Recycling of catalyst coated membrane components

Abstract

A method for recycling a catalyst coated membrane comprising an ionomer membrane coated on either side by catalyst layers, the method comprising: soaking the catalyst coated membrane in a first solvent; transferring the catalyst coated membrane from the first solvent into a second solvent which is different to the first solvent; and subjecting the catalyst coated membrane in the second solvent to an ultrasonic treatment to delaminate the catalyst layers from the ionomer membrane.

Description

[0001] RECYCLING OF CATALYST COATED MEMBRANE COMPONENTS

[0002] Field

[0003] This specification relates to recycling methods for components of catalyst coated membranes such as those used in fuel cells and hydrogen producing water electrolysers.

[0004] Background

[0005] Fuel cell and hydrogen producing water electrolyser production is set for rapid growth as investment is placed into the global hydrogen economy. Catalyst coated membranes (CCMs) are a major functional component of both fuel cells and electrolysers. Such CCMs generally comprise a conductive polymer membrane coated on either side by a catalyst containing layer. The CCMs are configured to drive oxidation and reduction reactions and support proton and electron transport, these processes being required for the fuel cell and electrolyser technologies to function.

[0006] While variations in CCM component materials and configurations exist according to functional performance requirements in end use applications, they generally contain several components of value including one or more platinum group metal (PGM) catalysts and one or more proton conducting polymers.

[0007] Typically, the membrane is formed of one or more ionomers such as perfluorosulfonic-acid (PFSA) ionomers. Ionomer may also be provided in one or both of the catalyst layers. The ionomer in the catalyst layers may be the same or different to the ionomer in the main membrane component and / or in other catalyst layer(s).

[0008] A CCM may comprise two different catalysts, one for driving an oxidation reaction on one side of the CCM and one for driving a reduction reaction on the other side of the CCM. A CCM may also comprise a recombination catalyst which is provided to catalyse the recombination of hydrogen and oxygen to form water, reducing the quantity of hydrogen crossing the membrane and mixing with oxygen to form a potentially explosive mixture. A CCM may also include a peroxide scavenger, e.g., a metal oxide such as CeO2.

[0009] CCM catalysts can be based on platinum group metals such as platinum, ruthenium, iridium, palladium, or mixtures thereof. The platinum group metals may be provided in elemental (metallic) form, in compound form (e.g., an oxide, such as an iridium oxide catalyst), or as a PGM-base metal alloy (e.g., PtCo). Furthermore, the PGM catalyst materials may be supported on a substrate material, such as a carbonaceous substrate material (e.g., carbon, such as a platinum-on-carbon catalyst comprising particles of carbon on which platinum is disposed or PtCo-on-carbon, or an organic material, e.g., nanostructured thin film catalyst (NTFC) technology as described in US2020102659 and WG2006089180).

[0010] Catalyst coated membranes (CCMs) can also be provided in combination with additional functional layers to form multi-layer membrane electrode assemblies (MEAs). Such MEAs may have 3, 5, or 7 layers for example.

[0011] With the increase in CCM manufacture for fuel cells and electrolysers, there is an associated increase in CCM waste materials, including a significant volume of scrap material created during CCM manufacture (e.g., due to failure at quality control) and also an increase in end-of-life (EoL) CCMs. Since CCMs contain several components which are rare and / or valuable, including platinum group metals (notably Pt, Pd, Irand Ru) and ionomer (both in the membrane and catalyst layers), there is a growing demand for methods of recycling such components from waste CCM materials.

[0012] Typically, greater than 80% of the ionomer to be recycled is present in the central membrane and greater than 80% of the catalyst materials are present in the catalyst layers coating the central membrane. As such, a process for delaminating the catalyst layers to separate them from the central ionomer membrane enables the bulk of the ionomer to be processed and recovered separately from the bulk of the catalyst material.

[0013] WO2024115878 (Johnson Matthey PLC) describes such a CCM recycling method for a waste catalyst coated membrane comprising a membrane including a membrane ionomer, a first catalyst layer disposed on one side of the membrane, the first catalyst layer comprising a first catalyst and a first catalyst layer ionomer, and a second catalyst layer disposed on an opposite side of the membrane, the second catalyst layer comprising a second catalyst and a second catalyst layer ionomer. The recycling method comprises the following steps:

[0014] contacting the waste catalyst coated membrane with a solvent to delaminate both of the first and second catalyst layers from the membrane without dispersing the membrane, wherein the first and second catalyst layers form a catalyst layer slurry comprising the first catalyst, the first catalyst layer ionomer, the second catalyst, and the second catalyst layer ionomer;

[0015] separating the membrane from the catalyst layer slurry;

[0016] processing the membrane to recover the membrane ionomer; and

[0017] processing the catalyst layer slurry to disperse and recover the first and second catalyst layer ionomers in a solvent, and separate and recover the first and second catalysts or components thereof.

[0018] It is described that the solvent used to delaminate both of the first and second catalyst layers from the membrane can be a mixture of an alcohol and water and that the solvent can be agitated to aid delamination of the catalyst layers, e.g., by sonication / insonation / sonification / ultra-sonification. Furthermore, while the solvent may be heated to aid delamination, it can be maintained at a temperature which is sufficiently low such that the catalyst layers delaminate without dispersing the fluorinated polymer membrane which remains in solid, undispersed form. Several other known CCM recycling methods also use a catalyst layer delamination step. For example, a paper entitled “PEM water electrolysis: Innovative approaches towards catalyst separation, recovery and recycling” (International Journal of Hydrogen Energy 44 (2019) 3450-3455) discloses a method of recycling a CCM including mounting the CCM across a reactor and subjecting both sides of the CCM to a circulation of a solution consisting of deionized water and alcohol such that delamination of the catalyst layers from the membrane occurs after 10-30 minutes.

[0019] CN106898790 and US2007 / 292745 also disclose CCM recycling methods in which catalyst layers are delaminated from an ionomer membrane in a solvent (using an alcohol or an alcohol-water mixture as the solvent) and then the solid membrane is separated from the delaminated catalyst layer material. Ultrasonic treatment in the solvent is also envisaged.

[0020] Summary of Invention As described in the background section, various documents disclose CCM recycling methods comprising a delamination step which involves contacting the CCM with a solvent (e.g., an alcohol ora mixture of an alcohol and water) and, at least in some methods, this may also involve heating and / or agitation (e.g. by sonification) of the CCM in the solvent to encourage delamination of the catalyst layers.

[0021] However, the present inventors have identified that this type of delamination methodology may not be optimal for one or more of the following reasons:

[0022] (i) Catalyst layer material may not be cleanly and completely separated from the ionomer membrane.

[0023] (ii) Delamination of the catalyst layer material may be accompanied by potentially hazardous gas release.

[0024] (iii) The solvent used for the delamination process (e.g. an alcohol or an alcohol / water mixture) may not be one which is desired for further processing of the delaminated components.

[0025] Having identified these problems with the prior art approaches, the present inventors have developed an improved method for recycling catalyst coated membrane comprising an ionomer membrane coated on either side by catalyst layers, the method comprising:

[0026] soaking the catalyst coated membrane in a first solvent;

[0027] transferring the catalyst coated membrane from the first solvent into a second solvent which is different to the first solvent; and

[0028] subjecting the catalyst coated membrane in the second solvent to an ultrasonic treatment to delaminate the catalyst layers from the ionomer membrane.

[0029] The first solvent is preferably an organic solvent, optionally an alcohol, optionally ethanol.

[0030] The second solvent is preferably an aqueous solvent, optionally water.

[0031] It has been found that this two-step delamination process in which the CCM is soaked in a first solvent (e.g., ethanol) and then transferred and subjected to a sonification treatment in a second solvent (e.g., water) can address the aforementioned problems and provide a process which can completely delaminate the catalyst layers while safely avoiding any hazardous gas release and resultingin delaminated components in a desired (e.g. aqueous) solvent after delamination ready for further processing.

[0032] Brief Description of the Drawings

[0033] For a better understanding of the present invention and to show how the same may be carried into effect, certain embodiments of the present invention will now be described by way of example only with reference to the accompanying drawings.

[0034] Figure 1 (a) shows a photograph of a fuel cell CCM.

[0035] Figure 1 (b) provides an SEM image of the fuel cell cross-section.

[0036] Figures 1 (c) and (d) depict SEM imagesofa Pt / C cathode layer a nd a Pt / C, IrTaOx anode layer (the displayed element compositions were obtained via sampling across the entire image). Figures 1 (e) and (f) display TEM images of Pt / C cathode and Pt / C, IrTaOx anode particles within the fuel cell CCM, respectively.

[0037] Figure 2 (a) shows a photograph of a water electrolyser CCM.

[0038] Figure 2 (b) provides an SEM image of the water electrolyser CCM cross-section.

