Mechanisms of solvolytic elimination reactions of tertiary substrates: Stereospecific 1,2-elimination reactions

Article abstract of DOI:10.1039/a901296f

Solvolysis of (R,S)-1-chloro-1-(fluoren-9-yl)-2-methylcyclopentane (1-Cl) or the analogous 3,5-dinitrobenzoate ester 1-DNB in largely aqueous solutions yields alkenes 1-(fluoren-9-yl)-2-methylcyclopentene (4) and 1-(fluoren-9-yl)-5-methylcyclopentene (5)

Full text of DOI:10.1039/a901296f

View Article Online / Journal Homepage / Table of Contents for this issue
Mechanisms of solvolytic elimination reactions of tertiary
substrates: stereospecific 1,2-elimination reactions
Qingshui Meng and Alf Thibblin*
Institute of Chemistry, University of Uppsala, PO Box 531, S-751 21 Uppsala, Sweden
Received (in Cambridge) 16th February 1999, Accepted 11th May 1999
Solvolysis of (R,S)-1-chloro-1-(fluoren-9-yl)-2-methylcyclopentane (1-Cl) or the analogous 3,5-dinitrobenzoate ester
1-DNB in largely aqueous solutions yields alkenes 1-(fluoren-9-yl)-2-methylcyclopentene (4) and 1-(fluoren-9-yl)-
5-methylcyclopentene (5) as the main products. The chloride 1-Cl gives a product ratio 4:5 of 35:65 in 25 vol%
acetonitrile in water at 25 ЊC. The former product is formed by an anti elimination route, which shows that the
elimination reaction does not have a concerted unimolecular mechanism. The reaction to give 4 is suggested to occur
via a carbocation ion pair in which the leaving chloride ion abstracts the β-hydron. Alternatively, the reaction may
have an enforced uncoupled concerted mechanism in which water acts as the hydron-abstracting base. Also, solvolysis
of 2-methyl-1-phenylcyclopentyl p-nitrobenzoate (8-PNB) yields the more stable alkene 10 by anti stereochemistry.
We report here a study of the kinetics, product distributions,
and stereochemistry of solvolytic elimination reactions of the
chloride 1-Cl and the corresponding 3,5-dinitrobenzoate 1-
DNB (Scheme 1). The putative ion-pair intermediates of these
Introduction
It was recently pointed out that there is no consensus about the
dynamics of solvolytic substitution and elimination reactions at
tertiary carbon.1 For example, it is not clear what the lifetime of
a simple tertiary carbocation is in aqueous solution. The rate
H
OH
Fl
Fl
constant for reaction of the tert-butyl carbocation with water
has been estimated as kw ~ 4 × 109 sϪ1 2–4 and kw ~ 1 × 1010 sϪ1 5–7
which suggests that the ion pair undergoes diffusional separ-
ation (kϪd ~ 2 × 1010 Ϫ1) before reaction with nucleophiles. A
recent estimate is much larger, ~4 × 1012 Ϫ1.1 This rate constant
OH
,
Fl
+
+
H
H
Me
Me
OH
s
Me
kS
s
1-OH
2-OH
Fl
3-OH
refers to reaction of the tert-butyl carbocation with water in
which water has already undergone rotation into a reactive con-
formation. This very high reactivity of the tert-butyl carbo-
cation suggests that the carbocation does not form a liberated
intermediate in aqueous solution because the reaction of the
ion pair with the solvent is much faster than its separation to
free ions. The high reactivity of the carbocation has been used
as one of several arguments for a concerted unimolecular
mechanism for the competing elimination reaction.1,8
Fl
X
Fl
kE
+
Fl
+
H
6
4
5
Me
k
B[base]
1-X
X = Cl, DNB, OH
The proposed concerted unimolecular reaction mechanism is
controversial because the generally accepted solvolytic elimin-
ation reaction mechanism for tertiary substrates is the ion-pair
mechanism in which the leaving group acts as a base to abstract
the hydron within the ion pair.9,10 Experimental support for
such a mechanism was recently reported for the solvolytic
elimination of cumyl substrates having neutral leaving groups
of different basicity; a Brønsted parameter of β = 0.13 was
measured.11 However, as has also been pointed out by Toteva
and Richard,1 it is not clear why chloride anion, which is much
less basic than water, is such an efficient base in abstracting a
hydron within the ion pair.
Fl =
H
7
Scheme 1
reactions are concluded to have similar or shorter lifetimes
than that formed from tert-butyl chloride. The closely related
chloride A-Cl has been concluded to undergo solvolysis,
yielding substitution and elimination via a common ion-pair
intermediate as shown in Scheme 2.12 This mechanistic assign-
ment was based upon the measured kinetic deuterium isotope
effects using the hexadeuterated analog of A-Cl, which indi-
cated a branched mechanism via a common intermediate. The
anti stereochemistry now reported for the elimination reaction
to give alkene 4 from 1-Cl and 1-DNB supports this mechanistic
assignment and the alternative concerted unimolecular mech-
anism can be ruled out. However, the estimated lifetime of the
putative intermediate 1ϩXϪ is very short and an elimination
reaction mechanism of uncoupled concerted type is also dis-
cussed. The paper also addresses the question of why chloride
The concerted unimolecular elimination mechanism requires
a cyclic transition state such as 12 and 13 providing syn stereo-
chemistry. An investigation of the stereochemistry of the elim-
ination is therefore essential for the mechanistic assignment.
H
O
H
O
Cl
R
12
13
J. Chem. Soc., Perkin Trans. 2, 1999, 1397–1404
1397

View Article Online
25 vol% MeCN-H2O
25 °C
R+Cl–
+
H
H
H
OH
CH3
NHCOMe
CH3
Cl
H3C
H3C
H3C
CH3
4 parts
trace
A-Cl
H
CH2
H3C
7 parts
Scheme 2
ion shows such a high reactivity as a Brønsted base in many
solvolytic elimination reactions.
PNB
Ph
H
Me
8-PNB
Results
Synthesis and structure assignment
(R,S)-1-(Fluoren-9-yl)-2-methylcyclopentanol (1-OH) was syn-
thesized by treating fluorene with butyllithium in dry diethyl
ether, followed by the addition of 2-methylcyclopentanone.
The addition of fluorenyl anion to the carbonyl group of
2-methylcyclopentanone should occur by attacking the face
opposite to the methyl group due to steric hindrance providing
1-OH (Scheme 1) as the main diastereomer. This structure
assignment was confirmed by 1H NMR studies; NOESY spec-
tra in DMSO-d6 showed NOEs (5–10%) between the methyl
and the hydroxy protons. (R,S)-1-Chloro-1-(fluoren-9-yl)-2-
methylcyclopentane (1-Cl) was prepared from 1-OH and dry
hydrogen chloride gas in dichloromethane at 0 ЊC.
