| Home | E-Submission | Sitemap | Contact Us |  
Environ Eng Res > Volume 28(6); 2022 > Article
Wen, Shi, Wang, and Huang: Reacting flow field simulation of Hg0 oxidation in flue gas by ozone injection


Ozone injection is an effective way to oxidate Hg0 in flue gas. The reacting flow field for the oxidation of Hg0 by O3 is numerically studied in detail with a reduced chemical mechanism in this paper. Based on sensitivity analysis with detailed chemical mechanism, a 12-steps reduced chemical mechanism is obtained and used in the following three-dimensional simulations of the reacting flow field. Through the reacting flow field simulation, the influence of various factors on Hg0 oxidation efficiency is calculated and analyzed, such as O3/NO molar ratios, jet velocities, moisture, temperatures, NO concentrations and Hg0 concentrations. Results show that temperature and O3/NO molar ratio are the key factors, jet velocity and NO concentration are the important factors, while water content has little effect. It is also found that the oxidation efficiency of Hg0 is shown to be highly correlated with the amount of NO3 generated in the flue gas. Increasing the local generation of NO3 may be the key to achieve efficient oxidation of Hg0. This study is good help of subsequent engineering applications.

Graphical Abstract

1. Introduction

Coal combustion produces not only SO2, NOX, and CO2, but also various forms of mercury [1]. Mercury exists in flue gases as elemental mercury (Hg0), Oxidized mercury (Hg2+) and particulate mercury (Hgp) [2]. Although the form of mercury varies according to combustion conditions, absorption and oxidation are still important methods for improving the efficiency of mercury removal [3]. The particulate mercury (Hgp) in the flue gas can be removed by industrial dust collectors with other dust. Hg2+ has low volatility and can be adsorbed by fly ash to form particulate mercury, so it can be efficiently removed by WFGD. Hg0 is the most difficult to control as it is volatile and has low water solubility [4]. At the same time, Hg0 can remain in the atmosphere for long periods and contribute to the global spread of mercury with the atmospheric circulation. The oxidation of Hg0 to Hg2+ is the core of mercury removal technology [5]. Researches show that Cl2, O3, H2O2 can effectively oxidize Hg0. However, Cl2 [6] is difficult to produce and preserve and H2O2 [7] has a short oxidation life. Ozone (O3) differs from other oxidants in that it has a long oxidation life at ordinary flue gas temperature. O3 still has good oxidation capacity for mercury at 573 k [8]. Due to the problems of equipment maintenance and high operation cost in a single pollutant removal technology, a variety of pollutants removal technology has gradually become a development trend. As a strong oxidant, O3 can quickly oxidize NO and Hg0 to high-valence [9,10]. The gas-phase oxidation method combined with other methods [8] can effectively remove a variety of pollutants from flue gas. In the flue gas with NO and Hg0, NO is in the dominant position and O3 will be preferred to NO reaction. And thus, Hg0 can be also oxidized by high-valence nitrogen oxides, promoting Hg0 removal [11,12]. In our previous work, the activation energy and kinetic parameters are calculated and analyzed by quantum chemical method and reliable data were determined. Through the combination of kinetic simulation and experiment, some preliminary conclusions on the reaction mechanism are obtained. However, the simulation and experiment are focused on mechanism research and limited to small reactors. The previous work is quite far from the actual industrial application. Therefore, based on the previous mechanism research work, this paper further simulates the reaction flow field in a large scale reaction pipeline in order to provide guidance for the practical application of the project. It is expected to find the most suitable conditions for Hg0 removal from flue gas in industrial application. The above added content has been added to the paper.
In order to provide data reference for engineering application, the reacting flow field for the oxidation of Hg0 by O3 is numerically studied in detail with a reduced chemical mechanism in this paper. Based on sensitivity analysis with detailed chemical mechanism, a reduced chemical mechanism is analyzed and obtained. And then, this reduced chemical mechanism is used in the following three-dimensional simulations of the reacting flow field. Through the reacting flow field simulation, the influence of various factors on Hg0 oxidation efficiency is calculated and analyzed, such as O3/NO molar ratios, jet velocities, moisture, temperatures, NO concentrations and Hg0 concentrations. Furthermore, the better working conditions are analyzed and obtained. The present study is good help of subsequent engineering applications.

2. Numerical Method

2.1. Chemical Mechanism

Tubular flow reactors have been utilized for many years in the chemical process industries. Traditional uses have included both homogeneous (in an empty tube) and fluid-solid heterogeneous (in packed beds) reactions. The plug-flow reactor [13] is an idealized model with decreased dimensionality in the case of tube flow. The plug-flow reactor is assumed that no mixing occurs in the axial (flow) direction, but that perfect mixing occurs in the direction transverse to this. Because there is no axial mixing, the possible reactant conversion can be maximized. Similarly, the absence of transverse gradients indicates that mass-transfer limits are absent, which improves reactor performance once again. In addition to these practical advantages, the plug flow reactor is computationally efficient because it is modeled using first-order ordinary differential equations and does not require transport characteristics. PLUG model has sufficient gas mixing and can be well used for sensitivity analysis in this paper. The chemical kinetic studies here consider gas-phase processes, and following equations are solved (the symbols and their meanings as used in the equations are given at the end of the essay):
ρuS [gasnihndYndx+CpdTdx]=aiQi
where gas inlet parameter setting and component setting: ϕ (O2) =5%, ϕ (NO)=0.01%, ϕ (H2O)=1%, ϕ (Hg)=0.0001%, ϕ (O3)=0.15%, equilibrium gas is N2, temperature is taken as 150°C.

