New Perspectives on Catalytic Hydrogen Production by the Reforming, Partial Oxidation and Decomposition of Methane and Biogas

Abstract

The article provides a short review on catalyst-based processes for the production of hydrogen starting from methane, both of fossil origin and from sustainable processes. The three main paths of steam- and dry-reforming, partial oxidation and thermo-catalytic decomposition are briefly introduced and compared, above all with reference to the latest publications available and to new catalysts which obey the criteria of lower environmental impact and minimize the content of critical raw materials. The novel strategies based on chemical looping with CO₂ utilization, membrane separation, electrical-assisted (plasma and microwave) processes, multistage reactors and catalyst patterning are also illustrated as the most promising perspective for CH₄ reforming, especially on small and medium scale. Although these strategies should only be considered at a limited level of technological readiness, research on these topics, including catalyst development and process optimization, represents the crucial challenge for the scientific community.

1. Introduction

Hydrogen and synthesis gas (syngas) are a fundamental intermediate for several key industrial processes, particularly ammonia and methanol synthesis as well as the Fischer–Tropsch process for production of synthetic fuels and oils and hydrogenation reactions of the petrochemical and chemical industry [1]. Furthermore, hydrogen is considered the leading candidate for substituting fossil fuels in power generation and storage (power-to-fuels approach), this transition representing a key strategy for the decarbonization of the energy industry to achieve the scenarios of carbon mitigation [2,3]. H₂-favorable energy-to-mass storage capacity and its completely CO₂ free combustion are the main advantages that today drive the increasing political and economic interest in its large-scale implementation in the energy field [4,5], while its high versatility and reactivity arouse interest in novel applications in the pharmaceutical, metallurgical and chemical industries [6]. As a result of the prospected growth of the world’s energy and chemical demand, combined with the need to develop more sustainable processes with lower greenhouse gas emissions, hydrogen demand is predicted to increase dramatically in the near future, and new reliable and more sustainable pathways for its large and small-scale production are needed, with particular emphasis on production from renewable sources [7].

Hydrogen can be classified according to its material and energetic origin into grey hydrogen, black and brown hydrogen, blue hydrogen and green hydrogen [8]. Grey, black and brown hydrogen are produced from fossil sources, from natural gas (NG) reforming/partial oxidation, bituminous coal pyrolysis and lignite coal pyrolysis, respectively, with lowest prices between 1.0 and 2.0 USD/kg, and nowadays it accounts for over 90% of the total global production of hydrogen, with large associated CO₂ emissions [9]. Blue hydrogen is also derived from fossil fuels, but with integration of carbon capture, storage and utilization pathways included in the production process to abate greenhouse gas emissions, and finally, green hydrogen is derived from water electrolysis powered by renewable sources. Another possible source for renewable hydrogen production is biomass, through either pyrolysis or gasification [10]. An illustration of hydrogen color according to production pathway is reported in Figure 1. It has to be noted that, while the classification of grey, brown, black and blue hydrogen is generally well defined [8,11,12,13], there is still debate on the color designation of other production sources. For example, the term yellow hydrogen, here designating hydrogen produced by power grid-riven electrolysis, as defined by Arcos and Santos [8], is sometimes used to designate instead hydrogen obtained from thermochemical solar splitting of water [11]. There is also not a general agreement on color classification of biomass-derived hydrogen, which is considered a green hydrogen production method by some authors when considering carbon neutral pathways [5,14] but a brown or blue hydrogen production method by others when considering non-renewable energy sources and the high CO₂ emission if carbon capture technology is not implemented [8].

Catalytic processes remain a crucial pathway for hydrogen production, especially for large-scale application, and much focus has been invested in research both for optimization of the current processes and the development of innovative and more sustainable pathways. Several reviews have been dedicated to H₂ production with thermochemical processes and related catalysts, dealing with both catalyst [15,16,17,18] and reactor/process design [19,20] and with different raw material for fuel [21,22,23,24]. It should be noted that while noble metal catalysts have repeatedly proved their high activity and stability for steam reforming [25], dry reforming [26] and partial oxidation [27] reactions, their high cost and scarcity renders their industrial application unfeasible, and therefore it was chosen to not include them in the present review. Several of these processes can also be potentially applied to biogas as well as natural gas, biogas being a complex mixture constituted mainly of methane and carbon dioxide derived from a variety of biological sources such as biomass waste material [28], which is particularly interesting from the perspective of green chemistry in order to lower carbon dioxide emissions and replace fossil fuels with new environmentally friendly alternatives; therefore, some applications of catalytic processes to biogas have also been considered, even though the presence of impurities such as sulfur compounds in biogas complicates its utilization compared to methane and natural gas [29,30].

The present article intends to review the current state of the art and future perspectives of catalytic hydrogen production from methane and biogas, mainly analyzing material properties and process aspects aimed at improving productivity and yield for the product of interest. Thermochemical processes utilizing different feedstock, with particular focus on CO₂ reuse, and conventional and novel catalysts are analyzed, and their current advantages and limitations are critically discussed to put in evidence areas requiring further research. For details on the environmental and economic aspects of these processes, relevant considerations can be found in the following references [31,32,33], while more general reviews including also hydrogen transport, storage and utilization, as well as alternative hydrogen production sources such as water electrolysis, can be found in the following references [5,34,35,36]. This work presents itself as a continuation of reviews performed by Chen et al. [37] and Boretti et al. [38] by including the newest published results for catalyst development, with a secondary focus on process and reactor modeling and design. In particular, new perspectives on chemical looping, multistage processes, catalyst patterning and electrically assisted reforming are reviewed. For chemical looping processes, the promising application of cerium dioxide carriers is discussed.

2. Steam Reforming

Steam reforming of methane is the most common method for hydrogen and syngas production and it is responsible for the production of 47% of the global hydrogen [39,40]. A detailed review of the state-of-the-art reactor design for steam reforming reaction has been recently provided by Ganguli and Bhatt [20]. Most recently, a novel design for a composite scale microreactor was also presented, with potential for distributed production [41].

The main reaction for steam reforming of methane can be written as follows:

$$\mathrm{C}{\mathrm{H}}_4+{\mathrm{H}}_2\mathrm{O}⟶\mathrm{C}\mathrm{O}+3{\mathrm{H}}_2\Delta {\mathrm{H}}_{298\mathrm{K}}^0=+206\frac{\mathrm{k}\mathrm{J}}{\mathrm{m}\mathrm{o}\mathrm{l}}$$

(1)

Besides natural gas, heavy hydrocarbons and other feedstocks such as biogas can also be used. Typically, the process involves feeding methane and an excess of water vapor onto a Ni-based catalyst [42]. The process operates at a pressure between 3 and 25 bar, and the reaction is strongly endothermic; therefore, substantial heat has to be provided in order to maintain the process temperature at around 650–1100 °C [43]. This heat is usually provided by burning part of the hydrocarbon feed, but this approach is not environmentally friendly, as it leads to high carbon dioxide emissions [44]. The excess of feed vapor is required in order to avoid carbon deposition on the catalyst surface, but the need to produce large quantities of superheated vapor represents another large energy sink, limiting process efficiency and further contributing to high CO₂ emissions. Because of the high capital costs of equipment and its unsuitability, however, steam reforming technology is indicated for large-scale productions [45] (400 ton/d or above [46]), and is not suitable for medium- and small-scale plants (50 ton/d or below) [45], which would be needed to allow the distributed production of hydrogen.

As steam reforming is already a mature technology, at a TRL of 9 [9], current research is mainly focused on achieving further process optimization, by improving on current catalysts in order to increase process efficiency and reduce the problems correlated with carbon deposition [42], sintering [47] and catalyst poisoning [48] that limit the efficiency of conventional Ni-based catalysts. In particular, a Ni-based catalyst may be improved by providing adequate supports [49]. CeO₂ support, for example, has been investigated due to its potential for reducing coke deposition, thanks to its oxygen exchange properties [50]. The core-shell structure of the catalyst may provide a good solution for improving sintering resistance: for example, the core-shell of Ni/SiO₂ applied to both dry and steam reforming displayed no deactivation in 40 h during methane dry reforming, and only limited deactivation during the initial phase of steam reforming [51]. A porous yolk-shell-structured Ni/Al₂O₃ catalyst can offer good performance and improved resistance to alkali poisoning and carbon formation, improving its suitability for biogas reforming [52].

Bimetallic catalysts are another promising solution for steam reforming [53]. In particular, iron oxide can promote Ni activity and resistance to carbon deposition [54], allowing for the reduction of the content of Ni and improving the lifetime of the catalyst. However, it is worth noting that Ni oxides are toxic and carcinogenic [55], and thus do not conform to green chemistry principles.

3. Dry Reforming

Due to the increased appeal of carbon capture technologies and decarbonization, interest in converting carbon dioxide to valuable products, such as syngas, methane and methanol, is becoming increasingly relevant in the chemical and energy industry [56]. Utilization of captured CO₂ as raw material for the production of intermediate chemicals is considered an interesting solution [57], as opposed to geological storage [58] and the application for further fossil fuel extraction through Enhanced Oil Recovery [59].

The dry reforming reaction (2) is considered a very interesting route for valorizing carbon dioxide by its conversion to synthesis gas through a reaction with methane, as it involves the simultaneous conversion of two common greenhouse gases and is particularly interesting for biogas conversion, as biogas is already a mixture of carbon dioxide and methane and can be treated without the need for an additional separation step [60,61,62].

$$\mathrm{C}{\mathrm{H}}_4+\mathrm{C}{\mathrm{O}}_2⟶2\mathrm{C}\mathrm{O}+2{\mathrm{H}}_2\Delta {\mathrm{H}}_{298\mathrm{K}}^0=+247\frac{\mathrm{k}\mathrm{J}}{\mathrm{m}\mathrm{o}\mathrm{l}}$$

(2)

The resulting produced syngas, with a H₂/CO ratio of 1, is particularly suitable for the synthesis of oxygenated compounds [63], but it can also alternatively be further processed in a water–gas shift unit, where carbon monoxide reacts with water, thus increasing the hydrogen production.

