Xiaopeng Zhang,Cheng Gao,Ziwei Wang,Ximiao Wang,Jie Cheng,Xinxin Song,Xiangkai Han,Ning Zhang,Junjiang Bao,Gaohong He*

State Key Laboratory of Fine Chemicals,Dalian University of Technology,Dalian 116023,China School of Chemical Engineering,Dalian University of Technology,Panjin 124221,China

Keywords:Wood vessel Elemental mercury Ce3O4 Ordered pore structure

ABSTRACT Catalytic oxidation of Hg0 to HgO is an efficient way to remove Hg0 from coal-fired flue gas.The catalyst with ordered pore structure can lower mass transfer resistance resulting in higher Hg0 oxidation efficiency.Therefore,in the present work,wood vessels were used as sacrificial template to obtain Co3O4 with ordered pore structure.SEM and BET results show that,when the mass concentrations of Co(NO3)2·6H2O was 20%,the obtained catalyst (Co3O4 [20%Co(NO3)2]) possesses better pore structure and higher surface area.It will expose more available surface active sites and lower the mass transfer resistance.Furthermore,XPS results prove that Co3O4 [20%Co(NO3)2] has the highest ratio of chemisorbed oxygen which plays an important role in Hg0 oxidation process.These results lead to a better Hg0 oxidation efficiency of Co3O4 [20%Co(NO3)2],which is about 90% in the temperature range of 200 to 350 °C.Furthermore,Co3O4 [20%Co(NO3)2] has a stable catalytic activity,and its Hg0 oxidation efficiency maintains above 90% at 250 °C even after 90 h test.A probable reaction mechanism is deduced by the XPS results of the fresh,used and regenerated catalyst of Co3O4[20%Co(NO3)2].Chemisorbed oxygen can react with Hg0 forming HgO with the reduction of Co3+to Co2+.And lattice oxygen and gaseous oxygen can supplement the consumption of chemisorbed oxygen to oxidize Co2+ to Co3+.

1.Introduction

Due to its high mobility,high toxicity,persistence and bioaccumulation,mercury is extremely harmful to the human being and the environment,which has attracted the wide attention of researchers [1].Coal-fired boilers are the largest source of anthropogenic mercury emissions [2].Mercury in flue gas is mainly presented as three forms,elemental mercury(Hg0),oxidized mercury(Hg2+)and particulate-bound mercury(Hgp)[3].Among them,Hg0is the most difficult to remove due to its high volatility and insolubility in water,which is the key for mercury removal [4].

Activated carbon injection is a commonly used technology in industry to remove Hg0[5].However,activated carbon has a low adsorption capacity and is difficult to be regenerated.In addition,the activated carbon is mixed with fly ash after Hg0adsorption,which reduces the industrial value of fly ash[6].Catalytic oxidation of Hg0to HgO with a high economic efficiency and catalytic activity is considered to be a promising method for Hg0removal [7,8].Many transition metal oxides,such as Co3O4[9-11],TiO2[12-14]and FeOx[15-17],etc.have obtained good Hg0oxidation efficiency.Co has a special redox system of Co3+/Co2+and electrons shift very quickly between redox pair,which leads to a higher redox performance and a better Hg0oxidation performance[18].Zhu[19]et al.prepared Co3O4nanosheets catalyst and found that it showed an excellent Hg0oxidation efficiency over a wide temperature range of 100-350 °C.

The catalytic oxidation of Hg0belongs to a gas-solid catalytic reaction.Therefore,the mass transfer of Hg0from gas phase to the catalyst pores inside and its adsorption on the catalyst surface are very important for Hg0oxidation process.Traditional Co3O4-based catalysts usually have irregular pore channels and many blind holes,which results in large mass transfer resistance and limits the Hg0oxidation activity [20].Loading Co3O4on the porous support can reduce the mass transfer resistance to a certain extent.However,the pore blocking phenomenon is still serious when the load amount is high,which increases the mass transfer resistance.On the contrary,when the load is low,fewer surface active sites are generated resulting in lower Hg0oxidation activity.Therefore,the establishment of straight-through channel inside of Co3O4can effectively reduce the mass transfer resistance and provide more surface active sites to achieve a high Hg0oxidation efficiency.The wood vessel has a natural orderly and straight-through channel structure.And there are many micropores on the vessel wall to connect the straight-through channel forming a 3D connected channel structure.Our previous study has shown that the construction of Co3O4nanoparticles inside the wood vessel can reduce the mass transfer resistance of the Hg0removal process and obtain a higher Hg0removal activity [9].However,biochar has a poor thermal stability which is not suitable for working conditions with higher temperatures.If wood vessel can be used as templates to construct Co3O4catalyst with a 3D ordered and connected pore structure,it may reduce the mass transfer resistance and provide more surface active sites resulting in a better Hg0oxidation efficiency.

