Simultaneous removal of NO, SO2 and Hg0 with the WDRMRS

ABSTRACT Micro-nanobubbles can spontaneously generate hydroxyl free radicals (OH). Urea is a cheap reductant and can react with NOx species, and their products are nontoxic and harmless N2, CO2 and H2O. In this study, a Wet Direct Recycling Micro-nanobubble Flue Gas Multi-pollutants Removal System (WDRMRS) was developed for the simultaneous removal of NO, SO2 and Hg0. In this system, a micro-nanobubble generator (MNBG) was used to produce a micro-nanobubble gas–liquid dispersion system (MNBGLS) through recycling the urea solution from the reactor and the simulated flue gas composed of N2, NO, SO2 and Hg0. The MNBGLS, which has a large gas–liquid dispersion interface, was recycled continuously from the MNBG to the reactor, thus achieving cyclic absorption of various pollutants. All of the investigated parameters, including the initial pH and temperature of the absorbent as well as the concentrations of urea, NO and SO2 had significant effects on the NO removal efficiency but did not significantly affect the SO2 removal efficiency, whereas only the initial solution pH and NO concentration affected the Hg0 removal efficiency. The analysis results of the reaction mechanism showed that ·OH played a critical role in the removal of various pollutants. After the treatment by this system, the main removal products were Hg0 sediment, and NH4+ which could be easily recycled. The use of this system (MNBGLS) for the simultaneous removal of NO, SO2 and Hg0 is a new technology application and research. Recycling process based on MNBGLS succeeded in simultaneously removing NO, SO2 and Hg0. The system (MNBGLS) can provide a reference for commercial applications. The removal products are relatively simple and beneficial to recycling, which can reduce the cost of waste gas treatment. GRAPHICAL ABSTRACT


Introduction
Nitrogen oxides (NO x ) and sulfur dioxide (SO 2 ), the major air pollutants, are the main precursors for acid rain and photochemical smog, and mercury (Hg), which is considered to be one of the most toxic heavy metals, has received increasing attention because it has persistence, bio-accumulation and neurological toxicity Wei ZS et al. [48,[1][2][3][4][5]. NO x , SO 2 and Hg are the typical components in coal-fired flue gas. The Chinese government attaches great importance to air pollution control and takes effective measures to prevent and control NO x , SO 2 and Hg pollution from flue gas. At present, as a mature method for NO x , SO 2 and Hg removal in flue gas, the combined system of selective catalytic reduction (SCR) [6,7], wet flue gas desulfurization (WFGD) [8] and activated carbon injection (ACI) [2,9] has been widely applied. Although this system can achieve the deep removal of NO x , SO 2 and Hg 0 , it still has some disadvantages such as the high cost of construction and operation, high occupational area, high complexity, low stability, ammonia leakage, and secondary pollution [10,11]. Another method is an integrated removal process by which NO x , SO 2 and Hg 0 are simultaneously removed in the same device. This process can avoid the disadvantages of the aforementioned method, simplifying the system, reducing the cost, and still obtaining the high removal efficiency. The wet liquid phase oxidation-absorption process which can oxidize and absorb NO x , SO 2 and Hg 0 in one-step, is the main direction of the future research on the integrated removal of multi-pollutants from the flue gas [11].
As for the simultaneous removal of NO x , SO 2 and Hg 0 , the key issue is to remove nitric oxide (NO) and Hg 0 efficiently and rapidly because NO, which is sparsely soluble in water, accounts for 90∼95% of NO x in flue gas [12] and Hg 0 is insoluble in the water. According to the literature and our previous study, micro-nanobubbles (MNBs) can spontaneously generate a large amount of hydroxyl free radicals (OH) with a high oxidizing ability [13][14][15], which can not only oxidize NO into NO 2 but oxidize Hg 0 into mercuric hydroxide or mercuric oxide, as shown in reaction (1) [16][17][18]. Liu et al. [19] conducted the simultaneous removal of NO, SO 2 and Hg 0 by using a Fenton-like reagent based on Fe3+ in a spray reactor, and the results showed that OH generated from the Fenton-like regent played a critical role in the removal of NO, SO 2 and Hg 0 , and the removal efficiencies of NO, SO 2 and Hg 0 were 85.3, 100 and 100%, respectively, under the optimal conditions. Urea ((NH 2 ) 2-CO), which is cheap, nontoxic, harmless and easy to obtain, is a strong reductant with weak alkalinity. Fang et al. [1] found that urea could effectively promote the removal of NO 2 , and the main reaction products are the recyclable ammonium sulfate and nontoxic and harmless gases (CO 2 and N 2 ). Therefore, urea is an economic and eco-friendly reagent for the removal of multipollutants from flue gas.
In this work, we developed a Wet Direct Recycling Micronanobubble Flue Gas Multi-pollutants Removal System (WDRMRS), in which a micro-nanobubble gas-liquid dispersion system (MNBGLS) was generated by utilizing a micro-nanobubble generator (MNBG) to inject the simulated flue gas consisting of N 2 , NO, SO 2 and Hg 0 and recycle a urea aqueous solution of a certain concentration from a column reactor. The MNBGLS recycled continuously from the MNBG to the reactor. The continuous cyclic absorption of NO, SO 2 and Hg 0 was achieved through a series of oxidation-reduction reactions in the reactor. Moreover, the MNBGLS has a high gas-liquid dispersion interface and can enhance the gas solubility [15]. To explore the performance of the oxidation-reduction system of the MNBGLS based on urea, the effects of the important operational parameters on the simultaneous removal of NO, SO 2 and Hg 0 , such as initial pH and temperature of the absorbent, urea concentration, NO concentration and SO 2 concentration, were investigated. The potassium permanganate absorbent solution was used to determine the concentration of mercury entering the reactor and was not used as an absorbent. The absorbent used for the removal was a urea solution and the oxidizing effect was mainly the hydroxyl radicals released by the micro and nano bubbles. The choice of these additives was based on cost and ease of preparation for industrial applications.

