CO2 Activation and Hydrogenation on Cu-ZnO/Al2O3 Nanorod Catalysts: An In Situ FTIR Study

CuZnO/Al2O3 is the industrial catalyst used for methanol synthesis from syngas (CO + H2) and is also promising for the hydrogenation of CO2 to methanol. In this work, we synthesized Al2O3 nanorods (n-Al2O3) and impregnated them with the CuZnO component. The catalysts were evaluated for the hydrogenation of CO2 to methanol in a fixed-bed reactor. The support and the catalysts were characterized, including via in situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS). The study of the CO2 adsorption, activation, and hydrogenation using in situ DRIFT spectroscopy revealed the different roles of the catalyst components. CO2 mainly adsorbed on the n-Al2O3 support, forming carbonate species. Cu was found to facilitate H2 dissociation and further reacted with the adsorbed carbonates on the n-Al2O3 support, transforming them to formate or additional intermediates. Like the n-Al2O3 support, the ZnO component contributed to improving the CO2 adsorption, facilitating the formation of more carbonate species on the catalyst surface and enhancing the efficiency of the CO2 activation and hydrogenation into methanol. The synergistic interaction between Cu and ZnO was found to be essential to increase the space–time yield (STY) of methanol but not to improve the selectivity. The 3% CuZnO/n-Al2O3 displayed improved catalytic performance compared to 3% Cu/n-Al2O3, reaching a CO2 conversion rate of 19.8% and methanol STY rate of 1.31 mmolgcat−1h−1 at 300 °C. This study provides fundamental and new insights into the distinctive roles of the different components of commercial methanol synthesis catalysts.


Introduction
In the last two decades, global warming and related extreme weather conditions have become topical issues, and the mitigation of carbon dioxide (CO 2 ) release into the atmosphere is a top global priority. CO 2 is a major anthropogenic greenhouse gas (GHG) released during fossil fuel exploitation; thus, controlling CO 2 emissions is critical to maintaining the carbon-neutral state of the atmosphere [1][2][3]. The catalytic reduction of CO 2 with renewable hydrogen (H 2 ) to value-added chemicals and fuels such as hydrocarbons and alcohols represents a potential strategy to mitigate CO 2 emissions into the atmosphere [1,[4][5][6]. For example, the hydrogenation of CO 2 to methanol is an important reaction in CO 2 conversion, as methanol can be used as a feedstock chemical and fuel [7]. However, CO 2 is a thermodynamically inert molecule that needs high reaction temperatures to activate the C=O bond. At high temperatures, the undesired reverse water-gas shift (RWGS) reaction is thermodynamically favored, which reduces the product selectivity. Therefore, developing a catalyst that can efficiently drive the process under mild conditions is imperative.
Different types of catalysts have been investigated for the thermocatalytic hydrogenation of CO 2 to methanol, including supported metal [8,9] and metal oxide [10,11] catalysts. Cu-based catalysts have been reported to have the best activity for methanol production under industrially relevant conditions (5-10 MPa and 200-300 • C) [12,13]. The Cu-ZnO-Al 2 O 3 catalyst prepared via the co-precipitation method is the industrial catalyst used for methanol production from CO + H 2 ; thus, the Cu-ZnO-based catalysts are widely investigated for CO 2 hydrogenation to methanol as well [14,15]. Various supports and promoters are explored for these catalysts, and different catalyst structures and active sites are reported [15][16][17]. The common supports or promoters for Cu-based catalysts include ZnO, ZnO-Al 2 O 3 , ZrO 2 , TiO 2 , In 2 O 3 , CeO 2 , and SiO 2 [18,19]. At times the contribution of the support is complex. For example, Nitta et al. [18] observed significantly improved selectivity towards methanol formation at a low temperature upon introducing ZnO into Cu/ZrO 2 . However, at high temperatures, adding ZnO promoted the methanol decomposition into CO, leading to a reduced yield for methanol.
