Boosting Low-Temperature CO2 Hydrogenation over Ni-based Catalysts by Tuning Strong Metal-Support Interactions.

CO2Reduction Boosting Low-Temperature CO2Hydrogenation over Ni-based Catalysts by Tuning Strong Metal-Support Interactions Runping Ye+, Lixuan Ma+, Xiaoling Hong+, Tomas Ramirez Reina, Wenhao Luo, Liqun Kang, Gang Feng, Rongbin Zhang, Maohong Fan,* Riguang Zhang,* and Jian Liu* Abstract: Rational design of low-cost and efficient transition-metal catalysts for low-temperature CO2activation is significant and poses great challenges. Herein, a strategy via regulating the local electron density of active sites is developed to boost CO2methanation that normally requires >350°C for commercial Ni catalysts. An optimal Ni/ZrO2 catalyst affords an excellent low-temperature performance hitherto, with a CO2conversion of 84.0%, CH4selectivity of 98.6% even at 230°C and GHSV of 12,000 mLg1h1for 106 h, reflecting one of the best CO2methanation performance to date on Ni-based catalysts. Combined a series of in situ spectroscopic characterization studies reveal that reconstructing monoclinic-ZrO2supported Ni species with abundant oxygen vacancies can facilitate CO2activation, owing to the enhanced local electron density of Ni induced by the strong metal-support interactions. These findings might be of great aid for construction of robust catalysts with an enhanced performance for CO2emission abatement and beyond. Introduction Carbon dioxide (CO2) utilization technology is under intensive development, given its key role in achieving carbon neutrality.[1] Under this global scenario and extensive applications of hydrogenation reactions,[2] CO2hydrogena- tion with green hydrogen from renewable energy is an effective approach to transforming CO2into chemicals and fuels.[3] Particularly, affording nearly quantitative selectivity towards methane and over 90% equilibrium CO2conversion levels even under low temperatures (<300°C) and atmospheric pressure, CO2methanation brings a major superiority compared to other CO2hydrogenation processes.[4] In other words, CO2methanation is deemed as an efficient strategy to convert a large amount of captured CO2into value-added synthetic natural gas, which could be further converted into other valuable products. Furthermore, CO2methanation could be applied to coal-to-gas, power-to-gas and CO2 removal for spacecraft showcasing the versatility and broad range of applications of this catalytic process.[5] As CO2methanation is an exothermic reaction, it is thermodynamically favored in the low temperatures range. However, the reaction kinetics counterbalance the exothermic nature imposing some energy barriers for reactants activation.[5a,6] Most of the previous studies report that the actual reaction temperatures for CO2methanation are [*] Prof. Dr. R. Ye,+Prof. Dr. G. Feng, Prof. Dr. R. Zhang Key Laboratory of Jiangxi Province for Environment and Energy Catalysis, Institute of Applied Chemistry, School of Chemistry and Chemical Engineering, Nanchang University Nanchang 330031 (P. R. China) X. Hong,+Prof. Dr. J. Liu State Key Laboratory of Catalysis, Dalian Institute of Chemical Physics, Chinese Academy of Sciences Dalian 116023, Liaoning (P. R. China) E-mail: [email protected] L. Ma,+Prof. Dr. R. Zhang State Key Laboratory of Clean and Efficient Coal Utilization, College of Chemical Engineering and Technology, Taiyuan University of Technology Taiyuan 030024, Shanxi (P. R. China) E-mail: [email protected] Prof. Dr. J. Liu DICP-Surrey Joint Centre for Future Materials, Department of Chemical and Process Engineering, and Advanced Technology Institute, University of Surrey Guilford, Surrey GU2 7XH (UK) Prof. Dr. M. Fan College of Engineering and Physical Sciences, and School of Energy Resources, University of Wyoming Laramie, WY 82071 (USA) E-mail: [email protected] Prof. Dr. W. Luo, Prof. Dr. J. Liu College of Chemistry and Chemical Engineering, Inner Mongolia University Hohhot 010021 (P. R. China) E-mail: [email protected] Prof. Dr. T. R. Reina Department of Inorganic Chemistry and Material Sciences Institute of Seville, University of Seville-CSIC 41092 Seville (Spain) Dr. L. Kang Max Planck Institute for Chemical Energy Conversion Stiftstraße 34–36, 45470 Mülheim an der Ruhr (Germany) [+] These authors contributed equally to this work. © 2023 The Authors. Angewandte Chemie International Edition published by Wiley-VCH GmbH. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. Angewandte Chemie Research Articles www.angewandte.org How to cite: Angew. Chem. Int. Ed. 2024,63, e202317669 doi.org/10.1002/anie.202317669 Angew. Chem. Int. Ed. 2024,63, e202317669 (1 of 11) © 2023 The Authors. Angewandte Chemie International Edition published by Wiley-VCH GmbH higher than 350°C.[7] Indeed, CO2is an inert gas with its σ bond and π bond activation demand relatively high temperatures. However, when a commercial application is envisaged, high temperatures may represent excessive energy consumption and impact the overall process operational costs. Moreover, upon raising the temperature the competitive reaction (reverse water gas shift, RWGS) begins to operate leading to higher selectivity towards CO as a byproduct and also the catalyst deactivation due to active phase sintering.