ACPAtmospheric Chemistry and PhysicsACPAtmos. Chem. Phys.1680-7324Copernicus PublicationsGöttingen, Germany10.5194/acp-17-11605-2017Secondary organic aerosol from atmospheric photooxidation of indoleMontoya-AguileraJuliaHorneJeremy R.HinksMallory L.FlemingLauren T.https://orcid.org/0000-0001-6495-6261PerraudVéroniqueLinPenghttps://orcid.org/0000-0002-3567-7017LaskinAlexanderhttps://orcid.org/0000-0002-7836-8417LaskinJuliaDabdubDonaldhttps://orcid.org/0000-0002-5130-4122NizkorodovSergey A.nizkorod@uci.eduhttps://orcid.org/0000-0003-0891-0052Department of Chemistry, University of California, Irvine, CA 92697, USADepartment of Mechanical and Aerospace Engineering, University of California, Irvine, CA 92697, USADepartment of Chemistry, Purdue University, West Lafayette, IN 47907, USASergey A. Nizkorodov (nizkorod@uci.edu)28September20171718116051162122March201723August201720August201728March2017This work is licensed under the Creative Commons Attribution 3.0 Unported License. To view a copy of this licence, visit https://creativecommons.org/licenses/by/3.0/This article is available from https://acp.copernicus.org/articles/17/11605/2017/acp-17-11605-2017.htmlThe full text article is available as a PDF file from https://acp.copernicus.org/articles/17/11605/2017/acp-17-11605-2017.pdf
Indole is a heterocyclic compound emitted by various plant species under
stressed conditions or during flowering events. The formation, optical
properties, and chemical composition of secondary organic aerosol (SOA)
formed by low-NOx photooxidation of indole were investigated. The
SOA yield (1.3±0.3) was estimated from measuring the particle mass
concentration with a scanning mobility particle sizer (SMPS) and correcting
it for wall loss effects. The high value of the SOA mass yield suggests that
most oxidized indole products eventually end up in the particle phase. The
SOA particles were collected on filters and analysed offline with UV–vis
spectrophotometry to measure the mass absorption coefficient (MAC) of the
bulk sample. The samples were visibly brown and had MAC values of ∼2m2g-1 at λ=300nm and ∼0.5m2g-1 at λ=400nm, comparable to
strongly absorbing brown carbon emitted from biomass burning. The chemical
composition of SOA was examined with several mass spectrometry methods.
Direct analysis in real-time mass spectrometry (DART-MS) and nanospray
desorption electrospray high-resolution mass spectrometry (nano-DESI-HRMS)
were both used to provide information about the overall distribution of SOA
compounds. High-performance liquid chromatography, coupled to photodiode
array spectrophotometry and high-resolution mass spectrometry
(HPLC-PDA-HRMS), was used to identify chromophoric compounds that are
responsible for the brown colour of SOA. Indole derivatives, such as
tryptanthrin, indirubin, indigo dye, and indoxyl red, were found to
contribute significantly to the visible absorption spectrum of indole SOA.
The potential effect of indole SOA on air quality was explored with an
airshed model, which found elevated concentrations of indole SOA during the
afternoon hours contributing considerably to the total organic aerosol under
selected scenarios. Because of its high MAC values, indole SOA can contribute
to decreased visibility and poor air quality.
Introduction
Atmospheric particulate matter (PM) absorbs and scatters solar
radiation and is responsible for diminished visibility in urban areas and for
global changes in climate. A key component of PM is secondary organic aerosol
(SOA). While air quality model prediction capabilities have improved in
recent years, disagreements between SOA predictions and measurements remain
(Couvidat et al., 2013; Heald et al., 2005; Hodzic et al., 2010; Jiang
et al., 2012; Volkamer et al., 2006). Discrepancies may result from incorrect
or incomplete parameterizations of mechanisms for known SOA precursors, as
well as from unaccounted precursors of SOA. Atmospheric researchers have
investigated in detail the SOA generated from oxidation of basic
anthropogenic and biogenic volatile organic compounds (VOCs), such as
isoprene, monoterpenes, saturated hydrocarbons, and aromatic hydrocarbons.
Much less is known about SOA from nitrogen-containing VOCs, even though such
VOCs are also common in the atmospheric environment and can potentially
provide significant additional pathways for SOA formation. For example,
photooxidation of amines could serve as a possible SOA source (Price et al.,
2014; Silva et al., 2008).
Heterocyclic nitrogen-containing aromatic compounds based on pyrrole,
pyridine, imidazole, indole, diazines, purines, etc., have been detected in
biomass burning emissions (Laskin et al., 2009). Such compounds can also be
emitted by vegetation; for example, indole is produced by wide variety of
plants (Cardoza et al., 2003; De Boer et al., 2004; Gols et al., 1999; McCall
et al., 1993; Turlings et al., 1990; Zhuang et al., 2012). Indole is emitted
in response to physical or herbivore-induced stress (Erb et al., 2015; Frey
et al., 2004; Misztal et al., 2015; Niinemets et al., 2013; Schmelz et al.,
2003; Turlings et al., 2004) and during flowering events (Gentner et al.,
2014). Once emitted, indole performs critical roles in plant ecology, for
example in attracting pollinators (Zito et al., 2015). For decades, indole
and its derivatives (Fig. 1) have been utilized in agriculture, dyes,
perfumes, and pharmaceutical applications. One of the better-known
derivatives of indole is indigo dye (also known as indigotin), which is used
to give jeans their characteristic
deep blue colour.
Studies of maize plants under stress revealed that indole acts as an aerial
priming agent, released before terpenoids (Erb et al., 2015; Schmelz et al.,
2003). For example, Schmelz et al. (2003) examined insect induced volatile
emissions in Zea Mays seedlings and demonstrated direct positive
relationships between jasmonic acid levels and both sesquiterpene and indole
volatile emissions. Additionally, they showed that indole can reach maximal
emission levels during nocturnal herbivory and concluded that indole could
function as an early morning signal for parasitoids and predators searching
for insect hosts and prey. Niinemets et al. (2013) found evidence that
quantitative relationships exist between the severity of biotic stress and
induced volatile emissions, in addition to the previously demonstrated
dose–response relationships for abiotic stresses. Erb et al. (2015) showed
that herbivore induced indole emissions in maize plants precede the release
of mono-, homo-, and sesquiterpenes, supporting the conclusion that indole is
involved in the airborne priming of terpenoids. Different plant stress
mechanisms typically elicit release of the same ubiquitous stress
volatiles, such as indole, and more
stress-specific mono- and sesquiterpene blends (Erb et al., 2015; Gentner
et al., 2014; Niinemets et al., 2013; Schmelz et al., 2003).
Emissions of indole have also been well correlated with monoterpene emissions
during flowering events (Gentner et al., 2014). Ambient measurements
conducted by Gentner et al. (2014) showed that both daytime and nighttime
concentrations of indole at their field site in California's San Joaquin
Valley were similar to or greater than the dominant monoterpene
β-myrcene. The authors stressed the need for future laboratory and
modelling studies on the SOA formation potential of indole and other novel
compounds identified in their study. A later study by Misztal et al. (2015)
used a combination of laboratory experiments, ambient measurements, and
emissions modelling to show that plants emit a wide variety of benzenoid
compounds (including indole) to the atmosphere at substantial rates and that
current VOC inventories underestimate biogenic benzenoid emissions. They
concluded that emissions of benzenoids from plants are likely to increase in
the future due to changes in the global environment and stressed that
atmospheric chemistry models should account for this potentially important
precursor of SOA.
An additional potential source of indole is animal husbandry, but the
emission rate of indole from this source remains uncertain. In concentrated
animal feeding operations (CAFOs), indole is primarily emitted from animal
waste (Yuan et al., 2017) and can contribute significantly to the malodours in
cattle feed yards and swine facilities (Feilberg et al., 2010; Wright et al.,
2005). While Yuan et al. (2017) indicated that indole is emitted from dairy
operations, beef feed yards, sheep feed yards, and chicken feed yards, the
emission rate of indole from these sources was not quantified. Other studies
have quantified the emission rate of indole but only for pig facilities
(Feilberg et al., 2010; Hobbs et al., 2004). The United States Department of
Agriculture (USDA) 2012 census agriculture atlas maps show no hogs or pigs in
the model domain used in this study. Furthermore, Hobbs et al. (2004) showed
only trace emissions of indole from cattle slurry and did not detect indole
from laying hen manure. Thus, emissions of indole from animal husbandry are
not included in this study but should be considered when modelling areas
with active animal husbandry facilities.
Chemical structures, common names, molecular formulas, and nominal
molecular weights for indole and its oxidized
derivatives discussed in this work.
Despite the importance of indole in the atmospheric environment, only a few
studies exist on the mechanism of its photooxidation. Gas-phase oxidation of
indole by OH, O3, and NO3 was previously studied by Atkinson
et al. (1995). They found that indole reacts with OH and NO3 at
collision-limited rates, with rate constants of 1.5×10-10 and
1.3×10-10cm3molec-1s-1, respectively. The
rate for the reaction of indole with O3 (rate constant of 5×10-17cm3molec-1s-1) and the rate of direct
photolysis were found to be too low to compete with the OH and NO3
reactions. Atkinson et al. (1995) observed 2-formylformanilide (Fig. 1) as
the major primary product of oxidation of indole by both O3 and OH.
