10,234
Views
18
CrossRef citations to date
0
Altmetric
Reviews

A Review on Applicability, Limitations, and Improvements of Polymeric Materials in High-Pressure Hydrogen Gas Atmospheres

ORCID Icon, , ORCID Icon &
Pages 175-209 | Received 22 Sep 2020, Accepted 24 Feb 2021, Published online: 15 Mar 2021

Abstract

Typically, polymeric materials experience material degradation and damage over time in harsh environments. Improved understanding of the physical and chemical processes associated with possible damage modes intended in high-pressure hydrogen gas exposed atmospheres will help to select and develop materials well suited for applications fulfilling future energy demands in hydrogen as an energy carrier. In high-pressure hydrogen gas exposure conditions, damage from rapid gas decompression (RGD) and from aging in elastomeric as well as thermoplastic material components is unavoidable. This review discusses the applications of polymeric materials in a multi-material approach in the realization of the “Hydrogen economy”. It covers the limitations of existing polymeric components, the current knowledge on polymeric material testing and characterization, and the latest developments. Some improvements are suggested in terms of material development and testing procedures to fill in the gaps in existing knowledge in the literature.

1. Introduction

On the one hand, conventional energy sources, mainly, fossil fuels are accountable for most of the current environmental complications, i.e., climate change, global warming, acid rains, pollution, etc., which are transforming this planet to an unlivable state. On the other hand, the energy supply is declining and leading to more world conflicts day by day.[Citation1–8] Therefore, researchers and policymakers all over the world keep working on identifying alternative fuels, which can considerably reduce greenhouse gas (GHG) emissions while maintaining the energy supply to assist the economic growth and cost competitiveness.[Citation1,Citation3,Citation7–10] The transition from the world’s current energy infrastructure, which is almost entirely based on fossil fuels, to a more diversified energy mix, with significant use of renewable energy sources and nuclear energy, and encouraging the hydrogen more as an energy carrier, especially in transportation, is identified as a feasible solution.[Citation4,Citation11] Therefore, it is worth mentioning the attractive concept of “Hydrogen economy” in which hydrogen is widely identified not as an energy source in itself, but, as a promising clean “energy vector”. Hydrogen comprises advantages over conventional energy carriers, for example, electricity, due to its capability to store and deliver a tremendous amount of energy and the possibility of easy conversion to another form of energy.[Citation5,Citation6,Citation11–14] Hydrogen as a transportation fuel, especially in fuel cell vehicles (FCV), provides immediate solutions to problems associated with current fossil fuel-based vehicle technologies, for example, a reduction in GHG and pollutant emissions, energy independence, diversification of fuel feedstocks, and onboard fuel efficiency, which is three times that of petroleum. Furthermore, as the most abundant element in the universe and richest in energy per unit mass, hydrogen has endless other applications besides road vehicles, for example, space missions, aeronautics, naval, rails, industrial and domestic energy requirements, back up energy, hydrogen village concepts, etc.[Citation1,Citation3–5,Citation7–10,Citation15,Citation16] Generally, renewable energy sources, such as solar, wind, hydro, tidal, geothermal, biomass, etc. are highly weather dependent and the inherent intermittent nature and fluctuation over time mismatch the load-demand, if they are directly connected to an electric grid. Therefore, renewable energy storage as hydrogen is one of the main upcoming solutions, which can be generated by electrolyzer and stored in different forms depending on the end-use. Hence, the development of hydrogen as a low cost, large-scale, and long-term energy storage method encourages more renewable energy industries.[Citation5,Citation7,Citation11,Citation14,Citation17,Citation18] Hydrogen has every potential to be the mainstream energy carrier. However, to be effective, and competitive for daily use, it still needs a highly reliable system, which is safe in operation, for example, in delivery, storage, and end-use strategies.[Citation1,Citation8,Citation9,Citation15,Citation19–21] The performance of these steps is highly dependent on technological improvements and material development, which will play a key role in enabling a viable hydrogen economy. This emphasizes the need of a multi-materials approach for components associated with hydrogen distribution, storage, and dispensing infrastructure to achieve a cost-effective and safe working environment.[Citation4,Citation14,Citation22]

The distribution of the produced hydrogen can be conducted in liquid or gaseous forms by pipelines, trucks, or other carriers.[Citation1,Citation4,Citation23] Both mobile and stationary applications need to store hydrogen efficiently to achieve the required performance.[Citation24] For mobile applications, hydrogen energy storage is a decisive factor, which needs to be safe, compact, lightweight, and has enough storage to cover a fairly sufficient mileage. It can be stored physically as high-pressure gas or liquid, and chemically or physio-chemically in different solid and liquid compounds, for example, metal hydrides, complex hydrides, carbon nanostructures, methane, light hydrocarbons, etc.[Citation3,Citation4,Citation8,Citation16,Citation24–26] Generally, the traditional storage facilities find difficulties due to very low boiling point (- 253 °C), extremely low density in gaseous state (0.09 kg/NA m3), and exceptionally high density in the liquid phase (70.9 kg/NA m3) of hydrogen.[Citation3] However, considering the cost, density, maturity, and ease of operation, the compressed hydrogen gas storage technology has been the most well-established hydrogen energy storage technology at the moment.[Citation19,Citation26–30] Generally, hydrogen gas has a low energy density, therefore it should be compressed to high-pressure to store a sufficient amount of energy for transportation applications as well as fuel stations.[Citation21] Achieving the intended energy storage, onboard storage of FCVs expects to reach up to 70 MPa and hydrogen fuel stations (HFS) up to 100 MPa storage pressure.[Citation3,Citation4,Citation14,Citation21,Citation26,Citation30,Citation31] In terms of safety, this should be carefully handled, especially on FCVs, “Hysafe” permeation study released legal requirements and standard on allowable permeation rates.[Citation32] Generally, the danger of explosion and fire is greater due to inherent characteristics of hydrogen, for example, low minimum ignition energy and high flammability. As a precaution, every possible leakage should be avoided, while increasing the intended pressure. Thus, the challenge arises to develop storage facilities with minimum or no leakage.[Citation16,Citation33] Therefore, the safety of high-pressure hydrogen gas storage in pressure vessels, as well as its distribution through pipes, relies on proper material choice. Earlier versions of pressure vessels and pipelines were fully made from metal-based materials. Certain challenges in use, for example, hydrogen embrittlement, were identified with those materials due to hydrogen atoms dissolution in metals, however, polymers are not subjected to this phenomenon.[Citation8,Citation9,Citation24,Citation34–39]

Therefore, the polymeric materials are more in demand for applications, where there is direct contact with high-pressure hydrogen in storage vessels and pipelines. Lately, pressure vessels have been developed containing a polymeric composite shell covered with a plastic liner (Type IV) and pipelines are being constructed with a steel tube and a plastic liner. The liners generally provide the intended gas tightness and plastic liners are lightweight, cost-efficient, corrosion resistance, and avoid hydrogen embrittlement compared with metallic liners.[Citation10,Citation16,Citation19,Citation24,Citation29,Citation40] The latest, type V pressure vessel, is fully made of composites without a liner has 20% less weight compared with type IV, however, it is still only able to handle low-pressure ranges and still needs a breakthrough.[Citation16]

Generally, from production to the HFS facility the hydrogen gas is supplied through tube trailers or pipelines at relatively lower pressure. A compressor pressurizes the supplied gas into the buffer storage, which is one or more incorporated vessels/cylinders. To avoid the increase of the gas temperature significantly during the high-pressure hydrogen refueling to the pressure vessels, which generally leads up to a vessel failure (max temperature 85 °C), a gas pre-cooling step is widely adopted. This also provides the possibility of decreasing the refueling time as well as increasing the gas density in the vessel.[Citation41–43] Dispensing the hydrogen gas to an FCV is driven by the pressure differences between the accumulator in HFS and the fuel tank of the FCV. This system contains many devices, for example, accumulators, valves, filters, nozzles, pre-coolers, for regulating the process. Furthermore, the proton exchange membrane fuel cell operates at very low-pressure levels, for example, ∼ 0.16 MPa in hydrogen gas. Hence, the fuel cell system should also contain control valves regulating the gas pressure from the tank to the cell.[Citation44–47] The connections of these devices are typically mounted with high-pressure seals.[Citation31,Citation44,Citation45] Therefore, the high-pressure hydrogen gas storage, and distribution up to the fuel cell of the vehicle need a multi-material approach for the different components. Among them, polymeric materials play a vital role in high-pressure hydrogen gas exposure applications. illustrates the main stages of the hydrogen gas energy system from production to fuel cell of FCV and the main applications of the polymeric grades. Depending on the respective application, these polymeric components are generally expected to expose a wide range of pressure (up to ∼100 MPa, static or dynamic depending on the operation) and temperature (- 40 °C to + 85 °C), except within compressors in which up to ∼ 200 °C of temperatures can be expected.[Citation20,Citation36,Citation41–43] summarizes the main components used in the process of hydrogen gas production until reaching the fuel cell of the FCV and suitable material grades for the respective function.

