A new approach for studying aqueous phase OH kinetics: Application of Teflon waveguides

Article abstract of DOI:10.1039/b301145n

Bimolecular rate coefficients for the reactions of the hydroxyl radical, OH, with methanol, ethanol, tetrahydrofuran, dimethylmalonate [CH₃OC(O)CH₂C(O)OCH₃], dimethylsuccinate [CH₃OC(O)CH₂CH₂C(O)OCH₃], dimethylcarbonate [CH₃OC(O)OCH₃] and diethylcarbonate [CH₃CH₂OC(O)OCH₂CH₃] in aqueous solutions have been measured using a novel experimental approach. The centrepiece of the new experimental technique reported in this work is a Teflon AF 2400 liquid core waveguide. The physical properties of the Teflon AF 2400 liquid core waveguide allow for the construction of a micro-flowtube reaction photolysis cell with an extremely low volume and a very long optical pathway. Such a reaction system allows for a very sensitive detection of chemical transients in the aqueous phase. The micro-flowtube experiments involved competition kinetics of the OH radical with the organic reactant of interest and an SCN^- anion, (k_OH+SCN- = 1.29 × 10¹⁰ M^-1 s^-1). ¹The (SCN)^-₂ anion was detected using UV-visible spectroscopy following a medium pressure mercury lamp photolysis (λ ≥ 366 nm) of H₂O/H₂O₂/reactant/KSCN mixtures. All experiments were carried out at room temperature. Measured rate coefficients for the reaction of the OH radical with methanol, ethanol, tetrahydrofuran, dimethylmalonate, dimethylsuccinate, dimethylcarbonate and diethylcarbonate are (units are 10⁸ M^-1 s^-1): k_OH+methanol = 13±4, k_OH+ethanol = 19±5, k_OH+THF = 38±10, k_{OH+dimethylmalonate} = 2.7±0.9, k_{OH+dimethylsuccinate} = 5.3±2.9, k_{OH+dimethylcarbonate} = 0.51±0.22, k_{OH+diethylcarbonate} = 7.9±3.2. Uncertainties in the above expressions are ±2σ and represent precision only. The reported rate coefficients for the reactions of OH with ethanol, methanol and THF agree very well with the currently recommended values. To date, there is no kinetic data reported in the literature for the OH radical reaction with dimethylmalonate, dimethylsuccinate, dimethylcarbonate and diethylcarbonate. The reaction mechanism is briefly discussed as a function of bond energies.

Full text of DOI:10.1039/b301145n

View Article Online / Journal Homepage / Table of Contents for this issue
A new approach for studying aqueous phase OH kinetics: application
of Teflon waveguides
Christian George,* Davy Rousse, Emilie Perraudiny and Rafal Strekowski
Laboratoire d’Application de la Chimie a` l ’Environnement (UCBL-CNRS) 43 boulevard du 11
Novembre 1918, F-69622, Villeurbanne, France. E-mail: christian.george@univ-lyon1.fr;
Fax: +33 (0)4 72 44 8114; Tel: +33 (0)4 72 43 1489
Received 28th January 2003, Accepted 26th February 2003
First published as an Advance Article on the web 12th March 2003
Bimolecular rate coefficients for the reactions of the hydroxyl radical, OH, with methanol, ethanol,
tetrahydrofuran, dimethylmalonate [CH OC(O)CH C(O)OCH ], dimethylsuccinate
3
2
3
[
[
CH
3
OC(O)CH
2
CH
2
C(O)OCH
3
], dimethylcarbonate [CH
CH CH OC(O)OCH CH ] in aqueous solutions have been measured using a novel experimental approach.
3
3
OC(O)OCH ] and diethylcarbonate
3
2
2
3
The centrepiece of the new experimental technique reported in this work is a Teflon AF 2400 liquid core
waveguide. The physical properties of the Teflon AF 2400 liquid core waveguide allow for the construction of a
micro-flowtube reaction photolysis cell with an extremely low volume and a very long optical pathway. Such a
reaction system allows for a very sensitive detection of chemical transients in the aqueous phase. The
micro-flowtube experiments involved competition kinetics of the OH radical with the organic reactant of interest
ꢀ
10
ꢀ1 ꢀ1
1
ꢀ
and an SCN anion, (kOH+SCN
ꢀ
¼ 1.29 ꢁ 10
M
s
). The (SCN)
2
anion was detected using UV-visible
O/H /reactant/
spectroscopy following a medium pressure mercury lamp photolysis (l ꢂ 366 nm) of H
2
2 2
O
KSCN mixtures. All experiments were carried out at room temperature. Measured rate coefficients for the
reaction of the OH radical with methanol, ethanol, tetrahydrofuran, dimethylmalonate, dimethylsuccinate,
8
ꢀ1 ꢀ1
s
dimethylcarbonate and diethylcarbonate are (units are 10 M
kOH+ethanol ¼ 19 ꢃ 5, k ¼ 38 ꢃ 10, k
): kOH+methanol ¼ 13 ꢃ 4,
¼ 2.7 ꢃ 0.9, k
¼ 5.3 ꢃ 2.9,
OH+THF
OH+dimethylmalonate
OH+dimethylsuccinate
k
OH+dimethylcarbonate ¼ 0.51 ꢃ 0.22, kOH+diethylcarbonate ¼ 7.9 ꢃ 3.2. Uncertainties in the above expressions are
ꢃ2s and represent precision only. The reported rate coefficients for the reactions of OH with ethanol, methanol
and THF agree very well with the currently recommended values. To date, there is no kinetic data reported
in the literature for the OH radical reaction with dimethylmalonate, dimethylsuccinate, dimethylcarbonate
and diethylcarbonate. The reaction mechanism is briefly discussed as a function of bond energies.
obtain ‘‘new’’ environmentally friendly solvents. However,
Introduction
the atmospheric degradation of dibasic esters and carbonates
will potentially result in the formation of tropospheric ozone
and other secondary pollutants. Moreover, the C –C oxyge-
Both in the gas phase and in aqueous droplets (i.e., hydrome-
teors) the hydroxyl radical is the most important daytime oxi-
dizing agent responsible for the atmospheric degradation of
5
2
10
nated organics can potentially form low vapour pressure sec-
ondary reaction species that can be either taken up by clouds
2
most of the volatile organic compounds (VOCs). To date,
9
,10
the reaction kinetics of the OH radical toward a large number
of organic species have been reported (see, for example, the
readily available solution kinetic database ). Although many
or form secondary aerosols
(SOA). In turn, these com-
pounds can be removed from the atmosphere via rain, dry or
wet deposition and potentially act as a source of organic pol-
lutants in surface waters.
3
1 2
studies of the kinetics of the OH + C to C compounds or
OH + long chain hydrocarbons (i.e., of biological interest)
reactions are reported in the literature, very few studies have
been performed for the reactions of OH with the intermediate
sized aliphatic compounds (i.e., C –C10), and especially for the
2
reactions of OH with organics containing oxygenated poly-
functional groups.
