CHAPTER 1: Bilayer Electrolytes for Low Temperature and Intermediate Temperature Solid Oxide Fuel Cells – A Review

Fuel cell technology holds the potential to change the way power is generated, transmitted and utilised in our increasingly energy dependent lifestyles. State-of-the-art solid oxide fuel cells (SOFCs) operate at high temperatures (750–950 °C). Lower operating temperatures would significantly improve the economics of power generation using SOFCs. The aim of this chapter is to evaluate component materials for bilayer electrolyte-based SOFCs that could work efficiently at temperatures below 650 °C.

A fuel cell is an electrochemical device that converts the chemical energy of a fuel and an oxidant into electrical energy. The conversion is direct without the need for intermediate conversion into heat and mechanical energy, as in the case of conventional turbine/generator systems and is not limited by the Carnot cycle with up to 50% achievable electrical efficiency.¹  As fuel cells do not involve any combustion process, there is no formation of pollutants such as NO_x, SO_x, hydrocarbons and particulates. Also, fuel cells have limited moving parts (blower) and hence do not generate any noise pollution and require minimal maintenance. Due to these advantages over other energy generators, fuel cells are desired in stationary as well as in mobile applications. Among all of the different types of fuel cells, proton exchange membrane fuel cells (PEMFCs), molten carbonate fuel cells (MCFCs) and SOFCs are considered to be the most advanced and closest to wide-scale commercialisation. However, significant technological and engineering challenges remain. When running on hydrocarbon fuels in addition to external reforming, PEMFCs also require CO removal from the fuel feed as they are susceptible to CO poisoning, which results in lower conversion efficiencies. On the other hand, SOFCs run at high temperatures and ideally can internally reform any hydrocarbon fuel with high efficiencies without the need of expensive catalysts. In addition, the high-quality waste heat from the SOFCs can be utilised in cogeneration to further improve the overall efficiency – up to 85% – of the system.² 

State-of-the art SOFCs operate at temperatures between 750 and 950 °C using yttria-stabilised zirconia (YSZ) as the electrolyte material, La(Sr)MnO₃-YSZ composite as the cathode material and Ni-YSZ ceramic metal (cermet) composite as the anode material. However, the high operating temperatures of SOFCs put considerable limitations on the choice of materials for the various components and also on the lifetime of the cell. Over the years, research in industry and academia has resulted in the development of materials and fabrication techniques such that the cell components are able to perform well under extreme conditions, to withstand thermal mismatch, to be microstructurally stable and to counter interactions between adjoining components. However, this has led to high material and fabrication costs, making SOFCs less competitive with existing power generation technologies. Decreasing the operating temperatures will considerably improve the economics of power generation using SOFCs by enabling the use of cheap ferritic stainless steel alloys as the interconnect material instead of expensive alloys or ceramics, cheaper balance of plant and insulation along with increased lifetime.³  Lower operating temperatures will also result in faster start-up that is critical in certain applications. All these factors have led to a global drive towards reducing the operating temperature of SOFCs from 750 to 950 °C to low to intermediate temperatures of 500–750 °C. However, efficient operation at lower temperatures will require new electrolyte materials with higher conductivity and new electrode materials with better catalytic activity at lower temperatures. For the purpose of further discussion, the operating temperature of SOFCs is defined as low temperature (500–650 °C), intermediate temperature (650–800 °C) and high temperature (800–950 °C).

Oxide-ion conducting solid electrolytes are the backbone of SOFCs, allowing selective transport of oxide ions for electrochemical oxidation of fuel to generate electrical power. The electrolyte materials for SOFCs are oxides having a fluorite or perovskite structure because they are larger, loose-packed structures that have the ability to accept a wide range of dopants. Oxide-ion conductivity of common solid-oxide electrolytes is shown in Figure 1.1.³  The ionic conduction is possibly due to the presence of oxide-ion deficiency in the solid electrolyte, which typically is introduced by doping the host electrolyte with lower valent cations. The fluorite type unit cell is shown in Figure 1.2. CeO₂ and ThO₂ exhibit a cubic fluorite structure from room temperature up to their melting points. The open fluorite structure offers the possibility of achieving an unusually wide range of solid solutions with alkaline-earth and rare-earth oxides such as CaO and Gd₂O₃. Generally, when the cation size of the host and the guest are almost equal, the solid solution is easily formed. In the case of ZrO₂ and Bi₂O₃, the high-temperature cubic fluorite structure can be stabilised at lower temperatures by forming a solid solution and hence, the terms stabilised zirconia and stabilised bismuth oxide are used. The electrolyte in the fuel cell is exposed to both the oxidant and the fuel atmospheres, and hence, it should be thermodynamically stable under oxidising and reducing conditions. To avoid the generation of electronic conductivity, the electrolyte should have a large band gap and the dopants introduced into the lattice should not exhibit multiple oxidation states. As the oxygen oxide ion conduction is a thermally activated process, the performance of the electrolyte improves with temperature. However, there are limits in terms of operating temperatures due to material compatibility issues with other components and operational viability over extended periods of time. Reduced operating temperatures will allow the choice of a wider range of materials and significantly improve the economics of power generation using SOFCs. However, operation of SOFCs at intermediate to low temperatures requires better performance electrolyte and electrode materials.

Two approaches have been actively pursued towards reducing the operating temperatures of SOFCs. The first is finding a better electrolyte (with high ionic conductivity and high ionic transference number) and better electrode (with high catalytic activity and high mixed conductivity) materials. At the same temperature, doped ceria and stabilised bismuth oxide shows oxide-ion conductivity that is significantly higher than that of YSZ.⁴  However, doped ceria is thermodynamically unstable in the anode atmosphere of the fuel cell due to reduction from Ce⁴⁺ to Ce³⁺. The polaronic charge transfer of the mixed-valent Ce⁴⁺/Ce³⁺ results in electronic conduction inside the electrolyte, which under fuel cell conditions results in an internal short circuit and lower available power than theoretical.⁵  On the other hand, bismuth oxide-based electrolytes show the highest oxide-ion conductivity among all solid electrolytes and are of considerable interest for application in SOFCs, ion transport membranes and oxygen sensors.⁶  The high temperature δ-Bi₂O₃ has a cubic fluorite structure with 25% inherently vacant oxygen sites. The high concentration of disordered oxide ion vacancies in δ-Bi₂O₃ along with the high polarisability of Bi³⁺ cation results in oxide-ion conductivity that is one to two orders of magnitude higher than that of YSZ at comparable temperatures. Unfortunately, bismuth oxide is not thermodynamically stable under the reducing atmosphere of SOFCs and decomposes to metallic bismuth.⁷  Different combinations of bilayer and trilayer electrolytes for low temperature and intermediate temperature SOFCs have been proposed as shown in Figure 1.3. Yahiro et al.⁸  presented the concept of a stabilised zirconia/doped ceria bilayer electrolyte for SOFCs where the zirconia layer avoids the reduction of ceria and blocks the partial electronic conduction within the ceria layer. Wachsman et al.⁹  proposed a doped ceria/stabilised bismuth oxide bilayer electrolyte for low temperature operation. The primary advantage of this bilayer was that instead of using a less conductive YSZ electrolyte to protect the ceria-based electrolyte, a high conductive material (ESB) was used to block the electronic leakage current of ceria (Figure 1.4). At the same time, the ceria layer increases the oxygen partial pressure at the interface providing stability to the bismuth oxide layer. An important consideration in the design of the bilayer electrolyte is the choice of electrolyte layer facing the reducing anode atmosphere. For further discussion, the naming convention used for the anode/electrolyte-1/electrolyte-2/cathode cell was “electrolyte-1/electrolyte-2” in which electrolyte-1 faces the reducing atmosphere.