Dry reforming, on the other hand, is an even more endothermic reaction with respect to steam reforming, and the high operating temperature needed for the process (600–1000 °C [64]) represents a challenge in process design, particularly in avoiding catalyst deactivation by sintering. Carbon dioxide activation is very difficult, as it is a highly stable molecule [64]. Furthermore, the reaction is severely limited by high carbon deposition due to methane cracking (3) and the Boudouard reaction (4), which also causes rapid catalyst deactivation and parasitic reverse water–gas shift reactions between H₂ and CO₂ (5); this lowers the H₂/CO ratio of produced syngas below 1, making the produced syngas unsuitable for chemical synthesis and therefore lowering its economic value [49,65,66].

$$\mathrm{C}{\mathrm{H}}_4⟶\mathrm{C}+2{\mathrm{H}}_2\Delta {\mathrm{H}}_{298\mathrm{K}}^0=74.9\frac{\mathrm{k}\mathrm{J}}{\mathrm{m}\mathrm{o}\mathrm{l}}$$

(3)

$$\mathrm{C}{\mathrm{H}}_4⟶\mathrm{C}+2{\mathrm{H}}_2\Delta {\mathrm{H}}_{298\mathrm{K}}^0=74.9\frac{\mathrm{k}\mathrm{J}}{\mathrm{m}\mathrm{o}\mathrm{l}}$$

(4)

$$\mathrm{C}{\mathrm{O}}_2+{\mathrm{H}}_2⟶\mathrm{C}\mathrm{O}+{\mathrm{H}}_2\mathrm{O}\Delta {\mathrm{H}}_{298\mathrm{K}}^0=42.1\frac{\mathrm{k}\mathrm{J}}{\mathrm{m}\mathrm{o}\mathrm{l}}$$

(5)

Much work has been carried out on the development of catalysts for the dry reforming reaction, particularly to address the problem of deactivation due to coke deposition [67].

Ni-based catalysts are again the most common choice for the dry reforming reaction. Much interest is in the development of novel and optimized supports capable of limiting Ni sintering and carbon deposition, by evaluating new materials and more elaborated support structures. For example, Phan and Minh [68] investigated the use of wet precipitation-obtained calcium hydroxyapatite and Mg-doped calcium hydroxyapatite as support for wetness-impregnated Ni. They observed that catalytic activity and stability were affected by the Ca/P ratio of the support, with high ratios producing increased activity but reduced stability. They suggest the high activity of Ca-rich samples to be attributed to increased basicity and improved CO₂ adsorption. Mg addition showed no effect on activity, but it also improved stability. The maximum selectivity observed for syngas production was 85%, and all catalysts exhibited carbon deposition negatively affecting their stability, while sintering appeared negligible. Nakajima et al. [69] investigated the kinetics of dry reforming on Ni-impregnated mesoporous MCM-41 silica support, and observed conversions higher than 85% at 973 K. They found that the reaction on their substrate could be described by the Langmuir–Hinshelwood mechanism. It should be noted that their investigated catalyst had a high Ni load, of 20%. Zhang et al. [70] also evaluated the effect of SiO₂ structure on the catalytic activity of Ni-impregnated catalysts and found that the best performance was obtained for monodispersed silica spheres (MSS) with inverted conical-shaped porosity at a 10% Ni loading. When compared to MCM-41 and commercial silica, MSS demonstrated improved pore impregnation and segregation of dispersed Ni nanoparticles that inhibited sintering and coking and reduced apparent activation energies for CH₄ and CO₂ decomposition, but the influence of reverse water–gas shift led to a final H₂/CO ratio slightly lower than 1. Similar results were obtained [71] for mesoporous SBA-15 silica and microporous beta silica: it was observed that, for catalysts prepared through the ammonia evaporation method for Ni deposition, beta silica had too small porosities, which did not allow proper Ni penetration, with the accumulation of Ni outside the pores leading to sintering and coking during the DRM reaction. The conventional impregnation method for Ni deposition also led to worse dispersion compared to the ammonia evaporation method [71].

The effect of Ni-support interactions for aluminum-nitride supports was inspected and it was observed that the formation of an ultrathin Al₂O₃ layer on the catalyst surface improved catalyst stability by preventing sintering and coking, achieving up to 89% conversion of methane with a 1% weight Ni load [72]. Haug et al. [73] investigated the effect of phase boundary structure for Ni/ZrO₂ catalysts. Interestingly, they found that an “inverted” catalyst, with a ZrO₂ phase deposited on the bulk Ni surface through chemical vapor deposition exhibited better anti-coking properties. Zhang and colleagues [74] analyzed the effect of a triple interface in Ni-MgAlO_x/BN catalysts and found that a boron nitride surface successfully reduces coke formation by inhibiting the cleavage of the final C-H bond, while the MgAlO_x surface acts as a dispersant for Ni particles inhibiting sintering and offers activation sites for CO₂, leading to stable 73.9% CH₄ conversion at 750 °C. Georgiadis et al. [75] investigated lanthanide oxides as support and found that while La oxides displayed the highest activity, they were also negatively affected by carbon deposition. Samarium oxide was suggested as the better option, as while slightly less active than lanthanum, it displayed significantly reduced carbon deposition. Cerium and praseodymium oxides, while coke resistant, displayed much lower activity, due to much weaker Ni-support interaction, and showed significant activity loss with time on stream.

Cobalt catalysts are also of interest for the dry reforming reaction. The addition of 10% yttrium promoter to the Co catalyst supported on WC-activated Carbon composite was shown to increase lattice oxygen availability, inhibiting coke formation while at the same time avoiding sintering and oxidation of Co particles [76]. Also, the proper selection of the calcination temperature (600 °C) increases catalyst stability, while yttrium addition inhibits the reverse water–gas shift reaction, ultimately improving the syngas H₂/CO ratio.

Depending on the preparation method, controlled adsorption or dry impregnation, cobalt-based catalysts supported on γAl₂O₃ bring about different performances. Indeed, the samples prepared by controlled adsorption displayed increased activity due to better Co dispersion and a lower tendency to deactivate for coke formation [77]. However, in both cases, coking led to catalyst deactivation within 8 h of reaction at 700 °C and the reverse water– gas shift reaction led to a H₂/CO ratio much lower than 1. Improved results were obtained for mesoporous Al₂O₃ support promoted with Y₂O₃ [78], for its use in optimized conditions; a methane conversion of 88.97% was recorded at 900 °C, with H₂ yield of 31.0% and CO yield of 69.4% (H₂/CO ratio ~ 1). Owgi et al. [79] compared the performance of Ni and Co catalysts supported on fibrous silica alumina and observed that Ni samples exhibited much higher activity and stability compared to Co, with a maximum conversion of CH4 of 97.5% for Ni/FSA and 20.6% for Co/FSA.

Ni-Co bimetallic catalysts show improved coke resistance when compared to monometallic Ni or Co catalysts; indeed, when the bimetallic alloy, Ni-Co, is supported on mullite, an increased activity and a better coke/sintering stability, compared to monometallic catalysts, can be achieved with more than 90% CH₄ conversion at 800 °C and showing only limited deactivation in 40 h on stream [80]. A 1:1 ratio of Ni:Co load was suggested as the optimal solution, as a catalyst with a higher Ni fraction displayed less activity, due to aggregation problems. Liang et al. [81] prepared Ni-Co nanoparticles encapsulated in attapulgite-derived MFI-zeolite and obtained a lower methane conversion of 71%, but their catalyst showed stable operation for 100 h. They propose that the zeolite shell not only confines metal nanoparticles avoiding sintering, but also helps in coke removal, thanks to its high sorption capacity for CO₂.

Copper and iron are cheap, non-toxic materials; therefore, their use in a catalyst is interesting from a safety and environmental point of view. The iron addition to a nickel-based catalyst supported on MgAl₂O₄ can bring the formation of a Ni-Fe alloy that reduces carbon deposition and increases CO₂ conversion, with the best Fe/Ni ratio equal to 0.7, as higher amounts of iron led to lower methane conversion [82]. Basicity of support also affects the performance of a catalyst, favoring CO₂ activation. Li et al. [83] also determined the best Fe/Ni ratio of 1 on Ce-Al₂O₃ at 600 °C. Higher-Ni-load catalysts would be more favorable for methane activation; however, they are prone to dealloying, leading to the formation of agglomerated Ni particles prone to deactivation by coking.

Ni/Cu supported on γAl₂O₃ in a ratio equal to 8:1 was found to be optimal for reaction at 650 °C. While limited deactivation and coking were still present for a bimetallic catalyst, it was seen that, for a Ni-Cu catalyst, mainly re-oxidable carbon was formed, compared to a monometallic Ni catalyst forming mainly refractory graphitic carbon [84]. The Cu presence partially suppressed the RWGS reaction, allowing for a stable 0.9 H₂/CO ratio. A Ni/Cu ratio of 8 was also confirmed as optimal by Han et al. [85] for catalysts on SiO₂ support, with stable 77.5% conversion of CH₄ and 84.5% conversion of CO₂. Khan et al. [86] developed and validated a Density Function Theory (DFT)-derived kinetic model for dry reforming on Ni-Cu catalysts. Such a modeling approach is able to provide an accurate description at the molecular level of the geometry, stability, and reactivity of chemical species adsorbed onto catalytic surfaces, and it is thus suitable for describing the reaction kinetics [87].

Other metallic pairs have been studied: a Co-Ce composite with a 3:1 Ce/Co ratio prepared through a glucose-assisted method displayed good activity (the maximum conversions of CH₄ and CO₂ were 87.2% and 54%) and stability, thanks to strong Ce-Co interactions [88]; Co-Sm, with yields of syngas >80% at 900 °C and no detectable carbon deposition [89,90]; and a Ni-Mo catalyst supported on monocrystalline MgO, a synthesized catalyst that offers high stability (850 h on stream) [91].

