New all-European high-performance stack: design for mass production

Final Report Summary - NELLHI (New all-European high-performance stack: design for mass production)

Executive Summary:
NELLHI has been a collaborative effort to produce a new, all-European, high-performance SOFC stack, designed for mass production. NELLHI combines European know-how in single cells, coatings, sealing, stack designs and manufacturing technology to produce an innovative and modular 1 kW SOFC stack. In this project, improvements over the state of the art in cost, performance, efficiency, and reliability have been proven, as a combined result of high-performance cells and manufacturability designed for mass production at high yield. The NELLHI stack operates at reduced temperature, with unprecedented performance. Within the project, the stack module has been developed over 3 successive generations according to system integrators’ requirements guided by an industrial advisory group. The target application of the developed stack module is stationary and residential combined heat and power production, and forms the basis for Elcogen Oy’s commercial SOFC stack value proposition, for system integrators to use. A number of follow-up projects and applications have taken up the NELLHI results for further development and maturation. Increased market awareness and penetration has been achieved for component manufacturers Elcogen AS, Sandvik, Borit and Flexitallic. All manufacturing methods, stack designs, and materials are tailored to mass production to enable 1000 €/kW profitable stack price at 50MW/year production.
To achieve this high-level objective, a systematic approach has been undertaken, focusing on each key component of the stack assembly and establishing sub-objectives for each part. This has led to the following achievements:
• Cell development
o Upscaled the unit cell from 100x100 mm2 size to 120x120 mm2 to increase yield rate and specific performance
o Tailored and improved the production process to achieve unit cells capable of reaching 900 mV cell voltage with 0.4 Acm-2 current density at 650 ºC operating temperature and over 75% fuel utilization in a stack environment.
o Validated cell performances and investigated cell processes in operand with innovative techniques, for analysis of critical conditions and parameters
• Interconnect development
o Consolidated the once-through roll-to-roll process of pre-coated steel sheets for massive supply of ready material for interconnect manufacturing without further surface treatment
o Optimal substrate and coating combination selected for the NELLHI stack in terms of operating performance and cost
o Designed and optimized interconnect geometry for the NELLHI stack assembly focusing on minimizing leakage and weight
o Consolidated the hydroforming process for rapid shaping of the interconnect from supplied pre-coated steel sheets, based on optimized design
o Adapted cutting, welding and leak-testing techniques for maximum manufacturing effectiveness and speed
• Sealing development
o Developed and optimized gasket material, thickness and composition based on the hybrid sealing concept, leading to a new product patented
o Validated sealing performance in terms of leakage and conformability in stack environment
o Improved sealing designs for quick manufacturing and easy handling in stack assembly line
• Stack integration
o Incorporated the successive improvements of above components in 2 design-freezes before confirming the optimal design, assembly and integration
o Validated stack performance in representative and aggravated test conditions.

Project Context and Objectives:
Accelerating the transition to a low-carbon, competitive economy is both an urgent necessity and a tremendous opportunity for Europe. It is a central challenge of our time. Failure may put our welfare at stake. Success would open unprecedented economic opportunities and new avenues to prosperity, welfare and growth. Energy efficiency is the great, invisible source of energy that drives creativity, sustainability and economic independence: it is the energy that is saved, the primary source that isn’t consumed, and the trigger to smarter solutions for industry, transport and buildings. Europe's businesses are ahead in many world markets – where competition from global competitors is growing – and European scientists and innovators are pushing the frontiers of knowledge at the service of this inexhaustible energy resource.
The transition to a low-carbon, energy-efficient and climate-resilient economy will certainly require a more decentralised, open system with the involvement of all society. The energy system has traditionally been marked by the dominance of large companies, incumbents and large-scale, centralised technological projects. But in the future the consumer has to be at the centre of the energy system: demanding competitive low-carbon solutions; participating as producer and manager of decentralised energy networks; acting as an investor, through decentralised platforms; and driving change through user innovation. More involved citizens take greater responsibility for their own and the EU's energy security.
The European Union is committed to the mission described above through pursuit of the so-called Energy Union, foreseeing a growing integration of European Member States in terms of energy supply, exchange and utilization, embracing 5 dimensions:
• Energy security
• A comprehensive, free internal market for energy
• Energy efficiency
• Decarbonization of energy
• Pushing European competitiveness through research and innovation
What Solid Oxide Fuel Cells – SOFC – can do
Solid oxide fuel cells (SOFC) are a cutting-edge technology for converting the chemical energy in hydro-carbon fuels, of both fossil and renewable origin, directly to electrical power and heat, avoiding the inefficient and polluting stage of combustion, by means of an electrochemical reaction.
The SOFC is the most efficient device available for power generation: 60% net electrical efficiency for a small, 1 kW system fed with natural gas has already been achieved, and 70% is close at reach, which is double the efficiency of the most efficient combustion engines of the same size and with less than a tenth of the harmful emissions produced by burning any fuel.
And what is more, the SOFC can be reversed to work as an electrical power storage system, allowing to capture excess renewable power (for example generated by wind turbines or solar PV panels) in the form of chemical energy: in other words, regenerating a fuel for future use, sustaining the energy cycle.
The SOFC is a high-tech solution, integrating cutting-edge chemistry, material science, engineering, process control and manufacturing, all fields where Europe is at the forefront of development. The NELLHI project shows how SOFC technology can engage audacious, knowledge-based enterprises and forward-looking multinationals to drive European competitiveness in high value-added, strategic supply chains such as providing clean, cost-effective and smart energy to the end users of tomorrow.
NELLHI combines European know-how in single cells, coatings, sealing, stack designs and manufacturing technology to produce an innovative and modular 1 kW SOFC stack. In this project, improvements over the state of the art in cost, performance, efficiency, and reliability have been proven, as a combined result of high-performance cells and manufacturability designed for mass production at high yield. The NELLHI stack operates at reduced temperature, with unprecedented performance. Within the project, the stack module has been developed over 3 successive generations according to system integrators’ requirements guided by an industrial advisory group. The target application of the developed stack module is stationary and residential combined heat and power production, and forms the basis for Elcogen Oy’s commercial SOFC stack value proposition, for system integrators to use. A number of follow-up projects and applications have taken up the NELLHI results for further development and maturation. Increased market awareness and penetration has been achieved for component manufacturers Elcogen AS, Sandvik, Borit and Flexitallic. All manufacturing methods, stack designs, and materials are tailored to mass production to enable 1000 €/kW profitable stack price at 50MW/year production.
To achieve this high-level objective, a systematic approach has been undertaken, focusing on each key component of the stack assembly and establishing sub-objectives for each part:

Cell development
o Upscale the unit cell from 100x100 mm2 size to 120x120 mm2 to increase yield rate and specific performance
o Tailor and improve the production process to achieve unit cells capable of reaching 900 mV cell voltage with 0.4 Acm-2 current density at 650 ºC operating temperature and 75% fuel utilization in a stack environment.
o Achieve robust cell structures resistant to long-term operation and thermal cycles
o Validate cell performances and investigate cell processes in situ for analysis of critical conditions and parameters
Interconnect development
o Consolidate the once-through roll-to-roll process of pre-coated steel sheets for massive supply of ready material for interconnect manufacturing without further surface treatment
o Select the optimal substrate and coating combination for the NELLHI stack in terms of operating performance and cost
o Design and optimize interconnect geometry for the NELLHI stack assembly focusing on minimizing leakage and weight
o Consolidate the hydroforming process for rapid shaping of the interconnect from supplied pre-coated steel sheets, based on optimized design
o Adapt cutting and welding techniques for maximum manufacturing effectiveness and speed
• Sealing development
o Develop and optimize gasket material, thickness and composition based on the hybrid sealing concept
o Validate sealing performance in terms of leakage and conformability in stack environment
o Improve sealing designs for quick manufacturing and easy handling in stack assembly line
• Stack integration
o Incorporate the successive improvements of above components in 2 design-freezes before confirming the optimal design, assembly and integration
o Validate stack performance in representative and aggravated test conditions

Project Results:
In the following sections, for each work package, a summary of achievements in scientific and technological progress will be given as regards the stated objectives. The NELLHI project has put together the efforts of specialized industries focusing each on a specific, key component or aspect of the finalized stack.
2.1 Work packages and foreground overview
In Work Package 2, Elcogen AS has led the development of single cell manufacturing processes, improving material formulations and modernizing equipment to achieve the lowest scrap rates possible maintaining the highest standards for cell quality, performance and reliability. Thus the project target of 95% effective yield was achieved and surpassed, reaching 96%. ENEA supported the cell development process by conceiving and realizing a highly innovative set-up for locally resolved, in-depth, in-operando electrochemical characterization of produced cells.
In Work Package 4, Elcogen Oy, Sandvik and Borit worked together to improve the design of the interconnect plate, as well as tailor the mass-manufacturing processes to this design. Sandvik, with support from VTT and ENEA, has gone through extensive material combinations (of substrate and coating) with in-depth analysis of their behaviour in stack-relevant environment, which has led to radical new insights (self-healing properties as well as unforeseen diffusion phenomena). Taking the pre-coated, rolled coils of interconnect material from Sandvik’s manufacturing line, Borit has optimised its Hydrogate forming process for larger throughputs, achieving over 99% yield rates. The bottleneck in this process however remains the welding of different layers of extremely thin, shaped plate, where 96% of yield was nevertheless achieved.
In Work Package 5, Flexitallic has formulated a breakthrough sealing material, that combines the high-temperature resistance and excellent sealing properties of their standard Thermiculite with previously unimaginable compliance. This new formulation, patented during the NELLHI project under the name CL87, has been successfully tested in collaboration with VTT to assess different application means, including glass coating and design optimisation to reduce the number of manufacturing steps required as well as improve manageability of the finished, cut seals.
Finally, in Work Package 3, Elcogen Oy has been putting together all the developments above in the high-performance stack that gives the name to the project. This entailed much design optimisation of interconnects and seals, assembly and conditioning procedures, stack characterization protocol definitions and testing, deconvolution of the ultimate, compound stack performance into the contributions of each component. Elcogen Oy was supported by CUTEC and VTT in this evaluation, and numerous stacks were tested in several conditions, yielding comprehensive data for the assessment of the progress achieved in the successive stack generations.
Work Package 6 was dedicated to dissemination and communication of project results, and preparation for their exploitation after the project. Everything was coordinated in Work Package 1.

