The diurnal variability of sulfate and nitrate aerosols during wintertime in the Indo-Gangetic Plain: implications for heterogeneous phase chemistry

Abstract

We have conducted this study (November 09–February 10) during the daytime (average PM₁: 113 μg m⁻³; n = 51) and nighttime (average PM₁: 159 μg m⁻³; n = 49) in the Indo-Gangetic Plain (IGP). Air-mass back trajectories suggest the impact of local emission and long-range transport (predominantly from the north–west direction). Mass fractions of SO₄²⁻ and NO₃⁻ in PM₁ are significantly (p < 0.05) different during both the daytime and nighttime, whereas NH₄⁺/PM₁ were similar during the day and night. The relatively high concentration of SO₄²⁻ observed during the daytime was explained based on heterogeneous-phase reactivity due to the positive response of Fe and Mn species (inferred from correlation and multi-linear regression analysis: MLRA). Likewise, the lower concentration of NO₃⁻ was explained based on the negative response of Fe during heterogeneous phase formation. The role of wet-bulb temperature and solar flux has also been studied. From the field-based measurements our study shows that heterogeneous formation of SO₄²⁻ (involving Fe and Mn) and NO₃⁻ occur via selective endothermic pathways. The proposed mechanism for sulfate and nitrate formation via heterogeneous phase reactivity is in good agreement with the field-based measurements obtained during this study (IGP). The impact of the heterogeneous-phase reactivity via endothermic pathways relates to the uptake of various reactive species during the winter in IGP. This, in turn, has implications for fog-formation and tropospheric oxidative cleansing. Furthermore, the uptake of various species will lead to altered size, morphology and optical properties of aerosols. This would have impact on regional scale radiative forcing estimates and future climate projections.

---

1 Introduction

It has been reported earlier that during high pollution episodes the heterogeneous formation of SO₄²⁻ and NO₃⁻ species is the kinetically more favored pathway (higher rate of reactions).¹ Furthermore, the impact of heterogeneous phase chemistry on the uptake of various chemical species with varying mass concentrations of fine particles has been discussed in detail previously.^2–4 Over the last two decades, we have seen an unprecedented increase in levels of air-pollutants attributable to urbanization and the development in rural areas. As a matter of fact, one of the major research focuses of atmospheric studies is to assess the diurnal, temporal and spatial variability of ambient aerosols owing to their impact on ambient air-quality, climate change, hydrological cycle and health.^5–7 The Indo-Gangetic Plain (IGP), the most populated region in India, experiences a massive loading of organic and inorganic aerosols during November–February every year.⁸ This is attributable to fossil-fuel combustion, large-scale biomass and bio-fuel burning emissions and shallower boundary layer height during winters.^9–12 Reviewing the last decades chemical composition data over IGP suggests that secondary aerosols (organic and inorganic species) can contribute ca. 40% of the fine PM loading.^8,13,14 As far as fine-mode aerosols are concerned, most studies have reported only the PM_2.5 composition.^8,13–15 Recently, our group has started assessing the PM₁ composition with the help of indigenous PM₁ impactor samplers.^16–19

High loading of fine-mode aerosols and significant contributions from secondary species during the winter in IGP results in thick haze and fog formation.⁸ In this context, a previous study suggested that the uptake of various chemical species in the aerosol phase can lead to the formation of photochemical smog and tropospheric oxidative cleansing.² This in turn results in poor air-quality and presents a health hazard risk on a regional scale. Previous studies have also addressed the impact of air-masses long-range transport from IGP over the Indian Ocean affecting its biogeochemistry.^20,21 Source-apportionment and the impact of local air-masses and long-range transport during the study period has been reported recently.²² This study documents our attempts to assess the day–night variability of secondary aerosol species, particularly SO₄²⁻ and NO₃⁻, in PM₁ during wintertime. Our study documents the variability and provides the mechanism for the formation of SO₄²⁻ and NO₃⁻ during the day and night. Also, the role of heterogeneous phase chemistry in the formation of secondary species via selective endothermic pathways is evident from field-based measurements used in this study.

2 Materials and method

2.1. Sampling and meteorology

PM₁ sampling was carried out on the roof-top of a 12 m tall building (Western Lab Extension) in the premises of the Indian Institute of Technology Kanpur (26.30° N, 80.13° E, 142 m above mean sea level). Aerosol samples have been collected onto Teflon filters (47 mm diameter) using an indigenous PM₁ sampler (calibrated flow rate: 10 LPM).^18,19 Sampling was preferred for ∼10 h to collect sufficient aerosol mass for chemical analysis. Overall, we have collected n = 51 samples during daytime and n = 49 samples during nighttime. During sampling, the winds were predominantly north-westerly with wind-speeds centering on 0.6–0.8 m s⁻¹ during October–December (average wind speed 2.2 m s⁻¹), whereas average wind-speed is relatively high in the months of January (avg.: 2.5 m s⁻¹) and February (avg.: 2.8 m s⁻¹). The ambient temperature (T) varied between 20.1–29.8 °C and the RH between 18–79% during the daytime, whereas T ranges from 9.5–23.4 °C and RH from 29–95% during the nighttime in this study. The PM₁ mass concentration has been ascertained based on gravimetric determination on an analytical micro balance (Mettler Toledo) after equilibrating the filters (pre- and post-sampling) at 25 °C temperature and ∼35% RH.

