A Review of Accelerated Stress Tests for Enhancing MEA Durability in PEM Water Electrolysis Cells

During the past decades, a signi ﬁ cant amount of excellent scienti ﬁ c results has been generated in the ﬁ eld of polymer electrolyte membrane water electrolysis (PEMWE). Compared to current state-of-the-art technologies, PEMWE o ﬀ ers the opportunity to produce green hydrogen with zero carbon emissions. However, the membrane electrode assembly (MEA), whose price is still high for a rather limited lifetime, needs further improvement in terms of performance, cost, and durability. In order to e ﬃ ciently process novel materials, accelerated stress tests (ASTs) can be implemented to provoke and investigate cell ageing processes and assess failure modes under real-life conditions. In this review, the di ﬀ erent accelerated stressors of the main components of the MEA are discussed, and recent publications of ASTs in the study of PEMWE cell durability are summarized. Furthermore, a concise review of the degradation mechanisms for the individual MEA components depicted in recent publications is presented. The di ﬀ erent aspects identi ﬁ ed in this review serve as a roadmap to further advance the durability of novel stack materials.


Introduction
In the efforts to create a renewable and sustainable energy supply chain, hydrogen plays a key role. It can be produced from various feedstocks, stored, and efficiently transformed back into energy through fuel cells or be used in the chemical industry. However, the current hydrogen production (i.e., approximately 70 Mt/a) is limited compared to, for example, the annual global energy demand in the transport sector alone, which amounts to roughly 1012 Mt/a hydrogen equivalent (neglecting efficiency losses) in 2020. Furthermore, the vast majority of hydrogen is produced from natural gas and coal [1]. Combined with the increasing need to efficiently store energy from renewable sources (e.g., wind, solar, water, and tidal), electrolysis is capable of contributing a valuable share to filling this gap. Among the available electrolysis technologies, polymer electrolyte membrane water electrolysis (PEMWE) is -up to date-the most capable of coping with the relatively dynamic energy supply pattern renewable sources exhibit [2]. PEMWEs have a quoted life-time of 60.000 to 100.000 hours [3] and can achieve current densities of >3 A cm -2 , with commercial offerings at 1-3 A cm -2 [3][4][5]. Yet this relatively new and emerging technology faces several challenges. Improved understanding of the performance, cost, and durability trade-offs under predicted future dynamic operating modes using CO 2 -free electricity is necessary to meet the hydrogen shot clean hydrogen cost target of $1/kg H 2 by 2030 (and an interim target of $2/kg H 2 by 2025) [6]. Furthermore, new materials are slow to make their way into actual products, as qualification and validation are time-consuming processes with little harmonization in testing procedures [5]. The examination and understanding of performance losses will become even more critical with a shift to low-cost applications and intermittent power sources [7,8].
Another well-known key issue of PEMWE is the scalability from microscopic phenomena in laboratory cells to stack or system level [9]. Different measurement set-ups and cell geometries influence the measurement results and their reproducibility. A variety of MEA materials and preparation methods exist, that additionally influence their comparability. Therefore, it is important to introduce harmonized testing procedures to rate the comparability of different cell sizes assembled with similar materials.
In PEM fuel cells (PEMFC), the reverse process of PEMWE, accelerated stress tests (ASTs) have experienced much attention, as they are capable of accelerating the development and increasing the understanding of lifetimelimiting mechanisms [7]. The principle of an AST is to apply different degradation accelerating stressors, either via in situ testing or ex situ measurements, to project the durability of the whole cell or a desired component. In order to design ASTs for PEMWE that are capable of portraying real-life relevant ageing, a fundamental understanding of dominating degradation mechanisms as well as their dependency on stressors during operation is critical. Degradation mechanisms in PEMWEs are primarily dominated by irreversible efficiency losses. Reversible degradation has been reported as well, although it is understood to a far lesser extent [8]. Problems related to the higher voltage in PEMWEs are performance degradation and material deterioration and bring a challenge for their durability. Prolonged use under certain operation conditions can lower the electrochemical stability and accelerate the degradation of the materials. The most critical components in the PEMWE cell are represented by the membrane electrode assembly (MEA). A typical MEA is composed of a polymer electrolyte membrane (PEM), two catalyst layers (CLs), a carbon-based gas diffusion layer (GDL) on the cathode side, and a metal-based porous transport layer (PTL) on the anode side of the cell. Figure 1 shows a schematic of the components in a PEMWE cell and the half-cell reactions at the anodic and cathodic compartment. At the anode, the oxygen evolution reaction (OER) takes place where water is electrochemically split into oxygen (O 2 ), electrons (e -), and protons (H + ). These protons diffuse through a proton-conducting membrane to the cathode side, where they are recombined with electrons to evolve as gaseous hydrogen (H 2 ). On the cathode, platinum (Pt) is predominantly used as catalyst, whereas on the anode-side iridium (Ir) is typically used for the OER due to its relatively high activity and stability [10][11][12][13].
PEMs, such as those used in electrolysis cells, are typically made from perfluorosulfonic acid (PFSA) polymers such as Nafion™ with a thickness of about 100 μm. Efforts are being made to develop cheaper and more durable membranes with higher proton conductivity. However, as of now, developments in one direction typically limit the performance in other directions as compared to commercial membranes [14]. The PTLs act as current collectors and are located between the current collecting flow field and the reactive sites in the CL. The GDL on the cathode side of the catalyst coated membrane (CCM) consists of a standard carbon cloth. On the anode side of the CCM, metallic materials like Ti-based meshes are used as PTLs due to the high potential of the OER.
Extensive reviews of the durability and degradation mechanisms of PEMWE cells and its components exist. For instance, Feng et al. [8] presented a comprehensive review on the degradation mechanisms of key components in PEMWE and summarized the alleviating strategies. Khatib et al. [15] recently published a summary of the degradation of components in PEMWE and investigated the effects of degradation on the overall performance of the electrolyzer. Shirvanian and Berkel [16] introduced a survey of PEMWE status and reviewed state-of-the-art components and limitations in performance and durability. In other recent publications, the focus was set on specific components in the PEMWE, such as electrocatalysts [17,18], membranes [19][20][21], PTLs [22][23][24], and BPPs [25]. Furthermore, EU-harmonized protocols for testing of several lowtemperature water electrolysis systems are summarized in Joint Research Centre (JRC) reports by the Fuel Cells and Hydrogen Joint Undertaking (FCH JU) [26][27][28][29]. A summary of some articles related to the field of PEMWE degradation is given in Table 1.
These works provide a general overview of potential degradation mechanisms and testing procedures in PEMWE; however, to the authors' knowledge, no comprehensive reviews on AST strategies for PEMWE exist so far. For PEMFCs, a technology facing similar challenges, various AST protocols at component stack and system level have already been reviewed [30][31][32]. In this review, aspects of degradation and durability of the key components of the PEMWE are brought in context with their individual accelerating conditions, aiming to gather the information necessary for the qualification and evaluation of state-of-the-art and novel materials for PEMWE. The accelerating factors responsible for PEMWE performance decay are summarized for the membrane, catalysts, and diffusion media (GDL and PTL), respectively. Finally, AST protocols for lifetime prediction based on the functional properties of the materials and operational conditions are discussed for the individual components.

