Electrocatalysis (2017) 8:103–114 DOI 10.1007/s12678-016-0346-6

ORIGINAL RESEARCH

Performance and Mechanism of In Situ Electro-Catalytic Flue Gas Desulfurization via Carbon Black-Based Gas Diffusion Electrodes Doped with MWCNTs Ze Chen 1 & Heng Dong 1 & Hongbing Yu 1 & Han Yu 1 & Min Zhao 1 & Xi Zhang 1

Published online: 29 November 2016 # Springer Science+Business Media New York 2016

Abstract Flue gas desulfurization (FGD) based on electrochemical technologies is efficient and environment-friendly. However, the low dissolved oxygen concentration, insufficient active sites, and high cost of electrode materials are still three bottlenecks needed to be solved. In this work, a series of novel carbon black-based gas diffusion electrodes (GDEs) doped with different multi-walled carbon nanotube (MWCNT) amounts were developed to generate H2O2 via oxygen reduction reaction (ORR) for in situ FGD. The influence of current density and electrolyte concentration for desulfurization performance was investigated. As a result, the GDEs doped with 100 mg MWCNTs gave the highest desulfurization efficiency (98.0%) and the lowest energy consumption (1.7 kW h kg−1) owing to its good balance on the porous structure, conductivity, and ORR activity. An appropriate current density (3.54 mA cm−2) and electrolyte concentration (50 mM) are conducive to the two-electron ORR and desulfurization. While the excessive values would cause some side reactions and the occupations of active sites by the SO42−. This compact system exhibits advantages of high desulfurization efficiency and low-energy consumption with no

Electronic supplementary material The online version of this article (doi:10.1007/s12678-016-0346-6) contains supplementary material, which is available to authorized users. * Heng Dong [email protected]

1

MOE Key Laboratory of Pollution Processes and Environmental Criteria, Tianjin Engineering Centre for Cleaner Technology of Iron-Steel Industry, College of Environmental Science and Engineering, Nankai University, No. 38 Tongyan Road, Jinnan District, Tianjin 300350, China

secondary pollution, providing an alternative way for in situ FGD in industrial application. Keywords Electro-catalytic . Flue gas desulfurization . Gas diffusion electrode . Multi-walled carbon nanotubes . Carbon black . Oxygen reduction reaction

Introduction Sulfur dioxide (SO2) in the flue gas is generated from the combustion of fossil fuel in thermal power plants, chemical plant, etc. [1, 2], which leads to the global environmental problems such as haze, photochemical smog, and acid rain [3–6]. The control of SO2 emission at its sources has therefore become a critical issue. Over the last decades, lots of efforts have been made on the flue gas desulfurization (FGD) technology including dry flue gas desulfurization (DFGD), semidry flue gas desulfurization (SDFGD), and wet flue gas desulfurization (WFGD) [7]. Although these conventional FGD methods have been applied extensively with mature equipments, high removal efficiency and good operating stability; stubborn characteristics including complex system, high investment, and operation cost; and undesired secondary pollution make them labored to support harmonious development between economics and environment [1, 8–11]. In this case, an efficient, energy-saving, and environment-friendly FGD technology is needed urgently. Electro-chemical flue gas desulfurization (EFGD) has been gradually developed in recent years, which utilizes electron as a clean reagent for SO2 removal and presents advantages in desulfurization efficiency and energy saving with no secondary pollution [10–12]. In this technology, SO2 is absorbed from the gas phase into the liquid electrolyte then directly oxidized on the anode of the electrolytic cell or indirectly

104

oxidized or reduced via a redox mediator [13, 14]. For example, SO2 can be anodically oxidized to SO42− in aqueous solution by adsorbed O2 species [15], and electrode materials such as Pt, Pd, Au, Ag, Ru, Re, and Ce have good catalytic effect on a complete removal of SO2 [16–19]. In addition, SO2 can be electro-chemically oxidized to NaHSO4 by H2O2 produced through dissolved O2 reduction on the surface of the graphite rod cathode [20]. In spite of their outstanding desulfurization performance, large-scale application for these technologies is still up in the air and mainly hindered by the three factors: (1) the mass transfer limitation of O2 due to its very limited solubility in the aqueous electrolyte [13]; (2) the insufficient reactive sites on a solid electrode, which means low removal efficiency for SO2; and (3) the extremely high price of the precious metal electrode materials. On this occasion, several researchers proposed gas diffusion electrodes (GDEs) as an effective method to overcome the problems mentioned above [21–24]. The GDE is a porous three-dimensional electrode, which is made up of a gas diffusion layer (GDL), a current collector (CC), and a catalyst layer (CL) [25–28]. The GDL toward the atmosphere is usually composed of conductive carbon black (CB) and a large number of polytetrafluoroethylene (PTFE) binder (60–70 wt% content), providing functions of delivering O2 as well as preventing electrolyte leakage. The CC between the GDL and the CL is used to improve the electrical conductivity and the mechanical strength of the GDE. The CL toward the electrolyte consists of low-cost carbon catalyst and a little PTFE (10–20 wt% content), forming numerous gas channels for O2 transfer as well as abundant gas–liquid–solid three phase interfaces (TPIs, i.e., reactive sites) for efficiently electro-generation of H2O2 via two-electron O2 reduction reaction (2e-ORR). As O2 coexists with SO2 in the flue gas (volume ratio 30–200:1) [29–32], it is more likely to realize complete and high-efficient electro-chemical removal of SO2 by H2O2 oxidation in situ with a gas diffusion cathode. In addition, the costs derived from the equipment and energy for aeration can be saved substantially [33]. In the previous work, we have demonstrated that the CBbased GDE is an alternative one for electro-chemical FGD with good operating stability and low-energy consumption. Moreover, adding a little amount of multi-walled carbon nanotubes (MWCNTs) in the CL is able to enhance the SO2 removal efficiency by promoting 2e-ORR, probably owing to its high electro-conductivity and mesopore structure [34–36]. The aim of this work is to further explore the role of MWCNTs in the electro-catalytic FGD process and the desulfurization mechanism. The effects of the amounts of MWCNTs on the desulfurization performance of GDEs are investigated according to the microscopic characterization and electro-chemical behavior. The influence of some operational parameters, including current density and electrolyte concentration on desulfurization, is also investigated.

