- Email: [email protected]
A novel carbon black graphite hybrid air-cathode for efficient hydrogen peroxide production in bioelectrochemical systems
Journal of Power Sources 306 (2016) 495e502
Contents lists available at ScienceDirect
Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour
A novel carbon black graphite hybrid air-cathode for efficient hydrogen peroxide production in bioelectrochemical systems Nan Li a, Jingkun An a, Lean Zhou b, Tian Li b, Junhui Li b, Cuijuan Feng b, Xin Wang b, * a
Tianjin Key Lab of Indoor Air Environmental Quality Control, School of Environmental Science and Engineering, Tianjin University, No. 92 Weijin Road, Nankai District, Tianjin 300072, China b MOE Key Laboratory of Pollution Processes and Environmental Criteria, Tianjin Key Laboratory of Environmental Remediation and Pollution Control, College of Environmental Science and Engineering, Nankai University, No. 38 Tongyan Road, Jinnan District, Tianjin 300350, China
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
A hybrid air-cathode with the optimal carbon black/graphite ratio of 1:5 is made. The maximum H2O2 yield is 11.9 mg L1 h1 cm2. Continuous flow without H2O2 accumulation increases current efficiency. Oxygen for H2O2 synthesis is mainly contributed by air diffusion (66 e94%). The use of bioanode increases H2O2 yield and current efficiency.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 22 October 2015 Received in revised form 8 December 2015 Accepted 16 December 2015 Available online xxx
Carbon black and graphite hybrid air-cathode is proved to be effective for H2O2 production in bioelectrochemical systems. The optimal mass ratio of carbon black to graphite is 1:5 with the highest H2O2 yield of 11.9 mg L1 h1 cm2 (12.3 mA cm2). Continuous flow is found to improve the current efficiency due to the avoidance of H2O2 accumulation. In the biological system, the highest H2O2 yield reaches 3.29 mg L1h1 (0.079 kg m3day1) with a current efficiency of 72%, which is higher than the abiotic system at the same current density. H2O2 produced in this system is mainly from the oxygen diffused through this air-cathode (>66%), especially when a more negative cathode potential is applied (94% at 1.0 V). This hybrid air-cathode has advantages of high H2O2 yield, high current density and no need of aeration, which make the synthesis of H2O2 more efficient and economical. © 2015 Elsevier B.V. All rights reserved.
Keywords: Bioelectrochemical systems Microbial fuel cells Hydrogen peroxide Air-cathode Graphite Carbon black
1. Introduction Hydrogen peroxide (H2O2) has been widely known as an environmentally friendly chemical which leaves no hazardous residues,
* Corresponding author. E-mail address: [email protected] (X. Wang). http://dx.doi.org/10.1016/j.jpowsour.2015.12.078 0378-7753/© 2015 Elsevier B.V. All rights reserved.
since it decomposes only to water and oxygen. As a powerful and versatile chemical, H2O2 is effective throughout the pH range from 0 to 14 with high oxidation potential (E0 ¼ 1.763 V at pH ¼ 0 and E0 ¼ 0.878 V at pH ¼ 14) [1]. Hence, H2O2 is applied to numerous industrial areas such as chemical synthesis, pulp paper and textile bleaching, medical disinfection, treatment of wastewater and destruction of hazardous organic wastes [2e5]. In general, H2O2 is commercially produced on an industrial scale by anthraquinone
496
N. Li et al. / Journal of Power Sources 306 (2016) 495e502
oxidation, which needs multi-step procedure and significant energy input. Considering the transport, handling and storage of H2O2 are potentially hazardous, anthraquinone oxidation method is thought to be inefficient and insecure [6,7]. In recent decades, a lot of researches have demonstrated that H2O2 can be in situ generated by reduction of oxygen in alkaline [Eq. (1)], neutral or acidic solution [Eq. (2)] in an electrochemical system [8]. O2þH2Oþ2e/HO 2 þOH
(1)
O2þ2Hþþ2e/H2O2
(2)
Bioelectrochemical systems (BESs), also known as microbial electrochemical systems or MXCs, demonstrated the possibility of H2O2 production using renewable energy from wastewaters in recent years [9]. BES reactor consists of an anode, a cathode and a separator. Electrochemically active bacteria on the anode directly oxidize soluble organic waste to generate electron for the cathode. On the cathode, electrons can be used for four-electron oxygen reduction reaction (ORR) to water or through a two-electron pathway to produce H2O2 [10e12]. Due to the advantages of high conductivity, low price, good stability and low catalytic activity of H2O2 decomposition, carbonbased cathodes such as carbon clothes [10], graphite granules [11], spectrographically pure graphite (SPG) rods [12] and modified graphite [13] were usually used in the H2O2 electrosynthesis process in BESs. Most of these cathodes, such as our three-dimensional electrodes [11], were immersed in the electrolyte to provide the reaction zone for proton/electron and O2 that diffuses along with the electrolyte to synthesize H2O2. This kind of cathode has more potential use in refractory pollutant removal and can be easily scaled-up. However, the low solubility of O2 in the electrolyte is a key limitation where external power is also needed for cathodic aeration in these systems. Air-cathode is an advanced design where oxygen is passively supplied without aeration. It is usually composed of a catalyst layer (CL) facing to the electrolyte and a gas diffusion layer (GDL) facing to the air [14]. It had been demonstrated that H2O2 can be efficiently generated using air-cathodes [10] with a production rate 2 kg m3day1. However, in previous reports, most air-cathodes that utilized for H2O2 production didn't have a GDL [11,15], which may have a problem of electrolyte leakage and catalyst flooding when the system is scaled up. Apart from oxygen diffusion, the oxygen dissolved in electrolyte should be a considerable amendment for cathodic H2O2 produce [15,16]. However, to the best of our knowledge, the contribution of dissolved oxygen to H2O2 in aircathode BES has not been investigated yet. In the present work, BESs equipped with novel carbon black and graphite hybrid aircathodes made by rolling-press method were utilized to produce H2O2. The performances of cathodes with different carbon black/ graphite mass ratio were analyzed. The contributions of dissolved and diffused oxygen on H2O2 production were also evaluated. 2. Experimental 2.1. Carbon black and graphite (CB&G) hybrid air-cathode The CB&G air-cathode consisted of a CL and a GDL with the stainless steel mesh as the current collector and matrix. CL was prepared according to the following procedures. Carbon black (CB, 30 nm, Vulcan XC-72R, Cabot Corporation, US) and graphite (40 mm, HTF0325, >99.9%, Huatai Chemical Reagent Co. Ltd., Qingdao, China) were cleaned by ultrasonic in deionized water for 20 min at room temperature. After drying, a mixture of 6 g CB and graphite powders at different mass ratios (4.8 g CB and 1.2 g graphite as
CB&G4:1; 4 g CB and 2 g graphite as CB&G2:1; 3 g CB and 3 g graphite as CB&G1:1; 1.5 g CB and 4.5 g graphite as CB&G1:3; 1 g CB and 5 g graphite as CB&G1:5 and 0.75 g CB and 5.25 g graphite as CB&G1:7) was individually dispersed into 45 mL ethanol in a beaker in an ultrasonic bath at room temperature for 10 min. Polytetrafluoroethylene (PTFE) emulsion (60%, Horizon, Shanghai, China) as binder was slowly added to the mixture to fabricate the catalyst layer as previously described [17]. The mixture was then stirred at about 80 C to get a dough-like paste. The paste was firstly rolled to a 0.5 mm CL film and then roll-pressed on one side of the stainless steel mesh (4 cm 4 cm, Type 304N, 60 meshes, Detiannuo Commercial Trade Co. Ltd., Tianjin, China) to be a flat sheet (0.5 mm in thickness). CLs made of pure CB and graphite were also fabricated and marked as PCB and PG. The GDLs of each cathode were made in parallel by rolling a mixture of carbon black and PTFE emulsion with a mass ratio of 4:9 according to the procedure described previously [18]. After heating at 340 C for 25 min, GDLs were rolled onto the opposite side of different CLs to form a final air-cathode with a total thickness of 1 mm [18]. 2.2. Material characterization The specific surface area and pore parameters of carbon black and graphite powders were determined by the multipoint BrunauereEmmetteTeller (BET) measurement at 77 K using Autosorb1 (Quantachrome, USA). Porous properties and the pore size distribution were analyzed based on the BJH model. The porous structure of the CLs were characterized based on capillary law with a mercury porosimeter (Autopore IV, Micromeritics) which can analyze porous media with a pore size ranging from 6 nm to 300 mm. Samples were firstly dried in an oven at 105 C for 3 h in a stream of nitrogen, and then intruded volume analysis was carried out over the pressure ranging from 5 kPa to 414 MPa. After each change in pressure, the system equilibrated for 15 s to minimize overlap of the intra- (0.002e1 mm pore radius) and inter(1e100 mm pore radius) porosity regions as determined by the porosimeter. Images of the surface morphology of CLs were taken with a scanning electron microscope (SEM, Nanosem 430, USA). 2.3. BES configuration and operation The experiments were performed on dual-chamber reactors shown in Fig. 1 equipped with different CB&G air-cathodes. The anode chamber was 3 cm in diameter and 4 cm in length (net volume of 28 mL) while the cathode chamber was 3 cm in diameter and 2 cm in length (net volume of 14 mL) [19]. Both chambers were assembled tightly on two sides of a cation exchange membrane (Ultrex CMI-7000, Membranes International Inc., Glen Rock, NJ, USA) by four clamping bolts. Anodes were made of carbon fiber brushes (3 cm in diameter and 2 cm in length). Different CB&G aircathodes were assembled to the other side of cathode chambers. Two 4 mm diameter holes were drilled on the top and at the bottom of the cathode chamber connecting to the effluent beaker and the influent aeration tank as showed in Fig. 1. N2 gas or air was continuously aerated into the aeration tank to obtain an oxygen free or oxygen saturated influent. Peristaltic pump (BT100L, 8 channels, Longer Precision Pump Co. Ltd, Baoding, China) was used to feed the cathode with electrolyte from aeration tank at the rate of 10 mL min1. All anode chambers of BESs were inoculated with the effluent from MFCs and pre-acclimated under 1 kU of external resistance with activated carbon air-cathode for 2 months before coupling with cathode chambers [19]. The medium contained acetate (1.0 g L1), a phosphate buffer solution (50 mM PBS; NH4Cl 0.31 g L1, KCl 0.13 g L1, NaH2PO4 2H2O 2.772 g L1,Na2HPO4
N. Li et al. / Journal of Power Sources 306 (2016) 495e502
497
Fig. 1. Schematic diagram of the system utilized for H2O2 yield measurement using different carbon black/graphite hybrid air-cathodes.
