Recent Advances in Methanol Steam Reforming Catalysts for Hydrogen Production

Abstract

The pursuit of carbon neutrality has accelerated advancements in sustainable hydrogen production and storage methods, increasing the importance of methanol steam reforming (MSR) technology. Catalysts are central to MSR technology and are primarily classified into copper-based and noble metal-based catalysts. This review begins with an examination of the active components of these catalysts, tracing the evolution of the understanding of active sites over the past four decades. It then explores the roles of various supports and promoters, along with mechanisms of catalyst deactivation. To address the diverse perspectives on the MSR reaction mechanism, the existing research is systematically organized and synthesized, providing a detailed account of the reaction mechanisms associated with both catalyst types. The discussion concludes with a forward-looking perspective on MSR catalyst development, emphasizing strategies such as anti-sintering methods for copper-based catalysts, approaches to reduce byproduct formation in palladium-based catalysts, comprehensive research methodologies for MSR mechanisms, and efforts to enhance atomic utilization efficiency.

1. Introduction

The extensive use of non-renewable fossil fuels such as coal, oil, and natural gas has caused the world to face significant challenges, including energy crises and climate change. To overcome such challenges, hydrogen, as an ideal renewable and clean energy source, has become a powerful alternative, although challenges in storage and transport remain [1,2,3]. Therefore, hydrogen production technologies at end-use locations, to sustainably supply this clean energy source with high calorific value, have become a major trend. The emergence of liquid organic hydrogen carriers (LOHCs), which can produce hydrogen by activating certain chemical bonds in the presence of catalysts, offers unlimited possibilities for in situ H₂ generation [4]. Methanol is an ideal LOHC, because it has a high hydrogen storage density of 99 kg m⁻³ and is inexpensive and widely available, being derived from biomass or CO₂ hydrogenation [2]. Hydrogen production from methanol can be carried out through the thermocatalytic process of methanol steam reforming (MSR), with the overall reaction represented by Equation (1) [5]. This includes two side reactions: methanol decomposition (MD, Equation (2)) and water-gas shift (WGS, Equation (3)).

CH₃OH + H₂O ↔ CO₂ + 3H₂   ∆H^θ = +49.2 kJ/mol

(1)

CH₃OH ↔ CO + 2H₂   ∆H^θ = +91.0 kJ/mol

(2)

CO + H₂O ↔ CO₂ + H₂   ∆H^θ = −41.1 kJ/mol

(3)

In practical applications, polymer electrolyte membrane fuel cells (PEMFCs) running on hydrogen require a CO-free hydrogen source to avoid poisoning of the anode catalysts of the cells [6,7]. In addition, a continuous and stable hydrogen source is a necessary guarantee. Therefore, developing MSR catalysts that exhibit high methanol conversion and hydrogen yield, low CO selectivity, and long-term stability is a critical and challenging task. To date, different kinds of MSR catalysts have been developed, among which copper-based (represented by Cu/ZnO/Al₂O₃) and palladium-based (represented by Pd/ZnO) catalysts predominate. These two types of catalysts have been proven to exhibit reliability and high efficiency under demanding conditions, making them the catalysts of choice for industrial hydrogen production via MSR. Furthermore, packed-bed reactors are the most commonly used [8], and reaction conditions such as the water-to-methanol ratio, reaction temperature, and weight hourly space velocity (WHSV) have a direct impact on the kinetics of the MSR reaction [9,10]. Generally, catalytic activity increases with higher water-to-methanol ratios and rising reaction temperature, up to a certain limit. The WHSV of methanol is usually in the range of 1–10 h⁻¹. Therefore, when evaluating the performance of MSR catalytic materials, it is essential to consider the reaction conditions to ensure a comprehensive assessment.

The research progress on copper-based and noble metal-based catalysts applied in methanol steam reforming for hydrogen production is examined in this paper, with a focus on developments from the late 20th century to the most recent findings (Figure 1). The analysis includes a detailed exploration of the catalytic components of these two types of catalysts, specifically the active metals, supports, and promoters, with an emphasis on their structure–activity relationships. Following this, the study delves into the reaction mechanisms and deactivation processes involved. The performance of both catalyst types is assessed, highlighting key scientific challenges that persist in the field. An outlook on potential strategies for overcoming these challenges is also provided, aiming to contribute to a deeper understanding of MSR catalysts.

2. Copper-Based Catalysts

The development of copper-based catalysts for methanol steam reforming can be traced back to the late 20th century, when Takezawa et al. [11] discovered that metallic copper was active in this reaction. However, due to the complexity of the MSR process, the rich redox chemistry of copper, and the limitations of catalyst characterization techniques at that time, fundamental scientific questions such as the intrinsic active sites and structure–activity relationships of copper-based catalysts in MSR were not fully resolved [12]. Over the past decade, with the gradual emergence of related research findings (

Table 1. Performance summary of representative copper-based catalysts for methanol steam reforming.

Catalysts	Temperature (°C)	CH3OH Conversion (%)	CO Selectivity (%)	H2 Yield (mmol g−1 h−1)	Reaction Conditions	Stability	Ref.
Cu5-Al	240	98	1.3 b	187.2	H2 pretreatment; feed rate = 0.048 mL h−1; H2O/CH3OH = 1.5	\	[13]
Cu/Cu(Al)Ox	240	99.5	1 b	398.88	H2 pretreatment; feed rate = 2.4 mL h−1; H2O/CH3OH = 2	240 °C, 100 h, 14% drop in CH3OH conversion	[12]
Cu/Al2O3	250	89.7	0.9 a	531.36	H2 pretreatment; WHSV = 10.56 h−1; H2O/CH3OH = 1	200 °C, 100 h, 10% drop in H2 production rate	[14]
CuZnAl	350	98	0	60.02	GHSV = 15,500 h−1; H2O/CH3OH = 2	\	[15]
CuZnO/γ-Al2O3/Al	275	100	3.34 a	3580	GHSV = 4000 mL g−1 h−1; H2O/CH3OH = 2	275 °C, 100 h, 10% drop in CH3OH conversion	[16]
Cu/ZnO/Al2O3	225	67	0.07 a	\	CH3OH/H2O/H2 pretreatment; WHSV = 6 h−1; H2O/CH3OH = 1.3	225 °C, 40 h, 10% drop in CH3OH conversion	[17]
Cu/ZnO/Al2O3	240	90	0	\	H2 pretreatment; GHSV = 10,000 cm3 g−1 h−1; H2O/CH3OH = 1.5	240 °C, 90 h, 30% drop in CH3OH conversion	[18]
Cu/SiO2	280	80	\	105	GHSV = 300 kg L−1 s−1; H2O/CH3OH = 1.5	\	[19]
Cu-MCM-41	250	72.3	0.8 b	\	H2 pretreatment; GHSV = 2838 h−1; H2O/CH3OH = 3	250 °C, 48 h, no drop	[20]
CeCuZn/CNTs	300	94.2	2.6 a	H2 yield = 98.2%	H2 pretreatment; WHSV = 7.5 h−1; H2O/CH3OH = 2	300 °C, 48 h, 7% drop in CH3OH conversion	[21]
Cu/Ce-Cu(BDC)	250	99	2 a	H2 yield = 97%	WHSV = 9.2 h−1; H2O/CH3OH = 2	250 °C, 32 h, 7% drop in CH3OH conversion	[22]
CuO/CeO2	260	100	2.4 a	\	H2 pretreatment; GHSV = 800 h−1; H2O/CH3OH = 1.2	\	[23]
CuO/ZnO/CeO2/ Al2O3	200	100	0	\	H2 pretreatment; GHSV = 10,000 cm3 g−1 h−1; H2O/CH3OH = 1.5	200 °C, 24 h, no drop	[24]
ZrO2/Cu	200	32	0	190	H2 pretreatment; WHSV = 10 h−1; H2O/CH3OH = 1.0	200 °C, 200 h, no drop	[7]
Cu/ZnO/ZrO2	250	88.6	0	12,600 mmol gCu−1 h−1	H2 pretreatment; H2O/CH3OH = 1.0	\	[25]
Cu/Ce1−xZrxO2	240	23	0	316	H2 pretreatment; WHSV = 27 h−1; H2O/CH3OH = 1.5	240 °C, 90 h, no drop	[26]
CuO/ZnO/CeO2-ZrO2	240	95	0.46 a	1836 mL g−1 h−1	H2 pretreatment; GHSV = 1200 h−1; H2O/CH3OH = 1.2	230–260 °C, 360 h, no drop	[27]
Cu/ZnO/CeO2/ ZrO2/SBA-15	300	95.2	1.4 b	H2 yield = 90%	H2 pretreatment; WHSV = 43.68 h−1; H2O/CH3OH = 2	300 °C, 60 h, 12% drop in CH3OH conversion	[28]
CuZnGaOx	150	22.5	0	393.6 mL g−1 h−1	H2 pretreatment; feed rate = 6 mL h−1; H2O/CH3OH = 2	\	[29]
CuGaZn	200	\	0.2 a	118.1	H2 pretreatment; WHSV = 6 h−1; H2O/CH3OH = 1.3	200 °C, 24 h, no drop	[30]
CuZnGaZr	250	42.9	0.3	10,620 mL g−1 h−1	GHSV = 2200 h−1; H2O/CH3OH = 1	275 °C, 44 h, 7% drop in CH3OH conversion	[31]
Cu/MgAl2O4	300	96	2.8 b	\	H2 pretreatment; WHSV = 8.5 h−1; H2O/CH3OH = 1	200 °C, 30 h, 4% drop in CH3OH conversion	[32]
Mg/Cu-Al spinel	255	96.5	3.8 a	H2 yield = 96.54%	WHSV = 2.28 h−1; H2O/CH3OH = 2.27	255 °C, 500 h, no drop	[33]
CuZnAlMg	200	68.5	0.88 a	172	H2 pretreatment; WHSV = 3.84 h−1; H2O/CH3OH = 1	350 °C, 8 h, 18% drop in CH3OH conversion	[34]
CuCeMg/Al	250	100	0.29 c	H2 yield = 29.1%	H2 pretreatment; feed rate = 1 mL h−1; H2O/CH3OH = 1	250 °C, 72 h, no drop	[35]
La(CoCuFeAlCe)0.2O3	600	98.9	8 c	436.8	LHSV = 20 h−1; H2O/CH3OH = 4	600 °C, 50 h, no drop	[36]

