Determination of .OH, O2.-, and hydroperoxide yields in ozone reactions in aqueous solution

Article abstract of DOI:10.1021/jp022455b

In ozone reactions in aqueous solutions, ^.OH and O₂^.- are often generated as short-lived intermediates and hydroperoxides are formed as labile or stable final products. Tertiary butanol reacts with ozone only very slowly but readily with ^.OH. In the presence of dioxygen, formaldehyde is a prominent final product, 30 ± 4%, whose ready determination can be used as an assay for ^.OH. Although dimethyl sulfoxide reacts much more readily with ozone, its fast reaction with ^.OH which gives rise to methanesulfinic acid can also be applied for the determination of ^.OH, at least in fast ozone reactions. The formation of O₂^.- can be assayed with tetranitromethane (TNM), which yields nitroform anion (NF^-) at close to diffusion-controlled rates. TNM is stable in neutral and acid solution but hydrolyzes in basic solution (k = 2.7 M^-1 s^-1), giving rise to NF^- plus nitrate ion (62%) and CO₂ plus 4 nitrite ions (38%). TNM reacts with O₃ (it = 10 M^-1 s^-1), yielding 4 mol of nitrate (plus CO₂) and 4 mol of O₃ are consumed in this reaction. NF^- reacts with O₃ (k = 1.4 × 10⁴ M^-1 s^-1) by O-transfer. The resulting products, (NO₂)₃CO^- and (NO₂)₂C=O, rapidly hydrolyze (K > 10 s^-1), and most of the nitrite released is further oxidized by ozone to nitrate. In the case of slow ozone reactions, these reactions have to be taken into account; i.e. the NO₃^- yield has to be measured as well. For the determination of hydroperoxides, Fe²⁺-based assays are fraught with considerable potential errors. Reliable data may be obtained with molybdate-activated iodide. The kinetics of this reaction can also be used for the characterization of hydroperoxides. Reactive hydroperoxides undergo rapid O-transfer to sulfides, e.g., k(HC(O)OOH + (HOCH₂CH₂)₂S] = 220 M^-1 s^-1, and the corresponding reaction with methionine may be used for their quantification (detection of methionine sulfoxide by HPLC). Distinction of organic hydroperoxides and H₂O₂ by elimination of the latter by reaction with catalase can often be used with advantage but fails with formic peracid, which reacts quite readily with catalase (k = 1.3 × 10^-3 dm³ mg^-1 s^-1). Some examples of ^.OH and O₂^.- formation in ozone reactions are given.

Full text of DOI:10.1021/jp022455b

7242
J. Phys. Chem. B 2003, 107, 7242-7253
ARTICLES
Determination of OH, O2•-, and Hydroperoxide Yields in Ozone Reactions in Aqueous
Solution†
•
Roman Flyunt,‡,§,| Achim Leitzke,‡,§ Gertraud Mark,‡ Eino Mvula,‡,§, Erika Reisz,‡,3
Roland Schick,¶ and Clemens von Sonntag*,‡,§
Max-Planck-Institut fu¨r Strahlenchemie, Stiftstrasse 34-36, P.O. Box 101365,
D-45413 Mu¨lheim an der Ruhr, Germany, Institut fu¨r Oberfla¨chenmodifizierung (IOM), Permoserstrasse 15,
04318 Leipzig, Germany, and ZweckVerband Bodensee-WasserVersorgung, Betriebs- und Forschungslabor,
D-78354 Sipplingen, Germany
ReceiVed: NoVember 20, 2002; In Final Form: January 14, 2003
•
•-
In ozone reactions in aqueous solutions, OH and O2 are often generated as short-lived intermediates and
hydroperoxides are formed as labile or stable final products. Tertiary butanol reacts with ozone only very
•
slowly but readily with OH. In the presence of dioxygen, formaldehyde is a prominent final product, 30 (
•
4%, whose ready determination can be used as an assay for OH. Although dimethyl sulfoxide reacts much
•
more readily with ozone, its fast reaction with OH which gives rise to methanesulfinic acid can also be
•
•-
applied for the determination of OH, at least in fast ozone reactions. The formation of O2 can be assayed
with tetranitromethane (TNM), which yields nitroform anion (NF-) at close to diffusion-controlled rates.
TNM is stable in neutral and acid solution but hydrolyzes in basic solution (k ) 2.7 M-1 s-1), giving rise to
NF- plus nitrate ion (62%) and CO2 plus 4 nitrite ions (38%). TNM reacts with O3 (k ) 10 M-1 s-1), yielding
4 mol of nitrate (plus CO2) and 4 mol of O3 are consumed in this reaction. NF- reacts with O3 (k ) 1.4 ×
104 M-1 s-1) by O-transfer. The resulting products, (NO2)3CO- and (NO2)2CdO, rapidly hydrolyze (k > 10
s-1), and most of the nitrite released is further oxidized by ozone to nitrate. In the case of slow ozone reactions,
-
these reactions have to be taken into account; i.e. the NO3 yield has to be measured as well. For the
determination of hydroperoxides, Fe2+-based assays are fraught with considerable potential errors. Reliable
data may be obtained with molybdate-activated iodide. The kinetics of this reaction can also be used for the
characterization of hydroperoxides. Reactive hydroperoxides undergo rapid O-transfer to sulfides, e.g.,
k(HC(O)OOH + (HOCH2CH2)2S] ) 220 M-1 s-1, and the corresponding reaction with methionine may be
used for their quantification (detection of methionine sulfoxide by HPLC). Distinction of organic hydroperoxides
and H2O2 by elimination of the latter by reaction with catalase can often be used with advantage but fails
with formic peracid, which reacts quite readily with catalase (k ) 1.3 × 10-3 dm3 mg-1 s-1). Some examples
•
•-
of OH and O2 formation in ozone reactions are given.
Introduction
bacteria.1,2 In many aspects, it can hence substitute chlorine.
Obviously, the chlorination byproducts such as trihalomethanes,
which limit the use of chlorine, are not formed upon ozonolysis,
but there must be also unwanted byproducts which have not
yet been elucidated in detail.2,3
Ozonolysis is gaining importance in drinking water and
wastewater processing. In arid countries, where drinking water
resources are scarce, it is widely used in the so-called reclama-
tion process, where wastewater is processed in several steps
for reuse as drinking water. Besides acting as an oxidant, it is
a very good disinfectant that copes readily with viruses and
Surprisingly little is known at present about the reactions of
ozone in aqueous solutions, and a detailed knowledge of these
reactions is a prerequisite for any further evaluation of a wider
application of ozone, especially in the reclamation technology.
In the case of simple olefins, the Criegee mechanism4 is
practically strictly followed; i.e. 1 mol of carbonyl compound
and 1 mol of R-hydroxyhydroperoxide are formed per mol of
ozone.5 In these reactions, no free radicals are generated.
On the other hand, in the reaction of ozone with amines,
including the chelators NTA and EDTA, a nonnegligible
pathway proceeds via •OH and O2•- radicals as intermediates.6
An example where •OH formation becomes the only process is
the reaction of H2O2 with ozone. This reaction has found quite
† Part of the special issue “Armin Henglein Festschrift”.
* To whom correspondence should be addressed. Fax: +49-208-3951.
E-mail: clemens@vonsonntag.de.
‡ Max-Planck-Institut fu¨r Strahlenchemie.
§ IOM.
| On leave from: Institute of Physico-Chemistry, National Academy of
Science of the Ukraine, Naukova St. 3a, UA-79053 L’viv, Ukraine.
On leave from: Chemistry Department, University of Namibia,
Windhoek, Namibia.
∇ On leave from: Faculty of Industrial Chemistry and Environmental
Engineering Technical University of Timisoara, Bul. Victoriei 2, 1900
Timisoara, Romania.
¶ Zweckverband Bodensee-Wasserversorgung.
10.1021/jp022455b CCC: $25.00 © 2003 American Chemical Society
Published on Web 03/19/2003

