Embed Size (px)
Citation preview
at SciVerse ScienceDirect
Environmental Pollution 169 (2012) 204e209
Contents lists available
Environmental Pollution
journal homepage: www.elsevier .com/locate/envpol
Atmospheric fate of poly- and perfluorinated alkyl substances(PFASs): II. Emission source strength in summer in Zurich, Switzerland
Zhanyun Wang a, Martin Scheringer a,*, Matthew MacLeod b, Christian Bogdal a, Claudia E. Müller a,c,Andreas C. Gerecke c, Konrad Hungerbühler a
a Institute for Chemical and Bioengineering, ETH Zürich, Wolfgang-Pauli-Strasse 10, 8093 Zürich, SwitzerlandbDepartment of Applied Environmental Science (ITM), Stockholm University, SE-10691 Stockholm, Swedenc Empa, Swiss Federal Laboratories for Materials Science and Technology, Überlandstrasse 129, 8600 Dübendorf, Switzerland
a r t i c l e i n f o
Article history:Received 24 March 2012Accepted 31 March 2012
Keywords:Fluorotelomer alcohols (FTOHs)Perfluorooctane sulfonamides (FOSAs)Emission source strengthNocturnal boundary layerDiel pattern
* Corresponding author.E-mail address: [email protected] (M. Sch
0269-7491/$ e see front matter � 2012 Elsevier Ltd.doi:10.1016/j.envpol.2012.03.037
a b s t r a c t
Fluorotelomer alcohols (FTOHs) and perfluorooctane sulfonamides (FOSAs) are present in consumerproducts and are semi-volatile precursors of persistent perfluoroalkyl acids (PFAAs). The high variability oflevels of FTOHs and FOSAs in products makes it difficult to derive FTOH- and FOSA-emissions from urbanareas based on emission factors. Here we used a multimedia mass balance model that describes thedayenight cycle of semi-volatile organic chemicals in air to interpret measurements of 8:2 FTOH, 10:2FTOH, MeFOSA and EtFOSA from a sampling campaign in summer 2010 in Zurich, Switzerland. The esti-mated emission source strength of the four substances follows the sequence: 8:2 FTOH (2.6 g/h) > 10:2FTOH (0.75 g/h)>MeFOSA (0.08 g/h)> EtFOSA (0.05 g/h). There is no FTOHs- or FOSAs-related industry inZurich. Accordingly, our estimates are representative of diffusive emissions during use and disposal ofconsumer products, and describe noticeable sources of these PFASs to the environment.
� 2012 Elsevier Ltd. All rights reserved.
1. Introduction
For over a decade, perfluoroalkyl acids (PFAAs), including per-fluorooctanoic acid (PFOA) and perfluorooctane sulfonic acid(PFOS), have been detected in the environment and biota all overthe world (Giesy and Kannan, 2001; Rayne and Forest, 2009), aswell as in human food items (Clarke and Smith, 2011). PFAAs arehighly persistent (Parsons et al., 2008; Frömel and Knepper, 2010),and some are bioaccumulative (Conder et al., 2008). Therefore, theyhave attracted attention as environmental pollutants of highconcern. In 2000, the major producers started to phase out thelong-chain PFAAs, including PFOS and PFOA (Prevedouros et al.,2006). However, many long-chain poly- and perfluorinated alkylsubstances (PFASs), including fluorotelomer alcohols (FTOHs) andperfluorooctane sulfonamides (FOSAs), are still in production anduse and can degrade into PFAAs in the abiotic environment (Youngand Mabury, 2010), in biota (Frömel and Knepper, 2010) and in thehuman body (Nilsson et al., 2010). Hence, there is a current need tounderstand the sources, transport and environmental fate of theseprecursors to PFAAs.
Unlike PFAAs, which are strong acids and are mainly emittedfrom manufacturing sites, the precursors are neutral semi-volatile
eringer).
All rights reserved.
species and have numerous possible emission pathways. Besidesdirect emission from industrial production processes, a significantamount may be present as an ingredient or undesired residuals inpolymeric materials, chemicals or consumer products and bereleased during use or disposal (Prevedouros et al., 2006). The highvariability of levels in consumer products (Dinglasan-Panlilio andMabury, 2006; Fiedler et al., 2010) makes it difficult to estimatediffusive emissions on the basis of surveys of consumer products. Atthe same time, it is known that urban areas in industrializedcountries, where consumer products containing PFASs are widelyused, can be important diffusive sources of PFASs (Piekarz et al.,2007; Primbs et al., 2008; Dreyer and Ebinghaus, 2009; Dreyeret al., 2009).
