Correlation of the rates of Solvolysis of 2,1-benzoxathiol-3-one-1, 1-dioxide (2-sulfobenzoic acid cyclic anhydride)

Article abstract of DOI:10.3184/174751915X14412072517781

Solvolysis of acyclic mixed carboxylic-sulfonic anhydrides in hydroxylic solvents is known to involve displacement at the carbonyl carbon to produce a carboxylic acid (water as the nucleophile) and/or an ester (alcohol as the nucleophile) plus the anion of the sulfonic acid. Parallel solvolyses of the cyclic mixed anhydride 2-sulfobenzoic anhydride (structurally similar to phthalic anhydride but with one carbonyl group replaced by a sulfonyl group) involve expulsion from the carbonyl carbon of a sulfonate anion that remains attached as an orthosubstituent in the benzoic acid and/or benzoate ester produced. This complicates the choice of a solvent-ionising-power scale for use in an extended Grunwald-Winstein equation treatment. The YOTs scale, previously recommended as a good general purpose scale, is chosen and used in conjunction with the NT solvent nucleophilicity scale. An acceptable correlation is obtained, which is improved when the two solvents rich in the highly ionising 1,1,1,3,3,3-hexafluoro-2-propanol are excluded. As the solvent is varied, the sensitivities to the changes induced in the two scales are low, consistent with an early transition state, but their ratio has a value which is typical for a pathway involving addition-elimination, with addition rate-determining. Earlier reports, supporting aspects of the proposed mechanism, are reviewed.

Full text of DOI:10.3184/174751915X14412072517781

JOURNAL OF CHEMICAL RESEARCH 2015
VOL. 39 OCTOBER, 561–566
RESEARCH PAPER 561
Correlation of the rates of solvolysis of 2,1-benzoxathiol-3-one-1, 1-dioxide
(2-sulfobenzoic acid cyclic anhydride)
Dennis N. Kevilla* and Zoon Ha Ryub
aDepartment of Chemistry and Biochemistry, Northern Illinois University, DeKalb, Illinois 60115-2862, USA
bDepartment of Chemistry, Dong-Eui University, 995 Eomgwango, Busan-Jin-Gu, Busan 614-714, Republic of Korea
Solvolysis of acyclic mixed carboxylic–sulfonic anhydrides in hydroxylic solvents is known to involve displacement at the carbonyl carbon
to produce a carboxylic acid (water as the nucleophile) and/or an ester (alcohol as the nucleophile) plus the anion of the sulfonic acid.
Parallel solvolyses of the cyclic mixed anhydride 2-sulfobenzoic anhydride (structurally similar to phthalic anhydride but with one carbonyl
group replaced by a sulfonyl group) involve expulsion from the carbonyl carbon of a sulfonate anion that remains attached as an ortho-
substituent in the benzoic acid and/or benzoate ester produced. This complicates the choice of a solvent-ionising-power scale for use
in an extended Grunwald–Winstein equation treatment. The Y scale, previously recommended as a good general purpose scale, is
chosen and used in conjunction with the NT solvent nucleophilicOitTys scale. An acceptable correlation is obtained, which is improved when
the two solvents rich in the highly ionising 1,1,1,3,3,3-hexafluoro-2-propanol are excluded. As the solvent is varied, the sensitivities to
the changes induced in the two scales are low, consistent with an early transition state, but their ratio has a value which is typical for a
pathway involving addition–elimination, with addition rate-determining. Earlier reports, supporting aspects of the proposed mechanism,
are reviewed.
Keywords: 2-sulfobenzoic acid cyclic anhydride, Grunwald–Winstein equation, solvolysis, addition–elimination, ring-opening
We have previously reported on the solvolyses of acyclic mixed
carboxylic acid–sulfonic acid anhydrides. These studies, in
a wide variety of solvents, involved the solvolyses of acetyl
p-toluenesulfonate (acetic p-toluenesulfonic anhydride)1
and a similar study of the slower reacting benzoyl and
p-nitrobenzoyl p-toluenesulfonates.2 In addition to linear free
energy relationship (LFER) treatments with use of the extended
Grunwald–Winstein equation,3,4 there were also considerations
of deuterium isotope effects, of leaving-group effects relative to
halide, and of the activation parameters.
