FeCr2O4 Nanoparticles as Anode Catalyst for Ethane Proton Conducting Fuel Cell Reactors to Coproduce Ethylene and Electricity

Ethylene and electrical power are cogenerated in fuel cell reactors with FeCr2O4 nanoparticles as anode catalyst, La0.7Sr0.3FeO3-δ (LSF) as cathode material, and BaCe0.7Zr0.1Y0.2O3-δ (BCZY) perovskite oxide as proton-conducting ceramic electrolyte. FeCr2O4, BCZY and LSF are synthesized by a sol-gel combustion method. The power density increases from 70 to 240 mW cm−2, and the ethylene yield increases from about 14.1% to 39.7% when the operating temperature of the proton-conducting fuel cell reactor increases from 650◦C to 750◦C. The FeCr2O4 anode catalyst exhibits better catalytic performance than nanosized Cr2O3 anode catalyst.


Introduction
Solid oxide fuel cells (SOFC) are promising as clean power sources having high energy conversion efficiency and excellent fuel flexibility.Typically, SOFC are based on an oxygen-ion-conducting electrolyte, which usually requires high operating temperature.In the last few years, an ever-growing interest has been directed toward protonconducting electrolytes for intermediate or low-temperature SOFC applications [1].Among these, the most investigated proton-conducting electrolytes are doped perovskite oxides (ABO 3 ), in which the A sites are occupied by alkaline earth elements such as Ba, Sr, and Ca, and B sites are occupied by tetravalent elements (usually Ce or Zr), and these perovskites are doped with trivalent ions such as Y, Nd, Sm, Yb, and In to increase the density of oxide ion vacancies required for ion conductivity.Among these electrolytes Y-doped barium cerate (BCY) stands out for its high proton conductivity but it has insufficiently high chemical stability as, for example, it reacts readily with CO 2 or SO 2 .On the other hand, Ydoped barium zirconate (BZY) shows good chemical stability but suffers from insufficient proton conductivity.When codoped with Y and Zr, barium cerate (BCZY) has balanced conductivity and chemical stability in reactive environments [2,3].
Ethylene is an important feedstock in the petrochemical industry, and its demand is expected to increase significantly in the near future [4,5].Currently, ethylene is produced mainly by steam cracking of naphtha or ethane, FCC (fluid catalytic cracking), and catalytic dehydrogenation of ethane.However, when steam cracking is operated at high temperature the process consumes much energy and has a low selectivity to ethylene and unavoidable coke formation.Recently, ethane-fueled proton-conducting ceramic fuel cell reactors have been developed to cogenerate ethylene and electrical energy with high efficiency [6][7][8][9][10][11].This reaction system has several specific advantages including high selectivity to ethylene and low impact on the environment, as there are very little or no greenhouse gas (CO 2 ) emissions.In a proton-conducting fuel cell reactor, the dehydrogenation of ethane to ethylene and hydrogen (forming protons) is conducted over the anode catalyst, while the protons are conducted through the proton conducting electrolyte to the cathode side and reacted with oxygen to form water. Electrons are conducted through an external circuit during this reaction.

Experimental
2.1.Materials Preparation and Characterization.FeCr 2 O 4 nanopowders were prepared using a sol-gel method [16,17].Cr(NO 3 ) 3 •6H 2 O, and Fe(NO 3 ) 3 •6H 2 O were first dissolved in deionized water.Subsequently, citric acid as chelating agent was added in 2 : 1 molar ratio to metal ions.The resulting solution was adjusted to about pH 8 with ammonium hydroxide, then heated on a hot plate to evaporate water at 90 • C until it formed a gel which then was dried.The dry gel was calcined at 500 Subsequently, citric acid as chelating agent and NH 4 NO 3 as oxidant agent were added in molar ratio of citric acid: total metal ions: NH 4 NO 3 of 1.5 : 1 : 3. The resulting solution was adjusted to about pH 8 with ammonium hydroxide and heated on a hot plate to evaporate water until it changed into brown foam and then ignited.After combustion, the obtained ash was calcined at 1000 • C in air for 6 h to obtain BCZY powders.
The phase structures of materials were identified using a Rigaku Rotaflex X-ray diffractometer (XRD) with Co Kα radiation.The morphology of sintered BCZY disc cross sections was determined using a Hitachi S-2700 scanning electron microscope (SEM).