4.576 g L1), a trace mineral (12.5 mL L1) and a vitamin solution (5 mL L1) [17]. When the acclimation of anodic biofilm was completed, cathode chamber and the aeration tank were filled with a 50 mM Na2SO4 solution. For the abiotic tests of H2O2 production in electrochemical cells, a 1 cm2 Pt sheet, used as the counter electrode, was inserted into the anode chamber 1 cm away facing to the cation exchange membrane. Electrolyte (50 mM Na2SO4) [11,20] was aerated and fed at a rate of 1.5 L min1. Electrolysis was carried out at poised potentials (0.4 E 1.4 V vs. Ag/AgCl) with H2O2 concentration monitored every 10 min during 120 min. In order to evaluate the contribution of gas diffused through this hybrid air-cathode on H2O2 production, GDL uncovered and blocked systems were established. For the GDL uncovered system, the cathode chamber was continuously provided with oxygen saturated electrolyte at rate of 10 mL min1 (Dissolved þ Diffusion) or deoxygenated electrolyte (Only Diffusion). For GDL blocked systems, the GDL was covered with plexiglas and sealed with silica gel ensuring no air outside can diffuse through the GDL (Only Dissolved). This system was fed with oxygen saturated electrolyte. Cathode potentials were poised at 0.4, 0.6, 0.8 and 1.0 V vs. Ag/AgCl. H2O2 concentrations in cathodic effluent were monitored every 10 min.
2.4. Chemical and electrochemical analysis Dissolved oxygen (DO) value was measured by JPB-607A (INESA, Scientific Instrument Co. Ltd, Shanghai China). The concentrations of H2O2 in the effluent of cathode chamber were determined by a spectrophotometer (T6-1650F, Persee Instrument Co. Ltd, Beijing, China), using potassium titanium (IV) oxalate as a colored indicator. The current efficiency (CE) of the electrolytic cell, also called as Faradic efficiency, was calculated from Eq. (3)
CE ¼
nFCH2 O2 V 100% It
(3)
where n is the number of electrons transferred for oxygen reduction to H2O2, F is the Faraday constant (96,486C mol1), CH2O2 stands for the concentration of H2O2 (mol L1), V is the bulk volume (L), I is the current (A), and t is the time (s). Linear sweep voltammetry (LSV) analysis was performed using a potentiostat (CHI660D, ChenHua Instruments Co., Ltd., Shanghai, China) in 50 mM Na2SO4 solution at a scan rate of 10 mV s1 at ambient temperature with CB&G air-cathode as the working electrode, a platinum sheet of 1 cm2 as the counter electrode and an Ag/ AgCl (3.5 M KCl, 0.197 V vs. SHE) as the reference electrode. Potentials mentioned in this manuscript were using Ag/AgCl as the reference except as indicated. All the experiments were carried out at least in duplicate and the mean values were reported. 3. Results and discussion 3.1. Morphology and porosity of the different cathodes According to SEM images of eight different CLs, it can be clearly seen that the main skeleton was carbonaceous fragment banded by massive PTFE fibers (Fig. 2). PCB was consisted of non-uniform aggregations of CB particles and PTFE (Fig. 2a), forming a rough and porous surface. Dissimilarly, the PG consisted of sheet-like graphite particles exhibited a flat surface (Fig. 2h). Graphite sheet was non-uniformly embedded into hybrid CLs with a size approximately hundred times larger than the CB particles (Fig. 2beg). Graphite particles became predominate from Fig. 2beg with the decrease of CB ratios. N2 sorption isotherms on CB belongs to Type V according to the IUPAC classification (Fig. 3), which represents mesoporous
498
N. Li et al. / Journal of Power Sources 306 (2016) 495e502
Fig. 2. SEM images of PCB (a), CB&G4:1 (b), CB&G2:1 (c), CB&G1:1 (d), CB&G1:3 (e), CB&G1:5 (f), CB&G1:7 (g) and PG (h) at a magnification of 15 K.
adsorbent with weak affinities. Whereas, N2 sorption isotherms on graphite powder belongs to Type Ⅲ, indicating that graphite powder is macroporous adsorbent with weak affinities. Furthermore, the hysteresis loops of the carbon black powder were type H3, with a steeper shape at higher relative pressures, demonstrating more mesopores, and probably macropore adsorption [21], corresponding to the morphology shown in the SEM images. The porous properties and pore size distribution were summarized in Table 1. The results indicate that the specific area and pore volume of carbon black powder (79.5 m2 g1, 0.24 cm3 g1) were 14 times and 11 times larger than those of graphite powder (5.32 m2 g1, 0.02 cm3 g1).
Based on mercury intrusion porosimetry analysis, the total pore area of PCB (47.8 m2 g1) was 11 times higher than that of PG (4.05 m2 g1), resulting in a 65% increase in porosity from 46% to 76% (Table 2). The average pore diameter of PCB (168 nm) was 68% smaller than PG (530 nm), indicating that larger pores were formed by graphite-PTFE aggregations than CB-PTFE aggregations in CLs, although PCB has a much higher total pore area and porosity, which was consistent with the SEM images in Fig. 2. The increase of CB ratio brought an increase in total pore area and porosity, while a higher graphite content resulted in a larger average pore diameter (Table 2). It was showed that the addition of CB to graphite can produce more small pores between large graphite particles as well
N. Li et al. / Journal of Power Sources 306 (2016) 495e502
Fig. 3. N2 adsorption/desorption isotherms of carbon black powder and graphite powder.
499
activated carbon-PTFE and graphite-PTFE particle electrodes. It was found that the reduction of H2O2 on the porous surface like activated carbon and carbon black could lead to a greater loss of H2O2 than on the graphite, which was in high accordance with our results. The slopes of production rate curve of PG showed two stages including the fast increase from 0.4 to 1.0 V and a relatively slow increase from 1.0 to 1.4 V. Meanwhile, the slopes of another four hybrid air-cathodes (CB&G1:1, CB&G1:3, CB&G1:5, CB&G1:7) did not exhibit the second stage. This phenomenon can be attributed to the additional pore area introduced by CB addition. Thus, a great number of mesoporous pores and more inner sheet-like graphite particles could be provided as the reaction interface to generate H2O2 at a larger current and a more negative potential. Besides, these additional pores may also provide more channels for oxygen diffusion, which alleviated the oxygen concentration polarization at the high current region. Among all hybrid air-cathodes, CB&G1:5 showed the highest yield of H2O2 at all given potentials, indicating that CB&G1:5 kept a good balance on the optimized mass ratio of catalysts, specific area and porosity. At 1.4 V, the H2O2 production
Table 1 Porous structural characteristics of carbon black powder and graphite powder based on N2 adsorption/desorption analysis.