OH
Ph
H
Me
Ph
Ph
8-OH
+
Ph
11
OH
10
H
Me
9-OH
A mixture of the two diastereomers of 2-methyl-1-phenyl-
cyclopentanol was prepared as above. The configurations of the
1
two diastereomers were determined by H NMR NOE differ-
ence spectra. One of the isomers showed NOEs between the
methyl protons and the hydroxy proton, and between the ortho-
protons of the phenyl group and the methine proton. The
(R,S)-structure was assigned to this isomer (8-OH, Scheme 3).
(R,S)-1-(Fluoren-9-yl)-2-methylcyclopentyl 3,4-dinitrobenz-
oate (1-DNB) was synthesized from 1-OH, butyllithium, and
3,5-dinitrobenzoyl chloride. This reaction is expected to pro-
ceed with complete retention of configuration; only one
diastereomer was obtained. The same procedure was used for
preparation of 8-PNB from 8-OH, and 9-PNB from 9-OH.
Ph
PNB
H
Me
9-PNB
Scheme 3
effects (Table 1) on both kS and kE (Scheme 1). Fig. 1 shows the
effect of azide anion. It is not clear why the fraction of alkene 4
increases with increasing azide ion concentration in 40 vol%
acetonitrile in water. However, there is no such effect in 25 vol%
acetonitrile (Table 2) or with acetate ion, which is of similar
basicity as azide ion.
Kinetics and product studies
The solvolysis of 1-Cl or 1-DNB in mixtures of water with
acetonitrile, methanol, or trifluoroethanol at 25 ЊC provides
mainly elimination products (Scheme 1). The predominant
alkenes are those formed by 1,2-elimination: 1-(fluoren-9-yl)-2-
methylcyclopentene (4) and 1-(fluoren-9-yl)-5-methylcyclo-
pentene (5), the former by an anti elimination path. A minor
amount of the 1,3-elimination product 5-(fluoren-9-yl)-1-
methylcyclopentene (6) is also formed. The substitution prod-
ucts are (R,S)-1-(fluoren-9-yl)-2-methylcyclopentanol (1-OH),
(R,R)-1-(fluoren-9-yl)-2-methylcyclopentanol (2-OH), and the
rearranged alcohol 1-(fluoren-9-yl)-2-methylcyclopentanol
(3-OH), along with the corresponding ethers. The product
compositions and the kinetics of the reactions were studied by a
sampling high-performance liquid chromatography procedure
combined with GC analysis of the alkene compositions. The
measured rate constants are recorded in Table 1 and the results
of the product studies are shown in Table 2. Sodium perchlorate
has a small rate-enhancing effect on the solvolysis of 1-Cl. In
contrast, chloride, bromide, and azide anions show negative salt
The substitution with solvent water predominantly occurs
with retention of configuration, e.g. the amount of retention
was measured with 1-Cl as 81% in the water–acetonitrile
mixtures and as ca. 92% in 50 vol% trifluoroethanol in water
(Table 2). The bimolecular methoxide-promoted E2 reaction
providing 1-fluoren-9-ylidene-2-methylcyclopentane (7) is rela-
tively slow. The data in Table 1 yield an approximate second-
order rate constant of (34.7 Ϫ 8.2) × 10Ϫ6/0.10 = 265 × 10Ϫ6
M
Ϫ1 sϪ1, which is about 6 times smaller than the corresponding
rate constant for the closely related A-Cl.12,13 It is difficult to
study the hydroxide-promoted E2 reaction of 1-Cl since even
with a large concentration of base this reaction is relatively slow
compared with the solvolysis. Moreover, the alkenes produced
from the latter reaction are isomerized by the hydroxide ion to
give the thermodynamically more stable alkene 7. It is easier to
study the methoxide-promoted elimination owing to the slower
competing solvolysis in methanol (Tables 1 and 2).
1398
J. Chem. Soc., Perkin Trans. 2, 1999, 1397–1404

View Article Online
Table 1 Rate constants for the reactions of 1-Cla at 25 ЊC
Salt
Solvent (vol%)b
kobs/10Ϫ6 sϪ1 c,d
kS/10Ϫ6 sϪ1 d
kE/10Ϫ6 sϪ1 d
None
MeCN (40)
MeCN (40)
MeCN (40)
MeCN (40)
MeCN (40)
MeCN (25)
MeOH
519
743
312
450
22
33
11
17
497
710
301
433
0.50 M NaClO4
0.50 M NaCl
0.50 M NaBr
0.50 M NaN3
None
519
16
504
3.9 × 103
8.2
0.2 × 103
0.5
3.7 × 103
7.7
None
0.10 M NaOMe
MeOH
34.7
a Substrate concentration 0.01–0.1 mM. b By volume in water. c kobs = kS ϩ kE ϩ kB[base] (Scheme 1). d Estimated maximum error: 10%.
Table 2 Product compositions for the solvolysis reactions of 1-Cl, 1-DNB, and the acid-catalyzed solvolysis of 1-OH at 25 ЊC
Product/mol%c
Substratea
Solvent (vol%)b
Salt
1-OH
2-OH
3-OH
4
5
6
1-Cl
MeCN (25)
MeCN (25)
MeCN (25)
MeCN (40)
MeCN (40)
MeCN (40)
MeCN (40)
MeCN (40)
MeCN (60)
MeCN (75)
MeOH (50)d
CF3CH2OH (50)d
MeOH (100)d
MeCN (40)
MeCN (25)
MeCN (25)
none
2.8
2.4
2.9
2.1
2.4
1.5
1.8
1.7
1.6
1.3
1.5
2.5
0.6
0.6
0.5
0.5
0.5
0.2
0.4
0.5
0.3
0.2
0.3
0.2
2.5
2.5
2.8
1.6
1.5
1.3
1.4
1.3
1.0
0.6
1.4
0.9
30
25
31
24
22
38
33
28
18
13
39
15
34
8
55
58
56
61
61
50
53
58
70
76
50
72
58
81
15
7
9
11
7
11
13
9
11
11
9
8
7
10
8
8
1-Cl
0.75 M NaOAc
0.75 M NaN3
none
0.50 M NaClO4
0.50 M NaN3
0.50 M NaBr
0.50 M NaCl
none
none
none
none
none
1-Cl
1-Cl
1-Cl
1-Cl
1-Cl
1-Cl
1-Cl
1-Cl
1-Cl
1-Cl
1-Cl
1-DNBe
1-OH
1-OHg
none
0.75 M HClO4
0.75 M HClO4
1.8
0.4
~9f
<1
1.4
—
<1
60
84
16
7
a Substrate concentration 0.01–0.1 mM. b By volume in water. c Estimated maximum error is 10%. d Ethers are formed in small amounts, but they are
not included in the calculation of the product compositions. e T = 70 ЊC. f The yield of 2-OH was determined by extrapolation to zero time of a plot
of product yield ratio [2-OH]/([2-OH] ϩ [4] ϩ [5] ϩ [6]) against time for 2–15% reaction during which the ratio of the alkenes does not change within
experimental error. No 3-OH was detected by HPLC analysis. g Product yields after 3 months, only 1.3 mol% of 1-OH left.
under the neutral conditions used for the solvolysis of 1-Cl and
1-DNB.