Ozone Reactions in Aqueous Solution
J. Phys. Chem. B, Vol. 107, No. 30, 2003 7243
some interest considering its potential to eliminate ozone-
refractory compounds in drinking water processing with the help
of the highly reactive •OH radical (for reviews see, e.g., refs 7
and 8; for compilations of ozone and OH rate constants see
refs 9 and 10, respectively).
follow the reaction by measuring one of the ensuing products,
formaldehyde (for details see below). Alternatively, dimethyl
sulfoxide (DMSO) may be used as •OH scavenger (cf. reaction
21). Methanesulfinic acid is the measured product. This ap-
proach is limited to highly ozone-reactive substrates, since
DMSO itself reacts with ozone with a rate constant of 8.2 M-1
•
At low pH, where H2O2 predominates (pKa(H2O2) ) 11.6),
the reaction is slow (e0.01 M-1 s-1 11
,
k ) 0.036 M-1 s-1 12),
s-1 9
.
and in neutral solutions the reaction is mainly governed by the
fast reaction of HO2- in equilibrium (k ) 5.5 × 106 M-1 s-1).11
The initiation step may proceed via a short-lived adduct (a
The formation of O2•- is readily detected by its fast reaction
with tetranitromethane (TNM, reaction 9, k ) 2 × 109 M-1
s-1 22
)
by measuring the formation of the nitroform anion,
•
-
hydropentoxide anion) which decomposes into OH, O2, and
C(NO2)3 (NF-, ꢀ(350nm) ) 15 000 M-1cm-1), a common
technique in peroxyl radical chemistry23 but also applied to
ozone chemistry.24
•-
•-
O2 (reaction 1). The subsequent reaction of O2 with ozone
is very fast (reaction 2, k ) 1.6 × 109 M-1 s-1).13,14 The ozonide
radical anion is only stable at very high pH (equilibrium 3, k3
) 2.1 × 103 s-1, k-3 ) 3.9 × 109 M-1 s-1).15 Due to the fast
protonation of O•- by water (equilibrium 4; pKa(•OH) ) 11.8),10
equilibrium 4 is rapidly drawn to the left at neutral pH (k4 ≈
108 s-1). As a consequence, •OH is formed at pH g 7
•
O2•- + C(NO2)4 f O2 + C(NO2)3- + NO2
(9)
In the present paper, it will be shown that in ozone reactions
the latter assay is not as straightforward as in peroxyl radical
chemistry, and the rapid reactions of ozone with NF- as well
•-
approximately at the same rate as O3 decays (reaction 3). In
-
as NO2 (a further product) have to be taken into account.
acid solutions and in neutral solution in the presence of buffer,
Furthermore, the characterization and quantification of hydro-
peroxides which are common products in ozone reactions will
be reported, and a number of ozone reactions will be described,
•-
•
O3 is protonated and OH formation is caused via the short-
lived HO3 radical.13
•
•
•-
where OH/O2 play an important role.
HO2- + O3 f [HO5 ] f •OH + O2 + O2
(1)
(2)
(3)
(4)
-
•-
Although our prime aim is a better understanding of the
reactions that prevail in drinking water processing, some of our
data must also be helpful for the elucidation of ozone reactions
in clouds.
•-
O2•- + O3 f O2 + O3
O3•- H O•- + O2
Experimental Section
O
•- + H2O H •OH + OH-
Ozone was generated with a dioxygen-fed ozonator (Philaqua
Philoz 04, Gladbeck). The ozone content of the aqueous ozone
stock solutions was determined spectrophotometrically using ꢀ-
(260nm)25,26 ) 3300 M-1 cm-1. Solutions were made up in
Milli-Q-filtered (Millipore) water. Formaldehyde was deter-
mined by the Hantzsch method.27 In this assay, 2 mL of reagent
(0.2 mL of acetylacetone, 3 mL of acetic acid, and 25 g of
ammonium acetate in 100 mL of water) are mixed with 5 mL
of the sample, heated for 30 min at 50 °C in the dark. The
absorbance at 412 nm is read (ꢀ ) 7530 M-1 cm-1). Alterna-
tively, formaldehyde can be determined by HPLC of its 2,4-
dinitrophenylhydrazone.28 Here, 0.2 mL of a 9 mM solution of
2,4-dinitrophenylhydrazine in acetonitrile is mixed with 0.1 mL
of 1 M perchloric acid in acetonitrile and 1.7 mL of the sample.
After standing for 45 min in the dark, the solution is chromato-
graphed on a C-18 reversed phase column with acetonitrile/
water (1:1). This method is reported to be generally applicable
for the determination of aldehydes and ketones, but it failed to
give reliable results in the H2O2/O3/tertiary butanol system (see
below). An assignment of the various aldehydes and ketones
in a reaction mixture can be done by gas chromatography/mass
spectrometry.29 The most prominent peaks of the mass spectra
(CH5, Finnigan MAT) of the 2,4-dinitrophenylhydrazones of
the carbonylic products that arise from the oxidation of tertiary
butanol are as follows. Formaldehyde-2,4-DNPH. MW: 210.
m/z (%): 210 (58), 180 (18), 152 (8), 122 (17), 106 (9), 91
(25), 89 (26), 79 (100), 63 (99), 51 (64). Acetone-2,4-DNPH.
MW: 238. m/z (%): 238 (100), 222 (6), 208 (6), 181 (38), 152
(36), 122 (24), 115 (35), 105 (20), 91 (34), 79 (86), 59 (86).
2-Hydroxy-2-methylproprionaldehyde-2,4-DNPH. MW: 268.
m/z (%): 268 (28), 253 (3.5), 225 (31), 203 (7), 183 (9), 164
(13), 148 (5), 59 (81), 55 (83) 43 (100).
The reaction of ozone with organic substrates can lead to
•-
•-
O3 (by electron transfer, reaction 5) and to O2 (by an
addition/elimination process, reaction 6). Both processes lead
•
to the formation of OH (cf. reactions 2-4).
•-
RH + O3 f RH•+ + O3
(5)
(6)
RH + O3 f RO• + O2•- + H+
The situation can become very complex at low substrate
•
concentrations, since OH reacts rapidly with ozone (reaction
7, k ) 1 × 108 M-1 s-1)9 and thus can initiate a chain reaction
•
•-
•
with HO2 /O2 as intermediates (cf. equilibrium 8, pKa(HO2 )
) 4.8).16 This chain reaction may lead to an unwanted self-
destruction of ozone.
•OH + O3 f HO2 + O2
(7)
(8)
•
HO2 H O2•- + H+
•
•
For a better understanding of ozone reactions in water, OH
and O2 yields have to be known. For monitoring •OH
•-
formation, one approach is to follow the consumption of an
added ozone refractory compounds.17,18 This procedure has the
advantage that the low additive concentration does not es-
sentially alter the progress of the reaction, but it has also the
•
disadvantage that the ensuing reactions of the OH radical are
not effectively suppressed.
In the present study, we follow a different approach. The
addition of a large excess of tertiary butanol which readily reacts
with •OH (cf. reactions 10 and 11, k10,11 ) 6 × 108 M-1 s-1 10
)
Nitrite/nitrate and methanesulfinate/methanesulfonate were
analyzed by ion chromatography (Dionex DX-100; AS-14;
eluent, 4.75 × 10-4 M Na2CO3/4.63 × 10-4 M NaHCO3).30
but is practically inert against ozone (k ) 1 × 10-3 M-1 s-1 19
and 3 × 10-3 M-1 s-1 20) allows us to scavenge all •OH and to

7244 J. Phys. Chem. B, Vol. 107, No. 30, 2003
Flyunt et al.
For the study of the destruction of H2O2 and formic peracid,
10 µL of catalase (from beef liver, Boehringer Mannheim) was
added per 10 mL of sample. Formic peracid was generated by
the ozonolysis of an aqueous solution 1,2-dichloroethene (cf.
ref 5).
comparison to reaction 20. If reaction 21 contributes, it would
lead to the formation of 2-hydroxy-2-methylpropanal and HO2
(reactions 24 and 25).
•
2•OOCH2C(CH3)2OH f [HOC(CH3)2CH2OO]2 (15)
To follow the reaction of H2O2 with iodide, the absence of
dioxygen had to be ensured. Deoxygenated solutions were
placed into a spectrophotometric quartz cell which could be
further purged with argon (cf. ref 31), but the inlet tube did not
fully reach the bottom for the analyzing light of the UV
spectrophotometer (Lambda 16, Perkin-Elmer) to pass through
the cell. A serum cap allowed the injection of small amounts
of reactants. The reactions of hydroperoxides with aqua-Fe2+
have been monitored at 305 nm (ꢀ ) 2095 M-1 cm-1) and those
[HOC(CH3)2CH2OO]2 f
O2 + HOC(CH3)2CH2OH + HOC(CH3)2CHO (16)
[HOC(CH3)2CH2OO]2 f H2O2 + 2HOC(CH3)2CHO (17)
[HOC(CH3)2CH2OO]2 f O2 + 2CH2O + 2•C(CH3)2OH
(18)
4-
with Fe(CN)6 at 420 nm (ꢀ ) 1020 M-1 cm-1).
[HOC(CH3)2CH2OO]2 f O2 + 2HOC(CH3)2CH2O• (19)
HOC(CH3)2CH2O• f CH2O + •C(CH3)2OH (20)
HOC(CH3)2CH2O• f HOC(CH3)2C(H,OH)• (21)
Stopped-flow experiments were carried out using a Biologic
SFM3 setup. Conductance changes were measured with a
conductometer (CDM3, Radiometer).
Results and Discussion
•C(CH3)2OH + O2 f •OOC(CH3)2OH
•OOC(CH3)2OH f (CH3)2CdO + HO2
(22)
(23)
Tertiary Butanol System. Hydroxyl radicals react with
tertiary butanol by abstracting an H-atom mainly from carbon
(95%) and, to a much lesser extent, from oxygen (reactions 10
and 11).32 The tertiary butoxyl radical rapidly decomposes by
â-fragmentation into acetone and a methyl radical (reaction 12,
•
HOC(CH3)2C(H,OH)• + O2 f HOC(CH3)2C(H,OH)OO•
(24)
k ) 1.4 × 106 s-1 33
;
cf. also ref 34).
•OH + (CH3)3COH f H2O + •CH2C(CH3)2OH (10)
HOC(CH3)2C(H,OH)OO• f HOC(CH3)2CHO + HO2
•
•OH + (CH3)3COH f H2O + (CH3)3CO•
(CH3)3CO• f (CH3)2CdO + •CH3
(11)
(12)
(25)
In competition with the unimolecular reaction 25, the peroxyl
radicals HOC(CH3)2C(H,OH)OO• may undergo bimolecular
decay processes, leading to, e.g., the formation of 2-hydroxy-
2-methylpropionic acid.41 Thus, the product yields will vary
somewhat with the experimental conditions, notably the rate of
radical generation.
In the presence of dioxygen, the alkyl radicals add dioxygen,
yielding the corresponding peroxyl radicals (reactions 13 and
14). As with other simple alkyl radicals,23,35 the rate of these
reactions must be close to diffusion controlled (k ≈ 2 × 109
M-1 s-1).
The yields of the main products, formaldehyde, acetone,
2-hydroxy-2-methylpropanal, and 2-methyl-2-hydroxypropanol,
have been determined in the radiolytic study.37 However, the
•CH2C(CH3)2OH + O2 f •OOCH2C(CH3)2OH (13)
•
formation of considerable amounts of HO2 /O2•- in this system
which also interact with the tertiary-butanol-derived peroxyl
radicals prevents a quantitative comparison with the present
system, where any O2 formed in the course of the tertiary-
•CH3 + O2 f •OOCH3
(14)
•-
The chemistry of methylperoxyl radical has been investigated
in some detail,36 but since it plays only a minor role here, it
will no longer be discussed.
butanol-derived peroxyl radicals are immediately converted into
further OH by ozone (cf. reactions 2-4). The regeneration of
•
•OH, however, does not lead to an efficient chain reaction
because of the two competing pathways 16 and 17.
It has been recently shown that the tertiary-butanol-derived
peroxyl radicals react with ozone in competition with their
bimolecular decay.40 This leads to the formation of oxyl radicals
(reaction 26) and consequently to an increase in the yields of
formaldehyde and 2-hydroxy-2-methylpropanal (cf. reactions 20,
21, and 25).41
The alkyl peroxyl radicals decay bimolecularly, whereby a
tetroxide is formed as a short-lived intermediate (reaction 15;
for a review on peroxyl radical reaction in aqueous solution
see ref 23). In the case of the tertiary-butanol-derived peroxyl
radicals,37 reaction 16 leads to dioxygen, 2-methyl-2-hydroxy-
propanol, and 2-methyl-2-hydroxypropanal, while reaction 17
gives rise to hydrogen peroxide and 2 mol of 2-methyl-2-
hydroxypropanal. These reactions are concerted processes and
do not proceed via free radical intermediates. The third reaction
pathway leads to the formation of 2 mol of formaldehyde, 1
mol of dioxygen, and 2 mol of 2-hydroxyprop-2-yl radicals in
a concerted reaction (cf. ref 38; reaction 18). The same products
may be formed via oxyl radicals (reactions 19 and 20). The
2-hydroxyprop-2-yl radical readily adds to dioxygen (reaction
HOC(CH3)2CH2OO• + O3 f O2 + HOC(CH3)2CH2O•
(26)
Reaction 26 is slow (k ) 1.8 × 104 M-1 s-1 40
)
and has to
compete with the bimolecular decay of the tertiary-butanol-
derived peroxyl radicals (2k ) (8 ( 2) × 108 M-1 s-1).37
Moreover, alkyl radical also reacts with ozone at a rate
comparable to that of dioxygen.42 This reaction gains in
importance at high [O3]/[O2] ratios. Details have been discussed
elsewhere.41
•
22), and this reaction is followed by a fast elimination of HO2
(reaction 23, k ) 650 s-1).39 In competition with the â-frag-
mentation reaction 20, the 1,2-H-shift reaction 21 may also
occur, although recently suggested40 to be at the most minor in

