Seasonal Variability of PM10 Chemical Composition Including 1,3,5-triphenylbenzene, Marker of Plastic Combustion and Toxicity in Wadowice, South Poland

Received: May 14, 2020
Revised: October 20, 2020
Accepted: November 1, 2020

 Copyright The Author(s). This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are cited.

Cite this article:

Furman, Pr., Styszko, A., Skiba, l., Zięba, D., Zimnoch, M., Kistler, M., Kasper-Giebl, A., Gilardoni, S. (2021). Seasonal Variability of PM₁₀ Chemical Composition Including 1,3,5-triphenylbenzene, Marker of Plastic Combustion and Toxicity in Wadowice, South Poland. Aerosol Air Qual. Res. 21, 200223. https://doi.org/10.4209/aaqr.2020.05.0223

HIGHLIGHTS

• Seasonal variability of OC, EC, ions, PAHs were determined in PM at urban area.
• Co-combustion had an impact on emissions of tracer and PAHs.
• 1,3,5-triphenylbenzene tracer is useful for investigating plastic co-combustion.

Keywords: Atmospheric aerosols, PAHs, 1,3,5-triphenylbenzene, Ions, OC/EC

1 INTRODUCTION

Particulate matter is a complex mixture of different chemical compounds, including nitrates, sulphates, as well as elemental and organic carbon (OC, EC). Some components of organic carbon are particularly toxic, such as polycyclic aromatic hydrocarbons (PAHs) and their derivatives (nitro-PAHs, hydroxy-PAHs). Particulate matter origin can be both natural and anthropogenic. Anthropogenic sources include, above all, solid fuel combustion processes, transport and illegal waste incineration processes (Putaud et al., 2010). At the urban site, coal is used in ca. 50% of households for residential heating purposes (Lochno et al., 2013; Styszko et al., 2015). Over the last few decades, concern about the health effects has grown, mainly due to the occurrence of increasingly high concentrations of particulate matter in the atmosphere. In Poland, the daily concentration limit of PM₁₀ is 50 µg m^–3, the daily level alert is set to100 µg m^–3, whereas the daily level alarm is 150 µg m^–3. The limits of annual average concentrations of PM₁₀ and PM_2.5 are 40 µg m^–3 and 25 µg m^–3, respectively.

Carbonaceous aerosols (CA) constitute a major fraction of particulate matter (PM) present in urban atmospheres. Their contribution varies, depending on the location, between ca. 30 and 40% of the total mass of coarse fraction (PM₁₀ minus PM_2.5) (Szramowiat et al., 2016), and between ca. 10 and 50% of the total particulate matter (PM₁₀) (Querol et al., 2013). Carbonaceous fraction of PM is a complex mixture of different compounds, that can be grouped into elemental carbon (EC), organic carbon (OC) and carbonate or mineral carbon (CC). Organic carbon (OC) accounts for a large fraction of atmospheric aerosol and its molecular composition might change during atmospheric processesing and aging. The ratio of OC to EC is an important index that reflects the source type and strength (Ji et al., 2016), as well as atmospheric processing. As indicated in the literature (Sillanpää et al., 2006; Samara et al., 2014; Liu et al., 2016) a low OC/EC ratio can be associated with fresh traffic aerosol (1.7–2.3 and 0.5–0.8 for light-duty gasoline and heavy duty diesel vehicles, respectively) (Na et al., 2004; Pio et al., 2011; Samara et al., 2014), whereas higher ratios are observed in residential heating emissions or in areas dominated by secondary organic aerosol (SOA). Examples of OC/EC ratios are 4.2 for wood combustion and 12.7 for natural gas (Na et al., 2004), while Zhang et al. (2013) showed that emissions from fireplaces/woodstoves can vary from (2.8–7.5) to relatively high values (26–119) (Na et al., 2004; Zhang et al., 2013). Tian et al. reported the average OC/EC ratio for coal (1.4 ± 1.3 and 6.3 ± 1.3 for bituminous and anthracite, respectively) whereas agricultural biomass burning ratio was 13.7 ± 2.7 (Tian et al., 2017). Since SOA formation contributes exclusively to OC concentration, SOA dominated environments show generally OC/EC ratios larger than 4 (Zeng et al., 2011; Sandrini et al., 2014).

Polycyclic aromatic hydrocarbons (PAHs) are one of the many particulate matter components. A significant increase of PAHs in the atmosphere has a negative impact on the environment and human health. PAHs have mutagenic and carcinogenic effects (Rogula-Kozłowska et al., 2013; Kozielska et al., 2015; Kuskowska et al., 2018; Rogula-Kozłowska et al., 2018). The strongest carcinogenic compounds are benzo[a]pyrene and dibenzo[a,h]anthracene. Although aerosol health outcomes depend on different factors, including personal exposure and sensitivity, it is important to highlight that PAHs are present in smaller and inhalable aerosol particles, and enter the human body reaching the alveolar ephitelium. Thus, they are potentially serious threat for human health. (Ciecierska and Obiedziński, 2012; Kubiak, 2013; Parol et al., 2014). The biggest impact is observed at the location of their absorption into the body (Kubiak, 2013). Only the target value (1 ng m^–3) for benzo[a]pyrene in the air is specified by the legislation (DIRECTIVE 2004/107/EC), since benzo[a]pyrene is considered a representative substance for the PAHs group. This value is a marker of the carcinogenic risk set out in the Regulation of the Minister of the Environment of Poland from 13 September 2012 on setting limit values of substances in the air (Dz.U. 2012 nr 0 poz. 1032) (Pachurka et al., 2014). Agency for Toxic Substances and Disease Registry classified 17 PAHs as the most dangerous in terms of potential exposure and adverse impact on human health, while the Environmental Protection Agency (EPA) compiled a list of 16 of them, which are treated as the most toxic. This group includes: acenaphten, acenaphthylene, antracene, benzo[a]anthracene, benzo[a]pyrene, benzo[b]fluoranthene, benzo[ghi]perylene, benzo[k]fluoranthene, chrysene, dibenzo[a,h]anthracene, fluoranthene, fluorene, indeno[1,2,3-cd]pyrene, phenanthrene, pyrene, and naphthalene (U.S. EPA, 2003).

