DOI : 10.1002/celc.201402017 Nickel–Copper/Gadolinium-Doped Ceria (CGO) Composite Electrocatalyst as a Protective Layer for a Solid-Oxide Fuel Cell Anode Fed with Ethanol Massimiliano Lo Faro,* [a] Rafael M. Reis, [b] Guilherme G. A. Saglietti, [b] Andre G. Sato, [b] Edson A. Ticianelli, [b] Sabrina C. Zignani, [a, b] and Antonio S. Aricò [a] 1. Introduction A major worldwide challenge is the development and wide- scale use of sustainable energy sources to reduce
of 8
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  DOI: 10.1002/celc.201402017 Nickel–Copper/Gadolinium-Doped Ceria (CGO) CompositeElectrocatalyst as a Protective Layer for a Solid-Oxide FuelCell Anode Fed with Ethanol Massimiliano Lo Faro,* [a] Rafael M. Reis, [b] Guilherme G. A. Saglietti, [b] Andre G. Sato, [b] EdsonA. Ticianelli, [b] Sabrina C. Zignani, [a, b] and Antonio S. Aric [a] 1. Introduction A major worldwide challenge is the development and wide-scale use of sustainable energy sources to reduce the effectsof global warming and overcome problems related to thefinite nature of fossil fuel reserves. Promising energy conver-sion processes, including fuel cells, are being developed tosolve these problems. The performance of such energy devicesdepends crucially on the properties of their components andmaterials.State-of-the-art solid-oxide fuel cells (SOFCs) suffer from thedirect utilisation of fuels other than H 2  and reformed methane.The occurrence of carbon deposits on the anode with conse-quent electrocatalyst poisoning in the absence of proper pre-reforming steps is a well-known drawback. [1] To address thisproblem in the short and medium term, a possible solution isto apply a protective layer to the anode to mitigate carbondeposition and anode poisoning by undesired contaminants.SOFCs are conveniently used for distributed energy genera-tion; these devices can be directly fed with biofuels.For remote electrical energy generation, in terms of energydensity, it is useful if fuels are in a liquid form, such as alcohols.Ethanol is characterised by a high volumetric energy density(6100 Whl  1 ) and low environmental impact. [2] It can be ob-tained from the fermentation of agriculture products. [3] To preliminarily assess the feasibility of a SOFC to work withbio-fuels, conditions simulating operation with practical bio-fuel streams should be investigated. In particular, bio-ethanoltypically contains a significant percentage of water; [4] more-over, bio-fuels are generally rich in CO 2 . [5] The approach used herein was to compare the behaviour of a conventional anode-supported cell (ASC) in the presence andabsence of a protective layer based on a nickel–copper gadoli-nium-doped ceria (Ni  Cu/CGO) composite. This material hasbeen already investigated in SOFCs, but in the bulk form dueto its ability to oxidise organic fuels and resilience againstforming carbon deposits. [6] These properties derive from theexcellent synergy of nickel providing electrocatalytic activityand copper ensuring resilience to cooking. Copper breaks theassemblies of nickel atoms, which activate carbon deposition.Doped ceria produces an extension of the ionic domain intothe electronic conducting electrocatalytic layer and promotesoxidation. However, such electrocatalytic systems performmore poorly than a conventional nickel–Y 2 O 3 -stabilized ZrO 2 (Ni/YSZ) anode in the presence of syngas (H 2 + CO) and it ismechanically less robust. Another drawback is due to the pro-pensity of NiO/doped-ceria systems forming a solid solution. [7] Such effects generally cause a loss in activity for the electro-chemical reaction mainly due to the loss of surface area anddecrease in the electronic and ionic conductivity. [7a,8] This effectis generally faster in the presence of a large concentration of dopants in CeO 2 . [9] However, the occurrence of an interactionbetween nickel and copper and the presence of a moderateconcentration of dopant in CeO 2  (Ce 0.9 Gd 0.1 O 2 ) can mitigatenickel inclusion into the fluorite-type lattice of CeO 2 . The risksof poor conductivity may be addressed by the utilisation of A nickel–copper alloy is prepared by using the oxalate methodand subsequent in situ reduction. The bimetallic alloy is mixedwith gadolinium-doped ceria (CGO) to obtain a composite ma-terial with mixed electronic–ionic conductivity. The catalyticand electrocatalytic properties of the composite material forethanol conversion are described. Different conditions to simu-late bio-ethanol feed operation are