Introduction

Airborne fine particulate matter (PM2.5) is a prevalent type of ‘criteria pollutants’1. PM2.5 refers to aerosol particles of 2.5 microns or less in aerodynamic diameter. The substance is a heterogeneous mixture of solid and liquid components, which can contain harmful compounds. It is transported through the atmosphere, often covering significant distances2,3. PM₂.₅ consists of various components, including organic and elemental carbon, ions (sulfate, nitrate, ammonium, chloride), heavy metals (e.g., lead, nickel, chromium, cadmium), as well as organic compounds like polycyclic aromatic hydrocarbons (PAHs), volatile organic compounds (VOCs), and secondary organic aerosols. Some scientific research about air quality in Krakow were performed4,5,6,7. They focus mainly on the total mass of particulate matter and their emission sources. The annual concentration of PM2.5 in Kraków, Poland, decreased by 25% in 2020/2021 compared to 2018/2019. Concentrations of elements, ions, and black carbon were determined and the emission sources of PM2.5 were identified using Positive Matrix Factorization (PMF) by Ryś et al. (2022). PM2.5 levels in both years were influenced by two key factors: the implementation of a ban on coal and wood for residential heating in September 2019 and the reduction in human activities due to COVID-19 lockdown measures in March 2020. Seasonal variation of S, P, Cl, K, Cr, V, Ni, Br, Pb, NO3−, SO42−, NH4+ and black carbon concentrations were observed. The concentrations of S, Cl, K, Zn, Br, Pb, NO3−, SO42−, NH4+ and BC were higher in winter, while the concentrations of Al, Si and Ca were higher in summer. The contributions of emission sources to PM2.5 mass were determined by PMF and four main sources were identified (road dust/construction work/industry/soil; secondary inorganic aerosols; exhaust traffic and solid fuel combustion). To our knowledge no study was performed on chemical states of elements present in particulate matter in Krakow.

X-ray absorption fine structure (XAFS) is an effective method for studying the chemical state of elements8,9,10,11. X-ray absorption near edge structure (XANES) refers to the XAFS technique that focuses on the energy region near the absorption edge. It offers data on the electronic configuration of atoms that are involved in absorption and indicates the oxidation state and chemical speciation of elements. XANES spectroscopy can serve as a distinctive identifier to compare to reference materials in order to acquire speciation information. The procedure is nondestructive, preserving the original speciation in the materials, hence allowing for the acquisition of valid information.

In this study the XANES investigations at S, Cl, K, Fe and Zn K-edges were performed on PM2.5 samples and corresponding reference compounds. We believe that the XANES measurements of PM2.5 will help to differentiate emission sources of PM. Small particles present in PM2.5 can have different composition from coarse particles (PM2.5−10). Coarse particles mostly come from soil and road dust, and industry but PM2.5 particles mostly come from combustions processes and secondary aerosol formation12. Such an experiment using XANES can help describe the emission sources of PM, such as solid fuel combustion, secondary aerosols, and traffic exhaust. It can help preparing emission source profiles of PM2.5 and how the sources changed from the year 2018/219 to 2020/2021 as well as from summers to winters. For the XANES measurements the samples of PM2.5 collected by AGH in two seasons (summer, winter) in two years (2018/2019 and 2020/2021) were used. To our knowledge it will be a first characterization of particulate matter (PM) by XANES for such samples in Poland.

Materials and methods

Sampling

PM2.5 samples were collected at the AGH Krakow University research station in Krakow, Poland. The research station is a typical urban background site with residential and commercial buildings. The sampling place is about 2 km from the City Centre.

Sampling was performed over 24 h (i.e., 8:00 a.m. to 8:00 a.m. the next day). Samples from summer (06_08_2018, 09_08_2018, 05_08_2020, 08_08_2020) and winter (09_01_2019, 15_01_2019, 11_01_2021, 17_01_2021) were taken for the analyses. The PM2.5 were collected on 46.2 mm diameter PTFE Teflon filters (Whatman) using a low-volume sampler at flow rate of 2.3 m3/h. All samples were stored in a refrigerator at – 20oC before and after analyses.

