Characterization of Two Passive Air Samplers for Per- andPolyfluoroalkyl SubstancesLutz Ahrens,*,†,‡ Tom Harner,*,† Mahiba Shoeib,† Martina Koblizkova,† and Eric J. Reiner§,⊥

†Environment Canada, Air Quality Processes Research Section, Toronto, Ontario M3H 5T4, Canada‡Swedish University of Agricultural Sciences (SLU), Department of Aquatic Sciences and Assessment, Uppsala, Uppland SE-750 07,Sweden§Ontario Ministry of the Environment, 125 Resources Road, Toronto, Ontario M9P 3V6, Canada⊥University of Toronto, Department of Chemistry, 80 St. George Street, Toronto, Ontario M5S 3H6, Canada

*S Supporting Information

ABSTRACT: Two passive air sampler (PAS) media werecharacterized under field conditions for the measurement ofper- and polyfluoroalkyl substances (PFASs) in the atmos-phere. The PASs, consisting of polyurethane foam (PUF) andsorbent-impregnated PUF (SIP) disks, were deployed for overone year in parallel with high volume active air samplers (HV-AAS) and low volume active air samplers (LV-AAS). Sampleswere analyzed for perfluoroalkyl carboxylic acids (PFCAs),perfluoroalkane sulfonic acids (PFSAs), fluorotelomer alcohols(FTOHs), fluorotelomer methacrylates (FTMACs), fluoro-telomer acrylates (FTACs), perfluorooctane sulfonamides (FOSAs), and perfluorooctane sulfonamidoethanols (FOSEs).Sampling rates and the passive sampler medium (PSM)-air partition coefficient (KPSM−A) were calculated for individual PFASs.Sampling rates were similar for PFASs present in the gas phase and particle phase, and the linear sampling rate of 4 m−3 d−1 isrecommended for calculating effective air sample volumes in the SIP-PAS and PUF-PAS for PFASs except for the FOSAs andFOSEs in the PUF-PAS. SIP disks showed very good performance for all tested PFASs while PUF disks were suitable only for thePFSAs and their precursors. Experiments evaluating the suitability of different isotopically labeled fluorinated depurationcompounds (DCs) revealed that 13C8-perfluorooctanoic acid (PFOA) was suitable for the calculation of site-specific samplingrates. Ambient temperature was the dominant factor influencing the seasonal trend of PFASs.

■ INTRODUCTION

Per- and polyfluoroalkyl substances (PFASs) have receivedincreasing public attention due to their persistence, bioaccu-mulative potential, and possible adverse effects on humans andwildlife.1 PFASs comprise a diverse group of chemicalsincluding, for example, fluorotelomer alcohols (FTOHs),fluorotelomer acrylates (FTACs), perfluorooctane sulfona-mides (FOSAs), perfluorooctane sulfonamidoethanols(FOSEs), perfluoroalkyl carboxylic acids (PFCAs), andperfluoroalkyl sulfonic acids (PFSAs). They have been widelyused in a variety of consumer and industrial products such asmetal plating, semiconductors, polishing agents, paints,surfactants in textile coatings, paper treatments, and firefightingfoams.2,3 Once released into the environment, PFASs can beglobally transported by ocean currents and the atmosphere.4,5

However, few data are available for atmospheric PFASs, inparticular for the PFSAs and PFCAs due to their uniquecharacteristics (e.g., ionizability) and low concentrationlevels.6−10 Thus, there is a need for a simple samplingtechnique to improve our understanding of the temporaltrends and spatial distribution of PFASs in a global context.

High volume active air samplers (HV-AAS) are typically usedfor measuring PFASs in the atmosphere because of their abilityto provide information on the gas- and particle-phasedistribution and high temporal resolution. However, HV-AASdepend on power supplies, and sampling artifacts have beenreported for PFSAs and PFCAs using conventional HV-AAS.11,12 In contrast, passive air samplers (PAS) generate time-integrated data and are ideal due to their simplicity and lowcost, especially for the purpose of spatial and long-termtemporal trend studies.13,14 Polyurethane foam (PUF) disks arethe most widely used PAS for persistent organic pollutants(POPs).13,15 A new PAS type was developed by Shoeib et al.comprising sorbent-impregnated PUF (SIP) disks to increasethe sorptive capacity for more volatile chemicals like FTOHs.16

In general, the uptake of the chemical depends on its diffusivityin air and the passive sampler medium (PSM)-air partitioncoefficient (KPSM−A), which depends on the PSM andcharacteristics of the chemical.17 In addition, the chamber

Received: May 22, 2013Accepted: November 12, 2013

Article

pubs.acs.org/est

© XXXX American Chemical Society A dx.doi.org/10.1021/es4048945 | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

design has an influence on the amount of particles sampled bythe PAS.18 Ultimately, the sampling rate of the chemical is alsoinfluenced by meteorological conditions like wind speed andtemperature.17−19 To compensate varying meteorologicalconditions, depuration compounds (DCs) can be used tocalculate the site-specific sampling rate by assessing their lossduring the deployment period.20 However, it has to beconsidered that uptake rates of the chemical of interest arenot necessarily equal to the loss of the DC due toinhomogeneities in diffusivities within the PSM.21 Furthermore,there is some uncertainty regarding the ability of PAS tocapture PFASs (in particular, PFSAs and PFCAs) in air andhow to derive air concentrations for PFASs with a high particleassociated fraction (e.g., FOSEs, longer chained PFSAs andPFCAs).22 The addition of perfluorooctane sulfonic acid(PFOS) (and its salts and precursors) to the StockholmConvention on POPs in 2009 means that air monitoringnetworks reporting to the global monitoring plan (GMP) willbe required to measure these compounds in air.23 The results ofthis study are very relevant therefore and provide guidance onthe use of PUF-disks or SIP-disks for monitoring PFOS andprecursors (i.e., FOSE and FOSAs) in air. The specificobjectives of this study include (i) to assess the samplingrates and KPSM−A values for PFASs for PUF and SIP disks basedon field calibration against HV-AAS and low-volume active airsamplers (LV-AAS), (ii) to evaluate the suitability of threefluorinated DCs, and (iii) to assess the comparability of fourdifferent air sampling techniques for measuring PFASs.

■ EXPERIMENTAL SECTIONChemicals. The target analytes included C4, C6, C8, C10

(PFBS, PFHxS, PFOS, PFDS) PFSAs (CnF2n+1SO3H), C4−C12,C14 (PFBA, PFPeA, PFHxA, PFHpA, perfluorooctanoic acid(PFOA), PFNA, PFDA, PFUnDA, PFDoDA, PFTeDA)PFCAs (C nF2 n+1COOH), 6:2 , 8 :2 , 10:2 FTOHs(CnF2n+1CH2CH2OH), 6:2 fluorotelomer methacrylate(FTMAC, C6F13CH2CH2OC(O)C(CH3)=CH2), 8:2, 10:2FTACs (CnF2n+1CH2CH2OC(O)CHCH2), FOSA(C8F17SO2NH2), methyl and ethyl FOSAs (C8F17SO2N-(CnH2n+1)H), and methyl and ethyl FOSEs (C8F17SO2N-(CnH2n+1)CH2CH2OH). In addition, 17 mass-labeled internalstandards (IS), three injection standards (InjS) (i.e., N,N-dimethyl perfluorooctane sulfonamide (Me2FOSA,C8F17SO2N(CH3)(CH3)),

