Hydrogen Solar (United Kingdom)

Towards a CO2 free energy supply, hydrogen plays an important role as a chemical energy storage capable of storing renewable energy in large dimensions. In this context, fuel cells, such as e.g. polymer electrolyte membrane fuel cells (PEMFC) are going to evolve to a key technology for the conversion of the chemically stored energy back to usable electric energy. Therefore, after decades of fuel cell research, the phase of industrialization of fuel cells has begun recently to provide fuel cells in large quantity for a variety of applications, such as e.g. to power heavy duty vehicles, ships and even airplanes or as stationary power plants.For PEMFC mass production, the optimization of the manufacturing process of fuel cells is important to decrease the manufacturing time and costs. One of the most time and cost consuming steps during the production of fuel cells is the first time activation. During the so-called break-in, the fuel cell is operated for the first time in order to activate the membrane electrode assembly (MEA). During break-in, the proton conductivity of the membrane is established and the catalyst is activated, but also chemical residues of the production of the fuel cell components were removed. This procedure is necessary for a fuel cell stack to provide its maximum power output and maximum robustness of all single cells. Furthermore, the successful activation of the PEMFC strongly affects the overall lifetime of a fuel cell stack.To date, a typical break-in of a fuel cell stack takes in between 2 and 8 hours of time, in which a considerable amount of hydrogen and air is required. Thereby, the stack is commonly operated at a high relative humidity doing several cycles of current ramps from minimum to medium or high load interrupted by shutdowns, whereby the first current ramp is usually performed very slowly. Besides the media consumption, the rather long state-of-the-art break-in time calls for a high degree of parallelization of break-in stations at each production line to reach a reasonable output of fuel cell stacks, which is related to high investment costs. An optimization of the break-in process in respect to the necessary time and media consumption significantly reduces the production costs of a PEMFC stack. In this work, several new approaches to activate PEMFCs were introduced in order to decrease the duration and media consumption of the break-in process. Thereby, the common break-in doing several current ramps was replaced by more efficient and promising approaches, such as e.g., hydrogen starvation events or hydrogen pumping. This break-in experiments were performed in short stacks of small scale as well as automotive scale. Finally, the influence of each break-in variation on the performance and durability of the PEMFC was investigated.

In this technical report, we present our second-generation digital twin model for monitoring of fuel cells. This digital twin combines two data-driven machine learning models: a stationary model for stationary cell voltage prediction and a degradation model for degradation correction. This combination of both models results in a precise probabilistic prediction of the cell voltage over time.Furthermore, we present a web-based framework for automated fuel cell monitoring in real time. This framework allows training a digital twin on existing data and subsequently applying the twin to automatically check new data for anomalies.

Transferring Electrochemical Impedance spectroscopy from fundamental research towards close to application systems allows direct insights into kinetics, transport processes and a state of health indication. Since such stacked systems are optimized for application comprehensive monitoring with reference electrodes is hardly possible. In addition, due to the high power consumption suitable measurement devices are hard to find and the results can only be used to a limited extent. In this work such a test object was set up and tested in different measurement configurations to investigate the reasons for the high noise ratio of the impedance measurement data at more than 200A. It was found that a series circuit consisting of a DC power supply, a high-precision impedance load and the test electrolyser has the least influence on the measurement result of all the methods tested. Finally, the impedance spectra were checked for interpretability using DRT and equivalent circuit fit.

Over the last decade, significant progress has been achieved in the commercialization of Polymer Electrolyte Membrane (PEM) fuel cells for mobile applications. However, long-term durability and stability of such systems still needs further improvement. To overcome today’s time-consuming long-term durability tests, the development of so-called accelerated durability tests (ADT) is urgently necessary. In this work, we present our recent results on such ADT test-protocols based on research promoting a detailed understanding of the individual aging mechanisms in PEM fuel cells.Various stressors had been investigated separately to understand their impact on overall fuel cell ageing. Certain levels of cell potential, temperature and humidity could be identified as major stressors and were successively added to the test protocol. In addition to the electrochemical test results, additional focus was put on different in-situ and ex-situ characterization methods. This combination enables a detailed understanding of individual stressors and their interply. The PEM-stack used in this study is an automotive high-performance stack developed within the publically funded Auto Stack Industry (ASI) project. An extensive long-term test over 5,500 operating hours was carried out as baseline test using the New European Driving Cycle, but with additional features to mimic realistic automotive load-profiles. Based on the results of this reference protocol, four different ADT protocols of only 1,200 hours test duration were derived. All tests were performed using short stacks with automotive-size active area. In-situ measurements include spatially resolved current measurement, cyclic voltammetry and impedance spectroscopy. After each test, selected cells were additionally analyzed ex-situ using layer thickness measurement via SEM, IR-thermography and contact angle measurements.The results from ex-situ investigations underline the degradation models postulated from in-site measurements. The overall data show that detailed understanding of individual stressors is necessary to successfully develop reliable accelerated ageing tests for PEM fuel cells.

