This experiment used coal fly ash generated from coal-fired power plant without further treatment. The surface area of the powder sample was measured by using the multipoint Brunauer–Emmett–Teller (BET) analysis method with a BET apparatus (autosorb-iQ, Quantachrome Instruments, United States). The measured surface area was 1.796 m²g⁻¹. X-ray diffraction (XRD) experiment for this powder sample was performed via a Rigaku D/max-2500 diffractometer (Japan) with Cu–K radiation at 50 kV and 100 mA. Scans from 2.6° to 50° (2θ) at a rate of 1° (2θ) per minute were carried out. The obtained XRD pattern was processed using JADE software and is shown in Supplementary Figure S1. The major peaks were indexed to permit recognition of the coal fly ash, and the results showed that the main component of this coal fly ash was aluminosilicate (mullite and corundum), aluminum, and silicon were the main component elements, such as the results derived by Koukouzas et al. (2006); Skvara et al. (2009). Figure 1 displays the scanning electron micrograph (S-4800, Hitachi, Japan) of this coal fly ash. The particle size distribution of this coal fly ash was between 5 and 100 μm in the coarse particle mode.

The gaseous sulfur dioxide (99.9%, Beijing Huayuan Gas Chemical Industry Co., Ltd.,) was used as reactant gas without further purification. The high weight percentage hydrogen peroxide solution (∼95 wt%) was produced by rotary evaporating aqueous solutions of hydrogen peroxide (35 wt%, Alfa Aesar, China) and subsequently concentrated by bubbling dry nitrogen. The weight concentration of purified H₂O₂ was greater than 95% as described by Zhou et al. (2012)

The uptake coefficients of H₂O₂ and SO₂ on coal fly ash samples were carried out by using a Knudsen cell system combining with an electron ionization (EI) quadrupole mass spectrometer (QMS, Hiden, HAL 3F 501, United Kingdom), as illustrated in Figure 2. The experimental set up and the procedure used has been described in detail in previous publications (Wang et al., 2011; Zhou et al., 2012). In brief, the reactor of the Knudsen cell system consists of a chamber with four isolated sample compartments for parallel experiments and a small escape aperture linking to the QMS detector. To ensure that the mean free path of the molecules exceeded the dimensions of the Knudsen cell, the pressure was maintained low in the reaction chamber (∼10^–3 Pa). A vacuum gauge was applied to monitor the pressure in the chamber. Before the reactant gas was injected into the chamber, the pressure was kept stable and the pressure variation due to adding gas was used to calculate the partial pressure of reactant in the chamber. Prior to the experiments, the reactor chamber was passivated with reactant gases for more than 30 min until a steady state of the QMS signal was established, while the coal fly ash samples were isolated from the gas by the sample cover.

The EI-QMS was used to detect reactants and products of these heterogeneous reactions. The measured mass charge ratio (m/z) signals varied linearly with the concentrations of the compounds. The specific masses monitored to detect H₂O₂ and SO₂ on coal fly ash samples were m/z = 34 (H₂O₂ ⁺) and 64 (SO₂ ⁺), respectively. A blank experiment was performed first to make sure the sample holders and covers were passivated by the gas phase reactants passing through the reactor. No obvious reactants signals consume were observed during blank experiments. The powder samples with known masses were evenly tiled on the bottom of the holder by dispersing with deionized water and drying first. For the temperature effect experiments, a circulator was applied to adjust the temperature from 253 to 313 K for the Teflon-coated metal sample holders and for the chamber. During each experiment, the temperature variability was maintained within 1 K.

The Knudsen cell heterogeneous kinetics experiments were performed following our previous works (Zhou et al., 2014; Zhou et al., 2016). For all the kinetic experiments, the temperature of reaction chamber was kept at a stable value first, and the samples were kept isolated from the gas reagent. Then, the system was passivated with gas reactant until the signal of gas reactant reached a steady state. After that, one sample was exposed to gas reactant by quickly opening the holder cover. We found the signal of reactant parent ion monitored at m/z = 34 (H₂O₂ ⁺) or 64 (SO₂ ⁺) dropped below its original value suddenly. An observed uptake coefficient, γ_obs, can be derived from the following equation (Usher et al., 2002): γ obs = A h A s ( I o − I I ) . (1)

Here, A_h represents the effective area of the escape hole between the Knudsen cell chamber and the QMS detector; in this study, A_h is 0.211 mm²; A_s represents the geometric area of the sample holder, which is 5.3 cm² in our system; I₀ and I are the signal intensities detected by QMS when sample holder was covered and exposed. The parameters of this system could be found in our previous work (Zhou et al., 2016) and also summarized in Supplementary Table S1.

In this work, the measurements of initial and steady state uptake coefficients at 298 K for H₂O₂ and SO₂ on coal fly ash were conducted separately.