13C8−PFOS, and 13C8−PFOA),and three DCs (i.e., perfluoroheptylethanol (7:2 sFTOH,C7F15CH(OH)CH3),

13C8−PFOS, and 13C8−PFOA) wereused. Details are provided in Tables S1 and S2 of theSupporting Information (SI).Sampling. The calibration of the PAS was conducted from

March 30 to October 13, 2010 at a semiurban meteorologicalstation in Toronto (Environment Canada field site, 43°46′ N,79°28′ W). Two different PAS media (i.e., PUF and SIP disks)were evaluated against parallel samples collected using LV-AASand HV-AAS. After the completion of the calibrationcomponent (October 2010), sampling for determining ambientair concentrations of PFASs continued until the end of April2011, using both PAS media and HV-AAS.21

Under the calibration study component, the PUF-PAS andSIP-PAS were deployed for 7, 21, 28, 42, 56, 84, 112, 140, 168,and 197 days. Duplicate PUF-PAS were collected on days 28,84, and 197 to verify reproducibility. SIP-PAS were preparedaccording to the protocol from Shoeib et al. 16 Briefly,precleaned PUF disks (14 cm diameter ×1.35 cm thick; surface

area 365 cm2, mass 4.40 g, volume 207 cm3, TischEnvironmental, Cleves, OH, USA) were impregnated withfinely ground XAD-4 resin (Supelco, Bellefonte, PA) (∼0.5 gper PUF disk) (for details, see elsewhere16). During fielddeployment, the SIP and PUF disks were individually housedinside precleaned stainless steel chambers (“original chamber”,Model TE-200-PAS, Tisch Environmental) and deployed ∼2 mabove the ground.To compare different chamber configurations, four different

chambers were used with different gaps between the twostainless steel dome housings (i.e., “original chamber” with 1cm overlap, “flush chamber”, “1 cm gap chamber”, “2 cm gapchamber”) (see Figure S1 in the SI). The PUF and SIP diskswere deployed in the different chambers for 28 days over 5sampling periods (Table S5 and S6 in the SI). The comparisonof the different chamber configurations will provide informationabout the influence of the chamber design on the collection ofparticles on the PUF and SIP disks and the potential windspeed effects that we expect to be dampened by the moreprotective (less open) chamber configurations.The PUF-PAS and SIP-PAS (including blanks) were spiked

with 20 ng absolute of the DCs 13C8−PFOS, 13C8−PFOA, and7:2 sFTOH prior to field deployment. The loss of the DCs (i.e.,volatilization from the PAS) during the deployment period ofthe PAS provides information of the site-specific sampling ratesthat account for the wind and temperature effects.18,20 The HV-AAS were not spiked with DCs. DCs were not detected in theHV-AAS samples indicating that no migration of the DCsoccurred to other samplers via air transport.For active sampling, high volume air samples (∼330 m3 over

24 h periods, one to two times a week) were collected fromMarch, 2010 to April, 2011 using glass-fiber filters (GFFs)(Type A/E Glass, 102 mm diameter, Pall Corporation) forcollecting the particle phase (n = 70) followed by a PUF/XAD−2 cartridge for trapping the gas-phase compounds (n =70). In addition, low volume air samples (∼46 m3) werecollected from March to October, 2010 using LV-AAS (n = 14)to provide time-integrated concentrations (integrated over 14days) (for details, see SI).Field blanks for PUFs, SIPs, GFFs, and PUF/XAD−2

cartridges were collected by exposing them for 1 min at thesampling site and then treating them like real samples. Tocheck the efficiency of the collection of PFASs in the gas phase,breakthrough experiments were conducted by operating theactive air samplers in series (i.e., two sets of collection PUF/XAD−2 media) and analyzing the first and second setseparately. These tests were conducted for both HV-AAS (n= 3) and LV-AAS (n = 3). All samples were stored at −20 °Cuntil extraction within four weeks. Details of the sampling,dates, air volume, and meteorological data are presented inTables S3−6 in the SI.

Sample Extraction and Instrumental Analysis. Theextraction and instrumental analysis is based on the methodsdescribed elsewhere.12 Prior to extraction, the PUF/XAD-2sandwiches, GFFs, SIPs, and PUFs were spiked with 25 ng(mass-labeled FTOHs, FOSAs and FOSEs) and 5 ng (mass-labeled PFSAs and PFCAs) (absolute amount) of an IS mixturecontaining 16 mass-labeled PFASs (Table S2 in the SI).The high volume PUF/XAD-2 sandwiches were Soxhlet

extracted with petroleum ether/acetone (85/15, v/v) for ∼6 h,followed by a ∼16 h extraction with methanol. The GFFs wereextracted by sonication, three times with dichloromethane andthen three times with methanol. The low volume PUF/XAD-2

Environmental Science & Technology Article

dx.doi.org/10.1021/es4048945 | Environ. Sci. Technol. XXXX, XXX, XXX−XXXB

sandwiches, SIPs, and PUFs were extracted by using apressurized liquid extraction (PLE) system (ASE 350,Accelerated Solvent Extraction System from Dionex Corpo-ration, Sunnyvale, CA, USA). The extraction was carried outusing petroleum ether/acetone (83/17, v/v; 2 cycles) andthereafter acetonitrile (2 cycles) using the same ASE conditionsin either case as follows: 100 °C, 5 min static cycle with a 100%flush and 240 s purge.The petroleum ether/acetone and dichloromethane extracts

contain the more volatile PFASs (i.e., 6:2 FTMAC, FTACs,FTOHs, FOSAs, and FOSEs), and the methanol andacetonitrile extracts contain the PFCAs and PFSAs. Thepetroleum ether/acetone and dichloromethane extracts, andmethanol extracts were concentrated by rotary evaporationfollowed by gentle nitrogen blow-down to 0.5 using iso-octaneas a keeper solvent and 1 mL using methanol as a keepersolvent, respectively. Prior to injection, 10 ng absolute ofMe2FOSA was added to the iso-octane extracts, respectively,and 4 ng absolute of 13C8−PFOS and 13C8−PFOA were addedto the methanol extract in a polypropylene (PP) vial (CanadianLife Science, Peterborough, ON, Canada). The methanolextracts were further filtered using Mini-Uniprep PP filters (0.2μm pore size, Whatman, Piscataway, NJ, USA) and finallytransferred to PP vials (for details, see Figure S2 in the SI).The separation and detection of the 6:2 FTMAC, FTACs,

FTOHs, FOSAs, and FOSEs were performed using gaschromatography−mass spectrometry (Agilent 5975C; AgilentTechnologies, Palo Alto, CA, USA) (GC/MS) in selective ionmonitoring (SIM) mode using positive chemical ionization(PCI). Aliquots of 2 μL were injected on a DB-WAX column(30 m, 0.25 mm inner diameter, 0.25 μm film, J&W Scientific,Folsom, CA, USA). Analyses of PFCAs, PFSAs, and FOSAwere performed by liquid chromatography (Agilent 1100;Agilent Technologies, Palo Alto, CA, USA) using a triplequadrupole mass spectrometer interfaced with an electrosprayionization source in negative-ion mode (LC−(−)ESI−MS/MS;API 4000, Applied Biosystems/MDS SCIEX, Foster City, CA,USA). Aliquots of 25 μL were injected on a Luna C8(2) 100Acolumn (50 × 2 mm, 3 μm particle size; Phenomenex,Torrance, CA) using a gradient of 250 μL min−1 methanol andwater (both with 10 mM aqueous ammonium acetate solution(NH4OAc)) (for details, see Tables S7 and S8 in the SI).12

The isotope dilution method was used for quantification,which is based on the ratio of the peak-areas of the targetanalyte to the IS (Tables S7 and S8 in the SI). The blankconcentrations and limits of detection (LODs) (average of

blanks + 3× standard deviation (σ)) are given in Tables S9 andS10 in the SI. Average recoveries were 78%, 96%, 67%, 81%,and 93% for the LV-AAS, SIP-PAS, PUF-PAS, and gas phaseand particle using HV-AAS, respectively (Table S11 in the SI).Breakthrough experiments were conducted to check theefficiency of the PUF/XAD sandwich for trapping the gas-phase compounds using HV-AAS (n = 3, air volume ∼330 m3)and LV-AAS (n = 3, air volume ∼46 m3) (for details, see FigureS3 in the SI).