Based on previous research (Simoneit, 2015), it is urgent to include triphenylbenzene into the group of toxic and tracked PAHs, as it was proved to be promising marker for plastic and landfills burning emission. There are three possible isomers of triphenylbenzen C₂₄H₁₈: 1,2,3-triphenylbenzene, 1,3,5-triphenylbenzene, 1,2,4-triphenylbenzene. 1,3,5-triphenylbenzene (135TPB) is typically predominant compound, 1,2,4-triphenylbenzene usually present as a minor compound (Simoneit, 2015).

According to Simoneit et al. (2005) from the source emission test based on plastic bags (polyethylene), 63 µg triphenylbenzene per 1 g of smoke were emitted and during the same test with roadside plastic trash 208 µg g^–1 were emitted (Simoneit et al., 2005). Conversely, the tests conducted in the ambient air in Sapporo (Japan) and Gosan Island (Korea) in 2001 showed that 1 g of smoke contains 22 µg and 12 µg of triphenylbenzene, respectively (Simoneit et al., 2004).

In Poland, the problem of high particulate matter concentration is particularly evident in the Malopolska, where the high concentrations of PM are due to the terrain landform, the weather conditions, as well as the presence of large industrial plants emitting pollution directly into the atmosphere (Skiba et al., 2019; Styszko et al., 2019a, b). The highest concentration values are observed in winter, where the PM₁₀ limit (50 µg m^–3-daily) is significantly exceeded each year. Majority of the households, located in suburban and rural areas are heated individually by solid fuels combustion in manually operated boilers, often with low efficiency. In Poland and Czech Republic a new regulation has been adopted to allow environmental inspectors to break privacy and to control the solid combustion byproducts, to inspect burning residues in boilers. Thus, monitoring of 1,3,5-triphenylbenzene (135TPB) has been suggested as a suitable marker tracing burning of the polyethylene plastics (Tomsej et al., 2018; Furman et al., 2019a, b).

The objective of this study is to characterize the chemical composition of atmospheric particulate matter (PM₁₀) and quantify their contributions in the atmosphere of small city, strongly affected by individual residential boilers, as well as industrial emissions (during different seasons). Specific aims of the study are to determine the concentrations of 16 PAHs, 135TPB, ions and OC/EC in particulate matter. This is the first investigation in Poland on the occurrence of 135TPB in atmospheric aerosols in ambient air.

 
2 MATERIALS AND METHODS

2.1 Sampling Site

PM₁₀ samples were collected in the city center of Wadowice (49°88′N 19°49′E, 273 m n.p.m), from February to October 2017. Sampling was conducted during a pilot study characterizing air quality at sites not yet included within the routine air quality network. Due to the limited time period only part of the heating period could be sampled and even higher concentrations might be expected during the remaining months. In total, 82 PM₁₀ samples during heating season (HS, February–April and October) and 72 samples during non-heating season (NHS, May-September) were collected. Wadowice city is located in southern Poland, in Malopolska voivodeship, near to the biggest city of the region-Kraków (771,069 inhabitants). Sampling site is shown in Fig. 1. The population of Wadowice is about 19,000 inhabitants with population density of about 1800 people per sq. km. Each year the city is visited by about 300,000 tourists. Air quality of the sampling area is affected by numerous small coal-fired, low efficiency boilers distributed over the city and the traffic through the city.

Fig. 1. Maps showing the sampling point location in Wadowice.

PM₁₀ samples were collected using a low-volume sampler PNS-15 on quartz fiber filters (Whatman QM-A 47 mm diameter) with 24 h resolution. All quartz fiber filters had been preheated for 6 h at 550 ± 8°C and then kept at temperature 20 ± 1°C and relative humidity 50 ± 5% for at least 24 h prior to weighting and sampling. After sampling, all filters were conditioned for 48 h, weighted with a microbalance (A&D HM-202-EC) up to accuracy of 0.01 mg, and then stored in a freezer at –20°C until analysis. The masses of the filters before and after sampling were obtained as the average of three measurements (PN-EN 12341:2006a; PN-EN 14907:2006b). The meteorological data used in the study (air temperature, wind speed and direction, pressure and precipitation volume) were obtained from the platform: https://www.ekologia.pl/pogoda/polska/malopolskie/wadowice/archiwum and form the data collected by IMGW-PIB (Institute of Meteorology and Water Management): https://dane.imgw.pl/. Details on the meteorological conditions during the study period are summarized in Table 1.