2.2. Flow Simulation

In practical industrial applications [14,15], ozone is generated by the ozone generator (with air or oxygen as the gas source), and the ozone usage is controlled by adjusting the gas flow. The numerical simulation in this paper is based on the practical industrial application. The model is constructed as a cylindrical pipe with a length of 10 m and a diameter of 2 m. In order to make a complete mixing of O3 and the flue gas, 19 small circular ports with 0.1 m diameter are set as O3 inlets, while the remaining part is the flue gas inlet.
To improve computational efficiency, the whole cylindrical pipe is quartered. The cut surface is set as a symmetrical structure. This does not affect the overall calculation. Mesh is drawn using an O-block structured mesh, which is a hexahedral mesh. After building the block, set the overall mesh size to generate the model’s mesh. The final mesh show in Fig. S1.

2.2.1. Mesh quality inspection

Checking the quality and quantity of the model mesh is the key factor to support the simulation results [16,17]. The total number of meshes in this model is 301,694. Minimum mesh quality for this model is 0.415, the optimum mesh quality is 1, and approximately 80% of the mesh quality is greater than 0.8, with an average quality of 0.897.
The original grid is encrypted. As shown in Fig. S2, the oxidation efficiency is a large difference at the initial stage. However, the difference in oxidation efficiency calculated by the two models is not significant as the observation distance increases. The observation surface is chosen the one at 10 m. The initial observation surface is not considered. To improve the efficiency of the calculation, we chose the number of 30w model as the calculation model. The number of meshes can support the following simulation.

2.2.2. Solver setups

FLUENT model is chosen as transient, with a time step of 0.01 s and 1000 time-steps. The effect of gravitational acceleration is ignored. The energy equation is switched on to include the chemical reactions. The turbulence model is the realizable k-epsilon model, which has the advantage over the standard k-epsilon model of keeping the Reynolds stress to be consistent with real turbulence. The component transport model is used. The finite-rate chemistry model and well-stirred reactor model are adopted with first-order closure hypothesis of the turbulence-chemistry interactions. The jet the flue gas inlets are set to have constant inlet velocity, while the outlet is the atmospheric environment. Pipe wall surface is selected as no-slip wall surface, the inter is selected as symmetry surface. Standard initialization is used.
The concentration of ozone is calculated from the concentration ratio, the jet velocity and the incident area using the following equations.
where Vgas is the flue gas incidence velocity, Sgas is the surface area of the flue gas jet, Volno is the volume fraction of NO, nO3/NO is the concentration ratio of O3 to NO, SO3 is the surface area of the ozone jet and VO3 is the O3 jet velocity.

2.2.3. Observation surface setup

The observation surface is a cross-section at 1 m intervals along the height of the pipe. Observation surface is used to monitor the efficiency and uniformity of the flue gas as the distance increases.
where Ef represents the Hg0 oxidation efficiency in percentage, Cin represents the Hg0 concentration in ppm at the flue gas inlet, and Cout represents the Hg0 concentration in ppm at the flue gas outlet.
Uniformity calculation is to set the observation surface as standard deviation during the calculation setting. This observation surface calculates square root of oxidation efficiency variance, which represents the degree of uniformity. The standard deviation of a specified field variable on a surface is computed using the mathematical expression below:
where x is the cell value of the selected variables at each facet, is the mean of x.

3. Results and Discussion

3.1. Sensitivity Analysis

Sensitivity analysis [1820] of a problem solution allows quantitative understanding of how the solution depends on the various parameters contained in a model. CHEMKIN software is used to calculate sensitivity coefficient. Parameter setting refers the experiments in the literature [21]. The moment when Hg0 is oxidated 50% (the moment when Hg0 oxidation is the fastest) is chosen to see the sensitivity coefficient of each reaction.
As shown in Fig. 1 and Fig. S3, the sensitivity analysis identifies four dominated reactions during Hg0 oxidation: R61 (NO2+ NO3=N2O5); R70 (NO3+Hg=HgO+NO2); R8 (O3+NO2=O2+NO3); R7 (O3+NO=NO2+O2). R61 and R8 have a negative impact on Hg0 oxidation. However, R70 and R7 have a positive impact on Hg0 oxidation. R8 directly generates NO3 in the form of free radical [2224] and the presence of NO3 directly oxidizes Hg0, therefore it has a negative effect. R7 is the reaction between O3 and NO to produce NO2, which competes with R8 for the oxidation of NO2 by O3, influencing the volume fraction of NO3. It plays a positive role in the oxidation of Hg0. R61 is the reversible reaction between NO2 and NO3 to produce N2O5, which also influences the volume fraction of NO3.
As shown in Fig. S4, the only way for Hg0 to generate O2 is to react with O3 and the arrows for O2 production from Hg0 are extremely thin. It indicates that the reaction between Hg0 and O3 in flue gas dehumidification is likewise extremely weak. To summarize, O3 flue gas oxidizing Hg0 works indirectly through the reaction of O3 and NO to form NO3.

