Review on Copper and Palladium Based Catalysts for Methanol Steam Reforming to Produce Hydrogen

Abstract

Methanol steam reforming is a promising technology for producing hydrogen for onboard fuel cell applications. The methanol conversion rate and the contents of hydrogen, carbon monoxide and carbon dioxide in the reformate, significantly depend on the reforming catalyst. Copper-based catalysts and palladium-based catalysts can effectively convert methanol into hydrogen and carbon dioxide. Copper and palladium-based catalysts with different formulations and compositions have been thoroughly investigated in the literature. This work summarized the development of the two groups of catalysts for methanol steam reforming. Interactions between the activity components and the supports as well as the effects of different promoters were discussed. Compositional and morphological characteristics, along with the methanol steam reforming performances of different Cu/ZnO and Pd/ZnO catalysts promoted by Al₂O₃, CeO₂, ZrO₂ or other metal oxides, were reviewed and compared. Moreover, the reaction mechanism of methanol steam reforming over the copper based and palladium based catalysts were discussed.

1. Introduction

Hydrogen is a promising alternative fuel because of its high energy density and low greenhouse gas emission rate [1]. As an energy carrier, hydrogen can be used to produce electrical power in a fuel cell. The use of fuel cells is attractive for the vehicle industry due to their high energy conversion efficiency and low emissions [2]. However, refueling of hydrogen is a major obstacle for the deployment of fuel cell vehicles owing to the lack of infrastructure [3]. Onboard hydrogen production by a processing system which converts liquid hydrocarbon fuels into hydrogen is a practical solution [4,5]. Among various fuels, methanol receives attention for onboard processing applications because: (a) It is easy to handle and store; (b) It can be extracted from renewable sources and it is biodegradable; (c) Its reforming temperature is relatively low compared to the other fuels; and (d) Its tendency to form coke during reforming is low due to the high H/C ratio [6,7].

Three main onboard fuel processing mechanisms to produce hydrogen are steam reforming, partial oxidation and auto thermal reforming [8]. Steam reforming is the most industrial kind of mature technology and it is also the most commonly employed process to produce hydrogen from methanol reforming [9]. Among all the three mechanisms, steam reforming has the highest ratio of hydrogen and carbon monoxide in the reformate.

The performance of a methanol steam reforming process significantly depends on the catalyst. Ni-based catalysts are mostly used in large scale industrial processes. However, only around 20% of the catalysts can be effectively utilized in industrial processes [10]. Thereafter, excess amounts of catalysts need to be used to counterbalance the effect of the low catalyst-utilization rate. For onboard methanol steam reforming, both the mass and size of the catalyst are restricted. Thus catalysts with higher activity and a better utilization rate are investigated. Interactions between the metal components and the support materials are critical for catalyst performance [11]. Promoters such as alkaline metals or acidic oxides are also important because they could promote the reactions following expected paths and help remove undesired matters that are formed during the reactions such as the carbonate species. Deactivation mechanisms including sintering and coke formation are also important concerns [12]. Overall, the objective for the development of an effective catalyst is to achieve the highest activity and hydrogen selectivity while minimizing the activity for side reactions such as methanol decomposition and maintaining the longest possible durability [9]. For different catalyst formulations, the influence of interactions between the catalyst components and the support as well as the addition of promoters on the structural, morphological and redox properties of the active phase needs to be examined.

The present work summarized the development of copper-based catalysts and palladium-based catalysts for methanol steam reforming. Interactions between the activity components and the supports as well as the effects of different promoters were discussed in detail. Methanol steam reforming performances based on different catalysts and the reaction mechanism were also presented. Besides copper-based and palladium-based catalysts, other metals such as nickel and cobalt are also used in catalysts for methanol steam reforming. Ni-based catalysts supported on alumina, zirconia, ceria and those containing Zn-Al or Mg-Al were intensively investigated for ethanol steam reforming instead of methanol reforming. The control of metallic Ni particle size and shape, as well as the redox properties of the catalyst system, are key factors in the preparation of Ni-based catalysts. Co and NiCo catalysts supported on cerium, lanthanum, zinc oxides, alumina and lantana-alumina were also studied for ethanol steam reforming. The key factors of Co-based catalysts are the control of metal morphology properties and the redox state of the system. A recent review summarized the progress of Ni-based and Co-based catalysts for methanol steam reforming [9]. Therefore they are not discussed in the present study.

