A mechanism for efficient proton-transfer catalysis. Intramolecular general acid catalysis of the hydrolysis of 1-arylethyl ethers of salicylic acid

Article abstract of DOI:10.1139/v99-080

The tert-butyl (1) and 1-arylethyl ethers (2) of salicylic acid are hydrolyzed with efficient general acid catalysis by the ortho-COOH group. The half-life of the neutral COOH form of the tert-butyl ether is 15.2 min at 39°, and the estimated acceleration by the COOH group of 2, X = Me, Y = H is 2.13 × 10⁵. The salicylate leaving group from 2 (X = Me, Y = H) has an effective pK_a of 2.9, compared with a nominal pK_a of 8.52. Analysis of substituent effects in both arylethyl and leaving groups provides the most detailed available mechanistic insight into a reaction involving efficient intramolecular proton-transfer catalysis. The mechanism is very different from classical general acid-base catalysis. Proton transfer takes place very rapidly within a developing strong hydrogen bond, and though an integral part of the C - O cleavage process is practically uncoupled from it. "Strategic delay" of the proton-transfer step, relative to C - O cleavage, makes a significant contribution to efficiency by setting up the conditions for the formation of the strong, intramolecular hydrogen bond.

Full text of DOI:10.1139/v99-080

792
A mechanism for efficient proton-transfer
catalysis. Intramolecular general acid catalysis
of the hydrolysis of 1-arylethyl ethers of
salicylic acid
Sarah E. Barber, Kathryn E.S. Dean, and Anthony J. Kirby
Abstract: The tert-butyl (1) and 1-arylethyl ethers (2) of salicylic acid are hydrolyzed with efficient general acid
catalysis by the ortho-COOH group. The half-life of the neutral COOH form of the tert-butyl ether is 15.2 min at 39°,
and the estimated acceleration by the COOH group of 2, X = Me, Y = H is 2.13 × 105. The salicylate leaving group
from 2 (X = Me, Y = H) has an effective pKa of 2.9, compared with a nominal pKa of 8.52. Analysis of substituent
effects in both arylethyl and leaving groups provides the most detailed available mechanistic insight into a reaction
involving efficient intramolecular proton-transfer catalysis. The mechanism is very different from classical general
acid–base catalysis. Proton transfer takes place very rapidly within a developing strong hydrogen bond, and though an
integral part of the C—O cleavage process is practically uncoupled from it. “Strategic delay” of the proton-transfer
step, relative to C—O cleavage, makes a significant contribution to efficiency by setting up the conditions for the
formation of the strong, intramolecular hydrogen bond.
Key words: catalysis, carboxyl, hydrogen bond, proton transfer, enzyme mechanism.
Résumé : L’hydrolyse des éthers tert-butyle (1) et 1-aryléthyle (2) de l’acide salicylique se fait avec une catalyse acide
générale efficace par le groupe -COOH en ortho. La demi-vie de la forme COOH neutre de l’éther tert-butyle est de
15,2 min à 39° et l’accélération évaluée de groupe COOH de 2, X = Me, Y = H, est de 2,13 × 105. Le pKa effectif du
groupe salicylate partant de 2 (X = Me, Y = H) est égal à 2,9 par rapport à un pKa nominal de 8,52. L’analyse des
effets de substituants du groupe arylméthyle et des groupes partants fournit la compréhension mécanistique la plus
détaillée possible pour une réaction impliquant une catalyse efficace par transfert intramoléculaire de proton. Le
mécanisme est très différent de la catalyse générale acide–base classique. Le transfert de proton se produit très
rapidement dans la forte liaison hydrogène qui se développe et, même s’il fait partie intégrale du processus de clivage
de la liaison C—O, il n’existe pratiquement pas de couplage avec lui.
Mots clés : catalyse, carboxyle, liaison hydrogène, transfert de proton, mécanisme enzymatique.
[Traduit par la Rédaction] Barber et al. 801
We are interested in the efficiency and thus the detailed
intramolecular hydrogen bond as the reaction proceeds. For
optimum catalytic efficiency this H-bond should be absent
in the ground state and strongest in the transition state, a
state of affairs that depends critically on geometry and the
extent and timing of charge development during the reac-
tion. We report an in-depth investigation of general acid ca-
talysis by the carboxyl group of nucleophilic substitution at
carbon (SN1 and SN2), a reaction that is particularly well un-
derstood. The leaving group is the salicylate monoanion,
known to be stabilised by a strong H-bond of well-defined
geometry, and the substrates its methyl, tert-butyl 1, and
substituted 1-phenylethyl ethers 2. We can vary the degree
of charge development on the leaving group oxygen by
varying the carbon centre from which it leaves systemati-
cally in the series 2, by changing the pattern of substitution
on both the salicylate leaving group and on the aromatic ring
of the 1-phenylethyl group (4).
mechanism of intramolecular general acid–base catalysis, as
a simple model for the myriad of proton-transfer processes
involved in catalysis by enzymes (1). General acid–base ca-
talysis may be presumed to be efficient in enzyme reactions:
but effective molarities (EM) for intramolecular proton-transfer
catalysis are generally lower by many orders of magnitude
than those for intramolecular nucleophilic (cyclization) reac-
tions, rising above 1–10 M only in special cases (2). We
have recently reported systems showing intramolecular gen-
eral acid catalysis with EMs as high as 106 M, (3) and have
identified as a key factor the development of a strong
Received December 2, 1998.
This paper is dedicated to Professor A. Jerry Kresge on the
occasion of his 70th birthday.
S.E. Barber, K.E.S. Dean, and A.J. Kirby.1 University
Chemical Laboratory, Cambridge CB2 1EW, U.K.
Kinetic methods and results
1Author to whom correspondence may be addressed.
Telephone:+44 1223 336370 (direct). Fax: +44 1223 336913.
e-mail: ajk1@cam.ac.uk
The tert-butyl ether 1 was made directly, by carboxylation
of the ortho-lithiated tert-butyl phenyl ether: it disappeared
Can. J. Chem. 77: 792–801 (1999)
© 1999 NRC Canada

Barber et al.
