Metal–organic frameworks (MOFs) have drawn attention as potential catalysts owing to their unique tunable surface chemistry and accessibility. However, their application in thermal catalysis has been limited because of their instability under harsh temperatures and pressures, such as the hydrogenation of CO2 to methanol. Herein, we synthesized Cu nanoparticles encapsulated in a zeolitic imidazolate framework-8 (ZIF-8) using a controlled two-step method to obtain a Cu nanoparticle-containing Zn-based MOF catalyst that is highly active, selective, and remarkably stable for the CO2 hydrogenation to methanol reaction. Two different catalysts based on the Cu on ZIF-8 MOF were synthesized, thoroughly characterized experimentally, and studied theoretically. The catalyst prepared with the optimized pathway is active (2.2 gmethanol gmetal−1 h−1) and selective (>90%), and stable (for >200 h) at 523 K and 50 bar, displaying superior methanol space–time yields than the benchmark Cu–Zn–Al industrial catalyst for CO2 hydrogenation. Density functional theory-assisted calculations reveal that Cu–O–Zn sites were active for CO2 adsorption with an adsorption energy of −2.82 eV. The developed catalyst is an inexpensive, robust, and easy-to-synthesize on-scale active catalyst for selectively hydrogenating CO2 to methanol.
Article
Copper nanoparticles encapsulated in zeolitic imidazolate framework-8 as a highly stable and selective catalyst for CO2 hydrogenation to methanol
https://doi.org/10.21203/rs.3.rs-2591998/v1
This work is licensed under a CC BY 4.0 License
Journal Publication
published 06 Mar, 2024
Version 1
posted
You are reading this latest preprint version
Copper catalysis
MOF
CO2 hydrogenation
methanol
zeolitic imidazolate framework (ZIF)
Metal–metal oxide interfaces possess unique catalytic properties for various catalytic processes, and such phase boundaries have been demonstrated to be highly effective for hydrogenating CO2 to methanol.1–4 For the hydrogenation of CO2 to methanol reaction, the interfaces between metallic Cu and ZnO or ZrO2 are catalytically active;5–7 however, because of the highly dynamic behavior of Cu nanoparticles under working conditions, the Cu species undergo restructuring during CO2 hydrogenation, and this leads to the sintering of active phases and results in a pronounced drop in catalytic activity and selectivity.5,8–11 Although multiple strategies have been explored to address the sintering problem (e.g., multicomponent catalysts with different structural promoters), a radically new systematic approach is required to advance CO2 hydrogenation technology.1,12–14 Recently, metal–organic frameworks (MOFs) have been explored to solve the sintering problem by immobilizing Cu particles in the defect sites of MOFs.15–18 Capturing Cu in the porous structure of a MOF makes these particles more resilient to restructuring, thereby making them harder to sinter. To our knowledge, the pioneering work by Bing et al. was the first to use MOFs as support for Cu catalysts in CO2 hydrogenation, thus opening the door for use of MOFs as catalysts in many different applications. Bing et al. immobilized Cu/ZnOx nanoparticles in the pores of UiO-bpy, an extremely stable MOF with small cavities.18 Recently, other researchers have successfully explored Cu-containing Zr-based MOFs (UiO-66 and MOF-808) as catalysts.1,14,17 These authors demonstrated that MOFs have multiple positive effects when used as catalyst support for CO2 hydrogenation.
On the one hand, the porous MOF structure can prevent or reduce Cu particle sintering, which is advantageous for preserving the catalytic function; on the other hand, structural defects in the MOF can generate a MOF-nanoparticle interface with unique catalytic properties. To date, all studies have focused on Zr-based MOFs owing to their stability and Zr promotion effect. To our knowledge, no work has been performed on the immobilization of Cu nanoparticles on Zn-based MOFs to hydrogenate CO2 to methanol. However, ZnO is a promoter in the current industrial hydrogen setups.19,20 Furthermore, a portfolio of MOFs prepared with Zn that are stable at high temperatures and pressures exists, e.g., zeolitic imidazolate frameworks (ZIFs).
We developed a robust two-step synthesis of a Cu nanoparticle containing Zn-based MOF for CO2 to methanol hydrogenation catalyst that is highly active, selective, and remarkably stable. The catalyst is based on a ZIF-8 supporting MOF in which atomically dispersed Cu2+ species are reduced to stabilized Cu sub-nanometric clusters close to Zn2+ sites (Fig. 1 and Scheme 1). Different catalysts based on the Cu on ZIF-8 MOF were synthesized and thoroughly characterized by different techniques such as operando spectroscopy, theoretical calculations of the disposition of the Cu nanoparticles and adsorption capacity, and activity–stability test of CO2 hydrogenation to methanol. With a complete fundamental understanding of Cu structural and electronic features in CO2 hydrogenation, a sustainable methanol production solution will be acheivable.21
Synthesis and structure of the catalysts: Two methods were used to integrate Cu into the ZIF-8 MOF: ion exchange (Fig. 1a, Fig. S1-S2) and conventional wet impregnation. To prepare the ion exchanged catalyst (Cu/ZIF-8|IE|), we introduced the ZIF-8 MOF in an ethanol solution of Cu(NO3)2. Then, the sample was washed in a Soxhlet with ethanol to remove the metal not strongly anchored to the MOF. In this way, we incorporated up to 12 wt.% of Cu in the MOF (Table 1). During ion exchange treatment, defects are generated in the MOF structure, and parts of linkers can be released.20 These missing linkers can be observed using infrared spectroscopy (Fig. 1d), where the broadband centered at 1311 cm−1 is observed because of linker deficiency. In the same figure, no peak between 1100 and 1150 cm−1 was observed, indicating metallic defects were not present. These missing linkers were replaced via hydroxyl groups, as can be seen by IR (broad band between 3250 and 3700 cm−1, Fig. S3). Note that Cu is anchored to these hydroxyl group ligands (Fig. 1a).
