Quality control of SOC stacks using electrochemical impedance spectroscopy

Abstract

The present invention regards a method of performing quality control (QC) of a solid oxide cell stack, the method comprising operating the solid oxide cell stack in electrolysis mode under a fuel side QC atmosphere comprising a reducing agent and an electrolyser feed at a QC current density of below 50% of the full current load; recording an electrochemical impedance (EIS) spectrum of the stack during such operation; and then comparing the obtained spectrum with a QC reference.

Description

[0001] QUALITY CONTROL OF SOC STACKS USING ELECTROCHEMICAL IMPEDANCE SPECTROSCOPY

[0002] TECHNICAL FIELD

[0003] The present invention regards a method of performing quality control (QC) of a solid oxide cell stack, the method comprising operating the solid oxide cell stack in electrolysis mode under a fuel side QC atmosphere comprising a reducing agent and an electrolyser feed at a QC current density of below 50% of the full current load; recording an electrochemical impedance (EIS) spectrum of the stack during such operation; and then comparing the obtained spectrum with a QC reference.

[0004] BACKGROUND

[0005] Solid Oxide Cells (SOCs) hold the promise of effectively converting substantial amounts of electrical energy into chemical energy, in the form of fuels like hydrogen, or more complex, energy-dense synthetic fuels containing carbon, which require catalytic upgrades outside the SOC system. To ensure market success, it is critical that the SOC stacks exhibit consistent performance and long-term stability. This necessitates the development of robust quality control (QC) procedures covering the entire production chain to reduce the risk that faulty stacks are provided to the market, ensure product consistency, identify and prevent product defects, and optimise production processes.

[0006] Efforts are ongoing to find industrially applicable QC procedures for identifying faults in solid oxide cell stacks in a cost-efficient and resource-efficient manner. In particular, electrochemical procedures using e.g. open circuit voltage (OCV), current-voltage characteristics (Ui), electrochemical impedance spectroscopy (EIS), and operation under constant current (CC) have been proposed.

[0007] Where the use of OCV, Ui and CC in electrochemical quality control is well-established for QC purposes, the use of EIS as a QC tool is still under development as a quality control tool.

[0008] When performing EIS, a sinusoidal modulation current is applied to the cell across a wide range of frequencies, and the resulting voltage's amplitude and phase are measured to generate an impedance spectrum, which provides detailed information about the cell's electrochemical properties. Implementing EIS on stack level is generally challenging due to low impedance of stacks with large cell areas due to inverse linearity between cell area and resistance.

[0009] Some previous QC procedures of SOC stacks attempt to simulate the conditions under which SOC stacks are to be operated. However, such simulated conditions add significant requirements to the equipment used for quality control, in particular the power supply required to supply operational current to the SOC stack undergoing quality control.

[0010] There is thus still a need for improved methods of performing quality control of solid oxide cell stacks.

[0011] SUMMARY

[0012] The conventional way of testing a SOC stack is to take it to its operational current and evaluate voltages, temperatures, etc. However, as the cell area, stack sizes, and hence currents go up, this becomes a challenge in a production line and thus increases the costs (CAPEX and OPEX) significantly. The present invention addresses this challenge by introducing a quality control procedure for industrial scale testing of solid oxide cell stacks providing reliable results at reduced costs and resource consumption. Electrochemical impedance spectroscopy (EIS) is considered a reliable method for characterising solid oxide cells, but also quite time and resource consuming to perform. While EIS measurements on SOC single-cell level are commonplace, implementation of EIS on stack level has remained limited. In particular, the known EIS procedures are not suitable for implementation as quality control procedures in industrial scale SOC manufacturing facilities.

[0013] The inventors have developed a quality control procedure which is based on EIS measurements to successfully provide information on faults in cells and stacks such as

[0014] • individual losses in the stack, such as oxygen ion oxidation kinetics, mass-transport processes, fuel electrode issues, and electrolyte resistance losses,

[0015] • weak interfaces

[0016] • ohmic stack resistance

[0017] • potential starvation issues within the stack.

[0018] Accordingly, a method according to the invention of performing quality control (QC) of a solid oxide cell stack preferably operable at full current load, the method comprising : a. Operating the solid oxide cell stack in electrolysis mode under a fuel side QC atmosphere comprising a reducing agent and an electrolyser feed at a QC current density preferably below 50% of the full current load; b. Recording an electrochemical impedance (EIS) spectrum of the stack during such operation, and then c. Comparing the obtained spectrum with a QC reference.

[0019] Under the method according to the present disclosure, an EIS spectrum is recorded while the solid oxide cell stack is operated in electrolysis mode under a fuel side QC atmosphere at a QC current density which is below 50% of the full current load.

[0020] By operating the SOC stack at a low current during quality control, while subjecting the SOC stack to other conditions which simulate actual operation, such as a fuel side QC atmosphere comprising a reducing agent and an electrolysis feed, the performance of the SOC stack at actual operational conditions can be accurately predicted.

[0021] The contribution to resistance in the SOC stack due to fuel gas conversion can be well- described in terms of up to four parameters: temperature, partial pressures, current density, and conversion rates I flow rates. By applying a reduced current density while keeping one or more of the other parameters close to the actual operational conditions, the current density (and the performance of the SOC stack) can be extrapolated from the reduced value of the current density to the operational current density. The QC atmosphere is further provided at a temperature which may be substantially similar to a temperature fuel gas provided to the SOC stack during actual operation.

[0022] This quality control procedure is highly reliable and accurate, even for SOC stacks having large cell areas, and the procedure offers advantages such as reduced requirements to power consumption, reduced requirements for power equipment, and reduced requirements to electrolysis feed consumption while maintaining accuracy and reliability. Since the quality control procedure is carried out at reduced current density and potentially fuel feed rates, it is also much quicker to perform the quality control procedure compared to many conventional QC procedures. The inventors found that this method of performing quality control surprisingly allowed to accurately and reliably predict faults likely to occur in a stack during full current load even though the test procedure only involved operating the stack at a current load of less than 50% of the full current load. In particular, current loads of less than 40, 30, 20 or even 10% of the full current load could be used to accurately and reliably predict faults likely to occur in a stack during full current load. The QC procedures disclosed herein thereby offer a promising framework for industrial-scale testing of SOC stacks.

[0023] Further, the QC procedures disclosed herein may be applied for steam electrolysis, CO2electrolysis, or for co-electrolysis where both steam and CO2undergo electrolysis.

[0024] The quality control procedure may be combined with state-of-the-art quality control procedures based on open circuit voltage (OCV) and voltage-current curves (Ui-curves).

[0025] The QC reference may comprise any of: an equivalent circuit model, one or more reference spectra obtained from historical data of solid oxide cell stacks, an obtained spectrum from another stack undergoing quality control, or a previously obtained spectrum from the same stack. The comparison of the obtained spectrum with the QC reference may, for example, be based on a pre-set threshold, i.e., if the difference between an obtained spectrum and a QC reference is greater than the threshold, the solid oxide cell stack fails the quality control. Such quality control may be automated, for example, a processor may be configured to perform the comparison and the determination as to whether the stack passes the quality control. The comparison between the obtained spectrum may further be based, at least in part, on input from a human operator. Identification of non-compliance with a QC reference may prompt further investigations, such as subjecting a stack to one or more further QC points.

