- Email: [email protected]
Tubular direct carbon solid oxide fuel cells with molten antimony anode and refueling feasibility
Energy 95 (2016) 274e278
Contents lists available at ScienceDirect
Energy journal homepage: www.elsevier.com/locate/energy
Tubular direct carbon solid oxide fuel cells with molten antimony anode and refueling feasibility Nan-Qi Duan a, Yong Cao a, Bin Hua b, Bo Chi a, Jian Pu a, Jingli Luo b, Li Jian a, c, * a Center for Fuel Cell Innovation, State Key Laboratory of Coal Combustion, School of Materials Science and Engineering, Huazhong University of Science and Technology, Wuhan, Hubei 430074, China b Department of Chemical and Materials Engineering, University of Alberta, Edmonton, Alberta T6G 2G6, Canada c Research Institute of Huazhong University of Science and Technology in Shenzhen, Shenzhen, Guangdong 518000, China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 5 June 2015 Received in revised form 9 October 2015 Accepted 11 October 2015 Available online 31 December 2015
Tubular direct carbon SOFCs (solid oxide fuel cells) supported by YSZ (Y2O3 stabilized ZrO2) electrolyte are fabricated by slurry-casting, slurry-dipping and sintering processes with La0.6Sr0.4Co0.2Fe0.8O310 mol.% Gd2O3 doped CeO2 (LSCF-10GDC) as the cathode. Their electrochemical performance is examined at temperatures from 700 to 800 C using molten antimony (Sb) anode and activated carbon fuel. The ohmic resistance of the cell is between 1.01 and 0.37 U cm2 mainly originated from the thick YSZ electrolyte (150 mm); the polarization resistance ranges from 0.22 to 0.06 U cm2. The maximum power density at 800 C is 304 mW cm2 and can be greatly increased by using a thinner and/or more conductive electrolyte. With 1 g activated carbon as the fuel, the cell performance is stable at 200 mW cm2 at 800 C for more than 6 h by chemical consumption (oxidization) of the carbon, which reduces the electrochemically formed Sb2O3 to Sb. The cell performance decreases as the fuel is used up and is recovered by refueling. © 2015 Elsevier Ltd. All rights reserved.
Keywords: Carbonaceous fuel Tubular solid oxide fuel cell Molten antimony anode Electrochemical performance Refueling
1. Introduction Coal, composed primarily of carbon, is the largest energy source for the generation of electricity and one of the largest anthropogenic sources for carbon dioxide emission. Thus the clean and efficient utilization of carbonaceous fuel is a vital challenge for humankind. SOFC (Solid oxide fuel cells) electrochemically convert the chemical energy of carbon (C) into electricity and heat without involving combustions and mechanical motions. In this way the energy conversion is highly efficient with limited emission and easy capture of CO2 [1,2]. In the DC-SOFC (direct carbon solid oxide fuel cell), C is oxidized in the anode compartment by oxygen ions transported from the cathode through the electrolyte [3]. A key issue of this technology is the lack of contact between the solid carbonaceous fuel and the oxygen ion [4,5]. Some approaches have been reported to increase
* Corresponding author. Center for Fuel Cell Innovation, State Key Laboratory of Coal Combustion, School of Materials Science and Engineering, Huazhong University of Science and Technology, Wuhan, Hubei 430074, China. Tel.: þ86 27 87557694; fax: þ86 27 87558142. E-mail address: [email protected] (L. Jian). http://dx.doi.org/10.1016/j.energy.2015.10.033 0360-5442/© 2015 Elsevier Ltd. All rights reserved.
the contact area for the anode reaction, such as using molten carbonates (K2CO3eLi2CO3) in the anode to generate 3-dimensional contact between solid carbonaceous fuel and oxygen ions [6,7] and gasifying the solid carbonaceous fuel to produce CO via the Boudouard Reaction [8,9]. However, the carbonate ions in the molten salts are consumed gradually by side reactions with carbon [10,11]; and the catalyzed gasification reaction of C lowers the efficiency of energy conversion [12,13]. Recently, low melting point metals, such as tin (Sn) (505 K) [14,15], indium (In) (430 K), bismuth (Bi) (544 K), lead (Pb) (601 K) [16] and antimony (Sb) (903 K) [17], were reported as the anode of DC-SOFCs. At the operating temperature, the molten metal forms an intimate contact with the electrolyte. Oxygen ions transported through the electrolyte electrochemically oxidize the metal to form metal oxide, which is then chemically reduced by the solid carbonaceous fuel added into the anode. Among these low melting point metals, Sb is often selected as the anode material, since its oxide Sb2O3 also has a low melting point of 928 K. In this case, the anode reactions are
2Sb þ 3O2 /Sb2 O3 þ 6e and
(1)
N.-Q. Duan et al. / Energy 95 (2016) 274e278
Sb2 O3 þ 3=2C/2Sb þ 3=2CO2
275
(2)
The low density and insulating liquid Sb2O3 formed during cell operation migrates away from the anode/electrolyte interface, which ensures a good contact between unreacted molten Sb and electrolyte for further reaction without increasing interfacial resistance. This kind of novel liquid anodes can be used with various types of carbonaceous fuel with easy capture of emitted CO2 and are expected to be more sulfur resistant than Ni-based anodes, as the addition of Sb and Sn in a Ni-based anode can enhance its resistance to sulfur poisoning [18]. So far, the feasibility of using molten Sb anode in DC-SOFCs was only demonstrated by planar button-size cells supported by ScSZ (Sc2O3 stabilized ZrO2) [17], YSZ (Y2O3 stabilized ZrO2) [19,20] or SDC (Sm2O3 doped CeO2) [21] electrolytes. High power density of 350 mW cm2 was achieved at 700 C by Jayakumar et al. [17] using a thin (100 mm) ScSZ supported cell; however, ScSZ was found chemically unstable and dissolved in molten Sb/Sb2O3 [19]. Compared with ScSZ and SDC electrolytes, YSZ is more reliable due to its high stability in molten Sb/Sb2O3 [19] and reduced atmosphere and high mechanical strength [22]. With molten Sb anode, the planar cell design may not be suitable for stacking and scale-up due to the difficulties in cell sealing, interconnecting and refueling; and its possibility of practical applications is limited. In contrast, the tubular cell design possesses advantages in the aspects of cell sealing, cell-to-cell connection, thermal cycling and rapid start-up [23,24], and is promising for refueling for continuous power generation. In the present study, tubular cells, with YSZ electrolyte, molten Sb anode and La0.6Sr0.4Co0.2Fe0.8O3-10 mol.