We now report on the solvolyses of 2-sulfobenzoic acid cyclic
anhydride [SBA, 2-(or o-) sulfobenzoic anhydride,5,6 systematic
name: 2,1-benzoxathiol-3-one-1,1-dioxide]. Solvolyses were in
a pure or binary hydroxylic solvent, designated as ROH, where
R is hydrogen, alkyl, or fluoroalkyl. The overall reaction is
shown in Scheme 1.
to phthalic acid with one of the carbonyl groups replaced by
a sulfonyl group. In the laboratory, for example, it has been
used in the synthesis of a catalyst for enantioselective Michael
additions7 and as a capping agent in solid phase peptide
synthesis.8
In industrial applications, SBA has been proposed as an
additive to lithium ion batteries to control swelling,9-11 as a
reagent for the preparation of proton-conducting fuel cell
membranes,12,13 as a reagent for the preparation of sulfonate-
bearing polysiloxanes for household and health-care
formulations,14 for use in the synthesis of stabilisers in emulsion
polymerisation15 and for use in the synthesis of well-defined
diblock copolymers.16
In the previously studied 1,2,17 acyclic carboxylic acid–sulfonic
acid anhydrides (Scheme 2) solvolysis leads to two products,
whereas (Scheme 1) only one results from solvolysis, with ring-
opening, of a cyclic anhydride.
The SBA is readily available from several sources and it has
a wide variety of important applications both in the laboratory
and in industry. It can be considered as being formally related
Earlier studies have looked at the rates of solvolysis of both
cyclic5,6,17 and acyclic1,2,17 mixed sulfonic–carboxylic anhydrides.
O
COOR
2 ROH
O
+
+ ROH2
SO3—
S
O
O
Scheme 1
2 ROH
+
R'
C
O
O
SO2R''
R'
C
O
OR + R''SO3 + ROH2
Scheme 2
* Correspondent. E-mail: dkevill@niu.edu

562 JOURNAL OF CHEMICAL RESEARCH 2015
Laird and Spence5,6,17 studied a wide variety of substrates in
only a few solvents, primarily alcohols containing 6.7% ether
or aqueous dioxane with, by volume, 10–20% H2O (or D2O)
and some of the determinations of the specific rates were made
at several temperatures, allowing activation parameters to be
calculated.
In contrast, our more recent studies concentrated on three
acyclic substrates. These were studied in a wide variety of
solvents, with consideration of the influence upon the specific
rates of the solvent nucleophilicity and the solvent ionising
power. Parallel applications of the extended Grunwald–Winstein
equation,3,4 are in this communication used to correlate the
specific rates of solvolysis of SBA.
needed for the application of the extended Grunwald–Winstein
equation to the experimental specific rates. The choice of a YX
scale is a difficult one for ring-openings of cyclic anhydrides.
In the present instance, the sulfonate anion produced is leaving
from the α-carbon but it is not leaving from the product formed
(Scheme 1). Since different arenesulfonate leaving groups
have for most solvents, similar YX scale values,25 the use of
YOTs values is a reasonable compromise. Indeed, Bentley and
Llewellyn recommend the use of YOTs as a general-purpose
scale of solvent ionising power for cases where the appropriate
YX scale is not available.25
For solvolyses in all of the types of solvent used in the present
study, it was earlier concluded that benzoyl p-toluenesulfonate
solvolysed by an SN1 pathway with only minor assistance from
nucleophilic solvation,2 acetyl p- toluenesulfonate solvolysed by
an SN1 pathway with an increased assistance from nucleophilic
solvation (or by a loose SN2 process, which would be expected
to show similar characteristics in an extended Grunwald–
Winstein correlation),1 and p-nitrobenzoyl p-toluenesulfonate
showed very different characteristics in the correlation, such
that a bimolecular process was indicated2 in all solvents except
those rich in the highly ionising 2,2,2-trifluoroethanol (TFE)
or 1,1,1,3,3,3-hexafluoro-2 propanol (HFIP). The results from
these earlier studies of acyclic mixed carboxylic–sulfonic
acid anhydrides, together with those for solvolyses of several
carboxylic acid halides26,27 and a chloroformate ester,28 are
compared, in Table 2, to those for the presently reported mixed
carboxylic–sulfonic acid anhydride, SBA. Since the magnitudes
of the paired NT and YCI, values tabulated for use in Eqn (1)
are not normalised, the magnitudes of the sensitivity values (l
and m) do not directly relate to the relative contributions to the
LFER from the terms governed by the nucleophilicity and by
the ionising power of the solvents [Eqn (2)]. The ratio of the
values, l/m, has, however, been found to be a useful indicator of
the relative importance of the two terms for a series of substrates
undergoing solvolysis.29,30 The l/m ratios are, therefore, also
tabulated in Table 2.