Factors impacting the overall catalytic performance of the Cu-based catalysts include the nature of the support material, the Cu loading, the dispersion, and the preparation method [12,20]. However, their effects on the activation of CO 2 are still not clear. The specific roles of the different components within the industrial Cu-ZnO-Al 2 O 3 catalyst for methanol synthesis from CO 2 were assigned in the literature [12,18]. ZnO is generally believed to be a multitasked component, acting as a physical spacer for the Cu particles and facilitating H 2 dissociation; Al 2 O 3 plays the role of a structural promoter that increases Cu dispersion, whereas Cu drives the selectivity towards methanol [21,22]. However, further investigation is required to ascertain the individual roles of the catalyst components concerning CO 2 activation, as well as to better understand the observed synergetic effects within the complete system. In addition to the composition, the nature of the catalyst support and characteristics such as the morphology (size and shape), were also found to influence the CO 2 conversion efficiency and selectivity towards methanol formation [23][24][25][26]. An et al. [27] found that the catalytic activities of Cu/Zn/Al/Zr catalysts depended strongly on the morphology of the support, and the utilization of a fibrous shape improved the hydrogenation performance significantly. In addition, the proper selection of the catalyst support may provide additional advantages, such as reducing the amount of required active metal, inhibiting the sintering of the active metal, and prolonging the stability [22].
Diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) is a powerful technique for analyzing in situ substrate or adsorbate interactions, and has been previously utilized to characterize systems where the interactions of gases over the surface of a catalyst elucidated the catalytic surface chemistry [28][29][30][31].
In this work, we explored porous Al 2 O 3 nanorods as the catalyst support in the CuZnO/Al 2 O 3 catalysts for CO 2 hydrogenation. The Al 2 O 3 nanorods were synthesized via the steam-assisted solid wet-gel method and deployed as a support for CuZnO. The CuZnO/Al 2 O 3 catalyst was prepared using the incipient wetness impregnation method and evaluated for the hydrogenation of CO 2 to methanol. The catalysts supported on alumina nanorods exhibited better CO 2 reduction to methanol under the examined conditions than that supported on commercial alumina. Through a DRIFTS study, we gained a new understanding of the specific roles of Al 2 O 3 , ZnO, and Cu in CO 2 adsorption, activation, and hydrogenation reaction. This information will contribute to the design of catalysts with improved performance for industrial applications.

Sample Preparation
(i) Nanorods Al 2 O 3 preparation. Al 2 O 3 nanorods were synthesized using the steam-assisted solid wet-gel method according to the literature [32]. Typically, 30 g Al(NO 3 ) 3 ·9H 2 O (>99.9%, SCR, Sinopharm Chemical Reagent Co., Ltd., Shanghai, China) was dissolved in 50 mL of deionized water. Approximately 15% NH 4 OH aqueous solution (36%, SCR, Sinopharm Chemical Reagent Co., Ltd., Shanghai, China) was added dropwise to the aluminum nitrate solution at room temperature under stirring to control the pH at 5.0. The resulting solid precipitate was recovered by filtration. The as-prepared solid cake-like wet gel was transferred to a glass beaker (25 mL) sitting in a Teflon vessel (200 mL), where 2 g of water was poured into the bottom of the vessel but was physically separated from the solid gel sample. The Teflon vessel was sealed and heated at 200 • C for 48 h. The obtained white solid material was washed with deionized water and recovered by centrifugation, dried at 60 • C for 24 h, and subsequently calcined in air at 600 • C for 5 h to obtain Al 2 O 3 nanorods.

Characterization
The crystallinity of the samples was characterized via X-ray diffraction (XRD) experiments performed on a Bruker D8 Advance X-ray diffractometer (Cu Kα) (Bruker Corporation Inc., Billerica, MA, USA). The diffraction patterns were acquired at a scanning step of 0.01 • and a scanning speed of 10 • /min. Transmission electron microscopy (TEM) and scanning electron microscopy (SEM) images of the samples were recorded on a Talos F200X transmission electron microscope (Thermo Fisher Scientific Inc., Waltham, MA, USA) operated at 300 kV and a scanning electron microscope (GeminiSEM 450, Carl ZEISS, Oberkochen, Germany), respectively.