Electrocatalysis (2017) 8:103–114

Experimental GDEs Preparation The sandwich-type GDE consists of a GDL and a CL with a stainless steel mesh (SSM) as the CC. The detailed preparation method has been described in our previous work [25, 26]. The GDL was made up of conductive CB (Jinqiushi Chemical Co. Ltd., Tianjin, China) and PTFE binder (60 wt%, Hesen, Shanghai, China) with a mass ratio of 3:7. After sintered at 340 °C for 25 min, the GDL was roll-pressed onto a SSM (6.5 cm × 6.5 cm, 60 meshes, Detiannuo Commercial Trade Co. Ltd., Tianjin, China) to form a flat sheet (0.5 mm in thickness). The carbon catalysis of 3 g including CB (254 m2 g−1 specific surface area, Vulcan XC-72, Cabot Corporation, USA) and MWCNTs (10–20 nm diameters, XFNANO Materials Tech Co. Ltd., Nanjing, China) with different mass ratios (pure CB; 2.95 g CB and 0.05 g MWCNTs; 2.9 g CB and 0.1 g MWCNTs; 2.7 g CB and 0.3 g MWCNTs; 2.5 g CB and 0.5 g MWCNTs) was added into the CL with PTFE binder at a mass ratio of 6:1. Then, each CL was roll-pressed onto the opposite side of the same flat sheet to form five types of GDEs with a thickness of 0.6 mm named as C, CCNT50, CCNT100, CCNT300, and CCNT500 in turn. Material Characterization The surface morphology images of the CLs were observed by scanning electron microscopy (SEM) (S-3500N, Hitachi Limited, Japan). The pore size distribution and total pore area of the CLs were measured by using a mercury porosimeter (Autopore IV, Micromeritics). The surface elemental composition of the CLs after electrolysis was analyzed using X-ray photoelectron spectroscopy (XPS) (K-alpha, Thermo Fisher Scientific, USA) operating at 10−9 Pa with an AlKα radiation (1486.6 eV), and all the samples were dried for 2 h in a vacuum drying oven at 80 °C after rinsed with high purity water. Electro-chemical Measurements The electro-chemical tests involving electro-chemical impedance spectroscopy (EIS), linear sweep voltammetry (LSV), and Tafel plots were conducted using a potentiostat (Autolab PGSTAT 302N, Metrohm, Herisau, Switzerland) in a conventional self-assembled three electrode cell with a volume of 14 mL (Fig. S1) at ~298 K. The five different GDEs were used as the working electrodes. A platinum sheet (1 cm2, Aidahengsheng Tech Co. Ltd., Tianjin, China) and an Ag/ AgCl (saturated KCl, 0.197 V versus SHE) were used as the counter electrode and reference electrode, respectively, which were both fixed with 0.5 cm spacing from the working electrode. The reactor was filled with 50 mM Na2SO4 solution for all the tests, and before each test, all the GDEs should be

Electrocatalysis (2017) 8:103–114

immersed in 50 mM Na2SO4 for 24 h. EIS measurements were performed over a frequency range from 100 kHz to 0.1 Hz at open circuit potential with a sinusoidal perturbation signal amplitude of 5 mV. LSV measurements were conducted at a scanning rate of 2 mV s−1 over a potential range from 0 to −1 V. N2 (99.99%, Baisida, Tianjin, China) was bubbled into the electrolyte for 30 min before each measurement to remove the dissolved O2. Then, O2 (99.99%, Baisida, Tianjin, China) was continuously bubbled into the gas chamber at a flow rate of 90 mL min−1 during the LSV tests. O2 was replaced by N2 with the same rate when the control tests were done. The Tafel plots of a base-10 logarithm of |current density| (lg |j|, A cm−2) versus |overpotential| (|η|, V) were recorded with the overpotential from 0 to 100 mV at 1 mV s−1. The electron spin resonance (ESR) detection technique (a Magnet TechMS400 spectrometer) was employed to estimate the free radicals generated by the ORR on the GDEs. The electrolysis reactor and flow rate of O2 were the same with that in the LSV measurements and a constant current of 0.1 A was applied. 5,5-dimethyl-1-pyrroline N-oxide (DMPO) was mixed with 50 mM Na2SO4 forming 10 mM DMPO electrolyte to trap the free radicals. The sample of 1.0 mL was transferred to a capillary tube for analysis with the ESR instrument after electrolysis for 30 min.

105

Results and Discussion Physical Characterization of the GDEs Surface Morphology of the CLs As shown in Fig. 1, the surface morphology of the five CLs is observed according to the SEM images at magnification of 20 K. It can be clearly seen from the five CLs that the numerous particles with small diameters (1–2 μm) distribute uniformly and form interconnected porous structure, representing the CB-PTFE network by roll-pressing method. The PTFE is unable to be observed clearly because of its slight addition (14.3 wt%) in all the CLs compared with that in the GDL (70 wt%) [26, 38]. For the CLs of the CCNT50, CCNT100, CCNT300, and CCNT500, the appearance of some scattered tube-like structures can be attributed to the addition of MWCNTs, which connect the CB-PTFE granules as Bbridges^ and the number of them increases gradually as the increased addition amount. However, the excessive MWCNTs (in the CCNT300 and CCNT500) are huddled in the form of scattered aggregates and may block some large pores formed by CB-PTFE granules. Porous Structure of the CLs

Electro-Catalytic Desulfurization The electrolysis reactor was the same as our previous experiments with 50 mM Na2SO4 electrolyte [34]. The gas chamber was filled with pure O2 for a few minutes, and then it was switched to the simulative flue gas containing SO2 (99.99%, Liufang, Tianjin, China) and O2 (1:9, 100 mL min−1) at ~298 K. A constant current of 0.1 A (current density 3.54 mA cm−2) was applied for each batch with an electrolytic period of 180 min, using a DC source meter (M8812, 75V/2A, Maynuo Electric Co. Ltd., Nanjing, China). To determine the concentration of H2O2 generated by 2e-ORR under electrocatalysis of the five GDEs, only O2 was fed into the gas chamber (90 mL min−1) and the other electrolytic conditions were the same with that in the electro-catalytic desulfurization tests. A series of experiments with different current densities (1.77, 3.54, 7.08, and 10.62 mA cm−2) and electrolyte concentrations (25, 50, 100, and 200 mM Na2SO4) were performed to investigate the influence of the two factors for electro-catalytic desulfurization and ORR. The SO42− concentration was determined by an ion chromatograph (Dionex ICS-5000, Thermo Fisher Scientific, USA), and the H2O2 concentration was measured by a spectrophotometer (T6, Perkinje General Instrument Co. Ltd., Beijing, China) according to a potassium titanium(IV) oxalate method [37]. All the experiments were performed in triplicate for error analysis and the average values were reported.

