Low-temperature kinetics of the reaction of the OH radical with hydrogen peroxide

Article abstract of DOI:10.1021/jp030424q

The kinetics of the reaction of the OH radical with hydrogen peroxide, H₂O₂, are studied over a temperature range of 96-296 K. The low-temperature environment is provided by a pulsed Laval nozzle supersonic expansion of nitrogen with admixed H₂O₂. Hydrogen peroxide serves as both the OH radical photolytic precursor (λ = 248 nm) and a reactant. Laser-induced fluorescence of the OH radicals excited in the (1,0) band of the A²Σ⁺-X²Π₁ transition is used to monitor the kinetics of OH removal. The rate coefficient of the OH + H₂O₂ reaction (k₁) shows a negative temperature dependence within this temperature range, which can be expressed as k₁ = (6.8 ± 1.0) × 10^-13 exp[(285 ± 27)/T] cm³ molecule^-1 s^-1. The combined low- and high-temperature (literature) kinetic data form a U-shaped Arrhenius plot, which suggests that the reaction mechanism changes from direct abstraction (at high temperatures) to a mechanism involving formation of a hydrogen-bonded complex (at low temperatures). Atmospheric implications of the new low-temperature kinetic data are discussed.

Full text of DOI:10.1021/jp030424q

10642
J. Phys. Chem. A 2003, 107, 10642-10647
Low-Temperature Kinetics of the Reaction of the OH Radical with Hydrogen Peroxide†
‡
§
§
,|
Andrei B. Vakhtin, David C. McCabe, A. R. Ravishankara, and Stephen R. Leone*
JILA, National Institute of Standards and Technology and UniVersity of Colorado, Boulder, Colorado 80309,
NOAA Aeronomy Laboratory, 325 Broadway R/AL2, Boulder, Colorado 80305, and Department of Chemistry
and Biochemistry and CIRES, UniVersity of Colorado, Boulder, Colorado 80309
ReceiVed: April 10, 2003; In Final Form: September 30, 2003
The kinetics of the reaction of the OH radical with hydrogen peroxide, H
range of 96-296 K. The low-temperature environment is provided by a pulsed Laval nozzle supersonic
expansion of nitrogen with admixed H . Hydrogen peroxide serves as both the OH radical photolytic
precursor (λ ) 248 nm) and a reactant. Laser-induced fluorescence of the OH radicals excited in the (1,0)
2 2
O , are studied over a temperature
2
O
2
2
+
2
band of the A Σ -X Π
OH + H reaction (k
be expressed as k
i
transition is used to monitor the kinetics of OH removal. The rate coefficient of the
O
2 2
1
) shows a negative temperature dependence within this temperature range, which can
-
13
3
-1 -1
1
) (6.8 ( 1.0) × 10 exp[(285 ( 27)/T] cm molecule s . The combined low- and
high-temperature (literature) kinetic data form a U-shaped Arrhenius plot, which suggests that the reaction
mechanism changes from direct abstraction (at high temperatures) to a mechanism involving formation of a
hydrogen-bonded complex (at low temperatures). Atmospheric implications of the new low-temperature kinetic
data are discussed.
Introduction
While the rate coefficient for reaction R1 has been measured
1
-11
previously,
these measurements have been restricted to
Hydrogen peroxide, H2O2, is formed in the atmosphere
primarily via the reaction of two HO2 radicals, thus converting
two reactive HOX (defined as the sum of OH and HO2) radicals
to an inactive form. Once formed, H2O2 is removed from the
atmosphere via a variety of processes such as rainout, uptake
onto condensed matter where it acts as an oxidant (e.g., oxidizing
SO2 to H2SO4), solar UV photolysis, and reaction with
atmospheric free radicals. In the mid- to lower-troposphere,
removal via rainout and uptake are the dominant loss processes
for H2O2. However, in the upper troposphere and the strato-
sphere, the reaction of H2O2 with OH
temperatures above 240 K because of the very small vapor
pressure of H2O2 at low temperatures. As a consequence, k1
has not been measured at the temperatures at which the reaction
is most important in the atmosphere.
Furthermore, it is not clear whether extrapolation of the higher
temperature data for k1 to the 190-250 K temperature range is
valid. Previous studies of the kinetics of reaction R1 show some
divergence in the measured values of k1 below room temper-
ature. Reaction R1 exhibits a small activation energy (Ea/R )
1
2
1
60 K from 300 to 400 K ) although the kinetic isotope effect
is large (e.g., k1 is ≈3 times larger than the rate coefficient for
7
the reaction of OH with D2O2 at 298 K ), suggesting a complex
k1
OH + H O2
9
8 HO + H O
(R1)
reaction pathway.
