CO2 Hydrogenation to Methanol over Ce and Zr Containing UiO-66 and Cu/UiO-66

Direct hydrogenation of CO2 to methanol is an interesting method to recycle CO2 emitted e.g., during combustion of fossil fuels. However, it is a challenging process because both the selectivity to methanol and its production are low. The metal-organic frameworks are relatively new class of materials with a potential to be used as catalysts or catalysts supports, also in the reaction of MeOH production. Among many interesting structures, the UiO-66 draws significant attention owing to its chemical and thermal stability, developed surface area, and the possibility of tuning its properties e.g., by exchanging the zirconium in the nodes to other metal cations. In this work we discuss—for the first time—the performance of Cu supported on UiO-66(Ce/Zr) in CO2 hydrogenation to MeOH. We show the impact of the composition of UiO-66-based catalysts, and the character of Cu-Zr and Cu-Ce interactions on MeOH production and MeOH selectivity during test carried out for 25 h at T = 200 ◦C and p = 1.8 MPa. Significant increase of selectivity to MeOH was noticed after exchanging half of Zr4+ cations with Ce4+; however, no change in MeOH production occurred. It was found that the Cu-Ce coexistence in the UiO-66-based catalytic system reduced the selectivity to MeOH when compared to Cu/UiO-66(Zr), which was ascribed to lower concentration of Cu0 active sites in Cu/UiO-66(Ce/Zr), and this was caused by oxygen spill-over between Cu0 and Ce4+, and thus, the oxidation of the former. The impact of reaction conditions on the structure stability of tested catalyst was also determined.


Introduction
Methanol is one of the most valuable chemicals. It is used as a solvent, fuel, or a raw material for the synthesis of other chemicals, such as formaldehyde, acetic acid, methyl methacrylate, dimethyl terephthalate, dimethyl ether, chloromethanes, and others. In conventional process MeOH is produced from synthesis gas via hydrogenation of CO (Equation (1)) and CO 2 (Equation (2)) and the reversed water-gas shift (RWGS) reaction (Equation (3)).
CO + 2H 2 = CH 3 OH, ∆H 298 K = − 90.1 kJ/mol (1) better intermediates for methanol formation. Therefore, the addition of CeO x NPs to Cu is plausible because it leads to the formation of a metal-oxide interface that enables adsorption and activation of CO 2 . The reduction of ceria produces Ce 3+ centres that are active sites for CO 2 adsorption, which is later activated and hydrogenated to yield a carboxyl species (-COOH) that are next decomposed to CO and -OH. In a subsequent step, CO is hydrogenated to MeOH through the following intermediates: formyl (-CHO), formaldehyde (H 2 CO), and methoxy (-OCH 3 ). In this regard, until CO formation, the synthesis of methanol is realized according to the elementary steps of the RWGS. Allam et al. [17] studied the CuO-ZnO-CeO 2 system that revealed 20% CO 2 conversion and 65% selectivity to methanol at 240 • C and ambient pressure. MeOH synthesis via CO 2 hydrogenation is structure sensitive. The composition and metal-support interactions play a crucial role in catalyst performance. Most Cu-Zn-containing catalysts possess low surface area (about 30 m 2 /g), which limits their catalytic function in the reaction. Owing to their developed surface area, surface properties (such as pore size, hydrophobicity/hydrophilicity, Lewis or Brønsted acidity/basicity), and remarkable synergism between their constituents metal-organic frameworks, MOFs can overcome some of the flaws of the conventional catalysts for MeOH synthesis. Improved catalytic performance of MOFs can be achieved by doping them with metal nanoparticles. Electronic interactions between metal and MOF in such a hybrid catalysts are significant for their catalytic activity or selectivity.
Rungtaweevoranit et al. [9] studied copper nanocrystals-encapsulated UiO-66, MIL-101, and ZIF-8 in MeOH synthesis via CO 2 hydrogenation at 175 • C and 10 bar. They found out that only Zr containing UiO-66-based catalyst was active. Moreover, they observed that the location of Cu influences the catalytic activity. The Cu encapsulated in UiO-66 revealed two-fold higher activity and enhanced stability than Cu deposited on UiO-66. Better performance of the former was probably due to enhanced metal-support interaction that allowed formation of active Cu sites for CO 2 conversion to MeOH.