a: ${\mathrm{S}}_{\mathrm{CO}}={\displaystyle \frac{{\mathrm{y}}_{\mathrm{CO}}}{{\mathrm{y}}_{\mathrm{CO}}+{\mathrm{y}}_{\mathrm{C}{\mathrm{O}}_2}}}\times 100\%$; b: ${\mathrm{S}}_{\mathrm{CO}}={\displaystyle \frac{{\mathrm{y}}_{\mathrm{CO}}}{{\mathrm{y}}_{\mathrm{CO}}+{\mathrm{y}}_{\mathrm{C}{\mathrm{O}}_2}+{\mathrm{y}}_{{\mathrm{CH}}_4}}}\times 100\%$; c: ${\mathrm{S}}_{\mathrm{CO}}={\displaystyle \frac{{\mathrm{y}}_{\mathrm{CO}}}{{\mathrm{y}}_{\mathrm{CO}}+{\mathrm{y}}_{\mathrm{C}{\mathrm{O}}_2}+{\mathrm{y}}_{{\mathrm{H}}_2}}}\times 100\%$.

), significant progress has been made in the rational design of catalysts and the development of application technologies.

2.1. Performance

2.1.1. Active Sites

Under reaction conditions, various copper species (Cu⁰, Cu^δ+/Cu⁺) usually coexist on the copper-based catalysts, which adds complexity to identifying active sites. By using isotope-labeling experiments, in situ spectroscopy, and density functional theory (DFT) calculations [37], the role of Cu⁰ species on Cu@mSiO₂ (CuO core/mesoporous silica shell) catalysts was elucidated (Figure 2a). It was shown that Cu⁰ sites enabled the cleavage of the O-H bond and the C-H bond in methanol (CH₃OH), promoting the generation of the main intermediate methyl formate (HCOOCH₃) rather than the byproduct CO. Therefore, the activity and selectivity could be adjusted by controlling the ratio of Cu⁰ and Cu⁺. The process of transforming from methanol to methyl formate has generally been considered as a prototypical C₁ chemical reaction, such as in MSR. The 5 wt.%Cu5Zn10Al catalyst prepared by Mrad et al. [15] exhibited high activity, which was believed to be related to the presence of stable Cu⁺ ions, regarded as the most active species in the MSR reaction. These conflicting views on active sites prompted more researchers to delve into this area. Ma et al. [38] constructed a catalyst with dual Cu⁰ and Cu⁺ sites on an SBA-15 support, achieving a hydrogen production rate of up to 1145 mol kg_cat⁻¹ h⁻¹. Studies suggested that the significant performance enhancement was due to a Cu⁺-dominated dual-site reaction pathway replacing the traditional Cu⁰-dominated single-site pathway (Figure 2b). A synergistic effect was found between Cu⁰ and Cu⁺, with sufficient Cu⁺ accelerating the reaction rate, while Cu⁰ was crucial for H₂ desorption. Xu et al. [7] designed an inverse ZrO₂/Cu catalyst, achieving high activity, high stability, and CO-free production in MSR. On this inverse catalyst, experimental and theoretical studies indicated that its superior performance was closely related to the presence of ZrO(OH)-(Cu⁺/Cu) composite sites. These sites promoted the formation of H₂ and CO₂ via the formate (HCOOH) intermediate, thus preventing the continuous dehydrogenation of methanol to form CO. Meng et al. [12] constructed Cu⁰-Cu⁺ dual sites on Cu/Cu(Al)O_x catalysts, and, through in situ spectroscopic characterization and theoretical calculations, found that oxygen-containing intermediates, methoxy (CH₃O*) and formate, adsorbed with moderate strength at these sites during the MSR reaction process. This promoted the transfer of electrons from the catalyst to surface species, significantly reducing the energy barrier for the cleavage of the C-H bond in the methoxy and formate intermediates (rate-determining steps). The optimized catalyst exhibited a methanol conversion rate of 99.5%, with a corresponding hydrogen production rate of 110.8 μmol s⁻¹ g_cat⁻¹, and maintained stability at 240 °C for over 300 h. Ma et al. [26] designed Cu/Ce_1−xZr_xO₂ solid solution catalysts and demonstrated that optimal catalytic performance was achieved when the ratio of Cu⁺/Cu⁰ was approximately 1.0. Through the integration of diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) experiments and DFT calculations, they elucidated that the CH₃OH molecule was activated upon adsorption at the Cu⁺ site, while the H₂O molecule was activated at the Cu⁰ site. Liu et al. [13] proposed that a dynamic Cu^0/+ site and an adjacent Cu⁰ site were the active sites. Water is activated on the Cu⁰ site, oxidizing it to Cu⁺-OH species. These Cu⁺-OH species subsequently promote the formation of reactive formate that produces H₂ and CO₂ upon dehydrogenation, while reducing Cu⁺ back to Cu⁰. The synergistic interaction between these two types of sites governs the overall reaction performance. As seen in the above studies, the perspective that the Cu⁰-Cu⁺ dual sites synergistically catalyze the MSR reaction seems to be more reliable.

In addition to the view that various Cu species are active sites for the MSR reaction, many studies have focused on interface sites. Through ambient-pressure X-ray photoelectron spectroscopy (AP-XPS) and Auger electron spectroscopy (AES), Rameshan et al. [6] demonstrated that the Cu(Zn)⁰/Zn(ox) interface was the active site on Cu/ZnO catalysts. The Cu(Zn)⁰ region facilitated the selective dehydrogenation of methanol to formaldehyde (HCHO), while the redox-active Cu(Zn)⁰-Zn(ox) interface aided in the activation of water and the subsequent transfer of the resulting hydroxide to the formaldehyde, providing optimal conditions for high CO₂ activity and selectivity. Zhang et al. [39] prepared a series of CuZnO/γ-Al₂O₃/Al catalysts for the MSR reaction with enhanced durability. Further studies involving high-resolution transmission election microscopy (HRTEM), X-ray photoelectron spectra (XPS), and DFT calculations revealed that ZnO maintained an optimal Cu⁺/Cu⁰ ratio and formed numerous Cu₂O/ZnO synergistic sites (Figure 2c). This significantly reduced the activation energy for the rate-determining step of methoxy dehydrogenation in the MSR reaction (1.17 eV), compared to Cu/ZnO (1.25 eV) and Cu (1.36 eV). Through an adsorbate-induced strong metal–support interaction [17], the migration of ZnO_x to the surface of Cu⁰ nanoparticles occurred on commercial Cu/ZnO/Al₂O₃ catalysts, generating abundant Cu-ZnO_x interface sites (Figure 2d), which doubled the catalytic activity for the MSR reaction. Mao et al. [14] elucidated the importance of the Cu-Al₂O₃ interface site for the MSR reaction using Cu/Al₂O₃ and inverse Al₂O₃/Cu catalysts. Combining quasi-in situ X-ray photoelectron spectroscopy, in situ CO DRIFTS, and in situ temperature-programmed DRIFTS methods, they demonstrated that the formate species adsorbed on the interfacial site (HCOO-CuAl) dissociated more rapidly to CO₂ and H₂ than those adsorbed on Al₂O₃ (HCOO-Al). The optimal sample with abundant Cu-Al₂O₃ interface sites thus showed a high hydrogen production rate of 147.6 μmol g⁻¹ s⁻¹ at 250 °C. The aforementioned studies highlight the significance of interface sites in the MSR reaction, and the formation of these sites is often associated with the carrier materials. For example, Jin et al. [40] identified the Cu-O_V-Ce interface as the critical active site in their study of Cu/CeO₂ catalysts for the MSR reaction.

