The CuMgAl-x catalysts with different Mg/Al molar ratios were prepared by an improved coprecipitation method as to the literature reported (Xiao et al., 2017). A certain amount of Cu(NO₃)₂·3H₂O, Mg(NO₃)₂·6H₂O, and Al(NO₃)₃·9H₂O (Mg/Al molar ratio = 1,2,3,4,5) was dissolved in 100 mL deionized water as metal salt solution A. Solution B consists of 100 mL sodium carbonate solution (0.15 M). To obtain the LDH precursor, solution A was dropped into solution B under intense agitation at 65°C. The pH during precipitation was monitored by a pH electrode and adjusted with an alkali solution of NaOH (1.0 M), a constant value of 10.0 ± 0.5. The resulting suspension was aged at 65°C for 5 h. Then the precipitate was filtered and washed with deionized water to pH 7 and dried at 100°C for 12 h. The obtained solids were the hydrotalcite-structured CuMgAl catalysts precursors. The precursors were ground to fine powder and calcined in static air at 500°C for 5 h (ramp rate of 2°C/min) in a muffle furnace. The resulting material was extruded into a circular sheet, then crushed and sieved into particles with 20–40 mesh size. The synthesized catalysts are noted as CuMgAl-x, where x (x = 1,2,3,4,5) was the Mg/Al atomic ratio of the catalyst. The content of Cu in all CuMgAl-x catalysts was 10 wt%.

Using Rigaku SmartLab SE photoelectron spectrometer with Cu Kα radiation (40 kV, 40 mA) to obtain the diffraction characteristics of the prepared material. XRD patterns were recorded for 2 h values from 5° to 90° with a scanning rate of 10°C/min.

The pore structure characteristics of the synthesized materials were obtained by N₂ adsorption-desorption method: the synthesized samples were treated at 300°C vacua for 4 h by Beishide 3H-2000PS2 physical adsorption instrument, and then cooled to −196°C for N₂ adsorption and desorption experiments, and the changes of N₂ adsorption and desorption with pressure were recorded. The BET equation was used to calculate the specific surface area of the samples, and the BJH equation was used to calculate the pore size distribution and pore volume of the samples.

TEM detection was carried out on the FEI Tecnai G2 F30 transmission electron microscope equipped with X-ray energy spectroscopy (EDX) with a resolution of 0.14 nm and a voltage of 210 kV. Before detection, the catalyst was reduced by H₂, and the sample was dissolved in anhydrous ethanol and ultrasounded for 20 min. Then the sample was dropped onto a nickel net coated with a carbon film, which was thoroughly dried for testing.

The surface morphology and structural characteristics of the catalyst were ob-served by Hitachi SU8010 scanning electron microscope (SEM) with an operating volt-age of 5 kV.

H2 temperature-programmed reduction (H₂-TPR) was carried out on AutoChem1 Ⅱ 2,920 equipped with a thermal conductivity detector (TCD) to analyze the reduction characteristics of the synthesized samples. Typically, the samples (about 150 mg) need to be pretreated under flowing argon for 2 h at 300°C to remove physically adsorbed water. As the temperature cooled down to 50°C, the samples were heated from 50°C to 600°C at a rate of 5°C/min in 10% H₂ flow (balanced with Ar).

The surface basic sites or acid sites of the samples was analyzed by the CO₂ and NH₃ temperature-programmed desorption (TPD) experiment on ChemStar TPx chemisorption analyzer. 0.1 g sample was first reduced for 2 h at 300°C in a 10 vol% H₂/N₂ flow. After cooling to 50°C, the sample was saturated with CO₂ or NH₃ (30 mL/min) for adsorption for 0.5 h and then flushed with He (40 mL/min) for 0.5 h. Thereafter, the experiment was tested with a heating rate of 5°C/min under He flows.

X-ray photoelectron spectroscopy (XPS) was measured by Kratos XSAM800 spec-trometer, and the elements contained in the samples were tested by Al Kα rays (12 kV, 15 mA, hv = 1,486.6 eV). The experimental binding energies were calibrated according to C1s (284.6 eV).