The samples were then subjected to reduction treatment at different temperatures to reduce Cu2+ to small Cu metallic clusters and characterized using XRD (Fig. 1b). The diffractograms do not change even when the sample was reduced at 723 K. This indicates that the sample was extremely stable under the reduction conditions. Only a small shift of the peaks is seen before reducing the sample, thus reinforcing the idea that Cu is incorporated into the structure after reduction treatment. Cu phases were not reported even after reduction at 723 K, which could be attributed to the fact that Cu particles are highly dispersed in the catalyst. However, the diffraction pattern of the sample prepared by wet impregnation Cu/ZIF-8|IM| (Fig. S4) shows that the MOF degrades during the impregnation process and that the structural detriment is even more accentuated when reduced at 523 K. Peaks belonging to crystalline Cu can be seen using the wet impregnation method. It is interesting to know that the structural arrangement of the prepared Cu/ZIF-8|IE|R catalyst did not change after being stored (in a glass vial) for more than three years, explaining its long shelf-life (Fig. S5).
Thermogravimetric studies then demonstrated the high stability of these materials (Fig. S6). Nitrogen adsorption isotherms (Fig. 1c) of the catalysts demonstrated that the exchange method developed in this study was essential for achieving the proper dispersion of Cu species in the ZIF-8 structure without considerably decreasing the specific surface area of the MOF. Synthesis of the same catalyst using the traditional impregnation method led to MOF degradation and a loss of all accessible porosity (no proper hysteresis is seen).22–25
Morphology of the catalysts: From the TEM analysis of ZIF-8 before and after Cu exchange and reduction (Fig. 2a, Fig. 2b, and Fig. S7), we could not detect Cu nanoparticles of sufficient size that could be differentiated from that of the ZIF-8 crystals. This confirms that the Cu species are highly dispersed (ultra-small or sub-nanometric level) throughout the MOF structure and provides evidence that the Cu particles are stabilized (as metallic species observed by XPS; Fig. 3e) into the ZIF structure. For comparison, the sample prepared by the traditional impregnation method demonstrated large Cu particles (Fig. S8). HAADF (Fig. 2d and Fig. 2e), along with STEM-EDX of the ion exchange sample, showed a uniform distribution of all elements (Zn, O, N and C) in the ZIF-8 structure with rhombic dodecahedron morphology. Furthermore, the presence of Cu species within the MOF crystal was clearly visible.
Surface acid-base and metal properties of the catalysts: The basicity and acidity of the catalysts were analyzed using CO2 temperature programmed desorption (CO2-TPD) and NH3-TPD techniques. As shown in Fig. 3a, the sample prepared using the ion exchange method after reduction (Cu/ZIF-8|IE|R) had higher basicity than ZIF-8 alone. It is well known that ZIF-8 interacts very weakly with CO2. This increase in basicity after incorporating Cu may be attributed to the presence of –OH groups generated during the exchange reaction (Fig. S3 and Scheme 1). The sample prepared via impregnation also showed higher basicity than ZIF-8. We attributed this to the decomposition of the material via reduction. Using NH3-TPD (Fig. 3b), we observed an increase in the acidity of the samples with the introduction of Cu. However, this increase is much more accentuated in the case of the impregnated sample, as its decomposition leads to the formation of acid sites, as shown in Fig. 1.
XPS analysis of all the catalysts was performed for each element, and the results are shown in Fig. 3 and Fig. S9. The Zn 2p core level spectra of all samples (Fig. 3c) revealed the presence of Zn (II) species in the catalysts with binding energy (BE) peaks at 1021.4 eV and 1044.5 eV.27 In the Auger spectra of the ZIF-8 and Cu/ZIF-8|IE| samples (Fig. 3d), the Zn Auger region consisted of two peaks at 498 eV and 495 eV because of Zn2+ and Zn0, respectively. The presence of Zn0 in both samples before any reduction treatment can be ascribed to photoreduction from the highly energetic X-rays used in the XPS technique.18 After reduction at 523 K in H2, the main peak at 498 eV was more pronounced and contained a small shoulder at 495 eV, indicating that some of the Zn2+ sites might have been reduced during hydrogenation.
The Cu 2p spectrum (Fig. 3e) of the Cu/ZIF-8|IE| sample confirms the presence of oxidized Cu2+ species with satellite peaks.28,29 After reduction at 523 K (Cu/ZIF-8|IE|R), the CuO species were reduced to metallic Cu species, as confirmed by the presence of a peak at932 eV (Fig. 3e). The small peak is attributed to Cu2+ from CuO that can be attributed to the outer layer of the sample from atmospheric oxygen exposure during sample preparation and transfer to the XPS apparatus.30 The absence of a shakeup satellite peak at 947.6 eV confirms the absence of Cu+ species on the surface. Compared to bulk Cu0 with Cu 2p1/2 at 952.5 eV, the Cu-ion exchanged sample demonstrated a lower Cu binding energy (Cu 2p1/2 at 951.7 eV). This likely resulted from electron injection from the conduction band of ZnO2 to CuO, and provides evidence for a strong interaction between Cu and Zn species.18,31 The same was confirmed with Auger spectroscopy (Cu LMM) of these samples (Fig. 3f). No peak for Cu+ was observed near 570 eV, and the peak at 568 eV confirms the presence of predominately Cu0 species on the reduced catalyst.32
The relative surface compositions obtained from this analysis demonstrated remarkably higher surface oxygen concentration after the ion exchange procedure (Table S2). After reduction, a slight decrease in surface oxygen concentration was observed. Both results were in good correlation with the in situ DRIFTS results in which an intense vibrational band due to –OH stretching was observed after Cu exchange and the subsequent reduction reaction (Fig. 4b). Note that the reduction of the sample in H2 resulted in an additional O 1s peak near 533 eV because of the superficial –OH groups (Fig. S9).
