Relationship between the amount of black carbon particles deposited on the leaf surface and leaf surface traits in nine urban greening tree species

Abstract To select urban greening tree species suitable for the purification of the atmosphere polluted by black carbon (BC) particles, it is necessary to clarify the determinants of the amount of BC particles deposited on the tree leaves. In the present study, we investigated the relationship between the amount of BC particles that were deposited from the atmosphere and firmly adhered to the leaf epicuticular wax, and leaf surface traits in seedlings of nine tree species grown for two years under natural conditions (Fuchu, Tokyo, Japan). There was a significant interspecific difference in the maximum amount of BC particles deposited on the leaf surface, and the order was as follows: Ilex rotunda > Cornus florida > Osmanthus fragrans > Cornus kousa > Quercus glauca ≒ Quercus myrsinifolia > Magnolia kobus ≒ Zelkova serrata ≒ Styrax japonicus. In the nine tree species, significant highly positive correlations were observed between the amount of BC particles deposited on the leaf surface, and the hydrophobicity of leaf epicuticular wax determined by its chemical composition. Therefore, we concluded that the hydrophobicity of leaf epicuticular wax is an important determinant of the amount of BC particles deposited on the leaf surface of urban greening tree species. NOVELTY STATEMENT This is the first paper that shows that the hydrophobicity of leaf epicuticular wax is an important determinant of the amount of BC particles deposited on the leaf surface of urban greening tree species. This study will provide the basic and novel information for the selection of urban greening tree species suitable for the purification of the air polluted by BC particles.


