Formation of singlet molecular oxygen by the Radziszewski reaction between acetonitrile and hydrogen peroxide in the absence and presence of ketones

Article abstract of DOI:10.1039/b202620c

Measurements of the infrared phosphorescence of singlet molecular oxygen (¹O₂) at 1270 nm have been used to demonstrate the formation of ¹O₂ by the Radziszewski reaction between acetonitrile and hydrogen peroxide. The kinetics of the Radziszewski reaction either alone or in the presence of ketones have been studied by this technique. The rate-determining step of the ¹O₂ formation of the reaction in the absence of ketones was found to be independent of both the concentration of acetonitrile and that of hydrogen peroxide. The kinetic data, the results of the volumetric measurements of the oxygen liberated and the results of the determination of the amount of ¹O₂ generated by the reaction are consistent with the assumption that the reaction between acetonitrile and hydrogen peroxide occurs via the heterolytic decomposition of the intermediate, peroxyacetimidic acid (PAIA), forming 1 mol acetamide and 0.5 mol ¹O₂ according to the stoichiometric equation: CH₃CN + H₂O₂ → CH₃C(=O)NH₂ + 0.5 ¹O₂. The rate constant of the heterolytic decomposition of PAIA was determined to be k₈ = 1.2 × 10^-3 dm³ mol^-1 s^-1 at T = 30°C. From the measurements at different pH values in the range 9.1 < pH < 12.0 the pK_a(PAIA) value was estimated to be 11.1 at T = 30°C. The investigation of the reaction between acetonitrile and hydrogen peroxide by using N,N-dimethyl-4-oxopiperidinium nitrate as catalyst, has unequivocally shown that the rate of ¹O₂ formation is considerably enhanced by this ketone. For the ketone-catalysed decomposition of PAIA a rate law can be derived showing a first order dependence on the concentration of acetonitrile and hydrogen peroxide at a given pH. In accordance with the observed rate law are the results with acetonitrile in 50% acetone containing a tenfold excess of hydrogen peroxide at pH 8.2 and T = 60°C.

Full text of DOI:10.1039/b202620c

View Article Online / Journal Homepage / Table of Contents for this issue
Formation of singlet molecular oxygen by the Radziszewski
reaction between acetonitrile and hydrogen peroxide in the absence
and presence of ketones
2
Hans-Dieter Brauer,* Beate Eilers and Andreas Lange†
Institut für Physikalische und Theoretische Chemie, Universität Frankfurt/M.,
Marie-Curie-Straße 11, D-60439 Frankfurt/M., Germany.
E-mail: H.Brauer@chemie.uni-frankfurt.de
Received (in Cambridge, UK) 14th March 2002, Accepted 15th May 2002
First published as an Advance Article on the web 10th June 2002
1
Measurements of the infrared phosphorescence of singlet molecular oxygen ( O ) at 1270 nm have been used to
2
1
demonstrate the formation of O by the Radziszewski reaction between acetonitrile and hydrogen peroxide.
2
The kinetics of the Radziszewski reaction either alone or in the presence of ketones have been studied by this
1
technique. The rate-determining step of the O formation of the reaction in the absence of ketones was found to
2
be independent of both the concentration of acetonitrile and that of hydrogen peroxide. The kinetic data, the results
1
of the volumetric measurements of the oxygen liberated and the results of the determination of the amount of O2
generated by the reaction are consistent with the assumption that the reaction between acetonitrile and hydrogen
peroxide occurs via the heterolytic decomposition of the intermediate, peroxyacetimidic acid (PAIA), forming 1 mol
1
1
acetamide and 0.5 mol O according to the stoichiometric equation: CH CN ϩ H O
CH C(᎐O)NH ϩ 0.5 O .
2
3
2
2
3 2 2
᎐
Ϫ3
3
Ϫ1 Ϫ1
The rate constant of the heterolytic decomposition of PAIA was determined to be k = 1.2 × 10 dm mol
s
8
at T = 30 ЊC. From the measurements at different pH values in the range 9.1 < pH < 12.0 the pK (PAIA) value was
a
estimated to be 11.1 at T = 30 ЊC. The investigation of the reaction between acetonitrile and hydrogen peroxide by
1
using N,N-dimethyl-4-oxopiperidinium nitrate as catalyst, has unequivocally shown that the rate of O formation
2
is considerably enhanced by this ketone. For the ketone-catalysed decomposition of PAIA a rate law can be derived
showing a first order dependence on the concentration of acetonitrile and hydrogen peroxide at a given pH. In
accordance with the observed rate law are the results with acetonitrile in 50% acetone containing a tenfold excess
of hydrogen peroxide at pH 8.2 and T = 60 ЊC.
Introduction
The reaction of a nitrile with hydrogen peroxide in mildly
alkaline solution to produce an amide, water and oxygen is
1
known as the Radziszewski reaction. In the original paper the
overall reaction is given as
RCN ϩ 2 H O2
RC(᎐O)NH ϩ H O ϩ O
2
(1)
2
᎐
2
2
The kinetics of alkaline decomposition of hydrogen peroxide
and benzonitrile were first studied by Wiberg. He reported that
Scheme 1
2
the rate-determining reaction was the formation of peroxy-
Recently, an efficient ketone-catalysed epoxidation using
benzimidic acid (PBIA) from the addition of the perhydroxide
hydrogen peroxide as primary oxidant in combination with
Ϫ
(
OOH ) toward the nitrile. It was assumed that the inter-
6
acetonitrile has been developed. Using different ketones as
mediary-generated PBIA oxidizes another molecule of hydro-
gen peroxide by a successive fast reaction to evolve oxygen and
benzamide (Scheme 1)3.
catalysts it was found that trifluoroacetone (CF COCH ) gave
3
3
the best conversion of trans-β-methylstyrene as substrate in the
acetonitrile–hydrogen peroxide system.
