j ourna l homepage: www.e lsev ie r .com/ locate /atmos
Speciation of organic compounds in aerosols from urbanbackground sites in the winter season
Célia Alves⁎, Teresa Nunes, Ana Vicente, Cátia Gonçalves, Margarita Evtyugina, Telma Marques,Casimiro Pio, Frieda Bate-EpeyCentre for Environmental and Marine Studies, Department of Environment, University of Aveiro, 3810-193 Aveiro, Portugal
Article history:Received 24 March 2014Received in revised form 9 July 2014Accepted 11 July 2014Available online 18 July 2014
Winter aerosol samples were daily collected during one-month long campaign in Oporto andCoimbra. The high-volume PM2.5 samples were solvent extracted and their organic contentseparated into several functional groups, which were then analysed by gas chromatography–mass spectrometry. The organic compounds identified and quantified revealed somedifferences between samples from the two urban areas. In general, the levels of totalhydrocarbons in the urban background station of Oporto were higher than those of Coimbra.Concentration ratios between specific compounds and the presence of molecular markersderived from petroleum, such as hopanes, pristane and phytane, point out vehicles as the mainsource of pollutants. The contribution of biogenic compounds, mainly hydrocarbons associatedwith the waxy cuticle of vegetation, is also observable in both cities. The benzo[a]pyreneequivalent daily values were frequently higher than 1 ng m−3 in Oporto suggesting anadditional cancer risk for the population. The PM2.5 mass attributable to vehicle emissions ishigher in the background atmosphere of Oporto than in Coimbra. On weekends, biomassburning emissions could represent up to 74% of the organic carbon content of the urbanaerosols.
It has long been recognised that atmospheric aerosolsinteract both directly and indirectly with the Earth'sradiation budget and climate (Carslaw et al., 2010) andmay have detrimental effects on human health, such asimpairment of pulmonary function (Islam et al., 2007; Rivaet al., 2011). The organic content of aerosols accounts for asubstantial fraction of the global aerosol burden (Zhang etal., 2011).
More than half of the world's population live in townsand cities. Thus, most of the aerosol characterisation andrisk assessment studies have been carried out in urban
areas (e.g. Bi et al., 2008; Herlekar et al., 2012;Schnelle-Kreis et al., 2005; Tsapakis et al., 2002; Wanget al., 2006). Among the target organic compounds, greatefforts have been made to speciate, quantify, and under-stand the sources and reactivity of polycyclic aromaticcompounds (PAHs), particularly due to their carcinogenicpotential (Chen et al., 2011; Krumal et al., 2013;Mazquiarán and Pinedo, 2007; Wang et al., 2009a; Wei etal., 2012; Xie et al., 2009).
In Portugal, only a few reports have dealt with theorganic composition of aerosols in urban environments(Alves et al., 2001; Oliveira et al., 2007). Comprehensiveinformation on the chemical composition of atmosphericaerosols is useful to improve climate models, to proposeemission abatement strategies and to estimate publichealth impacts. Aiming at better understanding the organic
58 C. Alves et al. / Atmospheric Research 150 (2014) 57–68
composition and sources of urban aerosols, a monitoringcampaign was carried out in two cities: Oporto andCoimbra.
2. Methodology
2.1. Sampling
Fine (PM2.5) and coarse (PM2.5–10) aerosol samples weredaily collected onto pre-fired quartz fibre filters with highvolume samplers operating at 1.13 m3 min−1. Samplingwas carried out during one-month long campaign, simul-taneously in Oporto and Coimbra, between January 27 andFebruary 27, 2007. Urban background sites of the air qualitynetwork were, respectively, selected: Ermesinde (41°12.40′N;8°33.17′W) and Geophysical Institute of Coimbra (10°13.33′N;8°24.65′W). The urban area of Oporto, which extends beyondthe administrative limits of the city, has a population of about1.3 million in an area of 389 km2, making it the second-largesturban area in Portugal, after Lisbon. Coimbra is the third biggestand most important city in the country with an estimatedresident population of 150,000. It plays a chief role in thenorthern-central littoral and interior of Portugal.
2.2. Analytical determinations
After gravimetric determination of particle concentration,small punches of the filterswere analysed by a thermal–opticaltransmission technique to obtain the organic carbon (OC)content (Pio et al., 2011). The PM2.5 samples were extracted for24 h by refluxing dichloromethane and the nitrogen-driedextract was separated into 5 different organic fractions by flashchromatography with silica gel and various eluents of increas-ing polarity. The detailed description of the methodology forthe extraction of organic compounds could be found in Gogouet al. (1998) and Alves et al. (2001). The fractionated extractswere analysed by gas chromatography–mass spectrometry(GC–MS). Before injection, the compounds with hydroxylicand carboxylic groups were converted into the correspondingTMS ether or TMS ester derivatives, respectively, by addition ofa mixture of N,O-bis(trimethylsilyl)triflouroacetamide andtrimethylchlorosilane (BSTFA/TMCS; 99:1), followed by 3 h inan oven at 70 °C. Extracts were injected within 24 h afterthe derivatisation procedure. Two quadrupole GC–MS wereused, a HP 6890 MSD 5973 and a GC Trace Ultra, DSQ II fromThermo Scientific. Both instruments were operated withTRB-5MS 60 m × 0.25 mm × 0.25 μm columns. Helium wasused as carrier gas at a constant flow of 1.2 mL min−1. Theheating programme was as follows: 60 °C (1 min); 60–150 °C(10 °C min−1); 150–290 °C (5 °C min−1); and290 °C (27 min).The acquisition mode was electronic impact at 70 eV andthe scanned masses ranged from 50 to 850 m/z. Calibrationfor GC–MS analysis was based on a total of more than 200standards in five different concentration levels with relativeresponse factors determined individually for the majority ofcompounds. For those compounds with no authentic standardsavailable, relative response factors were obtained as an averagefrom the overall homologous series or from compounds ofsimilar chemical structure and retention time. Standards andsamples were both co-injected with two internal standards:tetracosane-D50 and1-chlorohexadecane or 1-chlorododecane.
In the case of PAHs, an internal standard mix was used:1,4-dichlorobenzene-D4, naphthalene-D8, acenaphthene-D10,phenanthrene-D10, chrysene-D12 and perylene-D12. Com-pound identification was based on comparison of resultingspectra with mass spectra libraries (Wiley and NIST),co-injection with authentic standards and analysis offragmentation patterns. Quantification was performed bysingle ion monitoring and total ion chromatogram analysis.
3. Results and discussion
3.1. Meteorological conditions
Meteorological data from Oporto and Coimbra wereobtained for the study period from the nearest airportweather stations, both located less than 10 km from thecities: Pedras Rubras and Cernache, respectively. In Coimbra,the temperatures varied from a minimum value of 1.6 to amaximum of 17 °C. Values ranging from −1 to 18 °C wereregistered in Oporto (Fig. 1). During the sampling campaign,the averages of the daily mean temperatures were similar forboth cities: 11.4 °C (Oporto) and 10.7 °C (Coimbra).
Lower temperatures and almost no rainfall were regis-tered in the first 11 days of the campaign. Both sites wereunder the influence of continental air masses, most of whichoriginating and staying over Spain and Portugal, or maritimetransformed air masses, when backward trajectories indicat-ed an Atlantic origin, with a final re-circulation through theIberian Peninsula (Supplementary Fig. A1 and Fig. 2). Afterthis period, the regime was generally characterised bylow-pressure systems centred to the west of the BritishIsles. This feature led to the prevalence of the westerlies andfrontal systems over the country, which transported mari-time air masses and established rain-generating conditions,milder temperatures and higher wind speeds.
