Polycyclic aromatic hydrocarbons (PAHs) are products of incomplete combustion of fossil fuels and other organic materials. PAHs originate from a large variety of natural and industrial processes. In cities without industry, the main sources of PAHs are motor vehicles and residential heating (Srogi 2007). Contribution of PAH sources to the atmosphere will depend on a number of factors, including the emission rate of sources, geography and local weather conditions (Ravindra et al. 2008). Some studies have found that PAHs with two or three benzene rings exist in the vapour phase, while PAHs with more than five rings are primarily in the particulate phase; the latter then move into the soil, water, sediments and vegetation through dry or wet deposition (Possanzini et al. 2004; Srogi 2007). It is generally accepted that PAHs are preferentially bound to small particles (Srogi 2007; Ravindra et al. 2008). Some PAHs are mutagenic while some are carcinogenic and they should not be present in the air, or their concentrations should be the lowest possible (WHO 2000). More than 500 PAHs have been identified in the air but routine measurements include only up to 20 PAHs. Benzo(a)pyrene (BaP) is the most commonly measured PAH and it is used as an indicator of carcinogenic hazard in polluted environments.

This paper presents PAH mass concentrations measured in PM10 particles (particles with aerodynamic diameter less than 10 μm), at two different measuring sites in Zagreb, Croatia. Samples were collected on working days in summer and winter periods. The aim of the study was to find out in which part of the city and in which part of the year residents of Zagreb are exposed to higher concentrations of PAHs. Contents of PAHs in PM10 particle fraction and ratios of some PAHs were calculated in order to determine main pollution sources at both measuring sites.

Materials and Methods

Samples were collected at two measuring sites, located in different parts of Zagreb. Measuring site A was located in the northern low-rise residential part of Zagreb with a low population density, and modest traffic. House heating around site A is mostly on gas, but some households still use oil, wood or coal for heating and cooking. Site B was located in the southern part of the city near a busy crossroads, parking lot, and entrance to the underground garage. It is very populated area with buildings connected to district heating system. At both sites, twenty-four hour samples of PM10 particles were collected on quartz filters with a low volume PM10 Sven Leckel sampler from about 50 m3 of air. The samplers were set at about 1.5–2 m above the ground, and about 10–20 m away from the road. Samples were taken only on working days in summer (July and August) and winter (November and December) of 2008. After collection, samples were wrapped in aluminium foil and kept in a deep freezer at −18°C until analysed. Concentrations of the PM10 fraction were determined gravimetrically by weighing on Mettler Toledo MX-5 microbalance, before placing the samples in deep freeze.

The analysis was performed using a Varian Pro Star high performance liquid chromatograph (HPLC) with a fluorescence detector and programmed changes in excitation and emission wavelength to optimize selectivity and sensitivity for individual PAH species (Šišovic and Fugaš 1991; Eiguren-Fernandez and Miguel 2003). PAHs were separated on a Varian stainless steel Pursuit 3 PAH column (100 × 4.6 mm). The mobile phase was a mixture of acetonitrile and water, and the flow rate was 0.5 mL/min. To prepare calibration curves, a commercial PAH standard was used (Supelco EPA 610 PAHs Mix). Standard working solutions were obtained by diluting the certified ones with Merck HPLC-grade acetonitrile. Samples were analysed for the following PAHs: fluoranthene (Flu), pyrene (Pyr), benzo(a)antrazene (BaA), chrysene (Chr), benzo(b)fluoranthene (BbF), benzo(k)fluoranthene (BkF), benzo(a)pyrene (BaP), benzo(ghi)perylene (BghiP), and indeno (1,2,3-cd) pyrene (Ind). The detection limit, which equalled the triple standard deviation of the lowest concentration included in the calibration curve, varied from 0.005 ng/μL for BaP to 0.008 ng/μL for BghiP. For every set of samples we performed two blank tests and no PAHs were detected. Quantification limit for PAHs in 500 μL of acetonitrile and 50 m3 of air varied from 0.05 ng/m3 for BaP to 0.08 ng/m3 for BghiP. Recoveries of PAHs were determined by adding a standard amount of a known PAH to a blank filter. The average recoveries of all analyzed PAHs varied from 94.5% to 98%. The detailed procedure for the preparation of samples was described earlier (Šišovic and Fugaš 1991).

