Hydroxyl radical self-recombination reaction and absorption spectrum in water up to 350°C

Article abstract of DOI:10.1021/jp065992v

The rate constant for the self-recombination of hydroxyl radicals ( ^.OH) in aqueous solution giving H₂O₂ product has been measured from 150 to 350°C by direct measurement of the ^.OH radical transient optical absorption at 250 nm. The values of the rate constant are smaller than previously predicted by extrapolation to the 200-350°C range and show virtually no change in this range. In combining these measurements with previous results, the non-Arrhenius behavior can be well described in terms of the Noyes equation k_obs^-1=k _act^-1+ k_diff^-1, using the diffusion-limited rate constant k_diff estimated from the Smoluchowski equation and an activated barrier rate k_act nearly equal to the gas-phase high-pressure limiting rate constant for this reaction. The aqueous ^.OH radical spectrum between 230 and 320 nm is reported up to 350°C. A weak band at 310 nm grows in at higher temperature, while the stronger band at 230 nm decreases. An isosbestic point appears near 305 nm. We assign the 230 nm band to hydrogen-bonded ^.OH radical, and the 310 nm band is assigned to free ^.OH. On the basis of the spectrum change relative to room temperature, over half of the ^.OH radicals are in the latter form at 350°C.

Full text of DOI:10.1021/jp065992v

J. Phys. Chem. A 2007, 111, 1835-1843
1835
Hydroxyl Radical Self-Recombination Reaction and Absorption Spectrum in
Water Up to 350 °C
Ireneusz Janik and David M. Bartels*
UniVersity of Notre Dame, Radiation Laboratory, Notre Dame, Indiana 46556
Charles D. Jonah
Chemistry DiVision, Argonne National Laboratory, Argonne, Illinois 60439
ReceiVed: September 13, 2006; In Final Form: January 5, 2007
•
The rate constant for the self-recombination of hydroxyl radicals ( OH) in aqueous solution giving H
2
O
2
•
product has been measured from 150 to 350 °C by direct measurement of the OH radical transient optical
absorption at 250 nm. The values of the rate constant are smaller than previously predicted by extrapolation
to the 200-350 °C range and show virtually no change in this range. In combining these measurements
with previous results, the non-Arrhenius behavior can be well described in terms of the Noyes
equation kobs ) kact + kdiff , using the diffusion-limited rate constant kdiff estimated from the Smoluchowski
equation and an activated barrier rate kact nearly equal to the gas-phase high-pressure limiting rate constant
-1
-1
-1
•
for this reaction. The aqueous OH radical spectrum between 230 and 320 nm is reported up to 350 °C.
A weak band at 310 nm grows in at higher temperature, while the stronger band at 230 nm decreases. An
•
isosbestic point appears near 305 nm. We assign the 230 nm band to hydrogen-bonded OH radical, and the
•
3
10 nm band is assigned to “free” OH. On the basis of the spectrum change relative to room temperature,
•
over half of the OH radicals are in the latter form at 350 °C.
I. Introduction
at elevated pressures and temperatures. Because of their
relevance to the corrosion issue, the reactions
•
The hydroxyl radical ( OH) is arguably the most important
free radical species in oxidation processes of technological
importance. It is a primary chain carrier in flame chemistry,
•
•
OH + OH f H O
(1)
2
2
1
and the main actor in water-solvent-based “advanced oxida-
and
2
tion processes”. It is a primary product of water radiolysis
•
•
3
4
with ionizing radiation, photolysis with UV photons, or even
OH + H
2
O
2
f HO
2
+ H
2
O
(2)
5
bubble collapse in sonolysis. The importance of the species
received significant attention.13
has stimulated many, many studies of its reactions in the
6
7-10
Temperature studies of reaction 1 in the gas phase cover a
wide range of conditions beginning from the low-pressure
gas phase and in aqueous solution.
To our knowledge,
however, few connections have been established between
1
4
•
limit, through the falloff range, and to high-pressure limiting
the reaction rates of OH in the gas phase and in condensed
1
5,16
values
(k1,∞). Unexpected behavior of the simple association
phases. Can we use gas-phase and ab initio results to accurately
predict reaction rates in solution, and particularly water?
The present study makes the first direct connection between
gas-phase and aqueous solution reaction rates for the important
•
reaction of two OH radicals was the subject of several
theoretical studies.
1
7-20
From ab initio calculations of the
potential energy surface for H2O2 it becomes clear that during
•
the approach of two OH radicals, at first dipole-dipole forces
•
self-recombination of OH radicals, giving hydrogen peroxide
are dominant, and a weakly hydrogen-bonded HOHO metastable
intermediate forms. This intermediate isomerizes over a small
potential barrier into HOOH to finally form hydrogen peroxide
in its singlet electronic ground state. Besides the ground state
there are electronically excited potentials that interact with the
ground-state potential in a complicated manner. To investigate
whether excited electronic states contribute to the kinetic
product.
•
In water-cooled nuclear power plants, OH self-recombination
is a primary source of corrosive hydrogen peroxide and oxygen
molecules. The need to model radiation-induced chemistry in
reactor heat transport piping has stimulated the present study.
In particular, one would like to predict feasibility of reactors
cooled with supercritical water (SCW), because this can
significantly improve economics of the power generation over
present-day reactors.1 Recent studies on SCW-cooled nuclear
reactors involve still unresolved questions about the water
radiolysis mechanism and the water radiolysis product yields
16
behavior, Fulle et al. performed measurements of high-pressure
limiting rate constants for reaction 1, k1,∞. In the temperature
range 200-700 K they measured saturated laser-induced
1,12
•
fluorescence (SLIF) of OH radicals in a helium bath gas up to
100 bar. In the conclusion of their studies they reported no
temperature effect between 200 and 400 K on the high-pressure
limiting rate k1,∞, and even a slight negative temperature effect
on the rate coefficients between 400 and 700 K. At that time
*
To whom correspondence should be addressed. E-mail: bartels@
hertz.rad.nd.edu. Phone: (574) 631-5561. Fax: (574) 631-8068.
1
0.1021/jp065992v CCC: $37.00 © 2007 American Chemical Society
Published on Web 02/20/2007

1836 J. Phys. Chem. A, Vol. 111, No. 10, 2007
Janik et al.
the negative temperature effect was explained as underestimation
of the high-pressure limit due to an insufficient range of bath
gas pressures used in the experiment. However, in a subsequent
review about quantitative analysis of association reactions in
the atmosphere, Troe21 presented detailed analysis of the
temperature dependence of capture rate constants for reaction
1
and concluded that reaction 1 should have a very weak
negative temperature coefficient over the range 30-1300 K.