Perovskite-based catalysts are another alternative that is attracting much interest, thanks to the stability and tunability of the perovskite structure and reactivity by cation substitution. LaFe_0.9Ni_0.1O₃ perovskite displays improved performance when compared to Ni, Fe, and Ni-Fe La₂O₃-supported catalysts for dry reforming of ethane [92], thanks to the high dispersion of Ni nanoparticles formed in the reducing reaction environment and the fact that it remained stable over time, as well as the presence of oxygen vacancies that could dissociate CO₂. Suttiumporn et al. [93] investigated the effect of different M metal ions for La_0.8Sr_0.2Ni_0.8M_0.2O₃ perovskite, and found that while Cu substitution increased initial activity, Fe substitution allowed much higher stability. The stabilizing effect of Fe substitution was also confirmed for La_0.9Sr_0.1NiO₃ perovskite [94]: while the undoped perovskite irreversibly decomposed under DRM conditions, leading to increased coking, a 50% substitution of Ni by Fe permitted the obtaining of a more stable, regenerable perovskite structure with stable CO₂ and CH₄ conversion. Cerium doping in La_1−xCe_xNi_0.5Fe_0.5O₃ enhances activity for dry reforming when x = 0.4–0.6, with high syngas selectivity [95]. More recently, a very stable (260 h on stream) Ni-Fe catalyst supported on PrBaMn_1.6Ni_0.3Fe_0.1 O_5+δ double-layered perovskite has been obtained [96]. Remarkably, this latter material displayed promising stability and activity also for high pressure reactions.

4. Partial Oxidation

In a partial oxidation reaction, a hydrocarbon (typically methane) is reacted with sub stoichiometric oxygen to achieve its conversion to synthesis gas (CO + H₂) [97], as reported below for the case of methane:

$$\mathrm{C}{\mathrm{H}}_4+\frac{1}{2}{\mathrm{O}}_2⟶\mathrm{C}\mathrm{O}+{2\mathrm{H}}_2 \Delta {\mathrm{H}}_{298\mathrm{K}}^0=-36\frac{\mathrm{k}\mathrm{J}}{\mathrm{m}\mathrm{o}\mathrm{l}}$$

(6)

Contrary to steam reforming, partial oxidation is an exothermic reaction, avoiding the need to provide heat by combusting part of the hydrocarbon feed. While it is possible to operate with a purely thermal process, the use of a catalyst allows for a lower reaction temperature (800–900 °C for a catalytic process compared to 1200–1500 °C for a pure thermal process [98]). Partial oxidation also requires the supply of pure oxygen to avoid N₂ dilution of reaction products and has significant problems correlated with temperature run-off [99], which lead to loss in selectivity in favor of complete oxidation. As for dry reforming, coking [100], catalyst poisoning [101] and sintering [102] are also major issues for this process.

Ni-based catalysts have been vastly researched for PO on a variety of supports. Barbero et al. [103] investigated a Ni catalyst prepared by wet impregnation of La₂O₃, MgO and ZrO₂. La₂O₃-supported samples evolved into a perovskite structure and displayed the highest conversion, selectivity, and stability, while the ZrO₂-supported catalyst was the least active, and rapidly deactivated due to Ni sintering due to weak metal–support interaction. Good performance was obtained for 4% mol MgO-ZrO₂ support with 30% Ni load. In particular, higher MgO loading was found to increase coking, as the MgO-ZrO₂ solid solution formation was impaired [104]. Recently, Ni catalysts supported on fibrous and monolithic γAl₂O₃ have been compared, and the fibrous catalysts were found to achieve above 10 times the yield of their monolithic counterparts, thanks to better Ni dispersion and better mass transfer [105]. Fibrous catalysts also display high conversion and selectivity, with low carbon deposition, even for a high Ni loading of 45%, while the conventional wet-impregnated monolith sintered at high loading. Addition of magnesium to mesoporous alumina supports was investigated by Özdemir and Öksüzömer [106]: both Ni supported on MgO and Ni supported on MgAl₂O₄ displayed improved activity compared to the Al₂O₃ support. Interestingly, their samples were prepared through a solid combustion process without calcination. More recently, Khaleel et al. [107] determined that a 3% Fe doping of γAl₂O₃ support successfully reduces coke formation, and in particular the formation of crystalline carbon, without compromising activity.

MgO and MgAl mixed oxides have been evaluated as support also for Co-based catalysts [108]. While MgAl support displayed good activity, in this case samples supported on MgO experienced intense sintering and oxidation of Co species, leading to fast deactivation. An optimal 20 wt% Co load on 63 wt% MgO MgAl support was determined as the best-performing solution; however, carbon deposition was visible even for this catalyst. Choya and coworkers [109] instead evaluated the performance of Co supported on CeO₂/Al₂O₃ and found that a 12 wt% CeO₂ surface coating of alumina successfully weakened Co-Al₂O₃ interactions, inhibiting the formation of the inert cobalt aluminate phase, while increasing oxygen mobility. The resulting catalyst was, however, very sensitive to the presence of water in the reaction feed, which lead to partially irreversible deactivation. It has been previously observed that performance of a Co catalyst on γAl₂O₃ support prepared by chemical reduction could be increased by increasing the synthesis pH to 13.68 [110]. This latter catalyst displayed an initial methane conversion of 71.02% at 750 °C and a CO selectivity of 61.49% with a H₂/CO ratio of 2.05. Catalyst activity was stable for 70 h on stream, but after 110 h on stream the catalyst underwent a 27.15% loss in activity, due to coking.

Bimetallic catalysts have also been extensively studied for partial oxidation reactions. Fakeeha et al. [111] compared monometallic and bimetallic Ni and Co catalysts on ZrO₂-Al₂O₃ supports and determined that while monometallic catalyst performed best for a high calcination temperature, the Ni-Co catalyst provided the best conversion when the calcination temperature was low. They attributed the negative effect of a high calcination temperature on the bimetallic catalyst to spinel phase formation. It should be noted that at 800 °C they observed selectivity of CO formation > 99% for all catalysts. Interestingly, they observed that the monometallic Co catalyst calcined at 800 °C provided the best activity and stability, outperforming both the monometallic Ni and bimetallic catalyst for this support. Cobalt addition was instead observed to be beneficial for the Ni catalyst on zeolite MCM-41 support [112], where formation of the Ni-Co alloy inhibited formation of NiO and prevented Ni atoms from destroying the molecular-sieve structure of the support. A 1% addition of Co offered the best performance, with stable 88% methane conversion for 100 h at 750 °C and GHSV = 18 L g⁻¹ h⁻¹ [113]. Synthetized Ni-Co catalysts supported on ZnO nanoprisms showed remarkable stability, with a methane conversion of 98.11% for 56 h on stream and a H₂/CO ratio of 2.11 [113].

The effect of Cu and Ce promoters for Ni catalysts supported on mesoporous silica SBA-16 was investigated too [114], and it was observed that while Ce addition provided a synergistic effect with Ni, resulting in high activity even for lower Ni loads, Cu addition was detrimental for catalytic performance [114]. Compared to the monometallic Ni catalyst, bimetallic Ni-Ce displayed the best activity for oxygen-rich feed, as the presence of Ce inhibited Ni oxidation. On the other hand, the H₂/CO ratio for these catalysts was consistently below the stoichiometric value of 2 and CO selectivity was low, which is attributed to the influence of the RWGS reaction and total combustion. Carbon formation was also present. This contrasts with previous research [115], where a high activity for a Ce-Cu catalyst prepared by the intermetallic route was observed, comparable to a noble metal catalyst. On the other hand, Ce-Cu catalysts obtained by the sol–gel method demonstrated much lower activity and poor selectivity. Ce addition was instead not beneficial on the SiO₂-supported Ni catalyst [116], displaying delayed methane activation, low conversion and high selectivity for total oxidation compared to monometallic Ni when the reaction was started at a low temperature and complete deactivating occurred after following the temperature ramp. The behavior of the cerium-doped catalyst was completely different when reaction was started above the methane reduction temperature, in which case Ni-Ce and Ni demonstrated similar behavior, with Ce not providing significant improvement apart from a slight enhancement in coke removal.

Perovskite catalysts have also attracted attention for PO. Loktev et al. [117] investigated PrNi_0.5Co_0.5O₃ and SmCoO₃ activity both for PO and DR reactions, and determined that while PrNi_0.5Co_0.5O₃ displayed high activity for DR but not for PO, SmCoO₃ instead exhibited high activity for both reactions, with syngas yield > 90% at 900 °C in both cases. The activity of samarium cobaltate for PO was recently confirmed [118]. Here, the surface modification through supercritical anti-solvent precipitation was found to provide higher coke resistance. La_1−xGd_xCrO₃ with a varying Gd content (x = 0, 0.2, 0.5 and 0.8) was supported on porous Al₂O₃ [119] and tested in a burner reactor. All Gd-doped samples exhibited increased peak reaction temperature, compared to undoped samples, with increase in Gd doping leading to a reduction of the peak. The presence of the porous support successfully enhanced mass and heat transfer. The optimal yield for H₂ production was obtained for a Gd doping of x = 0.2 and a catalyst-support pellet size of 8 mm.

Molybdenum phosphide (MoP) was also evaluated as a catalyst both for the DR and the PO reaction [120], and it was found to exhibit different behavior. While for DR the catalyst displayed a methane conversion of only 65% at 900 °C and rapidly deactivated due to the coking formation of the Mo₂C phase and surface oxidation, a stable performance was obtained for the PO reaction, with near 90% methane conversion and 80% H₂ selectivity at the same temperature.

As the partial oxidation reaction is an exothermic and fast reaction, proper reactor design is essential to avoid formation of hot spots and to maintain a short residence time, to avoid complete combustion and excessive coking. While the non-catalytic partial oxidation process, despite still being less competitive than steam reforming, has already reached considerable maturity, particularly for higher hydrocarbon feedstocks (a novel high-temperature reactor-design methodology was, for example, recently published and validated [121] for the partial oxidation of various hydrocarbon feedstocks), catalytic processes, while in theory allowing for a lower operation temperature, have so far only been demonstrated in pilot-scale plants; this is due to catalyst limitations [98], with only a few small-scale applications already commercially available [122], such as short-contact-time reactors [123].