Results and Discussion

Table 1 shows average mass concentrations and other statistically important parameters of PAHs and PM10 particle fractions measured in two different parts of the city. Average mass concentrations of the PM10 particle fraction in the summer at both sites were low and similar. At site A, the average concentration was 21.4 μg/m3 and varied from 8.5 to 54.9 μg/m3. At site B, the average concentration was 23.8 μg/m3 and varied from 1.9 to 67.3 μg/m3. Average concentrations of all measured PAHs were also low at both sites. They were all higher at site B, but it was not statistically significant for Pyr, BaA and Ind (Student t test, p < 0.05). The difference in PAH levels between sites was most pronounced for Flu, Chr and BkF (p < 0.01). Average mass concentration of BaP at site B was 0.073 ng/m3 while at site A the average BaP concentration was 0.041 ng/m3. Concentrations of BaP measured at the same sampling site ten years ago (Šišović et al. 2002) were very similar, the average concentration being 0.05 ng/m3. Those BaP levels are comparable with levels measured in Spain (Callén et al. 2008). For example, average BaP mass concentrations in Zaragoza was 0.089 ng/m3 during warm season, which is similar as in Zagreb on site B. Low concentrations of all PAHs in summer were caused by many factors, such as photochemical, thermal and chemical degradation of PAHs at higher temperatures (Ravindra et al. 2008). Higher concentrations of all PAHs at site B in summer may be a consequence of the sampling location (near a busy crossroads, parking lot, and entrance to the underground garage).

Table 1 Average mass concentrations and other statistically important parameters of PAHs (ng/m3) and PM10 particle fractions (μg/m3) measured in two different parts of the city
Full size table

In winter, average mass concentrations of PM10 particles at both sites were much higher than those in summer (Table 1). Average concentration of PM10 at site A was 36.8 μg/m3 (ranging from 19.2 to 65.1 μg/m3), and was significantly lower than those at site B, where it was 55.6 μg/m3 (ranging from 13.1 to 104.9 μg/m3). Mass concentrations of all PAHs from both sites were much higher in winter than those in summer. Concentrations of all PAHs (except Flu and Pyr) were statistically significantly higher at site B (p < 0.01). Average mass concentration of BaP at site A was 1.147 ng/m3 and ranged from 0.101 to 4.028 ng/m3. Average concentration of BaP at site B was 2.762 ng/m3, and ranged from 0.274 to 11.404 ng/m3. Chr had the highest average concentration at site B, 2.873 ng/m3. Concentrations of Chr, BbF and BaP at site A were the highest and very similar (about 1.145 ng/m3). Comparing with the study of Callén et al. (2008), BaP concentrations in winter in Zaragoza were much lower (0.500 ng/m3) than in Zagreb. In downtown Rome Possanzini et al. (2004) found that concentration of BaP in particulate phase was 2.4 ng/m3 in cold season, which is comparable with site B in Zagreb. Chr in downtown Rome was higher than in Zagreb at site B and it was 3.9 ng/m3. Higher concentrations of PM10 and all PAHs at both sites in winter depended on the strength of the source (house heating, higher traffic intensity), and on weather conditions.

