Cu and Cu-Fe Bi-Metal Nanoparticles Encapsulated in Hollow S-1 Zeolite for Reverse Water Gas Shift Reaction

Abstract

The hollow hierarchical structure Cu@S and CuFe_0.5@S catalysts were successfully synthesized through the “dissolution-recrystallization” (D-R) method for the reverse water gas shift reaction (RWGS). The encapsulated catalysts had a hierarchical porous structure and better dispersion of Cu particles than the Cu-S and CuFe_0.5-S samples prepared via the conventional impregnation method. Furthermore, CuFe_0.5-S and CuFe_0.5@S catalysts showed higher CO₂ conversion and 100% selectivity of CO at the entire temperature range investigated in this work compared to the monometallic catalysts Cu-S and Cu@S. Interestingly, the reaction activity of all the samples increased according to the sequence: CuFe_0.5@S > CuFe_0.5-S > Cu@S > Cu-S at 400–550 °C under atmospheric pressure. These results indicate that the higher dispersion of encapsulation structure and the enhanced surface basicity derived from the addition of Fe play crucial roles in enhancing the catalytic performance of Cu-based catalysts in the RWGS reaction.

1. Introduction

The capture, conversion and utilization of CO₂ have been popular issues which play important roles in reducing greenhouse gas (GHG) emissions for environmental protection and achieving carbon cycling for the development of the recycling economy [1,2]. One potential CO₂ utilization pathway is the reverse water gas shift (RWGS) reaction (CO₂ + H₂ = CO + H₂O), which contributes to the conversion of the highly stable CO₂ molecule into an active CO molecule. As the key building block, CO can be transformed into light olefins, long-chain hydrocarbons, methanol and other high-value-added oxygenates [3]. The RWGS reaction is an endothermic reaction that transforms CO₂ into CO at atmospheric pressure and high temperature (>300 °C) [4,5]. Taking into account the reaction pressure, there is no doubt that achieving efficient conversion of CO₂ at low pressure will be a considerable advantage. In addition, the RWGS reaction is a heterogeneous catalytic reaction. An ideal design of a catalyst system considers both excellent catalytic performance and fair cost to play an important role in large-scale CO₂ conversions. Among the many catalyst systems, Cu-based catalysts have been studied the most in CO₂ hydrogenation reactions for their low cost and optimal catalytic activity [6]. Jurovi’c et al. [7] reported that Cu-based catalysts are sensitive to their structural characteristics by establishing the exposed Cu sites and RWGS reaction rate. B.M. Weckhuysen et al. [8] have also revealed that the activity of Cu catalysts was related to increased metal dispersion, higher Cu⁰: CuO ratios and adequate surface basicity sites. Moreover, deactivation is closely related to metal sintering under a high reaction temperature. In this regard, the optimization of the catalyst structure via adding textural promoters [9,10], such as Fe or Zn or confining the metal nanoparticles in inner channels or pores of porous supports, helps to resist aggregation resulting in the high stability of catalysts [11,12].

The metal@zeolite catalysts have attracted much interest due to their superior enhancement of the catalytic performance in many fields in recent years [13,14]. Wang et al. [10] reported a maximum activiation selectivity of 45% over the hollow hierarchical S-1 zeolite encapsulated PtNi nanoparticle catalyst PtNi@S-1, synthesized by the dissolution recrystallization (D-R) method. The results showed that not only the selectivity of the products but also the stability of the PtNi@S-1 was significantly higher than that of the PtNi/S-1 catalyst synthesized via the conventional impregnation method. Kwok et al. [15] synthesized the Fe@h-ZSM-5 catalyst by the DR method and used it for the Fischer–Tropsch reaction. It was found that the Fe@h-ZSM-5 catalyst exhibited significant advantages, such as excellent activity, outstanding coking and a sinter-resistance ability over the Fe/ZSM-5. Ma et al. [16] researched the catalytic performance in dry reforming of methane with high loading and uniform Ni nanoparticles encapsulated in the Silicalite-1 zeolite of which the sizes of Ni NPs were maintained at ca.4–5 nm and the TOF was kept at 60 s⁻¹ (800 °C) by increasing the Ni loading from 3% to 20%. The evolution results exhibited that the catalysts with various Ni loadings all exhibited high stability upon running for 150 h. In addition, Kang et al. [17] reported a sodium-modified rhodium nanoparticles embedded in the Silicalite-1 zeolite catalyst Na-Rh@S-1 and revealed that the catalyst Na-Rh@S-1 with the encapsulated structure exhibited a higher dispersion of Rh nanoparticles and more outstanding activity for the CO₂ hydrogenation to ethanol reaction.