BET Surface Area (m2g1) t-Plot Micropore Area (m2g1) t-Plot External Surface Area (m2g1) BJH Desorption average pore diameter (nm) BJH Desorption cumulative volume of pores (cm3g1)
Carbon black powder
Graphite powder
79.5 1.25 78.3 12.1 0.24
5.32 0.04 5.28 14.9 0.02
Table 2 Mercury porosimetric analysis for CLs of air-cathodes with different mass ratio of carbon black and graphite. Sample 1
Total intrusion volume (mL g Average pore diameter (nm) Porosity (%) Total pore area (m2 g1)
)
PCB
CB&G 4:1
CB&G 2:1
CB&G 1:1
CB&G1:3
CB&G1:5
CB&G1:7
PG
2.11 168 75.9 47.8
2.10 194 75.6 43.4
2.07 213 75.2 38.8
1.48 251 68.4 31.0
1.23 288 66.2 17.1
1.18 330 62.9 14.3
0.58 355 48.7 9.2
0.53 530 45.5 4.1
as clog large pores, forming a new pore structure for CLs. 3.2. Electrochemical generation of H2O2 on different cathodes H2O2 yield in abiotic electrochemical cells had a nearly linear increase from 3.2 ± 0.2 mg L1 to 39.8 ± 10.4 mg L1 when the potential decreased from 0.4 to 1.4 V for all cathodes (vs. Ag/ AgCl, Fig. 4a). Although the specific surface area of PCB was 10.8 times larger than that of PG, the H2O2 yields of PG were 8e16% higher than PCB over the tested potential window. Fig. 4b shows the CE of eight different cathodes as a function of the applied potential after 120 min of electrolysis. It can be seen that the CE of PG at all given potentials were 26e72% higher than that of PCB, showing that graphite was a better catalyst for H2O2 production for air-cathode in spite of its relatively low surface area than CB. It is known that the oxygen diffused first to the surface of the macropores and mesopores, and then to the inner micropores [22]. Hence, more micropores in the PCB can provide additional ORR sites and result in better performance of ORR, which means more oxygen would be reduced to H2O instead of H2O2. On the contrast, ORR seemed to proceed in two-electron rather than four-electron pathway on the sheet-like surface of PG. In the Ref [11], Chen has evaluated and compared the property of carbon black-PTFE,
rate of CB&G1:5 was 50.2 mg h1, (11.9 mg L1 h1 cm2) at the current density of 12.3 mA cm2, with the CE of 92%. Comparing to literature under the similar current density, H2O2 generated by Sheng's acetylene black-PTFE air-bubbling cathode reached 46.6 mg L1 after 2.5 h of electrolysis at 12 mA cm2 (18.2 mg L1 h1 cm2) [15], which was slightly higher than our result, possibly because their better oxygen supplement (sparging air directly into the reactor at 2 L min1) and enhanced mass transfer (magnetic stirring and pH ¼ 3). LSV was performed over the potential range from 0 to 1.4 V to further investigate the electrochemical performance of hybrid cathodes. No obvious current was found under the deoxygenated condition, showing that the cathodic current was mostly from the ORR with the formation of H2O2 [Eq. (2)] and H2O [Eq. (4)] (Fig. 5). The current of PG was higher than that of PCB at the potential range of 0 to 0.65 V, revealed that graphite had a higher ORR catalytic activity than CB. When the potentials were more negative than 0.65 V, the current of PCB was higher than PG, probably due to the further reduction of H2O2 to H2O [11]. The currents of four hybrid air-cathodes (CB&G4:1, CB&G1:3 CB&G1:5 CB&G1:7) were lower than that of PG over 0 to 1.0 V, and surpassed that at potentials more negative than 1.0 V. For the graphite predominant cathodes (CB&G1:3, CB&G1:5 and CB&G1:7), it is reasonable to
500
N. Li et al. / Journal of Power Sources 306 (2016) 495e502
believe that the increase of the surface area resulted in stronger electrochemical activity for H2O2 production. Whereas for the porous carbon black predominant cathode CB&G4:1, strong current response at more negative potentials was thought to be the consequence of four-electron ORR. All hybrid air-cathodes exhibited a similar trend in CE (Fig. 4) and current (Fig. 5), indicating that the current was mainly from H2O2 over the tested potential range. However, for PCB and PG, much higher currents were produced especially at more positive potentials, showing that these two cathodes were totally different from hybrid electrodes in terms of electrochemical behavior. As the optimized hybrid cathode, CB&G1:5 would be used in the following experiments. O2 þ 4Hþ þ 4e/ 2H2O
(4)
The H2O2 yield did not approach a plateau at 1.4 V (Fig. 4a), and most of the air-cathodes had a CE of 90 ± 3% (PCB had a low CE of 70%, CB&G4:1 had a CE of 78%, Fig. 4b), indicating that our aircathode had a wide potential window for electrochemical synthesizing of H2O2 where the H2O2 yield was in proportion to the current intensity [12]. As showed in Fig. 4b, CB&G1:5 and CB&G1:7 exhibited similar CE curves, indicating that the porous structure and catalytic activity possibly be other factors limiting CE except porous characteristics showed in Table 2. In order to have more information about this hybrid cathode, CB&G1:5 was tested at more negative potentials. The CE began to decline after potential reached 1.6 V. The further increase of the current density or electrolysis time always resulted in the decrease of the CE, H2O2 yield and the useful life of cathodes especially when the cathodic potential was more negative than 0.75 ± 0.1 V vs. saturated calomel electrode (0.71 ± 0.1 V vs. Ag/AgCl) [23]. It was ascribed to two reasons: 1) The increase in the cathode polarization accelerated the four-electron ORR that would be competitive with the H2O2 production. 2) The accumulation of H2O2 could also induce H2O2 turn to H2O by Eq. (5) [15], or undergo a chemical decomposition to O2 in the medium (homogeneous process) [Eq. (6)] [24].
Fig. 4. Applied potential (vs Ag/AgCl) as a function of H2O2 yields (a) and current efficiencies (b) of different hybrid air-cathodes in abiotic system after 120 min of electrolysis. Supporting electrolyte: 0.05 M Na2SO4, continuous flow rate: 10 mL min1.
Fig. 5. Linear sweep voltammetry of the cathodes at the scan rate of 10 mV/s obtained in 0.05 M Na2SO4 aqueous solution, continuous flow rate: 10 mL min1, pH ¼ 7.
H2O2 þ 2Hþ þ 2 e/2H2O
(5)
H2O2 /H2O þ 1/2 O2
(6)
Dissimilarly, the CE did not exhibit an obvious decrease until the cathodic potential reached 1.6 V in our system, where the continuous flow was thought to mainly contribute to the improved CE at relatively negative potentials to some extent. In order to confirm the hypothesis mentioned above, LSVs were performed after 10 min of electrolysis at 1.0 V in continuous mode (curve b) or batch mode (curve c) as showed in Fig. 6. LSVs with a longer time of electrolysis (20 min) at batch mode (curve d) and in deoxygenated electrolyte (curve a) were also tested. Compared to the weak current observed in deoxygenated system (curve a), curves b, c and d proved that currents were mainly from ORR. H2O2 concentrations at the beginning of LSVs were 51.7 mg L1 (curve d) > 36.5 mg L1 (curve c) > 24.2 mg L1 (curve b), and correspondent current efficiencies were 68% < 84% < 92%. Currents increased with the retention time in batch mode system (curve c and d), with values all higher than the continuous mode system (curve b), indicating that accumulation of H2O2 in the cathode chamber indeed increased the current and decreased the CE of H2O2 production because fourelectron ORRs occurred as showed in Eq. (5). As a result, the yields of H2O2 could not proportionally increase with the current density in the batch mode system. In our continuous system, H2O2 was transferred out before being reduced to H2O, which improved the CE by at least 34%.
N. Li et al. / Journal of Power Sources 306 (2016) 495e502
501
Fig. 7. The applied voltages as a function of average anode potentials (vs Ag/AgCl) and H2O2 yields in bioelectrochemical systems using CB&G1:5. Fig. 6. Linear sweep voltammetry (LSV) of CB&G1:5 at the scan rate of 10 mV/s obtained in 0.05 M Na2SO4. Before LSV, the system was poised at 1.0 V for 10 min with deoxygenated electrolyte and blocked cathode (curve a), 10 min continuous mode fed with oxygen saturated electrolyte (curve b), 10 min batch mode with oxygen saturated electrolyte (curve c) and 20 min batch mode with oxygen saturated electrolyte (curve d).
