Magnetic Emissions from Brake Wear are the Major Source of Airborne Particulate Matter Bioaccumulated by Lichens Exposed in Milan (Italy)

: The concentration of selected trace elements and the magnetic properties of samples of the lichen Evernia prunastri exposed for 3 months in Milan (Italy) were investigated to test if magnetic properties can be used as a proxy for the bioaccumulation of chemical elements in airborne particulate matter. Magnetic analysis showed intense properties driven by magnetite-like minerals, leading to signiﬁcant correlations between magnetic susceptibility and the concentration of Fe, Cr, Cu, and Sb. Selected magnetic particles were characterized by Scanning Electron Microscope and Energy Dispersion System microanalyses, and their composition, morphology and grain size supported their anthropogenic, non-exhaust origin. The overall combination of chemical, morphoscopic and magnetic analyses strongly suggested that brake abrasion from vehicles is the main source of the airborne particles accumulated by lichens. It is concluded that magnetic susceptibility is an excellent parameter for a simple, rapid and cost-e ﬀ ective characterization of atmospheric trace metal pollution using lichens.


Introduction
It has been estimated that more than 4 million people die annually for health problems caused by poor air quality [1]. Air quality in urban areas is a great concern worldwide, since most of the human population lives in cities that quite often do not meet the air quality standards, thus being exposed to high levels of pollutants [2]. Particulate matter (PM) is by far the most important air pollutant in urban areas [3], and short-and long-term exposure to PM is responsible for a wide array of negative effects on human health, spanning from inflammations, cardio-vascular diseases, and lung cancer [1]. Therefore, monitoring air pollution in urban areas is crucial for human health.
Automated stations are often used for monitoring air quality over time and space, but a real high-resolution network is often lacking owing to the constraints of establishing and maintaining sophisticated and costly equipment. Thus, biological monitoring (biomonitoring) may turn very useful in supporting and complementing the limited data derived from instrumental monitoring. In addition, biomonitoring may in turn help to improve the knowledge on the response of organisms to pollutants.

Lichen Sampling
The epiphytic (tree-inhabiting) lichen species Evernia prunastri (L.) Ach was selected for this study, being widely used in similar biomonitoring studies [13][14][15] and in spite of its documented capacity to bioaccumulate great amounts of trace elements and to reflect atmospheric deposition [5]. In addition, the fruticose (shrubby) growth form of this lichen allows easy handling of thalli. Apparently healthy thalli were collected from deciduous oak trees in a forested remote area of Tuscany (Central Italy), far removed from roads and any local pollution sources. Thalli were left to acclimate for 24 h in a climatic chamber at 16 • C, 40 µmol m −2 s −1 photons PAR (photoperiod of 12 h) and RH = 65% before preparation of samples. After cleaning with plastic tweezers from extraneous material such as bark or insects, samples of ca. 2 g were arranged into lichen bags made of plastic net (Figure 1) that were kept in a climatic chamber at 16 • C, 40 µmol m −2 s −1 photons PAR (photoperiod of 12 h) and RH = 65% until transplant. Samples were transplanted within 1 week from collection. Before transplantation, sample vitality was randomly evaluated by analyzing the photosynthetic efficiency, which resulted in being optimal. efficiency, which resulted in being optimal.
Samples were transplanted at 25 sites in Milan, selected following a stratified random design according to distance from the center (three belts: central, semiperipherical, peripherical), and the nine administrative city districts. In addition, a control site was selected at a relatively unpolluted forested area 50 km N of Milan, far from roads and local pollution sources. At each exposure site (and at the control site), five lichen bags were exposed from December 2018 to February 2019 at a height of ca. 2 m from ground, tied to the branches of lime (Tilia sp.) trees located along roads or inside parks. The exposure period (winter) was selected in such a way to assure a high metabolic activity of samples [8,15]; the duration of exposure of 3 months is regarded as optimal for E. prunastri [9]. After the exposure, samples were retrieved, air-dried and stored in paper bags at −20 °C until analysis. For the analysis, the lichen material inside each lichen bag was pooled. Having a control site accounting also for the possible effect of transplantation, unexposed samples were not assayed. Figure 1. A lichen bag exposed in the study area.

