The SN3-SN2 spectrum. Rate constants and product selectivities for solvolyses of benzenesulfonyl chlorides in aqueous alcohols

Article abstract of DOI:10.1002/poc.1522

Rate constants for a wide range of binary aqueous mixtures and product selectivities (S) in ethanol - Water (EW) and methanol-water (MW) mixtures, are reported at 25 °C for solvolyses of benzenesulfonyl chloride and the 4-chloro - Derivative. S is defined as follows using molar concentrations: S =([ester product]/[acid product]) × ([water solvent]/[alcohol solvent]). Additional selectivity data are reported for solvolyses of 4-Z-substituted sulfonyl chlorides (Z - OMe, Me, H, Cl and NO₂) in 2, 2, 2-trifluoroethanol-water. To explain these results and previously published data on kinetic solvent isotope effects (KSIEs) and on other solvolyses of 4-nitro and 4-methoxybenzenesulfonyl chloride, a mechanistic spectrum involving a change from third order to second order is proposed. The molecularity of these reactions is discussed, along with new term 'S_N3-S_N2 spectrum' and its connection with the better established term 'S _N2-S_N1 spectrum'. Copyright

Full text of DOI:10.1002/poc.1522

Research Article
Received: 1 October 2008,
Revised: 4 December 2008,
Accepted: 15 December 2008,
Published online in Wiley InterScience: 4 February 2009
(www.interscience.wiley.com) DOI 10.1002/poc.1522
The SN3–SN2 spectrum. Rate constants and
product selectivities for solvolyses of
benzenesulfonyl chlorides in aqueous alcohols
a
a
b
b
*
T. William Bentley , Robert O. Jones , Dae Ho Kang and In Sun Koo
Rate constants for a wide range of binary aqueous mixtures and product selectivities (S) in ethanol–water (EW) and
methanol–water (MW) mixtures, are reported at 25 -C for solvolyses of benzenesulfonyl chloride and the
4-chloro-derivative. S is defined as follows using molar concentrations: S ¼ ([ester product]/[acid product]) T ([water
solvent]/[alcohol solvent]). Additional selectivity data are reported for solvolyses of 4-Z-substituted sulfonyl chlorides
(Z ¼OMe, Me, H, Cl and NO2) in 2,2,2-trifluoroethanol–water. To explain these results and previously published data
on kinetic solvent isotope effects (KSIEs) and on other solvolyses of 4-nitro and 4-methoxybenzenesulfonyl chloride, a
mechanistic spectrum involving a change from third order to second order is proposed. The molecularity of these
reactions is discussed, along with new term ‘SN3–SN2 spectrum’ and its connection with the better established term
‘SN2–SN1 spectrum’. Copyright ß 2009 John Wiley & Sons, Ltd.
Keywords: solvolysis; solvent effects; alcohols; product selectivity; mechanistic spectrum; general base catalysis
Equations such as Eqn (3), having one solvent parameter, are
INTRODUCTION
very useful for detecting mechanistic changes which occur as the
solvent composition changes (e.g. solvolyses of benzoyl
chloride[9] and 4-dimethylaminobenzoyl fluoride[10] in alcohol–
water mixtures). Equation (4) provides some of the best evidence
for a spectrum of mechanisms from bimolecular to unimolecular
(SN2 to SN1).[5,11–14]
Evidence for third order solvolytic reactions can be obtained
directly from rate laws for mixtures of aprotic solvents and
very small amounts of alcohol.[15] By combining rate and
selectivity data, rate-product correlations for the whole range of
alcohol–water mixtures have been shown to be consistent with
third order reactions for pseudo-first order solvolyses of
4-nitrobenzenesulfonyl chloride (1, Z ¼ NO2),[16] and of other
acid chlorides.[17–19] Recently, we contrasted the third order
solvolyses of 1, Z ¼ NO2 with the second order solvolyses of
4-nitrobenzyl tosylate (2).[20] The second order solvolyses of 2
correlated linearly with YOTs (in the appropriately modified
version of Eqn (3)), whereas the correlation for 1, Z ¼ NO2 curved
downwards at high values of YOTs. Also, 1, Z ¼ NO2 showed a
much higher kinetic solvent isotope effect (KSIE) than 2.[16,20]
Investigations of the rates and product selectivities for solvolyses
in binary alcohol–water mixtures have provided new insights into
reaction mechanisms and into solvent effects on reactivity. A
solvolysis reaction of a substrate (RX) in an alcohol–water mixture
(R0OH—H2O) gives two substitution products (Eqn (1)) and
product selectivities (S, Eqn (2), in which square brackets refer to
molar concentrations[1]), provide information about the product-
determining steps;[1] e.g. for SN1 reactions, S values can indicate
whether the products are formed from a solvent-separated ion
pair[1–3] or from more dissociated cations.[3]
RX þ R0OH=H2O ! ROR0=ROH þ HCl
S ¼ ð½ROR0ꢀ=½ROHꢀÞ ꢁ ð½H2Oꢀ=½R0OHꢀÞ
(1)
(2)
Solvent effects on rate constants can be interpreted using the
Grunwald–Winstein (GW) equation (Eqn (3))[4] and the various
extended GW equations (e.g. Eqn (4)).[5–8]
logðk=k0Þ ¼ mY þ c
logðk=k0Þ ¼ mY þ lN þ c
(3)
(4)
In Eqns (3) and (4), log k refers to the rate constant in any
solvent relative to 80% ethanol–water (EW) (k0), m is the
sensitivity of the substrate (RX) to the solvent ionizing power (Y)
and c is a small residual term. In Eqn (4), l is the sensitivity of the
substrate to solvent nucleophilicity (N). Solvolyses of adamantyl
substrates have led to the modified GW equation, in which YX
values replace Y to minimise nucleophilic solvation effects and to
allow for differences in solvation effects of leaving groups.[6]
There are also various N scales,[7] and an additional parameter (I)
can be included to allow for solvation effects in conjugated
substrates: e.g. when the developing positive charge can be
delocalised onto aromatic rings.[8]
*
Correspondence to: T. W. Bentley, Chemistry Unit, Grove Building, School of
Medicine, Swansea University, Singleton Park, Swansea SA2 8PP, Wales, UK.