It was Payne et al. who initially developed a method for
epoxidation and oxidation using hydrogen peroxide and an
organic nitrile under mildly alkaline conditions. Later, Bach
However, some doubts exist with respect to the generally
accepted mechanism of the Radziszewski reaction postulated
2
7
by Wiberg. Sawaki and Ogata have investigated the kinetics of
substituted benzonitriles with hydrogen peroxide in methanol–
water solutions and have found that the formation of the
peroxycarboximidic acids is not rate-determining in contrast to
the mechanism of Wiberg.
4
et al. reported a procedure for the large-scale epoxidation
of cyclooctene using the acetonitrile–hydrogen peroxide and
the more reactive trichloroacetonitrile–hydrogen peroxide sys-
tem under biphasic reaction conditions. Today it is generally
accepted that the ‘oxidation power’ of the nitrile–hydrogen per-
oxide system is caused by the intermediate, peroxycarboximidic
2
Wiberg has investigated the mechanism between hydrogen
peroxides and benzonitrile and substituted benzonitriles,
respectively, in 25–50% acetone–water solutions. At this time it
was not known that the heterolytic decomposition of peroxy
5
acid (PCIA).
8
9
acids like Caro‘s acid and peroxycarboxylic acids is efficiently
†
Present address: Hychem AG, Karl-Winnacker-Str. 22, D-36396
Steinau a. d. Str., Germany.
catalysed by ketones through the formation of dioxiranes as
1
288
J. Chem. Soc., Perkin Trans. 2, 2002, 1288–1295
This journal is © The Royal Society of Chemistry 2002
DOI: 10.1039/b202620c

View Article Online
intermediates combined with a change in the mechanism.
Whereas the uncatalysed decomposition of a peroxy acid shows
a second-order dependence on the concentration of the
peroxy acid, the ketone-catalysed decomposition is found to be
of first-order with respect to the concentration of the peroxy
Results and discussion
A
Investigation of the CH CN–H O system in water in the
3 2 2
absence of ketones
The question of whether the oxygen evolved in the Radzi-
szewski reaction is generated as O2 has been investigated
8,9
acid.
1
Our interest in the uncatalysed and the ketone-catalysed
decomposition of peroxy acids led us to a reinvestigation of
the Radziszewski reaction between acetonitrile and hydrogen
peroxide to clarify the mechanism. Furthermore we studied the
mentioned reaction in the presence of N,N-dimethyl-4-oxo-
7,17
already using chemical traps. However, neither Sawaki and
7
17
Ogata nor McKeown and Waters could prove the generation
1
18
of O . In contrast to this result our preliminary tests with the
2
IRL technique have unequivocally shown that in the reaction of
1
acetonitrile with hydrogen peroxide O is formed, whereby the
2
ϩ
Ϫ
piperidinium nitrate (DMOP NO3 ) as catalyst as well in 50%
1
generation of O strongly depends on pH.
2
acetone–water.
1
A.1 O Phosphorescence intensity as a function of time.
2
1
Fig. 1 shows the O2 phosphorescence intensity I (t) as a
p
function of time observed for the reaction of equimolar
quantities of acetonitrile and hydrogen peroxide at pH 9.9. For
1
comparison no O signal was observed at pH 8.2 using the
2
acetonitrile–hydrogen peroxide system. The plot of (I (t)/
p
Ϫ1/2
mV)
versus time resulted in a straight line [inset of Fig. 1]
indicating that the rate-determining step for the O generation
is of second-order (vide infra). At first sight this result seems
to confirm the mechanism postulated by Wiberg. According to
1
2
8f
2
this mechanism the rate-determining step should show at a
given pH a first-order dependence on the concentration of the
nitrile and hydrogen peroxide, i.e. the rate-determining step
should be of second-order.
Experimental
Materials
Acetonitrile (Fluka, for UV-spectroscopy), acetone (Aldrich,
HPLC grade) and 30% hydrogen peroxide (Merck, pract.) were
used as received. Reactions were carried out in buffered aque-
ous solutions in the pH region 8–12 in the presence of Dequest
Consequently, when hydrogen peroxide or acetonitrile is used
in a large excess the rate-determining step should be pseudo-
first-order with respect to the concentration of acetonitrile and
of hydrogen peroxide, respectively. If this holds true then the
2
066 (Monsanto) as effective metal-chelating agent. H O was
plot of ln (I (t)/mV) versus time must result in a straight line. To
p
2
distilled twice and the buffer components were of the highest
prove this prediction we have performed measurements using
one of the educts in a large excess.
8f
ϩ
Ϫ
grade as previously described. DMOP NO was synthesized
3
1
0
by the literature procedure followed by anionic exchange of
Fig. 2 reveals the I (t) curve obtained for the reaction at pH
p
11
the requisite iodide salt with silver nitrate.
10.5 with a tenfold excess of hydrogen peroxide. In contrast to
the expectation the plot of ln (I (t)/mV) versus time does not
p
yield a straight line (it is not shown). But as shown in the inset
Singlet oxygen infrared emission measurements
Ϫ1/2
of Fig. 2 the plot of (I (t)/mV)
versus time is found to be
p
1
The O phosphorescence at λ = 1270 nm was recorded with a
2
linear, i.e. also in the presence of a tenfold excess of hydrogen
laboratory made near-infrared luminescence (IRL) spectro-
peroxide a second-order reaction is observed for the generation
1
2–15
1
meter.
For the intensity of the O2 phosphorescence
1
of O .