Ozone Reactions in Aqueous Solution
J. Phys. Chem. B, Vol. 107, No. 30, 2003 7245
CH3S(O)O- + O3 f CH3S(O)2O- + O2(1∆g) (28)
This and other side reactions are of little consequence when
ion chromatography is used to assay the formation of meth-
anesulfinic acid, since methanesulfonic acid elutes nearby (see
Experimental Section), and their combined yields allow conclu-
sions as to the yield of •OH formation. It will be shown below
that the determination of methanesulfonic acid is a prerequisite
for the exact determination of •OH yields in ozone reactions by
the DMSO method. In fact, major parts of the methanesulfinic
acid formed may undergo autoxidation, possibly induced by
peroxyl radicals (see below).
Tetranitromethane/Nitroform Anion System. For the use
of tetranitromethane as an O2 scavenger in ozone reactions,
Figure 1. Formation of formaldehyde (2, 4) and 2-hydroxy-2-
methylpropanal (b, O) in the reaction of ozone with H2O2 (0.5 mM)
in the presence of tertiary butanol (10 mM). Closed and open symbols
are from two sets of experiments.
•-
it is of importance to know the ozone rate constants with TNM
and that with the primary product, the nitroform anion (NF-).
Moreover, the reaction products of ozone with TNM and NF-
may also react with ozone, and for this reason these products
and their yields must be known as well. To generate NF-, the
hydrolysis of TNM is the most convenient method. How-
ever, in the literature there are conflicting reports as to the
products.46,47 This required a reinvestigation of the hydrolysis
of TNM which is reported next.
Hydrolysis of Tetranitromethane. TNM is stable in acid and
neutral solutions but hydrolyzes quite rapidly in basic solu-
tions.46,47 While the earlier study46 reports that the hydrolysis
takes two routes, an OH--attack at nitrogen which yields the
NF- and nitrate (reaction 29) and an OH--attack at carbon
which leads to the formation of CO2 and nitrite (reaction 30),
in ref 47 only the former route is considered.
In the radiolytic system, the formaldehyde yield was found
to be ∼25% of •OH, and the ratio of the 2-hydroxy-2-
methylpropanal and formaldehyde yields was ∼1.5.37 We have
now determined the formaldehyde and 2-hydroxy-2-methylpro-
panal yields in the ozone/hydrogen peroxide/tertiary butanol
system (variation of [H2O2] ) 0.5 mM and 1 mM; [tertiary
butanol] ) 10 and 20 mM, respectively). According to our
experience, yield data in ozone reactions are fraught with a
somewhat higher intrinsic error (poorer dose yield plots, poorer
day-to-day reproducibility) than radiolytic yield data. Because
•
of the importance of these data for establishing an OH probe
in ozone reactions, we have measured these yields many times,
and we obtain 30 ( 4% for the formaldehyde yield and a value
of 1.3-1.4 for the 2-hydroxy-2-methylpropanal-to-formaldehyde
ratio (cf. Figure 1).
C(NO2)4 + 2OH- f (NO2)3C- + NO3- + H2O (29)
C(NO2)4 + 4OH- f CO2 + 4NO2- + 2H2O (30)
We have also tried to quantify the other carbonylic products,
acetone as 2,4-dinitrophenylhydrazone. It turns out that the
method reported by Lipari and Swarin28 allows an accurate
determination of the aldehydes but fails to determine acetone
adequately in this mixture of carbonylic products (its yield
comes out too low with respect to the formaldehyde yield).
When the reagent concentration is raised 3-fold (a further
increase would impede the quantification of 2-methyl-2-hy-
droxypropanal due to an overlap with the tail of the reagent
peak), formaldehyde and acetone yields approach one another,
but even with this improvement the method is not recommended
for quantification of acetone in our system.
For the present study, it was most important to know to which
extent reaction 30 takes place, because in this reaction nitrite is
formed which reacts very fast with ozone by O-atom transfer
(reaction 31, k ) 3.7 × 105,48 (3.5-4.0) × 105,49 and 6.08 ×
105 M-1 s-1 50
;
for the formation of O2(1∆g), see ref 45).
NO2- + O3 f NO3- + O2(1∆g)
(31)
•
After completion of the hydrolysis of TNM, we find [NO3-]/
[NO2-] ) 0.4 and an NF- yield of 62%, in good agreement
with the earlier report.46 In this context, the rate of TNM
hydrolysis was redetermined by following the build-up of NF-
as a function of the OH- concentration. In these experiments,
the OH- concentration was always in large excess over the TNM
concentration (ensuring first-order kinetics), and from the data
shown in Figure 2 a rate constant of k ) 0.27 M-1 s-1 is
obtained. The reported47 value of 0.4 M-1 s-1 (at 25 °C) is in
reasonable agreement with our data (for a compilation of rate
constants determined in the present study, see Table 1).
Reaction of Tetranitromethane with Ozone. TNM reacts
slowly with ozone. The reaction has been followed spectro-
photometrically at 260 nm as well as by changes in conductance
as a function of time (Figure 3). From these data, rate con-
stants of k ) 10.2 (conductometry) and 9.9 M-1 s-1 (spectro-
photometry) are calculated. Their mean value of 10 M-1 s-1 is
given in Table 1.
However, for the quantification of OH formation in ozone
reactions by its scavenging with tertiary butanol it usually
suffices to determine the formaldehyde yield formed in reactions
18 and 20 by the fast Hantzsch method.27 If colored products
or educts interfere with this assay, one always can use HPLC
of the 2,4-dinitrophenylhyrazones as discussed above.
•
Dimethyl Sulfoxide System. In the reaction of OH with
dimethyl sulfoxide (DMSO), methanesulfinic acid is formed in
92% yield (reaction 27).21
•OH + (CH3)2SdO f CH3S(O)OH + •CH3
(27)
Methanesulfinic acid is readily detected by ion chroma-
tography (see Experimental Section) or by a sensitive spectro-
•
photometric assay.43 The use of DMSO as OH scavenger in
ozone reactions has the disadvantage that DMSO reacts reason-
ably fast with ozone (k ) 8.2 M-1 s-1), and also the product of
reaction 27, the methanesulfinate ion, is rapidly further oxidized
by ozone to methanesulfonate in an O-transfer reaction (reaction
28, k ) 2 × 106 M-1 s-1),44 with singlet dioxygen O2(1∆g) as
the other product.44,45
This rate constant is low when compared with many other
ozone rate constants, but noticeably higher than the one with
tertiary butanol mentioned above and in the range of rate