Inductively coupled plasma optical emission spectroscopy (ICP-OES) was used to determine the element content in the catalyst.

The copper dispersion (D_Cu), the specific surface area of exposed Cu in the sample (S_Cu), and the average Cu particle size (d_Cu) were determined by the N₂O-TPR method. Typically, 100 mg samples were pretreated in N₂ flow at 300°C for 1 h, then cooled down to 50°C. Thereafter, the samples were heated to 300°C at a heating rate of 5°C/min. The samples then were reduced with 10 vol% H₂/N₂ (30 mL/min) at 300°C for 1 h, the cor-responding consumption of H₂ is defined as X. Then, the reduced samples were exposed to 2 vol% N₂O/He (30 mL/min) at 60°C for 1 h, ensuring the surface metal Cu was completely oxidized to Cu₂O. After that, the samples were purged again with N₂ for 30 min and cooled to room temperature. Finally, the sample was heated to 300°C in a 10 vol% H₂/N₂ (30 mL/min) at the same temperature ramping, where the amount of H₂ consumption was defined as Y. The D_Cu, S_Cu, and d_Cu were calculated by the following Eqs 1–3: D C u = 2 Y X × 100 % (1) S C u = 2 Y × N 1.4 × 10 19 × W × X (2) d C u = 1.1 D C u (3) where N represents the Avogadro constant (6.02 × 10²³ atom mol⁻¹), and W represents the atomic mass of copper (63.546 g/mol).

In situ diffusion reflectance infrared Fourier transform spectroscopy (in situ DRIFTS) was recorded on the Nicolet 6,700 spectrometer with an MCT detector. For CO adsorption analysis, the sample was reduced in situ at 300°C H₂ (20 mL/min) for 1 h, scanned in Ar for 30 min, and then cooled to room temperature for CO adsorption for 30 min. Then Ar scanning was performed again to remove the adsorbed CO substance for 30 min, and the spectrum was collected. The experimental procedure of For in-situ FTIR spectroscopy of CO₂ hydrogenation, the catalyst was reduced in situ in H₂ (20 mL/min) at 300°C for 1 h, then purged with He for 30 min to remove the physically adsorbed H₂, and then cooled to 40°C to collect the background spectrum. Finally, He was replaced with a mixture of raw materials (CO₂:H₂:N₂ = 24:72:4), and the process of the intermediate product changing with time was recorded, and the infrared spectrum with a resolution of 4 cm⁻¹ was obtained.

Methanol synthesis via CO₂ hydrogenation was carried out in a high-pressure continuous flow fixed-bed reactor. Before the reaction of CO₂ hydrogenation to methanol, 0.5 g catalyst mixed with 2.0 g quartz sand (both 20–40 mesh) was loaded into a stainless steel reactor tube with an inner diameter of 10 mm. Firstly, the catalyst was reduced by H₂ (50 mL/min) at 300°C for 4 h. Then, the reactor was cooled to 200°C, and the feed gas (24% CO₂, 72% H₂, and 4% N₂) with the gas hourly space velocity (GHSV) of 9,000 mL⋅g_cat ⁻¹⋅h⁻¹ was introduced in the reactor that was pressurized to 2.5 MPa. All pipelines and valves were heated to 140°C by a heating belt to prevent condensation of gas products. The products were quantitatively analyzed online by a Shimadzu GC-2014 gas chromatograph equipped with a thermal conductivity detector (TCD) and a flame ionization detector (FID). The CO₂ conversion rate (X_CO2), selectivity of CH₃OH(S_CH3OH) and space-time yields of CH₃OH(STY_CH3OH, g⋅kg_cat ⁻¹⋅h⁻¹) were defined as the following Eqs 4–6: X C O 2 = C O 2 i n − C O 2 o u t C O 2 i n × 100 % (4) S C H 3 O H = C H 3 O H o u t C O 2 i n − C O 2 o u t × 100 % (5) S T Y C H 3 O H = F C O 2 , i n × X C O 2 × S C H 3 O H × M C H 3 O H W c a t (6)

Where, [CO₂]_in and [CO₂]_out represent the input and output of carbon dioxide (mol), [CH₃OH]_out represents the output of methanol (mol), F_CO2,_in represents the molar input flow of carbon dioxide (mol⋅h⁻¹), and W_cat represents the mass of catalyst (g). M_CH3OH represents the molar mass of methanol (g⋅mol⁻¹).