The porous structure of the CLs was analyzed using a mercury porosimeter (Fig. 2 and Table 1). As shown in the magnified view inserted in Fig. 2a, all the curves possess an intense peak at about 40 nm, indicating that the predominant pore size of the five CLs all distributes at 40 nm, mainly formed by CB and PTFE during the roll-pressing process. As the MWCNTs content was increased from 0 to 100 mg, the corresponding peak values of the CCNT50 and CCNT100 at 40 nm are almost the same with that of the C, which is possibly attributed to the low MWCNTs amount in the CLs. But as the MWCNTs content was further increased from 300 to 500 mg for the CCNT300 and CCNT500, the peak values at 40 nm began to decrease. In addition, a series of relatively moderate peaks at 12 nm emerged on the curves of the CCNT50, CCNT100, CCNT300, and CCNT500 in Fig. 2a, and higher peak values of differential intrusion were obtained with the increase of MWCNTs content (from 50 to 500 mg). It manifests that the peaks at 12 nm are due to the addition of MWCNTs with diameters of about 10–20 nm. Combined with the variation trend of peak at 40 and 12 nm for the five CLs, we inferred that the addition of MWCNTs in the CLs could produce more small pores and clog a part of the large pores (major in 40 nm) on account of the aggregation effect, which is in good agreement with the results of SEM. Figure 2b shows the total pore area versus the pore size of the five CLs, and the amplified view for the pore size in the range of 0 to 50 nm is inserted. In the pore size scope

106 Fig. 1 SEM images of the CLs of the C, CCNT50, CCNT100, CCNT300, and CCNT500 at magnification of 20 K

Electrocatalysis (2017) 8:103–114

C

CCNT50

1 µm

CCNT100

1 µm

CCNT300

1 µm

1 µm

CCNT500

1 µm

from 35 to 50 nm, the total pore areas of the CLs for the C, CCNT50, and CCNT100 are very close to each other, which is possibly attributed to the low MWCNTs content in CCNT50 and CCNT100. While for the CCNT300 and CCNT500, the total pore area is diminished gradually because of the increased addition of MWCNTs which block some large pores (major in 40 nm) as we mentioned above. For the pore size smaller than 35 nm, the increase of MWCNTs ratio brought a gradual growth in the total pore area, owing to its mesopore structure and scattered distribution among the CB particles (Fig. 1). The total pore area of the CCNT50, CCNT100, CCNT300, and CCNT500 is 1.9, 6.4, 10.4, and 16.6% larger than that of the C (84.5 m2 g−1), respectively, and the largest total pore area is obtained from the CCNT500 (98.5 m2 g−1), corresponding to the results of pore size distribution in Fig. 2a. Analogously, the porosity of the five CLs are arranged in descending order as follows: 72.4% (CCNT500) > 69.0% (CCNT300) > 68% (CCNT100) > 66.8% (CCNT50) > 63.2% (C). It is known that the total pore area and porosity are important factors to the ORR efficiency, where larger total pore area and porosity in the CL may provide additional ORR sites and O2 transfer channels separately

leading to better ORR activity [28]. In this regard, the increase of MWCNTs content in the CLs is likely in favor of the 2e-ORR activity for the five GDEs to produce H2O2. However, in consideration of that an exorbitant ORR activity may result in further decomposition of H2O2 to OH· or H2O, other characterizations are still needed to clarify the role of MWCNTs in the electro-catalytic FGD process. Electro-chemical Characterization of the GDEs EIS Measurement The Nyquist plots of the five GDEs obtained from EIS measurement are shown in Fig. 3a and were modeled by an equivalent circuit (inset figure in Fig. 3a) using a software named ZsimpWin. The series of Nyquist plots all consist of a semicircle in the high-frequency region and a sloping line at the low frequency range. The fitting results are given in Table 2, and the total internal resistances of the five GDEs are consisted of a solution resistance (Rs), a charge transfer resistance (Rct), and a diffusion resistance (Rd) [39]. The Rs has almost no difference

Electrocatalysis (2017) 8:103–114

107

the angular frequency within the low frequency range, but W which means Warburg coefficient is a constant [23, 25, 43]. The W of the CCNT50, CCNT100, CCNT300, and CCNT500 is 70.8–84.4% lower than that of the C (112.78 Ω s−0.5), further illustrating the positive role of MWCNTs for promoting ORR. In addition, the W tends to keep stable for the CCNT100, CCNT300, and CCNT500, suggesting that the doping amount of MWCNTs is no longer the pivotal influencing factors of W when it reaches to a certain point. According to the fitting results above, the conductivity of the GDEs can be calculated by Eq. (1) [44]. σ¼

N Rct  A

ð1Þ

where σ is the conductivity of the GDE, N is the thickness of GDEs (0.6 mm), and A is the surface area of the GDEs (28.3 cm2). The σ values are arranged in descending order as follows: 3.85 × 10−3 S cm−1 (CCNT500) > 2.94 × 10−3 S cm−1 (CCNT300) > 2.44 × 10 − 3 S cm − 1 (CCNT100) > 2.19 × 10−3 S cm−1 (CCNT50) > 3.93 × 10−4 S cm−1 (C). As anticipated, the MWCNTs added in the CLs acted as Belectron bridges^ [45] among the CB particles, leading to a about 10time increase in the conductivity of the GDE compared to pure CB GDEs, which can promote the ORR by accelerating the electron transfer. It is in accordance with the results of the total pore area and porosity for the five GDEs analyzed above. Fig. 2 Pore size distributions (a) and cumulative pore area (b) of the CLs made from the C, CCNT50, CCNT100, CCNT300, and CCNT500. Inset figure in a: detail of the pore size distribution ranged from 0 to 90 nm. Inset figure in b: detail of the cumulative pore area ranged from 0 to 50 nm

LSV Measurement

among the C (15.31 Ω), CCNT50 (15.56 Ω), CCNT100 (15.40 Ω), CCNT300 (15.22 Ω ), and CCNT500 (15.60 Ω) due to the same reactor configuration, electrolyte, and reference electrode used in the measurements [23, 40]. Rct, the dominant one in the internal resistance [25, 41, 42], is decreased accompanied by the increased addition of MWCNTs in the CLs, and the lowest Rct is obtained from the CCNT500 (0.55 Ω). The Rct of the CCNT50, CCNT100, CCNT300, and CCNT500 have a significant decrease from 82.0 to 89.8% compared to that of the C, indicating that the doping of MWCNTs in the CB-based GDEs accelerates the charge transfer processes during ORR. The Rd (W ω−0.5) is dynamic depending on