2. Catalysts

2.1. Copper-Based Catalysts

Cu-based catalysts are commonly employed in methanol steam reforming reactions for onboard applications. Cu/ZnO catalysts are originally designed for methanol synthesis, and are also often used for methanol steam reforming because of the high activity at low operating temperature in the range of 250 °C and 300 °C. However, they need pre-treatment in hydrogen to obtain full activity, which is not convenient in small scale onboard systems [13]. Other disadvantages of the Cu/ZnO catalysts include the possibility of pyrophoric behavior occurring when exposed to air and the high sensitivity to a few ppm of sulfur [14].

Although it is confirmed that zinc leads to an improvement in the desired properties of the Cu/ZnO catalysts, the mechanism of the promotion effect of zinc is not completely understood [15]. Empirical experience indicates a particularly large Cu-ZnO contact is desired in order to improve the catalyst performance, whereas the casualty of the contact on the catalyst is still ambiguous [16]. Several different explanations of the mechanism can be found in the literature such as the spillover model, the metallic copper model and the Cu-Zn alloy model. The two-way transfer of hydrogen between Cu and ZnO is supposed to be crucial in the ZnO-promoted Cu system as explained in the spillover model [17]. Because ZnO has a large capacity for adsorbed hydrogen [18], it may trap H atoms originally produced on the Cu surface and serve as a reservoir to facilitate hydrogen spillover in the catalyst system. However, the hydrogen spillover from Cu to ZnO was found to be rapid from a partially oxidized Cu surface but limited from a fully reduced Cu surface [17]. In the metallic copper model, Cu particles supported on ZnO are confirmed in the metallic state by extended X-ray absorption fine-structure spectroscopy (EXAFS) [19]. However, the morphology of the Cu particle may reversibly change under different operating gaseous conditions, while no detectable change of the ZnO phase can be observed. An increase of the oxidation potential in the operating gas can lead to more spherical Cu particles with larger particle size. In contrast, flat dish-like Cu particles can be obtained when the operating gaseous condition is reduced as shown in Figure 1a [20,21]. The metallic copper model also argues that Cu-Zn alloy may play a role in the Cu/ZnO system, but Cu-Zn alloy is only formed at extreme reducing operating conditions. However, the Cu-Zn alloy model claims that the role of ZnO in the Cu/ZnO system is not to support Cu but to separate Cu particles from sintering as a spacer. Thus the Cu/ZnO catalyst is a mixed system of Cu and ZnO particles as shown in Figure 1b. The Cu-Zn or Cu-O-Zn surface alloy is supposed to be responsible for the promotion effect of ZnO depending on the reaction conditions and the amount of Zn [22]. Activity of the Cu/ZnO catalyst was proposed to be related to the copper metal surface area. Alejo et al. [23] found that the catalyst Cu₄₀Zn₆₀ (containing Cu 40 wt.% and Zn 60 wt.%) showed the highest copper area in the series of Cu-Zn contents with copper relative content of 20–70 wt.%. Correspondingly, Cu₄₀Zn₆₀ has the highest activity for methanol reforming to produce hydrogen. Therefore it can be confirmed that the Cu-ZnO interaction also depends on the catalyst compositions. Both the turnover frequencies (TOF) and apparent activation energy change with the catalyst compositions.

Lunkenbein et al. [24] were the first to report clear evidence for the formation of metastable graphite-like (GL) ZnO layers in industrial Cu/ZnO/Al₂O₃ catalysts induced by strong metal-support interactions. High-angle annular dark field scanning transmission electron microscopy (HAADF-STEM) was employed to analyze the microstructure of the fresh catalyst in the reduced state. The results show that the Cu nanoparticles (in pink color) are homogeneously dispersed within a porous ZnO (in yellow color) rod-like matrix as shown in Figure 2. High resolution TEM results also indicate that the ZnO layer is 1 to 2 nm thick comprising four to six sheets with interlayer distance between 1.8 and 2.8 Å. The proposed scheme of Cu and ZnO interaction is shown in Figure 3. Although the TEM imaging shows almost complete surface coverage of Cu particles, the accessibility of the Cu surface is still unclear due to the fact that only 2D projections are indicated in TEM imaging. Fichtl et al. [25] concluded that about 30% of the Cu surface area is accessible based on H₂-temperature programed reduction (H₂-TPD) results. Lunkenbein et al. [24] argued that these results discredit the direct correlation between catalyst activity and surface area.