[0039] Figures 2 (c) and (d) illustrate Pt / C and IrOx layers.

[0040] Figures 2 (e) and (f) present TEM images of the Pt / C and IrOx particles, respectively.

[0041] Figure 3 (a) shows an initial fuel cell CCM.

[0042] Figure 3 (b) shows a cleaned central fuel cell membrane.

[0043] Figures (c) shows an initial water electrolyser CCM.

[0044] Figure 3 (d) shows a cleaned central water electrolyser membrane.

[0045] Figures 4 (a) and (b) are graphs showing delamination percentage in water with an ultrasonic bath of a) fuel cell CCM and b) water electrolyser CCM after swelling in different solvents.

[0046] Figures 4 (c) and (d) shows pictures of (c) fuel cell CCM and (d) water electrolyser CCM after soaking in acetone, ethanol, and water, followed by sonication in water (10 min for fuel cell CCM and 12 min forwater electrolyser CCM).

[0047] Figures 5 (a) to (d) show SEM images of delaminated membranes: (a) and (b) are Pt / C and Pt / C, IrTaOx sides of the fuel cell membrane; (c) and (d) are Pt-C and IrOx sides of water electrolyser membrane.

[0048] Figures 6 (a) to (d) show TEM images of delaminated particles: (a) and (b) are a mixture of Pt / C and Pt / C, IrTaOx sides from the fuel cell membrane; (c)and (d) are a mixture of Pt / C and IrOx sides from water electrolyser membrane.

[0049] Figure 7 shows an example of a high-intensity ultrasonic delamination unit.

[0050] Figure 8 shows schematic images of ultrasonic delamination: (a) with a cylindrical sonotrode; and (b) with a custom-built blade sonotrode. The cylindrical sonotrode configuration is set up to delaminate CCMs in batches, whereas the blade sonotrode is optimised for continuous processing by incorporation of roller blades to facilitate continuous sample feed.

[0051] Figure 9 shows SEM images of: (a) an initial fuel cell at the Pt / C cathode side and the Pt / C, IrMOx anode side; and (b) a clean area of delaminated membrane surfaces at the Pt / C cathode and the Pt / C, IrMOx anode side (40 W cm'2, 5 mm sonotrode-to-sample distance, and 1 second of sonication for each side of the sample).

[0052] Figure 10 shows images taken from a high-speed camera (20k frames per seconds) during ultrasonic delamination of the fuel cell sample in water: (a) Pt / C cathode side facing sonotrode; (b) Pt / C, IrMOx anode side facing sonotrode. The sonotrode was 20 mm in diameter, with 40 W cm-2power intensity, at 5 mm from the substrate.

[0053] Figure 11 shows delamination of fuel cell CCM induced by ultrasonic cylinder sonotrode: (a) the Pt / C cathode side; and (b) the Pt / C, IrMOx anode side of a CCM sample after ultrasonic delamination in water (10% power, 1 -second sonication time, 5 mm sonotrode-sam pie distance). (i) Initial sample, (ii) sample after drying (image adjusted for brightness and contrast), and (iii) corresponding scanning electron microscopy (SEM) image of the delamination interface.

[0054] Figure 12 shows characterization of fuel cell CCM samples before and after the delamination process, (a) Comparative images of CCM samples before and after delamination, (b) Scanning electron microscopy (SEM) images of the membrane surface after delamination at the Pt / C cathode side and Pt / C, IrMOx anode side. Delamination conditions: blade sonotrode, 40 W cm'2power intensity, 1 cm s-1viscosity feeding speed, ambient temperature, 5 mm sonotrode-to-sample distance.

[0055] Detailed Description

[0056] As described in the summary of invention section, the present specification provides a method for recycling catalyst coated membrane comprising an ionomer membrane coated on either side by catalyst layers, the method comprising:

[0057] soaking the catalyst coated membrane in a first solvent;

[0058] transferring the catalyst coated membrane from the first solvent into a second solvent which is different to the first solvent; and

[0059] subjecting the catalyst coated membrane in the second solvent to an ultrasonic treatment to delaminate the catalyst layers from the ionomer membrane.

[0060] It has been found that this two-step delamination process in which the CCM is soaked in a first solvent and then transferred and subjected to a sonification treatment in a second solvent can address the aforementioned problems and provide a process which can completely delaminate the catalyst layers while safely avoiding any hazardous gas release and resulting in delaminated components in a desired (e.g. aqueous) solvent after delamination ready for further processing. Optionally, the first solvent comprises or consists of an organic solvent, optionally an alcohol, optionally ethanol.

[0061] Optionally, the catalyst coated membrane is soaked in the first solvent for a time period of: at least 5 seconds, 10 seconds, 20 seconds, 30 seconds, 40 seconds, 50 seconds, 60 seconds, 2 minutes, 5 minutes, 10 minutes; no more than 5 hours, 60 minutes, 30 minutes, 20 minutes, 10 minutes, 8 minutes, 6 minutes, 4 minutes, 2 minutes, 1 minutes; and / orwithin a range defined by any combination of the aforementioned lowerand upper limits.

[0062] Optionally, the catalyst coated membrane is soaked in the first solvent at a temperature of: at least 0°C, 5°C, 10°C, 15°C or 20°C; no more than 80°C, 60°C, 50°C, 40°C, or 30°C; and / or within a range defined by any combination of the aforementioned lower and upper limits.

[0063] Optionally, the catalyst coated membrane is soaked in the first solvent at room temperature and pressure.

[0064] Optionally, the second solvent comprises or consists of an aqueous solvent, optionally water, e.g., deionized water. The second solvent can be a mixture of water and an organic solvent, optionally a mixture ofwaterand an alcohol, e.g., a mixture of waterand propanol.

[0065] Optionally, the ultrasonic treatment is for a time period of: at least 0.1 second, 1 seconds, 2 seconds, 5 seconds, 10 seconds, 20 seconds, 40 seconds, 1 minute, 5 minutes, or 10 minutes; no more than 60 minutes, 30 minutes, 20 minutes, 15 minutes, 10 minutes, 5 minutes, 1 minute, 30 second, or 10 seconds; and / or within a range defined by any combination of the aforementioned lower and upper limits.

[0066] Optionally, the ultrasonic treatment is at a temperature of: at least 0°C, 5°C, 10°C, 15°C or 20°C; no more than 60°C, 50°C, 40°C, or30°C; and / or within a range defined by any combination of the aforementioned lower and upper limits.

[0067] Optionally, the ultrasonic treatment is at room temperature and pressure.

[0068] Optionally, after subjecting the catalyst coated membrane to the ultrasonic treatment in the second solvent to delaminate the catalyst layers from the ionomer membrane, the ionomer membrane is transferred to a third solvent, optionally water, and subjected to a further ultrasonic treatment to clean the ionomer membrane.

[0069] Optionally, the ultrasonic treatment is a low-power ultrasonic treatment having one or more of the following characteristics:

[0070] a frequency of no more than 1 kHz, 500 Hz, 100 Hz, 60 Hz; no less than 10 Hz, 20 Hz, 30 Hz, 40 Hz, 50 Hz; and / or within a range defined by any combination of the aforementioned upper and lower limits;

[0071] a power of no more than 1 kW, 800 W, 600 W; no less than 300 W, 400 W, or 500 W; and / or within a range defined by any combination of the aforementioned upper and lower limits.

[0072] Optionally, the ultrasonic treatment is a high-power ultrasonic treatment having one or more of the following characteristics:

[0073] a frequency of no less than 2 kHz, 5 kHz, 10 kHz, 15 kHz, or 20 kHz; no more than 100 kHz, 80 kHz, 60 kHz, 40 kHz, 30 kHz, or 20 kHz; and / or within a range defined by any combination of the aforementioned lower and upper limits;

[0074] a power of no less than 100W, 200 W, 400 W, 600 w, 800 W, 1 kW ; no more than 10 kHz, 5 kW, 3 kW, or2 kW; and / orwithin a range defined byany combination of the aforementioned lower and upper limits;

[0075] a power intensity of no less than 20 W cm'2, 40 W cm'2, 60 W cm'2, 80 W cm'2, 100 W cm'2, 150 W cm'2; no more than 500 W cm'2, 400 W cm'2, 300 W cm'2, 200 W cm'2, 150 W cm'2; and / or within a range defined by any combination of the aforementioned lower and upper limits.

[0076] Optionally, the ultrasonic treatment is performed using a sonotrode or an ultrasonic bath. A distance between the sonotrode and the catalyst coated membrane maybe no more than 50 mm, 20 mm, 10 mm, 5 mm, or 3 mm. The sonotrode may be a cylindrical sonotrode or a blade sonotrode.

[0077] The ultrasonic treatment can be applied to a batch of catalyst coated membrane or the catalyst coated membrane can be fed through the ultrasonic treatment for continuous processing.

