Characteristics and Aging of Traffic-Emitted Particles with Sulfate and Organic Compound Formation in Urban Air

Abstract

Traffic is a major source of anthropogenic aerosol in urban atmosphere. In this study, aerosol particles were measured with a TEM-EDX system at the roadside of a main road in the northwestern part of Beijing, China, under clear and hazy conditions. Soot, organic, sulfur-rich (S-rich), mineral, and metal particles, as well as the mixtures, were frequently encountered in aerosols. Under hazy conditions, S-rich particles coated with organic matter (S-OM particles) accounted for most of the total particles (15% to 24%), followed by soot particles (18% to 21%), organic particles (17% to 21%), non-mixed S-rich particles (10% to 18%), and S-rich particles with soot-, mineral-, or metal-inclusions (here referred to as S-inclusion particles) (11% to 15%). Under clear conditions, non-mixed S-rich and organic particles were dominant components, while mineral and soot particles were secondary components, among which, ~14% of the total particles had a sulfate core or OM coating; inclusions of mixture particles were often mixed with sulfate cores. In the sulfate core–OM shell structure particles, the ratio of core diameter to the whole particle diameter was ~0.52 under hazy conditions and ~0.60 under clear conditions, indicating a substantial sulfate and organic formation on the particles. Soot particles accounted for 18% to 21% of the total particles. The relative growth of aged soot particles was higher under hazy conditions than under clear conditions. In sum, particles from traffic emissions on a main urban road aged with the formation of sulfate and organic matter.

1. Introduction

Particulate matter (PM) encompasses a major group of atmospheric pollutants that lead to adverse health (e.g., affecting humans’ lungs and hearts) and environmental effects (e.g., causing visibility impairment, affecting solar radiation transport, and affecting the diversity of ecosystems) [1]. Vehicle emissions contribute a very large part of air pollutants in the urban atmosphere, such as PM_2.5 (particles with a mean aerodynamic diameter less than or equal to 2.5 μm), volatile organic compounds (VOCs), and nitrogen oxides (NO_x) [2,3,4]. For example, 5.6 to 37.6% of PM_2.5 in some large cities in China comes from traffic [5].

Traffic-related air pollutants include primary PM, originating from exhaust and non-exhaust emissions [6], and secondary PM, produced via chemical conversions, involving precursor gases, such as NO_x and VOCs emitted from vehicle engines [7]. PM directly emitted from vehicles mainly consists of primary organic aerosol (POA) and black carbon (BC) [8]. Engine loadings, fuels, driving conditions and technologies significantly affect PM emissions from vehicles [9]. Non-exhaust PM emissions associated with traffic include particles from wears of brakes, tires, and engines, and resuspended road dust [10]. Vehicle exhausts undergo coagulation, condensation, dilution, dispersion, and chemical reactions in air, and display different properties in the nearby roadway environment and the ambient environment [11,12,13]. Diverse characteristics of pollutants near roads have been investigated in various areas [14,15]. The pollutant levels and aerosol properties in near-road air vary drastically based on traffic types and movement, weather conditions, air stability, and chemical reactions [16,17]. In particular, the concentrations of traffic-emitted air pollutants, such as carbon monoxide (CO), nitrogen oxides (NO_x), black carbon (BC), ultrafine particles, and coarse particles, elevate at the roadside but quickly decrease downwind of the road within a few hundred meters [18]. All available data indicate that PM in road traffic emissions contributes an important part of anthropogenic PM in urban air and the PM is highly variable.

After being emitted into the air, fresh soot particles can adsorb gaseous species and water vapor, and enhance the production of secondary species, such as sulfate and nitrate, via catalyzing heterogeneous reactions on the particle surface in the urban atmosphere [19]. Aged soot particles usually have water-soluble coatings of secondary sulfate, nitrate, and organic matter [20]. For example, fresh soot particles near a major road were slightly coated with organic materials, whereas the aged soot particles were heavily coated with organics and sulfate [21]. In the polluted urban air, several factors are proposed to cause the rapid aging of soot particles, such as the concentration of gaseous pollutants from the vehicles, the strength of solar radiation, and relative humidity [22]. However, the complex formation mechanism of aged particles under hazy conditions is still poorly understood.