[0026] LEGENDS

[0027] Figure 1 shows the equivalent circuit used to model the TSP-1 and TSP-2 SOC stacks.

[0028] Figure 2 shows the cell voltages obtained during the Ui curve, plotted as a function of time for the investigated TSP-1 stack.

[0029] Figure 3 shows the results obtained from the linear Kramers-Kroniq compliance test for cell groups 19-24 and 46-51 for fuel side QC points 1,2, and 4.

[0030] Figure 4 shows the raw EIS data obtained from the investigated TSP-1 stack plotted in the Nyquist configuration for a) fuel side QC point 2 and b) fuel side QC point 4.

[0031] Figure 5 shows the results from the CNLS fitting routine from cell group 46-51 in fuel side QC point 2 (Figures a and c), and fuel side QC point 4 (Figures b and d) plotted in both Nyquist and the Bode configuration. Figure 6 shows an overview of the individual contributions obtained from the CNLS fitting routine for all 13 cell groups in the investigated TSP-1 stack. Figure 6a shows the results from fuel side QC point 2, while Figure 6b is for fuel side QC point 4.

[0032] Figure 7 shows the actual operation of the EIS QC-tested TSP-1 stack with regards to 7a) the cell voltages and 7b) the temperatures during the test.

[0033] Figure 8 shows the raw EIS data presented in the Nyquist configuration for a TSP-2 stack with provoked flow channel issues. Figure a and b for fuel side QC point 1, and Figure c and d for fuel side QC point 4.

[0034] Figure 9 shows EIS data for another TSP-2 stack with provoked flow channel issues.

[0035] Figure 10 shows the cell group voltages during the maximum conversion test (Fig. 10b), along with a graph illustrating how the conversion rate increases (Fig. 10a).

[0036] Figure 11 shows a Ui curve of a SOEC, indicating the linear relation between current and voltage in a window of operation of relevance to the present disclosure.

[0037] DETAILED DISCLOSURE

[0038] Definitions

[0039] In the present context the terms "density of the QC current" and "QC current density" are used interchangeably. The term "QC current" may be understood as the current fed to the SOC or the solid oxide electrolysis cell (SOEC) and the "QC current density" may to refer to the QC current fed to the SOC or the SOEC per unit area. In the present context the QC current is provided as a direct current (DC).

[0040] For the avoidance of doubt, a QC current density of ~0% refers to measurements at open circuit voltage (OCV), where no current is applied to the stack.

[0041] In the present context, the term "reducing agent" refers to a gas which may be added to the fuel inlet stream in order to keep the cell at the inlet in a reduced state. A reducing agent is generally a chemical species that donates an electron to an electro recipient. Examples of a reducing agent within the context of the present disclosure are CO, H2, or any combination thereof. In the present context, the term "electrolyser feed" refers to the fuel used for the electrolysis Redox reaction. The electrolyser feed can be H2O, CO2or any combination thereof.

[0042] In the present context, the term "full current load" of a stack is meant to refer to the current load which the stack is designed to accommodate during normal operation in an electrolyser plant. The full current load can thereby also or alternatively be referred to as an operational current (density). In the context of solid oxide electrolysis, currents are often described in terms of current density. As an example, a full current load can be a current density in a range from 0.65 A / cm2to 4 A / cm2, such as in a range from 0.70 A / cm2to 3 A / cm2, 0.75 A / cm2to 2 A / cm2, for example in a range from 0.80 A / cm2to 1.5 A / cm2. As an example, a SOC stack is designed to be operated at a current density of 1.0 A / cm2. Operation of such a solid oxide cell stack at a current density below 50% of the full current load thereby corresponds to operation at a current density below 0.50 A / cm2. Thus, methods within the scope of the present disclosure may hence be described in terms of absolute current omitting any reference to a "full current load". For example, the present disclosure relates to a method of performing quality control (QC) of a solid oxide cell stack, the method comprising : a. Operating the solid oxide cell stack in electrolysis mode at least partly under a fuel side QC atmosphere comprising a reducing agent and an electrolysis feed at a QC current density; b. Recording an electrochemical impedance (EIS) spectrum of the stack during such operation, and then c. Comparing the obtained spectrum with a QC reference; wherein the QC current density is of below 0.50 A / cm2, such as in a range from 0.05 to 0.40 A / cm2, such as from 0.07 to 0.40 A / cm2, such as from 0.08 to 0.30 A / cm2, for example from 0.10 to 0.20 A / cm2. In this context, relative currents provided in terms of percent of a full current load within the present disclosure can alternatively be expressed in terms of absolute currents calculated assuming a full current load of 1.0 A / cm2. For example, the QC current density of QC1 expressed in terms of full current load in the further below can also be expressed as 0.05 to 0.15 A / cm2. This applies to all relative currents expressed herein.

[0043] The fuel side QC atmosphere may refer to the gas or atmosphere composition in the fuel side inlet of the solid oxide cell stack during quality control, which in turn may be the same atmosphere as used during later operation, for example during operation at the operation temperature. During quality control, multiple different atmospheres can be used, for example in order to test for different types of faults, and / or to simulate multiple different operational conditions which the stack is intended to be exposed to during operation. In the context of steam-, CO2-, and co-electrolysis, it is generally preferred to maintain a slightly reducing atmosphere in the fuel side inlet to ensure that the nickel catalyst remains in its reduced form.

[0044] Solid oxide cells and stacks High temperature solid oxide cells (SOCs) are highly efficient and environmentally friendly electrochemical systems for the H2 / H2O and / or CO / CO2redox reactions. The cells operate in two reversible modes: electrolysis mode (SOEC) for converting electrical energy into chemical energy (e.g., hydrogen production) and fuel cell (SOFC) mode for converting chemical energy into electrical energy, which facilitates this technology for power-to-gas-to-power application in renewable energy storage systems. However, the successful market introduction and public acceptance of the SOEC technology require high quality, reliability, and reproducibility of the corresponding cells and stacks.

[0045] It is therefore of interest to develop quality control (QC) procedures for SOC stacks, which can be performed before, e.g., the stacks are integrated into an electrolyser plant.

[0046] The challenges in performing quality control procedures on solid oxide cell stacks suitable for industrial scale electrolyser plants is that the active cell area is quite large and as the area increases, the resistance decreases. This results in significantly reduced voltage response to the sinusoidal current perturbation, in accordance with Ohm's law. This small voltage response makes it challenging to obtain reliable EIS measurements, due to the signal-to- noise ratio. In the present context solid oxide cell stacks suitable for industrial scale electrolyser plants can generally have an active surface area above 100 cm2, such as above 200 or 300 cm2.

[0047] For conversion of steam into hydrogen, CO2into CO, or for co-electrolysis where CO2and steam is fed in a (volume) ratio between 0 and 1, electrolyser plants based on solid oxide cell stacks are envisaged to be operated in a temperature range of 500 to 1000 °C, for example in a temperature range of 700 to 1000 °C, normally around 700 to 750 °C. It is further envisaged that the fuel feed when an electrolyser plant is in operation comprises 85-100% electrolyser fuel to thereby ensure optimal conditions for electrolysis.