Inspired by this,in the present work,wood vessel of cedarwood,which is the widely distributed and fast-growing plant,was originally used as a sacrifice template.And the Co3O4precursor solution was impregnated into its internal vessel pores.After carbonization,the wood carrier is removed.The prepared Co3O4catalyst has a 3D ordered and connected pore structure with a Hg0oxidation efficiency of about 90% in the temperature range of 200-350 °C.

2.Experimental

2.1.Catalyst preparation

The wood sample used in the present work is collected from cedarwood,which is a fast-growing and widely distributed plant in China.All of the chemical reagents are of analytical grade.

The wood collected from cedarwood is cut into small pieces with a size of 5 mm×3 mm×2 mm.2 g of obtained wood samples was placed in a water bath at 90 °C for 6 h and then dried in an oven at 60 °C for 12 h.Appropriate amount of Co(NO3)2·6H2O was dissolved into deionized water to prepare cobalt nitrate solutions with mass concentrations of 10%,20%,30%,and 40%.The wood samples were impregnated into the above solution kept in 30 °C water bath for 24 h.After that,the samples were dried in an oven at 60 °C for 24 h and then carbonized at 500 °C at 2 °C·min-1in a tube furnace for 3 h under the N2atmosphere.Finally,the samples were calcined in a muffle furnace at 450 °C for 4 h with a heating rate of 5 °C·min-1under air atmosphere to remove the wood template.The obtained catalysts are denoted as Co3O4[x%Co(NO3)2],where × is the concentration of cobalt nitrate solution.

2.2.Characterization

The scanning electron microscopy (SEM)and energy dispersive Sspectrometer (EDS) was performed on S-4000 (Hitachi,Japan) to determine the morphology,microstructure and element distribution of the samples.

The BET surface areas and pore size distribution are measured by N2adsorption/desorption on autosorb IQ automated gas sorption system (Quantachrome Intenturments,USA).The pore size and pore volume were obtained by using the desorption branch of N2adsorption/desorption isotherm and Barrett-Joyner-Halenda formula.

X-ray diffractogram measurements were carried out on XRD-7000 S system using Cu Kα radiation (40 kV,100 mA) (SHIMADZU Corporation,Japan).The scanning range was from 15°to 90°with a step size of 5 (°)·min-1.

X-ray photoelectron spectroscopy (XPS) measurements were carried out by using an ESCALAB250(Thermo Fisher Scientific Corporation,USA) with monochromatic Al Kα radiation.The surface charging effects were eliminated by the C 1s binding energy value of 284.6 eV.

2.3.Experimental setup

The reaction system is shown in Fig.1.Simulate flue gas contains 50 μg·m-3Hg0,6% (vol) O2and balanced with N2.The flue gas flow rate is 600 ml·min-1and the catalyst loading amount is 0.2 ml (about 0.05 g) resulting in a gas hourly space velocity of about 180000 h-1.A mercury permeation tube is placed in a Ushaped tube and the U-shaped tube is put into a 38 °C water bath to generate mercury vapor.And then the mercury vapor is carried into the flue gas by the carrier gas N2.In each test,the Hg0concentration at the inlet of the reactor is stable for 1 h before it is passed through the reactor.In order to eliminate the effect of adsorption on catalytic activity,the catalyst reaches adsorption saturation in a mercury atmosphere before the oxidation efficiency test.The Hg0concentration is tested by an online mercury analyzer (VM-3000,Mercury Instruments,Germany).The Hg0oxidation efficiency is defined as:

whereEoxistands for Hg0oxidation efficiency,Hginand Hgoutstand for Hg0concentration of the inlet and outlet of the reactor.