793
Table 1. Kinetic data for hydrolysis of tert-butyl salicyl ether 1
and its methyl ester, in 10% (v/v) dioxan–water, ionic strength
1.0 M (KCl).
For comparison with the reactions of the salicylic acid de-
rivatives 2, we measured rates of hydrolysis under the same
conditions for a series of simple sterically comparable aryl
1-phenylethyl ethers 4, with good leaving groups and a 2-
NO2 in place of the o-COOH group. These compounds are
less reactive than the corresponding salicyl ethers and were
mostly studied at 80°C. The data appear in Table 3. The 2,4-
dinitrophenyl ethers (4, Y = NO2, X = H, and 4-Me) show
pH-independent hydrolysis between pH 4.5 and 9. These re-
actions are not catalysed by added buffer. (Some amine buff-
ers are an exception, but HPLC examination of the products
showed that the products were anilines, and thus that the
amines had attacked the aromatic ring, rather than the benzylic
centre.) The addition of the strongly nucleophilic thiosul-
phate (Swain–Scott (5) n value –6.4) had no effect on the
rate of hydrolysis of 4 (Y = NO2, X = H), and in the case of
4 (Y = NO2, X = 4-Me) a six-fold increase in the concentra-
tion of thiosulphate gave only a 20% increase in the rate
constant. This increase is so small that it is reasonable to
suppose that less powerful nucleophiles would have little or
no effect on the rates for these compounds. In other cases, a
small negative effect of increasing buffer concentration could
be ascribed to specific salt effects.
In only one case was there convincing evidence for buffer
catalysis, and thus for an SN2 reaction at the benzylic centre.
The least reactive 2,4-dinitrophenyl ether 4 (Y = NO2, X =
3-Br) also showed a pH-independent reaction between
pH 4.5 and pH 6.3, at 80°C. Pseudo-first-order behaviour
could be observed only at buffer concentrations of 0.08 M or
less, but at these concentrations the reaction was catalysed
by the free-base form of the buffer (linear dependence on ac-
etate or phosphate concentration, 52% acceleration for
0.08 M phosphate, 50% free base). Extrapolation of the ob-
served rate constants to zero buffer gave the same intercept
for both phosphate and acetate buffers (and second-order
rate constants of 3.04 × 10–4 and 5.66 × 10–5, respectively).
It seems clear that in this case, against the slowest spontane-
ous reaction, acetate and phosphate anions do provide some
nucleophilic catalysis.
This result suggested that the 1-(3-bromophenyl)ethyl
salicyl ethers would be the most likely to show second-order
reactions with nucleophiles, but the addition of bromide
(Swain–Scott n value –3.9) had no effect on the rate of hy-
drolysis of (2, Y = H, X = 3-Br), and the reaction was not
catalysed to a measurable extent by the base component of
added buffers. The reactions of (2, Y = 5-NO2, X = 3-Br)
were studied in the presence of substituted pyridines. These
were of particular interest because they catalyse the hydroly-
sis of salicyl phosphate (6), and we would be able to compare
the two βnuc values obtained. However, the rate of reaction
was independent of the concentration of cyanopyridine, and
a slight rate retardation was seen in the presence of increas-
ing concentrations of nicotinamide.
It is difficult to observe SN2 reactions with the 1-arylethyl
ethers 2 because of the dominant SN1 reactions of these sys-
tems, so we searched carefully for reactions of nucleophiles
with the methyl ether 5. A UV study in water (0.1 M HCl
and formate buffer) showed no detectable reaction with 0.9 M
iodide or thiocyanate (we should have detected a reaction
with a first-order rate constant ≥10–8 s–1), and we could de-
tect no reaction by 1H NMR with iodide in acetonitrile,
103kH
104k0
pKapp
k0H͞k0D
T (°C)
(dm3 mol–1 s–1)
(s–1)
25
39
50
65a
4.05
4.16
4.27
4.29
10.9 ± 0.4
85.0 ± 3.0
365 ± 8.5
2022 ± 73
1.04 ± 0.07
7.6 ± 0.3
30.9 ± 0.9
122 ± 4
1.36 ± 0.16
Methyl ester
90.0 ± 0.3
39
aVariable temperature data give for k0: ∆H‡ = 98 ± 3 kJ mol–1, ∆S‡ = 8
± 9 J K–1 mol–1; for kH: ∆H‡ = 107 ± 3kJ mol–1, ∆S‡ =77 ± 8 J K–1 mol–1.
rapidly in organic solvents, but was stable in aqueous solu-
tion as the anion. Compounds of the 1-phenylethyl series
2 were generated from the stable methyl esters (3): the esters
were made by the Williamson ether synthesis, as described
in the experimental section. Stock solutions of the stable
carboxylate anions of ethers 2 were prepared by hydrolysing
the esters 3 with 1.1 equiv. of KOH in 50% dioxan–water.
For all but the most reactive esters, hydrolysis required
an overnight reflux. The resulting solutions were diluted,
and used directly in kinetic studies (final concentration
10–4–10–5 mol dm–3).
Table 1 shows kinetic data for the hydrolysis of the tert-
butyl ether 1, and Table 2 summarizes results for the series
of 16 1-arylethyl ethers 2 studied in this work. Rates are
very sensitive to substitution in the aromatic ring of the 1-
arylethyl derivatives, and pH–rate profiles were measured
conveniently at 25, 39, and 65°C, respectively, for com-
pounds with X = 4-Me, H, and 3-Br. (In the case of the very
reactive compound 2, X = 4-MeO, Y = 5-NO2, hydrolysis
was too fast to follow below pH 4, even at 1°C, and we
could only estimate the plateau rate constant.) Derived rate
constants were extrapolated to a common 39°C for the 4-Me
and 3-Br systems from variable temperature measurements,
which gave the thermodynamic parameters also shown in
Table 2.
The reactions of all the salicylic acid derivatives show the
same basic features, illustrated in Fig. 1 for the hydrolysis of
1 and 2 (X = H, Y = H, and X = H, Y = NO2); namely, rapid
acid-catalysed hydrolysis at low pH, a plateau between
pH 2–4 corresponding to the pH-independent hydrolysis of
the COOH form, and no significant reaction above pH 6.