Stabilizers and promoters are often employed in Cu/ZnO catalysts. Alumina can be added as support of Cu-ZnO to increase the catalyst durability and selectivity of H₂ in the reformate [9]. Alejo et al. [23] found that the presence of alumina in Cu₄₀Zn₅₅Al₅ (containing Cu 40 wt.%, Zn 55 wt.% and Al 5 wt.%) had a slightly inhibiting effect on methanol conversion, whereas it significantly improved the catalyst stability and selectivity for H₂ and CO₂ with only traces of CO existing in the reformate. Without the addition of alumina, the Cu₄₀Zn₆₀ catalyst started to deactivate after 20 h at 503 K. In comparison, the durability of Cu₄₀Zn₅₅Al₅ was increased to more than 110 h at the same operating temperature [23]. A recent work reported that a Cu-Al spinel oxide catalyst can improve the activity for methanol steam reforming compared to the commercial Cu/ZnO/Al₂O₃ catalyst [26]. The Cu-Al spinel oxide catalysts can be used for methanol reforming without pre-reduction by H₂, which not only avoids the copper sintering prior to the reaction, but also slows down the sintering rate during reaction. The small amount of CuO phase in the catalyst can initiate the reforming reaction at a low temperature, and Cu can be released gradually from CuAl₂O₄ to sustain the reaction after initiation. Moreover, the activity of the Cu-Al spinel catalyst can be further improved by adding excess amount of Al₂O₃, which can increase the surface area and CuAl₂O₄ crystal dispersion [26].

Other alkaline oxides such as CeO₂ and ZrO₂ can also be added into the Cu-based catalysts as promoters. The addition of ceria oxide is beneficial for maintaining copper dispersion and coke suppression due to its oxygen storage property [27]. Cu/CeO₂ catalysts have been proven to have superior activities for steam methanol reforming compared to industrial Cu/ZnO catalysts [28]. The addition of cerium to Cu/ZnO/Al₂O₃ catalysts can effectively improve the activity and hydrogen selectivity. In any case, the content of CO in the reformate can be kept low due to the suppression of methanol decomposition and reverse water-gas shift reactions [29]. Men et al. [30] reported that the activities of Cu/CeO₂/Al₂O₃ catalysts were dependent on the loading of Cu. They found that while the molar ratio of Cu and Ce in the catalyst increases from 0.1 to 0.9, the catalytic activity monotonously decreases. They attributed this to the fact that copper disperses preferentially on the ceria surface at low copper loading and the aggregation of copper particles occurs at high copper loading. Thermal desorption spectroscopy (TDS) results indicate that the decrease of copper loading and increase of ceria loading can enhance the basicity of the catalyst [31]. The methanol steam reforming reaction was believed to occur at the Cu/CeO₂ interface and the proposed reaction mechanism is shown in Figure 4 [30,32]. It includes four steps to fulfill a catalysis cycle, which are the adsorption of methanol and water at the Cu/CeO₂ interface, followed by the surface reaction and the oxygen reverse spillover (the migration of surface oxygen from CeO₂ to Cu), and, finally, the regeneration of partially oxidized copper and oxygen vacancies. ZrO₂ added to Cu/ZnO catalysts can also enhance the copper dispersion on the catalyst surface and thus lead to higher activity [33,34]. Moreover, the introduction of ZrO₂ into Cu/ZnO improves the reducibility and stability of the catalysts [35].

2.2. Performances of Copper-Based Catalysts

Pan and Wang [36] experimentally tested a commercial Cu/ZnO/Al₂O₃ catalyst in a compact plate-fin reformer. The catalyst was grounded and sieved to the particle size of 0.442–0.681 mm before loading. The feeding reactants (CH₃OH and H₂O) flow rate was kept at a constant of 5 mL/min. Different molar ratios of H₂O and CH₃OH between 1.2 and 1.6 were examined. The results showed that the content of H₂ in the dry reformate was 74.5% regardless of the ratio of H₂O/CH₃OH, whereas the content of CO decreased from 1.0% to less than 0.5% as the ratio of H₂O/CH₃OH increased from 1.2 to 1.6. Thus, complete methanol conversion can be achieved in all the tests.