2.1.2. Supports

Although ZnO is generally used as a promoter in MSR catalysts at present, it was originally used as a support to load copper (Cu/ZnO), while it improved the MSR performance. Accordingly, many researchers investigated the promotion mechanism of ZnO [6]. To answer this question, different theoretical models have been developed for Cu/ZnO catalysts, including the spillover model, the morphology model, and the CuZn alloy model. The spillover model [41] suggested that ZnO served as a reservoir for hydrogen atoms, benefiting hydrogen spillover and back-spillover between Cu and ZnO, finally relating to the enhancement of catalytic activity. The morphology model [42] demonstrated that changes in the reaction atmosphere altered the interfacial free energy and affected oxygen vacancies at the Zn-O-Cu interface (Figure 3a). The partial reduction of ZnO enhanced its interactions with Cu, reduced the surface free energy, increased oxygen vacancies, and produced disk-like Cu particles with greater surface area and catalytic activity. The CuZn alloy model, supported by studies from Nakamura et al. [43], indicated that after reduction pretreatment, CuZn alloys were formed in the Cu/ZnO catalysts, increasing reaction activity. Additionally, ZnO could isolate Cu metal particles, preventing their agglomeration and sintering, thereby improving reaction stability. Later, when Al₂O₃ was introduced into the Cu/ZnO catalytic system as a support, the MSR performance was further enhanced [18,44,45,46]. Shokrani et al. [18] found through characterization by X-ray diffraction (XRD), Brunauer–Emmett–Teller (BET), and field emission scanning electron microscopy (FESEM) that adding Al₂O₃ to the traditional Cu/ZnO catalyst system reduced the crystallinity of the metal oxides, increased the specific surface area of the catalyst, and improved the dispersion of the metals. Performance test results also showed that the addition of Al₂O₃ increased methanol conversion and reduced CO production. Additionally, the proportions of Cu, Zn, and Al need to be carefully selected. Miyao et al. [47] adjusted the proportions of components in Cu/Zn/Al alloy catalysts, finding that a combination of 30 wt.% Cu, 20 wt.% Zn, and 50 wt.% Al achieved the highest hydrogen yield and selectivity.

The selection of support materials for copper-based catalysts extends beyond traditional supports such as ZnO and Al₂O₃, encompassing SiO₂, TiO₂, SBA-15, MCM-41, and carbon nanotubes (CNTs) as well. Díaz-Pérez et al. [19] studied copper-based catalysts supported on SiO₂, Al₂O₃-SiO₂, TiO₂-rutile, and TiO₂-anatase, finding that their MSR activity varied in the order Cu/SiO₂ > Cu/TiO₂-rutile > Cu/Al₂O₃-SiO₂ > Cu/TiO₂-anatase. It should be noted that a loss of MSR activity was observed over Cu/SiO₂ and Cu/Al₂O₃-SiO₂ catalysts due to the growth of Cu particle size and coke deposition, respectively. In contrast, Cu particle size growth on Cu/TiO₂-rutile and Cu/TiO₂-anatase was not significant, attributed to the strong interaction between copper clusters and TiO₂. As a result, deactivation was hardly observed over these two samples. Mesoporous SBA-15 was also employed as a support to prepare copper-based catalysts [28]. Thanks to its high specific surface area, the distribution of metal particles was improved, making it a preferred support material. Among the tested samples, Cu/SBA-15 exhibited good catalytic performance, with 91% methanol conversion and 2.8% CO selectivity at 300 °C. Similarly to SBA-15, MCM-41 mesoporous molecular sieves have become a support choice for MSR reaction catalysts. For instance, Deshmane et al. [20] prepared Cu-MCM-41 catalysts for the MSR reaction. Performance results showed that the 15% Cu-MCM-41 sample achieved 89% methanol conversion and 0.8% CO selectivity at 300 °C with gas hourly space velocity (GHSV) of 2838 h⁻¹, maintaining good stability for 48 h. Carbon nanotubes have garnered significant attention as support materials due to several advantages [21,48,49,50]. Firstly, their mesoporous structure reduces mass transfer resistance, allowing reactant and product molecules to easily diffuse within the catalyst channels, thus enhancing reaction rates. Secondly, CNT-supported catalysts have much higher specific surface areas compared to conventional catalysts, exposing reactant molecules to both the inner and outer surfaces of the CNT-supported catalysts, thereby increasing catalytic conversion rates. Additionally, CNTs have uniform pore size distributions and high thermal stability, allowing for even heat dissipation during high-temperature reactions, preventing hot spots that could lead to catalyst sintering. Shahsavar et al. [21] prepared CeCuZn/CNT catalysts for the MSR reaction, which maintained long-term stability (>48 h) at 300 °C and a WHSV of 7.5 h⁻¹, attributed to the high dispersion of metal particles and the strong interaction between the metals and the CNT support.

2.1.3. Promoters

Besides the previously mentioned Zn, the elements of Ce, Zr, Ga, and Mg are most commonly used as MSR catalyst promoters with a view on enhancing catalytic performance such as hydrogen yield, activity, selectivity, and stability. It has been reported that the addition of CeO₂ can facilitate the formation of oxygen vacancies (V₀) and adsorbed oxygen molecules (${\mathrm{O}}_0^{\mathrm{x}}$) on the catalyst surface, thereby promoting water dissociation through a hydration reaction (${\mathrm{V}}_0+{\mathrm{H}}_2\mathrm{O}+{ \mathrm{O}}_0^{\mathrm{x}}\to { 2\mathrm{OH}}_0$) [22,51]. As we all know, the dissociation of H₂O is a crucial step in the MSR reaction [17]. In detail, when Ce⁴⁺ is converted to Ce³⁺, oxygen vacancies are formed, accompanied by the production of adsorbed oxygen molecules [52,53]. This also increases the concentration of highly mobile oxygen species, which promotes carbon gasification and effectively enhances the anti-carbon stability of the MSR catalyst [54]. Men et al. [55] prepared the Cu/CeO₂/γ-Al₂O₃ catalyst and found that the copper–ceria interface was the active site for the MSR reaction, with its redox properties determining the catalytic activity. Oxygen vacancies at the interface served as the most active sites for water dissociation, whereas surface oxygen played a critical role in methanol decomposition. Liu et al. [56] enhanced the adsorption and dissociation of water in the MSR reaction by adding Ce and La to Cu-Al catalysts, which also strengthened the adsorption of the product CO₂. As a result, the WGS reaction was promoted, while the reverse-WGS reaction was suppressed, leading to a higher hydrogen production rate.

In the MSR reaction, the addition of ZrO₂ as a promoter significantly improved the stability of copper-based catalysts and reduced the concentration of CO byproducts. By using different characterization techniques, such as XRD, temperature-programmed oxidation (TPO), and temperature-programmed reduction (TPR), Agrell et al. [57] found that the introduction of ZrO₂ into the Cu/ZnO/Al₂O₃ catalyst effectively stabilized the copper by preventing crystal growth, thus enhancing catalyst stability. Similarly, Jeong et al. [58] noted that adding Zr to Cu/Zn catalysts increased copper dispersion, aiding in the formation of small copper particles on the catalyst surface. Wu et al. [59] suggested that in 10 wt.%ZrO₂/Cu catalysts, the presence of ZrO₂ increased the specific surface area of Cu, stabilized the grain size of Cu, and prevented the aggregation of Cu particles, thereby improving catalytic activity and stability. Additionally, the ZrO₂ surface had abundant hydroxyl groups that could stabilize Cu⁺ species, which was a major reason for the superior performance of the ZrO₂/Cu catalyst compared to the ZnO/Cu catalyst. A subsequent research study performed by Liu et al. [31] also supported the role of Zr in enhancing stability. Specifically, the CuZnGaZr catalysts showed excellent stability of hydrogen production (44 h) for the MSR reaction. Furthermore, the role of ZrO₂ in reducing CO selectivity has garnered significant attention. Early studies by Lindström et al. [60] on the role of ZrO₂ in Cu/Zn/Zr catalysts showed that ZrO₂ doping reduced hydrogen yield but increased CO₂ selectivity, keeping CO byproduct concentration below 1%. Jeong and Liu et al. [31,58] found that ZrO₂ played a role in suppressing CO formation. They achieved low CO selectivity over Cu/ZnO/ZrO₂/Al₂O₃ (0.1%) and CuZnGaZr (0.3%) catalysts, respectively. Research by Sanches et al. [25] indicated that the reduced CO selectivity was due to the strong adsorption of CO by monoclinic zirconia over Cu/ZnO/ZrO₂ catalysts.

In the MSR reaction, the promotion effects of gallium and magnesium on copper-based catalysts have also been widely discussed. Tsang’s team [29,61] pointed out that incorporating Ga into Cu-Zn oxides could promote the formation of a non-stoichiometric cubic spinel phase (ZnGa₂O₄). On the defect-rich surface of this ZnGa₂O₄ spinel, a large number of highly dispersed, extremely small copper clusters (5 Å) could be generated in situ, achieving a hydrogen yield of 393.6 mL g_cat⁻¹ h⁻¹ at 150 °C with almost no CO formation. Li et al. [30] found that the addition of Ga promoted the formation of a ZnO_x overlayer on the Cu/ZnO catalyst, enhancing strong metal–support interactions (SMSIs) and creating more abundant Cu-ZnO_x interfacial sites (Figure 3b), which increased the intrinsic activity of the MSR reaction by a factor of 4.6. In related research, the introduction of magnesium has often been associated with the formation of the spinel structure to improve MSR activity. Kamyar et al. [32] prepared four high-surface-area MAl₂O₄ (M = Ni, Co, Zn, Mg) spinel structures as supports for copper-loaded MSR catalysts. Among these, the Cu/MgAl₂O₄ catalyst exhibited a large specific surface area, high copper dispersion, and strong adsorption of reactants on its active sites, resulting in high methanol conversion (96%), high H₂ selectivity (86%), and low CO selectivity (2.8%). Hou et al. [33] discovered that although CuAl₂O₄ spinel catalysts could avoid the need for pre-reduction treatment before use, the catalyst activity and stability declined, as most of the Cu²⁺ in the lattice was released during the MSR reaction. To prevent this phenomenon, CuAl₂O₄ spinel catalysts were modified with MgO. The incorporation of Mg²⁺ into the spinel structure altered the environment around Cu²⁺, reducing its release rate and promoting the formation of small copper particles. The slower release rate of Cu²⁺ and the high specific surface area of copper nanoparticles were the main reasons for the improved activity and stability of MgO-doped CuAl₂O₄ spinel catalysts (Figure 3c). In addition, Liu et al. [62] pointed out that this sustained-release catalytic system was fundamentally different from traditional metal catalytic systems. In conclusion, future research should focus on systematically investigating the interactions between various promoters to achieve an optimal balance among these components, thereby tailoring catalyst properties for specific industrial applications.