Figure 1. Graphical illustration of a hydrogen gas energy system; main stages and typical applications of polymeric components.[Citation20,Citation33,Citation36,Citation41–43,Citation48]

Figure 1. Graphical illustration of a hydrogen gas energy system; main stages and typical applications of polymeric components.[Citation20,Citation33,Citation36,Citation41–43,Citation48]

Table 1. Polymeric materials used in high-pressure hydrogen applications.[Citation9,Citation36,Citation49]

Among the requirements to reach high-pressure ranges of gas storage facilities, the sealing component is one of the main technological advancements needed in storage equipment.[Citation50] Generally, elastomeric materials are widely used in sealing components and additional applications, for example, control valves, flexible hoses, and connectors, which are also directly exposed to high-pressure hydrogen in-service conditions.[Citation20,Citation38,Citation51,Citation52] Mostly, rubber O-rings are used as sealing components, with the need to deliver high endurance and seal integrity for a long duration.[Citation53] Hydrogen gas leakage during service conditions can be expected mainly in three different areas: (i) gap leakage, which occurs between the seal and the equipment due to insufficient contact, (ii) gas permeation through the sealing, and (iii) leakage due to mechanical damage of the sealing. Depending on the pressure and service conditions, these factors affect the overall sealing function and, generally, at high-pressure conditions the mechanical damage significantly affects possible leakage.[Citation54] Installation of an O-ring in a high-pressure hydrogen exposure vessel and possible gas leakage are schematically depicted in .

Figure 2. Schematic representation of an O-ring; (a) installation of an O-ring in a high-pressure hydrogen pressure vessel and exposure to high-pressure hydrogen gas, (b), (c) and (d) possible types of gas leakage through a sealing component, through gaps, by permeation, and by mechanical damage, respectively.[Citation38,Citation54]

Figure 2. Schematic representation of an O-ring; (a) installation of an O-ring in a high-pressure hydrogen pressure vessel and exposure to high-pressure hydrogen gas, (b), (c) and (d) possible types of gas leakage through a sealing component, through gaps, by permeation, and by mechanical damage, respectively.[Citation38,Citation54]

The selection of the right sealing and the sealing squeeze ratio is important for gap leakage. However, the permeation in the sealing material needs to be considered in transport properties of hydrogen.[Citation54] In terms of mechanical damage of sealing components, the rubber sealing experiences gas permeation and dissolution, which may drive up to the swelling, buckling, and overflow fracture.[Citation20,Citation31,Citation55,Citation56] Furthermore, if the high-pressure is suddenly depressurized, internal fractures within the component can be expected as the dissolved gas tends to expand and try to emit from the material. The possible blister initiation within the material can drive until the surface of the component and lead to complete failure of the component, depending on the condition. This phenomenon is called rapid gas decompression (RGD) or explosive decompression (XDF) in the literature, it is reported mostly for high-pressure carbon dioxide (CO2), methane (CH4), nitrogen (N2), and argon (Ar) gases under exposed conditions.[Citation38,Citation54,Citation55,Citation57–66] Not only elastomeric components but also the thermoplastic grades, which are used as liner materials in high-pressure vessels as well as in pipelines, face severe pressure changes due to rapid gas decompression. In these applications, the liners, which are exposed to many routine gas filling-emptying cycles, and therefore, the dissolved gas into the material may generate mechanical damage, for example, blistering, and it demonstrates as cracking or whitening of the liner material due to RGD.[Citation27] Blistering due to RGD may be less common in thermoplastic materials compared with rubber materials due to less gas dissolution and better mechanical properties in thermoplastics, but, they do exist.[Citation67] Furthermore, accumulated stresses on the liner/shell interface due to dissolved gas during exposure periods even lead to the collapse of the liner during depressurization phases.[Citation10,Citation27,Citation68,Citation69] Moreover, the atmospheres, which are apparent in real operating conditions, for example, a range of temperatures, the cyclic and complex loading (tensile, compressive, and shear), long working periods, surrounding media, etc., strongly affect the expected performance of the polymeric components.[Citation70,Citation71] Therefore, the degradation, which reduces the expected functionality, and life-time of polymeric components used in high-pressure hydrogen conditions, is governed not only by the material properties but also the extreme environmental conditions.[Citation20,Citation72]

This review puts forward the current knowledge of polymeric materials used in high-pressure hydrogen gas working conditions and offers an assessment of challenges, opportunities, and possible improvements of materials and relevant test methods. The broad topic of degradation of polymers exposed to high-pressure hydrogen gas is split into (i) RGD induced damage of elastomers, (ii) RGD induced damage of thermoplastic materials and liner components, (iii) swelling induced damages to elastomeric components in high-pressure hydrogen, and (iv) aging, degradation and long-term performance of polymers exposed to high-pressure hydrogen.

2. Discussion

In high-pressure gas-exposed conditions, the transport properties of the gas/polymer combination (i.e., permeability, diffusivity, and solubility) are decisive parameters in the desired engineering applications, when the polymeric component is in direct contact with gases. Generally, the high-pressure gas penetrates the macromolecular network of polymers, which is a major drawback to the main function in gas storage or gas distribution by pipes.[Citation73,Citation74] Compared with other gases, hydrogen as the smallest element in size and molecular weight can diffuse rather easily through polymer chains. In high-pressure hydrogen storage and distributing applications, gas leakage through permeation raises concerns about safety and economic efficiency, therefore minimum permeation through liners in storage tanks and pipes is demanded.[Citation73,Citation74] The gas permeability as a function of pressure may even plasticize the polymeric material if the penetrant gas concentration is high enough to expand the polymer matrices and subsequently increase the free volume.[Citation75] Generally, elastomers with a greater free volume eventually allow gas diffusion through the polymer chains rather easily and the higher segmental chain mobility allows easier permeation of the gas into the polymer.[Citation9] Further, semi-crystalline polymeric grades may reduce the crystallinity due to the rearrangement of polymer chains and subsequent possible modulation of the glassy polymer to rubber transition. The plasticization effect can change thermal, mechanical, relaxation properties of polymeric material, including viscosity and diffusivity. However, other than the environmental factors, gas diffusion in the material and plasticization effect is dependent on the type of gas (molecular size, polarity, boiling point, etc.) and the polymer grade (polarity, additives, etc.).[Citation9,Citation62,Citation75] Therefore, the gas transportation properties are decisive factors on gas-induced material deterioration and the failure of the components at the exposure to high-pressure hydrogen gas. The following chapters review the literature on the high-pressure hydrogen gas-induced damage modes and laboratory level developments identifying the hydrogen-induced damage.

2.1. RGD induced damage of elastomers

The RGD failure is a structural damage, which is most common in elastomers, upon the sudden release of high-pressure gas. This sudden drop in the ambient pressure leads to a great volume increase of dissolved gas and subsequently a material volume increase (even up to 20%), depending on the gas/elastomer composition and exposed conditions.[Citation38,Citation57,Citation59,Citation66,Citation70,Citation76,Citation77] The process of RGD phenomena can be explained in two steps; (i) compression phase, where the gas sorption into the material as a result of exposure to high-pressure gas, (ii) decompression phase, where the sudden release of the ambient pressure is experienced by the elastomeric component and subsequently there can be expected a simultaneous pressure gradient between inside and the outer surface, a volume increase of the gas inside and gas desorption from the material due to a state of supersaturation inside the material.[Citation57,Citation59,Citation66,Citation70] In the compression phase, the gas saturates the elastomer over time into free space and at inherent defects in the material even up to plasticization. The dissolved amount of gas in the material adversely affects the damage behavior in the decompression phase, which nucleates and expands the elastomer as a result of the sudden release of the ambient pressure.[Citation62,Citation66,Citation70] During this diffusion-controlled gas desorption process, bubbles grow internally within the elastomer, which raises the stress concentration and results in irreversible damage, initiates cracks and propagates, and even develops into surface blisters, causing complete seal failure, when the stresses and strains at bubble walls surpass the tearing criteria of the polymer.[Citation64,Citation66,Citation78,Citation79] The possible phenomenology of this failure during decompression after exposure to a high-pressure gas is shown in . If no cracks are initiated after the decompression, the formed bubbles are exterminated with the desorbing solute gas molecules.[Citation62,Citation70] The degree of RGD damage depends on the exposed gas, the elastomeric material, additives, material processing method, ambient conditions (temperature and pressure), cyclic pressure loading and unloading, depressurizing rate, etc.[Citation38,Citation57,Citation66,Citation70] When the gas has a lower diffusion coefficient, it stays for a longer time in the material and generates a constant loading in a pressure accumulation after decompression and favors irreversible damage. As hydrogen is a small molecule and has higher diffusivity, the internal damage generated by hydrogen gas decompression can possibly be less than by other gases, for example, CO2, CH4, N2, etc. However, it still can be catastrophic in extreme conditions.[Citation79] Therefore, some researchers have tested elastomeric materials in high-pressure hydrogen gas conditions to gain more knowledge on the RGD phenomena and influencing parameters, as discussed below.

