Effect of PM_2.5 chemical constituents on atmospheric visibility impairment

ABSTRACT

PM_2.5 sampling was conducted at a curbside location in Delhi city for summer and winter seasons, to evaluate the effect of PM_2.5 and its chemical components on the visibility impairment. The PM_2.5 concentrations were observed to be higher than the National Ambient Air Quality Standards (NAAQS), indicating poor air quality. The chemical constituents of PM_2.5 (the water-soluble ionic species SO₄^2-, NO₃⁻, Cl⁻, and NH₄⁺, and carbonaceous species: organic carbon, elemental carbon) were analyzed to study their impact on visibility impairment by reconstructing the light extinction coefficient, b_ext. The visibility was found to be negatively correlated with PM_2.5 and its components. The reconstructed b_ext showed that organic matter was the largest contributor to b_ext in both the seasons which may be attributed to combustion sources. In summer season, it was followed by elemental carbon and ammonium sulfate; however, in winter, major contributions were from ammonium nitrate and elemental carbon. Higher elemental carbon in both seasons may be attributed to traffic sources, while lower concentrations of nitrate during summer, may be attributed to volatility because of higher atmospheric temperatures.

Implications: The chemical constituents of PM_2.5 that majorly effect the visibility impairment are organic matter and elemental carbon, both of which are products of combustion processes. Secondary formations that lead to ammonium sulfate and ammonium nitrate production also impair the visibility.

Introduction

Visibility impairment has become a highly prevalent phenomenon in developing countries including India, due to exponential growth and development in all sectors. It has resulted from an increase in atmospheric pollutant concentrations that frequently violates the National Ambient Air Quality Standards (NAAQS), particularly the ambient PM_2.5. Visibility impairment is primarily caused due to extinction of light by scattering and/or absorption by fine particles. Visibility impairment due to urban aerosol has been the subject of numerous air pollution studies around the world over the past several decades. Previous studies revealed that the size, chemical composition, and mass concentration of airborne particles substantially affect visibility (Kim et al. 2006; Wang et al. 2006; Yuan et al. 2006). Although the extinction of visible light from gaseous species can also impair visibility, such species have a much weaker influence (Sloane et al. 1991).

Delhi, the capital city of India, has a population of more than 14 million. The city is surrounded by other major growth centers of adjoining states such as Haryana and Uttar Pradesh, known as the National Capital Region (NCR) (Gulia et al. 2015). Delhi is reported to have 29 planned industrial areas, which lead to intercity goods distribution by diesel-fueled heavy duty vehicles (HDVs). Approximately 44,000 freight vehicles enter Delhi city every day, of which 40% are HDVs (United Nations Environment Programme [UNEP] 2015). All these sources cause high emissions of PM_2.5, which exceed the prescribed National Ambient Air Quality Standards (NAAQS) and lead to visibility degradation. Very limited literature is available regarding the effect of PM_2.5 composition on visibility degradation in New Delhi (Khanna et al. 2014; Khare and Khanna 2016; Tiwari et al. 2014; 2013).

The major objectives of this study were twofold: one, to assess the relationship between PM_2.5 and visual range; and two, to assess the causes of visibility impairment by relating light extinction coefficients (b_ext) to PM_2.5 chemical composition.

Materials and methods

Site description

The PM_2.5 sampling has been conducted at a curbside location on national highway NH-2 (Figure 1). The sampling site is in the proximity of one of the busiest intersections of Delhi, Ashram Chowk. The site has been selected based on the high traffic activity on NH-2 during the day and movement of HDVs for interstate goods transport during the night, as this junction acts as a connecting link to other two states, Uttar Pradesh and Haryana (Khanna, Khare, and Gargava 2015). Approximately 10% of the freight vehicles enter into Delhi city through the NH-2 entry point (United Nations Environment Programme [UNEP] 2015). The traffic composition varies during the day, with dominance of cars and two-wheelers during the morning and evening peak hours. The prominent wind directions in summer and winter are from the southwest and west, respectively. The average temperature was 12.94 ± 1.86ºC during winter and 33.36 ± 3.24ºC during summer sampling.

