Geochemical and Mineralogical Characteristics of Airborne Particulate Matter in Relation to Human Health Risk

: The main objective of this research was the determination of the geochemical and mineralogical properties of particulate matter: TSP (total suspended particles) and, especially PM1 (particles with aerodynamic diameter not greater than 1 µ m) suspended in the air of a selected urban area in southern Poland. Identiﬁcation of the emission sources of metals and metalloids bound in TSP and PM1 as well as the assessment of potential risk of urban ambient air to human health using hazard indices was an additional aim of this investigation. The daily TSP and PM1 quartz ﬁber ﬁlters collected during heating season were subjected to mass magnetic susceptibility ( χ ) measurements, SEM (Scanning Electron Microscopy) observations and geochemical analyses. Obtained results revealed that the concentration of TSP and PM1 well correlated with their mass-speciﬁc magnetic susceptibility. The good relationship between the PM concentration and χ suggests that magnetic susceptibility measurements can be a good proxy of low-level atmospheric dust pollution. The rank order of potentially toxic elements (PTE) based on average concentration was Ba > Zn > Al > Fe > Pb > Mn > Ti > Cu > Cr > Ni > As > Cd > V > Tl, both for TSP and PM1. PM1 / TSP ratios for PTE concentrations and χ were around or slightly above unity, which indicated that PM1 was the main carrier of PTE (with the exception of cadmium, copper and lead) and technogenic magnetic particles. The non-carcinogenic and carcinogenic risks were conﬁrmed by very high values of human health indices.


Introduction
Human activities contribute to the air pollution, especially in industrial and urban areas. Airborne particulate matter (PM), consisting of a mixture of solid and liquid particles suspended in the air, varies in terms of physical properties (shape, dimension, density) and chemical composition [1][2][3]. As a result of atmospheric deposition, metals migrate to water and soil, where they can be taken up by plants and animals and further transferred through the food chain directly into humans. Once released into the environment, metals are continuously undergoing physical and chemical transformations that affect the forms of their occurrence and determine their toxicity. Currently observed changes in the The collected 70 filters of daily retained PM1 (35 filters) and TSP (35 filters) were divided by means of a guillotine specially designed for this purpose and adapted to the diameter of the filters and then weighted. One part of them was subjected to the non-destructive volumetric magnetic susceptibility (ĸ) measurements using an MFK1 Kappabridge device (Agico Advanced Geoscience Instruments Co., Brno, Czech Republic) that operated at a low frequency (976 Hz) and low magnetic field intensity (200 A·m −1 ). In order to ensure the exact positioning of the specimen, each filter was placed in the 10 mL cylindrical bottle (originally assigned for the Bartington MS2B device, Bartington Instruments Ltd., Witney, UK) which afterwards was placed in the central bottom of the plastic vessel of MFK1 for specimen fragments for use in static measurement mode (KLY VES40). Each sample was measured five times at least, and the mean value was calculated. Subsequently, the mass-specific magnetic susceptibility (χ, m 3 /kg) was computed by taking into account the weight of the sample as well as the mass and the χ (−0.02 × 10 −8 m 3 /kg) of the clean filters.
Afterwards, SEM/EDS analysis was carried out at the Institute of Earth Science, Faculty of Natural Sciences of University of Silesia in Katowice. Industrial dusts and fly ash samples from neighboring industrial emission sources collected directly at the source (power and metallurgical plants) were analyzed using a scanning electron microscope Philips XL30 ESEM/TMP equipped with an EDAX EDS Sapphire system, detectors: SE, BSE and EDS. This microscope can work in the environmental mode, i.e., it enables analyses to be carried out without the need to spray the sample with a conductive layer, e.g., carbon (metallurgical dusts). Electrifying fly ashes were dusted with a The collected 70 filters of daily retained PM1 (35 filters) and TSP (35 filters) were divided by means of a guillotine specially designed for this purpose and adapted to the diameter of the filters and then weighted. One part of them was subjected to the non-destructive volumetric magnetic susceptibility (k) measurements using an MFK1 Kappabridge device (Agico Advanced Geoscience Instruments Co., Brno, Czech Republic) that operated at a low frequency (976 Hz) and low magnetic field intensity (200 A·m −1 ). In order to ensure the exact positioning of the specimen, each filter was placed in the 10 mL cylindrical bottle (originally assigned for the Bartington MS2B device, Bartington Instruments Ltd., Witney, UK) which afterwards was placed in the central bottom of the plastic vessel of MFK1 for specimen fragments for use in static measurement mode (KLY VES40). Each sample was measured five times at least, and the mean value was calculated. Subsequently, the mass-specific magnetic susceptibility (χ, m 3 /kg) was computed by taking into account the weight of the sample as well as the mass and the χ (−0.02 × 10 −8 m 3 /kg) of the clean filters.