The plateau reaction is of primary interest for this work, and
derived kinetic data in the tables refer to rate constants k0 for
this process. Rate constants and apparent carboxyl-group
pKa values were calculated from eq. [1]:
k0 + kHaH
[1]
kobs =
1 + (Ka͞aH)
where kH is the rate constant for the acid-catalysed hydroly-
sis of the neutral acid form and k0 the first-order rate constant
calculated for its spontaneous hydrolysis. (The alternative
interpretation as the kinetically equivalent acid-catalysed hy-
drolysis of the carboxylate anion has been ruled out in many
similar cases.) The pKapp values agree with the values ex-
pected for the pKa values of salicylic acid derivatives.
© 1999 NRC Canada

Table 2. Kinetic data for hydrolysis of 1-phenylethyl salicyl ethers 2 at various temperatures, in 10% (v/v) dioxan–water, ionic strength 1.0 M (KCl).
Compound 2
∆H‡ (kJ mol–1)
∆S‡ (J K–1 mol–1)
pKapp
k0H͞k0D
104k0 (s–1)
103kH (dm3 mol–1 s–1)
k0(39) (s–1)a
X
Y
T (°C)
4-Me
4-Me
4-Me
4-Me
4-Me
H
H
H
H
H
3-Br
3-Br
3-Br
3-Br
3-Br
4-MeO
5-Me
H
4-MeO
5-Cl
5-NO2
5-Me
H
4-MeO
5-Cl
5-NO2
5-Me
H
4-MeO
5-Cl
25
25
25
25
25
39
39
39
39
39
65
65
65
65
65
1–26
4.53 ± 0.11
4.12 ± 0.03
4.83 ± 0.07
3.79 ± 0.06
3.36 ± 0.06
4.17 ± 0.08
3.94 ± 0.02
4.53 ± 0.05
3.40 ± 0.05
2.92 ± 0.03
4.16 ± 0.05
3.83 ± 0.07
4.43 ± 0.07
3.44 ± 0.06
3.13 ± 0.07
2.93b
0.96
0.24
0.36
0.42
0.67
1.51
2.35
3.44
4.30
8.10
33.6
0.97
1.33
2.16
3.60
13.4
—
1.55 ± 0.15 × 10–2
2.57 ± 0.15 × 10–2
2.49 ± 0.30 × 10–2
6.21 ± 1.00 × 10–2
0.160 ± 0.03
2.35 ± 0.23 × 10–4
3.44 ± 0.17 × 10–4
4.30 ± 0.21 × 10–4
8.10 ± 0.40 × 10–4
3.36 ± 0.17 × 10–3
3.0 ± 1.0 × 10–6
9.0 ± 1.5 × 10–6
2.1 ± 0.5 × 10–5
1.20 ± 0.13 × 10–5
1.05 ± 0.09 × 10–4
55 ± 10b
—
—
0.27
0.28
—
105 ± 6
96 ± 7
120 ± 10
129 ± 9
—
—
—
—
—
114 ± 6
88 ± 6
76 ± 10
110 ± 6
83 ± 10
—
63 ± 10
32 ± 23
116 ± 20
156 ± 30
—
—
—
—
—
14 ± 30
–59 ± 17
–91 ± 29
12 ± 12
–53 ± 4
—
—
9.54
4.56
8.66
3.46
—
11.6
4.63
9.63
—
0.97
5-NO2
5-NO2
1.36
—
aRate constant measured at (X = H and X = Y = Me) or extrapolated to 39°C from measurements over a range of temperatures.
bEstimate based on measurements between 1 and 26°C (too fast to measure below pH 4). Extrapolation of the data to 39°C allowed the construction of a partial pH–rate profile.

Barber et al.
795
DMSO (both run for 2 weeks at 40°C) or water (5 days at
60°). We find no measurable nucleophilic catalysis for any
of the salicyl ethers 1 and 2.
nol oxygen is not normally a viable leaving group in the SN1
or SN2 reaction in the absence of very strongly electron-
withdrawing substituents on the aromatic ring, and the rapid
cleavage of the ether C—O bonds confirms that catalysis by
the o-COOH group is efficient in these systems.
Figure 1 shows pH–rate profiles for the hydrolysis of
three representative alkyl salicyl ethers, the unsubstituted
tert-butyl (1) and 1-phenylethyl derivatives and the 5-nitro-
Discussion
Ethers 1 and 2 were designed to show intramolecular gen-
eral acid-catalysed hydrolysis at a measurable rate.2 A phe-
2 The terms general acid and general base catalysis are commonly used to denote either or both of (i) a kinetic dependence on the concentra-
tion of the general acid or base, and (ii) the mechanism of this catalysis, involving rate-determining proton transfer from or to the general
acid or base. The intramolecular reaction in the salicylic acid system is so efficient that catalysis by external general acids is not significant.
© 1999 NRC Canada

796
Can. J. Chem. Vol. 77, 1999
Table 3. Kinetic data for hydrolysis of aryl 2-nitro-4-substituted-phenyl ethers 4 at various temperatures, in 10% (v/v)
dioxan–water, ionic strength 1.0 M (KCl).
Compound 4
k0 (s–1)
X
Y
T (°C)
∆H‡ (kJ mol–1)
∆S‡ (J K–1 mol–1)
4-Me
4-Me
4-Me
4-Me
H
CF3
H
Cl
NO2
NO2
NO2
80
80
80
39
65
80
6.2 × 10–3
1.7 × 10–4
7 × 10–4
1.87 × 10–3
3.84 × 10–4
4.22 × 10–5
91 ± 9
–6.68 ± 3.7
–49.5 ± 2
3-Br
98.4 ± 1
Fig. 1. pH–rate profiles (at 39°C) for the hydrolysis of tert-butyl
(1, circles) and salicyl and 5-nitrosalicyl 1-phenylethyl ethers (2,
Y = H and NO2: squares and diamonds, respectively).
compounds 4, with X = 4-Me to the pKa of the leaving
group is –0.95 ± 0.03 at 80°C. This allows us to estimate an
acceleration of 2.13 × 105 by the COOH group of 2, X =
Me, Y = H (at 39°C, based on an estimated pKa of 8.52 (7)
for the phenol OH of salicylic acid). The observed rate of
the COOH-catalysed reaction of this compound is equivalent
to that expected for the spontaneous departure of the anion
derived from a phenol of pKa = 2.9. The effective pKa of the
leaving group is close to 3 for similar reactions of salicyl
acetals (8, 9) and of salicyl phosphate (6).
