Carbon Supports Regulated the Performances of Copper Catalysts for H₂ Production through Microwave-Initiated Methanol Steam Reforming

Abstract

To tackle the storage challenges for distributed hydrogen applications, the microwave-initiated methanol steam reforming (MSR) process was deployed to increase the flexibility and energy efficiency of the MSR hydrogen production. As the key, a series of carbon-supported CuZn catalysts were systematically investigated to reveal the impacts of dielectric properties, conductivity, and dispersion on the catalytic performance under microwave irradiation. In addition to the excellent dielectric properties and conductivity exhibited by the carbon supports, the efficient dispersion of active species will dominate the catalytic activity, which was evidenced by the fact that the graphite supported CuZnG-40 is less active than the carbon nanotube supported and well dispersed CuZnCNTs-10, even with CuZnG-40 presenting superior conductivity and dielectric properties. The high specific surface area, tubular structure, and abundant pores of the carbon nanotubes could significantly promote the dispersion of the CuZn active species, forming highly dispersed small CuO particles. The H₂-TPR indicated that the reduction temperature of the carbon nanotube supported CuZn can be as low as 175 °C and it is easy to generate highly active Cu⁰/Cu⁺ sites. With a 1:1 (molar ratio) CH₃OH/H₂O feed and a weight hourly space velocity (WHSV) of 3 h^–1, the microwave-initiated methanol conversion of the CuZnCNTs-40 reached 90% with a relatively low CO content at 250 °C.

Introduction

Hydrogen is commonly used as fuel or chemical raw material due to its high mass-basis energy density and zero emissions during utilization. , However, one of the keys for building a public accessible hydrogen energy-driven society is safe and convenient storage and transportation for hydrogen. Currently, hydrogen is mainly produced from coal gasification, methane steam reforming, water electrolysis, alcohol reforming, etc. As cheap hydrogen production feedstock with wide sources, methanol is in the liquid state at room temperature and thus easy to store and transport. Methanol is mainly produced from synthesis gas and can also be produced via the direct oxidation of methane or through CO₂ hydrogenation with renewable hydrogen, which originates from water electrolysis powered by low-cost and abundant solar or wind energy far from end-users. − Steam reforming of lower-carbon alcohols is actually a typical process of releasing hydrogen temporarily stored in the alcohols produced from renewable resources. With much easier storage for liquid alcohols, the facile and efficient methanol steam reforming process of producing hydrogen thus could help to tackle the challenges in hydrogen storage and reduce CO₂ emissions for the energy industry. ,

Methanol steam reforming (MSR) (eq ) mainly involves methanol decomposition (eq ) and water–gas shift reaction (WGS) (eq ). , Compared with large-scale hydrogen production processes through coal gasification, natural gas reforming, and water electrolysis, hydrogen production through MSR features low reaction temperature and feasibility for implementation in miniaturized reforming systems. It is more suitable for distributed hydrogen utilization scenarios and helps to overcome the difficulties in hydrogen storage and transportation.

$${\mathrm{CH}}_3\mathrm{OH}+{\mathrm{H}}_2\mathrm{O}\to 3{\mathrm{H}}_2+\mathrm{CO}\mathrm{\Delta }{H}_{298\mathrm{K}}=+49.7\mathrm{kJ}/\mathrm{mol}$$ 1

$${\mathrm{CH}}_3\mathrm{OH}\to 2{\mathrm{H}}_2+\mathrm{CO}\mathrm{\Delta }{H}_{298\mathrm{K}}=+90.2\mathrm{kJ}/\mathrm{mol}$$ 2

$$\mathrm{CO}+{\mathrm{H}}_2\mathrm{O}\to {\mathrm{CO}}_2+{\mathrm{H}}_2\mathrm{\Delta }{H}_{298\mathrm{K}}=-41.2\mathrm{kJ}/\mathrm{mol}$$ 3

As the key to the MSR process, Cu-based catalysts are commonly used for their advantages such as low cost, high activity, and lower selectivity for the CO byproduct compared with precious metal catalysts. However, Cu-based catalysts have relatively poor thermal stability and are prone to sintering and deactivation. Therefore, they are often modified with other elements to promote the dispersion and reduction of Cu and further decrease the selectivity for CO. − Tajrishi et al. found that ZnO could improve the reducibility and dispersion of CuO. Meanwhile, ZnO can promote the adsorption of CH₃OH on Cu sites, thereby enhancing the activity of Cu-based catalysts. In addition, the types of supports and the synthesis methods of catalysts also significantly affect the dispersion of active components. − Shishido et al. compared two synthesis methods, namely homogeneous precipitation and coprecipitation, and found that the catalysts synthesized by the homogeneous precipitation method had smaller Cu/Zn particles, better dispersibility, and improved catalytic performance. Supports with relatively large specific surface areas, such as carbon nanotubes (CNTs), mesoporous carbon (MC), metal–organic frameworks (MOFs), and zeolite molecular sieves, can also improve the dispersion of Cu/Zn active species. − Shahsavar et al. prepared Ce and Zr modified CuZn/CNTs catalysts using CNTs as the support by a microwave-assisted synthesis strategy. Both CNTs and microwave treatment promoted the dispersion of Cu species and enhanced the methanol reforming activity. Apart from the catalysts, the energy transfer mode can also lead to significant differences in reaction pathways and product distributions. Hsu et al. conducted comparative studies on the exothermic reaction process of CO₂ hydrogenation to methane under the traditional heating and microwave-driven conditions. The thermodynamic analysis results showed that microwave irradiation could significantly decrease the spontaneity of this exothermic reaction and strongly suppress the formation of CH₄.

