Introduction

Atmospheric particulate matter (PM) is a portion of air pollution that was classified by the International Agency for Research on Cancer (IARC) as carcinogenic to humans due to their harmful effects, representing a threat for human health (IARC 2013). Several epidemiological studies have associated the atmospheric PM exposure with an increase of morbidity and mortality due to respiratory and cardiovascular diseases (Taioli et al. 2007; Brook et al. 2010; Anderson et al. 2012; Hoek et al. 2013; Quarato et al. 2017; Burnett et al. 2018; Simonetti et al. 2018; Tobías et al. 2018; Muñoz et al. 2019). PM encompasses many associated pollutants which are potential contributors to adverse health effects after their entry into the organism through inhalation (Galvão et al. 2018). Among the pollutants that may either induce or increase PM toxicity, polycyclic aromatic hydrocarbons (PAHs) are notable because they constitute a large and diverse class of organic molecules that are widespread in the environment. In the atmosphere, PAHs are found in gas (lower molecular weight PAHs, i.e. 2- and 3-ring molecules) and particulate (medium-high molecular weight PAHs, i.e. 4-, 5- and 6-ring molecules) phases, depending on atmospheric conditions and vapour pressures of the compounds (Dat and Chang 2017). PAH sorption onto fine respirable particulate phase facilitates the transport of PAH over long distances and contributes to increase the adverse health effects of PAHs compounds, which can penetrate deep into the bronchioles and alveoli of the lungs. PAHs are mainly formed as a result of incomplete combustion or pyrolysis of organic matter such as fossil fuel burning (motor vehicle emissions, industrial fumes, domestic heating), incineration of organic waste and coal and biomass burning, whereas natural emissions include forest fires, volcanic eruptions and hydrothermal processes (Gozzi et al. 2017). The occurrence of PAHs in the environment is an increasing concern because of their potential persistence, bioaccumulation and carcinogenic, mutagenic and endocrine disrupting effects in human health (Kim et al. 2013; Karimi et al. 2015; Abdel-Shafy and Mansour 2016). Special interest and attention have been paid to 16 PAHs designated as priority pollutants by United States Environmental Protection Agency (USEPA) (USEPA 1984). Among these priority PAHs, benzo(a)pyrene (BaP) is used as a marker for carcinogenic/mutagenic health risk assessment due to its being considered to exhibit the most adverse effect. In addition, BaP is the only PAH subjected to regulatory guidelines in PM, setting a value of 1 ng m−3 (annual mean concentration) in PM10 fraction (mean particulate matter which passes through a size-selective inlet with a 50% efficiency cut-off at 10-μm aerodynamic diameters) by the European Union (EU), seeking to avoid adverse effects of PAHs to health and the environment (EU 2004).

In the last decades and due to their carcinogenic/mutagenic properties, PAH concentrations have been closely monitored at urban, rural and industrial atmospheres around the world to assess human exposure to PAHs. However, only few studies have been focused at Southwest Atlantic European Coast sites (Arruti et al. 2012; Slezakova et al. 2013a, b; Albuquerque et al. 2016; Alves et al. 2016; Elorduy et al. 2016; Cerqueira and Matos 2019). Concentrations, possible emission sources and associated health risks of particulate PAHs in urban atmospheres of Lisbon (Cerqueira and Matos 2019), Porto (Slezakova et al. 2013a, b; Albuquerque et al. 2016) and Braga cities (Alves et al. 2016) in Portugal have been reported. Also, some data can be also found at two Atlantic cities of Spain: Bilbao city, northern Spain (Elorduy et al. 2016), urban cities (Santander, Castro Urdiales and Reinosa) and rural areas (Los Tojos) in the region of Cantabria (Arruti et al. 2012). This paper aims to provide information of seasonal variation, sources and carcinogenic/mutagenic risks of PM10-associated PAHs at the Southwest Atlantic facade of Europe, where few data are available. Furthermore, the relationship between PM10-bound PAHs and PM10 components such as major ions, metal(oid)s, equivalent black carbon (eBC) and UV-absorbing particulate matter (UVPM) was also investigated.

Materials and methods

Details and description of the study area

The sampling site is located at 5 m above sea level inside the downtown of the city of A Coruña (43° 22′ 04″ N 08° 25′ 08″ W), an Atlantic coastal city with a quarter of a million inhabitants in the northwest of Spain which occupies a small peninsula (Fig. 1). The climate of the city is humid oceanic characterized by moderate temperatures, constant relative humidity and frequent rainfalls throughout the year (especially during winter). Then, two seasons could be considered in the present study attending to meteorological similarities: summer (warm season: April–September) and winter (cold season: October–March). Several anthropogenic sources (Fig. 1) are accounted due to the proximity of several industrial areas near the city (aluminium processing, refinery, power plants, and municipal solid waste and biomass incinerators). Moreover, emissions from maritime and road traffic could be important anthropogenic sources. Due to the sea proximity, marine aerosol influence is also important and sea salt presence is significant.