The effect of nucleophiles on the expulsion of salicylate
The evidence indicates that the enhanced salicylate leav-
ing group departs in a simple SN1 reaction, and we discuss
the details of this process below. We looked also for evi-
dence of catalysis of the attack of nucleophiles (the SN2 re-
action). We could find no detectable reaction at the methyl
group of salicylic acid methyl ether 5; and would expect
none for the reaction of the tert-butyl ether 1.
The work of Richard and Jencks (10, 11) has defined the
point at which the nucleophilic substitution reactions of 1-
phenylethyl derivatives become coupled concerted. No reac-
tions of 1-(4-methylphenyl) ethyl derivatives are concerted:
unsubstituted 1-phenylethyl derivatives show a coupled con-
certed reaction only with azide, but 1-(3-bromophenyl)ethyl
chloride shows coupled concerted reactions with azide and
with acetate ion, though not with solvent. Consistently, we
find evidence for weak catalysis by acetate and phosphate
anions of the cleavage of 1-(3-bromophenyl)ethyl 2,4-
dinitrophenylether 4 (Y = NO2, X = 3-Br) at 80°C, but for
no other derivative 4.
Thus, our 1-(3-bromophenyl)ethyl salicyl ethers 2 (Y = 3-
Br) are the most likely to show second-order reactions with
nucleophiles. We found no evidence of such reactivity. The
thermodynamic activation parameters for the cleavage reac-
tion (Table 2) similarly offer no evidence to support
nucleophilic participation by water in the transition state for
hydrolysis: entropies of activation for the most reactive
compounds 2, X = 4-Me are generally large and positive.
Those for the least reactive systems (2, Y = 3-Br) are signif-
icantly more negative, consistent with more ordering of sol-
vent in the transition states for hydrolysis. Nevertheless, we
do not consider that these reactions are solvent assisted,
since far better nucleophiles than water show no effect.
These activation parameters can be compared with those
obtained for the spontaneous hydrolyses of two 2,4-dinitro-
phenyl ethers 4 (Table 3). Here also the entropy of activation
for the hydrolysis of (4, X = 3-Br, Y = NO2), which shows
salicyl 1-phenylethyl ether (2, X = H, Y = NO2), with the ni-
tro group para to the leaving group oxygen. The salicylic
acid derivatives 2 studied in this work all showed similar be-
haviour: more or less well-defined pH-independent regions be-
tween pH 2–4, with rates falling off at high pH with the
fraction of COOH form present. The presence of this feature
for the salicylic acid derivatives, and its absence for their
methyl esters is prima facie evidence for catalysis by the
carboxylic acid group. The half-life of the neutral COOH
form of tert-butyl ether 1 is 15.2 min at 39°, less than half
that of the unsubstituted 1-phenylethyl ether 2 (X = Y = H,
in 10% dioxan:water); which is itself accelerated by a factor
of 10 by the introduction of the 5-nitro group.
At low pH specific acid catalysis is observed, for all ex-
cept the ethers derived from 5-nitrosalicylic acid. (This latter
effect is the simple consequence of a high rate of spontane-
ous hydrolysis combined with a lower kH for the compounds
with the most electronegative leaving groups.)
These rapid hydrolysis reactions are an indication of the
efficiency of intramolecular catalysis by the o-COOH group
in the reactions of salicylic acid derivatives 1 and 2. The
corresponding intermolecular reaction, general acid catalysis
of the hydrolysis of similar ethers by carboxylic acids, is too
slow to measure, so we cannot estimate effective molarities
(EM) (2). As an alternative. we can quantify efficiency by
the comparison with the spontaneous reactions of simple
ethers 4. The slope of the linear free energy relationship
(βLG) relating the rates of hydrolysis of the series of four
© 1999 NRC Canada

Barber et al.
797
Table 4. Linear free energy relationships for 1-phenylethyl salicyl ethers 2.
Compound 2
c
c
ρphenol
ρCOOH
X
Y
T (°C)
Hydrolysis reactions
ρ+a
Varied
Varied
Varied
Varied
Varied
5-Me
H
4-MeO
5-Cl
39
39
39
39
39
–5.27 ± 0.1
–4.91 ± 0.5
–4.34 ± 0.7
–5.29 ± 0.4
–4.53 ± 0.4
5-NO2
b
ρLG (OAr)
4-Me
H
3-Br
Varied
Varied
Varied
25
39
65
0.84 ± 0.08
1.25 ± 0.08
1.23 ± 0.09
0.97 ± 0.10
0.96 ± 0.10
0.89 ± 0.30
–0.08 ± 0.15
0.23 ± 0.14
0.08 ± 0.40
Ionization of COOH groupd
d
Salicylic acids
ρCOOH
Variede
Variedf
Varied
Varied
Varied
25
25
25
39
65
0.89 ± 0.06
0.96 ± 0.04
1.45 ± 0.19
1.63 ± 0.09
1.32 ± 0.16
0.10 ± 0.05
0.25 ± 0.11
0.03 ± 0.44
0.50 ± 0.38
0.26 ± 0.57
0.82 ± 0.04
1.04 ± 0.07
1.31 ± 0.30
1.96 ± 0.26
1.82 ± 0.38
4-Me
H
3-Br
aBased on standard Hammett equation using substituent constants σ+.
bBased on standard Hammett equation using substituent constants σ, relative to leaving group phenol O.
cBased on extended Hammett equation (17) log (kY/kH) = ρCOOHσCOOH + ρphenolσphenol, using substituent constants σ relative to
COOH and leaving group phenol O, respectively. All substituent constants are taken from Jaffé’s compilation.
dBased on standard Hammett equation using substituent constants σ, relative to the COOH group.
eCorrelation for pKa values of 17 substituted acids (28).
fCorrelation using pKa values of the five substituents used in this work (28).
evidence of nucleophilic participation with nucleophiles, is
more negative than that for the hydrolysis of 4 (X = 4-Me, Y
= NO2), which does not. However, the differences are small
and entropies of activation are unreliable mechanistic indica-
tors, particularly in the borderline region which we are con-
sidering. It seems clear that the cleavage of the C—O bond
of the salicyl ethers 2 generally involves — at the carbon
centre — simple, uncoupled SN1 mechanisms.
mechanism as the cation lifetimes gets close to the point at
which its existence may be called into question.
