An efficient acid hydrolysis of the ether bond assisted by the neighbouring benzamide group. Part 3

Article abstract of DOI:10.1039/b008748n

The acid hydrolytic cleavage of 1 to 2, assisted by a neighbouring benzamide group, was kinetically investigated in a wide range of HCl concentrations (1.67 - 8.7 M) and at various temperatures. Kinetic measurements were performed also in the presence of LiCl at constant ionic strength (I = 5 M). The rate determining step of the acid hydrolysis of 1 was recognized through the investigation of the acid catalyzed cyclization of 3 to 2. Thermodynamic activation parameters were calculated and the experimetal data discussed. The participation of the neighbouring benzamide group was evaluated to be about 7.3 * 10³.

Full text of DOI:10.1039/b008748n

View Article Online / Journal Homepage / Table of Contents for this issue
An efficient acid hydrolysis of the ether bond assisted by the
neighbouring benzamide group. Part 31,2
Antonio Arcelli,* Gianni Porzi,* Samuele Rinaldi and Sergio Sandri
2
Dipartimento di Chimica “G.Ciamician”, Via Selmi 2, 40126 Università di Bologna, Italy.
E-mail: porzi@ciam.unibo.it
Received (in Cambridge, UK) 30th October 2000, Accepted 3rd January 2001
First published as an Advance Article on the web 23rd January 2001
The acid hydrolytic cleavage of 1 to 2, assisted by a neighbouring benzamide group, was kinetically investigated in a
wide range of HCl concentrations (1.67–8.7 M) and at various temperatures. Kinetic measurements were performed
also in the presence of LiCl at constant ionic strength (I = 5 M). The rate determining step of the acid hydrolysis of
1 was recognized through the investigation of the acid catalyzed cyclization of 3 to 2. Thermodynamic activation
parameters were calculated and the experimental data discussed. The participation of the neighbouring benzamide
group was evaluated to be about 7.3 × 103.
order rate constant (kobs) versus the HCl concentration is not
Introduction
linear, i.e. the hydrogen ion does not have a constant effect on
We have previously investigated the acid hydrolysis of the
the reaction rate, and at all the investigated temperatures the
methyl ether linkage of N-(2-ethyl-6-methylphenyl)-N-(1-
methoxypropan-2-yl)acetamide (1a) assisted by the vicinal
amide group.1,2 To the best of our knowledge the acid cleavage
of a dialkyl ether intramolecularly assisted by an amide
function has not previously been observed. The hypothesized
mechanism involves two kinetically indistinguishable pathways.1
In order to obtain further information about such assisted
reactions and to investigate the influence of the acyl group on
the reaction rate, we extended our study to the substrate N-
(2-ethyl-6-methylphenyl)-N-(1-methoxypropan-2-yl)benzamide
(1b). Thus, in this paper, the results of a kinetic investigation of
the acid hydrolysis of 1b to the oxazolidine derivative 2b, easily
converted to the open derivative 3b (Scheme 1), are reported.
Results and discussion
The kinetic measurements were performed over a wide acidity
range (1.67–8.7 M HCl) and at various temperatures (Table 1).
The experimental data for the acid hydrolysis of 1b at 71.2 ЊC,
plotted in Fig. 1, show that the dependence of the pseudo first
Scheme 1
Table 1 Rate constants and experimental conditions for acid hydrolysis of 1b
104 × kobs/sϪ1
[HCl]/
mol dmϪ3
45.4 ЊC
55.6 ЊCa
55.6 ЊCb,c
59.8 ЊC
66 ЊC
71.2 ЊCd
71.2 ЊCb,e
1.67
2.42
3.03
0.407
0.595
0.743
0.43
1.03
2.39
3.64
4.09
0.041
0.0714
0.16
0.26
0.545
0.984
1.58
1.66f
11.1g
3.47
4.01
4.70
4.92
5.44
5.90
5.99
6.59
7.42
8.70
0.169
0.35
0.58
1.13
2.28
0.973
2.24
3.9
5.70
6.47
1.85
4.10
6.4
8.96
0.808
7.39
13
17.5
30.5
47
1.85
3.25
6.25
10.40
9.58
15.5
16.8
27.90
Ϫ5
a The slope calculated in the range 2.42–4.01 M HCl is kH = 2.62 × 10 dm3 molϪ1 sϪ1 (r = 0.987). b At I = 5 M (LiCl). c The slope is
ϩ
Ϫ5
Ϫ4
kH = 2.42 × 10 dm3 molϪ1 sϪ1 (r = 0.987). d The slope calculated in the range 1.67–4.01 M HCl is kH = 1.76 × 10 dm3 molϪ1 sϪ1 (r = 0.976).