The data from the analysis for solvolyses of SBA reported in
Table 2 show an improved correlation when the two solvents
rich in HFIP are excluded. This situation usually results
from an ionisation mechanism, which is making a negligible
contribution towards rate-determining nucleophilic attack
mechanisms in the other solvents, becoming a significant
contributor in the low nucleophilicity-high ionising power
solvents rich in HFIP. The correlation in the 19 studied solvents
is shown in Fig. 1, where the deviation of the two HFIP-
containing solvents from the correlation line can be noted.
Figure 2 shows a simple Grunwald–Winstein plot against YOTs
values [use of Eqn (2) without the l NT term]. It can be seen that,
due to multicollinearity, reasonable correlations are obtained in
a given binary solvent but the overall correlation is very poor.
The solvolyses of phenyl chloroformate27 can be taken as a
standard for an addition–elimination (A–E) reaction with the
addition process rate-determining (Scheme 3). A variant would
involve attack by a pre-associated ROH dimer and the collapse
of the first two forward steps into one.
The extended Grunwald–Winstein equation18-20 can be
expressed as in Eqn (1), where a substrate RX is undergoing
solvolysis and k and k0 are the specific rates of solvolysis in the
solvent under investigation and in a standard solvent (chosen21
as 80% ethanol). The l value represent the sensitivity to change
in solvent nucleophilicity (NT),22,23 the m value represents the
sensitivity to changes in solvent ionising power (YX for a leaving
group X)24,25 and c is a constant (residual) term.
log (k/k0)RX = l NT + mYX + c
(1)
Results and discussion
Table 1 reports the first-order rate coefficients for the solvolysis
o
of SBA in a variety of hydroxylic solvents at –10.0 C. Also
presented within Table 1 are those NT and YOTs values that are
Table 1 Specific rates of solvolysis (k) of 2-sulfobenzoic cyclic anhydride
(SBA)a at –10.0°C and the appropriate solvent nucleophilicity (NT) and
solvent ionising power (YOTs) values
Solventb
104k/s-1
NT
YOTs
c
d
100% EtOH
90% EtOH
80% EtOH
100% MeOH
95% Acetone
90% Acetone
100% TFE
97% TFE
90% TFE
80% TFE
70% TFE
97% HFIP
90% HFIP
90T–10E
80T–20E
60T–40E
80.1 0.7e
343 16
652 18
187 2e
23.2 0.4
108 1
0.81 0.02
3.84 0.15
13.6 0.2
35.0 0.4
88.1 1.0
3.02 0.04
16.7 0.7
17.5 0.2
28.1 0.6
51.4 0.8
64.5 0.4
73.5 0.4
94.1 0.6
0.37
0.16
0.00
–1.95
–0.77
0.00
–0.92
–2.95
–1.99
1.77
1.83
1.90
1.94
2.00
3.61
2.90
1.32
0.98
0.21
0.14
0.17
–0.49
–0.35
–3.93
–3.30
–2.55
–2.19
–1.98
–5.26
–3.84
–2.62
–1.76
–0.94
–0.64
–0.34
0.08
50T–50E
40T–60E
20T–80E
–0.44
–1.18
a Determined conductimetrically after injection of 20 µL of a 0.10 mol dm-3 solution of the
substrate in dry acetonitrile into 2.0 mL of the indicated solvent (concentration of SBA
of ca 1.0×10-3 mol dm-3). Specific rates, averaged from duplicate runs, reported together
with standard deviations.
For solvolyses of phenyl chloroformate,28 values are obtained
of 1.66 for l and 0.56 for m, with the l/m ratio at a value of 2.96.
Similar values are obtained for the solvolyses of p-nitrobenzoyl
chloride of 1.78, 0.54, and 3.30, respectively.26 For the solvolyses
of p-nitrobenzoyl tosylate, with the p-nitro group disfavouring the
ionisation with loss of the tosylate anion, values are obtained of
1.19, 0.66 and 1.80, respectively, believed to indicate essentially
the same mechanism but with somewhat less involvement of the
nucleophile at the transition state.2 For the chlorides, replacement
b Binary solvents on a volume-volume basis at 25.0°C, except for TFE–H2O and HFIP–H2O
mixtures, which are on a weight-weight basis. When not specified the second component
is H2O. T–E represents 2,2,2-trifluoroethanol–ethanol mixtures.
c Values from refs 22 and 23.
d Values from ref. 25.
e For solvents also containing 6.7% diethyl ether, values of 54.9×10-4 s-1 and of 191×10-4 s-1
can be calculated from the Arrhenius parameters given in Table 3 of ref. 6 for solvolyses at
–10.0°C in ethanol and methanol, respectively.