[0078] Two sets of examples of the present methodology will now be described: (i) examples which us a low-power ultrasonic treatment; and (ii) examples which use a high-intensity ultrasonic treatment. (i) Catalyst Coated Membranes for Fuel Cell and Water Electrolyser Delamination Induced by Organic Solution Swelling and Water Ultrasonication

[0079] This study presents a novel room-temperature, two-step process for separating catalyst-coated membranes (CCMs) used in fuel cells and water electrolysers. The method not only achieves a clean separation of the central membrane from the catalyst materials but also preserves the catalyst, thereby avoiding any potential hazardous gas release. The process involves a brief one-minute soak in an optimised solution, followed by a 10-12 minute low-power ultrasonic treatment in water. The effectiveness of various organic (acetone, ethanol, ethylene glycol, hexane, and toluene) and aqueous (CaCl2, HCl, NaOH, NH4Cl) soaking solutions were thoroughly investigated to identify the optimal conditions for achieving near-pristine, separated membranes. This safe and efficient approach offers a strategy for CCM recycling, promoting resource recovery and economic benefits in clean energy technologies.

[0080] Introduction

[0081] In the endeavour to decarbonise energy production, there is a global shift towards renewable energy technologies. Hydrogen is one alternative energy source and electrolysis is an important tool for the production of hydrogen and oxygen and fuel cells are important to recombine them to make clean electricity. Both of these technologies require significant use of platinum group metals and proton exchange membranes. The most commonly used fluorinated membranes are made from perfluorinated sulfonic acid (PFSA) ionomers such as Nation®, which contain a polytetrafluoroethylene backbone with sulfonic acid terminated sidechains, typically 8 to 80 pm thick. The perfluorinated backbones provide chemical stability whilst the sulfonic acid groups facilitate proton conductivity. However, the sulfonated tetrafluoroethylene-based polymer significantly reduces conductivity beyond 90°C due to dehydration and subsequent loss of water molecules. Aquivion® has a similar structure to Nation but features a shorter side chain. This modification enhances its glass transition temperature, water retention capability, thermal stability and ion conductivity while also reducing hydrogen crossover. Non-fluorinated membranes, including sulfonated poly(arylene ether ketone), sulfonated poly(arylene ether ketone), sulfonated polyethersulfone, polybenzimidazole based polymer, and polystyrenesulfonic acid, also present viable options for proton exchange membrane applications. However, these are less common than the fluorinated polymers. The increased cost and demand for PFSA ionomers coupled with the longevity of the product in the environment makes its recovery at end-of-life (EOL) important.

[0082] In both proton exchange membrane fuel cells and water electrolysers, the proton exchange membrane is ‘sandwiched’ by catalysts. This catalyst-coated membrane (CCM) has different coating compositions, based on the desired function. For fuel cells, platinum nanoparticles (Pt NPs) of ca. 2-5 nm diameter adsorbed onto carbon particles form the catalytic material on both the cathode and anode sides of the membrane. However, the CCM for the water electrolyser generally has Pt NPs on the cathode side, with iridium (Ir) and ruthenium (Ru) oxide particles being the primary choice for the anode side. Pt NPs have good catalytic activity, the ability to endure demanding operational conditions, and high resistance against corrosion. Similarly, Ir and Iridium oxide (IrOx) are preferred to Ru or Pt due to their favourable catalytic characteristics and superior corrosion resistance, particularly towards the oxygen evolution reaction. A carbon support can ensure electrical conductivity within the catalytic layers, although it will be noted that carbon supports are not usually used with iridium based catalysts on the anode side of water electrolysers as they will corrode quickly under oxidising conditions. Current recycling technologies focus on recovering the platinum group metal (PGM) catalysts from proton exchange membrane stacks, as this is where the majority of the value is present. Recovery of the other components such as the polymer and the carbon support are often incidental to PGM recovery, even though the value of the ionomer in the devices can often be as much as that of the PGM. Recovery methods for the PGMs include pyrometallurgical and hydrometallurgical routes. The pyrometallurgical routes involve very high temperature smelting processes to separate and concentrate PGM from the oxides of other metals. Other combustible materials such as PFSA ionomer can also be removed through incineration during pyrometallurgical operations. The PGM concentrates from smelting operations are then processed through hydrometallurgical refining processes which involve dissolution and subsequent separation of the individual metals. Alternatively, PGM can be recovered directly using hydrometallurgical routes where metals or metal oxide NPs are leached from the CCM using caustic or acidic agents, followed by separation of the different components via solvent extraction, distillation, ion exchange, cementation, or filtration.

[0083] The principal advantages associated with hydrometallurgical approaches include their notable selectivity towards metals, relatively modest energy requirements, and the potential for reusing reaction components. However, these can degrade the membrane polymers and form secondary pollutants, posing environmental issues.

[0084] Organic solvents, such as alcohols can be used to recover the membrane polymers as the catalytic coating is detached from the membrane by disrupting the adhesive bond. Carmo et al. separated the cathode and anode catalysts by circulating a deionized water and alcohol mixture, resulting in a nearly pristine fluoropolymer membrane within 10-30 minutes (M. Carmo, G. P. Keeley, D. Holtz, T. Grube, M. Robinius, M. Muller and D. Stolten, International Journal of Hydrogen Energy, 2019, 44, 3450-3455). Similarly, Xu and co-workers demonstrated the isolation of a Nation 115 membrane from proton exchange membrane fuel cells by boiling in isopropanol for 20 minutes, followed by mechanical removal of catalyst traces (H. F. Xu, X. Wang, Z. G. Shao and I. M. Hsing, Journal of Applied Electrochemistry, 2002, 32, 1337-1340). Supercritical media have also been used to separate fluorinated polymers and ionomers from precious metal-containing fuel cell components. This occurs in a pressure reactor at 350-450 °C and 200-400 bar pressure for 1-10 hours, with no reported emissions of fluorine-containing substances (HF, F2, or fluorides). However, the majority of these recycling processes require elevated temperatures and pressures which can be energy intensive. To that end, Johnson Matthey have recently developed the HyRefine™ process which enables recovery of the PGM and the valuable PFSA ionomer together. This process uses chemical routes to improve the efficiency and sustainability of CCM recycling compared to conventional PGM refining.

[0085] This study presents a complimentary technology which may be integrated with the Johnson Matthey HyRefine™ process. The technology can be implemented at room temperature to separate PFSA ionomer membranes from catalyst coatings in fuel cell and water electrolyser CCMs (including end-of-life materials and production scrap). The method involves two stages: firstly, soaking the CCMs in a range of organic solvent to expand the membrane and, secondly, application of ultrasound in water to delaminate the catalyst-coated layers from the membrane. The ultrasound process has low energy consumption, as it is performed at room temperature and the coating is delaminated within a few minutes. Ultrasound with low power levels (<1000 W) has previously been employed in cleansing, blending, and expediting mechanical and chemical processes, whereas high power ultrasound (> 1,000 W) has previously been used for delaminating more challenging substrates such as metal oxide coatings from electric vehicle battery electrodes. The method of action for removal of material from substrates is thought to be due to the collapse of cavitation bubbles, inducing the formation of micro-jets, shockwaves, and micro-streaming. The enhanced mass transport will improve molecular-scale mixing or cleaning, while the cavitation can fracture brittle coatings, or the interfacial connections between layers. Experimental

[0086] Materials

[0087] Scrap (production scrap) fuel cell CCM, water electrolyser CCM, and pristine PFSA ionomer membrane were provided by Johnson Matthey. The CCMs were composed of a polymer membrane coated on both sides with an active catalyst material containing different mixtures of Pt NPs on carbon and either IrTaOx or IrOx particles. The solvents used for swelling the polymers were acetone (FisherScientific, £99.5%), ethanol (VWR, £99.8%), toluene (FisherScientific, Extra Pure, low in sulfur), hexane fraction from petroleum (Fisher Scientific, laboratory-grade), and ethylene glycol (Sigma-Aldrich, 99.8%). Aqueous solutions of ammonium chloride (VWR, 98%), sodium hydroxide (NaOH) (Sigma-Aldrich, >98%), hydrochloric acid (Fisher Scientific, 37%), calcium chloride hexahydrate (Thermo scientific, 98+%), or choline chloride (Thermo scientific, 99%) were also investigated.

[0088] Soaking and delamination methods

[0089] The methodology involves two primary steps: the soaking of fuel cell and water electrolyser catalyst-coated membrane (CCM) samples, approximately 10x10 mm2in size (equivalent to ca.

