Singlet dioxygen formation in ozone reactions in aqueous solution

Article abstract of DOI:10.1039/b101230o

In ozone reactions, singlet dioxygen [O₂(¹Δ_g)] is formed when ozone reacts by O-atom transfer. O₂(¹Δ_g) yields were determined for more than 50 compounds using as reference the reaction of hydrogen peroxide with hypochlorite. Close to 100percent yields were found in the reaction of O₃ with sulfides, disulfides, methanesulfinic acid, and nitrite. In accordance with this, the only products are: methionine sulfoxide, methanesulfonic acid, and nitrate for the reaction of O₃ with methionine, methanesulfinic acid, and nitrite, respectively. In the case of aliphatic tertiary amines (trimethylamine, triethylamine, and DABCO), the O₂(¹Δ_g) yields range between 70 and 90percent, the aminoxide being the other major product. With EDTA and nitrilotriacetic acid (NTA), the O₂(¹Δ_g) yield is around 20percent. The interpretation of the data with DABCO required the determination of the quenching constant of O₂(¹Δ_g) by this amine, k_q = 1.8 × 10⁹ dm³ mol^-1 s^-1 in H₂O and D₂O, two orders of magnitude lower than previously reported. Aromatic tertiary amines give even lower O₂(¹Δ_g) yields [N,N-dimethylaniline (7percent), N,N,N′,N′-tetramethylphenylenediamine (9percent)]. Substantial amounts of O₂(¹Δ_g) are also formed with the DNA model compounds, 2′-deoxyguanosine (40percent) and 2′-deoxyadenosine (15percent, in the presence of tert-butyl alcohol as ^·OH scavenger). The pyrimidine nucleobases only yield O₂(¹Δ_g) when deprotonated at N(1). O₂(¹Δ_g formation is also observed with hydrogen sulfide (15percent), azide (17percent), bromide (56percent), iodide (12percent), and cyanide ions (20percent). The O₂(¹Δ_g yield from the reaction of O₃ with phenols and phenolates is typically around 20percent, but may rise closer to 50percent in the case of pentachloro-and pentabromo-phenolate. Low O₂(¹Δ_g yields are found with unsaturated acids such as dihydroxyfumarate (6percent), muconate (2percent), and acetylenedicarboxylate (15percent). With saturated compounds, also, no O₂(¹Δ_g (e.g. with propan-2-ol, acetaldehyde, acetaldehyde dimethylacetal and glyoxal) or very little O₂(¹Δ_g (formic acid; 6percent, at high formate concentrations) was detected. As shown with some examples, knowledge of the O₂(¹Δ_g yield (in combination with that of other products) is a prerequisite for the elucidation of the mechanisms of O₃ reactions in aqueous solutions.

Full text of DOI:10.1039/b101230o

View Article Online / Journal Homepage / Table of Contents for this issue
Singlet dioxygen formation in ozone reactions in aqueous solution
Florinella Muñoz,† Eino Mvula,‡ Silvia E. Braslavsky and Clemens von Sonntag*
Max-Planck-Institut für Strahlenchemie, Stiftstraße 34–36, PO Box 101365, D-45470,
Mülheim an der Ruhr, Germany. E-mail: vonsonntag@mpi-muelheim.mpg.de;
Fax: ++49-(0)208-306-3951; Tel: ++49-(0)208-306-3529
2
Received (in Cambridge, UK) 7th February 2001, Accepted 18th May 2001
First published as an Advance Article on the web 12th June 2001
1
1
In ozone reactions, singlet dioxygen [O ( ∆ )] is formed when ozone reacts by O-atom transfer. O ( ∆ ) yields were
2
g
2
g
determined for more than 50 compounds using as reference the reaction of hydrogen peroxide with hypochlorite.
Close to 100% yields were found in the reaction of O with sulfides, disulfides, methanesulfinic acid, and nitrite. In
3
accordance with this, the only products are: methionine sulfoxide, methanesulfonic acid, and nitrate for the reaction
of O with methionine, methanesulfinic acid, and nitrite, respectively. In the case of aliphatic tertiary amines
3
1
(
trimethylamine, triethylamine, and DABCO), the O ( ∆ ) yields range between 70 and 90%, the aminoxide being the
other major product. With EDTA and nitrilotriacetic acid (NTA), the O ( ∆ ) yield is around 20%. The interpretation
of the data with DABCO required the determination of the quenching constant of O ( ∆ ) by this amine,
k = 1.8 × 10 dm mol
tertiary amines give even lower O ( ∆ ) yields [N,N-dimethylaniline (7%), N,N,N Ј,N Ј-tetramethylphenylenediamine
9%)]. Substantial amounts of O ( ∆ ) are also formed with the DNA model compounds, 2Ј-deoxyguanosine (40%)
and 2Ј-deoxyadenosine (15%, in the presence of tert-butyl alcohol as OH scavenger). The pyrimidine nucleobases
only yield O ( ∆ ) when deprotonated at N(1). O ( ∆ ) formation is also observed with hydrogen sulfide (15%), azide
2
g
1
2
g
1
2
g
9
3
Ϫ1 Ϫ1
s
in H O and D O, two orders of magnitude lower than previously reported. Aromatic
q
2 2
1
2
1
g
(
2 g
ؒ
1
1
2
g
2
g
1
(
17%), bromide (56%), iodide (12%), and cyanide ions (20%). The O ( ∆ ) yield from the reaction of O with phenols
2
g
3
and phenolates is typically around 20%, but may rise closer to 50% in the case of pentachloro- and pentabromo-
1
phenolate. Low O ( ∆ ) yields are found with unsaturated acids such as dihydroxyfumarate (6%), muconate (2%), and
2
g
1
acetylenedicarboxylate (15%). With saturated compounds, also, no O ( ∆ ) (e.g. with propan-2-ol, acetaldehyde,
2
g
1
acetaldehyde dimethylacetal and glyoxal) or very little O ( ∆ ) (formic acid; 6%, at high formate concentrations) was
2
g
1
detected. As shown with some examples, knowledge of the O ( ∆ ) yield (in combination with that of other products)
2
g
is a prerequisite for the elucidation of the mechanisms of O reactions in aqueous solutions.
3
1
It is well-known that in O reactions singlet dioxygen [O ( ∆ )]
3
2
g
1,2
may be formed in high yields. This is always the case when
O-transfer reactions dominate. Spin conservation rules then
demand that the resulting dioxygen molecule must be in
its (excited) singlet state, because educt and O as well as the
3
product are in their singlet (ground) states. A case in point is
the reaction of O with tertiary amines, whereby aminoxides
3
3–6
are generated [reaction (1)].