JOURNAL OF CHEMICAL RESEARCH 2015 563
Table 2 Correlations using the extended Grunwald–Winstein equation of the specific rates of solvolysis of SBA and a comparison with literature values
for other solvolyses at acyl carbon
Substrate (T°C)
SBA (–10.0)
na
lb
mb
cb
l/m
Rc
Fd
46
58
19e
17f
36g
21h
13i
0.65 0.09
0.74 0.08
0.11 0.04
1.19 0.05
0.56 0.10
0.31 0.05
0.47 0.03
1.78 0.08
1.58 0.09
1.66 0.05
0.30 0.08
0.31 0.07
0.80 0.04
0.66 0.06
0.61 0.05
0.81 0.02
0.79 0.02
0.54 0.04
0.82 0.05
0.56 0.03
–0.38 0.12
–0.32 0.11
0.11 0.06
–0.13 0.06
–0.29 0.08
–0.08 0.21
–0.49 0.17
0.11 0.37
–0.09 0.10
0.15 0.07
2.17
2.39
0.14
1.80
0.92
0.38
0.59
3.30
1.93
2.96
0.922
0.945
0.978
0.983
0.966
0.989
0.990
0.969
0.953
0.980
C6H5CO2Ts (–10.0)
p-NO2C6H4CO2Ts (–10.0)
CH3CO2Ts (–39.6)
p-MeOC6H4COCl (25.0)
C6H5COCl (25.0)
p-NO2C6H4COCl (25.0)
C6H5COF (25.0)
C6H5OCOCl (25.0)
364
265
70
738
680
237
186
568
37j
32k
34l
41m
49n
a Number of solvents (data points).
b From Eqn (1), with associated standard errors, using NT and YOTs for first four entries and NT and YCl for last five entries.
c Correlation coefficient.
d F-test value.
e Using all available data points.
f Omitting the data points for the two HFIP-containing solvents.
g From ref. 2, using all solvents.
h From ref. 2, omitting acetone-H2O, TFE-ethanol, 100% TFE and 97% HFIP data points.
i From ref. 1, using all solvents.
j From ref. 26, using all solvents.
k From ref. 26, using the solvolyses believed to follow an ionisation pathway (from a total of 47 measurements).
l From ref. 26, omitting the 97% HFIP data point.
m From ref. 27, using all solvents.
n From ref. 28, using all solvents.
Fig. 1 Plot of log(k/k0) for solvolyses of 2-sulfobenzoic cyclic anhydride
(SBA) in 19 solvents against (0.65 NT + 0.30 YOTs).
Fig. 2 Plot of log(k/k0) for solvolyses of 2-sulfobenzoic cyclic anhydride
(SBA) in 19 solvents against YOTs values.
of the strongly electron-withdrawing p-nitro group by the
strongly electron-supplying p-methoxy group leads to a major
change in solvolysis mechanism as indicated by corresponding
values of 0.31, 0.81 and 0.36, consistent with extensive ionisation
character and only a rather weak nucleophilic assistance within
the rate-determining step (Scheme 4).26
An alternative picture of the process in Scheme 4 would
involve a very loose concerted SN2 process, with at the
transition state rather weak bonding from the carbon at the
reaction centre to both the incoming and outgoing nucleophilic
species (Scheme 5). Such a process is proposed by Bentley for
the solvolyses of acetyl chloride.1,31

564 JOURNAL OF CHEMICAL RESEARCH 2015
R
R
H
O
O
+
ROH
ROH
fast
PhO
C
O
Cl
PhO
C
O
Cl
PhO
C
O
Cl
Ph
O
C
OR
—
—
O
—
+
+
+ ROH2
+ ROH2 + Cl
(R = alkyl, fluoroalkyl or H)
Scheme 3
Slow
ROH
ROH
+
Ar
C
O
Cl
Ar
C
O
Ar
C
O
Ar
C
O
OR
+ Cl—
+
O
+
—
H
R
+ Cl + ROH2
+ Cl—
(Ar = p-CH3OC6H4)
Scheme 4 Ionisation (assisted by solvation of the developing cation).
‡
Ar
R
R
—
δ+ δ+
δ
Cl
ROH
ROH
+
O
C
O
Ar
C
O
Cl
Ar
RO
C
C
O
Ar
[ ]
H
H
O
O
+ Cl —
—
+
+ Cl + ROH2
Scheme 5
Although we willsubsequently refer to an ionisation reaction
when the correlations show a significant l value but a low
l/m ratio, it is possible that the detailed pathway could be as
in Scheme 5, with the positive charge at the transition state
dispersed.