Materials and instruments
Materials and instruments are respectively provided in Tables 1 and 2 in Supplementary materials in Appendix.

Experimental procedure
The experimental setup is shown in Figure 1, and the detailed experimental procedures are described as follows.
First, urea aqueous solution (190 L) is prepared and fed into a column reactor (22) from the absorbent inlet (24) on the top of the reactor. To obtain the experimental requirements, a heating tube (26) and temperature controller (25) are used to adjust the solution temperature, and HCl and NaOH are used to adjust the solution pH. Then, the mixed solution (0.8 L), in which the concentrations of potassium permanganate and urea are 7 mmol/L and 5 wt.%, respectively, and the pH is 8, is prepared and fed into a three-neck flask (19). The digital display constant temperature water baths (8) and (18) are turned on and their temperature readings are set as 303 and 328 K, respectively.
Second, the gas control valves (13) and (16) are opened, and the gas control valves (14), (15) and (17) are closed. The gas relief valves on the cylinders (1-3) are all opened so that NO, SO 2 , and N 2 flow into a surge flask (10) by passing through the rotameters (4)(5)(6). Meanwhile, the carrier gas of Hg 0 and N2 passes through a rotameter (7) (flow rate: 200 ml/min) to bring Hg 0 vapour from a mercury permeation tube (9) into the surge flask (10), thus forming the simulated flue gas, which flows into a flue gas analyser (12), with which the concentrations of NO and SO 2 are measured by passing through a rotameter (11) (flow rate: 2 L/ min). The excess gas is released after being absorbed in the tail gas absorption tank (20). The desired gas concentration is achieved by adjusting the gas flow.
Third, the gas control valves (13) and (16) are closed and the gas control valves (14), (15) and (17) are opened after the gas concentration stabilizes. Meanwhile, the MNBG (21) is turned on to simultaneously inhale the simulated flue gas from the surge flask (10) and the absorbent from the column reactor (22) to form the MNBGLS, which is injected back into the reactor again for recycling. The excess gas enters the three-neck flask (19) and is absorbed. Then, the absorbed tail gas is released after being absorbed again in the tail gas absorption tank (20). The concentration of the gas released from the reactor is measured with the flue gas analyser (12). Finally, the system is turned off after running for 60 min. The test sample is then collected from the absorbent outlet (23) at the bottom of the column reactor (22). After the end of every single experiment, the absorbed liquid is discharged from the absorbent outlet (23) into the waste liquid storage tank and treated as the waste liquid.

NO/SO 2 removal efficiencies
The removal efficiencies of NO and SO 2 can be calculated as follows: where η (%) represents the removal efficiency, C in (ppm) represents inlet NO/SO 2 concentrations, and C out (ppm) represents outlet NO/SO 2 concentrations.