Chemical Analysis
The concentration of selected elements (Al, As, Cd, Cr, Cu, Fe, Pb, Sb, Zn) was assessed by ICP-MS (Sciex Elan 6100, Perkin-Elmer) after wet acid mineralization (3 mL 70% HNO3, 0.2 mL of 60% HF and 0.5 mL H2O2) in a microwave digestion system (Ethos 900, Milestone). A procedural blank and a sample of the certified reference material IAEA-336 "lichen" were included in each batch of samples. Recoveries were in the range of 90%-113% and the precision of analysis, expressed as relative standard deviation of five replicates, was within 10% for all elements. Results are expressed on a dry weight basis.

Magnetic Analysis
Dry lichen samples were placed in standard 8 cm 3 palaeomagnetic plastic cubes for the analysis of magnetic susceptibility and in pharmaceutical gel caps #4 for the hysteresis measurements. Mass magnetic susceptibility (k) was calculated dividing the values measured with a KLY5 Agico meter Samples were transplanted at 25 sites in Milan, selected following a stratified random design according to distance from the center (three belts: central, semiperipherical, peripherical), and the nine administrative city districts. In addition, a control site was selected at a relatively unpolluted forested area 50 km N of Milan, far from roads and local pollution sources. At each exposure site (and at the control site), five lichen bags were exposed from December 2018 to February 2019 at a height of ca. 2 m from ground, tied to the branches of lime (Tilia sp.) trees located along roads or inside parks. The exposure period (winter) was selected in such a way to assure a high metabolic activity of samples [8,15]; the duration of exposure of 3 months is regarded as optimal for E. prunastri [9]. After the exposure, samples were retrieved, air-dried and stored in paper bags at −20 • C until analysis. For the analysis, the lichen material inside each lichen bag was pooled. Having a control site accounting also for the possible effect of transplantation, unexposed samples were not assayed.

Chemical Analysis
The concentration of selected elements (Al, As, Cd, Cr, Cu, Fe, Pb, Sb, Zn) was assessed by ICP-MS (Sciex Elan 6100, Perkin-Elmer, Waltham, MA, USA) after wet acid mineralization (3 mL 70% HNO 3 , 0.2 mL of 60% HF and 0.5 mL H 2 O 2 ) in a microwave digestion system (Ethos 900, Milestone, Bomby Municipality, Denmark). A procedural blank and a sample of the certified reference material IAEA-336 "lichen" were included in each batch of samples. Recoveries were in the range of 90%-113% and the precision of analysis, expressed as relative standard deviation of five replicates, was within 10% for all elements. Results are expressed on a dry weight basis.

Magnetic Analysis
Dry lichen samples were placed in standard 8 cm 3 palaeomagnetic plastic cubes for the analysis of magnetic susceptibility and in pharmaceutical gel caps #4 for the hysteresis measurements. Mass magnetic susceptibility (k) was calculated dividing the values measured with a KLY5 Agico meter for the net weight of the samples. The coercive force (Bc), the saturation remanent magnetization by mass (Mrs) and the saturation magnetization by mass (Ms) were measured using a vibrating sample magnetometer (Micromag 3900, PMC) equipped with a carbon fibre probe, under cycling in a maximum field of 1.0 T. Concentration dependent hysteresis parameters were calculated subtracting the high field paramagnetic linear trend before dividing the magnetic moments for the net weight of the samples. The coercivity of remanence (Bcr) values were extrapolated from backfield remagnetization curves up to −1 T, following forward isothermal remanent magnetization up to a +1 T field.
The percentage decay of Mrs after 100 s was calculated as: where SP refers to the superparamagnetic fraction, Mrs 0 is the remanent magnetization measured as soon as the magnetic field is reduced to noise levels after the application of a 1 T field, and Mrs 100 is the remanence measured 100 s later. The values of Mrs (SP)% are indicatory of the fraction of remanent magnetization due to viscous magnetic components, which are usually carried by ultrafine magnetic particles, dimensionally in the superparamagnetic/stable single domain boundary, which is around 20-35 nm for magnetite [16,17]. The domain state and magnetic grain-size of the samples were compared to theoretical magnetite according to the hysteresis ratios Mrs/Ms vs. Bcr/Bc in the "Day plot" [18][19][20].
First order reversal curves (FORCs) [21,22] were measured using the Micromag operating software; FORC diagrams were processed, smoothed and drawn with the FORCINEL Igor Pro routine [23]. FORCs were measured in steps of 2.5 mT, with 300 ms averaging time and maximum applied field being 1.0 T. The optimum smoothing factor was evaluated by FORCINEL software.