Elemental analysis

Chemical element concentrations (Al, Si, P, S, Cl, K, Ca, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, As, Br, Rb, Sr and Pb) were determined by energy dispersive X-ray fluorescence spectrometer (ED-XRF) at the AGH Krakow University. The spectrometer consists of 3 kW Mo-anode water cooled X-ray tube with line focus (300 μm Be exit window, take off angle 6°, effective source size 0.4 mm x 0.8 mm), silicon drift detector (SDD active area 70 mm2 collimated to 50mm2, FWHM = 124 eV at 5.9 keV, 12.5 μm Be window), 8-positions sample changer integrated with secondary target holder with switchable secondary targets. The measurement geometry is defined by three axes perpendicular to each other: (1) X-ray source-secondary target, (2) secondary target-sample, (3) sample-detector. The acquisition process was controlled with the use of in-house developed LabView program. The X-ray tube was operated at 55 kV/30 mA exciting Ni and Mo secondary targets. Chemical elements such as: Al, Si, P, S, Cl were quantified with Ni secondary target; the measurement time for a single sample was 1 000 s. K, Ca, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, As, Br, Rb, Sr, Pb were quantified with Mo secondary target and the measurement time for one sample was 2 400 s. The specimens are placed on the plate with 8 positions with the automatic changer of samples. The measurements for both targets were performed under atmosphere air pressure.

XANES analysis

X-ray absorption spectroscopy measurements were conducted at the ASTRA beamline of the SOLARIS Synchrotron in Krakow, Poland13,14. We measured XANES spectra at the S, Cl, K, Fe, and Zn K-absorption edges using a photon beam from a double-bend achromatic 1.3 Tesla bending magnet (Ec≈2 keV). For the K, Fe, and Zn edges, a Ge(220) crystal pair was used to monochromatize the beam, while an Si(111) crystal pair was used for the S and Cl edges. The beam size at the sample position was set to 7 × 1 mm using slits. Spectra were obtained in fluorescence mode at room temperature. For Fe and Zn K-edge measurements, ionization chambers and sample chamber were filled with N2 gas at atmospheric pressure. For S, Cl, and K K-edge measurements, the pressures were 25 torr, 37 torr, and 82 torr, respectively. Fe and Zn foil from Exafs Company (Danville, USA) were used for beamline calibration at the Fe and Zn K-edges. Calibration at the S K-edge was done using ZnSO4∙7H2O (whiteline at 2481.4 eV), and for the Cl and K K-edges using KCl (whiteline positions at 2825.0 eV for Cl K-edge and 3611.0 eV for K K-edge). Filters with PM2.5 were cut to fit into the sample holders. Reference compounds expected to be present in particulate matter, including (NH4)2SO4, NH4HSO4, Na2SO4, NaHSO4, K2SO4, KHSO4, K2CO3, CaSO4, FeSO4, CuSO4, ZnSO4, ZnS, KCl, KHCO3, KNO3, K2HPO4, KH2PO4, Fe2O3, and FeOOH, were dilluted in cellulose to have concentration of the element of interest < 1% and pellets with diameter 13 mm were made. The number of reference compounds used for different elements in the Linear Combination Analysis (LCA) varies based on the chemical behavior of each element. For some elements, more reference compounds are required because they exhibit a wider range of oxidation states, coordination environments, or other chemical forms. This diversity necessitates a broader selection of reference compounds to capture the full range of possible spectral characteristics. Data processing and analysis were performed using the Athena program (version 0.9.26) from the Demeter software package for X-ray Absorption Spectroscopy (XAS) data analysis. The software is freely available and documented at https://bruceravel.github.io/demeter/15.