Hg 0 removal efficiency
According to the results of Fang et al. [20], under the conditions of 7 mmol/L potassium permanganate, 5 wt.% urea, 52.1 μg/m3 Hg 0 , 650 mg/m 3 NO, 2900 mg/ m3 SO 2 , 7% O2 and a solution temperature of 328 K, the Hg 0 removal efficiency was 100% when the solution was alkaline. Thus, five groups of experiments were designed to test the absorption efficiency of 0.8 L of mixed solution of potassium permanganate and urea for Hg 0 . ICP was used to test the concentration of Hg 0 in the absorbed solution, and the absorption efficiency of Hg 0 was calculated by formula (3): where E (%) is the absorption efficiency of Hg 0 , C Hg (mg/ L) is the concentration of Hg 0 in the absorption solution, and v(L) is the volume of the absorption solution. 150 refers to a mercury permeation rate of 150 ng/ min and 60 refers to minutes. The detailed experimental conditions and results are provided in Table 3 which is showed in Supplementary materials in Appendix.
Based on these experimental results, ICP is utilized to test the sample liquid from the column reactor and three-neck flask to determine the Hg 0 concentration before entering the reactor and after the removal reaction, respectively, and the Hg 0 removal efficiency is calculated by the formula (4): where h ′ (%) is the Hg 0 removal efficiency, C ′ in (mg/L) is the concentration of Hg 0 in the three-neck flask, V 1 (L) is the volume of the solution in the three-neck flask, C ′ out (mg/L) is the concentration of Hg 0 in the column reactor, and V 2 (L) is the volume of the solution in the column reactor.

Detection of the reaction products
and SO 2− 4 in the reaction products are detected by IC (ICS-1000), and NH + 4 in the reaction products is detected by IC (DX-600).

Detection of OH
The existence of OH has been confirmed by fluorescence spectrum experiments from another paper [15].

Effects of the urea concentration
As shown in Figure 2, the urea concentration had little effect on the removal efficiencies of SO 2 and Hg 0 . No matter whether urea was added or not, the SO 2 removal efficiency was greater than 99.6% and the Hg 0 removal efficiency was maintained at approximately 86%. However, the urea concentration significantly affected the NO removal efficiency, which first increased and then slightly decreased as the urea concentration increased. This was because the chemical reactivity and physical properties of urea caused changes in the NO removal efficiency. On the one hand, urea could react with the main N species in the solution including NO, NO 2 , NO − 2 , and NO − 3 , thus accelerating the oxidation and absorption of NO and promoting NO removal. On the other hand, with the increase in the urea concentration, the viscosity of the absorption solution increased so that the diffusion rate and solubility of NO in the liquid phase decreased Wei JC et al. [49], thus inhibiting the oxidation and absorption of NO. Therefore, the NO removal efficiency decreased when the urea concentration increased from 3 g/L to 5 g/L. To obtain the best removal effects, the urea concentration was designed as 3 g/L in subsequent experiments.

Effects of the initial solution pH
The effects of various initial solution pH values (3.47, 5.51, 7.46, 9.46 and 10.47) on the removal efficiencies of NO, SO 2 and Hg 0 were investigated (Figure 3). With the increase in the initial pH of urea solution, the SO 2 removal efficiency did not change significantly but only increased slightly from 99.4% to 99.9%. Meanwhile, the NO removal efficiency increased first and then decreased, and it reached the highest value (96.7%) at pH = 9.46, which was, however, only 0.6 higher than that at a pH = 7.46 (the original pH of the 3 g/L urea solution). Moreover, the Hg 0 removal efficiency decreased slightly as the pH increased. Therefore, to give consideration to the removal effects and convenience of the system operation, the initial solution pH of the subsequent experiments was designed to be 7.46.
According to the relevant literature, under strong acidic conditions, the presence of a great deal of H + was not conducive to the reaction of OH and NO [21], and it also accelerated the decomposition of HNO2 so that NO was regenerated in the removal reaction [1], thus inhibiting NO removal. Hence, as shown in reaction (1), more OH participated in the reaction of Hg 0 so that the removal efficiency of Hg 0 increased. However, the weak acidic conditions were conducive to the hydrolysis of urea, which could promote the oxidation and absorption of NO [20]. Moreover, with the increase in the solution pH, more and more OH − could absorb H + through the acid-base neutralization reaction (5) to promote the oxidation of OH to NO [22,23], thus improving the NO removal efficiency. Some researchers indicated that OH − could consume OH via reaction (6) at a very high rate (1.3 × 1010 M −1 s −1 ) [24][25][26], and radical ions (O − ) reacted more slowly with the same substrate than OH [27]; therefore, too high of a solution pH weakened the oxidation capacity of the MNBGLS, resulting in the inhibition of NO removal. In addition, the MNBs could not only grow in size but collapse faster because the excessive OH − repelled OH − on the surface of the MNBs to aggravate the movement of bubbles, thus decreasing the retention time of the bubbles in the water [28] and further inhibiting the absorption of NO. Therefore, the NO removal efficiency decreased when the initial solution pH increased from 9.46 to 10.47. Meanwhile, due to the competition between the reaction of OH and NO and the reaction of OH and Hg 0 , the Hg 0 removal efficiency decreased as the solution pH increased.