Membrane Degradation and Accelerated Stressors
Due to the simultaneous presence of hydrogen on one side and pure oxygen on the other side, gas separation is critical for the safe operation of an electrolyzer. The state-of-the-art materials used to separate the gaseous reaction products (oxygen and hydrogen) are PFSA-based polymer membranes (Nafion™). Nafion™ comprises of a PTFE backbone with sulfonic acid end-groups that provide protonic conductivity ( Figure 2) [16]. Historically, Nafion™ membranes are identified with a three-digit number in which the first two digits refer to the equivalent weight and the third digit refers to the membrane thickness in 25.4 μm-units (e.g., Nafion™ 115: equivalent weight = 1100 and thickness = 127 μm).
Applying thick electrolytic membranes to PEMWE cells improve the physical and chemical durability and reduce safety risks in the system [33]. However, the driving needs to reduce the overall transport resistance results in increasingly thinner PFSA membranes which tend to accelerate gas crossover in PEMWE cells [20]. Regarding long-term performance, the PFSA membrane is the weakest component in PEMWE systems, as performance deterioration is mostly attributed to membrane 2 International Journal of Energy Research pollution or chemical degradation [8]. Apart from Nafion™, a few other membranes can be used in hydrogen production processes [34]. For instance, polybenzimidazoles (PBI), sulfonated polyether ether ketones (SPEEK), polyoxadiazole, polysulfone (PSF), or polyimides have recently been considered as cheaper and more effective alternatives in membrane-based electrolytic systems [34].

Membrane Degradation Mechanisms.
In this section, the supposed degradation mechanisms refer to PFSA membranes, unless otherwise stated. Apart from safety and operating issues, one issue that affects the stability of the membrane and thus accelerates chemical degradation is the crossover of gases. The crossover of product gases (hydrogen and oxygen) can result in the formation of hydrogen peroxide or radicals at the catalysts [16,35]. An overview of the reactions involving free radicals in PEM membranes as presented by Gubler et al. [36] is given in Table 2.
The reactions 1-8 that are listed in Table 2 show that the hydroxyl radical is the most aggressive intermediate as its reaction with carboxylic end-groups in the main chain of the PFSA ionomer can lead to a gradual "unzipping" of the PFSA main chain (reaction 1) [37]. Primary sources of HO • , which leads to the chain scission, are the slow homolysis of H 2 O 2 (reaction 2) and the Fenton reaction (reaction 8    Figure 2: Chemical structure of Nafion™, x * 6.5.  (reaction 4). Hydrogen radicals are produced in the reaction of H 2 with HO • (reaction 5) that are further converted into hydroperoxyl radicals (HOO • ) in reaction 6. In reaction 8, highly reactive hydroxyl radicals (HO • ) are formed through the produced H 2 O 2 . This reaction is also known as the so-called Fenton's reaction and strongly catalyzed by Fe 2+ -ions [38]. If no ferrous iron-ion impurities are present, HOO • mainly decays via disproportionation (reaction 7) [36]. The radical attack of the perfluorinated backbone and perfluorosulfonic side chains of the PEM can cause degradation resulting in fluoride and sulphur release and thinning of the membrane [8]. The proposed reaction is as follows: In reaction 9, the carboxylic acid end-groups of the PEM membrane react with two hydroxyl radicals. As a result, two HF molecules are released and one CO 2 and one CF 2 molecule are formed [39]. Therefore, in durability tests, the fluoride-ion release rate (FRR) of the PEM electrolyzer exhaust can be seen as indicator for the rate of chemical degradation of the membrane.
Another degradation mechanism that poses a critical safety issue is the diffusion of evolved gases (H 2 and O 2 ) in water through the membrane to the opposite compartment of the electrolyzer [40]. This phenomenon is driven by differences in the dissolved gas concentrations in the liquid water phases across the membrane [19,40,41].
The membrane thickness strongly correlates with the gas crossover and therefore affects the lifetime of the electrolyzer. Although very thin membranes offer a reduced transport resistance, they cannot guarantee safe operation due to higher diffusion of H 2 and O 2 through the PFSA membrane. Therefore, relative thick membranes (120-200 μm) like Nafion™ 115 and Nafion™ 117 are used in PEMWE to avoid the crossover of gases [42]. An overview of the performance of different PFSA membranes in respect to their thickness from comparable PEMFC studies [43,44] is presented in Table 3.
The major reason for the larger performance decay for thin membranes can be explained by higher permeation effects across the MEA. However, thicker membranes exhibit a higher thickness loss, as they have more material to degrade [43]. Chemical degradation processes like membrane thinning or pinhole formation can increase the overall permeation and subsequently reduce the efficiency of the PEMWE [20]. Furthermore, enhanced operation conditions like high temperature [45][46][47], current density [40,46], and operation pressure [48][49][50] lead to an increase of the gascrossover rate. High-pressure operation results in significant stress on the membrane, and the differential pressure in particular favors permeation. Therefore, the high-operation pressure that is targeted in PEMWE requires the use of thicker materials, because high pressure driven by the pressure gradient increases the probability of H 2 or O 2 gas diffusion across the membrane [14,25,42].
Additional stress may originate from preexisting defects and foreign objects introduced during manufacturing or from components such as the PTLs and current collectors. Punctures, tears, and cracks are known to function as starting points for further degradation, which inevitably result in a reduced lifetime. In the presence of pure and usually pressurized oxygen, the thereby posed risk is, however, a multitude higher than in, e.g., fuel cells, where typically ambient air is fed. Even though the combustion of hydrogen and oxygen is in the presence of catalyst favored over explosion in the case of crossover between the electrodes, severe destruction of the materials can occur rapidly [51,52]. Therefore, a sufficient water supply to the anode is necessary to prevent hot spots and local drying of the membrane [51].