In this study, we measured the aerosol particles at the roadside of a main road under clear and hazy conditions during rush hours in Beijing, China. We utilized a transmission electron microscope equipped with an Oxford energy-dispersive X-ray spectrometer (TEM-EDX) and nanoscale secondary ion mass spectrometry (NanoSIMS) to identify the morphology, size, elemental composition, and mixing state of the collected particles.

2. Experimental Methods

2.1. Sample Collection

The observation was made when the air conditions turned from hazy to clear on 6–8 January 2017. Under hazy conditions, the atmospheric stability is high, resulting in the residence time of air pollutants possibly being longer than in clear conditions. The study main road was Xueyuan Road adjacent to the east boundary of China University of Mining and Technology Beijing campus, which is in the northwestern part of Beijing. The observation site (39°59′46″ N and 116°20′146″ E) was on the sidewalks of the Xueyuan Road (Figure 1). During the observation period, the average number of vehicles passing by the roadside site was about 1000 per hour, and gasoline- and diesel-powered vehicles accounted for 89% and 11%, respectively. A single-stage cascade impactor (KB-2, Qingdao Jinshida Company, Qingdao, China) was used to collect the particles onto the 300-mesh copper TEM grids, which were covered with a carbon-coated formvar film. The sampling flow rate was 1.0 L min⁻¹, and the cut-off diameter of the impactor for 50% collection efficiency was 250 nm if the density of the particles was 2 g cm⁻³. For each sample, the collection time was 150–900 s depending on the air pollution. Meteorological conditions were recorded by a portable meteorological device (NK4500, Kestrel, Guangzhou, China), including temperature (T), relative humidity (RH), and pressure (P). The details of sampling periods, meteorological conditions, and the concentrations of PM_2.5 are illustrated in

Table 1. Time and weather conditions of sample collection collected.

ID	Sampling Time	T (°C)	RH (%)	P (hPa)	Wind (m s−1)	PM2.5 (µg·m−3)	Weather Condition
N1	6 January 2017 9:00	3.3	64.0	102.0	1.6	173	hazy
N2	6 January 2017 18:00	3.7	61.9	101.6	1.8	237	hazy
N3	7 January 2017 9:00	3.7	60.0	101.6	1.6	182	hazy
N4	7 January 2017 18:00	3.2	83.0	101.4	1.2	103	hazy
N5	8 January 2017 9:00	2.9	39.9	101.5	5.8	25	sunny
N6	8 January 2017 18:00	4.2	34.2	101.3	5.4	19	sunny

N: Roadside site.

.

2.2. Microscopic Analysis

Particles were examined with a Tecnai G2 F30 field emission high-resolution transmission electron microscope (FE-HRTEM) equipped with an Oxford EDX. The system was operated at 300 kV. The EDX can detect elements with atom numbers larger than 5 (B) in a single particle. EDX spectra were first collected for 20 live seconds to minimize the influence of radiation exposure and potential beam damage and then collected for another 90 live seconds for a whole range of possible elements. Copper was excluded from the analysis because of interference of copper in the TEM grids. Individual particles are unevenly distributed on TEM grids, with coarser particles in the center of sampling spot and finer particles on the periphery. Therefore, to guarantee that the analyzed particles are representative, more than four random areas are selected from the sampling center to periphery on each TEM grid. More than 100 particles per sample are guaranteed to be analyzed.