Afterwards, SEM/EDS analysis was carried out at the Institute of Earth Science, Faculty of Natural Sciences of University of Silesia in Katowice. Industrial dusts and fly ash samples from neighboring industrial emission sources collected directly at the source (power and metallurgical plants) were analyzed using a scanning electron microscope Philips XL30 ESEM/TMP equipped with an EDAX EDS Minerals 2020, 10, 866 4 of 19 Sapphire system, detectors: SE, BSE and EDS. This microscope can work in the environmental mode, i.e., it enables analyses to be carried out without the need to spray the sample with a conductive layer, e.g., carbon (metallurgical dusts). Electrifying fly ashes were dusted with a thin layer of carbon and analyzed in the high vacuum mode. Magnifications from several to several thousand times were used, and the voltage accelerating the beam was 15 keV.
TSP/PM1 filters were analyzed using a Quanta 250 scanning microscope with Thermo-Fisher EDS Ultra Dry analyzer (Thermo Fisher Scientific Inc., Waltham, MA, USA). BSE images were recorded, in which, apart from the habit and grain size, the chemical composition variation was visible both between and within the sample grains. EDS microanalysis was performed at selected sample points to determine the chemical composition of the grains present in the samples.
The second parts of the filters were digested in a mixture of nitric acid and hydrogen peroxide solution in microwave oven in order to determine total content of the following elements, in majority considered as potentially toxic (PTE): V, Mn, Ni, Cu, Zn, As, Cd, Tl, Pb, Cr, Ti and Al using High-Resolution Inductive Coupled Plasma-Mass Spectrometry (HR-ICP-MS, 6100 DRC-e Perkin Elmer, Waltham, MA, USA). Iron content was determined by Inductively Coupled Plasma Optical Emission Spectroscopy (ICP-OES; Avio 200 Perkin Elmer, Waltham, MA, USA).
The homogeneity of the loaded halves of filters was previously examined by the magnetic susceptibility measurements of experimental three filters (six halves). Obtained results varied from 51.25 to 61.32 × 10 −8 m 3 /kg (mean 57.54 × 10 −8 m 3 /kg). The favorable low relative error (7.9%) indicates the homogeneity of the collected material and the possibility of using filter halves for further analyses. Sysalová et al. [49] tested homogeneity of divided filters in determination of trace elements in airborne particulate matter and they yielded sufficiently homogenous results (errors for particular elements between 1.7% and 12.5%).
The potential health impacts associated with environmental exposures to inhaled contaminants were estimated on the basis of USEPA's guidance [50]. The non-carcinogenic risk assessment (mutagenicity, neurotoxicity, developmental and reproductive toxicity) posed by As, Cd, Cr, Ni, Pb and Mn and carcinogenic one posed by As, Cd, Cr and Ni (classified as class 1 carcinogenic agents) [51] were determined by their hazard quotient (HQ) and carcinogenic risk (CR), respectively [52,53]. The hazard quotient (HQ) was computed as the ratio of the exposure concentration of the inhalable chemical (EC) to a reference concentration (RfC) meaning an estimate of a continuous inhalation exposure to the human population (including sensitive subgroups) that is likely to be without an appreciable risk of deleterious effects during a lifetime USEPA [54]. RfC values for As, Cd, Cr, Ni, Pb and Mn were as follows: 0.000015, 0.00001, 0.0001, 0.00009, 0.008 and 0.003 mg/m 3 , respectively: The carcinogenic risks (CR) were computed by multiplying the exposure concentration of the inhalable chemical (EC) for the inhalation unit risk (IUR): Values of IUR were specified by USEPA [55]: 0.0043, 0.0018, 0.012, 0.00024 for As, Cd, Cr and Ni ((µg/m 3 ) −1 ), respectively. EC was calculated according to the following equations [50]: All symbols concerning hazard parameters used in the above equations are explained in Table 1. To assess the overall potential non-cancer and cancer health risk caused by multiple compounds the hazard index (HI) was introduced as the sum of hazard quotients of individual elements [50]. In the present work HI values were calculated as follows: HI non-cancer = ΣHQ = EC As /RfC As + EC Cd /RfC Cd + EC Cr /RfC Cr + EC Ni /RfC Ni + EC Pb /RfC Pb + EC Mn /RfC Mn (4) HI cancer = ΣCR = EC As × IUR As + EC Cd × IUR Cd + EC Cr × IUR Cr + EC Ni × IUR Ni (5) Hazard index values exceeding unity provides evidence that potential health effects may occur. Otherwise (HI < 1), it is assumed that the risk is at an acceptable level. Carcinogenic risk is the chance of an individual causing any kind of cancer from lifetime exposure to carcinogenic hazardous substances. The acceptable or tolerable risk for regulatory purposes is 1 × 10 −6 -1 × 10 −4 [51].