Knowledge of the gas-phase high-pressure limiting rate
constants can benefit the analysis of partially diffusion-limited
condensed-phase reaction rates when observed rate constants
13,22
follow the Noyes equation:
kobs- ) kact + kdiff
1
-1
-1
(3)
Here kobs is the observed rate constant in the condensed phase
at a given temperature, kact is the hypothetical limiting reaction
2
3
rate in absence of a diffusional concentration gradient, and
kdiff is a diffusion-limited reaction rate determined on the basis
of the Smoluchowski equation:
Figure 1. Experimental setup for measuring high-temperature pres-
surized water UV transient absorption.
kdiff ) âσ4πRD
(4)
where D ) Da + Db is the relative diffusion coefficient of
reactants a and b, R is the reaction distance, â is a spin statistics
factor, and σ is /2 for identical reactants but unity otherwise.
scattering of the abundant longer-wavelength light in a single
monochromator, and interference from the intense Cerenkov
light emitted in the UV region. To overcome these problems in
the present work, an optical system was set up to provide
optimum performance in the UV (Figure 1).
1
The earlier temperature studies of reaction 1 up to 200 °C
•
were performed by measuring the decay of OH radical transient
2
4-26
absorption in the deep UV.
The latest work by Elliot et
The light source used was an XBO 150 W/S xenon arc lamp
2
6
9
-1 -1
al. reports a rate constant of 4.0 × 10 M
s
at 25 °C and
(Osram G.m.b.H., Germany). The arc was intensified ap-
-
1
average activation energy of 7.0 kJ mol . In a later review,
proximately 30-fold using a transistorized pulser designed to
produce a flat-topped 300 µs light pulse. By means of a
magnesium fluoride coated parabolic reflector (PTI-033-
13
Elliot concluded that either an Arrhenius equation or the Noyes
equation (3) describes the data to within the experimental error,
and essentially the same value is given by either equation when
extrapolated to 300 °C (within 10%). This extrapolation suggests
a slight increase of the reaction 1 rate constant over the range
0
008-A), the arc was imaged through the absorption cell and
suprasil lens pair onto the entrance slit of a double mono-
chromator (DH-10, Instruments, SA. Inc.). Gratings with
1
3
of temperature 200-350 °C .
1
200 groove/mm blazed for maximum efficiency at 250 nm
were used; the measured scattered light at 225 nm was less than
.3%. The entrance, intermediate, and exit slit widths were set
•
In the present investigation, the OH radical recombination
has been studied at temperatures up to 350 °C using pulse
radiolysis. The system chosen to study this reaction was
0
to 0.25, 1, and 2 mm, respectively, giving a half-maximum
intensity bandwidth of 4.8 nm (for all slits being 2 mm, the
bandwidth increases to 8 nm). The light emerging from the
monochromator exit slit was directed onto the photocathode of
a solar-blind photomultiplier (Hamamatsu R1319).
•
N2O-saturated water. Radiolysis produces mainly OH radical
and e aq, with a smaller amount of H atom. In the presence
of N2O, solvated electrons are quickly converted to OH radicals
via reaction 5, giving the desired OH radical as ca. 90% of the
-
•
3,7
•
•
Pulse radiolysis/transient absorption experiments were carried
out using 4-20 ns pulses from the Argonne Chemistry
Division’s 20 MeV electron linac accelerator. A lead shield was
constructed to reduce the X-ray interference on the photomul-
tiplier operation. The average dose per pulse was monitored by
charge integration on a thick copper shutter that could be
lowered in front of the cell. The sample cell, flow system, and
e- + N O + H O f OH + N + OH
•
-
(5)
aq
2
2
2
total at room temperature. To extract second-order recombina-
tion rate constants, an absolute concentration scale must be
established via the free radical extinction coefficients. In the
sections below we describe first the work required to establish
the extinction coefficients at high temperature, and then the
procedure followed in analyzing the rate constants. We discuss
the changes in the OH radical spectrum with temperature and
interpret the recombination rates by comparison with the
gas-phase results.
basic experimental setup and characteristics were described in
28,29
previous publications.
To improve the signal-to-noise ratio,
•
signals were averaged over 30 consecutive traces. Due to
transient absorption of sapphire windows after the pulse of
electrons (see section IIIB), blank traces for a given dose and
temperature were collected in the empty cell purged with
flowing argon gas. These blank traces were subtracted from the
corresponding traces collected for aqueous samples at the same
temperature.
II. Experimental Section
At wavelengths 250 nm and below, microsecond kinetic
spectrophotometry becomes difficult because the intensities of
the available analyzing light sources are low, and optical
transmission losses are high, both from the sapphire of the
SCW cell windows and from the water absorption in this region
Samples were prepared by saturating water with N2O (AGA
Gas, Ultra-High Purity grade) at ambient conditions resulting
-
2
in a final concentration of N2O of 2.5 × 10 molal (m). The
solution was then pumped with the high-pressure liquid
chromatography pump at flow rates from 1.6 to 2.0 mL/min
27
at higher temperatures. These adverse conditions lead to a high
photodetector shot noise level, poor spectral purity due to

Hydroxyl Radical Self-Recombination
J. Phys. Chem. A, Vol. 111, No. 10, 2007 1837
TABLE 1: G Values and Extinction Coefficients for
Primary Absorbing Species Applied in Data Analysis
G(X) (mol J- ) × 10
1
-7
ꢀ (dm mol cm-1
3
-1
)
temp (°C) e-aq
H•
•OH •OH H• HO
•/O
•-
2
2
2 2
H O
2
5
2.86
3.32
3.43
3.48
3.50
3.49
3.44
3.33
3.13
0.57
0.71
0.85
0.93
1.04
1.20
1.43
1.77
2.25
2.90 535
3.82 465
4.03 425
4.13 390
4.24 364
4.35 331
4.46 296
4.59 260
30
57
68
73
78
84
89
95
1880
1810
1760
1523
1415
1262
1082
871
22
39
45
48
52
55
58
62
65
1
2
2
2
2
3
3
3
50
00
25
50
75
00
25
50
4.73 224 100
720
through the preheater and then into the high-temperature,
corrosion-resistant, hastelloy C-276 cell.30 Auxiliary cartridge
heaters compensated any heat loss inside the insulated cell
housing. To minimize light scattering, the temperature of the
solution introduced to the cell was kept within 0.5 °C of the
destination set point temperature. All the measurements were
performed at the same pressure 250 ( 0.2 bar.
Figure 2. •OH radical spectrum change vs temperature: (b) 30 °C;
(O) 200 °C; (9) 300 °C; (0) 350 °C. Extinction coefficients were
•
calculated from measured G × ꢀ on the basis of G( OH) for the given
temperature as described in the text. Inset: changes of ꢀ2 nm for OH
50
•
radicals vs temperature.
