Synergetic effect of metal–support for enhanced performance of the Cu–ZnO–ZrO₂/UGSO catalyst for CO₂ hydrogenation to methanol

Abstract

Novel Cu–ZnO–ZrO₂/UGSO catalysts for CO₂ hydrogenation to methanol were developed using a metallurgical residue (UGSO) as the catalytic support, with special focus on understanding the effect of (i) synergy between Cu/Zn/Zr and UGSO composition on the catalytic activity and stability, and (ii) UGSO modification on the structure–activity relationship of the developed catalysts. Zr acts as a structure promoter to enhance Cu dispersion and facilitates CO₂ conversion. The highest methanol yield is achieved for CZ9Zr/UGSO (10 wt% Cu, 7.5 wt% Zn, 9 wt% Zr), attributable to the highest (i) Mg/Cu surface ratio and (ii) interfacial contact Cu–ZrO₂. Further modification of UGSO by H₂O₂ (UGSO-H) increases by 3 times the surface area and endows CZ9Zr/UGSO-H with the best CO₂ conversion and 57% higher methanol yield at 240 °C/20 bar compared to a commercial Cu–ZnO-based catalyst. These findings may pave the way for effective utilization of the no-marketable-value residue as a catalytic support in other CO₂ conversion processes.

---

1. Introduction

The unprecedented improvement in industrialization has undoubtedly led to the deterioration of natural resources and intensified many environmental problems such as the ceaseless emission of anthropogenic greenhouse gases. The main contributor of greenhouse gases, CO₂, is considered as a cheap and abundant C1-compound that can be converted into chemicals and fuels.¹ Among them, methanol is not only a clean alternative fuel but a crucial raw material in the chemical and energy industries owing to its usage as a building block to produce other chemicals such as dimethyl ether, light olefins, dimethyl carbonate, formaldehyde, methyl tert-butyl ether, etc. It can be also used as a liquid energy-carrier in transportation and fuel cell application.² The transformation of CO₂ into methanol through the CO₂ hydrogenation process has become an appealing research interest in recent decades.^3,4 Three catalyst categories have been widely used in CO₂ hydrogenation into methanol synthesis, namely: Cu-based, noble metal-based (such as Pd, Ag) and bimetallic catalysts (such as Ni–Ga, Pt–Co).⁵ Due to their hydrogen spill-over ability, long-term stability and sintering resistance, precious metal-based materials have remained attractive catalysts.^6,7 However, compared to Cu-based catalysts, their lower methanol selectivity, expensive cost and rare availability are the main factors that limit their utilization.⁸ Bimetallic materials are among other competitive catalysts because of their high thermal and mechanical stability although their lower activity in methanol synthesis compared to Cu-based catalysts is one of their drawbacks.⁹ Therefore, among catalysts that have been extensively used in this process, the upsurge of copper catalysts rooted from their high activity, low cost and wide availability has been widely reported.³

Despite having high activity, Cu particles are thermally unstable and easily subjected to sintering and agglomeration. Numerous strategies have been attempted to overcome this drawback including the incorporation of suitable promoters. Among them, ZrO₂ is considered as an appealing candidate because it can (i) act as a texture promoter which enhances Cu dispersion and redox ability,^10,11 (ii) improve the basicity which facilitates CO₂ adsorption and activation,^12–14 and (iii) create oxygen vacancies to favor the activation of CO₂ and the transformation of formate to methoxy.¹⁵ It was reported that the appropriate addition of Zr (1 wt%) to the CuO–Fe₂O₃ catalyst could not only lower the reduction temperature from 209 to 198 °C of CuO species by increasing the exposure of Cu on the catalytic surface, but also help prevent CuO agglomeration.¹¹ As a result, CO₂ conversion increased by a factor of 4 and methanol selectivity improved by 15-fold. Similarly, Xaba et al. reported that Zr addition offers a 3-fold higher methanol yield due to a better CO₂ adsorption capacity and strength, better H₂ uptake, and higher Cu reducibility.¹⁶ L'hospital et al.¹⁷ revealed a 3-fold increase in methanol yield through the addition of 23.8 wt% ZrO₂, attributed to the interaction between ZnO and ZrO₂ that improves Cu dispersion. In addition, the interaction between Cu species and metal oxides can create active sites for methanol formation.^18–20 Stangeland et al. found that the incorporation of 1 mol% of Zr into the Cu/ZnO catalyst leads to the formation of Cu–ZrO₂ interface sites and enhances the methanol formation rate from 14.7 to 15.6 mmol g_cat⁻¹ h⁻¹ (at 230 °C, 30 bar, and 38 000 cm³ g_cat⁻¹ h⁻¹).²⁰ Palomino et al.²¹ have found that the rate of methanol formation of the deposited ZnO on Cu(100) was 15–45 times higher than that of Cu(100), being attributed to the ZnO–Cu interface stimulating the formation and binding of formate species, which are finally hydrogenated into methanol. Other studies have reported that the atomic hydrogen needed for the hydrogenation of adsorbed carbonate was transferred from the Cu to ZnO surface at the Cu/ZnO interface.^22,23 The same conclusion was drawn by Larmier et al.²⁴ who highlighted that the Cu–ZrO₂ boundary was essential for the formation of formate species, which further transforms into methoxy, and finally methanol. Understanding the impact of ZrO₂ and interfacial sites between Cu species and metal oxides is thus crucial for achieving optimum methanol production in CO₂ hydrogenation.

A cost-effective catalytic support plays a crucial role in developing an effective catalyst for large scale application, as it not only enhances the active site properties, but also meets the objective of economic sustainability. Also, the implementation of environmentally friendly policies addressing the severe issues related to industrial residues has led to a surge of interest in the exploitation of solid waste as heterogeneous catalysts for the transformation of chemicals.²⁵ Salient examples of this are CaO rooted from metallurgical wastes and fly ash for biodiesel production,^26,27 iron oxides derived from red mud for hydrogen production²⁸ and slag-based geopolymer for reverse water-gas shift reaction.²⁹ Currently, at Rio Tinto Fer & Titane Company situated in Sorel-Tracy (Québec, Canada), the so-called Upgraded Titania-rich slag (UGS®), containing 94.5% of titanium dioxide, is produced from hemo-ilmenite ore (FeTiO₃–Fe₂O₃) by a high-pressure and high temperature acid leaching process (T > 900 °C and P > 5 bar).³⁰ At the end of this process, the beads of the metal oxides, containing mainly Fe, Mg, and Al in the form of magnetite (Fe₃O₄), magnesioferrite (MgFe₂O₄), hercynite (FeAl₂O₄) and free periclase (MgOH), are generated, crystalized, discarded, and referred to as UGS oxide (UGSO) residue. This waste has zero marketable use and is being disposed of in a mining waste site. Its use as the main component of heterogeneous catalytic materials would be a low cost and environmentally friendly concept. As example, our recent studies on hydrogen production by glycerol steam reforming^31–33 revealed a superior performance of Ni/UGSO catalysts in terms of glycerol conversion and H₂ yield in comparison with the commercial catalyst. These findings open therefore a door to the prospective use of this waste as a potential catalytic support and promoter in other patterns. In particular, the UGSO features made us believe that it could be a good candidate in CO₂ hydrogenation into methanol because of (1) thermal and mechanical resistance, since it is produced under high temperature and pressure conditions, and (2) the presence of MgO that could favor methanol formation owing to the preferable adsorption of formate species, which further hydrogenate into methanol rather than dissociate into CO on basic sites.^34,35 On the other hand, given that the surface area of the support dictates the dispersion of active metals, it becomes highly appealing to shed light on the possibility of a simple but efficient treatment of UGSO with the aim of increasing its surface area. In this respect, certain studies have found that treating solid materials with H₂O₂ could result in a substantial increase in surface area.^36–39 As an example, Jain et al.³⁶ reported an increase by 2 times in the mesopore area of activated carbon by employing H₂O₂-mediated hydrothermal pre-treated coconut shells.

In the contexts mentioned above, novel Cu–ZnO–ZrO₂/UGSO catalysts for CO₂ hydrogenation into methanol have been developed using a metallurgical residue as the catalytic support, with special focus on the understanding of the effect of (i) synergy between Cu/Zn/Zr and UGSO on the catalytic activity and stability, and (ii) UGSO modification on the structure–catalytic activity relationship. First, Cu and Zn fractions are fixed at 10 and 7.5 wt%, respectively, while the Zr content is varied from 3 to 12 wt% to identify the optimized Zr loading for methanol synthesis. Then, with the aim to increase the surface area of the support, we propose a simple modification of UGSO by H₂O₂ treatment. To elucidate the effect of UGSO modification and the relationship between the catalytic structure and activity of Cu–ZnO–ZrO₂ deposited over raw (as-received) and modified UGSO supports, numerous characterization techniques are employed, such as X-ray diffraction (XRD), H₂ temperature programmed reduction (H₂-TPR), CO₂ temperature programmed desorption (CO₂-TPD), N₂ physisorption, N₂O chemisorption, X-ray photoelectron spectroscopy (XPS), scanning electron microscopy (SEM), and transmission electron microscopy (TEM). To the best of the authors' knowledge, no similar investigation on the utilization of UGSO as a support for the development of Cu–ZnO–ZrO₂-based catalysts for CO₂ hydrogenation into methanol is available in the open literature.