In water, the reactions of O may differ considerably from
3
those encountered in the gas phase and in organic solvents
(
cf. ref.7 and 8). Moreover, the high permittivity of water
may allow electron-transfer reactions to occur [e.g. reaction (2)]
which are not possible in an organic solvent due to the lower
solvation energies of the resulting radical ions.
1
ratio of the O ( ∆ ) yield and the O-transfer-product yield
2
g
may fall below unity [cf. reaction (4)]. Upon diffusing out
of the cage [reaction (5)], the ozonide radical can induce a
decomposition of the O via a chain reaction [reactions (6)–(9)].
In water, the energy gained upon solvation of the resulting
ions reduces the ionization potential by about 3 eV as compared
to the gas phase ionization potential, while in an organic
3
O3ؒϪ
2
ؒ
2
Ϫ
ϩ H O → OH ϩ O ϩ OH
(6)
(7)
(8)
(9)
9,10
solvent only 1 to 2 eV are gained. Since the electron affinity
11
of O is only 2.1 eV, such an electron-transfer reaction will,
3
ؒ
ؒ
OH ϩ O → HO2 ϩ O2
3
even in water, only be possible with molecules that have a
12
low ionization potential (a recent suggestion to revise the
HO2 → O2ؒ ϩ H pK (HO ) = 4.8
ؒ
Ϫ
ϩ
ؒ
2
a
currently accepted mechanism of O reactions has therefore
3
to be rejected). However, should electron transfer occur
ؒ
Ϫ
ؒ
O2 ϩ O → O ϩ O
3
Ϫ
[cf. reaction (2)], ensuing free-radical reactions would lead also
3
2
to the O-transfer product [e.g., reactions (3) and (4)], but the
In competition, the ensuing OH radicals formed in reaction (6)
also attack the substrate.
Ozone is widely used in drinking-water processing, but the O3
reactions in water are as yet not adequately understood. In
most studies, material balances are missing, and the formation
†
Present address: Escuela Politécnica Nacional, Instituto de Ciencias
Nucleares, Quito, Ecuador.
On leave from Department of Chemistry, University of Namibia,
‡
Windhoek, Namibia.
DOI: 10.1039/b101230o
J. Chem. Soc., Perkin Trans. 2, 2001, 1109–1116
This journal is © The Royal Society of Chemistry 2001
1109

View Article Online
1
Fig. 1 Schematic representation of the O ( ∆ ) detection system
2
g
1
of O ( ∆ ) may give important information, besides mechanistic
2
g
aspects, on the formation of potential products that have
escaped attention. For this reason, we attempted to quantify
1
Fig. 2 Signals of O ( ∆ ) emission during the reaction of H O with
2
g
2
2
Ϫ4
Ϫ3
1
HOCl (7.5 × 10 mol dm each). Inset: signals of O ( ∆ ) emission
2
g
1
Ϫ4
Ϫ3
Ϫ3
the O ( ∆ ) yields in a number of systems relevant to drinking-
in the reaction of O (1 × 10 mol dm ) with EDTA (2.5 × 10 mol
3
Ϫ3
2
g
water processes, notably EDTA and related complexing
dm ). Measurements in H O.
2
6,13
agents.
This study was then put on a wider footing, e.g.
1
with sulfur-containing compounds, phenols, inorganic ions,
and DNA constituents.
Table 1 Integrated emission of O ( ∆ ) in the reference system
2
g
(H
O
ϩ NaOCl)
2
2
1
In the determination of the O ( ∆ ) yield in aqueous
2
g
Run
Area
solutions, a number of problems may arise as has been pointed
1
out by Kanofsky and his group in their pioneering work.
1
2
3
4
5
6
7
8
9
0
0.3386
0.3420
0.3101
0.3352
0.3242
0.2875
0.2737
0.2965
Here, we will report on some additional complications, but
will show that, despite the difficulties, useful data may be
obtained.
Experimental
All chemicals were of the highest purity commercially available.
Ozone solutions were made-up in Milli-Q-purified (Millipore)
0.2898
0.3200
0.3118 ± 0.024
1
water using an O -fed ozonator (SWO-70, WEDECO). Ozone
X
2
concentrations were determined spectrophotometrically taking
3
Ϫ1
Ϫ1 14
ε(260 nm) = 3314 dm mol cm .
(dd; J = 13.3, 9.2 Hz; one 3-H, one 6-H), in agreement with
17
The set-up used in the present study consists of a liquid
N -cooled germanium photodiode (EO-817L, North Coast
reported data.
2
Scientific Corporation, rise time 5–10 ms) sensitive to the
1
1
270 nm phosphorescence emitted by O ( ∆ ). In front of the
2
g
Results and discussion
detector, a Suprasil quartz cell (1 × 1 cm), a bandpass filter
1270 ± 70 nm, Laser Components) and a silicium cut-off filter
< 1050 nm, Laser Components) were placed at a distance
of one cm (see Fig. 1). The signal was electrically triggered
simultaneously with the injection of the substrate solution,
recorded by a storage oscilloscope (Tektronix 2221A), and the
data were evaluated with the help of laboratory-produced
software.
1
Determination of O ( ꢀ ) yields
(
(
2
g
The arrangement shown in Fig. 1 allowed us to carry out
the experiments in H O despite the fact that the photon yield
2
1
is low due to the intrinsic O ( ∆ ) lifetime of only 3–4 µs in this
2
g
18
solvent. A volume of 1.2 ml of an O solution of known
3
concentration (determined spectrophotometrically, see Experi-
mental section) was placed in the cuvette and 0.4 ml of the
substrate solution (containing buffer if required) injected at
the bottom of the cuvette. As a reference, the formation of
Ozone rate constants were determined by competition
4
3
Ϫ1 Ϫ1 7
(
cf. ref. 15) using buten-3-ol (k = 7.9 × 10 dm mol s ) as
competitor. The measured entity is formaldehyde which was
1
O ( ∆ ) from the reaction of hydrogen peroxide with hypo-
2
g
16
determined by the Hantzsch reaction.
chlorite [reaction (10)] was used, and its quantum yield was
taken as unity (cf. also ref.1 and 2).
Methionine and its sulfoxide were separated by HPLC on
a 15 cm C-18 reversed phase column using water as eluent
(
retention times 5 and 3.6 min, respectively).
The product of the ozonation of trans-1,2-dithiane-4,5-diol
1
H O ϩ HOCl → O ϩ H O ϩ HCl
(10)
2
2
2
2
were identified by GC/MS after trimethylsilylation. There is
only one product, and its mass spectrum is compatible with that
of bis-TMS ether of 1-oxo-1λ -1,2-dithiane-4,5-diol (MW 312);
It is unavoidable that there are volumes within the cuvette,
from where the emissions do not reach the photodiode. Should
complete mixing be achieved before a major advance of the
reaction, this does not matter because only a fraction of the
emitted light is recorded anyway. However, when the mixing is
incomplete the reaction can, to a certain extent, preferentially
take place in volumes seen (or not seen) by the detector.
