Achieving Large-Capability Adsorption of Hg0 in Wet Scrubbing by Defect-Rich Colloidal Copper Sulfides under High-SO2 Atmosphere

This paper reports on a novel method to remove Hg0 in the wet scrubbing process using defect-rich colloidal copper sulfides for reducing mercury emissions from non-ferrous smelting flue gas. Unexpectedly, it migrated the negative effect of SO2 on mercury removal performance, while also enhancing Hg0 adsorption. Colloidal copper sulfides demonstrated the superior Hg0 adsorption rate of 306.9 μg·g−1·min−1 under 6% SO2 + 6% O2 atmosphere with a removal efficiency of 99.1%, and the highest-ever Hg0 adsorption capacity of 736.5 mg·g−1, which was 277% higher than all other reported metal sulfides. The Cu and S sites transformation results reveal that SO2 could transform the tri-coordinate S sites into S22− on copper sulfides surfaces, while O2 regenerated Cu2+ via the oxidation of Cu+. The S22− and Cu2+ sites enhanced Hg0 oxidation, and the Hg2+ could strongly bind with tri-coordinate S sites. This study provides an effective strategy to achieve large-capability adsorption of Hg0 from non-ferrous smelting flue gas.


Introduction
Mercury has been a contaminant of global concern due to its toxicity, persistence, and bioaccumulation since the implementation of the Minamata Convention on August, 2017 [1]. Among anthropogenic sources, the non-ferrous smelting sector is a major contributor, accounting for 14.9% of all worldwide emissions, as reported in the Global Mercury Assessment Report 2018 [2,3]. Therefore, it is urgent to efficiently remove Hg 0 from non-ferrous smelting flue gas to abate the significant challenge of global mercury pollution control.
There are various forms of mercury in non-ferrous smelting flue gas, including elemental mercury (Hg 0 ), oxidized mercury (Hg 2+ ), and particulate mercury (Hg p ) [4][5][6]. In the existing flue gas purification system, Hg p would be captured by the electrostatic precipitator (ESP) [7][8][9]. Hg 2+ has strong water solubility and can be removed by wet scrubber, whereas the capture of Hg 0 is difficult due to its strong stability, high volatility, and low solubility [10]. Wet methods for Hg 0 removal promote the conversion of Hg 0 to Hg 2+ , such as the Boliden-Norzink and advanced oxidation process, thereby improving mercury removal efficiency [11,12]. Hence, oxidation is considered as a vital way for Hg 0 removal in the wet scrubbing of non-ferrous smelting flue gas [13].
Oxidative demercurization remains a major bottleneck in the application of wet scrubbing from non-ferrous smelting flue gas [14,15]. There is a large amount of reductive SO 2 (>6% vol) in the flue gas, which is thousands of times higher than Hg 0 concentration [16,17]. Selective oxidation of Hg 0 poses a great challenge since the standard oxidation potential of Hg 0 (0.85 V) is higher than that of SO 2 (0.17 V) [18,19]. Consequently, SO 2 is preferentially oxidized over Hg 0 , bringing about the depletion of oxidants and making it difficult to oxidize Hg 0 . Therefore, the efficient oxidation of Hg 0 from high-SO 2 flue gas is key to overcoming the bottleneck of Hg 0 removal in wet scrubbing.
Transition metal sulfides (TMSs) have become a hot material for Hg 0 capture due to the strong affinity of sulfur sites with Hg 0 [20][21][22]. Previous studies had shown that SO 2 could react with defective S 2− on TMSs to generate highly active S n 2− , which might promote Hg 0 oxidation [23]. O 2 dissolves into aqueous phase and forms dissolved oxygen, which has a high oxidation activity due to hydrogen bonding and Van Der Waals force with water molecules, providing an oxidative environment for TMSs to capture Hg 0 [24]. Therefore, if TMSs are added to the wet scrubber, it is expected to eliminate the negative impact of SO 2 on the removal of Hg 0 , which could enhance the ability of TMSs to oxidize mercury, thereby greatly improving the adsorption capacity of mercury. However, the mechanism of SO 2 and O 2 in wet scrubbing for mercury capture by TMSs has not been reported.
Based on the above considerations, this paper involves a novel method of using metal sulfides in the wet scrubbing process for large-capability adsorption of Hg 0 from high-SO 2 flue gas. The optimal mercury adsorbent was prepared by sulfide precipitation and optimized systematically. The morphology and structure of the adsorbents were characterized. The excellent Hg 0 removal performance of colloidal copper sulfides (c-CuS) was examined in the flue gas wet scrubber. The effects of SO 2 and O 2 on Hg 0 removal by c-CuS were investigated, and the mechanism of the activation of oxidizing sites (S 2 2− and Cu 2+ ) by SO 2 and O 2 was proposed after identifying the structural changes of c-CuS under different atmospheres.