3.1.1. Sensitivity analysis at different temperatures and O3/NO molar ratios

Fig. S5 shows the sensitivity analysis of Hg0 oxidation at different temperatures. R70 (NO3+Hg=HgO+NO2) directly determines the oxidation efficiency of Hg0, and its sensitivity coefficient increases with the increment of temperature. The production of N2O5 is affected by temperature, which is also shown in literatures [6]. The reaction between NO2 and NO3 form N2O5 is easier at a lower temperature. This depletes that the oxidation of Hg0 will be affected at a lower temperature. At 523 k, the sensitivity coefficient of R61 (NO2+NO3=N2O5) is almost zero, which means that negative effect on oxidizing Hg0 is negligible. We discover an additional R10 (O3+O=2O2) at 473 k and 523 k compared to 423 k and 373 k in the temperature sensitivity analysis, which is the self-decomposition of O3 [25]. Ozone will self-decompose at higher temperatures. This affects the production of NO3 gas, so it plays a positive role in Hg0 oxidation.
Fig. S6 shows the sensitivity analysis of Hg0 oxidation at different O3/NO ratios. The higher O3 concentration leads to a higher reaction rate of R8 (O3+NO2=O2+NO3). As a result, the oxidation of Hg0 by NO3 is intensified. Therefore, the sensitivity coefficient of R70 (NO3+Hg=HgO+NO2) increases with the increase of concentration. R7 (O3+NO=NO2+O2) is positive for O3/NO ratios of 1.2 and 1.4. Ratio of 1.2 sensitivity coefficient is larger than that of 1.4. In the case of low concentration of O3, R7 and R8 have weak competition. Therefore, the sensitivity coefficient of R7 is positive. With the increase of O3 concentration, this competition relationship gradually decreases.

3.2. Mechanisms Simplified

3.2.1. Mechanism simplification methods and steps

The detailed mechanism [26] adopted in the present study includes 70 reactions (Table S1). The kinetic parameters for the elementary reactions are derived from the NIST database and selected references. The mechanism simplification methods include direct relationship graph method [2728] and sensitivity analysis method [29], etc. In this paper, the reaction with small sensitivity coefficient is deleted. The target conditions are temperatures of 373 k to 523 k, NO concentrations of 100 ppm to 300 ppm, Hg0 concentrations of 1 ppm to 100 ppm and O3/NO molar ratios of 1 to 1.6. The target parameters are set to Hg0 molar fraction, NO molar fraction, and NO3 molar fraction. 10% error threshold is used. A reduced mechanism file for 15 species and 12 reactions is obtained, as shown in Table S2.

3.2.2. Validation of the reduced mechanism

Comparison of reduced and detailed chemical mechanisms on the temporal evolution of Hg0 molar fraction is shown in Fig. 2 [30][32]. There is no significant difference between the reduced and detailed chemical mechanisms. Therefore, in the following studies, we will use the reduced mechanism to increase the computational efficiency.

3.3. FLUENT Simulation Results

3.3.1. Pipeline pollutant distribution

Based on the sensitivity analysis, due to the presence of NO in the flue gas, the main component of oxidized Hg0 is NO3. NO2 will continue to be oxidized to NO3 when O3 is excessive. Therefore, the initial molar ratio of O3/NO is chosen as 1.6 for this simulation. In the actual flue gas, the volume fraction of NO is much greater than the volume fraction of Hg0, so the molar ratio of O3 and Hg0 is ignored in this simulation. The simulation inlet gas parameters are: The flue gas inlet velocity is 3 m/s, the temperature is 423 k, and the input composition is as follows: O2 volume fraction is 5%, water volume fraction is 1%, NO volume fraction is 0.1%, and Hg0 volume fraction is 1 ppm. The jet inlet has a velocity of 10 m/s and a temperature of 423 k. The input gas has a volume fraction of 21 percent O2, and the volume fraction of O3 is calculated using the total surface area of the jet inlet, the jet velocity, and the molar ratio to NO, with N2 as the equilibrium gas. Hg0 degradation efficiency is calculated.
Fig. 3 (a) shows that as the simulation time approaches 10 seconds. Hg0 is uniformly distributed along the interior and Hg0 concentration no longer varies greatly with time. At this time, the efficiency of O3 oxidation of Hg0 can be calculated by observing the Hg0 concentration at the cross-section of the pipe. As shown in Fig. 3 (b), it can be found that the concentration of Hg0 decreases with the distance increases. Hg0 distribution becomes more and more uniform with the increase of pipe distance. O3 distribution and concentration of the pipeline profile is shown in Fig. 3 (c). As the distance inside the pipeline increases, the gas mixes more uniformly and the O3 volume fraction is lower.