Microwave heating is a process in which microwaves interact with the molecules of materials and are directly converted to thermal energy. Compared with the traditional heating, the selective and volumetric heating of microwaves can significantly increase the heating speed and decrease the overall reaction temperature. − Meanwhile, it does not need the preactivation of the catalysts using external hydrogen sources. − Araia et al. utilized the microwave driven process for ammonia synthesis. The overall ammonia synthesis temperature was significantly reduced from 400–500 °C to 150–300 °C, achieving rapid start-up and shut-down for the reaction. The above advantages of the microwave-driven processes show great potential for improving the performance of catalytic reactions such as for the hydrogen production by MSR. In microwave-driven catalytic processes, the catalyst’s ability to absorb microwaves and convert them into thermal energy greatly affects the reaction efficiency. Normally, carbon nanomaterials or carbides that could absorb and convert microwaves, such as carbon nanotubes (CNTs), graphite (G), conductive carbon black (CB), silicon carbide (SiC), mesoporous carbon (MC), or activated carbon, are used as supports to enhance the microwave absorption performance of catalysts. − Attributed to the rich pore structures of carbon nanomaterials, the polarization loss and the conductive network can be enhanced or optimized. In addition, these structures may facilitate multilevel reflection and scattering for the microwaves, thereby enhancing microwave absorption and conversion. Moreover, the abundant surface structures of carbon materials can also help to anchor more active species and thus improve their dispersion to provide more active sites for reactions. Li et al. used KOH to regulate the porosity and specific surface area of corncob charcoal and studied the changes in the electromagnetic wave absorption ability of the composite material of this corncob charcoal and nanonickel. The results indicated that corncob charcoal with a large specific surface area and small pore size could induce multiple reflections and scatterings of microwaves and thus enhance the microwave absorption ability. However, changes in the surface and pore structures of carbon materials are accompanied by their graphitization degree. A lower degree of graphitization will weaken the penetration of microwaves into carbon materials, resulting in a poor microwave absorption ability. In addition, electrical conductivity, as another important parameter reflecting the electromagnetic characteristics of carbon materials, is strongly and positively correlated with their degree of graphitization. Good electrical conductivity of catalysts can promote electron migration, which is beneficial for improving the activity of redox reactions such as MSR. , Zhong et al. investigated the microwave absorption abilities of softwood treated at different carbonization temperatures. The results revealed that the carbon formed by high-temperature carbonization obtained higher graphitization, electrical conductivity, and microwave absorption ability, while its specific surface area decreased significantly. The above studies indicate that the degree of graphitization, electrical conductivity, and microstructure of carbon materials are complexly correlated and significantly affect their microwave absorption and response abilities. In a microwave catalytic system with carbon materials as supports, factors such as physical properties, structures of carbon materials, active components, and interactions between the active components and carbon supports have a more complex impact on the performance of microwave catalysts, which needs further exploration to guide the rational design of catalysts used in microwave-initiated reactions.

Therefore, a series of CuZn-supported catalysts were rationally prepared using carbon materials or carbides such as CNTs, CB, G, and SiC as supports, and the hydrogen production performances of these catalysts in the microwave driven MSR process were subsequently evaluated in this work. The influences of supports on the dispersion of active components, as well as the effects of supports’ electrical conductivity, dielectric properties, and degree of graphitization on the microwave absorption, conversion, and catalytic performance of the catalysts, were investigated, trying to reveal the synergy and effects lying in the dielectric properties, electrical conductivity, and microstructures of the catalysts under microwave fields and providing guidance for the design and performance regulation of catalysts applicable to the microwave-initiated MSR reaction process.

2. Experimental Section

2.1. Materials and Catalyst Preparation

The catalysts were prepared by an impregnation method. The general procedures for each preparation are as follows: (1) First, certain amounts of Cu­(NO₃)₂·3H₂O and Zn­(NO₃)₂·6H₂O with a molar ratio of 1:1 were dissolved in 50 mL of deionized water. (2) With the rational designed support content in the catalysts, the corresponding amounts of the microwave-sensitive supports such as SiC, CNTs, G, and CB were added to the above-mentioned precursor solution. Subsequently, the mixture was ultrasonicated for 1 h and magnetically stirred for 3 h to form homogeneous slurry and then dried in oven at 100 °C. (3) The dried slurry was calcined at 400 °C for 2 h with a ramping rate of 5 °C/min in a N₂ atmosphere. (4) After calcination, the samples were naturally cooled and collected for use.

In this study, SiC, G, Cu­(NO₃)₂·3H₂O, and Zn­(NO₃)₂·6H₂O were purchased from Shanghai Xianding Biotechnology Co., Ltd. The carbon nanotubes (CNTs) were purchased from Chengdu Institute of Organic Chemistry (Chinese Academy of Sciences), and CB was purchased from Cabot Chemical (Tianjin) Co., Ltd. The physicochemical properties of SiC and various carbon material supports are listed in Table . The catalyst is designated as “CuZn + support type + support mass content”. For example, a copper–zinc catalyst with a CNT support content of 10 wt % and the rest of CuO/ZnO formed by the thermal decomposition of Cu­(NO₃)₂·3H₂O and Zn­(NO₃)₂·6H₂O precursors in a N₂ atmosphere is designated as CuZnCNTs-10. In each preparation, 5 g of the final catalyst was rationally designed, and the corresponding usage of Cu­(NO₃)₂·3H₂O, Zn­(NO₃)₂·6H₂O, and the support was theoretically calculated. Taking the CuZnCNTs-10 as an example, Cu­(NO₃)₂·3H₂O, Zn­(NO₃)₂·6H₂O, and CNT support used in the above-mentioned procedures (1) and (2) were 6.75, 8.32, and 0.50 g, respectively. For comparison, the unsupported CuZn catalyst (with a Cu/Zn molar ratio of 1:1) was prepared with the above-mentioned general procedures and used as the control catalyst.