Chemicals and Reagents
The chemicals in analytical grade were used in this study, including copper chloride (CuCl 2 ), copper nitrate (CuNO 3 ·6H 2 O), Copper sulfate (CuSO 4 ·5H 2 O), sodium sulfide (NaS·9H 2 O), sodium hydroxide (NaOH), sodium chloride (NaCl), and nitric acid (HNO 3 ). They were purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China) All the agents were directly used without any further purification. Ultrapure water was used in all experiments unless otherwise stated.

Synthesis of TMSs
TMSs suspensions were synthesized by the double-jet liquid-phase sulfidation precipitation method. Simply, 50 mL of a 10 mmol·L −1 metal chloride solution and an equal volume of 10 mmol·L −1 Na 2 S·9H 2 O solution were added simultaneously to the beaker at a speed of 1.2 mL min −1 . These solutions were mixed by stirring at a speed of 380 rpm for 0.5 h, and then TMSs suspensions were obtained. c-CuS was synthesized using a single-jet liquid-phase sulfidation method. Then, 50 mL of Na 2 S·9H 2 O was rapidly added into a CuCl 2 solution of 1.0 mmol·L −1 , and then they were mixed by stirring. The experimental investigation on raw material concentration and ratio was undertaken to judiciously optimize the synthesis conditions.

Sample Characterization
These suspensions and colloids were filtered with nanopore filter membranes. The filtered residues were washed three times with ultrapure water and then dehydrated at −89 • C in a vacuum freeze-dryer for 8 h. The obtained filtered residues of TMSs suspensions and c-CuS were analyzed for crystalline structure by X-ray diffraction (XRD, Empyrean 2, PANalytic, Malvern, UK) with Cu-Kα radiation. High-resolution transmission electron microscopy (HRTEM, Titan G260-300, FEI, Lausanne, Switzerland) was used to observe the morphology of c-CuS particles. The scanning electron microscope (SEM, JSM-6360LV, Jeol, Tokyo, Japan) of c-CuS was characterized at the accelerating voltage of 200 kV after splashing a thin Au layer. The structure information of samples under different atmospheres was characterized by X-ray photoelectron spectroscopy (XPS, Escalab 250 Xi, Thermo Fisher Scientific, Waltham, MA, USA). The ultimate vacuum degree of the sample analysis room was 5 × 10 −7 Pa, and the Al Kα X-ray was used as the excitation source with a power of 16 mA × 12.5 kV. Correction was made using C ls = 284.6 eV as the internal standard for electron binding energy. The fingerprint and structural properties of c-CuS were analyzed by three-dimensional excitation-emission matrix (3D-EEM, F7000, Hitachi, Ibaraki, Japan) fluorescence spectroscopy. 3D-EEM spectra were generated by scanning in the range of 200 nm to 600 nm, while excitation and emission sampling wavelengths were 10 nm, and the scanning speed was 12,000 nm·min −1 . Finally, temperature-programmed desorption (TPD) was employed to identify the form of mercury species adsorbed on c-CuS. The TPD tests were carried out under pure nitrogen at the flow rate of 500 mL·min -1 from 50 • C to 600 • C at a rate of 5 • C·min −1 .