Theory of PAS. The uptake of chemicals by PUF disks andother PSMs has been shown to be controlled by the air-sidemass-transfer coefficient (kA), when the sampling medium has ahigh sorption capacity for the chemical.17 However, the uptakeof chemicals may be also influenced by sampler-side resistancewithin the PSM for some compounds.24 The uptake profile canbe described by the following equation:

= × × − − ×

×

−−

⎛⎝⎜⎜

⎡⎣⎢⎢⎛⎝⎜

⎞⎠⎟

⎤⎦⎥⎥⎞⎠⎟⎟

c K cAV

kK

t

1 expPSM PSM A APSM

PSM

A

PSM A

(1)

where cPSM is the concentration of chemical in the PSM (pgm−3), KPSM−A is the PSM-air partition coefficient, cA is the totalconcentrations of the target analyte in air measured by HV-AAS(pg m−3), APSM is the planar area of the passive sampler in cm2

(i.e., 370 cm2), VPSM is the volume of the PSM in cm3 (i.e., 210cm3), kA is the air-side mass-transfer coefficient (cm d−1), and tis the exposure time in days.The sample air volume is chemical specific and based on

KPSM−A for each chemical. The uptake profile of the chemical tothe PSM can be divided into three sections. Initially, the uptakeis linear because the amount in the PSM is small. As cPSMincreases over time, the term cPSM/KPSM−A becomes moreimportant and the uptake is reduced and becomes curvilinear,and finally, cPSM reaches an equilibrium plateau (equal fugacity).In addition to depending on the properties of the PSM, theuptake profile also depends on the chamber housing design andmeteorological factors such as temperature and wind speed.18

Further details for calibration of PAS are described else-where.17,25

Figure 1. Uptake profiles of PFASs for (A) PUF-PAS and (B) SIP-PAS.

Environmental Science & Technology Article

dx.doi.org/10.1021/es4048945 | Environ. Sci. Technol. XXXX, XXX, XXX−XXXC

■ RESULTS AND DISCUSSION

Uptake Profile for PUF-PAS and SIP-PAS for PFASs.The equivalent air volume for a passive air sampler is a measureof the amount of air that it has sampled after a given exposureperiod. It can also be regarded as equivalent to the samplingrate of the passive sampler (R, m3 d−1) × the number of days ofexposure assuming a linear uptake during the deploymentperiod. It can be calculated by dividing the amount of chemicalin the PSM (cPSM, pg sample−1) by the total concentrations ofthe target analyte in ambient air using the HV-AAS (cA, pgm−3).17 The uptake profiles of PFASs for PUF-PAS and SIP-PAS over the deployment time are given in Figure 1 andFigures S4 and S5 in the SI. The HV-AAS concentrations(average concentration over one month) were used to calculatethe uptake profiles for the PAS because of lower detectionfrequency of PFASs in the LV-AAS (see Table S12 in the SI).For the PUF-PAS, Me- and EtFOSA, and Me- and EtFOSE

had short linear uptake curves (<28 days) and equilibrated aftera few weeks. Although Me- and EtFOSA equilibrated faster(i.e., equilibration after ∼56 days) in comparison to the Me-and EtFOSE (i.e., equilibration after ∼120 days), the equivalentair volume at equilibrium was similar for both PFAS classes in arange of 60−70 m3, reflecting similar KPUF‑A values for thesetwo compound classes. In contrast, the PFSAs constantlyincreased and the uptake profile was still in the curvilinearphase until the end of the deployment period of 168 days (thePUF disks deployed over 197 days were not analyzed forPFSAs and PFCAs). It is important to note that 8:2 and 10:2FTOH, 6:2 FTMAC, PFCAs, FTACs, and FOSA were notdetected in the PUF-PAS which shows that PUF-PAS have avery low sorptive capacity for these compounds. A low uptakecapacity for the FTOHs in PUF disks was also observed in aprevious indoor uptake study with equilibration of FTOHs aftera few days resulting in an effective air volume of only a fewcubic meters.16

For the SIP-PAS, almost all analyzed PFAS classes (i.e.,PFSAs, PFCAs, FTOHs, FOSAs, and FOSEs) were detected,except the 6:2 FTMAC and FTACs (see Table S12 in the SI).The uptake for the PFCAs and FTOHs was greatly improvedusing SIP-PAS compared to the PUF-PAS for which no PFCAsand FTOHs were detected. Furthermore, greater sorptioncapacities were also observed for the FOSAs and FOSEsshowing longer linear uptake curves for these compounds inSIP-PAS (>56 days) in comparison to the PUF-PAS (<28days). In contrast, the uptake profile for the PFSAs was similarfor the PUF-PAS and SIP-PAS with equivalent air volume of∼300 m3 after 160 days of deployment (Figure S6 in the SI).

Time required for equilibrium to be established varieddepending on the chain length for the PFCAs. For example, thelonger chained C12−C16 PFCAs had a longer linear orcurvilinear phase compared to the C4−C11 PFCAs (Figure 1).A similar trend was previously observed for the uptake ofpolychlorinated biphenyls (PCBs) in PUF-PAS showing alonger linear uptake phase for the higher molecular weightPCBs (i.e., penta- to heptachlorinated biphenyls).19

As mentioned previously, the uptake profile is chemicaldependent and also influenced by variability of the chemicalconcentration during the deployment period and by meteoro-logical conditions like the wind speed and temperature that alsovary with time.17−19,26 The average wind speed was relativelyconstant throughout the deployment period with a mean ofabout 16 km h−1 and did not show a temporal bias, whereas thetemperature was lower at the beginning and in the end of thestudy (on average ∼14 °C) and higher in summer during themiddle of the uptake study (on average ∼22 °C) (Tables S5and S6 in the SI). However, no significant correlation wasobserved between the sampling rate and the range of windspeeds, temperature, and other meteorological parameters (i.e.,relative humidity, wind direction, and amount of rain) for thestudy (p > 0.05, Pearson Correlation).