Oxidation of indole was also studied by Iddon et al. (1971) in
γ-irradiated aqueous solutions, where oxidation by OH was the dominant
reaction mechanism. The reaction produced
3-oxindole, indoxyl red, indirubin, and indigo dye and eventually resulted in
a trimer of 3-oxindole and two
indole molecules as the major products.
Until now, the formation of SOA from indole has not been investigated. One
of the motivations for investigating indole SOA is that it may possess
unusual optical properties. Many of the indole-derived products are brightly
coloured and have distinctive absorption bands in visible spectrum. If these
products are formed during atmospheric oxidation of indole and partition into
aerosol particles, they could contribute to the pool of organic
light-absorbing species. Such organic compounds that absorb radiation
strongly in the near-UV and visible spectral ranges are collectively known as
“brown carbon” in the atmospheric literature (Andreae and Gelencser, 2006;
Laskin et al., 2015).
In this work, we investigate the formation of indole SOA in a smog chamber
and characterize its molecular composition and optical properties. We
incorporate these results into an airshed model with detailed SOA chemistry
to estimate the effect of indole on the total SOA and on the light-absorbing
components of SOA. We show that indole can measurably contribute to SOA
loading even in urban environments, where anthropogenic emissions dominate
over biogenic ones, such as the South Coast Air Basin of California (SoCAB).
Furthermore, we show that indole SOA contains unique strongly absorbing
compounds and can contribute to decreased visibility, especially under
plant-stressed conditions or during flowering events.
Experimental methods
SOA was generated in a 5 m3 Teflon chamber under low relative
humidity (RH <2 %; Vaisala HMT333 probe). No inorganic seed aerosol
was used because it would interfere with offline mass-spectrometric analysis
of SOA. Hydrogen peroxide was introduced into the chamber by evaporation of
a 30 weight percent solution of H2O2 in water (Fisher Scientific)
into a flow of clean air to achieve an initial mixing ratio of 2 part per
million by volume (ppmv). Indole (99 % purity, Sigma-Aldrich) was
dissolved in methanol (LC-MS grade, 99.9 % purity, Honeywell) and was
evaporated into the chamber to obtain an initial mixing ratio of 200 parts
per billion by volume (ppbv), which is equivalent to
960 µgm-3. The injector and inlet lines were heated to
70 ∘C to minimize losses on the surfaces. At room temperature, the
reported vapour pressure of indole is 0.012 mmHg (Das et al., 1993), which
is equivalent to ∼16ppmv. Therefore, most of the injected
indole should have remained in the gas phase although some of it could remain
adsorbed to the injection line and chamber wall surfaces, contributing to the
variability in the SOA yield (see below). The content of the chamber was
mixed with a fan for 10 min following the injection. After mixing was
stopped, UV-B lamps were turned on to initiate the photooxidation. In some
experiments, complete mixing was achieved only after the lamps were turned on
as evidenced by the measured indole concentrations continuing to increase in
the initial photooxidation period. Although mixing was not fast, it was
faster than the timescale of the reaction, so it should not have affected the
SOA mass yield calculations. Throughout the experiment, size and number
concentration of particles were monitored with a scanning mobility particle
sizer (SMPS; TSI 3936). A proton-transfer-reaction time-of-flight mass
spectrometer (PTR-ToF-MS; Ionicon model 8000) monitored the decay of indole
and the formation of volatile photooxidation products. The PTR-ToF-MS had
a resolving power of m/Δm∼5×103 and was operated with
the following settings: drift tube temperature of 60 ∘C, drift tube
voltage of 600 V, field strength of ∼135 Td, and inlet flow of
0.2 SLM. When the SOA particles reached a peak concentration in the chamber,
the UV irradiation was stopped, and the polydispersed particles were
collected on one Teflon filter (47 mm diameter, Millipore FGLP04700)
at 20 Lmin-1 for 3 h. One filter was collected per each chamber
run; therefore, each replicate sample was collected from a separate
experiment run under the same conditions. The amount of the collected SOA
material on each filter was estimated from SMPS data assuming 100 %
collection efficiency by the filters and SOA material density of
1.4 gcm-3. The density of indole SOA is unknown, but the adopted
value of 1.4 gcm-3 is similar to the reported range of densities
of 1.47–1.55 gcm-3 for SOA prepared from another bicyclic
aromatic compound, naphthalene (Chan et al., 2009; Chen et al., 2016). In
addition, densities of known indole oxidation products, for example isatin
(1.47 gcm-3), anthranilic acid (1.40 gcm-3), indigo
dye (1.20 gcm-3), isatoic anhydride (1.52 gcm-3),
and 3-oxindole
(1.20 gcm-3), range from 1.2 to 1.5 gcm-3,
suggesting that 1.4 gcm-3 should be a reasonable guess for
indole SOA.
The SOA yield was calculated from Eq. ().
Yield=ΔSOAΔVOC
The increase in the mass concentration of particles, ΔSOA,
was obtained from SMPS measurements and corrected for the particle wall loss
as described in the Supplement. The change in the mass concentration of
indole, ΔVOC, was equated to the initial indole concentration
because PTR-ToF-MS measurements suggested complete removal of indole during
the photooxidation.
The filter with the collected sample was cut in half. The first half was used
for UV–vis measurements. The sample was extracted by placing the filter half
in a covered petri dish containing 3 mL of methanol (LC-MS grade,
99.9 % purity, Honeywell) and shaken vigorously on a shaker for five
minutes. Assuming a complete extraction of the SOA material, this would
result in a mass concentration of 0.03–0.22 mgmL-1. The filter
colour changed from brown to white, suggesting that most of the
light-absorbing compounds were extracted. The assumption of a full extraction
is supported by the solubility of 21 mgmL-1 reported for isatin
in methanol (Baluja et al., 2013); methanol solubilities of other indole
oxidation products are expected to be similarly high. The SOA methanol
extract was then analysed by UV–vis spectrophotometry in a dual-beam
spectrophotometer (Shimadzu UV-2450), with pure methanol used as reference.
Wavelength-dependent mass absorption coefficient (MAC) was calculated for
indole SOA from the base-10 absorbance, A10, of an SOA extract, the path
length, b (cm), and the solution mass concentration, Cmass
(gcm-3):
MAC(λ)=A10solutionλ×ln(10)b×Cmass.
The main uncertainty in the calculated MAC values comes from the uncertainty
of the mass concentration, which arises from uncertainties in the SMPS
measurement of aerosol mass concentration, filter collection efficiency, and
solvent extraction efficiency. We estimate that MAC values should be accurate
within a factor of 2 (Romonosky et al., 2015a).
The second half of the filter was used for direct analysis in real-time mass
spectrometry (DART-MS) measurements. The filter half was extracted in the
same way with 3 mL of acetonitrile (LC-MS grade, 99.9 % purity,
Honeywell). Assuming a complete extraction of the SOA material, the mass
concentration of 0.03–0.22 mgmL-1 is much lower than the
solubility of 19 mgmL-1 reported for isatin in acetonitrile (Liu
et al., 2014). (We elected to use different solvents for UV–vis and DART-MS
because methanol afforded measurements deeper in the UV region and
acetonitrile gave cleaner background spectra in DART-MS. Based on visual
inspection, the samples appeared to dissolve fully in both acetonitrile and
methanol.) Aliquots from the acetonitrile SOA extracts were transferred onto
a clean stainless-steel mesh, dried in air, and manually inserted between the
DART ion source and mass spectrometer inlet. The DART-MS consisted of a Xevo
TQS quadrupole mass spectrometer (Waters) equipped with a commercial DART ion
source (Ion-Sense, DART SVP with Vapur® Interface).
It was operated with the following settings: 350 V grid electron
voltage, 3.1 Lmin-1 He gas flow, 350 ∘C He gas
temperature, and 70 ∘C source temperature. The samples were analysed
with DART-MS in both positive and negative ion modes. Background spectra from
the pure solvent were also collected and subtracted from the DART mass
spectra.
Additional sample filters were analysed via nanospray desorption electrospray
ionization high-resolution mass spectrometry (nano-DESI-HRMS) and
high-performance liquid chromatography, coupled to photodiode array
spectrophotometry and high-resolution mass spectrometry (HPLC-PDA-HRMS). The
former method provides a spectrum of the entire mixture without prior
separation; it is useful for providing an overview of the types of compounds
present in SOA. The latter method is suited for advanced detection of
individual light-absorbing components in SOA (Lin et al., 2015a, b; Lin
et al., 2016). Both methods employ an LTQ-Orbitrap mass spectrometer (Thermo
Corp.) with a resolving power of 105 at m/z 400, sufficient for
unambiguous characterization of SOA constituents.
For the HPLC-PDA-HRMS measurements, one-quarter of the filter was extracted
using 350 µL acetonitrile (CH3CN, gradient grade, ≥99.9 % purity) and the change in filter colour from brown to white
suggested that most light-absorbing compounds were extracted into the
solution. Separation of the SOA extract was achieved with a Scherzo SM-C18
column (Imtakt USA). The gradient elution protocol included a 3 min
hold at 10 % of CH3CN, a 45 min linear gradient to 90 %
CH3CN, a 16 min hold at this level, a 1 min return to
10 % CH3CN, and another hold until the total scan time of
90 min. The column was maintained at 25 ∘C and the sample
injection volume was 8 µL. The UV–vis spectrum was measured using
the PDA detector over the wavelength range of 250 to 700 nm. The ESI
settings were positive ionization mode, +4.5kV spray potential,
35 units of sheath gas flow, 10 units of auxiliary gas flow, and 8 units of
sweep gas flow.