III. Results
•
We present results that were obtained after many iterations
of fitting both kinetics and spectra of the several species
involved. First we present the spectra and extinction coefficients
and describe how they were extracted. Then we describe the
salient kinetics observations and the analysis approach that was
followed. Finally, we show the rate constants needed to fit the
kinetics.
250 nm is shown in Figure 2 (inset). The values of OH radical
extinction coefficient are also summarized in Table 1.
H . The room-temperature H atom extinction coefficient at
•
•
-
1
-1 33
250 nm is less than or equal to 30 M cm . The shape of
the spectrum is known to remain the same up to 250 °C,
essentially an exponential decrease toward the red.36 If the
amplitude changes with temperature, its contribution to observed
absorption is expected to be more pronounced above 300 °C
where the ratio G(H )/G( OH) increases. At 350 °C the H atom
yield reaches up to 30% of the total yield of OH radicals in
our system, and uncertainty of its extinction coefficient can
impact the measured rate of reaction 1. This absorption has been
identified as a red shift of those water molecules in the
•
A. Extinction Coefficients. OH. The prompt (ca. 100-120
•
•
•
ns after radiolysis) UV transient radical spectrum (G × ꢀ) was
•
measured in N2O saturated water at ambient temperature as well
•
as at higher temperatures. This spectrum is dominated by OH
•
absorption but includes contributions from H atoms, H2O2, and
•
•-
•
•
HO2 /O2 . The G values for OH and H were calculated on
3
1
33
the basis of data reported previously by Lin et al. up to
50 °C. From the total G( OH + e aq + H ) reported by
Lin et al. we subtracted the sum of G(H ) and G(e aq)
determined in our laboratory from gas product analysis in
irradiated N2O-saturated aqueous solutions of phenol and
EtOH-d6.32 These numbers are tabulated in Table 1. The G
values of H2O2 were taken from γ radiolysis experiments
immediate solvation shell of the H atom. The absorption edge
of water itself shifts to the red by 0.5 eV from room temperature
to 350 °C, and a similar shift may be expected for the H atom
•
-
•
3
•
-
27
•
absorption. If a 0.5 eV shift occurs, the H atom extinction
-
1
-1
coefficient will reach a value of about 190 M cm at
250 nm (at room temperature this is the extinction coefficient
3
3
at 230 nm ). The values between room temperature and 350
°C were scaled with the temperature according to this assump-
tion and are summarized in Table 1.
1
3
reported up to 300 °C and were extrapolated up to 350 °C.
The measured absorption at room temperature was cor-
•
rected for H atom absorption using the spectrum reported by
H2O2. Hydrogen peroxide, whose extinction coefficient at
3
3
-1
-1 33,34
Nielsen et al. At higher temperatures spectra were corrected
250 nm is roughly 22 M cm ,
needs to be included in
•
•
for H atom absorption assuming the same red shift as was found
the fitting as it is a final product of OH decay. The apparent
increase of the relative final absorption (see section B) implied
that the hydrogen peroxide UV spectrum also becomes more
intense or shifts to the red at elevated temperature. This spectrum
is very difficult to measure accurately at elevated temperature
because hydrogen peroxide catalytically decomposes on metal
oxide surfaces. We estimate the H2O2 extinction coefficient by
assuming a red shift of the H2O2 absorption spectrum with
temperature similar to that of water. The result is consistent
with the magnitude of the product absorption at 250 nm. The
values are summarized in Table 1. It has subsequently been
pointed out to us that the long-wavelength (longer than 230 nm)
tail of H2O2 absorbance in the gas phase does increase
27
for the absorption edge of water itself (see below). We also
made a small correction for H2O2 absorption3 assuming that
its spectrum also red shifts with the temperature. Figure 2 shows
3,34
•
resulting spectra of OH radicals in the range of 30-350 °C.
All spectra were obtained at the same dose within (5%
deviation in day to day experiments. The spectrum obtained at
room temperature agrees very well with the spectra recorded
previously by Nielsen et al.33 and others.
9,35
With an increase
in temperature we observed an apparent decrease of absorption
at the wavelength of our interest, i.e., 250 nm. However, above
3
2
00 °C we observed a sudden increase of the absorption at
50 nm and shorter wavelengths. After correction for the red-
•
37
shifting H atom absorption, it became obvious that with the
increase of temperature, OH absorption continuously decreases
substantially with temperature, and this may also contribute
•
to the liquid-phase result.
HO2 /O2•-. As a product of reaction 2, the effective extinction
coefficient of this radical at neutral pH was expected to be very
important for data analysis. We have measured the effective
•
at the wavelength of interest. Therefore we collected also
G × ꢀ values at 250 nm for all remaining temperatures in the
experimental range. The change of extinction coefficient at

1838 J. Phys. Chem. A, Vol. 111, No. 10, 2007
Janik et al.
2
Figure 4. UV absorption spectral time evolution for N O aqueous
solution after a 10 ns electron pulse at 350 °C: (b) 60 ns after electron
pulse; (O 2 µs after electron pulse; (2) 10 µs after electron pulse.
Inset: same spectra normalized at 250 nm. The relative absorption
decrease is especially pronounced near 310 nm.
•
Figure 3. Comparison of OH radical 250 nm absorption decay profiles
recorded at 225 °C (upper) and 350 °C (lower) with global fit traces
superimposed.
The product absorption comes from one of the two mecha-
nisms mentioned above. We tested both. Fitting reaction 2 with
room-temperature hydrogen peroxide extinction coefficient
extinction coefficient for HO2 /O2•- in the temperature range
•
-
2
3
6
0-350 °C in neutral water containing 5.0 × 10 m H2 and
-
4
38
•
.5 × 10 m of oxygen. The change of effective extinction
required larger rate constants for OH + H2O2 than one would
expect from Arrhenius extrapolation of literature data.42 That
approach led to k2 being almost 3 times higher than the
extrapolation at 350 °C, and the temperature dependence in the
entire experimental range deviated distinctly from the previous
coefficient at 250 nm with temperature is summarized in Table
. With the increase of the temperature we observed gradual
decrease of the extinction coefficient from roughly 1880 to
1
-1
-1
7
20 M cm . This is the result of the shifting from ionic form
to protonated form, as the pKa value of the radical is expected
to change from 4.8 at room temperature to about 7.05 at
50 °C. (This room-temperature value is roughly 15% lower
than the number reported by Bielski et. al at pH ) 6 but agrees
with the later Elliot and Buxton observations. ) More detailed
analysis of HO2 /O2 spectral changes with temperature have
been presented in a paper dealing with the reaction of H with
4
2
results. On the other hand, forcing Arrhenius behavior of
reaction 2 and allowing for peroxide absorbance adjustment led
to roughly 3 times larger H2O2 extinction coefficient at 350 °C
than at room temperature. This result is not impossible if the
spectrum red-shifts, as was already discussed. (It has now also
been pointed out to us that the gas-phase absorption of H2O2
3
9
3
40
41
•
•-
•
37
increases with temperature in this wavelength region. ) For
3
8
O2.