Zhang et al. [37,38] investigated the methanol steam reforming performances of the commercial Cu/ZnO, Cu/ZnO/Al₂O₃ (CB-7), Cu/ZnO/CeO₂, Cu/ZnO/ZrO₂, and Cu/ZnO/CeO₂/ZrO₂ catalysts. The experiments were conducted in a fixed-bed quartz tube reactor with the gas hourly space velocity (GHSV) of 1200 h⁻¹. The ratio of H₂O and CH₃OH in the reactants was kept at 1.2. At the reforming temperature of 250 °C, the methanol conversion is 97%, 97%, 80%, 75%, and 68% over Cu/ZnO/ZrO₂, Cu/ZnO/CeO₂/ZrO₂, Cu/ZnO/CeO₂, Cu/ZnO/Al₂O₃, and Cu/ZnO catalysts, respectively. The corresponding CO concentration in the dry reformate is 0.8%, 0.55%, 0.25%, 1.8%, and 0.3%, respectively. The results also show that no noticeable deactivation was observed during a 360 h continuous operation over the Cu/ZnO/CeO₂/ZrO₂ catalyst.

Baneshi et al. [39] compared the methanol steam reforming performances over Cu/CeO₂/Al₂O₃, Cu/ZrO₂/Al₂O₃, and Cu/CeO₂/ZrO₂/Al₂O₃ catalysts. The catalysts were evaluated in a fixed bed reactor with a constant GHSV of 10,000 cm³ h⁻¹/g-catalyst. The ratio of H₂O and CH₃OH in the reactant flow was kept at 1.5. At the reforming temperature of 240 °C, the methanol conversion is 100%, 95% and 94% over the Cu/CeO₂/Al₂O₃, Cu/CeO₂/ZrO₂/Al₂O₃, and Cu/ZrO₂/Al₂O₃ catalysts, respectively. For Cu/CeO₂/Al₂O₃, the H₂ content in the dry reformate is around 65% and only a trace amount of CO is detectable. In a long run test, the activity of Cu/CeO₂/Al₂O₃ remains at 100% after 110 h of continuous operation.

Zeng et al. [40] examined the Cu/ZnO/ZrO₂/Al₂O₃ catalyst in cube-post micro reactors. The catalyst was prepared by co-precipitation method and coated in the plate by the wash coating method. The fed reactant mixture had a flow rate varying from 2 to 6 mL/h. The ratio of H₂O and CH₃OH was 1.3. The results indicate that the highest methanol conversion of 70.27% was achieved at 280 °C. The H₂ concentration in the reformate were around 74.4%, and the CO concentration was about 1%. The low methanol conversion in this study resulted from the microreactor structure. With optimized microchannel structure and dimensions, the methanol conversion can be improved to around 95% [41].

Jones et al. [42] studied the performance of a Cu/ZnO/ZrO₂/Al₂O₃ catalyst containing ZrO₂ from a monoclinic nanoparticle ZrO₂ precursor instead of a Zr(NO₃)₂ precursor. The catalyst was prepared by the co-impregnation method and its compositions were 15 wt.% CuO-15 wt.% ZnO-10 wt.% ZrO₂-60 wt.% Al₂O₃. The total reactants fed rate was 0.8 mL/h and the ratio of H₂O and CH₃OH was 3. At 305 °C, the methanol conversation rate can achieve 80%, and the corresponding H₂ content in the dry reformate is almost 80%. Only a trace amount of CO exists in the reformate.

Chang et al. [43] compared the oxidative steam reforming of methanol performances over the Cu/ZnO/ZrO₂/Al₂O₃ catalyst and the commercial Cu/ZnO/Al₂O₃ catalyst. The Cu/ZnO/ZrO₂/Al₂O₃ catalyst was prepared by the co-precipitation method. Under the reaction conditions of fed rate = 3.5 mL/h, H₂O/CH₃OH = 1.1, O₂/CH₃OH = 0.1, temperature = 250 °C, Cu/ZnO/ZrO₂/Al₂O₃ (20/50/20/10 wt.%) shows higher activity than the commercial Cu/ZnO/Al₂O₃ (30/60/10 wt.%) catalyst. Methanol conversion rates are 75.3% and 65.5% over the two catalysts, respectively.

Ahmadi et al. [44] investigated the nanostructured CuO/ZnO/ZrO₂/Al₂O₃ catalyst synthesized by the sonochemically-assisted co-precipitation method using different ultrasonic power. BET results show that higher ultrasonic power results in higher surface area and better activity. When the catalyst is synthesized and assisted by 90 W ultrasonic power, it can achieve 100% methanol conversion at a relatively low temperature of 200 °C. The H₂ and CO concentrations in the dry reformate is above 75% and below 1%, respectively. No deactivation is observed after continuously running for 24 h.

