Hyperaromatic stabilization of arenium ions: A remarkable cis stereoselectivity of nucleophilic trapping of β-hydroxyarenium ions by water

Article abstract of DOI:10.1021/ja207160z

Cis- and trans-1,2-dihydrodiol isomers of benzene undergo acid-catalyzed dehydration to form phenol. In principle the isomeric substrates react through a common β-hydroxybenzenium (cyclohexadienyl) carbocation. Notwithstanding, the isomers show a large difference in reactivity, k_cis/k _trans = 4500. This difference is reduced to k_cis/k _trans = 440 and 50 for the 1,2-dihydrodiols of naphthalene and 9,10-dihydrodiols of phenanthrene, respectively, and to 6.9 for the dihydrodiols of the nonaromatic 7,8-double bond of acenaphthylene. Because the difference in stabilities of cis- and trans-dihydrodiols should be no more than 2-3-fold, these results imply a high cis stereoselectivity for nucleophilic trapping of a β-hydroxyarenium cation by water in the reverse of the carbocation-forming reaction. This is confirmed by studies of the 10-hydroxy-9-phenanthrenium ion generated from aqueous solvolyses of the trans-9,10-bromohydrin derivative of phenanthrene and the monotrichloroacetate ester of the phenanthrene cis-9,10-dihydrodiol. The cis stereoselectivity of forward and reverse reactions is explained by the formation (in the "forward" reaction) of different conformations of carbocation from cis- and trans-dihydrodiol reactants with respectively β-C-H and β-C-OH bonds in pseudoaxial positions with respect to the charge center of the carbocation optimal for hyperconjugation. Formation of different conformations is constrained by departure of the (protonated) OH leaving group from a pseudoaxial position. The difference in stability of the carbocations is suggested to stem (a) from the greater hyperconjugative ability of a C-H than a C-OH bond and (b) from enhanced conjugation arising from the stabilizing influence of an aromatic ring in the no-bond resonance structures representing the hyperconjugation (C ₆H₆OH⁺ ? C₆H₅OH H⁺). This is consistent with an earlier suggestion by Mulliken and a demonstration by Schleyer that the benzenium ion is subject to hyperconjugative aromatic stabilization. It is proposed that, in analogy with the terms homoconjugation and homoaromaticity, arenium ions should be considered as "hyperaromatic".

Full text of DOI:10.1021/ja207160z

ARTICLE
pubs.acs.org/JACS
Hyperaromatic Stabilization of Arenium Ions: A Remarkable Cis
Stereoselectivity of Nucleophilic Trapping of β-Hydroxyarenium
Ions by Water
David A. Lawlor,† Jaya Satyanarayana Kudavalli,† Aoife C. MacCormac,† Dara A. Coyne,† Derek R. Boyd,‡
and Rory A. More O’Ferrall*,†
†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
S Supporting Information
b
ABSTRACT: Cis- and trans-1,2-dihydrodiol isomers of benzene undergo acid-
catalyzed dehydration to form phenol. In principle the isomeric substrates
react through a common β-hydroxybenzenium (cyclohexadienyl) carbocation.
Notwithstanding, the isomers show a large difference in reactivity, kcis/ktrans
=
4500. This difference is reduced to kcis/ktrans = 440 and 50 for the 1,2-
dihydrodiols of naphthalene and 9,10-dihydrodiols of phenanthrene, respec-
tively, and to 6.9 for the dihydrodiols of the nonaromatic 7,8-double bond of
acenaphthylene. Because the difference in stabilities of cis- and trans-dihydrodiols should be no more than 2ꢀ3-fold, these results
imply a high cis stereoselectivity for nucleophilic trapping of a β-hydroxyarenium cation by water in the reverse of the carbocation-
forming reaction. This is confirmed by studies of the 10-hydroxy-9-phenanthrenium ion generated from aqueous solvolyses of the
trans-9,10-bromohydrin derivative of phenanthrene and the monotrichloroacetate ester of the phenanthrene cis-9,10-dihydrodiol.
The cis stereoselectivity of forward and reverse reactions is explained by the formation (in the “forward” reaction) of different
conformations of carbocation from cis- and trans-dihydrodiol reactants with respectively β-CꢀH and β-CꢀOH bonds in
pseudoaxial positions with respect to the charge center of the carbocation optimal for hyperconjugation. Formation of different
conformations is constrained by departure of the (protonated) OH leaving group from a pseudoaxial position. The difference in
stability of the carbocations is suggested to stem (a) from the greater hyperconjugative ability of a CꢀH than a CꢀOH bond and (b)
from enhanced conjugation arising from the stabilizing influence of an aromatic ring in the no-bond resonance structures
representing the hyperconjugation (C6H6OH+ T C6H5OH H+). This is consistent with an earlier suggestion by Mulliken and a
demonstration by Schleyer that the benzenium ion is subject to hyperconjugative aromatic stabilization. It is proposed that, in
analogy with the terms homoconjugation and homoaromaticity, arenium ions should be considered as “hyperaromatic”.
Scheme 1
’ INTRODUCTION
Arene cis-dihydrodiols are products of oxidative biotransfor-
mations of aromatic molecules in bacterial cells.12 They are
prepared on a large scale3 by employing bacterial cultures of
genetically modified microorganisms, such as Pseudomonas puti-
da UV4, where the enzyme responsible for further conversion to
catechols during normal metabolism is blocked.4 Their structure
and chirality (e.g., of the dihydrodiols of monosubstituted
benzenes) have made them valued as synthons in the preparation
of a wide range of molecules including carbasugars, terpenes,
inositols, and pyrethroids.5
A characteristic reaction illustrated for benzene cis-dihydrodiol
1 in Scheme 1 is dehydration in the presence of aqueous acid to
yield phenol. Studies of this reaction,6,7 and of the related
dehydration of arene hydrates,8,9 establish that it proceeds
through formation of a carbocation intermediate (e.g., 2) and
that, in contrast to the analogous dehydration of alcohols
yielding nonaromatic products, formation rather than reaction
of the carbocation is rate-determining.10 This is because loss
of a proton from an arenium ion is accelerated by formation
of a highly stabilized aromatic product or, in the case of
β-hydroxyarenium ions such as 2, by a hydride (NIH) shift
leading to the keto tautomer of the phenolic product, which
rapidly tautomerizes.11
Arene trans-dihydrodiols are not accessible by bacterial fer-
mentation but are accessible by chemical synthesis.12ꢀ14 Ben-
zene trans-1,2-dihydrodiol 3 for example may be prepared in four
steps from 1,4-cyclohexadiene.12 A remarkable finding is that
although, like their cis isomers, they undergo acid-catalyzed
dehydration to form phenols, they do so much more slowly.
Thus, for the benzenedihydrodiols, kcis/ktrans = 4500.
Received: July 30, 2011
Published: October 24, 2011
r
2011 American Chemical Society
19718
dx.doi.org/10.1021/ja207160z J. Am. Chem. Soc. 2011, 133, 19718–19728
|

Journal of the American Chemical Society
ARTICLE
Scheme 2
Table 1. Rate Constants for Carbocation Formation from
Benzene and Phenanthrene Dihydrodiols in Aqueous Solu-
tion at 25 °C
rate constant per
substrate
OH group (Mꢀ1 sꢀ1
)
cis-benzene-1,2-dihydrodiol
trans-benzene-1,2-dihydrodiol
cis-phenanthrene-9,10-dihydrodiol
trans-phenanthrene-9,10-dihydrodiol
a From ref 6.
0.055a
1.21 ꢁ 10ꢀ5
2.2 ꢁ 10ꢀ6
4.4 ꢁ 10ꢀ8
Scheme 3
much more favorable thermodynamically and kinetically. The same
is true of β-hydroxynaphthalenium ions. However, in the case of the
10-hydroxy-9-phananthrenium ion, a little less than 10% of dihy-
drodiols is formed in competition with loss of a proton to yield the
aromatic product.
In this paper we show that the β-hydroxyphenanthrenium ion
may be generated as a reactive intermediate from hydrolysis of
the trans-9,10-bromohydrin of phenanthrene or the mono
trichloracetate ester of the phenanthrene-cis-9,10-dihydrodiol.
