Heterogeneous chemistry of the NO3 free radical and N 2O5 on decane flame soot at ambient temperature: Reaction products and kinetics

Article abstract of DOI:10.1021/jp0670891

The interaction of NO₃ free radical and N₂O ₅ with laboratory flame soot was investigated in a Knudsen flow reactor at T = 298 K equipped with beam-sampling mass spectrometry and in situ REMPI detection of NO₂ and NO. Decane (C₁₀H₂₂) has been used as a fuel in a co-flow device for the generation of gray and black soot from a rich and a lean diffusion flame, respectively. The gas-phase reaction products of NO₃ reacting with gray soot were NO, N ₂O₅, HONO, and HNO₃ with MONO being absent on black soot. The major loss of NO₃ is adsorption on gray and black soot at yields of 65 and 59%, respectively, and the main gas-phase reaction product is N₂O₅ owing to heterogeneous recombination of NO₃ with NO₂ and NO according to NO₃ + {C} → NO + products. HONO was quantitatively accounted for by the interaction of NO₂ with gray soot in agreement with previous work. Product N ₂O₅ was generated through heterogeneous recombination of NO₃ with excess NO₂, and the small quantity of HNO ₃ was explained by heterogeneous hydrolysis of N₂O ₅. The reaction products of N₂O₅ on both types of soot were equimolar amounts of NO and NO₂, which suggest the reaction N₂O₅ + {C} → N₂O_3(ads) + products with N₂O_3(ads) decomposing into NO + NO ₂. The initial and steady-state uptake coefficients γ₀ and γ_ss of both NO₃ and N ₂O₅ based on the geometric surface area continuously increase with decreasing concentration at a concentration threshold for both types of soot. γ_ss of NO₃ extrapolated to [NO ₃] → 0 is independent of the type of soot and is 0.33 ± 0.06 whereas γ_ss for [N₂O₅] → 0 is (2.7 ± 1.0) × 10^-2 and (5.2 ± 0.2) × 10^-2 for gray and black soot, respectively. Above the concentration threshold of both NO₃ and N₂O₅, γ_ss is independent of concentration with γ _ss(NO₃) = 5.0 × 10^-2 and γ_ss-(N₂O₅) = 5.0 × 10^-3. The inverse concentration dependence of γ below the concentration threshold reveals a complex reaction mechanism for both NO₃ and N₂O₅. The atmospheric significance of these results is briefly discussed.

Full text of DOI:10.1021/jp0670891

1914
J. Phys. Chem. A 2007, 111, 1914-1926
Heterogeneous Chemistry of the NO3 Free Radical and N2O5 on Decane Flame Soot at
Ambient Temperature: Reaction Products and Kinetics
†
Federico Karagulian and Michel J. Rossi*
Laboratoire de Pollution Atmosph e´ rique et Sol (LPAS), Station 6, CH H5, Ecole Polytechnique F e´ d e´ rale
de Lausanne (EPFL), CH-1015 Lausanne, Switzerland
ReceiVed: October 29, 2006; In Final Form: January 10, 2007
The interaction of NO
reactor at T ) 298 K equipped with beam-sampling mass spectrometry and in situ REMPI detection of NO
and NO. Decane (C10 22) has been used as a fuel in a co-flow device for the generation of gray and black
soot from a rich and a lean diffusion flame, respectively. The gas-phase reaction products of NO reacting
with gray soot were NO, N , HONO, and HNO with HONO being absent on black soot. The major loss
of NO is adsorption on gray and black soot at yields of 65 and 59%, respectively, and the main gas-phase
reaction product is N
{C} f NO + products. HONO was quantitatively accounted for by the interaction of NO
in agreement with previous work. Product N was generated through heterogeneous recombination of NO
with excess NO , and the small quantity of HNO was explained by heterogeneous hydrolysis of N
reaction products of N on both types of soot were equimolar amounts of NO and NO , which suggest the
reaction N + {C} f N 3(ads) decomposing into NO + NO . The initial and
steady-state uptake coefficients γ and γss of both NO and N based on the geometric surface area
continuously increase with decreasing concentration at a concentration threshold for both types of soot. γss of
NO extrapolated to [NO ] f 0 is independent of the type of soot and is 0.33 ( 0.06 whereas γss for [N
f 0 is (2.7 ( 1.0) × 10 and (5.2 ( 0.2) × 10 for gray and black soot, respectively. Above the concentration
threshold of both NO and N , γss is independent of concentration with γss(NO
) ) 5.0 × 10 and γss
3 2 5
free radical and N O with laboratory flame soot was investigated in a Knudsen flow
2
H
3
2
O
5
3
3
2
O
5
owing to heterogeneous recombination of NO
3
with NO
2
and NO according to NO
3
+
2
with gray soot
2 5
O
3
2
3
2 5
O . The
2
O
5
2
O
2 5
2
O
3(ads) + products with N
2
O
2
0
3
2 5
O
3
3
2
2 5
O ]
-
-2
-
2
3
O
2 5
3
-
-
3
(N
2
O
5
) ) 5.0 × 10 . The inverse concentration dependence of γ below the concentration threshold reveals
a complex reaction mechanism for both NO
briefly discussed.
3 2 5
and N
O
. The atmospheric significance of these results is
atmosphere.9 These reaction products of intermediate oxida-
tion state can then bring about other redox reactions not possible
without the occurrence of reducing soot particles. There exist
many heterogeneous chemical reactions involving carbonaceous
particles of soot substrates of potential relevance to the
atmosphere that have been compiled and evaluated in the past.13
Last, combustion aerosol emitted into the boundary layer has
been recently recognized to be a public health issue.1
-12
1. Introduction
Atmospheric aerosols are recognized to be important as a
global climate factor whose uncertainty has given rise to
numerous climate change scenarios for time horizons of 50-
1
00 years.1 In addition, they may affect the atmospheric
composition through their interaction with trace gases in cases
where they attain significant surface-to-volume ratios that enable
many heterogeneous or multiphase reactions. Soot is a signifi-
4,15
2-4
Soot is a carbon-containing aerosol resulting from incomplete
combustion of hydrocarbon fuel of varying stoichiometry,
defined by the λ-ratio of fuel to oxygen. Soot usually implies
only the condensed phase and excludes the volatile organic
fraction as well as the surrounding gas phase. It is formed
through a complex network of chemical processes from small
C2-C4 hydrocarbon fragment free radicals that results in large
stacked molecular condensates that in part contain substituted
aromatic structures depending on the combustion conditions.1
Despite long-standing efforts at elucidating the soot formation
mechanism, the molecular details have not emerged yet even
cant component of many atmospheric aerosols
that plays a
unique role for several reasons. Despite the hydrophobic nature
of freshly emitted soot, it is thought to cause cloud formation
processes by providing cloud condensation (CCN) or ice (IN)
nuclei depending on the age, and thus surface composition, of
5
the aerosol particles and is therefore an important factor in
6
global climate change. In addition, it is the only globally
occurring aerosol that has strong absorbing properties from the
near IR to the UV spectral range and therefore significantly
affects the global radiative balance by absorbing both terrestrial
6-18
(
long wave) as well as solar (short wave) components of the
7,8
18
actinic spectrum. Soot particles have reducing properties and
therefore generate interesting reaction products such as HONO
not normally resulting from gas-phase processes in the oxidizing
though useful structural models have been proposed. Soot
particles are composed of insoluble amorphous carbon, some-
times called elemental (EC) or black carbon (BC), and a soluble
organic fraction (OC) containing up to 50% of the soot particle
mass under certain conditions, of partially oxidized hydrocarbons
as a bulk phase associated with the EC core or forming an
*
To whom correspondence should be addressed. E-mail:
michel.rossi@epfl.ch or mj.rossi@bluemail.ch.
†
Present address: Department of Chemistry, UC Irvine, Irvine, CA
16
9
2697-2025.
adsorbed layer on the particle surface. This soluble OC fraction
1
0.1021/jp0670891 CCC: $37.00 © 2007 American Chemical Society
Published on Web 02/22/2007

NO3 Free Radical and N2O5 on Decane Flame Soot
J. Phys. Chem. A, Vol. 111, No. 10, 2007 1915
may span the gamut of high molecular weight compounds from
complex organic matter such as humic acid-like substances
TABLE 1: Characteristic Parameters of Flame Soot Used
To Produce Decane Soot
(
HULIS) to highly annellated polycyclic aromatic hydrocarbons
flame
type
fuel duct
BET surface diam of soot
whose reactivity with trace gases has been the recent focus of
air flow (pore Ø)
area
particle
(decane) [L min-1] [µm]
soot type [m g ]22
2
-1
[nm]
21
research.1
9,20
This organic fraction may play a significant role
in the reactivity of soot and may take an active part in the
formation of reaction products such as HONO resulting from
the reduction of NO2 on soot according to eq 1.
rich
lean
1.2-1.4 17-40
1.3-1.5 11-16
“gray”
“black”
69
218
40
20
of flame soot. The samples were collected from the burnt gases
at 1 cm above the visible flame on ambient temperature Pyrex
glass plates of 19.6 cm surface area. The different soot
NO + {C-H} f HONO + {C}ox
(1)
2
red
2
HONO formation is an example of an interesting reaction
substrates used in this work are meant to show the change in
the chemical and physical properties of fresh soot emitted into
the atmosphere by combustion sources when two limiting
conditions of combustion, such as rich vs lean, are chosen.
The reactivity of the soot samples toward NO3 and N2O5 was
examined in a low-pressure (Knudsen) flow reactor using
modulated molecular beam mass spectrometry (MMBMS) for
the measurement of the rate of reactant uptake and product
release into the gas phase. To unambiguously monitor the
concentration of NO and NO2 in situ, resonance enhanced
multiphoton ionization (REMPI) was employed as part of a
multi-diagnostic experimental technique in addition to MMBMS.
Design and operation of the Knudsen flow reactor have been
product that previous studies have identified as a result of a
heterogeneous reaction of NO2 with soot.1
6,21,22,31-39
However,
it was conjectured that there must be additional sources of
HONO in the atmosphere because the quantities of soot
necessary to support the observed levels of atmospheric HONO
are not sufficient. Three recent examples are the light-supported
(
actinic) formation of HONO in the presence of humic acids
40
that are ubiquitous in the environment, the generation of
HONO from irradiated nitrate-containing ice in a hitherto
41
unknown reaction mechanism, and the interaction of NO2 with
NH4Cl aerosol.42 It is very likely that there are several significant
additional day and/or nighttime sources of HONO. The present
work contributes an additional mechanism of HONO formation
in the presence of N2O5 and a specific type of soot substrate.
The primary particle size of soot aerosol may range from 10
to 150 nm and a lifetime of several to tens of days before being
scavenged in wet deposition processes. During their lifetime as
suspended aerosol they are actively partaking in heterogeneous
43
described in detail in the literature as well as the in situ REMPI
44
detection of NO and NO2 performed using a Nd:YAG-pumped
dye laser. The gas-kinetic parameters of the Knudsen flow
reactor as well as details pertaining to in situ REMPI detection
of NO and NO2 may be found in refs 43 and 44. The
measurement of molecular flux and concentration was performed
chemical reactions with gaseous pollutants such as HNO3, NO2,
44
using MS and REMPI detection, as reported in previous work.
and O39
,21-23
and therefore reduce oxidized species in the
The mass of soot was varied between 1.5 and 20 mg spread
atmosphere. These processes are noncatalytic, consume the
reactive sites of soot and thus modify its structure through
surface oxidation in a typical aging process that generally leads
2
out over 19.6 cm . Each sample was pumped for 10 min before
performing an uptake experiment. The gas under study, NO3,
was generated by thermal decomposition of N2O5 at 530 K in
a 6 mm diameter glass tube that directly extends into the reactor.
Due to the low residence time and pressure in the reactor, there
is no recombination of NO3 with NO2 to form N2O5. Production
of NO3 and characterization of the NO3 source have been
16,24-29
to an increase of soot surface oxygen functionalities.
Until
now, no experimental work has been reported on the interaction
of NO3 on soot in contrast to laboratory experiments of N2O5
1
1,47
reacting with soot.
The present work deals with the interaction of NO3 and N2O5
with soot substrates from a laboratory combustion aerosol
simulating an exhaust gas plume of a combustion engine. Active
nitrogen or NOy, which includes NOx, is connected to the
abundance of global ozone because NOx directly partakes in
atmospheric oxidation processes being a catalytic species,
whereas NOy species act as NOx reservoir. The relative
abundance of NOx and NOy in the planetary boundary layer
can change significantly with altitude even within 10 m, and
4
4
reported in detail in a recent study. Hydrolysis of N2O5 may
occur on internal surfaces of the inlet line before flowing across
the hot glass tube thus generating HNO3 as an impurity on the
order of a few to 15%. However, HNO3 does not thermally
decompose inside the hot glass tube of the NO3 source because
we did not observe any change in the MS signal amplitude at
+
m/e 63 (HNO3 ) when increasing the source temperature from
ambient to 530 K. Both NO3 and N2O5 were introduced into
the reactor as a continuous flow from the NO3 source which
was heated for NO3 and was run at ambient temperature for
N2O5. The used free radical NO3 source consistently generated
a 3-fold excess of NO2 owing to heterogeneous secondary
processes of NO3 to NO2 within the source, as reported before.44
This associated and unfortunately unavoidable NO2 flow is in
part responsible for the obtained reaction products (see below).
2.1. NO3. The relative concentration of NO3 decreases with
decreasing escape orifice diameter because it undergoes increas-
ing wall loss with increasing residence time in the Knudsen
reactor. The measured first-order loss of NO3 given by k′ at a
given escape orifice size of the reactor may be expressed as
the sum of two components, namely, kesc and kdec, representing
escape, that is, physical, and chemical loss, respectively: k′ )
3
0
with it its characteristic influence on tropospheric ozone.
2. Experimental Method
A custom-designed co-flow system has been used to produce
2
2
flame soot from decane fuel in a reproducible way. It leads
to a diffusion flame maintained in a controlled flow of air. To
regulate the fuel flow feeding the flame by capillary forces, two
types of ceramics of different porosity were used. One type of
soot has been generated in a lean flame (low fuel/oxygen ratio)
and will be referred to as “black” soot; the second has been
generated in a rich (high fuel/oxygen ratio) and will be referred
to as “gray” soot. These correspond to two types of soot obtained
under two limiting flame conditions for the given combustion
device used whose internal surface measured by BET and
elemental composition have been given.22 Table 1 displays the
characteristic parameters we have used to produce the two types
4
4
kesc + kdec. As reported in recent work, the measurement of k′
at three different aperture sizes enabled the determination of
-
1
kdec ) 0.6 ( 0.27 s at 300 K. NO3 has a measurable parent