Fuel Cell Reactor Fabrication.
Dense BCZY electrolyte pellets were fabricated via a pressing method.The nanosized powder was first pressed at 5 tonnes in a stainless-steel die with 1.86 cm ID to form a substrate disc, which was sintered at 1500 • C for 10 h to obtain a nonporous, dense BCZY thick film.An intimate mixture comprising similar weights of FeCr 2 O 4 , Cu powder, and BCZY was dispersed in terpineol mixed with 10% polyethylene glycol (PEG) as screen printing binder to form a paste.This paste was screen painted onto one side of the electrolyte and dried under infrared light to form the anode.A mixture of 50% LSF and 50% BCZY was pasted onto the opposite side to form the cathode.The cell then was heated at 950 • C for 4 h in air to achieve strong bonding between the electrolyte and electrode materials.

Fuel Cell Reactor System Fabrication and Test.
The fuel cell reactor setup and the testing system were as described previously [6].The fuel cell reactor was set up by securing the MEA (membrane electrode assembly) between coaxial pairs of alumina tubes and sealed using ceramic sealant, which was cured by heating in a vertical Thermolyne F79300 tubular furnace.Au paste and mesh were used to make the current collector at both electrodes.10% H 2 (balance with He) was fed into the anode chamber as the temperature was increased from room temperature to 750 • C at 1 • C/min.Then, ethane was fed into the anode chamber to replace the 10% H 2 feed.The cathode feed was oxygen.
The electrochemical performance of fuel cell reactors was measured using a Solartron 1287 electrochemical interface together with 1255B frequency response analysis instrumentation.The outlet gases from the anode chamber were analyzed using a Hewlett-Packard model HP5890 GC equipped with a packed bed column (OD: 1/8 ; length: 2 m; packing: Porapak QS) operated at 80 • C and equipped with a thermal conductivity detector.The ethane conversion and ethylene selectivity were calculated according to the previously reported method [9].

Textural and Structural
Properties.The phases of the electrode and electrolyte materials were determined using Xray diffraction (XRD).Figure 1(a) shows the XRD patterns of the La 0.7 Sr 0.3 FeO 3-δ cathode and the BaCe 0.7 Zr 0.1 Y 0.2 O 3-δ electrolyte after firing.It can be seen that all samples were a single perovskite phase, without detectable amounts of any impurities.Figure 1(b) compares the XRD spectra of fresh and reduced Fe,Cr-based anode catalyst.XRD confirmed that pure FeCr 2 O 4 phase (JCPDS Card No. 34-0140) was formed after calcination of the Fe,Cr-citrate complex gel at 500 • C for 4 h in air.When this material was reduced in H 2 at 750 • C for 4 h, some iron ions were reduced to metallic iron nanoparticles.However, the remaining ions remained present as the FeCr 2 O 4 phase.The XRD peaks of fresh FeCr 2 O 4 phase were broadened due to the small crystalline size.From the line broadening of the diffraction peak at 42 • , the average crystallite sizes of the catalysts were calculated using the Scherrer Formula (T = 0.89 λ/βcosθ).The fresh catalyst crystallites averaged only 4.2 nm.Thus, using citrate as a chelating agent provided good distribution of Fe and Cr ions throughout the dry gel, generated a large amount of gas during the citrate-nitrate gel combustion and decomposition, and so the combination of these effects resulted in the formation of very fine particles.When the catalysts were reduced in H 2 at high temperature of 750 • C for 4 h, the crystallite size increased to 33.6 nm.

Advances in Physical Chemistry
Figure 2 shows a typical scanning electron microscopy (SEM) image of the electrolyte supported fuel cell.The BCZY electrolyte was dense, thus it separates the anode and cathode feeds very well.The image also shows the typical porous microstructure of the electrodes, required to achieve good diffusion of feed to the electrochemically active triple phase boundary sites (TPB).

