T.; Bax, A.; Loncharich, R. J.; Sayer, J. M.; Jerina, D. M.
J. Am. Chem. Soc. 1993, 115, 1711–1722.
Rao, S. N.; Sharma, N. D. J. Chem. Soc., Chem. Commun. 1994, 313–314.
10.1021/ol1014027  2010 American Chemical Society
Published on Web 11/05/2010

that a greater stability of the trans than cis reactant can
less stable cis product (Scheme 2). The cis/trans ratio of 14 is
significantly lower than kcis/ktrans ) 50 for carbocation formation.
This would imply 50/14 ) 3.6 as the ratio of stabilities of trans
and cis reactants. This ratio seems a little large, but it is difficult
to exclude formation of some trans product from reaction of
an epoxide intermediate, and generation of the carbocation
from solvolysis of the monotrichloroacetate ester of the
phenanthrene-9,10-cis-dihydrodiol (5-cis) led to no detectable
fraction of trans-dihydrodiol, implying a cis/trans ratio >20
for this reactant (Table S2, SI).
account for more than a factor of 2 or 3 in the observed
5
,6
difference in reactivity.
Remarkably, the magnitude of the rate difference depends
on the aromaticity of the ring formed in the dehydration product.
Values of kcis/ktrans for Vic-dihydrodiols of benzene, naphthalene,
and phenanthrene are compared in Figure 1, and it can be seen
6,7
The implication of these results is that, in the case of
phenanthrene, the ratio of cis to trans products substantially
matches the rate constant ratio for reaction of the cis- and
trans-dihydrodiols (allowing for their difference in thermo-
dynamic stabilities). This strongly suggests that the same
would be true of the much higher rate ratios in the case of
benzene and naphthalene (4500 and 440) if their carbocations
could be similarly trapped. Direct measurements for car-
bocations derived from dihydrodiols of nonaromatic dou-
ble bonds, such as dihydronaphthalene, consistently give
small trapping ratios favoring the cis product. This was
demonstrated for the 1,2-dihydrodiols of 6-methoxy-1,2-
Figure 1. Cis/trans reactivity ratios for arene dihydrodiols.
that this ratio decreases to 440 for the naphthalene-1,2-
dihydrodiols and to 50 for the phenanthrene-9,10-dihydrodiols.
For the dihydrodiol of the nonaromatic double bond of 3,4-
dihydronaphthalene, kcis/ktrans drops to 5.7 (of which a factor of
1.5 stems from the difference in stabilities of the cis and trans
5
dihydronaphthalene and 5-methoxyindene by Whalen.
reactants). The ratios are not appreciably affected by replace-
6
Hyperconjugation. What is the explanation of this
behavior? We believe it lies in the formation of different
initial conformations of carbocations from cis- and trans-
dihydrodiol precursors. For example, from the cis-isomer of
ment of hydroxy by methoxy groups.
Nucleophilic Trapping of ꢀ-Hydroxycarbocations. A
corollary of these results is that generating an arenium ion
independently and trapping it with water to give dihydrodiols
in the back reaction should yield a much higher proportion
of the cis-dihydrodiol than its more stable trans isomer.
Generation of arenium ions derived from benzene or
naphthalene 1,2-dihydrodiols leads almost exclusively to
aromatic products. However, as shown in Scheme 2, aqueous
(O-protonated) benzene-1,2-dihydrodiol 8, a carbocation is
formed with a conformation 9 in which a pseudoaxial ꢀ-C-H
bond is optimally aligned for hyperconjugation with a
p-orbital of the carbocation. In this conformer, the ꢀ-OH
group is in a pseudoequatorial position. For carbocation
formation from the trans-benzene dihydrodiol (11) we
suppose that the positions of the ꢀ-OH and H are reversed
and that hyperconjugation of a C-OH bond in the resultant
carbocation 10 is much less favorable than that of a C-H
Scheme 2. Aqueous Solvolysis of Phenanthrene Bromohydrin
bond. The differentiation depends on the protonated hydroxyl
+
leaving group (OH
2
) departing from a pseudoaxial position
in the diol reactant. This is an expected constraint arising
from the requirement of efficient overlap of the developing
p-orbital of the carbocation with the π-bonds or benzene rings
8
of the arenium ion intermediate. It implies that the trans-
1
,2-dihydrodiol 11 will react from its diaxial rather than
diequatorial conformation (Scheme 3).
solvolysis of phenanthrene-9,10-bromohydrin 3 forms 8%
of 9,10-dihydrodiols in addition to the predominant 9-hydroxy-
phenanthrene 4. The ratio of cis- to trans-dihydrodiols (5
and 6) was found to be 14 ( 2 (Table S1, Supporting
Information (SI)).