The nucleophilic discrimination between azide ion and water
is very small and only a very small amount of azide product is
observed in the solvolysis of 1-Cl in 40 vol% MeCN in water
containing 0.50 M NaN3 (kN /kHOH = 3). Thiocyanate also gives
only one product; a discrimi3nation value of kSCN/kHOH = 3 was
measured with 0.75 M NaSCN in 25 vol% acetonitrile in water.
These are dimensionless ratios of second-order rate constants
calculated by means of eqn. (1). No traces of other isomeric
kNu/kHOH = (Σ[RNu]/Σ[ROH])([H2O]/[NuϪ])
(1)
azide or thiocyanate products are detected by HPLC, 1H NMR,
and GC analyses.
The solvolysis of 8-PNB and of the more reactive isomer
9-PNB in 25 vol% acetonitrile in water yields the elimination
products 2-methyl-1-phenylcyclopentene (10), 5-methyl-1-
phenylcyclopentene (11) and the substitution products (R,S)-2-
methyl-1-phenylcyclopentan-1-ol (8-OH) and the correspond-
ing (R,R)-diastereomer 9-OH (Scheme 3). The product
compositions for the solvolysis reactions of the esters and the
acid-catalyzed reactions of the alcohols, determined by HPLC
analysis, are shown in Table 3. Alkene 10 is the thermo-
dynamically more stable product, and after long reaction time
in perchloric acid solution the alcohols have been converted
to 94% of 10 and 6% of 11 (Table 3). No interconversion of
the products occurs under the neutral conditions used for the
solvolysis of 8-PNB and 9-PNB. The nucleophilic discrimin-
ation between thiocyanate ion and water is small: a discrimin-
ation ratio of kSCN/kHOH = 63 was measured with 0.75 M
NaSCN in 25 vol% acetonitrile in water.
Fig. 1 The effect of added sodium azide on the reaction rates of 1-Cl
in 40 vol% acetonitrile in water at 25 ЊC. Ionic strength was kept con-
stant (0.50 M) with sodium perchlorate. The rate constant k14 refers to
formation of alkene 4 from 1-Cl.
The acid-catalyzed solvolysis of 1-OH in 25 vol% aceto-
nitrile in water containing 0.75 M HClO4 gives the alkenes 4, 5,
and 6 as shown in Table 2. The product composition is time-
dependent, showing that the initially formed products undergo
slow isomerization, a long reaction time favouring the more
stable alkene 4. No interconversion of the products occurs
J. Chem. Soc., Perkin Trans. 2, 1999, 1397–1404
1399

View Article Online
Table 3 Product compositions for the solvolysis reactions of 8-PNB,
9-PNB and the acid-catalyzed solvolysis of 8-OH and 9-OH in 25 vol%
acetonitrile in water at 25 ЊC
to significant reaction of azide ion through a preassociation
mechanism.†
Despite the uncertainties discussed above, let us use the
results with azide ion and thiocyanate ion, kN /kHOH = kSCN
/
3
Product/mol%b
kHOH = 3, as a clock to estimate the lifetime of the carbocation.
Thus, the carbocation 1ϩ reacts to give products through the ion
Substratea
[HClO4]/M
8-OH
9-OH
10
11
pair with a rate constant of ≥1.2 × 1012 Ϫ1, which is the sum of
s
the rate constants for formation of alcohol (kw ≥ 7 × 1010 sϪ1
)
8-PNB
9-PNB
8-OH
8-OH
9-OH
9-OH
—
—
0.10
0.75
0.10
0.75
9
9
66
6
9
20
94e
26
94e
19
22
17
and alkenes (ke ≥ 11 × 1011 Ϫ1). A considerably longer lifetime
s
60
63c,d
of the intermediate is obtained if the main routes to the prod-
ucts 5 and 6 do not proceed through the same intermediate as
the reactions giving 1-OH and 4, i.e. there are parallel reactions
to these products. Thus, if the main reaction routes providing 5
and 6 do not proceed through the carbocation intermediate as
the reactions which give 4 and 1-OH, the lifetime of the puta-
6e
51d,f
23
6e
a Substrate concentration 0.01–0.1 mM. b Estimated maximum error is
10%. c The yield of 9-OH was determined by extrapolation to zero
time of a plot of product yield ratio [9-OH]/([10] ϩ [11]) against time
for 6–21% reaction, during which the ratio of the two alkene stereo-
isomers does not change within experimental error. d Estimated maxi-
mum error is 20%. e Product yields after 3 months, no alcohol left.
f The yield of 8-OH was determined by extrapolation to zero time of a
plot of product yield ratio [8-OH]/([10] ϩ [11]) against time for 36–82%
reaction, during which the ratio of the two alkene stereoisomers does
not change within experimental error.
tive intermediate may be much longer (kw ~ 4 × 1010 Ϫ1 and ke ~
s
3 × 1011 Ϫ1) which is far below the borderline (~1013 Ϫ1) where
s
s
the intermediate does not exist as a discrete intermediate owing
to the disappearance of the barrier for departure of the leaving
group.
A roughly similar reactivity was estimated for the substitu-
tion reaction of the ion pair AϩClϪ, kw = 4 × 1010 Ϫ1 in 25 vol%
s
acetonitrile in water; the rate constant for the elimination was
measured as ke = 7 × 1010 sϪ1 12
.
For comparison, the nucleo-
Discussion
philic selectivity for PhCH2C(Me)2Cl under the same reaction
The solvolysis of 1-Cl in 25 vol% acetonitrile in water (Scheme
1) is about 27 times faster than that of the closely related sub-
strate A-Cl (Scheme 2), and the elimination-to-substitution
ratio is larger, 94:6 compared with 7:4.12 Thus, the solvolytic
substitution of 1-Cl is about 4 times faster than that of A-Cl
and the elimination reactions of 1-Cl are 40 times faster than
that of A-Cl. The reason is presumably steric; there is more
crowding in 1-Cl than in A-Cl, which is decreased in the elimin-
ation products. Consistently, alkenes with only traces of alco-
hols are the final products in acidic solution (vide infra).