5. Autothermal Dual/Tri-Reforming

As evidenced previously, steam- and dry-reforming reactions are highly endothermic, requiring considerable heat to be provided to the reformer reactor, while partial oxidation is an exothermic process which can be negatively affected by temperature runaway due to heat exchange impairments in the reactor. Furthermore, each process is only able to produce a syngas within a set range of H₂/CO ratio (maximum of one for dry reforming, two for partial oxidation and three for steam reforming), limiting the flexibility of the process and requiring further downstream treatments such as the water gas–shift reaction to tune the ratio to the desired value for further synthesis. These issues can both be addressed by combining the reforming and oxidation reactions in a single process. By properly tuning O₂/CH₄ and H₂O/CH₄ ratios in the feed stream, steam reforming and partial oxidation can be combined to obtain an autothermal steam-reforming (ATSR) process where the partial oxidation reaction provides the heat needed for the reforming without the need for external heat exchange [34]. Similarly, dry reforming and partial oxidation can also be combined in an autothermal dry-reforming process (ATDR) [124], while the steam-reforming and dry-reforming reaction can be combined in a dual-reforming process (2-R) to adjust the H₂/CO ratio (in this case, the process remains endothermic) [125]. The combination of all three reactions of SMR, DRM and PO, utilizing a combined CH₄/CO₂/H₂O/O₂ feed, is instead referred to as tri-reforming (TR) [126]. Both dual- and tri-reforming processes are particularly interesting for biogas, as well as, for flue gas conversion [62,127], as they avoid the need for separation of methane and CO₂ before reaction and, furthermore, the addition of water and oxygen to the reaction environment reduces the issue of carbon deposition that is predominant in pure dry reforming and allows good tuning of the H₂/CO ratio of produced syngas [14]. As previously seen for reforming and partial- oxidation reactions, catalyst design plays a crucial role also for coupled processes.

Ni-based catalysts are the most common choice for ATSR process [35,128], which typically operate in a temperature range of 800–1200 °C [129]. Support selection and preparation is crucial to ensure catalyst performance. For example, mesoporous γ-Al₂O₃ prepared by an easily scalable solid-state method was tested for autothermal reforming and displayed high surface area and methane conversion when calcined at 500 °C, with stable operation for 20 h on stream [130]. Further investigation by the same authors also investigated the effect of La₂O₃, CeO₂, ZrO₂ and SrO promoters on γ-Al₂O₃ catalysts and found that Ce promotion leads to the highest catalytic activity towards hydrogen production, as it also promotes WGS reaction [131]. Finally, a 3%-weight Ce addition was found to achieve optimal activity and resist carbon deposition [132]. Matus and colleagues investigated the effect of noble and non-noble metal promoters (M) on a Ni-M/Ce_0.5Zr_0.5O₂/Al₂O₃ catalyst and found that a Rhenium-promoted catalyst 10Ni-0.9Re/Ce_0.5Zr_0.5O₂/Al₂O₃ displayed self-activation capacity without the need for previous reduction, and stable reactivity with a high H₂ yield of 70%, close to that of a noble-metal-promoted catalyst [133]. Aurajo et al. [134] compared Ni catalysts supported on alumina and titanate perovskite and found that while all catalysts were active for the autothermal reforming, only perovskite-supported catalysts displayed stable operation, while alumina-supported catalysts deactivated due to sintering and nickel oxidation. Barium and calcium titanate displayed a high methane conversion of around 70%, while strontium titanate was much less active and exhibited coke formation due to weak metal–support interaction. Reactor design is particularly important for autothermal reformers and it has been the focus of substantial research. Considerable work has been dedicated to process optimization through numerical simulation, one of the most recent examples being the work by Tariq et al. [135] for optimization of autothermal reforming on a NiO/Al₂O₃ catalyst through the response surface method. Their work evidenced that temperature is the most influential parameter for autothermal reforming, followed by steam-to-carbon and steam-to-oxygen ratio, and they obtained an optimal performance at 699.85 °C, 1 bar, steam-to-carbon ratio of 3 and oxygen-to-carbon ratio of 0.45. Murmura and colleagues [136] numerically compared three different reactor configurations: the conventional co-feed autothermal reformer, the spatially separated thermally coupled methane combustion and the reforming and distributed-oxygen-feed reformer. In the conventional reformer, oxygen is introduced to the reactor directly mixed with the methane and water feeds: this configuration allows for the reaching of equilibrium conversion in a very limited reactor space, but suffers from very high temperature peaks at the reactor inlet, which may negatively affect the catalyst and reactor durability and reduce process safety and controllability. On the other hand, the spatially separated configuration featuring a first catalytic combustion step thermally coupled to a consecutive reforming step and distributed-oxygen-feed configuration can obtain the same hydrogen yield as the conventional ATSR process at the cost of slightly larger reactors but avoiding dangerous temperature peaks. Folded reactors were also found to provide the opportunity for higher conversion and hydrogen yield in a numerical simulation by Chen and Wang [137] when compared to conventional and heat-pipe reactors. In particular, conventional tubular reactors suffer from heat exchange limitation that leads to inefficient heat transfer and poorer catalytic activity. In an alternative to conventional tubular reactors, a novel radial-flow spherical reactor was also recently proposed [138], minimizing CO₂ production and allowing higher flowrates and catalyst loading, thanks to a negligible pressure drop. Finally, Gul and colleagues [139] performed a numerical evaluation of a CO₂ sorption-enhanced autothermal dry-reforming process with CaO as CO₂ sorbent coupled with a Ni/MgO catalyst, allowing for a methane conversion as high as 94% with a 97% hydrogen purity.

Autothermal steam reforming has so far reached considerable technological maturity [14,122], and applications are already available on the pilot scale. The start-up and long-term operation of a 50 Nm³/h pilot plant for biogas autothermal reforming has displayed high plant efficiency (68%) even for incomplete heat integration, without significative damage to the nickel catalyst employed during the start-up operation [140]. When operated with a noble-metal catalyst [141], the plant demonstrated flexible operation in a range of 20–100% of the design workload with a maximum plant efficiency of 75% and a production cost as low as 2.90 EUR/kg, which is advantageous compared to electrolysis and comparable to conventional SR. An economic analysis of biogas autothermal reforming for a 100 Nm³/h plant also evidenced the possibility of reaching costs of hydrogen as low as 2.50 EUR/kg after 0 years’ amortization [142].

Autothermal dry reforming has also recently received attention from researchers, particularly for biogas conversion, and a variety of catalysts and reactors have been investigated. Kelling et al. [143] proposed a multitubular ceramic counterflow reactor with easy scalability and increased capability of withstanding thermal stresses due to a high reaction temperature, but they used a noble-metal rhodium catalyst. Akri and colleagues prepared honeycomb monoliths of MgO-promoted Ni catalyst incorporated in illite clay and tested them for the reforming of a simulated biogas mixture [144]. The tested catalyst proved to be a cheap alternative to conventional wash-coated catalysts and displayed best performance for a 3% Mg loading and 8% Ni loading at 800 °C, with magnesium loading favoring Ni dispersion and preventing sintering. Nickel-upgraded slag oxide obtained from mining waste has been also positively investigated at 850 °C, obtaining a CH₄ conversion of 98% with 98.8% and 95.5% H₂ and CO yields and no observable loss of activity or coke deposition [145]. This catalyst was also found to be particularly resistant to hydrogen sulfide poisoning [146]. Rosha et al. [147] compared the catalytic performance of pure Ni nanoparticles in DR and ATDR and evidenced increased activity and lower carbon deposition (0.03% weight vs. 0.40% weight) in ATDR compared to DR. CoAl₂O₄-supported metal catalysts (Ni, Co, Rh and Ru) were tested in a cordierite monolith reactor [148]. The monolithic reactor performance compared favorably with the conventional Ni/Al₂O₃ fixed-bed reactor, with about 20% higher reaction rates and up to 40% higher methane conversion at the high flowrate, thanks to better distribution and greater availability of metal active sites due to the increased surface area of monoliths. Of the metal species investigated, rhodium displayed the highest catalytic activity, followed by nickel. As one of the limitations of ATDR is that addition of O₂ to the reaction mixture can lower CO₂ conversion, dual-site catalysts capable of improving CO₂ activation are of interest for the dry-reforming reaction: for example, addition of CeO₂ and ZrO₂ to a SiO₂ support for Ni catalysts was observed to remarkably include catalytic performance, providing CO₂ conversion close to the equilibrium value in a temperature range of 600–800 °C [149]. More recently, the effect of the addition of CeO₂ and ZrO₂ promoters to the alumina support for the Ni catalyst was also investigated, with cerium dioxide performing the best in terms of performance improvement [150]. Setting themselves apart from conventional thermochemical process, Fan et al. [151] proposed instead an innovative design for a novel Solid Oxide Fuel Cell reactor for the cogeneration of syngas and electrical power through ATDR. This process may be particularly advantageous, in that it does not require oxygen purification and avoids forming potentially dangerous CH₄/O₂ mixtures. The researchers demonstrated a stable operation for 120 h at 800 °C, with a hydrogen selectivity of 80% and stable 65% methane conversion.

Nickel-based catalysts are also the most common choice for the dual-reforming of methane [152], and have been recently reviewed [153,154]. Most recently, Jin et al. [155] demonstrated significantly improved catalytic performance of a non-stoichiometric cerium oxide-overcoated Ni/Al₂O₃ catalyst prepared through atomic layer deposition compared to an uncoated catalyst, with an increase in reaction for methane conversion from 61.9% for the uncoated catalyst to 87.2% for the coated sample at 800 °C. A bimetallic Ni-Co catalyst supported on cerium dioxide- and magnesia-promoted alumina was found to exhibit stable operation for 500 h with negligible carbon formation, 97% methane conversion and 89% CO₂ conversion at 850 °C [156]. As the 2-R process is highly endothermic, solar energy is potentially a green alternative for providing the required reaction heat: Storch et al. [157], however, evaluated that dual reforming of methane can only be effectively applied for a CO₂/CH₄ ratio up to 0.35. Thermodynamic analysis determined that for dual reforming optimal reaction conditions require atmospheric pressure, temperature above 800 °C, oxygen-to-carbon ratio greater than 1.1 and a carbon dioxide-to-steam ratio of 1.5–2 [158].