3.2. Particulate matter concentrations and organic constituents
The daily mean (±standard deviation) PM2.5 concentra-tions during the sampling campaign were 29.9 ± 25.0 and19.3 ± 13.3 μg m−3, in Oporto and Coimbra, respectively.PM2.5 represented 66.7 ± 17.8% of the PM10 mass in Oporto,whilst the percentage was 70.6 ± 16.1% in Coimbra. On somedays, in Oporto, the daily mean PM10 concentrations exceededthe limit value of 50 μg m−3 set by the Air Quality Directive2008/50/CE. Organic carbon accounted for 32.5 ± 14.6% and33.9 ± 18.5% of the PM2.5 concentrations measured in Oportoand Coimbra, respectively. In general, the lowest concentra-tions of both PM2.5 and OC were associated with rainy eventsand air masses from the Atlantic, whilst stagnant conditions,regional atmospheric circulation and/or continental air massescoincided with the most polluted days. The chromatographi-cally resolved organic compounds encompassed aliphatic andaromatic hydrocarbons, carbonyls, methyl esters of carboxylicacids, alcohols, anhydrosugars and several types of acids.
3.2.1. Aliphatic compoundsThe aliphatic fraction of particulate matter comprised
n-alkanes (Fig. 3), n-alkenes, hopanes, the unresolved complexmixture (UCM) of cyclic branched and unsaturated hydrocar-bons and acyclic isoprenoids (pristane and phytane). Levels
Fig. 1. Evolution of the temperature, precipitation and wind speed during the sampling campaign.
59C. Alves et al. / Atmospheric Research 150 (2014) 57–68
of total n-alkanes in the urban background station of Oporto(34.0 ± 31.1 ng m−3) were higher than those of Coimbra(16.5 ± 14.4 ng m−3). The contribution of biogenic hydrocar-bons associated with the waxy cuticle of vegetation wasobservable in both cities. The wax n-alkanes represented, onaverage, 19 and 25% of the total concentrations in Oporto andCoimbra, respectively. Wax percentages from 14 to 25 havebeen described as typical for urban areas (Mazquiarán andPinedo, 2007).
Fig. 2. Clustering (180 h) of backward trajectories arriving at Oporto from 3 to 6 Fethe NOAA HYSPLIT Lagrangian model (Draxler and Rolph, 2013). The first and secolevels recorded for both sampling sites.
The highest n-alkane concentrations were observed forthe homologues C25, C27, C29 and C31. Whilst the C25homologue has been associated with fossil fuel emissions,lubricating oils and unburnt heating oil, the other carbonnumbers derive from epicuticular waxes from higher plantswaxes, which can be released into the atmosphere uponbiomass burning (Alves et al., 2012). Cooking may alsorepresent a significant source of n-alkane emissions in theC14–C33 range (Zhao et al., 2007). The carbon preference
bruary (left panel) and from 7 to 11 February (right panel) obtained throughnd periods correspond, respectively, to the highest and lowest PM2.5 and OC
Fig. 3. Concentrations of n-alkanes (C14–C34), acyclic isoprenoid hydrocarbons (Pr — pristine, Ph — phytane) and unresolved complex mixture (UCM) in PM2.5
from Oporto and Coimbra.
60 C. Alves et al. / Atmospheric Research 150 (2014) 57–68
index (CPI) is a diagnostic tool that identifies contributionsarising from biogenic and anthropogenic inputs (Alves, 2008;Yadav et al., 2013). The n-alkane CPI values obtained for bothcities are indicative of a petrogenic signature (Table 1). Muchhigher CPI values are generally obtained in rural and forestenvironments (Alves et al., 2012; Oliveira et al., 2007; Yadavet al., 2013). The petrogenic input is confirmed by the highproportion between the chromatographically unresolved andresolved compounds (UCM/R). Pure hydrocarbon mixtures
Table 1Diagnostic parameters applied to concentrations of organic compounds.
CPIalkanes = Σ(C13–C35) / Σ(C12–C34); CPIalkanols = Σ(C12–C34) / Σ(C13–C35);Wax Cn = |[Cn] − [(Cn + 1 + Cn − 1) / 2)]|; homohopane index =C31αβ[S / (S + R)]; bishomohopane index = C32αβ[S / (S + R)]; CPAHs/TPAHs = concentrations of nine major nonalkylated compounds to the totalconcentration of PAHs; BFs — sum of concentrations of benzofluoranthenes;BaPE = BaA × 0.06 + BFs × 0.07 + BaP + DBahA × 0.6 + IcdP × 0.08. SeeFigs. 4 and 5 for other acronyms.
from plant waxes have UCM/R b 0.1, i.e. they have no UCM(Alves, 2008).
The hopanoid hydrocarbon series found in aerosols (Fig. 4)is derived from precursors in the cell membranes of prokary-otes (bacterial source) and cyanobacteria (blue-green algaesource) in sedimentary organic matter over geological time(Krumal et al., 2013). C30 hopanoids are also known to bepresent in certain higher plants (e.g. fern and moss). Accom-panying the geological evolution, plant remains undergophysical, chemical and biochemical transformations yielding aseries of coals of increasing rank of maturity (Simoneit et al.,2007). In this study, the total hopane concentrations in PM2.5
from Oporto and Coimbra were 6.17 ± 3.35 and 2.61 ±1.73 ng m−3, respectively. These concentrations are of thesame order as those determined in PM1 aerosols in urban areasof the Czech Republic (Krumal et al., 2013). Concentrations 10times higher (65 ± 24 ng m−3) have been reported forwintertime samples collected in the urban atmosphere ofBaoji, inland China (Wang et al., 2009b). Depending on the site,Herlekar et al. (2012) obtained hopane concentrations from7.81 ± 2.00 to 94.1 ± 28.7 ng m−3 in Mumbai, India.
The homohopane and bishomohopane indexes obtain-ed in both Portuguese cities are in the ranges described forvehicular emissions, whilst the C29αβ/C30αβ is within thevalues suggested for coal combustion. Rogge et al. (1993)obtained indexes from 0.5 to 0.6 for gasoline and dieselemissions. Oros and Simoneit (2000) reported C29αβ/C30αβ, C31αβ[S/(S + R)] and C32αβ[S/(S + R)] ratios,respectively, in the ranges of 0.6–2.0, 0.1–0.4 and 0.2–0.4for coal combustion. The Ts/Tm ratio is an additionalindicator for the type of hopane emission sources, becauseduring catagenesis 18α(H)-22,29,30-trisnorneohopane(Ts) is more stable than 17α(H)-22,29,30-trisnorhopane(Tm) (Schnelle-Kreis et al., 2005). The ratio decreaseswhen the impact of less mature fossil fuel (e.g. coal)increases or when instantaneous thermal maturation ofbiogenic hopanoids by biomass burning occurs. Ts/Tm
Fig. 4. Concentrations of hopanes in Oporto and Coimbra. 18α(H)-22,29,30-trisnorneohopane (Ts); 17α(H)-22,29,30-trisnorhopane (Tm); 17α(H),21β(H)-29-norhopane (29αβ); 17β(H),21α(H)-30-norhopane (29βα);17α(H),21β(H)-hopane (30αβ); 17β(H),21α(H)-hopane(30βα); 17α(H),21β(H)-22S or R-homohopane (31αβ-S/R); 17α(H),21β(H)-22S orR-bishomohopane (32αβ-S/R).
61C. Alves et al. / Atmospheric Research 150 (2014) 57–68
values from 0.3 to 0.7 have been found in aerosolsimpacted by biomass burning (Fang et al., 1999). Song etal. (2005) obtained Ts/Tm ratios of 1.33 and 1.09 for theparticulate matter accumulated on the ceiling surfaces ofvehicular tunnels in Hong Kong and Guangzhou, respec-tively. An average ratio of 1.07 has been reported for PM2.5
samples collected in a roadway tunnel in Marseille, France(Haddad et al., 2009). Thus, the Ts/Tm ratios around 0.9obtained in the Portuguese cities indicate a mixed imprintof emissions from the combustion of both fossil and lessthermally mature fuels.