To date, the atmospheric fate of C –C dibasic esters and
2
10
carbonates remains uncertain. In the atmosphere these com-
pounds are likely to be involved in multiphase chemistry and
be removed via the attack by OH. One large volume chamber
study of the reaction of the OH radical with the dibasic ester
CH OC(O)CH CH CH C(O)OCH , DBE-5, in the presence
3
2
2
2
3
The interest in the OH reactions mentioned above, stems
from the fact that the intermediate length carbon chain and
oxygenated hydrocarbon compounds are expected to be
increasingly used as new industrial solvents to favour water
of nitric oxide suggests that organic nitrate(s), mono-methyl
glutarate, [CH OC(O)CH CH CH C(O)OH and dimethyl
3
2
2
2
1,3-acetonedicarboxylate account for 77 ꢃ 27% of the total
1
1
reaction products and reaction pathways. In this paper, we
present the results of the rate coefficient measurements for
the liquid phase reactions of the OH radical with several oxy-
genated compounds of atmospheric interest using a novel
experimental technique. This new technique takes advantage
of the pioneering waveguide technology where the reaction cell
is a micro-flowtube made of Teflon AF. To our knowledge,
this is the first study to use the new Teflon AF liquid core
4
–6
based technology. Currently, C –C dibasic esters and car-
2
10
7
bonates are being considered to replace ‘‘old’’ additives to
,8
y Present address: Laboratoire de Physico-Chimie Mol e´ culaire (Uni-
versit e´ Bordeaux I/CNRS) 351 cours de la Lib e´ ration, 33405 Talence
Cedex, France. Fax: +33 (0)5 56 84 66 45; Tel.: +33 (0)5 56 84 28 68;
E-mail: e.perraudin@lpcm.u-bordeaux1.fr.
1
562
Phys. Chem. Chem. Phys., 2003, 5, 1562–1569
This journal is # The Owner Societies 2003
DOI: 10.1039/b301145n

View Article Online
waveguide technology to study liquid phase chemical kinetics
of atmospheric interest. Comparison of data using different
experimental techniques is important because each technique
is subject to different systematic errors.
Details of the experimental procedure that were employed to
study the reactions of OH with methanol, ethanol, tetrahydro-
furan, dimethylmalonate, dimethylsuccinate, dimethylcarbo-
nate and diethylcarbonate in aqueous solutions are given
below.
polymer is very porous its physical properties allow for trap-
ping of gases which in turn can be derivatized and products
2
analyzed using, e.g., UV-visible spectroscopy. Furthermore,
8
the Teflon AF LCW has also been applied in Raman and
fluorescence spectroscopies.
2
9,30
The experimental technique described below has been
inspired from investigations using waveguide technology
described above. To our best knowledge, this is the first work
to apply the liquid core waveguide as a type of photolysis reac-
tor to study kinetic reactions of atmospheric interest.
Again, the centrepiece of the experimental set-up is a liquid
core waveguide made of Teflon AF 2400 (from BioGeneral,
San Diego, CA). The Teflon AF 2400 guide had an inner
and outer diameter of 200 mm and 800 mm, respectively. The
length of the LCW was varied in the course of the study from
30 to 90 cm. In all cases, the internal volume of the liquid core
waveguide was less than 0.1 mL for a maximum optical path
length of 90 cm. Using very small liquid volumes in the
LCW is one of the major advantages of using waveguides com-
pared to a standard White cell design for studying radical
liquid phase reactions. A good White cell can achieve a maxi-
mum path length of a few meters (which often requires expen-
sive lasers) in a volume of ca. 50–100 mL. Such a large reaction
volume requires much larger flow rates to replenish the cell
between laser flashes. Under typical experimental conditions
employed, the LCW can achieve much longer optical path
lengths in much smaller volumes without requiring powerful
light sources as those required in standard White cell systems
for liquids. The application of the liquid core waveguide tech-
nology for laser flash photolysis studies will be presented
elsewhere.
Experimental
In order to study aqueous phase radical chemistry, a number
of precautions have to be taken. First, radical species must
be produced using the cleanest possible way (for example,
photo-Fenton reactions with iron oxides may generate poten-
1
2
tial interferences). Second, transient radical concentrations
have to be small enough in order to avoid unwanted complica-
tions from secondary chemistry (for example recombination
reactions). Such considerations have led over the years to the
development and validation of well established experimental
methods such as pulse radiolysis and laser flash photolysis
1
3–21
techniques.
In most cases, radical species produced using
the above techniques are coupled with UV-visible time
resolved spectroscopy.
Teflon AF waveguides
Often, when studying liquid phase reactions, when light passes
through a reaction cell between two White cell mirrors, energy
(or light) is lost on the windows. This problem may be over-
come by using long-path cells without windows.
A ‘‘windowless’’ reaction cell can be achieved by using a
liquid core waveguide (henceforth called LCW). For practical
reasons (discussed in the following section), the LCW must (i)
have a relatively small diameter, (ii) be flexible, (iii) be trans-
parent to UV radiation (used to produce OH radicals) and
Experimental approach
The Teflon waveguide photolysis system used for the reaction
of OH with methanol, ethanol, tetrahydrofuran, dimethylma-
lonate, dimethylsuccinate, dimethylcarbonate and diethylcar-
bonate is shown schematically in Fig. 1. The highly flexible
Teflon AF 2400 tubing was loosely coiled (ca. 4 cm diameter)
around a 3 cm od Pyrex tube. The radiation from a medium
pressure mercury UV lamp (Hereaus Noblelight, TQ150, 150
W) located within the Pyrex tube served as the photolytic
light source. The cell was maintained at a constant tempera-
ture (ꢃ0.5 K) by circulating water that was allowed to flow
from a thermostatically controlled cryostat bath (Huber
cc230). The temperature of the system (i.e., the Teflon coil
and its contents) was constantly monitored using a Pt100 sen-
sor. The Pt100 sensor/probe was placed in close contact with
the waveguide coil. It was assumed that the reaction cell tem-
perature was the same as the LCW and that no temperature
gradient existed. Further, experiments were carried out to
make sure that the mercury lamp did not heat the coil and
(
iv) have a refraction index lower than the one of water i.e.,
n ¼ 1.33. The latter condition is essential for the LCW in order
to allow total internal reflection and, therefore, conduction of
light.
To date, the only commercially available materials allowing
for the conduction of light in water at room temperature are
copolymers of tetrafluoroethylene and 2,2-bistrifluoromethyl-
2
2,23
4
,5-difluoro-1,3-dioxole sold by DuPont as Teflon AF.
Tubing made from these polymers has been shown to exhibit
excellent optical properties, such as high optical clarity (at
l ꢄ 200 nm more than 80% of light is transmitted through a
2
20 mm thick film of the polymer described above) and very
low refractive index (i.e., n ¼ 1.29 for Teflon AF 2400 grades).
The refractive index observed for the Teflon AF 2400 tube is
lower than virtually all standard temperature and pressure
liquids. In addition, the refractive index of n ¼ 1.29 for The
Teflon AF 2400 tube is close to the theoretical minimum for
2
4
an organic polymer as predicted by Groh and Zimmermann.
As a result, tubing made from the Teflon AF 2400 polymer
and filled with water will conduct light. It has been shown that
the optical attenuation for such tubing is less than 50% per
2
3
meter of tubing. Therefore, very long LCW lengths can be
used without large energy losses.
Capillary tubing made from Teflon AF polymer has been
applied previously as LCW. For example, the Teflon AF
2
400 LCW has been used to enhance sensitivity in UV-visible
spectroscopic measurements. As stated above, long lengths of
LCW can be used that result in a long optical pathway through
the liquid (to more than 25 m!) without large energy losses
(see above): the signal is improved and the noise remains the
same.
2
5,26
Teflon AF capillaries have also been applied in
HPLC and in gas analysis systems. Since the Teflon AF
2
7
Fig. 1 Schematics of the LCW micro-flowtube experimental setup.