1916 J. Phys. Chem. A, Vol. 111, No. 10, 2007
Karagulian and Rossi
Figure 1. N
(8.0 ( 1.3) × 10 cm and at an orifice diameter of 4 mm. Curves
a and b correspond to the raw MS signals monitored at m/e 46 and 30,
respectively. Curve c corresponds to the raw REMPI signal for NO at
2 5 2 5
O uptake on a sample of 10 mg of gray soot at [N O ]
1
2
-3
)
3 3
Figure 2. NO uptake on a sample of 10 mg of gray soot at [NO ] )
11
-3
(
7.0 ( 1.0) × 10 cm in the 8 mm aperture reactor. Curves a, b,
2
and d-f correspond to the raw MS signals monitored at m/e 30, 46,
62, 47, and 63, respectively. Curve c corresponds to the raw REMPI
λ
NO2 ) 511 nm (Coumarin 307, Radiant Dyes) scaled to a MS signal
at m/e 46. Curve d corresponds to the excess MS signal at m/e 30 related
signal for NO at λNO2 ) 511 nm scaled to a MS signal at m/e 46.
2
to the production of NO.
+
62(NO3)
62(NO3)
corresponding to NO2 production for a typical uptake experiment
of N2O5 on soot scaled to a MS signal amplitude and thus to a
flow rate. For both uptake experiments performed with NO3
and N2O5 the absolute NO2 concentration, [NO2]0,r, in the
absence and presence of the soot sample, respectively, has been
peak at m/e 62 (NO3 ) denoted as I0
and Ir
corre-
sponding to the MS signal before and during reaction, respec-
tively. The measured rate constant kobs for NO3 disappearance
from the gas phase is given by eq 2.
44
6
0
2(NO3)
determined by means of in situ REMPI detection of NO2 . The
I
4
6(NO2)
kobs
)
- 1 (k + kdec
esc
)
(2)
procedure used to calculate its MS signal contribution I
0,r(REMPI)
(
62(NO3)
)
Ir
at m/e 46 in the absence (subscript “0”) and presence (subscript
“
r”) of the soot sample has been reported in Appendix A2.
In the following, the subscripts 0 and r will refer to continuous
gas uptake experiments in the absence and presence, respec-
tively, of the sample. N2O5 does not have a measurable parent
If one assumes that the rate law for uptake is first order, the
uptake coefficient γobs for NO3 and N2O5 is given by γobs )
kobs/ω, where ω is the collision frequency of N2O5 or NO3 with
the soot sample. In the present data analysis γobs was calculated
using the geometric surface area As of the sample holder which
will be justified below; γobs became γss, the steady-state uptake
coefficient, once steady-state conditions were achieved after an
exposure time of 500 s or so; γobs is γ0 at t ) 0 s, that is,
immediately after lifting the plunger and thus exposing the soot
sample to N2O5. The details of the product analysis arising from
the heterogeneous reaction of N2O5 on soot may be found in
Appendix A2.
+
+
and fragment peak at m/e 108 (N2O5 ) and 62 (NO3 ),
respectively, under the present experimental conditions; the most
intense peaks are its fragment NO2 at m/e 46 followed by the
less intense fragment NO at m/e 30. The details regarding the
contribution of the HNO3 impurity to the MS amplitude at m/e
+
+
4
6 and 30 may be found in Appendix A1. In the NO3 uptake
experiments only [NO2] was detected by REMPI whereas [NO]
was evaluated at m/e 30 after successive subtraction of the
detected reaction products contributing at m/e 30 (see eq A2 in
Appendix A1).
Continuous flow uptake experiments were carried out at
ambient temperature (298 ( 2 K) under molecular flow
conditions. The concentration of NO3 inside the Knudsen reactor
2
.2. N2O5. When N2O5 is exposed to soot, it is taken up and
undergoes a heterogeneous reaction which results in a decrease
4
0
6(N2O5)
46(N2O5)
of the MS signal from I
to Ir
. Upon N2O5 uptake
11
-3
ranged between (2.7 ( 0.5) × 10 cm and (2.4 ( 0.5) ×
on both gray and black soot, we observe the formation of both
NO and NO2 in the gas phase as detected by in situ REMPI at
λ ) 452.5 and 511 nm, respectively. For NO2 REMPI detection
a Coumarin 307 (Radiant Dyes) methanol solution was used
whereas NO was multiphoton ionized using a methanol solution
of Coumarin 47 (Radiant Dyes). We have determined the rate
constant kobs for the disappearance of N2O5 following eq 3,
assuming that the rate law is first order in N2O5,
12
-3
1
0
cm . The associated [NO2] without reaction on soot
11
-3
determined by REMPI was (8.0 ( 1.0) × 10 cm for the
11
-3
lower limit of [NO3] ) (2.7 ( 0.5) × 10 cm and (6.2 (
12
-3
1
1
.5) × 10 cm for the upper limit of [NO3] ) (2.4 ( 0.5) ×
cm .
12
-3
0
3
. Results and Discussion
.1. NO3 Reaction with Gray Decane Soot: Reaction
3
4
0
6(N2O5)
I
Products. The interaction of NO3 with gray decane soot was
investigated in a series of uptake experiments performed at
different masses of soot and at different [NO3]. Figure 2 shows
a typical uptake experiment of NO3 on 10 mg of gray soot at
kobs
)
- 1 k
(3)
esc
(
46(N2O5)
)
Ir
4
0
6(N2O5)
46(N2O5)
+
where I
and Ir
are the intensities of the NO2
fragment of N2O5 before and during heterogeneous reaction,
11
-3
[NO3] ) (7.0 ( 1.0) × 10 cm . Once a steady flow of NO3
was established, the isolation plunger was lifted at t ) 550 s
and the substrate exposed to the NO3 flow. Because of the
respectively, and kesc is the measured rate constant of effusion
for N2O5 out of the flow reactor. To determine Ir
4
6(N2O5)
, the raw
uptake of NO on soot, the number of molecules effusing
3
MS signal at m/e 46 was corrected for the product NO2
through the escape orifice and thus the MS signal at m/e 62
generated during uptake of N2O5 on soot because NO2 contrib-
utes to the total MS signal Ir at m/e 46, its parent peak.
Figure 1 shows the raw REMPI signal at λ ) 511 nm
immediately decreases (curve d). At t ) 1050 s the sample
compartment is sealed by lowering the plunger and the MS
signal at m/e 62 returns to its initial value. The slight decrease
4
6
45