Pristane (2,6,10,14-tetramethylpentadecane) and phy-tane (2,6,10,14-tetramethylhexadecane) were only detectedin some samples from Oporto at an average Pr/Ph ratio of0.69. Since it is highly infrequent to find phytane in biologicalmaterials (excepting some bacteria), most biogenic samplespresent Pr/Ph much higher than 1 (Alves, 2008). Low ratiosare indicative of a petrogenic hydrocarbon signature.
3.2.2. Polyaromatic hydrocarbonsPolycyclic aromatic hydrocarbons are a class of com-
pounds that are considered ubiquitous. They are formedduring the incomplete combustion and pyrolysis of organicmatter, such as coal, wood, fossil fuels, and cooking oil (Chenet al., 2007; Fine et al., 2004; Gonçalves et al., 2011; Rogge etal., 1993; Simoneit et al., 2007). PAHs were found atconcentrations of 17.7 ± 17.6 ng m−3 in Oporto, whilstlower levels of 7.3 ± 7.4 ng m−3 were obtained in Coimbra(Fig. 5). In contrast with other PAH species, retene, a trace ofwood combustion (Gonçalves et al., 2011), was found atsimilar concentrations in both cities. Retene can also bereleased into the atmosphere during cooking (Jørgensen etal., 2013), but it accounts for different PM2.5 mass fractions,depending on the source. The benzo[a]pyrene equivalentconcentration (BaPE) has been introduced instead of the solebenzo[a]pyrene, a classical carcinogen and one of the mostpowerful mutagens, since the latter is easily decomposed inreactive air. It tries to parameterise the health risk forhumans related to ambient PAH exposure and is calculatedby adding weighted concentrations of each carcinogenic
congener (Alves et al., 2012). Among the PAHs measured, themass percentage of carcinogenic species was 63–64% in bothcities. However, the average values of BaPE indicate a highercancer risk in Oporto than in Coimbra. In Oporto, on mostdays, the BaPE daily values exceeded the maximum permis-sible risk level of 1 ng m−3 set by the World HealthOrganisation (Fig. 6). The BaPE concentrations obtained inOporto and Coimbra are within the values reported forBarcelona and Zurich (Alves et al., 2012) and much lowerthan those obtained for Chinese cities (Chen et al., 2011;Wang et al., 2009a, 2009b; Xie et al., 2009).
Various diagnostic concentration ratios of PAHs have beenused to differentiate between petrogenic and pyrogenicsources (Table 1). The CPAHs/TPAHs, Pyr/BaP and BghiP/BaPratios are consistent with the ranges of PAH emitted byvehicles, especially diesel-powered, whilst BFs/BghiP is closerto the values reported for wood burning (Alves, 2008; Alveset al., 2012; Tsapakis et al., 2002). A Fl/(Fl + Py) ratio around0.51 is used as a tracer of domestic heating gas oil (Yunker etal., 2002). The IcdP/(IcdP + BghiP) and BaA/(BaA + Chry)ratios reveal mixed influences of wood burning and vehicleemissions (Alves, 2008).
3.2.3. Carbonyl compoundsIn both cities, n-alkanones from C13 to C31, with mixed
origins, were present in the aerosol samples at similarconcentrations (~0.5 ng m−3). Discontinuous series ofn-alkanals, encompassing some of the homologues in therange of C10–C29, were also observed. An even carbonnumber predominance for the homologues N C20 wasgenerally registered in Oporto. The homologue with highestmean concentrations was tetradecanal (1.1 ng m−3).The higher molecular weight homologues were not detect-ed in most samples of Coimbra. The major carbonylcompound present in all aerosol samples was 6,10,14-trimethylpentadecanon-2-one, an oxidation product offarnesol and phytol in chlorophyll, also known as phytone(Rontani et al., 1996). It was found at average concentra-tions of 3.4 and 4.1 ng m−3 in Oporto and Coimbra,respectively. In PM2.5 samples from Augsburg, Germany,Schnelle-Kreis et al. (2005) detected maximum levels(101 ng m−3) in July, whilst a 10-fold decrease wasobserved in wintertime. Thus, phytone was classified as agood marker for secondary biogenic aerosol. However,Alves et al. (2012) have detected this carbonyl compoundat appreciable amounts in the winter aerosol from Zurich,pointing out biomass burning as another plausible source.The pyrogenic origin is corroborated by the detection of theisoprenoid ketone (225–2176 ng m−3) in the smokeplumes of prescribed fires (Alves et al., 2010).
A multifunctional carbonyl present in some samples, atconcentrations up to 200 pg m−3, was 3,5-di-tert-butyl-4-hydroxybenzaldehyde (BHT-aldehyde). It has been previ-ously detected in hourly in situ measurements made atChebogue Point, Nova Scotia (Williams et al., 2007) and inPM2.5 samples collected in Birmingham, UK (Alam et al.,2013). An origin in emissions from food industries wherethis chemical is used as food additive has been pointed out(Alam et al., 2013). However, an origin in emissions frommotor vehicles is also plausible since many multifunctional
62 C. Alves et al. / Atmospheric Research 150 (2014) 57–68
carbonyls have been identified in PM2.5 collected in ahighway tunnel (Rao et al., 2001).
Polycyclic aromatic ketones and aldehydes, such as 7H-benzo[c]fluoren-7-one, benzo[a]anthracene-7,12-dione, 9H-fluoren-9-one, hydroxy-9H-fluoren-9-one, 9,10-anthraquinone,phenanthrene-carboxaldehyde, 9-anthracenecarboxaldehydeand pyrenecarboxaldehyde, constituted oxy-PAHs present inalmost all samples at individual levels in the order of tens orhundreds of pg m−3.
α-Hexylcinnamaldehyde was found in many samples(from undetectable to 400 pg m−3). It is a flagrance used invarious consumer products and occurs naturally in the bark
Fig. 6. Percentages of carcinogenic PAHs and be
of some species (Gursale et al., 2010). It has also beendetected in particle emissions from fireplace combustion ofdifferent wood species (Fine et al., 2001).
Nopinone and pinonaldehyde are secondary carbonylaerosol products, formed from the ozonolysis of pinenes(Jaoui and Kamens, 2003). These pinenic products werepresent in PM2.5 at concentrations up to 0.35 and 9.2 ng m−3,respectively. In spite of their high vapour pressures, theseconstituents may partition into organic layers condensedonto pre-existing particles. Moreover, acid-catalysed poly-merisation reactions can produce secondary organic aerosol(SOA) products that condense onto particles even though
nzo[a]pyrene equivalent concentrations.
63C. Alves et al. / Atmospheric Research 150 (2014) 57–68
vapour pressure of the original compound is itself too highfor partitioning (Kallio et al., 2006).
3.2.4. Hydroxyl compoundsThe homologous series of n-alkanols (C10–C32) accounted
for concentrations in the ranges of 3–736 ng m−3 in Oportoand 21–630 ng m−3 in Coimbra. Around 70% of the levels ofthe higher molecular weight n-alkanols derived from vege-tation waxes, whilst the lower homologues originated mainlyfrom anthropogenic sources with petrogenic origin (Alves etal., 2012). The n-alkanol levels followed the same temporalpatterns as those observed for PM2.5. Good (r = 0.92) orreasonable (r = 0.65) correlations were observed betweenboth variables in Oporto and Coimbra, respectively. Ingeneral, the CPI values decreased with increasing PM2.5
levels. This suggests that PM2.5 spikes were associated withanthropogenic pollution sources.
