Final Report Summary - EVOLVE (Evolved materials and innovative design for high-performance, durable and reliable SOFC cell and stack)
Beyond the state of the art, the EVOLVE cell concept aims at combining the beneficial characteristics of the previous cell generations, the so called Anode Supported Cells (ASC) and Metals Supported Cells (MSC) while tackling key challenges like sulfur poisoning and redox stability, typical for state of art cells implementing nickel as electro-catalyst and structural component in the nickel / zirconia standard anodic cermet. The innovation of the EVOLVE cell concept remains in its anode compartment avoiding the use of pure nickel as structural component. The substrate, providing mechanical strength to the fuel cell is based on a robust metal alloy 3D porous backbone allowing the formation of a protective alumina layer enhancing stability during re-oxidation cycles and an electronic conducting material based on perovskite oxides.
An European consortium including academics and industrial partners with complementary backgrounds and expertise has been build up, and has been granted a financial support in 2011 by the Fuel Cells and Hydrogen Joint Understanding under the Grant Agreement n° 303429, for the development of the EVOLVE cell concept. The consortium EVOLVE included: the German Aerospace Center (Germany) assuming the leadership and the coordination, Alantum Europe GmbH (Germany), ARMINES (France), Ceramic Powder Technology AS (Norway), Consiglio Nazionale delle Ricerche (Italy), Institut Polytechnique de Grenoble (France), Saan Energi AB (Sweden) and Ceraco Ceramic Coating GmbH (Germany). The project ran from November 2012 until January 2017.
In a first step, set of materials were selected based on their physico-chemical properties and inter-compatibility for manufacturing a fuel cell. Namely a NiCrAl based metal foam provided by Alantum GmbH, as 3D metal alloy porous backbone, La0,1Sr0,9TiO3-δ, produced by CerPoTech AS, as electronic conducting material for the substrate and anodic electro-catalyst in combination with Gd doped ceria in the functional anode layer, yttria stabilized zirconia for the electrolyte, in combination with GDC as cathodic barrier layer and applied as thin film by EB PVD (Ceraco GmbH), and finally a LSCF based cathode.
Taking into consideration the properties of the material a specific modular manufacturing route was specifically designed for producing the EVOLVE cell. This route was divided into 5 consecutive modules consisting in 1) substrate manufacturing, 2) processing of anode functional layer (AFL) 3) Thin film Electrolyte Processing, 4) Cathode manufacturing, and 5) activation of the anode layer.
As level of electronic conductivity and electrocatalytic activity were limited, performance of the pure nickel free EVOLVE cells were limited to ca. 120 mW/cm² at 0.7V and 750°C with hydrogen as fuel. After the midterm evaluation of the project, it was thus decided to enhance the electronic conductivity in the substrate by replacing the LST material by a cermet LST-Ni (50:50), and boost the electro-catalytic activity of the anode functional layer by surface modification of the LST-CGO composite AFL with ca. 5 to 10wt % of nickel. With those modifications, the power density was increased up to 440 mW/cm² at 0.7V and 750°C with hydrogen as fuel. Comparable level of performance were achieved for cells built up on a standard ferritic stainless porous substrate, which confirmed the modularity of the processing route i.e. materials can be easily exchanged without redesigning a full manufacturing route.
The EVOLVE cell architecture with electrolyte thickness of less than 3µm could be successfully up-scaled to a size of 90 mm x 100 mm and could be successfully integrated into 1 cell stack, similarly to anode supported cells. The tested 1 cell stack achieved a power density up to 170 mW/cm² at 0.7V and 750°C with simulated syngas as fuel and air at the cathode. The stack could be successfully operated for more than 100 hours under constant current load. With less than 2% variation for 50 redox cycles, the Open Circuit Voltage showed an excellent stability, demonstrating that thin film electrolytes maintain integrity despite the harsh conditions.
At that point, the feasibility of the EVOLVE cell architecture has been demonstrated at the stack level. Further work would require assembly and tests of short stack fur further assessment, as well as implementation of technological improvements in the cell to enhance performance and durability.
The modular approach developed in the project for the cell manufacturing showed very promising characteristics for a quick implementation of advanced materials in the cell design and the possibility to involve different actors and processes in the fabrication process, which contrast with centralized cell production that can be observed nowadays in Europe. This modularity of the manufacturing route provides huge potential in terms of cost reduction potential on the basis of the principle of best performance / cost ratio and the flexibility offered regarding the materials that can be implemented.