[0048] Quality control procedure

[0049] In the method of performing quality control of a solid oxide cell stack according to the invention the stack is first operated at the temperature of operation in a QC point (fuel side QC atmosphere, oxy-side QC atmosphere, QC current density). In the present context, the QC current density refers to a direct current (DC) density. Once steady state is obtained, a device for measuring impedance applies a sinusoidal current (EIS current) on top of the already applied QC current and measures the voltage response of the SOC stack and reports the current and voltage magnitudes along with the current and voltage phases back. From these parameters the phase difference and amplitude difference can be calculated. From the differences in phase and amplitude reported, the impedance can then be calculated. A detailed description of the impedance technique follows.

[0050] The impedance measurement

[0051] Electrochemical impedance spectroscopy (EIS) is a well-known technique to study electrochemical systems. This technique is a powerful and non-invasive tool to investigate the dielectric and transport properties of materials, investigate interfaces and mechanisms, or electrochemical reactions and to explore the properties of porous electrodes.

[0052] In an EIS measurement, the response of a system, subjected to a small amplitude, sinusoidal current signal for galvanostatic mode and sinusoidal voltage in potentiostatic mode, in a frequency range of 0.1Hz to 40 kHz is studied. This sinusoidal current can be applied on top of a direct current (DC) as well as under open circuit potential (OCV).

[0053] The amplitude of the applied sinusoidal current is usually small, generally a small percentage of the DC in the galvanostatic EIS measurement (except for OCV, where the QC current density is ~0A / cm2). The voltage response, of the system, to the sinusoidal current is then measured. The impedance response Z can then be calculated from the perturbation current and the voltage response by using the generalized Ohm's law. When the EIS data is plotted in the complex plane, it is referred to as a Nyquist plot. A Nyquist plot shows the real part of the impedance in the abscissa and the imaginary part of the impedance in the ordinate. Notably, the imaginary part is presented with the opposite sign, to bring it into the first quadrant. Notably, when performing an EIS measurement, two currents are needed, the amplitude of the sinusoidal current (EIS current) and the direct current (QC current). I.e. adding a 5A AC on top of a 100A DC, would make the total current on a SOC stack variate between 95A and 105A.

[0054] Accordingly, the electrochemical impedance spectrum of the method is preferably recorded within a frequency range of 0.1 Hz to 80 kHz, such as 0.1 Hz to 40 kHz, 0.1 Hz to 30 kHz, 0.1 Hz to 20 kHz or 1 Hz to 20 kHz.

[0055] An advantage of recording within such frequency ranges is that more faults (or issues) related to individual losses in the stack (including oxygen ion oxidation kinetics, masstransport processes, and electrolyte resistance losses), ohmic stack resistance, and starvation issues in the stack may be identified. Some faults can remain unnoticed in some frequency ranges whereas in other frequency ranges the faults show. For example, some faults / issues are detectable in the frequency range of 0.1 to 5 Hz, others in the frequency range of 10 to 1000 Hz, and yet others are detectable in the frequency range of 1 to 40 kHz. An advantage of operating at low current densities is that a single power supply may be shared between two to six stacks or more during the quality control procedure. Another advantage of operating at low current densities is that the device used for measuring EIS may be shared between two to six stacks or more during the quality control procedure.

[0056] Accordingly, the method is preferably carried out within the QC current densities defined in claim 2.

[0057] QC points and faults

[0058] The inventors have found that certain combinations of a) reducing agent to electrolysis feed ratios in fuel side QC atmospheres, b) QC current densities and c) conversion percentages, are more promising than others to identify faults in cells and stacks related to individual losses in the stack (including oxygen ion oxidation kinetics, mass-transport processes, and electrolyte resistance losses), ohmic stack resistance, and starvation issues in the stack. Such combinations of a) reducing agent to electrolysis feed ratios in fuel side QC atmospheres, b) QC current densities and c) conversion percentage, may in the present context be referred to as QC points or QC operation points.

[0059] Accordingly, the method is preferably carried out under the conditions as defined in any one of claims 3 to 6.

[0060] It is thus preferable to conduct the QC procedure in one or more of the QC points of table 1 below. These QC points are highly suitable for performing the quality control procedure according to the invention. Notably, this QC procedure and this table considers a positive current, and hence a positive conversion, to be in electrolysis mode.

[0061] Table 1. Suitable QC points for performing quality control according to the invention. These exemplary values generally apply for steam electrolysis. For CO2electrolysis, the Boudouard reaction may influence the obtainable conversion percentage depending on the conditions. When performing co-electrolysis, the fuel side QC atmosphere ratio may apply to hydrogen and steam, and / or apply to CO and CO2. Similarly, when performing co-electrolysis, the conversion ratio may apply to hydrogen and steam, and / or apply to CO and CO2. The product may refer to H2 or CO.

[0062] Accordingly, the method is preferably carried out in at least one of the QC point, for example as defined in any one of claims 12 to 16.

[0063] Accordingly, the method is preferably carried out in at least two QC points as defined in any one of claims 9 to 11.

[0064] Identification of faults / issues

[0065] The inventors have identified combinations of QC points and EIS frequency ranges which are suitable for identifying various faults / issues in cells and stacks related to individual losses in the stack (including oxygen ion oxidation kinetics, mass-transport processes, and electrolyte resistance losses), ohmic stack resistance, and starvation issues in the stack. Examples of issues and suitable EIS frequency range QC point combinations for detection are given in table 2 below. The frequency ranges described in table 2 are for a Ni-YSZ anode-supported SOEC. Hence, comparing such frequency ranges of the obtained spectrum with corresponding frequency ranges of the QC reference may provide the most accurate identification of faults. For other types of solid oxide cells, other frequency ranges may potentially provide more accurate identification.

[0066] Table 2. Examples of issues and suitable EIS frequency range QC point combinations for detection.

[0067] More details regarding the different QC points are provided in the following.

[0068] QC point 1 simulates at least partly actual operational conditions, except for the conversion rate and the current density, and thereby also the pressure and the flow rates. Preferably, the thermodynamic behaviour, the ratio between electrolysis feed and reducing agent on the fuel electrode, the nitrogen to oxygen ratio on the oxy-electrode, and the temperature may however be substantially the same between the QC point and the actual operation of the stack. This QC point can thereby provide an improved scenario to investigate the oxygen electrode under operation, since if a crossover leakage is present, an insignificant amount of hydrogen crossover is present, and hence, the performance of the oxygen electrode can be evaluated. In addition, a high ohmic resistance in combination with a high oxygen electrode contribution may indicate delamination of the oxygen electrode. Hence, operational conditions substantially similar to those of QC 1 may be employed to identify oxygen electrode issues. A relevant frequency range is 102-103Hz, at least for a Ni-YSZ anodesupported SOEC.

