Environ. Sci. Technol. 2007, 41, 3850-3855
Source Apportionment of Atmospheric PAHs in the Western Balkans by Natural Abundance Radiocarbon Analysis ZDENEK ZENCAK,† JANA KLANOVA,‡ IVAN HOLOUBEK,‡ AND O ¨ R J A N G U S T A F S S O N * ,† Department of Applied Environmental Science (ITM), Stockholm University, 10691 Stockholm, Sweden, and Research Centre for Environmental Chemistry and Ecotoxicology (RECETOX), Masaryk University, Kamenice 126/3, 625 00 Brno, Czech Republic
Progress in source apportionment of priority combustion-derived atmospheric pollutants can be made by an inverse approach to inventory emissions, namely, receptorbased compound class-specific radiocarbon analysis (CCSRA) of target pollutants. In the present study, CCSRA of the combustion-derived polycyclic aromatic hydrocarbons (PAHs) present in the atmosphere of the countries of the former republic of Yugoslavia was performed. The carbon stable isotope composition (δ13C) of PAHs varied between -27.68 and -27.19‰, whereas ∆14C values ranged from -568‰ for PAHs sampled in Kosovo to -288‰ for PAHs sampled in the Sarajevo area. The application of an isotopic mass balance model to these ∆14C data revealed a significant contribution (35-65%) from the combustion of non-fossil material to the atmospheric PAH pollution, even in urban and industrialized areas. Furthermore, consistency was observed between the isotopic composition of PAHs obtained by high-volume sampling and those collected by passive sampling. This encourages the use of passive samplers for CCSRA applications. This marks the first time that a CCSRA investigation could be executed on a geographically wide scale, providing a quantitative fieldbased source apportionment, which points out that also nonfossil combustion processes should be targeted for remedial action.
Introduction Combustion of fossil fuels is the major source of energy for today’s global economy (1) but is also largely responsible for the increased global temperature (thus climate change) that has been observed over the last 150 years, as well as for air pollution of significant public health concern (2, 3). While statistical information exists for fossil fuel usage and emissions of combustion pollutants in different countries, the contribution of biomass combustion (as an energy source or simply as open fires, forest fires, waste incineration, etc.) to atmospheric pollution is much more difficult to constrain (4). Predictions based on traditional emission inventory approaches fail because of the very large variations in * Corresponding author phone: +46 8 6747317; fax: +46 8 6747638; e-mail:
[email protected]. † Stockholm University. ‡ Masaryk University. 3850
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 11, 2007
emission factors for different biomass combustion processes (4-6). The challenge of apportioning the pollution sources is further exacerbated for countries where data about energy usage and the extent of combustion of non-fossil material is limited due to difficult socioeconomic conditions. Epidemiological studies have shown that particle-borne air pollution contributes significantly to human morbidity and mortality (2, 3). Currently, 30-40% of Europeans live in cities where the air pollution level is equal or above the guidelines of the European Union (7). The World Health Organization estimates that every year over 300 000 humans die prematurely due to European air pollution (8). The ubiquitous polycyclic aromatic hydrocarbons (PAHs) are excellent molecular markers of combustion particles (9) and together with substituted PAHs (e.g., nitro-PAHs) account for most (35-82%) of the total mutagenic activity of ambient aerosols (10). Recognizing the adverse effects of atmospheric pollution by PAHs, the European Commission established a directive enforcing activities to such pollution with the aim of minimizing harmful effects on human health (11). Major well-known sources of PAHs and other particle-borne air pollution include traffic, power plants, residential heating, and industrial processes. In contrast, PAH emissions from biomass combustion are hard to constrain and easily overlooked. Nevertheless, efficiently decreasing PAH emissions and human exposure requires quantitative knowledge of the relative importance of all major PAH sources. Since traditional emission inventory strategies are challenged significantly by highly variable PAH emission