Sampling Rates for PUF-PAS and SIP-PAS for PFASs.The sampling rate (R, m3 d−1) was derived from the linearuptake phase of the uptake profiles, by taking the slope of theplot of cPSM/cA versus time. As a general rule for estimating R-values, ideally there should be at least 3 data points within thelinear region, which we define here as the time up to 25% ofequilibrium (t25). Inclusion of data points in the curvilinearregion, defined here as the time in the range of 25%−90% ofequilibrium (t25− t90), will result in underestimates of R.The KPSM−A at equilibrium can be described as the volume of

ambient air (VAIR) that contains an equivalent amount ofchemical contained in a PSM having a volume (VPSM). It is alsothe ratio of the concentration of the chemical in the PSM(cPSM) divided by the concentration of the target analyte in air(cA) when the system is at equilibrium.17

= =−KVV

ccPSM A

AIR

PSM

PSM

A (2)

The KPSM−A for a chemical in the PUF-PAS and SIP-PAS(i.e., KPUF−A and KSIP−A, respectively) can be determined fromits concentration in the PSM when it has reached equilibriumrelative to air. This is reflected by a flattening of the uptakeprofile with time. However, the PFSAs in the PUF-PAS andSIP-PAS and the C12−16 PFCAs, FTOHs, FOSAs, and FOSEsin the SIP-PAS did not equilibrate, and therefore, a minimum

Table 1. Calibration Results for PUF-PASa

cA (pg m−3) cPUF (pg disk−1) cPUF (pg m−3 disk) KPUF‑A and QPUF‑Ab log KPUF‑A and log QPUF‑A

b kA (m d−1) R (m−3 d−1)

PFBS 0.24 ± 0.18 62 2.94 × 105 >1.22 × 106 b >6.09b 48 1.8PFHxS 0.13 ± 0.11 21 9.84 × 104 >7.39 × 105 b >5.87b ncc ncc

PFOS 0.96 ± 0.46 287 1.37 × 106 >1.43 × 106 b >6.16b 55 2.0PFDS 0.09 ± 0.08 41 1.98 × 105 >2.23 × 106 b >6.35b 70 2.6MeFOSA 1.24 ± 0.81 86 4.12 × 105 3.31 × 105 5.52 ncc ncc

EtFOSA 0.89 ± 0.58 47 2.15 × 105 2.53 × 105 5.40 ncc ncc

MeFOSE 3.37 ± 2.07 134 6.38 × 105 1.90 × 105 5.28 ncc ncc

EtFOSE 1.77 ± 1.16 208 9.90 × 105 5.58 × 105 5.75 ncc ncc

aPFCAs, FTOHs, 6:2 FTMAC, FTACs, and FOSA were not detected in the PUF-PAS and are therefore not shown. bAverage temperature at 18 °Cbetween 20/07/2010 and 13/10/2010. For some analytes, the PUF-PAS did not reach equilibrium by the end of the 197 day uptake study (168 daysfor PFOS) so the lower limits of the partition coefficients were calculated as QPUF‑A.

cnc = not calculable, because of insufficient data points.

Environmental Science & Technology Article

dx.doi.org/10.1021/es4048945 | Environ. Sci. Technol. XXXX, XXX, XXX−XXXD

value for the partition coefficient is estimated for thesecompounds (in this case, QPSM−A is used instead of KPSM−A,to indicate that it is not a true partition coefficient) (see Tables1 and 2). It is important to note, that KPSM−A can increase by afactor of 2.5−3.0 with every 10 °C decrease in temperature.17

This means longer linear phases for PAS when operating atcolder temperatures.For the PUF-PAS, the sampling rates for the PFSAs (except

PFHxS) ranged between 1.8 and 2.6 m3 d−1 (PFCAs were notdetected in the PUF-PAS). The sampling rates for the FOSAs,FOSEs, and PFHxS could not be reported due to rapidequilibration of these compounds in the PUF-PAS and lack ofsufficient data points to allow for a reliable estimate of thelinear sampling rate. For the SIP-PAS, the sampling rates forFTOHs (3.3−4.3 m3 d−1), FOSAs (4.2−4.4 m3 d−1), andFOSEs (2.9 m3 d−1) in this study are in reasonable agreementwith previous reported indoor derived sampling rates (i.e., 4.6m3 d−1 for FTOHs, 2.6 for FOSAs, and 1.4−1.5 m3 d−1forFOSEs).16 Differences can be explained by different environ-mental conditions (outdoor vs indoor). PFSAs were similarcompared to the PUF-PAS, ranging from 1.8 to 2.5 m3 d−1. Thesampling rates for PFCAs were, on average, 4.2 m3 d−1 for theSIP-PAS.The uptake profile for PAS can be described by the uptake

constant kU (day−1).17

= ×−

kAV

kKU

PSM

PSM

A

PSM A (3)

The kU can be used to calculate the extent of the linearuptake phase as t25 = ln(0.75)/kU (i.e., time when the PSM hasaccumulated 25% of the equilibrium value) and the time of 95%of equilibrium value as t95 = ln(0.05)/kU. These values that

define the uptake profile can be used to make decisionsregarding ideal deployment times for particular chemicals orwhen groups of chemicals are being investigated. The estimatedt25 ranged between a few weeks and three months, whereas theestimated t95 ranged between several months and two and a halfyears for both PAS types (Table S13 in the SI). Generally, theestimated times to equilibrate for the SIP-PAS were a factor of∼2 higher compared to the PUF-PAS. This is a relatively smalldifference when compared to other compound classes such asPCBs and organochlorine pesticides (OCPs) where capacity ofthe SIP-PAS was up to 2 orders of magnitude greater comparedto PUF-PAS.14,27,28

This finding suggests that the PFASs partition differently toPUF-PAS and SIP-PAS compared to semivolatile organiccompounds like PCBs and OCPs, which sorb much less tosurfaces due to their nonpolar hydrophobic characteristics. Thisis consistent with observations from particle-gas partitioninginvestigations of the PFASs, that indicate that they do not obeythe typical KOA-based and subcooled liquid vapor pressure (pL

o)-based relationships that have been derived for nonpolarhydrophobic chemicals.22 However, more work is required tofurther investigate the sorption mechanism of PFASs.The correlation of log KPSM−A (or log QPSM‑A) against log

KOA for PFASs in PUF-PAS and SIP-PAS was investigated (seeSI), and the results are shown in Table S14 and Figure S7 inthe SI. Overall, only a weak correlation was observed indicatingthat the PFASs do not undergo KOA-driven partitioning intoPSM. Thus, there might be other factors influencing thesorption of PFASs to PSM, for example, amount of particles/aqueous aerosols sampled by the PAS and influence of thehydrophobic fluorocarbon chain length and hydrophilicfunctional groups of PFASs.11,22,29

Table 2. Calibration Results for SIP-PASa

cA (pg m−3) cSIP (pg disk−1) cSIP (pg m−3 disk) KSIP‑A and QPUF‑Ab log KSIP‑A and log QPUF‑A

b kA (m d−1) R (m−3 d−1)

PFBS 0.24 ± 0.18 76 3.60 × 105 >1.50 × 106 b >6.18b 53 2.0PFHxS 0.13 ± 0.11 35 1.69 × 105 >1.26 × 106 b >6.10b 55 2.0PFOS 0.96 ± 0.46 511 2.44 × 106 >2.55 × 106 b >6.41b 68 2.5PFDS 0.09 ± 0.08 20 9.49 × 104 >1.07 × 106 b >6.03b 50 1.8PFBA 4.80 ± 3.09 3268 1.56 × 107 3.25 × 106 6.51 110 4.1PFPeA 1.28 ± 0.93 448 2.14 × 106 1.67 × 106 6.22 91 3.4PFHxA 0.45 ± 0.33 446 2.13 × 106 4.68 × 106 6.67 135 4.2PFHpA nd 291 1.39 × 106