The obtained results were analyzed using Microsoft Office Excel software and the package of Statistica for Windows, version 12 (StatSoft Polska Sp. z o.o., Kraków, Poland). The basic statistical functions of mean, median, standard deviation, as well as factor analysis were applied in order to interpret results and explain variations in the data. For data evaluation and presentation mean values of obtained results were used. As the data did not exhibit a normal distribution, a Spearman's rank correlation coefficient was applied.

PM1/TSP Data Relationship
The TSP concentration averaged at 45.58 µg/m 3 with a wide range of 4.67-150.60 µg/m 3 , while PM1 mean concentration was 29.88 µg/m 3 with a range of 4.03-118.87 µg/m 3 ( Figure 2). The highest concentration of both PM fraction occurred on 7 November 2018. This remarkable increase in concentration of both PM fractions can be explained by the air temperature in the range of 7-18 • C and low wind speed (8-10 km/h) leading to temperature inversion creating a low atmospheric convection. This phenomenon causes the retention of pollutants within a height of 2-4 m from the ground [56]. A similar temporal distribution of PM1 and TSM (supported by a correlation coefficient of 0.83) was generally observed, although on some days ( Moreover, the concentration of TSP and PM1 corresponded well with mass-specific magnetic susceptibility of the same fractions, suggesting higher concentration of technogenic magnetic particles (Fe-rich minerals) on days with increasing PM concentration ( Figure 2). A similar relationship was observed for TSP collected in the city of Querétaro, where the magnetic susceptibility (χ) measurements of the filters showed a good linear correlation with the concentration of TSP (r = 0.928) [39]. In this study, the correlation coefficient was not that high and equaled 0.46 and 0.63 for TSP and PM1 concentrations, respectively. Nevertheless, on some days with high humidity  Table 2). This phenomenon was recently described by Petrovský et al. [57], who studied the magnetic properties of particulate matter during smoggy period, obtaining negative correlation between saturation magnetization (concentration of ferrimagnetic iron oxides) and both PM1 and PM10 concentrations. Table 2. Meteorological data (as daily average) for study site in Zabrze.

Date (Month Day Year) Temperature (°C) Wind Speed (m/s) Humidity (%) Pressure (hPa) Rain (mm)
10  Moreover, the concentration of TSP and PM1 corresponded well with mass-specific magnetic susceptibility of the same fractions, suggesting higher concentration of technogenic magnetic particles (Fe-rich minerals) on days with increasing PM concentration ( Figure 2). A similar relationship was observed for TSP collected in the city of Querétaro, where the magnetic susceptibility (χ) measurements of the filters showed a good linear correlation with the concentration of TSP (r = 0.928) [39]. In this study, the correlation coefficient was not that high and equaled 0.  Table 2). This phenomenon was recently described by Petrovský et al. [57], who studied the magnetic properties of particulate matter during smoggy period, obtaining negative correlation between saturation magnetization (concentration of ferrimagnetic iron oxides) and both PM1 and PM10 concentrations.