The N₂ adsorption-desorption isotherm and pore size distribution of CuMgAl-x catalysts are shown in Figure 1. Figure 1A showed that all samples had typical type IV adsorption isotherms with different hysteresis curves, indicating that CuMgAl-x catalysts had a typical mesoporous structure (Zeng et al., 2013). CuMgAl-1 and CuMgAl-2 catalysts exhibited typical H3-type hysteretic loops, which were related to the porosity and slit caused by the aggregation of lamellar particles. In addition, H2-type hysteretic loops can be observed on CuMgAl-4 and CuMgAl-5 samples, which were often associated with complex and interconnected pore structures (León et al., 2010). However, the isotherms of CuMgAl-3 samples showed an intermediate shape between CuMgAl-2 and CuMgAl-4. The pore structure of the sample changed obviously with the increase of the Mg/Al molar ratio. The chemical composition and structural characteristics of the samples are listed in Table 1. CuMgAl-3 had the highest S_BET of 192.3 m²g⁻¹ among all samples, and the S_BET value decreased significantly with the further increase of the Mg/Al ratio. On the whole, the change of pore diameter was opposite to S_BET (Figure 1B). The variation of total specific surface area, pore volume, and average pore size with Mg/Al molar ratio may be related to the difference in pore structure.

The morphology of calcined CuMgAl-x catalysts was characterized by scanning electron microscope (SEM), and the results are shown in Figure 2. Figure 2A shows the SEM images of the CuMgAl-1 catalyst, the surface of the sample was inlaid with many irregularly shaped nanosheets. In contrast, the existence forms of nanosheets on CuMgAl-2 and CuMgAl-3 samples were quite different, which was reflected in the cross of nanosheets and the formation of flower-like spheres, which was related to the formation of pores (Shao et al., 2020). It can be found that the flower-like sphere structure formed by nanosheets on CuMgAl-3 samples was more regular, which was conducive to the exposure of pores. When the Mg/Al ratio increased to 4 and 5, the flower-shaped spheres gradually disappeared and formed a more complex pore structure. These results were consistent with BET results, indicating that the morphology of samples varied with the Mg/Al ratio.

The CuMgAl-x catalysts were characterized by X-ray diffraction (XRD) to detect the crystallinity of the phases, and the results are shown in Figure 3. Figure 3A shows the XRD patterns of the CuMgAl-x precursors. The X-ray characteristic diffraction peaks at 2θ = 11.4°, 22.6°, 34.8°, 38.8°, 45.3°, 60.4° and 61.7° were respectively attributed to the (003), (006), (012), (015), (018), (110) and (113) crystal planes of Mg-Al hydrotalcite structure (JCPDS35-0965), indicating that the hydrotalcite structure was formed successfully in the CuMgAl-x catalysts precursors (Shao et al., 2019). The peak strength of the hydrotalcite crystal plane of CuMgAl-x catalysts precursors gradually increased with the decrease of the Mg/Al molar ratio and reached the maximum value when Mg/Al = 3, which means that CuMgAl-3 had the best crystallinity. Since the charge density of Al³⁺ was higher than that of Mg²⁺, the increase of Al³⁺ contents increased the charge density on the layer, resulting in the regular layered structure and strong interlayer interaction of CuMgAl-3 (Shi et al., 2021). The crystallinity of the sample gradually decreases with the further decrease the Mg/Al ratio, which may be caused by the decrease of stability of hydrotalcite caused by excessive Al. The diffraction peak of aluminum hydroxide was only observed in the CuMgAl-1 sample, which may be due to the large amount of Al³⁺ can not enter the CuMgAl hydrotalcite structure and precipitated in the form of Al(OH)₃.