% Gd2O3 doped CeO2 (LSCF-10GDC) composite cathode, were employed for the first time to demonstrate electrical power generation at intermediate temperatures from 700 to 800 C by using solid carbon fuel. During the test, fuel refilling was also conducted to prove the capability of continuous power generation. 2. Experimental The YSZ substrate tube of the tubular cell was prepared by a slurry-casting method [25] using a ball-milled slurry containing 8% mol Y2O3 stabilized ZrO2 powder (Tosoh) with xylene/ethanol (Sinopharm) as the solvent and polyvinyl butyral (Sinopharm) as the binder. The homogeneous slurry was poured into a tubular plastic mold and degassed centrifugally at a speed of 2000 rpm for 2 min. The viscosity of the slurry was carefully adjusted to between 14,000 and 16,000 mPa s, allowing the slurry to perfectly wet the inner wall of the plastic mold. The slurry hanging on the wall was gradually dried, while the mold was rotated vertically on a rotating plate at a constant angular speed, and detached from the mold due to shrinkage. The green body of the YSZ tube obtained was debound in air at a slow heating rate in a box furnace and then sintered at 1500 C for 5 h. To prevent the reaction between LSCF and YSZ, porous baffle layer of 10GDC and LSCF-10GDC composite cathode (1:1 weight ratio) were built in sequence on the outer surface of the sintered YSZ substrate tube by slurry dipping and sintering in air for 2 h at 1250 C and 1000 C respectively. Both the LSCF and 10GDC powders used in the slurries were synthesized by a wet chemical method, using metal nitrates (Sinopharm) as the precursors and ethylene diamine tetraacetic acid (Sinopharm) as the chelating agent. The dried gels were calcined in air at 850 C for 2 h. The electrochemical evaluation of the tubular cells was carried out using an in-house developed setup shown in Fig. 1. The tubular cell was loaded with 5 g of Sb powder (Sinopharm) and 1 g of activated carbon powder (Sinopharm) as the anode and fuel, respectively, and lap-sealed to an alumina tube using a ceramic
Fig. 1. Schematic drawing of the in-house developed setup for tubular DC-SOFC evaluation.
sealant (Ceramabond 668, Aremco Products). Rhenium (Re) wires of 0.5 mm in diameter were used as the anode current collector; and platinum (Pt) paste and twined silver (Ag) wire were used as the cathode current collector. Pure N2 was flowed into the tube to prevent the oxidation of Re, Sb and C at testing temperatures from 700 to 800 C. A Solartron 1260 frequency response analyzer (Solartron Analytical) was employed to measure the impedance of the cell in a frequency range between 100 kHz and 0.1 Hz with signal amplitude of 10 mV at open circuit. The currentevoltageepower (IeVeP) and voltageetime (Vet) curves were obtained using a Solartron 1287 electrochemical interface (Solartron Analytical) and a DC power source (IT-6720, iTech). The cell microstructure was examined by means of scanning electron microscopy (SEM, Sirion 200 and Quanta 200, FEI). 3. Results and discussion 3.1. Cell structure and initial electrochemical performance Fig. 2 shows the YSZ-supported tubular cell and its crosssectional microstructure. The substrate was 10 mm in outside diameter and 60 mm in length; and the thickness of the YSZ substrate, 10GDC baffle and LSCF-10GDC cathode was 150, 2.5 and 30 mm, respectively. The cathode was 6 mm in height with an active area of 2 cm2. The level of the molten Sb was slightly higher than that of the cathode to maximize the reaction area. Four cells were evaluated with molten Sb anode and activated carbon fuel, and similar performance was obtained within a fluctuation of 5%. This slight inconsistence was possibly caused by the unrealized condition variations in the preparation and testing of each cell. Fig. 3 shows the representative results of electrochemical impedance spectra (EIS) at open circuit voltage and IeVeP curves. The ohmic resistance was 1.01, 0.54 and 0.37 U cm2 at 700, 750 and 800 C, respectively, which mainly arises from the thick (150 mm) YSZ electrolyte [26], while the corresponding polarization resistance was 0.22, 0.11 and 0.06 U cm2, which includes the
276
N.-Q. Duan et al. / Energy 95 (2016) 274e278
contributions from both the anode and cathode (Fig. 3a). Compared with that of a similar cathode at 700 and 750 C [27,28], the polarization loss associated with the molten Sb anode is expected to be much less than the half of the total polarization resistance and even smaller than that of Ni-based anodes [29,30]. The open circuit voltage (OCV) was 0.741, 0.712 and 0.690 V at 700, 750, and 800 C, respectively (Fig. 3b), which is close to those reported previously [16,21] and the theoretical Nernst potential for the equilibrium between Sb and Sb2O3, suggesting that the cell was well prepared and sealed. The corresponding maximum power density was 114, 196, and 304 mW cm2 (Fig. 3b), which is slightly lower than that obtained from the planar cells [17,21] due to the use of thick YSZ electrolyte as the cell support. If the ohmic loss contributed by the thick (150 mm) YSZ electrolyte were neglected, the maximum power density would have been 624, 1150 and 1980 mW cm2 at 700, 750, and 800 C, respectively. This assumptive result indicates that the cell performance can be greatly improved simply by using a thinner YSZ or a more conductive electrolyte. 3.2. Cell performance at constant current
Fig. 2. Appearance (a) and cross-sectional microstructure (b) of the fabricated YSZsupported tubular cell with LSCF-10GDC cathode.