[0069] QC point 2 provides a higher reducing agent to electrolyser feed ratio, thereby lowering the change in the Nernst potential across the cell while having an outlet product concentration similar to that of actual operation of the cell. This QC point can investigate the fuel electrode impedance contributions, namely the ionic rail in the fuel electrode, the triple phase boundary, and the diffusion in the fuel support layer, since all other larger contributions are minimized in the conditions of this QC point. Relevant frequency ranges are 104to 2 x 104Hz, 103to 104Hz, and 10-100 Hz, at least for a Ni-YSZ anode-supported SOEC in steam electrolysis. In addition, operational conditions substantially similar to those of QC 2 may also provoke issues on the oxygen electrode if significant crossover of reducing agent or product is present. In that case, the oxygen electrode may to some extent, be reduced locally, which will be reflected in the impedance by an increased oxygen electrode contribution alongside an abnormal pattern in the data (in a Nyquist plot) in a frequencies range of 102-103Hz, at least for a Ni-YSZ anode-supported SOEC operated in steam electrolysis.

[0070] QC point 3 is similar to QC case 2, except that this QC point has an even lower conversion rate, below 5%. This minimizes electrochemical contributions to the impedance of the cell. This allows for a detailed investigation of the ohmic resistance, thereby being able to evaluate contact-related issues between stack components and the ionic conductivity of the electrolyte material as well. As an example, the high-frequency intersection between the impedance data and the abscissa can be identified (in a Nyquist plot), above 2 x 104Hz, at least for a Ni-YSZ anode-supported SOEC operated in steam electrolysis.

[0071] QC point 4 is designed to provoke potential fuel distribution issues. This is done by emulating the stack under real operation conditions, similar to QC point 1. However, the conversion under QC point 4 is higher than the envisaged actual conversion percentage at the operational point. A combination of higher QC current density, low flows, and higher conversion can ensure that if there is leakage inside the stack or if areas that get less fuel, due to incorrect manufacturing of interconnect, these areas will have an increased fuel conversion arc in the low frequency part of the obtained spectrum. This is because the starved areas will have a local conversion rate above the intended conversion rate. A relevant frequency range for identifying an enlarged fuel conversion resistance is 0.1-5 Hz, at least for a Ni-YSZ anode-supported SOEC running in steam electrolysis.

[0072] QC point 5 is similar to QC point 4 except that the conversion rate is lower, while keeping the current density similar to that of QC point 4. This enables probing of fuel distribution issues with reduced risk of damaging the cells locally. On the other hand, a signal in the observed spectrum indicating a fault may also be smaller compared to a signal provided under the conditions QC point 4. A relevant frequency range of QC point 5 for identifying an enlarged fuel conversion resistance is 0.1-5 Hz, at least for a Ni-YSZ anode-supported SOEC running in steam electrolysis.

[0073] Different cell designs, electrolysis modes, or SOC technologies can have different frequency range for the individual electrochemical contributions, it is generally accepted that the kinetics of electrolysis involves the reaction kinetics described in the following : A fuel electrode ionic rail contribution, a triple phase boundary contribution from the fuel electrode, an oxygen ion oxidation in a mixed ionic-electronic conductor or at least a oxygen electrode contribution, a diffusion element coming from the diffusion through the fuel electrode structure (notably, this can be close to zero for some electrolyte-supported cells), and a conversion resistance mostly coming from the fuel electrode. In the field, the gas diffusion and the gas conversion resistances can sometimes be referred to as one contribution called concentration impedance.

[0074] Combining several QC points may provide a further improved quality control. In particular, the different QC conditions disclosed herein can respectively be utilized to identify isolated types of faults, while testing stacks under several different conditions can potentially ensure a holistic control of manufactured stacks.

[0075] In some of the following examples, QC procedures disclosed herein are verified using steam electrolysis. Note, however, that the QC procedures disclosed herein also apply well to CO2electrolysis or co-electrolysis.

[0076] EXAMPLES:

[0077] Example 1 : The solid oxide cell stacks

[0078] Two types of stacks were tested.

[0079] The first type was a commercial Topsoe Stack Platform version 1 (TSP-1), the first commercial stack design from Topsoe containing the solid oxide cell (SOC) technology. The stack consists of 75 repeating units, i.e., containing 75 solid oxide cells. For the experiments the cells in the TSP-1 stacks were divided into 12 groups of six cells, and one group of three cells. The total active cell area was above 100 cm2.

[0080] The second type was the second generation Topsoe Stack Platform version 2 (TSP-2). The

[0081] TSP-2 is envisaged to contain 25 to 150 repeating units (solid oxide cells) and will be divided into sub-stacks of 25 repeating units. For the experiments a stack of 25 cells were divided into five groups, each containing five cells. The total active cell area was above 300 cm2.

[0082] Regardless of the stack platform, a voltage wire probe was attached to the interconnect separating the cell groups. These voltage wire probes allowed internal testing and monitoring of the individual cell groups.

[0083] The five TSP-1 stacks tested comprised various cell batches. However, each stack contained only a single cell batch, implying that all cell groups within each stack should exhibit very similar performance.

[0084] The TSP-2 stack comprised cells from a single cell batch but incorporated three different interconnect materials (ICs). Electrochemically, the stack should hence perform very similar, but differences in ohmic resistance were expected due to the varying interconnect materials.

[0085] An overview of the TSP-2 stack differences can be seen in Table 3.

[0086] Table 3. Overview of IC distribution in the tested TSP-2 stack. The details of the different IC materials are typically not of importance for the quality control procedures disclosed herein, but more accurate identification of faults may be achieved by relying on a QC reference which has been established based on the same material(s) as the solid oxide cell to be controlled. Example 2 : Conditioning and reduction of the solid oxide cell stacks

[0087] During the conditioning and reduction of the SOC stack, the stack was heated to high temperatures and afterwards the nickel-oxide in the fuel electrode was reduced, thereby activating the SOC. The EIS QC procedure was performed after this step.

[0088] Example 3: quality control (OQ procedure, steam electrolysis

[0089] The stacks were tested under different conditions with only a relatively small DC (direct current) compared to the operational current (referred to as different QC points). The cells were operated at a temperature in the range of 700 to 750 °C, however during a single quality control procedure the temperature was kept constant. The oxy side QC atmosphere of the solid oxide cell stack was air or air mixed with an inert gas such as nitrogen, however during a single quality control procedure the oxy side QC atmosphere was kept constant. The quality control procedure involved testing the stacks under fuel side QC atmospheres comprising various ratios of hydrogen to steam and optionally an inert gas such as nitrogen. Also, the QC current density through the stack was varied.

[0090] Electrochemical impedance spectra were obtained for one or more QC points. In the present context it is to be understood that while the QC points are selected within the ranges given below, it is to be understood that during a quality control procedure in a QC point, the selected fuel side QC atmosphere, QC current density and resulting conversion is kept constant to ensure reliable data material.

[0091] In the present example the stacks were tested in five QC points within the following ranges:

[0092] The EIS quality control procedure tested the stacks under different conditions with only a small DC. The conditions were different concentration ratios between hydrogen and steam at different current densities and conversion rates. Notably, all measurements were performed with only a fraction of the operational current densities for both of the industrial scale stacks TSP-1 and TSP-2. In combination with the EIS QC procedure, hydrogen OCVs and currentvoltage curves (Ui curves) were also performed.

[0093] Example 4: device for measuring EIS

[0094] The device for measuring EIS should preferably be able to provide reliable results in the frequency range of 0.1 - 40.000 Hz, or at least 0.1- 20.000 Hz. Only the reliable range can be used in a quality control procedure. In prior art it has only been possible to obtain reliable results in the range of 0.1 - 10.000 Hz on an actual SOC stack. In prior art it has been necessary to measure each repeat unit (cell) individually.