In situ DRIFTS measurements were performed on a Nicolet iS50 FTIR instrument (Thermo Fisher Scientific Inc., Waltham, MA, USA). The scans were recorded from 4000 to 600 cm −1 . Here, 50 mg catalyst powder was placed in a high-pressure (0-10 MPa) DRIFT cell (HVC-DRP-5, Harrick Scientific Products Inc., NY, USA) equipped with ZnSe windows and reduced at 300 • C in a pure H 2 stream (30 mL/min) for 2 h, after which the sample was flushed with N 2 (40 mL/min) for 1 h and cooled to room temperature. The background subtractions were executed over different samples for testing in 40 mL/min N 2 under 2 MPa. Then, the reactant gas mixture (H 2 + CO 2 , 3:1 ratio) at a flow rate of 40 mL/min was introduced into the sample cell and the spectra were collected at different temperatures, increasing by 3 • C/min at an interval of 19 s. For the CO 2 activation study, the sample cell was cooled to 250 • C. The background subtractions were similarly performed under 2 MPa. The CO 2 was switched into the reaction chamber and the spectra were recorded for up to 120 min, after which the inlet gas was changed to H 2 and the process lasted for another 1 h 40 min.

Catalytic Performance Evaluation
The CO 2 hydrogenation to methanol was performed in a high-pressure fixed-bed flow reactor. Here, 0.1 g catalyst was fixed in a quartz tube using quartz wool and then packed into a stainless steel tubular reactor. Prior to the catalytic measurements, the catalyst was reduced in a gas stream (10% H 2 /N 2 ) at 300 • C for 2 h with a flow rate of 10 mL/min under atmospheric pressure. Then, the temperature was decreased to 200 • C and the reducing gas was replaced by reaction gas (CO 2 + H 2 , 1:3 ratio). The reaction was X(CO 2 ) gas hourly space velocity (GHSV) of 7800 mL g −1 h −1 . The products flowing out from the reactor passed through a tube connected to a temperature control box to maintain the temperature at 120 • C and were then analyzed using an online gas chromatographer (Shimadzu GC-2014C) equipped with a TCD and an FID. The data obtained from the GC measurements were used to calculate the CO 2 conversion (X(CO 2 )), CH 3 OH selectivity (S i ), and methanol space-time yield (STY CH3OH ) using Equations (1)-(3).

Catalyst Structure, Morphology and Textural Properties
The XRD patterns of the n-Al 2 O 3 support with various levels of CuZnO loadings are presented in Figure 1   shown). The observed morphology is similar to that of the alumina nanorods reported in the literature [32,33]. Figure 2b shows the morphology of the n-Al 2 O 3 impregnated with 3% CuZnO, and the corresponding EDX result presented in Table 1 shows a Cu/ZnO ratio of 2.25, which is reasonably close to the theoretical value of 2. The TEM micrographs of the n-Al 2 O 3 and 3% CuZnO/n-Al 2 O 3 catalysts are shown in Figure 3. After impregnating the n-Al 2 O 3 support with 3% CuZnO, the morphology and structure of n-Al 2 O 3 remained unchanged. Both materials consist of nanorods with pores of about 3-10 nm in diameter. Due to their low loadings and low contrast, it is difficult to visualize the CuO and ZnO particles on the surface or inside the pores of the alumina nanorods with traditional widefield TEM [34]. In Figure 4a, the linear isotherms of Cu/n-Al 2 O 3 , CuZnO/n-Al 2 O 3 , ZnO/n-Al 2 O 3 , and CuZnO/c-Al 2 O 3 with corresponding BET surface areas of 94.4, 89.4, 67.6, and 99.3 m 2 /g, respectively, are shown. Except for CuZnO/c-Al 2 O 3 , other samples exhibit type IV isotherms with H1-type hysteresis loops, suggesting a typical mesoporous structure related to the n-Al 2 O 3 support [35]. CuZnO/c-Al 2 O 3 exhibits a type II isotherm with H4-type hysteresis loops, indicating an otherwise smaller mesoporous structure. According to the PSD curves (Figure 4b), CuZnO/c-Al 2 O 3 possesses pores exhibiting a broad range but with an average pore width of 68 Å. Despite the n-Al 2 O 3 -supported samples exhibiting a wider pore size distribution, more of the pores are concentrated in the low-diameter region up to 150 Å. Cu/n-Al 2 O 3 and CuZnO/n-Al 2 O 3 have a nearly similar total pore volume of approximately 0.5 cm 3 /g, which is greater than those of ZnO/n-Al 2 O 3 and CuZnO/c-Al 2 O 3 , 0.35 and 0.21 cm 3 /g, respectively.