Figure 3B shows the XRD patterns of the calcined CuMgAl-x catalysts at 500°C. After the calcination of precursors, the hydrotalcite structure disappeared in the XRD patterns, indicating that the hydrotalcite structure was destroyed by high-temperature calcination. The X-ray characteristic diffraction peaks at 2θ = 36.5°, 43.1°, 62.7°, and 78.9° correspond to the (111), (200), (220) and (222) crystal planes of MgO (Huang et al., 2015), respectively, and the characteristic peak intensity of MgO increased with the increase of Mg/Al molar ratio. The X-ray diffraction pattern of the CuMgAl-1 calcined sample showed some peaks corresponding to Al₂O₃, which was consistent with the diffraction pattern for the analysis of hydrotalcite precursors. It seemed that after the calcination process, some of the Al³⁺ in the laminates became part of the lattice while others remained separated, resulting in the production of aluminum oxide. However, as the Mg/Al molar ratio increases, no characteristic peaks associated with CuO, CuAl₂O₄ spinel, and Al₂O₃ are detected, possibly because Cu²⁺ combined with Al³⁺ into the MgO structure to form Mg(Cu, Al)O solid solutions, indicating that CuO has good dispersion on Mg(Al)O carriers (Li et al., 2016).

The morphology and Cu particle size distribution of CuMgAl-x catalysts after reduction were observed by transmission electron microscopy (TEM), as shown in Figure 4. The average Cu particle sizes for each sample are shown in Table 2. It was evident that CuMgAl-3 exhibited a minimum mean particle size of 5.57 nm and a good distribution of Cu particles due to the hydrotalcite structure with high crystallinity, which was conducive to the reduction of Cu particles. In summary, the average particle size of all samples was sorted as follows: CuMgAl-5 (12.02 nm)>CuMgAl-1 (10.13 nm)>CuMgAl-4 (8.25 nm)>CuMgAl-2 (7.95 nm)>CuMgAl-3 (5.57 nm). It was worth noting that neither too high nor too low Mg/Al ratio can obtain satisfactory particle distribution and small particle size of Cu, which indicates that the Mg/Al ratio has a significant effect on the size and distribution of Cu metal particles. The dispersion of active components can be further analyzed by EDX characterization. As shown in Supplementary Figure S1, Cu, Mg, Al, and O elements were uniformly dispersed in the CuMgAl-3 catalyst. Figures 4F, G showed the HRTEM images of CuMgAl-3, in which it can be observed that the lattice fringe of Cu(111) and MgO(220) were in the same region and presented a state of mutual combination, which proved that there is a strong interaction between Cu and MgO (Chen et al., 2024).

The reduction performance of CuMgAl-x catalysts was analyzed through the characterization of hydrogen temperature programmed reduction (H₂-TPR). As shown in Figure 5, all samples exhibited two reduction peaks, where the α peak at around 181°C was attributed to the reduction of highly dispersed CuO particles, while the β peak at higher temperatures (>210°C) was attributed to the reduction of bulk CuO interacting with MgO or Al₂O₃ (Chen et al., 2024). Obviously, CuO species on the surface of CuMgAl-3 catalyst had better dispersion than other catalysts. When the Mg/Al molar ratio increased from 1 to 3, the reduction temperature of the β peak decreased from 235°C to 224°C, which may be due to the strong interaction between CuO and MgO, which weaken the strength of the Cu-O bond. However, as the Mg/Al molar ratio increased from 3 to 5, the reduction temperature of CuO increased again, which was related to the accumulation of CuO.

It is well known that the performace of CO₂ hydrogenation to methanol is also closely related to the acidity of the catalyst surface (Zhou et al., 2021), which can be detected by NH₃ temperature-programmed desorption (NH₃-TPD). The NH₃-TPD profiles of CuMgAl-x catalysts are illustrated in Figure 7. Using the Gaussian curve fitting method, the peaks are also deconvoluted to obtain the semiquantitative analysis. The calculated results are listed in Table 4. The desorption peaks at about 155°C belong to weak acid sites (Ⅰ peak), the desorption peaks at 301°C belong to medium acid sites (Ⅱ peak), and the desorption peaks above 446°C belong to strong acid sites (Ⅲ peak), which should be attributed to the Lewis acidity of MgO-Al₂O₃ mixed oxide (Xia et al., 2016). It is evident that when the Mg/Al molar ratio rises, the amount of total acidic sites of CuMgAl-x catalysts falls noticeably, and the CuMgAl-1 catalyst with the lowest Mg concentration showed the highest amount of total acidic sites. The explanation for this is that the presence of MgO can lessen the acidity of catalysts. Moreover, Table 4 shows that the proportion of acid sites at the two high temperature peaks (II + III) of CuMgAl-3 is 82.4%, which is higher than other catalysts.