The recorded emission will thus be higher (or lower) than it
would be for homogeneous mixing. This situation leads to
4
m/z (%): 312 (1), 297 (3), 217 (17), 204 (3), 196 (19), 180 (4), 147
1
(
72), 133 (22), 116 (90), 101 (34), 73 (100). H NMR (D O):
2
δ = 4.17 (ddd; J = 11.5, 9.4, 3.6 Hz; 5α-H), 3.89 (ddd; J = 10.9,
.4, 3.7 Hz; 4β-H), 3.73 (dd; J = 13.5, 3.6 Hz; 6α-H), 3.44 (dd;
J = 14, 10.9 Hz; 3α-H), 3.17 (dd; J = 13.5, 11.5 Hz; 6β-H), 3.14
dd; J = 14, 3.7 Hz; 3β-H). For comparison, the mass spectrum
9
(
of the bis-TMS ether of trans-1,2-dithiane-4,5-diol (MW 296)
shows a similar pattern as concerns the major fragmentation
19
considerable scatter of the data even in the case of the slow
H O –HOCl reaction (cf. Fig. 2, main graph), but signal-
pathway; m/z (%): 296 (26), 281 (1), 203 (13), 180 (82), 147 (58),
2
2
1
1
33 (11), 116 (100), 101 (25), 73 (86). H NMR (D O): δ = 3.59
averaging over a reasonable number of runs improves the
2
(
m; 4-H, 5-H), 3.09 (br d; J = 13.3 Hz; one 3-H, one 6-H), 2.88
quality of the data (cf. Tables 1 and 2). However, when the rate
1
110 J. Chem. Soc., Perkin Trans. 2, 2001, 1109–1116

View Article Online
1
Ϫ3
Ϫ3
Table 2 Integrated emission of O ( ∆ ) in the reaction of O with EDTA (2.5 × 10 mol dm , pH 4.5)
2
g
3
Ϫ4
Ϫ3
1
Ϫ5
Ϫ3
1
Run
[O ]/10 mol dm
Area
[O ( ∆ )]/10 mol dm
% O ( ∆ )
2 g
3
2
g
1
2
3
4
5
6
7
X
2.26
2.20
2.13
2.04
1.99
1.92
1.72
0.0208
0.0227
0.0230
0.0253
0.0227
0.0292
0.0254
5.00
5.45
5.52
6.09
5.45
7.03
6.10
22.1
24.8
26.0
29.9
27.4
36.5
35.5
28.9 ± 5.4
of reaction becomes very fast as in the O –EDTA reaction
3
(
cf. inset in Fig. 2), complete mixing is not achieved within the
short reaction time, and hence the error becomes noticeably
larger.
On the other hand, when the reaction becomes very slow, the
signal may disappear in the noise. A case in point is amines
where only the free base is reactive toward O , and thus the rate
3
of reaction strongly depends on the pH (cf. ref. 15). Thus, at low
amine concentrations and/or a pH quite distant from the pK ,
a
1
the O ( ∆ ) signal may become small or disappear altogether.
2
g
Further errors result from the instability of the aqueous O3
stock solution. Therefore, repeated measurements of its O
3
1
content were required. The O ( ∆ ) yields were determined by
2
g
integrating the areas under the signal (in volts) vs. time curves
cf. Fig. 2) and are given in Table 3. They are based on the
average H O –HOCl- reference value (cf. Table 1), and the error
(
2
2
1
Fig. 3 Stern–Volmer plots for the quenching of O ( ∆ ) by DABCO in
2
g
given is only the statistical error of the measurements for the
1
H O (main graph) and D O (inset). O ( ∆ ) was generated by the
reaction of H O with HOCl (each 7.5 × 10 mol dm in H O and
2
2
2
g
Ϫ4
Ϫ3
individual compounds (cf. Table 2).
2
2
2
1
Ϫ5
Ϫ3
The O ( ∆ ) yields from the O reactions with a large number
7.5 × 10 mol dm in D
2
O). Solutions were basic due to the high
2
g
3
concentration of DABCO.
of compounds are compiled in Table 3. At an O : substrate
3
1
ratio = 1, the measured O ( ∆ ) yield is often very low. This can
2
g
be rationalized if there are fast free-radical processes causing
chain reactions which consume both O3 and the substrate.
In addition, the reaction may become so slow that the signal
disappears within the noise. Thus, only values for a ten-fold
substrate concentration may be used with some confidence.
However, often the low solubility of the substrate did not allow
the use of high concentrations.
8
3
Ϫ1 Ϫ1
from the literature which range from 5.8 × 10 dm mol
s
to
9
3
Ϫ1 Ϫ1
2
× 10 dm mol
Although this approach worked fine in the case of DABCO
and azide, a cautionary remark needs to be made. The H O –
s
(for a compilation see ref. 18).
2
2
1
HOCl reaction may not always be a good O ( ∆ ) source
for quenching experiments whenever either HOCl or H O2
or both) reacts rather fast with the quencher under study. In
2
g
2
(
our hands, this seemed to be the case with several amines (but
not with DABCO).
1
Quenching of O ( ꢀ ) by the substrate
2
g
1
Amines are known to be good O ( ∆ ) quenchers. A large
number of rate constants for O ( ∆ ) reactions have been com-
piled, but relatively few rate constants have been measured
in aqueous solutions. The value reported for O ( ∆ ) quenching
by DABCO in aqueous solution is 2.8 × 10 dm mol s ,
as derived from the analysis of a rather complex sequence of
reactions. Should this value be indeed that high, we would
have observed a substantial drop in the O ( ∆ ) yields at high
DABCO concentrations. This not being the case, a redeter-
mination of the O ( ∆ ) quenching rate constant by a more
direct method was required. We produced O ( ∆ ) by the H O –
HOCl reaction (see above) and measured the O ( ∆ ) yields as a
function of the DABCO concentration, both in H O and in
D O. Since the lifetime of O ( ∆ ) is much longer in D O than
in H O [k(in H O) ≈ 3 × 10 s , k(in D O) ≈ 1.8 × 10 s ],
2
g
1
2
g
Some mechanistic aspects
18
1
Amines. In the case of tertiary amines the formation of
2
g
8
3
Ϫ1 Ϫ1 20
1
O ( ∆ ) is connected with the formation of the aminoxide,
2
g
mainly reaction (1), although contributions of reaction (3)
may also play a small role (for studies in organic solvents see
refs. 3,4,21–23; in the gas phase the postulated aminoxide
1
2
g
24
intermediate further decomposes ). Due to a potential contri-
1
1
bution of reaction (4), the yield of O ( ∆ ) may be slightly lower
2
g
2
g
1
6
than that of the aminoxide. In the case of trimethylamine
and triethylamine, the products resulting from the reactions
following reaction (5), i.e. the corresponding dialkylamines
2
g
1
2
2
2
g
2
1
6,13
and aldehydes, complete the material balance. Only the free
2
2
5
g
Ϫ1
2
4
Ϫ1 18
amine reacts with O , and the protonated amine does not show
2
2
2
3
1
6,13,15,25,26
quenching of O ( ∆ ) in D O must occur at much lower
any reactivity.