ϩ
ϩ
e The slope is kH = 1.34 × 10Ϫ4 dm3 molϪ1 sϪ1 (r = 0.996). f At 63.2 ЊC. g At 79.5 ЊC.
ϩ
296
J. Chem. Soc., Perkin Trans. 2, 2001, 296–301
This journal is © The Royal Society of Chemistry 2001
DOI: 10.1039/b008748n

View Article Online
Fig. 1 Dependence of kobs vs. [HCl] at 71.2 ЊC for the hydrolysis of 1b
(ᮀ), for the hydrolysis of 1b at ionic strength I = 5 M with LiCl (ꢀ) and
for the acid cyclization of 3b (✭). Points are experimental and straight
lines are calculated.
Scheme 2
involves two reaction steps (cyclization, then water attack),
while in the conversion of 3 into 2 the process rate is given by
the cyclization, the subsequent proton transfer being very fast.5
To this end kinetic measurements of the acid cyclization of 3b
to 2b were carried out at 71.2 ЊC and we found that at [HCl] >
5 M the rate constants are very similar, or coincident, to those
obtained for the acid hydrolysis of 1b to 2b (Fig. 1 and Tables 1,
2). This behaviour is consistent with the Bunnett–Olsen plot
built for the cyclization of 3b to 2b which is linear in the whole
acidity interval, the slope (0.5) being very similar to that found
for the hydrolysis of 1b at high acidity (Fig. 2). This result
suggests that the first step in the acid hydrolysis of 1, i.e. the
cyclization to the oxonium ion, should be the rate determining
step of the whole process. In this way, the different behaviour
between 3b and 1b at low acidity could be due to the fact that
the acidity function H0 correctly describes the protonation of
3b, but does not describe so well the substrate 1b.
Fig. 2 Bunnett–Olsen plots of the acid hydrolysis of 1b at 71.2 ЊC (ꢀ),
at 55.6 ЊC (᭜), at 45 ЊC (᭹) and of the acid cyclization of 3b at 71.2 ЊC
(᭛). The straight lines are calculated.
In favour of the mechanism hypothesized in Scheme 2 are the
Molecular Electrostatic Potential (MEP) calculations showing
that the preferential site of protonation for both the substrates
1b and 3b is the carbonyl oxygen. The geometries of the lowest
energy conformers for both the substrates 1(a,b) and 3(a,b)
were calculated by the AM1 semiempirical method (Hyper-
Chem Program, 1994) by using the “Polak-Ribiere” algorithm
(RMS gradient 0.01 kcal).
plots evidence two distinct regions with different slopes (see
Fig. 1). In actual fact, the kobs steeply increases when the [HCl]
is raised from about 5 to 8.7 M, while the increase is smaller in
the range 1.67–4.92 M HCl (see Fig. 1).
To explain these experimental data, we have elaborated the
kinetic results by means of the Bunnett–Olsen treatment.3 As
shown in Fig. 2, similar behaviours were obtained at different
temperatures: straight lines of almost identical slopes (0.3–0.4)
at high acidities, then a break and a second region, at low acidi-
ties, in which the slopes are close to zero or slightly negative.
The changes in the plot slopes could be interpreted as a con-
sequence of: a) a change in the reaction mechanism or in the
rate determining step, b) an inaccurate description of the sub-
strate protonation, by means of the H0 acidity function. For the
acid hydrolysis of 1b a mechanism different from that shown in
Scheme 2 and already hypothesized for N-(2-ethyl-6-methyl-
phenyl)-N-(1-methoxypropan-2-yl)acetamide (1a)1 appears
very unlikely.