2.2. Chemical analysis

A portion of each sampled filter was extracted with 30 mL (10 mL × 3; 5 min each) of Milli-Q water (resistivity: 18.2 MΩ cm). Subsequently, the particles were allowed to settle under gravity and then the aqueous extracts were gently transferred into pre-cleaned vials for quantification of the anthropogenic ionic species (NH₄⁺, SO₄²⁻ and NO₃⁻) on an Ion-Chromatograph (Metrohm, Compact IC 761). A mixture of Na₂CO₃ and NaHCO₃ has been used as an eluent for anion separation, whereas dipolinic acid was used as an eluent for cation separation. Furthermore, a separate portion of the sample filter was subjected to hot-plate digestion (at ∼180 °C) using 20 mL of conc. HNO₃ (15.2 N, Seastar, 67–70% GR grade, Suprapure, Fisher, Leicestershire, UK). Several metals including major elements (Na, Mg, K, Ca and Fe) and trace metals (V, Cr, Mn, Ni, Co, Cu, Zn, As, Se, Cd and Pb) have been quantified using Inductively Coupled Plasma-Optical Emission Spectrometry (ICP-OES; iCAP 6300, Thermo Scientific).^17,22 Quality control and assessment pertaining to the data has been ascertained routinely by analyzing the procedural blanks and replicate analysis of the samples. We report herein the blank corrected concentrations of all the chemical species. Meteorological data (RH, T and wind speed) has been adopted from National Center for Environmental Prediction (NCEP) reanalysis information (National Oceanic & Atmospheric Administration; 27.5° N; 80° E).²³ Solar flux data for the site was retrieved from AERONET (AErosol RObotic NETwork).

3 Results and discussion

3.1. Air-mass back trajectories (AMBTs)

Hybrid Single Particle Lagrangian Integrated Trajectory (HYSPLIT) model analysis was carried out to assess the impact of AMBTs (local/short and long-range) at the receptor site during the study period.²⁴ A 5 d AMBT at 1000 m above mean sea level (amsl) during the sampling is shown in Fig. 1. Basically all the AMBTs have been categorized into four bins depending on their origin. For example, air-masses of local origin are separated out with those associated to long-range transport (Fig. 1a and b). Furthermore, the air-masses of long-range transport have been classified into three categories: westerly, south-westerly and north-westerly. It is important to mention here that number of long-range AMBTs pointing their origin from the westerly (n = 3; both day and nighttime) and south-westerly (n = 5; both day and nighttime) directions is quite small. These data points (n = 8, associated with westerly and south-westerly air-masses) have been omitted for further discussion from both the daytime (total n = 51) and nighttime samples (total n = 49). Moreover, after discussing the diurnal variability of the PM₁ mass concentration, the data points associated with north-westerly AMBTs have been considered to represent long-range transport. This data set (NW-long-range AMBTs) was compared with those associated to local emissions systematically through the entire manuscript.

Fig. 1 HYSPLIT air-mass (5 d) back trajectories (1000 m above mean sea level) during sampling at Kanpur (shown by an open circle) depicting the impact of: (a) local air-masses and (b) long-range transport predominantly from the NW-direction.

3.2. Diurnal variability of fine-mode aerosols in the Indo-Gangetic Plain

Assessment of the ambient atmospheric concentrations of PM₁ has recently gained more importance due to their impact on atmospheric chemistry and health hazard risk.^25,26 Real-time measurements of PM₁ on an Aerosol Mass Spectrometer (AMS) have really assisted in understanding many in situ chemical transformations in the ambient atmosphere. For example, the formation of organosulfates under acidic conditions is actually one of the most important discoveries of this decade.²⁷ Toxicological research due to submicron aerosols (PM₁) has also gained more importance for quantifying the impact of an individual chemical constituent as a target dose.^5,28 In conjunction with the HYSPLIT AMBTs (discussed above), we have shown the diurnal variability of PM₁ (Fig. 2a). The maximum concentration of PM₁ during the daytime (348 μg m⁻³) and nighttime (388 μg m⁻³) was associated with local emissions. However, the minimum concentration of PM₁ [daytime: 18 μg m⁻³ and nighttime: 35 μg m⁻³] was linked with the north-westerly long-range AMBT. Thus, the large variability of PM₁ over central IGP was attributed to the impact of local emissions and long-range transport. The overall variability in PM₁ mass concentration has been recorded from 18–348 (avg. ± SD: 113 ± 72) μg m⁻³ during the daytime in this study (Fig. 2a). However, the PM₁ concentrations varied from 35–388 (159 ± 89) μg m⁻³ during the nighttime (Fig. 2a). Thus, the PM₁ mass concentration during the nighttime was relatively high over the study location at Kanpur. This was also reflected in the overall day/night PM₁ ratio and its variability: centered at 0.73 ± 0.19 (range: 0.33–0.99; Fig. 2b). The relatively high aerosol loading during the nighttime was mainly attributed to the shallower boundary layer height and low temperature. It is important to highlight here that a couple of data points exhibit a day/night PM₁ ratio ≈1 (Fig. 2b). A previous study (November 2011–March 2012), assessing the day–night variability of chemical constituents in PM_2.5 from an upwind location in IGP (Patiala), also reported a high variability during the day (30–330 μg m⁻³) and nighttime (48–400 μg m⁻³) with a day/night PM_2.5 ratio ≈0.70.¹⁵ They also observed in their study that on several occasions the nighttime PM₁ overlaps with that observed in the daytime. Thus, our observation of the day–night variability of PM₁ (from central IGP) shows quite similar features as that reported in a previous study of PM_2.5 (from the Patiala site upwind of the IGP). This is mainly attributed to the similar meteorological conditions and boundary layer dynamics over central and upwind IGP.