2.3. Palladium Based Catalysts

Iwasa et al. [45] first experimentally tested Pd/ZnO, Pd/Al₂O₃, and Pd/ZrO₂ catalysts for methanol steam reforming. The results show that the catalyst characteristics were greatly affected by the support. For catalysts with 1.0 wt.% Pd loading, Pd/ZrO₂ has higher dispersion (28.9%) than Pd/ZnO (10.7%) and Pd/Al₂O₃ (13.1%). However, the TOF is the higher for Pd/ZnO (0.829 s⁻¹) compared to Pd/ZrO₂ (0.148 s⁻¹) and Pd/Al₂O₃ (0.152 s⁻¹). Pd/ZnO also has the highest selectivity for the steam reforming reaction of methanol. Only Pd/Al₂O₃ gave dimethyl ether in addition to hydrogen, carbon dioxide and carbon monoxide in the reformate.

Further investigation concludes that the catalytic performance of Pd/ZnO for methanol steam reforming can be greatly improved by reducing the catalyst prior to the reaction. Iwasa et al. [46] proposed that the Pd-Zn alloy formed in hydrogen treatment has positive effects on the selectivity of methanol steam reforming. Formaldehyde species formed in the reaction were able to be effectively transformed into H₂ and CO₂ over the catalyst containing Pd-Zn alloys. In contrast, the formaldehyde species selectively decomposed to H₂ and CO over metallic-palladium-dominated catalysts without reducing treatment.

Chin et al. [47] confirmed the formation of the Pd-Zn alloy in the reduced Pd/ZnO catalyst by XRD, TPR and TEM characterization techniques. Figure 5 shows the high resolution TEM graph of a spent 16.7 wt.% Pd/ZnO catalyst. The presence of the Pd-Zn alloy was verified by the distinct d-lines of the small crystallites. It can also be concluded that Pd and Zn formed a highly ordered crystallites structure because the d-lines are continuous over an extended area as shown in Figure 5.

Chin et al. [48] also studied the effects of the preparation and pretreatment procedures on the Pd/ZnO catalyst. Using highly acidic Pd nitrate aqueous precursors such as the Pd(NO₃)₂ solution can alter the textural properties of the ZnO support including porosities and crystalline structures. The dissolution of ZnO into Zn²⁺ influenced the extent of mixing between Zn²⁺ and Pd²⁺ in pretreatment. As mentioned earlier, the formation of the Pd-Zn alloy was verified in the reducing pretreatment. Karim et al. [49] reported that the high temperature pretreatment (>300 °C) not only completely transformed Pd to Pd-Zn alloy, but also lead to particle growth. It is necessary to eliminate the small particles (<2 nm) because they lower the CO₂ selectivity of the Pd/ZnO catalyst.

XRD characterization of the spent catalyst indicates that Pd-Zn crystallites are strongly dependent on Pd loading, as shown by the catalytic performance [50]. Surface concentration of the Pd-Zn alloy increases as the Pd loading increases, but larger Pd-Zn crystallites resulting from higher Pd loading also inhibit the further increase of the surface concentration of the Pd-Zn alloy. Excess amounts of Pd loading could decrease the catalyst activity and increase the CO content in the reformate [51]. Liu et al. [52] suggested that 5 wt.% Pd loading is appropriate with regard to the catalytic activity. However, Pfeifer et al. [53] reported that the catalyst with 10 wt.% Pd loading showed the highest activity. Liu et al. [54] also studied the deactivation mechanisms of the Pd/ZnO catalyst. Two major routes that caused deactivation are carbon deposited on the catalyst surface and surface oxidation of the Pd-Zn alloy.

Pd/ZnO supported on alumina was also investigated for methanol steam reforming [55]. The activity of Pd/ZnO catalysts can be greatly improved by supporting them on alumina, and the Pd/ZnO/Al₂O₃ catalysts show better stability than the commercial Cu/ZnO/Al₂O₃ catalysts [56]. Zn-modified Pd catalysts supported on CeO₂ also show high selectivity for methanol steam reforming. XRD results indicate the formation of the Pd-Zn alloy, as well as the complete disappearance of the metallic Pd after reduction [57]. Methanol steam reforming performances over Pd/ZnO and Pd/CeO₂ catalysts were compared. The results show that Pd/ZnO has a lower methanol conversion rate and higher CO₂ selectivity than the Pd/CeO₂ catalyst. The average TOF at 230 °C for Pd/ZnO and Pd/CeO₂ was 0.8 ± 0.3 s⁻¹ and 0.4 ± 0.2 s⁻¹, respectively. Pd/ZnO has a higher density of acidic sites while Pd/CeO₂ has a higher density of basic sites [58].