Some sterol compounds were present in the PM2.5
samples. Among them, phytosterols, such as sitosterol andstigmasterol, are plant-derived compounds, which are intro-duced into smoke by volatilisation during burning(Gonçalves et al., 2011). Sitosterol was detected at concen-trations ranging from undetectable levels to 29.8 and63.5 ng m−3 in Oporto and Coimbra, respectively. Stigmas-terol was only found in aerosols from Coimbra, neverexceeding 13.0 ng m−3. Lupeol, a tracer of angiospermsmoke, was also identified in both cities (Oporto: 0.112–10.3 ng m−3; Coimbra: undetectable — 18.6 ng m−3).
Of the hundreds of organic constituents present in woodsmoke, 4-substituted methoxylated phenolic compounds(methoxyphenols) have been suggested as potentialmolecularmarkers (Simpson et al., 2005). Some of these compoundswere present in aerosols from Oporto and Coimbra (Fig. 7).Gymnosperm smoke is composed almost exclusively of4-hydroxy-3-methoxy-phenyl (vanillyl) compounds, whilstangiosperms produce both vanillyl and 4-hydroxy-3,5-dimethoxyphenyl (syringyl) constituents (Kjällstrand et al.,1998).
3.2.5. SugarsThree anhydrosugars from the thermal breakdown of
cellulose and hemicellulose were among the most abundantcompounds in aerosols (Table 2). The levoglucosan concen-trations are in the range of values reported for urbanbackground sites in Italy during winter, as well as for otherregions in Europe, the USA or China (Cheng et al., 2013;Giannoni et al., 2012; and the references therein). Thelevoglucosan to mannosan ratios obtained from Oporto arewithin the ranges reported for hardwood combustion insmall-scale units, whilst lower ratios from Coimbra suggest arelatively higher consumption of softwood in this city. Thelevoglucosan to OC ratios have been used to evaluate thecontribution of biomass burning to OC (Alves et al., 2012;Puxbaum et al., 2007). However, this estimation may beproblematic due to the large differences in the levoglucosanto OC ratios among emissions from different biomass speciesand from different combustion conditions, making it verydifficult to determine the most suitable ratio for biomassburning emissions of a given region. Taking into account themain wood species burned in Portugal, Gonçalves et al.(2011) obtained an average levoglucosan to OC ratio of
420 mg g−1 in PM2.5 emitted from traditional combustionappliances commonly used in the country. Based on thisratio, the inputs from wood burning to OC were estimated tobe, on average, 18 and 7.5% in Oporto and Coimbra,respectively. The highest contributions were registered onthe weekends, reaching 66 and 74% at the end of a weekwhen the lowest temperatures of the whole campaign werefelt.
3.2.6. AcidsA series of n-alkanoic acids up to C30 were detected in the
aerosol samples. The average concentrations were 114 and54 ng m−3 in Oporto and Coimbra, respectively. The pre-dominance of even carbon numbered homologues stresses asignificant influence from biological sources of aerosols suchas microbial activities and epicuticular waxes of vascularplants (Ho et al., 2011). Whilst in Oporto, the homologousseries maximised at C16 and C18, the highest concentrationsin Coimbra belonged to C16 and C24. In the latter city a greaterplant wax contribution to the fatty acid fraction wasobserved. It has been observed that, regardless of woodspecies, aerosols from biomass combustion contain highlevels of both hexadecanoic and tetracosanoic acids(Gonçalves et al., 2010, 2011). These fatty acids are basicunits of plant fats, oils and phospholipids (Oros and Simoneit,1999).
Some low molecular weight carboxylic acids (LMWCA)were also present in samples from both cities (Table 3). Theyare commonly detected in urban, suburban, rural, forest,polar, and remote coastal regions (e.g. Fu et al., 2013;Kawamura et al., 2012; Miyazaki et al., 2009; Tsai and Kuo,2013; Yao et al., 2004). LMWCA increase the acidity of rainwater and, because of their hygroscopicity and role as activecloud condensation nuclei (CCN), absorb and disperse solarradiation to change the global thermal balance, and furtherinfluence climate (Acker et al., 2002; Hsieh et al., 2009; Tsaiet al., 2011). These compounds can be directly emitted byfossil fuel combustion and biomass burning, but a substantialfraction is secondarily produced (Kawamura et al., 2012; andreferences therein). Among the LMWCA present in thesamples, glycolic acid, for example, has been described as aSOA product from the reaction between acetic acid and OHradicals (Tan et al., 2012). Benzoic acid is a secondary productfrom the photochemical degradation of aromatic hydrocar-bons emitted by automobiles, such as toluene, and has alsobeen measured as primary pollutant in the exhaust of motorvehicles (Ho et al., 2011). Also, the formation of levulinic acidhas also been recently observed in smog chamber experi-ments involving the toluene photooxidation by NOx (Whiteet al., 2014). Other LMWCA, such as azelaic acid, have beenconsidered as photo-induced oxidation products of biogenicunsaturated fatty acids containing a double bond predomi-nantly at C-9 position (Kawamura and Gagosian, 1987).
Pinic and pinonic acids, formed through the reaction ofozone with α-pinene (Ma et al., 2007), were detected insamples collected in the hottest and sunniest days. InCoimbra, pinic acid achieved maximum concentrations of54.4 ng m−3 (average = 8.27 ng m−3), whilst it was absentin samples from Oporto. Pinonic acid ranged from undetect-able levels to 8.74 ng m−3 in the latter city, and from 0.889to 15.8 ng m−3 in Coimbra. The observation of higher
Fig. 7. Biomass burning compounds (phenolic acids — left panel; resin acids — right panel) in PM2.5 from both cities.
64 C. Alves et al. / Atmospheric Research 150 (2014) 57–68
concentrations of these SOA constituents in Coimbra may berelated to the presence of coniferous trees, which are strongpinene emitters, in the gardens surrounding the samplingsite.
Resin acids represent another group of compoundspresent in all samples. They encompass some diterpenoidcarboxylic acids that are emitted in combustion processes ofconiferous woods (Gonçalves et al., 2011), such as isopimaric,pimaric, abietic and dehydroabietic acids. Ursolic acid wasthe only triterpenic acid detected in some samples.
3.2.7. Other compoundsMethyl esters of n-alkanoic acids represent other sub-
group of compounds present in atmospheric aerosols fromthe Portuguese cities. The homologous series ranged frommethyl tetradecanoate to methyl hexacosanoate. Maximumtotal concentrations of 11.3 and 156 ng m−5 and meanvalues of 15.3 and 3.5 ng m−3 were registered in Oportoand Coimbra, respectively. Methyl alkanoates have beenpreviously detected in fine particles from distinct Europeanregions and have been apportioned to biomass burning(Alves et al., 2012; Schnelle-Kreis et al., 2005). The detectionof methyl alkanoates in PM2.5 emissions from residentialwood combustion (Fine et al., 2004), wildfires (Alves et al.,2011) and prescribed fires (Alves et al., 2010), and theirabsence from vehicular emissions (Rogge et al., 1993),support this assignment.
Tracers for the burning of plastics (Simoneit et al., 2005)have been detected at trace levels in the PM2.5 samples ofboth cities: Irganox 1076 and Irgafos 168. Their presence,although sporadic, is more visible in Coimbra. Somephthalates were also detected in the aerosol samples (1–
Table 2Anhydrosugar concentrations (minimum, maximum and average), levoglucosan tobiomass burning experiments.
100 ng m−3). Terephthalates were the most representativespecies. Phthalates have been classified as toxic, carcino-genic and/or endocrine disruptors (Kluwe, 1986; Sax,2010).
