HCl-Induced Hg 0 Transformation over CuMn 2 O 4 Sorbent

CuMn 2 O 4 spinel has been regarded as a highly efficient sorbent for Hg 0 capture from flue gas. The regenerability and recyclability of CuMn 2 O 4 sorbent are mainly associated with the mercury speciation adsorbed on its surface. However, the effect mechanism of HCl on Hg 0 transformation over CuMn 2 O 4 sorbent is still elusive. Experiments were conducted to understand the effect of HCl on Hg 0 transformation over CuMn 2 O 4 sorbent. The results indicate that CuMn 2 O 4 sorbent is a mesoporous material and possesses a good thermal stability. CuMn 2 O 4 shows >95% Hg 0 removal efficiency in a wide temperature window of 50–350 ◦ C. The favorable electron-transfer environment caused by the mixed valence states of Cu and Mn cations is responsible for the excellent Hg 0 removal performance of CuMn 2 O 4 sorbent. CuMn 2 O 4 shows a higher Hg 0 adsorption capacity of 4774.57 μ g/g. Hg 0 adsorption process over CuMn 2 O 4 sorbent can be well described by the developed kinetic model. Hg 0 removal efficiency of CuMn 2 O 4 sorbent does not depend on the presence of HCl. Mercury species adsorbed on the CuMn 2 O 4 sorbent in the presence of HCl mainly exist in the forms of HgO and HgCl 2 O 8 · H 2 O. HCl shows a significant effect on mercury speciation over CuMn 2 O 4 sorbent. Most of HgO species will be transformed into HgCl 2 O 8 · H 2 O in the presence of HCl.


Introduction
Mercury is a toxic air pollutant with high volatility, and is harmful to human health and ecosystem [1][2][3][4][5]. Thermal power plant is considered to be a main source of gaseous anthropogenic mercury emissions in China [6][7][8]. The Minamata Convention on Mercury has officially come into force in August 2017 to restrict the global anthropogenic mercury emissions [9]. In order to meet the requirements of the Minamata Convention on Mercury, the mercury emission limit of thermal power plants in China will be set to 1 μg/m 3 in 2030 [10]. Therefore, mercury emission control in thermal power plants becomes increasingly urgent.
To date, mercury emission control technologies of thermal power plants mainly include: sorbent injection [11][12][13][14][15][16], catalytic oxidation [17][18][19][20] and bromide addition [21][22][23]. The technical strategy of catalytic oxidation and bromide addition is to oxidize Hg 0 into Hg 2+ , which is subsequently removed by the wet flue gas desulfurization equipment. However, Hg 2+ may be reduced to Hg 0 in the wet desulfurization solution [24], leading to the secondary mercury pollution [25]. Therefore, sorbent injection is considered as an effective mercury emission control technology [26,27]. Currently, bromine-modified activated carbon is the commercial mercury sorbent. However, there are some problems during the application of activated carbon [28], such as high price, adverse effects on the quality of fly ash as the cement raw material, and non-recyclable utilization. Therefore, it is necessary to develop the cost-effective and regenerable mercury sorbents.
Currently, copper-manganese spinel oxide has received more and more attention due to its unique structure and physical-chemical properties [29]. Copper-manganese oxide has been widely used in the field of catalysis, such as CO oxidation [30], water-gas shift reaction [31], and pollutant removal [32]. The excellent catalytic activity of copper-manganese oxide is closely related to the flexible valence states of Cu +/2+ and Mn 2+/3+/4+ [33]. The valence state changes of copper and manganese cations easily lead to an electron-transfer environment [34], which is beneficial to the adsorption and oxidation of Hg 0 . Therefore, our previous work synthesized copper-manganese spinel-type oxide and used to remove Hg 0 from simulated flue gas (4% O 2 + 12% CO 2 + N 2 ) [35][36][37]. HCl usually exists in flue gas, and has a great influence on the mercury transformation [38]. However, to date, no relevant research has been reported to investigate the effects of HCl on Hg 0 transformation over CuMn 2 O 4 sorbent.
In this work, the thermogravimetric (TG) analysis, BET specific surface area, and pore size distribution were used to evaluate the thermal stability and textural properties of CuMn 2 O 4 sorbent. The effect of HCl on Hg 0 transformation over CuMn 2 O 4 sorbent was systematically investigated. The related effect mechanism was clarified based on the temperature-programmed desorption (TPD) analysis. This study can provide a theoretical basis for the application of CuMn 2 O 4 sorbent.