Although metal@zeolite catalysts showed excellent catalytic performance in many fields, the research into Cu embedding in zeolites was relatively less than other transition metals. In addition, there have been rare reports related to Cu confined in the ZSM-5 zeolite catalysts for C1 chemistry, especially for RWGS reaction. Therefore, in this work, in order to adequately integrate the unique Cu-based catalyst for the RWGS reaction with the promoter effect of Fe and the confinement effect of metal@zeolite, the Cu and Cu-Fe bi-metal nanoparticles embedding in the hollow S-1 zeolite were prepared via the D-R method, which was denoted as Cu@S and CuFe_0.5@S. Meanwhile, the catalytic activity over different catalysts was investigated. The influence of the hollow encapsulated structure on the CO₂ conversion and CO selectivity were emphatically focused on.

2. Results and Discussion

2.1. Textural Characteristics and N₂ Physisorption

The pore structure properties of the parent S-1 zeolite and the catalysts were characterized with N₂-BET. As shown in Figure 1A, the parent S-1, Cu-S and CuFe_0.5-S all showed the same type of isotherms, which indicated that, the loading of active metals has little effect on the structure of the silicalite-1 zeolite. For Cu@S and CuFe_0.5@S, the type-IV isotherms with an H2 hysteresis loop, which consists of an abrupt step at around P/P₀ = 0.45, were characteristic of the cavitation phenomenon or pore-blocking effect [18,19]. This feature implies that there existed an ink-bottle pore system in the samples, which means that the zeolite has cavities connected to the outside through narrow channels (˂4 nm) [19,20]. In addition, for all the samples, the steep increase in adsorption quantities and the hysteresis curves at high relative pressure P/P₀ > 0.9 demonstrated the existence of mesopores derived from the aggregation of zeolite crystals [21]. The NLDFT pore size distribution curves presented in Figure 1B further confirmed that the Cu@S and CuFe_0.5@S catalysts possessed a hierarchical porous structure, and the size of mesopores generated from TPAOH treatment was concentrated at 2~4 nm. Except for the parent S-1, Cu-S and CuFe_0.5-S samples, the catalysts Cu@S and CuFe_0.5@S mainly presented microporous structures and a small quantity of mesoporous structures can be observed at 2~4 nm, which were inter-crystalline mesopores generated by the aggregation of the zeolite crystals.

The textural properties of all samples which were derived from the N₂-BET experiment were summarized in

Table 1. Textual Properties of Parent S-1 Zeolite and Series Catalysts.

Samples	SBET a (m2/g)	Smicro b (m2/g)	Smeso/ext c (m2/g)	Vtotal d (cm3/g)	Vmicro b (cm3/g)
Parent S-1	390.6	356.2	34.4	0.238	0.174
Cu-S	342.6	310.2	32.4	0.220	0.154
Cu@S	336.9	291.2	45.7	0.260	0.151
CuFe0.5-S	340.6	305.1	35.5	0.218	0.152
CuFe0.5@S	369.2	325.5	43.7	0.256	0.158

Notes: ^a Determined by the BET method; ^b Calculated by t-plot method; ^c Calculated using “S_BET-S_micro”; ^d Results Taken from the N₂ adsorbed at P/P₀ = 0.95.

