Mechanisms of hydrolysis and rearrangements

of epoxides

DALE L. WHALEN

Department of Chemistry and Biochemistry, University of Maryland, Baltimore

County (UMBC), Baltimore, MD, USA

1 Introduction 248
2 Limiting mechanisms of epoxide reactions 248
3 Mechanisms of hydrolysis of epoxides derived from simple alkenes and cycloalkenes 250
Kinetic studies 250
Acid-catalyzed hydrolysis of aliphatic epoxides 251
pH-independent and hydroxide ion-catalyzed hydrolyses of aliphatic primary and
secondary epoxides 254
4 Acid-catalyzed hydrolyses of alkyl- and vinyl-substituted epoxides 254
Relative reactivities 254
Simple vinyl epoxides 255
Cyclic vinyl epoxides 257
5 Mechanisms of hydrolysis of styrene oxides 258
Acid-catalyzed hydrolysis of styrene oxides 258
Substituent effects on addition of amines and hydroxide ion to styrene oxides 262
Substituent effects on the pH-independent reactions of styrene oxides 263
6 Mechanisms of acid-catalyzed hydrolysis of 1-phenylcyclohexene oxides, indene oxides
and 1,2,3,4-tetrahydronaphthalene-1,2-epoxides 264
Acid-catalyzed hydrolysis of 1-phenylcyclohexene oxides 264
Acid-catalyzed hydrolysis of indene oxides: transition-state effects on stereochemistry
of diol formation 266
Tetrahydronaphthalene epoxide hydrolysis: conformational effects on stereochemistry
of diol formation 267
7 General acid catalysis in epoxide reactions 271
Ethylene oxide and simple primary and secondary epoxides 271
Tertiary epoxides 272
Acetals and epoxy ethers 272
Vinyl epoxides 273
Benzylic epoxides and arene oxides 274
8 pH-independent reactions of epoxides 277
Simple alkyl epoxides 277
Arene oxides 277
Cyclic vinyl epoxides 279
Benzylic epoxides that undergo rate-limiting 1,2-hydrogen migration 280
Benzo[a]pyrene 7,8-diol 9,10-epoxides 281
Summary of pH-independent mechanisms 283
9 Epoxide isomerization accompanying pH-independent reactions 283
10 Benzylic epoxides that exhibit complicated pH-rate profiles 286
Precocene I 3,4-oxide 286
Benzo[a]pyrene 7,8-diol 9,10-epoxide (80) 288
Specific effects of chloride ion in epoxide hydrolysis 290

247
ADVANCES IN PHYSICAL ORGANIC CHEMISTRY r 2005 Elsevier B.V.
VOLUME 40 ISSN 0065-3160 DOI: 10.1016/S0065-3160(05)40006-4 All rights reserved
248 D.L. WHALEN

11 Partitioning of hydroxycarbocations 291

12 Overall summary 294
Acknowledgments 294
References 295

1 Introduction

The epoxide functional group has played very important and versatile roles in or-
ganic synthesis, and a number of early review articles on their reactions have been
published.1,2 Numerous natural products contain the epoxide group,3–7 and the
biological activities of certain compounds are linked to the facile reactivity and
unique reactions of this group.
A large number of nonpolar, unsaturated compounds are metabolized by path-
ways that involve epoxidation of the unsaturated substance by the cytochrome P-
450 enzymes, followed by epoxide hydrolase-catalyzed hydrolysis or glutathione
transferase-catalyzed addition of glutathione to the epoxide.8,9 The resulting meta-
bolites are then secreted as glucuronide or glutathione derivatives. Although these
metabolic pathways serve as a detoxification mechanism for some compounds, the
intermediate epoxide metabolites of many unsaturated materials undergo covalent
binding with DNA, proteins and other biological macromolecules. The reactions of
compounds containing the epoxide functional group with biological molecules hav-
ing acidic and nucleophilic groups are critical steps in the initiation of biosynthetic
pathways that on the one hand may lead to the elimination of exogenous chemicals,
but on the other hand may cause mutations that lead to disease and other detri-
mental physiological effects.10–18
An understanding of the mechanisms by which epoxides react with acidic and
nucleophilic species is therefore of special importance. The purpose of this chapter is
not to provide an extensive survey of the literature of epoxide reactions, but rather
to summarize how pH, concentrations of added acidic and nucleophilic reagents,
and epoxide structure determine the mechanisms by which epoxides undergo reac-
tions in aqueous solutions. A review of the hydrolysis reactions of simple epoxides is
presented, followed by discussions of the effects of charge-stabilizing groups such as
vinyl and aryl on the mechanisms of epoxide reactions. Special attention is given to
the lifetimes of carbocation intermediates formed from the reactions of epoxides
with acidic reagents, and to the roles that carbocation conformation and transition-
state effects play in determining the stereochemistry of the products formed from
reaction of solvent with these carbocations.

2 Limiting mechanisms of epoxide reactions

The epoxide (oxirane) functional group has special features that allow it to play very
important roles in synthesis and as an electrophilic species in biologically reactive
intermediates. The epoxide group has an estimated strain energy of 27 kcal mol
1,19
and its reactions with both acidic and nucleophilic reagents result in epoxide ring
MECHANISMS OF HYDROLYSIS AND REARRANGEMENTS OF EPOXIDES 249

opening and the release of strain energy. The strain energy of an epoxide group,
however, is very similar to that of an oxetane. Yet epoxides are much more reactive
than oxetanes toward acidic and nucleophilic reagents. The greater reactivity of a
three-membered ring compared to a four-membered ring has been rationalized in
several ways. Houk and co-workers proposed that the transition state for opening of
a three-membered ring has aromatic stabilization.20 Hoz and co-workers attributed
the greater reactivity of three-membered rings, including the epoxide group, to
‘‘earlier’’ transition-state structures in which the strain energy is lost faster than in
ring-opening reactions of four-membered rings.21
In an acidic environment, the epoxide oxygen acts as a base, and several reaction
pathways are available that result in cleavage of a C–O bond and release of the
strain energy of the epoxide ring. One possible pathway involves protonation of the
epoxide oxygen by an acid HA, followed by C–O bond cleavage to yield a reactive
carbocation that undergoes subsequent reaction with nucleophilic reagents (Scheme
1). For this to occur, the carbocation must be stabilized sufficiently by adjacent
groups to exist as an intermediate. If the adjacent groups do not stabilize a
carbocation sufficiently, then alternate pathways in which either a nucleophile

H
HO R
O +O HO −
R HA R Nuc
R R + R R
R R
R
R R R R R R Nuc

Scheme 1.

O +O HO R
R HA R R
R R
R
R R R R Nuc
R
−
Nuc

Scheme 2.

− HO
O O R R
R R HA R
R
R R
R R R Nuc R Nuc
−
Nuc

Scheme 3.
250 D.L. WHALEN

attacks a protonated epoxide (Scheme 2) or proton transfer concerted with

nucleophile addition are followed. In basic solution, either nucleophile addition to
the epoxide group precedes addition of a proton by a weakly acidic solvent or other
molecule HA (Scheme 3), or nucleophile addition concerted with proton donation
by a weakly acidic group occurs. In an aqueous solution, the actual mechanism of
reaction of a given epoxide will depend on the pH of the solution, the structure of
the epoxide, and the concentrations of any added acidic and nucleophilic reagents.

3 Mechanisms of hydrolysis of epoxides derived from simple

alkenes and cycloalkenes

KINETIC STUDIES

Early mechanistic studies by Brønsted et al. of the reactions of ethylene oxide,

glycidol and glycidyl chloride in aqueous HClO4 solutions established that there are
two kinetically distinct hydrolysis reactions.22 These two reactions for ethylene oxide
are given by Equations (a) and (b) in Scheme 4, and correspond to hydronium
ion-catalyzed hydrolysis and ‘‘pH-independent’’ hydrolysis, respectively. Upon sub-
stitution of HCl for HClO4, two additional kinetically distinct mechanisms are ob-
served and are given by Equations (c) and (d) in Scheme 4.
Kinetic and isotopic labeling studies by Long and Pritchard on the hydrolyses of
ethylene oxide, propylene oxide and isobutylene oxide to their corresponding glycols
provide additional insight on the hydrolysis mechanisms of simple aliphatic epox-
ides.23 In each case, three kinetically distinct mechanisms for their reactions in water
solutions were observed: acid-catalyzed hydrolysis, pH-independent hydrolysis and
hydroxide-catalyzed hydrolysis. Thus, the rate law for hydrolyses of these simple

O H3O+
+ H2O HOCH2CH2OH
(a)

O
+ H2O HOCH2CH2OH
(b)

O
+ Cl + H3O+ HOCH2CH2Cl + H2O
(c)

O
+ Cl + H2O HOCH2CH2Cl + OH
(d)

Scheme 4.
MECHANISMS OF HYDROLYSIS AND REARRANGEMENTS OF EPOXIDES 251

Fig. 1 Observed first-order rate coefficient, k1 (s
1), for hydrolysis of isobutylene oxide at
25 1C plotted against
log C(H+). Reproduced with permission from Refs 23 and 24. Copy-
right 1956 American Chemical Society.

aliphatic epoxides is given by Equation (1), where kH is the bimolecular rate con-
stant for the acid-catalyzed hydrolysis, ko is the first-order rate constant for the pH-
independent reaction and kOH is the bimolecular rate constant for the hydroxide-
catalyzed hydrolysis:
kobsd ¼ kH ½Hþ  þ ko þ kOH ½HO
 (1)
The rate expression given in Equation (1) also applies to the hydrolyses of many
other epoxides, and the shape of the pH-rate profile for a given epoxide will reflect
the relative magnitudes of kH, ko and kOH. In the hydrolyses of some epoxides, ko is
small relative to kH[H+] and kOH[HO
], and cannot be easily detected. In other
cases, ko is very large relative to kOH[HO
], and consequently hydroxide ion-
catalyzed epoxide hydrolysis may not be detected. The pH-rate profile for hydrolysis
of isobutylene oxide is given in Fig. 1. For this compound, acid-catalyzed hydrolysis
predominates at pHo  6, the pH-independent reaction predominates at pH8–11,
and hydroxide ion-catalyzed hydrolysis predominates at pH4  12.

ACID-CATALYZED HYDROLYSIS OF ALIPHATIC EPOXIDES

Simple primary and secondary epoxides

Several conclusions are drawn from the product studies of the reactions of propylene
oxide and isobutylene oxide in H18
2 O.
23
In the acid-catalyzed hydrolysis of propylene
oxide, cleavage of the secondary C–O bond is favored by a factor of 2–3 over
cleavage of the primary C–O bond. In the acid-catalyzed hydrolysis of isobutylene
oxide, cleavage of the tertiary C–O bond is highly favored (499%). These results
show that in acid-catalyzed epoxide hydrolysis, cleavage of the C–O bond leading to
252 D.L. WHALEN

H H
+ r.d.s. O H2O OH
O O
+ H+ +
H+
OH
A-1 Mechanism

H OH
+O OH
r.d.s. H+
O
+ H+
OH OH
+ 2
OH2

A-2 Mechanism

Scheme 5.

the carbon that can better stabilize positive charge is favored. For a number of
simple alkyl-substituted epoxides, log kobsd for acid-catalyzed hydrolysis correlates
more closely with H0, the Zucker–Hammett acidity function, than with log [H+].24
On the basis of the Zucker–Hammett postulate, it was proposed that the activated
complex does not contain a water molecule, thus supporting the A-1 mechanism
involving a carbocation intermediate for acid-catalyzed hydrolysis of these epoxides,
including ethylene oxide (Scheme 5). However, the validity of the Zucker–Hammett
hypothesis was later brought into question,25,26 and an A-2 mechanism in which
there is concerted addition of a water molecule to the protonated epoxide was
suggested for the acid-catalyzed hydrolysis of epoxides containing only primary and
secondary carbons.27
Simple tertiary carbocations react with water/trifluoroethanol (H2O/TFE) solvent
with an estimated rate constant of 1012 s
1, which is somewhat faster than bulk
solvent reorganizes.28 Simple primary and simple secondary carbocations are pre-
dicted to be even more reactive with nucleophilic solvent than tertiary carbocations.
In water solutions, therefore, the rates at which simple primary and secondary
carbocations are predicted to react with solvent would exceed the rate of a bond
vibration (1013 s
1), and consequently they would not be sufficiently stable to exist
as an intermediate. An A-2 mechanism would therefore be enforced on the acid-
catalyzed hydrolysis of those epoxides that potentially undergo ring opening to form
primary or secondary carbocations.

Simple tertiary epoxides

The estimated rate at which a simple tertiary carbocation reacts with water solvent
(1012 s
1)28 is very close to the rate at which solvent reorganizes (1011–1012 s
1).29
It has been suggested that simple tertiary carbocations that are formed as part of ion
pairs or ion-molecule pairs react with a solvent molecule that is already present
within the solvent shell that is present at the time of formation of the carbocation.30
MECHANISMS OF HYDROLYSIS AND REARRANGEMENTS OF EPOXIDES 253

In the solvolysis of tertiary substrates in which there is a negative or neutral

leaving group other than solvent, the leaving group will shield one side of the
reaction center, and therefore attack of a solvent molecule leading to inversion of
configuration at the reaction center is more likely to occur. In the ionization of
RX to form a carbocation Rþ and X
, the first step is the heterolytic cleavage of the
RX bond to form an ion pair (Rþ X
) or ion-molecule pair. For retention of
configuration at carbon to be observed, the lifetime of the ion pair or ion-molecule
pair would have to be long enough for X
to be replaced by a water molecule.
However, C–O bond cleavage of a protonated epoxide yields a hydroxycarbocation
that is completely solvated, without a counter ion or molecule, and if the carbo-
cation has a significant lifetime, then some retention of configuration might be
expected.
Several pieces of information suggesting an A-1 mechanism for the acid-catalyzed
hydrolysis of isobutylene oxide are (1) isobutylene oxide is several orders of mag-
nitude more reactive than propylene oxide and (2) the entropy of activation
for hydrolysis of isobutylene oxide is 13 units more positive that the entropies of
activation for acid-catalyzed hydrolysis of propylene oxide and epichlorihydrin.31
However, the acid-catalyzed hydrolyses of tertiary epoxide 132 and di-tertiary
epoxides 2 and 333 are reported to proceed with complete inversion of configuration
at carbon to yield only trans diols, which suggest an A-2 mechanism. Since the
lifetime of a simple tertiary carbocation is very similar to the time required for
solvent reorganization, epoxide ring opening may be nearly complete at the trans-
ition state, with little bond formation from the solvent to the carbocation.
A mechanism in which proton transfer from hydronium ion to the epoxide oxygen
and epoxide O–C bond breaking are more complete than bond formation between
solvent and carbon has been referred to as ‘‘borderline SN2.’’34 Although complete
inversion at carbon in the acid-catalyzed hydrolyses of 1–3 is predicted if there is
some bond formation between solvent and carbon at the transition state, complete
inversion might also be observed if carbocation formation is complete, and solvent
collapses from within a solvent shell before solvent reorganization or conforma-
tional change occurs.

H
CH3 CH3 CH3
O O O
CH3 H CH3
1 2 3

- ‡ - ‡
Oδ Oδ
H H
H H H H
CH3 δ- CH3 δ + OH
OH 2

4 5
254 D.L. WHALEN

PH-INDEPENDENT AND HYDROXIDE ION-CATALYZED HYDROLYSES OF ALIPHATIC

PRIMARY AND SECONDARY EPOXIDES

In the second-order reaction of propylene oxide with sodium hydroxide in water

solution, hydroxide ion attack at the primary carbon (80%) is favored over attack
at the secondary carbon (20%).23 The ratio of attack by hydroxide ion at the
primary versus secondary carbons in hydroxide-catalyzed epoxide hydrolysis (4) is
therefore lower than the ratio of second-order rate constants for displacement
of bromide by labeled bromide ion in the reactions of propyl and isopropyl brom-
ides in aqueous acetone (15).35 In the pH-independent reaction of propylene oxide,
solvent also adds preferentially to the primary carbon, but to a slightly lesser extent
of 60–70%.23 The pH-independent reactions and the second-order reactions of
simple primary and secondary epoxides with hydroxide ion, which yield diol prod-
ucts, can be regarded as following very similar mechanistic pathways in which
epoxide C–O bond breaking is concerted with nucleophile O–C bond making.
Transition states for attack of hydroxide ion and for attack of water at the primary
carbon of propylene oxide are given by structures 4 and 5. The newly developing
oxyanion is presumably stabilized by strong hydrogen bonding with the solvent, or
perhaps proton addition to the oxyanion is concerted with epoxide C–O bond
breaking and nucleophile O–C bond making. Reactions involving nucleophilic at-
tacks of hydroxide ion or water on simple tertiary epoxide carbons have not been
characterized.