E-mail: t.w.bentley@swansea.ac.uk
a
b
T. W. Bentley, R. O. Jones
Chemistry Unit, Grove Building, School of Medicine, Swansea University,
Singleton Park, Swansea SA2 8PP, Wales, UK
D. H. Kang, I. S. Koo
Department of Chemistry Education and Research Institute of Natural
Science, Gyeongsang National University, Jinju, 660-701, Korea
J. Phys. Org. Chem. 2009, 22 799–806
Copyright ß 2009 John Wiley & Sons, Ltd.

T. W. BENTLEY ET AL.
Table 2. Rate constants (k) for solvolyses of benzenesulfonyl
chloride (1, Z ¼ H) in aqueous binary mixtures at 25 8Ca
k/10ꢂ3 sꢂ1
We now show that published kinetic data for solvolyses of
4-methoxybenzene sulfonyl chloride (1, Z ¼OMe) correlate well
with data for 2, and we report new rate and product data for
solvolyses of (1, Z ¼OMe, Me, H, Cl and NO2). These and other
published results for these solvolyses (particularly KSIEs[21,22]) are
consistent with a spectrum of reactivities from second order for 1,
Z ¼OMe to third order for 1, Z ¼ NO2). The concept of an SN3–SN2
spectrum of mechanisms will be discussed.
Solvent
% v/v
EtOH
MeOH
Me2CO
MeCN
100
90
80
70
60
40
20
Water
0.021b
0.077
0.136
0.103b
0.272
0.478
0.0077c
0.026e
0.063f
0.12e
0.517e
1.67
3.07h
0.012d
0.041d
0.080d
0.143d
0.266
1.34
0.299
0.882
2.73g
3.07h
1.14
2.01
3.08
3.07h
3.07h
RESULTS
a Determined conductimetrically in duplicate; average devi-
ation typically ꢃ3%; additional data at 25 8C for dioxan–water
are as follows: calculated from data in Reference [26], 90%
(0.0033), 80% (0.019), 60% (0.136), 40% (0.675); interpolated
from data in Reference [27], 70% (0.065), 50% (0.31); interp-
olated from data in Reference [23], 30% (1.17), 20% (1.80), 10%
(2.45).
Rate constants for solvolyses of 4-chlorobenzenesulfonyl chloride
(1, Z ¼Cl) are shown in Table 1 and those for benzenesulfonyl
chloride (1, Z ¼ H) are in Table 2. Product selectivities (S, Eqn (2))
for solvolyses of five substrates (1, Z ¼OMe, Me, H, Cl and NO2) in
ethanol-, methanol- and 2,2,2-trifluoroethanol–water are in
Table 3.
b Data from Reference. [24]
c Reference [26] reports 103k ¼ 0.0058.
d Calculated from data in Reference [28].
DISCUSSION
e In satisfactory agreement with Reference. [26]
f Reference. [26]
There is a general consensus that solvolyses of typical sulfonyl
chlorides are SN2 reactions with dominant bond making to the
incoming solvent nucleophile.[29–32] Although the possibility of
some extent of addition to the sulfonyl group cannot be
excluded,[21,33] there is currently little support for this proposal.
g Quadruplicate determination of rate constant.
h Solvent contained 0.2–0.4% acetonitrile; six previously pub-
lished value of k range from 2.97 to 3.10 (Reference [25]).
Table 1. Rate constants (k) for solvolyses of
4-chlorobenzenesulfonyl chloride (1, Z ¼Cl) in aqueous binary
mixtures at 25 8Ca
Also, the possible involvement of one additional solvent molecule
as a general base is often considered,[16,21,22,31,33] and even
higher kinetic orders in solvent have also been suggested.[34]
Highly electron rich sulfonyl chlorides (e.g. 1, Z ¼ NMe2 and
mesitylene sulfonyl chloride, 3) undergo a mechanistic change to
a more polar ‘exploded’ SN2 transition state.[35–37] The main focus
of the following discussion is the mechanistic role of only one
additional solvent molecule for solvolyses of the less electron rich
substrates.
k/10ꢂ3 sꢂ1
Solvent
% v/v
EtOH
MeOH
Me2CO
MeCN
100
90
80
70
60
50
40
30
0.022b
0.094
0.171
0.249
0.357
0.526
0.85
1.41
1.87
2.02
2.00c
0.083
0.255
0.467
0.733
1.05
1.44
1.87
2.15
2.29
0.017
0.051
0.111
0.203
0.370
0.598
0.975
1.38
0.0089
0.0324
0.0633
0.0985
0.160
0.286
0.547
0.99
20
10
Water
Product selectivities (S)
2.01
2.00c
1.65
2.00c
1.55
2.00c
As is often observed,[1] values of S increase in more aqueous
solvents and are greater for EW mixtures than for corresponding
methanol–water (MW) mixtures (Table 3). Also S increases
gradually as the substituent (Z) becomes more electron donating;
the effect is largest for 97% TFE and for the less aqueous EW and
MW mixtures (the results for 1, Z ¼ Me may be slightly low
because of uncertainties in the ester/acid response factor).
In contrast, solvolyses of 4-substituted benzoyl chlorides (4,
Table 4 of Reference[38]) and 4-substituted benzyl chlorides (5,
a Determined conductimetrically in duplicate; average devi-
ation typically ꢃ3%; additional data at 25 8C for dioxan–water
interpolated from data in Reference [23] are as follows: 30%
(1.14), 20% (1.40), 10% (1.66).
b Literature value: 0.0237 (Reference [24]).
c Solvent contained 0.2–0.4% acetonitrile; literature value 2.02
(Reference [25]).
www.interscience.wiley.com/journal/poc
Copyright ß 2009 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2009, 22 799–806

SN3–SN2 SPECTRUM
to a process lacking significant general base catalysis[21,39]), but
reasonable doubts have already been raised[40] about our
interpretation[21] of rate and product data in highly aqueous
media. The new selectivity data (Table 3) support the possibility
that a small decrease in S can occur in the absence of a
mechanistic change, but rate–rate profiles still support our
viewpoint[21] (refer to last paragraph of this discussion).