[0095] The inventors have found, that when the device for measuring EIS includes a debiasing unit to ensure that the transformer of the device operates efficiently and safely by preventing saturation and allowing for real-time adjustments based on the current flowing through the system to improve the efficiency of the transformer, then reliable results can be obtained in the frequency range of 0.1 - 40.000 Hz, or at least 0.1 - 20.000 Hz.

[0096] To be more specific, a preferred device for measuring EIS on a SOC stack comprises

[0097] - a transformer with at least one primary winding and two secondary windings, where:

[0098] - the primary winding of the transformer connects in series a direct-current (DC) power supply with the SOC stack and the DC power supply generates a DC current across the primary winding to power the SOC stack;

[0099] - the first secondary winding is connected in series to a variable DC voltage generator suitable for generating a DC de-biasing current across the first secondary winding to reverse a DC flux bias generated in the transformer by the DC current flowing across the primary winding;

[0100] - the second secondary winding is connected in series to a variable alternating-current (AC) voltage generator suitable for generating at least one SOC stack analysis signal; and the device further comprises

[0101] - at least one sensor suitable for being connected to the SOC stack to measure the impedance response of the SOC stack;

[0102] - a controller connected to one or more of the at least one sensor, the direct-current (DC) power supply, the variable DC voltage generator, and the variable alternating-current (AC) voltage generator, the controller being suitable for receiving a signal from the at least one sensor and using this information to determine the impedance properties of the SOC stack.

[0103] In particular this type of device for measuring EIS provides reliable data in the high frequency range of 10.000-40.000 Hz which is useful for reliably testing for loss of contact within and between the repeat units of the SOC stack as well as the conditioning and the reduction of the SOC stack. This type of device is described in e.g. WO 21 / 003577.

[0104] An advantage of such a device is that the first secondary winding producing the de-biasing current extends the frequency range where the apparatus produces reliable results.

[0105] Accordingly, a system comprising an arrangement for continuously operating a solid oxide cell stack in electrolysis mode and a device for measuring EIS may be envisaged, wherein: the arrangement comprises means for adjusting the DC current applied to the stack and means for adjusting the fuel side feed to the stack the device comprises a de-biasing unit

[0106] The de-biasing unit functions to reverse a DC flux bias generated in a transformer of the device by a DC current flowing in the transformer;

[0107] Example 5: Equivalent circuit model used for the complex non-linear least squared (CNLS) fitting

[0108] All data underwent a data quality control process, and the resulting EIS data were then fitted to an equivalent circuit to separate the various electrochemical processes. The equivalent circuit shown in Figure 1 was used to represent the data. The inductance from the test station and stack is modeled as an inductor (L) at the highest frequencies. All ohmic resistance contributions, from both the cell and stack, are represented by a single Rs. The first two RQ elements, at high frequencies, represent a combination of fuel-electro-reduction, ionic transport, and gas diffusion in the fuel-electrode functional layer. These two contributions are known as the ionic rail and the TPB resistances. At mid-frequencies, a Gerisher element represents the oxygen ion oxidation in a mixed ionic-electronic conducting oxygen-electrode. Following this, a finite-length Warburg diffusion element represents gas diffusion in the fuel-electrode support layer, including the contact mesh and fuel-electrode flow field. Lastly, at the lowest frequencies, two additional RQ elements represent the gas conversion resistance in the fuel- and oxy-electrode. Equivalent circuit models as exemplified above are well-known in the art.

[0109] Figure 1 shows the equivalent circuit used to model the TSP-1 and TSP-2 SOC stacks.

[0110] Example 6: Results

[0111] Test of TSP-1

[0112] The TSP-1 stack, consisting of 13 cell groups, generates an extensive volume of data across the various QC atmospheres that were tested. To streamline the analysis disclosed in the following, the focus is primarily directed to three specific QC points and two representative cell groups (cell group 19-24 and cell group 46-51). These cell groups were selected to highlight interesting observations and the overall performance of the SOC TSP-1 stack. The chosen atmospheres are fuel side QC point 2, 3 and 4, seen in table 1, with fuel side QC point 2 and 3 having a higher hydrogen to steam ratio than fuel side QC point 4. The chosen cell groups presented from the stack will be cell group 19-24 and cell group 46-51. For comparison, a small Ui curve / voltage measurement of all cell groups was performed before and after the EIS QC procedure. The results from such a Ui curve can be found in Figure 2. The current for this Ui curve is 0-15% of the normalized operation load.

[0113] Figure 2 shows the cell voltages obtained during the Ui curve, plotted as a function of time for the investigated TSP-1 stack.

[0114] The first and last cell group of the stack has a lower voltage than the remaining cell groups, which is likely due to their position in the stack where they experience a different temperature, as they receive more radiation heating from the furnace. Notably, cell group 46-51 continuously has a higher voltage than the rest of the stack. However, this increased voltage is, at the maximum current, only lOmV higher than the rest of the stack. Hence, if only a partial load DC quality control is performed, this minor difference in performance would likely not have been sufficient to identify the issues present in cell group 46-51, due to the performance difference between cell groups being so minor, and hence have passed the stack through QC.

[0115] Once the EIS QC procedure is finalized, data analysis begins. The first checkpoint for the data is to look at the Kramers-Kronig compliance tests since a non-compliant data set should be discarded. This data, from cell groups 19-24 and 46-51, can be seen in Figure 3, for fuel side QC points 2, 3, and 4, respectively.

[0116] Figure 3 shows the results obtained from the linear Kramers-Kroniq compliance test for cell groups 19-24 and 46-51 for fuel side QC points 1, 2, and 4.

[0117] As can be observed, all data are Kramers-Kroniq compliant, however, the low-frequency part of cell group 46-51 in fuel side QC point 2 is affected by underlying instabilities, however, the QC procedure can still be performed on the data. The next step in the QC data procedure is to investigate the data in both the Bode and Nyquist configurations. In Figure 4 the data, plotted in the Nyquist configuration, from fuel side QC points 2, and 4 can be seen for the whole stack.

[0118] Figure 4 shows the raw EIS data obtained from the investigated TSP-1 stack plotted in the Nyquist configuration for a) fuel side QC point 2 and b) fuel side QC point 4.

[0119] From Figure 4a, it is clear that cell group 46-51 are performing different to the rest of the stack. Interestingly, the same thing is not observed in test atmosphere 4, seen in Figure 4b, where cell group 46-51 is no longer an extreme outlier. However, to investigate whether this cell group 46-51 is behaving differently in one QC atmosphere and not in the other, the data was fitted using the CNLS fitting procedure. The results from the CNLS fitting routine, using the equivalent circuit described in Figure 1, can be seen in Figure 5, for cell group 46-51 in fuel side QC points 2, and 4.

[0120] Figure 5 shows the results from the CNLS fitting routine from cell group 46-51 in fuel side QC point 2 (Figures a and c), and fuel side QC point 4 (Figures b and d) plotted in both Nyquist and the Bode configuration.