Catalysts for the methane tri-reforming process are also prevalently nickel-based [159,160], and a recent review on thermodynamic evaluation, operation conditions and reactor configuration has been provided by Soleimani and colleagues [161]. Kozonoe et al. [162] compared carbon nanotubes and silica support for Ni catalysts and found that while both supports showed stable activity, carbon nanotubes offered better methane and carbon dioxide conversions, and also greater selectivity for hydrogen formation at 750 °C, mainly attributed to lower carbon dioxide activation by silica due to the acidic nature of the support. The deactivation behavior of bimetallic molybdenum catalysts supported on lanthana-alumina and on niobia determined that only NiMo catalysts are active on a lanthana-alumina support, while PtMo is active on a niobia support [163]. On a lanthana-alumina support, the Pt bimetallic catalyst rapidly deactivated due to sintering, while the CoMo catalyst deactivated due to oxidation of the cobalt species both on the lanthana-alumina and the niobia support. The nickel bimetallic catalyst on the niobia support instead deactivated due to intense carbon deposition. It should be noted that while lanthana addition to the alumina support for nickel catalysts initially increases hydrogen yield, an excess of lanthana causes increased carbon deposition due to the formation of an inactive LaAlO₃ phase [164]. A study of catalytic activity of LaNiO₃ perovskite with partial Ni substitution by Ce or Sr dopants demonstrated that conversion of methane was similar in undoped and doped samples (75%) but the cerium-substituted catalyst displayed greater CO₂ conversion and lower hydrogen selectivity due to higher CO formation [165]. This can be attributed to the improved activation of carbon dioxide due to improved oxygen exchange by the Ce³⁺/Ce⁴⁺ easy interconversion. At the time of this review, reactor studies for the tri-reforming process are mostly theoretical. Jang and Han [166] proposed that a distributed feed of oxygen to the reformer reactor can offer better process control and flexibility by achieving better control of the temperature profile and reducing mass and heat-exchange limitations. Aboosadi et al. [167] produced a model for a slurry bubble column reactor with a NiO-Mg/Ce-ZrO₂/Al₂O₃ catalyst and obtained a 92% methane conversion with a 1.76 H₂-to-CO ratio in optimized conditions. The main advantage of the slurry bubble column reactor is the prevention of hot spot formation. A 2-D axisymmetric model was produced for estimating process optimization of a fixed-bed reactor for different final objectives: maximization of H₂/CO ratio and energetic efficiency (case 1), maximization of H₂/CO ratio and carbon dioxide conversion (case 2), and maximization of carbon dioxide conversion and energy efficiency (case 3) [168]. The decision variables selected were inlet temperature, oxygen–methane ratio and carbon dioxide–methane ratio. For case 1 the scenario requires an increase in the oxygen inlet, a decrease in the carbon dioxide fraction and a reduced inlet temperature. At optimized conditions, methane conversion reaches 60%, but carbon dioxide conversion is below 10%. For case 2, maximizing carbon dioxide conversion requires increasing the inlet temperature to favor the dry-reforming reaction and inhibit the water–gas shift, while also lowering the oxygen content in the feed. An increase in carbon dioxide conversion inevitably leads to the lowering of the hydrogen-to-carbon monoxide ratio, as more CO is produced. For case 3, if a compromise of 1% is accepted for the requirement for energetic efficiency, a carbon dioxide conversion of 26.2% and a methane conversion of 80% can be achieved with a H₂/CO ratio of 1.5. Finally, the substitution of conventional steam reforming with tri-reforming for syngas production in methanol synthesis was evaluated and it was concluded that improved methane conversion and reduced carbon dioxide emissions can be obtained with this approach [169].

In contrast with autothermal steam reforming, ATDR, 2-R and TR processes are still in the early stages of development and have not been applied yet on a commercial scale.

6. Thermo-Catalytic Decomposition

While for the steam/dry reforming and partial-oxidation reaction coke formation is an undesired parasitic reaction, the direct decomposition of methane (3) is another possible pathway for producing hydrogen and carbon. In fact, because of inert conditions, no carbon oxides are expected to be formed.

Reaction (3) is thermodynamically favored at a high temperature (>1300 °C) and low pressure (atmospheric) [170]. Thermo-catalytic decomposition (TCD) in the presence of suitable catalysts allows the process to be carried out at lower temperatures of 500 to 1000 °C [170,171] but with more limited CH₄ conversion, because of thermodynamic limits.

Methane TCD has been recently reviewed [172,173], also addressing catalytic solar-assisted decomposition, as well as the innovative CH₄ cracking in molten salts (e.g., NaCl, NaBr, Na₂CO₃-KCO₃, KCl-NiCl₂, and MnCl₂-KCl) [174,175,176].

As far as catalysts are concerned, metals like Ni, Co and Fe are active for lowering the process temperature down to 600 °C, but the deactivation by carbon deposition is even larger as the temperature decreases. Iron seems to be more resistant to deactivation, as well as the support playing a relevant role, with good interaction and dispersion of the metal particles on the support preventing the agglomeration and the sintering of the catalyst. Msheik et al. [172] also reviewed the use of carbonaceous catalysts (chars of various origins) that, despite lower effectiveness than metals, boast several advantages, such as lower cost, resistance to high temperature, safe storage, tolerance to impurities such as sulfur, no contamination of the carbon byproduct, generally no need for regeneration, additional self-catalysis, and mitigation of CO₂ emissions.

The influence of the addition of various transition metals (Cu, Cr, Co, Zn, and Mn) on the properties and efficiency of the NiO(50)/FeAl₂O₄ catalyst was investigated in a micro fixed-ed reactor [177]. The Mn-based catalyst was the most effective among others in terms of catalytic activity and stability, with production of high-purity hydrogen and nanostructured carbon particles, at a temperature of 700 °C with CH₄ conversion of 62.3%.

Owing to the allo-thermal character of TCD, fluidized bed technology was proposed and applied, thanks to the high heat-transfer coefficient, of up to 300 W m⁻²K⁻¹ [178]. Ni-Cu-Al oxide systems were synthesized and tested in an FB reactor for the methane decomposition yielding H₂ and carbon fibers, with methane conversion of 40% at 675 °C [179]. Ammendola et al. [180] developed different CuO/Al₂O₃ catalysts by co-precipitation and wet impregnation of supporting granules (300–400 μm), for fluidized-bed operation. The active phase and the best catalyst superficial catalyst composition was determined by TPR analysis, reporting an optimal composition for a copper–aluminum spinel. The applicability of catalysts in a bench-scale fluidized bed was investigated at 800 °C. The catalyst was in the class B of Geldart classification [181], thus well-suited for fluidization, and characterized by a limited attrition rate, i.e., 3.0 × 10^–5 %/min. A two-stage operation for H₂ production in FB was investigated [182], consisting of first operating TCD until a defined catalyst deactivation degree is approached, and then operating the fluidized bed as a combustor or a gasifier for catalyst regeneration by carbon removal. Three different strategies of catalyst regeneration, by air, CO₂ and steam have been assessed, along with the optimization of the times for decomposition and regeneration steps. The authors concluded that regeneration time, preferably conducted in air for kinetic reasons, is a compromise between reducing catalyst re-oxidation and removing large amounts of deposited carbon.

7. Discussion

Table 1. Classification of reviewed articles.

Process	Technology Readiness	Catalysts	CRM	Harmfulness	Refs.
Steam reforming	Commercial	Ni	N	Y	[39,40,41]
		Ni-Fe	N	Y	[53]
Dry reforming	Lab scale	Ni	N	Y	[58,59,60,61,62,63,64]
		Co	Y	Y	[65,66,67,68]
		Ni-Co	Y	Y	[69,70]
		Ni-Cu	N	Y	[73,74,75]
		Ni-Fe	N	Y	[71,72]
		Co-Ce	Y	Y	[88]
		Co-Sm	Y	Y	[89,90]
		Ni-Mo	Y	Y	[91]
		Perovskite *	Y	Y-N	[80,81,82,83,84]
Partial oxidation	Pilot–Commercial Lab scale	Ni	N	Y	[91,92,93,94,95]
		Co	Y	Y	[96,97,98]
		Ni-Co	Y	Y	[99,113]
		Ni-Cu	Y	Y	[102,103]
		Ni-Ce	Y	Y	[102,104]
		Cu-Ce	Y	N	[115]
		Perovskite *	Y	Y-N	[105,106,107,108,109,110,111,112,113,114,115,116,117,118,119]
		MoP	Y	N	[120]
Thermo-catalytic decomposition	Lab scale	Ni-Mn	Y	Y	[177]
		Ni-Cu-Al	Y	Y	[179]
		Cu	Y	N	[180]

* Variable structure.

reports the classification of the reviewed articles. Critical raw materials (CRM) rely on the recent European Commission Communication of 2023 [183], including 34 elements. CRMs have been selected when an element reaches or exceeds the thresholds for both economic importance and supply risk for European countries, the threshold for supply risk being lower than that for economic importance. The harmfulness is instead based on the safety-data sheets of the individual compounds.

Overall, the level of technological readiness is advanced for steam reforming and good for partial oxidation, whilst the other two processes (DR and TCD) are still to be implemented at an industrial level. The studies of the catalysts mainly concern laboratory characterization of the developed materials, which are often tested with very small-scale instrumentation (TGA, TPR, etc.) and for limited catalyst time on stream (no more than 200 h). So far, the optimization of catalysts and supports has provided many novel solutions, but the issue of ensuring a good compromise between catalyst lifetime and activity has yet to be solved.

Table 2. Main advantages and disadvantages of methane-conversion technologies.