NO3 Free Radical and N2O5 on Decane Flame Soot
J. Phys. Chem. A, Vol. 111, No. 10, 2007 1917
2 2
Figure 3. NO uptake on a sample of 10 mg of gray soot at [NO ] )
1
2
-3
(
2.3 ( 0.5) × 10 cm (orifice diameter ) 8 mm). Curves a-c
correspond to the raw MS signals monitored at m/e 30, 46, and 47,
respectively. Curve c describes HONO production.
6
0
2(NO3)
of I
over extended periods of time such as displayed in
Figure 2 may be explained by a slight decrease of the
corresponding flow rate into the reactor. In all experiments we
observed uptake of NO2 that only occurred in the presence of
NO3, which led to a net decrease of the REMPI signal for NO2
Figure 4. (a) Uptake of NO
3
on 10 mg of gray decane soot and
1
1
resulting reaction products at [NO
3
] ) (7.0 ( 1.0) × 10 molecule
at λNO ) 511 nm (Figure 2, curve c). The major loss of NO3
2
-
3
cm from data displayed in Figure 2 that are electronically smoothed
is simple adsorption on gray and black soot. We did not perform
experiments at higher temperatures to measure the reversibility
of the reaction of NO3 with soot.
and interpolated (orifice diameter ) 8 mm). Curve a represents the
raw MS signal at m/e 62 for the NO uptake on the soot sample. Curve
3
b represents the raw MS signal at m/e 47 related to the production of
gas-phase HONO. Curve c represents the corrected MS signal at m/e
30 corresponding to the production of gas-phase NO. (b) Uptake of
NO2 in the absence of soot arises from the thermal decom-
position of N2O5 and NO3 in the NO3 source or in the flow
reactor. Reference uptake experiments with pure NO2 in the
presence of gray soot were performed at [NO2] ) (2.3 ( 0.5)
NO
3
on 10 mg of gray soot and resulting reaction products at [NO
3
] )
11
-3
(
2
7.0 ( 1.0) × 10 molecule cm (from experiment displayed in Figure
but electronically smoothed and interpolated, orifice diameter ) 8
1
2
-3
×
10 cm (Figure 3). A large and instantaneous rate of uptake
mm). Curve a represents the raw MS signal at m/e 62 for the NO
uptake on the soot sample. Curve b is the calculated MS signal at m/e
3
was observed and attained steady-state conditions after 3 min
of interaction. The initial and steady-state uptake coefficients
46
4
6, I , corresponding to N
O
2 5
formation. Curve c represents the raw
MS signal at m/e 63 for the impurity HNO uptake on soot. Positive
flow MS signal at m/e 63 indicated by the hatched area under curve c
represents the net amount of generated HNO
exc
-
2
-3
were γ0 ) (3.0 ( 0.6) × 10 and γss ) (1.3 ( 0.2) × 10 ,
3
respectively, and were similar to previous results.3
2,34
Simul-
taneously to the uptake of NO2 a large product peak of HONO
appears, which shows that the conversion of NO2 into HONO
is a fast process (Figure 3, curve c). The observed HONO yields
defined as the ratio of the integral of HONO released to the
amount of NO2 taken up during the reaction time tended toward
3
.
to the reduction of NO2 by reducing sites on gray soot according
3
4,46
to reactions 4a and 4b.
NO f NO
(4a)
NO2(ads) or NO + {C-H} f HONO + {C}ox (4b)
2
2(ads)
1
3
00 ( 10% and were independent of integration time (Figure
). No significant formation of NO was observed.
2
red
Identical behavior has been observed at low [NO2] in recent
2
2
work on the reactivity of NO2 on decane flame soot and
confirms that NO2 strongly interacts with flame soot. This is in
contrast to uptake experiments of NO2 on mineral dust, which
showed that NO2 does not significantly adsorb on dust except
The species listed in curved brackets refer to surface adsorbates.
C-H}red represents a surface site on gray soot that reduces
{
NO2 to HONO and {C}ox is the reaction product after surface
oxidation by NO2 on soot. The interaction of NO2 with an
adsorption site in reaction 4a must be weak to be sufficiently
mobile to subsequently interact with other surface sites for
reaction 4b in a Langmuir-Hinshelwood mechanism.
4
4
on Saharan dust. In that case NO2 reacted only on NO3 that
was previously adsorbed on the substrate which has also been
observed on the present soot substrate (see below).
During the uptake of NO3 on gray soot the MS signal at m/e
For all gray and black soot samples we have observed the
formation of small amounts of gas-phase N2O5. We have
calculated the yield of N2O5 from the increase of the corrected
62 (Figure 4a, curve a) partially recovered as the exposure time
increases, indicating a decrease in the rate of uptake of NO3,
presumably owing to a decrease of the net number of available
surface sites for reaction. As a consequence, we observe an
apparent reduction of the uptake coefficient γss. As displayed
in Figure 4a, a large burst of HONO at m/e 47 co-incident with
the uptake of NO3 has been observed immediately after the
exposure of the sample which reaches steady state after 500 s
46
MS signal at m/e 46, I , discussed in Appendix A2 and
exc
displayed in Figure 4b (curve b). The yield of N2O5 following
the uptake of NO3 continuously increases and reaches steady
state after 500 s.
Under our experimental conditions the formation of N2O5 may
be explained by NO2 reacting with adsorbed NO3 on the soot
substrate. It is important to note that NO2 itself reacts both on
gray and on black decane soot.
uptake for both NO3 and NO2 suggests the formation of N2O5(ads)
through the heterogeneous recombination reaction 5a.
(
Figure 4a, curve b). NO has been observed at the beginning of
the reaction and tended to small values at steady-state conditions
Figure 4a, curve c). A fast initial rate of formation of HONO
2
2,32
The observed simultaneous
(
and NO is observed immediately after exposure of soot to NO3
as displayed in Figure 4a, curves b and c. Following results
with pure NO2 interacting with the same type of soot as
displayed in Figure 3 the observed HONO formation is attributed
NO3(ads) + NO2(ads) f N O
(5a)
2
5(ads)

1918 J. Phys. Chem. A, Vol. 111, No. 10, 2007
Karagulian and Rossi
TABLE 2: Summary Yields of Reaction Products during the Heterogeneous Reaction of NO
3
on Decane Soot Samples of 19.6
2
12
-3
12
-3
cm Geometric Surface Area at [NO
3
]
0
) (2.3 ( 0.5) × 10 cm and [NO
2
]
0
) (6.0 ( 1.0) × 10 cm (Orifice Diameter ) 8
e,f
mm)
gray soot
decane)
%
yield
%
yield
%
yield
%
yield
average yieldd
g
(
5 mg
10 mg
20 mg
1
1
1
1
1
1
8
18
18
18
NO3(lost)
5(g)
(3.2 ( 0.7) × 10
(7.0 ( 1.7) × 10
(2.4 ( 0.6) × 10
(3.2 ( 0.5) × 10
(4.8 ( 0.5) × 10
(1.6 ( 0.2) × 10
(3.3 ( 0.7) × 10
(8.5 ( 1.4) × 10
(2.5 ( 0.5) × 10
(3.3 ( 0.7) × 10
(5.0 ( 0.3) × 10
(1.6 ( 0.5) × 10
(3.3 ( 0.7) × 10
(8.4 ( 1.2) × 10
(3.7 ( 0.7) × 10
(4.6 ( 1.0) × 10
(6.6 ( 0.6) × 10
(2.5 ( 0.8) × 10
(3.3 ( 0.7) × 10
(8.5 ( 1.4) × 10
(3.1 ( 0.6) × 10
(4.0 ( 0.9) × 10
(5.8 ( 1.0) × 10
(2.0 ( 0.7) × 10
7 a
8
17 a
18
17 a
18
17 a
18
N
2
O
22
25
22
23 ( 7
NO2(lost)
NO(g)
HNO3(g)
HONO(g)
7 a
6 a
8 b
17 a
16 a
18 b
17 a
16 a
18 b
17 a
16 a
18 b
10
1.5
94
10
1.5
98
13
2
87
11 ( 5
1.8 ( 0.5
96 ( 8
black soot
%
yield
%
yield
%
yield
%
yield
average yieldd
g
(decane)
5 mg
10 mg
20 mg
1
8
18
18
18
NO3(lost)
5(g)
(3.2 ( 0.5) × 10
(1.0 ( 0.6) × 10
(2.8 ( 0.4) × 10
(4.5 ( 0.7) × 10
(4.8 ( 0.5) × 10
(3.4 ( 0.8) × 10
(7.2 ( 1.2) × 10
(3.0 ( 0.6) × 10
(5.1 ( 0.4) × 10
(6.8 ( 0.7) × 10
-
(3.5 ( 0.5) × 10
(7.0 ( 1.0) × 10
(3.0 ( 0.8) × 10
(6.3 ( 0.5) × 10
(5.2 ( 0.5) × 10
-
(3.5 ( 0.5) × 10
(7.0 ( 1.0) × 10
(3.0 ( 0.6) × 10
(5.7 ( 0.5) × 10
(6.0 ( 0.7) × 10
-
1
1
1
1
8 a
8
17 a
18
17 a
18
17 a
18
N
2
O
31
20
20
24 ( 7
NO2(lost)
NO(g)
HNO3(g)
HONO(g)
7 a
6 a
17 a
16 a
17 a
16 a
17 a
16 a
14
1.5
15
2
18
1.5
16 ( 6
1.6 ( 0.4
c
-
a
taken up on soot during a reaction time of 500 s. b Yield
Yield of NO
y
given as percentage with respect to the total number of molecules of NO
3
c
of HONO given as a percentage with respect to the total number of molecules of NO2(lost) - N
2
O5(g) taken up during a reaction time of 500 s. Dash
d
(
-) indicates that no reaction product has been observed. The average yield is calculated on the basis of the results obtained for 10 and 20 mg
e
f
g
of soot corresponding to a coherent soot layer. The average uncertainty of the yields is (20%. Uncertainty includes systematic error. Includes
a potential systematic uncertainty for N and NO owing to successive subtraction (see eqs A1 and A2).
2
O
5
The conversion of NO3 to N2O5 occurs via a Langmuir-
Hinshelwood mechanism where NO3 and NO2 first adsorb onto
the soot surface and subsequently react forming N2O5 which in
part desorbs from the surface.
Once N2O5 has been formed in the adsorbed state, a fraction
may desorb into the gas phase, and another may subsequently
react, as will be discussed below in section 3.3.
point out that the experimental uncertainties for N2O5 and NO
reported in the last column of Table 2 also include potential
systematic errors owing to the evaluation of the product yields
by successive subtraction (see eqs A1 and A2). The mass
balance reveals that 16-30% of NO2 taken up on gray soot
has been converted into gas-phase N2O5 according to reactions
5a and 5b under the assumption of excess adsorbed NO3 that is
converted only in moderate yields to gas-phase NO (see below).
The remaining 70-84% of reacted NO2 equals the yield of
HONO produced due to reaction 4b according to the NO2 mass
balance, namely N(N2O5) + N(HONO) ) N(NO2), where N is
the number of molecules taken up after a given reaction time.
The essentially 100% yield of HONO resulting from pure NO2/
N O
f N O
5(g)
(5b)
2
5(ads)
2
HNO3, present as an impurity, reacts on the soot substrate as
well. This reaction has been studied in recent laboratory work
using the same experimental apparatus.23 As displayed in Figure
4
b, a significant initial uptake of HNO3 was observed at m/e
22
gray soot experiments, in agreement with previous results,
6
3 (curve c). After a reaction time of 170 s, an excess in the
leads to the amount of generated N2O5 calculated from the
amount of NO2 taken up minus the yield of N2O5. We may
conclude that NO3 reacting on gray soot does not generate
HONO because the amount of HONO may be entirely accounted
for by the loss of NO2. Therefore, the slow production of NO
must be attributed to the reaction of NO3 with gray soot.
MS signal at m/e 63 corresponding to formation rather than loss
of HNO3 is observed, as indicated by the hatched area in Figure
4
b, curve c. The increasing rate of formation of HNO3 in the
presence of soot coincides with the trend of N2O5 formation
displayed in Figure 4b, curve b and shows that part of N2O5(ads)
formed in reaction 5a undergoes hydrolysis on the surface of
soot. HNO3 is then released back into the gas phase after
heterogeneous hydrolysis, reaction 6.
Previous work has shown that HONO is the only gas-phase
22,34
product from the reaction of pure NO2 with gray soot
in
contrast to NO formation resulting from the NO2 reaction on
3
2
amorphous carbon or black soot (see below). Therefore, we
attribute the slow formation of NO exclusively to the hetero-
geneous reaction of NO3 on gray soot according to reaction 7
2
N O
+ H O f 2HNO3(g)
(6)
2
5(ads)
2
(ads)
The yield of HNO3 observed in uptake experiments performed
12
-3
knowing that NO does not lead to NO on gray soot.
at [NO3] ) (2.5 ( 0.5) × 10 cm on different amounts of
gray and black soot is only 1.5-2% of the total number of NO3
molecules taken up (see Table 2).
NO + soot f NO + products
(7)
3
An increase of gas-phase HNO3 formation in the presence
of soot aerosol has been observed in recent experimental work
at very low humidities in which NO2, HNO3, and NO3/N2O5
The amount of NO released into the gas phase has been
30
calculated from the excess MS signal at m/e 30, I according
exc
1
1
reacted on soot particles in a large aerosol chamber. In the
same study the reaction probability for reaction 6 was assumed
to eq A2. The yield of NO did not show a significant variation
with [NO3] and the amount of soot. Table 2 displays the fact
that NO corresponds to nearly 6-16% of NO3 taken up on the
substrate. In addition, NO is nonreactive toward fresh unexposed
-
5
to be time independent and resulted in γ ) (4.0 ( 2.0) × 10 .
Apparently there exists sufficient adsorbed water even at low
relative humidity to enable reaction 6 such that H2O is not the
limiting reagent.
2
2,32
soot samples,
a result obtained in ancillary experiments for
the present soot samples.
In Table 2 we report the product yields for the NO3 reaction
on three different masses of gray soot at [NO3] ) (2.5 ( 0.5)
To explain the temporal trend of the observed product yields
on gray soot displayed in Figure 4, we need to analyze the time
dependence of reactants and products at the beginning (between
12
12
-3
×
10 and an excess of NO2 of (6.0 ( 1.0) × 10 cm . We

