Quantum Yields of Hydroxyl Radical and Nitrogen Dioxide from the Photolysis of Nitrate on Ice

Article abstract of DOI:10.1021/jp0349132

Nitrate photolysis proceeds via two major channels at illumination wavelengths above 290 nm: NO₃^- + hv (+H⁺) → NO₂ + ^.OH (1) and NO₃^- + hv → NO₂^- + O(³P) (2). A recent study determined the quantum yield of reaction 1 on ice by measuring NO₂ production, but suggested their values might be lower bounds because of incomplete recoveries of NO₂. We measured the quantum yield of pathway 1 using an alternate approach, i.e., by following the formation of ^.OH. Our quantum yields for ^.OH (φ_OH) at 263 K were independent of nitrate concentration and illumination wavelength (λ > 300 nm), but were dependent upon pH. Values of φ _OH decreased from (3.6 ± 0.6) × 10^-3 at pH 7.0 to (2.1 ± 0.8) × 10^-3 at pH 2.0, where the listed pH values are those of the sample solution prior to freezing. Temperature dependence experiments (239-318 K; pH 5.0) showed that values of φ _OH in ice pellets and aqueous solutions were both well described by the same regression line, ln(_OH) = ln(φ₁) = -(2400 ± 480)(1/T) + (3.6 ± 0.8) (where errors represent ±1σ), suggesting that the photolysis of nitrate on ice occurs in a quasi-liquid layer rather than in the bulk ice. Our ice quantum yields between 268 and 240 K are 3-9 times higher, respectively, than φ ₁ values determined previously in ice. Applying our quantum yields to past field experiments indicates that nitrate photolysis can account for the flux of NO_x from sunlit snow in the Antarctic and at Summit, Greenland, but that nitrate was only a minor source of the snowpack NO _x measured during the Alert 2000 campaign in the Canadian Arctic. Additional calculations show that the photolysis of nitrate on cirrus clouds in the upper troposphere is a minor source of NO_x that cannot account for the apparent underestimation of the ratio of NO_x/HNO₃ in current numerical models.

Full text of DOI:10.1021/jp0349132

9594
J. Phys. Chem. A 2003, 107, 9594-9602
Quantum Yields of Hydroxyl Radical and Nitrogen Dioxide from the Photolysis of Nitrate
on Ice
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: April 7, 2003; In Final Form: August 21, 2003
-
Nitrate photolysis proceeds via two major channels at illumination wavelengths above 290 nm: NO3
+
hν (+H+) f NO2 + •OH (1) and NO3- + hν f NO2- + O(3P) (2). A recent study determined the quantum
yield of reaction 1 on ice by measuring NO2 production, but suggested their values might be lower bounds
because of incomplete recoveries of NO2. We measured the quantum yield of pathway 1 using an alternate
approach, i.e., by following the formation of •OH. Our quantum yields for •OH (ΦOH) at 263 K were independent
of nitrate concentration and illumination wavelength (λ > 300 nm), but were dependent upon pH. Values of
Φ
OH decreased from (3.6 ( 0.6) × 10-3 at pH 7.0 to (2.1 ( 0.8) × 10-3 at pH 2.0, where the listed pH values
are those of the sample solution prior to freezing. Temperature dependence experiments (239-318 K; pH
5.0) showed that values of ΦOH in ice pellets and aqueous solutions were both well described by the same
regression line, ln(ΦOH) ) ln(Φ1) ) -(2400 ( 480)(1/T) + (3.6 ( 0.8) (where errors represent (1σ),
suggesting that the photolysis of nitrate on ice occurs in a “quasi-liquid layer” rather than in the bulk ice. Our
ice quantum yields between 268 and 240 K are 3-9 times higher, respectively, than Φ1 values determined
previously in ice. Applying our quantum yields to past field experiments indicates that nitrate photolysis can
account for the flux of NOx from sunlit snow in the Antarctic and at Summit, Greenland, but that nitrate was
only a minor source of the snowpack NOx measured during the Alert 2000 campaign in the Canadian Arctic.
Additional calculations show that the photolysis of nitrate on cirrus clouds in the upper troposphere is a
minor source of NOx that cannot account for the apparent underestimation of the ratio of NOx/HNO3 in current
numerical models.
NO3- + hν (+H+) f NO2 + •OH
(1)
1. Introduction
Recent measurements indicate that the photochemically
mediated production and release of NOx is a common occurrence
in surface snowpacks. Measurements at Summit (Greenland),1-3
Alert (Nunavut, Canada),4,5 and Neumayer (Antarctic)6,7 all
reveal that sunlit polar snowpacks produce NOx. Significant
photochemical production of NOx has also been found in snow
at midlatitudes (Michigan).8 Since nitrate is a common trace
constituent of snow and ice, and NO3- photolysis in solution is
known to yield NO and NO2, a number of previous reports have
shown or suggested that nitrate photolysis is at least partially
responsible for the measured releases of NOx.1-10 This photo-
chemistry could be important for a number of reasons:1,10-12
as a source of NOx to the snowpack interstitial air and overlying
while the minor channel produces nitrite and oxygen atoms
NO3- + hν f NO2- + O(3P)
(2)
In aqueous solution at 298 K the quantum yields for reactions
1 and 2 are approximately 0.017 and 0.0011,14,15 respectively,
although other values have been reported for each reaction.16
In contrast to the solution case, the photochemical behavior
of nitrate in ice and snow is relatively poorly understood.
Dubowski et al. determined the quantum yield of reaction 1 on
ice by measuring the release of NO2,18 but suggested that their
values might be lower bounds because of incomplete recoveries
of this product. This same group has also measured Φ2, with
the more recent value of (1.5 ( 0.3) × 10-3 at 263 K (ref 19)
superseding an earlier value of (4.8 ( 1.5) × 10-3 (ref 18).
The goal of our study was to measure the quantum yield for
•
boundary layer, as a source of highly reactive OH to snow
grains, and as a sink for snowpack nitrate. These reactions (and
their subsequent effects) will not only alter the chemistry
occurring in the snowpack and boundary layer, but could also
affect the ice core records of trace species (e.g., NO3-, CH4,
and HOOH) that are used to infer the composition of past
atmospheres.
•
reaction 1 in ice by following the formation of OH to avoid
complications with quantitatively collecting photoproduced NO2.
We have determined quantum yields for •OH (ΦOH ) Φ1) as a
function of nitrate concentration and pH in ice pellets. In
addition, we have measured the temperature dependence of ΦOH
for both ice pellets and aqueous solutions.
Nitrate photolysis in aqueous solution has been extensively
studied and proceeds via two main pathways.13-17 At wave-
lengths above 290 nm, the major photolysis channel produces
•
NO2 and OH
2. Experimental Methods
2.1. Materials. Sodium nitrate (certified ACS), hydrogen
peroxide, acetonitrile (Optima), sodium borate (certified ACS),
* To whom correspondence should be addressed. Phone: (530) 754-
6095. Fax: (530) 752-1552. E-mail: canastasio@ucdavis.edu.
10.1021/jp0349132 CCC: $25.00 © 2003 American Chemical Society
Published on Web 10/21/2003

Photolysis of Nitrate on Ice
J. Phys. Chem. A, Vol. 107, No. 45, 2003 9595
-
sulfuric acid (Optima), and perchloric acid (Optima) were
obtained from Fisher. Benzoic acid (99%), sodium benzoate
(99%), and m-hydroxybenzoic acid (99%) were obtained from
Aldrich, while p-hydroxybenzoic acid (98%) was obtained from
TCI America. All chemicals were used as received. Purified
water (“Milli-Q water”) was obtained from a Milli-Q Plus
system (g18.2 MΩ cm).
above: for sample solutions with 40-300 µM NO3 the BA
concentration was 450 µM, while solutions containing 300-
-
3000 µM NO3 had 4500 µM benzoic acid.
