Micro-morphological analysis of foliar uptake and retention of airborne particulate matter (PM)-bound toxic metals: implications for their phytoremediation

ABSTRACT While airborne particulate matter (PM) pollution is a serious problem for urban environments, it can be reduced through uptake by plant leaves. In this study, we investigated and compared the uptake of PM-bound toxic metals by different plant species. Enrichment factor (EF) and correlation analyses across different sample types indicated anthropogenic origins of these toxic metals with airborne source signatures. Scanning electron microscopy (SEM) analyses of both leaf surfaces (adaxial and abaxial) suggested that the micro-morphological properties of the leaf surface (e.g. stomata, trichomes, epicuticular wax, and epidermal appendages) control the accumulation of PM and associated metals in plant leaves. Senna siamea leaves showed the most micro-morphological variation as well as the maximum concentration of toxic metals. It was found that foliar uptake of PM-bound toxic metals is affected by leaf surface morphological characteristics. Our results imply that plant speciation strategies can be used to help decrease PM pollution.

composition and structures, texture of both surfaces (adaxial and abaxial), and stomata type) and quantitative (numbers of stomata and trichomes) micro-morphological leaf features [30][31][32]. Hence, micro-morphological studies of leaf surfaces can be employed to assess the PM pollution load in urban environments [29,32].
Although the accumulation of these toxic metals in plant roots via soil has been extensively studied, relatively limited research has focused on the uptake pattern of toxic metals by plant leaves through the airborne route. The aim of the present study was to assess the accumulation potential of airborne toxic metals by plant leaves from different locations of an urban environment in India. Plants naturally growing in six different sites of Bilaspur city were selected for detailed study. To this end, Pb, Cd, and Cu were selected as target metals for analysis in different plant and environmental samples. This is because Pb, Cd, and Cu were commonly found in PM deposited at leaf surfaces of different plant species from various urban locations, especially from the study area [29,[33][34][35][36][37]. Furthermore, it has been reported that Pb and Cd are listed as hazardous air pollutants (HAPs) in the list of USEPA [38,39]. Traffic and industrial emissions were found responsible for Cu build-up in urban road side environments [40,41]. Vehicular activities including exhaust emissions, tyre wear, brake wear emission, metallic corrosion of engine wears of running vehicles, and coal combustion in industrial activities also release these metals in urban environment [42][43][44][45]. Therefore, these toxic metals can easily bind to suspended PM and further deposit leaf surfaces of roadside plants [25,46]. As such, the objectives of this study were (1) to assess the abundance of toxic metal content (Pb, Cd, and Cu) relative to Fe content in order to derive an enrichment factor in the leaves of the selected plants for foliar dust, road dust, and soil samples; (2) to explore the metal source processes; and (3) to assess the effect of micro-morphological leaf surface properties (adaxial and abaxial) on the accumulation of PM and associated toxic metals. Our findings offer valuable insights into the role of plant species in controlling PM pollution in urban environments.

Study site
The study was conducted in Bilaspur (E = 82°08`28.32, N = 22°05`9.6), which is located in the state of Chhattisgarh in central India. The selected area included the city centre and covered both commercial and residential areas of the city. All sites are major squares of the city with strong traffic source activity ( Figure S1). Site 1 is the traffic square located near Govt. Girls Higher Secondary School. This site was affected by transport activities of both light and heavy vehicles. Site 2 is a market square located near a bridge that connects the city. Site 3 and site 4 are located on National Highway (NH) 49, which has high traffic-related emission sources. Site 5 is located near Apollo Hospital Square, which connects the Rajkishor Nagar, Vasant Vihar, and Lingyadih residential areas. Site 6 is located on a city bypass road that connects NTPC (National Thermal Power Corporation) Sipat to NH130 Raipur Road.