NO3 Free Radical and N2O5 on Decane Flame Soot
J. Phys. Chem. A, Vol. 111, No. 10, 2007 1919
3
Figure 6. Uptake coefficient γ vs time for NO on 10 mg of black (a)
11
-3
and gray (b) soot. [NO
) 8 mm).
3
] ) (7.0 ( 1.0) × 10 cm (orifice diameter
2 2
Figure 5. NO uptake on a sample of 10 mg of black soot at [NO ] )
1
2
-3
(
2.3 ( 0.5) × 10 cm (orifice diameter ) 8 mm). Curves a-c
correspond to the raw MS signals monitored at m/e 30, 46 and 47,
respectively. Curve d (broken line) reports the NO formation.
t ) 62 and 210 s) and at t > 210 s of the uptake process,
respectively. At first the fast disappearance of gas-phase NO2
leads to fast HONO production (curve b in Figure 4a, reaction
4b). Concurrently, the fast uptake of NO3 leads to rapid
decomposition on the substrate resulting in a significant rate of
formation of NO (curve c in Figure 4a, reaction 7). Therefore,
owing to competitive reactions 4b and 7, only a fraction of
NO3(ads) reacts with NO2 and recombines to N2O5 according to
reaction 5a whereas all of the NO2 reacts either to HONO or to
N2O5. At t > 210 s the number of adsorbed NO3 molecules
increases, thereby leading to partial saturation of the NO3 uptake
and thus inhibiting the turnover of available surface sites where
NO3 decomposition can take place. Therefore, reactions 5a and
3
Figure 7. NO rate of formation vs time for NO on 10 mg of (a)
11
-3
black and (b) gray soot. [NO
diameter ) 8 mm).
3
] ) (7.0 ( 1.0) × 10 cm (orifice
observed in recent work on the reactivity of NO2 on black
22
decane flame soot.
5b become predominant with respect to reactions 4b and 7 at
Figure 6 shows the temporal variation of γ for NO3 on a 10
mg sample of black and gray soot at [NO3] ) (7.0 ( 1.0) ×
later reaction times. On the other hand, HONO production is
ongoing in the steady state by continuous NO2 reaction on gray
soot, in contrast to NO production according to reaction 7.
We therefore conclude that according to Table 2 ap-
proximately 35 ( 10% of NO3, which reacted on gray soot has
been converted to N2O5 and NO. The remaining amount of NO3,
that is, 65 ( 10%, is irreversibly taken up on the substrate.
However, no saturation of the uptake of NO3 has been observed
under the current experimental conditions despite the vanishing
NO production at late reaction times, as displayed in Figure
11
-3
1
0
cm in the presence of NO2. The value of γ0 for black
soot is twice that observed for gray soot, as will be discussed
below. At steady-state conditions the situation is reversed, where
γss for black soot is smaller by a factor of 2 with respect to γss
for gray soot. However, the number of NO3 molecules taken
up on both substrates during the exposure time of 500 s is
approximately the same. As already observed for gray soot,
simultaneous uptake of both NO3 and NO2 on black soot leads
to the formation of small amounts of gas-phase N2O5 according
to reactions 5a and 5b.
4
a.
3
.2. NO3 Reaction with Black Decane Soot. The reaction
Figure 7 displays the NO product flow generated during
uptake of NO3 on 10 mg of black and gray decane soot at the
same experimental conditions. NO2 effusing from the hot NO3
source reacts with black soot to produce mainly NO, as observed
of NO3 on black soot was examined in the same manner as
gray soot discussed above. A fast initial uptake coefficient γ0 was
observed for all experiments performed within [NO3], which
1
1
12
ranged between (2.7 ( 0.5) × 10 and (2.3 ( 0.5) × 10
cm , as reported in Table 2.
Ancillary uptake experiments with pure NO2 in the presence
32
in the interaction of NO2 with amorphous carbon or black
-
3
22
decane soot. NO2 presumably interacts with black soot,
resulting in the formation of HONO akin to the reaction on
gray soot, which to a large extent decomposes into NO according
to the following reaction mechanism proposed by Stadler and
1
2
of black soot were performed at [NO2] ) (2.3 ( 0.5) × 10
-3
cm . As observed in the companion experiment with gray soot,
the observed uptake of NO2 attained steady-state conditions after
22
Rossi:
3
min of interaction. The initial and steady-state uptake
-
2
coefficient was γ0 ) (1.5 ( 1.0) × 10 and γss ) (2.7 ( 0.6)
NO + {C - H} f HONO(ads) + {C}ox
(4b)
(8)
2
red
-
3
22,32
×
10 , respectively, similar to previous results. Simulta-
neous to the uptake of NO2 a significant quantity of product
NO appears. The amount of product NO has been calculated
2HONO(ads) f NO + NO2(ads) + (H O)
2 ads
3
0
from the excess MS signal at m/e 30, I , according to eq A2.
The difference in the NO yields between black and gray
decane flame soot following reaction with NO3, curves a and b
in Figure 7 is related to the NO formation from HONO
decomposition, reactions 4b and 8. NO from HONO decom-
position, reaction 8, may therefore occur in addition to the NO
formation from heterogeneous decomposition of NO3 on black
soot according to reaction 7 in contrast to gray soot where
exc
This signal has been converted to a flow rate using a calibration
factor for pure NO and subsequently integrated over the reaction
time. The observed NO yields defined as the ratio of the amount
of NO released to the amount of NO2 taken up during the
reaction time tended toward 40% in contrast to the vanishing
yield of HONO (Figure 5). Similar behavior at low [NO2] was

1920 J. Phys. Chem. A, Vol. 111, No. 10, 2007
Karagulian and Rossi
TABLE 3: Summary of NO Yield Resulting from the
it was kept at ambient temperature for N2O5. Figures 1 and 8
display typical uptake experiments of N2O5 on gray and black
soot, respectively. Both NO and NO2 were detected as the only
reaction products resulting from the interaction of N2O5 with
soot. Figure 1 shows that the NO2 concentration measured using
REMPI detection is zero within experimental uncertainty before
exposure of N2O5 to soot at t ) 350 s, which is proof for the
absence of NO2 in the N2O5 flow emanating from the source.
The same is true for the NO displayed in Figure 8 where REMPI
detection of NO was employed. Owing to the dye change for
the REMPI detection of NO and NO2, simultaneous REMPI
detection of both reaction products was not possible such that
separate uptake experiments using identical [NO3] had to be
undertaken, one set for NO and another for NO2 detection.
However, analysis of the MS signal amplitudes at m/e 30 and
Uptake of NO
3
on 10 mg of Gray and Black Soot at
Different [NO
3
] (Orifice Diameter ) 8 mm)
%
%
, cm-
3
NO (gray soot)a yield NO (black soot)a yield
[
NO
3
]
0
1
1
17
17
17
17
17
17
17
17
17
17
(
(
(
(
(
2.7 ( 0.5) × 10 (1.6 ( 0.4) × 10
13 (2.0 ( 0.5) × 10
11 (2.6 ( 0.7) × 10
20
19
16
14
15
1
1
3.8 ( 1.8) × 10 (2.0 ( 0.6) × 10
1
1
7.0 ( 1.0) × 10 (2.6 ( 0.5) × 10
9
8
(4.3 ( 0.4) × 10
(5.8 ( 0.8) × 10
1
2
1.5 ( 0.5) × 10 (3.0 ( 0.7) × 10
1
2
2.3 ( 0.5) × 10 (3.3 ( 0.4) × 10
10 (5.1 ( 0.7) × 10
a
Yield of NO given as a percentage with respect to the total number
taken up during the reaction time of 500 s.
of molecules of NO
3
HONO is kinetically stable and where NO exclusively stems
from reaction 7. Therefore, contributions of NO formation from
HONO and NO3 decomposition cannot be experimentally
separated for black in contrast to gray decane soot. Table 4
reveals a NO yield higher by approximately 50% for black
compared to gray decane soot that is consistent with the HONO
decomposition leading to additional NO in the case of black
decane soot.
4
6 after correction for the contribution of N2O5 and NO2
detected by REMPI following the procedure outlined in Ap-
30
pendix A2 based on I , revealed the presence of NO as a
exc
reaction product (see eq A6). Conversely, the MS signal
amplitudes at m/e 30 and 46 revealed the presence of NO2 as a
reaction product once the contribution of N2O5 and NO whose
absolute concentration was measured using REMPI were
accounted for (see eq A5). Qualitatively one observes an
increase of both NO and NO2 during the first 100 s or so after
exposing the soot to the N2O5 flow that is roughly co-incident
with the establishment of steady-state conditions of the MS
signal at m/e 46. This indicates that a certain amount of N2O5
must first be adsorbed onto soot before the rate of product
desorption becomes significant. We also note that unlike the
situation for NO3 the differences between gray and black soot
are minimal.
In recent work performed by Stadler,22 black decane soot was
exposed to a flow of pure HONO to decide whether or not
HONO produced on black soot may irreversibly adsorb on the
surface or decompose to NO. In that work fast uptake of HONO
and production of NO was observed according to reaction 8.
The total NO product yield at limiting high concentration was
5
0% with respect to HONO taken up, whereas for [HONO] )
11
-3
(3.7 ( 0.4) × 10 cm the total product yield consisted of
4
0% NO and 10% NO2. The open balance of nitrogen was
attributed to a reservoir of HONO or its reaction product
adsorbed on the soot substrate. The NO rate of formation
resulting from NO3 decomposition on gray soot tends to zero
under steady-state conditions, as shown in Figure 7, whereas
the rate of NO formation on black soot is sustained owing to
continuous HONO decomposition following reaction 8. Table
Table 4 displays the quantitative aspects in terms of the yields
of NO and NO2 using REMPI detection referenced to the
number of N2O5 molecules taken up within a given time, usually
500 s. The corresponding twin product not directly detected by
REMPI, for instance NO in the case of REMPI detection of
NO2, was evaluated as described above from the MS spectra.
2
reveals that the sum of the yields of NO and N2O5 directly
attributable to the reaction of NO3 on black soot amounts to 40
10% based on NO3 taken up, a value that is quite comparable
to that on gray soot when the contribution of HNO3 is ignored
12
-3
(
At low concentrations up to [N2O5] ) 2.0 × 10 molecule cm
the yields of both NO and NO2 are identical within experimental
uncertainty and equal to between 80 and 100%. Beyond this
threshold value of [N2O5] the NO2 yield detected by REMPI
and the NO yield inferred from the MS spectra seem to decrease
significantly. This decrease by roughly a factor of 2 seems to
depend primarily on the concentration and not on the dose of
N2O5 taken up by soot, as revealed by a comparison between
lines 3 and 4 for gray soot, and between lines 5 and 6 for black
soot of Table 4. The subtle difference between the effect of
concentration and dose undoubtedly points toward a complex
mechanism for the heterogeneous interaction of N2O5 with soot.
A similar situation prevails for the interaction of N2O5 with ice
at low temperatures. We conclude that both the NO and NO2
yields tend toward 100% for the reaction of N2O5 on soot under
atmospheric conditions, that is, for small N2O5 concentrations.
(
1
1.8% yield). The remainder of adsorbed NO3, namely 60 (
0% must stay adsorbed on the black soot substrate.
Table 3 reports the NO yield resulting from the uptake of
NO3 on 10 mg of gray and black decane soot at [NO3] between
1
1
12
-3
(
2.7 ( 0.5) × 10 and (2.3 ( 0.5) × 10 cm . The yield of
NO for black soot over an integration period of 500 s was 5-8%
larger than for gray soot. There is no significant dependence of
the NO yields on [NO3]. The NO yield expressed as a percentage
with respect to the total number of molecules of NO3 taken up
during the same reaction time resulted in a value of 10.2 (
1
.9% for gray and 16.8 ( 2.6% for black soot when averaged
over uptake experiments at different values of [NO3]. We
attribute the larger value of the NO yield found for black
compared to gray soot to HONO decomposition on black soot.
The small difference in the NO yields reported in Table 3
suggests that most of the putative HONO remains adsorbed on
black soot rather than decompose to form NO. In agreement
with this conclusion, the HONO rate of formation is negligible
for the interaction of NO2 with black soot, as displayed in Figure
In addition, neither HNO nor HONO was observed in any
3
of the uptake experiments on both gray and black soot. The
absence of significant amounts of HNO is somewhat of a
3
surprise because N O is known for its efficient hydrolysis in
2
5
the presence of small amounts of adsorbed water. On Norit A,
a commercial carbon black product that holds important
quantities of adsorbed water, the reaction of N2O5 generated a
5
, curve c.
3.3. N2O5 Reaction with Decane Soot: Reaction Products.
47
As far as the experimental arrangement is concerned, the only
difference between the NO3 and N2O5 uptake experiments was
that the source temperature was raised to 530 K for NO3 whereas
large yield of HNO3 that was generated through heterogeneous
hydrolysis of N2O5 in addition to the reduction of N2O5 resulting
in NO. The equimolar formation of NO and NO2 in the uptake