Effects of the initial solution temperature
As shown in Figure 4, with the increase in the initial solution temperature, the removal efficiencies of SO 2 and Hg 0 were insignificantly affected, but the NO removal efficiency gradually decreased, indicating that this MNBGLS based on urea had good performance in the removal of NO, SO 2 and Hg 0 under normal temperature conditions. According to the research, an increase in the solution temperature had four effects: the hydrolysis of urea was promoted to improve the oxidation and absorption of NO [20]; the active molecules in the solution increased and their reactivity was enhanced, thus increasing the reaction rate [29]; the gas solubility decreased [24,30]; and the stability of the MNBs declined, which further affected the gas solubility [31,32]. Amongst these effects, the former two effects promoted NO removal, while the latter two effects inhibited NO removal. In this work, due to the dominant inhibitory effect, the NO removal efficiency decreased as the solution temperature increased.

Effects of the NO concentration
As shown in Figure 5, with the increase in the NO concentration, the SO 2 removal efficiency changed insignificantly, while the removal efficiencies of NO and Hg 0 gradually decreased. It is worth noting that the performance of the MNBGLS remained unchanged because of the rated inlet gas flow and recycling solution volume of the MNBG. Therefore, more OH generated from the MNBs participated in the oxidation and absorption of NO due to the less NO entering the reactor under low NO concentrations. As the NO concentration increased, the quantities of NO molecules entering the reactor  per unit time raised to increase the gas to liquid relative molar ratio of NO and solution [33] so that NO was released more easily because more NO was not oxidized in time. Additionally, since the reaction of OH and NO competed with the reaction of OH and Hg 0 , more NO entered the system to consume more OH, resulting in less OH involved in the oxidation reaction of Hg 0 . Consequently, increasing the NO concentration decreased the removal efficiencies of NO and Hg 0 .

Effects of the SO 2 concentration
The effects of the SO 2 concentration on the simultaneous removal efficiencies of NO, SO 2 and Hg 0 were also investigated ( Figure 6). Due to the high solubility of SO 2 , the removal efficiencies of SO 2 and Hg 0 were not affected by the changes in the SO 2 concentration and were maintained above 99.8% and at approximately 86%, respectively. Nevertheless, the increase in the SO 2 concentration had a significant effect on the NO removal efficiency. The NO removal efficiency increased sharply from 85% to 92.9% when the SO 2 concentration increased from 0 to 950 ppm and then slowly increased as the SO 2 concentration increased from 950 ppm to 2750 ppm, but it declined significantly when the SO 2 concentration increased from 2750 to 3550 ppm. The reasons could be explained as follows: SO 2 entered the liquid phase and was quickly dissolved into the water to form HSO − 3 and SO 2− 3 [34,35], both of which reacted with NO, NO 2 , NO − 2 and NO − 3 to promote the oxidation and absorption of NO [36,37]. Moreover, HSO − 3 and SO 2− 3 could also combine with NO to form (ONSO 3 ) 2− , which further improved the absorption of NO [38]. Hence, the NO removal efficiency increased as the SO 2 concentration increased. However, HSO − 3 and SO 2− 3 were also the scavengers of OH, thus consuming more OH as the SO 2 concentration increased and inhibiting the removal of NO. When the SO 2 concentration rose from 2750 to 3550 ppm, this inhibition played a leading role in the removal of NO, causing the decrease of the NO removal efficiency. In the system, the NO removal efficiency in the presence of SO 2 is generally higher than that in the absence of SO 2 .