Phys. Chem. Chem. Phys., 2003, 5, 1562–1569
1563

View Article Online
ꢀ2
its contents. We did not observe any heating due to the mer-
cury lamp since no increase in temperature was observed when
the lamp was switched on. However, the Pt100 sensor used in
The (SCN) anion spectra were recorded in the wavelength
range from 300 to 800 nm.
Given the experimental setup described above, the absorp-
tion spectra were integrated over the entire length of the
LCW (i.e., over time). However, FACSIMILE computer simu-
lations showed that the steady state conditions were reached
within a few milliseconds (i.e., a few millimetres) of the
LCW. This was below the observed experimental sensitivity
limit and found not to affect the kinetic determinations of the
rate coefficients and, therefore, ignored. We, therefore, believe
that the integrated spectra can be safely attributed to the
ꢅ
this work had a stated uncertainty of ꢃ0.15 C. Given the
uncertainty of the Pt100 sensor used in this work, we con-
servatively estimate that the uncertainty in the temperature in
the Teflon coil and its contents resulting from heating by the
ꢅ
lamp was ꢃ0.30 C. Therefore, we can safely state that
the temperature rise caused by the UV lamp was less than
0
ꢅ
.30 C.
A major fraction of the lamp output was filtered out by the
water cooling liquid from the cryostat and the mercury lamp
Pyrex housing. A fraction of the light (l < 366 nm) was trans-
mitted through the LCW and its contents. The OH radicals
were generated within the LCW following a mercury lamp
steady-state conditions achieved within the micro-flowtube.
The observed (SCN)
ꢀ
2
anion spectra were corrected for the
mercury lamp atomic lines trapped within the LCW during the
photolysis of H O .
2
2
3
1,32
photolysis of H
2
O
2
(F ¼ 1 at l ꢄ 366 nm).
A quantum
yield of F ¼ 1 at l ꢄ 366 nm was adopted to determine the
Competition kinetics
3
1,32
initial OH radical concentration.
However, the effective
quantum yield for the OH radical production in aqueous solu-
Since the hydroxyl radical absorbs weakly in the visible region
of the spectrum, it cannot be readily measured (at least when
other absorbing compounds are present). Therefore, the OH
radical concentration (or its temporal profile) has to be probed
indirectly. This is commonly achieved using competition reac-
tion kinetics.
tions may in fact be lower than 1. Thus, Zellner and coworkers
and Zellner and Herrmann have reported F
OH
¼ 0.98 and
3
3,34
0
.96 at l ¼ 308 and 351 nm, respectively.
Values for
much less than 1 have been reported by Warneck, refer-
OH
F
ring to the loss of H
3
1,32
2
O
2
.
However, the exact concentration
Indeed, the OH radical can react either with a reactant RH
a generic organic compound for which the OH kinetics is
of OH radicals is not very important for the outcome of this
work, where nothing more is required than a major production
of OH radicals.
(
unknown) or with the competitor (where the reaction coeffi-
cient is well established) where the reaction produces a transi-
ent species which absorbs strongly in the wavelength region of
interest. Accordingly, the concentration of the transient species
is then directly related to the number of OH radicals reacting
with the competitor species. If the first reaction is fast, then the
concentration of the absorbing species is low. Conversely, if
the first reaction is slow then OH will have time to react with
the competitor resulting in a high concentration of the absorb-
ing species.
The concentration of OH radicals in the aqueous phase was
not directly measured but derived from the ‘‘titration’’ reac-
tion resulting from a competition reaction of the OH radical
ꢀ
with the SCN anion (see below). Under the experimental
conditions employed, typical OH radical concentration was
ꢀ
8
ca. 10 mol L .
ꢀ1
The geometry of the experimental setup (see Fig. 1) was such
that it allowed for the reactant solution and the xenon lamp
light to enter perpendicular to one another through a T-con-
nector and the photolysing mercury lamp to be orthogonal
to the waveguide reactor coil.
The solution content of the LCW was analyzed using UV-
visible absorption spectroscopy (Andor Technology Fast
Kinetics CCD) located in series to the direction of the propa-
gation of both the solution and the xenon lamp light. The
liquids were continuously pumped into the LCW through
one entry of the T-connector, using a peristaltic pump
Among several potential competitors (for examples, see
Table 1), the thiocyanate anion has been chosen because the
ꢀ
product of its reaction with the OH radical [i.e., the (SCN)
anion] exhibits the strongest extinction coefficient (e475 nm
7
applied in this work is given below (R1–R5).
2
¼
ꢀ1
ꢀ1
1
600 M cm ). The particular competition reaction system
H2O2 þ hn ! 2OH
ðR1Þ
ðR2Þ
ðR3Þ
ðR4Þ
ðR5Þ
OH þ SCNꢀ ! OH þ SCN
ꢀ
(
Watson–Marlow, type 323). Typical flow rates were in the
ꢀ
1
range from 0.1 to 3.0 ml min . That is, the LCW was used
as a micro-flowtube. In all cases, the time needed for the liquid
to pass through the LCW was less than 2 s. As a result, quite
small irradiation times were achieved without expensive
and technically involving experimental requirements. Under
these experimental parameters, steady-state conditions were
obtained within the LCW and verified experimentally (see
below) but also theoretically using numerical simulations of
the chemistry within the LCW. Theoretical calculations were
ꢀ
ꢀ
2
SCN þ SCN 0 ðSCNÞ
ðSCNÞꢀ2 ! products
OH þ RH ! products
The reaction mechanism (R1)–(R4) has been studied by Chin
and Wine. These investigators report that the thiocyanate
1
ion is a reliable competitor species that may be used to study
OH radical reactions in the liquid phase.
1
3
5
performed using a Gear type solver (FACSIMILE ). All
known rate and photolysis coefficients were directly taken
from the solution kinetics database.
It must be underlined that this reaction mechanism is
already simplified based on the work of Chin and Wine. In
1
3
fact, reaction (R2) proceeds in several steps but the corre-
sponding details are not necessary for the competition
approach used in this work and are, therefore, ignored. In fact,
In order to probe the solution contents of the LCW tube, the
output of a 75 W xenon lamp was focused on the entry of a 100
mm diameter fused silica optical fibre. The fused silica optical
fibre was held in the liquid content of the LCW tube. Then,
the light was allowed to escape the solid optical fibre and col-
lected in the liquid core waveguide. Again, the liquid core
waveguide conducted the light up to its end where another
fused silica optical fibre (located in the liquid) collected most
of this transmitted light. This second solid optical fibre then
conducted the light to the entry of a spectrograph (Lot-Oriel,
these authors showed that the OH radical is quantitatively
converted into (SCN) which may then be used as an OH
ꢀ
2
radical concentration indicator. In fact, reaction (R3) is an
equilibrium reaction and hence the (SCN) anion is effectively
ꢀ
2
the product of reaction (R2). As a result, for a given experi-
anion concentration is proportional to
ꢀ
ment, the (SCN)
the hydroxyl radical concentration.
2
In our study we adopted the temperature dependent rate
coefficient for reaction (R2) reported by Chin and
1
27i) coupled to a CCD camera (Andor Technology). The
1
0
ꢀ1 ꢀ1
1
at 298 K). Also, the inher-
CCD allowed to monitor the evolution of the UV-visible spec-
trum as a function of wavelength and time within the LCW.
