1. Introduction

Hydrogen (H₂) has the highest energy content per unit weight (142 kJ/g), surpassing that of biofuels such as methane, methanol, and ethanol[1]. H₂ is also used for direct combustion in internal combustion engines and fuel cells. H₂ combustion is environmentally friendly as it produces only water. However, carbon-based nonrenewable sources such as natural gas, coal, and heavy oil that are used for H₂ production are unsustainable and emit greenhouse gases[2]. Hence, the generation of H₂ from renewable sources such as biohydrogen is attracting increasing interest as a sustainable energy carrier.

Biohydrogen production needs renewable resources such as biomass and microorganisms to generate H₂. Unlike conventional commercial methods, biohydrogen production needs ambient temperature and pressure, resulting in reduced energy and cost[3,4-6]. The use of these technologies can reduce the generation and release of greenhouse gases such as CO₂ in the environment. However, their commercial and pilotscale production has not yet been achieved, because scaling up of these technologies poses significant challenges such as low hydrogen yields, unresolved hydrodynamic issues in bioreactors, and the optimization of operating conditions that require further research and development[7]. To overcome these challenges, many laboratories and pilot-scale studies are being conducted to develop the most suitable approach for generating biohydrogen[7].

Generally, H₂ production from biohydrogen occurs via three mechanisms: light-dependent biophotolysis, light-dependent photofermentation, and light-independent dark fermentation[5]. Biophotolysis involves the utilization of sunlight and carbon sources to generate H₂. Photofermentation uses photosynthetic bacteria with organic substrates and light as energy sources. Dark fermentation utilizes carbohydrates such as glucose to produce H₂ along with fermentation byproducts[8].

Unstable and insufficient H₂ production explains why biohydrogen fuel cell systems have not yet been commercialized[ 9]. For example, proton exchange membrane fuel cells (PEMFCs) require high hydrogen purity of up to 99.99%. Biohydrogen production typically yields mixed biogas containing H₂, CO₂, and possibly small amounts of CO, CH₄, and H₂S. The presence of CO and H₂S is detrimental to the fuel cell stack and reduces PEMFC performance[9]. Hence, it is essential to investigate H₂ separation and purification systems to produce high-purity H₂, with a particular focus on the removal of H₂S and CO₂.

Many hydrogen separation and purification technologies have been developed so far, including cryogenic distillation, pressure swing adsorption (PSA), absorption, and membrane technologies[10-12]. Cryogenic separation cools the mixture to remove unwanted gases and obtain purified H₂; however, it is impractical for small portable hydrogen plants because of limitations in purification and compression. PSA could also be considered for large-scale biohydrogen production in the future. Membrane separation, specifically by using nonporous polymeric membranes, is promising for biohydrogen production[13]. However, scaling up these membranes for large-scale use for H₂ separation and purification requires careful consideration of various factors such as selectivity, durability, and operating conditions. Therefore, further research and development is necessary to advance membrane technology for effective H₂ separation and purification in biohydrogen production.

This review aims to broadly describe general biohydrogen production and purification technologies. It specifically focuses on various polymer membrane gas technologies for H₂ separation and purification.

2. Biohydrogen-generation Methods

Biohydrogen-production technologies can be divided into two categories: light-dependent and light-dependent[ 7]. Dark fermentation is a light-independent process, whereas biophotolysis and light fermentation are light-dependent. Biophotolysis, which occurs in cyanobacteria and microalgae, involves splitting of water using solar energy and CO₂ as the carbon source. Light fermentation involves the degradation of organic compounds by photosynthetic bacteria in the presence of light[14,15].

Biophotolysis is a simple biohydrogen-production process(Table 1). It uses solar energy and CO₂ to convert water and substrates into oxygen and hydrogen [2]. However, biophotolysis affords relatively low H₂ yields owing to limited light absorption and conversion efficiency in darkness[16,17]. Light fermentation utilizes biomass feedstock such as organic waste and is used as an alternative method for waste treatment and generating renewable energy[18]. Dark fermentation, which does not rely on light energy, is the most profitable process, as it yields higher H₂ production than that achieved using the other two processes[19]. However, large-scale H₂ production through dark fermentation requires both fundamental and applied research and development efforts[20].

Direct biophotolysis produces H₂ by utilizing solar energy and algal photosynthetic systems, specifically those of Chlamydomonas reinhardtii (green algae) and Synechocystis (cyanobacteria)[21]. In this process, photosystem I (PSI) absorbs light and transports electrons to ferredoxin, whereas photosystem II (PSII) absorbs light and generates electrons from water. The light absorbed by PSII is essential for the oxidation of water, resulting in the production of electrons, protons (H⁺), and oxygen (O₂). These electrons are then transferred through the electron-transport chain using the light energy absorbed in PSI to reach the hydrogenase enzyme. This process does not involve any intermediate CO₂ fixation[21]. Hydrogenases catalyze the recombination of protons and electrons to evolve H₂ gas. The overall reaction of direct biophotolysis, also known as onestage direct biophotolysis, involves direct production of biohydrogen from water and solar energy through the algal photosynthesis system.

Biohydrogen is generated by photofermentation using photosynthetic bacteria and light energy to convert organic compounds[22]. Photosynthetic bacteria, classified as purple bacteria (purple non-sulfur and purple bacteria) and green bacteria (gliding bacteria and green sulfur), lack PSII; hence, they depend on PSI for photosynthesis during H₂ production[22]. However, similar to biophotolysis, these bacteria are not capable of water splitting. Instead, they use organic acids, such as acetic acid, as electron donors to generate H₂. For instance, purple non-sulfur bacteria produce H₂ in the photofermentation process through nitrogenase enzymes under nitrogen-deficient conditions using organic acids and light energy[22]. Despite extensive research in this field, the practical application of this process is hindered by several technical barriers, including low volumetric production rates, inefficient H₂ production by nitrogenases, low photosynthetic conversion efficiency, and low light conversion efficiency[16-18].

The complex process of dark fermentation comprises a series of biochemical reactions involving various bacteria[ 7]. H₂ production occurs during hydrolysis and acetogenesis during anaerobic degradation. In the absence of methanogenic bacteria, organic matter is degraded into H₂, CO₂, and volatile fatty acids (VFAs). To promote H₂ production, the environment must support the growth of H₂ producers and minimize the population of hydrogen consumers[19]. In dark fermentation, feedstock, such as biomass and waste, is subjected to pretreatment to eliminate methanogens and enhance the hydrogen yield[19]. For example, cellulose- containing agricultural waste, when pretreated, undergoes size reduction and delignification, whereas food processing waste may require only physical treatment[ 23]. Glucose, a preferred carbon source, undergoes glycolysis and breaks down into pyruvate and nicotinamide adenine dinucleotide (NADH) during the early stages of dark fermentation[23].

Dark fermentation is an attractive method for renewable hydrogen production because it can utilize sugars and carbohydrates from various wastes and biomasses as substrates[23]. This process utilizes hydrogen-producing enzymes that generate hydrogen without consuming oxygen. In addition, well-established fermentation reactor technologies and bioprocess controls can be utilized. As the name implies, dark fermentation does not require light energy input. In addition, as this method requires the use of plant-derived carbohydrates, it aligns with the optimized agricultural practices of the developed countries[7].

Dark fermentation also affords some additional advantages for H₂ production, including high efficiency, low cost (as it does not require light energy input), the ability to use diverse substrates (including the low-cost agricultural waste), and the use of stable H2-producing enzymes because of the absence of oxygen evolution [24-26]. However, it generates a significant amount of byproducts (e.g., butyrate, acetone, and organic acids). Approximately two-third of its substrate produces byproducts, leaving only one-third of the substrate to produce H₂. Additionally, dark fermentation primarily generates mixed biogas containing H₂ and CO₂, along with potential traces of CO, CH₄, and H₂S[22, 29]. To utilize H₂ as a fuel in internal combustion engines or fuel cells, it is necessary to use a separation and purification system to obtain high-purity H₂ gas. CO₂ present in the biogas can negatively impact the fuel cell performance.