[0090] 7 mg for fuel cell and 25 mg for water electrolyser CCMs), in various solutions, followed by delamination using an ultrasonic bath in water. For the soaking step, organic solvents such as acetone, ethanol, methanol, and ethylene glycol were tested, as well as aqueous solutions includingdeionised water, 1 M ammonium chloride, 1 M sodium hydroxide, 1 M hydrochloric acid, 1 M calcium chloride hexahydrate, and 1 M choline chloride. The soaking durations ranged from 10 seconds to 6 minutes. It was found that methanol was not preferred for membrane swelling due to its observed flammability when in contact with the catalytic coating. During the delamination process, the soaked catalyst-coated membrane (CCM) samples were subjected to sonication in 10 mL of water. Sonication was performed using an ultrasonic bath (Fisherbrand™ FB15055, 200-240V, 50 / 60 Hz, 550W) to detach the catalyst-coated material from the ionexchange membrane. The sonication times varied between 2 and 12 minutes at ambient temperature (15 to 20°C). This was followed by an additional 2 minutes of sonication in fresh deionised water to remove any contamination. The delamination percentage was evaluated by comparing the mass of the cleaned membrane sample to its initial mass before sonication. To make 100% clean membrane reference samples for comparison purposes, the cleaned membranes were immersed in ethanol for 1 minute, followed by sonication in deionised water for 30 minutes (one sample, 10x10 mm2, per 10 mL of water). A second sonication in fresh deionised water was conducted for an additional 10 minutes to ensure thorough cleaning.

[0091] The swelling behaviour of the PFSA ionomer membrane was investigated using membrane obtained from the uncoated edge of the water electrolyser CCM (ca. 80 pm thick). Samples measuring approximately 20 x 5 mm2were immersed in a range of solutions, both organic and aqueous, to assess their expansion ratios. Measurements were conducted by capturing images using a Inspex HD 1080p instrument at intervals ranging from 2 to 6 minutes. The length changes were then measured and compared to the original length.

[0092] Instrumentation Scanning Electron Microscopy (SEM, Zeiss Gemini 360 FEGSEM) and energy-dispersive X-ray Spectroscopy (EDS, Oxford Instruments Ultim Extreme windowless detector) was used to characterise the morphology and elemental composition of the samples. The SEM was operated in in-lens mode with an accelerating voltage of 1 kV with a 5 nm spot size for imaging. EDS analysis was performed at an accelerating voltage of 5-15 kV, controlled using Aztec software. The particle size and the contribution of catalytic materials were analysed by transmission Electron Microscopy (TEM). The initial samples were soaked in acetone for one minute and physically separated on each side in water. The delaminated particles obtained post-experiment were directly sampled from the delaminated aqueous solution. A JEOL JEM-1400 TEM instrument with an accelerating voltage of 120 kV was utilised to examine the morphology of the samples and ascertain the size of Pt NPs, both sides of CCMs. To facilitate this analysis, samples were prepared by depositing colloidal suspension drops onto copper grids which were air-dried for approximately 1 hour before imaging.

[0093] Fourier transform infrared (FTIR) spectroscopy was used to identify the polymer present in the membrane and binder materials. The pristine and water electrolyser membranes could be directly analysed, butthe fuel cell membrane had to be measured after delamination dueto there being no exposed membrane available. The samples were placed above the light source in a Bruker Alpha II spectrometer instrument, controlled by the corresponding Bruker software on the computer monitor. The scan range was 4000 to 400 cm-1, and the spectrum containing peaks was related to the magnitude of transmittance. The polymers were identified by comparing the measured spectra for the end-of-life membranes to existing literature data, and the identification confirmed by recording standard samples.

[0094] A Mettler Toledo TGA / DSC1 instrument with a resolution of ±1 pgand a maximum temperature of 1100 °C was used to measure sample mass change and heat flow, controlled by STARe software (version 12.10). The balance used toweigh the samples was a Mettler Toledo Semi-Micro Balance (MS105DU), with a resolution of 0.1 mg. The samples were placed in 100 pL aluminium crucibles in all experiments, with no lid. The temperature that operated in the program was from 25 to 600°C with a heating rate of 5 K min-1and a nitrogen flow rate of 75 mL min-1.

[0095] Results and discussion

[0096] Raw materials characterisation

[0097] Scrap (production scrap) fuel cell and water electrolyser CCMs were characterised using SEM / EDS to determine the morphology and chemical composition within the two CCMs. The disassembled fuel cell and water electrolyser CCMs have three distinct layers (central membrane and cover with two layers of catalyst-coated materials). Figure 1a shows a photograph of the fuel cell CCM and Figure 1 b shows a cross-sectional SEM image. The overall thickness is ca. 35 pm and the central polymer membrane layer is 14.57 (±1 .24) pm thick. EDS analysis indicated that the central membrane layer contains carbon (C), oxygen (O), fluorine (F), and sulfur(S), indicating the presence of the sulfur-containing fluorinated polymer. FTIR spectroscopy confirmed this to be the PFSA ionomer, which is consistent with literature. The catalytic coatings were identified to be coated on the anode side with a layer 10.98 (±2.29) wt% Ir, 35.53 (±4.09) wt% Pt, 4.11 (±1.29) wt% Ta and carbon measuring 4.50 (±0.75) pm. The cathode side has a thicker Pt / C layer containing 42.47 (±2.54) wt% Pt, measuring 14.53(±1.37) pm (Figures 1c and 1d). Catalyst materials correspond to Pt-on-carbon and IrTaOx. As illustrated in Figures 1e and 1 f , the TEM analysis reveals that the average sizes of the Pt-on carbon catalyst and the Pt, IrTaOx nanoparticles cell reversal catalyst are 5.33 (±1.12) nm and 5.12 (±1.81) nm, respectively. The presence of fluorine in these catalytic coatings indicates the PFSA ionomer is also used as the binder.

[0098] The water electrolyser is shown in Figure 2. In contrast to the fuel cell CCM, the total thickness of the water electrolyser CCM is considerably larger (ca. 100 pm). It is composed of a polymer membrane measuring80.37 (±0.85) pm, coated on the cathode side with a layercontaining39.06 (±2.37) wt% Pt, measuring 7.90 (±0.70) pm, and on the anode side with an IrOx layer containing 75.73 (±1.35) wt% Ir measuring 12.83 (±0.76) pm.

[0099] The polymer is composed of 27.91 (±2.69) wt% C, 1.93 (±2.72) wt% O, 69.06 (±0.27) wt% F, and 1.12 (±0.29) wt% S, indicative of a PFSA ionomer composition. This was confirmed using FTIR spectroscopy. The average sizes of the Pt NPs on the carbon substrate and IrOx are 5.35 (±1.53) nm and 18.32 (±3.52) nm (Figures 2e and 2f), respectively.

[0100] TGA / DSC and FTIR analysis were conducted to identify the PFSA ionomer (presence of fluorine detected by EDS in all catalyst layers) in isolated catalytic coating materials. TGA analysis revealed distinct weight loss patterns for isolated particles derived from both fuel cell and water electrolyzer catalyst-coated membranes (CCMs) compared to the central membrane itself. Furthermore, the weight loss profile of the IrOx particles (inclusive of PFSA ionomer) is diverged from that of other isolated CCM particles, due to the catalytic thermal decomposition of the ionomer by the IrOx catalyst particles. The typical polymer transitions expected for fluorinated binders and / or reinforcement, such as PVDF, PTFE and Nation, were not observed in the DSC curves due to the PGM which reduce the temperature at which thermal decomposition of these materials is usually observed. Typically, PFSA ionomer is used as the binder and expended PTFE is used as a mechanical support. TGA mass loss of the PFSA from the catalyst layers will differ from that of the membrane because the PGM catalyse combustion.

[0101] Effect of catalyst coated membrane soaking on delamination with an ultrasonic bath

[0102] The effect of soaking the CCMs with various organic solvents on the delamination efficiency was investigated using water in an ultrasonic bath. Firstly, samples of fuel cell CCM were immersed in a selection of organic solvents (acetone, ethanol, and ethylene glycol) and aqueous solutions (CaCl2, HCl, NaOH, NH4Cl) for a fixed duration of 1 minute. After 1 minute of soaking, the CCM sample was transferred to pure water and sonicated for 10 minutes. It was found that CCMs treated with 1 minute in ethanol followed by 30 minutes of insonation, produced pristine PFSA ionomer membranes. An additional 10 minutesof insonation with replaced fresh deionised water was utilized to ensure thorough cleaning, as shown in Figure 3. This was considered effective removal of the catalytic material from the PFSA ionomer membrane and is therefore used as a reference point to compare other pre-treatment and insonation conditions.

[0103] The subsequent changes in mass were quantified and compared with those of the reference samples (Figure 3). The mass variations of the reference samples were 36.18 (±1.67) % for the fuel cell catalyst coated membrane (CCM) and 23.59 (±0.77) % for the water electrolyser CCM, relative to their initial mass prior to delamination. Figure 4a depicts delamination percentages of fuel cell CCM. This study demonstrates that acetone and ethanol are more efficient solvents, ensuringthe successful detachment of the central membrane while preserving a clean surface within 10 minutes. In contrast, water, when used without a prior soaking step in acetone or ethanol, resulted in an unclean membrane surface, as shown in Figure 4c. Note that the rest of the solutions (ethylene glycol, CaCl2, HCl, NaOH, NH4Cl) demonstrate a significantly lower efficacy of delamination, potentially attributed to low solution absorption during solution immersion, resulting in failure to yield a clean membrane surface.