Wine (kR2 ¼ 1.29 ꢁ 10
M
ent sensitivity of waveguides allowed us to work with low
s
1
564
Phys. Chem. Chem. Phys., 2003, 5, 1562–1569

View Article Online
Table 1 Comparison of the room temperature rate coefficients for the
reaction of the OH radical with methanol obtained in this work with
All experiments were carried out under ‘‘slow flow’’ condi-
tions with a liquid linear flow rate through the micro-flowtube
reactor of about 1 m s . With this range of linear flow rates,
a
ꢀ1
literature values
the contents of the LCW were exposed to UV light only for
short times (less than 2 s) ensuring that all reactant concentra-
tions stayed constant, i.e., secondary chemistry and degrada-
ꢀ
1
ꢀ1
s
Technique
Competitor
k
5
/M
Ref.
2
ꢀ3
ꢀ
ꢀ
8
PR
PR
CO
8.8 ꢁ 10
7.8 ꢁ 10
1.0 ꢁ 10
1.0 ꢁ 10
9.5 ꢁ 10
8.3 ꢁ 10
8.3 ꢁ 10
9.7 ꢁ 10
1.0 ꢁ 10
9.7 ꢁ 10
36
36
37
38
39
38
38
40
41
42
ꢀ
8
9
9
8
8
8
8
9
8
tion of H
O
2
or SCN was negligible and, therefore, ignored.
SCN
SCN
2
All error ranges, reported below, correspond to 2s uncer-
tainties as simply derived from a linear fitting procedure of
eqn. (4) to our data. As no specific statistical treatment was
applied to our data (as Student’s t factor considerations), these
uncertainties are lower limits to the real uncertainty of our
measurements.
PR
PR
ꢀ
C
(CH
6
H
5
COO
6 4
NC H NO
PR
PR
3
)
2
Nitrobenzoate ion
ꢀ
PR
PR
6 5 2
C H CH COO
4
ꢀ
[Fe(CN) ]
6
2
ꢀ b
PR
ABTS
Recommended
value
Literature
review
PR
Reagents
ꢀ
9
8
9
SCN
1.2 ꢁ 10
8.3 ꢁ 10
1.3 ꢁ 10
43
44
Under the experimental conditions employed, typical hydro-
gen peroxide and thiocyanate concentrations (in units of mol
ꢀ
GR
MFT
C
SCN
6
H
5
ꢀ
COO
This
work
ꢀ1
ꢀ3
ꢀ4
L
the organic reagents used ranged from 10 to 10 mol L
) were 10 and 2 ꢁ 10 , respectively. Concentrations of
ꢀ4 ꢀ2 ꢀ1
a
PR: pulse radiolysis; GR: gamma radiolysis; LP: laser photolysis;
b
depending on the corresponding reactivities. The reagents used
in this study had the following stated minimum purities:
2ꢀ
0
MFT: micro-flowtube.
line-6-sulfonate).
ABTS : 2,2 -azinobis-(3-ethylbenzthiazo-
KSCN (Aldrich, > 99%); H
Merck, > 99.9%) ethanol (Prolabo, > 99.85%), tetrahydro-
furan (Aldrich, > 99.9%), dimethylmalonate (Aldrich, > 99%),
dimethylsuccinate (Aldrich, > 98%), dimethylcarbonate
2 2
O (Aldrich, > 30%), methanol
(
ꢀ
concentration of (SCN2) anions. Under typical experimental
ꢀ
(Fluka, > 99%) and diethylcarbonate (Fluka, > 99.5%) and
were used without any further purification. All solutions were
freshly prepared with non-degassed deionised water with resis-
tivity > 18 MO. Deionised water was prepared by passing tap
water through a reverse osmosis demineralization filter (ATS
Groupe Osmose) followed by a commercial deionizer (Milli-
pore, Milli-Q). Dissolved oxygen within the water solution
was shown not to be a problem and have any effect on the
results presented in this work. Experiments were carried out
in ‘‘regular’’ and degassed water. No difference in kinetics
was observed.
conditions employed, the (SCN
2
) anion concentration was ca.
mol L . As a result, the (SCN2) anions decayed by a
ꢀ
8
ꢀ1
ꢀ
1
first order process [see reaction (R4)]. At higher concentra-
0
1
ꢀ
ꢀ
tions of (SCN2) anions, the (SCN2) anion decay may be
governed by second order processes. However, the second
order process for the decay of the (SCN) anions was never
2
observed under the experimental conditions employed in this
work. In fact, Chin and Wine reported that the rate coefficient
ꢀ
ꢀ1 1
for reaction (R4) is 870 s , while for the second order decay
the rate coefficient was reported by these authors to be
9
ꢀ1 ꢀ1
1
at 298 K. Since the (SCN) anion con-
ꢀ
1
centrations were typically ca. 10 mol L , the recombination
.2 ꢁ 10 L mol
s
2
ꢀ
8
ꢀ1
ꢀ
2
of (SCN) was not expected to be an important pathway for
its removal. In the reaction mechanism (R1)–(R5), the
ion removal is believed to be governed by reaction
R4). This was confirmed experimentally in this work and is
Results and discussion
ꢀ
(
SCN
2
)
Validation of the technique
(
discussed in the result section of this paper.
In order to check the consistency of the results obtained by this
new technique, several well established rate constants have
been remeasured.
ꢀ
The SCN anion was detected using UV-visible spectro-
scopy following a medium pressure mercury lamp photolysis
2 2 2
of H O/H O /reactant/KSCN mixtures. All experiments for
the reaction of the OH radical with oxygenates were carried
out at room temperature (i.e., T ¼ 295K). Assuming steady-
state conditions for our flow conditions, the following expres-
sions were derived from the reaction mechanism (R1)–(R5) (all
processes are assumed to be first order or pseudo-first order
under the experimental conditions employed).
OH + methanol
The kinetics of the reaction of OH with methanol (R6) have
been studied before (see for example Table 1).
OH þ CH3OH ! H2O þ CH2OH
ðR6Þ
ꢀ
2
k1½H2O2ꢆ ꢀ k2½OHꢆ½SCN ꢆ ꢀ k5½OHꢆ½RHꢆ ¼ 0
ð1Þ
However, the agreement among the reported values for the
rate coefficient for the reaction of OH with methanol is less
clear than the reported values for the reaction of OH with
ethanol (see below). As a result, further investigations are war-
ranted to better define the rate coefficient for the removal of
OH by methanol.
ꢀ
ꢀ
ꢀ
k2½OHꢆ½SCN ꢆ þ kꢀ3½ðSCN2Þ ꢆ ꢀ k3½SCNꢆ½SCN ꢆ ¼ 0 ð2Þ
ꢀ
ꢀ
ꢀ
k3½SCNꢆ½SCN ꢆ ꢀ kꢀ3½ðSCN2Þ ꢆ ꢀ k4½ðSCN2Þ ꢆ ¼ 0
ð3Þ
Using simple algebra, eqns. (1)–(3) can be rearranged to give
the following relationship (eqn. 4).
ꢀ2
Typical data for the absorption profiles of the (SCN) con-
centration is shown in Fig. 2. The absorption profiles of the
ꢀ
ꢁ
1
1
k5½RHꢆ
ꢀ
(
conditions are effectively reached within the LCW. The metha-
SCN)
2
anion shown in Fig. 2 illustrate that the steady-state
¼
1 þ
ð4Þ
ꢀ
Abs Abs0
k2½SCN ꢆ
ꢀ3
ꢀ1
nol concentration was varied from 0 to 2 ꢁ 10 mol L . In
the data shown in Fig. 2, under typical experimental condi-
tions employed the mercury lamp was triggered off for the first
20–30 s in order to flush the LCW with the solution of interest.