Due to an increasing demand for sustainable energy, the polymer electrolyte membrane fuel cell (PEMFC) is considered a key technology in energy conversion with hydrogen as an energy carrier. The performance characteristics of a PEMFC is significantly influenced by the morphology of the catalyst layer[1], which is formed predominantly during the drying step in the manufacturing process.In order to improve the performance of PEMFC and hence reducing production costs, research effort as well as industrial interest has focused on the drying process. Moreover, energy efficiency of the drying step in the production of thin-film catalyst layers is considered a key driver of cost perspective for the mass-market introduction of PEM fuel cells and electrolyzers due to increasing energy costs and the need for reduction of environmental impact.In the presented work, an innovative drying approach for thin-film catalyst layers, used in PEMFC and electrolyzer applications, is investigated. The thin-film is either coated on a substrate and transferred onto a polymer electrolyte membrane in a second process step or is directly coated onto the polymer electrolyte membrane by doctor blade coating. Drying is performed by coupling of microwave radiation into the wet film, leading to a homogeneous and energy-efficient heating of the coated film. Energy losses of the microwave drying process are significantly reduced compared to conventional convective drying. Additionally, the microwave induced drying step offers maximum control during the process.Dried thin-films, produced in the lab scale, are transferred to a polymer electrolyte membrane by a hot-pressing step and are completed with gas diffusion layers to membrane electrode assemblies (MEA). Electrochemical performance of the MEA is characterized on a fuel cell test bench in single cell configuration. The morphology of the structure obtained from microwave drying is analyzed and compared to CCM samples dried in a conventional convective oven.The effect of a microwave-based drying process is discussed with regard to fuel cell performance and structure-property relations of the electrodes.

The operation of fuel cell stacks on test benches today is typically monitored by constantalarm threshold values for selected operating conditions. However, due to the wide range ofoperating conditions of the stacks, this type of monitoring only works for extreme maximumand minimum operating conditions. Wide setting ranges for thresholds prevent the detectionof minor faults, whereas too narrow limits will interrupt the tests unnecessarily. A particularchallenge is the monitoring of faults that either do not result in a directly measured responsefrom the fuel-cell stack or that only become noticeable with a time delay. The continuousimprovement of fuel cell stacks over the last years necessarily requires much-improvedmonitoring methods, especially for durability tests with an operating time of thousands ofhours. For this reason, a novel monitoring concept using AI-based methods for the operationof PEM fuel cells on test benches was developed and is being presented.Firstly, machine-learning based mechanisms for monitoring the operating conditions as setand controlled by the test bench are shown. Due to the cyclic operation during durabilitytests, the operating conditions set at a load point can be compared to the past operatingconditions at the same load point. This proceeding allows the early and accurate detection offaults caused by the test bench and thereby ensures the usability of the measured data aswell as an early alarm in case of problems.In addition, a digital twin based method for monitoring the condition of the fuel cell stack ispresented. The deep-learning based digital twin calculates a probabilistic prediction of theexpected voltage of the fuel cell stack based on the current and past operating conditions.The comparison of expected and measured cell voltage taking the model’s confidence intoaccount enables the early and precise detection of unforeseen events such as contaminationand thus averts consequential damage to the fuel cell stack. In contrast to physical models,the digital twin represents a data driven model-ling approach for fuel cells.The presented methods are applied to real testing data to demonstrate the detection of faultsduring the operation of fuel cell stacks on test benches that would have remainedundiscovered by today’s monitoring mechanisms. It shows that the data driven digital twin isable to predict the fuel cell’s stack voltage with an accuracy of 2.5 mV over 1000 h of unseendata. Figure 1

