Atmospheric Environment 138 (2016) 99e107

Contents lists available at ScienceDirect

Atmospheric Environment journal homepage: www.elsevier.com/locate/atmosenv

Online molecular characterization of fine particulate matter in Port Angeles, WA: Evidence for a major impact from residential wood smoke Cassandra J. Gaston a, 1, Felipe D. Lopez-Hilfiker a, 2, Lauren E. Whybrew a, Odelle Hadley b, Fran McNair b, Honglian Gao c, Daniel A. Jaffe a, c, Joel A. Thornton a, * a b c

Department of Atmospheric Sciences, University of Washington, Seattle, WA 98195, USA Olympic Region Clean Air Agency, Olympia, WA 98502, USA Physical Sciences Division, University of Washington, Bothell, WA 98011, USA

h i g h l i g h t s
A novel analytical technique was used to quantify wood smoke markers in near real-time.
Levoglucosan, nitroaromatics, and methoxyphenols dominated PM, particularly at night.
Residential wood smoke is a major source of wintertime PM in Port Angeles, WA.
Levoglucosan forms low volatility products in the presence of acids.

a r t i c l e i n f o

a b s t r a c t

Article history: Received 1 February 2016 Received in revised form 30 April 2016 Accepted 8 May 2016 Available online 10 May 2016

We present on-line molecular composition measurements of wintertime particulate matter (PM) during 2014 using an iodide-adduct high-resolution, time-of-flight chemical ionization mass spectrometer (HRTOF-CIMS) coupled to a Filter Inlet for Gases and AEROsols (FIGAERO). These measurements were part of an intensive effort to characterize PM in the region with a focus on ultrafine particle sources. The technique was used to detect and quantify different classes of wood burning tracers, including levoglucosan, methoxyphenols, and nitrocatechols, among other compounds in near real-time. During the campaign, particulate mass concentrations of compounds with the same molecular composition as levoglucosan ranged from 0.002 to 19 mg/m3 with a median mass concentration of 0.9 mg/m3. Wood burning markers, in general, showed a strong diurnal pattern peaking at night and in the early morning. This diurnal profile combined with cold, stagnant conditions, wind directions from predominantly residential areas, and observations of lower combustion efficiency at night support residential wood burning as a dominant source of wintertime PM in Port Angeles. This finding has implications for improving wintertime air quality in the region by encouraging the use of high efficiency wood-burning stoves or other cleaner home heating options throughout the relevant domain. © 2016 Published by Elsevier Ltd.

Keywords: Biomass burning aerosol Residential wood burning Volatility Air quality Levoglucosan Nitrophenol

1. Introduction In the U.S. Pacific Northwest, PM episodes that approach or exceed regulatory air quality thresholds often occur in the

* Corresponding author. E-mail address: [email protected] (J.A. Thornton). 1 Now at the Rosenstiel School of Marine and Atmospheric Science, University of Miami, Miami, FL 33149, USA. 2 Now at the Paul Scherrer Institut, 5232 Villigen PSI, Switzerland. http://dx.doi.org/10.1016/j.atmosenv.2016.05.013 1352-2310/© 2016 Published by Elsevier Ltd.

wintertime, due to suppressed vertical mixing and local emissions (Hand et al., 2012; Larson et al., 2004; Ogulei, 2010). Winter episodes occur on synoptic timescales, associated with prolonged cold periods and shallow mixed layers, particularly during nighttime, which promote enhanced residential heating often by wood stoves and fireplaces. However, direct compositional links to residential biofuel sources during these episodes are lacking. Pollution events induced from residential wood burning emissions are globally important for both human health and climate (Fine et al., 2004; Gorin et al., 2006; Hedberg et al., 2006; Khalil and

100

C.J. Gaston et al. / Atmospheric Environment 138 (2016) 99e107

Rasmussen, 2003; Larson et al., 2004; Naeher et al., 2007; Ogulei, 2010; Rogge et al., 1998; Schauer and Cass, 2000; Szidat et al., 2007; Ward et al., 2006). Exposure to wood smoke particles is associated with decreased lung function, asthmatic symptoms, and is estimated to be responsible for 1e2 million premature deaths annually (Bolling et al., 2009; Naeher et al., 2007). Further, black carbon and light-absorbing “brown” carbon released from wood burning impact climate by strongly absorbing light and affecting budgets of cloud condensation and ice nuclei (Carrico et al., 2010; Mohr et al., 2013; Petters et al., 2009; Ramanathan and Carmichael, 2008; Saleh et al., 2014; Sun et al., 2007; Weimer et al., 2009). The impact of wood combustion on human health and climate underscore the need to better understand and characterize various sources of wood smoke PM. Flaming and smoldering combustion conditions volatilize and thermally breakdown compounds found in plant materialdnamely cellulose, hemicelluloses, lignin, proteins, and plant waxesdforming compounds that can act as unique wood smoke tracers in PM (Andreae and Merlet, 2001). The most universally used tracer for biomass burning is levoglucosan, formed from the combustion of cellulose (Fine et al., 2002; Fraser and Lakshmanan, 2000; Hennigan et al., 2010; Hoffmann et al., 2010; Nolte et al., 2001; Rogge et al., 1998; Simoneit et al., 1999; Sullivan et al., 2014). Levoglucosan represents a significant fraction of PM impacted by residential wood combustion during winter. This compound is one of the most abundant individual organic compounds found in atmospheric aerosols (Fine et al., 2002; Nolte et al., 2001). However, levoglucosan provides little detail about the type of biomass burning (e.g., woodstove vs. fireplace vs. open biomass burning), combustion conditions (e.g., flaming vs. smoldering), and fuel being combusted, which are all needed to accurately apportion the various biomass burning related sources of PM. Guaiacol- and syringol-derived methoxyphenols, produced from the pyrolysis of lignins, and resin acids are typically used to distinguish between softwood (e.g., conifers such as pine) and hardwood combustion (e.g., oak) (Edye and Richards, 1991; Hawthorne et al., 1992; McDonald et al., 2000; Rogge et al., 1998; Schauer and Cass, 2000; Schauer et al., 2001; Simoneit, 2002; Simoneit et al., 1993; Simpson et al., 2005). In addition to levoglucosan and methoxyphenols, other wood smoke compounds used as chemical tracers include plant wax esters, carboxylic acids, and toxic compounds including polycyclic aromatic hydrocarbons (PAHs), nitrophenols, nitrocatechols, nitriles, and amides (Elias et al., 1999; Harrison et al., 2005; Iinuma et al., 2010; Ma and Hays, 2008; Mohr et al., 2013; Oros and Simoneit, 2001; Simoneit, 2002; Simoneit et al., 1993, 2003). A key challenge associated with quantifying chemical tracers in wood smoke is a lack of methods that can unambiguously identify and quantify multiple highly oxygenated and nitrogen-containing markers with a time resolution capable of distinguishing between common time-varying sources such as traffic, home heating, and combustion phases. Chemical measurements of wood smoke PM have typically been off-line, involving the collection of PM on filters for several hours, followed by extraction and analysis using gas chromatography (GC) (Bari et al., 2009; Edye and Richards, 1991; Elias et al., 1999; Fine et al., 2004; Fraser and Lakshmanan, 2000; Hawthorne et al., 1992; Hedberg et al., 2006; Ma and Hays, 2008; McDonald et al., 2000; Nolte et al., 2001; Oros and Simoneit, 2001; Rogge et al., 1998; Schauer and Cass, 2000; Schauer et al., 2001; Simoneit et al., 1993; Simpson et al., 2005). This method is quantitative and well-suited for analyzing hydrocarbons; however, the drawbacks include the sample preparation, which involves extraction and derivatization of oxygenated and high molecular weight compounds, unresolved complex mixtures (UCM), and low time resolution. Recently, Engling and co-workers developed a