HPLC analyses of the products of the reactions of these
substrates in water or aqueous acetonitrile confirm that the
dihydrodiols formed are very predominantly of cis stereochem-
istry. The same is true of trapping with azide ions, which yields a
cis-azido alcohol as product.21
These results point to an explanation for the difference in
reactivity of cis- and trans-dihydrodiols in terms of hyperconju-
gative stabilization of an initially formed conformation of the
carbocation which is more effective for a suitably oriented CꢀH
than CꢀOH bond. The unusually large magnitude of the
stabilization is suggested to come from the contribution of an
aromatic component to the hyperconjugation, as previously
proposed by Mulliken22,23 and Schleyer.24 As explained below,
this reflects the stabilizing presence of an aromatic no-bond
structure when the hyperconjugation is represented in terms of
limiting resonance forms. In the following paper25 further
experimental evidence and computational studies supporting
this interpretation are presented.
This large difference in reactivity is unexpected because
apparently the two isomers react through a common carbocation
intermediate 2 (Scheme 2). The difference in rates might there-
fore have been expected to reflect the difference in stabilites of cis
and trans reactants which, based on MP2 calculations15 and
comparisons with indane and tetrahydronaphthalene dihydrodiols,16
probably amounts to no more than a factor of 2- or 3-fold in reactivity.
The much larger difference observed experimentally implies a
substantial difference in energies of isomeric transition states.
A corollary of the high cis/trans reactivity ratio is that trapping
of the β-hydroxybenzenium ion 2 with water as a nucleophile in
the reverse of the carbocation-forming reaction should be subject
to a comparable difference in selectivity, i.e., that a cis-dihydrodiol
product 1 should be formed at least 1000 times more rapidly
than its trans isomer 3. This is a remarkable conclusion because
β-hydroxycarbocations have been extensively studied as reactive
intermediates, especially in connection with the mammalian
metabolism of carcinogenic polycyclic aromatic hydrocarbons.17
Typically they are formed from acid-catalyzed ring-opening of
epoxides and react to form either predominantly trans or a mild
excess of cis product.18 Thus Whalen has shown that the epoxide
of the 1,2-dihydro-7-methoxynaphthalene 4 reacts via benzylic cation
5 to yield 81% cis-diol and 19% trans product (Scheme 3).19 Pref-
erence for the cis-diol 6 over the more stable trans isomer 7 was
explained in terms of conformational factors or stabilization of the
transition state for cis attack by hydrogen bonding between the
attacking water nucleophile and β-hydroxy group.16ꢀ20
’ RESULTS
Dehydration of Arene Dihydrodiols. Acid-catalyzed dehy-
drations of cis- and trans-benzene and naphthalene-1,2-dihydrodiols
and cis- and trans-phenanthrene-9,10-dihydrodiols were monitored
spectrophotometrically. For the benzene and phenanthrenene
dihydrodiols, the measured or extrapolated second-order rate con-
stants are shown in Table 1. First-order rate constants are listed in
Tables S1 and S2 (Supporting Informaion).
As shown in Scheme 4, for reactions of the naphthalene
dihydrodiols 8 and their dimethyl ethers 9, two products are
possible corresponding to loss of the OH group at the 1- or
2-position to yield 2- and 1-naphthol or the corresponding ethers
as products, respectively. Individual rate constants kα and kβ were
derived by combining the measured rate constants with measure-
ments of product ratios for the two phenols or phenolic ethers by
HPLC or GC. Measured product ratios and factored second-
order rate constants for the dihydronaphthalene substrates are
summarized in Table 2. First-order rate constants are in Tables
S3 and S4.
The ratio of cis to trans products in this example is small
compared with the >1000-fold difference implied for trapping of
the β-hydroxybenzenium ion 2. It seems hardly possible that in
aqueous solution such a large difference could be attributed to
hydrogen bonding or simple conformational effects. In this paper
we compare cis/trans reactivity ratios for acid-catalyzed dehydra-
tion of dihydrodiols of benzene, naphthalene, and phenanthrene
and show that the magnitude of the cis stereoselectivity is
correlated with the aromaticity of the double bond from which
the dihydrodiol is derived.
It is not possible to confirm by direct experiment that
nucleophilic trapping of a β-hydroxybenzenium ion yields the
expected high ratio of cis/trans dihydrodiol products because the
competing loss of a proton to form the aromatic phenolic product is
It is noteworthy that the 2-position of the dihydronaphthalene
reactants is consistently more reactive than the 1-position.
19719
dx.doi.org/10.1021/ja207160z |J. Am. Chem. Soc. 2011, 133, 19718–19728

Journal of the American Chemical Society
ARTICLE
Scheme 4
Table 2. Rate Constants and Product Fractions for Reactions of cis- and trans-2-Hydroxy-1,2-dihydro-1-naphthols 8 and Their
Dimethyl Ethers 9 in Aqueous HClO4 at 25 °C
c
d
substituent
[H+]a (M) reactionb time (h) kobs (Mꢀ1 sꢀ1
)
reaction at 1-position (%) reaction at 2-position (%) kα (Mꢀ1 sꢀ1
)
kβ (Mꢀ1 sꢀ1
)
cis-1,2-OH, 8-cis
0.1
2.0
2.0
2.0
5
15
3
1.48 ꢁ 10ꢀ3
4.00 ꢁ 10ꢀ6
1.70 ꢁ 10ꢀ4
9.4 ꢁ 10ꢀ7
4.7
3.8
95.3
96.2
83.6
83.3
7.0 ꢁ 10ꢀ5
1.41 ꢁ 10ꢀ3
3.85 ꢁ 10ꢀ6
1.42 ꢁ 10ꢀ4
7.83 ꢁ 10ꢀ7
trans-1,2-OH, 8-trans
cis-1,2-OMe, 9-cis
1.5 ꢁ 10ꢀ7
2.8 ꢁ 10ꢀ5
1.6 ꢁ 10ꢀ7
16.4
trans-1,2-OMe, 8-trans
12
16.6
c
a Acid concentration for product analysis. b Reaction time for product analyses. Rate constant for loss of OH or OMe from the 1-position. d Rate
constant for loss of OH or OMe from the 2-position.
Scheme 5
Scheme 6
The difference is large enough that factored rate constants kβ for
the 2-position are significantly more accurate. However, derived
reactivity ratios for cis and trans reactants show rather small
differences between the two positions: kcis/ktrans = 357 (β)
compared with 440 (α) for the diols 8 and kcis/ktrans = 181 (β)
compared with 175 (α) for the methyl ethers 9. Thus, while a
comparison with benzene and phenanthrene dihydrodiols seems
most appropriate for reactions at the 1-position of the naphthalene
dihydrodiols, it is unlikely that uncertainty in these measurements
has a significant influence on the results.
In the case of the cis-phenanthrene-9,10-dihydrodiol 10,
the measured rate constant for formation of phenanthrol 12 re-
quires a minor correction for the reversibility of formation of the
10-hydroxy-9-phenanthrenium ion intermediate 11 (Scheme 5)
to yield the rate constant kcis for formation of this ion. The value
of kꢀ1/k2 needed for the correction was determined from
partitioning of the carbocation 11 generated by solvolysis of
the phenanthrene-9,10-bromohydrin as described below. In
principle, formation of the same ion from the trans-phenan-
threne-9,10-dihydrodiol is also reversible. In practice, as also
shown below, the “reverse” reaction yields very predominantly
the cis-dihydrodiol, which reacts much more rapidly than the
trans, so the measured rate constant needs no correction.
Isomerization of Acenaphthylene Dihydrodiols. Rate con-
stants for acid-catalyzed carbocation formation from cis- and
trans-7,8-dihydrodiols of acenaphthylene 13-cis and 13-trans
were also measured. For these reactants the final product in
concentrated solutions of aqueous perchloric acid is acenaphthe-
none 16. The obvious pathway for formation of this product is
acid-catalyzed dehydration of the dihydrodiols to form ace-
naphthenol 15, followed by tautomerization to its keto tautomer.
It might then have been expected that, in contrast to reaction of
the arene dihydrodiols, and by analogy with dehydration of
benzylic alcohols, carbocation formation and isomerization of
cis- to trans-dihydrodiols would be substantially faster than
formation of (acenaphthenol and) acenaphthenone. Consistent
with this expectation, spectrophotometrically determined rate
constants for dehydration of the cis- and trans-dihydrodiols have
similar values in the acid concentration range 4ꢀ6 M HClO4
(Table S5). However, HPLC analysis showed that, on the
contrary, cisꢀtrans isomerization was only a little faster than
product formation. The most likely explanation for this behavior
is that the acenaphthenone is formed directly from a carbocation
intermediate by hydride rearrangement, as shown in Scheme 6.
