Evidence for a New Oxidation Mechanism for Sulfur Dioxide from Laboratory Measurements

1. Introduction

The oxidation of sulfur dioxide (SO₂) produces products that are significant sources of atmospheric particulate matter and acid precipitation (acid rain) [1,2,3,4]. The products, small droplets of liquid sulfuric acid (H₂SO₄), bisulfate (HSO₄^-) and sulfate (SO₄⁼) have consequences for health, agriculture, climate, etc. SO₂ is emitted from many natural and anthropogenic sources [4,5]. The greatest source of sulfur emissions into the atmosphere comes from biological processes in the Earth’s oceans and this means that these sources are stronger in the southern hemisphere. Dimethyl sulfide (DMS, CH₃SCH₃) constitutes a significant fraction of the sulfurous emissions from the oceans [5,6,7,8,9,10]. Hydrogen sulfide (H₂S) is another biogenically emitted compound. H₂S and DMS are converted to SO₂ by a complicated chemical reaction mechanism [11].

Natural sources of SO₂ include volcanoes [5] and its photochemical production from biogenic emissions such as hydrogen sulfide (H₂S) and dimethyl sulfide (CH₃SCH₃) Biomass burning, i.e. wildfires, is an underappreciated source of sulfurous compound emissions [12]. The wildfire sources of sulfur compound emissions is increasing because of the increasing number and intensity of wildfires [13]. The major anthropogenic sources include the production of metals from sulfur containing ores (smelting) and the combustion of fossil fuels [14]. Coal is an example of a fossil fuel with sources that have large differences in sulfur content [15].

Sulfur dioxide is oxidized in the atmosphere to produce sulfate particles, i.e., H₂SO₄, HSO₄^- or sulfate SO₄⁼, that are chemical components of particulate matter or aerosols [3]. This oxidation of sulfur dioxide to produce particulate matter is very important because of their effects on climate, precipitation patterns, amounts and acidification, and their health effects [4]. The chemistry may have possible applications to exoplanetary atmospheres too [16].

Sulfur containing particles have very significant effects on climate because they strongly scatter solar radiation which effects the Earth’s radiation budget [4]. In fact, the emission of sulfate containing particles has been suggested as a method of cooling the Earth’s atmosphere to counteract global warming [17].

Sulfate containing particles, along with other forms of particulate matter, contribute to the secondary aerosol effect [4]. Fine aerosol particles provide nucleation sites for the formation of cloud water droplets. If there is an increase in the concentration of fine aerosol particles, then there will be more, but smaller droplets formed for a given level of water vapor in the atmosphere. Depending upon conditions the secondary aerosol effect can affect the amount and geographical distribution of precipitation.

Sulfate compounds are acidic in liquid water, and they contribute strongly to the acidification of precipitation along with nitrates (NO₃^-) and organic acids such as formic (HCOOH) and acetic acid (CH₃COOH) [18]. For this reason and direct health effects, emissions of SO₂ have regulated in the United States by the Clean Air Act Amendments since 1970 [19]. Acid precipitation was recognized as a major problem in Europe, Canada and the United States [20,21] during the late 1970s and 1980s. One of the major driving political and scientific questions for public policy regarding acid precipitation was to determine if the production of sulfate was limited by the available sulfur dioxide (SO₂) or by the available oxidant; this uncertainty was known as the oxidant limitation question. This was a major issue of several international programs including the American National Acid Precipitation Program (NAPAP) and the U.S. Department of Energy’s Processing of Emissions by Clouds Program (PRECP) [22,23,24,25,26].

Figure 1 shows the difference between a situation where the oxidation of SO₂ is limited by the available oxidant or not [27]. If there is a limited amount of oxidant relative to the amount of SO₂, left plot, then moderate reductions in SO₂ will not result in less sulfate produced; this condition was known as oxidant limited. The right panel shows that if there is more oxidant available then SO₂ then reductions in SO₂ will result in reductions in the amount of sulfate produced. Finding the chemical mechanisms for the oxidation of SO₂ is important for determining the emission reduction policies needed to reduce acid precipitation.