Process	Advantages	Disadvantages	Refs.
SR	High H2/CO ratio (=3) Mature technology	Excess high-pressure steam needed (steam/C: 2–5; 3) Endothermic reaction High CO2 emission High capital costs	[49,50,51,52,53]
DR	CO2 utilization Applicable to biogas without previous separation	Low H2/CO ratio (=1) RWGS parasitic reaction can decrease the syngas ratio below 1 Strong tendency for carbon deposition Endothermic reaction High reaction temperature needed (risk of catalyst sintering)	[68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,88,89,90,91,92,93,94,95,96]
PO	Exothermic process More compact reactors, thanks to fast kinetics Moderate syngas H2/CO ratio (=2) Non-catalytic operation is possible	Pure O2 is needed (expensive ASU unit) Difficult temperature and selectivity control Risk of explosion (CH4-O2 mixture) Coking can deactivate catalyst	[103,104,105,106,107,108,109,110,111,112,113,114,115,116,117,118,119,120]
ATSR	Mature technology Compact reactors Increased energy efficiency Reduced capital costs compared to SR Lower steam and oxygen requirements compared to SR and PO Lower operating cost compared to SR	Pure O2 needed Risk of explosion (CH4-O2 mixture) Coking and sintering can deactivate catalyst Low H2 yield compared to SR	[130,131,132,133,134,135]
ATDR	CO2 utilization Applicable to biogas without prior separation Autothermal process Lower oxygen and no steam requirement	Pure O2 needed Risk of explosion (CH4-O2 mixture) Coking and sintering can deactivate catalyst	[144,145,146,147,148,149,150]
2-R	CO2 utilization Applicable to biogas without prior separation Tunable H2/CO ratio Lower carbon deposition compared to pure SR and DR	Strongly endothermic process Coking and sintering can deactivate catalyst	[152,153,154,155,156]
TR	CO2 utilization Applicable to biogas without prior separation Tunable H2/CO ratio Presence of oxygen lowers endothermicity Lower carbon deposition compared to pure SR, DR and PO	Pure O2 needed Risk of explosion (CH4-O2 mixture) Coking and sintering can deactivate catalyst Difficult management of competitive oxidation–reforming reaction	[159,160,162,163,165]
TCD	Zero or near-zero carbon dioxide emission (no CO2 formation) Highest yield of hydrogen (no CO formation) Less complex separation of produced hydrogen (easy separation from solid carbon) Carbon can be produced in value-added form (nanotubes, nanosheets, etc.) No oxygen or steam requirement	Endothermic reaction Coke formation leads to difficult continuous operation due to catalyst deactivation and reactor clogging Harsh reaction conditions cause problems for durability of reactor materials	[177,179,180]

reports the main advantages and disadvantages of each catalytic process, while

Table 3. Literature data for process efficiency, operation conditions (T and P) and H₂-production cost for catalytic hydrogen production from methane. (CCS is short for Carbon Capture and Storage, technologies for CO₂ emission abatement [184], N.A. indicates data that were not available in the literature).

Process	Efficiency (%)	T (°C)	P (Bar)	H2 Cost
SR	70–85 (no CCS) [5,9,14,35,124] 60 (with CCS) [185]	650–1100 [44]	3–25 [44]	0.9–1.8 USD/Kg [14] 2.08 USD/Kg (no CCS)–2.27 USD/Kg (with CCS) [34] 1.83–2.35 USD/Kg [35] 1.54–2.30 USD/Kg [9]
DR	76 (estimated for biogas reforming) [186] 59 (estimated for power and H2 cogeneration) [187]	600–1000 [53]	N.A.	0.15 EUR/Nm3 (power and H2 co-generation) [187] 2.38–3.27 USD/Kg (coke-oven-gas reforming) [188] 1.07–1.32 USD/Kg (no CCS)—1.91 USD/kg (with CCS) [189]
PO	N.A.*	800–900 [100]	N.A. *	N.A.
ATSR	90 [14] 60–75 [190]	800–1200 [131]	1–30 [131]	1.48 USD/Kg (with CCS) [34]
ATDR	N.A.	N.A.	N.A.	N.A.
2-R	82 (natural and biogas co-reforming) [188]	N.A.	N.A.	N.A.
TR	N.A.	N.A.	N.A.	N.A.
TCD	58 [187,191]	500–1000 [162,163]	1 [162]	2 USD/Kg [8] 2.55–5 USD/Kg [12] 3.53–3.82 USD/kg [192] 1.72 USD/kg

* Data available in literature refer to non-catalytic process.

summarizes operation conditions, process efficiency and hydrogen production cost.

As far as issues of economic and environmental impact are concerned, almost all the elements used belong to the CRM category, except for Ni, which, however, is considered harmful. Iron, manganese, and copper should be considered of great relevance for future application, both being easily available also from waste recycling.

8. Novel Perspectives

8.1. Chemical Looping

In recent years, a lot of attention has been given to the study of transport properties of solid compounds, the chemical looping technologies and their mutual integration, to overcome the limits of traditional processes [193].

In a looping configuration the reaction scheme can be split into two separate steps. allowing for a higher flexibility with respect to that achievable in the traditional processes. Figure 2 provides some examples of typical looping configurations [194].

In a looping process, in the first step the fuel (methane in the reported example) is selectively oxidized to syngas or hydrogen (7), while in the second step the OC is re-oxidized through reaction with H₂O (8), CO₂ (9) or oxygen/air (10).

Step 1

$$\mathsf{\delta }\mathrm{C}{\mathrm{H}}_4+{\mathrm{M}\mathrm{e}\mathrm{O}}_{\mathrm{x}}⟶\mathsf{\delta }\mathrm{C}\mathrm{O}+2\mathsf{\delta }{\mathrm{H}}_2+{\mathrm{M}\mathrm{e}\mathrm{O}}_{\mathrm{x}-\mathsf{\delta }}$$

(7)

Step 2

$${\mathrm{M}\mathrm{e}\mathrm{O}}_{\mathrm{x}-\mathsf{\delta }}+\mathsf{\delta }{\mathrm{H}}_2\mathrm{O}⟶2\mathsf{\delta }{\mathrm{H}}_2+{\mathrm{M}\mathrm{e}\mathrm{O}}_{\mathrm{x}}$$

(8)

$${\mathrm{M}\mathrm{e}\mathrm{O}}_{\mathrm{x}-\mathsf{\delta }}+\mathsf{\delta }{\mathrm{C}\mathrm{O}}_2⟶\mathsf{\delta }\mathrm{C}\mathrm{O}+{\mathrm{M}\mathrm{e}\mathrm{O}}_{\mathrm{x}}$$

(9)

$${\mathrm{M}\mathrm{e}\mathrm{O}}_{\mathrm{x}-\mathsf{\delta }}+\frac{\mathsf{\delta }}{2}{\mathrm{O}}_2⟶{\mathrm{M}\mathrm{e}\mathrm{O}}_{\mathrm{x}}$$

(10)

The chemical looping operation can be achieved either in a continuous circulating fluidized-bed configuration or through a semi-continuous process where the looping is achieved by the switching of the gaseous atmosphere over a fixed or fluidized bed [195]. In all cases, the product streams are inherently separated, allowing greater process flexibility, as the CO stream produced in oxidation can be either used alone or mixed with the stream of syngas produced during reduction to correct the H₂/CO ratio.

The use of a looping configuration also allows for the decrease in the likelihood of undesired reactions. An example of this can be provided by the chemical looping dry reforming of methane (CL-DRM). Here, the use of the carrier for the transfer of oxygen allows the avoidance of the effect of the reverse water–gas shift reaction, as the reactant CO₂ and the product H₂ do not come in direct contact. CL is also beneficial for partial oxidation, as the inherent separation between reactants reduces the risks of explosion and removes the need of an air-separation unit to produce pure oxygen (Figure 2A).

Among a variety of options, thermochemical water splitting brings high conversion of hydrogen at high purity and if it is practiced in association with the utilization of a fuel (steam-reforming process), the temperature and energy demand considerably decrease, thanks to the enhanced reactivity of the oxygen carrier in presence of a reducing agent (Figure 2B). In this respect, selection of proper oxygen carrier material is essential for chemical looping applications: key parameters for an oxygen carrier are fast redox kinetics, high oxygen-storage capacity, resistance to sintering and carbon deposition, and thermal and mechanical stability (for fluidized- and circulating-bed applications) of the structure over repeated cycles [196]. Several different materials have been evaluated for CL-DRM: iron-based OCs have the advantages of relatively low costs and environmental safety and they can be successfully re-oxidized by carbon dioxide displaying high oxygen-storage capacity, but they tend to suffer from severe sintering problems and generally slow reaction kinetics with methane [193,197]; Chuayboon et al. [198] tested the performance of an iron oxide OC during a methane reforming–water splitting process heated by solar radiation, and found that the material performance depended strongly on temperature and degraded rapidly over only five cycles. As CO₂ is a much weaker oxidant compared to water, these limitations would presumably become more pronounced for a CL-DRM process. On the other hand, Ni-based oxygen carriers display high methane conversion, but they suffer from high carbon deposition, higher costs, and toxicity, and cannot be re-oxidized directly by CO₂ [67,193,199]. Mixed Ni-Fe materials have also been investigated [197,200], in an attempt to combine the properties of both metals to overcome their respective limitations, and perovskite oxides have also received attention [201].

Noble-metal catalysts (Rh, Ru, Pd, Pt, and Ir) show high catalytic activity and stability for dry reforming, with scarce carbon deposition, but they can hardly find large-scale application, due to their high price. On the contrary, transition metals such as Ni-, Co-, Cu-, Mn-, and Fe-based catalysts, are highly promising, due to their abundance, low cost, and good thermodynamic properties for reforming processes [202,203,204].

Solid transition-metal (Ni, Co, Fe, Cu)-based catalysts interact with hydrocarbons, breaking the C-H bond and favoring the decomposition at a lower temperature, being suitable for the direct production of hydrogen if mixed in alloy and applied molten [46] (Figure 2C). In this case, the material acts as carbon carrier and pure streams of H₂ and CO are produced [205].

Cu-based oxygen carriers show suitable redox behavior to ensure a good oxygen transfer capacity, along with good reactivity for the reforming reaction. They are also capable of directly releasing oxygen by thermal decomposition at medium-high temperatures (800–1000 °C) [206]. Also in favor of these systems is their low price and limited toxicity, but they can give rise to sintering and melting even at temperatures below 800 °C and be subjected to friction, due to limited mechanical strength. These limitations are partially overcome by employing support, for example, with γ-Al₂O₃, MgO, SiO₂, CeO₂, and ZrO₂, or promoters such as alkaline and alkaline earth metals, as well as rare earth metals, which remarkably improve catalytic performance and reduce coke formation, especially in the dry-reforming processes.

Among the various available oxygen carrier materials, cerium dioxide (CeO₂) shows some of the most promising properties for methane-reforming processes, both as support for Fe [207,208], Ni [203,209] or Co [204] or applied individually without inclusion of supports; therefore, its use has been analyzed in more detail in this review.