Fig. 3. Electrochemical impedance spectra (a) and IeV/IeP curves (b) of the tubular cell with molten Sb anode at 700, 750 and 800 C.
To understand the anode reactions and cell performance at constant current, three cells were operated at 0.8 A (0.4 A cm2) and 800 C with various amounts of carbon loaded into the Sb anode, as shown in Fig. 4. The voltage fluctuations demonstrated in the curves was caused by the real-time change of anode interface resistance, which was mainly due to the insulated Sb2O3 formation on and subsequent migration away from the molten Sb/YSZ electrolyte interface. For the cell with only 5 g Sb loaded (Cell A), the anode reaction proceeded according to the Reaction (1) and the cell voltage decreased to zero within around 2 h, corresponding to the oxidation of 53% of the added Sb to Sb2O3. The fast decrease of cell voltage with time was caused by the decrease in active anode area that was gradually covered by the resistant Sb2O3 floating on the top of molten Sb, as confirmed by SEM observation and EDS compositional analysis (Fig. 5). In contrast, with 5 g Sb and 1 g carbon loaded (Cell B), the cell performed rather stably for more than 6 h beyond which the voltage started to decrease rapidly. This result is similar to what reported by Jayakumar A et al. [17]. In this case, both Reaction (1) and (2) are expected to take place in the anode simultaneously; the electrochemically formed Sb2O3 was reduced by the carbon fuel. The longer than 6 h performance of Cell B suggests that Reaction (2) proceeded at a faster rate than that of Reaction (1), and the rapid decrease in cell voltage beyond 6 h indicates that 1 g carbon added was completely consumed by forming CO2 and the cell performance returned to the same level as that without adding 1 g of carbon. Refueling the anode with additional 1 g carbon at the operating temperature (Cell C), the cell perform resumed after an
Fig. 4. Time dependence of cell voltage at 0.8 A and 800 C with various amounts of carbon loaded in molten Sb anode.
N.-Q. Duan et al. / Energy 95 (2016) 274e278
277
Fig. 5. Vertical cross-sectional microstructure (a) (SEM in back scattering mode) and corresponding compositional analysis (b) of tested Cell A.
interruption of 0.5 h, during which the electrochemically formed Sb2O3 was fully reduced to Sb by the refilled carbon. The small running time difference of Cell-B and the first fueled Cell-C might be caused by the fuel weight error in the loading process. The performance of Cell C demonstrates that refueling a tubular cell is feasible and power generation can be maintained uninterruptedly as long as carbon fuel is supplied continuously. Fig. 6 compares the open circuit EIS and IeVeP curves of Cell C before the first and the second discharge. Such highly repeatable electrochemical performance suggests that the state of cell components was not altered by refueling. 4. Conclusions YSZ electrolyte-supported tubular SOFC cells with flexibility of stacking, scale-up and refueling were fabricated and evaluated with molten Sb anode and LSCF-10GDC cathode for the first time at various temperatures from 700 to 800 C. The polarization resistance of the molten Sb anode was significantly low, similar to that of the Ni-based anodes. The maximum power density was in the
Fig. 6. EIS spectra (a) and IeV/IeP curves (b) of Cell C before the first and second discharge at 800 C.
range between 114 and 304 mW cm2. With 1 g activated carbon added in the molten Sb anode, the cell performed for more than 6 h at 200 mW cm2 at 800 C without noticeable degradation. The cell performance recovered to the previous level for another 6 h with the addition of extra 1 g carbon, demonstrating the feasibility of refueling a tubular DC-SOFC with molten Sb anode.
Acknowledgment This research was financially supported by the National “863” Project of China (2011AA050702), National Natural Science Foundation of China (U1134001) & Financial Support from Guangdong Province & Shenzhen (JCYJ20140419131733975). The SEM characterizations were assisted by the Analytical and Testing Center of Huazhong University of Science and Technology.