Figure 3. Overview of general phenomenology during the decompression phase.[Citation66,Citation70]

Figure 3. Overview of general phenomenology during the decompression phase.[Citation66,Citation70]

2.1.1. Impact of geometry and working conditions on RGD damage

Generally, the sealing components are in the squeezed condition during the working environments, therefore, Yamabe et al.[Citation80] analyzed the fracture of O-rings exposed up to 100 MPa of hydrogen gas in constrained conditions as well as in unconstrained conditions. They considered the analysis of stress behavior during volume increase and subsequent influence due to crack initiation and growth behavior using an NBR grade (CB filled 95 phr). Specimens were squeezed up to 30% and tested at room temperature (RT) with cyclic exposure (compression and decompression) up to 20 cycles. After the experiments, the observation of the cross-sections revealed macroscopic crack distribution as shown in ; whereas unconstrained (not squeezed) specimens showed randomly distributed cracks in the core and cracks arranged in the shape of a ring along with the circumference near the surface. The former cracks are expected to initiate from stress concentrations at the randomly distributed bubbles formed by saturated hydrogen in the rubber structure after decompression. As for the latter cracks, it is assumed that the shear stress due to the hydrogen concentration gradient of a solute gas molecule at decompression influences the generation of these cracks along the circumference.[Citation80,Citation81] When the specimen is constrained, cracks are aligned to the constrained direction due to stresses generated inside of the specimen perpendicular to the compressed direction as shown in . The volume increase of the O-ring during the decompression raises the tensile stresses inside the O-ring due to external compression, and it accelerates crack initiation and growth. Conventionally, the squeeze ratio (see Equation 1) ranges from 8 to 30%, as within this range the sealing performance of O-rings increases with increasing the squeeze ratio, however, above 30% there is a risk of easy damage in working conditions and subsequent leakage.[Citation62,Citation80] Eq. 1 Squeeze ratio = (σ/w) × 100%Eq. 1

Figure 4. Illustration of cracks generated after decompression; (a) inner cracks in unconstrained conditions, (b) inner cracks in constraint conditions, (c) crack propagation even up to the surface.[Citation38]

Figure 4. Illustration of cracks generated after decompression; (a) inner cracks in unconstrained conditions, (b) inner cracks in constraint conditions, (c) crack propagation even up to the surface.[Citation38]

Where, σ = W-H, W is the cross-section diameter, and H is the depth of the groove. If the decompression rate is high enough, cracks can be expected even up to the outer surface as shown in .[Citation38,Citation80,Citation81]

Kane-Diallo et al.[Citation82] conducted tests to identify the exposure pressure (7 − 15 MPa) and the controlled decompression rate (0.75 − 30 MPa/min) influence on damage behavior and they observed a set of cavities initiation and evolution in different conditions over time using a transparent EPDM (unfilled) and used this information in modeling and predicting damage.[Citation82] The in-situ captured images suggested that the number of cavities and the size of the cavities were raised with increasing saturation pressure and decompression rate. In slower releasing rates, the cavities appeared as separated damage, however, at higher rates they collided with each other and generated larger cavities. Furthermore, the second generation of cavities appeared in the material at higher decompression rates, which could be due to local stretches developed as a result of primary bubbles. A similar trend has been observed with increasing saturation pressures. Jarval et al.[Citation61] also tested a range of controlled decompression rates (0.2 − 90 MPa/min) at different hydrogen saturation pressures (0.1 − 27 MPa) in order to identify the influence of parameters on explosive decompression failure using a transparent silicon rubber. They also identified a second generation of cavities, which popped up around the first generation of cavities when higher saturation pressure and higher decompression rates were used.[Citation61] Generally, material cavitation is influenced by the hydrostatic stress component. Therefore, Jarval et al.[Citation61] also experimented with the influence of mechanical loading on the sample for RGD behavior by stretching the specimen inside an autoclave and then exposing it with the 9 MPa of hydrogen gas. In results, they clearly identified that the stretched sample generates cavities much earlier than the unstretched sample, and when the higher the extension, the cavities occurred even earlier.[Citation61] As these test series considered one-time exposure to high-pressure hydrogen, Ono et al.[Citation60] conducted tests to identify the internal damage evolution in a transparent EPDM grade as a result of cyclic exposure to 9 and 15 MPa of high-pressure hydrogen. In general, the cyclic pressure exposures generate different damage advances. Therefore, an image processing technique was utilized in differentiating each damage process, for example, separate cavities from the aggregate of cavities, extraction of cavity advancement in consecutive cycles, extraction of actual cracks from cavities.[Citation60] The key observation from this study was that the damage evolution was not a cumulative process, but it could appear at any cycle to finally coalesce.[Citation60] However, there are limitations in these in-situ methods, as only transparent grades can be observed by these methods and only 2 D projections of cavities are given by these observations through the full thickness of the specimen. Therefore, Castagnet et al.[Citation83] attempted to implement an in-situ computer tomography test set up along with high-pressure hydrogen gas-induced decompression and identified the cavity distribution through-out the entire sample during high-pressure gas decompression. The measurement set up allowed to reach 12 MPa of high-pressure hydrogen gas and a pressure release of 2.5 MPa/min.[Citation83]

A seal behavior evaluation apparatus was developed by Yamabe, Nishimura, and their research team, especially for the component level O-rings called “O-ring durability tester”, which can be used to test the actual working conditions of an elastomeric O-ring used in hydrogen energy infrastructures. This apparatus allowed them to measure the leakage from the seal in different working conditions and evaluate the intensity of different environmental parameters.[Citation20,Citation81] In one study, an EPDM grade (filled with 90 phr silicon dioxide) was exposed to high-pressure hydrogen to identify the influence of different temperatures (30 − 100 °C), saturation pressures (10 − 70 MPa), and cycle frequencies.[Citation20,Citation72] With increasing pressure and temperature more severe damage on the cross-section as well as up to the outer surface was observed. Further, the measurements at two different exposure cycle frequencies (1 min/cycle and 160 min/cycle) clearly showed different RGD resistance showing the time-dependent loading influence on crack initiation during decompression.[Citation72] Koga et al.[Citation84] extended the research work with the durability tester in cyclic exposure with additional material grades, silicon rubber (VMQ), and HNBR, identifying the influencing factors for RGD behavior in the exposure to 90 MPa of high-pressure hydrogen. They evaluated the crack damage due to the RGD phenomenon by applying tension loads to the O-rings after decompression and rating the fracture force.[Citation84] Continuing the research on identifying the RGD behavior and influencing parameters on blistering and fracture initiation, Koga et al.[Citation63] conducted tests on transparent elastomeric grades (EPDM and VMQ) exposing them to 10 MPa of high-pressure hydrogen (H2), helium (He), and nitrogen (N2) gases, separately. The barrier properties were measured for all gas/rubber combinations and the RGD tests were conducted using O-rings in the durability tester and the results clearly showed the influence of barrier properties on blister fracture. For blister forming after decompression, the internal pressure, ∏, acts proportionally to the concentration of gas dissolved in the material around the blister, C. Considering the gas solubility coefficient, S, and gas exposure pressure, P, the relationship can be simply shown as Equation 2.[Citation63] Eq. 2 =CS PEq. 2

This clearly shows that when the desorption of the dissolved gas molecule out of the material is slow, a higher internal pressure applies on the blister for a long time after decompression. On the other hand, constant loading behavior shows a time-dependent crack growth in rubber-like materials. Therefore, if the load prevails for a longer period, the possibility of crack propagation is higher.[Citation63]

