Secondary organic aerosol formation via the isolation of individual reactive intermediates: Role of alkoxy radical structure

Article abstract of DOI:10.1021/jp506562r

The study of the chemistry underlying secondary organic aerosol (SOA) formation is complicated by the large number of reaction pathways and oxidation generations available to a given precursor species. Here we simplify such complexity to that of a single alkoxy radical (RO), by forming SOA via the direct photolysis of alkyl nitrite (RONO) isomers. Chamber experiments were conducted with 11 C₁₀RONO isomers to determine how the position of the radical center and branching of the carbon skeleton influences SOA formation. SOA yields served as a probe of RO reactivity, with lower yields indicating that fragmentation reactions dominate and higher yields suggesting the predominance of RO isomerization. The largest yields were from straight-chain isomers, particularly those with radical centers located toward the terminus of the molecule. Trends in SOA yields can be explained in terms of two major effects: (1) the relative importance of isomerization and fragmentation reactions, which control the distribution of products, and (2) differences in volatility among the various isomeric products formed. Yields from branched isomers, which were low but variable, provide insight into the degree of fragmentation of the alkoxy radicals; in the case of the two β-substituted alkoxy radicals, fragmentation appears to occur to a greater extent than predicted by structure-activity relationships. Our results highlight how subtle differences in alkoxy radical structure can have major impacts on product yields and SOA formation.

Full text of DOI:10.1021/jp506562r

Article
pubs.acs.org/JPCA
Secondary Organic Aerosol Formation via the Isolation of Individual
Reactive Intermediates: Role of Alkoxy Radical Structure
Anthony J. Carrasquillo,† James F. Hunter,† Kelly E. Daumit,† and Jesse H. Kroll*,†,‡
‡
†Department of Civil and Environmental Engineering and Department of Chemical Engineering, Massachusetts Institute of
Technology, Cambridge, Massachusetts 02139, United States
S
* Supporting Information
ABSTRACT: The study of the chemistry underlying secon-
dary organic aerosol (SOA) formation is complicated by the
large number of reaction pathways and oxidation generations
available to a given precursor species. Here we simplify such
complexity to that of a single alkoxy radical (RO), by forming
SOA via the direct photolysis of alkyl nitrite (RONO) isomers.
Chamber experiments were conducted with 11 C10 RONO
isomers to determine how the position of the radical center
and branching of the carbon skeleton influences SOA
formation. SOA yields served as a probe of RO reactivity,
with lower yields indicating that fragmentation reactions dominate and higher yields suggesting the predominance of RO
isomerization. The largest yields were from straight-chain isomers, particularly those with radical centers located toward the
terminus of the molecule. Trends in SOA yields can be explained in terms of two major effects: (1) the relative importance of
isomerization and fragmentation reactions, which control the distribution of products, and (2) differences in volatility among the
various isomeric products formed. Yields from branched isomers, which were low but variable, provide insight into the degree of
fragmentation of the alkoxy radicals; in the case of the two β-substituted alkoxy radicals, fragmentation appears to occur to a
greater extent than predicted by structure−activity relationships. Our results highlight how subtle differences in alkoxy radical
structure can have major impacts on product yields and SOA formation.
they will continue to react in the gas phase and form later-
generation oxidation products that may themselves condense.5
The resulting aerosol thus contains many product species,
formed by numerous pathways over several generations of
oxidation.
To simplify this complex chemistry, we recently demonstrated
the utility of forming SOA by the direct photolytic generation of a
single organic radical species.6 This approach limits both the
number of reactive pathways available (because only one radical
species is formed) and the number of oxidation generations
(because no oxidant is added). In that initial work, alkyl iodides
(RI) were photolyzed at 254 nm in the presence of oxygen to give
an alkylperoxy radical (RO2). Here we expand this approach to a
second key intermediate, the alkoxy radical, by the photolysis of
alkyl nitrites (RONO):
INTRODUCTION
■
Organic aerosol (OA) particles have important implications for
human health, visibility, and climate; however, our ability to
accurately predict ambient OA loadings, chemical composition,
and properties is limited.1−3 Particularly uncertain is the
chemistry associated with secondary organic aerosol (SOA),
formed from the oxidation of gas-phase VOCs and the
subsequent condensation of low-volatility products. SOA
formation is immensely complex, involving a large number of
chemical species, reaction pathways, and oxidation genera-
tions.3,4 This chemical complexity poses major challenges in
understanding the detailed chemistry and specific reaction
pathways that govern SOA formation and evolution, particularly
when the reaction of a single oxidant with a single VOC can lead
to the formation of a multitude of products.
Figure 1 illustrates the large number of reaction pathways and
intermediates possible for even a very simple reaction system, the
oxidation of an n-alkane by the OH radical. The reaction is
initiated by H-abstraction by OH at one of several sites, followed
by O2 addition to form a number of distinct alkylperoxy radicals
(RO2, not shown). In the presence of NO these will form
organonitrate (RONO2, not shown) species or alkoxy (RO)
radicals. The RO radicals are important branch points in
oxidation reactions, and for large species will either fragment or
isomerize. If the resulting products are sufficiently low in
volatility, they will partition into the aerosol phase; otherwise,
(1)
RONO + hv → RO + NO
Reaction 1 generally occurs in the 300−400 nm range and
therefore involves much lower-energy photons than required for
RI photolysis.7,8 RO radicals may still be initially formed
vibrationally9 or electronically7 excited, but for the large (C10)
species considered here, collisional deactivation by bath gas
Received: July 1, 2014
Revised: August 21, 2014
Published: August 22, 2014
© 2014 American Chemical Society
8807
dx.doi.org/10.1021/jp506562r | J. Phys. Chem. A 2014, 118, 8807−8816

The Journal of Physical Chemistry A
Article
Figure 1. Formation of alkoxy radical isomers from n-decane. The OH radical can abstract any available H, yielding a distribution of five different radical
isomers that then react to form a wide range of fragmentation or isomerization products; these then can oxidize further to form new products. Alkyl
nitrite photolysis, shown in red, yields one specific alkoxy radical, with little secondary oxidation, greatly limiting the number of products formed.
Figure 2. Alkoxy radical isomers examined in this study, listed according to their structural characterization (primary, secondary, or branched).
molecules is likely to remove this excess energy prior to reaction.8
In addition, because the NO photofragment is far less reactive
than I atoms formed from RI photolysis, any undesired (and
uncertain) secondary chemistry is less likely than in our previous
work.
alcohols used are 1-decanol and 2-decanol (Sigma-Aldrich), 3-
decanol, 4-decanol, 5-decanol, 2-methyl-5-nonanol, 2-methyl-4-
nonanol, 4-methyl-5-nonanol, 2-methyl-2-nonanol, 5-methyl-5-
nonanol, and 3,4-diethyl-5-hexanol (Chemsampco). The alkoxy
radicals formed from the 11 C10 nitrite isomers synthesized are
listed in Figure 2. These radicals are grouped by degree of
substitution: primary (straight-chain molecules with the radical
center at the terminal carbon atom), secondary (straight-chain
molecules with the radical center at one of the secondary/
internal carbon atoms), and branched (molecules with at least
one methyl or ethyl group branching off the carbon chain).