3.Results and Discussion

3.1.Catalytic performance

The Hg0oxidation efficiency of the catalysts at different temperatures is shown in Fig.2.The Hg0oxidation efficiency over the catalysts presents a trend of first rising and then falling with the increase of temperature.All of the catalysts have good catalytic performance in the temperature range of 150-300 °C suggesting a wide temperature window.Especially,for Co3O4[20%Co(NO3)2]and Co3O4[30%Co(NO3)2],the Hg0oxidation efficiency is above 78% in the temperature range of 150-350 °C.In contrast,Co3O4[10%Co(NO3)2] and Co3O4[40%Co(NO3)2] have lower catalytic activity in high and low temperature regions,respectively.Combining Hg0oxidation efficiency and temperature window,Co3O4[20%Co(NO3)2] has the best catalytic performance with a Hg0oxidation efficiency of 95% at 250 °C.

The effects of flue gas components such as O2,NO,SO2and H2O on Hg0oxidation of Co3O4[20%Co(NO3)2] were investigated at 200 °C.As shown in Fig.3,O2has important effects on Hg0oxidation efficiency.When the O2concentration increases from 3%to 6%,the Hg0oxidation efficiency increases from 75% to 88%,which shows that O2plays a positive role in the Hg0oxidation process.According to previous studies,gaseous O2can continuously supplement the consumed chemisorbed oxygen,which can maintain the whole oxidation cycle of Hg0[21].

NO can enhance the Hg0oxidation efficiency.After 100 × 10-6and 200×10-6(volume ratio)NO addition,the Hg0oxidation efficiency increases from 88%to 91%and 93%,respectively.NO can be oxidized to NO2by surface chemisorbed oxygen,and then NO2will react with oxygen and Hg0forming Hg(NO3)2resulting in a higher Hg0oxidation efficiency [22].

It is obvious that SO2has serious inhibitory effects on the Hg0oxidation process.When 100×10-6(volume ratio)of SO2is added into the flue gas,the Hg0oxidation efficiency declines from 88%to 76%,and it declines further from 76%to 53%when 200×10-6(volume ratio)of SO2is added.The decrease of Hg0oxidation efficiency may be due to the competitive adsorption between SO2and Hg0,which can poison the active centers [23,24].

Fig.1.Schematic diagram of the fix-bed experiment system.

Fig.2.Hg0 oxidation efficiency of the catalysts at different temperatures.

Fig.3.Effects of individual flue gas components on Hg0 oxidation efficiency over Co3O4 [20%Co(NO3)2].

When H2O is added into the flue gas,the oxidation efficiency of Co3O4[20%Co(NO3)2] decreases significantly.This result is mainly due to the competitive adsorption of Hg0and H2O [25].

Co3O4[20%Co(NO3)2] was selected for a long-term test to evaluate the stability of the catalyst due to its higher catalytic activity.As shown in Fig.4,the Hg0oxidation efficiency of the catalyst remains stable at above 95% during the first 70 h.After that,the catalytic activity slowly decreases,and the Hg0oxidation efficiency is still maintained more than 83%after 100 h test.After a 2 h in-situ thermal treatment at 500 °C under N2atmosphere,the catalytic performance of the catalyst is recovered to the initial level.The above results show that Co3O4[20%Co(NO3)2] has good stability and recyclability.

3.2.Characterization

3.2.1.Physical properties

Fig.4.A relatively long time activity test for Co3O4 [20%Co(NO3)2] at 250 °C.

Fig.5.SEM (a1,a2,b1,b2,c1,c2,d1 and d2) and EDS (a3,b3,c3 and d3) images of Co3O4 [10%Co(NO3)2] (a),Co3O4 [20%Co(NO3)2] (b),Co3O4 [30%Co(NO3)2] (c),Co3O4 [40%Co(NO3)2] (d).