1. Physicochemical Properties of the Supports.

support	SBET (m2/g)	purity	specifications	pore volume (cm3/g)	pore width (nm)
CNTs	350.85	>99.9%	diameter <8 nm, length: 10–30 μm	0.076	1.340
SiC	9.04	≥99.0%	particle size: 0.5–0.7 μm	0.003	1.688
CB	75.12	≥99.5%	∼10,000 mesh	0.038	2.040
G	24.73	≥99.9%	5000 mesh	0.011	2.647

2.2. Catalyst Characterization

Thermal gravity-differential thermal gravity analysis (TG-DTA) was conducted using a NETZSCH STA449F5 simultaneous thermal analyzer (Germany) to investigate the stability of the carbon supports. The samples were heated from room temperature to 800 °C with a 10 °C/min ramping rate in a flow of air (20 mL/min), and the weight of the samples was simultaneously monitored by the analyzer. X-ray diffraction (XRD) spectra of the samples were collected on a Bruker D8 ADVANCE X-ray diffractometer using Cu Kα radiation in the 5–90° range (2θ) with a scanning rate of 10°/min. N₂ physical adsorption isotherms were collected using an Autosotrb-IQ2-MP-XR physical adsorption analyzer to obtain the BET surface areas, BJH pore volumes, and pore size distributions of the samples. Prior to analysis, the samples were outgassed under a vacuum at 300 °C for 3 h. Temperature-programmed hydrogen reduction (H₂-TPR) and desorption (H₂-TPD) were tested with a ChemStar-TPx chemisorption analyzer from Quantachrome Instruments. The pretreatment procedures for both H₂-TPR and H₂-TPD were the same. Typically, 50–100 mg of samples was treated in 50 mL/min Ar flow at 300 °C for 1 h and cooled to room temperature. For the H₂-TPR tests, the pretreated sample was then heated to 600 °C with a 10 °C/min ramping rate in 10 vol % H₂/Ar flow, with the signal of the consumed hydrogen collected by the thermal conductivity detector (TCD). Meanwhile, for the H₂-TPD test, the pretreated sample was switched to the 10 vol % H₂/Ar flow and kept for 1 h to achieve saturated H₂ absorption. After the absorption, the system was purged with Ar for 1 h and then the samples were heated to 800 °C at a 10 °C/min ramping rate under Ar flow, and the signal of the desorbed H₂ was recorded by the TCD. To further investigate the dispersion of the active species, the CO pulse adsorption experiment was conducted on a Micromeritics AutoChem II 2920 chemisorption analyzer with the operating conditions as follows: the samples were first reduced in H₂ at 300 °C for 1 h and then cooled to 50 °C with high-purity He purging to obtain stable baseline, then a CO pulse with 3 min intervals was introduced, and pulse absorption curves were collected. The X-ray electron spectroscopy (XPS) of the catalysts before and after reaction was performed on a Thermo ESCALAB 250 with Al Kα radiation (hv = 1486.6 eV) as the excitation source, and charge correction was carried out with C 1s (284.80 eV) as the binding energy standard. Peak fitting was performed using the Avantage software. The dielectric properties of the samples were tested by using the coaxial method with an Agilent E5071C vector network analyzer. Typically, the sample powder was mixed evenly with the same weight of the melted paraffin wax and then poured into a mold to form a ring after cooling, and then the ring was placed on the sample holder to carry out the dielectric property collection. The microscopic morphology of the catalyst was obtained by a Zeiss Gemini 300 scanning electron microscope with a 1 kV acceleration voltage. The electrical conductivity of the sample powder was measured with the four-probe method on a ST2742B powder resistivity analyzer (Suzhou Lattice Electronics Co., Ltd.). For each test, 0.2 g of sample powder was taken and placed in a sample chamber, and then, the resistance of the compressed sample power in the 0–30.0 MPa pressure range was simultaneously collected with 2.0 MPa intervals.

2.3. Catalytic Performance Testing

The MSR performance of the catalyst was evaluated by using a vertical microwave tubular reactor, and the process is shown in Figure . The reactor is equipped with a built-in single-mode magnetron that emits microwaves at a fixed frequency of 2.45 GHz, while the microwave power can be flexibly tuned within the range of 0–2000 W. In this study, a constant microwave power of 300 W was used throughout all experiments, with the reforming temperature being simultaneously monitored and controlled by a silver-coated K-type thermocouple, which was placed in the middle of the catalyst bed. The temperature control system was programmed to automatically pause the microwave radiation when the catalyst bed temperature reached the setting temperature and restart it when the temperature dropped below the set point, ensuring the reaction temperature stably maintained at the testing temperature throughout the experiments. The gaseous products from the reforming reactions were analyzed using a GC9790Plus gas chromatograph (Fuli Instruments).

1.

Schematic diagram of the microwave-initiated MSR reaction process: 1 - feed liquid; 2 - constant-pressure flow pump; 3 - microwave tubular reactor; 4 - gas-washing apparatus; 5 - gas chromatograph; 6 - gas cylinder.

Typically, 1.50 g of the catalyst was loaded into a quartz tube with an inner diameter of 12 mm, and the catalyst bed was centered at the middle of the vertical tubular microwave reactor. After the system was purged with Ar, a methanol/water solution with a molar ratio of 1:1 was pumped directly into the reactor at a weight hourly space velocity (WHSV) of 1 h^–1 using a constant-flow pump without any inert carrying gas. Then, microwave irradiation was turned on to heat the catalyst to 250 °C for in situ catalyst activation using the directly injected methanol/water feedstock solution. After the catalyst was heated to 250 °C and activated for 15 min, the microwave irradiation was paused to allow the catalyst bed temperature to drop below 190 °C. Subsequently, the microwave was restarted to keep the catalyst bed at the testing temperature. The microwave-driven MSR performance of each catalyst was investigated at 190, 210, 230, 250, and 270 °C, and the reaction was maintained for 2 h at each temperature. For measuring and analyzing the flow rate and composition of the reformed gas product every 30 min, a gas flowmeter and a calibrated gas chromatograph were used, respectively. Then, methanol conversion, product selectivity, and yield were calculated according to the following equations. First, the molar flow rate of each component (n _i) in the gas product was calculated by the ideal gas law with the volume concentration determined by gas chromatography (eq ). Then, the molar flow rate of converted methanol (n _{CH3OH, converted}) was derived based on carbon balance (eq ). Combined with the feed molar flow rate of methanol, the methanol conversion (X _CH3OH), H₂ selectivity (S _H2 ), and yield (Y _H2 ) were subsequently calculated (eqs –). To further explore the effect of WHSV on the catalysts’ performances, the 40 wt % carbon support loaded catalysts were systematically tested at 250 °C with WHSVs of 1, 2, 3, 4, and 5 h^–1.