2.1.4. Preparation and Activation

Improvements in catalyst synthesis methods are often pursued to achieve better catalytic performance. For the MSR reaction, traditional preparation methods for copper-based catalysts include coprecipitation, sol–gel, and impregnation methods. Yang et al. [23] synthesized CeO₂ with different morphologies including nano-rod (R), nanoparticle (P), and sponginess (S) using hydrothermal, precipitation, and sol–gel methods, respectively, to serve as supports for copper-based catalysts (Figure 4a). The results showed that CuO/CeO₂-R had stronger interactions between copper oxide and the cerium dioxide support, as well as the highest content of surface oxygen vacancies, exhibiting better MSR catalytic activity than CuO/CeO₂-P and CuO/CeO₂-S. This indicated that adjusting the morphology of the support could enhance catalytic performance. In recent years, some new preparation methods and improved techniques of traditional methods have been used for synthesizing copper-based MSR catalysts. Wen et al. [63] developed a hydrolysis precipitation method to prepare Cu-ZnO@Al₂O₃ catalysts for the MSR reaction and compared them with those prepared by the traditional coprecipitation method. The former catalysts showed better catalytic performance due to their more developed pore organization, surface enrichment of Cu and Zn, and higher reducibility. Haghighi’s team [24,64] synthesized CeO₂- and ZrO₂-doped CuO/ZnO/Al₂O₃ catalysts using the sonochemical coprecipitation method. Compared to traditional coprecipitation samples, the ultrasonic-assisted coprecipitation samples displayed clearer morphologies, more uniform active component distribution, smaller CuO crystallite size, and higher specific surface areas, resulting in higher methanol conversion rates (100%) at lower temperatures (200 °C) and lower CO byproduct concentrations. Additionally, CuO/ZnO/Al₂O₃ catalysts prepared by oxalate gel-coprecipitation [65] and urea hydrolysis homogeneous precipitation [66] methods have also shown advancements.

The dynamic structural evolution of heterogeneous catalysts during different stages of their life cycle (activation, reaction, and deactivation) is a common phenomenon [17]. Among them, the modulation of the activation process to improve the catalytic performance has great potential. Traditionally, copper-based catalysts require pretreatment to convert inactive components into active states through hydrogen reduction at high temperatures. Recently, Zhu’s team [17] optimized the activation stage of commercial Cu/ZnO/Al₂O₃ catalysts and analyzed the structure–activity relationship. CuZnAl-R10, which was reduced in a H₂/N₂ mixture for 50 min and then in a H₂/H₂O/CH₃OH/N₂ mixture for 10 min, had a methanol conversion rate of 65.5% and a CO selectivity of 0.07% after 8 h at a reaction temperature of 225 °C (Figure 4b). However, CuZnAl-H, which was reduced in a H₂/N₂ mixture for 1 h, exhibited a methanol conversion rate of 55.1% and CO selectivity of 0.11% under the same reaction conditions. This indicated that the induced activation method improved the activity, selectivity, and stability of commercial Cu/ZnO/Al₂O₃ catalysts. A series of in situ characterization studies and DFT calculations revealed that a surface reconstruction process occurred during activation. This adsorbate-induced strong metal–support interaction accelerated the migration of ZnO_x species to the surface of Cu⁰ nanoparticles, creating abundant Cu-ZnO_x interfacial sites (Figure 2d). These interfacial sites demonstrated faster CH₃O* dehydrogenation and H₂O dissociation kinetics. Subsequently, they applied the same method by altering the activation gas for Cu/ZnO catalysts, thereby modulating the strong metal–support interaction between Cu and ZnO to enhance the catalytic activity for the MSR reaction [67]. During activation, more oxygen vacancies were generated on the ZnO support, promoting the dissociation of H₂O and hydroxylation of ZnO species, the occurrence of which facilitated the migration of ZnO to Cu, forming abundant interfacial sites and promoting the reaction. Zhang et al. [39] conducted an in situ self-activation process under reaction conditions (CH₃OH/H₂O/N₂) before evaluating the performance of CuZnO/γ-Al₂O₃/Al catalysts. This process resulted in the formation of a Cu/Cu₂O/ZnO tri-layer core–shell structure (Figure 2c), where ZnO stabilized the optimal ratio of Cu⁺/Cu⁰ and created numerous Cu₂O/ZnO synergistic sites. These sites lowered the activation energy for the rate-determining step of CH₃O* dehydrogenation in the MSR reaction (Figure 4c), thereby enhancing catalytic performance. This study of the process of improving catalysts, from the improvement of preparation methods to the activation pretreatment of catalysts, is an innovative quest for superior catalysts by researchers.

2.2. Reaction Mechanism

After decades of development, two possible reaction pathways for methanol steam reforming on copper-based catalysts have been essentially identified as the following: the HCOO* route [14,17,68,69,70,71,72] and the HCOOCH₃* route [12,68,69,70,73,74] (Figure 5).

By using in situ temperature-programmed DRIFTS studies, Li et al. [17] elucidated the MSR reaction pathway over a commercial Cu/ZnO/Al₂O₃ catalyst, where surface methoxy species (*CH₃O) sequentially transform to formaldehyde (*CH₂O), methylene dioxygen (*CH₂OO), and formate (*CHOO), with formate ultimately decomposing to CO₂ and H₂. DFT calculations (Figure 6a) showed that the two steps with the highest energy barriers on Cu(111) were H₂O dissociation (1.06 eV) and the dehydrogenation of *CH₃O to *CH₂O (1.03 eV). Li et al.’s elucidation of the HCOO* route aligns with previous studies [71,72] and is supported by Mao et al. [14].

Meng et al. [12] conducted operando pulse experiments equipped with a mass spectrometer detector to elucidate the MSR reaction route on the Cu/Cu(Al)O_x catalyst. The MSR reaction on this catalyst follows the HCOOCH₃* route, consistent with previous studies [73,74], which involves the following steps: CH₃OH* firstly undergoes dehydrogenation to form CH₃O* and CH₂O* species; then, CH₂O* dimerizes or reacts with CH₃O* to generate HCOOCH₃*; subsequently, HCOOCH₃* hydrolyses to form HCOOH* and CH₃O*, and CH₃O* re-participates in the catalytic cycle; finally, the decomposition of HCOOH* occurs to produce CO₂ and H₂. Combined with DFT calculations (Figure 6b), it was demonstrated that the cleavage of the C-H bonds in the intermediates CH₃O* and HCOO* is the rate-determining step. Additionally, the study showed that water molecules promote the decomposition of HCOOCH₃* but do not directly participate in cleavage of the C-H bonds.

In fact, Frank et al. [68] had earlier summarized the HCOO* route and the HCOOCH₃* route (Figure 5) and pointed out that the ratio of methanol to water affects the selection of these pathways. When water is in excess, the HCOO* route is preferred, involving the one-step oxidation of HCHO* to HCOO* by hydroxyl groups or reactive oxygen species from H₂O dissociation, followed by the decomposition of HCOO* to produce CO₂ and H₂. Conversely, when methanol is in excess, the HCOOCH₃* route is preferred, involving the dehydrogenation of CH₃OH to HCOOCH₃, which is then hydrolyzed to HCOO*, followed by further decomposition to produce CO₂ and H₂.

2.3. Deactivation

Improving the stability of copper-based catalysts and addressing issues including sintering and carbon deposition are crucial steps to enhance their practical application value in the MSR reaction. Sintering of metal nanoparticles proceeds through Ostwald ripening (OR), or through particle migration and coalescence (PMC) [75]. Once particles are sintered, it is almost impossible to restore the original catalytic activity of the catalysts through methods such as redispersion; so, it is necessary to fundamentally prevent sintering from occurring. From a thermodynamic perspective, the sintering of nanoparticles in catalysts is driven by the reduction in surface free energy, with the rate significantly increasing at higher temperatures [76]. Typically, the onset of sintering is estimated using the Hüttig temperature, where surface atoms become mobile, and the Tammann temperature, where bulk atom diffusion occurs [77]. Therefore, controlling the reaction to occur at lower temperatures is a direct strategy to inhibit the sintering of copper-based catalysts. However, for the endothermic MSR reaction, high activity is typically associated with elevated reaction temperatures up to a certain limit, making it challenging to prevent sintering.