2.2 Work package 2 – Cells
In order to increase the power per cell and to decrease the final price of SOFC stack, dimensions of Elcogen’s cells were changed from 100×100 mm2 to 120×120 mm2, that allowed an increase in electroactive surface up to 49.4 % per cell. These changes consist of optimization of tape-casting, screen-printing, sintering and overall manufacturing processes.
Cell manufacturing process itself consists of several steps, and every step needs to be optimized for different cells geometry and structure. This optimization was done based on analytical data (TGA, DSC, SEM etc.) and on series of experiments to adjust equipment features to achieve the highest yields possible.

Task 2.1 Cell manufacturing and optimization of process parameters
Thanks to the NELLHI project, the main objective for cells manufacturing partner Elcogen AS was successfully achieved. At the same time, electrochemical efficiency was kept at the same range, even increased after manufacturing process optimization.

Moreover, in the frame of the project, Elcogen had a possibility to optimize cells production increasing productivity and minimizing product variation and rejection rate at the same time. The yields rose at every step, decreasing cells cost and scrap quantity.

These results were obtained after a series of experiments optimizing tape-casting, screen-printing and sintering parameters. Each cell passed quality assurance control confirming successful manufacturing parameters.Also new manufacturing rules in production area combined with special coats allowed to implement ISO7 clean-room and to decrease the overall influence of dust during all manufacturing steps.
Result to reduce scrap rate below 5 % in GT production was achieved. Cost target for cells was also achieved and it is 7 € /cell which makes 280 €/ kWe . Further increasing of yields and decreasing in price are possible after complementary raw materials purification because of different impurities decreasing the yields. This step may be done by implementing supplementary raw materials treatment at Elcogen AS or by changing the supplier.

Another factor to achieve better results is installation of automated printing lines to improve reproducibility, production capacity and to decrease influence of human factor. In these cases, the final yields may exceed 96 % with considerably reduced price per cell and higher quality with better reproducibility at the same time.

During the NELLHI project, Elcogen AS delivered in total 85 cells of various dimensions.

Task 2.2 Cell characterization

This Task was supposed to be concluded early in the project as a quality assurance of the larger cells being produced by Elcogen AS. However, since this reliability in manufacturing was already proven at the beginning of the project, a new, more challenging objective was agreed upon by the consortium as regards cell testing. Indeed, current, conventional techniques for cell characterization have limited resolution, relying exclusively on “black box” measurements of the single cell (current, voltage, and chemical/thermodynamic conditions at the inlet and outlet only).
ENEA was charged therefore to develop an innovative set-up where the conditions (at the fuel electrode in particular) could be measured locally, in-situ and in-operando, generating data that can be used for detailed mapping of cell performance distribution, granting insight into process evolution across the cell (also over time), as well as enabling numerical model validation. This set-up has already been shown and explained in the first reporting period.

The ELCOGEN ASC 400B single cells (12x12 cm2 and 11x11 cm2 active area) produced in this last phase of the NELLHI project were characterized using the innovative setup developed in ENEA research center, which is able to perform an in-depth characterization of the anode surface through a localized sampling of the temperature and gas composition of an operating cell, without afflicting the cell performance.

The innovative setup, described in detail in the Deliverable 2.2 was improved to solve the surface oxidation of the Crofer mesh employed as the anode current collector that provoked the internal resistance increment shown in the first generation of tested cells. Employing gold mesh as anodic current collector, the internal resistance of the cell was reduced, as confirmed by measurements.

Differently from the first validation of the set up reported in Del 2.2 after the system improvement, it is clear that the response of the 120x120 mm2 single cell is in very good agreement with the behaviour of a 120x120 mm2 cell inside the 15-cell stack: the slightly lower open circuit voltage (OCV) is due to the inherently less gas-tight assembly in a single-cell configuration as compared to a stacked assembly. Considering the less favoured conditions of a single cell exposed to interfaces to the environment on all sides, as compared to a cell sandwiched between other cells in a stack assembly, the validation is considered excellent.