Cerium dioxide, despite being a lanthanide oxide, is rather abundant in the Earth’s crust, being comparable to copper [210]. The use of CeO₂ as an oxygen carrier for methane reforming was first proposed by Otsuka et al. [211], who studied the behavior of this oxygen carrier in a temperature range between 570 and 750 °C. They demonstrated that cerium dioxide could selectively produce syngas with a H₂/CO ratio of 2, with limited production of water and CO₂ at the beginning of the reduction step, and that CO₂ can be used effectively for regeneration. Cerium dioxide, in fact, displays very high oxygen-storage capacities associated with the Ce⁴⁺/Ce³⁺ redox pair, its reduction leading to the formation of a continuous range of non-stoichiometric species (CeO₂₋x, with 0 < x ≤ 0.5) that nonetheless preserve the original cubic fluorite structure [212,213]. Thermal stability of cerium dioxide may be improved by doping with tetravalent ions, for example Zr⁴⁺, which also has the advantage of improving oxygen storage and exchange by facilitating the migration of bulk oxygen species during reduction [213,214].

CeO₂ reduction and oxidation are easily reversible [215,216,217] and the material also displays fast internal oxygen diffusion [218] and high resistance to coke formation, being able to catalyze the combustion of deposited carbon [219,220].

Temperature-programmed reduction experiments with methane as the reducing agent [216,220,221,222] show that cerium dioxide is active for methane partial oxidation at temperatures above 700 °C, and the selectivity towards partial oxidation is high, with limited total oxidation of methane occurring at the beginning of reduction when more surface oxygen species are available. A limited influence of the methane cracking reaction was observed in all the examined studies, occurring once a high oxygen depletion of ceria was reached.

Chuayboon et al. [198] compared its performance with that achievable by employing iron oxides for methane reforming coupled with water splitting. The CeO₂ carrier was found to provide improved performances compared to iron oxides in terms of reactivity and stability over repeated cycles. Stable operation was also observed for 10 cycles at T = 1000 °C with stable syngas yield (5.67–6.80 $\frac{\mathrm{m}\mathrm{m}\mathrm{o}\mathrm{l}}{{\mathrm{g}}_{\mathrm{C}\mathrm{e}{\mathrm{O}}_2}}$) and selectivity (96.5–98.2%) and a notable methane conversion (46.9–60.9%).

Methane reforming in looping, in presence of cerium dioxide, has been carried out employing solar reactors as a source of heat [192,200,201]. Nair and Abanades [214] tested the performances of three different cerium dioxide structures, consisting of unmodified commercial cerium powder, ceria prepared through a hydrothermal templating method and ceria prepared through a self-assembly method over two/three isothermal reduction/oxidation cycles at 1000 °C. TGA experiments on these materials revealed that all of them could be almost completely re-oxidized by CO₂, but unmodified commercial dioxide showed structural instability, demonstrating rapid sintering and a decrease in CO yield during CO₂ splitting (from 0.365 $\frac{\mathrm{m}\mathrm{o}{\mathrm{l}}_{\mathrm{C}\mathrm{O}}}{\mathrm{m}\mathrm{o}{\mathrm{l}}_{\mathrm{C}\mathrm{e}{\mathrm{O}}_2}}$ for the first cycle to 0.322 $\frac{\mathrm{m}\mathrm{o}{\mathrm{l}}_{\mathrm{C}\mathrm{O}}}{\mathrm{m}\mathrm{o}{\mathrm{l}}_{\mathrm{C}\mathrm{e}{\mathrm{O}}_2}}$ for the second cycle). The self-assembled and hydrothermally prepared samples overall displayed more stable performances (with CO yields of 0.360–0374 $\frac{\mathrm{m}\mathrm{o}{\mathrm{l}}_{\mathrm{C}\mathrm{O}}}{\mathrm{m}\mathrm{o}{\mathrm{l}}_{\mathrm{C}\mathrm{e}{\mathrm{O}}_2}}$ and 0.341–0.347 $\frac{\mathrm{m}\mathrm{o}{\mathrm{l}}_{\mathrm{C}\mathrm{O}}}{\mathrm{m}\mathrm{o}{\mathrm{l}}_{\mathrm{C}\mathrm{e}{\mathrm{O}}_2}}$ respectively), while also displaying overall faster kinetics. Oxidation with carbon dioxide, as expected, was found to be slower than oxidation with water.

Warren et al. [223] instead investigated the performance of commercial CeO₂ powder, evaluating the effect of different operating conditions: temperature, flow rates and initial extent of CeO₂ oxygen non-stoichiometry. Interestingly, they observed that starting the reduction phase over oxygen-deficient cerium dioxide leads to higher selectivity toward syngas formation, lowering CO₂, H₂O and coke formation. It also allows for the achievement of faster oxidation during the regeneration phase with CO₂. A performance over 10 redox cycles at 1170 °C with a 4 min reduction time and 2 min oxidation resulted in a stable performance of the material, with a significant averaged conversion of reactants (69% for CH₄, and of 88% for CO₂), and a high selectivity towards H₂ and CO (99% and 93%, respectively). Also, carbon formation was found to be absent after a proper selection of the reaction time.

Finally, Chuayboon et al. [224] compared commercial cerium powder with structured porous foams and mixed powders of CeO₂ and Al₂O₃. Oxidation with CO₂ was performed for CeO₂ and Al₂O₃ mixed powders at different temperatures (950–1050 °C), confirming the increase in syngas yield and reaction kinetics during reforming with increasing temperature. Temperature was found to have no relevant effect on oxidation. Overall maximum yields of H₂ and CO were found to be of 5.11 and 5.04 $\frac{\mathrm{m}\mathrm{m}\mathrm{o}\mathrm{l}}{{\mathrm{g}}_{\mathrm{C}\mathrm{e}{\mathrm{O}}_2}}$ at 1050 °C, also considering methane cracking and coke combustion side reactions. The shape of the material was found to have no relevant impact on syngas yield, as CeO₂ powder and foams showed no relevant differences.

The thermodynamics of this reforming process was also investigated [225], and it was demonstrated that cerium dioxide can be efficiently used for isothermal operation at a temperature of 950 °C. The need for reaction temperatures above 900 °C was also determined [226], utilizing ASPEN Plus as a software simulator for thermodynamic predictions.

The mechanism of reaction of CeO₂ with methane and carbon dioxide was investigated by Warren and Scheffe [227], who found that partial oxidation of methane over cerium oxide follows different reaction mechanisms according to the degree of reduction and to the presence of oxygen vacancies in the structure of the material, with different kinetic limitations for low- or high-oxygen-depleted ceria. Interestingly, they found that hydrogen presence inhibited the reaction rate during reduction only for high-oxygen non-stoichiometries cerium dioxide, and that reaction rates seem not to be limited by bulk diffusion [114]. Both higher temperatures and higher methane partial pressure were found to increase reaction kinetics.

The kinetics of cerium dioxide oxidation with CO₂ were instead studied by Ackermann et al. [228], who found that the rate of reaction is suppressed for high-oxygen non-stoichiometries in a temperature range between 420 and 1000 °C. Welte et al. [229] performed an interesting study of CeO₂ reduction in a moving-bed particle reactor. They observed that a co-current flow of CH₄ and CeO₂ is needed to avoid further oxidation of syngas with cerium dioxide, which may also be possible near the exit of fixed-bed reactors where the methane concentration is lower than syngas. They also observed that carbon formation was suppressed at temperatures above 1300 °C.

In conclusion, this technique represents an interesting possibility for carrying out processes that require the supply of an oxidizing reagent without contamination with inert species (e.g., N₂), especially if conducted on a limited scale, for example, such as those situations in which the use of oxygen from air distillation is not economically convenient.

8.2. Electrically Asssited Reforming

Recently, the application of an electrical current/field to chemical reactors has attracted considerable interest, at least for lab-scale investigations. Compared to conventional purely thermo-catalytical processes, these techniques (e.g., microwaves, hot/cold plasma) enhance the performance of the reforming process allowing for catalyst activity at a reduced temperature.

The effect of support structure on Ni/γAl₂O₃ catalysts in plasma-assisted catalytic reforming was evaluated [209] and it was concluded that Al₂O₃ nanosheets are the best support compared to nanorods or spherical flowers, displaying smaller and better-dispersed Ni particles and achieving the best plasma–catalyst synergy. thanks to the high number of exposed Ni sites, improving charge transfer [230]. Higher than 90% methane conversion without external heating was achieved in their experiment.

Microwave (MW)-assisted methane DR was investigated [231] in a custom-designed reactor housed inside an MW apparatus. The authors reported the advantage of process intensification and reaction stability in their process compared to the conventional heated reactor, probably for the generation of local hotspots (micro-plasma) in the microstructure of the Ni-La catalyst. Similarly, catalytic dry reforming of CH₄ was experimentally tested in a plasma micro-reactor using noble metals (Pt, Ag) or Ni [232]. Hot plasma was generated by an AC power source at 10 kHz and tunable voltage of 0–30 kV. The highest conversion of reactants (CO₂ and CH₄), up to 27.6%, was obtained for the Ag catalyst. The authors reported the need to achieve a good understanding of the mechanism of plasma activation of the catalysts.

Microwave reforming was also tested for biogas, investigating the effect of the presence of nitrogen and oxygen in the reaction mixture, and it was found that their presence may improve methane conversion [233].

Gray et al. [234] applied a positive electric field on a pure metallic Ni-foam-catalyst bed and observed that the positive field promotes the oxidation of the Ni catalyst and increases methane activation and conversion while preventing coking. They attributed the improved performance to the water adhesion to the catalyst surface and water activation in the bulk catalyst, both promoted in the presence of an electric field. Oxide formation, as well as the presence of the electric field itself, also inhibits carbon polymerization.

Electrically assisted processes represent a very interesting challenge for exploiting electricity produced in excess during peak periods, especially of solar origin, and thermo-chemically converting methane or biomethane into hydrogen with high-scale flexibility.

8.3. Multistage Processes

The implementation of complex process schemes may override limited performances in heat/mass transfer or catalysts, according to the more general principles of process intensification. For instance, a multistage FB plant can largely reduce the typical back-mixing behavior of such technology [235], leading to well-controlled residence times and product yields. A multistage strategy can also contribute to complex reaction decoupling, when different reaction paths are possible, by selecting at each stage optimal operating conditions and a suitable catalyst [236].