 
3 CHEMICAL ANALYSES

 
3.1 Analysis of Carbonaceous Fraction

Circular punches with 1 cm diameter were analyze to measure the concentration of organic and elemental carbon by the thermal-optical method using the Sunset Laboratory OCEC Lab Aerosol Analyzer and the EUSAAR2 (European Supersites for Atmospheric Aerosol Research) protocol (Cavalli et al., 2010).

Accuracy and repeatability of results were regularly controlled on the basis of the determination of a sucrose solution containing 50 µg of carbon in 10 µL or by using reference filters. Limits of detection for OC and EC were equal to 0.30 µg m^–3 and 0.015 µg m^–3, respectively.

 
3.2 Inorganic Ions Analysis

Concentrations of inorganic anions and cations (NO₃^–, Cl^–, SO₄^2–, PO₄^3–, Na⁺, K⁺, Mg²⁺, Ca²⁺, NH₄⁺) were analyzed with isocratic ion chromatography: two circular punches (ø 8 mm) per filter were extracted under ultrasonic agitation for 40 min, in either 1.5 mL of extra pure water (anions) or in 1.5 mL of the 12 mM methane sulfonic acid (MSA) (cations), in order to determine concentrations of anions and cations, respectively.

Ion chromatography (IC) analysis was performed with a ICS-1100 instrument (Thermo Scientific), equipped with an autosampler AS-DV and ion-exchange columns:

- Ion Pac AS22 (4 × 250 mm) for anions, mobile phase: 4.5 mM Na₂CO₃ + 1.4 mM NaHCO₃,

- CS16 (5 × 250 mm) for cations, mobile phase: 12 mM MSA.

After electrochemical suppression (AERS 500-4 mm and CERS 500-4 mm suppressors), quantification was achieved with a conductivity detector (injection volume 25 µL, flow rate 1.2 mL min^–1).

Calibration was performed against external standards diluted from stock solutions supplied by VWR. The limit of detection (defined as 3 × standard deviation for the field blank samples) of the method for atmospheric samples is presented in a previous publication (Samek et al., 2018).

 
3.3 Analysis of PAHs and 135TPB

The standards of known concentration EPA 525 PAH Mix A (Sigma Aldrich) were used for the calibration of Thermo Scientific GC Trace 1310 gas chromatograph coupled with a ITQ 900 ion trap mass spectrometer and a TriPlus RSH auto-sampler. The mixture contained 16 PAHs in cyclohexane: acenaphtene (Acn), acenaphthylene (Acy), anthracene (Ant), benzo[b]fluoranthene (BbF), benzo[a]anthracene (BaA), benzo[a]pyrene (BaP), benzo[ghi]perylene (BghiP), benzo[k]fluoranthene (BkF), chrysene (Chry), dibenzo[ah]anthracene (DahA), fluoranthene (Flt), fluorene (Flu), indeno[1,2,3-cd]pyrene (IP), naphthalene (Nap), and phenanthrene (Phen) and pyrene (Pyr). 1,3,5-triphenylbenzene (135TPB) and deuterium-labelled PAH, benzo[a]pyrene-d12 (d12-BaP, 98%) were also provided by Sigma-Aldrich (St. Louis, USA). Organic solvents, dichloromethane, cyclohexane, both for GC analysis, were purchased from Avantor Performance Materials Poland S.A.

The isolation and enrichment of PAHs and 135TPB from the filters were performed using solvent extraction. In the first stage, a punch with 22 mm diameter was cut from each filter. In non-heating season, due to low concentration of PM and to obtain concentrations of PAHs above MQL, punches of multiple filters were joined to obtain mean, weekly samples. The individual punch or ensamble of punches were placed in separate vials (20 mL) and flooded with a mixture consisting of 100 µL of internal standard (benzo[a]pyrene D12, 0.50 µg mL^-1) and left under hood for 20 min in order to evaporate the organic solvent. Following, samples were extracted twice with 3 ml of dichloromethane and 2 ml of cyclohexane for 40 min, using a horizontal shaker at 50 rpm. The volume of combined extracts was reduced to 250 µL in a thermoblock (AccuBlock Digital Dry Bath Labnet, Woodbridge, USA), using gentle stream of argon at 35°C. The extract were centrifuged (12,000 rpm) to separate the solid impurities that separated during the reduction of the extract volume. At the end, the concentrates (50 µL) were transferred into chromatographic vials and analysed with GC/MS/MS. The flow of helium through a GC column was constant and set at 1 mL min^–1. The programmable temperature of the vaporization injector was maintained at 250°C, the transfer line at 250°C and the ion source at 250°C. The injector was operated at splitless conditions for 2 min, then turned to the split mode at the ratio of 50:1. Volume of injections was 1 µL. Separations was performed on a Rxi®-5Sil MS capillary column from Restek that had 30 m x 0.25 mm inner diameter and film thickness of 0.50 µm (5% diphenyl/ 95% dimethylpolysiloxane). The temperature program for PAHs analysis was as follows: 70°C for 2 min, from 70°C to 150°C at 20°C min^–1, to 300°C at 10°C min^–1, and finally 13 min at 300°C.