[0104] It is evident that the soaking pre-treatment of the CCMs in an organic solvent is key to an effective separation of catalytic materials from the ionomer membrane. However, interestingly, direct insonation in the same organic solvent resulted in lower efficiency delamination compared to the solvent followed by water steps. Note that, within minutes of soaking in the various solvents, the thickness of ionomer membrane was found to increase by ca. 4-8% (for acetone, ethanol and ethylene glycol). It is likely that the swelling of soaked ionomer membrane, the changes in the rigidity of the solvent-modified film, and the ultrasound cavitation effectiveness in water all contribute to the highly efficient, room-temperature delamination of CCMs.

[0105] This process was applied to the delamination of water electrolyser CCM, yielding comparable results to the delamination of fuel cell CCM; however, the delamination step needed about 12 minutes (Figure 4b) possibly due to a different ionomer type, catalyst loading and increased membrane thickness. The central membrane surface following sonication is presented in Figure 4d. Furthermore, a one-minute ethanol soak achieves better results than acetone regarding membrane surface delamination within the operational timeframe.

[0106] Conversely, delamination using pure ethanol, pure acetone, a 50-50 mixture of ethanol-water or acetone-water, combined with ultrasonication, is ineffective at achieving a clean membrane surface. Investigations into the cavitation behaviour of different aqueous media under ultrasonic conditions show that the cavitation shockwaves at low kHz frequencies are more vigorous in water than in ethanol, which are highly likely to be due to the lower density, surface tension, and vapour pressure of the alcohol solutions resulting in unstable cavitation bubbles.

[0107] The change in volume of the membrane during delamination with ethanol probably traps nanoparticles, complicating their separation. Additionally, the high alcohol content in contact with highly reactive metals during the delamination process can pose safety issues.

[0108] To optimise the soaking time in acetone and ethanol, a range of times from 10 seconds to 6 minutes were evaluated to minimise processing time whilst still maintaining a high delamination efficiency. The investigation revealed no statistically significant difference in delamination outcomes across the tested soaking times for acetone. However, after 4 minutes of soaking in ethanol, a lower delamination efficiency was observed due to variations in the swelling and shrinking behaviour of the polymer membrane, especially for the thicker membrane of water electrolyser CCM in ethanol for 6 minutes. This suggests a rapid interaction between the solvents and the membrane material. As a result, a soaking time of 1 minute in ethanol was implemented, ensuring adequate sample hydration as a standard pre-treatment protocol for subsequent experiments. Additionally, the influence of solvents on membrane swelling was investigated. PFSA Ionomer membranes were submerged in various organic solvents (acetone, ethanol, ethylene glycol, hexane, and toluene) and aqueous solutions (CaCl2, HCl, NaOH, and NH4Cl). However, these observed variations in swelling did not demonstrably influence the delamination process.

[0109] Membranes and particles after ultrasonic delamination.

[0110] The delaminated membranes exhibited deformation due to softening during soaking and delamination, leading to bending and increased foldability on a macro scale after drying. Despite this, they retained their sheet-like form, similar to the initial samples, and no fractures were observed in this study, as evidenced by the images from Figures 4c and 4d. The delaminated membranes were examined using SEM to assess their surface characteristics. Figure 5 presents SEM images of the membrane surfaces after a one-minute ethanol soak and subsequent delamination in an ultrasonic water bath. Figure 5a shows the delaminated membrane surface on the Pt / C cathode side of the fuel cell sample, which appears very clean. The membrane surface on the Pt / C, IrTaOx anode side of the fuel cell sample (Figure 5b) is also immaculate at a 10 pm scale and contains only traces of embedded particles on a 200 nm scale. Figures 5c and 5d show the delaminated membranes on the Pt / C and IrOx sides of the water electrolyserCCM. Both sides appear very clean, as confirmed by the absence of catalyst-coated particles detectable by SEM. This observation substantiates the high efficiency of the delamination method employed in this study. Note that the holes observed in the figures result from the polymer melting under vacuum conditions duringSEM analysis.

[0111] Following delamination of the CCMs, the PFAS ionomer membrane was removed from solution and the remaining aqueous solution was dried in the oven at 60 °C for 24 hours. The dried particles were examined usingTEM, as shown in Figure 6. Notably, no fracturing of the carbon substrate or detachment of metal or metal compounds from the carbon surface was observed. This suggests that the low-power ultrasonic bath lacks the energy required to dislodge the nanoparticles from the carbon substrate and PFSA ionomer binder. This study was designed to delaminate both sides of the catalyst materials. Additional process steps can then be implemented to separate the catalyst layers into different PGM streams and recover the ionomer from the membrane.

[0112] Conclusions

[0113] This study demonstrates an effective bulk separation of the catalytic materials from PFSA ionomer membranes. Simply pre-soaking the scrap catalyst-coated membranes (CCM) with ethanol for 1 minute, enables the active catalyst particles to be separated from the ionomer membrane following 12 minutes of insonation in water. The pre-soaking of CCM with ethanol is key to the rapid removal of active catalytic particles. This is likely due to a change in the ionomer structure of the particle-loaded ionomer film as evidenced by a non-insignificant degree of swelling of the membrane.

[0114] This technology is a non-destructive method allowing a rapid separation of ionomer films from active catalytic particles in scrap fuel cell and water electrolyser CCMs. It is envisaged that this approach can be readily scaled up and adopted to first recover ionomer films prior to catalytic upcycle treatments which would otherwise destroy or deform ionomer membranes.

[0115] (ii) Fast Delamination of Fuel Cell Catalyst-Coated Membranes using High-Intensity Ultrasonication

[0116] This study demonstrates a rapid and facile method for separating the central membrane and catalyst-coated material from production scrap fuel cell catalyst-coated membranes (CCMs), facilitating a circular economy of technologically critical metals. A novel approach is presented using high-intensity ultrasonication with two distinct sonotrode configurations for rapid delamination at ambient temperature in water. This technique utilizes cavitation, where high-frequency sound waves create, expand, and collapse microbubbles, generating high-speed jets, shockwaves, and acoustic streaming. This process effectively separates the membrane and catalyst while maintaining their overall integrity of the former. A cylindrical sonotrode (20 mm diameter) was used to optimize process parameters for smaller CCM samples to minimize time and energy consumption. To scale up the delamination process for industrial-size CCMs, a blade sonotrode (15 mmx210 mm) was employed to enable a flow process for rapid and continuous delamination (see Figure 7). Cavitation at the sonotrode-CCM interface was shown to facilitate the selective and rapid breakdown of the catalyst layers, enabling full delamination of the catalyst-loaded membrane within tens of seconds. This efficient and fast delamination approach offers a promising strategy for CCM recycling.

[0117] Introduction

[0118] Proton exchange membrane fuel cells (PEMFCs) are a prominent focus in renewable energy conversion technologies. They offer a method of greener, more cost-effective, and portable energy solutions. PEMFCs use hydrogen as fuel to generate electricity, with water as the sole byproduct. In PEMFCs, a central membrane (8-18 pm) is sandwiched between layers of cathode and anode catalyst material, forming a composite structure known as a catalyst-coated membrane (CCM). The chemical composition of these catalyst layers can be tailored to achieve specific functionalities within the fuel cell. Platinum supported on carbon (Pt / C) is the predominant electrocatalyst for PEMFCs owing to its fast hydrogen oxidation kinetics. To facilitate rapid proton transport to the catalyst active sites, perfluorosulfonic acid (PFSA) polymers are commonly employed as the ionomeric phase. A significant challenge lies in the recovery of these precious catalytic materials as fuel cells at end-of-life (typically 10-15 years) or from production scrap material. Therefore, the ability to recover and reuse precious metal catalysts (such as platinum and iridium) as well as expensive PFSA-based ionomer membranes becomes critical. This underscores the need for sustainable recovery and reuse of materials for the long-term sustainability of PEMFC technology.