Accordingly, there is a large driving force favouring reaction
of the carbocation to give the alkenes. This is in contrast to
the partitioning of
thermodynamics generally favours formation of the alcohol
product.
Addition of azide anion does not increase the rate of dis-
appearance of the substrate (kobs = kS ϩ kE, Fig. 1), and the
conditions was determined as kN /kHOH = 19,19 and the reaction
3
p-MeOC6H4CH2C(Me)2Cl in 50% trifluoroethanol or methanol
in water shows selectivities of kN /kHOH = 11–15.1 Nucleophilic
selectivity values are expected to 3be somewhat smaller in these
solvents because the viscosity of the alcohol–water mixtures are
higher than those of acetonitrile–water.‡
Indirect evidence for stepwise carbocation reactions is also
provided by the formation of significant amounts of 3-OH and
6. Formation of these products by concerted reactions would
require a large number of changes in bonding occurring at
single reaction steps. However, it is plausible that intra-
molecular transfer of a hydride is concerted with cleavage of
the bond to the leaving group. This produces a somewhat more
stable tertiary carbocation having a methyl instead of a fluor-
enyl substituent connected to the carbocation center. Also the
predominant retention of substitution with solvent water
supports the conclusion of stepwise reactions.
a simple acyclic carbocation where
amount of azide substitution product is very small, kN
/
The fluorenyl hydron of 1-Cl, despite its high acidity (pKa of
fluorene is ~22.5),20 is abstracted much more slowly than the
other β-hydrons. The reason for this behaviour, which is in
sharp contrast to the behaviour with added strong bases but
is similar to that of A-Cl and A-Br,12 is probably that the
β-hydrons of the cyclopentane ring provide stabilization of the
ionization transition state as well as of the carbocation inter-
mediate by hyperconjugation. Hyperconjugative stabilization
from the fluorene 9-hydrogen is sterically unfavourable in this
system. This small partial bond-breaking of the carbon-
hydrogen bonds increases the acidity of the hydrogens in the
carbocation ion pair. Therefore, it should be relatively easy for
3
kHOH = 3 [see Results, dimensionless ratio of second-order rate
constants in 25 vol% acetonitrile in water, eqn. (1)]. The
amount of substitution product with thiocyanate ion, which is
also expected to give substitution products with a diffusion-
limited rate, is also very small (kSCN/kHOH = 3). This is con-
sistent with the results obtained for A-Cl, which showed an
extremely low nucleophilic discrimination of kN /kHOH = 5 in 25
vol% acetonitrile in water.12 The nucleophilic3 discrimination
between methanol and water for reaction with the carbocation
was also found to be unusually small (kMeOH/kHOH ~ 0.6).12,14 The
corresponding bromide A-Br showed a nucleophilic selectivity
in 70 vol% methanol of kSCN/kHOH = 3 and kN /kHOH ~ 4.14
These results suggest that the carbocation3 1ϩ is very short-
lived. The lifetime should be close to that of Aϩ, which reacts
with solvent water as a nucleophile with a rate constant
† The ion pair does not necessarily show the same partitioning as the
free carbocation. A shielding effect that decreases the rate constant for
attack of negatively charged nucleophiles, even the potent azide ion,
more than attack of a neutral nucleophile is likely. Experimental sup-
port for such a shielding has been discussed previously; it results in an
overestimation of the rate constant for the water reaction.18 The shield-
ing effect should be more important in reactions showing predominant
retention of configuration like that of 1-Cl (Table 1) than in reactions
which mostly give inversion.
‡ The viscosity of 50 vol% methanol in water is about 1.8 times higher
than that of water. This means that the rate of diffusion is about 1.8
times slower in the methanol–water mixture than in the mixtures with
acetonitrile. Ethanol–water mixtures show even higher viscosity.
Accordingly, the value of kd = 5 × 109 MϪ1 sϪ1 for diffusion in water–
alcohol solvent mixtures is probably too high. A value of 7 × 109 MϪ1
sϪ1 has been estimated for the diarylmethyl carbocation in 0–50 vol%
acetonitrile water.16
estimated at kw ~ 4 × 1010 sϪ1 12
This rate constant was deter-
.
mined by the “azide-clock” method,15 i.e. it is based upon the
measured product ratio [A-N3]/[A-OH] and an assumed rate
constant for reaction of Aϩ with azide ion of 5 × 109 MϪ1 sϪ1
.
Direct measurement of second-order rate constants for
diffusion-limited reaction of azide ion with more stable carbo-
cations in aqueous acetonitrile confirms that this is a good
estimate.16 However, since the diffusional separation of an ion
pair in water has a rate constant of about 2 × 1010 sϪ1 17
the
,
ion pair reacts with nucleophiles to some extent before separ-
ation. Therefore, the azide-clock method for estimation of kw
for the carbocation probably results in a lower limit owing
1400
J. Chem. Soc., Perkin Trans. 2, 1999, 1397–1404

View Article Online
the chloride leaving group to abstract one of these hydrons to
occur from the rearranged carbocation intermediate which is
give the elimination products 4 and 5. The large secondary
formed either by hydride transfer from the initially formed
carbocation, or is formed by a direct ionization involving
concerted hydride transfer.
D6
kinetic deuterium isotope effect of kobsH/kobs = 2.2 measured
for the hexadeuterated analog of A-X (Scheme 2) is consistent
with a large amount of hyperconjugation in the rate-limiting
ionization transition state.12 Furthermore, the isotope effects on
the solvolytic elimination and substitution reactions vary with
solvent composition in a way which suggests coupled reactions
via a common intermediate.12
The esters 8-PNB and 9-PNB yield about 25% and 30%,
respectively, of alkenes 10 and 11. The fraction of the former,
which is formed by an anti elimination route from 8-PNB, is
10/(10 ϩ 11) = 24%. This is not much lower than the fraction
of 10 (28%) formed from 9-PNB by a syn elimination route.
This shows that the elimination reaction to give 10 does not
have a concerted unimolecular mechanism with a pericyclic
transition state as 13, but indicates a carbocation mechanism.
Anti elimination has also been observed in the solvolysis
of 2-aryl-endo-2-norbornyl trifluoroacetates in trifluoro-
ethanol.24
The substitution reactions of the two diastereomeric esters
with solvent water yield predominantly the alcohol 9-OH
(Table 3). The amount of retention for the reaction of 8-PNB
is 12% which might suggest a carbocation reaction with some
leaving-group shielding. However, the two diastereomers give
similar product compositions, indicating a common carbo-
cation intermediate. Attack of the nucleophile from the less
sterically hindered side gives rise to alcohol 9-OH as the main
product.
