Formation of hydroxyl radical from the photolysis of frozen hydrogen peroxide

Article abstract of DOI:10.1021/jp051415f

Hydrogen peroxide (HOOH) in ice and snow is an important chemical tracer for the oxidative capacities of past atmospheres. However, photolysis in ice and snow will destroy HOOH and form the hydroxyl radical (·OH), which can react with snowpack trace species. Reactions of ·OH in snow and ice will affect the composition of both the overlying atmosphere (e.g., by the release of volatile species such as formaldehyde to the boundary layer) and the snow and ice (e.g., by the ·OH-mediated destruction of trace organics). To help understand these impacts, we have measured the quantum yield of ·OH from the photolysis of HOOH on ice. Our measured quantum yields (Φ(HOOH → ·OH)) are independent of ionic strength, pH, and wavelength, but are dependent upon temperature. This temperature dependence for both solution and ice data is best described by the relationship ln(Φ(HOOH → ·OH)) = -(684 ± 17)(1/T) + (2.27 ± 0.064) (where errors represent 1 standard error). The corresponding activation energy (E_a) for HOOH (5.7 kJ mol^-1) is much smaller than that for nitrate photolysis, indicating that the photochemistry of HOOH is less affected by changes in temperature. Using our measured quantum yields, we calculate that the photolytic lifetimes of HOOH in surface snow grains under midday, summer solstice sunlight are approximately 140 h at representative sites on the Greenland and Antarctic ice sheets. In addition, our calculations reveal that the majority of ·OH radicals formed on polar snow grains are from HOOH photolysis, while nitrate photolysis is only a minor contributor. Similarly, HOOH appears to be much more important than nitrate as a photochemical source of ·OH on cirrus ice clouds, where reactions of the photochemically formed hydroxyl radical could lead to the release of oxygenated volatile organic compounds to the upper troposphere.

Full text of DOI:10.1021/jp051415f

6264
J. Phys. Chem. A 2005, 109, 6264-6271
Formation of Hydroxyl Radical from the Photolysis of Frozen Hydrogen Peroxide
Liang Chu and Cort Anastasio*
Atmosphere Science Program, Department of Land, Air, and Water Resources, UniVersity of California,
One Shields AVenue, DaVis, California 95616-8627
ReceiVed: March 17, 2005; In Final Form: May 16, 2005
Hydrogen peroxide (HOOH) in ice and snow is an important chemical tracer for the oxidative capacities of
past atmospheres. However, photolysis in ice and snow will destroy HOOH and form the hydroxyl radical
•
(•OH), which can react with snowpack trace species. Reactions of OH in snow and ice will affect the
composition of both the overlying atmosphere (e.g., by the release of volatile species such as formaldehyde
•
to the boundary layer) and the snow and ice (e.g., by the OH-mediated destruction of trace organics). To
•
help understand these impacts, we have measured the quantum yield of OH from the photolysis of HOOH
•
on ice. Our measured quantum yields (Φ(HOOH f OH)) are independent of ionic strength, pH, and
wavelength, but are dependent upon temperature. This temperature dependence for both solution and ice data
•
is best described by the relationship ln(Φ(HOOH f OH)) ) -(684 ( 17)(1/T) + (2.27 ( 0.064) (where
errors represent 1 standard error). The corresponding activation energy (Ea) for HOOH (5.7 kJ mol-1) is
much smaller than that for nitrate photolysis, indicating that the photochemistry of HOOH is less affected by
changes in temperature. Using our measured quantum yields, we calculate that the photolytic lifetimes of
HOOH in surface snow grains under midday, summer solstice sunlight are approximately 140 h at representative
•
sites on the Greenland and Antarctic ice sheets. In addition, our calculations reveal that the majority of OH
radicals formed on polar snow grains are from HOOH photolysis, while nitrate photolysis is only a minor
contributor. Similarly, HOOH appears to be much more important than nitrate as a photochemical source of
•OH on cirrus ice clouds, where reactions of the photochemically formed hydroxyl radical could lead to the
release of oxygenated volatile organic compounds to the upper troposphere.
•
•
1. Introduction
impact on concentrations of HOOH, OH, and HO2 in the
boundary layer, which in turn affects local atmospheric photo-
chemistry. In addition to this thermal control on the air-snow
partitioning of HOOH, there is some preliminary evidence that
photochemical reactions on snow grains might also slowly form
HOOH in polar snowpacks.8
Atmospheric hydrogen peroxide (HOOH) is both an important
oxidant as well as a significant reservoir for the short-lived
•
hydroxyl (•OH) and hydroperoxyl (HO2 ) radicals.1 This is
•
especially significant, since OH levels largely determine the
oxidizing capacity of the troposphere and thus the lifetimes of
many atmospheric trace species.2 Because of its high solubility,
HOOH is a common trace constituent of snow and ice that
serves as a relatively stable tracer of the oxidative capacity of
past atmospheres.3,4 Ice core records described by Sigg et al.
have shown that HOOH concentrations in snow at Summit,
Greenland have increased about 50% over approximately the
last 200 years.5 These records, as well as other studies, have
also shown that hydrogen peroxide in snow has a very strong
seasonality, with summer concentrations approximately 10 times
higher than those in winter, making HOOH a useful tracer for
dating snow layers and ice cores.
Results from McConnell et al.6 indicate that HOOH concen-
trations in surface snow depend mainly on temperature and
atmospheric concentrations of HOOH, while snow grain size
also has some effect. Their long-term snowpack simulations
indicate that the firn exchanges HOOH with the atmosphere
for a decade or more after deposition. The air-snow exchange
of HOOH has been further explored by Hutterli and co-workers,
who have found that the snowpack can be a strong temperature-
dependent source or sink for HOOH.7 Their work has revealed
that this diurnal emission and deposition of HOOH has a large
Snowpack hydrogen peroxide is also potentially important
as a source of the highly reactive hydroxyl radical. It has already
been shown that the photolysis of nitrate, a common trace
•
species in snow and ice, is a source of snowpack OH and
•
NOx.9-12 This photoproduced OH could affect snowpack and
boundary layer chemistry by reacting with organic matter in
the snow to form volatile organic compounds, such as formal-
dehyde, which could be sources of HOx to the snowpack and
overlying atmosphere.8,13 Oxidation by •OH could also convert
snowpack halides into reactive gaseous halogens, which could
alter ozone and hydrocarbon chemistry in both the snowpack
and the atmosphere.14 Similarly, the photoformation of •OH from
HOOH could affect ice core records of not only HOOH but
•
also trace species that react with OH such as methane.
While the photolysis of hydrogen peroxide is a known source
•
of OH in solution,15-18 its photochemistry on ice has not yet
been quantitatively explored. In aqueous solution, ultraviolet
radiation splits hydrogen peroxide into •OH with a quantum yield
near unity
HOOH + hν f 2•OH
(1)
This quantum yield is similar both for tropospherically relevant
illumination wavelengths above 290 nm15 as well as for higher-
* To whom correspondence should be addressed. Phone: (530) 754-
6095. Fax: (530) 752-1552. E-mail: canastasio@ucdavis.edu.
10.1021/jp051415f CCC: $30.25 © 2005 American Chemical Society
Published on Web 06/29/2005

Hydroxyl Radical Formation from HOOH
J. Phys. Chem. A, Vol. 109, No. 28, 2005 6265
energy 253.7 nm radiation.16-18 In addition to this work in
aqueous solution, Jacobi and co-workers recently studied the
photochemical decomposition of HOOH on ice grains produced
in the laboratory using radiation from a 1000 W mercury lamp.19
They found that HOOH was relatively rapidly destroyed during
illumination and suggested that this reaction could be an
important sink for hydrogen peroxide as well as a source of
hydroxyl radical on polar snow grains.
frozen. Sample solutions contained known concentrations of
HOOH and benzoate in Milli-Q water (typically 100 µM HOOH
and 200 µM BA) and were adjusted to the desired pH using
sulfuric acid (pH e 5) or borate (pH g 6). All listed
concentrations (e.g., BA and HOOH) and pH values in this paper
are for the sample solutions prior to freezing unless specified
otherwise.
