Analysis of Particulate Matter Concentration Variability and Origin in Selected Urban Areas in Poland

The work presents the results of research and analyses related to measurements of concentration and chemical composition of three size fractions of particulate matter (PM), PM10, PM2.5 and PM1.0. The studies were conducted in the years 2014–2016 during both the heating and non-heating season in two Polish cities: Wrocław and Poznań. The studies indicate that in Wrocław and Poznań, the highest annual concentrations of particulate matter (PM1.0, PM2.5, and PM10) were observed in 2016, and the mean concentrations were respectively equal to 18.16 μg/m3, 30.88 μg/m3 and 41.08 μg/m3 (Wrocław) and 8.5 μg/m3, 30.8 μg/m3 and 32.9 μg/m3 (Poznań). Conducted analyses of the chemical composition of the particulate matter also indicated higher concentrations of organic and elemental carbon (OC and EC), and water-soluble ions in a measurement series which took place in the heating season were studied. Analyses with the use of principal component analysis (PCA) indicated a dominating percentage of fuel combustion processes as sources of particulate matter emission in the areas considered in this research. Acquired results from these analyses may indicate the influence of secondary aerosols on air quality. In the summer season, a significant role could be also played by an influx of pollutants—mineral dust—originating from outside the analyzed areas or from the resuspension of mineral and soil dust.


Introduction
In the last few decades, concentrations of sulfur dioxide or nitrogen oxide within Europe, including Poland, decreased significantly [1]. However, there are many regions in Europe where standards established for particulate matter (PM) are still being exceeded, which negatively influences the health of people inhabiting those regions. Undesired effects are brought forth even by the mere presence of fine particles in the organism and in addition, their toxic influence is multiplied by heavy metals and polycyclic aromatic hydrocarbons, which are their carriers [2][3][4][5][6][7].
According to the European Environment Agency (EEA), in 2018, the standard permissible limit of days (in which the 24 hour average value of PM 10 concentration is higher than 50 µg/m 3 (the so called acceptable frequency of exceedance)), was exceeded in as many as 19 among 28 member states (19% of all measurement stations) [8]. The highest air pollution levels in Europe are observed in the eastern The measurement point in the studies representing urban background was stationed in the Biskupin estate in Wrocław at Kosiby 6/8 Street (Figure 1, Station 1) located in the southeastern part of the Śródmieście district, which encompasses a large number of green areas. The surroundings of the research station are an area of diverse character, which include parks, garden plots, low-rise buildings (semi-detached houses, villas) and high-rise buildings (blocks of flats). The measurement point was located about 5 km (in a straight line) from an electrical power and heating plant, about 0.1 km from the closest multi-family development (multi-family tenement house), about 180 m from another multi-family development (blocks of flats), and about 130 m from a road, whose traffic included passenger cars, trams, buses and bicycles.
The remaining measuring points are stations belonging to the VIEP in Wrocław and represent stations of urban background (Station 2, 3 and 5) and a station considered a road traffic station (Station 4). Stations 2 and 3 are located far from main traffic routes and are surrounded by green areas and multi-family dwellings. Station 4 at Wiśniowa Alley is located close to the main communication route in Wrocław, the intersection of Haller Street and Wiśniowa Alley, and is also located in the neighborhood of multi-family housings.
In the case of studies that took place in Poznań, two measurement campaigns were carried out: in the fall-winter season (25 October-22 November 2016) and in the summer season (5 June-2 July 2017) at three measurement stations (Figure 1, . Two measurement stations (at Szmanowskiego Street and at Polanka Street) belonged to the Provincial Environment Protection Inspector in Poznań; these were used as urban background stations. The station at Szymanowskiego Street (Station 6) borders with service and multi-family The measurement point in the studies representing urban background was stationed in the Biskupin estate in Wrocław at Kosiby 6/8 Street (Figure 1, Station 1) located in the southeastern part of thé Sródmieście district, which encompasses a large number of green areas. The surroundings of the research station are an area of diverse character, which include parks, garden plots, low-rise buildings (semi-detached houses, villas) and high-rise buildings (blocks of flats). The measurement point was located about 5 km (in a straight line) from an electrical power and heating plant, about 0.1 km from the closest multi-family development (multi-family tenement house), about 180 m from another multi-family development (blocks of flats), and about 130 m from a road, whose traffic included passenger cars, trams, buses and bicycles.