Selection of plant species
Six angiospermic plant species with different leaf surface characteristics were selected for this study (Ficus religiosa, Mangifera indica, Butea monosperma, Alstonia scholaris, Azadirachta indica, and Senna siamea). These species are widespread, remain green yearround, and are very common in subtropical areas.
F. religiosa (Family: Moraceae) is a species of large evergreen trees. Its leaves are generally 5-12 cm long and 5-8 cm wide, with alternate reticulate intercostae, a leathery appearance, an ovate-lanceolate (caudate) shape, and a green colour when they are mature. M. indica (Family: Anacardiaceae) is considered to be the most delicious Indian fruit plant.
Its evergreen trees are erect with broad canopies, and its leaves are alternate and clustered at the tip of the branch. Its leaves are 12-20 cm long and 6-8 cm wide, pinnate, and have reticulate/intercostae venation. B. monosperma (Family: Fabaceae) is a medium-sized deciduous tree with a bent trunk and irregularly arranged branches. The leaves are compound, alternate/spirally arranged on the stem, trifoliate, 8-12 cm long, and 5-7 cm wide. They also have an acute apex with an ovate shape, reticulate venation, and a prominent midrib. A. scholaris (Family: Apocynaceae) is a tall ornamental tree with rough grey bark. The branches emerge from the same place in the main stem and are arranged in a whorled pattern. The leaves are leathery, and one whorl contains 4-7 leaves. Leaf shape is similar to that of leaves on the mango tree. A. indica (Family: Meliaceae) is an evergreen tree that grows vertically erect and has greyish brown bark. Its leaves are imparipinnate and alternate with 12-20 cm long rachis. The leaflets generally contain 10-14 leaves, which are arranged in opposite/ sub-opposite positions. The leaves are 4-6 cm long, 2-3 cm wide, and have a serrate margin. One pair contains 12-20 leaves. The leaves also display intercostae reticulate venation with a prominent midrib. S. siamea (Family: Fabaceae) is a deciduous tree of medium height, up to 15-20 m. It has a wide spreading canopy, and its leaves are pinnately compound, 25-35 cm long, and 10-15 cm wide. The leaflets are arranged in the opposite position in 10-14 pairs. The leaves are dark green and have an ovate shape, entire margins, and reticulate venation [47,48].
In the present study, all plants showed similar phenological conditions. Therefore, at the time of collection, the leaves of the different plants were similarly matured. Special care was also taken at the time of sampling to select leaves equally in terms of morphological damage, bird faeces, dynamics such as chlorosis and necrosis, and uneven dust deposition.
Sampling of leaves, foliar dust, road dust, and soil samples and sample preparation Weather conditions were relatively stable within 2 weeks before sampling. The average temperature was 23.3°C and average relative humidity was 49.8%. Rainfall was not recorded prior to two weeks of sampling (Source: https://www.worldweatheronline. com/). Dusted leaves from each site were collected at a height of 2.5 m above the ground. Ten leaves were collected each study site from branches facing towards the road side. Steel trays and fine plastic hair brushes were used to collect road dust from the edge of road. Soil samples from each site were collected from the surface at a depth of 40 cm by the scraper plate method [18]. Considering the average depth of the plant fine roots, the soil-surface was dug to the same depth. Four soil samples (10 g each) were collected from each site from the same profile. All four samples were mixed to create a single uniform sample that represented the soil of that particular site. All samples were kept in air-tight Ziploc polyethylene bags and brought to the laboratory. The sampled dusted leaves were well stirred in a 200 mL glass beaker with 100 mL micro-distilled water for 10 min. In this suspension, foliar dust particles sticking to the surfaces of the leaves were separated. To obtain dust, the suspension was vaporised at 50°C on a hot plate. The washed leaves were oven dried for 24 h at 50°C. After drying, the leaves were crushed with a mortar and pestle and passed through the sieve to obtain a fine powder. Road dust and soil samples were also kept at room temperature for 2 days, after which they were ground with a mortar and pestle and collected into separate beakers by weighing 0.5 g of each. All samples were digested separately into aquaregia solution (3HNO 3 : HCl); sample preparation for metal analysis was performed according to the method of [18].

Enrichment factor (EF)
The EF assesses the relative contribution of manmade sources to toxic metal levels in collected samples. The EF was calculated according to equation (1) where S (E) is the concentration of a target metal (E) in the examined environmental sample, S (R) is the concentration of the reference metal in the examined environmental sample, C (E) is the concentration of a target metal (E) in the crust, and C (R) is the concentration of the reference metal in the crust. The C (E) values for Pb, Cd, and Cu were 20, 0.098 (98 ppb), and 25 ppm, respectively. Fe was used as the reference metal (C (R) ) with a crustal value of 35,000 ppm (3.5 wt. %) [49]. Based on EF values, the following categories were used for pollution levels and source (natural or anthropogenic): EF < 2 = low enrichment, 2 ≤ EF < 5 = moderate enrichment, 5 ≤ EF < 20 = significant enrichment (low anthropogenic emission), 20 ≤ EF < 40 = very high enrichment (moderate anthropogenic emission), and EF ≥ 40 = extremely high enrichment (high anthropogenic emission) [50].