Figure 1. Sampling location.

Sampling details

PM_2.5 sampling has been conducted during December 2013–January 2014 and May–June 2014 for winter and summer seasons, respectively. Samples have been collected in a sampling cycle of 12–12 hours (i.e., morning sampling 9 a.m. to 9 p.m. and night 9 p.m. to 9 a.m.) every day in alternate weeks using a collocated PM_2.5 sampler (Ecotech Instrument AAS-127), which operates at a constant flow rate of 16.67 L/min. The samples have been collected on 47-mm polytetrafluoroethylene (PTFE) Teflon filters (Whatman PTFE filters, with pore size 2 µm and PP ring supported) and quartz filters (Pall Life Sciences, Tissuquartz 2500QAT-UP) for inorganic and carbonaceous analysis, respectively. The samples have been analyzed for metallic and ionic species and carbon fractions. Forty-one metals have been analysed using energy-dispersive x-ray fluorescence (ED-XRF; PANalytical Epsilon 5) from Si to Pb, at the Central Pollution Control Board (CPCB), New Delhi. The ionic species have been analyzed using ion chromatography (IC; Dionex ICS-1000) for cationic (sodium, ammonium, potassium) and anionic (sulfate, nitrate, chloride) species. The carbon fractions have been analyzed using the IMPROVE_A Thermal Optical/reflectance protocol (DRI model 2001 thermal/optical carbon analyzer) for four fractions of organic carbon (OC1, OC2, OC3, OC4; temperature: ambient to 580ºC) and three fractions of elemental carbon (EC1, EC2, EC3; temperature >580ºC, ≤840ºC) at the National Environmental Engineering Research Institute (NEERI), New Delhi. The pyrolyzed carbon has been corrected by subtracting from OC4 and adding to the EC1 component.

The daily visual range (VR) data have been collected from the Delhi International airport. The VR is measured by using a transmissometer, Drishti, developed by CSIR-National Aerospace Laboratories. It uses a novel light intensity modulation technique with synchronous demodulation detection, which provides hourly VR data. Daily optical light extinction coefficients (b_ext) are estimated from the well-known Koschmieder equation as shown in eq 1 (Koschmieder 1924):

(1)

Results and discussion

Chemical composition of PM_2.5

The daily mean PM_2.5 12-hr concentrations at the sampling site have been observed to be 60.5 ± 21.4 μg m⁻³ and 293.1 ± 36.7 μg m⁻³ in summer and winter seasons, respectively. These concentrations have been observed to be higher than the NAAQS of 60 μg m⁻³ as prescribed by the NAAQS (CPCB 2009). OC has attributed highest percentage of PM_2.5 in both summer (31%) and winter (33%), as shown in Figure 2. OC concentrations seem to be significantly higher in winter than summer, due to the presence of higher emissions from combustion sources during the former. This was followed by metals and EC, with a higher percentage of EC in summer than in winter. OC and EC have shown a high correlation in both seasons, indicating common sources such as vehicular emissions. However, their regression analysis indicates the presence of other sources of OC, as well such as combustion sources. During the winter season, biomass, wood, and waste combustion is often used for heating across the city under uncontrolled conditions in the open areas (Fu et al. 2010; Pant et al. 2015; Yadav, Tandon, and Attri 2013). Among the metal species, sulfur was the most abundant in both seasons, followed by crustal elements like Si, K, Ca, Fe, and Zn as shown in Figure 3. Metallic analysis shows higher emissions from industrial and vehicular emissions during summer, and from coal and biomass burning during winter. A high correlation has been obtained for S with As and Ni and moderate correlation with Pb and Br, which may indicate coal combustion, as Ni, V, K, Cd, Se, Cl, and As are the tracers for coal combustion and oil combustion (Yu et al. 2013). Pb along with Mn and Ba are emitted from fuel additives, and K from biomass burning, which has been higher during winter than during summer (Begum, Biswas, and Hopke 2007; Satsangi et al. 2013). Zn, Cu, and Ni are used as tracers for metallurgical industries; however, Zn and Cl are also emitted from incineration (Harrison et al. 1997; Moreno et al. 2013; Srimuruganandam and Nagendra 2012).