[0119] Current recycling technologies for PEMFCs prioritize the recovery of platinum group metal (PGM) catalysts due to their higher market value. However, the ionomer, another essential PEMFC component, is often disregarded despite its significant economic value. Recognizing this discrepancy, Johnson Matthey developed their innovative HyRefine™ technology to achieve corecovery of valuable platinum group metals (PGMs) and ionomers for reuse in new catalyst-coated membranes (CCMs). This chemical approach enhances CCM recycling efficiency and sustainability compared to conventional PGM refining. Separation of the ionomer membrane is a critical step in the recycling process from CCMs, allowing for the independent recovery of both the membrane and the metal components. A common method utilizes alcohol-water mixtures, such as isopropyl alcohol, 1 -butanol, or 2-butanol, heated to an elevated temperature of 100-200 °C. These conditions are believed to disrupt the bonds between the fluorocarbon-based ionomer and the catalyst particles, causing the membrane to disperse in the solvent solution, facilitating separation. Xu et al. report a method for Nation membrane recycling via immersion in boiling isopropanol (20 min) for swelling, followed by mechanical removal of catalyst residues (H. F. Xu, X. Wang, Z. G. Shao and I. M. Hsing, Journal of Applied Electrochemistry, 2002, 32, 1337-1340). The pretreated membranes were then boiled in 3-5% H2O2solution (1 h) for complete decolorization. Alternatively, supercritical water (350-450°C, 200-400 bar, 1-10 h) offers selective separation of fluorine-containing components (ionomers, etc.) from precious metals in spent fuel cells.

[0120] Ultrasonication, using high-frequency sound waves (20-100 kHz)19, has emerged as a versatile tool across variousscientificdisciplines. It has been used in numerous applications including the dispersion of nanoemulsions and nanocomposite materials, separate water from crude oil, generating Hg(OH)2from Hg°(l), as well as the crucial task of recycling and extracting valuable metals. Applications for this technology span a wide range, including pharmaceutical and food sciences, biodiesel production, water treatment, cleaning, and removal of fine particles. Notably, ultrasonic soundwaves introduce forced convection through three primary mechanisms: cavitation collapse, microjet formation, and acoustic streaming. High-power ultrasound also enables rapid mass transport within a system, removing passivating surface layers during electrochemical stripping of metals to expose fresh reaction sites, and even facilitating the delamination of multi-layered materials. Recent advancements include the development of an ultrasonic blade designed explicitly for the delamination of lithium-ion battery anode and cathode sheets. The current research explores the potential of using a high-power ultrasound technique as a fast, simple, and ambient temperature process for the efficient separation of CCMs. By strategically positioning the CCM directly under the sonotrode, we aim to achieve delamination in under ten seconds. This rapid process facilitates the separation of the catalyst layers from the central membrane, enabling the recovery of both components. The concentrated ultrasonic cavitation and acoustic pressure generated at the sonotrode interface are expected to be sufficient to break the bond between the catalyst layers and the membrane. We further optimise this method for large scale production scrap membrane processing by changing the sonotrode configuration from a small cylinder to a novel blade configuration to facilitate continuous processing. This efficient and rapid delamination approach offers a promising strategy for CCM recyclingwith significant scale-up potential.

[0121] Experimental section

[0122] Materials

[0123] Production scrap fuel cell CCMs were provided by Johnson Matthey. These CCMs comprised a polymer membrane with a PFSA ionomer membrane coated on each side with a catalyst layer. The catalyst layers were composed of combinations of platinum on carbon (Pt / C) or a combination of Pt / C and iridium-based metal oxide (lrMOx) nanoparticles. Ethanol (£99.8%, VWR) was used as the solvent for soaking the fuel cell CCMs before delamination.

[0124] Delamination using the ultrasonic systems

[0125] A commercial ultrasonic delamination system (Branson Sonics, 1 ,25DCXa20-V, Danbury, CT, USA) was employed to separate the CCM of the fuel cell. The CCM sample was first cut into a 30 mm x 30 mm square (45.3±1.7 mg). Subsequently, the sample was submerged in ethanol for a predetermined duration, ranging from 5 to 60 seconds. After ethanol treatment, the CCM sample was placed onto the smooth surface of a stainless-steel rod and then positioned within a beaker filled with deionized water (200 ml). The delamination process utilized a vertical configuration with a cylinder sonotrode (horn tip, diameter of 20 mm) placed at a fixed distance directly above the CCM sample as illustrated in Figure 8a. The ultrasonic system operated at a frequency of 20 kHz and offered adjustable power ranging from 125 W to 1250 W, translating to a variable power intensity of c.a. 40 W cm'2to 400 W cm'2.

[0126] The work herein optimized parameters for the ultrasonic delamination process. The parameters explored included ultrasonic power, sonication time, and the perpendicular distance of the sonotrode-to-sample distance. The ultrasonic power was varied at four levels: 10% (ca. 40 Wcm'2), 20% (ca. 80 Wcm-2), 30% (ca. 120 Wcm-2), and 40% (ca. 160 Wcm-2) of the system's maximum capacity. Each power level was combined with three sonication times: 1 second, 2 seconds, and 5 seconds. Additionally, the effect of sonotrode-to-sample distance was evaluated at two settings: 2.5 mm and 5 mm.

[0127] The optimal parameters identified during ultrasonic delamination with a cylindrical horn were subsequently employed for a scaled-up delamination process. This involved a rapid and continuous delamination of a large fuel cell membrane using a high-power ultrasonic unit. This custom-designed unit incorporated a blade sonotrode with a rectangular cross-section measuring 15 mm x 210 mm, as shown in Figure 8b. The blade sonotrode could deliver a maximum power intensity of 70 W cm'2. A sonication bath was constructed using a stainless-steel tray positioned within a tank containing approximately 6 liters of deionized water. The sonotrode was placed 5 mm above the tray. A large, industrial-scale CCM sample measuring 80 mm x 170 mm was fed through the blade sonotrode. Prior to sonication, samples were immersed in ethanol for 30 seconds to undergo a pre-treatment process. Plastic sheets were affixed to both ends of the CCM sample to maintain sample integrity and to facilitate a continuous feeding du ring the sonication process. The prepared sample was passed through the sonication bath at a constant velocity of about 1 cm s'1with the blade sonotrode operating at 40 W cm'2of power intensity. Catalyst-coated materials were dispersed within the aqueous medium (DI water). Following complete delamination of both sample surfaces, the delaminated membrane and suspended catalyst materials in water were retrieved and characterised using SEM and EDS. To investigate the impact of fuel cell delamination without the support of a stainless-steel rod, a 30x50 mm2sample was secured using a custom-designed clamping apparatus. Ultrasonic delamination was induced under optimal conditions using a cylindrical sonotrode, with a power intensity of 40 W / cm2and a sonotrode-to-sample distance of 5 mm. Additionally, to establish a baseline for delamination efficiency, CCM sample sections (30 mm x 30 mm) were initially immersed in ethanol for 30 seconds, followed by ultrasonic treatment in deionized (DI) water for 30 minutes using a Fisherbrand® FB15055 ultrasonic bath (200-240V, 50 / 60 Hz, 550W). The DI water was then replaced, and sonication was continued for an additional 10 minutes. These fully delaminated samples served as a reference point against which the delamination efficiency of the experimental conditions was compared. Control samples of identical dimensions were prepared for this purpose.

[0128] Other instrumentation

[0129] High-speed imaging (Fastcam SA-Z 2100 K, Photron, Bucks UK) was employed to capture the cavitation activity and streaming phenomena associated with fuel cell delamination. Images were acquired at a frame rate of 20,000 frames per second using a macro lens (Milvus 100 mm f / 2 M, Zeiss, Oberkocken Germany). Synchronous 10 ns collimated laser pulses (632 nm, red) (CAVI LUX Smart, Cavitar, Finland) were utilized for illumination, enabling shadow graphic imaging and providing precise temporal resolution.

[0130] Scanning electron microscopy (SEM) was performed using a Zeiss Gemini 360 FEGSEM to characterize the morphology and elemental composition of the samples, as well as any particles remaining on the membrane substrate after delamination. The SEM was operated in in-lens mode with an accelerating voltage of 1 kV and a spot size of 5 nm for imaging. Energy-dispersive X-ray spectroscopy (EDS) analysis was performed using an Oxford Instruments Ultim Extreme windowless detector (Oxford Instruments, Abingdon, UK) at an accelerating voltage of 5-15 kV, controlled using Aztec software.

[0131] Results and Discussion

[0132] The effectiveness and scalability of ultrasound facilitated delamination of production scrap fuel cell CCM is studied herein. Firstly, the delamination parameters were optimised using a commercially available cylindrical sonotrode with a horn diameter of 20 mm. Subsequently, the optimised parameters obtained from the first part of this study is transferred to a custom-built blade sonotrode (15 mm x 210 mm) equipped with rollers to facilitate a continuous delamination of CCM materials to demonstrate scalability.