[8] Hence the bottleneck issue for CO2 methanation is to develop a robust non-noble catalyst to activate CO2molecules and further hydrogenate them yielding methane at low reaction temperatures. Generally, noble metal-based catalysts present efficient low-temperature performance of CO2methanation, but they are not practical options for large-scale applications due to their elevated costs.[9] Alternatively, Ni-based catalysts present high performance for CO2methanation at a generally high reaction temperature (>350°C).[10] To improve their activity, many strategies have been developed from different perspectives, such as enhancing Ni loading (ωNi),[11] changing the Ni particle size (dNi),[12] doping of the second metal,[13] improving the strong metal-support interactions (SMSI),[14] and optimizing the catalyst structure.[15] Among them, tuning the properties of supports, such as their crystal phases, specific surface areas, surface acidity or basicity, pore structures, support’s particle sizes and oxygen vacancies could obviously affect CO2methanation performance.[16] We have previously demonstrated that Ni/ monoclinic-ZrO2catalyst could afford a superior activity for CO2hydrogenation compared to Ni/cubic-ZrO2catalyst.[16a] Similarly, a series of Co/CeO2-ZrO2catalysts with different particle sizes of CeO2-ZrO2support could lead to various CO2methanation performance owing to the size-dependent metal-support interactions and oxygen vacancies.[16b] The particle size of CeO2-ZrO2with �20 nm supported Co nanoparticles exhibited better performance due to higher Co dispersion and facile formation of oxygen vacancies. Moreover, the properties of supports could further tune the SMSI and then influence catalytic performance.[17] For example, the effect of SMSI on the methane selectivity and catalytic activity during CO2methanation have been demonstrated over Ni/Ru/Co-based catalysts.[9b,16b,18] However, the optimal reaction temperatures over these reported catalysts for CO2methanation were still higher than 300°C. In this contribution, three Ni-based catalysts (Ni/SiAlZr) with different supports (SiO2, Al2O3, ZrO2) and strong metal-support interactions were prepared by sol-gel method for CO2methanation. Also, a reference Ni/ZrO2-IM sample was prepared by impregnation method (IM). As a result, the Ni/ZrO2catalyst exhibited an extraordinary low-temper- ature property with a CO2conversion of 84.0% and methane selectivity of 98.6% even at 230°C. The effect of local electron density on Ni active sites and the exceptional performance of Ni/ZrO2has been investigated by a series of experimental characterizations. Results and Discussion Three Ni-based catalysts with SiO2, Al2O3and ZrO2as supports have been synthesized and evaluated in fix-bed reactor. CO2hydrogenation over a series of Ni/SiAlZr catalysts was firstly performed under different reaction temperatures. As shown in Figure 1a, the Ni/ZrO2presents the highest CO2conversion (Conv.CO2) of 91.9% at 240°C, while the Ni/SiO2, Ni/ZrO2-IM and commercial Ni samples exhibit much lower Conv.CO2 under identical reaction conditions. The CH4selectivity (Sel.CH4) over Ni/SiAlZr catalysts is almost 100%, except that Ni/SiO2presents an obvious CO product under low reaction temperatures. It has been reported that the CO selectivity over Ni/SiO2catalysts would be affected by many factors, such as reaction conditions, preparation method, Ni loading, and silica structure (Table S1).[19] Furthermore, the Conv.CO2 over Ni/ ZrO2could reach up to 50% at only 210°C while that for Ni/Al2O3, Ni/SiO2and commercial Ni catalysts is 234, 273 and 268°C, respectively (Figure 1b and Figure S1). Therefore, the activity order over the Ni/SiAlZr catalysts is Ni/ ZrO2>Ni/Al2O3@Ni/SiO2—Commercial Ni catalyst>Ni/ ZrO2-IM. The effect of Ni loading on Ni/ZrO2has been further investigated. It indicates that the Conv.CO2 increases significantly, while the Sel.CH4 is almost the same in spite of the increase of Ni loading (Figure S2). Thus, the Ni-based catalysts with high Ni loading show better CO2activation activity and the 40% Ni/ZrO2catalyst exhibits the best catalytic performance here. As the Ni/ZrO2in this work displays exceptionally high activity for CO2methanation under low reaction temperatures, the comparison of Ni/ZrO2and some reported representative Ni-based catalysts are illustrated in Figure 1c and Table S2. The Ni/ZrO2reported here exhibits high Conv.CO2 and space-time yield (STY, 311.2 mmolCH4 gcat1h1) of methane, which is much higher (3–115 times) than those over the previously reported Ni/ZrO2catalysts (2.7– 89.9 mmolCH4 gcat1h1) under similar reaction conditions. For the other typical Ni-based catalysts listed in Table S2, the Ni/ZrO2presented herein still shows a much lower reaction temperature and higher STYCH4. Therefore, to the best of our knowledge, our engineered catalyst takes the edge over the state-of-the-art materials in this field. Beyond the advanced catalytic activity showcased by the Ni/ZrO2, its stability tests were performed at 230°C and gas hour space velocity (GHSV) of 12,000 mLg1h1(Figure 1d). The Conv.CO2 and Sel.CH4 were able to maintain above 80% and 98% even after a time on steam of 106 h, respectively. Furthermore, the Ni/ZrO2was stable under more severe reaction conditions (220°C, 24,000 mLg1h1). However, the Conv.CO2 could still maintain at around 45% without obvious deactivation (Figure 1d). In addition, we repeated the stability test of Ni/ZrO2catalyst under low temperatures of 200, 230 300°C and high GHSV of 48,000 mLg1h1, and it still presented prominent stability (Figure S3). Based on various characterization results of the fresh and spent Ni/ZrO2(Figure S4 and Figure S5), no apparent issues of metal sintering and coke formation can Angewandte Chemie Research Articles Angew. Chem. Int. Ed. 2024,63, e202317669 (2 of 11) © 2023 The Authors. Angewandte Chemie International Edition published by Wiley-VCH GmbH 15213773, 2024, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/anie.202317669 by Universidad De Sevilla, Wiley Online Library on [28/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License be indicated upon catalysis, confirming the superior stability of Ni/ZrO2catalyst. We further test their structural and textual properties based on the remarkable activity of Ni/ZrO2catalysts. The actual Ni loading over Ni/SiO2was only 32.91 wt%, which was much lower than that (�40 wt%) over other Ni/SiAlZr catalysts (Table 1). The reason was probably that there was not enough hydroxyl on the surface of silicon oxide to anchor the high concentration of nickel species, which resulted in some nickel species being evaporated during the process of evaporating the water to form gel. Would the lower Ni loading on Ni/SiO2lead to much lower catalytic performance than Ni/ZrO2? The 30% Ni/ZrO2exhibited a much higher Conv.CO2 (64.9%) than that over Ni/SiO2 (11.5%) at 240°C. In our previous work, a Ni/SiO2catalyst with 44.4 wt% content of Ni still only presented a low Conv.CO2 of �30% at 250°C.[20] Liu et al. also deeply investigated why Ni/CeO2with 10.0 wt% of Ni was more active than Ni/SiO2with 25.9 wt% of Ni for CO2 methanation.[21] The reason was attributed to the tiny Ni particle size for large H2uptake capacity and more oxygen vacancies for CO2activation on the Ni/CeO2. Therefore, we can exclude the influence of Ni content on the catalytic results and the support species would significantly affect the catalytic performance. Figure S6 and Table 1 display the N2physisorption results of Ni/SiAlZr catalysts. The shape of the isotherm on Ni/SiO2is much different from the others, but its specific Figure 1. CO2hydrogenation performance over Ni/SiAlZr catalysts. (a) Catalytic performance at 240°C and GHSV of 12,000 mLg1h1. (b) Conv.CO2 and Sel.CH4 over Ni/ZrO2and commercial Ni catalysts under different reaction temperatures. (c) The Ni/ZrO2compared to other reported Ni/ZrO2 catalysts at 240–250°C. More detailed information on the reported Ni/ZrO2catalysts is listed in Table S2. (d) The long-term stability of Ni/ZrO2 catalyst at 220–230°C and GHSV of 12,000–24,000 mLg1h1. Table 1: The physicochemical properties of Ni/SiAlZr catalysts. Catalysts ωNi [wt.%] SBET [m2/g] Dp [nm] VP [cm3/g] DNi [%] SANi [m2/g] dNi [nm] TOF [×103s1] Ni/SiO232.91 156.3 11.9 0.45 6.45 14.13 14.7 1.6 Ni/Al2O341.18 216.0 3.0 0.14 6.08 16.68 11.0 8.7 Ni/ZrO242.83 84.3 3.2 0.06 3.15 8.98 14.6 39.6 Ni/ZrO2-IM 40.03 20.8 7.5 0.04 1.17 3.12 39.3 1.7 Angewandte Chemie Research Articles Angew. Chem. Int. Ed. 2024,63, e202317669 (3 of 11) © 2023 The Authors. Angewandte Chemie International Edition published by Wiley-VCH GmbH 15213773, 2024, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/anie.202317669 by Universidad De Sevilla, Wiley Online Library on [28/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License surface area (SBET) and average pore diameter (Dp) as well as pore volume (Vp) were much larger than those over Ni/ ZrO2. The Ni/Al2O3exhibits the largest SBET of 216.0 m2/g, while the Ni/ZrO2-IM displays only 20.8 m2/g. It suggests that some nickel species were covered in the pores of Ni/ ZrO2-IM, and the SBET would not affect the catalytic performance. H2pulse chemisorption was carried out to determine the Ni dispersion (DNi) and Ni surfaces area (SANi). As a result, the Ni/ZrO2shows a DNi of 3.15%, which is lower than Ni/ SiO2and Ni/Al2O3catalysts while higher than Ni/ZrO2-IM catalyst (Table 1). The particle size of Ni nanoparticles on Ni/ZrO2(14.6 nm) is similar to that on the Ni/SiO2and Ni/ Al2O3, but much smaller than that on the Ni/ZrO2-IM (39.3 nm). Based on the above data, turnover frequency (TOF) values were further employed to compare their catalytic performance at 200°C (Table 1). It shows that the TOF value over Ni/ZrO2is much higher than the others. The crystalline phases of calcined and reduced samples were measured by XRD. The small diffraction peaks of Ni metal could be observed over the samples prepared by the sol-gel method, while the Ni/ZrO2-IM presents sharp diffraction peaks of NiO or Ni, indicating that much larger nickel species were formed on the Ni/ZrO2-IM. A weak amorphous silica or aluminum oxide peak could also be observed over the Ni/SiO2and Ni/Al2O3samples (Figure 2a). Interestingly, the ZrO2support over the calcined Ni/ ZrO2sample displays diffraction peaks of cubic zirconia (c- ZrO2, PDF#03-0640, Figure S7), while the reduced and especially the used Ni/ZrO2samples present some obvious peaks of monoclinic zirconia (m-ZrO2, PDF#05-0543, Figure 2a). It suggests that the