Project Context and Objectives:
Emission of greenhouse gases (GHG) from industry, transportation and agriculture has played a major role in the recently observed global warming. In order to reduce the impact on environment, zero or near-zero emission energy systems, converting fuel efficiently into useful energy are increasingly investigated. Unprecedented rise in the fossil fuel prices has led to a growing economic concern towards the need of such energy systems. Additionally, geopolitical priority of reducing the dependence on foreign energy resources emphasizes the development of systems that may work with a variety of fuels and eventually with renewable energy resources such as hydrogen or bio-fuels.
Solid Oxide Fuel Cells (SOFCs) are one of the most attractive energy conversion devices, owing to the potential of operating at high efficiency of about 60% in standalone condition and over 80% (net) if waste heat is used for cogeneration. SOFCs do not require noble metals for catalysis in electrodes and may use a variety of fuels including hydrocarbons, CO and bio-fuels, besides hydrogen. These low-noise convertors thus offer very high potential in stationary application and combined heat and power units (CHP) for decentralized energy. Despite all the promising advantages and the unparalleled progress in its power output, SOFC faces critical challenges in term of its poor reliability, low durability and higher cost. Unless addressed meticulously, these obstacles will impede large-scale commercialization of fuel cells. Reliability and durability are adversely affected by a number of factors of which the following two can be considered as the route cause: 1) high operating temperatures (800-1000°C) of SOFC and 2) the need to use materials that provide multiple functionalities. This includes structural support, electrochemical activity, and electrical or ionic conduction, as well as at the same time compatibility with neighbouring components during the manufacturing process and fuel cell operation. This has so far made an optimization of each individual functionality impossible.
Improving the anode compartment of the SOFC can address several issues of cell degradation in static and cyclic conditions including: Ni coalescence and agglomeration, carbon and sulphur tolerance, robustness of the cell during thermal and redox cycling. We propose to split the functionality of anode materials, thus relaxing the stringent materials demands and opening the way for these necessary improvements. This project focuses on an innovative concept for SOFC, particularly for the anode compartment, incorporating advanced materials with an approach in which each material performs only one functionality and aims at cell operation at reduced temperature of 750 °C. This concept, which has led to a patent application , is expected to enhance the durability and reliability of SOFC while exhibiting performance level equivalent to main-stream anode-supported cells. The scope of the project aligns itself with the objectives set in the call under the topic of SP1-JTI-FCH.2011.3.1. It is thus targeted:
-to reduce the amount of Nickel in the anodic substrate also called current collector,
-to replace Nickel within the Anode Functional Layer by a composite LST (La(1-x)SrxTiO3-x)-GDC modified by catalysts
The main objectives of EVOLVE are:
- the demonstration at the stack level of a SOFC implementing an innovative substrate resilient toward redox cycles with higher durability than mainstreams Metal Supported Cells implementing porous ferritic stainless steel substrates and cyclability than mainstreams anode supported cells implementing the Ni based cermet.
- the identification of innovative combinations of advanced materials with reduced amount of nickel, showing improved tolerance against sulphur poisoning compared to mainstreams nickel based cermet Anode and higher resilience toward redox cycles.
The cell, represented schematically, will consist of functional layers supported on an innovative anode current collector. This current collector will be a ceramic metal composite in form of metallic foams, between 0.6 and 1 mm in thickness, impregnated with conductive ceramics. In the current collector, no Ni particles will appear as structural component or for current collection. Because percolation of Ni is not necessary in the new concept, the nuisance of Ni coalescence and redox failure in state-of-the-art cells will be avoided. The metallic foam will be of an alloy composition that can form a thin continuous stable oxide layer of Al2O3. The main advantage of this oxide is in its thermal stability and its ability to slow down the oxidation kinetics of the underlying metal in case of exposure to oxidant environment. Al2O3-forming alloys are therefore significantly more durable compared to Cr-forming alloys at elevated temperatures. Metallic foams will be first anodized in order to form a continuous thin protective layer of Al2O3. The metallic component will provide a robust structural integrity of the cell whereas thin alumina scale will enhance its chemically stability under cell fabrication and operating conditions. This formation of Al2O3 will, however, render the metal foam electronically non-conductive (or poorly conductive). The backbone of oxidized metal foam will therefore be impregnated with conductive ceramics such as LST (La(1-x)SrxTiO3-x), which will lead to a percolating phase of electron conductor. In this manner, the mechanical and chemical stability will be attained by oxidized metallic foam whereas the electronic (and ionic) conductivity will be guaranteed by the impregnated ceramic. Using doctor blade method or a modified screen printing technique the ink will be then pushed into the upper surface to create a very low roughness topography on which functional layers can be produced. An anode layer of 10 to 30 µm in thickness and composed of LST-GDC will then be coated on top of the impregnated foam by colloidal suspension spraying or screen printing (this layer will be infiltrated with catalyst in the final stage). A thin electrolyte layer of YSZ (thickness between 1 to 5 µm) will be coated using fine-grain-sized powder in order to decrease the sintering temperature below the upper limit of the stability of the metal substrate. The half-cell will be sintered. GDC of 0.1 to 0.5 µm in thickness will be deposited on top of zirconia electrolyte as diffusion barrier layer. A 15 to 30 µm thick cathode, based on LSCF, will be finally coated on top of the interlayer. At the last stage, suitable catalysts such as Co, Ni will be infiltrated (in restricted quantity) within the electrodes as salts before integration and testing of cells.