Fig. 1
figure 1

Sampling point situation in A Coruña city (Spain). Source: Google Maps (satellite maps, without annotations) and Mapchart.net (https://mapchart.net/spain.html)

Full size image

Atmospheric particulate matter sample collection

Sampling and PM10 determination were according to the European Norm 12341 (EN 12341:2015) (UNE 2015). Samplings were performed during January to December 2017 using an automatic high-volume sampler DIGITEL DHA-80 (Hegnau, Switzerland) provided with a 10-μm diameter cut-off particle separator and pre-heated micro quartz fibre filters of 15 cm of diameter (Ahlstrom Munksjö MK 360, Falun, Sweden). Atmospheric particulate matter was collected at 30 m3 h−1 for 24 h, 0 to 24 h (UTC). Before and after sampling, filters were stabilized at 20 ± 1 °C and relative humidity of 45–50% for 48 h, for mass weighing by means of an analytical balance (Sartorius Genius, Goettingen, Germany) with a sensitivity of 0.01 mg. Once the gravimetric determination of PM10 was performed, PM10 filters were folded and wrapped in aluminium foils inside envelopes, put inside sealed plastic bags and stored in the lab in a freezer (−18 °C) until analysis to avoid the losses of volatile compounds, especially during summer season. A total of 65 samples were selected covering all the months of the campaign (one/two randomized samples at week, distributed over the year). Samples collected are in agreement with minimum sample number for indicative measurements, according to European Directive 2008/50/EC (EU 2008). Additionally, field blanks were also collected to decrease gravimetric bias due to filter handling during and/or after sampling, which were analysed following the same procedure as samples (procedure blanks).

PAH extraction procedure and quantification by HPLC-FLD

PAH content in PM10 samples was assessed following the method described by Piñeiro-Iglesias (Piñeiro-Iglesias et al. 2003, 2004) with some modifications. Briefly, four circular portions of each PM10 sample (area of 5.31 cm2) were cut and extracted by using hexane:acetone 1:1 assisted with microwave energy (ETHOS SEL microwave system, Milestone, Sorisole, BG, Italy). Extracts were cleaned up (Supelclean™ LC-Si SPE Tubes (Supelco, Steinheim, Germany)), concentrated (TurboVap® II Concentration Workstation (Biotage, Uppsala, Sweden)) and then to dryness using a gentle stream of N2. Extracts were reconstituted by using 500 μL of acetonitrile for subsequent analysis. PAHs were determined in PM10 extracts by using a chromatographic system (Waters, Milford, MA, USA) equipped with programmable fluorescence detector (Waters 2475) and a Waters® PAH C18 (250 × 4.6 mm i.d., 5 μm) column. A chromatographic separation and fluorescence conditions were performed basing on the work of Fernández-Amado et al. (Fernández-Amado et al. 2016). In supporting Information (SFigure 1) shows a chromatogram obtained for a mixture of target PAHs at 25 μg L−1 HPLC-FLD conditions used in this study. Moreover, details of quality assurance and control of PAH extraction and quantification procedures are also shown in Supporting Information.

Exposure assessment and health risk assessment

The carcinogenicity of environmental exposure to PM10-associated PAHs relatively to BaP equivalent (BaPTeq) concentration of 16 PAHs (Σ16BaPTeq) was calculated using the following formula (Gao et al. 2019):

$$ {\sum}_{16}\left[{\mathrm{BaP}}_{\mathrm{Teq}}\right]={\sum}_{i=1}^{i=16}\left(\left[{\mathrm{PAH}}_i\right]\ x\ {\mathrm{TEF}}_i\right) $$

where [PAHi] represents the concentration of each PAH in PM10 samples (expressed as ng m−3), and TEFi is the toxic equivalence factor of the PAHi relative to BaP. TEFs used in this study were 0.001 for naphthalene (Naph), acenaphthene (Ace), fluorene (Fl), phenanthrene (Phe), fluoranthene (Ft) and pyrene (Pyr); 0.01 for anthracene (Ant), chrysene (Chry) and benzo(g,h,i)perylene (BghiP); 0.1 for benzo(a)anthracene (BaA), benzo(b)fluoranthene (BbF), benzo(k)fluoranthene (BkF) and indeno(1,2,3-cd)pyrene (IP); 1.0 for benzo(e)pyrene (BeP) and BaP; and 5.0 for dibenz(a,h)anthracene (DBahA) (Samburova et al. 2017).

