- Part 1 - Coating tests and coating combinations
In a joint project, a sensor fuel cell was developed that can qualify material and monitor hydrogen systems. Its central components consist of 3D-printed polyamide-12 (PA12), which was functionalized using galvanic and electroless processes. A combination of chemical nickel deposition and subsequent galvanic copper, nickel and gold deposition was chosen for the coating. The coatings showed good adhesion to the plastic substrate, high conductivity and complete hydrogen impermeability.
The finished sensor fuel cells developed as part of the IGF project 22754 N showed performance values of over 0.12 W/cm2 in cell operation without external humidification and temperature control. The stable voltage signals of the cells are highly sensitive to pollutant emissions, making them suitable for material qualification and monitoring tasks for hydrogen systems with high purity requirements.
Introduction
Hydrogen energy is a flexible and environmentally friendly energy source that is increasingly seen as the key to the transition to a sustainable and future-proof energy supply [1-3]. Especially in the transportation sector, fuel cell technology in combination with green hydrogen offers a promising option for the transition to sustainable mobility [4, 5]. Significant advantages in this area result from the high efficiency and range of fuel cells combined with short refueling times. In addition, the relatively simple scalability of the technology allows it to be adapted to different performance requirements [6]. However, the high thermal and chemical loads to which the fuel cell components are exposed during operation, as well as the long operating times required, place considerable demands on a large number of materials in the fuel cell systems [7]. For this reason, cost-intensive high-performance materials are usually used, which have the required corrosion and temperature resistance, but also lead to higher costs compared to competing energy systems. There is therefore enormous potential for savings in fuel cell technology by replacing expensive high-performance materials with more cost-effective materials such as plastics [8]. Simple methods for material characterization must ensure that these new materials are suitable for fuel cells in the long term. In addition to the known factors such as the required temperature stability and hydrogen permeability, unknown aspects such as impurities and chemical residues, for example from the manufacturing process, which can damage the fuel cell, pose a problem. The rapidly growing field of 3D printing of plastics in particular offers great potential due to the wide variety of materials and processes. At the same time, material qualification in fuel cell technology is playing a greater role in identifying possible sources of pollutants. For in-situ testing, (sensor) fuel cells can be exposed to various potentially problematic materials and possible pollutant sources in special test chambers in order to analyze their effects. These sensor cells must react extremely sensitively to contamination from air or hydrogen in order to enable precise identification of pollutant sources. As the test cells are contaminated by material outgassing and can be irreversibly damaged in the process, they act as consumables that need to be replaced regularly. To ensure the cost-effectiveness and sustainability of the test procedures, it is therefore crucial that sensor cells not only work effectively but can also be produced cost-effectively. In addition to being used for material qualification, these sensor cells can also be used to monitor the constantly growing hydrogen infrastructure and in the field of rapid prototyping.
The results presented in this paper were produced as part of the funded IGF project 22754 N between the fem, the Center for Fuel Cell Technology ZBT GmbH and the Institute of Microstructure Technology at the Karlsruhe Institute of Technology (KIT). The aim of the project was to develop a test/sensor fuel cell with a reduced active surface area that can be used as a central test, inspection and monitoring element. This was achieved through a combined process of 3D-printed bipolar plates, surface activation and electroplating processes in order to functionalize the 3D-printed plastic products with electrically conductive areas. In addition, reliable metallization ensures hydrogen impermeability and prevents pollutants from the 3D component from reaching and damaging the membrane electrode unit of the fuel cell.
Experimental
Coating tests
As part of the work, various sample body geometries made of polyamide-12 (PA12) were used for the coating tests. Initially, simple flat samples without channel structures were examined. In some cases, glass beads were added to the plastic in order to improve the mechanical properties. In addition, selected samples were chemically smoothed to reduce the surface roughness caused by 3D printing. The effect of the material composition and chemical smoothing on the coatability of the surface was investigated. Test specimens with geometrically varying channel structures were constructed to investigate the influence of the flow field design on the galvanic and chemical coating processes. The optimized coating tests were finally carried out on the 3D-printed, shaping test specimens of the fuel cell. Details on the test specimens and associated CAD images can be found in the chapter on material and specimen selection.