XPS analysis of all catalysts was carried out to investigate the important influence of catalyst surface Cu state on the catalytic behavior. Figure 8A shows the XPS spectra of all CuMgAl-x catalysts after reduction. In the Cu 2p spectrum, the characteristic peaks at 932.8 and 952.5eV corresponded to the reduced Cu⁰/Cu⁺, and the characteristic peaks at 934.8 and 954.7 corresponded to the Cu²⁺ species (Song et al., 2023b). The binding energy of Cu 2p_3/2 of CuMgAl-x catalyst (932.8–933.1 eV) was higher than the standard value (932.6 eV), which indicated the existence of electron transfer from Cu to MgO, forming a metal-support strong interaction (MSI) (Wang et al., 2023b). In order to further study the chemical state on the surface of the reduction catalyst, the distribution of Cu species was further studied by in-situ infrared spectroscopy using CO as the probe molecule of the irreversible adsorption reaction, and the results were shown in Figure 8B. After CO adsorption and He purification, the physically adsorbed CO disappeared, while the chemisorbed CO bands remained near 2,148 cm⁻¹ and 2096 cm⁻¹, which were attributed to the linear stretching of Cu⁺-CO and Cu⁰-CO species (Nielsen et al., 2021), respectively. This indicated that copper oxide was reduced to Cu⁺/Cu⁰ pairs rather than fully reduced to Cu⁰ species under the reduction conditions in this study. Supplementary Figure S2 displays the Cu LMM Auger transition spectra of the reduced CuMgAl-x catalysts. The characteristic peaks near 912.7 eV and 918.5 eV are responsible for the signals of Cu⁺ and Cu⁰, respectively. The percentage of Cu⁺ and Cu⁰ content can be estimated by integrating the two characteristic peaks (Lu et al., 2019; Zhao et al., 2021). As shown in Figure 8C, a volcanic trend was presented between the ratio of Cu⁺/(Cu⁺+Cu⁰) and the molar ratio of Mg/Al. Among all the samples, CuMgAl-3 reached the highest value of 36.7%. This difference allowed us to relate the interaction between CuO and MgO, and the stronger interaction was conducive to the formation of more Cu⁺ species in the Cu-MgO interface, which has been confirmed by many researchers (Wang et al., 2023a; Chen et al., 2023; Zhang et al., 2023). Figures 8D–F) showed the XPS spectra of Mg 2p, Al 2p and O 1s, respectively. With the increase of Mg content, Mg 2p gradually moved to the direction of low binding energy, and Al 2p basically remained unchanged. In the O 1s spectrum, two characteristic peaks were observed at 531.9 and 533.0 eV, belonging to lattice oxygen (O_latt) and adsorbed oxygen (O_ads), respectively (Zhang et al., 2023). Obviously, with the increase of Mg content, O_latt and O_ads moved to the direction of low binding energy, which was caused by the strong interaction between Cu and MgO (Chen. et al., 2024; Sha et al., 2021).