An extended Grunwald–Winstein treatment of the solvolyses
of SBA leads, for all 19 solvents, to values for l of 0.65 and for m
of 0.30, with a rather low correlation coefficient of 0.922 (Table
2). Inspection of the plot (Fig. 1) of log k/k0 against 0.65 NT +
0.30YCl [using Eqn (1)] shows that the two points for HFIP–
H2O mixtures lie above the plot. Omission of these two points,
leads to a moderate improvement of the correlation with an l
value of 0.74 and m value of 0.31, with a correlation coefficient
of 0.945. As can be seen by a comparison with the data for the
solvolysed related compounds given in Table 2, this correlation,
although marginally acceptable, is significantly inferior to the
other correlations within the table. This is consistent with the
observation earlier that the ring-opening leads to the breaking of
one bond to the sulfonyl group but the formed sulfonate ion is
still bonded as an ortho-substituent to the formed acid or ester
group (Scheme 1).
The values for both l and m are lower than for related solvolyses
believed to follow the addition–elimination pathway, with
addition rate-determining, but the l/m ratio of 2.39 is consistent
with such a pathway. One way of rationalising the low values,
coupled with an l/m ratio compatible with an A–E pathway, is
to assume that a much earlier transition state is present along the
reaction pathway for the solvolyses of the cyclic SBA.
The solvolyses of benzoyl chloride are delicately balanced26 in
the solvents used for the investigation with concurrent operation
of both ionisation and A–E mechanisms. For the solvent region
favouring ionisation, values are obtained of 0.47 for l, 0.79
for m, and 0.59 for the l/m ratio. The corresponding tosylate,2
with a leaving group stabilised by resonance, solvolysed by an
ionisation mechanism for the full range of solvents with values
of 0.11 for l and 0.80 for m, such that the l/m ratio is 0.14. Indeed,
an excellent correlation for these solvolyses was obtained2 when
only solvent ionising power was considered.
To give a measure of the specific solvolysis rates relative
to earlier studied mixed carboxylic–sulfonic acid anhydrides,
we can look at specific rates measured (or estimated using the
Arrhenius equation) at –10.0 °C in 95% acetone. Relative values
are: SBA (1.0), acetyl tosylate (99),1 benzoyl tosylate (0.0032)2
and p-nitrobenzoyl tosylate (0.20).2 The solvolysis of SBA is
appreciably slower than for the open-chain aliphatic substrate
and faster than for the open-chain aromatic substrates, which
are appreciably stabilised by the interaction of the aromatic ring
with the carbonyl group.
In an earlier study, a consideration of the effect of a
substituent in the 4- or 5- position of the aromatic ring of SBA

JOURNAL OF CHEMICAL RESEARCH 2015 565
was considered.6 A Hammett equation32 treatment was carried
out with the substituent (Cl, Br, I, NO2) in the 4-position (para
to the carboxyl group and meta- to the sulfonyl group). For
a chlorine substituent, a study was also carried out with the
substituent in the 5-position. Remarkably, at –25°C, the specific
rates for the two isomeric substrates (4-chloro- and 5-chloro-
derivatives of SBA) were essentially identical to each other for
the solvolyses in methanol, ethanol, 1-propanol, and 2-propanol.
Unfortunately, the influence of the other substituents was only
studied when in the 4-position and it would be dangerous to
generalise from this one observation.
A common situation, in applications of the Hammett equation,
is to have two (or more) aromatic ring substituents and such
situations can be considered as additive as regards the σ values
of the substituents.32 A much less common situation is to have
one substituent but two reaction sites, as in the study here under
consideration. Laird and Spence6 attempted to analyse this
situation using an Hammett-equation32 approach. The extended
form used involved two terms.33, 34 These govern the influence
of the substituent for each of the two directions leading into the
anhydride ring, which is ortho-fused to the benzene ring [Eqn
(2)].
spontaneous (no acid catalysis) solvolyses of acyclic44 and
cyclic45,46 carboxylic anhydrides in water at 25 °C.
One possible explanation for the higher values in water than in
90% dioxane–10% H2O could be that dioxane is to some degree
replacing H2O as a general-base assisting the hydrolysis. Both
dioxane and acetone have been shown to be able to function
as nucleophiles in their mixtures with water47 and, presumably,
they could also function as bases. However, the observation
that runs in 60% dioxane–40% water give48 essentially the
same solvent deuterium isotope effect for hydrolyses of acetic
anhydride and glutaric anhydride as their values in 100%
H2O would argue against this suggestion. A more probable
explanation could be that the more reactive mixed carboxylic
sulfonic anhydride, SBA, is less selective as regards whether
it reacts with H2O or D2O than the less reactive acylic or cyclic
carboxylic acid anhydrides.