4 Acid-catalyzed hydrolyses of alkyl- and vinyl-substituted

epoxides

RELATIVE REACTIVITIES

A summary of bimolecular rate constants for the acid-catalyzed hydrolysis of a

series of alkyl-, vinyl- and phenyl-substituted epoxides is given in Table 1. Propylene
oxide (7) is 6.6 times more reactive than ethylene oxide, and from a study of its
reaction in H2O18, it was shown that 70% of the glycol product results from
addition of solvent to the secondary carbon and 30% from addition of solvent to
the primary carbon. The reactivity per primary carbon of ethylene oxide is one-
half of the observed reactivity of ethylene oxide, and thus the introduction of a
methyl group on ethylene oxide results in an increase in reactivity at the primary
carbon by a factor of 4 and an increase in reactivity at the secondary carbon by a
factor of 9. These results are consistent with A-2 mechanisms for the acid-
catalyzed hydrolyses of ethylene oxide and propylene oxide, in which some amount
of positive charge generated on carbon at the transition state is stabilized by a
methyl group.
MECHANISMS OF HYDROLYSIS AND REARRANGEMENTS OF EPOXIDES 255

Table 1 Bimolecular rate constants for hydronium ion-catalyzed hydrolysis of a series of

epoxides in water, 25 1C

Compound kH (M
1 s
1) Compound kH (M
1 s
1)

O 9  10
3a O 6  10
2a

6 7

O 6.8a 2.7b

8
9

O 1.7b 3.6b
O

10
11

O 3.7  103b O 1.1  104b

12
13

a
Taken from Ref. 23.
b
Taken from Ref. 36.

SIMPLE VINYL EPOXIDES

Acid-catalyzed hydrolysis of isobutylene oxide (8) is 4750 times faster than that of
ethylene oxide (6), and 499% of the glycol product is from addition of solvent at
the tertiary carbon.23 These results are consistent with a mechanism in which there is
significant positive charge on the tertiary carbon at the transition state, as discussed
in the previous section. Butadiene monoepoxide (10) is slightly less reactive than
isobutylene oxide,36 and its acid-catalyzed hydrolysis can potentially proceed via a
resonance-stabilized allyl cation (Scheme 6). However, the acid-catalyzed hydrolysis
of 10 yields 96% of 3-buten-1,2-diol (15) and only 4% of 2-butene-1,4-diol (16),36
and the acid-catalyzed methanolysis of 10 is reported to yield only 2-methoxy-3-
buten-1-ol.37 An A-2 mechanism proceeding via transition state 17 may account for
the observation that 1,2-diol 15 is the predominant product from acid-catalyzed
hydrolysis of 10. The minor yield of the 1,4-diol 16 may be formed from reaction of
256 D.L. WHALEN

OH
O OH OH
H+ H2O
+ +
A-1
HO
10 14 15 HO 16

H+

‡
H
O+ OH
-H+
15
A-2 +
H2O
OH2
17 18

Scheme 6.

CH3CR CH CH2Cl CH3CR CH CH2OH

H
19a, R=H 22a, R=H
b, R=CH3 H2O/Ag2O C H2O b, R=CH3
CH3CR + CH2
CH3CRClCH CH2 CH3CR(OH)CH CH2
20a, R=H 21a, R=H 23a, R=H
b, R=CH3 b, R=CH3 b, R=CH3

Scheme 7.

allylic carbocation 14 with solvent, although a concerted reaction for formation of

this diol from attack of water on 17 is also possible.
It is interesting to compare the hydrolysis of butadiene oxide (10) with those of 1-
chloro-2-butene and 3-chloro-1-butene.38 1-Chloro-2-butene (19a) solvolyzes in wa-
ter to yield 45% of 2-butene-1-ol (22a) and 55% of 3-butene-2-ol (23a) (Scheme 7).
3-Chloro-1-butene (20a) also solvolyzes to yield alcohols 22a and 23a, but in a
different ratio (34%:66%, respectively). The lifetime of the allylic carbocation in-
termediate 21a is therefore not sufficient to allow dissociation of chloride ion and
equilibration of solvent about the carbocation. However, hydrolyses of 19b and 20b
(R ¼ CH3 instead of H) yield identical ratios of allylic alcohol products 22b and 23b
(15%:85%, respectively).38 Allylic cation 21b, with a second methyl group to sta-
bilize positive charge, must have a longer lifetime, which allows departure of the
leaving group and equilibration of solvent about the carbocation.
Whereas the hydrolysis of 1-chloro-2-butene gives comparable yields of products
from attack of solvent at both primary and secondary carbons, very little product
from the acid-catalyzed hydrolysis of butadiene oxide 10 is formed from the attack
of solvent at the primary carbon. The transition state for acid-catalyzed epoxide ring
opening has a relatively reactant-like geometry, and therefore positive charge de-
localization into the adjacent double bond at the transition state is expected to be
less than that for allyl chloride hydrolysis.
MECHANISMS OF HYDROLYSIS AND REARRANGEMENTS OF EPOXIDES 257

CYCLIC VINYL EPOXIDES

The mechanisms and rates of acid-catalyzed vinyl epoxide hydrolysis depend mark-
edly on whether the vinyl group is part of a ring system. When the epoxide group
and vinyl group are both part of a five- or six-membered ring, the rate of acid-
catalyzed hydrolysis (Table 1) is increased dramatically, and both 1,2- and 1,4-diols
are formed as products.36 The bimolecular rate constant kH for acid-catalyzed hy-
drolysis of cyclopentadiene oxide (12) and cyclohexadiene oxide (13) are 2:2  103
times larger and 6:5  103 times larger, respectively, than that for acid-catalyzed
hydrolysis of butadiene monoxide (10). Acid-catalyzed hydrolysis of cyclopenta-
diene oxide (12) yields major amounts of both cis and trans 1,2- and 1,4-diols
(Scheme 8).36b The presence of all four diols in the product mixture provides strong
evidence that a discrete cyclopentadienyl carbocation 24 is an intermediate, and that
the reaction proceeds via an A-1 mechanism.
Cyclohexadiene oxide (13) is even more reactive toward acid-catalyzed hydrolysis
than cyclopentadiene oxide (12), and comparable yields of both 3-cyclohexene-trans-
1,2-diol (30) (60%) and 2-cyclohexene-trans-1,4-diol (31) (36%) are formed (Scheme
9), along with minor amounts of cis 1,2- and cis 1,4-diols (total 4%).36a,c As a
significant yield of trans 1,4-diol 31 is formed, it was suggested that 13 also hy-
drolyzes via the A-1 mechanism (Scheme 9).36 The cyclohexenyl carbocation 29
should have a lifetime in water comparable to that of the cyclopentenyl carbocation
24. The greater stabilities of the cyclic allylic carbocations 24 and 29 compared to
that of acyclic allylic carbocation 14 are partly due to the stabilizing effect of sub-
stitution: both of the terminal carbons in cyclic allylic carbocations 24 and 29 are
secondary, whereas one terminal carbon atom in allylic carbocation 14 is primary
and one is secondary. The observation that cyclohexadiene oxide yields mostly trans
1,2- and trans 1,4-diols and very low yields of cis 1,2- and cis 1,4-diols is attributed to
conformational factors36 that will be discussed separately. Cycloheptadiene oxide

OH OH OH OH OH
O
H+ H 2O OH OH
+ + +
+ -H+
HO HO
12 24 25 (25%) 26 (16%) 27 (16%) 28 (43%)

Scheme 8.

OH OH HO
O
H+ H2O OH
+
+ H+

OH
13 29 30 31

Scheme 9.
258 D.L. WHALEN

and cyclooctadiene oxide (11) also undergo acid-catalyzed hydrolysis to give sub-
stantial yields of cis and trans 1,2- and 1,4-diols, suggesting that they, too, react via
A-1 mechanisms. The low reactivity of 11 is rationalized by the fact that the allylic
C–O bond in each of its two principal conformations is twisted substantially with
respect to the p-orbitals of the C–C double bond,39 and substantial strain is de-
veloped in the transition state for the formation of the allylic carbocation in which
there is maximum overlap of the newly developing p-orbital with the C–C p-bond.

5 Mechanisms of hydrolysis of styrene oxides

ACID-CATALYZED HYDROLYSIS OF STYRENE OXIDES

Acid-catalyzed hydrolysis of styrene oxide and of its cis and trans b-methyl deriv-
atives in H182 O are reported to yield glycol products in which all of the
18
O label
40a
is located in the benzyl positions. Also, the acid-catalyzed methanolysis of p-
nitrostyrene oxide yields a mixture of hydroxy ethers in which the major isomer
(95%) is the result of benzyl C–O bond cleavage.40b,c Thus, an aryl group, even the
electron-withdrawing p-nitrophenyl group, stabilizes the transition state in which
positive charge is developed at the benzyl carbon. The rates of acid-catalyzed hy-
drolysis and methanolysis of substituted styrene oxides give good Hammett corre-
lations, yielding a rþ value for hydrolysis of
4.241 and a r value for methanolysis
of
4.1.42 Thus, significant positive charge is developed on the benzylic carbon at
the transition state for both hydrolysis and methanolysis. For comparison, rþ values
for the solvolysis of 1-phenylethyl halides and esters in aqueous organic solvents
range from
5 to
6,43,44 indicating a greater extent of carbocation formation at the
transition state for solvolysis reactions.
In a study by Berti et al., acid-catalyzed hydrolysis of styrene oxide was reported
to occur with 67% inversion and 33% retention at the benzyl carbon.45 In a later
study, it was reported that the styrene glycol product formed in the acid-catalyzed
hydrolysis of chiral styrene oxide is completely racemic, which would indicate an
A-1 mechanism.46 As these two results indicate quite different mechanisms for this
reaction, the glycol product from acid-catalyzed hydrolysis of chiral styrene oxide
was converted to its bis-(+)-a-(methoxy-a-trifluoromethyl)phenylacetate diester de-
rivative, and the composition of the diastereomeric diester mixture was determined
by 1H NMR.47 This study agreed with those of Berti et al. and showed that acid-
catalyzed hydrolysis of styrene oxide occurs with 67% inversion and 33% retention
at the benzyl carbon. Acid-catalyzed methanolysis of styrene oxide is reported to
occur with 89% inversion at the benzyl carbon.48 The fact that the diol product from
acid-catalyzed hydrolysis of chiral styrene oxide is not completely racemic demon-
strates that the lifetime of the carbocation is not sufficiently long for it to become
symmetrically solvated.
A mechanistic scheme for the acid-catalyzed hydrolysis of styrene oxide (32) is
given in Scheme 10. If reaction of H+ with 32 occurs with benzyl C–O bond
MECHANISMS OF HYDROLYSIS AND REARRANGEMENTS OF EPOXIDES 259

H + H
OH H2O
H OH
O OH
H H H + 34 CH CH2OH
α β H+ α β H
H H H2O
H 36 (racemic)
32 33 +
H OH
H2O ?
H
32 (mostly inverted) 35

H
+
δ OH
O H
H+ H H
+ H
H2O δ H
OH2 H
HO
37 36 (inverted)

Scheme 10.

breaking, but without rotation about the Ca–Cb bond, then a carbocation with
structure 33 will be formed. Quantum chemical calculations of the 2-hydroxy-1-
phenylethyl carbocation in the gas phase at the MP2/6-31G* level of theory indicate
that structure 33 collapses without activation to either structure 34 or 35.49 Structure
34 is calculated to be 3.9 kcal mol
1 more stable than structure 35 in the gas phase.
After inclusion of a solvation energy term, structure 34 is still calculated to be
2.3 kcal mol
1 more stable than structure 35. Structure 33, in which the C–O bond is
constrained to be co-linear with the empty p-orbital on the benzylic carbon, is
calculated to be 6.2 kcal mol
1 higher in energy than structure 34. Rotation of the
Ca–Cb bond of the initially formed carbocation structure 33 in one direction will
yield structure 34, and rotation of this bond in the opposite direction will yield
structure 35. Conformers 34 and 35 possess planes of symmetry, however, and will
react with solvent to yield completely racemic diol.
When the stability of the benzylic carbocation 33 is such that the rate of solvent
attack on the carbocation is similar to the rate of rotation about the Ca–Cb bond of
33, then reaction pathways other than fully concerted attack of solvent at the ben-
zylic carbon accompanied with epoxide C–O bond breaking (A-2 mechanism) and
attack of solvent on a symmetrically solvated carbocation intermediate (A-1 mech-
anism) become possible. The rate constant for reaction in 50:50 TFE/H2O solvent of
the 1-phenylethyl carbocation, which is similar in structure to the 2-hydroxy-1-
phenylethyl carbocation, is estimated to be approximately 1011 s
1.43 Reorganiza-
tion of water solvent is on the order of 1011–1012 s
1,29 and therefore an unstable
species such as 33 might react with water to give either retained or inverted diol
product at a rate that is comparable to the rate at which it undergoes Ca–Cb bond
rotation to form 34 or 35. In one mechanism, a water molecule that is hydrogen
260 D.L. WHALEN

bonded to the newly formed hydroxyl group may be in position to collapse with the
carbocationic center in an encounter mechanism to yield diol with retention of
configuration at the benzyl carbon. However, attack of water from the side of the
carbocationic center opposite to the newly formed hydroxyl group would involve
fewer nonbonding interactions at the transition state, and this reaction would lead to
inversion of configuration.
In a comparison of the rates of acid-catalyzed racemization of and 18O exchange
in chiral 1-phenylethanol, it was concluded that the departing –OH2 group shields
that side of the benzyl carbocation from attack by other solvent molecules.26 The
lifetime of the b-phenylethyl carbocation, which should be similar to that of the b-
hydroxy benzylic carbocation formed in the acid-catalyzed hydrolysis of styrene
oxide 32, is therefore too short to allow complete solvent equilibration about the
carbocation.
The benzylic carbocation formed from the reaction of styrene oxide with H+, like
the 1-phenylethyl carbocation and simple tertiary carbocations, most likely reacts
with water molecules from within the inner solvent shell that are present when
the carbocation is formed.30 The lifetime of the carbocation formed in the acid-
catalyzed hydrolysis of styrene oxide must be similar to the time required for solvent
relaxation and rotation about the Ca–Cb bond.
The lifetimes of 2-hydroxy-1-phenylethyl carbocations in which there are sub-
stituents in the phenyl ring are expected to vary with the ability of the substituent to
stabilize positive charge at the benzyl carbon, and therefore the mechanism of acid-
catalyzed hydrolysis will depend on the substituent. For example, a methoxy group
substituted at the para position of the phenyl ring in 33 is expected to stabilize the
carbocation and increase its lifetime, whereas a nitro group substituted at the para
position of the phenyl ring is expected to destabilize the carbocation and decrease its
lifetime. If the p-nitro group destabilizes charge at the benzyl carbon sufficiently,
then addition of water at the benzyl carbon will be concerted with benzyl C–O bond
cleavage (A-2 mechanism). Such a range in mechanism has been observed in the
solvolysis of ring-substituted 1-phenylethyl derivatives in 50:50 H2O/TFE solution,
where strongly electron-donating substituents stabilize benzyl carbocations suffi-
ciently so that the reaction occurs via a carbocation intermediate (SN1 mechanism),
and strongly electron-withdrawing substituents destabilize benzyl carbocations suf-
ficiently so that concerted addition of solvent is enforced (SN2 mechanism).50
The acid-catalyzed hydrolysis of p-methylstyrene oxide (38a) and p-met-
hoxystyrene oxide (38b) in solutions of varying azide ion concentrations have been
interpreted by the mechanism shown in Scheme 11. Azide ion reacts with 38 by
direct nucleophilic opening of the epoxide ring at pH 9.5 to yield azido alcohols and
also by trapping of an intermediate carbocation 39, subsequent to its rate-limiting
formation, at lower pH.41 At pH 5.75, the rate of reaction of 38b in dilute sodium
azide solutions is independent of the concentration of azide ion. However, the yield
of an azide product increases with increasing azide ion concentration, suggesting
that an intermediate carbocation is trapped, subsequent to its rate-limiting forma-
tion. By using the azide ‘‘clock’’ technique43 to estimate carbocation lifetimes, the
rate constant for reaction of 39b with water (ks) is estimated to be 1:5  108 s
1 . This
MECHANISMS OF HYDROLYSIS AND REARRANGEMENTS OF EPOXIDES 261

OH
O
H H H +
H+
C CH2OH ks CHCH2OH
H
H2O
X X X
38 a, X = CH3 39 40
b, X = OCH3
kaz[N3-]
kN[N3-]

N3
Azido Alcohols
CHCH2OH

X 41

Scheme 11.