On the other hand, it is unwise to take the Bunnett–Olsen
criteria alone as evidence to ascertain a possible change in the
reaction mechanism or rate determining step.4a Therefore, in
order to clarify the hydrolytic process, we thought it useful
to investigate the substrate N-(2-ethyl-6-methylphenyl)-N-(1-
hydroxypropan-2-yl)benzamide (3b) which in acid medium
cyclizes to the oxazolidine derivative 2b (Scheme 2). We
believed the alcohol 3b a suitable compound for providing very
useful information about the rate determining step of the over-
all hydrolytic process of 1b. In fact, the conversion of 1 into 2
The acid hydrolysis of the substrate 1 (depicted by S) to the
cyclic product 2 can be schematized:
S ϩ Hϩ
SHϩ → [Cyclic intermediate]ϩ ϩ H2O → 2
Since in our case the unprotonated species S is unreactive, the
overall low rate can be expressed, according to the transition
state theory,6a by eqn. (1) where k0 is the rate constant for the
ϩ
ϩ
#
ϩ
ϩ
kobs = [k0/(KSH ϩ [H ])] [H ] fSH /f
(1)
ϩ
slow step, KSH the dissociation constant for the protonated
substrate, fSH and f # the activity coefficients of the protonated
ϩ
substrate and the transition state, respectively. Rearrangement
of eqn. (1) yields eqn. (2).
#
kobs = [k0[SHϩ]/([S] ϩ [SHϩ])] fSH /f
(2)
ϩ
If the substrate is essentially unprotonated stoichiometrically,
i.e. [S] ӷ [SHϩ], the expression (2) assumes the form given in
eqn. (3) while when [S] Ӷ [SHϩ] eqn. (4) is applied.
J. Chem. Soc., Perkin Trans. 2, 2001, 296–301
297

View Article Online
Table 2 Rate constants and experimental conditions for acid cyclization of 3b to 2b
104 × kobs/sϪ1
[HCl]/
mol dmϪ3
55.3 ЊC
59.8 ЊC
65.8 ЊC
68.4 ЊC
71.2 ЊC
79.1 ЊC
1.74
3.03
4.65
5.22
5.99
7.42
8.70
1.87
2.35
5.53
9.27
13.7
29.5
46.4
0.348
0.63
1.37
1.72
5.51
5.35
16
20.2
62.8
Table 3 Thermodynamic activation parametersa,b for acid hydrolysis of 1a, 1b and for cyclization of 3b to 2b at different conditions
Entry
[HCl]/mol dmϪ3
∆G# c/kJ molϪ1
∆H# d/kJ molϪ1
∆S#/J molϪ1 KϪ1
T∆S#/kJ molϪ1
1a→2a
1
2
3.03 I = 5 M (LiCl)
7.42
110
109
3
5
99
98
2
4
Ϫ35
Ϫ33 11
6
Ϫ11
Ϫ11
2
4
1b→2b
3
4
5
6
7
3.03
4.92
7.42
8.70
112
109.5
106
105
109.5
4
2
3
2
5
108
101 1.5
95
91
106
3
Ϫ13
Ϫ27 4.5
Ϫ35
Ϫ44 4.5
Ϫ11 11
8
Ϫ4 2.5
Ϫ8.5 1.5
2
1
4
6
Ϫ11
Ϫ14
2
1
3.03 I = 5 M (LiCl)
Ϫ3.5 3.5
3b→2b
8
9
3.03
7.42
111.5
107
3
3
108.5
97
2
2
Ϫ9.5
Ϫ30
6
6
Ϫ3
Ϫ10
2
2
#
a Obtained from kH = kobs/[Hϩ]. b Calculated at 50 ЊC from the equation ln kH = ∆S /R ϩ ln (kT/h) Ϫ Ea/RT plotting ln kH vs. 1/T by means of a
linear least square; standard errors are calculated according to ref. 7. c Calculated at 50 ЊC from ∆G# = ∆H# Ϫ T∆S#. d Calculated at 50 ЊC from the
equation ∆H# = Ea Ϫ RT.
ϩ
ϩ
ϩ
#
kobs = (k0[Hϩ]/KSH ) fSH /f
(3)
(4)
above, in the range 5–8.7 M HCl the kobs depends on both the
ϩ
ϩ
acidity and the fSH /f # ratio [eqn. (1)], the latter has to increase
ϩ
#
ϩ
kobs = k0 fSH /f
considerably at [HCl] > 5 M, i.e. at higher ionic strengths.
Really, the influence of the ion atmosphere on the fSH /f # ratio
ϩ
For the purpose of explaining the dependence of kobs vs.