As verified in Fig. 3b, LSV tests were performed over the potential window ranged from 0 V to −1 V for the five GDEs. No obvious current was found under the deoxygenated condition for the C and CCNT100, illuminating that the cathode current was mostly from the ORR under the oxygenated condition [28]. Besides, an intense reduction peak appeared at ca. −0.55 V for all the GDEs should be attributed to the formation of H 2O2 via 2eORR [38]. The absolute values of the peak currents are arranged in descending order as follows: 39.54 mA (CCNT500) > 33.31 mA (CCNT300) > 24.69 mA (CCNT100) > 19.78 mA (CCNT50) > 9.98 mA (C), indicating that the positive role of MWCNTs for electrocatalyzing ORR probably is due to the enhancement in

Table 1 Structural characteristics of porous CLs

Catalytic layers

C

CCNT50

CCNT100

CCNT300

CCNT500

Total pore area (m2 g−1) Total intrusion volume (mL g−1) Porosity (%)

84.5 1.386 63.2

86.1 1.392 66.8

89.9 1.392 68.0

93.3 1.512 69.0

98.5 1.879 72.4

108

Electrocatalysis (2017) 8:103–114 Table 2

Fitting results of EIS spectra for the five GDEs

Gas diffusion electrodes

Rs (Ω)

C (F)

Rct (Ω)

W (Ω s−0.5)

C CNT50 CNT100 CNT300 CNT500

15.31 15.56 15.40 15.22 15.60

1.77 × 10−6 3.40 × 10−6 6.31 × 10−6 5.92 × 10−6 6.51 × 10−6

5.39 0.97 0.87 0.72 0.55

112.78 32.95 17.60 19.95 18.65

relatively weak reduction peak at ca. −0.7 V on the curves of the CCNT300 and CCNT500 (peak 1 and peak 2), which represents the generation of H2O resulted from the decomposition of H2O2. Tafel Plots The Tafel plots (Fig. 3c) were performed to better explain the different catalytic behaviors of the five GDEs during ORR according to the simplified Butler–Volmer equation [26, 46]:




βnF







lg
j
¼ lg
j0
þ ð2Þ
η
2:303RT


Fig. 3 Nyquist plots of EIS spectra for the five GDEs at open circuit potential (a). The frequency range 100 kHz to 0.1 Hz; the sinusoidal perturbation signal amplitude 5 mV; the electrolyte 50 mM Na2SO4 solution. Inset figure: an equivalent circuit for Nyquist plot fitting. LSV curves for the five GDEs in the atmosphere of O2 or N2 over a potential range from 0 to −1 V (b) and Tafel plots of the five GDEs with the overpotential from 0 to 100 mV (c)

the total pore area, porosity, and conductivity for the GDEs which leads to more active sites, faster O2, and electron transfer during ORR. Particularly, we can see a

In this equation, j is the current density, j0 is the exchange current density, β which means the transfer coefficient is a constant depending on the pH of electrolyte (≤0.5, approximately 0.036 in this work [26]), n is the electron transfer number in the cathodic ORR, F is the Faraday constant (96,485 C mol−1), R is the ideal gas constant (8.314 J mol−1 K−1), T is the absolute temperature (298 K), and η is the cathodic overpotential. The values of j0 and n depend on the intercept and slope of the linear area on the Tafel curve (overpotential interval of 80–100 mV), respectively. As listed in Table 3, the value of j0 rose up rapidly with the MWCNTs content increased from 0 to 500 mg in the CLs. Notably, the CCNT500 gave the highest j0 with a value of 7.18 × 10−4 A cm−2, which was 1.3, 1.7, 1.9, and 3.0 times higher than that of the CCNT300, CCNT100, CCNT50, and C, respectively, illustrating that the CCNT500 have the best ORR activity [46–49]. Combined with the previous results of total pore area, it can be adequately demonstrated that the GDE with larger pore size in the CL has better activity in catalyzing ORR mainly due to the additional active site. The corresponding n values for the C, CCNT50, and CCNT100 are 2.2, 2.3, and 2.6, which are close to the classical two-electron ORR process. But for the CCNT300 and CCNT500, the n values of 2.8 and 3.3 suggest a hybrid ORR with two-electron and fourelectron process simultaneously, which is adversed to the FGD experiments based on H2O2 oxidation.

Electrocatalysis (2017) 8:103–114 Table 3 Linear fit equations, exchange current densities, and electron transfer number calculated from the linear region of the Tafel plots for the five GDEs

109

GDEs

fit linear equation

R2

j0 (10−4 A cm−2)

Number

C

y = −3.61650 + 1.34416x

0.99395

2.42

2.2

CNT50 CNT100

y = −3.42051 + 1.39351x y = −3.36197 + 1.60649x

0.99679 0.99174

3.80 4.35

2.3 2.6

CNT300

y = −3.24732 + 1.73377x

0.99022

5.66

2.8

CNT500

y = −3.14391 + 2.00909x

0.99537

7.18

3.3

Desulfurization Performance of the GDEs The H2O2 yields of the five GDEs are plotted in Fig. 4a. It was accumulated continuously as the electrolysis time for all the GDEs. The value at each sample time was increased with MWCNTs doped from 0 to 100 mg, which was in good agreement with the results of the porous structure and electrical conduction. However, it fell down gradually with the increased MWCNTs doped from 300 to 500 mg. According to the results of LSV and Tafel tests, it is mainly caused by the decomposition of H2O2. The maximum H2O2 concentration (1002.44 mg L−1) was obtained from the CCNT100 after 180 min, which is one of the top production compared to many literatures using carbon-based GDEs under the similar electrolytic conditions [21, 22, 50, 51], followed by the CCNT50 (920.42 mg L−1), C (821.10 mg L−1), CCNT300 (755.19 mg L − 1 ), and CCNT500 (438.46 mg L − 1 ) successively. The desulfurization results of the five GDEs were evaluated on the basis of the concentration of SO42− in the electrolyte. After an electrolytic period of 180 min, the largest SO42− concentration (86.46 mM) was obtained from the CCNT100, giving a highest removal efficiency of 98.0% (Fig. 4b, c). For the other samples, the SO42− concentration and desulfurization efficiency followed the order of CCNT50 (79.65 mM, 91.1%) > C (75.75 mM, 89.1%) > CCNT300 (73.43 mM, 84.6%) > CCNT500 (65.58 mM, 75.6%). This consequence is consistent with the H2O2 yields above and further supported the inference in our previous study that the desulfurization results of the GDEs are dominated by H2O2 production [34]. Considering the industrial application in the future, the energy consumption during the electro-catalytic desulfurization test was another important indicator for evaluating desulfurization performance of the GDEs. It can be calculated according to Eq. (3): Z 3 U ⋅I⋅dt 0 E¼ ð3Þ M SO2 ⋅C⋅V where U, I, MSO2, C, and V represent the real-time voltage, supplying current (0.1 A), relative molecular mass of SO2 (64 g mol−1), SO42− concentration in the electrolyte, and volume of the electrolytic cell (123.5 mL), respectively. As