$$n_{\mathrm{i}}=PV{C}_{\mathrm{i}}/RT$$ 4

$$n_{{\mathrm{CH}}_3\mathrm{OH},\mathrm{converted}}=n_{{\mathrm{CH}}_4}+n_{\mathrm{CO}}+n_{{\mathrm{CO}}_2}$$ 5

$$X_{{\mathrm{CH}}_3\mathrm{OH}}=n_{{\mathrm{CH}}_3\mathrm{OH},\mathrm{converted}}/n_{{\mathrm{CH}}_3\mathrm{OH},\mathrm{in}}\times 100\%$$ 6

$$S_{{\mathrm{H}}_2}=n_{{\mathrm{H}}_2}/(3\times n_{{\mathrm{CH}}_3\mathrm{OH},\mathrm{converted}})\times 100\%$$ 7

$$Y_{{\mathrm{H}}_2}=X_{{\mathrm{CH}}_3\mathrm{OH}}\times S_{{\mathrm{H}}_2}=n_{{\mathrm{H}}_2}/(3\times n_{{\mathrm{CH}}_3\mathrm{OH},\mathrm{in}})\times 100\%$$ 8

where C _i is the volume concentration of each component in the gas product stream determined by the calibrated gas chromatography (i = H₂, CH₄, CO, CO₂) and n _CH3OH, in is the methanol molar flow rate in the methanol/water feedstock stream.

3. Results and Discussion

3.1. Evaluation of Catalytic Performance

Under a WHSV of 1 h^–1, the variation of the methanol conversion and the compositions (CH₄, CO, CO₂, and H₂) of the reforming product stream for each catalyst at different temperatures are shown in Figure . As illustrated in Figure a, the methanol conversion of all catalysts increases with the rising temperature. Compared with the unsupported CuZn catalyst (control catalyst), the carbon-supported catalysts exhibit a more pronounced increase in activity. However, the enhancement varies with the different carbon-supported CuZn catalysts. For example, the methanol conversion of the unsupported CuZn at 270 °C is only 66.1%, while those of the CuZnG-10 and CuZnSiC-10 increase to 77.5% and 86.7%, respectively, and the CuZnCNTs-10 and CuZnCB-10 can achieve near complete methanol conversion. The content of support significantly impacts the catalytic activity. Within the 190–270 °C range, the activity of the catalysts at a 40 wt % support content is generally higher than that of the catalysts with the same support at a 10 wt % content. For the CNT and CB supported catalysts with a 40 wt % support content (CuZnCNTs-40 and CuZnCB-40), near complete conversion of methanol can be achieved at only 230 °C. Meanwhile, for the 40 wt % SiC and G supported catalysts, although the activities were significantly improved compared to those of their 10 wt % supported catalysts, their conversions are still much lower than those of the CuZnCNTs-40 and CuZnCB-40. Compared with the SiC and G supports, the CNTs and CB are more suitable for supporting the CuZn active species to provide much higher activity and help to decrease the methanol steam reforming temperature.

2.

(a) Variation of the methanol conversion and (b) compositions of the reforming gas product with different reaction temperatures.

Figure b illustrates the variation of the gas product compositions over the catalysts synthesized with different reaction temperatures. As the reaction temperature gradually increases from 190 to 270 °C, the CO content in the product gas stream rises. This might be attributed to the following: (1) the endothermic methanol decomposition reaction (eq ) proceeds more forward with rising temperature, thereby releasing more CO; (2) the rise in temperature activates the reverse water–gas shift reaction (RWGR), which tends to generates more CO. Although the CuZn catalysts supported on CNTs and CB can enhance the activity of CuZn catalysts, the fluctuations of the CO byproduct caused by the carbon supports are not negligible. For the CuZn, CuZnG, and CuZnSiC catalysts, the increase in the CO content is not remarkable and it remains below 1.8 vol % within the temperature range of 190–270 °C. Thus, with the above-mentioned performance results, the catalysts with 40 wt % support were selected and tested at 250 °C (typical catalyst bed temperature profile as Figure S1 in the Supporting Information) in the subsequent experiments to conduct further investigation for achieving both enhanced methanol conversion and a relatively low CO content in the gas product.

To assess the stability of the carbon supports, thermogravimetric analysis (TGA) of the carbon supports in air was performed, and the results are shown in Figure . The oxidation temperatures of CNTs, CB, G, and their supported catalysts all exceed 500 °C, which are all significantly higher than the MSR operating temperature. Since the oxidation ability of H₂O is much weaker than air, suggesting that the carbon supports will be stable and not be gasified to generate CO and H₂ under the MSR conditions and thus could serve as supports for the microwave-initiated MSR reactions. However, compared to natural graphite, artificially synthesized carbon materials often have surface defects. When microwaves irradiate carbon materials with surface defects, the interface polarization effect generates local ″hot spots″ with temperatures higher than the overall system temperature, which lasts only a fraction of a second. These hot spots often promote heterogeneous reactions. ,

3.

