Characteristics of Aerosol Water Content and Its Implication on Secondary Inorganic Aerosol Formation during Sandy Haze in an Inland City in China

Abstract

Sand events continue to occur frequently and affect the North China region. Under unfavorable meteorological conditions, they can easily combine with haze pollution, forming sandy haze events that have a significant impact on human health. Aerosol water content (AWC) is known to have a significant impact on PM_2.5, but its effect is still unclear in sandy haze. In this work, sandy haze and haze periods were observed in Zhengzhou using a series of high-time-resolution instruments. The AWC calculated by the ISORROPIA-II model reached 11 ± 5 μg m⁻³, accounting for 10% of the PM_2.5, in the sandy haze period. Sensitivity tests show that AWC was mainly relative humidity (RH)-dependent. Additionally, elevated SO₄²⁻, TNO_3, and TNH₄ were crucial in the increase in AWC. The increase in Ca²⁺ ions in the sandy haze led to lower AWC than that in the haze periods. Specifically, (NH₄)₂SO₄ was the major contributor to the AWC when the RH was between 30 and 46% in the sandy haze period, and NH₄NO₃ gradually became the main contributor with the increase in RH. In turn, AWC could enhance the formation of sulfate and nitrate, even during the sandy haze period. Therefore, the emergency control of gaseous precursors should also be implemented before the sand events.

1. Introduction

Fine particulate matter (PM_2.5, aerodynamic diameter ≤ 2.5 μm), a major atmospheric pollutant, has significant adverse effects on human health [1,2], reduces visibility [3], damages ecosystems, and alters regional climates [4]. PM_2.5 can absorb water from the air to form aerosol water content (AWC) on its surface, especially when relative humidity (RH) meets or exceeds the deliquescence relative humidity (DRH). AWC is an important component of PM_2.5 and generally accounts for more than 50% of ambient PM_2.5 mass, especially during haze events [5,6]. In addition, the AWC may affect the optical properties of PM_2.5 [7], particle volumes [8], gas-particle transformations, and the formation of inorganic aerosol (nitrate, sulfate, and ammonium, SNA) and organic matter (OM) through direct participation in atmospheric complex physical and chemical mechanisms [9,10], thereby playing profound roles on air quality degradation [11].

The formation of AWC mainly depends on aerosol mass concentration, particle hygroscopicity, the aerosol phase, and RH [12]. Field surveys by Wang et al. [13] showed that simultaneously elevated RH levels and SNA resulted in an abundant AWC during haze episodes. Wu et al. [14] and Speer et al. [15] quantified that SNA was associated with 62–80% of aerosol AWC. Laboratory experiments by Ansari and Pandis [16] confirmed that inorganic salts like (NH₄)₂SO₄ were the key factor in AWC formation. Moreover, crustal species play an important role in the AWC, in which Ca²⁺ can form insoluble CaSO₄ with sulfate, lowering AWC concentrations [17,18]. On the other hand, AWC can affect the formation of secondary aerosol by offering a medium for heterogeneous gas/particle reactions and the partitioning of water-soluble compounds in the aerosol phase, such as the major aqueous-phase formation paths for sulfate, which are the S(IV) referred to as SO₂·H₂O + HSO₃⁻ + SO₃⁻ oxidation by H₂O₂, NO₂, O₃, the catalysis by transition metal ions (TMIs, e.g., Fe (III) and Mn (II)), and the heterogeneous hydrolysis reaction of NO₂ and N₂O₅ with water to form nitrate [19,20]. In addition, AWC is a crucial factor in determining aerosol acidity. High AWC and excess NH₃ have been linked to the low acidity of aerosols during severe haze in Northern China [21]. In addition, mineral dust particles play an important role in SO₂ and NO₂ conversion by supplying TMIs and undergoing heterogeneous reactions [22,23]. Moreover, mineral dust can modify the particle pH, and eventually affect SNA formation [17,23]. In summary, there are complex effects between AWC, particle pH, and mineral dust. Several mixing processes of sand and haze, accompanied by high concentrations of secondary and crustal species as well as high RH, have been observed in Northern China [24,25], which can supply a specific scenario for an in-depth study of their relationships.