2. Experimental

2.1. Catalyst preparation

2.1.1. Cu–ZnO–ZrO₂/UGSO catalysts. The as-received beads of UGSO were first crushed and sieved to particle sizes less than 90 μm prior to use. A series of Cu–ZnO–ZrO₂/UGSO catalysts were prepared by a deposition–coprecipitation method. First, UGSO and de-ionized water were stirred in a round-bottom flask placed in an oil bath. A solution containing copper nitrate trihydrate (Cu(NO₃)₂·3H₂O, Acros Organics, 99.0%, 0.1 M), zinc nitrate hexahydrate (Zn(NO₃)₂·6H₂O, Acros Organics, 98.0%, 0.1 M) and zirconium oxynitrate hydrate (ZrO(NO₃)₂·xH₂O, Acros Organics, 99.5%, 0.1 M) and a solution of NH₄HCO₃ (Alfa Aesar, 99.0%) (0.3 M) were added dropwise simultaneously into the above suspended UGSO under stirring at 85 °C until the pH value equal to 8.5. After precipitation, ethanol (AnalaR NORMAPUR® ACS) was added into the above mixture (ethanol/Cu molar ratio of 12) and stirring was kept for 60 min at 90 °C. Subsequently, the final precipitate was cooled down to room temperature and washed 3 times with de-ionized water in vacuum filtration. The collected solid powder was then dried overnight at 85 °C in an oven before being calcined at 430 °C for 3 h in an inert atmosphere (pure N₂, 300 mL min⁻¹) (heat rate: 3 °C min⁻¹). Finally, the product was milled and sized to obtain particles sizes which are less than 90 μm before characterizing and testing for the CO₂ hydrogenation reaction. The percentages of Cu and Zn were kept at 10 wt% and 7.5 wt%, respectively, while the percentage of Zr was varied from 3 to 12 wt%. The catalysts were denoted at CZ3Zr/UGSO, CZ6Zr/UGSO, CZ9Zr/UGSO and CZ12Zr/UGSO, corresponding to the percentage of Zr loading (C, Z, and Zr denote Cu, Zn, and Zr, respectively).

2.1.2. Cu–ZnO–ZrO₂/modified UGSO catalyst. First, a mixture of UGSO powder (<90 μm) and de-ionized water was stirred in a round-bottom flask placed in an oil bath. A solution of H₂O₂ (30% w/w, Honeywell Fluka®) was slowly added into the suspended UGSO (3 mL H₂O₂ 30% w/w per g UGSO) until no cloudy oxygen could be seen. The mixture was then stirred at 70 °C for 3 hours and cooled down to room temperature. The resulting solid material was collected by vacuum filtration and washed with de-ionized water until the pH value of the filtrate reached 8. The filter cake was then dried at 110 °C overnight and calcined at 600 °C for 2 h under air atmosphere (300 mL min⁻¹) (heat rate: 3 °C min⁻¹). The obtained material was denoted as UGSO-H.

To evaluate the influence of different supports (as-received and H₂O₂-modified UGSO) on CO₂ hydrogenation, a catalyst containing 10 wt% Cu, 7.5 wt% Zn, and 9 wt% Zr was prepared using the deposition–coprecipitation described in section 2.1.1., where the UGSO was substituted by UGSO-H. The obtained catalyst was denoted as CZ9Zr/UGSO-H.

2.2. Catalyst characterization

Elemental chemical composition was analysed by X-ray fluorescence (XRF) using a Thermo Niton XL3t 980 GOLDD+ (2 W power) apparatus equipped with a Ag anode. Major elements were presented in weight percentage (wt%) while trace elements were expressed in ppm (parts per million).

The X-ray diffraction (XRD) patterns were recorded on a Bruker D8 Advance, using Cu Kα radiation (λ = 1.542 Å) operated at 40 kV and 40 mA. The catalysts were scanned from 10° to 90° (2θ) with a step size of 0.01° and a time/step of 0.5 s. Phase identification was performed using DIFFRAC.EVA software.

The Brunauer–Emmett–Teller (BET) surface area and Barrett–Joyner–Halenda (BJH) pore size distribution were measured by N₂ physisorption at −196 °C using a Micromeritics ASAP 2020 analyser. Prior to the analysis, the sample was degassed under vacuum at 250 °C for 3 h.

The morphology properties were examined by transmission electron microscopy (TEM) using a JEOL electron microscope (JEM 1230) with an accelerating voltage of 80 kV. Before characterization, the catalysts were dispersed in ethanol and deposited on a carbon film. The samples were then dried naturally at room temperature.

The surface chemical composition and electronic states of the element in the calcined and reduced catalysts were evaluated by X-ray photoelectron spectroscopy (XPS) analysis and the X-ray induced Auger spectra of Cu species were obtained by atomic emission spectroscopy (AES) using a Kratos Axis Ultra spectrometer equipped with a monochromatic Al Kα X-ray radiation source operating at 300 W. The binding energy (BE) scale was corrected internally by setting the BE of C 1s at 284.8 eV. Peak fitting and calculation of relative atomic concentrations were performed using the Origin software.

The surface morphology was observed by scanning electron microscopy (SEM-VEGA3) via a retractable solid state backscattered electron detector. The working distance of the SEM was 10 mm with a 10.0 kV accelerating voltage and a 6 A beam current. Prior to the analysis, the sample was coated with gold and palladium to prevent the charging effect.

The reducibility of the developed materials was evaluated by temperature program reduction (H₂-TPR) using a Micromeritics Autochem II 2920. Prior to the measurement, the sample was loaded in a U-shape tube reactor and pre-treated in a pure Ar flow (50 mL min⁻¹) for 10 min. Subsequently, the gas was switched to 10.2 vol% H₂/N₂ flow (50 mL min⁻¹) and the sample was heated from room temperature to 900 °C at a heating rate of 10 °C min⁻¹ and held at 900 °C for 30 min to get the baseline. H₂ consumption was monitored using a thermal conductivity detector (TCD).

Temperature programmed desorption (CO₂-TPD) was conducted on the same apparatus to analyse the surface basicity of catalysts. A pre-reduction of the sample at 350 °C was first performed in a nitrogen-diluted hydrogen stream for 2 h (10.2 vol% H₂/N₂ flow, 50 mL min⁻¹). After the reduction step, the sample was cooled down to 40 °C under a pure He flow and then saturated with CO₂ for 1 h under a 10.0 vol% CO₂/He flow (50 mL min⁻¹) for 1 h. Afterward, the flow was switched back to He (50 mL min⁻¹) to eliminate any physiosorbed CO₂ on the catalyst surface. The temperature was subsequently increased to 700 °C under the same He flow with a heating rate of 5 °C min⁻¹. The desorbed CO₂ signal was continuously monitored by a TCD.

The copper surface area was analysed by N₂O chemisorption using the same apparatus. Approximately 30 mg of catalyst was first reduced at 350 °C in a flow of 10.2 vol% H₂/N₂ (50 mL min⁻¹). After reduction, the sample was cooled down to 60 °C in pure He flow (50 mL min⁻¹) and exposed to N₂O flow (50 mL min⁻¹, 10 vol% N₂O/He) for 1 h at the same temperature so that Cu⁰ in the surface can be oxidized into Cu₂O. Then the sample was flowed with pure Ar flow for 30 min in order to remove the physically adsorbed N₂O. Subsequently, the catalyst was reduced again following the procedure mentioned above.

The copper surface area S_Cu (m² g_Cu⁻¹) and dispersion D_Cu (%) are determined based on the following equations:⁴⁰

where n_H2 is the amount of H₂ uptake in the second reduction, N_av is Avogadro's constant (6.02 × 10²³ atom per mol), M_cat is the mass of the catalyst (g), A is the number of Cu surface atoms per unit area (1.46 × 10¹⁹ atom_Cu per m²), M_Cu is the Cu molar mass and W_Cu is the weight percentage of Cu in the catalyst.

2.3. Catalyst evaluation in methanol synthesis

Methanol synthesis was carried out in a milli-fixed bed stainless steel reactor (inner diameter of 6.35 mm and a length of 27 cm) placed in a thermostated chamber. The reaction tube was loaded with 0.5 g of catalyst (<90 μm) installed above quartz wool in the center of the tube. The temperature of the reactor was measured by a thermocouple located inside the catalytic bed. Prior to CO₂ hydrogenation, the catalyst was reduced at 350 °C for 2 h in a 50 vol% H₂/N₂ flow (50 N mL min⁻¹). After reduction, the temperature was cooled down to the reaction temperature under a N₂ flow (50 N mL min⁻¹). When the reaction temperature was reached, a mixture of H₂/CO₂/N₂ (9/3/1) was fed to the reactor at 20 bar (gas space velocity (GHSV) of 5000 N mL g_cat⁻¹ h⁻¹). The reaction stream was analyzed online using an Agilent 6890N GC equipped with a thermal conductivity detector connected with a SUPELCO Carboxen 1010 plot column (30.0 m × 320 μm × 50.0 μm) for H₂, CO₂, N₂ and CO analysis, and a flame ionization detector connected with an Agilent JW CP-PoraBond Q column (25 m × 320 μm × 45.0 μm) for methanol analysis. The reactant and carrier gaseous flows were controlled using mass flow controllers (Bronkhorst®). Methanol and CO were the only detected products of the reaction. The catalyst activity was analysed in a temperature range between 200 and 300 °C with gaps of 20 °C. Three different as-received UGSO batches were used to evaluate the possible influence of the UGSO source on the catalytic performance. The results are consistent, indicating that fluctuation in raw UGSO from batch-to-batch is not a significant factor influencing catalytic activity.

Limitations related to the external diffusion and internal diffusion in the catalyst pores were considered negligible due to the moderate conversions reported in this study, which support a reaction-limited regime.⁴¹

CO₂ conversion (X_CO2, %), CO selectivity (S_CO, %), methanol selectivity (S_CH3OH, %), and methanol yield (Y_CH3OH, %) were determined based on eqn (3)–(6).

where F_CO2,in and F_CO2,out are the molar flow rate of CO₂ feeding and exiting the reactor, respectively. F_CH3OH,out and F_CO,out are the molar flow rate of methanol and CO, respectively.