Introduction
In East Asia, the amount of pollutants emitted into the atmosphere has rapidly increased due to the increase in fuel consumption associated with rapid economic development since the 1980s (Kurokawa and Ohara 2020). Black carbon (BC) particle with an aerodynamic diameter of 2.5 mm or less (PM 2.5 ) is one of the air pollutants that can affect human health in Asian megacities (Mori et al. 2020). In general, BC particles are generated by combustion engines, such as diesel engine, power stations, field burning of agricultural wastes, forest and vegetation fires, and residential burning of coal and wood (WHO 2012). Air pollution caused by BC particles originating from automobile exhaust gas has adverse effects on the health of pedestrians and residents in Asian urban areas (Pani et al. 2020). Relatively high concentrations of particulate matter (PM) contribute to the occurrence of respiratory and cardiovascular diseases (Liu et al. 2015;Ravindra 2019), and air pollution caused by BC particles is feared to adversely affect human health (Janssen et al. 2011;WHO 2016). Therefore, it is necessary to take measures against the source of BC particles, and reduce the atmospheric concentration in Asian urban areas.
In general, leaves can absorb gases through the stomata and deposit PM on the leaf surface from the atmosphere. Vegetation in urban areas is important because it can reduce the atmospheric concentration of air pollutants (Beckett et al. 1998;Zhang et al. 2017;Xu et al. 2019). Air purification by urban greening tree species is expected to mitigate particulate air pollutants, such as BC particles, because trees can purify PM from the urban atmosphere (Blanusa et al. 2015;Mullaney et al. 2015;Chaudhary and Rathore 2018). Brantley et al. (2014) reported that a maximum of 22% of BC particle concentration in the atmosphere was reduced by placing trees along the road. Currently, however, very little information is available on Asian urban greening tree species that are suitable for the purification of PM, such as BC particles from the urban atmosphere (Xu et al. 2018;He et al. 2020;Li et al. 2021;Park et al. 2022).
To select urban greening tree species suitable for the purification of the urban atmosphere polluted by BC particles, it is necessary to clarify the interspecific differences and seasonal changes in the amount of BC particles deposited on leaf surfaces. However, very few information is available on the interspecific difference and seasonal change in the amount of BC particle on leaf surface, and its relating factors of Japanese tree species Takahashi et al. 2023). Several researchers have reported that interspecific differences in the amount of PM deposited on the leaf surface are correlated with leaf surface traits, such as the amount of cuticular wax, water-repellence, microstructure of the cuticular wax layer, incidence of leaf surface micro-roughness due to epicuticular waxes and the presence of stomata and trichomes (Burkhardt et al. 1995;Neinhuis and Barthlott 1998;Dzierzanowski et al. 2011;Saebø et al. 2012;Terzaghi et al. 2013;Song et al. 2015;Perini et al. 2017). Lin and Schuepp (1996) investigated the relationship between the leaf PM retention ability and the microstructure of leaf surfaces of 10 tree species, and reported that leaf surfaces with dense and narrow grooves, strip-like projections, high roughness and high wettability had strong retention abilities of PM. Popek et al. (2013) and Łukowski et al. (2020) reported that there was a significant positive correlation between the leaf PM amount and leaf cuticular wax amount. He et al. (2020) reported that a significant negative correlation was observed between the leaf PM amount and contact angle of water droplets on the leaf surface. Chaudhary and Rathore (2019) and Chiam et al. (2019) reported that deposition of PM on leaves depends on the morphological and anatomical leaf traits, such as specific leaf area and leaf surface roughness. Recently, we reported that seasonal variations in the amount of BC particles on the leaves of deciduous and evergreen broad-leaved urban trees are mainly regulated by leaf surface traits, such as water-repellence and the amount of epicuticular wax on the leaf surface, and environmental factors, such as atmospheric concentration of BC particles, respectively (Takahashi et al. 2023). Therefore, leaf surface traits are closely related to the deposition of BC particles on tree leaves. However, very little information is available on the interspecific difference in the amount of BC particles deposited on the leaf surface of urban greening tree species, and its relationship with leaf surface traits. At the present time, therefore, leaf traits determining the interspecific difference in the amount of BC particles deposited on the leaf surface of urban greening tree species are not consistent in the previous studies and identified. Furthermore, it is not clarified the relationship between the amount of BC particles deposited on the leaf surface, and the composition and chemical property of epicuticular wax.
In the present study, we investigated the interspecific difference in the amount of BC particle on the leaf surface and its related leaf surface traits of typical nine Japanese urban greening tree species (four evergreen and five deciduous broad-leaved tree species, NILIM 2018). Based on the previous studies suggesting that the amount and property of epicuticular wax is one of the important determinants relating to the amount of PM deposited on the leaf surface (Burkhardt et al. 1995;Lin and Schuepp 1996;Terzaghi et al. 2013;Song et al. 2015;Takahashi et al. 2023), we hypothesized that the leaf surface characteristic that determines the amount of BC particles deposited on the leaf surface of urban greening tree species is the hydrophobicity of epicuticular wax. To test this hypothesis, in the present study, we investigated the relationship between the amount of BC particles deposited on the leaf surface, and the composition and chemical property of epicuticular wax in seedlings of the nine urban greening tree species.

Plant materials
As described in Takahashi et al. (2023), seedlings with a height of 1.0-1.5 m of four evergreen (Quercus glauca Thunb., Quercus myrsinifolia Blume, Osmanthus fragrans Lour. var. aurantiacus Makino, and Ilex rotunda Thunb.) and five deciduous (Zelkova serrata (Thunb.) Makino, Styrax japonicus Siebold & Zucc., Magnolia kobus DC., Cornus kousa Buerger ex Hance subsp. kousa, and Cornus florida L.) broad-leaved tree species were grown in 2019 and 2020 at the Field Museum Fuchu (FM Fuchu) of Tokyo University of Agriculture and Technology (Fuchu, Tokyo, Japan) under natural conditions. On 6 October 2017, the four evergreen broad-leaved tree seedlings were planted in 7-L pots filled with black soil (Andisol), and Z. serrata, S. japonicus, M. kobus, and C. kousa seedlings were planted in 7-L pots filled with the same black soil on 10 February 2018. On 4 March 2018, C. florida seedlings were planted in 14-L pots filled with black soil. During the growth period, all seedlings were irrigated as necessary with tap water and fertilized at 2-week intervals with 100 mL of liquid fertilizer diluted 1,000 times (Hyponex, N:P:K ¼ 6:10:5, Hyponex Japan Co. Ltd., Osaka, Japan). Although there were no specific sources of BC particle near the tree growing site, automobile traffic like diesel engine car in the surrounding area is considered to be the main source of BC particle.