The analyses were performed in positive mode, with electron impact ionization at 70 eV and 250 µA emission current. Helium (99.999%) was used as a carrier gas at 0.3 mL min^–1. Mass spectrometry analyses was performed in the multiple reaction monitoring (MRM) mode detecting the fragmentation of the precursor ions. The choice of fragmentation products for each substance was based on the most intense signal. For the acquisition of the MRM transitions the analytical run was split into time windows considering the expected retention time of the selected compounds. Data were collected, analysed and processes using Xcalibur® software.

Specific and intense product ions of each target analyte were used for quantification and a secondary product ion was used as qualifier ion for confirmatory purposes. The retention times, characteristic ions of tested analytes and validation parameters of the method are shown in Table 2. Data quality was determined based on limits of detection (LOD), limits of quantification (LOQ), and the linearity of the calibration line that was estimated by analyzing the 7 points calibration curves (6–7500 ng mL^–1 for PAHs and 5–5600 ng mL^–1 for 135TPB). The calibration curves of all the organic compounds were highly linear (0.994 < R² > 0.999). Per-deuterated benzo[a]pyrene d12-BaP was also added to each level of standards as an internal standard to account for components losses due to volatilization or instrument bias in standards. For further quality assurance, known amount of mixed standards were spiked on blank filters and analyzed in the same steps as samples. Blank samples analysis showed none of the PAHs and 135TPB of interest. The percent recoveries for all the target compounds were in the range of 78–95%. Recoveries corrections were made to the measured concentrations. The accuracy of 95% was determined by the error obtained between the mean values of replicates of the standard solution 10 ng mL^–1 taken as reference. The precision of 5% was calculated be means of percent relative standard deviation (% RSD).

 
4 RESULTS AND DISCUSSION

 
4.1 Concentrations and Chemical Composition of PM₁₀

The mass concentrations of PM₁₀ collected during the heating season (HS, February–April and October) were higher than during non-heating season (NHS, May–September), as expected. The average PM₁₀ HS concentrations was 43.3 µg m^–3, while the average NHS concentration was 27.1 µg m^–3. PM₁₀ concentrations during HS varied between 10.9 and 116.8 µg m^–3 and the range between the lowest and highest concentration was lowerin the NHS (10.8–53.7 µg m^–3). The highest concentration in the NHS season was noted in September. The variation of PM₁₀ concentrations is reported in the Fig. 2.

Fig. 2. Time series of monthly average PM₁₀ concentrations during the sampling period.

Mean concentrations of PM₁₀ chemical components during HS and NHS are listed in the Table 3. The analysed components accounted on average for 81% and 88% for the NHS and HS PM₁₀ mass, respectively. Carbonaceous fraction (basen on organic and elemental carbon, OC+EC and the conversion of OC to organic matter (OM) as explained later) accounted for the highest share in both NHS (36%) and HS (42%) PM₁₀ mass. The average contribution of inorganic ions varied from 34% to 42%, in HS and NHS, respectively, while the absolute sum of concentrations were 15 µg m^–3 and 12 µg m^–3. The highest concentration was observed for Na⁺ among cations and SO₄^2– among anions. It should be noted that average concentrations of SO₄^2– was similar in HS and NHS, 5.6 µg m^–3 and 5.3 µg m^–3, respectively. However, average concentrations of NO₃^– and Cl^–, were up to 3 times higher in HS in comparison to NHS, due to the lower ambient temperatures that avoid ammonium nitrate volatilization.

 
4.2 Chemical Mass Closure

Chemical characterization of the main components of aerosols showed that carbonaceous components and inorganic ions exhibited the highest share in the collected aerosols, in both seasons. Similarly, Fig. 3. shows the chemical composition of PM₁₀ collected in Wadowice in form of chemical mass balance. Organic matter (OM) mass, i.e., the mass of organic carbon and the associated heteroatoms (H, O, S, N, etc.), was estimated on the base of OC concentrations multiplied by a factor of 1.6, as proposed for urban environments (Putaud et al., 2004). The OM/OC ratio is still a matter of discussion in recent atmospheric studies. Previous observations showed already a wide possible range for the OM/OC ratio, between 1.4–2.1 depending on the aerosol type (Turpin and Lim, 2001; Sandrini et al., 2014). More recent techniques based on the high resolution aerosol mass spectrometry (HR-AMS), which allows the direct measurement of O/C; H/C and N/C ratios confirmed the previous findings reporting the range 1.6–2.1 for Mexico City (Aiken et al., 2008) and an average OM/OC of 1.56 for a traffic-related site (Brown et al., 2013). The lowest OM/OC ratios are usually seen for fresh combustion aerosols (hydrocarbon-like organic aerosols), while the highest values are usually associated with aged organic aerosols and reflect the presence of oxygenated functional groups formed during atmospheric transport. Higher ratios are indeed observed at background sites or at sites affected by temperature inversion (Turpin and Lim, 2001; Malm and Hand, 2007). Here we use the value of 1.6 as the lower bound of the ratio in areas strongly impacted by primary sources. The OM concentration estimated represents the lower bound of the actual OM. The contribution carbonaceous aerosols and inorganic components was higher in the HS, which can be attributed to the enhanced emissions from residential combustion and traffic (the difference between HS and NHS was statistical significant in terms of T-Test at p < 0.001) and meteorological conditions that prevent pollution dispersion (Witkowska et al., 2016; Zimnoch et al., 2020). OC was dominant in both seasons-18% in the NHS and 34% in the HS. EC mass ranged between 5% in the NHS up to 8% in the HS.