PFOA 1.71 ± 1.15 572 2.73 × 106 1.60 × 106 6.20 104 3.8PFNA 0.50 ± 0.35 345 1.65 × 106 3.32 × 106 6.52 96 3.6PFDA 0.31 ± 0.27 215 1.02 × 106 3.28 × 106 6.52 105 3.9PFUnDA 0.54 ± 0.85 244 1.17 × 106 2.17 × 106 6.36 100 3.7PFDoDA 0.09 ± 0.06 77 3.69 × 105 >3.94 × 106 b >6.60b 110 4.1PFTrDA 0.06 ± 0.04 102 4.86 × 105 >8.16 × 106 b >6.91b 149 5.5PFTeDA 0.03 ± 0.02 45 2.15 × 105 >6.27 × 106 b >6.80b 142 5.3PFPeDA 0.02 ± 0.02 37 1.75 × 105 >7.12 × 106 b >6.85b 125 4.6PFHxDA 0.03 ± 0.02 29 1.38 × 105 >5.33 × 106 b >6.73b 120 4.4PFODA nd 20 9.50 × 104

8:2 FTOH 53.5 ± 23.8 27024 1.29 × 108 >2.41 × 106 b >6.38b 93 3.410:2 FTOH 21.7 ± 11.9 7838 3.74 × 107 >1.72 × 106 b >6.24b 89 3.3MeFOSA 1.24 ± 0.81 825 3.94 × 106 >3.17 × 106 b >6.50b 118 4.4EtFOSA 0.89 ± 0.58 642 3.06 × 106 >3.45 × 106 b >6.54b 113 4.2MeFOSE 3.37 ± 2.07 1493 7.13 × 106 >2.12 × 106 b >6.33b 79 2.9EtFOSE 1.77 ± 1.16 840 4.01 × 106 >2.26 × 106 b >6.35b 77 2.9

and = not detected. 6:2 FTMAC, FTACs, and FOSA were not detected in the SIP-PAS and are therefore not shown. bAverage temperature at 18 °Cbetween 20/07/2010 and 13/10/2010. For some analytes, the SIP-PAS did not reach equilibrium by the end of the 197 day uptake study so thelower limits of the partition coefficients were calculated as QSIP‑A.

Environmental Science & Technology Article

dx.doi.org/10.1021/es4048945 | Environ. Sci. Technol. XXXX, XXX, XXX−XXXE

Loss of Depuration Compounds. The use of various DCswas explored to cover the range of PFASs. DCs are useful forcalculating site-specific R-values under different meteorologicalconditions. The R-values (m3 d−1) can be calculated bymultiplying the kA value, which was derived from the loss of theDCs, by the surface area of the PUF-PAS (for details, seeelsewhere15,20). Conditions under which DCs are selected andapplied include:

(i) DCs should not be present in ambient air.(ii) DCs should belong to the same compound class as the

target analytes.(iii) DCs should not degrade in the PSM during the

deployment period. Losses should only be due tovolatilization to air (i.e., air-side mass transfer).

(iii) Target losses for DCs during the deployment period is inthe range of >40% to up to 90%. This ensures that lossesare large enough to be distinguished from analyticalvariability.

(iv) Several DCs should be used so that an average samplingrate can be determined. This also helps to reducevariability associated with analysis of individual com-pounds.20

In Figure 2, the loss of the DCs (i.e., 13C8−PFOS, 13C8−PFOA, and 7:2 sFTOH) is shown during the deploymentperiod. For 13C8−PFOS, no substantial loss was observed forthe PUF-PAS and SIP-PAS during the deployment period of197 days. This demonstrates the low volatility and highadsorption and stability of 13C8−PFOS (and therefore also forthe native PFOS) to both PAS media. The high adsorptionpotential of PFOS is also observed by its strong sorption toairborne particles (i.e., 46%).12 Given the negligible loss of 13C8PFOS from the PAS, this chemical is better suited as a qualitycontrol/recovery surrogate for the DCs. In contrast, the 13C8−PFOA concentration decreased linearly for both PAS typesduring the deployment period. Overall, the loss of 13C8−PFOAwas ∼85% for the PUF-PAS and ∼45% for the SIP-PAS after168 days. For 7:2 sFTOH, a different adsorption behavior wasobserved between the PUF-PAS and SIP-PAS. For the SIP-PAS, 7:2 sFTOH had a linear loss of ∼17% during thedeployment period of 197 days showing the high sorptioncapacity of SIP-PAS for more volatile chemicals like 7:2sFTOH. In contrast, 7:2 sFTOH was only found in the spikedblank PUF disk samples but was not detected in any deployedPUF-PAS which means that 7:2 sFTOH was completelyvolatilized within seven days of deployment which representsthe time when the first uptake sample was collected. This is inagreement with the uptake results for the FTOHs, which have a

similar structure to the sFTOHs, showing that they have a verylimited sorption capacity in PUF-PAS.Overall, of the DCs tested in this study, only 13C8 PFOA was

suitable for the calculation of the site specific sampling rates.The R-value of 13C8 PFOA was determined to be 3.7 ± 0.9 m3

d−1 for the SIP-PAS. This is in agreement with the R-valuecalculated for native compounds in this uptake study (see Table2). This is in accordance to a previous study showingcomparable R-values determined from time integrated activesampling and the DC approach.30 More work is required tounderstand the influence of the PSM-side kinetic resistance onthe loss of the DC and identify additional PFASs that aresuitable as DCs.

Implications for the Calculation of the EquivalentSampling Air Volume. The approach for calculating theequivalent sampling air volume (VAIR) for the PAS wasdescribed previously17 (for the uptake parameters, see Tables1 and 2).

= × × − − ×

×

−−

⎛⎝⎜⎜

⎡⎣⎢⎢⎛⎝⎜

⎞⎠⎟

⎤⎦⎥⎥⎞⎠⎟⎟

V K VAV

kK

t

1 expA

AIR PSM A PSMPSM

PSM

A

PSM

(4)

For analytes which are still in the linear phase, VAIR cansimply be calculated by multiplying the R-value of the analytewith the days of deployment.

= ×V R tAIR (5)

These same expressions are applied to both gas- and particle-phase compounds.31−33 In order to investigate the robustnessof the passive sampling chamber design, different configurationswere tested by varying the gap (opening) between the upperand lower domes. A larger opening between the upper andlower domes could result in increased sampling of gas-phasecompounds if the wind effect is important, over the range ofwind speeds at the sampling site.34 The larger gaps should alsoallow for unimpeded movement of particles into the sampler. Ifparticle sampling is somehow reduced by the conventionalchamber design, we should see greater sampling of particle-bound PFASs in the larger gap configurations.18 The resultsindicate that the concentration difference for PFSAs, FOSAs,and FOSEs that are associated with particles12 and for otherPFASs that are mainly in the gas phase was not significantbetween the different chambers for SIP-PAS and PUF-PAS (p >0.05, Student’s t-test) (Figures S8 and S9 in the SI).

Figure 2. Loss of DCs during deployment as fraction of the starting amount. The concentration of the chemical in the PSM during the deploymenttime (c) is divided by the concentration of the chemical at deployment time t = 0 (c0).