NO3 Free Radical and N2O5 on Decane Flame Soot
J. Phys. Chem. A, Vol. 111, No. 10, 2007 1921
TABLE 4: Yield of NO
2
and NO Given as Percentage with Respect to the Total Number of Molecules of N
2
O
5
Taken up on
2
a
Gray and Black Soot during a Reaction Time of 500 s (19.6 cm Geometric Surface Area)
%
yield
%
yield
[N O
2 5
]
0
, cm-
3
N
2
O5(lost) (gray soot)
NO2(g) (gray soot)
NO(g) (gray soot)
1
1
1
1
1
2 b
2 b
2 b
2 b
2 c
17
17
17
(1.0 ( 0.5) × 10
(1.6 ( 0.5) × 10
(2.0 ( 0.7) × 10
(2.7 ( 1.0) × 10
(8.0 ( 1.3) × 10
(6.0 ( 1.5) × 10
(8.9 ( 1.3) × 10
(9.8 ( 1.1) × 10
(1.0 ( 0.2) × 10
(2.0 ( 0.3) × 10
(4.8 ( 1.2) × 10
80
86
94
45
33
(4.3 ( 1.3) × 10
(8.0 ( 1.1) × 10
(8.6 ( 0.9) × 10
(3.8 ( 0.7) × 10
(7.0 ( 1.0) × 10
72
90
88
38
35
17
17
18
18
17
17
17
17
17
(7.7 ( 1.6) × 10
17
(9.3 ( 1.5) × 10
17
(4.5 ( 0.2) × 10
17
(7.0 ( 1.2) × 10
%
yield
%
yield
[N
2
O
5
]
0
, cm-
3
2
N O5(lost) (black soot)
NO2(g) (black soot)
NO(g) (black soot)
1
1
1
1
1
1
1
2 b
2 b
2 b
2 b
2 b
2 b
2 c
17
17
18
17
18
18
18
17
17
(1.0 ( 0.5) × 10
(1.3 ( 0.7) × 10
(1.5 ( 0.5) × 10
(1.9 ( 0.8) × 10
(2.0 ( 0.8) × 10
(2.9 ( 1.0) × 10
(8.0 ( 1.3) × 10
(5.3 ( 1.4) × 10
(9.5 ( 1.5) × 10
(1.0 ( 0.7) × 10
(9.7 ( 1.2) × 10
(1.0 ( 0.2) × 10
(1.0 ( 0.3) × 10
(1.7 ( 0.5) × 10
(5.2 ( 1.3) × 10
98
87
95
90
92
45
39
(5.2 ( 1.5) × 10
100
81
100
84
87
50
17
17
(8.2 ( 1.3) × 10
(7.7 ( 1.2) × 10
17
17
(9.5 ( 1.9) × 10
(9.9 ( 0.5) × 10
17
17
(8.7 ( 1.0) × 10
(8.2 ( 1.0) × 10
17
17
(9.6 ( 1.1) × 10
(8.7 ( 1.7) × 10
17
17
(4.2 ( 1.2) × 10
(5.0 ( 1.5) × 10
17
17
(6.6 ( 1.1) × 10
(7.1 ( 2.1) × 10
42
a
The data for NO and NO
2
given in straight font correspond to yields obtained by REMPI detection whereas data listed in italics have been
obtained using the analysis of MS signals at m/e 30 and 46 (see Appendix A2). Orifice diameter ) 8 mm. c Orifice diameter ) 4 mm.
b
Figure 9. Uptake coefficient γss of NO
3
based on the geometric sample
] (orifice diameter ) 8 mm): NO on
surface as a function of [NO
3
3
2 5 2 5
Figure 8. N O uptake on a sample of 15 mg of black soot at [N O ]
black (triangles) and gray soot (open circles).
1
2
-3
)
(2.0 ( 0.7) × 10 cm and at an orifice diameter of 8 mm. Curves
a and b correspond to the raw MS signals monitored at m/e 46 and 30,
respectively. Curve c corresponds to the raw REMPI signal for NO
detection at λNO ) 452.6 nm (Coumarin 47, Radiant Dyes) scaled to a
MS signal at m/e 30.
3
TABLE 5: Summary of Uptake Experiments of NO on 10
a
mg of Gray and Black Decane Soot
, cm-
3
γ
0
(gray soot)
γss (gray soot)
3
[NO ]
0
1
1
1
1
1
1
1
1
2
2
(
(
(
(
(
2.7 ( 0.5) × 10
3.8 ( 1.8) × 10
7.0 ( 1.0) × 10
1.5 ( 0.5) × 10
1.8 ( 0.5) × 10
0.44 ( 0.05
0.35 ( 0.06
0.15 ( 0.07
0.12 ( 0.05
0.12 ( 0.05
0.12 ( 0.04
0.2 ( 0.02
0.15 ( 0.03
of N2O5 on both gray and black soot suggests the occurrence
of reactions 9 and 9a.
0.11 ( 0.02
-2
-2
-2
(5.7 ( 1.3) × 10
(6.5 ( 1.0) × 10
(7.2 ( 1.5) × 10
N O
+ {C} f N O
+ {C ‚ O2}
(9)
12
2
5(ads)
2
3(ads)
(2.3 ( 0.5) × 10
3
[
NO
3
]
0
, cm-
γ
0
(black soot)
γss (black soot)
N O
f NO + NO2
(9a)
2
3(ads)
1
1
1
1
1
1
1
2
(
(
(
(
(
2.7 ( 0.5) × 10
5.5 ( 1.0) × 10
8.0 ( 0.5) × 10
1.4 ( 0.5) × 10
2.3 ( 0.5) × 10
0.5 ( 0.07
0.2 ( 0.03
0.17 ( 0.02
(6.7 ( 1.4) × 10
(6.5 ( 1.5) × 10
(4.8 ( 1.0) × 10
0.35 ( 0.08
0.3 ( 0.04
0.2 ( 0.05
0.18 ( 0.03
This mode of decomposition is similar to NO3 decomposition
according to eq 7 in that two O atoms are transferred to the
soot substrate after reaction corresponding to a reduction
involving the transfer of two electrons per N atom in the case
of N2O5. In case the reduction of N2O5 on soot occurs via NO2
and subsequent secondary reduction of NO2 to NO akin to NO2
reduction on amorphous carbon32 it would be highly improbable
to observe equimolar amounts of NO and NO2.
-2
-2
-2
12
a
Initial (γ
0
) and steady-state (γss) uptake coefficients (orifice diameter
8 mm, 19.6 cm geometric surface area).
2
)
between (7.0 ( 1.0) × 10 cm-3 and (2.3 ( 0.5) × 10 cm
1
1
12
-3
-
2
result in a constant value of γss ) (6.5 ( 1.5) × 10 ,
independent of [NO3]. In addition, there is no significant
difference in γss between gray and black soot.
3
.4. Uptake Kinetics of NO3 on Decane Soot. The interac-
tion of NO3 with gray and black decane soot shows that after
an initial fast uptake of NO3 there is no saturation of the sample
at steady-state conditions. Figure 9 and Table 5 display the
values of γss as a function of [NO3] for gray and black soot
based on the geometric sample surface of 19.6 cm . The
uncertainties for NO3 uptake experiments were determined from
the signal-to-noise ratio of the MS signal at m/e 62.
From this series of measurements it is evident that γss follows
a rate law pseudo first order in NO3 for [NO3] > (7.0 ( 1.0) ×
2
11
-3
11
-3
10 cm . Conversely, for [NO3] < (7.0 ( 1.0) × 10 cm
the inverse dependence of γss on [NO3] suggests that the
mechanism of NO3 uptake is complex and does not correspond
to a simple first-order rate law for uptake. A similar behavior
has been observed for the interaction of NO3 with mineral dust
We note that for a variation of [NO3] between (2.7 ( 0.5) ×
11
11
-3
1
0
0
and (7.0 ( 1.0) × 10 cm γss decreases from 0.2 (
4
4
.03 to 0.11 ( 0.01 for both types of soot. Variations of [NO3]
substrates such as Kaolinite. Akin to γss , γ0 for black and