7246 J. Phys. Chem. B, Vol. 107, No. 30, 2003
Flyunt et al.
TABLE 1: Compilation of Rate Constants (M-1 s-1) Determined in the Present Study and Related Values Taken from the
Literature
reaction
k
ref
C(NO2)4 + OH- f products
0.27 (20 °C)
0.4 (25 °C)
10
this work
47
this work
this work
this work
this work
51
this work
this work
this work
this work
this work
this work
5
52
this work
5
this work
5
5
C(NO2)4 + O3 f products
C(NO2)3- + O3 f products
1.4 × 104
>10 s-1
48 (18 °C)
60 (25 °C)
39 (18 °C)
13 (18 °C)
220
(NO2)2CdO + H2O f 2HNO2 + CO2
Fe2+ + H2O2 f Fe3+ + •OH + OH-
Fe2+ + HOCH2OOH f Fe3+ + HOCH2O• + OH-
Fe3+ + t-BuOOH f Fe3+ + t-BuO• + OH-
HC(O)OOH + (HOEt)2S f HC(O)OH + (HOEt)2SO
HC(O)OOH + HOEtSSEtOH f HC(O)OH + HOEtS(O)SEtOH
HC(O)OOH + 2Fe(CN)64- f HC(O)OH + 2Fe(CN)63- + 2H+
HC(O)OOH + catalase f products
10
4.3
(1.3 × 10-3
)
a
H2O2 + I-/molybdate-activated f I3
2.1
-
H2O2 + I- f I3
very slow
1.8 × 10-3
1.7 × 103
1.7 × 103
2.6 × 10-2
0.39
-
HC(O)OOH + I-/molybdate-activated f I3
-
-
HC(O)OOH + I- f I3
HOCH2OOH + I-/molybdate-activated f I3
-
CH3CH(OH)OOH + I-/molybdate-activated f I3
-
(CH3)2C(OH)OOH + I-/molybdate-activated f I3
1.5 × 10-4
7.5
5
53
-
5-hydroperoxy-5-methylhydantoine + I- f I3
-
a In units of dm3 mg-1 s-1
.
conductance measurements (Figure 3) and by ion chromatog-
raphy. The formation of NF- is not observed, and nitrite yields
are very small (<2% of nitrate). It is tentatively suggested that
this reaction proceeds by a bipolar addition of ozone at nitrogen.
The subsequent elimination of dioxygen may be followed by a
rearrangement into trinitromethyl nitrate (reaction 32).
C(NO2)4 + O3 f (NO2)3CONO2 + O2
(32)
Nitrate is a better leaving group than nitrite, and it is expected
that (NO2)3CONO2 hydrolyzes (reaction 33) much faster than
C(NO2)4 (for its hydrolysis see above).
(NO2)3CONO2 + H2O f NO3- + (NO2)3COH + H+ (33)
Figure 2. Hydrolysis of tetranitromethane at 20 °C. Rate of nitroform
formation as a function of the OH- concentration.
The trinitromethanol thus formed immediately loses nitrous
acid (reaction 34) (HCl release from trichloromethanol occurs
within less than a few microseconds),54 and the hydrolysis of
dinitro carbon monoxide (reaction 35) will also be fast.
(NO2)3COH f H+ + NO2- + (NO2)2CdO
(34)
(35)
-
(NO2)2CdO + H2O f CO2 + 2H+ + 2NO2
Nitrite is then readily oxidized by ozone (reaction 31). The
fact that practically no nitrite is observed after completion of
the reaction is a strong indication that the bona fide intermediate
(NO2)3CONO2 must have hydrolyzed quite rapidly, faster than
the time required for the completion of the reaction, which was
12 min in the above experiment. This also requires that the
lifetime of (NO2)2CdO must be short (see below).
Reaction of the Nitroform Anion with Ozone. In the prepara-
tion of NF- by the hydrolysis of TNM, nitrite is a major product
which reacts very fast with ozone (see above). Hence, for the
determination of the rate constant of ozone with the NF-, nitrite
could be destroyed most conveniently by ozonolysis. The
remaining nitrite concentration was ∼3% of that of NF-. Using
the stopped-flow technique, the decay of NF- at 350 nm was
followed in the presence of a large excess of ozone (Figure 4),
and from such data a rate constant of k ) 1.4 × 104 M-1 s-1
is obtained.
Figure 3. Reaction of ozone (7.2 × 10-5 M) with tetranitromethane
(6.7 × 10-4 M) as a function of time as followed by the buildup of
conductance, starting at pH 5. Inset: first-order plot of these data (b)
and of data (4), where the rate of reaction was followed by the decay
of the 260 nm absorption of ozone ([O3] ) 4.5 × 10-5 M; [tetra-
nitromethane] ) 8.2 × 10-4 M) at pH 3.
constants of H-abstraction reactions of saturated compounds
having activated C-H functions or olefins deactivated by
electron-withdrawing substituents. Thus, this reaction deserves
some further comments.
In the reaction of TNM with ozone, 4 mol of ozone is
consumed and 4 mol of nitrate ion is formed (C(NO2)4 + 4O3
+ 2H2O f 4HNO3 + 4O2 + CO2). This is shown by

Ozone Reactions in Aqueous Solution
J. Phys. Chem. B, Vol. 107, No. 30, 2003 7247
Figure 5. Rate of reaction of Fe2+ in 0.1 M H2SO4 with H2O2 (b)
hydroxymethylhydroperoxide (O), and tertiary butylhydroperoxide (2)
at 18 °C as a function of the Fe2+ concentration. The kinetics of the
reaction was followed by stopped-flow.
Figure 4. Reaction of ozone (2.4 × 10-4 M) with the nitroform anion
(2 × 10-5 M) as followed by the decay of the nitroform anion at 350
nm in a stopped-flow experiment. Inset: plot of kobs vs the ozone
concentration ([nitroform anion] ) 2 × 10-5 M).
Detection of Hydroperoxides. In ozone reactions, hydro-
peroxides are common products. They are sometimes only very
short-lived, too short-lived to be detectable by a peroxide assay,
and their formation can only be inferred from the final
products.58 However, there is a wide span of hydroperoxides
that are stable for seconds, minutes, or “forever”. Hydrogen
peroxide is such a long-lived hydroperoxide, but it may react
with an ozonolysis product, and then it can fade away even on
the time scale of typical analytical procedures. A case in point
is the ozonolysis of cinnamic acid.58 Here, benzaldehyde,
hydrogen peroxide (via an R-hydroxyhydroperoxide that is too
short-lived to be detected), and glyoxylic acid are the products
(reaction 39).58 The latter two react with one another (reaction
40), and the hydroperoxide thus formed decomposes into formic
acid and carbon dioxide (reaction 41).
This value is ∼30 times lower than the rate constant of ozone
with nitrite (in agreement with a stationary ratio of [NF-]/
[NO2-] ≈ 30 observed upon ozonolysis of NF-). This has a
the yield of NF- formation. This aspect will be discussed below.
It is suggested that the reaction of ozone with the NF-
proceeds by O-atom transfer (reaction 36).
•-
bearing on the determination of the O2 yield by following
(NO2)3C- + O3 f (NO2)3CO- + O2
(36)
This reaction is followed by a rapid release of nitrite (reaction
37 followed by the hydrolysis of (NO2)2CdO, reaction 35, see
above].
-
(NO2)3CO- f (NO2)2CdO + NO2
(37)
Ph-CHdCH-CO2- + O3 f
An alternative pathway, the reaction by electron transfer
(reaction 38), can be excluded.
-
Ph-C(O)H + H2O2 + HC(O)-CO2 (39)
-
H2O2 + HC(O)-CO2- f HOOC(OH)H-CO2 (40)
(NO2)3C- + O3 f (NO2)3C• + O3
(38)
•-
-
HOOC(OH)H-CO2- f CO2 + HCO2
(41)
•-
•
This reaction would give rise to O3 and thus to OH. The
formation of OH can be readily detected by the formation of
•
For any mechanistic study, a material balance is a prerequisite.
Hence for the study of ozone reactions, reliable assays for
hydroperoxides which allow their quantitative determination are
required. In the following, some established assays will be
discussed and some new ones presented.
methanesulfinic acid in the presence of DMSO (see above). The
absence of methanesulfinic (and its oxidation product meth-
anesulfonic acid) indicates that reaction 38 does not take place.
The lifetime of the bona fide intermediates (NO2)3CO- and
(NO2)2CdO must be indeed very short. According to the
proposed mechanism, ozonolysis of NF- leads to the formation
of 3 mol of NO2-, but 3 mol of NO3- and no NO2- are detected
after the completion of the reaction. The stoichiometry of the
reaction (as followed by conductance and optical measurements)
is 4 mol of ozone/mol of NF-. Thus, NO2- released in reactions
37 and 35 is immediately oxidized to NO3-. This requires that
these hydrolyses (i.e. release of nitrite which subsequently reacts
very rapidly with ozone) are much faster than the reaction of
NF- with ozone under such conditions. The first step, reaction
34, will certainly be very fast, typically in the microsecond time
range (cf. the release of HCl from geminal chlorohydrins).54-56
In analogy with phosgene, the resulting (NO2)2CdO is expected
to hydrolyze on the subsecond time scale (reaction 35). Due to
the much stronger electron-withdrawing properties of the -NO2
substituent compared to the -Cl substituent, one even may
suspect that reaction 35 is faster than the hydrolysis of phosgene
under such conditions (k ∼ 10 s-1).54,57 Indeed, if the hydrolysis
of (NO2)2CdO were as slow as that of phosgene, considerable
Fe(II)-Based Assays. Aqua-Fe2+ is known to reduce H2O2
and organic hydroperoxides. For H2O2, the rate constant is ∼60
M-1 s-1 at 25 °C.51 The rate of reaction is strongly temperature
dependent,51 and our value of 48 M-1 s-1 determined at 18 °C
(b in Figure 5) agrees very well with these data. The rate
constant of the reaction of hydroxymethylhydroperoxide (pro-
duced in the reaction of 1,1-dichloroethene with ozone; cf. ref
5) with Fe2+ is not much slower (k ) 39 M-1 s-1; O in Figure
5). Noticeably slower is the reaction with tertiary butylhydro-
peroxide (k ) 13 M-1 s-1; 2 in Figure 5), but these differences
are too small for the characterization of a given hydroperoxide
based on its kinetics with aqua-Fe2+
.
This ready reaction of Fe2+ with hydroperoxides has often
been used for their quantification (e.g., refs 59 and 60), but the
question has to be asked, whether one can rely on such data.
In principle, the stoichiometry of the reaction hydroperoxides
with Fe2+ is simple; i.e. 1 mol of hydroperoxide may yield 2
mol of Fe3+ (reactions 42-45; k44 ≈ 3 × 108 M-1 s-1 61
,
k45
≈
3 × 108 M-1 s-1; cf. ref 10). For organic hydroperoxides, two
-
amounts of NO2 should have been found after ozonolysis.
pathways are conceivable, reactions 42 and 43, but only the