The deposition tests were carried out both galvanically and without external current. In the electroless nickel and gold deposition tests, the plastic surface was first pre-treated with acetone and then treated with a reduction and activator solution suitable for PA12 (direct metallization Slotosit SEA 2800 from Schlötter). Various methods were used to produce the metallic starting layers for the electroplating tests. On the one hand, the samples were coated with conductive Cr-Au or Al2O3-Cr-Au starting layers. Alternatively, Cu starter layers were used, which were applied using physical vapor deposition (PVD).
In the electroplating process, target layer thicknesses of around 30 µm copper, 5 µm nickel and 1-2 µm gold were achieved by adjusting the deposition times using the following electrolytes: copper 830 (Umicore), Cu 375 (RIAG), nickel sulphamate MS (Schlötter) and Auruna 556 (Umicore). The experiments without external current were carried out using the electrolytes Slotosit KM 3400 for nickel and Slotogold IG 2420 for gold (both from Schlötter).
Fig. 1: SEM images of the surface of the PA12 plastic with glass beads a) untreated surface
and b) chemically smoothed surface
Material characterization
Scanning electron micrographs of the surfaces and cross-sections were taken (Gemini SEM 300, Zeiss) to assess the coating quality and thickness of the different coating structures. Surface resistance and contact resistance were measured (contact resistance according to DoE criteria: Bipolar plate between two layers of GDL, current flow via metallic coating) were determined using four-wire measurement in order to minimize measurement errors due to lead resistances and the contacting of the measurement leads. Data acquisition and evaluation were carried out on a conductivity measuring stand equipped with SORAYA software specially adapted to the measurement task, which handles both the control and the acquisition of resistance and force measurement data. Copper and stainless steel were used as reference materials to determine the contact resistance. The contact resistance was determined as a function of the applied contact force. The test setup for measuring the contact resistance is described in Wang et al [9]. The sample was clamped between two Toray papers and two cylindrical copper blocks. The diameter of the upper copper block was 2.5 cm. In this way, a defined force could be exerted on the sample.
Hydrogen permeability measurements were carried out in a Devanathan cell. For this purpose, the different layer stacks were deposited on a reference steel sheet and the signal of the cell was compared over time. A current measurement on the other side of the sheet indicates whether hydrogen has diffused through the sample. An uncoated steel sheet was used as a reference. Corrosion measurements of the coatings were carried out using the aggravated acetic salt spray test (AASS) DIN 9227 over 48 hours.
Selection of materials and test specimens for the tests
As part of the project, the plastic polyamide 12 (PA12), both in the version with and without glass beads, was selected as the starting material for the shaping 3D printed component of the fuel cell. PA12 is a highly resistant polyamide that is characterized by particularly advantageous properties such as low moisture absorption, high chemical resistance and excellent flexibility. It also offers outstanding mechanical stability, which can be further increased by adding glass beads. The corresponding test specimens were manufactured using the 3D printing process Multi Jet Fusion (MJF) from Hewlett Packard. This process offers several advantages, such as high temperature stability up to 170 °C, excellent H2 pressure tightness and high mechanical stability. To further improve the mechanical properties, it was analyzed to what extent the addition of glass beads strengthens the material and what effects this has on the galvanic coating. In addition, it was investigated whether chemical smoothing can reduce the pronounced surface roughness of the printed material and thus positively influence the coating quality.
Various sample geometries were used as part of the work. Investigations into the influence of the glass beads, the smoothing and activation of the plastic surface as well as initial coating tests were carried out on simple flat samples without channel structures. First, the surface properties of the 3D-printed components were thoroughly analyzed, particularly with regard to their chemical resistance, surface roughness, morphology and wettability. The chemical resistance was tested at pH 1 (sulphuric acid) and pH 13 (sodium hydroxide solution) and the material showed good resistance.
The roughness measurements showed Ra values of 6.8-10.2 µm and Rz values of 42.2-56.9 µm for the directly printed PA12 flat samples. For chemically smoothed samples, the Ra values were 3.1-5.5 µm and the Rz values 15.4-25.7 µm. Corresponding SEM images of the surface are shown in Figure 1.
Both the SEM images of the surface and the measured roughness values confirm that the chemical smoothing process leads to a significant leveling of the surface. In the further course of the project, it was also necessary to check whether and to what extent this smoothing affects the chemical and galvanic metallization as well as the operation in the fuel cell.