1922 J. Phys. Chem. A, Vol. 111, No. 10, 2007
Karagulian and Rossi
γBET ) 1.2 × 10 and γBET ) 3.7 × 10-8 for gray and black
-7
soot, respectively.
The amount of adsorbed NO3 during an uptake experiment
on gray and black soot turned out to be the same. At [NO3] )
1
2
-3
18
(
2.3 ( 0.5) × 10 cm (3.3 ( 0.7) × 10 molecules of NO3
were adsorbed on 10 mg of soot during a reaction time of 500
s. During this time the samples did not saturate. As discussed
in recent work on the interaction of NO3 on mineral dust,44 NO3
may be represented by a 4.5 Å diameter sphere with a projected
-
15
2
surface area of 1.6 × 10 cm /molecule, which leads to a full
1
4
-2
surface coverage of 6.3 × 10 molecules cm . The 10 mg
3
2
Figure 10. Uptake coefficient of NO
geometric sample surface: dependence of the initial (open circles) and
steady-state (full circles) uptake coefficient on sample mass at [NO
3
on gray soot based on the
sample of gray soot has a total surface area of 6.9 × 10 cm
2
based on a BET surface area of 69 m /g (see Table 1). This
3
]
16
19
leads to 1.3 × 10 and 3.5 × 10 NO3 forming a monolayer
on 10 mg of soot based on the geometric and BET surface area,
respectively. Therefore, the total number of NO3 taken up on
1
2
-3
)
(2.3 ( 0.5) × 10 cm (orifice diameter ) 8 mm).
gray soot seems to follow the same trend as a function of [NO3]
within experimental uncertainty.
1
8
19
the substrate during 500 s has consumed 3.3 × 10 /3.5 × 10
)
9.4% of reactive sites based on the BET surface area. On
Figure 10 displays the steady-state uptake coefficient γss of
NO3 as a function of the mass of gray and black soot (Table
the basis of this estimate, we may expect to observe increasing
levels of saturation of NO3 uptake on soot beyond 50% coverage
6
). A linear dependence of γss as a function of the substrate
12
-3
at 2500 s of exposure at [NO3] ) (2.3 ( 0.5) × 10 cm .
mass is clearly visible for soot masses up to 7 mg. The identical
situation holds for black soot (not shown). We note that for the
interaction pair NO3-soot there is no mass dependence beyond
the mass of 7 mg, which we interpret as the minimum mass
required for forming a coherent monolayer of soot. The
alternative interpretation of NO3 penetrating into deeper layers
of the sample is less applicable owing to the large absolute
values of γ. This value of 7 mg corresponds to a soot loading
3
.5. Uptake Kinetics of N2O5 on Decane Soot. Figure 11
and Table 7 display the kinetic results of the heterogeneous
reaction of N2O5 on gray and black soot in terms of γ0 and γss
based on the geometric sample surface. We note from Figure
1
1 that both γ0 and γss decrease with increasing [N2O5] down
-
3
to a constant value of 5 to 10 × 10 that is independent of
[N2O5] for both γ0 and γss. The concentration dependence of
γ0 and γss shows the same trend as for NO3 albeit with two
differences: for one the threshold value for [N2O5] below which
the uptake coefficient increases with decreasing [N2O5] in an
inverse γ/concentration dependence is higher than for NO3. This
threshold value for black soot seems to lie around [N2O5] ) 3
2
of 7/19.6 ) 0.37 mg/cm and is identical to that for black soot.
21
In earlier work a value for the threshold loading of 0.41 (
2
0
.1 mg/cm has been determined. Soot is a porous material,
and on the time scale of a few seconds given by the orifice
diameter of the flow reactor, we do not expect NO3 to explore
the internal surface of the pores as given for instance by the
BET surface area. We do not think that the surface porosity of
soot plays an important role in the determination of the uptake
coefficient in view of the large value of γss of NO3 on gray and
black soot. In contrast, for the reaction of N2O5 on soot we
may have to take into account possible diffusion of N2O5 into
deeper sample layers due to smaller values of γss with respect
to NO3.
1
2
-3
11
×
10 molecule cm , whereas it lies around (7-8) × 10
-3
molecule cm for NO3 according to Figure 9. The situation is
less clear for gray soot, which seems to follow an inverse
γ/concentration relationship throughout the examined concentra-
tion range where the values of γ0 and γss coincide at the highest
examined [N2O5] of 8 × 10 molecule cm . The second
difference with NO3 is the fact that the γ dependence on [N2O5]
for gray and black soot is different from that with N2O5 whereas
it is identical for both types of soot in the case of NO3 as
displayed in Figure 9. In any case, the concentration independent
γ value for N2O5 beyond the threshold value is a factor of 10
lower than for the corresponding γ value for NO3. As for NO3
we report the calculated value of γ using the BET surface area
thus regarding it as a lower bound. The steady-state uptake
1
2
-3
We therefore take the geometric surface area for the evalu-
ation of the gas-surface collision frequency ω under the
constraint of the present experimental conditions of low gas-
phase residence times for NO3. The linear mass dependence of
γ0 and γss observed below 7 mg for gray and black soot and
displayed in Figure 10 is interpreted that no coherent layer of
soot on the sample holder has been formed at those sample
masses yet. Beyond 7 mg of soot the initial uptake coefficient
γ0 is very large and stays constant at approximately 0.2 such
that NO3 does not have sufficient time to explore the internal
surface area of the substrate given by the BET surface.
Therefore, as already discussed for N2O5 on mineral dust
coefficient for N2O5 reacting on 20 mg of soot is γBET ) 3.2 ×
-8
1
0
for gray and black soot (see Table 7).
4
. Conclusions and Atmospheric Significance
The reaction products resulting from the heterogeneous
reaction of N2O5 and NO3 with soot support the reduction of
both trace gases according to reactions 9, 9a, and 7. N2O5 yields
an equimolar amount of NO + NO2 whereas NO3 exclusively
leads to NO independent of the type of decane soot. The nature
of the reaction products reveals a heterogeneous renoxification
reaction following a mechanism that effectively converts NOy
to NOx. The science community is looking for renoxification
pathways other than the photolysis of HNO3 to adjust the
[HNO3]/[NOx] concentration ratio downward, as all models
48
substrates, we do not think that the pore diffusion model should
be applied to correct γ0 or γss. At steady-state conditions we
note that γss follows the same trend as γ0 as a function of mass
of the soot sample. The correction to γss afforded by the pore
diffusion theory, if feasible, may serve as a lower limit for γ
whereas γss based on the geometrical surface area may be
regarded as an upper limit to the true value of γ. In addition, it
is noteworthy that the pore diffusion theory has never been
successfully applied to soot substrates in view of their structural
complexity. If we use the BET surface area to provide a lower
bound for γss of NO3 reacting on 20 mg of soot, we obtain
4
9
substantially overpredict HNO3. The present work does not
lay any claim to the importance of these renoxification pathways
but does emphasize its potential importance in certain situations.

NO3 Free Radical and N2O5 on Decane Flame Soot
J. Phys. Chem. A, Vol. 111, No. 10, 2007 1923
on Gray and Black Soot as a Function of Sample Mass ([NO ] ) (2.3 (
ss (gray soot) (black soot) ss (black soot)
TABLE 6: Summary of Uptake Experiments with NO
3
3
1
2
-3
0
.5) × 10 cm , Orifice Diameter ) 8 mm)
mass (mg)
γ
0
(gray soot)
γ
γ
0
γ
-
2
-2
-2
-2
-2
-2
-2
1
3
5
7
0
5
0
.5
(7.0 ( 4.0) × 10
(1.8 ( 1.2) × 10
(3.0 ( 1.7) × 10
(5.5 ( 1.6) × 10
(6.5 ( 1.6) × 10
(7.2 ( 1.2) × 10
(8.0 ( 2.0) × 10
(8.5 ( 2.0) × 10
(5.0 ( 3.5) × 10
(8.0 ( 3.0) × 10
0.15 ( 0.035
(1.5 ( 1.3) × 10
(2.5 ( 1.3) × 10
(5.0 ( 1.5) × 10
-
-
-
-
-
-
2
2
2
2
2
2
0.1 ( 0.05
0.16 ( 0.035
0.17 ( 0.035
0.12 ( 0.030
0.2 ( 0.032
0.2 ( 0.032
-2
-2
-2
1
1
2
0.18 ( 0.03
0.2 ( 0.03
0.2 ( 0.03
(6.7 ( 1.4) × 10
(7.5 ( 2.3) × 10
(8.4 ( 2.0) × 10
NO3 are identical on both types of soot. Somewhat surprisingly,
only a very low rate of hydrolysis of N2O5 to HNO3 is observed
on both types of soot, which is at variance with results on N2O5
reacting on carbon black, Norit A, a carbonaceous material
4
7
produced from pyrolysis. For both trace gases significant
amounts remain adsorbed on the soot substrate. Only 11 and
1
6% of the lost NO3 are converted to NO on gray and black
decane soot, respectively. We attribute the difference of 5%
excess” yield of NO on black vs gray soot to HONO
“
decomposition (see above). Considering the fraction of NO3 that
goes on to yield N2O5 we see that approximately two-thirds of
the lost NO3 remains adsorbed on soot during uptake vs 1/3
generating gaseous N2O5. In contrast, we observe an essentially
quantitative conversion of N2O5 to an equimolar mixture of NO
+ NO2 for concentrations lower than 3 × 10 molecules cm ,
which means that no significant amounts of N2O5 remain
adsorbed on both types of soot, as indicated in Table 4. At higher
concentrations of N2O5 the rate of surface reaction according
to reactions 9 and 9a obviously slows down compared to N2O5
adsorption which leads to a significant reduction of the product
yield.
Figure 11. Uptake coefficient γss of N based on the geometric
2
O
5
sample surface as a function of [N ]: N O on black (circles) and
2
O
5
2 5
gray soot (open squares). For all measurements we used the 8 mm
orifice diameter except for the point marked by the arrow where a 4
mm orifice diameter has been used.
12
-3
TABLE 7: Summary of Results for Uptake Experiments of
on 10 mg of Gray and Black Decane Soota
2 5
N O
3
[N O
2 5
]
0
, cm-
γ
0
(gray soot)
γss (gray soot)
1
1
1
1
1
2 a
2 a
2 a
2 a
2 b
-2
-2
-2
-2
-2
-2
-2
-2
-2
-2
(
1.0 ( 0.5) × 10
1.6 ( 0.5) × 10
2.0 ( 0.7) × 10
2.7 ( 1.0) × 10
8.0 ( 1.3) × 10
(8.5 ( 0.7) × 10
(5.4 ( 0.3) × 10
(5.2 ( 0.2) × 10
(4.6 ( 0.5) × 10
(4.4 ( 0.7) × 10
(2.5 ( 1.0) × 10
(2.3 ( 0.7) × 10
(2.3 ( 0.4) × 10
(2.0 ( 0.3) × 10
(1.0 ( 0.5) × 10
(
(
(
(
In contrast to the interaction of NO2 and HNO3 with gray
21,23
and black hexane and decane soot
whose product branching
ratio strongly depended on the type of soot, we see the opposite
behavior, namely a lack of dependence on the type of soot in
the case of NO3 and N2O5. This finding is somewhat unexpected
compared to both the behavior of NO2 and HNO3 and is in
3
[N O
2 5
]
0
cm-
γ
0
(black soot)
γss (black soot)
1
1
1
1
1
1
1
2 a
2 a
2 a
2 a
2 a
2 a
2 b
-2
-2
-2
-2
-2
-3
-3
(
(
(
(
(
(
(
1.0 ( 0.5) × 10
1.3 ( 0.7) × 10
1.5 ( 0.5) × 10
1.9 ( 0.8) × 10
2.0 ( 0.8) × 10
2.9 ( 1.0) × 10
8.0 ( 1.3) × 10
0.1 ( 0.05
(3.3 ( 0.8) × 10
(3.4 ( 0.5) × 10
(3.0 ( 0.6) × 10
(2.0 ( 0.8) × 10
(2.0 ( 0.5) × 10
(8.0 ( 1.4) × 10
(3.6 ( 1.0) × 10
0.1 ( 0.05
24
disagreement with the results of Smith and co-workers who
0.1 ( 0.08
-2
-2
-2
-2
have observed a strong correlation of the reactivity of O3 and
of the interaction of H2O vapor with several types of combustion
soot as a function of the air/fuel ratio that is comparable to the
fuel/oxygen ratio used in the present work. In general agreement
(8.9 ( 0.5) × 10
(7.5 ( 0.5) × 10
(7.3 ( 0.5) × 10
(4.4 ( 0.5) × 10
5
0-52
a
with the present results Longfellow et al.
NO2 of 104 ( 10% and a small HONO yield on the order of
% for the interaction of N2O5 on propane soot whose fuel to
find a yield of
Initial (γ
0
) and steady-state (γss) uptake coefficients (orifice diameter
8 mm; orifice diameter ) 4 mm).
)
1
oxygen ratio was not controlled. The small HONO yield was
In contrast to NO221 and HNO323 the mass balance for both N2O5
Table 4) and NO3 (Table 2) indicates that the yields of reaction
believed to originate from the secondary reaction of NO on
2
(
propane soot. On the side, the present work yields additional
products are independent of the type of soot, in agreement with
the measured uptake coefficients that are also essentially
independent of the type of soot, as displayed in Figures 9 and
confirmation of HONO formation of NO2 on decane soot
compared to previous results obtained on hexane soot21 and
lends support to the importance of the fuel to oxygen ratio λ in
2
1-24
11. We attribute the higher NO yield in the case of NO3 reacting
determining the reactivity of soot toward NO2.
with black compared to gray decane soot to the decomposition
of HONO that is formed as a consequence of the presence of
NO2 on black soot. This is an artifact of the used NO3 source
that delivers 3 times as much NO2 compared to NO3 and that
forces us to take into account the interaction of NO2 with soot
in detail. In contrast to NO3 and N2O5, the type of soot matters
for the NO2/soot interaction.32
Finally, we would like to point out that the heterogeneous
reduction reaction of both NO3 and N2O5 on both types of soot
formally corresponds to a 4 electron transfer in regard to NO
formation from NO3 and NO + NO2 formation from N2O5. This
is accompanied with the formal transfer of two O atoms from
the reacting trace gas to the reducing soot substrate. In contrast,
HONO formation from NO2 implies a formal transfer of one
2
1,22,40,41
The N2O5 yield of 23% is observed to be independent of the
type of soot and is a consequence of the presence of adsorbed
NO3 reacting with excess gaseous NO2. Adsorbed NO2 is
reacting to HONO and NO on gray and black decane soot,
respectively.22 We conclude that the extents of adsorption of
electron,
3
whereas HONO formation from HNO /soot
2
1
interaction requires the transfer of two electrons. In this
analysis we regard the NO formation from the NO2/black decane
soot interaction as a secondary product of HONO decomposition
22
on soot, i.e., a one-electron reduction.