Hourly data, including PM_2.5, PM₁₀, the chemical composition of PM_2.5, gaseous pollutants, and meteorological parameters, were collected from January 6 to 26, 2021, using a series of online monitoring instruments. The Thermodynamic ISORROPIA-II model and Zdanovskii−Stokes−Robinson (ZSR) mixing rule were used to estimate the total AWC values of particles and AWC of major components of particles. The main purposes of this study are as follows: (1) to study the characteristics of AWC under sandy haze and haze pollution periods, and (2) to use the sensitivity tests to quantify the influence of major driving forces on AWC concentration during the observation period. To further understand the effects of AWC on secondary inorganic compound formation, we calculated the contributions of main composition forms to AWC. This work is useful to awaken awareness of the atmospheric chemistry during the sandy haze episodes and guide the measures for the emergency control of severe pollution in the future.

2. Experiment and Method

2.1. Observation Site

Field sampling was conducted from 6 to 26 January 2021, at Zhengzhou Environmental Monitoring Station (34.75° N, 113.61° E) (Figure S1), which is surrounded by busy roads, with the sampling equipment located on the roof of the monitoring station, at a height of 15 m. Moreover, this site is near a large personnel flow of the region such as the Zhengzhou Institute of Technology, residential areas, and hospitals.

2.2. Measuring Instruments

The working principle of measuring instruments, quality assurance/quality control, and measurement uncertainties can be found in the Supplementary Materials (Text S1) and previous research by Wang et al. [26]. Briefly, water-soluble inorganic ions (NH₄⁺, SO₄²⁻, NO₃⁻, Cl⁻, Na⁺, Mg²⁺, Ca²⁺, and K⁺) in PM_2.5 and trace gases (NH₃, HNO₃, HONO, and HCl) were continuously measured by an ambient ion monitor (URG-9000D, Thermo Science, Waltham, MA, USA). The hourly concentrations of metal elements such as Fe and Mn in PM_2.5 were detected by an energy-dispersive X-ray Fluorescence instrument (Xact 625, Cooper Environmental Science, Portland, OR, USA). Hourly organic carbon (OC) and elemental carbon (EC) concentrations in PM_2.5 were obtained by a thermo-optical transmission method operating a semi-continuous Carbon Analyzer (Model 4, Sunset Laboratory Inc., Tigard, OR, USA). The mass concentrations of PM_2.5, PM_10, and gaseous pollutants (SO₂, NO₂, CO, and O₃) were measured by a series of online analyzers (Model 1400a, 1405 42i, 43i, 48i, and 49i, Thermo Fisher Scientific, Waltham, MA, USA). Meteorological parameters including temperature (Temp), RH, wind direction (WD), and wind speed (WS) were simultaneously monitored by the city station (LUFFT WS500, Sutron Corporation, Sterling, VA, USA).

2.3. Date Analysis

2.3.1. ISORROPIA II Model

The pH values and AWC_inorg of particles were calculated using the ISORROPIA-II model (http://isorropia.eas.gatech.edu, accessed on 1 January 2020), which had been implemented and reported in previous research [27]. In this study, ISORROPIA-II operates in forward mode and a “metastable” state. The calculations using total (gas and aerosol) measurements in the forward model are less affected by measurement errors. The particle water associated with the organics (AWC_org) was estimated by the organic hygroscopicity and the mass concentrations of organic matter (OC × 1.6). Detailed descriptions of pH values and AWC calculations can be found in previous studies [28,29] and Text S2 in Supplementary Materials.

2.3.2. Calculating for Aqueous-Phase Sulfate Production Rates

Refer to Cheng et al. [19] to calculate the formation rate of sulfate by liquid-phase reactions, such as S(IV) oxidation by O₃, TMI, and NO₂. These methods and their reaction formulations are detailed in Supplementary Materials (Tables S1–S3, Texts S3 and S4).

2.3.3. Sensitivity Test of AWC

To explore the main influencing factors of aerosol AWC and quantify the influence of inputs on AWC, sensitivity tests were conducted. In the sensitivity analysis, SO₄²⁻, TNO₃ (total nitrate (gas + aerosol) represented by equivalent HNO₃), TNH₃ (total ammonium (gas + aerosol) represented by equivalent NH₃), Ca²⁺, K⁺, Na⁺, Mg²⁺, Cl⁻, RH, and temperature were selected as the variables to assess how a variable affects PM_2.5 AWC, and the real-time measured values of this variable and the average values of other ions in different pollution periods were input into ISORROPIA II. The relative standard deviation (RSD) of calculated aerosol AWC can reflect the influence of variable variation on aerosol water content. AWC affects these variables, but RH affects AWC. The sensitivity analysis in this work is only for PM_2.5. Considering that part of the SNA may be transported in the sandy haze and haze periods [30], by comparing the correlation between nitrate, sulfate, and EC, as well as crustal material (CM) and SNA in different periods (Figure S2), we found that the high correlations between nitrate, sulfate, and EC and low correlations between SNA and CM indicated that SNA was generated locally and was less affected by sand and dust transport.