CO 2 Hydrogenation to Methanol
The catalytic performances, including the CO 2 conversion efficiency, selectivity towards methanol, and space-time yield for methanol formation, are presented in Figure 5. The influence of the temperature was examined in the range of 200 • C to 300 • C for the catalysts with CuZnO loadings from 3 to 20%, and the results are presented in Figure 5a. As expected, the CO 2 conversion increased with the temperature for all of the examined catalysts because the thermodynamically stable CO 2 molecule can be more efficiently broken down at high temperatures. However, at high reaction temperatures, the selectivity towards CH 3 OH formation presented in Figure 5b decreased, given the competitive advantage of the RWGS reaction, resulting in CO formation [36]. The increasing CuZnO loading increased the CH 3 OH selectivity from 58.5% for 3% CuZnO to 79.5% for 20% CuZnO at 200 • C, yet the improved selectivity did not essentially originate from increased CO 2 conversion efficiency. The catalysts with 6-20% CuZnO had relatively similar CO 2 conversion levels, higher than that of the catalyst containing 3% CuZnO by 12.6% at 300 • C. In addition, the methanol selectivity and STY of the former were higher than that of the latter at all temperatures (Figure 5b,c).
Since our aim was not to pursue a high efficiency for CO 2 conversion to methanol but rather to explore and understand the influence of the catalyst structure on its catalytic performance, as well as the fact that the catalyst with 3% CuZnO showed a more homogeneous distribution of CuZnO on the catalyst support than the catalyst with higher loadings, we focused on the catalytic performance of the 3% CuZnO catalyst. As shown in Figure 5d-f, different samples containing 3% Cu, 3% ZnO, or 3% CuZnO were compared to better ascertain the role of each component and the synergistic effect. In addition, the catalytic performance of 3% CuZnO/n-Al 2 O 3 was compared with that of 3% CuZnO/c-Al 2 O 3 (commercial Al 2 O 3 ). As shown in Figure 5a-f, improved catalytic performance was observed, with better CO 2 conversion (19.8%) and STY CH3OH (1.31 mmol·g cat −1 h −1 ) at 300 • C while utilizing the 3% CuZnO/n-Al 2 O 3 sample. This combination outperformed those of the separate components, as well as the sample supported on the commercial Al 2 O 3. However, the methanol selectivity on 3% CuZnO/n-Al 2 O 3 was lower than those of the Cu/n-Al 2 O 3 and CuZnO/c-Al 2 O 3 at temperatures above 220 • C (Figure 5e). The 3% ZnO/n-Al 2 O 3 exhibited the lowest catalytic activity under all studied temperatures. This observation was not unexpected because Cu is the main component that contributes to methanol selectivity. The pure n-Al 2 O 3 support was found to have negligible activity for CH 3 OH production (data not shown). Overall, CuZnO exhibited better reactivity than either Cu or ZnO. While Cu is the methanol-selective component, ZnO and Al 2 O 3 play other roles in the catalytic process [12].