Some particles were selected for the nanoscale secondary ion mass spectrometer (NanoSIMS, 50 L, CAMECA Instruments, Gennevilliers, France) analysis. In this study, signal intensity mappings of ¹²C⁻, ¹⁶O⁻, ¹²C¹⁴N⁻ and ³²S⁻ ions were obtained after the Cs⁺ primary ion beam ionized the atoms of the particle surface. Ion signal intensity mapping of individual particles with nanometer resolution clearly shows the ion distribution in particles. For example, ¹²C¹⁴N⁻ and ¹²C⁻ signals imply the existence of organic components in the particles [23].

The TEM images were digitized in the Microscopic Particle Size of Digital Image Analysis System (an automated fringe image processing system, UK) to project the surface areas of individual particles. For single particles, the equivalent spherical diameter was calculated as $\sqrt{4A/\pi }$ (where A is the projected area).

The fractal dimension (D_f) of soot particles was used to indicate the aging degree. It describes the morphology and compactness of soot particles and was calculated by the scaling law according to [24,25] (Equation (1)).

$$N=k_g{\left(\frac{2R_g}{d_p}\right)}^{D_f}$$

(1)

where N is the total number of monomers in each soot aggregate; R_g is the radius of gyration of the soot aggregate; k_g is the fractal factor, and d_p is the diameter of monomer, which was directly obtained from the TEM images. N was scaled with the aggregate projected area in the power law relationship in Equation (2). In this study, D_f and k_g are estimated from a power law fit of a scatter plot of N versus the values of 2R_g/d_p.

$$N=k_a{\left(\frac{A_a}{A_p}\right)}^{\alpha }$$

(2)

where A_a is the projected area of the soot aggregate; A_p is the mean projected area of the monomers; k_a is a constant, and α is an empirical projected area exponent. The exact values of k_a and α depend on the overlap parameter δ, which was calculated with Equation (3).

$$\delta =\frac{2\alpha }{L}$$

(3)

where α is the monomer radius and L is the lattice spacing of the monomer in TEM images. The actual radius of gyration (R_g) was calculated with Equation (4).

$$\frac{L_{max}}{2R_g}=1.50\pm 0.05$$

(4)

where L_max is the maximum length of the soot aggregate.

3. Results

3.1. Major Types of Traffic-Derived Individual Particles

Six samples were collected at the roadside site under clear and hazy conditions. A total of 829 particles were analyzed with the TEM-EDX. In each sample, about 105–205 particles were taken to measure the morphology, size, and element composition. According to the morphologies and chemical compositions, the traffic-derived particles were divided into six types: soot, organic, sulfur-rich (S-rich), mineral, metal, and the mixture particles [26,27].

Soot particles mostly had chain-like (Figure 2a) or densely compacted (Figure 2b) shapes, with main elements of C, with minor O and Si. At the roadside site, most of the soot particles (78%) were chain-like shapes under clear conditions, while 52% of the soot particles were densely compacted shapes (thickly coated with other compounds) under hazy conditions. The proportion of chain-like soot particles is higher in fresh emissions than in the road environment or open air [28]. Organic particles in this study are primary organic particles (POA). TEM observations showed that the POAs were spherical (Figure 2c,d), while the secondary organic particles (SOA) exhibited domelike or core-shell structures (Figure 2e) and are often mixed with other particles. S-rich particles were mainly composed of O, S, and N (Figure 2f); they were produced as a result of reactions that involve gaseous SO₂, NO_x, and NH₃ on the particle surface [23]. Mineral particles were irregularly shaped (Figure 2g–i). Based on the element compositions, mineral particles were classified into Ca-rich, Al-rich, and Mg-rich particles. The Ca-rich particles (Figure 2h) were frequently spherical, with major element components of Ca, O, and P. Such particles could be from vehicles under the driving state of acceleration, steady speed, and engine braking [29]. Metal particles were mainly composed of O, C, and metallic elements (e.g., Fe, Zn, Mn, and Ca). The three major types of metal particles were Zn-rich (Figure 2j), Fe-rich (Figure 2k), and Mn-rich (Figure 2l) particles. The metal particles were nearly spherical, indicating that they likely came from vehicle emissions or coal-fired power generation, via high-temperature combustion followed by fast cooling [30].

