Observation of gas-phase peroxynitrous and peroxynitric acid during the photolysis of nitrate in acidified frozen solutions

Article abstract of DOI:10.1016/j.cplett.2011.06.055

The photolysis of nitrate embedded in ice and snow can be a significant source of volatile nitrogen oxides affecting the composition of the planetary boundary layer. In this work, we examined the nitrogen oxides evolved from irradiated frozen solutions containing nitrate. Products were monitored by cavity ring-down spectroscopy (CRDS), NO-O₃ chemiluminescence (CL), and chemical ionization mass spectrometry (CIMS). Under acidic conditions, the nitrogen oxides volatilized were mainly in the form of NO_z, i.e., nitrous (HONO), nitric (HONO₂), peroxynitrous (HOONO), and peroxynitric acid (HO₂NO₂). Identification of acidic nitrogen oxides by CIMS and possible HOONO, HONO₂ and HO ₂NO₂ formation pathways are discussed.

Full text of DOI:10.1016/j.cplett.2011.06.055

Chemical Physics Letters 511 (2011) 187–192
Contents lists available at ScienceDirect
Chemical Physics Letters
journal homepage: www.elsevier.com/locate/cplett
Observation of gas-phase peroxynitrous and peroxynitric acid
during the photolysis of nitrate in acidified frozen solutions
1
⇑
Otman Abida , Levi H. Mielke, Hans D. Osthoff
Department of Chemistry, University of Calgary, 2500 University Drive NW, Calgary, Alberta, Canada T2N 1N4
a r t i c l e i n f o
a b s t r a c t
Article history:
Available online 23 June 2011
The photolysis of nitrate embedded in ice and snow can be a significant source of volatile nitrogen oxides
affecting the composition of the planetary boundary layer. In this work, we examined the nitrogen oxides
evolved from irradiated frozen solutions containing nitrate. Products were monitored by cavity ring-
down spectroscopy (CRDS), NO-O3 chemiluminescence (CL), and chemical ionization mass spectrometry
(CIMS). Under acidic conditions, the nitrogen oxides volatilized were mainly in the form of NOz, i.e.,
nitrous (HONO), nitric (HONO2), peroxynitrous (HOONO), and peroxynitric acid (HO2NO2). Identification
of acidic nitrogen oxides by CIMS and possible HOONO, HONO2 and HO2NO2 formation pathways are
discussed.
Ó 2011 Elsevier B.V. All rights reserved.
NOꢀ3 þ h
m
! ONOOꢀ; ONOOꢀ þ Hþ ꢀ HOONO; pKa ¼ 6:6 ½18ꢁ
1. Introduction
ð3Þ
It is well established that photolysis of nitrate anions embedded
in snow or ice can be a significant source of gas-phase nitrogen oxi-
des (NOx = NO + NO2) and hence affect the composition of the plan-
etary boundary layer [1]. Even though the production of volatile
nitrogen oxides by this pathway has been the subject of numerous
laboratory, field, and theoretical studies (e.g., [2–16]) and of a re-
cent review [17], some mechanistic details have yet to be eluci-
dated. Factors known to affect the yield of gas-phase nitrogen
oxides include temperature, presence of organic impurities (e.g.,
[12,13]), ionic strength, and pH. For example, the rate of NOx and
HONO production is increased at lower pH (e.g., [9,10]) for reasons
that are not fully understood.
NO2 and HONO are known to partition to the gas-phase [17].
Observations of nitric oxide (NO) have been rationalized by photo-
dissociation of the nitrite anion formed via reaction (2), whose
absorption spectrum exhibits greater overlap with the actinic spec-
trum than nitrate.
Under actinic conditions (i.e., irradiation wavelengths k >
300 nm), the quantum yield for reaction (3) is small [21]. However,
recombination of the photo fragments produced in reaction (1) can
be an alternate pathway for the photoisomerization of nitrate as
reaction (1) is promoted by the so-called ‘solvent cage-effect’ (e.g.,
[19,22]).
Under conditions encountered in the troposphere, ice surfaces
are usually covered by a thin liquid layer, referred to as the ‘quasi li-
quid layer’ (qll). The photochemistry of nitrate in ice is thus similar
to that in aqueous solution, for which photolysis of nitrate anion has
NO2 þ OH ! HONO2; HONO2 ꢀ NOꢀ þ Hþ; pKa ꢂ ꢀ2
ð4aÞ
3
NO2 þ OH ! HOONO
ð4bÞ
been shown to produce nitrogen dioxide (NO2), nitrite (NOꢀ), and
2
Reaction (4b) can also occur in the gas phase. The latter process
is significant since in the lower troposphere HOONO dissociates to
form NO2 and OH [23], key ingredients in photochemical ozone
production, whereas nitric acid (HONO2), produced via reaction
(4a), is generally considered a permanent sink of NOx. In polluted
environments, the ‘recycling’ of NOx via reaction (4b) can have con-
siderable impacts on ozone budgets [24].