Sorbent Synthesis and Characterization
CuMn 2 O 4 sorbent was synthesized by our previously proposed low-temperature sol-gel autocombustion synthesis method [35]. The detailed synthesis procedure has been provided in our previous work [35]. The thermal stability of sorbent was evaluated on a thermogravimetric analyzer (TA SDT-Q600). The temperature was increased from 25 to 850 • C in the air atmosphere at a heating rate of 10 • C/min. The flow rate of air was set as 100 mL/min. The textural properties (such as specific surface area and pore size distribution) were investigated on a N 2 physical adsorption apparatus (ASAP2020, Micromeritics). Sorbents were degassed at 200 • C for 2 h before the N 2 adsorption experiments.

Hg 0 Removal Experiments
The Hg 0 removal experiments were carried out in a fixed-bed reactor system, including a gas feed subsystem, a Hg 0 generation subsystem, a quartz reactor, and an online mercury analyzer (VM3000, Mercury Instruments, Germany). The simulated flue gas (including 4% O 2 , 12% CO 2 , 0-10 ppm HCl, and N 2 as the balance gas) were fed into the quartz reactor. In order to avoid the interference of other gas components (such as NO, SO 2 , and H 2 O) in the study of the effect mechanism of HCl on Hg 0 transformation, these gas components were not included in the simulated flue gas. The total flow rate of the simulated flue gas was 1 L/min. Stable Hg 0 vapor was continuously generated from a mercury permeation tube, and was fed into the simulated flue gas by N 2 . Hg 0 concentration in the flue gas was maintained at 90 μg/m 3 . 0.2 g of sorbent was mixed with 1.8 g of quartz sand to reduce the pressure drop of flue gas through the sorbent bed and to prevent pipeline blockage. The height of sorbent bed was 19 mm, and the gas hourly space velocity (GHSV) was 5 × 10 4 h −1 . Before the flue gas entering the mercury analyzer, the acid gas HCl was removed by a 10% NaOH solution to prevent the HCl corrosion of mercury measurement instrument. The experimental details have been clearly described in our previous work [35].

TGA Analysis
Thermogravimetric analysis was used to investigate the thermal stability of CuMn 2 O 4 sorbent. The TG and DTG curves of CuMn 2 O 4 spinel in air atmosphere are displayed in Fig. 1. It can be seen that the weight of sorbent decreases slightly when the temperature increases and the weight loss curve of CuMn 2 O 4 is flat. As the temperature increased from 25 to 350 • C (the maximum temperature of Hg 0 removal experiments), the weight loss was 0.61%, as shown in Fig. 1a. A small weight loss of 0.17% was obtained when the temperature further increased from 350 to 850 • C. Differential thermogravimetric (DTG) curve was used to calculate weight loss rates, as shown in Fig. 1b. CuMn 2 O 4 sorbent showed three endothermic peaks at 498, 583 and 825 • C. Moreover, two exothermic peaks were also observed at 527 and 628 • C. The largest weight loss rate of CuMn 2 O 4 (0.0084%/min) occurs at 583 • C. Obviously, these results indicate that CuMn 2 O 4 sorbent has a good thermal stability at high temperatures.

BET Surface Area and Pore Size Distribution
The BET surface area and pore size distribution were measured to understand the textural properties of CuMn 2 O 4 sorbent. The N 2 adsorption-desorption isotherms and pore size distribution of CuMn 2 O 4 sorbent is shown in Fig. 2. According to the physical adsorption isotherm classification proposed by the International Union of Pure and Applied Chemistry (IUPAC), the N 2 adsorption-desorption isotherm of CuMn 2 O 4 sorbent is the type IV curve (Fig. 2a). There is an obvious hysteresis loop in the relative pressure range of 0.60-0.95, which is the typical feature of mesopores. The pore size of CuMn 2 O 4 sorbent is mainly distributed in the range of 5-50 nm, suggesting the existence of mesoporous structure. The BET surface area, average pore size, and pore volume are 29.33 m 2 /g, 24.85 nm, and 0.18 cm 3 /g, respectively. Pores with a size between 2-50 nm are defined as mesopores [39]. Therefore, CuMn 2 O 4 sorbent is a mesoporous material, which is favorable for the mass transfer during Hg 0 capture.