2.1.2. Impact of filler strategies on RGD damage

Yamabe, Nishimura, and the research team conducted several experiments identifying the relationships between the degree of blister fracture, the material properties, and the solute hydrogen content on RGD failure when exposed to high-pressure hydrogen gas. In one study,[Citation85] the tests were conducted to identify the bulk volume increase and amount of dissolved hydrogen in CB-filled and silica-filled sulfur vulcanized NBR in different high-pressure hydrogen conditions (0.7 − 100 MPa). The bulk volume change was measured right after decompression by digital micrometer, while the residual gas amount in the material was measured by thermal desorption analysis (TDA) after the decompression. With the tests up to 100 MPa at 30 °C, they revealed that at high-pressure hydrogen exposure, the amount of hydrogen gas dissolved is proportional to the saturated pressure, obeying Henry’s law. Further, this suggests that the H2 gas exists in molecular form in the elastomer material and behaves as an ideal gas.[Citation20,Citation85] The residual hydrogen amount in CB-filled NBR was found to be twice as high compared with non-filled or silica filled NBR due to the influences of the filler-matrix interface. However, this dissolution of gas does not significantly change the bulk volume of the composite during the gas compression phase.[Citation53,Citation85] The RGD resistance was worst in unfilled conditions, a bit better, however, with a considerable number of cracks in the CB-filled material, and the best RGD resistance in silica-filled compounds in the tested conditions.[Citation77,Citation86] Another study by Yamabe and Nishimura focused on identifying the influence of the CB grade in RGD behavior in high-pressure hydrogen conditions (10 − 70 MPa), with EPDM material grade. In this study, five types of CB (N110, N220, N330, N550, and N744) and eight types of material grades, varying the degree of each CB type, were conducted.[Citation20,Citation38,Citation58] They experienced that permeability and diffusivity of hydrogen decreased with increasing the degree of filler content, and further that the permeability was not affected based on the filler grade when the filler volume fraction was the same in their tested conditions (25 phr and 50 phr). However, they found that the solubility of hydrogen gas increased when CB was added, and solubility and diffusivity were influenced by the different surface areas of the CB. The higher the surface area of the CB, higher the hydrogen gas solubility and lower the diffusivity; they assumed that this is due to the trapped small gas molecules on the filler-rubber interface. However, no significant difference in RGD resistance was observed in different CB filled material grades in the tested conditions, because the larger surface area of CB increases the gas solubility into the material, and at the same time, it reinforces the mechanical properties of the composite.[Citation38,Citation58,Citation77] Nevertheless, Balasooriya et al.[Citation87] identified an influence on RGD resistance due to different CB grades using an HNBR material grade in high-pressure CO2 in which CB with the middle range of surface area improved the RGD resistance, however, CB with high surface area showed a negative effect on RGD resistance due to possible easy breakage of large agglomeration of filler during the decompression phase.

2.1.3. Observations on bubble formation and fracture initiation

Further, Yamabe and Nishimura[Citation88] conducted tests on EPDM material grades to identify the influence of the microstructure of EPDM grade on RGD behavior and compared that with the microstructural changes after the exposure up to 10 MPa of hydrogen gas. Based on this procedure, they estimated the required hydrogen pressure for crack initiation in terms of a fracture mechanical point of view. They assumed that a crack was initiated when a grown bubble reached a threshold tearing energy of static crack growth. If the bubble size is considered as the maximum size of the defects (filler agglomerate or voids observed by SEM), the estimated internal pressures at crack initiation are similar to the experimental hydrogen pressure. In this method, by testing at different pressure ranges, possibly the hydrogen pressures at crack initiation for different rubber materials can be estimated.[Citation20,Citation63,Citation77,Citation86,Citation88,Citation89] The SEM observations of fracture surfaces revealed that the RGD behavior originated from a site of micrometer-sized defects, filler agglomerates, and voids, or from nothing visibly unusual at all.[Citation54,Citation72,Citation77,Citation80,Citation87,Citation88] Therefore, the fractures observed after the decompression are assumed to be initiated from the bubbles, which were formed around those micrometer-sized defects or at low-strength sites. These low strength sites are related to inhomogeneities of cross-link density. The generated bubbles grow further over time and conditions, and subsequently, the stress concentration around the developed bubble prompts the initiation of micrometer-sized cracks. The inherent defect may accommodate more hydrogen gas during compression and, subsequently, during the decompression phase, nucleate and expand as bubbles. Generally, rubber materials contain micrometer-sized structural defects and therefore, the micrometer-sized bubbles hardly influence the macroscopic mechanical properties of the material if these bubbles do not grow and create an irreversible crack.[Citation88] Yamabe and co-researchers attempted to detect these sub-micrometer-level bubble formations and crack initiation by the acoustic emission (AE) method as they may not be visible by optical methods.[Citation90] Further, the research was extended to observe the relationship between the microstructure of the rubber material and the nanoscale fracture process before and after the exposure to high-pressure hydrogen (10 MPa). The atomic force microscope (AFM) observations revealed nanoscale line-like structures before exposure and the number and length of those line-like structures increased with hydrogen exposure. The increase of intensity is considered as the sub-micrometer size fracture, which originated from bubbles formed by dissolved gas molecules after decompression. Based on these observations and a fracture mechanical approach, Yamabe and Nishimura determined the requirements for the nano-level fracture initiation in the rubber material.[Citation88]

2.1.4. Theory of bubble formation and modeling of fracture initiation

In the literature, there are some attempts to model the fracture initiation of elastomeric materials during the decompression phase. A mechanical cavitation model for elastomeric material, when the middle of the specimen experiences a hydrostatic tension, was first introduced by Gent and Lindley[Citation91] utilizing a poker chip test. They managed to determine a critical hydrostatic pressure, ΔPc, which is required for the initiation of cavitation damage based on the material properties. For this, they determined equation 3 based on the Young’s modulus (E) of rubber grade, and equation 4 based on the shear modulus (μ) of the rubber, demonstrating the material requirements to be fulfilled on the onset of cavitation damage in rubber-like materials.[Citation91] Eq. 3 ΔPc5E6Eq. 3 Eq. 4 ΔPc5μ2Eq. 4

Later, some more attempts were made by some authors to replicate the mechanical load requirements in cavitation damage considering the three-dimensional stress state inside the material considering the incompressibility or rubber grades.[Citation92,Citation93] Extending the modeling of cavitation onset, Jaravel et al.[Citation94] examined the influence of time-dependent behavior of the load, especially considering the real-time gas exposure and decompression data for modeling purposes using a vinyltrimetoxysilane grade exposing to a range of 3 − 27 MPa of high-pressure hydrogen inside an autoclave. Furthermore, a tensile test machine was also mounted inside the autoclave to understand the response of an existing cavity by coupling load gas/mechanical loading behavior. The controlled decompression rate (0.1 − 90 MPa/min) and the size of the existing cavity radius influence on blistering fracture were studied and compared with the developed numerical model. This study attempted to predict a critical stretch ratio criterion when a small cavity grows to a visible blister in a hyper-elastic material considering the gas exchanges between an existing cavity and the material, pressure inside the cavity, and gas concentration.[Citation94] The model predicted reasonably the threshold saturation pressure and decompression rates for cavitation damage, however, this model had some limitations in estimating the crack initiation in higher decompression rates, because this hyper-elastic material model did not consider the local viscoelastic behavior of used rubber. Further, this model lacked predicting the cavitation damage after the decompression period because it considered the cavity wall expansion only during the decompression period. But, this modeling attempt seems indispensable since they took the cavitation in time-dependent loading conditions into account considering the gas diffusion through the existing cavitation during decompression.

For modeling purposes, researchers worked up to the nanometer scale for understanding the high-pressure hydrogen gas-induced fracture initiation mechanisms. They[Citation60,Citation61,Citation94] assumed that an infinitesimal bubble initially existed in the rubber composite at the beginning and they supposed this bubble is inflated by the penetrated hydrogen gas and leads to fracture. Ono et al.[Citation95] extended the research work to understand the forming mechanism of these infinitesimal bubbles in the polymer matrix at the blister fracture initiation using an NBR grade. It is necessary to detect the dissolved state of hydrogen in the polymer matrix, and therefore, they utilized infrared spectroscopy (FT-IR) to analyze the hydrogen molecular interaction based on molecular vibration. The molecular hydrogen can exist in a polymer matrix either as isolated molecules that are entangled by the polymer chains in the matrix or the hydrogen molecules collide or interact with each other in preexisting voids in the polymer matrix.[Citation95] This research work led them to conclude that it is probable that isolated hydrogen molecules exist between matrix polymer chains rather than a coalition of more than one hydrogen molecule at high-pressure hydrogen exposure.[Citation95] Further, for identifying the origin of a bubble, Nishimura and Fujiwara[Citation50,Citation96] attempted to detect the dissolved hydrogen in rubber material by solid-state nuclear magnetic resonance (NMR). They identified new peaks, which were originated from dissolved hydrogen molecules in the material after the exposure to 100 MPa of hydrogen gas for 24 h at RT, and they could assume the existence of hydrogen molecules in the free volume of rubber as well as constrained by the molecular chains of the rubber material.[Citation50,Citation53,Citation96] Simmons et al.[Citation97] also conducted similar tests on NBR exposed up to 28 MPa hydrogen gas for 24 h at RT and detected the dissolved hydrogen within the matrix observing the peaks in the NMR spectrum. Further investigations of bubbling and initiation of cracks at RGD were conducted by identifying the change of submicron-scale morphological structures during the hydrogen elimination process by Ohyama et al.[Citation98] In this study, they used a SAXS, observing the submicron-scale voids in peroxide-vulcanized NBR composites during hydrogen elimination after exposure to 90 MPa of high-pressure hydrogen gas at 30 °C for 24 h. Based on the SAXS and Debye-Bueche function, they managed to identify two phases in NBR with a clear interface, which originated from penetrated hydrogen; one phase corresponds to voids with hydrogen, which is identified as a low-density phase and the other phase is from rubber chain molecules, which are closely packed and identified as a high-density phase.[Citation98] Ohyama et al.[Citation98] assumed the voids generated within the matrix had started from the precursors, which already existed in the matrix due to the inhomogeneity of vulcanized rubber. Further, they assumed that the initiation of the bubble occurs at these low-density phases of the matrix. When the external pressure is released then the dissolved gas is concentrated at the low-density phases and inflates until it generates submicron-sized voids, and it expands to bubbles and even up to fracture, depending on the environmental conditions.[Citation98] shows this model of void and bubble formation at low strength sites and leading up to the blister formation. Moreover, the blisters in rubber composites as a result of high-pressure hydrogen decompression may also be initiated easily as sub-micrometer-sized bubbles in the vicinity of inherent defects as well as at irregular shaped filler agglomerates as illustrated in .[Citation38]