2
2
2
Hippler et al.10,11 studied this reaction in shock tubes at
temperatures up to 1680 K. They found that the reaction rate
coefficient dramatically increases with temperature at T > 900
K. This behavior is indicative of a change in the reaction
mechanism, probably from a mechanism involving complex
formation at low temperatures to direct hydrogen abstraction
at high temperatures.
and photolysis of H2O2 to yield two OH radicals become the
dominant loss processes for H2O2.
Formation of H2O2 and subsequent photolysis do not lead to
a net loss of HOX radicals. However, H2O2 formation followed
by reaction R1 leads to the loss of two HOX radicals. Since
HOX radicals play key roles in ozone formation in the upper
troposphere and catalytic ozone destruction in the lower
stratosphere, it is important to quantify any process, such as
reaction R1, that significantly affects HOX concentration.
Therefore, the rate coefficients for reaction R1 at the temper-
atures of the upper troposphere and the lower stratosphere (T
Curved Arrhenius plots have been observed in studies of
various reactions that involve the OH radical,
that the OH radical readily forms hydrogen-bonded complexes
1
3-15
indicating
1
5
with many H- and O-containing molecules. Recently Brown
1
4
et al. reported that the rate coefficient for the reaction of OD
)
190-250 K) are needed.
with DNO forms a U-shaped Arrhenius plot. They also reported
3
that the rate coefficients of the OD + HNO3 and OH + HNO3
*
Corresponding author. E-mail: srl@cchem.berkeley.edu.
reactions show weak sigmoidal pressure dependence, particularly
below room temperature. To explain the observed kinetic
behavior, Brown et al. successfully applied a simple kinetic
model that incorporates a reaction mechanism involving the
formation of an intermediate cyclic six-membered ring hydrogen-
†
‡
§
Part of the special issue “Charles S. Parmenter Festschrift”.
JILA. Present address: Southwest Sciences, Inc., Santa Fe, NM 87505.
NOAA and Department of Chemistry and Biochemistry and CIRES,
University of Colorado.
|
JILA and Department of Chemistry and Biochemistry, University of
Colorado. Present address: Departments of Chemistry and Physics and
Lawrence Berkeley National Laboratory, University of California, Berkeley,
CA 94720.
1
4
bonded complex. It is reasonable to expect that the OH +
H2O2 reaction may proceed through a similar cyclic complex.
1
0.1021/jp030424q CCC: $25.00 © 2003 American Chemical Society
Published on Web 11/11/2003

Reaction of OH with Hydrogen Peroxide
J. Phys. Chem. A, Vol. 107, No. 49, 2003 10643
appropriately collimate the supersonic flow. The background
pressure is measured by a capacitance manometer.
The expansion through the Laval nozzle results in a collimated
supersonic gas flow, which is characterized by a uniform Mach
number, gas number density, and temperature along the flow
1
7-19
axis for ≈20 cm.
ized by using the Pitot tube method,
temperature of the OH radicals from their LIF excitation
The expansion conditions are character-
17,18
measuring the rotational
1
8
19
spectra, and directly measuring the flow velocity. In this
work, three different Laval nozzles are used to create supersonic
flows of different temperatures: 96 ( 4 K, 110 ( 7 K, and
1
9
1
65 ( 14 K.
The reaction and pre-expansion chambers have quartz win-
dows, enabling introduction of laser beams along the axis of
the gas flow to generate radicals photolytically (using an excimer
laser) and for laser-based diagnostic techniques (using a
frequency-doubled tunable dye laser) within the cooled expan-
sion. In this work, LIF of OH is used as a probe. The laser
beams exit the reaction chamber through an arm containing light
baffles and a quartz Brewster-angle window, minimizing
scattered light.
Figure 1. Schematic diagram of the experimental apparatus used to
measure k .
1
Indeed, quantum chemical calculations predict that OH does
form a five-membered ring complex with H2O2, bound by 17.1
kJ/mol (4.1 kcal/mol).16 If such a complex is formed in reaction
R1, one would expect a significant deviation of k1 from
Arrhenius behavior at lower temperatures.
The OH radicals are produced by pulsed photolysis of H2O2
at 248 nm with the unfocused beam of a KrF excimer laser
For these reasons, it is important to measure the rate
coefficient for reaction R1 at low temperatures. However, as
mentioned earlier, conventional kinetic methods cannot be used
because the saturation vapor pressure of H2O2 is too low at these
temperatures. The Laval nozzle technique, however, is appropri-
ate to measure k1 at low temperatures.