Parallel tests
Parallel tests were conducted under the optimal conditions (3 g/L urea concentration, initial solution pH of 7.46, initial solution temperature of 298 K, 750 ppm NO, 2750 ppm SO 2 , and a Hg 0 penetration rate of 150 ng/min). As shown in Table 1, the removal efficiencies of NO, SO 2 and Hg 0 exhibited good reproducibility, indicating that the system had stable performance and could provide a reference for its industrial application. Table 2 shows the detection of the reaction products.   As shown in Table 2, NO −  2 and SO 2−   3 were not detected, and the reaction products were mainly SO 2− 4 and NH 4+ , with only a small amount of NO − 3 . This was because the reaction between nitrous acid and urea was stronger than that between nitric acid and urea [1], and nitric acid was relatively steady in the solution. SO 2− 3 could not only be oxidized by OH to SO 2− 4 but also participated in the oxidation-absorption of NO, thus being consumed. NH 4+ was produced by the hydrolysis of urea.

Reaction mechanism
The chemical reactions of NO, SO 2 and Hg 0 in this system are complicated, but the main reaction mechanisms can be inferred from the results in Table 2 and are depicted in Figure 7. When the MNBs collapse in the MNBGLS, OH is generated and NO, SO 2 and Hg 0 are also released from the bubbles. SO 2 quickly dissolves into the water and is converted to HSO − 3 , SO 2− 3 and H + (reactions (7) and (8)) [39]. Part of the SO 2 in the gas phase is also oxidized by OH to SO 2− 4 (reactions (9) and (10)) [40].  (10) NO is oxidized by OH and absorbed in the water to form NO − 3 and NO − 2 (reactions (11)(12)(13)(14)(15)(16)) [21]. Additionally, NO can also combine with NO 2 to form N 2 O 3 , and NO 2 and N 2 O 4 can convert into each other. In the liquid phase, the highly soluble compounds N 2 O 3 and N 2 O 4 also dissolve and react rapidly with water to produce equimolar amounts of HNO 2 and HNO 3 [1]. Therefore, the main NO x species in the gas and liquid phases may be NO, NO 2 , N 2 O 3 , N 2 O 4 , HNO 2 and HNO 3 . The potassium permanganate solution is only used to test the concentration of mercury entering the reactor and is not a sorbent for experimental stripping. The·OH released by the micro-nano bubbles and urea play an important role in the simultaneous Urea in the MNBGLS hydrolyzes to produce ammonium carbamate (NH 2 COONH 4 ) (reaction 17). Urea can react with NO and NO 2 , NH 2 COONH 4 can react with HNO 2 and HNO 3 , and all of these reactions produce N 2 and CO 2 (reactions (18)(19)(20)(21)) [1,20].
With regard to Hg 0 , it is oxidized by OH to Hg(OH) 2 (reaction (1)), which is easily heated to produce HgO sediment, thus achieving the removal of Hg 0 .

Conclusions
The recycling oxidation-reduction process of MNBGLS based on urea, which was proposed in this study, succeeded in simultaneously removing NO, SO 2 and Hg 0 . Under the optimal conditions, the average removal efficiencies of NO, SO 2 and Hg 0 were 97.5, 99.8 and 87.2%, respectively. Increasing solution temperature and NO concentration decreased the removal efficiencies of NO and Hg 0 , whereas with the increase in solution pH and concentrations of urea and SO 2 , Hg 0 removal efficiency did not change obviously but NO removal efficiency increased firstly and then decreased. All of the investigated parameters had an insignificant effect on SO 2 removal efficiency. The reaction mechanism was deduced according to the analysis of the removed products and it was found that the oxidation of OH generated from the MNBs played a dominant role in the removal of NO, SO 2 and Hg 0 , meanwhile urea further promoted the removal of NO through reduction reaction. The main removal products were SO 2− 4 , NH 4+ , a small amount of NO − 3 , as well as non-toxic and harmless N 2 . N 2 can be directly discharged and other removal products can be made into nitrogen fertilizer after Hg2+ in them is separated by heating, so as to improve the economic benefits of flue gas treatment and reduce the treatment costs.

Data availability statement
The datasets and materials used and/or analysed during the current study are available from the author and corresponding author on reasonable request.