Under conditions relevant to the troposphere, HOONO has been
a rather elusive molecule. In aqueous solution, both the protonated
and deprotonated forms rapidly decompose or isomerize to the
more stable nitrate anion [18,25]. In the gas-phase, HOONO is
predicted to have a very short lifetime (ꢃ10 s) with respect to ther-
mal decomposition at ambient temperature and pressure [23]. As a
consequence, HOONO has not been observed in the troposphere
peroxynitrite (ONOOꢀ) as primary photoproducts [18,19]; the lat-
ter two are protonated under acidic conditions.
NOꢀ3 þ h
NOꢀ3 þ h
m
m
! NO2 þ OÅꢀ
;
O
Åꢀ þ Hþ ꢀ OHÅ; pKa ¼ 11:9 ½20ꢁ ð1Þ
! NOꢀ2 þ Oð PÞ; NO2ꢀ þ Hþ ꢀ HONO; pKa
¼ 2:8 ½8ꢁ
3
ð2Þ
⇑
Corresponding author. Fax: +1 403 289 9488.
E-mail address: hosthoff@ucalgary.ca (H.D. Osthoff).
Present address: Wadsworth Center, New York State Department of Health,
1
Department of Environmental Health and Toxicology, State University of New York,
Albany, NY 12201-0509, USA.
0009-2614/$ - see front matter Ó 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.cplett.2011.06.055

188
O. Abida et al. / Chemical Physics Letters 511 (2011) 187–192
and neither its gas- or liquid-phase lifetime at tropospherically
relevant temperatures nor its Henry’s law constant are constrained
by observations.
power (20 and 40 W). This attenuated the output intensity equally
at all wavelengths (data not shown).
The experiments were conducted using a custom-built Pyrex
photo reactor (internal volume 275 cm3). Samples of frozen solu-
tions were prepared by placing approximately 4 mL of pre-cooled
aqueous solution onto the bottom of the photo reactor tube, which
was cooled to ꢀ20 °C using a circulating chiller (VWR). The reactor
was flushed during sample freezing and irradiation with ultrapure,
or ‘zero’, air (Praxair) at a flow rate of (1.4 0.1) L minꢀ1. Tubing
and compression fittings outside of the photo reactor were con-
structed from PFA or FEP Teflon. The zero air entered the reaction
chamber near the top and exited via a side arm whose opening is
located 1 cm above the ice surface, minimizing the residence time
of gas-phase products and their contact with the inner walls of the
reaction chamber. Upon freezing, polycrystalline ice with an aver-
age thickness of about 2 mm and area of approximately 19.6 cm2
were obtained. The output of the photo reactor was diluted by an
additional zero air dilution flow (approximately 8 L/min). The res-
idence time of the effluent within the Teflon tubing connecting the
photo reactor to the instruments was <1 s.
Studies on the evolution of volatile nitrogen oxides from irradi-
ated frozen solutions have so far relied on chemiluminescence (CL)
detection of NO and of NOy (=NOx + HONO + HNO3 + HNO4....), on
laser-induced fluorescence detection of NO2 (e.g., [4,5]), and/or on
measurement of gas-phase acids using a denuder (also called mist
chamber) coupled to ion chromatography (e.g., [14,26]). It is not
known if peroxynitric and peroxynitrous acid would be detected
by these techniques or if they are (falsely) accounted for as either
NO2 (in case of LIF) or as nitric acid (in case of the mist chamber) [26].
In this work, we investigated the nature of gas-phase nitrogen
oxides evolved from irradiated frozen solutions containing nitrate
anion and show that nitric (HNO3), nitrous (HONO), peroxynitric
(HO2NO2) and peroxynitrous (HO2NO) acid are volatilized under
acidic conditions. We used CL detection to quantify NO and NOy,
blue diode laser cavity ring-down spectroscopy (CRDS) to measure
NO2, and iodide ion chemical ionization mass spectrometry (CIMS)
to monitor acidic nitrogen oxides.
2.2. Detection and quantification of gas-phase nitrogen oxides
2. Experimental
Gas-phase nitrogen oxides were monitored (in parallel) using a
commercial NOy CL analyzer (Thermo 42i), the University of
Calgary blue diode laser CRDS [30], and the University of Calgary
Iꢀ-CIMS [31]. The CRDS uses a blue diode laser to quantify NO2
by optical absorption at 405 nm. The instrument is equipped with
2.1. Ice photolysis experiments
All reagents were analytical grade (>99%) and were used as re-
ceived. Dilute (0.05 M) buffer solutions were prepared by mixing
the appropriate amounts of oxalic acid and sodium oxalate
(Sigma–Aldrich, pH 2.5) in de-ionized water (Barnstead Easypure
multiple detection channels which allows parallel determination
P
of NO2 and either total peroxynitrates ( PAN) by thermal dissoci-
II; 18 MX cm). Aqueous solutions containing up to 1 M sodium ni-
ation (to NO2) at an inlet temperature of 250 °C or total peroxy plus
P
trate (Sigma–Aldrich, 99%) were prepared immediately prior to the
experiments in buffer solution. The pH of the solutions was mea-
sured prior to freezing. Oxalic acid was chosen as a buffer as it is
usually one of the most abundant species in organic aerosol in
the Arctic [27]. The pH chosen is at the lower end of the range typ-
ically found in nature. For example, in heavily polluted air, snow
flake pH values as low 3.0 have been reported [28].
total alkyl nitrates ( AN) at an inlet temperature of 450 °C [30,32].