Similarly, the total mutagenicity of PM10-associated PAHs relatively to BaP mutagenic equivalent (BaPMeq) concentration of 8 USEPA priority PAHs (Σ8BaPMeq) was calculated using the following formula (Durant et al. 1996):

$$ {\sum}_8\left[{\mathrm{BaP}}_{\mathrm{Meq}}\right]={\sum}_{i=1}^{i=8}\left(\left[{\mathrm{PAH}}_i\right]\ x\ {\mathrm{MEF}}_i\right) $$

where MEFi is the mutagenic equivalence factor (MEF) value of the PAHi. MEFs used in this study were 0.082, 0.017, 0.25, 0.11, 1, 0.29, 0.19 and 0.31, for BaA, Chry, BbF, BkF, BaP, DBahA, BghiP and IP, respectively (Durant et al. 1996).

Incremental lifetime cancer risk refers to the probability of an individual who is exposed to PM10-associated PAHs by inhalation during his or her lifetime were also estimated (USEPA 2005; WHO 2010). The ILCRinh values were calculated using the following formula:

$$ {\mathrm{ILCR}}_{\mathrm{inh}}={\mathrm{IUR}}_{\mathrm{BaP}}\ x\ {\sum}_{16}\left[{\mathrm{BaP}}_{\mathrm{Teq}}\right] $$

where ILCRinh is the incremental lifetime cancer risks resulting from the PM10 inhalation pathway and IURBaP is the inhalation unit cancer risk factor (expressed as ng m−3)−1 for BaP. IURBaP describes the theoretical upper limit possibility of contracting cancer calculated for a daily exposure to BaP concentration in air of 1.0 μg m−3 for a 70-year average life span (OEHHA 2005; USEPA 2011). Quantitative cancer risk estimates of PAHs as air pollutants are very uncertain because of the lack of useful and good-quality data and due to an estimated IURBaP based on epidemiology studies of populations from A Coruña is not available, several international accepted values for estimated IURBaP, based on different approaches (epidemiology study on coke-oven workers in Pennsylvania, IURBaP = 8.7 × 10−5 (ng m−3)−1 (WHO 2000); and studies on the data respiratory tract tumours in hamsters chronically exposed by inhalation to BaP, 1.1 × 10−6 (ng m−3)−1 (OEHHA 2005) and 6.0 × 10−7 (ng m−3)−1 (USEPA 2017)) were used in this work to calculate the PM10-bound PAH ILCRinh. When ILCRinh ≤10−6 denotes negligible or virtual safety under most regarding regulatory programmes, ILCRinh among 10−6 and 10−4 suggests a potential risk and ILCRinh > 10−4 imply potentially high risk (Wang et al. 2011; Cao et al. 2019; Ghanavati et al. 2019).

Additional data

Major ion and trace metal(oid) quantification

Major ions in PM10 samples were measured by zone capillary electrophoresis (ZCE) after aqueous extraction and metal(oid)s were measured by inductively coupled plasma mass spectrometry (ICP-MS) after acid extraction. Both procedures were previously optimized (Blanco-Heras et al. 2008; Moreda-Piñeiro et al. 2015). Statistical summary for the concentrations of major ions and metals in 65 PM10 samples are shown in STable 1–2.

Equivalent black carbon and UV-absorbing particulate matter quantification

eBC and UVPM were measured by using a Magee SootScan™ OT-21 (Berkeley, CA, USA) transmissometer at two wavelengths: measurements at 880 nm are interpreted as a measure of light-absorbing carbon analogous to black carbon present on the filter, while measurements at 370 nm are designated as UVPM indicator of aromatic organic compounds (Davy et al. 2017; Greilinger et al. 2019). Results obtained in PM10 samples are shown in STable 1.

Backward trajectory analysis

Backward trajectories were calculated at 2500, 1500 and 750 m above mean sea level (AMSL) 120 h before the time of the arrival to study site using the NOAA Hybrid Single-Particle Lagrangian Integrated Trajectory Model (HYSPLIT) model (Draxler and Rolph 2003).