2
14
11
emission, I , eqn. (4) holds, where c = (1.0 ± 0.1) × 10 mV
dm mol s is a constant of the IRL spectrometer, k is the rate
p
As Fig. 3 demonstrates a similar result is obtained for
the reaction with a tenfold excess of acetonitrile. Obviously the
results found for the reaction between acetonitrile and hydro-
gen peroxide in water are in contrast to the mechanism of
3
Ϫ1
p
constant of the
2
Wiberg. Our results suggest that neither acetonitrile nor
(
4)
hydrogen peroxide is involved in the rate-determining step of
1
the O -generation and as a consequence the formation of the
2
1
Ϫ1 15
intermediate, PAIA, cannot be the rate-determining step, in
O phosphorescence emission (in water: k = 0.21 s
)
and τ∆
2
p
7
1
16
agreement with the observation made by Sawaki and Ogata.
is the lifetime of O (in water: τ = 3.1 µs). The errors of kp
and τ , respectively, amount to about 10%. d[ O ]/dt denotes the
rate of singlet oxygen formation.
2
∆
1
1
Considering the fact that O
2
is evidently formed via a rate-
∆
2
determining second-order step independently of the ratio of the
1
concentrations of acetonitrile and hydrogen peroxide, it can be
The concentration of O2 can be calculated according to
1
deduced that O is generated by the heterolytic decomposition
2
eqn. (5), where C represents the product ck τ = (6.5 ± 1.3) ×
1
p
∆
4
3
Ϫ1 Ϫ1
of two molecules of PAIA, i.e. by a mechanism generally
0 mV dm mol
s .
observed for the decomposition of inorganic peroxy acids and
19
peroxycarboxylic acids, respectively. However, in this case
only one molecule of hydrogen peroxide should be involved
(
5)
1
in the generation of O . Thus for the stoichiometric equation
2
of the reaction between acetonitrile and hydrogen peroxide
eqn. (6) should be valid.
Reactions were carried out in thermostatted 1 cm quartz
cuvettes at constant temperature. To the buffer/Dequest
066/acetonitrile solutions and to the buffer/Dequest 2066/
1
CH CN ϩ H O
CH C(᎐O)NH ϩ 1/2 O
(6)
2
3
2
2
3
᎐
2
2
acetonitrile/ketone solutions, respectively, 30% hydrogen
peroxide was added by a micro syringe. Vigorous stirring
was conducted. The initial concentrations at time t = 0 of
the used educts are given in the legends of the figures or in
the text.
A.2 Determination of the amount of oxygen. Neither
1
2
Radziszewski nor Wiberg has reported quantitative measure-
ments of the oxygen evolved in the reaction between benzo-
20
nitrile and hydrogen peroxide. Up to now only McIsaac et al.
J. Chem. Soc., Perkin Trans. 2, 2002, 1288–1295 1289

View Article Online
Ϫ3
Fig. 1 I (t) as a function of time for the reaction between acetonitrile and hydrogen peroxide. Conditions: [CH CN](0) = [H O ](0) = 0.94 mol dm
,
;
p
3
2
2
Ϫ4
Ϫ3
Ϫ1/2
Ϫ4 Ϫ1
[
Dequest] = 1.0 × 10 mol dm , pH = 9.9 and T = 30 ЊC. The inset shows the corresponding plot of (I (t)/mV)
versus time (sl = 1.49 × 10
s
p
int = 1.49).
Ϫ3
Fig. 2 I (t) as a function of time for the reaction between acetonitrile and hydrogen peroxide. Conditions: [CH CN](0) = 0.5 mol dm , [H O ](0) =
p
3
2
2
Ϫ3
Ϫ4
Ϫ3
Ϫ1/2
5
.0 mol dm , [Dequest] = 1.0 × 10 mol dm , pH = 10.5 and T = 30 ЊC. The inset shows the corresponding plot of (I (t)/mV)
versus time;
p
Ϫ4 Ϫ1
(
sl = 1.82 × 10
s ; int = 1.13).
have conducted gas volumetric measurements of the evolved
oxygen from the reaction of p-cyanobenzoic acid with hydrogen
peroxide. The authors reported that these measurements would
confirm the stoichiometry of eqn. (1). However, in this study
ethylenediaminetetraacetate (EDTA) was used as chelating
agent with [EDTA] = 8 × 10 mol dm . It is of interest that a
significant curvature of the second-order and the first-order
plots of the given data was observed. It seems that the decom-
position of the in-situ generated peroxyimidic acid is extra-
ordinarily sensitive to transition-metal ions and some radical
decomposition of the peroxyimidic acid may have occurred
using EDTA as chelating agent and resulted in an enhancement
of the liberated oxygen.
accordance with eqn. (6) it was found that 0.52 ± 0.06 mol of
oxygen is generated from 1 mol acetonitrile.
1
Additionally, the amount of O was determined with the
2
IRL technique. The measurement was also carried out at
Ϫ3
pH 9.9 and at T = 60 ЊC using 0.5 mol dm acetonitrile and
Ϫ5
Ϫ3
Ϫ3
1
5.0 mol dm hydrogen peroxide in solution. The O phosphor-
2
escence curve obtained is depicted in Fig. 4. The area under the
1
curve corresponds to the amount of the evolved O . The area
2
was estimated to be (18914 ± 4%) mV s and according to
4
Ϫ3
Ϫ1
eqn. (5) with C = 6.5 × 10 mV dm mol s for the concen-
1
1
Ϫ3
tration of O2 a value of [ O ] = 0.29 ± 0.05 mol dm is
2
obtained. From this value and the initial concentration
Ϫ3
[CH CN](0) = 0.5 mol dm it follows that the efficiency η =
3
1
To prove the prediction of eqn. (6) we have carried out gas
volumetric measurements according to the procedure described
[ O ]/[CH CN](0) amounts to about 0.58 ± 0.10. This value
2
3
is consistent with eqn. (6) and indicates that the oxygen evolved
21
by Rebek et al. In the experiments at pH 9.9 concentrations of
in the reaction between acetonitrile and hydrogen peroxide is
Ϫ3
Ϫ3
1
[
CH CN](0) = 0.94 mol dm and of [H O ](0) = 9.4 mol dm
quantitatively generated as O .