We also used a red diode laser CRDS recently constructed in our
laboratory to verify the absence of NO3 and N2O5. This instrument
is similar to the one described by [33] and will be described in a
forthcoming publication [34]. It has a limit of detection (1
of 5 part-per-trillion by volume (pptv).
r, 10 s)
The CL analyzer is equipped with a heated Mo catalyst to reduce
the various components of NOy to NO and was calibrated against
CRDS by simultaneous measurements of NO2 in zero air and scaling
the response of the CL analyzer to that of the CRDS. The conversion
efficiency of other NOy species (i.e., HONO and HNO3) was assumed
to be equal to that of NO2.
The CIMS and its operation have been described elsewhere [31]
with the only difference being that the inlet was operated at room
temperature to minimize dissociation of thermally unstable mole-
cules. Iodide reagent ions were generated by passing methyl iodide
pasta210Po ionsource. Sincethe numberofionsgeneratedscaleswith
the amount of(excess) iodidereagent ion, all ion countswerenormal-
ized to 1 ꢄ 106 iodide reagent ion counts prior to presentation.
Table 1 summarizes major ions observed when the CIMS is ex-
posed to various nitrogen oxides. Iodide ions react with many nitro-
gen oxides to form cluster ions (e.g., ClNO2ꢅIꢀ, N2O5ꢅIꢀ, HONOꢅIꢀ, or
HONO2ꢅIꢀ) which are well-suited for quantification due to low back-
ground counts and high specificity. Proton exchange reactions, e.g.,
The photolysis setup is shown in Figure 1. The photolysis light
source was a 150 W Xenon arc lamp (ORIEL ‘Solar Simulator’, mod-
el 96000) equipped with an IR filter (ORIEL 81096) and a digital
exposure controller (ORIEL 68951). Its output was directed through
a cut-off filter window (Pyrex, k > 300 nm) at the top of reactor. At
the maximum power setting, the incident light flux was measured
by ferrioxalate actinometry [29] to be 1.5 ꢄ 1016 photons sꢀ1 cmꢀ2
.
In the experiments shown here, the lamp was operated at lower
to
IR filter
CIMS,
Xe arc
lamp
CRDS,
and NOy CL
Pyrex
window
glycol
out
HONO2 þ Iꢀ ¢ NOꢀ þ HI;
D
G298K ¼ þ8:8 kcal=mol ½35ꢁ
ð5Þ
3
and
Zero
Air
in
HONO þ Iꢀ ¢ NOꢀ2 þ HI;
D
G298K ¼ þ24:7 kcal=mol ½35ꢁ
ð6Þ
are not thermodynamically favored at room temperature. Reactions
of iodide anion with peroxyacids, e.g.,
glycol
in
ice
bypass
valve
HO2NO2 þ Iꢀ ꢀ NOꢀ þ HOI;
ð7Þ
3
are well-known from solution phase chemistry (e.g., [36,37]) and
can also occur in the gas phase. Hence, The ions NOꢀ2 or NO3ꢀ are
Figure 1. Experimental setup. For a discussion, see Section 2.1 in the text.

O. Abida et al. / Chemical Physics Letters 511 (2011) 187–192
189
Table 1
Detection of nitrogen oxides by iodide ion chemical ionization mass spectrometry.
Species
Major ion
Ion forming reaction(s)
Minor ion (m/z, rel. intensity)
Ion forming reaction(s)
Ref.
(m/z, rel. intensity)
NO
–
–
62
–
–
–
–
–
–
–
–
This work
This work
[38]
NO2
NO3
N2O5
NO3 þ Iꢀ ! I þ NOꢀ3
N2O5 ! NO3 þ NO2
62 (variable)
235 (variable)
N2O5 þ Iꢀ ! N2O5 ꢅ Iꢀ
[39]
NO3 þDIꢀ ! I þ NOꢀ3
H2
O
H2O
!
ClNO2
208 (3)
210 (1)
[45]
35ClNO2 þ Iꢀ
!