Results and discussion

Elemental analysis

Table S1 describes the collected PM2.5 samples. Table S2 and Fig. 1 present concentrations of PM2.5 and elements in PM2.5 samples used later for XANES analysis. The highest concentrations were observed for sulphur, which were in the range 554–3526 ng/m3. The S concentrations did not depend on the season of the year. Sulphur appeared in aerosols mainly as secondary inorganic aerosols. Chlorine concentrations were high during wintertime, and they were in the range 520–3724 ng/m3. The sources of chlorine during winter can be connected to solid fuel combustion and de-icing of the pavements and roads. In summertime the concentrations of chlorine were slightly higher than the detection limit. Potassium concentrations varied daily, but no significant correlation with the day of the week was observed. We did not observe any seasonal variations of K concentrations. The source of potassium in winter can be related to heating season while in summer K can come from mineral and soil dust. The determined concentrations of iron were low, and they occurred in the range 16–202 ng/m3. Fe appeared mainly in aerosols from industry, soil dust, wearing tires and roads and from the engines. Zinc concentrations changed from 5.5 to 88 ng/m3 depending on the day of collecting the samples. Industry, wearing of tires and brakes, working engines and soil dust are the main sources of zinc in the atmosphere.

Fig. 1
figure 1

Elemental concentrations in PM2.5 samples collected in summer and winter. The arrows refer to the elements selected for the analysis.

Full size image

XANES analysis

Sulphur, chlorine, potassium, iron and zinc were chosen for analysis because firstly they are markers of emission sources of PM2.5 as industry, combustion, traffic or are responsible for oxidative reactions.

XANES analysis can be performed on samples with element concentrations as low as several tens of ppm, depending on the specific element and experimental conditions. Our XRF studies provides quantitative elemental analysis and serves as a preliminary tool to estimate whether the element of interest is present and at which concentrations. In our work we decided to focus on the chemical speciation of elements that have relatively high concentrations (whose spectra could be measured at the beamline in a reasonable time) and have been investigated by other groups worldwide using different research methods. This approach allows us to determine if our results complement existing research.

