- Email: [email protected]
Effectiveness of paper-structured catalyst for the operation of biodiesel-fueled solid oxide fuel cell
Journal of Power Sources 283 (2015) 320e327
Contents lists available at ScienceDirect
Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour
Effectiveness of paper-structured catalyst for the operation of biodiesel-fueled solid oxide fuel cell Tran Quang-Tuyen a, b, *, Taku Kaida a, Mio Sakamoto a, Kazunari Sasaki a, b, c, d, Yusuke Shiratori a, b, c, d a
Faculty of Engineering, Kyushu University, Motooka 744, Nishi-ku, Fukuoka 819-0395, Japan International Research Center for Hydrogen Energy, Kyushu University, Motooka 744, Nishi-ku, Fukuoka 819-0395, Japan International Institute for Carbon-Neutral Energy Research (WPI), Kyushu University, Motooka 744, Nishi-ku, Fukuoka 819-0395, Japan d Next-Generation Fuel Cell Research Center, Kyushu University, Motooka 744, Nishi-ku, Fukuoka 819-0395, Japan b c
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 model biodiesel fuel (BDF) was applied as an alternative fuel for SOFC. Novel paper-structured catalyst (PSC) containing Mg/Al-hydrotalcite was synthesized. PSC exhibited fuel conversion comparable to SOFC anode with less than 1/100 Ni weight. Long-term stability of SOFC fueled by BDF was achieved by the application of PSC.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 22 January 2015 Received in revised form 19 February 2015 Accepted 22 February 2015 Available online 26 February 2015
Mg/Al-hydrotalcite (HDT)-dispersed paper-structured catalyst (PSC) was prepared by a simple papermaking process. The PSC exhibited excellent catalytic activity for the steam reforming of model biodiesel fuel (BDF), pure oleic acid methyl ester (oleic-FAME, C19H36O2) which is a mono-unsaturated component of practical BDFs. The PSC exhibited fuel conversion comparable to a pelletized catalyst material, here, conventional Niezirconia cermet anode for solid oxide fuel cell (SOFC) with less than onehundredth Ni weight. Performance of electrolyte-supported cell connected with the PSC was evaluated in the feed of oleic-FAME, and stable operation was achieved. After 60 h test, coking was not observed in both SOFC anode and PSC. © 2015 Elsevier B.V. All rights reserved.
Keywords: Paper-structured catalyst Mg/Al-hydrotalcite Steam reforming Oleic-FAME Solid oxide fuel cell
1. Introduction
* Corresponding author. Faculty of Engineering, Kyushu University, Motooka 744, Nishi-ku, Fukuoka 819-0395, Japan. E-mail addresses: [email protected], [email protected] (T. Quang-Tuyen). http://dx.doi.org/10.1016/j.jpowsour.2015.02.116 0378-7753/© 2015 Elsevier B.V. All rights reserved.
High temperature solid oxide fuel cell (SOFC) can accept direct feed of hydrocarbon fuels. SOFC is expected to be the most efficient device for direct conversion of chemical energy into electricity [1,2]. Reforming of hydrocarbons can proceed in the porous cermet anode to produce H2 and CO, which electrochemically oxidized to
T. Quang-Tuyen et al. / Journal of Power Sources 283 (2015) 320e327 Table 1 Composition of raw papers (RPs) after drying. Type of RP
HDT/wt% Cf/wt% Zs/wt% Cationic Anionic Pulp/wt% polymer/wt% polymer/wt%
HDT-free 0 HDT-dispersed 14.7
86.1 73.3
8.6 7.3
0.5 0.5
0.5 0.5
4.3 3.7
Table 2 Ni contents in the HDT-dispersed PSCs (for two pieces of each type of PSC with 16 mm in diameter and 1.1 mm in thickness) and pelletized NieScSZ cermet anodes. Paper-structured catalysta
Ni loading/mg
NieScSZ cermet anodeb (Reference catalyst)
Ni loading/mg
PSC-1 PSC-2 PSC-3 PSC-4 PSC-5
0.6 2.4 2.8 3.2 5.6
Cermet-1 Cermet-2 Cermet-3 Cermet-4 Cermet-5
170 240 300 430 1200
a Weights of HDT, Cf and Zs in two pieces of each type of PSC are 18.9, 95 and 9.5 mg, respectively. b Weight ratio of Ni:ScSZ is 1:1.
generate electricity and heat [3,4]. This operation is called direct internal reforming (DIR) operation, which brings us several advantages such as downsizing and cost reduction of fuel cell system and the reduction of energy consumption for the stack cooling by the supply of excess air due to strong endothermicity of the reforming reaction [5,6]. Nickel is commonly-used anode material for SOFC due to its excellent electro-catalytic properties. However, coking on Ni is a main drawback of DIR operation of SOFC fueled by light hydrocarbons [7e9] as well as heavy hydrocarbons [10e12]. Carbon deposition on and inside of porous anode material is one of the most crucial issues of DIR-SOFC, because it significantly degrades fuel cell performance as the results of the coverage of electrochemical and reforming reaction sites. To solve this problem, many approaches
321
have been reported. The first approach is careful control of steam to carbon ratio (S/C). Excess amount of water is essential to promote reforming reaction, however high S/C ratio results in the reduction of electromotive force, furthermore deteriorates catalytic activity of the porous cermet anode due to oxidation of Ni [13,14]. The second approach is incorporation of coking-resistant catalyst materials such as copper [15,16], but low melting temperatures of copper and its oxides limit the use of them for SOFC anode [17,18]. Zhan et al. applied a thin RueCeO2 catalyst layer on the top of a conventional NieYSZ anode [19]. They obtained stable cell operation without coking by the direct feed of fuel mixture of 5% iso-octane and 9% air diluted by 86% CO2 to a SOFC single cell at 770 C under current density of 0.6 A cm2 for 50 h, although difficulties of fuel diffusion and current collection must be overcome. Recently, flexible paper structured-catalyst (PSC) based on inorganic fiber network was developed to solve the problems associated with the DIR operation of SOFC mentioned above. The PSC exhibited excellent catalytic performance for dry reforming of methane [20,21]. Long term stability of SOFC fueled by pure oleic acid methyl ester (oleic-FAME), which is a main component of practical biodiesel fuels (BDFs), was confirmed by the connection of SOFC with PSC [22]. However, for the practical application of PSCs to SOFC systems, further improvement of catalytic performance is necessary. In this study, Ni loaded PSC with improved catalytic performance in which layered double hydroxide particles containing Mg2þ and Al3þ called Mg/Al-hydrotalcite (HDT) were dispersed in the paper-making process was tested for the stream reforming of oleic-FAME. Then, effectiveness of the HDT-dispersed PSC for the operation of biodiesel-fueled SOFC was demonstrated. 2. Experimental 2.1. Model fuel Biodiesel fuel (BDF) is produced from biomass resources such as vegetable oil, waste cooking oil and animal fat by alkali catalyzed
Table 3 Component materials of anode- and electrolyte-supported cells prepared in this study. Anode
Anodesupporteda Electrolytesupportedb a b
Electrolyte
Current collector
Substrate
Functional layer
NiOeScSZ (30 mm) NiOeScSZ (30 mm)
NiOeScSZ (800 mm)
e NiOeScSZ (60 mm)
ScSZ (22 mm) ScSZ (220 mm)
Cathode Functional layer
Current collector
LSM-ScSZ (30 mm) LSM-ScSZ (30 mm)
LSM (30 mm) LSM (30 mm)
Anode contains 550 mg of Ni metal. Anode contains 6.8 mg of Ni metal.
Fig. 1. Experimental setup for testing SOFC connected with PSC in the feed of oleic-FAME.