Except for Zn, the catalytic functions of the Pd/Al₂O₃ and Pd/CeO₂ catalysts can be greatly modified by the addition of some other metals such as In and Ga due to the formation of Pd-In and Pd-Ga alloys [59,60,61]. Similarly to the Pd/ZnO catalysts, the alloys are responsible for the high CO₂ selectivity and the metallic Pd is responsible for CO selectivity. The ratios of Pd and the second metal as well as the loadings on the support have significant effects on the catalyst activity [62,63].

Pd/ZrO₂ catalysts were found to have high CO selectivity of approximately 50% at the reforming temperature of 260 °C [64,65]. Azenha et al. [64] first investigated the effect of Cu as a promoter of Pd/ZrO₂ catalysts. The results show that the addition of Cu into Pd supported on ZrO₂ with monoclinic crystalline structure leads to a remarkable decrease of the CO selectivity from 50% to 5%. SEM images shown in Figure 6 indicate the Cu and Pd particles are in intimate contact and highly dispersed in the support. The superior methanol steam reforming performance of the Cu-Pd/ZrO₂ catalyst is due to the strong interaction between Pd and Cu as well as the enhanced dispersion of the metal phase on the support.

2.4. Performances of Palladium Based Catalysts

Echave et al. [66] experimentally examined the methanol steam reforming performance of the 2.5 wt.% Pd/ZnO catalyst. The catalyst was prepared by incipient wetness impregnation of pre-synthesized ZnO with a palladium nitrate solution. The catalyst was evaluated at the reforming temperature of 350 °C with a reactant flow rate of 0.064 mL/min and a steam to carbon ratio of 1.5. No obvious deactivation was observed during a 4000 min continuous operation. Methanol conversion higher than 90% can be reached. H₂ and CO contents in the dry reformate are 70% and less than 1%, respectively. Sanz et al. [67] washcoated 2.5 wt.% Pd/ZnO catalyst in microchannel reformers, and similar results were reported.

Xia et al. [68] investigated the methanol steam reforming performance over a series of Pd/ZnO/Al₂O₃ catalysts. Under the reforming conditions of 220 °C, GHSV = 14,400 h⁻¹, and H₂O/CH₃OH = 1.78, the methanol conversion is 14.3% and 46.5% over Pd/ZnO (Pd loading 8.6 wt.%) and Pd/ZnO/Al₂O₃ (Pd loading 8.9 wt.%), respectively. Both catalysts achieve CO₂ selectivity higher than 99%. While elevating the reforming temperature to 270 °C, 100% methanol conversion is reached over the Pd/ZnO/Al₂O₃ catalyst. Pd loading and the Pd/Zn ratio were also varied and the performances were compared with respect to the methanol conversion and CO₂ selectivity. The results showed that the catalyst with a Pd loading of 8.9 wt.% and a Pd/Zn ratio of 0.38 had the highest methanol conversion and CO₂ selectivity. At the reforming temperature of 250 °C, 80% methanol conversion can be obtained over this catalyst.

Barrios et al. [69] tested the performances of Pd/CeO₂ and Pd/ZnO/CeO₂ catalysts for methanol steam reforming. Pd/CeO₂ can only decompose methanol to CO at the reforming temperature of 250 °C and the steam to carbon ratio of 1. The methanol steam reforming reaction only occurs at the presence of zinc oxide. At 300 °C, the methanol conversion rate is 65% and 95% over Pd/ZnO and Pd/ZnO/CeO₂ catalysts, respectively. Although the Pd/ZnO/CeO₂ catalyst significantly improves the methanol conversion compared to Pd/ZnO, its CO₂ selectivity (80%) is lower than that of Pd/ZnO (90%).