Thus, when most of the amine is proto-
2
g
2
DABCO concentrations than in H O. This is indeed observed
nated (e.g. at pH 7) the rate of reaction slows down (at equal
total amine concentration). As a consequence, the chain
reactions induced by the free-radical pathway consume both O3
2
1
(
Fig. 3). The O ( ∆ ) quenching rate constant in H O and in
2 g 2
6 3 Ϫ1 Ϫ1
D O is calculated as k = 1.8 × 10 dm mol s .
2
q
1
This value is two orders of magnitude lower than the
and amine and hence reduce the yield of O ( ∆ ). Here, it is
2
g
1
previously reported one. Our low value of O ( ∆ ) quenching
of interest that not only the ozonide radical may start the chain
2
g
by DABCO now explains why no significant reduction of
but also the amine radical cation (cf. studies on trimethyl-
amine and EDTA ).
1
27
28
O ( ∆ ) yields was found in the O reactions at high DABCO
2
g
3
concentrations (see Table 3).
In the case of the aromatic tertiary amines, N,N-diethyl-
aniline and N,N,N Ј,N Ј-tetramethylphenylenediamine, the
Similar experiments were carried out with azide (data not
1
1
shown), and its O ( ∆ ) quenching rate constant was found to
yield of O ( ∆ ) is comparatively low (<10%, cf. Table 3). These
2
g
2
g
8
3
Ϫ1 Ϫ1
be 9 × 10 dm mol s , which compares well with the data
aromatic tertiary amines have lower reduction potentials than
J. Chem. Soc., Perkin Trans. 2, 2001, 1109–1116 1111

View Article Online
1
Table 3 O ( ∆ ) yields in O reactions in % of O consumed at different pH and at various substrate : O ratios (in parentheses). The data are average
2
g
3
3
3
Ϫ4
values of at least five individual determinations (scatter less than ± 15%). The O concentrations in these experiments were typically (1–2) × 10 mol
3
Ϫ3
dm
Substrate
pH
7
Yield (%) (substrate : ozone)
Trimethylamine
Triethylamine
63 (10 : 1)
9
7
8
9
.5
60 (10 : 1), 69 (100 : 1)
No signal (1 : 1), 42 (10 : 1), 72 (10 : 1)
No signal (1 : 1), 63 (10 : 1), 80 (100 : 1)
52 (1 : 1), 73 (10 : 1), 85 (100 : 1)
59 (1 : 1)
78 (100 : 1)
70 (1 : 1), 80 (10 : 1)
90 (10 : 1)
No signal (1 : 1), 19 (10 : 1)
No signal (1 : 1), 40 (10 : 1), 39 (100 : 1)
39 (1 : 1), 59 (10 : 1), 43 (100 : 1)
31 (1 : 1), 33 (10 : 1), 37 (100 : 1)
15 (10 : 1)
No signal (1 : 1), no signal (10 : 1)
No signal (1 : 1), 18 (10 : 1)
22 (1 : 1), 21 (10 : 1)
20 (1 : 1)
.5
1
0.5
1.5
7
1
DABCO
EDTA
9
4.5
5
7
9
.5
.5
2
ϩ
EDTA–Ca
EDTA–Fe
Nitrilotriacetic acid (NTA)
∼3
∼3
7
3
ϩ
9
.5
Diethylamine
9
1
1
1
1
0.5
9
0.5
9
0.5
7
0.5
8.5
3.5
6
24 (1 : 1), 20 (10 : 1)
17 (1 : 1)
20 (1 : 1), 18 (10 : 1)
No signal (1 : 1)
11 (1 : 1), 17 (10 : 1)
No signal (1 : 1), (10 : 1)
4 (10 : 1)
7 (17 : 1)
9 (3 : 1)
4 (3 : 1)
Iminodiacetic acid (IDA)
Ethylamine
Glycine
N,N-Diethylaniline
N,N,N Ј,N Ј-Tetramethylphenylenediamine
Uracil
3.5
7
No signal (9 : 1)
6 (9 : 1)
1
1
7 (9 : 1)
1
6
,3-Dimethyluracil
-Methyluracil
3.5
1
No signal (4 : 1)
No signal (4 : 1)
No signal (10 : 1)
12 (10 : 1)
1
3.5
7
1
1
3.5
7
0
3.5
7
15 (10 : 1)
No signal (4 : 1)
45 (4 : 1)
43 (4 : 1)
No signal (4 : 1)
4 (4 : 1)
5
-Chlorouracil
1
Thymine
1
1
1
7
7
0
7
0
7
8 (4 : 1)
Thymidine
Cytosine
No signal (10 : 1)
No signal (4 : 1)
No signal (16 : 1)
10 (16 : 1)
No signal (10 : 1)
No signal (10 : 1)
40 (10 : 1)
37 (1 : 1)
16 (10 : 1)
10 (1 : 1), 21 (10 : 1)
No signal (10 : 1)
6 (10 : 1)
1
2
Ј-Deoxycytidine
Cytidine
Ј-Deoxyguanosine
Guanosine
Ј-Deoxyadenosine
2
7
7
b
2
7
7
b
Adenosine
Phenol
1.8
7
9
0
9 (10 : 1)
8 (10 : 1)
1
1
Tyrosine
7
9 (1 : 1), 12 (1 : 1)
2
,4,6-Trimethylphenol
7
9
10 (1 : 1)
17 (10 : 1)
Pentachlorophenol
Pentabromophenol
8
8
9
7
58 (1 : 1), 68 (10 : 1)
48 (1 : 1), 59 (10 : 1)
19 (10 : 1)
6 (1 : 1)
30 (1 : 1)
2
1
1
,4,6-Triiodophenol
,4-Dimethoxybenzene
,3,5-Trimethoxybenzene
7
Nitrite
Azide
7
7
96 (100 : 1)
11 (1 : 1)
a
Cyanide
Bromide
Iodide
> 11
21 (10 : 1)
Weak broad signal (40 : 1), 56 (400 : 1)
12 (3 : 1)
7
7
7
Hydrogen sulfide
16 (100 : 1)
1
10
7
7, 11
7
1
15 (100 : 1)
96 (1.4 : 1)
Weak signal (23 : 1), 6 (230 : 1)
No signal (25 : 1)
6 (23 : 1)
Methanesulfinate
Formate
Glyoxylate
Dihydroxyfumarate
1
112
J. Chem. Soc., Perkin Trans. 2, 2001, 1109–1116

View Article Online
Table 3 (Contd.)