(a) TG and (b) DTG profiles of the carbon-supported CuZn catalyst and its supports

At 250 °C, the variations of the reforming performance of CuZnG-40, CuZnCB-40, and CuZnCNTs-40 with different space velocities were investigated, and the results are illustrated in Figure . Within the range of WHSV from 1 to 5 h^–1, both the methanol conversion and the CO content show a decreasing trend as the space velocity increases, which might be attributed to the shortening of contact time between the catalyst and reactants, resulting in a large amount of feedstock accumulating on the catalyst surface and directly passing through the microwave tubular reactor without reacting. It is noteworthy that the methanol conversion of CuZnCNTs-40 decreases by only 12.4%, and its activity still remains at around 95% when the space velocity is 3 h^–1. However, the methanol conversions of CuZnCB-40 and CuZnG-40 decrease by about 50% under the same conditions. Within the space velocity range of 1–5 h^–1, the CO contents of CuZnCNTs-40, CuZnCB-40, and CuZnG-40 all exhibit a decreasing trend. Among which, the CuZnCNTs-40 exhibits the best reforming performance and presents a CO content reduction to 1.07 vol % when the space velocity increases from 1 to 3 h^–1. Through comprehensive comparisons, it is revealed that CuZnCNTs-40 exhibits superior initial performance in microwave-driven MSR for hydrogen production under high space velocities (with the stability of CuZnCNTs-40 illustrated in Figure S2, Supporting Information). Though all the carbon materials (CNTs, CB, and G) exhibited good enough microwave response ability, their impacts on the microwave reforming activity of the supported CuZn catalyst vary significantly. These differences will be analyzed in detail, focusing on the microwave absorption characteristics of carbon supports and their impact on the CuZn active species.

4.

(a) Methanol conversion and (b) gas product compositions at 250 °C under different space velocities for CuZnCNTs-40, CuZnCB-40, and CuZnG-40.

3.2. Influence of Catalyst Dielectric Properties on Reaction Performance

Dielectric properties significantly affect the microwave absorption and conversion capabilities of the materials. The microwaves absorbed by materials are mainly converted into two forms of energy: (1) the electromagnetic field guides an induced current inside the material, and an internal electric field is established to store the absorbed microwave in the form of electric potential energy; (2) heat is generated due to the dielectric loss and is used to maintain the temperature of the reaction system. The dielectric properties of materials are measured by complex permittivity ε, as shown in eq :

$$\epsilon ={\epsilon }^{\prime }-j{\epsilon }^{″}$$ 9

Among which, ε′ denotes the dielectric constant, representing the ability of the material to generate an internal built-in electric field under microwave induction and store the absorbed microwaves in the form of electric potential energy in the material. In contrast, the imaginary part ε″ represents the loss factor, quantifying the ability of the material to convert the absorbed microwaves into heat, where j is the imaginary unit and j ² = −1. To accurately evaluate the wave-absorbing ability of materials, the loss tangent tan δ (loss coefficient) is usually used as the key parameter. The larger the tan δ, the stronger the ability of the material to convert the absorbed microwaves into heat, as illustrated in eq :

$$\tan \delta ={\epsilon }^{″}/{\epsilon }^{\prime }$$ 10

The dielectric property parameters such as the dielectric constant ε′, dielectric loss factor ε″, and loss coefficient tan δ of each catalyst are measured by the vector network analyzer and shown in Figure . The results indicate that each dielectric property parameter generally decreases with the increase of the test frequency. As the microwaves that can be used for the civil industry are generally two frequencies, which are 2.45 and 915 MHz, the dielectric property parameters recorded at 2.45 GHz are of certain significance.

5.

(a) Dielectric constant; (b) loss factor; (c) loss coefficient; (d) heating rate curves.

As compared, it is revealed that there is no significant improvement in the dielectric properties of the catalysts when the addition of each support is just 10 wt %. At the working frequency of 2.45 GHz, the ε′ and ε″ of the unsupported CuZn catalyst are 3.11 and 0.38, respectively, and tan δ is only 0.12. After adding 10 wt % carbon supports, only the CuZnG-10 presents a relatively obvious change in the dielectric property parameters (ε′ = 4.53, ε″ = 0.82, tan δ = 0.18). However, when the support content increases to 40 wt %, the dielectric property parameters of each catalyst at 2.45 GHz change very significantly. Especially, for the CuZnCNTs-40 and CuZnCB-40, their ε″ reach 45.36 and 19.36, while their tan δ increase to 0.83 and 0.71, and the dielectric constants of the CuZnCNTs-40 and CuZnCB-40 are 78.18 and 34.60 times of the catalysts with 10 wt % of the corresponding carbon support, respectively. Moreover, the dielectric loss coefficient tan δ of the CuZnCNTs-40 and CuZnCB-40 are 6 and 4.32 times those of the corresponding catalysts with 10 wt % carbon supports, respectively. The results reveal that the dielectric properties of the catalysts could be significantly improved with an increase in the carbon support content.

Figure d shows the microwave heating characteristics of each catalyst under the same conditions (microwave power is 300 W and catalyst dosage is 1.50 g per batch). The results indicate that the heating rate of the catalysts is highly consistent with their dielectric property parameters. Taking the heating of the catalyst bed to 250 °C as the setting limit, the CuZnCNTs-40, which has the optimal dielectric properties, only needs 1.3 min to reach 250 °C, while the CuZnCB-40 requires 2 min, and CuZnG-40, with the worst dielectric properties, needs 6.5 min. Moreover, the heating rate of CuZnG-40 shows a significant downward trend when the temperature rises above 150 °C. However, as mentioned above, 10 wt % of the support does not significantly improve the dielectric properties of the catalysts. The CuZnG-10 catalyst, which has the optimal dielectric properties among the catalysts with 10 wt % supports, only reaches 160 °C after heating for 12 min. Based on the comprehensive results of Figures and , it can be concluded that the reforming performance of the catalysts is generally consistent with the quality of their dielectric properties. It is attributed to the improvement of the dielectric properties enhancing the absorption and conversion of microwaves by the catalysts, which is extremely beneficial to the endothermic MSR reaction. The results further demonstrate that CNTs, CB, and G are highly beneficial for improving the dielectric properties of the CuZn catalyst and enhancing the hydrogen production performance of the MSR.