The HRMS data analysis was performed by methods summarized by Romonosky
et al. (2015b). Briefly, the mass spectra were clustered together, the m/z
axis was calibrated internally with respect to expected products of
photooxidation, and the peaks were assigned to formulas
CcHhOoNnNa0-1+
or CcHhOoNn-
constrained by valence rules and elemental ratios (c, h, o, n refer
to the number of corresponding atoms in the ion). These were then converted
to formulas of the corresponding neutral species, obtained by removing Na or
H from the observed positive ion formulas or adding H to the negative ion
formulas. The HPLC-PDA-HRMS analysis was done as described in Lin
et al. (2015b, 2016).
Modelling methods
Air quality simulations were performed to complement laboratory experiments
and to assess the formation of indole SOA in a coastal urban area. The
University of California, Irvine, and California Institute of Technology
(UCI-CIT) regional airshed model with a state-of-the-art chemical mechanism
and aerosol modules was used in this study. The model domain utilized 4970
computational cells (five vertical layers with 994 cells per layer) with
a 5km×5km horizontal grid size and encompassed the
SoCAB, including the Pacific Ocean on
the western side, heavily populated urban areas around Los Angeles, and
locations with a high density of plant life such as the Angeles National
Forest on the eastern side. The model included spatially and temporally resolved
emissions and typical meteorological conditions for this region. The emission
inventory used in this study was based on the 2012 Air Quality Management
Plan (AQMP) provided by the South Coast Air Quality Management District
(SCAQMD, 2013). Boundary and initial conditions were based on historical
values. Simulations were performed for a 3-day summer episode. Two days of
model spin-up time were used to reduce the influence of initial conditions
and allow sufficient time for newly added emissions to drive changes in air
quality. Results shown below are for the third day of the simulations.
The UCI-CIT model utilizes an expanded version of the Caltech atmospheric
chemical mechanism (CACM; Dawson et al., 2016; Griffin et al., 2002a, b,
2005) and has been used in numerous other studies to simulate air quality in
the SoCAB (Carreras-Sospedra et al., 2006; Carreras-Sospedra et al., 2010;
Chang et al., 2010; Nguyen and Dabdub, 2002). The CACM includes
a comprehensive treatment of SOA known as the Model to Predict the Multiphase
Partitioning of Organics (MPMPO) (Griffin et al., 2003, 2005). MPMPO is
a fully coupled aqueous–organic equilibrium-partitioning-based model and is
used to calculate gas–particle conversion of secondary organic species. The
SIMPOL.1 group-contribution method of Pankow and Asher (2008) is used to
calculate vapour pressures of SOA species for use in MPMPO. Activity in both
the aqueous and organic phases is determined using the UNIFAC model of Hansen
et al. (1991). Henry's law constants are calculated according to the group
contribution method of Suzuki et al. (1992). Several studies have used the
UCI-CIT model to investigate SOA formation, dynamics, reactivity, and
partitioning phase preference in the SoCAB (Carreras-Sospedra et al., 2005;
Chang et al., 2010; Cohan et al., 2013; Dawson et al., 2016; Griffin et al.,
2002b; Vutukuru et al., 2006). For a more detailed description of recent
model developments incorporated into the UCI-CIT model and its SOA modules,
the reader is referred to Dawson et al. (2016).
Summary of modifications made to the UCI-CIT model chemical
mechanism and aerosol modules. In the simulation that included oxidation by
NO3 an additional similar reaction was added with the
indole +NO3 rate constant of 1.3×10-10cm3molec-1s-1.
For the present study, the chemical mechanism was modified from the base case
version to include species and processes shown in Fig. 2. Two new gas-phase
species were added: indole and one representative oxidation product, indigo
dye. Because of the high mass yield of indole SOA, with most of the products
ending up in the particle phase, any reasonable indole oxidation product with
a low vapour pressure would be suitable. We elected to use indigo dye
(C16H10O2N2) because it is a very common derivative of indole
and because its formula was reasonably close to the average formula of SOA
compounds determined from nano-DESI (C15H11O3N2). One new
gas-phase reaction was added, which forms gas-phase indigo dye via oxidation
of gas-phase indole by hydroxyl radical. Lastly, indigo dye was also added to
the model as a new SOA species. Gas-phase indigo dye was assumed to partition
into the aerosol phase based on its calculated vapour pressure and Henry's
law constant. After the modifications described here, the model contained
a total of 202 gas-phase species, 607 gas-phase reactions, and 18 SOA
species. Each SOA species was sorted into eight distinct size bins based on
particle diameter, up to a maximum of 10 µm. The activity
coefficient of indigo dye was assumed to be 1.
Because gas-phase indole was not included in the base case emissions
inventory, its emission rate was estimated based on available literature
data. As discussed in the introduction section, emissions of indole have been
shown to be well correlated to emissions of monoterpenes in a variety of
plant species (Erb et al., 2015; Gentner et al., 2014; Niinemets et al.,
2013). However, most existing data were obtained from controlled laboratory
experiments and emissions of indole at the regional scale are not well
constrained. In this work, emissions of gas-phase indole were added to the
base case emissions inventory by using a ratio to “BIOL”, an existing
gas-phase species in the emission inventory. BIOL is representative of lumped
biogenic monoterpenes and contains spatially and temporally resolved
emissions in the base case inventory. Therefore, the spatiotemporal
distribution of indole emissions follows that of BIOL, with the magnitude of
the emissions set to a given percentage of BIOL emissions. Please note that no
emissions of indole derived from agriculture and animal husbandry activities
were added in the model because these sources remain uncertain. In addition,
no direct emissions of gas-phase indigo dye were added to the model. Because
of the uncertainty and episodic nature of gas-phase indole emissions,
simulations were performed with a range of possible emission factors to
determine the sensitivity of indole SOA formation to gas-phase indole
emissions.
Five scenarios were considered for model calculations. The first scenario had
zero emissions of gas-phase indole. This scenario will be referred to as the
“base case” and serve as the reference scenario to which the other
scenarios are compared to determine changes in air quality. The second,
third, and fourth scenarios had emissions of gas-phase indole set to 5, 10,
and 25 % of BIOL emissions, referred herein as “low”, “medium”, and
“high” emissions, respectively. When averaged over the entire simulation
domain, the corresponding average emission factors for indole were 0.25,
0.51, and 1.27 µgm-2h-1, respectively. A comparable
emission factor of 0.6 µgm-2h-1 for indole was used in
a previous study of Misztal et al. (2015), in which indole emissions under
average stress conditions were incorporated in the MEGAN 2.1 biogenic VOC
emissions model to estimate total global emissions. Therefore, the medium
emission scenario considered in this study should be representative of the
emissions of indole under average stress conditions, while the high emission
scenario is more likely to represent episodic emission events such as those
during springtime flowering or herbivore infestation.
The focus of this modelling work was to study the formation of SOA from the
photooxidation of indole by OH, in order to complement the experimental data
reported in this work. While the SOA formation from oxidation of indole by
NO3 was not experimentally studied, the reaction of indole with
NO3 is fast and may provide an additional source of indole SOA. The
fifth and final scenario explored the potential impact of including an
additional oxidation pathway for gas-phase indole via reaction with
NO3. This scenario corresponds to the high emission scenario with
one new gas-phase reaction included in the model in addition to those
described previously. For this new reaction, it is assumed that gas-phase
indole reacts with NO3 to produce indigo dye, the same representative
oxidation product assumed for the reaction of gas-phase indole with hydroxyl
radical. A rate constant of 1.3×10-10cm3molec-1s-1 is used following Atkinson
et al. (1995). No other changes were made to the model or its inputs in this
scenario.
The mass concentration of indole (solid trace), the mass
concentration of particles (open circles), and the wall-loss-corrected mass
concentration of particles (solid circles) over time. Indole was not yet
fully mixed in the chamber by the time photooxidation started at t=0,
resulting in an apparent initial rise in the measured indole concentration.
Results and discussionProperties of indole SOA
Figure 3 illustrates the time dependence of mass concentrations of indole and
particulate matter in a typical chamber experiment. According to PTR-ToF-MS
measurements, indole decayed with a half-life of approximately
60 min, which translates into an average OH concentration in the
chamber of 1.4×106moleccm-3, similar to ambient
levels (Fig. S2.1 in the Supplement). The PTR-ToF mass spectrum of indole
before photooxidation (Fig. S2.2) was dominated by the protonated indole at
m/z 118.067 (the m/z values cited in the text correspond to the measured
m/z values; the corresponding exact m/z values are listed in Table S2 in
the Supplement). After photooxidation, a few other prominent peaks appeared.