The correlations are generally good and give values of ρ+
from –4.5 to –5.3. A number of literature values of ρ+ are
available for comparison, for the solvolysis of various
1-phenylethyl derivatives. For substituted 1-phenylethyl chlo-
rides in 80% aqueous acetone, ρ+ = –5.5 (14); for 1-phenyl-
ethylacetates in 30% aqueous ethanol, ρ+ is –5.7 (15); and
for 1-phenylethyl diphenylphosphinates in 80% ethanol–wa-
ter, ρ+ is –5.1 (16). Also Richard et al. (13) correlated rate
constants for the acid-catalysed cleavage of 1-phenylethyl
alcohols to form the corresponding cations in 1:1 trifluoro-
ethanol:water), obtaining ρ+ = –4.5. Similar values were ob-
served in this work for the hydrolyses of simple 2,4-
dinitrophenyl ethers 4 (Table 3). These are correlated by the
Hammett equation with ρ+ = –4.9 ± 0.5.
The effect of changing cation lifetime
As the substituent X becomes less electron-donating, the
lifetime of the 1-phenylethyl cation (6) decreases. From the
work of Richard et al. (11–13), we know the lifetimes of
such cations in 1:1 trifluoroethanol:water. These lifetimes
were determined by competition experiments between sol-
vent and azide, or for short-lived cations, by extrapolation of
the data obtained for long-lived cations. In water, the life-
times are reduced by a factor of two.3 Thus, we can study
the effect of the substituent X on the rate of hydrolysis of 1-
phenylethyl salicyl ethers and relate this to the cation life-
time.
For a fixed salicylate leaving group (2, constant Y) the
plateau rate constants, corrected to 39°C, are correlated
by the Hammett equation using substituent constants σ+
(Table 4). The plots are linear over a range of cation life-
times, from 10–8 s (4-MeO substituent) to 5 × 10–13 s (3-Br
substituent). There is no evidence of a change in catalytic
These large, negative values of ρ+ indicate similar, sub-
stantial charge development on carbon. The fact that the val-
ues obtained for the hydrolysis of 1-phenylethyl salicyl
ethers are close to those obtained for the solvolyses of other
1-phenylethyl derivatives show that charge development on
carbon is also similar.
A more meaningful estimate of the extent of charge devel-
opment on carbon is obtained by comparing ρ+ with the
equilibrium value of for complete C—O cleavage, i.e., for
the formation of 1-phenylethyl carbocations, ρ+eq. Richard et
al. (13) determined ρ+ not only for the acid-catalysed cleav-
3 J.P. Richard. Personal communication.
© 1999 NRC Canada

798
Can. J. Chem. Vol. 77, 1999
Fig. 2. More O-Ferrall – Jencks potential energy diagram for the
intramolecular general acid-catalysed cleavage of 1-arylethyl
ethers (2) of salicylic acid to salicylate anion and the 1-arylethyl
cation 6. A and B represent, respectively, the approximate
locations of the transition state for the reverse reaction, the
general base-catalysed hydrolysis of the cations (21), and the
forward reaction studied in this work. For details see the text.
bond cleavage to oxygen is accompanied by very little pro-
ton transfer from the COOH group in the transition state for
C—O cleavage. These data complement those for charge de-
velopment at carbon: rather less than half a full negative
charge appears to have developed on the leaving group oxy-
gen in the transition state, as expected if the bond is about
half broken, with an apparently very small amount of the de-
veloping charge neutralized by proton transfer.
Further evidence that proton transfer in the transition state
is not far advanced comes from the solvent deuterium iso-
tope effects. These were measured for four compounds, and
are presented in Tables 1 and 2. The values of kH0͞kD0 are
similar to those obtained for other salicylate derivatives (6,
9), and lead, as before, to the conclusion that proton transfer
in the transition state is not as far developed as in a classical
general acid-catalysed reaction.
The ρcarboxyl values could be converted into α-values by
dividing them by ρ for the ionization of the carboxylic acid,
but in this case the operation is of little significance since
the ρ values are already close to zero. (The apparent pKa val-
ues for the five substituted compounds are correlated reason-
ably well by the simple Hammett equation but, surprisingly,
give ρ values substantially greater than 1, especially for the
set with X = H (Table 4). The value ρ for the ionization of
benzoic acids is 1 by definition, and generally close to 1 for
substituted benzoic acids.)
There is no evidence for any change in α as the stability
of the cation changes (X is varied), indicating that the inter-
action coefficient Pxy is zero. It is not possible to assess
whether α changes with changing pKa of the phenol, because
the Jaffé treatment assumes ρphenol and ρcarboxyl to be constant
over the range of compounds studied. If this assumption
were not valid the Jaffé equation would not give linear plots.
In fact, for one subset of ethers (with X = 3-Br) the plot
does not give a good straight line (see the data in Table 3).
However, if the extent of proton transfer to the leaving
group oxygen were changing over this series, we would not
expect the simple Hammett plot (which assumes that the
substituent Y affects the phenol oxygen only) to give a good
straight line; but it does (r = 0.992). We conclude that the
nonlinear Jaffé plot is not indicative of varying extents of
proton transfer, but rather reflects experimental error.
A final question is whether the extent of bond cleavage to
the leaving group in the transition state changes with the sta-
bility of the cation. For values of ρphenol obtained from the
simple Hammett correlation (Table 4), ρphenol is significantly
smaller for X = 4-Me than for X = H and X = 3-Br. How-
ever, the difference disappears when the Jaffé treatment is
applied, and we conclude that this variation also is unlikely
to be significant. Our general conclusion is that, as far as
C—O cleavage is concerned, the expulsion of salicylate
closely resembles the spontaneous expulsion of good leaving
groups.
age of substituted 1-phenylethyl alcohols, but also for the re-
verse reaction, the addition of water to the cations, for which
ρ+ is +4.5. The equilibrium value ρ+ (ρ+
– ρ+reverse) is
eq
therefore –9. Using this as an estimate forfρor+weaqrdfor the forma-
tion of phenylethyl cations from our ethers suggests that the
cleavage of the C—O bond in our salicyl ethers 2 has pro-
ceeded about halfway towards the cation in the transition
state.4
The effect of changing leaving group
When the substituent Y on the aromatic ring of salicylate
is varied, both the leaving group capability of the phenol and
the strength of the catalytic carboxylic acid group change.