Fig. 2 (a) The diurnal variability of PM₁ mass concentration and (b) PM₁ mass ratio during the day with respect to night at Kanpur (IGP). The different color bars represent the origin of air-masses (as shown in Fig. 1).

In a previous study,²² using positive matrix factorization (PMF) we have shown quantitatively the contribution of six potential sources (coal combustion, leather tanneries, industrial and construction activities, diesel emission, secondary aerosols and biomass burning emission) of both local emissions and long-range transport impacting the site during study period. One of the important features of that study related to the observation that long-range NW-AMBTs were associated with a 10% higher contribution of biomass burning emissions when compared to the local sources. Thus, having noticed some differences in the source contributions of aerosols from local emissions and those associated with long-range transport (NW-direction), we have sorted our data set into two categories of local air-masses and long-range transport (only considered NW-air-masses owing to sufficiently large number of data points during day and nighttime).

3.3. Mass fractions of the secondary aerosol species in the IGP

We have shown the percentage contribution of secondary inorganic species in PM₁ during the daytime and nighttime for both the cases of local air-mass and long-range transport (north-westerly; Fig. 3a and b). Accordingly, the daytime and nighttime PM₁ concentration in local air-masses averages at 136 ± 86 μg m⁻³ and 193 ± 109 μg m⁻³, respectively. However, the average PM₁ mass concentration was relatively low in the air-masses indicating their origin from long-range transport (upwind, NW-direction): daytime PM₁ averages at 93 ± 59 μg m⁻³ whereas that in the nighttime it is 138 ± 66 μg m⁻³. In local air-masses during the daytime and nighttime the anthropogenic ionic species (∑ANTH = NH₄⁺ + SO₄²⁻ + NO₃⁻) average concentration was 38 and 45 μg m⁻³, respectively (Fig. 3a). The percentage contributions of NH₄⁺, SO₄²⁻ and NO₃⁻ in ∑ANTH are 36%, 52% and 12%, respectively. In the nighttime, the contribution of NH₄⁺ looks similar to that at daytime. However, SO₄²⁻ (41%) and NO₃⁻ (23%) were distinctly different in nighttime when compared to those in the daytime (Fig. 3a and b).

Fig. 3 The water-soluble ionic composition of NH₄⁺, SO₄²⁻ and NO₃⁻ in PM₁ during daytime and nighttime under the influence of: (a) local sources and (b) long-range transport from the upwind (north–west) direction. The mass concentrations of ∑ANTH (sum of ANTH: NH₄⁺ + SO₄²⁻ + NO₃⁻) are given in μg m⁻³.

Likewise, air-masses of distant origin show a similar mass fraction of NH₄⁺ in ∑ANTH during the day versus nighttime. The contribution of SO₄²⁻ and NO₃⁻ to ∑ANTH (day: 23 μg m⁻³ and nighttime: 27 μg m⁻³) are different in the day and nighttime. Summing up, the relatively high mass concentration of PM₁ and associated ∑ANTH during the nighttime was attributable to the shallower boundary layer height. In the daytime samples (Fig. 3a and b), the percentage contribution of NH₄⁺, SO₄²⁻ and NO₃⁻ look identical between local versus long-range transport. The nighttime samples were similar. In other words, we have observed conspicuous day–night variability in the abundance of secondary aerosol species over central IGP. However, irrespective of the influence from local air-masses and long-range transport, the daytime composition of ANTH looks similar. This observation is also true for the nighttime samples. Thus, it is logical to group all the data set (local emissions and long-range transport) into two bins of daytime and nighttime (Table 1). The impact of excess ammonium on the formation of nitrate was shown through the scatter plot of NH₄⁺/SO₄²⁻ equivalent ratio and NO₃⁻/SO₄²⁻ mass ratio (Fig. S2†).

Table 1 Mass fraction (range with average ± 1σ) of anthropogenic ionic species during the daytime versus nighttime over central IGP (Kanpur) with statistical analysis (two-tailed t-test)^a

Parameter	Daytime (n = 43)	Nighttime (n = 41)	t-Test	Difference	p-Value
a Daytime PM1: 114 ± 76 μg m^−3 and nighttime PM1: 165 ± 93 μg m^−3.
NH4^+/PM1	0.01–0.29 (0.10 ± 0.07)	0.004–0.23 (0.08 ± 0.05)	1.5	Not significant	p = 0.1374
SO4^2−/PM1	0.04–0.24 (0.11 ± 0.05)	0.01–0.16 (0.07 ± 0.04)	4.0	Significant	p = 0.0001
NO3^−/PM1	0.004–0.12 (0.03 ± 0.03)	0.002–0.11 (0.05 ± 0.03)	3.0	Significant	p = 0.0030