Figures S2.3 and S2.4 show the time-dependence profiles of several peaks of
interest detected by PTR-ToF-MS during the photooxidation of indole, and
Table S2 contains their proposed assignments. Peaks at m/z 120.072,
131.062, and 132.050 (Fig. S2.4) appeared simultaneously with indole
injection, suggesting that the indole sample contained small amounts (<2 %) of indoline, diazanaphthalene, and 3-oxyindole impurities,
respectively, which may have contributed to SOA formation. From the ions that
first appeared and then were consumed during photooxidation (Fig. S2.3), the
one at m/z 122.061 had the largest peak abundance. It corresponds to
protonated 2-formylformanilide [M+H]+ ion (Fig. 1), a major
gas-phase product of indole oxidation by OH (Atkinson et al., 1995). Another
significant product was detected at m/z 148.041 and was tentatively
assigned to the [M+H]+ ion from isatin (Fig. 1). Isatin also was
observed as an abundant peak in both DART(+) and nano-DESI(+) mass
spectra, suggesting that it can partition between the gas and particle
phases. Smaller peaks produced and then consumed in photooxidation included
indoxyl, benzonitrile, and phenylamine. A few peaks at smaller m/z grew
during the photooxidation including cyanic acid, acetaldehyde, acetone, and
acetic acid.
Wavelength-dependent mass absorption coefficient (MAC) of indole
SOA. The inset shows the log–log version of the same data used to determine
the absorption Angstrom exponent (fitted from 300 to 600 nm) and photographs of the indole SOA collected on a filter and extracted in
methanol.
The particles had a geometric mean diameter of approximately
0.3 µm when the filter collection started. The terminal
wall-loss-corrected mass concentration of particles (Fig. 3) was higher than
the initial concentration of indole, suggesting that the SOA yield, defined
by Eq. (), was high. For five experiments repeated under the same
conditions on separate days, the SOA yields calculated from Eq. ()
were 1.21, 1.10, 0.86, 1.77, and 1.46 with an average of 1.3±0.3. We
normally obtain much more reproducible yields for more volatile precursors,
such as monoterpenes; it is not clear to the authors why the yield of indole
SOA was so variable. Indole
oxidation products could be lost to the walls reducing the apparent yield and
contributing to its scatter. However, this effect is probably minor given
that the apparent yield is quite high. The yield is higher than that for SOA
formed from another bicyclic aromatic compound, naphthalene, which has
a reported yield range of 0.04–0.73 under low-NOx conditions (Chan
et al., 2009; Chen et al., 2016). The high yield suggests that the major
fraction of indole oxidation products ends up in the particle phase at the
concentrations used in this work. The yield of 1.3 would require that, on
average, at least two oxygen atoms should add to the indole during oxidation
(C8H7N→C8H7NO2.2), which is quite
reasonable and qualitatively consistent with mass spectrometric observations.
Nano-DESI and DART mass spectra of indole SOA plotted as a function
of the molecular weights of the neutral compounds. The nano-DESI mass spectra
contain only peaks assignable to specific formulas, while DART mass spectra
contain all observed peaks. The five most abundant peaks in each mass
spectrum are indicated with letters: (a) 248 Da,
C15H8O2N2; (b) 250 Da, C15H10O2N2,
tryptanthrin; (c) 262 Da, C16H10O2N2, indirubin
and/or indigo dye; (d) 264 Da, C16H12O2N2,
dihydro indigo dye; (e) 266 Da, C15H10O3N2; (f)
147 Da, C8H5O2N, isatin; (g) 252 Da,
C15H12O2N2; (h) 280 Da, C16H12O3N2;
(i) 282 Da, C16H14O3N2; (j) 121 Da,
C7H7ON, 2-formylformanilide; (k) 162 Da,
C8H6O2N2; (l) 163 Da,C8H5O3N2, isatoic
anhydride; (m) 165 Da, C8H7O3N.
Figure 4 shows the MAC values measured for an extract of indole SOA in
methanol. MAC values reached ∼2m2g-1 at λ=300nm. At λ=400–700 nm, the MAC values ranged
from 0.5 to 0.02 m2g-1. These high MAC values are comparable
to values of strongly absorbing SOA derived from naphthalene or
methylpyrroles (Romonosky et al., 2015a) and to MAC values of biomass burning
organic aerosol (Sun et al., 2007). The wavelength dependency of MAC deviates
from the power law commonly observed for brown carbon (e.g. see reviews of
Laskin et al., 2015, and Moise et al., 2015) and has a reproducible broad
band at ∼350nm, possibly due to the well-known derivatives of
indole (indirubin, indigo dye, and indoxyl red) that have characteristic
absorption bands at this wavelength (see below). For the wavelength range of
300–600 nm, the absorption Angstrom exponent was ∼6,
comparable to the value of ∼5 reported for brown carbon from biomass
burning (Kirchstetter et al., 2012).
We used two offline MS methods (DART and nano-DESI) and both negative (-)
and positive (+) ion modes to characterize the SOA composition to detect
a broader range of compounds than possible with a single method. Figure 5
shows the DART and nano-DESI mass spectra of indole SOA in both positive and
negative modes. The high resolving power of nano-DESI-HRMS afforded
unambiguous formula assignments for all peaks up to m/z 500, and the
molecular weights (MWs) of the neutral compounds could be determined from the
corresponding ion formulas. About half of the ions observed in nano-DESI
(+) mass spectra were [M+Na]+ adducts, and the remaining
compounds were protonated ions, [M+H]+. The DART mass spectra were
acquired on a triple quadrupole mass spectrometer with only unity mass
resolution. As a result, only the nominal m/z values for the observed peaks
could be determined. It was assumed that the dominant mechanism of ionization
was protonation ([M+H]+ ions formed; nominal MW = nominal
m/z-1) in the positive ion mode and deprotonation ([M-H]- ions
formed; nominal MW = nominal m/z+1) in the negative ion mode (Nah
et al., 2013). For ease of comparison, all the mass spectra were plotted as
function of the exact mass of the corresponding neutral compounds.
For a given ion mode, the DART and nano-DESI mass spectra were qualitatively
similar, although nano-DESI appeared to favour larger, more oxidized
compounds compared to DART. Both DART and nano-DESI mass spectra showed
a clear separation into distinct clusters of peaks corresponding to monomer,
dimer, trimer, and tetramer oxidation products. For a given ion mode, the
major monomer peaks were the same in DART and nano-DESI, strongly suggesting
that they correspond to more abundant indole oxidation products (as opposed
to minor SOA compounds that happened to have unusually high ionization
efficiencies). There was also good correspondence between the major dimer
peaks recorded in DART and nano-DESI. In the nano-DESI mass spectra, the peak
abundances in the negative ion mode spectra were shifted towards higher
MWs compared to the positive ion mode mass spectra. The
preferential negative ion formation from more oxidized compounds was
previously observed in ESI mass spectra of limonene SOA (Walser et al.,
2008). Although we cannot assign formulas to the DART-MS peaks, it is evident
from Fig. 5 that this ionization method also favours larger, and presumably
more oxidized, compounds in the negative ion mode. For example, carboxylic
acids are more readily observed in the negative ion mode DART mass spectra
(Nah et al., 2013).
Table 1 lists the most abundant peaks observed in the monomer and dimer
ranges of nano-DESI-HRMS and DART-MS spectra, as well as additional smaller
peaks for the specific compounds discussed in this paper. Isatin
(C8H5O2N, MW =147Da) was the single dominant peak
in the monomer range observed in both nano-DESI(+) and DART(+); it was
also detected in the negative ion mode mass spectra. Isatoic anhydride
(C8H5O3N; MW =163Da) was the second most abundant
monomeric peak in all four mass spectra but with much lower abundance. Other
abundant monomeric products included 3-oxyindole (C8H5ON;
MW =131Da) and 2-formylformanilide (C7H7ON;
MW =121Da). Of the compounds shown in Fig. 1, tryptanthrin
(C15H10O2N2; MW =250Da), indirubin
(C16H10O2N2; MW =262Da), indigo dye
(C16H10O2N2; MW =262Da), and dihydro indigo dye
(C16H12O2N2; MW =264Da) were the most abundant
dimer peaks. Meanwhile, indoxyl red (C16H10ON2;
MW =246Da) was detected with lower but appreciable abundances in
nano-DESI(-) and in both DART mass spectra. The prominent dimer compounds
listed in Table 1 contained additional oxygen atoms compared to indoxyl red,
indirubin, indigo dye, and dihydro indigo dye and could be formed by further
oxidation of these compounds.
Distribution of the number of C atoms in the indole SOA compounds
detected in both positive and negative ion modes by nano-DESI-HRMS.
Figure 6 shows the distribution of the number of C atoms in the indole SOA
compounds, as detected by nano-DESI-HRMS (for each group of compounds with
the same number of C atoms, the abundances in the positive and negative ion
mode mass spectra were added together). Most of the observed compounds
contained 8, 16, or 24 C atoms, corresponding to the monomer, dimer, and
trimer derivatives of indole. Peaks with 7 and 15 carbon atoms were also
prominent, suggesting an important role of the primary C7 oxidation
product 2-formylformanilide in the formation of low volatility species. Minor
peaks containing other C numbers were also present, suggesting further
fragmentation of the primary oxidation products. The average formula for all
observed SOA compounds was C15H11O3N2.
Figure S3 shows the distribution of the N/C ratios in
indole SOA compounds. Many of the compounds had the same
N/C ratio as indole (N/C=1/8),
indicating that the oxidation and oligomerization reactions conserved both N and C
atoms in many of the products. However, some products had a slightly larger
ratio consistent with a loss of C atoms (e.g.