[0121] By comparing Figures 5a, and 5c, with Figures 5b, and 5d, it can be seen that the CNLS fitting results from fuel side QC point 4, give a much better fit. However, by comparing the size of the individual contributions for the two atmospheres, additional information about what has gone wrong in cell group 46-51, can be obtained. Figure 6 shows an overview of the individual contributions obtained from the CNLS fitting routine for all 13 cell groups in the investigated TSP-1 stack. Figure 6a shows the results from fuel side QC point 2, while Figure 6b is for fuel side QC point 4.

[0122] From the data presented in Figure 6b, it is apparent that none of the individual contributions for cell group 46-51 (Group no. 9) are behaving differently compared to the other cell groups. However, if the attention is turned to Figure 6a, especially the ohmic, conversion, and oxygen electrode contributions are significantly higher for cell group 46-51, compared to the rest of the stack. These results strongly indicate issues with the oxygen electrode, due to the increased ohmic resistance in combination with higher oxygen electrode contribution. In addition, the increased gas conversion resistance gives signs of higher leakage or areas being starved in cell group 46-51. The reason for the issues in the oxygen electrode, not being visible in fuel side QC point 4, could be explained by the fact that, if a crossover happens from the fuel electrode to the oxygen electrode when operated at higher hydrogen concentrations (fuel side QC points 2 and 3) hydrogen would be transferred to the oxygen electrode. This hydrogen would then either locally reduce the oxygen electrode or spontaneously react and combust with oxygen to produce water. If a locally reducing atmosphere is present in the oxygen electrode, a decrease in activity and hence an increase in resistance would be expected. However, this decrease in activity seems to be reversible, since, when the fuel electrode is investigated at lower hydrogen to steam ratios, the same decrease in activity is not observed. This is due to the crossover now being mostly steam, which is also detrimental to the oxygen electrode, however only with time. That the crossover gas changes from hydrogen to steam, will allow the oxygen electrode to partly regenerate. The overall conclusion of this QC investigated stack is hence that leaks are present in cell group 46-51, which affects an already weakened oxygen electrode, hence it is more apparent at higher hydrogen concentrations.

[0123] The interesting question to ask is how this stack then performed when handled in operation. The stack was operated in the standard TSP-1 conditions reported elsewhere. As seen in Figure 7a, the test was short and brutal, despite OCV values looking normal during heat-up. Interestingly, the first cells started short-circuiting in cell group 46-51, see Figure 7a, combined with an increase in temperature, see Figure 7b, already 45 minutes after being put into operation. And, as time progressed, more cell groups started to short-circuit, now also in the surrounding cell groups until the test was finally stopped after only 140 hours of operation.

[0124] Figure 7a shows the cell voltages during the actual operation of the EIS QC-tested TSP-1 stack. Figure 7b shows the temperatures during the test. Conclusion

[0125] During the QC experiments performed on the TSP-1 it was possible to identify issues in the oxygen electrode interfaces, resulting in weakened or even partly delaminated oxygen electrode, since both the oxygen electrode and the ohmic resistance was affected. In addition, a significant cross-over leakage was identified, by comparing the different QC points with each other. That the stack afterwards only was in operation for 45 minutes before cells started to short circuit, validates the EIS QC procedure.

[0126] Test of TSP-2

[0127] Performing EIS on a stack the size of the TSP-2 stack proved to be quite a challenge, due to the stringent requirements of both equipment and test setup. To support the development of the QC procedure, TSP-2 stacks were intentionally manufactured with provoked faults, to test if the QC EIS procedure captured all these issues and also to test that the operator made the correct conclusions. One of the provoked issues was fuel starvation of specific areas in the cells. This was done by producing interconnects with smaller flow channels than specified.

[0128] Figure 8 shows the raw EIS data presented in the Nyquist configuration for a TSP-2 stack with provoked flow channel issues. Figure a and b for fuel side QC point 1, and Figure c and d for fuel side QC point 4.

[0129] Figure 8a and b present the EIS data obtained from cell group 11-15 and 16-20, in fuel side QC point 1, respectively. Figure 8c and 8d present the data from the same cell groups, however in fuel side QC point 4, where the conversion was increased, by decreasing flows, compared to fuel side QC point 1.

[0130] The EIS data from the investigated TSP-2 stack can be seen in figure 8. Notably, the data is presented with the same abscissa and ordinate for easier comparison. From Figures 8b and 8d, the EIS response of the cell group with smaller flow channels in the interconnects changes significantly, especially the conversion resistance. Meanwhile, the response of the reference cell group, seen in Figures 8a and 8c, does not change significantly when changing the conversion. This gives strong indications that the EIS QC procedure can be used on the TSP-2 stack design with similar success. These results suggest that flow distribution issues on the fuel side, leading to starvation and thereby limiting maximum conversion percentage in operation, can be identified.

[0131] Conclusion: The EIS QC results from the TSP-2 experiment showed that it is possible to identify fuel distribution issues on the fuel side, something that would limit the maximum conversion percentage which the stack would be able to endure during actual operation (under full current load). In addition, the EIS QC procedure on TSP-2 is also able to identify stacks which are not gas tight, production and performance issues in the SOC, cross-over leakage and delamination issues.

[0132] Figure 9 shows EIS data for another TSP-2 stack with provoked flow channel issues. Further, this stack is additionally tested via a maximum conversion test. Figure 9a shows the EIS data for all five cell groups at fuel side QC point 1, and Figure 9b shows the EIS data for all five cell groups at fuel side QC point 4, where the conversion was increased by decreasing the flow compared to QC point 1. In both sub-figures, the data is presented in as Nyquist plot with the same abscissa and ordinate for easier comparison.

[0133] Cell group 1, which contained interconnects with smaller flow channels, exhibited a significant change in EIS response, particularly in the conversion resistance.

[0134] The reference cell groups (those without provoked faults) did not show significant changes in response when the conversion was altered

[0135] These results further demonstrate that the EIS QC procedure can reliably identify fuel distribution issues on the fuel side, which can lead to fuel starvation and limit the maximum conversion percentage the stack can achieve during operation.

[0136] To further test whether the results obtained at less than 50% of the operational current density could be correlated to performance at full operational current density, a maximum conversion test was also conducted on the same TSP-2 stack. In this test, the steam flow to the fuel electrode is gradually lowered while maintaining a constant current density. An increase in voltage of one cell group relative to the other cell groups indicates that this cell group has reached its maximum operational conversion rate.

[0137] Figure 10 shows the cell group voltages during the maximum conversion test (Fig. 10b), along with a graph illustrating how the conversion rate increases (Fig. 10a).

[0138] The cell group that exhibited fuel distribution issues in the EIS QC procedure (cell group 1) was the first to react during the maximum conversion test, confirming the correlation between the EIS QC results and the operational performance of the stack.

[0139] The equivalent circuit element representing the fuel electrode conversion resistance was quantified and used to evaluate the fuel distribution issue. This allowed for a direct correlation between the fuel side conversion resistance determined by the EIS QC procedure and the maximum conversion rate that the cell group could tolerate.

[0140] The EIS QC procedure applied to the TSP-2 stack successfully identified fuel distribution issues on the fuel side, which would limit the maximum conversion percentage the stack could endure during full current load operation. Further, the EIS QC procedure can also potentially detect gas-tightness issues, production and performance issues in the SOC, crossover leakage, and delamination issues.