The remaining measuring points are stations belonging to the VIEP in Wrocław and represent stations of urban background (Station 2, 3 and 5) and a station considered a road traffic station (Station 4). Stations 2 and 3 are located far from main traffic routes and are surrounded by green areas and multi-family dwellings. Station 4 at Wiśniowa Alley is located close to the main communication route in Wrocław, the intersection of Haller Street and Wiśniowa Alley, and is also located in the neighborhood of multi-family housings.
In the case of studies that took place in Poznań, two measurement campaigns were carried out: in the fall-winter season (25 October-22 November 2016) and in the summer season (5 June-2 July 2017) at three measurement stations (Figure 1, Stations 6-8).
Two measurement stations (at Szmanowskiego Street and at Polanka Street) belonged to the Provincial Environment Protection Inspector in Poznań; these were used as urban background stations. The station at Szymanowskiego Street (Station 6) borders with service and multi-family developments from the west, with multi-family buildings from the north and south and from the west and southwest directions, it is bordered by Podolany (a district of Poznań), which is characterized by single-family developments. At Podolany, only a manual measurement of PM 10 particulate matter is carried out. The station at Polanka Street (Station 7) is surrounded by multi-family buildings, heated by local boiler rooms. Close to the north, there is also a man-made lake called Jezioro Maltańskie. From the east, there is a multi-family development heated by a district heating system and from the southwest, there is a large shopping mall. In the distance, about 800 m in the southeastern direction, there is a point emission source-the Malta-Décor company-which deals with the production of decorative paper. This work used the filters from the manual collection of PM 2.5 from this station. Our own measurement station was located at Jana Pawła II 10 Street (Station 8). It is a divided street with separated tram trackage between the lanes. This segment is usually gridlocked. The selected location was supposed to reflect the influence of emissions from road traffic on air quality in this area. The area is directly bordered by the street. The device was installed at a distance of about 50 m from the street, on a green terrain belonging to the Poznan Supercomputing and Networking Center (PCSS). From the west, the station bordered directly with the street and right behind it is the Jezioro Maltańskie lake. From the north, the station bordered with green terrains, from the west with the Cybina river and from the south with the Poznan University of Technology Campus. At the station, the measurements were conducted manually by collection of three particulate matter fractions: PM 10 , PM 2.5 and PM 1.0 .
The dates of measurements were intentionally planned for the heating season and the non-heating season in order to compare the chemical composition of particulate matter in two thermally different periods.

Collection of Samples and Determination of Particulate Matter Concentration
In order to collect PM 1.0 , PM 2.5 and PM 10 samples, Harvard type impactors (MS&T Area Samplers, Air Diagnostics and Engineering, Inc., Harrison, ME, USA) were used. The airflow was forced by ultra-silent, oil-free vacuum pumps (Air Diagnostics and Engineering, air sampling pump, SP-280E model, Harrison, USA). Quartz fiber filters-Whatman QM-A (Wathaman Healthcare UK Limited, Amersham Place, Little Chalfont, Buckinghamshire, UK)with diameter of 37 mm were used as a substrate. The flow rate during sample collection was respectively equal to: samples ≤1.0 µm-23 dm 3 /min, samples ≤2.5 µm and ≤10 µm-10 dm 3 /min. The volume of pumped air was controlled by an Actaris type flowmeter. The time of sample collection was 24 h. The measurement of particulate matter was carried out with use of the gravimetric method in accordance with the PN-EN 12341: 2014 standard.