Hg 0 Removal Test
An experiment was conducted in a bubbling reactor to remove Hg 0 . Pure N 2 was used as the carrier gas for the Hg 0 generator (VICI Metronics) with a flow rate of 200 mL·min −1 . The simulated flue gas (SFG) containing N 2 , SO 2 , and O 2 at a flow rate of 300 mL·min −1 was mixed with the Hg 0 carrier gas. The mixture gas was then blown into a bubbling reactor containing the scrubbing solution. Then, the SFG was treated with 5 mol·L −1 NaOH and quartz wool before the mercury detection to reduce the influence of SO 2 and water vapor. The mercury analyzer (RA-915M, Lumex Zeeman) was used to monitor the outlet Hg 0 concentration. All data were obtained by averaging three measurement results. Finally, the remaining Hg 0 in the tail gas was absorbed by KMnO 4 solution and activated carbon. The operating conditions in the experiment are listed in Table S1. The Hg 0 removal efficiency and adsorption capacity were calculated according to Equations (1) and (2), respectively.
where η (%) represents Hg 0 removal efficiency, and C in and C out (µg·m −3 ) are the Hg 0 concentration at the inlet and outlet of the experimental system, respectively. Q t (mg·g −1 ) is the adsorption capacity of Hg 0 , M (mg) is the mass of adsorbent, t 1 and t 2 (min) are the start and end times, ∆t (min) is the scrubbing time, and u is the gas flow rate (m 3 ·min −1 ).

DFT Calculation
Perdew-Burke-Ernzerhof (PBE0) functional was used to characterize the exchangecorrelation effects in the density functional theory (DFT) by CP2K [25,26]. The CuS (110) system was simulated using an orthorhombic box with dimensions of 16.44 × 13.15 × 34.49 Å3, and the dipole correction technique was applied in the XY orientation. BROYDEN-MIXING was used to optimize the geometries, while the DZVP-MOLOPT-SR-GTH basis set was used to describe the basis. A plane wave cutoff of 400 Ry was set, and Goedecker-Teter-Hutter (GTH) pseudopotentials were used to represent all electrons [27,28]. In addition, the D3 dispersion correction by Grimme et al. was employed, and the adsorption energy (E ad ) was calculated using Equation (3), based on the optimized structure.
where E Hg is the state energy of a free Hg atom, E CuS represents the total energy of the CuS of different configurations, and E (Hg+CuS) is the total energy of the Hg atoms being adsorbed on CuS.

Selection and Optimization of Transition Metal Sulfides for Hg 0 Removal
Hg 0 removal performance was evaluated by adding TMSs suspensions into a bubble reactor as the scrubbing solution. Figure 1a shows that TMSs suspensions had a good ability to capture Hg 0 in wet scrubbing. Especially, copper sulfides suspension was the preferred Hg 0 removal scrubbing agent, with the highest efficiency of 87.6%. Furthermore, the effect of regulating feeding methods on the dispersibility of detergent was investigated to improve the Hg 0 removal performance of copper sulfides. Figure 1b displays that c-CuS prepared by single-jet was homodispersed in solution, and the Tyndall effect was evident, while the precipitation of suspensions copper sulfides (s-CuS) was visible. The removal efficiency gradually decreased from 65.8% to 46.4% due to the poor contact between gas and s-CuS, while the mercury removal efficiency of c-CuS could be maintained at 97.5~98.9%. Therefore, the c-CuS synthesized by single-jet has better Hg 0 capture performance.  Figure 2 shows the XRD patterns of the obtained c-CuS and s-CuS filter residues. The phase was confirmed to be Cu4(S2)2(CuS)2 (JCPDS card NO. 74-1234), with overlapping diffraction peaks that could be divided into lattice planes such as (101), (102), (103), (110), and (116) [29]. The diffraction peaks of s-CuS were stronger and sharper, indicating a better crystallinity. According to Equation (4), the crystallinity of s-CuS was calculated to be 70.9%. Surprisingly, the crystallinity of the c-CuS sample was only about 3.4% with severely broadened diffraction peaks. It indicated that c-CuS might contain abundant defects, which was beneficial for exposing more active adsorption sites to capture Hg 0 [15,30].