O CH3 O
H CH3 H + H H
C CHOH
H H+ H+ CH3

X X X

42a, X = CH3 43a, X = CH3 44a, X = CH3

b, X = OCH3 b, X = OCH3 b, X = OCH3

Scheme 12.

value of ks for reaction of 39b with water solvent is slightly larger than that es-
timated for reaction of 1-(p-methoxyphenyl)ethyl cation with 50:50 TFE/H2O solv-
ent ð5  107 s
1 Þ,43 and indicates that the lifetime of 39b is longer than the time
required for solvent reorganization and most likely the time required for conforma-
tional equilibration.
The acid-catalyzed hydrolysis of p-methylstyrene oxide (38a) in azide solutions
has also been studied, but the yields of azido alcohols are only very slightly greater
than the yields predicted if all of the azido alcohol is formed from the bimolecular
(kN) pathway.41 For example, in water containing 0.025 M azide ion, pH 5.75, the
observed yield of azido alcohol from 38a is 22%. The calculated yield from the
bimolecular reaction is 19%. This yield is very close, perhaps within experimental
error, to that expected from only the bimolecular pathway (kN). If it is assumed that
3% of azido alcohol is actually from trapping of the intermediate carbocation 39a
with azide ion, then a lower limit of  3  109 s
1 can be estimated for ks. This value
is fortuitously close the value of ks estimated for reaction of 1-(p-methylphenyl)ethyl
carbocation with 50:50 TFE/H2O (4  109 s
1 ), and is most likely an underestimate.
In closely related studies, the acid-catalyzed hydrolyses of cis- and trans-b-me-
thylstyrene oxides (42a and 44a)51–52 and of cis- and trans-anethole oxides (42b and
44b)53 have been reported (Scheme 12). Reactions of the cis and trans epoxides in
262 D.L. WHALEN

each series with H+ yield the same carbocation 43, and if the lifetime of this car-
bocation is sufficiently long to allow solvent reorganization and conformational
equilibration to occur, then the same mixture of erythro and threo diols should be
formed from both cis and trans epoxides. However, the isomeric cis and trans ep-
oxides 42a and 44a undergo acid-catalyzed hydrolysis to yield different erythro/
threo diol product mixtures resulting from syn and anti addition of solvent, and
therefore the lifetime of carbocation 43a must be shorter than the time required for
conformational equilibration of the intermediate carbocation. The lifetime of 43a is
expected to be comparable to that of the hydroxybenzylic carbocation that is formed
in the acid-catalyzed hydrolysis of styrene oxide 32. In contrast, 42b and 44b un-
dergo acid-catalyzed hydrolysis to yield identical mixtures of erythro and threo diols
resulting from retention and inversion of configuration at the benzyl carbon. Thus,
carbocation 43b has a lifetime sufficient for both solvent relaxation and conforma-
tional equilibration of the carbocation to occur. By using the azide ‘‘clock’’
technique43 to estimate carbocation lifetimes, the rate constant for reaction of 43b
with water (ks) is calculated to be 2  108 s
1 . This rate constant is very similar to
that calculated for reaction of the 2-hydroxy-1-(p-methoxyphenyl)ethyl carbocation
39b with water (1:5  108 s
1 ).

SUBSTITUENT EFFECTS ON ADDITION OF AMINES AND HYDROXIDE ION TO STYRENE

OXIDES

Laird and Parker studied54 the reactions of benzylamine with m- and p-substituted
styrene oxides in ethanol and showed that the Hammett r value for attack of amine
at the primary b-carbon is positive and opposite in sign to that for attack of amine at
the a-benzylic carbon. Thus, the regiochemistry of addition of amines to substituted
styrene oxides (cf. Scheme 13) is dependent on the nature of the phenyl substituent.
In the reaction of styrene oxide with benzylamine, 83% of product arises from
attack at the primary carbon and only 17% arises from attack at the benzyl carbon.
In the reaction of 3,4-dimethylstyrene oxide with benzylamine, only 19% of product
arises from attack of amine on the primary carbon and 81% arises from attack of
amine at the benzylic carbon.
It was originally reported that attack of hydroxide ion on styrene oxide occurs
only at the primary carbon.55 Later studies on the hydroxide ion-catalyzed hydro-
lysis of styrene oxide in H18
2 O and of chiral styrene oxide, however, showed that

O HO H
H OH H
H X H H
Nu:
H H + Nu
ROH
Nu H
X β-Attack X
α-Attack
45 46 47

Scheme 13.
MECHANISMS OF HYDROLYSIS AND REARRANGEMENTS OF EPOXIDES 263

attack of hydroxide ion occurs equally at the benzylic a-carbon and primary b-
carbon.41,47 The effects of p-substituents on the rate of addition of hydroxide ion to
the a-carbon is opposite to that for addition of hydroxide ion to the b-carbon, as
found earlier for the addition of benzylamine to substituted styrene oxides. Addition
of hydroxide ion and methoxide ion41 to the a-carbon is favored by electron-
donating groups, and addition of these nucleophiles to the b-carbon is favored by
electron-withdrawing groups.

SUBSTITUENT EFFECTS ON THE PH-INDEPENDENT REACTIONS OF STYRENE OXIDES

The only products from the pH-independent reactions of styrene oxide and its p-
chloro and p-methyl derivatives are the corresponding styrene glycols.41 From
analysis of the mass spectra of glycol products from the pH-independent reactions of
styrene oxide and p-methylstyrene oxide in H2O18, it was determined that 495% of
glycol was derived from attack of water at the benzylic a-carbon. Attack of water at
the benzylic carbon of chiral styrene oxide also occurs with inversion of stereo-
chemistry.47 The pH-independent reactions of styrene oxide, p-methylstyrene oxide
and most likely p-chlorostyrene oxide, therefore, appear to involve attack of water at
the benzyl carbon concerted with epoxide C–O bond breaking as shown in 48.
Hydrogen bonding of solvent to the epoxide oxygen at the transition state is also
expected, or proton transfer from solvent to the epoxide oxygen might be coupled
with C–O bond breaking.
δ
O
H H
H
X
H2O +
δ
48

The observation that water attacks only the benzylic a-carbon of styrene oxide,
whereas hydroxide ion attacks equally at both the a-carbon and b-carbon indicates
that the aromatic ring stabilizes the transition state for a-attack of water (relative to
b-attack) more than it does the transition state for a-attack of hydroxide ion. The
transition state for attack of water must be a ‘‘looser’’ structure in which the nuc-
leophile is less bonded to the benzylic carbon.
The pH-independent reactions of p-methoxystyrene oxide and of p-nitrostyrene
oxide have different mechanisms than those for reactions of styrene oxides without
strongly electron-donating or strongly electron-withdrawing groups. Both of these
compounds are much more reactive than predicted from a Hammett sigma–rho plot
of log kobsd versus s for the pH-independent reactions of styrene oxide and its p-
methyl and p-chloro derivatives.41 The most likely explanation for the enhanced
reactivity of p-nitrostyrene oxide is that attack of water at the b-carbon, which is
264 D.L. WHALEN

expected to be favored by strongly electron-withdrawing groups in the phenyl ring,

becomes the major reaction instead of attack of water at the benzylic carbon. The
pH-independent reaction of p-methoxystyrene oxide is clearly different because the
major product is p-methoxyphenylacetaldehyde, and p-methoxystyrene glycol is
only a minor product. The mechanisms of pH-independent reactions of epoxides will
be discussed in more detail in a later section.

6 Mechanisms of acid-catalyzed hydrolysis of 1-phenylcyclohexene

oxides, indene oxides and 1,2,3,4-tetrahydronaphthalene-1,2-
epoxides

ACID-CATALYZED HYDROLYSIS OF 1-PHENYLCYCLOHEXENE OXIDES

The acid-catalyzed hydrolysis of p-phenyl-substituted cyclohexene oxides 49a–d

yields diols resulting from ‘‘cis’’ and ‘‘trans’’ addition of water to the epoxide group
(Scheme 14). It was initially reported that the cis/trans diol ratio correlates well with
the electronic effect of the p-substituent, and varied from 7.5:93.5 for p-nitro-sub-
stituted oxide 49d to 95.3:4.7 for p-methoxy-substituted epoxide 49a.56–58 Later
work established that methoxy-substituted diols 50a and 51a underwent isomer-
ization under the conditions of acid-catalyzed epoxide hydrolysis, and that the cis/
trans diol ratios for hydrolysis of 49a–c are quite similar.59
A mechanism that explains some of the more important observations in the acid-
catalyzed hydrolysis of epoxides 49a–d is outlined in Scheme 15. The cis/trans diol
product ratios from the acid-catalyzed hydrolysis of 49a–c, which have either hy-
drogen- or electron-donating groups in the para position of the phenyl ring, are
74:26, 83:17 and 65:35, respectively. An intermediate carbocation 52a is trapped by
azide ion in the acid-catalyzed hydrolysis of 49a and the rate constant for reaction of
52a with water in 10:90 dioxane–water solvent is estimated, by the azide ‘‘clock’’
technique, to be 1:7  108 s
1 . Azide ion also traps an intermediate 52b in the acid-
catalyzed hydrolysis of 49b, but somewhat less efficiently. The rate constant ks for
reaction of 52b with solvent is estimated to be  2  109 s
1 . The somewhat greater
reactivity of 52b compared to that of 52a is consistent with the observation that

HO HO
O OH
H+
X X
H2O OH

49 a, X = OCH3 50 51
b, X = CH3
c, X = H
d, X = NO2

Scheme 14.
MECHANISMS OF HYDROLYSIS AND REARRANGEMENTS OF EPOXIDES 265

X
OH
O
H+ ks
X + 50 + 51
H2O H2O

49 a, X = OCH3 52
b, X = CH3
c, X = H kaz[N3-]
d, X = NO2

azides

Scheme 15.

OH X
H+ H OH
49
+ +
X H
52ax 52eq
a, X = OCH3
b, X = CH3
H2O, -H+ c, X = H H2O, -H+
trans diol 51 d, X = NO2 cis diol 50

Scheme 16.

tertiary a-substituted p-methylbenzyl carbocations are more reactive with solvent

than the corresponding a-substituted p-methoxybenzyl carbocations.60 The acid-
catalyzed hydrolysis of 49a–c can therefore be regarded as proceeding via discrete
carbocation intermediates 52a–c.
In a study of the acid-catalyzed equilibration of cis and trans diols 50a, b and 51a,
b, it was determined that the cis diol in each system is more stable than the trans
diol.59 A mechanism for hydrolysis of 49 that takes into account the relative
stabilities of the diol products and also conformational factors was proposed, and is
given in Scheme 16. In this mechanism, the reaction of 49 with H+ occurs with axial
C–O bond breaking to yield a carbocation with an axial hydroxyl group (52ax).
Inversion of the ring yields a second carbocation conformation 52eq, in which the
hydroxyl group occupies an equatorial position. It was suggested that axial attack of
solvent on each conformation would be favored, and that those factors contributing
to the greater stability of the cis diol product would also contribute to lowering
of the transition-state energy leading to it. However, this assumption is not always
valid, and factors other than the stabilities of the hydrolysis products may play very
important roles.61
The mechanism of acid-catalyzed hydrolysis of 49d, which yields 492% of trans
diol,57 is clearly different than that for hydrolysis of 49a–c. The lifetime of the
carbocation 52d is expected to be considerably shorter than the lifetimes of 52a and
266 D.L. WHALEN

b. If 52d were sufficiently unstable, then addition of solvent, concerted with proton
transfer and benzylic C–O bond breaking, would occur and would yield only trans
diol. Another possibility is that carbocation conformation 52d(ax) is formed from
the reaction of 49d with H+, and undergoes axial attack by solvent faster than it
undergoes ring inversion to form 52d(eq), i.e. the lifetime of 52(ax) is short relative
to the time required for conformational change.

ACID-CATALYZED HYDROLYSIS OF INDENE OXIDES: TRANSITION-STATE EFFECTS ON

STEREOCHEMISTRY OF DIOL FORMATION

The cis/trans hydrolysis ratio from the acid-catalyzed hydrolysis of indene oxide
(53a, Scheme 17) is 75:25,62,63 and that from acid-catalyzed hydrolysis of 5-met-
hoxyindene oxide (53b) is 80:20.61 Thus, the introduction of a methoxy group into
position 5 does not result in any significant change in the cis/trans hydrolysis ratio.
There is evidence that 54b, which is stabilized by the 5-methoxy group, is sufficiently
stable to be trapped by external nucleophiles. The very similar cis/trans diol product
ratio from acid-catalyzed hydrolyses of both 53a and b suggest that 54a is also a
discrete intermediate.
The cis and trans diols 55b and 56b are sufficiently reactive in dilute acid solution
to undergo equilibration, and the trans diol was determined to be more stable than
the cis diol by a factor of 3. In this system, therefore, the major diol from acid-
catalyzed hydrolysis is the less stable isomer. This result contrasts with the results
from acid-catalyzed hydrolysis of phenyl-substituted cyclohexene oxides, where the
major cis diol product is also more stable than the corresponding trans diol. In the
hydrolysis of indene oxides 53a, b, therefore, effects that are present in the transition
state but absent in the products must play major roles in controlling the cis/trans
product ratio.
Quantum chemical calculation of carbocation 54 suggests that it occupies only
one ground-state conformation in which the five-membered carbon ring is near
planar.61 If there is no hydroxyl substitution at the b-carbon of 54, the resulting
carbocation would have a plane of symmetry, and solvent would attack equally from
both sides of the electron-deficient benzylic carbon. The presence of the b-hydroxyl
group is therefore responsible in some way for lowering the transition-state energy
for cis attack of water leading to the less stable cis diol (or perhaps increasing the
transition-state energy for trans attack of water). As pointed out in a review article

X X X X
H+ H2O
OH OH + OH
O -H+
+
OH OH
53 a, X = H 54 55 56
b, X = OCH3

Scheme 17.
MECHANISMS OF HYDROLYSIS AND REARRANGEMENTS OF EPOXIDES 267

by Richard et al., the Marcus intrinsic barrier leading to the less stable cis diol must
be lower than the intrinsic barrier leading to the trans diol.64 Intramolecular hy-
drogen bonding between the attacking water molecule and the b-hydroxyl group is
possible in the transition state 57 leading to the cis diol, but not in the transition
state 58 leading to the trans diol. This intramolecular hydrogen bonding may con-
tribute to lowering the intrinsic barrier in a number of ways. One possibility is that
this internal hydrogen bond in 57 is stronger than the hydrogen bonding between the
attacking water molecule and solvent in the transition state for trans attack of water
(58). Nearby solvent molecules other than the one attacking the benzylic carbo-
cation must also play a role in the selective stabilization of 57. A second factor that
may contribute to the greater stabilization of the transition state for cis attack of
water is that intramolecular hydrogen bonding may result in an ‘‘earlier’’ transition
state with less hybridization change of the benzylic carbon than that occurring in the
formation of the transition state for trans attack of water.61 A greater change in
hybridization of the benzylic carbon for trans attack of water would result in a
greater imbalance between loss of resonance stabilization and lowering of energy
due to bond formation at the transition state, resulting in a raising of the intrinsic
barrier for trans attack of water.
Transition-state effects in which a b-hydroxy group stabilizes the transition state
for cis attack of water on a b-hydroxycarbocation may also be responsible for the
observation that cis diols are the major diols formed in the acid-catalyzed hydrolysis
of acenaphthylene oxide and cyclopenta[cd]pyrene oxide.65 Due to this rate-
enhancing effect, b-hydroxycarbocations may be more reactive with solvent than
their parent carbocations that do not contain the b-hydroxy group.
X
H δ+ H O H
X
O H
O δ+

δ+ δ+
O
57 H H
58

TETRAHYDRONAPHTHALENE EPOXIDE HYDROLYSIS: CONFORMATIONAL EFFECTS

ON STEREOCHEMISTRY OF DIOL FORMATION

The stereochemical outcomes of the acid-catalyzed hydrolyses of 1,2,3,4-tetrahy-

dronaphthalene 1,2-oxide 59a and its 6-methoxy derivative 59b (Scheme 18) are
quite different from those of the corresponding indene oxide systems. For example,
acid-catalyzed hydrolysis of 59a yields only 5% of cis diol 61a and 95% of trans diol
62a.66 However, acid-catalyzed hydrolysis of its 6-methoxy derivative 59b yields
80% of cis diol 61b and 20% of trans diol 62b.67 One might argue that carbocation
60a does not have a sufficiently long lifetime compared to solvent relaxation,
whereas 60b does, and that this difference in lifetime leads to different mechanisms
268 D.L. WHALEN