liquid-based method for improved quantification of oxygenated compounds in biomass burning aerosol, namely levoglucosan, using high performance anion-exchange chromatography with pulsed amperometric detection (HPAEC-PAD) (Engling et al., 2006). This technique requires no sample preparation or derivatization, has faster separation than GC, and a low limit of detection (Christian et al., 2010; Engling et al., 2006; Gorin et al., 2006; Sullivan et al., 2014). Electrospray ionization-mass spectrometry (ESI-MS) has recently been used to detect high molecular weight oxygenated and nitrogen-containing polar compounds found in biomass burning aerosol that cannot be detected by GC-MS (Laskin et al., 2009; Smith et al., 2009), but lacks reproducible quantification. While both HPAEC-PAD and ESI-MS methods have improved the detection of polar compounds found in wood smoke particles, neither technique, as employed so far, provides high time resolution. Online techniques, such as aerosol mass spectrometry (AMS), have been used with positive matrix factorization (PMF) to quantify biomass burning aerosol in real-time (Aiken et al., 2010; Weimer et al., 2008); however, the technique utilizes electron impact ionization resulting in extensive fragmentation, which precludes detection and identification of specific molecular tracers. Here we present measurements of ambient fine PM using highresolution, time-of-flight chemical ionization mass spectrometry (HR-TOF-CIMS) coupled to a Filter Inlet for Gases and AEROsols (FIGAERO) (Lopez-Hilfiker et al., 2014). This online technique combines soft ionization with a high-resolution, time-of-flight mass analyzer allowing the detection and quantification of oxygenated and nitrogen-containing compounds on the molecular level at near real-time with no sample preparation (Lee et al., 2014; Lopez-Hilfiker et al., 2014). Further, gas- and particle-phase compounds are detected simultaneously with this technique, and particulate compounds are thermally resolved providing additional information on volatility. A predecessor of this technique was used to quantify nitrophenols in submicron PM in Detling, UK at hourly time resolution (Mohr et al., 2013). In this work, diel transformations and volatility relevant to wood burning organic aerosol are probed by simultaneously detecting compounds with molecular compositions consistent with levoglucosan, methoxyphenols, nitrophenols, resin acids, lignans, and other oxygenated and nitrogen-containing markers in wood smoke particles. We provide insight into the dominant sources of fine PM during several poor air quality episodes in Port Angeles, WA. 2. Materials and methods Measurements were conducted in Port Angeles, WA at the Port Angeles Fire Station. The site is located at 48.115 N, 123.436 W (see Figure S1 of the Supporting Information (SI) for a site map of the location). Residential communities are primarily to the south and southwest while industrial and port sources are to the north of this site. Continuous measurements of gases, particle size distributions and mass concentrations, and meteorological conditions have been conducted at this site since December 2013. Gas-phase measurements of CO (Thermo Fisher Scientific, Model 48C TLE) and CO2 (LICOR Inc., Model LI-840A) that were zeroed every 2 h and calibrated daily and measurements of PM2.5 using an optical particle counter (OPC, MetOne Model 212 Profiler) were conducted during this intensive campaign. Additional details can be found in the SI; we focus on measurements made using HR-TOF-CIMS described below. 2.1. FIGAERO HR-TOF-CIMS measurements An iodide-adduct high-resolution, time-of-flight chemical ionization mass spectrometer (HR-TOF-CIMS) (Lee et al., 2014) operated with a custom-built Filter Inlet for Gases and AEROsols

C.J. Gaston et al. / Atmospheric Environment 138 (2016) 99e107

(FIGAERO) was deployed from January 21-March 6, 2014. The instrument detects oxygenated, halogenated, and nitrogencontaining compounds using iodide-adduct CIMS, which is a soft ionization technique that nearly exclusively forms molecular ions clustered with iodide (Lee et al., 2014). The mass analyzer is a highresolution, time-of-flight mass spectrometer with high massaccuracy (<20 ppm) and mass-resolving power (R > 5500), which when combined with the large mass defect of iodide, allows for the determination of elemental formulas (Lee et al., 2014). The instrument was housed in a waterproof box placed directly on the roof of the building, allowing short sampling inlets for both particles and gases. The gas-phase inlet was 3/400 OD Teflon and ~1 m in length while the particle-phase inlet was 1/200 OD stainless steel and ~1 m in length. A custom-built inertial impactor with an aerodynamic cut-point diameter of ~3 mm was installed at the tip of the particle inlet. Ambient air was continuously drawn through the two sampling inlets using an auxiliary scroll pump. The inlets were directly coupled to the FIGAERO manifold, described in detail elsewhere (Lopez-Hilfiker et al., 2014). Briefly, the FIGAERO mounts directly to the ion-molecule reaction (IMR) region of the mass spectrometer, and consists of two sampling ports for separate particle and gas-phase analysis. In gasphase sampling mode, ambient air is sub-sampled from the gasphase inlet through a critical orifice directly into the IMR at a flow of 2 lpm; the total flow through the gas-phase inlet is 20 lpm giving a residence time of <1 s. As there is not additional heating and air spends ~100 ms in the IMR, only gas-phase species are ionized and detected. Gas-phase concentrations have not been corrected for possible losses to inlet walls due to diffusion, which would require a maximum correction of 27% based on residence time, tube diameter, and tube length. While ambient gases are being analyzed with the mass spectrometer, particles are collected on a 24 mm Teflon filter held in a movable tray within the FIGAERO. In this case, ~6 slpm of ambient air was drawn through the particlephase inlet for 30 min giving a residence time of ~1.3 s. After the particle collection period, a linear actuator slides the tray and filter into the thermal desorption region. In this position, the tray prevents gas-phase inlet flow from entering the IMR, and opens an orifice directly in line with the filter and heater. A 2 slpm stream of dry UHP N2 is continually passed across the filter thermally ramped at a rate of 10  C/minute to 200  C, inducing thermal desorption of compounds from the collected particles into the gas-phase. The desorption flow is drawn through the orifice, located immediately downstream of the filter, and into the IMR for ionization and detection by the mass spectrometer. The ramped temperature program provides additional separation of compounds based on their effective relative volatility. A full desorption cycle takes 70 min giving near hourly time resolution for the data presented herein. After three consecutive desorptions, particle-phase background determinations were performed by actuating a second filter upstream of the particle-phase inlet, and then performing a full sampling period and thermal desorption as described above. This background subtraction method corrects for adsorbed semivolatile gases without the use of a denuder. Particle-phase data is unavailable from 02/02-02/05 and 02/09-02/18 when the linear actuator failed. This work focuses on organic compounds detected in the particle-phase, which had compositions consistent with organic acids, methoxyphenols, nitrophenols, and monosaccharide derivatives (e.g., levoglucosan), as expected if biomass burning was a prevalent aerosol source during the measurements. Compounds assigned to the individual ions detected by the HR-TOF in this work denote the most probable compound for a given elemental formula fit to an observed ion peak. No structural information is derived from this technique. This technique cannot resolve isomeric/