Mixture particles were mainly mixed with S-rich, soot, organic, mineral, or metal particles, with a nearly spherical shape or a core-shell structure (Figure 3). The majority of mixtures were S-OM particles and S-inclusion particles. S-OM particles were S-rich ones with secondary organic components and had core-shell structures with a sulfate core and OM coating. The ion maps of particles at the roadside site from the NanoSIMS showed ¹²C¹⁴N⁻ signals in the coating and ³²S²⁻ signals in the core (Figure 4). This underlines that the coating of particles was by an organic component, and the core of the particles was made up of a sulfate component. S-inclusion particles were S-rich, with inclusions of soot, minerals, or metals (Figure 3).

3.2. Relative Abundance

At the roadside site, PM_2.5 decreased from 237 to 19 µg m⁻³ when the air turned from hazy on 6–7 January (PM_2.5 ≥ 150 µg m⁻³ and visibility < 10 km) to clear on 8 January (PM_2.5 < 50 µg m⁻³ and visibility > 10 km). Under hazy conditions, the dominant particles were S-OM particles and soot particles, which accounted for 15–24% and 18–21% of the total particles, respectively; organic particles, non-mixed S-rich particles, and S-inclusion particles accounted for 17–21%, 10–18%, and 11–15% of the total particles, respectively (Figure 5). Under clear conditions, the dominant particles were non-mixed S-rich particles, organic particles, mineral particles, and soot particles, which accounted for 22–23%, 20–21%, 18–21%, and 19–20% of the total particles, respectively. In sum, soot particles, organic particles, and non-mixed S-rich particles were major aerosols at the roadside site. It has also been reported that OC, EC, and SO₄²⁻ were more abundant in the fine fraction in the roadside environment [31]. The amount of mixture particles, including S-OM and S-inclusion particles, was greater under hazy conditions than under clear conditions. This indicates that the formation of secondary aerosol was more active under hazy conditions. Approximately 14% of the particles had a sulfate core or an OM coating in all the samples. Compared with the clear conditions, there were more particles with OM coatings under hazy conditions and the thickness of the OM coating was greater.

4. Discussion

4.1. Mixture of Organic and Sulfate

The mixing state of aerosols is a key factor affecting the aerosol–radiation and aerosol–cloud interactions [32]. We found two types of mixture particles at the study site: S-OM particles and S-inclusion particles. Approximately 14% of the particles had a sulfate core or an OM coating in all the samples. Organic compounds and sulfuric acid were frequently involved in particle formation and growth, leading to their appearance in aerosol particles [33]. We also found that more particles had OM coatings under hazy conditions than under clear conditions, and furthermore, the OM coatings were thicker under hazy conditions than under clear conditions. The coating on particles is usually caused by aging processes, such as coagulation, condensation, and heterogeneous chemical reactions [34]. The core-to-shell ratio (R) of a particle, which is the ratio of the diameter of the core to the diameter of the whole particle, was used to assess the aging degree of particles [28]. Small R means a relatively thick coating on particles, thus, reflecting a higher degree of aging [35]. The R of the S-OM particles was smaller under hazy conditions (ranging from 0.38 to 0.62, with an average of 0.52) than under clear conditions (ranging from 0.53 to 0.67, with an average of 0.60) (Figure 6). The density distribution also showed that the R of the S-OM particles was smaller under hazy conditions than that under clear conditions, indicating more sulfate and organic formation on the particle surfaces under hazy conditions.

The aging of aerosol particles could be impacted by several factors, such as the concentration of gaseous pollutants, strength of solar radiation, relative humidity (RH), the atmospheric oxidizing capacity, and different formation mechanism [22]. Because of the complex formation mechanism and processes, organic aerosols might display a mixing state different from that of secondary inorganic aerosols. At the study sites, the soot/organic/mineral/metal inclusions in the mixture particles were frequently mixed with sulfate cores instead of OM coatings (Figure 3 and Figure 4), indicating that the inclusions likely had acted as nuclei for secondary sulfate or organic uptake in the air. These changes are particularly important for the particles because they are key factors influencing the physical and chemical characteristics of the particles, such as hygroscopicity, cloud condensation nuclei potential, and the ability of solar radiation absorption and scatterings [36].