The second approach towards reducing the operating temperatures of SOFCs is by reducing the thickness of the electrolyte and attaining higher power densities at lower temperatures with electrode-supported electrolyte unit cells. De Souza et al.^10,11  demonstrated an anode-supported SOFC by co-firing the anode and electrolyte and subsequent insertion of the cathode and showed a significant improvement in performance. For the processing of thick films on porous supports, the colloidal route has advantages over other film deposition techniques such as vapour deposition (CVD/EVD) and chemical routes (sol–gel) in its simplicity, cost effectiveness, flexibility (thickness ranging from 10 to 100 µm can be deposited) and scalability.¹²  Following the advent of high-tech lithography and deposition devices, sub-micron and free-standing electrolyte micro SOFCs have also recently gained significant interest.^13–19 

Electrochemical reactions at the cathode and anode are thermally activated and operation at low temperatures result in reduced cell performance. The oxygen reduction reaction at the cathode has a larger activation energy than fuel oxidation at the anode and thus new cathode materials are required. Candidate materials for cathodes should be mixed-ionic electronic conductors (MIEC) with high catalytic activity for oxygen molecule dissociation and oxygen reduction. The material should be compatible with other cell components with respect to chemical reaction and thermal expansion coefficient (TEC). Oxides with a perovskite structure (ABO₃) have been the material of choice for the cathodes; where A is a rare-earth element and B is a transition metal (Mn, Co, Fe). On doping the A site with alkaline-earth cations (Sr²⁺, Ca²⁺), the charge compensation occurs by valence change of the transition metal cations and under certain conditions by oxide vacancy formation resulting in an MIEC.²⁰  La(Sr)MnO₃ (LSM)-YSZ composites have been the material of choice for YSZ-based SOFCs. The oxygen reduction happens at the triple phase boundaries (TPBs) between the electrocatalyst, the electrolyte and the gas phase. Furthermore, the TPBs are active only if electrons, oxide ions and oxygen gas can transport to or away from the TPBs. Hence, the performance of the composite cathode depends critically on the relative ratio, particle size, and spatial distribution of the two solid phases, so as to achieve high concentration of TPBs and percolation for both phases. In addition, the electrode microstructure should also be porous to provide for gas diffusion. It is often found that the high polarisation losses are primarily associated with inadequate ionic transport within the electrode structure. Hence, at lower temperatures LSM (a poor ionic conductor) is replaced by La(Sr)Co(Fe)O_3−δ (LSCF) or Sr(Sm)CoO₃ (SSC), which have considerable ionic conductivities.^21,22  YSZ is replaced by doped ceria, with higher oxide ion conductivity, as the electrolyte phase to develop better performance LSCF-GDC cathodes.^23,24 

LSCF has been extensively studied as a candidate cathode material for SOFCs based on ceria electrolytes.^25,26  LSCF is a good electronic conductor (e.g. La_0.6Sr_0.4Co_0.2Fe_0.8O₃ – 300 S cm⁻¹ at 750 °C) and also shows fast oxygen surface exchange with higher oxide-ion conductivity than that of LSM (e.g. 10⁻³ S cm⁻¹ at 750 °C).²⁴  LSCF also appears to be chemically stable with ceria electrolytes as the pyrochlore compound La₂Ce₂O₇ does not exist.²⁷ 

Among this family, the Fe-rich compositions are more attractive than Co-rich compositions as they have a lower thermal expansion coefficient and hence, are better matched with ceria electrolytes.²⁵  La_0.6Sr_0.4Co_0.2Fe_0.8O₃-Ce_1−xGd_xO_2−δ composite cathodes have been studied by Dusastre et al.²³  and Murray et al.²⁴  and were found to be very effective in improving the performance over single-phase La_0.6Sr_0.4Co_0.2Fe_0.8O₃ cathodes. Optimum composition was found to be between 30–50 wt% Ce_0.9Gd_0.1O_2−δ with area specific resistance (ASR) values of 0.33–0.6 Ω cm² at 600 °C. Recently, Çelikbilek et al.²⁸  studied the effect of sintering temperature on the microstructure and performance of LSCF-GDC composite cathodes. By using electrostatic spray deposition and synthesis of a nano-structure and lowering the sintering temperature to 800 °C they achieved an ASR of 0.13 Ω cm² at 600 °C.²⁸  A Ba_0.5Sr_0.5Co_0.8Fe_0.2O₃ (BSCF) based cathode was reported to exhibit low ASR due to its high catalytic activity and high oxygen diffusivity but BSCF shows fast degradation rates due to its low stability. In comparison, Sm_0.6Sr_0.4CoO₃ (SSC) showed reasonably high stability and catalytic activity. Ag-(Bi₂O₃)_0.75(Er₂₅O₃)_0.2 and Ag-BIMEVOX cermet cathodes have been also reported for low temperature applications.^29,30  Recently, pyrochlores based on bismuth ruthenate, lead ruthenate and yttrium ruthenate have been studied for application as cathodes in SOFCs.^31–33  Pyrochlore ruthenates are electrically conductive and ruthenium oxide is known to be catalytically active towards oxygen reduction and it can be expected that solid solutions containing ruthenium oxide will be beneficial as cathodes.

The anode material should be chemically compatible with the electrolyte material under reducing atmosphere and elevated temperatures. TPB, porosity, particle size and grain size distribution impact the performance of the anode significantly and must be tuned carefully.³⁴  Ni-YSZ-based cermet composites have been adopted as an anode material by most of the SOFC groups for high temperature application. Using Ni as the metal component satisfies the major requirements for the electrochemical oxidation of H₂ and CO fuels. The addition of a ceramic component makes the TEC of the composite comparable to that of the electrolyte.²¹  However, with reduced operating temperatures, the surface activity of nickel for electrochemical oxidation, as well as for coking and sulfur tolerance, reduces. In this respect, Ni-ceria cermet provides an advantage due to high activity compared to Ni-zirconia for anode reactions. Doped ceria is a MIEC under the reducing conditions at the anode, and the reaction zone is not restricted to the electrocatalyst/electrolyte interface, which results in enhanced performance. Chemical composition and microstructure play important roles in the performance of the anodes as shown by Ohara et al.³⁵  They studied the performance of Ni-samaria doped ceria cermet anodes as a function of Ni content and found that a cermet with a Ni content of around 50 vol% showed the lowest anodic polarisation (∼30 mV at 300 mA cm⁻², 800 °C).

The earliest demonstration of a bilayer electrolyte SOFC dates back to 1988 when Yahiro et al.⁸  showed the potential of using an electron blocking layer (zirconia) at the anode side of a ceria-based electrolyte to enhance the open circuit voltage (OCV) of the cell. Although the primary focus of their work was to evaluate the conductivity of various ceria-based electrolytes and the effect of cathode on the cell performance, they found that coating the ceria electrolyte on the anode side with a thin layer of zirconia deposited by RF-sputtering significantly improved OCV as shown in Figure 1.5, resulting in 33% higher power density. Following that, many researchers implemented the same approach to empower the concept of replacing the less conductive zirconia-based electrolyte with the more conductive ceria-based electrolytes and lowering the operating temperature of the cell. The initial structural design of SOFCs was constrained to electrolyte-supported cells. Such cells suffered from large ohmic losses across the electrolyte layer and thus did not show reasonable power density output. With the advent of sophisticated thick and thin film fabrication methods such as tape casting and pulsed laser deposition (PLD), respectively, the thickness of the electrolyte was significantly reduced, which resulted in much smaller ohmic losses and higher power density output. In this work electrolyte-supported SOFCs will be reviewed briefly. However, electrode-supported SOFCs will be discussed in more detail.

The results of Yahiro et al.⁸  spurred several research activities based on zirconia/ceria bilayer electrolytes for application in intermediate temperature solid oxide fuel cells (IT-SOFCs). There were two key considerations for the development work: firstly, in the design of the zirconia/ceria bilayer, whether the ceria layer should be on the anode side or on the cathode side and the minimum thicknesses of the zirconia layer to block electronic conduction in the ceria layer. Ceria on the cathode side allows the possibility of using cobalt containing cathodes (LSC, LSCF, SSC, BSCF), with higher performance at lower temperatures compared to LSM, which tend to react with zirconia. Secondly, fabrication of the bilayer to match the shrinkage rates of zirconia and ceria layers and to avoid formation of resistive secondary phases at the interface during sintering.

Lim and Virkar³⁶  measured the chemical potential of oxygen in YSZ/GDC bilayer electrolyte SOFCs and found that the position of the YSZ layer, whether on the anode side or the cathode side, affected the chemical potential profile and the cell performance. The thickness of the GDC and YSZ electrolyte was 45 and 5 µm, respectively, and the electric potential inside the GDC electrolyte was measured using Pt-embedded electrodes. LSM-YSZ and LSC-GDC were used as the cathode for the bilayer cell with GDC at the anode side and the cathode side, respectively. The embedded probe measurements indicated that with GDC at the cathode, electric potential inside GDC was a function of position and load; oxygen partial pressure at the GDC/cathode interface showed a small change (Figure 1.6). However, with GDC at the anode, electric potential inside GDC was independent of position and load and a sharp change in oxygen partial pressure was observed at the YSZ/cathode interface. In terms of performance, the cell with GDC at the anode showed >3 times higher maximum power density (MPD) than the cell with GDC at the cathode (0.52 vs. 0.14 W cm⁻² at 650 °C). However, a majority of the work that followed focused on zirconia/ceria bilayer electrolytes with ceria on the cathode side as it allowed the use of cobalt containing cathodes, which improve performance at lower temperatures. In another study, Kwon et al.³⁷  suggested that a YSZ/GDC thickness ratio of about 2 × 10⁻⁷ is sufficient to block the electronic leakage current in the GDC electrolyte. However, fabrication limitation prevents the development of bilayers with such an ultra-thin YSZ layer. Cell configuration, processing and the performance of zirconia/ceria bilayer-based IT-SOFCs are summarised in Table 1.1.