Local environments of Cu and Zn: To determine the local arrangements of Cu and Zn species, XAS analysis was performed on Cu-ZIF-8|IE| and after aging the sample in the CO2+H2 reaction mixture under working conditions (Fig. 4a) for 100 h. The two resulting spectra are completely different. The freshly prepared sample and the sample after CO2+H2 treatment show the typical XANES profiles of Cu2+ and Cu0, respectively.33,34 This is also observed in the curves of the first derivative of the XANES spectra. As shown in Fig. 4b, the derivative spectrum of the freshly prepared sample presents a weak peak at ~8977 eV for the dipole-forbidden 1s → 3d transition and the main peak at ~8986.6 eV for dipole-allowed 1s → 4p transition, which is characteristic of Cu2+.35 For the characteristics of the 1s → 4p transition, the Cu2+ species in the freshly prepared sample demonstrate a positive energy shift of ~1.0 eV compared to Cu2+ in Cu(OH)2, indicating that a part of the Cu2+ species is in a distorted state.36 The derivative spectrum of the used sample demonstrates an edge-energy feature at ~8979 eV for the 1s → 4p transition, which is characteristic of reduced Cu.37 All these results agree with the XPS results (Fig. 3).
For Zn K-edge XANES spectra (Fig. 4c), we see that both spectra are highly similar and correspond to the published spectra of ZIF-8. This indicates that the local environment of Zn mainly comprises ZnN4.38 In other words, most Zn is part of the ZIF-8 structure. After subtracting the spectrum of the catalyst before using it from the sample after treatment, we see that a small peak appears at 9662 eV, indicating that the Zn white line is altered during the CO2 reduction reaction. This Zn white line observation has been previously reported in other Cu/ZnO systems.33,39 In this study, this change was attributed to the formation of interfaces between reduced Cu and partially reduced Zn<2+.39 These results agree with the Zn 2p XPS spectra in Fig. 3c.
Activity and stability in the CO2 hydrogenation to methanol reaction: The prepared catalysts were screened below 10% CO2 conversion levels under 50 bar pressure (Fig. S10 and Fig. S11). The main reduction products were methanol and CO at every temperature screened. Irrespective of the different temperatures and pressures used for the Cu/ZIF-8|IE|R catalyst, we did not detect any methane production, and the testing with different particle sizes (Fig. S12) confirmed no diffusion limitations for our catalyst. The synthesized Cu/ZIF-8|IE| catalyst by the ion exchange method showed sustained activity for more than 100 h of reaction time with no significant changes in selectivity after the initial induction period of around 70 h (Fig. 5a). Depending on the exact catalyst, various induction periods were required before methanol formation to allow the catalysts to complete the necessary structural rearrangements.12,40 Remark: The Cu/ZIF-8|IE|R synthesized by the ion exchange method required approximately 70 h for induction. Increasing the reduction time during the catalyst synthesis step did not result in any differences in terms of Cu morphology, as shown in Fig. 1b. However, after more than 70 h in the reaction mixture (CO2+H2), Cu nanoparticles could be seen from powder XRD and HAADF-STEM-EDX analyses (Fig. 5d and Fig. 5f, respectively). The catalyst showed >90% selectivity throughout the course of the CO2 reduction reaction, greatly outperforming the standard Cu-Zn-Al catalyst for both selectivity (>90% vs 60%) and reaction rate (2.2 gmethanol gmetal-1 h-1 vs 1.4 gmethanol gmetal-1 h-1) under similar reaction conditions.
The prepared Cu/ZIF-8|IE| displayed significantly higher methanol selectivity and corresponding space-time yield (STY) than Cu/ZIF-8|IM| and the benchmark Cu-Zn-Al. The parent ZIF-8 did not show any activity (results not presented) under the experimental conditions employed. The higher selectivity of CO over the Cu/ZIF-8|IM| catalyst is due to the agglomeration of large-sized Cu particles on the ZIF-8 MOF (Fig. S8). The higher methanol selectivity observed over the Cu-ion-exchanged sample should be attributed to the presence of finely dispersed Cu nanoparticles (Fig. 5e, Fig. 5f), leading to a higher CO2-adsorption capacity (Table S1) from the increased availability of surface hydroxyl species observed from both, XPS (Table S2) and in situ DRIFTS (Fig. 6b).
As shown in Fig. S11, the reaction temperature significantly affected the reduction product selectivity over the Cu/ZIF-8|IE| catalyst. The trend observed here correlates well with equilibrium concentrations under a similar protocol (Fig. S13). As observed in both cases, an inverse relationship is observed between methanol and CO selectivity as the reaction temperature increases from 498 to 573 K. Optimized methanol production was observed at 523 K, above which the catalyst showed more CO production with high CO2 conversion. The decline in methanol selectivity at higher temperatures can be explained by the endothermic reverse water–gas shift (rWGS) pathway (CO production) is thermodynamically more favorable at high temperatures, enhancing the production of CO over methanol.41,42
The chemical environment (XPS; Fig. 5c and Fig. S14), catalyst structure (from XRD; Fig. 5c), and catalyst morphology (Fig. 5e, Fig. 5f) after 100 h reaction time are compared with that of the reduced form of the catalyst. The powder XRD analysis of Cu/ZIF-8|IE|R after the reaction showed a small diffraction peak at 2θ = 43.5° (Fig. 5d), confirming metallic Cu presence. The surface compositions obtained from the XPS analysis (Table S2) showed a slight decrease in the surface oxygen with no changes to the Cu/Zn ratio before and after the CO2 reduction reaction, confirming no loss of Cu species during CO2 conversion. Therefore, after exposure to the reaction mixture (in situ reduction), further clustering of the finely dispersed Cu in the ZIF-8 MOF occur, leading to the formation of Cu nanoparticles (Fig. 5g) as seen in the HAADF-STEM-EDX analysis of the catalyst after 100 h of time on stream.