Removal Efficiency
The mercury removal efficiency of CuMn 2 O 4 sorbent in the temperature range of 50-350 • C is shown in Fig. 3. It can be seen that CuMn 2 O 4 sorbent shows >95% Hg 0 removal efficiency over the wide temperature range. The excellent Hg 0 removal efficiency may be attributed to the high concentrations of surface Cu atom and chemisorption oxygen. The XPS analysis of our previous work [35] found that the surface Cu and Mn atoms exist in the form of flexible valence (Cu + , Cu 2+ , Mn 2+ , Mn 3+ , and Mn 4+ ). A part of Cu + cations occupied the tetrahedral position of CuMn 2 O 4 spinel, and some Mn 4+ cations were located at the octahedral position of CuMn 2 O 4 spinel. Thus, the different valence states and position distributions of metal cations (Cu + , Cu 2+ , Mn 2+ , Mn 3+ and Mn 4+ ) leaded to lattice distortion of CuMn 2 O 4 spinel, resulting in an electrontransfer environment which can accelerate Hg 0 adsorption and oxidation over CuMn 2 O 4 sorbent. Therefore, CuMn 2 O 4 sorbent exhibits excellent Hg 0 removal performance in a wide temperature window.

Adsorption Kinetics
A mathematical model is an effective tool to predict the Hg 0 adsorption capacity of sorbent and to evaluate the performance of sorbent for mercury removal under different operating conditions [9]. The mathematical model can provide a rational basis for describing and characterizing the effectiveness of sorbent injection for mercury removal. In addition, the mathematical model can provide theoretical guidance for the development of new sorbents and for the optimization of sorbent injection process. Therefore, it is necessary to study the adsorption kinetics of mercury on the sorbent surface. In a fixed-bed reactor, as the flue gas flows through a constant-temperature sorbent bed, the differential mass conservation equation of mercury can be described as: where u represents the flow velocity of flue gas. C represents the Hg 0 concentration in the flue gas. ε represents the porosity of sorbent bed. q represents the Hg 0 adsorption capacity of sorbent. D ax represents the axial diffusion coefficient. z represents the axial coordinate. t represents time.
The differential mass conservation equation is closely related to the constitutive equation of Hg 0 adsorption rate on the sorbent.
It was reported that the axial diffusion of mercury can be ignored if the ratio of the sorbent bed diameter (d bed ) to the sorbent particle diameter (d particle ) is greater than 35 and the ratio of the sorbent bed thickness (h bed ) to the sorbent particle diameter is greater than 75 [27]. In this work, d bed /d particle and h bed /d particle are 120 and 253, respectively. In addition, the Hg 0 removal experiments were conducted under the condition of a larger flue gas flow rate of 1000 mL/min. These indicated that Hg 0 removal process of CuMn 2 O 4 sorbent is controlled by reaction kinetics. The chemisorption mechanism is responsible for Hg 0 adsorption over CuMn 2 O 4 sorbent. The chemical adsorption reaction has been incorporated into the mathematical model (i.e., Langmuir isotherm equation) of adsorption kinetics. Therefore, the mass diffusion resistance of mercury in the sorbent particle and the external boundary layer can be ignored. The Langmuir isotherm equation can be used to describe the adsorption process of Hg 0 on the CuMn 2 O 4 surface: where k a represents the rate constant of Hg 0 adsorption on the sorbent surface. C in represents the stable Hg 0 concentration in the flue gas at the inlet of the fixed bed reactor. q e represents the equilibrium Hg 0 adsorption capacity of CuMn 2 O 4 sorbent. q represents the Hg 0 adsorption capacity at time t. Integrating Eq. (2), q can be expressed as: k 1 is used to replace k a C in , Eq. (3) can be expressed as: The increment of Hg 0 adsorption amount on the sorbent surface within the time range of t can be calculated by the following formula: where Q represents the flow rate of flue gas. m represents the weight of sorbent. C out (t) represents the Hg 0 concentration in the flue gas at the reactor outlet at time t. During the whole Hg 0 adsorption experiment, Hg 0 adsorption amount of sorbent can be expressed as: In this work, Hg 0 adsorption experiments were conducted for 2 h. Therefore, the integration time of Eq. (6) is 2 h. Eq. (4) was used to fit the experimental data at 150 • C, as shown in Fig. 4. The adsorption kinetic parameters of Eq. (3) can be obtained: q e = 4774.57 μg/g, k a = 9.81 × 10 −7 m 3 /(μg · min). Therefore, CuMn 2 O 4 sorbent has a higher Hg 0 adsorption capacity. Hg 0 adsorption kinetics on the CuMn 2 O 4 surface can be expressed as: In order to further assess the accuracy and reliability of the adsorption kinetic model, Eq. (7) was used to predict the experimental results under the conditions of different reaction temperature. The comparison between the prediction results and experimental results at the reaction temperatures of 50, 100, 200, 250, 300 and 350 • C is shown in Fig. 5. It can be seen that model predictions are in good agreement with the experimental results. Therefore, the kinetic model can accurately predict the Hg 0 adsorption amount of CuMn 2 O 4 sorbent in a wide reaction temperature range.