An electrolyte-supported YSZ/YDC bilayer electrolyte cell was developed by Kim et al.³⁸  using a thick 1.6 mm YDC pellet made by the sintering of co-precipitated powders at 1600 °C. The YSZ layer was deposited by spin coating and then was sintered at 1400 °C. Compared to a monolayer YDC cell, the YSZ/YDC bilayer cell increased the OCV slightly by 0.05 V between 700 and 1050 °C. However, the MPD of the YSZ/YDC bilayer cell at 800 °C (0.12 W cm⁻²) was comparable to that of a monolayer YSZ cell at 1000 °C (0.14 W cm⁻²) due to the higher conductivity of YDC. Based on the model developed by Virkar,³⁹  p_O2 at the YSZ/YDC interface was estimated to be two orders of magnitude higher than the stability limit of ceria at 800 °C.

Hsieh et al.⁴⁰  fabricated and tested electrolyte-supported micro-tubular SOFCs based on a ScSZ/GDC bilayer electrolyte. GDC micro-tubes were extruded and pre-sintered at 1100 °C. A NiO-GDC anode (60 : 40) and LSCF-GDC cathode (80 : 20) were deposited on the inner surface and outer surface of the micro-tubes and sintered at 1400 °C and 1100 °C, respectively. For the bilayer cell, ScSZ was deposited prior to the anode deposition and sintered at 1400 °C. The thickness of the sintered ScSZ and GDC was 8 and 285 µm, respectively. The ScSZ layer in the bilayer cell significantly improved the OCV from 0.92 to 1.06 V at 650 °C by blocking the electronic conduction in GDC, but decreased MPD from 0.20 to 0.16 W cm⁻² due to high interfacial resistance across ScSZ/GDC and anode/ScSZ interfaces. At 800 °C, the bilayer cell showed a better performance compared to the monolayer GDC cell and the MPD of the cell increased from 0.36 W cm⁻² to 0.42 W cm⁻².

Electrolyte-supported YSZ/GDC bilayer electrolyte SOFCs with an ultra-thin (28–150 nm) YSZ film was fabricated by Jee et al.⁴¹  using atomic layer deposition (ALD). A thick 0.43 mm GDC electrolyte support was fabricated by pressing and sintering, and ultra-thin YSZ film was deposited by ALD at 230 °C. The NiO-GDC anode and LSCF-GDC cathode were sintered at 1200 °C and 1400 °C, respectively. High sintering temperatures resulted in significant inter-diffusion of cerium into the YSZ layer as observed by XPS. Furthermore, morphological changes in the YSZ thin film were observed – YSZ grains coalesced at the GDC grain boundaries, uncovering the underlying GDC layer. In an electrochemical test with H₂ as fuel, OCV was found to be stable only for YSZ/GDC bilayer SOFCs with a 150 nm YSZ protection layer, suggesting that an ultra-thin YSZ protective layer may not be practically feasible.

An electrolyte-supported GDC/YSZ/GDC trilayer electrolyte-based SOFC was developed by Timurkutluk et al. to suppress the leakage current between the anode and cathode.⁴²  A thin layer of doped zirconia (YSZ, ScSZ, ScCeSZ) was laminated between two thick GDC layers made by tape casting. A NiO-GDC (60 : 40) anode and LSCF-GDC (50 : 50) were then screen printed. A doped zirconia layer improved the OCV of the trilayer cells at 550–650 °C but the MPD was found to be lower than that of the monolayer GDC cell due to higher resistance. At 650 °C, the MPD of the monolayer GDC cell was ∼0.25 W cm⁻², while the trilayer cell with ScCeSZ showed ∼25% lower MPD (0.19 W cm⁻²).

In another study, Wu et al.⁴³  measured the conductivity of a YSZ/SDC bilayer consisting of a thin 250 nm YSZ film deposited on a SDC substrate by electron-beam evaporation at 200 °C. YSZ/SDC bilayer samples were sintered at 1000 °C and showed lower conductivity than single crystal YSZ, as presented in Figure 1.7. TEM-EDS analysis indicated the formation of a solid solution layer at the YSZ/SDC interface due to interdiffusion of cerium and zirconium resulting in lower ionic conductivity (Figure 1.8).

Ceria/bismuth oxide bilayer electrolyte-based SOFCs were first proposed by Wachsman et al. to combine two low-temperature high-conductivity electrolytes, which cannot operate efficiently as a single electrolyte in a SOFC environment.⁹  With ceria on the anode side and bismuth oxide on the cathode side, a functional bilayer electrolyte-based SOFC was demonstrated. It was suggested that the partial electronic conduction in the ceria layer could be blocked by the bismuth oxide layer, while the ceria layer increased the oxygen partial pressure at the interface above the decomposition potential of bismuth oxide. Practical devices require the development of cathodes compatible with stabilised bismuth oxide as platinum and the conventional LSCF cathode for IT-SOFCs show significant reactivity with bismuth oxide. This resulted in the development of new alternative cathodes such as bismuth ruthenate-ESB^44,45  and LSM-ESB⁴⁶  composites for low temperature operation. The initial concept demonstration of ceria/bismuth oxide bilayer cells was done with electrolyte-supported cells using noble metal (gold, silver) electrodes, which was then followed with the development of anode-supported cells with thick/thin film bilayer electrolytes showing high performance at low temperatures. Low temperature operation of SOFCs allows the possibility to use stable nanostructured electrodes (without undue sintering during operation) with higher surface-area-to-volume ratio and higher catalytic activity. By shifting the effective particle diameter of the catalytic phase from the micro (10⁻⁶) to the nano (10⁻⁹) regime, TPB density [(10⁻⁶/10⁻⁹)³ = 10⁹] is dramatically increased and can be used to compensate for the exponentially decreasing area-specific electrode reaction rates (activation polarisation) with decreasing temperature.⁴⁷  Some of these advances have been employed to improve the performance of bilayer electrolyte SOFCs. Recent work on ceria/bismuth oxide bilayers has focused on utilising new bismuth oxide compositions such as double-doped cubic bismuth oxides (DWSB, DGSB) with high oxide ion conductivity. Cell configuration, processing and performance of ceria/bismuth oxide bilayer-based LT-SOFCs are summarised in Table 1.2.

Wachsman et al.^9,48  used different stabilised bismuth oxides (5–60 µm⁴⁸ ), including (Y₂O₃)_0.2(Bi₂O₃)_0.8 and (Er₂O₃)_0.2(Bi₂O₃)_0.8, to block the electronic transport across thick self-supporting pellets (∼0.8–1 mm) of doped ceria – (CeO₂)_0.9(MO_1.5)_0.1 where M is Y, Sm and Gd. The samples were coated with Au and Pt paste on the cathode and anode side, respectively, and tested with humidified H₂ and O₂ as fuel and oxidant, respectively. The results of the experiments indicated that all bilayer cells have higher OCV compared to the cell with the single layer ceria cell of the same composition. The higher OCV of the bilayer cells was suggested to be due to the electron blocking capability of the bismuth oxide layer, which was stable due to high interfacial p_O2 provided by the ceria layer. SDC cells (both monolayer and bilayer electrolyte) showed higher OCV compared to YDC cells at all temperatures, due to the higher transference number (t_i) of SDC compared to YDC. For the SDC/ESB bilayer cell, increasing the density of the ESB layer increased the OCV and at 500 °C, the OCV was higher by 90–160 mV depending on the temperature compared to the monolayer SDC cell.⁴⁸  The effect of ESB thickness on the performance of the bilayer electrolyte cells was studied using GDC electrolyte and it was found that thick ESB resulted in lower OCV compared to thin ESB, possibly due to the lower interfacial p_O2 in the case of a thick ESB cell (Figure 1.9). The SDC/ESB cell was operated under an open circuit condition for 1400 h and was found to be stable. Moreover, it was found that the addition of ESB not only increased the OCV but also decreased the cell area specific resistance (ASR), resulting in 33% higher power density for the SDC/ESB bilayer cell.⁴⁸  This result favourably compared a ceria/bismuth oxide bilayer to a zirconia/ceria bilayer in which resistive secondary phases form leading to higher resistance and lower performance.⁴³ 

Following the initial promising results with ceria/bismuth oxide bilayer cells, Park et al.⁴⁹  fabricated and studied the electrical conductivity of a SDC/ESB bilayer electrolyte. SDC powder was produced by conventional solid-state reaction and pressed into ∼2 mm thick pellets. The ESB layer was made by two methods: (1) PLD for thin film, and (2) dip coating for thick film. For a PLD target, ESB powder was made by the solid-state route, while for dip coating, ESB powder was fabricated by a citrate process. The ESB powder made by the citrate process required lower calcination temperature (650 °C) compared to the solid-state route (800 °C) and enabled densification of the dip-coated film. The thickness of the ESB layer was controlled by the deposition time and solid loading of the slurry for PLD and dip-coating method, respectively. The XRD pattern of the SDC/ESB bilayer showed only peaks from the SDC and ESB phases and no other secondary phase was identified. An absence of reaction between SDC and ESB was also confirmed through EDX scans across the bilayer interface. Conductivity measurements showed that the SDC/ESB bilayer electrolyte has a slightly higher conductivity compared to the SDC monolayer electrolyte, which was likely due to the formation of a more conductive solid solution phase at the SDC-ESB interface. Further analysis as shown in Figure 1.10 indicates that the grain boundary conductivity of the SDC/ESB bilayer electrolyte was higher than the SDC monolayer electrolyte, which was attributed to the scavenging effect of bismuth oxide for silica-rich impurities at the grain boundaries in SDC resulting in a lower activation energy for oxygen transport through the bilayer electrolyte.

In another study, Park and Wachsman⁵⁰  looked at the effect of the relative thickness of bilayer electrolyte layers and anode oxygen partial pressure (p_O2) on the OCV and t_i of SDC/ESB bilayer SOFCs at 600–800 °C. The thin ESB layer (1.5 µm) was made by PLD and the thick ESB (9–30 µm) layer was made by dip coating in ESB suspension. At high anode p_O2 levels, the t_i of the SDC monolayer electrolyte and ESB/SDC bilayer electrolyte was almost identical. However, as the anode p_O2 was decreased, the bilayer electrolyte showed a higher t_i compared to the SDC monolayer electrolyte and the increase in t_i with the bilayer electrolyte was more pronounced at higher temperatures. The effect of relative thickness of the two layers was also examined and it was found that t_i and OCV increased with an increase in the relative thickness of ESB/SDC, showing a sigmoidal (s-shaped) relationship. The t_i of the bilayer electrolyte reached ∼0.9 at ESB/SDC relative thickness of 0.03. The performance of the Au and Ag-ESB cathode was also studied and it was found that the Ag-ESB cathode has a lower overpotential resulting in a higher OCV. In another study,⁵¹  Park et al. looked at the application of SDC/ESB bilayer electrolytes for ceramic oxygen generators.

Results with electrolyte-supported zirconia/ceria bilayer electrolyte cells were not promising in terms of power density, which led researchers to evaluate anode-supported configurations using both thick and thin film processing for the fabrication of zirconia and ceria layers. Some studies with cathode-supported and metal-supported configurations have also been reported but the performance was found to be lower compared to anode-supported cells.

To understand the effect of a zirconia/ceria bilayer on the ionic conductivity of the electrolyte, Brahim et al.⁵²  measured the conductivity of the YSZ/GDC bilayer deposited by DC magnetron sputtering on a LSM cathode. Five bilayer samples with different thickness ratios of YSZ and GDC layers were studied. The results of impedance spectroscopy studies showed that the bilayer electrolyte resistance increased with an increase in thickness of both the YSZ and GDC layers and, as expected, the effect of the YSZ thickness on bilayer resistance was more pronounced. Overall, the effect of the thickness ratio was found to be more important than the electrolyte thickness and the bilayer sample with a YSZ/GDC thickness ratio of 1 : 2.2 showed lower conductivity than all the other samples with ratio of 1 : 5 or lower but with higher thicknesses.

Liu et al. fabricated Ni-YSZ anode-supported YSZ/GDC bilayer SOFCs with a LSCF-GDC composite cathode. A Ni-YSZ cermet anode was made by powder pressing and pre-sintering at 900 °C. Spray coating was used to deposit YSZ and GDC layers of 3 and 7 µm thickness, respectively. The OCV of the bilayer cell was 1.05 V at 800 °C, which was significantly higher than that of the 10 µm monolayer GDC cell (0.59 V), as shown in Figure 1.11; suggesting that the YSZ layer significantly suppressed the electronic leakage in GDC. However, the MPD of the bilayer cell was low (0.68 W cm⁻² at 800 °C), which was attributed to gas diffusion limitation inside the Ni-YSZ anode.⁵³ 

Wang et al. fabricated and tested YSZ/GDC bilayer electrolyte SOFCs using a BSCF (Ba_0.5Sr_0.5Co_0.8Fe_0.2O₃) cathode.⁵⁴  A NiO-YSZ (50 : 50) anode substrate was prepared by pressing and pre-sintering at 1000 °C. AYSZ layer was screen printed and sintered at 1400 °C, followed by screen printing and sintering at 1400 °C of the SDC layer. The thickness of the YSZ and SDC layers was 2–8 and 12 µm, respectively. A reaction between the YSZ and SDC layers after sintering at 1400 °C was confirmed by a shift in XRD peak positions and EDX showed a significant diffusion of cerium and samarium into the YSZ layer. Moreover, conductivity of the YSZ/SDC pellet was found to be seven times lower than that of the SDC pellet. At 750 °C, the OCV of a bilayer cell with 2 µm thick YSZ was 0.88 V, while cells with 5–8 µm thick YSZ showed ∼1.05 V, suggesting that the YSZ layer below 5 µm may not be able to completely block SDC leakage current at higher temperatures. The MPD of the bilayer cell with 5 µm YSZ was 0.82 W cm⁻² at 750 °C, but after only two thermal cycles non-ohmic ASR increased resulting in 35% lower MPD.

Zhang et al.⁵⁵  used tape casting to fabricate thin ScSZ/SDC bilayer electrolyte SOFCs. NiO-YSZ support was fabricated by tape casting followed by screen printing of a NiO-SSZ (56 : 44) active anode layer. The thickness of the ScSZ and SDC electrolyte layers after co-firing at 1400 °C was 2 and 4 µm thickness, respectively. As shown in Figure 1.12, the OCV of the bilayer cell was 1.03 V at 650 °C indicating that the ScSZ layer successfully blocked the electronic leakage current in the SDC layer. The MPD of the ScSZ/SDC bilayer cell was 0.54 and 1.8 W cm⁻² at 650 and 800 °C, respectively. Further analysis suggested that a 1 µm thick interdiffusion layer at the ScSZ/SDC interface accounted for more than 90% of the electrolyte ohmic ASR, as the conductivity of the interfacial layer was 2–3 times lower than ScSZ and 1–2 orders of magnitude lower than SDC. Long-term stability studies showed that MPD decreased with time largely due to an increase in non-ohmic ASR at an approximately constant rate, following an initial 6–7 h decrease. It was suggested that cathode performance degraded over time due to a coarsening of the cathode microstructure and diffusion of strontium away from the SSC cathode.

Thin film ScSZ/SDC bilayer electrolyte-based SOFCs were developed by Yang et al.⁵⁶  using PLD. PLD process parameters for the two layers were optimised to achieve a cubic phase for ScSZ and SDC at 400 and 200 °C, respectively, and no reaction was observed because of the low processing temperatures. With an increase in substrate temperature, the density of the ScSZ layer was found to decrease while the density of the SDC layer was found to increase due to the different band gaps vis-à-vis the laser photons. For the fabrication of full cells, a ScSZ film of 1 µm thickness was first deposited at 200–400 °C on NiO-SDC substrate, followed by deposition of SDC film of 6–7 µm thickness at 400–600 °C. The OCV of the ScSZ/SDC bilayer cell with a SSC-SDC (75 : 25) cathode was 1.04 V and the MPD was 0.5 and 1.6 W cm⁻² at 550 and 700 °C, respectively.

Solovyev et al.⁵⁷  used bipolar reactive magnetron sputtering to fabricate YSZ/GDC bilayer electrolyte SOFCs with a NiO-YSZ anode and LSCF-GDC cathode. Monolayer YSZ and GDC cells were also fabricated and tested. It was shown that a bipolar reactive magnetron has almost a three times higher deposition rate than DC sputtering. The deposited YSZ and GDC layers showed a cubic structure without the presence of any secondary phase. Monolayer YSZ cell studies indicated that below 2 µm YSZ thin film did not provide sufficient gas tightness and OCV decreased significantly. Among the different monolayer YSZ cells, a 4 µm thick YSZ cell showed the highest MPD of 0.25 and 0.58 W cm⁻² at 650 and 800 °C, respectively. In comparison, a 2.8 µm thick monolayer GDC cell showed a higher MPD of 0.54 mW cm⁻² at 650 °C. On the other hand, a bilayer cell with YSZ and GDC thickness of 4 and 1.5 µm, respectively, showed an OCV of 1.13 V and MPD of 1.25 W cm⁻² at 800 °C. Unfortunately, the performance of the bilayer samples below 750 °C was not reported in this study.

Lu et al.⁵⁸  tested YSZ/SDC bilayer electrolyte SOFCs made by PLD. A NiO-YSZ anode substrate was prepared by tape casting and lamination. A YSZ and SDC thin film of 0.5–1 and 5 µm, respectively, were deposited at 1000 °C. A La_0.6Sr_0.4Co_0.2Fe_0.8O₃ layer was screen printed and then coated with a La_0.8Sr_0.2CoO₃ layer to act as a cathode functional layer and cathode current collector, respectively. At 650 °C, the OCV and MPD of the YSZ/SDC bilayer cell was 0.95 and 0.99 V and 0.95 and 1.37 W cm⁻² in air and oxygen at the cathode, respectively. The power density improvement in oxygen was due to the significantly lower non-ohmic ASR of the cell. The stability of the YSZ/SDC cell was tested at 1 A cm⁻² and 650 °C for 400 h and only a slight decrease in cell voltage was observed.

The effect of the anode functional layer (AFL) on YSZ/SDC bilayer electrolyte-based SOFCs was studied by Qian et al.⁵⁹  Both electrolyte films were deposited on a pressed NiO-YSZ (60 : 40) anode by PLD at 600 °C, followed by a post-anneal at 800 °C. To achieve smoother, denser electrolyte films, NiO-YSZ (50 : 50) AFL of 12 µm thickness was deposited on the anode substrate by drop-coating of AFL slurry. The thickness of the YSZ and SDC films was 2 and 6 µm, respectively. Using a SSC-SDC (70 : 30) cathode, the OCV of the YSZ/SDC cell without AFL was 0.85 V, which was higher than that of the monolayer SDC cell (0.78 V); the MPD of the cell without AFL was 0.75 W cm⁻². In comparison, the YSZ/SDC bilayer cell with AFL showed higher OCV and MPD of 1.01 V and 0.9 W cm⁻² at 650 °C, respectively. The increase in MPD was attributed to the lower polarisation resistance of the cell because of AFL addition. The stability of the YSZ/SDC bilayer cell was tested at 600 °C for 50 h and the OCV slightly dropped by 0.98%.

Myung et al.¹³  fabricated and tested ultra-thin YSZ/GDC bilayer electrolyte SOFCs by PLD. NiO-YSZ AFL (2 µm thick) was deposited by PLD to modify the surface roughness of the NiO-YSZ anode substrate; a LSC cathode was also deposited by PLD. The thickness of the YSZ film varied between 0 and 0.2 µm, while the thickness of the GDC film was 1.0 µm. YSZ grains in the thin film demonstrated a homoepitaxial relationship with YSZ grains in the anode and a heteroepitaxial relationship with GDC grains. The YSZ/GDC bilayer electrolyte was found to increase OCV from 0.6 V to over 1 V at 600 °C (Figure 1.13). Furthermore, the MPD of bilayer cells was significantly higher (>1 W cm⁻² at 600 °C) compared to monolayer GDC cells (0.38 W cm⁻² at 600 °C) as shown in Figure 1.14. It was noted that roughness of the anode substrate played an important role in the reproducibility of OCV among cells with the same thickness.

Ultra-thin YSZ/GDC bilayer electrolyte-based SOFCs using a chemical solution deposition (CSD) technique were fabricated by Oh et al.¹⁴  A nanoparticle-based solution with 5% volume loading was spin-coated on NiO-YSZ substrate (56 : 44). The CSD required that the film thickness be greater than both the grain size of the film and pore size of the substrate but allowed the fabrication of thin films of YSZ (0.1 µm) and GDC (0.1 µm) at substantially lower sintering temperatures. It was noted that the addition of slow-sintering nanoparticles to the chemical solution was effective in suppressing the generation of macro-defects and directional development of the film microstructure, which typically is observed in CSD due to constrained sintering. Full cells were made with a LSCF-GDC cathode. The OCV of the YSZ/GDC bilayer cells was > 1 V at 500–650 °C and the MPD was 1.3 W cm⁻² at 650 °C. The authors claimed to have developed the thinnest bilayer SOFCs and suggested that CSD can effectively replace vacuum deposition of thin film SOFCs.

Cathode-supported micro-tubular design was used to evaluate performance of a ScSZ/GDC bilayer electrolyte by Yamaguchi et al. in multiple studies, but unfortunately the performance was comparatively lower than that of anode-supported designs.^60–62  LSM cathode support was fabricated by extrusion and a ScSZ/GDC bilayer electrolyte was fabricated by dip coating in GDC and ScSZ slurries and co-firing at 1300 °C. A NiO-GDC anode layer was also made by dip coating. The final sintered bilayer electrolyte cell consisted of 15 µm ScSZ and 20 µm GDC thick layers. The OCV of the bilayer cell was 1.1 V at 500 °C and the MPD was 0.07 and 0.20 W cm⁻² at 600 and 750 °C, respectively. The electrode ASR accounted for the low power density and a LSM-GDC interlayer was added at the GDC/LSM interface⁶¹  that reduced the electrode ASR from 20 to 5 Ω cm² at 600 °C and the MPD of the cell was increased from 0.23 to 0.38 W cm⁻² at 750 °C. The tubular design allowed rapid heating and cooling cycles without performance degradation after five cycles. In another study, the anode composition was optimised and NiO/GDC at 50 : 50 ratio was found to show six times higher MPD than that at 70 : 30 ratio,⁶²  due to lower mass transport limitations.

Metal-supported ScSZ/SDC bilayer electrolyte SOFCs with a SSC-SDC cathode were developed by Hui et al.⁶³  Stainless steel (SS430 grade) was used as the metallic support and a NiO-SDC anode was deposited by spin-coating. ScSZ (2 µm) and SDC (20 µm) layers were deposited by PLD and sintered at a comparatively low temperature of 850 °C to minimise oxidation of the metallic support. The OCV of the ScSZ/SDC bilayer cell was found to be 1.00–1.03 V at 400–600 °C. The MPD was low (0.16 W cm⁻² at 600 °C) but the performance at 500 °C was stable for over 11 days. It was found that most of the voltage loss was due to the ohmic resistance of the cell.

A ceria/zirconia bilayer electrolyte-based micro-SOFC with ceria on the anode side was developed by Lee et al.⁶⁴  Stainless steel (STS-434L) was used as a substrate to deposit a nano-porous Ni layer by screen printing and sintering in dry hydrogen at 575 °C. A NiO-GDC (60 : 40) anode, GDC, YSZ and La_0.7Sr_0.3CoO₃ layers were deposited by PLD. The OCV and MPD of the cell was 0.91 V and 0.03 W cm⁻² at 450 °C and remained stable for over 100 h of testing. It was noted that the ohmic resistance of the cell remained constant, indicating that delamination between the cell components did not occur. Non-ohmic resistance of the cell significantly decreased during the first 100 h and then remained stable, which was attributed to slow kinetics of NiO reduction at 450 °C. The authors claimed their work to be the longest lasting operation of a micro-SOFC based on a STS-NiO metallic support.

Ceres Power is working on the commercialisation of metal-supported trilayer electrolyte-based SOFC technology.⁶⁵  The company's ‘Steel Cell’ is based on a thick film GDC layer that provides mechanical integrity and gas tightness, a thin film electron blocking layer (most likely zirconia based) to block leakage current in GDC, and a thin film GDC layer to avoid reaction between the electron blocking layer (which would be necessary if that electron blocking layer is zirconia based as described in the section below) and a perovskite-based cathode (Figure 1.15). The layers including a nickel-ceria anode are deposited on a ferritic stainless-steel substrate using conventional ceramic processing techniques. Additionally, metal support provides improved mechanical robustness compared to the anode-supported planar design. The OCV of the trilayer electrolyte-based cell was improved to 1.17 V by reduction in electrolyte defects and other improvements. Over the years, power per cell was increased from 10 W in 2010 to 20 W in 2015 largely by the optimisation in materials and mechanical design within the stack as shown in Figure 1.16.⁶⁵  Lastly, the metal-supported design showed encouraging long-term stability and thermal cycling results.

The primary purpose of ceria-based bilayer electrolytes paired with zirconia is to protect the GDC layer from reduction and enhance the OCV of the cell. However, zirconia/ceria-based bilayers have been used for other reasons. Within the structure of zirconia-based bilayer cells, the thinner GDC layer is normally inserted as a buffer layer to limit the diffusion of strontium and cobalt from the cobalt-containing cathode such as LSC and LSCF into the YSZ, and thus preventing the formation of resistive phases. However, YSZ and GDC are known to have chemical incompatibility. During the conventional high temperature (HT) sintering of the electrolyte at 1450 °C, resistive compounds with two orders of magnitude lower conductivity than YSZ form and surface porosity as a result of the Kirkendhal effect occurs. Hence, the major focus of different works in this category is to develop methods for low temperature fabrication of zirconia/ceria electrolytes such as using the effect of additives for lowering the sintering temperature of the bilayer electrolyte,^66,14,67  using sputtering for deposition of electrolytes,^68,69,15  spray pyrolysis,⁷⁰  PLD^71,72  and electron beam physical vapour deposition (EB-PVD).¹⁶ 

Due to the relatively low conductivity of YSZ and sluggish oxygen reduction reaction kinetics at lower temperatures, many works targeting 500 °C and lower temperature SOFCs are focused on the fabrication of ultra-thin electrolytes and using noble metal electrodes such as platinum. In such zirconia-based YSZ/GDC bilayer electrolyte SOFCs, a GDC layer is usually implemented to reduce electrode overpotential and achieve high OCV despite the extremely thin (few hundreds nm) electrolytes. It is well known that oxide ion conductivity of the electrolytes adjacent to the cathode and anode impacts the non-ohmic polarisation and a faster oxide ion conductor such as doped ceria compared to zirconia can enhance the performance of the cell, particularly at lower temperature. Although such thin bilayer electrolyte micro-SOFCs showed high performance at lower temperatures (1.3 W cm⁻² at 450 °C¹⁹ ), they are typically fabricated using semi-conductor processing equipment such as plasma-enhanced atomic layer deposition (PEALD),¹⁷  aerosol-assisted chemical vapour deposition (AA-CVD),¹⁸  and a combination of ALD and PLD.¹⁹  They also require noble metals such as platinum as their electrodes and thus are very expensive to fabricate. Following the same rationale, a few works in the literature showed the application of a thin GDC layer at the anode side of a zirconia-based SOFC to reduce the anodic overpotential of the cell.^73–75 

Some researchers have looked at improving the performance of YSZ-based SOFCs for intermediate temperature operation by incorporating a bismuth oxide layer on the cathode side. Joh et al.⁷⁶  fabricated YSZ/ESB bilayer electrolyte SOFCs using a LSM-ESB cathode. A NiO-YSZ anode (65 : 35) was made by tape casting and pre-sintered at 900 °C. Both YSZ and ESB electrolytes were deposited via colloidal drop coating and sintered at 1400 and 800 °C, respectively. The LSM-ESB cathode (50 : 50) was screen printed and sintered at 800 °C. No reaction between the YSZ and ESB layers was observed after heat treatment at 800 °C and after testing of unit cells by EDS analysis. The OCV of the bilayer cell with a 7 µm YSZ layer and 5 µm ESB layer was >1.10 V at all temperatures due to the unity transference number (t_i) of the YSZ electrolyte and at 650 °C the MPD of the bilayer cell was 1.62 W cm⁻². The MPD of the bilayer cell was 2.08 W cm⁻² at 700 °C, which compared favourably to the monolayer 7 µm YSZ cell which showed a MPD of 0.86 W cm⁻² (Figure 1.17). The significant improvement in the power density of the bilayer cell was attributed to a faster oxygen reduction reaction at the cathode/electrolyte interface in the presence of the ESB layer. The enhanced electrode kinetics was confirmed by the reduction in cathodic overpotential and a higher exchange current density.

In another work, Wang et al.⁷⁷  looked at YSZ/21NbSB bilayer electrolyte SOFCs made by RF magnetron sputtering. Ni-SDC and silver were used as the anode and cathode, respectively. The YSZ layer was deposited at 300 °C using an Ar/O2 mixture of 30 : 1. For the 21NbSB layer, the composition of the deposited film depends greatly on the gas mixture composition which with an Ar/O2 mixture of 31 : 7 yielded the closest film composition to the target composition. It was found that the bilayer electrolyte deposited at 300 °C was thermally stable and did not crack after a 700 °C anneal. However, the MPD of the cell was rather low at 0.01 W cm⁻² at 600 °C using a 4 µm YSZ and 1.5 µm NbSB bilayer electrolyte.

As reviewed in the previous section, the initial bilayer electrolyte cells were built using thick electrolyte pellets and noble metal electrodes, and the cell performance was low. To improve the bilayer cell performance, research focused on developing anode-supported cells using a thin-film electrolyte and new cathode materials compatible with bismuth oxide electrolyte has been carried out. Traditional cathode materials for IT-SOFCs such as perovskite LSCF show extreme reactivity with bismuth oxide, and among noble metals, Pt is known to form secondary phases such as Bi₂Pt₂O₇ with bismuth oxide.⁷⁸  One of the cathodes discovered was a bismuth ruthenate (BRO)-ESB composite by Jaiswal.⁴⁴  It was shown that the BRO and ESB phases are compatible with each other by XRD and DSC studies. Another system under investigation is the LSM-ESB composite;⁷⁹  however, the compatibility results are inconclusive with respect to the reaction between LSM and ESB and the consequent effect on the electrode performance and stability.

Ahn et al.^80,81  fabricated anode-supported cells with a thick film bilayer electrolyte by different fabrication routes and studied cell performance using different cathodes. Ni-GDC/GDC half-cells were made by either co-pressing of the anode and GDC powders or by spray coating of the GDC layer on a tape-casted anode. The co-pressing method resulted in a relatively thick (∼50 µm) GDC layer, while the spray coating method resulted in a 10–20 µm GDC layer. For the cells made by the co-pressing method, an ESB layer was made by screen printing using solid state powders. For the cells made by spray coating, an ESB layer was made by either screen printing using co-precipitated powders or by hot/cold PLD. In the last step, a BRO-ESB cathode was brush-painted to complete the cell. A SEM cross-section image of a GDC/ESB bilayer cell is shown in Figure 1.18. The performance of the bilayer cells was compared with a monolayer GDC cell using a standard LSCF-GDC cathode. For the co-pressed cells, the screen-printed ESB layer did not sinter completely into a dense layer due to the large size of the solid state powders and thus was found to not effectively increase the OCV of the bilayer cell. On the other hand, the BRO-ESB cathode reduced the ASR of the bilayer cell resulting in a MPD of 0.59 W cm⁻² at 650 °C, which was higher than that of the monolayer GDC cell with a LSCF-GDC cathode (0.34 W cm⁻²). For the cell made by spray coating, the OCV was significantly lower due to the lower thickness (10–20 µm) of the GDC layer, but the screen-printed ESB layer was found to be effective in increasing OCP from 0.67 to 0.75 V, resulting in an increase in MPD of the cell from 0.41 to 0.61 W cm⁻² at 650 °C. Moreover, the ESB layer made by hot PLD exhibited lower ohmic and non-ohmic ASR and higher OCV (0.77 V vs. 0.72 V at 650 °C) compared to that of the monolayer GDC cell, resulting in a 93% increase in MPD from 1.03 W cm⁻² for the monolayer cell to 1.95 W cm⁻² for the bilayer cell with a ∼4 µm thick ESB layer made by hot PLD (Figure 1.19). The improvement in cell performance was ascribed to the combined effect of higher OCV using an ESB layer and 43% lower ASR using a BRO-ESB composite cathode. This performance, ∼2 W cm⁻² at 650 °C, set the benchmark for high power density ceria/bismuth oxide bilayer electrolytes and a subsequent publication³  laid out the potential and issues for this technology.

As a result, numerous groups have investigated this approach and continued the improvement in OCP and power density. However, while the result of ceria/bismuth oxide bilayer electrolyte-based SOFCs was impressive, the OCV was lower than the theoretical estimate, suggesting that further optimisation of the relative thickness and total thickness of the bilayer electrolyte needs to be carried out. As such, there is a trade-off in optimising the total thickness of the bilayer electrolyte between reducing ASR with thinner electrolytes and increasing OCV with thicker electrolytes.⁴⁷  In addition, for bilayer electrolytes, the OCV further depends on the relative thickness of the two layers. Lee et al.⁸²  studied an anode-supported GDC/ESB bilayer cell with a BRO-ESB cathode to understand the effect of relative thickness of the two electrolyte layers on the cell performance. A GDC electrolyte layer was deposited by spin coating of a GDC slurry on a NiO-GDC anode support. An ESB layer was fabricated by drop coating using powders made from solid state and co-precipitation methods. It was found that powders made by co-precipitation were finer and resulted in a dense ESB layer after sintering at 800 °C. Cells with different GDC and ESB thicknesses were fabricated and it was observed that by increasing the thickness of the GDC and ESB layers, the OCV of the cell increased. The MPD of a bilayer cell with 10 µm GDC and a 5 µm thick ESB reached over 1.5 W cm⁻² at 650 °C. The increase in MPD compared to the monolayer GDC cell (0.87 W cm⁻²) was due to both higher OCV (0.8 vs. 0.75 V) and 47.6% lower ASR.

Lee et al.⁸³  studied the performance of GDC/ESB bilayer cells with a LSM-ESB cathode using a tape-cast anode and anode functional layer (AFL). GDC and ESB layers were deposited with spin coating of a colloidal solution and the sintered thickness was 25 and 2 µm, respectively. The MPD of the cell was 1.02 W cm⁻² at 650 °C, which was higher than all previously reported LSM cathode-based SOFCs. In addition, cathode and cell performance was stable for over 80 and 200 h of testing, respectively. Further improvement in cell performance was suggested by replacing the GDC/ESB bilayer combination with higher conductivity doped ceria and stabilised bismuth oxide electrolytes such as SNDC and DWSB.⁴⁷  Redox Power Systems is utilising these developments in ceria/bismuth oxide bilayer electrolytes to develop commercial SOFCs operating at low temperatures.

Recent reports on ceria/bismuth oxide bilayer electrolyte-based SOFCs by Zhang et al.,⁸⁴  Lee et al.⁸⁵  and Hou et al.^86,87  have looked at SDC and SNDC-based ceria electrolytes, YSB-based bismuth oxide electrolytes and LSM-based cathodes. For SDC/YSB bilayer SOFCs using a LSM-YSB cathode by Zhang et al.,⁸⁴  a NiO-SDC (60 : 40) anode and SDC electrolyte were co-pressed and co-fired at 1250 °C. The YSB electrolyte was deposited using DC magnetron sputtering in an argon/oxygen mixture at room temperature and then annealed at 800 °C for 2 h. The thickness of the SDC and YSB layers was 26 and 6 µm, respectively. Using the bilayer electrolyte, the OCV of the bilayer electrolyte cell increased to 0.85 V compared to 0.78 V of the monolayer SDC cell at 650 °C. MPD also increased from 0.29 W cm⁻² for the monolayer cell to 0.38 W cm⁻² for the bilayer cell. Ohmic resistance of the bilayer cell did not change with the addition of the YSB layer due to its high conductivity. However, the non-ohmic resistance of the bilayer cell under load was much lower than that of the monolayer cell due to the positive effect of the YSB electrolyte on cathodic polarisation. The overall increase in MPD was attributed to the higher OCV and lower cathodic polarisation of the bilayer electrolyte.

Lee et al.⁸⁵  reported on their work on GDC/YSB bilayer SOFCs using a LSM-YSB cathode. A NiO-GDC anode support was fabricated by powder pressing and sintering at 1100 °C. A GDC electrolyte was deposited by drop coating and sintering at 1400 °C, while a YSB electrolyte and LSM-YSB (50 : 50) cathode were screen printed and sintered at 800 °C and 850 °C, respectively. The thicknesses of the GDC and YDB layer were 20 and 10 µm, respectively. For the reference monolayer GDC cell, LSM-GDC was used as the cathode material. The bilayer cell showed a slightly higher ohmic ASR but a significantly lower non-ohmic ASR compared to the monolayer cell with a LSM-GDC cathode. The lower non-ohmic ASR was attributed to the lower resistance of LSM-YDB for charge transfer at the electrode/electrolyte interface. It was also noted that despite a significantly lower ionic conductivity of LSM compared to LSCF, the LSM-YSB cathode showed a similar performance to that of the LSCF cathode at 650 °C, suggesting the beneficial effect of YSB in the composite cathode. The bilayer electrolyte enhanced the OCV from 0.50 to 0.89 V and the MPD improved from 0.21 to 0.53 W cm⁻² at 650 °C.

Hou et al.^86,87  fabricated Sm_0.075Nd_0.075Ce_0.85O₂ (SNDC)/ESB and GDC/ESB bilayer electrolyte SOFCs using La_0.74Bi_0.1Sr_0.16MnO₃(LBSM)-ESB (50 : 50) as the cathode. A Ni-ceria (60 : 40) anode and ceria electrolyte were prepared by co-pressing and sintering at 1400 °C. An ESB electrolyte was deposited by drop coating and sintering at 800 °C. The thickness of the ceria and ESB layers after sintering was 20 µm each. Bismuth doping of LSM was found to be effective in limiting any reaction between LBSM and ESB after high temperature sintering as confirmed by XRD diffraction pattern analysis. Impedance spectroscopy showed that above 550 °C, ohmic losses accounted for the majority of the cell resistance. At 650 °C, the OCV and MPD of the SNDC/ESB bilayer cell reached 0.80 V and 0.98 W cm⁻², respectively, as shown in Figure 1.20.⁸⁶  At 450 °C, the MPD of the bilayer cell was 0.13 W cm⁻², which the authors claimed to be the highest recorded value at 450 °C.⁸⁷  In another work, Hou et al. compared the electrochemical performance of SNDC/ESB to that of GDC/ESB bilayer electrolyte SOFCs.⁸⁷  At 650 °C, the OCV and MPD of the SNDC/ESB and GDC/ESB cells were 0.81 V and 0.93 W cm⁻² and 0.83 V and 0.90 W cm⁻², respectively. The higher MPD of the SNDC/ESB cell was attributed to the lower ohmic ASR of the SNDC/ESB cell, as confirmed by the impedance spectroscopy. In addition, it was noted that the conductivity of the bilayer electrolyte calculated from the overall thickness of the electrolyte and the ohmic ASR of the cell was higher than that of SNDC and GDC. The authors argued that the measured conductivity in an operating cell was higher than the conductivity measured in air due to consumption of oxygen at the anode. However, impedance measurements were performed at an open circuit condition and hence, consumption of oxygen at the anode should be negligible and cannot be expected to affect the conductivity of the bilayer electrolyte.

Other researchers have studied ceria/bismuth oxide bilayer electrolytes using different cathode materials – Pt,⁸⁸  Ag-YSB⁸⁹  and Pr_0.5Ba_0.5MnO₃ (PBM).⁹⁰  Leng and Chan⁸⁸  fabricated GDC/YSB bilayer SOFCs using a Pt cathode. A NiO-GDC anode and GDC electrolyte were fabricated by co-pressing and sintering at 1350 °C. YSB was applied by screen printing and sintered at 900 °C. A SEM micrograph of the bilayer cell showed a dense, 84 µm thick GDC layer and a porous, 16 µm thick ESB layer. The porous ESB layer was able to marginally enhance the OCV of the bilayer cell from 0.75 to 0.81 V at 700 °C. Although the ohmic resistance of the cell did not change, the non-ohmic polarisation ASR of the bilayer decreased by 36%. The MPD of the bilayer cell was relatively modest but increased from 0.16 to 0.26 W cm⁻² at 700 °C due to the higher OCV and lower ASR of the bilayer cell. Zhang et al.⁸⁹  fabricated and tested SDC/YSB bilayer electrolyte SOFCs with an Ag-YSB cathode. A NiO-SDC anode and SDC electrolyte were pressed and co-fired at 1250 °C for 5 h. A YSB layer was deposited using DC magnetron sputtering and annealed at 800 °C for 2 h. The thickness of the SDC and YSB layer was 26 and 6 µm. An Ag-YSB (60 : 40) composite cathode was deposited by screen printing. The bilayer SDC/YSB electrolyte improved the OCV from 0.80 to 0.89 V at 500 °C, while the MPD of the cell improved from 0.13 to 0.22 W cm⁻². At 600 °C, the MPD of the bilayer cell was 0.57 W cm⁻². Impedance studies using symmetrical cells showed that the ASR of the Ag-YSB cathode on the YSB electrolyte was lower than that on the SDC electrolyte and suggested that the increase in MPD of the bilayer cell was due to both higher OCV and lower interfacial polarisation compared to the monolayer SDC electrolyte cell. Hou et al.⁹⁰  developed a new PBM perovskite and ESB (50 : 50) based composite cathode for SNDC/ESB bilayer electrolyte SOFCs. PBM and ESB powders were made using a citric acid-nitrate gel combustion method. PBM showed a perovskite structure composed of cubic and hexagonal phases but after firing with ESB at 750 °C for 3 h showed minor secondary phase in the XRD diffraction patterns. The activation energy for the PBM-ESB cathode was estimated to be 1.38 eV. A NiO-SNDC (60 : 40) anode substrate and SNDC electrolyte were co-pressed and fired at 1400 °C. An ESB electrolyte was deposited by drop coating and sintered at 800 °C. The OCV and MPD of the bilayer cell with a 22 µm thick SNDC layer and a 20 µm thick ESB layer were 0.80 V and 0.99 W cm⁻² at 650 °C, respectively.