Recently, a novel method has been developed to estimate thestrength of diffusive emissions of semi-volatile organic chemicals(SVOCs) from urban areas (MacLeod et al., 2007). The methodcombines a multimedia mass balance model that includes atmo-spheric boundary layer dynamics with high-temporal-resolutionmeasurements and it has been successfully applied to poly-brominated diphenyl ethers (PBDEs) and polychlorinated biphenyls(PCBs) in different cities (Moeckel et al., 2008, 2010; Gasic et al.,2009, 2010).
In this work, we further extended the application of thisestimation technique for diffusive emission source strength tofour semi-volatile poly- and perfluorinated precursors to PFAAs
Z. Wang et al. / Environmental Pollution 169 (2012) 204e209 205
(8:2 fluorotelomer alcohol (8:2 FTOH), 10:2 fluorotelomer alcohol(10:2 FTOH), N-methyl perfluorooctane sulfonamide (MeFOSA) andN-ethyl perfluorooctane sulfonamide (EtFOSA); molecular struc-tures can be found in Table S1 in the Supplemental Material). Wereport their diffusive emission source strengths from thecity of Zurich (Switzerland), which is an urban area typical ofWestern industrialized countries. Our modeling study exploits themeasurements from a sampling campaign carried out in August2010 that is presented in detail in an accompanying article (Mülleret al., in this issue). We use a Monte Carlo analysis and estimates ofthe uncertainty and variability in input parameters to assess therange of emission strengths of PFASs that results in modeledconcentrations that are consistent with the measurements. To getmore insight into the emission pathways of these semi-volatilePFASs, our emissions estimates are compared with literature dataestimated from emission-factor-based methods.
2. Materials and methods
2.1. Model design
The mass balance model is a derivative of the Berkeley-Trent (BETR) contami-nant fate modeling framework (MacLeod et al., 2001) that has been further devel-oped in previous studies (MacLeod et al., 2007; Moeckel et al., 2008, 2010; Gasicet al., 2009, 2010) and in the current work.
The model tracks the mass balance of a chemical in environmental compart-ments that include atmosphere, soil, vegetation, water and sediment. Specifically, ithas been designed to describe the dayenight cycle of SVOC concentrations in air byincluding atmospheric boundary layer dynamics (Holtslag and Duynkerke, 1998;Whiteman, 2000). After sunset, the land surface cools more quickly than the over-lying air. Thus, a nocturnal stable boundary layer (SBL) forms with cold air near thesurface underlying warmer air. The SBL prevents vertical mixing and thus contam-inants being emitted near the surface are trapped (see Fig. 1, right). When the sunrises, solar radiation heats the ground and thus heats the air adjacent to the surface.This heating causes convectional vertical movement of air, which mixes thecontaminants enriched within the SBL with less contaminated upper air (dilution).
Our model divides the atmospheric compartment into three parts (the lowerand upper air compartment and the free troposphere) and uses mass transfercoefficients (MTC, m/h) between the layers to describe the boundary layer dynamics.To represent formation and dissipation of the SBL, the MTC between the lower andupper air compartment is set at 0m/h at nightwhereas during the daytime it is set ata higher value selected so that the lower air compartment is fully mixed with theupper air compartment on the timescale of 1 h.
2.2. Model parameterization
In the current study, the model was parameterized to describe a 72-h samplingcampaign performed from 00:00, 20.08.2010 to 24:00, 22.08.2010 (local
Fig. 1. Schematic diagram of the dayenight evolution the boundary layer over a cityand the enrichment of contaminants within the nocturnal stable boundary layer.
time¼ UTC þ 2h) at a sampling site located in the city of Zurich (47.474 �N, 8.536 �E,446 m above sea level). Details about the sampling campaign are described in theaccompanying paper (Müller et al., in this issue).
The physicochemical properties and estimated degradation half-lives of fourPFASs, the environmental parameters describing the city of Zurich, and the kineticparameters describing interphase transport processes of chemicals were set asconstants in all modeling scenarios. In the current study, the total surface area of themodel regionwas set to 100 km2 to describe the most densely populated area of thecity of Zurich. Previous applications of the model used an area of 1000 km2, whichwould also include less-densely populated areas near Zurich. More information onthese input parameters is listed in Tables S1 to S3 in the Supplemental Material.
Temperature and wind speed data (see Fig. 2, left two panels) for Zurich werecollected during the sampling period from four nearbymeteorological stations (datafrom NABEL and IDAWEB; geographical information can be found in Table S5 andFig. S1 in the Supplemental Material). To describe the wind conditions, we usedthe average from four stations (Affoltern, Kaserne, Kloten and Fluntern), becauseeach station alone was likely to be influenced by the turbulence caused by buildingsin the city and thus would not be representative of the conditions in the city asa whole.