SEM was used to investigate the morphology of the catalysts.As shown in Fig.5,all of the catalysts have an orderly pore structure,which is similar to the pore structure of the wood vessel(shown in Fig.S1,Supplementary Material).It indicates that the catalysts well remain the pore structure of the wood vessel.The pore wall of Co3O4[10%Co(NO3)2] is thin and dense,which may be related to the less impregnation amount of the Co precursor during the preparation.As the concentration of the Co precursor solution increases,the pore walls of the catalysts become thicker,fluffy and porous.For Co3O4[40%Co(NO3)2],the pore diameter is significantly smaller than that of the other catalysts.These results indicate that the amount of Co precursor impregnation can significantly affect the catalyst structure.Co3O4[20%Co(NO3)2]has a suitable amount of impregnation,a large pore diameter,and a porous structure on the pore wall.It will reduce mass transfer resistance and provide more surface active sites,which is beneficial to Hg0oxidation process[26].The EDS images show that no C signal can be observed in four samples.It confirms that all WB has been burned out during calcination process in air atmosphere.

Fig.6.XRD pattern of the catalysts.

XRD was used to investigate the crystal phase and crystallinity of the catalysts.As shown in Fig.6,five peaks around 2θ=31.27°,36.85°,44.81°,59.36° and 65.24° can be detected in all of the catalysts which are corresponding to the diffraction of(2 2 0),(3 1 1),(4 0 0),(5 1 1) and (4 4 0) in Co3O4(PDF 42-1467).It proves that the CoOxin the catalyst exists as the crystalline phase of Co3O4.The FWHM of the strongest diffraction peaks of the four catalysts Co3O4[10%Co(NO3)2],Co3O4[20%Co(NO3)2],Co3O4[30%Co(NO3)2],and Co3O4[40%Co(NO3)2]are 0.484,0.374,0.311 and 0.332,respectively.The crystal grain sizes calculated by the Scherrer Equation are 17.8,22.9,27.6 and 25.9 nm,respectively.Previous studies have shown that the lower the crystallinity is corresponding to the smaller the grain size.It will give a higher degree of dispersion,leading to more surface active sites,which is beneficial to the Hg0oxidation process [27].

N2adsorption/desorption was used to investigate the specific surface area and pore structure of different catalysts.As shown in Fig.7,H3hysteresis loops appeared in all catalysts,indicating the presence of mesopores in the samples,which is consistent with the pore size distribution curves in Fig.8.Mesoporous structure is beneficial to mass transfer in the reaction process.In addition,micropore structures can be observed in all catalysts from Fig.8,which is conducive to increase the specific surface area.The specific surface area,pore volume and average pore diameter of the samples are shown in Table 1.It can be seen that Co3O4[20%Co(NO3)2]has the largest specific surface area,pore volume and average pore diameter.A larger specific surface area can expose more surface active sites[28].The above results are consistent with SEM.

Fig.7.N2 adsorption/desorption isotherms of the catalysts.

Fig.8.Pore size distribution curves of the catalysts.

Table 1 BET surface area and pore volume of the catalysts

3.2.2.Chemical properties

XPS was performed to investigate the oxidation state and atomic concentration ratio of the surface elements and the results are shown in Fig.9 and Table 2.As shown in Fig.9(a),both of Co3+and Co2+are observed in all of the catalysts.The peaks at 779.75 eV and 794.65 eV are assigned to Co3+species,while those at 781.55 eV and 796.3 eV are ascribed to Co2+[29].The O 1s spectrum can be fitted into two different overlapping peaks at 529.75 eV and 530.7 eV originating from lattice oxygen (denoted as Oα) and chemisorbed oxygen (denoted as Oβ) [30].As shown in Table 2,Co3O4[20%Co(NO3)2] has the highest ratio of Co3+/Co2+and Oβ/Oα.Previous studies have shown that Co3+can produce more oxygen vacancy and improve the oxidation capacity of thecatalyst [31].Oβas the active site in Hg0oxidation process can react with Hg0forming HgO [32].