Ice samples (section 2.2) were illuminated for known time
intervals, using 313-nm (or, in a few cases, 334 nm) light from
a 1000-W Hg/Xe monochromatic system.22 Samples were held
in a custom-designed (Paige Instruments), Peltier-cooled, tem-
perature-controlled, Teflon-coated copper chamber (38 mm ×
26 mm × 70 mm outer dimensions) enclosed in an aluminum
housing purged with dry air. In the center of the copper chamber
was a 13 mm × 13 mm × 62 mm rectangular opening (with a
removable aluminum top) to hold the ice sample, which was
placed at the side of the opening nearest the illumination system.
A quartz window at this “upstream” side of the opening
transmitted light from the illumination system. The light beam
(9 mm × 12 mm) from the illumination system passed through
the quartz window, into the ice sample, through the quartz slide,
and into the back of the sample chamber. The temperature of
the sample chamber was monitored with a thermocouple placed
into the base of the copper block, just outside the inner sample
cavity. Because of the relatively small actinic fluxes in our
system (see section 2.4), the low concentrations of nitrate used,
and the short pathlengths employed (∼1 mm for the sample
solution ice and ∼2 mm for the water ice base), we calculate
that illumination would cause negligible heating of the ice
sample (<0.3 K over the course of an experiment). Thus the
temperature measured in the copper block near the sample cavity
should be essentially that of the ice pellet.
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 5-10 min and then the melted
mixture was analyzed for p-HBA, using HPLC with UV
detection.20 HPLC conditions were the following: column )
Keystone BetaBasic-18 with associated guard column; eluent
) 30% CH3CN and 70% H2O, mixture adjusted to pH 2 with
HClO4; flow rate ) 0.60 mL min-1; detection wavelength )
256 nm. Concentrations of p-HBA were determined based on
calibration standards made in Milli-Q water and run during the
day of the illumination experiment.
2.2. Ice Sample Preparation. Ice samples were prepared in
a frame consisting of a PVC or Teflon template placed on a
quartz backing. The template was a 45 mm × 12 mm × 5 mm
piece of material containing a 12 mm × 9 mm × 5 mm hole
corresponding to the size and position of the light beam in the
illumination system. Samples were prepared by placing the
template onto a 45 mm × 12 mm × 1 mm quartz slide and
then pipetting 150 µL of Milli-Q water into the well created by
the hole in the template. This pure water ice “base” was then
frozen by cooling from room temperature to the temperature of
the experiment at a rate of -1 K min-1 in the ice preparation
chamber described below. After the base was frozen, the
chamber was kept at the experiment temperature and 50 µL of
sample solution was pipetted onto the base and allowed to sit
until frozen. This volume of sample solution was small enough
that the solution contacted only the pure ice base and did not
touch the PVC or Teflon walls of the template. Sample solutions
contained known concentrations of NaNO3 and benzoic acid
•
(to trap photoformed OH; see section 2.3) in Milli-Q water
and were adjusted to the desired pH with sulfuric acid (pH 2 to
5) or borate (pH 6 and 7). Note that the terms “benzoic acid”
and “BA” here are both used to refer to the sum of benzoic
acid and benzoate; stock solutions of 14 mM BA contained 1.3
mM benzoic acid and 12.7 mM sodium benzoate. In addition,
all listed concentrations (e.g., BA and NO3-) and pH values
are for the sample solutions prior to freezing unless specified
otherwise.
The ice preparation chamber (100 mm × 100 mm × 50 mm)
was a custom-designed unit (Paige Instruments) consisting of
a copper block base, Teflon walls, and a removable aluminum
top with a glass window. The chamber was cooled by two Peltier
units that were attached to the bottom of the copper block base
and were run by a custom-designed, programmable, computer-
controlled temperature controller. Temperatures in the chamber
(239-268 K) were monitored by a thermocouple embedded in
the copper block, near its top.
Initial rates of p-HBA formation during illumination (R*p-HBA,λ
)
were determined from plots of [p-HBA] versus illumination time
by using a linear regression fit or, in a very few cases, an
exponential fit, depending on the shape of the data.20 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
nitrate was added (RBlank,λ). The corrected formation rate of
p-HBA in the illuminated sample was then calculated as
2.3. •OH Measurements. Hydroxyl radicals were character-
ized by using a chemical probe technique where photoformed
•OH reacts with benzoic acid (BA) to form stable, measurable
products including p-hydroxybenzoic acid (p-HBA) and m-
hydroxybenzoic acid (m-HBA):20,21
BA + •OH f p-HBA + m-HBA + other
(3)
We used high concentrations of benzoic acid to trap all of the
hydroxyl radicals produced during illumination. To determine
the concentration of BA required to scavenge all OH, we
R
p-HBA,λ ) R*p-HBA,λ - RDark - RBlank,λ
(4)
•
As shown in Figure 1, there was generally no formation of
p-HBA in samples kept in the dark and rates of p-HBA
formation in the illuminated blanks were typically very small
compared to the rates in the corresponding illuminated samples.
Corrected rates of p-HBA formation were converted to rates of
•OH formation (ROH,λ) by using the yield of p-HBA formed from
performed two experiments where we measured the rate of
p-HBA formation as a function of added benzoic acid. The first
experiment used a pH 5.0 sample solution (100 µM NO3- and
40-1000 µM BA) frozen and illuminated at 263 K; the second
test used similar conditions (same pH and temperature) except
that [NO3-] ) 3000 µM and [BA] ) 250-6000 µM. In both
cases the rate of p-HBA formation increased with benzoic acid
concentration until [BA]:[NO3-] g 1, above which the rate of
p-HBA formation was constant. This indicates that all of the
•OH radicals were scavenged by BA at [BA]:[NO3-] g 1. All
experiments in this study used [BA]:[NO3-] ratios of 1.5 or
the reaction of OH with BA (Yp-HBA):20
•
R
OH,λ ) Rp-HBA,λ/Yp-HBA
(5)
Values of Yp-HBA were measured in illuminated (313 nm) ice
samples (section 2.2) made from sample solutions containing

9596 J. Phys. Chem. A, Vol. 107, No. 45, 2003
Chu and Anastasio
Figure 1. Photoformation of p-hydroxybenzoic acid (p-HBA) during
•
Figure 2. Yield of p-HBA formed from the reaction of OH with
313-nm illumination of a frozen (263 K) sample solution composed of
benzoic acid (Yp-HBA) as a function of the sample solution ionic strength
(pH 5.0) in solutions containing HOOH, BA, and variable amounts of
Na2SO4. Values of Yp-HBA were determined by two methods: Rp-HBA
calcd ROH (eq 6; solid circles) or Rp-HBA/RBA (eq 7; open triangle). The
dashed line is an exponential fit to all of the experimental data points.