N/C=1/7 and 2/15), whereas some products gained
extra C atoms. One product with a relatively large abundance,
C12H14O4, had no N atoms left in it. In addition, there were
several C8-9HhOoN2 products that gained an additional N
atom. The mechanism of photooxidation is clearly complex involving a large
number of secondary reactions. The full mechanism of indole photooxidation
cannot be obtained from this data set. In the discussion that follows, the
focus will be on the mechanism of formation of light-absorbing products.
HPLC-PDA chromatogram of indole SOA. The absorbance is plotted as
a function of both retention time and wavelength. Peaks are labelled by their
PDA retention time followed by their proposed assignment. Bold-faced
assignments are specific isomers that are discussed further in the text. Note
the reference line for dihydro indigo points to a small peak between two
larger peaks that obscure it in this projection.
Comparison between measured PDA absorption spectra at selected
retention time (RT) and reference spectra of proposed chromophores in the
literature (reference spectrum is not available for dihydro indigo dye; panel
e).
Monomer and dimer peaks with the largest peak abundance observed in
DART-MS and nano-DESI-HRMS spectra. Selected peaks corresponding to the
compounds shown in Fig. 1 are also included. Proposed assignments are based
on the formulas from nano-DESI-HRMS. Peak abundances are normalized with
respect to the most abundant peak in each spectrum.
Figure 7 shows the HPLC-PDA chromatogram of an indole SOA sample
demonstrating its components with strong light-absorbing properties near-UV
and visible spectral ranges (above 300 nm). To identify specific
chromophores from the HPLC-PDA-HRMS data, the methods described by Lin
et al. (2015b, 2016) were followed. High-resolution mass spectra were
examined to identify m/z values that appear at the retention times
associated with the peaks in the HPLC chromatograms. The PDA absorption
spectra associated with these retention times were then compared with
possible candidates constrained by their molecular formula determined from
the mass spectra.
Tentative mechanism for the formation of observed chromophores in
the photooxidation of indole. (a) Processes leading to indigo dye
and indoxyl red based on Iddon et al. (1971). (b) Processes leading
to tryptanthrin based on Novotna et al. (2003). “Ox” denotes an oxidation
step.
Figure 8 shows a comparison of the absorption spectra for the key peaks in
the HPLC-PDA chromatogram with absorption spectra of selected compounds
reported in the literature. The match is excellent in terms of the absorption
peak maxima: 280, 310, 334, and 392 nm for tryptanthrin; 240, 283, 335,
and 610 nm for indigo dye; 242, 290, 365, and 540 nm for indirubin;
and 217, 273, 350, and 520 nm for indoxyl red. The shapes of the spectra
do not match perfectly because the chromophores are not fully separated by
the HPLC column (Fig. 7) and may co-elute with additional minor compounds.
Likely, more than one chromophore contributed to the absorbance at any given
retention time. However, the power of the method is clear, as illustrated,
for example, by the separation of the structural isomers indigo dye and
indirubin (Fig. 7).
The precursors to indoxyl red and indigo dye – dihydro indoxyl red and
dihydro indigo dye, respectively – were also identified by this analysis and
were observed in nano-DESI mass spectra. This observation supports
a mechanism similar to the aqueous-phase indole oxidation proposed by Iddon
et al. (1971), in which indole first oxidizes to 3-oxindole, then to dihydro
indigo dye or dihydro indoxyl red, and then finally to indigo dye and indoxyl
red (Fig. 9a). We note that the mechanism by Iddon et al. (1971) was
developed for the aqueous oxidation of indole. While our experiments were
performed under dry conditions, it is conceivable that similar
oligomerization processes can occur in the gas phase and/or organic particle
phase. For example, Healy et al. (2012) observed efficient dimerization of
naphthoxy radicals in the gas phase, leading to rapid formation of SOA
following photolysis of 1-nitronaphthalene. The dimerization of
3-oxindole to dihydro indigo dye, as
well as other oligomerization processes in indole SOA, could follow
a mechanism similar to the one described by Healy et al. (2012).
Several products were assigned based on previous observations by Novotna
et al. (2003), who studied photodegradation of indigo dye in dichloromethane
solution. They proposed the mechanism shown in Fig. 9b to explain the
production of tryptanthrin and anthranilic acid from ambient indigo dye
oxidation. In this mechanism, hydroxyl radicals attack the carbonyl carbon
atoms of isatin ultimately opening the five-membered N-heterocyclic ring to
yield anthranilic acid. Although anthranilic acid does not show up in Fig. 7
because it is not a chromophoric species, it was detected by nano-DESI-HRMS.
As shown in Fig. 9b, anthranilic acid can react with another molecule of
isatin to produce tryptanthrin. This mechanism is relevant to indole SOA,
because isatin can be produced not only from the oxidation of indigo dye but
also directly from indole, through the intermediacy of 3-oxindole (Fig. 9a).
Moreover, Novotna and colleagues suggested that isatoic anhydride should also
be formed from indigo dye oxidation. A compound with this formula had large
abundance in both HPLC-PDA-HRMS (Fig. 7) and nano-DESI-HRMS and DART-MS
(Fig. 5, Table 1). Combined with the evidence that tryptanthrin is a major
secondary chromophore, this could be a significant pathway to brown carbon
formation in the oxidation of indole.
We emphasize that the mechanism outlined in Fig. 9 is tentative and
is based on the limited information from our experiments and previous
experimental data from the literature. Multiple unresolved questions remain.
For example, formation of tryptanthrin was very slow in experiments by
Novotna et al. (2003), and it is not at clear how this compound could form in
just a few hours of photooxidation in the chamber. Furthermore, it is not
clear which processes occur in the gaseous phase vs. the particle phase.
Although indeterminable from the current experiments, at least some of the
dimer formation pathways described in Fig. 9 likely occur in the particle
phase.
Twenty-four-hour average concentrations (µgm-3) of (a)
total SOA in the base case and additional SOA resulting from indole
photooxidation in (b) the low emission scenario, (c) the
medium emission scenario, and (d) the high emission scenario.
Potential effects of indole SOA
The spatiotemporal distribution of indole SOA is likely controlled by
a combination of (i) the spatiotemporal distribution of gas-phase indole and
its emissions sources; (ii) the availability of hydroxyl radical for
gas-phase oxidation chemistry; and (iii) meteorological conditions in the
region, including temperature, humidity, and wind direction. Once emitted,
indole reacts with hydroxyl radical to form gas-phase indigo dye. Gas-phase
indigo dye can then partition into the aerosol phase to form indole SOA. The
presence of a sea breeze in the SoCAB results in a prevailing wind direction
of north-northeast, transporting pollutants inland during the daytime hours.
As a result, peak concentrations of indole SOA should be located further
inland than peak concentrations of gas-phase indole and occur in areas that
are already burdened with poor air quality.
Figure S4 shows the spatial distribution of 24 h average gas-phase indole
concentrations in the SoCAB for the low, medium, and high emission scenarios
considered in this study. The amount of indole SOA formed in the model, and
thus the impact of indole on the total predicted SOA concentrations, depends
strongly on the emissions of gas-phase indole. In the high emission scenario,
hourly gas-phase indole concentrations peaked at 0.3 ppbv, with the
highest concentrations occurring in the early morning hours before sunrise
(Fig. S5). For comparison, during a field measurement campaign in the San
Joaquin Valley of California, Gentner et al. (2014) reported gas-phase indole
concentrations of about 1–3 ppbv in ambient air during a springtime
flowering event. Measured concentrations of indole were slightly higher
during the late-night and early morning hours than during the daytime,
consistent with the model results obtained in this study. Gentner
et al. (2014) also showed that flowering was a major biogenic emissions
event, causing emissions of many compounds to increase by several factors to
over an order of magnitude. Therefore, episodic emissions of indole in rural
areas are likely to be significantly greater than the emissions used in this
study. Based on the high SOA yield from gas-phase indole found in this study
we propose that biogenic emissions events such as springtime flowering may
degrade local air quality.
Figure 10a shows 24 h average SOA concentrations in the base case model
simulation, and Fig. 10b–d show the additional SOA resulting from indole in
the three emission scenarios. The highest SOA concentrations occurred
directly east of Riverside where a combination of biogenic and anthropogenic
precursors accumulated during days one and two and into day three. The 24 h
average indole SOA concentrations peaked at about 0.13 µgm-3
in the high emission scenario (Fig. 10d). The highest concentrations of
indole SOA occurred north of Los Angeles and Riverside. To put this number in
perspective, aerosol with mass concentration of 0.1 µgm-3
and MAC of 0.5 m2g-1 will have an absorption coefficient of
0.05 Mm-1 (we neglect the particle size effects in this
estimation). Thompson et al. (2012) reported an absorption coefficient of
4 Mm-1 at 532 nm during the 2010 CalNex campaign in
Pasadena, California, with the absorption being dominated by black carbon.
The average absorption coefficients reported for “average urban USA” and
“average remote USA” by Horvath et al. (1993) were 22 and
0.7 Mm-1, respectively. While the absorption by indole SOA is
unlikely to compete with that by black carbon in urban areas, it may
contribute to the aerosol absorption in more remote areas, where the black
carbon concentrations are smaller.