X X X X
H+ H2O +
-H+
OH OH OH
O +
OH OH
59 a, X = H 60 61 62
b, X = OCH3

Scheme 18.

and different stereochemical outcomes. However, the cis/trans diol ratio from hy-
drolysis of 59a at acidic pH in KCl solution is different from hydrolysis of 59a in
NaClO4 solution, although the rate of hydrolysis does not change.66 This obser-
vation indicates that there is an intermediate in the reaction, formed in a rate-
limiting step. This intermediate is presumed to be a carbocation. Further, carbo-
cation 60a should have about the same reactivity with solvent as 54a, which appears
to have a lifetime longer than the time required for solvent relaxation. To reconcile
the different stereochemical outcomes of the hydrolyses of indene oxides and tet-
rahydronaphthalene oxides, it was proposed that conformational factors play im-
portant roles in the stereochemistry of diol formation from 60a and b.67
Whereas indene oxide and its 2-hydroxy-1-indanyl carbocation are each calcu-
lated to occupy one stable conformation, tetrahydronaphthylene oxide 59a and
carbocation 60a are each calculated at the B3LYP/6-31G* level of theory to have
two minimum-energy conformations, shown in Scheme 19. The mechanism for hy-
drolysis of this system should therefore take into consideration these additional
conformations.
Tetrahydronaphthalene oxide 59a has two conformations, 63a and 63b. From
analysis of vicinal H–C–C–H coupling constants in the 1H NMR spectrum of 59a, it
was concluded that conformation 63a, in which the bonds of the angular ring have
fewer eclipsing interactions than those of 63b, is the more stable.67 This conclusion is
supported by quantum chemical calculations in which conformation 63a is calcu-
lated at the B3LYP/6-31G* level of theory to be 44 kcal mol
1 more stable than 63b
in the gas phase.49
Addition of H+ to the more stable conformation 63a yields carbocation 64ax, in
which the newly formed hydroxyl group occupies an axial position. Conformational
inversion of the cyclohexenyl ring in 64ax leads to a second conformation of the
carbocation in which the hydroxyl group occupies an equatorial position. This
carbocation can also be formed directly by the reaction of the less stable epoxide
conformation 63b with H+. These two carbocation conformations are related
in structure to the cyclohexenyl carbocation. From the work of Goering and
Josephson, it was shown that the cyclohexenyl carbocation undergoes energetically
favorable axial attack of solvent, which converts the intermediate carbocation to the
product possessing an energetically favorable half-chair conformation.68 Equatorial
attack of water on the cyclohexenyl carbocation, on the other hand, leads to product
with an energetically unfavorable half-boat conformation.
MECHANISMS OF HYDROLYSIS AND REARRANGEMENTS OF EPOXIDES 269

Scheme 19.

Application of these principles to Scheme 19 leads to the prediction that confor-

mation 64ax should react with solvent to preferentially form trans diol, whereas
reaction of conformation 64eq with water should lead preferentially to cis diol.
Conformation 64eq, in which the hydroxyl group occupies an equatorial position, is
calculated to be more stable than conformation 64ax in the gas phase by
1.25 kcal mol
1.49 A possible explanation for the observed results is that tetra-
hydronaphthalene oxide reacts with H+ only from the more stable ground conforma-
tion 63a to yield the less stable carbocation conformation 64ax. Carbocation 64ax
then reacts with water to form trans diol faster than it undergoes conformation
inversion to form carbocation 64eq. When a methoxy group occupies position 6, it
stabilizes the carbocation, resulting in a smaller rate constant for reaction of 64ax(6-
OCH3) with solvent. However, substitution of the 6-methoxy group should not sig-
nificantly change the energy barrier for inversion of the carbocation ring. Therefore,
ring inversion that transforms 64ax(6-OCH3) to 64eq(6-OCH3) may occur faster than
the attack of water on 64ax(6-OCH3), and axial attack of water on this carbocation
conformation would yield cis diol as the major product.
The mechanism in Scheme 19 is somewhat speculative, and relies heavily on the
assumption that axial attack of solvent is more energetically favorable than equat-
orial attack of solvent on carbocations similar in structure to 64. However, the
stereochemistry of acid-catalyzed hydrolysis of two diastereomeric hexa-
hydrophenanthrene 9,10-epoxides (65 and 67) provides support for this proposal.
These two epoxides contain transfused cyclohexane moieties that restrict the geo-
metry of each epoxide to a single conformation. Reaction of 65 with H+ yields a
single carbocation conformation 66, in which the hydroxyl group is forced to occupy
an axial position (Scheme 20).
270 D.L. WHALEN

Scheme 20.

Scheme 21.

Acid-catalyzed hydrolysis of 65a yields trans diol as the only detectable product,69
and this diol product is rationalized by axial attack of water on carbocation 66a.
Since the lifetime of a simple benzyl carbocation is very short, an argument might be
made that the trans diol is instead formed by an A-2 mechanism. However, we have
also synthesized the methoxy derivative 65b, and its acid-catalyzed hydrolysis also
yields 498% trans diol.70 The lifetime of carbocation 65b is sufficiently long that it
can be captured, subsequent to its rate-limiting formation, by azide ion. This result
clearly shows that the trans diol product from acid-catalyzed hydrolysis 65b is
formed by axial attack of water on a discrete carbocation (66b).
Whereas acid-catalyzed hydrolyses of 65a and b yield almost exclusively trans diol
products, the acid-catalyzed hydrolysis of 67a yields 85% of cis diol and 15% of
trans diol.69 The reaction of 67 with H+ also yields a single carbocation conforma-
tion 68, in which the hydroxyl group is forced to occupy an equatorial position
(Scheme 21). The major cis diol product can be rationalized by energetically favored
axial attack of water on carbocation 68. In unpublished work, we have also de-
termined that the acid-catalyzed hydrolysis of 67b, which is much more reactive than
67a and yields a stabilized carbocation 68b, also yields cis and trans diols in a ratio
of 85:15. Axial attack of water on 68 to yield cis diol is therefore favored over
equatorial attack to yield trans diol, but not to the same extent that axial attack of
water on carbocation 66 is favored.
MECHANISMS OF HYDROLYSIS AND REARRANGEMENTS OF EPOXIDES 271

7 General acid catalysis in epoxide reactions

ETHYLENE OXIDE AND SIMPLE PRIMARY AND SECONDARY EPOXIDES

In the early studies by Brønsted et al. of the rates of hydrolysis of epichlorohydrin

(69) in water solutions containing formic acid, acetic acid, benzoic acid and
trimethylacetic acid buffers, it was observed that the rate of reaction depends on the
concentration of the buffer base, but not the buffer acid.22 Therefore, nucleophilic
addition of carboxylate ions to epichlorohydrin was observed, but general acid
catalysis was not observed. However, the maximum concentration of the buffer
acids used in this study was only 0.02 M, and the observation of general acid cata-
lysis might not be expected. In a later study by Swain, it was observed that the rate
of reaction of epichlorohydrin (69) in 0.15 M NaI/4.0 M HOAc/1.0 M NaOAc so-
lution is 2.2 times faster that its rate in 0.15 M NaI/0.4 M HOAc/0.1 M NaOAc/
0.90 M NaClO4 solution.71 General acid-catalyzed addition of iodide ion to epi-
chlorohydrin by a mechanism outlined in Scheme 22 was proposed. This mechanism
involves concerted proton transfer from acetic acid to the epoxide oxygen, coupled
with C–O bond breaking and C–I bond making at the transition state; a carbocation
is not involved in this proposed mechanism. Control experiments in which it was
shown that added sodium perchlorate actually resulted in a lowering of the rate of
hydrolysis of epichlorohydrin and that addition of other organic solvents such as
dioxane and acetamide had little effect on the rate provide evidence that the pro-
posed mechanism is correct. The concentration of epichlorohydrin is quite high
(0.13 M), and the reaction presumably involves attack of iodide ion on an associ-
ation complex of epichlorohydrin and acetic acid. Although this study presents
evidence for general acid-catalyzed addition of iodide ion to a simple epoxide, gen-
eral acid-catalyzed addition of solvent to a primary or secondary epoxide has not
been observed, and may not be a viable mechanism because solvents such as water
are much less nucleophilic than iodide ion.
The interpretation of the observed rate increase of the reaction of epichlorohydrin
with iodide ion in the presence of acetic acid was questioned by Long and Paul, who
suggested that specific effects of the medium brought about by 4.0 M acetic acid may
be responsible for the rate increase rather than general acid catalysis.72 However, the
alternative mechanism for acid-catalyzed epoxide hydrolysis proposed by these
workers, an A-1 mechanism involving a carbocation intermediate, has been ruled

OAc +
+
H OH
O
+ HOAc + I O ClCH2CHCH2I + AcO
ClCH2 71
69 ClCH2 I
70

Scheme 22.
272 D.L. WHALEN

out for primary and secondary epoxides (Section 3). It is important to note that
Swain’s results provide evidence for general acid-catalyzed addition of iodide ion to
a simple epoxide, epichlorohydrin, but not for general acid-catalyzed addition of
solvent to simple epoxides.

TERTIARY EPOXIDES

Although general acid-catalyzed hydrolysis of a simple tertiary epoxide by formic

acid is reported, the kinetic term in formic acid is difficult to detect over the back-
ground hydronium ion-catalyzed reaction. The rate of reaction of tetramethyl-
ethylene oxide in solutions with ionic strength held constant at 2.0 M and buffer
ratio held constant increases by only 22% as the concentration of formic acid is
varied up to 1.0 M.73 It was estimated that two-thirds of the rate increase could be
accounted for by a formate ester product, formed by nucleophilic attack of formate
ion on protonated epoxide. The remaining one-third of the rate increase (7%) was
attributed to general acid catalysis. A specific-acid–general-base mechanism, in
which there is a pre-equilibrium protonation of the epoxide group followed by rate-
determining formate-assisted attack of water on protonated epoxide, was suggested
for this reaction. Additional studies of the reactions of this and other simple tertiary
epoxides with general acids, including but not limited to formic acid, would be
helpful in confirming the interpretation of these results.

ACETALS AND EPOXY ETHERS

General acid catalysis in the hydrolysis of simple acetals is usually not observed.74
However, general acid catalysis has been observed in the hydrolysis of acetals with
phenolic leaving groups when the intermediate oxocarbocations possess moderate
stability.75 General acid catalysis has also been observed in the hydrolysis of acetals
that yield an alkoxytropylium ion76 or an oxocarbocation stabilized with a 4-met-
hoxyphenyl group.77 The mechanism proposed for the general acid-catalyzed hy-
drolysis of 2-(4-nitrophenoxy)tetrahydropyran (72) is concerted, with proton
transfer from the general acid coupled with C–O bond breaking at the transition
state 73 (Scheme 23).78
General acid catalysis by dihydrogen phosphate ion is also observed in the hy-
drolysis of an epoxy ether 74.79 Although the authors did not discuss the nature of
the dihydrogen phosphate-catalyzed reaction, they proposed that in the hydronium
ion-catalyzed reaction of 74, proton transfer from hydronium ion to the epoxide
oxygen is concerted with epoxide O–C bond breaking in a reactant-like transition
state. Although the reaction of H+ with 74 does not yield a highly stabilized oxo-
carbocation, the free energy change associated with the breaking of the C–O bond of
the epoxide ring provides a driving force for the reaction. Thus, 74 is very reactive
toward acid-catalyzed hydrolysis; the bimolecular rate constant for its hydronium
ion-catalyzed hydrolysis in 10:90 dioxane/water at 25 1C is 1:57  105 M
1 s
1 .
MECHANISMS OF HYDROLYSIS AND REARRANGEMENTS OF EPOXIDES 273

+
+

O + O NO2
+ RCO2H δ
O O NO2 H

72 O
R C δ-
O
73

Scheme 23.

O
CH3 Ph

CH3 OCH3

74

VINYL EPOXIDES

General acid catalysis has also been observed in the hydrolysis of epoxides that react
with H+ to yield stabilized carbocations. For example, the hydrolysis of 1,3-
cyclohexadiene oxide 13 is catalyzed by dihydrogen phosphate and cacodylic acid.36a
Other general acids such as cyanomethylamine hydrochloride are not catalytic. The
bimolecular rate constant for general acid catalysis of 13 by dihydrogen phosphate
ion (kHA) is 0.074 M
1 s
1, a rate constant sufficient to increase the rate of hydro-
lysis of 13 by 60% at pH 5.78 in 0.1 M H2PO
4 . The allylic carbocation 29 is not
highly stabilized, and the rate increase due to general acid catalysis is not large. The
fact that dihydrogen phosphate ion catalyzes the hydrolysis whereas ammonium
ions do not may be due to electrostatic stabilization of the partially formed allylic
cation by the negatively charged dihydrogen phosphate ion at the transition state 75
(Scheme 24).80–82
Cyclohexadiene oxide 13 is very reactive toward hydronium ion-catalyzed hy-
drolysis; the bimolecular rate constant for acid-catalyzed hydrolysis
kH ¼ 1:6  104 M
1 s
1 . General acid catalysis by dihydrogen phosphate or other
general acids was not observed in the hydrolysis of 1,3-cyclooctadiene oxide 11,
which is much less reactive toward hydronium ion-catalyzed hydrolysis
(kH ¼ 3:6 M
1 s
1 at 25 1C) than 13.36a The magnitude of general acid catalytic
terms for hydrolysis of a particular epoxide is expected to correlate with its reactivity
with hydronium ion. So it is not surprising that general acid catalysis in the hy-
drolysis of simple tertiary epoxides such as tetramethylethylene oxide
(kH ¼ 38 M
1 s
1 at 36 1C)73 and isobutylene oxide (kH ¼ 6:8 M
1 s
1 at 25 1C)23
274 D.L. WHALEN

+
- +
OPO3

H OH OH OH
O OH
- O H2O
+ H2PO4 +
+
+
13 δ 29 30 OH
31
75

Scheme 24.

is difficult to detect. Values of kHA for general acid-catalyzed hydrolysis of simple

tertiary epoxides by general acids with pK a s44 are expected to be orders of mag-
nitude smaller than kH.

BENZYLIC EPOXIDES AND ARENE OXIDES

Relatively little is published on general acid-catalyzed hydrolysis of benzylic epox-

ides and arene oxides. The hydrolysis of phenanthrene 9,10-oxide is catalyzed by a
number of general acids,83 and general acid catalysis by dihydrogen phosphate in the
hydrolysis of cis- and trans-anethole oxides (4-methoxy-b-methylstyrene oxides) is
readily observed.84 General acid catalysis in the hydrolysis of benzene oxide
(kH ¼ 32 M
1 s
1 ), however, is not detectable.85 The magnitude of general acid cata-
lysis observed in epoxide hydrolysis is correlated with the ability of substituent
groups to stabilize the carbocation resulting from epoxide ring opening.
Precocene I oxide (76) exhibits a particularly interesting pH-rate profile, and
shows striking general acid catalysis.86 Reaction of this epoxide with H+ yields a
highly stabilized benzylic carbocation 77 (Scheme 25). The hydronium ion-catalyzed
hydrolysis of 76 is very facile (kH ¼ 1:2  105 M
1 s
1 ), and yields only cis and trans
diols 78 and 79. A most interesting observation is that general acid catalysis by MES
(2-[N-morpholino]ethanesulfonic acid, pK a ¼ 6:17), and BES (N,N-bis[2-hydroxy-
ethyl]-2-aminoethanesulfonic acid, pK a ¼ 7:08) in the hydrolysis of 76 results a lin-
ear dependence of rate on the concentration of general acid, whereas plots of kobsd
versus general acid concentration for Tris (2-amino-2-[hydroxymethyl]-1,3-propane-
diol, pK a ¼ 8:06), HEPES (N-[2-hydroxyethyl]piperazine-N0 -2-ethanesulfonic acid,
pK a ¼ 7:49) and CHES (2-[N-cyclohexylamino]ethanesulfonic acid, pK a ¼ 9:28),
respectively, are nonlinear. Implications of this observation and the complicated
pH-rate profile will be discussed in a later section. Catalytic constants for general
acid catalysis in the hydrolysis of 76 are especially large and, of course, vary with the
pKa of the general acid. The bimolecular rate constant for general acid-catalyzed
hydrolysis of 76 by MES, e.g., is 152 M
1 s
1. Reaction of the acid form of MES or
MECHANISMS OF HYDROLYSIS AND REARRANGEMENTS OF EPOXIDES 275

CH3O O CH3 + CH3O CH3 CH3O O CH3 CH3O O CH3

CH3 H O
CH3 CH3 CH3
kH H2O +
OH OH OH
O +
OH OH
76 77 78 79

Scheme 25.