One data point (I ) 7.1 mM, Yp-HBA ) 0.088; Table S1) is not shown
in the figure, but is included in the exponential fit.
-
200 µM NO3 and 450 µM BA, adjusted to pH 5.0. The solid circles
represent data for illuminated frozen samples, while the squares
represent p-HBA concentrations in an identical frozen sample kept in
the dark. The inverted triangles correspond to the formation of p-HBA
in a blank control (i.e., an illuminated frozen sample solution that
contained no nitrate, 450 µM BA, and was adjusted to pH 5.0).
/
•
HOOH (as a photochemical source of OH) and benzoic acid
from 0.039 at I ) 428 µM to 0.084 at I > 600 µM. This
and adjusted to a given pH with H2SO4. We used two methods
to determine Yp-HBA. In the first, we prepared sample solutions
with 450 µM BA and 100 µM HOOH and measured the
formation of p-HBA during illumination. Yields of p-HBA from
relationship is well described by the equation Yp-HBA
)
-1172.5550 + 1172.6388(1 - exp(-0.0237 × I)) (R2 ) 0.962),
where the ionic strength is the sample solution value in µM.
To examine the dependence of Yp-HBA on pH we used ice
(263 K) prepared from sample solutions of pH 2.0-7.0 and
with 67 µM added Na2SO4 to replace the ionic strength
contribution from the 200 µM NaNO3 used for our pH
experiments. These measurements revealed that Yp-HBA has a
relatively weak pH dependence, increasing from 0.058 ( 0.016
at pH 2.0 to 0.084 ( 0.014 at pH 7.0, and follows the formula
Yp-HBA ) -0.0011(pH)2 + 0.0154(pH) + 0.0318 (R2 ) 0.978;
data in Table S1 of Supporting Information). In contrast to the
pH 5.0 case, at 263 K in ice prepared from sample solutions of
pH 2.0 there was no ionic strength dependence; the average
value of Yp-HBA in these experiments (0.063 ( 0.014) was
consistent with that from the pH-dependence experiment.
Finally, we examined the temperature dependence of Yp-HBA
(pH 5.0 sample solution with 67 µM added Na2SO4) and found
that there was no significant difference over the range of 243-
268 K, with an average value of 0.081 ( 0.014, consistent with
the result from Figure 2. Values of Yp-HBA are tabulated in the
Supporting Information (Table S1).
•
the OH + BA reaction were determined by using the ratio of
the measured (corrected) rate of p-HBA formation to the
calculated rate of •OH formation from HOOH photolysis (“calcd
R
OH, λ”):
Rp-HBA, λ
Yp-HBA
)
)
calcd ROH, λ
Rp-HBA, λ
2.303IλΦHOOHfOH, λꢀHOOH, λ[HOOH]l
(6)
where Iλ is the actinic flux in the sample chamber (section 2.4),
HOOHfOH,λ is the quantum yield of •OH from HOOH photolysis
Φ
(extrapolated from the temperature dependence reported in ref
15), ꢀHOOH,λ is the molar absorptivity of HOOH (assumed to be
the same as the room-temperature value of 0.37 M-1 cm-1 at
313 nm; ref 23), [HOOH] is the molar concentration of hydrogen
peroxide, and l is the effective path length of the sample (cm).
The product Iλ × l in eq 6 was determined by using 2-nitro-
benzaldehyde actinometry performed under the same conditions
as the Yp-HBA experiment (see section 2.4). In the second method
we employed low concentrations of benzoic acid (10-14 µM)
and HOOH (4-6 µM) and measured the rate of formation of
p-HBA (subsequently corrected for RDark and RBlank,λ) and the
rate of loss of BA (RBA,λ) simultaneously in the illuminated
samples. Yields of p-HBA were then determined as
2.4. Calculation of •OH Quantum Yield. On the day of each
experiment the surface-area normalized actinic flux (Iλl) in the
sample chamber was measured by using ice pellets of the same
size and composition as the nitrate-containing samples except
for the addition of 4 µM 2-nitrobenzaldehyde (2NB) as a
chemical actinometer. For the low light-absorbing conditions
of our actinometry, the measured rate constant for 2NB loss
(j2NB,λ) is related to actinic flux through24
Yp-HBA ) Rp-HBA,λ/RBA,λ
(7)
j
2NB,λ ) 2.303Iλꢀ2NB,λ
Φ
2NB,λl
(8)
•
Note that this assumes that all BA loss was a result of OH
formed from HOOH photolysis, which should be the case in
our experiments.
To ensure that Yp-HBA values could be applied to our nitrate
photolysis experiments, we determined p-HBA yields under the
same range of conditions used in the nitrate experiments. As
shown in Figure 2, Yp-HBA varied nonlinearly with the ionic
strength (I) of the sample solution (pH 5.0, 263 K), increasing
where ꢀ2NB,λ
Φ2NB,λ is the product of the molar absorptivity and
quantum efficiency of 2NB (640 M-1 cm-1 at 313 nm and
300 M-1 cm-1 at 334 nm, both at 293 K; ref 24) 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 (i.e., NaNO3, BA, and H2SO4)
showed that the presence of these chemicals had no effect on

Photolysis of Nitrate on Ice
J. Phys. Chem. A, Vol. 107, No. 45, 2003 9597
measured values of j2NB,λ. We also found that j2NB,λ was
independent of temperature in the range 243-268 K, suggesting
that room-temperature values for the product ꢀ2NB,λΦ2NB,λ can
be used at lower temperatures. The average value of j2NB,313 in
our experiments was 0.266 s-1, corresponding to a surface-area
normalized actinic flux of 2.5 × 1017 photons cm-2 s-1 (160
mW cm-2). As mentioned above, only a small fraction (∼0.1%)
of this incident light was absorbed by the nitrate-containing ice
pellets in our experiments.
•
Under these low light-absorbing conditions the rate of OH
formation can be expressed as
-
R
OH,λ ) 2.303 IλꢀNO -,λ
Φ
OH,λl[NO3 ]
(9)
3
where ꢀNO ,λ is the molar absorptivity of nitrate (5.3 M-1 cm-1
-
3
Figure 3. Base-10 molar absorptivities of aqueous nitrate at 278 K
at 313 nm; see section 2.5), ΦOH,λ is the quantum yield of •OH
from nitrate photolysis, and [NO3-] is the molar concentration
of nitrate. Rearranging eq 8 to solve for Iλ and substituting that
into eq 9 produces an expression for the quantum yield of •OH
-
(solid line) and 293 K (dashed line). Values of ꢀNO as a function of
,λ
3
temperature and wavelength are tabulated in the Supporting Information
(Table S2).
illuminated with 313-nm light in stirred 1-cm cells; small
aliquots were removed at measured time intervals and analyzed
for p-HBA. Rates of •OH formation were determined with use
of eq 5, using a newly determined value for Yp-HBA (0.19 (
0.01) that is consistent with that reported previously.20 The
quantum yield of OH was calculated by using eq 10, where
j2NB,λ was the value measured in aqueous 2NB (4 µM) on the
same day as the nitrate photolysis experiment.