References [1] Cao D, Sun Y, Wang G. Direct carbon fuel cell: fundamentals and recent developments. J Power Sources 2007;167(2):250e7. [2] Giddey S, Badwal SPS, Kulkarni A, Munnings C. A comprehensive review of direct carbon fuel cell technology. Prog Energy Combust Sci 2012;38(3): 360e99. [3] Gür TM. Critical review of carbon conversion in “carbon fuel cells”. Chem Rev 2013;113(8):6179e206. [4] Nabae Y, Pointon KD, Irvine JTS. Electrochemical oxidation of solid carbon in hybrid DCFC with solid oxide and molten carbonate binary electrolyte. Energy Environ Sci 2008;1(1):148e55. [5] Gür TM. Utilization modes for solid carbon conversion in fuel cells. ECS Trans 2009;25(2):1099e108. [6] Hao W, He X, Mi Y. Achieving high performance in intermediate temperature direct carbon fuel cells with renewable carbon as a fuel source. Appl Energy 2014;135(0):174e81. [7] Xu X, Zhou W, Liang F, Zhu Z. A comparative study of different carbon fuels in an electrolyte-supported hybrid direct carbon fuel cell. Appl Energy 2013;108(0):402e9. MO, Shao Z. A carboneair battery for high [8] Yang B, Ran R, Zhong Y, Su C, Tade power generation. Angew Chem Int Ed 2015;54(12):3722e5. [9] Jiao Y, Tian W, Chen H, Shi H, Yang B, Li C, et al. In situ catalyzed Boudouard reaction of coal char for solid oxide-based carbon fuel cells with improved performance. Appl Energy 2015;141(0):200e8. [10] Jiang C, Ma J, Bonaccorso AD, Irvine JTS. Demonstration of high power, direct conversion of waste-derived carbon in a hybrid direct carbon fuel cell. Energy Environ Sci 2012;5(5):6973e80. [11] Pointon K, Lakeman B, Irvine J, Bradley J, Jain S. The development of a carboneair semi fuel cell. J Power Sources 2006;162(2):750e6. [12] Yu X, Shi Y, Wang H, Cai N, Li C, Ghoniem AF. Using potassium catalytic gasification to improve the performance of solid oxide direct carbon fuel cells: experimental characterization and elementary reaction modeling. J Power Sources 2014;252:130e7. [13] Antunes R, Skrzypkiewicz M. Chronoamperometric investigations of electrooxidation of lignite in direct carbon bed solid oxide fuel cell. Int J Hydrogen Energy 2015;40(12):4357e69. [14] Tao T, Bateman L, Bentley J, Slaney M. Liquid tin anode solid oxide fuel cell for direct carbonaceous fuel conversion. ECS Trans 2007;5(1):463e72.
278
N.-Q. Duan et al. / Energy 95 (2016) 274e278
[15] McPhee WA, Bateman L, Koslowske M, Slaney M, Uzep Z, Bentley J, et al. Direct JP-8 conversion using a liquid tin anode solid oxide fuel cell (LTA-SOFC) for military applications. J Fuel Cell Sci Technol 2011;8(4):041007. [16] Jayakumar A, Vohs JM, Gorte RJ. Molten-metal electrodes for solid oxide fuel cells. Ind Eng Chem Res 2010;49(21):10237e41. [17] Jayakumar A, Kungas R, Roy S, Javadekar A, Buttrey DJ, Vohs JM, et al. A direct carbon fuel cell with a molten antimony anode. Energy Environ Sci 2011;4(10):4133e7. [18] Marina OA, Coyle CA, Engelhard MH, Pederson LR. Mitigation of sulfur poisoning of Ni/Zirconia SOFC anodes by antimony and tin. J Electrochem Soc 2011;158(4):B424e9. [19] Jayakumar A, Javadekar A, Gissinger J, Vohs JM, Huber GW, Gorte RJ. The stability of direct carbon fuel cells with molten Sb and SbeBi alloy anodes. AIChE J 2012:3342e8. [20] Xu X, Zhou W, Zhu Z. Samaria-doped ceria electrolyte supported direct carbon fuel cell with molten antimony as the anode. Ind Eng Chem Res 2013;52(50): 17927e33. [21] Wang H, Cao T, Shi Y, Cai N, Yuan W. Liquid antimony anode direct carbon fuel cell fueled with mass-produced de-ash coal. Energy 2014;75(0):555e9. [22] Minh NQ, Takahashi T. Science and technology of ceramic fuel cells. Elsevier; 1995.
[23] Singh P, Minh NQ. Solid oxide fuel cells: technology status. Int J Appl Ceram Technol 2004;1(1):5e15. [24] George RA. Status of tubular SOFC field unit demonstrations. J Power Sources 2000;86(1e2):134e9. [25] Duan N-Q, Yan D, Chi B, Pu J, Li J. High performance anode-supported tubular solid oxide fuel cells fabricated by a novel slurry-casting method. Sci Rep 2015;5. [26] Sasaki K, Maier J. Re-analysis of defect equilibria and transport parameters in Y2O3-stabilized ZrO2 using EPR and optical relaxation. Solid State Ion 2000;134(3e4):303e21. [27] Chen J, Liang F, Chi B, Pu J, Jiang SP, Li J. Palladium and ceria infiltrated La0.8Sr0.2Co0.5Fe0.5O3d cathodes of solid oxide fuel cells. J Power Sources 2009;194(1):275e80. [28] Liu Y, Wang F, Chi B, Pu J, Li J, Jiang SP. A stability study of impregnated LSCFeGDC composite cathodes of solid oxide fuel cells. J Alloys Compd 2013;578(0):37e43. [29] Koide H, Someya Y, Yoshida T, Maruyama T. Properties of Ni/YSZ cermet as anode for SOFC. Solid State Ion 2000;132(3e4):253e60. [30] Guo W, Liu J. The effect of nickel oxide microstructure on the performance of NieYSZ anode-supported SOFCs. Solid State Ion 2008;179(27e32):1516e20.