(
≈30 mJ/pulse at a repetition rate of 10 Hz, fluence ≈8 mJ/
2
cm ). The frequency-doubled output radiation of a pulsed dye
laser pumped by the second harmonic of a Nd:YAG laser is
used for excitation of OH on the Q1(1) line of the (1,0) band of
2
+
2
1
7-19
the A Σ r X Πi transition. LIF from OH is detected in the
1,1) and (0,0) bands of the A f X transition by a photomul-
In this work, we use a pulsed Laval nozzle apparatus
combined with pulsed photolysis and laser-induced fluorescence
LIF) of the OH radical to measure low-temperature rate
coefficients for reaction R1 over the temperature range of 96-
65 K. The rate coefficient was also measured at 296 K, using
(
tiplier tube (PMT) equipped with a UV-band-pass filter and a
narrow-band interference filter (310 ( 10 nm). The PMT,
mounted 15 cm downstream of the Laval nozzle, detects the
light through a quartz lens (5-cm diameter, 5-cm focal length).
The lens focuses fluorescence from an ≈2 cm segment of the
irradiated zone onto the PMT photocathode. The optics are
appropriately shielded to minimize the effect of scattered light.
(
1
pulsed photolysis and LIF with a slow-flow system. The reaction
rate coefficient shows a strong negative temperature dependence,
which suggests that a complex-formation mechanism dominates
at low temperatures.
Temporal profiles of OH are recorded by monitoring the LIF
intensity versus the delay between the photolysis and probe laser
pulses. Typically, the signal from the photomultiplier is
integrated over a 200-ns gate after a delay of 200 ns with respect
to the probe laser pulse by a gated integrator. Active background
subtraction is performed as described previously.19 Normally,
Smith and co-workers have used the rate coefficient for the
removal of a vibrationally excited radical by another species as
a probe of the high-pressure limit (k∞) for the association of
the radical and the second species.1
5,20
An unusually large rate
coefficient of relaxation of a high-frequency vibration of the
radical may indicate that the energy transfer occurs via a
transient complex. In this work, we have also measured the rate
coefficient for the removal of OH (V ) 1) by H2O2
2
0-100 experimental runs were averaged to obtain one OH
temporal profile. The triggering of all units and devices is
provided by a four-channel digital delay/pulse generator. Both
the gated integrator and the pulse generator are interfaced to a
computer, which is used to control the experiment and acquire
data.
k2
98 loss of OH (V ) 1)
(R2)
2
OH (V ) 1) + H O2
at 296 K to further investigate the mechanism of reaction R1.
Nitrogen (99.999%), used as supplied from the cylinder, flows
through a stainless steel line and a calibrated mass flow
controller. An adjustable fraction of the N2 flow is bubbled
through a concentrated aqueous solution of hydrogen peroxide
(>95 weight %, measured by titration with potassium perman-
ganate) and then mixed into the main flow of the carrier gas.
Downstream of the bubbler, all the tubing supplying the N2/
H2O2 mixture is made of PFA fluoropolymer. The concentration
of hydrogen peroxide in the flow is determined by measuring
the absorbance at 213.9 nm (Zn lamp, σ213.9nm(H2O2) ) 3.2 ×
Experimental Section
The apparatus used in this work has been described
17-19
previously.
here. A schematic of the experimental setup is shown in Figure
. The low-temperature environment is created by supersonic
Therefore, only a brief description will be given
1
expansion of gas through a pulsed Laval nozzle, mounted in a
nozzle block inside a stainless steel reaction chamber. Two
pulsed solenoid valves supply an ≈5 ms gas pulse (nitrogen
with admixed hydrogen peroxide) to a pre-expansion chamber
in the nozzle block. The gas expands from the pre-expansion
chamber through the Laval nozzle into the reaction chamber,
which is pumped by a mechanical pump (60 L/s). A variable
gate valve is used to limit the pumping speed in order to
optimize the background pressure in the reaction chamber to
-
19
2 12,21
10
cm
) through a 100-cm absorption cell (see Figure
1). The total gas number density in the collimated supersonic
flow is measured using the Pitot tube method, as described
18,19
elsewhere.
The gas number density of H2O2 in the supersonic
collimated flow is calculated from the total number density in
the flow and the mole fraction of H2O2 in the gas mixture,

10644 J. Phys. Chem. A, Vol. 107, No. 49, 2003
Vakhtin et al.
Figure 3. First-order OH loss rate coefficients versus concentration
of H O obtained at room temperature. For each concentration, the error
2 2
bar indicates two standard deviations of the fit of the corresponding
experimental OH decay profile by the single-exponential function. The
2
line shows the 1/σ -weighted linear least-squares fit.
concentration. From the slope of the straight line that fits the
experimental data, the room-temperature rate coefficient is
-
12
3
-1
obtained as k1(296 K) ) (1.78 ( 0.19) × 10 cm molecule
-
1
s , where the quoted uncertainty represents 2σc (σc is the
“
combined standard uncertainty,” see below). This value is in
2-9
good agreement with most available literature data,
and the
Figure 2. Examples of the OH decay profiles obtained at room
temperature at different concentrations of H , plotted in normal (a)
and semilogarithmic (b) coordinates: 1 [H
-12
3
current recommendations, k1(298 K) ) 1.7 × 10
cm
2
O
2
-1 -1 12,21
molecule s .