Table 3. Product selectivities (S, Eqn (2)) for solvolyses of 4-Z
substituted benzenesulfonyl chlorides (1) in aqueous alcohol
mixtures at 25 8C
Substituent (Z)
Alcohol
c
% v/v
OMea
Mea
Hb
Clb
NO2
Rate–rate profiles and the correlation of solvolysis rates with
solvent parameters
Ethanol
90
80
70
60
50
40
30
20
10
5
Methanol
90
80
70
60
50
40
1.6
2.3
1.0
1.5
0.92
1.5
2.0
2.6
3.1
3.4
3.5d
3.5d
2.9e
2.9f
0.61
1.0
1.5
1.8
2.4
2.9
3.3d
3.2d
2.6e
2.8f
0.40
0.68
0.94
1.2
1.5
1.7
1.9
2.0
2.0
2.1
Very recent work[41] on correlations for various sulfonyl chlorides
utilising Eqn (4) gave m values of 0.6–0.7 and l values of 1.1–1.3 for
aromatic sulfonic acid chlorides. When compared with earlier
correlations,[40] lower errors in the quoted parameters were
found by including additional data[42] for fluorinated alcohols; the
importance of a diversity of solvents can be illustrated by
solvolyses of 1, Z ¼OMe and Me, for which the error in l was
halved by the addition of only one data point for 97% TFE to a
data set for over 30 other aqueous mixtures. Even with the improved
data set, it is difficult to make a case for mechanistic changes using
Eqn (4), and the reactions were all classified as SN2.[41]
For solvolysis reactions, some the most important evidence for
mechanistic changes is obtained by comparing solvolysis rates
with the rates of ‘well understood’ reference reactions for
substituent or solvent effects. In the case of solvent effects, it is
sometimes difficult to distinguish between solvation effects and
changes in mechanism,[43] and multi-parameter treatments are
not helpful in resolving these difficulties because they are
relatively insensitive.[44]
3.8
3.9
4.2
3.9
3.6
3.4
3.2
2.6
3.0
2.9
3.0
2.7
2.5
4.2
1.3
2.2
3.2
3.9
4.5
5.0
5.2d
4.8d
4.7e
4.7f
0.86
1.4
2.1
2.9
3.3
3.9
4.1d
3.7d
3.6e
3.7f
0.88
1.2
1.6
1.9
2.3
2.6
2.9
3.2
3.4
3.4
2.2
3.5
4.5
4.4
6.0
30
20
10
5
5. 9
5.2
5.1
4.2
4.5
4.4
Solvolyses of methyl, ethyl and 4-nitrobenzyl tosylate (2) are
suitable ‘well understood’ reference reactions for SN2 solvo-
lyses.[5,13,20] Log (k/k0) for solvolyses of 2 correlate linearly with
log (k/k0) for solvolyses of methyl tosylate with only small
deviations for weakly nucleophilic solvent such as formic acid and
97% TFE (Fig. 2 of Reference[13]); importantly, there are no large
deviations for other aqueous mixtures such as acetonitrile–water.
In contrast, the data point for 90% acetonitrile–water deviates
over one order of magnitude from the correlation line for
aqueous ethanol in a plot versus YOTs (Fig. 1 of Reference[13]). The
deviation can be explained by the relatively low nucleophilicity of
acetonitrile–water mixtures, reflected in the values of NOTs, which
decrease as the percentage acetonitrile increases.[12]
2,2,2-trifluoroethanol–water (% w/w) at 75 8C
97 0.2 0.09 0.07
0.05
0.02
a Data from Reference [21].
b This work determined by at least duplicate analyses of two
independent solutions; substrate injected as a 10% solution in
acetonitrile (10 ml) into 5 ml of thermostatted solvent, except
where stated otherwise; typical errors 2–6%.
c Data from Reference [16].
d Injected 20 ml of a 5% solution of the substrate.
e Injected 30 ml of a 1% solution of the substrate.
f Injected 30 ml of a 0.5% solution of the substrate.
In order to discuss the data for sulfonyl chlorides, we can now
select 2 as a reference system for SN2 reactions, knowing that a
link to methyl tosylate is established.[13] Also, we showed recently
that rate constants for solvolyses of 2 in EW and MW mixtures
could be correlated with the product compositions, assuming
that the reactions were second order.[20] Logarithms of rate
constants for solvolyses of 1, Z ¼OMe correlate linearly
(slope ¼ 1.09 ꢃ 0.04, r ¼ 0.981) with those for 2, in all solvents
including 97% TFE (Fig. 1), showing that their responses to
changes in solvent nucleophilicity are almost identical; using Eqn
(4), l values of 0.94 ꢃ 0.03[13] and 1.07 ꢃ 0.08[41] were reported for
2 and 1, Z ¼OMe, respectively.
Table 1 of Reference[20]) do not show consistent trends in either
substituent or solvent effects. These more complex trends are
explained by larger substituent effects leading to mechanistic
changes. Consequently, the simpler pattern of substituent effects
on S for solvolyses of sulfonyl chlorides (1, Z ¼OMe, Me, H, Cl,
NO2) is in agreement with the accepted[29–32] SN2 mechanism.
In previous work on sulfonyl chlorides,[21,39–41,45] the YCl para-
meter was employed in Eqns (3) and (4) to investigate the
responses to changes in solvent ionising power. Substituent
effects on rates of solvolyses of sulfonyl chlorides are small, even
in weakly nucleophilic media such as TFE,[42] supporting the
judgement made earlier[21] that relatively little positive charge is
delocalised onto the benzene ring during typical SN2 solvolyses
Although the selectivity data for sulfonyl chlorides (Table 3) do
not show any clear evidence for mechanistic changes, it is not
unusual for values of S to decrease slightly (ca. 10%) from 40 to
5% alcohol–water solvents – in Table 3, only solvolyses of 1,
Z ¼ NO2 do not show this effect, and 1, Z ¼OMe shows a larger
effect. The decrease in S may indicate a mechanistic change (e.g.