ꢀ2
In eqn. (4), Abs is the absorbance of the (SCN) anion at
a known concentration of the organic compound RH and
ꢀ
Abs
0
is the maximum (SCN)
2
anion absorbance in the
absence of the reactant RH. Therefore, given the absorbencies,
and the concentration of the reactant RH, the rate coeffi-
cient k can easily be determined.
Then the mercury lamp was triggered back on again and the
(SCN) ion monitored for about 150 s before triggering the
2
lamp off again. The rise time of the (SCN) ion signal (see
ꢀ
k
2
2
ꢀ
5
Phys. Chem. Chem. Phys., 2003, 5, 1562–1569
1565

View Article Online
and not a second order process. However, the data presented
in Fig. 3 prove that the decay of the (SCN) anion is a first
order process and not a second order process. Since the decay
ꢀ
2
ꢀ
of the (SCN)
and not second order, reaction (R4) is believed to be the major
2
anion shown in Fig. 3 is shown to be first order
ꢀ
reactive channel responsible for the removal of (SCN)
2
anions. As a result, eqn. (4) can be used to calculate the relative
rate constant for the reaction of OH radical with organic
species.
The rate coefficient for the reaction of OH with methanol
at T ¼ 295 K obtained from a linear fit to eqn. (5) gives
9
ꢀ1 ꢀ1
s . The uncertainty is
k
OH+methanol ¼ (1.3 ꢃ 0.4) ꢁ 10
M
ꢃ2s and represents precision only. Again, all uncertainties
have been calculated based on uncertainties in temperature
and absorption measurements.
The reported room temperature rate coefficient for the reac-
tion of OH with methanol falls within the upper limit of the
previously reported room temperature data (ranging from
ꢀ2
Fig. 2 Typical (SCN) anion concentration temporal profiles for 6
different methanol concentrations (from the top to the bottom line
9
ꢀ1 ꢀ1
ꢀ5
ꢀ4
ꢀ4
0.78 to 1.2 ꢁ 10 M
s ). However, certain room tempera-
the methanol concentrations were 0, 5 ꢁ 10 , 2 ꢁ 10 , 5 ꢁ 10
,
ꢀ3
ꢀ3
ꢀ1
ture studies failed to give specific experimental conditions or
the specific experimental conditions were impossible to extract.
At least part of the variations among previously reported data
for the rate coefficient for the reaction of OH with methanol
can be attributed to temperature effects. Many other relative
rate measurements of the rate coefficient for the reaction of
OH radical with methanol were carried out (see Table 1). In
the majority of these studies the reported rate coefficients for
the OH + methanol reaction is lower than that obtained in this
study. However, all previous (with the exception of the work
1
off at ca. 30 and 180 s, respectively. The absorption for the (SCN)
ꢁ 10 and 2 ꢁ 10 mol L ). The UV lamp is switched on and
ꢀ
2
anion was taken in the ‘plateau’ region in the temporal range 150–
80 s.
1
Fig. 2) was observed to depend on the mercury lamp tempera-
ture. Since the mercury lamp was triggered at different times
for different durations depending on the experiment we did
not control the starting temperature of the lamp. As a result,
ꢀ
the (SCN)
trace to another. This did not pose a problem since all experi-
2
ion signal rise times differed slightly from one
43
by Adams et al. ) investigations of the rate coefficient for
the OH + methanol reaction were reported at higher pH
values than the ones used in this work. One relative rate
measurement of kOH+methanol is reported in the literature that
was carried out at pH values similar to the ones employed in
mental conditions were reproducible leading to comparable
absorbance profiles (see Fig. 2). Fig. 2 shows that the (SCN)
ꢀ
2
anion concentration, after the initial rise, decreases slightly
with time, with an approximately 5% decrease in absorbance
from 50 to 180 s. This seems to be more problematic when
lower reactant concentrations (top two traces in Fig. 2) are
43
this work. Thus, Adams et al. have studied the reaction
OH + methanol using a pulse radiolysis technique. These
9
investigators obtained a value for kOH+methanol ¼ 1.2 ꢁ 10
used. However, this ‘sloping’ effect was weaker in the temporal
range 150–180 s. As a result, the (SCN) anion absorbance
ꢀ1 ꢀ1 43
Similar to the experimental conditions employed
in this work, Adams et al. obtained the rate coefficient for
ꢀ
M
s
.
2
used in eqn. (4) was taken in the temporal range 150–180 s
shown in Fig. 2. As mentioned, this sloping effect was mainly
introduced by the stability of our mercury lamp.
ꢀ
the reaction of the OH radical with methanol using the SCN
ion as the competitor species. Moreover, similar to this work,
Adams et al. performed their studies in solutions at low pH in
order to minimize any adverse effects from secondary chemis-
ꢀ
The inverse absorbance plot for the (SCN) anion versus
2
the organic species concentration is shown in Fig. 3. The fit
is obtained from linear least squares analysis to eqn. (4) and
gives the relative rate constant, k . The major assumption used
43
try. As a result, the rate coefficient for the reaction of OH
with methanol reported in this study should only be compared
with the work of Adams and co-workers whose work was
performed under similar experimental conditions. Therefore,
we find that the rate coefficient for the reaction of OH with
methanol reported in this work is in excellent agreement
5
ꢀ
in eqn. (4) is that the decay of the (SCN) anion is first order
2
9
ꢀ1 ꢀ1
s by
with the previously reported value of 1.2 ꢁ 10
M
4
3
Adams et al.
Two other relative rate measurement of kOH+methanol are
reported in the literature that were carried out at lower pH
values but used the SCN ion in their competition kinetics
ꢀ
mechanism. Thus, Adams et al. and Elliot and McCracken
8
ꢀ1 ꢀ1 36
s and
s , respectively. These investigators used a
obtain values for k
¼ 7.8 ꢁ 10
M
OH+methanol
8
ꢀ1 ꢀ1 37
1
.0 ꢁ 10 M
1
0
ꢀ1 ꢀ1
value of 1.1 ꢁ 10
M
s for the rate coefficient for the com-
ꢀ
petition reaction of the OH radical with SCN . In this work
9
ꢀ1 ꢀ1
we adopted a rate coefficient of 1.29 ꢁ 10 M
s reported
1
ꢀ
by Chin and Wine for the OH + SCN reaction. This rate
coefficient is 17% higher than the value used by Adams et al.
and Elliot and McCracken. A correction of Adams’s et al.
and Elliot and McCracken’s result upward by 17% gives
9
9
ꢀ1 ꢀ1
k
OH+methanol ¼ 0.91 ꢁ 10 and 1.2 ꢁ 10 M
s , respectively,
a range of values which agree well with the rate coefficient for
the OH + methanol reaction obtained in this work.
ꢀ2
Fig. 3 An inverse absorbance plot for (SCN) as a function of the
concentration of methanol (circle), ethanol (square) and THF (trian-
gle). The fit is obtained from linear least squares analyses to eqn. (4).
The slope of a linear fit to eqn. (4) yields the ratio of the rate coeffi-
Based on the results obtained in this work for the reaction of
OH with methanol we can report with confidence that the new
experimental approach used in this study gives results that are
ꢀ
cients for the OH + organic reactant and OH + SCN reactions.