Bipolar plates in a fuel cell usually have a structured channel structure (flow field) to ensure uniform distribution of the reaction gases, removal of the reaction products and regulation of the temperature. When electroplating bipolar plates with a complex flow field design, problems such as overcoating at the edges and insufficient material coverage on the walls of the channels often occur, especially in deep, relatively vertical structures. These inconsistencies can significantly affect the corrosion resistance and service life of the bipolar plate. Electroless coating processes such as electroless nickel offer advantages here, as the process is based on chemical reduction and thus enables a more even coating, even in areas that are difficult to access. However, the diffusion limitation of the electrolyte, for example, must also be taken into account in this process. In order to investigate the influence of the flow field design on the galvanic and chemical coating processes, test specimens with geometrically varying channel structures were constructed as part of the work.
Figure 2 shows a CAD model created with different test channels and gas inlets for the design of anode and cathode-side flow field structures in the sensor cell components.
Finally, the optimized coating tests were carried out on the 3D-printed shaping test bodies of the fuel cell from Figure 3 and Figure 4. Figure 3 shows the outside and Figure 4 the cell side of the CAD model of the 3D-printed base body created for the anode side of the sensor cell.
Fig. 2: CAD model of a 3D printed test specimen for evaluating different channel and web properties for coating tests
Fig. 3: CAD model of the 3D printed body of the anode side of the sensor cell with flow field
Fig. 4: CAD model of the 3D printed body of the anode side of the sensor cell with connections
The external dimensions are determined by the desired active cell area and the space required for assembly and connection technology and are 49 mm (X) × 49 mm (Y) × 8 mm (Z). The material thickness along the Z-axis should be sufficient for PA12 and the necessary contact pressures even without additional reinforcement with glass beads. The flow field is designed in the form of a single-channel meander and supplies an active cell area of 5 cm2. The depth of the flow field is 0.5 mm on the anode side and 1 mm on the cathode side. The channel and web widths vary between 0.8 and 1.2 mm, depending on the angle of the side walls. The active cell surface is surrounded by a square sealing groove.
Metallization of the shaping 3D printed component of the sensor cell
As already explained in the chapter on material and specimen selection, the metal deposition tests were carried out iteratively on different specimen geometries. Corrosion-resistant, electrically conductive layers are required to ensure the functionality of the bipolar plates in the fuel cell. As part of the present work, various coating systems and sequences were initially investigated on the simple flat samples. The work can be roughly divided into electroplated and electroless plated layers. Electroplated layers of copper, nickel and gold were applied. Copper has excellent conductivity for the cell current, while the gold layer ensures ideal cell contact with the gas diffusion layer of the MEA and at the same time offers a high level of corrosion protection. A nickel layer was integrated between the copper and gold layers as a diffusion barrier. In addition, electroless nickel and gold depositions and combinations of chemical and galvanic processes were investigated. The aim here was to achieve the optimum in terms of material use, H2 tightness, corrosion resistance, conductivity and connection to the gas diffusion layer of the fuel cell with optimized process times and stability.
Both electroless and galvanic metal deposition on plastics require activation of the surface for optimum layer adhesion. Various methods and systems were also evaluated as part of the investigations.
The following chapters present the results of the various activation methods: the galvanic tests, the results of the electroless deposition tests and finally the summary of the main results and the presentation of the most promising coating combinations and their transfer to the final test specimen geometry.
Activation of the plastic surface for the coating tests
The first step was to activate the plastic surfaces in preparation for the coating tests. Activation is necessary for both the galvanic and the electroless build-up of the metal layers. However, while the electrochemical processes require a continuous conductive starting layer in order to carry the current during the process, chemical deposition processes do not require a continuous conductive starting layer.
Initial activation tests were carried out on various flat samples. Both the chemical smoothing of the printed surface and the addition of glass beads in various combinations were taken into account. This was followed by a transition to the structured samples in Figure 2, before the final samples in Figure 3 were coated with the optimum activation and coating parameters.
For the purely electroless deposition tests of nickel and then gold, the surface of the plastic was pre-treated with acetone and then treated with a commercially available reducer and activator solution suitable for PA12 (direct metallization Slotosit SEA 2800 from Schlötter). Palladium nuclei act as catalysts that initiate the nickel deposition process by promoting the adsorption of nickel ions and thus enabling a uniform and stable nickel coating.