Ni/ZrO2could be reconstructed by changing the crystal form of ZrO2during hydrogen activation and CO2methanation. In our previous work,[16a] the Ni/m-ZrO2possessed high activity for this reaction than the Ni/c-ZrO2due to their diverse local electron density of Ni. To further analyze the nickel species and the reduction behavior of Ni/SiAlZr catalysts, H2-TPR profiles were obtained and shown in Figure 2b. The reduction peak of Ni/ SiO2is obviously smaller than that of the others, which is consistent with the ICP result that shows lower Ni loading. Furthermore, the large asymmetric reduction peaks from 100 to 600°C could be fitted by a Gaussian-type function into three small reduction peaks (Figure S8). The first one (α) centered at around 200°C is attributed to the reduction of NiO weakly interacting with oxide supports and active adsorbed oxygen species. The second reduction peak (β) centered at around 300°C can be assigned to the hydrogen consumption on the reduction of nickel species in the interfacial structure of Ni-O-Si (Al or Zr). The last one (γ) reduction peak centered at around 425°C mainly corresponds to highly dispersed NiO with strong metal-support interactions.[22] The H2-TPR suggests that the nickel species over Ni/ZrO2could be easier to be reduced than the other two samples, demonstrating that the metal interacts strongly with support on the Ni/SiO2and Ni/Al2O3catalysts. In addition, the CO2-TPD tests were operated to determine the surface basicity of Ni/SiAlZr catalysts. Figure 2c shows that the peaks belonging to the weak basic sites at 50–150°C are Figure 2. The physicochemical properties of Ni/SiAlZr catalysts. (a) XRD patterns of reduced Ni/SiAlZr catalysts. (b) H2-TPR profiles of Ni/SiAlZr catalysts. (c) CO2-TPD profiles of Ni/SiAlZr catalysts. (d-f) HAADF and HRTEM images with corresponding elemental mapping of Ni/ZrO2catalyst. Angewandte Chemie Research Articles Angew. Chem. Int. Ed. 2024,63, e202317669 (4 of 11) © 2023 The Authors. Angewandte Chemie International Edition published by Wiley-VCH GmbH 15213773, 2024, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/anie.202317669 by Universidad De Sevilla, Wiley Online Library on [28/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License very weak, while most of the desorption peaks are attributed to the medium-strength (150–500°C) and strong basic sites (500–750°C). The Ni/ZrO2possesses the strongest CO2 desorption peaks while the Ni/Al2O3is the opposite, indicating that the Ni/SiO2and Ni/Al2O3mainly exhibit medium-strength basic sites while those over the Ni/ZrO2 are the strong basic sites. From previous literature reports,[22a] we know that medium-strength basic sites are mainly associated with acid-base metal oxides and Ni-O2 pairs, while the strong basic sites can be caused by surface coordinatively unsaturated Lewis basic sites of O2species. The peaks around 500–750°C can be attributed to the CO2 adsorption on the strong basic sites and surface oxygen vacancies.[23] The Ni/ZrO2sample has more surface oxygen vacancies and thus presents a larger peak around 500– 750°C. More importantly, the CO2at the high-temperature desorption peak would be converted to the formation of HCO3or CO32, leading to the strong basicity.[24] As the CO2-TPD test did not flow hydrogen, the HCO3or CO32 could not be further converted to methane. During CO2 hydrogenation, the HCO3or CO32could be converted to other intermediates and the final methane product. In addition, it was reported that the CO2methanation performance was correlated with the strength of the basicity of the catalyst.[24] The Ni/ZrO2catalyst with stronger basic sites presented better catalytic properties. Therefore, the metal species are various among the Ni/SiAlZr catalysts, with the Ni/ZrO2exhibiting more reduced nickel species and strong basic sites while metal interacts the strongest with support on the Ni/Al2O3with medium-strength basic sites. Figure 2d–f and Figures S9–S13 present the electron microscope images of Ni/SiAlZr catalysts. The SEM images show that the Ni/ZrO2exhibits abundant large pores (Figure S9), probably produced through citric acid decomposition. These large pores are beneficial for gas flow and mass transfer. TEM images were further obtained to observe the Ni nanoparticles (NPs) and other elemental distribution. Figure 2e–f and Figures S10–S11 display that Ni NPs are highly dispersed over the SiO2, Al2O3, and ZrO2supports. The mean particle size over Ni/SiO2is slightly higher than that over Ni/Al2O3(Figure S10). For the Ni/Al2O3, the TEM and corresponding EDS mapping images show that the Ni nanoparticles are encapsulated by a thin Al2O3layer (Figure S11), which results in slightly larger Ni particle sizes determined by H2pulse chemisorption than the XRD results (11.0 nm vs. 16.7 nm). It was reported that the palladium or Ni nanoparticles were decorated with a thin layer of an aluminate phase or nickel aluminate layer, which indicates that the SMSI exhibits in the Al2O3-based catalysts.