13
2
O
2
] ) 9.1 × 10 molecules
-3 14
2 2
-
3
14
At 296 K, the OH temporal profiles were recorded over
several milliseconds (see Figure 2) in order to observe the decay
of the OH signal over about 2 orders of magnitude, since the
inverse decay time (k′) is limited by the absolute value of k1
cm ; 9 [H
molecules cm ; 2 [H
the fits to the single-exponential function (1).
2
O
2
-
] ) 4.5 × 10 molecules cm ; b [H
O
] ) 8.6 × 10
3
15
-3
O
2 2
] ) 1.3 × 10 molecules cm . Lines show
23
obtained from the absorption measurements. The room-tem-
perature value of k1 is obtained in an identical manner, using
the same apparatus, except that the N2/H2O2 mixture continu-
ously flows directly into the reaction chamber, bypassing the
Laval nozzle.
and the H2O2 saturation vapor pressure. While using the Laval
nozzle expansion, the time window available for kinetic
measurements is limited to about 200 µs by the length of the
uniform collimated supersonic flow and the flow velocity.1
Therefore, even with the highest attainable concentrations of
H2O2 (limited by its saturation vapor pressure at the temperature
of the laboratory and the throughput of the solenoid valves) in
the Laval nozzle flow, we would not expect to be able to observe
the decay of the OH signal over as large a dynamic range. As
expected, in the low-temperature experiments, even with the
highest obtainable H2O2 concentrations, we could detect only
the initial parts of the OH temporal profiles. Figure 4a shows
examples of OH temporal profiles obtained at “low” (1.7 ×
8,19
The rate coefficient for reaction R2 was measured in a slow-
flow system at 296 K using pulsed photolysis to produce OH
2
2
(V ) 1) and LIF to detect it, as described in detail elsewhere.
Briefly, H2O2 is photolyzed with 248-nm laser radiation; a small
fraction (on the order of 1%) of the OH produced by photolysis
of H2O2 at this wavelength is produced in the (V ) 1) state.
OH temporal profiles are measured using LIF: the Q1(1) line
of the (0, 1) band of the OH (A r X) electronic transition (≈346
nm) is excited by the probe laser; fluorescence in the (0, 0)
band of the same electronic transition is detected by a PMT,
mounted orthogonally to the photolysis and probe lasers and
shielded with a band-pass filter to reject scattered laser light.
1
2
-3
14
-3
1
0
molecules cm ) and “high” (2.2 × 10 molecules cm )
concentrations of hydrogen peroxide at T ) 110 K. It is seen
from Figure 4a that the “tail” (t > 30 µs) of the OH trace
obtained at “high” [H2O2] decays more steeply than that of the
trace obtained at “low” [H2O2]. Although the difference is not
large, it is consistent and reproducible. In numerous low-
temperature experiments performed at different concentrations
of H2O2, we always observed that the decay was steeper at
higher [H2O2]. The rise during the first 10-30 µs could be due
Results and Discussion
First, the room-temperature rate coefficient of reaction R1
was measured. Figure 2 shows examples of OH temporal
profiles obtained at different concentrations of H2O2. Since the
concentration of H2O2 was significantly larger (by a factor of
at least 100) than the initial concentration of OH, the kinetics
are pseudo-first-order in OH concentration. All measured
temporal profiles could be fitted reasonably well to a single-
exponential function. Figure 3 shows the first-order rate
coefficient for OH loss (i.e., k′) plotted versus the H2O2
2
4,25
to rotational relaxation of OH.
The ripples in the OH decay
profiles are mostly due to the inhomogeneity of the collimated
1
8
flow; they are reproducible and cannot be avoided.
To reduce the contribution of these ripples and of radial
diffusion to the OH temporal profiles, we normalized the
primary OH temporal profiles in each set of experiments by
dividing them by a reference OH temporal profile, usually the

Reaction of OH with Hydrogen Peroxide
J. Phys. Chem. A, Vol. 107, No. 49, 2003 10645
2 2
Figure 6. Pseudo-first-order plots for the OH + H O reaction obtained
at T ) 96 K (2), 110 K (O), and 165 K (9). The error bars indicate
two standard deviations of the fits of the corresponding normalized
experimental OH decay profiles by the single-exponential function (1).
The lines show the 1/σ -weighted linear fits for T ) 96 K (solid line),
110 K (dashed line), and 165 K (dotted line).
Figure 4. (a) Examples of the OH decay profiles obtained in Laval
nozzle expansion at T ) 110 K with “high” (b) and “low” (O)
2
concentrations of H
the OH decay profiles obtained at “high” [H
b) and the single-exponential fit (line).