The partial breaking of the bonds to the β-hydrogens is also
manifested by a 1,2-hydride shift which gives rise to the
rearranged alcohol 3-OH and the alkene 6. The π-orbitals of
the fluorene moiety have been concluded to assist the ionization
to the primary carbocation which is formed in the solvolysis of
9-(4Ј-bromophenylsulfonyloxymethyl)fluorene.21 Stabilization
of this type is expected to be small in the tertiary carbocations
1ϩ and Aϩ for steric reasons and there is also less need for such
participation. Therefore, the large amount of retention is most
reasonably attributed to the larger stability of the 1-OH dia-
stereomer, and not to significant participation. Attack of the
nucleophile from the less sterically hindered side of the carbo-
cation gives rise to the more stable 1-OH as the main substi-
tution product. However, this side of the ion pair holds the
negatively charged leaving group that may slow down the attack
of the negatively charged nucleophile more than the reaction
with the uncharged water molecule (which may involve the
solvent-separated ion pair). The result is an overestimation of
kw for short-lived carbocations.18
The lifetime of the carbocation has been determined by
trapping experiments with 9-PNB in 25 vol% acetonitrile in
water. The measured nucleophilic discrimination ratio of kSCN
/
kHOH = 63 corresponds to a reactivity of kw = 3 × 109 sϪ1. It is,
ϩ
,
The stereochemistry of the elimination reactions of 1-Cl and
1-DNB to give alkene 4 is anti, which shows that these reactions
are not of the concerted unimolecular type, as suggested by
Richard and co-workers for the elimination of HCl from tert-
butyl chloride and related derivatives.1,22
as expected, close to that of the cumyl carbocation, PhC(Me)2
which has a reactivity of kw = 5 × 109 sϪ1 based upon a meas-
ured nucleophilic discrimination ratio of kN /kHOH = 42 in the
3
same aqueous solvent.25
The product data (Table 3) show that the carbocation is not
a completely free, solvent-equilibrated carbocation since 9-OH
gives a quite different product spectra than the two esters
8-PNB and 9-PNB. Also solvolysis of cumyl compounds
PhC(Me)2X gives an elimination fraction which varies with
leaving group X.11,25 This clearly shows that the leaving group
is involved in the elimination process. The thermodynamic
stabilities of the products are, however, quite different: the
cyclic alkenes are thermodynamically much more stable relative
to the alcohols than the corresponding acylic alkenes. This is
reflected in the product compositions.
The major elimination product from the esters 8-PNB and
9-PNB is 11 owing to kinetic control; important factors here are
the number of available β-hydrons and the extent of steric
hindrance. In contrast, under thermodynamic control alkene 10
is predominant. The same behaviour is observed with 1-Cl; the
alkene product ratio 5 to 4 is much larger under kinetic control
than under thermodynamic control (Table 2).
Is a concerted anti elimination reaction a possible explan-
ation for the formation of alkene 4? An E2 reaction having a
central transition state with a solvent water molecule acting as
the hydron-abstracting base should involve loss of the more
acidic of the β-hydrons, which is the 9-hydron of the fluorene
moiety of the substrate (pKa ~ 22.5). Moreover, strong base,
e.g. hydroxide anion, abstracts the latter hydron giving rise
to alkene 7 by a base-promoted E2 mechanism and there is no
indication of a bimolecular reaction to give 4. We have previ-
ously concluded that concerted solvent-promoted elimination
reactions in aqueous solvents are only significant for a substrate
having an acidic β-hydrogen and which relatively slowly ionizes
to a carbocation intermediate.21 However, in principal, an E2
reaction having a very carbocation-like transition state may
show similar characteristics to an elimination reaction through
a very unstable ion-pair intermediate. Accordingly, a concerted
elimination reaction may be enforced by the non-existence of a
barrier for departure of the leaving group. This corresponds to
a reaction across a flat potential maximum with no energy well
for the intermediate. Such a reaction may account for the
relatively large amount of alkene 5 that is formed from 1-Cl
under kinetic control, compared to the more stable alkene 4
(Table 2). Alternatively, 5 may be formed by a concerted uni-
molecular mechanism.22
Why is chloride ion such an efficient base in solvolytic elimination
reactions?
An important driving force of the ionization is of course the
solvation of the chloride ion but it is not expected to be com-
plete at the contact ion-pair stage. Thus, the chloride anion in
the ion pair is not solvent-equilibrated and the pKa of HCl in
water may not reflect its kinetic basicity. A halide anion, which
is not fully solvated, should be a relatively strong base. For
example, it is well known that halide anions in aprotic solvents
are strong bases. The stabilization of the positive charge of the
carbocation partner by the negative leaving chloride ion of
the ion pair may decrease the basicity of the leaving group but
this effect is probably less important than the positive effect
originating from incomplete solvation. The low catalytic
activity of water as a hydron acceptor in E1 reactions and in
water-promoted E2 reactions has been observed previously;26
Brønsted plots generally show a negative deviation for water as
base of about one order of magnitude.
The elimination of HCl from the ion pair to give alkene 4 is
apparently very fast, about 3 × 1011 Ϫ1, which is in fair agree-
s
ment with the expected rate constant for such a process. The
rate constant for rotation of the leaving group into reaction
position should be in the order of 1011 Ϫ1 23 The estimated rate
s .
constant for formation of 5 from the putative intermediate is
larger, about 6 × 1011 Ϫ1. The studied system does not give any
s
information about the stereochemistry of this elimination, i.e.
which of the hydrons is abstracted. Therefore, it is not possible
to exclude that the major reaction path to give 5 does not
employ a carbocation intermediate but is a one-step reaction of
the concerted unimolecular type.22 The reaction to give alkene 6
is slower, about 1 × 1011 Ϫ1, but this hydron abstraction should
s
J. Chem. Soc., Perkin Trans. 2, 1999, 1397–1404
1401

View Article Online
It has been concluded that even substrates which solvolyze to
give relatively stable carbocations in mostly aqueous solution
produce alkene predominantly from the ion pair; the solvent-
equilibrated carbocation yields almost exclusively the alcohol
product.27 Moreover, Bunton and co-workers have found that,
even in highly aqueous solvents, added ClϪ promotes elimin-
ation from relatively stable carbocations.28
1H NMR (DMSO-d6) δ 7.20–7.90 (m, 8 H), 4.61 (s, 1 H), 4.10 (s,
1 H), 1.93 (m, 1 H), 1.56 (m, 1 H), 1.41 (m, 2 H), 1.04 (d, J 6.8,
3 H), 0.97 (m, 2 H), 0.79 (m, 1 H).
(R,S)-1-Chloro-1-(fluoren-9-yl)-2-methylcyclopentane (1-Cl)
A solution of 1-OH (1 g) in dichloromethane (50 ml) contain-
ing anhydrous calcium chloride and lithium chloride was
cooled to 0 ЊC. Dry hydrogen chloride was bubbled through the
solution for 5 h. After filtration, the solvent was removed.