[25] Different catalyst structures could be inferred from TEM results with a further comparison between Ni/ZrO2 and Ni/ZrO2-IM by the HRTEM and HAADF-STEM images (Figure 2d–f and Figures S12–S13). For the HRTEM images of Ni/ZrO2, crystal planes belonging to the m-ZrO2 and c-ZrO2could be obviously observed while that for the Ni or NiO are difficult to observe (Figure 2d). The dvalues of the distances between adjacent lattice fringes in Ni/ZrO2 catalsyt are slightly shorter than those in the bulk m-ZrO2 (Figure 2d). The elemental mapping further shows that Ni metals are separated and embraced by the Zr and O elements on the Ni/ZrO2(Figure 2f and Figure S12). The situation for the Ni/ZrO2-IM is opposite, which presents many obvious crystal planes belonging to Ni or NiO (Figure S13). Therefore, the TEM results indicate that some Ni NPs are uniformly incorporated into the lattice of ZrO2and surrounded by the Zr and O elements on the Ni/ZrO2. The literatures have already reported that Ni nanoparticles could be incorporated into mesoporous amorphous ZrO2matrix and ZrO2-CeO2supports.[26] The extent of SMSI over Ni/ SiAlZr catalysts is different and the SMSI is obvious on the Ni/ZrO2and weak on the Ni/Al2O3while SMSI is lacking on the Ni/SiO2catalyst. The X-ray photoelectron spectroscopy (XPS) was also employed to investigate the chemical state and surface compositions of the Ni/SiAlZr catalysts. Firstly, the quasi in situ XPS spectroscopy on the Ni/ZrO2catalyst under different processes was also investigated and shown in Figure 3 and Table S3. For the Ni 2p spectra, it shows that most of the nickel species over the calcined Ni/ZrO2sample are Ni2+ species at �855.3 eV and corresponding satellite peaks at �860.6 eV. After reduction by hydrogen, the ratio of Ni0/ (Ni0+Ni2+) increased from 9.6% to 44.1% and then increased up to 55.2–58.3% during CO2methanation. It proves that most of the nickel species are Ni0species during CO2hydrogenation reaction. The binding energy (BE) of Ni02p3/2 on Ni/ZrO2is �852.5 eV, which is much lower than the others and reported bulk Ni0, demonstrating that the electronic density of Ni was increased by the SMSI.[27] Therefore, some of Ni atoms (Niδ) were negatively charged by electron transfer from ZrO2support, which is consistent with the chemical state and surface compositions of Ni 2p in the Ni-Bi catalyst.[28] The defective ZrO2support with abundant oxygen vacancies can trap negative electrons at its center and then transfer to Ni at the interface. In addition, the negatively charged metals from the supports due to metal-support interactions have also been observed over Ru/TiO2,[29] Cu/ZnO,[30] and Pd/Ga2O3catalysts for other hydrogenation reactions.[31] Furthermore, the O 1s spectra can be divided into three peaks. Namely, the former two peaks (Oαand Oβ) can be attributed to the surface adsorbed oxygen species (Oor O22, O2), and the last peak (Oγ) is belonged to lattice oxygen (O2).[32] The surface adsorbed oxygen species may come from the surface hydroxyl groups, transformation of the surface adsorbed O2molecules, defect oxide, and CO32.[33] The surface lattice oxygen ratio over the calcined Ni/ZrO2sample (54.9%) was slightly decreased to around 50.2% while the ratio of Zr3+/(Zr3++Zr4+) increased from 18.2% to around 22.8% after reduction and CO2methanation. For the Zr 3d spectra, the Zr4+species have two strong peaks at �181.5 eV and �184.0 eV while the Zr3+species from partial reduction of Zr4+have two weak peaks at �180.5 eV and �182.5 eV.[22b,34] The BE of Zr 3d5/2 on Ni/ZrO2catalyst (180.5–181.5 eV) was much lower than the value of Zr4+ions in ZrO2(183.5 eV), which was because that surface Zr centers are coordinatively unsaturated and have lower oxidation states due to the dangling bond, indicating that Zr4+on the surface has been partly reduced to Zr3+.[33b] The ratio of Zr3+/(Zr3++Zr4+) Angewandte Chemie Research Articles Angew. Chem. Int. Ed. 2024,63, e202317669 (5 of 11) © 2023 The Authors. Angewandte Chemie International Edition published by Wiley-VCH GmbH 15213773, 2024, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/anie.202317669 by Universidad De Sevilla, Wiley Online Library on [28/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License suggests the relative concentrations of the oxygen vacancies. It indicates that more oxygen vacancies are produced after reduction and the oxygen vacancies are stable because the O 1s and Zr 3d peaks are stable during reaction. Thus, oxygen vacancies on the Ni/ZrO2catalyst are involved in CO2 methanation reaction. The oxygen vacancies in the Ni/ZrO2 sample was also demonstrated by the pyridine-infrared spectroscopy (Py-IR). The peaks at 1445 cm1is attributed to the Lewis acid sites, which are from coordinatively unsaturated Zr3+and Zr4+as well as the oxygen vacancies.