Sulphur

Figure 2, Figures S1, S2 show sulphur K-edge XANES spectra, and their derivatives collected for PM2.5 samples and 11 reference compounds. Additionally, XANES spectra and the first derivatives of XANES spectra of sulphur at K-edge of reference samples only were presented on Figure S3. For summer samples XANES peak at S K-edge appeared at the energy 2481.2 eV and corresponded to sulphates (S+ 6). The comparison to reference samples shows that the sulphur compounds can be (NH4)2SO4, Na2SO4, K2SO4, CaSO4, CuSO4, FeSO4 and ZnSO4. It corresponds to our results from X-ray fluorescence analysis. Sulphur precursor SO2 can come from natural or anthropogenic sources and in the atmosphere, it reacts with oxygen, producing SO3. At the presence of water and chemical elements, SO3 produces H2SO4. The reactions need to be at the presence solar radiation (higher ambient temperature) and metals like iron, mangan. Ultraviolet radiation causes appearing OH radicals, which catalyse the process. Ammonium ions present in the atmosphere react with sulphuric acid producing ammonium sulphate. The sources of ammonium ions are industry, agriculture and combustion of liquid fuels. Secondary sulphates are in many cases attributed to long-range transport events and are frequently associated with “aged air masses” due to the slow oxidation of SO2 to SO4 2−4,16. In summer SO2, H2SO4, and metals can come from combustion of liquid fuels like for example diesel, industry as well as power plants. The emitted diesel particles are in size range from 20 nm to 200 nm17,18. The reactions of H2SO4 with metals produce sulphates of these metals. The road dust is the source of Cu, Zn, and Fe, which could be present in the form of sulphates. Copper comes from brake wear, zinc and iron come from tires and asphalt wearing6 and also from the engines due to the Zn content of lubricants. Based on the spectra (Fig. 2 and Figures S1a–c, S2a–c), the peaks were observed at energies of 2481.4 and 2473 eV in the winter samples. The intense asymmetric broad peak corresponds to bisulphates and sulphates (S+ 6), the weaker second one at the lower energy (enlarged part at the Fig. 2 and Figures S1a–c, S2a–c) corresponds to sulphides (S− 2) and aromatic organic sulphur, most probably thiophenic or similar aromatic sulphur derivatives. The presence of sulphates in both seasons was also confirmed by Cozzi et al.21. In our study the presence of sulphides (S− 2) in the form of ZnS phase in winter samples confirmed by Linear Combination Analysis (LCA) of XANES spectra at Zn K-edge (Fig. 4d and Figure S9b,d,f). The presence of these phases of sulphur in PM2.5 can be connected to coal combustion during winter. In coal sulphur is present in organic and inorganic form. There was an increase of the secondary sulphate contribution in the winter probably due to the increase of its precursors, which are associated with the combustion sources used for residential heating whose impact in Krakow has already been described in previous studies19. Moreover, sulphate is to a large extent originating also from cloud phase SO2 oxidation, which occurs during winter as well4. Low temperature and high humidity in winter influence the heterogenic oxidation reactions in liquid phase neutralizing NH3. Similar results were obtained by Huggins et al. (2000)10,11 for standards reference material Urban PM SRM 1648. Sulphides and organic sulphur peak appeared in XANES spectra of different coal samples by Huffman et al. (1991)20, where 0–35% of sulphur in coal exists as sulphides and 30–75% of sulphur as thiophene. Only few percentages of sulphur in coal appears in the form of sulphate, it is in the range 0–12%. Also, the comparison to spectra of reference compounds suggests that for winter samples the bisulphate forms in PM2.5 are mostly NH4HSO4, KHSO4 and NaHSO4, which are the main products of solid fuel combustion in more acidic atmosphere20,21. The sulphates are presented by (NH4)2SO4, Na2SO4, K2SO4, CaSO4, CuSO4, FeSO4 and ZnSO4. The presence of potassium, iron and zinc sulphates for summer and winter samples were confirmed by LCA analysis of the XANES spectra at K, Fe and Zn K-absorption edges (Figs. 3 and 4 and Figures S6, S7, S9).

Fig. 2
figure 2figure 2

The XANES spectra (a) and the first derivatives of XANES spectra (b) at sulphur K-edge of PM2.5 collected in summer 06 August 2018 and in winter 9 January 2019 and the reference compounds. The inset represents the enlarge part of the energy range characteristic for the sulphides and sulphur organic compounds.

Full size image
Fig. 3
figure 3

Linear combination analysis results of potassium K-edge XANES spectra of winter PM2.5 samples.

Full size image
Fig. 4
figure 4

Linear combination analysis results of iron K-edge (a, b) and zinc K-edge (c, d) XANES spectra of summer and winter samples.

Full size image

Chlorine

At the K- absorption edge of Cl only XANES spectra of four PM2.5 samples collected in winter were analysed because samples from summer had very low concentrations of chlorine. For three samples collected on 9 January 2019, 10 January 2019 and 11 January 2021 the same features of the XANES spectra (Figure S4 a-c) were observed what allows one to conclude on similarity of Cl containing phases in the air. The main features of the Cl K-edge spectra for the samples appeared at the energies 2825 eV and 2826.7 eV (Figure S4.) and can be related to many pollutants containing Cl present in the air during wintertime. Fine Cl-containing particles originate from human activities such as industry, fuel combustion and waste burning, whereas the main source of coarse Cl-containing particles is the sea spray and earth crust22. Chlorine is also released during volcanic eruptions. Most common form of chlorine is HCl which is released by volcanoes. Elemental chlorine is also emitted, but most Cl2 reacts with H2 or VOCs to form HCl or chlorinated organic compounds. Compounds created during volcanic eruptions are emitted into the atmosphere and can be transported over distances of thousands of kilometers and can remain in the atmosphere for a long period of time23. Based on the comparison (see Figure S4 a-c) of our experimental XANES spectra with XANES spectra of Cl containing phases available in the literature24,25,26,27, we could conclude that the main Cl containing phases in the three samples mentioned above were inorganic components of CaCl2 × 2H2O and NaCl. On the other hand, the XANES spectrum of the sample collected on 17 January 2021 (Figure S4 d) contains well pronounced feature at 2823.8 eV and 2826.7 eV, which allows us to conclude that apart from the content of the aforementioned inorganic phases, the sample contain organic chlorine compounds such as C6O2Cl4. The sources of organic chlorine can be coal, rubber burning and industry9,24. Calcium and sodium chlorides are used for de-icing of roads and pavements during wintertime.