1
566
Phys. Chem. Chem. Phys., 2003, 5, 1562–1569

View Article Online
Table 2 Comparison of the room temperature rate coefficients for
the reaction of the OH radical with ethanol obtained in this work with
obtained from linear least squares analyses to eqn. (4). The
observed rate coefficient for the removal of OH by THF at
a
9
ꢀ1 ꢀ1
s . The uncer-
literature values
T ¼ 298 K is found to be (3.8 ꢃ 1.0) ꢁ 10 M
tainty is ꢃ2s and represents precision only.
ꢀ1
ꢀ1
s
Technique competitor Competitor
k
5
/M
Ref.
The reaction of OH with tetrahydrofuran (R8) was the focus
of only one previous investigation carried out by Eibenber-
ger. This investigator reported a rate coefficient for (R8) to
ꢀ
9
GR
PR
C
SCN
6
H
5
ꢀ
COO
1.7 ꢁ 10
1.8 ꢁ 10
1.9 ꢁ 10
2.2 ꢁ 10
2.0 ꢁ 10
1.6 ꢁ 10
1.8 ꢁ 10
45
36
43
46
47
39
38
38
38
48
40
49
41
42
5
2
9
9
9
9
9
9
9
9
9
9
9
9
9
9
ꢀ1 ꢀ1
s
2ꢀ3
be 3.8 ꢁ 10 M
at T ¼ 298 K.
OH þ C4H8O ! H2O þ C4H7O
PR
PR
CO
[Fe(CN)
4ꢀ
6
]
ðR8Þ
ꢀ
PR
PR
HSO4
(CH
C
3
)
2
NC
6
H
4
NO
ꢀ
The value for the OH + THF reaction rate coefficient reported
in this work is in excellent agreement with the data obtained by
Eibenberger.
PR
PR
6
H
5
CH COO
2
5
2
Nitro benzoate ion 1.8 ꢁ 10
ꢀ
PR
PR
C
CO
6
H
5
ꢀ3
COO
1.8 ꢁ 10
2.1 ꢁ 10
2.1 ꢁ 10
1.9 ꢁ 10
1.9 ꢁ 10
1.9 ꢁ 10
Since the reported rate coefficients for the reactions of OH
with ethanol, methanol and THF agree very well with the lit-
erature data, we report with confidence that the new experi-
mental approach used in this study is scientifically sound and
reliable to be applied to carry out the following and any rela-
tive future liquid phase kinetic experiments.
2
4
ꢀ
ꢀ
PR
PR
[Fe(CN)
[Fe(CN)
6
]
]
4
6
ꢀ b
2
PR
ABTS
Recommended
value
Literature
review
PR
4ꢀ
9
9
9
9
[Fe(CN)
COO
6
]
1.9 ꢁ 10
2.2 ꢁ 10
2.1 ꢁ 10
1.9 ꢁ 10
50
44
ꢀ
GR
LP
C
6
H
5
ꢀ
ꢀ
Determination of new rate constants
SCN
SCN
51
This
In the atmosphere, the degradation of the dimethylmalonate,
dimethylsuccinate, dimethylcarbonate and diethylcarbonate
will most likely be influences by the OH radical attack [i.e.,
reactions (R9)–(R12)].
MFT
work
a
PR: Pulse radiolysis; GR: Gamma radiolysis; LP: Laser photolysis;
2ꢀ
b
0
MFT: Micro-flowtube. ABTS : 2,2 -azinobis-(3-ethylbenzthiazo-
line-6-sulfonate).
OH þ CH3OCðOÞCH2CðOÞOCH3
!
H2O þ CH3OCðOÞCH2CðOÞOCH2
ðR9Þ
ðR10Þ
ðR11Þ
ðR12Þ
OH þ CH3OCðOÞCH2CH2CðOÞOCH3
! H2O þ CH3OCðOÞCH2CHCðOÞOCH3
in good agreement with the previously reported data using
other well established and proven techniques.
OH þ CH3OCðOÞOCH3
OH + ethanol
! H O þ CH OCðOÞOCH
2
3
2
Many room temperature studies of the OH + ethanol reaction
kinetics are reported in the literature (see Table 2).
OH þ CH3CH2OCðOÞOCH2CH3
! H2O þ CH3CH2OCðOÞOCHCH3
ꢀ
Absorbance data of the (SCN)
ent concentrations of ethanol obtained in this work, are similar
2
ion concentration at differ-
The experimental approach described above was used to mea-
sure the bimolecular rate coefficients for the reactions of the
OH radical with dimethylmalonate (R9), dimethylsuccinate
to the one shown in Fig. 2 for methanol. The absorbance
ꢀ
profile of the (SCN)
increasing ethanol concentration. This is in agreement with
2
anion was observed to decrease with
(
R10), dimethylcarbonate (R11) and diethylcarbonate (R12).
ꢀ
the fundamentals of the competition kinetics.
An inverse absorbance plot for (SCN)
Again, absorbance data of the (SCN)
2
ion concentration at
ꢀ
2
at l ¼ 480 nm as a
different concentrations of dimethylmalonate, dimethylsucci-
nate, dimethylcarbonate and diethylcarbonate obtained in this
work are similar to the one shown in Fig. 2 for methanol.
function of ethanol concentration is shown in Fig. 3. The fit is
obtained from linear least squares analyses to eqn. (4). The
slope of a linear fit to eqn. (4) yields the ratio of the rate coeffi-
ꢀ
Inverse absorbance plots for (SCN)
2
at l ¼ 480 nm at dif-
ꢀ
cients for the OH + ethanol and OH + SCN reactions. Since
the rate coefficient for the OH + SCN reaction is known,
ferent concentrations of dimethylmalonate, dimethylsuccinate,
dimethylcarbonate and diethylcarbonate are shown in Fig. 4.
The obtained rate coefficients for the reactions of the OH
radical with dimethylmalonate, dimethylsuccinate, dimethyl-
carbonate and diethylcarbonate carried out at T ¼ 295 K
are summarized in Table 3. To our best knowledge, there are
no other kinetic data for reactions (R9)–(R12) reported in
the literature with which to compare our results. However,
bond energies for the reactions of the OH radical with the 7
oxygenates shown in Table 4, tend to support the idea that
the reactivity of the oxygenates toward the attack by the OH
radical is linked to the steric environment of the –C(O)O–
functional group. This important point is discussed in the
following section.
ꢀ
the rate coefficient for the reaction of OH with ethanol [i.e.,
reaction (R7)] can easily be determined using simple algebra.
OH þ CH3CH2OH ! H2O þ CH3CHOH ðR7Þ
The rate coefficient for the aqueous phase reaction of the OH
radical with ethanol at T ¼ 295 K obtained in this work is
9
ꢀ1 ꢀ1
s . The uncertainty is
k
OH+ethanol ¼ (1.9 ꢃ 0.5) ꢁ 10
M
ꢃ2s and represents precision only. The rate coefficient for
the OH + ethanol reaction reported in this study compares
very well with the available literature data (see Table 2) inde-
pendent of the competitor species used.