An et al. [18] produced in situ ultra-small Cu/ZnO x nanoparticles in the UiO-bpy cavities under CO 2 /H 2 . The agglomeration of Cu nanoparticles and phase separation between Cu and ZnOx was prevented because of Zr 6 SBUs, strong Cu-ZnO x interaction, and the bpy moieties on the 2,2'-bipyridine-5,5'-dicarboxylate ligands. After 30 min of CO 2 hydrogenation at 250 • C and p = 4 MPa, the catalyst showed 7.2% conversion of CO 2 with 100% selectivity to MeOH and production of 1.97 g MeOH /kg Cu /h.
Recently, Kobayashi et al. [19] reported the Cu supported on Zr-UiO-66, Zr-UiO-66-NH 2 , Zr-UiO-66-COOH, and Hf-UiO-66 (with a Cu loading ranging from 13 to 19 wt.%) as hybrid catalysts for CO 2 hydrogenation to methanol. They observed that -COOH group in the BDC linker as well as the replacement of Zr 4+ with Hf 4+ had a beneficial effect on MeOH production, which was higher than that of commercially used Cu/ZnO/Al 2 O 3 . The MeOH production over Zr-UiO-66-COOH and Hf-UiO-66 was ca. 380 µmol/g Cu /h. In addition, the activity of studied MOF-based catalysts was not associated to the amount of defects in the Zr 6 -clusters.
The zirconium containing UiO-66 was first reported by Cavka et al. [20] and owing to its thermal and chemical stability as well as easy functionalization [21][22][23], it attracted attention of many other researchers. The stability of UiO-66 structure explains its wide use in heterogeneous catalysis. It was studied, e.g., in transesterification of triglyceride with methanol [24], sulphide oxidation [25], desulfurization [26], and other reactions [27]. A number of papers concerning catalytic performance of the UiO-66 modified in Zr nodes or the BDC linker, or decorated with metals, was published. For example, the Cu(II)-and Cu 0 -loaded UiO-66-NH 2 hybrids were found to be highly efficient heterogeneous catalysts for oxidation of cyclohexene and hydrogenation of styrene under mild conditions [28].
In this work we discuss the features and the catalytic properties of Cu supported on UiO-66(Zr) and UiO-66(Ce/Zr). Our intention was to exchange half of Zr 4+ ions with Ce 4+ in the UiO-66 structure and combine it with copper to examine-for the first time-the performance of such a system in CO 2 hydrogenation to MeOH.

Characterization of UiO-66
Textural properties of UiO-66 and Cu/UiO-66 materials were determined using the N 2 sorption experiments. Figure 1 and Table 1 present the N 2 adsorption-desorption isotherms, pore size distribution, and textural parameters of the obtained UiO-66 and Cu-UiO-66 materials. The courses of N 2 uptake vs. P/P 0 (Figure 1a) indicate type I isotherms which is characteristic for microporous materials. As is shown in Figure 1b, the majority of pores in obtained materials are smaller than 20 Å. From Table 1 it can be seen that the mean pore width ranges from 23.1 to 23.8 Å. The surface area (S BET ) of UiO-66(Zr) is 1380 m 2 /g, which is higher than that reported in [20][21][22][23] and similar to [19]. The decrease in S BET to 810 m 2 /g is observed when part of Zr cations is replaced with cerium. Further S BET decrease occurs after UiO-66 impregnation with copper nitrate. The decrease in S BET of Zr-UiO-66, Hf-UiO-66, Zr-UiO-66-NH 2 , and Zr-UiO-66-COOH after impregnation with Cu(OAc) 2 was also reported [7,19].