Environmental Science & Technology Article

dx.doi.org/10.1021/es4048945 | Environ. Sci. Technol. XXXX, XXX, XXX−XXXF

In this study, all PFAS (except FOSAs and FOSEs in thePUF-PAS) showed a lengthy linear/curvilinear uptake phasewith an average linear-phase R-value ranging from 1.8−5.5 m3

d−1 (on average, 3.5 m3 d−1) for PUF-PAS and SIP-PAS andthe R-value derived from the DC 13C8 PFOA was determinedto be 3.7 m3 d−1 for the SIP-PAS. These sampling rates are veryclose to the suggested R-value of 4 m3 d−1 reported previouslyfor the classical POPs indicating a similar uptake characteristicfor POPs and PFASs.15 Minor variations in derived samplingrates, between compounds classes, are likely due to analytical orexperimental variability. Thus, to simplify the assessment ofequivalent sample air volumes, eq 5 was applied using acommon linear sampling rate of 4 m3 d−1 (with the exception ofFOSAs and FOSEs in the PUF-PAS, see below). This results ina VAIR of 112 m3 for a one month deployment period. Thesample volumes for FOSAs and FOSEs in the PUF-PAS werecalculated using eq 4 to account for their approach toequilibrium and reduced sample air volumes. VAIR rangedfrom 39 to 72 m3 for the FOSAs and FOSEs in PUF-PAS.Ultimately, the calculated VAIR can be used to calculate theconcentration of the analyte in air (cA) by dividing cPSM (pgdisk−1) with VAIR (m3).

=c c V/A PSM AIR (6)

Overall, VAIR can be calculated by applying a R-value of 4 m3

d−1 for all PFASs except for the FOSAs and FOSEs in the PUF-PAS for which the full uptake expression of eq 4 has been used.It is interesting to note that there was no significant correlationof the R-value for the individual analytes (see Tables 1 and 2)with the particle associated fraction of the analyte in air12 (p >0.05, Pearson Correlation). This indicates that the SIP-PAScapture gas-phase and particle-phase PFASs with similarefficiency, consistent with the recent findings by Harner et al.33

Comparison of Four Different Sampling Techniques.PFAS were measured in air using four different samplingtechniques: (i) HV-AAS to measure gas and particle phaseseparately, (ii) LV-AAS comprising the sum of the gas andparticle phase, (iii) SIP-PAS, and (iv) PUF-PAS. In general, theaverages agree generally within a factor of 2 and no significantdifferences were found for the PFAS concentrations measuredby the PUF-PAS, SIP-PAS, LV-AAS, and HV-AAS (p > 0.05,Kruskal−Wallis test) (Figure S10 in the SI).The performance of the PUF-PAS and SIP-PAS for

measuring FOSAs/FOSEs and PFSAs in the atmosphere wascompared using linear regression. Both the FOSA/FOSE andPFSA concentrations were generally within a factor of 2 for thetwo PAS types (r2 = 0.66 and r2 = 0.98, respectively) (FigureS11 in the SI).The air concentration of PFASs measured by the HV-AAS

(representing 14−29% of the time for the monthly average)was compared with the air concentration derived by the SIP-PAS and PUF-PAS (which sample 100% of the time) usinglinear regression (Figure 3). Generally, the SIP-PAS showed agood agreement with the air concentration determined by theHV-AAS for all PFAS classes (r2 > 0.9). The average differencebetween the two sampling techniques was less than 50% forindividual PFASs. For the PUF-PAS, the FOSA/FOSEconcentrations showed a higher scattering of the data (r2 =0.76) but the linear regressions with the HV-AAS measure-ments were close to unity. The higher scattering of the FOSAand FOSE concentrations can be explained by loweraccumulation of these compounds in the PUF-PAS due to alimited uptake capacity (compared to the SIP-PAS). Con-sequently, these lower concentrations in the PUF-PAS mayapproach detection limits for these compounds, which results inderived air concentrations that have greater analytical

Figure 3. Comparison between total air concentrations (gas and particle phase) using HV-AAS and concentrations derived by PUF-PAS and SIP-PAS in pg m−3 using linear regression. Each dot represents average concentration over one month for HV-AAS and integrated concentration overone month for PAS for individual PFAS.

Environmental Science & Technology Article

dx.doi.org/10.1021/es4048945 | Environ. Sci. Technol. XXXX, XXX, XXX−XXXG

uncertainty. In contrast, the PFSA concentrations derived fromthe SIP-PAS and PUF-PAS showed a good linear regressionwith r2 of 0.91 and 0.90, respectively (p < 0.05, PearsonCorrelation). However, the concentrations were lowercompared to the total air concentration measured by the HV-AAS, which can be due to the high association of PFSAs to theparticle phase22 which might be less efficiently collected by thePAS. The PFCA concentrations derived from the SIP-PAS alsoshowed a good linear regression with r2 of 0.92 (p < 0.05,Pearson Correlation), but the concentrations were slightlyhigher indicating a higher collection efficiency of PFCAs bySIP-PAS compared to HV-AAS. However, the PFOAconcentrations derived from the SIP-PAS were generallylower compared to HV-AAS which might be due toconcentrations close to the detection limit for SIP-PAS (seeFigure 4). Overall, the difference for individual PFASs waswithin a factor of 2 using PUF-PAS and SIP-PAS which can beconsidered to be good agreement, especially considering thatsome variability is expected, due to the HV-AAS not operating100% of the time.Atmospheric Composition and Seasonal Trends of

PFASs. Overall, all of the 29 targeted PFASs were detected inair samples (Table S12 in the SI). The most abundant PFASclass for the total air concentration (sum of gas and particlephase measured by HV-AAS) was the FTOHs representing onaverage ∼80% of the ΣPFASs, followed by PFCAs (∼7%) andFTACs/fluorotelomer methacrylates (FTMACs) (∼7%) (Fig-ure S12 in the SI). The other PFAS classes represented lessthan 3% of the ΣPFASs. Total air concentrations (HV-AAS) forΣFTOHs ranged from 20−182 pg m−3 with 8:2 FTOH as the

dominant compound (48% of the ΣFTOHs). The total ΣPFCAconcentration ranged from 0.7 to 20 pg m−3 with PFBA as thedominant compound (∼54% of the ΣPFCA) and with atendency of decreasing air concentrations for the longer chainPFCAs. The C4-based PFASs (e.g., PFBA) are the mainreplacement compound of the voluntary phase-out C8-basedproducts (i.e., PFOA and PFOS) which may explain theelevated concentration of PFBA in air in Toronto.2,3 It isinteresting to note that the average ΣFTAC concentrations inthis study were higher (8.3 pg m−3) compared to the ΣFOSAsand ΣFOSEs (1.4 and 3.4 pg m−3, respectively). Thisdemonstrates the importance of FTACs as the next mostrelevant precursor class after the FTOHs. However, a recentstudy of PFASs in the Asian atmosphere showed that 8:2fluorotelomer olefin (FTO) is the second most abundant PFASclass after the FTOHs.35 The ΣFOSA and ΣFOSE concen-trations were about 5 times lower than previously reported forsuburban or urban areas5,36 which might be due to the phaseout of perfluorooctyl sulfonyl fluoride (POSF), reduced PFASemissions by optimization of the production process,2 or aproduction shift to shorter chain PFASs and new fluorinatedchemicals.37,38 Lower total air concentrations were observed forΣPFSAs (on average 1.0 pg m−3) with PFOS as thepredominant compound in this PFAS class (74% of theΣPFSAs). This is in agreement with recent measurement in theToronto atmosphere.39,40

The air concentrations of individual PFASs in Toronto werecompared over one year, using the four different samplingapproaches used in this study (see Figure 4 and Figure S13 inthe SI). For the PFASs in LV-AAS and the PFSAs and PFCAs

Figure 4. PFOS, PFOA, 8:2 FTOH, and MeFOSE concentrations in air measured by four different sampling techniques over one year: HV-AAS(sum of gas and particle phase; integrated over 24 h (black bars) and average concentration over one month) and LV-AAS (sum of gas and particlephase; integrated over 14 days), and SIP-PAS and PUF-PAS (integrated over one month).