35ClNO2 ꢅ Iꢀ
37ClNO2 þ Iꢀ
37ClNO2 ꢅ Iꢀ
HONO
HONO2
HO2NO2
Photo reactor
effluent (pH 2.5)
174 (8)
HONO þ Iꢀ ! HONO ꢅ Iꢀ
HONO2 þ Iꢀ ! HONO2 ꢅ Iꢀ
HO2NO2 þ Iꢀ ! HOI þ NO3ꢀ
HONO2 þ Iꢀ ! HONO2 ꢅ Iꢀ
46 (1)
62 (1)
46 (1a)
62 (31)
HONO þ Iꢀ ! HI þ NOꢀ2
HONO2 þ Iꢀ ! HI þ NOꢀ3
?
This work
This work
This work
This work
190 (6.6)
62 (420)
190 (1)
<1%: HONO2 þ Iꢀ ! HI þ NO3ꢀ
>99%: HO2NO2 þ Iꢀ ! HOI þ NO3ꢀ
<1%: HONO þ Iꢀ ! HI þ NO2ꢀ
>99%: HO2NO þ Iꢀ ! HOI þ NO2ꢀ
174 (1.2)
HONO þ Iꢀ ! HONO ꢅ Iꢀ
46 (18)
This work
a
Upper limit.
non-specific since they can arise from proton or electron transfer
reactions or fragmentation of other ions.
oxide (data not shown). Photolysis of nitrate in acidified frozen
solutions yielded a greater amount of gas-phase nitrogen oxides
(NOy), with the major fraction in the form of NOz (=NOy–NOx).
For example, the top panel of Figure 2 shows a time series of NO,
NO2 and NOy emitted from an irradiated frozen sample containing
1.0 M nitrate and 0.05 M oxalate buffer (pH 2.5). In this experi-
ment, the lamp was switched on at 10:35 local time at a power set-
ting of 20 W, switched off briefly at 11:02, and switched on again
at 11:06 at a power setting of 40 W. When the lamp was turned
off, the NO2 signal rapidly returned to baseline. The response of
the CL analyzer was slower, signifying the presence of one or more
‘sticky’ NOy component(s). When the lamp power was increased,
the amount of NO2 and NOy evolved also increased.
The middle panel of Figure 2 shows the CIMS mass counts ob-
served at m/z 62 and 190, respectively. The ratio of the ion counts
observed at m/z 62 relative to m/z 190 (31:1) far exceeded the
amount expected if only HONO2 were present (m/z
62:190 ꢂ 1:6.6; Table 1 and Figure 3). Furthermore, the responses
at m/z 62 and 190 were not well correlated (Figure 2), in particular
during period 1, i.e., from 10:40 to 10:50. Thus, different molecules
were contributing to each of these mass counts. Species that could
contribute to m/z 62 (other than HONO2) are NO3, N2O5 [38], PANs
[31], and HO2NO2 (Table 1). We briefly operated the CIMS with a
heated inlet [31,38] and verified the absence of PANs. Using the
red diode laser CRDS, we verified that neither NO3 nor N2O5 were
formed in the experiment. Thus, the most plausible explanation
is that the counts at m/z 62 were mainly due to HO2NO2, consistent
with the results with the synthetic standard.
The response at both masses was delayed relative to the modu-
lation of the lamp intensity, and was sustained for some time after
the lamp was turned off. This suggests that these mass counts ar-
ose from a ‘sticky’ compound, i.e., compounds whose sorption and
desorption kinetics on the inner walls of the connecting tubing is
slow (HONO2 at m/z 190), or that these compounds continued to
outgas from the ice surface ‘in the dark’ (HO2NO2 at m/z 62).
The bottom panel shows the corresponding time series at m/z 46
and 174. The response at both masses to lamp intensity changes was
more rapid than for m/z 62 and 190. The counts at m/z 46 are corre-
lated with the amount of NO2 observed by CRDS (r2 = 0.89). Since
NO2 does not give a signal at m/z 46, the ratio of ion counts (Figure
3) at m/z 46 relative to m/z 174 is more than 100ꢄ greater than ob-
served for HONO, and the mass spectrometer’s ion chemistry pro-
motes conversion of peroxyacids to anions (Eqs. (7) and (8)), we
conclude that the high counts at m/z 46 are mainly due to HOONO.
We considered that matrix effects may be affecting the ratio of
the cluster ions at m/z 174 and 190 relative to the fragment ions at
2.3. Generation of gas standards to verify response and determine
calibration factors of CIMS
Nitric oxide and nitrogen dioxide were generated by diluting
the output of an NO cylinder (Praxair, 2.28 ppm NO in N2, certified)
and periodically adding ozone (generated by photolysis of oxygen
using a Hg pen ray lamp) to partially oxidize NO to NO2. No ele-
vated ion counts were observed by CIMS in either case. Addition
of stoichiometric amounts of ozone generated a response at m/z
62, interpreted as arising from NO3 or N2O5 [38,39]. We note that
neither NO, NO2, nor NO3/N2O5 produced a signal at m/z 46 (NOꢀ).