either explanation of the product absorption, the temperature
dependence of reaction 1 was basically the same and the
numbers shifted only 10-15%. From this exercise we conclude
that we are able to achieve good estimation of the rate of the
•
B. Data Analysis. Initially, we analyzed OH decay traces
using the simplest model possible. We applied just the two
reactions (1) and (2), and we were able to achieve reasonably
good global fitting for data collected in the entire experimental
temperature range. The first observation resulting from collected
data was that on increasing temperature to 350 °C we saw a
gradual increase in the product absorption after completion of
reaction 1. A comparison of signals collected at 225 and 350
•
second-order decay of OH radicals, but the experiment cannot
deliver exact information about temperature changes of the rate
coefficients for reaction 2.
It is worth noting that an additional obstacle comes from
uncertainty in the sapphire transient absorption contribution at
higher temperatures. As mentioned in the Experimental Section
all the measurements in the SCW cell required subtraction of
sapphire absorption signals. The signal amplitude from sapphire
windows at longer times after the electron pulse increases
dramatically with temperature, and at 350 °C it is comparable
with the transient absorption signal amplitudes resulting from
the irradiation of aqueous solutions. The comparison of sapphire
absorptions at temperatures varying from 200 to 380 °C for two
commercial sapphire windows is presented in the Supporting
Information. One can expect that if the temperature of the
sapphire windows in the empty cell differs enough from the
temperature of the sapphire windows with the fluid in it (the
thermocouple placed in the cell does not touch the window
surface but rather the liquid phase in the center of the cell and
indirectly the body of the metal cell), an error could result in
the corrected residual absorption at longer times after the pulse
of electrons. Therefore we exchanged sapphire windows and,
instead of commercial synthetic sapphire windows (ESCO
°
C is presented in Figure 3. With an increase of the temperature
the ratio between initial and final signal amplitude decreases,
suggesting that either the H2O2 product extinction coefficient
increases with temperature or the rate constant for reaction (2)
is higher than extrapolated from data collected in the tem-
perature range 25-150 °C. (The OH radical spectrum was
already corrected for temperature, as explained above.)
42
•
•
To check that the UV signal observed after OH decay does
not arise from the OH but rather from the products, i.e., both
•
•
•-
H2O2 and HO2 /O2 radicals, we performed an experiment
showing the time evolution of the spectrum at 350 °C.
Figure 4 presents the result of this experiment. Proceeding from
time zero after the electron pulse toward longer times, we
observed a gradual decay of the spectrum. However, in the
region of 310 nm we found almost complete decay of the band
that is related to the gas-phase OH radical absorption
see Discussion below). This seems to confirm that the product
absorption completely dominates by 10 µs.
•
43
(

Hydroxyl Radical Self-Recombination
J. Phys. Chem. A, Vol. 111, No. 10, 2007 1839
Products), we used heat exchange method manufactured vacuum-
UV-grade sapphire windows (HEM VUV sapphire, Crystal
Systems) that were expected to have a lower level of both
impurities and crystal defects.44 In fact, the transient absorption
profile of vacuum-UV-grade windows in the deep UV was much
more convenient for transient absorption measurements at
elevated temperatures (The comparison of the transient absorp-
tion signals of the two different types of sapphire windows at
3
50 °C is shown in Supporting Information along with
some other data comparing the two types of windows).
Where inexpensive commercial windows show increase of
absorption after the pulse, the vacuum-UV-grade windows
show a small bleaching with no apparent change in the
transmittance level on the time scale of experiment. To compare
how this difference affects our data at the highest temperatures,
•
we performed a series of tests. The result of OH decay analysis
at 300-350 °C showed that, by using either vacuum-UV-grade
or commercial grade windows, we were able to obtain compa-
rable results for k1. However, signals related to the product
formation, i.e., several microseconds after the electron pulse,
differed enough to change by 100% the numbers for H2O2 ꢀ or
k2 (depending on the fitting approach; see above) at 350 °C.
We concluded once again that the level of confidence for the
numbers reported for reaction 1 can be relatively high, but to
avoid the problem of sapphire absorption, the vacuum-UV-grade
sapphire windows should be used in the range 300-350 °C.
Ultimately, we could not determine from fitting whether the
Figure 5. Temperature dependence of reaction 1 rate constants: (0)
values reported by Elliot;26 (b) experimental values from our studies;
(---) SACM/CT limiting high-pressure gas-phase values reported by
Troe and Ushakov;57 (s) diffusion-limited rate constants calculated
on the basis of the Smoluchowski equation with â ) 1; (O) limiting
k
act calculated on the basis of the Noyes eq 3 with use of given kdiff
and kobs
.
•
•-
long-time absorption originates from H2O2 or the HO2 /O2
TABLE 2: Fitted Rate Constants for Reaction (1)
radical or both together. To settle the issue, we performed back-
-10
(M-1 s-1) ( 20%
temp (°C)
k
3
× 10
to-back experiments at 300 °C at neutral pH and at pH ) 11
1
2
2
50
00
25
1.06
1.25
1.14
1.21
1.20
1.14
1.07
0.99
-
3
•-
(
i.e., 10 m NaOH) where we expected to see O2 rather than
•
•-
HO2 in the case of reaction (2) dominance. The O2 extinction
coefficient at 250 nm is twice as large as that of HO2 . The
result of the experiment shows no change, to within 5%, in the
•
38
250
275
300
•
product absorption after OH decay. It convinced us that the
3
3
25
50
final absorption originates mainly from H2O2 at this temperature.
C. Fitting Model. Even though we were able to fit our data
with a very simple model including only reactions 1 and 2, there
are certainly other processes contributing to the kinetics, and
the fitting parameters of the simple model may not give correct
rate constants. We built a more extensive fitting model that
accounts for all other reactions occurring during the water
radiolysis as well as including all extinction coefficients for the
other absorbing species present in the system. With extensive
testing, we converged on the following list as the only significant
processes in the N2O-saturated water system on the time scale
of interest:
et. al42 up to 150 °C. As explained above, the HO2 product
makes little contribution to the final absorbance at high
temperature, and our data are consistent with the Arrhenius
extrapolation. Reaction 6 rate constants were calculated as by
Reactions 7,-7 and 8 were estimated as described
in our recent publication. These reactions all represent minor
corrections to the OH recombination kinetics. The extinction
coefficients and G values were applied as stated above. For
studying second-order decay of OH radicals, three doses were
chosen for each temperature, varying from 11 to 70 Gy.