In the earlier study5 of the solvolyses of SBA, the effect on the
specific rate of temperature variation was reported in terms of
activation energies and frequency factors for dioxane–water (or
D2O) and five alcohols (containing 6.7% ether). The frequency
factors correspond to entropy of activation (ΔS‡) values from
–10 cal mol-1 K-1 (in 2-propanol) to –29 cal mol-1 K-1 (in 90%
dioxane–10% H2O). These values are similar to the values of
–23 cal mol-1 K-1 in ethanol and –16 cal mol-1 K-1 in methanol
for the solvolyses of p-nitrobenzoyl tosylate, solvolyses believed
to be following an addition–elimination pathway, with the
addition rate-determining.2 The values are much more negative
than the values of -3 cal mol-1 K-1 in ethanol and -4 cal mol-
log(k/k0) = ρ1σ1 + ρ2σ2 + c
(2)
It is the ρσ terms that are additive and they represent the
sensitivities of the specific rates towards the changes at the two
“reaction sites” of the substrate: in this instance, to the changes
taking place at the carbonyl carbon and at the developing
sulfonate anion during the heterolytic ring-opening process.
It was found6 that the two ρ values obtained from the
multiregression analysis approach were considerably
lower than the values for addition to the carbonyl group of
substituted benzaldehydes35 and for the ethanolysis of ethyl
arenesulfonates.36 It was concluded that an early transition state
was present with bond making running ahead of bond breaking.
This conclusion is consistent with the A–E mechanism presently
suggested by the extended-Grunwald–Winstein treatment.
It should, however, be mentioned that six data points are
considerably less than the number recommended for a two-term
regression analysis.37 Also, this seems to be a type of situation
where second-order effects, involving a third term of the ρ σ1σ2
form might make a noticeable contribution. In contrast, 1w2 hen
a series of substituents are in two aromatic rings, one in the
attacking nucleophile and one in the substrate, then only a very
small contribution from a cross-interaction term of this type is
observed.38-40
1
K-1 in methanol for benzoyl tosylate solvolyses, believed to
follow an ionisation pathway.2 Johnson49 has proposed that large
negative entropies of activation coupled with appreciable solvent
deuterium isotope effects can be considered as giving excellent
support to claims that general-base catalysis is operating.
The solvolysis of the cyclic carboxylic anhydride, succinic
anhydride, in 60% dioxane48 also involves an appreciably
negative entropy of activation of –31 cal mol-1 K-1 and the
appreciably slower reaction follows mainly from an enthalpy of
activation of 14 kcal mol-1, somewhat higher than the 11.5 kcal
mol-1 for solvolysis of SBA in 90% dioxane.5 This difference
is consistent with a favourable lowering of the electron density
at the carbonyl carbon when it is attached to a sulfonate group
rather than when attached to a carboxylate group: the Hammett
σm values are 0.31 for COMe and 0.65 for SO2Me.33
Conclusions
Application of the extended Grunwald–Winstein equation to
the solvolyses of the mixed cyclic anhydride 2-sulfobenzoic
anhydride (SBA; systematic name: 2,1,benzoxathiol-3-one-
1,1-dioxide) leads to sensitivities towards changes in solvent
nucleophilicity (l value) and towards change in solvent ionising
power (m value) that are lower than those for previously studied
acyclic mixed anhydrides of carboxylic and sulfonic acids. The
l/m ratio was, however, within the range previously observed
for reactions believed to follow an addition–elimination
(association–disassociation) pathway, with the addition rate-
determining (Table 2).
An earlier extended Hammett treatment of the parent and
4-substituted (para to the carbonyl group and meta to the
sulfonyl group) derivatives gave what were considered to be
low ρ values for the influences at these two reaction centres.6
An early transition state was proposed in which bond making
had proceeded further than bond breaking, consistent with the
conclusion reached from the Grunwald–Winstein treatment.
A previous study5 of the solvent deuterium isotope effect
for hydrolysis (kH2O/kD2O) of SBA leads to a value in 90%
A study of solvent deuterium isotope effects5 in the solvolyses
of SBA in 90% dioxane–10% H2O (or D2O) showed no trend
in the kH O/kD2O value at three temperatures in the range of 9.7–
17.2 °C, 2and an average value of 1.45 0.05 was observed. This
value is slightly higher than for the corresponding solvolyses
of benzenesulfonic anhydride (1.20) and methanesulfonic
anhydride (1.22) at 23.0 °C. For the cyclic 3-sulfopropionic
anhydride at 10.2 °C, a kH O/kD ratio of 1.27 was observed.17
O
2
These ratios indicate simil2ar solvent deuterium isotope effects
for sulfonic anhydrides and cyclic mixed sulfonic–carboxylic
anhydrides, with slightly higher values for the cyclic mixed
anhydrides.