2
Gas-phase nitric and nitrous acid were eluted from passive dif-
fusion sources containing either concentrated nitric acid or a con-
centrated solution of sodium nitrite and measured in parallel by
CIMS and CL. For HONO2, the major ion observed was at m/z 190
due to the iodide cluster ion, with a normalized response factor
of 1.2 counts per pptv. Ion counts were also observed at m/z 62
due to fragmentation with an intensity of 1:6 relative to the cluster
ion. For HONO, the major ion observed was HONOꢅIꢀ at m/z 174
with a normalized response factor of 0.14 counts pptvꢀ1. The frag-
ment ion, at m/z 46, had an intensity of 1:8 relative to m/z 174.
Peroxynitric acid was synthesized using the method described
in Ref. [40] and sampled using a passive diffusion source yielding
mass counts at only m/z 62. Since the synthesis method did not
yield a pure sample and m/z 62 is not specific to HO2NO2, a re-
sponse factor was not determined; judging from the parallel CL
measurements, the response factor is orders of magnitudes higher
than that of HONO at m/z 174 or HONO2 at m/z 190.
We also attempted to synthesize gas-phase HO2NO. The syn-
thetic route used was inspired by [41] and involved passing a gas
stream containing HONO over an aqueous solution containing
50% H2O2. The resulting mass spectrum (not shown) contained
peaks at m/z 190, 174, 62 and 46. The presence of m/z 46 can be
rationalized by a reaction analogous to reaction (7):
HO2NO þ Iꢀ ꢀ NOꢀ þ HOI:
ð8Þ
2
3. Results
3.1. Production of gas-phase nitrogen oxides
Photolysis of nitrate anion in either unbuffered or basic frozen
solutions yielded NO2 as the major (>90%) gas-phase nitrogen

190
O. Abida et al. / Chemical Physics Letters 511 (2011) 187–192
1.0
0.8
0.6
0.4
0.2
62
1000
800
600
400
200
0
Illuminated ice
HONO
46
NO
0W
190
174
NO2
NOy
HNO3
Blank (in the dark)
0W
40W
20W
0.0
200
4
3
2
1
0W
40
60
80
100
120
m/z
140
160
180
m/z 62
m/z 190
150
100
50
Figure 3. CIMS mass spectra, offset in increments of 100 counts for clarity. The
bottom (red) trace and the top (blue) scans were taken during the time periods
indicated in Figure 2. The scans offset by 100 and 200 counts (gold and black,
respectively) are reference mass spectra of HONO2 and HONO, respectively. (For
interpretation of the references in color in this figure legend, the reader is referred
to the web version of this article.)
40W
0W
20W
P
channel is included, the sum of nitrogen oxides, NOy (hot), equals
101% of the observed NOy (Table 2).
0
200
0
4
3.2. Radical quenching experiments
0W
m/z 46
m/z 174
150
100
50
3
2
1
0
To gain insight into the mechanism for HO2NO and HO2NO2 for-
mation, radical quenching experiments were carried out in which
either hexane was periodically added to the zero air entering the
photolysis chamber or 0.02 M isopropanol was added to the solu-
tions prior to freezing. The presence of isopropanol reduced the
overall yield of nitrogen oxides without affecting the fractionation
of NOy. The addition of hexane to the gas-phase had no statistically
relevant effect on NOy.
40W
0W
20W
0
10:15
10.Aug.2010
10:30
10:45
Time (local)
11:00
11:15
4. Discussion
4.1. Acid-promoted production of HONO(g) and NO(g)
Figure 2. Production of gas-phase nitrogen oxides from an irradiated frozen
solution containing 1.0 M nitrate anion and 0.05 M oxalate buffer (pH 2.5). (top)
Time series of NO, NO2, and NOy mixing ratios. The power levels indicated are those
of the Xe arc lamp. (middle) Time series of m/z 190 (HONO2ꢅIꢀ) and m/z 62 (NO3ꢀ) for
the same experiment. The pale red and blue underlays are the times during which
mass spectra (Figure 3) were acquired. (bottom) Time series of m/z 174 (HONOꢅIꢀ)
and m/z 46 (NOꢀ2 ) for the same experiment. (For interpretation of the references in
color in this figure legend, the reader is referred to the web version of this article.)
Under acidified conditions, photolysis of nitrate produced a
considerable amount of gas-phase NOz species. The observed for-
mation of HONO is consistent with previous reports (e.g.,
[6,12,15,16]) and can be rationalized as follows: Photolysis of ni-
trate anion is known to produce nitrite (reaction 2), which in acidic
medium is protonated and partitions to the gas-phase (Henry’s law
constant at room temperature = 49 M atmꢀ1 [42]). Beine and
coworkers have reported positive fluxes of nitrous acid under
acidic conditions [16], while under basic conditions, emission of ni-
trous acid was found to be suppressed [15]. The experiment here
demonstrates acid-promoted release of HONO to the gas-phase.