Catalytic Performance.
At elevated temperatures the main reaction in the anode chamber was dehydrogenation of ethane to ethylene with high selectivity, with cogeneration of electricity.Figure 3 shows the ethane conversion and ethylene selectivity at different reaction temperatures.The conversion of ethane increased from 14.6% to 29% and 43.7% whilst  the selectivity to ethylene decreased from 96.6% to 94.2% and 90.8% as the operating temperature increased from 650 • C to 700 • C and then 750 • C. The ethane conversions over this nanosized FeCr 2 O 4 catalyst were higher than the values over Cr 2 O 3 catalysts under the same conditions, described in our previous study [6].In that study, conversion of ethane increased from 8.5% to 35.3% whilst the selectivity of ethylene decreased from 98.6% to 88.2% as the operating temperature increased from 650 to 750 • C. Cr 2 O 3 is widely used as an excellent dehydrogenation catalyst, and so dehydrogenation of ethane to ethylene occurred readily [18].However, the present FeCr 2 O 4 catalyst showed even better catalytic performance than Cr 2 O 3 nanoparticles.
The differences between GC analyses of the anode feed and effluent showed the formation of small but increasing amounts of methane and hydrogen and traces of carbon oxides, as the rate of dehydrogenation of ethane increased with temperature.The hydrogen in the anode outlet was attributable to either or both of the gas phase cracking of ethane and the more rapid catalytic production of H 2 than could be accommodated by conduction of protons through the thick electrolyte membrane used in this study.Methane was produced from the thermal cracking of ethane which, as expected, was more favoured at the higher temperature, so that ethylene selectivity decreased while methane selectivity increased with temperature.The amounts of CO 2 formed were traces.The proton-conducting ceramic electrolyte membrane prevented contact with any oxygen source other than the oxide materials themselves and the very small amount of oxide conductivity in the electrolyte, as the proton ceramic electrolyte membrane conducted the protons to the cathode to react with oxygen and form water, thus providing the thermodynamic driving force and removing equilibrium limitation of the dehydration reaction.

Electrochemical Performance.
To investigate the electrochemical performance of the FeCr 2 O 4 anode under fuel cell operating conditions a single cell based on BCZY electrolyte was tested in the range from 650 • C to 750 • C using different types of fuel. Figure 4 shows the I-V and power density curves (after manual compensation) of the single cell using pure hydrogen (Figure 4 The open-circuit voltages (OCVs) of 1.04, 1.07, and 1.1 V for H 2 fuel and 0.98, 0.95, and 0.88 V for C 2 H 6 fuel at 750, 700, and 650 • C, respectively, indicated that the BCZY electrolyte was dense and impermeable.The OCV value decreased as the temperature increased for the tests with H 2 as fuel.In contrast, the trend was the inverse when C 2 H 6 was used as fuel.The OCV for C 2 H 6 fuel was lower than that when using H 2 as fuel.These trends suggest that the dissolution rate of a proton generated from C 2 H 6 fuel into the electrolyte bulk was slower than that of a proton generated from H 2 as fuel, probably as a consequence of the relative rates of generation of protons from the respective feeds.The maximum power densities were 510, 360, and 250 mW cm −2 for H 2 fuel and 240, 160, and 70 mW/cm −2 for C 2 H 6 fuel at 750, 700, and 650 • C, respectively.The high power density values indicated that the FeCr 2 O 4 material is a suitable candidate as an anode catalyst for proton-conducting SOFC.The high performance of the cell is attributable mainly to the high activity of H 2 on the FeCr 2 O 4 surface and the Fe metallic sites.Comparison of these power density values suggested that the FeCr 2 O 4 -based anode had higher catalytic activity for the electrochemical oxidation of H 2 than for that of C 2 H 6 .
In order to further compare the different contributions to the total resistance with H 2 or C 2 H 6 as fuel in the single cell test, electrochemical impedance spectroscopy measurements were performed under open-circuit conditions at different temperatures, as shown in Figure 5.The intercept with the real axis at high frequency represented the ohmic resistance (R ohm ) of the cell which, reasonably, was usually taken as the overall electrolyte resistance of the cell.The difference between the high-frequency and low-frequency intercepts with the real axis represented the total interfacial polarization resistance (R p ) of the cell.The overall electrolyte  [20].It was suggested that deposition of hydrocarbons in the anode improves the connectivity between catalyst particles and decreases the ohmic resistance [19,20].
In each of the fuel cell tests the activation polarization was very similar in value, which meant that the activation ability of FeCr 2 O 4 catalysts for H 2 and C 2 H 6 varied little with temperature.Overall, the plots (Figure 5) each consisted of a small higher-frequency depressed arc and a large lower-frequency arc, the sizes of both of which increased with decreasing temperature, as typically observed for electrochemical processes such as charge transfer or surface diffusion.However, in each case the arcs were envelopes of a series of overlapping arcs.The overlapping arcs may arise from the presence of both nano-Fe particles and FeCr 2 O 4 in the anode catalyst (Figure 1).While it is not possible to unequivocally determine this from the present data and further investigation is required, it is known that Fe particles are active catalysts for alkane dehydrogenation, including isobutane to isobutene [21].
The high values of the ohmic resistance were attributable to the overall configuration of the single cell, in particular the large thickness of the BCZY electrolyte (about 0.9 mm), suggesting that higher performance should be attainable from cells with much thinner electrolyte layers.