Figure 5. Model of blister initiation; (a) right after decompression, (b) bubble formation by clustering supersaturated hydrogen molecules, and (c) blister initiation due to stress concentration caused by bubbles.[Citation77]

Figure 5. Model of blister initiation; (a) right after decompression, (b) bubble formation by clustering supersaturated hydrogen molecules, and (c) blister initiation due to stress concentration caused by bubbles.[Citation77]

Figure 6. Model of blister initiation after decompression.[Citation38]

Figure 6. Model of blister initiation after decompression.[Citation38]

2.2. RGD induced damage of thermoplastic materials and liner components

At high-pressure hydrogen exposure, gas can dissolve in liner materials (semi-crystalline polymers), especially in amorphous regions, and upon rapid gas decompression they may face blistering (localized damage) and cavitation (microscopic scale damage) similar to rubbers as explained in section 2.1. However, the dissolution is significantly less as the liner materials comprise high barrier and mechanical properties compared with the elastomeric materials.[Citation79,Citation99,Citation100] Furthermore, the hydrogen gas molecules can penetrate through the polymer and accumulate at the interface of the liner and hosting composite shell or tube in a pressure vessel or pipe. This accumulated gas puts pressure on the outer layer of the liner when the vessel or pipeline is emptied rapidly and moves away from the host shell/tube. This structural damage is identified as a major issue due to rapid gas decompression and it is commonly called “buckling collapse” or “liner collapse”.[Citation99,Citation101–103] The distortion can be plastically deformed or regain its shape after full desorption of the gas from the system, depending on the environmental conditions as well as liner thickness and the ratio of the yield stress to the stiffness of the liner material.[Citation101,Citation103] The cavitation damage in liner materials as a result of RGD may also increase the probability of buckling collapse due to the degradation of the material properties.[Citation79,Citation99,Citation100]

Generally, HDPE is widely used as the liner material due to low gas permeability and low cost. However, PA grades are also widely used due to their low gas permeability properties. Therefore, almost all the research in the literature on liner degradation is based on HDPE and PA grades. Yersak et al.[Citation27] and Baldwin[Citation104] exposed specimens of HDPE grades and PA grades to 87.5 MPa and 65 MPa of high-pressure hydrogen, respectively, and upon the RGD, they observed cavitation damage. Yersak et al.[Citation27] developed a blistering model for semi-crystalline materials in high-pressure hydrogen based on Henry's law, a yield criterion derived from continuum mechanics and Fick’s 2nd law. Based on this model, they determined the requirements of proper design of liner to compensate between a thin liner preventing blistering and a thick liner limiting hydrogen loss as well as preventing the liner buckling depending on the working conditions.[Citation27] Higher yield stresses and lower H2 solubility of PA compared with HDPE showed better RGD resistance and both grades showed increasing extent and intensity of blistering as the depressurization rate increases.[Citation27,Citation104] Similar to elastomers, in semi-crystalline materials, the RGD is a diffusion-controlled phenomenon, therefore even after decompression cavitation may appear, and the cavitation is mainly dependent on the diffusivity and mechanical performance.[Citation99] Ono et al.[Citation79] conducted tests to identify the internal damage initiation due to RGD in an HDPE grade exposed to 90 MPa of high-pressure hydrogen by analyzing transmitted light digital images. They revealed internal damage evolution quantitatively as a result of cyclic high-pressure exposure. Cross-section investigations by an optical microscope revealed that the damage due to RGD is concentrated in the middle of the specimen rather than the outer surface of the specimen.[Citation79] Melnichuk et al.[Citation99] proposed a numerical model considering non-dimensional parameters to estimate the cavitation damage of thermoplastic liner materials in high-pressure hydrogen gas conditions. Based on the model outcomes, they suggested polymers with a higher diffusion coefficient to minimize the cavitation risk. However, in this study,[Citation99] they considered only the diffusion coefficient; but the solubility of the gas in the material grade has a significant influence on the amount of gas in the material, which significantly affects the RGD damage behavior. Generally, the cavitation in semi-crystalline polymers, especially with high crystallinity ratios, is a well-known phenomenon at higher temperatures above the Tg by applied mechanical loading on the material, such as positive hydrostatic stresses.[Citation75,Citation105] Therefore, Gerland, Boyer, and their team[Citation75,Citation105] conducted tests on PVDF material grades identifying the early stage of cavity nucleation as a result of both stretching and exposure up to 12 MPa of high-pressure hydrogen gas, utilizing a tensile test apparatus mounted in a pressure chamber. The results showed that the decompression seemingly increased the size of the preexisting cavities as well as those created during tension.[Citation105]

The liner collapse behavior during the high-pressure hydrogen loading and unloading conditions was tested at a laboratory-scale by Pépin et al.[Citation10,Citation68] As the actual pressure vessel liner collapse is an expensive test, they developed a laboratory-scale liner collapse test method. A thermoplastic polymer (PA 6, Tg ∼42 °C) plate was used as the liner material and the testing specimens were prepared as liner/composite/liner layered specimens for tests at 30 MPa at 50 °C in an autoclave with high-pressure hydrogen. They confirmed the possibility to replicate the liner collapse as a result of rapid gas decompression at laboratory scale and the possibility to evaluate the quality of bonding between the layers.[Citation10] The additional test results on identifying key parameters on liner collapse and major mechanisms on collapse initiation were used for the modeling of collapsing damage behavior.[Citation68] The numerical modeling revealed that the swelling of the material as a result of gas compression or decompression has no significant effect on liner collapse. Therefore, the possible collapse phenomenon is mainly affected by the dissolved gas into the interface layer between the liner and composite.[Citation68] They revealed that liner collapse damage behavior after the depressurization of exposed high-pressure hydrogen can be explained similarly to RGD in rubber materials. During the compression phase and saturation time, the gas is dissolved into the material following Henry’s law until an equilibrium is reached where there is no pressure gradient between the free surface of the specimen and the interface between liner/composite. As a result of releasing the ambient pressure, if the decompression rate is faster than the diffusion rate of the hydrogen gas out from the specimen, a pressure gradient is generated. When this driving force is large enough to detach the liner/composite interface, then the liner collapse takes place and the trapped hydrogen gas can fill the newly created cavities and continue to increase the damage until there is no longer a pressure gradient. After fully desorbing the gas, the opening between liner and composite interface may get smaller, however, permanent deformation of liner can be expected depending on the conditions. This liner collapse damage model can be illustrated as shown in . Therefore, the mechanical properties, gas transport properties, geometry of the liner, and liner/shell adhesion properties are decisive factors in liner collapse damage besides the environmental conditions (temperature, saturation pressure, decompression rate, and decompression pressure gradient, etc.) in high-pressure gas decompression.[Citation68]

Figure 7. Schematic representation of the evolution of the amount of gas inside the liner/composite assembly; (a) during the exposure to saturation pressure, (b) during the depressurization phase, (c) at the end of the depressurization phase, and (d) after complete gas desorption.[Citation19,Citation68]

Figure 7. Schematic representation of the evolution of the amount of gas inside the liner/composite assembly; (a) during the exposure to saturation pressure, (b) during the depressurization phase, (c) at the end of the depressurization phase, and (d) after complete gas desorption.[Citation19,Citation68]

As an extension of specimen-level liner collapse test,[Citation10,Citation68] the same research group, Blanc-Vann et al.[Citation101] implemented full component tests on actual pressure cylinders made of the same material grades, focusing on the determination of the diffusion properties of the gas through the liners. The cylinders were pressurized either up to 35 MPa or 87.5 MPa and depressurized at a range of 0.007 − 0.7 MPa/min. The CT scans of cross-sections revealed the collapse damage, and they managed to identify the threshold pressure and decompression rates for the onset of damage and degree of damage for tested conditions.[Citation101] However, the liner/composite bond is not perfectly even in these prototype cylinders, which gave some limitations to precisely identifying the intensity of the liner collapse damage in different operational conditions. Therefore, to accurately identify the parameters for the onset of liner collapse damage, more tests are needed with liner/composite adhesive properties comparable to the actual pressure vessels. Furthermore, the repeating cycles of exposure similar to real-time applications seem necessary to replicate the actual pressure vessel conditions.