7248 J. Phys. Chem. B, Vol. 107, No. 30, 2003
Flyunt et al.
Figure 7. Formation of Fe3+ in the reaction of tertiary butylhydro-
peroxide with Fe2+ (3 mM) at pH 1.5 as a function of the tertiary
butylhydroperoxide concentration in the presence of air (b) and in
deoxygenated solution (O).
Figure 6. Reaction of H2O2 (varied between 0.2 and 2 mM) with Fe2+
(6 mM) in the presence of 2-propanol at pH 1.5. Formation of Fe3+ as
a function of the 2-propanol concentration.
former route is taken in a one-electron reduction process.62 The
reason for this is the much higher solvation energy of OH- as
compared to that of RO-.
[Fe2+]/[2-propanol] ratios (k(•OH+2-propanol) ) 1.9 × 109 M-1
s-1),10 where the Fe3+ yield approaches the value 2Fe3+/H2O2.
When Fe2+ reacts with organic hydroperoxides, it is no longer
guaranteed that even at low substrate concentrations 2 mol of
Fe3+ are formed/(mol of hydroperoxide reacted). The reason
for this is the occurrence of the fast unimolecular processes.
An example is tertiary butylhydroperoxide. Here, 4 mol of Fe3+
are formed/(mol of hydroperoxide) in the presence of dioxygen,
but only 0.8 mol of Fe3+ in its absence (Figure 7).
Fe2+ + ROOH f Fe3+ + RO• + OH-
Fe2+ + ROOH f Fe3+ + RO- + •OH
Fe2+ + RO• + H+ f ROH + Fe3+
Fe2+ + •OH f Fe3+ + OH-
(42)
(43)
(44)
(45)
This low yield, which is even below unity, deserves a
comment. The rapid fragmentation of the tertiary butoxyl radical
(reaction 12) prevents a further oxidation of Fe2+ according to
reaction 43. In reaction 12, methyl radicals are formed. They
are capable of abstracting an H-atom from tertiary butylhydro-
peroxide (cf. k(•CH3+H2O2) ) 2.7 × 104 M-1 s-1].68 This
reaction consumes tertiary butylhydroperoxide, which thus is
no longer capable of reacting with Fe2+. The fast termination
reactions of the methyl and tertiary butylperoxyl radicals will
eliminate the latter before it can be reduced by Fe2+ to a major
extent.
A system which seems to follow this sequence of reactions
with practically no side reaction is the reaction of formic peracid
with Fe(CN)64-. At pH ∼3.7 the rate constant of this reaction
3-
is 4.3 M-1 s-1, and ∼1.9 mol of Fe(CN)6 and ∼0.9 mol of
formic acid are formed. These yields were independent of the
4-
Fe(CN)6 concentration (0.4-4 mM) and were not affected
•
by an addition of tertiary butanol. This indicates that no OH
are formed in this electron-transfer reaction (confirming the
earlier62 conclusion) and that the HC(O)O• radical undergoes
the otherwise very fast 1,2-H-shift (cf. reaction 48) only slowly
if at all. The reason for this could be a resonance stabilization
of the alkoxyl radical by the carbonyl group.
It is thus concluded that Fe2+-based hydroperoxide assays,
although most useful for their detection, are often not well-
suited for their quantification (cf. also ref 69).
Reaction 44 does not consider competing reactions. In reality,
however, the RO• radicals are usually very short-lived and may
react with the substrate RH (reaction 46, k46 ≈ 3 × 105 M-1
s-1 for the tert-butoxyl63 and the methoxyl64 radicals with
methanol), decompose by â-fragmentation (e.g., reaction 12, k
) 1.4 × 106 s-1),33 or undergo an 1,2-H-shift65-67 when
R-hydrogens are available (e.g., reaction 47). In water, the 1,2-
Iodide Assay. Iodide is known to react very slowly with
hydrogen peroxide and many other hydroperoxides, but this
reaction can be considerably speeded up by molybdate cataly-
sis,70 for example with H2O2 by 3 orders of magnitude (Table
1). Similar to the reaction of sulfides with hydroperoxides (see
below), the reaction may proceed via a two-electron reduction
process (nucleophilic substitution; reaction 50). The measured
entity is I3- (reactions 51 and 52). To shift equilibrium 52 fully
to the right side, high iodide concentrations are required.
shift reaction is also very fast, typically ∼106 s-1
RO• + RH f ROH + R•
.
(46)
(47)
CH3O• f •CH2OH
H2O2 + I- f IOH + OH-
IOH + I- f I2 + OH-
(50)
(51)
(52)
In the presence of dioxygen, the alkyl radicals are rapidly
converted into the corresponding peroxyl radicals (reaction 48)
which are subsequently reduced by Fe2+ into hydroperoxides
(reaction 49).
-
I2 + I- H I3
R• + O2 f ROO•
(48)
(49)
The potential alternative to reaction 50, a one-electron-transfer
process (H2O2 + I- f I + OH + OH-), must not be
considered, since this (and also the proton-assisted) reaction is
endothermic by 31 kcal mol-1 (12 kcal mol-1 at pH 0; G.
Mere´nyi, private communication); i.e. side reactions involving
•OH have not been considered.
•
•
ROO• + Fe2+ + H+ f ROOH + Fe3+
Thus, a short chain reaction is induced. To exemplify this
point, the formation of Fe3+ in the reaction of H2O2 with Fe2+
in the presence of 2-propanol is shown in Figure 6. Here,
quantitative oxidation of Fe2+ by •OH is only achieved at high
An advantage of this assay is that the kinetics of the reaction
varies considerably from hydroperoxide to hydroperoxide.5 This

Ozone Reactions in Aqueous Solution
J. Phys. Chem. B, Vol. 107, No. 30, 2003 7249
Table 1). The reaction with the sulfide can be used to reduce
reactive organic hydroperoxides (a case in point is the ozonolysis
of thymine).53
Although they can be separated readily from their parent
sulfides by HPLC, simple aliphatic sulfoxides do not have
sufficiently strong absorptions at λ > 200 nm to use them for
the quantitative determination of reactive hydroperoxides. The
amino acid methionine also gives this reaction. Its corresponding
sulfoxide absorbs more strongly in the UV, and, therefore,
methionine can be used with advantage to quantify reactive
hydroperoxides (see Experimental Section).
Factors To Be Taken into Account in the Determination
•
of OH in Ozone Reactions with the Tertiary Butanol/
Figure 8. Rate of reaction of formic peracid with (HOCH2CH2)2S (b)
and HOCH2CH2SSCH2CH2OH (O) as a function of the sulfide/disulfide
concentration. The kinetics were followed by the buildup of conductance
due to the formation of formic acid.
Formaldehyde Assay. It has been shown above that in the
•
reaction of OH with tertiary butanol the formation of formal-
dehyde is connected with the concomitant release of O2•-. Since
•
the latter reacts rapidly with ozone giving rise to further OH,
can be used to characterize a given hydroperoxide, and if the
rate constants of two hydroperoxides with this reagent differ
significantly (ga factor of ∼10), this assay can be used to
quantify the yields of these hydroperoxides by a kinetic
analysis.71 Rate constants for the reaction of some hydroper-
oxides that are often formed in the ozonolysis of olefinic
compounds are given in Table 1. As can be seen from this table,
the more reactive hydroperoxides such as formic peracid react
with iodide also without molybdate catalysis.
Titanyl Sulfate Assay. It must be mentioned that the com-
plexation of H2O2 by Ti(IV) in sulfuric acid solution that gives
rise to a yellow color is often used for the determination of
H2O2. However, the molar absorption coefficient is only low
(ꢀ(410nm) ) 710 M-1 cm-1);72,73 i.e. the method is not very
sensitive. Organic hydroperoxides give rise to this assay only
after hydrolysis, which may be slow and incomplete due to
competing reactions.73 In favorable cases, however, the kinetics
of this assay may also be used for the characterization of a given
hydroperoxide.
this additional source of formaldehyde has to be taken into
•
account when the primary OH yield is calculated from such
data. In the above study, •OH was generated by the reaction of
•-
ozone with an excess of H2O2; i.e. it did not matter if any O2
•
•
gave rise to OH via its reaction with ozone or whether OH
formation was entirely due to the reaction of ozone with H2O2.
However, in any other system, the formaldehyde yield will be
higher than the above value of (30 ( 4%)/(1 mol of primary
•OH) due to the short chain reaction. The relationship between
the yield of the chain carrier forming reaction x and the total
yield of chain reaction product (y), y ) x/(1 - x), was
established by solving the initial value problem of the set of
ordinary differential equations, modeling the reaction mechanism
and taking the limes to infinite time. Applied to our system, a
correction factor of 0.48 ( 0.08 is obtained by which the
measured formaldehyde yield has to be divided to arrive at the
precursor •OH yield; i.e. as a rule of thumb one can take twice
the formaldehyde yield.
•
When the initial OH yield is high, ozone depletion due to
Destruction of H2O2 by Catalase. The destruction of H2O2
with the help of catalase is often a most useful tool for
distinguishing between H2O2 and organic hydroperoxides. The
addition of 10 µL of catalase to 10 mL of 0.1 mM H2O2 solution
eliminates H2O2 in less than 2 s (the time required for adding
catalase and subsequently molybdate-activated iodide), while
organic hydroperoxides are typically not or only very little
affected. For example, at a catalase concentration of 0.8 mg
dm-3 methylhydroperoxide decays with a half-life of a day.36
In contrast, we now have found that formic peracid is very
rapidly destroyed by catalase (kobs ) 1.2 × 10-2 s-1 at a catalase
concentration of 8.9 mg L-1, pH ∼ 7). In the reaction with a
sulfide (two-electron reductant), formic peracid is fully trans-
the secondary formation of •OH in this assay can no longer be
neglected and has to be taken into account in material balance
calculations. Furthermore, when the rate of reaction of ozone
•-
with the substrate is very fast, O2 released in the tertiary-
butanol-derived peroxyl radical reactions may no longer be able
to react with ozone and thus the correction factor mentioned
above may not have to be applied. A case in point may be
N,N,N′,N′-tetramethylphenylenediamine (cf. data in Table 2).
In addition, when •OH formation in ozone reactions becomes a
major process, there are always substrate radicals to be taken
into account which will interact with the tertiary-butanol-derived
peroxyl radicals. This must have an influence on the formal-
dehyde yield. Whether this will lead to an increase or a decrease
cannot be predicted. As a consequence, the tertiary butanol assay
is an excellent tool for the detection of •OH in ozone reactions,
but as its quantitative aspect is concerned, it is fraught with an
error beyond the experimental one, especially when the form-
aldehyde (•OH) yield is found to be high.
Formation of •OH in Ozone Reactions. It has been
mentioned above that •OH may be formed in an electron-transfer
process (reaction 5). This reaction always has to compete with
other reactions of ozone. Thus, the differences in the reduction
potentials of substrate and ozone are not expected to be a
measure of the electron-transfer efficiency, although a larger
difference will favor electron transfer and thus •OH formation.
Most experiments reported here were carried out at pH ∼ 7.
For assessing the reduction potential difference, one has to take
the ozone reduction potential at pH 7. In the compilation of
4-
formed into formic acid and with Fe(CN)6 (one-electron
reductant) to at least 90% (see above). In contrast, catalase
treatment yields only little if any formic acid. This very
surprising observation was not further investigated.
Oxidation of Sulfides and Disulfides by Reactive Hydro-
peroxides. Reactive hydroperoxides such as formic peracid and
hydroperoxides of a similar reactivity are capable of oxidizing
sulfides and disulfides by O-transfer (cf. reaction 53).
HC(O)OOH + R2S f HC(O)O- + H+ + R2SO (53)
The reaction with formic peracid can be readily followed by
conductance measurements, because formate ions and protons
are liberated in this reaction (pKa(formic peracid) ) 7.1, pKa-
(formic acid) ) 3.8). The rate constant with the sulfide is
considerably faster than that of the disulfide (cf. Figure 8 and