OH + tetrahydrofuran
Discussion
ꢀ
Absorbance data of the (SCN) ion concentration at different
2
concentrations of THF obtained in this work are similar to the
one shown in Fig. 2 for methanol. An inverse absorbance plot
A major assumption in the reactions studied above is that the
OH attack on the alcohol, ester or carbonate proceeds via a H-
abstraction mechanism. This is consistent with the large
volume chamber study of the reaction of the OH radical with
DBE-5 in the presence of NO carried out by Atkinson and
ꢀ
of the inverse for (SCN)
concentration ranging from 0 to 1 ꢁ 10 mol L is shown in
2
at l ¼ 480 nm as a function of THF
ꢀ ꢀ1
3
Fig. 3. Similar to the situation described above, the fit is
Phys. Chem. Chem. Phys., 2003, 5, 1562–1569
1567

View Article Online
Table 4 Number of equivalent weak C–H bonds (n ) with the bond
dissociation energy (BDE) and the logarithm of the rate constant per
C–H bond caption
H
ꢀ1
Species
n
H
BDE/kJ mol
H
Log(k )
Methanol
Ethanol
THF
3
2
4
6
4
6
4
385
393
382
444
424
431
424
8.637
8.978
8.978
7.653
8.122
6.929
8.296
Dimethylmalonate
Dimethylsuccinate
Dimethylcarbonate
Diethylcarbonate
the OH radical with oxygenates proceeds via a H-atom
abstraction mechanism. However, dimethyl carbonate does
not seems to fit the pattern. The reason for this is not clear.
As a consequence, the weakness of the correlation does not,
currently, allow any reliable attempt of building an estimation
method for the OH radical rate coefficients. While it seems
unlikely that the reaction of the OH radical with oxygenates
does not proceed via a H-atom abstraction mechanism, this
issue remains to be conclusively resolved. Clearly, more theo-
retical and experimental research is needed to understand the
H-atom abstraction mechanism of the OH + oxygenates reac-
tion. Accordingly, any future research is warranted on the
reactivity of free radicals in aqueous solutions.
Fig. 4 An inverse absorbance plot for the oxygenates as a function of
their concentration [according to eqn. (1)]. The slope of the fitted line is
a measure of the ratio of the reference to oxygenate rate constants.
(
Bottom axis: S dimethylmalonate, L dimethylsuccinate, and /
diethylcarbonate; top axis: ; dimethylcarbonate).
co-workers. These investigators report that the OH radical
reaction with DBE-5 proceeds by three pathways, all involving
1
1
H-atom abstraction from the C–H bonds. If the OH attack
proceeds via a H-abstraction, then the observed rate coefficient
constants should be correlated with the type of bond and bond
strength of the ‘‘leaving’’ hydrogen atom. The ‘‘leaving’’
hydrogen atom is probably the one having the weakest C–H
bond strength (it was assumed that no H-abstraction occurs
at the O–H group for alcohols). Bond dissociation energies
Conclusion
This work addressed the oxidation of several oxygenated com-
pounds by the hydroxyl radical using a newly developed
experimental technique. The experimental approach described
herein takes advantage of the novel liquid core waveguide
technology of Teflon AF 2400. The physical nature of wave-
guides allows for the construction of photolysis reaction cells
with very low volumes but potentially very long optical path
lengths. As a result, the Teflon waveguide photolysis system
described here is potentially a very powerful and cost-effective
technique to study liquid phase chemistry with potentially very
low detection limits of chemical transients.
In summary, we have successfully applied optical wave-
guides to measure rate coefficients for the liquid phase reaction
of the OH radical with methanol, ethanol, tetrahydrofuran,
dimethylmalonate, dimethylsuccinate, dimethylcarbonate and
diethylcarbonate. The reported rate coefficients for the reac-
tions of OH with ethanol, methanol and THF agree very well
(
BDE) of the weakest C–H for methanol, ethanol, tetrahydro-
furan, dimethylmalonate, dimethylsuccinate, dimethylcar-
bonate and diethylcarbonate are listed in Table 4. Other
parameters listed in Table 4 are: (i) the number of equivalent
H
hydrogen atom n (i.e., those having the weakest bond
strength) and (ii) logarithm of the rate per H atom kH which
is derived by:
kH ¼ kobs=nH
ð5Þ
Many of the bond dissociation energies have not been experi-
mentally measured and were, therefore, estimated using semi-
5
3,54
empirical group contribution methods.
It might be
assumed that the observed reactivity differences of the OH
radical toward the organic species considered are due to
changes in the activation energies which are expected to be
linked to the bond dissociation energy of the bond being rup-
5
5
tured. If this is the case, there should be a linear dependence
of logarithm of k on the bond strength. A plot of log(k ) ver-
sus BDE is shown in Fig. 5. As can be seen in Fig. 5, there is
indeed a (weak) correlation between the logarithm of k and
H
H
H
the bond strength, i.e., a decrease of the rate coefficient is
observed with increasing BDE. As a result, the reaction of
Table 3 Summary of the rate coefficients for the reactions of the OH
radical with four oxygenates obtained in this work
a b c
Organic species
k
7
Dimethylcarbonate
Diethylcarbonate
Dimethylmalonate
Dimethylsuccinate
k
k
k
k
9
¼ (5.1 ꢃ 2.2) ꢁ 10
8
8
8
10 ¼ (7.9 ꢃ 3.2) ꢁ 10
11 ¼ (2.7 ꢃ 0.9) ꢁ 10
12 ¼ (5.3 ꢃ 2.9) ꢁ 10
a
b
ꢀ1 ꢀ1
All experiments performed at T ¼ 295 K. Units are k(M
s
).
c
Fig. 5 Plot of log(k
H
) vs. bond dissociation energy. Uncertainties are
Uncertainties are ꢃ2s and represent precision only.
ꢃ2s and represent precision only.
1
568
Phys. Chem. Chem. Phys., 2003, 5, 1562–1569

View Article Online
with the literature values. To date, there is no kinetic data
reported for the OH radical reaction with dimethylmalonate,
dimethylsuccinate, dimethylcarbonate and diethylcarbonate
with which to compare our results.
23 R. Altkorn, I. Koev, R. P. Van Duyne and M. Litorja, Appl. Opt.,
1
997, 36, 8992–8998.
W. Groh and A. Zimmermann, Macromolecules, 1991, 24, 6660–
663.
R. D. Waterbury, W. S. Yao, R. H. Byrne, B. J. Marquardt, K. P.
Turney and L. W. Burgess, Anal. Chim. Acta, 1997, 357, 99–102.
C. Gooijer, G. P. Hoornweg, T. deBeer, A. Bader, D. J. vanIperen
and U. A. T. Brinkman, J. Chromatogr., 1998, 824, 1–5.
R. H. Byrne, W. S. Yao, E. Kaltenbacher and R. D. Waterbury,
Talanta, 2000, 50, 1307–1312.
24
25
26
27
6
Acknowledgements
Support of this work by the Programme National de Chimie
Atmosph e´ rique (PNCA) from the CNRS is gratefully
acknowledged and the project PRIMEQUAL by the French
Minister for Environment. The authors also thank the anon-
ymous reviewers for their critical and helpful comments and
G. V. Buxton and G. A. Salmon for having provided the initial
push into that field.
28 P. K. Dasgupta, S. Y. Liu and H. Fein, Eur. Pat. Appl., 1999, 15.
29
30
31
B. J. Marquardt, P. G. Vahey, R. E. Synovec and L. W. Burgess,
Anal. Chem., 1999, 71, 4808–4814.
J. Z. Li, P. K. Dasgupta and Z. Genfa, Talanta, 1999, 50,
6
17–623.
P. Warneck, Heterogeneous and Liquid Phase Processes, Labora-
tory Studies Related to Aerosols and Clouds, Springer, Berlin,
1
996.