101

isobaric structures. We also note that oligomers are subjected to thermal decomposition upon heating during a desorption causing accretion products to break down into monomeric units (LopezHilfiker et al., 2014). As such, compounds corresponding to the detected elemental formulas described in this work do contain contributions from isomeric structures and from the breakdown of accretion products. Calibrations were performed using standard compounds found in wood smoke; the details can be found in the SI and in Table S1. All data are presented in local Pacific Standard Time (PST). 3. Results 3.1. Residential wood smoke contribution to wintertime PM2.5 in Port Angeles, WA On average, one of the single highest ion intensities in the mass spectra of particle-phase data taken with FIGAERO-HR-TOF-CIMS was at an exact mass of 288.9578 Th, which was often on par with the reagent ion (I) intensity (see Figure S2). The best high resolution fits for this ion peak and its isotope required C6H10O5I as the main compound, likely corresponding to iodide clustered with the biomass burning tracer levoglucosan (see Figure S3 for the high resolution fit). For simplicity, we refer to all compounds represented by the measured composition C6H10O5 as levoglucosan. Related isomers, including galactosan and mannosan are also found in wood smoke, although typically in much lower concentrations compared to levoglucosan (Fine et al., 2004; Nolte et al., 2001). The volatility dimension, which when calibrated using authentic standards, further constrains the chemical assignment. For example, the measured volatility exhibited by C6H10O5I in ambient particles was identical to that for a levoglucosan standard deposited by injection to the FIGAERO manifold supporting the assignment of this ion peak to levoglucosan and closely related exact mass isomers (see Figure S4). Figure S2 and Table 1 lists compositions and representative potential compounds associated with such compositions observed simultaneously with levoglucosan, including methoxyphenols and nitrocatechols (Harrison et al., 2005; Hawthorne et al., 1992; Iinuma et al., 2010; Mohr et al., 2013). The time series of levoglucosan mass concentrations for the first 5 days of the campaign, when some of the highest concentrations were observed, is shown in Fig. 1. The inset shows the average diurnal profile of levoglucosan as a percentage of total PM2.5 measured by a co-located Optical Particle Counter (OPC). See SI for details on converting levoglucosan signals and OPC measurements into mass concentrations. The time series in Fig. 1 highlights the ability of the technique to provide molecular-level detection and quantification of levoglucosan at ~ hourly resolution. We estimate an uncertainty of a factor of 2 associated with the reported mass concentrations of levoglucosan mostly due to errors during calibration, and to unresolved isomeric interferences in the ambient data from compounds such as mannosan and galactosan. It is also important to note that during desorptions reagent ion concentrations were partially titrated suggesting that our levoglucosan concentrations are likely lower estimates. The desorption signal is normalized to the total reagent ion current (I þ I(H2O)), which corrects for titrations on the order of 20%. Several lines of evidence support a residential wood burning source for the observed mass concentrations of levoglucosan and related isomers. First, the temporal profile exhibits persistently high mass concentrations throughout the campaign. During the campaign no nearby wildfires, which typically burn in the summertime in Washington State, or agricultural burning was detected by MODIS, and we did not identify any long-range transport as indicated by 48-h HYSPLIT back trajectories. A diurnal pattern in

102

C.J. Gaston et al. / Atmospheric Environment 138 (2016) 99e107

Table 1 Molecular formulas, molecular weights (MW), molecular weights of ion peaks coupled to iodide, and the most likely compounds that correspond to those ion peaks are presented. The different markers are divided up by different classes of biomass burning tracers (e.g., monosaccharide derivatives, methoxyphenols, etc.). Detected formula

Example compound

Monosaccharide derivatives C6H10O5 Levoglucosan/Mannosan/Galactosan C6H12O6 Glucose/Mannose/Galactose Guaiacol derivatives C7H8O2 Guaiacol C8H10O2 Creosol C9H10O2 Vinyl guaiacol/Phenyl propanoic acid C9H12O2 Ethyl guaiacol Vanillin C8H8O3 C10H12O2 Eugenol C9H10O3 Acetovanillone/Veratraldehyde C8H8O4 Vanillic acid C10H10O3 Coniferyl aldehyde C10H12O3 Coniferol/Guaiacyl acetone C10H14O3 Vanillyl propanol C9H10O4 Homovanillic acid/Veratric acid C10H14O3 Homovanillyl alcohol/Guaiacyl Propanol Lignans C17H20O4 Divanillylmethane C18H22O4 Divanillylethane C20H24O5 Shonanin C20H22O6 Matairesinol Syringol derivatives C8H10O3 Syringol C9H12O3 Methyl syringol/Trimethoxybenzene C11H16O3 Propylsyringol C10H12O4 Acetosyringone/Homoveratric acid C9H10O5 Syringic acid C11H14O4 Syringyl acetone/Propionyl syringol C17H20O5 Vanillyl syringyl Aromatics C6H6O Phenol C6H6O2 Catechol Resin acids/Diterpenoids C20H28O2 Dehydroabietic acid C20H30O2 Pimaric acid/Abietic acid C20H26O3 Oxodehydroabietic acid Nitroaromatics C6H5NO3 Nitrophenol C7H7NO3 Methyl nitrophenol C6H5NO4 Nitrocatechol C7H7NO4 Methyl nitrocatechol C6H4N2O5 Dinitrophenol

MW

MW þ I

162 180

289 307

124 138 150 152 152 164 166 168 178 180 182 182 182

251 265 277 279 279 291 293 295 305 307 309 309 309

288 302 344 358

415 429 471 485

154 168 196 196 198 210 304

281 295 323 323 325 337 431

94 110

221 237

300 302 314

427 429 441

139 153 155 169 184

266 280 282 296 311

the planetary boundary layer height was observed using the Modern-Era Retrospective analysis for Research and Applications (MERRA) and the Goddard Earth Observing System Model, Version 5 (GEOS-5) product. The boundary layer height was shallower at night when levoglucosan and other biomass burning tracers peaked, which would concentrate wood burning emissions near the surface. We do note, however, that not all compounds exhibited a nighttime maximum, including nitrophenols. Therefore, stagnant conditions in combination with nighttime wood burning caused the observed trends in PM. Second, diurnal profiles of levoglucosan, and quantified methoxyphenols (e.g., vanillin, guaiacol, vanillic acid) and nitrocatechols (e.g., nitrocatechol, methyl nitrocatechol), all peak during the evening, starting at 17:00 and again in the early morning starting at 5:00 (see Fig. 2). These trends are consistent with residential wood burning practices (Gorin et al., 2006; Hedberg et al., 2006; Mohr et al., 2013; Szidat et al., 2007; Ward et al., 2006). Third, meteorological conditions during this campaign consisted of cold, stagnant conditions (see Figure S5) with a dominant wind direction that was southerly and southwesterly from nearby residential communities. Fig. 3 shows that the highest mass concentrations of levoglucosan and its related