Measurement of environmental factors
From January 2019 to December 2020, air temperature, relative air humidity, precipitation amount, precipitation duration, wind speed, and atmospheric concentration of BC particle were measured at the FM Fuchu by the methods described in Takahashi et al. (2023), and the monthly mean values of environmental factors are shown in supplemental Table S.1. In Japan, an air quality standard for PM 2.5 was introduced in 2009, with an annual mean value of 15 mg m À3 and a daily mean value of 35 mg m À3 . In 2019 and 2020, the monthly mean atmospheric concentration of BC particle was relatively low in summer and autumn, and relatively high in spring and winter (supplemental Table S.1). The minimum and maximum monthly mean atmospheric concentrations of BC particle were 0.70 mg m À3 in June and 0.93 mg m À3 in December in 2019, and 0.56 mg m À3 in July and 0.85 mg m À3 in February in 2020, respectively (supplemental Table S.1).

Quantification of BC particles deposited on the leaf surface
The leaves attached 50-100 cm above the surface of the potted soil were randomly selected for BC deposition measurement to prevent the difference in the BC amount on the leaves depending on the vertical positions of the leaves among the nine tree species. On April 21-22, May 22-23, June 17-19, July 15-16, August 15-18, September 14-16, October 14-18, November 20-22, December 16-18, 2019, January 21-22, February 18-20, March 18-19, June 15-17, July 15-16, August 11-14, September 22-24, October 14-17, andNovember 11-13, 2020, the leaves were randomly sampled from the seedlings, and then the sampled leaves were separated by the leaf emergence time (1st, 2nd or 3rd flush leaves for deciduous tree species, and 1st flush leaves in 2019 or that in 2020 for evergreen tree species, which can be easily identified by distinguishing branching and leaf position).
The amount of BC particles deposited on the leaf surface was measured following the methods (Takamatsu et al. 2001;Takahashi et al. 2023). The particles targeted in the present study were BC particles that firmly adhered to the epicuticular wax on the leaf surface even when the leaf was washed with water (in-wax BC particle). The sampled leaves were first washed with 150 mL of deionized water for 3 min. The water-washed leaves were then dried for 30 min in an oven (DX402, Yamato Scientific Co., Ltd., Japan) at 40 C to remove water from their surfaces. Subsequently, the area of the leaves (adaxial and abaxial surfaces) was measured using a digital camera (IXY DIGITAL 920 IS, Canon Inc., Japan) and image analysis software (LIA32 ver. 0.3781, Nagoya University, Japan). Then, the leaves were washed with 150 mL of chloroform for 20 s. The particles suspended in chloroform were collected on a quartz fiber filter (QR-100, Advantec MFS, Inc., Japan) by gravitational filtration. After air-drying the filters for 30 min, their absorbance at 510 nm was measured using a spectrophotometer with an integrating sphere (U-4100, Hitachi High Technology Corp., Japan). The BC amount collected on the filters was determined as the amount of regular elemental carbon (EC) following the thermal optical reflectance (TOR) method (IMPROVE protocol) using an OC/EC carbon analyzer (Model 2001 A, DRI, USA). The BC amount on the filter was determined for 13-25% and 18-25% of the leaf samples obtained from the four evergreen and five deciduous broad-leaved tree species, respectively. According to the linear relationship between the BC amounts measured by the OC/EC carbon analyzer and the absorbance at 510 nm of BC on the filter in each tree species (supplemental Figure S.1), the BC amount on the leaves of each tree species was calculated. The BC amount on the leaves was expressed as the amount of elemental carbon (EC) on the basis of double-sided leaf area (mg C m À2 double-sided leaf area).

Observation of leaf surface
During the growing period, the leaves of the nine tree species were randomly sampled to observe their surfaces using scanning electron microscopy (SEM). The harvested leaves were air-dried for two days in desiccators, fixed onto specimen stubs, and coated with gold using a sputter coater. The leaf surfaces were observed under a scanning electron microscope (JCM-5000, JEOL Ltd., Tokyo, Japan) at an accelerating voltage of 10 kV.