Subsequent photolysis of nitrous acid, whose absorption spectrum
has a greater overlap with the lamp’s emission spectrum than ni-
trate, yields NO (as observed).
m/z 46 and 62. To verify the absence of such effects, standard addi-
tion experiments were carried out in which the effluent of the
photo reactor was mixed with either nitrous or nitric acid. The
ion ratios observed (not shown) were identical to those in the off-
line experiment shown in Figure 3, corroborating that the ion
counts at m/z 46 and 62 are mainly due to molecules other than ni-
trous or nitric acid.
Using the calibration factors for HONO2 at m/z 190 and for
HONO at m/z 174, the fractionation of NOy was calculated (Table
2). For period 2, the major NOy species evolved are HONO (35% rel-
4.2. Acid-promoted production of HONO2(g), HO2NO2(g), and
HO2NO(g)
ative to measured NOy), NO2 (33%), NO (12%) and HONO2 (6%). The
P
sum of these nitrogen oxides, NOy (cold), accounts for 86% of the
observed NOy. The amount of NO2 observed in the heated CRDS
channel is slightly greater than that observed in the room temper-
ature channel (data not shown), approximately 47% relative to NOy
observed. The difference between the ‘hot’ and ‘cold’ CRDS channel
observed is approximately 13% (relative to NOy). Because NO3,
N2O5, and PANs were absent, the difference signal was likely due
to peroxy nitrates. If the amount of NO2 observed in the ‘hot’ CRDS
The data provide strong evidence that nitrogen oxides other
than NO, NO2 and HONO are emitted by the irradiated ice. It is cer-
tain that HONO2 is generated. It is unlikely that HONO2 forms in
the ice and is then volatilized because its pKa is approximately
ꢀ2, i.e., deprotonation to the nonvolatile nitrate anion is strongly
favored in ice, and because its Henry’s law constant for water at
room temperature is large (ꢃ105 M atmꢀ1 [42]). Hence, it is more

O. Abida et al. / Chemical Physics Letters 511 (2011) 187–192
191
Table 2
Yields of NOy species (in ppbv) for the experiment shown in Figure 2. The uncertainties are
1r, i.e., indicate the precision of the measurements.
Species
Period 1 (10:40–10:50)
Period 2 (11:13–11:18)
Percentage rel. to NOy (obs)
(%)
NO
0.06
0.05
0.10
0.04
0.03
0.04
0.07
0.01
0.08
0.22
0.32
0.24
0.04
0.58
0.68
0.67
0.03
0.04
0.07
0.11
0.02
12
NO2 (cold CRDS inlet)
NO2 + (heated CRDS inlet)
HONO
33
47
35
6
86
101
–
0.09
HONO2
ꢀ0.01
0.19
0.24
0.29
P
NOy (cold CRDS inlet)
NOy (heated CRDS inlet)
NOy (obs)
P
0.05
0.07
HO2 þ NO2 ! HO2NO2:
ð9Þ
likely that gas-phase HONO2 is produced by a secondary process,
e.g., by reaction (4a). It is unclear if reaction (4a) occurs in the
gas phase, either in the tubing connecting the reactor to the CIMS
or just above the ice surface, or in the condensed phase (i.e., within
the qll), even though the latter two are more likely since addi-
tion of hexane to the zero air had no obvious effect on the produc-
tion of nitrogen oxides and addition of isopropanol lowered the
yield of nitrogen oxides.
Peroxynitric acid has a pKa of 5.85 [43], and will be present
mainly in the protonated form under the conditions of the experi-
ment. It’s Henry’s law constant for room temperature solutions is
approximately 104 M atmꢀ1 [36]. Considering that volatilization of
HONO2 was observed whose Henry’s law constant is one order of
magnitude larger, it is not surprising that volatilization of HO2NO2
is also observed.
If reaction (4a) is used to rationalize production of HONO2, it
implies that NO2 and OH are there in abundance. Thus, reaction
(4b) should also occur and HOONO should be produced. This is
consistent with the observation of very high ion counts at m/z
46, assigned to HOONO. One interesting feature of the counts
at m/z 46 is that they respond rapidly to changing light intensity,
unlike those at the other nitrogen oxide masses observed by
CIMS. This is, at first, surprising, since one might expect rela-
tively slow sorption and desorption kinetics on the inner walls
of the connecting tubing. However, since HOONO is thermally
unstable, it is unlikely that slow desorption kinetics from the in-
ner walls will give rise to inlet memory effects or a sustained
background signal.
The source of HO2 is somewhat uncertain. It is known that oxa-
late in the presence of Fe3+ produces superoxide (Oꢀ2 ) [29]. The
hydroperoxyl radical has a pKa of approximately 4.8 [44]; thus,
HO2 could be produced from protonation of Oꢀ2 . However, Fe3+ is
present at most in trace quantities (arising from impurities in the
chemical reagents); hence, we expect this pathway to be negligi-
ble. Another potential HO2 source is photolysis of ONOOꢀ [25].
Since ONOOꢀ is rapidly protonated at pH 2.5, we speculate that
HO2 is a product of HOONO photolysis.