DRIFTS Study on the CO 2 Activation
The CO 2 activation was investigated via in situ DRIFT spectroscopy [37,38]. First, CO 2 was adsorbed on the various fresh catalysts (at 250 • C, 2 MPa) that were reduced in H 2 at 300 • C and the spectra were recorded. After sufficient adsorption of CO 2 and attaining a stable state for up to 120 min, the inlet gas was changed to H 2 . The spectra were again recorded for up to 220 min under the same conditions. The different catalyst components were studied during the experiments, and their specific roles in CO 2 activation and hydrogenation were elucidated.  Figure 6a). Specifically, the bands at 1661, 1585, and 1447 cm −1 are those of bicarbonates (HCO 3 − ) species [39]. Other bands at 1548 and 1327 cm −1 can be assigned to polydentate carbonate, while the band at 1393 cm −1 belongs to monodentate carbonate. The 1228 cm −1 band can be attributed to the δ(OH) of bicarbonate species. This band decreases or even disappears with time. On the C-H stretching region in Figure 6b, bands can be observed at 3002, 2929, 2899-2896, 2875, 2769, and 2741 cm −1 . According to the literature, the bands at 3002 and 2769 cm −1 were assigned to bidentate formate species on Al 2 O 3 -supported catalyst [40][41][42]; however, these bands were formed on pure alumina nanorods in this study. Here, we question the formation of formate species, since no H 2 gas was passed through the sample during the first 120 min, except that certain reactions exist with OH groups on the Al 2 O 3 surface [37]. Upon passing H 2 gas through the sample, the band frequencies in the spectra did not change significantly, indicating that no substantial reaction occurred with the absorbed carbon species and pointing to the fact that Al 2 O 3 cannot or can poorly activate H 2 . In addition, the presence of H 2 may promote CO 2 desorption from the Al 2 O 3 surface, since there was no strong reaction between them (CO 2 and Al 2 O 3 ), as indicated by the disappearance of the band at 1228 cm −1 . This result indicates that nanorod Al 2 O 3 can be used as a support for the CO 2 conversion catalyst due to its ability to adsorb CO 2 in its activated forms.

(ii) CO 2 activation on Cu/Al 2 O 3 and ZnO/n-Al 2 O 3
The formation of carbonate species on the surface of Cu/Al 2 O 3 was obvious at 120 min, as evidenced by the bands at 1644, 1442, and 1228 cm −1 [39,42] (Figure 6c). Interestingly, these carbonate species disappeared and formate species formed after switching to H 2 . This indicates that Cu can activate H 2 and promote its reaction with adsorbed CO 2 species [43]. The formate species showed bands at 1662, 1585, 1395, and 1329 cm −1 . These bands were attributed to bidentate formate species on both Cu and Al 2 O 3 and could be confirmed by the new bands at 2999, 2952, and 2742 cm −1 in the C-H stretching spectra (Figure 6d) [41,44]. Wang et al. [37] assigned the band at around 2999 cm −1 to the C-H bonds in unsaturated CH x on CuNi catalysts, and its presence facilitated the deep hydrogenation of CO 2 to form methanol. For ZnO/n-Al 2 O 3 ( (Figure 6e,f), after switching the feed gas from CO 2 to H 2 , there was no strong indication of H 2 activation; however, the appearance of bands of higher intensities and at nearly similar frequencies to that of pure n-Al 2 O 3 is an indication that ZnO promoted the CO 2 adsorption on the catalyst surface. The prominence of bands at all frequencies, which originated when only CO 2 was adsorbed and maintained in the presence of H 2 , prove that ZnO activates CO 2 mainly as carbonates. However, some bands exhibited slight shifts in their positions, which indicates some sort of interaction and formation of formate species. In a previous study, the bands at 1385 and 1328 cm −1 were assigned to bidentate formate on ZnO and Al 2 O 3 , whereas that at 1613 cm −1 was attributed to V as (OCO) of a bidentate carbonate species on ZnO [45]. The formation of formate species on ZnO can be confirmed by the evolution of 2946 and 2841 cm −1 bands in the C-H stretch spectra.