Fig. 3. Average chemical composition of particulate matter for two seasons.

Ionic balance (expressed as ratio of equivalent concentrations of anions to cations) was calculated to check for any indication that not all ions were covered by the direct analysis. During the HS and NHS the ionic balance values were 1.1 and 1.2, respectively. Nitrates, sulfates and ammonium ions, forming the secondary inorganic aerosols (SIA), dominated the inorganic components in PM₁₀ aerosols collected in Wadowice. SIA accounted for 18% of PM₁₀ mass in NHS and 21% in HS. NaCl-calculated based on either Na⁺ or Cl^–, depending which of those species were less in the equivalent concentration, contributed on average to 2% in NHS and 6% in HS. Chlorine containing compounds can be emitted during coal combustion (McCulloch et al., 1999) and it cannot be excluded that the observed chloride originated from coal combustion. In previous studies conducted in Krakow an excess of Cl^– ions in comparison to the equivalent Na⁺ concentration was found during the heating season (Junninen et al., 2009; Styszko et al., 2015). In addition, it was seen that the concentration of the excess Cl^– ions correlated with particulate Mercury, which was a hint for the coal-burning origin. In this study the excess of chloride was observed only in nine samples, mainly those collected in the HS. The amount of chloride not neutralized by Na⁺ ranged between 0.4 and 9.6 µg m^–3. The highest concentration was very far from the average (2.8 µg m^–3) and was noted during the NHS (September), pointing rather to a single local event. If this high value is excluded, a positive correlation (R² = 0.66) between the excess of Cl^– and PM₁₀ concentration can be observed. The small number of samples with this phenomenon does not allow the recognition of other significant relations.

Not identified PM₁₀ fraction might include other than analyzed inorganic ions and trace metals, as well as oxygen atoms from their oxides. Moreover, a certain amount of water associate with hygroscopic components (mainly salts) should be accounted. The further part might originate from underestimation of organic matter, derived from the assumed OM/OC ratio. Based on mineral dust and ions analysis, certain seasonality can be expected also for the missing categories. Aerosol bound water will be more relevant during HS, as considerably more ions were found during this time. Oxygen atoms will play a significant role (on the relative scale) during NHS, as the mineral fraction was higher during that time. Not identified fraction was 12% and 19% for HS and NHS, respectively. It is similar to values reported elsewhere (Hueglin et al., 2005), the average sum for those components was around 19%.

 
4.3 Particulate Matter Sources

Chemical groups presented in mass closure can be originating from mixed sources. Therefore, the trends of a few source characteristic chemical tracers have been analyzed.

The variability of ratios of organic carbon to elemental carbon (OC/EC) reflects different carbon sources and atmospheric processing. Elemental carbon is mainly emitted from incomplete combustion of fossil fuels, biomass burning and other carbon containing materials. Because organic carbon can originate from fresh emissions (described as primary organic carbon), as well as can be formed through photochemical activity involving gaseous or liquid phase precursors (described as secondary organic carbon)-we analyzed the OC/EC ratio. According to literature (Sillanpää et al., 2006; Samara et al., 2014; Ji et al., 2016; Liu et al., 2016), a low OC/EC ratio can be related with fresh sources of traffic aerosol (2.2 and 0.8 for light-duty gasoline and heavy-duty diesel vehicles, respectively) (Na et al., 2004; Pio et al., 2011; Samara et al., 2014), while higher ratios are rather expected for heating sources (coal combustion: 2.6–6.0; wood combustion: 4.15; natural gas: 12.7), forest fires (14.5), dust from paved roads (13.1) (Na et al., 2004), and secondary organic carbon (winter-17% of total OC, summer-65% OC (Na et al., 2004)). During this study the average OC/EC ratios were 3.7 and 3.8 for NHS and HS, respectively. Ranges of OC/EC in both seasons were very close, 2.1–6.5 in NHS and 1.8–6.5 in HS. The relatively high ratios can be explained by the presence of local sources like wood and coal combustion, as well as contribution of secondary organic carbon, especially during the NHS. Previous studies reported similar OC/EC ratios during HS and NHS in areas affected by wood burning during the cold season ( Cao et al., 2007; Samek et al., 2020, 2017; Sandrini et al., 2014; Viana et al., 2006). Similarly, Cao et al., (2007) observed similar OC/EC ratios in summer and winter (average about 4) in areas impacted by coal combustion emissions. Ranges of OC/EC in both seasons were very close, 2.1–6.5 in NHS and 1.8–6.5 in HS. The variations of OC/EC obtained for each sample show that local combustion sources are most relevant for the investigated urban area. High correlations (R² from 0.685 in NHS to 0.712 in HS) were observed (Fig. 4).

Fig. 4. Scatterplots of OC and EC concentrations and linear regression for two seasons.