Scheme 3. Conformations of the ꢀ-Hydroxybenzenium Ion
This measurement confirms that trapping of the 9-hydroxy-
phenanthrenium ion 7 leads to predominant formation of the
(
4) Platt, K. L.; Oesch, F. Synthesis 1982, 45, 9–461, and refs cited
therein.
(
5) Sampson, K.; Paik, A.; Duvall, B.; Whalen, D. L. J. Org. Chem.
If the stabilization of the carbocation arises from more
favorable C-H than C-OH hyperconjugation, why should
this depend on the stability of the π-bond formed in the
2
004, 69, 5204–5211.
(
6) Kudavalli, J. S.; Lawlor, D. A.; Coyne, D.; Keeffe, J. R.; More
O’Ferrall, R. A., in preparation.
Org. Lett., Vol. 12, No. 23, 2010
5551

product, which should have no influence on the stabilities
of either the dihydrodiol reactants or the arenium ion
intermediates? A simple explanation is that it depends on
the stability of the valence bond structures associated with
hyperconjugation treated as “no-bond resonance”. As il-
lustrated for the ꢀ-hydroxybenzenium ion 2a in Figure 2,
Figure 2
. “No-bond resonance” of the 2-hydroxybenzenium ion
2
; cyclo-2-hydroxycyclohexen-5,6-yl cation 12.
the contributing no-bond valence structure 2b includes a
benzene (phenol) ring. It is thus understandable that the
stereochemical constraint (implied by the cis/trans rate ratio)
tracks the aromatic stabilization of the final product because
this mirrors the structure contributing to the resonance.
The supposed order of stabilities of different conformations
of arenium ions is reproduced by MP2 calculations (6-
Figure 3
.
Taft plot of log k versus σ
I
I
(log k ) -0.49 - 8.84σ ) for
the acid-catalyzed dehydration of cis (filled circles) and trans (open
circles) 2-substituted 1,2-dihydro-1-naphthols in aqueous solution
at 25 °C; rate constants are listed in Table S4 (SI).
3
11+G** or 6-31G*) summarized in Table S5 (SI). In
practice, only one conformation corresponding to a minimum
energy, in which the ꢀ-C-H bond is pseudoaxial (as in 9),
is identified by the calculations. However, constraining dihedral
angles to give a structure close to that in which a ꢀ-OH group
rather than a ꢀ-H atom occupies a pseudoaxial position (as in
For clarity, other trans-substituents are not included in the
figure. For 2-alkyl and 2-aryl substituents, the trans isomer
is less reactive than the cis, but the ratios of rate constants
t
are much smaller than for OH or OMe (Me, 8.4; Bu , 12.7;
-1
Ph, 3.8). This is consistent with the groups exerting a
stabilizing hyperconjugative effect which, however, is less
than that of hydrogen. Other trans-substituents such as EtS,
10) increases the energy by 8.8, 6.0, and 4.3 kcal mol ,
respectively, for benzenium, 1-naphthalenium, and 9-phenan-
threnium ions bearing a hydroxyl group ꢀ to the carbocation
center. For the cation 12 derived from reaction of the dihy-
drodiol of 1,3-cyclohexadiene, the difference is 0.4 kcal. Com-
3
PhS, and N show positive deviations consistent with
neighboring group participation. Only protonated 2-amino
substituents exhibit cis/trans rate ratios comparable to (and
perhaps surprisingly smaller than) OH or OMe; e.g., for
parison of hyperconjugating groups gives the expected order
6
of stabilization H
3
Si > H > CH
3
> OH > F. An independent
+
6
9
NH , kcis/ktrans ) 315.
3
estimate of magnetic ring currents suggests a value for the
benzenium ion close to half that for benzene.