J. Phys. Org. Chem. 2009, 22 799–806
Copyright ß 2009 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/poc

T. W. BENTLEY ET AL.
Table 4. Statistical data for correlations of logarithms of rate constants for solvolyses of 4-Z substituted benzenesulfonyl chlorides
(1) with those for 4-methoxybenzoyl chloride (4, Z=OMe) in binary aqueous mixtures at 25 8Ca
Substituent
(Cosolvent) m, c, rb
OMe
Me
H
Cl
NO2
(Ethanol)
m
0.468 (0.011)
0.037 (0.025)
0.997
0.424 (0.013)
0.063 (0.029)
0.996
0.410 (0.025)
0.088 (0.054)
0.991
0.352 (0.026)
0.106 (0.058)
0.976
0.266 (0.047)
0.171 (0.105)
0.885
ꢂc
r
nc
(Methanol)
11
11
7
11
11
m
c
r
0.404 (0.028)
0.326 (0.066)
0.979
0.370 (0.030)
0.295 (0.068)
0.974
0.346 (0.042)
0.206 (0.095)
0.964
0.305 (0.043)
0.136 (0.101)
0.921
0.249 (0.050)
ꢂ0.10 (0.12)
0.854
nc
(Acetone)
11
10
7
11
11
m
0.496 (0.013)
0.268 (0.029)
0.997
0.508 (0.014)
0.329 (0.032)
0.997
0.452 (0.018)
0.256 (0.037)
0.996
0.355 (0.019)
0.138 (0.041)
0.989
0.213 (0.029)
0.004 (0.064)
0.932
ꢂc
r
nc
(Acetonitrile)
10
10
7
10
10
m
ꢂc
r
nc
0.567 (0.021)
0.566 (0.048)
0.996
0.472 (0.020)
0.448 (0.045)
0.996
0.467 (0.015)
0.654 (0.034)
0.997
0.301 (0.026)
0.428 (0.058)
0.982
7
6
7
7
a Equation (3) was used, with the following modified Y values derived from log (k/k0) for 4-methoxybenzoyl chloride (4, Z=OMe) from
Reference:[43] EtOH, ꢂ1.34; 90% EtOH, ꢂ0.56; 80% EtOH, 0.0 (by definition); 70% EtOH, 0.48; 60% EtOH, 0.94; 50% EtOH, 1.40; 40%
EtOH, 2.08; 30% EtOH, 2.81; 20% EtOH, 3.20; 10% EtOH, 3.59; water, 3.86; MeOH, ꢂ0.31; 90% MeOH, 0.22; 80% MeOH, 0.71; 70% MeOH,
1.17; 60% MeOH, 1.60; 50% MeOH, 2.04; 40% MeOH, 2.44; 30% MeOH, 2.93; 20% MeOH, 3.29; 10% MeOH, 3.59; 90% acetone, ꢂ2.02;
80% acetone, ꢂ0.99; 70% acetone, ꢂ0.25; 60% acetone, 0.39; 50% acetone, 1.02; 40% acetone, 1.65; 30% acetone, 2.31; 20% acetone,
2.81; 10% acetone, 3.39; 90% acetonitrile, ꢂ1.15; 80% acetonitrile, ꢂ0.22; 60% acetonitrile, 0.76; 40% acetonitrile, 1.78; 30%
acetonitrile, 2.50; 20% acetonitrile, 2.94.
b Slope (m) and intercept (c), with associated ꢃ errors in brackets, and r is the correlation coefficient.
c n ¼ Number of data points in each set of data, including the data for pure water.
Figure 1. Logarithms of relative rate constants log k/ko (log k0 refers to
80% ethanol–water) for solvolyses of 4-methoxybenzene sulfonyl chloride
(1, Z ¼OMe) at 25 8C versus log k/ko for solvolyses of 4-nitrobenzyl
tosylate (2) at 45 8C; data from References [13,20,21]
Figure 2. Logarithms of rate constants for solvolyses of 4-Z substituted
benzene sulfonyl chlorides (1) at 25 8C versus YCl[6] for a range of binary
mixtures between 60% methanol–water and water; kinetic data from
Tables 1 and 2 and References [16,21]
www.interscience.wiley.com/journal/poc
Copyright ß 2009 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2009, 22 799–806

SN3–SN2 SPECTRUM
of benzenesulfonyl chlorides. Consequently, it may not be
necessary to make any allowances for solvation effects adjacent
to the reaction site (e.g. using the I parameter[8] or alternative
similarity models). Nevertheless, an alternative Y scale, based on
solvolyses of 4-methoxybenzoyl chloride (4, Z ¼OMe), was
chosen to investigate substituent effects on m values for the
sulfonyl chlorides (1) – one practical reason is that insufficient YCl
values were available for acetonitrile–water mixtures.
The results (Table 4) show slopes (m) and intercepts (c) for each
substrate in each binary mixture (note: all intercepts are shown as
– c, except for methanol). Good correlations are observed for 1,
Z ¼OMe. As the substituents (Z) become less electron-donating
(changing from OMe to NO2), m decreases significantly, and the
correlations become progressively less precise. Also, more precise
correlations are observed for acetonitrile–water mixtures, the
least nucleophilic solvent mixture.[7,12] Correlations for acetone–
water, EW and MW become progressively less precise. An
alternative plot (Fig. 2) for MW, showing log k (not log k/k0) clearly
shows that curvature is significant for solvolyses of 1, Z ¼ NO2, Cl
and H.
adamantyl derivatives[6]). The remaining difficult questions are
how sensitive or reliable are the tests for nucleophilic solvent
assistance (NSA). One recurring criticism of the tests based on l
values (Eqn (4)) is that nucleophilic effects of solvents are
overestimated because they have not been adequately separated
from electrophilic effects; this complex issue was discussed in
detail recently, and we concluded that the various criticisms
based on multi-parameter correlations were unreliable, partly
because the parameters used for solvent nucleophilicity were too
insensitive.[44] Another test for NSA, the apparent absence of rate
enhancements by ‘strong’ anionic nucleophiles, is insensitive
because the energy required to desolvate a nucleophile is
significant;[58] (e.g. the nucleophilicity of halide ions is inversely
related to the Y value of the solvent – see Figs 6 and 7 of
Reference[59]).