Measurement of roughness on the leaf surface
During the growing period, the leaves were randomly sampled every two months from the seedlings (May 14, July 17-19, September 24-26, November 12-15, 2019, January 15-16, March 10, June 24-26, August 18-21 and October 19-22, 2020), and two pieces (approximately 0.5 Â 0.5 cm 2 ) were excised from the center of the lamina of each leaf using scissors and a razor. The leaf pieces obtained from the adaxial and abaxial leaf surfaces were fixed on a glass slide using double-sided adhesive tape, and leaf surface roughness was measured using a confocal laser microscope (VK-X100/X200, Keyence Corp., Osaka, Japan) with a 100Â objective lens. The arithmetic average roughness (R a ) was automatically generated using the measurement and analysis software (VK-X100/X200, Keyence Corp., Osaka, Japan), using the following formula: where N and Z n denote the number of data points in the measured area and the difference between the height of each point and the height of the reference plane (mm), respectively.

Measurement of density of stomata or leaf hair and trichome
On May 13, July 7, August 19, September 11-12, October 7, November 22, 2019, June 17, August 24, and October 12, 2020, the leaves were randomly sampled from the seedlings to determine the density of stomata, or leaf hair and trichome by the SEM. The SEM images (resolution of a micrograph, 1,280 Â 1,024 pixels; area, 780 Â 624 mm or 234 Â 187 mm; magnification, Â150 or Â500) of leaf pieces were used for calculation of the numbers of stomata and trichomes by eye, and their densities on the leaf surface were calculated based on one-sided leaf area. November 11, 2020, the leaves were randomly sampled from the seedlings to measure the contact angle of a water droplet on the leaf surface according to the method of Takamatsu et al. (2001) and Takahashi et al. (2023).

Measurement of the amount of epicuticular wax
The amount of epicuticular wax on the leaf surface was determined gravimetrically based on the method described by Sase et al. (1998) and Takahashi et al. (2023). After the BC particles in the chloroform extract were collected on a quartz fiber filter, the chloroform extract was placed in a petri dish and its tare weight was measured with a balance (CPA225D, Sartrius, Germany). Then, the petri dish was heated at 80 C on a hot plate (HTP452AA, Advantec MFS, Inc., Japan) until the extract was evaporated to dryness. The petri dish was then cooled to room temperature under darkness and again weighed with the balance. The amount of epicuticular wax on the leaf surface was determined as the difference in the weight of the petri dish after heating and the tare weight.
In view of the fact that the higher the hydrophobicity of the chemical component, the smaller the retention time (Figure 1), the hydrophobic factor (Hn) of each chemical component is defined as the reciprocal of the retention time (RT, min) of each chemical component as follows: In the present study, the hydrophobicity index of leaf epicuticular wax (Hw) in each tree species was evaluated using the following equation, based on the contents of five chemical components in the leaf epicuticular wax (C alkane , C ester and aldehyde , C fatty acid , C alcohol , and C phospholipid ; mg cm À2 ) and their Hn:

Statistical analyses
Statistical analyses were performed using Statistical Package for the Social Sciences software (SPSS Inc., Chicago, IL, USA). The measurements for leaf traits are presented as means of four individuals per tree species, and their standard deviation in the figures and tables. Analysis of variance (ANOVA) was applied to determine whether significant differences in the leaf surface traits occurred among the different tree species. To identify significant differences among the means of the nine tree species, Tukey's Honestly Significant Difference (HSD) test was performed as a posthoc test. Pearson's correlation analysis was conducted to test the relationship between the amount of BC particles deposited on the leaf surface, and leaf surface traits. A given effect was assumed to be significant at p < 0.05.