Characterization of UiO-66
Textural properties of UiO-66 and Cu/UiO-66 materials were determined using the N2 sorption experiments. Figure 1 and Table 1 present the N2 adsorption-desorption isotherms, pore size distribution, and textural parameters of the obtained UiO-66 and Cu-UiO-66 materials. The courses of N2 uptake vs. P/P0 (Figure 1a) indicate type I isotherms which is characteristic for microporous materials. As is shown in Figure 1b, the majority of pores in obtained materials are smaller than 20 Å. From Table 1 it can be seen that the mean pore width ranges from 23.1 to 23.8 Å. The surface area (SBET) of UiO-66(Zr) is 1380 m 2 /g, which is higher than that reported in [20][21][22][23] and similar to [19]. The decrease in SBET to 810 m 2 /g is observed when part of Zr cations is replaced with cerium. Further SBET decrease occurs after UiO-66 impregnation with copper nitrate. The decrease in SBET of Zr-UiO-66, Hf-UiO-66, Zr-UiO-66-NH2, and Zr-UiO-66-COOH after impregnation with Cu(OAc)2 was also reported [7,19].  The crystallographic structures of obtained materials were determined by powder X-ray diffraction (XRD). The XRD patterns of UiO-66 and Cu/UiO-66 materials are presented in Figure 2. It can be observed that all of synthesized materials are crystalline and show sharp reflexes at 2⍬ of ca. 7.38, 8.52, 12.06, 14.15, and 14.78°, that correspond to reflections for the (111), (200), (220), (311), and (222) planes, respectively. Those reflexes are typical for the UiO-66(Zr) structure [29]. A 1° shift towards lower 2θ is observed for the bimetallic UiO-66(Ce/Zr) which is due to partial replacement of smaller-radii Zr 4+ (0.84 Å) with a bigger-radii Ce 4+ (0.97 Å) [30]   The crystallographic structures of obtained materials were determined by powder X-ray diffraction (XRD). The XRD patterns of UiO-66 and Cu/UiO-66 materials are presented in Figure 2. It can be observed that all of synthesized materials are crystalline and show sharp reflexes at 2θ of ca. 7. 38, 8.52, 12.06, 14.15, and 14.78 • , that correspond to reflections for the (111), (200), (220), (311), and (222) planes, respectively. Those reflexes are typical for the UiO-66(Zr) structure [29]. A 1 • shift towards lower 2θ is observed for the bimetallic UiO-66(Ce/Zr) which is due to partial replacement of smaller-radii Zr 4+ (0.84 Å) with a bigger-radii Ce 4+ (0.97 Å) [30] UiO-66(Zr) and UiO-66(Ce/Zr) ( Figure 2) was retained after impregnation with copper nitrate. No reflexes coming from Cu phase (Cu, Cu2O, or CuO) are observed for Cu/UiO-66(Zr) and Cu/UiO-66(Ce/Zr), which can be due to relatively low Cu loading and its good dispersion over the framework. The X-ray photoelectron spectroscopy (XPS) proved that in Cu/UiO-66(Zr) and Cu/UiO-66(Ce/Zr) samples copper was present both in the form of CuO and Cu2O. The Cu2p XPS spectra presented in Figure 3a,c display the Cu 2p3/2 and Cu 2p1/2 peaks of Cu 2+ at ca. 934 and 954 eV, respectively, with a strong shake-up satellite at ca. 944 eV [31]. Whereas peaks at ca. 932 and 952 eV correspond to Cu + . The Ce3d spectra for UiO-66(Ce/Zr) and Cu/UiO-66(Ce/Zr) (Figure 3b,  The X-ray photoelectron spectroscopy (XPS) proved that in Cu/UiO-66(Zr) and Cu/UiO-66(Ce/Zr) samples copper was present both in the form of CuO and Cu 2 O. The Cu2p XPS spectra presented in Figure 3a,c display the Cu 2p 3/2 and Cu 2p 1/2 peaks of Cu 2+ at ca. 934 and 954 eV, respectively, with a strong shake-up satellite at ca. 944 eV [31]. Whereas peaks at ca. 932 and 952 eV correspond to Cu + . The Ce3d spectra for UiO-66(Ce/Zr) and Cu/UiO-66(Ce/Zr) (Figure 3b,  Determination of the morphology of the obtained materials and distribution of particular metals within and over the UiO-66 framework was performed with scanning electron microscope (SEM) and high-resolution transmission electron microscope (HRTEM). The elemental composition of the obtained in this work UiO-66(Ce/Zr) is presented in Table 2, displaying lower than assumed concentration of Ce. It proves that less than 50% of Zr 4+ was replaced with Ce 4+ . It is important to notice that mapping of the UiO-66(Ce/Zr) ( Figure 4) showed uniform dispersion of Ce and Zr in the sample.  Determination of the morphology of the obtained materials and distribution of particular metals within and over the UiO-66 framework was performed with scanning electron microscope (SEM) and high-resolution transmission electron microscope (HRTEM). The elemental composition of the obtained in this work UiO-66(Ce/Zr) is presented in Table 2, displaying lower than assumed concentration of Ce. It proves that less than 50% of Zr 4+ was replaced with Ce 4+ . It is important to notice that mapping of the UiO-66(Ce/Zr) ( Figure 4) showed uniform dispersion of Ce and Zr in the sample.  