Ice samples were illuminated for known times using 313 nm
(or 334 nm) light from a 1000 W Hg/Xe monochromatic
system.23 Samples were held in a custom-designed (Paige
Instruments), Peltier-cooled, temperature-controlled, Teflon-
coated copper chamber with a quartz window to transmit light
onto the sample. Because of the relatively small photon fluxes
in our system, the low concentrations of hydrogen peroxide used,
and the short path lengths employed (∼1 mm for the sample
solution ice), illumination should have caused only negligible
heating of the ice sample.10 At the end of illumination, the
complete ice sample (frozen sample solution and pure water
ice base) was melted in the dark at room temperature in about
Our goal in this current work was to quantitatively determine
the quantum yields of •OH from HOOH photolysis on ice. These
values are needed to understand the impacts of HOOH pho-
tolysis as a source of snowpack •OH (and as a sink for HOOH)
in the snowpack, in deeper ice, and on atmospheric ice particles.
•
We have determined quantum yields for OH (i.e., Φ(HOOH
•
f OH)) on illuminated ice pellets as a function of pH, ionic
strength, and wavelength. In addition, we have measured the
•
temperature dependence of Φ(HOOH f OH) for both ice
pellets and aqueous solutions.
•
5-10 min, and then the melted mixture was analyzed for OH
2. Experimental Methods
and HOOH as described above.
2.1. Materials. Hydrogen peroxide, acetonitrile (Optima),
sodium borate (certified ACS), sulfuric and perchloric acids
(Optima), and potassium hydrogen phthalate were from Fisher.
Benzoic acid (99%), sodium benzoate (99%), and horseradish
peroxidase were from Aldrich, while p-hydroxybenzoic acid
(98%) and p-hydroxyphenylacetic acid were from TCI America.
All chemicals were used as received. Purified water (“Milli-
Q”) was obtained from a Milli-Q Plus system (g18.2 MΩ cm).
•
2.4. Calculation of OH Quantum Yield. Initial rates of
p-HBA formation during illumination (R/p-HBA,λ) were deter-
mined from plots of [p-HBA] versus illumination time using a
linear regression fit. For each illuminated set of ice samples,
rates of p-HBA formation were also measured in two controls:
(1) in a dark control under identical conditions (sample
composition and temperature) except for no illumination (RDark
)
and (2) in an illuminated blank control with identical conditions
as the sample except that no HOOH was added (RBlank,λ). The
corrected formation rate of p-HBA in the illuminated sample
•
2.2. OH and HOOH Measurements. Hydroxyl radicals
were measured using a chemical probe technique where photo-
formed •OH reacts with benzoate/benzoic acid (together referred
to as “BA” or “benzoate” for simplicity) to form stable,
measurable products that include p-hydroxybenzoate (p-
HBA).10,20,21 Stock solutions of BA contained 1.4 mM benzoic
acid and 12.6 mM sodium benzoate. The BA concentrations
used in all sample solutions in this study were high enough to
was then calculated using Rp-HBA,λ ) R/p-HBA,λ - RDark
-
R
Blank,λ. These blank corrections were small: On average,
((1 σ) RDark and RBlank,λ represented 7 ( 4% and 2 ( 2%,
respectively, of the value for R/p-HBA,λ. Corrected rates of
•
p-HBA formation were converted to rates of OH formation,
•
R(HOOH f OH)λ, by using
•
scavenge all photoformed OH (i.e., [BA]:[HOOH] g 1). We
R(HOOH f •OH)λ ) Rp-HBA,λ/Yp-HBA
(2)
also tried a number of experiments with benzene (instead of
benzoate) as a chemical probe for •OH. However, these
experiments were unsuccessful, because a significant fraction
of the added benzene evaporated from the ice pellet either during
sample preparation or subsequent sample handling.
•
where Yp-HBA is the molar yield of p-HBA from the OH +
BA reaction. We determined values of Yp-HBA using ice pellets
made from pH 5.0 solutions containing 10 µM BA, 4 µM
HOOH, and Na2SO4 to adjust the total ionic strength to between
198 and 628 µM. During 313 nm illumination of the pellets at
263 K, we simultaneously measured the formation rate of
p-HBA (Rp-HBA,λ) and loss rate of BA (RBA,λ) and then
Concentrations of p-HBA were measured using an HPLC/
UV system described earlier,20 using calibration standards made
in Milli-Q and run during each experiment. HOOH was
measured using an HPLC technique with postcolumn deriva-
tization and fluorescence detection.22 Analyses were performed
with a Shimadzu LC-10 AT VP pump and RF-551 fluorescence
detector (excitation at 320 nm; emission at 400 nm) using an
eluent of 0.1 mM H2SO4 and 1 mM Na2EDTA in Milli-Q at a
flow rate of 0.60 mL min-1. Analyses were performed with a
5001-CS guard column and PEEK P#0296-250 × 046 analyti-
cal column (MetaChem). Concentrations of HOOH were
determined on the basis of calibration standards made in Milli-Q
and run during the same day of the experiment.
determined yields of p-HBA using Yp-HBA ) Rp-HBA,λ/RBA,λ
.
The average ((σ) value of Yp-HBA from these four experiments
was 0.083 ( 0.011, which is consistent with the value (0.081
( 0.014) we determined previously by a somewhat different
method.10 Since this yield is independent of temperature between
243 and 268 K (ref 10), we used Yp-HBA ) 0.083 for the
temperature-dependence and ionic-strength-dependence experi-
ments with frozen pH 5.0 sample solutions. For the pH-
dependence experiments, we used the relationship Yp-HBA
)
2.3. Ice Sample Preparation and Illumination. The tech-
niques used for ice pellet preparation and illumination are
described fully in Chu and Anastasio10 and are summarized here.
Samples were prepared in a refrigerated chamber by first
pipetting 300 µL of Milli-Q water into the well created by a
Teflon template placed on a quartz slide backing. After this
pure water ice “base” was frozen, the chamber was kept at the
experiment temperature (239-268 K), and 100 µL of sample
solution was pipetted onto the ice base and allowed to sit until
-0.0011(pH)2 + 0.0154(pH) + 0.0318 (R2 ) 0.978; ref 10).
As we reported previously,10 we also found here that values of
Yp-HBA decrease with decreasing ionic strength, in this case,
for ice made from sample solutions with I < 198 µM. To avoid
the uncertainty introduced by this variation in Yp-HBA, our
sample solutions in this work all had ionic strengths of at least
198 µM.