For future Lithium-ion battery applications such as electric vehicles, the battery manufacturing costs and their energy density are key limiting factors that have to be improved, in order to open this technology for a broader market. To address both challenges at once, one promising strategy is to increase the mass loading of state of the art electrodes to extreme values, thus, yielding ultra-thick electrodes. This approach allows to reduce the amount of passive materials in the cell stack in favor of more active material.[1,2] However, an increase in electrode film thickness is usually associated with a variety of drawbacks such as lower rate capability and higher mechanical stress during drying and processing.[3,4] In this contribution, different concepts to overcome these drawbacks are presented and compared. Ultra-thick NMC 622 cathodes with a mass loading of 50 mg/cm2 (≈ 8 mAh/cm2) were prepared and the influence of different manufacturing processing steps (mixing, drying and calendaring) on the electrochemical properties of these electrodes was investigated. Furthermore, the electrode architecture was optimized by pursuing different structuring strategies, in order to improve the Lithium-ion transport. Firstly, the introduction of electrolyte channels by pore-forming agents and electrode perforation is evaluated. Secondly, a multilayer design is explored[5], whereby local porosity and active material particle sizes can be controlled specifically. Finally, the electrolyte concentration was shown to have a significant influence on the electrochemical performance of ultra-thick electrodes and advantageous structures are identified by 3D microstructure resolved simulations based on stochastic structure models, which were calibrated to tomographic image data.[6–8] By application of combined processing and structuring measures, the specific discharge capacity of ultra-thick cathodes at 8 mA/cm2 (≈ 1C) was enhanced by more than 60%. Acknowledgement: The presented work was financially supported by BMBF within projects HighEnergy and PRODUKT under the reference numbers 03XP0073 and 03XP0028.

Novel mobile and stationary applications require batteries with increasing energy density / specific energy, fast-charging capability,1 and long life-time – while maintaining a high safety level. However, there is a dearth of knowledge on the interactions between these parameters. Additionally, the data sheets of commercial cells contain usually only a limited set of parameters, such as a wide temperature range and only one maximum charging C-rate. Therefore, it is highly interesting to correlate electrochemical results with material and electrode parameters. In the present study, we investigated nine types of commercial Li-ion cells (including 18650 and 21700 cells) as well as cells built in our labs. The commercial cells were firstly subject to tests showing their suitability for further experiments. Our experiments reveal interesting specific influences and general trends of materials, electrode and cell properties on specific energy, fast-charging capability, cell resistance, and aging effects. Especially, Li deposition which becomes thermodynamically favorable at negative anode potentials vs. Li/Li+ is to be avoided due to its negative impact on safety and aging.2 In order to judge the fast-charging capability we defined three criteria which have to fulfilled: Charged capacity after fast-charging in comparison to slow charging at 0.1C (>80%)Temperature rise on cell surface (<60°C)Occurrence of Li deposition, estimated from voltage relaxation curves.3,4 Since voltage relaxation curves give only hints on Li deposition while averaging over the whole anode of the cell, it is compared with Post-Mortem analysis,5 measurements of the anode potential vs. Li/Li+ in 3-electrode full cells6 for exemplary cells, and with neutron diffraction data.4 The determination of fast-charging capability is applied to nine types of cylindrical cells for different ambient temperatures (0-40°C) and charging C-rates (0.1-3C), resulting in large data sets for further evaluation. Passing/failing of the three criteria as a function of C-rate and ambient temperature is visualized in fast-charging capability maps (Figure 1c,e). Although the tested commercial cells are from different manufacturers and contain different cathode (NMCs, LFP, NCA, NMC/LMO blend) and anode materials (graphite with different amounts of Si <5%) it is possible to find interesting general trends. For example there is an interaction between the fulfillment of the three fast-charging criteria (Figure 1a) corresponding to either high-energy or high-power cells (Figure 1b,d). In contrast, the change from the 18650 to the 21700 format shows only minor effects on specific energy7 and fast-charging capability. However, the cell format has a significant effect on heating behavior, energy content per cell (~50% more energy in 21700 cells with same electrodes), and therefore most likely a positive impact on costs per produced Wh.7 The study reveals specific influences and general trends in Li-ion cells and gives a perspective for necessary material and electrode improvements to fulfill the requirements for future battery generations.