Determination of the morphology of the obtained materials and distribution of particular metals within and over the UiO-66 framework was performed with scanning electron microscope (SEM) and high-resolution transmission electron microscope (HRTEM). The elemental composition of the obtained in this work UiO-66(Ce/Zr) is presented in Table 2, displaying lower than assumed concentration of Ce. It proves that less than 50% of Zr 4+ was replaced with Ce 4+ . It is important to notice that mapping of the UiO-66(Ce/Zr) ( Figure 4) showed uniform dispersion of Ce and Zr in the sample.   The SEM and high-angle annular dark-field imaging (HAADF-STEM) of Cu/UiO-66 materials presented in Figure 5 show that both samples are homogeneous and are composed of small UiO-66 particles that overlap each other. The energy dispersive X-ray spectroscopy (EDS) detected 3.47 and 12.09 wt.% of Cu in Cu/UiO-66(Zr) and Cu/UiO-66(Ce/Zr), respectively. This discrepancy is most probably caused by the inhomogeneous dispersion of copper over UiO-66 particles. Nevertheless, as was also determined with EDS mapping (Figure 6), the Cu dispersion in scanned areas in both samples was very good.  The SEM and high-angle annular dark-field imaging (HAADF-STEM) of Cu/UiO-66 materials presented in Figure 5 show that both samples are homogeneous and are composed of small UiO-66 particles that overlap each other. The energy dispersive X-ray spectroscopy (EDS) detected 3.47 and 12.09 wt.% of Cu in Cu/UiO-66(Zr) and Cu/UiO-66(Ce/Zr), respectively. This discrepancy is most probably caused by the inhomogeneous dispersion of copper over UiO-66 particles. Nevertheless, as was also determined with EDS mapping (Figure 6), the Cu dispersion in scanned areas in both samples was very good.    The SEM and high-angle annular dark-field imaging (HAADF-STEM) of Cu/UiO-66 materials presented in Figure 5 show that both samples are homogeneous and are composed of small UiO-66 particles that overlap each other. The energy dispersive X-ray spectroscopy (EDS) detected 3.47 and 12.09 wt.% of Cu in Cu/UiO-66(Zr) and Cu/UiO-66(Ce/Zr), respectively. This discrepancy is most probably caused by the inhomogeneous dispersion of copper over UiO-66 particles. Nevertheless, as was also determined with EDS mapping (Figure 6), the Cu dispersion in scanned areas in both samples was very good.   The Fourier transform infrared spectroscopy (FT-IR) was used for observations of the surfaces of the obtained UiO-66 and Cu/UiO-66 materials. The FT-IR spectra (Figure 7) of all samples display bands at the same wavenumber. The presence of vibration mode at 664 and 473 cm −1 is due to µ 3 -O and µ 3 -OH stretching vibrations of the Zr 6 O 4 (OH) 4 (CO 2 ) 12 in the UiO-66 framework. Whereas the band at 551 cm −1 is ascribed to Zr-(OC) asymmetric stretching vibrations. The small band visible at 1507 cm −1 in all samples can be associated with the vibration of the C=C in benzene ring in the deprotonated, and thus coordinated, BDC linker. Bands indicating the occurrence of asymmetric and symmetric ν(OCO) stretchings in the linker are visible at 1578 and 1390 cm −1 , respectively. At low frequencies the modes due to OH and CH bending in H 2 BDC are mixed with Zr-O modes. Bands at ca. 812, 743, and 714 cm −1 correspond to the mix of C-H vibration, C=C stretch, OH bend, and OCO bend in terephthalic acid [33]. The presence of the latter, together with the bands at 1507, 1017 cm −1 , can be due to BDC coordination with Zr [34] or Ce nodes. Whereas the presence of free -COOH groups in partially uncoordinated linker is proven by the occurrence of the band at 1673 cm −1 . However, that band is not present on the spectra, which implies that obtained in this work UiO-66-based structures are terminated on the surface with Zr and/or Ce clusters. The existence of residual dimethylformamide (DMF) within the pores of obtained UiO-66 and Cu/UiO-66 is normally displayed by the band at 1658 cm −1 coming from the C=O asymmetric stretching vibrations. That band is hardly observable on the FT-IR spectra of obtained materials, which indicates successful exchange of solvent after the synthesis of UiO-66(Zr) and UiO-66(Ce/Zr).  [33]. The presence of the latter, together with the bands at 1507, 1017 cm −1 , can be due to BDC coordination with Zr [34] or Ce nodes. Whereas the presence of free -COOH groups in partially uncoordinated linker is proven by the occurrence of the band at 1673 cm −1 . However, that band is not present on the spectra, which implies that obtained in this work UiO-66based structures are terminated on the surface with Zr and/or Ce clusters. The existence of residual dimethylformamide (DMF) within the pores of obtained UiO-66 and Cu/UiO-66 is normally displayed by the band at 1658 cm −1 coming from the C=O asymmetric stretching vibrations. That band is hardly observable on the FT-IR spectra of obtained materials, which indicates successful exchange of solvent after the synthesis of UiO-66(Zr) and UiO-66(Ce/Zr). Thermal stabilities and the average number of linker defects per one [Zr6O4(OH)4] 12+ cluster in the obtained materials were studied using thermogravimetric analysis (TGA). The low temperature mass decrease with a maximal rate at ca. 100 °C (Figure 8) is as a result of desorption of H2O from the sample (the FTIR showed that the DMF was successfully removed during solvent exchange step). The H2O content in the samples, calculated from TGA, increased in the following order: UiO-66(Ce/Zr) < Cu/UiO-66(Ce/Zr) < UiO-66(Zr) < Cu/UiO-66(Zr), being 16.13, 18.21, 19.98, and 23.65%, respectively. Hence, the amount of water was lower for the samples where 50% of Zr 4+ was replaced with Ce 4+ , and it slightly increased after impregnation with copper nitrate. Continuous but less significant mass decrease from ca. 120 °C can be due to framework dehydroxylation, i.e., removal of the -OH groups in the form of H2O from the Zr (or Ce) clusters. Simultaneously, solvent molecules (H2O and/or DMF-if present) that coordinated with metal clusters to compensate linker deficiencies are removed. The UiO-66(Zr) and UiO-66(Ce/Zr) are thermally stable up to 490 and 450 °C, respectively. Lower thermal resistance of UiO-66(Ce/Zr) compared to UiO-66(Zr) arises from partial exchange of Zr to Ce. The beneficial influence of zirconium on thermal stability of cerium oxide is generally known [35]. As was reported by Lammert et al. [36], increasing amount of Ce in the UiO-66(Ce/Zr) negatively influenced thermal stability of the material, which decreased from 450 °C for UiO-66(Zr) to 220 °C at 40 At.% content of Ce. However, in the present work, the stability of UiO-66(Ce/Zr) with a theoretical Ce/Zr atomic ratio of 1:1 is lower only by 50 °C compared to UiO-66(Zr). As was reported in [36,37] the UiO-66(Ce) was stable only up to 220−240 °C. Also our studies revealed that complete exchange of Zr for Ce in the UiO-66 shifted the temperature of BDC decomposition to Thermal stabilities and the average number of linker defects per one [Zr 6 O 4 (OH) 4 ] 12+ cluster in the obtained materials were studied using thermogravimetric analysis (TGA). The low temperature mass decrease with a maximal rate at ca. 100 • C (Figure 8) is as a result of desorption of H 2 O from the sample (the FTIR showed that the DMF was successfully removed during solvent exchange step). The H 2 O content in the samples, calculated from TGA, increased in the following order: UiO-66(Ce/Zr) < Cu/UiO-66(Ce/Zr) < UiO-66(Zr) < Cu/UiO-66(Zr), being 16.13, 18.21, 19.98, and 23.65%, respectively. Hence, the amount of water was lower for the samples where 50% of Zr 4+ was replaced with Ce 4+ , and it slightly increased after impregnation with copper nitrate. Continuous but less significant mass decrease from ca. 120 • C can be due to framework dehydroxylation, i.e., removal of the -OH groups in the form of H 2 O from the Zr (or Ce) clusters. Simultaneously, solvent molecules (H 2 O and/or DMF-if present) that coordinated with metal clusters to compensate linker deficiencies are removed. The UiO-66(Zr) and UiO-66(Ce/Zr) are thermally stable up to 490 and 450 • C, respectively. Lower thermal resistance of UiO-66(Ce/Zr) compared to UiO-66(Zr) arises from partial exchange of Zr to Ce. The beneficial influence of zirconium on thermal stability of cerium oxide is generally known [35]. As was reported by Lammert et al. [36], increasing amount of Ce in the UiO-66(Ce/Zr) negatively influenced thermal stability of the material, which decreased from 450 • C for UiO-66(Zr) to 220 • C at 40 At.% content of Ce. However, in the present work, the stability of UiO-66(Ce/Zr) with a theoretical Ce/Zr atomic ratio of 1:1 is lower only by 50 • C compared to UiO-66(Zr). As was reported in [36,37] the UiO-66(Ce) was stable only up to 220−240 • C. Also our studies revealed that complete exchange of Zr for Ce in the UiO-66 shifted the temperature of BDC decomposition to ca. 260−280 • C (depending on the synthesis method) which was followed with the collapse of MOF structure [38]. Compared to UiO-66(Zr) and UiO-66(Ce/Zr), a significant decrease in thermal stability, by ca. 150 • C, is observed for their copper-containing analogues. The presence of well dispersed Cu nanoparticles facilitates thermal decomposition of the framework at high temperature by weakening the interaction between -COO − and Zr 4+ or Ce 4+ .