Men et al. [62] studied the methanol steam reforming performance of Pd-In/Al₂O₃ catalysts with different Pd loadings and Pd/In ratios. The results show that the catalyst with 7.5 wt.% Pd and 10 wt.% In₂O₃ has the overall best performance considering both the methanol conversion rate and the CO content in the reformate. About 90% methanol conversion and 0.6% CO concentration are achieved at the reforming temperature of 375 °C, with total flow rate of 30 mL/min and a steam to carbon ratio of 1.5.

Kuc et al. [70] experimentally investigated methanol steam reforming over perovskite-type oxides catalysts of LaCo_0.87Pd_0.13O_3±δ (denoted as LCPO) and LaCo_0.85Pd_0.075Zn_0.075O_3±δ (denoted as LCPZO). At the reforming temperature of 250 °C, with a total flow rate of 100 mL/min and a H₂O to CH₃OH ratio of 1.3, the methanol conversion rate is about 90% and 30% for LCPO and LCPZO, respectively. The corresponding CO₂ selectivity is 10% and 60%, respectively.

2.5. Reaction Mechanism of Methanol Steam Reforming

Regarding the reaction mechanism of methanol steam reforming over copper-based catalysts, four schemes have been proposed [7] including a methanol decomposition-water gas shift reaction scheme [71], a 1-step methanol steam reforming scheme [72,73], a methanol steam reforming-methanol decomposition-reverse water gas shift reaction scheme [74,75] and a methyl formate scheme [76,77,78]. The methanol steam reforming-methanol decomposition-reverse water gas shift reaction scheme is also adopted for methanol steam reforming over palladium catalysts [79,80].

The reaction scheme proposed by Peppley et al. [74,75] claimed the three reactions of methanol decomposition, water gas shift reaction, and methanol steam reforming all occur in parallel. This scheme has been supported and experimentally validated by different researchers [81,82,83,84,85,86,87]. The three parallel reactions are shown in Equations (1)–(3).

$${\mathrm{CH}}_3\mathrm{OH}+{\mathrm{H}}_2\mathrm{O}\underset{k_{-1}}{\overset{k_1}{\leftrightarrow }}3{\mathrm{H}}_2+{\mathrm{CO}}_2$$

(1)

$${\mathrm{CO}}_2+{\mathrm{H}}_2\underset{k_{-2}}{\overset{k_2}{\leftrightarrow }}{\mathrm{H}}_2\mathrm{O}+\mathrm{CO}$$

(2)

$${\mathrm{CH}}_3\mathrm{OH}\stackrel{k_3}{\to }2{\mathrm{H}}_2+\mathrm{CO}$$

(3)

As indicated above, both the methanol steam reforming and reverse water gas shift reactions are reversible, and the methanol decomposition reaction is irreversible. k₁, k₂ and k₃ are the forward rate constants, while k₋₁ and k₋₂ are the backward rate constants.

Mastalir et al. [82] evaluated the reaction rate coefficients of the above three reactions. The Arrhenius equations are shown in Equations (4)–(6).

$$r_1=k_1{C}_{{\mathrm{CH}}_3\mathrm{OH}}^{0.6}{C}_{{\mathrm{H}}_2\mathrm{O}}^{0.4}\exp \left(-\frac{E_{a1}}{RT}\right)-k_{-1}{C}_{{\mathrm{CO}}_2}{C}_{{\mathrm{H}}_2}\exp \left(-\frac{E_{a1}}{RT}\right)$$

(4)

$$r_2=k_2{C}_{{\mathrm{CO}}_2}{C}_{{\mathrm{H}}_2}\exp \left(-\frac{E_{a2}}{RT}\right)-k_{-2}{C}_{\mathrm{CO}}{C}_{{\mathrm{H}}_2\mathrm{O}}\exp \left(-\frac{E_{a2}}{RT}\right)$$

(5)

$$r_3=k_3{C}_{{\mathrm{CH}}_3\mathrm{OH}}^{1.3}\exp \left(-\frac{E_{a3}}{RT}\right)$$

(6)

Chein et al. [88] proposed a different model including the methanol steam reforming reaction, methanol decomposition reaction and water gas shift reaction as shown in Equations (7)–(9). All the reactions are assumed to be irreversible.

$${\mathrm{CH}}_3\mathrm{OH}+{\mathrm{H}}_2\mathrm{O}\to 3{\mathrm{H}}_2+{\mathrm{CO}}_2$$

(7)

$${\mathrm{CH}}_3\mathrm{OH}\to 2{\mathrm{H}}_2+\mathrm{CO}$$

(8)

$${\mathrm{H}}_2\mathrm{O}+\mathrm{CO}\to {\mathrm{CO}}_2+{\mathrm{H}}_2$$

(9)