The objectives of the project will be targeted by correlating in detail the manufacturing processes for different cell components with their microstructure and with the electrochemical performance of electrodes. The latter two will be undertaken by sophisticated experimental tools and by modelling approaches. 3D-microstructural analyses coupled with microstructural and electrochemical modelling will be extensively employed to generate an effective map of process-microstructure-performance relationships. Based on this map, the optimal cells will be produced – split into intermediate milestones defined in the work packages. The processes considered for the electrode layers are colloidal spraying, screen printing and suspensions plasma spraying. In the case of the electrolyte, the ultra-thin layer of thickness of 1 to 5 µm, assisting in reducing the ohmic ASR of the cell, will be coated by colloidal spraying or sol-gel techniques. The degradation of the cell will be targeted by introducing a two-way strategy: 1) the fuel tolerance of the anode will be targeted by incorporating materials reported to have better tolerance against coking and sulphur poisoning such as LST along with dispersed nickel. Doped cerium oxide will be added by infiltration techniques in the ceramic matrix of anode to further enhance carbon and sulphur tolerance. These materials will be tested in carbon and sulphur containing fuel gases. 2) Lower operating temperature of 750 °C will be considered as high operating temperature is one of the major causes of the degradation. 750 °C is opted to maintain a respectable performance level of the cell and to avoid enhanced sensitivity against carbon and sulphur. On the cathode side, LSCF-CGO composite cathodes will be produced after depositing an electron beam physical vapour deposition (EB-PVD) barrier layer of CGO on the zirconia-based electrolyte. Alternatively nickelate based cathode materials that do not react readily with the zirconia based electrolyte will be exploited. Besides, 3D analysis and modelling will be utilized to optimize the microstructure of electrodes leading to lowest degradation and highest performance. The tolerance of the SOFC in transient conditions (thermal and redox) will be improved by avoiding Ni as the structural component. Instead, the structural stability will be given by the metallic foam forming a stable and protective thin oxide layer (Al2O3) at elevated temperature. Redox cycling is thus expected not to introduce stresses in the cell. The performance criteria will be still fulfilled by the ceramic current collector impregnated within the substrate, like LST. However, it needs to be confirmed if LST loose conductivity during reducing and oxidizing cycles. Resistance against thermal cycling will be enhanced by adjusting precisely the coefficient of thermal expansion (CTE) the components. This will be done by adopting the stoichiometry of perovskite materials and adjusting the volume percentage of different phases in the anode. The process development for the functional layers will be again based on modelling to recommend the appropriate microstructure of layers to reduce the risks of failure under transient conditions. The developed cell will be up-scaled to industrially relevant size of a footprint of around 100 cm² and will be integrated into stacks.
Project Results:
Reference materials for the anodic compartment, namely La0,1Sr0,9TiO3-α (LST) as anode material was synthesized and successfully combined with a metal foam based on the NiCrAl alloy for manufacturing the current collector (Figure 1).
Figure 1 NiCrAl foam from Alatum GmbH showing a cell size of 450µm.
A first prototype has been developed using the Plasma Spraying technique for electrolyte deposition, La0,1Sr0,9TiO3-α (LST) as reference material (Figure 2) for the anode and the cermet NiCrAl / LST as current collector. Implemented in full cell with a LSM based cathode and thick YSZ electrolyte (ca. 100µm) to achieve sufficient gas tightness, the performance of the first prototype was limited to 30mW/cm² at 750°C but remarkably stable over a period of time of 180h in tested conditions. As a further development step an approach using thin film electrolyte has been developed aiming at reducing the cell polarization.