Fig. 4 The plots of H2O2 yield in the electrolyte generated by ORR (a) and SO42− production in desulfurization experiments (b) versus time using the C, CCNT50, CCNT100, CCNT300, and CCNT500 and SO2 removal efficiency and energy consumption in desulfurization experiments using the five GDEs (c)

110

illustrated in Fig. 4c, the CCNT100 still shows the highest desulfurization efficiency with energy consumption of 1.70 kW h for converting 1 kg SO2, saving 13.3, 8.6, 12.4, and 20.2% electric energy compared with the C, CCNT50, CCNT300, and CCNT500, respectively. Moreover, this energy consumption for the CCNT100 is 33.6% lower than the data reported by Wang et al. [20] with similar desulfurization conditions via electro-chemistry methods.

The Mechanism of Electro-Catalytic Desulfurization for the GDEs As described above, the desulfurization performance of the GDE is greatly influenced by H2O2 generated by 2e-ORR, which depends on the electron transfer number (n), conductivity, total pore area, and porosity orderly in reference to the literature [24, 33, 52, 53]. Hence, the better 2e-ORR and desulfurization performance of the CCNT50 and CCNT100 can be attributed to the addition of MWCNTs which enriches the ORR sites in the CL and promotes O2 and electron transfer by optimizing the porous structure and electrical conductivity of the GDEs. Although the CCNT300 and CCNT500 have higher conductivity, larger total pore area, and porosity, the H2O2 yield of them was 24.7 and 56.3% less than that of the CCNT100. It can be speculated that H 2O 2 was further decomposed to OH· or H 2O under the catalysis of the CCNT300 and CCNT500 during ORR. To deeply understand the mechanism of electro-catalytic FGD processes for the GDEs, ESR spectra of the C and CCNT500 was conducted using DMPO as OH· trapper under electrolysis. As depicted in Fig. S2, no DMPO–OH· signals (with intensity 1:2:2:1 [54]) were observed during the ORR after 30 min, indicating that the ORR process using this CBbased GDE doped with MWCNTs does not produce any OH·. Combined with the results of physical characterization, electro-chemical and ESR tests, it can be proved that excess Fig. 5 The schematic diagram of desulfurization mechanism for the GDEs

Electrocatalysis (2017) 8:103–114

MWCNTs added in the CLs greatly provided additional ORR sites and accelerated the electron transfer process, resulting in higher activity of ORR. It means that a part of H2O2 was further decomposed to H2O, going against the desulfurization performance of the five GDEs. The possible desulfurization mechanism of the GDEs is proposed in Fig. 5. First of all, the simulated flue gas (mixture of SO2 and O2) was put into the gas chamber of the reactor and then contacted with the GDL which can spontaneously absorb the gas mixture and transport it to the CL by the capillary action. At the second step, on the gas (SO2 and O2)–liquid (electrolyte)–solid (carbon material) three-phase boundary, SO2 was easy to dissolve in electrolyte to form SO32− and O2 is electro-catalytically reduced to H2O2. Finally, H2O2 oxidize SO32− to SO42−, achieving the purpose of desulfurization. As for the CCNT300 and CCNT500, H2O2 was further decomposed to H2O by the catalysis of excess MWCNTs, which led to a poorer desulfurization performance and H2O2 yield. According to these, the electro-catalytic desulfurization process should be described as follows: SO2 þ H2 O→ 2Hþ þ SO3 2−

ð4Þ

O2 þ 2Hþ þ 2e→ H2 O2

ð5Þ

þ

ð6Þ

H2 O2 þ 2H þ 2e→ 2H2 O SO3

2−

þ H2 O2 →SO4

2−

þ H2 O

ð7Þ

The Influence of Operational Parameters Current density and electrolyte concentration are the two important operational parameters for the 2e-ORR which has a close relationship with SO2 removal. So, it is necessary to explore the influence of current density and electrolyte concentration for the 2e-ORR and desulfurization performance. In view of the optimal doping amount of MWCNTs in the CLs is 100 mg, the CCNT100 was chosen as the cathode.

Electrocatalysis (2017) 8:103–114

111

The Influence of Current Density In order to investigate the influence of current density on the electro-catalytic desulfurization, a series of current density (1.77, 3.54, 7.07, 10.61 mA cm−2) was applied in the electrolysis experiments with 50 mM Na2SO4. As indicated in Fig. 6a, the H2O2 concentration in the electrolyte after 180 min gradually increased with the increasing current density and the maximum yield of 2047.06 mg L−1 was reached when the current density was 10.61 mA cm−2, which is the top one compared to literature under the similar current using carbon-based GDE [21, 22, 50, 51]. On the other hand, the trend of SO42− concentration found in Fig. 6b (55.98, 86.46, 88.61, and 90.39 mM, respectively, with the increasing current density) was non-accidentally in good agreement with that of H2O2 concentration. Figure 6c shows the current efficiency (CE) versus time within 180 min electrolysis, which can be determined using the following equation: CEðH2 O2 Þ ¼

n⋅F⋅C H2 O2 ⋅V  100% M H2 O2 ⋅I⋅t

ð8Þ

where n is the electron transfer number in ORR (n = 2), F is the Faraday constant (96,485 C mol−1), CH2O2 is H2O2 concentration in the electrolyte, and MH2O2 is the relative molecular mass of H2O2 (34 g mol−1). Accordingly, CEs were

45.98, 65.08, 53.06, and 44.30%, respectively, with increasing current densities, whose trend was nearly contrary in comparison with that of SO42− and H2O2 concentrations. This observation was in accordance with the preceding literatures [22, 55, 56]. Accordingly, the energy consumption under the current density of 7.07, and 10.61 mA cm−2 was 3.0–5.2 times higher than that under the current density of 3.54 mA cm−2 (Fig. 6d). Although the increase in the current density is able to accelerate the electron transfer on the GDE to generate more H2O2, some side reactions such as 4e-ORR and hydrogen evolution reaction may occur and become more distinct when a higher current density (≥7.07 mA cm−2 ) is applied [22]. This can explain why the CE under the current density of 3.54 mA cm−2 was the highest and give us a notice that an appropriate current density is necessary for 2e-ORR and desulfurization. The Influence of Electrolyte Concentration The H2O2 and SO42− concentrations obtained with different Na2SO4 concentrations (25, 50, 100, and 200 mM) are given in Fig. 7a, b. As the Na2SO4 concentration increased from 25 to 50 mM, the H2O2 and SO42− concentrations slightly increased probably due to the increasing conductivity of the electrolyte and then began to decrease when the Na2SO4