Structural Characterizations of c-CuS
where Ic is the crystal diffraction intensity and Ia is the amorphous scattering intensity. Optimization of the raw material concentration and ratio to synthesize copper sulfides is crucial for enhancing Hg 0 removal. As shown in Figures 1c and S1, when the concentration of c-CuS was 1/32 mmol·L −1 , the efficiency of Hg 0 removal was only 65% due to the insufficient content of active components in the solution. Increasing the concentration of CuCl 2 and Na 2 S fourfold to synthesize c-CuS could immediately improve the removal efficiency of Hg 0 to 93%. When the concentrations were optimized to 1/2 mmol·L −1 and 2 mmol·L −1 , the efficiencies were higher than 96.3%. Unfortunately, a large amount of precipitation appeared in the solution of 2 mmol·L −1 concentration, which was not conducive to sufficient contact with Hg 0 . Finally, the Cu/S ratio of copper sulfides was optimized, as shown in Figure 1d. Copper sulfides had the best removal efficiency for Hg 0 , which was achieved by reaching 98.1% when the Cu:S ratio was 1:1. However, as shown in Figure  S2, further increasing the S 2− ratio caused possible instability of colloid, which would have a negative effect on the Hg 0 removal. In summary, the c-CuS was prepared with the concentration of 1/2 mmol·L −1 and the Cu:S raw material ratio of 1:1, which was the optimal mercury capture agent. Figure 2 shows the XRD patterns of the obtained c-CuS and s-CuS filter residues. The phase was confirmed to be Cu 4 (S 2 ) 2 (CuS) 2 (JCPDS card NO. 74-1234), with overlapping diffraction peaks that could be divided into lattice planes such as (101), (102), (103), (110), and (116) [29]. The diffraction peaks of s-CuS were stronger and sharper, indicating a better crystallinity. According to Equation (4), the crystallinity of s-CuS was calculated to be 70.9%. Surprisingly, the crystallinity of the c-CuS sample was only about 3.4% with severely broadened diffraction peaks. It indicated that c-CuS might contain abundant defects, which was beneficial for exposing more active adsorption sites to capture Hg 0 [15,30].

Structural Characterizations of c-CuS
where I c is the crystal diffraction intensity and I a is the amorphous scattering intensity.  Figure 3 shows the XPS full spectrum of c-CuS. The peak analysis result indicated that the Cu:S atomic ratio was 0.98:1 on the surface of c-CuS filter residue (the O, C peak belongs to the peak of carrier sheet). The mismatched stoichiometric ratio suggested that rich sulfur sites were exposed. Figure 4a shows that c-CuS was composed of near-spherical nanoparticles with the diameter of about 8.0 nm, similar to those of SEM in Figure S3. HRTEM images in Figure 4b display many orange-marked vacancies and amorphous regions in the lattice, indicating that rich defect that could provide more active sites for mercury capture. The measured lattice spacing of 1.89 nm was consistent with the (110) plane of c-CuS. The broad diffraction rings in the selected-area electron diffraction (SAED) shown in Figure 4c further confirmed the existence of amorphous regions in c-CuS [31]. Figure 4d-f shows the high-angle annular dark-field image (HADDF), Cu element, and S element scanning images of c-CuS, respectively. It confirms that Cu and S were evenly distributed. In summary, c-CuS with abundant defect sites has the potential for high-activity mercury capture.  Figure 3 shows the XPS full spectrum of c-CuS. The peak analysis result indicated that the Cu:S atomic ratio was 0.98:1 on the surface of c-CuS filter residue (the O, C peak belongs to the peak of carrier sheet). The mismatched stoichiometric ratio suggested that rich sulfur sites were exposed. Figure 4a shows that c-CuS was composed of near-spherical nanoparticles with the diameter of about 8.0 nm, similar to those of SEM in Figure S3. HRTEM images in Figure 4b display many orange-marked vacancies and amorphous regions in the lattice, indicating that rich defect that could provide more active sites for mercury capture. The measured lattice spacing of 1.89 nm was consistent with the (110) plane of c-CuS. The broad diffraction rings in the selected-area electron diffraction (SAED) shown in Figure 4c further confirmed the existence of amorphous regions in c-CuS [31]. Figure 4d-f shows the high-angle annular dark-field image (HADDF), Cu element, and S element scanning images of c-CuS, respectively. It confirms that Cu and S were evenly distributed. In summary, c-CuS with abundant defect sites has the potential for high-activity mercury capture. cury capture. The measured lattice spacing of 1.89 nm was consistent with the (110) plane of c-CuS. The broad diffraction rings in the selected-area electron diffraction (SAED) shown in Figure 4c further confirmed the existence of amorphous regions in c-CuS [31]. Figure 4d-f shows the high-angle annular dark-field image (HADDF), Cu element, and S element scanning images of c-CuS, respectively. It confirms that Cu and S were evenly distributed. In summary, c-CuS with abundant defect sites has the potential for high-activity mercury capture.