Lithium-ion battery (LIB) technology is the key enabler of energy storage and electric propulsion. Today´s LIB technology was primarily developed for applications in consumer electronics and is now transferred into automotive applications. These applications demand high specific capacity in order to provide a long driving range as well as high power densities for acceleration, recharging and recuperation. So far, state-of-the-art Lithium-ion batteries do not fully meet the requirements of automotive applications with respect to energy density, power density, extended life time, cost and safety. The commercialized LiFePO4 exhibits good rate capability. Energy density is below that of commercially used layered oxides due to its redox plateau at 3.5 V. A lot of interest is therefore drawn to the isostructural LiMnPO4 due to its higher redox plateau of 4.1 V vs. Li/Li+, which is able to provide higher energy densities [1,2]. Unfortunately, pure LiMnPO4 suffers from low rate capability. On material level, Yamada et al. demonstrated that the partial substitution of manganese in LiMnPO4 with iron, Li(MnxFe1-x)PO4 (LFMP) leads to an improved electrochemical performance and increased energy density [3,4]. On electrode level, a promising strategy to further adjust the electrochemical behavior is to custom-tailor properties by blending complementary types of cathode materials [5-7]. In order to enhance the limited kinetics of manganese based olivine, blending with LiMn2O4-Spinel is expected to result in a promising cathode material for two reasons: a) spinel show fast kinetics, as can be observed from a less pronounced voltage decay at higher C-rates and b) provide a redox plateau in a range comparable to LFMP. Blend electrodes are prepared by mixing the olivine LiFe0.3Mn0.7PO4 (LFMP) and LiMn1.9Al0.1O4 spinel (LMO) in order to obtain a composite electrode, combining the high capacity of LFMP and the rate capability of the spinel. Electrode characteristics are compared with theoretical calculations based on experimental data obtained from electrodes of the single components LFMP and LMO. While tap density of the powders and electrode parameters show a linear dependency on the blend ratio, remarkable synergetic effects can be observed regarding the electrochemical performance at high rates (3C). Potential curves of blend electrodes at rates of 3C reveal a less pronounced voltage decay for the Mn2+/3+plateau than expected from theoretical calculations as depicted in Figure 1a. This buffer effect is also observed for high current pulses (5C) where blend electrodes resemble the behavior of pure spinel electrodes. In terms of power density at high states of charge (SoC), the performance of the investigated blends exceeds even that of pure spinel. In addition, the spinel-related manganese dissolution can be drastically reduced by blending spinel with LFMP. Structural changes in blend electrodes during pulse power discharge of the manganese redox plateau are investigated with in-situ XRD measurements. Shifts in XRD pattern during relaxation (I = 0) gives us indication of particle-to-particle interactions between LFMP and LMO. Figure 1b illustrates a comparison between the battery relevant features of LFMP, spinel and a blend with LFMP:LMO = 50:50 cap%. This study shows the expected and synergetic effects of LFMP/spinel blends and compares the results with theoretical calculations [8]. The complementary properties of the single component materials LFMP and spinel result in a cathode material which can be promising for 48V starter batteries in automotive applications, where high power densities at high states of charge are required in order to start the internal combustion engine.

Rising demands for high-energy density Li-ion batteries require the development of new cathode active materials with enhanced specific capacity and energy, increased safety and reduced cost. Currently used high-cost metals, such as cobalt and nickel, need to be replaced with more abundant and less toxic elements. A Co-free cathode material with high-voltage and high-capacity has been developed in our labs. The lithium‐nickel‐manganese oxide with composition Li1+xNi0.5Mn1.5O4.0 (0<x<0.5) was tailored with respect to morphology, particle size and shape distribution. The particles consist of densely packed nano‐sized primary crystallites and the free‐flowing powder shows a tap‐density of 2.4 g cm-3. The cathode material shows a specific capacity of about 210 mA h g-1 in the potential window 2.4-4.9 V. The working mechanism involves two distinct voltage plateaus at 4.7 and 2.9 V vs. Li/Li+. Very good cycling stability is obtained for more than 200 cycles in half-cell. Compared to other high-capacity materials, such as lithium-manganese-rich layered oxides, their spinel integrated variants and the cation disordered materials [1-3], this material does not need an activation cycle; it shows good rate capability and low voltage hysteresis. In full-cell vs. graphite anode, a stable capacity of 160 mA h g-1 is reached. For full-cell application no electrochemical pre-treatment or structural activation has to be applied in order to achieve the full capacity; therefore, the material can be directly assembled vs. graphite anode. Cathode material features will be described, electrochemical performances in half- and full-cell will be presented and the influence of the operating voltage window will be shown. The effects of particle shape, grain architecture and structural features on the electrochemical behavior will be discussed. Phase transitions in the low-potential region have been studied by HRTEM, EELS and XRD and the involved mechanism will be discussed. Acknowledgements This work was supported by the German Federal Ministry of Education and Research (BMBF) in the project Li-EcoSafe (03X4636A).