[HCl] displayed in Fig. 1, in view of the above eqns. (1) or (2),
the dissociation constant at 70 ЊC was measured and the value
should be larger at such high ionic strengths because intra-
molecular reaction occurs between a neutral-dipolar (-OCH3)
and a charged group (the protonated carbonyl).6b
ϩ
of pKSH = Ϫ2.6 was obtained (see Experimental). Thus, this
implies that at H0 > Ϫ1.6 (i.e. [HCl] < 5 M) the substrate will be
essentially unprotonated and therefore eqn. (3) can be applied.
In the range H0 = Ϫ1.6 to Ϫ2.75 (i.e. 5–8.7 M HCl), where
appreciable quantities of both SHϩ and S exist in the pre-
equilibrium mixture, eqn. (2) will be followed. At H0 < Ϫ3.6
(i.e. [HCl] > 11.5 M) the substrate will be essentially protonated
and eqn. (1) will reduce to eqn. (4).
Thermodynamic activation parameters
As is well known, rather than the acidity function dependence
such as the Bunnett–Olsen relationship, thermodynamic acti-
vation parameters are useful for providing further confirmatory
evidence of the possible reaction mechanism.
The thermodynamic activation parameters for the acid
catalyzed hydrolysis of 1b to 2b and cyclization of 3b to 2b were
determined from the kobs reported in Tables 1, 2, respectively.
The values, evaluated from the linear relationship ln(kobs/[Hϩ])
vs. 1/T, are reported in Table 3. The data obtained show that the
decrease in free activation energy due to both an increase in
HCl and the addition of LiCl is accompanied by an enthalpic
advantage which exceeds the unfavourable entropy change.
Since the investigated intramolecular reaction occurs between
a charged and a dipolar group, the electrostatic contribution to
the overall free activation energy is not relevant.6c Really, the
change in relative permittivity (by adding LiCl) does not signifi-
cantly modify the ∆G#: to this end compare entries 3, 4 and 7 in
Table 3, taking into account that HCl and LiCl have a different
molar depression of relative permittivity.8 Thus, we believe that
the decrease in ∆G# observed on going from 4.92 to 8.7 M HCl
(entries 4, 5 and 6) can be ascribed to a significant increase in
From the plot reported in Fig. 1, in the range 1.67–4 M HCl a
linear dependence of kobs against HCl concentration can be
#
ϩ
observed, suggesting that on the basis of eqn. (3) the fSH /f
ratio is almost constant. This observation is strengthened by the
kinetic measurements conducted at various HCl concentrations
in the presence of LiCl at constant ionic strength (I) in order to
minimize the change in activity coefficients (Fig. 1). In fact, at
I = 5 M in the range 1.67–4.7 M HCl a good linear plot was
ϩ
ϩ
obtained according to the equation kobs = kH [H ] ϩ kЈ. The
Ϫ4
slope kH = 1.34 × 10 dm3 molϪ1 sϪ1 is the second-order
ϩ
rate constant (Table 1) and the intercept kЈ = 2.09 × 10Ϫ5 sϪ1
,
about 30-fold smaller than kobs measured at 1.67 M HCl, agrees
with the not observed uncatalyzed competitive reaction. The
similarity of this slope with that found in the absence of LiCl
Ϫ4
(kH = 1.76 × 10 dm3 molϪ1
s
Ϫ1) in the same acidity range
ϩ
#
ϩ
confirms that the fSH /f ratio remains practically constant. A
divergent behaviour is shown in the rate–acidity profile at
[HCl] > 5 M where the kobs is not linearly dependent on the
acidity, but it follows an apparently exponential trend. A
reasonable explanation can be the following. Since, as reported
the fSH /f # ratio [eqn. (1)].
ϩ
As shown in Table 3, the thermodynamic activation param-
eters calculated for the acid cyclization of 3b to 2b (entries 8
and 9) are very similar, within experimental error, to those
298
J. Chem. Soc., Perkin Trans. 2, 2001, 296–301

View Article Online
Table
4
Rate constants and experimental conditions for acid
hydrolysis of 1a
104 × kobs/sϪ1
57.8 ЊC,
I = 5 M
(LiCl)c
[HCl]/
mol dmϪ3
50 ЊC
57.8 ЊCa,b
69.9 ЊC
79.9 ЊC
1.67
2.42
3.03
4.01
7.42
0.026
0.205
0.382
0.697
2.86
0.596
0.726
0.894
1.1
1.1
11.3
26
a In the range 1.67–4.01 M HCl the slope is kH = 2.9 × 10Ϫ5 dm3 molϪ1
ϩ
sϪ1 (r = 0.997). b The kobs measured in the range 4.75–8.37 M HCl are
Ϫ5
reported in ref. 1 and the respective slope is kH = 6.18 × 10 dm3
ϩ
molϪ1 sϪ1 (r = 0.994). c The slope is kH = 2.2 × 10 dm3 molϪ1 sϪ1
Ϫ5
Fig. 4 Dependence of log ([SHϩ]/[S]) vs. H0 for 1a (ꢀ) and for 1b (᭹).