The mass-specific magnetic susceptibility (χ) of collected daily filters varied from 2.0 to 168.1 × 10 −8 m 3 /kg for TSP and from 0.7 to 139.2 × 10 −8 m 3 /kg for PM1 (Figures 2-4; Table 3). The mean and median values for the data were 22.4 and 12.4 × 10 −8 m 3 /kg for TSP filters, and 18.3 and 11.1 × 10 −8 m 3 /kg for PM1 samples (Table 3). These values were relatively low, in the range of rural background area obtained in Latium (Italy) by Sagnotti et al. [40], who for a high-traffic industrial area received an extremely high value of the χ above 1000 × 10 −8 m 3 /kg. In presented case, the results of both fractions were quite similar, suggesting a comparable content of magnetic iron minerals in both TSP and respirable PM1. Simplifying, it seems that PM1 was the main carrier of most technogenic magnetic particles in the total aspirated dust. A high PM1/TSP ratio ( Table 4) may indicate that the TSP is mostly influenced by the small size distribution sources such as combustion processes and secondary particles sources, while lower PM1/TSP ratio indicates that the TSP was mostly influenced by natural sources and mechanical processes generating larger particles, e.g., non-exhaust particles [58]. Similar results concerning the ratio between fine (PM 2.5 ) and coarser (PM 2.5-10 ) particular matter were reported for the industrial area of Beijng, China; however, the χ values were much higher [42].        Generally, enhanced values of χ indicated that PM on filters may contain relevant amounts of Fe-rich particles originated probably from anthropogenic sources, but natural dusts and re-suspended soil may have also significant contribution to the magnetic properties. Previous studies have shown that the magnetic properties of airborne particles may be strongly associated with potentially toxic elements (PTE) [59][60][61][62].

Concentration of Potentially Toxic Elements and Their Source Apportionment
The total PTE concentrations (range, mean, median and standard deviation) determined in this study are summarized in Table 3 along with corresponding χ values. The rank order of PTE based on average concentration was Ba > Zn > Al > Fe > Pb > Mn > Ti > Cu > Cr > Ni >As > Cd > V > Tl, both for TSP and PM1 (with only exception: Mn occurred in higher concentration than Pb in PM1) ( Table 3). Comparing the content of elements in TSP and PM1, it was observed that content of V, Mn, Ni, As, Tl, Cr, Ti, Ba and Al was slightly higher in PM1, but the daily variability of elemental concentrations in TSP and PM1 looks similar (Figures 3 and 4). On these days, stagnant weather conditions and residual mists were noted, i.e., conditions that encourage concentration of pollutants by capturing particles and reducing the degree of their dispersion in the atmosphere [63].
Calculated averaged ratios between all PTE (mass concentration and χ) determined in PM1 and TSP were around or above unity (Table 4), which unequivocally indicated that, although PM1 exhibited lower mass concentration than TSP, it was a main carrier of PTE (with the exception of cadmium, copper and lead) and technogenic magnetic particles (expressed by higher values of magnetic susceptibility) as well. The computed empirical correlation coefficients between magnetic susceptibility of diurnal filters and the particular PTE were relatively high and statistically significant (0.34-0.99) for the majority of elements, suggesting that the largest part of magnetic minerals in particulate matter originates from anthropogenic combustion processes. From Spearman's correlations presented in Tables 5 and 6 we observed that in the case of TSP, magnetic susceptibility showed a significant correlation with As, Cd, Cu, Fe, Mn, Pb and Tl. In the case of PM1, magnetic susceptibility exhibited a significant positive correlation only with Cu and Mn. In addition, PM1 showed a negative significant correlation between magnetic susceptibility and Ba. Such a specific "negative" binding to very small particles can also be observed for Al and Zn. In case of TSP, no association with χ and Al, Ba and Zn was found. Therefore, in case of such elements as Al, Ba and Zn, magnetic susceptibility cannot be used as reliable proxy reflecting PTE concentration in PM filters.   The bold font-correlation significant at the 0.05 level.  The bold font-correlation significant at the 0.05 level.