3.3.2. Effect of different factors on oxidation efficiency

Based on the preliminary insights and calculation presented above, the three-dimensional simulations for the reacting flow field of Hg0 oxidation by O3 are performed. the influence of various factors on Hg0 oxidation efficiency is calculated and analyzed, such as O3/NO molar ratios, jet velocities, moisture, temperatures, NO concentrations and Hg0 concentrations. Temperature effect on Hg0 oxidation

The parameter settings are shown in section 3.3.1. Temperature varies from 373 k to 523 k. The simulation time is 10 seconds. As shown in Fig. 4, the oxidation efficiency increases gradually with the increase of temperature. NO3 plays a crucial role in the oxidation of Hg0. Increasing temperature will promote the reaction between NO3 and Hg0, therefore increasing the oxidation efficiency of Hg0. Temperature rises from 373 k to 423 k and the oxidation efficiency rises from 69% to 80%. As shown in Fig. S7, the largest increase in the forward reaction rate of R70 (The elementary reaction that directly affects the oxidation efficiency) occurs when the temperature is increased from 373 k to 423 k. After 423 k, the increase of R70 (HgO+NO3=NO2+HgO) reaction rate tends to be flat. In order to more clearly show the effect of temperature on the reaction rate of R70, the maximum reaction rate with the oxidation efficiency is compared. As shown in Fig. 5, the acceleration of reaction rate and the improvement of oxidation efficiency are synchronous, and the rapid increase of reaction rate brings about the rapid improvement of oxidation efficiency. The forward reaction rate of R70 increases most in the range from 373 k to 423 k and the oxidation efficiency increases significantly (69%–80%). Therefore, the oxidation efficiency increases the most in the temperature range from 373 k to 423 k. Whereas temperature increase from 473 k to 523 k only increased oxidation efficiency by about 3%, which is not a substantial increase. Combined with the above sensitivity analysis, although the reaction between NO3 and Hg0 is weaker at 523 k than that at 473 k, the reaction between NO2 and NO3 to form N2O5 is inhibited at 523 k. Therefore, the oxidation efficiency will be higher at 523 k than that at 473 k. But R70 reaction is limited by temperature, the oxidation efficiency increase is not significant. Taking into accounting the clustering effect of the jet duct, the oxidation efficiency in duct from 1–3m decreases and then gradually increases. O3 entering pipeline first reacts with NOx to produce enough NO3. Hg0 has not been oxidized and accumulated. Then as the jet gas continues to rush in, NO3 becomes more concentrated in the middle and rear part of the pipeline and reacts with Hg0. Therefore, in the rear part of the pipeline, the oxidation efficiency increases. O3 is injected into the cylindrical pipeline as a jet and is not thoroughly mixed with the surrounding gases. The cross-sectional view in Fig. 3 also clearly shows that, even at 10 m in the pipe, the Hg0 does not spread equally to every area of the cross-section. Oxidation efficiency is constrained by gas diffusion. The uniformity of Hg0 oxidation in each cross section at different temperatures is shown in Fig. S8, with smaller values indicating more uniform oxidation. The effect of varied O3/NO molar ratios on Hg0 oxidation efficiency

The parameter settings are shown in section 3.3.1. O3/NO molar ratios varies from 1 to 2. As shown in Fig. 6, the oxidation efficiency of Hg0 increases as the molar ratio increases. The higher molar ratio, the faster increase in oxidation efficiency. The increase of O3/NO molar ratio not only promotes the formation of NO3, but also promotes the reaction between NO3 and Hg0. Therefore, the oxidation efficiency is obviously increased. When the O3/NO molar ratio at 1.0, the oxidation efficiency of Hg0 increases very slowly with increasing measurement distance. Because O3 preferentially reacts with NO in flue gas. At the molar ratio of 1.0, there is little excess O3. This results in the generation of NO3 is particularly difficult and slow, affecting the oxidation efficiency of Hg0 further. Whereas increasing the molar ratio to 1.2 boosts the oxidation efficiency by about 18%. The excess O3 facilitates the reaction between O3 and NO2 to produce NO3. Increasing the molar ratio from 1.2 to 2.0 will continue to improve oxidation efficiency, but the increase will gradually slow down. When the molar ratio is increased from 1.8 to 2, the oxidation efficiency is only raised by 5%. At a molar ratio of 1.8, the pipeline already has the conditions to produce sufficient NO3. Fig. S9 shows that at any molar ratio, the uniformity increases swiftly and remains stable in the middle halves of the pipeline, likely because the Hg0 oxidation reaction is fast and can be accomplished in the first half of the pipe. The effect of jet velocity on Hg0 oxidation

The parameter settings are shown in section 3.3.1. Jet velocity varies from 3 m/s to 10 m/s. As shown in Fig. S10, the effect of jet velocity on Hg0 oxidation efficiency. The oxidation efficiency increases with the increment of jet velocity, and the oxidation efficiency is about 80% at 10 m/s. When jet velocity is 3 m/s, the difference with other jet velocities is too large. It should be noted that the velocity of flue gas is also 3 m/s. Therefore, the mixing and residence time maybe insufficient to complete all oxidation within 10 seconds. The increase of oxidation efficiency is still noticeable when the jet velocity is increased from 5 m/s to 7 m/s. While further increment of the jet velocity from 7 m/s results in a quite small increment of the oxidation efficiency. When the velocity reaches a certain level, the gas diffusion is less affected by the velocity. In Fig. S11, it is clear shown that the uniformity of the jet velocities above 7 m/s eventually reaches a high level. When the low velocity case is excluded, the effect of jet velocity on oxidation efficiency is relatively minor compared with that of the temperature. The effect of Hg0 concentration on Hg0 oxidation