Figure 2. Chemical constituents of PM_2.5.

Figure 3. Metallic constituents of PM_2.5.

Among the ions, sulfate has been found to be the maximum in summer, followed by ammonium and nitrate; however, in winter, ammonium has been found to be the maximum, followed by nitrate and sulfate. This is in accordance with observations from previous studies with sulfate being higher in summer and nitrate in winter, as ammonium nitrate is volatile in nature and thus has been lost in summer due to volatilization (Chatterjee et al. 2010; Chow et al. 2005; Kim, Choi, and Ghim 2015). In summer, the concentration of ions has been higher in the day than at night as the ionic formation is favored by high temperature and strong solar radiation (Zhang et al. 2006). Ion equivalency has been estimated using anions (sulfate, nitrate, and chloride) and cations (ammonium) for both seasons to understand the neutralization of the ions in the atmosphere, and the anions have been found to be neutralized by ammonium in both seasons (Figure 4).

Figure 4. Ion balance between anions and cations.

Correlation of visual range with PM_2.5

Figure 5 shows the VR as a function of PM_2.5 mass concentrations for summer and winter seasons. In both the seasons, VR shows a significant correlation with PM_2.5, although they are negatively correlated. VR showed an exponential fit correlation with PM_2.5 concentrations; with sulfate and nitrate concentrations, it showed a power profile. VR showed a negative correlation with all of the PM_2.5 constituents, that is, sulfate, nitrate, ammonium, OC, and EC (Table 1).

Table 1. Correlation between visual range and PM_2.5 chemical constituents.

	VR	Sulfate	Nitrate	Ammonium	OC	EC
VR	1
Sulfate	−0.68	1
Nitrate	−0.60	0.82	1
Ammonium	−0.75	0.90	0.93	1
OC	−0.48	0.53	0.71	0.67	1
EC	−0.36	0.57	0.60	0.59	0.83	1

Figure 5. Visual range as a function of PM_2.5 mass concentrations.

Effect of chemical components on light extinction coefficients

The degradation of visibility occurs as a result of the scattering and absorption of light by particles and gases in the atmosphere. The light extinction (b_ext), which is the sum of light scattering by particles (b_sp), absorption by particles (b_ap), scattering by gas (b_sg), and absorption by gas (b_ag), is reconstructed according to the revised IMPROVE algorithm (Pitchford et al. 2007). The revised algorithm reduces bias for high and low light extinction extremes. It reduces the underprediction of high haze periods and the overprediction of low haze periods compared with the performance of the original algorithm (Pitchford et al. 2007). This revised algorithm has been applied by a large number of studies with concentrations comparable to New Delhi concentrations, particularly in China (Chen et al. 2016; Cheng et al. 2017; Fu et al. 2016; Jung et al. 2009; Wang et al. 2015). The algorithm is shown in eq 2:

(2)

where f_S(RH) and f_L(RH) are the water growth factors for small- and large-sized distribution of sulfate and nitrate, respectively, and f_SS(RH) is the water growth factor for sea salt. These water growth factors vary as a function of relative humidity (RH). The water growth factors have been used from Pitchford et al. (2007) based on the hourly relative humidity data. These have been averaged out for the 12-hr period. The constant numbers in the preceding equation are extinction efficiencies for each chemical species under dry conditions. Sulfate and nitrate refer to ammonium sulfate and ammonium nitrate complexes, which are estimated as 1.38 times the mass of sulfate and 1.29 times the mass of nitrate, assuming that sulfate and nitrate are fully neutralized by ammonium. Organic mass is estimated as 1.8 times the OC concentrations to account for unmeasured fractions. Small and large modes of sulfate, nitrate, and organics indicate formation through dry and aqueous mechanisms and are determined by the mass of each of these components, where the large mode is estimated by dividing the total concentration of the component by 20 μg m⁻³ and the small mode is estimated by subtracting the large mode concentrations from the total concentration. If the total concentration of a component exceeds 20 μg m⁻³, all of it is assumed to be in the large mode.