selected. Electrochemicaltests are performed by utilizing the Ni  Cu/CGO cermet asa barrier layer in a conventional anode-supported solid-oxidefuel cell (AS-SOFC). A comparative study between the modifiedcell and a conventional AS-SOFC without the protective layer ismade. A maximum power density of 277 mWcm  2 @0.63 V isrecorded in the presence of a mixture of ethanol–water fora cell containing the protective anodic layer compared with231 mWcm  2 @0.64 V for a bare cell under the same condi-tions. This corresponds to a 20% increase in performance. [a]  Dr. M. Lo Faro, Dr. S. C. Zignani, Dr. A. S. AricCNR-ITAE, Via Salita S. Lucia sopra Contesse 598126 Messina (Italy)E-mail:  [b]  Dr. R. M. Reis, Dr. G. G. A. Saglietti, Dr. A. G. Sato, Prof. E. A. Ticianelli,Dr. S. C. Zignani Institution USP-IQSC, Av. Trab. S¼o-carlense400 CEP 13560-970 S¼o Carlos, SP (Brazil) Supporting Information for this article is available on the WWW under manuscript is part of a Special Section devoted to the GEI 2013.  2014 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim  ChemElectroChem  2014 ,  1 , 1395–1402  1395 CHEM ELECTRO CHEM  ARTICLES  a thin film as a protective layer characterised by lowseries resistance.The studies performed herein thus focus on theutilisation of such an electrocatalytic system as a bar-rier layer against carbon formation on the conven-tional Ni/YSZ anode under operation in the presenceof ethanol. Accordingly, the basic idea is to use sucha composite as a thin protective layer for a thick con-ventional SOFC anode. 2. Results and Discussion 2.1. Physico-chemical Analysis of the Ni  Cu/CGOElectrocatalyst The crystal structure and morphology of the as-calcined Ni  Cuelectrocatalyst were investigated (Figure 1). The main phaseafter calcination was a mixed oxide of nickel and copper. Crys-tallisation mainly occurs in the cubic phase (similar toNi 0.8 Cu 0.2 O JCPDS no. 25-1049). However, the simultaneouspresence of Ni  Cu oxide in a tetragonal phase (NiCuO 2 , JCPDSno. 6-720) cannot be ruled out. Peak positions, in particular, re-flect the occurrence of the Ni  x  Cu 1   x  O cubic phase, whereaspeak intensities indicate a good match with the tetragonalNiCuO 2  phase. However, the different relative intensities of theexperimentally recorded diffraction peaks for the equimolarNi  Cu oxide and the cubic Ni 0.8 Cu 0.2 O phase in the literaturemay also be related to the different Ni/Cu ratio. No phase sep-aration into NiO and CuO was observed by XRD. However,such a result may be the consequence of large peak broaden-ing caused by the fine particles. A crystallite size of about 4–5 nm was derived from the broadening of the XRD peaks.The morphologies of calcined Ni  Cu electrocatalyst andcomposite Ni  Cu/CGO electrocatalyst after ball-milling areshown in the Figure 2. Fine particles with a typical particle sizeof about 5–10 nm were present in both materials. Large aggre-gates of Ni  Cu nanoparticles were observed in the samplemilled for 20 h. Only a low amount of Ni  Cu phase is properlydispersed on the support (Figure 2b).X-ray fluorescence (XRF) analysis (not shown) indicated thatthe final composition of the electrocatalyst was Ni/Cu/CGO25:25:50 (w/w/w). 2.2. Ex Situ Catalytic Test for the Ni  Cu/CGO Composite Catalytic experiments were performed for 10 h to investigatethe reliability of the Ni  Cu/CGO electrocatalyst for ethanol re-forming under autothermal conditions. This procedure pro-vides an ex situ screening analysis of the catalytic propertiesunder conditions close to those occurring during SOFC opera-tion. [10] Pre-treatment of the electrocatalysts at 800 8 C in thepresence of H 2 , similar to the SOFC conditioning procedure,was performed to activate the electrocatalyst. The ex situ cata-lytic tests, in addition to serving as a screening method, allowinsights into the reaction process.Figure 3 shows the outlet gas composition from ethanol au-tothermal reforming (ATR) at the Ni  Cu/CGO composite, as de-termined by gas chromatography (GC). The outlet gas was pre-ventatively dried before GC analysis. The experiment demon-strates that this electrocatalyst operates with good stabilityunder the reaction stream. However, it gives rise to a significantdifference in the composition of outlet gas relative to the ther-modynamic equilibrium, especially concerning CO 2  and CO.The calculated equilibrium composition for the external etha-nol autothermal reformer at 800 8 C is 40 mol% H 2 O, 40 mol%H 2 , 10 mol% CO 2 , and 10 mol% CO with traces of CH 4  andothers sub-products. Thus, in the equilibrium composition, theratio of H 2 /CO is 4:1, whereas the experimentally obtainedvalue is about 6:1. On the contrary, the equilibrium value of H 2 /CO 2  is 4:1, whereas the experimental value is about 3:1. Figure 1.  