3
2
B. J. Finlayson-Pitts and J. N. Pitts, Chemistry of the Upper and
Lower Atmosphere, Theory, Experiments, and Applications, Aca-
demic Press, San Diego, 2000.
References
1
2
M. Chin and P. H. Wine, J. Photochem. Photobiol., A, 1992, 69,
7–25.
R. E. Huie, in Laboratory Studies of Atmospheric Heterogeneous
Chemistry; Current Problems in Atmospheric Chemistry, Advances
in Physical Chemistry Series, ed. J. R. Barker, World Scientific,
Singapore, 1994, vol. 3, pp. 374–419.
A. B. Ross, W. G. Mallard, W. P. Helman, G. V. Buxton, R. E.
Huie and P. Neta, 1998, NDRL-NIST Solution Kinetics Data-
base, National Institute of Standards and Technology (NIST),
see also http://www.rcdc.nd.edu.
33 R. Zellner, M. Exner and H. Herrmann, J. Atmos. Chem., 1990,
10, 411–425.
34 R. Zellner and H. Herrmann, in Spectroscopy in Environmental
Science, ed. R. J. H. Clark and R. E. Hester, John Wiley & Sons
Ltd., 1995, pp. 381–451.
35 AEA Technology, 1998, FACSIMILE for Windows, A Program
for Modelling Kinetics and Transport Under Windows 95 and
NT, http://www.mcpa-software.com/facsimileframe.html.
36 G. E. Adams, J. W. Boag and B. D. Michael, Trans. Faraday Soc.,
1965, 61, 1417–1424.
1
3
4
5
6
7
M. Uusi-Penttila, R. J. Richards, P. Blowers, B. A. Torgerson and
K. A. Berglund, J. Chem. Eng. Data, 1996, 41, 235–238.
F. Cavalli, I. Barnes and K. H. Becker, Int. J. Chem. Kinet., 2001,
37 A. J. Elliot and D. R. McCracken, Radiat. Phys. Chem., 1989, 33,
69–74.
38 P. Neta and L. M. Dorfman, J. Phys. Chem., 1969, 73, 413–417.
39 J. H. Baxendale and A. A. Khan, Int. J. Radiat. Phys. Chem.,
1969, 1, 11–24.
40 R. L. Willson, C. L. Greenstock, G. E. Adams, R. Wageman and
L. M. Dorfman, Int. J. Radiat. Phys. Chem., 1971, 3, 211–220.
41 B. S. Wolfenden and R. L. Willson, J. Chem. Soc., Perkin Trans.
2, 1982, 805–812.
33, 431–439.
J. Wenger, E. Porter, E. Collins, J. Treacy and H. Sidebottom,
Chemosphere, 1999, 38, 1197–1204.
M. Bilde, T. E. M o¨ gelberg, J. Sehested, O. J. Nielsen, T. J.
Wallington, M. D. Hurley, S. M. Japar, M. Dill, V. L. Orkin,
T. J. Buckley, R. E. Huie and M. J. Kurylo, J. Phys. Chem. A,
1
997, 101, 3514–3525.
42 G. V. Buxton, C. L. Greenstock, W. P. Helman and A. B. Ross,
J. Phys. Chem. Ref. Data, 1988, 17, 513–886.
8
9
M. A. Pacheco and C. L. Marshall, Energy Fuels, 1997, 11, 2–29.
J. R. Odum, T. P. W. Jungkamp, R. J. Griffin, R. C. Flagan and
J. H. Seinfeld, Science, 1997, 276, 96–99.
J. H. Seinfeld and S. N. Pandis, Atmospheric Chemistry and Phy-
sics: From Air Pollution to Climate Change, Wiley, New York,
43 G. E. Adams, J. W. Boag, J. Currant and B. D. Michael, in Pulse
Radiolysis, ed. M. Ebert, J. P. Keene, A. J. Swallow and J. H.
Baxendale, Academic Press, New York, 1965, pp. 131–143.
44 N. Motohashi and Y. Saito, Chem. Pharm. Bull., 1993, 41,
1842–1845.
1
0
1
998.
11
12
13
14
15
16
17
18
19
20
21
22
E. C. Tuazon, S. M. Aschmann and R. Atkinson, Environ. Sci.
Technol., 1999, 33, 2885–2890.
C. von Sonntag and H.-P. Schuchmann, Angew. Chem. Int. Ed.
Engl., 1991, 30, 1229–1253.
P. H. Wine, R. L. Mauldin and R. P. Thorn, J. Phys. Chem., 1988,
45 R. W. Matthews and D. F. Sangster, J. Phys. Chem., 1965, 69,
1938–1946.
46 G. E. Adams, J. W. Boag and B. D. Michael, Trans. Faraday Soc.,
1965, 61, 492–505.
47 E. Heckel, A. Henglein and G. Beck, Ber. Bunsen-Ges. Physik.
Chem., 1966, 70, 149–154.
48 G. V. Buxton, Trans. Faraday Soc., 1970, 66, 1656–1660.
49 M. S. Matheson, A. Mamou, J. Silverman and J. Rabani, J. Phys.
Chem., 1973, 77, 2420–2424.
50 H. R. Park and N. Getoff, Z. Naturforsch., A: Phys. Sci., 1992, 47,
985–991.
51 B. Ervens and H. Herrmann, Phys. Chem. Chem. Phys., 2002,
submitted.
52 J. Eibenberger, PhD Thesis, Pulse Radiolytic Investigations
Concerning the Formation and Oxidation of Organic Radicals in
Aqueous Solutions, University of Vienna, 1980.
53 J. Gasteiger, in Physical Property Prediction in Organic Chemistry,
ed. C. Jochum, M. G. Hicks and J. Sunkel, Springer Verlag,
Heidelberg, 1988, pp. 119–138.
92, 1156–1162.
G. V. Buxton, M. Bydder and G. A. Salmon, J. Chem. Soc., Fara-
day Trans., 1998, 94, 653–657.
G. V. Buxton, T. N. Malone and G. A. Salmon, J. Chem. Soc.,
Faraday Trans., 1997, 93, 2889–2891.
G. V. Buxton, T. N. Malone and G. A. Salmon, J. Chem. Soc.,
Faraday Trans., 1996, 92, 1287–1289.
G. V. Buxton, S. McGowan, G. A. Salmon, J. E. Williams and
N. D. Wood, Atmos. Environ., 1996, 30, 2483–2493.
H. Herrmann, M. Exner, H. W. Jacobi, G. Raabe, A. Reese and
R. Zellner, Faraday Discuss., 1995, 100, 129–153.
H. Herrmann, H. W. Jacobi, G. Raabe, A. Reese and R. Zellner,
Fresenius J. Anal. Chem., 1996, 355, 343–344.
H. Herrmann, A. Reese, F. Wicktor and R. Zellner, J. Mol.
Struct., 1997, 408, 539–542.
H. Herrmann, A. Reese and R. Zellner, J. Mol. Struct., 1995, 348,
54 J. Gasteiger, 2000, PETRA Parameter Estimation for the
Treatment of Reactivity Applications, http://
www2.chemie.uni-erlangen.de/software/petra/.
55 C. L. Clifton and R. E. Huie, Int. J. Chem. Kinet., 1989, 21,
677–687.
1
83–186.
R. Altkorn, I. Koev and M. J. Pelletier, Appl. Spectrosc., 1999, 53,
169–1176.
1
Phys. Chem. Chem. Phys., 2003, 5, 1562–1569
1569