Scanning electron microscopy
Fresh leaves were thoroughly washed with ionised water, blotted dry, and then cut into square sections (about 0.5 cm wide × 0.5 cm long) using razor blades. The leaves were cut from the margins and the mid-rib area. The leaf samples were kept in a mixture of 2.5% glutaraldehyde solution overnight to enable prefixation, after which the samples were kept in 2% osmium tetroxide for post-fixation for one hour. The samples were washed twice with phosphate buffer solution for 15 min and then passed through a series of acetone solutions (30%, 50%, 70%, 95%, and 100%) for dehydration. Drying was performed in a critical point drier (CPD) using CO 2 as a carrier gas. An aluminum stub with an adhesive surface was used to mount the leaf samples. The stub was coated with 40-60 nm gold using a Denton Vacuum. SEM analysis of samples was carried out with an FEI NOVA NANO SEM-450 scanning electron microscope equipped with different detectors.

Statistical analysis
The statistical analyses were carried out using the statistical package for the social science, version 22 (SPSS-22, IBM, Chicago, USA). One-way Analysis of variance (ANOVA) was applied to analyze the significant differences in concentrations of target toxic metals measured in different sample types. The level of significance (p value) was set at different levels (<0.05 and <0.001). The correlation coefficient was used to determine different significance level (p < 0.05 to p < 0.01) between leaf vs. foliar dust, road dust, and soil samples.

Concentrations of metals in different matrices
The concentrations of the four target metals in the four sample types are shown in Figure 1. Site 6 (S. siamea) showed the highest concentrations of all metals in foliar dust (Table S1). The maximum concentration of Fe in foliar dust in the present study was more than 10 times higher than that observed in a previous study in Bilaspur, whereas it was nearly 2 times higher for Pb [20]. The maximum concentration of Cd in foliar dust was comparable with that observed in Miskolc (industrial city), Hungary (4 ± 1 mg kg −1 : Platanus × acerifolia) [27]. This comparative study was performed with samples from high traffic and active urban sites and showed traffic-related sources for these toxic metals.

Source apportionment of metals and relationships between the different matrices
The results of ANOVA showed the significant differences (p < 0.05 and 0.001 level) among concentration of metals in road dust, leaf, and foliar dust sample within study sites ( Figure  1). A correlation analysis was conducted between metal concentrations within different samples to determine their sources of origin. In foliar dust, significant correlations were found between Pb vs. Cd (r = 0.93, p < 0.01), Pb vs. Cu (r = 0.85, p < 0.05), and Cd vs. Cu (r = 0.89, p < 0.01) [ Table S2a (i)]. Similar results were found in leaf and road dust samples [Tables S2a (ii) and (iii)]. However, in soil, the correlations were not significant, implying negligible anthropogenic influence. A correlation analysis was also conducted between leaf and other samples (foliar dust, road dust, and soil) ( Table S2b). All metals (Fe, Pb, Cd, and Cu) showed significant correlations (p < 0.05, 0.01) between leaf vs. foliar dust and leaf vs. road dust.

Enrichment factor (EF)
To identify the sources of toxic metals in the four samples (foliar dust, leaves, road dust, and soil) and thus determine whether the toxic metals are derived from anthropogenic activity or natural sources, the EFs were calculated (Table S3).
In foliar dust, the EFs for Pb, Cd, and Cu showed moderate to very high enrichment, which suggests anthropogenic origin. In leaf samples, the EFs of Pb and Cu showed similar levels of enrichment (significant enrichment). This finding was comparable with the EF value of Tilia spp. and Aesculus hippocastanum leaves (EF > 10) isolated from an urban area in Belgrade, Serbia [59]. However, for Cd, the EFs ranged from 22.  [50,60]. The high EF value of Cd in the S. siamea leaf samples is comparable with a previous study [18]. In road dust, Pb and Cu showed moderate to significant EFs. The EFs of Cd ranged from 9.20 (B. monosperma) to 25.7 (S. siamea), indicating significant to very high enrichment. A similar pattern of Pb and Cu EFs in road dust was observed by [61], which was mainly associated with traffic emission and vehicular part corrosion. The road dust EFs of Pb and Cu were comparable with the road dust EFs from Delhi, India, which indicated an anthropogenic source of origin [62]. In soil samples, all selected metals had an EF of 2 or below, suggesting that they originated from natural sources or were background concentrations.