Fig. 1. Temporal profile of the mass concentration of C6H10O5 (levoglucosan) measured in the particle-phase using FIGAERO-HR-TOF-CIMS from 01/22-01/27. Inset is the diurnal average of levoglucosan observed during the entire campaign as the percentage of total mass measured by a co-located OPC.

isomers were measured during the evening and early morning when southerly and southwesterly winds were prevalent, consistent with a residential as opposed to an industrial source for the observed biomass burning aerosol. The diurnal ratio of (UV) light absorbing organic carbon (OC) to black carbon (BC) observed by the aethalometer also increased at night when biomass burning markers increased, consistent with a wood burning source of PM (see Figure S6). Fourth, the enhancement ratio of CO/CO2 also peaked in the morning and in the evening, consistent with residential wood burning practices (see Fig. 4). While the morning peak in the CO/CO2 ratio could be due to traffic, we note that levoglucosan and other wood smoke tracers were also elevated during this time of day including gas-phase guaiacol (see Fig. 4). Levoglucosan related tracers were also highly correlated with the CO/CO2 ratio (R2 ¼ 0.68, see Fig. 4) suggesting a wood burning source with a lower combustion efficiency consistent with residential wood burning (Andreae and Merlet, 2001; Dhammapala et al., 2006; Hedberg et al., 2006). As shown in Table 1, ion compositions consistent with methoxyphenols (C8H8O3I, e.g., vanillin; C8H8O4I, e.g., vanillic acid; C10H10O3I, e.g., coniferyl aldehyde) and syringol functional groups were observed (C9H10O5I, e.g., syringic acid). However, a sinapyl aldehyde composition was not observed, typically a unique marker to hardwood combustion (Fine et al., 2002), while guaiacolderived methoxyphenol compositions were more abundant, suggesting that softwoods were the dominant local biofuel consistent with residential heating practices in the Western United States (Fine et al., 2002). Further, the detection of compositions that correspond to resin acids (C20H28O2I, e.g., dehydroabietic acid), lignans derived from coniferyl alcohol (C20H22O6I, e.g., matairesinol), and C8H12O6I that has been observed during controlled burns of ponderosa pine needles and sticks are all consistent with residential combustion of softwoods (Bari et al., 2009; Edye and Richards, 1991; Elias et al., 1999; Fine et al., 2002, 2004; Hawthorne et al., 1992; McDonald et al., 2000; Nolte et al., 2001; Rogge et al., 1998; Schauer and Cass, 2000; Schauer et al., 2001; Simoneit, 2002; Simoneit et al., 1993; Simpson et al., 2005; Smith et al., 2009). The dominance of residential wood burning in Port Angeles, WA is also not surprising given the high contribution of wood smoke to wintertime PM2.5 levels in other regions of Washington State including Pierce County/Tacoma, Olympia, and Seattle

C.J. Gaston et al. / Atmospheric Environment 138 (2016) 99e107

Fig. 2. Diurnal mass concentration profiles of C6H10O5 (levoglucosan) (magenta line with circles), methoxyphenol derivatives (MPs) (orange line with diamonds), and nitrocatechols (NCs) (blue line with squares) found in wood smoke particles. Error bars represent one standard deviation. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 3. Levoglucosan mass concentration as a function of wind direction.

(Khalil and Rasmussen, 2003; Larson et al., 2004; Ogulei, 2010; Simpson et al., 2005).

3.2. Comparison to wood smoke marker measurements in other locations Fig. 1 shows levoglucosan concentrations of 0.002e19 mg/m3 (median 0.9 mg/m3) which represents up to 40% of total fine particle mass (median 19%) (see also Table 2). By comparison, ammonium nitrate mass concentrations ranged from <1 ng/m3 to 5 mg/m3, with a median mass concentration of 0.3 mg/m3, constituting 4% of the observed PM2.5, on average, with most of the ammonium nitrate mass peaking during the daytime rather than at night (see SI for ammonium nitrate mass concentration calculations). Previous studies report that levoglucosan represents 1.9e46% of PM2.5 emitted from biomass burning with most of the spread depending on the fuel type and burn conditions (see Table S2 and SI for discussion) (Christian et al., 2010; Engling et al., 2006; Fine et al., 2004; Hays et al., 2002; Mazzoleni et al., 2007; Nolte et al., 2001; Schauer

103

Fig. 4. Diurnal profile of the CO/CO2 ratio (dashed green line) and guaiacol (dashed purple line) measured by the FIGAERO-TOF-CIMS (top panel) and correlation between levoglucosan mass concentrations and the CO/CO2 ratio (bottom panel). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

and Cass, 2000; Schauer et al., 2001; Ward et al., 2006). Combining data on heating appliances, wood burned, and emissions profiles of different chemical tracers from wood burning, Fine et al., 2002 estimated that levoglucosan represented 16% of PM2.5 from residential wood burning in Washington State in 1995 using a chemical mass balance receptor model (Fine et al., 2002). The fraction of levoglucosan and related tracers reported here (up to 40%) is on the higher end of these reported values. As noted earlier, FIGAERO-ToFCIMS cannot separate isomers with the same elemental formulas meaning the detected signal at C6H10O5I likely contains contributions from mannosan and galactosan, both biomass burning tracers, in addition to levoglucosan. However, previous work using techniques that can separate these compounds report that mannosan represented, at most, 20% of the emitted mass from a wood stove with an elemental formula of C6H10O5 while galactosan only represented up to 5% (Bari et al., 2009; Fine et al., 2004; Nolte et al., 2001). This means that of the 40% of PM2.5 that is explained by C6H10O5I, up to 8% could be mannosan and 2% could be galactosan while levoglucosan would still represent up to 30% of PM2.5. Notably, Engling and co-workers detected higher mass fractions of levoglucosan using HPAEC-PAD (17e49% of PM2.5) compared to previous work, which quantified levoglucosan using GC sometimes without prior derivatization. We speculate that our method of detecting and quantifying levoglucosan, which does not suffer from losses due to extraction and is more sensitive to polar compounds, similarly explains the higher fractions of levoglucosan observed herein compared to previous work. In addition to levoglucosan, the instrument response to several other biomass burning related compounds was quantified including a suite of methoxyphenols, nitrophenols, and nitrocatechols (see Table S1 for calibration factors and Table 2 for the median and maximum mass concentrations). The methoxyphenols quantified were vanillin, vanillic acid, and guaiacol or 4methoxyphenol. Applying the instrument response for the specific isomers to the methoxyphenol compositions measured in the field resulted in concentrations ranging from <1 to 366 ng/m3, <1e683 ng/m3, and <1e1150 ng/m3, respectively, and an overall range for the 3 combined of <1e1654 ng/m3 with a median mass concentration of 63 ng/m3 for all of the quantified methoxyphenols. These concentrations compare well to methoxyphenols measured