Contents lists available at ScienceDirect
Energy journal homepage: www.elsevier.com/locate/energy
Tubular direct carbon solid oxide fuel cells with molten antimony anode and refueling feasibility Nan-Qi Duan a, Yong Cao a, Bin Hua b, Bo Chi a, Jian Pu a, Jingli Luo b, Li Jian a, c, * a Center for Fuel Cell Innovation, State Key Laboratory of Coal Combustion, School of Materials Science and Engineering, Huazhong University of Science and Technology, Wuhan, Hubei 430074, China b Department of Chemical and Materials Engineering, University of Alberta, Edmonton, Alberta T6G 2G6, Canada c Research Institute of Huazhong University of Science and Technology in Shenzhen, Shenzhen, Guangdong 518000, China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 5 June 2015 Received in revised form 9 October 2015 Accepted 11 October 2015 Available online 31 December 2015
Tubular direct carbon SOFCs (solid oxide fuel cells) supported by YSZ (Y2O3 stabilized ZrO2) electrolyte are fabricated by slurry-casting, slurry-dipping and sintering processes with La0.6Sr0.4Co0.2Fe0.8O310 mol.% Gd2O3 doped CeO2 (LSCF-10GDC) as the cathode. Their electrochemical performance is examined at temperatures from 700 to 800 C using molten antimony (Sb) anode and activated carbon fuel. The ohmic resistance of the cell is between 1.01 and 0.37 U cm2 mainly originated from the thick YSZ electrolyte (150 mm); the polarization resistance ranges from 0.22 to 0.06 U cm2. The maximum power density at 800 C is 304 mW cm2 and can be greatly increased by using a thinner and/or more conductive electrolyte. With 1 g activated carbon as the fuel, the cell performance is stable at 200 mW cm2 at 800 C for more than 6 h by chemical consumption (oxidization) of the carbon, which reduces the electrochemically formed Sb2O3 to Sb. The cell performance decreases as the fuel is used up and is recovered by refueling. © 2015 Elsevier Ltd. All rights reserved.
Keywords: Carbonaceous fuel Tubular solid oxide fuel cell Molten antimony anode Electrochemical performance Refueling
1. Introduction Coal, composed primarily of carbon, is the largest energy source for the generation of electricity and one of the largest anthropogenic sources for carbon dioxide emission. Thus the clean and efficient utilization of carbonaceous fuel is a vital challenge for humankind. SOFC (Solid oxide fuel cells) electrochemically convert the chemical energy of carbon (C) into electricity and heat without involving combustions and mechanical motions. In this way the energy conversion is highly efficient with limited emission and easy capture of CO2 [1,2]. In the DC-SOFC (direct carbon solid oxide fuel cell), C is oxidized in the anode compartment by oxygen ions transported from the cathode through the electrolyte [3]. A key issue of this technology is the lack of contact between the solid carbonaceous fuel and the oxygen ion [4,5]. Some approaches have been reported to increase
* Corresponding author. Center for Fuel Cell Innovation, State Key Laboratory of Coal Combustion, School of Materials Science and Engineering, Huazhong University of Science and Technology, Wuhan, Hubei 430074, China. Tel.: þ86 27 87557694; fax: þ86 27 87558142. E-mail address: [email protected] (L. Jian). http://dx.doi.org/10.1016/j.energy.2015.10.033 0360-5442/© 2015 Elsevier Ltd. All rights reserved.
the contact area for the anode reaction, such as using molten carbonates (K2CO3eLi2CO3) in the anode to generate 3-dimensional contact between solid carbonaceous fuel and oxygen ions [6,7] and gasifying the solid carbonaceous fuel to produce CO via the Boudouard Reaction [8,9]. However, the carbonate ions in the molten salts are consumed gradually by side reactions with carbon [10,11]; and the catalyzed gasification reaction of C lowers the efficiency of energy conversion [12,13]. Recently, low melting point metals, such as tin (Sn) (505 K) [14,15], indium (In) (430 K), bismuth (Bi) (544 K), lead (Pb) (601 K) [16] and antimony (Sb) (903 K) [17], were reported as the anode of DC-SOFCs. At the operating temperature, the molten metal forms an intimate contact with the electrolyte. Oxygen ions transported through the electrolyte electrochemically oxidize the metal to form metal oxide, which is then chemically reduced by the solid carbonaceous fuel added into the anode. Among these low melting point metals, Sb is often selected as the anode material, since its oxide Sb2O3 also has a low melting point of 928 K. In this case, the anode reactions are
2Sb þ 3O2 /Sb2 O3 þ 6e and
(1)
N.-Q. Duan et al. / Energy 95 (2016) 274e278
Sb2 O3 þ 3=2C/2Sb þ 3=2CO2
275
(2)
The low density and insulating liquid Sb2O3 formed during cell operation migrates away from the anode/electrolyte interface, which ensures a good contact between unreacted molten Sb and electrolyte for further reaction without increasing interfacial resistance. This kind of novel liquid anodes can be used with various types of carbonaceous fuel with easy capture of emitted CO2 and are expected to be more sulfur resistant than Ni-based anodes, as the addition of Sb and Sn in a Ni-based anode can enhance its resistance to sulfur poisoning [18]. So far, the feasibility of using molten Sb anode in DC-SOFCs was only demonstrated by planar button-size cells supported by ScSZ (Sc2O3 stabilized ZrO2) [17], YSZ (Y2O3 stabilized ZrO2) [19,20] or SDC (Sm2O3 doped CeO2) [21] electrolytes. High power density of 350 mW cm2 was achieved at 700 C by Jayakumar et al. [17] using a thin (100 mm) ScSZ supported cell; however, ScSZ was found chemically unstable and dissolved in molten Sb/Sb2O3 [19]. Compared with ScSZ and SDC electrolytes, YSZ is more reliable due to its high stability in molten Sb/Sb2O3 [19] and reduced atmosphere and high mechanical strength [22]. With molten Sb anode, the planar cell design may not be suitable for stacking and scale-up due to the difficulties in cell sealing, interconnecting and refueling; and its possibility of practical applications is limited. In contrast, the tubular cell design possesses advantages in the aspects of cell sealing, cell-to-cell connection, thermal cycling and rapid start-up [23,24], and is promising for refueling for continuous power generation. In the present study, tubular cells, with YSZ electrolyte, molten Sb anode and La0.6Sr0.4Co0.2Fe0.8O3-10 mol.