3. Results and Discussion

3.1. Sandy Haze and Haze Events

Figure 1 illustrates the temporal variation in particle mass concentration, gas pollutants, and meteorological factors from January 6 to 26, 2021. The observation period is divided into three scenes based on pollutant concentration under varying meteorological conditions with their average values listed in Table S4. During the sandy haze period (January 11–16), PM₁₀ and Ca²⁺ concentrations rose sharply from 71 μg/m³ in the clean period to 274 μg/m³ and from 1.4 to 3.9 μg/m³, respectively, driven by high WS likely due to sandstorms originating from Northwest China (Figure S3). After that, an extended haze event developed over the January 16–26 period (labeled haze in Figure 1), with an average PM_2.5 concentration of 111 ± 63 μg/m³. During the sandy haze period, the average concentrations of gaseous pollutants (SO₂, NO₂, and CO) in the sandy haze periods were about twice as high as those in the clean period. In addition, the Temp and RH slightly increased in the sandy haze period. These results indicate that atmospheric pollutants accumulated during the sandy haze period accompanied by sandstorm effects.

3.2. Role of Aerosol Water Content on PM_2.5

Figure 2a shows the 2D kernel density graphs between the PM_2.5 and the ratio of AWC/PM_2.5 under different pollution levels. The ratio of AWC/PM_2.5 increased as PM_2.5 levels increased. Ratios were three times higher during the most polluted periods (0.67, PM_2.5 > 225 μg m⁻³) than those during clean and moderately polluted conditions (ratios of 0.2 when PM_2.5 < 75 μg m⁻³). This is consistent with previous studies that enhanced PM_2.5 mass concentration was accompanied by the persistent growth of AWC [31,32,33]. The mass fractions of the PM_2.5 components and calculated AWC are grouped by the different pollution periods, as displayed in Figure 2b. SNA dominated the PM_2.5 mass concentration and accounted for 38 and 50% of the sandy haze and haze periods, respectively. In addition, the proportion of CM in the sandy haze period (29 ± 10 μg m⁻³) was more than four times that in the haze period (7 ± 3 μg m⁻³), accounting for 27 and 3% in PM_2.5 mass concentration, respectively.

The AWC was 11 ± 5 and 73 ± 68.4 μg m⁻³ in the sandy haze and haze periods accounted for 10 and 31% of the PM_2.5, respectively. The RH condition was primarily responsible for the low AWC in the sandy haze period. Furthermore, we divided the periods into five groups according to the actual humidity range (Figure 3). When RH < 30% represented the dry conditions, each 5% rise in RH was taken as a group to represent the gradual transformation of dry conditions into humid conditions. It can be found that the average concentration of PM_2.5 in the sandy haze period (67 ± 9 μg m⁻³) was twice that in the haze period (36 ± 3 μg m⁻³) under dry conditions, but the average concentration of PM_2.5 in the haze period under humid conditions was about twice as high as in the sandy haze period and has quadrupled, and except for CM elements, the proportions of components are similar. Meanwhile, with an increasing RH, the AWC increased significantly during different periods. With the increase in RH, AWC increased from 8 to 15% in the sandy haze period and from 11 to 19% in the haze period. The differences between 8% and 15% (differed by 7%) and between 11% and 19% (differed by 8%) are limited, showing that the AWC increase in sandy haze and polluted haze is similar. Thus, high RH might determine the AWC concentrations under both haze and sandy haze periods and lead to an increase in PM_2.5, which is consistent with previous studies [34,35].

3.3. Driving Factors of AWC

Sensitivity tests were conducted to assess the influence of chemical components and meteorological parameters on AWC. As shown in Figure 4, the most important factor for AWC was RH, followed by SO₄²⁻ and TNO₃. The AWC concentrations rose slowly over the RH range from 30 to 80% (Figure 4e). Significant explosive growth in AWC occurred when RH exceeded 80%, which was higher than the DRH of most chemical components in PM_2.5 and reached the point of dampness to form a liquid state [36,37]. The temperature had little effect on AWC in the range of −3–13 °C. However, when the temperature increases above 20 °C, the AWC will exhibit an obvious downward trend [38,39].