2.7. Estimation of Nitrate Concentration and pH in the
Quasi-liquid Layer. Slow freezing of an aqueous solution
segregates the components into frozen pure water ice and a
disordered, possibly heterogeneous, “quasi-liquid” layer (QLL)
containing the solutes.26-29 This interfacial QLL has not been
well characterized and its properties often seem to depend on
the analytical method employed to study it.26-29 As described
below, we estimated the nitrate concentration and pH in the
QLL for our experiments using the freezing-point depression
due to solutes and the acid-dissociation equilibria for sulfuric,
nitric, and benzoic acids. Although the quasi-liquid layer is
structurally different from liquid water at the temperatures used
here,26,29 Cho et al.28 found that, at temperatures above the NaCl
eutectic point of 252 K, the measured liquid water fraction in
a frozen NaCl solution was similar to that calculated by freezing
point depression. Based on this, we have estimated the molality
of total solute (m, mol kg-1) in the QLL in our experiments
from30
ꢀ2NB,λ
Φ
2NB,λROH,λ
ΦOH,λ
)
(10)
-
j
2NB,λꢀNO ,λ[NO3 ]
-
3
•
Note that values of ROH,λ might include small contributions from
the photolysis of nitrite and nitrous acid, which are formed
during nitrate photolysis (reaction 2). However, a kinetic model
of the nitrate, nitrite, and nitrous acid reactions in our system
indicates that <8% and <14% of our experimentally measured
•OH was from nitrite and nitrous acid in ice from sample
solutions of pH 5.0 and 2.0, respectively. Because these upper
bound contributions are smaller than our experimental uncer-
tainties (the average relative standard deviation in our ice
experiments was 20%), we did not adjust our nitrate ΦOH,λ
-
values for possible contributions from NO2 and HNO2 pho-
tolysis. Finally, because values of ΦOH,λ are independent of
illumination wavelength in the 300-nm band (see section 3.3),
we use the term “ΦOH” in this paper instead of wavelength-
specific nomenclature such as ΦOH,313
.
2.5. Molar Absorptivities of Nitrate as a Function of
Temperature. While the 200-nm absorption band of nitrate on
ice has been measured,25 there are no published ice measure-
ments for the weaker 300-nm band. Because the weak absorption
in this latter band makes it difficult to measure molar absorp-
tivities on ice, we instead used values extrapolated from the
temperature dependence of absorbance in aqueous solution.
Absorbance spectra of four aqueous nitrate solutions (27-108
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
(276-298 K) with chilled water from a recirculating water bath.
The molar absorptivity at each wavelength (250-360 nm) was
determined as the slope of the linear regression fit to the data
of absorbance versus nitrate concentration.
m ) ∆T/1.86
(11)
∆T here is the freezing point depression, i.e., 273 K - Texp
,
where Texp is the ice temperature during the experiment (K).
Because we froze our sample solutions slowly, we assume that
all solutes are excluded from the bulk ice and are present in a
(homogeneous) QLL; based on results from the Aerosol
Inorganics Model,31 there should be no formation of nitrate
hydrates or solids in our ice pellets. The resulting rough
estimates of QLL composition under typical experimental
conditions are listed in Table 1.
3-,λ
As shown in Figure 3, ꢀNO changes very little across the
temperature range examined. In particular, the molar absorptivity
of nitrate at 313 nm exhibited no significant dependence on
temperature, with an average value of 5.31 ( 0.03 M-1 cm-1
between 276 and 298 K. The best fit values of the molar
absorptivities at all temperatures and wavelengths are listed in
the Supporting Information (Table S2).
3. Results and Discussion
3.1. Quantum Yield of •OH in Ice as a Function of Nitrate
•
Concentration. Quantum yields of OH (ΦÃΗ) from nitrate
2.6. Measurements of ΦOH in Aqueous Solution. Aqueous
solutions contained 200 µM NaNO3 and 450 µM benzoic acid,
and were adjusted to pH 5.0 with sulfuric acid. Samples were
photolysis in ice were measured at 263 K in frozen sample
solutions with different nitrate concentrations at pH 2.0 and 5.0.
As shown in Figure 4, ΦÃΗ in the frozen pH 5.0 sample

9598 J. Phys. Chem. A, Vol. 107, No. 45, 2003
Chu and Anastasio
TABLE 1: Typical Sample Solution Conditions and the
Corresponding Estimated Quasi-liquid Layer (QLL)
Composition for Ice Pellets at 263 K
sample solution values
ice quasi-liquid layer compositiona
-
- b
[BA] [NO3
pH (µM)
]
concn
ionic
[NO3 ]
(µM) factorc strength (M)
(M)
pH
5.0
5.0
5.0
5.0d
5.0
2.0
2.0
450
450
450
450
4500
450
40
100
200
200
3000
150
5673
5036
4241
330
362
482
2.6
2.6
2.6
5.2
2.7
1.4
1.5
0.23
0.50
0.85
0.066
1.1
0.072
0.65
4.9
4.9
4.9
5.0
4.7
-0.36
0.17
4500
3000
216
a The total molality in the quasi-liquid layer at 263 K was calculated
to be 5.38 mol kg-1, based on the freezing point depression (eq 11)
and assuming all solutes were present in the QLL. b Total concentration
-
of NO3 and HNO3 in the QLL. c Ratio of the total concentration of
-
NO3 and HNO3 in the QLL to the total concentration in the sample
solution. d Sample solution also contained 5.0 mM added Na2SO4 to
adjust the ionic strength.
•
Figure 5. Quantum yields of OH in ice pellets (263 K, 313-nm
illumination) as a function of pH in the unfrozen sample solution.
Sample solutions contained 200 µM of either NaNO3 (solid circles) or
HNO3 (open square). The italicized pH values in parentheses are the
estimated values in the quasi-liquid layers of the ice pellets; the HNO3
samples had an estimated QLL pH of 4.5. Each data point represents
the average value of 2-24 individual measurements, with error bars
of (1σ. The solid line is an exponential regression fit to the
experimental data (R2 ) 0.961): ΦOH ) 0.000587 + 0.00374(1 -
exp(-0.257 × pH)), where the pH is the value in the sample solution.
pH 5.0) resulting from the Na2SO4-adjusted pH 5.0 sample was
much higher than that resulting from the pH 2.0 solution (QLL
pH -0.36), suggesting that differences in pH might be the cause
of the observed ΦOH differences.
•
3.2. Quantum Yield of OH in Ice as a Function of pH.
To explore the pH dependence of ΦÃΗ on ice, we measured
values in frozen sample solutions containing 200 µM NO3- and
with pH values from 2.0 to 7.0, as shown in Figure 5. The
estimated pH values in the QLL, which are also listed on the
figure, ranged from pH -0.4 (from the pH 2.0 sample solution)
to 5.2 (pH 7.0 solution). ΦÃΗ was approximately constant in
ice pellets made from sample solutions of pH 5.0 to 7.0, but
decreased in ice made from sample solutions with lower pH
values. ΦOH in the ice prepared from the pH 2.0 sample solution,
(2.1 ( 0.8) × 10-3, was approximately 40% lower than that
obtained for ice made with the pH 7.0 solution, (3.6 ( 0.6) ×
10-3. As shown in Figure 5, the •OH quantum yield in samples
prepared with HNO3 (sample solution pH 4.6) was consistent
with the pH trend observed in the other samples, which were
prepared with NaNO3.