To overcome the limitations of reaction temperature, other strategies to inhibit sintering include [76] controlling the uniformity of nanoparticles to eliminate chemical potential differences between particles, enhancing the chemical bonding between metal nanoparticles and the support, physically confining the metal nanoparticles, and constructing energy barriers to prevent the surface diffusion of metal nanoparticles. Li et al. [17] achieved surface reconstruction by adjusting the composition of the reducing agent, causing ZnO_x to migrate to the surface of Cu⁰ nanoparticles in the Cu/ZnO/Al₂O₃ catalyst. This moderate encapsulation tripled the long-term stability of the copper-based catalyst (Figure 4b). Cheng et al. [34] improved the stability of Cu/ZnO/Al₂O₃ catalysts by adding Mg, which enhanced the Cu-ZnO synergistic effect, thus inhibiting the sintering of the Cu and ZnO phases. Sanches et al. [25] found that the presence of Zr in Cu/ZnO/ZrO₂ catalysts increased the microstrain in CuO and ZnO, reducing their grain size and limiting their growth. Siriruang et al. [78] prepared Cu-Zn/ZrO₂-doped Al₂O₃ catalysts using a urea impregnation method, which exhibited anti-sintering capability and maintained high hydrogen yield even after accelerated sintering treatment. Clearly, exploring more potential anti-sintering strategies based on fundamental scientific principles of sintering inhibition remains an area of ongoing effort for the future.

Carbon deposition can block catalyst pores and cover the active species on the MSR catalyst surface, leading to deactivation. Carbon deposits originate partly from hydrocarbons generated during the MSR reaction [79] or from carbon–oxygen species [54]; another source is elemental carbon produced from CO by the disproportionation reaction [54,79,80]. Unlike deactivation caused by sintering, catalysts deactivated by carbon deposition can be regenerated to restore their initial activity. A common regeneration strategy involves adding CeO₂ to copper-based catalysts, utilizing CeO₂’s oxygen storage capacity to remove carbon deposits from the catalyst surface [27,35,54,81]. Specifically, under the reducing conditions of the MSR reaction, the partial reduction of CeO₂ generates mobile oxygen, which promotes the gasification of carbon deposits, thereby enhancing the stability of the catalysts. Industrial practices also highlight the critical role of catalyst regeneration techniques in maintaining performance and longevity. Methods such as advanced passivation and controlled-atmosphere treatments during handling and storage are widely employed to minimize sintering during downtime. To address carbon deposition, strategies including the optimization of reaction conditions (e.g., water-to-methanol ratio and operating temperature) and periodic catalyst regeneration through controlled oxidation cycles are commonly implemented [12]. Incorporating these industrial insights bridges the gap between fundamental research and practical application, providing effective solutions for ensuring the long-term stability and efficiency of catalysts.

In summary, copper-based catalysts remain the most prevalent catalysts used in methanol steam reforming. Regarding the identification of their active sites, many studies indicate a synergistic interaction between Cu⁰ and Cu⁺, although this conclusion often does not take into account the influence of the commonly used ZnO and Al₂O₃ supports. When these supports are present, Cu-Zn and Cu-Al interface sites are more likely to form. To elucidate the role of the widely used ZnO support in copper-based catalysts, three models are commonly referenced: the spillover model, the morphology model, and the CuZn alloy model. The discussion then shifted to Cu/ZnO/Al₂O₃ catalysts, emphasizing their advancements over Cu/ZnO catalysts. Additionally, the use of alternative supports, such as SiO₂, TiO₂, SBA-15, MCM-41, and CNTs, as well as the incorporation of promoters like Ce, Zr, Ga, and Mg, was also discussed. Furthermore, the impact of preparation methods and reaction pretreatment activation processes on catalytic performance was summarized. Two potential reaction pathways for copper-based catalysts in the MSR reaction were highlighted, namely the HCOO* route and the HCOOCH₃* route, with pathway selection influenced by factors like the water-to-alcohol ratio and the nature of the active metal in the catalyst. Finally, the challenges of sintering and coking in copper-based catalysts during the MSR reaction were analyzed, identifying contributing factors such as Ostwald ripening, particle migration and coalescence, and the origins of carbon species, while presenting practical solutions to address these issues.

3. Noble Metal-Based Catalysts

In addition to the copper-based catalysts, noble metal-based catalysts are another large class of MSR catalysts. The noble metals most often considered are Pd and Pt, followed by Ru and Rh. The performance of representative catalysts in this category is listed in

Table 2. Performance summary of representative noble metal-based catalysts for methanol steam reforming.

Catalysts	Temperature (°C)	CH3OH Conversion (%)	CO Selectivity (%)	H2 Yield (mmol g−1 h−1)	Reaction Conditions	Stability	Ref.
Pd/ZnO	400	94	0.5 a	1628	H2 pretreatment; GHSV = 12,000 h−1; H2O/CH3OH = 1.2	\	[82]
Pd/ZnAl2O4	250	35	3.0 a	41.04	H2 pretreatment; Pmethanol = 6.4 mol%; H2O/CH3OH = 1.1	250 °C, 100 h, no drop	[83]
ZnPd/MoC	160	40.3	0.9 b	68.9	CH4/H2 pretreatment; feed rate = 1.2 mL h−1; H2O/CH3OH = 3	240 °C, 170 h, initial deactivation only	[84]
Pd/Zn1Zr1Ox	330	46	0	\	H2 pretreatment; GHSV = 17,000 h−1; H2O/CH3OH = 1.3	330 °C, 30 h, no drop	[85]
Pd/ZrO2-TiO2	300	98	37 c	\	H2 pretreatment; GHSV = 30,000 h−1; H2O/CH3OH = 0.16	\	[86]
Pd/In2O3/CeO2	375	96	1.3 c	250	H2 pretreatment; GHSV = 13,809.6 h−1; H2O/CH3OH = 1.4	400 °C, 30 h, no drop	[87]
Pd-Cu/ZnAl2O4	240	100	\	H2 yield = 84%	H2 pretreatment; GHSV = 2400 h−1	\	[88]
CuPd/ZrO2	220	63	5 a	86.3	H2 pretreatment; GHSV = 295 mol g−1 h−1; H2O/CH3OH = 1.4	240 °C, 80 h, no drop	[89]
Pt/In2O3/CeO2	325	98.7	2.6 a	333	feed rate = 1.2 mL h−1; H2O/CH3OH = 1.4	325 °C, 32 h, no drop	[90]
Pt/In2O3/CeO2	350	99.9	2.5 c	H2 yield = 64.7%	WHSV = 99,500 mL g−1 h−1; H2O/CH3OH = 1.4	350 °C, 100 h, 8% drop in CH3OH conversion	[91]
Pt/MoC	200	100	3 c	\	CH4/H2 pretreatment; WHSV = 9000 cm3 g−1 h−1; H2O/CH3OH = 1	200 °C, 20 h, no drop	[92]
Zn-Pt/MoC	160	65.9	\	106.9	Carburizing treatment; feed rate = 1.2 mL h−1; H2O/CH3OH = 3	120 °C, 25 h, 4% drop in CH3OH conversion	[93]
Pt1/ZnO	390	43	\	\	WHSV = 55,200 cm3 g−1 h−1; H2O/CH3OH = 1.5	\	[94]
Pt-K@S-1	250	15	<1.9% c	4308	H2 pretreatment; WHSV = 45 h−1; H2O/CH3OH = 3	400 °C, 50 h, no drop	[95]
Ru/TiO2	300	98.9	5.4 b	\	H2 pretreatment; WHSV = 1.8 h−1; H2O/CH3OH = 1.2	\	[96]
RuCe	400	98	0.13 c	882 mmol cm−3 h−1	feed rate = 3.47 mL h−1; H2O/CH3OH = 2	400 °C, 115 h, no drop	[97]
Ru1/CeO2	350	25.6	2.2 a	139.6	feed rate = 3 mL h−1; H2O/CH3OH = 3	350 °C, 72 h, no drop	[98]
Rh1/CeO2	350	21	36 a	100	feed rate = 3 mL h−1; H2O/CH3OH = 3	\	[98]

a: ${\mathrm{S}}_{\mathrm{CO}}={\displaystyle \frac{{\mathrm{y}}_{\mathrm{CO}}}{{\mathrm{y}}_{\mathrm{CO}}+{\mathrm{y}}_{\mathrm{C}{\mathrm{O}}_2}}}\times 100\%$; b: ${\mathrm{S}}_{\mathrm{CO} }={\displaystyle \frac{{\mathrm{y}}_{\mathrm{CO}}}{{\mathrm{y}}_{\mathrm{CO}}+{\mathrm{y}}_{\mathrm{C}{\mathrm{O}}_2}+{\mathrm{y}}_{{\mathrm{CH}}_4}}}\times 100\%$; c: ${\mathrm{S}}_{\mathrm{C}\mathrm{O}}={\displaystyle \frac{{\mathrm{y}}_{\mathrm{C}\mathrm{O}}}{{\mathrm{y}}_{\mathrm{C}\mathrm{O}}+{\mathrm{y}}_{\mathrm{C}{\mathrm{O}}_2}+{\mathrm{y}}_{{\mathrm{H}}_2}+{\mathrm{y}}_{\mathrm{C}{\mathrm{H}}_4}}}\times 100\mathrm{\%}$.

.