Fig. 6 The H2O2 yield (a), SO42− production (b), and current efficiency (c) versus time and SO2 removal efficiency and energy consumption obtained from the CCNT100 with different current density (d). The electrolyte 50 mM Na2SO4 solution; the O2 flow rate 90 mL min−1

112

Electrocatalysis (2017) 8:103–114

Fig. 7 The H2O2 yield (a) and SO42− production (b) versus time obtained from the CCNT100 with different electrolyte concentrations; the XPS spectra of S2p detected in the surface of a fresh CCNT100; an immersed CCNT100 and four CCNT100 used with different electrolyte

concentration after ORR tests (c); and SO2 removal efficiency and energy consumption using the CCNT100 with different electrolyte concentrations (d). The current density 3.54 mA cm−2; the O2 flow rate 90 mL min−1

concentration increased further to 100 and 200 mM. In order to further explain this phenomenon, XPS was performed for the four CCNT100 using in 25, 50, 100, and 200 mM Na2SO4 electrolyte, respectively, after 180-min ORR tests. By contrast, a fresh CCNT100 and a CCNT100 immersed in 200 mM Na2SO4 solution for 180 min without the ORR were also conducted by XPS and the results are summarized in Fig. 7c and Table S1. The S2p spectra in Fig. 7c are fitted with two peak contributions, and one peak at 163.8 eV can be found on the curve of all the samples with almost identical intensities, indicating that these are the organic sulfur impurities existing in the CB powders inherently. Another peak at 168.5 eV can be assigned as SO32− or SO42− [57, 58], of which the intensity was gradually incremental for the CCNT100 electrolyzed with 25 mM (S atomic 0.26%), 50 mM (S atomic 0.41%), 100 mM (S atomic 0.65%), and 200 mM (S atomic 0.81%) Na2SO4 solution in turn after 180 min. In addition, no S signals can be identified for the fresh CCNT100 and immersed CCNT100 at 168.5 eV, indicating that more reactive sites on the GDE may be occupied by the SO42− with the increase of electrolyte concentration during ORR and thus hindering the formation of H2O2 [59], which is harmful to the electro-catalytic desulfurization. As a result, even though the energy consumption of desulfurization with Na2SO4

concentration of 100 and 200 mM was 4.1 and 14.7% less than that with 50 mM Na2SO4, respectively, due to the rising conductivity of the electrolyte, the SO2 removal efficiency with 50 mM Na2SO4 was reversely 14.1 and 19.1% higher than that with 100 and 200 mM Na2SO4 orderly (Fig. 7d).

Conclusions With the aim to explore the function of MWCNTs in the CBbased GDE for in situ electro-catalytic FGD, CB-based GDE doped with different amounts of MWCNTs was prepared. Although the total pore area, porosity, and conductivity were increased with the increase in the MWCNTs content from 0 to 500 mg, the highest SO2 removal efficiency (98.0%) with the lowest energy consumption (1.7 kW h kg−1) was obtained from CB-based GDE doped with 100 mg MWCNTs. It is because that, in this in situ FGD system, the SO2 removal is mainly realized by the oxidation of H2O2 which requires a good balance among the porous structure, conductivity, and catalysis to ensure a 2e-ORR route. Otherwise, excess MWCNTs will lead to a decomposition of H2O2 to H2O. In addition, an appropriate current density (3.54 mA cm−2) and electrolyte concentration (50 mM) is necessary for 2e-ORR

Electrocatalysis (2017) 8:103–114

and desulfurization, while some side reactions may occur and more reactive sites on the GDE may be occupied by the SO42− with the excessive current density and electrolyte concentration, respectively, during ORR, which are unfavorable to the formation of H2O2. Considering the superiority in desulfurization efficiency, environmental friendliness, and energy conservation, this electro-catalytic system based on CB-based GDE doped with MWCNTs provides great potential for in situ FGD in prospective industrial application. Acknowledgments The authors gratefully acknowledge financial support by the Major National Science and Technology Projects of China on Water Pollution Control and Treatment (2012ZX07501002-001), Research Project of Tianjin City for Application Foundation and Advanced Technology (BE026071), and the Fundamental Research Funds for the Central Universities.

References 1.

2.

3.

4.

5.

6.

7.

8.

9.

10. 11.

12.

13.

F. L. K. Lam, M. O. Fu, X. J. Hu, Superior adsorption capacity of film typed carbon for the abatement of sulfur dioxide. Catal. Today 158, 269 (2010) F. Rahmaninejad, V. S. Gavaskar, J. Abbasian, Dry regenerable CuO/γ-Al2O3 catalyst for simultaneous removal of SOx and NOx from flue gas. Appl. Catal. B-Environ. 119, 297 (2012) H. He, Y. S. Wang, Q. X. Ma, J. Z. Ma, B. W. Chu, D. S. Ji, G. Q. Tang, C. Liu, H. X. Zhang, J. M. Hao, Mineral dust and NOx promote the conversion of SO2 to sulfate in heavy pollution days. Sci. Rep. 4, 1 (2014) N. Rastogi, A. Singh, D. Singh, M. M. Sarin, Chemical characteristics of PM2.5 at a source region of biomass burning emissions: evidence for secondary aerosol formation. Environ. Pollut. 184, 563 (2014) S. C. Ma, J. J. Yao, X. Y. Ma, L. Gao, M. Guo, Removal of SO2 and NOx using microwave swing adsorption over activated carbon carried catalyst. Chem. Eng. Technol. 36, 1217 (2013) R. T. Yang, A. J. Hernandez-Maldonado, F. H. Yang, Desulfurization of transportation fuels with zeolites under ambient conditions. Science 301, 79 (2003) Y. Mathieu, L. Tzanis, M. Soulard, J. Patarin, M. Vierling, M. Molière, Adsorption of SOx by oxide materials: a review. Fuel Process. Technol. 114, 81 (2013) F. J. Gutierrez Ortiz, F. Vidal, P. Ollero, L. Salvador, V. Cortes, A. Gimenez, Pilot-plant technical assessment of wet flue gas desulfurization using limestone. Ind. Eng. Chem. Res. 45, 1466 (2006) C. L. Chen, H. S. Wen, Nanosized CeO2-supported metal oxide catalysts for catalytic reduction of SO2 with CO as a reducing agent. Appl. Catal. B-Environ. 55, 115 (2005) Y. Hu, C. L. Yang, L. M. Cao, J. Yang, An electrochemical membrane reactor for a recycled FGD process. Chem. Eng. J. 223, 563 (2013) C. H. Huang, T. W. Xu, Comparative study on the regeneration of flue-gas desulfurizing agents by using conventional electrodialysis (ED) and bipolar membrane electrodialysis (BMED). Environ. Sci. Technol. 40, 5527 (2006) C. L. Yang, Y. Hu, L. M. Cao, J. Yang, Circulating regeneration and resource recovery of flue gas desulfurization residuals using a membrane electroreactor: from lab concept to commercial scale. Environ. Sci. Technol. 46, 11273 (2012) U. T. Un, A. S. Koparal, U. B. Ogutveren, Sulfur dioxide removal from flue gases by electrochemical absorption. Sep. Purif. Technol. 133, 13 (2007)