3
2
2
2
1
were employed. Since the velocity of oxygen liberation was very
low at T = 30 ЊC the experiments were performed at T = 60 ЊC.
Under these conditions reproducible values were obtained. In
It should be mentioned that the amount of O estimated
2
from the area under the I (t) curve of Fig. 2 observed at
p
1
T = 30 ЊC is smaller than the amount of O expected according
2
1
290
J. Chem. Soc., Perkin Trans. 2, 2002, 1288–1295

View Article Online
Ϫ3
Fig. 3 I (t) as a function of time for the reaction between acetonitrile and hydrogen peroxide. Conditions: [CH CN](0) = 5.0 mol dm , [H O ](0) =
p
3
2
2
Ϫ3
Ϫ4
Ϫ3
Ϫ1/2
0
.5 mol dm , [Dequest] = 1.0 × 10 mol dm , pH = 12.0 and T = 30 ЊC. The inset shows the corresponding plot of (I (t)/mV)
versus time;
p
Ϫ4 Ϫ1
(
sl = 2.83 × 10
s ; int = 0.69).
In this mechanism the bimolecular reaction of PAIA with
Ϫ
the corresponding peroxy acid anion, PAIA , is assumed to be
the rate-determining step [eqn. (8)]. Furthermore the per-
hydration equilibrium (7) is postulated to occur very fast.
Ϫ
Of course the HOO is a better nucleophile than the
undissociated H O , especially at pH = pK = 11.7, so that it is
2
2
a
considerably more effective than the latter in carrying out the
Ϫ
addition step to acetonitrile forming PAIA . However, this has
1
Ϫ
no influence on the rate of O formation, because HOO is
2
also not involved in the rate-determining step.
On the basis of the result that at a given pH the values of the
experimental rate constants, kexp, obtained at different concen-
trations of hydrogen peroxide agree very well within the error
limits (vide infra), it can be assumed that equilibrium (7) is
practically shifted to [PAIA]. In this case eqn. (9) can be derived
for the formation of singlet oxygen:
Fig. 4 I (t) as a function of time for the reaction between acetonitrile
p
Ϫ3
and hydrogen peroxide. Conditions: [CH CN](0) = 0.5 mol dm
,
3
Ϫ3
Ϫ4
Ϫ3
[
H O ](0) = 5.0 mol dm , [Dequest] = 1.0 × 10 mol dm , pH = 9.9
2 2
17000
and T = 60 ЊC. 0͐
Ip(t) dt = 18914 mV s (±4%).
(
9)
to eqn. (6). In this connection it must be noted that in aqueous
1
solution O can only be detected by our IRL spectrometer, if
2
1
1
Ϫ6
where kexp is defined by eqn. (10)
the rate of O formation amounts to d[ O ]/dt ≥ 3.0 × 10 mol
s . Thus it seems likely that at T = 30 ЊC after a longer
reaction time the rate of O formation becomes smaller than
2
2
Ϫ3 Ϫ1
dm
1
kexp = k
8
F
H
(10)
2
Ϫ6
Ϫ3 Ϫ1
3
.0 × 10 mol dm
s
although the reaction is not completed.
k8 denotes the second-order rate constant of the rate-
determining step. The factor F is a constant at a given pH and
is defined by eqn. (11)
H
A.3 The mechanism of the Radziszewski reaction between
CH CN and H O . On the basis of our results presented so far
3
2
2
for the reaction between acetonitrile and hydrogen peroxide the
mechanism given in Scheme 2 can be deduced.
(
11)
where K (PAIA) represents the dissociation constant of PAIA.
a
1
FH describes the pH dependence of the rate of O formation.
Finally, [PAIA] = [PAIA] ϩ [PAIA ] denotes the total concen-
2
Ϫ
T
tration of peroxyacetimidic acid. For [PAIA]T the following
relations should be valid:
i) for [CH CN](0) = [H O ](0): [PAIA] (0) = [CH CN](0) or
3
2
2
T
3
[
PAIA] (0) = [H O ](0)
T
2
2
ii) for [CH CN](0) ӷ [H O ](0): [PAIA] (0) = [H O ](0)
3
2
2
T
2
2
iii) for [H O ](0) ӷ [CH CN](0): [PAIA] (0) = [CH CN](0).
2
2
3
T
3
On the basis of these assumptions eqn. (12) can be derived
1
8f
for the O phosphorescence intensity I (t):
2
p
(
12)
Scheme 2
J. Chem. Soc., Perkin Trans. 2, 2002, 1288–1295
1291

View Article Online
a
Table 1 Second-order rate constant, kexp , for the decomposition of
which are generally by a factor of about 4 higher than the
peroxyacetimidic acid in the hydrogen peroxide–acetonitrile system as a
1
corresponding values of Table 1. In addition the O phosphor-
2
function of pH
escence intensity I (0) is generally found to be higher at a given
p
b
3
Ϫ1 Ϫ1
c
3
Ϫ1 Ϫ1
pH than that observed for the system with a tenfold excess of
pH
kexp /dm mol
s
kexp /dm mol
s
Ϫ3
hydrogen peroxide ([H O ] = 5.0 mol dm ). Obviously solvent
2
2
Ϫ5
Ϫ5
effects give rise to both results. Water solution containing
5.0 mol dm of acetonitrile represents a 26% acetonitrile solu-
9
9
9
0.5
1.0
2.0
.1
.3
.9
2.2 × 10
2.2 × 10
5.2 × 10
2.6 × 10
2.9 × 10
1.4 × 10
3.7 × 10
3.6 × 10
7.8 × 10
1.5 × 10
2.1 × 10
1.3 × 10
Ϫ3
Ϫ5
Ϫ5
Ϫ5
Ϫ5
tion. In such a solution both the constant k and the lifetime
p
Ϫ4
Ϫ4
1
1
1
τ increase compared to water. With the values of k (water) =
∆
p
Ϫ4
Ϫ4
Ϫ1
Ϫ1 15
0
.21 s , k (acetonitrile) = 0.44 s
,
τ∆(water) = 3.1 µs and
p
Ϫ4
Ϫ4
15
τ∆(acetonitrile) = 75 µs valid at T = 25 ЊC and the molarities
a
b
Ϫ3 c
T = 30 ЊC. [CH CN](0) = [H O ](0) = 0.94 mol dm
.5 mol dm , [H O ](0) = 5.0 mol dm
.