3.3. Contribution of gas diffusion on H2O2 production The H2O2 yields of CB&G1:5 from 0.4 to 1.0 V were summarized in Table 3. It can be seen that H2O2 yields by only air diffusion were 0.9e12 times higher than those by only dissolved oxygen at all given potentials. Especially at more negative potentials, oxygen provided by air diffusion through the GDL became predominant (92% at 0.8 V and 94% at 1.0 V), indicating that the only diffusion was sufficient and sustainable for a high yield of H2O2 at a high current where dissolved oxygen had limited contribution. The DO value of the electrolyte in the aeration tank was 7.2 ± 0.1 mg L1, which was much lower than that in the air (310 mg L1 under standard condition). Despite of oxygen mass transfer resistance from gas to water, it was demonstrated in Table 3 that the air diffusion had a much higher rate of oxygen supplement than the electrolyte mediated oxygen transfer. Therefore, for a fast production of H2O2 at more negative potentials for industries, sparging devices (air bubbling device, magnetic stirring device et al.) were not necessarily needed when using this hybrid air-cathode.
3.4. H2O2 production at different applied voltages in BESs The bioelectrochemical production of H2O2 through twoelectron pathway ORR was proven to be spontaneous process accompanied with an energy output [10e12]. However, under those conditions the H2O2 production proceeded at a relatively low rate. In order to improve the production rate of H2O2 in BES,
external voltage was usually applied by DC power sources. Here, before voltages of 0.2e0.8 V were applied, all single chambered cubic BESs with activated carbon air-cathode reached 0.59 ± 0.01 V at 1 kU and operated for more than 2 months. The H2O2 yield rose up rapidly with the applied voltage from 0.2 to 0.6 V (Fig 7), with values increased from 1.05 to 3.69 mg L1 h1. Notably, the production rate of H2O2 at 0.6 V reached 3.29 mg L1h1 (0.079 kg m3day1) with current density of 0.61 mA cm2 and CE of 72%, which was 4.7% higher than that in abiotic system (69%) at the same current density. It was interesting that the bacteria based system has a better performance than the traditional chemical cells, which could be attributed to the relatively lower cathode potential in BESs. When it was normalized to cathode area, the yield here was 0.47 mg L1h1cm2, which was comparable to those obtained by Fu et al. (0.37 mg L1h1cm2) and Rozendal et al. (0.44 mg L1h1cm2) using 3.8 and 19 times larger graphite granule anodes at the applied voltage of 0.5 V [10,12]. The further increase of the voltage to 0.7 and 0.8 V decreased the H2O2 yield to 2.9 and 2.59 mg L1h1. In previous reports, the decrease of the production rate of H2O2 at a high input voltage was attributed to the further reduction of H2O2 to water that competed with two-electron ORR [20]. However, the cathode in our system was proven to have relatively stable performance over the potential range from 0.2 to 1.6 V vs. Ag/AgCl with current densities ranged from 1.1 to 12.3 mA cm2, exhibited a high and stable H2O2 production activity. The possible limitation of the H2O2 yield in our system should be from the polarization of the bioanode, not the cathode in previous studies. As an evidence, a nonlinear faster increase in anode potential was observed when the applied voltage was higher than 0.6 V (Fig. 7). Comparing to the abiotic electrochemistry system, H2O2 production using BESs has advantages of low cost and high energy efficiency, especially when we use CB&G hybrid air-cathodes.
Table 3 Contributions of diffusion and dissolved oxygen to H2O2 yield at different cathode potentials after 120 min. Cathode potential
Dissolved þ diffusion
Only dissolved
vs Ag/AgCl
H2O2 yield mg L1 h1
H2O2 yield mg L1 h1
Fraction %
H2O2 yield mg L1 h1
Fraction %
0.4 0.6 0.8 1
5.1 18.47 31.77 43.65
1.78 2.53 2.88 3.13
35 15 9 7
3.37 16.2 29.35 41.15
66 88 92 94
Only diffusion
502
N. Li et al. / Journal of Power Sources 306 (2016) 495e502
Wastewater can be simultaneously treated in the anode chamber, making BESs environmentally friendly and sustainable. The key limitation of BESs is the current density provided by microbes on the anode. It has been proven that in the presence of biocatalyst, the performance of BESs for in situ H2O2 production relies on efficient microbial electron transfer to the anode [25,26]. And among the parameters that determine the anode performance, the nature of anode material and the configuration are crucial, which can influence both the microbial community and the interfacial electron transfer resistance [27]. Therefore, besides the development of cathodes, the advanced design of a scalable and efficient anode for in situ H2O2 production in BESs is also needed in the future.
4. Conclusion High H2O2 yield of 11.9 mg L1 h1 cm2 is obtained when the mass ratio of carbon black and graphite is optimized to 1:5 for hybrid air-cathode, which is mainly due to the balance of specific area and catalytic activity in catalyst layer. It was interesting that both the H2O2 yield and CE can be improved when a bioanode is utilized, indicating that the anodic behavior in this system needs more attention in future works. Air diffusion has a dominant contribution (>90%) at a high current density (high H2O2 yield), which shows that this hybrid air-cathode has a bright future for industrial H2O2 production.
Acknowledgments This research was supported by the MOE Innovative Research Team in University (IRT13024), National Natural Science Foundation of China (NSFC, Nos. 51208352 and 21577068), Tianjin Research Program of Application Foundation and Advanced Technology (No. 13JCQNJC09100) and Independent Innovation Research Fund of Tianjin University (No. 2014XRG-0095).
References [1] M. Panizza, G. Cerisola, Electrochim. Acta 54 (2008) 876e878. [2] A.R. Khataee, M. Safarpour, M. Zarei, S. Aber, J. Electroanal. Chem. 659 (2011) 63e68. [3] R. Hage, A. Lienke, Angew. Chem. Int. Ed. 45 (2005) 206e222. [4] E.M. Siedlecka, P. Stepnowski, Water Environ. Res. 78 (2006) 852e856. [5] M. Ksibi, Chem. Eng. J. 119 (2006) 161e165. [6] Y. Wang, X. Li, L. Zhen, H. Zhang, Y. Zhang, C. Wang, J. Hazard. Mater. 229e230 (2012) 115e121. [7] J.M. Campos-Martin, G. Blanco-Brieva, J.L. Fierro, Angew. Chem. Int. Ed. 45 (2006) 6962e6984. [8] W.R.P. Barros, T. Ereno, A.C. Tavares, M.R.V. Lanza, ChemElectroChem 2 (2015) 714e719. [9] H. Wang, H. Luo, P. Fallgren, S. Jin, Z.J. Ren, Biotechnol. Adv. 33 (2015) 317e334. [10] R.A. Rozendal, E. Leone, J. Keller, K. Rabaey, Electrochem. Commun. 11 (2009) 1752e1755. [11] J.Y. Chen, N. Li, L. Zhao, J. Power Sources 254 (2014) 316e322. [12] L. Fu, S.J. You, F.L. Yang, M.M. Gao, X.H. Fang, G.Q. Zhang, J. Chem. Technol. Biotechnol. 85 (2010) 715e719. [13] H. Ding, Y. Li, A. Lu, S. Jin, C. Quan, C. Wang, X. Wang, C. Zeng, Y. Yan, Bioresour. Technol. 101 (2010) 3500e3505. [14] H. Dong, H. Yu, H. Yu, N. Gao, X. Wang, J. Power Sources 232 (2013) 132e138. [15] Y. Sheng, Y. Zhao, X. Wang, R. Wang, T. Tang, Electrochim. Acta 133 (2014) 414e421. [16] Y. Sheng, S. Song, X. Wang, L. Song, C. Wang, H. Sun, X. Niu, Electrochim. Acta 56 (2011) 8651e8656. [17] H. Dong, H. Yu, X. Wang, Q. Zhou, J. Feng, Water Res. 46 (2012) 5777e5787. [18] Y. Zhang, X. Wang, X. Li, N. Gao, L. Wan, C. Feng, Q. Zhou, RSC Adv. 4 (2014) 42577e42580. [19] N. Li, Y.N. Liu, J.K. An, C.J. Feng, X. Wang, J. Power Sources 272 (2014) 895e899. [20] J.Y. Chen, L. Zhao, N. Li, H. Liu, J. Power Sources 287 (2015) 291e296. [21] H. Dong, H.B. Yu, X. Wang, Q.X. Zhou, J.W. Sun, J. Chem. Technol. Biotechnol. 88 (2013) 774e778. [22] H. Dong, H. Yu, X. Wang, Environ. Sci. Technol. 46 (2012) 13009e13015. [23] C.H. Feng, F.B. Li, H.J. Mai, X.Z. Li, Environ. Sci. Technol. 44 (2010) 1875e1880. s, J.A. Garrido, R.M. Rodríguez, E. Brillas, N. Oturan, M.A. Oturan, Appl. [24] I. Sire Catal. B Environ. 72 (2007) 382e394. [25] B.E. Logan, J.M. Regan, Trend. Microbiol. 14 (2006) 512e518. [26] R.P. Ramasamy, Z.Y. Ren, M.M. Mench, J.M. Regan, Biotechnol. Bioeng. 101 (2008) 101e108. [27] J.J. Sun, H.Z. Zhao, Q.Z. Yang, J. Song, A. Xue, Electrochim. Acta 55 (2010) 3041e3047.