Figure 3 presents the experimental data collected in N2O
saturated water at 250 nm with global fitting curves superim-
posed at two temperatures: 225 and 350 °C. Figure 5 shows a
plot of the rate constant for reaction 1 studied in the range
•
1
3,26
Elliot.
3
8
•
•
•
•
OH + OH f H O
(1)
(2)
2
2
•
•
2
OH + H O f HO + H O
2
2
2
1
50-350 °C. The new rate constants are summarized in
Table 2. The data points shown represent an average of the fitted
rate constants obtained for the set of fitting parameters varied
within their accuracy limits at each temperature. The (20%
error bars are indicative of the range of rate constants obtained
when the fit was “good”, but this is a qualitative rather than a
quantitative measure. The data below 200 °C were obtained by
e- + N O + H O f OH + N + OH
•
-
(5)
(6)
aq
2
2
2
•
•
OH + H f H O
2
HO2• T O2•- + H+
(7)
(8)
26
Elliot et al. Our new measurements match quite well with the
previous results, when the extinction coefficients are adjusted
•
•
H + N O f OH + N
to common values. Above 200 °C, the rate constant stays
2
2
1
0
-1 -1
basically on the same level around 1.2 × 10
M
s . The
The reaction 2 rate constants were extrapolated using the
Arrhenius parameters determined previously by Christensen
value of this rate constant at 350 °C was overestimated by a
factor of 2.5 by the best previous extrapolations.
1
3

1
840 J. Phys. Chem. A, Vol. 111, No. 10, 2007
Janik et al.
IV. Discussion
spectra, one can see changes at 310 nm showing up already at
9
0 °C, and becoming more pronounced at 150 °C. It is there-
Although it was never our intention, we find that we have
gathered significant new information on the UV absorption
spectrum of OH radical in water during the course of our study.
We first review this subject and provide an interpretation of
the spectrum changes consistent with our kinetics data. We then
consider the new rate constant information for reaction 1 and
make a connection with high-pressure gas-phase measurements.
fore no surprise that we were able to record this change as well.
•
By analogy with the band observed in the gas phase around
•
3
07 nm, and interstitial OH in crystalline ice at 280 nm, the
simplest conclusion is that the observed absorption at 310 nm
is related to OH radicals that are not hydrogen bonded, and
the 230 nm band corresponds to hydrogen-bonded OH. With
•
•
increasing temperature the population of hydrogen-bonded
•
•
OH Radical Spectrum. At room temperature the OH radical
has a broad absorption band with a maximum around 230 nm
and a long-wavelength tail that extends beyond 320 nm even
•
OH radicals should eventually decrease, and the 230 nm band
should decrease as the 310 nm band increases. This is actually
the case in our experiment, and it makes sense of our preliminary
observation that the 250 nm extinction coefficient of OH
decreases at elevated temperature. Simple inspection of the
spectra in Figure 2 then suggests that at 350 °C less than half
of OH radicals are hydrogen bonded! This is also the conclusion
reached in a fully classical MD simulation of OH in SPC/E
water by Svishchev and Plugatyr.
out to 400 nm. Our result, shown in Figure 2, corresponds well
•
to the results reported over the last 30 years.9
,33,35,45,46
As we
were not able to extend our measurements below 230 nm at
higher temperatures, we cannot discuss any changes in the band
•
9
maximum or the existence of an additional band described
•
4
5
previously in the 188-200 nm region. The most recent
5
5
4
7
studies by Alam et. al have shown no special features at
•
However, we still have the problem of the origin of the
1
90-250 nm of the OH radical spectrum, in contrast to the
5
3
35
48
230 nm band. We agree with Treinin and Hayon that 1.4 eV
is far too large a solvent shift for the intramolecular Π f ∑
spectrum reported by Czapski et. al and other researchers.
2
2
+
This discrepancy has been neither resolved nor discussed
recently but is most likely explained by the larger spectral
bandwidth (not reported in the paper) in Alam’s experiment
transition. The smaller but already very large 0.8 eV blue shift
of the lowest water band has its origin in a ground-state to
Rydberg transition, where a very large condensation effect is
plausible. As for the claim there is no OH/ OD isotope effect
for the 230 nm band, this is not clearly supported by the
published spectra. The comparison of spectra is made very
difficult by overlapping bands in the deep UV, a difference in
spectrum shapes, and possible absorbance by other transients.
It makes more sense that the 230 nm band is a charge-transfer
4
5
compared with the narrow bandwidth (0.95 nm) used earlier.
•
•
Nielsen et al. suggested this band below 200 nm was similar in
•
origin to the H atom absorptionsa red shift of the water
3
3
molecule absorption in the first solvation shell of the radical.
Though it has not been noted in this context, a bound to
(
2
2 -
unbound Π f Σ transition should also be found in the
vacuum UV around 170 nm. )
49
•
•
band of OH radicals that are hydrogen bonded to water, and
that the 310 nm band corresponds to OH which is not hydrogen
The 230 nm band observed with OH in liquid water is not
•
•
found for OH in the gas phase, which only has one near UV
absorption band around 307 nm,
a Π f Σ intramolecular transition. In crystalline H2O ice
two forms of OH radical were reported. One was reported in
4
3,49,50
bonded. Then the question remains, where is the intramolecular
corresponding to
2
2
+
•
2
2
+
Π f ∑ absorbance for OH that is hydrogen bonded to
•
water? We suggest this is the origin of the weak long tail
extending to 400 nm in the room-temperature spectrum.35 Our
experiments have not probed this spectral region, but we predict
that the long tail will attenuate at higher temperature much like
the 230 nm band.
•
the region above 200 nm very much like OH dissolved in liquid
•
H2O and was believed to be hydrogen-bonded OH (in a
substitutional position),51 and the other with maximum at
•
52
2
80 nm was believed to be OH in “interstitial” positions.
Treinin and Hayon argued that the apparent blue shift
1.4 eV) of the 230 nm band in water relative to the 307 nm
The sudden increase in the initial deep UV absorption below
250 nm observed above 300 °C (one example is shown on
Figure 4) is not explained by our hydrogen-bonding inter-
pretation. However, it can be understood on the basis of
three trends: (i) the H atom yield increases in the range 300-
350 °C by more than 60%; (ii) the H atom spectrum, originating
(
band in vacuum is too large to be explained by a solvation
effect.5 They ascribe the OH absorption at 230 nm to a charge-
3
•
•
•
transfer transition involving H2O as donor and OH radical as
acceptor, much like the H atom absorption. Nielsen et al.