Analysis of OC and EC Content in Particulate Matter Samples
In order to determine organic carbon (OC) and elemental carbon (EC) in particulate matter samples, a thermo-optic analyzer of organic and elemental carbon (Model 4L Main Oven Assembly, Sunset Laboratory Inc., Tigard, OR, USA) with flame ionization detection (FID) was used. Samples of particulate matter, in the form of quartz filter specimens with an area of 1 cm 2 , were subjected to an analysis composed of two stages. First, by the gradual heating of the sample in a stream of helium, the organic fraction (OC) was released from the filter. In the following stage, a specimen of the filter was heated in an oxidizing mixture of helium and oxygen in order to release elemental carbon (EC). The heating of the examined samples progressed in accordance with the parameters determined in a specific measurement protocol. Compounds separated from the sample during subsequent heating stages in the furnace, in oxidizing conditions and with the presence of magnesium oxide, were transformed in stoichiometric ratio in relation to CO 2 . Carbon dioxide produced during the analysis was further reduced in a methanizer, in the presence of a hydrogen and a nickel catalyst, to methane, which was measured with the use of FID. Based on transmittance measurement which was done during the thermo-optic analysis with the use of a red-light laser beam (thermal optical transmittance (TOT)), the corrections for the determined contents of OC and EC were made. The last stage of the measurement process was the inclusion of a fixed volume loop in order to inject an external standard at the end of each analysis, which enabled us to provide optimal and stable conditions for the measurement, increasing the repeatability of the whole analytical method. With the use of the carbon analyzer, it is possible to take measurements in accordance with several standard temperature protocols, which differ from each other in the values of temperature thresholds, their number and duration. The analysis for this work was carried out using the "EUSAAR_2" protocol, which was designed as the suggested standard method for European measurement stations as a part of a European Supersites for Atmospheric Aerosol Research (EUSAAR) project.
Certified gases from Air Liquide (Cracow, Poland) were used to conduct analyses. Validation parameters for the methods used are as follows: 1.

Analysis of the Content of Ions Soluble in Water in Samples of Particulate Matter
Determination of water-soluble ions in samples of particulate matter deposited on filters consisted of extracting anions and cations from particulate matter and in deionized water and later determining their content in water extracts. The extraction of ions was performed with the use of a vortexer (12 h) in a cold room (≤18 • C) at 60 cycles per minute. Moreover, before the agitation process, the samples of particulate matter were subjected to ultrasound for 1 h in a water bath (at ≤27 • C). After the agitation process, acquired extracts were cleansed with the use of micropore polyethersulfone (PES) syringe filters.
The content of ions acquired from water extracts was determined with the use of the ion chromatography method. In order to determine anions (F − , Cl − , NO 3 − , PO 4 3− , SO 4 2− ) and cations (Na + , NH 4 + , K + , Ca 2+ , Mg 2+ ) in samples of particulate matter, authors used chromatographs of Dionex ICS 1100 type (Thermo Scientific company, USA) with a dual piston serial pump, automatic electrolytic suppression, and vacuum eluent degassing before the guard column and thermostating columns. Chromeleon ® SE (version 7, Thermo Scientific company, USA) dedicated software was used in order to operate the chromatograph. Only reagents with purity of at least p.a. grade were used in the analysis and the solutions for the preparation of calibration curves and control samples were done based on LGC standards of known uncertainty and concentration equal to 1 g/l.
Limits of detections were at the following levels: 0.01 mg/dm 3

Analysis of Main Components
This work presents the results of the PCA analyses carried out based on the determination of concentrations of organic carbon, elemental carbon, the following anions: In the case of Poznań, the analysis was performed for two measurement stations with separation for two thermal seasons. non-heating season (April-September) are presented in Figure 2. Pursuant to the regulation of the Minister for Environmental Protection [22], the permissible annual average concentration of PM 2.5 in the air is equal to 25 µg/m 3 ; in 2014 the applicable margin of tolerance was equal to 1 µg/m 3 . In the case of PM 10 particulate matter, the law sets out two permissible values: annual average value equal to 40 µg/m 3 and permissible number of days with 24 h average value exceeding 50 µg/m 3 . In accordance with the numbers described above, in 2014-2016, the analysis of data indicates that the standard for PM 2.5 particulate matter was exceeded three-fold in Wrocław (each time at the station at Wiśniowa Alley-Station 4). In Poznań, the limits were not exceeded. However, the annual average values were closing in on the permissible value threshold. In Poznań as in Wrocław, significantly higher values of particulate matter were observed in the heating season. The values were twice as high than in the summer season. A similar situation occurs in other Polish urban areas [10,18,32,33].