104

C.J. Gaston et al. / Atmospheric Environment 138 (2016) 99e107

Table 2 Molecular formulas, molecular weights (MW), molecular weights of ion peaks coupled to iodide, and median (maximum) mass concentrations of calibrated biomass burning tracer compounds likely observed in Port Angeles are presented. Detected formula Monosaccharide derivatives C6H10O5 Methoxyphenol derivatives C7H8O2 C8H8O3 C8H8O4 Nitroaromatics C7H7NO3 C6H5NO3 C6H4N2O5 C6H5NO4 C7H7NO4

Example compound

MW

MW þ I

Median (max) (ng/m3)

Levoglucosan

162.14

289.14

949 (18,900)

4-methoxyphenol/Guaiacol Vanillin Vanillic acid Sum methoxyphenols

124.14 152.15 168.14

251.14 279.15 295.14

28 10 28 63

4-methyl-2-nitrophenol 4-nitrophenol Dinitrophenol Sum nitrophenols 4-nitrocatechol Methyl-nitrocatechol Sum nitrocatechols

153.14 139.11 184.11

280.14 266.11 311.11

155.11 169.14

282.11 296.14

5.38 (209) 0.57 (7) 1.06 (243) 6 (455) 27 (736) 31 (678) 58 (1414)

in Fresno and Bakersfield (up to 876 ng/m3) (Schauer and Cass, 2000), Salt Lake City and Minneapolis (231e1290 ng/m3 for guaiacol-derived methoxyphenols) (Hawthorne et al., 1992), but are higher than methoxyphenols measured in the Seattle region, which ranged from <0.1e22 ng/m3 (Simpson et al., 2005). The higher methoxyphenol concentrations in Port Angeles during the wintertime likely reflect the higher contribution of wood burning to PM in Port Angeles than in the Seattle area. As shown in Table S1, 4-nitrophenol and 4-nitrocatechol were used to calibrate the instrument response to compositions corresponding to nitrophenols (e.g., nitrophenol, dinitrophenol, and methyl nitrophenol) and nitrocatechols (e.g., nitrocatechol and methyl nitrocatechol) (see SI for details). Based on these calibrations, nitrophenols were less abundant than nitrocatechols with a mass concentration range of <1e455 ng/m3 compared to <1e1414 ng/m3, respectively. Overall, the mass concentration for all quantified nitrophenol and nitrocatechol compositions ranged from <1 to 1446 ng/m3 with a median concentration of 55 ng/m3, which is higher than the total nitrophenols and nitrocatechols quantified in Detling, UK that ranged from <1 to 90 ng/m3 using a similar instrument (Mohr et al., 2013). To compare the sources of phenolic compounds found in the particle-phase, Fig. 5 shows the percentages of nitrated aromatics (e.g., nitrophenols and nitrocatechols), phenol, and catechol detected in the particle-phase.

Fig. 5. Contributions of particulate phenols in gray (e.g., C6H6OI, C6H6O2I), nitrophenols in red (e.g., C6H5NO2I, C6H5NO3I, C6H4N2O5I, C7H7NO3I), and nitrocatechols in blue (e.g., C6H5NO4I, C7H7NO4I) to the total budget of observed phenolic compounds in the particle-phase. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

(1150) (366) (683) (1654)

Nitrated compounds represent <90% of phenolic compounds with nitrocatechol and methyl nitrocatechol representing 85% of total phenolic compounds observed in the particle-phase underscoring the importance of residential wood burning as a source of this class of compounds. Notably, as shown in Fig. 6, nitrophenols and nitrocatechols exhibited distinct diurnal profiles with nitrophenols peaking in mass concentration in the afternoon while nitrocatechols had a larger peak at night. This behavior is in contrast to the diurnal profiles observed in Detling where both nitrophenols and nitrocatechols peaked at night due to a common wood burning source (Mohr et al., 2013). The daytime observations of nitrophenols in Port Angeles are likely due to direct emission from automobile exhaust, and daytime nitration of phenol from OH and NO2 radicals (Harrison et al., 2005), which is supported by the increased afternoon concentration of gas-phase phenol observed in Fig. 6. Phenol also peaks at night when residential wood burning occurs. Nitrocatechols are produced from reactions of cresols, emitted during residential wood burning, with OH in the presence of NOx (Iinuma et al., 2010; Kahnt et al., 2013). Gas-phase cresol was not detected by the FIGAERO due to the low sensitivity of this compound using Iodide-adduct chemistry (Lee et al., 2014).

Fig. 6. Diurnal profile of particle-phase nitrophenols (red line with asterisks), nitrocatechols (blue line with squares), and gas-phase phenol (yellow dashed line). Error bars represent one standard deviation. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

C.J. Gaston et al. / Atmospheric Environment 138 (2016) 99e107

3.3. Low volatility organic components in ambient biomass burning particles Continuous, on-line characterization of PM in Port Angeles allowed for the detection of residential wood smoke particle composition during fresh and aged biomass burning conditions. Fig. 7 shows thermograms of the levoglucosan and related tracers ion (C6H10O5I) from four different times of day averaged for the entire campaign: 0:00e5:00 (period 1), 5:00e10:00 (period 2), 10:00e17:00 (period 3), and 17:00e0:00 (period 4). The inset in Fig. 7 shows the normalized mass concentration profile of the levoglucosan tracers and the normalized signal for sulfuric acid during the different times of day. Time period 1 corresponds to a decrease in levoglucosan components as residential heating practices are complete for the evening. Time period 2 is when tracers for residential wood burning increase during the daytime. Time period 3 is when residential wood burning tracers are relatively subdued and the photochemical production of sulfuric acid occurs. Time period 4 is when nighttime residential wood burning tracers peak in the evening. The value of the maximum temperature (Tmax) of the thermogram provides information regarding the volatility of the compound desorbing (or fragmenting) from the particle collection filter. The Tmax has been shown to be correlated to the enthalpy of sublimation of individual compounds (Lopez-Hilfiker et al., 2014). During time periods 1, 2, and 4, the thermograms in Fig. 7 generally show a Tmax between 58 and 70  C, which has previously been observed for a pure levoglucosan standard (Lopez-Hilfiker et al., 2014) (see Figure S4). Also shown in Fig. 7 are the best fits to the observed thermogram data (thick gray lines with dashed line indicating Tmax), where the fit is based on the thermogram profiles of a set of desorptions of standard compounds. For periods 1, 2, and 4, the best fits to the thermograms are a single mode with Tmax~60  C, as expected for pure levoglucosan desorptions. However, thermograms for time period 3 show that the Tmax for the representative ion shifts to near 100  C and the thermogram is best fit to two modes, one occurring at Tmax~60  C and a second mode at