The oxidation of SO₂ may be important on local, urban scales as well as regional and global. In general, the particulate matter produced through atmospheric chemical reactions are fine particles with an aerodynamic dynamiter of 2.5 µm and these particles are known as PM_2.5. Larger particles with an aerodynamic dynamiter of 10 µm are known as PM₁₀ [19]. As an example, the city of El Paso Texas in the Paso Del Norte Border Region of the United States and Mexico has had problems with high concentrations of PM_2.5 and PM₁₀ [28,29]. High emissions of SO₂, NO_x and volatile organic compounds and wind-swept mineral dust have occurred in El Paso which contributes to particulate pollution. Emissions of nitrogen oxides are significant because the region contains the American – Mexican border. Truck and auto traffic are queued near the border until they are cleared to pass the border into the United States.

1.1. SO₂ Oxidation Mechanisms

Many atmospheric chemical reactions produce products that condense to form particulate matter, PM_2.5 or PM₁₀. In general, SO₂ is oxidized in either gas or aqueous phases or it may be deposited directly to the ground. In this paper we present evidence that there many be chemical reactions that may involve nitrogen oxides and/or heterogeneous processed involving sulfate aerosols.

1.1.1. The Gas-Phase Oxidation of Sulfur Dioxide

There are two gas-phase mechanisms known mechanisms for the gas-phase oxidation of sulfur dioxide. The hydroxyl radical (HO•) is the oxidant [30]. As shown by the reactions given below the HO• reacts with SO₂ to produce the adduct HOSO₂•, note that M is any molecule of air that stabilizes the formation of the adduct by absorbing excess collision energy.

HO• + SO₂ (+ M) → HOSO₂• (+ M)

(1)

HOSO₂• + O₂ → HO₂• + SO₃

(2)

SO₃ + H₂O → H₂SO₄ (+ liquid aerosol)

(3)

The HOSO₂• adduct reacts with molecular oxygen to produce the hydroperoxyl radical (HO₂•) and sulfur trioxide (SO₃) [30]. SO₃ reacts with water vapor to produce sulfuric acid. In a nitrogen oxide polluted atmosphere, HO₂• reacts with NO to reproduce HO• as shown below.

HO₂• + NO → HO• + NO₂

(4)

The hydroxyl radical produced by Reaction 4 can react with SO₂ making this mechanism a chain mechanism. When the concentration of nitric oxide (NO) is low, HO₂• reacts with another HO₂• or with organic peroxy radicals to produce hydrogen peroxide (H₂O₂) or an organic peroxide.

Another gas-phase process for the oxidation of SO₂ reaction with Criegee intermediates [31,32]. Criegee intermediates are produced by reactions of ozone (O₃) with alkenes. An example of the mechanism for the production of Criegee intermediates from ethene (CH₂CH₂) and its oxidation of SO₂ is given below, Reactions 5–6.

O₃ + CH₂CH₂ → CH₂OOOCH₂

(5)

CH₂OOOCH₂ → CH₂OO + CH₂O

(6)

CH₂OO + SO₂ → SO₃ + CH₂O

(7)

The average total rates of the HO• radical and the Criegee intermediate mechanisms are a few percent per hour for a range of realistic conditions with a maximum rate of 6.13 % hr^-1 [33].

In cloud water or water coated aerosols the H₂O₂ may react with SO₂ to produce sulfate [1,4,34] and in the gas-phase SO₂ does not affect the concentration of the HO•. Either way there is little or no oxidant limitation to sulfate production from SO₂ in the lower troposphere.

However, there is reason to suspect that could be additional oxidation processes. Relative to observations, air quality models underestimate sulfate production across the Eastern U.S and the models underestimate the effect of temperature on its production [35,36]. These newer modeling studies for United States, Alaska and Beijing, China indicate that estimates of particulate sulfur may be improved by the addition of additional heterogeneous chemical reactions [36]. In this paper we present an analysis of experimental data that suggests a process for the oxidation of SO₂ that is not included in models.