113 14.

M. Aurousseau, T. Hunger, A. Storck, F. Lapicque, Electrochemical scrubbing of SO2-containing gas: coupling absorption to electrochemical reaction. Chem. Eng. Sci. 48, 541 (1993) 15. A. J. Appleby, B. Pichon, The mechanism of the electrochemical oxidation of sulfur dioxide in sulfuric acid solutions. J. Electroanal. Chem. 95, 59 (1979) 16. P. Hoffmann, C. Roizard, F. Lapicque, S. Vénot, A. Maire, A process for the simultaneous removal of SO2 and NOx using Ce (IV) redox catalysis. Process Saf Environment 75, 43 (1997) 17. K. C. Pillai, S. J. Chung, T. Raju, I. S. Moon, Experimental aspects of combined NOx and SO2 removal from flue-gas mixture in an integrated wet scrubber-electrochemical cell system. Chemosphere 76, 657 (2009) 18. P. W. T. Lu, R. L. Ammon, An investigation of electrode materials for the anodic oxidation of sulfur dioxide in concentrated sulfuric acid. J. Electrochem. Soc. 127, 2610 (1980) 19. J. A. O'Brien, J. T. Hinkley, S. W. Donne, The electrochemical oxidation of aqueous sulfur dioxide I. Experimental parameter influences on electrode behavior. J. Electrochem. Soc. 157, 111 (2010) 20. C. Wang, H. Liu, X. Z. Li, J. Y. Shi, G. F. Ouyang, M. Peng, C. C. Jiang, H. N. Cui, A new concept of desulfurization: the electrochemically driven and green conversion of SO2 to NaHSO4 in aqueous solution. Environ. Sci. Technol. 42, 8585 (2008) 21. H. J. Luo, C. L. Li, C. Q. Wu, W. Zheng, X. Q. Dong, Electrochemical degradation of phenol by in situ electrogenerated and electro-activated hydrogen peroxide using an improved gas diffusion cathode. Electrochim. Acta 186, 486 (2015) 22. X. M. Yu, M. H. Zhou, G. B. Ren, L. Ma, Catalysis kinetics and porous analysis of rolling activated carbon-PTFE air-cathode in microbial fuel cells. Chem. Eng. J. 263, 92 (2015) 23. Q. N. Wang, H. Dong, H. B. Yu, Development of rolling tin gas diffusion electrode for carbon dioxide electrochemical reduction to produce formate in aqueous electrolyte. J. Power Sources 271, 278 (2014) 24. J. J. Wu, P. P. Sharma, B. H. Harris, X. D. Zhou, Electrochemical reduction of carbon dioxide: IV dependence of the Faradaic efficiency and current density on the microstructure and thickness of tin electrode. J. Power Sources 258, 189 (2014) 25. H. Dong, H. B. Yu, X. Wang, Q. X. Zhou, J. L. Feng, A novel structure of scalable air-cathode without Nafion and Pt by rolling activated carbon and PTFE as catalyst layer in microbial fuel cells. Water Res. 46, 5777 (2012) 26. H. Dong, H. B. Yu, X. Wang, Catalysis kinetics and porous analysis of rolling activated carbon-PTFE air-cathode in microbial fuel cells. Environ. Sci. Technol. 46, 13009 (2012) 27. H. Dong, H. Yu, H. B. Yu, N. S. G. Gao, X. Wang, Enhanced performance of activated carbon-polytetrafluoroethylene air-cathode by avoidance of sintering on catalyst layer in microbial fuel cells. J. Power Sources 232, 132 (2013) 28. N. Li, J. K. An, L. Zhou, T. Li, J. H. Li, C. J. Feng, X. Wang, A novel carbon black graphite hybrid air-cathode for efficient hydrogen peroxide production in bioelectrochemical systems. J. Power Sources 306, 495 (2016) 29. F. Vidal Barrero, P. Ollero, A. L. Villanueva Perales, A. GomezBarea, Catalytic seawater flue gas desulfurization model. Environ. Sci. Technol. 43, 9393–9399 (2009) 30. G. S. Qi, R. T. Yang, Low-temperature selective catalytic reduction of NO with NH3 over iron and manganese oxides supported on titania. Appl. Catal. B-Environ. 44, 217 (2003) 31. D. H. Xia, L. L. Hu, C. He, W. Q. Pan, T. S. Yang, Y. C. Yang, D. Shu, Simultaneous photocatalytic elimination of gaseous NO and SO2 in a BiOI/Al2O3-padded trickling scrubber under visible light. Chem. Eng. J. 279, 929 (2015) 32. Y. Zhao, R. L. Hao, T. H. Wang, C. Y. Yang, Follow-up research for integrative process of pre-oxidation and post-absorption cleaning flue gas: absorption of NO2, NO and SO2. Chem. Eng. J. 273, 55 (2015)

114 33. 34.

35.

36.