322
T. Quang-Tuyen et al. / Journal of Power Sources 283 (2015) 320e327
Fig. 2. FESEM images of (a) porous NieScSZ cermet anode and (b) PSC-4.
trans-esterification reaction. Now, BDF is used as an alternative fuel to petro-diesel in transportation sector. Oleic acid methyl ester (Oleic-FAME, C19H36O2), which is mono-unsaturated fuel and one of the main components of popular BDFs such as Palm-, Jatropha-, and Soybean-BDFs [23], was chosen as a model BDF in this study. 2.2. Preparation of paper-structured catalyst (PSC) PSC was prepared by an established papermaking technique through a dual polyelectrolyte retention system. Detailed procedure has been reported elsewhere [20]. A water suspension of the mixture of ceramic fiber denoted as Cf (SiO2:52%, Al2O3:48%; IBIDEN Ltd., Japan) with 500 mm in length and 2 mm in diameter and HDT powder ([Mg6Al2(OH)16CO3]4H2O; Wako Pure Chemical
Industries Ltd., Japan) was prepared. Then, cationic polymer, PDADMAC (Sigma Aldrich LCC, USA), zirconia sol (ZrO2 without any stabilizer) denoted as Zs (Daiichi Kigenso Kagaku Kogyo Co., Ltd., Japan) and anionic polymer, A-PAM FA-405S (Fujikasui Engineering Co., Ltd., Japan), were added to the suspension in that order and mixed for 3 min. The mixture was poured into a pulp fiber suspension, and solidified by dewatering using a 200-mesh wire. The wet-state paper sheet was pressed at 350 kPa for 3 min, and dried at 105 C for 2 h. The dried raw paper denoted as RP whose composition is summarized in Table 1 was subsequently heattreated at 800 C for 5 h. The heat-treated HDT-dispersed RP was cut into disc-shaped pieces (16 mm in diameter and 1.1 mm in thickness), and were immersed in a Ni(NO3)2 aqueous solution with a certain Ni(NO3)2 concentration for Ni loading, followed by drying
Fig. 3. Microstructure of reduced PSC-4: (a) EDX elemental mapping corresponding to Fig. 2(b), (b) STEM images focusing on an HDT-derived particle, showing nano-sized Ni dispersion.
T. Quang-Tuyen et al. / Journal of Power Sources 283 (2015) 320e327
323
process at 105 C for 3 h, and subsequent heat treatment at 800 C for 5 h to burn out nitrate. In this study, 5 types of PSC with different Ni loading (PSC-1, 2, 3, 4 and 5) listed in Table 2 were prepared using the HDT-dispersed RP. 2.3. Preparation of reference catalyst Nickel yttria-stabilized zirconia (NieYSZ) and nickel scandiastabilized zirconia (NieScSZ) are common SOFC anode materials so-called cermet anode. NieScSZ is known to be coke-tolerant compared to NieYSZ [24e26]. In addition, NieScSZ interaction is stronger than NieYSZ interaction, a strong interaction is against Ni agglomeration [27]. Therefore, NieScSZ was selected as a reference catalyst in this study. 5 types of pelletized NieScSZ (Cermet-1, 2, 3, 4 and 5) listed in Table 2 were prepared from commercial NieScSZ half-cell (Japan Fine Ceramics) by removing ScSZ electrolyte thin film sintered on NieScSZ anode substrate using abrasive papers. Ni content in the pelletized NieScSZ cermet anode was controlled by the abrasive level. Fig. 4. XRD patterns of (a) raw HDT powder, (b) HDT powder calcined at 800 C for 5 h, (c) HDT-free RP heat-treated at 800 C for 5 h, (d) HDT-dispersed RP heat-treated at 800 C for 5 h and (e) PSC-4 after reduction treatment.
2.4. Cell fabrication Anode- and electrolyte-supported cells were fabricated for the evaluation of cell performance in the feed of oleic-FAME. NieScSZ anode-supported half-cell with a diameter of 20 mm (Japan Fine Ceramics, Japan) was used to fabricate anode-supported single cell. Current collector paste based on the mixture of 80 wt% NiO and 20 wt% ScSZ was screen-printed on the anode surface of the halfcell and subsequently sintered at 1300 C for 3 h. A mixture of 50 wt% LSM and 50 wt% ScSZ was adopted as a cathode functional layer, and coarse LSM was applied as a cathode current collector layer. These pastes were screen-printed on the dense ScSZ electrolyte thin film of half-cell, subsequently sintered at 1200 C for 5 h to have porous cathode with the area of 0.64 cm2. ScSZ electrolyte plate with a thickness of 220 mm and a diameter of 20 mm were used to make electrolyte-supported single cells. Anode pastes based on the powder mixtures of 56 wt% NiO e 44 wt % ScSZ and 80 wt% NiO e 20 wt% ScSZ were screen-printed on the dense ScSZ plate as anode functional layer and current collector layer, respectively, subsequently sintered at 1300 C for 3 h. Cathode was prepared by the same procedure as anode-supported cell mentioned above. Electrode area of the obtained single cell was 0.64 cm2. Component materials of the anode- and the electrolytesupported single cells were summarized in Table 3. 2.5. Steam reforming test To evaluate catalytic performances of HDT-dispersed PSC and cermet anode (cf. Table 2), these catalysts were mounted in an alumina cylindrical reactor as described in left side of Fig. 1. Prior to steam reforming test, the mounted catalyst sample was reduced in the reactor in the flow of pure hydrogen at 800 C for 15 and 1 h for PSC and cermet anode, respectively. Then, Oleic-FAME and deionized water were supplied into the evaporator at 500 C by micro liquid pumps (LC-20AD, Shimadzu, Japan) with molecular flow rates of 1 and 67 mmol h1, respectively, so that steam to carbon ratio (S/C) becomes 3.5. Then, the gaseous mixture was fed to the mounted catalyst set at 800 C using 50 ml min1 of N2 as carrier gas. During steam reforming test, composition of reformate was monitored using an automatic gas chromatograph GC-20B (Shimadzu Corp., Japan).
Fig. 5. Comparison of catalytic activity between PSC and porous NieScSZ cermet anode for steam reforming of oleic-FAME (S/C ¼ 3.5) at 800 C: (a) fuel conversion and (b) C2H4 concentration in reformate. The equilibrium values (dry basis), which were estimated by HSC 7.1 software (Outokumpu Research Oy, Finland), were also plotted.
2.6. Electrochemical measurement Electrochemical measurement setup for the anode-supported
324
T. Quang-Tuyen et al. / Journal of Power Sources 283 (2015) 320e327
Fig. 6. Steam reforming of oleic-FAME (S/C ¼ 3.5) using PSC-4 (W/F ¼ 3.2 gNi h mol1) and cermet-3 (W/F ¼ 300 gNi h mol1) at 800 C: (a) fuel conversion, and concentrations of (b) H2, (c) CO and CO2, and (d) CH4 and C2H4 in reformate. The equilibrium values (dry basis) were also plotted.
single cell fueled by BDF was described elsewhere [23]. The setup for the electrolyte-supported cell connected with PSC is shown in Fig. 1. The reforming and electrochemical reaction zones were separated and series connected. In the reforming zone, two circular PSCs (PSC-4) were stacked and placed in an alumina cylinder. In the electrochemical reaction zone, electrolyte-supported single cell was placed. After a complete purge of the system with pure nitrogen, the PSC and the anode of the single cell were subjected to reduction treatment at 800 C for 15 h in pure H2 flow. After the reduction treatment, oleic-FAME and deionized water were supplied into the evaporator at 500 C by micro liquid pumps. Molecular flow rates of oleic-FAME and deionized water were set at 1 and 67 mmol h1, respectively. Then, the gaseous mixture was fed to the reforming zone with PSC set at 800 C using 50 ml min1 of N2 as carrier gas, and the reformate gas was supplied to SOFC anode. After the analysis of anode off-gas composition under open circuit condition, currentevoltage curve (IeV curve) was measured. To evaluate polarization resistance of a single cell, impedance measurement was carried out under open-circuit condition with a VersaSTAT3 frequency response analyzer (Princeton Applied Research, USA) in a frequency range between 0.1 and 100 kHz with voltage amplitude of 10 mV. After the impedance measurement, current load was increased up to 128 mA (corresponding to 0.2 A cm2) and then the terminal voltage was galvanostatically measured.