Contents lists available at ScienceDirect
Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour
A novel carbon black graphite hybrid air-cathode for efficient hydrogen peroxide production in bioelectrochemical systems Nan Li a, Jingkun An a, Lean Zhou b, Tian Li b, Junhui Li b, Cuijuan Feng b, Xin Wang b, * a
Tianjin Key Lab of Indoor Air Environmental Quality Control, School of Environmental Science and Engineering, Tianjin University, No. 92 Weijin Road, Nankai District, Tianjin 300072, China b MOE Key Laboratory of Pollution Processes and Environmental Criteria, Tianjin Key Laboratory of Environmental Remediation and Pollution Control, College of Environmental Science and Engineering, Nankai University, No. 38 Tongyan Road, Jinnan District, Tianjin 300350, China
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
A hybrid air-cathode with the optimal carbon black/graphite ratio of 1:5 is made. The maximum H2O2 yield is 11.9 mg L1 h1 cm2. Continuous flow without H2O2 accumulation increases current efficiency. Oxygen for H2O2 synthesis is mainly contributed by air diffusion (66 e94%). The use of bioanode increases H2O2 yield and current efficiency.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 22 October 2015 Received in revised form 8 December 2015 Accepted 16 December 2015 Available online xxx
Carbon black and graphite hybrid air-cathode is proved to be effective for H2O2 production in bioelectrochemical systems. The optimal mass ratio of carbon black to graphite is 1:5 with the highest H2O2 yield of 11.9 mg L1 h1 cm2 (12.3 mA cm2). Continuous flow is found to improve the current efficiency due to the avoidance of H2O2 accumulation. In the biological system, the highest H2O2 yield reaches 3.29 mg L1h1 (0.079 kg m3day1) with a current efficiency of 72%, which is higher than the abiotic system at the same current density. H2O2 produced in this system is mainly from the oxygen diffused through this air-cathode (>66%), especially when a more negative cathode potential is applied (94% at 1.0 V). This hybrid air-cathode has advantages of high H2O2 yield, high current density and no need of aeration, which make the synthesis of H2O2 more efficient and economical. © 2015 Elsevier B.V. All rights reserved.
Keywords: Bioelectrochemical systems Microbial fuel cells Hydrogen peroxide Air-cathode Graphite Carbon black
1. Introduction Hydrogen peroxide (H2O2) has been widely known as an environmentally friendly chemical which leaves no hazardous residues,
* Corresponding author. E-mail address: [email protected] (X. Wang). http://dx.doi.org/10.1016/j.jpowsour.2015.12.078 0378-7753/© 2015 Elsevier B.V. All rights reserved.
since it decomposes only to water and oxygen. As a powerful and versatile chemical, H2O2 is effective throughout the pH range from 0 to 14 with high oxidation potential (E0 ¼ 1.763 V at pH ¼ 0 and E0 ¼ 0.878 V at pH ¼ 14) [1]. Hence, H2O2 is applied to numerous industrial areas such as chemical synthesis, pulp paper and textile bleaching, medical disinfection, treatment of wastewater and destruction of hazardous organic wastes [2e5]. In general, H2O2 is commercially produced on an industrial scale by anthraquinone
496
N. Li et al. / Journal of Power Sources 306 (2016) 495e502
oxidation, which needs multi-step procedure and significant energy input. Considering the transport, handling and storage of H2O2 are potentially hazardous, anthraquinone oxidation method is thought to be inefficient and insecure [6,7]. In recent decades, a lot of researches have demonstrated that H2O2 can be in situ generated by reduction of oxygen in alkaline [Eq. (1)], neutral or acidic solution [Eq. (2)] in an electrochemical system [8]. O2þH2Oþ2e/HO 2 þOH
(1)
O2þ2Hþþ2e/H2O2
(2)
Bioelectrochemical systems (BESs), also known as microbial electrochemical systems or MXCs, demonstrated the possibility of H2O2 production using renewable energy from wastewaters in recent years [9]. BES reactor consists of an anode, a cathode and a separator. Electrochemically active bacteria on the anode directly oxidize soluble organic waste to generate electron for the cathode. On the cathode, electrons can be used for four-electron oxygen reduction reaction (ORR) to water or through a two-electron pathway to produce H2O2 [10e12]. Due to the advantages of high conductivity, low price, good stability and low catalytic activity of H2O2 decomposition, carbonbased cathodes such as carbon clothes [10], graphite granules [11], spectrographically pure graphite (SPG) rods [12] and modified graphite [13] were usually used in the H2O2 electrosynthesis process in BESs. Most of these cathodes, such as our three-dimensional electrodes [11], were immersed in the electrolyte to provide the reaction zone for proton/electron and O2 that diffuses along with the electrolyte to synthesize H2O2. This kind of cathode has more potential use in refractory pollutant removal and can be easily scaled-up. However, the low solubility of O2 in the electrolyte is a key limitation where external power is also needed for cathodic aeration in these systems. Air-cathode is an advanced design where oxygen is passively supplied without aeration. It is usually composed of a catalyst layer (CL) facing to the electrolyte and a gas diffusion layer (GDL) facing to the air [14]. It had been demonstrated that H2O2 can be efficiently generated using air-cathodes [10] with a production rate 2 kg m3day1. However, in previous reports, most air-cathodes that utilized for H2O2 production didn't have a GDL [11,15], which may have a problem of electrolyte leakage and catalyst flooding when the system is scaled up. Apart from oxygen diffusion, the oxygen dissolved in electrolyte should be a considerable amendment for cathodic H2O2 produce [15,16]. However, to the best of our knowledge, the contribution of dissolved oxygen to H2O2 in aircathode BES has not been investigated yet. In the present work, BESs equipped with novel carbon black and graphite hybrid aircathodes made by rolling-press method were utilized to produce H2O2. The performances of cathodes with different carbon black/ graphite mass ratio were analyzed. The contributions of dissolved and diffused oxygen on H2O2 production were also evaluated. 2. Experimental 2.1. Carbon black and graphite (CB&G) hybrid air-cathode The CB&G air-cathode consisted of a CL and a GDL with the stainless steel mesh as the current collector and matrix. CL was prepared according to the following procedures. Carbon black (CB, 30 nm, Vulcan XC-72R, Cabot Corporation, US) and graphite (40 mm, HTF0325, >99.9%, Huatai Chemical Reagent Co. Ltd., Qingdao, China) were cleaned by ultrasonic in deionized water for 20 min at room temperature. After drying, a mixture of 6 g CB and graphite powders at different mass ratios (4.8 g CB and 1.2 g graphite as
CB&G4:1; 4 g CB and 2 g graphite as CB&G2:1; 3 g CB and 3 g graphite as CB&G1:1; 1.5 g CB and 4.5 g graphite as CB&G1:3; 1 g CB and 5 g graphite as CB&G1:5 and 0.75 g CB and 5.25 g graphite as CB&G1:7) was individually dispersed into 45 mL ethanol in a beaker in an ultrasonic bath at room temperature for 10 min. Polytetrafluoroethylene (PTFE) emulsion (60%, Horizon, Shanghai, China) as binder was slowly added to the mixture to fabricate the catalyst layer as previously described [17]. The mixture was then stirred at about 80 C to get a dough-like paste. The paste was firstly rolled to a 0.5 mm CL film and then roll-pressed on one side of the stainless steel mesh (4 cm 4 cm, Type 304N, 60 meshes, Detiannuo Commercial Trade Co. Ltd., Tianjin, China) to be a flat sheet (0.5 mm in thickness). CLs made of pure CB and graphite were also fabricated and marked as PCB and PG. The GDLs of each cathode were made in parallel by rolling a mixture of carbon black and PTFE emulsion with a mass ratio of 4:9 according to the procedure described previously [18]. After heating at 340 C for 25 min, GDLs were rolled onto the opposite side of different CLs to form a final air-cathode with a total thickness of 1 mm [18]. 2.2. Material characterization The specific surface area and pore parameters of carbon black and graphite powders were determined by the multipoint BrunauereEmmetteTeller (BET) measurement at 77 K using Autosorb1 (Quantachrome, USA). Porous properties and the pore size distribution were analyzed based on the BJH model. The porous structure of the CLs were characterized based on capillary law with a mercury porosimeter (Autopore IV, Micromeritics) which can analyze porous media with a pore size ranging from 6 nm to 300 mm. Samples were firstly dried in an oven at 105 C for 3 h in a stream of nitrogen, and then intruded volume analysis was carried out over the pressure ranging from 5 kPa to 414 MPa. After each change in pressure, the system equilibrated for 15 s to minimize overlap of the intra- (0.002e1 mm pore radius) and inter(1e100 mm pore radius) porosity regions as determined by the porosimeter. Images of the surface morphology of CLs were taken with a scanning electron microscope (SEM, Nanosem 430, USA). 2.3. BES configuration and operation The experiments were performed on dual-chamber reactors shown in Fig. 1 equipped with different CB&G air-cathodes. The anode chamber was 3 cm in diameter and 4 cm in length (net volume of 28 mL) while the cathode chamber was 3 cm in diameter and 2 cm in length (net volume of 14 mL) [19]. Both chambers were assembled tightly on two sides of a cation exchange membrane (Ultrex CMI-7000, Membranes International Inc., Glen Rock, NJ, USA) by four clamping bolts. Anodes were made of carbon fiber brushes (3 cm in diameter and 2 cm in length). Different CB&G aircathodes were assembled to the other side of cathode chambers. Two 4 mm diameter holes were drilled on the top and at the bottom of the cathode chamber connecting to the effluent beaker and the influent aeration tank as showed in Fig. 1. N2 gas or air was continuously aerated into the aeration tank to obtain an oxygen free or oxygen saturated influent. Peristaltic pump (BT100L, 8 channels, Longer Precision Pump Co. Ltd, Baoding, China) was used to feed the cathode with electrolyte from aeration tank at the rate of 10 mL min1. All anode chambers of BESs were inoculated with the effluent from MFCs and pre-acclimated under 1 kU of external resistance with activated carbon air-cathode for 2 months before coupling with cathode chambers [19]. The medium contained acetate (1.0 g L1), a phosphate buffer solution (50 mM PBS; NH4Cl 0.31 g L1, KCl 0.13 g L1, NaH2PO4 2H2O 2.772 g L1,Na2HPO4
N. Li et al. / Journal of Power Sources 306 (2016) 495e502
497
Fig. 1. Schematic diagram of the system utilized for H2O2 yield measurement using different carbon black/graphite hybrid air-cathodes.