4.2. Aging of Soot Particles

Soot particles accounted for a significant fraction (18–21%) of the total particles at this roadside site. Under clear conditions, most soot particles (78%) near the roadside site were fresh, consisting of small spherical primary particles combined with branched aggregates (Figure 2a). Under hazy conditions, about 52% of the soot particles were thickly coated with other compounds at the roadside site (Figure 2b). It is widely acknowledged that the D_f of soot particles reflects their sources and aging processes [24]. Compact soot particles often have D_f values larger than those of branched aggregates. Here, we calculated that the mean D_f for soot particles was smaller under clear conditions (with an average of 1.75) than under hazy conditions (with an average of 2.08) (Figure 7). Therefore, our result imply that the aging degree of soot particles on the studied main road is higher in the polluted urban air than under clear conditions. The wide range of D_f in soot particles was expected because they originate from different sources, combustion conditions and have undergone different atmospheric aging processes. Previous studies showed that the D_f of soot particles of the primary sources, such as the D_f from biomass burning in the range of 1.74−1.85 [25,37], the D_f from light-duty passenger vehicles in the range of 1.70−1.78 [37], and the D_f from diesel vehicles in the range of 1.80−2.0 [38] (

Table 2. Mean values of fractal dimension of soot particles from different sources and in the ambient environment.

Fractal Dimension	Probable Dominant Source	Literatures
1.75–2.08	roadside environment	This study
1.70–1.78	light-duty passenger vehicle	Chakrabarty et al. (2006) [37]
1.74	biomass burning	Chakrabarty et al. (2006) [37]
1.85	biomass burning	China et al. (2013) [25]
2.4	Asian dust soot	Adachi et al. (2007) [24]
1.8–2.0	diesel vehicle	China et al. (2014) [41]
2.00	ambient environment	Wang et al. (2017) [40]
1.80	Tunnel environment	Wang et al. (2017) [40]
1.88	ambient environment (Background)	Yuan et al. (2019) [42]
1.89–1.99	diesel vehicle	Guarieiro et al. (2017) [38]

), become larger when soot is coated by other components during atmospheric processes. Especially, at a higher humidity, the coating components liquefy, and are, therefore, more likely to cause soot particle reformation [39]. Tunnel studies showed that D_f values for soot particles in a tunnel are around 1.80 [40]. They became more compact as evidenced by the larger D_f (>2) in background air and in the ambient air of an Asian dust episode [24]. By comparing with the fresh soot particles emitted by various sources (Table 2), we found that particles in the roadside environment showed a slightly higher D_f and particles in the ambient air showed an even higher D_f. This indicates that a small amount of the fresh soot particles aged in the road environment, but more fresh soot particles aged in open air, producing particles in more spherical and compacted states.

5. Conclusions

We analyzed particles at the roadside of an urban main road and identified soot, organic, S-rich, mineral, and metal particles, as well as the mixtures, among which, ~14% of particles had a sulfate core or an OM coating. The average core-to-shell ratio (R) of the S-OM particles was 0.52 under hazy conditions and 0.60 under clear conditions, indicating a substantial sulfate and organic formation among the particles. Soot particles accounted for 18% to 21% of the total particles. The mean D_f for soot particles was smaller under clear conditions (with an average of 1.75) than under hazy conditions (with an average of 2.08). By comparing the measurement data under clear and hazy conditions, we concluded that the aging degree of particles was higher under hazy conditions than under clear conditions. Our experimental results could be used to guide future measurements of targeted particles from traffic emissions in the local area, modeling their roles in urban environment activities.