[CH CN](0) =
of water and acetonitrile, respectively, the values of k (26%
3
2
2
3
p
Ϫ3
Ϫ3
0
.
Ϫ1
2
2
CH CN) = 0.27 s and τ (26% CH CN) = 4.1 µs can be esti-
mated. The increase of kp and τ∆ is accompanied by an
3
∆
3
15
increase of I (0) according to eqn. (4). Furthermore in a 26%
Conversion of eqn. (12) yields eqn. (13) which describes the
p
Ϫ1/2
acetonitrile solution most probably the solvation, especially of
observation that the plot of (I (t)/mV)
versus time results
p
Ϫ
PAIA -anion, is not as strong as in water and consequently the
in a straight line (see Figs. 1–3). From the slope (sl) and
the intercept (int) the second-order rate constant kexp can be
estimated.
Ϫ
nucleophilic attack of the PAIA -anion either at the carbon
atom of the >C᎐NH group of PAIA or at the outer peroxide
᎐
oxygen of PAIA occurs faster than in pure water.
Inspection of Table 1 reveals that the values of kexp deter-
mined under different conditions agree very well within the
error as mentioned above. In accordance with the mechanism
postulated in Scheme 2 the values of kexp run through a maxi-
mum with increasing pH. The maximum of the pH–rate
(
13)
Alternatively, it is also conceivable that equilibrium (7) is
shifted to the side of the educts. It was pointed out by Sauer
and Edwards that the constant of formation of 2-hydroxy-2-
hydroperoxypropane in the primary addition reaction between
profile (Fig. 5) is reached at about pH 11 indicating that the pKa
22
hydrogen peroxide and acetone was calculated to be K = 6.4 ×
f
Ϫ2
3
Ϫ1
1
0
dm mol at T = 32 ЊC. However, if the formation con-
2 2
H O
stant K7 of PAIA were of similar magnitude then Kf
-
1
(
acetone), the rate for the formation of O , would show a
2
second-order dependence in the concentration of acetonitrile
and hydrogen peroxide. In this case e.g. for systems with a large
1
excess of hydrogen peroxide for the rate of O formation the
2
pseudo-second-order rate law (14) should be valid:
(
14)
where kЈexp is defined by eqn. (15):
Fig. 5 Calculated rate constants, k , represented by the curve
Ϫ12
calc
Ϫ3
Ϫ3
Ϫ1 Ϫ1
(
estimated with k = 1.2 × 10 dm mol
s
and K (PAIA) = 8 × 10
a
2
7
2
8
kЈexp = K
k F [H O ] (0)
(15)
Ϫ3
8
H
2
2
mol dm ) and the experimental rate constants, kexp, as a function
of pH. The circles represent the kexp-values of the second column of
Table 1, the squares denote the kexp-values of the third column.
According to eqn. (15) the pseudo-second-order rate con-
stant kЈexp should be proportional to the square of the concen-
tration of hydrogen peroxide. Thus at a given pH and a given
value of PAIA should amount to about 11. The experimental
1
concentration of acetonitrile the rate of O formation and as a
data (kexp) can be well reproduced by assuming K (PAIA) =
2
a
Ϫ12
Ϫ3
Ϫ3
3
Ϫ1 Ϫ1
consequence the value of the phosphorescence intensity, I (0),
8 × 10
mol dm and k = 1.2 × 10 dm mol
s
for
p
8
should considerably increase with increasing excess of hydrogen
peroxide. However, this is not observed. Our results are only
consistent with the assumption that equilibrium (7) is practic-
the second-order rate constant for the self-decomposition of
peroxyacetimidic acid (see Fig. 5).
It is obvious that the rate constant k (PAIA) for the hetero-
8
ally shifted to PAIA, i.e. K must be several orders of magnitude
higher than Kf (acetone).
lytic self-decomposition of PAIA is of the same order as the
corresponding rate constant observed for the decomposition of
7
2 2
H O
23
peroxyacetic acid (PAA) by Evans and Upton [k (PAA) =
8
Ϫ3
3
Ϫ1
Ϫ1
A.4 Second-order rate constant kexp as a function of pH. The
values of kexp were determined with either equimolar quantities
of hydrogen peroxide and acetonitrile or with a tenfold excess
of hydrogen peroxide at T = 30 ЊC in the pH range 9.1 < pH <
2.0 × 10 dm mol
s ] as well as the rate-constant of
19a Ϫ3 3
PAA determind by Koubek et al. [k (PAA) = 5.4 × 10 dm
8
Ϫ1 Ϫ1
mol
s ].
Although there is a remarkably similar geometry of peroxy-
24
1
2.0 and are summarized in Table 1.
carboxylic acids and peroxycarboximidic acids we conclude
that PAA is the more reactive epoxidizing agent as compared to
PAIA. However, PAIA has a significantly higher pKa value
The values displayed are mean values calculated from at least
three different measurements. The error of kexp amounts to
about 25%. The considerable error results mainly from the fact
(11.1) than PAA [pK (PAA) = 8.2]. Thus PAIA can be used as
a
1
that the O phosphorescence intensity detected under the used
an epoxidation agent up to about pH 8.5, whereas using PAA at
pH ∼ 8.5 the self-decomposition of PAA strongly competes
with the epoxidation reactions.
2
conditions is relatively low (see the signal/noise ratio of the I (t)
p
curves in Figs. 1 and 2).