Various methods were used to produce the metallic starting layers for the purely galvanic coating tests and their suitability was evaluated. On the one hand, samples were coated with conductive Cr-Au or Al2O3-Cr-Au starting layers by electron beam evaporation (Leybold Univex 400). On the other hand, Cu starting layers were applied to samples by physical vapor deposition. The results show that all investigated starting layers can be successfully applied to the surfaces. Figure 5 shows various samples that were provided with a Cr-Au starter layer.
As part of the further work, the optimized starting layer or activation layer was selected for each coating method or layer sequence. The different approaches allow a comprehensive variation of the layer structure. The influence of the respective starting layer on the coating adhesion and its effect on the final result is examined and evaluated in detail in the rest of the chapter.
Fig. 5: PA12 samples with Cr-Au starting layers a) with chemical smoothing of the plastic and b) without chemical smoothing (sample size 50 mm x 20 mm x 2 mm, CAD model see Figure 2)
Electroplating process
Initially, a purely galvanic process was carried out in which galvanic Cu, Ni and Au layers were applied to the various metallic starting layers (Cu-PVD, Cr-Au and Al2O3-Cr-Au starting layers). When looking at the different substrate materials in the chapter on material and specimen selection, it became clear that chemical smoothing and the addition of glass beads in particular have a negative effect on the coating process and coating adhesion.
On the rough surface of the unsmoothed material, the coating can penetrate into the microstructures, which can significantly improve adhesion due to the so-called "push-button effect". Another potential advantage of the rough surface could be the improved connection with the gas diffusion layer (GDL) in fuel cell operation. In the further course of the investigations, however, it must be checked whether corrosion-resistant, closed layers can also be deposited despite the roughness. As the material was sufficiently stable for the fuel cell application even without glass beads, they were not used in the further course.
In the electroplating process, target layer thicknesses of around 30 µm copper, 5 µm nickel and 1-2 µm gold were deposited using the electrolytes copper 830 (Umicore), Cu 375 (RIAG), nickel sulphamate MS (Schlötter) and Auruna 556 (Umicore). Adhesion-resistant layers were produced on the Cu PVD starting layers. In the course of the depositions, however, it became apparent that reliable adhesion of the coatings could not be achieved on either smoothed or unsmoothed samples to which the various Cr-Au starter layers were applied. This was particularly problematic on the smoothed samples.
To characterize the different layers, surface images were taken using SEM and cross-sections were prepared. In addition, adhesion tests and layer thickness measurements were carried out using XRF and in the cross-section.
Figure 6 shows scanning electron micrographs of electrodeposited Cu, Ni and Au layers. The substrate material used was unsmoothed PA12 plastic with glass beads, to which a Cu PVD starting layer was applied.
It was shown that galvanically crack-free and adherent layers can be deposited on the flat samples. The electrolytic copper electrolyte also leads to a significant leveling of the substrate material.
Before being transferred to the final specimen geometry, the nickel and gold deposits without external current were first examined. The aim was to determine the optimum layer sequence, possibly also through a combination of chemical and galvanic processes.
Fig. 6: Scanning electron micrographs of (left) cross-section and (right) the surface of electrodeposited Cu, Ni and Au layers on a Cu PVD starting layer (unsmoothed PA12 plastic with glass bead additive)
Chemical coating process
In the further course of the work, the purely electroless deposition of nickel and gold on the flat samples was investigated. One advantage of this method for coating channel structures is that chemical electrolytes enable a much more uniform layer formation, especially in recesses and areas that are difficult to access. In addition, chemical deposition does not require a continuous metallic starting layer, but only nucleation, which significantly reduces the process costs. However, one disadvantage of chemical nickel deposition is the lower electrical conductivity compared to electrolytically deposited copper. In addition, the layer thicknesses in electroless depositions are generally lower than in electrolytic processes, which can produce thicker layers. The ability to level out unevenness is also much more limited with chemical electrolytes, as these tend to deposit the base material and are less able to level out existing unevenness. Due to the different advantages and disadvantages, it is therefore likely that a combination of chemical and galvanic processes will ultimately be the most sensible solution. To characterize the layers, surface images were also taken using SEM and cross-sections were prepared. In addition, adhesion tests and coating thickness measurements were carried out using XRF and in the cross-section. Even with purely chemical deposition, it was not possible to achieve a reliably adhesive coating on the chemically smoothed samples. Therefore, the chemical smoothing step was also omitted in this case.