Synthesis of the nitrites was carried out via O-nitrosation of the
alcohol.12 In this procedure, all glassware, laboratory utensils, and
aluminum foil were combusted at 550 °C for 10 h to remove all
organic contaminants. Sodium nitrite (Sigma-Aldrich) and the
alcohol were combined in a 1.1:1.0 molar ratio and stirred with a
magnetic stirrer at 10 °C. To this mixture was added chilled 6 M
sulfuric acid (Sigma-Aldrich) dropwise until a 0.5:1.0 acid:-
alcohol molar ratio was achieved. The resulting clear yellow
liquid was dried over sodium sulfate (Sigma-Aldrich) and
neutralized with excess sodium bicarbonate (Sigma-Aldrich).
Purity was confirmed by analysis with GC-FID and UV−vis. For
GC analysis, 1 μL of nitrite solution (1 μg RONO/1 mL hexane)
was injected into an Agilent 7890 GC-FID (50 m, 0.25 mm i.d.
DB-5 column; held at 40 °C for 3 min, then ramped to 250 °C at
20 °C/min), which provided sufficient separation of the RONO
(10−12 min retention time) from the precursor alcohol (12−14
In studies of OH + alkanes under high-NO conditions,
Ziemann and co-workers have shown that the chemistry of
alkoxy radicals is crucial in determining the yields and
composition of SOA.5,10,11 Such chemical systems can involve
a complex mixture of several RO isomers. Here, using the
photolysis of a series of C10 RONO isomers, we are able to
examine RO radical chemistry in even more detail, by isolating
the role of individual alkoxy radical species in SOA formation.
Furthermore, in the absence of an added oxidant, the effects of
multigenerational oxidative chemistry are minimized. The study
of aerosol formation from individual alkoxy radical isomers thus
allows for a highly detailed examination of the role of chemical
structure (position and degree of substitution of the radical
center) on SOA yields and composition.
EXPERIMENTAL SECTION
■
Selection and Synthesis of Alkyl Nitrite Species. RONO
species are not commercially available, and so were synthesized
in the laboratory. This synthesis, described below, requires the
corresponding alcohol as the starting material; a series of C10
RONO isomers (formula C10H21ONO) was chosen on the basis
of the large number of precursor alcohols available. The C10
8808
dx.doi.org/10.1021/jp506562r | J. Phys. Chem. A 2014, 118, 8807−8816

The Journal of Physical Chemistry A
Article
min retention time). For all syntheses, the ratio of FID signals for
the RONO compound to the unreacted alcohol was >95:1,
indicating a conversion of at least 95% of the starting material to
the nitrite. Confirmation of conversion to RONO was also made
by UV−vis spectroscopy of the same RONO mixture, with
measured spectra similar to those reported by Heicklen.8
Synthesized RONO species were stored in the dark and used
within 1−2 h of synthesis to maintain its integrity.
ΔcOA(t)
ΔRONO(t)·0.840
Y (%) =
·100%
(2)
where ΔcOA(t) is the increase in organic aerosol concentration
(μg/m3) and ΔRONO(t) is the mass concentration of nitrite
photolyzed (the multiplicative factor 0.840 is to exclude the mass
of the NO photofragment from the yield calculation). ΔRONO-
(t) is calculated as the difference between the initial amount
present, RONO(t0), and the amount remaining at time t,
corrected for flushing using measured C6F6 concentrations. For
all RONO measurements the FID calibration factor was
determined by averaging the GC-FID response across all
experiments at t0 and equating this to 200 ppbv.
The AMS and an SMPS (TSI model 3080) were used to
determine organic aerosol formation (ΔcOA). Aerosol growth
was corrected for wall loss and dilution using the method of
Hildebrandt et al.,16 in which the organic-to-sulfate mass ratio (as
measured by the AMS) is scaled by the initial sulfate mass
(calculated from the SMPS volume and a density of 1.77 g/cm3).
Because the sulfate mass concentration will change only by
physical loss processes (deposition to the walls and flushing), this
approach corrects for losses of organic particle mass to walls and
dilution, losses of organic vapors to OA deposited on the
chamber walls, and any changes to the AMS collection efficiency
or sensitivity over the course of the experiment. This approach
does not account for any partitioning of semivolatile species to
the chamber walls,17 an effect that might lower the measured
yields; however, the overall trends are unlikely to be affected by
such wall-loss processes.
Chamber Studies. Experiments were conducted in a 1 m3
Teflon PFA (40 mil) chamber (Ingeniven) surrounded by four
40 W black lights (Sylvania BL350). The spectral output of these
lamps (300−400 nm, centered at ∼350 nm) overlaps well with
the absorption spectrum of RONO species,8 allowing for
efficient photolysis. The chamber was operated at a constant
volume such that the air sampled was balanced with dry (RH <
7%) makeup air from a clean air generator (Aadco, <5 ppbv
hydrocarbons). Gas-phase RONO concentration was monitored
with a GC-FID (SRI 8610C) equipped with a Tenax trap and a
30 m 0.53 mm DB-5 column. Prior to each experiment, the
chamber was flushed with clean air for a minimum of 10 h, giving
a background particle concentration of <10 particles/cm3.
The RONO species and a dilution tracer hexafluorobenzene
(C6F6, Sigma-Aldrich) were injected into the chamber via a
stream of air to attain initial chamber concentrations of 200 ppbv
and 40 ppbv, respectively. For the experiments described here,
high levels of NO (∼1 ppm) were added to the chamber, both to
ensure that RO2 + NO reactions dominate the RO2 chemistry
across all experiments and to inhibit the formation of O3 and
NO3. Ammonium sulfate ((NH4)2SO4) seed particles were
added to promote condensation of low-volatility species onto
particles, and also to correct for losses of particles to dilution or
wall loss (as described below). Polydisperse (NH4)2SO4 seed was
generated by atomizing a 1 g/L aqueous solution with a constant
output atomizer (TSI Instruments), dried by a drierite denuder,
and neutralized using two 210Po static eliminators (Nuclecel, 10
mCi) in series. To minimize the contribution of organic
background from the seed, the aqueous (NH4)2SO4 solution
was extracted five times with high-purity dichloromethane
(Omnisolv, VWR) and boiled for 4 h to remove any residual
solvent. Seed surface area concentrations were high enough
(1000−1800 μm2/cm3 at ∼30 μg/m3) to inhibit new particle
formation within the chamber. RONO, C6F6, NO, and seed
particles were allowed to mix in the dark chamber for 30 min
before the blacklights were turned on, initiating photochemistry.