Due to slower reactions, studies with carboxylic anhydrides
have been made with more aqueous dioxane–H2O mixtures.
These reactions are subject to acid catalysis and many of the
numerous studies have been of this catalysis.41-43 The deuterium
isotope effect values are appreciably larger than those for SBA,
with values in the range of 2.7–3.9 having been recorded for

566 JOURNAL OF CHEMICAL RESEARCH 2015
8
9
J. W. Drijfhout and W. Bloemhoff, Peptide Chemistry, 1988, Volume Date
1987, 191 (Chem. Abstr., 1988, 109, 211458).
Y. G. Kim, S. K. Kim, M. S. Kim and Y. B. Roh, PCT Int. Appl., 2003, WO
2003021707 A1 20030313.
dioxane–10% H2O (or D2O) at 9.7 to 17.2°C of 1.45, which is
appreciably lower than the range of 2.7 to 3.944-46 observed for
the solvolyses of cyclic carboxylic anhydrides in more aqueous
solvents. While alternative explanations are possible, this is
consistent with the higher reactivity of the mixed anhydrides
leading to a lower selectivity as regards reaction with H2O or
D2O.
The bimolecular (or higher) nature of the solvolyses of
SBA is supported by studies at several temperatures leading
to frequency factors5 which convert to entropies of activation
in the –10 to –29 cal mol-1 K-1 range. These values are similar
into those obtained for p-nitrobenzoyl tosylate solvolysis1 and
considerably more negative than those for benzoyl tosylate
solvolysis,1 consistent with an addition–elimination pathway,
rather than an ionisation pathway, for the solvolyses of SBA.
10 N. I. Kim, U. S. Pat. Appl. Publ., 2005, US 20050233207 A1 20051020.
11 N. -R. Park, J. -B. Kim, J. -S. Kim and J. H. Lim, US Pat. Appl. Publ., 2009, US
20090155697 A1 20090618.
12 E. P. Jutemar and P. Jannasch, J. Membr. Sci., 2010, 351, 87.
13 L. E. Karlsson and P. Jannasch, Electrochim. Acta., 2005, 50, 1939.
14 J. J. Kennan and M. K. -J. Lee, PCT Int. Appl., 2006, WO 2006065467 A2
20060622.
15 J. Zhang, M. R. Dubay, C. J. Houtman and S. J. Severtson, Macromolecules,
2009, 42, 5080.
16 C. -D. Vo, P. D. Iddon and S. P. Ames, Polymer, 2007, 48, 1193.
17 R. M. Laird and M. J. Spence, J. Chem Soc. (B), 1971, 1434.
18 S. Winstein, E. Grunwald and H. W. Jones, J. Am. Chem. Soc., 1951, 73, 2700.
19 T. W. Bentley, F. L. Schadt and P. v. R. Schleyer, J. Am. Chem. Soc., 1972, 94,
992.
20 F. L. Schadt, T. W. Bentley and P. v. R. Schleyer, J. Am. Chem. Soc., 1976,
98, 7667.
Experimental
21 E. Grunwald and S. Winstein, J. Am. Chem. Soc., 1948, 70, 846.
22 D. N. Kevill and S. W. Anderson, J. Org. Chem., 1991, 56, 1845.
23 D. N. Kevill in Advances in quantitative structure-property relationships, Vol.
1, ed. M. Charton, JAI Press, Greenwich, CT, 1996, pp. 81-115.
24 T. W. Bentley and G. E. Carter, J. Am. Chem. Soc., 1982, 104, 5741.
25 T. W. Bentley and G. Llewellyn, Prog. Phys. Org. Chem., 1990, 17, 121.
26 D. N. Kevill and M. J. D’Souza, J. Phys. Org. Chem., 2002, 15, 881.
27 D. N. Kevill and M. J. D’Souza, J. Org. Chem., 2004, 69, 7044.
28 D. N. Kevill, F. Koyoshi and M. J. D’Souza, Int. J. Mol. Sci., 2007, 8, 346.
29 D. N. Kevill and M. J. D’Souza, Recent Res. Devel. Org. Chem., 2013, 13, 1-38.
30 M. J. D’Souza, B. Sandosky, G. A. Fernandez-Bueno, M. J. McAneny and D.
N. Kevill, Can. Chem. Trans., 2014, 2, 160.
The 2-sulfobenzoic acid cyclic anhydride (SBA, TCI >95%) was used
as-received. For determination of the specific rates of solvolysis, a 0.1
M solution of SBA in dry acetonitrile was prepared and 20 µL at room
temperature was added to 2 mL of the appropriate solvent contained
within a conductivity cell at –10.0 °C to give a 1×10-3 mol L-1 solution
of the reactant.