Some experiments were also performed with systems contain-
The pK (PAIA) value of peroxyacetimidic acid is compar-
a
Ϫ3
25
ing a tenfold excess of acetonitrile ([CH CN] = 5.0 mol dm ).
able with the pKa values of hydroperoxides, e.g. of H O2
3
2
For these systems values of kexp were observed at a given pH
(pK = 11.75), of CH OOH (pK = 11.08) and of C H OOH
a
3
a
2
5
1
292
J. Chem. Soc., Perkin Trans. 2, 2002, 1288–1295

View Article Online
(
pK = 11.05). But, most probably within the peroxycarbo-
where kobs is given by eqn. (21):
kobs = K K K k FЈ [ketone](0)
a
ximidic acids PAIA is no exception. From bleaching experi-
ments on tea using PAA generated by the reaction of
tetraacetylethylenediamine with hydrogen peroxide it is well
known that the bleaching power runs through a maximum with
increasing pH. The maximum bleaching effect is found at
pH = pK (PAA) = 8.2. Recently Reinhardt and Löffler
have investigated the bleaching of tea stain as a function of pH
using the N,N-dimethylmorpholiniumacetonitrile–H O system
and the cyanopyrrolidine–H O system. For the first system the
(21)
7
16 17 18
H
K , K and K17 are the constants of the equilibria (7), (16)
7
16
26
and (17) and k18 denotes the first-order rate constant of
the rate-determining step for the formation of the reactive
dioxirane III from the intermediate II. The factor FЈH is a
constant at a given pH and is defined by eqn. (22).
27
a
2
2
2
2
maximum bleaching effect was found to be at pH 9.5 and for
the second system at pH 11.0, respectively.
(22)
B. Investigation of the CH CN–H O sytem in water in the
presence of ketones
3
2
2
Here KW is the ionic product of water. Finally, [ketone](0)
represents the initial concentration of the ketone at starting
time.
As discussed in section A the decomposition of PAIA occurs
via a mechanism generally observed for peroxy acids. Thus by
analogy with the ketone-catalysed decomposition of peroxy
Interestingly, eqn. (20) is consistent with the observation
2
made by Wiberg for the reaction between benzonitrile deriv-
8,9
atives and hydrogen peroxide in 25–50% acetone solutions.
According to eqn. (20) the rate of consumption of hydrogen
peroxide should show a first-order dependence in [R–CN] and
acids the mechanism presented in Scheme 3 can be postulated
Ϫ
in [H O ]. However, this rate does not show a first-order in OH
2
2
concentration, but is directly proportional to the factor FЈ , as
H
demonstrated for the ketone-catalysed decomposition of
8f
Caro’s acid.
To test whether ketones catalysed the reaction of aceto-
nitrile with hydrogen peroxide we have performed an
ϩ
Ϫ
experiment with DMOP NO3 . Taking into account the
ϩ
Ϫ
electron-withdrawing ammonium group of DMOP NO3 it is
one of the most reactive catalysts in the ketone-catalysed
decomposition of peroxy acids and has been shown to
display a dramatic acceleration of peroxy acid decompos-
8a,8f,8g,11
ition.
Inspection of the I (t) curve presented in Fig. 6 clearly
p
Scheme 3
also for the ketone-catalysed decomposition of PAIA and in
general for peroxycarboximidic acids. For simplification the
hydration equilibrium between the ketone and its gem H O-
2
Fig. 6 I (t) as a function of time for the reaction between acetonitrile
8g
p
adduct
and additionally the perhydration equilibrium
ϩ
Ϫ
and hydrogen peroxide in the presence of DMOP NO3 (curve a) and
ϩ
Ϫ
between the ketone and its gem H O -adduct were not taken
2
2
in the absence of DMOP NO
3
(curve b). Conditions: [CH CN](0) =
3
Ϫ4 Ϫ3
Ϫ3
into account. Furthermore, the dissociation equilibria of the
different adducts were neglected (vide infra).
[H O ](0) = 0.94 mol dm ; [Dequest] = 1.0 × 10 mol dm , pH = 9.5,
2 2
ϩ Ϫ Ϫ2 Ϫ3
T = 30 ЊC and [DMOP NO
] = 5.0 × 10 mol dm .
3
Analogous to the peroxy acids, PAIA is assumed to react very
fast with the ketone to form the adduct I, i.e. the successive
reaction of PAIA is not the rate-determining step as observed
for the uncatalysed reaction of acetonitrile with hydrogen per-
ϩ
Ϫ
demonstrates the influence of DMOP NO3 in the reaction of
acetonitrile with hydrogen peroxide. As shown in curve a in
comparison with curve b, the addition of a small amount
ϩ
Ϫ
of DMOP NO3 to a system containing equimolar quantities
of the educts leads to a considerable increase of I (t). This may
be explained as a consequence of an enhancement of the
decomposition of PAIA.
oxide. According to Scheme 3 eqn. (20) can be derived for the
1
rate of O formation.
p
2
According to eqn. (20) one can expect that the difference
∆
I (t) = I (t) [curve a] Ϫ I (t) [curve b] can be described by a
p p p
Ϫ1/2
(
20)
second-order plot (∆I (t)/mV)
versus time. However, such a
p
plot does not yield a straight line over the whole reaction time
J. Chem. Soc., Perkin Trans. 2, 2002, 1288–1295
1293

View Article Online
Fig. 7 I (t) as a function of time for the reaction between acetonitrile and hydrogen peroxide in 50% acetone solution. Conditions: [CH CN](0) =
p
3
Ϫ3
Ϫ3
Ϫ4
Ϫ3
0
.1 mol dm , [H O ](0) = 1.0 mol dm , [Dequest] = 1.0 × 10 mol dm , pH = 8.2 and T = 60 ЊC. The inset shows the corresponding plot of ln (I (t)/
mV) versus time; (sl = 1.4 × 10
2
2
p
Ϫ4 Ϫ1
s
).