2
O
2
. (b) Normalized OH profile, i.e., the ratio of
O
2 2
2 2
] and “low” [H O ]
(
TABLE 1: The Rate Coefficients of the OH + H
Reaction (k ) Measured at Different Temperatures; Buffer
Gas N
2 2
O
1
2
total gas density,
[H
2
O
2
], 1012
k
1
,a 10-12 cm3
1016 molecules cm
-3
molecule cm
-3
molecule
-1 -1
s
T, K
9
6 ( 4
1.9 ( 0.2
2.65 ( 0.45
3.5 ( 0.7
9.9 ( 0.1
1.2-160
1.7-220
9.8-590
3.4-1300
12.3 ( 4.5
10.1 ( 2.8
3.4 ( 1.4
1
1
2
10 ( 7
65 ( 14
96 ( 2
1.78 ( 0.19
a
The indicated uncertainties represent 2σ
c
, where σ
c
is a “combined
standard uncertainty” that includes both statistical and systematic errors.
where S0 and τ are the fitting parameters. It is assumed that the
OH decays exponentially in the presence of H2O2 at low
temperatures because pseudo-first-order conditions exist for OH.
This assumption could not be directly verified at low temper-
atures by obtaining a full OH temporal profile (reduction of
the OH signal by 2-3 orders of magnitude) with the present
Laval nozzle apparatus. We compared the fit of the full
normalized temporal profiles and truncated normalized temporal
profiles, which had the first three points removed, to eq 1 in
order to check for possible influence of the rise in the first part
of the temporal profiles (see Figure 4a) upon the obtained values
of k’. We observed no significant difference and used the fits
to the full temporal profiles.
Figure 6 is a plot for three different temperatures of the first-
order OH loss rate coefficients (k′), obtained by fitting the
normalized experimental OH temporal profiles to eq 1, versus
H2O2 concentration. The rate coefficients, obtained from the
slopes of the linear fits of the experimental data in Figure 6,
are presented in Table 1 along with the conditions under which
the rate coefficients were obtained. The reported uncertainties
of the measured rate coefficients represent 2σc, where σc is the
Figure 5. Normalized OH decay profiles obtained in Laval nozzle
. The
profiles are scaled to coincide at zero time. Lines show the single-
exponential fits to the experimental data.
2 2
expansion at T ) 110 K with different concentrations of H O
one acquired with the smallest H2O2 concentration. This
2
6
procedure has been described and justified by Orkin et al.
Figure 4b shows the normalized “high” H2O2 trace from Figure
4
5
a (i.e., the ratio of the two traces shown in Figure 4a). Figure
presents examples of normalized OH decay profiles obtained
“
combined standard uncertainty”, which incorporates both
statistical errors and the systematic errors. The values of σc were
at different concentrations of H2O2 at T ) 110 K. The lines
show the fits of the normalized decay profiles by the single-
exponential function
27
calculated according to NIST recommendations, as described
18,19
elsewhere.
For the room-temperature rate coefficient the
main contributions to σc are the statistical scatter of the data
and the uncertainty in the absorption cross section of H2O2 at
213.9 nm (assumed to be 5% ). For the low-temperature rate
S(t) ) S exp(-t/τ)
(1)
7
0

10646 J. Phys. Chem. A, Vol. 107, No. 49, 2003
Vakhtin et al.
(
k1 is ≈3 times larger than the rate coefficient for the reaction
7
of OH with D2O2 at 298 K ).
The rate coefficient for the removal of OH (V ) 1) by H2O2,
-
11
3
-1 -1
k2, is measured to be (3.5 ( 0.3) × 10
cm molecule
s
at 296 K. This rate coefficient includes an estimated correction
of about 20% to account for the quenching of OH (V ) 1) by
H2O, which is introduced into the flow, along with H2O2, when
N2 is bubbled through the H2O2/H2O solution. With a 95 weight
%
H2O2 solution, the effluent from the bubbler contains H2O2
2
3
and H2O in a mole ratio of about 2:1. For the quenching of
OH (V ) 1) by H2O, we use an average of several reported rate
2
8-31
-11
3
-1 -1
coefficients,
1.4 × 10
cm molecule s . This large
rate coefficient may corroborate the formation of the OH‚‚‚
1
5
H2O2 complex. However, loss of OH (V ) 1) by resonant
intermolecular energy transfer (i.e., without formation of the
complex) cannot be ruled out, since OH and H2O2 are closely
vibrationally resonant. For more definitive conclusions, mea-
surements of the rate coefficient for removal of OD (V ) 1) by
H2O2, where intermolecular V-V energy exchange is much less
important, are necessary. These measurements are now in
progress.