Statistical treatment of data

For analytical data treatment, the ANOVA test was performed in order to verify the statistical significance of the differences between the means assessment; Pearson coefficients were calculated in order to investigate the existence of possible associations; principal component analysis (PCA) and cluster analysis (CA) were performed in order to identify the PAHs emission sources in the study area; and multivariate analysis based on quadratic surface model was performed to assess statistically significant relationship between variables. All statistical procedures were performed using the statistical package Statgraphics version 7.0 routine (Statgraphics Graphics Corporation, ST.SC., USA) at the 95% confidence level. In this study, PCA analysis was performed, after half-range and central value transformation, by utilising the orthogonal transformation method with Varimax rotation and retention of principal components which show eigenvalues higher than 1.0. Finally, CA was performed after half-range and central value transformation, using Ward’s clustering method and squared Euclidian distance metric. When PAH concentrations were below the quantification limits, values were approximated to LOQs/2 for statistical calculations. All PAHs, except for Naph, were used to estimate total PAHs bound to PM10 (Σ15PAH i.e. the sum of concentrations of individual species) assessment. However, Naph, Ace, Fl and Ant were not included in univariate, multivariate, cluster and PCA analyses because of having found concentrations below LOQs for the majority of PM10 samples.

Results and discussion

PM10 mass concentrations

According to the gravimetric measurements, average PM10 mass concentrations during the whole period were 23.5 μg m−3 (Table 1). The highest PM10 mass concentration was measured for one sample (94 μg m−3) during the winter season (on 15 October 2017), which exceeds the daily limit value of 50 μg m−3 set in the European Directive 2008/50/EC (EU 2008). This was the only daily exceedance observed in analysed samples, which cannot be exceeded more than 35 times per year (EU 2008). The Saharan dust incursion during 12 to 17 October 2017 at the sampling site could explain the PM10 mass concentration of 94 μg m−3 during 15 October 2017 (see Backward Trajectory Analysis, SFigure 2). Despite, the average PM10 mass concentration during the winter season corresponds with 25.7 μg m−3. Also, PM10 concentrations in summer and winter seasons showed no statistically significant differences (the p-value of the F-test is higher than 0.05) at 95.0% confidence level (STable 3).

Table 1 Average, maximum (max), minimum (min) and RSD (%) values of PM10 mass (μg m-3), individual PM10-bound PAHs and PAH summations (of 15 PAHs (Σ15PAHs), carcinogenic PAHs (ΣPAHc), non-carcinogenic PAHs (ΣPAHnc), 4 rings PAHs ΣPAH4rings and 5–6 rings PAHs ΣPAH5–6rings) concentrations (ng m−3)
Full size table

Atmospheric particle-bound PAH concentrations in PM10

Summary data of PAHs concentrations (mean, maximum, minimum and RSD) in PM10 samples during the whole sampling period studied and during summer and winter seasons are reported in Table 1. BbF and BeP were predominant PAHs in all samples collected with average concentrations of 1.6 and 1.5 ng m−3, respectively (Table 1). Other PAHs followed the sequence BghiP (0.88 ng m−3) > Chry (0.85 ng m−3) > IP (0.61 ng m−3) > BaP (0.48 ng m−3) > BkF (0.44 ng m−3) > Ft (0.33 ng m−3) > Pyr (0.31 ng m−3) > BaA (0.25 ng m−3) > Phe (0.18 ng m−3) > DBahA (0.08 ng m-3) ~ Ace (0.08 ng m−3) > Fl (0.02 ng m−3) and Ant (0.009 ng m−3) (Table 1). Naph (most volatile PAH mainly present in the gas phase) were the least abundant with concentrations lower than LOQ in all PM10 samples. Other volatile PAHs (Ace, Fl and Ant) offer concentrations lower than their LOQs in almost all PM10 samples (mainly during warm months). Target PAHs demonstrated high variation from event to event (Table 1), which reflect inherent heterogeneity of the atmospheric particles (PM10 sources) and the variability of meteorological conditions.

Despite differences in the number of target PAHs measured and sampling period times, the sum of PAH concentrations (Σ15PAHs) achieved in this study was compared with ΣPAH values reported in PM10 samples from other urban sites collected at Southwest Atlantic European Coast. The annual range (0.58 to 33.4 ng m−3, Table 1) is higher than those found in Lisbon (Portugal), 0.11 to 8.2 ng m−3 range (for Σ10PAHs) (Cerqueira and Matos 2019) and Bilbao (Spain), 1.2 to 9.8 ng m−3 range (diurnal range, sampling period of 8 h, for Σ13PAHs) (Elorduy et al. 2016) and lower than reported in Oporto (Portugal), from 30.8 (annual mean during 2014) to 126.5 (annual mean during 2004) ng m−3 (for Σ10PAHs) (Albuquerque et al. 2016) and 16.8 to 149 ng m−3 range, with a mean of 70 ng m−3 range (for Σ18PAHs in air, i.e. sum of concentrations of all PAHs both in gas phase and in PM10 particles) (Slezakova et al. 2013a, b).