Potassium

Two characteristic XANES peaks at the K K-edge for samples collected in winter appeared at the energies in the ranges 3612.5 eV − 3613.35 eV and 3618.2 eV – 3618.8 eV (Fig. 4). We also measured reference compounds, which could be expected in the urban air11 (Figure S5). Linear combination fitting analysis delivered the information about contribution of the chemical compounds containing potassium in PM2.5 collected at winter season (Table 1).

Table 1 The contribution of compounds to K in PM2.5 collected in winter.
Full size table

The highest contribution (35–46%) of potassium compounds had KHSO4, which is in correlation with data obtained for sulphur. It is produced in acidic atmosphere. The source of potassium can be biomass burning. At higher temperatures during the char combustion, potassium can be released as K2SO428. The contribution from 3 to 11% was observed for K2HPO4. Airborne Potassium bicarbonate has widespread use in crops, especially for neutralizing acidic soil. The contribution KHCO3 to K in PM2.5 is in the range from 17 to 25%. We were not able to reconstruct the K K-edge XANES spectra for summer samples, which have an intense peak at around 3612.7 eV using the references which we apply for winter samples analysis (Figure S6). From the literature review we found that such minerals as nepheline, olivine, AlK(SO4)2 × 12H2O, K2SiO3, leucite29,30 have intense peak at the positions close to the intense peak for the samples of interest from summer. These minerals could be a part of a soil dust, which is a standard component in the summer air among K2SO4, K2CO3 and KNO3.

Iron

XANES spectra of all samples of PM2.5 and reference compounds FeSO4, Fe2O3, FeOOH were measured at Fe K-edge. LCA analysis of the obtained spectra showed that that phases of Fe+ 3 (Fe2O3 and FeOOH) and Fe+ 2 oxidation states (FeSO4) are present in the samples of PM2.5 (Fig. 3a-b, Figure S7a-f, S8). The contribution of FeSO4 to iron in PM2.5 in summer was 7% for 3 samples and 26% for one sample (Table 2). During winter these contributions were 4% for three samples and 8% for one sample. FeSO4 appeared in the atmosphere probably by the reaction of Fe with sulphuric acid and sulfure oxides (SO2 and SO3). The highest contributions were observed for FeOOH in summer, and they were in the range 54–70% of iron in PM2.5. During winter FeOOH contributions to Fe were 32–54%. α-FeOOH is a mineral goethite and the origin of this compound in natural as soil dust. α-Fe2O3 has natural origin, as a mineral hematite. It belongs also to a road and soil dust source. Similar contributions of FeOOH and Fe2O3 to iron in PM2.5 were obtained in former studies10,31,32,33. Table 2 presents the contributions of chemical compounds to iron in PM2.5 during winter and summer. Iron has also anthropogenic origin, it can come from brake and tire wear and industrial activities. Iron particles during oxidation can form iron oxides. O’Day et al.32, described that road grinding, ambient weathering, and atmospheric processing of Fe(0) and Fe+ 2-bearing particles will produce oxidized Fe+ 3 phases such as ferrihydrite (amorphous Fe(OH)3), lepidocrocite (γ-FeOOH), goethite (α-FeOOH), and sulfate-rich goethite (schwertmannite). Particle surfaces modified by emission generation processes (i.e., combustion or abrasion), and oxidized and modified by environmental weathering would have surface chemistries reflecting alteration and mixed-metal oxidation states, and they would mix with a large and complex suite of organic compounds from combustion emissions and volatilization.

