Selection of cathode materials for forsterite supported solid oxide fuel cells – Part I: Materials interactions

Highlights

• A forsterite supported SOFC was evaluated.

• Seven different cathodes and their interaction with forsterite have been investigated.

• All cathode materials have a certain tendency to interact with forsterite.

• LSCF, LSC, PSCF, LCCF and LSFM decompose at the co-sintering temperature.

• Despite that LSF shows also strong materials interaction performance data are promising (cf. Part II).

Abstract

An inert-supported cell (ISC) was developed by Bosch with the aim of lowering the manufacturing costs of SOFCs and thus increasing their marketability and prolonging their lifetime. This ISC concept uses forsterite, a magnesium silicate doped with Zn and Ca, as support material. The cell can be described as air side inert-supported cell, since forsterite faces the air compartment.

Forsterite was chosen as a support material, as it is abundant and therefore relatively inexpensive. All functional layers are subsequently applied and co-sintered at T < 1300 °C to further reduce cell manufacturing costs.

At present, LSM is used as a cathode. However, the performance of the cell is drastically reduced due to the formation of a Zn–Mn spinel at the triple-phase boundaries during co-firing.

Based on these findings, seven different cathodes were synthesized to identify a cathode that is less reactive with forsterite. In order to investigate their reactivity, different types of samples were prepared: mixed pellets, double-layered pellets and screen-printed cathode inks on forsterite green substrates. These samples and their cross sections were then investigated by using XRD, SEM, EDX, and WDX. Their reactivity was as follows (ascending order): LSFM > LSF > LSC > PSCF > LSCF > LCCF.

Introduction

Solid oxide fuel cells (SOFCs) are highly efficient, low-emission, flexible-fuel conversion devices that play a crucial role in reducing greenhouse gases for a future renewable energy environment [1]. A huge variety of different fuel cell types have been developed over the last 20–30 years [[1], [2], [3], [4], [5], [6]]. SOFCs can be subdivided into four major cell types: electrode, electrolyte, metal and inert-supported cells. Each cell type has its own advantages and disadvantages [4]. In general, a low-cost SOFC is targeted in order to increase its marketability. The number of manufacturing steps required (e.g. shaping, casting, coating, cutting, handling, etc.), the sintering conditions and the costs of each layer therefore play a crucial role in the costs of the different cell types [7].

The anode-supported cell (ASC) gains its mechanical stability from a relatively thick, porous anode layer (>200 μm). This porous tape cast support enables a high fuel gas feed charge. A thin (5–10 μm), fine-graded functional layer consisting of NiO and 8YSZ is tape cast or screen printed on top of the porous support in order to increase the number of triple-phase boundaries. As the mechanical stability of the cell is based on the anode support layer, the electrolyte thickness is rather low at 2–20 μm. This enables low ohmic resistances and low operating temperatures [8]. However, after each manufacturing step, a separate sintering step is typically required. After applying the diffusion barrier layer (if high performance cathodes are used), the cathode, and the current collector, the total number of sintering steps amounts to three up to five, depending on the chosen materials [9]. This number of heat treatment steps can be further reduced by using, for example, the promising method of multilayer sequential tape casting as described by Menzler et al. [7].

Similar to the multi-sequential tape cast ASC, the electrolyte-supported cells (ESCs) can also be realized in just two heat treatment steps. The ESC manufacturing route starts with the tape casting of the electrolyte followed by cutting in the desired shape and a first sintering step. The electrolyte is then screen-printed with the electrodes and finalized in a second heat treatment step [10]. With a relatively thick electrolyte of ≥100 μm, the electrolyte-supported cell gains its mechanical stability from the electrolyte layer. An advantage of the ESC design is its very low leakage rate, although the desired ionic conductivity for such a thick electrolyte requires relatively high operating temperatures of ~900 °C due to the high ohmic resistance [8].

Another method of reducing material and manufacturing costs is the utilization of a porous metal support. The metal-supported cell (MSC) takes advantage of cheap metals (cheap compared to ceramic supports based on e.g. YSZ and/or Ni), mostly containing Fe and Cr, meaning that thick – and therefore expensive – anode or electrolyte layers are not required [11]. The operating temperatures of MSCs are typically rather low (≈600 °C e.g. Ceres Power [12]). However, using metal as a support poses new challenges, as the sintering of the cell must be performed in a reducing atmosphere or vacuum so that no corrosion takes place. Furthermore, interdiffusion between the metal support and the anodic nickel plays a crucial role, while oxide scale formation at high water vapor pressures might be detrimental [13,14].

Comparison of these different cell designs with respect to the scaling up of a low-cost cell serves to underline the crucial factor: Material prices [15]. Therefore, alternative materials that are cost-efficient, abundant and fulfill the respective requirements are also desired.

Upon closer inspection of the ASC, ESC, and MSC cell designs, it becomes clear that a cost-effective support material (the thickest layer in the cell), a thin electrolyte (to lower the ohmic losses and the operating temperatures), and the absence of metal to avoid cost intensive vacuum sintering steps are all part of the requirements of a low-cost SOFC [16]. With these aspects in mind, the inert-supported cell (ISC) potentially fulfils all of these criteria. It utilizes cheap porous materials as a support, which can either be applied at the fuel [17] or air side [18]. This concept has been used in the past, for instance by Rolls-Royce [17] and TOTO [19].

Within a public funded project an ISC concept based on forsterite, an abundant and cheap magnesium silicate, as support material on the air side was developed [18]. To enhance the marketability of this ISC concept, a reasonable approach would be to further reduce the number of sintering steps by co-firing all the layers in just one heat treatment step.

However, during co-firing, the different layers might form undesired reaction phases with the support material forsterite. Furthermore, co-firing has an impact on the microstructure and porosity of each layer. These changes can have a huge impact on the output power density of the SOFC [18]. Therefore, one of the focal points of the present work is the selection of a cathode material that exhibits low tendency to react with forsterite and which does not lose its catalytic activity.

To our knowledge, forsterite has not yet been used for SOFCs. In the literature, most publications concerning forsterite focus on the synthesis and the effect of different dopants on the mechanical and electrical properties. This is because forsterite is mostly used for medical purposes and for radar systems. Forsterite is a material belonging to the mineral group olivine with superior electrical and thermal insulating properties [20,21]. Its basic oxides are abundant, making forsterite (Mg₂SiO₄) a low-cost material [22].

The first synthesis methods for forsterite were developed in the mid-20th century. Different groups reported the synthesis of an olivine single crystal (forsterite) using the Czochralski process. This is a costly process due to the melting of the educts MgO and SiO₂, which requires high temperatures (T > 1800 °C) [[22], [23], [24]]. To avoid high temperatures, various synthesis approaches such as the solid-state reaction or the sol–gel synthesis can be found in the literature. For the solid-state reaction, Nurbaiti et al. [25] reported the synthesis of forsterite based on MgO and SiO₂ at 1200 °C. Tavangarian and Emadi [26] lowered the synthesis temperature to 1000 °C using talc, magnesium carbonate, and ammonium chloride as educts. Other groups [20,[27], [28], [29], [30]] reported the sol–gel synthesis of Mg₂SiO₄ or forsterite-like compositions at T < 1100 °C using different educts.

Within the R&D project forsterite doped with Zn and Ca which is in contact with LSM as cathode material is used. To date, the performance of the cell is rather low due to Zn–Mn spinel formation at the triple-phase boundaries and the impact of co-firing on the microstructure [18,31].

The aim of this study is to identify a cathode material exhibiting either no or limited reaction and interdiffusion with forsterite or which interaction phases do not influence the cell performance too much. To do so, different cathode materials were screen printed on forsterite, heat-treated at <1300 °C, and analyzed in terms of their interdiffusion and crystallographic interactions. In the second part of the paper, the cathodes are electrochemically characterized using impedance spectroscopy.