Both the OH radical concentration and the mixing height were modeled bytruncated sinusoidal forcing functions. The nighttime OH radical concentration wasset to zero, and the daytime concentrations were calibrated to correspond to the24-h global average of 9.7 � 105 molecules/cm3 (Anderson and Hites, 1996). Themixing height data shown in Fig. 2 were derived from solar radiation measurementsand compared with three other indicators, including results from the Hybrid SingleParticle Lagrangian Integrated Trajectory (HYSPLIT) Model (Draxler and Hess, 1997,see the empty dots in the Fig. 2, right panel), and the diel patterns of potentialtemperature and air pollutants levels (see Figs. S2 and S3 in the SupplementalMaterial). We assumed the median height of the SBL at night was 100 m. Detailsare provided in the Supplemental Material.
The model was run dynamically to describe 72 h of real time. The initialconcentrations were set at the results from a steady-state calculation using themedian values of the meteorological data and the forcing functions. The integrationtime step was selected to resolve the fastest process in the model and is less than2min. AMonte Carlo analysis with 1500 discrete runs for each chemical was appliedto estimate the range of modeled concentrations resulting from the estimateduncertainty and variability in input parameters (MacLeod et al., 2002). Detailedinformation on the confidence factor used for each parameter can be found inTables S1 to S4 in the Supplemental Material.
2.3. Estimation of diffusive emission source strength
The only adjustable input parameter for the model is the emission sourcestrength, whereas all other input parameters were derived from empiricalmeasurements and were not treated as adjustable. In the model, the emission intothe model regions was divided into two types: background air inflow and localemission into the air compartment of the region. It was assumed that the city ofZurich is influenced by both emission types. The background air inflow is a result oftransport of PFASs into the city from remote emission sources. Thus, typical back-ground levels in the range of measurements in areas that are remote from sources,which are listed in the accompanying article (Müller et al., in this issue), were takenfor the input with inflowing air. The diffusive emission source of each PFAS species inthe city of Zurich was modeled as a pool of pure-phase PFAS that is a surrogate for allvolatilization sources to the atmosphere from various materials present in thesource region. The volatilization flux (kg/year) was calculated as the product of (i)a diffusion mass transfer coefficient (5 m/h), (ii) the surface area of the pure-phasePFAS-pool available as a volatilization source (m2), and (iii) the concentrationgradient between the surface of the PFAS-pool and the air (mol/m3). The concen-tration gradient was calculated as the concentration at the surface of the pure-phasePFAS-pool (estimated as the vapor pressure (VP) divided by RT by assuming equi-librium conditions) minus a negligible concentration in air (Mackay, 2001). Thus,temperature is the only driving force of diel variation in emission strength and thesurface area of pure-phase PFAS-pool was the only adjustable parameter used to fitthe model results to the measured concentrations (by means of visual inspection),whereas default values of all other input parameters were used. The correspondingvolatilization flux (g/h) represented the estimated diffusive emission sourcestrength in the city.
3. Results
3.1. Model results
Modeled concentrations are in good agreement with measure-ments of all four PFASs during the sampling campaign (see Fig. 3top row). Both measurements and modeled results follow thesame diel pattern with maxima at night and minima duringthe daytime. The diel boundary layer dynamics was identified to be
Fig. 2. Left: median values of measured temperature in Zurich city. Middle: median values of wind speed (solid line) derived from measurements at four meteorological stations(Kloten, Affoltern, Fluntern, Kaserne). Right: median values of mixing height data were derived from solar radiation measurements and compared with other indicators, includingthe HYSPLIT model results, which are shown as empty dots. The other indicators, including diel pattern of potential temperature and air pollutant levels are provided in Fig. S2 andS3 in the Supplemental Material. The x-axis shows the 72-hour period of the sampling campaign conducted between 20.08.2010 and 22.08.2010. Grey areas here and in the figurebelow represent the time period when the nocturnal boundary layer was present. Dashed lines indicate the 2.5%-to-97.5% confidence interval of the temperature, wind speed andmixing height used for the Monte Carlo analysis.
Z. Wang et al. / Environmental Pollution 169 (2012) 204e209206
the major factor responsible for this pattern, with the small volumeof air receiving the PFASs leading to enrichment of PFASs atnight. Low outflow due to the low wind speed also contributes tohigher concentrations at night relative to the daytime (for amore detailed discussion see section S4 in the SupplementalMaterial).
3.2. Emission estimates of the four PFASs in Zurich city
We were able to achieve good fits between the model andthe measurement data for all four PFASs by adjusting only thesize of the hypothetical pool of PFASs, i.e., the emission sourcestrength. The geometric means of the estimated emission sourcestrength of the four PFASs during the 72-h period follow thesequence: 8:2 FTOH (2.6 g/h) > 10:2 FTOH (0.75 g/h) > MeFOSA(0.08 g/h) > EtFOSA (0.05 g/h) (see Fig. 3, bottom row).