Table 2 Surface atom concentration of Co and O

Table 3 Atomic concentration on catalyst surface

3.2.3.Structure-activity relationship

The oxidation performance of the catalysts is closely related to their structures and physical properties.SEM results elucidate that the catalyst remains the 3D ordered and connected pore structure of wood vessels.Co3O4[20%Co(NO3)2] catalyst has a large amount of pore on its channel wall.At the same time,XRD and BET results show that Co3O4[20%Co(NO3)2] has low crystallinity and large specific surface area,which provides more surface active sites for Hg0oxidation.Furthermore,XPS results show that Co3O4[20%Co(NO3)2] has a higher concentration of chemisorbed oxygen which as the main active site in the process of Hg0oxidation can easily oxidize Hg0into HgO.These results are the main reasons for the higher Hg0oxidation efficiency of Co3O4[20%Co(NO3)2].It should be noted that,Co3O4[10%Co(NO3)2]has a better Hg0oxidation efficiency at lower temperature.It may be due to that,at lower temperature,physical absorption of Hg0can hardly transformed into chemisorption of Hg0,and then lower active sites can participate in Hg0oxidation reaction.In this condition,Hg0oxidation will be mainly controlled by diffusion process.SEM results show that Co3O4[10%Co(NO3)2] has an even larger pore size compared with Co3O4[20%Co(NO3)2],which will effectively decrease the mass transfer resistance leading to a higher catalytic performance.When the temperature goes high,more active sites can participate in Hg0oxidation reaction,Co3O4[20%Co(NO3)2] will have a better catalytic performance.

3.3.Hg0 oxidation mechanism

Fig.9.XPS spectra of the catalysts Co 2p (a) and O 1s (b).

Fig.10.XPS spectra of fresh,used and regenerated Co3O4 [20%Co(NO3)2] 2p (a) and O 1s (b).

In order to reveal the reaction mechanism,XPS was used to characterize the oxidation state and atomic concentrations of the elements on the fresh,used and regenerated Co3O4[20%Co(NO3)2] surface.As shown in Fig.10 and Table 3,the intensity of the peak attributed to Co3+and Oα decreases after Hg0oxidation process.Compared with the fresh catalyst,the ratio of Co3+/Co2+for the used catalyst decreases from 1.05 to 0.837,indicating that Co3+participates in the Hg0oxidation process by converting to Co2+.The ratio of Oβ/Oα increases from 1.15 to 1.58,which may be due to that some Oα near the surface convert to Oβin the Hg0oxidation process to supplement the consumption of Oβ.As has been reported,Oβas a kind of active site can react with the adsorbed Hg0forming HgO with the transform of Co3+to Co2+[33].And then gaseous oxygen will replenish the consumed Oβand reoxidize the catalyst surface.With accumulation of HgO,gaseous oxygen can not efficiently transfer to catalyst surface and Oα will convert to Oβresulting in the decrease of Co3+/Co2+.After regeneration,Co3+/Co2+ratio and Oα concentration increase.It is due to the decomposition of HgOviaHgO →Hg+O,which can give active oxygen to oxidize Co2+to Co3+and the active oxygen will convert to Oα.Based on the above analysis,a possible reaction pathway of Hg0oxidation process is deduced.Firstly,the gaseous Hg0is adsorbed on the surface of catalyst forming adsorbed Hg0(-ad)(1).Subsequently,it reacts with Oβto form HgO(ad)(2).Finally,the adsorbed HgO shift to gas-phase HgO and the catalyst surface is reoxidized,the Oβis supplemented by gaseous oxygen and lattice oxygen (3,4) .The reaction pathway of Hg0oxidation can be described as follows:

4.Conclusions

In the present work,the catalyst with ordered pore structure was derived from cedarwood vessels.Co3O4[20%Co(NO3)2] has better pore structure and a large surface area leading to more available surface active sites and low mass transfer resistance.Furthermore,Co3O4[20%Co(NO3)2]has a higher ratio of Co3+/Co2+and Oβ/Oα,which empower it a good redox property.As a consequence,Co3O4[20%Co(NO3)2] shows the best Hg0oxidation performance with a Hg0oxidation efficiency of 95% at 250 °C.XPS results of the fresh,used and regenerated catalyst demonstrated that Oβis the active sites and it can be generated from gaseous oxygen and also Oα.With Hg0oxidation,the ratio of Co3+/Co2+decreases because of the reduction of the catalyst surface,but it increases after regeneration process due to the oxygen generated from HgO decomposition.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

We gratefully acknowledge the financial supports from the National Natural Science Foundation of China(51978124),Science Fund for Creative Research Groups of the National Natural Science Foundation of China (22021005),and the Cheung Kong Scholars Programme of China (T2012049).

Supplementary Material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.cjche.2022.06.018.