Table 2 The contributions of FeSO4, Fe2O3 and FeOOH to iron in PM2.5.
Full size table

Zinc

XANES spectra of all samples of interest and three reference compounds (ZnSO4, ZnS, ZnO) were measured at zinc K-edge (Fig. 3c-d, Figure S9 a-f, S10). The whiteline for summer samples and for sample from 17 January 2021 appeared at the energy 9667.15 eV. For the rest of the winter samples the whiteline appeared at the energy 9664.6 eV. LCA analysis of the measured Zn K-edge spectra showed that main phases present in the winter samples are ZnSO4 (24–46%) and ZnS (54–76%) (Table 3), where for summer samples Zn containing phases are mostly ZnSO4 (88–96%) with small contribution from ZnO (4–12%). These data are in correlation with the data obtained for the at S K-edge. Wang et al.34 obtained similar results regarding the presence of ZnSO4 in PM2.5 and PM10. Knowledge on the chemical speciation of Zn in particulate matter is important on one hand for assessment of toxicity after inhalation34, on the other hand airborne particulate matter emitted by specific sources can contain Zn in characteristic chemical forms. For example, ZnO can be related to fly-ash particles formed during combustion processes of solid fuels (coal or biomass)35, since ZnO is not present in soils or rocks, it can only be formed during high-temperature processes. The sources of zinc are road dust (tire wear) and soil dust (natural origin). For winter sample 17,012,021 the contribution of ZnS (11%) and ZnSO4 (89%) was also in correlation with the data obtained at S K-edge, where intense signal from sulphates were detected. ZnSO4 appeared in the atmosphere as a result of reaction of Zn with H2SO4 and SO3, which are produce in industrial processes. ZnS comes from combustion of coal, industrial activities as well as it has natural origin as a mineral sphalerite. ZnS can be released from residual oil combustion36.

Table 3 The contributions of ZnSO4, ZnO and ZnS to zinc in PM2.5.
Full size table

Conclusions

Samples of PM2.5 were measured using XANES technique. Different spectra were obtained for summer and winter at all measured edges. At the edge of sulphur for samples collected in summer mainly sulphates were identified. On the other hand, for winter samples additionally bisulphates, sulphides and organic sulphur compounds were present. These results were confirmed by measurements of PM2.5 at K, Fe and Zn edges. In winter potassium exists mainly as KHSO4, K2SO4, KHCO3 and KNO3. Spectra for summer were more complicated and additionally the K compounds of aluminosilicates and other not stochiometric compounds of potassium from natural origin were assumed also to be present in the samples. Iron exists on + 2 and + 3 oxidation states in particulate matter. The main compounds were FeSO4, Fe2O3 and FeOOH, which were present in winter and summer in different contributions. From measurements at Zn K-edge different compounds were observed for winter and summer. In summer ZnSO4 and ZnO were identified and in winter ZnSO4 and ZnS. Inorganic and organic compounds of chlorine were observed in XANES spectra. The XANES study is highly useful for determining the chemical phases of elements in PM2.5, allowing for a more precise identification of emission sources such as combustion processes, industrial activities, and road dust. Moreover, understanding the oxidation states and chemical forms of these elements provides insight into their potential toxicity and health effects, as different species exhibit varying reactivity in the human respiratory system. It is crucial for activities which aims the decreasing the pollution at any site in the World. The study was performed for one city, but the results and conclusions can be applied to many sites.