Hydration Conditions and Elevated Temperature.
Compared to PEMFC, where the membrane has to be humidified by gases, the PEMWE membrane is in contact with liquid water during the whole electrolyzer operation [14,19]. Several researchers found, that drying at elevated temperatures (>100°C) has a negative impact on the protonic conductivity and the water-uptake properties of PEM membranes [53,54]. The water-uptake properties are affected by the irreversible reorganization of the membrane structure, and any decrease in water-uptake results in the degradation of the proton conductivity [19]. During electrolyzer operation, the water not only functions as reactant but also is used for cooling purposes. Consequently, water flowrate and temperature are important parameters that should not be neglected. These findings agree with a study by Selamet et al. [55], where the effect of water flowrate on the performance of a 10-cell PEM electrolyzer stack (100 cm −2 active area MEAs) is investigated. The PEMWE stack is operated in a range of flowrates from 750 to1500 mL/min at 1.35 A cm −2 for 2000 hours. According to the authors, the flowrate has no major influence on the  [43,44]. The thicknesses are listed at begin of testing (BoT) and end of testing (EoT), and the degradation rate is given in mV h -1 . performance of the PEM electrolyzer. However, if the flowrate of water is too low, the cell temperature rises uncontrollably, which affects the efficiency of the PEMWE. Therefore, the flow rate is generally chosen for cooling purposes at an operating current density [55]. The researchers also tested the influence of other parameters during PEMWE operation and found that the operating temperature is the most important factor affecting the performance. At high operating temperatures, the activation barrier is lowered, which results in enhanced membrane properties like faster reaction kinetics, improved mass transport (MT), and higher conductivity [55]. Temperatures above 90°C accelerate the rapid degradation of the membrane, and the increase of chemical degradation has reported to be directly proportional to the temperature increase [56]. According to Kusoglu and Weber [20], the thermal decomposition of the PFSA ionomer occurs in three stages: (1) Loss of (residual) water (from 100  (2) Cleavage of the C-S bond (280 ± 30°C to 400 ± 20°C) resulting in sulfonate-group degradation ending with oxidation of the side-chain terminus, and (3) The final decomposition (oxidation) of the perfluorinated matrix (400 ± 20°C to 600 ± 40°C) [20] Therefore, PEMWE operation at higher temperatures of above 90°C can limit the lifetime of the membrane significantly. According to Frensch et al. [57], a temperature increase from 60 to 80°C results in a moderate increase of the degradation rate from 1.2 to 3.0 μV h −1 , while operating at 90°C has a significantly more detrimental effect of 183.8 μV h −1 . Furthermore, operation at elevated temperatures significantly increases the FRR and membrane thinning over time. Chandesris et al. [35] developed a model that investigates the oxygen crossover and temperature-induced membrane degradation in PEMWE. The authors revealed that the lifetime of an electrolyzer cell is significantly decreased at high temperatures. Only 8.700 h operation at 353 K were necessary to thin 50% of the membrane [35].

OCV.
It has been reported that idle or open-circuit voltage (OCV) conditions cause largely uniform degradation across the MEA active area in PEMFC and accordingly lead to near uniform thinning of the membrane [30]. Singh et al. [58] showed that in situ hold at OCV causes chemical degradation of PEMFC membranes. The results were verified via infrared (IR) and scanning electron microscopy (SEM) images that show that the pure chemical membrane degradation proceeds generally uniform across the whole active region of MEA. Therefore, in situ hold at OCV is the accepted means to accelerate membrane chemical degradation of PEM cells, while the change of OCV is a useful index to evaluate the membrane health, because the OCV decreases with the increase of gas permeability caused by membrane thinning, crack, and pinhole development [58,59].
Cationic contamination has been reported as an accelerating factor for membrane degradation. While improperly treated, feed water is the major cause of PEMWE failures in this field [60,61]. Weiß et al. [62] demonstrated the ion exchange of metal impurities into the membrane/ionomer phase during an OCV-AST. The corrosion processes at low potentials resulted in the formation of metal cations which not only reduced the membrane/ionomer conductivity but also increased the high-frequency resistance (HFR) through the displacement of protons in the ionomeric membrane [62].

Load Cycling and Constant Current Operation.
Another lifetime-limiting factor for membrane ageing in PEMWE is the operating current. Chandesris et al. [35] highlighted the impact of current density on the chemical degradation rate in PEMWE membranes. They investigated chemical membrane degradation in a 25 cm 2 single cell setup at various temperatures and current densities. According to the researchers, at low-current densities, two competing HO • radical-consuming reactions take place: a first reaction that requires hydrogen peroxide, and the second reaction that accelerates the membrane attack. As the peroxide concentration is quite high at low current densities, the first reaction becomes predominant, and the degradation rate is reduced. High-current density operation results in a decrease of the degradation rate, since the molar percentage of oxygen at the cathode is reduced, and peroxide and free radicals are formed [35].
These trends in low-current density operation have also been confirmed by more recent research conducted by Li et al. [63]. The researchers designed AST protocols comprising of dynamic load profiles of 0-0.5 A cm −2 and 1.2-2 A cm −2 , respectively. The experiments revealed that low-current cycling had a far greater influence on the voltage degradation, whereas the ohmic resistance decreased significantly during the high-load cycles, which could imply membrane thinning and failure.
In a recent study, Fouda-Onana et al. [64] measured the FRR as indicator for the membrane state-of-health. The researchers performed in situ measurements of PEMWE single cells that consisted of an alternating current profile between 0 and 1 A cm -2 . A set of electrochemical measurements was performed at the end of each current sequence to measure the performance deterioration of the cell. The researchers additionally investigated the effect of temperature on the MEA. Therefore, the same ageing test (load cycling between 0 and 1 A cm -2 ) was carried out on a MEA at 60°C and at 80°C. The FRR shows an extremum between 0.2 and 0.4 A cm -2 . The extremum can be explained by the mentioned competitive reaction between radical (HO • ) formations and the two reactions that consume those radicals. In contrast to ageing at 60°C, ageing at 80°C has a more severe impact on the membrane ageing (see Table 4). The FRR shows a 5-fold increase at 80°C compared to that one measured at 60°C (operation at 1 A cm -2 ). These findings indicate that the temperature has 5 International Journal of Energy Research much more severe effect on the membrane degradation than the operating current [64].
Lettenmeier et al. [65] performed a study on the degradation mechanisms of a commercially available 8-cell 120 cm 2 stack electrolyzer. The researchers designed a protocol of measurements comprising of three setpoints (T1-T3). After an initial activation for 300 h, at nominal condition setpoint with 2 A cm −2 was applied for more than 400 h (T1). Thereafter, the electrolyzer operated dynamically for 50 h between 2 A cm −2 and 0.15 A cm −2 (T2). At setpoint (T3), the electrolyzer was operated at 4 A cm −2 for 250 h. Operation at high current densities resulted in a reduction of the ohmic drops, while longer operation gradually deactivated the anode. After the test sequences, the researchers performed ex situ measurements of the MEAs (SEM and AFM) and water resin (XPS). The results indicate that the loss of ionomer and catalyst material in the anode is dependent on the set current [65]. Interestingly, no lowering of the performance of the PEM electrolyzer was observed over time, despite the degradation of the MEA components that was reported [65].