Within the class of labdanes, manoyl oxide and13-epi-manoyl oxide have been identified in almost all theurban background samples, reaching daily concentrations upto 2.6 ng m−3. These natural bicyclic diterpenes are abun-dant components of the essential oils of tens of plants nativeof the Mediterranean region (Demetzos et al., 2002).
3.3. Contribution of major sources to PM2.5
The contributions of the major primary sources toambient PM2.5 can be roughly estimated as follows:
%PM2:5 ¼ Ci½ �= Ci½ �=PM2:5ð Þsource
where [Ci] is the concentration of a source-specific markerobtained in the environmental samples and ([Ci]/PM2.5)sourceis the organic marker to PM2.5 ratio (in mass percentage)determined in source emission tests. The ([Ci]/PM2.5)sourcevalues adopted to estimate the relative input of biomassburning, fossil fuel combustion and cooking emissions tothe PM2.5 concentrations were adopted from Chow et al.(2007). To apportion the contribution of diesel andpetrol-powered vehicles, values of 0.0118 and 0.0146 for17α(H),21β(H)-29-norhopane (percent of emitted PM2.5)were, respectively, used. Although levoglucosan has beenregarded as a universal tracer for biomass burning, whenusing this anhydrosugar to PM2.5 ratios reported for thecombustion of hard- and softwood species in the USA (Chow
mannosan ratios in PM2.5 from Oporto and Coimbra and values reported for
65C. Alves et al. / Atmospheric Research 150 (2014) 57–68
et al., 2007) or values measured in experiments with biofuelsand residential burning appliances of Southern Europe(Gonçalves et al., 2011), unrealistically contributions (some-times higher than 100%) from this source to PM2.5 wereobtained for some days. Thus, retene to PM2.5 ratios of 0.0272and 0.0140 were adopted to apportion the input of hardwoodcombustion and softwood combustion, respectively. Sincethis compound is also emitted during cooking processes, aratio of 0.0059 was used to account for kitchen andrestaurant emissions. It should be stated that this apportion-ment may present some uncertainties, due to the hithertounavailable European emission profiles for major sources.
The contribution of primary vehicular emissions to PM2.5
in Oporto was estimated at around 50%, which can besubdivided into 21% and 28% as inputs from gasoline anddiesel emissions, respectively (Fig. 8). Lower PM2.5 massfractions assigned to petrol (17%) and diesel engines (21%)were obtained in Coimbra. The input from biomass burningrepresented 1 to 12% of PM2.5 in Oporto and 3 to 40% inCoimbra, averaging 3 and 13%, respectively. The highestcontributions were observed on the last weekend of January,when the lowest temperature values of the whole campaignwere attained. In Coimbra, softwood was emitted at a greaterproportion than hardwood smoke, whilst the opposite wasseen in Oporto.
Together with secondary aerosol formation processes,vehicle exhaust and biomass burning emissions have beengenerally pointed out as major sources of airborne PM2.5 inmany parts of the world. Yu et al. (2013) applied PositiveMatrix Factorisation (PMF) to the dataset obtained in anurban environment in Beijing, where daily PM2.5 samplinghad been carried out throughout the year of 2010. Vehicleexhaust and fossil fuel combustion were found to account for38% of the PM2.5 concentrations, whilst biomass burningemissions contributed to 11.2%. At a centrally located urbanmonitoring site in Seoul, Korea, Heo et al. (2009) found, alsoby applying PMF, that secondary constituents representedthe major PM2.5 mass fraction (around 42%), whereas thecontributions from other sources were less important:gasoline-fuelled vehicles (17.2%), biomass burning (12.1%),and diesel emissions (8.1%).
Lee et al. (2008) applied both PMF and CMB (ChemicalMass Balance) to 3-year PM2.5 data at two urban sites(Atlanta, GA and Birmingham, AL) and two rural sites(Yorkville, GA and Centreville, AL). Both CMB and PMFresults showed that secondary constituents are the dominantcontributors (33–45%) of ambient PM2.5 mass in urban andrural areas. Motor vehicles and wood burning were identifiedas the two major primary sources. The former contributed17–25% in urban areas and 7–9% in rural areas; the lattercontributed 6–13% in urban areas and 6–30% in rural areas.Marmur et al. (2006) estimated the daily source impacts atthe same four sites by applying an extended chemical massbalance receptor model (CMB-LGO) and an emissions-basedair-quality model (CMAQ). Biomass burning was identifiedas a major source of primary PM2.5 in the region, withcontributions ranging between 27% and 77% (higher fractionsin the rural sites). The average contribution from dieselvehicles ranged between 5% and 31% of the primary PM2.5,with higher fractions at the urban sites. The solutionsobtained by CMB indicate that, even at the rural sites, thecontributions from mobile sources comprised approximately50% of the primary PM2.5, and that the gasoline-vehiclecontribution at Birmingham was nearly 70% of the primaryPM2.5, with an extremely high gasoline-to-diesel ratio of 10.4.On the other hand, solutions obtained by CMB-LGO andCMAQ indicated that the gasoline-to-diesel ratio was b1 andthat wood burning was the major source of primary PM2.5 atthe rural sites. These results demonstrated the collinearityproblem often encountered with regular CMB sourceapportionment.
Similar contributions from cooking emissions were ob-tained in both Portuguese cities. PM2.5 mass fractions from 2to 23% and from 1 to 27% were obtained in Oporto andCoimbra, respectively. On average, the input of this sourceaccounted for 10% of the aerosol levels. The cookingcontribution was estimated to be 5–19% of PM2.5 at Fresno,California (Chow et al., 2007). Allan et al. (2010) estimatedthat traffic represented, on average, 40% of the primaryorganic aerosol collected in London and Manchester duringcolder conditions, whilst cooking emissions accounted for34%. Estimates based on AMS measurements indicated that
0
20
40
60
80
100
27
-01
-20
07
28
-01
-20
07
29
-01
-20
07
30
-01
-20
07
31
-01
-20
07
01
-02
-20
07
02
-02
-20
07
03
-02
-20
07
04
-02
-20
07
05
-02
-20
07
06
-02
-20
07
07
-02
-20
07
08
-02
-20
07
09
-02
-20
07
10
-02
-20
07
11
-02
-20
07
12
-02
-20
07
13
-02
-20
07
14
-02
-20
07
15
-02
-20
07
16
-02
-20
07
17
-02
-20
07
18
-02
-20
07
19
-02
-20
07
20
-02
-20
07
21
-02
-20
07
22
-02
-20
07
23
-02
-20
07
24
-02
-20
07
25
-02
-20
07
26
-02
-20
07
27
-02
-20
07
% o
f PM
2.5
mas
sOPORTO Gasoline Diesel Softwood Hardwood combustion Cooking Other sources/processes
0
20
40
60
80
100
27
-01
-20
07
28
-01
-20
07
29
-01
-20
07
30
-01
-20
07
31-0
1-2
007
01
-02
-20
07
02
-02
-20
07
03
-02
-20
07
04
-02
-20
07
05
-02
-20
07
06
-02
-20
07
07
-02
-20
07
08
-02
-20
07
09
-02
-20
07
10
-02
-20
07
11
-02
-20
07
12
-02
-20
07
13
-02
-20
07
14
-02
-20
07
15
-02
-20
07
16-0
2-2
007
17
-02
-20
07
18
-02
-20
07
19
-02
-20
07
20
-02
-20
07
21
-02
-20
07
22
-02
-20
07
23
-02
-20
07
24
-02
-20
07
25-0
2-2
007
26
-02
-20
07
27
-02
-20
07
% o
f PM
2.5
mas
s
COIMBRA Gasoline Diesel Hardwood combustion Softwood combustion Cooking Other sources/processes
2.5
OPORTO
COIMBRA
Fig. 8. Mass fraction of PM2.5 (%) assigned to different sources.