37. 38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

Electrocatalysis (2017) 8:103–114 B. E. Logan, M. Elimelech, Membrane-based processes for sustainable power generation using water. Nature 488, 313 (2012) Z. Chen, H. Dong, H. B. Yu, H. Yu, In-situ electrochemical flue gas desulfurization via carbon black-based gas diffusion electrodes: performance, kinetics and mechanism. Chem. Eng. J. 307, 553 (2017) Z. H. Yang, N. Nakashima, An electrocatalyst based on carbon nanotubes coated with poly(vinylpyrrolidone) shows a high tolerance to carbon monoxide in a direct methanol fuel cell. ChemCatChem 8, 600 (2016) Z. H. Yang, A. Nagashima, T. Fujigaya, N. Nakashima, Electrocatalyst composed of platinum nanoparticles deposited on doubly polymer-coated carbon nanotubes shows a high COtolerance in methanol oxidation reaction. Int. J. Hydrog. Energy 41, 19182 (2016) R. M. Sellers, Spectrophotometric determination of hydrogen peroxide using potassium titanium(IV) oxalate. Analyst 105, 950 (1980) J. Y. Chen, N. Li, L. Zhao, Three-dimensional electrode microbial fuel cell for hydrogen peroxide synthesis coupled to wastewater treatment. J. Power Sources 254, 316 (2014) F. Zhang, D. Pant, B. E. Logan, Long-term performance of activated carbon air cathodes with different diffusion layer porosities in microbial fuel cells. Biosens. Bioelectron. 30, 49 (2011) F. Zhang, M. D. Merrill, J. C. Tokash, T. Saito, S. A. Cheng, M. A. Hickner, B. E. Logan, Mesh optimization for microbial fuel cell cathodes constructed around stainless steel mesh current collectors. J. Power Sources 196, 1097 (2011) X. Zhang, K. X. Li, P. Y. Yan, Z. Q. Liu, L. T. Pu, N-type Cu2O doped activated carbon as catalyst for improving power generation of air cathode microbial fuel cells. Bioresour. Technol. 187(299) (2015) X. J. Li, X. Wang, Y. Y. Zhang, N. Ding, Q. X. Zhou, Opening size optimization of metal matrix in rolling-pressed activated carbon air–cathode for microbial fuel cells. Appl. Energy 123, 13 (2014) R. Vedalakshmi, V. Saraswathy, H. W. Song, N. Palaniswamy, Determination of diffusion coefficient of chloride in concrete using Warburg diffusion coefficient. Corros. Sci. 51, 1299 (2009) Y. Cui, X.L. Zhao, R.S. Guo, Improved electrochemical performance of La0.7Sr0.3MnO3 and carbon co-coated LiFePO4 synthesized by freeze-drying process, Electrochim. Acta 55, 922 (2010) W.J. Liu, K. Tian, Y.R. He, H. Jiang, H.Q. Yu, High-yield harvest of nanofibers/mesoporous carbon composite by pyrolysis of waste biomass and its application for high durability electrochemical energy storage. Environ, Sci. Technol. 48, 13951 (2014) P. Y. Liu, S. C. Sun, Y. S. Chen, M. C. Chuang, Efficient electrocatalytic reduction and detection of hydrogen peroxideat an IrIVOx·H2O nanostructured electrode prepared by electroflocculation. Electrochim. Acta 187, 256 (2016) X. L. Zhou, Y. Liu, H. X. Ju, B. C. Pan, J. F. Zhu, T. Ding, C. D. Wang, Q. Yang, Design and epitaxial growth of MoSe2–NiSe vertical heteronanostructures with electronic modulation for enhanced hydrogen evolution reaction. Chem. Mater. 28(1838) (2016)

48.

Z. Q. Liu, B. C. Ge, K. X. Li, X. Zhang, K. Huang, The excellent performance and mechanism of activated carbon air cathode doped with different type of cobalt for microbial fuel cells. Fuel 176, 173 (2016) 49. J.W. Miao, F.X. Xiao, H.B Yang, S.Y. Khoo, J.Z. Chen, Z.X. Fan, Y.Y. Hsu, H.M. Chen, H. Zhang, B. Liu, Hierarchical Ni-Mo-S nanosheets on carbon fiber cloth: a flexible electrode for efficient hydrogen generation in neutral electrolyte, Sci. Adv. 1, (2015) 50. R. B. Valim, R. M. Reis, P. S. Castro, A. S. Lima, R. S. Rocha, M. Bertotti, M. R. V. Lanza, Electrogeneration of hydrogen peroxide in gas diffusion electrodes modified with tert-butyl-anthraquinone on carbon black support. Carbon 61, 236 (2013) 51. H. J. Luo, C. L. Li, C. Q. Wu, X. Q. Dong, In situ electrosynthesis of hydrogen peroxide with an improved gas diffusion cathode by rolling carbon black and PTFE. RSC Adv. 5, 65227 (2015) 52. N. Duteanu, B. Erable, S. S. Kumar, M. M. Ghangrekar, K. Scott, Effect of chemically modified Vulcan XC-72R on the performance of air-breathing cathode in a single-chamber microbial fuel cell. Bioresour. Technol. 101, 5250 (2010) 53. L. Nie, J. Liu, Y. Zhang, M. Liu, Effects of pore formers on microstructure and performance of cathode membranes for solid oxide fuel cells. J. Power Sources 196, 9975 (2011) 54. J. C. Zhao, T. X. Wu, K. Q. Wu, K. Oikawa, H. Hidaka, N. Serpone, Photoassisted degradation of dye pollutants. 3. Degradation of the cationic dye rhodamine B in aqueous anionic surfactant/TiO2 dispersions under visible light irradiation: evidence for the need of substrate adsorption on TiO2 particles. Environ. Sci. Technol. 32, 2394 (1998) 55. J. Y. Rau, J. C. Chen, M. D. Lin, M. Y. Wey, Removal the coal ash, NO, and SO2 simultaneously by the fluidized-bed catalyst reactor. Energ. Fuel 24, 1711 (2010) 56. H. Z. Zhao, Y. Zhang, B. Zhao, Y. Y. Chang, Z. S. Li, Electrochemical reduction of carbon dioxide in an MFC-MEC system with a layer-by-layer self-assembly carbon nanotube/cobalt phthalocyanine modified electrode. Environ. Sci. Technol. 46, 5198 (2012) 57. Z. H. Yang, T. Fujigaya, N. Nakashima, Homogeneous coating of ionomer on electrocatalyst assisted by polybenzimidazole as an adhesive layer and its effect on fuel cell performance. J. Power Sources 300, 175 (2015) 58. K. Punyawudho, S. Ma, J. W. Van Zee, J. R. Monnier, Effect of O2 on the adsorption of SO2 on carbon-supported Pt electrocatalysts. Langmuir 27, 7524 (2011) 59. J. M. Chen, Y. J. Xia, Q. Z. Dai, Electrochemical degradation of chloramphenicol with a novel Al doped PbO2 electrode: performance, kinetics and degradation mechanism. Electrochim. Acta 165, 277 (2015)