4.576 g L1), a trace mineral (12.5 mL L1) and a vitamin solution (5 mL L1) [17]. When the acclimation of anodic biofilm was completed, cathode chamber and the aeration tank were filled with a 50 mM Na2SO4 solution. For the abiotic tests of H2O2 production in electrochemical cells, a 1 cm2 Pt sheet, used as the counter electrode, was inserted into the anode chamber 1 cm away facing to the cation exchange membrane. Electrolyte (50 mM Na2SO4) [11,20] was aerated and fed at a rate of 1.5 L min1. Electrolysis was carried out at poised potentials (0.4 E 1.4 V vs. Ag/AgCl) with H2O2 concentration monitored every 10 min during 120 min. In order to evaluate the contribution of gas diffused through this hybrid air-cathode on H2O2 production, GDL uncovered and blocked systems were established. For the GDL uncovered system, the cathode chamber was continuously provided with oxygen saturated electrolyte at rate of 10 mL min1 (Dissolved þ Diffusion) or deoxygenated electrolyte (Only Diffusion). For GDL blocked systems, the GDL was covered with plexiglas and sealed with silica gel ensuring no air outside can diffuse through the GDL (Only Dissolved). This system was fed with oxygen saturated electrolyte. Cathode potentials were poised at 0.4, 0.6, 0.8 and 1.0 V vs. Ag/AgCl. H2O2 concentrations in cathodic effluent were monitored every 10 min.
2.4. Chemical and electrochemical analysis Dissolved oxygen (DO) value was measured by JPB-607A (INESA, Scientific Instrument Co. Ltd, Shanghai China). The concentrations of H2O2 in the effluent of cathode chamber were determined by a spectrophotometer (T6-1650F, Persee Instrument Co. Ltd, Beijing, China), using potassium titanium (IV) oxalate as a colored indicator. The current efficiency (CE) of the electrolytic cell, also called as Faradic efficiency, was calculated from Eq. (3)
CE ¼
nFCH2 O2 V 100% It
(3)
where n is the number of electrons transferred for oxygen reduction to H2O2, F is the Faraday constant (96,486C mol1), CH2O2 stands for the concentration of H2O2 (mol L1), V is the bulk volume (L), I is the current (A), and t is the time (s). Linear sweep voltammetry (LSV) analysis was performed using a potentiostat (CHI660D, ChenHua Instruments Co., Ltd., Shanghai, China) in 50 mM Na2SO4 solution at a scan rate of 10 mV s1 at ambient temperature with CB&G air-cathode as the working electrode, a platinum sheet of 1 cm2 as the counter electrode and an Ag/ AgCl (3.5 M KCl, 0.197 V vs. SHE) as the reference electrode. Potentials mentioned in this manuscript were using Ag/AgCl as the reference except as indicated. All the experiments were carried out at least in duplicate and the mean values were reported. 3. Results and discussion 3.1. Morphology and porosity of the different cathodes According to SEM images of eight different CLs, it can be clearly seen that the main skeleton was carbonaceous fragment banded by massive PTFE fibers (Fig. 2). PCB was consisted of non-uniform aggregations of CB particles and PTFE (Fig. 2a), forming a rough and porous surface. Dissimilarly, the PG consisted of sheet-like graphite particles exhibited a flat surface (Fig. 2h). Graphite sheet was non-uniformly embedded into hybrid CLs with a size approximately hundred times larger than the CB particles (Fig. 2beg). Graphite particles became predominate from Fig. 2beg with the decrease of CB ratios. N2 sorption isotherms on CB belongs to Type V according to the IUPAC classification (Fig. 3), which represents mesoporous
498
N. Li et al. / Journal of Power Sources 306 (2016) 495e502
Fig. 2. SEM images of PCB (a), CB&G4:1 (b), CB&G2:1 (c), CB&G1:1 (d), CB&G1:3 (e), CB&G1:5 (f), CB&G1:7 (g) and PG (h) at a magnification of 15 K.
adsorbent with weak affinities. Whereas, N2 sorption isotherms on graphite powder belongs to Type Ⅲ, indicating that graphite powder is macroporous adsorbent with weak affinities. Furthermore, the hysteresis loops of the carbon black powder were type H3, with a steeper shape at higher relative pressures, demonstrating more mesopores, and probably macropore adsorption [21], corresponding to the morphology shown in the SEM images. The porous properties and pore size distribution were summarized in Table 1. The results indicate that the specific area and pore volume of carbon black powder (79.5 m2 g1, 0.24 cm3 g1) were 14 times and 11 times larger than those of graphite powder (5.32 m2 g1, 0.02 cm3 g1).
Based on mercury intrusion porosimetry analysis, the total pore area of PCB (47.8 m2 g1) was 11 times higher than that of PG (4.05 m2 g1), resulting in a 65% increase in porosity from 46% to 76% (Table 2). The average pore diameter of PCB (168 nm) was 68% smaller than PG (530 nm), indicating that larger pores were formed by graphite-PTFE aggregations than CB-PTFE aggregations in CLs, although PCB has a much higher total pore area and porosity, which was consistent with the SEM images in Fig. 2. The increase of CB ratio brought an increase in total pore area and porosity, while a higher graphite content resulted in a larger average pore diameter (Table 2). It was showed that the addition of CB to graphite can produce more small pores between large graphite particles as well
N. Li et al. / Journal of Power Sources 306 (2016) 495e502
Fig. 3. N2 adsorption/desorption isotherms of carbon black powder and graphite powder.
499
activated carbon-PTFE and graphite-PTFE particle electrodes. It was found that the reduction of H2O2 on the porous surface like activated carbon and carbon black could lead to a greater loss of H2O2 than on the graphite, which was in high accordance with our results. The slopes of production rate curve of PG showed two stages including the fast increase from 0.4 to 1.0 V and a relatively slow increase from 1.0 to 1.4 V. Meanwhile, the slopes of another four hybrid air-cathodes (CB&G1:1, CB&G1:3, CB&G1:5, CB&G1:7) did not exhibit the second stage. This phenomenon can be attributed to the additional pore area introduced by CB addition. Thus, a great number of mesoporous pores and more inner sheet-like graphite particles could be provided as the reaction interface to generate H2O2 at a larger current and a more negative potential. Besides, these additional pores may also provide more channels for oxygen diffusion, which alleviated the oxygen concentration polarization at the high current region. Among all hybrid air-cathodes, CB&G1:5 showed the highest yield of H2O2 at all given potentials, indicating that CB&G1:5 kept a good balance on the optimized mass ratio of catalysts, specific area and porosity. At 1.4 V, the H2O2 production
Table 1 Porous structural characteristics of carbon black powder and graphite powder based on N2 adsorption/desorption analysis.