% Gd2O3 doped CeO2 (LSCF-10GDC) composite cathode, were employed for the first time to demonstrate electrical power generation at intermediate temperatures from 700 to 800 C by using solid carbon fuel. During the test, fuel refilling was also conducted to prove the capability of continuous power generation. 2. Experimental The YSZ substrate tube of the tubular cell was prepared by a slurry-casting method [25] using a ball-milled slurry containing 8% mol Y2O3 stabilized ZrO2 powder (Tosoh) with xylene/ethanol (Sinopharm) as the solvent and polyvinyl butyral (Sinopharm) as the binder. The homogeneous slurry was poured into a tubular plastic mold and degassed centrifugally at a speed of 2000 rpm for 2 min. The viscosity of the slurry was carefully adjusted to between 14,000 and 16,000 mPa s, allowing the slurry to perfectly wet the inner wall of the plastic mold. The slurry hanging on the wall was gradually dried, while the mold was rotated vertically on a rotating plate at a constant angular speed, and detached from the mold due to shrinkage. The green body of the YSZ tube obtained was debound in air at a slow heating rate in a box furnace and then sintered at 1500 C for 5 h. To prevent the reaction between LSCF and YSZ, porous baffle layer of 10GDC and LSCF-10GDC composite cathode (1:1 weight ratio) were built in sequence on the outer surface of the sintered YSZ substrate tube by slurry dipping and sintering in air for 2 h at 1250 C and 1000 C respectively. Both the LSCF and 10GDC powders used in the slurries were synthesized by a wet chemical method, using metal nitrates (Sinopharm) as the precursors and ethylene diamine tetraacetic acid (Sinopharm) as the chelating agent. The dried gels were calcined in air at 850 C for 2 h. The electrochemical evaluation of the tubular cells was carried out using an in-house developed setup shown in Fig. 1. The tubular cell was loaded with 5 g of Sb powder (Sinopharm) and 1 g of activated carbon powder (Sinopharm) as the anode and fuel, respectively, and lap-sealed to an alumina tube using a ceramic
Fig. 1. Schematic drawing of the in-house developed setup for tubular DC-SOFC evaluation.
sealant (Ceramabond 668, Aremco Products). Rhenium (Re) wires of 0.5 mm in diameter were used as the anode current collector; and platinum (Pt) paste and twined silver (Ag) wire were used as the cathode current collector. Pure N2 was flowed into the tube to prevent the oxidation of Re, Sb and C at testing temperatures from 700 to 800 C. A Solartron 1260 frequency response analyzer (Solartron Analytical) was employed to measure the impedance of the cell in a frequency range between 100 kHz and 0.1 Hz with signal amplitude of 10 mV at open circuit. The currentevoltageepower (IeVeP) and voltageetime (Vet) curves were obtained using a Solartron 1287 electrochemical interface (Solartron Analytical) and a DC power source (IT-6720, iTech). The cell microstructure was examined by means of scanning electron microscopy (SEM, Sirion 200 and Quanta 200, FEI). 3. Results and discussion 3.1. Cell structure and initial electrochemical performance Fig. 2 shows the YSZ-supported tubular cell and its crosssectional microstructure. The substrate was 10 mm in outside diameter and 60 mm in length; and the thickness of the YSZ substrate, 10GDC baffle and LSCF-10GDC cathode was 150, 2.5 and 30 mm, respectively. The cathode was 6 mm in height with an active area of 2 cm2. The level of the molten Sb was slightly higher than that of the cathode to maximize the reaction area. Four cells were evaluated with molten Sb anode and activated carbon fuel, and similar performance was obtained within a fluctuation of 5%. This slight inconsistence was possibly caused by the unrealized condition variations in the preparation and testing of each cell. Fig. 3 shows the representative results of electrochemical impedance spectra (EIS) at open circuit voltage and IeVeP curves. The ohmic resistance was 1.01, 0.54 and 0.37 U cm2 at 700, 750 and 800 C, respectively, which mainly arises from the thick (150 mm) YSZ electrolyte [26], while the corresponding polarization resistance was 0.22, 0.11 and 0.06 U cm2, which includes the
276
N.-Q. Duan et al. / Energy 95 (2016) 274e278
contributions from both the anode and cathode (Fig. 3a). Compared with that of a similar cathode at 700 and 750 C [27,28], the polarization loss associated with the molten Sb anode is expected to be much less than the half of the total polarization resistance and even smaller than that of Ni-based anodes [29,30]. The open circuit voltage (OCV) was 0.741, 0.712 and 0.690 V at 700, 750, and 800 C, respectively (Fig. 3b), which is close to those reported previously [16,21] and the theoretical Nernst potential for the equilibrium between Sb and Sb2O3, suggesting that the cell was well prepared and sealed. The corresponding maximum power density was 114, 196, and 304 mW cm2 (Fig. 3b), which is slightly lower than that obtained from the planar cells [17,21] due to the use of thick YSZ electrolyte as the cell support. If the ohmic loss contributed by the thick (150 mm) YSZ electrolyte were neglected, the maximum power density would have been 624, 1150 and 1980 mW cm2 at 700, 750, and 800 C, respectively. This assumptive result indicates that the cell performance can be greatly improved simply by using a thinner YSZ or a more conductive electrolyte. 3.2. Cell performance at constant current
Fig. 2. Appearance (a) and cross-sectional microstructure (b) of the fabricated YSZsupported tubular cell with LSCF-10GDC cathode.