ϩ
(r = 0.997).
findings show that in the range 1–4 M HCl, analogously to that
observed for 1b, the ionic strength variation has a negligible
ϩ
effect on the reaction rate. From the pKSH value Ϫ1.9, meas-
ured for 1a at 60 ЊC (see Experimental), it can be deduced that
in the range H0 = Ϫ0.9 to Ϫ2.9 (i.e. 2.7–9.3 M HCl) appreciable
quantities of both protonated and unprotonated substrate exist
in the pre-equilibrium mixture. Then, in this range eqn. (2)
should be followed, i.e. the profile of kobs vs. acidity should be
not linear, but rather resemble that observed for 1b in the range
5–8.7 M HCl. Conversely, the rate–acidity profile is linear also
at [HCl] > 4 M where, furthermore, the sensitivity of the rate to
the medium acidity appears considerably smaller with respect
to 1b. This behaviour is difficult to interpret. Naively, for the
substrate 1a the acidity increasing, also at high ionic strength,
causes a small change in the fSH /f # ratio.
ϩ
These findings suggest that the unlike behaviour of two sub-
strates at high acidity does not appear ascribable to the differ-
Fig. 3 Dependence of kobs vs. [HCl] at 57.8 ЊC for the hydrolysis of 1a
(ᮀ) and for the hydrolysis of 1a at ionic strength I = 5 M with LiCl (᭹).
The values at [HCl] > 4.75 M are taken from the ref. 1. Points are
experimental and straight lines are calculated.
ϩ
ences in the pKSH values, but more probably to the unequal
effect of the ionic strength on the fSH /f # ratio of 1a and 1b.
ϩ
The thermodynamic activation parameters calculated for 1a
at 3.03 M HCl and I = 5 M (LiCl), from data reported in Table
4, compare well with those obtained for 1b under the same
conditions (Table 3). The hydrolysis of 1a and 1b occurs with
the same ∆G# though the former substrate requires a smaller
∆H# and a greater ∆S#. The identical value (within experi-
mental error) of free activation energy could be consistent with
the similar stabilization of the two protonated reactants, con-
sidering that the transition states will probably be very similar
in energy. In this regard, semi-empirical quantum mechanical
calculations, performed by the AM1 method (from HyperChem
program, 1994) allowed us to deduce that the positive charge of
the protonated substrate 1b is poorly delocalized because the
aromatic ring is about 40Њ out of the carbonyl plane. Thus, on
the protonated 1b the weak spreading out of the charge over the
phenyl ring will be comparable with that induced by the methyl
group on the protonated 1a through hyperconjugation.
In any case the data reported in the present work and in the
previous paper1 shows that the lowering in ∆G#, with respect to
the unassisted reaction,1 is essentially due to an enthalpic gain.
This is in agreement with a recent report9 that for ring sizes
three through six the intramolecular reaction rate commonly
depends on the enthalpic change occurring as the reactant
structure is converted into the transition state.
obtained for the acid hydrolysis of 1b to 2b (entries 3 and 5)
under the same conditions. This considerable similarity sub-
stantiates the above assertion that the rate determining step of
the acid hydrolysis of 1 is its cyclization to the oxonium salt
intermediate (Scheme 2): thus, in the HCl concentration range
investigated, any change in the mechanism can be excluded.
Comparison between the substrates 1a and 1b
To better interpret the kinetic results we thought it may be help-
ful to compare the acid hydrolysis of 1b and 1a, the latter
having been previously investigated1 in the range 4.75–8.37 M
HCl. To this end kinetic measurements of acid hydrolysis of 1a
were also performed in the range 1.67–4.01 M HCl at 57.8 ЊC
(Table 4 and Fig. 3). The Bunnett–Olsen plot (not reported)
built for 1a showed a very similar behaviour to that observed
for 1b, with a breaking point that generates two slopes, 0.721
and 0.15. It is also interesting to note that the Bunnett–Olsen
plot (not reported) built for the cyclization of N-(2-ethyl-6-
methylphenyl)-N-(1-hydroxypropan-2-yl)acetamide (3a) is
linear, the slope being 0.9, in the entire range of [HCl] investi-
gated as found for 3b.