7250 J. Phys. Chem. B, Vol. 107, No. 30, 2003
Flyunt et al.
TABLE 2: Reduction Potentials at pH 7 of Various Substrates,74 Their Difference against the Reduction Potential of Ozone
•
(1.01 V), Formaldehyde Yields in the Presence of Tertiary Butanol, and the Yields of OH Calculated Therefrom
substrate
E7/V
∆E
CH2O yield/%
OH yield/%
N,N,N′,N′-tetramethylphenylenediamine
o-phenylenediamine
hydroquinone
0.24
0.34
0.45
0.53
0.77
0.77
0.67
0.56
0.48
0.34
0.15
0.15
34
14
20
11
see text
30 ( 5
43 ( 7
24 ( 4
∼70a
catechol
Fe(CN)6
3-
N,N-diethylaniline
phenol
0.86
0.86
13
11
28 ( 5
24 ( 4
15
triethylamine6
1.04
1,3,5-trimethoxybenzene
(1.23)b
(1.27)b
1.28
(0.22)b
(0.26)b
0.27
0.28
0.29
0.43
0.55
0.61
0.99
7.4
16 ( 2.5
iodide
<0.5
0
8.4
6.8
20
0
0
guanosine49
1,4-dimethoxybenzene
1,2-dimethoxybenzene
adenosine
1.2981
1.3082
1.4482
1.5681
1.6282
2.00
18 ( 3
15 ( 2
43 ( 7
9 ( 1.5
0
anisole
bromide
4.1
0
a From experiments with DMSO. b Based on values in acetonitrile of 1.49 V vs SCE83 and 1.25 V vs a silver electrode.84
reduction potentials, a value of E7 ) 1.8 V is cited.74 This value
halide ions, bromide and iodide. Kinetic evidence for such an
intermediate has been found with bromide which does not react
very fast with ozone (reactions 57 and 58). Reaction 57 is close
to isoenergetic, and thus the first step is even reversible.50
had been calculated75 by taking pKa(HO3 ) ) 10.4 from ref 76.
•
Considering that the pKa value of •OH is 11.910 and that of HO2
•
is 4.8,16 it is very unlikely that the pKa of the species richer by
one oxygen atom, HO3 , should be as high as 10.4. Also for the
•
Br- + O2 H BrOOO- f BrO- + O2(1∆g) (57/58)
corresponding parents, an increasing acidity H2O < H2O2 <
H2O3 has been calculated.77 That the pKa of HO3• must indeed
be low has already been discussed by Czapski.78 It is thus more
reasonable to take the value E(O3•-/O3) ) 1.01 V measured79
at high pH also for experiments at pH 7. Thus, ozone is a rather
weak one-electron oxidant. This low reduction potential does
not allow the conclusion80 that the Criegee mechanism for the
reaction of ozone with olefins has to be revised and proceeds
by an electron transfer in the first step, although the rate
constants of the ozone reaction and the gas-phase ionization
potentials of selected (electron rich) olefins correlate.80 This
correlation only reflects the electrophilic nature of ozone.
For a number of substrates, the formaldehyde yields in the
presence of an excess tertiary butanol, the •OH yields calculated
(see above) from theses data, together with the substrate
reduction potentials at pH 7 are given in Table 2.
Of the substrates compiled in Table 2, the compound with
the lowest reduction potential is N,N,N′,N′-tetramethylphen-
ylenediamine. Although the reduction potential difference is as
much as 0.77 V, and thus the only reaction to take place has
been considered85 to be electron transfer, it is not. As is evident
from the 9% yield of singlet dioxygen, O2(1∆g), O-transfer
competes.45 A short-lived (zwitterionic) adduct to nitrogen is
the likely precursor for both reaction channels. A tertiary amine
with a higher reduction potential is triethylamine. Here, an
adduct has also been assumed to be the primary product
(reaction 54), but now the route to the N-oxide and singlet
dioxygen (reaction 55) is the preferred pathway, and electron
transfer (reaction 56) occurs with a much lower yield (Table
2).6
The reaction of I- with ozone is, however, very fast (k )
1.2 × 109 M-1 s-1),50 and the endothermic electron transfer
does not compete effectively (cf. Table 2) with O-transfer
(reactions 59 and 60).
I- + O3 f [IOOO-] f IO- + O2(1∆g) (59/60)
Support for a certain lifetime of such adducts comes from
the O2(1∆g) yields in these systems. While with other systems,
e.g. various sulfur- and nitrogen-containing compounds, O2(1∆g)
yields are 100% when an O-transfer occurs; it is only 54% in
the case of Br- and 14% in the case of I-.45 This low O2(1∆g)
yield is readily explained if the adducts (IOOO- and BrOOO-)
are sufficiently long-lived to allow heavy-atom-assisted spin
conversion and release of triplet (ground-state) dioxygen. This
effect would be more pronounced with I- than with the lighter
Br-, as observed.
The Criegee mechanism, followed by olefins, also involves
a short-lived zwitterion, and there is now experimental evidence
that under favorable conditions, e.g. in the case of maleic and
fumaric acids, the zwitterion reverts again, thereby producing
also the other isomer.86
The preference of forming adducts rather than undergoing
outer-sphere electron transfer is also given by the much stronger
oxidant, the •OH radical (E7 ) 2.73 V),74 and even in the very
few cases, where electron transfer is observed (e.g., ref 87),
short-lived adducts are not fully excluded.
It is reasonable to assume that short-lived trioxides are also
intermediates in other O-transfer reactions and that they compete
with electron transfer whenever this reaction is thermodynami-
cally possible, e.g. in the case of aqua-Fe2+ (reaction 61, k )
8.5 × 105 M-1 s-1).88
N(CH2CH3)3 + O3 f +N(CH2CH3)3OOO-
(54)
+N(CH2CH3)3OOO- f (CH2CH3)3NO + O2(1∆g) (55)
Fe2+ + O3 f FeO2+ + O2
(61)
+N(CH2CH3)3OOO- f (CH2CH3)3N•+ + O3
(56)
•-
In this context, it may be worth mentioning that short-lived
adducts are common in ozone chemistry. Cases in point are the
Although the reduction potential of E°(Fe3+/Fe2+) ) 0.77 V
would allow the electron-transfer reaction to occur in acid