Jaffé’s simple extension of the Hammett equation (17)
(eq. [2]) allows the two effects to be separated.
[2]
log (kH/kY)/σcarboxyl
= ρcarboxyl + (ρphenol × σphenol)/σcarboxyl
Values of σphenol and ρcarboxyl obtained from the Jaffé plots
are also given in Table 4. It is evident that in the transition
state for the hydrolysis of these 1-phenylethyl salicyl ethers,
there is extensive bond cleavage to the leaving group oxy-
gen. The values of ρphenol may be converted to βLG values by
dividing by ρ for the ionization of a phenol, –2.28 (18). The
values obtained should be close to normalized βLG, since
substituent effects on the carbon basicity of phenolate anions
are the same, within experimental error, as those on their
proton basicity (19, 20). They are also close to those
(0.4–0.5) reported for other salicylic acid acetals (9). So too
are the very low values of ρcarboxyl, indicating that substantial
The mechanism for efficient general acid catalysis
The detailed mechanism which emerges for proton-
transfer catalysis appears to be very different from the mech-
anism proposed by Jencks and his co-workers (21, 22) for
4 There is some variation in the values of ρ+ as the leaving group changes (Table 4), but the variation is not systematic, and the errors too
large to convince us that the effect is significant.
© 1999 NRC Canada

Barber et al.
799
Scheme 1.
the Bronsted α coefficient) lag behind changes in local bond
order.
Mechanism and efficiency in intramolecular general
acid catalysis
This mechanism in which a rapid proton-transfer step forms
an integral part of C—O cleavage, yet is largely uncoupled
from it, is the most detailed available for a reaction involv-
ing efficient intramolecular proton-transfer catalysis. An im-
portant question is whether it is peculiar to the salicylic acid
leaving group: the potential delocalization of the developing
negative charge of the leaving group oxygen into the COOH
group is a relevant special feature of the salicylate system.
Comparisons of ρ values for the leaving group suggest that
this is not a major factor.
This mechanism also differs significantly from a classical
general acid catalysis mechanism. Most well-characterized
general species catalysed reactions appear to involve proton
transfers strongly coupled to bond making or breaking pro-
cesses, as indicated by Bronsted coefficients in the region
0.5 ± 0.2. However, intramolecular versions of these reac-
tions are characteristically inefficient. Features of the pro-
posed mechanism which clearly favour efficient catalysis are
the development of a strong hydrogen bond in the transition
state, and its absence in the ground state (1). This makes
possible the horizontal section of the pathway (Fig. 2) for ef-
ficient catalysis.
This strategic delay in setting up the hydrogen bond that is
central to efficient proton-transfer catalysis is incidental to
the salicylate system, but could clearly make an important
contribution to catalysis in an enzyme active site, where dis-
tance and geometry are under much closer control. The pre-
liminary stages of C—O bond breaking, not a problem in
reactive model systems, require separate, perhaps nucleo-
philic assistance in the case of the poor leaving groups in-
volved in most enzyme-catalysed cleavages of C—OR and
P—OR bonds, so details of systems where the two modes of
catalysis might work together are of great potential interest.
In this context, we find no evidence for catalysis of nucleo-
philic attack on any ether of salicylic acid. Such negative ev-
idence is not strong, but the leaving group is good enough
for nucleophilic attack to be expected for simple 1-(3-bro-
mophenyl)ethyl ethers, on the basis of its observation for the
2,4-ditrophenyl derivative. We can speculate that the timing
of the proton-transfer step may be too critical to be coupled
with the attack of an outside nucleophile in this very simple
model.
the reverse reaction, the general base-catalysed addition of
alcohols to 1-phenylethyl cations. The contrast is illustrated
in Fig. 2, a More O-Ferrall – Jencks diagram based on the
published plot of Richard and Jencks (21). These authors
found that addition of alcohols and phenols to moderately
stable arylethyl cations 6 is general base catalysed, and shows
very small β values: corresponding to large α-values, close
to unity, for general acid catalysis of ether cleavage. We by
contrast observe very small α-values for the cleavage of the
salicyl ethers 2. Richard and Jencks proposed a reaction co-
ordinate close to vertical, and thus very close to the pathway
for specific acid catalysis (broken arrow, transition state A
in Fig. 2) for the reaction of the 1-(4-methoxyphenyl)ethyl
cation with trifluoroethanol, suggesting that general base ca-
talysis will not be observed at all for the addition of alcohols
to less stable carbocations. Thus, the evidence indicates that
the expulsion of the phenol oxygen catalysed by the ortho-
carboxylic acid group of salicylate, for which the transition
state (B) appears to be located on the opposite side of Fig. 2,
is quite different from the same reaction catalysed by exter-
nal general acids.
This apparent divergence can be resolved if the (vertical)
C—O cleavage and (horizontal) proton-transfer processes
(Fig. 2) are only very weakly coupled, as indicated by the
bold, broken reaction coordinate, with the transition state at
B. We have pointed out previously, for the similar departure
of salicylate from salicyl phosphate (6), that in the ground
state of salicylic acid derivatives in aqueous solution the car-
boxyl group is likely to be rotated out of the plane of the ar-
omatic ring (as it is in crystal structures of salicylic acid
derivatives), while proton transfer takes place most efficiently
in the plane, within a developing strong hydrogen bond. At a
given stage of the C—O cleavage process of a reactive sys-
tem 2, the COOH group must rotate into plane to form the
hydrogen bond to the developing oxyanion. If this happens
at, or close to, the point where proton transfer becomes ther-
modynamically favourable, it marks the transition state for
the overall process. The apparent contradiction between the
observation of efficient proton-transfer catalysis and the evi-
dence that there is little proton transfer from the COOH
group in the transition state arises because the proton trans-
fer itself takes place within a strong intramolecular — and
thus low barrier (23–25) — hydrogen bond. Because the
proton-transfer step is so fast it (i) does not contribute sig-
nificantly to the free energy barrier to reaction, and (ii) takes
place faster than the solvation shell can respond (Scheme 1).