Data collection continued until particle growth finished.
RESULTS
■
Aerosol Formation and Yields. Results for a representative
experiment, the photolysis of 200 ppbv of decyl-3-nitrite, are
shown in Figure 3. When the lights are turned on, RONO decays
exponentially, with a decay constant of 2.1 × 10−4 s−1. Photolysis
rates for all nitrites are similar, with an average decay constant of
2.5 × 10−4 s−1 for all species studied. Significant GC interferences
Aerosol Composition. Particle composition was measured
in real time using an Aerodyne high-resolution aerosol time-of-
flight mass spectrometer (AMS),13 run in “V mode” (resolving
power of ∼3000) to maximize sensitivity. Time series for the
major aerosol families (sulfate, organic, nitrate) were generated
from high-resolution fitting of individual mass spectra, and
elemental ratios of the organics (O/C, H/C, N/C) were
determined using the approach described in Aiken et al.14,15
Filter blanks were taken for 10 min at the start and end of every
experiment to measure and correct for the gas-phase CO2+ signal
within the AMS. All N from NO+ and NO2+ were included in the
calculation of N/C, because any nitrate formed is likely to be
from organonitrates; however, only O atoms directly bonded to
the carbon were included in the determination of O/C.
Figure 3. Time-dependent results from a representative alkoxy radical
experiment, using decyl-3-nitrite as the precursor RONO species. GC-
FID measurements of RONO concentrations (scaled by 1/50) are
shown as black squares, with the exponential fit to the data as a dashed
line. The organic aerosol mass concentration (cOA), corrected for wall
loss and dilution, is shown in green. Elemental ratios (O/C, H/C, and
N/C) exhibit low signal-to-noise at low mass loadings and so are shown
only after ∼4 μg/m3 of OA is formed.
Yield Calculations. The time-dependent SOA yield (Y) is
given by eq 2:
8809
dx.doi.org/10.1021/jp506562r | J. Phys. Chem. A 2014, 118, 8807−8816

The Journal of Physical Chemistry A
Article
from gas-phase products with retention times similar to that of
RONO preclude the direct measurement of RONO concen-
tration after 3−5 h, so fitted values (dotted line in Figure 3) were
used for all values of ΔRONO. Roughly 25 min after irradiation is
initiated, OA mass concentration (cOA) begins increasing. A
significant amount of SOA is formed, with corrected loadings
reaching a plateau of ∼25 μg/m3 after 4−5 h. Also shown in
Figure 3 are the elemental ratios of the SOA formed; these vary
little over the course of the experiment and are discussed in detail
in a later section.
SOA formation was observed for all alkoxy radical species
studied, with the exception of the 3,4-diethyl-3-hexyloxy radical.
Yield curves (plots of aerosol yield as a function of cOA) for all
alkoxy radicals that formed SOA are shown in Figure 4. Yields
The mass-based yields for each bin (α1, α10, and α100) are listed in
Table 1, both for the individual alkoxy radicals and for averages of
all secondary and branched radicals.
Aerosol Composition. Elemental ratios (O/C, H/C, and
N/C) of the SOA generated from all alkoxy radical species
(averaged over the last 2 h of each experiment) are also given in
Table 1. Because our determination of N/C is dominated by the
+
NO+ and NO2 ions, here N/C is a measure of organonitrate
abundance. Similarly, because reported O/C ratios do not
include the two terminal O atoms bonded to the nitrate-N,23 they
account only for oxygen atoms directly bonded to the carbon
skeleton, and thus represent a measure of the degree of carbon
oxidation.
As illustrated in Figure 3, elemental ratios do not vary
dramatically over the course of an experiment; on average, O/C
and N/C values increased slightly (by <10% and ∼5%
respectively) while H/C decreased (by ∼3−5%) over 2−3 h of
irradiation. This absence of “aging” is expected for the present
system, whose chemistry is dominated by a single rate-limiting
reaction step (the photolysis of the RONO species); the
subsequent drift toward slightly more oxidized products may
result from secondary oxidation reactions (described below) or
slow repartitioning effects that result from dilution of the
chamber.
SOA formed from the primary alkoxy radical is less oxidized
(O/C = 0.19, H/C = 1.62) than SOA formed from secondary
and branched alkoxy radicals (average values: O/C = 0.25, H/C
= 1.63). This corresponds to differences in the degree of
functionalization of the carbon skeleton: assuming that
condensed species retain their C10 carbon skeleton (i.e., that
the SOA includes no fragmentation products), SOA from the
primary alkoxy radical contains on average two oxygen-
containing moieties per molecule, whereas the SOA from
secondary and branched alkoxy radicals can have up to three such
moieties per molecule on average. At the same time, the SOA
formed from isomers with their radical centers closest to the
terminus of the molecules tended to have the highest abundances
of organonitrates (N/C of 0.05 for decyl-1-oxy and decyl-2-oxy
vs ∼0.03 for all others). Because SOA from these isomers also
had the lowest O/C values, nitrates appear to account for a
relatively large fraction of the functional groups in the less-
oxidized SOA.
Figure 4. Yield % curves as a function of cOA (μg/m3) for the C10 alkoxy
radical species studied. Curves are colored by structural type: red
indicates the primary radical (1: decyl-1-oxy); blue the secondary
radicals (2: decyl-2-oxy, 3: decyl-3-oxy, 4: decyl-4-oxy, and 5: decyl-5-
oxy); and black the branched radicals (A: 2-methyl-2-nonyl-oxy, B: 2-
methyl-4-nonyl-oxy, C: 2-methyl-5-nonyl-oxy, D: 5-methyl-5-nonyl-
oxy, and E: 4-methyl-5-nonyl-oxy). No aerosol formation was observed
from the 3,4-diethyl-3-hexyl-oxy radical. The SOA yield of the analogous
reaction, n-decane + OH, as measured by Presto et al.18 is shown as the
green diamond. Yields from additional studies19−21 are not shown here
as they were conducted at higher cOA levels.