Fenton's Test.
Fenton's reagent is commonly used to study the ex situ chemical degradation of membranes in PEMFC research [30]. Although PEMWE is the reverse technology of PEMFC, the degradation mechanisms of the membrane are the same, since the state-of-the-art membrane material used in both technologies is Nafion™. In general, Fenton's test is carried out in a solution of hydrogen peroxide (H 2 O 2 ) with ferrous iron (Fe 2+ ions) as a catalyst: It is commonly agreed that trace Fe 2+ or other contaminants accelerate the membrane attack in the presence of H 2 O 2 [38]. In practice, these metallic impurities come from the balance of plant [8]. The interaction of Fe 2+ ions with the PEMWE system is not well documented, i.e., the ions may cross the membrane, stay within it, or be flushed out at one of the outlets. The radicals formed through the Fenton's reaction attack the membrane and produce further radicals, such as R-CF 2 in the case of Nafion™ [38]. A systematic investigation of the degradation of Nafion™ membranes using two variations of Fenton's accelerated ageing experiments was conducted by Kundu et al. [66]. A solution method where Fe 2+ ions and peroxide exist together in solution prior to the addition of Nafion™, and an exchange method where Nafion™ in the Fe 2+ form is exposed to hydrogen peroxide. Figure 3 shows a comparison of the sur-face morphologies of Nafion™ membranes after exposure to Fenton's reagent by two different test methods (solution and exchange method).
A modelling framework for PEM degradation in electrolysis mode was simulated by Frensch et al. [38]. The researchers developed a Fenton model in which all major involved electrochemical reactions were visualized. Additionally, experiments were performed that revealed a high dependence of the FRR on the interaction of hydrogen peroxide and Fe impurities in the system. The iron impurities were found to catalyze the reaction, while hydrogen peroxide is necessary for the formation of destructive hydroxyl radicals [38]. Further studies showed that Fenton-like metal catalysts can replace Fe 2+ ions to produce OH • radicals [67,68]. Laconti et al. [69] proposed an order for the rate of chemical degradation caused by common contaminants: has the highest impact on the degradation rate and Ni 2+ the lowest. In state-of-the-art PEMWE cells, titanium is the material of choice for anodic components such as BPPs and PTL [38].

Summary Membrane
Degradation. The Nafion™ membrane represents a core component of the MEA. Membrane failure is mainly accelerated by undesirable operation conditions such as high temperature, OCV, and mechanical stress and results in changes of thickness (membrane thinning) and the formation of defects (e.g., cracks or perforation). The deterioration of these parameters can be measured as a result of the FRR, voltage drop, or changes in the HFR in membrane specific failure analysis [59]. An overview of membrane failure in lifetime tests is given in Table 5.
Other AST strategies than the experiments presented in Table 5 exist focus on the mechanical degradation of the membrane. For instance, Borgardt et al. [70] tested the influence of the clamping pressure on the performance of different PEMWE cell designs. To minimize the negative impact of membrane degradation on the cell performance, novel approaches like the reinforcement or modification of membranes exist. Although the reduction of membrane thickness in PEMWE is an important route to achieve proper efficiency at high-current density, thinner membranes can bring the risk of an increase of gas crossover and therefore compromise the mechanical stability of the MEA. Recently, Siracusano et al. [71] demonstrated a reinforced short-sidechain PTFE-based membrane that enables stable electrolysis operation at 4 A cm −2 . Still, at the present stage, novel materials to replace PFSA membranes need to undergo durability studies for the evaluation of their long-term performance.

Catalyst Degradation and Accelerated Stressors
The harsh operation conditions, especially at the anodic side of the PEMWE, limit the choice of electrocatalysts to a small range of precious metal compounds. For monometallic catalysts during the OER, a link between the activity and stability has been established in which the most active oxides (Au ≪ Pt < Ir < Ru ≪ Os) show the lowest stability  [61]. Therefore, the most popular materials that are currently used for the OER in commercial anodes are Ir-based and Ru-based electrocatalysts [10][11][12]72]. For the hydrogen-evolution reaction (HER) that occurs on the cathode side of the PEMWE, carbonsupported platinum (Pt/C) is recognized as the state-ofthe-art material [16]. The performance of the catalysts is largely credited to the morphology, particle size, crystal structure, and substrate materials [5]. Various material combinations have been tested, especially for the OER to reduce the amount of catalyst loading and enhance the stability of the materials. In this review, the focus is set on catalysts that comprise of Pt-based metals as cathode and IrO x as anode catalysts.

Catalyst Degradation Mechanisms.
Catalyst degradation occurs on both anode and cathode catalysts, whereas the fast kinetics of the HER on widely used Pt catalysts in PEMWE cathodes allow loading reduction down to 0.025 mg cm −2 and have a minor impact on the performance [73]. Therefore, the main challenge remains the stability of the anode catalyst. According to Spöri et al. [74], the degradation of the CLs during the harsh PEMWE operating conditions includes catalyst dissolution, catalyst migration and agglomeration, CL detachment, and support passivation [74]. Catalyst degradation can also be related to the instability of the PFSA membrane and PTLs. During long-term electrolyzer operation, the instability of the PFSA-membrane highly affects the anode and cathode CLs. Grigoriev    Ex situ degradation Fenton's reagent The fluoride emission is highly dependent on the iron concentration, which catalyzes the reaction [38] Hydration conditions Feed water flowrate between 750 and 1500 sccm The water flowrate indirectly influences the performance because it is also used for cooling purposes in the system [55] Thermal stress Operation at 1 A cm −2 for 80.700 h at 353 K and for 380.500 h at 333 K Membrane chemical degradation (fluoride emission and membrane thinning) is strongly affected by the operating temperature [35] OCV-AST Cycling between OCV and operating potentials (up to 3 A cm −2 ) Cationic contamination of the membrane is a possible reason for the increase of the high-frequency resistance (HFR) [62] Load cycling Current between 0 and 2 A cm −2 The temperature was found to have a more critical effect on the membrane degradation compared to the current density in PEMWE operation [64] Load cycling Dynamic load of 0-0.5 A cm −2 and 1. The current density has a minor effect on the degradation rate. It follows a complex mechanism, and a maximum degradation rate is observed at quite low-current densities [35] 7 International Journal of Energy Research reported membrane thinning during PEMWE long-term operation as a result of interactions with oxidizing species formed at the electrodes. The stability issue of the PTLs has also been studied. According to Rakousky et al. [75], the Ti in the PTL corrodes and causes contamination of the anode CL, leading to a decrease in the anodic exchange current density. In addition, a low-conductive Ti-oxide layer is formed between the anode and the PTL that increases the ohmic resistance of the cell [75,76]. Other mechanisms for catalyst degradation are related to mechanical damage (i.e., uneven contact pressure or improper membrane swelling). For instance, bubble formation on the catalyst surface is reported to induce particle loss or layer detachment and disruption of the catalyst-membrane interface [8].
3.1.1. Catalyst Agglomeration and Migration. Nanoparticles (NPs) tend to agglomerate into bigger particles with respect to time due to their high-specific surface energy [8,15]. As a result, the active catalyst area available for the electrochemical reaction is reduced, resulting in a decrease of the catalytic activity [77]. The interaction between catalyst nanoparticles (NPs) is a common phenomenon in PEMFC, and there are several mechanisms that are believed to drive the Pt surface area loss with time: (i) coalescence via crystal migration, (ii) growth via modified Ostwald ripening, and (iii) reprecipitation at other nucleation sites [74,78]. These particle growth mechanisms may also be applied to PEMWE, because changes in the morphology of the CLs like migration, agglomeration, coalescence, or detachment of catalyst NPs have been reported at the cathode and anode during the long-term operation of PEMWE cells [8,15,79]. For instance, IrO x agglomeration is accelerated under harsh operation conditions and contributes to the drop of mass activity toward the OER [79]. Paciok et al. [80] showed that the migration of Pt particles is correlated to the applied overpotential at the cathode. In their study, XPS analysis revealed that with a more negative overpotential, the hydrogen coverage rate of the Pt particles increases, which results in a bigger gap between the Pt particles and the carbon support. Thus, the weakened Van der Waals forces allow an increase of the mobility of Pt particles, which further results in their agglomeration [80].
Yu et al. [81] confirmed that catalysts undergo degradation through agglomeration during the long-term operation of PEMWE. Pt was profiled via EDX mapping, and the results show the particle agglomeration and shape transformation from spherical to dendritic. Additionally, Claudel et al. [79] revealed that simulated PEMWE anode operating conditions result in a loss of OER activity due to the growth of catalyst particles and oxidation of species along with the decrease in the number of individual NPs concomitantly [79].