[35] The specific amount of Lewis acid sites can be quantified by the area of the peak at 1445 cm1with the calculation formula at the end of Table S4. The Ni/ZrO2 exhibits a slightly lower amount of Lewis acid sites than pure ZrO2support but a higher amount of Lewis acid sites than Ni/ZrO2-IM sample, suggesting that a moderate amount of oxygen vacancies are present over Ni/ZrO2 sample (Figure S14). For the control samples, the XPS results are illustrated in Figure S15 and Table S5. Firstly, the surface Ni atom ratio (24.3%) on the Ni/ZrO2-IM is obviously higher than the others (6.7–9.9%), indicating that more nickel species sintering on Ni/ZrO2-IM catalyst prepared by impregnation method while the other catalysts would have more Ni species encapsulated by the supports. Secondly, the Ni 2p and O 1s on the Ni/ZrO2-IM is much different from the others. Namely, the Ni/ZrO2-IM has more obvious peaks of Ni0species and lattice oxygen. The main nickel species from ex situ XPS test over the control samples are Ni2+, which is probably from the quick re-oxidation of active Ni0to Ni2+or unreduced Ni2+as proved by the above H2-TPR. It is obvious that Ni/ZrO2-IM exhibits more lattice oxygen (52.7%) from crystalline ZrO2than Ni/SiO2and Ni/Al2O3. The latter two catalysts exhibit 9.3–11.0% of lattice oxygen that probably originates from the crystalline NiO. Although the Ni/SiO2and Ni/Al2O3possess more surface adsorbed oxygen species for CO2methanation, their catalytic performances are still lower than Ni/ZrO2. In addition, X-ray absorption spectra (XAS) were performed to account for the difference in the Ni-O-Zr coordination structure between the main sample and the control samples, and the results are shown in Figure 4, Figures S16–S18 and Table S6. For the Ni K-edge X-ray absorption near-edge structure (XANES) and extended X- ray absorption fine structure (EXAFS) spectra, the trends in edge positions and the whiteline intensities are all in line with the actual compositions of the Ni catalysts obtained from the EXAFS analysis. The first shell Ni-O and Ni-Ni paths are characteristic for Ni oxides and Ni metal. The second shell Ni-Ni path is taken to complete the NiO structure. Based on the XRD characterization results, it is reasonable to believe the majority of Ni species are present in the form of Ni metal and Ni oxides, which is consistent with the XPS results. From the fitting results, we could see the bond lengths are in good agreement with the Ni-O and Ni-Ni distances in bulk Ni metal and NiO phase (Table S6). The only noticeable difference is that the two Ni-Ni distances are slightly longer than those in Ni metal and NiO, suggesting the expanded lattice distances due to the formation of mixed phases. The Debye–Waller factor is significantly bigger than those in typical Ni foil and NiO, which means the coordination structure of the Ni sites is strongly disordered, corresponding well with the broad FWHM of the XRD diffraction peaks. Furthermore, it is worth mentioning that in addition to the shorter Ni-Ni scattering path, the longer Ni-Ni scattering path could be Figure 3. Quasi in situ XPS spectra of (a) Ni 2p, (b) O 1s, and (c) Zr 3d over Ni/ZrO2catalysts after calcination, quasi in situ reduction by H2,quasi in situ reaction under the atmosphere of CO2and H2at room temperature (RT), 200°C and 240°C. Angewandte Chemie Research Articles Angew. Chem. Int. Ed. 2024,63, e202317669 (6 of 11) © 2023 The Authors. Angewandte Chemie International Edition published by Wiley-VCH GmbH 15213773, 2024, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/anie.202317669 by Universidad De Sevilla, Wiley Online Library on [28/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License replaced by a Ni-Zr scattering. With Zr as a slightly heavier back scatter, the mathematical EXAFS fitting could still get converged, showing similar coordination numbers and Debye–Waller factors. But the calculated Ni-Zr distance is around 2.85 Å, which is significantly shorter than typical Zr- Zr distances in Zr oxides. Therefore, most of the Ni species are still loaded on the surface of ZrO2. For the Zr K-edge XANES and EXAFS (Figure 4 and Figure S18), the Zr-O and Zr-Zr distances are in good agreement with the average Zr-O and Zr-Zr distances in the ZrO2structure. Actually, they were slightly shorter than the ideal bond lengths. Similarly, the huge Debye–Waller factors for the Zr-Zr paths suggest that the ZrO2structure is highly disordered. The coordination number (C.N.) for the Zr-O path is less than typical cubic ZrO2(C.N.=8) and monoclinic ZrO2 (C.N.=7). This indicates the ZrO2supports exhibit surface coordinatively unsaturated Zr centers with highly disordered structure. To explore the reaction mechanism, a series of experimental measurements were operated. Firstly, the temperature programmed surface reaction (TPSR) of CO2metha- nation was carried out to investigate CO2activation and possible reaction products under different reaction temperatures. The detailed test procedures are shown in the Supporting Information. As depicted in Figure S19, all the peak areas of both the thermal conductivity detector (TCD) signal and mass spectrometry (MS) signal on Ni/ZrO2-IM are lower than those appeared on Ni/ZrO2, indicating that the surface of Ni/ZrO2is better for CO2adsorption and activation than that over the Ni/ZrO2-IM. We could obtain the following important information from Figure 5a and Figure S19c. CO2could be detected only at a temperature lower than �150°C, indicating that the left CO2at the higher temperature was consumed by hydrogen. Also, when the temperature was higher than 100°C, the peak of CH4at m/z of 16 on Ni/ZrO2began to increase and reached two vertexes at 180 and 626°C, respectively. This result suggests that CO2could be activated to form methane at the extremely low reaction temperature, which is consistent with the activity test and the temperature is also much lower than that reported over the mesoporous Ni/CeO2-ZrO2 catalyst.