2.7. Characterization Microstructures of PSC and SOFC anode were investigated using a field emission scanning electron microscope (FESEM-S5200, Hitachi High-Technologies, Japan) equipped with an energy dispersive X-ray spectrometer (EDS) and a scanning transmission electron microscope (STEM-HD2300A, Hitachi High Technology, Japan). Crystal structures of the samples were analyzed by D8 Advance powder X-ray diffractometer (Bruker AXS GmbH, Germany) with CuKa radiation (l ¼ 1.54056 Å) at 40 kV and 40 mA in the scan range between 2q ¼ 10 and 80 (scan rate: 0.05 /step). Ni content in each PSC was analyzed by ICP apparatus (iCAP 6500 DUO, Thermo Fisher Scientific Inc., Japan). 3. Results and discussion 3.1. Microstructure of the prepared PSC Fig. 2 shows FESEM images of (a) porous NieScSZ cermet anode and (b) PSC-4. Cermet anode produced by powder compaction had the porosity of 32% and the pore size diameter of approx. 1 mm. Particle size of component materials (Ni and ScSZ) was around 0.5e1.0 mm. PSC has completely different microstructure in which ZrO2 sol (Zs) acts as binder for bonding inorganic fibers to form fiber network. Porosity and average pore diameter of PSC are above 80% and ca. 20 mm, respectively. In PSC, as shown in Fig. 3(a) Ni was
T. Quang-Tuyen et al. / Journal of Power Sources 283 (2015) 320e327
325
Fig. 8. Galvanostatic tests of SOFC single cells operated with oleic-FAME (S/C ¼ 3.5) at 800 C under 0.2 A cm2.
Hydrotalcite-like structure can be reconstructed from this MgO in an aqueous solution by taking up anion [28,29,31]. For the heattreated HDT-free RP, strong monoclinic and weak cubic zirconia peaks reflecting ZrO2 sol (Zs) dispersed in the inorganic fiber
Fig. 7. (a) IeV curves and (b) electrochemical impedance spectra (for the whole cell) in the feed of wet-H2 and oleic-FAME (S/C ¼ 3.5) at 800 C, showing the comparison between anode-supported cell and electrolyte-supported cell connected with PSC-4.
loaded mainly on HDT-derived Mg-containing oxide particles due to strong affinity between HDT and Ni in the impregnation process [21,28e30]. Fine Ni particles with the size of around 6 nm were observed on HDT-derived oxide dispersed in the inorganic fiber network (see Fig. 3(b)). XRD patterns of (a) raw HDT powder, (b) calcined HDT powder, (c) heat-treated HDT-free RP, (d) heat-treated HDT-dispersed RP and (e) PSC-4 after reduction treatment are shown in Fig. 4. Raw HDT powder was confirmed to have typical layered structure of Mg/ Al-hydrotalcite ((a)) [28,29]. This layered structure was destroyed by the heat-treatment at 800 C to form Al-dissolved MgO, although small peak of hydrotalcite was still observed ((b)).
Table 4 Anode off-gas compositions for anode-supported cell and electrolyte-supported cell connected with PSC-4 in the feed of oleic-FAME (S/C ¼ 3.5) measured under opencircuit condition at 800 C. In this table, measured reformate composition after PSC4 and equilibrium gas composition are listed for the comparison. Gas concentrations (dry basis)/% H2
CO
CO2
CH4 C2H4 C>3
Anode-supported cell 57.8 19.2 9.1 8.0 Electrolyte-supported 69.3 18.8 11.5 0.4 cell connected with PSC-4 PSC-4 66.2 18.9 12.8 2.1 Equilibrium 70.6 14.5 14.9 0.0
5.9 0.0
0.1 0.0
0.0 0.0
0.0 0.0
Total Ni applied/mg 550 10 3.2 e
Fig. 9. Post-test observation of (a, b) anode-supported cell and (cee) electrolytesupported cell connected with PSC-4 operated by oleic-FAME (S/C ¼ 3.5); (a) and (c) are optical images of anode-side, (b) and (d) are FESEM images of inside of anode material, and (e) is FESEM image of PSC-4 after the tests shown in Fig. 8.
326
T. Quang-Tuyen et al. / Journal of Power Sources 283 (2015) 320e327
Fig. 10. Schematic illustration of DIR-operation of SOFC fueled by liquid biofuel; (a) anode-supported cell and (b) electrolyte-supported cell connected with PSC.
network as a binder for connecting fibers were detected ((c)). On the other hand, for the heat-treated HDT-dispersed RP, cubic zirconia phase emerged, indicating Mg dissolution into Zs due to similar ionic radii of Mg2þ (0.72 Å) and Zr4þ (0.72 Å) to promote stabilization of zirconia phase ((d)) [32]. For the PSC ((e)), broad peak of Ni metal corresponding to fine Ni particles formed on the MgO-based basic oxide support derived from HDT (see Fig. 3(b)) which acts as a promoter of reforming reaction [33e35]. Cubic zirconia phase caused by the Mg dissolution into Zs further emerged due to heat-treatment to burn out nitrate after the impregnation process and subsequent high-temperature reduction treatment. The Mg dissolution, which may cause the loss of positive effect of HDT as a promoter, has to be prevented for higher catalytic performance. The importance of chemical compatibility between HDT and inorganic binder, Zs, for higher catalytic performance of PSC will be discussed in elsewhere. No peak reflecting Cf, which forms inorganic fiber network, was detected due to its amorphous structure. 3.2. Steam reforming of model biodiesel using PSC Catalytic activities of PSC and porous NieScSZ cermet anode for steam reforming of oleic-FAME were evaluated under S/C ¼ 3.5 at 800 C. Fig. 5(a) and (b) show fuel conversion and C2H4 concentration in reformate, respectively, as a function of W/F (weight of Ni catalyst per fuel molecular flow rate, gNi h mol1). The off-gas composition for the thermal decomposition of oleic-FAME without any catalyst was also plotted as “blank”. Fuel conversion increased with W/F and converged maximum (81%) at W/F of around 3.2 and 300 gNi h mol1 for PSC and cermet anode, respectively. PSC exhibited fuel conversion comparable to Ni-based cermet anode with approx. 1/100 catalyst weights at 800 C. The rather low maximum fuel conversion in this study is due to coke formation on the wall of alumina reactor and fuel feeding line. Formation of C2H4, well-known as a precursor of coking [36,37], was not detected above 3.2 and 300 gNi h mol1 for PSC and cermet anode, respectively. Fig. 6 shows the results of steam reforming of oleic-FAME over PSC-4 and cermet-3 at W/F of 3.2 and 300 gNi h mol1, respectively, corresponding to the conditions where fuel conversion reaches maximum and C2H4 concentration approaches zero (see Fig. 5). As
shown in this figure, PSC exhibited stable catalytic performance comparable to cermet anode even under approx. 1/100 contact time. In the case of cermet-3, concentrations of CO and CH4 started to increase accompanied by the decrease in CO2 and H2 concentrations at 4 h, indicating the occurrence of C þ 2H2 / CH4 and C þ CO2 / 2CO, whereas concentrations of all species were kept constant for PSC-4. 3.3. Electrochemical performance of SOFC fueled by model biodiesel Electrochemical performances of anode-supported cell and electrolyte-supported cell connected with PSC-4 were compared in the feed of wet-H2 (97 vol.% H2) and oleic-FAME (S/C ¼ 3.5) at 800 C. Fig. 7(a) and (b) show IV curves and electrochemical impedance spectra of SOFC single cells, respectively. In these impedance plots, high frequency intercept with the real axis corresponds to ohmic resistance of the tested cells, and the distance between low and high frequency intercepts with the real axis is related to polarization resistance of the tested cells. The impedance plots were measured between the anode and the cathode. Considering that only the condition of anode side was changed by switching fuel from wet-H2 to oleic-FAME, plots for the frequency region lower than 1 kHz may reflect gas diffusion and gas conversion resistances of the anode [38]. For anode-supported cell, gradient of IV curve in the feed of oleic-FAME (S/C ¼ 3.5) was larger than that in the feed of wet-H2 due to slightly larger polarization resistance for oleic-FAME than that for wet-H2 (see Fig. 7(b)). Considering the fact that only the impedance of the low frequency region increased, slower mass transport process compared to wetH2 is a main reason of the larger polarization resistance due to low H2 concentration for the case of oleic-FAME [39]. On the other hand, for the electrolyte-supported cell connected with PSC-4, gradient of IV curve in the feed of oleic-FAME was smaller than that in the feed of wet-H2. Anode off-gas compositions in the feed of oleic-FAME were summarized in Table 4. Under the present condition, PSC-4 can contribute to convert the feedstock to nearly equilibrium composition. Formation of H2 close to equilibrium (70.6% in dry basis), suppression of CH4 concentration and prevention of C2H4 formation were achieved by the application of PSC. As far as H2 is main reactant of anode reaction, higher humidity will result in lower charge transfer resistance [40,41]. In the feed of oleic-FAME,