Recrystallization several times from a mixture of ethanol and
Conclusions
Our results show that the solvolysis of 1-Cl does not produce
alkene 4 by the recently proposed concerted unimolecular
mechanism. The results are consistent with a mechanism in
which the leaving chloride ion abstracts the β-hydron within an
ion pair. The mechanistic conclusions for the solvolysis of the
esters 8-PNB and 9-PNB are similar.
1
pentane gave pure material: mp 79–81 ЊC; H NMR (CDCl3)
δ 7.26–8.22 (m, 8 H), 4.55 (s, 1 H), 2.43 (m, 1 H), 1.79 (m, 1 H),
1.68 (m, 2 H), 1.49 (m, 1 H), 1.41 (d, J 6.5, 3 H), 1.32 (m, 1 H),
1.09 (m, 1 H); 13C NMR (CDCl3) δ 145.48, 142.59, 142.09,
141.23, 127.93, 127.60, 127.19, 126.75, 126.73, 126.29, 119.81,
119.44, 89.42, 54.99, 43.29, 36.49, 31.33, 19.35, 14.93.
(R,S)-1-(Fluoren-9-yl)-2-methylcyclopentyl 3,5-dinitrobenzoate
(1-DNB)
Experimental
General procedures
A solution of 1-OH (0.34 g) in dry diethyl ether (20 ml) was
cooled to Ϫ20 ЊC, and butyllithium (0.8 ml of 1.6 M solution in
hexane) was added under nitrogen by means of a syringe. The
mixture was stirred at this temperature for 30 min. Then 3,5-
dinitrobenzoyl chloride (0.30 g), dissolved in dry diethyl ether
(10 ml), was added at Ϫ20 ЊC. The mixture was stirred for
another 1 h and was then poured into ice and water, followed by
extraction three times with diethyl ether. The combined diethyl
ether fractions were washed with saturated aqueous potassium
carbonate solution, followed by washing with water and brine,
and drying over sodium sulfate. After removal of solvent,
the crude product was purified by flash chromatography (silica
gel) with 5% ethyl acetate–pentane as eluent. Recrystallization
twice from CH2Cl2–pentane–ethanol gave pure material: mp
NMR spectra were recorded at 25 ЊC with a Varian Unity 400
1
spectrometer, for H at 400 MHz and for 13C at 100.6 MHz.
Chemical shifts are indirectly referenced to TMS via the solvent
signal (chloroform-d1 7.26 ppm and 77.0 ppm; DMSO-d6 2.49
and 39.5 ppm). J Values are given in Hz. The high-performance
liquid chromatography analyses were carried out with a
Hewlett-Packard 1090 liquid chromatograph equipped with a
diode-array detector on an Inertsil 5 ODS-2 (3 × 100 mm)
reversed-phase column. The mobile phase was a solution of
acetonitrile in water. The reactions were studied at constant
temperature in a HETO 01 PT 623 thermostat bath. The pH
was measured using a Radiometer PHM82 pH meter with an
Ingold micro glass electrode.
The GC analyses were carried out with a Varian 3400 capil-
lary gas chromatograph equipped with a flame ionization
detector. Nitrogen was used as carrier gas. The column was a
fused-silica capillary column (Rescom, SE54, 25 m, 250 µm).
The injection temperature was 250 ЊC, and the column temper-
ature was maintained constant at 200 ЊC.
1
149–151 ЊC; H NMR (CDCl3) δ 9.32 (t, J 2.2, 1 H), 9.27 (d,
J 2.2, 2 H), 7.12–7.82 (m, 8 H), 5.44 (s, 1 H), 3.06 (m, 1 H),
2.52 (m, 1 H), 1.71 (m, 1 H), 1.62 (m, 1 H), 1.35 (m, 3 H), 0.62
(d, J 6.5, 3 H).
1-(Fluoren-9-yl)-2-methylcyclopentene (4), 1-(fluoren-9-yl)-5-
methylcyclopentene (5) and 5-(fluoren-9-yl)-1-methylcyclo-
pentene (6)
Materials
Merck silica gel 60 (240–400 mesh) was used for flash chrom-
atography. Diethyl ether and tetrahydrofuran were distilled
under nitrogen from sodium and benzophenone. Methanol and
acetonitrile were of HPLC grade. All other chemicals were of
reagent grade and used without further purification. The
aqueous solutions of sodium azide, sodium chloride, sodium
bromide, and sodium perchlorate were adjusted to pH ~6.5
with 1 M aqueous perchloric acid before use.
These were prepared by treatment of the alcohol 1-OH with
ZnCl2–HCl in chloroform. After the reaction mixture had been
stirred for 1 h at room temperature, it was extracted with
pentane, followed by washing with water, brine, and drying over
sodium sulfate. After removal of solvent, the residue was separ-
ated by flash chromatography (silica gel) with 1% ethyl acetate–
pentane as eluent. The early fractions contained a mixture of
1
4 and 5 (4:1); 4: H NMR (CDCl3) δ 7.20–7.90 (m, 8 H), 4.96
(s, 1 H), 2.45 (m, 2 H), 2.05 (m, 3 H), 1.67 (m, 4 H); 5: 1H NMR
(CDCl3) δ 7.20–7.90 (m, 8 H), 5.76 (m, 1 H), 4.80 (s, 1 H), 1.32–
2.70 (m, 5 H), 0.51 (d, J 7.0, 3 H).
(R,S)-1-(Fluoren-9-yl)-2-methylcyclopentanol (1-OH)
A solution of butyllithium (12.5 ml of a 1.6 M solution in
hexane, diluted with 30 ml of dry diethyl ether) was added to
fluorene (3.3 g), dissolved in dry diethyl ether (60 ml), at room
temperature under nitrogen. After addition, the reaction
mixture was refluxed for 1 h. Then it was cooled to Ϫ40 ЊC, and
a solution of 2-methylcyclopentanone (2.4 g, dissolved in 10 ml
of dry diethyl ether) was added. The reaction mixture was
stirred at this temperature for 20 min, and was then poured into
a mixture of ice and 2 M hydrochloric acid. The mixture was
extracted three times with diethyl ether. The combined diethyl
ether fractions were washed with water to neutrality, followed
by washing with brine and drying over sodium sulfate. After
removal of solvent, the crude product was purified by flash
chromatography (silica gel) with 10–15% ethyl acetate–pentane
as eluent. Recrystallization twice from CH2Cl2–pentane gave
The later fractions contained only 6: 1H NMR (CDCl3)
δ 7.20–7.80 (m, 8 H), 5.54 (m, 1 H), 4.16 (d, J 3.5, 1 H), 3.50 (m,
1 H), 2.04 (m, 3 H), 1.99 (m, 1 H), 1.72 (m, 1 H), 1.53 (m, 1 H),
0.67 (m, 1 H); 13C NMR (CDCl3) δ 147.33, 144.68, 142.09,
141.41, 140.67, 127.79, 126.90, 126.74, 126.49, 125.37, 123.93,
119.57, 119.46, 51.81, 47.98, 31.31, 24.20, 15.38.