The estimated b_ext using the IMPROVE equation was 484.39 ± 28.8 Mm⁻¹ and 2875.8 ± 321.6 Mm⁻¹ in summer and winter seasons, respectively. The b_ext was significantly correlated with PM_2.5 concentrations (r² = 0.94, p < 0.001), demonstrating the strong influence of PM_2.5 particles on visibility degradation. The reconstructed light extinction coefficient was compared with that derived from the visual range in eq 1 using the Koschmieder equation. The estimated b_ext using the equation was 832 ± 56.1 Mm⁻¹ and 3139 ± 632.9 Mm⁻¹ in summer and winter seasons, respectively. There is significant difference in the values estimated using the two methods; however, the reconstructed chemical b_ext correlated strongly with the optical b_ext estimated from the Koschmieder equation (r² = 0.78, p < 0.001), as shown in Figure 6. Lowenthal and Kumar (2016) recommended using a factor of 2.1 instead of 1.8 for calculating organic matter (OM) from OC, which allows for accounting of water-soluble OM. Using this factor significantly increases the b_ext to 550.58 ± 30.4 Mm⁻¹ and 3110.04 ± 283.1 Mm⁻¹ in summer and winter, respectively reducing the variation from the estimated b_ext using Koschmeider equation (summer 832 ± 56.1 Mm⁻¹ and winter 3139 ± 632.9 Mm⁻¹).

Figure 6. Scatter plot of reconstructed b_ext versus measured b_ext from Koschmieder formula.

On average, OM is the largest contributor to the light extinction coefficient, accounting for 40% of chemical b_ext, as shown in Figure 7. The fractional contribution of OM to b_ext is the highest in summer (44%), followed by winter (37%). The next most important contributors to b_ext are elemental carbon (31%) and ammonium sulfate [(NH₄)₂SO₄], with minor contributions from ammonium nitrate (NH₄NO₃) and fine soil (~7%). However, in winter, major contributions are from ammonium nitrate (22%), ammonium sulfate (20%), and elemental carbon (13%). Higher elemental carbon in both seasons may be attributed to traffic sources, while lower concentrations of nitrate during summer may be attributed to volatility because of higher atmospheric temperatures. The values obtained in this study are in accordance with previous studies conducted in India (Ram et al. 2012; Singh et al. 2008; Tiwari et al. 2014).

Figure 7. Relative source contributions to chemical light extinction.

Conclusion

The PM_2.5 mass concentrations were observed to be threefold higher than the NAAQS for the winter season. PM_2.5 and its components were strongly correlated with the visual range, although the correlation was negative in nature. The reconstructed b_ext was 484.39 ± 28.8 Mm⁻¹ and 2875.8 ± 321.6 Mm⁻¹ in summer and winter seasons, respectively, which indicates poor air quality for visibility in Delhi city. Based on a method established by the IMPROVE program, organic matter (OM) is found to be the largest contributor to visibility impairment overall, accounting for 40% of b_ext, followed by ammonium sulfate and nitrate, elemental carbon, and fine dust. The particle scattering coefficient, b_sp, and particle absorption coefficient, b_ap, contributed the majority of the light extinction coefficient, and thus, particle emissions must be regulated for better air quality and visibility.

Acknowledgment

The authors thank CPCB, Delhi, and NEERI, Delhi Zonal office, for providing necessary laboratory support.

Additional information

Funding

The authors gratefully acknowledge UKIERI, India, for funding the project (UKIERI-TRP-2012/13-0089 IND/CONT/E/12-13/724). The participation of Isha Khanna in this study has been supported by the University Grant Commission (UGC), Government of India.

Notes on contributors

Isha Khanna

Isha Khanna is a research associate at Department of Civil Engineering, Indian Institute of Technology Delhi, Hauz Khas, New Delhi, India.

Mukesh Khare

Mukesh Khare is a professor at Department of Civil Engineering, Indian Institute of Technology Delhi, Hauz Khas, New Delhi, India.

Prashant Gargava

Prashant Gargava is an additional director and head of Air Quality Management Division at Central Pollution Control Board, Delhi, India.

Anwar Ali Khan

Anwar Ali Khan is an additional director at the Environment Department, Uttar Pradesh, India.