Identification of reaction intermediates: The structure and stability of the catalysts, and the reaction mechanism, were studied using in situ DRIFT spectroscopy at different pressures (1–25 bar) and temperatures (323–523 K; Fig. 6). ZIF-8 and Cu/ZIF-8|IE|R share identical bands with less intense bands observed in the Cu/ZIF-8|IM|R sample, primarily due to the structural collapse of the MOF. Additionally, a highly intense and broad vibrational band is observed in the range of 3150 to 3700 cm-1 that can be assigned to the –O–H stretching vibrations in –OH/OH2 groups, even after reduction at 523 K (Fig. 6b and Fig. S3). Exploring more deeply, we also collected the IR spectra after introducing the reaction mixture CO2:H2 under 25 bar pressure at the reaction temperature after in situ reduction of the sample. This resulted in new peaks between 3150 and 3700 cm-1, mostly due to the adsorption of reactants and the interaction of CO2 with the surface hydroxyl groups (Fig. 6c and Fig. S3).
We analyzed the reaction mechanism by monitoring the intermediates and products from Cu/ZIF-8|IE|R in CO2+H2 at reaction temperature (523 K) under 25 bar pressure by in situ IR spectroscopy (Fig. 6). We propose the reaction mechanism pathway as depicted in Fig. 6a based on the observed reaction intermediates. These experiments were performed at the limit allowed by the cell for safe operation, which was 25 bar (Fig. S10). The in situ IR spectra of Cu/ZIF-8|IE| (Fig. 6b and Fig. 6c) exposed to the CO2+H2 mixture at different pressures (10–25 bar) showed the formation of CO adsorbed on different Cu sites (Fig. 6e, and Fig. 6f). The increase in the intensity of the vibrational band for CO adsorption with pressure confirms the high activity of CO2 conversion at high pressures. The bands at 2093, 2078, and 2060 cm−1 can be assigned to linearly adsorbed CO molecules on metallic Cu species.26 We also observed methoxy species, i.e., two pairs of bands at 2960–2910 cm−1 and 2865 cm−1, corresponding to the ν(CH3) and δs(CH3) vibrations and bicarbonate/carbonates (Fig. 6d) in the region of 1750–1690 cm−1 and 1680–1600 cm−1.26,43 The bands at 2971, 2930, 2888, and 2750 cm−1 were assigned to the formate (HCOO)−metal species(Fig. 6g).17,44 Additionally, the difference IR spectra also show a gradual increase in CO vibrational band intensity with increasing pressure, confirming high conversion at higher pressures (Fig. 6f). Due to the superimposing features of methanol vibrational bands (–OH and –CH) with that of the surface hydroxyl and methoxy groups, it was challenging to monitor methanol formation in the IR cell directly. Thus, the observed HCOO− species was identified as a possible key reaction intermediate in the current reaction pathway.
Adsorption behavior of CO2 studied by density functional theory (DFT): To understand the adsorption properties of CO2 for a given site, we carried out ab initio calculations using DFT (Fig. 7). Based on the results from XRD, IR, XAS, and XPS, we optimized the structure of the Cu cluster encapsulated on the ZIF-8 through linker vacancies (Fig. S15). In this connection, 4 different (key and possible from the obtained results and reported literature)1,20 cases are considered here to study the adsorption of CO2 on (1) Cu sites on the Cu cluster, (2) Zn2+–O2-–Zn2+ sites in a tridentate complex, (3) Cu/ZnO interfacial Cu–Zn2+ sites, and (4) Cu–O–Zn interfacial site (Fig. 7).
The binding of CO2 is not energetically favored on the Cu metal particle (Fig. 7a), as the adsorption energy (ΔEads) is slightly positive (0.21 eV) and the studied case is also consistent with the weak adsorption of CO2 found on Cu surfaces.26 In addition, no adsorption was observed when CO2 was exposed to bare ZIF-8 (no Cu cluster or linker vacancy), consistent with experimental results where no activity was observed in CO2 hydrogenation. Insignificant CO2 uptake was observed in the CO2-TPD experiments on bare ZIF-8 (Fig. 3a). When considering CO2 adsorption on to the Zn2+-O2- sites (to form a bi/tridentate carbonate with an adjacent Cu cluster; Fig. 7b), an adsorption energy of –1.39 eV was found. The adsorption of CO2 at the Cu–Zn2+ interfacial sites (Fig. 7c) was significantly stronger (ΔEads= −1.62 eV) than the adsorption on the Cu nanoparticles and ZnO nodes. CO2 is stabilized via electron transfer from the Cu nanoparticle to CO2 upon binding C to the metal, along with the stabilization of the negatively charged oxygens by the interfacial Lewis acid Zn2+ and Cu sites. Finally, we can conclude that the Cu-O–Zn site was the most preferred site for CO2 adsorption with an adsorption energy of -2.82 eV (Fig. 7d), whereas the Cu site was the weakest adsorption site.