PAH contents in the PM10 particle fraction (ng/mg) were calculated for summer and winter period (Fig. 1). Average PAH contents in summer were low at both sites. They were higher at site B, but the difference was statistically significant only for Chr, BbF and BghiP (p < 0.05). Content of BbF was the highest at both sites; it was 2.74 ng/mg at site A, and 6.69 ng/mg at site B. Content of Pyr was the lowest at both sites; it was about 0.45 ng/mg. Average BaP content was 2.26 ng/mg at site A, and 4.17 ng/mg at site B. In winter, average PAH contents of all PAHs measured at both sites were much higher than those measured in summer. Contents of all PAHs were higher at site B, but only contents of Chr, BghiP and Ind were statistically significant at p < 0.05. Content of BbF and BaP was highest at site A, about 34.66 ng/mg. At site B, content of Chr was the highest, 47.46 ng/mg. Average BaP content at site A was 34.72 ng/mg. Average BaP content at site B was higher and was 46.69 ng/mg.

Fig. 1
figure 1

Average PAH content in the PM10 particle fraction (ng/mg) in summer (a) and winter (b) period

Full size image

Ratios of some PAHs were calculated in order to determine the main pollution sources (Table 2). The ratio Flu/(Flu + Pyr) did not differ significantly between seasons and between locations (p < 0.01). According to the previous studies, the ratio Flu/(Flu + Pyr) between 0.2 and 0.5 is characteristic for exhaust gases, 0.4–0.5 for liquid fossil fuel (vehicle or crude oil) combustion, while >0.5 suggest coal, grass or wood combustion (Yunker et al. 2002; Zhang et al. 2004). Values about 0.60 found in this study are comparable with values from the study of Zhang et al. (2004) and Fu et al. (2010), and higher than those found by Kishida et al. (2008) in Hanoi. Ratio Flu/(Flu + Pyr) obtained in this study suggest that PAHs in particles in Zagreb originate not only from vehicle exhaust gases but also from coal and wood combustion. Ratio BaP/BghiP is used to distinguish traffic exhaust from coal combustion sources. Ratio between 0.3 and 0.4 indicates auto exhausts, 0.46–0.81 diesel combustion and 0.9–6.6 coal combustion, and it is usually higher in winter than in summer (Simcik et al. 1999; Fu et al. 2010). In Zagreb, ratio BaP/BghiP was about 0.77 in summer (traffic and diesel combustion) and it was much higher in winter (1.25–1.67) suggesting contribution from coal and wood combustion due to winter heating. Ratios Ind/BghiP and Ind/(Ind + BghiP) were higher in winter, and during both seasons they were higher at site A. The InP:BghiP ratio in gasoline exhaust is 0.4, whereas this ratio for diesel exhaust is approximately 1.0 (Moon et al. 2006). In Zagreb both sources contribute to PAH levels and diesel is more represented at site A. According to the study of Yunker et al. (2002) ratio Ind/(Ind + BghiP) <0.5 indicate petroleum combustion and <0.7 combustion of wood. Similar results were obtained for urban influenced samples Fraser River and Strait of Georgia (Yunker et al. 2002). Ratios BaA/Chr in Zagreb were three times higher in winter, and no difference between sites was found. Winter ratios correspond to vehicle and coke oven, and summer values to diesel and gasoline (Simcik et al. 1999). Finally, BaP/(BaP + Chr) was about 0.51 except for site B in summer (0.44). Values less than 0.33 indicates vehicle with catalyst, 0.49 gasoline, 0.73 diesel (Ravindra et al. 2008). At site B in summer gasoline is the main source, while at site A, and during the winter at both site there is higher contribution from diesel. During the winter wood and coal combustion contribute to PAH levels in the air at both sites.

Table 2 Diagnostic ratios between some PAHs in PM10 (average ± standard deviation)
Full size table

PM10 particle fraction and PAH mass concentrations measured in the PM10 particle fraction in Zagreb, at both measuring sites show seasonal variations with higher concentrations in the winter. PAH concentrations at site B were about two times higher. Higher concentrations at site B may be due to the busy crossroads, parking lots and entrance to the garage. PAH content in PM10 particles was also higher in winter. Ratios of some PAHs indicate that PAHs originate mostly from traffic and diesel combustion, but also from wood and coal combustion, especially at residential part of the city.