It was found that the RSD of AWC variations (Figure 4e) in the haze period (74.7%) was higher than that in the sandy haze period (26.2%), which may be due to the difference in components. Figure S4 suggests that AWC was well correlated with SNA (R² > 0.5), but poorly correlated with other components (R² < 0.5), indicating a strong interplay between inorganic composition and AWC. Figure 4a reveals that elevated SO₄²⁻ was crucial in the increase in AWC and had a larger influence on AWC concentration in the haze period (RSD = 17.7%) than in the sandy haze period (RSD = 3.4%). The effect of TNO₃ and TNH₄ on AWC was similar to that of SO₄²⁻, but the effect flattens out after a typical value, which is consistent with previous studies [38]. This phenomenon may be because there is not enough TNO₃ to react with the excessive levels of TNH₄. In the process of increasing the concentration of NH₃ in the ammonia−nitrite−sulfate−water system, NH₃ reacts with sulfuric acid first and then with HNO₃ to form ammonium nitrate. After most of the nitric acid is converted to ammonium nitrate, it is difficult to dissolve more ammonia into aerosol droplets [39]. In contrast, the increase in Ca²⁺ ions led to a decrease in AWC. Therefore, the low AWC concentration in the sandy haze period may be affected by low RH, low SNA concentration, and high Ca²⁺ concentration.

To reduce the complexity of the aerosol system, two important parameters (sulfate ratio: RSO₄² = [Na⁺+NH₄⁺]/[SO₄²⁻], sodium salt ratio: RNa⁺ = [Na⁺]/[SO₄²⁻]) are introduced to divide different sub-systems [40]: (i) if RSO₄²⁻ < 1, excessed sulfate existed as the chemical forms of NaHSO₄ and NH₄HSO₄; (ii) 1 ≤ RSO₄²⁻ < 2, sulfate existed as NaHSO₄, NH₄HSO₄, Na₂SO₄, (NH₄)₂SO₄, and (NH₄)₃H(SO₄)₂; (iii) if RSO₄²⁻ ≥ 2 and RNa⁺ < 2, then insufficient sulfate and sodium existed as Na₂SO₄, (NH₄)₂SO₄, NH₄NO₃, and NH₄Cl; and (iv) if RSO₄²⁻ ≥ 2 and RNa⁺ ≥ 2, insufficient sulfate and excess sodium, the constituent forms include Na₂SO₄, NaNO₃, NaCl, NH₄NO₃, and NH₄Cl. Figure 5 shows the contribution of the chemical composition to the AWC quantified by changing the RH value using the method of ZSR [41]. CaCl₂ and (NH₄)₂SO₄ were the major contributors to the AWC, accounting for 39% and 49%, when RH was between 30 and 46% in the sandy haze period and when RH was between 30 and 57% in the haze period, respectively. With the increase in RH, the contribution of (NH₄)₂SO₄ and Na₂SO₄ decreased, and NH₄NO₃ gradually became the main contributor to the sandy haze period (RH range: 47–90%) and haze period (RH range: 58–90%), contributing 45% and 56%, respectively. It is evident that (NH₄)₂SO₄ has higher water uptake than NH₄NO₃ at an RH lower than 50%, which is consistent with the previous study in India [42].

3.4. Impact of the AWC on Secondary Aerosol Formation

The sulfur oxidation ratio (SOR) and nitrogen oxidation ratio (NOR) were defined as the molar ratio of sulfate to total sulfur and nitrate to total nitric acid [43,44]:

$$\mathrm{SOR}={\displaystyle \frac{n{(\mathrm{SO}}_4^{2-})}{n{(\mathrm{SO}}_4^{2-})+n{(\mathrm{SO}}_2)}}$$

(1)

$$\mathrm{NOR}={\displaystyle \frac{n({\mathrm{NO}}_3^{-})}{n({\mathrm{NO}}_3^{-})+n({\mathrm{NO}}_2)}}$$

(2)

where n represents the molar concentration (mol m⁻³).

To understand the impacts of AWC on nitrate and sulfate, Figure 6 plotted the correlation of AWC with sulfate and nitrate at day and night mapped SOR and NOR, respectively. The results show higher AWC concentrations and correlation coefficients between SO₄²⁻ and NO₃⁻ with AWC at night than those during the day. Therefore, to analyze the effect of AWC on SNA generation, nighttime data were selected for further analysis. Research on the formation of sulfate has determined that the formation of sulfate during haze was mainly from the aqueous-phase reactions [26].