3.3. Uncertainty analysis
The uncertainty in modeled emission strength can be derivedfrom the Monte Carlo analysis. A lower bound for the emissionstrength is obtained by dividing the median estimate by the devi-ation of the lower bound of concentrations derived from the MonteCarlo analysis from the median (a factor of 2.9, 3.5, 2.2 and 2.3,respectively, for 8:2 FTOH, 10:2 FTOH, MeFOSA and EtFOSA),because emissions that were higher by that factor would bring thelower bound of the possible model results into agreement with themeasurements. Likewise, an upper bound on the emission esti-mates can be derived from the deviation of the upper bound ofconcentrations (a factor of 2.7, 2.1, 2.2, and 2.5, respectively) (seeFig. 3, top row, from left to right).
In addition, the five most influential model parameters thatcontribute to the variance of the model output were identified inthe Monte Carlo analysis (see Fig. S5 and S6 in the SupplementalMaterial). Among all input parameters, the meteorological condi-tions (wind speed and boundary layer dynamics) generallycontribute the most to uncertainty in calculated concentrations inlower air. The wind speed contributes around 20% (in the daytime)
to 60% (at night) to the overall variance over the time, especially atnight when advection is the only loss process. The nocturnalboundary layer dynamics, including themixing height and the timeof formation and break-up of the layer, contributes<1% (during themiddle of the day) to about 60% (during the period when theboundary layer forms or breaks up) to the overall variance, becauseit affects the volume of air available to dilute PFASs.
For all four PFASs, reactions with OH radicals are too slow to becompetitive with other removal processes. Thus, uncertainty indegradationerelated parameters (diel OH radical concentrationdynamics, activation energy and degradation half-lives) do notsignificantly contribute to the uncertainty in calculated concen-trations of PFASs.
4. Discussion
To compare our results to other emission estimates reported inliterature, the geometric means of the hourly-resolved emissionsource strength (g/h) were extrapolated to yearly average emissionsource strengths (kg/year) by assuming constant emissions over theyear (see Table 1). A recent one-year global passive sampling studyshows levels of different PCBs and PBDEs in the samples collectedduring winter in the Northern Hemisphere were generally lowerthan the ones collected in the other seasons (Pozo et al., 2009). PCBsand PBDEs likely have similar emission pathways as the PFASsestimated here, i.e. through volatilization from reservoirs includingconsumer products. Hence, the simplified extrapolation made heremay introduce an overestimation of the true annual average; theseasonal variation of emissions of the PFASs investigated stillremains unknown.
4.1. Estimated diffusive emission strengths of FTOHs in Zurich city
For FTOHs, we compared our emission estimates with valuesfrom three available literature sources. Buser and Morf (2009)reported a substance flow analysis of PFOA, PFOS and someprecursors including FTOHs in Switzerland in 2007. Based on theresidual contents of FTOHs detected in samples including
Fig. 3. Top: modeled (solid lines) and measured (triangles) concentrations of the four PFASs in the air near surface in Zurich city; the dashed lines indicate the 2.5%-to-97.5%confidence interval of the modeled PFAS concentrations derived from the Monte Carlo analysis. Bottom: the emission strength (g/h) that yields the median values of the modeledconcentrations of the four PFASs in Zurich city; the horizontal dashed line depicts the geometric mean of the emission strength during the 72-h period.
Z. Wang et al. / Environmental Pollution 169 (2012) 204e209 207
dispersions of fluorotelomer polymers and other fluorosurfactants,they estimated the emission source strength for 8:2 and 10:2 FTOHsinto air in Switzerland in 2007 (Buser and Morf, 2009). In order tocompare their estimates with our results, emission strengths wereexpressed on a per-capita basis (see Table 1), as PFASs are onlyemitted by human activities. The median estimates from Buser andMorf (2009) are generally higher by a factor of three than the onesfrom our work. This is likely caused by the fact that, in their study,emission source strengths along the entire life cycle includingproduction, use, and disposal were considered whereas in thecurrent study, the estimated emission source strength representsmainly diffusive emissions from use and disposal of consumerproducts in Zurich city where there is no fluoropolymer-relatedindustry. However, our estimated emission source strengths arestill within the uncertainty range of the estimates of Buser andMorf(2009) (close to the lower bound).