BES buffers (pK a o7:5) with 76 to form benzylic carbocation 77 is solely rate lim-
iting, whereas increase in Tris, HEPES and CHES (pK a 47:5) buffer concentrations
bring about a change in mechanism from rate-determining epoxide ring opening at
low buffer concentration to attack of solvent on the carbocation at high buffer
concentration.
Pronounced general acid catalysis by acetic acid (pK a ¼ 4:8), dihydrogen phos-
phate ion (pK a ¼ 6:9), Tris H+ (pK a ¼ 8:2), hydroxyethylammonium ion
(pK a ¼ 9:5) and phenol (pK a ¼ 10:0) has been observed in the hydrolysis of ben-
zo[a]pyrene 7,8-diol 9,10-epoxides 80 and 81.87 Only the acid form of each buffer
was found to be catalytic, and it was proposed that the mode of general acid cata-
lysis involves rate-determining reactions of general acid with diol epoxides 80 and 81
to yield benzylic carbocations 82 and 83, respectively. Plots of kobsd for reactions of
80 and 81 versus [H2PO
4 ] are provided in Fig. 2. A concentration of 0.05 M H2PO4

is sufficient to increase the rate of reaction of diol epoxide 81 by a factor of 20–25.

Brønsted plots of log kHA versus pKa for general acid-catalyzed hydrolysis of
diol epoxide 81 by a series of ammonium ions88 with pKas varying between 5.4 and
10.6, and of both 80 and 81 by dihydrogen phosphate and a series of phosphonate
anions89 provide reasonable Brønsted a values of 0.45–0.50.

O O HO + +
HO

HO HO HO HO
OH OH OH OH
80 81 82 83

Kinetic solvent deuterium isotope effects for the hydronium ion-catalyzed hy-
drolyses (k(H3O+)/k(D3O+)) of diol epoxides 80 and 81 are 0.67 and 0.70, respect-
ively, which are considerably larger than the solvent kinetic deuterium isotope ef-
fects observed for the hydronium ion-catalyzed hydrolyses of acetals that are
thought to hydrolyze via specific acid catalysis (k(H3O+)/k(D3O+)  0.37).75 The
observed solvent deuterium isotope effects for the acid-catalyzed hydrolyses of 80
and 81 are very close to the value of 0.75 observed for the acid-catalyzed hydrolysis
of 2-(p-nitrophenoxy)-tetrahydropyran, an acetal that is thought to hydrolyze by a
mechanism in which hydronium ion acts as a general acid.75 The kinetic isotope
effects and the fact that hydronium ion fits rather well on a Brønsted plot of log kHA
for reactions of 80 and 81 catalyzed by a series of phosphonate ions versus pKa
of the general acid suggest that hydronium ion is acting as a general acid in the
276 D.L. WHALEN

0.03
80

81
ko bsd , s-1 0.02

0.01

0.00
0.00 0.01 0.02 0.03 0.04 0.05 0.06

[H2PO4 ], M

Fig. 2 Plot of kobsd versus ph for reaction of 80 and 81 in 10:90 dioxane/water, pH 6.9.
Reproduced with permission from Ref. 87. Copyright 1979 American Chemical Society.

Fig. 3 Plots of kobsd versus total concentration of monophosphate ester, 10:90 dioxane/
water, 25 1C. Reproduced with permission from Ref. 90b. Copyright 1987 American Chemical
Society.

hydrolyses of these diol epoxides and may act as a general acid in the hydrolysis of
other epoxides that hydrolyze to form sufficiently stable carbocations.
Even more pronounced general acid catalysis by nucleoside monophosphates in
the hydrolysis of diol epoxides 80 and 81 is observed.90 In Fig. 3 are plots of kobsd
versus total concentrations of guanosine (G), ribose 50 -phosphate, 50 -cytosine mono-
phosphate (50 -CMP), 50 -adenosine monophosphate (50 -AMP) and 50 -guanosine
MECHANISMS OF HYDROLYSIS AND REARRANGEMENTS OF EPOXIDES 277

monophosphate (50 -GMP). Whereas guanosine does not catalyze the hydrolysis
of 81, monophosphate esters of cytosine, adenosine and guanosine are catalytic.
Ribose 50 -phosphate (pK a ¼ 6:6) is slightly more effective than dihydrogen phos-
phate (pK a ¼ 6:9), but 50 -AMP and 50 -GMP (pKas  6.3–6.5) are 50–60 times
more effective than dihydrogen phosphate as general acids. The greater catalytic
efficiencies of 50 -AMP and 50 -GMP compared to dihydrogen phosphate ion
(similar pKas) are attributed to a favorable stacking interaction or association of
the nucleotide base with the polycyclic arene portion of the diol epoxide at the
transition state.90

8 pH-independent reactions of epoxides

SIMPLE ALKYL EPOXIDES

Ethylene oxide, propylene oxide and isobutylene oxide each undergoes hydrolysis to
form a 1,2-glycol by a reaction whose rate is independent of pH.22,23 This reaction
has been referred to as a ‘‘spontaneous’’ reaction,22 a ‘‘water’’ reaction23 and a ‘‘pH-
independent’’ reaction. The pH-independent reaction of propylene oxide in H2O18
yields glycol with 65% of O18 located on the primary carbon and 35% located on
the secondary carbon.23 These results suggest that the pH-independent reaction of
simple epoxides with primary and secondary carbon centers occurs by nucleophilic
addition of water, with attack at a less hindered primary carbon slightly favored
over attack at a secondary carbon. An attempt to study the pH-independent re-
action of isobutylene oxide in H182 O was made, but the pH of the solution drifted
from its initial value where the pH-independent reaction predominates to a value of
4, where the acid-catalyzed reaction predominates, over the long 80-h reaction
time.23 Some 18O (10–20%) was incorporated at the primary carbon in these
experiments, whereas 499% 18O is located at the tertiary carbon of the glycol
formed in the acid-catalyzed reaction. The only conclusion that can be made from
this result is that attack of water at the primary carbon in the pH-independent
reaction of isobutylene oxide is at least a minor reaction pathway (420%).

ARENE OXIDES

A number of aromatic hydrocarbons are metabolized to phenols via intermediate

arene oxides.91 In a very important study of the rates of rearrangements of benzene
oxide and 1,2-naphthalene oxide to their corresponding phenols, it was demon-
strated that the rearrangement reaction occurs by two kinetically distinct pathways,
an acid-catalyzed route and a pH-independent route.92 From deuterium label-
ing studies, it was shown that each pathway involves a 1,2-hydrogen migration to
yield an intermediate ketone, which then rearranges rapidly to phenol product.93
This 1,2-hydrogen migration has been referred to as the ‘‘NIH’’ shift.91 Proposed
278 D.L. WHALEN

mechanisms for the acid-catalyzed rearrangement of benzene oxide to phenol are

outlined in Scheme 26. One mechanism involves reversible protonation of benzene
oxide, followed by rate-limiting epoxide C–O bond breaking to yield a carbocation
86, rearrangement with 1,2-hydrogen migration of 86 to oxocarbocation 87, and loss
of a proton from 87 to yield phenol (88). A second possible mechanism is concerted,
with hydrogen migration and epoxide C–O bond breaking occurring concurrently.
On the basis of a great sensitivity of the reaction to substituent effects and the lack of
a primary kinetic deuterium isotope effect on hydrogen migration, the concerted
mechanism was ruled out.94
The kinetic deuterium isotope effects on the pH-independent reactions of per-
deuterobenzene oxide and 1-2H-dihydronaphthalene-1,2-oxide to form phenols have
been determined.94 The lack of primary kinetic isotope effects on these reactions was
taken as evidence that hydrogen migration does not occur in the rate-limiting step. If
the reaction were concerted and 1,2-hydrogen migration were part of the rate-
H D
limiting step, then kH D
o /ko should be 2–3. Instead, values for k o =k o ¼ 1:05 for each
reaction were obtained. On the basis of a lack of a primary kinetic deuterium isotope
effect and a very large negative Hammett r value (
7.3), a stepwise mechanism for
the pH-independent reactions of these arene oxides was proposed as outlined in
Scheme 27 for the rearrangement of dihydronaphthalene oxide 89 to 1-naphthol. In
this mechanism there is rate-limiting epoxide C–O bond breaking in 89 to form a
zwitterion intermediate 90, with 1,2-hydrogen migration occurring after the rate-
limiting step. The observation that 80% of the original deuterium is located at C-2
of the 1-naphthol product 92 requires a 1,2-deuterium migration and the interme-
diacy of ketone 91 in the reaction.93 An isotope effect on the reaction of 91 to form
phenol 92 is expected, which will result in a greater loss of hydrogen than deuterium
from 91.

H OH OH
H+ r.d.s. + OH
+ H
O O H H
A-1
+ H
H H
84 85 86 87 88

concerted

Scheme 26.

D D O- O OH
O D
H r.d.s. H D(H)
+ H

89 90 91 92

Scheme 27.
MECHANISMS OF HYDROLYSIS AND REARRANGEMENTS OF EPOXIDES 279

Arene oxides are also known to undergo an NIH shift in which there is 1,2-alkyl
group migration. For example, 1,4-dimethylbenzene oxide undergoes acid-catalyzed
and pH-independent reactions to yield both 2,4-dimethylphenol and 2,5-dimethyl-
phenol.95 2,5-Dimethylphenol is formed from 1,2-hydrogen migration, and 2,4-di-
methylphenol is formed from 1,2-methyl migration. A new mechanism for the NIH
shift involving a diol intermediate was discovered in this study. The pH-independent
reaction of 8,9-indane oxide to form 4-indanol occurs in part via a 1,2-alkyl mi-
gration to form an intermediate spiro ketone, which undergoes secondary isomer-
ization to the observed phenol.96

CYCLIC VINYL EPOXIDES

Whereas the acid-catalyzed hydrolysis of cyclohexadiene oxide 13 gives comparable

yields of 1,2- and 1,4-diols (Scheme 9), the pH-independent reaction of 13 yields only
3-cyclohexene-trans-1,2-diol.36a,c The hydroxide ion-catalyzed hydrolysis of 13 also
yields only trans 1,2-diol. The observation that the pH-independent reaction of 13
yields only trans 1,2-diol suggests that epoxide C–O bond breaking and water O–C
bond making are concerted. There is a negative salt effect on the pH-independent
reaction of 13,36c so the transition state must not have significant charge separation.
The newly developing oxyanion is expected to be either strongly hydrogen bonded
with solvent, or proton transfer from solvent to the oxyanion may also be part of the
transition state (i.e. structure 93).
-
HO δ
H

δ+
OH2
93

The pH-independent reaction of cyclopentadiene oxide 12 in water is clearly dif-

ferent from that of 13.97 This reaction yields 33% of 3-cyclopentenone, 35% of cis-
2,4-pentadienal and 35% of a mixture of cis and trans 1,2- and 1,4-diols (Scheme
28). When the reaction is carried out in D2O instead of H2O, no deuterium is
incorporated into the ketone product. Thus, 1,2-hydrogen migration is required for
this reaction just as it is in the rearrangement of arene oxides to phenols. The
mechanisms of product formation in this reaction are not fully understood. The
observation that the diol mixture is similar to that from the acid-catalyzed hydro-
lysis of 12 suggests that an allylic carbocation may be involved in the diol-forming
reaction. Ketone (96) and dienal (97) products are potentially formed either
by stepwise or concerted mechanisms, and there is insufficient evidence to rule
out either one. There is a significant normal salt effect on this pH-independent
280 D.L. WHALEN

OH O
OH O
O ko H
OH
+ + +

HO
12 94 (16%) 95 (19%) 96 (33%) 97 (32%)

Scheme 28.

reaction,36c so there must be substantial charge separation in the transition states

leading to products.

BENZYLIC EPOXIDES THAT UNDERGO RATE-LIMITING 1,2-HYDROGEN MIGRATION

Many benzylic epoxides undergo rearrangements to ketones in their pH-indepen-

dent reactions. For example, the pH-independent reactions of indene oxide63 and 5-
methoxyindene oxide61 yield 75–80% of 2-indanones, along with cis and trans diols.
The pH-independent reaction of 6-methoxy-1,2,3,4-tetrahydronaphthalene-1,2-
epoxide (98, Scheme 29) also gives 76% of ketone (99), along with cis and trans
diols 100 (2.4:1.0 cis/trans ratio).67 However, the pH-independent reaction of
1,2,3,4-tetrahydronaphthalene-1,2-epoxide yields only a trans 1,2-diol, presumably
from nucleophilic attack of water at the benzylic carbon.66 The transition state for
the rearrangement pathway from benzylic epoxide to ketone must possess significant
positive charge on the benzylic carbon, since an electron-donating p-substituted
methoxy group promotes the rearrangement pathway over the diol-forming path-
way.
The kinetic deuterium isotope effect kH D
o /ko on the pH-independent reaction of
98
98 is reported to be 1.59. This isotope effect is in the range expected for rate-
determining 1,2-hydrogen migration. If it is assumed that the isotope effect on the
minor diol-forming reaction of 98 is near unity, then the isotope effect on the
ketone-forming reaction is calculated to be 2.0. This observation indicates that
migration of hydrogen in the ketone-forming reaction of 98 is rate limiting. The
observed kinetic deuterium isotope effect on the reaction of precocene I oxide 76 at
pH 11, which yields 75% of rearranged ketone, is 2.15. The deuterium isotope effect
on 1,2-hydrogen migration is estimated to be 4, which is somewhat larger than
isotope effects on most 1,2-hydrogen migrations and indicates substantial hydrogen
migration at the transition state.86
The observed kinetic isotope effect kH D
o /ko on the pH-independent reaction of 5-
methoxyindene oxide (101), which yields ketone 102 as the major product,63 is 2.22
(Scheme 30).99 This observed kinetic isotope effect is even larger than those for
reaction of 98 and 76. The large substituent effect of the methoxy group on this
reaction suggests that epoxide C–O bond breaking is more advanced than 1,2-
hydrogen migration, leaving substantial positive charge on the benzylic carbon at
MECHANISMS OF HYDROLYSIS AND REARRANGEMENTS OF EPOXIDES 281

HO
H H H(D) H OH
O
H(D) ko O H(D)
+

CH3O CH3O
CH3O
98 99 100

Scheme 29.

δ- δ-
H H H O O
H
O ko H(D)
H H
H(D) O δ+ δ+
CH3O CH3O CH3O CH3O
101 102 103 104

Scheme 30.

the transition state as shown in structure 103. Another possibility is that epoxide
C–O bond breaking is completed before 1,2-hydrogen migration occurs, as shown in
104. Calculations at the MP2/6-31G*//MP2/6-31G* level of theory for the rear-
rangement of protonated propene oxide to protonated propanal in the gas phase
show a concerted, asynchronous pathway in which the epoxide C–O bond is broken
before hydrogen migration starts.100
The observations that the pH-independent reactions of deuterium-labeled 5-met-
hoxyindene oxide and 6-methoxy-1,2,3,4-tetrahydronaphthalene-1,2-oxide show
significant primary kinetic deuterium isotope effects for the ketone-forming reac-
tions, whereas the pH-independent reactions of deuterium-labeled naphthalene ox-
ide and benzene oxide do not, are quite puzzling. Clearly, more work needs to be
done to fully understand why transition-state structures for rearrangement of arene
oxides to phenols differ from those for rearrangement of benzylic epoxides to
ketones.