Morphoscopic Observations
Fragments from representative thalli were cut into pieces, mounted onto carbon stubs and then coated with carbon in an Automatic Carbon Coater (Agar Scientific, Essex, UK). The sample surfaces were observed under a Scanning Electron Microscopy (SEM Jeol JSM-5310) equipped with an INCA Energy Dispersive Spectroscopy (EDS) system. Micrographs and spectra were collected with a 15 KV operating voltage, 50-70 µA current, 20 mm work distance, and a 15-17 spot size (JSM-5310 data). Micrographs were acquired by an INCA imagine capture system; spectra were collected by an INCA X-stream pulse processor.

Statistical Analysis
The magnitude of the relationship (effect size) between magnetic parameters and each bioaccumulated element was evaluated by the coefficient of determination R 2 , which estimates the proportion of variance shared by the two variables, as explained by a linear model. As recommended for environmental data [24], the probability of the effect size was estimated using the highest posterior density intervals following a Bayesian approach [25]. In addition, Cohen's ƒ 2 was also calculated [26] to class the effect size into small, medium and large [27]. Only R 2 values at the lower point of the 95% credible interval scored at least as medium size effect according to Cohen's ƒ 2 were retained. Prior to analysis, data were transformed to logarithms to achieve normal distributions, and centered to zero means in such a way that the intercept represents the value of the response variable at the mid-point of the predictor range, and the original variance is retained. All calculations were done with the free software R [28].

Bioaccumulation of Trace Elements
Lichen thalli transplanted at the control site showed values (Table 1) well within the range of background concentrations for this species [29]. After 3 months of exposure in Milan, samples of E. prunastri exhibited accumulation for all elements at all sites, with very few exceptions (Table 1), with mean values exceeding those of the control sample by a factor 1.2 (Cd) to 6.3 (Sb). Differences in concentrations across sites were in some cases remarkable, e.g., for Cu, which ranged between 6.5 and 55.4 µg/g dw. Table 1. Concentrations (µg/g dw) of trace elements in samples of the lichen Evernia prunastri exposed for 3 months at 25 sites in Milan, as well as at a control site (ctrl).

Sample
Mrs (SP)% was estimated for samples 11 and 22 (Figure 3e,f), the most and the least intense of the dataset, respectively, and showed that the contribution of rapidly decaying components to the overall magnetization is about 5%-6%. This result is only indicative, as errors are large due to the relatively low values of magnetization and the short averaging time (100 ms).
The Mrs/Ms vs. Bcr/Bc ratios ( Figure 4) indicated that the samples were distributed in the middle-right side of the plot, between the theoretical curves calculated for mixtures of single domain (SD) and multidomain (MD) magnetite grains and that calculated for a mixture of SD and superparamagnetic (SP) magnetite grains [19,20]. Overall, the hysteresis parameters indicated slightly variable concentrations of very similar soft ferrimagnetic minerals, presumably ascribable to magnetite. Consistently with susceptibility measurements, Ms and Mrs values were always higher than in the control sample. It was not possible to estimate the Curie temperature by means of magnetothermic curves, given the low susceptibility values and the noisy behaviour caused by burning of organic matter during heating.
Mrs (SP)% was estimated for samples 11 and 22 (Figure 3e,f), the most and the least intense of the dataset, respectively, and showed that the contribution of rapidly decaying components to the overall magnetization is about 5%-6%. This result is only indicative, as errors are large due to the relatively low values of magnetization and the short averaging time (100 ms).
The Mrs/Ms vs. Bcr/Bc ratios ( Figure 4) indicated that the samples were distributed in the middle-right side of the plot, between the theoretical curves calculated for mixtures of single domain (SD) and multidomain (MD) magnetite grains and that calculated for a mixture of SD and superparamagnetic (SP) magnetite grains [19,20].  [30], air filters from monitoring stations in Rome (green diamonds and squares) [31], different kinds of fuel exhausts (orange, yellow and pink dots) and brake dusts (purple squares) [32]. The SD (single domain), PSD (pseudo-single domain) and MD (multidomain) fields and the theorethical mixing trends for SD-MD and SP-SD pure magnetite particles (SP, superparamagnetic) are taken from Dunlop [19,20]. Modified after [32]. FORC diagrams ( Figure 5) were made for two samples (11 and 20), selected for their relatively intense magnetic properties; the distribution peaked close to the origin of the diagram and was dominated by viscous components of magnetization, typical for MD grains, without the asymmetrical features which usually suggest the presence of SP ultrafine particles. The two FORC diagrams are substantially identical, confirming that the samples contain variable concentrations of alike magnetic particles.