1924 J. Phys. Chem. A, Vol. 111, No. 10, 2007
Karagulian and Rossi
6
3(HNO3)
63(HNO3)
From the present kinetic uptake experiments performed on
gray and black decane soot, we may extrapolate γ to vanishing
In the absence of a substrate, f I
30 0
and f I
have
4
6 0
46
30
been subtracted from the total MS signals I0 and I0 at m/e 46
and 30, respectively, to assign the remaining MS amplitude to
NO3 concentration as displayed in Figure 9 and obtain an
1
1
-3
estimate of γss for [NO3] < 7.0 × 10 molecule cm
+
+
46(N2O5)
0
46
the NO2 and NO fragments of N2O5: I
) I0 - f46
independent of the type of soot. With tropospheric [NO3] at a
63(HNO )
0
30(N O )
30
3
63(HNO )
3
2
5
I
and I0
) I0 - f30I0
. The same type of
9
-3 30
typical value of 2.0 × 10 molecule cm , γss tends toward
procedure has been used for experiments in the presence of a
sample once all the reaction products were known, starting at
the high mass end (HNO3, NO3, HONO, N2O5, NO2, and NO).
values larger than 0.3 according to the results displayed in Figure
9
. The NO3 loss rate constant (khet) due to heterogeneous uptake
-
1
onto aerosols is given by khet ) τ (NO3) ) γA jc /4V as far as
the surface resistance is concerned without consideration of mass
transfer, where γ is the uptake coefficient of NO3, A/V is the
surface-to-volume ratio of the aerosol, and cj is the mean
molecular speed of NO3.
4
6(NO2)
(REMPI)
63(HNO3)
and f46I
0
In the absence of the soot substrate, I0
have been subtracted from the total MS signal I0 at m/e 46 to
attribute the remaining signal to the NO2 fragment of the
electron-impact ionization of NO3 once the absence of nondis-
sociated N2O5 from the NO3 source was established. The
temperature of the NO3 source was increased to the point where
all non dissociated N2O5 fell below the detection limit. The
46
+
The uptake of NO3 on soot was studied in a large aerosol
-
4
chamber and yielded an upper limit of γ e 3.0 × 10 at very
low relative humidity (rh) (H2O < 10 ppm), whereas at 50%
4
0
6(NO3)
46
) I0
-3
resulting MS signal at m/e 46 pertaining to NO is I
rh an upper limit of γ e 1.0 × 10 was obtained. These values
are 2 (50% rh) and 3 orders of magnitude lower than the γ
3
4
6(NO2)
63(HNO3)
- I0
- f46I0
.
(REMPI)
1
1
-3
value of 0.3 for [NO3] < 7.0 × 10 molecule cm . The impact
When soot is exposed to the effluents of the NO3 source,
NO3 is taken up and reacts on soot, resulting in a decrease of
[NO ] that leads to a concomitant decrease of the MS signal Ir
of soot aerosol surface reactions on the rate of formation of
photochemical ozone was also investigated using a box model.
11
46
3
The results showed that soot has a minor impact on ozone
formation at low [NOy]. In contrast, soot may cause an ozone
reduction of up to 10% at high [NOy]. However, we must point
out that the nature of the carbonaceous aerosol from a spark
at m/e 46. To correct for the contribution of NO at m/e 46 for
3
the following series of uptake experiments, we have determined
4
6(NO3) 62(NO3)
r46 ) I0
/I0
) 8.5 ( 1.5 as the ratio of the MS signal
46(NO3)
+
62(NO3)
, the molecular ion peak at
0
I
at m/e 46 (NO2 ) and I
m/e 62 (NO3 ), both for NO3 free radical in the absence of the
soot sample.
0
1
1
ignition generator is significantly different from flame soot
used in the present study. At variance, the work of Saathoff et
+
1
1
al. shows that the uptake of NO3 on soot seems to saturate
and tend to a small γss value at steady state. This may perhaps
be due to the graphitic core of the used soot that proves to be
unreactive toward NO3 free radical attack once the reactive
As a result of the exposure of the sample to NO3 in the
presence of NO2, we expect four possible reaction products
besides changes of NO2: HNO3, N2O5, HONO, and NO. Under
our experimental conditions HNO3 may possibly be formed at
high densities by heterogeneous recombination of NO2 and NO3
“
soot” surface functionalities are consumed and the “bare”
graphitic core is exposed. On the other hand, the present
experiments are not sensitive to saturation phenomena owing
to the shorter time scale used such that critical comparison of
results has to await further experiments.
to N2O5 and subsequent heterogeneous hydrolysis. To find other
46
exc
possible reaction products contributing to the I MS signal
intensity at m/e 46 not due to HNO3, we have subtracted the
4
6
following known contributions from the total MS signal I :
r
4
6(NO2)
62(NO3)
63(HNO3)
(a) I
for NO2, (b) r46Ir
for NO3, and (c) f46Ir
Acknowledgment. We gratefully acknowledge the Fonds
National Suisse de la Recherche Scientifique (FNS) for generous
support of this research under project no. 200020-105471/1 as
well as OFES, now SER (Secr e´ tariat de l’enseignement et de
la recherche), for financing the Swiss participation of the
MINATROC project within the fifth framework program of the
European Union (EU).
r(REMPI)
for the possible HNO3 formation during the reaction which result
in the following remaining MS signal called excess MS intensity
at m/e 46:
4
exc
6
46
r
46(NO2)
r(REMPI)
62(NO3)
46 r
63(HNO3)
I
) I - I
- r I
- f I
46 r
(A1)
It is reasonable to expect that N2O5 may be the only reaction
Appendix A: Data Analysis for the Detection of Reaction
Products
product contributing to an excess at m/e 46, as will be discussed
4
exc
6
46(N2O5)
below. Therefore, in the following I will be named Iexc
.
A1. NO3 Reaction and Detection. The absolute NO2
concentration [NO2]0(REMPI) originating from the NO3 source has
been determined by means of in situ REMPI detection akin to
the method outlined in ref 44. Subsequently, the corresponding
Using REMPI detection at λNO ) 452.6 nm, we could not
detect any NO product formation because its concentration
dropped below the detection limit given by the chosen experi-
mental conditions. Therefore, to establish the amount of NO
4
6(NO2)
30
MS signal of NO2 at m/e 46, I0
, was calculated accord-
from the excess MS signal intensity at m/e 30, I
(REMPI)
exc, during the
ing to eq A3 (see below). We point out that in contrast to MS
signals that are proportional to an ion current, and thus to flux,
the REMPI signals scale with the concentration of the detected
trace gas as explained in detail in ref 44.
Through a calibrated mass spectrum of pure HNO3 we have
accurately determined the contribution of the HNO3 impurity
at m/e 46 and 30 by using the fragmentation pattern expressed
as the ratios
exposure of soot to NO3, we have accurately determined all
30
the possible contributions to the total MS signal Ir at m/e 30.
30
The major contribution to Ir comes from the mixture of NO2
and NO3 originating from the hot NO3 source given that the
amount of HNO3 and N2O5 in the presence of the sample is
small. Using a reference mass spectrum of pure NO2, we have
calculated the effective contribution of NO2 at m/e 30 by using
the fragmentation pattern expressed as the ratio
4
0
6(HNO3)
3(HNO3)
30(HNO3)
0
I
I
I
I
30(NO2)
0
I
f )
) 52 ( 8
f )
) 33 ( 4
4
6
30
z1 )
) 2.0 ( 0.2
6
o
63(HNO3)
o
46(NO2)
0
I