---

We have assessed the mass fractions of NH₄⁺, SO₄²⁻ and NO₃⁻ in PM₁ during daytime versus nighttime (Table 1). The difference in the mass fraction of these secondary species in conjunction with a statistical two-tailed t-test was studied to infer about their predominant formation during daytime versus nighttime (Table 1). Accordingly, no significant difference (t = 1.5; p > 0.05) in the mass fraction of NH₄⁺ was observed between day (avg.: 10%) versus nighttime (8%), respectively. The average mass fraction of SO₄²⁻ was 11% and 7% during the day and nighttime, respectively. These mass fractions are significantly different (t = 4.0; p < 0.05). Furthermore, the average mass fraction of NO₃⁻ during the day (3%) and nighttime (5%) was also significantly different (t = 3.0; p < 0.05). For the sake of comparison within IGP, we have adopted the data set from a study conducted in upwind IGP.¹⁵ Their study also reported a nearly equal NH₄⁺/PM_2.5 ratio of ∼0.03 during the day and nighttime. They found the SO₄²⁻/PM_2.5 ratio was relatively high during the day (0.09) when compared to that found during the night (0.07). In a sharp contrast, the NO₃⁻/PM_2.5 ratio of their study was relatively high during nighttime (0.09) when compared to that found during the day (0.03). Another study from southern India (Chennai) reported a similar observation for the day–night variability of secondary aerosol species; the SO₄²⁻ during the daytime was ∼23% higher, whereas NO₃⁻ during the night was higher by 66% and the abundance of NH₄⁺ was almost the same during the day and night.²⁹ Thus, the day–night abundance pattern of secondary inorganic species reported herein is quite similar to that reported previously from other geographical locations in India.^15,29

Furthermore, a quite similar result for the day–night abundance pattern of the mass fraction of secondary aerosols has been reported recently in PM₁₀ from a rural site at San Pietro Capofiume (Po Valley; Table 2).³⁰ However, from an urban location at Bologna in the Po Valley, the mass fractions of NH₄⁺, SO₄²⁻ and NO₃⁻ in PM₁₀ are similar in the daytime and nighttime.³⁰ Furthermore, long-term monitoring at the Nepal Climate Observatory-Pyramid (NCO-P; foot-hills of Mt. Everest) suggested that the mass fraction of NH₄⁺ in ∑ANTH (=NH₄⁺ + SO₄²⁻ + NO₃⁻) was relatively higher in the nighttime and NO₃⁻ is more in daytime, whereas SO₄²⁻ in nighttime is quite similar to that in daytime (Table 2).³¹ Moreover at Mangshan (near Beijing) was noticed that the mass fractions of NH₄⁺, SO₄²⁻ and NO₃⁻ in PM₁ during nighttime were ∼1.5 times higher when compared to those observed in the daytime.³² More interestingly, the nighttime PM₁ mass concentrations (avg.: 146 μg m⁻³) at Mangshan were found to be relatively lower than those in the daytime (267 μg m⁻³). Summing up, the day–night abundance pattern in the mass fraction of secondary aerosol species of this study is quite similar to that reported from a rural location at San Pietro Capofiume in the Po Valley. However, the mass fractions of secondary aerosol species at other locations do not show a unique pattern. The difference in the day–night abundance patterns of aerosol species at locations such as Bologna, NCO-P and Mangshan was attributed to the impact of air-masses from different sources.

Table 2 Average mass fraction (daytime/nighttime) of anthropogenic ionic species from different geographical locations in the world^a

Site	Location	Period	Mass fraction
			NH4^+	SO4^2−	NO3^−
a SPC (San Pietro Capofiume, 11 m amsl) and Bologna (54 m above mean sea level: amsl) are situated in the Po Valley.^30 NCO-P (Nepal Climate Observatory-Pyramid, 5079 m amsl) is at the foot-hills of Mt. Everest^31 and Mangshan (187 m amsl) is near Beijing.^32b Mass fraction of species in PM10.c Mass fraction in ∑ANTH: NH4^+ + SO4^2− + NO3^−.d Mass fraction in PM1.
SPC^b	Rural	June–July 2012	0.09/0.10	0.22/0.16	0.15/0.24
Bologna^b	Urban		0.07/0.07	0.19/0.19	0.16/0.14
NCO-P^c	High-altitude	2006–08	0.30/0.34	0.49/0.50	0.21/0.16
Mangshan^d	Sub-urban	Sept–Oct 2007	0.18/0.27	0.40/0.66	0.42/0.60

---

Summing up, we observe a very high variability in the mass fractions of secondary aerosol species (NH₄⁺, SO₄²⁻ and NO₃⁻) both during the daytime and nighttime in this study. Based on a one-to-one comparison (Table 1), it is evident that SO₄²⁻ was formed predominantly during daytime. In a sharp contrast, NO₃⁻ forms predominantly during the nighttime. The diurnal variability in aerosol chemical composition is basically governed by various physico-chemical activities. It is important to mention here that within the daytime or nighttime the rate of formation of secondary aerosol species can vary significantly depending on their precursor concentration, ambient atmospheric conditions and aerosol loading (heterogeneous phase reactions).³³ Owing to the large surface-area and high mass loading of PM₁ (avg. > 100 μg m⁻³), it is important to assess the secondary aerosol formation in the air-shed of the IGP.

3.4. The proposed reaction mechanism for the heterogeneous formation of SO₄²⁻ and NO₃

We reiterate that SO₄²⁻ formation takes place predominantly during the daytime and NO₃⁻ during the nighttime in this study in PM₁ (high loading) over the IGP. Furthermore, the role of xs-NH₄⁺ (xs: excess) is of utmost importance in governing the NO₃⁻ concentrations. It has been suggested previously that under massive emission of fine-mode aerosols the heterogeneous phase reactivity can be the prominent formation pathway of secondary aerosol species.¹ Furthermore, the impact of Fe³⁺ and Mn²⁺ has been highlighted earlier for the heterogeneous formation of secondary aerosol species.^34,35 With this background information, we propose herein the heterogeneous mechanism of sulfate and nitrate formation: the dominant formation pathway under massive emission of particulate matter.