Dissolution of IrO x .
Regardless of its high activity, dissolution of IrO x during the OER has been reported [82][83][84]. The dissolution rate enhances with an increase of potential, and stable IrO 2 starts to dilute at approximately 1.8 V [8]. Cherevko et al. [85] assume that dissolution processes at high potentials are a consequence of the formation of one or several unstable Ir(III) complexes which rapidly dissolve in an electrochemical process. The tentative mechanism, as reported by the authors, is represented in the following chemical reactions: The suggested mechanism for the OER is shown in Figure 4. The whole pathway for the second reaction can be rewritten as following: Degradation of IrO x through the evolution of oxygen from the lattice strongly depends on the structure and morphology of the electrocatalyst [86]. While crystalline, rutile-type iridium oxides are metallic conductors with a high degree of stoichiometry; their amorphous counterparts are permeable to water and more vulnerable to corrosion in the OER potential region [86,87]. Kasian et al. [83] found that lattice oxygen evolution may occur in both materials; however, amorphous hydrous IrO x exhibits a higher number of defects, and the porosity and the presence of -Ir(III)OOH-groups in the CL increase the probability of the formation of molecular oxygen from the lattice and the dissolution of catalyst [83].

Start-Up/Shut-Down Cycling and OCV.
Factors such as operational currents and the frequency of changes in load, temperature, and pressure make real-world operation of PEMWEs challenging [88]. Many stressors are involved, and coupling effects make it difficult to distinguish between their impacts on material degradation [89]. According to recent studies, AST profiles for the investigation of PEMWE catalyst durability frequently mimic operation from renewable energy, like wind, solar, or water sources. In this case, the fluctuating power supply results in PEMWE operation between periods of hydrogen generation, when power is available, and idle periods, where no electricity is supplied (OCV) [73]. For instance, Weiß et al. [62] evaluated the catalyst stability for such realistic operating conditions. The proposed test protocol is shown in Figure 5. According to the researchers, an AST that cycles between OCV and operating potential has a much severe effect on the degradation rate compared to experiments where no OCV periods are applied.
The study also revealed that IrO x can easily be reduced to metallic Ir when IrO x -based MEAs are held at OCV. Crossover H 2 from the cathode side between ∼0 V and ∼1.6 V during the OCV-AST reduces the surface of the IrO x catalyst at the anode to a state closer to less stable hydrous IrO x [62]. Considering that hydrous IrO x is much more likely to corrode in the OER potential region compared to dry, crystalline IrO x , these results represent a critical degradation mechanism in the study of PEMWE [86].
These findings are supported by a recent study conducted by Rheinländer et al. [86]. The researchers have tested events similar to operation interrupts in a PEMWE cell at 80°C. The AST profile consisted of three steps: in the first step, the cathode pressure was set to 5.5 bara and a current of 1 A cm −2 was applied. In the second step, the operation was interrupted by switching off the power supply, waterflow, and cell heating on the anode side. In the third step, the cell was held at OCV for 2 h in a transient period. After these steps, the cell was restarted [86]. Dodwell et al. [88] reported open-circuit Pt dissolution during intermittent operation of a PEMWE for the first time. The researchers found that the underlying mechanism of Pt dissolution at the cathode is similar to cathode degradation in PEMFCs, where the passivation of catalyst surfaces gradually hinders the Pt dissolution process. During a 90 h OCV, hold test 152 ng cm −2 or 0.005% of the total catalyst loading was lost [88] in their study.
The effect of CL degradation at low catalyst loading and dynamic operation was observed by Alia et al. [90]. The researchers analyzed various test variables including the stressor pattern/severity, the cycling frequency, and potential ramping [90].
The dynamic operation leads to thinning of the Ir CLs which significantly contributed to the overall durability loss [90]. Life tests of an electrolysis cell that were carried out by Grigoriev et al. [60] showed that current reversals can occur during shut-down procedures in PEMWE. In these shut-down periods, operation in FC mode is possible by the consumption of the remaining electrolysis gases (H 2 and O 2 ). Consequently, Pt may be carried away from the cathodic electrocatalytic layer into the membrane. The decreased Pt concentration and the loss of cohesion of Pt NPs result in a lowered activity of the cathodic CL. Besides Pt, several other contaminants (Ti, Ir, Fe, Ni, and Si) from feed water or, e.g., the corrosion of pipes that have lower mobility compared to hydrogen is deposited in the membrane's near-cathode region, leading to an increase of the electrolysis potential over the course of durability tests [60].