10
Evidence for “aromaticity-enhanced” stabilization of are-
nium ions by hyperconjugation is not confined to stereo-
chemical comparisons. It comes also from measurements of
thermodynamic stabilities. These stabilities are conveniently
The above energy differences imply that the transition
states and carbocation conformations (e.g., 10) formed from
the arene-trans-dihydrodiols are exceptionally unstable. This
is confirmed and the alternative possibility that the cis-
conformation is unusually reactive ruled out, by a Taft
correlation for dehydration of 2-substituted 1,2-dihydro-1-
naphthols 13 shown in Figure 3. The filled circles refer to
expressed by the equilibrium constant K
stability of an ion (ArH ) to that of its formal hydrolysis
product (ArHOH) as shown in eqs 1 and 2.
R
, which relates the
+
+
+
cis-substituents, and these form a good correlation (F
I
) 8.8)
ArH + H O h ArHOH + H
(1)
(2)
2
with small but distinct deviations for hydroxy and methoxy
groups. The open circles represent trans-hydroxy and -meth-
oxy groups, and these show pronounced negative deviations
consistent with the interpretation offered above.
[
ArHOH][H+]
K )
R
ArH+]
[
Values of pK
R
(-log K
R
) for the benzenium, 1-naphthale-
(
7) Whalen, D. L.; Ross, A. M.; Dansette, P. M.; Jerina, D. M. J. Am.
Chem. Soc. 1977, 99, 5672–5676.
8) Goering, H. L.; Josephson, R. R. J. Am. Chem. Soc. 1962, 84, 2779–
nium, and 9-phenanthrenium ions are shown under the relevant
structures in Figure 4. It is apparent that pK decreases along
the series from -2.3 to -8.0 to -11.6. This implies a decrease
in stability of nearly ten powers of ten in going from protonated
benzene to protonated phenanthrene.
(
R
11
2
785.
(
9) Steiner, E.; Fowler, P. A. J. Chem. Soc., Chem. Commun. 2001,
2
220–2221.
(
10) Fowler, P. A.; Bean, D., to be published.
5552
Org. Lett., Vol. 12, No. 23, 2010

1
4,19
of benzene bound to tin, germanium, and lead.
Here the
hyperconjugation is for C-Si, C-Sn, C-Ge, and C-Pb
20,21
bonds. These indeed are much more effective than C-H.
Yet no serious suggestion of an aromatic component to the
stabilization was entertained.
In 1993, Schleyer provided computational evidence for
the aromatic character of σ-π electron delocalization in the
2
2
phenonium and benzenium ions. Subsequent studies,
including calculations of aromaticity indices for cyclopen-
tadienes substituted at the methylene group with Si, Sn, and
2
3
Ge substituents, have identified him as a consistent
advocate of hyperconjugative aromaticity. His conclusions
24,25
Figure 4
.
Influence of benzoannelation on carbocation stability
have not escaped criticism
by the present results.
but are emphatically endorsed
(
cycloheptadienyl charges indicate sites of reaction with water).
In view of its theoretical and experimental significance, it
seems appropriate to characterize aromatic hyperconjugation
by the term “hyperaromaticity”. This implies an analogy
between hyperconjugation and hyperaromaticity on the one
hand and homoconjugation and homoaromaticity on the
This might not seem remarkable. However, solvolysis of
substrates yielding benzylic carbocations normally occurs more
readily than that of those yielding structurally comparable allylic
12
cations. This difference in reactivity is opposite to that
expected from the effect of benzoannelation on the stability of
the benzenium ion.
2
6
other. The term would seem to be consistent with Mul-
27
liken’s original definition of hyperconjugation as “conjuga-
tion over and above that usually recognized”. It remains
remarkable that so apparently important a concept should
only now find extensive experimental support.
The unusual stability of the benzenium ion is confirmed by
6
comparison with the cycloheptadienyl cation in Figure 4. The
difference in pK for these two ions, -2.3 and -12.1, seems
R
too large to be accounted for by a difference in strain energies
but is consistent with a difference of aromatic from nonaromatic
hyperconjugation. Moreover, as indicated by the inferred values
Acknowledgment. The work was supported by the Sci-
ence Foundation Ireland (Grant No. 04/IN3/B581). The
authors thank Yitzhak Appeloig (Technion, Israel Institute
of Technology), Herbert Mayr (Ludwig Maximilian Uni-
versity of Munich), Paul Schleyer (University of Georgia),
and Patrick Fowler (University of Sheffield) for helpful
discussion and suggestions.