2 2
Figure 7. The rate coefficient of the OH + H O reaction obtained in
this work and taken from the literature, plotted in Arrhenius coordi-
nates: (solid circle) this work; (box with crossed lines) Baldwin and
1
2
3
Formation of an intermediate weakly bound complex, fol-
lowed by competing dissociation of the complex back to the
reactants and to the reaction products, may result in unusual
Walker; (triangle down) Keyser; (triangle left) Sridharan et al.;
4
5
(
triangle up) Wine et al.; (tilted square) Kurylo et al.; (open box)
Lamb et al.; (open circle) Vaghjiani et al.; (open star) Lovejoy et
6
7
8
9
al.; (triangle right) Turnipseed et al.; (crossed lines in circle) Hippler
temperature and pressure dependence of the overall reaction rate
1
0,11
et al.
The line shows the fit of the data obtained in this work by the
13-15
coefficient.
The particular kinetic behavior strongly depends
expression k ) A exp(-T*/T) (see text). The inset presents an expanded
view of the data obtained within the temperature range of 165-500 K.
on the barriers to the backward and forward dissociation of the
complex and the structure and “rigidity” of the corresponding
transition states. Quantum chemical calculations aimed at
evaluating the height and the profile of the “exit” barrier
(associated with dissociation of the complex into the reaction
products HO2 and H2O) are underway at NOAA. Knowledge
of these features of the potential energy surface will be helpful
in evaluating a kinetic model for this reaction, similar to that
coefficients, σc is dominated by the uncertainty in the total gas
number density in the Laval flow.18,19
Figure 7 is a plot of the rate coefficient k1 measured in this
work, together with the available literature data, as a function
of 1/T. At low temperatures k1 shows a negative temperature
dependence. The measured values of k1, obtained in this work,
can be described by the following expression:
1
4
successfully applied by Brown et al. for the OH + HNO3
reaction.
k1 ) (6.8 ( 1.0) × 10-13 exp[(285 ( 27)/T]
Given the dramatic non-Arrhenius behavior of reaction R1
over the temperature range shown in Figure 7, and the
uncertainty of the rate coefficients reported here, we must note
that the quality of the fit of our data to an Arrhenius expression
(with negative activation energy) may be simply fortuitous. In
other words, our data certainly do not prove that k1 reaches its
minimum near 296 K. Nevertheless, these data do show that k1
begins increasing with decreasing temperature at some point.
Some previous measurements of k1, particularly the measure-
cm molecule s-1
3
-1
T ) 96-296 K
(2)
2
This expression is obtained by 1/σ -weighted fitting of the
experimental data to the formula k ) A exp(-T*/T) with the
fitting parameters A and T*. The uncertainties indicated for A
and T* represent two standard deviations. The result of the fit
is shown in Figure 7 as a solid line. Although the fitting formula
is formally identical to the Arrhenius expression, A and T*
should be considered purely empirical parameters valid only
within the indicated temperature range, rather than physically
meaningful quantities.
6
ments of Lamb et al., do show an increase below room
temperature; however, these measurements are not considered
1
2,21
in current recommendations.
While it is possible that the
rate coefficient continues decreasing below room temperature
to ≈240 K or even lower, it is also possible that some previous
Within the full temperature range where experimental data
are available, 96 < T < 1680 K, the Arrhenius plot of the rate
2-5
measurements below room temperature have underestimated
k1 due to loss of H2O2 to the walls of the reactor at low
temperature.
1
1
coefficient k1 is U-shaped. As discussed by Hippler et al., the
sharp change at around 900 K is probably due to the reaction
mechanism changing from direct abstraction at high tempera-
tures to the formation of an intermediate complex at lower
For atmospheric purposes, direct measurements of k1 between
190 and 240 K are needed. However, the Laval nozzle expansion
method may not be successful at these temperatures, because
of the limited time scale available to monitor the kinetics with
this method. It will be particularly difficult to measure the rate
coefficient accurately at 190-240 K, because it will be smaller
at these temperatures than at the temperatures studied in this
work.
temperatures. This is in accord with quantum-chemical calcula-
tions by Wang et al.16 They used the CCSD(T)//B3LYP method
to locate a hydrogen-bonded OH‚‚‚H2O2 five-membered ring
-
1
complex with a binding energy of 17.1 kJ mol (4.1 kcal
-
1
mol ). The complex-formation mechanism is also supported
by the observations that, although reaction R1 at T < 900 K
apparently proceeds with little or no activation barrier, the
absolute value of k1 is relatively low in this temperature range,
and the kinetic isotope effect for the reaction is relatively strong
Interpolation of expression 2 for k1 to the temperature range
190-240 K can provide an estimate for k1 in the absence of
direct measurements of the rate coefficient at these temperatures.