Highly dispersed Cu nanoparticles encapsulated in a ZIF-8 MOF were synthesized by a simple ion exchange method via post-synthetic metalation of ZIF-8 with Cu(NO3)2. The Cu species formed after reduction were stabilized by the ZIF-8 structure, thereby conferring high stability to the Cu site during the CO2 hydrogenation reaction. In addition, the chemical environment surrounding these nanometric particles gives the catalyst a previously unknown catalytic activity per gram of active phase. The synthesis strategy employed here opens a new pathway for chemically grafting Cu particles, and likely of other metals, that retain metal nanoparticle character at the interface while only directly replacing a fraction of the cations in the lattice. We show that the direct interaction between Zn-MOF crystals and Cu particles is mandatory for the improved catalytic reduction of CO2 to methanol with the prepared Cu/ZIF-8|IE| catalysts, as no activity was observed over the bare ZIF-8. The preliminary in situ DRIFTS studies conducted for the Cu/ZIF-8|IE|R catalyst showed a significant increase of hydroxyl groups formed through linker vacancies in the ZIF-8 structure. In contrast, in situ DRIFTS analysis of the Cu/ZIF-8|IM| catalyst showed signs of both linker vacancies and Zn vacancies that led to the collapse of the MOF network. The in situ adsorption of the reaction mixture on the Cu/ZIF-8|IE| sample in the DRIFTS cell confirmed the presence of formate, which was identified as the main reaction intermediate from CO2 reduction to methanol over this catalyst.
As further confirmed after DFT calculations, the constituents of the Zn–O–Cu interface formed through significant metal-support interactions are at least part of the active site that strongly adsorbs CO2. The high concentration of hydroxyl groups (required for high CO2 chemisorption) and the highly dispersed metallic Cu (necessary for CO2 reduction) at the interface are hypothesized to explain the much higher activity of the ion exchanged catalyst produced in this study.
In this study, we encapsulated Cu nanoparticles in a Zn-based ZIF-8 MOF through a simple ion exchange method without breaking a significant portion of the MOF network. Defects were purposefully created for Cu2+ loading in the created linker vacancies via the deprotonation of–OH groups. During the subsequent reduction step, these Cu species reorganized into well-dispersed nanoparticles stabilized in the MOF structure through the same –OH linker groups. We demonstrated that the resulting material had extremely high activity in the CO2 hydrogenation to methanol reaction and outstanding stability compared with other MOF-derived catalysts. The Cu/ZIF-8|IE|R catalyst presented here is active (2.2 gmethanol gmetal−1 h−1) and selective (>90%) for 200 h at 523 K and 50 bar.
The developed catalyst displays higher methanol space–time yields than the benchmark Cu–Zn–Al industrial catalyst in CO2 hydrogenation at 523 K. Crucial to this higher activity, selectivity, and stability are the (i) highly uniform distribution of active Cu0 sites encapsulated in a Zn-based MOF that act as CO2 and H2 adsorption sites and (ii) the stabilization of these sites by the framework and–OH groups remaining in the catalyst after Cu reduction and the CO2 to methanol reaction.
The developed catalyst is exciting because it is relatively inexpensive to synthesize on larger scale and opens novel possibilities for the industrial implementation of MOFs in thermal catalysis.
Catalyst synthesis: The parent ZIF-8 (Basolite® or Z1200; Zn(C4H5N2)2 is a Zn 2-methylimidazolate material) and Cu–ZnO–Al2O3 (Cu–Zn–Al; pellets having a diameter of 5.5 mm and height of 3.6 mm) were commercially purchased from Merck (produced by BASF) and Alfa Aeser (product No. 45776), respectively. Cu2+ exchange in ZIF-8 was performed by suspending 3 g of the commercial ZIF-8 (Basoliote® Z1200) in a solution prepared with 12.58 g of Cu nitrate hexahydrate (Cu(NO3)2∙6H2O, purity >99%,12.58 g Sigma-Aldrich) and 100 mL of ethanol. The suspension was placed in a glass bottle, secured with a screw cap, and maintained at 323 K for three days. The resulting solid was separated by filtration and washed in a Soxhlet extractor with ethanol to remove the unreacted Cu and the extracted Zn species. Then, the sample was dried in an oven at 373 K for 24 h to obtain Cu/ZIF-8|IE|. The resulting Cu/ZIF-8|IE| was then reduced to Cu/ZIF-8|IE|R in H2/Ar (6%; vol/vol) at 523 K for 3 h with a ramp rate of 10 K min–1.
ZIF-8 was impregnated by Cu (in similar loading, 12% as confirmed by ICP-OES analysis) using the wetness impregnation method. ZIF-8 support was impregnated with Cu nitrate using ethanol as a solvent with constant stirring for 2 h at 300 K. The solvent was evaporated by drying at 353 K and the resulting Cu/ZIF-8|IM| was reduced following the same method as above to obtain Cu/ZIF-8|IM|R. The contents of Cu and Zn metals for all the samples were determined from ICP-OES analysis (Table S1).
Material characterization: Crystallographic analyses were performed via powder X-ray diffraction (PXRD) in a Bruker D8 advanced diffractometer equipped with a Bragg–Brentano geometry fitted with a Cu tube operating at 40 kV and 40 mA. Diffractograms were acquired over a 2θ range of 10–80° with a step size of 0.1° and a scan speed of 0.5 s per step. PDF-4+ (2019) was the database used for phase identification. Crystallite size was determined with the Debye–Scherrer equation.