The substrate 1a was submitted to kinetic measurements also
at ionic strength I = 5 M (LiCl) (Table 4 and Fig. 3).
The data obtained suggest some interesting considerations.
In the range 1.67–4.01 M HCl at variable ionic strength, where
Conclusions
the rate changes linearly with increasing [HCl], the slope, i.e.
Although some difficulties in interpreting experimental data
have to be resolved (from both a chemical and mathemat-
ical point of view), the above mentioned findings allow the
following conclusions.
Ϫ5
dm3 molϪ1 sϪ1
, is comparable to kH =
ϩ
ϩ
kH = 2.9 × 10
2.2 × 10Ϫ5 dm3 molϪ1 sϪ1 found at I = 5 M (LiCl) (Table 4). In
addition, these values are similar to those measured for the
Ϫ5
substrate 1b (kH = 2.62 × 10 dm3 molϪ1
s
Ϫ1) at 55.6 ЊC both
For both substrates investigated, 1a and 1b, the mechanism
ϩ
in the absence and in the presence of LiCl (Table 1). These
involved is that reported in Scheme 2. In this way, the possible
J. Chem. Soc., Perkin Trans. 2, 2001, 296–301
299

View Article Online
alternative pathway proposed in the previous paper1 can there-
fore be excluded.
127.3, 127.7, 127.8, 128.0, 128.7, 128.8, 130.0, 135.4, 135.5,
140.2, 141.1, 170.9.
The rate determining step is the cyclization of the protonated
substrate, the subsequent attack of H2O on the cationic cyclic
intermediate being faster.
Synthesis of N-(2-ethyl-6-methylphenyl)-N-(1-hydroxypropan-2-
yl)benzamide (3b)
The substrates 1a and 1b show a very similar sensitivity to the
acidity or to the ionic strength at values below 5 M, while at
higher values the compound 1b shows a remarkably greater
sensitivity than 1a. This behaviour does not appear ascribable
The acid aqueous solution of 2b was made alkaline with
NaOH, then extracted with ethyl acetate. The oil product,
recovered after total evaporation in vacuo of the organic
solvent, was pure by the TLC analysis. 1H-NMR δ 1.15 (2t, 3H,
J = 7.5 Hz), 1.47 (2d, 3H, J = 7.1 Hz), 2.3 (2s, 3H), 2.52 (m, 1H),
2.68 (m, 1H), 3.85 (m, 2H), 4.16 (m, 1H), 4.8 (m, OH), 7.15 (m,
8ArH); 13C-NMR δ 13.7, 14.0, 19.2, 19.4, 23.6, 23.7, 62.0, 62.2,
66.3, 126.6, 126.7, 127.4, 127.9, 128.0, 128.9, 130.1, 135.3, 135.8,
ϩ
to the different pKSH values (Ϫ1.9 and Ϫ2.6, respectively), but
#
ϩ
rather to a greather effect of the ionic strength on the fSH /f
ratio [eqn. (1)] for 1b than 1a.
The kH values of hydrolysis of 1b (5.4 × 10Ϫ4 dm3 molϪ1 sϪ1
at 71.2 ЊC) and reference substrate1 (7.7 × 10Ϫ8 dm3 molϪ1 sϪ1
,
,
ϩ
141.0, 172.5. IR (thin film): ν = 1626 (C᎐O), 3430 cmϪ1 (OH).
᎐
at 69.9 ЊC), allow us to estimate that the participation of the
neighbouring benzamide group is about 7.3 × 103. This value
is lowered to about 3 × 103 by comparison with diethyl ether.
In fact, the polar effect of the protonated amino-ether of the
unassisted model can be estimated as 2.5-fold.1
؉
pKSH measurements
ϩ
The pKSH measurements of 1a and 1b were performed spec-
trophotometrically in HCl solutions. Samples were prepared by
adding 30 µL of a stock solution of 1a (0.11 M in MeOH) and
1b (9.8 × 10Ϫ3 M in MeOH) at 3 mL of acid solution thermo-
statted at 60 and 70 ЊC, respectively. The optical density (OD)
was monitored at λ = 266 nm for 1a and at λ = 258 nm for 1b.