Environmental Science & Technology Article

dx.doi.org/10.1021/es4048945 | Environ. Sci. Technol. XXXX, XXX, XXX−XXXH

in PUF-PAS, results are presented for 29 consecutive weeksfrom the start of the study. The individual PFAS concentrationsbased on HV-AAS varied over time by up to a factor of 5. Thisvariability is lower than previously reported at a site close toHamburg, Germany.41 In particular, we did not see peak eventswith extremely high PFAS concentrations (i.e., 1 order ofmagnitude higher than baseline levels) as reported previously.41

The main factor governing the variability in air concen-trations of PFASs over time proved to be temperature. Themajority of PFASs classes measured by HV-AAS wassignificantly correlated with ambient temperature (p < 0.05,Pearson Correlation) (for details, see Table S15 in the SI). Thissuggests an important influence from local/regional sourcesthat exhibit seasonality which may be partly attributed totemperature (i.e., enhanced volatilization). Generally, the PFASconcentrations decreased in the order of summer, spring, fall,and winter. Potential emission sources in Toronto for PFASsinclude inter alia WWTPs and landfills, which are consideredpoint sources,42 and residential homes, which can beconsidered as diffuse sources.6 Ultimately, all four samplingapproaches (i.e., HV-AAS, LV-AAS, SIP-PAS, and PUF-PAS)are deemed suitable for capturing temporal trends of PFASs inair.

■ ASSOCIATED CONTENT*S Supporting InformationAdditional details on sampling sites, meteorological data, QA/QC data, PFAS concentrations, and predicted log KOA forindividual PFAS. This material is available free of charge via theInternet at http://pubs.acs.org.

■ AUTHOR INFORMATIONCorresponding Authors*E-mail: lutz.ahrens@slu.se; phone: +46 70-2972245; fax: +4670-2972245.*E-mail: tom.harner@ec.gc.ca; phone: +1 416-739-4837; fax:+1 416-739-4281.NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSPartial funds for this work were provided through theChemicals Management Plan (Government of Canada), theChemicals Management Division (Environment Canada), andthe United Nations Environment Programme (UNEP).

■ REFERENCES(1) Houde, M.; Martin, J. W.; Letcher, R. J.; Solomon, K. R.; Muir, D.C. G. Biological monitoring of polyfluoroalkyl substances: A review.Environ. Sci. Technol. 2006, 40, 3463−3473.(2) Paul, A. G.; Jones, K. C.; Sweetman, A. J. A first globalproduction, emission, and environmental inventory for perfluorooc-tane sulfonate. Environ. Sci. Technol. 2009, 43, 386−392.(3) Prevedouros, K.; Cousins, I. T.; Buck, R. C.; Korzeniowski, S. H.Sources, fate and transport of perfluorocarboxylates. Environ. Sci.Technol. 2006, 40, 32−44.(4) Ahrens, L. Polyfluoroalkyl compounds in the aquatic environ-ment: A review of their occurrence and fate. J. Environ. Monit. 2011,13, 20−31.(5) Shoeib, M.; Harner, T.; Vlahos, P. Perfluorinated chemicals in theArctic atmosphere. Environ. Sci. Technol. 2006, 40, 7577−7583.(6) Shoeib, M.; Harner, T.; Webster, G. M.; Lee, S. C. Indoor sourcesof poly- and perfluorinated compounds (PFCS) in Vancouver,

Canada: Implications for human exposure. Environ. Sci. Technol.2011, 45, 7999−8005.(7) Ahrens, L.; Shoeib, M.; Harner, T.; Lee, S. C.; Guo, R.; Reiner, E.J. Wastewater treatment plant and landfills as sources of polyfluor-oalkyl compounds to the atmosphere. Environ. Sci. Technol. 2011, 45,8098−8105.(8) Wang, Z.; Scheringer, M.; MacLeod, M.; Bogdal, C.; Muller, C.E.; Gerecke, A. C.; Hungerbuhler, K. Atmospheric fate of poly- andperfluorinated alkyl substances (PFASs): II. Emission source strengthin summer in Zurich, Switzerland. Environ. Pollut. 2012, 169, 204−209.(9) Muller, C. E.; Gerecke, A. C.; Bogdal, C.; Wang, Z.; Scheringer,M.; Hungerbuhler, K. Atmospheric fate of poly- and perfluorinatedalkyl substances (PFASs): I. Day-night patterns of air concentrationsin summer in Zurich, Switzerland. Environ. Pollut. 2012, 169, 196−203.(10) Barber, J. L.; Berger, U.; Chaemfa, C.; Huber, S.; Jahnke, A.;Temme, C.; Jones, K. C. Analysis of per- and polyfluorinated alkylsubstances in air samples from Northwest Europe. J. Environ. Monit.2007, 9, 530−541.(11) Arp, H. P. H.; Goss, K. U. Irreversible sorption of traceconcentrations of perfluorocarboxylic acids to fiber filters used for airsampling. Atmos. Environ. 2008, 42, 6869−6872.(12) Ahrens, L.; Shoeib, M.; Harner, T.; Lane, D. A.; Guo, R.; Reiner,E. J. Comparison of annular diffusion denuder and high volume airsamplers for measuring per- and polyfluoroalkyl substances in theatmosphere. Anal. Chem. 2011, 83, 9622−9628.(13) Pozo, K.; Harner, T.; Wania, F.; Muir, D. C. G.; Jones, K. C.;Barrie, L. A. Toward a global network for persistent organic pollutantsin air: Results from the GAPS Study. Environ. Sci. Technol. 2006, 40,4867−4873.(14) Genualdi, S.; Lee, S. C.; Shoeib, M.; Gawor, A.; Ahrens, L.;Harner, T. Global pilot study of legacy and emerging persistentorganic pollutants using sorbent-impregnated polyurethane foam diskpassive air samplers. Environ. Sci. Technol. 2010, 44, 5534−5539.(15) Pozo, K.; Harner, T.; Lee, S. C.; Wania, F.; Muir, D. G.; Jones,K. C. Seasonally resolved concentrations of persistent organicpollutants in the global atmosphere from the first year of the GAPSstudy. Environ. Sci. Technol. 2009, 43, 796−803.(16) Shoeib, M.; Harner, T.; Lee, S. C.; Lane, D.; Zhu, J. Sorbent-impregnated polyurethane foam disk for passive air sampling ofvolatile fluorinated chemicals. Anal. Chem. 2008, 80, 675−682.(17) Shoeib, M.; Harner, T. Characterization and comparison ofthree passive air samplers for persistent organic pollutants. Environ. Sci.Technol. 2002, 36, 4142−4151.(18) Klanova, J.; Eupr, P.; Kohoutek, J.; Harner, T. Assessing theinfluence of meteorological parameters on the performance ofpolyurethane foam-based passive air samplers. Environ. Sci. Technol.2008, 42, 550−555.(19) Chaemfa, C.; Barber, J. L.; Gocht, T.; Harner, T.; Holoubek, I.;Klanova, J.; Jones, K. C. Field calibration of polyurethane foam (PUF)disk passive air samplers for PCBs and OC pesticides. Environ. Pollut.2008, 156, 1290−1297.(20) Moeckel, C.; Harner, T.; Nizzetto, L.; Strandberg, B.; Lindroth,A.; Jones, K. C. Use of depuration compounds in passive air samplers:Results from active sampling-supported field deployment, potentialuses, and recommendations. Environ. Sci. Technol. 2009, 43, 3227−3232.(21) Zhang, X.; Tsurukawa, M.; Nakano, T.; Lei, Y. D.; Wania, F.Sampling medium side resistance to uptake of semivolatile organiccompounds in passive air samplers. Environ. Sci. Technol. 2011, 45,10509−10515.(22) Ahrens, L.; Harner, T.; Shoeib, M.; Lane, D. A.; Murphy, J. G.Improved characterization of gas-particle partitioning for per- andpolyfluoroalkyl substances in the atmosphere using annular diffusiondenuder samplers. Environ. Sci. Technol. 2012, 46, 7199−7206.(23) United Nations Environment Programme (UNEP), Genevahosts Stockholm Convention on Persistent Organic Pollutants from 4to 8 May, 2009, from http://www.unep.org/newscentre/ (22/09/2010).