Hg 0 Removal Performance of c-CuS under SO2 and O2 Atmospheres
SO2 is typically considered to be a negative effects gas that inhibits Hg 0 removal in industrial flue gas [23,32]. SO2 resistance has a crucial impact on the mercury capture performance of adsorbents. SO2-contained simulated flue gas was inputted into a bubbling reactor with 10 mL of c-CuS to investigate the effect of SO2 on the Hg 0 removal performance of c-CuS. As shown in Figure 5a, the removal efficiency was 80.1% in the absence of SO2. The efficiency significantly increased to 95.1~97.4% under 3~9% vol SO2. Furthermore, the performance of c-CuS for Hg 0 removal was compared with and without SO2 pretreatment, as shown in Figure 5b. The outlet concentration of Hg 0 was about 10 μg·m −3 after passing through c-CuS solution pretreated without SO2, but it decreased immediately to ~0 μg·m −3 after turning on SO2. When c-CuS solution was pretreated with SO2, the outlet Hg 0 concentration was about 3 μg·m −3 after opening gas pipeline of Hg 0 . Meanwhile, the outlet concentration exhibited almost no fluctuation after supplying SO2 and Hg 0 simultaneously. This might mean that SO2 did not occupy the adsorption sites, and SO2 could interact with c-CuS to enhance adsorption with Hg 0 . The 3D-EEM was used to compare the fluorescence fingerprint spectra of c-CuS under N2 + 6% vol SO2 atmospheres. As shown in Figure 5c,d, the region II area of the sample treated with SO2 was smaller than