T. Quang-Tuyen et al. / Journal of Power Sources 283 (2015) 320e327
by the effect of PSC, fuel composition in the vicinity of anode functional layer is considered to be close to equilibrium (cf. Table 4), and this fuel composition can be regarded as highly-humidified H2, leading to lower interfacial impedance as shown in Fig. 7(b). Fig. 8 shows the cell voltages of anode-supported cell and electrolyte-supported cell connected with PSC-4 in the feed of oleic-FAME (S/C ¼ 3.5) under 0.2 A cm2 at 800 C. Electrolytesupported cell connected with PSC-4 exhibited stable cell voltage, whereas stable operation could not be achieved for the anodesupported cell. Post-test observations of the anodes were performed after the tests of Fig. 8. As shown in Fig. 9, in the case of anode-supported cell severe coking occurred not only on the anode surface but also inside the porous anode. On the contrary, anode material of electrolyte-supported cell was quite clean after the test, indicating that the application of PSC enable us to operate BDFfueled SOFC with lower risk of coking. Microstructure of anode material for anode-supported cell, which is optimized to fulfill functions as a support material, is not suitable for the DIR operation, especially in the feed of heavyhydrocarbons such as liquid-biofuels. By the use of PSC technology, both of electrochemical and reforming reactions can be separately optimized so that best performance can be brought out. 4. Conclusions This study demonstrated stable operation of SOFC fueled by model biodiesel by the application of paper-structured catalyst (PSC). PSC, in which Mg/Al-hydrotalcite (HDT) powder is dispersed, exhibited catalytic activity for the steam reforming of oleic-FAME comparable to the conventional NieScSZ cermet anode with about one-hundredth Ni weights due to its adequate microstructure for the reforming of heavy-hydrocarbons with large pore size diameter and well-dispersed catalyst particles. In the DIR operation, when anode-supported cell is selected (see Fig. 10(a)), complex reaction including three reactions, (I) thermal decomposition of heavy hydrocarbons, (II) steam reforming of heavy and light hydrocarbons (light HCs) and (III) electrochemical oxidation of H2 and CO, take place in anode material. However, by the application of PSC (see Fig. 10(b)), functions of reforming and electrochemical reaction sites can be well-separated without losses of both functions, leading to stable operation with lower risk of coking. Acknowledgments This study was supported by Industrial Technology Research Grant Program in 2011 from New Energy and Industrial Technology Development Organization (NEDO) of Japan. References [1] B.C.H. Steele, Nature 400 (1999) 619e621.
327
[2] E.D. Wachsman, C.A. Marlowe, K.T. Lee, Energy Environ. Sci. 5 (2012) 5498e5509. [3] V.M. Janardhanan, V.V. Heuveline, O. Deutschmann, J. Power Sources 172 (2007) 296e307. [4] D. Mogensen, J.D. Grunwaldt, P.V. Hendriksen, K. Dam-Johansen, J.U. Nielsen, J. Power Sources 196 (2011) 25e38. [5] N.Q. Minh, Solid State Ionics 174 (2004) 271e277. [6] P. Leone, A. Lanzini, G.A. Ortigoza-Villalba, R. Borchiellini, Chem. Eng. J. 191 (2012) 349e355. [7] C.M. Finnerty, N.J. Coe, R.H. Cunningham, R.M. Ormerod, Catal. Today 46 (1998) 137e145. [8] Y. Shiratori, T. Ijichi, T. Oshima, K. Sasaki, Int. J. Hydrogen Energy 35 (2010) 7905e7912. [9] K. Yamaji, Y.P. Xiong, T. Horita, N. Sakai, M.E. Brito, H. Yokokawa, J. Power Sources 159 (2006) 885e890. [10] K. Kendall, M. Slinn, J. Preece, J. Power Sources 157 (2006) 750e753. [11] G.J. Saunders, J. Preece, K. Kendall, J. Power Sources 131 (2004) 23e26. [12] T. Quang-Tuyen, Y. Shiratori, K. Sasaki, Int. J. Energy Res. 37 (2013) 609e616. [13] K. Eguchi, Y. Kunisa, K. Adachi, H. Arai, J. Electrochem. Soc. 143 (1996) 3699e3703. [14] T. Matsui, R. Kishida, J.Y. Kim, H. Muroyama, K. Eguchi, J. Electrochem. Soc. 157 (2010) B776eB781. [15] S. Park, R. Cracium, J.M. Vohs, R.J. Gorte, J. Electrochem. Soc. 146 (1999) 3603e3605. [16] S. Park, M. Vohs, R.J. Gorte, Nature 404 (2000) 265e267. [17] S. McIntosh, R.J. Gorte, Chem. Rev. 104 (2004) 4845e4865. [18] J.M. Vohs, R.J. Gorte, Curr. Opin. Colloid Interface Sci. 14 (2009) 236e244. [19] Z. Zhan, S.A. Barnett, Science 308 (2005) 844e847. [20] Y. Shiratori, T. Ogura, H. Nakajima, M. Sakamoto, Y. Takahashi, Y. Wakita, T. Kitaoka, K. Sasaki, Int. J. Hydrogen Energy 38 (2013) 10542e10551. [21] Y. Shiratori, M. Sakamoto, T. Uchida, H. Le, T. Quang-Tuyen, K. Sasaki, Int. J. Hydrogen Energy, (under review). [22] Y. Shiratori, T. Quang-Tuyen, K. Sasaki, Int. J. Hydrogen Energy 38 (2013) 9856e9866. [23] Y. Shiratori, Q.T. Tran, Y. Takahashi, K. Sasaki, ECS Trans. 35 (2011) 2641e2651. [24] H. Sumi, K. Ukai, Y. Mizutani, H. Mori, C. Wen, H. Takahashi, O. Yamamoto, Solid State Ionics 172 (2004) 151e156. [25] T. Iida, M. Kawano, T. Matsui, R. Kikuchi, K. Eguchi, J. Electrochem. Soc. 154 (2007) B234eB241. [26] H. Sumi, P. Puengjinda, H. Muroyama, T. Matsui, R. Kikuchi, K. Eguchi, J. Power Sources 196 (2011) 6048e6054. [27] K. Ke, A. Gunji, H. Mori, S. Tsuchida, H. Takahashi, K. Ukai, Y. Mizutani, H. Sumi, M. Yokoyama, K. Waki, Solid State Ionics 177 (2006) 541e547. [28] T. Sato, K. Kato, T. Endo, M. Shimada, React. Solids 2 (1986) 253e260. [29] T. Shishido, M. Sukenobu, H. Morioka, R. Furukawa, H. Shirahase, K. Takehira, Catal. Lett. 73 (2001) 21e26. [30] D. Li, M. Koike, J. Chen, Y. Nakagawa, K. Tomishige, Int. J. Hydrogen Energy (2014) 10959e10970. [31] M. Mokhtar, A. Inayat, J. Ofili, W. Schwieger, Appl. Clay Sci. 50 (2010) 176e181. [32] A.R. West, Basic Solid State Chemistry, second ed., John Wiley & Sons Ltd., UK, 1999, pp. 227e228. [33] K. Tomishige, Y.G. Chen, K. Fujimoto, J. Catal. 181 (1999) 91e103. [34] E. Ruckenstein, Y.H. Hu, Appl. Catal. A 183 (1999) 85e92. [35] J.T. Richardson, R.M. Scates, M.V. Twigg, Appl. Catal. A 267 (2004) 35e46. [36] J.R. Rostrup-Nielsen, T.S. Christensen, I. Dybkjaer, Stud. Surf. Sci. Catal. 113 (1998) 81e95. [37] S. Yoon, I. Kang, J. Bae, Int. J. Hydrogen Energy 33 (2008) 4780e4788. [38] R. Barfod, M. Mogensen, T. KlemensØ, A. Hagen, Y. Liu, P.V. Hendriksen, J. Electrochem. Soc. 154 (2007) B371eB378. [39] H. Sumi, T. Yamaguchi, K. Hamamoto, T. Suzuki, Y. Fujishiro, T. Matsui, K. Efuchi, Electrochem. Acta 67 (2012) 159e165. [40] S.P. Jiang, S.P.S. Badwal, J. Electrochem. Soc. 144 (1997) 3777e3784. [41] T. Hosoi, T. Yonekura, K. Sunada, K. Sasaki, J. Electrochem. Soc. 162 (2015) F136eF152.
