3
Ϫ1 Ϫ1
Table 1 Rate constants (in units of dm mol
s ) of ozone with some
Results and discussion
cinnamic acids
Rate constants
Substrate
Acid
Anion
Reference
The cinnamic acids under investigation have pK values
a
9,19
5
6
between 4.1 and 4.6.
The free acids, especially 4-nitro-
Cinnamic acid
1 × 10 1.2 × 10 Ref. 9
5
4
5
× 10 3.8 × 10 This work
cinnamic acid, are poorly soluble in water, and for this reason
only the reactions of their anions have been investigated in
some detail. The rate constants for the reactions of ozone with
5
5
4
4
3
3
-Methoxycinnamic acid
-Nitrocinnamic acid
1.3 × 10 6.8 × 10 This work
5
—
1.2 × 10 This work
6
6
-Methoxy-4-hydroxycinnamic acid 1.1 × 10 7.9 × 10 Ref. 9
17
the cinnamate ions were determined by competition with
6
7
,4-Dihydroxycinnamic acid
2 × 10 1.2 × 10 Ref. 9
buten-3-ol by following the formaldehyde yield [cf. reaction (6)]
as a function of the [cinnamic acid]/[buten-3-ol] ratio (cf. ref.
1
7). The rate constant for the reaction of ozone with buten-3-ol
8
,10,11
ؒ
earlier studies on ozonation reactions
and on the OH-
4 3 Ϫ1 Ϫ1 2
is 7.9 × 10 dm mol s .
12
induced degradation of EDTA, where the involvement of this
short-lived intermediate was postulated.
CH ᎐CH–CH(OH)–CH + O →
2
3
3
CH O + CH –CH(OH)–CHO + H O (6)
2
3
2
2
Experimental
Cinnamic acid (>99%, Merck), 4-methoxycinnamic acid
The rate constants thus obtained are compiled in Table 1.
As expected, the electron-donating methoxy group in a para-
position enhances the rate of reaction while the electron-
withdrawing nitro group reduces it. From these three values it is
estimated that the substituent effect on the rate of reaction, as
expressed by the ρ value, is an order of magnitude lower than in
(
≥98%, Fluka) and 4-nitrocinnamic acid (≥97%, Fluka),
glyoxylic acid monohydrate sodium salt (>99%, Fluka) and
fumaric acid (disodium salt, 98%, Aldrich) were used as
received. Benzaldehyde and its 4-methoxy and 4-nitro deriva-
tives, required as reference materials, were also commercially
available. Benzaldehyde contained some benzoic acid and was
redistilled. Catalase (from beef liver) was obtained from
Boehringer Mannheim.
Solutions were made-up in Milli-Q-filtered (Millipore) water.
Ozone stock solutions were prepared by bubbling ozone from
a dioxygen-fed ozonator (Wedeco SWO-70 or Philaqua Philoz
the case of the ozone reaction with para-substituted benzenes
1
(
ρ = Ϫ3.1). This clearly shows that the para-substituent at the
aromatic ring has only a small effect on the reactivity of the
conjugated double bond. In contrast, there can be quite a
marked influence of the substituent on the fate of the Criegee
intermediate (see below).
0
4) for some minutes through water, and its ozone content
In the reactions of ozone, the reactivity of anions is generally
13,14
was determined spectrophotometrically taking ε(260 nm)
=
higher than that of the conjugate acids, due to a higher electron
3
Ϫ1
Ϫ1
3
300 dm mol cm . Experiments were carried out at room
15
density in the C᎐C double bond. This is also observed for the
᎐
temperature (20–21 ЊC) and in the absence of buffer. In
most experiments, the cinnamic acid was in large excess over
the added ozone concentration (cinnamic acid consumption
present system (cf. Table 1).