Load Cycling and Constant Current Operation.
The effect of constant current operation and load cycling on catalyst degradation in PEMWE was studied by Rakousky et al. [91]. The researchers investigated a PEMWE cell at 80°C and found that high-static current densities (up to 3 A cm −2 ) resulted in the largest performance losses. Siracusano et al. [92] investigated PEMWE catalyst dissolution at high-current densities. Galvanostatic operation at 1 A cm −2 leads to a quite low-degradation rate of 5 μV h −1 , whereas operation at 3 A cm −2 caused a three-fold higher degradation rate of 15 μV h −1 [92]. In particular, the anodic dissolution of IrO x has been reported at high current densities [14,17,30]. A case study involving anode dissolution has been performed by Grigoriev et al. [93]. In order to maintain a higher cell voltage and anode potential in their degradation study, Pt was used as the electrocatalyst at both anode and cathode CL. The researchers found that a corrosion process in the CL is the first MEA degradation mechanism followed by thinning of the Nafion™ membrane. After 100 h of activation the current density was set to 1 A cm −2 ; afterwards, on/off galvanostatic cycles (between 0 and 1 A cm −2 ) for approximately 5500 h of operation were applied. Additionally, SEM, TEM, and EDX analyses were performed to detect the observed performance decay. TEM pictures representing cross-sections of the anodic area after the experiments are shown in Figure 6. The results indicate that the dissolution and precipitation of anode catalyst particles have a major impact on to the overall degradation of the cell performance [93].  [94]. The researchers used a reference electrode to study the catalyst degradation in an operating PEMWE. The results revealed that the cathode contributes more to changes in OCV than the anode during shut-down cycling. The researchers also demonstrated that potential cycling has a more severe effect on the catalyst durability than current cycling. The Pt surface area barely decreased during the current cycling procedure, whereas a significant loss of the ECSA (electrochemically active surface area) was observed during potential cycling. Therefore, potential cycling can be considered as a useful AST for the evaluation of catalyst stability in Pt-based cathode CLs in PEMWE [94].
In a recent study, Spöri et al. [95] presented a transient ADT (accelerated degradation test) protocol for RDE tests that consist of potential square-wave cycles between 0.05 and 2 V. Additionally, the researchers ran static CA (chronoamperometric) tests at the same potentials for comparison reasons. A much higher electrode potential dependence of catalyst dissolution was observed during the transient ADT protocols compared to the static CA protocols [95]. In order to verify the proposed ADTs for their applicability in PEMWE systems, the researchers also ran PEMWE single cell tests. Therefore, CCMs were subjected to (i) 24 h potentiostatic operation at 1.75 V; (CCM-CA), (ii) 24 h power cycling between 0 A cm −2 and a potential of 1.75 V (CCM-PC), and (iii) 15.000 cycles between 0 and 1.75 V (CCM-ADT) [95]. Figure 7 represents the ADT protocols (i)-(iii) and the polarization curve during (iii) [95]. According to Spöri et al. [95], transient fast potential cycling and static experiments have a lower impact on the degradation in single cell PEMWE measurements compared to the RDE measurements in their study. This might be due to the higher temperature of 80°C in CCM vs. ∼25°C in RDE experiments and higher catalyst loading in CCMs. Still, the researchers observed a higher degree of mass-activity losses in transient compared to static CCM measurements. Additionally, the results imply that static constant current or constant potential stability measurements (CCM-CA) are insufficient to test the catalyst stability of state-ofthe-art PEMWE applications [95].

Cationic
Contamination. The accumulation of metallic cations like Fe 3+ [96], Cu 2+ and Al 3+ [97], Ca 2+ [98], and Na + [99] in the MEA can range from being a contributing factor to being the dominating mechanism in PEM electrolysis performance degradation. The foreign cationic impurities may originate from impurities in stack component materials, impurities of feed water, and other sources [96,100]. According to Cheng et al. [100], even trace amounts of impurities present in either fuel or air streams of PEMFCs can have a severely poisoning effect on MEA components, including the anode and cathode CL [100]. This effect is anticipated to be identical, if not more severe in electrolyzer operation if impurities enter the stack with the feed water. The distribution of dissolved contaminants is promoted by the over stoichiometric feed rate (λ > 4) of liquid water [101] in PEMWE. Cationic contamination from Fe 3+ , Cu 2+ , and Al 3+ has been reported as a result from degradation of materials and corrosion of pipes in the system [96,97]. Li et al. [102] investigated the effect of long-term

11
International Journal of Energy Research contamination with iron ions on the performance of a PEMWE single cell. The researchers observed a significant performance decay in which the degradation rate reached to 128.9 μV h −1 after the 829 h contamination test [102]. In a more recent publication, Li et al. [97] investigated the effect of further cationic impurities (Fe 3+ , Cu 2+ , and Al 3+ ions) in single PEMWE cells and found that all impurity ions have a severe impact on the cell performance. Cell poisoning and associated performance breakdown can already occur at low concentrations of 3 ppm Al 3+ on the cathode side [97]. Cathode CL thinning due to the accumulation of Ca 2+ ions in PEMFC cells has been investigated by Banas et al. [103]. Results showed that the accumulation of Ca 2+ ions can lead to proton depletion in the cathode CL [97,103]. As a result, carbon corrosion can occur, which then accelerates cathode CL thinning (Figure 8)

12
International Journal of Energy Research As shown in Figure 8, foreign cations accumulate in the cathode side due to the balance of migration and diffusion. Lack of proton access accelerates the carbon corrosion [97]. On the cathode CL, the same state-of-the-art materials are used in PEMWE and PEMFC cells. Therefore, the effect of cationic impurities on the performance of the PEMWE cell is expected to be even more detrimental due to the higher potentials [97]. The contamination effect of Na + on PEMWE has been studied by Zhang et al. [104]. Different poisoning modes (anode poisoning and cathode poisoning) were compared, and the results imply that Na + poisoning has a more severe effect on the anodic than on the cathodic cell components. In the experiments, the anode potential increased by about 0.160 V because protons were replaced by Na + ions at the anode and membrane with increasing cell voltage. At the cathode, the potential decreased with decreasing proton concentration [104]. Generally, reactions (5) and (6) occur at anode and cathode in PEMWE systems. However, if there are not enough protons for reaction (6), the reaction (7) takes place at the cathode. As a result, the cathode potential declines significantly due to the lack of protons. Additionally, Na + ions and OH − can enter the feed water, which increases the pH at the cathode compartment [104].
Anode : Cathode : Cathode : 3.3. Summary Catalyst Degradation. The main failure modes for catalyst degradation, primarily OCV, dynamic operation, load cycling, and cationic contamination can be identified by various AST methods. Considering the fact that during the lifetime of an electrolyzer operation interrupts can be expected to occur frequently, AST profiles that represent realistic operation conditions are needed. In Table 6 a summary of lifetime tests for PEMWE catalysts, including test protocols simulating an intermittent power supply, are presented.
Coupling of an electrolyzer unit with fluctuating renewable energy sources results in intermittent operation with idle periods when no power is available and hydrogen generation, when enough current is supplied [62]. This recurring transition between reducing and oxidizing conditions significantly affects the performance of the PEMWE system [62]. According to the currently understood decay mechanisms, the sluggish kinetics and instability of the OER catalyst materials constitute an obstacle toward commercialization of PEMWE technology [16]. Therefore, current research focuses on lowering the Ir loading and improving the degradation resistance [105][106][107]. In particular, understanding the degradation mechanism on the catalyst surface during OER is a cornerstone in the development of novel catalysts for PEMWE, since the state-of-the-art Ir-based anode materials are still affected by the harsh reaction environment [108].