Contents lists available at ScienceDirect
Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour
Effectiveness of paper-structured catalyst for the operation of biodiesel-fueled solid oxide fuel cell Tran Quang-Tuyen a, b, *, Taku Kaida a, Mio Sakamoto a, Kazunari Sasaki a, b, c, d, Yusuke Shiratori a, b, c, d a
Faculty of Engineering, Kyushu University, Motooka 744, Nishi-ku, Fukuoka 819-0395, Japan International Research Center for Hydrogen Energy, Kyushu University, Motooka 744, Nishi-ku, Fukuoka 819-0395, Japan International Institute for Carbon-Neutral Energy Research (WPI), Kyushu University, Motooka 744, Nishi-ku, Fukuoka 819-0395, Japan d Next-Generation Fuel Cell Research Center, Kyushu University, Motooka 744, Nishi-ku, Fukuoka 819-0395, Japan b c
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 model biodiesel fuel (BDF) was applied as an alternative fuel for SOFC. Novel paper-structured catalyst (PSC) containing Mg/Al-hydrotalcite was synthesized. PSC exhibited fuel conversion comparable to SOFC anode with less than 1/100 Ni weight. Long-term stability of SOFC fueled by BDF was achieved by the application of PSC.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 22 January 2015 Received in revised form 19 February 2015 Accepted 22 February 2015 Available online 26 February 2015
Mg/Al-hydrotalcite (HDT)-dispersed paper-structured catalyst (PSC) was prepared by a simple papermaking process. The PSC exhibited excellent catalytic activity for the steam reforming of model biodiesel fuel (BDF), pure oleic acid methyl ester (oleic-FAME, C19H36O2) which is a mono-unsaturated component of practical BDFs. The PSC exhibited fuel conversion comparable to a pelletized catalyst material, here, conventional Niezirconia cermet anode for solid oxide fuel cell (SOFC) with less than onehundredth Ni weight. Performance of electrolyte-supported cell connected with the PSC was evaluated in the feed of oleic-FAME, and stable operation was achieved. After 60 h test, coking was not observed in both SOFC anode and PSC. © 2015 Elsevier B.V. All rights reserved.
Keywords: Paper-structured catalyst Mg/Al-hydrotalcite Steam reforming Oleic-FAME Solid oxide fuel cell
1. Introduction
* Corresponding author. Faculty of Engineering, Kyushu University, Motooka 744, Nishi-ku, Fukuoka 819-0395, Japan. E-mail addresses: [email protected], [email protected] (T. Quang-Tuyen). http://dx.doi.org/10.1016/j.jpowsour.2015.02.116 0378-7753/© 2015 Elsevier B.V. All rights reserved.
High temperature solid oxide fuel cell (SOFC) can accept direct feed of hydrocarbon fuels. SOFC is expected to be the most efficient device for direct conversion of chemical energy into electricity [1,2]. Reforming of hydrocarbons can proceed in the porous cermet anode to produce H2 and CO, which electrochemically oxidized to
T. Quang-Tuyen et al. / Journal of Power Sources 283 (2015) 320e327 Table 1 Composition of raw papers (RPs) after drying. Type of RP
HDT/wt% Cf/wt% Zs/wt% Cationic Anionic Pulp/wt% polymer/wt% polymer/wt%
HDT-free 0 HDT-dispersed 14.7
86.1 73.3
8.6 7.3
0.5 0.5
0.5 0.5
4.3 3.7
Table 2 Ni contents in the HDT-dispersed PSCs (for two pieces of each type of PSC with 16 mm in diameter and 1.1 mm in thickness) and pelletized NieScSZ cermet anodes. Paper-structured catalysta
Ni loading/mg
NieScSZ cermet anodeb (Reference catalyst)
Ni loading/mg
PSC-1 PSC-2 PSC-3 PSC-4 PSC-5
0.6 2.4 2.8 3.2 5.6
Cermet-1 Cermet-2 Cermet-3 Cermet-4 Cermet-5
170 240 300 430 1200
a Weights of HDT, Cf and Zs in two pieces of each type of PSC are 18.9, 95 and 9.5 mg, respectively. b Weight ratio of Ni:ScSZ is 1:1.
generate electricity and heat [3,4]. This operation is called direct internal reforming (DIR) operation, which brings us several advantages such as downsizing and cost reduction of fuel cell system and the reduction of energy consumption for the stack cooling by the supply of excess air due to strong endothermicity of the reforming reaction [5,6]. Nickel is commonly-used anode material for SOFC due to its excellent electro-catalytic properties. However, coking on Ni is a main drawback of DIR operation of SOFC fueled by light hydrocarbons [7e9] as well as heavy hydrocarbons [10e12]. Carbon deposition on and inside of porous anode material is one of the most crucial issues of DIR-SOFC, because it significantly degrades fuel cell performance as the results of the coverage of electrochemical and reforming reaction sites. To solve this problem, many approaches
321
have been reported. The first approach is careful control of steam to carbon ratio (S/C). Excess amount of water is essential to promote reforming reaction, however high S/C ratio results in the reduction of electromotive force, furthermore deteriorates catalytic activity of the porous cermet anode due to oxidation of Ni [13,14]. The second approach is incorporation of coking-resistant catalyst materials such as copper [15,16], but low melting temperatures of copper and its oxides limit the use of them for SOFC anode [17,18]. Zhan et al. applied a thin RueCeO2 catalyst layer on the top of a conventional NieYSZ anode [19]. They obtained stable cell operation without coking by the direct feed of fuel mixture of 5% iso-octane and 9% air diluted by 86% CO2 to a SOFC single cell at 770 C under current density of 0.6 A cm2 for 50 h, although difficulties of fuel diffusion and current collection must be overcome. Recently, flexible paper structured-catalyst (PSC) based on inorganic fiber network was developed to solve the problems associated with the DIR operation of SOFC mentioned above. The PSC exhibited excellent catalytic performance for dry reforming of methane [20,21]. Long term stability of SOFC fueled by pure oleic acid methyl ester (oleic-FAME), which is a main component of practical biodiesel fuels (BDFs), was confirmed by the connection of SOFC with PSC [22]. However, for the practical application of PSCs to SOFC systems, further improvement of catalytic performance is necessary. In this study, Ni loaded PSC with improved catalytic performance in which layered double hydroxide particles containing Mg2þ and Al3þ called Mg/Al-hydrotalcite (HDT) were dispersed in the paper-making process was tested for the stream reforming of oleic-FAME. Then, effectiveness of the HDT-dispersed PSC for the operation of biodiesel-fueled SOFC was demonstrated. 2. Experimental 2.1. Model fuel Biodiesel fuel (BDF) is produced from biomass resources such as vegetable oil, waste cooking oil and animal fat by alkali catalyzed
Table 3 Component materials of anode- and electrolyte-supported cells prepared in this study. Anode
Anodesupporteda Electrolytesupportedb a b
Electrolyte
Current collector
Substrate
Functional layer
NiOeScSZ (30 mm) NiOeScSZ (30 mm)
NiOeScSZ (800 mm)
e NiOeScSZ (60 mm)
ScSZ (22 mm) ScSZ (220 mm)
Cathode Functional layer
Current collector
LSM-ScSZ (30 mm) LSM-ScSZ (30 mm)
LSM (30 mm) LSM (30 mm)
Anode contains 550 mg of Ni metal. Anode contains 6.8 mg of Ni metal.
Fig. 1. Experimental setup for testing SOFC connected with PSC in the feed of oleic-FAME.