Results
Interspecific difference in the BC amount deposited on the leaf surface Table 1 shows the amount of BC particles deposited on the leaf surface of nine tree species from April 2019 to November 2020. There was a significant interspecific difference in the amount of BC particles deposited on the leaf surface among the nine tree species. The order of the maximum amount of BC particles per unit leaf area was as follows: I. rotunda (9.3 mg C m À2 in October 2020) > C. florida (7.3 mg C m À2 in December 2019) > O. fragrans (6.3 mg C m À2 in November 2019) > C. kousa (5.9 mg C m À2 in November 2019) > Q. glauca (5.1 mg C m À2 in October 2020) Լ Q. myrsinifolia (5.1 mg C m À2 in October 2020) > M. kobus (2.4 mg C m À2 in September 2020) Լ Z. serrata (2.4 mg C m À2 in August 2020) Լ S. japonicus (2.3 mg C m À2 in August 2019). The tendency of interspecific differences in the amount of BC particles deposited on the leaf surface among the tree species was different each month. The amount of BC particles on the leaf surface of C. kousa was significantly high in May 2019, while that of I. rotunda was significantly high in November 2019 as compared with those of the other tree species.
Interspecific difference in the leaf surface traits As an example, typical SEM images of the surface of the current-year leaves or 1st flush leaves of the nine tree species in October 2020 are shown in Photograph S.1. According to the classification of epicuticular wax structures into 23 types by Barthlott et al. (1998), epicuticular wax on the adaxial leaf surface of Q. glauca and Q. myrsinifolia, and M. kobus was in the form of film-like smooth layers that smoothly covered the leaf surface. Epicuticular wax with a thick layer called the crust was confirmed on the leaf surfaces of O. fragrans and I. rotunda. An epicuticular wax structure with a rod shape and a circular cross section called rodlets was observed on the leaf surfaces of Z. serrata, S. japonicus, C. kousa, and C. florida.
Leaf surface roughness values of the nine tree species from May 2019 to October 2020 are indicated in supplemental Table S.2. In Q. glauca and O. fragrans, leaf surface roughness in May was significantly higher than that in the other months in 2019. The leaf surface roughness of Z. serrata and M. kobus was significantly higher than that of the other tree species in most measurement months.
Stomatal density in the leaves of the nine tree species from May 2019 to October 2020 are shown in supplemental Table S.3. The stomatal densities of Q. glauca, Q.       Each value is the mean of four determinations, and the standard deviation is shown in parenthesis. In each month, the values followed by different letters indicate significant difference among the tree species at p < 0.05 (Tukey's honestly significant difference test). myrsinifolia, and O. fragrans were higher than those of the other tree species throughout the year. The density of leaf hair and trichome on the leaf surface of the nine tree species from May 2019 to October 2020 are indicated in supplemental Table S.4. The densities of leaf hair and trichome on the leaf surfaces of Q. glauca and C. florida were higher than those of the other tree species throughout the year. No trichomes were found in Q. myrsinifolia or I. rotunda.
The contact angle of water droplet on the leaf surface of the nine tree species from April 2019 to November 2020 are shown in supplemental Table S.5. Cornus florida showed the highest contact angle of water droplet on the leaf surface, followed by Q. myrsinifolia, Q. glauca, and C. kousa throughout the year.
Interspecific difference in the amount and composition of leaf epicuticular wax Table 2 shows the amount of epicuticular wax on the leaves of the nine tree species from April 2019 to November 2020. Throughout the year, the amounts of epicuticular wax in the four evergreen broad-leaved tree species and C. florida were greater than those in the other tree species. The amount of epicuticular wax on the leaves of Q. glauca, Q. myrsinifolia, O. fragrans, and I. rotunda tended to increase from April to August or September, decrease in January or February, and then increase again. Figure 2 indicates the chemical composition of the epicuticular wax on the leaves of the nine tree species from May 2019 to October 2020. Phospholipids accounted for over 70% of the epicuticular wax in O. fragrans, I. rotunda, S. japonicus, C. kousa, and C. florida. In I. rotunda and M. kobus, the proportion of alkanes in epicuticular wax was significantly high as compared with that in the other tree species (p < 0.05, Tukey's HSD test). Table 3 shows the correlation coefficient of the relationship between the amount of BC particles deposited on the leaf surface, and leaf surface roughness, stomatal density, leaf hair and trichome density, water droplet contact angle on the leaf surface, or the amount of epicuticular wax in the nine tree species. In September, October, and November 2019 and, August, September, and October 2020, a significant positive correlation was found between the amount of BC particles deposited on the leaf surface, and the amount of epicuticular wax. In August and September 2020, a significant positive correlation was observed between the amount of BC particles deposited on the leaf surface, and the contact angle of the water droplets on the leaf surface. In contrast, there were few significant correlations between the amount of BC particles deposited on the leaf surface, and leaf surface roughness, stomatal density, or leaf hair and trichome density in the nine tree species. Table 4 indicates the correlation coefficient between the contact angle of water droplet on the leaf surface, and other leaf surface traits of the nine tree species. A significant positive correlation was found between the contact angle of water droplet, and the density of leaf hair and trichome on the leaf surface, but a significant negative correlation was obtained between the contact angle of water droplet, and leaf surface roughness in the nine tree species. In addition, a significant positive correlation was found between the contact angle of water droplet on the leaf surface, and the amounts of epicuticular wax, esters and aldehydes, and fatty acids in the epicuticular wax of the nine tree species. Table 5 shows the correlation coefficient between the amount of BC particles deposited on the leaf surface, and the amount of five chemical components in the epicuticular wax in the nine tree species. A significant positive correlation was found between the amount of BC particles deposited on the leaf surface, and the amount of alkanes or phospholipids in the epicuticular wax of the nine tree species.