Figure 1 shows the fractions of cis- and trans-dihydrodiols 13
and acenaphthenone 16 plotted as a function of time for reaction
of the cis isomer in 6 M perchloric acid at 25 °C. It is clear that the
19720
dx.doi.org/10.1021/ja207160z |J. Am. Chem. Soc. 2011, 133, 19718–19728

Journal of the American Chemical Society
ARTICLE
Scheme 8
Information. The measurements with sodium azide led to
trapping of the ion 14 as cis- and trans-azido alcohols and will
be described fully elsewhere.21 There was no indication of
formation of an epoxide (acenaphthylene-6,7-oxide) accompa-
nying solvolysis of the acenaphthylene halohydrins, as was
observed for solvolysis of the phenanthrene 9,10-bromohydrin
described below.
Figure 1. Plots against time of % cis (9) and trans (O) acenaphthylene-
6,7-dihydrodiols (13-cis and 13-trans) and acenaphthenone (16, b) for
reaction of the cis-dihydrodiol isomer in 6 M aqueous HClO4 at 25 °C.
The full lines represent best fits of the points to a series first-order kinetic
analysis based on Schemes 6 and 7.
Evaluation of the rate constants ka and kb in Scheme 7 from
kinetic measurements at 5 and 6 M HClO4 gave an average value
of the constant for equilibration of the isomers, Keq
=
Scheme 7
[trans]eq/[cis]eq = ka/kb = 2.7 ( 0.2. Combination of this value
with kꢀ1/k2 (Scheme 6) then gave as a ratio of rate constants for
reaction of cis and trans dihydrodiols to form the common
carbocation intermediate, kcis/ktrans = Keqkꢀ1/kꢀ2 = 6.9. This is
the most important result of the analysis in that it provides a
comparison between values of kcis/ktrans for dihydrodiols derived
from aromatic and nonaromatic hydrocarbons.
It should be noted that partitioning of the β-hydroxyace-
naphthenium cation 14 fomed from hydrolysis of the 7,8-
halohydrins (17-cis or 18-trans) in the absence of acid did not
lead to formation of acenaphthenone. As indicated below,
measurements at 5 and 6 M HClO4 gave some indication that
the rate of formation of acenaphthenone increased relative to
cisꢀtrans equilibration of the dihydrodiol isomers (13-cis and 13-
trans) as the acid concentration increased. Insofar as there seems
little reason to expect the equilibrium ratio of cis- to trans-
dihydrodiols to depend strongly on acid concentration, the above
rate constant ratio (kcis/ktrans) should be close to that in aqueous
solution at 25 °C.
Partitioning of 10-Hydroxy-9-phenanthrenium Ions in
Water. Solvolysis of the trans-9,10-bromohydrin of phenan-
threne 19 gave as principal products 9-phenanthrol 12 and
9,10-phenanthraquinone 20, together with 8% of predominantly
cis-phenanthrene-9,10-dihydrodiols 10 (Scheme 9). The phe-
nanthraquinone is an oxidation product of phenanthrol, and its
proportion in HPLC samples increased with time. Under the
reaction conditions, no phenanthraquinone is formed from the
diol products. For the purpose of the analysis, therefore, it was
assumed that the sum of concentrations of phenanthrol and
phenanthraquinone represents the fraction of phenanthrol in-
itially formed in the solvolyses.26 Thus the ratio of concentrations
of phenanthrol plus phenanthraquinone to cis-dihydrodiol was
used to determine the ratio of forward to reverse rate constants,
k2/kꢀ1 = 14, for reaction of the phenanthrenium ion intermedi-
ate 11 in the acid-catalyzed dehydration of the cis-dihydrodiol, as
described above (Scheme 5).
isomers are equilibrated in competition with formation of ac-
enaphthenone product, which in this case amounts to 70% of the
reaction mixture at the longest reaction time. The reactions may
be analyzed as series first-order processes involving two expo-
nential terms based on the simplified reaction scheme shown in
Scheme 7. Alternatively, the data may be treated by numerical
integration. Details of the analysis are provided in the Experi-
mental Section. The HPLC measurements are given in Tables
S6ꢀS9. Figure 1 shows a best fit of calculated to experimental
data based on numerical integration.
The rate constants ka, kb, kc, and kd in Scheme 7 may be
expressed in terms of the microscopic rate constants of Scheme 6,
in which the cisꢀtrans isomerization and formation of ace-
naphthenone are shown as occurring through a common 7-hy-
droxyacenaphthenium ion intermediate, 14. In principle, this
leads to determination of rate constants kcis and ktrans and rate
constant ratios kꢀ1/kꢀ2 and kꢀ1/k3. In practice, the ratio k
/
ꢀ1
k
ꢀ2, which corresponds in Scheme 6 to the ratio of rate constants
for partitioning of the 7-hydroxyacenaphthenium ion 14 between
formation of cis- and trans-dihydrodiol isomers 13-cis and 13-
trans was measured independently. As shown in Scheme 8, the
ion was generated by solvolysis of bromo- or chlorohydrins of
acenaphthylene (17-cis or 18-trans) in water in the absence of
acid. The ratio of cis to trans products formed, which corresponds
to kꢀ1/kꢀ2, was measured by HPLC in acetate buffers and in the
presence of unbuffered sodium azide (Tables S6 and S7). Both
halohydrins furnished the same ratio, kꢀ1/kꢀ2 = 2.3, and this was
used in the kinetic analysis based on Schemes 6 or 7 to reduce
the number of unknowns from four to three. Details of this
analysis are given in the Experimental Section and Supporting
The solvolyses were carried out in acetate buffers at pH 3.70,
4.55, and 5.10. Product analyses under these conditions are
19721
dx.doi.org/10.1021/ja207160z |J. Am. Chem. Soc. 2011, 133, 19718–19728

Journal of the American Chemical Society
ARTICLE
Scheme 9
Scheme 10
Table 3. Product Percentages from Reaction of trans-9-Bro-
mo-10-hydroxy-9,10-dihydrophenanthrene (19) in Acetic
Acid Buffers at 25 °Ca
Table 4. Fractions of cis-Phenanthrene-9,10-dihydrodiol
(10-cis) Formed from Solvolysis and Hydrolysis of cis-9-Tri-
chloroacetoxy-10-hydroxy-9,10-dihydrophenanthrene (21)
in Aqueous Acetonitrile at 25 °Ca
% cis-dihydro- % trans-dihydro-
% phenanthrol
pH
diol, 10-cis
diol, 10-trans [c]/[t]b
+ quinone, 12 + 20
5.10
4.55
3.70
6.65
6.73
6.70
1.25
0.80
0.60
5.3
9.0
92.1
92.5
92.7
% MeCN (v/v)
0
30
50
70
90
11.2
% cis-dihydrodiolb
21.2
15.6
9.72
6.04
3.09
a Measurements by HPLC at 270 nm with a substrate concentration of
a Measurements by HPLC at 270 nm with a substrate concentration of
1.14 ꢁ 10ꢀ3 M. b The other products are phenanthrol 12 and phenan-
thraquinone 20.
1.0 ꢁ 10ꢀ3 M. b [10-cis]/[10-trans].
summarized in Table 3, and it can be seen that significantly more
cis- than trans-dihydrodiol is formed, but that the proportion of
trans product increases with pH. This behavior is consistent with
formation of part of the trans-dihydrodiol from an epoxide
intermediate26 formed in a competing base-catalyzed reaction
of the bromohydrin 19, which should be favored with an
increase in pH.
The intermediacy of an epoxide in the formation of the trans-
dihydrodiol product 10-trans was confirmed by measurements in
azide buffers (to be reported in full elsewhere)21 in which
addition of 0.2 M sodium azide caused a decrease in the amount
of trans dihydrodiol and an accompanying formation of the trans
isomer of the azido alcohol 9-hydroxy-10-azido-9,10-dihydro-
phenanthrene. The amount of azido alcohol showed little
increase as the concentration of sodium azide was increased
from 0.2 to 0.8 M, consistent with its predominant formation
from ring-opening of the epoxide and complete conversion of
the epoxide to azido alcohol at >0.2 M concentration of
sodium azide.
If this assessment is correct, the ratio of cis- to trans-dihydro-
diols of 11.7 in Table 3 at the lowest pH represents a minimum
value. This was confirmed by examination of phenanthrol and
dihydrodiol products from the solvolysis of the monotrichlor-
oacetate ester 21 of the cis-dihydrodiol of phenanthrene 10
(Scheme 10). These products are shown in Table 4 for solvolysis
in water and acetonitrileꢀwater mixtures. For this substrate there
is no possibility of forming an epoxide. A complication is that
there is an additional pathway to cis-dihydrodiol 10-cis from
hydrolysis of the trichloroacetate ester arising from attack of
water at the carbonyl group. However, on the assumption that
the ratio of diol to phenanthrol products formed from a phen-
anthrenium ion intermediate (11) is similar to that from the
intermediate in the hydrolysis of the bromohydrin (19), the
large fraction of phenanthrol in the products implies that a
substantial proportion of the cis-dihydrodiol arises from attack
of water on a phenanthrenium ion intermediate of the
solvolysis reaction rather than the ester hydrolysis pathway.