322
T. Quang-Tuyen et al. / Journal of Power Sources 283 (2015) 320e327
Fig. 2. FESEM images of (a) porous NieScSZ cermet anode and (b) PSC-4.
trans-esterification reaction. Now, BDF is used as an alternative fuel to petro-diesel in transportation sector. Oleic acid methyl ester (Oleic-FAME, C19H36O2), which is mono-unsaturated fuel and one of the main components of popular BDFs such as Palm-, Jatropha-, and Soybean-BDFs [23], was chosen as a model BDF in this study. 2.2. Preparation of paper-structured catalyst (PSC) PSC was prepared by an established papermaking technique through a dual polyelectrolyte retention system. Detailed procedure has been reported elsewhere [20]. A water suspension of the mixture of ceramic fiber denoted as Cf (SiO2:52%, Al2O3:48%; IBIDEN Ltd., Japan) with 500 mm in length and 2 mm in diameter and HDT powder ([Mg6Al2(OH)16CO3]4H2O; Wako Pure Chemical
Industries Ltd., Japan) was prepared. Then, cationic polymer, PDADMAC (Sigma Aldrich LCC, USA), zirconia sol (ZrO2 without any stabilizer) denoted as Zs (Daiichi Kigenso Kagaku Kogyo Co., Ltd., Japan) and anionic polymer, A-PAM FA-405S (Fujikasui Engineering Co., Ltd., Japan), were added to the suspension in that order and mixed for 3 min. The mixture was poured into a pulp fiber suspension, and solidified by dewatering using a 200-mesh wire. The wet-state paper sheet was pressed at 350 kPa for 3 min, and dried at 105 C for 2 h. The dried raw paper denoted as RP whose composition is summarized in Table 1 was subsequently heattreated at 800 C for 5 h. The heat-treated HDT-dispersed RP was cut into disc-shaped pieces (16 mm in diameter and 1.1 mm in thickness), and were immersed in a Ni(NO3)2 aqueous solution with a certain Ni(NO3)2 concentration for Ni loading, followed by drying
Fig. 3. Microstructure of reduced PSC-4: (a) EDX elemental mapping corresponding to Fig. 2(b), (b) STEM images focusing on an HDT-derived particle, showing nano-sized Ni dispersion.
T. Quang-Tuyen et al. / Journal of Power Sources 283 (2015) 320e327
323
process at 105 C for 3 h, and subsequent heat treatment at 800 C for 5 h to burn out nitrate. In this study, 5 types of PSC with different Ni loading (PSC-1, 2, 3, 4 and 5) listed in Table 2 were prepared using the HDT-dispersed RP. 2.3. Preparation of reference catalyst Nickel yttria-stabilized zirconia (NieYSZ) and nickel scandiastabilized zirconia (NieScSZ) are common SOFC anode materials so-called cermet anode. NieScSZ is known to be coke-tolerant compared to NieYSZ [24e26]. In addition, NieScSZ interaction is stronger than NieYSZ interaction, a strong interaction is against Ni agglomeration [27]. Therefore, NieScSZ was selected as a reference catalyst in this study. 5 types of pelletized NieScSZ (Cermet-1, 2, 3, 4 and 5) listed in Table 2 were prepared from commercial NieScSZ half-cell (Japan Fine Ceramics) by removing ScSZ electrolyte thin film sintered on NieScSZ anode substrate using abrasive papers. Ni content in the pelletized NieScSZ cermet anode was controlled by the abrasive level. Fig. 4. XRD patterns of (a) raw HDT powder, (b) HDT powder calcined at 800 C for 5 h, (c) HDT-free RP heat-treated at 800 C for 5 h, (d) HDT-dispersed RP heat-treated at 800 C for 5 h and (e) PSC-4 after reduction treatment.
2.4. Cell fabrication Anode- and electrolyte-supported cells were fabricated for the evaluation of cell performance in the feed of oleic-FAME. NieScSZ anode-supported half-cell with a diameter of 20 mm (Japan Fine Ceramics, Japan) was used to fabricate anode-supported single cell. Current collector paste based on the mixture of 80 wt% NiO and 20 wt% ScSZ was screen-printed on the anode surface of the halfcell and subsequently sintered at 1300 C for 3 h. A mixture of 50 wt% LSM and 50 wt% ScSZ was adopted as a cathode functional layer, and coarse LSM was applied as a cathode current collector layer. These pastes were screen-printed on the dense ScSZ electrolyte thin film of half-cell, subsequently sintered at 1200 C for 5 h to have porous cathode with the area of 0.64 cm2. ScSZ electrolyte plate with a thickness of 220 mm and a diameter of 20 mm were used to make electrolyte-supported single cells. Anode pastes based on the powder mixtures of 56 wt% NiO e 44 wt % ScSZ and 80 wt% NiO e 20 wt% ScSZ were screen-printed on the dense ScSZ plate as anode functional layer and current collector layer, respectively, subsequently sintered at 1300 C for 3 h. Cathode was prepared by the same procedure as anode-supported cell mentioned above. Electrode area of the obtained single cell was 0.64 cm2. Component materials of the anode- and the electrolytesupported single cells were summarized in Table 3. 2.5. Steam reforming test To evaluate catalytic performances of HDT-dispersed PSC and cermet anode (cf. Table 2), these catalysts were mounted in an alumina cylindrical reactor as described in left side of Fig. 1. Prior to steam reforming test, the mounted catalyst sample was reduced in the reactor in the flow of pure hydrogen at 800 C for 15 and 1 h for PSC and cermet anode, respectively. Then, Oleic-FAME and deionized water were supplied into the evaporator at 500 C by micro liquid pumps (LC-20AD, Shimadzu, Japan) with molecular flow rates of 1 and 67 mmol h1, respectively, so that steam to carbon ratio (S/C) becomes 3.5. Then, the gaseous mixture was fed to the mounted catalyst set at 800 C using 50 ml min1 of N2 as carrier gas. During steam reforming test, composition of reformate was monitored using an automatic gas chromatograph GC-20B (Shimadzu Corp., Japan).
Fig. 5. Comparison of catalytic activity between PSC and porous NieScSZ cermet anode for steam reforming of oleic-FAME (S/C ¼ 3.5) at 800 C: (a) fuel conversion and (b) C2H4 concentration in reformate. The equilibrium values (dry basis), which were estimated by HSC 7.1 software (Outokumpu Research Oy, Finland), were also plotted.
2.6. Electrochemical measurement Electrochemical measurement setup for the anode-supported
324
T. Quang-Tuyen et al. / Journal of Power Sources 283 (2015) 320e327
Fig. 6. Steam reforming of oleic-FAME (S/C ¼ 3.5) using PSC-4 (W/F ¼ 3.2 gNi h mol1) and cermet-3 (W/F ¼ 300 gNi h mol1) at 800 C: (a) fuel conversion, and concentrations of (b) H2, (c) CO and CO2, and (d) CH4 and C2H4 in reformate. The equilibrium values (dry basis) were also plotted.
single cell fueled by BDF was described elsewhere [23]. The setup for the electrolyte-supported cell connected with PSC is shown in Fig. 1. The reforming and electrochemical reaction zones were separated and series connected. In the reforming zone, two circular PSCs (PSC-4) were stacked and placed in an alumina cylinder. In the electrochemical reaction zone, electrolyte-supported single cell was placed. After a complete purge of the system with pure nitrogen, the PSC and the anode of the single cell were subjected to reduction treatment at 800 C for 15 h in pure H2 flow. After the reduction treatment, oleic-FAME and deionized water were supplied into the evaporator at 500 C by micro liquid pumps. Molecular flow rates of oleic-FAME and deionized water were set at 1 and 67 mmol h1, respectively. Then, the gaseous mixture was fed to the reforming zone with PSC set at 800 C using 50 ml min1 of N2 as carrier gas, and the reformate gas was supplied to SOFC anode. After the analysis of anode off-gas composition under open circuit condition, currentevoltage curve (IeV curve) was measured. To evaluate polarization resistance of a single cell, impedance measurement was carried out under open-circuit condition with a VersaSTAT3 frequency response analyzer (Princeton Applied Research, USA) in a frequency range between 0.1 and 100 kHz with voltage amplitude of 10 mV. After the impedance measurement, current load was increased up to 128 mA (corresponding to 0.2 A cm2) and then the terminal voltage was galvanostatically measured.