During the experiments, we found that H₂O₂ was first taken up on the surface of the fly ash, and after a certain period, the reactant signal (H₂O₂ ⁺, m/z = 34) became balanced and a steady state uptake process could be measured. The uptake phenomenon of SO₂ on the surface of coal fly ash was a slightly different from those of H₂O₂: only the initial uptake was obvious, while the steady state uptake was quite small (<8 × 10⁻⁵); this steady state uptake was similar to that of Chinese Inner Mongolia desert dust, which was <5 × 10⁻⁵ (Zhou et al., 2014). Besides each gas phase reactant, some possible products, including O₂ (m/z = 32), CO₂ (m/z = 44). were also monitored by QMS synchronously, while there was no remarkable undulation during the uptake experiments such as catalytic oxidation reactions.

According to Eq. 1, uptake coefficients were derived by assuming the gas-surface collision only on the top layer of the solid reactants. To test if all the surface area of coal fly ash samples was involved in the uptake processes, the uptake coefficients γ_obs were calculated using the geometric surface area of the sample versus a series of sample masses. As illustrated in Figure 3, the observed uptake values grew linearly with increasing sample mass, which indicated that reactions were not only on the top layer of the sample surface, but the total surface areas of particles could also participate in the reactions. The observed mass dependence phenomena supported that the diffusion of the reactants can go through the underlying layers of these samples. In this work, the experiments were carried out by using the sample masses in the linear range to make sure the entire surface areas of the coal fly ash were accessible to the reactants (H₂O₂ and SO₂). BET surface areas of each sample were treated as the entire sample participates during the reactions (Underwood et al., 2000).

The plots in Figure 3 show the region where the observed uptake coefficients (γ_obs) are linearly dependent on the mass of samples. From the plot, a mass independent uptake coefficient can be derived as follows: γ BET = A h A B E T ( I o − I I ) = γ obs A s A B E T . (2)

Here, A_BET represents the surface area of the coal fly ash powder sample, the BET area of the powder times the sample mass is taken to calculate this value. The γ_BET values obtained from each experiment are given in Table 1 (298 K).

Because the uptake processes on aerosol surfaces might be subjected to the pressure, namely, the mixing ratios of reactants, influence from H₂O₂/SO₂ mixing ratio to the uptake coefficient was considered in this work by varying the initial mixing ratio of H₂O₂/SO₂ from 10 to 60 ppbv, respectively. The uptake experiments of H₂O₂/SO₂ on coal fly ash showed similar phenomena. This means that there was limited dependence of initial uptake coefficient on different mixing ratios of H₂O₂ or SO₂ in this range.

The lack of dependence on mixing ratio observed in this study were similar with previous work conducted on mineral dust samples (El Zein et al., 2014; Zhou et al., 2016). Therefore, according to the results obtained from the experimental measurements, the H₂O₂ heterogeneous uptake on coal fly ash was assumed to be passing through physical adsorption and heterogeneous reactions, while SO₂ heterogeneous uptake on coal fly ash was through physical adsorption at room temperature.

In this study, we first measured the uptake coefficient of H₂O₂ on coal fly ash samples with a constant mixing ratio of H₂O₂ and SO₂ mixing ratio of 0–52.8 ppbv. The initial uptake coefficients of H₂O₂ showed little difference between measuring with only H₂O₂ and with both the reactive gases (H₂O₂ and SO₂) present (The relative deviation was less than 10%.). This phenomenon likely since the physisorption rate was much faster than the rate of chemical transformation. The main processes of initial uptake were only related to physisorption independent of the active sites for adsorption seemed to be independent. As illustrated in Figure 4, no apparent growth in steady state uptake coefficients of hydrogen peroxide on coal fly was shown with increasing of SO₂ mixing rations in the reaction chamber. It was expected that on these coal fly ash samples, SO₂ did not influent the reactive site for H₂O₂ uptake regularly. Further investigation would need to evaluate the effects of SO₂ on coal fly ash. We next measured the uptake coefficient of SO₂ on coal fly ash samples when the mixing ratio of SO₂ was kept constant and the H₂O₂ mixing ratio ranged from 0 ppbv to around 40 ppbv. The initial uptake coefficients of SO₂ were independent of the H₂O₂ mixing ratio. As mentioned above, this process was also dominated by the physisorption. However, hydrogen peroxide can easily increase the measured steady state uptake coefficients of SO₂, as shown in Figure 5. It was implied that the absorbed hydrogen peroxide can provide a new reactive surface for SO₂ uptake.