Figure 9 shows the catalytic performance of CO₂ hydrogenation to methanol on the prepared CuMgAl-x catalysts. It can be noted in Figure 9A that the conversion of CO₂ increased with the increase in reaction temperature, which indicated that high tem-perature was conducive to the activation and conversion of CO₂. It can be found in Figure 9B that the selectivity of CH₃OH decreased with the increase in temperature, which was contrary to the change trend of CO₂ conversion. In fact, there are two important competitive reactions in CO₂ hydrogenation to methanol. The first one is the methanol synthesis, and the second one is the reverse water gas shift (RWGS) reaction (Singh et al., 2021; Stangeland et al., 2021). The equilibrium of these reactions can be described as Eqs 7, 8: CO 2 + 3 H 2 ↔ CH 3 OH + H 2 O   Δ H = ‐ 49.5 kJmol ‐ 1 (7) CO 2 + H 2 ↔ CO + H 2 O   Δ H = + 41.2 kJmol ‐ 1 (8)

As shown in reaction (7), methanol synthesis was an exothermic reversible reaction and its equilibrium constant decreased with the increase in reaction temperature. In addition, compared with the RWGS reaction (8), the methanol synthesis reaction had lower apparent activation energy. Therefore, in the whole temperature range, the selectivity of CH₃OH decreased with the increase in reaction temperature. As shown in Figure 9C, according to the data of CO₂ conversion and methanol selectivity, the relationship between space-time yield of methanol (STY_CH3OH) and Mg/Al molar ratio was obtained with a volcanic relationship, which indicated that the Mg/Al molar ratio had a greater influence on the performance of CuMgAl-x catalysts. The order of STY_CH3OH of all catalysts at this temperature was as follows: CuMgAl-3 (418.9 g⋅kg_cat. ⁻¹⋅h⁻¹)>CuMgAl-2 (323.8 g⋅kg_cat. ⁻¹⋅h⁻¹)>CuMgAl-4 (269.4 g⋅kg_cat. ⁻¹⋅h⁻¹)>CuMgAl-1 (209.6 g⋅kg_cat. ⁻¹⋅h⁻¹)>CuMgAl-5 (147.6 g⋅kg_cat. ⁻¹⋅h⁻¹). The catalytic performance of all catalysts was measured under the conditions of temperature(T) = 240°C, pressure(P) = 2.5 MPa, WHSV = 9,000 mL⋅g_cat ⁻¹⋅h⁻¹, CO₂/H₂/N₂ (molar ratio) = 24:72:4. The apparent activation energy (E) was calculated by the Arrhenius equation (Lu et al., 2018; Zhang et al., 2024a), as shown in Figure 9D. Among all catalysts, CuMgAl-3 showed the lowest apparent activation energy of 20.1 kJ/mol.

The operational stability of catalysts is one of the key issues in catalyst development. The catalytic performance of the CuMgAl-3 catalyst changes over time, as shown in Supplementary Figure S3. Both CO₂ conversion and methanol selectivity showed good stability during the catalysis test for 150 h, indicating that the active site in the CuMgAl-3 catalyst remained stable. As shown in Supplementary Figure S4, the spent-CuMgAl-3 catalyst was characterized by XRD and compared with the fresh-CuMgAl-3 catalyst. It can be seen from the results that the XRD pattern of the spent-CuMgAl-3 catalyst was similar to that of the fresh-CuMgAl-3 catalyst, and no obvious characteristic peak of metal Cu was found, indicating that Cu particles were well dispersed after the reaction. As shown in Supplementary Figure S5, the TEM characterization of the used CuMgAl-3 catalyst showed that the Cu particle size after the reaction (7.48 nm) was slightly increased compared with that before the reaction (5.57 nm), and there was no obvious sintering, indicating that the Cu particles in CuMgAl-3 catalyst still had good dispersion after the reaction.