No trans isomer was observed, and from the minimum amount
that could have been detected it was inferred that the ratio
of cis/trans product from trapping the phenanthrenium ion
must be g20.
Solvolysis Rate Constants. Rate constant for solvolysis
of the halohydrins of phenanthrene and acenaphthylene
based on single measurements in aqueous solution were 5.8 ꢁ
10ꢀ3 sꢀ1 for the trans-9,10-bromohydrin of phenanthrene 19,
(9.6 ( 1.0) ꢁ 10ꢀ5 sꢀ1 for the cis-7,8-chlorohydrin of ace-
naphthylene 17-cis, and (1.48 ( 0.15) ꢁ 10ꢀ3 sꢀ1 for the trans-
bromohydrin 18-trans of acenaphthylene. Solvolysis of the cis-
trichloroacetate ester 21 in water occurred with a rate constant
1.9 ꢁ 10ꢀ3 sꢀ1. Rate constants for the phenanthrene derivatives
were determined spectrophotometrically and for the ace-
naphthylene halohydrins from HPLC measurements given in
Table S8. For the phenanthrene substrates, it is at first surprising
that the phenanthrene trans-bromohydrin is only 3 times as
reactive as the trichloroacetate ester, despite bromide being a
much better leaving group than trichloroacetate ion. This
probably reflects the different relative configurations of the two
compounds and the greater reactivity of cis than trans stereo-
chemistry for SN1 reactions yielding an arenium ion, which is the
principal topic of this paper.
’ DISCUSSION
Cis/Trans Rate Ratios (kcis/ktrans) for Arene Dihydrodiols.
Reactivity ratios for the acid-catalyzed dehydration of cis- and
trans-dihydrodiols of benzene, naphthalene, and phenanthrene
are shown in Chart 1, together with the corresponding ratio for
the dihydrodiols of acenaphthylene. The rate constants refer to
rate-determining carbonꢀoxygen bond-breaking of the O-pro-
tonated reactants to form carbocation intermediates. As ex-
plained in the Introduction, this is the rate-determining step of
the directly measured dehydration reaction of the arene dihy-
drodiols. For the dihydrodiols of acenaphthylene, the ratio of rate
constants was accessed indirectly from measurements of the ratio
of cis- and trans-dihydrodiol products (13-cis and 13-trans)
formed from reaction with water of the acenaphthylenium ion
14 generated from a bromo- or chlorohydrin precursor and the
equilibrium constant between cis and trans isomers. The dihy-
drodiols of acenaphthylene were chosen for the comparison as
being both representative of dihydrodiols of nonaromatic double
19722
dx.doi.org/10.1021/ja207160z |J. Am. Chem. Soc. 2011, 133, 19718–19728

Journal of the American Chemical Society
ARTICLE
Chart 1. Cis/Trans Rate Ratios for Acid-Catalyzed Carbocation Formation from Arene Dihydrodiols
Scheme 11
bonds and otherwise similar in structure to the “aromatic”
dihydrodiols.
The comparison of values of kcis/ktrans in Chart 1 illustrates
their striking dependence on the aromatic character of the double
bond from which the dihydrodiol is derived. Thus kcis/ktrans
decreases from 4500 for the dihydrodiols of benzene to 440 and
50 for those of naphthalene and phenanthrene and to 6.9 for
those of acenaphthylene. The last comparison includes a factor of
3.0 arising from the difference in stabilities of the cis and trans
dihydrodiol reactants, and this difference seems to be typical of
such isomers,16,27 for which measurements range from 4.3 for the
7,8-dihydrodiol of 4-methoxyacenaphthylene27 to 1.5 for the 1,2-
dihydrodiol of 6-methoxytetrahydronaphthalene.16 A compari-
son of the naphthalene dihydrodiols with cis- and trans-1,2-
dimethoxy-1,2-dihydronaphthalene, for which kcis/ktrans = 175
(Table 2 and Scheme 4), suggests that the high cis/trans rate
ratios are characteristic of β-oxygen substituents in general rather
than specifically hydroxyl groups. The cis/trans rate ratios refer to
the 1-position in the case of the dihydronaphthalenes, but similar
ratios are observed for reactions at the 2-position (366 and 180
for 1,2-(OH)2 and 1,2-(OMe)2, respectively).
Before attempting to interpret the observed magnitudes and
variation in kcis/ktrans, we recall an implication of the small
difference in stabilities of cis- and trans-dihydrodiol isomers
noted in the Introduction, that the major part of the reactivity
difference between the isomers should be expressed in the
reverse reaction with water of a carbocation intermediate com-
mon to the reaction of cis and trans isomers for which the
stereochemical difference of the reactants has been lost.
In principle, the carbocations may be generated from solvo-
lysis reactions of diol derivatives with better leaving groups than
OH at a sufficiently mild pH that the dehydration reaction does
not occur. This is not practicable for the carbocation intermedi-
ates from the dehydration of benzene and naphthalene dihydro-
diols because the carbocation loses a proton to form phenol or
naphthol much more rapidly than it undergoes a nucleophilic
reaction with water. However, in the case of phenanthrene 9,10-
dihydrodiols, the carbocation leads not only to loss of a proton to
form phenanthrol but also to competing nucleophilic attack to
form cis- and trans-dihydrodiols.
Nucleophilic Trapping of the 10-Hydroxy-9-phenanthre-
nium Ion 11. The required carbocation may be generated
solvolytically from reactions of the trans-9,10-bromohydrin of
phenanthrene 19 or 9-monotrichloracetate of cis-9,10-dihydro-
diol 21. There is also extensive literature data on its formation
from acid-catalyzed ring-opening of phenanthrene 9,10-oxide
2211,25,28 and its benzoannelated analogues,11 to which we return
below. In this work examination of the products of hydrolysis in
aqueous solution of the bromohydrin 19 by HPLC showed,
in addition to 9-phenanthrol 12, 7% of cis-phenanthrenedi-
hydrodiol 10-cis. The latter product we infer was derived from
stereoselective nucleophilic attack of water on the 10-hydroxy-9-
phenanthrenonium ion 11, as shown in Scheme 11.
In practice, the 9-phenanthrol 12 in Scheme 11 is accompanied by
its oxidation product 9,10-phenanthraquinone 20, which was pre-
sumed to be formed from it.26 More importantly, a small fraction of
trans-dihydrodiol 10-trans is also formed. However, the amount of
trans-dihydrodiol depends on the pH, increasing at higher pH. From
this dependence, and from the observation that the fraction of trans-
dihdyrodiol is reduced by the addition of azide ion, it was inferred that
this isomer was formed partly via conversion of the bromohydrin to
phenanthrene-9,10-oxide 22, followed by opening of the epoxide ring
as shown in the lower pathway of Scheme 11. Studies of the pH
dependence of products from the ring-opening of phananthrene
oxide in aqueous solution by Bruice28 and Whalen26 have shown that,
at the relevant pH, about 15% of trans-dihydrodiol is formed, in
addition to phenanthrol as the predominant product and a small
amount of cis-dihydrodiol.
From measurements at low enough pH to prevent formation
of an epoxide, or in the presence of 0.2 M or greater concentra-
tion of sodium azide which traps the epoxide as an azido alcohol,21
the ratio of cis/trans-dihydrodiols [10-cis]/[10-trans] arising
from solvolysis of the trans-bromohydrin is 14.4 ( 1.9. The
predominant formation of cis-dihydrodiol is confirmed by HPLC
analysis of the products of solvolysis of the 9-trichloroacetate
ester of phenanthrene-9,10-dihydrodiol (21), a reaction which
cannot lead to formation of an epoxide. While part of the
cis-dihydrodiol formed must arise from direct hydrolysis of the
trichloroacetate ester through water attack at the carbonyl group
(Scheme 10), provided that the ratio of diol/phenanthrol formed
from a carbocation intermediate (11) is similar to that from the
intermediate in the hydrolysis of the bromohydrin 19, the large
fraction (79%) of phenanthrol in the products confirms that the
major part of the reaction proceeds via the carbocation. Failure to
detect the formation of any trans-dihydrodiol in this case implies
that the ratio of cis/trans products is g20. This indeed is greater
19723
dx.doi.org/10.1021/ja207160z |J. Am. Chem. Soc. 2011, 133, 19718–19728

Journal of the American Chemical Society
ARTICLE
than the ratio of dihydrodiols formed from the trans-bromohy-
drin, for which the trans isomer was clearly detectable. It is
noteworthy too that trapping with azide ion led to the formation
of no trans-azido alcohol from trichloroacetate ester.21
An alternative possibility is that the trans product arises from
nucleophilic attack on the protonated epoxide, leading to a concerted
mechanism for ring-opening. This would be consistent with the
probable instability of the 9-hydroxy-10-phenanthrenium ion inter-
mediate 11.7,21 However, it is not easily reconciled with the absence of
a significant SN2 pathway for hydrolysis of the 9,10-bromohydrin 19
implied by a lack of kinetic dependence on the strongly nucleophilic
azide ion.21 Such a combination of poorer leaving group and stronger
nucleophile should allow detection of a bimolecular reaction.