The mass ratio of sulfate to nitrate has been used as an indicator of the relative importance of stationary vs. mobile sources of sulfur and nitrogen in the atmosphere. The equivalent ratios of sulfate to nitrate were 5.7 in NHS and 3.7 in HS; therefore, the stationary source emissions, particularly coal burning, were important for SIA formation in these area (Xiu et al., 2005). In addition during the HS we observed high relative humidity and low temperature, which favored sulfate formation in the heterogeneous phase through cloud processing. Much lower concentration of NO₃^– during NHS could be driven by the equilibrium between nitrate and nitric acid moved to gas phase due to higher temperatures. The high concentrations of SO₄^2– during sampling period, average above 5 µg m^–3 in HS and NHS (see Table 3) is caused by the coal combustion through whole year for residential, water heating and industry (power station) (Baykara et al., 2019; Błaszczak et al., 2020; Dai et al., 2019; Kozáková et al., 2018).

 
4.4 Concentration Trends and Distribution of PAHs and 135TPB in PM₁₀

The monthly variations in the total PAHs, 135TPB and PM₁₀ mean concentrations, plotted in Fig. 5. shows a strong variability of the target compounds throughout the sampling period. As was expected, the highest concentrations were observed in the HS.

 Fig. 5. Variation of the average concentrations of the PAHs and 135TPB throughout the sampling period. Bars represent standard deviations between sampling days.

The total concentration of the PAHs followed closely the 135TPB concentration, throughout the sampling period. Table 3 provides monthly minimum, average and maximum concentrations of PAHs and 135TPB analyzed in collected samples from February to October 2017. Concentrations of determined PAHs varied from 0.08 ng m^–3 to 25 ng m^–3 in whole sampling period and average value of 1 ng m^–3 in NHS and average value 3.6 ng m^–3 in HS. The more volatile Naphthalene was not detected, because it is mostly partitioned into the gas phase. The concentration of 135TPB varied from 0.3 ng m^–3 to 2.6 ng m^–3. The average concentration of 135TPB in NHS was 0.3 ng m^–3 and 0.8 ng m^–3 in HS.

For both periods the heavier 4 and 5 ring PAHs accounted for almost 80% (90% for 11 PAHs in Siudek and Frankowski (2018)) of the total amount of PAHs. The lighter PAHs (3 rings) contribution was up to 7% (6% for 11 PAHs in Siudek and Frankowski (2018), due to extensively partitioning in the atmosphere between the gas and particulate matter Fig. 6.

Fig. 6. Ring number distribution of PM₁₀ associated PAHs during heating and non-heating seasons.

Fig. 7 shows the changes in the mean concentrations of PAHs and 135TPB during heating and non-heating seasons. Only benzo[ghi]perylene's concentrations are similar or higher in the NHS than in the HS (Fig. 7). A relationship profile was created between the concentrations of all PAHs and 135TPB relative to the Pearson coefficient. This coefficient takes values from –1 to 1, respectively: < 0.2-no linear relationship, 0.2–0.4-weak relationship, 0.4–0.7-moderate relationship, 0.7–0.9-strong relationship, > 0.9-very strong relationship. The values of the coefficient for PAHs are shown in Table 4, with the exception of compounds that were not present in some months. The highest relationship (the highest value of Pearson coefficient) was observed for fluorene (0.96), indeno[1,2,3-cd]pyrene (0.89) and benzo[k]fluoranthene (0.86).

Fig. 7. The mean concentrations of PAHs and 135TPB in heating and non-heating seasons.

According to Simoneit (2015) 135TPB was found in both types of plastic materials that had been analyzed. Studies have concluded that PAHs, 135TPB and other unknown compounds are derived from PET combustion (Tomsej et al., 2018). Co-combustion of both plastics with fuels resulted in a higher abundance of fluorene and most of 4–6 ring PAHs, including indeno[1,2,3-cd]pyrene and benzo[k]fluoranthene. The same relationship between fluorene and 135TPB (Pearson 0.96-very strong relationship) was observed in our studies carried out in Wadowice area, where the main sources of pollution are transportation and residential heating. The presence of 135TPB in analysed samples confirms the assumption that plastic waste is burned in the studied area together with other fuels. Due to this fact,135TPB can be used as a good PET combustion marker in air quality monitoring programs.

Fig. 8. shows the dependence of 135TPB concentration on temperature. Increased concentrations of 135TPB were noted for lower temperatures (the trends are the same as for PM₁₀ yearly concentration-higher particulate matter concentrations at low temperatures promote higher sorption of compounds), which may confirm that in densely populated areas, plastics are co-burned with fuels in residential furnaces. On the other hand, the higher velocity of wind speed contributed to the decrease in 135TPB concentration regardless of wind direction, Fig. 9. The analysis did not take into account the non-heating period due to many days when 135TPB was below detection limit.

Fig. 8. Trend between concentration of 135TPB and temperature through the heating season (Some days were not presented due to the value of 135TPB below LOQ).

Fig. 9. Trend between concentration of 135TPB and wind speed through the heating season (Some days were not presented due to the value of 135TPB below LOQ).

4.5 Diagnostic Ratios

Ambient concentration of specific air pollutants depend on their emission sources. Those pollutants that are sufficiently stable in the atmosphere and are associated to one source can be used as tracers of that specific source. For example, PHE, FLU and PYR are tracer of coal combustion, BaP and FLU are emitted during wood combustion, whereas FLU, PYR, BbF and BkF are characteristic for diesel engines exhausts. Diagnostic ratios (DR) are ratio between tracers’ concentrations. Although diagnostic ratios should be used with caution (Galarneau et al., 2008), their analyses give insights into the identification of particulate matter sources and their relative contribution (e.g., diesel/gasoline engine exhaust gases, biomass/wood/coal combustion) (Yunker et al., 2002). Table 5 presents characteristic DRs from various sources (Finardi et al., 2017; Kulshrestha et al., 2019).