Reaction of OH with Hydrogen Peroxide
J. Phys. Chem. A, Vol. 107, No. 49, 2003 10647
-
12
12
cm3
At 200 K, expression 2 gives k1 ) (2.8 ( 0.9) × 10
Acknowledgment. The authors gratefully acknowledge the
support of this research by the National Aeronautics and Space
Administration (Grant NAG5-8923, NAG5-13339, and the
Upper Atmosphere Research Program). D.C.M. acknowledges
a National Science Foundation Graduate Research Fellowship.
S.R.L. wishes to acknowledge the many profound interactions
with Professor Charles Parmenter throughout his career.
-
1
-1
molecule s , while extrapolation of the current JPL and
2
1
-12
3
IUPAC recommendation gives k1 ) (1.3 ( 0.6) × 10 cm
-
1
-1
molecule s . Clearly, extrapolation of the currently recom-
mended value of k112,21 to the lower temperatures of the
atmosphere may result in a value of k1 that is quite inaccurate.
If k1 is indeed higher than currently recommended, atmospheric
models underestimate the loss of HOX in the upper troposphere
and lower stratosphere due to reaction R1. It is also noteworthy
that expression 2 yields values of k1 around 240-250 K that
agree well with the value of k1 at these temperatures reported
References and Notes
(
1) Baldwin, R. R.; Walker, R. W. J. Chem. Soc., Faraday Trans. 1
1979, 75, 140.
(2) Keyser, L. F. J. Phys. Chem. 1980, 84, 1659.
3) Sridharan, U. C.; Reimann, B.; Kaufman, F. J. Chem. Phys. 1980,
3, 1286.
4) Wine, P. H.; Semmes, D. S.; Ravishankara, A. R. J. Chem. Phys.
1981, 75, 4390.
6
8
by Lamb et al. and Lovejoy et al.
(
It is possible that the reaction of OH with H2O2 that we
observe at low temperatures does not yield the products quoted
for reaction R1; i.e., the reaction we observe is only producing
an OH‚‚‚H2O2 complex that is not in turn dissociating to form
H2O and HO2. However, we do not believe that this is the case.
7
(
(5) Kurylo, M. J.; Murphy, J. L.; Haller, G. S.; Cornett, K. D. Int. J.
Chem. Kinet. 1982, 14, 1149.
(6) Lamb, J. J.; Molina, L. T.; Smith, C. A.; Molina, M. J. J. Phys.
The binding energy of the OH‚‚‚H2O2 complex is predicted to
Chem. 1983, 87, 4467.
be quite small, 17.1 kJ/mol,16 so that the rate coefficient for
(7) Vaghjiani, G. L.; Ravishankara, A. R.; Cohen, N. J. Phys. Chem.
989, 93, 7833.
1
formation of the OH‚‚‚H2O2 complex, if that complex could
only dissociate back to the reactants, would be quite small at
(8) Lovejoy, E. R.; Murrels, T. P.; Ravishankara, A. R.; Howard, C. J.
J. Phys. Chem. 1990, 94, 2386.
the pressures of these experiments. Using the Troe method for
(9) Turnipseed, A. A.; Vaghjiani, G. L.; Gierczak, T.; Thompson, J.
E.; Ravishankara, A. R. J. Chem. Phys. 1991, 95, 3244.
barrierless association reactions32,33 with the binding energy,
(10) Hippler, H.; Troe, J. Chem. Phys. Lett. 1992, 92, 333.
structure, and vibrational frequencies of the OH‚‚‚H2O2 complex
reported by Wang et al.,16 we estimated an upper limit for the
termolecular rate coefficients for formation of the OH‚‚‚H2O2
complex from reactants. The estimates show that, at the gas
number densities used in the experiments (see Table 1), the
bimolecular association rate coefficient would be at least 3
orders of magnitude smaller than the observed k1 between 96
and 296 K. This is evidence that the reaction of OH with H2O2
that we are observing is producing HO2 and H2O. We attempted
to experimentally verify that HO2 is formed by reaction R1 at
low temperatures by adding NO, which should react rapidly
with HO2 to form OH,34 to the gas mixture flowing through
the Laval nozzle. These attempts were not successful due to
experimental problems and were not further pursued.
(
11) Hippler, H.; Neunaber, H.; Troe, J. J. Chem. Phys. 1995, 103,
3510.
(
12) Sander, S. P.; Friedl, R. R.; Golden, D. M.; Kurylo, M. J.; Huie,
R. E.; Orkin, V. L.; Moortgat, G. K.; Ravishankara, A. R.; Kolb, C. E.;
Molina, M. J.; Finlayson-Pitts, B. J. Chemical Kinetics and Photochemical
Data for Use in Atmospheric Studies, EValuation 14; JPL Publication No.