X-ray diffraction (XRD) spectrum of the as-calcined Ni  Cu electro-catalyst. Figure 2.  Transmission electron microscopy (TEM) micrographs of the as-calcined Ni  Cucatalyst (a) and ball-milled Ni  Cu/CGO catalyst (b). Figure 3.  Reaction yield for the main products of the ethanol ATR process at800 8 C for 10 h in the presence of the Ni  Cu/CGO electrocatalyst.  2014 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim  ChemElectroChem  2014 ,  1 , 1395–1402  1396 CHEM ELECTRO CHEM  ARTICLES  From these considerations, it is derived that part of CO is oxidised to CO 2  in comparison to the expectedequilibrium composition. This effect is ascribed to ox-idative properties of the composite electrocatalyst es-pecially due to the presence of CGO. Such character-istics should mitigate carbon deposition phenomenain the anodic fuel cell compartment.To obtain information on the thickness and cohe-sion between the various layers and interfaces of themultiple component SOFC, scanning electron micros-copy (SEM) analysis was performed on the cell crosssection (Figure 4). The bare cell is shown in Figure 4a.The three layers consist of the anode support(  150  m  m, bottom), the full density thin electrolyte(  10  m  m, middle) and the cathode layer (40–50  m  m, top). Thinfunctional layers are present at the electrode–electrolyte inter-faces (not shown in detail). Figure 4b shows the interface be-tween the supporting anode and pre-layer in the modified cell.The pre-layer thickness was about 5  m  m. The pre-layer appearsto be slightly less porous than the supported anode; this is es-sentially due to different particle sizes and relative contents of metal and ceramic phases in these layers (Ni/YSZ supportinganode 70: 30 wt%, Ni  Cu/CGO pre-layer 50:50 wt%). No de-lamination effects between the protective layer and the sup-port were observed. This indicates that the thermal expansionof the composite layers is similar. The thermal expansion coeffi-cient of Y 0.08 Zr 0.92 O 2  is similar to Ce 0.9 Gd 0.1 O 2  (11.610  6 vs.14.410  6 K   1 in the temperature range of 200–980 8 C). [11] Figure 5 shows the XRD spectrum collected in grazing anglemode of the pre-layer modified cell after sintering at 1000 8 C(Figure 5a) and subsequent reduction at 800 8 C in H 2  (Fig-ure 5b). After the firing step in air (Figure 5a), X-ray reflectionsrelated to the fluorite phase of CGO (CeO 2  JCPDS card no. 4-593) and to the cubic structure of Ni  x  Cu 1   x  O (similar toNi 0.8 Cu 0.2 O, JCPDS no. 25-1049) were present. It is observedthat further crystallisation of the Ni  Cu oxide occurs after thethermal treatment of the protective layer at 1000 8 C. In thiscase, there is a good match with the Ni 0.8 Cu 0.2 O cubic phasefrom the literature, but shifts for the peak positions, which re-flect that a different composition (Ni/Cu = 50:50 at%) is ob-served. According to the XRD spectrum, there is a significantinteraction between the nickel and copper metals in the layerdeposited on the SOFC anode and a growth of the particleswith a single cubic phase compared with the as-calcined mate-rial. The crystallite size in the Ni  x  Cu 1   x  O layer after cell manu-facturing, as determined by the Scherrer equation, was approx-imately 35 nm. After reduction in hydrogen, an equimolarNi  Cu metal alloy is obtained together with CGO. 2.3. Electrochemical Studies Polarisation and electrochemical-impedance spectroscopy (EIS)were initially recorded in the presence of H 2  feed under open-circuit voltage (OCV) and at 0.8 V. The latter operating condi-tions correspond to about 70% electrical efficiency, which isgenerally of practical interest for SOFCs. Figure 6a shows theEIS spectra collected under OCV (1.17 V for both cells) for thetwo types of cells investigated. These impedance spectra havebeen fitted by using the equivalent circuit model with seriesRQ elements. [12] Complete fitting results are reported in theSupporting Information. In the following discussion, the mostrelevant parameters, such as series and polarisation resistances,are described. A comparison of the two spectra clearly evi-denced an increase in the total resistance ( R t ), obtained fromthe intercept at low frequency, in the case of the protectivelayer coated on the anode; this was ascribed mainly to thelarger semicircle at low frequency. No large change in theseries resistance at high frequency was observed as a conse-quence of the additional layer. Figure 6b shows the EIS behav-iour at 0.8 V for the two cells. In this case, a slight increase inthe series resistance ( R s ), as determined from the intercept athigh frequency, was registered in the case of the modified cell.The value of   R s  recorded in the case of the bare cell was0.21  W cm 2 , whereas it was 0.26  W cm 2 in the case of the cell Figure 4.  