Relationship between the BC amount and leaf surface traits or leaf epicuticular wax
Supplemental Table S.6 indicates the hydrophobicity index of leaf epicuticular wax (Hw) of the nine tree species from May 2019 to October 2020. The Hw in O. fragrans and I. rotunda was significantly higher than that in the other tree species throughout the year. Table 6 shows the correlation coefficient between the amount of regular elemental carbon (EC) on the leaves and the hydrophobicity index of leaf epicuticular wax of the nine tree species from May 2019 to October 2020. A significant positive correlation was found between the amount of BC particles deposited on the leaf surface, and Hw across the nine tree species in September 2019 and all months of 2020.

Discussion
In the present study, an interspecific difference was found in the maximum amount of BC particles deposited from the atmosphere and firmly adhered to the epicuticular wax on the leaf surface among the nine tree species (Table 1). Leaf surface morphological traits, such as surface roughness, stomatal density, and leaf hair and trichome density, had little relationship with the amount of BC particles deposited on the leaf surface (Table 3). However, a significant positive correlation was observed between the amount of BC particles deposited on the leaf surface, and the amount of epicuticular waxes, or the contact angle of water droplet on the leaf surface in the nine tree species (Table 3). The micromorphology of the leaf surface and epicuticular wax has been reported to influence the deposition ability of PM on the leaves (Perini et al. 2017;Wang et al. 2019). It has been reported that PM is trapped in leaf wax (Dzierzanowski et al. 2011), and high deposition of PM on the leaves has been observed in tree species with a high amount of leaf cuticular wax (Saebø et al. 2012;Weerakkody et al. 2017). Furthermore, Rasanen et al. (2013) reported that leaf surfaces with relatively large water droplet contact angle can capture many particles. The contact angle of water droplet on the leaf surface is used as an index of leaf wettability; the Table 2. The amount of leaf epicuticular wax in nine tree species from April 2019 to November 2020.
Tree species Each value is the mean of four determinations, and the standard deviation is shown in parenthesis. In each month, the values followed by different letters indicate significant difference among the tree species at p < 0.05 (Tukey's honestly significant difference test).
lower the contact angle, the higher the wettability (Aryal and Neuner 2010). Rasanen et al. (2013) reported that leaf surface with low wettability increased the particle deposition on broad-leaved trees. In the present study, therefore, it is considered that more BC particles can be deposited on the leaf surface with high water repellency, and the interspecific difference in the amount of BC particles deposited on the leaves among the nine tree species is related to leaf surface traits, such as water repellency. Unlike present study, Li et al. (2021) reported significant negative relationship between the amount of PM and contact angle of water droplet on the leaf surface, indicating the importance of leaf surface wettability for PM retention. They measured the amount of PM as the sum of the masses of leaf surface PM and in-wax PM, which can be collected by washing with water and chloroform, respectively. The total amount of PM was largely accounted for by leaf surface PM, which is hydrophilic, in their report. In the present study, on the contrary, we excluded hydrophilic PM, and measured BC particles firmly adhered to the leaf epicuticular wax (i.e.,   hydrophobic). Therefore, such difference in the hydrophobicity of targeted PM could result in the inconsistent results between Li et al. (2021) and the present study. Similarly, several researchers reported significant positive relationship between the amount of PM and leaf surface roughness unlike our present study (Weerakkody et al. 2018;Shao et al. 2019;Li et al. 2021). In their reports, PM was not divided by its hydrophobicity. Therefore, the importance of leaf surface physical traits for PM retention could be different among the targeted PMs depending on their traits such as hydrophobicity.