Substrate
pH
Yield (%) (substrate : ozone)
Muconate
Acetylenedicarboxylate
10
10
4.8
2 (100 : 1)
Weak signal (100 : 1), 15 (1000 : 1)
46 (38 : 1)
1
,4-Dithiothreitol
7
9
22 (38 : 1)
17 (38 : 1)
trans-1,2-Dithiane-4,5-diol
Methionine
Glyoxal
Acetaldehyde
Acetaldehyde dimethylacetal
Propan-2-ol
7
7
105 (31 : 1)
104 (42 : 1)
No signal (25 : 1)
No signal (1000 : 1)
No signal (1000 : 1)
No signal (500 : 1)
7, 11
7
7
7
a
b
Not corrected for quenching by substrate; upon correction the value is 17%. In the presence of tert-butyl alcohol.
This reaction is notably faster than the rate constant of O
3
3
Ϫ1 Ϫ1
with cystine for which a value of 550 dm mol
s
has been
25
reported. In order to see whether the reaction of trans-1,2-
dithiane-4,5-diol is unusually high, we also determined the
rate constant for bis(2-hydroxyethyl) disulfide and find a value
5
3
Ϫ1 Ϫ1
of 1.7 × 10 dm mol
s
for this disulfide, i.e. we have no
explanation for the low rate of reaction observed with cystine.
In both cases, the final products are the corresponding S-alkyl
sulfinates, corroborating that the O reaction proceeds by O-
3
transfer.
Compared with sulfides and disulfides, thiols show relatively
1
low O ( ∆ ) yields (< 50%), especially when deprotonated
2
g
[
pK (H S) = 7.0; pK (1,4-dithiothreitol) = 9.1, 10.5]. Thiolates
a 2 a
25
react much more readily with ozone than thiols, and changes
in O ( ∆ ) yields are already observed in a pH range where the
1
2
g
thiol still predominates. A case in point is 1,4-dithiothreitol
1
which already shows a strong drop in the O ( ∆ ) yield on
2
g
1
going from pH 4.8 to 7. This low O ( ∆ ) yield has, of course, a
2
g
Fig. 4 Reaction of ozone with methionine at pH 3. Consumption of
methionine and formation of its sulfoxide as a function of the ozone
concentration.
bearing on the reactions of O with e.g. H S, where it has been
noted that sulfate is formed as the only detectable product.
3
2
Interestingly, this reaction requires only ∼2.3 mol O per mol
3
32,33
H S.
This stoichiometry requires that an efficient chain
2
the aliphatic ones [e.g., E(N,N-dimethylaniline) ≈ 0.770 V,
29
reaction must occur (for an EPR study on the reactions of
E(N,N,N Ј,N Ј-tetramethylphenylenediamine) = 0.265 V], and
ؒ
ؒ
Ϫ
dioxygen with HS and S with O see ref. 34).
2
most probably the electron-transfer process becomes the major
ؒ
Ϫ
29
process [E(O /O3 ) = 1.01 V].
3
Phenols. Phenolate reacts as much as six orders of magnitude
25
faster with O than phenol. Thus, even at pH ∼7 (especially
3
Sulfur-containing compounds. The methanesulfinate ion
reacts rapidly with O by O-transfer to give methanesulfonate
in the presence of a buffer that re-establishes the equilibrium
3
upon phenolate depletion) the phenolate [pK (phenol) = 10] is
a
1
6
3
Ϫ1 Ϫ1 30
1
and O ( ∆ ) [reaction (11); k = 2 × 10 dm mol s ]. An
2
g
the dominating reactant. No O ( ∆ ) has been observed in the
reaction of phenol at pH 1.8, i.e. the formation of O ( ∆ ) at
2
g
1
2
g
Ϫ
Ϫ
1
pH 7 is likely due to a reaction with the phenolate present
CH S(O)O ϩ O → CH S(O) O ϩ O ( ∆ ) (11)
3
3
3
2
2
g
1
in the equilibrium. Much higher O ( ∆ ) yields are observed
2
g
with the halogenated phenolates. In view of its importance
electron-transfer reaction that would give rise to a very efficient
30
in water purification, the ozone chemistry of phenol has
chain reaction is not observed.
The sulfide methionine reacts very fast with O (k = 1.8 × 10
35–48
6
been widely studied,
but it is not yet fully understood. At
3
3
Ϫ1
Ϫ1 26
1
present, one can only speculate on the nature of the products
dm mol s ), the O ( ∆ ) yield being practically 100%
2
g
1
that may be formed in the course of O ( ∆ ) release, although
2
g
(
cf. Table 3 and ref. 2) [reaction (12)]. Its sulfoxide has been
catechol and hydroquinone are the most likely ones [cf.
reactions (14)–(16)]. These products are indeed formed, but
31
reported as the only detected product, and we find that the
methionine sulfoxide yield is indeed 100% (Fig. 4).
The disulfide trans-1,2-dithiane-4,5-diol in its reaction with
1
O3 also gives rise to 100% O ( ∆ ) [cf. reaction (13); k =
2
g
5
3
Ϫ1 Ϫ1
2
.1 × 10 dm mol s ].
J. Chem. Soc., Perkin Trans. 2, 2001, 1109–1116
1113

View Article Online
ؒ
Ϫ
1
Ϫ
largely originate from a chain reaction involving OH as
O ϩ NO2 → O ϩ ONOO
(25)
3
2
49
intermediates.5
0
51
ؒ
59
The ozone and OH-induced chemistry of halogenated
phenols has also been studied, but the analysis of the products
is even more difficult in these systems. A material balance has
not been achieved.
Phenols quench O ( ∆ ) with rate constants of ∼10 dm
mol s . This low quenching rate will not influence our
observation that peroxynitrite is not formed (< 2%) is in good
agreement with the thermochemistry of these two reactions
[reaction (25) is endothermic].
The reaction of O with azide has attracted some attention
because of the formation of peroxynitrite in this system.