Hg 0 Removal Performance of c-CuS under SO 2 and O 2 Atmospheres
SO 2 is typically considered to be a negative effects gas that inhibits Hg 0 removal in industrial flue gas [23,32]. SO 2 resistance has a crucial impact on the mercury capture performance of adsorbents. SO 2 -contained simulated flue gas was inputted into a bubbling reactor with 10 mL of c-CuS to investigate the effect of SO 2 on the Hg 0 removal performance of c-CuS. As shown in Figure 5a, the removal efficiency was 80.1% in the absence of SO 2 . The efficiency significantly increased to 95.1~97.4% under 3~9% vol SO 2 . Furthermore, the performance of c-CuS for Hg 0 removal was compared with and without SO 2 pretreatment, as shown in Figure 5b. The outlet concentration of Hg 0 was about 10 µg·m −3 after passing through c-CuS solution pretreated without SO 2 , but it decreased immediately to~0 µg·m −3 after turning on SO 2 . When c-CuS solution was pretreated with SO 2 , the outlet Hg 0 concentration was about 3 µg·m −3 after opening gas pipeline of Hg 0 . Meanwhile, the outlet concentration exhibited almost no fluctuation after supplying SO 2 and Hg 0 simultaneously. This might mean that SO 2 did not occupy the adsorption sites, and SO 2 could interact with c-CuS to enhance adsorption with Hg 0 . The 3D-EEM was used to compare the fluorescence fingerprint spectra of c-CuS under N 2 + 6% vol SO 2 atmospheres. As shown in Figure 5c,d, the region II area of the sample treated with SO 2 was smaller than the region I area of c-CuS under N 2 atmosphere, confirming that SO 2 leads to the formation of more vacancies. The formation of vacancies meant that low-active S 2− had been converted to more active S 2 2− [33,34]. The specific conversion process of activity sites will be further studied in the following.
c-CuS and strengthen the oxidation of Hg 0 .
The Hg 0 adsorption capacity and rate of c-CuS were investigated and calculated by penetration experiments. The prepared 10 mL c-CuS (containing 0.48 mg CuS particles) was input with 6% vol SO2 + 6% vol O2 flue gas containing an inlet Hg 0 concentration of 890.0 µ g·m −3 . As shown in Figure 7, the outlet concentration of Hg 0 gradually increased from 303.2 µ g·m −3 to 835.5 µ g·m −3 after 2400 min of scrubbing experiment. The Hg 0 adsorption capacity of c-CuS was calculated to be 736.5 mg·g −1 , and the average adsorption rate was 306.9 μg·g −1 ·min −1 . Figure 8 and Table S2 demonstrate the comparison of the adsorption capacity and rate of c-CuS with previously reported metal sulfides. c-CuS has the highest adsorption capacity and rate among the reported sulfide adsorbents. Compared to all other adsorbents, the Hg 0 adsorption capacity of c-CuS is exceptionally high, exceeding them by an impressive 277%, such as nano-CuS, Co3S4, S/FeS2, ZnS, and so on. Additionally, the Hg 0 adsorption rate of c-CuS is much higher than that of other mineral sulfides, which is due to the abundant adsorption sites under the positive effect of SO2 and O2 [7,14,[36][37][38][39][40].  In Figure S4, it could be observed that the capture of mercury was almost unaffected under an atmosphere of 3~9% vol O 2 . This was attributed to the fact that c-CuS solution could dissolve only 0.076 mmol·L −1 oxygen, which was much higher than the amount of mercury and much lower than the O 2 proportion in the flue gas. Therefore, we conducted further investigations to explore the influence of dissolved oxygen on Hg 0 oxidation. As displayed in Figure 6a, the dissolved oxygen concentration in c-CuS solution increased from 0.6 mg·L −1 to 7.6 mg·L −1 . The redox potential of the c-CuS solution gradually increased from 166 mV to 222 mV, which meant enhancing the oxidation ability of the scrubbing solution. It led to a significant increase in the removal efficiency of mercury from 45.0% to 95.3%. As shown in Figure 6b, the results displayed that the Cu LMM auger spectral peaks included Cu 2+ of binding energy at 568.1 eV and Cu + at 565.8 eV [35]. The Cu + spectral peak of c-CuS in oxygen-free water was visible, while it was weakened for c-CuS in oxic water, conversely indicating the oxidation of Cu + to Cu 2+ by O 2 . It could be inferred that dissolved oxygen in an aqueous solution could increase the Cu 2+ content of c-CuS and strengthen the oxidation of Hg 0 .    The Hg 0 adsorption capacity and rate of c-CuS were investigated and calculated by penetration experiments. The prepared 10 mL c-CuS (containing 0.48 mg CuS particles) was input with 6% vol SO 2 + 6% vol O 2 flue gas containing an inlet Hg 0 concentration of 890.0 µg·m −3 . As shown in Figure 7, the outlet concentration of Hg 0 gradually increased from 303.2 µg·m −3 to 835.5 µg·m −3 after 2400 min of scrubbing experiment. The Hg 0 adsorption capacity of c-CuS was calculated to be 736.5 mg·g −1 , and the average adsorption rate was 306.9 µg·g −1 ·min −1 . Figure 8 and Table S2 demonstrate the comparison of the adsorption capacity and rate of c-CuS with previously reported metal sulfides. c-CuS has the highest adsorption capacity and rate among the reported sulfide adsorbents. Compared to all other adsorbents, the Hg 0 adsorption capacity of c-CuS is exceptionally high, exceeding them by an impressive 277%, such as nano-CuS, Co 3 S 4 , S/FeS 2 , ZnS, and so on. Additionally, the Hg 0 adsorption rate of c-CuS is much higher than that of other mineral sulfides, which is due to the abundant adsorption sites under the positive effect of SO 2 and O 2 [7,14,[36][37][38][39][40].