Domain-wide average SOA concentrations in the base case (black line,
left axis) and the relative increase in domain-wide average SOA
concentrations (right axis) driven by the oxidation of indole by OH in the
(a) low emission scenario, (b) medium emission scenario, and (c) high
emission scenario. Figure S6 additionally shows the effect of inclusion of
the indole +NO3 reaction on the model.
SOA concentrations averaged over the entire domain are shown in Fig. 11 for
the first four modelled scenarios. The averaged SOA concentrations were
computed by averaging the concentration of total SOA in all
computational cells in the domain.
Therefore, changes in the averaged SOA concentrations are representative of
the overall impact on total SOA concentrations for the entire basin. In the
high emission scenario, the averaged SOA concentrations increased by about
4–13 %, indicating that indole SOA can contribute significantly to total
organic aerosol concentrations. While base case SOA concentrations peaked
during the early morning and late-night hours when metrological conditions
were favourable, the largest changes in SOA concentrations occurred
during the late morning and afternoon hours. The formation of gas-phase
indigo dye and indole SOA depends on the availability of the hydroxyl
radical, which reaches peak concentrations during daylight hours, when
photochemistry is active. Therefore, increased production of hydroxyl radical
during the daytime accelerates the oxidation of gas-phase indole, ultimately
resulting in increased formation of indole SOA. Increases in total SOA are
due mostly to the formation of indole SOA, with only small changes in the
concentration of other SOA species.
Figures S5 and S6 suggest that the oxidation of indole via reaction with
NO3 may also be an important pathway to indole SOA formation during
the late-night and early morning hours when photochemistry is inactive. When
the oxidation of indole via reaction with NO3 is included in the
fifth modelled scenario, gas-phase indole concentrations are lower (Fig. S5)
and indole SOA concentrations are higher (Fig. S6) than in the fourth
scenario. The differences are most pronounced during the late-night and early
morning hours due to the different diurnal profiles of OH and NO3; OH
concentrations peak during the daytime hours when photochemistry is active,
whereas NO3 concentrations peak at night. Thus, the relative increase
in total SOA concentrations due to indole SOA shows less temporal variability
throughout the day when the reaction of gas-phase indole with NO3 is
included in the model, but peak indole SOA concentrations remain essentially
unchanged.
The amount of indole SOA formed in each scenario was found to be directly
proportional to the emissions of gas-phase indole. In the low emission
scenario, gas-phase indole and indole SOA concentrations were about factor of
5 lower than those seen in the high emission scenario, with 24 h average
indole SOA concentrations peaking at about 0.025 µgm-3.
Similarly, relative increases in the averaged SOA concentrations ranged from
1 to 3 % in the low emission scenario. In the medium emission scenario,
24 h average indole SOA concentrations reached about
0.05 µgm-3, causing total SOA concentrations to increase by
2–6 %. In all three emission scenarios, the spatial distribution of
indole SOA remained essentially the same, with peak concentrations occurring
in the northeast portion of the basin, an area already burdened with poor air
quality.
Conclusions
This work demonstrates that indole is an effective precursor to
SOA. At the concentrations used in this chamber study, the majority of indole
oxidation products ended up in the particle phase, with an effective SOA
yield of ∼1.3±0.3. The resulting SOA was found to be strongly
light absorbing, with MAC values ranging from 0.5 to
0.02 m2g-1 across the visible spectrum and approaching those
of strongly absorbing brown carbon from biomass burning. The high MAC values
were due to well-known chromophoric products of indole oxidation, including
tryptanthrin, indirubin, indigo dye, and indoxyl red, which were identified
by their molecular formulas and characteristic peaks in their absorption
spectra. These observations suggest that N-heterocyclic compounds may be
important contributors to secondary brown carbon.
Contribution of indole to SOA formation can potentially result in reduced
visibility, particularly in regions where plants are exposed to biotic and
abiotic stresses. When combining the experimental MAC values with peak SOA
concentrations predicted in the model, the estimated maximum absorption
coefficient is 0.05 Mm-1 due to indole SOA. This is smaller than
the values typically reported for SoCAB but comparable to values reported in
more remote areas. Thus, despite its large MAC, indole SOA is not likely to
contribute to particle absorption in urban areas, where anthropogenic black
carbon dominates the aerosol absorption. However, the situation could be
different in remote areas, where black carbon does not contribute to aerosol
absorption, and indole emissions are higher.
The UCI-CIT regional airshed model showed significant potential for indole
SOA formation driven by the oxidation of indole by OH. Simulations indicate
that the oxidation of indole via reaction with NO3 may also be an
important SOA formation pathway during the late-night and early morning hours
when photochemistry is inactive. While the mass loading of indole SOA in the
SoCAB was relatively low in all scenarios, it represents a previously
unconsidered source of SOA in air quality models, which have been improved in
recent years but still tend to disagree with measured SOA concentrations
(Couvidat et al., 2013; Heald et al., 2005; Hodzic et al., 2010; Jiang
et al., 2012; Volkamer et al., 2006). Indole SOA can interact with other
aerosol-phase species, causing indirect changes in the concentration of total
SOA. Such interactions were not considered in this study because an activity
coefficient of unity was used for indole SOA in the model simulations. Rural
or agricultural regions with significant biomass burning or a high density of
plant life likely have much higher emissions of gas-phase indole than the
SoCAB. For example, field measurement studies (Gentner et al., 2014) reported
ambient indole concentrations up to an order of magnitude greater than the
peak modelled concentrations employed in this study, indicating a significant
potential for indole SOA formation in rural areas. Furthermore, future
climate change is likely to increase gas-phase indole emissions through
environmental and physical stress factors such as drought, elevated
temperatures, increased CO2 and O3 concentrations, and
enhanced herbivore feeding (Yuan et al., 2009). Therefore, indole represents
a potentially important source of biogenic SOA that should be included in
regional and global models.
The raw data from DART-MS, nano-DESI-HRMS, HPLC-PDA-HRMS,
and PTR-MS instruments are available from the corresponding author upon
request.
The Supplement related to this article is available online at https://doi.org/10.5194/acp-17-11605-2017-supplement.
The authors declare that they have no conflict of interest.
Acknowledgements
This publication was developed under assistance agreement no. EPA 83588101
awarded by the US Environmental Protection Agency to the regents of the
University of California. It has not been formally reviewed by EPA. The views
expressed in this document are solely those of the authors and do not
necessarily reflect those of the agency. EPA does not endorse any products or
commercial services mentioned in this publication.
Julia Montoya-Aguilera acknowledges additional support from the
California LSAMP Bridge to the Doctorate Program at the University of
California, Irvine, which is funded by grant NSF-1500284. The DART-MS and
PTR-ToF-MS instruments used in this work were previously purchased with
grants NSF CHE-1337080 and NSF MRI-0923323, respectively. Peng Lin, Julia
Laskin, and Alexander Laskin were supported by the US Department of Commerce,
National Oceanic and Atmospheric Administration through Climate Program
Office's AC4 program, awards NA16OAR4310101 and NA16OAR4310102. The HRMS
measurements were performed at the W.R. Wiley Environmental Molecular
Sciences Laboratory (EMSL), a national scientific user facility located at
PNNL, and sponsored by the Office of Biological and Environmental Research of
the US DOE. PNNL is operated for US DOE by Battelle Memorial Institute under
contract no. DE-AC06-76RL0 1830.