3.1. Palladium-Based Catalysts

The research on palladium-based catalysts for hydrogen production through methanol steam reforming began in the late 20th century. The Takezawa team [99] prepared a series of supported palladium-based catalysts using metal oxides as carriers, including Pd/SiO₂, Pd/Al₂O₃, Pd/La₂O₃, Pd/Nb₂O₅, Pd/Nd₂O₃, Pd/ZrO₂, and Pd/ZnO. Among the aforementioned catalysts, Pd/ZnO stood out for its high activity and selectivity in the MSR reaction. Therefore, the Takezawa team subsequently conducted a decade-long study on the Pd/ZnO catalyst [100,101,102,103]. They found that a PdZn alloy formed on the Pd/ZnO catalyst after high-temperature hydrogen reduction treatment, which greatly improved the MSR performance, especially in terms of selectivity.

3.1.1. Alloys

Notably, reduction temperature and palladium loading are important factors influencing the formation of the PdZn alloy. The PdZn alloy begins to form at 420 K, and as the reduction temperature increases, the proportion of the alloy increases, which corresponds to an enhancement in selectivity [100]. Additionally, the research indicated that an equivalent molar amount of ZnO to the palladium loading could be reduced to form the PdZn alloy; thus, the formation of the alloy was not limited by the palladium loading [101]. In the Takezawa team’s series of studies, the palladium loading was mostly 10%. This prompted subsequent researchers to consider the possibility of forming alloys with lower palladium loadings. Studies by Halevi and Peterson et al. [104,105,106] indicated that the PdZn_β phase (Pd/Zn molar ratio = 1) exhibited low CO selectivity, while the PdZn_α phase (Pd/Zn molar ratio > 1) favored CO formation during MSR. Meanwhile, the formation of different PdZn alloy phases was closely related to the palladium loadings. For example, the PdZn_α phase was predominant when the palladium loading was below 4.8% on the conventional Pd/ZnO catalyst [107]. This explains the use of high noble metal loadings in many studies on palladium-based catalysts for MSR. Based on these findings, Liu et al. [83] synthesized a PdZn alloy catalyst with a palladium loading of only 0.1% using ZnAl₂O₄ as the support. Due to the limited number of Zn atoms provided by ZnAl₂O₄ and the enhanced interaction between Pd and Zn by the polar facets exposed on ZnAl₂O₄, the PdZn_β alloy predominated on the Pd/ZnAl₂O₄ catalyst. This enabled high CO₂ selectivity in MSR.

Subsequently, there have been more studies on PdZn alloy catalysts for MSR, focusing on the alloy formation process and the impact of alloy crystallite size on performance. Föttinger et al. [108] monitored the dynamic formation process of the active phase PdZn alloy for catalysis in real time by means of the operando quick-EXAFS (extended X-ray absorption fine structure) technique (Figure 7a). It was demonstrated that alloy formation started at the nanoparticle surface and then proceeded from the surface inward, leaving a metallic Pd core at the end. Moreover, alloying was reversible, and treatment with oxygen led to alloy decomposition and the formation of metallic Pd, which was due to the preferential oxidation of Zn. Similarly, Wang et al. [109,110] inferred that the reduction process followed the sequence PdO/ZnO → Pd/ZnO → Pd/ZnO_1−x → PdZn alloy/ZnO on the Pd/ZnO catalyst based on temperature-programmed reduction and X-ray diffraction results. Among them, the PdZn alloy and Pd/ZnO_1−x formed by partial oxidation during the reaction might be the real active species. Their study also suggested that the best MSR performance was achieved when the particle size of the PdZn alloy was 5–14 nm. Karim et al. [111] showed that high selectivity in the MSR reaction could still be achieved by reducing the size of the small particles of palladium to below 2 nm, despite the lack of complete PdZn alloying. In order to avoid the dissolution of ZnO or the alteration of ZnO morphology brought about by conventional aqueous impregnation, Dagle et al. [112] synthesized Pd/ZnO catalysts using an organic preparation method. It was demonstrated that large-sized PdZn crystallites could significantly inhibit CO generation while exhibiting high activity for the MSR reaction. This conclusion was also supported by Lim et al. [113] and indicated that the decrease in defect sites due to the growth of PdZn alloy particles was the main reason for the suppression of CO generation. Although the size of the active metal components is the most prominent structural factor in the study of the structure–property relationship, their coordination environment, valence states, and geometric configuration are also changed along with the size [114]. Therefore, the effects of these structural change factors on catalytic performance need to be considered together.

The stability issues associated with copper-based catalysts and the high CO selectivity observed in palladium-based catalysts have led researchers to consider combining the advantages of both, resulting in the development of PdCu bimetallic catalysts. Azenha et al. [89,116] synthesized CuPd/ZrO₂ bimetallic catalysts via the wet impregnation method and found that the order of the impregnation of Cu and Pd affected the catalytic performance. When Pd was impregnated first, it facilitated a uniform distribution of Pd and Cu on the support surface, forming closely bonded PdCu nanoparticles or alloys. Additionally, the synergistic interaction between Pd and Cu adjusted the electronic structure of the system, thereby enhancing MSR activity and reducing CO selectivity. Ruano et al. [117] also demonstrated through in situ mass spectrometry that the formation of PdCu alloys was crucial for reducing CO selectivity. Mierczynski et al. [88] discovered that during the reduction process of the 1%Pd-20%Cu/ZnAl₂O₄ catalyst (300 °C, 5% H₂/Ar), hydrogen spillover from metallic Pd to Cu species promoted the reduction of copper oxide, leading to the formation of PdCu alloys. The formation of PdCu alloys increased the concentration of Cu⁰ or Cu⁺ active species (Figure 7b), thereby improving hydrogen yield and selectivity.

3.1.2. Supports and Promoters

Researchers have made many other attempts to develop palladium-based catalysts regarding supports (such as MoC, ZrO₂, TiO₂, and CeO₂) or promoters (such as In₂O₃ and Ga₂O₃). Tang et al. [84] introduced a small amount of Zn into Pd/MoC catalysts, which promoted the formation of the α-MoC_1−x phase and improved the dispersion of Pd on the surface of MoC, thereby enhancing the low-temperature performance, including the methanol conversion and hydrogen production rate, although the selectivity still needed improvement. Pérez-Hernández et al. [86] prepared Pd/ZrO₂-TiO₂ catalysts using ZrO₂-TiO₂ synthesized by the sol–gel method as a support. Compared to palladium catalysts supported on either ZrO₂ or TiO₂ oxide, Pd/ZrO₂-TiO₂ showed higher activity, but the CO selectivity remained high, around 30%. Matsumura et al. [118] coprecipitated PdO/ZnO/Al₂O₃ on an amorphous ZrO₂ support, forming ultrafine PdO particles. The obtained PdZnAl/ZrO₂ with 3 wt.% Pd content had activity as high as that of 10 wt.% PdZnAl, and it remained stable even after the reaction at 550 °C. Wang et al. [85] prepared Pd/Zn₁Zr₁O_x catalysts for the MSR reaction using nanoscale Zn₁Zr₁O_x mixed oxides as supports. The presence of ZrO₂ facilitated the dispersion of ZnO clusters, which in turn favored PdZn alloying and alloy stability, resulting in excellent selectivity. Barrios et al. [119] prepared ZnO-CeO₂ nanocomposite-supported Pd catalysts and found that although the reaction activity and selectivity of Pd/ZnO-CeO₂ were lower than those of Pd/ZnO, they exhibited higher stability, possibly due to the reducibility of CeO₂ and its ability to generate oxygen vacancies. Zhang et al. [87] found that the morphology of the CeO₂ support significantly affected the MSR performance of Pd/In₂O₃ catalysts. A Pd/In₂O₃/CeO₂ (rod-shaped) catalyst promoted the reaction of the intermediate formaldehyde with hydroxyl to produce CO₂ and H₂, due to the presence of large palladium nanoparticles and a high density of oxygen vacancies created by the strong interaction between In and Ce, showing good reactivity and stability (Figure 7c).

Penner and Lorenz et al. [120,121] confirmed that Pd-Ga bimetallic particles formed on Pd/Ga₂O₃ at appropriate reduction temperatures effectively inhibited methanol dehydrogenation to CO, in comparison with Pd/ZnO catalysts. Haghofer et al. [115] explored the relationship between intermetallic compounds formed on Pd/Ga₂O₃ catalysts and MSR performance. Their research indicated that the intermetallic compound Pd₂Ga formed within the temperature range of 548–673 K favored the MSR reaction, while Ga-rich PdGa formed at a higher reduction temperature of 773 K resulted in poorer MSR performance (Figure 7d). Föttinger et al. [122] demonstrated that the formation of intermetallic compounds (Pd₂Ga or PdZn) with Ga or Zn was crucial for reducing the amount of the byproduct CO from methanol dehydrogenation on palladium-based catalysts. However, the surface degradation of these intermetallic compounds at low temperatures was one reason affecting their stability in catalyzing the MSR reaction. Rameshan et al. [123] found that on Pd-Ga₂O₃-In₂O₃ catalysts, as long as the appropriate intermetallic phases were present and exhibited optimized intermetallic-support phase boundary dimensions, the presence of various supported intermetallic InPd and GaPd₂ phases would not adversely affect the activity or selectivity in MSR. Overall, forming appropriate intermetallic compounds and optimizing the catalyst structure are crucial for improving MSR reaction performance and stability on palladium-based catalysts.