SEM-based micro-morphological study
The micro-morphological parameters considered for each plant are listed in Table 1.

F. religiosa
In F. religiosa, emergence was visible at the adaxial surface of the mid-rib portion of the leaves, which can facilitate deposition of PM (Figure 2(a-c)). On the abaxial surface, epicuticular waxy platelets were found. Waxy platelets were arranged in rosette-like structures on which PM was deposited in a range of 3.244-4.169 µm (Figure 2(d,e)). Moreover, stomata were found on the abaxial surface. These stomata were covered by PM and had a mean opening of 1.719 µm × 904.6 nm (Figure 2(f)).

M. indica
The adaxial surface of M. indica leaves shows rough morphological variations due to sparse arrangement of the cuticular layer [62]. This roughness permits the deposition of PM of different sizes (929.8 nm to 14.08 µm, Figure 3(a,b)). For instance, fine PM (0.1-2.5 µm) was deposited on the surface (Figure 3(c)). The abaxial surface was found to be rougher, presumably due to the presence of the resin secretor gland, stomata, and epicuticular wax with deposited PM (Figure 3(d,e)). The abaxial surface of M. indica has been reported to have a granular form of epicuticular wax surrounding the stomata, a wrinkled cuticular surface, and reticulate-shaped glands, all of which encourage the retention and accumulation of PM [63]. As such, the stomata were able to uptake fine PM with a diameter below 1000 nm (e.g. 931.4 and 840.2 nm) (Figure 3(f)).   (Figure 4(a,c)). This morphology facilitated deposition of large particles up to 173.4 × 55.76 µm (Figure 4(b)) on the adaxial surface and below 17.32 µm on the abaxial surface (Figure 4(d)). The stomata were not seen clearly in the SEM images as they may have been covered due to the excessive density of the trichomes and deposited PM.

A. scholaris
The leaves of A. scholaris have projection-like structures on both surfaces near the milk secretor glands with epicuticular wax (Apocynaceae). The presence of trichomes and appendages facilitated high deposition and accumulation of PM in the fine (PM 2.5 ) to respirable (PM 10 ) range on the adaxial surface (1.41-4.40 µm) ( Figure 5(a,b)). However, the significant deposition of PM damaged the leaf surface, leading to accumulation of PM inside the leaf ( Figure 5(c)). In addition, trichomes were found in higher density on the abaxial surface than on the adaxial surface ( Figure 5(d,e)), and fine PM deposits (0.54 µm to 3.21 µm) were observed on the stomata and rough surface of the groove in the mid-rib portion of the leaf ( Figure 5(f)). The waxy layer covering the surface of the gland facilitated entrapment of PM ( Figure 5(g,h)). The deposited PM aggregated and accumulated on the abaxial surface, thus forming a deposition layer over the surface ( Figure 5(i)).

A. indica
The leaves of A. indica also showed different micro-morphological characteristics when compared to other plants. On the adaxial surface, trichomes were present only on the midrib portion ( Figure 6(a,b)). The morphological features of the trichome surface facilitated the accumulation of PM on the surface (Figure 6(c)). The presence of stomata and the raised surface of the epidermal cell wall make the abaxial surface rougher; this surface showed more PM deposition than the adaxial surface ( Figure 6(d,e)). Moreover, the abaxial surface had stomata that facilitated PM accumulation through stomatal pores ( Figure 6(e)). Granulated epicuticular wax and the raised epidermal cell wall structure also have been shown to enable PM accumulation on both surfaces [64].