DISCUSSION
■
Reaction Mechanism. Reaction pathways available to the
alkoxy radicals studied in these experiments are shown in Figure
5. The top panel in Figure 5 shows the initial gas-phase reactions
possible: the two key channels are isomerization (green arrows),
which forms a hydroxyalkyl radical, and fragmentation (red
arrows), which forms a carbonyl and an alkyl radical. A third
channel, the reaction with O2 to form a carbonyl and HO2, is
expected to be negligible for straight-chain species but can be
significant for some of the branched isomers.24 The isomer-
ization pathway tends to form multifunctional, lower-volatility
products (generally promoting SOA formation), whereas
fragmentation forms two smaller, relatively volatile species
(generally inhibiting SOA formation).10 Isomerization forms a
peroxy radical, which will react with NO to form either an alkyl
nitrate (blue arrow) or another alkoxy radical. This newly formed
alkoxy radical will mostly “back-isomerize” to form a
hydroxycarbonyl, though some fraction (∼10%) will instead
“forward-isomerize”, ultimately leading to the formation of a
dihydroxycarbonyl or dihydroxynitrate.19,24
(Y) from the three groups of alkoxy radicals decreased from
Yprimary > Ysecondary > Ybranched. SOA yields for the straight-chain
species generally decreased as the radical center was located
closer to the center of the carbon skeleton (Ydecyl‑1‑oxy > Ydecyl‑2‑oxy
> Ydecyl‑4‑oxy > Ydecyl‑3‑oxy > Ydecyl‑5‑oxy). The branched alkoxy
radicals exhibited lower yields than the straight-chain isomers,
with yields showing no obvious trend with the location of the
branching methyl group relative to the radical center (α, β, γ, δ).
No aerosol formation was observed from the α,β-disubstituted
isomer (3,4-diethyl-3-hexyloxy). These results are examined in
detail in the Discussion.
Product volatility distributions were derived from yield curves
using the volatility basis set approach,22 with fixed saturation
vapor concentration (c*) bins of 1, 10, and 100 μg/m3. These
bins span the range of cOA accessed in this study, and thus
including additional bins did not improve the fits significantly.
8810
dx.doi.org/10.1021/jp506562r | J. Phys. Chem. A 2014, 118, 8807−8816

The Journal of Physical Chemistry A
Article
a
Table 1. Summary of SOA Yield Parameters and Elemental Ratio Measurements
RO isomer
decyl-1-oxy
substitution
primary
O/C
H/C
N/C
α1 (c* = 1 μg/m3) α2 (c* = 10 μg/m3) α3 (c* = 100 μg/m3)
0.19
1.62
0.05
0.33
2.15
7.95
decyl-2-oxy
secondary
secondary
secondary
secondary
0.23
0.32
1.60
1.61
0.05
0.03
0.23
0.27
0.47
0.58
9.55
6.25
decyl-4-oxy
decyl-3-oxy
0.32
1.61
0.03
0.22
0.00
7.89
decyl-5-oxy secondary
secondary average
0.23
1.70
0.04
0.20
0.90
3.07
0.28 0.05
1.63 0.07
0.04 0.01
0.23 0.03
0.49 0.37
6.69 2.77
2-methyl-2-nonyloxy
2-methyl-4-nonyloxy
2-methyl-5-nonyloxy
5-methyl-5-nonyloxy
4-methyl-5-nonyloxy
α-substituted
γ-substituted
δ-substituted
α-substituted
β-substituted
α,β-disubstituted
0.23
0.28
1.64
1.61
0.04
0.09
0.11
0.13
0.00
6.67
5.99
0.03
0.25
1.66
0.03
0.11
0.00
5.99
0.21
1.66
0.03
0.14
0.00
4.76
0.25
1.64
0.03
−
0.06
0.39
2.68
b
3,4-diethyl-3-hexyloxy
−
−
−
−
−
c
branched average
0.24 0.03
1.64 0.02
0.03 5 × 10−3
0.10 0.03
0.10 0.17
5.22 1.58
a
Species are listed in order of decreasing yields. Mass-based yields (α) were derived using a three-product fitting with the volatility basis set for c*
b
c
(μg/m3) = {1, 10, 100}. Averages are reported as 1σ. SOA formation not observed. Average of values for all branched alkoxy radicals except 3,4-
diethyl-3-hexyloxy.
Figure 5. Formation of first-generation products from a single alkoxy radical species. Shown are the major gas-phase (top panel) and condensed phase/
heterogeneous (bottom panel) reactions and products formed from a representative alkoxy radical, decyl-2-oxy. Alkoxy radicals will fragment (red
arrow) or isomerize (green arrow); isomerization leads to the formation of a peroxy radical, which in the presence of NO can form an organonitrate
(blue arrow) or a new alkoxy radical, opening up additional isomerization and fragmentation channels.
The lower panel of Figure 5 shows the subsequent
heterogeneous and condensed-phase reactions available to
(di)hydroxycarbonyls, which can undergo heterogeneous
cyclization to form (hydroxy)cyclic hemiacetals.10,11,19 Such
compounds can dimerize to form hemiacetal dimers or
dehydrate to form a (hydroxy)dihydrofuran. This latter
compound is volatile and will likely partition into the gas
phase, where (if oxidants are present) it will rapidly react to form
second-generation oxidation products.25
High molecular weight (m/z > 150 Da) ion fragments were
observed in all mass spectra, and the chemical formulas of major
ions were determined by high-resolution analysis (Figure 6). All
8811
dx.doi.org/10.1021/jp506562r | J. Phys. Chem. A 2014, 118, 8807−8816

The Journal of Physical Chemistry A
Article
Figure 6. High-mass ions (m/z 150−235) in the AMS spectrum of SOA formed from decyl-1-nitrite photolysis. Ion formulas given were confirmed by
high-resolution analysis; tentative molecular assignments are from Lim and Ziemann.19
+
such ions have 10 carbon atoms, and most are the C10 analogs of
ions previously observed in alkane + OH experiments by Lim and
Ziemann;19 the exact masses measured using high-resolution
mass spectrometry confirm the identification of the ion formulas
suggested in that study. Three of the most abundant fragments
correspond to species with 10 carbon atoms and 1−2 oxygen
dimer (m/z 171.138), respectively. The C10H17O2 ion (m/z
169.122) has previously been identified as the cyclichemiacetal
nitrate, a second-generation oxidation product.11 This may
indicate the importance of multigenerational oxidation pro-
cesses, as discussed below, though this fragment ion may not be
unique to this one product. Furthermore, several prominent N-
containing ions of m/z > 200 are present, pointing to the
presence of organonitrate species; the detailed molecular
assignments for such species are unclear at present. The absence
of abundant ions corresponding to the hydroxycarbonyl,
hydroxynitrate, and dihydrofuran (which would be found at
m/z 154, 138, and 140 respectively)19 is likely due to the high
volatilities of such species. Unfortunately, low ion intensities and
unknown ionization efficiencies preclude the detailed inter-
pretation of the high-mass ions in the AMS spectra, and in
particular their relative product yields in different alkoxy radical
experiments. Nonetheless, the observation of ions similar to
those observed by Lim and Ziemann19 provides strong
confirmation that the present experiments access the chemistry
important in the early generation oxidation of alkanes.
atoms, m/z 155.143 (C10H19O+), 169.122 (C10H17O2 ), and
+
+
171.138 (C10H19O2 ). The time series of each fragment from
representative primary, secondary, and branched nitrite
photolysis experiments are shown in Figure 7. Fragment
Role of Multigenerational Oxidation. The mechanism
shown in Figure 5 includes no oxidation of the reaction products;
in standard (oxidant-initiated) studies of SOA, reaction products
will react further with oxidants, leading to a complex mixture of
multigenerational products. By contrast, in the present chemical
system (the photolytic generation of radicals), all chemistry is
initiated without oxidants, allowing for individual oxidation steps
to be studied in greater detail than in oxidant-initiated studies.