Furthermore, the average annual concentration of BaP (0.48 ng m−3) was significantly lower than the annual value set in 1.0 ng m−3 by the European Directive (EU 2004). However, this limit was exceeded by 9 samples during the studied period (2 and 7 samples during summer and winter seasons, respectively). However, the average BaP concentration is higher than previous reported data, i.e. <LOD to 0.85 ng m−3 (Cerqueira and Matos 2019), 0.15 to 0.94 ng m−3 (Albuquerque et al. 2016), 0.06 to 0.70 ng m−3 (Elorduy et al. 2016) and 0.04 to 0.15 ng m−3 (Arruti et al. 2012) in PM10 samples collected at several Southwest Atlantic European cities (Lisbon (Cerqueira and Matos 2019), Oporto (Albuquerque et al. 2016), Bilbao (Elorduy et al. 2016) and Santander and Reinosa cities (Arruti et al. 2012) respectively).

Higher mean PAH concentrations were obtained during the winter than summer season. However, after applying ANOVA (STable 3), statistically significant differences (at 95.0% confidence level) between summer and winter seasons were not found (p-value > 0.05). The continuous harbour activity, as well as maritime and road traffic in the area, might explain the occurrence of PM10-bound PAHs throughout the year, with no differences between seasons.

BbF and BeP amounted to 41, 45 and 38% of the total PAHs bound to PM10 (Σ15PAH i.e. the sum of concentrations of individual species) during the whole sampling period and during summer and winter seasons, respectively (Fig. 2). In Fig. 2 is also shown the relative percentage contributions of other priority PAHs (which contribution is less than 1%) during the whole sampling period and during summer and winter times.

Fig. 2
figure 2

Compositions in percentages of PAHs during 1-year sampling and summer and winter seasons

Full size image

The carcinogenic PAHs (BbF, Chry, IP, BaP, BkF, BaA and DBahA) concentrations (ΣcPAH) and non-carcinogenic PAHs (Ace, Fl, Phe, Ant, Ft, Pyr, BeP, BghiP) concentrations (ΣncPAH) represented 58–60% and 40–42% of PAHs, respectively. The distribution of carcinogenic and non-carcinogenic PAHs in PM10 was similar for both seasons. BbF was the most abundant carcinogenic PAH, representing 21, 23 and 20% of PAH in mean during a year, summer season and winter season, respectively.

Distribution of PAHs on molecular weight basis shown that lower ring-number molecules, 2–3 rings, Σ2–3ringsPAH, (Ace, Fl, Phe and Ant) are partitioned mainly into the gas phase; thus, Σ2–3ringsPAH accounted for only 2–4%. Middle ring-number molecules, (Σ4ringsPAH), (Ft, Pyr, BaA and Chry) accounted for 37–44%, while high ring-number molecules, (Σ5–6ringsPAH), (BeP, BbF, BkF, BaP, IP, DBahA, BghiP) accounted for 71–76% (Fig. 2). The distribution of low, middle and high ring-number molecules was constant in the two seasons. Also, the high proportions of high ring-number molecules and the good correlation between Σ15PAH and BghiP content during the whole sampling period (R2=0.9771) and both seasons (R2=0.9817 and 0.9701 for summer and winter seasons, respectively) (SFigure 3) indicate that high-temperature processes such as combustion of fuels derived from harbour activities as well as road and ship traffic could be the main sources of particulate PAHs at sampling site.

Source apportionment of PAHs

A study about the association between particle-bound PAH concentrations and PAH sources was assessed. Molecular indices (PAH ratios) and the relationship between PAH concentrations and the contents of major ions, metal(oid)s, eBC and UVPM in PM10 samples were studied by univariate and multivariate analyses, principal component analysis (PCA) and cluster analysis (CA).

Molecular indices

Molecular indices based on PAH physical-chemical behaviour covariability were usually used as diagnostic ratios to identify and estimate the contribution of main PAHs sources: pyrogenic combustion (incomplete combustion of petroleum fuel and vehicular exhaust emission) and petrogenic sources (unburned crude oil and petroleum products) (Yunker et al. 2002). Nonetheless, the overlapping of different PAH sources and the transformation/decomposition of some PAH species due to their reactivity into the atmosphere make difficult the identification of potential emission sources of PAHs. Then, the analysis of more PAHs such as alkylated PAHs would provide a more reliable source identification.