Fig. 3. Electrochemical impedance spectra (a) and IeV/IeP curves (b) of the tubular cell with molten Sb anode at 700, 750 and 800 C.
To understand the anode reactions and cell performance at constant current, three cells were operated at 0.8 A (0.4 A cm2) and 800 C with various amounts of carbon loaded into the Sb anode, as shown in Fig. 4. The voltage fluctuations demonstrated in the curves was caused by the real-time change of anode interface resistance, which was mainly due to the insulated Sb2O3 formation on and subsequent migration away from the molten Sb/YSZ electrolyte interface. For the cell with only 5 g Sb loaded (Cell A), the anode reaction proceeded according to the Reaction (1) and the cell voltage decreased to zero within around 2 h, corresponding to the oxidation of 53% of the added Sb to Sb2O3. The fast decrease of cell voltage with time was caused by the decrease in active anode area that was gradually covered by the resistant Sb2O3 floating on the top of molten Sb, as confirmed by SEM observation and EDS compositional analysis (Fig. 5). In contrast, with 5 g Sb and 1 g carbon loaded (Cell B), the cell performed rather stably for more than 6 h beyond which the voltage started to decrease rapidly. This result is similar to what reported by Jayakumar A et al. [17]. In this case, both Reaction (1) and (2) are expected to take place in the anode simultaneously; the electrochemically formed Sb2O3 was reduced by the carbon fuel. The longer than 6 h performance of Cell B suggests that Reaction (2) proceeded at a faster rate than that of Reaction (1), and the rapid decrease in cell voltage beyond 6 h indicates that 1 g carbon added was completely consumed by forming CO2 and the cell performance returned to the same level as that without adding 1 g of carbon. Refueling the anode with additional 1 g carbon at the operating temperature (Cell C), the cell perform resumed after an
Fig. 4. Time dependence of cell voltage at 0.8 A and 800 C with various amounts of carbon loaded in molten Sb anode.
N.-Q. Duan et al. / Energy 95 (2016) 274e278
277
Fig. 5. Vertical cross-sectional microstructure (a) (SEM in back scattering mode) and corresponding compositional analysis (b) of tested Cell A.
interruption of 0.5 h, during which the electrochemically formed Sb2O3 was fully reduced to Sb by the refilled carbon. The small running time difference of Cell-B and the first fueled Cell-C might be caused by the fuel weight error in the loading process. The performance of Cell C demonstrates that refueling a tubular cell is feasible and power generation can be maintained uninterruptedly as long as carbon fuel is supplied continuously. Fig. 6 compares the open circuit EIS and IeVeP curves of Cell C before the first and the second discharge. Such highly repeatable electrochemical performance suggests that the state of cell components was not altered by refueling. 4. Conclusions YSZ electrolyte-supported tubular SOFC cells with flexibility of stacking, scale-up and refueling were fabricated and evaluated with molten Sb anode and LSCF-10GDC cathode for the first time at various temperatures from 700 to 800 C. The polarization resistance of the molten Sb anode was significantly low, similar to that of the Ni-based anodes. The maximum power density was in the
Fig. 6. EIS spectra (a) and IeV/IeP curves (b) of Cell C before the first and second discharge at 800 C.
range between 114 and 304 mW cm2. With 1 g activated carbon added in the molten Sb anode, the cell performed for more than 6 h at 200 mW cm2 at 800 C without noticeable degradation. The cell performance recovered to the previous level for another 6 h with the addition of extra 1 g carbon, demonstrating the feasibility of refueling a tubular DC-SOFC with molten Sb anode.
Acknowledgment This research was financially supported by the National “863” Project of China (2011AA050702), National Natural Science Foundation of China (U1134001) & Financial Support from Guangdong Province & Shenzhen (JCYJ20140419131733975). The SEM characterizations were assisted by the Analytical and Testing Center of Huazhong University of Science and Technology.