Considering the high-pressure hydrogen pipeline materials, Rueda et al.[Citation69,Citation102,Citation103] conducted experiments to identify the external pressure induced buckling damage to HDPE and developed a numerical model and simulation routine to replicate the buckling collapse of liners. They introduced a three-network model (TNM) for semi-crystalline materials to model the material behavior in different temperatures (0 − 60 °C) and under typical loading conditions. With the simulation results, they estimated the critical pressure based on parameters such as temperature and thickness: diameter ratio of the liner.[Citation69,Citation103] The simulation of the liner collapse based on the TNM model and their laboratory test observations on component level liner collapse were correlated to validate the developed simulation tool for time-dependent problems in the liner collapse behavior.[Citation103]

2.3. Swelling induced damages to elastomeric components in high-pressure hydrogen

If the exposure to high-pressure gas and subsequent depressurization causes the volume inflation of a rubber sealing in the groove resulting in overflow from the groove and leading to surface cracks and the rupture of the sealing, it is determined as extrusion.[Citation31,Citation71] Koga et al.[Citation81,Citation84] identified the extrusion as well as bending damage of O-rings made of EPDM when they tested with various filling ratios in the durability tester. They observed the extrusion damage and bending damage of O-rings with higher and lower filling ratios, respectively, after exposing these to 35 MPa of hydrogen, at 100 °C for 15 hours and after decompression. In bending, damage also occurs due to swelling of the rubber sealing and this leads to surface cracks. These two possible damage modes can be illustrated as shown in . Since both types of damage lead to crack propagation in the component, identifying an optimum level of filling ratio for the O-ring and the groove is vital, and it is well documented in the literature.[Citation20,Citation71,Citation81,Citation84] Further, the knowledge of the saturated hydrogen content and subsequent swelling ratio of a rubber grade is also important information to be considered in designing and selecting purposes of sealings to avoid these failures.[Citation31,Citation71]

Figure 8. Influence of O-ring filling ratios on possible fracture behavior of rubber O-ring; (a) extrusion-fracture at high filling ratio, (b) buckling-fracture at low filling ratio.[Citation84]

Figure 8. Influence of O-ring filling ratios on possible fracture behavior of rubber O-ring; (a) extrusion-fracture at high filling ratio, (b) buckling-fracture at low filling ratio.[Citation84]

Windslow and Busfield[Citation106] researched the extrusion damage phenomena of elastomeric O-ring seals and modeled this based on the compression creep test results. They developed a test rig to replicate the extrusion damage and compared these test results with FEA analysis based on the developed model. The reliability of the developed model to predict the extrusion damage was observed and further, they validated the compression creep data for modeling the extrusion damage in seals.[Citation106] Zhou et al.[Citation107–109] introduced more design attempts that avoid the extrusion damage due to excessive swelling, suggesting a combined seal structure, which is constructed by merging a rubber sealing ring and a wedge. They proposed three structures, O-ring-wedge,[Citation107] X-ring-wedge,[Citation108] and D-ring-wedge[Citation109] which are more likely to prevent the extrusion damage in high-pressure hydrogen conditions. They also developed a numerical model and a simulation routine to identify the sealing performance of every combined seal used in high-pressure hydrogen vessel by considering the swelling behavior as a result of dissolved gas. shows the design of the combined seal and its use in high-pressure vessels. The simulation outcomes suggested that the sealing capacity is enhanced with a combined seal irrespective of the shape of the sealing ring cross-section. In terms of contact stress, the D-ring combined seal has a slight advantage over the O-ring and X-ring up to 100 MPa of hydrogen pressure with the better sealing performance.[Citation107–109]

Figure 9. The schematic of the seal-wedge combined sealing ring used in a high-pressure hydrogen vessel (O-ring, X-ring, and D-ring can be replaced by each other).[Citation107]

Figure 9. The schematic of the seal-wedge combined sealing ring used in a high-pressure hydrogen vessel (O-ring, X-ring, and D-ring can be replaced by each other).[Citation107]

2.4. Aging, degradation, and long-term performance of polymers exposed to high-pressure hydrogen

Generally, designing of rubber material grades for high-pressure conditions, the soft, nearly elastic, and nearly incompressible nature, as well as a great toughness under static or dynamic loadings and high abrasion resistance compared with other material classes are expected. Additionally, the thermoplastic polymers utilized in high-pressure hydrogen gas applications are expected to deliver the desired mechanical performances and subsequent gas tightness in liner applications. However, the performance of those polymers depends on factors in operation, for example, application temperature, viscoelastic effects, and surrounding media.[Citation9,Citation70,Citation71,Citation76,Citation110,Citation111] Elevated working temperatures can alter the elastomer properties by aging, especially through chemical degradations, when the high-energy barriers for further reactions are surpassed; the degree of cross-link density and chain scissions are supposed to increase and the dominance of either mechanism depends on the rubber grade, the additives of the rubber recipe and the exposure temperature. Generally, the increase of cross-link density raises the Tg and decreases the flexibility of polymer chains, which subsequently, increases the stiffness and tear resistance of the material. However, this increases the possible scissions of macromolecular chains, reducing the strength of the rubber.[Citation110,Citation112,Citation113] Therefore, the increase of cross-link densities due to aging either minimizes the RGD damage due to a possible lower gas absorption and increase of tear resistance[Citation76,Citation113] or affects adversely in the RGD damage if a relatively higher cross-link density reduces the strength and tear resistance.[Citation114] Viscoelastic effects on polymer components depend on working time and loading history. Therefore, other than the external mechanical loadings, the high-pressure hydrogen exposure time and exposure cycles may degrade the expected mechanical properties of rubber components in a longer run.[Citation71,Citation110] Generally, when polymeric materials are exposed to high-pressure gas or liquid in-service conditions, mainly three phenomena can be expected; (i) the diffusion of fluid molecules into the polymer may modify the molecular mobility and mechanical properties (e.g., plasticization), (ii) chemical aging if the fluid is chemically active, changing the properties of polymer permanently, (iii) physical aging may degrade the properties temporarily and the transport properties of polymers, especially to polymers in a glassy amorphous state or in semi-crystalline polymers due to constraints in the amorphous phase.[Citation35,Citation71,Citation76] In contact with solvents, if the material is swollen, the material may soften and the magnitude of degradation is correlated with the degree of swelling.[Citation71,Citation76,Citation112]

2.4.1. Aging of elastomers in high-pressure hydrogen

The possible structural changes of elastomers as a result of exposure to high-pressure hydrogen gas were examined by Yamabe et al.[Citation80] and Fujiwara et al.[Citation111] using an NBR material grade at a pressure of 100 MPa at RT. Generally, the hydrogenation is the most probable reaction of NBR, which partly converts to HNBR, if the required conditions are met. However, within these studies, they observed neither any new peaks in NMR, IR, or Raman spectroscopy observations emerging within the range of CNH groups, which should have appeared if the cyano groups had been hydrogenated, nor Tg changes.[Citation111] Menon et al.[Citation9] exposed an NBR grade and an FKM grade to similar high-pressure hydrogen conditions, and they witnessed no difference in Tg of NBR, but the Tg of FKM was lowered after high-pressure hydrogen exposure due to reduced crystallinity.[Citation9] Bhattacharjee et al.[Citation115] observed hydrogenation of NBR to more than 70%, with even lower-pressure hydrogen exposures (5.6 MPa), but only in the presence of rhodium catalysts. Further, Fujiwara et al.[Citation55] continued research on evaluating the changes in the chemical structure and mechanical properties of an NBR grade as a result of cyclic exposure to high-pressure hydrogen. After 40 cycle exposures to 90 MPa hydrogen gas at 30 °C, in CB filled as well as silica filled NBR grades, no structural changes occurred due to the exposure to hydrogen.[Citation55] The influence of high-pressure hydrogen exposure on the static crack growth behavior of rubber materials was examined by Yamabe and Nishimura[Citation116] using an unfilled EPDM grade, exposing up to 10 MPa at RT. Initially, the SENT specimens were mounted on the jig inside the pressure vessel and the static crack propagation under hydrogen pressure was monitored with a video camera through a glass window and no influence on crack growth behavior due to high-pressure hydrogen exposure was observed.[Citation116] Recently, Simmons et al[Citation97] aged two NBR grades (unfilled and filled with a combination of silica and CB, dioctyl sebacate (DOS) as a plasticizer in both grades) exposing up to 90 MPa of hydrogen gas for 22 hours at 110 °C and compared the compression set resistance of aged and unaged samples. Compressing a disk specimen with a 22.2 mm and 2.9 mm of diameter and thickness, respectively, they recognized ∼ 40% of compression set increase in both aged samples compared with unaged samples. Based on the observations by helium ion microscopy (HeIM), transmission electron microscopy (TEM), and chemical analysis of both grades, they postulated that the high-pressure hydrogen gas exposure moved the plasticizer particles and agglomerated, subsequently created a phase separation within the material. This is a vital study as the compression set is a highly responsible parameter on the sealing performance of sealing components and observations of this study possibly speculate the reasoning of the increased compression set in tested conditions.[Citation97] However, this opens up some new study areas to investigate the possible influence of thermal degradation and other possible factors for compression set resistance in these conditions.