Ozone Reactions in Aqueous Solution
J. Phys. Chem. B, Vol. 107, No. 30, 2003 7251
methyl radical from reaction 27 with dioxygen), can induce this
chain reaction albeit with a much slower rate.44 We tentatively
suggest that the observed partial conversion of methanesulfinic
acid into methanesulfonic acid is due to an methylperoxyl-
induced autoxidation. The formation of high yields of meth-
anesulfonic acid besides methanesulfinic acid seems to be quite
general, and this is also observed when probing with DMSO
•OH-formation in the phenol/phenolate system.90
Use of Tetranitromethane for the Detection of Superoxide
in Ozone Reactions. When the ozone rate constant with the
substrate is much lower than that with NF-, the TNM assay
when only based on the formation of NF- may underestimate
the O2 yield because of a consumption of NF- by ozone.
•-
Figure 9. Ozonolysis of Fe(CN)64- (1 mM) in the presence of DMSO
(10 mM). Formation of methanesulfonic acid (b) and methanesulfinic
acid (O) as a function of the ozone concentration.
Raising the substrate concentration to account for this may not
be always possible due to solubility limitations. To minimize
the reaction of ozone with O2•-, TNM should be present at a
∼10-fold concentration with respect to ozone. At a TNM
concentration of 1 mM (close to the solubility limit), the ozone
concentration should be kept below 10-4 M.
solution (pH e 4), there is no indication that it takes place to
a noticeable extent.88 In agreement with this conclusion, in the
presence of an excess of DMSO the yield of methanesulfinic
•-
Often, it does not matter, whether TNM reacts with O2 or
•
acid and thus the OH yield is only 10 ( 5% at pH 4. These
a reducing radical that is formed in the course of an ozone
reaction. A case in point is the triethylamine system,6 where
NF- may be formed directly by the reaction of TNM with
reducing carbon-centered radical (cf. reactions 56, 61, and 63)
or at the later stage, when dioxygen is converted by reducing
carbon-centered radical into O2•- which further reacts with TNM
(reactions 64 and 9).91,92
measurements are fraught with a considerable error, since Fe2+
2-
and SO4 cause some problems in the determination of
methanesulfinic acid by ion chromatography, and it was
impossible to extend these measurements to lower pH values.
However, in contrast to aqua-Fe2+, Fe(CN)6 cannot ac-
4-
commodate a further ligand and has to react by electron transfer
and/or by an oxidation of the ligand. The rate of reaction of
4-
(CH3CH2)3N•+ f (CH3CH2)2N-CH•-CH3 + H+ (62)
Fe(CN)6 with ozone was determined by competition with
indigosulfonate using the stopped-flow setup for rapid mixing
of an ozone solution with a solution containing Fe(CN)64- and
indigosulfonate at varying concentration ratios. Based on
(CH3CH2)2N-CH•-CH3 + TNM f
(CH3CH2)2N+dCH-CH3 + NF- + •NO2 (63)
k(O3+indigosulfonate) ) 9.4 × 107 M-1 s-1 89
,
we arrived at
an ozone rate constant with Fe(CN)6 of 1.2 × 106 M-1 s-1
(data not shown), a value close to that for aqua-Fe2+. When
ozone is reacted with Fe(CN)64-, Fe(CN)63- is formed in 140%
yield (as determined by UV spectrophotometry at 420 nm, ꢀ )
1020 M-1 cm-1). Addition of DMSO halves this value, and
methanesulfinic plus methanesulfonic acid are formed in a total
yield of ∼60% (Figure 9).
4-
(CH3CH2)2N-CH•-CH3 + O2 f
(CH3CH2)2N+dCH-CH3 + O2 (64)
•-
There is, however, also the possibility that O2•- is a primary
product in an ozone reaction (reaction 65),93 and this reaction
is not related to an electron-transfer reaction.
Reaction 27 has an efficiency of 92%. Correcting for this
brings the •OH yield as determined by the DMSO assay to 65%.
This value is in agreement with a value of ∼70% as concluded
from the oxidation of Fe(CN)64-. Thus, electron transfer with
concomitant formation of •OH occurs in this system to an extent
of only 65-70%. In competition to electron transfer, the ligand
may have been oxidized, but experimental evidence for this is
still lacking. The ratio of methanesulfinic acid to methane-
sulfonic acid remained constant when the DMSO concentration
was varied between 0.01 and 0.1 M. We have recently shown
that methanesulfinic acid reacts very fast with ozone (k ) 2 ×
O3 + RH f O2•- + RO• + H+
(65)
There has never been an example for this type of reaction.
Adenine and its derivatives react very slowly with ozone
(k(Ado+O3) ) 12 M-1 s-1 49 and is known94,95 to produce •OH
)
in remarkable yields49 despite its high reduction potential (Table
2). It is thus a potential candidate for undergoing reaction 65.
•
•-
The OH that were detected would then have O2 as the
precursor.
•
106 M-1 s-1), but it also undergoes a very fast OH-induced
We have now set up an experiment where the concentration
of adenosine was ∼10 times higher than that of TNM in order
to protect TNM from being attacked by ozone. The TNM
concentration was chosen as high as possible, and the ozone
concentration was kept as low as possible in order to prevent
chain reaction in the presence of dioxygen.44 It has the
uncommon peculiarity that its chain length does not depend on
the substrate (methanesulfinic acid) concentration. It can be
readily shown by computer-assisted modeling that under the
prevailing experimental conditions the formation of methane-
sulfonic acid cannot be due to a reaction of methanesulfinic
acid with ozone to a significant extent. The fact that the
methanesulfinic to methanesulfonic acid ratio was not affected
much by the DMSO concentration when varied from 0.01 to
0.1 M also excludes an •OH-induced chain reaction as a cause
of the partial conversion of methanesulfinic acid into methane-
sulfonic acid. There was also an indication that other radicals,
e.g. the methylperoxyl radical (formed in the reaction of the
•-
O2 from reacting with ozone to a major extent. Moreover,
tertiary butanol was added at a ∼100-fold excess over adenosine
•
to ensure that OH would react with tertiary butanol. Experi-
mental details and the data for the key products nitrate and
nitroform anion (for O2•-) and formaldehyde (for •OH) are given
in Table 3. The 2-hydroxy-2-methylpropanal yield was also
determined.
The nitrate, nitroform anion, and formaldehyde yields in this
system have been modeled on the basis of a series of equations

7252 J. Phys. Chem. B, Vol. 107, No. 30, 2003
Flyunt et al.
(15) Elliot, A. J.; McCracken, D. R. Radiat. Phys. Chem. 1989, 33, 69-
(16) Bielski, B. H. J.; Cabelli, D. E.; Arudi, R. L.; Ross, A. B. J. Phys.
TABLE 3: Product Yields (%) in the Reactions of
Adenosine in the Presence of Tertiary Butanol and TNM of
Ozone Consumed ([adenosine] ) 10 mM, [tertiary butanol]
) 0.95 M, and [TNM] ) 1.1 mM; for Calculations, See
Text)
74.
Chem. Ref. Data 1985, 14, 1041-1100.
(17) Hoigne´, J. In Process Technologies for Water Treatment; Stucki,
S., Ed.; Plenum: New York, 1988; pp 121-143.
(18) Elovitz, M. S.; von Gunten, U. Ozone: Sci. Eng. 1999, 21, 239-
260.
calcd
calcd
•-
products
exptl (43% •OH) (43% O2
)
(19) Yao, C. C. D.; Haag, W. R. Water Res. 1991, 25, 761-773.
(20) Hoigne´, J.; Bader, H. Water Res. 1983, 17, 173-183.
(21) Veltwisch, D.; Janata, E.; Asmus, K.-D. J. Chem. Soc., Perkin
Trans. 2 1980, 146-153.
(22) Asmus, K.-D.; Henglein, A.; Ebert, M.; Keene, J. P. Ber. Bunsen-
Ges. Phys. Chem. 1964, 68, 657-663.
nitrate
nitroform anion
formaldehyde
43
1
5.3
45
0.2
7.6
56
6.8
0.2
2-hydroxy-2-methylpropionaldehyde 16
-
-
(23) von Sonntag, C.; Schuchmann, H.-P. In Peroxyl Radicals; Alfassi,
Z. B., Ed.; Wiley: Chichester, U.K., 1997; pp 173-234.
(24) Ragnar, M.; Eriksson, T.; Reitberger, T. Holzforschung 1999, 53,
292-298.
using the rate constants determined above. From the experiments
in the absence of TNM, it was concluded that in this system
•OH are generated in ∼43% yield (Table 2). As mentioned
above, these could have been formed directly by electron transfer
and subsequent decay of O3•- or via O2•- and its reaction with
ozone (which gives O3•-; see above) that may be produced in
the not yet established reaction 65. As is seen from Table 3,
(25) Forni, L.; Bahnemann, D.; Hart, E. J. J. Phys. Chem. 1982, 86,
255-259.
(26) Hart, E. J.; Sehested, K.; Holcman, J. Anal. Chem. 1983, 55, 46-
49.
(27) Nash, T. Biochem. J. 1953, 55, 416-421.
(28) Lipari, F.; Swarin, S. J. J. Chromatogr. 1982, 247, 297-306.
(29) Olson, K. L.; Swarin, S. J. J. Chromatogr. 1985, 333, 337-347.
(30) Babbs, C. F.; Gale, M. J. In Free Radicals, Methodology and
Concepts; Rice-Evans, C., Halliwell, B., Eds.; Richelieu Press: London,
1988; pp 91-121.
the simulations based on OH (O3•-) as the primary product
•
are in very reasonable agreement with the experimental data in
•-
contrast to O2 as the primary species. This rules out a
•
significant contribution of reaction 65 to the OH yield in this
(31) Weeke, F.; Bastian, E.; Schomburg, G. Chromatographia 1974, 7,
163-170.
system. We thus conclude that ozone reacts with adenosine by
electron transfer despite the fact that the reaction is endothermic.
The very low rate constant of this reaction; i.e. the absence of
other very efficient competing reactions allows this reaction to
proceed nevertheless. We have, therefore, not yet an example
for reaction 65.
(32) Asmus, K.-D.; Mo¨ckel, H.; Henglein, A. J. Phys. Chem. 1973, 77,
1218-1221.
(33) Erben-Russ, M.; Bors, W.; Saran, M. Int. J. Radiat. Biol. 1987,
52, 393-412.
(34) Gilbert, B. C.; Marshall, P. D. R.; Norman, R. O. C.; Pineda, N.;
Williams, P. S. J. Chem. Soc., Perkin Trans. 2 1981, 1392-1400.
(35) Neta, P.; Huie, R. E.; Ross, A. B. J. Phys. Chem. Ref. Data 1990,
19, 413-513.
In Table 3, the yield of 2-hydroxy-2-methylpropanal is also
given. It is seen that the ratio of the 2-hydroxy-2-methylpropanal
to formaldehyde yields is above the value of ∼1.5 typical for
the pure tertiary butanol system. This is an indication that the
adenosine-derived radicals also interact with the tertiary-butanol-
derived ones, as already suggested above. As a result, the
formaldehyde yield is lower than expected, and, in principle,
(36) Schuchmann, H.-P.; von Sonntag, C. Z. Naturforsch. 1984, 39b,
217-221.
(37) Schuchmann, M. N.; von Sonntag, C. J. Phys. Chem. 1979, 83,
780-784.
(38) Zegota, H.; Schuchmann, M. N.; von Sonntag, C. J. Phys. Chem.
1984, 88, 5589-5593.
(39) Bothe, E.; Behrens, G.; Schulte-Frohlinde, D. Z. Naturforsch. 1977,
32b, 886-889.
•
the OH yield might be even somewhat higher than the value
(40) Lind, J.; Mere´nyi, G.; Johansson, E.; Brinck, T. J. Phys. Chem. A
2003, 107, 676-681.
given in Table 2.
(41) Reisz, E.; Schmidt, W.; Schuchmann, H.-P.; von Sonntag, C.
EnViron. Sci. Technol., in press.
Acknowledgment. E.M. and E.R. would like to thank the
German Academic Exchange Service (DAAD) for stipends. The
continuing support by Prof. Dr. R. Mehnert, enabling us to
finalize the work at the IOM, is highly appreciated.
(42) Sehested, K.; Holcman, J.; Bjergbakke, E.; Hart, E. J. J. Phys. Chem.
1987, 91, 2359-2361.
(43) Babbs, C. F.; Gale Steiner, M. Methods Enzymol. 1990, 186, 137-
147.
(44) Flyunt, R.; Makogon, O.; Schuchmann, M. N.; Asmus, K.-D.; von
Sonntag, C. J. Chem. Soc., Perkin Trans. 2 2001, 787-792.
(45) Mun˜oz, F.; Mvula, E.; Braslavsky, S. E.; von Sonntag, C. J. Chem.
Soc., Perkin Trans. 2 2001, 1109-1116.
(46) Schmidt, E. Ber. Dtsch. Chem. Ges. 1919, 52, 400-413.
(47) Henglein, A.; Jaspert, J. Z. Phys. Chem. (Munich) 1957, 12, 324-
334.
References and Notes
(1) Ellis, K. V. Crit. ReV. EnViron. Control 1991, 20, 341-407.
(2) von Gunten, U. Water Res. 2003, 37, 1443-1467.
(3) von Gunten, U. Water Res. 2003, 37, 1469-1487.
(4) Criegee, R. Angew. Chem. 1975, 87, 765-771.
(5) Dowideit, P.; von Sonntag, C. EnViron. Sci. Technol. 1998, 32,
1112-1119.
(6) Mun˜oz, F.; von Sonntag, C. J. Chem. Soc., Perkin Trans. 2 2000,
2029-2033.
(7) Hoigne´, J.; Bader, H. In Proceedings of the 8th Ozone World
Congress September 1987, Zu¨rich; Nauf, H. R., Ed.; 1987; pp K83-K97.
(8) von Sonntag, C.; Mark, G.; Mertens, R.; Schuchmann, M. N.;
Schuchmann, H.-P. J. Water Supply Res. Technol.-Aqua 1993, 42, 201-
211.
(9) Neta, P.; Huie, R. E.; Ross, A. B. J. Phys. Chem. Ref. Data 1988,
17, 1027-1284.
(10) Buxton, G. V.; Greenstock, C. L.; Helman, W. P.; Ross, A. B. J.
Phys. Chem. Ref. Data 1988, 17, 513-886.
(11) Staehelin, J.; Hoigne´, J. EnViron. Sci. Technol. 1982, 16, 676-
681.
(12) Sehested, K.; Corfitzen, H.; Holcman, J.; Hart, E. J. J. Phys. Chem.
1992, 96, 1005-1009.
(13) Bu¨hler, R. E.; Staehelin, J.; Hoigne´, J. J. Phys. Chem. 1984, 88,
2560-2564.
(48) Hoigne´, J.; Bader, H.; Haag, W. R.; Staehelin, J. Water Res. 1985,
19, 993-1004.
(49) Theruvathu, J. A.; Flyunt, R.; Aravindakumar, C. T.; von Sonntag,
C. J. Chem. Soc., Perkin Trans. 2 2001, 269-274.
(50) Liu, Q.; Schurter, L. M.; Muller, C. E.; Aloisio, S.; Francisco, J.
S.; Margerum, D. W. Inorg. Chem. 2001, 40, 4436-4442.
(51) Barb, W. G.; Baxendale, J. H.; George, P.; Hargrave, K. R. J. Chem.
Soc., Faraday Trans. 1951, 47, 462-500.
(52) Hart, E. J.; Henglein, A. J. Phys. Chem. 1985, 89, 4342-4347.
(53) Flyunt, R.; Theruvathu, J. A.; Leitzke, A.; von Sonntag, C. J. Chem.
Soc., Perkin Trans. 2 2002, 1572-1582.
(54) Mertens, R.; von Sonntag, C.; Lind, J.; Mere´nyi, G. Angew. Chem.,
Int. Ed. Engl. 1994, 33, 1259-1261.
(55) Ko¨ster, R.; Asmus, K.-D. Z. Naturforsch. 1971, 26b, 1108-1116.
(56) Latif, N.; O’Neill, P.; Schulte-Frohlinde, D.; Steenken, S. Ber.
Bunsen-Ges. Phys. Chem. 1978, 82, 468-471.
(57) Manogue, W. H.; Pigford, R. L. AIChE J. 1960, 6, 494-500.
(58) Leitzke, A.; Reisz, E.; Flyunt, R.; von Sonntag, C. J. Chem. Soc.,
Perkin Trans. 2 2001, 793-797.
(14) Sehested, K.; Holcman, J.; Hart, E. J. J. Phys. Chem. 1983, 87,
1951-1954.
(59) Mihaljevic, B.; Katusin-Razem, B.; Razem, D. Free Radicals Biol.
Med. 1996, 21, 53-63.