. It can be observed that the parent S-1 possessed the largest specific surface area (390.6 m².g⁻¹) and microporous specific surface area (356.2 m².g⁻¹). After impregnation, the specific surface and the total volume of Cu-S and CuFe_0.5-S decreased slightly since the metal deposited in the micropores and on the external surface of catalysts. In addition, the specific surface of Cu@S decreased by 1.67% compared to the Cu-S sample due to its smoother crystal surface caused by the recrystallization [22]. Compared to Cu-S, the Cu@S exhibited both a decrease (6.13%) in the microporous specific area and an increase (41.11%) in the external specific area, which suggests the formation of mesopores during the encapsulation process [20]. However, the specific area of CuFe_0.5@S increased by 8.41% compared to CuFe_0.5-S, which was contrary to the previous situation. It was speculated that the Fe originally deposited in the micropores and external surface entered the skeleton of zeolite during the TPAOH treatment process, which was consistent with the result that CuFe_0.5@S exhibited larger microporous (increased by 6.69%) and external specific areas (increased by 23.17%). These results further confirmed that Cu@S and CuFe_0.5@S possess a hierarchical porous structure, which were in agreement with the experimental results reported for hollow MFI zeolites [20,23,24]^.

2.2. XRD and TEM Experiments

The XRD data obtained for all samples are shown in Figure 2. All the samples exhibited characteristic diffraction peaks at 2θ = 7.96°, 8.82°, 23.08°, 23.97° and 24.43°corresponding to the (011), (200), (501), (033) and (133) facets of the Silicalite-1 zeolite (PDF#44-969) [15]. To characterize the change in relative crystallinity, the 2θ = 22.5°~25°of the standard ZSM-5 sample was selected as reference, and the relative crystallinity of all the samples is calculated as follows and shown in Figure 2A. The parent S-1 presented a high relative crystallinity close to the standard sample. However, the other catalyst samples presented a lower crystallinity, indicating that some zeolite frameworks were destroyed during the process of loading metals. In addition, there was little change in the relative crystallinity among the different catalysts. A closer observation of the XRD patterns (Figure 2B) showed that all catalysts possessed the peaks at around 2θ = 35.5°, 38.7° and 48.8° which were consistent with the (002), (111) and (−202) facets of the CuO phase (PDF#89-5895). After that, the other two low-intensity peaks at 37.7° and 42.61°might be attributed to the Cu₂O phase (PDF#34-1354). No obvious diffraction peaks from iron oxide were observed due to the low content. The average size of CuO for all the samples was calculated based on the CuO (111) facet using Scherrer’s equation. Therefore, the results of Cu crystal sizes for Cu-S, Cu@S, CuFe_0.5-S and CuFe_0.5@S are 22.6 nm, 17.6 nm, 23.2 nm and 21.7 nm, respectively, demonstrating that the catalyst encapsulated structures have lower particle sizes. Compared to Cu-S, the addition of Fe had no obvious effects on the CuO size. Nevertheless, it might be the association of Cu and Fe oxide during the encapsulation process, because CuFe_0.5@S displayed a larger particle size compared to Cu@S.

Figure 3 shows the TEM micrographs obtained from the fresh catalysts. The images of Figure 3A–D, respectively, correspond to the Cu-S, CuFe_0.5-S, Cu@S and CuFe_0.5@S samples. Compared to the encapsulated structural catalysts, Cu-S and CuFe_0.5-S did not display visible metal oxide nanoparticles uniformly distributed on the zeolite frameworks, which might be attributed to aggregation of metal particles under the calcination process of the samples. Nevertheless, Cu@S and CuFe_0.5@S showed obvious uniform nanoparticles embedded in the cavity of the zeolite, which was consistent with the N₂-BET results. The CuO nanoparticle size of the Cu@S and CuFe_0.5@S were 12.83 nm and 17.7 nm, respectively, resulting from counting at least 80 particles of several TEM images, which were slightly smaller than the results (17.6 nm for Cu@S and 22.2 nm for CuFe_0.5@S) calculated via XRD. Additionally, the metal dispersion (D) was calculated from the equation $D=\frac{k}{d}$ according to the methodology proposed in [25] by using the relationship between the number of surface (n_s) and total (n_t) atoms. The coefficient k can be expressed via the equation [5] $k=\frac{M.n_s}{\rho .N_A}$, where M is the molar mass, $\rho$ accounts for the density, N_A denotes Avogadro’s number and d is the particle size. Based on the above equation, the dispersions for Cu-S, CuFe_0.5-S, Cu@S and CuFe_0.5@S were 7.96%, 7.74%, 14.63% and 10.38% which were summarized in

Table 2. Metal loading and Dispersion.