SEM images of the complete bare cell (a) and the interface between the anodeand pre-layer in the modified cell (b). Scale bars: 100  m  m (a) and 20  m  m (b). Figure 5.  XRD spectrum of the pre-layer anode coating in grazing angle con-figuration: a) after calcination at 1000 8 C and b) after subsequent reductionin H 2  at 800 8 C.  2014 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim  ChemElectroChem  2014 ,  1 , 1395–1402  1397 CHEM ELECTRO CHEM  ARTICLES  modified with the pre-layer. A slight re-oxidation of the pre-layer may have occurred under these conditions. Such aneffect may be positive in terms of tolerance to carbon deposi-tion. A slightly oxidised electrocatalyst phase can reduce theanode propensity to promote carbon fibre growth, whichoccurs for a pure metallic electrocatalyst. A remarkable differ-ence in the EIS profile for the two cells was observed and as-cribed to the semicircle at very low frequencies related tomass transport constraints. This effect occurs at low frequencyin the Nyquist plot for the cell covered by the protective layer(Figure 6b). The performance of the SOFCs was evaluated bymeasuring the current–voltage ( I  – V  ) characteristics and calcu-lating the current–power curves ( I  – P  ), as shown in Figure 6c.It can be seen that, in the presence of H 2  feed, both cellsshowed the same OCV (1.17 V), whereas the modified cell per-formance was lower than that of the bare cell. Significant diffu-sion constraints were recorded for the modified cell at a currentdensity higher than 700 mAcm  2 . However, at a cell voltage of 0.8 V, the difference in performance between the two cells isvery small.A comparison of the behaviour of both cells in the presenceof diluted ethanol (1:3 in water) is shown in Figure 7. A signifi-cant decrease in the OCV down to 1 V was evident for bothcells. Slight noise in the recorded curves was possibly due tothe use of the isocratic pump for feeding ethanol. As shown inFigure 7, the modified cell had a slightly higher open-circuitpotential (0.97 V instead of 0.92 V recorded for the bare cell).Accordingly, the polarisation resistance ( R p = R t  R s ) for thespectra collected at OCV (Figure 7a) was affected by the differ-ent direct current (DC) voltage conditions, whereas the seriesresistance ( R s ) was similar (0.23  W cm 2 ) for these cells. Similar al-ternating current (AC)–impedance profiles between the twocells were observed at 0.8 V (Figure 7b). In this case, the DCvoltage is the same for the cells under investigation. Interest-ingly, the value of   R s , which is generally affected by electronicpercolation in the anode, was similar in the presence of dilutedethanol for the two cells. The  R s  value of the bare cell in-creased a slightly from 0.21  W cm 2 in hydrogen to 0.23  W cm 2 in diluted ethanol due to the less-reducing environment,whereas  R s  in the modified cell decreased from 0.26  W cm 2 inhydrogen to 0.23  W cm 2 in diluted ethanol. It was likely that, inthe case of Ni/YSZ exposed to the mixture of ethanol/water,there was a partial re-oxidation of Ni on the surface, whereas,in the case of Ni  Cu/CGO exposed to the same solution of eth-anol, a slight formation of carbon deposits may have increasedthe conductivity of the protective layer. Figure 6.  Comparison of the electrochemical experiments performed in thepresence of H 2  at 800 8 C for modified and bare cells: a) EIS spectrum at OCV,b) EIS spectrum at 0.8 V, and c) current–voltage and current–power curves. Figure 7.  Comparison of the electrochemical experiments performed in thepresence of diluted ethanol (1:3 in water) at 800 8 C for modified and barecells: a) EIS spectrum at OCV, b) EIS spectrum at 0.8 V, and c) current–voltageand current–power curves.  2014 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim  ChemElectroChem  2014 ,  1 , 1395–1402  1398 CHEM ELECTRO CHEM  ARTICLES
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks