Air plasma spray processing and electrochemical characterization of Cu–SDC coatings for use in solid oxide fuel cell anodes

Abstract

Air plasma spraying has been used to produce porous composite anodes based on Ce_0.8Sm_0.2O_1.9 (SDC) and Cu for use in solid oxide fuel cells (SOFCs). Preliminarily, a range of plasma conditions has been examined for the production of composite coatings from pre-mixed SDC and CuO powders. Plasma gas compositions were varied to obtain a range of plasma temperatures. After reduction in H₂, coatings were characterized for composition and microstructure using EDX and SEM. As a result of these tests, symmetrical sintered electrolyte-supported anode–anode cells were fabricated by air plasma spraying of the anodes, followed by in situ reduction of the CuO to Cu. Full cells deposited on SS430 porous substrates were then produced in one integrated process. Fine CuO and SDC powders have been used to produce homogeneously mixed anode coatings with higher surface area microstructures, resulting in area-specific polarization resistances of 4.8 Ω cm² in impedance tests in hydrogen at 712 °C.

Introduction

Fuel cells convert the chemical energy of a fuel, such as hydrogen, into electrical energy very efficiently on many size scales, without combustion and with little or no emission of pollutants. Solid oxide fuel cells (SOFCs) are highly efficient, entirely solid-state fuel cells that operate at high temperatures. They can be used for large-scale central power generation, for distributed generation, or for auxiliary power in transportation. When hydrogen is used to power solid oxide fuel cells, water is the only local emission produced, and the fuel cells experience relatively low degradation rates and fast electrochemical kinetics. However, hydrogen must be generated, compressed, and transported, thus creating high energy requirements and correspondingly higher costs compared to more readily available fuels. Using carbon-containing fuels, such as coal gas, methanol, ethanol, natural gas, gasoline, diesel, carbon monoxide, or renewable bio-fuels to power SOFCs may allow faster commercialization and more widespread use. Widespread power generation from carbon containing fuels using fuel cells rather than combustion engines will result in substantial reductions in greenhouse gas and acid gas emissions due to the absence of nitrogen oxides and the higher efficiency of electrochemical energy conversion.

When using hydrocarbon fuels such as methane, the fuel is typically converted through steam reforming to CO and H₂, which are then consumed electrochemically within the fuel cell, or in the case of CO, reacted with steam to form CO₂ and H₂ through a water gas shift reaction. The reforming reaction can take place externally, in a reformer placed prior to the fuel cell inlet. This configuration, however, increases the overall cost and complexity of the system. Reforming of the fuel in a high temperature SOFC system can also take place within the fuel cell itself, utilizing a reforming catalyst, commonly nickel, in the SOFC anode. This process is known as internal reforming. Internal reforming eliminates the requirement for an external reformer and therefore simplifies the balance of plant system and reduces costs. In addition to reduced costs, internal reforming is endothermic for methane, and therefore it can assist in thermal management of the cell. However, steam reforming of petroleum-based hydrocarbons such as naphtha is less endothermic, and exothermic when the temperature is lowered below 600 °C [1].