2.4.2. Aging of thermoplastics in high-pressure hydrogen

Generally, the thermoplastic materials experience significantly less gas permeation compared with elastomeric materials, but, the degree of crystalline and amorphous phases decide the permeation properties. The gas diffuses through the amorphous phases (above Tg) while the crystalline phase hinders the permeation. Due to the less intake of gas into the material, thermoplastics do not experience changes in Tg due to plasticization. However, Menon et al.[Citation9] identified improved mechanical properties in high-pressure hydrogen gas exposed PTFE at 100 MPa, RT for one week compared with the unexposed conditions. The Young’s modulus, yield stress, and strength values were increased, due to changing the polymer chain alignments as a result of gas exposure.[Citation9] Identifying the hydrogen gas diffusion into semi-crystalline thermoplastic materials and its long-term influence on mechanical properties, Castagnet et al.[Citation35,Citation73,Citation117] conducted tests using specimens of a PE grade and a PA 11 grade and multi-layered PE/EVOH/PE or PA 11/EVOH/PA 11. In these studies, they aged the specimens in pure hydrogen gas at different pressures (0.5 − 3 MPa) and different temperatures (20 − 80 °C) up to 13 months. They observed the increase of crystallinity ratios, but it did not show any changes in mechanical properties. However, the experimental set up inside an autoclave for tensile tests had some constraints as the load cell is not sensitive for small scale load levels and testing under high-pressure scattered the results. Alvine et al.[Citation118] also conducted tensile tests inside an autoclave for the HDPE grade after exposure to 28 MPa, 31 MPa, and 35 MPa pressure of hydrogen. Despite high scattering in every test, they found decreasing ultimate tensile strength (UTS) of HDPE as the pressure of hydrogen increased. Furthermore, they assumed that this effect of reducing the UTS will rise with increasing pressure due to the plasticizing effect. This phenomenon is alarming for the real-time hydrogen storage or delivery conditions with involved liner materials as they face much higher-pressure conditions.[Citation118] Klopffer et al.[Citation74] also aged the HDPE and PA 11 polymer membranes and pipe sections at different pressures (0.5 and 2 MPa) and temperatures (20 − 80 °C) for one year, and along the aging period, they monitored the permeation coefficient by the in-situ method without disturbing the aging process. However, they found the permeation coefficient unchanged under hydrogen exposure for more than one year for both HDPE and PA 11 irrespective of the aged pressure and temperature in tested conditions.[Citation74] As the plasticization of the polymers in high-pressure gas conditions is a serious concern in many ways for designing components (as discussed in section 2), more research on liner materials to get in-situ material properties exposing to hydrogen application conditions, different high-pressures and temperatures, is indispensable.

2.4.3. Influence on tribological properties in high-pressure hydrogen

Sealing components, as a result of exposure to high-pressure hydrogen gas, may swell, which changes the dimensions of the sealing, and increases the contact with the housing;[Citation31,Citation81,Citation84] further, with the possible changes in the surface properties due to exposure to the high-pressure hydrogen, the lifetime of the sealing and intended gas tightness can deteriorate.[Citation119] In addition, valves, compressors, liners in storage tanks, and pipes are made from polymeric materials and cyclic gas exposures and temperature changes cause possible back and forth motions between the polymeric and counter metallic surfaces, repeatedly.[Citation120] Duranty et al.[Citation121,Citation122] developed an in-situ tribometer and test method to be used in high-pressure hydrogen atmospheres, measuring the wear and friction properties of polymeric materials. This set-up allows the reciprocating motion of a steel ball or a pin on the polymeric specimen in a rubbing or sliding motion. They observed a higher friction and wear behavior for specimens made of an EPDM[Citation120] and an NBR[Citation121] grades in 28 MPa of high-pressure hydrogen compared with experiments in air. Nevertheless, they did not mention this behavior as an effect of high-pressure hydrogen gas exposure, as they had less information to explain this whole phenomenon within the tested conditions.[Citation120,Citation121] Therefore, this study needs to be expanded to understand the pressure, temperature effects as well as the chemical interaction between hydrogen and the polymeric material. Using the same in-situ tribometer test set up, Kuang et al.[Citation123] conducted pin-on-flat tribology measurements under high-pressure hydrogen (27.6 MPa) exposure at RT for EPDM and NBR grades identifying the effects of blended filler (CB) and/or plasticizer on the surface properties. Generally, both NBR and EPDM grades showed a higher coefficient of friction (COF) and drastically reduced wear behavior in all hydrogen exposed tests compared with the test results in ambient air conditions. The possibility of dissolved hydrogen act as a plasticizer was discussed as the possibility of lower wear behavior in high-pressure hydrogen exposed conditions. Further, based on HeIM observations, they identified a phase separation in the material as a result of high-pressure hydrogen exposure on NBR grades due to aggregation of used DOS plasticizer.[Citation123] A similar observation was reported by Simmons et al.[Citation97] However, tested EPDM grade did not show this behavior and subsequently, they postulated this phenomenon as a possible reason to experience a relatively higher COF in high-pressure hydrogen exposed NBR grades compared with EPDM grades.[Citation123] Sawae et al.[Citation124] developed a pin-on-disk method tribometer to place inside an autoclave to conduct in-situ tests in high-pressure hydrogen conditions identifying sliding friction and wear properties. They tested PTFE disk specimens (graphite filled[Citation124,Citation125] and bronze filled[Citation126]) and exposed these to 40 MPa of high-pressure hydrogen at 100 °C. The coefficient of friction (COF) results did not seem significantly changed for tests at different hydrogen pressures as well as in reference test conditions with high-pressure helium or atmospheric conditions. However, a lower wear behavior was experienced in every PTFE grade when they were exposed to high-pressure hydrogen compared with other conditions. The XPS spectra revealed some differences in chemical compositions in both transfer films and the stainless-steel counterpart surfaces after tests. Using the same test set-up and methodology, Nakashima et al.[Citation127,Citation128] examined the wear and friction behavior of PTFE in high-pressure hydrogen (40 MPa) to identify the effect of humidity in gaseous hydrogen[Citation127] and higher hydrogen pressure with longer duration.[Citation128] They observed a lower wear rate for tested unfilled PTFE grades with a higher humidity content in the high-pressure hydrogen.[Citation127] A lower wear behavior in high-pressure hydrogen exposed PTFE was monitored compared with tests in air. Therefore, the PTFE transfer film on metal counterpart was examined with XPS, and it revealed a higher peak related to metal fluoride on the thicker transfer film from the exposed PTFE. Hence, they conclude that in high-pressure hydrogen exposed conditions, steel counterparts do not form protective metal oxide layers and the pure metal composite is exposed to the active PTFE radicals, which form metal fluoride stable bonds as a transfer film and this transfer film hinders the subsequent wear due to the relative motion.[Citation128] Theiler et al. conducted tribology tests on PI[Citation129] and PEEK[Citation130] grades in hydrogen gas conditions (up to 0.1 MPa) to identify the influence on tribological properties by exposure to hydrogen. They also assumed no chemical reaction of PI or PEEK matrix in high-pressure hydrogen, however, they found tribologically induced reactions on the fresh metal surface of the counterpart and surface of PI and PEEK in the presence of hydrogen. Therefore, the hydrogen atmosphere had been favorable in the tribological properties of polymers in contact with metal counterparts.[Citation129,Citation130] But, more tests are needed to identify the influence of metallic counterparts in high-pressure hydrogen conditions and chemical analyses with different material grades to identify this phenomenon thoroughly. In general, further improvements of characterization techniques and facilities of in-situ tribological experimental set-ups are needed to get a deeper understanding on influencing parameters of wear and friction behavior of polymers exposed to high-pressure hydrogen conditions.