In the case of CuZnO/n-Al 2 O 3 , in the CO 2 gas stream, similar spectral features to Cu/n-Al 2 O 3 were observed. The peaks at 1643 and 1434 cm −1 were those of V s (OCO) of bicarbonate species [42]. The band at 1228 cm −1 was assigned to carboxylate (CO 2 δ− ) species in previous studies [38,46], but we related the band to the δ(OH) of bicarbonate species, as earlier mentioned. However, the peak intensities were found to be higher than those obtained from Cu/n-Al 2 O 3 . This was attributed to the contribution of ZnO, which promotes CO 2 adsorption. After switching the gas to H 2 , most of the formed carbonate species converted into the formate species by reacting with H-atoms. This was evident from analyzing the bands at 1616, 1585, 1388, and 1326 cm −1 , which indicated formate species on Al 2 O 3 , ZnO, and Cu [41,45]. The increased band intensities at 2897, 2921, and 2995 cm −1 indicated that the evolution of bidentate formate on the catalyst increased with time. It follows that CuZnO/n-Al 2 O 3 exhibited the best CO 2 activation, good H 2 activation, and catalytic activity for converting carbonate into formate, which is an important intermediate for methanol production. The increase in peak intensity means that the CO 2 adsorption intensified with time and reached a maximum (steady-state) at 120 min. Figure 6. In situ DRIFT spectra at different times for the fresh catalysts (reduced in H 2 at 300 • C, purged in N 2 , and then switched to CO 2 gas): (a,b) n-Al 2 O 3 , (c,d) 3% Cu/n-Al 2 O 3 , (e,f) 3% ZnO/n-Al 2 O 3 , and (g,h) 3% CuZnO/n-Al 2 O 3 at 250 • C under 2 MPa for both pure CO 2 gas (0-120 min) and after switching the feed gas to H 2 (120-220 min).

DRIFTS Study of Methanol Formation from CO 2 Hydrogenation
The key surface species and intermediates involved in forming methanol over the catalysts are shown in Figure 7. The in situ DRIFT spectra collected at different temperatures on n-Al 2 O 3 , ZnO/Al 2 O 3 , Cu/n-Al 2 O 3 , and CuZnO/n-Al 2 O 3 dosed with the reaction gas mixture indicated that the hydrogenation of CO 2 was initiated at a temperature above 50 • C. The initiation temperature decreases when the catalyst contains an active phase (Figure 7d,f,h). Except for the band around 1228 cm −1 , which is associated with δ(OH) of bicarbonate species or weakly adsorbed carbonate species (CO 3 2− ) that are desorbed at high temperatures, there is an increase in the intensities of the bands of hydrogenated species as the temperature rises. This means that the formation of key surface species increases with the temperature. The peaks at 1644 and 1416 cm −1 are attributed to bicarbonate species formed when the reaction gas mixture passed through n-Al 2 O 3 at 50 • C (Figure 7a). At higher temperatures, formate species formation is evident, typically indicated by the bands at 1612, 1583, 1391, and 1329 cm −1 . The formate species are also confirmed with the C-H stretching bands at 2997 and 2902 cm −1 in Figure 7b, which are identical to those observed when formic acid is adsorbed and evacuated at 473 K on Al 2 O 3 [41]. In Figure 7e,f, the bands at 1661 , 1608, 1582, 1389, and 1325 cm −1 are attributed to formate species on Cu and Al 2 O 3 [40]. Specifically, the band at 1325 cm −1 is for V s (OCO) of bidentate formate species on Cu [40,41]. Other bands also evolved at 2931 and 2740 cm −1 , related to v(C-H)+v as (OCO) and v(C-H)+v s (OCO) of bidentate formate on Al 2 O 3 [40]. On CuZnO/n-Al 2 O 3 (Figure 7g,h), the reaction gas is adsorbed as carbonate species up to 100 • C, as indicated by the adsorption bands around 1526 and 1415 cm −1 for polydentate carbonate and bicarbonate species, respectively [40]. The formation of formate species can be related to the bands at 1666 and 1628 cm −1 . The bidentate formate species on ZnO and Al 2 O 3 showed bands of higher intensities at 1575, 1388, 1314, 2888, and 2740 cm −1 [41,45]. Meanwhile, the 2921 cm −1 band can be assigned to the v(C-H) and δ(C-H)+v as (OCO) of bidentate formate on Cu [41,44]. The peaks at 2948 and 2843 in the CuZnO/n-Al 2 O 3 spectra and at 2931 and 2846 in the Cu/n-Al 2 O 3 spectra indicate the formation of methoxy species [30]. At 150 • C or higher temperatures, the band intensities increased compared with those of Cu/n-Al 2 O 3 . For these catalysts, the carbonate conversion to formate species is seen on both CuZnO and Cu, but the formation of formate species begins at a low temperature (100 • C) exclusively on Cu/n-Al 2 O 3 . The key intermediate species formed at different temperatures are summarized in Table 2. The introduction of ZnO enhances the CO 2 adsorption and facilitates more carbonate species formation. This, in turn, boosts the formate production through hydrogenation via dissociated H atoms, which spillover to the catalyst surface. The formation of formate at a lower temperature on the CuZnO/n-Al 2 O 3 catalyst might be ascribed to a more effective H 2 dissociation in the presence of ZnO [40]. Generally, two reaction pathways for the hydrogenation of the CO 2 to methanol have been identified: RWGS + CO-hydro, along with formate pathways [47]. However, many studies on Cu-based catalysts conform to the formate pathway [30,38]. In this study, CO 2 is activated as carbonate species on the catalyst surface under the examined reaction conditions. As the temperature increases, H 2 is better activated and spills over to the catalyst surface, where it reacts with the carbonate and forms formate and methoxy species. Further hydrogenation of the surface species produces methanol. This mechanism is, thus, in good agreement with literature reports for other Cu-based systems.