Figure 2 : SEM image of the LST powder produced by CerPoTech.
The dip-coated YSZ layer and the EB-PVD GDC layer have been homogenously deposited on top of the LST-GDC anode layer. The thickness of the calcined LST-GDC anode layer is about 20 µm. The thickness of the thin-film YSZ layer and the gas-tight GDC layer was approximately 1 µm and 2 µm, respectively. With the homogenous YSZ layer and the dense GDC layer, some of the produced half cells realized excellent gas-tightness. Despite the extremely thin electrolyte, button cells with size up to 25 cm² were produced showing an air leakage rate down to 7.0*10-4 (hPa*dm3)/(s*cm2). Full cells with LSCF cathodes were produced for further electrochemically tests. It is interesting to mention that a yield of 100% was obtained on the last produced batch. This demonstrates the reliability of the coating approach for producing the electrolyte. However, it must be mentioned, that co-firing the half cells at elevated temperature (1000 oC) in air, the leak rate of the cells increase significantly due to the structural changes in the bi-layer electrolyte. Further work needs to be performed to stabilize the microstructure and avoid degradation of leak-rate in case of thermal ageing. It was found that the success rate of the thin film deposition approach is largely dominated by the density of macroscopic defects (>200nm in size) in the underlying layers.
By reducing electrolyte thickness down to 3 µm in total, the performance of EVOLVE cells could be increased to about 100 mW /cm² at 750°C and 0,7V. From the electrochemical investigation, it resulted that high ohmic resistance coupled with high polarization resistance were jointly responsible for the moderate level of performance achieved. Posttest analysis of the electrolyte revealed the presence of cobalt in the porous YSZ supporting layer and originating from the cathode. For this reason, an additional dense YSZ layer produced by EB-PVD was incorporated in the multi-layer structure of the electrolyte between the porous YSZ layer and the dense and gas tight GDC layer.
In this first nickel free prototype, Electronic conductivity of the LST was evaluated below 10 S/cm-1 at our operating conditions which makes questionable its relevance for use as current collector. For this reason, it was considered to incorporate Nickel (up to 57wt % of NiO in the slurry) both in the current collector and as catalytic material in the functional anode layer (between 5 to 10 wt%).
By modifying the original concept with further addition of Ni within the current collector – for enhancing the electronic conductivity – and in the functional anode layer – for enhancing the electro-catalytic behavior – a power density of 340 mW/cm² at 750 °C and 0.7 V is demonstrated. Further modification of substrate consisting in the addition of pore former led to an increase of power density up to 440 mW / cm² at 750 °C at 0.7 V. This corresponds to an increase of performance of about 30 % at reference voltage compared to previous generation. The implementation of a 0,5µm thick Yttria stabilized zirconia layer by PVD in the multi layered structure of the electrolyte impacted positively the Open Circuit Voltage of the cell, and durability of the performance. The precise quantification of its impact on Open Circuit Voltage remains nonetheless difficult to quantify exactly as this parameter still remains highly sensitive to density of defects in the electrolyte. The same performance level was achieved by transferring the functional layers - Anode functional layer, Electrolyte and Cathode – on a porous ferritic stainless steel substrate with the same manufacturing route. This result highlighted the modularity of the processing route and its compatibility with other substrate. This provides evidence that a modular manufacturing of the cell is possible with possibility to exchange elementary components without having to redesign a manufacturing route. The redox tolerance was assessed by exposing for 30mn the anode functional layer to a stream of pure oxygen. With less than 2% variation for 50 cycles, the Open Circuit Voltage showed and excellent stability as a function of redox cycles, demonstrating that thin film electrolyte maintain integrity. Electrochemical measurements were performed on two cells by adding 5 ppm H2S at the anode side at 750 °C. Operating the EVOLVE cell with 5 ppm H2S did not result in significant decrease of performances for both cells. During sulfur exposure, nickel free cell showed better performances but a higher degradation rate was recorded during recovery. On the other side, only a small continuous performance drop was recorded on cell modified with nickel. Though the reason are not yet clearly understood, this result clearly indicates that modifying the original concept of the EVOLVE cell by introducing nickel is beneficial to operate in harsh conditions.