[0141] All of these issues can be identified using a QC current density less than 50% of the full operational current density and can be directly correlated to issues observed under full operational current densities.

[0142] These results thereby validate the effectiveness of the EIS QC procedure in identifying critical faults.

[0143] Figure 11 shows a Ui curve of a SOEC, indicating the linear relation between current and voltage in a window of operation.

[0144] Preferably, methods according to the present disclosure rely on quality control performed with a QC current density within this range of current densities with a substantially linear voltage-current relationship (while preferably also below 50% of the full current load). Similarly, the full current load is also, preferably, within this linear range of current densities. As a result, behaviour of the SOC stack observed at low current during quality control can be extrapolated (i.e., are indicative of) behaviour of the SOC stack at operational current densities. The range of current densities with a linear voltage-current relationship can be further quantified in terms of, e.g., the second derivative of the voltage with respect to the current or current density, and if the absolute deviation of this second derivative from zero is within a threshold, then the voltage-current relationship is linear for a given current or current density. The size of the threshold can be set relative to the second derivative at zero current, for example, the threshold is 50% of the second derivative at zero current.

[0145] Thermoneutral operation

[0146] During the actual operation of solid oxide cell stacks in an electrolysis plant, these may preferably be operated under thermoneutral conditions, for example at the full current load. Since the quality control disclosed herein generally relies on using a QC current density below 50% of the full current load, thermoneutral operation may not be possible to achieve during the quality control. To nevertheless ensure conditions which are as close as possible to thermoneutral operation, the stack may be operated at near-thermoneutral conditions. Such conditions may be quantified in terms of a temperature difference across the stack (including both resistive heating and endothermic cooling) relative to thermoneutral conditions, the thermoneutral conditions corresponding to a balancing of resistive heating and thermal cooling resulting in a temperature difference across the stack of 0°C. Preferably, the total temperature difference across the stack has a magnitude less than 15 °C relative to thermoneutral conditions, for example less than 10 °C , such as less than 5°C.

[0147] The quality control procedures disclosed herein are performed under particular conditions, preferably using a relatively low current. Preferably, actual subsequent operation of the stack is performed under one or more of the same conditions (except for the current density which is preferably greater) in, e.g., an operational electrolysis plant. By utilizing the same conditions in quality control and operation, the risk of not identifying relevant faults can be reduced. In case one or more faulty cells or stacks are identified, these should preferably not be used, at least not without mitigation of these faults. Thereby, the actual operation of the stack can be based on the quality control. The stack is only operated once it successfully passes the quality control.

[0148] Measurement setup

[0149] Various adaptations of the QC measurement setup may be performed to ensure sufficient data quality. Such adaptations may in particular be of relevance for stacks comprising more than just a few cells, for cells having a large cell area, and for measurement at high frequencies, such as above 10 or 20 kHz.

[0150] Typically, a solid oxide cell stack may have a central axis extending perpendicularly to the cells of the stack, and the solid oxide cell stack is operated by supplying current via a first conductor in contact with a first end of the solid oxide cell stack and by discharging current via a second conductor in contact with a second end of the solid oxide cell stack, the first end of the solid oxide cell stack being opposite to the second end of the solid oxide cell stack.

[0151] To improve data quality, particularly for stacks comprising more than just a few cells, for cells having a large cell area, and for measurement at high frequencies, one or more of the following adaptations may be implemented in any combination:

[0152] - the first conductor is parallel to the central axis at a portion of the first conductor proximate to the solid oxide cell stack; and / or the second conductor is parallel to the central axis at a portion of the first conductor proximate to the solid oxide cell stack. - a contact point of said first conductor with the solid oxide cell stack is at the central axis; and / or a contact point of said second conductor with the solid oxide cell stack is at the central axis.

[0153] - the electrochemical impedance spectrum is recorded via voltage contacts of the solid oxide cell stack, the respective voltage contacts providing electrical contact to respective cells of the stack, wherein the voltage contacts are collectively arranged in parallel to the central axis.

[0154] - neighbouring wires connected to said respective voltage contacts are pairwise twisted into twisted wire pairs, wherein neighbouring twisted wire pairs are pairwise twisted into twisted wire bundles.

[0155] Further remarks

[0156] Preferably, the QC current density is supplied from a current source which is separate from the device for measuring EIS, the device for measuring EIS supplying a signal for recording the electrochemical impedance spectrum of the stack. Hence, preferably, any contribution to a total current density in the electrolysis stack from the device for measuring EIS may not be considered as part of the QC current density. This can for example be relevant in case a device for measuring EIS further supplies a (minor) DC current, for example because a small direct current is required such that the alternating current does not cross 0A.

[0157] Preferably, the QC current density is in a range from 5% to below 50% of the full current load, and / or the solid oxide cell stack is not operated under open circuit voltage while the QC current density is supplied. Thereby, the operational conditions of the stack can be more accurately simulated during the quality control.

[0158] Open circuit voltage may alternatively be defined as a QC current density within a range from -2% to 2% of the full current load, such as from 0% to 2% of the full current load.

Claims

CLAIMS1. A method of performing quality control (QC) of a solid oxide cell stack which is operable at full current load, the method comprising: a. Operating the solid oxide cell stack in electrolysis mode at least partly under a fuel side QC atmosphere comprising a reducing agent and an electrolyser feed at a QC current density of below 50% of the full current load; b. Recording an electrochemical impedance spectrum of the stack during such operation, and then c. Comparing the obtained spectrum with a QC reference.

2. The method according to claim 1 wherein the QC current density is below 40%, such as below 35, 30, 25, 20, 15, or 10% of the full current load.

3. The method according to any one of claims 1 to 2 wherein the fuel side QC atmosphere comprises the reducing agent and the electrolyser feed at a volume ratio of agent:feed in the range of from 0.01 to 3, such as 0.01 to 0.9, 0.02 to 0.7, 0.02 to 0.5, 0.03 to 0.4, 0.03 to 0.25, or 0.03 to 0.2; or such as 0.3 to 3, 0.5 to 2, 0.6 to 1.7, or 0.7 to 1.5.

4. The method according to any one of claims 1 to 3 wherein the fuel side QC atmosphere comprises the reducing agent and the electrolyser feed at a ratio to provide a conversion of the electrolyser feed into the reducing agent, at the QC current density, of 0-90%, such as 0-80%, 5-70%, 20-40%, 0-5%, or 50-80%.

5. The method according to any one of claims 1 to 4 wherein in step a. the solid oxide cell stack is operated with a temperature difference across the stack of less than 15 °C.

6. The method according to any one of claims 1 to 5, wherein the solid oxide cell stack has a range of current densities with a substantially linear voltage-current relationship, and wherein the QC current density and / or the full current load is within said range of current densities.

7. The method according to any one of claims 1 to 6, wherein the QC current density is in a range of from 5% to below 50% of the full current load; and / or wherein in step a. the solid oxide cell stack is not operated under open circuit voltage.

8. The method according to any one of claims 1 to 7, wherein:- the reducing agent is or comprises H2, and the electrolyser feed is or comprises H2O and / or CO2;- the reducing agent is or comprises CO, and the electrolyser feed is or comprises CO2 and / or H2O; or- the reducing agent comprises H2 and CO, and the electrolyser feed comprises H2O and / or CO2.