Structural properties were studied using N2 adsorption-desorption (Micromeritics ASAP 2040) at 77 K. Before the analysis, the samples were degassed for 10 h at 523 K. The BET equation computed the specific surface area (SBET) from the resulting isotherms. Thermogravimetric analysis was carried out using a gas controller GC 200 model TGA (Mettler Toledo) device. As-synthesized sample (approximately 10 mg) was placed in an Al2O3 crucible and heated from 303 K to 1123 K with a heating rate of 10 K min-1 in a 30 mL min-1 flow of 6% H2 in Ar (vol/vol). An empty crucible was used as the reference material. Metal-support interactions were studied using temperature-programmed H2 reduction (H2 -TPR, Altamira AMI-200ip) with a thermal conductivity detector. Samples were pretreated in Ar at 473 K for 20 min to remove any adsorbed species. The measurements were done in a 5% H2in Ar with a total flow of 50 mL min−1 and a heating rate of 5 K min−1.
The NH3– and CO2–TPD experiments were carried out on an Autochem 2950 instrument with a TCD detector. The effluent mixture was analyzed by mass spectrometry (Hiden Analytical). The U-shaped quartz reactor was loaded with approximately 0.1 g of sample, using a heating rate of 10 K min-1. After an in situ reduction at 523 K under 10% H2/Ar for 1 h, the sample was cooled to 323 K before introducing a mixture of 10% NH3 in He or 10% CO2 in He for 60 min. Finally, the desorption step was carried out under pure flowing He, increasing temperature from 323 to 823 K.
X-ray photoelectron spectroscopy (XPS) was performed with a K-ALPHA spectrometer (Thermo Scientific). All spectra were collected using Al-Kα radiation (1486.6 eV), monochromatized by a twin crystal monochromator, yielding a focused X-ray spot with a diameter of 400 nm, at 3 mA × 12 kV. The alpha hemispherical analyzer was operated in constant energy mode with survey scan pass energies of 200 eV to measure the entire energy band and 50 eV in a narrow scan for selective measurements. Charge compensation was achieved with the system flood gun, providing low-energy electrons and low-energy argon ions from a single source. The powder samples were pressed, mounted on the sample holder, and placed in the vacuum chamber. Before recording the spectrum, the samples were maintained in the analysis chamber until a residual pressure of ca. 5×10−7 N m−2 was reached. Quantitative analysis was estimated by calculating the integral of each peak, after subtracting the S-shaped background and by fitting the experimental curve to a combination of Lorentzian (30%) and Gaussian (70%) lines.
Transmission electron microscopy (TEM) of the samples was performed with a Titan Themis-Z microscope from Thermo Fisher Scientific, operating at an accelerating voltage of 300 kV with a beam current of 0.5 nA. Darkfield imaging was performed by scanning TEM (STEM) coupled to a high-angle annular dark-field (HAADF) detector. The STEM-HAADF data were acquired with a convergence angle of 29.9 mrad and a HAADF inner angle of 30 mrad. Furthermore, an X-ray energy dispersive spectrometry (FEI SuperX, ≈0.7 sR collection angle) was utilized in conjunction with DF-STEM imaging to acquire STEM-EDS spectrum-imaging data sets (image size: 1024 × 1024 pixels, dwell time 5 μs).
Elemental analysis was performed using ICP-OES (Varian, Inc./Agilent Model 7200-ES). For microwave digestion, approximately 10 mg of catalyst was added to a mixture of concentrated HCl (1 mL), concentrated HNO3 (3 mL), and concentrated HF (1 mL) and subjected to a microwave-assisted digestion with a15 min ramp time and 30 min hold time at 1000 W and 493 K.
In situ IR studies: In situ diffuse reflectance Fourier transform infrared spectroscopy (DRIFTS) measurements were performed on a Nicolet 6700 FTIR spectrometer with a liquid-nitrogen cooled MCT detector using a high-temperature reaction cell (Harrick) equipped with a temperature programmer and connected to a gas-dosing device with mass flow controllers (Bronkhorst). The cell was filled with catalyst particles (150–300 μm). The sample was pretreated with 6% H2 in Ar at 523 K for 1 h with a ramp of 10 K min-1. The temperature decreased to 323 K, and a baseline spectrum was recorded after switching to CO2+H2 gas. Thereafter, the sample was exposed to the same gas mixture and pressurized to 25 bar before heating from 323 to 573 K. Spectra were recorded every 5 min. The presented difference spectra were obtained by subtracting the baseline spectrum from the adsorbate spectrum. The background spectrum of dried KBr was collected in flowing He at the measurement temperature. Due to the temperature gradient in the Harrick cell, a temperature calibration at the measured surface was done (against the set temperature in the programmer), as the temperatures at the surface of the catalyst bed are generally lower than that measured at the bottom.
Density functional theory (DFT) calculations: Adsorption energies were calculated using the Vienna Ab Initio Simulation Package (VASP). A plane wave basis set with a 500 eV energy cut-off was used for optimizing carbon dioxide adsorption on defect-containing primitive cells of ZIF-8. Note that 13 atom Cu clusters were selected and encapsulated via linker vacancies, as observed experimentally and shown in Fig. 1a. We selected the Cu cluster based on the unit cell used in this study. DFT with periodicity was selected because it is capable of accurately predicting structural transformations such as unit cell dimension in ZIF-8 (Cu cluster/Zn). In this study, the Perdew–Burke–Ernzerhof (PBE) functional is used to model electron exchange and correlation interactions, with a generalized gradient approximation. Using the projector-augmented wave (PAW) methodology, electron–ion interactions were determined. For the geometry optimization process, atoms in the linker vacancy connected to two Zn atoms, Cu cluster, and adsorbates on the ZIF-8 structure were relaxed, and atoms in the other ZIF-8 structure were fixed. The convergences for energy and forces were set to 1.0 × 10−6 eV and <0.05 eV Å−1, respectively.