The OD value of unprotonated 1a was evaluated in aqueous
alcoholic solutions (1–2 M HCl–MeOH = 2:1), the product
being insoluble in low HCl concentration. The addition of
MeOH does not produce spectral change. Since at HCl > 5 M
the hydrolysis of 1b proceeds significantly, the absorbance was
recorded as function of time and it was extrapolated to t = 0.
The OD values of protonated 1a and 1b species were measured
in 9–11 M H2SO4 because concentrated HCl does not allow
total protonation. The absorbances at the λ employed did not
change when H2SO4 instead of HCl was used. In addition,
owing to the medium effects, the OD values of protonated and
unprotonated 1b were estimated by using the Katritzky
method.10 Sigmoid curves of OD vs. H0 (values corrected at
60 ЊC for 1a and extrapolated to 70 ЊC for 1b)4b,c were obtained.
The linear plots of log ([SHϩ]/[S]) against H0 are reported in
In conclusion, in order to achieve a more satisfactory solu-
tion to the complex problem about the influence of the acidity
and ionic strength on the reaction rate, further investigation is
in progress providing additional and useful data especially on
the salt effects.
Experimental
General
1H- and 13C-NMR spectra were recorded with a Varian Gemini
300 (300 MHz) instrument by using CDCl3 as solvent. UV
spectra and kinetic measurements were recorded on a Perkin-
Elmer Lambda 6 spectrophotometer.
Synthesis of N-(2-ethyl-6-methylphenyl)-N-(1-methoxypropan-
2-yl)benzamide (1b)
It was obtained by treating the N-(2-ethyl-6-methylphenyl)-N-
(1-hydroxypropan-2-yl)amine [1H-NMR δ 1.07 (d, 3H, J = 6.5
Hz), 1.25 (t, 3H, J = 7 Hz), 2.3 (s, 3H), 2.45–2.75 (m, 4H), 3.36
(m, 1H), 3.5 (dd, 1H, J = 6.8, 11 Hz), 3.69 (dd, 1H, J = 4.1, 11
Hz), 6.85–7.1 (m, 3ArH)], synthesized starting from 6-ethyl-o-
toluidine and ( )-ethyl 2-bromopropionate followed by lithium
borohydride reduction, with an equivalent amount of NaH
(80% dispersion in mineral oil) under an inert atmosphere.
After about 2 h, CH3I (1.5 equivalent) was added dropwise and
the reaction mixture stirred for about 4 h at rt. The crude reac-
tion product [1H-NMR δ 1.21 (d, 3H, J = 6.7 Hz), 1.26 (t, 3H,
J = 7.6 Hz), 2.31 (s, 3H), 2.68 (q, 1H, J = 7.6 Hz), 3.37 (m, 7H),
6.9 (m, 1ArH), 7.05 (m, 2ArH)] was isolated and reacted in
CHCl3 with benzoyl chloride in the presence of triethylamine.
The reaction mixture was refluxed for about 8 h and worked up
by the usual procedure. After purification by silica gel chrom-
atography eluting with hexane–ethyl acetate, the pure product
was isolated as an oil in 70% overall yield. 1H-NMR δ 1.13 (2t,
3H, J = 7.5 Hz), 1.38 (2d, 3H, J = 6.7 Hz), 2.28 (s, 3H), 2.58
(2m, 2H), 3.33 (2s, 3H), 3.79 (m, 1H), 3.97 (m, 2H), 7.1 (m,
8ArH); 13C-NMR δ 13.6, 13.8, 15.4, 15.6, 19.3, 19.5, 23.7, 57.6,
58.7, 75.1, 75.3, 126.2, 127.3, 127.7, 127.8, 128.5, 129.7, 135.9,
136.1, 136.5, 140.6, 141.5, 141.6, 170.9.
ϩ
ϩ
Fig. 4. The pKSH values, calculated from the equation pKSH
=
H0 ϩ n log ([SHϩ]/[S]), are: Ϫ1.9 0.15 (n = 0.73) for 1a and
Ϫ2.6 0.20 (n = 0.82) for 1b.
Kinetic experiments
The kinetic measurements of acid catalysed hydrolysis of 1a
(Table 4) were followed as described in ref. 1. The hydrolysis
of 1b was investigated at various temperatures, in a large range
of HCl concentrations (1.0–8.7 M) and in the presence of
LiCl (Table 1). The reaction was followed spectrophoto-
metrically by measuring the change in the OD at λ = 258 nm.