The above results indicated that a suitable molar ratio of Mg/Al was beneficial to improve the catalytic performance of CuMgAl-x catalyst for CO₂ hydrogenation to methanol. CuMgAl-3 catalyst with Mg/Al molar ratio of 3 had the maximum CO₂ conversion (14.3%) and methanol selectivity (94.5%) at 240°C and 2.5 MPa. CuMgAl-3 catalyst with the best performance had the highest Cu dispersion (32.4%) and the largest Cu surface area (S_Cu, 31.1 m²/g), exposing the most Cu species, which was conducive to the interaction of Cu particles with MgO and formed more Cu-MgO active interfaces. As shown in Figure 10A, there was a positive correlation between STY_CH3OH and S_Cu. The high S_Cu provided more active sites for the adsorption and activation of CO₂ and H₂, thus obtaining higher catalytic activity (Dharmalingam et al., 2023). However, the relationship between STY_CH3OH and S_Cu was not linear, which proved that not only does the S_Cu have an effect on catalytic activity, but other factors also have an effect on catalytic activity. In addition, the electronic state of Cu also affected the catalytic performance of Cu-based catalysts (Zheng et al., 2023). It can be seen from Figure 8 that both Cu⁰ and Cu⁺ existed on the CuMgAl-x catalyst, and the ratio of Cu⁺/(Cu⁺+Cu⁰) can be adjusted by changing the Mg/Al molar ratio. Generally, Cu⁰ is exposed to the surface of the catalyst and played a fundamental role in the adsorption and activation of H₂, while Cu⁺ produced by the strong interaction between Cu and the carrier plays a positive role in the adsorption of CO, which effectively improves the selectivity of methanol. (Samson et al., 2014; Song et al., 2023a). Figure 10B shows the relationship between STY_CH3OH and the ratio of Cu⁺/(Cu⁺ +Cu⁰), and the relationship between the two was linear, which indicated that Cu⁺ had a greater influence on the performance of the catalyst.

In order to further analyze the catalytic performance of the CuMgAl-3 catalyst, CuMgAl-3 catalyst were compared with some typical catalysts for CO₂ hydrogenation to methanol. Based on the STY_CH3OH, the results were listed in Supplementary Table S1. The catalytic performance of the CuMgAl-3 catalyst was better than that of traditional Cu-based catalysts (Lei et al., 2016; Xiao et al., 2017; Chen et al., 2022) and the catalysts containing Mg and Al elements at the same time (Ren et al., 2015; Fang et al., 2019a; Fang et al., 2019b; Cored et al., 2022). The STY_CH3OH of CU-0.5-300 (Liu et al., 2020) and Pd/In₂O₃ (Rui et al., 2017) was higher than that of CuMgAl-3, mainly because the activity tests of these catalysts were performed at higher GHSV, pressure, or temperature. In particular, the increase in the GHSV can significantly increase the methanol production capacity, resulting in higher STY_CH3OH. In summary, CuMgAl-3 has excellent catalytic performance with the advantages of low cost, green environmental protection, and easy industrial production, so it can be used as an ideal catalyst for methanol synthesis.

In situ DRIFTS experiments were conducted on CuMgAl-1, CuMgAl-3 and CuMgAl-5 catalysts using CO₂ + H₂ as reactants to explore the possible mechanism of CO₂ hydrogenation. Figure 11 shows the transient evolution of major surface substances during CO₂ hydrogenation of these catalysts at 40°C–320°C. For the CuMgAl-1 catalyst (Figure 11A), three characteristic peaks were found at 1431,1538 and 1,646 cm⁻¹ are attributed to carbonate (CO₃*) and bicarbonate (HCO₃*) at 40°C, and no other characteristic peaks were found at this temperature (Hartadi et al., 2016). When the temperature rises further to 120°C, the characteristic peak intensity of formate (the peaks at 1,601, 1,362, 2,872 and 2,968 cm⁻¹attributed to the symmetric vibration of ν_s (OCO), the symmetric OCO stretching vibrations ν_s (OCO), the CH stretching vibrations ν(CH), the CH bending δ(CH) and asymmetric OCO stretching vibration ν_as (OCO), respectively) increases rapidly, but the characteristic peak of the methoxy species (the peaks at 2,928 and 1,040 cm⁻¹ attributed to the asymmetric CH₃ stretching vibration ν_as (CH₃) and the OCO stretching vibrations ν(CO), respectively) does not appear until 160°C (Yan et al., 2022; Wang et al., 2023b; Tian et al., 2023). When the temperature gradually increases to 240°C, the conversion of formate (HCOO*) to CH₃O* substance was promoted, which is characterized by a gradually stronger signal of CH₃O* and a gradually weaker band of HCOO*. At the same time, it was found that the strength of bicarbonate (HCO₃*) reached the maximum when the temperature rose to 120°C, but there was no obvious change in strength when the temperature continued to rise, indicating that the ability of HCO₃* to convert into HCOO* on CuMgAl-1 was weak. In addition, the characteristic peak of CO was found in 2013 cm⁻¹, indicating the existence of RGWS reaction, and high temperature was conducive to the generation of CO (Kattel et al., 2016).