•
•
disagreed with this assessment and noted a small isotope effect
with water molecules in the solvation shell, can be expected to
shift toward the red in the same manner as H2O with the
•
•
on the 230 nm band for OH in H2O relative to OD in D2O,
•
•
56
•
whereas isotope shifts for H atom vs D atom and for solvated
electron deep UV absorption in the two solvents are quite
temperature; (iii) the OH radical room-temperature band
observed at 180 nm could shift to the red in the same manner
3
3
•
•
apparent. These authors instead compared the OH spectrum
with the (0.8 eV) blue shift of the UV spectrum of H2O on
as the H atom. It has been found that the H2O UV absorption
edge shifts by 0.5 eV from room temperature to 350 °C,
33
27
going from the vapor to liquid. This disagreement over the
origin of the OH spectrum has apparently not been discussed
corresponding to roughly a 20 nm red shift in the UV. If we
•
•
assume similar behavior for the H atom, we estimate that at
•
further since that time.
350 °C and 230 nm, the extinction coefficient of H atom would
-
1
-1
•
•
Spectra have since been obtained at higher temperatures. In
their very first paper about high-temperature pulse radiolysis,
be 613 M cm . Comparing yields of H and OH at 350 °C
and their extinction coefficients indicates that the initial absorp-
•
•
Christensen and Sehested reported OH spectrum changes from
tion at 230 nm originates roughly 50% from the H atom and
0 to 200 °C.48 A decade later in their high-temperature analysis,
•
•
2
50% from OH radicals. Using the shift of the H atom spectrum
by 20 nm to correct the apparent initial absorption spectrum of
•
41
•
54
Buxton and co-workers reported OH and OD spectra and
claimed no significant changes with temperature. However, these
published OH radical spectra do show evidence of a change
with temperature. The spectra normalized at 250 nm very
clearly show temperature-related differences in the 310 nm
•
OH radicals below 250 nm, we were able to recover a shape
of the OH spectrum at 350 °C that resembles very well the
•
•
41
shape observed at lower temperatures. Nevertheless, kinetic
5
4
experiments performed above 300 °C, among them measure-
4
8
•
region. Looking back again to Christensen and Sehested’s
ments of the H + O2 reaction rates that are the subject of

Hydroxyl Radical Self-Recombination
J. Phys. Chem. A, Vol. 111, No. 10, 2007 1841
another article,38 lead us to the conclusion that a 0.5 eV shift
of the H atom spectrum is somewhat too high, as the fitting of
In a preceding paper, we reported a very similar result for
38
•
•
reaction of H atoms with O2. Once the diffusion limit was
kinetic traces indicates that its extinction coefficient at 230 nm
accounted for using the Smoluchowski equation and the Noyes
equation, an activated rate constant in “perfect” agreement with
the gas-phase high-pressure limit was found. In that system,
•
cannot reach values larger than those of the OH radical. This
•
leads us to point (iii), that the absorption spectrum of OH
•
radicals changes not only in response to the hydrogen bonding
but also because of a red shift of the third band below 200 nm,
the result may be expected because both H atom and O2 are
hydrophobic. The water solvent has very little effect on the
intermolecular potential near the reaction “bottleneck” and
simply acts as a very efficient “third body”. (There is no barrier
4
5,53
which could originate with the solvating water molecules.
Another explanation for the deep UV result above 300 °C
could simply be a distortion error related to the broad mono-
chromator bandwidth used at 230 nm. This was not expected
•
to reaction for approximately 45° approach of H with respect
to the O-O internuclear axis, but for other angles a substantial
barrier is present.) In the present case the close agreement of
gas-phase and aqueous-phase reaction rates is surprising. The
gas-phase reaction path is expected to involve a metastable
HO‚‚‚HO hydrogen-bonded pair, with a transition state that
•
to affect the data at the time of experiment because the H atom
spectrum was not expected to contribute much at 230 nm.
Further studies are clearly needed to clarify these spectrum
changes in the deep UV.
•
•
involves rotation of the central H atom out from between the
The evident spectral changes of OH radicals are reflected in
•
•
O atoms. In liquid water we can expect OH to be hydrogen
the determination of OH radical absorption at the wavelength
bonded to water much of the time, and other angles of approach
probably dominate. The good agreement indicated in Figure 5
means that the barrierless inner reaction channel for oxygen-
oxygen approach dominates in both phases, and that this channel
is not strongly perturbed by the water solvent.
of interest, i.e., 250 nm. The changes of extinction coefficient
for 250 nm versus the temperature from our experimental data
are presented in Figure 2 (inset). Given insufficient information
•
about the H atom spectrum, one might expect a systematic
•
underestimation of the OH extinction at 250 nm. In this
eventuality, the rate constants obtained here will be lowered
by at most 25% at 350 °C and deviate less than this at lower
temperatures.
•OH Radical Recombination. Our reaction rate measure-
The most surprising aspect of the comparison with the gas-
phase result is the slightly negative temperature dependence of
kact, in apparent agreement with the gas-phase negative tem-
perature coefficient. In the gas phase this behavior comes
2
2
primarily from the statistical factor for Π1/2 and Π3/2 spin-
ments fall within 20% of numbers reported previously up to
•
57
2
6
orbit states of the OH radical. The spin-orbit splitting is
2
00 °C by Elliot et al. When we correct the previous results
-1
2
•
139 cm , with Π3/2 being lower in energy. The ground singlet
state of H2O2 product correlates with a pair of separated Π3/2
OH radicals. At very low temperatures only Π3/2 is populated,
and /4 of the collisions will correlate with the singlet ground
state, the rest correlating with repulsive triplet excited states.
for the OH absorbance changes vs temperature as described
above, we have an excellent match between the data sets. The
Noyes eq 3 was used to correlate the previous data, assuming
2
•
2
1
13
that the kact rate constant follows an Arrhenius law. Extrapola-
tion of this fit does not match our experimental numbers above
2
At higher temperatures, Π1/2 becomes populated, resulting in
2
00 °C. It seems somewhat surprising that the studied rate
a larger percentage of nonreactive collisions. In the limit of
constant does not change above 200 °C and remains basically
unchanged up to 350 °C.
-
1
1
kT . 139 cm , the statistical factor will reach /16. The negative
temperature coefficient comes from this transition of the
statistical factor between /4 at very low temperature and /16 at
very high temperature.