3. Results and discussion

3.1. Effect of Zr content

3.1.1. Structure and physical characterizations. The influence of Zr on the physicochemical properties of CZxZr/UGSO catalysts was explored to elucidate the structure–property relationship. First, the crystallinity and phase structure of the samples with different Zr contents were examined using XRD analysis. Fig. 1 shows the XRD patterns of calcined UGSO and CZxZr/UGSO. As can be seen, the diffraction peaks observed at 2θ of 18.4°, 30.3°, 35.7°, 43.4°, 53.9°, 57.4°, and 63.1° correspond to the main phase of UGSO, including magnesioferrite Mg(Al_xFe_2−x)O₄ solid solution (PDF: 04-014-3055) and magnetite Fe₃O₄ (PDF: 04-005-4319). Periclase MgO (PDF: 00-045-0946) is also detected. These findings are in line with the results from Sahraei et al.³¹ After the addition of Cu, Zn and Zr into the UGSO by the deposition–coprecipitation method, the appearance of a new peak at 2θ = 38.9° ascribed to the formation of tenorite CuO (PDF: 04-007-0518) is only found in CZ3Zr/UGSO and CZ6Zr/UGSO. When the Zr content is above 6 wt%, the intensity of diffraction peaks of CuO becomes weaker and disappears, suggesting the existence of an amorphous phase or the high dispersion of CuO in these samples, so that their particles are too small to be identified by the XRD method. The two shoulders at 2θ = 31.7° and 34.4° can be assigned to the presence of zincite ZnO (PDF: 04-003-2106). The same tendency in the diffraction peaks of CuO is observed for those of ZnO, indicating the high dispersion of ZnO or its existence in amorphous phase in the catalysts of 9 and 12 wt% Zr. In addition, the formation of mixed oxides of iron, copper, and zinc can be created in: (1) CuFe₂O₄ (PDF no.: 00-034-0425), and/or copper iron mixed oxides (Cu_xFe_3−xO₄) such as Cu_0.86Fe_2.14O₄ (PDF: 01-073-2315), and/or Cu(Fe_2−xAl_x)O₄ such as Cu_0.5Fe_1.67Al_0.67O₄ (PDF: 04-007-6615); and (2) ZnFe₂O₄ (PDF: 04-006-1542), and/or zinc iron mixed oxides (Zn_xFe_3−xO₄) such as Zn_1.1Fe_1.9O₄ (PDF: 04-018-7642), and/or Zn(Fe_2−xAl_x)O₄ such as ZnFe_1.5Al_0.5O₄ (PDF 04-007-6615). The presence of these mixed oxides was found to facilitate the dispersion of Cu²⁺ and Zn²⁺ (ref. 42 and 43) because Cu²⁺ ions occupy predominantly octahedral sites in copper ferrite spinel, while Zn²⁺ ions occupy tetrahedral sites in zinc ferrite spinel.^44,45 The reflections at 2θ = 30.4° and 34.3° can be ascribed to the presence of t-ZrO₂ (PDF 04-017-7044). However, due to the overlap of these diffraction peaks with the characteristic peaks of mixed iron and magnesium oxides, the crystallinity degree of ZrO₂ in each catalyst cannot be described precisely. The average crystallite size of CuO (which is estimated using the Scherrer equation) decreases from 19.2 nm for CZ3Zr/UGSO to 17.4 nm for the CZ6Zr/UGSO catalyst.

Fig. 1 XRD patterns of UGSO and CZxZr/UGSO catalysts.

The textural properties of the prepared samples were investigated using the N₂ adsorption–desorption isotherms which are displayed in Fig. S1.† As shown, the isotherm of UGSO is a classical type II, characteristic of a non-porous or macroporous material according to IUPAC (International Union of Pure and Applied Chemistry) classification. The presence of mesopores in small quantity is indicated by a slight hysteresis loop with a steep desorption branch.⁴⁶ After the deposition–coprecipitation of Cu, Zn, and Zr into UGSO, the N₂ adsorption–desorption isotherms of CZxZr/UGSO catalysts show a typical type II, accompanied by a well-defined hysteresis loop with steep adsorption and desorption branches that belong to the H3 type, indicating the presence of mesopores with aggregates of plate-like particles with slit-shaped pores.⁴⁶ The pore size distribution is determined by the BJH method from the adsorption branch of the isotherms (Fig. S1B†). It is observed that a small quantity of intrinsic pores of UGSO centers at around 3 nm. For CZxZr/UGSO catalysts, the increase of Zr content creates more pores also centered at 3 nm. With the increase of Zr content from 3 to 12 wt%, the BET surface area of CZxZr/UGSO increases significantly from 35.0 to 100.9 m² g⁻¹, together with the increase of pore volume from 0.14 to 0.23 m³ g⁻¹ (Table 1). The increase in surface area and pore volume could be an indication of the new pores created by the addition of Zr. This can be beneficial for the dispersion of CuO particles, as evidenced by the XRD results described above, in which the diffraction peak of CuO is weaker with the addition of Zr.

Table 1 Physical properties of CZxZr/UGSO catalysts

Sample	S BET (m^2 g^−1)	Pore volume^a (m^3 g^−1)	Pore diameter^a (nm)	S Cu (m^2 g^−1)	D Cu (%)
a Determined from the BJH pore size distribution adsorption branch.
Fresh UGSO	5.6	0.04	26.7	—	—
CZ3Zr/UGSO	35.0	0.14	16.1	14.9	20.7
CZ6Zr/UGSO	43.2	0.14	12.4	15.9	22.0
CZ9Zr/UGSO	53.8	0.15	10.9	17.1	25.7
CZ12Zr/UGSO	100.9	0.23	9.0	21.1	33.5

---

The metallic Cu surface area was determined by N₂O chemisorption and the results are presented in Table 1. With the increase of Zr loading content, S_Cu increases from 14.9 to 21.1 m² g⁻¹ and D_Cu rises from 20.7 to 33.5%. These results are clearly consistent with the XRD and N₂ physisorption analyses presented above, indicating that Zr can act as a physical promoter to enhance the dispersion of Cu species.

3.1.2. Surface chemical properties. H₂-TPR was performed to analyse the redox behaviors of the samples. As shown in Fig. S2,† the onset reduction temperature of mixed metal oxides in UGSO is at 500 °C. The sample exhibits two overlapping peaks at approximately 660 °C and 800 °C, which could be assigned to the reduction of iron oxide species.^43,47 After the deposition–coprecipitation, the reducibility of iron oxide species in UGSO is improved, as evidenced by the shift to lower temperature of the reduction peak from 660 °C to 500–600 °C. This can be attributed to the hydrogen spill-over effect of metallic copper.⁴⁸ Regarding the H₂-TPR profiles of CZxZr/UGSO catalysts presented in Fig. 2, the onset reduction temperature decreases significantly from 169 to 152 °C when the amount of Zr rises from 3 to 9 wt%, and then slightly increases to 154 °C with the further increase of Zr content from 9 to 12 wt%. This indicates that the appropriate addition of Zr (9 wt%) can enhance the reducibility of Cu species. All the profiles in the low temperature range (100–350 °C) are deconvoluted into several Gaussian peaks. The peak positions and their relative concentrations are summarized in Table S1.† The α, β, and γ peaks can be assigned to different copper species that are related to different aggregation states and the interaction between active metals and the support. It was reported that ZnO and ZrO₂ are not reduced in this experimental region,^49,50 suggesting that these three reduction peaks are linked to the reductions of three different copper species. The low temperature peak at around 190 °C (α peak) can be attributed to the reduction of mixed oxides of copper and iron,⁵¹ while the peak at higher temperature (β peak) can be assigned to the reduction of highly dispersed CuO species which strongly interact with ZnO and ZrO₂.⁵² The third peak (γ peak) around 205–240 °C can be related to the free or bulk CuO species,^34,52 while the small peak (λ peak) at around 252 °C in CZ3Zr/UGSO is associated with the reduction of larger-sized bulk CuO particles, probably due to the lowest surface area of this catalyst. The position of the α peak for all the catalysts follows the same tendency as the onset temperature. It is reported that the movement in reduction peaks can be associated with the dispersion of particles and the higher dispersion is linked to the lower reduction temperature.⁵³ The shift of the β peak position from 210 to 200 °C with the elevation of Zr content from 3 to 12 wt% shows that higher CuO dispersion is found in the catalyst with higher Zr loading. This result is in good agreement with those obtained from XRD and BET analysis, suggesting the role of Zr as a texture promoter in enhancing Cu reducibility and preventing CuO agglomeration.³⁴ As summarized in Table S1,† the maximum relative contribution of the β peak is observed for CZ9Zr/UGSO, suggesting the highest amount of dispersed CuO species in close contact with ZnO and ZrO₂ in this catalyst. It can be suggested that the introduction of an appropriate amount of Zr can facilitate the amount of CuO that interacts with ZnO and ZrO₂, consequently leading to an increase of Cu–ZnO and Cu–ZrO₂ interface sites. Since the interfaces between Cu–ZrO₂ or Cu–ZnO are the main active sites for methanol formation while the RWGS primarily occurs on the metallic Cu surface,^18,24 CZ9Zr/UGSO is projected to show the highest methanol formation. The largest γ peak is observed for CZ3Zr/UGSO, which is associated with the biggest amount of bulk CuO species in this catalyst. It is observed that the position of the γ peak shifts to lower temperatures from 236 to 205 °C with the elevation of Zr content from 3 to 12 wt%, indicating that the bulk CuO particles become smaller with higher Zr addition.