We used 2-nitrobenzaldehyde (2NB) as a chemical actinom-
eter to measure the surface-area-normalized photon flux (Iλ ×

6266 J. Phys. Chem. A, Vol. 109, No. 28, 2005
Chu and Anastasio
l) in the sample chamber during illumination. The loss of 2NB
was measured in ice pellets of the same size and composition
as the hydrogen peroxide-containing samples, except that 4 µM
2NB was also added. Under the low light-absorbing conditions
of our actinometry, the measured rate constant for 2NB loss
(j2NB,λ) is related to the photon flux through23
j
2NB,λ ) 2.303Iλlꢀ2NB,λΦ2NB,λ
(3)
where (ꢀ2NB,λΦ2NB,λ) is the product of the molar absorptivity
and quantum efficiency for 2NB (640 M-1 cm-1 at 313 nm
and 300 M-1 cm-1 at 334 nm, both at 293 K; ref 23), Iλ is the
volume-normalized photon flux (mol-photons L-1 s-1), and l
is the effective path length of the sample (cm). Actinometry
performed on ice pellets with and without the chemical
components used for the •OH measurements (HOOH, BA, and
H2SO4) showed that the presence of these chemicals had no
effect on measured values of j2NB,λ. We also found that j2NB,λ
was independent of temperature in the range 243-268 K,
indicating that room-temperature values for the product
Figure 1. Base-10 molar absorptivities of aqueous hydrogen peroxide
at 274 K (solid line) and 298 K (dashed line). Values of ꢀHOOH,λ as a
function of temperature and wavelength are tabulated in the Supporting
Information (Table S1).
ice pellets made from low ionic strength solutions, the thin QLL
will result in higher concentrations of HOOH; if the concentra-
tion is greater than the ice-air partitioning equilibrium value,
then a portion of the HOOH will evaporate.
(ꢀ2NB,λΦ
2NB,λ) can be used at lower temperatures.10
Under our low light-absorbing conditions, the initial rate of
•OH formation during illumination of HOOH with radiation of
wavelength λ is
Rearranging eq 3 to solve for Iλ and substituting that into eq
4 produces the expression used to calculate the quantum yield
•
of OH
R(HOOH f •OH)λ )
ꢀ2NB,λ
Φ
2NB,λR(HOOH f •OH )λ
2NB,λꢀHOOH,λ[HOOH]0
2.303Iλ l ꢀHOOH,λΦ(HOOH f •OH)λ[HOOH]0 (4)
Φ(HOOH f •OH)λ )
(5)
j
where ꢀHOOH,λ is the molar absorptivity of hydrogen peroxide
(0.37 M-1 cm-1 at 313 nm and 274 K; see Table S1 of
Supporting Information) and Φ(HOOH f •OH)λ is the quantum
Note that we express the quantum yield here as the number of
moles of OH produced per mole of photons absorbed by
•
•
yield of OH from hydrogen peroxide photolysis. [HOOH]0 is
HOOH, and thus, the maximum value is 2 because of the
reaction stoichiometry (eq 1). Thus, the quantum efficiency for
HOOH destruction (i.e., the number of moles of HOOH lost
per mole of photons absorbed) is one-half the value of Φ(HOOH
the initial molar concentration of hydrogen peroxide, determined
as the HOOH concentration measured after illumination plus
the small amount lost during the illumination, that is, [HOOH]0
1
•
•
) [HOOH]meas + /2R(HOOH f OH)λ × t, where t is the
illumination time. We found that it was important to measure
HOOH concentrations in the ice pellets, rather than assume the
values were the same as those initially delivered in the solution
(100 µM), because the pellets often gained or lost a small
amount (<10%) of HOOH during experiments. This change in
HOOH depended both upon the ionic strength of the sample
solution as well as the ice pellet temperature. For ice made from
a sample solution with I ) 180 µM, the ratio of measured
[HOOH] to added [HOOH] increased from 1.04 to 1.10 with
decreasing temperature between 268 and 239 K, but the
correlation was weak (R2 ) 0.20). We also tested HOOH in a
series of pellets at 263 K made from sample solutions with ionic
strengths from 45 µM to 6180 µM. Here, the ratio of measured
[HOOH] to added [HOOH] increased from 0.96 to 1.10 over
this range of increasing ionic strength (R2 ) 0.53). We also
tested ice made from pure water (with no added HOOH) and
found that, after melting, the samples contained only relatively
small amounts of HOOH (<0.5 µM). This indicates that the
uptake of HOOH on our ice pellets is not due to contamination
during sample preparation. Instead, our results suggest that the
sample pellets are exchanging HOOH with the ambient atmo-
sphere. The greater uptake of HOOH at lower temperatures is
consistent with the temperature dependence of ice-air partition-
ing,24 while the ionic strength dependence likely represents the
influence of the thickness of the quasi-liquid layer (QLL). At
high ionic strength, the resulting QLL will be thicker and the
concentration of HOOH in the QLL relatively low, making it
more likely for gas-phase HOOH to dissolve into the QLL. For
f OH)λ.
2.5. Molar Absorptivities of Hydrogen Peroxide as a
Function of Temperature. Absorbance spectra of four aqueous
HOOH solutions (6.5-26 mM) and a Milli-Q blank were
measured in a Shimadzu UV-2501PC spectrophotometer using
a stirred 1.0 cm quartz cell and Milli-Q as reference. The sample
cell holder was cooled (274-298 K) using chilled water from
a recirculating water bath. The molar absorptivity at each
wavelength (240-400 nm) was determined as the slope of the
linear regression fit to the data of absorbance versus [HOOH].
Our resulting molar absorptivities for aqueous HOOH at 298
K (Figure 1) are very similar to those reported previously.25
Comparing the 298 K measurements with those made at 274 K
(Figure 1) shows that ꢀHOOH,λ changes very little as a function
of temperature across the range examined here. In particular,
the molar absorptivity of hydrogen peroxide at 313 nm exhibits
no significant dependence on temperature, with an average value
of 0.386 ( 0.016 M-1 cm-1 between 274 and 298 K. The best
fit values of the molar absorptivities at five temperatures in this
range are listed in the Supporting Information (Table S1).
Interestingly, although the HOOH molar absorptivities have little
temperature dependence, a plot of this relationship (i.e., the slope
of ꢀHOOH,λ vs T) as a function of wavelength reveals some
intriguing structure at wavelengths below 300 nm (Figure S1
of the Supporting Information).
•
2.6. Measurements of Φ(HOOH f OH)λ in Aqueous
Solution. Aqueous solutions contained 100 µM HOOH and 200
µM BA and were adjusted to pH 5.0 using sulfuric acid. Samples
were illuminated with 313 nm light in stirred, airtight, 1 cm

Hydroxyl Radical Formation from HOOH
J. Phys. Chem. A, Vol. 109, No. 28, 2005 6267
•
Figure 2. Quantum yields for OH in ice pellets (263 K; 313 nm
•
illumination) as a function of ionic strength in the sample solution.