Figure 4. Bilogarithmic Day plot of the hysteresis ratios Mrs/Ms vs. Bcr/Bc for transplanted lichens
Evernia prunastri (red dots) compared with Quercus ilex leaves from high traffic areas of Rome (green triangles) [30], air filters from monitoring stations in Rome (green diamonds and squares) [31], different kinds of fuel exhausts (orange, yellow and pink dots) and brake dusts (purple squares) [32]. The SD (single domain), PSD (pseudo-single domain) and MD (multidomain) fields and the theorethical mixing trends for SD-MD and SP-SD pure magnetite particles (SP, superparamagnetic) are taken from Dunlop [19,20]. Modified after [32]. FORC diagrams ( Figure 5) were made for two samples (11 and 20), selected for their relatively intense magnetic properties; the distribution peaked close to the origin of the diagram and was dominated by viscous components of magnetization, typical for MD grains, without the asymmetrical features which usually suggest the presence of SP ultrafine particles. The two FORC diagrams are substantially identical, confirming that the samples contain variable concentrations of alike magnetic particles.

Morphoscopic Observations
Morphoscopic observations ( Figure 6) clearly showed the accumulation of Fe-rich particles, ranging from clusters of submicrometric grains (Figure 6a) to >10 μm, as well as coarser and sometimes flaky metallic residuals (Figure 6b). Their chemical composition, as determined by EDS X-ray microanalysis, in addition to Fe, highlighted the widespread presence of elements such as Zn, Cu and Ba (Figure 6c). The shape and the recurrent inclusion of various elements other than Fe make these magnetic particles different from natural stoichiometric magnetite. Most morphologies were irregular, with rough and/or rounded and frictional surfaces, dissimilar to the spherical shapes of industrial fly ashes originated by the combustion of black and brown coal (e.g., [33][34][35][36]) or waste electrical and electronic equipment [37].

Morphoscopic Observations
Morphoscopic observations ( Figure 6) clearly showed the accumulation of Fe-rich particles, ranging from clusters of submicrometric grains (Figure 6a) to >10 µm, as well as coarser and sometimes flaky Appl. Sci. 2020, 10, 2073 9 of 15 metallic residuals (Figure 6b). Their chemical composition, as determined by EDS X-ray microanalysis, in addition to Fe, highlighted the widespread presence of elements such as Zn, Cu and Ba (Figure 6c). The shape and the recurrent inclusion of various elements other than Fe make these magnetic particles different from natural stoichiometric magnetite. Most morphologies were irregular, with rough and/or rounded and frictional surfaces, dissimilar to the spherical shapes of industrial fly ashes originated by the combustion of black and brown coal (e.g., [33][34][35][36]) or waste electrical and electronic equipment [37].
Appl. Sci. 2020, 10, x FOR PEER REVIEW 10 of 17 Figure 6. FESEM images and EDS spectra of selected particles embedded in the lichens exposed in Milan; (a) Spectrum 1; cluster of submicrometric Fe-rich particles; (b) coarse and flaky iron-rich residuals; (c) micrometric particles, rich in Fe, Cu and Ba. Figure 6. FESEM images and EDS spectra of selected particles embedded in the lichens exposed in Milan; (a) Spectrum 1; cluster of submicrometric Fe-rich particles; (b) coarse and flaky iron-rich residuals; (c) micrometric particles, rich in Fe, Cu and Ba. Figure 7 shows the relationships between magnetic susceptibility and element concentrations, along with the 95% credible intervals (95CIs) of the R 2 values. The latter, along with the respective Cohen's f 2 output, are detailed in Table 3.