Nonetheless, some additional oxidation may occur in the present
studies via the secondary generation of oxidants. Oxidation by
NO3 and O3 is unlikely, due to the high (ppm) levels of NO
present. However, because HO2 can be formed via isomerization
reactions (Figure 5), some OH formation may occur via the HO2
+ NO reaction. Though some of the OH formed will react with
NO2 to form HNO3, most will likely react with the organic
products formed (and early in the reaction, possibly even the
nitrite as well). The importance of OH-initiated secondary
chemistry can be examined from the behavior of the C10H19O+
(m/z 155.143) ion, previously assigned to the cyclic hemiacetal
species.11,25 Because this compound is in equilibrium with the
highly reactive dihydrofuran (Figure 5), the presence of oxidants
will lead to a rapid depletion of that ion. However, as shown in
Figure 7, no such depletion is observed, in stark contrast to
results from alkane oxidation studies.11,25 Although this might
indicate that OH-initiated chemistry is unimportant in the
present system, it is also possible that the stable concentration of
Figure 7. Ion time series for three of the major ions, m/z 155.143,
169.122, and 171.138, for a representative primary (decy-1-nitrite),
secondary (decyl-3-nitrite), and branched (2-methyl-4-nonylnitrite)
experiment. Traces are shown as a fraction of the total organic mass.
abundances (shown as the fraction of the organic signal)
increase immediately following irradiation and level off after 1.5−
+
3 h, with the exception of C10H19O2 , which peaks and then
decreases slightly.
Although definitive assignments of the parent molecular
product species are not possible using the AMS, tentative
assignments can be made on the basis of the work of Lim and
Ziemann.19 The C10H19O+ and C10H19O2+ ions may correspond
to two major first-generation products, the cyclic hemiacetal (m/
z 155.143) and the hydroxy-cyclic hemiacetal or hemiacetal
8812
dx.doi.org/10.1021/jp506562r | J. Phys. Chem. A 2014, 118, 8807−8816

The Journal of Physical Chemistry A
Article
the species arises from the absence of sufficient HNO3 to catalyze
the dehydration of the cyclichemiacetal. Furthermore, the
assignment of the C10H19O+ ion to the cyclichemiacetal is
tentative, and this ion instead may correspond to some other
studied. Such calculations take into account the fate of not only
the original alkoxy radical but also any alkoxy radical formed later
in the reaction sequence (Figure 5); calculation details are given
in the Supporting Information.
The calculated product distributions for the major first-
generation products of each alkoxy radical isomer are shown in
Figure 8. For most of the alkoxy radicals studied, hydroxynitrates
+
product. By contrast, the C10H19O2 ion (m/z 171.138) does
indeed decrease to some extent; this could indicate oxidation of
the hydroxy-cyclic hemiacetal, though it may instead be a result
of nonoxidative chemistry of the hemiacetal dimer.
Differences in SOA yields and properties measured in the
present study and those in the analogous n-decane + OH
reaction18,20,21,26 provide additional insight into the role of
multigenerational oxidation in our experiments. In particular,
SOA yields from OH + n-decane are substantially higher than
those from the RONO photolysis. Shown in Figure 4 is the SOA
yield measured by Presto et al.18 (green diamond). This yield falls
between those measured in this study from decyl-1-oxy and
decyl-2-oxy radicals, and thus is substantially higher than the
yields inferred from the RONO experiments (because these two
radicals together account for only ∼24% of the first alkoxy radical
isomers formed in the n-decane + OH reaction27). The difference
in aerosol formation between these two studies is even more
pronounced than is implied from these yield differences, given
that decyl nitrate, a volatile first-generation product from OH +
n-decane, is not formed from RONO photolysis. Studies of OH +
n-decane carried out at higher OH exposures than accessed by
Presto et al.18 found even higher SOA yields (>10%),20,21
consistent with additional OH oxidation leading to more SOA
formation (the experiments from Lim and Ziemann19 were
carried out at much higher cOA, preventing a direct comparison
with our results here). Similar differences are seen in the degree
of oxidation of SOA; O/C ratios measured in this work are lower
(0.19−0.32) on average than those from OH + n-decane (0.33−
0.35),21,26 whereas H/C ratios are higher (1.6−1.7 versus 1.5−
1.6).21,26 These results all suggest that, relative to OH-initiated
SOA studies, multigenerational oxidation chemistry is sup-
pressed in the present experiments, with the formation of SOA
that is more heavily weighted toward first-generation oxidation
products.
Effect of Alkoxy Radical Structure on SOA Formation.
The large variability in yields and chemical composition of SOA
formed from various C10 alkoxy radical isomers (Figure 3 and
Table 1) underscores the high degree of sensitivity of SOA
formation to the exact chemical structure of the precursor
species. The location and degree of substitution of the alkoxy
radical center appears to have a governing effect on observed
yields this is particularly evident in the large differences in yields
among the various radicals with straight-chain carbon skeletons
and between the structurally similar 2-methyl-2-nonyloxy and 5-
methyl-5-nonyloxy radicals. As described below, differences in
the yield and composition of SOA from the different isomeric
alkoxy radicals can be explained in terms of two general factors,
the branching between fragmentation and isomerization of the
alkoxy radicals (including those formed subsequent to isomer-
ization) and the relative volatilities of the different isomeric
products formed from each radical.
Figure 8. Distribution of major first-generation products from each
alkoxy radical isomer. Isomers are listed in order of decreasing yields,
and the identifiers in parentheses correspond to those used in Figure 4.
Product distributions are estimated using structure−activity relation-
ships, as described in the Supporting Information.
and hydroxycarbonyls are the major products. The predicted
product distributions from decyl-1-oxy, decyl-2-oxy, and decyl-3-
oxy are identical. For these species, the formation of
fragmentation products is negligible (<3%), whereas products
arising from multiple isomerization reactions (dihydroxycarbon-
yls and dihydroxynitrates) represent a small but significant
(∼10%) fraction of the total products. The predicted products
from the other two straight-chain radicals (decyl-4-oxy and decyl-
5-oxy) are similar, though they do not include the dihydroxy
species, because “forward-isomerization” reactions either cannot
occur (due to the lack of H atoms in the δ -position) or are highly
unfavorable (due to the δ-H atoms residing on methyl groups).