Advanced contact materials for the air side in SOC stacks were investigated. They consist in flexible and inexpensive metal foams produced by Alantum GmbH, able to be converted into an electronic conductive oxide when exposed to air at high temperature. CoNi based metal foams were tested as contact material for the air side for more than 1000 hours of operation and showed very promising performance: No quantifiable performance loss could be attributed to the contact degradation. This material was considered for stack integration in the EVOLVE project, nonetheless considering the importance of the commissioning of the foam to achieve his full function and the time constraints of the project, this solution could not be implemented within the project. This material provide nonetheless huge potential in terms of performance/cost ratio for the future, in comparison with the main stream contact solution based on ceramic pastes.The EVOLVE cell architecture could be successfully up-scaled to a size of 90 mm x 100 mm (Figure 3) and could be successfully integrated into 1 cell stack (Figure 4). The tested 1 cell stack could show a power density up to 170 mW/cm² at 0,7V and 750°C with simulated syngas as fuel at the anode and air at the cathode. The stack could be successfully operated for more than 100 hours under constant current load. At that point, the feasibility of the EVOLVE cell architecture at the stack level. Further work would require assembly and tests of short stack fur further assessment, as well as implementation of technological improvements in the cell.
Figure 3 Evolve cell for stack assembly (left), SEM cross section (right)
Figure 4 i-V characteristic of a 1 level stack at different temperature
Among several improvement routes, one could list following guidelines to improve durability and performance of cells:
- Stabilizing the interface between the ionic conductors in the multi-layer electrolyte, especially from the nano-porous supporting layer. This could be performed by increasing the gap between the manufacturing temperature (annealing, PVD deposition) and the operating temperature. Use of sol-gel process might be beneficial to stabilize the microstructure.
- Enhancing contact of the cathode with the electrolyte, by sintering the cathode at low temperature. The activation of the surface of the electrolyte by plasma etching showed improvement of the adhesion of the cathode and deserves further improvement.
- The set of electro-catalysts to be used as functional anode material is still subject to improvement as infiltrated nickel and La0,1Sr0,9TiO3-α are known to react under certain conditions with each other which hinder the performance. Pd could be an alternative.
Potential Impact:
At current technology, extrapolated to large scale production (ca. 10 MW/year) the cost for cells at stack level is estimated below 2500 € / kW, with perspective to reduce well below 1000 € / kW. The main cost drivers being the yield mainly governed by the surface quality of the substrate, the cost of the substrate itself and manufacturing of the electrolyte layer. Performance increase through further component improvement as well as automation of the PVD process might bring the cost down to a value comparable to the cost estimation of other competing ceramic cell architecture like ASCs.
Respect to the manufacturing route, the production of the substrate with a smooth surface remains the key challenge for the successful manufacturing of complete cell. The yield in terms of manufacturing for the stackable cells was around 21% for the premium quality. Improvement in terms of quality and performance could be achieved through implementation of a rationalized manufacturing route. It appears from modeling and experimental work that, with the current set of materials, there is an upper temperature limit above which keeping flatness of the substrate is difficult. This upper limit ranges between 1000°C and 1100°. Therefore it is important to design a rational and smart manufacturing route at low to mild temperature to produce the cells. The production of an advanced metallic substrate with tope pore size below 30µm would be highly advantageous for manufacturing at higher yield the cell.
The extended testing at operating conditions, including redox and thermal cycling has shown excellent stability of the composite EVOLVE substrate consisting in a mixture of LST and NiO infiltrated in a NiCrAl foam structure. The structure seems unaffected by testing and conductivity of the substrate is not deteriorated. Contacting also did not appear as a critical issue. Nonetheless the evolution of the interface with the bi-polar plate is subject to caution as an oxide scale growth was noticed when contact with CroFer 22 APU was used as interconnect material. The protection against corrosion of the steel surface even at the fuel electrode appears meaningful.
The manufacturing route was divided into module for each elementary component of the cell. The maximal firing and sintering temperature achieved overall did not exceed 1000°C, and were always performed in air atmosphere which contrast significantly with the usual hydrogen atmosphere used in the processing of standard metal supported cells. This modular approach of cell manufacturing route is expected to provide subsequent advantages for future development, by offering a huge flexibility in terms of used materials. This would for example allow the implementation of optimized materials designed for specific application, reduce the time to prototype for new advanced materials, and provide significant possibility for cost reduction by offering flexibility in the coating technology to be used.
List of Websites:
Dr. Rémi Costa, Deutsches Zentrum für Luft- und Raumfahrt e.V. Institute of Technical Thermodynamics - Electrochemical Energy Technology. Pfaffenwaldring 38-40, 70569 Stuttgart, Germany.
Tel: +49 711 6862 733
Fax: +49 711 6862 747
E-mail: remi.costa@dlr.de
Project website address: www.evolve-fcell.eu