66 C. Alves et al. / Atmospheric Research 150 (2014) 57–68
17% of the organic aerosol in Barcelona originates fromcooking (Mohr et al., 2012).
The mass fractions not assigned to the major primaryanthropogenic sources ranged between 24 and 77% andbetween 12 and 74% in Oporto and Coimbra, respectively. Ingeneral, higher values were obtained on days under theinfluence of air masses with Atlantic origin, suggesting asignificant contribution from sea salt. Moreover, this fractionalso includes other primary inputs, as well as secondaryinorganic aerosol (e.g. particulate sulphate, ammonium andnitrate) and SOA constituents.
4. Conclusions
Some differences between Oporto and Coimbra wereobserved. The urban background atmosphere of Oportoshowed the highest levels of PM2.5, OC and hydrocarboncompounds. Molecular ratios and the presence of petroleummarkers point out vehicle exhausts, especially from dieselpowered engines, as the major pollutant source. Biomassburning also represents a significant emission source of
organic compounds in both cities. A rough source apportion-ment indicated that the primary contributions of petrol- anddiesel-powered vehicle emissions to PM2.5 in Oporto are, onaverage, 21 and 28%, respectively, whilst the correspondingmass fractions in Coimbra are 17 and 21%. On the coldestdays, emissions from biomass burning may account for up to40% and 12% of the PM2.5 levels in Coimbra and Oporto,respectively. A similar contribution, around 10%, fromcooking activities was estimated for both cities. Taking intoaccount the rising prices of fossil fuels and electricity, it isexpected that a growing band of residents who want cheaperand locally-available fuels will be increasingly looking atwood fuel boilers and stoves instead of those traditionalforms of energy. This will certainly change the atmosphericpollution patterns. To devise cost-effective mitigation mea-sures, monitoring programmes, carried out in a regular basis,are recommended. This is even more important because, inthis study, PM2.5 and the carcinogenic content exceeded therecommended levels on several days. Thus, it is imperative tocontrol regional combustion sources to reduce the healthrisks associated with air pollution.
67C. Alves et al. / Atmospheric Research 150 (2014) 57–68
Acknowledgements
The sampling campaign was performed within the projectPOCI/AMB/60267/2004 funded by the Portuguese ScienceFoundation (FCT). The analytical work was also funded byFCT through the project “Source apportionment of URBanEmissions of primary particulate matter”, PTDC/AAC-AMB/117956/2010 (URBE).
Appendix A. Supplementary data
Supplementary data to this article can be found online athttp://dx.doi.org/10.1016/j.atmosres.2014.07.012.
References
Acker, K.,Mertes, S.,Moller, D.,Wieprecht, W., Auel, R., Kalass, D., 2002. Casestudy of cloud physical and chemical processes in low clouds at Mt.Brocken. Atmos. Res. 64, 41–51.
Alam, M.S., West, C.E., Scarlett, A.G., Rowland, S.J., Harrison, R.M., 2013.Application of 2D-GCMS reveals many industrial chemicals in airborneparticulate matter. Atmos. Environ. 65, 101–111.
Alves, C., 2008. Characterisation of solvent extractable organic constituentsin atmospheric particulate matter: an overview. An. Acad. Bras. Cienc.80, 21–82.
Alves, C.,Pio, C.A.,Duarte, A., 2001. Composition of extractable organic matterof air particles from rural and urban Portuguese areas. Atmos. Environ.35, 5485–5496.
Alves, C.A., Gonçalves, C., Evtyugina, M., Pio, C.A., Mirante, F., Puxbaum, H.,2010. Particulate organic compounds emitted from experimentalwildland fires in a Mediterranean ecosystem. Atmos. Environ. 44,2750–2759.
Alves, C.A.,Vicente, A.,Monteiro, C.,Gonçalves, C.,Evtyugina, M.,Pio, C., 2011.Emission of trace gases and organic components in smoke particles froma wildfire in a mixed-evergreen forest in Portugal. Sci. Total Environ.409, 1466–1475.
Alves, C., Vicente, A., Pio, C., Kiss, G., Hoffer, A., Decesari, S., Prevot, A.S.H.,Minguillón, M.C., Querol, X., Hillamo, R., Spindler, G., Swietlicki, E., 2012.Organic compounds in aerosols from selected European sites — biogenicversus anthropogenic sources. Atmos. Environ. 59, 243–255.
Bi, X., Simoneit, B.R.T., Sheng, G., Ma, S., Fu, J., 2008. Composition and majorsources of organic compounds in urban aerosols. Atmos. Res. 88,256–265.
Carslaw, K.S., Boucher, O., Spracklen, D.V.,Mann, G.W., Rae, J.G.L.,Woodward,S., Kulmala, M., 2010. A review of natural aerosol interactions andfeedbacks within the Earth system. Atmos. Chem. Phys. 10, 1701–1737.
Chen, Y.,Ho, K.F.,Ho, S.S.H.,Ho, W.K.,Lee, S.C.,Yu, J.Z.,Sit, E.H.L., 2007. Gaseousand particulate polycyclic aromatic hydrocarbons (PAHs) emissionsfrom commercial restaurants in Hong Kong. J. Environ. Monit. 9,1402–1409.
Chen, Y., Feng, Y., Xiong, S., Liu, D.,Wang, G., Sheng, G., Fu, J., 2011. Polycyclicaromatic hydrocarbons in the atmosphere of Shanghai. Environ. Monit.Assess. 172 (235–247), 2011.
Chow, J.C.,Watson, J.G.,Lowenthal, D.H.,Chen, L.W.A.,Zielinska, B.,Mazzoleni,L.R., Magliano, K.L., 2007. Evaluation of organic markers for chemicalmass balance source apportionment at the Fresno Supersite. Atmos.Chem. Phys. 7, 1741–1754.
Demetzos, C., Anastasaki, T., Perdetzoglou, D., 2002. A chemometricinterpopulation study of the essential oils of Cistus creticus L. growingin Crete (Greece). Z. Naturforsch. 57c, 89–94.
Draxler, R.R., Rolph, G.D., 2013. HYSPLIT (HYbrid Single-Particle LagrangianIntegrated Trajectory) Model Access Via NOAA ARL READY(Website(http://www.arl.noaa.gov/HYSPLIT.php)) NOAA Air Resources Labora-tory, College Park, MD.
Fang, M.,Zheng, M.,Wang, F.,To, K.L.,Jaafar, A.B.,Tong, S.L., 1999. The solvent-extractable organic compounds in the Indonesia biomass burningaerosols — characterization studies. Atmos. Environ. 33, 783–795.
Fine, P.M.,Cass, G.R., Simoneit, B.R.T., 2001. Chemical characterization of fineparticle emissions from fireplace combustion of woods grown in thenortheastern United States. Environ. Sci. Technol. 35, 2665–2675.
Fine, P.M.,Cass, G.R., Simoneit, B.R.T., 2004. Chemical characterization of fineparticle emissions from the fireplace combustion of wood types grownin the midwestern and western United States. Environ. Eng. Sci. 21,387–409.
Fu, P., Kawamura, K., Usukura, K., Miura, K., 2013. Dicarboxylic acids,ketocarboxylic acids and glyoxal in the marine aerosol collected duringa round-the-world cruise. Mar. Chem. 148, 22–32.
Giannoni, M.,Martellini, T.,Del Bubba, M.,Gambaro, A., Zangrando, R., Chiari,M., Lepri, L., Cincinelli, A., 2012. The use of levoglucosan for tracingbiomass burning in PM2.5 samples in Tuscany (Italy). Environ. Pollut.167, 7–15.