The lack of formation of these low-volatility products may
contribute somewhat to their lower SOA yields. However,
because the SOA yield from the decyl-4-oxy radical is somewhat
greater than that from the decyl-3-oxy radical, this appears to be a
minor effect.
Estimated product distributions from the δ- and γ-substituted
branched species (2-methyl-5-nonyloxy and 2-methyl-4-non-
yloxy, respectively) are similar to those of the straight-chain
species, because the branch points are located relatively far from
the alkoxy radical center and thus have little influence on the
chemistry. The other branched alkoxy radicals, however, are
predicted to form very different sets of products; these can
include fragmentation products, which can at least partly explain
their lower SOA yields. The two α-substituted branched species
(2-methyl-2-nonyloxy and 5-methyl-5-nonyloxy), despite being
quite similar structurally, have dramatically different product
distributions, due to the number of isomerization reactions
available to each. 2-Methyl-2-nonyloxy is unable to form a
Alkoxy Radical Products. The exact chemical structure of
each alkoxy radical has a governing influence on the reactive
pathways available to it, which in turn can affect the amount and
composition of SOA formed.5,10,11 Using a structure−activity
relationship (SAR) to describe alkoxy radical reactivity,24 and in
particular the branching between isomerization (functionaliza-
tion) and C−C bond scission (fragmentation), we calculated the
distribution of first-generation products from each alkoxy radical
8813
dx.doi.org/10.1021/jp506562r | J. Phys. Chem. A 2014, 118, 8807−8816

The Journal of Physical Chemistry A
Article
hydroxycarbonyl, because the α-substituted methyl group
prevents “back-isomerization”; however, it is able to undergo a
second “forward-isomerization” that 5-methyl-5-nonyloxy can-
not. Instead, the second alkoxy radical formed from 5-methyl-5-
nonlyoxy will react with O2 to form a significant amount of
hydroxycarbonyl species (∼50%). The result is that the products
of the 2-methyl-2-nonyloxy radical are dominated by the low-
volatility dihydroxy species (∼80%), whereas those from 5-
methyl-5-nonyloxy are fragmentation products (25%), alkyl
nitrates (∼25%), or hydroxycarbonyls (∼50%) formed from the
RO + O2 pathway. This is consistent with 2-methyl-2-nonyloxy
having the highest yields among the branched isomers, and 5-
methyl-5-nonyloxy among the lowest.
The lowest SOA yields measured were from the β-substituted
species, 4-methyl-5-nonyloxy (mono-β-substituted) and 3,4-
diethyl-3-hexyloxy (α,β-disubstituted, for which no SOA
formation was observed), suggesting a high degree of radical
fragmentation in those cases. However, the amount of
fragmentation predicted by the SAR is relatively modest (4%
from 4-methyl-5-nonyloxy and 30% from 3,4-diethyl-3-hexyl-
oxy). This difference suggests that fragmentation of β-substituted
alkoxy radicals may be underestimated using the SAR, which
estimates fragmentation based upon reaction enthalpy;24 this
may result from the lack of accurate values of reaction enthalpies
of secondary β-substituted alkoxy radicals. An alternate SAR for
describing alkoxy radical fragmentation, based on the degree of
substitution of the α- and β-carbon atoms,28 predicts
substantially more fragmentation (57% and 94% from 4-
methyl-5-nonyloxy and 3,4-diethyl-3-hexyloxy, respectively).
This is in better agreement with the SOA trends observed,
though fragmentation of 4-methyl-5-nonyloxy may still be
underestimated. Moreover, this SAR also predicts increased
fragmentation of α-substituted radicals, which is not consistent
with the relatively high SOA yields from those species
(Supporting Information). Both SARs are based upon measure-
ments of fragmentation rates of relatively small alkoxy radicals;
this work suggests the need for the direct study of the fate of
larger, more substituted radicals to better constrain such
descriptions of alkoxy radical chemistry.
straight-chain radicals, yields tend to increase as the alkoxy
radical center is located closer to the terminus of the molecule.
(The one exception is decyl-3-oxy, which has lower yields than
decyl-4-oxy; however, such a difference is small.) This effect
likely arises from the fact that functional groups located closer to
the terminus of molecules generally have larger influences on
vapor pressure, because the molecules can “stack” more
efficiently. For example, for straight-chain C9 alcohols, the
vapor pressure of 1-nonanol is 5 times lower than that of 2-
nonanol, which in turn is 1.4 times lower than the vapor pressures
of the remaining isomers (the differences among 3-nonanol, 4-
nonanol, and 5-nonanol are much smaller, under 5%).29 The
relative ordering of SOA yields among the straight-chain alkoxy
radicals is broadly consistent with this trend, pointing to the role
of the volatilities of product isomers on SOA formation. In
addition, the formation of aldehydes (terminal carbonyls), which
have slightly lower volatilities than ketones formed by secondary
alkoxy radicals,1 and which are also more susceptible to
oligomerization reactions (such as hemiacetal formation),30
further decreases volatility. Because the primary decyl-1-oxy
species is the only isomer capable of forming aldehydes, its
products are lower in volatility, explaining its higher values of α1/
α100 relative to those from the secondary and branched isomers.
Similarly, the structure of the carbon skeleton may play a role,
because branched species generally have higher volatilities than
straight-chain isomers. This “P25 effect”10 may contribute to the
low yields from branched alkoxy radicals; however, differences in
alkoxy radical fragmentation (as described above) are likely the
dominant factor for most of the isomers studied.
Such differences in volatilities of different product isomers may
also explain some of the trends in O/C ratios, particularly among
the straight-chain radicals, because products from the decyl-1-
oxy and decyl-2-oxy are lower in volatility relative to
corresponding isomers from the other alkoxy radicals. This
enhances the gas-to-particle partitioning of less oxidized (less
functionalized) products, resulting in lower O/C of the aerosol.31
It may also explain the increased abundance of organonitrate
species, evident by higher N/C, in SOA that is the least oxidized
(lowest O/C).
Volatility of SOA Components. As described above, the
extent of alkoxy radical fragmentation and isomerization are in
good qualitative agreement with many of the observed
differences in yields, particularly for the branched species.