BENZO[a]PYRENE 7,8-DIOL 9,10-EPOXIDES

Rate data for the reaction of benzo[a]pyrene diol epoxide 80 between pH 4 and 10 in
water fit the equation kobsd ¼ kH ½Hþ  þ ko .101 The changeover in mechanism from
acid-catalyzed hydrolysis to pH-independent hydrolysis occurs at pH  5. Products
of the pH-independent reaction of 80 are tetrols resulting from cis and trans addition
of water to the benzylic epoxide group and ketone resulting from 1,2-hydrogen
migration (Scheme 31). The ketone product is quite unstable, especially in base
solution, and the original yield of ketone product was underestimated.102 The
ratio of cis and trans tetrols (105 and 106) from this pH-independent reaction is
identical with the cis/trans tetrol ratio from the acid-catalyzed hydrolysis of 80. An
282 D.L. WHALEN

OH OH
O + HO HO
HO ks
k1
+
pH 8 H2O HO
HO HO HO
OH OH OH
OH
80 82 (55%) 105 (55%) 106 (5%)
k2
k3[HO-]

HO
OH
107 (40%)

Scheme 31.

intermediate leading to tetrol products is trapped, subsequent to its rate-limiting

formation, by azide ion.89 This intermediate is presumed to be carbocation 82, the
same intermediate formed in the acid-catalyzed hydrolysis of 80, since it leads to the
same ratio of cis and trans tetrols. Ketone 107 is formed from either attack of
hydroxide ion on carbocation 82, coupled with 1,2-hydrogen migration (k3[HO
]),
or from a separate reaction pathway (k2) by either a concerted mechanism or a
mechanism that involves a zwitterion intermediate. Attack of hydroxide ion on 82 as
a pH-independent route leading to ketone can be ruled out on the following basis.
The rate of this k3 route would have to be comparable to or greater than the rate of
the ks route. The value of ks is known to be  2  108 s
1 , so k3[HO
] would have to
be comparable in magnitude. However, at pH 6–11 the value of k3 would have to
exceed the diffusional limit of 1010 M
1 s
1. Yet in this pH range, the yield of
ketone is 40%. Carbocation 82 is therefore not an intermediate in the formation of
ketone 107. The pH-independent reaction of diol epoxide 80 is therefore a com-
bination of two nonintersecting reaction pathways, one stepwise pathway leading to
tetrols via a carbocation intermediate and a second pathway (k2) leading to ketone.
The mechanism of carbocation formation in the pH-independent reaction of 80 is
especially interesting. A neutral water molecule must serve at the proton source, i.e.
a general acid, in the carbocation-forming reaction. Concerted general acid catalysis
is to be expected if the pKa of the catalyst is between the pKa of the substrate site and
the pKa of the final substrate site (Jenck’s rule for concerted catalysis).103 General
acid-catalyzed formation of carbocation 82 by acids with pKa less than that of 82
(13) is in accord with Jenck’s rule and supports a concerted mechanism.89 How-
ever, the pKa of water is higher than the pKa of carbocation 82, and Jenck’s rule is
violated if water acts as a general acid in a concerted mechanism as outlined in
Scheme 32. A concerted mechanism for this reaction is enforced, however, if the
zwitterion 108 is too unstable to exist.104 If the reaction is concerted, a physical
process such as ion dissociation, not proton transfer, must occur in the rate-limiting
MECHANISMS OF HYDROLYSIS AND REARRANGEMENTS OF EPOXIDES 283

HO
H
O + -O +
HO
concerted + HO-
HO HO HO
HO HO HO

80 82 108

Scheme 32.

step. In this sense, water is not a ‘‘true general acid catalyst’’ in the pH-independent
reaction of 80.
The pH-independent reaction of diol epoxide 81 is quite different from that of diol
epoxide 80, although their chemical structures are similar. Subtle differences in
conformation clearly are sufficient to cause different pH-independent mechanisms.
Whereas one of the pH-independent reaction pathways of 80 involves a carbocation
intermediate, carbocation 83 cannot be detected in the pH-independent reaction of
81.89 The mechanism of the diol-forming reactions in the pH-independent reactions
of 81 are not clear, but may involve concerted reactions of 81 with solvent.

SUMMARY OF PH-INDEPENDENT MECHANISMS

The pH-independent reaction of a given epoxide may follow one or more of a

number of possible mechanisms. For simple alkyl-substituted primary and secondary
epoxides, nucleophilic addition of a water molecule, concerted with epoxide C–O
bond breaking, occurs. The pH-independent reactions of simple vinyl- and phenyl-
substituted epoxides also follow this mechanism. When epoxide C–O bond breaking
yields a stabilized carbocation, a number of other reaction mechanisms are possible.
Water may act as a proton donor in converting some epoxides to stabilized carbo-
cations, which react with solvent in a second step to form cis and trans diols. Com-
peting with diol-forming reactions of epoxides are reactions involving 1,2-hydrogen
migrations that yield carbonyl compounds. Certain epoxides also undergo rear-
rangement to isomeric epoxides by mechanisms outlined in the following section.

9 Epoxide isomerization accompanying pH-independent reactions

8,9-Indane oxide (109) undergoes a pH-independent reaction to form 4-indanol

(115), and part of this product is formed by a 1,2-alkyl migration.96 However, most
of the 4-indanol is formed as a result of the isomerization of 8,9-indane oxide to 4,9-
indane oxide (111) via an ‘‘oxygen walk.’’96a,b A minor amount of 5-indanol (116) is
284 D.L. WHALEN

O- O-
O O -O
+
O
+ +
109 110 111 112 113 114

OH
HO

115 116

Scheme 33.

O O- D
H+ H CH3O H H
H D D H +
H H D
O-
O
CH3O CH3O CH3O
117 118 119 120

Scheme 34.

formed as a result of an additional ‘‘oxygen walk’’ in which 4,9-indane oxide (111)

rearranges to 4,5-indane oxide (113). Scheme 33 outlines the mechanism, involving
zwitterions 110, 112 and 114, proposed to explain the observed products. Zwitter-
ions 110, 112 and 114 must be very short-lived intermediates, or the interconversions
of 109, 111 and 116 might also take place by concerted suprafacial 1,5-sigmatropic
rearrangements, which are thermally allowed.105
The pH-independent reaction of p-methoxystyrene oxide yields p-methoxypheny-
lacetaldehyde as a major product and p-methoxystyrene glycol as a minor prod-
uct.106 If deuterium is substituted for hydrogen at the trans-b position and the
reaction is followed by 1H NMR, scrambling of deuterium between the trans-b and
cis-b positions during the course of the reaction is observed (Scheme 34). The
mechanism must involve benzylic C–O bond breaking, rotation about the Ca–Cb
bond via structures 118 and 119, and epoxide ring closure to form 120 at a rate
somewhat faster than the final products, p-methoxyphenylacetaldehyde and p-met-
hoxystyrene glycol, are formed. The rearrangement of 117 to 120 is either stepwise
via a very short-lived intermediate zwitterion, or it is concerted.
Another example of an epoxide isomerization occurring by a similar mechanism is
the pH-independent reaction of cis-anethole oxide 121, which yields (p-methoxy-
phenyl)acetone 125 along with threo and erythro 1-(p-methoxyphenyl)-1,2-pro-
panediols 126.107 During the course of the reaction, trans-anethole oxide (124) builds
up to a level that is detected by 1H NMR (Scheme 35). The mechanism for this
reaction involves benzylic C–O bond breaking, rotation about the Ca–Cb bond via
zwitterionic structures 122 and 123, and epoxide ring closure to form trans-anethole
MECHANISMS OF HYDROLYSIS AND REARRANGEMENTS OF EPOXIDES 285

O O- H CH3O
H H H + H + CH3 H CH3
H
α β CH CH3
H
3 O-
O
CH3O CH3O CH3O
121 122 123 124

~H H2O

O OH
CH3O CH2CCH3 CH3O CHCHCH3
OH
125 126

Scheme 35.

oxide 124 in a mechanism similar to that outlined in Scheme 34. Whereas zwitterion
structures 118 and 119 in Scheme 34 are equivalent except for deuterium labeling,
structures 122 and 123 in Scheme 35 have very different steric interactions.
The acid-catalyzed hydrolyses of both cis-anethole oxide (121) and trans-anethole
oxide (124) yield identical product mixtures of 20% erythro and 80% threo 1-(p-
methoxyphenyl)-1,2-propanediols, suggesting that there is a common benzylic car-
bocation intermediate (127) and common product-forming steps.108 These results
indicate that rotation about the Ca–Cb bond of the carbocation intermediate 127 is
faster than attack of water on the carbocation. However, the ratio of diol and ketone
products from the pH-independent reaction of cis-anethole oxide is very different
than that from trans-anethole oxide, so rotation about the Ca–Cb bond in the
transformation of 121 to 124 is not rapid compared to the rates of ketone and diol
product formation.

OH
H +
C CHCH3
α β
CH3O
127

The rates and products of the reaction of 121 were compared with the rates and
products of the reaction of b-deuterium-labeled cis-anethole oxide 121-d.107 The
reaction pathway to ketone 125 involves a 1,2-hydrogen migration, and this step
should exhibit a primary kinetic isotope effect of 2–3. The diol-forming step,
however, does not involve a 1,2-hydrogen migration, and the kinetic isotope effect
on this step should be near unity. Since ketone 125 is the major product from the
pH-independent reaction of 121, then if the reaction of 121 to form 125 is either
concerted or occurs via an intermediate 122/123 in a rate-limiting step, then the
observed kinetic isotope effect kH D H D
o /ko should be  2. However, the value of ko /ko
was measured to be 1.06, which is too much small for rate-limiting hydrogen mi-
gration in which the hydrogen is substantially migrated at the transition state.
286 D.L. WHALEN

If there is a kinetic isotope effect on the ketone-forming step but not the diol-
forming step, there will be a partitioning isotope effect that would result in a decrease
in ketone yield relative to diol yield in the reaction of 121-bd. However, the relative
yields of ketone and diol products from the pH-independent reactions of both 121
and 121-bd are the same. Therefore, ketone and diol products from the pH-inde-
pendent reaction of 121 must be formed from nonintersecting reaction pathways.
The absence of both kinetic and partitioning deuterium isotope effects was ration-
alized by a mechanism in which benzylic C–O bond breaking is followed by rotation
about the Ca–Cb bond either in a clockwise or in a counterclockwise direction.
Rotation of this bond in one direction must give one product mixture, and rotation
in the other direction must give a different product mixture. In other words, epoxide
ring opening and rotation of the Ca–Cb bond in a given direction commits to the
formation of a given product mixture. Thus, the hydrogen migration step will not be
rate limiting. The rearrangement pathways for the isomerization of 121 to 124 and
ketone product may be examples of asynchronous concerted reactions.100

10 Benzylic epoxides that exhibit complicated pH-rate profiles

PRECOCENE I 3,4-OXIDE

Several highly reactive benzylic epoxides show more complicated pH-rate profiles
that indicate a change in mechanism or rate-limiting step in the intermediate pH
range. Instead of a simple biphasic profile indicating the presence of acid-catalyzed
hydrolysis at low pH and pH-independent hydrolysis at intermediate pH for the
reaction of precocene-I 3,4-oxide (76), a more complicated profile with a negative
inflection point at pH  10 is observed (Fig. 4).86 There is also a nonlinear de-
pendence of rate on buffer concentration for buffers with pK a 47:5. These obser-
vations were attributed to a change in rate-limiting step from epoxide ring opening
at low pH (Regions A and B in Fig. 4) and low buffer concentrations to rate-limiting
carbocation capture at higher pH and higher buffer concentrations of those buffers
with pK a 47:5.
In Region A of Fig. 4, there is rate-limiting carbocation formation by the reaction
of 76 with H+ to yield 77 (Scheme 36). The carbocation is then captured by water to
yield cis and trans diols 78 and 79. In Region B, reaction of 76 yields 20% of ketone
128, in addition to diols 78 and 79. An intermediate in the diol-forming reaction
is captured, subsequent to its rate-limiting formation, by acetylhydrazine. Rate-
limiting reaction of 76 with water to form carbocation 77, which is the intermediate
captured by acetylhydrazine, was proposed (Scheme 37, k3 4k
2 [HO
]). This mech-
anism is similar to that proposed for the pH-independent reaction of benzo[a]pyrene
diol epoxide 80 (Section ‘‘Benzylic epoxides and arene oxides’’). At pH49, however,
k3 ok
2 [HO
], and carbocation formation becomes reversible. In Region D of Fig.
4, capture of carbocation 77 by hydroxide or its kinetic equivalent becomes rate
limiting. In Region D of Fig. 4, ketone 128 is the major product (75%).
MECHANISMS OF HYDROLYSIS AND REARRANGEMENTS OF EPOXIDES 287

Fig. 4 Plot of log kobsd (solid circles) versus pH for the hydrolysis of precocene I oxide.
Reproduced with permission from Ref. 86. Copyright 1988 American Chemical Society.

CH3O O CH3 CH3O O CH3 CH3O O CH3 CH3O O CH3

CH3 CH3 CH3 CH3
k1[H+] H2O +
OH OH
O + OH
r.d.s. OH OH
76 77 78 79

Scheme 36.

CH3O O CH3
CH3
CH3O O CH3 OH
CH3 k2 CH3O O CH3 k3 + k4[HO-] OH
CH3 + OH-
78 + 79
O k-2[HO-] + OH +
76 77 CH3O O CH3
CH3
O
128

Scheme 37.
288 D.L. WHALEN

A number of very interesting conclusions are arrived at in this study.86 (1) Re-
action of hydroxide ion as a base on carbocation 77 to form epoxide (k
2[HO
]) is
energetically more favorable than its reaction with 77 as a nucleophile to form diols.
(2) The reaction pathway for formation of ketone 128 is completely separate from
the stepwise mechanism for diol formation. (3) The observed kinetic deuterium iso-
tope effect for reaction of 76-b-d at pH 11 is 2.15; the kinetic deuterium isotope
effect on the 1,2-hydrogen migration is estimated to be 4. Hydrogen migration
must be occurring at the transition state for ketone formation.

BENZO[a]PYRENE 7,8-DIOL 9,10-EPOXIDE (80)

A more complicated pH-rate profile is also observed for the hydrolysis reactions of
benzo[a]pyrene diol epoxide epoxide 80, and is shown in Fig. 5.102 This profile shows
Regions A–D that are similar to those for reaction of precocene I oxide 76 (Fig. 4),
except that Region B reaches a full plateau that extends from pH 5 to 9 in water. The
interpretation of this pH-rate profile is essentially the same as the interpretation of
the profile for hydrolysis of precocene I oxide (Fig. 4). The pH-independent reaction
of 80 in Region B (discussed in detail in Section ‘‘Benzylic epoxides and arene
oxides’’) yields 60% tetrols in a stepwise mechanism involving a carbocation inter-
mediate and 40% ketone from a completely separate pathway (Scheme 31). The
negative inflection of the profile at pH 10–11.5 indicates that hydroxide ion reacts as
a base with the intermediate carbocation to reform diol epoxide 80 and thus slow the
reaction rate. There is a corresponding increase in the yield of ketone 107 at pH411.

Water
A
5:95 Dioxane-Water
-1
1:9 Dioxane-Water
log k obsd , s- 1

B E
C
-2 D

-3
4 6 8 10 12

pH

Fig. 5 Plots of log kobsd versus pH for the reaction of diol epoxide 80 in water, 5:95 dioxane/
water (v/v), 0.1 M NaClO4, 25 1C. Reproduced with permission from Ref. 102. Copyright
2001 American Chemical Society.
MECHANISMS OF HYDROLYSIS AND REARRANGEMENTS OF EPOXIDES 289

The reaction of hydroxide ion with carbocation 82 is expected to occur at the

diffusion-controlled limit because proton transfer from the b-OH is thermodynam-
ically favorable, and the activation barrier for collapse of the resulting zwitterion
(108, Scheme 32) to reform the epoxide ring should be absent or very small. A
concerted mechanism for this reaction is ‘‘enforced’’ if zwitterion 108 is too unstable
to exist as an intermediate.104 The attack of hydroxide ion as a base on 82 as shown
in Scheme 38 must be energetically more favorable than attack as a nucleophile to
yield tetrol product. The exact yield of tetrol from direct combination of hydroxide
ion with carbocation 82 could not be estimated, but is negligible or very small.
The reactions of stable carbocations with water are generally base catalyzed,109,110
and their reactions with hydroxide ion are slower than their reactions with azide ion
and sulfhydryl ions. The less favorable reaction of hydroxide ion with carbocations
has been attributed to the fact that deprotonation of a water molecule by hydroxide
ion is not a thermodynamically favorable reaction, and activation energy to generate
a desolvated hydroxide ion is required.110 These factors would also account for the
less favorable reaction of hydroxide ion as a nucleophile with 82 to form tetrols
instead of as a base to bring about epoxide ring closure.
At pH412 (Region E, Fig. 5), the rate of reaction of 80 increases due to the
incursion of a second-order reaction of 80 with hydroxide ion. The ratio of trans/cis
tetrols 106/105 (Scheme 31) increases somewhat, indicating that hydroxide ion at-
tacks diol epoxide 80 to give some trans tetrol product.102 However, the yield of
trans tetrol does not increase substantially, indicating that the reaction of hydroxide
ion with 80 may form some other product(s) such as unstable ketone 107.
The pH-rate profile for diol epoxide 81 does not exhibit a negative inflection in the
intermediate pH region similar to that for diol epoxide 80, and its pH-independent
reaction does not proceed via an intermediate carbocation, even though the car-
bocation formed from reaction of this epoxide with H+ has a sufficient lifetime to be
detected.88 Tetrol products from the pH-independent reaction of 81 must occur by
some other mechanism(s), possibly concerted.
General acid catalysis in the hydrolysis of 81 is quite facile. This reaction, as
discussed in Section ‘‘Benzylic epoxides and arene oxides’’ and shown in Scheme 39,
involves proton transfer to the epoxide oxygen concerted with epoxide C–O bond
breaking to form a carbocation 83. For primary ammonium ions with pK a o  8,
only the acid form of the amine is reactive, and carbocation formation is irreversible,

HO-
HO
H H
O O
+

HO
HO
OH
OH
82 80

Scheme 38.
290 D.L. WHALEN

O HO
k1[RNH3+] k2 + k4[RNH2]
+ Products
RNH3 + + + RNH2
HO k-1[RNH2] HO
OH OH
81 83

Scheme 39.

i.e. 83 reacts with water (k2) faster than it reacts with RNH2 to reform epoxide
(k
1[RNH2]) or with RNH2 to form product(s) (k4[RNH2]. However, for ammo-
nium ions with pK a 4  8, the base form of the buffer reacts with 81 as a nuc-
leophile and with carbocation 83 as a base to reform epoxide 81 (k
1[RNH2]). For
these amine buffers, the rate at which RNH2 reacts with carbocation 83 to reform
epoxide (k
1[RNH2]) exceeds the rate that it reacts with 83 as a nucleophile or
general base to form other products (k4[RNH2), and, under these conditions, car-
bocation formation is partially reversible. Thus, although there is a change in
mechanism and rate-limiting step for the general acid-catalyzed hydrolysis of diol
epoxide 80 by water as a function of pH and not for diol epoxide 81, there is a
change in mechanism and rate-limiting step in the general acid-catalyzed hydrolysis
of diol epoxide 81 by amine buffers as a function of amine basicity.