Nevertheless, it should be emphasized that the values of ε″ and tan δ for CuZnG-40 are 5.98 and 0.39, respectively, which are significantly higher than those of CuZnCNTs-10. In line with this, the heating performance of CuZnG-40 under microwave irradiation is also superior to that of CuZnCNTs-10. However, as demonstrated by the reforming performances (Figure ), the microwave reforming activity of the CuZnG-40 is remarkably lower than that of CuZnCNTs-10. The reforming performance is apparently contradictory to the observations from the dielectric properties of CuZnG-40 and CuZnCNTs-10. It clearly indicates that, though the dielectric properties of the catalysts are extremely important, they are not the primary determinants of the overall catalytic performance of the catalysts in the microwave-initiated reforming process.

3.3. Influence of the Conductivity of Catalysts on the Reaction Performance

Microwaves are electromagnetic waves. Under the influence of the microwave electric field, any mobile charges within a material are induced to oscillate back and forth, thereby generating an electric current. This motion of the charges causes collisions of the charges with adjacent molecules or atoms, giving rise to a heating effect. Notably, materials with excellent electrical conductivity (σ) make a significant contribution to the imaginary part of the complex permittivity ε″ at microwave frequencies through the relationship ε″ ≈ σ/(ωε₀), where ω is the angular frequency and ε₀ is the vacuum permittivity, respectively. The enhanced ε″ directly translates into a more efficient microwave energy absorption and heat generation. Moreover, apart from the heat generation mechanism, the electromagnetic field induced current formed can directly traverse the active sites inside the catalyst, facilitating rapid electron transfer between reactants and catalytic centers. , To validate the role of electrical conductivity of the catalysts, the resistivities of both the supports and catalysts under a pressure range of 2–30 MPa were measured using the four-probe method.

As illustrated in Figure , a conspicuous trend emerges wherein the resistances of the supports and catalysts exhibit a sharp decline with increasing pressure. In the context of the reaction system, the catalyst bed assumes a naturally compacted state. Consequently, the resistivity value obtained under a minimum test pressure of 2 MPa could be used as a representative metric to gauge the electrical conductivity of the bed. In its prereduced state, CuZn exists in an oxidized form and exhibits insulating properties. CNTs, CB, and G all exhibit excellent electrical conductivity, with the maximum resistivity being only 3.0 × 10^–3 Ω·m. It is noteworthy that the electrical conductivity of the catalysts increases in parallel with the increased support contents. Among the various catalyst series with the same support content, the CNT supported CuZn (CuZnCNTs) catalysts manifest the highest electrical conductivity level, followed by the CuZnCB series. The graphite supported CuZn catalyst series demonstrates the poorest conductive performance in this regard. Particularly, when the support content reaches 40 wt %, the resistance of CuZnG-40 is still approximately 27.75 Ω·m, while those of the CuZnCNTs-40 and CuZnCB-40 are lower than 0.02 Ω·m. The CuZnCNTs-40 and CuZnCB-40 are much more conductive than the CuZnG-40. Thus, it could be well explained that the H₂ yields of the CuZnCNTs-40 and CuZnCB-40 are significantly higher than that of CuZnG-40. It should be attributed to the fact that the highly conductive networks composed of CNTs and CB enable the electromagnetic field induced current to pass through the catalyst rapidly, effectively facilitating the electron migration and the collision combination of electrons and H⁺, thus accelerating the formation and desorption of H₂. Meanwhile, for the comparison of CuZnG-40 and CuZnCNTs-10, though the former exhibits slightly better electrical conductivity than the latter, the activity of the latter catalyst is higher in the microwave-initiated MSR reactions (Figure a). It is suggested that the electrical conductivity is also not the dominant factor for enhancing the microwave-initiated MSR performance when the catalysts reaches a sufficiently conductive level.

6.

Variation of resistivity of different supports and catalysts with pressure.

The dielectric properties and electrical conductivity of the supported CuZn catalysts primarily originate from the carbon supports, and these properties are closely associated with the structural features of the carbon supports. Especially, the conductivity of the carbon supports exhibits a significant positive correlation with the graphitization of the carbon supports, while the graphitization degree directly depends on the structural order of the carbon supports. Thus, the CNTs, CB, and G were characterized by XRD and Raman spectroscopy, and the results are illustrated in Figure. . By combination of the Bragg formula (eq ) and the Franklin formula (eq ) as well as the ratio of the D peak to the G peak, the graphitization degrees of the carbon supports were calculated and compared. As shown in Table , a larger g value or a smaller I _D/I _G value indicates a higher graphitization degree of the carbon support. Figure a presents that the natural graphite has the sharpest peak and its half-peak width is the smallest when compared with the artificially synthesized CNTs and CB, representing the high crystallinity of graphite. According to the g value, the G exhibits the highest graphitization degree, followed by CNTs and CB in turn. The Raman spectroscopy can also verify the graphitization degree of the carbon supports. As shown in Figure b, the D peak and G peak of the carbon supports appear at 1350 and 1570 cm^–1, respectively. According to the calculation of the peak intensity ratio I _D/I _G, the I _D/I _G ratio of CB is the largest at 0.89, followed by 0.57 for CNTs and 0.21 for G, respectively. This result is consistent with that of XRD analysis.

$$d(002)=\lambda /2\sin \theta$$ 11

$$g=(0.344-d)/(0.344-0.335)\times 100\%$$ 12

7.