However, the differences in yields among the straight-chain
species (which are predicted to have very similar product
distributions) and the relatively low O/C values for decyl-1-oxy
and decyl-2-oxy, are not captured by this explanation alone,
pointing to additional factors controlling SOA formation. These
appear to be related to differences in the volatilities of the
(isomeric) products formed. As shown in Table 1, SOA from the
secondary and branched radicals exhibit volatility distributions
that are skewed toward higher c* relative to the SOA from
primary radicals, as reflected in differences in both α1/α100 and
α10/α100 ratios. This effect can also be seen in Figure 4, which
shows that the relative differences among yields from different
CONCLUSIONS
■
In the present work we highlight several important structural
features of alkoxy radical structure that influence the SOA-
forming potential of straight-chain and branched alkanes. The
use of photolytically generated alkoxy radicals allows for the
isolation of individual reaction pathways and oxidation
generations, permitting the examination of the role of chemical
structure in more detail than is typically possible in oxidant-
initiated SOA studies. Measurements of SOA yields and
composition broadly agree with previous work on the OH
initiated oxidation of alkanes, indicating the relevance of this
work to that important chemical system. However, both SOA
yields and chemical composition are found to vary dramatically
with alkoxy radical structure, even for those sharing the same
carbon skeleton (e.g., the five unbranched isomers, decyl-1-oxy
through decyl-5-oxy). Such differences can largely be explained
in terms of (1) the complex branching between reaction
pathways available to the alkoxy radical and (2) the role of
functional group position and carbon skeleton on the volatility of
the oxidation products. These SOA measurements also provide
some insight into the role of structure on alkoxy radical reactivity;
in particular, the very low SOA yields from the β-substituted
radical types are most pronounced at low values of cOA
.
Such differences in SOA volatility, even when formation
occurs via similar chemistry (i.e., from straight-chain isomers),
likely arise from differences in vapor pressures of the isomeric
product species. Although the various precursor alkoxy radicals
and their products (hydroxycarbonyls, hydroxynitrates, etc.)
have the same chemical formulas, their spatial arrangement can
lead to large differences in their relative volatilities. Among the
8814
dx.doi.org/10.1021/jp506562r | J. Phys. Chem. A 2014, 118, 8807−8816

The Journal of Physical Chemistry A
Article
from the Direct Photolytic Generation of Organic Radicals. J. Phys.
Chem. Lett. 2011, 2, 1295−1300.
alkoxy radicals suggest more fragmentation than is predicted by
commonly used structure−activity relationships.28,32
(7) Morabito, P.; Heicklen, J. Primary Photochemical Processes in the
Photolysis of Alkyl Nitrites at 366nm and 23C in the Presence of 15NO.
Int. J. Chem. Kinet. 1985, 17, 535−546.
More generally, the study of aerosol formation from single
radical species enables the importance of individual reaction
pathways and products in SOA formation to be examined in
detail. For systems whose yields of condensable species are low
(such as the first-generation oxidation of C10 hydrocarbons,
studied here), products that are formed in small yields may still
play a major role in SOA formation. For example, for the straight-
chain structures studied, the primary radical (decyl-1-oxy) was
found to form more SOA than any of the secondary ones. In the
analogous OH + n-decane reaction, the formation of this radical
(via abstraction at the terminal carbon) accounts for only ∼4% of
the total reaction;27 however, the fractional contribution of this
channel to the total SOA formed is likely to be substantially
higher than this. Ultimately, the identification of these key
channels and their products is necessary for relating the amounts
and properties of SOA to the structure of a given precursor
compound.
(8) Heicklen, J. The Decomposition of Alkyl Nitrites and the Reaction
of Alkoxy Radicals. Adv. Photochem. 1988, 14, 177−272.
(9) Orlando, J. J.; Tyndall, G. S.; Bilde, M.; Ferronato, C.; Wallington,
T. J.; Vereecken, L.; Peeters, J. Laboratory and Theoretical Study of the
Oxy Radicals in the OH- and Cl-Initiated Oxidation of Ethene. J. Phys.
Chem. A 1998, 102, 8116−8123.
(10) Lim, Y. B.; Ziemann, P. J. Effects of Molecular Structure on
Aerosol Yields from OH Radical-Initiated Reactions of Linear,
Branched, and Cyclic Alkanes in the Presence of NOx. Environ. Sci.
Technol. 2009, 43, 2328−2334.
(11) Lim, Y. Bin; Ziemann, P. J. Products and Mechanism of Secondary
Organic Aerosol Formation from Reactions of n-Alkanes with OH
Radicals in the Presence of NOx. Environ. Sci. Technol. 2005, 39, 9229−
9236.
(12) Noyes, W. A. Explanation of the Formation of Alkyl Nitrites in
Dilute Solutions; Butyl and Amyl Nitrites. J. Am. Chem. Soc. 1933, 55,
3888−3889.
(13) DeCarlo, P. F.; Kimmel, J. R.; Trimborn, A.; Northway, M. J.;
Jayne, J. T.; Aiken, A. C.; Gonin, M.; Fuhrer, K.; Horvath, T.; Docherty,
K. S.; et al. Field-Deployable, High-Resolution, Time-of-Flight Aerosol
Mass Spectrometer. Anal. Chem. 2006, 78, 8281−8289.
(14) Aiken, A. C.; DeCarlo, P. F.; Jimenez, J. L. Elemental Analysis of
Organic Species with Electron Ionization High-Resolution Mass
Spectrometry. Anal. Chem. 2007, 79, 8350−8358.
(15) Aiken, A. C.; Decarlo, P. F.; Kroll, J. H.; Worsnop, D. R.; Huffman,
J. A.; Docherty, K. S.; Ulbrich, I. M.; Mohr, C.; Kimmel, J. R.; Sueper, D.;
et al. O/C and OM/OC Ratios of Primary, Secondary, and Ambient
Organic Aerosols with High-Resolution Time-of-Flight Aerosol Mass
Spectrometry. Environ. Sci. Technol. 2008, 42, 4478−4485.
(16) Hildebrandt, L.; Donahue, N. M.; Pandis, S. N. High Formation of
Secondary Organic Aerosol from the Photo-Oxidation of Toluene.
Atmos. Chem. Phys. 2009, 9, 2973−2986.
(17) Zhang, X.; Cappa, C. D.; Jathar, S. H.; McVay, R. C.; Ensberg, J. J.;
Kleeman, M. J.; Seinfeld, J. H. Influence of Vapor Wall Loss in
Laboratory Chambers on Yields of Secondary Organic Aerosol. Proc.
Natl. Acad. Sci. U. S. A. 2014, 111, 5802−5807.
(18) Presto, A. A.; Miracolo, M. A.; Donahue, N. M.; Robinson, A. L.
Secondary Organic Aerosol Formation from High-NOx Photo-
Oxidation of Low Volatility Precursors: n-Alkanes. Environ. Sci. Technol.
2010, 44, 2029−2034.