Figure 7 shows scanning electron micrographs of the cross-section of a chemically produced sample. A nickel layer about 500 nm thick and a gold layer about 150 nm thick can be seen. Due to the unsmoothed surface, a clear roughness is visible, which is not leveled by the electroless nickel layer. The possible layer thickness of the chemical gold is also very limited, so that it had to be clarified in the further course whether this was sufficient for the required corrosion resistance.
Fig. 7: Scanning electron micrographs of a PA12 flat sample after chemical activation and the layer structure of chemical Ni (Slotosit KM 3400) and Au (Slotogold IG 2420)
Summary of the coating tests and coating combinations
It was successfully demonstrated that both galvanic and electroless processes are suitable for depositing various coatings on the flat samples. The chemical smoothing step of the printed PA12 plastic posed a particular challenge, which is why this process step was ultimately omitted.
As a metallic starting layer is always required for electroplated coatings, Cr-Au and Cu (using PVD) were initially used as possible variants. However, both methods proved to be more complex in terms of process technology, and Cr-Au also showed insufficient adhesion to the plastic substrate.
For this reason, chemically deposited nickel was chosen as the starting layer for all further tests. This can be applied reliably and firmly after simple activation of the plastic with an activator-reducer solution. Different coating systems and sequences can then be easily built up on the nickel layer.
Two different coating sequences in particular were selected for further intensive investigation and transfer to the final test specimen geometries: Firstly, a coating of pure electroless nickel, supplemented by a thin layer of gold, was applied. On the other hand, a layer sequence consisting of a thin electroless nickel layer followed by an electroplated Cu/Ni/Au coating was realized.
In the next chapter, which will appear in Galvanotechnik 8/2025, these two layer combinations will be comprehensively characterized in order to test the conductivity and corrosion resistance of the coatings and to ensure that the layers are hydrogen-tight. In particular, the lower conductivity of chemically deposited nickel compared to copper could represent a critical point for the purely chemical layers. In addition, the chemical gold layer could be too thin for sufficient corrosion resistance. After the thorough characterization, the transfer of the optimized coating system to the final test specimen geometries is described.
The second and final part will be published in Galvanotechnik 8/2025.44
Literature
[1] Hassan, Q., Azzawi, I. D. J., Sameen, A. Z. & Salman, H. M. Hydrogen Fuel Cell Vehicles: Opportunities and Challenges. Sustainability 15, 11501 (2023).
[2] Mustafa Inci. Future vision of hydrogen fuel cells: A statistical review and research on applications, socio-economic impacts and forecasting prospects. Sustain. Energy Technol. Assess. 53.
[3] Staffell, I. et al. The role of hydrogen and fuel cells in the global energy system. Energy Environ. Sci. 12, 463-491 (2019).
[4] Dash, S. K., Chakraborty, S., Roccotelli, M. & Sahu, U. K. Hydrogen Fuel for Future Mobility: Challenges and Future Aspects. Sustainability 14, 8285 (2022).
[5] Ephraim Bonah Agyekum, Flavio Odoi-Yorke, Agnes Abeley Abbey, Godwin Kafui Ayetor. A review of the trends, evolution, and future research prospects of hydrogen fuel cells - A focus on vehicles. Int. J. Hydrog. Energy 72, 918-939.
[6] Cullen, D. A. et al. New roads and challenges for fuel cells in heavy-duty transportation. Nat. Energy 6, 462-474 (2021).
[7] Ahmad, S. et al. An overview of proton exchange membranes for fuel cells: Materials and manufacturing. Int. J. Hydrog. Energy 47, 19086-19131 (2022).
[8] Sazali, N., Wan Salleh, W. N., Jamaludin, A. S. & Mhd Razali, M. N. New Perspectives on Fuel Cell Technology: A Brief Review. Membranes 10, 99 (2020).
[9] Wang, H. Stainless steel as bipolar plate material for polymer electrolyte membrane fuel cells. J. Power Sources 115, 243-251 (2003).