The parameter settings are shown in section 3.3.1. Hg0 volume fraction varies from 1 ppm to 50 ppm. As shown in Fig. S12, the effect of Hg0 concentrations on oxidation efficiency is not obvious before the first 4 m of the pipeline. After 4 m of the pipeline, the effect of Hg0 concentration on oxidation efficiency gradually appears. The oxidation efficiency gradually decreases with the increase of Hg0 concentration. Because NO3 can’t fully diffuse before the pipeline 4 m. Hg0 in the pipeline can only react with the NO3 at the edge of the jet gas. With the increase of the distance, NO3 in the jet quickly diffuses towards the edge of the pipeline and oxidizes Hg0. As shown in Fig. S13, the uniformity of each Hg0 concentration decrease rapidly. It is worthy noted that the oxidation efficiency remained above 70% even at a high Hg0 concentration of 50 ppm. Practically, Hg0 does not reach the concentration of 50 ppm in the industrial flue gas. And so, a molar ratio of 1.6 is sufficient to oxidize the most of Hg0 present in the actual flue gas. The effect of moisture in the flue gas on Hg0 oxidation

The parameter settings are shown in section 3.3.1. Water volume fraction varies from 1% to 20%. As shown in Fig. S14, the moisture has marginal effect on the final oxidation efficiency. Effect of moisture on oxidation efficiency is overall negligible. Previous sensitivity analysis and the reaction path analysis have shown that the oxidation of Hg0 mainly depends on the reaction with NO3 instead of H2O2, which is generated by H2O/O3 reaction. In the second half of the pipeline, most of NO3 and Hg0 react completely, which makes the oxidation of H2O2 to Hg0 gradually manifest. Uniformity of Hg0 oxidation under different moisture conditions as Shown in Fig. S15. The effect of NO concentration in flue gas on Hg0 oxidation

The parameter settings are shown in section 3.3.1. NO volume fraction varies from 50 ppm to 250 ppm. Fig. S16 shows that higher NO concentration in the flue gas results in higher Hg0 oxidation efficiency and the faster the rate of improvement of the oxidation efficiency. When the volume fraction of NO reaches 150 ppm, the oxidation efficiency of Hg0 at 1–3m will no longer decrease. Because high concentration of NO accelerates the formation reaction of NO3, Hg0 oxidation can be carried out faster. Higher NO concentration leads to higher NO3 concentration when the O3/NO molar ratio remains constant. Therefore, Hg0 in the flue gas can be oxidized more easily. The impact of NO concentration on the degree of oxidation uniformity is illustrated in Fig. S17.

3.4. Discussion On Key Influencing Factors

According to the above calculation and discussions, O3/NO molar ratio and temperature have the dominated effect on the Hg0 oxidation. The influence of these two key factors is analyzed and discussed in detail in the following text.

3.4.1. Different O3/NO molar ratios

As shown in Fig. 7, NO3 concentration increases with increasing O3/NO molar ratio. Whereas NO2 concentration increases and then decreases with increasing O3/NO molar ratio, peaking at O3/NO=1.4. According to the reaction mechanism, there are two ways to generate NO2: the reaction between O3 and NO and the reaction between NO3 and Hg0. Before the O3/NO molar ratio reaches 1.4, NO2 concentration rapidly increases. Combined with the trend of slowly increasing NO3 concentration in the graph, it is easy to see that the increase of O3/NO molar ratio causes the oxidation of NO2 by O3 in flue gas. NO3 further oxidises Hg0 to produce NO2. It results NO2 content increasing rapidly before the O3/NO molar ratio reaching 1.4. This stage is also the time when Hg0 oxidation efficiency rises rapidly and NO3 at the section remains at a low concentration. When the molar ratio is 1.6, NO2 doesn’t increase, but slightly decrease compared with 1.4. At the same time, NO3 significantly increases. Excessive O3 reacts again with the NO2 in the flue gas. It reduces NO2 in the flue gas while significantly increasing the remaining NO3 concentration at the 10m cross section. After a molar ratio of 1.6, NO2 decreases significantly while NO3 increases dramatically. Combined with Fig. 9 (a), we can see that the oxidation efficiency of Hg0 is slowly enhanced at this stage, and the concentration of O3 no longer has any effect on the elementary reaction of NO2+O3=NO3+O2.

3.4.2. Different temperatures

The previous simulation results show that at an O3/NO=1.6 molar ratio, the efficiency of Hg0 oxidation increases with increasing temperature. As shown in Fig. S18, the NO3 concentration at the 10 m cross section increases with temperature. When combined with the NO3 sensitivity analysis in Fig. S19, we can easily find that R7 which is a reaction to generate NO3 directly will increase with the temperature sensitivity coefficient. It is easier for O3 to oxidize NO2 to NO3 with the temperature increasing. At O3/NO=1.6 and temperature range 423–523 k, the reaction of R70 which affects directly the efficiency of Hg0 oxidation is more likely to occur. Intuitively, increasing temperature leads to improve oxidation efficiency. Although the R70 sensitivity coefficient is greater at 373 k than at 423 k, the NO3 generation reaction is less intense at this temperature, resulting in a lower oxidation efficiency at 373 k than at 423 k.