In this situation (related to the “imperfect synchronisation of
resonance” identified by Bernasconi (26) in reactions in-
volving proton transfer to carbon), substituent effects (here,
Experimental
Materials
Dry THF used in anion reactions was refluxed over so-
dium before use. Substituted methyl salicylates were ob-
tained commercially (Aldrich) or prepared by esterification
of the corresponding salicylic acid (BF3–methanol). Substi-
tuted acetophenones and chlorobenzenes were obtained
commercially (Aldrich or Fluka).
Column chromatography was performed on Merck Kieselgel
60 (7734) and flash chromatography on Merck Kieselgel 60
(9385). NMR spectra were recorded on a Varian EM390 in-
strument and IR spectra on a Perkin–Elmer 297 spectro-
© 1999 NRC Canada

800
Can. J. Chem. Vol. 77, 1999
Table 5. Summary of synthesis and characterization of substrate methyl esters 3.
Compound 2
X
Y
Method
Yield (%)
Found, M+
Required; for
4-Me
4-Me
4-Me
4-Me
4-Me
H
H
H
H
3-Br
3-Br
3-Br
3-Br
3-Br
5-Me
H
4-MeO
5-Cl
5-NO2
5-Me
4-MeO
5-Cl
5-NO2
5-Me
H
A
A
A
A
B
A
A
A
B
A
A
A
A
B
66
58
67
63
38
66
17
23
30
84
79
15
77
40
253.1233
270.1252a
300.1347
304.0863
284.0965a
270.1262
286.1218
259.0548a
301.0957b
348.0343
334.0129
364.0327
371.9756
379.0031
253.1229; C17H17O2
270.1256; C17H18O3
300.1362; C18H20O4
304.0866; C17H17ClO3
284.0944; C16H14NO4
270.1255; C17H18O3
286.1205; C17H18O4
259.0526; C17H17ClO2
301.09501; C16H15NO5
348.0360; C16H15BrO3
334.0144; C16H15BrO3
364.0310; C17H17BrO4
371.9767; C16H14BrClO3
379.0056; C16H14BrNO5
4-MeO
5-Cl
5-NO2
aLargest fragment, [M – Ome]+.
bCrystallized from ether–hexane as cubes, mp 95–96°C. Found: C 64.0, H 5.20, N 4.6; C16H15NO5 requires C 63.8, H 5.05, N 4.6%.
photometer. Mass spectra were determined on a Kratos MS
30 instrument. Melting points (uncorrected) were determined
on a Kofler hot stage apparatus.
1-Phenyl ethyl salicyl ethers 2 were prepared in situ by
the alkaline hydrolysis of their methyl esters 3. The esters 3
(Table 5) were in most cases made by the Williamson syn-
thesis, by alkylation of the sodium salt of the methyl
salicylate (Method A). Details follow for a representative
example.
cooled. Sodium 2-chloro-4-nitrobenzoate was prepared sepa-
rately by treating 2-chloro-4-nitro benzoic acid (0.61 g) with
sodium hydride (0.12 g) in dry THF to give a white slurry of
the sodium salt. Tetrabutylammonium bromide (0.1 g) was
added and the slurry poured quickly into the solution of
alcoholate anion. The mixture was stirred at 50°C for 24 h.
Trimethyloxonium tetrafluoroborate (0.89 g) was then added
and the mixture stirred for 1 h longer. During this time, the
cloudy solution of the acid salt clarified as the carboxylate
ion was methylated. The THF was removed under reduced
pressure and dichloromethane added. The dichloromethane
was extracted three times with 1 M HC1, three times with
1 M potassium bicarbonate, and once with saturated brine.
The organic layer was dried over sodium sulphate and the
solvent removed under reduced pressure to leave a brown
oil. The oil was purified by column chromatography on sil-
ica gel under pressure using 25% ether–hexane as eluant to
1-Phenylethyl (2-carbomethoxy)phenyl ether (3, X = H,
Y = H)
Methyl salicylate (4 mL) was added to a suspension of so-
dium hydride (1.36 g) in THF. When effervescence had
ceased the mixture was refluxed for 30 min, then cooled. 1-
Phenylethyl bromide (5.73 g) and tetrabutyl ammonium bro-
mide (0.97 g) were added and the mixture refluxed for 24 h.
THF was removed under reduced pressure and water and
ether added. The ether layer was separated and the water
layer washed twice more with ether. The combined organic
extracts were dried over sodium sulphate and solvent re-
moved under reduced pressure. The residue was purified by
column chromatography on silica gel under pressure using
10% ether–hexane as eluant, to give the ether (3, X = H, Y =
H; 5.68 g, 71%) as an oil. νmax: 1730, 1610, 1580, and
1500 cm–1. δH (90 MHz, CDCl3): 7.77 (IH, dd, 8, 2 Hz),
7.5–7.2 (6H, m), 7.0–6.7 (2H, m), 5.37 (lH, q, 6 Hz), 3.91
(3H, s), 1.61 (3H, d, 6 Hz). m/z: 256 (5%, M+), 105 (100%,
1-phenylethyl cation) (found: M 256.1095; C16H16O3 re-
quires 256.1099).
give 2 (X = 4-OMe, Y = 5-NO2, 0.46 g, 47%) as an oil. νmax
:
1740 (CϭO), 1620, 1600, 1500, 1520, and 1360 cm–1. δH
(90 MHz, CDCl3): 8.71 (1H, d, 3Hz), 8.20 (lH, dd, 9, 3Hz),
7.37 (2H, d, 7.5 Hz), 6.95 (1H, d, 9 Hz), 6.91 (2H, d,
7.5 Hz), 5.50 (1H, q, 6Hz), 4.00 (3H, s), 3.81 (3H, s), 1.70
(3H, d, 6Hz). m/z: 331 (10%, M + 1), 135 (100%, 1-(4-
methoxyphenyl)ethyl cation) (found: M+ 331.1056;
C17H17NO6 requires 331.1056).
Aryl 1-phenylethyl ethers 4 were prepared by the same
method, by SNAr substitution on the chloro- or fluoronitro-
benzene (e.g., by 1-(4-methylphenyl)ethoxide on 2,4-dini-
trofluorobenzene for 4, X = 4-Me, Y = 4-NO2). Details are
summarized in Table 6.