9. The method according to any one of claims 1 to 8, wherein the method comprises operating the solid oxide cell stack and recording the electrochemical impedance spectrum under several different sets of QC conditions, for example wherein said different sets of QC conditions have any of: different fuel side QC atmospheres, different QC current densities, and different conversions of the electrolyser feed into the reducing agent(s).

10. The method according to any one of claims 1 to 9, wherein said different sets of QC conditions are respectively indicative of different faults in the solid oxide cell stack.

11. The method according to any one of claims 1 to 10, wherein said different faults are indicated by features in the obtained spectrum obtained at said different respective QC conditions.

12. The method according to any one of claims 1 to 11, wherein in step a. the solid oxide cell stack is operated under conditions where: the fuel side QC atmosphere comprises the reducing agent and the electrolyser feed at a volume ratio of agent:feed in a range of from 0.01 to 2.0, for example from 0.02 to 1.0, for example in a range from 0.03 to 0.75, for example in a range from 0.03 to 0.50, such as in a range from 0.04 to 0.25; the QC current density is in a range from 5% to 40% of the full current load, for example in a range from 5% to 35%, for example in a range from 5% to 30%, for example in a range from 5% to 25%, for example in a range from 5% to 20%, such as in a range from 5% to 15%; and the QC atmosphere optionally comprises the reducing agent and the electrolyser feed at a ratio to provide a conversion of the electrolyser feed into the reducing agent at the QC current density in a conversion range from 10% to 65%, for example in a conversion range from 12% to 60%, for example in a conversion range from 14% to 55%, for example in a conversion range from 16% to 50%, for example in a conversion range from 18% to 45%, such as in a conversion range from 20% to 40%.

13. The method according to any one of claims 1 to 12, wherein in step a. the solid oxide cell stack is operated under conditions where: the fuel side QC atmosphere comprises the reducing agent and the electrolyser feed at a volume ratio of agent:feed in a range of from 0.25 to 3.0, for example from 0.35 to 2.50, for example in a range from 0.45 to 2.10, for example in a range from 0.55 to 1.80, such as in a range from 0.65 to 1.50; the QC current density is in a range from 5% to 40% of the full current load, for example in a range from 5% to 35%, for example in a range from 5% to 30%, for example in a range from 5% to 25%, for example in a range from 5% to 20%, such as in a range from 5% to 15%; andthe QC atmosphere optionally comprises the reducing agent and the electrolyser feed at a ratio to provide a conversion of the electrolyser feed into the reducing agent at the QC current density in a conversion range from 10% to 65%, for example in a conversion range from 12% to 60%, for example in a conversion range from 14% to 55%, for example in a conversion range from 16% to 50%, for example in a conversion range from 18% to 45%, such as in a conversion range from 20% to 40%.

14. The method according to any one of claims 1 to 13, wherein in step a. the solid oxide cell stack is operated under conditions where: the fuel side QC atmosphere comprises the reducing agent and the electrolyser feed at a volume ratio of agent:feed in a range of from 0.25 to 3.0, for example from 0.35 to 2.50, for example in a range from 0.45 to 2.10, for example in a range from 0.55 to 1.80, such as in a range from 0.65 to 1.50; the QC current density is in a range below 10% of the full current load, for example in a range below 8%, for example in a range below 6%, for example in a range below 4%, for example in a range below 2%, such as operation at open circuit voltage; and the QC atmosphere optionally comprises the reducing agent and the electrolyser feed at a ratio to provide a conversion of the electrolyser feed into the reducing agent at the QC current density in a conversion range from 10% to 65%, for example in a conversion range from 12% to 60%, for example in a conversion range from 14% to 55%, for example in a conversion range from 16% to 50%, for example in a conversion range from 18% to 45%, such as in a conversion range from 20% to 40%.

15. The method according to any one of claims 1 to 14, wherein in step a. the solid oxide cell stack is operated under conditions where: the fuel side QC atmosphere comprises the reducing agent and the electrolyser feed at a volume ratio of agent:feed in a range of from 0.01 to 2.0, for example from 0.02 to 1.0, for example in a range from 0.03 to 0.75, for example in a range from 0.03 to 0.50, such as in a range from 0.04 to 0.25; the QC current density is in a range from 5% to 45% of the full current load, for example in a range from 6% to 40%, for example in a range from 7% to 35%, for example in a range from 8% to 30%, for example in a range from 9% to 25%, such as in a range from 10% to 20%; and the QC atmosphere optionally comprises the reducing agent and the electrolyser feed at a ratio to provide a conversion of the electrolyser feed into the reducing agent at the QC current density in a conversion range from 0% to 95%, for example in a conversion range from 10% to 95%, for example in a conversion range from 20% to 90%, for example in a conversion range from 30% to 90%,for example in a conversion range from 40% to 85%, such as in a conversion range from 50% to 80%.

16. The method according to any one of claims 1 to 15, wherein in step a. the solid oxide cell stack is operated under conditions where: the fuel side QC atmosphere comprises the reducing agent and the electrolyser feed at a volume ratio of agent:feed in a range of from 0.01 to 2.0, for example from 0.02 to 1.0, for example in a range from 0.03 to 0.75, for example in a range from 0.03 to 0.50, such as in a range from 0.04 to 0.25; the QC current density is in a range from 5% to 45% of the full current load, for example in a range from 6% to 40%, for example in a range from 7% to 35%, for example in a range from 8% to 30%, for example in a range from 9% to 25%, such as in a range from 10% to 20%; and the QC atmosphere optionally comprises the reducing agent and the electrolyser feed at a ratio to provide a conversion of the electrolyser feed into the reducing agent at the QC current density in a conversion range from 10% to 65%, for example in a conversion range from 12% to 60%, for example in a conversion range from 14% to 55%, for example in a conversion range from 16% to 50%, for example in a conversion range from 18% to 45%, such as in a conversion range from 20% to 40%.

17. The method according to any one of claims 1 to 16, wherein said method of performing quality control of the solid oxide cell stack is a method of performing quality control of the solid oxide stack and operating the solid oxide cell stack, wherein the method comprises a further step of: d. Operating the solid oxide cell stack in electrolysis mode at an operational current density greater than the QC current density, for example at the full current load, as part of an operational electrolysis plant.

18. The method according to claim 17, wherein in step d. the solid oxide cell stack is operated under thermoneutral conditions.

19. The method according to any one of claims 17 to 18, wherein step d. is performed using a fuel side operational atmosphere, with a ratio between electrolyser feed and redcuing agent, which is substantially the same as at least one of the fuel side QC atmosphere(s) used during step a.

20. The method according to any one of claims 17 to 19, wherein the operational atmosphere has substantially the same temperature as at least one of the fuel side QC atmosphere(s) used during step a.

21. The method according to any one of claims 17 to 20, wherein step a. is performed to provide a QC conversion rate of the electrolyser feed into the reducing agent, wherein step d. is performed to provide an operational conversion rate of the electrolyser feedinto the reducing agent, wherein the QC conversion rate is substantially the same as the operational conversion rate.

22. The method according to any one of claim 17 to 21, wherein step d. is performed based on step c.

Images