Table 6 summarises the results obtained for the DRs within this study. The analysis assumes the threshold of 50% as an indicator of the effective determination of the source of pollution. In Table 6, only the ratio of FLU/(FLU + PYR), FLT/(FLT + PYR), BaP/(BaP + CH) and BaA/(BaA + CH) can be taken into account because their percentage ratios represent more than the set threshold. 

The ratio of FLU/(FLU + PYR) indicates petrol combustion as an air pollution source for all measured days in every month taken into consideration. From the FLT/(FLT + PYR) ratio the sources were identified mostly as a fuel combustion. It is worth to note that similar observations for FLT/(FLT + PYR) were reported in other studies (Simoneit, 2015) for areas mainly characterized by the intense transport activities (cars, motorcycles, etc.) and the cities households were mainly heated using wood or coal (Kulshrestha et al., 2019). The BaP/(BaP + CH) ratio is used to identify gasoline and diesel combustion. In addition, at residential area during the heating season it allows the identification of firewood and oil from heating radiators as a source of pollution (Manoli et al., 2004). BaP/(BaP + CH) in this study indicates the dominance of household heating pollution sources in the winter. In non-heating period sources from road transport dominated. Values of BaA/(BaA + CH) indicates that sources of pollution include vehicles transport (Alves et al., 2017). Other authors reported discordant values, e.g., 0.5 for diesel vehicles and 0.73 for gasoline combustion (Khalili et al., 199; 5Finardi et al., 2017). BaP/(BaP + CH) has been considered as household heating pollution sources in the winter and road transport in non-heating period. However, the IP/(IP + BghiP) ratio was not taken into account due to the low percentage of days on which the value of the ratio overlapped the numbers presented in Table 5.

DahA came out more often in lower concentrations than BghiP and ideno[1,2,3-cd]pyrene, but the DahA concentration values are much higher during the days when transport dominates (which was also confirmed in this work in Table 6). An example of this is the work of Kozielska et al. (2015) as well as Siudek and Frankowski (2018) where the DahA values exceed the concentrations of BghiP and indeno[1,2,3-cd]pyrene (studies conducted also for Poland-Katowice, Złoty Potok, Poznań).

To investigate the impact of long range transported emissions on Wadowice, and especially industrial emissions from Krakow, we analysed air back-trajectory using the Hybrid Single-Particle Lagrangian Integrated Trajectory model (HYSPLIT), developed by NOAA’s Air Resources Laboratory. The created pollution transport maps Fig. 10 show the possible trajectory and directions of inflow air masses (8 h backwards, height AGL: 100 m (red line), 300 m (blue line), 500 m (green line)). 

Fig. 10 reports back trajectories calculated for the days when the highest PM₁₀ concentrations were recorded over the entire measurement period during the HS (95.0 µg m^–3 on 09.03.17, Fig. 10(A), 116.8 µg m^–3 on 13.03.17, Fig. 10(B), 101.5 µg m^–3 on 27.03.17, Fig. 10(C), 61.4 µg m^–3 on 05.04.17, Fig. 10(D), 60.0 µg m^–3 on 21.04.17, Fig. 10(E)) and NHS (45.3 µg m^–3 on 01.06.17, Fig. 10(F), 33.3 µg m^–3 on 25.08.17, Fig. 10(G), 107.3 µg m^–3 on 20.09.17, Fig. 10(H)). The maps indicate mostly W, NW, SW and N directions of pollution inflow. NE (Fig. 10(B)) wind directions can possibly bring pollution masses from Krakow-the nearest biggest city (over 771,000 citizens). Krakow is located in a basin stretching in the Vistula Valley. In locations of this type, pollution sources are very often concentrated in the city centre, and the long-term movement of air masses leads to its transport to other areas. Examples of PM₁₀ emissions sources from the N direction are coal-fired power plant (Siersza), refinery (Trzebinia) and metallurgical factories (Bukowno, Bolesław, Olkusz) located on the pollution transport route shown in the above graphs. An additional aspect affecting the air quality are domestic furnaces used in transport areas, where solid and high-quality fuels can be burned, as well as waste not intended for this type of furnace. Trajectories from SW/W/NW can be affected by emissions from industrial areas located in Silesia. In fact, this region is well known for many hard coal mines and high density of population. Analyzes of archival data (Environmental Protection Inspectorate in Krakow and Katowice) regarding PM₁₀ concentration in backward trajectory areas, show that high concentrations of PM₁₀ on a particular day in Wadowice coincided with high PM₁₀ concentrations appearing in trajectory areas up to 24 h earlier.

Fig. 10. The air masses trajectory and directions during selected days of the measured period: A-09.03, B-13.03, C-27.03, D-05.04, E-21.04, F-01.06, G-25.08, H-20.09. 