Mechanism for Hg 0 Adsorption
As mentioned above, SO 2 and O 2 have positive effects on c-CuS for Hg 0 capture. To determine the mechanism of Hg 0 adsorption under SO 2 and O 2 atmospheres, as shown in Figure 9, the XPS spectra of spent c-CuS samples were analyzed under different atmospheres. Shown in Figure 9a-d are the S 2p peaks of samples under N 2 , 6% vol SO 2 , 6% vol SO 2 + 6% vol O 2 , and 6% vol SO 2 + 6% vol O 2 + Hg 0 , respectively. The forms of S on the c-CuS surface are multiple, such as tri-coordinated S 2− (CN=3) , tetra-coordinated S 2− (CN=4) , and sulfur-sulfur-coordinated S 2 2− [23]. After peak fitting, the peaks at 161.1 eV and 162.2 eV were assigned to S 2− (CN=3) , while the peaks at 161.9 eV and 163.0 eV were attributed to S 2− (CN=4) . The peaks at 163.2 eV and 164.4 eV were attributed to S 2 2− , while the peaks at 167.9 eV and 168.9 eV belonged to SO 4 2− [37,38,41]. Under N 2 atmosphere, the ratio of S and Cu total amount (S t , Cu t ) on c-CuS was found to be 1.01:1, with S 2p consisting of 26.8% S 2− (CN=3) , 39.0% S 2− (CN=4) , 33.5% S 2 2− , and 0.7% SO 4 2− . Upon treatment with 6% vol SO 2 , the S t :Cu t ratio increased to 1.79:1, with a decrease in S 2− (CN=3) content to 10.4%, an increase in S 2 2− to 49.5%, and a slight change in other forms of S, indicating the formation of new S 2 2− by the combination of S 2− (CN=3) with SO 2 . Additionally, the Cu 2p peak in Figure 10 shifted towards higher binding energy, indicating the reduction of a part of Cu 2+ to Cu + by SO 2 . Similarly, under 6% vol SO 2 + 6% vol O 2 conditions, S t :Cu t decreased with the oxidation of Cu + into Cu 2+ by O 2 and inhibition of S 2− (CN=4) combination with SO 2 adsorbed on the c-CuS surface. The S 2− (CN=3) content decreased with an increase in S 2− (CN=4) , which led to less S 2 2− . The further supply of Hg 0 resulted in a decrease in the ratio of S 2 2− and S 2− (CN=4) , but an increase in S 2− (CN=3) , indicating the transformation of S 2 2− and S 2− (CN=4) into S 2− (CN=3) after capture of Hg 0 . The results of Cu 2p shifting towards higher binding energy confirmed that Cu 2+ was also an active site for Hg 0 oxidation. It suggests that S 2 2− and Cu 2+ serve as oxidation sites, while S 2− (CN=3) is the binding site with Hg after adsorption. The TPD results in Figure 11 indicated that the decomposition temperature of Hg on the c-CuS was about 195~220 • C, which suggested the presence of black HgS [38].
The experimental results were inconclusive about the main binding site of the adsorbed mercury. Therefore, the E ad at each site on the c-CuS (110) crystal was analyzed at the molecular level using DFT calculations. The result in Figure 12 indicated that the tricoordinate sulfur sites had the highest adsorption energy of −125.23 kJ·mol −1 . It means that the tri-coordinate sulfur site might be the main binding site for mercury, which is consistent with XPS results. The experimental results were inconclusive about the main binding site of the adsorbed mercury. Therefore, the Ead at each site on the c-CuS (110) crystal was analyzed at the molecular level using DFT calculations. The result in Figure 12 indicated that the tricoordinate sulfur sites had the highest adsorption energy of −125.23 kJ·mol −1 . It means that the tri-coordinate sulfur site might be the main binding site for mercury, which is consistent with XPS results.  The experimental results were inconclusive about the main binding site of the adsorbed mercury. Therefore, the Ead at each site on the c-CuS (110) crystal was analyzed at the molecular level using DFT calculations. The result in Figure 12 indicated that the tricoordinate sulfur sites had the highest adsorption energy of −125.23 kJ·mol −1 . It means that the tri-coordinate sulfur site might be the main binding site for mercury, which is consistent with XPS results.     Figure  S4).

Conclusions
In this paper, the optimal conditions for preparing c-CuS with a Cu-S ratio of 1:1 and a concentration of 0.5 mmol·L −1 were synthesized by a double-jet liquid-phase sulfidation precipitation method. Defect-rich c-CuS with low crystallinity was used as a scrubbing agent to remove mercury from flue gas in wet scrubbing. It migrated the negative effect of SO 2 on mercury removal performance, while also enhancing mercury capture. The Hg 0 removal efficiency of c-CuS was 99.1% under 6% vol SO 2 + 6% vol O 2 atmosphere. The Hg 0 adsorption capacity of c-CuS reached 736.5 mg·g −1 , and the average adsorption rate was 306.9 µg·g −1 ·min −1 , which was far better than other reported metal sulfides adsorbents. Based on structural characterization and DFT calculation, it was found that SO 2 and O 2 can enhance the formation of Cu 2+ and S 2 2− sites from Cu + and S 2− (CN=3) to promote the oxidation of Hg 0 to Hg 2+ , and then Hg 2+ could strongly bind with S 2− (CN=3) . However, reusability of the material should be developed in future studies.