Temperature and gas sampling analysis
The gas sampling analysis presented in the first report (Del 2.2) were affected by errors most likely due to condensation of water within the sampling capillary and/or infiltration of ambient air during the sampling. Insulation of the gas sampling lines have been improved, providing more reliable results from the gas analysis.
The same compositions presented in the Deliverable 2.2 were tested using the improved setup with the second generation 12x12 cm2 single cell. Hereby only the result obtained for the reformate natural gas composition are presented, since the usage of natural gas is the final application of the NELLHI project.
In accordance with what stated above the composition employed in the tests reported in this document was fed to the cell operating at OCV and with current loads (165 mA/cm2 and 330 mA/cm2) corresponding to 27% and 54% of fuel utilization, respectively.
In order to monitor a thermal equilibrium, the temperature analysis was carried out for 3 hours for each condition.
Results confirmed that the chemical gradient variation could be considered negligible along the vertical axes (across the direction of flow). For this reason, all the following assessments about the composition analysis are referred to the direction of flow.

The axial development of the composition, under OCV condition, shows how CO2 and H2O compete in CH4 reforming leading to the presence of two different types of methane reforming: Steam Reforming and Dry Reforming (Equations 1 and 2), producing in both cases H2 and CO as reaction products.

CH4 + H2O ⇌ CO + 3H2 〖〖-∆H〗^0〗_298=-206 kJ/mol [1]

CH4 + CO2 ⇌ 2CO + 2H2 〖〖-∆H〗^0〗_298=-247 kJ/mol [2]

Observing the component gradients seems that the dry reforming reaction is dominant respect to steam one, due to a major consumption of CO2 respect the consumption of H2O going from the inlet to outlet. Furthermore, the increment of CO and the decrease of CO2 seems to underline a not key role for the Water Gas Shift reaction (WSG) in OCV condition.
As current is drawn, corresponding to 165 mA/cm2 of current density, some changes emerge in the axial distribution of compound concentrations: CO increases (less markedly than at OCV) and at the same time the CO2 presents a particular trend showing an initial decrement in the first half of the anode surface associated to the presence of the dry reforming reaction, and a slight increase in the second half, showing the occurrence of the water gas shift reaction (Equation 3).

CO + H2O ⇌ CO2 + H2 〖〖-∆H〗^0〗_298=41kJ/mol [3]

H2 + O2- → H2O + 2e- 〖〖-∆H〗^0〗_298=286kJ/mol [4]

The behavior of H2O changes with respect to OCV condition, increasing as an effect of the electrochemical oxidation (Equation 4). H2 presents a slight increment despite of the electro-chemical oxidation that should leads to its decrement, demonstrating that reforming reactions are still predominant compared to the electrochemical oxidation. Regarding to CH4 behavior, it shows the same trend present in OCV condition. The chemical trend changes completely when double the current (40A) is drawn, with the exception of the CH4 that shows an identical trend as for the in OCV and 20A cases, suggesting that the reforming reaction is unaffected by the current load.
The effective consumption of H2 from the inlet to the outlet coupled with the increment of the H2O, confirms the strong relevance of the electro-chemical oxidation reaction in this condition. It is possible to note also a net decrement of CO coupled with an increase of CO2, pointing to a greater contribution of the WGS reaction (possibly also CO electrochemical oxidation).

The local temperature analysis reflects the results obtained in the composition analysis. In OCV condition it is possible to observe a general temperature decrement due to the endothermic nature of the internal reforming reactions, reflecting the result obtained for the composition analysis.
At 20A the thermo-dynamical equilibrium, represented by the curve plateau, is achieved after one hour in overall the anode surface. The slightly exothermic trend indicates that, despite the endothermic contribution of the internal reforming, the effect of the exothermal reactions (WGS reaction and electrochemical oxidation of H2) is dominant for this value of fuel utilization.
The temperature evolution over time on the anode surface under 40A of current shows a more marked exothermic trend respect to 20A, highlighting the predominant contribution of exothermic reactions respect to endothermic ones. Moreover, it is possible to observe that the increment of temperature in the sampling points located close to the anodic inlet is greater than the increment of those located close to the anodic outlet. This non-homogeneous temperature, risen inside the cell, was also verified in a previous results obtained and as is explained in Deliverable 2.2 the reason should be sought in the depletion of H2 due to the evolution of the electrochemical oxidation, provoking a non-uniform current generation (and therefore heat generation) over the anodic surface, decreasing from inlet (rich in H2) to outlet (depleted of H2).