Catalytic tests: Catalytic tests were performed using a Flowrence® Avantium parallel reactor comprising 4 or 16 tubular fixed-bed quartz reactors (2 mm ID, length 300 mm). Reactors were placed in a furnace and the flow was equally distributed over every channel using a microfluidic glass distributor. In each reactor, ~0.05 g of sieved catalyst particles (150–300 mm) were loaded on 300 mL of a SiC (particle grit 40) bed to ensure the catalyst bed rested in the isothermal zone of the reactor. One reactor was always used without a catalyst as a blank. The reactors were then pressurized with a mixed feed containing 20 vol.% of CO2 and 80 vol.% of H2 to 50 bar using a membrane-based pressure controller in the temperature range of 500 to 573 K. Moreover, 0.5 mL min–1 of He was added in the mixed feed and as internal standard. Reactor outputs were then analyzed using an Agilent 7890B gas chromatograph equipped with two sample loops, one TCD, and two FIDs detectors. The special configuration and columns of this analytical system permit us to identify and quantify H2, He, CO, CO2, or methanol, among other possible carbon-based products. Conversion (X, %), selectivities (S, %), and space–time yields (STY, gmethanol·gmetal−1·h−1) are expressed as follows:
Acknowledgments
King Abdullah University of Science and Technology (KAUST) funded this work: BAS/1/1403. The authors acknowledge the KAUST Supercomputing Laboratory for providing high-performance computational resources and support from the KAUST Core Labs.
- Zhu, Y. et al. Copper-zirconia interfaces in UiO-66 enable selective catalytic hydrogenation of CO2 to methanol. Nat. Commun. 2020 111 11, 1–11 (2020).
- Zhu, Y. et al. Inverse iron oxide/metal catalysts from galvanic replacement. Nat. Commun. 2020 111 11, 1–7 (2020).
- Wang, X., Shi, H. & Szanyi, J. Controlling selectivities in CO2 reduction through mechanistic understanding. Nat. Commun. 2017 81 8, 1–6 (2017).
- Graciani, J. et al. Highly active copper-ceria and copper-ceria-titania catalysts for methanol synthesis from CO2. Science (80-. ). 345, 546–550 (2014).
- Sehested, J. Industrial and scientific directions of methanol catalyst development. J. Catal. 371, 368–375 (2019).
- Bonura, G., Cordaro, M., Cannilla, C., Arena, F. & Frusteri, F. The changing nature of the active site of Cu-Zn-Zr catalysts for the CO2 hydrogenation reaction to methanol. Appl. Catal. B Environ. 152–153, 152–161 (2014).
- Kuld, S. et al. Quantifying the promotion of Cu catalysts by ZnO for methanol synthesis. Science (80-. ). 352, 969–974 (2016).
- Martin, O. et al. Operando Synchrotron X-ray Powder Diffraction and Modulated-Excitation Infrared Spectroscopy Elucidate the CO2 Promotion on a Commercial Methanol Synthesis Catalyst. Angew. Chemie Int. Ed. 55, 11031–11036 (2016).
- Beck, A. et al. Following the structure of copper-zinc-alumina across the pressure gap in carbon dioxide hydrogenation. Nat. Catal. 2021 46 4, 488–497 (2021).
- Prieto, G., Zečević, J., Friedrich, H., De Jong, K. P. & De Jongh, P. E. Towards stable catalysts by controlling collective properties of supported metal nanoparticles. Nat. Mater. 2012 121 12, 34–39 (2012).
- Martin, O. & Pérez-Ramírez, J. New and revisited insights into the promotion of methanol synthesis catalysts by CO2. Catal. Sci. Technol. 3, 3343–3352 (2013).
- Bavykina, A. et al. Turning a Methanation Co Catalyst into an In–Co Methanol Producer. ACS Catal. 9, 6910–6918 (2019).
- Frei, M. S. et al. Nanostructure of nickel-promoted indium oxide catalysts drives selectivity in CO2 hydrogenation. Nat. Commun. 2021 121 12, 1–9 (2021).
- Rungtaweevoranit, B. et al. Copper Nanocrystals Encapsulated in Zr-based Metal-Organic Frameworks for Highly Selective CO2 Hydrogenation to Methanol. Nano Lett. 16, 7645–7649 (2016).
- Jiang, H. L. et al. Au@ZIF-8: CO oxidation over gold nanoparticles deposited to metal-organic framework. J. Am. Chem. Soc. 131, 11302–11303 (2009).
- Jiang, H. L., Akita, T., Ishida, T., Haruta, M. & Xu, Q. Synergistic catalysis of Au@Ag core-shell nanoparticles stabilized on metal-organic framework. J. Am. Chem. Soc. 133, 1304–1306 (2011).
- Zhang, J. et al. Neighboring Zn-Zr Sites in a Metal-Organic Framework for CO2Hydrogenation. J. Am. Chem. Soc. 143, 8829–8837 (2021).
- An, B. et al. Confinement of Ultrasmall Cu/ZnOx Nanoparticles in Metal-Organic Frameworks for Selective Methanol Synthesis from Catalytic Hydrogenation of CO2. J. Am. Chem. Soc. 139, 3834–3840 (2017).
- Sholl, D. S. & Lively, R. P. Defects in Metal–Organic Frameworks: Challenge or Opportunity? J. Phys. Chem. Lett. 6, 3437–3444 (2015).
- Zhang, C., Han, C., Sholl, D. S. & Schmidt, J. R. Computational Characterization of Defects in Metal-Organic Frameworks: Spontaneous and Water-Induced Point Defects in ZIF-8. J. Phys. Chem. Lett. 7, 459–464 (2016).