2.7. Characterization Microstructures of PSC and SOFC anode were investigated using a field emission scanning electron microscope (FESEM-S5200, Hitachi High-Technologies, Japan) equipped with an energy dispersive X-ray spectrometer (EDS) and a scanning transmission electron microscope (STEM-HD2300A, Hitachi High Technology, Japan). Crystal structures of the samples were analyzed by D8 Advance powder X-ray diffractometer (Bruker AXS GmbH, Germany) with CuKa radiation (l ¼ 1.54056 Å) at 40 kV and 40 mA in the scan range between 2q ¼ 10 and 80 (scan rate: 0.05 /step). Ni content in each PSC was analyzed by ICP apparatus (iCAP 6500 DUO, Thermo Fisher Scientific Inc., Japan). 3. Results and discussion 3.1. Microstructure of the prepared PSC Fig. 2 shows FESEM images of (a) porous NieScSZ cermet anode and (b) PSC-4. Cermet anode produced by powder compaction had the porosity of 32% and the pore size diameter of approx. 1 mm. Particle size of component materials (Ni and ScSZ) was around 0.5e1.0 mm. PSC has completely different microstructure in which ZrO2 sol (Zs) acts as binder for bonding inorganic fibers to form fiber network. Porosity and average pore diameter of PSC are above 80% and ca. 20 mm, respectively. In PSC, as shown in Fig. 3(a) Ni was
T. Quang-Tuyen et al. / Journal of Power Sources 283 (2015) 320e327
325
Fig. 8. Galvanostatic tests of SOFC single cells operated with oleic-FAME (S/C ¼ 3.5) at 800 C under 0.2 A cm2.
Hydrotalcite-like structure can be reconstructed from this MgO in an aqueous solution by taking up anion [28,29,31]. For the heattreated HDT-free RP, strong monoclinic and weak cubic zirconia peaks reflecting ZrO2 sol (Zs) dispersed in the inorganic fiber
Fig. 7. (a) IeV curves and (b) electrochemical impedance spectra (for the whole cell) in the feed of wet-H2 and oleic-FAME (S/C ¼ 3.5) at 800 C, showing the comparison between anode-supported cell and electrolyte-supported cell connected with PSC-4.
loaded mainly on HDT-derived Mg-containing oxide particles due to strong affinity between HDT and Ni in the impregnation process [21,28e30]. Fine Ni particles with the size of around 6 nm were observed on HDT-derived oxide dispersed in the inorganic fiber network (see Fig. 3(b)). XRD patterns of (a) raw HDT powder, (b) calcined HDT powder, (c) heat-treated HDT-free RP, (d) heat-treated HDT-dispersed RP and (e) PSC-4 after reduction treatment are shown in Fig. 4. Raw HDT powder was confirmed to have typical layered structure of Mg/ Al-hydrotalcite ((a)) [28,29]. This layered structure was destroyed by the heat-treatment at 800 C to form Al-dissolved MgO, although small peak of hydrotalcite was still observed ((b)).
Table 4 Anode off-gas compositions for anode-supported cell and electrolyte-supported cell connected with PSC-4 in the feed of oleic-FAME (S/C ¼ 3.5) measured under opencircuit condition at 800 C. In this table, measured reformate composition after PSC4 and equilibrium gas composition are listed for the comparison. Gas concentrations (dry basis)/% H2
CO
CO2
CH4 C2H4 C>3
Anode-supported cell 57.8 19.2 9.1 8.0 Electrolyte-supported 69.3 18.8 11.5 0.4 cell connected with PSC-4 PSC-4 66.2 18.9 12.8 2.1 Equilibrium 70.6 14.5 14.9 0.0
5.9 0.0
0.1 0.0
0.0 0.0
0.0 0.0
Total Ni applied/mg 550 10 3.2 e
Fig. 9. Post-test observation of (a, b) anode-supported cell and (cee) electrolytesupported cell connected with PSC-4 operated by oleic-FAME (S/C ¼ 3.5); (a) and (c) are optical images of anode-side, (b) and (d) are FESEM images of inside of anode material, and (e) is FESEM image of PSC-4 after the tests shown in Fig. 8.
326
T. Quang-Tuyen et al. / Journal of Power Sources 283 (2015) 320e327
Fig. 10. Schematic illustration of DIR-operation of SOFC fueled by liquid biofuel; (a) anode-supported cell and (b) electrolyte-supported cell connected with PSC.
network as a binder for connecting fibers were detected ((c)). On the other hand, for the heat-treated HDT-dispersed RP, cubic zirconia phase emerged, indicating Mg dissolution into Zs due to similar ionic radii of Mg2þ (0.72 Å) and Zr4þ (0.72 Å) to promote stabilization of zirconia phase ((d)) [32]. For the PSC ((e)), broad peak of Ni metal corresponding to fine Ni particles formed on the MgO-based basic oxide support derived from HDT (see Fig. 3(b)) which acts as a promoter of reforming reaction [33e35]. Cubic zirconia phase caused by the Mg dissolution into Zs further emerged due to heat-treatment to burn out nitrate after the impregnation process and subsequent high-temperature reduction treatment. The Mg dissolution, which may cause the loss of positive effect of HDT as a promoter, has to be prevented for higher catalytic performance. The importance of chemical compatibility between HDT and inorganic binder, Zs, for higher catalytic performance of PSC will be discussed in elsewhere. No peak reflecting Cf, which forms inorganic fiber network, was detected due to its amorphous structure. 3.2. Steam reforming of model biodiesel using PSC Catalytic activities of PSC and porous NieScSZ cermet anode for steam reforming of oleic-FAME were evaluated under S/C ¼ 3.5 at 800 C. Fig. 5(a) and (b) show fuel conversion and C2H4 concentration in reformate, respectively, as a function of W/F (weight of Ni catalyst per fuel molecular flow rate, gNi h mol1). The off-gas composition for the thermal decomposition of oleic-FAME without any catalyst was also plotted as “blank”. Fuel conversion increased with W/F and converged maximum (81%) at W/F of around 3.2 and 300 gNi h mol1 for PSC and cermet anode, respectively. PSC exhibited fuel conversion comparable to Ni-based cermet anode with approx. 1/100 catalyst weights at 800 C. The rather low maximum fuel conversion in this study is due to coke formation on the wall of alumina reactor and fuel feeding line. Formation of C2H4, well-known as a precursor of coking [36,37], was not detected above 3.2 and 300 gNi h mol1 for PSC and cermet anode, respectively. Fig. 6 shows the results of steam reforming of oleic-FAME over PSC-4 and cermet-3 at W/F of 3.2 and 300 gNi h mol1, respectively, corresponding to the conditions where fuel conversion reaches maximum and C2H4 concentration approaches zero (see Fig. 5). As
shown in this figure, PSC exhibited stable catalytic performance comparable to cermet anode even under approx. 1/100 contact time. In the case of cermet-3, concentrations of CO and CH4 started to increase accompanied by the decrease in CO2 and H2 concentrations at 4 h, indicating the occurrence of C þ 2H2 / CH4 and C þ CO2 / 2CO, whereas concentrations of all species were kept constant for PSC-4. 3.3. Electrochemical performance of SOFC fueled by model biodiesel Electrochemical performances of anode-supported cell and electrolyte-supported cell connected with PSC-4 were compared in the feed of wet-H2 (97 vol.% H2) and oleic-FAME (S/C ¼ 3.5) at 800 C. Fig. 7(a) and (b) show IV curves and electrochemical impedance spectra of SOFC single cells, respectively. In these impedance plots, high frequency intercept with the real axis corresponds to ohmic resistance of the tested cells, and the distance between low and high frequency intercepts with the real axis is related to polarization resistance of the tested cells. The impedance plots were measured between the anode and the cathode. Considering that only the condition of anode side was changed by switching fuel from wet-H2 to oleic-FAME, plots for the frequency region lower than 1 kHz may reflect gas diffusion and gas conversion resistances of the anode [38]. For anode-supported cell, gradient of IV curve in the feed of oleic-FAME (S/C ¼ 3.5) was larger than that in the feed of wet-H2 due to slightly larger polarization resistance for oleic-FAME than that for wet-H2 (see Fig. 7(b)). Considering the fact that only the impedance of the low frequency region increased, slower mass transport process compared to wetH2 is a main reason of the larger polarization resistance due to low H2 concentration for the case of oleic-FAME [39]. On the other hand, for the electrolyte-supported cell connected with PSC-4, gradient of IV curve in the feed of oleic-FAME was smaller than that in the feed of wet-H2. Anode off-gas compositions in the feed of oleic-FAME were summarized in Table 4. Under the present condition, PSC-4 can contribute to convert the feedstock to nearly equilibrium composition. Formation of H2 close to equilibrium (70.6% in dry basis), suppression of CH4 concentration and prevention of C2H4 formation were achieved by the application of PSC. As far as H2 is main reactant of anode reaction, higher humidity will result in lower charge transfer resistance [40,41]. In the feed of oleic-FAME,
T. Quang-Tuyen et al. / Journal of Power Sources 283 (2015) 320e327
by the effect of PSC, fuel composition in the vicinity of anode functional layer is considered to be close to equilibrium (cf. Table 4), and this fuel composition can be regarded as highly-humidified H2, leading to lower interfacial impedance as shown in Fig. 7(b). Fig. 8 shows the cell voltages of anode-supported cell and electrolyte-supported cell connected with PSC-4 in the feed of oleic-FAME (S/C ¼ 3.5) under 0.2 A cm2 at 800 C. Electrolytesupported cell connected with PSC-4 exhibited stable cell voltage, whereas stable operation could not be achieved for the anodesupported cell. Post-test observations of the anodes were performed after the tests of Fig. 8. As shown in Fig. 9, in the case of anode-supported cell severe coking occurred not only on the anode surface but also inside the porous anode. On the contrary, anode material of electrolyte-supported cell was quite clean after the test, indicating that the application of PSC enable us to operate BDFfueled SOFC with lower risk of coking. Microstructure of anode material for anode-supported cell, which is optimized to fulfill functions as a support material, is not suitable for the DIR operation, especially in the feed of heavyhydrocarbons such as liquid-biofuels. By the use of PSC technology, both of electrochemical and reforming reactions can be separately optimized so that best performance can be brought out. 4. Conclusions This study demonstrated stable operation of SOFC fueled by model biodiesel by the application of paper-structured catalyst (PSC). PSC, in which Mg/Al-hydrotalcite (HDT) powder is dispersed, exhibited catalytic activity for the steam reforming of oleic-FAME comparable to the conventional NieScSZ cermet anode with about one-hundredth Ni weights due to its adequate microstructure for the reforming of heavy-hydrocarbons with large pore size diameter and well-dispersed catalyst particles. In the DIR operation, when anode-supported cell is selected (see Fig. 10(a)), complex reaction including three reactions, (I) thermal decomposition of heavy hydrocarbons, (II) steam reforming of heavy and light hydrocarbons (light HCs) and (III) electrochemical oxidation of H2 and CO, take place in anode material. However, by the application of PSC (see Fig. 10(b)), functions of reforming and electrochemical reaction sites can be well-separated without losses of both functions, leading to stable operation with lower risk of coking. Acknowledgments This study was supported by Industrial Technology Research Grant Program in 2011 from New Energy and Industrial Technology Development Organization (NEDO) of Japan. References [1] B.C.H. Steele, Nature 400 (1999) 619e621.