Yarwood et al. (2007) estimated the emission of FTOHs in NorthAmerica in 2004 and 2007, and Wania (2007) estimated the globalemission of FTOHs over 50 years in the past based on the review ofPrevedouros et al. (2006). In these two cases, a regional model anda global model, respectively, was used to evaluate the reliability ofthe emission estimates. The model-predicted FTOH air concentra-tions were generally in agreement with available monitoringdata, suggesting that the emission estimates of Yarwood et al.(2007) and Wania (2007) are plausible. Recently, Shoeib et al.(2011) measured PFASs in indoor air in Vancouver, Canada, andcompared the results with other literature data; they concludedthat the levels of FTOHs are likely to be broadly similar acrosshomes in industrialized regions (Shoeib et al., 2011). Therefore, weagain normalized the estimates from Yarwood et al. (2007) andWania (2007) per capita and compared them with our estimates(see Table 1).
The estimated per-capita emission of sum FTOHs (including 6:2,8:2 and 10:2 FTOH) in 2005 from Wania (2007) is very close to thevalues reported by Buser and Morf (2009), supporting theassumption that emissions of FTOHs in industrialized regions arelikely to have similar patterns. Hence, the discrepancy between
Wania (2007) and the current study is likely caused by the differentscope of studies as well, as discussed above.
In comparison to the emission source strength estimated byYarwood et al. (2007), our estimates are close to their emissionscenario for 2004 andmuch higher than their emission scenario for2007. This finding is consistent with a monitoring study (Shoeibet al., 2010) reporting that concentrations of 8:2 FTOH in airsamples collected on the Gulf of Mexico and East Coast CarbonCruise in 2007 are close to the model results from Yarwood et al.(2007) based on the 2004-emission scenario. Therefore, a closerlook at both emission scenarios in Yarwood et al. (2007) was taken.The 2004-emission scenario was estimated based on productiondata from DuPont manufacturing locations and the estimatedcustomer, consumer and disposal data from Prevedouros et al.(2006), whereas the 2007-emission scenario was estimated basedon the producer’s reduction plan. Moreover, both emissionscenarios were estimated solely based on the data gained from oneproducer (DuPont) and then extended to the whole market usingan extrapolation factor of 2.5. On the other hand, PFASs productionhas shifted to other countries with less stringent regulations since2002. For example, companies in China began large-scale ofproduction of PFOS and related products in 2003. The 2006 annualproduction volume is estimated to be more than 200 t, mainly foruse in the textile industry and for export (Wang et al., 2009).Therefore, the 2007-emission scenario in Yarwood et al. (2007)mayunderestimate emissions, whereas the 2004-emission scenario isconsistent with measurements and modeling reported here and inother studies.
In summary, our estimated emission source strengths of FTOHsare generally in good agreement with the available estimates,which are all based on emission factors along the life cycle ofFTOHs. The discrepancy between the current estimates and the2007-emission scenario in Yarwood et al. (2007) suggests thatthere are emissions from the remaining consumer productsmanufactured prior to the execution of the reduction plan, or fromongoing production in countries where legislation is less restrictive,or from some uncounted sources missing in those emission-factor-
Table 1The emission strength data estimated in this study and derived from other literature sources. All estimates refer to emissions to air.
Source Chemical Year Area Population in the area Normalized emission flux per capita[kg/year capita] a
Estimated emission fluxin Zurich city[kg/year] b
Median Min. Max. Median Min. Max.
this study 8:2 FTOH 2010 Zurich city 3.77 � 105 c 5.97 � 10e5 2.07 � 10e5 1.61 � 10e4 22.5 7.8 60.810:2 FTOH 2010 3.77 � 105 c 1.72 � 10e5 4.91 � 10e6 3.61 � 10e5 6.5 1.9 13.7MeFOSA 2010 3.77 � 105 c 1.86 � 10e6 8.45 � 10e7 4.09 � 10e6 0.7 0.3 1.5EtFOSA 2010 3.77 � 105 c 1.06 � 10e6 4.61 � 10e7 2.65 � 10e6 0.4 0.2 1.0
Buser and Morf, 2009 8:2 FTOH 2007 Switzerland 7.55 � 106 d 1.72 � 10e4 2.65 � 10e5 1.19 � 10e3 64.9 10.0 449.110:2 FTOH 2007 7.55 � 106 d 6.62 � 10e5 1.32 � 10e5 5.30 � 10e4 25.0 5.0 199.6
Yarwood et al., 2007 8:2 FTOH 2004 North America 3.26 � 108 e 6.21 � 10e5 23.42007 3.35 � 108 e 1.76 � 10e6 0.7
10:2 FTOH 2004 3.26 � 108 e 3.20 � 10e5 12.12007 3.35 � 108 e 9.35 � 10e7 0.4
Wania, 2007 S FTOHs f 2005 global 1.01 � 109 g 2.19 � 10e4 82.010:2 FTOH 2005 1.01 � 109 g 2.19 � 10e5 h 5.47 � 10e5 h 8.2 20.6
a Normalized emission flux per capita ¼ estimated emission flux/population in the area.b Estimated emission flux in Zurich city ¼ normalized emission flux per capita � population in Zurich city.c Official website of Zurich city (http://www.stadt-zuerich.ch/content/prd/en/index/statistik.html).d OECD database (http://stats.oecd.org/Index.aspx).e Official website of United nations (http://esa.un.org/wpp/unpp/panel_population.htm).f The data was read from the Fig. S1 in the supporting information of Wania (2007), which was estimated based on the review of Prevedouros et al. (2006).g These global population numbers are three times the population in North America, since in Yarwood et al. (2007), it was assumed that the global emission of FTOHs was
three times stronger than the one in North America. Hence, the global population that is related to the PFASs emission is assumed to be three times the population in NorthAmerica.