Semi-empirical kinetic models proposed by Suh et al. [89] were used for the methanol steam reforming and decomposition reactions, which are,

$$r_R=\left(1-\epsilon \right){\rho }_s{k}_R{C}_{{\mathrm{CH}}_3\mathrm{OH}}$$

(10)

$$r_D=\left(1-\epsilon \right){\rho }_s{k}_D$$

(11)

where C is the molar concentration, ε is the catalyst layer porosity, and ρ_s is the catalyst density. The rate constants k_R and k_D can be expressed as,

$$k_R=5.5\left(1.15\times {10}^6+9.41\times {10}^5\ln \phi \right)e^{-E_R/RT}$$

(12)

$$k_D=5.5\times 7.09\times {10}^7{e}^{-E_D/RT}$$

(13)

where ϕ is the molar ratio of water to methanol, and E_R and E_D are the activation energies of the reforming and decomposition reactions, respectively.

Studies of Purnama et al. [90] and Chen et al. [91] were used for the water gas shift reaction. The reaction rate is expressed as,

$$r_{WGS}=11.2k_{WGS}\left(p_{\mathrm{CO}}{p}_{{\mathrm{H}}_2\mathrm{O}}-p_{{\mathrm{CO}}_2}{p}_{{\mathrm{H}}_2}/{\mathrm{K}}_{\mathrm{eq}}\right)$$

(14)

$$k_{WGS}=1.78\times {10}^{22}\left(1-0.154\delta +0.008{\delta }^2\right)T^8{e}^{-E_{WGS}/RT}$$

(15)

$${\mathrm{K}}_{\mathrm{eq}}=\exp \left(\frac{4577.8}{T}-4.33\right)$$

(16)

where k_WGS and K_eq are the rate constant and the equilibrium constant of the water gas shift reaction, respectively. p is the partial pressure of a species, and δ is the molar ratio of water to carbon monoxide.

3. Conclusions

Methanol steam reforming is a promising technology to produce hydrogen-rich syngas with only a trace amount of carbon monoxide for onboard fuel cell applications. The performance of methanol steam reforming is significantly dependent on the reforming catalyst. The present work reviewed the commonly used copper-based and palladium-based catalysts for methanol steam reforming. Compositional and morphological characteristics, along with the methanol steam reforming performances of the two groups of catalysts were summarized.

Concerning the copper-based catalysts, it is confirmed that the Cu/ZnO catalysts have high activity towards methanol steam reforming reactions. However, the mechanism of the promotion effect of zinc oxide is still under discussion. One widely accepted theory is that Cu-Zn alloys formed in the Cu/ZnO catalysts are responsible for the promotion effect of ZnO. Alumina can be added to the Cu/ZnO catalysts as a stabilizer to increase the durability and selectivity of hydrogen, although it slightly reduces the methanol conversion rate. Other alkaline oxides such as CeO₂ and ZrO₂ can also be added as promoters. The addition of ceria oxide is beneficial for maintaining copper dispersion and coke suppression. The catalytic activity and hydrogen selectivity of the Cu/ZnO catalysts are also enhanced by the addition of ceria oxide. The introduction of ZrO₂ into Cu/ZnO improves the reducibility and stability of the catalysts.

As for the palladium-based catalysts, the Pd/ZnO catalysts were found to have similar activities with the copper-based catalysts for methanol steam reforming. Moreover, they have higher carbon dioxide selectivity due to the formation of Pd-Zn alloys. The Pd-Zn crystallites and the catalytic performance of the catalysts are both strongly affected by the loading of Pd. An optimum range of Pd loading between 5% and 10% exists as reported in the literature. The activity of Pd/ZnO catalysts can be greatly improved by supporting them on alumina, and the Pd/ZnO/Al₂O₃ catalysts show better stability than the commercial Cu/ZnO/Al₂O₃ catalysts.

Several schemes have been proposed regarding the reaction mechanism of methanol steam reforming. The methanol steam reforming-methanol decomposition-reverse water gas shift reaction scheme is generally accepted for methanol steam reforming reactions occurring over the copper-based or palladium-based catalysts. In this scheme, the three reactions of the methanol decomposition, water gas shift reaction and methanol steam reforming are assumed to occur in parallel. Two different models based on this scheme were also presented in this work.