Several commonly used diagnostic ratios, i.e. Ft/(Ft+Pyr), BaA/(BaA+Chry), BaP/BghiP, IP/(IP+BghiP) and BaP/(BaP+Chry) ratios were selected to distinguish the major source PAHs in the study area. Ft and Pyr are pyrogenic products derived from high-temperature condensation of lower molecular weight aromatic compounds, Ft being less thermodynamically stable than Pyr. The predominance of Ft over Pyr is characteristic of a pyrogenic process, while in petroleum-derived PAHs, Pyr is more abundant than Ft. A ratio of Ft/(Ft+Pyr) > 0.5 suggests emissions from coal and biomass burning, Ft/(Ft+Pyr) < 0.4 petrogenic sources, while Ft/(Ft+Pyr) ratios between 0.4 and 0.5 would point out vehicle emissions (Yunker et al. 2002). Chry, BaA, BghiP and IP are derived from processes of organic matter combustion at high temperature. Thus, BaA/(BaA+Chry) ratio > 0.35 suggests combustion of fossil fuels, < 0.2 petrogenic sources (unburned fossil material) and the range between 0.2 and 0.35 would indicate the mixture of petroleum and combustion sources (Yunker et al. 2002). Also, the BaP/BghiP ratio is used to differentiate traffic and non-traffic sources, then BaP/BghiP < 0.6 would indicate vehicles emissions (Jamhari et al. 2014). The IP/(IP+BghiP) ratio is used to distinguish liquid fossil fuel combustion from biomass and coal combustion sources, IP/(IP+BghiP) ratios < 0.5 suggests fossil fuel combustion, while IP/(IP+BghiP) ratios > 0.5 suggests coal/biomass combustion sources (Yunker et al. 2002; Vicente et al. 2018). Finally, the BaP/(BaP+Chry) ratio was used to distinguish diesel combustion (< 0.5) and gasoline combustion (> 0.5) sources (Teixeira et al. 2012).

Despite the limitations of this approach and the fact that the averages of BaA/(BaA+Chry), BaP/BghiP and BaP/(BaP+Chry) ratios for the warm season were statistically significant different (p < 0.05) from that of the cold season (Table 2), Ft/(Ft+Pyr), BaA/(BaA+Chry) and IP/(IP+BghiP) ratios would suggest pyrogenic sources (incomplete fuel combustion) for PAHs during summer and winter periods (Table 2). BaA/(BaA+Chry), IP/(IP+BghiP) and BaP/BghiP ratios also suggested mixed sources, fuel combustion source and vehicle emissions during both seasons, respectively. Also, most of the samples shown a BaP/(BaP+Chry) ratio lower than 0.5, which suggests that PAH emissions from diesel combustion prevailed over gasoline combustion during total period (Table 2). Finally, in Fig. 3a, b is presented a plot of BaA/(BaA+Chry) versus IP/(IP+BghiP) ratios and IP/(IP+BghiP) versus BaP/BghiP ratios respectively, suggesting pyrogenic sources (fuel combustion) during both seasons. Furthermore, diesel combustion (most likely due to the harbour activities and vehicular emissions) was observed to be the predominant source of PAHs during both seasons when plotting BaP/(BaP+Chry) ratio versus Ft/(Ft+Pyr) ratio (Fig. 3c).

Table 2 Values (average ± SDa) estimated for diagnostic PAHs ratios and p-values achieved after statistical comparison of average PAH ratios during summer and winter seasons
Full size table
Fig. 3
figure 3

Cross-plots for selected PAHs diagnostic ratios: a BaA/(BaA+Chry) ratio vs. IP/(IP+BghiP) ratio. b IP/(IP+BghiP) ratio vs. BaP/BghiP ratio. c BaP/(BaP+Chry) ratio vs. Ft/(Ft+Pyr) ratio

Full size image

Univariate analysis

Pearson product moment correlations of Σ12PAH, ΣcPAH, ΣncPAH, Σ4ringsPAH and Σ5–6ringsPAH contents with major ions, metal(oid)s, eBC and UVPM (at a confidence level of 95%) were assessed for samples collected for 1 year long (STable 4). Results show that Σ12PAH, ΣcPAH, ΣncPAH, Σ4ringsPAH and Σ5–6ringsPAH contents are correlated with NO3−, As, Bi and Zn (anthropogenic sources, including fossil fuel combustion and road traffic). Finally, a positive correlation between eBC (p-value < 0.0002) and UVPM (p value < 0.0008) were found for Σ12PAHs, ΣcPAHs, ΣncPAHs, Σ4ringsPAHs and Σ5–6ringsPAHs. Moreover, a correlation was found for Σ4ringsPAHs with anthropogenic tracer such as Pb (p value=0.0220) and Sb (p value=0.0282). Similar conclusions were achieved for samples collected during winter and summer seasons.