Fig. 2 H₂-TPR profiles of CZxZr/UGSO catalysts. Solid lines indicate raw data while dotted lines present separated peaks. The α peak is assigned to the reduction of CuFe₂O₄, the β peak is appointed to the reduction of highly dispersed CuO species, the γ peak is ascribed to the reduction of separated CuO species, and λ is associated with the reduction of larger-sized bulk CuO particles.

The XPS spectra are shown in Fig. 3 to assess the oxidation state, surface composition, and electronic interaction between Fe, Cu and Zr species present in the calcined CZxZr/UGSO samples. Fe 2p, Cu 2p and Zr 3d spectra were all peak-fitted into several peaks using the Gaussian distribution. Considering the spectra of Fe 2p for the UGSO (Fig. 3A), the dominant peak at BE of 710.1 eV can be assigned to Fe 2p_3/2, while the peak at BE of 724 eV can be associated with Fe 2p_1/2. The satellite band at around 719 eV can characterize the presence of Fe³⁺.^54,55 After the deposition of Cu, Zn and Zr, the 2p_3/2 peak can be fitted into two peaks at 710–712 eV and 712–714 eV, which are assigned to the presence of Fe³⁺ in octahedral sites (in Fe₃O₄ and ZnFe₂O₄) and Fe³⁺ in tetrahedral sites (in Fe₃O₄ and CuFe₂O₄), respectively.^56,57 As shown in Fig. 3B, the Zr 3d spectra displays two characteristic peaks at 181.7 and 183.4 eV, being attributed to Zr 3d_5/2 and Zr 3d_3/2 of Zr⁴⁺ species.⁵⁸ Regarding Cu 2p_3/2 (Fig. 3C), the BE at 933.8 eV and the satellite band in the region of 940 and 945 eV indicate the presence of Cu²⁺ in the mixed iron copper oxides and CuO structure, respectively.^45,59 In addition, it was reported that the change in chemical environment is predicted by the shift of binding energy.⁶⁰ The Cu 2p spectra of the calcined CZxZr/UGSO catalysts shift to higher BE value with the incorporation of Zr from 3 to 9 wt% and slightly decrease to a lower binding energy value for 12 wt% Zr. The inverse tendency is, however, noted for the Zr 3d spectra, with the lowest binding energy value observed for the CZ9Zr/UGSO catalyst. These positive and negative binding energy shifts could be attributed to the electron transfer from Cu to Zr at the Cu–ZrO_x interfaces, indicating a strong interaction between Cu and Zr in CZxZr/UGSO samples.⁶¹ The appropriate addition of Zr into Cu–ZnO/UGSO (9 wt% Zr) could promote the synergetic effect between Cu and Zr, resulting in the highest amount of dispersed CuO species which are in close contact with ZrO₂, as confirmed by H₂-TPR analysis.

Fig. 3 XPS spectra of surface elements of calcined samples: (A) Fe 2p, (B) Zr 3d, (C) Cu 2p. Solid brown, yellow, green, blue, and purple lines indicate raw data while dotted green, red, and blue lines are separated peaks.

The relative atomic surface ratios of different elements in the calcined and reduced samples are presented in Table 2. As shown, the surface atomic Zn/Cu ratio of all the samples is considerably higher than their nominal atomic value, which is 0.74. The value of the Zn/Cu atomic ratio decreases with the increase of Zr, proposing that the addition of Zr leads to a partial separation of Cu onto the surface of the samples.⁶² This confirms that the addition of Zr results in a higher amount of Cu exposed on the surface, as proved by N₂O chemisorption results (Table 1).

Table 2 Surface atomic ratio derived from XPS analysis of calcined samples

Sample	Zn/Cu	Mg/Cu
CZ3Zr/UGSO	3.55	0.11
CZ6Zr/UGSO	2.73	0.17
CZ9Zr/UGSO	2.25	0.37
CZ12Zr/UGSO	2.16	0.13

---

The surface basicity of the catalysts was analyzed by temperature programmed desorption of CO₂ (CO₂-TPD) and their profiles are presented in Fig. S3.† Based on the strength of basic sites, the basicity can be divided into three categories. The weakly basic sites (50–200 °C) are assigned to the OH– group, the moderately basic sites (200–500 °C) are attributed to the metal oxygen M–O pairs (M: Zr, Mg, Cu or Zn) and the strongly basic sites are associated with the unsaturated O₂ anion (>500 °C).²⁰ The variation in the proportion of the moderately strong basic area to the total basic area (CZ9Zr/UGSO (0.51) > CZ6Zr/UGSO (0.49) > CZ12Zr/UGSO (0.41) > CZ3Zr/UGSO(0.4)) follows a similar order of the Mg/Cu ratio (Table 2). The highest amount found for CZ9Zr/UGSO suggests that the presence of MgO is related to the surface moderately basic strength. In addition, the highest fraction of moderately strong basic sites to the total basic sites and the highest amount of the dispersed CuO species strongly interacting with the ZnO and ZrO₂ observed for CZ9Zr/UGSO, as provided by H₂-TPR results, indicate that the moderately strong basicity can also be related to the interface between Cu–ZnO and Cu–ZrO₂.⁶³ Since the moderately basic strength is associated with CH₃OH formation,³⁵ CZ9Zr/UGSO is expected to have the highest methanol yield among the CZxZr/UGSO samples. Moreover, an increasing trend in the area and desorption temperature of strong basic sites for the catalyst containing higher Zr loading suggests a similar order of CO₂ adsorption capacity for these catalysts.

3.2. Effect of UGSO modification by H₂O₂

3.2.1. Structure and physical characterizations. H₂O₂ was used to modify the as-received UGSO with the aim of increasing the surface area of the support. To evaluate the influence of different supports (H₂O₂-modified UGSO and as-received UGSO) on CO₂ hydrogenation, the CZ9Zr/UGSO-H catalyst was prepared using basically the same synthesis method as CZ9Zr/UGSO. XRF analysis was used to analyze the change in chemical composition of UGSO after the H₂O₂ treatment. Table 3 presents the main chemical compositions of UGSO, UGSO-H, CZ9Zr/UGSO, and CZ9Zr/UGSO-H samples.

Table 3 Chemical composition of samples (wt%)

Sample	Al	Mg	Fe	Mn	Si	Ca	Cr	V	Cu	Zn	Zr
UGSO	7.20	16.73	42.44	3.55	0.11	1.35	0.45	0.87	—	—	—
UGSO-H	6.46	15.33	39.30	3.27	0.15	0.76	0.42	0.84	—	—	—
CZ9Zr/UGSO	3.06	4.53	23.58	1.78	0.21	0.15	—	—	9.93	7.15	8.41
CZ9Zr/UGSO-H	4.44	6.62	23.78	1.10	0.19	0.12	—	—	11.38	7.61	9.45

---

Considering the as-received UGSO, the main elements detected were Fe (42.44 wt%), followed by Mg (16.73 wt%), Al (7.20 wt%) and Mn (3.55 wt%). Other elements such as Si, Ca, Cr and V were observed below 2 wt%, while the weight percentage of O was the rest. After H₂O₂ activation, the weight percentages of the main elements in UGSO slightly decreased (Fe (39.30 wt%), Mg (15.33 wt%), Al (6.46 wt%) and Mn (3.27 wt%)), indicating that the weight percentage of O increased. This can be attributed to the partial oxidation of these metal oxides by H₂O₂.⁶⁴

The morphology and topography of the calcined samples were observed by scanning electron microscopy and their images are presented in Fig. 4. As shown in Fig. 4A and B, the average diameter of UGSO particles is 22.3 μm (particle sizes ranging from 3.1 to 85.1 μm), while it is 17.0 μm for UGSO-H particles (particle sizes ranging from 2.0 to 78.2 μm), demonstrating a decrease in particle sizes. It is observed that H₂O₂-activated UGSO particles maintain a similar block-like morphology to the as-received UGSO particles. However, higher magnification images show that the surface of UGSO particles presents an overall platelet-like morphology (Fig. 4C), while a sphere-like morphology is observed for UGSO-H (Fig. 4D). This indicates a slight change of the particles from platelet to a more spherical shape after H₂O₂ activation. In addition, a rougher and more porous surface is observed for UGSO-H in comparison with UGSO. This could be related to the transformation of magnesium oxide into its correspondent peroxide caused by its interaction with H₂O₂.^65,66 MgO₂ can be decomposed into MgO through the thermal treatment because O₂˙⁻ is destroyed at temperature higher than 200 °C.⁶⁶ The more porous and defect-containing surface of USGO-H leads to its higher surface area in comparation with its counterpart, which is in good agreement with the N₂ physisorption results. Similarly, after deposition of Cu, Zn and Zr, the surface of CZ9Zr/UGSO-H consisted of rougher sphere-like particles in comparison to CZ9Zr/UGSO (Fig. 4E and F).

Fig. 4 Scanning electron microscopy of UGSO (A) and (C), UGSO-H (B) and (D), CZ9Zr/UGSO (E) and CZ9Zr/UGSO-H (F).

The influence of H₂O₂ treatment on the structure properties of the supports and their corresponding catalysts was investigated by XRD analysis. Fig. S4A† shows the XRD patterns of the calcined UGSO, UGSO-H, CZ9Zr/UGSO and CZ9Zr/UGSO-H catalysts. As can be seen, the diffraction peaks of all phases in UGSO-H are similar to those of UGSO, indicating no major change in crystalline phase by H₂O₂ activation.

In order to investigate the structures of the samples prior to the reaction, the reduced CZ9Zr/UGSO and CZ9Zr/UGSO-H samples (which were treated in a mixture of 50 vol% H₂/N₂ prior to the CO₂ hydrogenation reaction) were characterized by XRD and their patterns are presented in Fig. S4B.† The diffraction peak at 2θ = 43.3° is observed for the two reduced samples, indicating the presence of Cu (PDF: 04-004-0836)) and/or CuZn alloy (such as Cu₅Zn₈, PDF 04-007-1117). The average crystal sizes of Cu particles (which is estimated using Scherrer equations at 2θ = 43.3°) are 16.2 nm and 15.7 nm for the reduced CZ9Zr/UGSO and CZ9Zr/UGSO-H, respectively.