The reaction mechanism for the formation of SO₄²⁻ via the heterogeneous pathway is proposed below:

SO₂ + H₂O → HSO₃⁻ + H⁺ (1)

M⁽ⁿ⁺¹⁾⁺ + HSO₄⁻ + 5H₂O → [M(HSO₄)(H₂O)₅]ⁿ⁺ (3)

[M(HSO₄)(H₂O)₅]ⁿ⁺ + H₂O → SO₄²⁻ + [M(H₂O)₆]⁽ⁿ⁺¹⁾⁺ + H⁺ (4)

SO₄²⁻ + 2NH₄⁺ ⇋ (NH₄)₂SO₄ (5)

The mechanism for NO₃⁻ formation via the heterogeneous pathway is proposed below:

NO₂ + O₃ → NO₃ + O₂ (6)

NO₃ + NO₂ ⇋ N₂O₅ (7)

[M(H₂O)₆]⁽ⁿ⁺¹⁾⁺ + N₂O₅ + OH⁻ → 2HNO₃ + [M(OH)(H₂O)₅]ⁿ⁺ (8)

HNO₃ + NH₃ ⇋ NH₄NO₃ (9)

where, M⁽ⁿ⁺¹⁾⁺ ≡ Fe³⁺ or Mn²⁺ (3d⁵ system). Here Fe³⁺ can exist also as FeOH²⁺, Fe(OH)₂⁺ and Fe₂(OH)₂⁴⁺. The formation of FeOH²⁺, Fe(OH)₂⁺ and Fe₂(OH)₂⁴⁺ requires reaction with an H₂O molecule. Likewise, the formation of NO₃⁻ requires reaction of N₂O₅ with H₂O (refer to eqn (8)). In this context, a previous study has reported that the reaction of atmospheric species with H₂O is endothermic.³⁶ Thus, previous literature based on laboratory studies suggested that NO₃⁻ and SO₄²⁻ formation via heterogeneous phase reactivity undergo via endothermic pathways.^36,37 Our study reporting the heterogeneous formation of SO₄²⁻ and NO₃⁻ via selective endothermic pathways (discussed below) from field-based measurements is in good agreement with the previous inferences made from laboratory based studies.

In this study, we have measured the ambient concentrations of total Fe and Mn in PM₁. However, the reaction mechanism for heterogeneous-phase catalysis for the formation of SO₄²⁻ and NO₃⁻ involves specifically Fe³⁺ and Mn²⁺ (as proposed above). In this context, it is very important to understand the typical mass fractions of Fe³⁺ and Mn²⁺ in their total abundance reported previously in atmospheric aerosol samples. To start with, aerosol Fe and Mn can originate from the upper continental crust and are also derived from anthropogenic emission sources.^38–40 Fe can exist in two oxidation states +II (water-soluble) and +III (water-insoluble), whereas the most stable state of Mn is +II (water-soluble). These species have not been measured previously in the entire IGP. Furthermore, Mn²⁺ has been studied relatively less when compared to Fe³⁺ (=Fe_total − Fe²⁺) on a global scale.⁴⁰ However, assessment of continental outflow from the IGP (during wintertime; January 2009) in the marine atmospheric boundary layer over the Bay of Bengal has revealed that a large fraction of total Fe exists as Fe³⁺; Fe³⁺/Fe_total ratio >0.75 in PM_2.5 (fine-mode) and >0.90% in PM₁₀ (coarse-mode).²⁰ Furthermore, their study was consistent with a previous study over the Sargasso Sea near Bermuda (North Atlantic Ocean) on TSP (total suspended particles; July–August 2003; April–June 2004), which documented more than 85% of the total Fe to be Fe³⁺.³⁹ These studies have attributed the difference in mass fraction of Fe³⁺ to the variability in the characteristic nature of mineral dust (alluvial versus Thar desert versus Sahara desert) and anthropogenic emissions. Furthermore, Mn²⁺ contribution assessed from an urban background location (Slovenia; March–April 2006) suggests that the Mn²⁺/Mn_total ratio was ∼0.40.⁴⁰ Summing up, previous studies have revealed that the mass fraction of Fe³⁺ is very high in ambient aerosols and the mass fraction of Mn²⁺ is quite significant. Thus, taking in account for the high-loading of pollutants and significant fractions (as per aforementioned studies) of Fe³⁺ and Mn²⁺, the heterogeneous formation of SO₄²⁻ and NO₃⁻ can be the dominant pathway over the IGP. We urge modeling based studies to incorporate the heterogeneous formation pathways (proposed herein) for simulating the global scenario on the abundance of anthropogenic species.

3.5. The heterogeneous formation of SO₄²⁻ and NO₃⁻

We now discuss the field-based measurements of transition metals (Mn and Fe), NH₄⁺ and ambient meteorological parameters (wet-bulb temperature and solar flux) to understand the heterogeneous formation of SO₄²⁻ and NO₃⁻ during the daytime and nighttime. Previous literature reports that Fe³⁺ and Mn²⁺ can catalyze SO₄²⁻ formation in the aqueous-phase.³⁴ Recently our group has discovered that Fe and Mn can catalyze the formation of secondary organic aerosols.³⁵ Previous studies have assessed multi-component reactions and the heterogeneous and homogeneous reactivity of organic compounds in the presence of transition metals.^41,42 The ambient concentrations of Fe and Mn along with the other assessed metals in this study are shown in the ESI (Fig. S1a and b†). The ambient temperature varied from 20.1–29.8 °C during the daytime and 9.5–23.4 °C during the nighttime, whereas the RH varied from 18–79% during the daytime and 29–95% during the nighttime in this study. The solar flux during the study period was centered at 355 W m⁻².