[0133] The production scrap fuel cell CCM was used for delamination studies. The CCM comprised three distinct layers with a total thickness of approximately 35 pm: a Pt / C cathode layer (14.53 pm ± 1.37 pm), a central PFSA ionomer membrane layer (14.57 pm ± 1.24 pm), and Pt / C, lrMOxanode layer (4.50 pm ± 0.75 pm). Square-shaped samples (30 mm x 30 mm) were prepared from the CCM to facilitate optimisation of the delamination process. A high-intensity ultrasonic delamination technique employing a cylindrical sonotrode submerged in a water bath was utilized at ambient temperature to assess the influence of various parameters on delamination efficiency. Before sonication, the samples underwent a pre-treatment stage involving immersion in ethanol for 30 seconds. As detailed in the previous section, this pre-treatment soaking step was implemented to modify, swell and soften the rigidity of the particle-loaded Nation film, significantly reducing the insonation time required to fully separate central membranes from catalyst active materials.

[0134] The parameters, including ultrasonic horn power, the spacing between the sonotrode and the sample, and the sonication time, were optimised to enhance operational efficiency and minimize energy consumption. Experiments were conducted with a sonotrode with a diameter of 20 mm at a fixed sonotrode-to-sample distance. Ultrasonic horn power output was varied from 10% to 40% of the maximum power setting in 10% increments, and sonication time of 1 second, 2 seconds, and 5 seconds were evaluated. As expected, the delamination area increased when sonication power increased. Interestingly, delamination of the CCM with the anode side (a thin layer composed of Pt / C and lrMOxparticles) facing the sonotrode results in a more transparent central membrane as compared to the case if the cathode side (a thicker layer composed of Pt / C particles) was facing the sonotrode. In all cases, under excessively high power and long sonication times (e.g., 40% power and 5 seconds at a 5 mm distance), irreversible damage (ripping) to the middle membrane layer was seen, and this potentially hinders the reusability of the recovered membrane. Similarly, reducing the sonotrode-to-sample distance to 2.5 mm also caused similar damage. This observation aligns with the established principle that sound wave pressure and intensity increases at distances closer to the sonotrode surface. Notably, a setting of 10% power and 1 second of insonication of both sides of the membrane at a 5 mm distance away from the sonotrode was sufficient for effective delamination.

[0135] Figure 9 shows SEM images of the CCM sample before and after two-sided delamination using the above optimised parameters. In this case, both sides of the CCM were subject to 1 second of insonication in deionised water at 10% power output (ca. 40 W.cnr2) and a 5 mm sonotrode-to-sample distance. As can be seen in Figure 9b, both layers of catalyst active materials were successfully removed revealing the quasi-transparent PFSA membrane. SEM-EDS analysis was used to characterise the material before and after sonication. Figure 9a shows the morphologies of the Pt / C cathode, and Pt / C, lrMOxanode layers before delamination. The elemental identity of ‘M’ is commercially sensitive and its identity is therefore not revealed as it is unimportant for this work. The delaminated surfaces of membrane surfaces after the insonication treatment are presented in Figure 9b. Post-delamination EDS analysis of the recovered membrane revealed only carbon, oxygen, fluorine, and sulfur, with no detectable traces of catalyst materials.

[0136] High-speed imaging was employed to investigate the effects of cavitation of bubbles during the sonication of a fuel cell sample. Many studies have utilized this technique to directly observe cavitation activity, using high frame rates to capture the intricate dynamics of bubble formation and collapse near surfaces so as to provide new physical insights to the delamination processes. Figure 10 presents the delamination behaviour of pristine, untreated fuel cell samples with the Pt / C (Figure 10a) and, separately, Pt / C, lrMOxside (Figure 10b) facing the sonotrode. The sonotrode was operated at a power intensity of 40 W cm-2at a fixed distance of 5 mm from the CCM substrate suspended in water. Sonication initiation at t = Os generates stochastic bubble clouds at the surface of the cylindrical ultrasonic horn tip. After ca. 5 ms, smaller, variable-sized bubble clusters near the sample surface are observed. During sonication, the violent collapse of bubble clouds and cavitation implosions within the cavitation region between the sonotrode and the sample resulting in the generation of shockwaves and micro-jets to enhance mass transport to and from the surface of the sample at a short distance of 5 mm from the sonohorn. Concomitantly, acoustic streaming, characterised by fluid circulation in the vicinity of bubbles, forces fluid toward the sample surface (observed at 20 ms). After 50 ms, cavitation led to delamination of catalyst active materials from the CCM substrate. Interestingly, the delamination behaviour of the Pt / C cathode side and Pt / C, lrMOxanode side facing the sonotrode exhibited significant differences as revealed by the high-speed camera. While the delamination of the Pt / C cathode side resulted in small but sizeable flakes (evidenced at 0.9 s), the delamination of the Pt / C, lrMOxanode side emanate a ‘cloud’ of finer particles (at 0.1 s). Moreover, the delamination of the Pt / C, lrMOxanode side appeared to be more facile and resulted in a larger delamination area. These disparities in detachment mechanisms may be attributed to compositional differences in materials and layer thickness differences between the cathode and anode catalyst layers. Full length delamination videos can be found online as part of the Supplementary Information.

[0137] Figure 11 a (i) shows the optical image of the delaminated CCM with the Pt / C cathode facing the sonotrode and shown in Figure 11 b(i) is a separate experiment with the anode side (Pt / C, lrMOx) facing the sonotrode. To enhance the visibility, Figure 11 a(ii) and Figure 11 b(ii) is a replica of Figure 11 a(i) and Figure 11 a(ii), respectively, but the brightness and contrast have been enhanced to better display the extent of delamination on the membrane surface. While successful delamination occurred on the Pt / C cathode side, the opposite side of the membrane exhibited significant residual material (not shown). Regardless of which side of the CCM is facing the sonotrode, there is a consistent radial delamination pattern revealing the part of membrane that is closest to the sonotrode. This is consistent with the above physical insights obtained from highspeed imaging, where self-organized conical cavitation fields directly under a sonotrode operating at low frequencies and high power are evidenced. While a clean area is observed on the Pt / C, lrMOxanode side after delamination, the opposite side of the membrane (not shown) still retains coated material. Note that the weakly coated materials may detach from the central membrane during the sample drying. Furthermore, the interfacial regions between delaminated and intact catalyst layers, SEM images shown in Figures 1 la(iii) and 11 b(iii), exhibit distinct morphological characteristics. On the Pt / C cathode side, the catalyst coating appears to detach in cohesive failure, with particles which had not been successfully delaminated remaining on the surface in a flake-like structure. In contrast, the Pt / C-lrMOxanode interface reveals a more adhesive failure mode, where small, micron-sized agglomerates of individual particles are observed to remain on the surface. These findings are consistent with the high-speed imaging which revealed flakes of material emanating from Pt / C surface and clouds of fine-particles emanating from the Pt / C-lrMOxsurface.

[0138] Additionally, the findings suggest the possibility of achieving one-sided delamination (without flipping the sample for further sonication) when sonicating commences from the Pt / C, lrMOxanode layer facing the sonotrode. This effect is particularly evident at higher power levels and extended sonication times, such as 20-30% of power for 5 s, resulting in cleaned areas on both sides. This is supported by the observation of cleaned areas on both sides of the membrane after delamination due to the propagation of ultrasonic waves through the membrane, resulting in delamination of both sides. However, initiating sonication from the Pt / C side resulted in an unclean membrane surface, potentially due to the thicker and different material properties of the Pt / C layer compared to the Pt / C, lrMOxanode layer, which could hinder effective ultrasonic wave transmission and delamination. Consequently, due to the practical challenge of consistently identifying the anode layer from the cathode, which both appears black for the naked eye, a two-sided delamination approach is adopted to ensure clean membrane surfaces on both sides of the CCM sheet.

[0139] Ethanol pre-soaking is critical to achieve successful delamination. As anticipated, sonication without this pre-treatment resulted in incomplete separation, as evidenced by the substantial presence of undelaminated material in the sample. This observation reinforces the hypothesis that the pre-soaking stage is essential for modifying the material properties of the CCM, facilitating effective delamination and ultimately leading to the acquisition of fully separated membranes. Furthermore, while the investigation explored the influence of pre-soaking time on delamination efficiency, the findings revealed no statistically significant variations in results between Wand 60 seconds. This suggests that the soaking time within this range may be a minor factor for delamination success. However, to maintain consistency throughout the study, a constant soaking time of 30 seconds will be adopted for further experiments.