Analysis of Variability in Concentration of Particulate Matter in Wrocław and Poznań
The analyses of the studies carried out in Wrocław (Station 1) indicate that the largest concentrations of PM 10 particulate matter were recorded during the studies carried out in the heating season in the winter of 2016. The maximum concentration of the PM 10 fraction at that time was equal to 97.81 µg/m 3 (Table 1). Moreover, for the fraction studied in that research period, the authors observed the highest standard deviation (± 22.98) for specific concentration values of the analyzed period. A large standard deviation (± 10.66) was recorded for PM 1.0 particulate matter. The results acquired in that measurement series for the studied fraction (PM 1.0 ) were contained within the range of 2.84-38.20 µg/m 3 [34].
In the case of Poznań, the average concentration of PM 10 in the winter season was higher at the Szymanowskiego Street station. However, during the measurement series, the concentration levels at both stations were at a similar level ( Table 2). For the station at Jana Pawła II Street (Station 8), an average concentration of PM 10 in the summer season was 35% lower than in the first measurement series; moreover, in the case of the station at Szymanowskiego Street (Station 6), the concentration in the second measurement series was lower than the average concentration in the first measurement series by as much as 55%. Perhaps this phenomenon is related to the phenomenon of low-stack emissions. The surroundings of the station at Szymanowskiego Street are composed mainly of single-family buildings, and the method of heating these houses in the heating season affects the air quality in this case. However, it must be noted that the average concentration of PM 10 in the summer season at the station at Szymanowskiego Street was lower than at the station at Jana Pawła II Street. This situation also confirms that elimination of this factor (heating of the buildings) affects the air quality in a given area. The levels of PM 10 pollution at both stations were close to the concentrations of PM 10 observed in the same time periods in the previous years of measurements carried out by VIEP in the cities of Wielkopolska, with the exception of Kalisz (where there are specific conditions for the dispersion of pollutants) [17,32]. Concentrations of PM 2.5 during the first measurements series ranged from 2.2 µg/m 3 to 39.9 µg/m 3 at the station at Jana Pawła II Street (Station 8). At the station at Polanka Street (Station 7), the concentrations ranged from 7.2 µg/m 3 to 88.5 µg/m 3 . The average concentration was respectively equal to 15.2 µg/m 3 for the station at Jana Pawła II Street and 30.8 µg/m 3 for the station at Polanka Street. Concentrations noted during the second measurement series (summer period) were lower and ranged from 1.2 µg/m 3 to 40.3 µg/m 3 at the station at Jana Pawła II Street (Station 8) and from 7.2 µg/m 3 to 17.1 µg/m 3 at the station at Polanka Street (Station 7). The average concentrations of PM 2.5 during the second measurement series at both stations were very similar and equal to 10.4 µg/m 3 at Jana Pawła II Street and 11.0 µg/m 3 at Polanka Street. Thus, the average concentration of PM 2.5 at Jana Pawła II Street, in the summer season, was 32% lower than in the winter season, and at the station at Polanka Street, the concentration was lower by as much as 64%. At the station at Jana Pawła II Street, which is treated as a so-called road traffic station and was established to determine how emission from road traffic influences the air quality, we did not observe as distinctive differences in the concentrations of PM 10 and PM 2.5 between seasons, as we did for the two remaining background stations. However, it should be highlighted that the levels of particulate matter concentration in the winter season at every station were considerably higher than concentrations determined by World Health Organization (WHO) as safe for the health and life of people (PM 10 -20 µg/m 3 , PM 2.5 -10 µg/m 3 ) [35]. In the case of Poznań, the average concentration of PM10 in the winter season was higher at the Szymanowskiego Street station. However, during the measurement series, the concentration levels at both stations were at a similar level (Table 2). For the station at Jana Pawła II Street (Station 8), an average concentration of