Also consistent with the relative instability of the 9-hydroxy-10-
phenanthrenium ion, particularly for a conformation in which the
OH group occupies a pseudoaxial position (see below), is the
probability that nucleophilic trapping of the ion occurs at or close to
the limit of solvent relaxation.7,21 Whalen has suggested that in these
circumstances the reaction may reflect the structure of the solvation
shell of the reactant or the presence of a “preassociated” solvent
molecule in the solvation shell of the carbocation.20 Indeed the
discrepancy in behavior is consistent with the established principle
that a leaving group may provide some protection from “frontside”
attack for a short-lived carbocation. As illustrated in structure 24, in
the case of the epoxide this naturally leads to formation of trans
product from the initial conformation of the carbocation.
While we do not illustrate the corresponding conformations of
carbocation (ion pairs) from reaction of the bromohydrin and
trichloroacetate ester, it is evident that shielding from the de-
parting bromide ion in the former case would lead to cis product.
We will see below that the trichloroacetate ester is expected to
yield directly a stable conformation for the carbocation, in which
a hyperconjugating hydrogen in a β-axial position renders it
uniquely prone to nucleophilic trapping leading to cis product. In
terms of Sayer’s account of ring-opening of phenanthrene oxide,
this would also correspond to the more stable structure into
which the initially formed conformation from the epoxide (24)
was converted in competition with trapping as trans-dihydrodiol
10-trans and 9-phenanthrol 12.11
Hyperconjugation of Arenium Ions. The above results
confirm that there is a complementary relationship between
cis/trans reactivity ratios for acid-catalyzed conversion of arene
dihydrodiols to β-hydroxycarbocations and the cis/trans selec-
tivity of nucleophilic attack by water (and azide ion21) on the ions
themselves. Similar product ratios of cis/trans dihydrodiols and
azido alcohols from the trapping reactions implies that the
cis/trans selectivity depends primarily on the influence of the
β-hydroxyl group of the carbocation rather than on the nature of
the nucleophile or leaving group.
Although the complementary relationship between reactivity
and product partitioning is established only for the 10-hydroxy-9-
phenanthrenium ion 11, it is a reasonable inference that it applies
also to β-hydroxybenzenium and -naphthalenium ions, even
though nucleophilic trapping of these cations cannot be observed
directly because of competing loss of a proton to form the
aromatic molecule. This brings us back to the question, why are
these rate ratios correlated with the stability of the aromatic
molecule from which the dihydrodiols derive?
For comparison, an independent study of solvolysis of the
methylated bromohydrin trans-9-methoxy-10-bromo-9,10-di-
hydrophenanthrene 23, for which there is also no possibility of
forming an epoxide, gave a ratio of cis- to trans-9-methoxy-10-
hydroxy-9,10-dihydrophenanthrene products of 14 ( 2.21 The
measured reactivity ratio for dehydration of these solvolysis pro-
ducts was kcis/ktrans = 50 ( 20. The large uncertainty in this ratio
derives in part from the need to separate the dehydration reaction
from the competing acid-catalyzed loss of methanol.21
It is clear that these results are consistent with the high cis/
trans reactivity ratio for acid-catalyzed dehydration of the 9,10-
dihydrodiols of phenanthrene 10 if allowance is made for a
reasonable difference in stability of the isomeric reactants. A
trapping ratio of 14 between nucleophilic attack of water on the
9-hydroxy-10-phenanthrenonium ion 11 to yield cis- and trans-
dihydrodiols implies a rather high ratio of 55/14 = 3.9 for the
equilibrium between these diols. However, as we have seen, the
ratio is measurably less if based on trapping of the carbocation
generated from solvolysis of the cis-trichloroacetate ester of
phenanthrene-9,10-dihydrodiol 21 rather than 9,10-trans-bro-
mohydrin 19.
Indeed a much larger discrepancy is apparent if we compare the
carbocation generated from these substrates with that from acid-
catalyzed ring-opening of phenanthrene-9,10-oxide 22.11,26 This
reaction too has been presumed to occur with the intermediacy of
the 9-hydroxy-10-phenanthrenonium ion 11. However, in contrast
to the solvolytic reactions, the predominant dihydrodiol is not the cis
isomer but the trans. In the pH range of our own studies (3.5ꢀ5.5),
the products comprise 15% trans-dihydrodiol, ∼5% cis-dihydrodiol,
and 80% phenanthrol.26 This is markedly different from the product
distribution from the bromohydrin 19 (7% cis- and <0.5% trans-
dihydrodiol) despite the fact that formally the same carbocation
intermediate is formed. In methanolic solution the fraction of trans-9-
methoxy-10-hydroxy product from acid-catalyzed ring-opening of
phenanthrene oxide increases to 54% (and cis to 9%).11
Differences in product distribution from carbocations gener-
ated from epoxide and halohydrin precursors have been well
documented by Whalen.26,29ꢀ31 Sayer has suggested that acid-
catalyzed ring-opening of the epoxide yields intitially a carboca-
tion in an unstable conformation (24) which is trapped by water
to form a trans-dihydrodiol in competition with conversion to a
more stable conformer which gives mainly the cis isomer.11 This
is consistent with an interpretation offered for the predominant
formation of trans-diols from hydrolysis of benzylic epoxides
which ring-open to form relatively unstable carbocations.19
The small difference in stabilities of the cis- and trans-arene
dihydrodiols themselves implies that differences in reactivity are
expressed in the transition states for formation (or in the reverse
direction nucleophilic trapping) of a β-hydroxycarbocation in-
termediate. The dilemma for attempts to interpret the differences
is how it can be linked to aromaticity for transition states of
reactions in which neither the dihydrodiol reactant nor carboca-
tion intermediate appears to be aromatic.
19724
dx.doi.org/10.1021/ja207160z |J. Am. Chem. Soc. 2011, 133, 19718–19728

Journal of the American Chemical Society
ARTICLE
As briefly indicated in the Introduction, the dilemma may be
resolved by recognizing that the carbocation intermediates
themselves are “aromatic”. This is most simply expressed by
writing valence bond resonance structures for hyperconjugation
of the CH2 group of the benzenium ion (25a T 25b). The
effectiveness of this resonance can then be supposed to be a
consequence of the stabilizing influence of an aromatic ring for
the no-bond resonance structure 25b. As already indicated, the
importance of this was recognized by Mulliken at an early stage in
his consideration of hyperconjugation.22 However, it was not
possible to assess reliably the magnitude of this stabilization until
a much later date. As indicated in the following paper25 and
demonstrated earlier by Schleyer and co-workers,24 a calculated
structure shows the two methylene hydrogens of the benzenium ion
to be equivalent, with both implicated in the hyperconjugation, and
a stabilization energy about half that of benzene.
Scheme 12
hyperconjugating capacity” for CꢀOH from positive destabiliza-
tion by this group. However, we defer discussion of this point to
the following paper.25
It seems likely that the unstable conformation of the carboca-
tion initially formed from trans-arene dihydrodiols relaxes rapidly
to the more stable conformation. Nevertheless the influence of
the higher energy of this conformation is expressed in the
transition states of reactions leading directly to it. In the reverse
reaction of the carbocation with a nucleophile, we infer that this
will normally occur by attack of the nucleophile on the more
stable conformation and will yield a product in which the
bonding of the nucleophile and β-OH is cis. However, a fraction
of reaction can be expected to occur through the small concen-
tration of the energetically less favorable conformation to yield
trans product.