h Wania (2007) reviewed three papers on the monitoring of PFASs in air and concluded that the relative contribution of 10:2 FTOH to the total emissions is likely on theorder of 10e25%.
Z. Wang et al. / Environmental Pollution 169 (2012) 204e209208
based methods. It is also likely that it takes several years until theresidual content of FTOHs in consumer products on the market orbeing used is actually reduced.
4.2. Estimated diffusive emission strengths of FOSAs in Zurich city
For MeFOSA and EtFOSA, to our knowledge there are no otheremission strength estimates available. Our emissions of MeFOSAand EtFOSA estimated for the year 2010 are order(s) of magnitudelower than the ones of FTOHs as shown in the Results section,whereas in 2000 FOSAs were produced in similar orders ofmagnitude amount as FTOHs globally (Lange et al., 2006;Prevedouros et al., 2006). Such a decrease of FOSAs emissionsinto air was also observed in some other indoor or outdoor airsampling campaigns (Piekarz et al., 2007; Shoeib et al., 2011) and isconsistent with the phase-out of production of these perfluorooctylsulfonyl-based PFASs in 2002 by the principal manufacturers(Prevedouros et al., 2006) and their listing under part 1 of “AnnexB” in the Stockholm Convention on POPs, which allows onlyrestricted uses for certain applications such as firefighting foams,metal plating and textiles (Wang et al., 2009). These exceptionsinclude, however, irreplaceable applications of PFASs. A likelyexplanation for our findings is that there is still low but ongoingvolatilization of MeFOSA and EtFOSA from consumer productsmanufactured prior to 2002. It may still take years until they are alldisappeared in consumer products and in the environment.
4.3. Limitations and outlook
In this work, the diffusive emission source strengths of foursemi-volatile PFASs from a typical urban area have been estimated.These estimates are similar to the low-end of estimates that havebeen made at the global level, possibly because they represent onlydiffusive emissions from the city of Zurich, where there is nomanufacturing or processing of PFASs or PFAS-containing products.Therefore, emissions from processes in the PFAS industry are notincluded in our results. To estimate more comprehensive overallglobal emission source strengths of these precursors to PFAAs,studies using this combined monitoring-and-modeling technique
should be carried in multiple places, including urban areas withfluoropolymer-based industries, especially in new productioncountries like China.
Acknowledgments
We are grateful to the Swiss Federal Office for the Environment(FOEN) for funding of this research. We thank AndreasM. Buser andBojan Gasic for helpful discussions, Erol Dedeoglu for IT support,and NABEL and MeteoSchweiz for providing meteorological data.Some elements in Fig. 1 are courtesy of the Integration and Appli-cation Network (ian.umces.edu/symbols/)
Appendix A. Supplementary material
Supplementary material associated with this article can befound, in the online version, at doi:10.1016/j.envpol.2012.03.037.
References
Anderson, P.N., Hites, R.A., 1996. OH radical reactions: the major removal pathwayfor polychlorinated biphenyls from the atmosphere. Environmental Science &Technology 30, 1756e1763. doi:10.1021/es950765k.
Buser, A.M., Morf, L., 2009. Substance Flow Analysis for Switzerland e Per-fluorinated Surfactants Perfluorooctanesulfonate (PFOS) and PerfluorooctanoicAcid (PFOA). Federal Office for the Environment (FOEN), 22/09, Bern. www.environment-switzerland.ch/uw-0922-e.
Clarke, B.O., Smith, S.R., 2011. Review of ‘emerging’ organic contaminants inbiosolids and assessment of international research priorities for the agriculturaluse of biosolids. Environment International 37, 226e247. doi:10.1016/J.ENVINT.2010.06.004.
Conder, J.M., Hoke, R.A., De Wolf, W., Russell, M.H., Buck, R.C., 2008. Are PFCAsbioaccumulative? A critical review and comparison with regulatory criteria andpersistent lipophilic compounds. Environmental Science & Technology 42,995e1003. doi:10.1021/es070895g.