We follow the previous approach13 in estimating the diffusion
limit for reaction 1 from the Smoluchowski equation, using
1
1
-
9
2
-1
relative diffusion coefficient D ) 4.6 × 10 m s at room
temperature, and scaling the temperature dependence by the ratio
of T/viscosity. The reaction distance R is assumed to be 4.4
How can the same statistical factor apply to the aqueous-
•
phase reaction? If the OH radical is hydrogen bonded, this
1
3
should quench the orbital angular momentum, and the spin-
Å, independent of temperature. (We note immediately that
2
2
2
•
1/2 3/2
orbit states Π and Π are no longer eigenstates of the
these estimates could be up to 25% off.) Isolated Π OH radical
-
1
potential. We have to expect that this is the situation in liquid
water much of the time, in between brief moments when the
is characterized by a large spin-orbit splitting of 139 cm ,
due to the orbital angular momentum about the internuclear
•
5
0
hydrogen bond breaks and OH exists as a freely rotating
axis. Hydrogen bonding is expected to quench this angular
momentum and the resulting spin-orbit coupling. However, the
fluctuations of hydrogen bonding within liquid water probably
result in a very fast and large modulation of the local magnetic
molecule. (The UV absorption spectrum discussed above
suggests that such moments do occur, and are even very
common at higher temperature and lower densities.) When the
•
•
OH radical breaks free of hydrogen bonds, the probability of
field in the OH radical; this can be expected to cause very fast
2
2
forming Π1/2 or Π3/2 eigenstate is dictated by the Boltzmann
factor for the given temperature. It would seem that the only
way to get exactly the same temperature-dependent statistical
factor for both gas- and aqueous-phase rate constants is if the
reaction path involves breaking hydrogen bonds to both radicals,
momentarily producing the Π1/2 or Π3/2 spin-orbit eigenstates
of the free OH. Note that this is consistent with â ) 1 used
spin relaxation within the radical encounter pair and justifies
the use of a spin factor â ) 1. In Figure 5 we plot the estimated
diffusion limit along with the observed rate constants from this
26
study and from the earlier study of Elliot et al. Using the Noyes
equation, we then deduce the rate constant kact, which is plotted
in Figure 5 as the open circles. The dashed line in Figure 5
represents the limiting high-pressure gas-phase values calculated
2
2
•
by Troe and Ushakov.2
1,57
above in the Smoluchowski equation, because essentially each
re-collision of an encounter pair could give a new random pair
One can see that the calculated kact
corresponds very well to the limiting high-pressure values,
with nearly the same magnitude,and above 200 °C the same
weakly negative temperature dependence. By tweaking the
estimate of kdiff to a lower value, we could get nearly the same
negative temperature coefficient over the whole temperature
range.
2
2
of Π1/2 or Π3/2 states.
It seems just as likely that the weak negative temperature
dependence in our data above 200 °C is due to other factors
(not the least of which could be experimental error), producing
a serendipitous resemblance to the gas-phase result. If the

1842 J. Phys. Chem. A, Vol. 111, No. 10, 2007
Janik et al.
•
reacting OH radicals are hydrogen bonded to water, and orbital
(3) Draganic, I. G.; Draganic, Z. D. The Radiation Chemistry of Water;
Academic Press: New York, 1971.
1
angular momentum is quenched, a statistical factor of /16
0.0625) would be expected for formation of the ground state
singlet, independent of temperature. That is, by reducing spin-
orbit splitting to a small average value, the aqueous system is
always in the high-temperature limit. The gas-phase statistical
factor in our experimental temperature range is approximately
10. If we normalize our aqueous-phase results with the factor
.6, they lie nearly on top of the high-pressure limit result,
(
4) Gonzalez, M. G.; Oliveros, E.; Worner, M.; Braun, A. M. J.
Photochem. Photobiol. C 2004, 5, 225.
5) Adewuyi, Y. G. Ind. Eng. Chem. Res. 2001, 40, 4681.
(
(
(6) Baulch, D. L.; Bowman, C. T.; Cobos, C. J.; Cox, R. A.; Just, T.;
Kerr, J. A.; Pilling, M. J.; Stocker, D.; Troe, J.; Tsang, W.; Walker, R. W.;
Warnatz, J. J. Phys. Chem. Ref. Data 2005, 34, 757.
(
7) Buxton, G. V.; Greenstock, C. L.; Helman, W. P.; Ross, A. B. J.
Phys. Chem. Ref. Data 1988, 17, 513.
8) Ervens, B.; Gligorovski, S.; Herrmann, H. Phys. Chem. Chem. Phys.
2003, 5, 1811.
1
/
1
(
2
suggesting that reaction of Π3/2 OH radicals on the ground state
singlet surface occur with almost identical rate constants in water
and the gas phase.
(9) Herrmann, H. Chem. ReV. 2003, 103, 4691.
(
10) von Sonntag, C. Free-radical-induced DNA damage and its repair.
A chemical perspectiVe; Springer: Heidelberg, 2006.
11) Koshizuka, S.; Oka, Y. Prog. Nucl. Energy 1998, 32, 547.
(12) Oka, Y.; Koshizuka, S. J. Atom. Eng. Soc. Jpn. 2002, 44, 600.
13) Elliot, A. J. Rate Constants and G-Values for the Simulation of
(
V. Conclusion
(
Pulse radiolysis of N2O-saturated water combined with UV
transient absorption detection has allowed determination of the
rate for OH radical self-recombination up to 350 °C. The values
of the rate constant are smaller than predicted previously in the
the Radiolysis of Light Water over the Range 0-300 °C. AECL report
11073, 1994.
(
14) Zellner, R.; Ewig, F.; Paschke, R.; Wagner, G. J. Phys. Chem. 1988,
2, 4184.
15) Forster, R.; Frost, M.; Fulle, D.; Hamann, H. F.; Hippler, H.;
9
(
200-350 °C range and show no apparent change in this regime.
Schlepegrell, A.; Troe, J. J. Chem. Phys. 1995, 103, 2949.
(16) Fulle, D.; Hamann, H. F.; Hippler, H.; Troe, J. J. Chem. Phys. 1996,
105, 1001.
The temperature behavior and magnitude of the rate constant
are in excellent agreement with the gas-phase high-pressure
limiting rate, suggesting very little perturbation of the reaction
path by the water solvent. In fact, above 200 °C where the
diffusion limit is unimportant, the only obvious effect of the
(17) Brouwer, L.; Cobos, C. J.; Troe, J.; Dubal, H. R.; Crim, F. F. J.
Chem. Phys. 1987, 86, 6171.
(
(
(
18) Harding, L. B. J. Phys. Chem. 1989, 93, 8004.
19) Harding, L. B. J. Phys. Chem. 1991, 95, 8653.
20) Maergoiz, A. I.; Nikitin, E. E.; Troe, J. J. Chem. Phys. 1995, 103,
2
2
water solvent is to modify the statistical ratio of Π1/2 or Π3/2
spin-orbit sublevels, and to act as a third body. This suggests
that high-level ab initio results are perfectly adequate for
prediction of OH radical recombination rates in the condensed
phase.
2083.
(
(
(
21) Troe, J. Chem. ReV. 2003, 103, 4565.
22) Noyes, R. M. Prog. React. Kinet. 1961, 1, 129.