Correlation analysis of SO₄²⁻ and NO₃⁻ with Fe, Mn, NH₄⁺, wet-bulb temperature and solar flux during the daytime exhibits positive correlations (Fig. 4a–e and 5a–e). Interestingly, we observe two distinctly different slopes while assessing the catalysis involving Fe and Mn (Fig. 4a and b and 5a and b). Data points with a lower slope are considered to be associated predominantly with the aqueous-phase formation of SO₄²⁻ and NO₃⁻ due to the higher rate of heterogeneous-phase reactivity during the high pollution period.¹ In contrast, the data points associated with a steeper slope are considered to be representing the predominant role of the heterogeneous-phase reactivity in the formation of SO₄²⁻ aerosols during the daytime. During the nighttime the correlation analyses of SO₄²⁻ with Fe, Mn, NH₄⁺ and wet-bulb T also show the pre-dominant role of heterogeneous-phase sulfate formation over the aqueous-phase pathway (Fig. 4f–i and 5f–i). Thus, it can be inferred from this study over the IGP that Mn and Fe catalyze both the heterogeneous- and aqueous-phase formation of SO₄²⁻ and NO₃⁻ aerosols, with a higher rate for the previous one.

Fig. 4 Correlation analyses of SO₄²⁻ with: (a, f) Fe, (b, g) Mn, (c, h) NH₄⁺ and (d, i) wet-bulb temperature during the day and night. (e) Also shown is the scatter plot of sulfate with solar flux during the daytime.

Fig. 5 Correlation analyses of NO₃⁻ with: (a, f) Fe, (b, g) Mn, (c, h) NH₄⁺ and (d, i) wet-bulb temperature during the day and night. (e) Also shown is the scatter plot of NO₃⁻ with solar flux during the daytime.

3.6. Statistical analysis: reaffirming the role of catalysis in the heterogeneous formation of sulfate and nitrate

We have performed multi-linear regression analysis (MLRA) to affirm the dependence of SO₄²⁻ and NO₃⁻ formation on the several key parameters discussed above. The results of MLRA for SO₄²⁻ and NO₃⁻ aerosol formation, pertaining to heterogeneous and aqueous-phase reactivity, during the daytime and nighttime are shown through the following equations. The correlation coefficient (r value) and level of significance (p-value) for the MLRA fit are mentioned in each equation.

During the daytime the formation of sulfate via heterogeneous (H) and aqueous phase reactivity (A) can be expressed using MLRA as follows:

[SO₄²⁻_H, r: 0.98; p: 0.000] = 5.27[Fe] + 0.24[Mn] + 0.11[NH₄⁺] + 0.31[T_WB] − 0.01[F] (10)

[SO₄²⁻_A, r: 0.98; p: 0.000] = 2.09[Fe] + 0.06[Mn] + 0.30[NH₄⁺] − 0.92[T_WB] + 0.03[F] (11)

here F and T_WB represent the solar flux (W m⁻²) and wet-bulb temperature, respectively. H and A in the subscripts represent the heterogeneous and aqueous phase reactivity, respectively.

We can observe here that the correlation coefficient was high (r = 0.98) and significant (p < 0.05) for the above fits. During the daytime Fe, Mn and NH₄⁺ show positive response with SO₄²⁻ for both the heterogeneous as well as aqueous phase reactivity. Interestingly, between wet-bulb T (T_WB) and solar flux (F) only one parameter exhibits a positive response. Moreover, eqn (10) suggests that the heterogeneous formation pathway is endothermic, whereas the aqueous phase formation of sulfate depends on solar flux.

During nighttime the equations deduced from the MLRA for the formation of sulfate via heterogeneous and aqueous phase reactivity are given below:

[SO₄²⁻_H, r: 0.97; p: 0.000] = −0.50[Fe] + 0.53[Mn] + 0.25[NH₄⁺] + 0.60[T_WB] (12)

[SO₄²⁻_A, r: 0.97; p: 0.000] = 0.25[Fe] + 0.04[Mn] + 0.30[NH₄⁺] + 0.22[T_WB] (13)

During nighttime, all the predictors (eqn (12) and (13)) show a positive response (r = 0.97, p < 0.05) for SO₄²⁻ formation, the exception being Fe in the heterogeneous pathway (eqn (12)). The concentration of Fe was three orders of magnitude higher when compared to Mn in the aerosol samples (this study) and its negative response for sulfate formation via heterogeneous phase reactivity (higher rate when compared to the aqueous phase; Fig. 4a, b, f and g) during the night could be one of the plausible reasons for the higher concentration of SO₄²⁻ during the daytime. Also, it is evident from eqn (12) and (13) that both the formation pathways are endothermic during the nighttime.

Nitrate formation during the daytime via heterogeneous and aqueous phase reactivity:

[NO₃⁻_H, r: 0.98; p: 0.000] = −0.09[Fe] + 0.13[Mn] + 0.32[NH₄⁺] − 0.16[T_WB] + 0.005[F] (14)

[NO₃⁻_A, r: 0.99; p: 0.000] = 0.56[Fe] + 0.001[Mn] + 0.26[NH₄⁺] − 0.37[T_WB] + 0.01[F] (15)

In the daytime, Mn, NH₄⁺ and solar flux show a positive response (r > 0.98, p < 0.05) whereas T_WB exhibits a negative response (exothermic) for NO₃⁻ formation via both pathways (eqn (14) and (15)). The negative response of Fe for NO₃⁻ formation via heterogeneous phase reactivity (eqn (14)) during the daytime could be one of the plausible reasons for the lower concentration of NO₃⁻ in the day.