Correlated PTE could be related to a common source of emission; however, principal component analysis, a valuable multivariate statistical tool was applied as a complementary attempt of the anthropogenic origin assessment. Two main groups of components were separated in both TSP and PM1 ( Figure 5). A distinct group consisting of Al, Ba and Zn was distinguished for both size fractions of PM. As the filters were collected in an urban area, high concentration of these elements, especially Ba and Zn, was probably due to the influence of tyre abrasion [64]. Moreover, Ba is a common component in automobile brake pads [65], as well as in diesel fuel as a smoke suppressant [66], but, in the case of Silesia, Ba originated from fly ashes of the coal used for heating [67,68]. These elements could be also related to fossil fuel combustion, but the influence of vehicular and industrial (non-ferrous metallurgy in short-distant Miasteczko Slaskie) emissions cannot be excluded [69,70]. Simultaneously, Al along with Ti (and also Fe and Mn) are considered as crustal metals and their enhanced concentration can be an effect of road dust re-suspension and the local geology/soil composition [71,72]. The second TSP cluster was characterized by high loadings of As, Cd and Pb which may be related to coal combustion [73,74]. The rest of the elements in TSP have not been assigned to any of the clusters, indicating their diverse origin. While the second PM1 cluster was associated with χ, As, Cd, Fe, Mn, Pb and Tl, elements which are produced as a by-product of nonferrous metal smelting and refining. Mn, Cd and Ni may be associated with exhaust emissions [74][75][76]. Furthermore, such elements as Pb, Zn, Cu, Cd, Cr and Ni in TSP and PM1 are the markers of diesel engine exhaust, oil combustion, tire and brake abrasion, brake wear debris and burning of lubricating oil [77][78][79]. Simultaneously, Al along with Ti (and also Fe and Mn) are considered as crustal metals and their enhanced concentration can be an effect of road dust re-suspension and the local geology/soil composition [71,72]. The second TSP cluster was characterized by high loadings of As, Cd and Pb which may be related to coal combustion [73,74]. The rest of the elements in TSP have not been assigned to any of the clusters, indicating their diverse origin. While the second PM1 cluster was associated with χ, As, Cd, Fe, Mn, Pb and Tl, elements which are produced as a by-product of non-ferrous metal smelting and refining. Mn, Cd and Ni may be associated with exhaust emissions [74][75][76]. Furthermore, such elements as Pb, Zn, Cu, Cd, Cr and Ni in TSP and PM1 are the markers of diesel engine exhaust, oil combustion, tire and brake abrasion, brake wear debris and burning of lubricating oil [77][78][79].

Mineral and Morphological Characteristics of Dusts
The results of SEM observations confirmed the above-mentioned relationships between particular elements, showing that in both PM fractions natural fly ashes are mostly composed of silicate and carbonate minerals occur together with characteristic technogenic particles. Additionally, remarkably alike technogenic particles visible on SEM images of fly ashes from power plants as well as metallurgical dusts confirmed that they were sources of particles accumulated on filters (Figures 6  and 7). Amorphous glass (mullit and quartz), quick coke, calcium sulfate and aluminosilicates with inclusions of spherical iron and manganese oxides were the main components of fly ashes ( Figure 6A), while oxides of Fe, Zn, Ca and Mn with varied morphology (spherules, stamens, aggregates) were characteristic for metallurgical dusts ( Figure 6B,C).
Dusts after smelting of Pb and Zn ores consisted of small rods and irregular beads of Pb and Zn oxides, chlorides and sulfates, often enriched with Cd ( Figure 6D). The mineral composition of these dusts was reflected in dusts collected on the filters, both TSP and PM1, indicating their origin. Aluminosilicate glaze characteristic for fly ashes from combustion processes in power plants and combined heat and power plants were visible on SEM image of TSP filters ( Figure 7A). In central parts of SEM of TSP particulate, an aggregate consisting of oxides containing Zn, Fe, Mg, Mn as well as Al and Si was detected ( Figure 7B). The fine gray particles on Figures 6C and 7B were soot aggregates. The composition of the particles in the PM1 ( Figure 7D) containing Zn, As and Pb oxides corresponds to the dust of the non-ferrous metal industry. SEM observations showed the presence of metal chlorides in both industrial dust samples and those aspirated on filters ( Figure 6B,D and Figure 7C).