3. Biohydrogen-purification Technologies

The purification process for biohydrogen involves treating the impure hydrogen gas by compressing and densely storing it and allowing it to compete with other gases such as gasoline and natural gas. Various technologies, including cryogenic distillation[12], PSA [10], and membrane technologies[11], have been employed for H₂ separation(Table 2).

Cryogenic distillation involves cooling a mixture to low temperatures to remove unwanted gases by liquefaction. This process yields highly purified H₂ because of its low boiling point in the gas mixture. However, it is not suitable for small portable hydrogen purification and compression units. Cryogenic processes are commonly used in large-scale production of hydrogen owing to cost considerations[30]. When applied to biological mixtures containing CO₂, the transitioning of CO₂ from the gaseous state to the solid state at 1 atm pressure becomes challenging, potentially clogging the system[31]. To overcome this problem, the operating pressure was increased to enable the change of the CO₂ phase from gas to liquid[31]. Therefore, the cryogenic process is conducted at very low temperatures (120 K) and high pressures, making it energy-intensive. In addition, it requires various instruments for compression, refrigeration, and separation, which, however, increases the construction and operational costs[30]. The efficiency of the purification process depends on the partial condensation of the gas mixture at high pressures and low temperatures, typically affording an overall efficiency of 95~98%[32]. Owing to these challenges, the industry is currently moving toward adopting alternative separation systems and moving away from cryogenic methods.

PSA is one of the most extensively employed technologies for H₂ separation[33]. It is considered a cutting- edge technology for producing pure H₂ in the petrochemical industry[34]. The PSA process involves four steps: adsorption, depressurization, blowdown, and pressurization(Fig. 1)[35]. PSA can be applied in medium- to-large industries as well as in pilot and small portable systems, yielding H₂ purities ranging from 99.5 to 99.9%[35]. Primarily used for separating H₂ from other gases including CO₂, the efficiency of H₂ purification in PSA depends on the adsorption capacity of the adsorber at both high and low pressures[36].

In this PSA, gases are selectively adsorbed at high pressures and desorbed at low pressures. Higher pressure enhances H₂ adsorption onto the adsorbent, whereas lower pressure facilitates the detachment of H₂. The purity of H₂ depends on the binding forces between the gas molecules and the adsorbent, making the selection of the adsorbent crucial for achieving effective separation[34]. Various adsorbents have so far been developed, including silica gel, activated carbon, and zeolites[ 34]. These adsorbents exhibit different capabilities for removing impurities(Table 3)[34]. For instance, silica gel is effective in removing impurities from water vapor, but is ineffective with CO₂ separation. Conversely, activated carbon is suitable for CO₂ removal, but it is less effective for removing water vapor.

Multiple bed absorbers should be used to enhance the purity of adsorbed H₂ in practical applications. This configuration, also known as the polybed PSA process, usually comprises 7~16 beds (depending on the mixture conditions) and a cycle configuration with at least three pressure-equalization steps to maximize H₂ recovery and throughput[34]. The primary focus of the research on polybed PSA systems is the development of an optimal bed-to-bed configuration[37]. Although the PSA system is a well-established technology, the biohydrogen process operates under conditions such as ambient pressure and moderate temperature (~40°C). Consequently, the use of the PSA process is quite challenging for biohydrogen application. As shown in Table 2, PSA is the most widely used technology for H₂ separation, particularly for large-scale production. It is currently being tested for both laboratory and pilot-scale biohydrogen production[38]. If proven successful on a larger scale, PSA could be used for biohydrogen separation during production.

Membrane technology is a promising technique for the separation and purification of biohydrogen. Membrane systems are advantageous because of their modular nature, fewer operating units, and low energy consumption[39]. Various types of membranes are available for enriching H₂. These membranes can be classified as porous or nonporous. They can also be further divided into organic or inorganic materials. In general, porous membranes have pore diameters of < 2 nm[40-42]. They employ a molecular sieving mechanism as the primary means of separating H₂ from larger species such as CO₂, CO, CH₄, H₂O, and H₂S. Consequently, the size and configuration of the pores play pivotal roles in this separation process, exhibiting enhanced H₂/CO₂ selectivity at elevated temperatures. Zeolite, silica, and carbon-based membranes are the most commonly employed porous membranes[40-42]. However, the precise control of the pore size and shape still remains a major challenge. As an alternative, a porous polymeric membrane incorporating ionic liquids has been proposed as a supporting material to form supported ionic liquid membranes (SILMs), which can be employed for biohydrogen separation [43,44]. One of the key benefits of employing SILMs for biohydrogen separation is the difference in the solubility of various gas components within ionic liquids (ILs) (Fig. 2). This attribute has significant potential for the effective separation of H₂ and CO₂.

The gas-separation efficiency of SILMs strongly depends on the ionic liquid, supporting membrane, and separation conditions[45-47]. In addition, the SILM configuration resembles that of dense polymeric membranes, with separation based on the gas mixture diffusivity and solubility that are affected by pressure and temperature[47]. Increasing the transmembrane pressure gradient enhances gas permeability, but may cause IL loss, defects, or leakage. Thus, it is crucial to study the effect of transmembrane pressure, which determines the critical displacement pressure based on the pore structure, IL interface tension, and contact angle, to improve the mechanical durability[48,49].

Traditionally, researchers tend to use dense membranes for H₂ separation[50-53]. Nonporous metallic membranes, such as palladium, exhibit exceptional H₂-separation performance; however, their high operating temperatures (approximately 400°C) are not favorable for biohydrogen separation[50,51]. Therefore, the development of metallic membranes specifically for biohydrogen purification is crucial, because it could reduce the material cost and improve the separation efficiency under biohydrogen-operating conditions. Other dense membranes, particularly polymeric materials, are also promising candidates for biohydrogen separation (Fig. 3). These materials demonstrate reasonable separation properties in similar biohydrogen-separation systems, which operate at ambient temperature and pressure[ 54,55]. Furthermore, their low costs align well with the biohydrogen-production process, making them potentially useful technologies for biohydrogen separation.

4. Polymeric Membranes for Biohydrogen Separation

Various materials have been proposed for producing dense membranes(Table 4). Promising results have been observed with the testing of organic membranes (polymers) for the separation of H₂/CO₂ mixtures (Table 5)[54,55]. Typically, these membranes operate under conditions similar to those required for biohydrogen production (ambient temperature and pressure). In addition, their low costs make them promising technologies for biohydrogen separation.

Nonporous polymeric membranes generally facilitate mass transfer based on the solution-diffusion model [56,57]. Considering the transport mechanism, two material design approaches have been proposed(Fig. 4): one focuses on diffusion-dominant materials, while the other involves the synthesis of solution-dominant materials. Interestingly, H₂ and CO₂ have distinct preferences for transport mechanisms[55]. H₂ transport relies on diffusion, whereas CO₂, a condensable gas, is transported based on its solubility[57,58]. Some recent studies have suggested the possibility of using unique membrane materials that combine both features, that is, they can allow diffusion and solubility-based separation[59,60]. Most materials allow this possibility. The selectivity of polymeric membranes significantly affects the gas-separation efficiency. In H₂-selective polymeric membranes, H₂ extensively permeates the membrane, impeding the passage of other components, including that of CO₂ [57,58]. High selectivity is achieved when H₂ exhibits high diffusivity, while CO₂ has low solubility[57,58].

Despite exhibiting different dominant transport behaviors, both H₂ and CO₂ are fast-permeating gases [61]. Therefore, the primary objective of polymeric membranes is to enhance the H₂/CO₂ selectivity. The nature of the polymer used allows for the production of membranes with either sorption selectivity for CO₂ or diffusion selectivity for H₂[61]. Generally, polar rubbery polymers exhibit the maximum sorption selectivity for CO₂ and minimal diffusion selectivity for H₂, resulting in an overall selectivity of up to 10 for CO₂ over H₂. In contrast, highly rigid glassy polymers such as polybenzimidazole (PBI) maximize the diffusion selectivity, favor H₂, and can achieve an overall selectivity of 10–20 for H₂ over CO₂(Table 5)[55].