Fig. 7. Average thermograms of C6H10O5 (levoglucosan) binned by four different time periods. Thermograms are averaged from 0:00e5:00 (line with open squares), 5:00e10:00 (dashed line with open triangles), 10:00e17:00 (dashed black line with open circles), and 17:00e0:00 (solid line). Thermogram fits are also show in gray with dashed lines indicating the Tmax of the fits. Inset is the diurnal profile of levoglucosan concentrations (pink line with open circles) and sulfuric acid (yellow line with open circles) with dashed lines to denote the different time bins used to average thermograms. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

105

Tmax~100  C. In contrast, the thermograms for a molecular composition consistent with vanillin, another biomass burning tracer, did not exhibit any similar shifts in Tmax or shape during this period. A comparison of the combined volatility of all biomass burning compounds not including levoglucosan and its related isomers also showed no significant shift in Tmax. The only observed difference in thermograms of other biomass burning tracers was the appearance of an additional mode containing higher volatility components during time period 4 (see Figure S7) likely due to increased contributions from semi-volatile compounds that partitioned to the particle-phase at night. The most probable explanation for the shift in Tmax for levoglucosan and related tracers during period 3 is the formation or preferential persistence in the particle-phase of low volatility oligomers that undergo thermal decomposition to their monomeric units at higher temperature, and are therefore detected at higher values of Tmax than actual monomers of levoglucosan. Thermal decomposition of high molecular weight compounds has also been shown to occur for secondary organic aerosol (SOA) formed from apinene ozonolysis in the FIGAERO (Lopez-Hilfiker et al., 2014). The formation of oligomers from levoglucosan has been shown to occur in the presence of acid catalysts (Holmes and Petrucci, 2006, 2007). Significant shifts in levoglucosan thermograms only occurred during time period 3 when sulfuric acid was also found to desorb from the particle-phase (see Fig. 7). This is likely due to an acid-catalyzed mechanism for the formation of low volatility products from levoglucosan. While the exact cause of the observed shifts in volatility for levoglucosan is not known at this juncture, this work highlights the potential formation of low volatility products during the chemical aging of ambient biomass burning aerosol and the dynamic partitioning of semi-volatile compounds. 4. Conclusions and implications Wintertime PM was characterized in Port Angeles, WA during an intensive study in 2014 that was part of a larger effort to characterize PM and ultrafine particle sources in the region. The diurnal pattern of wood burning tracers, light absorbing organic carbon, and indicators of combustion efficiency (e.g., CO/CO2 ratio) combined with cold, stagnant meteorological conditions and wind directions from residential areas all provide strong evidence of a residential rather than an industrial source for the observed wood smoke PM in Port Angeles, WA during the wintertime. Residential heating practices have been shown to dominate wintertime PM2.5 in several urban and rural locations (Bari et al., 2009; Fine et al., 2002; Gorin et al., 2006; Hedberg et al., 2006; Khalil and Rasmussen, 2003; Larson et al., 2004; Schauer and Cass, 2000; Simpson et al., 2005) leading to non-attainment levels of PM2.5 in some cases, particularly when cold, stagnant conditions trap wood smoke near ground-level (Ogulei, 2010; Ward et al., 2006). Upgrading old woodstoves to newer models has been shown to reduce levels of PM2.5 in Libby, MT (Bergauff et al., 2009) and have likely contributed to improved wintertime air quality in WA state as well. Further upgrades throughout the airshed affecting Port Angeles during southerly flow would likely further reduce levels of wood smoke PM and improve wintertime air quality there. During this intensive study, a novel technique, FIGAERO HRTOF-CIMS, was used to detect and quantify several classes of biomass burning tracers, including levoglucosan. High mass concentrations of levoglucosan and related isomers (up to 19 mg/m3) for an area impacted by residential wood burning were detected. Evidence was also presented for the existence of low volatility aging products, possibly due to acid-catalyzed oligomer formation (Holmes and Petrucci, 2006, 2007). In addition to levoglucosan, guaiacol-containing methoxyphenols and resin acids indicative of

106

C.J. Gaston et al. / Atmospheric Environment 138 (2016) 99e107

softwood wood burning, and nitrocatechols were detected in near real-time. Nitrated phenols are phytotoxic, negatively impact human health, and strongly absorb light (Harrison et al., 2005; Iinuma et al., 2010; Mohr et al., 2013). The detection of high ng/m3 levels of these compounds during this intensive study highlights both the health and climate impacts of wood smoke. A recent review on the health effects of wood smoke by Naeher et al. (2007) indicated the need for novel analytical techniques to characterize more markers in wood smoke and to track their fate to better assess the air quality impacts of wood smoke particles. The results presented herein represent a promising step toward fulfilling this goal through the use of on-line techniques such as FIGAERO HR-TOF-CIMS to characterize wood smoke PM. Acknowledgements We acknowledge the City of Port Angeles Fire Department for assistance provided to take these measurements. Funding was provided by the Washington State Legislature 2013e2015 biennial budget sponsored by Washington State Representative Tharinger and Representative Van De Wege. The funding was directed through the Department of Ecology through contracts G1400165 and UW OSP #A86899. We thank Jonathan Hee for assistance during set-up. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.atmosenv.2016.05.013. References Aiken, A.C., de Foy, B., Wiedinmyer, C., DeCarlo, P.F., Ulbrich, I.M., Wehrli, M.N., Szidat, S., Prevot, A.S.H., Noda, J., Wacker, L., Volkamer, R., Fortner, E., Wang, J., Laskin, A., Shutthanandan, V., Zheng, J., Zhang, R., Paredes-Miranda, G., Arnott, W.P., Molina, L.T., Sosa, G., Querol, X., Jimenez, J.L., 2010. Mexico city aerosol analysis during MILAGRO using high resolution aerosol mass spectrometry at the urban supersite (T0) e Part 2: analysis of the biomass burning contribution and the non-fossil carbon fraction. Atmos. Chem. Phys. 10 (12), 5315e5341. Andreae, M.O., Merlet, P., 2001. Emission of trace gases and aerosols from biomass burning. Glob. Biogeochem. Cy. 15 (4), 955e966. Bari, M.A., Baumbach, G., Kuch, B., Scheffknecht, G., 2009. Wood smoke as a source of particle-phase organic compounds in residential areas. Atmos. Environ. 43 (31), 4722e4732. Bergauff, M.A., Ward, T.J., Noonan, C.W., Palmer, C.P., 2009. The effect of a woodstove changeout on ambient levels of PM2.5 and chemical tracers for woodsmoke in Libby, Montana. Atmos. Environ. 43 (18), 2938e2943. Bolling, A.K., Pagels, J., Yttri, K.E., Barregard, L., Sallsten, G., Schwarze, P.E., Boman, C., 2009. Health effects of residential wood smoke particles: the importance of combustion conditions and physicochemical particle properties. Part. Fibre Toxicol. 6 (29) http://dx.doi.org/10.1186/1743-8977-6-29. Carrico, C.M., Petters, M.D., Kreidenweis, S.M., Sullivan, A.P., McMeeking, G.R., Levin, E.J.T., Engling, G., Malm, W.C., Collett, J.L., 2010. Water uptake and chemical composition of fresh aerosols generated in open burning of biomass. Atmos. Chem. Phys. 10 (11), 5165e5178. Christian, T.J., Yokelson, R.J., Cardenas, B., Molina, L.T., Engling, G., Hsu, S.-C., 2010. Trace gas and particle emissions from domestic and industrial biofuel use and garbage burning in central Mexico. Atmos. Chem. Phys. 10, 565e584. Dhammapala, R., Claiborn, C., Jimenez, J., Corkill, J., Gullett, B., Simpson, C., Paulsen, M., 2006. Emission factors of PAHs, methoxyphenols, levoglucosan, elemental carbon and organic carbon from simulated wheat and Kentucky bluegrass stubble burns. Atmos. Environ. 41, 2660e2669. Edye, L.A., Richards, G.N., 1991. Analysis of condensates from wood smoke e components derived from polysaccharides and lignins. Environ. Sci. Tech. 25 (6), 1133e1137. Elias, V.O., Simoneit, B.R.T., Perreira, A.S., Cabral, J.A., Cardoso, J.N., 1999. Detection of high molecular weight organic tracers in vegetation smoke samples by hightemperature gas chromatography-mass spectrometry. Environ. Sci. Tech. 33, 2369e2376. Engling, G., Carrico, C.M., Kriedenweis, S.M., Collett, J.L., Day, D.E., Malm, W.C., Lincoln, E., Hao, W.M., Iinuma, Y., Herrmann, H., 2006. Determination of levoglucosan in biomass combustion aerosol by high-performance anion-exchange chromatography with pulsed amperometric detection. Atmos. Environ. 40, S299eS311.