Dinglasan-Panlilio, M.J.A., Mabury, S.A., 2006. Significant residual fluorinatedalcohols present in various fluorinated materials. Environmental Science &Technology 40, 1447e1453. doi:10.1021/es051619þ.
Draxler, R.R., Hess G.D., 1997. Description of the HYSPLIT_4 modeling system.NOAA Tech. Memo. ERL ARL-224, NOAA Air Resources Laboratory, SilverSpring, MD, p. 24.
Dreyer, A., Ebinghaus, R., 2009. Polyfluorinated compounds in ambient air fromship- and land-based measurements in northern Germany. Atmospheric Envi-ronment 43, 1527e1535. doi:10.1016/j.atmosenv.2008.11.047.
Z. Wang et al. / Environmental Pollution 169 (2012) 204e209 209
Dreyer, A., Matthias, V., Temme, C., Ebinghaus, R., 2009. Annual time series of airconcentrations of polyfluorinated compounds. Environmental Science & Tech-nology 43, 4029e4036. doi:10.1021/es900257w.
Fiedler, S., Pfister, G., Schramm, K.W., 2010. Poly- and perfluorinated compounds inhousehold consumer products. Toxicological & Environmental Chemistry 92,1801e1811. doi:10.1080/02772248.2010.491482.
Frömel, T., Knepper, T.P., 2010. Biodegradation of fluorinated alkyl substances.Reviews of Environmental Contamination and Toxicology 208, 161e177.doi:10.1007/978-1-4419-6880-7_3.
Gasic, B., Moeckel, C., MacLeod, M., Brunner, J., Scheringer, M., Jones, K.C.,Hungerbühler, K., 2009. Measuring and modeling short-term variability of PCBsin air and characterization of urban source strength in Zurich, Switzerland.Environmental Science & Technology 43, 769e776. doi:10.1021/es8023435.
Gasic, B., MacLeod, M., Klanova, J., Scheringer, M., Ilic, P., Lammel, G., Pajovic, A.,Breivik, K., Holoubek, I., Hungerbühler, K., 2010. Quantification of sources ofPCBs to the atmosphere in urban areas: a comparison of cities in North America,Western Europe and former Yugoslavia. Environmental Pollution 158,3230e3235. doi:10.1016/j.envpol.2010.07.011.
Giesy, J.P., Kannan, K., 2001. Global distribution of perfluorooctane sulfonate inwildlife. Environmental Science & Technology 35, 1339e1342. doi:10.1021/es001834k.
Holtslag, A.A.M., Duynkerke, P.G., 1998. Clear and Cloudy Boundary Layers:Proceedings of the Colloquium “Clear and Cloudy Boundary Layers”, Amster-dam, 26e29 August, 1997. Royal Netherlands Academy of Arts and Sciences,Amsterdem. 10: 9069842351.
IDAWEB, 2011: the data portal of MeteoSwiss, http://www.meteoschweiz.admin.ch/web/en/services/data_portal/idaweb.html
Lange, F.T., Schmidt, C., Brauch, H.J., 2006. Perfluoroalkylcarboxylates andeSulfonates: Emerging Contaminants for Drinking Water Supplies? RhineWater Works, Association of River Waterworks, The Netherlands.
MacLeod, M., Woodfine, D.G., Mackay, D., McKone, T., Bennett, D., Maddalena, R.,2001. BETR North America: a regionally segmented multimedia contaminantfate model for North America. Environmental Science and Pollution Research 8,156e163. doi:10.1007/BF02987379.
MacLeod, M., Fraser, A.J., Mackay, D., 2002. Evaluating and expressing the propa-gation of uncertainty in chemical fate and bioaccumulation models. Environ-mental Toxicology and Chemistry 21, 700e709. doi:10.1002/etc.5620210403.
MacLeod, M., Scheringer, M., Podey, H., Jones, K.C., Hungerbühler, K., 2007. Theorigin and significance of short-term variability of semivolatile contaminantsin air. Environmental Science & Technology 41, 3249e3253. doi:10.1021/es062135w.
Mackay, D., 2001. Multimedia Environmental Models: the Fugacity Approach. LewisPublisher, Boca Raton, Florida. 10: 0873712420.
Moeckel, C., MacLeod, M., Hungerbühler, K., Jones, K.C., 2008. Measurement andmodeling of diel variability of polybrominated diphenyl ethers and chlordanesin air. Environmental Science & Technology 42, 3219e3225. doi:10.1021/es702341q.
Moeckel, C., Gasic, B., MacLeod, M., Scheringer, M., Jones, K.C., Hungerbühler, K.,2010. Estimation of the source strength of polybrominated diphenyl ethersbased on their diel variability in air in Zurich, Switzerland. EnvironmentalScience & Technology 44, 4225e4231. doi:10.1021/es1001049.