Human Health Risk Assessment
The composition of PM strongly determines its toxicity. The majority of elements contained in TSP and PM1 have a very negative impact on human health, even in very small quantities, causing serious diseases, including cancer. Non-cancer effects of metals include effects on the neurological, cardiovascular, hematological, gastrointestinal, musculoskeletal, immunological and epidermal systems. Epidemiological research has found consistent and coherent associations between long-term exposure and various health outcomes, such as reduced lung function, respiratory symptoms, chronic bronchitis and cardiopulmonary mortality [11]. Analyzing values of HQ, CR and HI (Table 7), i.e., the possible adverse health impact on humans (adults) due to inhaled metals and metalloids of TSP and PM1, the PM1 fraction exhibited only a slightly higher risk. However, it should be taken into account that the basis for the calculation of HQ and CR was only the concentration of particular elements, but the particle size was not taken into consideration. Furthermore, the finer the dust fraction, the deeper the penetration of impurities in the respiratory system and then into the human blood system [5,6]. However, all obtained indices are extremely high and the highest HQ values were obtained for toxic As, followed by Cd > Cr = Ni > Mn > Pb for TSP. Taking into account PM1 fraction the decreasing order of HQ was As > Cr > Cd > Ni > Mn > Pb, whereas the CR value of Cr was almost ten times as high as that of other elements followed by Ni, As and Cd. Moreover, as the HI in the area of study exceeded unity by several times, the health hazard posed by TSP and PM1 is very high. The children are more endangered then adults, thus USEPA [50] recommends application of a 10-fold adjustment for exposures during the first 2 years of life and 3-fold adjustment for exposures from ages 2 to 16, when assessing adverse impact of inhaled PTE on children's health. In this light, the non-cancer and cancer risk to children's health seems to be extremely high.
remarkably alike technogenic particles visible on SEM images of fly ashes from power plants as well as metallurgical dusts confirmed that they were sources of particles accumulated on filters ( Figures  6 and 7). Amorphous glass (mullit and quartz), quick coke, calcium sulfate and aluminosilicates with inclusions of spherical iron and manganese oxides were the main components of fly ashes ( Figure  6A), while oxides of Fe, Zn, Ca and Mn with varied morphology (spherules, stamens, aggregates) were characteristic for metallurgical dusts ( Figure 6B,C).

Conclusions
Obtained results revealed that the concentration of TSP and PM1 well correlated with their mass-specific magnetic susceptibility. The good relationship between the PM concentration and χ suggests that magnetic susceptibility measurements can be a good proxy of low-level atmospheric dust pollution.
The difference in χ for TSP and PM1 was generally low, indicating a comparable content of magnetic minerals in both TSP and PM1. Thus, PM1 was the main carrier of technogenic magnetic particles in the total deposited dust. Generally, enhanced values of χ indicated an increased number of Fe-rich particles in PM filters, but obtained results concerning Fe concentration and its correlation with χ did not confirm this statement. Probably, a foggy weather condition occurring during almost the whole investigated period can be the reason for an unexpected negative correlation between magnetic susceptibility (concentration of ferrimagnetic iron oxides) and Fe concentration [57]. On the other hand, χ exhibited a positive and significant correlation with the majority of determined PTE in the case of TSP, but not for the PM1 fraction-Only with Cu, Mn, and a negative significant correlation with Ba. So, the other elements were linked to susceptibility when the particles were larger than 1 µm, as it happens for non-exhaust emissions. Comparing the content of chemical elements in TSP and PM1, it was observed that concentration of V, Mn, Ni, As, Tl, Cr, Ti, Ba and Al was slightly higher in PM1, but the daily variability of elemental concentrations in TSP and PM1 looked similar. PM1/TSP ratios for PTE concentrations and χ were around or slightly above unity, which indicated that PM1 was the main carrier of PTE (with the exception of cadmium, copper and lead) and technogenic magnetic particles.
Correlation matrices and principal component analyses were applied in order to determine the sources of dusts aspirated on filters. In the case of both PM fractions two main groups of components were separated. A distinct group consisting of Al, Ba and Zn was distinguished for both size fractions of PM. High concentration of these elements was probably due to the influence of tyre abrasion, and Ba most likely originated from fly ashes of the coal used for heating in the study area. The second TSP cluster was characterized by high loadings of As, Cd and Pb which may be related to coal combustion, while the second PM1 cluster was associated with χ, As, Cd, Fe, Mn, Pb and Tl, elements which are produced as a by-product of non-ferrous metal smelting and refining. Obviously, the influence originated from steel metallurgy and vehicular emissions, but also that from soil and road dust re-suspension should not be excluded. SEM images and results of EDS analysis confirmed the presence of particles and minerals characteristic for the above-mentioned sources.
Surprisingly, despite not very high values of magnetic susceptibility and the PTE concentration, the human health risk indices (cancer and non-carcinogenic) determined on the basis of PTE concentration were very high, which confirms the relevance of this study and suggests that they should be continued.