The temperature variation on the anode surface is visualized by the temperature measurement distribution in the three conditions studied. In fact, in OCV condition the lower temperature is localized near the inlet zone, in accordance with the faster kinetics of the steam and dry reforming. Under 20 A condition, the temperature inside the cell seems to be very homogeneous reporting only a ΔT of 4°C, showing that a faster equilibrium between exothermic and endothermic reactions is achieved. At 40 A the contour graph confirms a non-homogenous temperature distribution between the inlet zone and outlet zone., due to the dominance of the exothermic reactions.

All in all, the innovative experimental set-up has proven to provide radically innovative measurements and insights into the anode processes of the SOFC.

2.3 Work Package 3 – Stack design, assembly, and evaluation
The final target in this WP was the proof of concept of a new Elcogen 1 kW stack. Performance target of the stack was set to 900 mV at 0.4 A.cm-2 with fuel utilization capability measured to be 85 %, operation temperature 650 °C and pressure losses at both gas compartments less than 5 mbar. Testing should have been at least 4000 h indicating less than 0.2% / 1000 h degradation. Furthermore, stack degradation in at least 10 thermocycles should be tested with less than 0.5 % leakage rate and 0.5 % degradation. In additions, a 10 kW power class stack module should have been designed and verified. Module verification could be done in reduced power class. The final results are summarized as follows.

Unit cell footprint 120 x 120 cm
Amount of unit cells >10 1 kW
Stack operating temperature 600 – 700 °C
Current density @ 0.9 V >0.41 A.cm-2 (0.92V @ 0.4 A.cm-2)
Fuel utilization capability 91 %
Pressure loss on fuel side 4.5 mbar°
Pressure loss on air side 3.8 mbar°
Leak rate after 10 cycles <0.5%
Degradation 0.9 %/1000h
Degradation after 10 thermocycles 0.6 % /10 TC
Stack production yield 97 %
Stack price 1000 €/kW price level is estimated to be reached at around 50 MW annual production volume
Stack module Designing conducted and module validated in 6 kW power class

The 5-cell stacks were tested by VTT. The 15-cell stack testing was conditioned by Elcogen and tested by CUTEC. The 39-cell stack was conditioned by Elcogen and tested by VTT. In addition to the results shown in this report, Elcogen and VTT conducted testing for altogether 17 stacks during the project period.

All stacks were showing above 900 mV mean cell voltage. At 0.4 A.cm-2 mean cell voltage was 916 and 925 for Elcogen and Sandvik post-coatings, respectively. CUTEC tested Elcogen 3rd generation stack with Elcogen coating for its fuel utilization capability. Stack voltage was stable in the measurement up to 91.4 % fuel utilization level.
CUTEC and VTT further tested Elcogen’s 2nd generation stacks for their efficiency. Tests were conducted at VTT with 1 kW stack (39 cells) and with 15 cell stack at CUTEC. In both places, a hardware-in-a-loop testing concept was utilized for a system. The gas composition matrix was predefined for operation conditions and these gas compositions were mixed with mass flow controllers and at VTT fuel reformer was placed between gas feed and stack inlet. The gross efficiency is determined as the electrical power output divided by the chemical input power of natural gas. Test conditions were: inlet air flow temperature = 590 °C, fuel = natural gas, reformer outlet temperature = 600 °C, current = 30 A, fuel utilization = 87 – 90 %, oxygen utilization [air] = 22 %. Stack was located in a furnace with set point temperature of 600 °C. Stack was measured to reach at it nominal operating conditions gross efficiency of 74 %-LHV.

Elcogen stack durability for thermocycles was tested by Elcogen. Testing included three phases: thermocycles down to 300 °C (TC 1 – 5), thermocycles down to 50 °C (TC 6 – 8), and complete removal of the stack from the test system (TC 9) and reinstallation and repeating phase 1 (TC 10 – 14). During thermocycles, 1/1 H2 / N2 mixture was supplied to anode and air to cathode. Heating and cooling rates were 60 K/h. The leakage level and the polarization of the stack was tested after each thermocycle. Test conditions for the polarization measurement were: fuel flow rate (H2) = 0.52 lN/min/cell and (N2) = 0.52 lN/min/cell, flow rate (air) = 2.2 lN/min/cell, furnace temperature = 650 °C, current ramp rate = 1 A.min-1 maximum current 30 A. At 30 A, the stack voltage was let to stabilize for 60 min in order to determine degradation after which current was ramped back to 0 A. Stack voltage was let to stabilize after the polarization measurement for 60 min after which open circuit voltage was recorded in order for leakage level analysis. The leakage level was calculated by using the Nernst equation which revels the ratio between hydrogen and steam, calculated to Fig. 14 (right). As the inlet humidity is known for the anode side (0.6 hPa), the increase in humidity level is resulting from the oxygen leakage and combustion with H2. It is assumed here that nitrogen is leaking in the ratio of 79/21 to the anode and thus the leakage between the cathode and anode is calculated as V(leak) = [p(H2O)_OCV – p(H2O)_inlet]*V(H2) / [2* p(atm)*x(O2,air)] and the relative leakage as leakage = V(leak)/V(fuel, tot). The relative oxygen leakage is depicted in Fig. 15, left. As the leakage can be determined from the open circuit voltage and that is increasing in thermocycles, leakage value is decreasing in thermocycles. The leakage value was below 0.4 % during the 14 thermocycles. The stack degradation was determined from the stack voltage measured at 30 A after each thermocycle and compared to the voltage before the thermocycles (908 mV). After 10 thermocycles, stack voltage was decreased 0.6 % compared to the initial value.