An example of analogous behavior well known in carbocation
chemistry is provided by the difference in reactivity of norbornyl
reactants possessing exo or endo leaving groups. Although the
formation of structurally differentiated carbocation intermedi-
ates from endo and exo reactants 30 and 31 was envisaged by
Winstein and Trifan,33 it seems possible in this case that there is no
energy minimum for a less stable (unbridged) conformation of the
norbornyl cation and that the difference in reactivity arises from an
inability to express fully the stabilization arising from σ-bond
delocalization in the endo transition state.34 Nevertheless, the
analogy between CꢀC bond participation in formation of the
norbornyl cation from exo reactant and CꢀH participation for
formation of a β-hydroxycarbocation from a cis-dihydrodiol remains
close. Indeed it should be noted that the discrete existence of the
conformation 28 is by no means definitely established. Thus
attempts to calculate MP2 energies for this conformation without
geometric constraint, at least in the gas phase, indicate the absence of
a local minimum and give the more stable conformation instead.25
In this case too, therefore, there may exist only one stable con-
formation for the carbocation in solution, but nevertheless a marked
difference in energy of isomeric transition states.
The main purpose of this and the following paper25 is to
present experimental evidence of the aromaticity of arenium
ions. An attempt at overall assessment of the work and, to a lesser
extent, computational results (including estimates of magnetic
ring currents) will also be presented in the following paper. In the
present paper we aim to show that the stereochemical results can
indeed be interpreted by supposing that the arenium ion inter-
mediates are strongly stabilized by hyperconjugation
Origin of the Dependence of the Reactivity of Arene-1,2-
dihydrodiols on Stereochemistry. The very high cis/trans rate
ratios for reaction of vicinal arene dihydrodiols may be under-
stood if reactions of cis and trans isomers lead initially to different
conformations of a β-hydroxy carbocation intermediate. The
argument that they do so depends on an SN1 displacement of the
protonated hydroxyl leaving group of the dihydrodiol reactant
occurring from a pseudoaxial orientation in the transition state.
This was proposed by Goering for carbocation-forming reactions
of cyclohexenol derivatives because it maximizes overlap be-
tween the incipient orbital of the carbocation and the adjacent
π-bond.32
As shown in Scheme 12 for reaction of the of the cis isomer of
the dihydrodiol reactant 26, departure of the leaving group in an
axial trajectory leads directly to a conformation of the carboca-
tion 27 in which the CꢀH bond of the β-CHOH group is in a
pseudoaxial position, favorably oriented for hyperconjugation
with the adjacent positive charges. In this conformation the
CꢀOH occupies a pseudoequatorial location. The conformation
is consistent with MP2 calculations for the gas phase, which show
the tetrahedral carbon to be out of the plane of the cyclopenta-
dienyl cation fragment of the ring by 10.5°.
Hyperaromaticity. The term “hyperaromaticity” has been
used in the titles of this and the following paper for two reasons.
First, it emphasizes the importance of a concept which hitherto
has received only limited attention. Second, it captures its key
features by implying an analogy between the relationship of
hyperaromaticity to hyperconjugation and of homoaromaticity
to homoconjugation.
For reaction of the trans isomer 28, the positions of CꢀH and
CꢀOH bonds are reversed. Now it is the C2ꢀOH bond which is
in a pseudoaxial position while the C2ꢀH is pseudoequatorial
(29). We infer that the difference in reactivity arises as a
consequence of the much less favorable hyperconjugating capa-
city of a CꢀOH than CꢀH bond and the fact that the aromatic
character of the hyperconjugation makes the energies of the two
conformations particularly sensitive to this change. In principle,
the experimental results do not distinguish “less favorable
The term “homoaromaticity” was introduced by Winstein35 as
an extension of homoconjugation to apply to structures which,
19725
dx.doi.org/10.1021/ja207160z |J. Am. Chem. Soc. 2011, 133, 19718–19728

Journal of the American Chemical Society
ARTICLE
from their stability and number of electrons involved in homo-
cyclic delocalization, could be inferred to be aromatic. Thus
the relationship between homo (homologous) conjugation and
aromaticity parallels that between ordinary (π) conjugation and
aromaticity. It should be noted that while both homoaromaticity
and hyperaromaticity are conveniently described in terms of no-
bond resonance structures, as illustrated above for the benze-
nium ion 26 and below for cycloheptatriene 32 (both of which
are six-electron pseudoaromatics), the analogy does not imply a
closer relationship than between homo- and hyperconjugation.
In general the latter two concepts have been treated as distinct.
However, we can recognize that in terms of energetic or magnetic
effects, as well as breadth of implications for organic chemistry, a
case can be made that hyperaromaticity is of significance
comparable to if not greater than that of homoaromaticity.25
6.07 (dd, J = 9.8, 2.95 Hz, 1H), 6.54 (dd, J = 9.7, 1.75 Hz, 1H),
7.12ꢀ7.38 (m, 4H); 13C NMR (101 MHz, CDCl3) δ 56.7, 56.8, 76.3,
77.9, 126.9, 127.1, 127.7, 128.1, 128.2, 132.9, 133.0.
cis-1,2-Dimethoxy-1,2-dihydronaphthalene. The preparation was as
described for the trans isomer above. The product was obtained as a
mixure of the two monomethylated products, which were separated by
column chromatography (diethyl etherꢀhexane mixture) to yield the
desired product (0.25 g, 43% from 0.5 g of dihydrodiol, Rf = 0.71, 20%
ethyl acetate in pentane): 1H NMR (CDCl3, 300 MHz) δ 3.38 (s, 3H),
3.48 (s, 3H), 4.21 (m, 1H), 4.28 (d, J = 3.9 Hz, 1H), 6.07 (dd, J = 9.9, 2.0
Hz, 1H), 6.53 (d, J = 9.6 Hz, 1H), 7.12ꢀ7.38 (m, 4H); m/z (EI) 191.2
(M+).
Kinetics and Product Analyses. First-order rate constants kobs
,
for reaction of benzene, naphthalene, and phenanthrene dihydrodiols
measured in aqueous solutions of perchloric or hydrochloric acids, are
recorded as a function of acid concentration in Tables S1ꢀS3. For
benzene and naphthalene cis-dihydrodiols, the measurements were in
dilute solutions of acid, and second-order rate constants were obtained
from the slopes of plots of kobs against acid concentration or the
intercepts of plots of kobs/[H+] against acid concentration. For the
corresponding trans-dihydrodiols and for the cis- and trans-9,10-dihydro-
diols of phenanthrene, the first-order rate rate constants were measured in
concentrated solutions of perchloric acid, and second-order rate constants
were extrapolated to dilute solution as (antilogs of) intercepts of plots of
log(kobs/[H+]) against the acidity parameter Xo.42,43
’ SUMMARY
The greater reactivity of cis- than trans-arene dihydrodiols
toward acid-catalyzed formation of β-hydroxy carbocation inter-
mediates, and the marked dependence of this reactivity differ-
ence upon the aromatic stability of the arene (benzene 4500 >
naphthalene 440 > phenanthrene 50), offers evidence of aroma-
ticity of the carbocations themselves. The difference in reactivity
is suggested to stem from (a) the initial formation of intermedi-
ates (or at least transition states) in which CꢀH and CꢀOH
bonds respectively are located in axial positions with respect to
the carbocation center optimal for hyperconjugation and (b) the
greater hyperconjugating ability of the CꢀH than CꢀOH bond.
As shown in Scheme 5, for the cis-dihydrodiol of phenanthrene kobs
=
kcis[H+]/(1 + kꢀ1/k2), where k1 is the rate constant for formation of the
carbocation 11 and k2 and kꢀ1 correspond to rate constants for reaction
of the carbocation to form phenanthrol 12 and re-form reactant,
respectively. The ratio kꢀ1/k2 was determined as described below by
generating the 9-hydroxyphenanthrenium ion 11 from solvolysis of the
trans-9,10-bromohydrin of phenanthrene or the monotrichloroacetate
ester of the cis-phenanthrene-9,10-dihydrodiol 10 and measuring the
product ratio of phenanthrol 12 to dihydrodiol 13 at sufficiently mild pH
that the dihydrodiol was not converted to the phenol. Based on
formation of 7% cis-phenanthrene dihydrodiol and 92% phenanthrol,
’ EXPERIMENTAL SECTION
k
ꢀ1/k2 = 1.08.