(a) XRD and (b) Raman spectra of the carbon support.

2. Graphitization Degree of Carbon Supports.

support	2θ (°)	d002 (nm)	g (%)	ID/IG
CNTs	26.37	0.338	66.667	0.57
CB	25.90	0.344	0.000	0.89
G	26.56	0.335	100	0.21

3.4. The Effect of Dispersion on the Catalytic Performance

Besides the positive impacts imposed by the carbon supports on the dielectric properties and conductivity, dispersion of the active species is also important. Though the Cu-based catalysts have the advantages of low cost and good selectivity, Cu nanoparticles are prone to sintering and agglomeration due to their relatively low Tammann temperature. Improving the dispersion of CuZn species can ensure its long-term stability. ,− In the microwave-driven MSR system, BET, XRD, and XPS were utilized to study the dispersion effect of different carbon supports on the CuZn species.

Figure and Table illustrate the results of the N₂ adsorption–desorption and pore size distribution tests on the catalysts. The adsorption–desorption isotherms of CuZnCNTs and CuZnCB exhibit typical Type IV characteristics and are accompanied by a remarkable H3 hysteresis loop, indicating that the catalysts possess a certain pore volume, with the pore channel structure presenting an irregular morphology. In contrast, the isotherms of the CuZnG, CuZnSiC, and CuZn samples conform to Type II characteristics, suggesting that there are basically no obvious pores within the catalysts. Meanwhile, as shown in Table , when the contents of CNTs, CB, and G are increased from 10 to 40 wt %, the specific surface area of the CNT supported catalysts increases from 27.24 to 150.77 m²/g, and the specific surface area of the catalyst with CB as the support increases from 15.66 to 48.64 m²/g, and that of the catalyst with G as the support only increases from 4.63 to 15.78 m²/g. Obviously, the relatively large specific surface areas of CNTs and CB provide favorable conditions for the better dispersion of CuZn active components.

8.

(a) N₂ adsorption–desorption isotherms of the carbon-supported CuZn catalysts; (b) pore size distributions of the carbon-supported CuZn catalysts.

3. Catalyst Specific Surface Area, Pore Volume, and Pore Width Data.

item	SBET (m2/g)	pore volume (cm3/g)	pore width (nm)
G	24.73	0.011	2.647
CNTs	350.85	0.076	0.076
CB	75.12	1.340	1.340
CuZn	27.49	0.011	2.419
CuZnCNTs-10	27.24	0.107	28.746
CuZnCNTs-40	150.77	0.480	14.160
CuZnCB-10	15.66	0.077	40.540
CuZnCB-40	48.64	0.167	26.460
CuZnG-10	4.63	0.004	36.000
CuZnG-40	15.78	0.030	2.897

The SEM images, TEM images, and particle size distributions for the catalysts with 10 and 40 wt % support are shown in Figure . With 10 wt % support, the CuZn particles formed on the surface of CNTs are smaller than those of the CB and G, with the dimension in the 30–55 nm range. Granular CuZn active species particles are also formed on the surface of CB. While, the CuZn species supported on the G exhibit a massive coral-like structure, with their particle sizes distributed between 95 and 110 nm. With the support content increasing to 40 wt %, fine particles of the CuZn active species are formed on the CNT surface, with the particle size distribution mainly concentrated in the range of 9–12 nm. Small CuZn particles are also formed on the CB surface, with the particle size mainly concentrated in the 25–35 nm range. Meanwhile, for the CuZnG-40, the CuZn active species still maintain a coral-like morphology, and the graphite support is stacked in flakes, with the particle size distribution concentrated in 40–100 nm, indicating that it is still difficult to ensure the uniform dispersion of CuZn active species with a large amount of graphite support. The above results reveal that the large specific surface area and abundant pores of CNTs can more effectively promote the dispersion of the CuZn active species compared with CB and the tightly stack-layered graphite, which allows more active sites to be exposed on the CNT-supported catalysts surfaces, thereby effectively improving the catalytic performance. The active species dispersion determined by the CO pulse adsorption is shown in Table . The results further evidence that with the same support content, the dispersion of the active species supported on the CNTs is significantly higher than those of on CB and G.

9.

(a) SEM, (b) TEM, and (c) particle size distributions of the catalysts.

4. Dispersion of Metal Cu in Different Catalysts.

catalyst	Cu metal dispersion
CuZn	4.61%
CuZnCNTs-10	14.64%
CuZnCB-10	11.78%
CuZnG-10	6.45%
CuZnCNTs-40	20.64%
CuZnCB-40	16.46%
CuZnG-40	8.30%

The XRD patterns of each catalyst shown in Figure could also further confirm the significant dispersion effect of the CNTs on the CuZn active components. As seen in Figure , the diffraction peak intensities of the CuZn active species (mainly CuO and ZnO) continuously decrease with an increase in the support contents. Compared with the catalysts supported on the G support, the diffraction peak intensities of CuO and ZnO supported on the CNTs and CB are significantly weakened and their half-peak widths are obviously increased. In particular, the characteristic diffraction peaks of CuO at 35.60° and 38.96° in the CuZnCNTs-40 almost disappear, indicating the high dispersion of CuZn species on the carbon nanotube support, which is highly consistent with the BET and SEM results presented above. In addition, the XRD patterns of CuZnCNTs-40 and CuZnG-40 after the reforming reaction show that the CuO species are transformed into metallic Cu species (Cu⁰/Cu⁺), and the characteristic peak intensity of Cu⁰/Cu⁺ in CuZnCNTs-40 is very low and broad, suggesting that the reduced Cu⁰/Cu⁺ species in the used catalysts also maintain much better dispersion on the CNT support.

10.

(a) XRD spectra of support-10; (b) XRD spectra of support-40 before and after the reaction.