Edited by: Jason Surratt
Reviewed by: four anonymous referees
ReferencesAndreae, M. O. and Gelencsér, A.: Black carbon or brown carbon? The
nature of light-absorbing carbonaceous aerosols, Atmos. Chem. Phys., 6,
3131–3148, 10.5194/acp-6-3131-2006, 2006.Atkinson, R., Tuazon, E. C., Arey, J., and Aschmann, S. M.: Atmospheric and indoor chemistry of gas-phase indole,
quinoline, and isoquinoline, Atmos. Environ., 29, 3423–3432, 10.1016/1352-2310(95)00103-6, 1995.Baluja, S., Bhalodia, R., Bhatt, M., Vekariya, N., and Gajera, R.: Solubitily
of a pharmacological intermediate drug isatin in different solvents at
various temperatures, Int. Lett. Chem. Phys. Astron., 17, 36–46,
10.18052/www.scipress.com/ILCPA.17.36, 2013.Cardoza, Y. J., Lait, C. G., Schmelz, E. A., Huang, J., and Tumlinson, J. H.: Fungus-induced biochemical changes in
peanut plants and their effect on development of beet armyworm, Spodoptera exigua Hübner (Lepidoptera: Noctuidae) larvae,
Environ. Entomol., 32, 220–228, 10.1603/0046-225X-32.1.220, 2003.Carreras-Sospedra, M., Griffin, R. J., and Dabdub, D.: Calculations of incremental secondary organic aerosol
reactivity, Environ. Sci. Technol., 39, 1724–1730, 10.1021/es0495359, 2005.Carreras-Sospedra, M., Dabdub, D., Rodriguez, M., and Brouwer, J.: Air quality modeling in the south coast air basin
of California: What do the numbers really mean?, J. Air Waste Manage., 56, 1184–1195, 10.1080/10473289.2006.10464530,
2006.Carreras-Sospedra, M., Vutukuru, S., Brouwer, J., and Dabdub, D.: Central power generation versus distributed
generation–an air quality assessment in the South Coast Air Basin of California, Atmos. Environ., 44, 3215–3223,
10.1016/j.atmosenv.2010.05.017, 2010.Chan, A. W. H., Kautzman, K. E., Chhabra, P. S., Surratt, J. D., Chan, M. N., Crounse, J. D., Kürten, A.,
Wennberg, P. O., Flagan, R. C., and Seinfeld, J. H.: Secondary organic aerosol formation from photooxidation of naphthalene and
alkylnaphthalenes: implications for oxidation of intermediate volatility organic compounds (IVOCs), Atmos. Chem. Phys., 9,
3049–3060, 10.5194/acp-9-3049-2009, 2009.Chang, W. L., Griffin, R. J., and Dabdub, D.: Partitioning phase preference for secondary organic aerosol in an urban
atmosphere, P. Natl. Acad. Sci. USA, 107, 6705–6710, 10.1073/pnas.0911244107, 2010.Chen, C., Kacarab, M., Tang, P., and Cocker, D. R.: SOA formation from naphthalene, 1-methylnapthalene, and
2-methylnaphthalene photooxidation, Atmos. Environ., 131, 424–433, 10.1016/j.atmosenv.2016.02.007, 2016.Cohan, A., Eiguren-Fernandez, A., Miguel, A. H., and Dabdub, D.: Secondary organic aerosol formation from naphthalene
roadway emissions in the South Coast Air Basin of California, Int. J. Environ. Pollut., 52, 206–224,
10.1504/IJEP.2013.058461, 2013.Couvidat, F., Kim, Y., Sartelet, K., Seigneur, C., Marchand, N., and Sciare, J.: Modeling secondary organic aerosol
in an urban area: application to Paris, France, Atmos. Chem. Phys., 13, 983–996, 10.5194/acp-13-983-2013, 2013.Das, A., Frenkel, M., Gadalla, N. A. M., Kudchadker, S., Marsh, K. N., Rodgers, A. S., and Wilhoit, R. C.:
Thermodynamic and thermophysical properties of organic nitrogen compounds: Part II. 1- and 2-butanamine, 2-methyl-1-propanamine,
2-methyl-2-propanamine, pyrrole, 1-, 2-, 3-methylpyrrole, pyridine, 2-, 3-,
and 4-methylpyridine, pyrrolidine, piperidine, indole, quinolone,
isoquinoline, acridine, carbazole, phenanthridine, 1- and 2-naphthalenamine,
and 9-methylcarbazole, J. Phys. Chem. Ref. Data, 22, 659–782,
10.1063/1.555925, 1993.Dawson, M. L., Xu, J., Griffin, R. J., and Dabdub, D.: Development of aroCACM/MPMPO 1.0: a model to simulate
secondary organic aerosol from aromatic precursors in regional models, Geosci. Model Dev., 9, 2143–2151,
10.5194/gmd-9-2143-2016, 2016.De Boer, J. G., Posthumus, M. A., and Dicke, M.: Identification of volatiles that are used in discrimination between
plants infested with prey or nonprey herbivores by a predatory mite, J. Chem. Ecol., 30, 2215–2230,
10.1023/B:JOEC.0000048784.79031.5e, 2004.Erb, M., Veyrat, N., Robert, C. A. M., Xu, H., Frey, M., Ton, J., and
Turlings, T. C. J.: Indole is an essential herbivore-induced volatile primal
signal in maize, Nature Communications, 6, 6273, 10.1038/ncomms7273,
2015.Feilberg, A., Liu, D., Adamsen, A. P. S., Hansen, M. J., and Jonassen, K. E. N.: Odorant emissions from intensive pig
production measured by online proton-transfer-reaction mass spectrometry, Environ. Sci. Technol., 44, 5894–5900,
10.1021/es100483s, 2010.Frey, M., Spiteller, D., Boland, W., and Gierl, A.: Transcriptional activation of Igl, the gene for indole formation
in Zea mays: a structure–activity study with elicitor-active N-acyl glutamines from insects, Phytochemistry, 65, 1047–1055,
10.1016/j.phytochem.2003.11.021, 2004.Gentner, D. R., Ormeño, E., Fares, S., Ford, T. B., Weber, R., Park, J.-H., Brioude, J., Angevine, W. M., Karlik,
J. F., and Goldstein, A. H.: Emissions of terpenoids, benzenoids, and other biogenic gas-phase organic compounds from
agricultural crops and their potential implications for air quality, Atmos. Chem. Phys., 14, 5393–5413,
10.5194/acp-14-5393-2014, 2014.Gols, R., Posthumus, M. A., and Dicke, M.: Jasmonic acid induces the production of gerbera volatiles that attract the
biological control agent Phytoseiulus persimilis, Entomol. Exp. Appl., 93, 77–86, 10.1046/j.1570-7458.1999.00564.x, 1999.Griffin, R. J., Dabdub, D., and Seinfeld, J. H.: Secondary organic aerosol 1. Atmospheric chemical mechanism for
production of molecular constituents, J. Geophys. Res., 107, 4332, 10.1029/2001JD000541, 2002a.Griffin, R. J., Dabdub, D., Kleeman, M. J., Fraser, M. P., Cass, G. R., and Seinfeld, J. H.: Secondary organic
aerosol 3. Urban/regional scale model of size- and composition- resolved aerosols, J. Geophys. Res., 107, 4334,
10.1029/2001JD000544, 2002b.Griffin, R. J., Nguyen, K., Dabdub, D., and Seinfeld, J. H.: A coupled hydrophobic-hydrophilic model for predicting
secondary organic aerosol formation, J. Atmos. Chem., 44, 171–190, 10.1023/A:1022436813699, 2003.Griffin, R. J., Dabdub, D., and Seinfeld, J. H.: Development and initial evaluation of a dynamic species-resolved
model for gas phase chemistry and size-resolved gas/particle partitioning associated with secondary organic aerosol
formation, J. Geophys. Res., 110, D05304, 10.1029/2004JD005219, 2005.Hansen, H. K., Rasmussen, P., Fredenslund, A., Schiller, M., and Gmehling, J.: Vapor–liquid equilibria by UNIFAC
group contribution. 5. Revision and extension, Ind. Eng. Chem. Res., 30, 2352–2355, 10.1021/ie00058a017, 1991.Heald, C. L., Jacob, D. J., Park, R. J., Russell, L. M., Huebert, B. J., Seinfeld, J. H., Liao, H., and Weber, R. J.:
A large organic aerosol source in the free troposphere missing from current models, Geophys. Res. Lett., 32, L18809,
10.1029/2005GL023831, 2005.Healy, R. M., Chen, Y., Kourtchev, I., Kalberer, M., O'Shea, D., and Wenger, J. C.: Rapid formation of secondary
organic aerosol from the photolysis of 1-nitronaphthalene: role of naphthoxy radical self-reaction, Environ. Sci. Technol., 46,
11813–11820, 10.1021/es302841j, 2012.Hobbs, P., Webb, J., Mottram, T., Grant, B., and Misselbrook, T.: Emissions of volatile organic compounds originating
from UK livestock agriculture, J. Sci. Food Agr., 84, 1414–1420, 10.1002/jsfa.1810, 2004.Hodzic, A., Jimenez, J. L., Madronich, S., Canagaratna, M. R., DeCarlo, P. F., Kleinman, L., and Fast, J.: Modeling
organic aerosols in a megacity: potential contribution of semi-volatile and intermediate volatility primary organic compounds to
secondary organic aerosol formation, Atmos. Chem. Phys., 10, 5491–5514, 10.5194/acp-10-5491-2010, 2010.Horvath, H.: Atmospheric light-absorption – a review, Atmos. Environ., 27, 293–317,
10.1016/0960-1686(93)90104-7, 1993.Iddon, B., Phillips, G. O., Robbins, K. E., and Davies, J. V.: Radiation chemistry of aqueous solutions of indole and
its derivatives, J. Chem. Soc. B., 1887–1892, 10.1039/J29710001887, 1971.Jiang, F., Liu, Q., Huang, X., Wang, T., Zhuang, B., and Xie, M.: Regional modeling of secondary organic aerosol over
China using WRF/Chem, J. Aerosol Sci., 43, 57–73, 10.1016/j.jaerosci.2011.09.003, 2012.Kirchstetter, T. W. and Thatcher, T. L.: Contribution of organic carbon to wood smoke particulate matter absorption