Conclusions
An electrolyte-supported proton-conducting SOFC fabricated using FeCr 2 O 4 as the anode catalyst, and LSF as the cathode has high performance for conversion of ethane to cogenerate ethylene and electrical power.The fuel cell reactor provides high selectivity, over 90%, for ethane dehydrogenation and ethylene yields of 14.1, 27.3 and 39.7%, while cogenerating 70, 160, and 240 mW cm −2 power densities at 650, 700, and 750 • C, respectively.The ethane conversions over this nanosized FeCr 2 O 4 catalyst are notably higher than the values over previously described Cr 2 O 3 catalysts.

Figure 3 :
Figure 3: Ethane conversion and ethylene selectivity for a FeCr 2 O 4 -BCZY|BCZY|LSF-BCZY single cell as a function of temperature.The flow rates of ethane and oxygen are each 150 mL/min.

Figure 4 :
Figure 4: I-V curves and power density output of the FeCr 2 O 4 -BCZY|BCZY|LSF-BCZY single cell at different temperatures with (a) hydrogen and (b) ethane as the fuel.
[19]re 5: Electrochemical impedance spectra of the FeCr 2 O 4 -BCZY|BCZY|LSF-BCZY single cell at different temperatures, with (a) hydrogen and (b) ethane as the fuel.resistances of the cell were 1.19, 1.42, and 1.75 Ω cm 2 for H 2 fuel and 0.94, 1.12, and 1.24 Ω cm 2 for C 2 H 6 fuel at 750, 700, and 650 • C, respectively.The corresponding polarization resistances (R E ) of the cell were 0.35, 0.48, and 0.78 Ω cm 2 for H 2 fuel and 0.91, 1.02 and 1.44 Ω cm 2 for C 2 H 6 fuel.Apparently, as the operating temperature was decreased, R ohm and R p , the two specific contributions to the total resistance, both increased dramatically.R ohm was predominant for the test using H 2 as fuel, while R ohm and R p had comparable values for the tests using C 2 H 6 as fuel.The impedance arcs for both the H 2 fuel and C 2 H 6 fuel decreased significantly with the increase in temperature.It is interesting that the impedance responses of the ethane fuel tests became clearly separated at high frequencies.This indicated that the H 2 oxidation reaction on FeCr 2 O 4 anodes was controlled by at least two distinct electrode processes.In addition, it is obvious that the R ohm in C 2 H 6 fuel is a little lower than that in H 2 fuel.A similar phenomenon was observed for the Cu-Ce-YSZ fuel cell fed by hydrogen and n-butane[19].It has been reported that the value of R ohm of Cu-Ni alloy cermets decreased with time after long-time exposure to methane