Principal component and cluster analysis

PCA has been first attempted with a data set in which Phe, Ft, Pyr, BaA, Chry, BeP, BbF, BkF, BaP, DBahA, BghiP and IP concentrations were the discriminating variables and 65 (1-year period), 32 (summer season) and 33 (winter season) PM10 samples were the objects. Results (Fig. 4) shown that over 91.45, 96.21 and 91.07% of the variance can be explained by 2 principal components (PCs) for 1-year period, summer and winter season, respectively. The first factor (PC1) was associated with high loadings of high molecular weight PAHs including Chry, BeP, BbF, BkF, BaP, DBahA, BghiP and IP, which were usually abundant in gasoline vehicle emissions and coal combustion (Iakovides et al. 2019). The second factor (PC2) with high loadings of Phe, Ft, Pyr and BaA indicated emissions from diesel vehicles (Teixeira et al. 2015) and coal-biomass combustion (Iakovides et al. 2019). However, high factor loadings were achieved for Pyr in the PC1 during summertime.

Fig. 4
figure 4

Proportion of variance contributions in percentage from PCA analysis for Phe, Ft, Pyr, BaA, Chry, BeP, BbF, BkF, BaP, DBahA, BghiP and IP concentrations in PM10 samples collected during the whole period (N=65) and summer (N=32) and winter (N=33) seasons

Full size image

PCA has been also attempted with a data set in which ion (Cl−, NO3−, SO42−, NH4+, K+, Na+, Ca2+ and Mg2+), metal(oid) (Al, Ba, Bi, Cu, Fe, Sb, Pb and Zn) concentrations and Σ12PAHs were the discriminating variables and 65 (1-year period) PM10 samples were the objects. Results show that 72.68% of the total variance was explained by 4 principal components (PCs) which show eigenvalues higher than 1.0 (Fig. 5). K+, Ba, Ca2+, Mg2+, Al and Fe are the main features in PC1 (crustal/terrestrial source), explaining 36.24% of total variance. The PC2 (vehicle traffic source and fossil fuel combustion) was loaded with Σ12PAHs, NO3−, Ba, Bi, Cu, Pb, Sb and Zn (18.04% of the total variance); Pb and Zn are tracers of traffic emission and Cu and Zn could originate from road traffic (Moreda-Piñeiro et al. 2015). PC3 (sea salt source) offers the highest weights for Cl−, Na+ and Mg2+ (10.59% of the total variance). Also, PC4 is composed of NO3−, SO42− and NH4+ (secondary anthropogenic source) explaining 7.81% of total variance. Finally, cluster analysis (CA) was also performed by using major ion, metal(oid) and Σ12PAH concentrations as variables and 65 (1-year period) PM10 samples as objects. CA (Fig. 6) shows that three clusters are formed; the first cluster is formed between Σ12PAHs, NO3−, SO42− and NH4+, Bi, Cu, Pb, Sb and Zn (fuel combustion and vehicle emission source); Al, Ba, Ca2+, Fe and K+ are the second cluster (crustal source); and the third cluster is formed between Cl−, Mg2+ and Na+ (marine source). Similar conclusions were achieved after PCA and CA for samples collected during winter and summer seasons.

Fig. 5
figure 5

Source composition and proportion of variance contributions in percentage from PCA analysis for PM10 samples collected during the study period (N=65). Σ12PAH= Phe, Ft, Pyr, BaA, Chry, BeP, BbF, BkF, BaP, DBahA, BghiP and IP

Full size image
Fig. 6
figure 6

Dendrogram of the cluster analysis (CA) of major ions, metal(oid)s and Σ12PAH. Σ12PAH = Phe, Ft, Pyr, BaA, Chry, BeP, BbF, BkF, BaP, DBahA, BghiP and IP

Full size image

Multivariate analysis

A multivariate analysis based on linear regression model was performed to test the relationship of major ion and metal(oid) contents and PAH concentrations. As shown in Table 3, predicted PAH concentrations are very close to the measured/observed PAH concentrations (p-value <0.05 for Phe and BaA; and p-values <0.01 for Ft, Pyr, Chry, BeP, BbF, BkF, BaP, DBahA, BghiP, IP and Σ12PAHs) when anthropogenic and traffic road tracers (NO3−, Ba, Bi, Cu, Pb, Sb and Zn) were fitted to a quadratic model. The quadratic regression coefficient (R2) and p-values are also listed in Table 3.