Each sample solution contained 100 µM HOOH, 200 µM BA, was
adjusted to pH 5.0 with H2SO4, and contained various amounts of
Na2SO4 to adjust the ionic strength. Error bars represent (1 σ,
calculated on the basis of propagated uncertainties in Rp-HBA,λ, Iλ,
Figure 3. Quantum yields of OH from HOOH photolysis (open
triangles) in ice pellets (263 K, 313 nm illumination) as a function of
pH in the sample solution (100 µM HOOH and 200 µM BA). For
comparison, the quantum yields of •OH from nitrate photolysis are also
plotted: open circles represent data from Chu and Anastasio,10 while
the two closed circles were determined in this present work. Error bars
represent (1 σ, calculated as described in Figure 2. The pH values
listed in parentheses are the estimated values in the quasi-liquid layer
of the ice pellets. The solid line through the HOOH data is the average
value (0.70 ( 0.031), while the dashed line through the nitrate data is
an exponential regression fit.10
Yp-HBA, ꢀHOOH,λ, [HOOH], and the p-HBA calibration curve. The
horizontal line is the average value (0.72 ( 0.017).
quartz cells and small aliquots were removed at measured time
intervals and analyzed for p-HBA. Rates of •OH formation were
determined using the same procedure described for the ice pellets
(eq 2), except that the value of Yp-HBA in aqueous solution is
(0.19 ( 0.01).10 The quantum yield of •OH was calculated using
eq 5, where j2NB,λ was the value measured in aqueous 2NB
(4 µM) on the same day as the hydrogen peroxide photolysis
experiment.
3. Results and Discussion
•
3.1. Quantum Yield of OH in Ice as a Function of Ionic
•
•
Strength. Quantum yields of OH (Φ(HOOH f OH)) from
hydrogen peroxide photolysis on ice were first measured at 263
K in frozen sample solutions at pH 5.0 that contained 100 µM
HOOH, 200 µM BA, and variable amounts of Na2SO4 to adjust
the ionic strength (I). As shown in Figure 2, the plot of Φ-
•
(HOOH f OH) versus I has a slope ((1 standard error) that
is not significantly different from zero ((5.1 ( 26) × 10-7; p
< 0.84), indicating that the ionic strength of the sample solution
has no effect on •OH quantum yields within the range of ionic
strengths tested (198-6198 µM). The average value of
Figure 4. Temperature dependence (T ) 239-318 K) of Φ(HOOH
•
f OH) in aqueous and frozen sample solutions containing 100 µM
HOOH, 200 µM BA, and adjusted to pH 5.0. The circles are measured
data from the present study in both aqueous solution and ice pellets,
with errors of (1σ. The solid line is a linear regression fit to our
combined solution and ice data (eq 6). The inverted triangles (with
dotted regression line) are solution data from Zellner et al.15
•
Φ(HOOH f OH) in these experiments was 0.72 ( 0.017,
which is over 200 times higher than the corresponding value
obtained from nitrate photolysis at the same pH and temperature
((3.4 ( 0.6) × 10-3; ref 10).
in Figure 3, these new measurements are essentially identical
to those previously reported. When combined, these results for
•
3.2. Quantum Yield of OH on Ice as a Function of pH.
•
To explore the pH dependence of Φ(HOOH f OH) on ice,
-
•
HOOH and NO3 indicate that the pH dependence for OH
formation during nitrate photolysis is real (e.g., it is not an
artifact of our BA technique), although its cause is still
unexplained.10
we measured values in frozen sample solutions containing 100
µM HOOH and 200 µM BA and with pH values from 2.0 to
7.0. As shown in Figure 3, the •OH quantum yield was
essentially constant in these samples, with a slope for the
•
3.3. Temperature Dependence of OH Quantum Yields.
•
regression line of Φ(HOOH f OH) versus pH of 0.006 (
As the first step in determining the temperature dependence of
0.004 (p < 0.93). The average value ((1 standard error) for
Φ(HOOH f •OH) in these samples was 0.70 ( 0.031,
consistent with the value from the ionic strength experiments
described above. This absence of a pH effect on Φ(HOOH f
•OH) is different from the previously reported case for nitrate,
•
HOOH photolysis, we measured values of Φ(HOOH f OH)
in aqueous solutions (100 µM HOOH, 200 µM BA, pH 5.0)
•
between 278 and 318 K. As shown in Figure 4, our OH
quantum yields in solution are very similar to those previously
reported by Zellner and co-workers 15 and follow the regression
•
-
where quantum yields of OH from NO3 photolysis on ice
increased from 2.1 × 10-3 at pH 2.0 to 3.6 × 10-3 at pH 7.0
(Figure 3; ref 10). As a check on this pH dependence for nitrate,
we remeasured Φ(NO3- f •OH) in two sets of ice pellets made
from a pH 2.0 sample solution. As shown by the filled circles
•
ln(Φ(HOOH f OH)) ) -(679 ( 53)(1/T) + (2.25 ( 0.016)
(where errors represent 1 standard error and n ) 12). The
corresponding activation energy (Ea ) 5.6 ( 0.44 kJ mol-1
)
and change in entropy (∆S ) 18.7 ( 0.13 J mol-1 K-1) for our

6268 J. Phys. Chem. A, Vol. 109, No. 28, 2005
Chu and Anastasio
TABLE 1: Values of Ea and ∆S ((1 standard error) for the
While the temperature data described above indicate that the
photolysis of HOOH occurs in a quasi-liquid layer in our
experiments, it should be noted that the distribution of hydrogen
peroxide in our ice pellets might be different than that in natural
snow. Because we freeze the aqueous hydrogen peroxide
solution relatively slowly (over the course of several minutes),
essentially all of the solutes in the ice pellet probably reside in
quasi-liquid layers at grain boundaries and the air-ice interface.
In natural snow, it has been suggested that most snow grain
HOOH resides in the bulk ice as a result of co-condensation
with water vapor during snow formation,26 although appreciable
amounts (∼20%) are also present in the QLL.27 Although
HOOH in bulk ice will still undergo photolysis to produce free
•OH, the quantum yield for this process is likely lower than
that reported here (eq 6) because of reduced diffusion and
enhanced •OH recombination in the ice matrix. These potential
differences in distribution are not an issue in the case of frozen
nitrate, since NO3- is present in the QLL both in natural snow27
as well as in our previous experiments.10
•
formation of OH from hydrogen peroxide photolysis in ice
and aqueous solutiona
Ea
∆S
(kJ mol-1
)
(J mol-1 K-1
)
Solution Values
5.6 ( 0.44
6.0 ( 0.40
this study
18.7 ( 0.13
19.9 ( 1.3
Zellner et al.b
Ice Values
5.8 ( 0.34
this study
this study
19.5 ( 0.83
18.9 ( 0.53
Combined Solution and Ice Data
5.7 ( 0.14
a The activation energy (Ea) and the change in entropy (∆S) were
calculated on the basis of the linear regression fits to the data of
•
ln(Φ(HOOH f OH)) vs 1/T: Ea ) -slope × R × 10-3 (kJ mol-1
)
and ∆S ) y-intercept × R (J mol-1 K-1), where R is the gas constant
(8.314 J mol-1 K-1). b Values calculated on the basis of the pH 4-9
data of Zellner et al.15
solution data are essentially identical to those previously reported
(Table 1).
3.4. Wavelength Dependence. We also measured Φ(HOOH
f •OH) during illumination with 334 nm radiation at 263 K to
test whether the quantum yield is wavelength dependent.