S. siamea
On the adaxial surface of S. siamea leaves, dust-loaded epidermis, irregular patches of an epicuticular waxy layer, and dense trichomes were observed (Figure 7(a,b)). Grooves were found at the surface of the mid-rib area; these grooves were covered with a wax lining that facilitates the deposition of small PM (Figure 7(c)). The mean length of the observed trichomes was 60.76 µm, with deposition of PM with sizes up to the fine range (712.2 nm to 10.32 µm) (Figure 7(d)). Moreover, PM was deposited in and around stomatal openings on the adaxial surfaces (Figure 7(e)). Hence, PM-bound toxic metals can potentially accumulate in stomata on the leaf surface [20]. Similarly, the abaxial surfaces showed high trichome density, epicuticular wax, and stomata that facilitate deposition and increase the efficiency of PM accumulation (Figure 7(f-h)). Fine PM (e.g. 152.8 nm to 714.5 nm) was deposited over the stomata and readily accumulated in the internal stomatal pores (7.823 × 2.127 µm) (Figure 7(i,j)). The concentration levels of target toxic metals varied considerably across the different sample types such as foliar dust, road dust, and leaves. Comparison of the data with those of previous studies indicated that the samples were significantly enriched with the target metals. The lack of correlation between leaf vs. soil samples indicates that metals can accumulate in leaves from foliar dust or road dust resuspension, suggesting the  significance of metals of airborne origin [20]. Likewise, the EF results demonstrated that the Pb, Cd, and Cu in the foliar dust, leaf, and road dust samples arose from anthropogenic activities. The EFs of the various samples can be ranked as follows: (leaf > foliar dust > road dust > soil). Thus, the metals in the leaves came mainly from airborne sources (e.g. traffic sources and other urban sources) and not from soil.
Leaf surface roughness is determined by grooves and bulges of arranged epidermis cells, the presence of epicuticular wax, secretor glands, veins, and number/density of stomata and trichomes on both surfaces [65,31,18]. S. siamea showed more diverse leaf surface micro-morphological features (epicuticular wax, stomata, and trichomes) followed by A scholaris, A. indica, M. indica, F. religiosa, and B. monosperma, respectively (Figure 7(aj), Table 1). F. religiosa and M. indica showed the similar characters, i.e. presence of epicuticular wax and stomata on abaxial surfaces. Whereas, B. monosperma has only trichomes on both surfaces with high density. In addition, leaf surfaces of M. indica did not show the presence of trichomes as compared to A. indica. However, due to the more numbers and large opening of stomata, they show more accumulation of metals [66]. It has been previously observed that in comparison to the density and size of the stomata, the pattern and dynamics of the stomatal opening is rather more important factor for foliar PM accumulation [67]. This suggests that the high complexity in external morphology is helpful in the greater deposition of PM (Figure 4). On the other hand, the foliar accumulation of the metals from PM depends on a number of factors. For instance, the dynamics of stomatal opening and its arrangement is the most important factor in accumulation of metals. In addition, external supporting morphology is also helpful. The deposited PM entered directly into the stomata according to its ultra fine size ranges and accumulated inside (e.g. Figure 3 and 7). Moreover, it can also convert toxic metals in ionic state that enters into the leaves through the aqueous pores which are present on cuticular edge of stomatal guard cells, and with the help of the epidermal cuticular anticlinal wall [68,69].

Conclusions
Based on the results of concentration levels of target metals in different plant and environmental samples, EFs, and correlation analyses in this study, it was evident that the target metals bound with PM have anthropogenic source signatures. Moreover, there was a line of confirmation that these metals were partitioned through airborne route which is thought to be less prevalent than their accumulation from soil via roots of the plants. The results of this study further revealed that leaves of different plants have unique micro-morphological traits that facilitate the retention and accumulation of PM in different size ranges. Based on these micro-morphological features of the leaves and the accumulation potential of PM (or PM-bound toxic metals), the six plants are ranked as follows: S. siamea > A. scholaris > M. indica > F. religiosa > A. indica > B. monosperma. Hence, it was found that the particle type, particle diameter class, and assessment scale should all be considered in order to accurately evaluate the comprehensive particle retention abilities of an urban tree species. Based on our findings, these plants can be used for phytomonitoring and management of PM (PM-bound toxic metals) in urban environments. The study also recommends more research covering diverse sampling locations and plant species with variable source characteristics and environmental conditions. Before choosing plant species for urban forestry and greening programmes, it is also important to consider socioeconomic and aesthetic factors.

Disclosure statement
No potential conflict of interest was reported by the author(s).