Even when evidence for NSA is established, it is not agreed that
the evidence for molecularity follows automatically, although
we then apply the term ‘SN2 character’ to all cases.[5,57,58]
Others[60,62] try to distinguish between: (i) ‘nucleophilic solvation’
in a transition state leading to a nucleophilically solvated cationic
intermediate and (ii) a conventional SN2 reaction. Another
possibility is to refer to the effects simply as ‘solvation’.[62] Our
viewpoint is now supported by recent DFT calculations,
indicating that NSA does involve the formation of a weak
covalent bond for solvolyses of t-butyl substrates.[63,64] Con-
sequently, it is reasonable to propose that the molecularity of the
reaction has increased from 1 to 2.
Molecularity, SN1, SN2 and SN3
Before discussing the concept of an SN3–SN2 spectrum of mecha-
nisms, some general comments are appropriate. The agreed
definition of an SN2 reaction is ‘two molecules simultaneously
undergoing covalency change’[46] during the rate-determining
step of the reaction, and the reaction is bimolecular.[46] When a
substrate RX reacts with an anionic nucleophile, a second order
rate law provides good evidence for an SN2 reaction.
Ingold’s definition of the molecularity of a reaction stage is ‘the
number of molecules necessarily undergoing covalency change’.
It is not expected that the reactions in solution will be
‘termolecular’, with a meaning that requires simultaneous
collision between three molecules. By analogy with the SN2
transition state as an assembly of two separate molecules, an ‘SN3
reaction’ could occur when two molecules preassociate prior to
reaction with the third (followed by the covalency changes).
Many reactions are thought to proceed in this way, including
hydrolytic reactions of anhydrides,[47] acid chlorides,[15–19]
esters[48–51] and amides.[52] Although the label ‘3’ is relatively
rare, the AdE3 electrophilic addition reaction is the reverse of the
E2 elimination.[53,54]
Research[55,56] on methoxymethyl substrates was important in
establishing the idea that some SN2 reactions occur via relatively
cationic transition states – these may be referred to as ‘SN2
reactions with high carbocation character’[57] or the more graphic
equivalent ‘exploded SN2 transition states’.[56] In other words, it is
also agreed that there can be varying extents of (and even quite
small) rate enhancements due to nucleophilic attack, although
the terminology may vary.
More serious disagreements arise for solvolytic reactions, in
which the only external nucleophile present in the reaction
mixture is the solvent. Most show pseudo-first order kinetics, and
the assignment of molecularity has been controversial for decades.
As a large dynamic pool of solvent molecules contributes to the
observed solvent effect, molecularity is much more difficult to
assign.
Evidence supporting SN3 reactions and a spectrum
of mechanisms
Solvolyses of typical primary substrates occur by SN2 reactions
(e.g. the product ratios for solvolyses of 2 in alcohol–water
mixtures[20]) can be explained by two competing second order
reactions, and a medium effect. Third order solvolyses of
carboxylic acid chlorides are well established directly from the
rate equations for alcoholyses in a large excess of relatively inert
solvents such as ether and acetonitrile, because the alcohol is not
present in large excess.[15] Corresponding reactions of various
acid chlorides (including 1, Z ¼ NO2) in pure methanol, show
large KSIEs[22] – the rate ratio (kMeOH/kMeOD) is 2.3 (Table 5),
whereas SN1 solvolyses and SN2 solvolyses of methyl and (2)
show a KSIEs of <1.3.[20] In MW and EW mixtures, four competing
third order reactions explain the product compositions quanti-
tatively for 1, Z ¼ NO2, consistent with a transition state in which
three molecules simultaneously undergo covalency change (i.e.
an SN3 mechanism – Scheme 1).
Various comparisons between solvolyses of 1, Z ¼OMe and 2
(Table 5) further illustrate the strong similarities (also refer Fig. 1).
In contrast to solvolyses of 1, Z ¼OMe, solvolyses of 1, Z ¼ NO2
show:
(i) higher solvent kinetic isotope effects;
(ii) higher rate ratios for 90% alcohol/pure alcohol despite over-
all lower m values (Table 4) and lower selectivities (Tables 3
and 5);
(iii) a greater response to changes in solvent nucleophilicity
(compare 40% ethanol and 97% trifluoroethanol – k40E
/
k
97T ratios – Table 5 or l values Eqn (4))[40,41]
.
Consider the distinction between uni- and bimolecular
reactions. For a substrate (RX), when nucleophilic assistance to
heterolysis of the RX bond by solvent is negligible, the reaction is
unimolecular and SN1 (a situation approached with solvolyses of
Numerical values of the above mechanistic indicators for
solvolyses of 1, Z ¼ Me, H and Cl are consistent with a spectrum of
J. Phys. Org. Chem. 2009, 22 799–806
Copyright ß 2009 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/poc

T. W. BENTLEY ET AL.
Table 5. Comparisons of various mechanistic indicators for
solvolyses of 4-nitro and 4-methoxybenzenesulfonyl chlorides
(1, Z ¼ NO2 and OMe) and 4-nitrobenzyl tosylate (2) in
alcohol–water mixtures
Substrate
a
Indicator
1, Z ¼ NO2
1, Z ¼OMeb
2c
kMeOH/MeOD
kH2O/D2O
k90M/MeOH
S90M
2.31d
1.76d
3.3
0.88
7.0
0.4
1.5 ꢁ 104
1.58d
1.37d
2.4
2.5
2.7
1.6
3.0 ꢁ 102
1.27
1.17e
2.0f
2.1
2.6
1.1
3.3 ꢁ 102
Figure 3. Plot of the solvent nucleophilicity parameter (NT) versus YCl in
ethanol and methanol–water mixtures; data from References [6,7]; the
two lines are drawn through the data points for methanol–water mixtures
only (filled triangles)
k90E/EtOH
S90E
g
k40E/97T
a Data at 25 8C from Reference [16].
b Data at 25 8C from Reference [21].
c Data at 45 8C from References [13,20].
d Data from Reference [22].
e Refers to mesylate.
f Interpolated.
g Data for 1, Z ¼OMe and NO2 from Reference [42].
by general base catalysis is small or negligible. Examples include
most solvolyses of alkyl substrates, because deprotonation of the
solvent nucleophile is relatively fast.[66] Solvolyses of 1, Z ¼OMe
in water appear to be close to the borderline between SN2
and SN3.