The durability of Elcogen stacks were tested by VTT for five different stack types. Stack degradation was determined between 20 h – 1100 hours as a linear fitting to the voltage data resulting in degradation value of 0.9% / 1000 hours. After thermocycle conducted at 1100 hours, voltage decay rate increased as the degradation rate was 1.8 % / 1000 hours for the remaining test period. It was noticed that voltage decay could be at least partly explained by reformer degradation. It was seen that H2 content was decreasing and CH4 increasing. Reformer heating was increased after 2400 hours and 4000 hours where stack voltage is seen to be increased. Stack resistance against voltage decay was tried to increase by adding protective coatings to the interconnect parts. In the first experiments, pre-coated AISI 441 was tested, i.e. interconnect was manufactured by Borit out of Sandvik’s pre-coated steel. Two different conditioning processes were studied, i.e. Elcogen standard conditioning process (G2.1.1) and high temperature conditioning process (G2.1.2). In the second experiments, post-coated Crofer was used, i.e. interconnect was manufactured by Borit and Elcogen (G3.0.1) and Sandvik (G3.0.2) applied coatings with their internal processes on the ready interconnect plates. Both the initial voltage at nominal operation conditions (NOC) and degradation rate with the pre-coated steel stack (G2.1.1 and G2.1.2) was high, over 15 % / 1000 hours. Test was terminated after 750 hours, both stacks were opened and post-mortem analysis conducted. Post mortem analysis revealed that the interconnect structures were severely corroded in both stacks. Post-coated stacks were operated for 1500 hours, until the end of the project, and 3 % / 1000 hours degradation was measured for both stacks. Post mortem analysis are not yet conducted for these stacks.

Elcogen stack price against manufacturing volumes was estimated based on the values received from the project partners and raw material suppliers. The prediction contains many uncertainties for each individual component and stack manufacturing steps but it is concluded that Elcogen stack price can reach 1000 €/kW level around 50 MW annual production volumes.
Elcogen stack production yield was calculated from the stacks produced for the project including 30 stacks. One stack was broken during conditioning process and reason shown to be surface imperfections of the current collector structure. New pre-treatment process was developed for the part and taken into production use. The production yield is thus 29 / 30 = 97 %. It should be noted that as the production volume is small, production yield figure is statistically not well defined.

A concept for a 6 kW stack module based on Elcogen stacks was designed. Nominal power of each stack in the tower is 3 kW. Gas manifold is located between the stacks and the tower is located inside a metallic cylindrical compartment. The tower concept has been tested by VTT in their real fuel cell system.

2.4 Work Package 4 – Interconnect manufacture and Coating
Task 4.1 Steel preparation, coating development and manufacturing

During the 1st reporting period, two less expensive steels AISI441 (EN 1.4509) and Outokumpu 4622 (EN 1.4622) were selected to be used in combination with precoating concept. Sandvik Materials Technology has developed a PVD (physical vapor deposition) process with roll to roll type continuous evaporation coating of full width strip steel. The process significantly shortens the value chain for bipolar plate/interconnect production. It also ensures homogenous coating of the whole strip surface and repeated quality from batch to batch. The coating process consists of the following steps:
Cleaning/inspection. Cleaning is important for good adhesion. To avoid harmful defects, the whole area of the strip surface is inspected – both sides.
Coating. Coating of metal layers by Sandvik's continuous evaporation coating process.
Inspection. Automatic X-ray inspection devices measure the thickness and quality of the coating.
Testing, slitting and packaging. After coating, the steel strip is inspected and tested (for example tensile, hardness and surface properties). The process is completed by slitting to the required width, control of burr height, width and shape, and finally by packaging and shipment.

Interconnect coatings were tested by e.g. stretching the substrate and by observing self-healing of the coating: after 500 h at 800 °C the coating has diffused over the FIB-milled trench and covered it completely.

Coated interconnect materials were tested for endurance both at ENEA and at VTT. The experimental campaign included oxidation measurements in air, cyclic oxidation tests in air and dual atmosphere tests both in H2/air and syngas/air.