Because of the lower molar absorptivity of hydrogen peroxide
at 334 nm (0.083 M-1 cm-1; Table S1), we used ice pellets
made from sample solutions containing a relatively high
hydrogen peroxide concentration (1.0 mM) along with 1.0 mM
BA and with a pH of 5.0. The resulting value ((1σ) of
To complete our temperature dependence study, we also
•
measured quantum yields for OH in a series of ice pellets (T
) 239-268 K) prepared from sample solutions identical to those
described above. As shown in Figure 4, our values ((1σ) of
Φ(HOOH f •OH) on ice decreased monotonically with
decreasing temperature, ranging from 0.76 ( 0.11 at 268 K to
0.55 ( 0.093 at 239 K. The corresponding regression line for
•
the ice pellet data is ln(Φ(HOOH f OH)) ) -(702 ( 41)-
•
Φ(HOOH f OH)334 at 263 K was 0.69 ( 0.0082, which is
(1/T) + (2.34 ( 0.10) (where errors represent 1 standard error
and n ) 27), which is almost identical to the regression equation
for our aqueous solution data described above. Similarly, the
activation energy and entropy change calculated from the ice
data are statistically indistinguishable from the corresponding
values from our aqueous solution data and from those of Zellner
and co-workers (Table 1). The fact that our ice pellet and
aqueous solution data have the same temperature dependence
suggests that the photolysis of hydrogen peroxide on ice between
239 and 268 K in our experiments occurs in a quasi-liquid, or
disordered, layer rather than in the bulk ice. This is the same as
the case for the photolysis of frozen aqueous nitrate, as described
previously.9,10
statistically indistinguishable from the average value (0.72 (
0.017) obtained with 313 nm illumination of the frozen pH 5.0
sample solutions at 263 K. This suggests that the quantum yield
•
of OH from HOOH photolysis is independent of illumination
wavelength throughout at least the UV-A and UV-B portions
of the HOOH absorbance tail (i.e., for at least λ > 280 nm;
Figure 1).
4. Environmental Implications.
•
4.1. OH Formation in Arctic and Antarctic Snow. Our
goal here is to use our measured quantum yields to estimate
the relative importance of NO3- and HOOH as photochemical
sources of •OH in polar snowpacks. The rate of •OH formation
from the photolysis of chromophore i on snow grains can be
expressed as
A regression fit to all of our solution and ice data yields the
relationship
ln[Φ(HOOH f •OH)] ) -(684 ( 17)(1/T) +
(2.27 ( 0.064) (6)
R(i f •OH)λ ) 2303/NA × I′ꢀi,λΦ(i f OH)λ[i] )
λ
j(i f •OH)λ[i] (7)
(where errors represent 1 standard error and n ) 39), which is
very similar to the regressions for the separate ice and aqueous
data and which yields similar thermodynamic parameters (Table
1). Given these similarities, we recommend that the regression
R(i f •OH)SUN ) 2303[i]/NA ×
[I′ꢀ , Φ(i f •OH) ∆λ] ) [j(i f •OH) ∆λ][i] (8)
•
from the combined data (eq 6) be used to determine OH
∑
∑
λ
i λ
λ
λ
quantum yields for both aqueous solutions and ice grains. As
shown in Figure 4, this regression line is very similar to that
previously reported for aqueous solution. For example, calcu-
where R(i f •OH)λ is the rate during illumination with radiation
within a given wavelength interval (mol L-1 s-1 nm-1) and
R(i f •OH)SUN is the corresponding rate under sunlight
•
lated values of Φ(HOOH f OH) from our recommended
regression (eq 6) at 298 and 239 K are 0.4% lower and 2.7%
higher, respectively, compared to values calculated from the
extrapolated regression of Zellner et al.15 Finally, it is interesting
to note that the temperature dependence for •OH formation from
HOOH photolysis (Ea ) 5.7 ( 0.14 kJ mol-1) is much smaller
than that from nitrate photolysis (Ea ) 20 ( 0.89 kJ mol-1; ref
illumination (mol L-1 s-1). In these equations, I′ is the
λ
spherically integrated actinic flux (photons cm-2 s-1 nm-1), ∆λ
is the wavelength interval (i.e., the width of the actinic flux
interval centered at wavelength λ in units of nm), NA is
Avogadro’s number, ꢀi,λ is the base-10 molar absorptivity for
-
chromophore i (e.g., NO3 or HOOH, see Table S1 of
10). Thus, not only is Φ(HOOH f OH) much higher than
Supporting Information and ref 10; units of M-1 cm-1), Φ(i f
•OH)λ is the quantum yield of •OH (eq 6 and ref 10), and [i] is
the molar bulk concentration of chromophore in the snow. As
shown in the top panel of Figure 5, while both hydrogen
•
-
•
Φ(NO3 f OH) in aqueous solution, but this difference is
magnified on snow and ice and increases with decreasing
temperature.

Hydroxyl Radical Formation from HOOH
J. Phys. Chem. A, Vol. 109, No. 28, 2005 6269
shown in Table 2, we have calculated the rate constants (j(i f
•OH)) and rates (R(i f •OH)) of •OH formation on snow grains
for four polar locations on days typical for photochemistry
experiments at a given site. At all sites, the measured or
estimated snow concentrations of HOOH are comparable to, or
larger than, the corresponding concentrations of nitrate. In
•
addition, calculated rate constants for OH formation from
HOOH photolysis are 9-22 times greater than those from
nitrate, with the greatest differences occurring under conditions
with the coldest temperatures. Combining these factors reveals
that HOOH is much more important than NO3- as a source of
•
snow grain OH under the conditions examined. Midday rates
•
of OH formation from hydrogen peroxide photolysis are 11-
140 times greater than the corresponding rates from nitrate
•
photolysis, that is, HOOH accounts for 93-99% of the OH
photochemically formed from these two chromophores (Table
2). While the absolute rates of •OH formation from HOOH and
NO3- at any given location will depend on snow composition,
temperature, and actinic flux, our simple overview in Table 2
indicates that hydrogen peroxide is, overall, much more
•
important than nitrate as a source of OH in both Arctic and
Antarctic snowpacks. It should be noted, however, that there
could be additional photochemical (or thermal) sources of •OH
in snow and that our calculations need to be tested by
•
measurements of OH in the field.
Table 2 also reveals the approximate photolytic lifetimes of
HOOH on polar snow grains. Given that 2 •OH are formed from
the photolysis of every HOOH molecule (eq 1), the rate constant
for the photochemical destruction of HOOH will be one-half
•
of the rate constant for OH formation. Thus, midday rate
constants for HOOH destruction under the conditions listed in
Table 2 range from 0.29 × 10-6 s-1 at Alert on the equinox to
2.0 × 10-6 s-1 at Summit on the boreal summer solstice.
Corresponding lifetimes of HOOH with respect to direct
photolysis are 960 h at Alert and 140 h at Summit under these
conditions. In addition, our calculated lifetime for photolysis
of HOOH on surface snow at Neumayer on the austral summer
solstice is 150 h, which is approximately 4 times longer than
that estimated by Jacobi et al. from their experiments of
snowbound HOOH illuminated with an optically filtered lamp.19
It should be noted that all of these lifetimes are for midday
sunlight conditions and that snow grain HOOH lifetimes in the
field will be several times longer as a result of the diurnal
changes in actinic flux (except for the South Pole where the
diurnal changes in actinic flux are extremely small).