A significant positive correlation was observed between the contact angle of water droplet on the leaf surface, and the amount of esters and aldehydes or fatty acids in the epicuticular wax (Table 4), although the amount and composition of epicuticular wax were quite different among the nine tree species (Table 2 and Figure 2). It has been pointed out that the chemical composition and structure of the leaf cuticle layer are related to the difference in the amount of particles deposited on the leaf surface among tree species (Jouraeva et al. 2002;Terzaghi et al. 2013). Previous studies have shown that leaf cuticular wax, aldehydes, and fatty acids are hydrophobic (Tang et al. 2017;Ziarati et al. 2019;Paraskar and Kulkarni 2021). Therefore, it is considered that leaf surfaces with high amounts of hydrophobic substances have high water repellency, and a high contact angle of water droplet on the leaf surface. Wang et al. (2015) showed that the capacity of leaves to accumulate PM depends on the adhesion force between the leaf surface and particles. Furthermore, the adhesion force is derived from the intermolecular forces between the leaf cuticular wax and the particle, which is determined by the chemical constitution of the molecules and the interfacial area . Therefore, the difference in the amount of BC particles firmly adhered to the epicuticular wax on the leaf surface among the nine tree species is considered to be related to the amount and chemical composition of epicuticular wax, which determines the hydrophobicity of the leaf surface.   To clarify the relationship between the amount of BC particles firmly adhered to the epicuticular wax, and the hydrophobicity of leaf surface in the nine tree species, in the present study, the hydrophobicity index of leaf epicuticular wax (Hw) in each tree species was evaluated based on the contents of five chemical components in the epicuticular wax, and their hydrophobicity (supplemental Table S.6). A significant positive correlation was found between the amount of BC particles deposited on the leaf surface, and Hw across the nine tree species especially in 2020 (Table 6). Therefore, the hydrophobicity index of leaf epicuticular wax determined by its chemical compositions is considered to be closely related to the capturing capacity of BC particles by the leaves of the nine tree species. These results support the hypothesis that, among various leaf traits, the hydrophobicity of leaf epicuticular wax is an important determinant of the amount of BC particles deposited on the leaf surface of urban greening tree species.

Conclusions
In the present study, there was a great difference in the amount of BC particles deposited from the atmosphere and firmly adhered to the epicuticular wax on the leaf surface among the nine Japanese typical urban greening tree species, and the order of the maximum amount of BC particles deposited on the leaf surface was as follows: I. rotunda > C. florida > O. fragrans > C. kousa > Q. glauca Լ Q. myrsinifolia > M. kobus Լ Z. serrata Լ S. japonicus. A significant positive correlation was found between the amount of BC particles deposited on the leaf surface, and the amount of epicuticular wax, the amount of alkanes or phospholipids in the epicuticular wax, or the hydrophobicity index of leaf epicuticular wax determined by its chemical compositions in the nine tree species. Therefore, tree species with relatively high hydrophobicity of leaf epicuticular wax are suitable as urban greening tree species for the purification of urban air polluted by BC particles.
Further studies are needed on the mechanisms of the relationship between the amount or chemical composition of leaf epicuticular wax, and the capturing capacity of BC particles by the leaves of many urban greening tree species.
To select greening tree species suitable for the purification of the urban atmosphere polluted by BC particles, we must clarify the interspecific and seasonal variations in the amount of BC particles on the leaves at different leaf positions in the crown of mature trees and their related factors not only at the leaf micromorphological level, but also at the whole-plant level.