ϩ
Ϫ
indicating that the DMOP NO3 -catalysed decomposition of
PAIA is more complex than the mechanism given in Scheme 3
For kobs a value of k = K K K17 k18 FЈ [acetone](0) =
obs
Ϫ1 Ϫ1
7
16
H
Ϫ5
3
7.0 × 10 dm mol
10 mol dm at pH 8.2 and with [acetone](0) = 6.8 mol dm
s
can be estimated. With FЈH = 1.7 ×
Ϫ3
ϩ
Ϫ
Ϫ6
3
predicts. Obviously for DMOP NO3 the hydration equi-
librium and the perhydration equilibrium, respectively, between
DMOP NO3 and the corresponding adducts must be taken
into account.
the fourth-order rate constant, k , for the formation of
Di
ϩ
Ϫ
dimethyldioxirane is calculated to be k = K K K17 k18 =
Di 7 16
3 Ϫ1 3 Ϫ1
6.1 (dm mol ) s
.
To further illustrate the dioxirane formation in the
acetonitrile–hydrogen peroxide–ketone systemϩ we Ϫ used
acetone as catalyst. In contrast to DMOP NO3 the
simplified mechanism should be valid using acetone as catalyst,
The formation of a dioxirane in the acetonitrile–hydrogen
peroxide–ketone system is in agreement with the recent study of
Shu and Shi who demonstrated the efficient epoxidation of
6
olefins using CH CN–H O and different ketones as catalyst.
3
2
2
2
H O
since the formation constant Kf (acetone) for the H O-
It was found that trifluoroacetone (CF COCH ) as catalyst
2
3 3
2 2
H O
adduct and the formation constant Kf (acetone) for the
H O -adduct are sufficiently small that the corresponding equi-
gave the highest conversion in a 1 : 1 mixture of CH CN and
3
aqueous EDTA solution using trans-β-methylstyrene as sub-
strate, whereas in the absence of ketone catalyst, the epoxid-
ation was minimal under the same conditions. Interestingly,
upon the determination of the optimal reaction pH, the best
2
2
libria can be neglected. However, high concentrations of acet-
one are necessary to demonstrate the influence of acetone as
catalyst in the hydrogen peroxide–acetonitrile system because
acetone is by a factor of about 1000 (pH ≤ 9) less efficient as
epoxide conversion was found to be at pH around 11, i.e. at
ϩ
Ϫ 8f,8g
28
catalyst than DMOP NO3
uncatalysed reaction we have performed measurements at pH
.2 and T = 60 ЊC with a tenfold excess of hydrogen peroxide
.
To avoid a contribution of the
the pK value of PAIA. This result seems to be no excep-
a
6
tion. Shu and Shi have also investigated the epoxidation of
8
trans-β-methylstyrene as substrate using the Caro’s acid–
Ϫ3
1
(
[H O ] = 1.0 mol dm ) in 50% acetone. Accordingly, no O2
CF COCH3 system as a function of pH. Obviously the pH
2
2
3
was detected under the same conditions only with the hydrogen
peroxide–acetonitrile system itself in the absence of acetone.
Fig. 7 shows that eqn. (20) describes the acetone-catalysed
decomposition of PAIA. If a large excess of hydrogen peroxide
has a large effect on the epoxidation and a maximum conver-
sion of the substrate was observed at pH 10, i.e. also around
Ϫ
the pKa2 value of the monoanion of Caro’s acid (HSO5
with pKa2 = 9.4.
)
is used eqn. (20) yields for the time dependence of I (t)
eqn. (23).
For a more detailed study of the ketone-catalysed decom-
position of peroxycarboximidic acids (PCIA) a water soluble
nitrile–H O system should be used generating a more reac-
p
2
2
Ϫ2kЈobs
t
tive PCIA than PAIA. For example, with a nitrile quat like
29
I (t) = I (0)e
(23)
(24)
(25)
p
p
N-methylmorpholiniumacetonitrile it should be possible to
ϩ
measure the ketone-catalysed decomposition of the PMIA -
where I (0) is defined by eqn. (24)
p
cation with relatively small concentrations of ketones like acet-
one, cyclohexanone and fluoroacetone, i.e. with concentrations
of the ketones which are not accompanied by a remarkable
change in the properties of the solvent.
Ip(0) = CЈkЈobs [CH CN](0)
3
Here kЈobs is given by eqn. (25)
kЈobs = kobs [H O ](0)
2
2
Conclusions
and CЈ represents the product of ck (50% acetone)τ (50%
acetone).
Our results indicate that peroxyacetimidic acid (PAIA) does not
p
∆
1
react with hydrogen peroxide to form acetamide and O . This is
2
According to eqn. (23) the pseudo-first-order plot of ln (I (t)/
mV) versus time resulted in a straight line up to 14000 s as
true for both the uncatalysed and the ketone-catalysed decom-
position of PAIA. The observation that even the reaction of
p
Ϫ
shown in the inset of Fig. 7. From the slope of the straight
PAIA with the HOO anion can obviously not compete with
Ϫ4 Ϫ1
Ϫ
line 2kЈobs is determined to be 1.4 × 10
s
. Consequently
the reaction of PAIA and the PAIA -anion indicates that
Ϫ
Ϫ
the pseudo-first-order plot is linear through three half-lives.
PAIA is more nucleophilic than HOO .
1
294 J. Chem. Soc., Perkin Trans. 2, 2002, 1288–1295

View Article Online
2 R. Schmidt and H.-D. Brauer, J. Am. Chem. Soc., 1987, 109, 6976.
3 R. Schmidt, Chem. Phys. Lett., 1988, 151, 369.
1
1
Acknowledgements
Financial support by the Deutsche Forschungsgemeinschaft
is gratefully acknowledged. Special thanks are expressed to
M. Hild for a critical review.
14 K. Böhme and H.-D. Brauer, Inorg. Chem., 1992, 31, 3468.
1
1
1
5 R. Schmidt, F. Shafii and M. Hild, J. Phys. Chem. A, 1999, 103,
599.