In contrast to the downward curvature for 1, Z ¼ NO2 and Cl,
the plot (Fig. 2) is approximately linear for 1, Z ¼ Me and OMe.
These results are consistent with our proposal, based on
upward curvature previously observed for mesitylene sulfonyl
chloride (3),[21] that there is a mechanistic change e.g. using
the terminology now proposed, from the SN3–SN2 spectrum
for solvolyses of 1, Z ¼ NO2 – OMe to the SN2–SN1 spectrum
for more electron rich substrates (e.g. 3). Whilst we accept
the warning that caution is required,[40] an alternative expla-
nation[40] of the upward curvature for 3 is incorrect as discussed
below.
transition state structures, as shown by the substituent effects on
KSEIs,[22] as well as selectivities (Table 3) and m values (Table 4).
An interdependence of mechanistic indicators is established
for the SN2–SN1 spectrum of mechanisms. As NSA increases (i.e. l
increases), m decreases because the positive charge is more
delocalised (e.g. by gradually changing from a hindered
secondary alkyl to methyl,[5] or by gradually decreasing electron
donation in substituted benzyl substrates[11–14]). An SN3–SN2
spectrum of mechanisms involves more subtle changes in
transition state structure, but similar trends may arise. As the
importance of general base catalysis increases, l increases and m
decreases further (because the positive charge is more
delocalised), continuing the trend from SN1 to SN2. The same
trends may continue further to solvolyses of organo phosphorous
compounds, which show high l values, high KSIEs and relatively
low S values,[65] (but low S values have also been explained by
increased steric effects[65]).
It was proposed[40] that the nonlinearity for 3 could be
explained using Eqn (4) because of the nonlinearity of a plot of NT
(based on solvolyses of S-methyldibenzothiophenium ions[7]
)
versus YCl; for alcohol–water mixtures, there is an increasing
dependence of NT on YCl at higher YCl values (Fig. 3). Plausible l
and m values combined with an l/m ratio >3 (Eqn (4), e.g. with
l ¼ 1.5, m ¼ 0.5) can explain the decrease in rate constant for
solvolyses of 1, Z ¼ NO2 in water (Fig. 2). As l decreases, the
predictions from Eqn (4) more closely approach those of Eqn (3)
(a linear plot vs. YCl), so the increase in slope observed for
solvolyses of 3 cannot be explained using Eqn (4) (assuming that
only one value of l and one of m are proposed). Also, the most
marked changes in slope (Fig. 3) occur between alcohol and 80%
alcohol–water mixtures, whereas the changes in slopes shown in
Fig. 2 occur between 40% MW and water.
At the SN2 end of the SN3–SN2 spectrum of mechanisms, a
bimolecular (SN2) reaction occurs when the extent of assistance
CONCLUSIONS
For solvolyses of 1, Z ¼OMe, an SN2 mechanistic classification is
justified, e.g. on the basis of the correlation shown in Fig. 1, and
previous research on 4-nitrobenzyl tosylate (2) and methyl
tosylate.[13,20] In contrast, solvolyses of 1, Z ¼ NO2 are consistent
with third order kinetics.[16] Other mechanistic criteria such as
selectivity (Table 3), changes in m values (Table 4) and KSIEs
Scheme 1. A simplified picture of some differences between second
(SN2) and third order (SN3) hydrolyses for ArSO2Cl; the SN3 process could
alternatively be referred to as SN2 with general base catalysis by a second
molecule of solvent
www.interscience.wiley.com/journal/poc
Copyright ß 2009 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2009, 22 799–806

SN3–SN2 SPECTRUM
(Table 5) indicate no sudden mechanistic changes for the other
three compounds (1, Z ¼ Me, H and Cl), so there is a mechanistic
spectrum from third order for 1, Z ¼ NO2 to second order for 1,
Z ¼OMe. If the third order reactions are labelled SN3, referring to
an SN2 process with general base catalysis by a second molecule
of solvent (Scheme 1), the spectrum of mechanisms due to
varying extents of general base catalysis can be labelled SN3–SN2.
The consistent pattern of selectivities (Table 3) contrast with
those observed for solvolyses of benzoyl (4)[38] and benzyl
chlorides (5),[20] which show larger substituent effects and
undergo more extensive mechanistic changes over the same
range of substituents from Z ¼ NO2 to OMe. A mechanistic
change may occur for solvolyses of more electron rich sulfonyl
chlorides (1, Z ¼ NMe2 and 3) in highly polar and/or weakly
nucleophilic media.[35–37]
Kinetics and product studies
Kinetic data were obtained as described earlier.[21] Product
studies for rapid reactions at 25 8C were carried out by injections
of ml amounts of substrate dissolved in dry acetonitrile into
turbo-stirred solutions of thermostatted solvent (5 ml). Other
product studies at 25 8C were performed on a 5 ml scale by
suspending 8 ml sample tubes in an ultrasonic bath. Care was
taken to avoid excessive reaction times, because the sulfonate
ester products are further solvolysed to acid in aqueous alcohols;
this side reaction becomes increasingly significant in 90–99%
alcohol–water. Solvolyses in 97% TFE were performed by
injecting 3.5 ml of a 10% solution of sulfonyl chloride in dry
acetonitrile into 97% TFE (1 ml) in ampoules, prior to sealing and
placing in a PEG 400 bath at 75 8C for 10 half-lives.
Chromatography
EXPERIMENTAL
Products from the solvolyses of chlorides, a sulfonic acid and a
sulfonate ester, were analysed using a 5 mm Spherisorb ODS2
chromatography column (15 cm ꢁ 1/400); typical conditions were:
eluent (60% MW), flow rate (1 ml/min), UV detection (l ¼ 259 or
218 nm for 1, Z ¼ H, and 265 or 226 nm for 1, Z ¼Cl). Separations
could also be achieved by the addition of 1% acetic acid or 0.01 M
tetrabutyl ammonium bromide to the MW eluent.