(R,S)- and (R,R)-1-Phenyl-2-methylcyclopentanol (8-OH) and
(9-OH)
A solution of butyllithium in hexane (1.6 M, 8 ml) was added to
a solution of bromobenzene (2 g) in dry diethyl ether (20 ml) at
0 ЊC under nitrogen. After addition, the reaction mixture was
stirred for 30 min at 0 ЊC, followed by addition of a solution of
2-methylcyclopentanone (1.3 g, dissolved in 10 ml of THF).
The reaction mixture was then allowed to reach room temper-
ature. After 2 h, the reaction solution was poured into a mixture
of ice and 2 M hydrochloric acid, and extracted with diethyl
1
pure material: mp 102–103 ЊC; H NMR (CDCl3) δ 7.20–7.90
(m, 8 H), 4.17 (s, 1 H), 2.16 (m, 1 H), 1.73 (m, 1 H), 1.71 (s, 1 H),
1.52 (m, 2 H), 1.25 (m, 1 H), 1.13 (d, J 6.9, 3 H), 1.08 (m, 2 H);
1402
J. Chem. Soc., Perkin Trans. 2, 1999, 1397–1404

View Article Online
ether three times. The combined diethyl ether fractions were
washed with water to neutrality, followed by washing with brine
and drying over sodium sulfate. After removal of solvent, the
crude product was purified by flash chromatography (silica gel)
with 10–15% ethyl acetate–pentane as eluent. The early
was diluted with water to a final concentration of 25 vol%
acetonitrile in water. After two days at 25 ЊC, GC analysis
showed no change in the area ratio of the two alkenes 4 and 5.
The product compositions for the acid-catalyzed reaction of
1-OH, 8-OH and 9-OH were determined according to the
following method: At appropriate times, samples were trans-
ferred to a 2 ml HPLC flask and quenched with a mixture
of acetonitrile and aqueous sodium acetate (1 M), followed by
HPLC analyses giving the mol% of each component, except 4
and 5 which eluted as one peak. Pentane was added to a part of
the quenched reaction solutions. After vigorous shaking, the
organic phase was transferred to a pear-shaped flask, and solv-
ent was evaporated with a stream of nitrogen until ~10 µl were
left. Aliquots of this solution were injected directly onto the GC
column. The ratio of the two alkene stereoisomers 4 and 5 was
1
fractions contained only 8-OH: H NMR (CDCl3) δ 7.21–7.51
(m, 5 H), 2.23 (m, 1 H), 2.11 (m, 1 H), 1.97 (m, 3 H), 1.80 (m,
1 H), 1.66 (m, 1 H), 1.60 (s, 1 H), 0.86 (d, J 6.8, 3 H).
The later fractions contained only 9-OH: 1H NMR (CDCl3)
δ 7.20–7.55 (m, 5 H), 2.42 (m, 1 H), 2.22 (m, 1 H), 1.91 (m, 3 H),
1.60 (s, 1 H), 1.41 (m, 1 H), 0.87 (m, 1 H), 0.56 (d, J 7.1, 3 H).
(R,S)-1-Phenyl-2-methylcyclopentyl p-nitrobenzoate (8-PNB)
This was synthesized from 8-OH and p-nitrobenzoyl chloride
by the same method as described above for 1-DNB: mp 130–
1
determined from the peak areas. Comparison with H NMR
1
133 ЊC; H NMR (CDCl3) δ 8.20–8.34 (m, 4 H), 7.21–7.36 (m,
analysis showed that the relative response factors were the same
within experimental error assuming that the NMR integrals
exactly correspond to the expected number of protons. Control
experiments showed that the alcohol 1-OH and the alkenes 4, 5,
6 are stable under GC analyzing conditions, but 1-Cl and
1-DNB decompose to the alkenes. The fractions of 4 and 5 from
the solvolysis of 1-Cl and 1-DNB were determined after at least
five half-lives.
5 H), 2.73 (m, 2 H), 2.06 (m, 2 H), 1.86 (m, 2 H), 1.68 (m, 1 H),
1.15 (d, J 6.4, 3 H); 13C NMR (CDCl3) δ 163.19, 150.45, 141.99,
136.83, 130.54, 128.26, 126.91, 124.35, 123.58, 93.62, 48.80,
36.75, 32.35, 22.06, 12.89.
(R,R)-1-Phenyl-2-methylcyclopentyl p-nitrobenzoate (9-PNB)
This was synthesized from 9-OH as described above: mp
109–111 ЊC; 1H NMR (CDCl3) δ 8.14–8.30 (m, 4 H), 7.23–7.37
(m, 5 H), 2.69 (m, 3 H), 2.16 (m, 1 H), 1.91 (m, 2 H), 1.47 (m,
1 H), 0.65 (d, J 7.1, 3 H).
Determination of relative HPLC response factors
The relative response factors of 1-OH and alkene 6, 6-OH and
the mixture of alkene 4 and 5 (the relative response factors of 4
and 5 were assumed to be same), 8-OH and 9-OH, 8-OH and
alkene 10, 8-OH and 8-PNB, 8-OH and 9-PNB were deter-
mined by a combination of NMR analysis and HPLC analysis.
A mixture of an appropriate amount of the two materials was
dissolved in chloroform-d1 and analysed by 1H NMR. The peak
areas of the well-resolved signals, usually methyl signals, were
integrated and used to calculate the relative ratio of the two
components, assuming that the NMR integrals exactly corre-
spond to the expected number of protons. A few microlitres of
this solution were transferred to a 2 ml HPLC flask. The solvent
was evaporated under a stream of nitrogen. The residue was
dissolved in acetonitrile and analyzed by HPLC.
1-Phenyl-2-methylcyclopentene (10)
This was prepared by stirring a mixture of 8-OH with ZnCl2–
HCl in chloroform for 1 h at room temperature. The mixture
was extracted with pentane, followed by washing with water,
brine, and drying over sodium sulfate. Removal of solvent gave
an oil. Analysis by NMR showed about 5% of 1-phenyl-5-
methylcyclopentene (11) but no other impurity: 1H NMR
(CDCl3) δ 7.15–7.40 (m, 5 H), 2.77 (m, 2 H), 2.53 (m, 2 H), 1.93
(m, 2 H), 1.88 (m, 3 H); 13C NMR (CDCl3) δ 138.73, 135.17,
134.76, 127.96, 126.12, 125.96, 40.10, 37.26, 21.87, 15.47.