3.1.3. Reaction Mechanism

By comparison with the reaction mechanism of copper-based catalysts, Iwasa et al. [100,124] summarized three possible pathways of MSR on palladium-based catalysts (Figure 8a). On palladium alloys, pathways II and III, which primarily produced CO₂ and H₂, were preferred over pathway I. In contrast, for metallic palladium without alloy formation, a large amount of the byproduct CO, derived directly from the decomposition of HCHO, was produced. It appears that all the differences stem from HCHO, a key reaction intermediate of the MSR reaction. With the advancement of surface science research [125], researchers discovered that the adsorption configurations of formaldehyde on palladium alloys and metallic palladium differed (Figure 8b) [70,126]. This difference was likely the reason for the distinct reaction pathways observed on the two types of catalysts. As for the catalysts containing Pd alloys, the η¹(O)-HCHO configuration anchored on positively charged Pd sites shows preferential stability, where the carbonyl group is perpendicular to the surface and only oxygen interacts with the metal; this intermediate further reacts with the hydroxyl group to produce the formate species, which eventually decomposes into CO₂ and H₂. As for the catalysts containing metallic Pd, differently, formaldehyde adsorbs in an η²(C, O)-HCHO configuration where the carbonyl group is parallel to the surface and both C and O atoms are bound to the metallic Pd. In this case, electrons from metallic Pd strongly donate to the π*_CO antibonding orbital of formaldehyde, promoting the rapid decomposition of η²(C, O)-HCHO to CO and H₂. Unfortunately, there is still a lack of microscopic evidence supporting the MSR mechanism over palladium-based catalysts through the formaldehyde adsorption configuration theory.

Although the preferred bonding configuration of the formaldehyde intermediate has been considered to be a reasonable mechanism for the high selectivity of the PdZn alloy, this hypothesis was contradicted by the DFT calculations of Chen et al. [113,128,129], based on the interaction between formaldehyde and the PdZn (111) surface. Their calculations indicated that the η²(C, O) configuration was the most stable on the alloy surface, rather than the η¹(O) configuration proposed by Iwasa et al. [70,126]. These theoretical calculations also suggested that increasing the C-H bond dissociation barrier, thereby inhibiting formaldehyde dehydrogenation, was more likely due to the beneficial effect of forming the PdZn alloy. As shown in Figure 8c, the barrier for the dissociation of the C-H bond in formaldehyde on PdZn (111) is 40 kJ mol⁻¹ higher than on Pd (111) [113]. Gu and Li [130] also confirmed through DFT calculations that on Pd(111), formaldehyde tended to undergo direct dehydrogenation. Later, the above results of computation were also verified by experimental means such as temperature-programmed desorption (TPD) and high-resolution electron energy loss spectroscopy (HREELS) by Jeroro et al. [131].

3.1.4. Deactivation

In the MSR reaction, the stability of palladium-based catalysts is widely recognized and confirmed to be superior to that of copper-based catalysts. Iwasa et al. [105] compared the stability of Pd/Zn/CeO₂ and Cu/ZnO catalysts. The results showed that at 623 K, Pd/Zn/CeO₂ did not deactivate within 180 min; in contrast, the hydrogen production of Cu/ZnO gradually decreased over time, dropping by 20% of the initial value after 180 min. Conant et al. [127] tested the 60 h stability of Pd/ZnO/Al₂O₃ and commercial Cu/ZnO/Al₂O₃ catalysts (Figure 8d). The results showed that the Cu/ZnO/Al₂O₃ catalyst rapidly deactivated within the first 12 h, with the methanol conversion decreasing by 40% after 60 h; meanwhile, the Pd/ZnO/Al₂O₃ catalyst deactivated by only about 17% in the initial 20 h and remained stable thereafter up to 60 h. Additionally, the Pd/ZnO/Al₂O₃ catalyst could recover its initial activity after redox treatment, whereas Cu/ZnO/Al₂O₃ could not, indicating that the deactivation of palladium-based catalysts may not be due to sintering.

Despite the relatively low extent of deactivation observed for palladium-based catalysts, investigation into the underlying causes of their deactivation remains an ongoing area of research. Suwa et al. [132] confirmed that the deactivation of Pd/ZnO catalysts was due to the active PdZn alloy sites being covered by zinc carbonate hydroxide (Zn₄CO₃(OH)₆·H₂O) formed from ZnO. Liu et al. [133] suggested two reasons for the deactivation of Pd/ZnO catalysts: the surface carbon deposition contamination and surface oxidation decomposition of the PdZn alloy ($\mathrm{Pd}\text{-}\mathrm{Zn}+2{\mathrm{H}}_2\mathrm{O} \leftrightarrow \mathrm{Pd}+\mathrm{Zn}({\mathrm{OH}}_2)+{ \mathrm{H}}_2$). Therefore, the catalyst can be regenerated in an oxygen-containing atmosphere at relatively low temperatures to remove carbon deposits or in a hydrogen-containing atmosphere at higher temperatures to regenerate the PdZn alloy. From the above studies, it is clear that maintaining the stability of the PdZn alloy structure is crucial. Penner et al. [134] synthesized Pd/ZnO/SiO₂ catalysts, and the PdZn alloy particles on the surface could maintain good structural and thermal stability between 473 and 873 K. Even at temperatures above 873 K, the PdZn alloy only partially decomposed. This broad stability range is related to the strong interaction between Pd and ZnO and the high stability of the 1:1 PdZn alloy phase. These findings may explain the excellent stability of palladium-based catalysts, although their activity and selectivity still lag behind those of copper-based catalysts.

3.2. Other Catalysts

Following the success of palladium-based catalysts, researchers turned their attention to other noble metal-based catalysts, such as Pt, Ru, and Rh. Among them, Pt-based catalysts have proven to be promising and effective for the MSR reaction. Studies by the Men team [90,135,136] and Shanmugam et al. [91] have shown that the Pt/In₂O₃/CeO₂ catalyst exhibited low CO selectivity and good stability over 100 h of the MSR reaction. Over this catalyst, In₂O₃ improved the dispersion of Pt nanoparticles and enhanced the interaction between the metal and the support, while CeO₂ provided abundant active oxygen species with generating oxygen vacancies (Figure 9a). The combined effects of In₂O₃ and CeO₂ promoted the activation of water, favoring the main reaction and then reducing the formation of the byproduct CO. Additionally, many studies have used the α-MoC phase as support to increase the dispersion of loaded Pt atoms, achieving high performance in the MSR reaction [137]. Ma et al. [92] prepared Pt/α-MoC_1−x, which achieved 100% methanol conversion at the low temperature of 473 K. Cai et al. [93] modified Pt/MoC with Zn, promoting the formation of the α-MoC_1−x phase, thus enhancing Pt dispersion and improving the interaction between the metal and the support, thereby enhancing catalytic performance. The emergence of single-atom catalysts also offers the potential for further improving the MSR performance of Pt-based catalysts. Lin et al. [138] developed atomically dispersed Pt single-atom catalysts using α-MoC as the support. The well-dispersed Pt₁ atoms provided a high density of electron-deficient surface sites, promoting methanol adsorption and activation, while α-MoC offered high-activity sites for water dissociation, generating abundant surface hydroxyl groups. These resulted in excellent hydrogen production activity and stability. Gu et al. [94] reported that Pt single atoms were stabilized on the ZnO surface by lattice oxygen. This made the reaction intermediates bind more firmly to the active Pt₁ sites, altering the reaction energy and kinetics, and subsequently changing the reaction pathway. The turnover frequency (TOF) of the Pt₁ active sites was 1000 times higher than that of ZnO.

Many researchers have also made efforts to identify the active sites and reaction pathways of Pt-based catalysts in the MSR reaction. Shao et al. [95] analyzed the MSR reaction mechanism on Pt-K@S-1 catalysts using temperature-programmed surface reaction mass spectrometry (TPSR-MS) and DFT calculations (Figure 9b). The cleavage of the O-H bond in CH₃OH was activated over Pt⁰ sites to produce HCOOCH₃, whereas Pt^δ+ sites promoted the hydrolysis of HCOOCH₃ to HCOOH and ultimately CO₂ and H₂. The synergy between Pt⁰ and Pt^δ+ on the S-1 support, along with the promotion by K, endowed the catalyst with excellent activity, selectivity, and stability for the MSR reaction. Wang et al. [139] elucidated the reaction mechanism on Pt/NiAl₂O₄ catalysts through spectroscopy, kinetics, and isotope studies. Their research suggested that the MSR reaction involved a tandem process of methanol dehydrogenation and water-gas shift; the interface sites between Pt and NiAl₂O₄ were active for methanol dehydrogenation, while the sites on NiAl₂O₄ were active for the water-gas shift reaction (Figure 9c). Jin et al. [140] analyzed the MSR reaction pathway on Pt (111) surfaces theoretically using DFT and kinetic Monte Carlo (kMC) calculations to reduce CO selectivity. Their studies indicated that the energy barrier difference between the steps H₂O* + * → OH* + H* and CH₃OH* + * → CH₂OH* + H* was critical for CO selectivity. When the energy barrier for the former was 0.30 eV lower than that for the latter, CO formation was significantly suppressed. These studies provide valuable insights for designing high-performance Pt-based catalysts for MSR.