•
Figure 4. Quantum yields of OH in ice pellets (263 K; 313-nm
illumination) as a function of nitrate concentration in the sample
solution. Sample solution composition: pH 5.0 (circles) or 2.0 (squares);
[BA] ) 450 µM for [NO3-] e 450 µM, and [BA] ) 4500 µM for
[NO3-] > 450 µM. Error bars represent (1σ, calculated based on the
uncertainties in Rp-HBA,λ, Iλ, Yp-HBA, ꢀNO ,λ, [NO3-], and the p-HBA
-
3
calibration curve. Horizontal lines indicate the average values for ice
made from sample solutions of pH 5.0 ((3.4 ( 0.6) × 10-3) and 2.0
((2.2 ( 0.5) × 10-3). The two “×” points represent pH 5.0 sample
solutions that had been adjusted with Na2SO4 to the same ionic strength
(15 mM) as the corresponding pH 2.0 sample solution.
solutions was independent of the nitrate concentration and had
an average value ((1σ) of (3.4 ( 0.6) × 10-3. Values of ΦÃΗ
in the pH 2.0 sample solutions were also independent of [NO3-],
but the average value of (2.2 ( 0.5) × 10-3 was 35% lower
than the pH 5.0 average (Figure 4). As described in more detail
below, it is unclear why ΦÃΗ is lower in the ice prepared with
the pH 2.0 sample solution. To test whether the greater ionic
strength in the pH 2.0 solution was responsible, we measured
ΦÃΗ in a pH 5.0 sample solution that was adjusted with Na2-
SO4 to the same ionic strength as the pH 2.0 solution. As shown
in Figure 4, the addition of Na2SO4 had little effect and the
We examined four possible reasons why ΦOH decreases at
•
lower pH: (1) effects of ionic strength on OH formation, (2)
sensitivity of the yield of p-HBA (Yp-HBA) to varying ionic
strength or pH, (3) evaporation of HNO3 from the ice surface,
-
and (4) differences in the behavior of HNO3 compared to NO3
.
As discussed in section 3.1, we eliminated the first possible
reason since differences in ionic strength had no significant
effect on ΦOH under our experimental conditions. We can also
eliminate the second possibility since for any given sample
experiment we measured Yp-HBA under nearly identical condi-
tions of pH and ionic strength. The third possibility (HNO3
evaporation at low pH) was ruled out by measuring the
concentrations of total nitrate (i.e., HNO3 and NO3-) in ice
pellets after being held in the sample chamber in the dark for
20 min, the typical duration of a photochemistry experiment.
average ΦÃΗ in these two experiments, (3.6 ( 0.6) × 10-3
,
was identical (within the uncertainties) to the average ΦÃΗ for
the other pH 5.0 experiments. This indicates that the ionic
strength of the sample solution has no significant effect on ΦÃΗ
within the range of ionic strengths tested. Estimated quasi-liquid
layer (QLL) concentrations (Table 1) indicate that the pH 2.0
solution and the Na2SO4-adjusted pH 5.0 solution had similar
nitrate concentrations. However, the quasi-liquid layer pH (QLL
-
There was no significant loss (average of <1%) for NO3 in
ice made from sample solutions with pH values of 2.0 or 5.0.

Photolysis of Nitrate on Ice
J. Phys. Chem. A, Vol. 107, No. 45, 2003 9599
The final possibility we considered is that the •OH quantum
yield and/or molar absorptivity of HNO3 are lower than those
•
of NO3-. This would lead to less OH production and lower
calculated values of ΦOH under more acidic conditions where
HNO3 is more prevalent. However, based on the pKa for HNO3
(-1.2 at room temperature32), only about 14% of NO3- would
be converted to HNO3 under the most acidic conditions of this
study (i.e., QLL pH of -0.4 resulting from the pH 2.0 sample
solution). Therefore, even if the •OH quantum yield from HNO3
-
photolysis was zero, protonation of NO3 is unable to explain
the ∼40% decrease in measured values of ΦOH between the
highest and lowest pH values studied (Figure 5). Furthermore,
•
in aqueous solution the estimated values of the OH quantum
yield and molar absorptivity for HNO3 are not much different
-
from the corresponding values for NO3 (ref 33). Thus
protonation of NO3- to HNO3 at low pH does not appear to be
the reason for the reduction of ΦOH with decreasing pH in the
ice pellets. A related possibility is that in the more acidic
solutions HNO3 might be drawn to the air-QLL interface and
that this reduces apparent reactivity, perhaps because of separa-
tion of HNO3 from the benzoic acid probe, although we have
no mechanism to suggest why this would occur. This surface
enhancement of HNO3 has been observed in very acidic ternary
HNO3-H2SO4-H2O solutions (e.g., 25% HNO3 and 25%
H2SO4),34 but it is unclear whether a similar phenomenon occurs
under our milder experimental conditions. Furthermore, unless
the pKa of HNO3 is higher at the air-QLL interface at these
low temperatures (compared to the value in bulk solution at
298 K), the fraction of HNO3 should be too small (∼14%) to
produce the pH effect we have measured, as discussed above.
Thus at this point we are unable to explain the pH dependence
Figure 6. Temperature dependence (T ) 239-318 K) of ΦOH in
aqueous and frozen sample solutions containing 200 µM NO3-, 450
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 the average value of
ΦOH at each temperature for both our solution and ice data (eq 12).
The triangles, inverted triangles (with dotted regression line), and
squares are solution data from Zepp et al.,13 Zellner et al.,15 and Warneck
et al.,14 respectively. The diamonds are measured data of Dubowski et
al.18 in ice pellets. Error bars for these previous sets of data are the
unspecified uncertainties reported by the authors.
TABLE 2: Values of Ea and ∆S ((1σ) for the Formation of
•OH and NO2 from Nitrate Photolysis in Ice and Aqueous
Solutiona
Ea (kJ mol-1
)
∆S (J mol-1 K-1
)
solution values
this study
19 ( 4
16 ( 4
29 ( 6
19 ( 5
•
of the OH quantum yield. However, regardless of the mecha-
Zellner et al. (1990)b
ice values
nism responsible for the pH dependence of ΦOH, values
determined from sample solutions of pH g5 should be ap-
plicable for many field studies given that melted, bulk snow
typically has a pH value of 5 or greater (e.g., ref 35). It should
be noted, however, that there are conditions under which the
pH of the “disordered” layer on ice or snow surfaces can be
significantly lower (e.g., ref 28).
this study
20 ( 4
42 ( 6
33 ( 7
Dubowski et al. (2001)c
combined values
this study
103 ( 22
20 ( 4
31 ( 7
a The activation energy (Ea) and change in entropy (∆S) were
calculated based on the linear regression fits to the ln(ΦOH) versus
1/T data (Figure 6): Ea ) -slope × R × 10-3 kJ J-1 and ∆S )
y-intercept × R, where R is the gas constant (8.314 J mol-1 K-1).