Fine, P.M., Cass, G.R., Simoneit, B.R.T., 2002. Organic compounds in biomass smoke from residential wood combustion: emissions characterization at a continental scale. J. Geophys. Res. Atmos. 107 (D21). Fine, P.M., Cass, G.R., Simoneit, B.R.T., 2004. Chemical characterization of fine particle emissions from the wood stove combustion of prevalent United States tree species. Environ. Eng. Sci. 21 (6), 705e721. Fraser, M.P., Lakshmanan, K., 2000. Using levoglucosan as a molecular marker for the long-range transport of biomass combustion aerosols. Environ. Sci. Tech. 34 (21), 4560e4564. Gorin, C.A., Collett, J.L., Herckes, P., 2006. Wood smoke contribution to winter aerosol in Fresno, CA. J. Air Waste Manag. Assoc. 56 (11), 1584e1590. Hand, J.L., Schichtel, B.A., Pitchford, M., Malm, W.C., Frank, N.H., 2012. Seasonal composition of remote and urban fine particulate matter in the United States. J. Geophys. Res. Atmos. 117. Harrison, M.A.J., Barra, S., Borghesi, D., Vione, D., Arsene, C., Olariu, R.L., 2005. Nitrated phenols in the atmosphere: a review. Atmos. Environ. 39 (2), 231e248. Hawthorne, S.B., Miller, D.J., Langenfeld, J.J., Krieger, M.S., 1992. PM-10 high-volume collection and quantitation of semivolatile and nonvolatile phenols, methoxylated phenols, alkanes, and polycyclic aromatic-hydrocarbons from winter urban air and their relationship to wood smoke emissions. Environ. Sci. Tech. 26 (11), 2251e2262. Hays, M.D., Geron, C., Linna, K.J., Smith, N.D., Schauer, J.J., 2002. Speciation of gasphase and fine particle emissions from burning of foliar fuels. Environ. Sci. Tech. 36, 2281e2295. Hedberg, E., Johansson, C., Johansson, L., Swietlicki, E., Brorstrom-Lunden, E., 2006. Is levoglucosan a suitable quantitative tracer for wood burning? Comparison with receptor modeling on trace elements in Lycksele, Sweden. J. Air Waste Manag. Assoc. 56, 1669e1678. Hennigan, C.J., Sullivan, A.P., Collett, J.L., Robinson, A.L., 2010. Levoglucosan stability in biomass burning particles exposed to hydroxyl radicals. Geophys. Res. Lett. 37. Hoffmann, D., Tilgner, A., Iinuma, Y., Herrmann, H., 2010. Atmospheric stability of levoglucosan: a detailed laboratory and modeling study. Environ. Sci. Tech. 44 (2), 694e699. Holmes, B.J., Petrucci, G.A., 2006. Water-soluble oligomer formation from acidcatalyzed reactions of levoglucosan in proxies of atmospheric aqueous aerosols. Environ. Sci. Tech. 40 (16), 4983e4989. Holmes, B.J., Petrucci, G.A., 2007. Oligomerization of levoglucosan by Fenton chemistry in proxies of biomass burning aerosols. J. Atmos. Chem. 58, 151e166. Iinuma, Y., Boge, O., Grafe, R., Herrmann, H., 2010. Methyl-nitrocatechols: atmospheric tracer compounds for biomass burning secondary organic aerosols. Environ. Sci. Tech. 44 (22), 8453e8459. Kahnt, A., Behrouzi, S., Vermeylen, R., Shalamzari, M.S., Vercauteren, J., Roekens, E., Claeys, M., Maenhaut, W., 2013. One-year study of nitro-organic compounds and their relation to wood burning in PM10 aerosol from a rural site in Belgium. Atmos. Environ. 81, 561e568. Khalil, M.A.K., Rasmussen, R.A., 2003. Tracers of wood smoke. Atmos. Environ. 37, 1211e1222. Larson, T., Gould, T., Simpson, C., Liu, L.J.S., Claiborn, C., Lewtas, J., 2004. Source apportionment of indoor, outdoor, and personal PM2.5 in Seattle, Washington, using positive matrix factorization. J. Air Waste Manag. Assoc. 54 (9), 1175e1187. Laskin, A., Smith, J.S., Laskin, J., 2009. Molecular characterization of nitrogencontaining organic compounds in biomass burning aerosols using highresolution mass spectrometry. Environ. Sci. Tech. 43 (10), 3764e3771. Lee, B.H., Lopez-Hilfiker, F.D., Mohr, C., Kurten, T.C., Worsnop, D., Thornton, J.A., 2014. An iodide-adduct high-resolution time-of-flight chemical-ionization mass spectrometer: application to atmospheric inorganic and organic compounds. Environ. Sci. Tech. 48, 6309e6317. Lopez-Hilfiker, F.D., Mohr, C., Ehn, M., Rubach, F., Kleist, E., Wildt, J., Mentel, T.F., Lutz, A., Hallquist, M., Worsnop, D., Thornton, J.A., 2014. A novel method for online analysis of gas and particle composition: description and evaluation of a Filter Inlet for Gases and AEROsols (FIGAERO). Atmos. Meas. Tech. 7 (4), 983e1001. Ma, Y., Hays, M.D., 2008. Thermal extraction-two-dimensional gas chromatography-mass spectrometry with heart-cutting for nitrogen heterocyclics in biomass burning aerosols. J. Chromatogr. A 1200, 228e234. Mazzoleni, L.R., Zielinska, B., Moosmuller, H., 2007. Emissions of levoglucosan, methoxy phenols, and organic acids from prescribed burns, laboratory combustion of wildland fuels, and residential wood combustion. Environ. Sci. Tech. 41 (7), 2115e2122. McDonald, J.D., Zielinska, B., Fujita, E.M., Sagebiel, J.C., Chow, J.C., Watson, J.G., 2000. Fine particle and gaseous emission rates from residential wood combustion. Environ. Sci. Tech. 34 (11), 2080e2091. Mohr, C., Lopez-Hilfiker, F.D., Zotter, P., Prevot, A.S.H., Xu, L., Ng, N.L., Herndon, S.C., Williams, L.R., Franklin, J.P., Zahniser, M.S., Worsnop, D.R., Knighton, W.B., Aiken, A.C., Gorkowski, K.J., Dubey, M.K., Allan, J.D., Thornton, J.A., 2013. Contribution of nitrated phenols to wood burning brown carbon light absorption in Detling, United Kingdom during winter time. Environ. Sci. Tech. 47, 6316e6324. Naeher, L.P., Brauer, M., Lipsett, M., Zelikoff, J.T., Simpson, C.D., Koenig, J.Q., Smith, K.R., 2007. Woodsmoke health effects: a review. Inhal. Toxicol. 19 (1), 67e106. Nolte, C.G., Schauer, J.J., Cass, G.R., Simoneit, B.R.T., 2001. Highly polar organic compounds present in wood smoke and in the ambient atmosphere. Environ.