Gogou, A., Apostolaki, M., Stephanou, E., 1998. Determination of organicmarkers in marine aerosols and sediments: one-step flash chromatog-raphy compound class fractionation and capillary gas chromatographicanalysis. J. Chromatogr. A 799, 215–231.
Gonçalves, C.,Alves, C.,Evtyugina, M.,Mirante, F., Pio, C.,Caseiro, A., Schmidl, C.,Bauer, H., Carvalho, F., 2010. Characterisation of PM10 emissions fromwood stove combustion of common woods grown in Portugal. Atmos.Environ. 44, 4474–4480.
Gonçalves, C.,Alves, C.,Fernandes, A.P.,Monteiro, C.,Tarelho, L.,Evtyugina, M.,Pio, C., 2011. Organic compounds in PM2.5 emitted from fireplace andwoodstove combustion of typical Portuguese wood species. Atmos.Environ. 45, 4533–4545.
Gursale, A.,Dighe, V.,Parekl, G., 2010. Simultaneous quantitative determina-tion of cinnamaldehyde and methyl eugenol from stem bark ofCinnamomum zeylanicum blume using RP-HPLC. J. Chromatogr. Sci. 48,59–62.
Haddad, I.E., Marchand, N., Dron, J., Temine-Roussel, B., Quivet, E.,Wortham,H., Jaffrezo, J.L., Baduel, C., Voisin, D., Besombles, J.L., Gille, G., 2009.Comprehensive primary particulate organic characterization of vehicu-lar exhaust emissions in France. Atmos. Environ. 43, 6190–6198.
Heo, J.B.,Hopke, P.K.,Yi, S.M., 2009. Source apportionment of PM2.5 in Seoul,Korea. Atmos. Chem. Phys. 9, 4957–4971.
Herlekar, M., Joseph, A.E.,Kumar, R.,Gupta, I., 2012. Chemical speciation andsource assignment of particulate matter (PM10) phase molecularmarkers in Mumbai. Aerosol Air Qual. Res. 12, 1247–1260.
Ho, K.F., Ho, S.S.H., Lee, S.C., Kawamura, K., Zou, S.C., Cao, J.J., Xu, H.M., 2011.Summer and winter variations of dicarboxylic acids, fatty acids andbenzoic acid in PM2.5 in Pearl Delta River Region, China. Atmos. Chem.Phys. 11, 2197–2208.
Hsieh, L.Y., Kuo, S.C., Chen, C.L., Tsai, Y.I., 2009. Size distributions of nano/micron dicarboxylic acids and inorganic ions in suburban PM episodeand non-episodic aerosol. Atmos. Environ. 43, 4396–4406.
Islam, T., Gauderman, W.J., Berhane, K., McConnell, R., Avol, E., Peters, J.M.,Gilliland, F.D., 2007. Relationship between air pollution, lung functionand asthma in adolescents. Thorax 62, 957–963.
Jaoui, M., Kamens, R.M., 2003. Gaseous and particulate oxidation productsanalysis of a mixture of α-pinene + β-pinene/O3/air in the absence oflight and α-pinene + β-pinene/NOx/air in the presence of naturalsunlight. J. Atmos. Chem. 44, 259–297.
Jørgensen, R.B.,Strandberg, B.,Sjaastad, A.K., Johansen, A.,Svendsen, K., 2013.Simulated restaurant cook exposure to emissions of PAHs, mutagenicaldehydes, and particles from frying bacon. J. Occup. Environ. Hyg. 10,122–131.
Kallio, M.,Jussila, M.,Rissanen, T.,Antilla, P.,Hartonen, K.,Reissell, A.,Vreuls, R.,Adahchour, M., Hyötyläinen, T., 2006. Comprehensive two-dimensionalgas chromatography coupled to time-of-flight mass spectrometry in theidentification of organic compounds in atmospheric aerosols fromconiferous forest. J. Chromatogr. A 1125, 234–243.
Kawamura, K., Gagosian, R.B., 1987. Implication of ω-oxocarboxylic acids inthe remote marine atmosphere for photo-oxidation of unsaturated fattyacids. Nature 325, 330–332.
Kawamura, K., Ono, K., Tachibana, E., Charriére, B., Sempéré, R., 2012.Distributions of low molecular weight dicarboxylic acids, ketoacids andα-dicarbonyls in the marine aerosols collected over the Arctic Oceanduring late summer. Biogeosciences 9, 4725–4737.
Kjällstrand, J., Ramnäs, O., Petersson, G., 1998. Gas chromatographic andmass spectrometric analysis of 36 lignin-related methoxyphenolsfrom uncontrolled combustion of wood. J. Chromatogr. A 824,205–210.
Kluwe, W.M., 1986. Carcinogenic potential of phthalic acid esters and relatedcompounds: structure–activity relationships. Environ. Health Perspect.65, 271–278.
68 C. Alves et al. / Atmospheric Research 150 (2014) 57–68
Krumal, K., Mikuska, P., Vecera, Z., 2013. Polycyclic aromatic hydrocarbonsand hopanes in PM1 aerosols in urban areas. Atmos. Environ. 67, 27–37.
Lee, S., Liu, W., Wang, Y., Russell, A.G., Edgerton, E.S., 2008. Sourceapportionment of PM2.5: comparing PMF and CMB results for fourambient monitoring sites in the southeastern United States. Atmos.Environ. 42, 4126–4137.
Ma, Y., Willcox, T.R., Russell, A.T., Marston, G., 2007. Pinic and pinonic acidformation in the reaction of ozone with α-pinene. Chem. Commun.2007, 1328–1330.
Marmur, A.,Park, S.K.,Mulholland, J.A.,Tolbert, P.E.,Russell, A.G., 2006. Sourceapportionment of PM2.5 in the southeastern United States using receptorand emissions-based models: conceptual differences and implicationsfor time-series health studies. Atmos. Environ. 40, 2533–2551.
Mazquiarán, M.A.B.,Pinedo, L.C.O., 2007. Organic composition of atmospher-ic urban aerosol: variations and sources of aliphatic and polycyclicaromatic hydrocarbons. Atmos. Res. 85, 288–299.
Miyazaki, Y., Aggarwal, S.G., Singh, K., Gupta, P.K., Kawamura, K., 2009.Dicarboxylic acids and water-soluble organic carbon in aerosols in NewDelhi, India in winter: characteristics and formation processes. J.Geophys. Res. 114, D19206. http://dx.doi.org/10.1029/2009JD011790.
Mohr, C.,DeCarlo, P.F.,Heringa, M.F.,Chirico, R.,Slowik, J.G.,Richter, R.,Reche,C., Alastuey, A., Querol, X., Seco, R., Peñuelas, J., Jiménez, J.L., Crippa, M.,Zimmermann, R., Baltensperger, U., Prévôt, A.S.L., 2012. Identificationand quantification of organic aerosol from cooking and other sources inBarcelona using mass spectrometer data. Atmos. Chem. Phys. 12,1649–1665.
Oliveira, C., Pio, C., Alves, C., Evtyugina, M., Santos, P.,Gonçalves, V.,Nunes, T.,Silvestre, A., Palmgren, F., Wahlin, P., Harrad, S., 2007. Seasonaldistribution of polar organic compounds in the urban atmosphere oftwo large cities from the North and South of Europe. Atmos. Environ. 41,5555–5570.
Oros, D.R., Simoneit, B.R.T., 2000. Identification and emissions ratios ofmolecular tracers in coal smoke particulate matter. Fuel 79, 515–536.
Pio, C.,Cerqueira, M.,Harrison, R.M.,Nunes, T.,Mirante, F.,Alves, C.,Oliveira, C.,Sanchez de la Campa, A., Artiñano, B., Matos, M., 2011. OC/EC RatioObservations in Europe: re-thinking the approach for apportionmentbetween primary and secondary organic carbon. Atmos. Environ. 45,6121–6132.