Peroxynitric acid is thermally unstable. It is hence interesting to
observe that the response to changes in lamp power of m/z 62 is
slower than of m/z 46, which suggests either that HO2NO2 slowly
desorbs from the inner walls of the Teflon tubing, outgasses from
the ice surface ‘in the dark’, or both.
4.2.1. Proposed mechanism for formation of HOONO and HONO2
Since the quantum yield for HOONO formation from nitrate
photolysis is small in the actinic region [21], HOONO is most likely
formed by reaction (4b), i.e., recombination of the nitrate photo
products, Oꢅꢀ (which rapidly protonates to OH) and NO2. This reac-
tion is likely promoted by the solvent cage. Since HOONO’s pKa is
6.6 [18], the protonated form will dominate at pH 2.5. The Henry’s
law constant for the partitioning of HOONO between the ice and
gas phase is not known; however, it is reasonable to assume that
the Henry’s law constant of HOONO is more similar to that of
HONO than of HONO2. Considering that the rate of thermal decom-
position and isomerization of HOONO in ice is likely slower in ice
at ꢀ20 °C than in solution at room temperature, it is plausible that
HOONO is sufficiently long-lived to at least partially partition to
the gas-phase.
5. Conclusions
This work demonstrates that HONO2, HOONO and HO2NO2 can
be formed as volatile secondary products following photolysis of
nitrate anion in acidic ice and that these molecules can be observed
by mass spectrometry. The release of these nitrogen oxides is not
currently considered in mechanisms describing the photochemical
transformation of nitrate in snow (e.g., [9]); however, the results
presented here suggest that this pathway could be significant. Fur-
ther work, including measurements of the yields of nitrogen oxides
in intermediate pH ranges and at lower nitrate concentrations, will
be needed to assess the importance of this chemistry in the
troposphere.
Once partitioned to the gas-phase, the fate of HOONO is thermal
decomposition. This would have ‘masked’ HOONO formation in
previous experiments; we can observe HOONO mainly because of
the rapid analysis time (<1 s) used. The decomposition products,
OH and NO2, can combine to HONO2 (reaction (4a)) or recombine
to form HOONO (reaction (4b)); however, since OH is efficiently
‘lost’ (e.g., by wall reactions), a considerable fraction of the emitted
HOONO will end up as NO2. Hence, volatilization of HOONO is one
of the mechanisms by which NO2 is volatized from the ice surface.
This mechanism would consistent with observations by Honrath
et al. [14] and others, who have noted that the presence of OH rad-
ical quenchers in the ice suppresses the yield of NO2.
Acknowledgments
We gratefully acknowledge financial support by the Canadian
Foundation for Climate and Atmospheric Sciences (CFCAS) under
Grant No. GR-7054 and funding from the Alberta Ingenuity Fund
in the form of a New Faculty Award. The CIMS used in this work
was purchased using a National Science and Engineering Research
Council (NSERC) Research Tools and Instruments (RTI) Grant.
References
[1] R.E. Honrath, M.C. Peterson, S. Guo, J.E. Dibb, P.B. Shepson, B. Campbell,
Geophys. Res. Lett. 26 (1999) 695.
[2] L. Chu, C. Anastasio, J. Phys. Chem. A 107 (2003) 9594.
4.2.2. Proposed mechanism for formation of HO2NO2
The most likely HO2NO2 formation pathway is from reaction of
NO2 with HO2 within the qll:
[3] Y. Dubowski, A.J. Colussi, M.R. Hoffmann, J. Phys. Chem. A 105 (2001) 4928.
[4] C.S. Boxe, A.J. Colussi, M.R. Hoffmann, J.G. Murphy, P.J. Wooldridge, T.H.
Bertram, R.C. Cohen, J. Phys. Chem. A 109 (2005) 8520.

192
O. Abida et al. / Chemical Physics Letters 511 (2011) 187–192
[5] C.S. Boxe, A.J. Colussi, M.R. Hoffmann, I.M. Perez, J.G. Murphy, R.C. Cohen, J.
[27] K. Kawamura, H. Kasukabe, L.A. Barrie, Atmos. Environ. 30 (1996) 1709.
[28] J.L. Colin, D. Renard, V. Lescoat, J.L. Jaffrezo, J.M. Gros, B. Strauss, Atmos.
Environ. 23 (1989) 1487.
[29] J.G. Calvert, J.N. Pitts, Photochemistry, Wiley, New York, 1966.
[30] D. Paul, H.D. Osthoff, Anal. Chem. 82 (2010) 6695.
[31] A. Furgeson, L.H. Mielke, D. Paul, H.D. Osthoff, Atmospheric Environment 45
(2011). doi:10.1016/j.atmosenv.2011.03.072.
[32] D. Paul, A. Furgeson, H.D. Osthoff, Rev. Sci. Instrum. 80 (2009) 114101.
doi:10.1063/1.3258204.