SPECIFIC EFFECTS OF CHLORIDE ION IN EPOXIDE HYDROLYSIS

The pH-rate profiles for the reactions of a number of epoxides in solutions con-
taining chloride salts show the same general shape as that in Fig. 4 for reaction of
precocene I oxide. For example, indene oxide,62 phenanthrene 9,10-oxide,111
cyclohexadiene oxide,36c cyclopentadiene oxide,36c styrene oxide,36c 7-methyl-
benzo[a]anthracene 5,6-oxide,112 7,12-dimethylbenzo[a]anthracene 5,6-oxide112 and
diol epoxide 81113 all show more complicated pH-rate profiles resembling Fig. 4 or 5
when their rates are determined in solutions containing chloride ion, but not when
their rates are determined in solutions containing sodium perchlorate, a non-
nucleophilic salt. The pH-rate profiles for reactions of these epoxides in solutions
containing chloride ion are more complicated due to the nucleophilicity of chloride
ion. Bromide ion and iodide ion, being more nucleophilic than chloride ion, have
even greater effects on shapes of the pH-rate profiles of epoxides that are susceptible
to reaction with nucleophiles.
Scheme 40 shows the mechanism of reaction of diol epoxide 81 in solutions
containing chloride, bromide or iodide salts.113 At sufficiently low pH, kH[H+] is the
dominant term of the rate equation, and 81 reacts with H+ to form carbocation 83.
Carbocation 83 partitions between reaction with water (ks) and capture by halide ion
(k2[X
]). Iodide ion reacts with 83 at or near the diffusion-controlled limit, and is
3–4 times more reactive than bromide ion and 28 times more reactive than chloride
MECHANISMS OF HYDROLYSIS AND REARRANGEMENTS OF EPOXIDES 291

OH OH
O HO + HO
O HO ko + kOH[OH-] kH[H+] ks
+ H2O
HO HO HO HO HO
HO HO HO HO HO
107 129 81 83 129
(6% cis, 94% trans)
k1[X-] k-1 k2[X-] k-2

X X OH
-O HO
1/Ka HO k3
H2O
HO HO
HO HO
HO HO
130 131 129
(35% cis, 65% trans)

Scheme 40.

X
O HO k3
+ X- + H+ tetrols (129)
HO HO
HO HO

81 131

Scheme 41.

ion. Thus, chloride ion reacts with carbocation 83 within one or two orders of
magnitude of the diffusional limit.
At somewhat higher pH, direct nucleophilic attack of halide ion on diol epoxide
81 (k1[X
]) becomes important, and a rate plateau is reached in which this term is
the main one. If the pH is sufficiently low, the pH-dependent equilibrium between
halohydrin 131 and diol epoxide 81 (shown in Scheme 41) favors halohydrin, which
reacts via an SN1 reaction (k3) to form tetrols 129. As the pH is increased, however,
the pH-dependent equilibrium between halohydrin 131 and diol epoxide 81 shifts to
favor diol epoxide, and there is a resulting rate decrease that gives an inflection point
in the pH-rate profile at intermediate pH that resembles those in the profiles in Figs
4 and 5. Rate and product observations are rationalized by the mechanism of
Scheme 40, and comparable mechanisms can be expected for reactions of other
epoxides susceptible to reaction with nucleophiles.

11 Partitioning of hydroxycarbocations

The rates for reaction of highly stabilized carbocations such as triarylmethyl car-
bocations with water are sufficiently slow so that their reactions can be measured
292 D.L. WHALEN

directly.109,110 The ratios of their rate constants with azide ion and with water are
relatively constant, and are approximately 106–107 M
1. Other less stable carbo-
cations have reduced kaz/ks ratios, and these reduced ratios are interpreted to mean
that these carbocations react with azide ion at the diffusional limit, whereas the
reactions of more highly stabilized carbocations with solvent are activation limit-
ed.43,114,115 For a series of ring-substituted 1-phenylethyl carbocations, values of kaz/
ks are between 1 and 130 M
1.43 It was concluded that their reactions with azide ion
are diffusion limited. From direct measurements of these partitioning ratios, the
assumption that kaz ¼ 5  109 M
1 s
1 and some extrapolation of structure–activity
relationships, values of ks for reactions of 1-phenylethyl carbocations with solvent
were calculated and varied between  1013 s
1 for 1-(4-nitrophenyl)ethyl carbo-
cation and  2  103 s
1 for 1-(4-dimethylaminophenyl)ethyl carbocation. The val-
ue of ks for reaction of 1-(4-methoxyphenyl)ethyl carbocation was calculated to be
4  107 s
1 .
For benzo[a]pyrene benzylic carbocations (82 and 83)88,89 and 4-methoxyphenyl-
stabilized carbocations (52a, 54b, 66b, 68b and 127),59,63,70,108 values of kaz/ks are in
the range 25–600 M
1. It is reasonable to assume that these carbocations also react
with azide ion at the diffusional limit of 1010 M
1 s
1. The rate constants for
reaction of these carbocations with water (ks) are estimated to be  ð8  106 Þ to
ð2  108 Þ s
1 , which correspond to energies of activation of 6.5–8.0 kcal mol
1.
Benzo[a]pyrene 7,8-diol 9,10-epoxides (80 and 81) and tetrahydronaphthalene epox-
ide each has two optimized conformations similar to structures 132 and 134, which
are shown in Scheme 42. The carbocation formed from reaction of each of these

R2 OH R2
O kH1[H+] ks
R1 R1 Diol Mixture A
R1 H + R1
H H
H R2 R2
132 133
a; R1 = R2 = H
b; R1 = OH, R2 = H
c; R1 = H, R2 = OH

k2 k−2

R
R2 1 R1
kH2[H+] R2 k 's
O R2 Diol Mixture B
R2
R1 OH
H H + R1
H
H
134 135

Scheme 42.
MECHANISMS OF HYDROLYSIS AND REARRANGEMENTS OF EPOXIDES 293

epoxides with H+ also has two conformations, which are given by structures 133
and 135. Since axial attacks of solvent on carbocation conformations 133 and 135
are energetically favored, each will yield a different ratio of cis and trans diols or
tetrols from its reaction with water. When the kH1 pathway is faster than the kH2
pathway for reaction of epoxide with H+, then the partitioning of carbocation
conformation 133 and the relative stabilities of carbocation conformations 133 and
135 all play important roles in determining cis/trans product ratios. The value of ks
will depend on the ability of the aryl group to stabilize positive charge at the benzylic
carbon, whereas the value of k2 will depend on the size and nature of substituents R1
and R2. For systems in which the energy barrier for conformational interconversion
of carbocations 133 and 135 is small relative to the energy barriers for attack of
water on either carbocation, then the relative yields of diol or tetrol products are
determined only by the difference in energy of the transition states leading to prod-
ucts (ks versus k0 s steps). When the energy barrier for conformational interconver-
sion of carbocations 133 and 135 is large relative to the energy barriers for attack of
water on either carbocation, then the relative yields of diol or tetrol products
are determined only by the difference in energy of the transition states leading to
carbocation intermediates 133 and 135 (kH1 versus kH2 steps).
Conformation 132a ðR1 ¼ R2 ¼ HÞ is significantly more stable than 134a, pre-
sumably due to fewer eclipsing interactions in the angular ring.67 1H NMR evidence
also indicates that diol epoxide conformation 132b (R1 ¼ OH; R2 ¼ H) is substan-
tially more stable than conformation 134b.116 In order to explain the observations
that acid-catalyzed hydrolysis of diol epoxide 81 and solvolysis of its trans 9,10-
chlorohydrin yields different cis/trans tetrol mixtures, even though each reaction
proceeds via a carbocation intermediate with the same connectivity, it was proposed
that (1) the rate at which diol epoxide conformation 132b reacts with H+ to form
133b is greater than the rate at which diol epoxide conformation 134b reacts with
H+ to form 135b, and (2) the energy barrier for interconversion of carbocation 133b
to 135b is large relative that for reaction of 133b with water ðks 4k2 Þ.117 Therefore,
products from the acid-catalyzed hydrolysis of 81 result mainly from reaction of
133b with solvent, even though quantum chemical calculations suggest that carbo-
cation conformations 133b and 135b have comparable stabilities.117
Product studies of the acid-catalyzed hydrolysis of diol epoxide 80 are somewhat
more difficult to interpret. Epoxide conformation 132c (R1 ¼ H; R2 ¼ OH) is es-
timated to be somewhat more stable than 134c,116 and carbocation conformation
135c is calculated to be significantly more stable than 133c.118 The stereochemistry
of diol products from reaction of 80 suggest that all or most of the tetrol products
formed are from reaction of 135c with solvent. One possible explanation of the
results is that carbocation conformation 133c is formed faster than 135c and
the barrier for interconversion of 133c and 135c is small relative to the energy barrier
for attack of water on 133c. However, the energy barrier for interconversion of 133c
and 135c (R1 ¼ H; R2 ¼ OH) might be similar to that for interconversion of
133b and 135b (R1 ¼ OH; R2 ¼ H) because the substituent groups are the same,
and this energy barrier may be larger than the barriers for attack of solvent on
either carbocation conformation. A second possible explanation for the observed
294 D.L. WHALEN

results is that the transition-state energy of the kH2 pathway for reaction of 80 via
conformation 134c leading directly to the more stable carbocation conformation
135c is lower than that for the kH1 pathway, and tetrol products are formed from
reaction of this carbocation with solvent. Additional research is needed to clarify
this mechanism.
In summary, the reactions of aryl epoxides with H+ are generally irreversible.
In some cases where the energy barrier for conformational interconversion of the
intermediate carbocation is sufficiently high, then the relative rates of reaction of
each epoxide conformation (kH1[132] versus kH2[134]) determine the diol product
ratio, since each reaction pathway commits the products to be formed from a
given intermediate carbocation. However, when interconversion of the carbocation
conformations is rapid relative to the rate of carbocation capture by solvent,
then the product distribution is determined only by the transition-state energy dif-
ference of the product-forming steps (ks versus k0 s pathways, Curtin–Hammett
principle).119 The partitioning of a carbocation between reaction with solvent and
ring inversion to form a second carbocation conformation is expected to vary sig-
nificantly with both the ability of the aryl group to stabilize positive charge on
the benzylic carbon and the magnitude of the energy barrier for conformational
interconversion.

12 Overall summary

The mechanism of hydrolysis of a given epoxide is a function of many variables. The

pH of the solution will determine whether a given epoxide will react via an acid-
catalyzed reaction, a pH-independent reaction or a base-catalyzed reaction. The
relative rate of each kinetically distinct reaction will vary with the epoxide structure.
The relative rate of the acid-catalyzed reaction increases with the abilities of the
substituent groups to stabilize positive charge in the transition state for epoxide ring
opening. The pH-independent reaction has a number of completely different mech-
anisms as discussed in Section 8, and the mechanism followed will also depend on
the abilities of substituent groups to stabilize positive charge in the transition state
and, in many cases, an intermediate. The lifetimes of the intermediates in the acid-
catalyzed and pH-independent reactions of epoxides play crucial roles in determin-
ing the reaction mechanisms, and both conformational effects and transition-state
effects are important factors in affecting the stereochemistry of the hydrolysis
products.

Acknowledgments

I would like to acknowledge Prof. John P. Richard and Prof. Ram S. Mohan for
reading this chapter and making many valuable suggestions.
MECHANISMS OF HYDROLYSIS AND REARRANGEMENTS OF EPOXIDES 295

References

1. Rosowsky, A. (1964). In Heterocyclic Compounds with Three- and Four-Membered Rings.

Weissberber, A. (ed.), Part 1, pp. 1–523, Wiley-Interscience, New York
2. Buchanan, J.G. and Sable, H.Z. (1972). In Selective Organic Transformations,
Thyagarajan, B.S. (ed.), vol. 2, pp. 1–95, Wiley-Interscience, New York
3. Cross, A.D. (1960). Quart. Rev. 14, 317–335
4. Nicolaou, D.C. and Smith, A.L. (1992). Acc. Chem. Res. 25, 497–503
5. Bierl, B.A., Beroza, M. and Collier, C.W. (1970). Science 170, 87–89
6. Persoons, C.J., Verwiel, P.E.J., Ritter, F.J., Talman, E., Nooijen, P.J.F. and Nooigen,
W.J. (1976). Tetrahedron Lett. 24, 2055–2058
7. Dahm, K.H., Trost, B.M. and Roller, H. (1967). J. Am. Chem. Soc. 89, 5292–5294
8. El Masri, A.M., Smith, J.N. and Williams, R.T. (1958). J. Biochem. 68, 199–204
9. Boyland, E., Ramsay, G.S. and Sims, P. (1961). J. Biochem. 78, 376–384
10. Wogan, G.N. (1992). Cancer Res. 52, 2114–2118
11. Olson, J.H., Dragsted, L. and Autrup, H. (1988). Br. J. Cancer 58, 392–396
12. Hayes, R.B., Van Nieuwenhuize, J.P., Raatgever, J.W. and ten Kate, F.J.W. (1984).
Food Chem. Toxicol. 22, 39–43
13. Ivers, R.S., Coles, B.F., Raney, K.D., Thier, R., Guengerich, F.P. and Harris, T.M.
(1994). J. Am. Chem. Soc. 116, 1603–1609
14. Dipple, A., Moschel, R.C. and Bigger, C.A.H. (1984). In Chemical Carcinogens, Searle,
C.E. (ed.) (2nd edn), vol. 1, ACS Monograph 182, pp. 41–163. American Chemical
Society, Washington, DC
15. Harvey, R.G. (ed.) (1985) Polycyclic hydrocarbons and carcinogenesis, ACS Symposium
Series 283. American Chemical Society, Washington, DC
16. Thakker, D.R., Yagi, H., Akagi, H., Koreeda, M., Lu, A.Y.H., Levin, W., Wood, A.W.,
Conney, A.H. and Jerina, D.M. (1977). Chem. Biol. Interact. 16, 281–300
17. For historical reviews, see (a) Phillips, D.H. (1983). Nature 303, 468–472; (b) Rubin, H.
(2001). Carcinogenesis 22, 1903–1930
18. Luch, A., Platt, K.L. and Seidel, A. (1998). Carcinogenesis 19, 639–648
19. Greenberg, A. and Liebman, J.F. (1978). Strained Organic Molecules, p. 281. Academic
Press, New York
20. Sawicka, D., Wilsey, S. and Houk, K.N. (1999). J. Am. Chem. Soc. 121, 864–865
21. Wolk, J.L., Sprecher, M., Basch, H. and Hoz, S. (2004). Org. Biomol. Chem. 2,
1065–1069
22. Brønsted, J.N., Kilpatrick, M. and Kilpatrick, M. (1929). J. Am. Chem. Soc. 51, 428–461
23. Long, F.A. and Pritchard, J.G. (1956). J. Am. Chem. Soc. 78, 2663–2667
24. Long, F.A. and Pritchard, J.G. (1956). J. Am. Chem. Soc. 78, 2667–2670
25. (a) Bunnett, J.F. (1961). J. Am. Chem. Soc. 83, 4956–4967; (b) Bunnett, J.F. (1961). Ibid.
83, 4968–4973; (c) Bunnett, J.F. (1961). Ibid. 83, 4973–4977; (d) Bunnett, J.F. (1961).
Ibid. 83, 4978–4983; (e) Bunnett, J.F. and Olsen, F.P. (1966). Can. J. Chem. 44,
1899–1916
26. Grunwald, E., Heller, A. and Klein, F.S. (1957). J. Chem. Soc., 2604–2613
27. Wohl, R.A. (1974). Chimia 28, 1–5
28. Toteva, M.M. and Richard, J.P. (1996). J. Am. Chem. Soc. 118, 11434–11445
29. (a) Billing, G.D. and Mikkelsen, K.V. (1996). Molecular Dynamics and Chemical Kin-
etics, p. 108, Wiley, New York; (b) Castner Jr., E.W., Maroncelli, M. and Fleming, R.
(1987). J. Chem. Phys. 86, 1090–1097
30. Richard, J.P. (1992). J. Org. Chem. 57, 625–629
31. Pritchard, J.G. and Siddiqui, I.A. (1973). J. Chem. Soc. Perk. II, 452–457
32. Archer, I.V.J., Leak, D.J. and Widdowson, D.A. (1996). Tetrahedron Lett. 37,
8819–8822
33. Henbest, H.B., Smith, M. and Thomas, A. (1958). J. Chem. Soc. 3293–3298
296 D.L. WHALEN