Materials
Sulfonyl chlorides (1) and solvents were obtained and purified as
described earlier.[16,42] Purified sulfonic acids were hygroscopic
and so standard solutions for HPLC calibrations were prepared by
solvolysing the chlorides in acetonitrile–water; ethyl and methyl
esters were prepared similarly, and also by dissolving a twofold
excess of sodium hydroxide in alcohol and adding the sulfonyl
chloride at 0 8C (products were isolated by ether extraction after
about 5 min).[67] The required 4-toluenesulfonate esters were
available commercially, and were recrystallised from chloroform/
petroleum spirit (40–60) at ꢂ20 8C.
Acknowledgements
We are grateful to EPSRC (UK) for equipment grants, to the Korean
Ministry of Education, Science and Technology and Gyeongsang
National University for support of this work, and to D. N. Kevill for
helpful discussions. Additional exploratory research on solvolyses
in 90–99% alcohol–water was carried out by J. Buckley, J. L.
Gooding and P. O’Sullivan.
2,2,2-Trifluoroethyl 4-toluenesulfonate (6, Z ¼ Me, Lancaster)
was recrystallised from ether at low temperature (mp 36–38).
Other esters were prepared by reacting sodium hydroxide
(0.0228 mol) in excess TFE (10 ml) at 0 8C with a solution of
sulfonyl chloride (0.014 mol) in TFE (4 ml) added dropwise. After
about 5 min, the excess TFE was removed under reduced pressure,
and the products were extracted by stirring with dichloromethane
(DCM, 100 ml). The mixture was then filtered and the DCM was
shaken with sodium hydrogen carbonate solution, dried with
magnesium sulphate and the DCM was removed. The ester was
then dried overnight under vacuum in the presence of phosphorus
pentoxide. Two esters (6, Z¼ H, Cl) were obtained as colourless
liquids and two new solid esters were also obtained (as shown
REFERENCES
[1] R. Ta-Shma, Z. Rappoport, Adv. Phys. Org. Chem. 1992, 27, 239–291.
[2] J. M. Harris, D. C. Clark, A. Becker, J. C. Fagan, J. Am. Chem. Soc. 1974,
96, 4478–4484.
[3] T. W. Bentley, D. N. Ebdon, E.-J. Kim, I. S. Koo, J. Org. Chem. 2005, 70,
1647–1653.
[4] E. Grunwald, S. Winstein, J. Am. Chem. Soc. 1948, 70, 846–854.
[5] F. L. Schadt, T. W. Bentley, P. v. R. Schleyer, J. Am. Chem. Soc. 1976, 98,
7667–7674.
1
below); all were shown to be pure by HPLC and H NMR.
[6] T. W. Bentley, G. Llewellyn, Prog. Phys. Org. Chem. 1990, 17, 121–158.
[7] D. N. Kevill, In: Advances in Quantitative Structure-Property Relation-
ships, Vol. 1 (Ed.: M. Charton), Jai Press, Greenwich, CT, 1996, 81–115.
[8] D. N. Kevill, M. J. D’Souza, J. Chem. Res. 2008, 61–66.
[9] T. W. Bentley, G. E. Carter, H. C. Harris, J. Chem. Soc., Perkin Trans. 2
1985, 983–990.
[10] B. D. Song, W. P. Jencks, J. Am. Chem. Soc. 1988, 111, 8470–8479.
[11] D. N. Kevill, H. Ren, J. Org. Chem. 1989, 54, 5654–5655.
[12] M. Fujio, T. Susuki, M. Goto, Y. Tsuji, K. Yatsugi, Y. Saeki, S. H. Kim, Y.
Tsuno, Bull. Chem. Soc. Jpn. 1994, 67, 2233–2243.
[13] M. Fujio, T. Susuki, M. Goto, Y. Tsuji, K. Yatsugi, S. H. Kim, G. A.-W.
Ahmed, Y. Tsuno, Bull. Chem. Soc. Jpn. 1995, 68, 673–682.
[14] D. N. Kevill, M. J. D’Souza, H. Ren, Can. J. Chem. 1998, 76, 751–757.
[15] D. N. Kevill, C.-B. Kim, Bull. Chem. Soc. France 1988, 383–390.
[16] T. W. Bentley, R. O. Jones, I. S. Koo, J. Chem. Soc., Perkin Trans. 2 1994,
753–759.
2,2,2-trifluoroethyl-4-nitrobenzenesulfonate (6, Z ¼ NO2)
Prepared as described above, yield 91%, recrystallised from ether
to give pale yellow platelets, mp 80–81 8C. 1H NMR: d ¼ 4.4 (q), 8.1
(d), 8.45 (d). Anal. Calcd for C8H6NO5SF3: C, 33.67; H, 2.12; N, 4.91.
Found: C, 33.58; H, 2.24; N, 4.85.
2,2,2-trifluoroethyl-4-methoxybenzenesulfonate (6, Z ¼OMe)
Prepared as described above, recrystallised from ether to give
[17] T. W. Bentley, H. C. Harris, Z. H. Ryu, G. T. Lim, D. D. Sung, S. R. Szajda, J.
Org. Chem. 2005, 70, 8963–8970.
[18] T. W. Bentley, R. O. Jones, J. Chem. Soc., Perkin Trans. 2 1993,
2351–2357.
1
long white needles, mp 56–57 8C. H NMR: d ¼ 3.85 (s), 4.3 (q),
7.0(d), 7.8 (d). Anal. Calcd for C9H8O4SF3: C, 39.97; H, 3.36. Found:
C, 40.13; H, 3.44.
J. Phys. Org. Chem. 2009, 22 799–806
Copyright ß 2009 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/poc

T. W. BENTLEY ET AL.
[19] I. S. Koo, K. Yang, K. Kang, I. Lee, T. W. Bentley, J. Chem. Soc., Perkin
Trans. 2 1998, 1179–1183.
[41] M. J. D’Souza, L. Yaakoubd, S. L. Mlynarski, D. N. Kevill, Int. J. Mol. Sci.
2008, 9, 914–925.
[20] T. W. Bentley, I. S. Koo, H. Choi, G. Llewellyn, J. Phys. Org. Chem. 2008,
21, 251–256.
[21] I. S. Koo, T. W. Bentley, G. Llewellyn, K. Yang, J. Chem. Soc., Perkin Trans.
2 1991, 1175–1179.