- Ramos-Fernandez, E. V., Redondo-Murcia, A., Grau-Atienza, A., Sepúlveda-Escribano, A. & Narciso, J. Clean production of Zeolitic Imidazolate Framework 8 using Zamak residues as metal precursor and substrate. J. Clean. Prod. 260, 121081 (2020).
- Thommes, M. et al. Physisorption of gases, with special reference to the evaluation of surface area and pore size distribution (IUPAC Technical Report). Pure Appl. Chem. 87, 1051–1069 (2015).
- Ta, D. N. et al. Preparation of nano-ZIF-8 in methanol with high yield. Can. J. Chem. Eng. 96, 1518–1531 (2018).
- Nordin, N. A. H. M., Ismail, A. F., Mustafa, A., Murali, R. S. & Matsuura, T. The impact of ZIF-8 particle size and heat treatment on CO2/CH4 separation using asymmetric mixed matrix membrane. RSC Adv. 4, 52530–52541 (2014).
- Jamil, N. et al. Green one-pot synthesis and characterisation of hybrid reduced graphene oxide/zeolitic imidazole framework-8 (rGO/ZIF-8). J. Iran. Chem. Soc. 18, 363–373 (2021).
- Zhu, Y. et al. Copper-zirconia interfaces in UiO-66 enable selective catalytic hydrogenation of CO2 to methanol. Nat. Commun. 2020 111 11, 1–11 (2020).
- Yue, B., Li, Q., Iwai, H., Kako, T. & Ye, J. Hydrogen production using zinc-doped carbon nitride catalyst irradiated with visible light. http://www.tandfonline.com/action/journalInformation?show=aimsScope&journalCode=tsta20#.VmBmuzZFCUk 12, 34401–34408 (2011).
- Zhen, S. Y. et al. Metal–organic framework derived hollow porous CuO–CuCo2O4 dodecahedrons as a cathode catalyst for Li–O2 batteries. RSC Adv. 9, 16288–16295 (2019).
- La Rosa-Toro, A. et al. Preparation and Characterization of Copper-Doped Cobalt Oxide Electrodes. J. Phys. Chem. B 110, 24021–24029 (2006).
- Wang, W., Qu, Z., Song, L. & Fu, Q. CO2 hydrogenation to methanol over Cu/CeO2 and Cu/ZrO2 catalysts: Tuning methanol selectivity via metal-support interaction. J. Energy Chem. 40, 22–30 (2020).
- Liao, F. et al. Morphology-Dependent Interactions of ZnO with Cu Nanoparticles at the Materials’ Interface in Selective Hydrogenation of CO2 to CH3OH. Angew. Chemie Int. Ed. 50, 2162–2165 (2011).
- Li, D., Wang, G., Cheng, L., Wang, C. & Mei, X. Engineering the Self-Assembly Induced Emission of Copper Nanoclusters as 3D Nanomaterials with Mesoporous Sphere Structures by the Crosslinking of Ce3+. ACS Omega 3, 14755–14765 (2018).
- Frei, E. et al. Cu−Zn Alloy Formation as Unfavored State for Efficient Methanol Catalysts. ChemCatChem 12, 4029–4033 (2020).
- Newton, M. A. et al. On the Mechanism Underlying the Direct Conversion of Methane to Methanol by Copper Hosted in Zeolites; Braiding Cu K-Edge XANES and Reactivity Studies. J. Am. Chem. Soc. 140, 10090–10093 (2018).
- Nian, J. N., Chen, S. A., Tsai, C. C. & Teng, H. Structural feature and catalytic performance of Cu species distributed over TiO2 nanotubes. J. Phys. Chem. B 110, 25817–25824 (2006).
- Yuan, L. et al. Dynamic Evolution of Atomically Dispersed Cu Species for CO2 Photoreduction to Solar Fuels. ACS Catal. 9, 4824–4833 (2019).
- Ma, Q., Buchholz, D. B. & Chang, R. P. H. Local structures of copper-doped ZnO films. Phys. Rev. B - Condens. Matter Mater. Phys. 78, 214429 (2008).
- Boada, R. et al. Unraveling the Molecular Details of the ‘gate Opening’ Phenomenon in ZIF-8 with X-ray Absorption Spectroscopy. J. Phys. Chem. C 126, 5935–5943 (2022).
- Dalebout, R. et al. Insight into the Nature of the ZnOx Promoter during Methanol Synthesis. ACS Catal. 12, 6628–6639 (2022).
- Wang, Y. et al. Exploring the ternary interactions in Cu–ZnO–ZrO2 catalysts for efficient CO2 hydrogenation to methanol. Nat. Commun. 2019 101 10, 1–10 (2019).
- Graaf, G. H., Sijtsema, P. J. J. M., Stamhuis, E. J. & Joosten, G. E. H. Chemical equilibria in methanol synthesis. Chem. Eng. Sci. 41, 2883–2890 (1986).
- Hu, B. et al. Cu@ZIF-8 derived inverse ZnO/Cu catalyst with sub-5 nm ZnO for efficient CO2 hydrogenation to methanol. Catal. Sci. Technol. 9, 2673–2681 (2019).
- Beebe, T. P., Crowell, J. E. & Yates, J. T. Reaction of methyl chloride with alumina surfaces: a study of the methoxy surface species by transmission infrared spectroscopy. J. Phys. Chem. 92, 1296–1301 (2002).
- Fujita, S. ichiro, Usui, M., Ohara, E. & Takezawa, N. Methanol synthesis from carbon dioxide at atmospheric pressure over Cu/ZnO catalyst. Role of methoxide species formed on ZnO support. Catal. Lett. 1992 134 13, 349–358 (1992).
Table 1 is not available with this version.
There is NO Competing Interest.
- SuppInfo.docx
Supplementary information