(19) Lim, Y. B.; Ziemann, P. J. Chemistry of Secondary Organic
Aerosol Formation from OH Radical-Initiated Reactions of Linear,
Branched, and Cyclic Alkanes in the Presence of NOx. Aerosol Sci.
Technol. 2009, 43, 604−619.
(20) Lambe, A. T.; Onasch, T. B.; Croasdale, D. R.; Wright, J. P.;
Martin, A. T.; Franklin, J. P.; Massoli, P.; Kroll, J. H.; Canagaratna, M. R.;
Brune, W. H.; et al. Transitions from Functionalization to
Fragmentation Reactions of Laboratory Secondary Organic Aerosol
(SOA) Generated from the OH Oxidation of Alkane Precursors.
Environ. Sci. Technol. 2012, 46, 5430−5437.
(21) Hunter, J. F.; Carrasquillo, A. J.; Daumit, K. E.; Kroll, J. H.
Secondary Organic Aerosol Formation from Acyclic, Monocyclic, and
Polycyclic Alkanes. Environ. Sci. Technol. 2014, 48, 10227−10234.
(22) Donahue, N. M.; Robinson, A. L.; Stanier, C. O.; Pandis, S. N.
Coupled Partitioning, Dilution, and Chemical Aging of Semivolatile
Organics. Environ. Sci. Technol. 2006, 40, 2635−2643.
(23) Farmer, D. K.; Matsunaga, A.; Docherty, K. S.; Surratt, J. D.;
Seinfeld, J. H.; Ziemann, P. J.; Jimenez, J. L. Response of an Aerosol
Mass Spectrometer to Organonitrates and Organosulfates and
Implications for Atmospheric Chemistry. Proc. Natl. Acad. Sci. U. S. A.
2010, 107, 6670−6675.
(24) Atkinson, R. Rate Constants for the Atmospheric Reactions of
Alkoxy Radicals: An Updated Estimation Method. Atmos. Environ. 2007,
41, 8468−8485.
ASSOCIATED CONTENT
■
S
* Supporting Information
Details of the branching ratio calculation (as shown in Figure 8),
including a figure of radical branching, tables of isomerization
and fragmentation rate constants, and a graph of first-generation
product distributions calculated by an alternative SAR, are given
in the Supporting Information. The material is available free of
charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
*Jesse H. Kroll. E-mail: jhkroll@mit.edu.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by the National Science Foundation
under grant number CHE-1012809. We thank Roger Summons
for use of the GC for confirmation of nitrite identities following
synthesis and Paul Ziemann for helpful discussions.
REFERENCES
■
(1) Kroll, J. H.; Seinfeld, J. H. Chemistry of Secondary Organic
Aerosol: Formation and Evolution of Low-Volatility Organics in the
Atmosphere. Atmos. Environ. 2008, 42, 3593−3624.
(2) Hallquist, M.; Wenger, J. C.; Baltensperger, U.; Rudich, Y.;
Simpson, D.; Claeys, M.; Dommen, J.; Donahue, N. M.; George, C.;
Goldstein, A. H.; et al. The Formation, Properties and Impact of
Secondary Organic Aerosol: Current and Emerging Issues. Atmos. Chem.
Phys. 2009, 9, 5155−5236.
(3) Kroll, J. H.; Donahue, N. M.; Jimenez, J. L.; Kessler, S. H.;
Canagaratna, M. R.; Wilson, K. R.; Altieri, K. E.; Mazzoleni, L. R.;
Wozniak, A. S.; Bluhm, H.; et al. Carbon Oxidation State as a Metric for
Describing the Chemistry of Atmospheric Organic Aerosol. Nat. Chem.
2011, 3, 133−139.
(4) Goldstein, A. H.; Galbally, I. E. Known and Unexplored Organic
Constituents in the Earth’s Atmosphere. Environ. Sci. Technol. 2007, 41,
1514−1521.
(5) Ziemann, P. J.; Atkinson, R. Kinetics, Products, and Mechanisms of
Secondary Organic Aerosol Formation. Chem. Soc. Rev. 2012, 41, 6582−
6605.
(6) Kessler, S. H.; Nah, T.; Carrasquillo, A. J.; Jayne, J. T.; Worsnop, D.
R.; Wilson, K. R.; Kroll, J. H. Formation of Secondary Organic Aerosol
8815
dx.doi.org/10.1021/jp506562r | J. Phys. Chem. A 2014, 118, 8807−8816

The Journal of Physical Chemistry A
Article
(25) Lim, Y. Bin; Ziemann, P. J. Kinetics of the Heterogeneous
Conversion of 1,4-Hydroxycarbonyls to Cyclic Hemiacetals and
Dihydrofurans on Organic Aerosol Particles. Phys. Chem. Chem. Phys.
2009, 11, 8029−8039.
(26) Tkacik, D. S.; Presto, A. A.; Donahue, N. M.; Robinson, A. L.
Secondary Organic Aerosol Formation from Intermediate-Volatility
Organic Compounds: Cyclic, Linear, and Branched Alkanes. Environ.
Sci. Technol. 2012, 46, 8773−8781.
(27) Atkinson, R. A Structure-Activity Relationship for the Estimation
of Rate Constants for the Gas-Phase Reactions of OH Radicals with
Organic Compounds. Int. J. Chem. Kinet. 1987, 19, 799−828.
(28) Peeters, J.; Fantechi, G.; Vereecken, L. A Generalized Structure-
Activity Relationship for the Decomposition of (Substituted) Alkoxy
Radicals. J. Atmos. Chem. 2004, 48, 59−80.
(29) Verevkin, S. P.; Schick, C. Vapour Pressures and Heat Capacity
Measurements on the C7−C9 Secondary Aliphatic Alcohols. J. Chem.
Thermodyn. 2007, 39, 758−766.
(30) Jang, M.; Czoschke, N. M.; Lee, S.; Kamens, R. M. Heterogeneous
Atmospheric Aerosol Production by Acid-Catalyzed Particle-Phase
Reactions. Science 2002, 298, 814−817.
(31) Shilling, J. E.; Chen, Q.; King, S. M.; Rosenoern, T.; Kroll, J. H.;
Worsnop, D. R.; DeCarlo, P. F.; Aiken, A. C.; Sueper, D.; Jimenez, J. L.;
et al. Loading-Dependent Elemental Composition of α-Pinene SOA
Particles. Atmos. Chem. Phys. 2009, 9, 771−782.
(32) Atkinson, R. Rate Constants for the Atmospheric Reactions of
Alkoxy Radicals: An Updated Estimation Method. Atmos. Environ. 2007,
41, 8468−8485.
8816
dx.doi.org/10.1021/jp506562r | J. Phys. Chem. A 2014, 118, 8807−8816