Puxbaum, H., Caseiro, A., Sánchez-Ochoa, A., Kasper-Giebl, A., Claeys, M.,Gelencsér, A.,Legrand, M.,Preunkert, S.,Pio, C., 2007. Levoglucosan levelsat background sites in Europe for assessing the impact of biomasscombustion on the European aerosol background. J. Geophys. Res. 112,D23S05. http://dx.doi.org/10.1029/2006JD008114.
Rao, X., Kobayashi, R.,White-Morris, R., Spaulding, R., Frazey, P., Charles, J.M.,2001. GC/ITMS measurement of carbonyls andmultifunctional carbonylsin PM2.5 particles emitted from motor vehicles. J. AOAC Int. 84, 699–705.
Riva, D.R.,Magalhães, C.B.,Lopes, A.A.,Lanças, T.,Mauad, T.,Malm, O.,Valença,S.S., Saldiva, P.H., Faffe, D.S., Zin, W.A., 2011. Low dose of fine particulatematter (PM2.5) can induce acute oxidative stress, inflammation andpulmonary impairment in healthy mice. Inhal. Toxicol. 23, 257–267.
Rogge, W.F., Hildemann, L.M., Mazurek, M., Cass, G.R., 1993. Sources of fineorganic aerosol: 2. Noncatalyst and catalyst-equipped automobiles andheavy-duty diesel trucks. Environ. Sci. Technol. 27, 636–651.
Rontani, J.F.,Cuny, P.,Grossi, V., 1996. Photodegradation of chlorophyll phytylchain in senescent leaves of higher plants. Phytochem. 42, 347–351.
Sax, L., 2010. Polyethylene terephthalate may yield endocrine disruptors.Environ. Health Perspect. 118, 445–448.
Schnelle-Kreis, Sklorz, J.M., Peters, A., Cyrys, J., Zimmermann, R., 2005.Analysis of particle-associated semi-volatile aromatic and aliphatichydrocarbons in urban particulate matter on a daily basis. Atmos.Environ. 39, 7702–7714.
Simoneit, B.R.T., Medeiros, P.M., Didyk, B.M., 2005. Combustion products ofplastics as indicators for refuse burning in the atmosphere. Environ. Sci.Technol. 39, 6961–6970.
Simoneit, B.R.T., Bi, X., Oros, D.R., Medeiros, P.M., Sheng, G., Fu, J., 2007.Phenols and hydroxyl-PAHs (arylphenols) as tracers for coal smoke
particulate matter: source test and ambient aerosol assessments.Environ. Sci. Technol. 41, 7294–7302.
Simpson, C.D., Paulsen, M., Dills, R.L., Liu, L.J.S., Kalman, D.A., 2005.Determination of methoxyphenols in ambient atmospheric particulatematter: tracers for wood combustion. Environ. Sci. Technol. 39, 631–637.
Song, Z., You, M.,Duzgoren-Aydin, N.S., 2005. Characterization of particulateorganics accumulated on the ceiling of vehicular tunnels in Hong Kongand Guangzhou, China. Atmos. Environ. 39, 6398–6408.
Tan, Y., Lim, Y.B., Altieri, K.E., Seitzinger, S.P., Turpin, B.J., 2012. Mechanismsleading to oligomers and SOA through aqueous photooxidation: insightsfrom OH radical oxidation of acetic acid and methylglyoxal. Atmos.Chem. Phys. 12, 801–813.
Tsai, Y.I., Kuo, S.C., 2013. Contributions of low molecular weight carboxylicacids to aerosols and wet deposition in a natural subtropical broad-leaved forest environment. Atmos. Environ. 81, 270–279.
Tsai, Y.I.,Hsieh, L.Y.,Kuo, S.C.,Chen, C.L.,Wu, P.L., 2011. Seasonal and rainfall-type variations in inorganic ions and dicarboxylic acids and acidity ofwet deposition samples collected from subtropical East Asian. Atmos.Environ. 45, 3535–3547.
Tsapakis, M.,Lagoudaki, E.,Stephanou, E.G.,Kavouras, I.G.,Koutrakis, P.,Oyola,P., von Baer, D., 2002. The composition and sources of PM2.5 organicaerosol in two urban areas of Chile. Atmos. Environ. 36, 3851–3863.
Wang, G., Kawamura, K., Watanabe, T., Lee, S., Ho, K., Cao, J., 2006. Highloadings and source strengths of organic aerosols in China. Geophys. Res.Lett. 33. http://dx.doi.org/10.1029/2006GL027624 L22801.
Wang, G., Kawamura, K., Xie, M., Hu, S., Gao, S., An, Z.,Wang, Z., 2009a. Size-distributions of n-alkanes, PAHs and hopanes and their sources in theurban, mountain and marine atmospheres over East Asia. Atmos. Chem.Phys. 9, 8869–8882.
Wang, W., Tao, S., Wang, W., Shen, G., Zhao, J., Lam, K.C., 2009b. Airborneparticulates and polycyclic aromatic hydrocarbons (PAHs) in ambientair in Donghe, Northern China. J. Environ. Sci. Health A. 44, 854–860.
Wei, S.,Huang, B.,Liu, M.,Bi, X.,Ren, Z.,Sheng, G.,Fu, J., 2012. Characterizationof PM2.5-bound nitrated and oxygenated PAHs in two industrial sites ofSouth China. Atmos. Res. 109–110, 76–83.
White, S.J., Jamie, I.M.,Angove, D.E., 2014. Chemical characterisation of semi-volatile and aerosol compounds from the photooxidation of toluene andNOx. Atmos. Environ. 83, 237–244.
Williams, B.J., Goldstein, A.H., Millet, D.B., Holzinger, R., Kreisberg, N.M.,Hering, S.V., White, A.B., Worsnop, D.R., Allan, J.D., Jimenez, J.L., 2007.Chemical speciation of organic aerosol during the International Consor-tium for Atmospheric Research on Transport and Transformation 2004:results from in situ measurements. J. Geophys. Res. 112, D10S26. http://dx.doi.org/10.1029/2006JD007601.
Yadav, S., Tandon, A., Attri, A.K., 2013. Monthly and seasonal variations inaerosol associated n-alkane profiles in relation to meteorologicalparameters in New Delhi, India. Aerosol Air Qual. Res. 13, 287–300.
Yao, X., Fang, M., Chan, C.K., Ho, K.F., Lee, S.C., 2004. Characterization ofdicarboxylic acids in PM2.5 in Hong Kong. Atmos. Environ. 38, 963–970.
Yu, L.,Wang, G.,Zhang, R.,Zhang, L., Song, Y.,Wu, b,Li, X.,An, K.,Chu, J., 2013.Characterization and source apportionment of PM2.5 in an urbanenvironment in Beijing. Aerosol Air Qual. Res. 13, 574–583.
Yunker, M.B., Macdonald, R.W., Vingarzan, R., Mitchell, R.H., Goyette, D.,Sylvestre, S., 2002. PAHs in the Fraser River basin: a critical appraisal ofPAH ratios as indicators of PAH sources and composition. Org. Geochem.33, 489–515.
Zhang, Q.,Jimenez, J.L.,Canagaratna, M.R.,Ulbrich, I.M.,Ng, N.L.,Worsnop, D.R.,Sun, Y., 2011. Understanding atmospheric organic aerosols via factoranalysis of aerosol mass spectrometry: a review. Anal. Bioanal. Chem.401, 3045–3067.
Zhao, Y.,Hu, M.,Slanina, S.,Zhang, Y., 2007. The molecular distribution of fineparticulate organic matter emitted from Western-style fast foodcooking. Atmos. Environ. 41, 8163–8171.