The N₂ adsorption–desorption isotherms of UGSO, UGSO-H and their corresponding catalysts (Fig. 5A) show type II, characteristic of macroporous or non-porous materials according to the IUPAC classification. The well-defined hysteresis loops with steep adsorption and desorption branches at high relative pressure (P/P₀ = 0.6–0.9) belong to the H3-type, suggesting the presence of mesopores with aggregates of plate-like particles accompanied by slit-shaped pores.⁴⁶ After H₂O₂ activation, the BET surface area of the support increases by 135%, which is well consistent with the literature^36–39 showing that H₂O₂ activation can lead to the improvement in BET surface area of the sample. In addition, the presence of mesopores over the wide range size of 5–35 nm (Fig. 5B) means that the H₂O₂ treatment creates a new pore type for the UGSO-H support, as proved by SEM images, resulting in an increased total pore volume of nearly two fold compared to that of the as-received UGSO (Table 4). After the deposition of Cu, Zn and Zr into the modified UGSO, the pore sizes increased and recentered at 5 nm, resulting in an increase of 40% in the BET surface area and 60% in the pore volume of CZ9Zr/UGSO-H in comparison to CZ9Zr/UGSO.

Fig. 5 N₂ adsorption–desorption curves (A) and pore size distributions (B) of samples.

Table 4 Physical properties of samples

Sample	S BET (m^2 g^−1)	Pore volume^a (m^3 g^−1)	Pore diameter^a (nm)	S Cu (m^2 g^−1)	D Cu (%)
a Determined from the BJH pore size distribution adsorption branch.
UGSO	5.6	0.04	26.7	—
UGSO-H	13.9	0.10	27.4	—
CZ9Zr/UGSO	53.8	0.15	10.9	17.1	25.7
CZ9Zr/UGSO-H	75.4	0.23	12.1	19.1	31.1

---

Table 4 presents the metallic Cu surface area of the catalysts. As shown, S_Cu increases by 12% for CZ9Zr/UGSO-H, being mainly assigned to the higher BET surface area of modified UGSO, which facilitates the dispersion of CuO species. As a result, Cu dispersion increases by 21% for this catalyst in comparison with its counterpart.

3.2.2. Surface chemical properties. H₂-TPR analysis was conducted to evaluate the effect of H₂O₂ activation on the reducibility and the interaction between metal active phases and the support. Fig. 6 shows the H₂-TPR profiles of the as-received/modified supports and their corresponding catalysts. As presented in Fig. 6A, two overlapping reduction peaks are observed for UGSO-H, which is similar to those of the as-received UGSO. The positions of these two peaks, however, shift to lower reduction temperatures for the activated UGSO, indicating that the oxidized metal species in UGSO-H are more easily reduced than the metal species in the original UGSO. Concerning the profiles of the catalysts prepared from two different supports, as presented in Fig. 6B, three overlapping peaks are also clearly observed, indicating the presence of three different copper species, as discussed in section 3.1.2. The peak position and its relative contribution are summarized in Table 5. The position of the β peak shifts to lower reduction temperature from 202 °C for CZ9Zr/UGSO to 194 °C for CZ9Zr/UGSO-H. The same tendency is observed for the γ peak, which moves from 206 to 203 °C. These decreases in reduction temperature indicate that the dispersion of CuO species in the catalyst supported over the modified UGSO is higher than those of the one supported over the as-received UGSO. This could be attributed to the higher surface area and pore volume of UGSO-H compared to the UGSO support, which facilitates the dispersion of CuO species. In addition, the relative contribution of the β peak of CZ9Zr/UGSO-H is larger than that of CZ9Zr/UGSO, suggesting the higher amount of the dispersed CuO species strongly interacting with the ZnO and ZrO₂ in the former. Therefore, it can be concluded that H₂O₂ activation affects the texture and reducibility of the CZ9Zr/UGSO catalyst.

Fig. 6 H₂-TPR profiles of the samples (A) from 50 to 900 °C and (B) from 50 to 350 °C. Solid lines indicate raw data while dotted lines present separated peaks. The α peak is assigned to the reduction of CuFe₂O₄, the β peak is appointed to the reduction of highly dispersed CuO species, and the γ peak is ascribed to the reduction of separated CuO species.

Table 5 Reducibility of CZ9Zr/UGSO and CZ9Zr/UGSO-H catalysts

Sample	Peak α		Peak β		Peak γ
	T (°C)	% area	T (°C)	% area	T (°C)	% area
CZ9Zr/UGSO	178	6.3	202	74.9	206	18.7
CZ9Zr/UGSO-H	179	15.2	194	79.3	203	5.5

---

The influence of H₂O₂ on the surface oxidation state, chemical composition and electronic interaction of elements in the samples was investigated by XPS analysis. The spectra are presented in Fig. 7 and S5.† The Fe 2p, Cu 2p, Zr 3d, and O 1s spectra were all fitted into several peaks using Gaussian distribution. Concerning the UGSO support, the main peak of the Mg 2p spectra (Fig. S5A†) at 49.2 eV is attributed to Mg²⁺ in the form of MgO or MgFe₂O₄ structure.^55,67 When compared to the Mg 2p spectra of UGSO-H (49.4 eV), a slight chemical shift to higher BE is observed, which can be deduced from the partial oxidation of Mg by the H₂O₂ activation.⁶⁸ The same tendency is observed for Fe (Fig. S5B†), in which the BEs of Fe 2p shift to higher values for UGSO-H, suggesting that the Fe species in the as-received UGSO were partially oxidized by H₂O₂. After the deposition–coprecipitation, the presence of two peaks at 710–712 eV and 712–714 eV in CZ9Zr/UGSO and CZ9Zr/UGSO-H suggests the presence of Fe³⁺.^56,57

Fig. 7 XPS spectra of surface elements in reduced samples (A) Cu 2p, (B) O 1s, (C) XAES spectra of Cu LMM. Solid red, grey, blue, and pink lines indicate raw data while dotted green, red, and blue lines are separated peaks.

Considering the reduced CZ9Zr/UGSO and CZ9Zr/UGSO-H catalysts (ex situ reduced in 50% H₂/N₂, at 350 °C, for 2 h, i.e., the reduction condition as prior to the methanol synthesis reaction), the peak at 933.7 eV and the satellite band in the region of 940 and 945 eV demonstrate the presence of the Cu²⁺ ion⁶⁹ (Fig. S5C†). The additional peak at ∼932.6 eV can be attributed to the existence of Cu⁰ or Cu⁺ from the reduction of Cu²⁺ species (Fig. 7A).⁷⁰ Due to the overlapping peaks of Cu⁰ and Cu⁺ in XPS, the X-ray induced Auger spectrum was used to distinguish Cu valences. The spectra of Cu LMM (Fig. 7C) are fitted into two peaks at 918.1 eV and 915.8 eV, which are assigned to Cu⁰ and Cu⁺, respectively.^61,71 This result further confirms the existence of both Cu⁺ and Cu⁰ in the reduced catalysts. The Cu^α/(Cu^α + Cu²⁺) ratio (α = 0 or 1+) of CZ9Zr/UGSO (26%) is lower than that of CZ9Zr/UGSO-H (38%), indicating the higher concentration of Cu^α in the latter. This could be attributed to the higher surface area of CZ9Zr/UGSO-H, which facilitates the dispersion and reducibility of the Cu species. Since both Cu⁰ and Cu⁺ are reported as active sites for the CO₂ hydrogenation reaction,^72,73 the higher concentration of (Cu⁰ + Cu⁺) in CZ9Zr/UGSO-H is expected to lead to the higher CO₂ conversion of this catalyst. Concerning the O 1s spectra of the reduced catalysts (Fig. 7B), the main peak is divided into 2 fitted peaks: (1) the peak at BE of 529.8 eV is assigned to the lattice oxygen (O_I) and (2) the peak at BE of 531.0 eV is appointed to the surface-adsorbed oxygen (O_II) (which is related to the presence of oxygen vacancies on the sample surface).⁷⁴ The relative contribution of O_II (O_II/(O_I + O_II)) of CZ9Zr/UGSO (61.4%) is lower than that of CZ9Zr/UGSO-H (71.9%), indicating the higher concentration of oxygen vacancies on the latter. As the oxygen vacancies on the catalyst surface are considered as active sites for CO₂ adsorption and activation,⁷⁵ CZ9Zr/UGSO-H containing a higher concentration of oxygen vacancies is expected to have higher CO₂ adsorption and conversion.

Fig. 8 shows the CO₂-TPD profiles of the two catalysts with the three divided regions as discussed in §3.1.2. As can be seen, after UGSO modification, there is a slight increase in the area of moderately and strong basic sites. The higher fraction of moderately strongly basic sites of CZ9Zr/UGSO-H, which is attributed to the higher Mg/Cu (0.46 for CZ9Zr/UGSO-H versus 0.37 for CZ9Zr/UGSO), and its higher amount of the dispersed CuO species strongly interacting with ZnO and ZrO₂, suggest a higher methanol formation for this catalyst. In addition, the significant increase in the area of strongly basic sites when UGSO-H is used as a support reveals a relationship between basicity strength and CO₂ adsorption onto oxygen vacancies on the catalyst surface.⁷⁵ As a result, CZ9Zr/UGSO-H, which has a higher number of strongly basic sites than CZ9Zr/UGSO, is expected to exhibit higher CO₂ conversion than CZ9Zr/UGSO.