Materials. The cis-1,2-dihydrodiols of benzene and naphthalene
were prepared by biotransformations of the parent aromatic molecules
using whole cell oxidation by a mutant strain (UV4) of the bacterium
Pseudomonas putida.2 They were purchased from the School of Chem-
istry, Queen’s University of Belfast, as solutions in ethyl acetate. The cis-
dihydrodiols of phenanthrene (9,10)36 (10-cis) and acenaphthylene
(6,7)37 (13-cis) and the trans-dihydrodiols of benzene,13 naphthalene
(1,2),12 phenanthrene (9,10)14 (10-trans), and acenaphthylene (6,7)38
(13-trans) were synthesized using literature methods, as were cis-1-
chloro-2-hydroxyacenaphthene39 (17-cis), trans-1-bromo-2-hydroxya-
cenaphthene40 (18-trans), and the monotrichloroacetate ester of cis-
9,10-dihydroxy-9,10-dihydrophenanthrene41 (21). The preparation of
the trans-9,10-bromohydrin of phenanthrene was described in an earlier
paper.1
Because of extrapolation from concentrated acid solutions as plots of
log(kobs/[H+]) against Xo, individual rate constants for phenanthrene
dihydrodiols were subject to appreciable uncertainty, (4.1 ( 1.1) ꢁ
10ꢀ6 Mꢀ1 sꢀ1 for the cis-dihydrodiol and (8.2 ( 2.6) ꢁ 10ꢀ8 Mꢀ1 sꢀ1
for the trans, thus the implied uncertainty in the cis/trans ratio is large.
However, extrapolation of measured cis/trans rate constant ratios of
101, 95, and 85 at 6.0, 5.0, and 4.0 M HClO4 leads to a ratio of 47 ( 1.0,
which after correction for reversibity of dehydration of the cis isomer
becomes 50 ( 1.0.
Acenaphthylene Dihydrodiols. A more complex case is pre-
sented by the equilibration of cis- and trans-acenaphthylene dihydro-
diols, especially because, as shown in Scheme 6, isomerization is
accompanied by irreversible conversion of both isomers to acenaphthe-
none. However, the reactions are conveniently monitored by HPLC
(Tables S9ꢀS12), and a kinetic analysis may be based on Scheme 6
(shown in abbreviated form as Scheme 13), in which k denotes the rate
constant for (the presumed) reaction of the 2-hydroxyacenaphthenium
ion (R+) to form acenaphthenone.
As shown in Figure 1, reaction of the cis-dihydrodiol as reactant leads
to an intital increase in concentration of the trans isomer and then a
decrease as this is converted to acenaphthenone. Relative concentrations
of dihydrodiols and acenaphthenone as a function of time at 5 and 6 M
HClO4 for both cis- and trans-dihydrodiols as reactants are listed in
Tables S7 and S8. The acenaphthenone concentration is corrected for a
3:1 greater extinction coefficient than the dihydrodiols at the wavelength
of analysis.
Cis- and trans-1,2-dimethoxy-1,2-dihydronaphthalenes were pre-
pared as follows.
trans-1,2-Dimethoxy-1,2-dihydronaphthalene. Sodium hydride
(40 mg, 1.4 mmol) was added to a solution of trans-1,2-dihydroxy-
1,2-dihydronaphthalene (0.1 g, 0.61 mmol in DMF, 5 mL). Dimethyl
sulfate (0.18 g, 1.5 mmol) was added slowly over 10 min, and the
mixture was stirred for 20 h at room temperature. The reaction was
quenched with acetic acid (0.1 mL) and diluted with water (20 mL)
followed by extraction with diethyl ether (2 ꢁ 15 mL). After washing
with water and drying (Na2SO4), the ether was removed under reduced
pressure to yield the crude product, which was purified by column chroma-
tography to give a pale yellow liquid (80 mg, 73%): 1H NMR (CDCl3, 300
MHz) δ 3.38 (s, 3H), 3.48 (s, 3H), 4.2 (m, 1H), 4.28 (d, J = 4.36 Hz, 1H),
19726
dx.doi.org/10.1021/ja207160z |J. Am. Chem. Soc. 2011, 133, 19718–19728

Journal of the American Chemical Society
ARTICLE
Scheme 13
Table 6. Rate Constants (Mꢀ1 sꢀ1) from Isomerization of Cis
and Trans Isomers of Acenaphthylene 6,7-Dihydrodiol and
Conversion to Acenaphthenone at 5 and 6 M HClO4 at 25 °C
(with kꢀ1/k2 = 2.3)
[HClO4]
reactant
Keq
kcis/ktrans
kcis (sꢀ1
)
kꢀ1/k3
6 M
6 M
5 M
5 M
cis
2.3
2.2
3.4
2.6
5.3
5.0
7.7
6.0
4.3 ꢁ 10ꢀ4
3.8 ꢁ 10ꢀ4
5.0 ꢁ 10ꢀ5
(9.8 ꢁ 10ꢀ5
2.2
2.5
3.8
trans
cis
The kinetic behavior can be described in terms of a double exponen-
tial dependence of concentration on time for both the cis- and trans-
diols, as shown in eqs 1 and 2. As usual, the coefficients of the expo-
nentials are complicated but can nevertheless be expressed in terms of
the four rate constants ka, kb, kc, and kd in Scheme 7 or the microscopic
rate constants shown in Schemes 6 and 13, namely kcis and ktrans and rate
constant ratios kꢀ1/kꢀ2 and kꢀ1/k3. The relationships between the two
sets of rate constants are ka = kcis/x, kb = ktrans(kꢀ1/kꢀ2)/x, kc = kcis(k3/
trans
)
(7.3)
by concentration of the solution under reduced pressure. Analyses were
carried out on a reverse-phase C18 HPLC column with aqueous
acetonitrile (1:1 v/v) as mobile phase and a flow rate of 0.25ꢀ1 mL/
min or, in the case of the cis-1,2-dihydrodiol, by GC. Retention times and
response factors of products were checked from known samples. Acid-
catalyzed dehydration (or loss of MeOH) occurred more readily at the
2-position than the 1-position. The ratio of rate constants for reaction at
the 1- and 2- positions, kα/kβ, is given by the ratio of concentrations of
the 2- and 1-substituted phenolic or methyl ether products. These
product fractions are recorded in Table S3.
Ratios of cis- to trans-dihydrodiol products formed from solvolyses of
halohydrins of phenanthrene and acenaphthylene and from solvolysis of
the monotrichloroacetate of phenanthrene 9,10-cis-dihydrodiol were
also determined by HPLC. In each case the measurements were
combined with studies of trapping of the presumed β-hydroxy carboca-
tion intermediates in the solvolysis reactions by azide ions. As already
indicated, the azide trapping will be described in a separate paper.25
However, the ratios of cis- to trans-dihydrodiols obtained at different
concentrations of azide ions for the chloro- and bromo-6,7-halohydrins
of acenaphthylene are summarized in Tables S6 and S7.
k
ꢀ2)/x, and kd = ktrans(k3/kꢀ2)/x, where x = 1 + (kꢀ1/kꢀ2) + (k3/kꢀ2).
As explained above, the four unknowns are reduced to three by direct
determination of kꢀ1/kꢀ2 = 2.3 from partitioning of the 7-hydroxyace-
naphthenium ion (R+), generated from aqueous solvolysis of acenaph-
thylene halohydrins, between cis- and trans-dihydrodiol products.
The kinetic dependence of the concentration of the cis isomer may be
described by a best fit of eq 1,
t
t
½cisꢂ ¼ Q eꢀλ þ ðA ꢀ QÞ eꢀλ
ð1Þ
1
2
where Q = (a1 ꢀ (p ꢀ q)/2)A/q, A is the initial concentration of the cis
√
isomer as reactant taken as 100 (%), p = a1 + b2, q = (p2 ꢀ 4 (a1b2 ꢀ
b1a2)), a1 = kcis(1 ꢀ z), a2 = wzkcis, b1 = zktrans, b2 = (1 ꢀ wz)ktrans, z =
1/(1 + w + v), w = kꢀ2/kꢀ1, and v = k3/k
.
ꢀ1
Similarly the kinetic dependence of the trans isomer formed from
reaction of the cis and reacting on to acenaphthenone can be fitted by
eq 2,
’ ASSOCIATED CONTENT
t
t
½transꢂ ¼ P1 eꢀλ þ P2 eꢀλ
ð2Þ
1
2
S
Supporting Information. First-order rate constants for
b
acid-catalyzed dehydration of dihydrodiols (Tables S1ꢀS5);
product ratios of cis- to trans-dihydrodiols from solvolysis of
acenaphthylene halohydrins (Tables S6, S7); HPLC monitoring
of the solvolysis reactions (Table S8); HPLC data (Tables
S9ꢀS12) and rate constants (Table S13) for equilibration and
reaction of cis- and trans-acenaphthylene-6,7-dihydrodiols in 5
and 6 M HClO4; and NMR spectra. This material is available free
of charge via the Internet at http://pubs.acs.org.
where P1 = {a1 ꢀ (p + q)/2}Q/b1 and P2 = {a1 ꢀ (p ꢀ q)/2}(A ꢀ Q)/
b1. The fraction (%) of acenaphthenone is given by 100 ꢀ [cis] ꢀ
[trans]. In practice, the numerical integration program ‘Berkeley Ma-
donna’ used for Scheme 7 proved effective in providing a best fit to
concentrations of the cis- and trans-dihydrodiols and acenaphthenone
siumtaneously. Plots of percent reactant, intermediate, and product
against time for the cis-dihydrodiol reacting in 6 M HClO4 are shown in
Figure 1 and illustrate the quality of fit of calculated to experimental data.