It is noticed that there is a substantial difference between the
rate constants determined by competition (this study) and those
by continuous flow using indigotrisulfonic acid as a detector
for residual ozone (ref. 9). These differences (a factor of 2–3)
are larger than the confidence limits of these methods. In an
attempt to solve this discrepancy, we have redetermined the rate
constant of the reaction of ozone with buten-3-ol, but our more
<
15%). Due to the high rate constant of the reaction of
cinnamic acid with ozone (cf. Table 1) compared to those of the
3
Ϫ1 Ϫ1 15
products benzaldehyde (k = 2.5 dm mol s ) and glyoxylate
3
Ϫ1 Ϫ1 15
ion (k = 1.9 dm mol s ), the consumption of cinnamic acid
could also be determined at high conversion.
The concentrations of benzaldehydes were determined by
HPLC (Merck-Hitachi with diode array detection) on a 25 cm
RP18 reversed phase column [eluent: water containing 40%
methanol and 0.1% phosphoric acid; retention times (min):
cinnamic acid (19), 4-methoxycinnamic acid (22), 4-nitro-
cinnamic acid (16), benzaldehyde (8), 4-methoxybenzaldehyde
4
3
Ϫ1 Ϫ1
recent value (k = 9.1 × 10 dm mol s ) does not materially
alter the situation. Hence, we do not have an explanation for
this discrepancy.
Products
(
10), 4-nitrobenzaldehyde (7.5)]. The benzoic acids elute
With all of the cinnamic acids investigated, one mol of
cinnamic acid is consumed per mol ozone (data not shown),
and in the case of cinnamic acid one mol of benzaldehyde and
one mol of glyoxylic acid are formed (Fig. 1 and Table 2). This
also holds for 4-methoxycinnamic acid (data not shown), while
in the case of 4-nitrocinnamic acid some formic acid instead
sufficiently distant from the benzaldehydes, and low concen-
trations of these potential by-products can be detected without
interference. Glyoxylic and formic acids were determined by ion
chromatography [Dionex DX 100; column: AS9 HC; eluent:
Ϫ2
Ϫ3
1
× 10 mol dm NaHCO ; retention times (min): glyoxylic
3
acid (10.5), formic acid (11.5)]. Hydrogen peroxide was deter-
mined with molybdate-activated iodide.
Since H O reacts with glyoxylic acid giving rise to formic
acid and carbon dioxide (see below), catalase was added (10 µl
per 10 ml sample) right after ozonolysis to destroy H O . Under
our experimental conditions, its destruction was complete
within less than 3 s (absence of any residual H O ).
of glyoxylic acid (plus CO , not determined here) is formed;
2
16
the sum of the two acids, however, again corresponds to the
4-nitrobenzaldehyde yield (Fig. 1). In this context it is worth
mentioning that also in the case of 4-hydroxycinnamic acid
(p-cumaric acid), 4-hydroxybenzaldehyde and glyoxylic acid are
2
2
2
2
4
the major products.
2
2
17
In the competition between the cinnamic acids and buten-
-ol for ozone, formaldehyde is the measured product which
Formation and decay of 2-hydroperoxy-2-hydroxyacetic acid
3
18
was determined with the help of the Hantzsch reaction. The
error in these determinations was typically less than ±10%.
NMR data were taken on a Bruker DRX 400 instrument
and UV spectra on a Perkin Elmer Lambda 16 spectro-
photometer. The rate constant for the reaction of buten-3-ol
was redetermined by the stopped-flow (Biologic SFM3) tech-
nique by following the decay of the ozone absorption at 260 nm
as a function of time.
When an ozonated cinnamate solution (pH ∼6.5) is kept
for some time, glyoxylic acid is progressively converted into
formic acid due to the presence of H O which is thereby con-
2
2
sumed (data not shown, for the kinetics of this reaction see
Fig. 2).
Hydrogen peroxide and glyoxylic acid [in aqueous solution
mainly present as its hydrate, cf. Scheme 2, equilibrium (7), for
details see below] give rise to 2-hydroperoxy-2-hydroxyacetic
7
94
J. Chem. Soc., Perkin Trans. 2, 2001, 793–797