Table 3 Lineal model of the variables NO3−, Ba, Bi, Cu, Pb, Sb and Zn and PAH concentrations
Full size table

PAH toxicity evaluation

The annual average Σ15[BaPTeq] concentration was 2.7 ± 3.0 ng m−3, while the average concentrations during summer and winter seasons were 2.9 ± 3.4 and 3.2 ± 2.6 ng m−3, respectively (Table 4). Concentrations obtained in this study were generally higher than results reported in urban sites of the Southwest Atlantic facade of Europe (Lisbon, 0.185 ± 0.232 ng m-3 (Cerqueira and Matos 2019), Oporto, 0.21–3.8 ng m−3 (Slezakova et al. 2013a, b) and Santander, 0.11–0.23 ng m−3 (Arruti et al. 2012)). Averages of Σ16[BaPTeq] concentrations for the warm season were not statistically significant different from that of the cold season (p > 0.05). BeP (53–57% of the total Σ15[BaPTeq] concentrations) was the major contributor to Σ15[BaPTeq] concentrations in both seasons, followed by BaP (14–20%) and DBahA (14–17%), while Phe, Ft, Pyr, Chry and BghiP (< 0.3%) contributed to a lesser extent.

Table 4 Σ15[BaPTeq] and Σ8[BaPMeq] concentrations (average ± SD, ng m−3) and Incremental Lifetime Cancer Risk (ILCR) assessment regarding inhalation exposure of the PM10-bound PAH 70-year average life span
Full size table

Regarding mutagenic activity, the annual median Σ8[BaPMeq] was 1.4 ± 1.4 ng m−3, although average concentrations were higher during wintertime to those obtained for summer (Table 4). Any seasonal dependence was found for mutagenic activity (p > 0.05). BaP and BbF were major contributors of Σ8[BaPMeq] concentrations, accounting for 29–37% (BaP) and 28–35% (BbF). The contribution of IP and BghiP was also important (14–15% and 12–13% for IP and BghiP, respectively).

Estimated ILCRinh values for PM10-bound PAH inhalation exposure (Table 4) show similar values during both seasons. According to the data, average ILCRs exceeded the risk level of 1.0×10−4 when using the value of IURBaP of 8.7 × 10−5 (recommended by WHO (WHO 2000)), suggesting a potentially high risk for human health in the studied area. However, the lung cancer risk was reduced to a potential cancer risk when IURBaP values of 1.1 × 10−6 (recommended by OEHHA (OEHHA 2005)) and 6.0 × 10−7 (recommended by USEPA (USEPA 2017)) were considered. In addition, it is important to point out that PAH health risk assessment considering ILCRs strongly depends on TEFs, IURs and number of PAHs considered. Then, future efforts focused on the standardization of PAH health risk by using the ILCR approach would be interesting in order to facilitate interpretation and obtain more comparable data.

Conclusions

Sixteen PAH concentrations in PM10 collected at a European urban site (Northwest of Spain) mean novel contribution to the knowledge of sources of PAHs in an Atlantic coastal European region. In general, it was found that the concentration of PAHs found at this site was higher than most of other ones reported at Atlantic coast European sites. The average annual concentration of BaP concentration (0.48 ng m−3) did not exceed the target value in PM10 set in directive 2004/107/EC. BbF and BeP were the major PAHs found in PM10 samples analysed. BghiP, Chry and IP were the next most important contributors. Around 58–60% of the total PAH concentration is attributed to carcinogenic PAHs. Also, PAH concentrations were dominated by higher molecular PAHs (5-ring and 6-ring PAHs recorded around 71–76%). Statistically seasonal differences were not found for all target PAHs. Data from molecular indices and univariate, principal component and cluster analyses suggest a pyrogenic origin (due to the continuous harbour activity and vehicle emissions, which is present throughout the year) of PAHs at urban site during both seasons (summer and winter seasons). Also, statistical multivariate analysis has shown that PAH concentrations are correlated with PM10-bound NO3−, Ba, Bi, Cu, Pb, Sb and Zn. Furthermore, PCA suggested that sea salt, anthropogenic and crustal sources were the main PM10 formation sources in the studied area. Finally, the health risk assessment based on ILCRinh approach suggested high potential cancer risk for long-term exposure to PM10-associated PAHs in study area during the sampling period. However, this outcome should be viewed with some caution since it assumes the use of dose response functions based on epidemiology study on coke-oven workers in Pennsylvania and studies on the data from hamsters chronically exposed by inhalation to BaP, but not using data from epidemiology studies associated to PAH inhalation in population from A Coruña city. The present findings are a first step towards a better knowledge of PAH health risk assessment via inhalation in A Coruña city.