PM10 in the summer season was 35% lower than in the first measurement (solid and liquid) and biomass combustion [36]. It is confirmed that in the winter season, there is an increase in the emission of gaseous precursors of PM 1.0 ; therefore, it is obvious that the concentration of the finest ambient particles is increasing simultaneously. The concentrations of PM 1.0 in Poznań are comparable to concentrations measured in other Polish cities and also higher than in the countries of western Europe [35,37]. High concentrations of PM 1.0 in cities are related to the presence of many emission sources in those areas. It is however necessary to mention that high concentrations of PM 1.0 in Poland are also confirmed in rural and suburban stations. This is most likely related to the process of transportation of PM 1.0 from other areas, often very distant ones. The finest fraction of particulate matter may be present in the atmosphere for a long time; therefore, it may be transported a long distance [38]. Higher concentrations of particulate matter in the heating season in Poland are confirmed by numerous studies [29,39,40]. Studies carried out in Wrocław and Poznań confirmed that concentrations of all PM fractions are higher in the heating period than in the non-heating period.    The analysis of the composition of PM 2.5 in Poznań in the fall-winter season revealed that among the determined cations, the dominating percentage belonged to (Table 3) [12,44,45]. The content of OC and EC in PM 2.5 particulate matter at both of the stations in the summer season were at a similar level, however it was about four times lower than the content in the winter season. Higher concentrations of OC, EC and selected ions during the heating period are also observed in other Polish and European cities (Table 4).   Table 5 and Figure 3 present, respectively, the results of the PCA analysis factors for the PM 2.5 fraction and linear dependence of the sum of PM 2.5 -bound cations vs. the sum of PM 2.5 -bound anions obtained for the heating and non-heating season and for the two selected cities under study.  In the case of Wrocław, the conducted analyses (Table 4) enabled the identification of four main components in the heating season. In the Principal Component (PC) 1 (Factor 1), for the heating season, there was a correlation between the main component and F − and Cl − anions, and Na + , K + and Mg 2+ cations. Such a relationship was strengthened by the correlation between the main component and Ca 2+ cations in the PC 2 and NH 4 + , OC and EC in the PC 3. This proves the influence of combustion processes and soil resuspension as sources of PM 2.5 particulate matter [58]. In the non-heating season, there was a correlation for Mg 2+ , SO 4 2− and NH 4 + identified in the PC 1 that indicates the influence of emissions originating from fuel combustion processes and also the influence of secondary aerosols [48,58,59]. For stations located in Poznań, the analysis revealed that in the winter season at the station at Polanka Street (Station 7), the most significant for the first PC were OC, EC, F − , Cl − and K + , whereas at the station at Jana Pawła II (Station 8), the most important for the first PC were: F − , Cl − , SO 4 2− , K + , Mg 2+

Identification of the Sources of Origin of Particulate Matter in Wrocław and Poznań
and Ca 2+ , which indicates combustion processes as the source of the considered pollutants. A high percentage of Na + and Cl − in the fourth PC may also indicate the influence of salting (in the winter season) on the composition of particulate matter at the given station [60]. PC 2, singled out in the analysis for the station at Polanka Street, indicates the existence of mineral dust and vegetal fragments in the content of PM 2.5 particulate matter. PC 3, similarly to the case of PC 2, indicates the influence of emissions originating from fuel combustion processes in the analysis for the station at Jana Pawła II Street. In this case, the influence of road transport emissions indicates a high correlation coefficient for OC and EC. PC 4, in the case of the station at Polanka Street, as well as PC 3 from the station at Jana Pawła II Street, are related to the emission of the secondary aerosol NH 4 NO 3 , whose presence proves the that there are high concentrations of its precursors in atmospheric air [33,46]. In the case of