[36] The large peak of CH4formation at around 626°C indicates that the surface of Ni/ZrO2exhibits strong basic sites for CO2adsorption as also demonstrated by CO2- TPD. In addition, a large peak of CO formation at around 626°C on Ni/ZrO2is mainly due to the RWGS reaction that is favourable occurred at high reaction temperatures. Furthermore, in situ DRIFT spectra were collected to identify the surface species and speculate on the possible reaction mechanism, as depicted in Figure 5b and Figures S20–S23. As the Conv.CO2 is still very high at 240°C for Ni/ZrO2, its in situ CO2methanation was operated at 210°C Figure 4. Ni K-edge and Zr K-edge XAS spectra of Ni/SiAlZr catalysts. (a, c) XANES spectra (normalized) and (b,d) k2or k3-weighted R-space EXAFS spectra. Angewandte Chemie Research Articles Angew. Chem. Int. Ed. 2024,63, e202317669 (7 of 11) © 2023 The Authors. Angewandte Chemie International Edition published by Wiley-VCH GmbH 15213773, 2024, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/anie.202317669 by Universidad De Sevilla, Wiley Online Library on [28/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License while the other two samples were operated at 240°C to observe the possible reaction intermediates. As a result, the key surface species over Ni/SiAlZr catalysts are much different. For the Ni/SiO2, we can observe the strong peaks of CO* linearly adsorbed on Ni sites (2040, 1920 cm1) and bridged CO* on active sites (1852 cm1) as well as CHO* (1760 cm1). However, the formate peak at 1585 cm1is very weak. For the Ni/Al2O3, the intensities of CO* and CHO* peaks decrease much while the intensity of formate increases considerably. Furthermore, the CO* and CHO* peaks could not be observed for Ni/ZrO2, which only presents the formate peak at 1593 cm1as the key intermediate. The CO* mainly originated from dissociative adsorption of CO2and decomposition of formate species. Thus, the reaction mechanism changes from a dissociative mechanism to an associative mechanism when the support is switched from SiO2to ZrO2. The other important surface species such as CHx, methane, carbonates, and surface hydroxyl groups could also be detected, as shown in Figure 5b and Figures S20–S23. To further verify the reaction mechanism of Ni/ZrO2, the in situ DRIFTS spectra on the reference samples were also obtained (Figure S22). The Ni/ZrO2-IM only presents some hydroxyl groups and a weak formate peak, without showing the peaks of CO* and even the carbonates and methane. Similarly, only some peaks of hydroxyl groups and carbonates could be found on the pure ZrO2support. It suggests that the ZrO2support is significant for CO2 adsorption and activation, while the Ni metal is beneficial for hydrogen activation and further hydrogenation of intermediates. Therefore, the Ni/ZrO2possesses suitable exposed Ni metals and Ni-O-Zr as well as abundant oxygen vacancies that are active for CO2methanation through the direct formate hydrogenation pathway, which was also demonstrated by the transient in situ DRIFTS experiments. As shown in Figure S23 and Figure 5c, the Ni/ZrO2catalyst was firstly exposed to pure CO2then the system was switched to other reaction atmospheres, resulting in considerable change of the surface species. Different types of hydroxyl groups at around 3733–3590 cm1belonged to bicarbonates species that appeared and increased after the introduction of CO2gas. The bicarbonates were generated from CO2adsorption on OHgroups of ZrO2support.[37] The observation of obvious hydroxyl groups at around 3750–3500 cm1was also reported in Ni/Al2O3,[37] Ru/ CeCrOx,[38] and LaNiO3/CeO2catalysts during in situ DRIFTS test of CO2methanation.[39] Some other carbonates (1637, 1273, 1250, 1050 cm1) and formate (1593, 1369 cm1) Figure 5. The results of the potential reaction mechanism. (a) The TPSR profiles of Ni/ZrO2, Note: CO2:m/z=44, H2O: m/z=18, CH4:m/z=16, CO: m/z=28. (b) The results of in situ DRIFTS tests on CO2hydrogenation over Ni/SiAlZr catalysts. The reaction temperature for Ni/SiO2and Ni/ Al2O3was 240°C while for Ni/ZrO2was 210°C. (c) Transient in situ DRIFTS experiments on Ni/ZrO2catalyst collected at 210 °C when the reaction atmospheres were switched from CO2to H2, CO2+H2, and Ar. The peak intensity of HCOO* species at 1593 cm1as a function of time. Angewandte Chemie Research Articles Angew. Chem. Int. Ed. 2024,63, e202317669 (8 of 11) © 2023 The Authors. Angewandte Chemie International Edition published by Wiley-VCH GmbH 15213773, 2024, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/anie.202317669 by Universidad De Sevilla, Wiley Online Library on [28/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License peaks have also presented in Figure S23a.[40] It indicates that these surface species could not be further converted and there was no peak belonging to *CO and *CH4. The formate formation was because that chemisorbed CO2reacted with hydrogen available on the catalyst surface (Figure S23a), which was consistent with the in situ DRIFTS results over Ru/CeO2catalyst for the formate formation under CO2/He and H2/He atmospheres.[41] López-Rodríguez et al. also observed formate peak under CO2/He atmosphere and then decreased under H2/He atmosphere.