Roles of Al 2 O 3 , ZnO, and Cu in CO 2 Hydrogenation to Methanol
Understanding the structure-function relationship of a composite catalyst is important to develop more efficient catalysts for industrial applications. The individual roles of Al 2 O 3 , ZnO, and Cu and their synergistic effects in the CuZnO/Al 2 O 3 catalyst used for the CO 2 -to-methanol synthesis were investigated using DRIFT spectroscopy. Al 2 O 3 is a large surface area support material with an appropriate surface for dispersing and stabilizing the active phase and slowing the sintering during the reaction at high temperatures [40]. Herein, we found Al 2 O 3 to be a good material for CO 2 adsorption and activation to produce bicarbonate, carbonate, and formate species; the bicarbonate and carbonate species are converted into formate and methoxy in the presence of metal species that can activate H 2 . The 3% Cu/n-Al 2 O 3 and 3%CuZnO/n-Al 2 O 3 catalysts possess nearly similar BET surface areas of 94.4 and 89.4 m 2 /g, respectively. This shows that both active phases had the same impact on the surface area; however, the catalysts exhibit slight differences in their CO 2 adsorption and catalytic activities. Methoxy and increasing amounts of formate species are produced for both catalysts. The conversion of carbonates to formate is evident, as H-atoms are available for the reaction after activation on Cu particles following the spillover effect. According to the catalytic performance and in situ DRIFT results, CuZnO/n-Al 2 O 3 exhibited the best performance due to the synergistic effect resulting from the interaction of Cu and ZnO. Apparently, as ascertained from the in situ DRIFT spectra, the presence of ZnO with Cu promotes CO 2 adsorption and formate formation. The enhanced formate formation is thought to be due to additional H 2 dissociation activity. We, thus, suggest that ZnO also promotes H 2 activation. It was previously reported that a defective ZnO x overlayer formed in a Cu/ZnO catalyst could dissociate H 2 , even at room temperature [40]. From the above discussion, it will be reasonable to conclude from our study that the different components of the industrial catalyst for methanol synthesis from CO 2 play different roles but with a cooperative effect in methanol formation, as shown in Scheme 1.

Conclusions
Cu-based catalysts supported on Al 2 O 3 nanorods for CO 2 activation and subsequent hydrogenation to methanol are discussed in this study. Compared with the commercial Al 2 O 3 support, catalysts using Al 2 O 3 nanorods as the support show higher efficiency for CO 2 conversion between 200 and 300 • C, but the methanol selectivity becomes lower above 220 • C. There is improved CO 2 adsorption on the Al 2 O 3 nanorods. CO 2 is adsorbed and activated on the catalyst surface as carbonate species, which upon subsequent H 2 dissociation on Cu hydrogenates to formate species by reacting with the H atom. ZnO promotes CO 2 adsorption and probably facilitates H 2 dissociation on Cu, also leading to the more efficient hydrogenation of activated CO 2 species on CuZnO/Al 2 O 3 . This study will help to tailor the properties and composition of the Cu-based catalyst system to achieve higher efficiency for methanol synthesis from CO 2 .