C.J. Gaston et al. / Atmospheric Environment 138 (2016) 99e107 Sci. Tech. 35, 1912e1919. Ogulei, D., 2010. Sources of Fine Particles in the Wapato Hills-Puyallup River Valley PM2.5 Non-attainment Area, in Air Qualty Program, Was, Edited by W.S.D.o.E. Air Quality Program, pp. Publication No. 10-02-009. Oros, D.R., Simoneit, B.R.T., 2001. Identification and emission factors of molecular tracers in organic aerosols from biomass burning Part 1. Temperate climate conifers. Appl. Geochem. 16 (13), 1513e1544. Petters, M.D., Carrico, C.M., Kreidenweis, S.M., Prenni, A.J., DeMott, P.J., Collett, J.L., Moosmuller, H., 2009. Cloud condensation nucleation activity of biomass burning aerosol. J. Geophys. Res. Atmos. 114. Ramanathan, V., Carmichael, G., 2008. Global and regional climate changes due to black carbon. Nat. Geosci. 1 (4), 221e227. Rogge, W.F., Hildemann, L.M., Mazurek, M.A., Cass, G.R., Simoneit, B.R.T., 1998. Sources of fine organic aerosol. 9. Pine, oak and synthetic log combustion in residential fireplaces. Environ. Sci. Tech. 32 (1), 13e22. Saleh, R., Robinson, E.S., Tkacik, D.S., Ahern, A.T., Liu, S., Aiken, A.C., Sullivan, R.C., Presto, A.A., Dubey, M.K., Yokelson, R.J., Donahue, N.M., Robinson, A.L., 2014. Brownness of organics in aerosols from biomass burning linked to their black carbon content. Nat. Geosci. 7 (9), 647e650. Schauer, J.J., Cass, G.R., 2000. Source apportionment of wintertime gas-phase and particle-phase air pollutants using organic compounds as tracers. Environ. Sci. Tech. 34, 1821e1832. Schauer, J.J., Kleeman, M.J., Cass, G.R., Simoneit, B.R.T., 2001. Measurements of emissions from air pollution sources. 3. C1-C29 organic compounds from fireplace combustion of wood. Environ. Sci. Tech. 35, 1716e1728. Simoneit, B.R.T., Rogge, W.F., Mazurek, M.A., Standley, L.J., Hildemann, L.M., Cass, G.R., 1993. Lignin pyrolysis products, lignans, and resin acids as specific tracers of plant classes in emissions from biomass combustion. Environ. Sci. Tech. 27 (12), 2533e2541. Simoneit, B.R.T., Schauer, J.J., Nolte, C.G., Oros, D.R., Elias, V.O., Fraser, M.P., Rogge, W.F., Cass, G.R., 1999. Levoglucosan, a tracer for cellulose in biomass burning and atmospheric particles. Atmos. Environ. 33, 173e182.

107

Simoneit, B.R.T., Rushdi, A.I., Bin Abas, M.R., Didyk, B.M., 2003. Alkyl amides and nitriles as novel tracers for biomass burning. Environ. Sci. Tech. 37, 16e21. Simoneit, B.R.T., 2002. Biomass burning e a review of organic tracers for smoke from incomplete combustion. Appl. Geochem. 17 (3), 129e162. Simpson, C.D., Paulsen, M., Dills, R.L., Liu, L.J.S., Kalman, D.A., 2005. Determination of methoxyphenols in ambient atmospheric particulate matter: tracers for wood combustion. Environ. Sci. Tech. 39 (2), 631e637. Smith, J.S., Laskin, A., Laskin, J., 2009. Molecular Characterization of biomass burning aerosols using high-resolution mass spectrometry. Anal. Chem. 81 (4), 1512e1521. Sullivan, A.P., May, A.A., Lee, T., McMeeking, G.R., Kreidenweis, S.M., Agaki, S.K., Yokelson, R.J., Urbanski, S.P., Collett, J.L., 2014. Airborne characterization of smoke marker ratios from prescribed burning. Atmos. Chem. Phys. 14, 10535e10545. Sun, H.L., Biedermann, L., Bond, T.C., 2007. Color of brown carbon: a model for ultraviolet and visible light absorption by organic carbon aerosol. Geophys. Res. Lett. 34 (17), L17813. http://dx.doi.org/10.1029/2007GL029797. Szidat, S., Prevot, A.S.H., Sandradewi, J., Alfarra, M.R., Synal, H.-A., Wacker, L., Baltensperger, U., 2007. Dominant impact of residential wood burning on particulate matter in Alpine valleys during winter. Geophys. Res. Lett. 34, L05820. http://dx.doi.org/10.1029/2006GL028325. Ward, T.J., Rinehart, L.R., Lange, T., 2006. The 2003/2004 Libby, Montana PM2.5 source apportionment research study. Aerosol Sci. Tech. 40 (3), 166e177. Weimer, S., Alfarra, M.R., Schreiber, D., Mohr, M., Prevot, A.S.H., Baltensperger, U., 2008. Organic aerosol mass spectral signatures from wood-burning emissions: influence of burning conditions and wood type. J. Geophys Res. 113, D10304. http://dx.doi.org/10.1029/2007JD009309. Weimer, S., Mohr, C., Richter, R., Keller, J., Mohr, M., Prevot, A.S.H., Baltensperger, U., 2009. Mobile measurements of aerosol number and volume size distributions in an Alpine valley: influence of traffic versus wood burning. Atmos. Environ. 43 (3), 624e630.