of solar radiation, Atmos. Chem. Phys., 12, 6067–6072, 10.5194/acp-12-6067-2012, 2012.Laskin, A., Laskin, J., and Nizkorodov, S. A.: Chemistry of atmospheric brown carbon, Chem. Rev., 115, 4335–4382,
10.1021/cr5006167, 2015.Laskin, A., Smith, J. S., and Laskin, J.: Molecular characterization of nitrogen-containing organic compounds in
biomass burning aerosols using high-resolution mass spectrometry, Environ. Sci. Technol., 43, 3764–3771,
10.1021/es803456n, 2009.Lin, P., Laskin, J., Nizkorodov, S. A., and Laskin, A.: Revealing brown carbon chromophores produced in reaction of
methylglyoxal with ammonium sulfate, Environ. Sci. Technol., 49, 14257–14266, 10.1021/acs.est.5b03608, 2015a.Lin, P., Liu, J., Shilling, J. E., Kathmann, S. M., Laskin, J., and Laskin, A.: Molecular characterization of brown
carbon (BrC) chromophores in secondary organic aerosol generated from photo-oxidation of toluene, Phys. Chem. Chem. Phys., 17,
23312, 10.1039/C5CP02563J, 2015b.Lin, P., Aiona, P. K., Li, Y., Shiraiwa, M., Laskin, J., Nizkorodov, S. A., and Laskin, A.: Molecular
characterization of brown carbon in biomass burning aerosol particles, Environ. Sci. Technol., 50, 11815–11824,
10.1021/acs.est.6b03024, 2016.Liu, J. Q., Chen, S. Y., and Ji, B.: Solubility and thermodynamic functions of isatin in pure
solvents, J. Chem. Eng. Data, 59, 3407–3414, 10.1021/je500396b, 2014.McCall, P. J., Turlings, T. C., Lewis, W. J., and Tumlinson, J. H.: Role of
plant volatiles in host location by the specialist parasitoidMicroplitis
croceipes Cresson (Braconidae: Hymenoptera), J. Insect Behav., 6, 625–639,
10.1007/BF01048128, 1993.Misztal, P. K., Hewitt, C. N., Wildt, J., Blande, J. D., Eller, A. S. D., Fares, S., Gentner, D. R., Gilman, J. B.,
Graus, M., Greenberg, J., Guenther, A. B., Hansel, A., Harley, P., Huang, M., Jardine, K., Karl, T., Kaser, L., Keutsch, F. N.,
Kiendler-Scharr, A., Kleist, E., Lerner, B. M., Li, T., Mak, J., Nolscher, A. C., Schnitzhofer, R., Sinha, V., Thornton, B.,
Warneke, C., Wegener, F., Werner, C., Williams, J., Worton, D. R., Yassaa, N., and Goldstein, A. H.: Atmospheric benzenoid
emissions from plants rival those from fossil fuels, Sci. Rep.-UK, 5, 12064, 10.1038/srep12064, 2015.Moise, T., Flores, J.M, and Rudich, Y.: Optical properties of secondary organic aerosols and their changes by
chemical processes, Chem. Rev., 115, 4400–4439, 10.1021/cr5005259, 2015.Nah, T., Chan, M., Leone, S. R., and Wilson, K. R.: Real time in situ chemical characterization of submicrometer
organic particles using direct analysis in real time-mass spectrometry, Anal. Chem., 85, 2087–2095, 10.1021/ac302560c,
2013.Nguyen, K. and Dabdub, D.: NOx and VOC control and its effects on the formation of aerosols, Aerosol Sci. Tech., 36,
560–572, 10.1080/02786820252883801, 2002.Niinemets, U., Kannaste, A., and Copolovici, L.: Quantitative patterns between plant volatile emissions induced by
biotic stresses and the degree of damage, Front. Plant. Sci., 4, 1–15, 10.3389/fpls.2013.00262, 2013.Novotna, P., Boon, J. J., Horst, J., and Pacakova, V.: Photodegradation of indigo in dichloromethane solution,
Color. Technol., 119, 121–127, 10.1111/j.1478-4408.2003.tb00161.x, 2003.Pankow, J. F. and Asher, W. E.: SIMPOL.1: a simple group contribution method for predicting vapor pressures and
enthalpies of vaporization of multifunctional organic compounds, Atmos. Chem. Phys., 8, 2773–2796,
10.5194/acp-8-2773-2008, 2008.Price, D. J., Clark, C. H., Tang, X. C., Cocker, D. R., Purvis-Roberts, K. L., and Silva, P. J.: Proposed chemical
mechanisms leading to secondary organic aerosol in the reactions of aliphatic amines with hydroxyl and nitrate radicals,
Atmos. Environ., 96, 135–144, 10.1016/j.atmosenv.2014.07.035, 2014.Romonosky, D. E., Ali, N. N., Saiduddin, M. N., Wu, M., Lee, H. J., Aiona, P. K., and Nizkorodov, S. A.: Effective
absorption cross sections and photolysis rates of anthropogenic and biogenic secondary organic aerosols, Atmos. Environ., 130,
172–179, 10.1016/j.atmosenv.2015.10.019, 2015a.Romonosky, D. E.; Laskin, A.; Laskin, J.; and Nizkorodov, S. A.: High-resolution mass spectrometry and molecular
characterization of aqueous photochemistry products of common types of secondary organic aerosols, J. Phys. Chem. A, 119,
2594–2606, 10.1021/jp509476r, 2015b.SCAQMD: Final 2012 Air Quality Management Plan, Tech. rep., South Coast Air Quality Management District, Diamond Bar,
California, available at:
http://www.aqmd.gov/home/library/clean-air-plans/air-quality-mgt-plan/final-2012-air-quality-management-plan, last access:
15 January 2015.
Schmelz, E. A., Alborn, H. T., Banchio, E., and Tumlinson, J. H.:
Quantitative relationships between induced jasmonic acid levels and volatile
emission in Zea mays during Spodoptera exigua herbivory, Planta, 216,
665–673, 2003.Silva, P. J., Erupe, M. E., Price, D., Elias, J., Malloy, Q. G. J., Li, Q., Warren, B., and Cocker, D. R. III:
Trimethylamine as precursor to secondary organic aerosol formation via nitrate radical reaction in the atmosphere,
Environ. Sci. Technol., 42, 4689–4696, 10.1021/es703016v, 2008.Sun, H., Biedermann, L., and Bond, T. C.: Color of brown carbon: a model for
ultraviolet and visible light absorption by organic carbon aerosol, Geophys.
Res. Lett., 34, L17813, 10.1029/2007GL029797, 2007.Suzuki, T., Ohtaguchi, K., and Koide, K.: Application of principal components analysis to calculate Henry's constant
from molecular structure, Comput. Chem., 16, 41–52, 10.1016/0097-8485(92)85007-L, 1992.Thompson, J. E., Hayes, P. L., Jimenez, J. L., Adachi, K., Zhang, X., Liu, J., Weber, R. J., and Buseck, P. R.:
Aerosol optical properties at Pasadena, CA during CalNex 2010, Atmos. Environ., 55, 190–200,
10.1016/j.atmosenv.2012.03.011, 2012.
Turlings, T. C., Tumlinson, J. H., and Lewis, W. J.: Exploitation of herbivore-induced plant odors by host-seeking
parasitic wasps, Science, 250, 1251–1253, 1990.
Turlings, T. C. and Wäckers, F.: Recruitment of predators and parasitoids
by herbivore-injured plants, in: Advances in Insect Chemical Ecology,
Cambridge University Press, Cambridge, 21–75, 2004.Volkamer, R., Jimenez, J. L., San Martini, F., Dzepina, K., Zhang, Q., Salcedo, D., Molina, L. T., Worsnop, D. R.,
and Molina, M. J.: Secondary organic aerosol formation from anthropogenic air pollution: rapid and higher than expected,
Geophys. Res. Lett., 33, L17811, 10.1029/2006GL026899, 2006.Vutukuru, S., Griffin, R. J., and Dabdub, D.: Simulation and analysis of secondary organic aerosol dynamics in the
South Coast Air Basin of California, J. Geophys. Res., 111, D10S12, 10.1029/2005JD006139, 2006.Walser, M. L., Desyaterik, Y., Laskin, J., Laskin, A., and Nizkorodov, S. A.: High-resolution mass spectrometric
analysis of secondary organic aerosol produced by ozonation of limonene, Phys. Chem. Chem. Phys., 10, 1009–1022,
10.1039/B712620D, 2008.Wright, D. W., Eaton, D. K., Nielsen, L. T., Kuhrt, F. W., Koziel, J. A., Spinhirne, J. P., and Parker, D. B.:
Multidimensional gas chromatography – olfactometry for the identification and prioritization of malodors from confined animal
feeding operations, J. Agr. Food Chem., 53, 8663–8672, 10.1021/jf050763b, 2005.Yuan, B., Coggon, M. M., Koss, A. R., Warneke, C., Eilerman, S., Peischl, J., Aikin, K. C., Ryerson, T. B., and de
Gouw, J. A.: Emissions of volatile organic compounds (VOCs) from concentrated animal feeding operations (CAFOs): chemical
compositions and separation of sources, Atmos. Chem. Phys., 17, 4945–4956, 10.5194/acp-17-4945-2017, 2017.Yuan, J. S., Himanen, S. J., Holopainen, J. K., Chen, F., and Stewart, C. N.: Smelling global climate change:
mitigation of function for plant volatile organic compounds, Trends Ecol. Evol., 24, 323–331, 10.1016/j.tree.2009.01.012,
2009.Zhuang, X., Fiesselmann, A., Zhao, N., Chen, H., Frey, M., and Chen, F.: Biosynthesis and emission of insect
herbivory-induced volatile indole in rice, Phytochemistry, 73, 15–22, 10.1016/j.phytochem.2011.08.029, 2012.Zito, P., Doetterl, S. Sajeva, M.: Floral volatiles in a sapromyiophilous plant and their importance in attracting
house fly pollinators, J. Chem. Ecol., 41, 340–349, 10.1007/s10886-015-0568-8, 2015.