As the active centers, the reduced Cu species (Cu⁰ and Cu⁺) are originated from the reduction of CuO. , H₂-TPR was employed to investigate the reduction and activation of each catalyst. Figure demonstrates that the reduction temperature of the CuZn catalyst decreases significantly with the increase in the CNT content, and the reduction peak temperature of the CNT-supported catalyst is significantly lower than those of the other catalysts. Among which, the reduction peak of CuZnCNTs-40 drops to below 230 °C, presenting a main-peak and a secondary-peak structure at 177 and 230 °C, respectively, representing two different dispersion states of the CuO. Moreover, the reduction peak at 177 °C corresponds to the highly dispersed CuO with small particle sizes, while the reduction peak at 233 °C represents the CuO with larger particle sizes. ,− The results of the H₂-TPR test are consistent with those of the XRD and SEM.

11.

H₂-TPR plots for each catalyst.

In addition, XPS is used to analyze the valence state distribution of the Cu element before and after the reaction for the CuZnCNTs-40, CuZnCB-40, and CuZnG-40 catalysts. As illustrated in Figure , the binding energy of the Cu 2p_3/2 peak decreases after the reforming reaction, indicating that the Cu²⁺ is reduced to Cu⁰/Cu⁺ species during the reaction. , Table shows that the peak area ratio of the Cu⁰/Cu⁺ species to the Cu²⁺ species for CuZnCNTs-40 after the reaction is 1.19, which is significantly higher than those of CuZnCB-40 (0.88) and CuZnG-40 (0.77). The results suggest that the CNT support is beneficial to the reduction and activation of the Cu active species. Compared with other supports, CNTs exhibit enhanced capability to form highly dispersed CuZn active particles and promote easier reduction and activation of the active species, thereby providing more active sites and higher microwave-initiated MSR activity. , Thus, CuZnCNTs-40 presented the strongest catalytic activity. Even at a low temperature of 190 °C (as supplemented in Figure S3, Supporting Information), the methanol conversion over the CuZnCNTs-40 can reach 56%, which is much higher than the maximum methanol conversion of 23% for other catalysts at the same temperature. As recalled by the previous MSR performance results, the higher dispersion of the CuZn active species should be the main reason that the CuZnCNTs-10 obtained higher microwave-initiated MSR activity than CuZnG-40, even though the CuZnCNTs-10 presented poorer dielectric properties and conductivity than the CuZnG-40 catalyst.

12.

(a) XPS spectra of Cu 2p binding energies for CuZnCNTs-40, CuZnCNTs-40, and CuZnG-40 catalysts before the reaction; (b) XPS spectra of Cu 2p binding energies for CuZnCNTs-40, CuZnCNTs-40, and CuZnG-40 catalysts after the reaction.

5. XPS Peak Results of Cu Elements for Each Catalyst.

items	CuZnCNTs-40	CuZnCB-40	CuZnG-40
(Cu0/Cu+)/Cu2+ratio	1.19	0.88	0.77

In addition to the enhanced dispersion of active species, the adsorption and desorption effects of the supports could also impose a significant impact on the reaction performance of the catalysts. As the H₂-TPD curves for the carbon-supported catalysts illustrate in Figure , the H₂ desorption temperatures can be divided into three ranges: >500 °C, 250–500 °C, and <250 °C. Here, the H₂ desorption peak located in the >500 °C range is attributed to the dissociative hydrogen spillover from the larger-sized CuO particle surface to the ZnO promoter and then desorption on its surface. , It should be noted that, compared with CuZnCNTs-10, a H₂ desorption peak within the 250–500 °C range for the CuZnCNTs-40 appears, and it is attributed to the adsorption and activation effect of CNTs on hydrogen, which could promote the hydrogen spillover from the CuO active sites to the adsorption sites on the CNT surface. The high electrical conductivity could accelerate the electron migration along the walls of CNTs, facilitating the reduction of H⁺ to form H₂ and then desorption. − The positive hydrogen spillover effect of the CNTs helps to accelerate the surface dehydrogenation rate in the MSR reaction and thus improve the methanol conversion.

13.

H₂-TPD plots for each catalyst.

4. Conclusions

This work investigates the microwave-initiated catalytic performance of the CuZn catalysts supported on different types of carbon supports for methanol steam reforming. The influence of the carbon support’s dispersion ability, dielectric properties, and electrical conductivity on hydrogen production performance was systematically studied. The results reveal that the incorporation of carbon supports in the CuZn catalysts is beneficial for enhancing dielectric properties and conductivity, thereby facilitating the intensification of microwave-initiated MSR for hydrogen production. With adequate microwave absorption ability and conductivity, the dispersion of active species will be the critical factor determining catalytic performance, which was evidenced by the fact that CuZnCNTs-10 exhibited superior performance than CuZnG-40 even though the former is less conductive than the latter. With the highly dispersed CuZn species and the reducible CuO particles (with the reduction temperatures as low as 175 °C) on the CNT surface, the conductive CNTs with larger specific surface area, rich pore structures, and excellent microwave absorption ability could help to provide more active sites for enhancing the microwave-initiated MSR catalytic activity. With comprehensive consideration of the dielectric properties, conductivity, and dispersion effects, the CNT support would be optimal among the tested carbon supports. With the 1:1 (molar ratio) methanol/water feedstock being fixed at a WHSV of 3 h^–1, the CuZnCNTs-40 catalyst achieved a methanol conversion of 90% at 250 °C, with CO content controlled below 2 vol %, demonstrating excellent microwave-initiated methanol steam reforming performance.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.5c02772.

• Additional experimental details including typical catalyst bed temperature profile under microwave irradiation, the optimal catalyst stability test, and the performance comparisons of the catalysts between conventional heating and microwave heating (PDF)

#.

W.L. and Y.S. contributed equally.

The authors declare no competing financial interest.