Effect Size
Appl. Sci. 2020, 10, x FOR PEER REVIEW 11 of 17 Figure 7 shows the relationships between magnetic susceptibility and element concentrations, along with the 95% credible intervals (95CIs) of the R 2 values. The latter, along with the respective Cohen's ƒ 2 output, are detailed in Table 3. Only the correlations with Fe, Cr, Cu, and Sb showed a Cohen's ƒ 2 at least medium. In addition, 95CIs of the R 2 values as well as Cohen's ƒ 2 were calculated also between the four chemical elements above (data not shown), showing strong intercorrelations among all elements, with Cohen's ƒ 2 always > medium. Aluminium, which is a major element in the Earth's crust and has limited metabolic significance in lichens, is commonly used as a tracer of geogenic inputs, and correlations with this Only the correlations with Fe, Cr, Cu, and Sb showed a Cohen's ƒ 2 at least medium. In addition, 95CIs of the R 2 values as well as Cohen's ƒ 2 were calculated also between the four chemical elements above (data not shown), showing strong intercorrelations among all elements, with Cohen's ƒ 2 always > medium. Aluminium, which is a major element in the Earth's crust and has limited metabolic significance in lichens, is commonly used as a tracer of geogenic inputs, and correlations with this element are roughly taken as an indication of soil contamination of samples [38]. 95CIs of R 2 and Cohen's ƒ 2 of Al with Fe, Cr, Cu and Sb (data not shown) highlighted only the correlation Al-Fe (R 2 = 0.20-0.63, ƒ 2 = 0.25 "medium"). Magnetic susceptibility was also very well correlated with the other two concentration-dependent magnetic parameters Ms and Mrs (Table 3).