Samples	Metal loading a/%		Cu/Si b		XPS/XRF	dCu (nm)	DCu (%)
	Cu	Fe	XPS c	XRF
Cu-S	5.16		0.032	0.052	0.62	22.6	7.96
Cu@S	5.64		0.021	0.058	0.36	12.83 ± 3.35	14.63
CuFe0.5-S	5.26	0.53	0.048	0.054	0.89	23.2	7.74
CuFe0.5@S	6.82	0.70	0.037	0.072	0.51	17.74 ± 5.87	10.38

Notes: ^a Calculated by XRF; ^b Molar ratio of Cu to Si; ^c Determined using Cu/Si = (I_Cu/f_Cu)/(I_Si/f_Si), where I is the peak area of XPS and f is the atom sensitivity of factor. In this work, f_Cu = 5.321, f_Si = 0.328.

. The chemical compositions of samples, analyzed by XRF and XPS, are also listed in Table 2. XPS is a surface technique with an analysis depth of ˂10 nm and XRF can analyze the bulk composition, therefore, XPS/XRF can be used to characterize the distribution of Cu on the sample surface and inside the cavity of the zeolite [23,26]. By comparing the results of XPS/XRF, the encapsulated structural samples Cu@S and CuFe_0.5@S had smaller ratios than the Cu-S and CuFe_0.5-S samples, which further confirms the Cu species were successfully encapsulated into the S-1 zeolite. These results are consistent with the TEM images.

2.3. H₂-TPR and CO₂-TPD Experiments

The reducibility of different catalysts was exhibited in Figure 4A. All the samples present obvious reduction peaks near 200 °C, which could be ascribed to the characteristic peak of CuO. The difference was that the peaks of Cu@S and CuFe_0.5@S shifted to higher temperatures compared to Cu-S and CuFe_0.5@S, which might be due to the Metal-Si strong interaction resulting from the encapsulated structure catalysts that suppressed the reduction of the metal oxide [18,19]. Additionally, other reducibility peaks of Cu at higher temperatures could be attributed to the presence of highly dispersed Cu₂O species. For the Cu-Fe bi-metal catalysts, the peak between 300 °C and 400 °C is the reduction peak of FeO_x.

CO₂-TPD profiles were utilized to underline the surface basic characteristics of catalysts by the temperature and intensity of the desorption peaks which were shown in Figure 4B. The temperature and intensity of the desorption peak underline the surface basic characteristics of catalysts. In general, a higher the desorbed temperature corresponds to a stronger basicity of the catalysts. For instance, the lower temperature (˂150 °C), the moderate temperature (150~450 °C) and the higher temperature (>450 °C) could be, respectively, related to weak, moderate and strong basic sites in the samples [5]. The specific results of peak fitting by the deconvolution of curves based on a Gaussian function are summarized in

Table 3. The results of fitting peaks of different catalysts.

Samples	Weak		Moderate		Strong
	Peak 1 /°C	Peak 2 /°C	Peak 3 /°C	Peak 4 /°C	Peak 5 /°C	Peak 6 /°C
Cu-S	81.07 (20.52%)	91.43 (14.52%)			494.52 (59.65%)	582.32 (5.31%)
Cu@S	92.30 (15.59%)	126.95 (30.69%)			545.29 (53.72%)
CuFe0.5-S		101.85 (2.89%)			540.58 (40.27%)	685.23 (56.84%)
CuFe0.5@S	73.11 (2.92%)		163.69 (8.67%)	235.84 (22.74%)	535.59 (31.13%)	643.56 (34.54%)

Notes: The number in the bracket represents the percentage of area quantity.