Ozone Reactions in Aqueous Solution
J. Phys. Chem. B, Vol. 107, No. 30, 2003 7253
(60) Wagner, J. R.; Berger, M.; Cadet, J.; van Lier, J. E. J. Chromatogr.
1990, 504, 191-196.
(78) Czapski, G. Annu. ReV. Phys. Chem. 1971, 22, 171-208.
(79) Kla¨ning, U. K.; Sehested, K.; Holcman, J. J. Phys. Chem. 1985,
89, 760-763.
(61) Mihaljevic, B.; Medved-Rogina, B.; Razem, D. Radiat. Phys. Chem.
1999, 55, 593-597.
(80) Zhang, X.-M.; Zhu, Q. J. Org. Chem. 1997, 62, 5934-5938.
(81) Steenken, S.; Jovanovic, S. V. J. Am. Chem. Soc. 1997, 119, 617-
618.
(62) Phulkar, S.; Rao, B. S. M.; Schuchmann, H.-P.; von Sonntag, C.
Z. Naturforsch. 1990, 45b, 1425-1432.
(63) Paul, H.; Small, R. D.; Scaiano, J. C. J. Am. Chem. Soc. 1978,
100, 4520-4527.
(82) Jonsson, M.; Lind, J.; Reitberger, T.; Eriksen, T. E.; Mere´nyi, G.
J. Phys. Chem. 1993, 97, 11278-11282.
(64) Ellison, D. H.; Salmon, G. A.; Wilkinson, F. Proc. R. Soc. London
1972, A328, 23-36.
(83) Zweig, A.; Hodgson, W. G.; Jura, W. H. J. Am. Chem. Soc. 1964,
86, 4124-4129.
(65) Berdnikov, V. M.; Bazhin, N. M.; Fedorov, V. K.; Polyakov, O.
V. Kinet. Catal. (Engl. Transl.) 1972, 13, 986-987.
(66) Gilbert, B. C.; Holmes, R. G. G.; Laue, H. A. H.; Norman, R. O.
C. J. Chem. Soc., Perkin Trans. 2 1976, 1047-1052.
(67) Schuchmann, H.-P.; von Sonntag, C. J. Photochem. 1981, 16, 289-
295.
(84) Hu, K.; Niyatzymbetov, M. E.; Evans, D. H. Tetrahedron Lett. 1995,
36, 7027-7030.
(85) Cataldo, F. Eur. Polym. J. 2002, 38, 885-893.
(86) Leitzke, A.; Naumov, S.; von Sonntag, J.; von Sonntag, C.
Unpublished results.
(87) Fang, X.; Schuchmann, H.-P.; von Sonntag, C. J. Chem. Soc., Perkin
Trans. 2 2000, 1391-1398.
(68) Ulanski, P.; Mere´nyi, G.; Lind, J.; Wagner, R.; von Sonntag, C. J.
Chem. Soc., Perkin Trans. 2 1999, 673-676.
(88) Jacobsen, F.; Holcman, J.; Sehested, K. Int. J. Chem. Kinet. 1996,
29, 17-24.
(69) Frew, J. E.; Jones, P.; Scholes, G. Anal. Chim. Acta 1983, 155,
139-150.
(89) Mun˜oz, F.; von Sonntag, C. J. Chem. Soc., Perkin Trans. 2 2000,
661-664.
(70) Allen, A. O.; Hochanadel, C. J.; Ghormley, J. A.; Davis, T. W. J.
Phys. Chem. 1952, 56, 575-586.
(90) Mvula, E.; von Sonntag, C. Unpublished results.
(91) Das, S.; von Sonntag, C. Z. Naturforsch. 1986, 41b, 505-513.
(92) Das, S.; Schuchmann, M. N.; Schuchmann, H.-P.; von Sonntag,
C. Chem. Ber. 1987, 120, 319-323.
(71) Flyunt, R.; Leitzke, A.; von Sonntag, C. Radiat. Phys. Chem., in
press.
(72) Eisenberg, G. M. Ind. Eng. Chem., Anal. Ed. 1943, 15, 327-328.
(73) Uher, G.; Gilbert, E.; Eberle, S. H. Vom Wasser 1991, 76, 225-
234.
(74) Wardman, P. J. Phys. Chem. Ref. Data 1989, 18, 1637-1755.
(75) Koppenol, W. H. FEBS Lett. 1982, 140, 169-172.
(76) Buxton, G. V. Faraday Trans. 1969, 65, 2150-2158.
(77) Koller, J.; Plesnicar, B. J. Am. Chem. Soc. 1996, 118, 2470-2472.
(93) Ragnar, M.; Eriksson, T.; Reitberger, T.; Brandt, P. Holzforschung
1999, 53, 423-428.
(94) Ishizaki, K.; Shinriki, N.; Ueda, T. Chem. Pharm. Bull. 1984, 32,
3601-3606.
(95) Van der Zee, J.; Dubbelman, T. M. A. R.; Van Steveninck, J. Free
Radical Res. Commun. 1987, 2, 279-284.