From the endurance tests it can be concluded that the most promising Co/Ce-coated steel material for solid oxide fuel cell interconnects would be Crofer 22 APU. It had best coating adhesion and thinnest oxide layers. Also the formation of MnCo-spinel was one of the benefits of Crofer 22 APU. Outokumpu 1.4622 steel showed significant grain growth over a period of 15 000 hours, which could be a problem in long term operation. Of course, Crofer 22 APU is the most costly material and therefore stack manufacturers need to weigh the lifetime against cost when making a decision between Crofer and less expensive materials. For more details, see deliverable D4.4 Summary of steel pre-coating materials, characterization and coating and manufacturing process. The endurance testing will be continued until 08/2017 and the final outcomes will be reported in a dedicated publication in scientific journal.

Task 4.2 Interconnect shaping and manufacturing

During the 1st reporting period, Borit and Elcogen Oy designed and manufactured the forming and welding tools for interconnect production. The process was iterative and quality of the complete interconnect assemblies were analysed after each change to the tooling and process. The manufacturing processes and tooling designed and produced during 1st reporting period were then further developed and streamlined over 2nd reporting period to enable higher quality interconnect manufacturing.

Borit continued to develop the production processes for the interconnect plates during the 2nd reporting period. The Hydrogate forming process is now very stable for the production of interconnect plates as shown in the production chart below. The average yield for a production batch of almost 40000 parts is 99.41%.

Borit also formalised the welding set-up process. A check-list has been created based on a DOE to determine cross-influence between the various welding parameters/settings and lessons learned from previous welding process experiences, to formalise and document the set-up procedure. The result is that the welding process has become more mature but still remains less stable and robust as needed for high quantity production.

2.5 Work Package 5 – Sealing
Flexitallic is a well-established manufacturer of sealing materials and gaskets that are used in a variety of industries. The Company has a history of over 100 years of innovation. One of the innovative materials that Flexitallic has developed is Thermiculite. This was developed to combat a technical shortcoming with graphite that was well known in the market. Thermiculite has proven itself as a sealing material in a wide range of applications as it is temperature stable and resistant to a wide range of chemicals.
Once established in the industrial market, this proprietary material was modified to form a polymer free compression seal for the growing SOFC market. This material was known as TH866 and has been tested and evaluated by numerous SOFC research groups and is in use in several places.
As one of the key elements in the NELLHI project Flexitallic was challenged to provide a new material that had the sealing and thermal resistance properties of Thermiculite 866 but was softer and more compliant, leading to sealing at lower applied loads.
This high degree compression was useful for ceramic fuel cells where high loads could not be applied to the seal or gasket. Thermiculite 866 is a compression seal, and therefore required the application of compression, or load, to affect a seal. A further advantage of this high compression is that it allows the gasket to effectively compensate for some manufacturing tolerances elsewhere in the unit cell. This is obviously more important as the number of unit cells per stack increases, and also a relaxation in tolerances could help cost improvement initiatives.
A further element was this gasket material had to be compatible with the glass coating previously developed by VTT and Flexitallic, and sometimes used by Elcogen to produce a composite or hybrid gasket.
As described in the mid-term report, the development of a new compression gasket material has been successful. This material was known by the experimental designation CL87, but now it is being transitioned to launch as a commercial product, and it will be known as TH870.
As discussed above, the key initial characteristic was compression, and the need to improve this. A ten-fold improvement was achieved thanks to the new forulation: TH866 compressed 0.022mm under a load of 1MPa whereas TH870 compresses 0.20 mm. This is consistent with the material being considerably softer.

Clearly sealing must not be compromised when improving the compression: TH870 is as good as or slightly better than TH866 across a range of gasket stresses.

The long term test data show how sealing is maintained over the length of the test, with leak rates consistent over 1000hrs. TH870 has a lower leak rate than TH866, therefore a better seal.
This high temperature test work was carried out on equipment designed by VTT specifically for this purpose as part of the NELLHI project.
During this development work sample runs were made of TH870 and a variety of thicknesses (measured not by thickness but by weight per unit area), and shown that this material can be made consistently on Flexitallic's current prototyping equipment.
Gaskets were supplied to Elcogen to build in to stacks and as part of WP3 Elcogen reported that the leakage of gas through the seals was <0.2%, which exceeds the start of project state (1.0%) and the project target (0.5% leakage).