. For the lower temperature peaks, it can be speculated that the peaks at below 100 °C should be attributed to the physisorption of the CO₂ molecule’s interaction with the porous catalysts, which can be explained by the fact that the CO₂ hydrogenation reaction over Cu-based catalysts at low temperature has no activity. Compared Cu-S with Cu@S, the proportions of different types of basic sites were similar, which is the reason that the change of texture structure has little effect on the surface basicity of catalysts. However, introducing the Fe promoter changed the basicity of the catalyst, which can be seen from the increased number of strong basic sites in the CO₂-TPD results. Furthermore, compared to the CuFe_0.5-S catalyst, the CuFe_0.5@S sample exhibited superior amounts of moderate basic sites which were related to improved capacities of adsorbing and activating the highly stable CO₂ molecules and decomposition of intermediate species into reaction products achieving the RWGS reaction rate [5].

2.4. CO₂ Hydrogenation Activity

Figure 5 shows the RWGS catalytic performance for catalysts with different configurations under the conditions of 17.8 L/h and H₂/CO₂ = 3/1. Figure 5A displays the generally increasing conversion of CO₂ as the temperature rises from 400 °C to 550 °C. Furthermore, the reaction activity of all the samples increased according to the sequence: CuFe_0.5@S > CuFe_0.5-S > Cu@S > Cu-S. Comparing the CO₂ conversion on Cu-S and Cu@S catalysts, it can be speculated that the higher Cu dispersion and hierarchical porous structure contributed to enhancing the catalytic performance resulting from the high activity of nanoparticles and better diffusion performance of the mesoporous structure. Surprisingly, the catalysts containing the Fe agent exhibited significantly better performance than the Cu-S and Cu@S samples in both the CO₂ conversion and CO selectivity (as shown in Figure 5B). In fact, the TOF values of Cu-S, CuFe_0.5-S, Cu@S and CuFe_0.5@S (3.68 × 10⁻³, 5.17 × 10⁻³, 1.79 × 10⁻³ and 3.13 × 10⁻³, respectively) were estimated at the initial temperature of 400 °C. It can be inferred that the catalytic enhancement is related to the introduction of Fe, which is reported [5] to be associated with the presence of Cu-FeO_x domains with improved CO₂ dissociation abilities. CO₂-TPD results indicated that the increased surface basic sites are related to the Fe presence. Furthermore, it was reported [27] that the Cu-FeO_x interface contributed to achieving the RWGS reaction rates by increasing the electron density on Cu metal interfaces. The CO selectivity results displayed in Figure 5B demonstrated that the Cu-based catalyst has superior performance on the RWGS reaction than other the catalytic series, which the selectivity of Cu-S and Cu@S was 90.94% and 87.82% at 400 °C, respectively. These results further illustrated the advantage of samples containing FeO_x species over non-Fe Cu catalysts.

Overall, the Cu-based catalysts with hollow S-1 zeolite structure and Fe promoter indeed contributed to the RWGS reaction. The encapsulation structure has a better Cu dispersion and a better diffusion of molecules, which is related to the reaction activity. The incorporation of FeO_x species, either in co-impregnated or embedded in zeolite support, significantly enhances the transformation of CO₂ into a goal product which is attributed to the unique electronic effect and structure effect of Cu-FeO_x. In addition, the moderate surface basicity sites displayed in the CO₂-TPD results of CuFe_0.5@S further engaged the catalyst with a better catalytic performance over the entire temperature range.

In order to explore the stability of the Cu-based catalysts with different structures, the change in CO₂ conversion with time under the same experimental conditions was used as an indicator of the catalyst stability. The results are shown in Figure 6. With the extension of time, the CO₂ conversion of different catalysts was significantly different. The stability of Cu-S and Cu_0.5-S catalysts prepared by the impregnation method was lower than that of Cu @ S and CuFe_0.5-S catalysts encapsulated by the hierarchical porous zeolite during the whole evaluation period of 120 h. It fully showed that the confinement effect of the zeolite on the metal nanoparticles can effectively improve the stability of the catalysts. In addition, comparing the catalyst Cu-S with CuFe_0.5-S, it is not difficult to find that the activity of Cu_0.5-S was significantly higher than that of Cu-S within 50 h before the reaction because the Fe promoter can improve the RWGS reaction activity of copper-based catalysts. In general, the stability of the encapsulated catalyst was better than that of the supported catalyst prepared by the impregnation method, which was consistent with the initial idea of the confinement effect of zeolite metal nanoparticles.