Müller, C.E., Gerecke, A.C., Bogdal, C., Wang, Z., Scheringer, M., Hungerbühler, K.Atmospheric fate of poly- and perfluorinated alkyl substances (PFASs): I. dailypattern of air concentrations in summer in Zurich, Switzerland. EnvironmentalPollution, in this issue.
NABEL, 2011. the National Air Pollution Monitoring Network (NABEL), Switzerland.http://www.bafu.admin.ch/luft/luftbelastung/index.html?lang¼en
Nilsson, H., Kärrman, A., Rotander, A., van Bavel, B., Lindström, G., Westberg, H.,2010. Inhalation exposure to fluorotelomer alcohols yield perfluorocarboxylatesin human blood? Environmental Science & Technology 44, 7717e7722.doi:10.1021/es101951t.
Parsons, J.R., Saez, M., Dolfing, J., de Voogt, P., 2008. Biodegradation of per-fluorinated compounds. Reviews of Environmental Contamination and Toxi-cology 196, 1e19. doi:10.1007/978-0-387-78444-1_2.
Piekarz, A.M., Primbs, T., Field, J.A., Barofsky, D.F., Simonich, S., 2007. Semivolatilefluorinated organic compounds in Asian and western U.S. air masses. Envi-ronmental Science & Technology 41, 8248e8255. doi:10.1021/es0713678.
Pozo, K., Harner, T., Lee, S., Wania, F., Muir, D.C.G., Jones, K.C., 2009. Seasonallyresolved concentrations of persistent organic pollutants in the global atmo-sphere from the first year of the GAPS study. Environmental Science & Tech-nology 43, 796e803. doi:10.1021/es802106a.
Prevedouros, K., Cousins, I.T., Buck, R.C., Korzeniowski, S.H., 2006. Sources, fate andtransport of perfluorocarboxylates. Environmental Science & Technology 40,32e44. doi:10.1021/es0512475.
Primbs, T., Piekarz, A., Wilson, G., Schmedding, D., Higginbotham, C., Field, J.,Simonich, S.M., 2008. Influence of Asian and Western United States urban areasand fires on the atmospheric transport of polycyclic aromatic hydrocarbons,polychlorinated biphenyls, and fluorotelomer alcohols in the western UnitedStates. Environmental Science & Technology 42, 6385e6391. doi:10.1021/es702160d.
Rayne, S., Forest, K., 2009. Perfluoroalkyl sulfonic and carboxylic acids: a criticalreview of physicochemical properties, levels and patterns in waters andwastewaters, and treatment methods. Journal of Environmental Science andHealth Part A, Toxic/Hazardous Substances & Environmental Engineering 44,1145e1199. doi:10.1080/10934520903139811.
Shoeib, M., Vlahos, P., Harner, T., Peters, A., Graustein, M., Narayan, J., 2010. Surveyof polyfluorinated chemicals (PFCs) in the atmosphere over the northeastAtlantic Ocean. Atmospheric Environment 44, 2887e2893. doi:10.1016/j.atmosenv.2010.04.056.
Shoeib, M., Harner, T., Webster, G.M., Lee, S., 2011. Indoor sources of poly- andperfluorinated compounds (PFCs) in Vancouver, Canada: implications forhuman exposure. Environmental Science & Technology 45, 7999e8005.doi:10.1021/es103562v.
Wang, T., Wang, Y., Liao, C., Cai, Y., Jiang, G., 2009. Perspectives on the inclusion ofpperfluorooctane sulfonate into the Stockholm convention on persistentorganic pollutants. Environmental Science & Technology 43, 5171e5175.doi:10.1021/es900464a.
Wania, F., 2007. A global mass balance analysis of the source of perfluorocarboxylicacids in the Arctic Ocean. Environmental Science & Technology 41, 4529e4535.doi:10.1021/es070124c.
Whiteman, C.D., 2000. Mountain Meteorology: Fundamentals and Applications.Oxford University Press, Inc., New York. 10: 0195132718.
Yarwood, G., Kemball-Cook, S., Keinath, M., Waterland, R.L., Korzeniowski, S.H.,Buck, R.C., Russell, M.H., Washburn, S.T., 2007. High-resolution atmosphericmodeling of fluorotelomer alcohols and perfluorocarboxylic acids in the NorthAmerican troposphere. Environmental Science & Technology 41, 5756e5762.doi:10.1021/es0708971.
Young, C.J., Mabury, S.A., 2010. Atmospheric perfluorinated acid precursors:Chemistry, occurrence, and impacts. Reviews of Environmental Contaminationand Toxicology 208, 1e109. doi:10.1007/978-1-4419-6880-7_1.