Ru-based catalysts are widely used in homogeneous catalysis, such as low-temperature aqueous-phase methanol reforming, achieving efficient and stable hydrogen production [141,142,143]. By dispersing a liquid film over the large inner surface area of a porous solid, known as supported liquid phase (SLP) technology, these efficient Ru-based catalysts can be applied to the heterogeneous MSR reaction [144]. Schwarz et al. [144] deposited a basic and hygroscopic KOH coating on an Al₂O₃ support, effectively fixing the Ru-pincer complex on the support. In the temperature range of 130–170 °C, only trace amounts of CO were produced, demonstrating high hydrogen production activity without deactivation for 70 h. Tahay et al. [96] deposited metallic Ru on a monolithic TiO₂ support for the MSR reaction. Compared to other metals like Cu, Cu-Ni, and Pt, Ru/TiO₂ showed higher conversion and selectivity. This was attributed to the metal–support interaction, which enhanced the activity and dispersion of Ru metal particles on the TiO₂ support. Aouad et al. [97] prepared three catalysts, Ru/Ce, Ru/Al, and Ru/CeAl, using the impregnation method. Among them, the 5 wt.% Ru/Ce catalyst exhibited the best performance, with a CO selectivity of 0.81% at 400 °C, and remained active for 115 h without deactivation. This was due to the synergistic effect between Ru and Ce, promoting the formation of active sites with excellent redox properties. Chen et al. [98] studied the mechanism of the tandem MSR reaction on Ru₁/CeO₂ and Rh₁/CeO₂ single-atom catalysts (Figure 9d). The active centers, composed of metal single atoms and adjacent oxygen vacancies, exhibited a unique synergistic catalytic effect. Ru₁/CeO₂ showed a hydrogen production rate of up to 579 mL_H2 g_Ru⁻¹ s⁻¹ with 99.5% CO₂ selectivity, followed closely by Rh₁/CeO₂. Additionally, Lytkina et al. [145,146,147] used detonation nanodiamond (DND) as a support to load Ru-Rh bimetals, achieving a high hydrogen production rate and low CO selectivity. Their studies provided a new insight into the design of RuRh bimetallic catalysts, although economic costs need careful consideration.

In summary, palladium-based catalysts are a focal point of research within noble metal catalysts for the MSR reaction. The analysis began with the well-established Pd/ZnO catalyst used in methanol steam reforming, examining the factors that influence the formation of the PdZn alloy and the effect of alloy size on catalytic performance. Other commonly used promoters and supports, such as MoC, ZrO₂, TiO₂, CeO₂, In₂O₃, and Ga₂O₃, were also discussed. Considering the advantages and disadvantages of both copper-based and palladium-based catalysts, the application of the PdCu alloy in MSR reactions was reviewed. However, current research primarily focuses on enhancing activity and selectivity, with comparatively less emphasis on catalyst stability and the underlying microscopic mechanisms. The possible reaction pathways on palladium-based catalysts appear to be similar to those on copper-based catalysts, although the former tend to produce higher amounts of CO. This difference between metallic palladium and PdZn alloys is attributed to two factors: the adsorption configuration of the key intermediate formaldehyde (η¹(O)-HCHO or η²(C, O)-HCHO) and the energy barrier for C-H bond cleavage in formaldehyde. Despite the excellent stability exhibited by palladium-based catalysts, potential deactivation mechanisms were examined, including the coverage of active alloy sites, oxidation and decomposition of the alloy, and carbon deposition. Beyond palladium-based catalysts, other noble metal-based catalysts, such as those based on Pt, Ru, and Rh, were also discussed. The catalytic performance, active sites, and reaction mechanisms of Pt-based catalysts, as well as the applications of Ru-Rh bimetallic catalysts, were highlighted. Additionally, significant research interest has been drawn to the field of single-atom catalysts involving these noble metals.

4. Conclusions and Perspectives

The carbon neutrality initiative has driven the advancement of sustainable methods for hydrogen production and secure storage [1]. Methanol steam reforming technology for in situ hydrogen production perfectly aligns with the aforementioned requirements. In this technology, the catalysts serve as the core component. This review provides a detailed introduction to the two main categories of catalysts used in the MSR reaction—copper-based catalysts and noble metal-based catalysts. It not only covers the roles of each component but also delves into their active sites, structure–activity relationships, reaction mechanisms, and deactivation mechanisms.

In summary, copper-based catalysts, represented by the Cu/ZnO/Al₂O₃ catalyst, exhibit good low-temperature activity, high hydrogen yield, low CO selectivity as a byproduct, and low cost. However, they are prone to deactivation due to sintering and carbon deposition, and their stability needs further improvement. Noble metal-based catalysts, represented by Pd/ZnO catalysts, have high CO selectivity and high cost, but are less prone to deactivation and have excellent thermal stability as a standout advantage. In real-world applications, the selection between copper-based and noble metal-based catalysts is often influenced by the trade-off between cost and performance. Copper-based catalysts are typically the preferred choice for large-scale hydrogen production, where cost considerations are paramount. In contrast, noble metal-based catalysts are favored in niche applications that prioritize stability and durability under harsh reaction conditions. Hybrid systems that combine the cost effectiveness of copper with the high-temperature stability of noble metals may offer a balanced solution.

On the other hand, despite decades of research, the MSR reaction mechanism and the active sites of catalysts remain inconclusive, warranting further investigation. Firstly, the specific reaction intermediates, such as formate (HCOO*), methoxy (CH₃O*), and hydroxyl groups (OH*), and their contributions to the reaction pathway are still debated. While formate species are often observed in experimental studies, their direct role in the formation of CO₂ versus CO remains unclear. Secondly, the interplay between copper nanoparticles and supports (e.g., ZnO, CeO₂, ZrO₂) significantly impacts catalytic activity and stability. However, the exact nature of active sites, particularly at the metal–support interface, and how support modifications influence the reaction pathway, requires further exploration. Finally, the factors governing selectivity toward CO or CO₂, including the contributions of methanol decomposition versus steam reforming pathways, remain insufficiently understood. This lack of understanding hinders the design of catalysts that can effectively suppress CO formation. To address these challenges, a combination of experimental and computational approaches should be applied systematically.

Given the current research status, future studies on MSR for hydrogen production could focus on the following areas:

(1) Improving the stability of copper-based catalysts based on structure–activity relationships: To mitigate sintering induced by Ostwald ripening, it is crucial to address the chemical potential difference between particles. This can be achieved by ensuring uniform copper particle size during catalyst preparation, which can be facilitated by methods such as strong electrostatic adsorption (SEA). Furthermore, selectively anchoring copper nanoparticles to specific oxide facets can enhance chemical bonding strength and reduce the likelihood of coalescence. Utilizing specific physical structures—such as core–shell, core–sheath, lamellar, and mesoporous matrices—can effectively limit copper particle growth and aggregation. Additionally, employing a dual-oxide support composed of size-controlled nanoscale domains of two different oxides can create energy barriers that impede copper nanoparticles from sintering via surface diffusion.

(2) Rationally designing palladium-based alloy catalysts to reduce CO selectivity: The previously mentioned PdZn and PdCu alloy catalysts have demonstrated excellent performance in reducing CO selectivity in the MSR reaction. This phenomenon may be related to the optimization of the electronic structure of the core metal atom Pd by the second metal in alloys, as well as the modulation of the interactions between the core metal and key intermediates. The choice of the second metal is not limited to Zn and Cu; other transition metals, such as Fe, Co, Ni, and W, are also ideal candidates.

(3) Establish methods to quickly and accurately identify the reaction mechanism of MSR: Most existing studies on the MSR reaction lack systematic and comprehensive approaches to quickly and accurately identify the active sites on catalysts and the reaction pathways. This may involve advanced in situ characterization techniques, such as in situ X-ray absorption spectroscopy (XAS), Fourier transform infrared spectroscopy (FTIR), and scanning transmission electron microscopy (STEM), which enable real-time monitoring of catalytic processes. In addition, density functional theory calculations and kinetic models can offer theoretical support by providing insights into reaction pathways and active sites on the catalyst surface. By integrating these experimental and computational approaches, a clearer understanding of the MSR mechanism can be achieved, leading to the development of more efficient catalysts.

(4) Enhancing the atomic utilization efficiency of MSR catalysts: From an economic perspective, improving the atomic utilization efficiency of MSR catalysts is beneficial, whether dealing with noble metals like palladium or transition metals like copper, which often require high loadings for enhanced activity. Single-atom alloy (SAA) catalysts, which feature an atomically dispersed metal within a bi- or multi-metallic complex, have emerged as a promising system for heterogeneous catalysis. Based on this, constructing PdCu single-atom alloy catalysts not only enhances atomic utilization efficiency but also potentially combines the advantages of both metals to achieve improved MSR performance.

In addition, methanol occupies a central role in sustainable energy systems, not only as a hydrogen carrier and feedstock for steam reforming but also as a platform molecule that can be produced from CO₂ and methane. The catalytic hydrogenation of CO₂ to methanol has emerged as a promising technology for carbon capture and utilization (CCU), transforming greenhouse gasses into valuable fuels and chemicals. Similarly, methane, the primary component of natural gas, can be selectively converted into methanol via oxidative or non-oxidative routes, offering a cleaner pathway for utilizing abundant natural gas reserves. These production methods create a closed-loop system where CO₂, captured from industrial emissions or the atmosphere, can be transformed into methanol and subsequently converted into hydrogen through MSR. This integrated cycle not only supports the transition to a hydrogen economy but also aligns with global efforts to mitigate climate change. Recent advances in catalyst design for CO₂ and methane conversion, such as bimetallic catalysts and single-atom catalysts, parallel the catalyst development strategies discussed in this review, highlighting opportunities for cross-disciplinary innovation.