Poznań, their presence may be attributed to both road transport and gas heating, which is used by the downtown inhabitants. In the summer season, four main PCs were singled out for both of the analyzed stations. At the station at Jana Pawła II Street, it was found that the most important components for the first PC were SO 4 2− and NH 4 + , and it confirms the influence of fuel combustion on the air quality in Poznań. In the second factor, a high correlation coefficient was characteristic for Ca 2+ , Mg 2+ and PO 4 3− , and in the third factor, a high correlation coefficient was characteristic for F − . This indicates the influence of mineral dust and vegetal fragments on the composition of particulate matter in the summer season at that station [61,62]. At the station at Polanka Street, in the two first factors, the main components that were singled out were characteristic for both liquid and solid fuel combustion. In the third and fourth factors, the main components that were singled out were characteristic of mineral dust presence. The influence on this condition in the summer season may be attributed to the floating of mineral particles originating from the surface of roads and parking lots and also to the influx of air from the Saharan terrains [48]. Nonetheless, it is possible to observe a difference between the origin of particulate matter in the heating and non-heating seasons. In the winter season, fuel combustion remains a dominating source of emissions and at the same time a factor influencing the composition of particulate matter. In the summer season, the presence of mineral dust, which often flows from distant areas [46,60], is very important. In the case of Wrocław, the conducted analyses (Table 4) enabled the identification of four main components in the heating season. In the Principal Component (PC) 1 (Factor 1), for the heating season, there was a correlation between the main component and Fand Cl − anions, and Na + , K + and Mg 2+ cations. Such a relationship was strengthened by the correlation between the main component and Ca 2+ cations in the PC 2 and NH4 + , OC and EC in the PC 3. This proves the influence From analyzing the linear dependence sum of PM 2.5 -bound cations vs. the analyses sum of PM 2.5 -bound anions in all cases (locations and seasons), a statistically significant (p < 0.05) linear relationship of Σcations and Σanions was observed ( Figure 3). Generally, it can be seen that the Σcations/Σanions ratio reached values below one in the winter season and higher than one in the summer season. Specifically, in Poznań, the average values of the Σcations/Σanions ratio in winter and summer were at 0.33 and 2 at Polanka Street, and 2 and 1.1 at John Paul II Street, respectively. In the case of Wrocław, at Kosiby Street, the average values of the Σcations/Σanions ratio were at 1.4, and 0.74 in winter. It can be stated that in winter seasons in the studied atmospheric aerosol, both in Poznań and Wrocław, mainly sulfate and nitrogen anions forming inorganic secondary aerosols predominated. In the summer season, in both considered urban locations, mineral aerosol components (calcium, potassium and magnesium) had a clear contribution. During this period, the aerosol probably have an alkaline characteristic, and the emission of mineral dusts (road dust, soil) can definitely contribute to the PM 2.5 concentration.

Summary
In the analyzed measuring period, the concentrations of particulate matter were high; the highest PM 2.5 concentration in Wrocław was found in the winter of 2016 (70.12 µg/m 3 ) and in Poznań in the autumn-winter of 2016 (88.5 µg/m 3 ). Organic and elemental carbon were macro elements in both locations regardless of the measurement season. Among the ions determined, the highest concentrations were generally recorded for PO 4 3− , SO 4 2− , Na + , NH 4 + and F − .
It has been shown that the primary sources of fine PM emissions in both areas are the burning of fossil fuels, biomass and liquid fuels in automotive engines. In addition, in the measurement campaign conducted in Poznań, a clear impact of gaseous transformations of PM precursors on PM concentrations in the air was observed. In both locations, in selected seasons, a possible contribution of mineral matter (road and soil dust) in PM was also revealed.

Conflicts of Interest:
The authors declare no conflict of interest