b Values calculated based on the pH 4-9 data of Zellner et al.15
c Values calculated based on Figure 5 of Dubowski et al.18
3.3. Wavelength Dependence of ΦOH. To test whether the
•OH quantum yield is wavelength dependent, we also measured
ΦÃΗ during illumination with 334-nm radiation. Because of the
lower molar absorptivity of nitrate at 334 nm (0.34 M-1 cm-1
at 263 K, interpolated from values in Table S2), we used ice
pellets made from sample solutions containing a relatively high
nitrate concentration (10 mM) along with 10 mM BA and with
a pH of 5.0. The resulting average value of ΦÃΗ,334 at 263 K
was (3.5 ( 0.5) × 10-3 (N ) 3 determinations), which is
statistically indistinguishable from the average value ((3.4 (
0.6) × 10-3) obtained with 313-nm illumination of the frozen
pH 5.0 sample solutions at 263 K. This suggests that the
quantum yield of •OH is independent of illumination wavelength
within the 300-nm absorbance band, as found previously for
nitrate photolysis in aqueous solution.15
3.4. Quantum Yields of •OH in Aqueous Solutions and in
Ice Pellets as a Function of Temperature. We measured values
of ΦOH in aqueous sample solutions (200 µM NaNO3, 450 µM
BA, pH 5.0) between 278 and 318 K. As shown in Figure 5,
our •OH quantum yields in solution are very similar to
previously reported values, and range from ΦOH ) (7.2 ( 1.6)
× 10-3 at 278 K to (2.2 ( 0.5) × 10-2 at 318 K. We also
measured quantum yields for •OH in ice pellets (T ) 239-268
K) prepared from sample solutions identical with those described
above. As shown in Figure 6, our values of ΦOH in ice ranged
from (1.7 ( 0.3) × 10-3 at 239 K to (5.4 ( 1.0) × 10-3 at
268 K and were significantly higher than previously reported
values.18
It is clear from Figure 6 that our aqueous solution and ice
pellet results have essentially the same temperature dependence.
This is confirmed by a comparison of the calculated activation
energies (Ea) and entropy changes (∆S) (Table 2): values of
Ea and ∆S were statistically indistinguishable between our
solution and ice results. Since the same temperature dependence
applies to both the ice and aqueous data, we combined all of
our ΦOH measurements to obtain a “universal” relationship for
the quantum yield of reaction 1:
ln(Φ1) ) ln(ΦÃΗ) ) -(2400 ( 480)(1/T) + (3.6 ( 0.8)
(12)
where T is in units of Kelvin, errors represent (1σ, and the R2
value is 0.982. At 268 and 240 K, values of Φ1 determined
from eq 12 are 21% and 34% lower, respectively, than
corresponding values determined by extrapolating the temper-
ature dependence of the aqueous solution data of Zellner et al.15

9600 J. Phys. Chem. A, Vol. 107, No. 45, 2003
Chu and Anastasio
In contrast, values of Φ1 from eq 12 are 3 and 9 times higher
at 268 and 240 K, respectively, than values obtained from the
temperature dependence of the ice data of Dubowski et al.18
The fact that our ice pellet and aqueous solution data have
the same temperature dependence suggests that the photolysis
of nitrate on ice between 239 and 268 K occurs in a quasi-
1012 molecules m-2 s-1 (measured on April 21). These latter
authors suggest that the emitted HONO was primarily from
reactions of photoproduced NO2 in the snowpack; in this case
the total equivalent NOx flux should be (7-12) × 1012 molecules
m-2 s-1. On the basis of the data and approach of Simpson and
co-workers,45 but using the ice values for Φ1 and Φ2, we
calculate that the rates of formation of gaseous NO2 and NO in
mid-April at midday (solar zenith angle (θ) ) 71.5°) are 1.5 ×
1012 and 0.5 × 1012 molecules m-2 s-1 from reactions 1 and 2,
respectively. Assuming no loss due to reactions in the snowpack,
this would result in a total flux of NOx out of the snowpack of
2 × 1012 molecules m-2 s-1. (Note that this is close to the
calculated result of Simpson et al.,45 3 × 1012 molecules m-2
s-1, which was based on only reaction 1 with an estimated
quantum yield based on the temperature dependence reported
by Zellner et al.15) Therefore, if the Beine et al.5 flux is the
more representative value, nitrate photolysis in the snowpack
accounted for approximately 40% of the maximum flux of NOx.
However, if the HONO assertion of Zhou et al.44 is correct,
nitrate photolysis accounted for only a small portion (∼20%)
of the maximum measured NOx release at Alert. While there
are significant uncertainties in the flux measurements,5 as well
as in our assumptions of nitrate concentrations and actinic flux
in the snowpack, this comparison suggests either that chro-
mophores other than nitrate are responsible for the bulk of NOx
release from the snow at Alert, or that there are uncharacterized
thermal reactions in the snowpack that led to NOx emissions.
-
liquid, or disordered, layer rather than in the bulk ice. If NO3
photolysis occurred in the bulk ice we would expect the
temperature dependence of the ice experiments to differ from
the solution experiments in two ways: (1) there would be a
noticeable drop in ΦOH (i.e., a discontinuity in Figure 6) when
going from solution to ice, and (2) the slope of the ice data
would be steeper (more negative) than that of the solution data.
•
These effects are expected because measurement of OH from
nitrate photolysis requires that the initial photoproducts (NO2
and •O-) diffuse out of the solvent “cage” and away from each
other to avoid recombination (e.g., refs 15 and 36 and references
therein). Because nitrate diffusion is orders of magnitude slower
in ice compared to that in aqueous solution,37,38 recombination
of the initial photoproducts should be more efficient in a bulk
ice matrix, leading to significantly lower values of ΦOH in ice
(compared to that in solution) and a discontinuity in the ΦOH
Arrhenius plot. Furthermore, diffusion in the bulk ice has a
larger activation energy (Ea ) 50 kJ mol-1; ref 37) compared
to ΦOH in solution (Ea ) 19 kJ mol-1; Table 1). Therefore,
ΦOH in bulk ice would drop off more quickly with decreasing
temperature compared to values in solution, i.e., the slope of
the ice portion of the Arrhenius plot (Figure 6) would be steeper
(more negative) than that of the liquid water portion. The
suggestion that the photolysis of chromophores in ice occurs in
a quasi-liquid layer has been made previously based on results
for both nitrate18 and 4-nitrophenol.39 Other authors have shown
that a quasi-liquid layer exists on water ice at temperatures as
low as 200 K and that properties of the QLL are very sensitive
to the presence of solutes.28,29 This suggests that our reported
temperature dependence for Φ1 might be applicable to temper-
atures as low as 200 K.