2
6 S. Y. Egorov, V. F. Kamalov, N. I. Koroteev, A. A. Krasnowsky,
B. N. Tolentaev and S. V. Zinnkov, Chem. Phys. Lett., 1989, 163, 421.
7 (a) E. McKeown and W. A. Waters, Nature, 1964, 203, 1063;
References
(
b) E. McKeown and W. A. Waters, J. Chem. Soc. (B), 1966, 1040.
1
B. Radziszewski, Ber. Dtsch. Chem. Ges., 1885, 18, 355 . The often
cited publication in Ber. Dtsch. Chem. Ges., 1884, 17, 1289 describes
the oxidation of imidazole.
K. B. Wiberg, J. Am. Chem. Soc., 1953, 75, 3961.
(a) G. B. Payne, P. H. Deming and P. H. Williams, J. Org. Chem.,
18 A. Lange, PhD Dissertation, Frankfurt, 1997.
19 (a) E. Koubek, M. L. Haggett, C. J. Battaglia, K. M. Ibne-Rasa,
H. Y. Pyun and J. O. Edwards, J. Am. Chem. Soc., 1963, 85, 2263;
(b) E. Koubek, G. Levey and J. O. Edwards, Inorg. Chem., 1964, 3,
1331; (c) D. L. Ball, J. O. Edwards, M. L. Haggett and P. Jones,
J. Am. Chem. Soc., 1967, 89, 2331; (d ) E. Koubek and J. E. Welsch,
J. Org. Chem., 1968, 33, 445.
2
3
1
1
961, 26, 659; (b) G. B. Payne and P. H. Williams, J. Org. Chem.,
961, 26, 251; (c) G. B. Payne, Tetrahedron, 1962, 18, 763.
4
5
(a) R. D. Bach and J. W. Knight, Org. Synth., 1981, 60, 63;
b) L. A. Arias, S. Adkins, C. J. Nagel and R. D. Bach, J. Org. Chem.,
983, 48, 888.
H.-W. Hennig, G. Prescher and R. Siegmeier, in Methoden der
Organischen Chemie (Houben-Weyl), ed. H. Kropf, Georg Thieme
Verlag, Stuttgart, 1988, Vol. E 13/1.
20 J. E. McIsaac, R. E. Ball and E. J. Behrman, J. Org. Chem., 1971, 36,
3048.
21 J. Rebek, R. McCready, S. Wolf and A. Mossman, J. Org. Chem.,
1979, 44, 1485.
22 M. C. V. Sauer and J. O. Edwards, J. Phys. Chem., 1971, 75, 3004.
23 D. F. Evans and M. W. Upton, J. Chem. Soc., Dalton Trans., 1985,
1151.
24 T. J. Lang, G. J. Wolber and R. D. Bach, J. Am. Chem. Soc., 1981,
103, 3275.
25 R. Curci and J. O. Edwards, in Catalytic Oxidations with Hydrogen
Peroxide as Oxidant, ed. G. Strukul, Kluwer Academic Publishers,
Dordrecht, 1992, ch. 3.
26 G. Reinhardt and W. Schuler, Offizielle Kongress-Schrift der
SEPAWA, Vereinigung der Seifen-, Parfüm- und Waschmittel-
fachleute, e.V., Bad Dürkheim, 1989.
(
1
6
7
8
L. Shu and Y. Shi, J. Org. Chem., 2000, 65, 8807.
Y. Sawaki and Y. Ogata, Bull. Chem. Soc. Jpn., 1981, 54, 793.
(a) R. E. Montgomery, J. Am. Chem. Soc., 1974, 96, 7820; (b) W.
Adam, R. Curci and J. O. Edwards, Acc. Chem. Res., 1989, 22, 205;
(
c) R. W. Murray, Chem. Rev., 1989, 89, 1187; (d ) R. Curci, in
Advances in Oxygenated Processes, ed. A. L. Baumstark, JAI Press,
Greenwich, CT, 1990, vol. 2, ch. 1; (e) W. Adam, L. P.
Hadjiarapoglou, R. Curci and R. Mello, in Organic Peroxides,
ed. W. Ando, J. Wiley & Sons, New York, 1992, ch. 4; ( f ) A. Lange
and H.-D. Brauer, J. Chem. Soc., Perkin Trans. 2, 1996, 805;
27 G. Reinhardt and M. Löffler, Tenside, Surfactants, Deterg., 1997, 34,
(
g) A. Lange, M. Hild and H.-D. Brauer, J. Chem. Soc.,
404.
Perkin Trans.2, 1999, 1343.
28 It is worthy of note here that the uncatalysed decomposition of
9
A. Lange and H.-D. Brauer, in Peroxide Chemistry, Mechanistic and
Preparative Aspects of Oxygen Transfer . Research Report of the
Deutsche Forschungsgemeinschaft, ed. W. Adam, Wiley-VCH
Verlag, Weinheim, 2000, section B, ch. 7.
0 H. M. E. Cardwell and F. J. McQuillin, J. Chem. Soc., 1949, 708.
1 S. E. Denmark, D. C. Forbes, D. S. Hays, J. S. DePue and
R. G. Wilde, J. Org. Chem., 1995, 60, 1391.
peroxyacetimic acid with maximum rate at pH = pK = 11.1 forming
a
quantitatively singlet oxygen can be neglected in the acetonitrile–
hydrogen peroxide system catalysed by trifluoroacetone due to the
slow rate constant of the uncatalysed self-decomposition.
29 The nitrile quat with hydrogensulfate as counterion was until
recently commercially available under the trade name Sokalan® BM
G from BASF AG.
1
1
J. Chem. Soc., Perkin Trans. 2, 2002, 1288–1295
1295