•
Figure 5. Action spectra for OH formation from the photolysis of
HOOH and NO3- on ice. The top panel shows molar absorptivities of
aqueous hydrogen peroxide (solid line) and nitrate (dashed line) at 274
K as well as the modeled midday, actinic flux on the summer solstice
at Neumayer, Antarctica (70.7 °S, 8.3 °W; dotted line; ref 52). The
•
middle panel shows quantum yields of OH from the photolysis of
HOOH (squares; this work) and nitrate (circles; ref 10) in ice pellets
at 263 K. Symbols show data for pellets illuminated with 313 and 334
nm radiation, while the lines represent the recommended quantum yields
at 263 K for each chromophore at λ > 290 nm. The bottom panel
shows the wavelength dependence of the rate constants for •OH
formation from photolysis of HOOH (solid line) and nitrate (dashed
•
line). The area under each curve is the total rate constant for OH
formation from that chromophore (j(i f •OH)) at Neumayer under the
These relatively short photolytic lifetimes of HOOH are in
contrast with field measurements, which have shown that
hydrogen peroxide concentrations can be preserved both in the
near-surface snowpack as well as in deeper ice cores4,5 These
field observations reveal that the net loss of HOOH from snow
grains is much smaller than predicted from our photolysis
calculations. One possibility for this behavior is that the majority
of snow grain HOOH is sequestered in bulk ice where photolysis
is likely to be less efficient (section 3.3). However, since
photoproducts can migrate out of a water ice cage at temper-
atures as low as 140 K,28 even HOOH in the bulk ice of polar
snowpacks should undergo photodestruction. An alternate
possibility to explain the preservation of HOOH in polar snow
and ice is that there is a source of snowpack HOOH of
approximately the same magnitude as the photolytic sink.
Deposition of atmospheric HOOH appears to be too small to
account for this source, given that the amount of HOOH in the
boundary layer is much smaller than in the near-surface
snowpack and that net deposition fluxes are also relatively small,
conditions described above.
peroxide and nitrate absorb tropospheric UV-A and UV-B
radiation, the molar absorptivities for NO3- are 7.1 - 2.5 times
higher than those of HOOH between 300 and 340 nm, although
the HOOH absorption tail extends to longer wavelengths. The
situation is reversed, however, with the quantum yields (middle
panel of Figure 5): Φ(HOOH f •OH) is 212 times greater than
-
•
Φ(NO3 f OH) at 263 K, for example. The resulting action
spectrum for the rate constants of •OH formation (bottom panel
of Figure 5) shows that, for a given concentration, HOOH is a
•
much more effective source of OH than is nitrate, by a factor
of ∼9 for the temperature and actinic flux conditions of Figure
5.
•
While Figure 5 indicates that the rate constant for OH
formation from HOOH is about an order of magnitude greater
•
-
than that for OH from NO3 photolysis, comparing the rates
•
of OH formation from these two chromophores requires that
their snowpack concentrations also be taken into account. As

6270 J. Phys. Chem. A, Vol. 109, No. 28, 2005
Chu and Anastasio
•
TABLE 2: Calculated Rates of Formation of OH from the Photolysis of Snowpack Hydrogen Peroxide and Nitrate at Sites in
the Arctic and Antarctic.
typical surface snowpack
concentration of
rate constant for •OH
rate of •OH formation,
fraction
of OH from
HOOH
formation, j(i f •OH)SUN
R(i f •OH)SUN
chromophore i (µM)c
(10-6 s-1
)
(10-12 M s-1
)
d
locationa
dateb
HOOH
6
NO3
4.2
HOOH
NO3
HOOH
3.5
NO3
-
-
-
Alert,
Mar. 21
0.57
0.059
0.25
0.93
0.99
Nunavut
Summit,
Greenland
South Pole
Neumayer,
Antarctic
Jun. 21
18
4
4.0
0.26
72
1.0
Dec. 21
Dec. 21
10
4.8
1.6
1.4
1.9
3.7
0.088
0.42
19
18
0.14
0.59
0.99
0.97
a Latitudes and longitudes of sampling sites: Alert (82.5 °N, 62.3 °W), Summit (72.6 °N, 38.5 °W), South Pole (90 °S), and Neumayer (70.7
°S, 8.3 °W). b Calculations are for midday (solar noon) on the specified date. Temperatures used for calculations at Alert, Summit, South Pole, and
Neumayer were 268, 263, 253, and 268 K, respectively. c Values are from Anastasio and Jordan8 (estimated HOOH) and Toom-Sauntry and Barrie46
for Alert, Hutterli et al.7 and Dibb et al.47 for Summit, and Wolff et al.48 for South Pole. At Neumayer, we used an average of the typical 0-5 mm
value of Jacob and Klockow49 (6.0 µM) and that of Riedel50 (3.5 µM) for HOOH and the average values of Mulvaney et al.51 (1.9 µM) and Wolff
et al.48 (0.8 µM) for NO3
.
d The actinic fluxes used in our calculations are from the NCAR TUV model52 using a wavelength-independent albedo
-
of 0.93 for the snow surface53 and ozone columns of 306-309 Dobson units.54 Altitudes used for Alert, Summit, South Pole, and Neumayer were
63, 3200, 2800, and 45 m, respectively.
at least at Summit.7,29 Instead, we hypothesize that the “missing”
source of snowpack HOOH is production within the firn as a
result of vigorous HOx chemistry resulting from abundant levels
of precursors such as HCHO and other aldehydes.29-31 While
firn air reactions such as the disproportionation of HO2• to form
The photolysis of ice-bound HOOH is rapid enough that it
might significantly affect gas-phase composition and chemistry.
For example, photoformed OH on the ice could react with
•
organic matter on the particles to possibly form oxygenated
volatile organic compounds (OVOCs) which then evaporate to
•
•
HOOH are likely to be significant, the HO2 -mediated formation
the gas phase. This OH-mediated release could potentially be
of HOOH in the quasi-liquid-layers of snow grains is probably
also very important, on the basis of the analogous production
of HOOH in liquid water cloud drops.32
a significant source of gaseous oxygenated organics such as
formaldehyde, acetaldehyde, acetone, and methanol, which have
been found in the upper troposphere at surprisingly high
•
4.2. HOOH Photolysis on Cirrus Ice Clouds. Given its high
solubility in ice, especially at lower temperatures,24 we expect
that significant amounts of hydrogen peroxide are present on
ice particles in the atmosphere and that this species might play
an important role in the chemistry of ice clouds. To explore
this issue, we first examine the relative importance of hydrogen
concentrations.41-43 Analogous OH-mediated reactions have
been proposed for sunlit polar snowpacks, which emit a wide
range of OVOCs including aldehydes, ketones, and carboxylic
acids.29-31,44,45 We can estimate an upper-bound rate of this
process in cirrus clouds by assuming that every •OH formed on
the ice particles reacts with an organic molecule to form an
OVOC that then volatilizes to the gas phase. Under this
assumption, the rate of release of OVOCs from cirrus ice
particles is equal to rate of •OH formation on the ice particles,
which is 1 × 104 molecules cm-3-air s-1 if we include both
the photolysis of ice-bound HOOH (70% of total rate) and the
•
peroxide and nitrate as sources of OH on upper tropospheric
cirrus ice clouds under typical conditions (10 km altitude, 238
K, ice water content (L) ) 0.1 g m-3, 35 °N, midday, summer
solstice). On the basis of an air-ice partitioning coefficient for
HOOH at 238 K of 2.4 × 105 mol L-1-ice atm-1 (refs 6, 24,
33), and using an equation analogous to the Henry’s law gas-
aqueous partitioning,34 we calculate that ∼1% of HOOH is
present on the cirrus ice particles, while Thibert et al. calculate
that 12% of HNO3 under these conditions will be on the cirrus
ice particles.35 On the basis of typical gas-phase mixing ratios
of HOOH and HNO3 of 500 pptv and 150 pptv, respectively,36,37
and using the molar absorptivities and quantum yields illustrated
in Figure 5, we estimate that the •OH formation rate from HOOH
photolysis on ice particles is 7 × 1010 molecules cm-3-ice s-1
(equivalent to 7 × 103 molecules cm-3-air s-1), which is ∼30
times higher than that from nitrate photolysis. Furthermore, the
•
gas-to-particle partitioning of OH (30% of rate).