327
[2] E.D. Wachsman, C.A. Marlowe, K.T. Lee, Energy Environ. Sci. 5 (2012) 5498e5509. [3] V.M. Janardhanan, V.V. Heuveline, O. Deutschmann, J. Power Sources 172 (2007) 296e307. [4] D. Mogensen, J.D. Grunwaldt, P.V. Hendriksen, K. Dam-Johansen, J.U. Nielsen, J. Power Sources 196 (2011) 25e38. [5] N.Q. Minh, Solid State Ionics 174 (2004) 271e277. [6] P. Leone, A. Lanzini, G.A. Ortigoza-Villalba, R. Borchiellini, Chem. Eng. J. 191 (2012) 349e355. [7] C.M. Finnerty, N.J. Coe, R.H. Cunningham, R.M. Ormerod, Catal. Today 46 (1998) 137e145. [8] Y. Shiratori, T. Ijichi, T. Oshima, K. Sasaki, Int. J. Hydrogen Energy 35 (2010) 7905e7912. [9] K. Yamaji, Y.P. Xiong, T. Horita, N. Sakai, M.E. Brito, H. Yokokawa, J. Power Sources 159 (2006) 885e890. [10] K. Kendall, M. Slinn, J. Preece, J. Power Sources 157 (2006) 750e753. [11] G.J. Saunders, J. Preece, K. Kendall, J. Power Sources 131 (2004) 23e26. [12] T. Quang-Tuyen, Y. Shiratori, K. Sasaki, Int. J. Energy Res. 37 (2013) 609e616. [13] K. Eguchi, Y. Kunisa, K. Adachi, H. Arai, J. Electrochem. Soc. 143 (1996) 3699e3703. [14] T. Matsui, R. Kishida, J.Y. Kim, H. Muroyama, K. Eguchi, J. Electrochem. Soc. 157 (2010) B776eB781. [15] S. Park, R. Cracium, J.M. Vohs, R.J. Gorte, J. Electrochem. Soc. 146 (1999) 3603e3605. [16] S. Park, M. Vohs, R.J. Gorte, Nature 404 (2000) 265e267. [17] S. McIntosh, R.J. Gorte, Chem. Rev. 104 (2004) 4845e4865. [18] J.M. Vohs, R.J. Gorte, Curr. Opin. Colloid Interface Sci. 14 (2009) 236e244. [19] Z. Zhan, S.A. Barnett, Science 308 (2005) 844e847. [20] Y. Shiratori, T. Ogura, H. Nakajima, M. Sakamoto, Y. Takahashi, Y. Wakita, T. Kitaoka, K. Sasaki, Int. J. Hydrogen Energy 38 (2013) 10542e10551. [21] Y. Shiratori, M. Sakamoto, T. Uchida, H. Le, T. Quang-Tuyen, K. Sasaki, Int. J. Hydrogen Energy, (under review). [22] Y. Shiratori, T. Quang-Tuyen, K. Sasaki, Int. J. Hydrogen Energy 38 (2013) 9856e9866. [23] Y. Shiratori, Q.T. Tran, Y. Takahashi, K. Sasaki, ECS Trans. 35 (2011) 2641e2651. [24] H. Sumi, K. Ukai, Y. Mizutani, H. Mori, C. Wen, H. Takahashi, O. Yamamoto, Solid State Ionics 172 (2004) 151e156. [25] T. Iida, M. Kawano, T. Matsui, R. Kikuchi, K. Eguchi, J. Electrochem. Soc. 154 (2007) B234eB241. [26] H. Sumi, P. Puengjinda, H. Muroyama, T. Matsui, R. Kikuchi, K. Eguchi, J. Power Sources 196 (2011) 6048e6054. [27] K. Ke, A. Gunji, H. Mori, S. Tsuchida, H. Takahashi, K. Ukai, Y. Mizutani, H. Sumi, M. Yokoyama, K. Waki, Solid State Ionics 177 (2006) 541e547. [28] T. Sato, K. Kato, T. Endo, M. Shimada, React. Solids 2 (1986) 253e260. [29] T. Shishido, M. Sukenobu, H. Morioka, R. Furukawa, H. Shirahase, K. Takehira, Catal. Lett. 73 (2001) 21e26. [30] D. Li, M. Koike, J. Chen, Y. Nakagawa, K. Tomishige, Int. J. Hydrogen Energy (2014) 10959e10970. [31] M. Mokhtar, A. Inayat, J. Ofili, W. Schwieger, Appl. Clay Sci. 50 (2010) 176e181. [32] A.R. West, Basic Solid State Chemistry, second ed., John Wiley & Sons Ltd., UK, 1999, pp. 227e228. [33] K. Tomishige, Y.G. Chen, K. Fujimoto, J. Catal. 181 (1999) 91e103. [34] E. Ruckenstein, Y.H. Hu, Appl. Catal. A 183 (1999) 85e92. [35] J.T. Richardson, R.M. Scates, M.V. Twigg, Appl. Catal. A 267 (2004) 35e46. [36] J.R. Rostrup-Nielsen, T.S. Christensen, I. Dybkjaer, Stud. Surf. Sci. Catal. 113 (1998) 81e95. [37] S. Yoon, I. Kang, J. Bae, Int. J. Hydrogen Energy 33 (2008) 4780e4788. [38] R. Barfod, M. Mogensen, T. KlemensØ, A. Hagen, Y. Liu, P.V. Hendriksen, J. Electrochem. Soc. 154 (2007) B371eB378. [39] H. Sumi, T. Yamaguchi, K. Hamamoto, T. Suzuki, Y. Fujishiro, T. Matsui, K. Efuchi, Electrochem. Acta 67 (2012) 159e165. [40] S.P. Jiang, S.P.S. Badwal, J. Electrochem. Soc. 144 (1997) 3777e3784. [41] T. Hosoi, T. Yonekura, K. Sunada, K. Sasaki, J. Electrochem. Soc. 162 (2015) F136eF152.