BET Surface Area (m2g1) t-Plot Micropore Area (m2g1) t-Plot External Surface Area (m2g1) BJH Desorption average pore diameter (nm) BJH Desorption cumulative volume of pores (cm3g1)
Carbon black powder
Graphite powder
79.5 1.25 78.3 12.1 0.24
5.32 0.04 5.28 14.9 0.02
Table 2 Mercury porosimetric analysis for CLs of air-cathodes with different mass ratio of carbon black and graphite. Sample 1
Total intrusion volume (mL g Average pore diameter (nm) Porosity (%) Total pore area (m2 g1)
)
PCB
CB&G 4:1
CB&G 2:1
CB&G 1:1
CB&G1:3
CB&G1:5
CB&G1:7
PG
2.11 168 75.9 47.8
2.10 194 75.6 43.4
2.07 213 75.2 38.8
1.48 251 68.4 31.0
1.23 288 66.2 17.1
1.18 330 62.9 14.3
0.58 355 48.7 9.2
0.53 530 45.5 4.1
as clog large pores, forming a new pore structure for CLs. 3.2. Electrochemical generation of H2O2 on different cathodes H2O2 yield in abiotic electrochemical cells had a nearly linear increase from 3.2 ± 0.2 mg L1 to 39.8 ± 10.4 mg L1 when the potential decreased from 0.4 to 1.4 V for all cathodes (vs. Ag/ AgCl, Fig. 4a). Although the specific surface area of PCB was 10.8 times larger than that of PG, the H2O2 yields of PG were 8e16% higher than PCB over the tested potential window. Fig. 4b shows the CE of eight different cathodes as a function of the applied potential after 120 min of electrolysis. It can be seen that the CE of PG at all given potentials were 26e72% higher than that of PCB, showing that graphite was a better catalyst for H2O2 production for air-cathode in spite of its relatively low surface area than CB. It is known that the oxygen diffused first to the surface of the macropores and mesopores, and then to the inner micropores [22]. Hence, more micropores in the PCB can provide additional ORR sites and result in better performance of ORR, which means more oxygen would be reduced to H2O instead of H2O2. On the contrast, ORR seemed to proceed in two-electron rather than four-electron pathway on the sheet-like surface of PG. In the Ref [11], Chen has evaluated and compared the property of carbon black-PTFE,
rate of CB&G1:5 was 50.2 mg h1, (11.9 mg L1 h1 cm2) at the current density of 12.3 mA cm2, with the CE of 92%. Comparing to literature under the similar current density, H2O2 generated by Sheng's acetylene black-PTFE air-bubbling cathode reached 46.6 mg L1 after 2.5 h of electrolysis at 12 mA cm2 (18.2 mg L1 h1 cm2) [15], which was slightly higher than our result, possibly because their better oxygen supplement (sparging air directly into the reactor at 2 L min1) and enhanced mass transfer (magnetic stirring and pH ¼ 3). LSV was performed over the potential range from 0 to 1.4 V to further investigate the electrochemical performance of hybrid cathodes. No obvious current was found under the deoxygenated condition, showing that the cathodic current was mostly from the ORR with the formation of H2O2 [Eq. (2)] and H2O [Eq. (4)] (Fig. 5). The current of PG was higher than that of PCB at the potential range of 0 to 0.65 V, revealed that graphite had a higher ORR catalytic activity than CB. When the potentials were more negative than 0.65 V, the current of PCB was higher than PG, probably due to the further reduction of H2O2 to H2O [11]. The currents of four hybrid air-cathodes (CB&G4:1, CB&G1:3 CB&G1:5 CB&G1:7) were lower than that of PG over 0 to 1.0 V, and surpassed that at potentials more negative than 1.0 V. For the graphite predominant cathodes (CB&G1:3, CB&G1:5 and CB&G1:7), it is reasonable to
500
N. Li et al. / Journal of Power Sources 306 (2016) 495e502
believe that the increase of the surface area resulted in stronger electrochemical activity for H2O2 production. Whereas for the porous carbon black predominant cathode CB&G4:1, strong current response at more negative potentials was thought to be the consequence of four-electron ORR. All hybrid air-cathodes exhibited a similar trend in CE (Fig. 4) and current (Fig. 5), indicating that the current was mainly from H2O2 over the tested potential range. However, for PCB and PG, much higher currents were produced especially at more positive potentials, showing that these two cathodes were totally different from hybrid electrodes in terms of electrochemical behavior. As the optimized hybrid cathode, CB&G1:5 would be used in the following experiments. O2 þ 4Hþ þ 4e/ 2H2O
(4)
The H2O2 yield did not approach a plateau at 1.4 V (Fig. 4a), and most of the air-cathodes had a CE of 90 ± 3% (PCB had a low CE of 70%, CB&G4:1 had a CE of 78%, Fig. 4b), indicating that our aircathode had a wide potential window for electrochemical synthesizing of H2O2 where the H2O2 yield was in proportion to the current intensity [12]. As showed in Fig. 4b, CB&G1:5 and CB&G1:7 exhibited similar CE curves, indicating that the porous structure and catalytic activity possibly be other factors limiting CE except porous characteristics showed in Table 2. In order to have more information about this hybrid cathode, CB&G1:5 was tested at more negative potentials. The CE began to decline after potential reached 1.6 V. The further increase of the current density or electrolysis time always resulted in the decrease of the CE, H2O2 yield and the useful life of cathodes especially when the cathodic potential was more negative than 0.75 ± 0.1 V vs. saturated calomel electrode (0.71 ± 0.1 V vs. Ag/AgCl) [23]. It was ascribed to two reasons: 1) The increase in the cathode polarization accelerated the four-electron ORR that would be competitive with the H2O2 production. 2) The accumulation of H2O2 could also induce H2O2 turn to H2O by Eq. (5) [15], or undergo a chemical decomposition to O2 in the medium (homogeneous process) [Eq. (6)] [24].
Fig. 4. Applied potential (vs Ag/AgCl) as a function of H2O2 yields (a) and current efficiencies (b) of different hybrid air-cathodes in abiotic system after 120 min of electrolysis. Supporting electrolyte: 0.05 M Na2SO4, continuous flow rate: 10 mL min1.
Fig. 5. Linear sweep voltammetry of the cathodes at the scan rate of 10 mV/s obtained in 0.05 M Na2SO4 aqueous solution, continuous flow rate: 10 mL min1, pH ¼ 7.
H2O2 þ 2Hþ þ 2 e/2H2O
(5)
H2O2 /H2O þ 1/2 O2
(6)
Dissimilarly, the CE did not exhibit an obvious decrease until the cathodic potential reached 1.6 V in our system, where the continuous flow was thought to mainly contribute to the improved CE at relatively negative potentials to some extent. In order to confirm the hypothesis mentioned above, LSVs were performed after 10 min of electrolysis at 1.0 V in continuous mode (curve b) or batch mode (curve c) as showed in Fig. 6. LSVs with a longer time of electrolysis (20 min) at batch mode (curve d) and in deoxygenated electrolyte (curve a) were also tested. Compared to the weak current observed in deoxygenated system (curve a), curves b, c and d proved that currents were mainly from ORR. H2O2 concentrations at the beginning of LSVs were 51.7 mg L1 (curve d) > 36.5 mg L1 (curve c) > 24.2 mg L1 (curve b), and correspondent current efficiencies were 68% < 84% < 92%. Currents increased with the retention time in batch mode system (curve c and d), with values all higher than the continuous mode system (curve b), indicating that accumulation of H2O2 in the cathode chamber indeed increased the current and decreased the CE of H2O2 production because fourelectron ORRs occurred as showed in Eq. (5). As a result, the yields of H2O2 could not proportionally increase with the current density in the batch mode system. In our continuous system, H2O2 was transferred out before being reduced to H2O, which improved the CE by at least 34%.
N. Li et al. / Journal of Power Sources 306 (2016) 495e502
501
Fig. 7. The applied voltages as a function of average anode potentials (vs Ag/AgCl) and H2O2 yields in bioelectrochemical systems using CB&G1:5. Fig. 6. Linear sweep voltammetry (LSV) of CB&G1:5 at the scan rate of 10 mV/s obtained in 0.05 M Na2SO4. Before LSV, the system was poised at 1.0 V for 10 min with deoxygenated electrolyte and blocked cathode (curve a), 10 min continuous mode fed with oxygen saturated electrolyte (curve b), 10 min batch mode with oxygen saturated electrolyte (curve c) and 20 min batch mode with oxygen saturated electrolyte (curve d).
3.3. Contribution of gas diffusion on H2O2 production The H2O2 yields of CB&G1:5 from 0.4 to 1.0 V were summarized in Table 3. It can be seen that H2O2 yields by only air diffusion were 0.9e12 times higher than those by only dissolved oxygen at all given potentials. Especially at more negative potentials, oxygen provided by air diffusion through the GDL became predominant (92% at 0.8 V and 94% at 1.0 V), indicating that the only diffusion was sufficient and sustainable for a high yield of H2O2 at a high current where dissolved oxygen had limited contribution. The DO value of the electrolyte in the aeration tank was 7.2 ± 0.1 mg L1, which was much lower than that in the air (310 mg L1 under standard condition). Despite of oxygen mass transfer resistance from gas to water, it was demonstrated in Table 3 that the air diffusion had a much higher rate of oxygen supplement than the electrolyte mediated oxygen transfer. Therefore, for a fast production of H2O2 at more negative potentials for industries, sparging devices (air bubbling device, magnetic stirring device et al.) were not necessarily needed when using this hybrid air-cathode.