Effect of HCl
HCl is the most important component for Hg 0 oxidation in flue gas [40]. Therefore, it is necessary to investigate the effect of HCl on Hg 0 transformation over CuMn 2 O 4 sorbent. 10 ppm HCl was added to the simulated flue gas, Hg 0 removal efficiency of CuMn 2 O 4 sorbent at different reaction temperatures is shown in Fig. 6. In general, HCl shows little effect on the Hg 0 removal efficiency of CuMn 2 O 4 sorbent. It is well known that HCl can promote Hg 0 adsorption and oxidation in flue gas [41]. However, CuMn 2 O 4 sorbent can show >95% Hg 0 removal efficiency in the absence of HCl. Therefore, Hg 0 removal efficiency of CuMn 2 O 4 sorbent does not depend on the presence of HCl. Since CuMn 2 O 4 sorbent shows excellent Hg 0 capture capacity and its Hg 0 removal efficiency does not depend on the presence of HCl, HCl has almost no effect on Hg 0 removal by CuMn 2 O 4 sorbent. The slightly inhibitory effect (∼1.2%) at 200 • C and promotional effect (∼1.3%) at 300 • C of Fig. 6 are attributed to the experimental error rather than the effects of HCl. Although HCl has little effect on Hg 0 removal efficiency of CuMn 2 O 4 sorbent, HCl may affect the mercury speciation on the CuMn 2 O 4 surface.
To further explore the effect of HCl on the mercury speciation over the CuMn 2 O 4 surface, TPD experiments were carried out to identify the mercury speciation. The TPD curves of the used CuMn 2 O 4 sorbent after Hg 0 removal experiments in the absence and presence of HCl are shown in Fig. 7. After Hg 0 removal experiments in the absence of HCl, the used CuMn 2 O 4 sorbent only showed a peak at 385 • C (Fig. 7a). This peak is attributed to the decomposition of yellow HgO species [42].  (Fig. 7b). The TPD peak at 330 • C is attributed to the decomposition of red HgO species adsorbed on the sorbent surface. Yellow HgO and red HgO have different crystal structure, leading to the different decomposition temperature. The desorption peak at 470 • C is ascribed to the decomposition of HgCl 2 O 8 ·H 2 O species [42]. XPS analysis was further applied to illustrate the effect of HCl on the Hg 0 transformation on CuMn 2 O 4 sorbent surface. It can be seen in Fig. 8 that the peak of Hg 4f XPS spectra is observed at 104.61 eV. This is characteristic of Hg 4f 5/2 spin-orbit doublet of Hg 2+ . This suggests that Hg 0 is chemisorbed and transformed into oxidized mercury species (Hg-Cl compounds) on CuMn 2 O 4 sorbent surface. Therefore, it can be inferred that Hg-Cl compounds are formed on CuMn 2 O 4 sorbent in the presence of HCl.
The formation of HgCl 2 O 8 ·H 2 O species is governed by Eq. (10). The area of HgCl 2 O 8 ·H 2 O desorption peak is much larger than that of HgO desorption peak, indicating that HgCl 2 O 8 ·H 2 O is the main mercury specie on the used CuMn 2 O 4 sorbent surface. Based on the above TDP analysis, it can be found that the presence of HCl shows a larger effect on the mercury speciation over the CuMn 2 O 4 sorbent surface.

Conclusions
Experiments were conducted to investigate the effect mechanism of HCl on Hg 0 transformation over CuMn 2 O 4 sorbent. CuMn 2 O 4 sorbent is a mesoporous material, and has a good thermal stability at high temperatures during Hg 0 removal. The mesopores are favorable for mass transfer during Hg 0 capture. CuMn 2 O 4 sorbent shows >95% Hg 0 removal efficiency in the wide temperature range of 50-350 • C. The excellent Hg 0 removal performance of CuMn 2 O 4 sorbent is closely associated with the electron-transfer environment caused by the mixed valence states of Cu and Mn cations. The equilibrium adsorption capacity and adsorption rate constant are 4774.57 μg/g and 9.81