No N has been found among the reaction products. For the
formation of peroxynitrite, this has led to the suggestion of a
two-step process. In the first step, N O and an NOO species
are thought to be formed. The reaction of the latter with a
second molecule of O3 would then give peroxynitrite [and
60–62
3
1
6
3
61,62
2
g
Ϫ1 Ϫ1 52
61
2
1
O ( ∆ ) yield at pH 7. Upon deprotonation, the quenching rate
2
g
8
3
Ϫ1 Ϫ1
Ϫ
constant is raised to 1.8 × 10 dm mol s , i.e. quenching still
has no effect at pH 9, but may have lowered our pH 10 value
slightly.
2
1
Ϫ
O ( ∆ ); the postulated NOO species cannot be identical with
2
g
DNA constituents and related compounds. A considerable
nitrite which gives rise to nitrate and not peroxynitrite,
see above]. This mechanism has to be disregarded because the
effect of pH on the rate of the reaction of O with the nucleo-
3
53
6
3
Ϫ1 Ϫ1 63
bases and related compounds has recently been reported.
reaction of O with azide is fast (k = 4 × 10 dm mol s )
3
Here, we show that this may be connected with a change in
and peroxynitrite builds up without an induction period even
1
Ϫ3 61
mechanism, i.e. the O ( ∆ ) yield may become quite pronounced
at an azide concentration of 0.1 mol dm . Considering that
2
g
61
Ϫ
at high pH where the nucleobase is deprotonated. The most
striking example is 5-chlorouracil (cf. Table 1). We suggest that
when the neutral molecule predominates, the ozone reaction
follows the Criegee mechanism [overall reactions (18) and (19)],
no material balance has as yet been obtained, the N –O
3 3
system has to be re-investigated, and the question of N2
formation has to be re-addressed. Were it a major product,
peroxynitrite should be formed according to reactions (26) and
(27), compatible with what we know now about the formation
54,55
but upon deprotonation [reaction (17)] an isopyrimidine
is
formed [reaction (20)]. Preliminary experiments indicate that 5-
chlorouracil indeed yields HCl, parabanic acid (imidazolidine-
trione), and formic acid [reaction (19)], whereas in the reaction
of the 5-chlorouracilate ion with ozone the yields of parabanic
acid and formic acid are reduced and isodialuric acid (dihydro-
Ϫ
ؒ
ؒ
2
Ϫ
N3 ϩ O → NO ϩ O ϩ N2
(26)
(27)
3
ؒ
ؒ
Ϫ
Ϫ
NO ϩ O2 → ONOO
6
-hydroxypyrimidine-2,4,5(3H)-trione) is formed instead
and decay of this species (cf., e.g., ref. 64). The formation of
1
56
[reactions (20)–(23)].
O ( ∆ ) and N O, for which there is no adequate explanation
2 g 2
as yet, could also be reconciled by the sequence (28) and (29).
O-transfer to azide may lead to HNO [reaction (28)] which is a
well-known precursor of N O [reaction (29)].
2
Ϫ
1
Ϫ
O ϩ N3 ϩ H O → O ( ∆ ) ϩ HNO ϩ N ϩ OH (28)
3
2
2
g
2
2
HNO → N O ϩ H O
(29)
2
2
Ϫ
1
Bromide and iodide ions. In its reaction with Br , O ( ∆ )
2
g
is produced in a ∼50% yield [reaction (30)], and in the case of
Ϫ
1
I the O ( ∆ ) yield is only ∼12% [reaction (31)].
2
g
Ϫ
1
Ϫ
O ϩ Br → O ( ∆ ) ϩ BrO
(30)
(31)
3
2
g
Ϫ
1
Ϫ
O ϩ I → O ( ∆ ) ϩ IO
3
2
g
Reaction (30) is currently considered to be the only primary
Ϫ
65–69
Ϫ
Ϫ
process occurring in the Br system.
Since Br and I are
1
Ϫ
6
3 Ϫ1 Ϫ1 Ϫ
poor O ( ∆ ) quenchers [k(Br ) = 1 × 10 dm mol s ; k(I ) =
.2 × 10 dm mol s ], it is difficult to see why the O ( ∆ )
2
g
6
3
Ϫ1 Ϫ1 18
1
7
2 g
yields should not reach 100%. Despite this, one may consider
Ϫ
Compared to the other nucleobases, the rate of the reaction
that an O –Br -adduct, the bona-fide intermediate in reaction
3
1
of adenine and its derivatives with O is slow, and OH radicals
3
(30), is sufficiently long-lived to reduce the O ( ∆ ) yield by spin
2
g
are formed (possibly via superoxide radicals as intermediates).
To suppress subsequent reactions, tert-butyl alcohol was added
conversion due to the heavy atom effect caused by the bromine
Ϫ
atom. In the corresponding reaction with I , such a heavy
ؒ
1
as OH-scavenger in order to measure the intrinsic O ( ∆ ) as
2
g
atom effect should be even more pronounced. Indeed, the
1
1
accurately as possible. The formation of O ( ∆ ) is an indication
2
g
O ( ∆ ) yield is even lower in this system. However, these
2
g
1
that an O-transfer reaction must occur, but in the reaction of
low O ( ∆ ) yields also raise the question of whether there are
2
g
1
2
Ј-deoxyadenosine the N -oxide is not formed to a noticeable
additional free-radical pathways in these systems. This question
57
extent, indicating that an O-transfer to N(1) does not occur
this is the position of N-oxide formation by hydrogen peroxide,
is currently under investigation in our laboratory. The reaction
Ϫ
(
of O with I is one of the standard methods for measuring
3
1
70–72
cf. ref. 58). The reaction leading to O ( ∆ ) is thus not yet
2
g
O concentrations. In this assay, I is determined.
Its yield
3
2
known. This also holds for guanosine.
has been in dispute, cf. ref. 73, although there now seems to be
1
general agreement that the stoichiometry is 1:1. The low O ( ∆ )
2
g
Nitrite and azide ions. In its reaction with O , the nitrite
yield found in this study may be the basis for studying again
the mechanistic details of this interesting and certainly complex
reaction.
3
1
ion gives rise to O ( ∆ ) in ∼100% yield. In principle, one may
2
g
envisage two pathways of O-transfer, one leading to nitrate
reaction (24)], the other to peroxynitrite [reaction (25)]. The
[
Cyanide ion. In the cyanide–O system cyanate is practically
the only product, but besides the O-transfer which leads to
3
Ϫ
1
Ϫ
3
O ϩ NO2 → O ϩ NO
(24)
3
2
1
114
J. Chem. Soc., Perkin Trans. 2, 2001, 1109–1116

View Article Online
1
O ( ∆ ) there are two chain reactions (one can be interrupted by
7 P. Dowideit and C. von Sonntag, Environ. Sci. Technol., 1998, 32,
2
g
74
1
1112.
tert-butyl alcohol, the other one cannot). Thus, the O ( ∆ )
yields given in Table 3 only reflect its yields under very specific
experimental conditions.