To estimate whether this •OH-mediated release from the ice
particles is significant, we consider four representative oxygen-
ated organics: formaldehyde, acetaldehyde, acetone, and metha-
•
nol. As shown in Table 3, if we consider reaction with OH
and direct photolysis as sinks, estimated lifetimes for these
compounds in the gas phase range from 6 × 103 s for
formaldehyde to 2 × 106 s for methanol. On the basis of typical
gas-phase mixing ratios of these species, the corresponding rates
of destruction range from 3 × 103 molecules cm-3-air s-1 for
methanol to 5 × 105 molecules cm-3-air s-1 for formaldehyde
•
•
rate of OH formation from HOOH photolysis is also greater
(Table 3). To determine whether OH-mediated release from
•
than the flux of OH moving from the gas phase to the ice
ice particles could be a significant source of any of these
OVOCs, we assume that the gaseous oxygenated organics are
in approximately steady state and, therefore, that their rates of
formation are roughly equal to the calculated rates of destruction
shown in Table 3. Comparing these values with our upper-bound
estimate for •OH-mediated OVOC release from the ice particles
(1 × 104 molecules cm-3-air s-1) indicates that this mechanism
is an insignificant source of short-lived species such as
formaldehyde (Rd ) 5 × 105 molecules cm-3-air s-1) but that
it could be significant for species with intermediate or relatively
long lifetimes (e.g., methanol with Rd ) 3 × 103 molecules
cm-3-air s-1). While there are a number of important uncer-
surface, which we calculate to be 3 × 103 molecules cm-3-air
s-1 based on continuum regime mass transport34 of gas-phase
•OH (1 × 106 molecules cm-3; ref 38) to an ice particle with a
diameter of 25 µm (ref 39). Thus, as in polar snowpacks, it
appears that HOOH photolysis dominates over nitrate photolysis
•
as an OH source on cirrus ice particles. It should be noted,
however, that this formation of •OH on the ice particles
•
represents <1% of the rate of OH formation from HOOH
photolysis in the gas phase because of the air-ice partitioning
of HOOH and the fact that photolysis of gaseous HOOH is
somewhat more efficient.40

Hydroxyl Radical Formation from HOOH
J. Phys. Chem. A, Vol. 109, No. 28, 2005 6271
(16) Baxendale, J. H.; Wilson, J. A. Trans. Faraday Soc. 1956, 53, 344.
(17) Hunt, J. P.; Taube, H. J. Am. Chem. Soc. 1952, 74, 5999.
(18) Volman, D. H.; Chen, J. C. J. Am. Chem. Soc. 1959, 81, 4141.
(19) Jacobi, H.-W.; Kwakye-Awuah, B.; Schrems, O. Ann. Glaciol. in
TABLE 3: Estimated Rates of Loss of Gaseous Oxygenated
Organics in the Upper Troposphere (10 km, 238 K, 0.35
atm, 35 °N, midday, summer solstice)
pseudo-first-order
press.
rate constant
for loss
photolysis
(20) Anastasio, C.; McGregor, K. G. Atmos. EnViron. 2001, 35, 1079.
(21) Zhou, X. L.; Mopper, K. Mar. Chem. 1990, 30, 71.
(22) Kok, G. L.; McLaren, S. E.; Staffelbach, T. A. J. Atmos. Ocean.
Technol. 1995, 12, 282.
(23) Anastasio, C.; Faust, B. C.; Allen, J. M. J. Geophys. Res.-Atmos.
1994, 99, 8231.
(24) Conklin, M. H.; Sigg, A.; Neftel, A.; Bales, R. C. J. Geophys. Res.-
Atmos. 1993, 98, 18367.
(25) Morgan, M. S.; van Trieste, P. F.; Garlick, S. M.; Mahon, M. J.;
Smith, A. L. Anal. Chim. Acta 1988, 215, 325.
(26) Sigg, A.; Neftel, A. J. Atmos. Chem. 1992, 14, 223.
(27) Jacobi, H. W.; Bales, R. C.; Honrath, R. E.; Peterson, M. C.; Dibb,
J. E.; Swanson, A. L.; Albert, M. R. Atmos. EnViron. 2004, 38, 1687.
(28) Pursell, C. J.; Conyers, J.; Denison, C. J. Phys. Chem. 1996, 100,
15450.
(29) Jacobi, H.-W.; Frey, M. M.; Hutterli, M. A.; Bales, R. C.; Schrems,
O.; Cullen, N. J.; Steffen, K.; Koehler, C. Atmos. EnViron. 2002, 36,
2619.
(30) Sumner, A. L.; Shepson, P. B.; Grannas, A. M.; Bottenheim, J.
W.; Anlauf, K. G.; Worthy, D.; Schroeder, W. H.; Steffen, A.; Domine´, F.;
Perrier, S.; Houdier, S. Atmos. EnViron. 2002, 36, 2553.
(31) Boudries, H.; Bottenheim, J. W.; Guimbaud, C.; Grannas, A. M.;
Shepson, P. B.; Houdier, S.; Perrier, S.; Domine´, F. Atmos. EnViron. 2002,
36, 2573.
(32) Lelieveld, J.; Crutzen, P. J. J. Atmos. Chem. 1991, 12, 229.
(33) McConnell, J. R.; Winterle, J. R.; Bales, R. C.; Thompson, A. M.;
Stewart, R. W. Geophys. Res. Lett. 1997, 24, 441.
rate
rate of loss,
due to •OH,
kOH+C[•OH ]a
constant,b lifetime,c
Rd
compound
(C)
jC
τC
(s)
(mlc cm-3
-
(s-1
)
(s-1
)
air s-1
)
d
HCHO
9 × 10-6
2 × 10-5
1 × 10-3
5 × 10-3
2 × 10-4 6 × 103
2 × 10-5 2 × 104
3 × 10-6 3 × 105
5 × 105
3 × 104
1 × 104
3 × 103
CH3CHO
CH3COCH3
CH3OH
0
2 × 106
a Calculated using values of the second-order rate constant between
•
compound C and OH (kOH+C) adjusted to 238 K (ref 40) and an
estimated gas-phase •OH concentration of 1 × 106 molecules cm3-air
(ref 38). b Photolysis rate constants of the gas-phase oxygenated
organics are from the TUV model.52 c Lifetimes of the compounds
are calculated as 1/(kOH+C[•OH ] + jC). d The rate of loss is calculated
as [C]/τC, where [C] is the gas-phase concentration of C. Estimated
values of [C] were 3 × 109, 8 × 108, 5 × 109, and 6 × 109 mlc cm-3
-
air for HCHO, CH3CHO, CH3COCH3, and CH3OH, respectively.42
tainties in our calculations (e.g., the yield of a particular OVOC
•
from every OH formed on the ice surface), they do point out
that HOOH photolysis on ice particles could affect the gas-
phase composition of the upper troposphere. Similarly, the
photoformation of condensed-phase •OH likely has even greater
effects on the composition (and perhaps properties) of the ice
particles themselves.