2. Materials and Methods

The experiments that are reanalyzed here were made with the objective to measure the rate coefficient of the HO• radical with SO₂ (HO+SO₂) reaction) relative to the rate coefficient of the HO• radical with carbon monoxide (HO+CO) reaction [30,37]. The reaction of HO• with SO₂ (Reaction 1) is given above and the reaction of with HO• with CO is given below.

HO• + CO (+ O₂) → HO₂• + CO₂

(8)

The original analysis determined that HO• radical concentrations were not affected by SO₂ concentrations, and this observation led to the proposal of the mechanism given by Reactions 1 through 4 [30]. This study was cited as a landmark in sulfur oxide chemistry [38]. The proposed mechanism was verified by several independent studies [39,40].

Mixtures of SO₂, carbon monoxide (CO), nitrous acid (HONO), nitric oxide (NO) and nitrogen dioxide (NO₂) were photolyzed in molecular nitrogen (N₂) or synthetic air as a background gas, Table 1 [37]. The initial reactive nitrogen is the total molecular concentration of NO, NO₂ and HONO, Table 1 gives the initial conditions for the experiments made in N₂; note that there was a small trace of molecular (O₂) introduced during the filling of the chamber. Table 2 gives the initial conditions for the experiments made in synthetic air. Photolysis of nitrous acid produced HO• radicals.

HONO + hν → HO• + NO

(8)

The photolysis experiments were made using a long-path glass chamber that was 6m long and the concentration changes were measured with a Fourier Transform Spectrometer [30,37]. The concentrations of CO were sufficiently high so that they did not appreciably change during the experiments so the relative extent of the HO•+CO reaction was taken as the total CO₂ produced during each experiment. The rates of the HO•+SO₂ and HO•+CO reactions are given below where k_SO2 and k_CO are rate coefficients for these reactions.

$${\displaystyle \frac{\left[{SO}_2\right]}{dt}}=k_{HO+SO2}\left[HO\ast \right]\left[{SO}_2\right]$$

(9)

$${\displaystyle \frac{\left[CO\right]}{dt}}=k_{HO+CO}\left[HO\ast \right]\left[CO\right]$$

(10)

Equations 9 and 10 may be rearranged to derive the ratio, k_HO+SO2/k_HO+CO. To derive Equation 11 for the experiments, note that the HO• concentrations cancel out, that the concentrations of SO₂ and CO may be averaged over the photolysis period ([SO₂]_AVG, [CO]_AVG) and that the change in CO is equal to the amount of CO₂ produced (Δ[CO₂]. The change in the SO₂ concentration is given by Δ[SO₂].

$${\displaystyle \frac{k_{HO+SO2}}{k_{HO+CO}}}={\displaystyle \frac{{\left[CO\right]}_{AVG}}{{\left[{SO}_2\right]}_{AVG}}}{\displaystyle \frac{∆\left[{SO}_2\right]}{∆\left[{CO}_2\right]}}$$

(11)

The gas-phase rate coefficients for the Reactions 1 and 8 are well known now. These were used to calculate the literature k_HO+SO2/k_HO+CO ratio. The rate coefficients for the reaction of HO• with SO₂ is pressure and temperature dependent [41]. This reaction involves two small molecules forming a reactive intermediate that requires its collision energy to be dissipated through collisions with a third bodies. The rate coefficient n units of cm⁶molecule^-1s^-1 is calculated through the following equations.

$$k_0\left(T\right)=k_{\mathrm{0,298}}{\left({\displaystyle \frac{T}{298}}\right)}^{-n}$$

(12)

$$k_{\infty }\left(T\right)=k_{\infty ,298}{\left({\displaystyle \frac{T}{298}}\right)}^{-m}$$

(13)

$$k_{HO+SO2}\left(T,\left[M\right]\right)=\left\{{\displaystyle \frac{k_{\infty }\left(T\right)k_0\left(T\right)\left[M\right]}{k_{\infty }\left(T\right)+k_0\left(T\right)\left[M\right]}}\right\}{0.6}^{{\left\{1+{\left[{log}_{10}\left({\displaystyle \frac{k_0\left(T\right)\left[M\right\}}{k_{\infty }\left(T\right)}}\right)\right]}^2\right\}}^{-1}}$$

(14)

where: $k_{\mathrm{0,298}}=$2.90×10^-31, n = 4.1, $k_{\infty ,298}=$1.70×10^-12 and m = -0.2; and [M] is the concentration of the total background gas in molecules cm^-3 [41].