NO3 Free Radical and N2O5 on Decane Flame Soot
J. Phys. Chem. A, Vol. 111, No. 10, 2007 1925
3
0(NO3)
+
62(NO3)
The ratio of the MS signal I0
at m/e 30 (NO ) and I0
of the soot sample, the NO2 REMPI signal was scaled to a MS
signal using the following equation:
+
at m/e 62 (NO3 ) for NO3 radical has been defined as follows:
NO2
[
NO ]
k
esc(NO2)Vcell
3
0
0(NO3)
2(NO3)
30
0
46(NO2)
1 0
63(HNO3)
F
0,r
2 0,r(REMPI)
4
0
6(NO2)
,r(REMPI)
I
I
I - z I
- f I
30 0
I
)
)
r )
)
) 6.3 ( 0.8
C
C
cal(NO2)
3
0
cal(NO2)
6
0
62(NO3)
0
I
(A3)
3
0
46(NO2)
where Ccal(NO ) is the NO2 calibration factor that was directly
determined from an absolute measurement of [NO2] in an
ancillary experiment using a suitably calibrated MS signal for
2
where I is the total MS signal at m/e 30 and z1I0
I
3
and f30
0
6
0
3(HNO3)
are the contributions of NO2 and HNO3, both at m/e
0, respectively.
As explained above, it is reasonable to expect that N2O5 will
be a reaction product of the reaction of NO3 on soot. However,
we have to consider that pure N2O5 has fragment peaks at m/e
6 and 30 that are correlated by the fragmentation pattern
expressed as the ratio rN O ) I
This ratio was established after extensive passivation of the
inlet system and running the NO3 source at ambient temperature
to afford a stabilized flow of N2O5. As shown in previous studies
on soot,23 HNO3 reacts on the soot surface resulting in the
formation of volatile products such as HONO and NO that
pure NO2. kesc(NO ) is the NO2 effusive loss rate constant. One
2
should be reminded that the conversion following eq A3 is
necessary because the REMPI signal scales with the concentra-
tion whereas the MS signal corresponds to a flux (see above).
Equation A3 allows one to calculate the fraction of the MS
signal at m/e 46 owing to the presence of NO2 using the
measured REMPI signal for NO2 to establish its absolute
concentration. Therefore, we have corrected the MS signal at
4
4
0
6(N2O5) 30(N2O5)
/I0
) 1.36 ( 0.3.
2
5
m/e 46 for the presence of NO according to eq A4 when the
2
sample is exposed to N O and obtained the MS signal at m/e
2
5
46 for pure N O under the assumption that there are no
2
5
3
0
46
contribute to the total MS signal Ir at m/e 30. At the present
additional contributions to I .
r
experimental conditions HONO has a measurable, albeit low
intensity, parent peak at m/e 47 (HONO ). Therefore, the
4
6(N2O5)
46
0
46(NO2)
r(REMPI)
+
Ir
) I - I
(A4)
effective contribution of HONO at m/e 30 has been determined
by using its fragmentation pattern expressed as the ratio h )
which thus exclusively corresponds to N2O5 for this experiment
after correction of the MS signal I
contribution of NO2 produced during uptake of N2O5 on the
soot substrate. Specifically, significant amounts of HNO3 were
not detected in the present experiments.
To calculate the amount of NO2 formed during the reaction
of N O on gray and black soot, we have subtracted the
contribution of N2O5 from the total MS signal at m/e 46. The
MS signal at m/e 30 was corrected for its NO contribution when
NO was detected by REMPI, namely, I , after scaling to
a flow rate or MS signal. Therefore, the total contribution of
3
0
0(HONO) 47(HONO)
46
r
I
/I0
) 22 ( 0.5. As already explained above,
at m/e 46 for the
HNO3 present in the hot NO3 source also provides a contribution
to the MS amplitude at m/e 30 as expressed by f30.
Finally, during the exposure of soot to NO3 we have
subtracted the following known contributions from the total MS
3
0
46
exc
2
5
signal Ir at m/e 30: (a) I /rN O for the contribution owing to
2
4
5
6(NO2)
the presence of N2O5, (b) z1I
NO2 (see eq A3, next paragraph), (c) r30I
contribution of NO3, (d) hIr
for the contribution of
r(REMPI)
6
r
2(NO3)
for the
30(NO)
r(REMPI)
4
7(HONO)
63(HNO3)
for HONO, (e) f30Ir
for HNO3. The final expression for the residual ()excess) MS
3
r
0
30(NO)
3
0
N2O5 to the MS signal at m/e 46 was rN2O5(I
final expression for I attributed to the presence of NO2 is
- Ir(REMPI)). The
signal amplitude I attributed to the presence of NO resulted
exc
46
exc
from the following equation:
given by eq A5.
4
exc
6(N2O5)
4
exc
6
46
r
30
r
30(NO)
r(REMPI)
I
I
) I - r (I - I
)
(A5)
3
exc
0
30
r
(46)NO2
1 r(REMPI)
62(NO3)
-
r
N O
2
5
I
) I -
- z I
- r I
30
r
N O
2
5
Conversely, to calculate the amount of NO formed during the
reaction of N2O5 on gray and black soot, we have subtracted
the contribution of N2O5 from the total MS signal at m/e 30.
The MS signal at m/e 46 was corrected for its NO2 contribution
4
7(HONO)
63(HNO3)
30 r
hIr
- f I
(A2)
It is reasonable to expect that N2O5 and NO may be the only
reaction products contributing to m/e 46 and 30, respectively,
after subtraction of all known contributions to these masses
displayed in eqs A1 and A2.
(NO2)46
when NO was detected by REMPI, namely, I
2
after
r(REMPI)
scaling to a flow rate or MS signal. Therefore, the total
4
r
6
contribution of N2O5 to the MS signal at m/e 30 was (I
-
46(NO2)
30
A2. N2O5 Reaction and Detection. For N2O5 uptake experi-
ments performed on gray and black soot, both NO2 and NO
could be observed in the gas phase. The NO REMPI detection
sensitivity was increased in the N2O5 uptake experiments such
that both products could be detected in situ in back-to-back
experiments, albeit not concurrently. The reaction product not
determined using REMPI was quantified using MS signal
analysis of amplitudes at m/e 30 and 46. The absolute NO and
NO2 concentrations obtained using REMPI detection were then
scaled to the corresponding flow rates using eq A3 for NO2
and the analogous expression for NO. The relevant MS
calibration factors were previously established in ancillary
experiments (see above).
I
)/rN O . The final expression for I attributed to the
r(REMPI)
2 5
exc
presence of NO is given by eq A6.
4
r
6
46(NO2)
r(REMPI)
(I - I
)
3
exc
0
30
r
I
) I -
(A6)
r
N O
2
5
This analysis makes use of the observation that both HNO3 and
HONO were undetectable during uptake of N2O5 on gray and
black soot.
References and Notes
(1) Houghton, J., et al., Eds. Climate Change 2001, The Scientific Basis;
Cambridge University Press: Cambridge, U.K., 2001.
(2) Goldberg, E. D. Black Carbon in the EnVironment John Wiley and
Sons: New York, 1985.
4
6(NO2)
The corresponding MS signal contribution I0
6 in the absence (subscript “0”) and presence (subscript “r”)
at m/e
,r(REMPI)
4

1926 J. Phys. Chem. A, Vol. 111, No. 10, 2007
Karagulian and Rossi
(
3) Buseck, P. R.; Posfai, M. Proc. Natl. Acad. Sci. U.S.A. 1999, 96,
372.
4) Posfai, M.; Simonics, R.; Li, J.; Hobbs, P. V.; Buseck, P. R. J.
Geophys. Res.sAtmos. 2003, 108.
5) Hallett, J.; Hudson, J. G.; Rogers, C. F. Aerosol Sci. Technol. 1989,
0, 70.
(30) Brown, S. S.; Dube, W. P.; Osthoff, H. D.; Wolfe, D. E.; Angevine,
W. M.; Ravishankara, A. R. Atmos. Chem. Phys. Discuss. 2006, 6, 9431.
(31) Kleffmann, J.; Becker, K. H.; Lackhoff, M.; Wiesen, P. Phys. Chem.
Chem. Phys. 1999, 1, 5443.
(32) Tabor, K.; Gutzwiller, L.; Rossi, M. J. J. Phys. Chem. 1994, 98,
6172.
3
(
(
1
(
(
(
6) Jensen, E. J.; Toon, O. B. Geophys. Res. Lett. 1997, 24, 249.
7) Horvath, H. Atmos. EnViron. Pt. AsGen. Top. 1993, 27, 293.
8) Haywood, J. M.; Ramaswamy, V. J. Geophys. Res.sAtmos. 1998,
(33) Rogaski, C. A.; Golden, D. M.; Williams, L. R. Geophys. Res. Lett.
1997, 24, 381.
(
34) Gerecke, A.; Thielmann, A.; Gutzwiller, L.; Rossi, M. J. Geophys.
Res. Lett. 1998, 25, 2453.
35) Kalberer, M.; Ammann, M.; Arens, F.; G a¨ ggeler, H. W.; Balt-
ensperger, U. J. Geophys. Res.sAtmos. 1999, 104, 13825.
36) Kirchstetter, T. W.; Harley, R. A.; Littlejohn, D. EnViron. Sci.
Technol. 1996, 30, 2843.
1
03, 6043.
(9) Schurath, U.; Naumann, K. H. Pure Appl. Chem. 1998, 70, 1353.
(
(
10) Kamm, S.; M o¨ hler, O.; Naumann, K. H.; Saathoff, H.; Schurath,
U. Atmos. EnViron. 1999, 33, 4651.
11) Saathoff, H.; Naumann, K. H.; Riemer, N.; Kamm, S.; M o¨ hler,
O.; Schurath, U.; Vogel, H.; Vogel, B. Geophys. Res. Lett. 2001, 28, 1957.
12) Kirchner, U.; Scheer, V.; Vogt, R. J. Phys. Chem. A 2000, 104,
908.
(
(
(37) Finlayson-Pitts, B. J.; Pitts, J. N. Science 1997, 276, 1045.
(
(38) Staffelbach, T.; Neftel, A.; Blatter, A.; Gut, A.; Fahrni, M.; St a¨ helin,
8
J.; Prev oˆ t, A.; Hering, A.; Lehning, M.; Neininger, B.; B a¨ umle, M.; Kok,
G. L.; Dommen, J.; Hutterli, M.; Anklin, M. J. Geophys. Res.sAtmos. 1997,
(
(
13) IUPAC. http://www.iupac-kinetic.ch.cam.ac.uk/.
14) Pope, C. A.; Thun, M. J.; Namboodiri, M. M.; Dockery, D. W.;
1
02, 23345.
39) Sakamaki, F.; Hatakeyama, S.; Akimoto, H. Int. J. Chem. Kinet.
983, 15, 1013.
40) Stemmler, K.; Ammann, M.; Donders, C.; Kleffmann, J.; George,
C. Nature 2006, 440, 195.
41) Bartels-Rausch, T.; Donaldson, D. J. Atmos. Chem. Phys. Discuss
006, 6, 10713.
Evans, J. S.; Speizer, F. E.; Heath, C. W. Am. J. Respir. Crit. Care Med.
995, 151, 669.
15) Pope, C. A.; Burnett, R. T.; Thurston, G. D.; Thun, M. J.; Calle,
E. E.; Krewski, D.; Godleski, J. J. Circulation 2004, 109, 71.
16) Akhter, M. S.; Chughtai, A. R.; Smith, D. M. Appl. Spectrosc. 1985,
9, 143.
17) Bockhorn, H. In Soot Formation in Combustion: Mechanisms and
Models; Bockhorn, H. Ed.; Springer-Verlag: Berlin, Heidelberg, 1994.
18) Sergides, C. A.; Jassim, J. A.; Chughtai, A. R.; Smith, D. M. Appl.
Spectrosc. 1987, 41, 482.
19) P o¨ schl, U.; Letzel, T.; Schauer, C.; Niessner, R. J. Phys. Chem. A
001, 105, 4029.
20) Esteve, W.; Budzinski, H.; Villenave, E. Atmos. EnViron. 2006,
0, 201.
(
1
1
(
(
(
(
3
2
(
(
(
42) Takenaka, N.; Rossi, M. J. J. Atmos. Chem. 2005, 50, 171.
43) Caloz, F.; Fenter, F. F.; Tabor, K. D.; Rossi, M. J. ReV. Sci. Instrum.
(
1
3
2
997, 68, 3172.
(
44) Karagulian, F.; Rossi, M. J. Phys. Chem. Chem. Phys. 2005, 7,
(
150.
2
(45) Hanisch, F.; Crowley, J. N. Phys. Chem. Chem. Phys. 2001 b, 3,
(
474.
4
(
46) Ammann, M.; Kalberer, M.; Jost, D. T.; Tobler, L.; R o¨ ssler, E.;
Piguet, D.; G a¨ ggeler, H. W.; Baltensperger, U. Nature 1998, 395, 157.
47) Brouwer, L.; Rossi, M. J.; Golden, D. M. J. Phys. Chem. 1986,
0, 4599.
(
(
(
21) Salgado, M. S.; Rossi, M. J. Int. J. Chem. Kinet. 2002, 34, 620.
22) Stadler, D.; Rossi, M. J. Phys. Chem. Chem. Phys. 2000, 2, 5420.
23) Munoz, M. S. S.; Rossi, M. J. Phys. Chem. Chem. Phys. 2002, 4,
(
9
5
110.
(
48) Keyser, L. F.; Moore, S. B.; Leu, M. T. J. Phys. Chem. 1991, 95,
(
24) Chughtai, A. R.; Kim, J. M.; Smith, D. M. J. Atmos. Chem. 2002,
5
496.
4
3, 21.
(
25) Novakov, T.; Ramanathan, V.; Hansen, J. E.; Kirchstetter, T. W.;
Sato, M.; Sinton, J. E.; Sathaye, J. A. Geophys. Res. Lett. 2003, 30.
26) Penner, J. E.; Eddleman, H.; Novakov, T. Atmos. EnViron. Pt. As
Gen. Top. 1993, 27, 1277.
(49) Rivera-Figueroa, A. M.; Sumner, A. L.; Finlayson-Pitts, B. J.
EnViron. Sci. Technol. 2003, 37, 548.
(50) Longfellow, C. A.; Ravishankara, A. R.; Hanson, D. R. J. Geophys.
Res.sAtmos. 2000, 105, 24345.
(
(
(
(
27) Bodhaine, B. A. J. Geophys. Res.sAtmos. 1995, 100, 8967.
28) Clarke, A. D. Aerosol Sci. Technol. 1989, 10, 161.
29) Rau, J. A.; Khalil, M. A. K. Atmos. EnViron. Pt. AsGen. Top.
(51) Berner, A.; Sidla, S.; Galambos, Z.; Kruisz, C.; Hitzenberger, R.;
tenBrink, H. M.; Kos, G. P. A. J. Geophys. Res.sAtmos. 1996, 101, 19559.
(52) Martinez, M.; Perner, D.; Hackenthal, E. M.; Kulzer, S.; Sch u¨ tz,
L. J. Geophys. Res.sAtmos. 2000, 105, 22685.
1
993, 27, 1297.