Nitrate formation during the nighttime via heterogeneous and aqueous phase reactivity:

[NO₃⁻_H, r: 0.99; p: 0.000] = 0.83[Fe] + 0.13[Mn] + 0.35[NH₄⁺] + 0.26[T_WB] (16)

[NO₃⁻_A, r: 0.97; p: 0.000] = 0.35[Fe] + 0.003[Mn] + 0.52[NH₄⁺] − 0.005[T_WB] (17)

In the nighttime, all predictors (eqn (16) and (17)) show a positive response for NO₃⁻ formation, the exception being the case of T_WB for aqueous phase reactivity. The higher rate of heterogeneous phase reactivity and positive response of both Fe and Mn could be the plausible reasons for the higher concentration of NO₃⁻ during nighttime. Furthermore, heterogeneous formation of NO₃⁻ during nighttime is endothermic (eqn (16)). Summing up, SO₄²⁻ and NO₃⁻ formation via heterogeneous phase reactivity relates to an endothermic pathway, the exception being daytime heterogeneous formation of nitrate (eqn (14)). This suggests that the heterogeneous reactivity for SO₄²⁻ (both during the day and night) and NO₃⁻ (only at night) formation undergoes via selective endothermic pathways. Our observation (from field-based measurements) of the heterogeneous phase chemistry via endothermic pathways is quite similar to that previously reported (from laboratory-based experiments).³⁷ They suggest heat is required to dissociate the intermediate N₂O₅. Residual analyses for SO₄²⁻ and NO₃⁻ showed a satisfactory distribution pattern, varying between −3 to +3 (Fig. S3 and S4†). This suggests a good regression fit.

4 Conclusions

PM₁ samples have been collected during the daytime (n = 51) and nighttime (n = 49) during November 09–February 10 at an urban location (Kanpur) in the Indo-Gangetic Plain. PM₁ mass concentrations varied overall from 18–388 μg m⁻³ with the day/night ratio centered at 0.73. The highest concentration of PM₁ is associated with local emissions whereas the lowest is associated with long-range AMBTs (upwind, NW-direction). Sulfate showed predominant formation during the daytime whereas nitrate was during nighttime. The xs-NH₄⁺ relates to the enhanced formation of NO₃⁻ in the aerosol phase, relatively more during the night. Ambient T and RH appeared to influence the formation of SO₄²⁻ and NO₃⁻. Heterogeneous formation of SO₄²⁻ and NO₃⁻ (catalyzed by Fe and Mn) appear to occur via selective endothermic pathways. The proposed mechanism for heterogeneous phase formation of sulfate and nitrate species in atmospheric aerosols is in good agreement with the correlation analysis and MLRA. This study, based on a statistically significant data set, documents the heterogeneous formation process (endothermic pathways) of SO₄²⁻ and NO₃⁻ through field-based measurements in the Indo-Gangetic Plain. Our study has implications to the uptake of various reactive species and thus, radiative forcing estimations.

Acknowledgements

This study has been carried out utilizing the funds from IIT Kanpur. We thank Anil Mandaria for assistance in PM sampling. The authors acknowledge Dr Brent N. Holben and local PI for maintenance and data retrieval from AERONET station. We thank the anonymous reviewers for providing constructive comments and critical suggestions to improve the content of this manuscript. The authors acknowledge Prof. Yiu Fai Tsang for editorial handling of this MS.