For the CuMgAl-3 catalyst (Figure 11B), the signal strength of CO₃* and HCO₃* was stronger than that of CuMgAl-1 at 40°C, indicating that the CuMgAl-3 had stronger CO₂ adsorption capacity, which was consistent with the conclusion in Figure 6. With the increase of temperature, the conversion of CO₃* species to HCOO* species can still be found, but the signal strength of the CH₃O* group on CuMgAl-3 was significantly greater than that on CuMgAl-1, which was consistent with less CO₃* accumulation on CuMgAl-3. In addition, the HCO₃* characteristic peak on CuMgAl-3 completely disappeared at 200°C. This phenomenon indicated that CO₃* and HCO₃ ⁻ species on CuMgAl-3 were more easily converted to formates at low temperatures, which was conducive to further conversion to methanol, which may relate to the stronger Cu-MgO interaction on CuMgAl-3 catalyst. It can be seen from the H₂-TPD (Supplementary Figure S6) results that CuMgAl-3 had the largest H₂ desorption peak at low-temperature, indicating that it had the best H₂ activation capacity because the strong Cu-MgO interaction was conducive to the formation of highly dispersed Cu nanoparticles that promote H₂ dissociation, which will promote the hydrogenation of intermediate species CO₃* and HCO₃ ⁻ to HCOO*. At the same time, it was found that the accumulation of CO in CuMgAl-3 at high-temperature was smaller than that in CuMgAl-1, which was because the formation of more Cu⁺ enhances the adsorption of CO, thus improving the selectivity of methanol (Yao et al., 2019). For the CuMgAl-5 catalyst (Figure 11C), the HCOO* species and CH₃O* group strength were lower than CuMgAl-1 and CuMgAl-3 during temperature rise, and the CO strength was the highest at high-temperature, indicating that excessive Mg/Al has a negative effect on the catalytic performance.

Based on the above discussion, there is no doubt that the interaction between active Cu and MgO on CuMgAl-3 catalyst plays an irreplaceable role in the hydrogenation of CO₂ to the methanol synthesis. The Cu-MgO interaction on CuMgAl-x catalysts can be changed by changing the Mg/Al ratio. It was well known that catalysts with higher D_Cu are more dominant in stability, and smaller metal particles tend to provide more active sites, thus promoting adsorption and activation of reactants at the interface (Zhong et al., 2020). Figure 12 describes the mechanism comparison of CuMgAl-x catalysts for CO₂ hydrogenation to methanol before and after the optimization of Mg/Al ratio. In this study, the diversity of basic sites on CuMgAl-x catalyst was affected by MSI. The interaction between Cu and MgO was enhanced when the Mg/Al ratio increased from 1 to 3. However, the number of basic sites of the catalyst decreased when further increasing the Mg/Al ratio, which due to the reduction of crystallinity of the hydrotalcite structure in the catalyst precursor by excess Mg. Therefore, the good alkalinity of the CuMgAl-3 catalyst was conducive to the adsorption of CO₂. In addition, when the Mg/Al ratio was optimized to 3, the Cu with higher dispersion and smaller particle size was obtained on the CuMgAl-3 catalyst due to the strong interaction between Cu and MgO, which made H₂ more easily dissociated, the overflow of H allows CO₃* and HCO₃ ^* to be quickly converted to HCOO^*, which facilitates further conversion to CH₃O^* and CH₃OH. At the same time, the strong interaction between Cu and MgO was conducive to the formation of Cu⁺, which can inhibit the desorption of CO in RWGS reaction to a certain extent, and improve the selectivity of methanol. Therefore, the higher methanol production efficiency of CuMgAl-3 was the result of the adsorption of CO₂ by more basic sites, the high H activation ability, and the inhibition effect of Cu⁺ on CO desorption.