PTL/GDL Degradation and Accelerated Stressors
In PEMFCs, usually carbon materials are used as current collectors on anode and cathode side (GDL). Due to the harsh conditions on the anode in PEMWEs, carbon materials are in most cases only used on the cathode, where hydrogen is formed. The carbon corrosion and consumption due to the high potential of the oxygen electrode result in the rapid deterioration of interfacial contact, which has a major impact on the performance and efficiency of the cell [109]. Therefore, metallic materials like Ti-based meshes, foams, or sintered powders are usually used as PTLs at the anode side of the electrolyzer. In this section, the term "GDL" is used for the material on the cathode side, and the term "PTL" is used for the material on the anode side of the electrolyzer. In a recent review, Doan et al. [22] summarized some of the state-of-the-art PTL and GDL materials used in PEM technology (Table 7). Currently, only a limited number of studies have focused on the lifetime-limiting factors for the carbon-based GDL and even less for the Ti-PTL on the anode side of the PEMWE [30]. PTL/GDL degradation can generally be divided into two groups: on the one hand, chemical degradation that is mainly caused by corrosion and on the other hand, mechanical degradation that in most cases originates from compression force, dissolution, and erosion by hydrothermal effects [8].

Gas Diffusion Layer Degradation Mechanisms.
The GDL is placed on the cathode side of the CCM between the current collecting flow field and the reactive sites of the Pt catalyst. It provides mechanical support to the MEA and is responsible for the electron and gas transport to and from the CL with minimum voltage, current, thermal, interfacial, and fluidic losses [89]. Mechanical stress that results from the high compression of the cell components strongly affects the physical properties of the GDL. Carbon fibers can break or even be displaced under compression, which can change the structure morphology, and affect the current transport and permeability of reactants [8,122]. As a result, the obstruction of diffusion pathways can lead to ineffective gas transport or saturation of the pores, decreasing the overall performance of the cell [8]. Compared to the anode compartment, the cathodic overpotential (approximately 0.1 V SHE) is insufficient to cause rapid carbon corrosion of the GDL. Although carbon-based GDLs are used in the cathode side of the cell, further research is needed to gain a better understanding of the performance of these materials under real-life conditions over time [14].

Accelerated Stressors for Gas Diffusion Layer
Degradation. In PEMFC research, various AST protocols exist, targeting for instance the chemical or mechanical degradation of the GDLs [123]. GDL degradation can occur rapidly under accelerated conditions like immersion in 13 International Journal of Energy Research OCV-AST Cycling between OCV and operating potentials (up to 3 A cm −2 ) Cycling between OCV and operating potentials leads to Ir dissolution and an increased contact resistance between the electrode and the PTL [62] OCV-AST Succession of on/off galvanostatic cycles (ranging between 0 and 1 A cm −2 ) A corrosion process in the CL is the first MEA degradation mechanism followed by thinning of the Nafion™ membrane [93] OCV-AST 5 cm −2 electrolyzer is switched from average duty point (1 A cm −2 ) to idle mode (OCV) Operation transients during PEMWE can induce conversion of IrO 2 in the anode CL into a more hydrous form [86] OCV hold test 1 h 1 A cm −2 pre-OCV followed by OCV for 72 h and 1 h operation at 1 A cm −2 post-OCV During operation, the Pt dissolution is constant, whereas at idle periods, the Pt dissolution depends on the potential and duration of OCV [88] Shut-down procedures 4000 h life tests (6 h operation everyday (shut-down over night and for weekends and holidays) When the electrolysis cell is switched off, current reversal and operation in FC mode can cause degradation of the cathode CL [60] Dynamic operation Evaluation of various stressor profiles (potential hold, triangle/ square wave, and Sawtooth up/down profiles) Less severe performance losses are reported for wind and solar AST models compared to triangle and square-wave potential cycling [90] High and elevated current  [124,125]. For instance, Liu et al. [125] performed a carbon corrosion test in H 2 O 2 solution and analyzed the effect of a MPL on commercial carbon-based GDLs. For a pristine GDL with MPL at 2.5 A cm −2 , the potential decreased by 59% and an increased water content of up to 44% was observed [125]. The GDL is the component with the highest porosity in the PEMWE, and its cellular structure leaves it vulnerable to structural changes and permanent degradation under mechanical stress [123,126]. In the study of Chun et al. [127], the mechanical durability of different GDLs with commercial and home-made crack-free MPL has been tested in a dummy cell without CL and only air supply to avoid electrochemical degradation. The results imply that the mechanical degradation during long-term operation can be significantly reduced by the preparation of crack-free MPL [127]. Dotelli et al. [128] investigated the effect of compression pressure on the GDL, when the assembly of GDLs and CCMs was compressed to 30% and 50% of its initial thickness. The higher compression ratio at 50% resulted in a reduced cell performance. This effect was partially mitigated when the operating temperature was raised to 80°C. The presence of the MPL in the carbon cloth resulted in a reduced contact resistance and therefore proved extremely beneficial for the operations, especially at high current density [128]. Radhakrishnan and Haridoss [129] performed five compression cycles on a Toray carbon paper GDL at two different pressure setpoints (1.4 MPa and 3.4 MPa). The cyclic compression experiments resulted in changes in the structure like cracks of fibers and a decreased porosity and hydrophobicity. Furthermore, a gradual loss of PTFE was observed in course of the experiments. The elevated pressure experiments at 3.4 MPa showed a more severe effect during the initial cycles [123,129].

Porous Transport Layer Degradation Mechanisms.
The PTL, which is located between the anode CL and flow field, is responsible for the transport of the reactant water and removal of the produced oxygen gas in the PEMWE cell. Furthermore, it acts as a current collector to ensure stable electrochemical performance [24]. Generally, materials like Ti powders, foams, films, or meshes are used as PTL in PEMWE cells [130]. The metal-based PTLs can undergo hydrogen embrittlement or surface passivation and are therefore often coated with precious metals such as platinum or gold with obvious effects on the capital costs. The resistance for hydrogen embrittlement is given by Lu and Srinivasan [131] and follows the following order Ti = Ta > Nb > Zr > graphite (with Ti and Ta being the most prone to hydrogen embrittlement and graphite as the least) [131].
PTL properties like surface roughness have lately received much attention because of the effect on the Irbased anode CL which is in close contact with the PTL material [22]. The formation and growth of oxygen bubbles inside the PTL pores of the anode of a PEMWE cell can block active catalyst sites and decrease the effective ECSA. Approaches exist which aim to reduce the MT losses by modification of the PTL morphology to decrease the number of coalescence host sites [132]. The choice of PTL material influences the performance of PEMWE, for example, for sintered Ti powder, the morphology and surface properties can be the most important factors, while porosity is the most important factor for thin/tunable Ti films (TT PTL) [24].