(34) Seinfeld, J. H.; Pandis, S. N. Atmospheric Chemistry and Physics:
From Air Pollution to Climate Change; Wiley: New York, 1998.
(35) Thibert, E.; Domine´, F. J. Phys. Chem. B 1998, 102, 4432.
(36) Lee, M.; Heikes, B. G.; O’Sullivan, D. W. Atmos. EnViron. 2000,
34, 3475.
(37) Chatfield, R. B. Geophys. Res. Lett. 1994, 21, 2705.
(38) Tan, D.; Faloona, I.; Simpas, J. B.; Brune, W.; Olson, J.; Crawford,
J.; Avery, M.; Sachse, G.; Vay, S.; Sandholm, S.; Guan, H.-W.; Vaughn,
T.; Mastromarino, J.; Heikes, B.; Snow, J.; Podolske, J.; Singh, H. J.
Geophys. Res.- Atmos. 2001, 106, 32667.
(39) Baumgardner, D.; Chepfer, H.; Raga, G. B.; Kok, G. L. Geophys.
Res. Lett. 2005, 32, L01806; doi: 10.1029/2004GL021300.
(40) Sander, S. P.; Friedl, R. R.; Ravishankara, A. R.; Golden, D. M.;
Kolb, C. E.; Kurylo, M. J.; Orkin, V. L.; Huie, R. E.; Molina, M. J.;
Moortgat, G. K.; Finlayson-Pitts, B. J. Chemical Kinetics and Photochemical
Data for Use in Atmospheric Studies, EValuation Number 14; JPL
Publication 02-25: Jet Propulsion Laboratory: Pasadena, CA, 2003.
(41) Crawford, J.; Davis, D.; Olson, J.; Chen, G.; Liu, S.; Gregory, G.;
Barrick, J.; Sachse, G.; Sandholm, S.; Heikes, B.; Singh, H.; Blake, D. J.
Geophys. Res.-Atmos. 1999, 104, 16255.
(42) Singh, H. B.; Tabazadeh, A.; Evans, M. J.; Field, D. B.; Jacob, D.
J.; Sachse, G.; Crawford, J. H.; Shetter, R.; Brune, W. H. Geophys. Res.
Lett. 2003, 30, 1862; doi: 10.1029/2003GL017933.
(43) Singh, H.; Chen, Y.; Staudt, A.; Jacob, D. J.; Blake, D.; Heikes,
B.; Snow, J. Nature (London) 2001, 410, 1078.
(44) Dibb, J. E.; Arsenault, M. A. Atmos. EnViron. 2002, 36, 2513.
(45) Guimbaud, C.; Grannas, A. M.; Shepson, P. B.; Fuentes, J. D.;
Boudries, H.; Bottenheim, J. W.; Domine´, F.; Houdier, S.; Perrier, S.;
Biesenthal, T. B.; Splawn, B. G. Atmos. EnViron. 2002, 36, 2743.
(46) Toom-Sauntry, D.; Barrie, L. A. Atmos. EnViron. 2002, 36,
2683.
Acknowledgment. This work was funded by the Office of
Polar Programs at the National Science Foundation (OPP-
0230288). We thank Ingrid George and Tony Robles for
laboratory assistance and the anonymous reviewers.
Supporting Information Available: Base-10 molar absorp-
tivities for aqueous hydrogen peroxide (ꢀHOOH,λ) as a function
of temperature between 274 K and 298 K at 240 nm to 380 nm
are shown in Table S1. Figure S1 shows the temperature
dependence of HOOH molar absorptivities as a function of
wavelength between 240 nm to 380 nm. This material is
available free of charge via the Internet at http://pubs.acs.org.
References and Notes
(1) Finlayson-Pitts, B. J.; Pitts, J. N. Chemistry of the Upper and Lower
Atmosphere: Theory, Experiments, and Applications; Academic Press: San
Diego, 2000.
(2) Thompson, A. J. Geophys. Res. 1991, 96, 13089.
(3) Kamiyama, K.; Motoyama, H.; Fujii, Y.; Watanabe, O. Atmos.
EnViron. 1996, 30, 967.
(4) Neftel, A.; Jacob, P.; Klockow, D. Nature (London) 1984, 311,
43.
(5) Sigg, A.; Neftel, A. Nature (London) 1991, 351, 557.
(6) McConnell, J. R.; Bales, R. C.; Stewart, R. W.; Thompson, A. M.;
Albert, M. R.; Ramos, R. J. Geophys. Res. 1998, 103, 10561.
(7) Hutterli, M. A.; McConnell, J. R.; Stewart, R. W.; Jacobi, H. W.;
Bales, R. C. J. Geophys. Res.-Atmos. 2001, 106, 15395.
(8) Anastasio, C.; Jordan, A. Atmos. EnViron. 2004, 38, 1153.
(9) Dubowski, Y.; Colussi, A. J.; Hoffmann, M. R. J. Phys. Chem. A
2001, 105, 4928.
(10) Chu, L.; Anastasio, C. J. Phys. Chem. A 2003, 107, 9594.
(11) Honrath, R. E.; Peterson, M. C.; Guo, S.; Dibb, J. E.; Shepson, P.
B.; Campbell, B. Geophys. Res. Lett. 1999, 26, 695.
(12) Honrath, R. E.; Peterson, M. C.; Dziobak, M. P.; Dibb, J. E.;
Arsenault, M. A.; Green, S. A. Geophys. Res. Lett. 2000, 27, 2237.
(13) Domine´, F.; Shepson, P. B. Science 2002, 297, 1506.
(14) Peterson, M. C.; Honrath, R. E. Geophys. Res. Lett. 2001, 28, 511.
(15) Zellner, R.; Exner, M.; Herrmann, H. J. Atmos. Chem. 1990, 10,
411.
(47) Dibb, J. E.; Talbot, R. W.; Munger, J. W.; Jacob, D. J.; Fan, S. M.
J. Geophys. Res.-Atmos. 1998, 103, 3475.
(48) Wolff, E. W.; Jones, A. E.; Martin, T. J.; Grenfell, T. C. Geophys.
Res. Lett. 2002, 29, 1944; doi: 10.1029/2002GL015823.
(49) Jacob, P.; Klockow, D.; Fresenius, J. Anal. Chem. 1993, 346, 429.
(50) Riedel, K. Reports on polar and marine research 394; Riemann,
F., Wegener, A., Eds.; Institute for Polar and Marine Research, Bremer-
haven, Germany, 2001.
(51) Mulvaney, R.; Wagenbach, D.; Wolff, E. W. J. Geophys. Res. 1998,
103, 11021.
(52) Madronich, S.; Flocke, S. J. The role of solar radiation in
atmospheric chemistry. In Handbook of EnVironmental Chemistry; Boule,
P., Ed.; Springer-Verlag: Heidelberg, 1998; pp. 1-26. www.acd.ucar.edu/
TUV/.
(53) Dickerson, R. R.; Stedman, D. H.; Delany, A. C. J. Geophys. Res.
1982, 87, 4933.
(54) The_NASA/GSFC_TOMS_Team. Total Ozone Mapping Spec-
trometer. data http://toms.gsfc.nasa.gov.