References [1] Cao D, Sun Y, Wang G. Direct carbon fuel cell: fundamentals and recent developments. J Power Sources 2007;167(2):250e7. [2] Giddey S, Badwal SPS, Kulkarni A, Munnings C. A comprehensive review of direct carbon fuel cell technology. Prog Energy Combust Sci 2012;38(3): 360e99. [3] Gür TM. Critical review of carbon conversion in “carbon fuel cells”. Chem Rev 2013;113(8):6179e206. [4] Nabae Y, Pointon KD, Irvine JTS. Electrochemical oxidation of solid carbon in hybrid DCFC with solid oxide and molten carbonate binary electrolyte. Energy Environ Sci 2008;1(1):148e55. [5] Gür TM. Utilization modes for solid carbon conversion in fuel cells. ECS Trans 2009;25(2):1099e108. [6] Hao W, He X, Mi Y. Achieving high performance in intermediate temperature direct carbon fuel cells with renewable carbon as a fuel source. Appl Energy 2014;135(0):174e81. [7] Xu X, Zhou W, Liang F, Zhu Z. A comparative study of different carbon fuels in an electrolyte-supported hybrid direct carbon fuel cell. Appl Energy 2013;108(0):402e9. MO, Shao Z. A carboneair battery for high [8] Yang B, Ran R, Zhong Y, Su C, Tade power generation. Angew Chem Int Ed 2015;54(12):3722e5. [9] Jiao Y, Tian W, Chen H, Shi H, Yang B, Li C, et al. In situ catalyzed Boudouard reaction of coal char for solid oxide-based carbon fuel cells with improved performance. Appl Energy 2015;141(0):200e8. [10] Jiang C, Ma J, Bonaccorso AD, Irvine JTS. Demonstration of high power, direct conversion of waste-derived carbon in a hybrid direct carbon fuel cell. Energy Environ Sci 2012;5(5):6973e80. [11] Pointon K, Lakeman B, Irvine J, Bradley J, Jain S. The development of a carboneair semi fuel cell. J Power Sources 2006;162(2):750e6. [12] Yu X, Shi Y, Wang H, Cai N, Li C, Ghoniem AF. Using potassium catalytic gasification to improve the performance of solid oxide direct carbon fuel cells: experimental characterization and elementary reaction modeling. J Power Sources 2014;252:130e7. [13] Antunes R, Skrzypkiewicz M. Chronoamperometric investigations of electrooxidation of lignite in direct carbon bed solid oxide fuel cell. Int J Hydrogen Energy 2015;40(12):4357e69. [14] Tao T, Bateman L, Bentley J, Slaney M. Liquid tin anode solid oxide fuel cell for direct carbonaceous fuel conversion. ECS Trans 2007;5(1):463e72.
278
N.-Q. Duan et al. / Energy 95 (2016) 274e278
[15] McPhee WA, Bateman L, Koslowske M, Slaney M, Uzep Z, Bentley J, et al. Direct JP-8 conversion using a liquid tin anode solid oxide fuel cell (LTA-SOFC) for military applications. J Fuel Cell Sci Technol 2011;8(4):041007. [16] Jayakumar A, Vohs JM, Gorte RJ. Molten-metal electrodes for solid oxide fuel cells. Ind Eng Chem Res 2010;49(21):10237e41. [17] Jayakumar A, Kungas R, Roy S, Javadekar A, Buttrey DJ, Vohs JM, et al. A direct carbon fuel cell with a molten antimony anode. Energy Environ Sci 2011;4(10):4133e7. [18] Marina OA, Coyle CA, Engelhard MH, Pederson LR. Mitigation of sulfur poisoning of Ni/Zirconia SOFC anodes by antimony and tin. J Electrochem Soc 2011;158(4):B424e9. [19] Jayakumar A, Javadekar A, Gissinger J, Vohs JM, Huber GW, Gorte RJ. The stability of direct carbon fuel cells with molten Sb and SbeBi alloy anodes. AIChE J 2012:3342e8. [20] Xu X, Zhou W, Zhu Z. Samaria-doped ceria electrolyte supported direct carbon fuel cell with molten antimony as the anode. Ind Eng Chem Res 2013;52(50): 17927e33. [21] Wang H, Cao T, Shi Y, Cai N, Yuan W. Liquid antimony anode direct carbon fuel cell fueled with mass-produced de-ash coal. Energy 2014;75(0):555e9. [22] Minh NQ, Takahashi T. Science and technology of ceramic fuel cells. Elsevier; 1995.
[23] Singh P, Minh NQ. Solid oxide fuel cells: technology status. Int J Appl Ceram Technol 2004;1(1):5e15. [24] George RA. Status of tubular SOFC field unit demonstrations. J Power Sources 2000;86(1e2):134e9. [25] Duan N-Q, Yan D, Chi B, Pu J, Li J. High performance anode-supported tubular solid oxide fuel cells fabricated by a novel slurry-casting method. Sci Rep 2015;5. [26] Sasaki K, Maier J. Re-analysis of defect equilibria and transport parameters in Y2O3-stabilized ZrO2 using EPR and optical relaxation. Solid State Ion 2000;134(3e4):303e21. [27] Chen J, Liang F, Chi B, Pu J, Jiang SP, Li J. Palladium and ceria infiltrated La0.8Sr0.2Co0.5Fe0.5O3d cathodes of solid oxide fuel cells. J Power Sources 2009;194(1):275e80. [28] Liu Y, Wang F, Chi B, Pu J, Li J, Jiang SP. A stability study of impregnated LSCFeGDC composite cathodes of solid oxide fuel cells. J Alloys Compd 2013;578(0):37e43. [29] Koide H, Someya Y, Yoshida T, Maruyama T. Properties of Ni/YSZ cermet as anode for SOFC. Solid State Ion 2000;132(3e4):253e60. [30] Guo W, Liu J. The effect of nickel oxide microstructure on the performance of NieYSZ anode-supported SOFCs. Solid State Ion 2008;179(27e32):1516e20.