The progress of the reaction was monitored by following the
increases in conductivity as the mixed diacid (reaction with water) and/
or carboxylic acid ester derivative of the diacid (reaction with alcohol)
were produced (Scheme 1), as previously described.1 Details of the
conductivity apparatus and the computer procedure for calculation
of the specific rates have previously been reported.50,51 The multiple
regression analyses were performed using commercially available
statistical packages.
31 T. W. Bentley, G. Llewellyn and J. A. McAlister, J. Org. Chem., 1996, 61, 7927.
32 L. P. Hammett, Physical organic chemistry, 2nd edn. McGraw-Hill, New
York, NY, 1970, pp. 355-376.
33 H. H. Jaffé, Chem. Rev., 1953, 53, 227.
34 E. Andersen, E. A. Birkhimer and T. A. Bak, Acta Chem. Scand., 1960, 14,
1899.
35 J. W. Baker and H. B. Hopkins, J. Chem. Soc., 1949, 1089.
36 R. E. Robertson, Can. J. Chem., 1953, 31, 589.
37 M. H. Abraham, P. L. Grellier, J. -L. M. Abboud, R. M. Doherty and R. W. Taft,
Can. J. Chem., 1988, 66, 2673.
38 I. Lee, Chem. Soc. Rev., 1990, 19, 17.
39 I. Lee, Adv. Phys. Org. Chem., 1992, 27, 57.
40 D. N. Kevill and M. J. D’Souza, J. Chem. Soc. Perkin Trans. 2, 1994, 945.
41 V. Gold and J. Hilton, J. Chem. Soc., 1955, 843.
This research was supported by the donors of the American
Chemical Society Petroleum Research Fund (PRF#38163-
AC4). Z. H. R. thanks Dong-Eui University for support of
this work during the 2005 research year. We thank Professor
Malcolm J. D’Souza (Wesley College, Dover, Delaware,
USA) for helpful discussions and assistance with the multiple
correlation analyses.
42 J. Koskikallio, D. Pouli and E. Whalley, Can. J. Chem., 1959, 37, 1360.
43 C. A. Bunton, J. H. Fendler, N. A. Fuller, S. Perry and J. Rocek, J. Chem. Soc.,
1965, 6174.
44 A. R. Butler and V. Gold, J. Chem. Soc., 1961, 2305.
45 A. R. Butler and V. Gold, J. Chem. Soc., 1962, 2212.
46 K. R. Davis and J. L. Hogg, J. Org. Chem., 1983, 48, 1041.
47 A. Kivinen, The chemistry of the acyl halides, part of the series The chemistry
of the functional groups, ed. S. Patai. Wiley-Interscience, New York, 1972,
p. 185.
48 C. A. Bunton, N. A. Fuller, S. G. Perry and V. J. Shiner, J. Chem. Soc., 1963,
2918.
49 S. L. Johnson, Adv. Phys. Org. Chem., 1967, 5, 237.
50 I. Lee, H. W. Lee, T. S. Uhm, D. D. Sung and Z. H. Ryu, J. Korean Chem. Soc.,
1988, 32, 85.
Received 9 June 2015; accepted 23 August 2015
Paper 1503415 doi: 10.3184/174751915X14412072517781
Published online: 2 October 2015
References
1
2
3
D. N. Kevill and Z. H. Ryu, J. Phys. Org. Chem., 2013, 26, 983.
D. N. Kevill and Z. H. Ryu, J. Chem. Res., 2014, 38, 387.
S. Winstein, E. Grunwald and H. W. Jones, J. Am. Chem. Soc., 1951, 73,
2700.
4
5
6
7
D. N. Kevill and M. J. D’Souza, J. Chem. Res., 2008, 61.
R. M. Laird and M. J. Spence, J. Chem. Soc. (B), 1970, 388.
R. M. Laird and M. J. Spence, J. Chem. Soc. (B), 1971, 454.
M. Leven, J. M. Nuedoerfl, and B. Goldfuss, Beilstein J. Org. Chem, 2013,
9, 155.
51 H. J. Koh, K. L. Han, H. W. Lee and I. Lee, J. Org. Chem., 1998, 63, 9834.