The rate coefficient for the reaction of HO• with CO in units of cm⁶molecule^-1s^-1 is given is given by equation 15 [42,43]:

$$k_{HO+CO}=1.44\times {10}^{-13}\times \left(1.0+{\displaystyle \frac{0.8\times \left[N_2\right]}{4\times {10}^{+19}}}\right)$$

(15)

These were used to calculate literature ratio of the coefficients given in the Results section of this work.

3. Results

Table 3 shows experimental results for the photolysis experiments.

Table 3 shows the experimental k_HO+SO2/k_HO+CO ratios calculated using Equation 11 from the initial concentrations. The observed changes in SO₂ and CO₂ concentrations The ratios ranges from 9.1 to 14.9 for the experiments made with background N₂. The ratio ranges from 0.8 to 16.7 for the experiments made with background synthetic air.

Table 4 shows calculated k_HO+SO2/k_HO+CO ratios from the literature rate coefficients. The expected values of the ratios are between 4 and 4.5. The experimental values of the k_HO+SO2/k_HO+CO ratios are between 1.80 and 3.75 times greater than the literature calculated values.

Plots were made to examine the possibility that there was an effect of reactive nitrogen on the k_HO+SO2/k_HO+CO ratios. Figure 2 shows plots of the experimental k_HO+SO2/k_HO+CO ratios as functions of the initial reactive nitrogen oxides (HONO+NO+NO₂. Table 5 shows the average reactive nitrogen in the three sets of experiments, the average k_HO+SO2/k_HO+CO ratio and the shared variance between them.

4. Discussion

This new analysis of photolysis experiments made with relatively high concentrations of HONO, NO, NO₂, SO₂ and CO showed that there may be an uncharacterized oxidation process for SO₂. The literature value of the k_HO+SO2/k_HO+CO ratio for the experimental conditions is near 4 to 4.5 while the measured ratio is around three times greater for the N₂ only, synthetic air background gas experiments and both sets of experiments considered together.

The experimental ratio for the experiments with N₂ only as the background gas has a shard variance of 0.65 with the initial total reactive nitrogen concentration. However, the shard variance between the experimental ratio and the initial total reactive nitrogen concentration is much lower for the synthetic air case. Although there is much uncertainty, this could be due to an oxygen effect that is not taken into account in this analysis.

If there is a new oxidation process for SO₂ that is three times greater than the known hydroxyl radical reaction that occurs in NO_x polluted atmospheres then conversion rates in the range of 10 to 20% hr^-1 may occur in urban regions such as El Paso, Texas with sources of SO₂ and NO_x (using reference 33 as the baseline).

Sulfuric acid has been produced on a commercial scale from the direct reaction of SO₂ and NO_x and H₂O. These were mixed at very high concentrations in large chambers, made from lead, where reactions produced H₂SO₄. [44] Based on our experiments we show that similar reactions may occur at lower concentrations. These reactions are probably heterogeneous and appear to be faster than the gas-phase reactions. The reactions may be important for converting SO₂ to sulfate on urban scales and therefore contribute to local PM_2.5 formation. Improved representation of these processes may improve the agreement between measurements and air quality modeling results [35,36] for the United States, China and elsewhere. New research involving laboratory studies and field measurements are needed to better characterize the oxidation of SO₂.because it is possible that heterogeneous reactions occurred on sulfate particles as soon as they were formed ans/or other reactive nitrogen catalyzed processes occure. Therefore, better measurements of formation and properties of the sulfate particles and their surface reactions are needed.