3.4.3. Comparison between simulation and experiment

Our previous experiments [21] used a 950 mm long glass tube bolus flow reactor with an internal diameter of 10.5 mm. Gas flows rate of 1 L/min and a gas residence time of approximately 3.5 s. The numerical simulations use a 10 m long 1/4 cylindrical pipe with a flue gas jet velocity of 3 m/s and an O3 injection velocity of 10 m/s, and a flue gas residence time in the pipe of about 3.3 s.
The comparison of the simulation data with the literature experimental data under the same conditions is illustrated in Fig. S20 and Fig. S21. As shown, the simulated oxidation efficiency is larger than the experimental oxidation efficiency. The reason could lie in the following two points: (1) although the experimental preheating of the reaction gases takes time, the total gas residence time and final constant temperature reaction time in the glass tube bolus flow reactor are both very short. The gases can’t fully react with each other before leaving the reactor. In contrast, there is no preheating in the numerical simulations. The reaction can be performed directly at setting temperature. (2) The simulation is in a 10 m length of 1/4 cylindrical pipe with multiple round holes for O3 injection. There is a velocity difference between the O3 jet velocity and the flue gas jet velocity. O3 jet is faster than the flue gas. It makes flue gas diffusion into the O3 jet easily. The effect of different O3 jet velocities on the oxidation efficiency is simulated in the Section 3.3.2. We have shown that the oxidation efficiency of Hg0 is at low level when there is no velocity difference between the O3 jet velocity and flue gas jet velocity, c.f. Section 3.3.2.
Although the results of numerical simulations do not match well with the experimental data, the trend of the two factors on the oxidation efficiency is reproduced in numerical studies. Increasing the O3/NO molar ratio and the temperature both have a significant increase in the oxidation efficiency. Furthermore, as shown in Fig. S20, the increment of Hg0 oxidation efficiency due to temperature increase can be quite similar between the numerical simulation and the experiment.

3.5. Discussion Of Practical Application

Based on the relevant research and practical engineering of the removal of various pollutants by ozone [15], the corresponding models and process parameters are given in the key part of ozone reactor in this paper. It can be seen from the simulation results that the higher the O3/NO concentration ratio, the better the oxidation efficiency. However, considering the economy, 1.6 is selected as the best concentration ratio. In practical industry, ozone is produced by ozone generator [31] because it is easy to decompose. The incident temperature of ozone should be lower than 473 k. The reaction temperature of the overall flue should be controlled within the temperature range of 423–523 k. The longitudinal ozone injection is changed from the pipeline to the horizontal ozone injection. The volume fraction of ozone pipeline is controlled by orifice flow meter. The ozone nozzle in the flue is connected in series by the pipe. Flue gas injects from the rest of the pipe. It can be seen from the simulation results that when there is a speed difference between flue gas and ozone, the Hg0 oxidation efficiency is significantly improved. The nozzle of ozone pipe can ensure that there is a certain speed difference with flue gas. Our simulation mainly focuses on the flue gas removal in the second half of the pipeline (Ozone reactor). The design scheme of flue gas pipeline is shown in Fig. S22 and Fig. S23. The design scheme is consistent with the reaction flow field simulation in this paper.

4. Conclusion

Based on the sensitivity analysis, four key element reactions for Hg0 oxidation by O3 are found: R7 (O3+NO=NO2+O2); R8 (O3+NO2=O2+NO3); R61 (NO2+NO3=N2O5); R70 (NO3+Hg= HgO+NO2). For Hg0 oxidation, R7 and R70 have a positive impact while R8 and R61 have a negative impact. The reacting flow field simulation shows that, temperature and O3/NO molar ratio are the key factors, jet velocity and NO concentration are the important factors, while water content has little effect. Moreover, the oxidation efficiency of Hg0 is shown to be highly correlated with the amount of NO3 generated in the flue gas. The better working conditions for Hg0 oxidation by O3 in flue gas are obtained: temperature is 473 k, O3/NO molar ratio is 1.6, jet velocity is 10 m/s, flue gas velocity is 3 m/s. This result would guide the subsequent engineering applications.


This research was supported Zhejiang Provincial key research and development program (Grant No. 2020C03084) and Zhejiang Provincial Natural Science Foundation of China (Grant No. LY19E060002). We thank Dr. Shenghui Zhong for the fruitful and constructive discussions.


Symbol Meaning


Internal cross-sectional area per unit length along the flow direction in a reactive flow


The Hg0 concentration in ppm at the flue gas inlet


The Hg0 concentration in ppm at the flue gas outlet


Mean specific heat capacity of a gas


The Hg0 oxidation efficiency in percentage


Frictional resistance between the reactor wall and gas


Friction coefficient


Specific enthalpy of component n


Absolute pressure of a gas


Heat transfer between the system wall and gas


Gas constant for an ideal gas


Cross-sectional area in the flow direction


Surface area of the flue gas jet


The surface area of the ozone jet


Absolute temperature of a gas


Axial gas velocity


Flue gas incidence velocity


The O3 jet velocity


Volume fraction of NO


Mean molar mass of a gas


Molar amount of component n


The cell value of the selected variables at each facet


The mean of x




Molar production rate of component n for chemical reaction n

Supplementary Information


Conflict-of-Interest Statement

The authors declare that they have no conflict of interest.