During scale up work it was noted that the glass coating required some improvement to allow consistent processing. This was because the material was curing too fast and didn’t allow for effective or efficient processing, creating a lot of expensive waste material. A small reformulation was required to add a retarder to the formulation, but importantly the retarder had to be of a chemistry that did not compromise the cell chemistry. After the addition of the retarder, the coating formulation now has a good “pot life” meaning it can be used over an extended period of time, creating less waste.
During the testing of TH870 in stacks some degradation of the interconnect was observed by Elcogen, this was found to be related to sulphur, which in turn was traced back to the gasket. Investigations showed that the sulphur was originating in the carrier material used to make the TH870. A low cost method was found to overcome this by substituting the substrate used in gasket manufacture and subsequently this issue has not reoccurred.

The cell design requires the sealing element to have different compression characteristics in different areas of the seal. As development progressed, this was initially achieved by combining two TH870 gaskets of different thicknesses together in the stack. One of the gaskets is very thin, 0.3mm and can break during handling.
In order to facilitate improvements in stack manufacture by allowing robotic assembly, combining two gaskets into one would be beneficial, as one thicker gasket would be more robust when handled. This was a suggestion put forward by other partners within the project. Using one gasket instead of two can also cut down on the time taken to construct the stack.
Casting TH870 with different compression values within it is not possible. Laboratory work was carried out to investigate the possibility selectively compressing some areas of the gasket. While this worked to some degree in the laboratory, it was subsequently found that the in service compression of the pre-compressed areas would not be enough and also that this would be extremely difficult to replicate on a larger scale.
It was therefore determined that the only practical route forward was to stick two different thickness gaskets together. Possibilities to do this included using the pre-existing glass coating and CEV, one of the components of the gasket material. Early in the project it was determined that glass coating the two individual gaskets was beneficial in improving sealing and improving leak rates due to thermal cycling. This was achieved by the glass coating melting at the preconditioning step, eliminating any possible leak paths between the two gaskets. How important this glass coating was when the two gaskets were glued together using the CEV was not known. Several different combinations were produced with and without glass coating.

The leak rates of different combinations of laminated gasket show how the leak rate changes during thermal cycling, providing a good indication of gasket performance. When no glass is present between the two layers and CEV is used to glue the two layers together, the leak rates are comparable at high temperature and are the same as those where glass is used between the two gaskets. The worse combination was where CEV and glass were present in between the gaskets. Therefore it was concluded that CEV could be used to glue the two gaskets together.
Further work would be required to determine if this would provide sufficient sealing around the actual cell, which is the primary reason for applying the glass coating in the two gasket situation. This would include testing in a full stack.
Further production equipment and process development would be required in order to mass produce this laminated gasket.

Potential Impact:
Through the NELLHI cooperation, the project partners have been able to consolidate their positions as key component manufacturers in the European arena of SOFC technology. The focused activity in NELLHI has allowed bottlenecks in design and production processes to surface and be dealt with. Unexpected issues related to contamination through carrier belts, material diffusion and interaction mechanisms, leak testing and surface corrosion have all been identified and resolved in a plenary approach. This joint undertaking has led to a highly committed team of players that are fully aware of the interconnectedness of the development process and assembly of the highly sophisticated components that make up the NELLHI SOFC stack.
The most important evidence of the maturing supply chain has been the initiation of several “follow-up” projects, all within the framework and support of the FCH 2 JU, that have taken up the stack concept developed in NELLHI. Thus, INNOSOFC (Grant Agreement 671403) takes the NELLHI stack for optimization and implementation in a 50kWe CHP system, assessing the most promising market segments, DEMOSOFC (Grant 671470) demonstrates operation of three 50kWe SOFC systems fed with biogas from waste water treatment, one of which will be the INNOSOFC system, and qSOFC (Grant 735160) focuses on improving quality assurance of the mass-manufacturing processes of the NELLHI-INNOSOFC stack components, reducing costs. The intrinsic collaboration between these projects and their results were publicly presented at the final NELLHI stakeholder workshop, co-organised with the European-funded projects INNOSOFC, DEMOSOFC and HELTSTACK. This event (the SuperSOFC Workshop) was held in Stuttgart 31 January 2017, in conjunction with the FDFC conference, and the outcome was submitted as Deliverable 6.4.
In addition, the NELLHI activities and results have been actively disseminated and promoted via various channels, addressing the scientific community, as well as industrial uptakers and the interested general public, creating considerable awareness around the project and the project partners. The Layman’s Report, Deliverable 6.6 provides an eye-catching, informative and divulgating pamphlet for distribution, to summarize and showcase the achievements of the NELLHI project to a wide audience.
Finally, a joint article pooling contributions from VTT, Sandvik and ENEA on a systematic analysis of different interconnect-coating combinations in stack-representative conditions is being prepared for submission to a high-impact factor scientific journal, for which the last data are being gathered in the days of writing this report.

List of Websites:
www.nellhi.eu
stephen.mcphail@enea.it
Tel. +390630484926

Final Report Summary - NELLHI (New all-European high-performance stack: design for mass production)

Share this page

Download