Although internal reforming simplifies the SOFC balance of plant, it is still limited in practice due to technological issues. Experiments on internal methane steam reforming have shown that the strong activity of the nickel anode results in a marked temperature reduction in a very localized reaction zone that may lead to steep thermal gradients and thermal stresses [2]. Another disadvantage of steam reforming is that the reforming reaction is endothermic and has a high temperature requirement. When the fuel is reformed by steam, high equilibrium conversions require high temperatures. For example, the equilibrium conversion of methane for a H₂O/CH₄ ratio of one at 1 bar is only 37% at 600 °C, 68% at 700 °C, and 87% at 800 °C, and for a H₂O/CH₄ ratio of two at 1 bar is 62% at 600 °C, 77% at 700 °C, and 100% at 800 °C. A higher H₂O/CH₄ ratio increases the extent of reformation of CH₄ to H₂, but it dilutes the fuel and thereby decreases the overall cell efficiency. If reforming is to be performed internally in an SOFC, the high temperature requirement for equilibrium conversion limits the choice of materials that can be used to construct the fuel cell, since 700 °C is at the working limit for common metals [3]. Operation of SOFCs at low temperatures (LT-SOFC, 500–700 °C) may reduce material costs and sealing problems and thus is of great research interest [4]. However, the low conversion efficiencies of the steam reforming reaction at low temperatures lead to a drastic reduction in cell efficiency. An additional disadvantage of steam reforming is the large amount of steam that is needed to suppress carbon deposition when utilizing hydrocarbon fuels other than methane [5]. Moreover, it has been shown that even at high steam to carbon ratios (∼>3.5) in the case of internal reforming of ethane and ethylene, the power generation of SOFCs deteriorates with time due to carbon deposition [6]. Carbon deposition reduces the cell performance by blocking the anode reaction sites [7] and damaging the microstructure of Ni in the anode.

These disadvantages of internal reforming encourage current research attempts searching for a direct oxidation mechanism to utilize hydrocarbon fuels rather than internal reforming, with particular emphasis on development of coking resistant anode materials. Direct oxidation of hydrocarbon (HC) fuels may reduce the thermal gradients created by internal reforming and improve fuel conversion efficiency, particularly with heavier hydrocarbons than methane. However, when HC fuel is directly utilized on conventional nickel-based anodes, carbon deposited on the anode material due to a secondary cracking reaction blocks the reactants from reaching the reaction sites, degrades the nickel microstructure over time, and dramatically reduces the fuel cell performance and stability. Previous studies show that nickel can be utilized in direct oxidation of methane between 500 °C and 700 °C without carbon formation, but it is unlikely with higher hydrocarbons, since the temperature window for pyrolysis will be lower and carbon formation more severe [8]. This drives research attempts to find alternatives to the use of nickel in SOFC anodes for direct oxidation of hydrocarbon fuels.

In recent studies [9], [10], [11], [12], copper has been suggested as an alternative to nickel as the electronic conductor in SOFC anodes. Copper has been used as the metal for inclusion in the anode due to its high electrical conductivity and relatively low catalytic activity for hydrocarbon cracking [9]. However, copper also has a low catalytic activity for hydrogen or hydrocarbon electrochemical oxidation, and so in order to improve the cell performance, ceria and samaria doped ceria (SDC) [10], [11] have been utilized instead of yttria stabilized zirconia (YSZ), which is commonly used as the ionic conductor in SOFC anodes. The addition of a mixed conductor such as doped ceria has been shown to play an important role in improving anode performance, through improved catalytic activity and mixed ionic–electronic conductivity, which increases reaction surface area. Carbon deposition was not observed using this anode design. However, copper–SDC anodes tend to be unstable at high temperature due to the copper's relatively low melting temperature and high surface energy, resulting in rapid sintering of the copper and loss of electrode conductivity at temperatures above 800 °C [12].

Despite the advantages of copper-based anodes for direct oxidation, they are manufactured presently using a multi-step wet ceramic technique that requires even more processing and firing steps than needed to make nickel-based anodes by wet ceramic processing, which makes them less attractive for mass production.

SOFC processing typically includes a combination of wet powder compaction steps such as tape casting or extrusion, followed by deposition by a chemical or physical process such as spray pyrolysis, screen printing, or electrochemical vapor deposition, and then densification at elevated temperatures. The complex multi-step processing procedures are time consuming and require significant capital costs, particularly when scaled up for mass production. High sintering temperatures also increase the likelihood of inter-reactions between adjacent cell layers or of metal support oxidation in metallic interconnect-supported cells.

Recent studies (e.g. [9], [10], [12]) utilizing copper as the electronic conductor in SOFC anodes use impregnation of aqueous solutions of nitrate salts into a pre-sintered porous YSZ matrix. This procedure is used because the low melting temperature of copper oxide compared to that of nickel oxide prevents co-sintering of an anode layer containing copper oxide at a sufficiently high temperature to densify the YSZ electrolyte. This procedure adds further complexity to the wet ceramic manufacturing process, especially since multiple impregnation and firing steps are needed to obtain adequate connectivity of the anode metal phase.