Author Contributions

W.Z.C. (Associate Professor) conducted simulation supervision, wrote and revised the manuscript. S.Z.Y. (Master’s student) conducted the simulation work and wrote the manuscript. W.Z.H. (Professor) conducted simulation supervision and revised the manuscript. H.Q.X. (Professor) conducted conceptualization, manuscript review and financial support.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.


1. Finkelman RB, Orem W, Castranova V, et al. Health impacts of coal and coal use: possible solutions. Int. J. Coal. Geol. 2002;50:425–443. https://doi.org/10.1016/S0166-5162(02)00125-8

2. Ci ZJ, Zhang XS, Wang ZW. Enhancing atmospheric mercury research in China to improve the current understanding of the global mercury cycle: The need for urgent and closely coordinated efforts. Environ. Sci. Technol. 2011;46:5636–5642. https://doi.org/10.1021/es300137y
crossref pmid

3. Zhan FM, Li CT, Zeng GM, et al. Experimental study on oxidation of elemental mercury by UV/Fenton system. Chem. Eng. J. 2013;232:81–88. https://doi.org/10.1016/j.cej.2013.07.082

4. Pavlish JH, Sondreal EA, Mann MD, et al. Status review of mercury control options for coal-fired power plants. Fuel. Process. Technol. 2003;82:89–165. https://doi.org/10.1016/S0378-3820(03)00059-6

5. Zheng YJ, Jensen AD, Windelin C, Jensen F. Review of technologies for mercury removal from flue gas from cement production processes. Prog. Energ. Combust. Sci. 2012;38:599–629. https://doi.org/10.1016/j.pecs.2012.05.001

6. Jennifer W. A Kinetic Investigation of High-Temperature Mercury Oxidation by Chlorine. J. Phys. Chem. A. 2009;113(24)6633–6639. https://doi.org/10.1021/jp901050d
crossref pmid

7. Martinez AI, Deshpande BK. Kinetic modeling of H2O2-enhanced oxidation of flue gas elemental mercury. Fuel. Process. Technol. 2007;88:982–987. https://doi.org/10.1016/j.fuproc.2007.05.009

8. Wang ZH, Zhou JH, Zhu YQ, Wen ZC, Liu JZ, Cen KF. Simultaneous removal of NOx, SO2 and Hg in nitrogen flow in a narrow reactor by ozone injection: Experimental results. Fuel. Process. Technol. 2007;88:817–823. https://doi.org/10.1016/j.fuproc.2007.04.001

9. Calvert JG, Lindberg SE. Mechanisms of mercury removal by O3 and OH in the atmosphere. Atoms. Environ. 2005;39:3355–3367. https://doi.org/10.1016/j.atmosenv.2005.01.055

10. Sun CL, Zhao N, Zhuang ZK, et al. Mechanisms and reaction pathways for simultaneous oxidation of NOx and SO2 by ozone determined by in situ IR measurements. J. Hazard. Mater. 2014;274:376–383. http://dx.doi.org/10.1016/j.jhazmat.2014.04.027

11. Hutson ND, Krzyzynska R, Srivastava RK. Simultaneous Removal of SO2, NOX, and Hg from Coal Flue Gas Using a NaClO2-Enhanced Wet Scrubber. Ind. Eng. Chem. Res. 2008;47:5825–5831. https://doi.org/10.1021/ie800339p

12. Wang MY, Luo HJ, Zhu TL. Effects of reaction conditions on elemental mercury oxidation in simulated flue gas by DC nonthermal plasma. Ind. Eng. Chem. Res. 2011;50:5914–5919. https://doi.org/10.1021/ie200130m

13. Zhang YD, Lou C, Liu DH, Li Y, Ruan LF. Chemical Effects of CO2 Concentration on Soot Formation in Jet-stirred/Plug-flow Reactor. Chin. J. Chem Eng. 2013;21(11)1269–1283. https://doi.org/10.1016/S1004-9541(13)60624-2

14. Ma Q, Wang ZH, Lin FW, et al. Characteristics of O3 Oxidation for Simultaneous Desulfurization and Denitration with Limestone–Gypsum Wet Scrubbing: Application in a Carbon Black Drying Kiln Furnace. Energ. Fuel. 2016;30:2302–2308. https://doi.org/10.1021/acs.energyfuels.5b02717

15. Lin FW, Wang ZH, Zhang ZM, et al. Flue gas treatment with ozone oxidation: An overview on NOx, organic pollutants, and mercury. Chem. Eng. J. 2020;382:123030. https://doi.org/10.1016/j.cej.2019.123030

16. Wen C, Cao XW, Yang Y. Swirling flow of natural gas in supersonic separators. Chem. Eng. Process. 2011;50(7)644–649. https://doi.org/10.1016/j.cep.2011.03.008

17. Zhu JH, Frerich T, Herrmann AX. CFD modeling and validation of heat transfer inside an autoclave based on a mesh independency study. J. Compos. Mater. 2021;0(0)1–19. https://doi.org/10.1177/0021998320979043