We made a similar comparison for a study performed at
Summit, Greenland on June 27, 2000, where the measured 1-h-
average NOx and HONO fluxes at midday (θ ) 66.4°) were
6.2 × 1012 and 0.3 × 1012 molecules m-2 s-1, respectively,
and peak values were approximately two times higher.2 Our
calculated midday NOx flux is 14 × 1012 molecules m-2 s-1
,
based on a temperature of 263 K, the average nitrate content of
the snow (4.4 µM liquid equivalent),2 a snow density of 0.30
g cm-3 (ref 46), and a light attenuation coefficient in the snow
of 0.080 cm-1 (ref 47). Our calculated flux is approximately
two times higher than the 1-h-average measured NOx + HONO
flux, but is close to the peak values, suggesting that nitrate
photolysis accounts for essentially all of the emitted NOx and
HONO in this study. It is interesting to note that our calculated
4. Implications
4.1. NOx Release in Arctic and Antarctic Snow. Several
recent field studies have shown that photochemical reactions
in surface snow result in the release of NOx and have indicated
that nitrate photolysis is an important source of this NOx.1-10
Our goal here is to use our measured values of Φ1, in
combination with values of Φ2 from Dubowski and co-workers
(Figure 5 of ref 19; values for ice without formate), to estimate
the nitrate-derived flux of NOx from the snowpacks in several
-
value of the photolysis rate constant for nitrate (jNO ) at the
3
snow surface under these conditions is 2.6 × 10-7 s-1. This
value is approximately 5 times smaller than the 2-h-average
value (∼14 × 10-7 s-1) measured at midday at Summit on June
7 in the near-surface snow (0-2 cm) with use of quartz tubes
containing nitrate in acetonitrile with radical probe traps.47 This
-
-
comparison suggests that these field measurements of jNO are
of these studies. We consider both NO2 formed from NO3
3
too high, perhaps because of errors in the radical probe
calibration or because ΦOH is different in acetonitrile than in
water.47
photolysis as well as NO formed as a result of nitrite formation
via reaction 2. Given the much greater rate of sunlight absorption
by NO2- and HNO2,40 we assume that photolysis of nitrite and
nitrous acid are fast and that every NO2- formed is immediately
converted to NO and •OH. Values of aerosol optical depth (0.37
for Neumeyer, 0.38 for Alert and Summit) and actinic flux used
Finally, we compare calculated and measured NOx fluxes for
two studies performed at Neumayer, Antarctic in February 1999
by Jones and co-workers. In the first study, the authors measured
a NOx flux of ∼2.4 × 1012 molecules m-2 s-1 from the
snowpack at approximately midday (θ ) 55.3°). In comparison,
we calculate a flux of 2.2 × 1012 molecules m-2 s-1 at 268 K
based on the typical nitrate content of Neumayer snow (50 ng
g-1; ref 48), a snow density of 0.45 g/cm3 (ref 6), and a light
attenuation coefficient of 0.270 cm-1 in the snowpack.48 This
calculated NOx flux is very close to the measured value and to
a model value reported recently by Wolff and co-workers.48 In
the second study at Neumayer, Jones et al. measured the flux
of NOx from near the center of a ∼104 cm3 cube of snow held
in our calculations are from the NCAR TUV model.41
A
wavelength-independent albedo of 0.93 for the snow surface
42
was assumed and ozone columns ranged from 306 to 309
Dobson units.43
The first field campaign we consider is Alert 2000, where
NOx release was measured from April 9 to May 6 in a snowpack
containing the liquid equivalent of 8 µM nitrate and at a midday
temperature of ∼240 K.5 During this campaign Beine et al.5
measured midday NOx fluxes of (0-5) × 1012 molecules m-2
s-1, while Zhou et al.44 reported a midday HONO flux of 7 ×

Photolysis of Nitrate on Ice
J. Phys. Chem. A, Vol. 107, No. 45, 2003 9601
References and Notes
1 m above the ground.6 Measured fluxes of NO and NO2 at
15:00 were 1.1 × 106 and 2.1 × 106 molecules cm-3‚snow s-1
,
(1) Honrath, R. E.; Peterson, M. C.; Guo, S.; Dibb, J. E.; Shepson, P.
B.; Campbell, B. Geophys. Res. Lett. 1999, 26, 695.
(2) Honrath, R. E.; Lu, Y.; Peterson, M. C.; Dibb, J. E.; Arsenault, M.
A.; Cullen, N. J.; Steffen, K. Atmos. EnViron. 2002, 36, 2629.
(3) Dibb, J. E.; Arsenault, M. A.; Peterson, M. C.; Honrath, R. E. Atmos.
EnViron. 2002, 36, 2501.
(4) Ridley, B.; Walega, J.; Montzka, D.; Grahek, F.; Atlas, E.; Flocke,
F.; Stroud, V.; Deary, J.; Gallant, A.; Boudries, H.; Bottenheim, J.; Anlauf,
K.; Worthy, D.; Sumner, A. L.; Splawn, B.; Shepson, P. J. Atmos. Chem.
2000, 36, 1.
(5) Beine, H. J.; Honrath, R. E.; Domine´, F.; Simpson, W. R.; Fuentes,
J. D. J. Geophys. Res.-Atmos. 2002, 107, 4584.
(6) Jones, A. E.; Weller, R.; Wolff, E. W.; Jacobi, H. W. Geophys.
Res. Lett. 2000, 27, 345.
(7) Jones, A. E.; Weller, R.; Anderson, P. S.; Jacobi, H. W.; Wolff, E.
W.; Schrems, O.; Miller, H. Geophys. Res. Lett. 2001, 28, 1499.
(8) Honrath, R. E.; Peterson, M. C.; Dziobak, M. P.; Dibb, J. E.;
Arsenault, M. A.; Green, S. A. Geophys. Res. Lett. 2000, 27, 2237.
(9) Hoffman, M. R. In Chemical Exchange Between the Atmosphere
and Polar Snow; Wolff, E. W., Bales, R. C., Eds.; Springer-Verlag: Berlin,
Germany, 1996; Vol. 43; p 353.
respectively. We calculate a NOx flux of 6.8 × 106 molecules
cm-3‚snow s-1, using values described above, and an average
nitrate value of 40 ng g-1 in the snowblock.6 This calculated
NOx flux is approximately twice as high as the measured flux,
suggesting that nitrate was the dominant NOx source in the
snowblock and that much of the photoformed NOx escaped from
the snowblock rather than being sampled. Thus, in contrast to
the data from Alert, where nitrate photolysis accounted for only
∼20-40% of the measured NOx + HONO flux, it appears that
nitrate photolysis accounts for essentially all of the measured
snowpack flux of NOx at both Neumayer and Summit. In
addition to its importance as a source of NOx, nitrate photolysis
•
in snowpacks will also be a source of OH, which is likely to
have important effects on the chemistry of the snow and the
overlying atmosphere.11,49
4.2. NOx Release in Cirrus Clouds. Recent studies of the
upper troposphere have indicated that photochemical models
typically underestimate the [NOx]/[HNO3] ratio.50 Honrath and
co-workers2 have hypothesized that this might be due to the
photochemical conversion of HNO3 into NOx on cirrus ice
clouds. We have investigated the potential importance of this
mechanism under typical upper tropospheric conditions modeled
by Jaegle´ et al.:50 summer, ∼10 km altitude (P ) 0.35 atm),
35° N, 238 K, and steady-state mixing ratios of HNO3 and NOx
of 240 and 80 pptv, respectively. On the basis of the lifetime
of NOx under these conditions (6 days),50 the rate of loss of
NOx is 1600 molecules cm-3 s-1 (24-h average). If we (i)
assume all of the HNO3 is sorbed to ice particles, (ii) use our
(10) Honrath, R. E.; Guo, S.; Peterson, M. C.; Dziobak, M. P.; Dibb, J.
E.; Arsenault, M. A. J. Geophys. Res.-Atmos. 2000, 105, 24183.
(11) Domine´, F.; Shepson, P. B. Science 2002, 297, 1506.
(12) Dibb, J. E.; Talbot, R. W.; Munger, J. W.; Jacob, D. J.; Fan, S. M.
J. Geophys. Res.-Atmos. 1998, 103, 3475.
(13) Zepp, R. G.; Hoigne´, J.; Bader, H. EnViron. Sci. Technol. 1987,
21, 443.