34. Parker, R.E. and Isaacs, N.S. (1959). Chem. Rev. 59, 737–799
35. Data were taken from Streitwieser, A. (1962). Solvolytic Displacement Reactions, p. 12.
McGraw-Hill, New York
36. (a) Whalen, D.L. (1973). J. Am. Chem. Soc. 95, 3432–3434; (b) Whalen, D.L. and Ross,
A.M. (1974). J. Am. Chem. Soc. 96, 3678–3679; (c) Ross, A.M., Pohl, T.M., Piazza, K.,
Thomas, M., Fox, B. and Whalen, D.L. (1982). J. Am. Chem. Soc. 104, 1658–1665
37. (a) Bartlett, P.D. and Ross, S.D. (1948). J. Am. Chem. Soc. 70, 926–929; (b) Kadesch,
R.G. (1946). Ibid. 68, 41–45
38. DeWolfe, R.H. and Young, W.G. (1956). Chem. Rev. 56, 753–901
39. Anet, F.A. and Yavari, I. (1978). J. Am. Chem. Soc. 100, 7817–7819
40. (a) Audier, H.E., Dupin, J.F. and Jullien, J. (1968). Bull. Soc. Chim. Fr. 9, 3850–3856; (b)
Hanzlik, R.P. and Westkaemper, R.B. (1980). J. Am. Chem. Soc. 102, 2464–2467; (c)
Jacober, S.P. and Hanzlik, R.P. (1986). Ibid. 108, 1594–1597
41. Blumenstein, J.J., Ukachukwu, V.C., Mohan, R.S. and Whalen, D.L. (1993). J. Org.
Chem. 58, 924–932
42. Biggs, J., Chapman, N.B., Finch, A.F. and Wray, V. (1971). J. Chem. Soc. B, 55–63
43. Richard, J.P., Rothenberg, M.E. and Jencks, W.P. (1984). J. Am. Chem. Soc. 106,
1361–1372
44. (a) Tsuno, Y., Kusuyama, Y., Sawada, M., Fuji, T. and Jukawa, Y. (1975). Bull. Soc.
Chem. Jpn. 48 (1), 3337–3346; (b) Richard, J.P. and Jencks, W.P. (1984). J. Am. Chem.
Soc. 106, 1383–1396
45. Berti, G., Bottari, F., Ferrarini, P.L. and Macchia, B. (1965). J. Org. Chem. 30,
4091–4096
46. Dupin, C. and Dupin, J.-F. (1970). Bull. Soc. Chim. Fr. 11, 249–251
47. Lin, B. and Whalen, D.L. (1994). J. Org. Chem. 59, 1638–1641
48. Biggs, J., Chapman, N.B. and Wray, V. (1971). J. Chem. Soc. B 71–74
49. Whalen, D.L., unpublished results
50. Richard, J.P. and Jencks, W.P. (1982). J. Am. Chem. Soc. 104, 4689–4691
51. Audier, H.E., Dupin, J.F. and Jullien, J. (1966). Bull. Soc. Chim. Fr. 7, 2811–2816
52. Audier, H.E., Dupin, J.F. and Jullien, J. (1968). Bull. Soc. Chim. Fr. 9, 3850–3855
53. Mohan, R.S. and Whalen, D.L. (1993). J. Org. Chem. 58, 2663–2669
54. Laird, R.M. and Parker, R.E. (1961). J. Am. Chem. Soc. 83, 4277–4281
55. Audier, H.E., Dupin, J.F. and Jullien, J. (1968). Bull. Soc. Chim. Fr. 9, 3844–3850
56. Battistini, C., Balsamo, A., Berti, G., Crotti, P., Macchia, B. and Macchia, F. (1974). J.
Chem. Soc. Chem. Commun., 712–713
57. Battistini, C., Crotti, P., Donatella, D. and Macchia, F. (1979). J. Org. Chem. 44,
1643–1647
58. Crotti, P., Dell’Omodarme, G., Ferretti, M. and Macchia, F. (1987). J. Am. Chem. Soc.
109, 1463–1469
59. Doan, L., Bradley, K., Gerdes, S. and Whalen, D.L. (1999). J. Org. Chem. 64, 6227–6234
60. Amyes, T.L., Stevens, L.W. and Richard, J.P. (1993). J. Org. Chem. 58, 6057–6066
61. Sampson, K., Paik, A., Duvall, B. and Whalen, D.L. (2004). J. Org. Chem. 69,
5204–5211
62. Balsamo, A., Berti, G., Crotti, P., Ferretti, M., Macchia, B. and Macchia, F. (1974). J.
Org. Chem. 39, 2596–2598
63. Whalen, D.L. and Ross, A.M. (1976). J. Am. Chem. Soc. 98, 7859–7861
64. Richard, J.P., Amyes, T.L. and Toteva, M.M. (2001). Acc. Chem. Res. 34, 981–988
65. Whalen, D., Doan, L., Poulose, B., Friedman, S., Gold, A., Sangaiah, R., Ramesha, A.,
Sayer, J. and Jerina, D.M. (2000). Polycycl. Aromat. Hydrocarb. 21, 43–52
66. Becker, A.R., Janusy, J.M. and Bruice, T.C. (1979). J. Am. Chem. Soc. 102, 5679–5687
67. Gillilan, R.E., Pohl, T.M. and Whalen, D.L. (1982). J. Am. Chem. Soc. 104, 4481–4482
68. Goering, H.L. and Josephson, R.R. (1962). J. Am. Chem. Soc. 84, 2779–2785
MECHANISMS OF HYDROLYSIS AND REARRANGEMENTS OF EPOXIDES 297

69. Sayer, J.M., Yagi, H., Silverton, J.V., Friedman, S.L., Whalen, D.L. and Jerina, D.M.
(1982). J. Am. Chem. Soc. 104, 1972–1978
70. (a) Doan, L. and Whalen, D.L., unpublished results; (b) Chini, M., Crotti, P., Ferretti,
M. and Macchia, F. (1988). Tetrahedron, 44, 2001–2004
71. Swain, C.G. (1951). J. Am. Chem. Soc. 74, 4108–4110
72. Long, F.A. and Paul, M.A. (1957). Chem. Rev. 57, 935–1010
73. Pocker, Y. and Ronald, B.P. (1978). J. Am. Chem. Soc. 100, 3122–3127
74. Fife, T.H. (1972). Acc. Chem. Res. 5, 254–272
75. (a) Fife, T.H. and Jao, L.K. (1968). J. Am. Chem. Soc. 90, 4081–4085; (b) Fife, T.H. and
Brod, L.H. (1970). J. Am. Chem. Soc. 92, 1681–1684; (c) Anderson, E. and Capon, B.
(1969). J. Chem. Soc. B, 1033–1037
76. Anderson, E. and Fife, T.H. (1969). J. Am. Chem. Soc. 91, 7163–7166
77. Fife, T.H. (1967). J. Am. Chem. Soc. 89, 3228–3231
78. Fife, T.H. and Hutchins, H.E.C. (1976). J. Am. Chem. Soc. 98, 2536–2543
79. Mori, A.L. and Schaleger, L.L. (1972). J. Am. Chem. Soc. 94, 5039–5043
80. (a) Dunn, B.M. and Bruice, T.C. (1971). J. Am. Chem. Soc. 93, 5725–5731; (b) Vitullo,
V.P. and Grossman, N.R. (1973). J. Org. Chem. 38, 179–180
81. Kresge, A.J. and Chiang, Y. (1973). J. Am. Chem. Soc. 95, 803–806
82. Whalen, D.L., Weimaster, J.F., Ross, A.M. and Radhe, R. (1976). J. Am. Chem. Soc. 98,
7319–7324
83. Bruice, P.Y., Bruice, T.C., Dansette, P., Selander, H.G., Yagi, H. and Jerina, D.M.
(1976). J. Am. Chem. Soc. 98, 2965–2973
84. Mohan, R.S. and Whalen, D.L., unpublished results
85. Kasperek, G.J. and Bruice, T.C. (1972). J. Am. Chem. Soc. 94, 198–202
86. Sayer, J.M., Grossman, S.J., Adusei-Poku, K.S. and Jerina, D.M. (1988). J. Am. Chem.
Soc. 110, 5068–5074
87. Whalen, D.L., Ross, A.M., Montemarano, J.A., Thakker, D.R., Yagi, H. and Jerina,
D.M. (1979). J. Am. Chem. Soc. 101, 5086–5088
88. Lin, B., Islam, N., Friedman, S., Yagi, H., Jerina, D.M. and Whalen, D.L. (1998). J.
Am. Chem. Soc. 120, 4327–4333
89. Islam, N.B., Gupta, S.C., Yagi, H., Jerina, D.M. and Whalen, D.L. (1990). J. Am.
Chem. Soc. 112, 6363–6369
90. (a) Gupta, S.C., Pohl, T.M., Friedman, S.L., Whalen, D.L., Yagi, H. and Jerina, D.M.
(1982). J. Am. Chem. Soc. 104, 3101–3104; (b) Gupta, S.C., Islam, N., Whalen, D.L.,
Yagi, H. and Jerina, D.M. (1987). J. Org. Chem. 52, 3812–3815
91. (a) Jerina, D.M., Daly, J., Witkop, B., Zaltaman-Nirenberg, P. and Udenfriend, S.
(1968). Arch. Biochem. Biophys. 128, 176–183; (b) Jerina, D.M., Daly, J.W., Witkop, B.,
Zaltaman-Nirenberg, P. and Udenfriend, S. (1970). Biochemistry 9, 147–156; (c) Jerina,
D.M., Daly, J.W. and Witkop, B. (1968). J. Am. Chem. Soc. 90, 6523–6525; (d) Jerina,
D.M., Daly, J.W., Witkop, B., Zaltaman-Nirenberg, P. and Udenfriend, S. (1968). J.
Am. Chem. Soc. 90, 6525–6527; (e) Boyd, D.R., Jerina, D.M. and Daly, J.W. (1970). J.
Org. Chem. 35, 3170–3172
92. Kasperek, G.J. and Bruice, T.C. (1972). J. Am. Chem. Soc. 94, 198–202
93. Boyd, D.R., Daly, J.W. and Jerina, D.M. (1972). Biochemistry 11, 1961–1966
94. Kasperek, G.J. and Bruice, T.C. (1972). J. Chem. Soc. Chem. Comm., 784–785
95. Kasperek, G.J., Bruice, T.C., Yagi, H., Kaubisch, N. and Jerina, D.M. (1972). J. Am.
Chem. Soc. 94, 7876–7882
96. (a) Bruice, P.Y., Kasperek, G.J. and Bruice, T.C. (1973). J. Am. Chem. Soc. 95,
1673–1674; (b) Kasperek, G.J., Bruice, P.Y., Bruice, T.C., Yagi, H. and Jerina, D.M.
(1973). J. Am. Chem. Soc. 95, 6041–6046; (c) Vögel, E. and Günther, H. (1967). Angew.
Chem. Int. Ed. Engl. 6, 385–401
97. Whalen, D.L. and Ross, A.M. (1974). J. Am. Chem. Soc. 96, 3678–3679
298 D.L. WHALEN

98. Gillilan, R.E., Pohl, T.M. and Whalen, D.L. (1982). J. Am. Chem. Soc. 104, 4482–4484
99. Sridharan, S. and Whalen, D. L., unpublished results
100. Coxon, J.M., Maclagan, R.G.A.R., Rauk, A., Thorpe, A.J. and Whalen, D.L. (1997). J.
Am. Chem. Soc. 119, 4712–4718
101. Whalen, D.L., Montemarano, J.A., Thakker, D.R., Yagi, H. and Jerina, D.M. (1977). J.
Am. Chem. Soc. 99, 5522–5524
102. Doan, L., Lin, B., Yagi, H., Jerina, D.M. and Whalen, D.L. (2001). J. Am. Chem. Soc.
123, 6785–6791
103. Jencks, W.P. (1972). J. Am. Chem. Soc. 94, 4731–4732
104. Jencks, W.P. (1980). Acc. Chem. Res. 13, 161–169
105. Woodward, R.B. and Hoffman, R. (1970). The Conservation of Orbital Symmetry.
Verlag Chemie, GmbH, Academic Press
106. (a) Ukachukwu, V.C., Blumenstein, J.J. and Whalen, D.L. (1986). J. Am. Chem. Soc.
108, 5039–5040; (b) Ukachukwu, V.C. and Whalen, D.L. (1988). Tetrahedron Lett. 29,
293–296
107. Mohan, R.S., Gavardinas, K., Kyere, S. and Whalen, D.L. (2000). J. Org. Chem. 65,
1407–1413
108. Mohan, R.S. and Whalen, D.L. (1993). Org. Chem. 58, 2663–2669
109. Ritchie, C.D. and Virtanen, P.O.I. (1972). J. Am. Chem. Soc. 94, 4966–4971
110. Ritchie, C.D. (1972). Acc. Chem. Res. 5, 348–354
111. Whalen, D.L., Ross, A.M., Dansette, P.M. and Jerina, D.M. (1977). J. Am. Chem. Soc.
99, 5672–5676
112. Nashed, N.R., Balani, S.K., Loncharich, R.J., Sayer, J.M., Shipley, D.Y., Mohan, R.S.,
Whalen, D.L. and Jerina, D.M. (1991). J. Am. Chem. Soc. 113, 3910–3919
113. Lin, B., Doan, L., Yagi, H., Jerina, D.M. and Whalen, D.L. (1994). Chem. Res. Toxicol.
11, 630–638
114. Richard, J.P. and Jencks, W.P. (1982). J. Am. Chem. Soc. 104, 4689–4691
115. For a review and discussion of literature data for reactions of carbocations with azide
ion, see: Ta-Shma, R. and Rappoport, Z. (1983). J. Am. Chem. Soc. 105, 6082–6095
116. (a) Yagi, H., Hernandez, O. and Jerina, D.M. (1975). J. Am. Chem. Soc. 97, 6881–6883;
(b) Whalen, D.L., Ross, A.M., Yagi, H., Karle, J.M. and Jerina, D.M. (1978). J. Am.
Chem. Soc. 100, 5218–5221
117. Doan, L., Yagi, H., Jerina, D.M. and Whalen, D.L. (2002). J. Am. Chem. Soc. 124,
14382–14387
118. Doan, L., Yagi, H., Jerina, D.M. and Whalen, D.L. (2004). J. Org. Chem. 69, 8012–8017
119. (a) Curtin, D.Y. (1954). Rec. Chem. Prog. 15, 111; (b) Eliel, E.L. (1962). Stereochemistry
of Carbon Compounds, pp. 151–152, McGraw-Hill, New York