In solid electrolytes, doping of aliovalent cation increases the oxide vacancy concentration but the oxide ion conductivity can be enhanced by increasing the dopant concentration only up to a certain level. Beyond that level, the ionic conductivity decreases due to the association of vacancies or formation of defect complexes in the dilute solution range and the formation of super-lattices or ordering of vacancies in the concentrated solution range.^91,92  The extent of anion sublattice ordering can be reduced by using highly polarisable dopants with ionic radii close to that of the host cation and randomising the site energetics by multi-cation doping.⁹³  This approach has been used to develop double-doped ceria, such as samarium and neodymium co-doped ceria (SNDC)⁹⁴  and dysprosium and tungsten co-stabilised bismuth oxide (DWSB),⁹³  which show higher oxide ion conductivity compared to ESB and GDC, respectively. In the bismuth oxide system, double doping reduces the amount of dopant required to stabilise the cubic phase to room temperature and hence, increases the ionic conductivity compared to traditional ESB. In addition, the cubic ESB electrolyte suffers from disorder–order transformation on the oxide-ion sublattice at ∼500 °C, resulting in decay of oxide ion conductivity with isothermal aging.⁹¹  Recent understanding on the interplay between dopant type/level and phase composition has resulted in the development of a new stabilised bismuth oxide composition, which shows stable ionic conductivity at ∼500 °C.^95,96  Ruth et al.⁹⁷  have studied the performance of bilayer SOFCs using double-doped bismuth oxides, i.e. GDC/DWSB and GDC/DGSB. At 650 °C, the conductivity of DGSB and DWSB was 0.94 and 0.53 S cm⁻¹, respectively, which was significantly higher than of ESB. Ni-GDC and LSM-ESB were used as the anode and cathode material, respectively. The OCV of the bilayer's cells with 20 µm GDC and 3–5 µm bismuth oxide layers at 650 °C was ∼0.8 V, while the MPD was 0.6, 0.9 and 1.3 W cm⁻² for ESB, DWSB and DGSB-based bilayers, respectively.

An interesting work on a layered heterostructure consisting of alternating GDC/ESB layers was reported by Sanna et al.⁹⁸  with the goal to improve the stability of the ESB phase to room temperature and in reducing environments. The alternating GDC/ESB layers were deposited on a MgO substrate using PLD to result in a MgO/GDC/[ESB/GDC]_N structure, where N was varied from 1 to 30. For a constant film thickness of 60 nm, N = 20 showed the highest conductivity in both air and reducing environments. Unlike bulk ESB, the order–disorder transition and related conductivity decay was not observed at 600 °C in air. Moreover, the conductivity of a heterostructure with N = 20 was found to increase by two orders of magnitude in 90% H₂/10% N₂ reducing environment at 533 °C. Lastly, the heterostructure was found to be stable in the reducing environment and the electrical conductivity was constant. The conductivity results for the heterostructure were extremely interesting but no full cell results using the heterostructured electrolyte were presented.

As mentioned previously, bilayer electrolytes serve distinct functions in a SOFC structure. In ceria/bismuth oxide or ceria-based ceria/zirconia bilayers, bismuth oxide and zirconia layers block the leakage current through ceria and increase the OCV. However, in other types of bilayer electrolyte SOFCs, the main electrolyte is not reduced and near theoretical OCV can be obtained. The major functionality of the secondary electrolyte is to suppress the reaction between the electrolyte and electrodes or protect the electrolyte from reaction with fuel constituents such as water, CO₂ or SO₂. For instance, LSGM has high oxide ion conductivity and shows near theoretical OCV. However, LSGM and Ni-based cermets react and resistive phases with low conductivity form at the interface. Hence, many researchers have studied the potential of using a thin ceria-based electrolyte such as SDC,^99–103  or LDC^104–107  as a diffusion barrier between the LSGM electrolyte and Ni-cermet anode. Proton conductor-based SOFCs typically use doped BaCeO₃ as their electrolyte. Despite its high ionic conductivity BaCeO₃ suffers from poor chemical stability under CO₂, acidic gases and steam, which are prevalent under the operating environment of a SOFC. On the other hand, BaZrO₃ has a conductivity one order of magnitude lower than that of BaCeO₃. However, it possess good chemical stability and hence the potential of protecting the BaCeO₃ with a thin layer of BaZrO₃ has been studied in the literature.^108–110  Interested readers are referred to other types of bilayers such as ceria/BZCY,^111,112  LBAYO/SDC¹¹³  or biphasic electrolytes^114–116  as a means to improve the OCV of ceria-based SOFCs.

As summarised in Tables 1.1 and 1.2, for similar total electrolyte thickness and temperature, ceria/zirconia bilayer IT and HT-SOFCs yield higher OCV values compared to ceria/bismuth oxide IT and LT-SOFCs. However, due to the significantly lower ohmic ASR across the ceria/bismuth oxide bilayer electrolyte layers, these bilayer electrolyte SOFCs have the potential to deliver higher power density at lower temperatures. For example, at 650 °C, ScSZ (6–7 µm)/SDC (6–7 µm) bilayer electrolyte SOFC⁵⁶  and GDC (10 µm)/ESB (4 µm) bilayer electrolyte SOFC⁸¹  (total electrolyte thickness of 14 µm) demonstrated OCVs of 1.04 V and 0.77 V, respectively; however, in terms of MPD, they achieved 1.38 and 1.95 W cm⁻², respectively.

As such, although many scholarly works in developing IT-SOFCs are primarily focused on depositing a thin stabilised zirconia layer at lower temperatures using alternative processing methods, these bilayers often suffer from higher ohmic losses and hence, despite having higher OCV, they deliver lower power output. Consequently, ceria/bismuth oxide bilayer electrolyte SOFCs are suggested as a more promising approach to reduce the SOFC operating temperature to 650–500 °C. Towards this goal, (1) new doped bismuth oxide compositions with stable oxide ionic conductivity at temperatures below 600 °C must be explored, (2) structure of the cell, i.e. bilayer electrolyte thickness ratio and fabrication processes must be engineered, and (3) new cathode materials with low ASR at low temperatures, that are compatible with bismuth oxide, should be developed.

Lowering the operating temperature of the solid oxide fuel cell below 650 °C reduces the manufacturing cost, decreases the performance degradation rate, increases efficiency and hence can promote the commercialisation of SOFC technology. Two major approaches to reduce the operating temperature are typically pursued: (1) using alternative electrolyte materials with higher ionic conductivity such as doped ceria instead of the traditional YSZ, and (2) few microns or sub-micron electrolyte SOFCs. Sub-micron electrolyte SOFCs show reasonable power density at lower temperatures. However, fabrication requires more expensive semi-conductor processing such as advanced lithography, ALD, or sputtering and hence may not be cost effective for large-scale power generation potential; therefore, they were not reviewed in detail in this work. Doped ceria such as GDC and SDC have ionic conductivity orders of magnitude higher than YSZ and hence have been explored extensively. However, ceria reduces under the fuel conditions of the SOFC, which causes an internal leakage current and power dissipation. Stabilised bismuth oxide compositions such as ESB, YSB and DWSB have been shown to be capable of blocking the leakage current and increasing the OCV of the cell. Although many scholars have investigated the potential of using a thin zirconia layer to block the leakage current, ceria/bismuth oxide bilayer SOFCs showed higher power density values due to the much higher ionic conductivity of stabilised bismuth oxide compared to YSZ. The ceria/bismuth oxide thickness ratio controls the interfacial p_O2 and the thermodynamic stability of the bismuth oxide layer, and hence is of great importance and must be tuned carefully. Bilayer electrolyte SOFCs can serve other functions (other than blocking the leakage current) such as a diffusion barrier layer between the electrolyte and adjacent electrodes or to enhance the catalytic activity of the cathode as a fast oxide ion conductor.

BRO –

Bismuth ruthenate

BZCY –

Barium cerium oxide doped with zirconia and yttria

DGSB –

Dysprosium gadolinium-stabilised bismuth oxide

DWSB –

Dysprosium tungsten-stabilised bismuth oxide

ESB –

Erbium-stabilised bismuth oxide

GDC –

Gadolinium doped ceria

LBAYO –

Lanthanum barium yttrium oxide

LBSM –

Lanthanum bismuth strontium manganese oxide

LSCF –

Lanthanum strontium cobalt iron oxide

LSGM –

Lanthanum gallate doped with strontium and magnesium

LSM –

Lanthanum strontium manganese oxide

LYBO –

Lanthanum yttrium-stabilised bismuth oxide

NbSB –

Niobium-stabilised bismuth oxide

NGSB –

Niobium gadolinium-stabilised bismuth oxide

ScSZ –

Scandia-stabilised zirconia

SDC –

Samarium doped ceria

SNDC –

Samarium neodymium doped ceria

xSB –

Double-doped bismuth oxide

YDC –

Yttrium doped ceria

YSB –

Yttrium-stabilised bismuth oxide

YSZ –

Yttria-stabilised zirconia