18. Rabitz H, Kramer M, Dacol D. Sensitivity analysis in chemical kinetics: Recent developments and computational comparisons. Int. J. Chem. Kinet. 1984;16:559–578. https://doi.org/10.1002/kin.550160506

19. Reuven Y, Smooke MD, Rabitz H. Sensitivity analysis of boundary value problems: application to nonlinear reaction-diffusion systems. J. Comput. Phys. 1986;64:27–55. https://doi.org/10.1016/0021-9991(86)90017-3

20. Tamds T. Applications of sensitivity analysis to combustion chemistry. Reliab. Eng. Syst. Safe. 1997;57:41–48. https://doi.org/10.1016/S0951-8320(97)00016-1

21. Wen ZC, Wang ZH, Li Y, Cen KF. Mechanism and Kinetic Study on Elemental Mercury Oxidation in Flue Gas by Ozone Injection. Ozone-Sci. Eng. 2017;40:29–36. https://doi.org/10.1080/01919512.2017.1330134

22. Gour NK, Begum SS, Deka RC. Computational study on night-time reaction of 1,1-Dichlorodimethylether (DCDME) CH3OCHCl2 with NO3 radical and the fortuity of alkoxy radical CH3OC(O·)Cl. Chem. Phys. Lett. 2018;701:157–164. https://doi.org/10.1016/j.cplett.2018.04.045

23. Yan YH, Wang SS, Zhu J, et al. Vertically increased NO3 radical in the nocturnal boundary layer. Sci. Total. Environ. 2021;763:142969. https://doi.org/10.1016/j.scitotenv.2020.142969
crossref pmid

24. Newland MJ, Ren YG, McGillen MR, Michelat L, Daële V, Mellouki A. NO3 chemistry of wildfire emissions: a kinetic study of the gas-phase reactions of furans with the NO3 radical. Atmos. Chem. Phys. 2021;22:1761–1772. https://doi.org/10.5194/acp-22-1761-2022

25. Li Y, Chen DF, Yang CL, et al. Engineering practice and economic analysis of ozone oxidation wet denitrification technology. Chin. J. Chem. Eng. 2020;29:401–408. https://doi.org/10.1016/j.cjche.2020.08.042

26. Wen ZC. Study on the mechanism of oxidation and degradation of multiple pollutants by ozone in flue gas. Zhejiang [dissertation]. Zhejiang: Zhejiang University; 2009.

27. Luo ZY, Lu TF, Maciaszek MJ, Som S, Longman DE. A Reduced Mechanism for High-Temperature Oxidation of Biodiesel Surrogates. Energ. Fuel. 2010;24:6283–6293. https://doi.org/10.1021/ef1012227

28. Chen YL, Chen JY. Application of Jacobian defined direct interaction coefficient in DRGEP-based chemical mechanism reduction methods using different graph search algorithms. Combust. Flame. 2016;174:77–84. https://doi.org/10.1016/j.combustflame.2016.09.006

29. Xi SH, Xue J, Wang F, Li XY. Reduction of large-size combustion mechanisms of n-decane and n-dodecane with an improved sensitivity analysis method. Combust. Flame. 2020;222:326–335. https://doi.org/10.1016/j.combustflame.2020.08.052

30. Tang AK, Huang QH, Li Y, Gao LJ, Ni Q. Development and verification of simplified methane/dimethyl ether mechanism for micro-combustion. Fuel. Process. Technol. 2022;226:107071-1–107071-10. https://doi.org/10.1016/j.fuproc.2021.107071

31. Parka SL, Moonb JD, Leeb SH, Shinc SY. Effective ozone generation utilizing a meshed-plate electrode in a dielectric-barrier discharge type ozone generator. J. Electrostat. 2006;64:275–282. https://doi.org/10.1016/j.elstat.2005.06.007

32. Pepiot-Desjardins P, Pitsch H. An automatic chemical lumping method for the reduction of large chemical kinetic mechanisms. Combust. Theor. Model. 2008;305:1089–1108. https://doi.org/10.1080/13647830802245177

Fig. 1
Hg0 dimensionless sensitivity coefficient.
Fig. 2
A comparison between reduced and detailed chemical mechanism.
Fig. 3
concentration distribution of Hg0 at pipeline section and its cross section. (a) Hg0 distribution and concentration variation with time in pipeline profile; (b) Hg0 distribution and concentration of pipeline cross section at 10 seconds; (c) O3 distribution and concentration in pipeline profile at 10 seconds.
Fig. 4
Different Temperatures
Fig. 5
Relation between reaction rate and oxidation efficiency
Fig. 6
Different Molar Ratio
Fig. 7
NO2 and NO3 concentration at outlet
PDF Links  PDF Links
PubReader  PubReader
Full text via DOI  Full text via DOI
Download Citation  Download Citation
Supplement  Supplement
Web of Science
Editorial Office
464 Cheongpa-ro, #726, Jung-gu, Seoul 04510, Republic of Korea
FAX : +82-2-383-9654   E-mail : eer@kosenv.or.kr

Copyright© Korean Society of Environmental Engineers.        Developed in M2PI
About |  Browse Articles |  Current Issue |  For Authors and Reviewers