The sulfate production rates of aqueous-phase reactions, such as S(IV) oxidation by O₃, TMI, and NO₂, are calculated to identify further the AWC effect on sulfate formation in different pollution periods, as seen in Figure 7a. The effect of AWC on the oxidation rate of TMI was negligible, but the oxidation rates of NO₂ and O₃ increased with the increase in AWC. However, the oxidation pathways of NO₂ and O₃ calculated in the datasets of the sandy haze and haze periods were lower than the S(IV) pathway of TMI oxidation shown in Figure 7b. In addition, there were significant differences in the rates of TMI oxidation between the sandy haze and haze periods, which cannot be explained by AWC. The pH, another factor affecting the sulfur oxidation rate, is analyzed in Figure 7c. The aqueous oxidation rates of S(IV) by NO₂, O₃, and TMI were strongly pH-dependent. The NO₂ and O₃ oxidation rates increased with increasing pH, and TMI-catalyzed oxidation decreased with pH. The trend is consistent with previous research on pH sensitivity [45,46]. Therefore, the rate of TMI-catalyzed (0.9 μg m⁻³ h⁻¹) in the haze period was more than two orders of magnitude higher than that of the sandy haze period (7.6 × 10⁻³ μg m⁻³ h⁻¹), which may be attributed to the pH (3.7–6) in the haze period. The high pH value in the sandy haze period might be due to the high Ca²⁺, Mg²⁺, and K⁺ concentrations since the low volatility of these cations allows them to preferentially neutralize sulfates and nitrates over NH₃ [47]. In summary, AWC provided a medium for the aqueous-phase reaction of sulfate, while high pH suppressed the dominant reactions (i.e., TMI oxidation) during the process of sandy haze.

The formation pathway of nitrate in the atmosphere is different during the day and night due to the light conditions. During the daytime, nitrate mainly comes from the oxidation of NO₂ by OH radicals in the gas phase, and at night, it mainly comes from the heterogeneous hydrolysis of N₂O₅ [48,49]. Considering that NO₃ radicals in the atmosphere are mainly produced by the oxidation of NO₂ by O₃, [50], the correlation between [NO₂]² × [O₃] and NOR can roughly represent the contribution of N₂O₅ to the heterogeneous hydrolysis of atmospheric nitrate at night. As shown in Figure 8, the variations in [NO₂]² × [O₃] in the nighttime were all positively correlated with NOR during both the haze (R² = 0.6) and sandy haze periods (R² = 0.11). Therefore, the heterogeneous hydrolysis of N₂O₅ made a remarkable contribution to atmospheric nitrate even during the sandy haze episodes.

4. Conclusions

A comprehensive analysis was performed using a series of high-time-resolution instruments in Zhengzhou during sandy haze and haze periods. The ISORROPIA-II model was used to investigate the AWC value and its driving factors. The AWC was 11 ± 5 and 73 ± 68 μg m⁻³ in the sandy haze and haze periods and accounted for 10% and 31% of the PM_2.5, respectively. The predicted AWC values of PM_2.5 were significantly affected by the particle components and meteorological parameters. AWC increased slowly at lower RH and evolved to a sharp rise above 80% RH. Elevated SO₄²⁻, TNO_3, and TNH₄ were crucial in the increase in AWC and had a larger influence on AWC concentration in the haze period than in the sandy haze period. (NH₄)₂SO₄ was the major contributor to the AWC when the RH was between 30 and 46% in the sandy haze period and when the RH was between 30 and 57% in the haze period, accounting for 28% and 49%, respectively. With the increase in RH, NH₄NO₃ gradually become the main contributor to the sandy haze period (RH range: 47–90%) and haze period (RH range: 58–90%), contributing 45% and 56%, respectively. The increase in Ca²⁺ ions led to a decrease in AWC. Therefore, the low AWC concentration in the sandy haze period may be affected by RH, and in turn, it will subsequently exhibit high Ca²⁺ concentration.

The effect of AWC on the oxidation rate of TMI was negligible, but the rates of NO₂ and O₃ oxidation increased with the increase in AWC. The low rate of TMI-catalyzed in the sandy haze period may be attributed to the high pH as a result of the high Ca²⁺, Mg²⁺, and K⁺ concentrations. In addition, AWC also promoted the formation of nitrate by affecting the heterogeneous hydrolysis of N₂O₅. Therefore, the AWC in PM and its impact on secondary particulate matter components during the sandy haze periods should not be ignored.