3. Experimental Section

3.1. Materials

Tetrapropylammonium hydroxide solution (TPAOH, 25%), ethyl orthosilicate (TEOS, AR), Cu(NO₃)₂ 3H₂O (AR) and Fe(NO₃)₃ 9H₂O were purchased from Shanghai Aladdin Bio-Chem Technology Co., Ltd., Shanghai, China.

3.2. Catalysts Preparation

3.2.1. Synthesis Parent S-1 Zeolite

Silicalte-1 zeolite was synthesized by the clear solution method with the molar ratio of 1SiO₂:0.3TPAOH:37H₂O. Typically, 10.00 g TEOS, 14.29 g TPAOH and 18.21 g deionized water were mixed in a 100 mL beaker under stirring at room temperature for 6 h. Then the above mixture was transferred to a 100 mL Teflon-lined steel autoclave for crystallization. After that, the product was obtained via centrifugation and dried overnight at 105 °C. Finally, the zeolite sample was calcined in a muffle furnace at 550 °C for 2 h to remove the organic template.

3.2.2. Preparation of Cu-S and CuFe_0.5-S Catalysts

Cu-S and CuFe_0.5-S catalysts were synthesized by an impregnation method. Briefly, the parent S-1 zeolite was impregnated with Cu(NO₃)₂ solution and Cu(NO₃)₂ -Fe(NO₃)₃ co-solution. After the impregnation process, the samples were dried at 105 °C for 12 h, and then calcined at 450 °C for 4 h. The obtained catalysts were named Cu-S-1 and CuFe_0.5-S-1, respectively. The ideal Cu, Fe loading was 5% and 0.5%, respectively.

3.2.3. Preparation of Cu@S and CuFe_0.5@S Catalysts

Cu@S and CuFe_0.5@S catalysts were synthesized by a dissolution-recrystallization (DR) method [18,20,28]. In brief, the as-prepared Cu-S-1 and CuFe_0.5-S catalysts were treated with TPAOH solution at 170 °C for 24 h and the molar ratio of SiO₂: TPAOH: H₂O was 1:0.25:22. After being cooled to room temperature, the products were recovered by centrifugation and washed at pH = 7.0 with deionized water. Then the samples were dried overnight at 105 °C and calcined at 450 °C for 4 h. The final obtained samples were donated as Cu@S and CuFe_0.5@S and the ideal Cu, Fe loading was 5% and 0.5%.

3.3. Catalysts Characterization

The textural properties of different configurational catalysts were measured through N₂ physisorption, which was performed on the Micro-meritics ASAP 2460 instrument. Before the experiment, all samples were degassed under a vacuum for 12 h. The Brunauer-Emmett-Teller (BET) method was applied for calculating the total surface area. The micro- and meso- were calculated via the t-plot method. In addition, the pore volume and average pore sizes of catalysts were calculated through the Barrett-Joyner-Halenda (BJH) model. Powder X-ray diffraction (XRD) patterns were collected on an EMPYREAN powder diffractometer (PANalytical Corporation, Netherland) using Cu Ka radiation (λ = 0.154 nm) at a scanning rate of 2°/min from 20° to 70°.

The elemental chemical compositions of configurational catalysts were measured using the X-ray Fluorescence (XRF) spectrometer (ZSX-Ⅱ, Rigaku). The X-ray photoelectron spectroscopy (XPS) characterizations were carried out on a Thermo ESCALAB 250XI spectrometer equipped with an Al Ka X-ray radiation source and each element was calibrated via C 1s peak at 284.8 eV.