The feed temperature and pressure play significant roles in gas diffusion through membranes. Theoretically, a higher pressure difference between the feed and the permeate sites results in an increased flux and selectivity[ 65], offering an opportunity to enhance the separation efficiency without requiring drastic material changes. For instance, a commercial Matrimid^® membrane demonstrates feed-partial pressure-dependent separation properties in terms of gas mixture permeability and selectivity[55]. The operating temperature is another critical factor influencing the selectivity and permeability of membranes during gas mixture separation [66,67]. The theoretical transport mechanism indicates that gas diffusion into the membrane is directly proportional to the operating temperature because of the increased molecular-diffusion activation energy[66,67]. Additionally, higher temperatures reduce gas solubility in the membrane, favoring diffusion-selective transport, as observed for H₂-separation membranes. A study utilizing a H₂-selective polyimide membrane showed higher separation factors as the operating temperature increased from 21 to 65°C[67]. However, determining the optimal temperature for H₂/CO₂ gas mixture separation is crucial, because both components exhibit different gas-transport behaviors.

5. Organic-based Membrane Reactor for Biohydrogen Separation

A membrane contactor (MC) employed to enhance mass transfer between gas and liquid phases upon contact. These devices facilitate mass transfer between the target gas and the absorbing liquid by utilizing a membrane. Consequently, MCs effectively merge membrane and absorption separation methods. A MC system comprises circularly operated membrane absorber and desorber units. The gas separation mechanism is the same as the present absorption–desorption separation system, but the difference is the presence of a membrane segregating gas and liquid phases in each separation unit. For H₂/CO₂ mixtures, the feed gas enters the absorption unit, where CO₂ permeates the polymeric membrane and pure H₂ exits as retentate. Absorbed CO₂ moves to the desorption unit, passes through the membrane, and exits the system. The liquid absorbent returns for continuous operation.

In general MC polymeric membranes are porous, however their use poses challenges due to bubble formation and liquid penetration, causing contamination. As an alternative, non-porous polymeric membranes offer greater potential because of higher mass transfer coefficients, resistasnce to feed pressure, and no contamination. Notably, Polyvinyltrimethylsilane (PVTMS) membrane displays high gas permeability, ideal for biohydrogen separation[68]. Furthermore, PVTMS and potassium carbonate (K₂CO₃) achieve efficient CO₂ removal in MC systems, extending to CH₄/O₂ mixtures. When used different liquid absrobents such as monoethanolamine, this system achieves higher CO₂ separation efficiency (99%) than K₂CO₃ solution (57%) in PVTMS membrane-based MCs[69]. As a result, chemical absorbents like MEA and K₂CO₃ offer better mass transfer, practicality, and eliminate the need for additional compression systems seen in physical absorbents.

The MC system's integration with biohydrogen production has been explored. Beggel et al. demonstrated a 99% CO₂ separation efficiency from real biogas (H₂/CO₂ mixture) and obtained 90% pure biohydrogen with 10% CO₂ by coupling an MC with photofermentation. They employed batch separation due to the slow biogas production rate of the photofermenter, achieving high purity biohydrogen. However, MC's adoption is less prominent compared to NPPM in this context[7].

6. Conclusions

Production of biohydrogen from renewable sources is vital for sustainable energy production. Biohydrogen production can be performed through biophotolysis, photofermentation, and dark fermentation. However, challenges related to low hydrogen yields and optimization of operating conditions need to be first addressed. Furthermore, the production of high-purity H₂ suitable for application in H₂-combustion engines and fuel cells requires efficient separation and purification systems. Cryogenic distillation is cost-intensive and unsuitable for small-scale applications. PSA offers high H₂ purity. Membrane technology affords various advantages, such as low energy consumption and modularity, making it a promising option for biohydrogen separation and purification. The study results show that dense polymeric membranes are promising candidates, although they require further research and development for large-scale application.