Accelerated Stressors for Porous Transport Layer
Degradation. Borgardt et al. [111] investigated the effect of pressure on the mechanical properties of Ti PTLs. In situ and ex situ durability tests were carried out on PTLs made from two differently shaped powders. Samples made of spherical powder were in a porosity range of 10-33%, while samples of irregular-shaped powder were in the range of 10-55%. According to the researchers, high porosity and thin thickness lower the physical stability of the PTL but exhibit high durability under high pressure. Furthermore, a porosity of at least 25% and 500 μm thickness enabled the PTL to resist 50 bar differential pressure [111]. Pushkarev et al. [133] investigated MEA combinations of different PTLs in PEMWE single cells. The PTL structure had only a minor impact on the activation and ohmic overpotential. However, at current densities of 2 A cm −2 and higher, structuredependent MT losses were reduced due to the higher volume of void phase and flow pore size. According to the researchers, PTLs with low porosity and small pore size could be responsible for MT limitations due to the worsened contact between the flow field channels and the PTL. In Figure 9, two SEM images of PTLs with different bulk porosities tested in this study are shown [133].
Passivation that has been reported on the anode side of the PEMWE for Ti-based BPPs will also influence the Ti-PTL if not coated with a protective layer [89]. A study  [75] shows a performance loss at constant current density (2 A cm −2 ) during long-term operation of a PEMWE due to semiconducting TiO x that is formed on the anode Ti-PTL. Coating of the Ti-PTLs with Pt achieved a substantially reduced degradation rate of only 12 μV h −1 , compared to 194 μV h −1 when no coating was applied [75].
The passivation of Ti components in PEMWEs was also observed in lifetime studies performed by Frensch et al. [57]. They investigated several MEAs with the same specifications for 500 h at different operation modes, including variations in the operating temperature and different current cycling modes (Table 8). Titanium-sintered PTLs (1 mm, 30% porosity) were used on the cathode and anode sides, respectively. In an attempt to separate the total ohmic resistance into its components, the researchers identified the passivation of the Ti-PTL and the accompanying positive effect of MEA adjustments as main PTL contributions [57].
Although the use of corrosion-resistant coatings for PTL materials is advantageous for long-term PEMWE operation, it makes small scale research hardware expensive and difficult to source and is hardly sustainable in industrial applications. Despite the fact that carbon materials are generally not used in anodic PTLs, several studies approve the use of cheaper, more abundant materials like carbon fibers in short term experiments [24].
Results from Young et al. [134] show that the stability of conventional carbon-based fuel cell GDLs is sufficient for cell-conditioning and break-in procedures as well as the initial characterization of PEMWE cells [134]. The performance of the carbon-based materials was tested in a 10 h cell-conditioning procedure followed by twelve polarization curve measurements (2.5 h each). Allowable operation of 40 h was reported that includes 15 h of operation between 2 V and 3 V [134]. Becker et al. [135] designed PEMWE single cell tests for anode PTLs made out of carbon-coated 316 L stainless steel. The substituted material was tested in situ for 30 days at 2 A cm −2 and potentials up to 1.2 V vs. RHE [135]. These results show that Ti-based PTLs are not necessarily required for initial performance characterization procedures in PEMWE.

Summary PTL/GDL Degradation.
Studies on the impact of different PTL/GDL materials on PEMWE performance are still under investigation, and the high cost of titanium is still a limiting factor that makes the development of commercialized materials challenging. To date, no effective characterization protocols exist for PEMWE, although much attention has been paid to the design and screening of PEMWE materials more recently. In Table 9, some of the available studies performed on GDL durability are summarized.
According to the currently understood decay mechanisms in GDLs, effects such as carbon corrosion and CL interaction under compression lead to a performance decay during PEMWE operation. Successful mitigation strategies so far have relied on scarce and expensive metal coatings that contribute to the already high capital cost. To date, only a limited number of degradation protocols for the PTL on the anodic side of the electrolyzer exist. Some of the available lifetime tests for PTL degradation are listed in Table 10.
Recently, great promise for PTL design optimization has been shown by Stiber et al. [136]. The researchers presented a PTL for PEMWE that enables operation at up to 6 A cm −2 , 90°C, and 90 bar H 2 output pressure. The PTL consisted of a (a) (b) Figure 9: Initial SEM images of a PTL with bulk porosity of 51.9% (a) and 25.4% (b). Reprinted from [133] with permission from Elsevier. Ti porous-sintered layer (PSL) with a low-cost Ti mesh. Interestingly, this approach did not require a flow field in the BPP [136]. Other studies suggest that traditional carbon-based GDL materials are sufficient for electrolyzer initial performance assessments [134,137]. To facilitate low-cost PTL material and component screening and development, standardized procedures for various types of materials and experimental designs are required [24].

Concluding Remarks
This document reviews and summarizes different AST protocols in PEMWE technology with a focus on the membrane electrode assembly. The component specific stressors under different operating conditions are outlined for the PFSA-membrane, catalyst layers, and GDL and PTL, respectively. Regarding material deterioration in catalyst degradation research, the current focus is set on the dissolution of the Ir-catalyst during the OER, whereas on the cathode-side Pt/C corrosion is a major issue, particularly during the dynamic operation of the PEMWE. The PFSA-membrane and ionomer degrade when the electrolyzer operates at elevated temperatures and at OCV. For the Ti-based PTL, surface corrosion and passivation are of concern. This is triggered by the high current density operation of the PEMWE, and once a passive film is formed on the surface, the contact resistance increases and the performance declines. A precious metal coating is therefore commonly applied to protect the PTL from corrosion and passivation of the metal surface. To conclude, novel materials and cell components for PEMWE are still slow to make their way into the market due to the limited access to standardized test equipment and procedures and the high cost of precious metal components. The implementation of AST protocols has been successful to some extent; however, a harmonized database needs to be created to move technology from lab scale to commercial deployment. This complementary investigation of accelerated stressors in PEMWE cells provides a basis for the development of a variety of ASTs, which can help the research community to quickly assess component durability and efficiently process novel materials. Puddle-shaped defects decrease the performance due to water accumulation [127] Static compression Mechanical compression of GDLs to 30% and 50% of its initial thickness at 60°C and 80°C, respectively A higher compression ratio results in lower cell performance. Introduction of a MPL onto the carbon cloth is extremely beneficial for the operation, especially at high current density [128] Cyclic compression Compression of GDL samples between a pair of graphite plates, which are sandwiched between aluminum end plates at 1.7 MPa and 3.4 MPa The impact of cyclic compression on the hydrophobicity of the GDL is seen to be significantly higher than that reported elsewhere due to electrochemical effects [129] Accelerated carbon corrosion  Substitution of Pt-Ti with cheaper PTL materials shows potential for significant cost savings in PEMWE stack design [135] Various operation modes/MEA See Table 8 Mitigation strategies are necessary to prevent Ti passivation from reducing the cell's efficiency [57] 17 International Journal of Energy Research