Plasma spray processing has also been studied recently as a processing procedure for the manufacturing of SOFCs [13], [14], [15], [16], [17]. Plasma spraying has the advantages of short processing time, material composition flexibility, and a wide range of controllable spraying parameters that can be used to adjust the properties of the coatings. Controlling the spraying and feedstock parameters during spraying allows control of the coating characteristics, creating the opportunity to vary the coating properties with thickness to obtain a functionally graded material (FGM) structure that may lead to better electrochemical performance and reduced thermal stresses [16], [17]. It also allows manufacturing of an entire cell in an integrated fabrication process [13], and spraying directly onto robust metallic interconnects, thus lowering material costs, with no requirement for sintering [14]. Nano-structured anodes have also been produced by plasma spraying of nano-agglomerated feedstock powder [18]. This approach can provide more triple phase boundaries for the hydrogen oxidation reaction and contribute to lowering polarization losses. Plasma spray processing can also be scaled up easily for rapid, automated mass production, which may allow further reduction of manufacturing costs.

To date, plasma sprayed Cu–SDC anode testing in a working cell has not previously been reported. The primary challenge in making such anodes by plasma spraying is the large difference between the melting temperature of CuO and SDC (1326 °C for copper oxide and 2600 °C for SDC), which makes it difficult to co-deposit these materials. Co-deposition of materials with a large melting temperature difference, such as tungsten carbide–cobalt coatings, has been conducted previously by plasma spraying [19]. However, that application does not require coatings with significant porosity, so both materials can be co-deposited in a high-energy plasma that fully melts both materials, resulting in dense coatings. However, the requirement for high porosity of SOFC anodes for good diffusivity makes it more difficult to co-deposit materials with a large difference between the melting temperatures in a coating having a porous structure. It is advantageous to co-spray CuO and SDC rather than Cu and SDC due to the higher melting temperature of CuO (1326 °C) and because the reduction of CuO to Cu after cell fabrication results in additional porosity.

It has been shown [20] that the addition of more catalytically active materials such as cobalt to Cu–SDC anodes is beneficial. Nevertheless, due to the higher cost of cobalt, direct oxidation anodes are likely to contain at least a portion of their structure consisting entirely of Cu–ceramic mixtures, to function primarily as a current collector/gas diffusion layer within the anode. Therefore, coarse layers containing only Cu and SDC are still of practical importance for application in direct oxidation SOFC anodes.

Cu–SDC anodes exhibit lower catalytic activity compared to conventional Ni–YSZ anodes due to the set of materials utilized in such anodes. Therefore, increased triple phase boundary surface area in an anode functional layer is important in improving anode performance. This study utilizes nano-agglomerated powders for the production of nano-structured plasma sprayed Cu–SDC coatings for use in an anode functional layer to achieve higher surface area and improved performance relative to coarser diffusion layers.

Previously, Cu–YSZ layers have been sprayed onto stainless steel coupons [21]. However, these layers have not been tested as anodes in a working cell. In addition, it is more beneficial to replace the YSZ with ceria or SDC, as ceria or SDC enhances the catalytic activity in Cu-based anodes [9]. CuO–SDC layers have also been sprayed onto ceramic substrates and the CuO has been successfully reduced to copper in hydrogen [22]; however, the layers were not tested as anodes in electrochemical cells.

We have recently applied plasma spray processing for the manufacturing of metal supported fuel cells with Cu–SDC (Ce_0.8Sm_0.2O_1.9) SOFC anode layers utilizing nano-agglomerated feedstock powders. In this article, we present the spraying and feedstock conditions developed and used to co-deposit CuO–SDC SOFC anode layers, their resulting elemental and phase compositions, their microstructures before and after reduction of CuO to Cu, and their electrochemical performance in a solid oxide fuel cell.