Hydrogen production via steam reforming in an interconnected fluidized bed has been modelled [237]. Although the authors were mainly focused on a good estimation of fluid-dynamic behavior and drag mode, the investigation lays the basis for a possible application of a multistage plant on an industrial scale.

A two-zone fluidized-bed-reactor coupling permselective Pd/Ag membranes was experimentally investigated at lab-scale [238] to intensify the process for producing pure hydrogen by biogas DR. Compared to a single-stage process, the 2-zone FB reactor is more stable, thanks to the regeneration of the catalyst in the lower section of the system. Coupling a membrane leads to improvements in H₂ yield and purity. Along these lines, a 2-zone FB reactor and 2-zone FB reactor with membranes were mathematically modelled [239] and the model predicted well the in situ regeneration of the catalyst in the lower zone of the fluidized bed.

For gas mixture purification, a multistage FB can effectively perform as an absorber by strongly limiting bypass phenomena and easily tackling issues connected to temperature control. This option was successfully investigated for sour gas purification by chemisorption [240] and would be applied to syngas treatment by tuning the adsorbent characteristics, e.g., for CO removal [241].

In conclusion, the multistage reactor configurations allow for the intensification of the plant productivity and for overcoming the fluid dynamic limits due to the back-mixing of the reactants and products, for example, in multiphase systems.

8.4. Hydrogen Purification by Membranes

Membrane-based processes for gas separation or purification are attractive for their low environmental footprint, intrinsic safety, and modular nature with respect to traditional technologies, essentially based on chemical absorption [242]. Hydrogen separation can be accomplished at low temperatures by polymeric membranes [243,244], whose performance, limits, and perspectives were recently reviewed [245]. In glassy polymeric membranes (e.g., cellulose acetate and polysulfone), the selectivity towards N₂ and CH₄ is high enough (50–100), but it drastically decays towards CO₂. Furthermore, the use at high temperatures is prevented by the decomposition of the polymers at medium-high temperatures.

One of the major concerns about the use of membrane technology for hydrogen separation is mainly related to the intrinsic limitation in meeting the requested purity of the hydrogen (well above 99%) and recovery (around 60%). Gas separation by membrane, indeed, can hardly reach those purity thresholds with an economically sustainable process (i.e., a limited number of membrane stages).

For this reason, the coupled membrane and PSA process was experimentally investigated [246] for recovery of H₂ from industrial streams (syngas, coke gas, petrochemical, etc.). A major part of the impurities (CO, N₂, etc.) was removed at the membrane stage, whilst CO₂ and H₂ were separated by PSA, and the hydrogen purity could thus be enhanced as high as 99.9 or 99.99%, as requested. The proposed process exhibited enhanced performance in comparison to the purely membrane step.

Membranes for high-temperature (higher than 600 °C) separation of H₂ via protonic exchange are based on a thin Pd film deposited over microporous ceramic or metallic support, achieving H₂/N₂ selectivity up to 10⁴ [247]. Unfortunately, the presence of CO in the gas mixture gives rise to poisoning and the lowering of hydrogen permeability [248]. Catalano et al. [249] theoretically investigated such a phenomenon, assuming a CO competition with H₂ in the covering of the active sites by adsorption, which is a preliminary stage for permeation through the metal film. Their model can estimate the effect of temperature, thickness and CO molar fraction on permeation rate in Pd membranes.

Ceramic membranes can represent a valid alternative compared to the use of precious metals for the separation of hydrogen at high temperatures, allowing for the configuring of a reforming reactor that is improved by the separation of the product. Microporous membranes are already commercially available, though with limited selectivity. Hybrid silica membranes were modified by chemical vapor infiltration [250] to increase the hydrogen selectivity towards CO₂ and CH₄, achieving values of 61.3 and 460.5 at 250 °C, respectively.

A novel proton–ceramic membrane (MPEC) has been developed and tested at 750 °C [251] for hydrogen separation from syngas. Stable permeation was measured in an asymmetric configuration for 100 h, using a sweep gas downstream. Despite the rather complex protocol of synthesis, the membrane cost could result in being largely lower than that of the Pd-based membranes, with respect to which they have similar high selectivity.

Residual CO₂ in syngas can be separated at low temperatures by well-established methods based on adsorption [252] or membranes [253]. In this respect, an innovative membrane-based module for the gas separation has been developed and tested at lab-scale by Minelli et al. [254]. The process was based on alternating steps of permeation and adsorption, resulting in an improvement in the effective separation factor. The characteristic times represent a further operation variable of the process that can purposely change the average purity or flow rate of the permeates, without modifying the membrane configuration. This device could find application in the separation of CO₂ from syngas, if applied downstream of a reformer.

In summary, the syngas purification can be carried out either by removing the non-hydrogen species or by separating H₂ through proton-transport membranes. The latter represent an expensive alternative because they are based on noble metals, but the recent development of ceramic membranes may represent a promising alternative for the in-line H₂ separation during the methane and bio-methane reforming.

8.5. Catalyst Patterning

In conventional tubular autothermal reformers, a partial oxidation reaction is carried out first at the inlet of the reactor and is then followed by the steam-reforming or dry-reforming reaction [136]. As mentioned previously in Section 5, the highly exothermic and fast oxidation reaction at the inlet of the reactor can cause the temperature of this zone to rapidly increase and creates a non-homogeneous temperature profile which hinders reactant conversion and catalyst stability by forming hot spots and cold zones [139,255]. Non-isothermicity is also a problem for conventional packed-bed steam-reforming reactors, due to heat-exchange limitations [256,257], and PO, DR and the dual-reforming process can be expected to suffer from similar problems. While fluidized beds can offer a solution in terms of heat- and mass-transfer optimization [258], they also add complexity to reaction operation and catalyst selection compared to fixed-bed reactors, as factors such as the control of a correct fluidization regime and attrition resistance of catalyst and reactor components have to be considered [259,260]. Therefore, improvement in the performance of fixed-bed reactors is particularly interesting for future developments. Patterned fixed beds [261,262], which is to say beds where layers of different catalytic activity are regularly distributed inside the reactor, may offer a simple solution to improving the performance of reforming reactors, with different possible configurations that have recently been evaluated. For steam reformers, axially alternating layers of catalyst material with metallic foam layers [263] can potentially greatly decreases the temperature gradient inside the reformer reactor. Even though a patterned-bed reactor intuitively contains less catalyst compared to a conventional packed bed of the same size, simulation results suggest that even halving the mass of catalyst reduced the methane conversion rate by only 15%, suggesting the possibility of obtaining high conversion even with a lower catalyst load. Radial patterning of the catalyst, as opposed to axial layering, appears to be a more promising solution, as numerical investigation indicates a capability of achieving an increase in methane conversion while lowering the required catalyst mass (10.6% increase in conversion with a 26.16% decrease in catalyst load) [264]. Crucial aspects of optimization of the process consist in determining the number of layers, their size and the catalyst-to-gap ratio [265]. For an optimized radially patterned biogas dual reformer, a reduction in catalyst load of 41% with no adverse effect on reactant conversion was predicted, displaying great benefits of this configuration also in reducing cost, compared to conventional reformers [266]. Lee et al. [267] observed that optimized axial alternation of layers of a steam-reforming and dry-reforming catalyst can greatly help to reduce the carbon deposition problem associated with the dry- or dual-reforming process for biogas, while at the same time displaying great potential for reduction in capital and energy costs of the reformer.

An interesting alternative to removing temperature and mass-transfer gradients is to pattern the distribution of the catalyst particle diameter in multichannel reactors [268]. Patterning alternating preheated catalyst sections has also been evaluated [269].

The patterning of catalysts also proved beneficial in simulation work on ATSR process, with radially patterned reactors offering best performance enhancement with reduced catalyst load and oxygen demand [270,271,272]. Interestingly, patterning of catalysts can also potentially be achieved in novel and compact plate reactors [273]. Finally, a patterned reactor could also be operated with the use of membranes, either for separation of reaction zones in dual reforming [274] or to improve hydrogen yield and separation in steam reforming [275].

In conclusion, the patterning of catalysts makes it possible to considerably reduce the non-isothermal effects due to the succession of different reactions during the CH₄ reforming. The development of such catalysts represents a very interesting challenge for improving the performance and efficiency of fixed-bed catalytic reactors.

9. Conclusions

Steam reforming is currently the most diffused process for H₂ production on an industrial scale and has the lowest price (between 1.0 and 2.0 USD/kg), making use of catalysts essentially based on nickel. The current research is principally oriented towards the development of new catalysts, with the aim of reducing the negative impact of Ni compounds with respect to the environment and health. Iron, in combination with Ni, appears to be an effective element for improving catalyst performance and safety.

Due to the increased relevancy of carbon capture and decarbonization, dry reforming of hydrocarbons is a suitable option for conversion of CO₂ in syngas. The process is highly allothermal, and needs an energy supply of heat or electricity. In this respect, microwave- or plasma-assisted reactors deserve a large amount of interest for future development. While substantial work has been carried out on catalysts, the process is yet to achieve sufficient maturity for industrial application.

Partial oxidation of methane is an exothermic reaction, where the use of a catalyst allows for a lower reaction temperature and increased selectivity. On an industrial scale, the process requires pure oxygen and may experience problems related to temperature run-off. Very interestingly, the application of chemical looping schemes with the use of an oxygen carrier can avoid the use of an air-separation unit, moving this process to a lower and more-distributed industrial scale for hydrogen generation, with high-scale flexibility.

Thermo-catalytic decomposition of methane is a further process for producing H₂ and carbon, with a slightly endothermic reaction. The technology is easy to implement, also at small scale, using catalysts based on Ni or Cu, the latter being non- toxic. The main issue is catalyst regeneration because of deactivation by carbon deposition, an aspect that poses challenges in process optimization and intensification.

For all the above reported processes, the implementation of more complex reactor schemes (multistage, membrane/adsorption coupling, chemical looping, etc.) could improve the process efficiency, scalability, and flexibility. The utilization of fluidized-bed technology would be helpful for some related catalyst/carrier issues, such as regeneration, circulation, and discharging, as well as for largely improving the heat-transfer rate, and also for the utilization of an external energy source (e.g., solar). In such a context, the application of catalyst patterning is another effective way to tackle problems of temperature distribution and instabilities along the reformer.