Discussion
The results of bioaccumulation showed that Milan is quite homogeneously affected by high deposition loads of elements such as Cr, Cu, Fe, and Sb; some more spotted peaks emerged also for Pb. This outcome is consistent with other lichen biomonitoring studies in urban areas of Italy [37,[39][40][41][42]. The results are indicative of a common origin of pollutants from non-exhaust sources of vehicular traffic, such as brake abrasion. As a matter of fact, the above elements are used as components of brake systems [43], and brake abrasion is reported as the main source of trace elements in PM [44]. This conclusion is supported by the results of EDS X-ray microanalysis, which pinpointed the massive presence of Fe, and the widespread occurrence of Zn, Cu and Ba, the latter usually being connected with emissions from car brakes and tyres [32,45]. In addition, as shown by SEM analysis, Fe-rich particles are generally different from industrial spherical combustion particles, further indicating that their source are motor vehicles and more specifically, relying on their composition, break abrasion [32].
The range of the magnetic susceptibility values (7.8 to 35.8 × 10 −8 m 3 kg −1 ), when compared with previous studies on transplanted lichens, is indicative of a notable concentration of magnetic minerals and, consequently, of iron-rich bioaccumulated particles. Transplants of the lichen Pseudovernia furfuracea exposed for 4 months in a complex and heavily polluted area, ranged from 7.1 to 17.1 × 10 −8 m 3 /kg −1 [37]. Transplants of the same lichen species exposed for 2 months in an industrial area showed values in the range 0.4 to 7.4 × 10 −8 m 3 /kg −1 [46]. Transplants of Evernia prunastri exposed for 6 months around a cement production plant in Slovakia showed magnetic susceptibility ranging 1.3 to 8.1 × 10 −8 m 3 /kg −1 [47]. Bags of Parmotrema pilosum exposed to atmospheric pollutants over the course of 1 year increased from an initial mass-specific magnetic susceptibility value (mean ± S.D.) of 24.1 ± 5.0 × 10 −8 m 3 kg −1 to higher values up to 51.2 ± 23.0 × 10 −8 m 3 kg −1 [48]. This dataset included three sites, two at metallurgical factories, and one influenced by vehicular emissions, where values increased by 100%.
As emerged from the hysteresis loops, which are very similar and differ only for the concentration-dependent magnetic parameters, which follow the same trend of magnetic susceptibility, the variations of magnetic susceptibility are linked to different concentrations of magnetic particles similar for composition and grainsize. The coercivities range, the low saturation field and the SIRM/k values, spanning 2.4-7.1 kA/m, all suggest that magnetite-like minerals are the main magnetic carriers.
In the Day Plot, the lichens transplanted in Milan have been compared to the previous points obtained for Quercus ilex leaves sampled in Rome, Italy, PM filters and dusts arising from fuel exhausts and brakes' emissions, as discussed in [32]. The transplants markedly overlap the "brake" samples and fall at the lower right end of the cluster defined by the particles accumulated on leaves and filters, in a region of the plot falling in-between the theoretical trends for SD-MD and SD-SP pure magnetites, far from "diesel" and "gasoline" exhausts, which instead follow the theoretical trends for mixtures of single domain (SD) and multidomain (MD) magnetites. This result is at variance with the observations of [37] for combustion dusts accumulated by lichens exposed in an area subjected to arsons, where the magnetic mineralogy corresponded to magnetite-like minerals in PSD magnetic domain state/grain size, similarly to the diesel exhaust emissions in [32].
Also the SIRM/k values (5.0 ± 1.2 kA/m) are compatible with the brake emissions (6.7 ± 2.5 kA/m) and distinct from gasoline and diesel emissions (13.8 ± 6.3 kA/m and 14.5 ± 12.1 kA/m, respectively) reviewed in [49]. Even FORC diagrams confirm the uniform prevalence, irrespective of the concentration, of viscous components of magnetization, closely resembling the diagrams for brakes and leaves reported by [32].
It is recalled that the critical magnetic grain size transitions, theoretically determined for equidimensional magnetite, are about 0.03 µm for SP to SD, 0.08 µm for SD to PSD, and 17 µm for PSD to true MD [50]. Overall, it is not easy to attribute these magnetic features to SP or MD particles. In this sense, none of the magnetic properties diagnostic of the relevant presence of SP particlese.g., the enhancement of magnetic susceptibility, very low SIRM/k values, asymmetry along the Bu axis in FORC diagrams-supported the presence of ultrafine magnetic particles, although the latter could be linked to the choice of a relatively high averaging time (300 ms) during the measurements. As a further difficulty, in traffic-related PM, the SP fraction may occur as coating of MD particles and is originated by localized stress in the oxidized outer shell surrounding the unoxidized core of magnetite-like grains; thus it cannot be considered as a direct proxy for the overall content of ultrafine <30 nm particles [17].
On the other hand, the presence of clusters of fine and ultrafine magnetic particles is clearly shown by SEM observations; thus, the "intermediate" position in the Day plot might depend both on the chemically-impure composition of the magnetite-like particles and on the coexistence of SP and MD granulometric fractions.
The strength of the statistical associations between magnetic parameters and bioaccumulation of chemical elements corroborated the above conclusions that the main source of air pollution in Milan is non-exhaust emission from vehicular traffic, notably brake abrasions. The deposition of PM enriched in Fe, Cr, Cu, and Sb caused the bioaccumulation of these chemical elements and determined the magnetic susceptibility of lichen samples.
According to their strong intercorrelations, k, Ms and Mrs are indicative of the same magnetic fraction; thus, magnetic susceptibility is a valid, representative and fast proxy for the concentration of ferromagnetic particles associated with trace metals in PM.
Brake wear debris comprises primarily carbonaceous and metal-bearing components, including, besides Fe, heavy metals such as Cu, Sb, Sn, Cd, Cr, Pb, and Zn [43]. In [49], it was shown that Fe-bearing particles result from the friction between a brake pad and the cast iron brake disc; at brake pad temperatures below 200 • C, abrasive processes dominate, and wear particles >1 µm are mostly generated. At higher operational temperatures (>190 • C), the concentration of nanoparticles (<100 nm) increases due to evaporation, condensation and aggregation processes, increasing by four to six orders of magnitude the emission rate of particles, thus constituting >90% of total brake dusts and constituting a major source of airborne magnetic nanoparticles. It is possible to speculate that the coexistence of SP and MD particles is the result of both processes, and that SP particles, for their intrinsically deciduous magnetic properties, are more difficult to be seen with standard room temperature magnetic measurements.
As suggested by the relationship between Al and Fe, some proportion of these two metals may also originate from soil resuspension. This is quite common, and in addition, it is known that, although it is well known that brake wear constitutes one of the main sources of non-exhaust roadside PM, brake wear accounts for 50%-60% of the total non-exhaust emissions [51].

Conclusions
From the present results, it is possible to conclude that magnetic properties of lichen transplants are a robust proxy for the bioaccumulation of trace metals. The linear regression model with magnetic susceptibility, as determined by the magnetic fraction of PM, explained large proportions of variance for Fe, Cr, Cu, and Sb accumulated by E. prunastri. Chemical, magnetic and morphoscopic analysis clearly pinpointed non-exhaust sources of vehicular traffic, notably brake abrasion, with a broad grain-size range, as the main driver of PM air pollution in Milan.
The above relationships and the strong association of magnetic susceptibility with the concentration-dependent hysteresis parameters suggest that magnetic susceptibility is a simple, fast and very useful parameter that allows time-and cost-effective analysis of air pollution using lichen transplants.