Fig. 8 CO₂-TPD profiles of (A) CZ9Zr/UGSO and (B) CZ9Zr/UGSO-H catalysts.

3.3. Catalyst performance in CO₂ hydrogenation

3.3.1. The influence of Zr content on the catalytic performance in CO₂ hydrogenation. The performance of each catalyst was first evaluated at 240 °C, 20 bar and GHSV = 5000 N mL g_cat⁻¹ h⁻¹ after 4 hours on stream. Only methanol and CO were the carbon-containing products under the reaction conditions. The CO₂ hydrogenation can be described by two competed equilibrium reactions: (R1) the methanol synthesis and (R2) the reverse water-gas shift reaction (RWGS), presented as below:

As shown in Fig. 9A, CO₂ conversion is found to increase from 2.9 to 5.3% with the increase of Zr content from 3 to 12%. This can probably be attributed to (i) the enhancement in BET surface area of the catalysts, which facilitates the dispersion of active sites, (ii) the increase in Cu surface area, and (iii) the increase in CO₂ adsorption capacity. Meanwhile, methanol selectivity decreases from 45 to 26.1% with the increase of Zr content. As a result, the methanol yield follows a volcanic shape with the increase of Zr loading, reaching the highest value at 1.51% for CZ9Zr/UGSO, because the acceleration in CO₂ conversion is higher than the decrease in methanol selectivity. The methanol yield obtained over the CZxZr/UGSO catalysts shows no linear relationship with the Cu surface area, indicating that S_Cu is not the key factor in methanol formation.

Fig. 9 (A) CO₂ conversion and CH₃OH yield, (B) CH₃OH and CO selectivity of CZxZr/UGSO catalysts. Reaction conditions: P = 20 bar, T = 240 °C, H₂/CO₂ = 3, GHSV = 5000 N mL g_cat⁻¹ h⁻¹.

CO₂-TPD results (§3.1.2) and Fig. 9B suggest that CZ9Zr/UGSO (with the highest Mg/Cu ratio and proportion of moderately strong basic area) shows the highest methanol yield among the CZxZr/UGSO samples. Therefore, Mg/Cu is one of the factors determining the methanol yield. In addition, it was pointed out that the interfaces between Cu–ZrO₂ and/or Cu–ZnO are the main active sites for methanol formation, while the RWGS is primarily favored on the metallic Cu surface.^18,24,76 As confirmed by H₂-TPR and XPS analysis, the strongest synergy between Cu and Zr leading to the highest interfacial contact between Cu–ZnO and Cu–ZrO₂ observed for CZ9Zr/UGSO could explain the maximum methanol yield of this catalyst.

To evaluate the activity–temperature relationship of the prepared samples, the catalytic performance was evaluated at 200–280 °C under the same pressure and space velocity conditions (20 bar, GHSV = 5000 N mL g_cat⁻¹ h⁻¹). A commercial Cu–ZnO-based methanol synthesis catalyst (Alfa Aesar, ThermoFisher (Kandel) GmbH, containing CuO (64 wt%), ZnO (24 wt%), Al₂O₃ (10 wt%) and MgO (2 wt%)) was tested under the same reaction conditions to compare with our prepared catalysts. The CO₂ conversion at equilibrium is determined from reaction (R1) and (R2) using ProSim software. The CO₂ conversion, methanol yield, methanol and CO selectivity are plotted in Fig. 10. As shown in Fig. 10A, the CO₂ conversion of CZxZr/UGSO samples increases with the addition of Zr. However, under all reaction conditions, it is lower than that obtained from thermodynamic simulation, indicating that the reactions worked under the prevailing kinetic regime.⁷⁷ Evidently, as temperature increases from 260–280 °C, CO₂ conversion shifts to a higher value and gradually reaches near equilibrium, suggesting that CO₂ conversion can be kinetically enhanced with the elevation of temperature. Generally, methanol selectivity decreases with the increase of Zr loading, while CO selectivity shows the opposite tendency. Concerning the effect of temperature, all the samples show a decreasing methanol selectivity (Fig. 10C) and an increasing CO selectivity (Fig. 10D) when the temperature increases from 200–280 °C, due to the exothermic and endothermic nature of reactions (R1) and (R2), respectively. As an example, as temperature rises from 200–280 °C, the methanol selectivity of CZ3Zr/UGSO and CZ12Zr/UGSO decreases from 100% and 82.7% to 12.8% and 5.8%, respectively. Notably, the methanol selectivity of CZxZr/UGSO catalysts is significantly higher than that of the commercial Cu–ZnO-based methanol synthesis catalyst at the reaction temperature (200–280 °C). For all the samples, the methanol yield (Fig. 10B) shows a volcanic shape with the increase of reaction temperature. The highest values are observed for CZ9Zr/UGSO at all investigated reaction temperatures, which are indeed higher than those of the commercial Cu–ZnO-based methanol synthesis catalyst. The highest methanol yield over CZ9Zr/UGSO is attributed to (1) the highest Mg/Cu ratio, which facilitates methanol formation, and (2) the highest amount of interfacial sites between Cu–ZnO and Cu–ZrO₂ which are considered as active sites for methanol synthesis.

Fig. 10 (A) CO₂ conversion, (B) CH₃OH yield, (C) CH₃OH selectivity and (D) CO selectivity as a function of reaction temperatures for CZxZr/UGSO and commercial Cu–ZnO-based methanol synthesis catalyst. Reaction conditions: P = 20 bar, T = 240 °C, H₂/CO₂ = 3, GHSV = 5000 N mL g_cat⁻¹ h⁻¹.

3.3.2. The influence of H₂O₂ treatment on the catalytic performance in CO₂ hydrogenation. The influence of H₂O₂ activation on the catalytic performance was evaluated using the CZ9Zr/UGSO-H catalyst. Overall, CO₂ conversion increases with the increase of temperature while methanol selectivity follows an opposite tendency, which is ascribed to the thermodynamic nature of the methanol synthesis. It is observed that the CO₂ conversion of CZ9Zr/UGSO-H is higher than that of CZ9Zr/UGSO at all reaction temperatures (Fig. 11A). As an example, the differences in CO₂ conversion between CZ9Zr/UGSO-H and CZ9Zr/UGSO are 45 and 23% at 240 and 260 °C, respectively. This discrepancy can be attributed to (1) the higher surface area of the UGSO-H support which results in a higher dispersion and reducibility of Cu species; (2) the higher ratio of Cu^α/(Cu^α + Cu²⁺) (α = 0 or 1+) and Cu metallic area on the surface of reduced CZ9Zr/UGSO-H; and (3) the higher number of strongly basic sites and concentration of oxygen vacancies on the catalyst surface (which act as active sites for CO₂ adsorption) of CZ9Zr/UGSO-H. At all reaction temperatures, the methanol selectivity of CZ9Zr/UGSO is, however, slightly higher than that of CZ9Zr/UGSO-H. As a result, the methanol yield reaches the maximum value at 1.83% at 240 °C for the CZ9Zr/UGSO-H catalyst (Fig. 11D). As confirmed by CO₂-TPD (§3.2.2), the proportion of moderately basic sites of CZ9Zr/UGSO-H, being attributed to the higher Mg/Cu, can be related to the interface between Cu–ZnO and Cu–ZrO₂. Therefore, the higher methanol yield for CZ9Zr/UGSO-H is probably due to (1) the higher Mg/Cu which facilitates methanol formation and (2) the higher amount of the dispersed CuO species strongly interacting with ZnO and ZrO₂ (79.3% for CZ9Zr/UGSO-H versus 74.9% for CZ9Zr/UGSO) which are considered as active sites for methanol synthesis. Again, compared with the commercial Cu–ZnO-based methanol synthesis catalyst, the result illustrates that CZ9Zr/UGSO-H has an advantage in methanol synthesis over the commercial catalyst at the reaction temperature ranging from 200–280 °C, with the maximum methanol yield being achieved at 240 °C. The CZ9Zr/UGSO-H catalyst also stands out from several other Cu-based^77–81 and even Pd-based catalysts^82,83 developed for CO₂ hydrogenation into methanol (Table S2†). For example, CZ9Zr/UGSO-H achieves higher CO₂ conversion and CH₃OH yield compared to some catalysts (Cu/ZnAl₂O₄,⁷⁹ Cu–In–Zr–O (ref. 80) and CuO–ZrO₂ (ref. 81)) evaluated at higher pressure/temperature, illustrating advantages associated with energy saving and lower operating cost. Interestingly, the CH₃OH yield of the developed catalysts surpasses that obtained over noble metal-based catalysts (such as Pd–Ga₂O₃/SiO₂ (ref. 82) and Pd–Ca/MCM41 (ref. 83)) at lower reaction temperature and pressure. The economic advantages in terms of lower material cost (Cu and UGSO compared to Pd) and milder operating conditions indicate the effectiveness of CZ9Zr/UGSO-H as a catalyst for CO₂ hydrogenation into methanol.

Fig. 11 (A) CO₂ conversion, (B) CH₃OH yield, (C) CH₃OH selectivity, and (D) CO selectivity as a function of reaction temperatures for CZ9Zr/UGSO, CZ9Zr/UGSO-H and the commercial Cu–ZnO-based methanol synthesis catalyst. Reaction conditions: P = 20 bar, T = 240 °C, H₂/CO₂ = 3, GHSV = 5000 N mL g_cat⁻¹ h⁻¹.