The possible trajectory and directions of inflow of 135TPB were presented below (Fig. 11). 135TPB trajectory analysis in Fig. 11 shows the days with the highest concentration of the plastic burning marker: 1.9 ng m^–3 04.04.17 (Fig. 11(A)), 2.0 ng m^–3 on 09.04.17 (Fig. 11(B)), 2.2 ng m^–3 10.04.17 (Fig. 11(C)), 2.3 ng m^–3 on 01.10.17 (Fig. 11(D)), 2.6 ng m^–3 on 16.10.17 (Fig. 11(E)). The maps show mainly the SW direction, which is an area with high population density, characterized by very high air pollution (Bielsko Biała, Szczyrk, Żywiec). However, the NE wind direction is also possible (see Fig. 11(A))-these regions are well know from aircraft industry and furniture factories (Mielec) and chemical industry (Tarnow).

Fig. 11. The 135TPB trajectory and directions during selected days of the measured period: A-04.04, B-09.04, C-10.04, D-01.10, E-16.10. 
 

4.6 Health Risk Assessment

From a perspective of human health, the total concentrations of benzo[a]pyrene was above the recommended guideline concentrations of 1 ng m^–3 (DIRECTIVE 2004/107/EC). The relative impact of the PAH mixture on human health can be determined by different types of indicators such as the carcinogenic equivalent (CEQ), mutagenic equivalent (MEQ), and toxic equivalent (TEQ) (Kozielska et al., 2015). The results for analysed months are presented in Table 7. whereas the patterns (1)–(3) for the calculations of MEQ, CEQ, TEQ are shown below:

Compare to Kozielska et al. (2015) data, the results obtained in Wadowice presented in this article can be classified as alarmingly high against the background of Europe. For instance, the results obtained for winter 2009 in Madrid (Spain) (Mirante et al., 2013) in the urban background are significantly lower than winter months measurements in Wadowice: 0.11 ng m^–3, 0.12 ng m^–3, 0 ng m^–3 (for MEQ, CEQ and TEQ, respectively). Going further, to the results from Florence (urban background, Italy) (Martellini et al., 2012) in the cold season 2009–2010 were lower than in summertime in Wadowice: 2.17 ng m^–3, 5.43 ng m^–3, 0.02 ng m^–3 (for MEQ, CEQ and TEQ, respectively). Among previously investigated cities in Europe, higher equivalent levels were obtained in winter 2010 in Brno (Czech Republic) (Křůmal et al., 2013): 5.85 ng m^–3, 8.71 ng m^–3 and 0.02 ng m^–3 (for MEQ, CEQ and TEQ, respectively), although they are still significantly lower than results presented for winter period in Wadowice. There are no strictly defined standards for MEQ, CEQ and TEQ values. That coefficients are used for comparison reasons only. However, these values should be as low as possible.

 
5 CONCLUSION

This work presents a detailed chemical characterization of PM₁₀ in Wadowice during period of 9 months. As sampling was not performed during November to January even higher concentration values might occur during these winter months. The analyzes showed higher concentrations of PM₁₀, PAHs, OC, EC and ions in the colder measurement period than in the warm one. The highest concentrations of particulate matter were recorded on the days when the wind was mostly from W, NW, SW and N directions of pollution inflow. Trajectory analysis suggests seasonal inflow of pollution from industrial areas of Poland. Other pollution sources impacting the measurement area include road transport, coal burning, fuel combustion, smelters. In the case of carbon analysis, organic carbon accounted for the largest fraction of PM₁₀ mass as stated in this work, OC may contain polycyclic aromatic hydrocarbons (PAHs) and other organic constituents that have potential mutagenic, teratogenic and carcinogenic effects. The method used for PAHs analysis did not allow to estimate the concentration for anthracene and naphthalene, due to the low concentration of these compounds in the particle-phase. Nevertheless, larger PAHs were determined and showed elevated concentrations. Strong correlation between PM₁₀, PAHs and emerging contaminants (135TPB) has been noted. The highest relationship (the highest value of Pearson coefficient) was observed between 135TPB and fluorene (0.96). 135TPB is considered as a PET plastic waste incineration marker. The implementation of this type of markers into the atmospheric air monitoring program can help in estimating the contribution of PET combustion in furnaces not adapted for the disposal of plastic waste. Analysis of emerging contaminants is useful tool for identifying sources of pollution, their elimination and air composition control. Health risk assessment showed that mutagenic, carcinogenic, and toxic equivalent (MEQ, CEQ, TEQ), calculated based on PAHs concentrations, are significantly higher than the values observed in other European urban areas. High MEQ, TEQ and CEQ values indicate a high content of carcinogenic compounds, which are a serious problem for human health and life. Purposive actions should be taken immediately to reduce the concentrations of harmful compounds in the atmosphere, and above all to eliminate sources of pollution.

 
ACKNOWLEDGEMENTS

This research was partially financed by the AGH UST grant 16.16.210.476 subsidy of the Ministry of Science and Higher Education. This work was partially funded by the Rector's Grant for the RedoX Students’ Research Group No. AGH90/ Grant/2019. This article is based upon work from COST Action COLOSSAL (CA16109) supported by COST (European Cooperation in Science and Technology). The infrastructure of the AGH Center of Energy in Kraków was applied in order to determine the concentration of ions. Authors thank Krakow Smog Alert for the help in the collection of samples. PF and AS have been partly supported by the EU Project POWR.03.02.00-00-I004/16.