13
of pK
R
above, for the cycloheptadienyl cation benzoannelation
decreases from -11.8 to -8.7).
does increase its stability (pK
R
Electrophilic Aromatic Substitution. The benzenium ion
(
cyclohexadienyl cation, protonated benzene) is the parent
Wheland intermediate of electrophilic aromatic substitution
and is itself implicated in aromatic hydrogen isotope ex-
Supporting Information Available: Experimental pro-
cedures, kinetic measurements, and tables of rate constants
and product ratios (S1-S4) and of computational details
1
4
change. Stabilization of this ion by hyperconjugation
associated with significant aromatic character was originally
1
5
envisaged by Mulliken in 1953. Perhaps because the
magnitude of this stabilization was later thought to be
(Table S5). This material is available free of charge via the
Internet at http://pubs.acs.org.
16
overestimated, and certainly because experimental implica-
tions were lacking or overlooked, the notion remained largely
dormant for 40 years. There is no mention of it in Dewar’s
OL1014027
1
7
(16) Ermier, W. C.; Mulliken, R. S.; Clementi, E. J. Am. Chem. Soc.
monograph on hyperconjugation in 1964 or Taylor’s
1
976, 98, 388–394
17) Dewar, M. J. S. Hyperconjugation; Ronald: NY, 1964
(18) Ahira, J. Bull. Chem. Soc. Jpn. 1981, 54, 2268–2273
.
14
account of electrophilic aromatic substitution in 1990, while
(
.
18
a discussion by Ahira in 1980 attracted only two citations.
Hyperconjugation was invoked in the 1960s to account
for the high reactivity of trimethylsilylbenzene toward
protodesilylation and the much greater reactivity with acid
.
(
19) Eaborn, C.; Pande, K. C. J. Chem. Soc. 1960, 1566–1571
.
(
20) Panisch, R.; Bolte, M.; Muller, T. Organometallics 2007, 26, 3524–
3529
21) Lambert, J. B.; Zhao, Y.; Emblidge, R. W.; Salvador, L. A.; Liu,
X.; So, J.-H.; Chelius, E. C. Acc. Chem. Res. 1999, 32, 183–190
22) Sieber, S.; Schleyer, P. von R. J. Am. Chem. Soc. 1993, 115, 6987–
6988.
(23) Nyulazi, L.; Schleyer, P von R. J. Am. Chem. Soc. 1999, 121, 6872–
6875.
(24) Stanger, A. Chem.sEur. J. 2006, 12, 2745–2751
(25) Olah, G. A.; Head, N. J.; Rasul, G.; Suraya Prakash, G. K. J. Am.
Chem. Soc. 1995, 117, 875–882
.
(
.
(
(
11) Lawlor, D. A.; More O’Ferrall, R. A.; Rao, S. N. J. Am. Chem.
Soc. 2008, 130, 17997–12807.
12) Rao, S. N.; More O’Ferrall, R. A.; Kelly, S. C.; Boyd, D. R.;
Agarwal, R. J. Am. Chem. Soc. 1993, 115, 5458–5465.
13) Ceccon, A.; Gambaro, A.; Romanin, A.; Venzo, A. Angew. Chem.,
Int. Ed. 1983, 22, 559–560.
14) Taylor, R. Electrophilic Aromatic Substitution; John Wiley and
Sons: Chichester, 1990.
15) Pickett, L. W.; Muller, N.; Mulliken, R. S. J. Chem. Phys. 1953,
1, 1400–1401.
(
.
(
.
(
(26) Winstein, S. In Carbonium Ions; Olah, G. A., Schleyer, P. von R.,
Eds.; Wiley: New York, 1972; Vol. III.
(
(27) Mulliken, R. S.; Riecke, C. A.; Brown, W. G. J. Am. Chem. Soc.
1941, 63, 41–56.
2
Org. Lett., Vol. 12, No. 23, 2010
5553