Kinetics and product studies
The relative response factor of 8-OH and alkene 11 was
determined as follows: Pure 8-OH in acetonitrile was analyzed
by HPLC. Then, 200 µl of this solution was transferred to a 2
ml measuring flask, after which aqueous perchloric acid solu-
tion (600 µl, 1 M) was added. After 1 h, the reaction was com-
plete (products: 10 and 11); aqueous sodium acetate solution
(1000 µl, 1 M) was then added, the volume was adjusted to 2 ml
with acetonitrile, and the sample was analyzed again. The data
were used to calculate the relative response factors of 8-OH and
the corresponding alkene product 11.
The response factors for 2-OH and 3-OH and for the substi-
tution products with SCNϪ were assumed to be the same as for
1-OH.
The estimated errors are considered as maximum errors
derived from maximum systematic errors and random errors.
The maximum errors of the directly measured quantities were
thus allowed to propagate as systematic errors into derived
quantities, e.g. reaction rate constants.
The reaction solutions were prepared by mixing acetonitrile or
methanol with water at room temperature, ca. 22 ЊC. The reac-
tion vessel was a 2 ml HPLC flask, sealed with gas-tight PTFE
septa, which was placed in an aluminium block in the water
thermostat bath. The reactions were initiated by fast addition,
by means of a syringe, of a few microlitres of the substrate
dissolved in acetonitrile. The concentration of the substrate
in the reaction solution was usually about 0.01–0.1 mM. At
appropriate intervals, samples were analyzed using the HPLC
apparatus. The rate constants for the disappearance of the sub-
strates were calculated from plots of substrate peak area versus
time by means of a nonlinear regression computer program.
Very good pseudo-first-order behaviour was seen for all the
reactions studied. The separate rate constants for the elimin-
ation and substitution reactions were calculated by combin-
ation of product composition data, obtained from the peak
areas and the relative response factors determined in separate
experiments, with the observed rate constants.
The pseudo-first-order rate constant for the reaction of 1-Cl
with sodium methoxide (0.1 M) in methanol was measured as
above, except that samples (100 µl) of the reaction solutions
were quenched by a mixture of aqueous acetic acid (20 µl, 1 M)
and ethanol (500 µl) before HPLC analysis. The alkenes 4 and 5
isomerize to alkene 7 in the presence of strong base.
Acknowledgements
We thank Dr A. Gogoll for help with the structure assignments
and the Swedish Natural Science Research Council for support-
ing this work.
A control experiment showed that the alkene 6 is stable under
solvolytic conditions. The stability of the alkenes 4 and 5 was
checked in the following way: Solvolysis of 1-Cl in 75 vol%
acetonitrile in water gave alkenes 4 and 5 (ratio 15:85, GC
analysis). After the reaction was finished, the reaction mixture
References
1 M. M. Toteva and J. P. Richard, J. Am. Chem. Soc., 1996, 118,
11434.
J. Chem. Soc., Perkin Trans. 2, 1999, 1397–1404
1403

View Article Online
2 E. H. Boyd, R. W. Taft, A. P. Wolf and D. R. Christman, J. Am.
Chem. Soc., 1960, 82, 4729.
3 E. M. Arnett and T. C. Hofelich, J. Am. Chem. Soc., 1983, 105, 2889.
4 R. A. McClelland, V. M. Kanagasabapathy, N. S. Banait and
S. Steenken, J. Am. Chem. Soc., 1989, 111, 3966.
16 R. A. McClelland, V. M. Kanagasabapathy, N. S. Banait and
S. Steenken, J. Am. Chem. Soc., 1991, 113, 1009.
17 Based upon an assumed association constant for ion pair formation
from free ions of Kas = 0.25 MϪ1, cf. J. E. Prue, J. Chem. Soc., 1965,
7534.
5 W. K. Chwang, P. Knittel, K. M. Koshy and T. Tidwell, J. Am.
Chem. Soc., 1983, 105, 3395.
6 Y. Chiang, W. K. Chwang, A. J. Kresge, M. F. Powell and S. Szilagyi,
J. Org. Chem., 1984, 49, 5218.
7 Y. Chiang and A. J. Kresge, J. Am. Chem. Soc., 1985, 107, 6363.
8 Bentley and co-workers recently proposed that both elimination and
substitution from the tert-butyl carbocation may be enforced
concerted reactions: T. W. Bentley, G. Llewellyn and Z. H. Ryu,
J. Org. Chem., 1998, 63, 4654.
9 M. Cocivera and S. Winstein, J. Am. Chem. Soc., 1963, 85, 1702.
10 A. Thibblin, Chem. Soc. Rev., 1993, 22, 427.
11 A. Thibblin and Y. Saeki, J. Org. Chem., 1997, 62, 1079.
12 A. Thibblin, J. Am. Chem. Soc., 1987, 109, 2071.
13 A. Thibblin, J. Am. Chem. Soc., 1988, 110, 4582.
14 A. Thibblin and H. Sidhu, J. Chem. Soc., Perkin Trans. 2, 1994, 1423.
15 (a) R. Ta-Shma and Z. Rappoport, J. Am. Chem. Soc., 1983, 105,
6082; (b) J. P. Richard and W. P. Jencks, J. Am. Chem. Soc., 1984,
106, 1373.
18 A. Thibblin, J. Org. Chem., 1993, 58, 7427.
19 A. Thibblin, J. Am. Chem. Soc., 1989, 111, 5412.
20 F. G. Bordwell and G. J. McCollum, J. Org. Chem., 1976, 41, 2391.
21 Q. Meng and A. Thibblin, J. Am. Chem. Soc., 1997, 119, 4834.
22 T. L. Amyes and J. P. Richard, J. Am. Chem. Soc., 1991, 113,
8960.
23 C. Paradisi and J. F. Bunnett, J. Am. Chem. Soc., 1985, 107,
8223.
24 X. Creary, V. P. Casingal and C. E. Leahy, J. Am. Chem. Soc., 1993,
115, 1734.
25 A. Thibblin, J. Phys. Org. Chem., 1989, 2, 15.
26 See ref. 21 and references therein.
27 A. Thibblin, J. Phys. Org. Chem., 1992, 5, 367.
28 C. A. Bunton, N. Carrasco, N. Cully and W. E. Watts, J. Chem. Soc.,
Perkin Trans. 2, 1980, 1859.
Paper 9/01296F
1404
J. Chem. Soc., Perkin Trans. 2, 1999, 1397–1404