factors (4), particularly for biomass combustion processes, alternative approaches should also be attempted. It is suggested here that the radiocarbon composition of the load of pollutants actually found in the environment may provide an independent approach to their source apportionment. Compound-specific radiocarbon analysis (CSRA) and compound class-specific radiocarbon analysis (CCSRA) have been previously applied in a few studies to investigate the sources of PAHs. Mandalakis et al. investigated the radiocarbon composition of PAHs present in surface sediments from the Stockholm metropolitan area (6). CSRA results showed that 17 ( 9% of the PAHs in the surface sediments originated from biomass burning. Furthermore, an increase of the relative contribution of biomass burning was observed in the suburban areas. In a second study, Mandalakis et al. used natural abundance radiocarbon analysis of pooled PAHs in air samples to show that biomass combustion contributes to ca. 50% of the atmospheric PAHs at a background area in Sweden, whereas it contributes to only ca. 10% to the PAHs levels at two background sites in south Europe (12). Kumata et al. applied CCSRA to PAHs associated with PM10 and PM1.1 aerosols from residential areas of suburban Tokyo, revealing that biomass burning contributes to as much as 17-45% of the PAH burden (13). The combined findings of these studies clearly indicate the importance of biomass burning as a source of pollution by PAHs. However, these studies are limited by the complexity and high demands of CSRA and CCSRA. This challenging technique requires the isolation of ca. 25-30 µg of PAHs to perform 14C analysis. This can be achieved only by taking very large samples (e.g., 0.6-9 kg of sediment as in ref 6) or by pooling a large number of samples (e.g., high-volume air samples collected monthly over 2 years as in ref 12). Furthermore, the resulting analytical cleanup and the effort to isolate the required amount of target compound by preparative capillary gas chromatography (pcGC) are very demanding and time-consuming. 10.1021/es0628957 CCC: $37.00
2007 American Chemical Society Published on Web 04/20/2007
FIGURE 1. Investigated area and percentage of total PAHs originating from combustion of non-fossil matter. The aim of this work was to apply CCSRA to determine the relative contributions of biomass versus fossil fuel combustion to airborne PAH pollution for the western Balkans using geographically distributed samples obtained through an extensive passive sampling campaign. Receptorbased assessment is particularly attractive in socioeconomic challenged regions as the traditional emission inventory approach breaks down with poor availability of energy statistics. To the best of our knowledge, this work represents the first CCSRA-based source apportionment of the combustion products PAHs for any large area such as the western Balkans.
Experimental Procedures Sampling. Passive air samplers consisting of polyurethane foam disks (15 cm diameter, 1.5 cm thick, density 0.030 g/cm3, type N3038, Gumotex Bøeclav) housed in protective chambers were employed. The theory of passive sampling using similar devices was described elsewhere (14). Calibration experiments in the field showed that the sampling rate of these devices can vary between 3 and 7 m3/day (data not shown). The sampling chambers were pre-washed and solvent-rinsed with acetone prior to installation. All polyurethane foam disks were pre-washed, cleaned (extracted for 8 h with acetone and 8 h with dichloromethane), dried in a desiccator under vacuum for 2 days, wrapped in two layers of aluminum foil, placed into zip-lock polyethylene bags, and kept in a freezer prior to deployment. Exposed samplers were wrapped in two layers of aluminum foil, labeled, placed into zip-lock polyethylene bags, and transported in a cooling box at 5 °C to the laboratory where they were stored in a freezer at -18 °C until analysis. Field blanks were obtained by installing and removing the PUF disks at all sampling sites. Passive samplers were deployed at various locations in Croatia (10 sites), Bosnia, Herzegovina (five sites in the Tuzla area and five sites in the Sarajevo area), Serbia, Montenegro (three sites in the Novi Sad area, three sites in the Pancevo area, and three sites in the Kragujevae` area), and Kosovo (two sites in the Pristina area). See Figure 1 for a map with marked sampling locations. Each passive sampler was deployed for 28 days as part of a separate project aiming at investigating the release of pollutants (e.g., PCBs) as a consequence of the recent war (15-17). Therefore, sampling locations included rural as well as urban and industrialized areas and industrial locations subject to bombing during the