References

1. G. Lammel and A. Leip, Environ. Sci. Pollut. Res., 2005, 12, 213–220.
2. C. E. Kolb, R. A. Cox, J. P. D. Abbatt, M. Ammann, E. J. Davis, D. J. Donaldson, B. C. Garrett, C. George, P. T. Griffiths, D. R. Hanson, M. Kulmala, G. McFiggans, U. Pöschl, I. Riipinen, M. J. Rossi, Y. Rudich, P. E. Wagner, P. M. Winkler, D. R. Worsnop and C. D. O'Dowd, Atmos. Chem. Phys., 2010, 10, 10561–10605.
3. P. V. Nidheesh, RSC Adv., 2015, 5, 40552–40577.
4. Q. Liu, W. Wang, Z. Liu, T. Wang, L. Wu and M. Ge, RSC Adv., 2014, 4, 19716–19724.
5. T. Gupta, P. Demokritou and P. Koutrakis, Inhalation Toxicol., 2004, 16, 851–862.
6. S. Menon, J. Hansen, L. Nazarenko and Y. Luo, Science, 2002, 297, 2250–2253.
7. S. Ghosh, T. Gupta, N. Rastogi, A. Gaur, A. Misra, S. N. Tripathi, D. Paul, V. Tare, O. Prakash, D. Bhattu, A. K. Dwivedi, D. S. Kaul, R. Dalai and S. K. Mishra, Aerosol Air Qual. Res., 2014, 14, 879–891.
8. P. Rajput, M. M. Sarin, D. Sharma and D. Singh, Tellus, Ser. B, 2014, 66, 21026.
9. P. Rajput, M. M. Sarin, D. Sharma and D. Singh, Environ. Sci. Pollut. Res., 2014, 21, 5724–5729.
10. P. Rajput, M. M. Sarin and R. Rengarajan, J. Environ. Prot., 2011, 2, 445–453.
11. P. Rajput, A. Mandaria, L. Kachawa, D. K. Singh, A. K. Singh and T. Gupta, Clim. Change, 2015, 1, 503–507.
12. A. Singh, P. Rajput, D. Sharma, M. M. Sarin and D. Singh, Advances in Meteorology, 2014, 2014, 10.
13. P. Rajput, M. M. Sarin, D. Sharma and D. Singh, Environ. Sci.: Processes Impacts, 2014, 16, 2371–2379.
14. P. Rajput and M. M. Sarin, Chemosphere, 2014, 103, 74–79.
15. N. Rastogi, A. Singh, D. Singh and M. M. Sarin, Environ. Pollut., 2014, 184, 563–569.
16. A. Kumar and T. Gupta, Aerosol Air Qual. Res., 2015, 15, 768–775.
17. A. Chakraborty and T. Gupta, Aerosol Air Qual. Res., 2010, 10, 433–445.
18. T. Gupta, A. Chakraborty and K. K. Ujinwal, Development and Performance Evaluation of an Indigenously Developed Air Sampler Designed to Collect Submicron Aerosol, Annals of the Indian National Academy of Engineering (INAE), 2010, vol. 7, pp. 189–193.
19. T. Gupta, Jaiprakash and S. Dubey, Sci. Total Environ., 2011, 409, 3500–3507.
20. B. Srinivas, M. M. Sarin and A. Kumar, Biogeochemistry, 2011, 1–12.
21. B. Srinivas, A. Kumar, M. M. Sarin and A. K. Sudheer, Atmos. Chem. Phys. Discuss., 2011, 11, 20667–20711.
22. P. Rajput, A. Mandaria, L. Kachawa, D. K. Singh, A. K. Singh and T. Gupta, Tellus, Ser. B, 2016, 68, 30659.
23. E. Kalnay, et al., Bull. Am. Meteorol. Soc., 1996, 77, 437–470.
24. A. F. Stein, R. R. Draxler, G. D. Rolph, B. J. B. Stunder, M. D. Cohen and F. Ngan, Bull. Am. Meteorol. Soc., 2015, 96, 2059–2077.
25. V. A. Lanz, A. S. H. Prévôt, M. R. Alfarra, S. Weimer, C. Mohr, P. F. DeCarlo, M. F. D. Gianini, C. Hueglin, J. Schneider, O. Favez, B. D'Anna, C. George and U. Baltensperger, Atmos. Chem. Phys., 2010, 10, 10453–10471.
26. A. Ripoll, M. C. Minguillón, J. Pey, J. L. Jimenez, D. A. Day, Y. Sosedova, F. Canonaco, A. S. H. Prévôt, X. Querol and A. Alastuey, Atmos. Chem. Phys., 2015, 15, 2935–2951.
27. J. D. Surratt, Y. GoÌmez-GonzaÌlez, A. W. H. Chan, R. Vermeylen, M. Shahgholi, T. E. Kleindienst, E. O. Edney, J. H. Offenberg, M. Lewandowski, M. Jaoui, W. Maenhaut, M. Claeys, R. C. Flagan and J. H. Seinfeld, J. Phys. Chem. A, 2008, 112, 8345–8378.
28. T. Gupta and A. Mandaria, Environ. Sci. Pollut. Res., 2013, 20, 5615–5629.
29. C. M. Pavuluri, K. Kawamura, S. G. Aggarwal and T. Swaminathan, Atmos. Chem. Phys., 2011, 11, 8215–8230.
30. S. Sandrini, D. van Pinxteren, L. Giulianelli, H. Herrmann, L. Poulain, M. C. Facchini, S. Gilardoni, M. Rinaldi, M. Paglione, B. J. Turpin, F. Pollini, N. Zanca and S. Decesari, Atmos. Chem. Phys. Discuss., 2016, 2016, 1–31.
31. S. Decesari, M. C. Facchini, C. Carbone, L. Giulianelli, M. Rinaldi, E. Finessi, S. Fuzzi, A. Marinoni, P. Cristofanelli, R. Duchi, P. Bonasoni, E. Vuillermoz, J. Cozic, J. L. Jaffrezo and P. Laj, Atmos. Chem. Phys., 2010, 10, 4583–4596.
32. N. He, K. Kawamura, Y. Kanaya and Z. Wang, Atmos. Environ., Part A, 2015, 123, 18–24.
33. C. R. Usher, A. E. Michel and V. H. Grassian, Chem. Rev., 2003, 103, 4883–4940.
34. J. H. Seinfeld and S. N. Pandis, Atmospheric Chemistry and Physics – From Air Pollution to Climate Change, John Wiley & Sons, New York, 2nd edn, 2006.
35. D. K. Singh and T. Gupta, J. Aerosol Sci., 2016, 94, 56–69.
36. A. R. Ravishankara, Science, 1997, 276, 1058–1065.
37. M. Hallquist, D. J. Stewart, S. K. Stephenson and R. A. Cox, Phys. Chem. Chem. Phys., 2003, 5, 3453–3463.
38. S. M. C. McLennan, Geochem., Geophys., Geosyst., 2001, 2, 2000GC000109.
39. E. R. Sholkovitz, P. N. Sedwick and T. M. Church, Geochim. Cosmochim. Acta, 2009, 73, 3981–4003.
40. J. Tursic, H. Radić, M. Kovacević and M. Veber, Arh. Hig. Rada Toksikol., 2008, 59, 111–116.
41. M. J. Climent, A. Corma and S. Iborra, RSC Adv., 2012, 2, 16–58.
42. K. T. Mahmudov, M. Sutradhar, L. M. D. R. S. Martins, M. F. C. Guedes da Silva, A. Ribera, A. V. M. Nunes, S. I. Gahramanova, F. Marchetti and A. J. L. Pombeiro, RSC Adv., 2015, 5, 25979–25987.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra19595d

---