[0140] Building upon the optimal delamination conditions identified with the cylinder sonotrode (30 seconds ethanol pre-soaking, 40 W cm'2power intensity, and 5 mm sonotrode-to-sample distance), these parameters were then applied to the custom-built blade sonotrode. A fuel cell CCM sample of 80 mm x 170 mm dimensions was fed under the blade sonotrode at a speed of approximately 1 cm s’1. This allows continuous delamination of large CCM which is otherwise not possible using the smaller cylindrical sonotrode. The resulting effects of ultrasonic delamination on both the anode and cathode sides of the CCM can be visualized in the SEM images of the membrane after the insonation treatment in Figure 12. A comparative analysis of the initial and post-delamination stages of the scrap fuel cell CCM (Figure 12a) reveals a successful removal of the catalyst-coated layers from the central membrane polymer through ultrasonic sonication. Subsequent SEM-EDS analysis of the delaminated membrane surfaces detected negligible catalyst residues (Figure 12b), mirroring the findings obtained with the cylinder sonotrode. The overall efficiency of PFSA ionomer membrane delamination in this process is quantified as 99.28 (±0.92)%, identified by comparison with fully delaminated samples that served as a reference point. Note that the ultrasonic delamination process experienced a temperature increase from ca. 25 °C to 35 °C within 5 minutes of insonation. This thermal escalation suggests that implementing a cooling system may be useful to ensure the process's stability and reliability during continuous operation. While the blade sonication method employed in this study demonstrated efficacy in delaminating intact sheet or roll-format samples, however, there might be other challenges associated with shredded materials with irregular or fragmented geometries. A potential alternative, continuous delamination within a high-intensity ultrasonic bath or tank, could be explored for processing irregular or fragmented materials, such as production waste, disassembled components, or materials with diverse geometries and conditions. The current investigation serves as a proof-of-concept, validating the potential of high-intensity sonication for complete sheet separation in proton exchange membrane recycling and recovery.

[0141] The delaminated PFSA ionomer membranes present an opportunity for closed-loop recycling of both the catalyst active materials and the PFSA membrane. Alternatively, other potential recycling pathways include membrane film reformation or catalyst ink production. The former involves dissolving the ionomer in an organic solvent system (e.g., ethanol-water, DMSO, DMF) under elevated temperature and pressure conditions, followed by casting. Alternatively, the recovered ionomer can be a constituent of the catalyst layer, incorporated into catalyst ink through dispersion with catalyst particles. Subsequent separation of catalyst particles from water enables their recovery and further purification according to established methodologies. Conclusion

[0142] This study demonstrated the potential of high-intensity ultrasonic delamination for separating catalyst layers from the PFSA ionomer membrane in fuel cell CCMs. The investigation identified vital parameters influencing the delamination process, including ultrasonic power, sonication time, and sonotrode-to-sample distance. The optimized protocol achieved clean delamination with minimal damage through a fast process. Firstly, a 30-second ethanol pre-treatment effectively modified the rigidity of the catalyst-loaded membrane, significantly impacting the subsequent delamination step. Secondly, sonication was performed using a conservative 10% ultrasonic horn power (40 W cm'2) for a brief duration (1 second per side) at a controlled sonotrode-sample distance (5 mm) at ambient temperature in water.

[0143] The efficacy of these optimised parameters was further validated using a blade sonotrode, demonstrating its capability of continuous delamination of large pieces of production scrap CCMs at a rate of 80 mm2s-1with minimal ionomer membrane damage. As the first generation of fuel-cells now approaching their end-of-life, the insights presented herein provides a green, rapid, highly scalable, low-cost close-loop separation, and recycling of valuable components like the PFSA ionomer membrane and catalyst materials.

[0144] Further Example

[0145] In this example a sonic bath rather than a sonic probe was used for the ultrasonic treatment and a water / organic solvent mixture was used in the sonic bath. A catalyst coated membrane was immersed in ethanol (10 seconds), then subsequently immersed in 50 / 50 1-propanol / water (at approximately 25°C) and subjected to a sonication process in an ultrasonic bath (10 minutes). Complete delamination of the catalyst layers was achieved.

[0146] While this invention has been particularly shown and described with reference to certain examples, it will be understood to those skilled in the art that various changes in form and detail may be made without departing from the scope of the invention as defined by the appended claims.

Claims

Claims1. A method for recycling a catalyst coated membrane comprising an ionomer membrane coated on either side by catalyst layers, the method comprising:soaking the catalyst coated membrane in a first solvent;transferring the catalyst coated membrane from the first solvent into a second solvent which is different to the first solvent; andsubjecting the catalyst coated membrane in the second solvent to an ultrasonic treatment to delaminate the catalyst layers from the ionomer membrane.

2. A method according to claim 1,wherein the first solvent comprises or consists of an organic solvent.

3. A method according to claim 1 or 2,wherein the first solvent comprises or consists of an alcohol.

4. A method accordingto any preceding claim,wherein the first solvent comprises or consists of ethanol.

5. A method accordingto any preceding claim,wherein the catalyst coated membrane is soaked in the first solvent for a time period of: at least 5 seconds, 10 seconds, 20 seconds, 30 seconds, 40 seconds, 50 seconds, 60 seconds, 2 minutes, 5 minutes, 10 minutes; no more than 5 hours, 60 minutes, 30 minutes, 20 minutes, 10 minutes, 8 minutes, 6 minutes, 4 minutes, 2 minutes, 1 minutes; and / or within a range defined by any combination of the aforementioned lowerand upper limits.

6. A method accordingto any preceding claim,wherein the catalyst coated membrane is soaked in the first solvent at a temperature of: at least 0°C, 5°C, 10°C, 15°C or 20°C; no more than 80°C, 60°C, 50°C, 40°C, or 30°C; and / or within a range defined by any combination of the aforementioned lower and upper limits.

7. A method accordingto any preceding claim,wherein the catalyst coated membrane is soaked in the first solvent at room temperature and pressure.

8. A method accordingto any preceding claim,wherein the second solvent comprises or consists of an aqueous solvent.

9. A method accordingto any preceding claim,wherein the second solvent comprises or consists of water or a mixture of water and an organic solvent, optionally a mixture of water and an alcohol, optionally a mixture of water and propanol.

10. A method accordingto any preceding claim,wherein the ultrasonic treatment is for a time period of: at least 0.1 second, 1 seconds, 2 seconds, 5 seconds, 10 seconds, 20 seconds, 40 seconds, 1 minute, 5 minutes, or 10 minutes; no more than 60 minutes, 30 minutes, 20 minutes, 15 minutes, 10 minutes, 5 minutes, 1 minute, 30 second, or 10 seconds; and / or within a range defined by any combination of the aforementioned lower and upper limits.

11. A method accordingto any preceding claim,wherein the ultrasonic treatment is at a temperature of: at least 0°C, 5°C, 10°C, 15°C or 20°C; no more than 60°C, 50°C, 40°C, or30°C; and / or within a range defined by any combination of the aforementioned lower and upper limits.

12. A method accordingto any preceding claim,wherein the ultrasonic treatment is at room temperature and pressure.

13. A method accordingto any preceding claim,wherein after subjecting the catalyst coated membrane to the ultrasonic treatment in the second solvent to delaminate the catalyst layers from the ionomer membrane, the ionomer membrane is transferred to a third solvent, optionally water, and subjected to a further ultrasonic treatment to clean the ionomer membrane.

14. A method accordingto any preceding claim,wherein the ultrasonic treatment is a low-power ultrasonic treatment having one or more of the following characteristics:a frequency of no more than 1 kHz, 500 Hz, 100 Hz, 60 Hz; no less than 10 Hz, 20 Hz, 30 Hz, 40 Hz, 50 Hz; and / or within a range defined by any combination of the aforementioned upper and lower limits;a power of no more than 1 kW, 800 W, 600 W; no less than 300 W, 400 W, or 500 W; and / or within a range defined by any combination of the aforementioned upper and lower limits.

15. A method accordingto any preceding claim,wherein the ultrasonic treatment is a high-power ultrasonic treatment having one or more of the following characteristics:a frequency of no less than 2 kHz, 5 kHz, 10 kHz, 15 kHz, or 20 kHz; no more than 100 kHz, 80 kHz, 60 kHz, 40 kHz, 30 kHz, or 20 kHz; and / or within a range defined by any combination of the aforementioned lower and upper limits;a power of no less than 100W, 200 W, 400 W, 600 w, 800 W, 1 kW ; no more than 10 kHz, 5 kW, 3 kW, or 2 kW; and / or within a range defined by any combination of the aforementioned lower and upper limits;a power intensity of no less than 20 W cm'2, 40 W cm'2, 60 W cm'2, 80 W cm'2, 100 W cm'2, 150 W cm'2; no more than 500 W cm'2, 400 W cm'2, 300 W cm'2, 200 W cm'2, 150 W cm'2; and / or within a range defined by any combination of the aforementioned lower and upper limits.

16. A method accordingto any preceding claim,wherein the ultrasonic treatment is performed using a sonotrode or an ultrasonic bath.

17. A method accordingto claim 16,wherein a distance between the sonotrode and the catalyst coated membrane is no more than 50 mm, 20 mm, 10 mm, 5 mm, or3 mm.

18. A method accordingto claim 16 or 17,wherein the sonotrode is a cylindrical sonotrode or a blade sonotrode.

19. Amethod accordingto any preceding claim,wherein the ultrasonic treatment is applied to a batch of catalyst coated membrane or wherein the catalyst coated membrane is fed through the ultrasonic treatment for continuous processing.

Images