war. Sampling was performed continuously between July and December 2004, yielding five samples per location. Additionally, active high-volume air samplers (PS-1, GrasebyAndersen, flow: 20-25 m3/h) equipped with Whatmann quartz fiber filters (fraction dae < 50 µm) and polyurethane foam plugs (Gumotex Bøeclav, density 0.03 g.m-3) were deployed at five locations in Tuzla (Bosnia and Herzegovina) in May 2004, collecting five 24 h samples of gaseous and particulate PAHs at each location to verify if data obtained from the passive samplers at this location would be consistent with those obtained by high-volume sampling. Samples obtained from several months of passive sampling at adjacent locations were combined to obtain the amount of PAHs necessary to perform 14C analysis (see Table 1 for a list of resulting samples and for a brief description of the character of the different sampling locations). The high amount of PAHs in the passive samplers from Tuzla allowed separation of these samples in both winter and summer. Sample Treatment. The samples were extracted with dichloromethane (Pestiscan, Labscan) in a Bu ¨ chi System B-811 automatic extractor. After extraction, the solvent volume was reduced to ca. 2 mL under a gentle nitrogen stream at ambient temperature. The extract was pre-cleaned using adsorption chromatography on a silica column (30 cm length, 1 cm i.d.) eluted with 40 mL of dichloromethanehexane (1:1). Samples from adjacent locations were pooled to provide the quantity of PAHs necessary for 14C analysis. The so-obtained samples were reduced in volume to ca. 50 µL under a nitrogen stream and dissolved in 5 mL of n-pentane (>98%, Fluka). Aliphatic compounds were removed using a dimethylformamide/n-pentane partitioning method (18) consisting of the following steps: 5 mL of dimethylformamide (>99.5% VWR Int.) containing 5% H2O (HPLC quality, VWR Int.) was added, and the mixture was shaken for 1 min and centrifuged for 2 min at 2000 rpm with a Sigma 4--15 centrifuge (Sigma Laboratory Centrifuges). The n-pentane phase was separated, and 5 mL of dimethylformamide containing 5% H2O was added to it. The mixture was shaken again for 1 min and centrifuged for 2 min at 2000 rpm, the n-pentane phase was discarded, and the so-obtained dimethylformamide extracts were combined. A total of 10 mL of water and 10 mL of n-hexane (HPLC quality, Fluka) was added to the dimethylformamide phase, the mixture was shaken for 1 min and centrifuged for 2 min at 2000 rpm, and the n-hexane phase was separated. The dimethylforVOL. 41, NO. 11, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3851
TABLE 1. Pooled Samples from Which PAHs Were Isolated and Investigated for Their 13C and 14C Isotopic Composition and Brief Description of Character of Sampling Locations from Which Pooled Samples Originate pooled sample
character of sampling locations
Croatia (Zadar + Zagreb + western Slavonia) Kosovo (Pristina) Serbia (Kragujevacˇ ) Serbia (Pancevo) Serbia (Novi Sad) Bosnia Herzegovina (Sarajevo) Bosnia Herzegovina (Tuzla July to Sept.) Bosnia Herzegovina (Tuzla Oct. to Dec.) Bosnia Herzegovina (Tuzla high vol, May 2004) Bosnia Herzegovina (Tuzla high vol, May 2004)
mix of urban, industrial, and recreational urban and industrial very industrial urban and industrial urban and industrial urban and industrial urban and industrial urban and industrial urban and industrial urban and industrial
TABLE 2. Ratios between Different PAHs Observed for Different Samples and End Member Valuesa sample
Fl/(Fl + Py)b
MP/P
Ret/(Ret + Chy)c
1,7/(1,7 + 2,6)
Croatia (Zadar + Zagreb + western Slavonia) Kosovo (Pristina) Serbia (Kragujevacˇ ) Serbia (Pancevo) Serbia (Novi Sad) Bosnia Herzegovina (Sarajevo) Bosnia Herzegovina (Tuzla July to Sept.) Bosnia Herzegovina (Tuzla Oct. to Dec.) Bosnia Herzegovina (Tuzla high vol, May 2004) Bosnia Herzegovina (Tuzla high vol, May 2004) petrogenic source petroleum combustion coal combustion softwood combustion
0.56 0.54 0.54 0.57 0.56 0.57 0.56 0.57 0.52 0.55 0.5 >0.5
0.29 0.35 0.26 0.54 0.44 0.17 0.26 0.27 0.32 0.23 ca. 5