23) Pilling, M. J.; Seakins, P. W. Reaction kinetics; Oxford University
Press: Oxford, U.K., 1995.
In the course of the study new information on the OH radical
UV absorption spectrum was obtained, showing a decrease in
the primary band at 230 nm and growth of a weak band at
(24) Christensen, H.; Sehested, K. Radiat. Phys. Chem. 1981, 18,
723.
(25) Elliot, A. J. Radiat. Phys. Chem. 1989, 34, 753.
(26) Elliot, A. J.; McCracken, D. R.; Buxton, G. V.; Wood, N. D. J.
3
3
10 nm at elevated temperature, with isosbestic point near
05 nm. We interpret the 230 nm band as due to hydrogen-
Chem. Soc., Faraday Trans. 1990, 86, 1539.
(27) Marin, T. W.; Takahashi, K.; Bartels, D. M. J. Chem. Phys. 2006,
bonded OH, and the 310 nm band corresponding to “free” OH.
125, 104314.
•
At the same time we find the absorption spectra of the H atom
(28) Cline, J.; Takahashi, K.; Marin, T. W.; Jonah, C. D.; Bartels, D.
M. J. Phys. Chem. A 2002, 106, 12260.
and H2O2 shift toward the visible.
(
29) Marin, T. W.; Cline, J. A.; Takahashi, K.; Bartels, D. M.; Jonah,
C. D. J. Phys. Chem. A 2002, 106, 12270.
30) Takahashi, K.; Cline, J. A.; Bartels, D. M.; Jonah, C. D. ReV. Sci.
Instrum. 2000, 71, 3345.
31) Lin, M. Z.; Katsumura, Y.; Muroya, Y.; He, H.; Wu, G. Z.; Han,
Z. H.; Miyazaki, T.; Kudo, H. J. Phys. Chem. A 2004, 108, 8287.
32) Janik, D.; Janik, I.; Bartels, D. M. Radiat. Res. Manuscript in
Acknowledgment. We thank Dr. Sergey Chemerisov for
maintaining and operating the linac accelerator used in this work.
We especially thank Dr. Juergen Troe for sharing an unpublished
manuscript concerning the reaction 1 gas-phase high-pressure
limit, and Dr. Larry Harding for useful discussions concerning
the relevant statistical factors. We thank Dr. Daniel Chipman
for very useful discussions regarding the UV absorption of OH
radical. Work at Argonne National Laboratory was performed
under the auspices of the Office of Science, Division of
Chemical Science, US-DOE under contract number W-31-109-
ENG-38. Additional support has been provided by US-DOE
NERI grant 02-060. The Notre Dame Radiation Laboratory is
supported by the Office of Basic Energy Sciences at the United
States Department of Energy. This is document number NDRL-
(
(
(
preparation.
(33) Nielsen, S. O.; Michael, B. D.; Hart, E. J. J. Phys. Chem. 1976,
80, 2482.
(34) Vaghjiani, G. L.; Ravishankara, A. R. J. Geophys. Res. [Atmos.]
1
989, 94, 3487.
(
(
35) Czapski, G.; Bielski, B. H. J. Radiat. Phys. Chem. 1993, 41, 503.
36) Sehested, K.; Christensen, H. Int. J. Radiat. Appl. Instrum. Pt. C:
Radiat. Phys. Chem. 1990, 36, 499.
(37) Kijewski, H.; Troe, J. HelV. Chim. Acta 1972, 55, 205.
(38) Janik, I.; Marin, T.; Jonah, C. D.; Bartels, D. M. J. Phys. Chem. A
2
007, 111, 79.
(
(
39) Christensen, H.; Sehested, K. J. Phys. Chem. 1988, 92, 3007.
40) Bielski, B. H. J.; Cabelli, D. E.; Arudi, R. L.; Ross, A. B. J. Phys.
4
689 from the Notre Dame Radiation Laboratory.
Chem. Ref. Data 1985, 14, 1041.
Supporting Information Available: Five figures and an
(41) Elliot, A. J.; Buxton, G. V. J. Chem. Soc., Faraday Trans. 1992,
8, 2465.
8
8
explanation of transient absorption observed in two sapphire
window samples at elevated temperatures. This material is
available free of charge via the Internet at http://pubs.acs.org.
(42) Christensen, H.; Sehested, K.; Corfitzen, H. J. Phys. Chem. 1982,
6, 1588.
(43) Pellerin, S.; Cormier, J. M.; Richard, F.; Musiol, K.; Chapelle, J.
J. Phys. D: Appl. Phys. 1996, 29, 726.
(
44) Khattak, C. P.; Schmid, F. J. Cryst. Growth 2001, 225, 572.
References and Notes
(45) Pagsberg, P.; Christensen, H.; Rabani, J.; Nilsson, G.; Fenger, J.;
(
1) Haber, L. C.; Vandsburger, A. Combust. Sci. Technol. 2003, 175,
859.
2) Pignatello, J. J.; Oliveros, E.; MacKay, A. Crit. ReV. EnV. Sci.
Technol. 2006, 36, 1.
Nielsen, S. O. J. Phys. Chem. 1969, 73, 1029.
1
(46) Janata, E. Proc. Ind. Acad. Sci., Chem. Sci. 2002, 114, 731.
(47) Alam, M. S.; Kelm, M.; Rao, B. S. M.; Janata, E. Radiat. Phys.
Chem. 2004, 71, 1087.
(

Hydroxyl Radical Self-Recombination
J. Phys. Chem. A, Vol. 111, No. 10, 2007 1843
(
83.
(
73.
48) Christensen, H.; Sehested, K. Radiat. Phys. Chem. 1980, 16,
(53) Treinin, A.; Hayon, E. J. Am. Chem. Soc. 1975, 97, 1716.
(54) Buxton, G. V.; Lynch, D. A.; Stuart, C. R. J. Chem. Soc., Faraday
Trans. 1998, 94, 2379.
(55) Svishchev, I. M.; Plugatyr, A. Y. J. Phys. Chem. B 2005, 109, 4123.
(56) Marin, T. W.; Takahashi, K.; Bartels, D. M. J. Chem. Phys. 2006,
125, 104314.
1
8
49) Vandishoeck, E. F.; Dalgarno, A. J. Chem. Phys. 1983, 79,
(50) Herzberg, G. Molecular Spectra and Molecular Structure. I. Spectra
of Diatomic Molecules, 2nd ed.; van Nostrand: New York, 1950.
(
(
51) Ghormley, J. A.; Hochanadel, C. J. J. Phys. Chem. 1971, 75, 40.
52) Taub, I. A.; Eiben, K. J. Chem. Phys. 1968, 49, 2499.
(57) Troe, J.; Ushakov, V. G. Phys. Chem. Chem. Phys. Submitted for
publication.