Zhou et al.[Citation131] studied the fretting behavior of rubber O-rings using an NBR grade exposed to high-pressure hydrogen (up to 100 MPa) and developed a numerical model describing the influence of hydrogen swelling on the fretting characteristic in terms of contact pressure and von Mises stress. They revealed that higher fretting damages are expected when the amplitude of reciprocating amplitude is higher and this occurs with an increasing swelling behavior of the sealing material as a result of gas exposure.[Citation131]

3. Outlook

The literature discussed in section 2 demonstrated that in high-pressure hydrogen atmospheres, as a result of rapid gas decompression, blister fracture, extrusion, and buckling of sealing components as well as liner collapse can be expected in polymeric components. In elastomeric components, the blistering fracture is identified as catastrophic damage, which depends on environmental factors, design aspects as well as the material properties. A material with minimum flaws is a primary requirement and from a mechanical properties point of view, a higher tensile strength, elastic modulus, and elongation at break will minimize the damage occurrence. Furthermore, increasing the barrier properties, diffusivity, and minimizing the gas solubility would decrease the pressure accumulation inside the material during the gas decompression and possibly lower the damage. Recently, Fujiwara et al.[Citation48] developed a test method to measure the transport properties of polymeric materials, especially in the equilibrium state under high-pressure hydrogen up to 100 MPa of pressure. Similar test set-ups and measurements will help to obtain precise and reliable information on selecting the materials and designing the components for high-pressure hydrogen energy systems.[Citation48] Generally, the fillers enhance the mechanical properties of polymers. Therefore, a better filler strategy would balance the mechanical properties and barrier properties, which can possibly optimize the durability of a sealing component in high-pressure hydrogen atmospheres. The controlling of transport properties is important not only in blister fracture but also for all sorts of other damage modes based on high-pressure hydrogen gas, for example, aging due to swelling, extrusion, liner collapse, etc.

The literature discussed above in section 2.1 on RGD related failure was mainly about unfilled, CB or silica filled material grades. However, specific filler strategies, which could be used to obtain required transport properties, can be found in the literature. The transport properties of a polymer/gas combination are affected by the free volume, segmental mobility and, degree of crystallinity of the polymer, as well as the size and the shape of the gas molecules.[Citation132,Citation133] Generally, adding fillers can significantly increase the tortuous path of the penetrating molecule into the polymer matrix by acting as an impermeable obstacle. This effect is very pronounced in two-dimensional fillers with a high aspect ratio, which is used to enhance barrier properties in polymers, and the best results can be obtained when fillers are arranged perpendicularly to the diffusion direction of the permeant as shown in .[Citation134] Graphene nanosheets or layered silicates, e.g., nano-clays, are well known as 2 D fillers, which are also responsible for increasing the mechanical properties. However, graphene nanosheets tend to agglomerate, and therefore surface functionalization is necessary to achieve better exfoliation of the stacked sheets and intercalation of polymer chains.[Citation134–140] Improving the filler distribution helps to minimize the filler agglomerations and generate better mechanical properties. Prioglio et al.[Citation141] studied the applicability of nano-sized graphite fillers in rubber with serinol pyrrole as surface functionalization to obtain the better filler distribution and in return, they received improved mechanical properties and fracture properties. Some research works can be found in literature in which 2 D filler strategies were used in rubber applications to reduce the gas permeability. Sun et al.[Citation40] filled a PA 6 grade with lamellar silicate nanofiller (LIC) and studied the hydrogen permeability in different pressure (25 MPa, 35 MPa, 50 MPa) and temperature conditions (-10 °C, 25 °C, 85 °C). They observed 3-5 times improved hydrogen barrier properties in filled grades compared with the unfilled PA 6 grade. The improvement was pronounced in lower temperatures due to the lower mobility of polymer chains and side groups. Further, Bandyopadhyay et al.[Citation135] examined the hydrogen permeability of nylon films coated with a polyurethane/graphene-oxide composite, Gatos et al.[Citation142] investigated an HNBR grade filled with organic layered silicates to understand the oxygen barrier properties, and Hwang et al.[Citation143] investigated an NBR filled with organoclays to minimize the water and methanol vapor permeability. Therefore, there is a high potential to introduce novel filler strategies to minimize the current limitations of polymeric materials in hydrogen energy systems. As discussed in section 2.1.2, the addition of filler enhances the mechanical properties, which is essential in RGD resistance, however, it may also minimize the RGD resistance due to possible increased gas solubility into the matrix and negative effects of gas diffusivity. Further, excessive filler agglomeration can soften the rubber material due to Mullin’s effect and Payne’s effect in cyclic exposure conditions, and at the same time, it can create some easy crack propagation paths due to not firmly attached filler-filler bonds, in case of blister initiation. Therefore, the filler size and morphology, the anisotropy of the filler interface, filler-polymer compatibility, inherent defects of filler during processing should be carefully considered during the selection process of the filler for avoiding every possible weak link. Some literature[Citation132,Citation144] also reported possible delamination at the interface of two-dimensional stacked fillers, at higher loading or inferior compatibility and these aspects should be considered for long-term applications.

Figure 10. Increased tortuous path length, generated by orthogonally arranged two-dimensional obstacles.[Citation134]

Figure 10. Increased tortuous path length, generated by orthogonally arranged two-dimensional obstacles.[Citation134]

As the above section 2.1.4 suggests, the blistering fracture originated in relation to extrinsic defects or inhomogeneity of the rubber microstructure. Considering this micrometer-sized defect as the stress concentration point for crack initiation and propagation, a fracture mechanical approach can be utilized in identifying the fracture behavior in rubber material as a corresponding tool to blister fracture in RGD resistance. However, no research was found in the literature using the fracture criterion of rubber materials in correlation with RGD resistance for considering the requirements of material properties. Specifically, impact-like loading conditions can be used to understand the blister fracture behavior at higher gas decompression rates, as the decompression damage initiation is most commonly seen as a single-cycle phenomenon. The elastic strain energy released rate, G, based on Griffith’s energy approach[Citation145,Citation146] would be most applicable for characterizing non-linear materials, and Rivlin and Thomas[Citation147] adapted this approach finding an approximation relation of fracture energy for elastomeric materials and derived the equation for the single-edge notched specimens (SENT) as shown in Equation 5. SENT is simple and easy to use and can be utilized in these experiments. Eq. 5 G = 2KWaEq. 5

Where, K, W, and a are the strain-dependent term, elastic strain energy density, and notch size, respectively. Lake et al.[Citation148,Citation149] experimentally proved the applicability of G at maximum load “Gmax” as a material property for elastomers, which is the tear resistance parameter. Recently, Agnelli et al.[Citation150] experimentally proved and provided the dimensional requirements and calculation techniques of rubber SENT samples for precise G determinations. Therefore, utilizing SENT specimens and impact-like loading tests maintaining constant loading rates can be utilized in characterizing the material requirements in RGD resistance. In general, the component level RGD measurements are expensive and time-consuming. Therefore, this initial laboratory specimen level characterizing tool for elastomers would be more convenient and encouraging in material development processes.

4. Summary and conclusions

Hydrogen as an energy carrier comprises the potential to be the leader of energy providers changing the existing energy infrastructure and our lifestyle, fulfilling the demands of all the stakeholders. However, to fully unveil the potential and benefits of hydrogen, the materials, which are used in these harsh conditions, can be a limiting factor and it will be a challenge to the academic and industrial communities to cope up with the demand. Already, elastomers and semi-crystalline polymers have been used in many vital components related to hydrogen gas energy systems with enormous progress, however, the development of these materials for pushing the limits for better performance, lifetime enhancement and to fulfill the industry-driven targets of high-pressure gas conditions is an ongoing process. Therefore, the recent progress of polymeric materials used in high-pressure hydrogen atmospheres, their limitations, the development of test methods, and some modeling attempts have been discussed here. Mainly, the influence of rapid gas decompression on rubber as well as semi-crystalline polymers was discussed including the aging of polymers in high-pressure hydrogen working conditions. These phenomena are responsible for attenuating the expected functional properties of polymeric components and even up to the fracture of sealing and subsequent failure, extrusion and buckling, further in blistering and liner collapse in liner applications in high-pressure hydrogen gas cylinders and distribution pipes. The RGD damage behavior is dependent on many factors. For this reason, up to date knowledge on high-pressure hydrogen conditions was discussed thoroughly in this review. Despite the great acknowledgment on current developments in test methods, novel materials, and research findings, the main criticism concerning most of the available research knowledge in literature is that they have been conducted in lab-scale conditions using experimental level material grades. Thus, there is a high potential in research work related to actual material grades in their actual working conditions and in industry-driven target conditions. Moreover, a fracture mechanical analysis on material properties, which can be possibly used as a ranking tool on RGD resistance, and possible improvements to the material transport properties by introducing novel filler strategies were also discussed as further improvements to fill in gaps in existing knowledge in the literature.

Acknowledgements

This research work was performed at the Polymer Competence Center Leoben GmbH (PCCL, Austria) and within the COMET-module “Polymers4Hydrogen” within the framework of the COMET-program of the Federal Ministry for Transport, Innovation and Technology and Federal Ministry for Economy, Family and Youth, with contributions by the Department of Polymer Engineering and Science (Montanuniversität Leoben). The PCCL is funded by the Austrian Government and the State Governments of Styria, Lower Austria and Upper Austria.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by the by the Austrian Research Promotion Agency (FFG) under grant numbers of 854178 and 21647053.

References