Stability is a crucial indicator that determines whether the prepared catalyst can be used for the CO₂ hydrogenation to methanol process in long term operation. Herein, we tested CZ9Zr/UGSO and CZ9Zr/UGSO-H for a long run (30 h, 20 bar, T = 240 °C) to access the stability of the catalysts. Fig. 12 shows the CO₂ conversion and methanol selectivity versus the time-on-stream over the two catalysts. For CZ9Zr/UGSO, the CO₂ conversion only slightly decreases from 4.8 to 4.5%, which is 93.7% of the initial value (Fig. 12A), while it is 95.4% of the initial value for CZ9Zr/UGSO-H (Fig. 12B). This observation suggests that the latter is slightly more stable and resistant to catalytic deactivation. It is widely reported that the Cu-based catalysts are commonly deactivated by the agglomeration of Cu particles.³ The higher stability of CZ9Zr/UGSO-H compared to CZ9Zr/UGSO therefore could be attributed to the higher surface area of the UGSO-H support, which facilitates Cu dispersion.

Fig. 12 Catalyst activity of CZ9Zr/UGSO (A) and CZ9Zr/UGSO-H (B) with time-on-stream at P = 20 bar, T = 240 °C, H₂/CO₂ = 3, GHSV = 5000 N mL g_cat⁻¹ h⁻¹.

3.4. Structural changes in reduced and used catalysts

Investigation of the microstructure properties of the reduced (ex situ reduced in 50 vol% H₂/N₂, at 350 °C, for 2 h) and spent catalysts (after 4 h and 30 h on stream) was carried out by TEM analysis and the images are presented in Fig. 13. The bright void spaces between particles are attributed to the interparticle mesopores for both CZ9Zr/UGSO and CZ9Zr/UGSO-H samples, which is proved by N₂ physisorption results. It is clearly seen that the morphology of the spent samples (after 4 h and 30 h on stream) is almost the same as the reduced ones. As presented in the particle size histograms, spherical particles with an average of 15.4 nm were formed after reducing CZ9Zr/UGSO, in comparison with the reduced CZ9Zr/UGSO-H, where smaller spherical particles with a particle size of 13.3 nm were created. The smaller particle size of CZ9Zr/UGSO-H compared to that of CZ9Zr/UGSO can be assigned to the higher surface area of the UGSO-H support. After 4 and 30 h in reaction, the average particle size of spent CZ9Zr/UGSO increased to 16.1 nm and 19.1 nm, while an increase to 14.1 nm and 15.8 m is observed for spent CZ9Zr/UGSO-H, respectively. This growth in particle sizes might be attributed to the agglomeration of Cu particles.

Fig. 13 TEM images and particle size distributions of the CZ9Zr/UGSO and CZ9Zr/UGSO-H samples. (A) and (B) reduced, (C) and (D) spent 4 h, and (E) and (F) spent 30 h.

4. Conclusion

Based on the concept of a sustainable development, a series of new Cu–ZnO–ZrO₂/UGSO catalysts for CO₂ catalytic valorization into methanol were developed using a metallurgical residue (UGSO) as the catalytic support. The study was centered on the understanding of the effect of synergy between Cu/Zn/Zr and UGSO composition on the catalytic activity and stability, as well as the UGSO modification by H₂O₂ (proposed to increase the surface area of the support) on the structure–activity relationship of the developed catalyst.

Concerning the CZxZr/UGSO catalysts, with the increase of Zr content from 3 to 12 wt%, the BET surface area of the CZxZr/UGSO increased significantly, being beneficial to Cu dispersion and the increase of CO₂ conversion. Further modification of UGSO by H₂O₂ (UGSO-H) increases by 3 times the surface area and endows CZ9Zr/UGSO-H with a higher CO₂ conversion, attributable to (1) the higher surface area of the UGSO-H support (which resulted in a higher dispersion and reducibility of Cu species), (2) higher ratio of Cu^α/(Cu^α + Cu²⁺) (α = 0 or 1+) and Cu metallic area on the surface of reduced CZ9Zr/UGSO-H, and (3) higher concentration of oxygen vacancies and number of strongly basic sites on the surface of reduced CZ9Zr/UGSO-H. The methanol yield, on the other hand, was found to depend on the Mg/Cu ratio and the amount of the dispersed CuO species that strongly interact with ZnO and ZrO₂. As a result, CZ9Zr/UGSO-H offered 57% higher methanol yield at 240 °C and 20 bar in comparison with the commercial Cu-ZnO-based methanol synthesis catalyst. In addition, the developed catalyst exhibited good stability during 30 hours of operation (CO₂ conversion above 95% of its initial value).

These findings can pave a way for the design and development of efficient catalysts for other CO₂ catalytic conversion processes.

Author contributions

Thi Thanh Nguyet Vu: conceptualization, methodology, investigation, visualization, writing – original draft preparation. Alex Desgagnés: investigation, visualization, writing – original draft preparation. Pascal Fongarland: resources, funding acquisition, review. Laurent Vanoye: assistance with laboratory work. Frédéric Bornette: assistance with laboratory work. Maria C. Iliuta: conceptualization, methodology, validation, project administration, resources, writing-review & editing, funding acquisition.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

Financial support from the Ministère de l'Europe et des Affaires Étrangères (Campus France program) and the Natural Sciences and Engineering Research Council of Canada (NSERC) is gratefully acknowledged.