Values of ka, kb, kc, and kd based on Scheme 7 are listed in Table S13.
The derived microscopic rate constants at 6 and 5 M HClO4 for cis-
and trans-dihydrodiol reactants are shown in Table 6. The agreement
between sets of measurements with cis- and trans-dihydrodiol isomers
respectively as reactants is satisfactory at 6 M HClO4 but not at 5 M. In
practice, the slower-reacting trans isomer was monitored to only 50%
reaction in the 5 M acid. The rate constants based on this measurement
are deemed less reliable than for the cis reactant and are shown in
parentheses. However, both measurements are consistent with the
expectation, based on failure to detect acenaphthenone from partition-
ing of the acenaphthenium ion in water in the absence of acid, that
k3/kꢀ1 should decrease with decreasing acid concentration.
’ AUTHOR INFORMATION
Corresponding Author
rmof@ucd.ie
’ ACKNOWLEDGMENT
This work was supported by the Science Foundation Ireland
(Grant No. 04/IN3/B581). We thank Joseph Murdoch for
helpful comments.
Product Analyses. Product analyses for reaction of cis- and trans-
naphthalene-1,2-dihydrodiols (9) and their dimethyl ethers were carried
out by GC or HPLC to determine the fraction of isomeric phenols and
methyl ethers formed from loss of OH or OMe at the 1- and 2-positions
of the reactant (Scheme 4). Typically 7ꢀ10 mg of substrate in 1 mL of
acetonitrile was added to 10 mL of aqueous HClO4. After an appropriate
reaction time (usually 10 half-lives) the acid was neutralized with
aqueous sodium carbonate and extracted with ethyl acetate followed
’ REFERENCES
(1) For a preliminary account of this work see: Kudavalli, J. S.; 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.
(2) Boyd, D. R.; Bugg, T. H. Org. Biomol. Chem. 2006, 4, 181–192.
(3) Boyd, D. R.; Sheldrake, G. N. Nat. Prod. Rep 1998, 15, 309.
19727
dx.doi.org/10.1021/ja207160z |J. Am. Chem. Soc. 2011, 133, 19718–19728

Journal of the American Chemical Society
ARTICLE
(4) Paralis, R. E. Resnick, S. M. In Pseudomonas Vol IV, Molecular
Biology of Emerging Issues; Amos, J.-L., Resnick, S. M., Eds.; Kluwer/
Academic/Plenum: New York, 2005; pp 287ꢀ340.
(5) Hudlicky, T. H.; Reed, J. W. Synlett 2009, 5, 685–703.
(6) 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.
(7) Lawlor, D. A; More O’Ferrall, R. A.; Rao, S. N. J. Am. Chem. Soc.
2008, 130, 17997–8007.
(42) Bagno, A.; Scorrano, G.; More O’Ferrall, R. A. Rev. Chem.
Intermed. 1987, 7, 313–352.
(43) Cox, R. A. Adv. Phys. Org. Chem. 2000, 35, 1–66.
(8) Rao, S. N.; More O’Ferrall, R. A.; Kelly, S. C.; Boyd, D. R.;
Agarwal, R. J. Am. Chem. Soc. 1993, 115, 5458–5465.
(9) Pirinccioglu, N.; Thibblin, A. J. Am. Chem. Soc. 1998, 120,
6512–6517.
(10) Boyd, D. R.; McMordie, R. A. S.; Sharma, N. D.; More
O’Ferrall, R. A.; Kelly, S. C. J. Am. Chem. Soc. 1990, 112, 7822–7823.
(11) Nashed, N. T.; Bax, R. J.; Loncharich, R. J.; Sayer, J. M.; Jerina,
D. M. J. Am. Chem. Soc. 1993, 115, 1711–1722.
(12) Platt, K. L.; Oesch, F. J. Org. Chem. 1983, 48, 265–268.
(13) Platt, K. L.; Oesch, F. Synthesis 1977, 449–450.
(14) Platt, K. L.; Oesch, F. Synthesis 1982, 459–461.
(15) Unpublished MP2 calculations by J. R. Keeffe indeed show the
benzene-cis-1,2-dihydrodiol to be more stable than its trans isomer in the
gas phase. However, this reflects intramolecular hydrogen bonding,
which is unlikely to persist in aqueous solution.
(16) Sampson, K.; Paik, A.; Duvall, B; Whalen, D. L. J. Org. Chem.
2004, 69, 5204–5211.
(17) Doan, L.; Lin, B.; Yagi, H.; Jerina, D. M.; Whalen, D. L. J. Am.
Chem. Soc. 2001, 123, 6785–6791.
(18) Whalen, D. L. Adv. Phys. Org. Chem. 2005, 40, 247–298.
(19) Gillilan, R. E.; Pohl, T. M.; Whalen, D. L. J. Am. Chem. Soc.
1982, 104, 4481–4482.
(20) Doan, L.; Whalen, D. L. J. Org. Chem. 2006, 71, 7252–7260.
(21) Kudavalli, J. S.; Lawlor, D. A.; More O’Ferrall, R. A., to be
published.
(22) Pickett, L. W.; Muller, N.; Mulliken, R. S. J. Chem. Phys. 1953,
21, 1400–1401.
(23) Ermier, W. C.; Mulliken, R. S.; Clement, E. J. Am. Chem. Soc.
1976, 98, 388–394.
(24) Sieber, S.; Schleyer, P. v. R.; Gauss, J. J. Am. Chem. Soc. 1993,
115, 6987–6988.
(25) 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, DOI: 10.1021/ja2071626 (following paper in this issue).
(26) Whalen, D. L.; Ross, A. M.; Dansette, P. M.; Jerina, D. M. J. Am.
Chem. Soc. 1977, 99, 5672–5676.
(27) Zhao, C.; Whalen, D. L. Chem. Res. Toxicol. 2006, 19, 217–222.
(28) Bruice, P. Y.; Bruice, T. C.; Dansette, P. M.; Selander, H. G.;
Yagi, H.; Jerina, D. M. J. Am. Chem. Soc. 1976, 98, 2965–2973.
(29) Doan, L; Whalen, D. L. J. Org. Chem. 2006, 71, 7252–7260.
(30) Doan, L; Yagi, H.; Jerina, D. M.; Whalen, D. L. J. Org. Chem.
2004, 69, 8012–8017.
(31) Doan, L; Yagi, H.; Jerina, D. M.; Whalen, D. L. J. Am. Chem. Soc.
2002, 124, 14382–14387.
(32) Goering, H. L.; Josephson, R. R. J. Am. Chem. Soc. 1962,
84, 2779–2785.
(33) Winstein, S.; Trifan, D. J. Am. Chem. Soc. 1952, 74, 1147–1154.
(34) Walling, C. Acc. Chem. Res. 1983, 16, 448–454.
(35) Winstein, S. In Carbonium Ions; Olah, G. A., Schleyer, P. v. R.,
Eds.; Wiley: New York, 1972; Vol. III.
(36) Shing, T. K.; Tai, V. W.-F.; Tam, E. K. W. Angew. Chem., Int. Ed.
1994, 33, 2312–2313.
(37) Jack, K. M.; Rule, G. H. J. Chem. Soc. 1938, 188–192.
(38) Hyward, L. D.; Czismidia, I. G. Tetrahedron 1963, 19, 2111–2121.
(39) De Luca, L.; Giacomelli, G.; Porcheddu, A. Org. Lett. 2002,
4, 553–555.
(40) Ghosh, A. K.; Kim, J.-H. Org. Lett. 2004, 5, 1063–1066.
(41) Clarke, P. A.; Holton, R. A.; Kayaleh, N. E. Tetrahedron Lett.
2000, 41, 2687–2690.
19728
dx.doi.org/10.1021/ja207160z |J. Am. Chem. Soc. 2011, 133, 19718–19728