3.4. H2O2 production at different applied voltages in BESs The bioelectrochemical production of H2O2 through twoelectron pathway ORR was proven to be spontaneous process accompanied with an energy output [10e12]. However, under those conditions the H2O2 production proceeded at a relatively low rate. In order to improve the production rate of H2O2 in BES,
external voltage was usually applied by DC power sources. Here, before voltages of 0.2e0.8 V were applied, all single chambered cubic BESs with activated carbon air-cathode reached 0.59 ± 0.01 V at 1 kU and operated for more than 2 months. The H2O2 yield rose up rapidly with the applied voltage from 0.2 to 0.6 V (Fig 7), with values increased from 1.05 to 3.69 mg L1 h1. Notably, the production rate of H2O2 at 0.6 V reached 3.29 mg L1h1 (0.079 kg m3day1) with current density of 0.61 mA cm2 and CE of 72%, which was 4.7% higher than that in abiotic system (69%) at the same current density. It was interesting that the bacteria based system has a better performance than the traditional chemical cells, which could be attributed to the relatively lower cathode potential in BESs. When it was normalized to cathode area, the yield here was 0.47 mg L1h1cm2, which was comparable to those obtained by Fu et al. (0.37 mg L1h1cm2) and Rozendal et al. (0.44 mg L1h1cm2) using 3.8 and 19 times larger graphite granule anodes at the applied voltage of 0.5 V [10,12]. The further increase of the voltage to 0.7 and 0.8 V decreased the H2O2 yield to 2.9 and 2.59 mg L1h1. In previous reports, the decrease of the production rate of H2O2 at a high input voltage was attributed to the further reduction of H2O2 to water that competed with two-electron ORR [20]. However, the cathode in our system was proven to have relatively stable performance over the potential range from 0.2 to 1.6 V vs. Ag/AgCl with current densities ranged from 1.1 to 12.3 mA cm2, exhibited a high and stable H2O2 production activity. The possible limitation of the H2O2 yield in our system should be from the polarization of the bioanode, not the cathode in previous studies. As an evidence, a nonlinear faster increase in anode potential was observed when the applied voltage was higher than 0.6 V (Fig. 7). Comparing to the abiotic electrochemistry system, H2O2 production using BESs has advantages of low cost and high energy efficiency, especially when we use CB&G hybrid air-cathodes.
Table 3 Contributions of diffusion and dissolved oxygen to H2O2 yield at different cathode potentials after 120 min. Cathode potential
Dissolved þ diffusion
Only dissolved
vs Ag/AgCl
H2O2 yield mg L1 h1
H2O2 yield mg L1 h1
Fraction %
H2O2 yield mg L1 h1
Fraction %
0.4 0.6 0.8 1
5.1 18.47 31.77 43.65
1.78 2.53 2.88 3.13
35 15 9 7
3.37 16.2 29.35 41.15
66 88 92 94
Only diffusion
502
N. Li et al. / Journal of Power Sources 306 (2016) 495e502
Wastewater can be simultaneously treated in the anode chamber, making BESs environmentally friendly and sustainable. The key limitation of BESs is the current density provided by microbes on the anode. It has been proven that in the presence of biocatalyst, the performance of BESs for in situ H2O2 production relies on efficient microbial electron transfer to the anode [25,26]. And among the parameters that determine the anode performance, the nature of anode material and the configuration are crucial, which can influence both the microbial community and the interfacial electron transfer resistance [27]. Therefore, besides the development of cathodes, the advanced design of a scalable and efficient anode for in situ H2O2 production in BESs is also needed in the future.
4. Conclusion High H2O2 yield of 11.9 mg L1 h1 cm2 is obtained when the mass ratio of carbon black and graphite is optimized to 1:5 for hybrid air-cathode, which is mainly due to the balance of specific area and catalytic activity in catalyst layer. It was interesting that both the H2O2 yield and CE can be improved when a bioanode is utilized, indicating that the anodic behavior in this system needs more attention in future works. Air diffusion has a dominant contribution (>90%) at a high current density (high H2O2 yield), which shows that this hybrid air-cathode has a bright future for industrial H2O2 production.
Acknowledgments This research was supported by the MOE Innovative Research Team in University (IRT13024), National Natural Science Foundation of China (NSFC, Nos. 51208352 and 21577068), Tianjin Research Program of Application Foundation and Advanced Technology (No. 13JCQNJC09100) and Independent Innovation Research Fund of Tianjin University (No. 2014XRG-0095).
References [1] M. Panizza, G. Cerisola, Electrochim. Acta 54 (2008) 876e878. [2] A.R. Khataee, M. Safarpour, M. Zarei, S. Aber, J. Electroanal. Chem. 659 (2011) 63e68. [3] R. Hage, A. Lienke, Angew. Chem. Int. Ed. 45 (2005) 206e222. [4] E.M. Siedlecka, P. Stepnowski, Water Environ. Res. 78 (2006) 852e856. [5] M. Ksibi, Chem. Eng. J. 119 (2006) 161e165. [6] Y. Wang, X. Li, L. Zhen, H. Zhang, Y. Zhang, C. Wang, J. Hazard. Mater. 229e230 (2012) 115e121. [7] J.M. Campos-Martin, G. Blanco-Brieva, J.L. Fierro, Angew. Chem. Int. Ed. 45 (2006) 6962e6984. [8] W.R.P. Barros, T. Ereno, A.C. Tavares, M.R.V. Lanza, ChemElectroChem 2 (2015) 714e719. [9] H. Wang, H. Luo, P. Fallgren, S. Jin, Z.J. Ren, Biotechnol. Adv. 33 (2015) 317e334. [10] R.A. Rozendal, E. Leone, J. Keller, K. Rabaey, Electrochem. Commun. 11 (2009) 1752e1755. [11] J.Y. Chen, N. Li, L. Zhao, J. Power Sources 254 (2014) 316e322. [12] L. Fu, S.J. You, F.L. Yang, M.M. Gao, X.H. Fang, G.Q. Zhang, J. Chem. Technol. Biotechnol. 85 (2010) 715e719. [13] H. Ding, Y. Li, A. Lu, S. Jin, C. Quan, C. Wang, X. Wang, C. Zeng, Y. Yan, Bioresour. Technol. 101 (2010) 3500e3505. [14] H. Dong, H. Yu, H. Yu, N. Gao, X. Wang, J. Power Sources 232 (2013) 132e138. [15] Y. Sheng, Y. Zhao, X. Wang, R. Wang, T. Tang, Electrochim. Acta 133 (2014) 414e421. [16] Y. Sheng, S. Song, X. Wang, L. Song, C. Wang, H. Sun, X. Niu, Electrochim. Acta 56 (2011) 8651e8656. [17] H. Dong, H. Yu, X. Wang, Q. Zhou, J. Feng, Water Res. 46 (2012) 5777e5787. [18] Y. Zhang, X. Wang, X. Li, N. Gao, L. Wan, C. Feng, Q. Zhou, RSC Adv. 4 (2014) 42577e42580. [19] N. Li, Y.N. Liu, J.K. An, C.J. Feng, X. Wang, J. Power Sources 272 (2014) 895e899. [20] J.Y. Chen, L. Zhao, N. Li, H. Liu, J. Power Sources 287 (2015) 291e296. [21] H. Dong, H.B. Yu, X. Wang, Q.X. Zhou, J.W. Sun, J. Chem. Technol. Biotechnol. 88 (2013) 774e778. [22] H. Dong, H. Yu, X. Wang, Environ. Sci. Technol. 46 (2012) 13009e13015. [23] C.H. Feng, F.B. Li, H.J. Mai, X.Z. Li, Environ. Sci. Technol. 44 (2010) 1875e1880. s, J.A. Garrido, R.M. Rodríguez, E. Brillas, N. Oturan, M.A. Oturan, Appl. [24] I. Sire Catal. B Environ. 72 (2007) 382e394. [25] B.E. Logan, J.M. Regan, Trend. Microbiol. 14 (2006) 512e518. [26] R.P. Ramasamy, Z.Y. Ren, M.M. Mench, J.M. Regan, Biotechnol. Bioeng. 101 (2008) 101e108. [27] J.J. Sun, H.Z. Zhao, Q.Z. Yang, J. Song, A. Xue, Electrochim. Acta 55 (2010) 3041e3047.