The morphology of all the samples and the dispersion of metal nanoparticles were measured via scanning electron microscopy (SEM) images (HITACHI, S-4800) and transmission electron microscopy (TEM) images (FEI, Tecnai G2 F20), respectively.

The reducibility of samples was characterized through a H₂ temperature programmed reduction (H₂-TPR) method conducted on a Micromeritics AutoChem II 2920 analyzer suited with a thermal conductivity detector (TCD). In total, 100 mg of the catalysts (40~60 mesh) were used in every H₂-TPR experiment. The samples were pretreated at 200 °C for 60 min with Argon (Ar) flow to remove water and other wastes. After that, the samples were heated from room temperature to 900 °C at the rate of 10 °C/min in a 10% H₂-Ar flow (50 mL/min) and the reduced data were recorded with TCD. In addition, the basicity of catalysts was also analyzed in a Micromeritics AutoChem II 2920 chemical adsorption instrument by a CO₂-temperature programmed desorption (CO₂-TPD) method. In this regard, 200 mg of the catalysts (40~60 mesh) were per-reduced at 600 °C for 2 h with H₂. After cooling to 50 °C, the samples were saturated by CO₂ gas with a flowing rate of 50 mL/min followed by sweeping with flowing Ar at the same temperature for 30 min to remove physically adsorbed CO₂. The desorption procedure was carried out between 50 and 800 °C at a rate of 10 °C/min under Ar flow (50 mL/min) and the desorption curve was also recorded with TCD.

3.4. Catalytic Evaluation

The catalytic performance of different configurational catalysts was performed in a fixed-bed quartz reactor (ID = 8 mm) at atmospheric pressure. Typically, 0.5 g of catalyst (40~60 mesh) was mixed with SiO₂ (40~60 mesh) to 5 mL. Then the above mixture was added into the isothermal region of the reactor. All the samples were reduced under a 60% H₂/N₂ flow at the rate of 120 mL min⁻¹ at 600 °C for 3 h under atmospheric pressure. After cooling down to room temperature, the reactors were pressurized employing a feed stream with composition of 12 vol% H₂/36 vol% CO₂/52 vol% N₂; the catalytic performance of the catalysts was evaluated at 400–550 °C with the heating rate of 2 °C·min⁻¹. The tail gas was analyzed by the Agilent 9000 B gas chromatograph equipped with a FID detector and a TCD detector. All data points were collected at a steady state after continuously reacting for 2 h. N₂ in the feed gas was assumed an internal standard during the reaction. The conversion and selectivity were calculated via the different equations as follows:

$$\begin{array}{l}{X}_{CO}{}_{{}_2}=\frac{F_{C{O}_2,in}-F_{C{O}_2,out}}{F_{C{O}_2,in}}\times 100\%\\ {\mathrm{S}}_{\mathrm{CO}}=\frac{F_{CO,out}}{F_{C{O}_2,in}-F_{C{O}_2,out}}\times 100\%\\ \end{array}$$

The “in” or “out” in the above equation represents gas that flows into or out of the reactor.

4. Conclusions

The hollow hierarchical structure Cu@S and CuFe_0.5@S catalysts were successfully synthesized through the dissolution recrystallization (D-R) method for the RWGS reaction. The encapsulated catalysts had a hierarchical porous structure and better dispersion of Cu particles compared to the Cu-S and CuFe_0.5-S catalysts prepared via a conventional impregnation method. Moreover, compared to monometallic Cu-S and Cu@S catalysts, CuFe_0.5-S and CuFe_0.5@S catalysts showed a higher CO₂ conversion and 100% selectivity of CO at the entire temperature range investigated in this work as a result of introducing FeO_x species. Furthermore, the reaction activity of all the samples increased according to the sequence: CuFe_0.5@S > CuFe_0.5-S > Cu@S > Cu-S at 400–550 °C under atmospheric pressure. The remarkably enhanced catalytic performance in the RWGS reaction was associated with the higher dispersion of encapsulation structure and the enhanced surface basicity derived from the addition of Fe. This work provides some optimistic prospects in designing high-exhibition catalysts for the RWGS reaction and even the utilization of CO₂ hydrogenation.