2
g
8
9
R. Atkinson and W. P. L. Carter, Chem. Rev., 1984, 84, 437.
E. Amouyal, A. Bernas and D. Grand, Photochem. Photobiol., 1979,
2
9, 1071.
0 A. Bernas and D. Grand, J. Phys. Chem., 1994, 98, 3440.
11 D. Yu, A. Rauk and D. A. Armstrong, J. Phys. Chem., 1992, 96,
031.
2 X.-M. Zhang and Q. Zhu, J. Org. Chem., 1997, 62, 5934.
13 F. Muñoz, Doctoral thesis, Ruhr Universität, Bochum, 1999.
1
Saturated compounds. The reactions of ozone with saturated
compounds are still poorly understood. In the case of propan-
6
1
2
-ol (neat or in an organic solvent) O evolves when ozone
2
75 1
reacts at low temperatures and some 3% of O ( ∆ ) is detected
2
g
1
4 L. Forni, D. Bahnemann and E. J. Hart, J. Phys. Chem., 1982, 86,
255.
15 F. Muñoz and C. von Sonntag, J. Chem. Soc., Perkin Trans. 2, 2000,
1
76
in the presence of an O ( ∆ ) probe. Due to incomplete
scavenging of O ( ∆ ) under these conditions, this must be a
2
g
1
2
g
lower value. In aqueous solution, the O reaction may proceed
661.
3
77
1
6 T. Nash, Biochem. J., 1953, 55, 416.
with free radicals as intermediates, but this has not been
78
17 P. K. Singh, L. Field and B. J. Sweetman, J. Org. Chem., 1988, 53,
reported for propan-2-ol. Hydrotrioxides are expected inter-
79,80
2608.
mediates.
Such intermediates may also account for the
1
8 F. Wilkinson, W. P. Helman and A. B. Ross, J. Phys. Chem. Ref.
Data, 1995, 24, 663.
19 A. M. Held, D. J. Halko and J. K. Hurst, J. Am. Chem. Soc., 1978,
100, 5732.
pronounced selectivity of the ozone reaction with -glucose in
ؒ
81
the presence of an OH scavenger. In a recent study on the
reaction of propan-2-ol with ozone in organic solvents and
at low temperatures, the formation of the hydrotrioxides
2
0 Y. Lion, E. Gandin and A. van de Vorst, Photochem. Photobiol.,
980, 31, 305.
21 H. B. Henbest and J. W. Stratford, J. Chem. Soc., 1964, 711.
1
(
CH ) C(OH)OOOH and HOOOH has now been established
3 2
17 82
by O NMR. Furthermore, it has been shown that the decay
of these species is catalysed by water. These experiments were
carried out at 0 ЊC in deuterated acetone up to a water content
of 4 mol dm . It is very daring to extrapolate these data to
neat water (55 mol dm ) and to room temperature, yet the low
2
2
2 P. S. Bailey and J. E. Keller, J. Org. Chem., 1968, 33, 2680.
3 P. S. Bailey, D. A. Lerdal and T. P. Carter, Jr., J. Org. Chem., 1978,
43, 2662.
24 E. C. Tuazon, R. Atkinson, S. M. Aschmann and J. Arey, Res.
Chem. Intermed., 1994, 20, 303.
5 J. Hoigné and H. Bader, Water Res., 1983, 17, 185.
6 W. A. Pryor, D. H. Giamalva and D. F. Church, J. Am. Chem. Soc.,
Ϫ3
Ϫ3
2
2
rate of the decay of the hydroperoxides under their conditions
does not exclude the possibility that the lifetimes of these hydro-
peroxides may be longer than the few seconds required for
1
984, 106, 7094.
2
7 S. Das, M. N. Schuchmann, H.-P. Schuchmann and C. von Sonntag,
1
an adequate determination of the O ( ∆ ) signal (cf. Fig. 2).
Chem. Ber., 1987, 120, 319.
28 B. Höbel and C. von Sonntag, J. Chem. Soc., Perkin Trans. 2, 1998,
2
g
However, the information available from radiation-chemical
5
09.
studies on the properties of HOOOH suggests a sufficiently
8
3,84
29 P. Wardman, J. Phys. Chem. Ref. Data, 1989, 18, 1637.
30 R. Flyunt, O. Makogon, M. N. Schuchmann, K.-D. Asmus and
C. von Sonntag, J. Chem. Soc., Perkin Trans. 2, 2001, 787.
31 J. B. Mudd, R. Leavitt, A. Ongun and T. T. McManus, Atmos.
Environ., 1969, 3, 669.
short lifetime.
The (water-catalysed) decay of (CH ) C(OH)-
3 2
1
OOOH and HOOOH is expected to give rise to O ( ∆ ). Thus,
the fact that we were unable to detect any O ( ∆ ) is very
2
g
1
2
g
intriguing.
3
2 P. K. Overbeck, Proc. Annu. Conf. Am. Water Works Assoc., 1995,
41.
2
Conclusions
3
3
3 G. Mark, E. Mvula and C. von Sonntag, unpublished results.
4 J. Zhu, K. Petit, A.-O. Colson, S. DeBolt and M. D. Sevilla,
J. Phys. Chem., 1991, 95, 3676.
35 H. R. Eisenhauer, J. Water Pollut. Control Fed., 1968, 40, 1887.
6 M. L. Konstantinova, S. D. Razumvskii and G. E. Zaikov, Bull.
Acad. Sci. USSR, Div. Chem. Sci., 1991, 266.
Through product analysis, we have shown that upon reaction
of O with methionine, methanesulfinic acid and nitrite O-
3
1
transfer is the only reaction. In these cases, the O ( ∆ ) yield
2
g
3
is also 100%, i.e. material balance is obtained. This not only
1
indicates that our O ( ∆ ) yield measurements can be used with
37 M. L. Konstantinova, S. D. Razumvskii and G. E. Zaikov, Bull.
Acad. Sci. USSR, Div. Chem. Sci., 1991, 271.
2
g
some confidence within the limitations discussed above, but
1
3
8 Y. Skarlatos, R. C. Barker and G. L. Haller, J. Phys. Chem., 1975, 79,
587.
9 Y. Yamamoto, E. Niki, H. Shiokawa and Y. Kamiya, J. Org. Chem.,
979, 44, 2137.
that O ( ∆ ) yields provide most valuable information as to
2
g
2
mechanistic details of ozone reactions in aqueous solution and
3
certainly also in non-aqueous solvents. Thus, the determination
1
1
of O ( ∆ ) yields is an indispensable tool in mechanistic studies
2
g
40 P. C. Singer and M. D. Gurol, Wasser ’81, 1981, 519.
41 J. A. Roth, W. L. Moench and K. A. Debalak, J. Water Pollut.
Control Fed., 1982, 54, 135.