In this study, although Knudsen cell reactor was kept in a low pressure, the experiments were performed in an almost dry condition, there would still be some adsorbed water present on the surface of the coal fly ash samples. The adsorbed water present on the surface could be in the form of O^2- or OH as water-related-reactive sites, but could not been quantified. The sulfate or bisulfate formation on the coal fly ash should be consistent with the reaction of gas-phase SO₂ with reactive surface (Usher et al., 2002). S O 2 ( g ) + O 2 − ( s ) → S O 3 2 − S O 2 ( g ) + O H − ( s ) → HS O 3 −

In previous study, the oxidation of SO₂ through uptake process was proposed a two steps mechanism. The first step was a reversible adsorption of SO₂ onto the surface, and the followed step was an irreversible reaction by which adsorbed SO₂ was oxidized to be sulfate (Ullerstam et al., 2002). So, the H₂O₂ may have the reactivity to oxidize adsorbed SO₂ on the surface of coal fly ash samples via the following reactions: H 2 O 2 ( g ) + S → S [ OH ] ( S + + O H − ) + OH, S [ OH ] ( S + + O H − ) + S O 2 ( g ) + → S [ OH ] − S O 2 ( S + + H S O 3 − ) , S [ OH ] − S O 2 ( S + + H S O 3 − ) + H O 2 → S [ H S O 4 ] ( S + + H S O 4 − ) + H 2 O .

Our result suggested that H₂O₂ uptake affects mineral surface-active site and promotes the further reactions.

The temperature dependence of the uptake coefficients for heterogeneous reactions on coal fly ash was investigated for the temperature ranging from 253K to 313 K with this temperature-controllable Knudsen cell system. The results are presented in Table 1.

The experimental results showed that the initial uptake coefficients of gas phase H₂O₂ and SO₂ on coal fly ash decreased evidently with increasing temperature. This trend of the initial uptake coefficients with temperature well match with the physisorption principle (Zhou et al., 2012). A similar temperature effect of initial uptake process had also been derived on some real mineral dust and typical mineral oxides (Zhou et al., 2012; Romanias et al., 2013; El Zein et al., 2014; Zhou et al., 2016). The steady state uptake coefficients of H₂O₂ on coal fly ash samples performed contrary phenomena. The γ_ss,BET values increased with the temperature increasing. At 253 K, the uptake phenomenon was too weak and generated large error for the steady state uptake coefficient calculation, so this value was not given in this work. The main component of coal fly ash was aluminosilicate. In our previous work, the interactions of H₂O₂ with Arizona test dust, Inner Mongolia desert dust and Xinjiang sierozem over the temperature range from 268 to 320 K have been investigated (Zhou et al., 2016). The certified chemical compositions of these mineral dust were also SiO₂ and Al₂O₃, but the temperature effects of their steady state uptake coefficients on mineral dust samples showed different phenomena, which may relate to the differences of their trace species. Therefore, the temperature dependence of steady state uptake coefficients indicating that the structure and the component of the solid samples may impact the uptake processes.

Now that the initial uptake coefficients showed negative correlation to the temperature, the changes of enthalpy (ΔH _obs) and entropy (ΔS _obs) for H₂O₂ and SO₂ adsorption on coal fly ash and the reactive energy (Ea) for H₂O₂ uptake on coal fly ash could be derived from the following Eqs 3, 4 (Jayne et al., 1990): ln ( γ init,BET 1 − γ init,BET ) = − Δ H obs RT + Δ S obs R , (3) ln ( γ ss,BET 1 − γ ss,BET ) = − Ea RT + A . (4)

From the plot on the left side of Eq. 3 versus inverse temperature, as shown in Figures 6, 7, the enthalpy (ΔH_obs) and entropy (ΔS_obs) were determined to be −(13.27 ± 1.62) KJ mol⁻¹ and −(124.0 ± 5.8) J K mol⁻¹ for H₂O₂ initial uptake process; −(19.57 ± 1.56) KJ mol⁻¹ and −(154.3 ± 5.6) J K mol⁻¹ for SO₂ initial uptake process. The reactive energy (Ea) for H₂O₂ uptake on coal fly ash was (11.01 ± 1.36) KJ mol⁻¹. Normally, the activation energy of a reaction exit in the atmosphere is larger than 20 KJ mol⁻¹, the importance of this reaction is negligible (Smith, 2003; Liu et al., 2008). This reveals that the heterogeneous uptake processes of H₂O₂ on coal fly ash are significant and should not be neglected.

The mechanism of H₂O₂ uptake on an oxide surface at low relative humidity (RH) condition are presented as below (Lin and Gurol, 1998; Pradhan et al., 2010a; Zhao et al., 2011). Herein, S represents the surface of solid sample. H 2 O 2 ( g ) + S → S [ OH ] ( S + + O H − ) + OH . OH . + H 2 O 2 + → H 2 O + H O 2 . H O 2 . + H O 2 . → H 2 O 2 + O 2

Based on the study results, we could suggest that the initial uptake coefficients of H₂O₂ and SO₂ on coal fly ash are analogous to the other mineral aerosols, which can provide active surface for sinking of H₂O₂ and SO₂, especially at low temperature.