[22] I. S. Koo, I. Lee, J. Oh, K. Yang, T. W. Bentley, J. Phys. Org. Chem. 1993, 6,
223–227.
[23] R. V. Vizgert, E. K. Savchuk, Zh. Obsh. Khim. 1964, 34, 3396–3402.
[24] E. Tommila, P. Hirsja¨rvi, Acta Chem. Scand. 1951, 5, 659–664.
[25] A. R. Haughton, R. M. Laird, M. J. Spence, J. Chem. Soc., Perkin Trans. 2
1975, 637–643.
[26] F. E. Jenkins, A. N. Hambly, Aust. J. Chem. 1961, 14, 190–204.
[27] R. V. Vizgert, Zh. Obsh. Khim. 1962, 32, 628–633.
[28] W. K. Kim, I. Lee, J. Korean Chem. Soc. 1974, 18, 8–11.
[29] I. M. Gordon, H. Maskill, M.-F. Ruasse, Chem. Soc. Rev. 1989, 18,
123–151.
[42] T. W. Bentley, R. O. Jones, J. Phys. Org. Chem. 2007, 20, 1093–1101.
[43] T. W. Bentley, I. S. Koo, S. J. Norman, J. Org. Chem. 1991, 56,
1604–1609.
[44] T. W. Bentley, M. S. Garley, J. Phys. Org. Chem. 2006, 19, 341–349.
[45] D. N. Kevill, B. C. Park, K. H. Park, M. J. D’Souza, L. Yaakoubd,
L. Mlynarski, J. B. Kyong. Org. Biomol. Chem. 2006, 4, 1580–1586.
[46] C. K. Ingold, Structure and Mechanism in Organic Chemistry, Cornell
University Press, Ithaca, New York, 1953, p. 310.
[47] A. R. Butler, V. Gold, J. Chem. Soc. 1961, 2305–2312.
[48] W. P. Jencks, J. Carriuolo, J. Am. Chem. Soc. 1961, 83, 1743–1750.
[49] E. K. Euranto, N. J. Cleve, Acta Chem. Scand. 1963, 17, 1584–1594.
[50] D. Stefanidis, W. P. Jencks, J. Am. Chem. Soc. 1993, 115, 6045–6050.
[51] M. A. Fernandez, R. H. De Rossi, J. Org. Chem. 1999, 64, 6000–6004.
[52] W. Blokzijl, J. Jager, J. P. F. N. Engberts, M. J. Blandamer, J. Am. Chem.
Soc. 1986, 108, 6411–6413.
[30] C. A. Bunton, J. Phys. Org. Chem. 2005, 18, 115–120.
[31] D. N. Kevill, B.-C. Park, K.-H. Park, M. J. D’Souza, L. Yaakoubd, S. L.
Mlynarski, J. B. Kyong, Org. Biomol. Chem. 2006, 4, 1580–1586.
[53] D. J. Pasto, G. R. Meyer, B. Lepeska, J. Am. Chem. Soc. 1974, 96,
1858–1865.
[54] T. W. Bentley, Ann. Reports, Chem. Soc. 1974, 71B, 111–133.
[55] G. A. Craze, A. J. Kirby, R. Osborne, J. Chem. Soc., Perkin Trans. 2 1978,
357–369.
´
[32] L. Garcia-R´ıo, M. Mendez, M. R. Paleo, F. J. Sardina, J. Phys. Chem. B
2007, 111, 12756–12764.
[33] S. N. Ivanov, B. G. Gnedin, V. V. Kislov, Russian J. Org. Chem. 2004, 40,
733–739.
[34] S. N. Ivanov, A. V. Mikhailov, B. G. Gnedin, A. Y. Lebedukho, V. P.
Korolev, Kinet. Catal. 2005, 46, 29–36.
[35] R. M. Forbes, H. Maskill, J. Chem. Soc., Chem. Commun. 1991, 854–856.
[36] S. Thea, C. Carpanelli, G. Cevasco, Eur. J. Org. Chem. 2001, 151–
154.
[37] T. W. Bentley, J. Org. Chem. 2008, 73, 6251–6257.
[38] T. W. Bentley, I. S. Koo, J. Chem. Soc., Perkin Trans. 2 1989, 1385–
1392.
[56] B. L. Knier, W. P. Jencks, J. Am. Chem. Soc. 1980, 102, 6789–6798.
[57] T. W. Bentley, C. T. Bowen, J. Chem. Soc., Perkin Trans. 2 1978, 557–562.
[58] T. W. Bentley, G. E. Carter, J. Am. Chem. Soc. 1982, 104, 5741–5747.
[59] S. Minegishi, R. Loos, S. Kobayashi, H. Mayr, J. Am. Chem. Soc. 2005,
127, 2641–2649.
[60] D. N. Kevill, M. H. Abduljaber, Croat. Chem. Acta 1992, 65, 539–546.
[61] D. N. Kevill, M. H. Abduljaber, J. Org. Chem. 2000, 65, 2548–2554.
[62] J. P. Richard, M. M. Toteva, T. L. Amyes, Org. Lett. 2001, 3, 2225–2228.
[63] A. G. Mart´ınez, E. T. Vilar, J. O. Barcini, S. dl, M. Cerero, J. Org. Chem.
2005, 70, 10238–10246.
[39] I. S. Koo, E. Kwon, H. Choi, K. Yang, J. K. Park, J. P. Lee, I. Lee, T. W.
Bentley, Bull. Korean Chem. Soc. 2007, 28, 2377–2380.
[40] D. N. Kevill, M. J. D’Souza, Collect. Czech. Chem. Commun. 1999, 64,
1790–1796.
[64] F. Ruff, O. Farkas, J. Phys. Org. Chem. 2008, 21, 53–61.
[65] D. N. Kevill, J. S. Carver, Org. Biomol. Chem. 2004, 2, 2040–2043.
[66] J. M. W. Scott, Can. J. Chem. 2005, 83, 1667–1719.
¨
[67] P. Hirsjarvi, E. Tommila, Acta Chem. Scand. 1951, 5, 1097–1104.
www.interscience.wiley.com/journal/poc
Copyright ß 2009 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2009, 22 799–806