2 D. Lesczynska, Environ. Protect. Eng., 1982, 8, 105.
3 P. C. Chrostowski, A. M. Dietrich and I. H. Suffet, Water Res., 1983,
of ozone reactions in general. The ease of such experiments
and the relatively low equipment cost may stimulate a more
abundant use of this technique.
4
4
1
7, 1627.
4
4
4
4
4 P. C. Singer and M. D. Gurol, Water Res., 1983, 17, 1163.
5 M. D. Gurol and P. C. Singer, Water Res., 1983, 17, 1173.
6 M. D. Gurol and R. Vatistas, Water Res., 1987, 21, 895.
7 J. P. Duguet, B. Dussert, J. Mallevialle and F. Fiessinger, Water Sci.
Technol., 1987, 19, 919.
Acknowledgements
This work has been supported by the German Bundesminister
für Bildung und Forschung (Project: 02-WT-9583). F. M. and
E. M. thank the DAAD for a stipend. The very able technical
help given by Sigrid Russell is warmly acknowledged. We
also thank Dr J. Leitich for the interpretation of the complex
NMR spectra.
4
4
8 S. Beulker and M. Jekel, Ozone: Sci. Eng., 1993, 15, 361.
9 E. Mvula, M. N. Schuchmann and C. von Sonntag, J. Chem. Soc.,
Perkin Trans. 2, 2001, 264.
50 X. Fang, Y. He, J. Liu and J. Wu, Radiat. Phys. Chem., 1998, 53, 441.
1 X. Fang, H.-P. Schuchmann and C. von Sonntag, J. Chem. Soc.,
Perkin Trans. 2, 2000, 1391.
5
5
5
2 F. E. Scully and J. Hoigné, Chemosphere, 1987, 16, 681.
3 J. A. Theruvathu, R. Flyunt, C. T. Aravindakumar and C. von
Sonntag, J. Chem. Soc., Perkin Trans. 2, 2001, 269.
References
1
2
3
4
5
J. R. Kanofsky and P. Sima, J. Biol. Chem., 1991, 266, 9039.
J. R. Kanofsky and P. D. Sima, Photochem. Photobiol., 1993, 58, 335.
L. Horner, H. Schaefer and W. Ludwig, Chem. Ber., 1958, 91, 75.
A. Maggiolo and S. J. Niegowski, Adv. Chem. Ser., 1959, 21, 202.
P. S. Bailey, Ozonation in Organic Chemistry. Vol. 2. Nonolefinic
Compounds, Academic Press, New York, 1982.
54 M. I. Al-Sheikhly, A. Hissung, H.-P. Schuchmann, M. N.
Schuchmann, C. von Sonntag, A. Garner and G. Scholes, J. Chem.
Soc., Perkin Trans. 2, 1984, 601.
55 M. N. Schuchmann, M. Al-Sheikhly, C. von Sonntag, A. Garner
and G. Scholes, J. Chem. Soc., Perkin Trans. 2, 1984, 1777.
56 R. Flyunt, J. A. Theruvathu and C. von Sonntag, unpublished
results.
6
F. Muñoz and C. von Sonntag, J. Chem. Soc., Perkin Trans. 2, 2000,
2
029.
J. Chem. Soc., Perkin Trans. 2, 2001, 1109–1116
1115

View Article Online
5
5
7 R. Flyunt and C. von Sonntag, unpublished results.
8 J. F. Mouret, F. Odin, M. Polverelli and J. Cadet, Chem. Res.
Toxicol., 1990, 3, 102.
70 D. H. Byers and B. E. Saltzman, Adv. Chem. Ser., 1959, 21, 93.
71 R. S. Ingols, R. H. Fetner and W. H. Eberhardt, Adv. Chem. Ser.,
1959, 21, 102.
5
6
6
9 G. Mark and C. von Sonntag, unpublished results.
0 K. Gleu and E. Roell, Z. Anorg. Allg. Chem., 1929, 179, 233.
1 W. A. Pryor, R. Cueto, W. H. Koppenol, M. Ngu-Schwemlein,
G. L. Squadrito, P. L. Uppu and R. M. Uppu, Free Radicals Biol.
Med., 1995, 18, 75.
72 E. D. Parry and D. H. Hern, Environ. Sci. Technol., 1973, 7, 65.
73 L. R. Schmitz, Environ. Sci. Technol., 1973, 7, 647.
74 F. Muñoz, P. Ulanski and C. von Sonntag, unpublished results.
75 M. C. Whiting, A. J. N. Bolt and H. R. Park, Adv. Chem. Ser., 1968,
77, 4.
76 R. W. Murray, W. C. Lumma, Jr. and J. W. P. Lin, J. Am. Chem. Soc.,
1970, 92, 3205.
77 J. Staehelin and J. Hoigné, Vom Wasser, 1983, 61, 337.
78 J. Hoigné and H. Bader, Water Res., 1983, 17, 173.
79 P. S. Nangia and S. W. Benson, J. Am. Chem. Soc., 1980, 102, 3105.
80 D. H. Giamalva, D. F. Church and W. A. Pryor, J. Am. Chem. Soc.,
1986, 108, 7678.
81 M. N. Schuchmann and C. von Sonntag, Aqua (Oxford), 1989, 38,
311.
6
6
6
2 R. M. Uppu, G. L. Squadrito, R. Cueto and W. A. Pryor, Methods
Enzymol., 1996, 269, 311.
3 P. Neta, R. E. Huie and A. B. Ross, J. Phys. Chem. Ref. Data, 1988,
1
7, 1027.
4 G. Merényi, J. Lind, S. Goldstein and G. Czapski, Chem. Res.
Toxicol., 1998, 11, 712.
6
6
6
5 W. R. Haag, J. Hoigné and H. Bader, Vom Wasser, 1982, 59, 238.
6 W. R. Haag and J. Hoigné, Environ. Sci. Technol., 1983, 17, 261.
7 P. Westerhoff, R. Song, G. Amy and R. Minear, Water Res., 1998,
3
2, 1687.
82 B. Plesnicar, J. Cerkovnik, T. Tekavec and J. Koller, Chem. Eur. J.,
2000, 6, 809.
83 G. Czapski and B. H. J. Bielski, J. Phys. Chem., 1963, 67, 2180.
84 G. Czapski, Annu. Rev. Phys. Chem., 1971, 22, 171.
6
6
8 U. von Gunten and J. Hoigné, Aqua (Oxford), 1992, 41, 209.
9 U. von Gunten and J. Hoigné, Environ. Sci. Technol., 1994, 28,
1
234.
1
116
J. Chem. Soc., Perkin Trans. 2, 2001, 1109–1116