[41] After 30 mins, the gas was switched to hydrogen. The peaks of hydroxyl groups disappeared quickly, and the other peaks of carbonates and formate also became much weaker (Figure S23a). Upon switching the reaction atmosphere from hydrogen to CO2+H2(1:4 v/v), new peaks can be observed at 3015 and 1301 cm1, which are attributed to *CHxand methane (Figure S23b).[42] With time increased, the intensities of *CHxand methane peaks increased while the intensities of peaks attributed to hydroxyl groups, carbonates and formate decreased, which suggests that these are the important reaction intermediates. In particular, the intensity of formate increased quickly and then decreased slightly. Finally, the gas flow was switched to Ar, and most of the peaks disappeared while a small formate peak at 1593 cm1still could be observed (Figure S23b), indicating that it could be stably adsorbed. This was maybe because the reaction time was not enough and the reaction temperature was low. The strong basic sites and surface oxygen vacancies over Ni/ZrO2 could adsorb these reaction intermediates. The variation tendency of formate intensity as a function of time is illustrated in Figure 5c. It clearly shows that the formate could be produced and increased in the CO2/CO2+H2 atmospheres due to CO2reaction with surface OH groups.[22b] During the whole test procedure, the peaks attributed to CO* could not be observed. In addition, the ion chromatography results show that the amount of formate was 3.1 mg/L at 230°C and increased to 15.7 mg/L at 200°C, suggesting that formate is an important intermediate. Therefore, the HCOO* path instead of CO* is the main reaction path for CO2methanation on the Ni/ZrO2catalyst, which was also consistent with the literature reports.[22b,43] Conclusion In summary, extraordinary low-temperature CO2-methana- tion catalysts have been designed and synthesized in this work. Characterizations like TEM, XPS, and Py-IR showcase regulation of the SMSI and oxygen vacancies, which further affect the local electron density of active sites. Furthermore, the chemisorption and activation of CO2and H2dissociation would be affected, and thus influence the catalytic results through different reaction pathways from a dissociative mechanism over Ni/SiO2to an associative mechanism over Ni/ZrO2. Experimental results further prove that the Ni/ZrO2catalyst with a re-constructed structure of more monoclinic ZrO2species displays outstanding catalytic behaviour outperforming the benchmark Ni-based catalysts and even superior to a wide range of noble metal-based catalysts reported so far. Indeed, methane formation on Ni/ZrO2is kinetically and thermodynamically way more advantageous compared to a Ni/SiO2 catalyst. This work opens new research avenues to design robust catalysts for CO2hydrogenation based on the regulation of SMSI and then the local electron density of reaction active sites. Therefore, it is crucial to tune SMSI by engineering the support properties and metal loading methods toward boosting low-temperature CO2hydrogena- tion performance. Acknowledgements This work was supported financially by the National Key R&D Program of China (Nos. 2022YFA1504500, 2021YFA1502804), National Natural Science Foundation of China (Nos. 22005296, 21868016), Natural Science Foundation of Liaoning Province, China (No. 2020-YQ-01), Natural Science Foundation of Shanxi Province for Distinguished Young Scholar (No. 20210302121005), Natural Science Foundation of Jiangxi Province for Distinguished Young Scholars (No. 20232ACB213001), Thousand Talents Plan of Jiangxi Province (No. jxsq2023101072), and Natural Science Foundation of Chongqing, China (No. CSTB2022NSCQ- MSX0231). This work was also partially funded by the Eurpean Commission through the BIOALL project (Grant agreement: 101008058). The authors would also like to thank Shiyanjia Lab (www.shiyanjia.com) for the XPS analysis. Conflict of Interest The authors declare no conflict of interest. Data Availability Statement The data that support the findings of this study are available from the corresponding author upon reasonable request. Keywords: CO2Methanation ·Low-Temperature Activity · Ni-Based Catalysts ·Reaction Pathway ·Strong Metal-Support Interactions [1] a) C. Hepburn, E. Adlen, J. Beddington, E. A. Carter, S. Fuss, N. Mac Dowell, J. C. Minx, P. Smith, C. K. Williams, Nature 2019,575, 87–97; b) J. Ding, Q. Tang, Y. Fu, Y. Zhang, J. Hu, T. Li, Q. Zhong, M. Fan, H. H. Kung, J. Am. Chem. Soc. 2022, 144, 9576–9585. [2] a) X. Ren, M. Guo, H. Li, C. Li, L. Yu, J. Liu, Q. Yang, Angew. Chem. Int. Ed. 2019,58, 14483–14488; b) C. Dong, Q. Yu, R. P. Ye, P. Su, J. Liu, G. H. Wang, Angew. Chem. Int. Ed. 2020,59, 18374–18379. [3] a) R. P. Ye, J. Ding, W. Gong, M. D. Argyle, Q. Zhong, Y. Wang, C. K. Russell, Z. Xu, A. G. Russell, Q. Li, M. Fan, Y. G. Yao, Nat. Commun. 2019,10, 5698; b) J. Wang, G. Li, Z. Li, C. Tang, Z. Feng, H. An, H. Liu, T. Liu, C. Li, Sci. Adv. 2017,3, Angewandte Chemie Research Articles Angew. Chem. Int. Ed. 2024,63, e202317669 (9 of 11) © 2023 The Authors. Angewandte Chemie International Edition published by Wiley-VCH GmbH 15213773, 2024, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/anie.202317669 by Universidad De Sevilla, Wiley Online Library on [28/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License