Ethers of 5-nitrosalicylic acids were made by aromatic
nucleophilic substitution reactions, of the substituted 1-phenyl-
ethanolate anion on the 2-chloro-5-nitrobenzoate (method
B). For example:
Kinetic measurements
Stock solutions (5 mM in dioxan) of aryl acetals were in-
jected into buffer solutions in 10% (v/v) dioxan–water (up to
0.4 M total buffer) made up with KCl to ionic strength 1.0
M. Rates of hydrolysis were measured by monitoring the ap-
pearance of the aryloxide anion over time, at 39.2°C. Reac-
tions were followed in 10% (v/v) dioxan–water, ionic
strength 1.0 M (KCl), for at least three half-lives and end
points taken after at least 10. Ether hydrolysis could in most
1-(4-Methoxyphenyl)ethyl (2′-carboxymethyl-4 -
nitrophenyl) ether 2 (X = OMe, Y = 5-NO2)
Sodium hydride (0.14 g of a 60% slurry in oil) was sus-
pended in dry THF and 1-(4-methoxyphenyl) ethanol
(0.46 g) added. The mixture was refluxed for 30 min, then
© 1999 NRC Canada

Barber et al.
801
Table 6. Summary of synthesis and characterization of ethers 4.
Compound 4
X
Y
Yield (%)
Found, M+
Required; for
mp
4-Me
4-Me
4-Me
4-Me
H
CF3
Ha
67
20
16
60
40a
50
278.09335
257.1062
291.0663
184 (5%), 119 (100%)
184 (51%), 105 (100%)
M+, 365.9870
278.0925; C16H13F3O
257.1052; C15H15NO3
291.0673; C15H14ClNO3
Oil
Oil
Oil
Oil
103–104°C
108–109°C
Cla
b
NO2
NO2
NO2
C6H4NO5; C9H11
b,c
C6H4NO5; C8H9
3-Br
365.9862; C14H11BrN2O5
aThe chloronitrobenzene was used.
1
bNo molecular ion observed. Correct by IR, H NMR, and fragmentation pattern.
cFound: C 58.2, H 4.1, N 9.6; C14H12N2O4 requires C 58.3, H 4.2, N 9.7%.
8. C. Buffet and G. Lamaty. Rec. Trav. Chim. Pays-Bas, 95, 1
(1976).
9. G.-A. Craze and A.J. Kirby. J. Chem. Soc. Perkin Trans. 2, 61
(1974).
10. J.P. Richard and W.P. Jencks. J. Am. Chem. Soc. 1383 (1984).
11. J.P. Richard. In Advances in carbocation chemistry. Edited by
X. Creary. JAI Press, London. 1989. pp. 121–169.
12. J.P. Richard and W.P. Jencks. J. Am. Chem. Soc. 104, 4689
(1982).
13. J.P. Richard, M.E. Rothenberg, and W.P. Jencks. J. Am. Chem.
Soc. 106, 1361 (1984).
14. Y. Tsuno, Y. Kusuyama, M. Sawada, T. Fujii, and Y. Yukawa.
Bull. Chem. Soc. Jpn. 48, 3337 (1975).
15. E.A. Hill, M.L. Gross, M. Stasiewicz, and M. Manion. J. Am.
Chem. Soc. 91, 7381 (1969).
16. D.S. Noyce and J.A. Virgilio. J. Org. Chem. 37, 2643 (1972).
17. H.H. Jaffé. Chem. Rev. 53, 191 (1953).
18. P.D. Bolton, F.M. Hall, and J. Kudrynski. Aust. J. Chem. 21,
1541 (1968).
cases be followed directly by measuring the increase in UV
absorbance as the salicylate anion was released, but for most
5-nitrosalicylic acid derivatives the change in absorbance
was too small, so a simple pH jump technique was used. Re-
actions were initiated by injecting a sample of stock solution
into 2 mL of buffer to give a final substrate concentration of
about 5 × 10–4 mol dm–3: aliquots were taken at intervals
and injected into 2 mL portions of pH 10.5 buffer. This
quenched the reaction and converted all the 5-nitrosalicylate
into the dianion. All pH measurements were made under the
conditions of the kinetic experiments using a Radiometer
PHM82 pH meter fitted with a Russell CTWL electrode.
Activities at pH < 2 were calculated by adding the appropri-
ate activity coefficient correction (e.g., 0.18 at 39°, ionic
strength 1.0 M with KCl) to –log [H+]. The pD of solutions
in D2O was measured by adding 0.4 to the meter reading ob-
tained using the same electrode standardized against H2O
buffers (27).
19. C.F. Bernasconi and G.D. Leonarduzzi. J. Am. Chem. Soc.
104, 5133 (1984).
20. A. Williams. Adv. Phys. Org. Chem. 27, 1 (1992).
21. J.P. Richard and W.P. Jencks. J. Am. Chem. Soc. 106, 1396
(1984).
References
1. A.J. Kirby. Acc. Chem. Res. 30, 290 (1997).
2. A.J. Kirby. Adv. Phys. Org. Chem. 17, 183 (1980).
3. C.J. Brown and A.J. Kirby. J. Chem. Soc. Perkin Trans. 2,
1081 (1997).
22. R. Ta-Shma and W.P. Jencks. J. Am. Chem. Soc. 108, 8040
(1986).
23. J.A. Gerlt and P.G. Gassman. J. Am. Chem. Soc. 115, 11 552
(1993).
4. S.E. Barber and A.J. Kirby. Preliminary communication: J.
Chem. Soc. Chem. Commun. 1775 (1987).
24. W.W. Cleland, P.A. Frey, and J.A. Gerlt. J. Biol. Chem. 273,
25 529 (1998).
5. C.G. Swain and C.B. Scott. J. Am. Chem. Soc. 75, 141 (1953).
6. R.H. Bromilow and A.J. Kirby. J. Chem. Soc. B, 149 (1972).
7. R.H. Bromilow, S.A. Khan, and A.J. Kirby. J. Chem. Soc. B,
911 (1972).
25. S. Scheiner. Acc. Chem. Res. 28, 402 (1995).
26. C.F. Bernasconi. Tetrahedron, 41, 3219 (1985).
27. P.K. Glasoe and D.A. Long. J. Phys. Chem. 64, 188 (1960).
28. G.E. Dunn and F.L. Kung. Can. J. Chem. 44, 1261 (1966).
© 1999 NRC Canada