2-25; Jet Propulsion Laboratory: Pasadena, CA, 2003.
(
0
13) Troe, J. J. Chem. Soc., Faraday Trans. 1994, 90, 2303.
14) Brown, S. S.; Burkholder, J. B.; Talukdar, R. K.; Ravishankara,
(
A. R. J. Phys. Chem. A 2001, 105, 1605.
(15) Smith, I. W. M.; Ravishankara, A. R. J. Phys. Chem. A 2002, 106,
4
798.
(16) Wang, B.; Hou, H.; Gu, Y. Chem. Phys. Lett. 1999, 309, 274.
17) Lee, S.; Hoobler, R. J.; Leone, S. R. ReV. Sci. Instrum. 2000, 71,
(
1816.
(18) Vakhtin, A. B.; Lee, S.; Heard, D. E.; Smith, I. W. M.; Leone, S.
R. J. Phys. Chem. A 2001, 105, 7889.
(
19) Vakhtin, A. B.; Murphy, J. E.; Leone, S. R. J. Phys. Chem. A, in
press.
(20) Smith, I. W. M. J. Chem. Soc., Faraday Trans. 1997, 93, 3741.
21) Atkinson, R.; Baulch, D. L.; Cox, R. A.; Hampson, R. F.; Kerr, J.
A.; Rossi, M. J.; Troe, J. J. Phys. Chem. Ref. Data 1997, 26, 521.
22) McCabe, D. C.; Brown, S. S.; Gilles, M. K.; Talukdar, R. K.; Smith,
I. W. M.; Ravishankara, A. R. J. Phys. Chem. A 2003, 107, 7762.
23) Scatchard, G.; Kavanagh, G. M.; Ticknor, L. B. J. Am. Chem. Soc.
952, 74, 3715.
24) Kliner, D. A. V.; Farrow, R. L. J. Chem. Phys. 1999, 110, 412.
(25) Sims, I. R.; Smith, I. W. M.; Bocherel, P.; Defrance, A.; Travers,
D.; Rowe, B. R. J. Chem. Soc., Faraday Trans. 1994, 90, 1473.
26) Orkin, V. L.; Huie, R. E.; Kurylo, M. J. J. Phys. Chem. 1996, 100,
907.
Since the supersonic expansion is greatly supersaturated with
H2O2, formation of dimers and larger clusters of H2O2 could
occur within the expansion. Also, H2O2 might form clusters with
water introduced into the flow, along with H2O2, when nitrogen
is bubbled through the H2O2/H2O solution. The rate coefficients
for the reactions of OH with the H2O2 dimers and clusters could
be significantly different than k1. However, formation of dimers
and clusters should be slow at the pressures of these experiments.
(
(
(
1
(
We again used the Troe method3
2,33
and the binding energy,
(
structure, and vibrational frequencies of the Ci dimer of H2O2
8
3
5
and the H2O2-H2O complex described by Gonz a´ lez et al. to
estimate upper limits for the rate coefficients for formation of
those complexes. The estimates show that these two processes
convert no more than 0.02% of H2O2 within the available time
at the temperatures of this study; therefore they cannot affect
the observed rate of OH decay.
(
27) Taylor, B. N.; Kuyatt, C. E. Guidelines for EValuating and
Expressing the Uncertainty of NIST Measurement Results; NIST Technical
Note 1297; US Government Printing Office: Washington, DC, 1994.
(
28) Kneba, M.; Wolfrum, J. Annu. ReV. Phys. Chem. 1980, 31, 47.
(29) Smith, I. W. M.; Williams, M. D. J. Chem. Soc., Faraday Trans.
2 1985, 81, 1849.
30) Raiche, G. A.; Jeffries, J. B.; Rensberger, K. J.; Crosley, D. R. J.
Chem. Phys. 1990, 92, 7258.
31) Silvente, E.; Richter, R. C.; Hynes, A. J. J. Chem. Soc., Faraday
(
In addition to direct studies of k1 in the temperature range
needed for atmospheric purposes, further elucidation of the
mechanism of this reaction could be provided by an investigation
of the primary and secondary H/D kinetic isotope effects on k1
at low and high temperatures. Direct study of the products of
reaction R1 at temperatures below 240 K would also be helpful.
(
Trans. 1997, 93, 2821.
(32) Troe, J. J. Chem. Phys. 1977, 66, 4745.
(
33) Troe, J. J. Chem. Phys. 1977, 66, 4758.
(34) Seeley, J. V.; Meads, R. F.; Elrod, M. J.; Molina, M. J. J. Phys.
Chem. 1996, 100, 4026.
(35) Gonzalez, L.; Mo, O.; Yanez, M. J. Comput. Chem. 1997, 18, 1124.