References

1. A. Álvarez, A. Bansode, A. Urakawa, A. V. Bavykina, T. A. Wezendonk, M. Makkee, J. Gascon and F. Kapteijn, Chem. Rev., 2017, 117, 9804–9838.
2. G. A. Olah, A. Goeppert and G. S. Prakash, J. Org. Chem., 2009, 74, 487–498.
3. T. T. N. Vu, A. Desgagnés and M. C. Iliuta, Appl. Catal., A, 2021, 118119.
4. S. Mbatha, R. C. Everson, N. M. Musyoka, H. W. Langmi, A. Lanzini and W. Brilman, Sustainable Energy Fuels, 2021, 5, 3490–3569.
5. J. Niu, H. Liu, Y. Jin, B. Fan, W. Qi and J. Ran, Int. J. Hydrogen Energy, 2022, 47(15), 9183–9200.
6. Z. Ou, J. Ran, J. Niu, C. Qin, W. He and L. Yang, Int. J. Hydrogen Energy, 2020, 45, 6328–6340.
7. X. Yang, S. Kattel, S. D. Senanayake, J. A. Boscoboinik, X. Nie, J. Graciani, J. A. Rodriguez, P. Liu, D. J. Stacchiola and J. G. Chen, J. Am. Chem. Soc., 2015, 137, 10104–10107.
8. X.-Y. Meng, C. Peng, J. Jia, P. Liu, Y.-L. Men and Y.-X. Pan, J. CO2 Util., 2022, 55, 101844.
9. O. Martin, A. J. Martín, C. Mondelli, S. Mitchell, T. F. Segawa, R. Hauert, C. Drouilly, D. Curulla-Ferré and J. Pérez-Ramírez, Angew. Chem., Int. Ed., 2016, 55, 6261–6265.
10. N. Kanoun, M. Astier and G. Pajonk, Catal. Lett., 1992, 15, 231–235.
11. R.-w. Liu, Z.-z. Qin, H.-b. Ji and T.-m. Su, Ind. Eng. Chem. Res., 2013, 52, 16648–16655.
12. P. Gao, F. Li, H. Zhan, N. Zhao, F. Xiao, W. Wei, L. Zhong, H. Wang and Y. Sun, J. Catal., 2013, 298, 51–60.
13. F. Arena, G. Italiano, K. Barbera, S. Bordiga, G. Bonura, L. Spadaro and F. Frusteri, Appl. Catal., A, 2008, 350, 16–23.
14. J. Xiao, D. Mao, X. Guo and J. Yu, Appl. Surf. Sci., 2015, 338, 146–153.
15. K. Samson, M. Sliwa, R. P. Socha, K. Góra-Marek, D. Mucha, D. Rutkowska-Zbik, J. Paul, M. Ruggiero-Mikołajczyk, R. Grabowski and J. Słoczynski, ACS Catal., 2014, 4, 3730–3741.
16. B. Xaba, A. Mahomed and H. Friedrich, J. Environ. Chem. Eng., 2021, 9, 104834.
17. V. L'hospital, L. Angelo, Y. Zimmermann, K. Parkhomenko and A.-C. Roger, Catal. Today, 2021, 369, 95–104.
18. L. Li, D. Mao, J. Yu and X. Guo, J. Power Sources, 2015, 279, 394–404.
19. S. Tada, S. Kayamori, T. Honma, H. Kamei, A. Nariyuki, K. Kon, T. Toyao, K.-i. Shimizu and S. Satokawa, ACS Catal., 2018, 8, 7809–7819.
20. K. Stangeland, H. H. Navarro, H. L. Huynh, W. M. Tucho and Z. Yu, Chem. Eng. Sci., 2021, 238, 116603.
21. R. M. Palomino, P. J. Ramírez, Z. Liu, R. Hamlyn, I. Waluyo, M. Mahapatra, I. Orozco, A. Hunt, J. P. Simonovis and S. D. Senanayake, J. Phys. Chem. B, 2018, 122, 794–800.
22. J. Nakamura, I. Nakamura, T. Uchijima, Y. Kanai, T. Watanabe, M. Saito and T. Fujitani, J. Catal., 1996, 160, 65–75.
23. C. Tisseraud, C. Comminges, T. Belin, H. Ahouari, A. Soualah, Y. Pouilloux and A. Le Valant, J. Catal., 2015, 330, 533–544.
24. K. Larmier, W. C. Liao, S. Tada, E. Lam, R. Verel, A. Bansode, A. Urakawa, A. Comas-Vives and C. Copéret, Angew. Chem., Int. Ed., 2017, 56, 2318–2323.
25. J. A. Bennett, K. Wilson and A. F. Lee, J. Mater. Chem. A, 2016, 4, 3617–3637.
26. R. Chakraborty, S. Bepari and A. Banerjee, Chem. Eng. J., 2010, 165, 798–805.
27. K. Wilson, C. Hardacre, A. F. Lee, J. M. Montero and L. Shellard, Green Chem., 2008, 10, 654–659.
28. M. Balakrishnan, V. Batra, J. Hargreaves, A. Monaghan, I. Pulford, J. Rico and S. Sushil, Green Chem., 2009, 11, 42–47.
29. L. Deng, Q. Su, Q. Ye, H. Wan, Y. He and X. Cui, Sustainable Energy Fuels, 2022, 6, 1436–1447.
30. M. Guéguin and F. Cardarelli, Miner. Process. Extr. Metall. Rev., 2007, 28, 1–58.
31. O. A. Z. Sahraei, F. Larachi, N. Abatzoglou and M. C. Iliuta, Appl. Catal., B, 2017, 219, 183–193.
32. M. Aissaoui, O. A. Z. Sahraei, M. S. Yancheshmeh and M. C. Iliuta, Int. J. Hydrogen Energy, 2020, 45, 18452–18465.
33. O. A. Sahraei, A. Desgagnés, F. Larachi and M. C. Iliuta, Appl. Catal., B, 2020, 279, 119330.
34. C. Zhong, X. Guo, D. Mao, S. Wang, G. Wu and G. Lu, RSC Adv., 2015, 5, 52958–52965.
35. C. Liu, X. Guo, Q. Guo, D. Mao, J. Yu and G. Lu, J. Mol. Catal. A: Chem., 2016, 425, 86–93.
36. A. Jain, R. Balasubramanian and M. Srinivasan, Chem. Eng. J., 2015, 273, 622–629.
37. K. Sonu, M. Sogani, Z. Syed, A. Dongre and G. Sharma, Environ. Prog. Sustainable Energy, 2021, ep13638.
38. Z. Tan, X. Zhang, L. Wang, B. Gao, J. Luo, R. Fang, W. Zou and N. Meng, Environ. Pollut. Bioavailability, 2019, 31, 198–207.
39. C. Moreno-Castilla, M. Ferro-Garcia, J. Joly, I. Bautista-Toledo, F. Carrasco-Marin and J. Rivera-Utrilla, Langmuir, 1995, 11, 4386–4392.
40. X. Dong, F. Li, N. Zhao, F. Xiao, J. Wang and Y. Tan, Appl. Catal., B, 2016, 191, 8–17.
41. H. S. Fogler, Elements of Chemical Reaction Engineering, Prentice Hall, 5th edn, 2016.
42. X. Lin, R. Li, Y. Zhang, Y. Zhan, C. Chen, Q. Zheng and J. Ma, Int. J. Hydrogen Energy, 2015, 40, 1735–1741.
43. Y. H. Choi, E. C. Ra, E. H. Kim, K. Y. Kim, Y. J. Jang, K. N. Kang, S. H. Choi, J. H. Jang and J. S. Lee, ChemSusChem, 2017, 10, 4764–4770.
44. D. S. Mathew and R.-S. Juang, Chem. Eng. J., 2007, 129, 51–65.
45. R. Dhanda and M. Kidwai, RSC Adv., 2016, 6, 53430–53437.
46. K. S. Sing, Pure Appl. Chem., 1985, 57, 603–619.
47. X. Mou, Y. Li, B. Zhang, L. Yao, X. Wei, D. S. Su and W. Shen, Eur. J. Inorg. Chem., 2012, 2012, 2684–2690.
48. M. Zhu, P. Tian, J. Chen, M. E. Ford, J. Xu, I. E. Wachs and Y. F. Han, AIChE J., 2020, 66, e16846.
49. W. Wang, Y. Xie, S. Zhang, X. Liu, M. Haruta and J. Huang, Catalysts, 2018, 8, 60.
50. R.-x. Zhou, T.-m. Yu, X.-y. Jiang, F. Chen and X.-m. Zheng, Appl. Surf. Sci., 1999, 148, 263–270.
51. S.-C. Yang, W.-N. Su, S. D. Lin, J. Rick, J.-H. Cheng, J.-Y. Liu, C.-J. Pan, D.-G. Liu, J.-F. Lee and T.-S. Chan, Appl. Catal., B, 2011, 106, 650–656.
52. X. Guo, D. Mao, G. Lu, S. Wang and G. Wu, J. Catal., 2010, 271, 178–185.
53. P. Gao, F. Li, F. Xiao, N. Zhao, N. Sun, W. Wei, L. Zhong and Y. Sun, Catal. Sci. Technol., 2012, 2, 1447–1454.
54. M. Zhu, J. Chen, L. Shen, M. E. Ford, J. Gao, J. Xu, I. E. Wachs and Y.-F. Han, Appl. Catal., B, 2020, 271, 118943.
55. N. Huo, Y. Yin, W. Liu, J. Zhang, Y. Ding, Q. Wang, Z. Shi and S. Yang, New J. Chem., 2016, 40, 7068–7074.
56. B. Zheng, S. Wu, X. Yang, M. Jia, W. Zhang and G. Liu, ACS Appl. Mater. Interfaces, 2016, 8, 26683–26689.
57. H. Guo, Q. Li, H. Zhang, F. Peng, L. Xiong, S. Yao, C. Huang and X. Chen, Mol. Catal., 2019, 476, 110499.
58. A. S. Malik, S. F. Zaman, A. A. Al-Zahrani, M. A. Daous, H. Driss and L. A. Petrov, Catal. Today, 2020, 357, 573–582.
59. F. Cocco, B. Elsener, M. Fantauzzi, D. Atzei and A. Rossi, RSC Adv., 2016, 6, 31277–31289.
60. Q. Zhang, H. Wang, P. Ning, Z. Song, X. Liu and Y. Duan, Appl. Surf. Sci., 2017, 419, 733–743.
61. T. Liu, X. Hong and G. Liu, ACS Catal., 2019, 10, 93–102.
62. R. Ladera, F. J. Pérez-Alonso, J. M. González-Carballo, M. Ojeda, S. Rojas and J. L. G. Fierro, Appl. Catal., B, 2013, 142, 241–248.
63. C. Huang, J. Wen, Y. Sun, M. Zhang, Y. Bao, Y. Zhang, L. Liang, M. Fu, J. Wu and D. Ye, Chem. Eng. J., 2019, 374, 221–230.
64. C. M. Lousada and M. Jonsson, J. Phys. Chem. C, 2010, 114, 11202–11208.
65. E. Giamello, P. Rumori, F. Geobaldo, B. Fubini and M. C. Paganini, Appl. Magn. Reson., 1996, 10, 173–192.
66. A. Amorelli, J. C. Evans and C. C. Rowlands, J. Chem. Soc., Faraday Trans. 1, 1988, 84, 1723–1728.
67. Y. Cao, L.-J. Wang, C. Feng, J.-Y. Zhang, F. Yang, S.-W. Li, S.-L. Jiang, Y. Liu, G. Yang and C.-J. Zhao, Mater. Lett., 2014, 137, 351–353.
68. B. Taleatu, E. Omotoso, C. Lal, W. Makinde, K. Ogundele, E. Ajenifuja, A. Lasisi, M. Eleruja and G. Mola, Surf. Interface Anal., 2014, 46, 372–377.
69. M. Zhu, P. Tian, M. E. Ford, J. Chen, J. Xu, Y.-F. Han and I. E. Wachs, ACS Catal., 2020, 10, 7857–7863.
70. S. Chawla, N. Sankarraman and J. Payer, J. Electron Spectrosc. Relat. Phenom., 1992, 61, 1–18.
71. J. Yu, M. Yang, J. Zhang, Q. Ge, A. Zimina, T. Pruessmann, L. Zheng, J.-D. Grunwaldt and J. Sun, ACS Catal., 2020, 10, 14694–14706.
72. M. Huš, D. Kopač, N. S. Štefančič, D. L. Jurković, V. D. Dasireddy and B. Likozar, Catal. Sci. Technol., 2017, 7, 5900–5913.
73. X.-M. Liu, G. Lu, Z.-F. Yan and J. Beltramini, Ind. Eng. Chem. Res., 2003, 42, 6518–6530.
74. S. Chen, J. Zhang, P. Wang, X. Wang, F. Song, Y. Bai, M. Zhang, Y. Wu, H. Xie and Y. Tan, ChemCatChem, 2019, 11, 1448–1457.
75. A. M. Bahmanpour, F. Héroguel, M. Kılıç, C. J. Baranowski, P. Schouwink, U. Röthlisberger, J. S. Luterbacher and O. Kröcher, Appl. Catal., B, 2020, 266, 118669.
76. E. L. Kunkes, F. Studt, F. Abild-Pedersen, R. Schlögl and M. Behrens, J. Catal., 2015, 328, 43–48.
77. Q. Jiang, Y. Liu, T. Dintzer, J. Luo, K. Parkhomenko and A.-C. Roger, Appl. Catal., B, 2020, 269, 118804.
78. L. Gao and C. Au, J. Catal., 2000, 189, 1–15.
79. M. Huš, V. D. Dasireddy, N. S. Štefančič and B. Likozar, Appl. Catal., B, 2017, 207, 267–278.
80. L. Yao, X. Shen, Y. Pan and Z. Peng, J. Catal., 2019, 372, 74–85.
81. S. Tada, K. Larmier, R. Büchel and C. Copéret, Catal. Sci. Technol., 2018, 8, 2056–2060.
82. D. L. Chiavassa, J. Barrandeguy, A. L. Bonivardi and M. A. Baltanás, Catal. Today, 2008, 133, 780–786.
83. N. Koizumi, X. Jiang, J. Kugai and C. Song, Catal. Today, 2012, 194, 16–24.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d2cy01317g

---