(14) Warneck, P.; Wurzinger, C. J. Phys. Chem. 1988, 92, 6278.
(15) Zellner, R.; Exner, M.; Herrmann, H. J. Atmos. Chem. 1990, 10,
411.
(16) Mack, J.; Bolton, J. R. J. Photochem. Photobiol. A: Chem. 1999,
128, 1.
(17) Alif, A.; Boule, P. J. Photochem. Photobiol. A: Chem. 1991, 59,
357.
(18) Dubowski, Y.; Colussi, A. J.; Hoffmann, M. R. J. Phys. Chem. A
2001, 105, 4928.
-
(19) Dubowski, Y.; Colussi, A. J.; Boxe, C.; Hoffmann, M. R. J. Phys.
Chem. A 2002, 106, 6967.
measured values of ꢀNO (Figure 3), (iii) use extrapolated values
3
for Φ1 (1.7 × 10-3) and Φ2 (6.0 × 10-4; ref 19) at 238 K, and
(iv) use actinic flux values extrapolated from Finlayson-Pitts
and Pitts,38 we calculate that the 24-h-average rate of NO2
production via photolysis of HNO3 on cirrus ice particles is
110 mlc cm-3 s-1. This rate is much slower than that needed
to maintain steady state for NOx (1600 mlc cm-3 s-1). In
contrast, we calculate that the rate of NOx formation from the
photolysis of 240 pptv HNO3 in the gas phase is fast enough
(∼1400 mlc cm-3 s-1) to approximately maintain steady state,
based on a quantum yield of 1 and absorption cross sections
tabulated in Finlayson-Pitts and Pitts.38 These results suggest
that photolysis of nitrate on cirrus ice particles is a minor source
of NOx that cannot resolve the underestimation of the [NOx]/
[HNO3] ratio in models. However, the large difference in the
values for ΦOH (and the corresponding rates of nitrate photolysis
and NOx formation) in the gas phase and on ice particles
illustrates the need to accurately determine the phase of
measured nitrate in field experiments to properly model NOy
chemistry in the upper troposphere.
(20) Anastasio, C.; McGregor, K. G. Atmos. EnViron. 2001, 35, 1079.
(21) Zhou, X. L.; Mopper, K. Marine Chem. 1990, 30, 71.
(22) Arakaki, T.; Faust, B. C. J. Geophys. Res.-Atmos. 1998, 103, 3487.
(23) Arakaki, T. Personal communication.
(24) Anastasio, C.; Faust, B. C.; Allen, J. M. J. Geophys. Res.-Atmos.
1994, 99, 8231.
(25) Berland, B. S.; Foster, K. L.; Tolbert, M. A.; George, S. M.
Geophys. Res. Lett. 1996, 23, 2757.
(26) Sadtchenko, V.; Ewing, G. E. J. Chem. Phys. 2002, 116, 4686.
(27) Doppenschmidt, A.; Butt, H.-J. Langmuir 2000, 16, 6709.
(28) Cho, H.; Shepson, P. B.; Barrie, L. A.; Cowin, J. P.; Zaveri, R. J.
Phys. Chem. B 2002, 106, 11226.
(29) Wei, X.; Miranda, P. B.; Shen, Y. R. Phys. ReV. Lett. 2001, 86,
1554.
(30) Adamson, A. W. Physical Chemistry of Surfaces, 4th ed.; John
Wiley: New York, 1982.
(31) Clegg S. L.; Brimblecombe P.; Wexler A. S. J. Phys. Chem. A
1998, 102, 2137. Used web version at: mae.ucdavis.edu/wexler/aim.html.
(32) Seinfeld, J. H.; Pandis, S. N. Atmospheric Chemistry and Physics:
From Air Pollution to Climate Change; Wiley: New York, 1998.
(33) Graedel, T. E.; Weschler, C. J. ReV. Geophys. Space Phys. 1981,
19, 505.
(34) Yang, H.; Finlayson-Pitts, B. J. J. Phys. Chem. A 2001, 105, 1890.
(35) Toom-Sauntry, D.; Barrie, L. A. Atmos. EnViron. 2002, 36, 2683.
(36) Chen, W.-J.; Lo, W.-J.; Cheng, B.-M.; Lee, Y.-P. J. Chem. Phys.
1992, 97.
(37) Thibert, E.; Domine´, F. J. Phys. Chem. B 1998, 102, 4432.
(38) Finlayson-Pitts, B. J.; Pitts, J. N. Chemistry of the Upper and Lower
Atmosphere: Theory, Experiments, and Applications; Academic Press: San
Diego, CA, 2000.
(39) Dubowski, Y.; Hoffmann, M. R. Geophys. Res. Lett. 2000, 27, 3321.
(40) Arakaki, T.; Miyake, T.; Hirakawa, T.; Sakugawa, H. EnViron. Sci.
Technol. 1999, 33, 2561.
Acknowledgment. This work was funded by the Office of
Polar Programs at the National Science Foundation (OPP-
9907434). Special thanks to Richard Honrath, Jack Dibb, and
Paul Shepson for getting us involved in the field of ice and
snow photochemistry. We also appreciate the thoughtful com-
ments of the anonymous reviewers.
(41) TUV model, version 4.1, www.acd.ucar.edu/TUV/.
(42) Dickerson, R. R.; Stedman, D. H.; Delany, A. C. J. Geophys. Res.
1982, 87, 4933.
Supporting Information Available: Table S1 containing
the yields of p-hydroxybenzoic acid (Yp-HBA) resulting from
(43) TOMS data, http://toms.gsfc.nasa.gov/.
•
(44) Zhou, X. L.; Beine, H. J.; Honrath, R. E.; Fuentes, J. D.; Simpson,
W.; Shepson, P. B.; Bottenheim, J. W. Geophys. Res. Lett. 2001, 28, 4087.
(45) Simpson, W. R.; King, M. D.; Beine, H. J.; Honrath, R. E.; Zhou,
X. L. Atmos. EnViron. 2002, 36, 2663.
the reaction of OH with benzoic acid (BA) on ice and Table
-
S2 containing molar absorptivities for aqueous nitrate (ꢀNO
)
3 ,λ
as a function of temperature and wavelength. This material is
available free of charge via the Internet at http://pubs.acs.org.
(46) Albert, M. R.; Shultz, E. F. Atmos. EnViron. 2002, 36, 2789.

9602 J. Phys. Chem. A, Vol. 107, No. 45, 2003
Chu and Anastasio
(47) Qiu, R.; Green, S. A.; Honrath, R. E.; Peterson, M. C.; Lu, Y.;
Dziobak, M. Atmos. EnViron. 2002, 36, 2563.
(50) Jaegle´, L.; Jacob, D. J.; Wang, Y.; Weinheimer, A. J.; Ridley, B.
A.; Campos, T. L.; Sachse, G. W.; Hagen, D. E. Geophys. Res. Lett. 1998,
25, 1705.
(51) Lide, D. R., Ed. CRC Handbook of Chemistry and Physics, 80th
ed.; CRC Press: Boca Raton FL, 1999.
(48) Wolff, E. W.; Jones, A. E.; Martin, T. J.; Grenfell, T. C. Geophys.
Res. Lett. 2002, 29, 1944.
(49) Anastasio, C.; Jordan, A. Submitted to Atmos. EnViron.