30 µL of the stock solution of 1b (0.02 M in MeOH) were
added to 3 mL of HCl thermostatted in a 1 cm path cell of the
spectrophotometer. The pseudo first order rate constants (kobs
)
were obtained from the equation OD = ODO ϩ (OD∞ Ϫ DO) ×
[1 Ϫ exp(Ϫt × kobs)] by plotting at least 200 values of OD with a
non-linear last square routine (FigP6.0 programme by Biosoft).
Excellent plots were obtained with correlation coefficients
above 0.999. OD∞ values were taken after at least ten half lives.
In all cases investigated the reaction followed a first order over
at least 90% of reaction.
After completion of the reaction, about ten half lives at
66 ЊC, the compound 3b, isolated by alkalinization, was the sole
reaction product and it was identified by its 1H-NMR spectrum
(see above). The structure of the cyclic product 2b was deduced
Synthesis of 3-(2-ethyl-6-methylphenyl)-2-hydroxy-4-methyl-2-
phenyloxazolidine (2b)
It was obtained by submitting 1b to hydrolysis in 5 M HCl at
80 ЊC for at least 2 h. The reaction solution was then evaporated
to dryness in vacuo and the oily residue was pure by TLC
analysis. 1H-NMR δ 1.12 (t, 3H, J = 7.5 Hz), 1.18 (t, 3H, J = 7.5
Hz), 1.47 (2d, 3H, J = 6.7 Hz), 2.28 (2s, 3H), 2.55 (m, 2H),
3.95 (m, 1H), 4.12 (m, 2H), 7.15 (m, 8ArH); 13C-NMR δ 13.6,
13.8, 14.6, 14.9, 19.1, 19.4, 23.6, 46.8, 47.0, 60.3, 60.4, 126.5,
1
from the H-NMR spectrum, as previously reported for the
analogous N-(2-ethyl-6-methylphenyl)-N-(1-methoxypropan-2-
yl)acetamide (1a).1
The kinetics of acid catalysed cyclization of 3b to 2b were
followed spectrophotometrically (at λ = 258 nm) at 3.03 and
7.42 M HCl and at various temperatures (Table 2).
300
J. Chem. Soc., Perkin Trans. 2, 2001, 296–301

View Article Online
1970, p. 194; (b) C. H. Rochester, Acidity Functions, Academic Press,
London, 1970, p. 39; (c) C. H. Rochester, Acidity Functions,
Academic Press, London, 1970, p. 28.
Acknowledgements
Financial support from the University of Bologna (Funds for
selected research topics and Fondi Ricerca Istituzionale, ex
60%).
5 M. Eigen, Angew. Chem., Int. Ed. Engl., 1964, 3, 1.
6 (a) C. H. Bamford and C. F. H. Tipper, Comprehensive Chemical
Kinetics, Elsevier, New York, 1969, vol. 2, p. 310; (b) C. H. Bamford
and C. F. H. Tipper, Comprehensive Chemical Kinetics, Elsevier,
New York, 1969, vol. 2, p. 333; (c) C. H. Bamford and C. F. H.
Tipper, Comprehensive Chemical Kinetics, Elsevier, New York, 1969,
vol. 2, p. 323.
7 L. L. Schaleger and F. A. Long, Adv. Org. Chem., 1963, 1, 1.
8 R. A. Robinson and R. H. Stokes, Electrolyte Solutions,
Butterworths, London, 1959, p. 18.
References
1 A. Arcelli, G. Porzi and S. Sandri, Tetrahedron, 1995, 51, 9729 and
references therein.
2 A. Arcelli, M. Papa, G. Porzi and S. Sandri, Tetrahedron, 1997, 53,
10513.
3 T. H. Lowry and K. Schueller Richardson, Mechanism and
Theory in Organic Chemistry, Harper and Row, New York, 1987,
p. 259.
9 F. C. Lightstone and T. C. Bruice, Bioorganic Chemistry, 1998, 26,
193.
10 A. R. Katritzky, A. J. Waring and K. Yates, Tetrahedron, 1963, 19,
465.
4 (a) C. H. Rochester, Acidity Functions, Academic Press, London,
J. Chem. Soc., Perkin Trans. 2, 2001, 296–301
301