Environmental Science & Technology Article

dx.doi.org/10.1021/es4048945 | Environ. Sci. Technol. XXXX, XXX, XXX−XXXI

(24) Zhang, X.; Wania, F. Modeling the uptake of semivolatileorganic compounds by passive air samplers: Importance of masstransfer processes within the porous sampling media. Environ. Sci.Technol. 2012, 46, 9563−9570.(25) Bartkow, M. E.; Booij, K.; Kennedy, K. E.; Muller, J. F.; Hawker,D. W. Passive air sampling theory for semivolatile organic compounds.Chemosphere 2005, 60, 170−176.(26) Petrich, N. T.; Spak, S. N.; Carmichael, G. R.; Hu, D.; Martinez,A.; Hornbuckle, K. C. Simulating and explaining passive air samplingrates for semivolatile compounds on polyurethane foam passivesamplers. Environ. Sci. Technol. 2013, 47, 8591−8598.(27) Koblizkova, M.; Genualdi, S.; Lee, S. C.; Harner, T. Applicationof sorbent impregnated polyurethane foam (SIP) disk passive airsamplers for investigating organochlorine pesticides and polybromi-nated diphenyl ethers at the global scale. Environ. Sci. Technol. 2012,46, 391−396.(28) Hayward, S. J.; Lei, Y. D.; Wania, F. Sorption of a diverse set oforganic chemical vapors onto XAD-2 resin: Measurement, predictionand implications for air sampling. Atmos. Environ. 2011, 45, 296−302.(29) McMurdo, C. J.; Ellis, D. A.; Webster, E.; Butler, J.; Christensen,R. D.; Reid, L. K. Aerosol enrichment of the surfactant PFO andmediation of the water-air transport of gaseous PFOA. Environ. Sci.Technol. 2008, 42, 3969−3974.(30) He, J.; Balasubramanian, R. A comparative evaluation of passiveand active samplers for measurements of gaseous semi-volatile organiccompounds in the tropical atmosphere. Atmos. Environ. 2010, 44,884−891.(31) Bohlin, P.; Jones, K. C.; Strandberg, B. Field evaluation ofpolyurethane foam passive air samplers to assess airborne PAHs inoccupational environments. Environ. Sci. Technol. 2010, 44, 749−754.(32) Chaemfa, C.; Wild, E.; Davison, B.; Barber, J. L.; Jones, K. C. Astudy of aerosol entrapment and the influence of wind speed, chamberdesign and foam density on polyurethane foam passive air samplersused for persistent organic pollutants. J. Environ. Monit. 2009, 11,1135−1139.(33) Harner, T.; Su, K.; Genualdi, S.; Karpowicz, J.; Ahrens, L.;Mihele, C.; Schuster, J.; Charland, J.-P.; Narayan, J. Calibration andapplication of PUF disk passive air samplers for tracking polycyclicaromatic compounds (PACs). Atmos. Environ. 2013, 75, 123−128.(34) Tuduri, L.; Harner, T.; Hung, H. Polyurethane foam (PUF)disks passive air samplers: Wind effect on sampling rates. Environ.Pollut. 2006, 144, 377−383.(35) Li, J.; Del Vento, S.; Schuster, J.; Zhang, G.; Chakraborty, P.;Kobara, Y.; Jones, K. C. Perfluorinated compounds in the Asianatmosphere. Environ. Sci. Technol. 2011, 45, 7241−7248.(36) Jahnke, A.; Ahrens, L.; Ebinghaus, R.; Temme, C. Urban versusremote air concentrations of fluorotelomer alcohols and otherpolyfluorinated alkyl substances in Germany. Environ. Sci. Technol.2007, 41, 745−752.(37) D’eon, J.; Crozier, P. W.; Furdui, V. I.; Reiner, E. J.; Libelo, E.L.; Mabury, S. A. Perfluorinated phosphonic acids in Canadian surfacewaters and wastewater treatment plant effluent: Discovery of a newclass of perfluorinated acids. Environ. Toxicol. Chem. 2009, 28, 2101−2107.(38) Moller, A.; Ahrens, L.; Surm, R.; Westerveld, J.; Van DerWielen, F.; Ebinghaus, R.; De Voogt, P. Distribution and sources ofpolyfluoroalkyl substances (PFAS) in the River Rhine watershed.Environ. Pollut. 2010, 158, 3243−3250.(39) Ahrens, L.; Shoeib, M.; Harner, T.; Lane, D. A.; Guo, R.; Reiner,E. J. Comparison of annular diffusion denuder and high volume airsamplers for measuring per- and polyfluoroalkyl substances in theatmosphere. Anal. Chem. 2011, 83, 9622−9628.(40) Vierke, L.; Ahrens, L.; Shoeib, M.; Reiner, E. J.; Guo, R.; Palm,W. U.; Ebinghaus, R.; Harner, T. Air concentrations and particle-gaspartitioning of polyfluoroalkyl compounds at a wastewater treatmentplant. Environ. Chem. 2011, 8, 363−371.(41) Dreyer, A.; Matthias, V.; Temme, C.; Ebinghaus, R. Annual timeseries of air concentrations of polyfluorinated compounds. Environ. Sci.Technol. 2009, 43, 4029−4036.

(42) Ahrens, L.; Shoeib, M.; Harner, T.; Lee, S. C.; Guo, R.; Reiner,E. J. Wastewater treatment plant and landfills as sources ofpolyfluoroalkyl compounds to the atmosphere. Environ. Sci. Technol.2011, 45, 8098−8105.

Environmental Science & Technology Article

dx.doi.org/10.1021/es4048945 | Environ. Sci. Technol. XXXX, XXX, XXX−XXXJ