[33] W.R. Simpson, Rev. Sci. Instrum. 74 (2003) 3442.
[34] C.A. Odame-Ankrah, H.D. Osthoff, Applied Spectroscopy (2011) submitted
(June 15), manuscript i.d. 11-06384.
[35] J.E. Bartmess, in: P.J. Linstrom, W.G. Mallard (eds.), National Institute of
Standards and Technology, Gaithersburg, MD, 2011, <http://webbook.nist.
gov> (retrieved 16.06.11).
[36] J.-M. Régimbal, M. Mozurkewich, The Journal of Physical Chemistry A 101
(1997) 8822.
Phys. Chem. A 110 (2006) 3578.
[6] H.J. Beine, A. Amoroso, F. Dominé, M.D. King, M. Nardino, A. Ianniello, J.L.
France, Atmos. Chem. Phys. 6 (2006) 2569.
[7] H.J. Beine, F. Domine, A. Ianniello, M. Nardino, I. Allegrini, K. Teinila, R. Hillamo,
Atmos. Chem. Phys. 3 (2003) 335.
[8] C. Anastasio, L. Chu, Environ. Sci. Technol. 43 (2009) 1108.
[9] H.W. Jacobi, B. Hilker, J. Photochem. Photobiol. A-Chem. 185 (2007) 371.
[10] C.S. Boxe, A. Saiz-Lopez, Atmos. Chem. Phys. 8 (2008) 4855.
[11] J.L. Thomas, M. Roeselova, L.X. Dang, D.J. Tobias, J. Phys. Chem. A 111 (2007)
3091.
[12] T. Bartels-Rausch et al., Atmos. Environ. 44 (2010) 5443.
[13] C. George, R.S. Strekowski, J. Kleffmann, K. Stemmler, M. Ammann, Faraday
Discuss. 130 (2005) 195.
[14] R.E. Honrath, S. Guo, M.C. Peterson, M.P. Dziobak, J.E. Dibb, M.A. Arsenault, J.
Geophys. Res.-Atmos. 105 (2000) 24183.
[15] H.J. Beine et al., Geophys. Res. Lett. 32 (2005) L10808.
[16] X.L. Zhou, H.J. Beine, R.E. Honrath, J.D. Fuentes, W. Simpson, P.B. Shepson, J.W.
Bottenheim, Geophys. Res. Lett. 28 (2001) 4087.
[37] J.-M. Régimbal, M. Mozurkewich, The Journal of Physical Chemistry A 104
(2000) 6580.
[17] A.M. Grannas et al., Atmos. Chem. Phys. 7 (2007) 4329.
[18] J. Mack, J.R. Bolton, J. Photochem. Photobiol., A 128 (1999) 1.
[19] G. Mark, H.-G. Korth, H.-P. Schuchmann, C. von Sonntag, J. Photochem.
Photobiol., A 101 (1996) 89.
[20] G.V. Buxton, C.L. Greenstock, W.P. Helman, A.B. Ross, J. Phys. Chem. Ref. Data
17 (1988) 513.
[21] S. Goldstein, J. Rabani, J. Am. Chem. Soc. 129 (2007) 10597.
[22] T.G. Koch, J.R. Sodeau, J. Phys. Chem. 99 (1995) 10824.
[23] D.M. Golden, J.R. Barker, L.L. Lohr, J. Phys. Chem. A 107 (2003) 11057.
[24] A.K. Mollner et al., Science 330 (2010) 646.
[25] R. Kissner, T. Nauser, P. Bugnon, P.G. Lye, W.H. Koppenol, Chem. Res. Toxicol.
10 (1997) 1285.
[38] D.L. Slusher, L.G. Huey, D.J. Tanner, F.M. Flocke, J.M. Roberts, J. Geophys. Res.-
Atmos. 109 (2004) D19315. doi:10.1029/2004JD004670.
[39] J.P. Kercher, T.P. Riedel, J.A. Thornton, Atmos. Meas. Tech. 2 (2009) 193.
[40] E.H. Appelman, D.J. Gosztola, Inorg. Chem. 34 (1995) 787.
[41] K.M. Robinson, J.S. Beckman, Nitric Oxide, Pt E, 2005. pp. 207–214.
[42] S.E. Schwartz, W.H. White, in: J.R.Z. Pfafflin (Ed.), Advances in Environmental
Science and Engineering, Gordon and Breach Science Publishers, NY, 1981, pp.
1–45.
[43] T. Loegager, K. Sehested, J. Phys. Chem. 97 (1993) 10047.
[44] B.H.J. Bielski, D.E. Cabelli, R.L. Arudi, A.B. Ross, J. Phys. Chem. Ref. Data 14
(1985) 1041.
[45] H.D. Osthoff et al., Nat. Geosci. 1 (2008) 324.
[26] R.W. Talbot, A.S. Vijgen, R.C. Harriss, J. Geophys. Res.-Atmos. 95 (1990) 7553.