Catalysts 2019, 9, x FOR PEER REVIEW 9 of 17 ca. 260−280 °C (depending on the synthesis method) which was followed with the collapse of MOF structure [38]. Compared to UiO-66(Zr) and UiO-66(Ce/Zr), a significant decrease in thermal stability, by ca. 150 °C, is observed for their copper-containing analogues. The presence of well dispersed Cu nanoparticles facilitates thermal decomposition of the framework at high temperature by weakening the interaction between -COO − and Zr 4+ or Ce 4+ . On the basis of TGA we have calculated the average number of linker molecules coordinated to one metal cluster. In an ideal UiO-66 structure to each metal cluster coordinates 12 linker molecules. It can be seen from Table 3 that the amount of linker defects was the highest for UiO-66(Zr) and decreased when half of Zr 4+ was replaced with Ce 4+ (i.e., for UiO-66(Ce/Zr)) or when the material was impregnated with copper. Defects in MOFs, either in the linkers or in the metal nodes, are catalytically active sites so their presence is somewhat desired. It is known that the UiO-66 structure remains stable up to 4.3 missing linker molecules per one [Zr6O4(OH)4] 12+ node [39]. Above that number the structure collapses.

Catalytic Tests of CO2 Hydrogenation to MeOH
The results of the catalytic tests of CO2 hydrogenation to MeOH over the UiO-66(Zr), UiO-66(Ce/Zr), Cu/UiO-66(Zr), and Cu/UiO-66(Ce/Zr) are presented in Figure 9. The MeOH production was changing during first few hours of test run for all catalysts. Selectivity to MeOH over particular On the basis of TGA we have calculated the average number of linker molecules coordinated to one metal cluster. In an ideal UiO-66 structure to each metal cluster coordinates 12 linker molecules. It can be seen from Table 3 that the amount of linker defects was the highest for UiO-66(Zr) and decreased when half of Zr 4+ was replaced with Ce 4+ (i.e., for UiO-66(Ce/Zr)) or when the material was impregnated with copper. Defects in MOFs, either in the linkers or in the metal nodes, are catalytically active sites so their presence is somewhat desired. It is known that the UiO-66 structure remains stable up to 4.3 missing linker molecules per one [Zr 6 O 4 (OH) 4 ] 12+ node [39]. Above that number the structure collapses.

Catalytic Tests of CO 2 Hydrogenation to MeOH
The results of the catalytic tests of CO 2 hydrogenation to MeOH over the UiO-66(Zr), UiO-66(Ce/Zr), Cu/UiO-66(Zr), and Cu/UiO-66(Ce/Zr) are presented in Figure 9. The MeOH production was changing during first few hours of test run for all catalysts. Selectivity to MeOH over particular catalysts was stable from the third hour of test run (Figure 9b). The gas chromatography (GC) analyses performed during the tests revealed that the only organic by-product of the reaction was dimethyl ether.
According to [18], the adsorption and activation of CO 2 and H 2 takes place on the unsaturated Zr and Cu sites, respectively. But in our conditions the CO 2 activation occurs on the Zr and Ce (if applicable) sites. The poorest selectivity to MeOH, of only 3.5%, was noticed for UiO-66(Zr), but the introduction of cerium to the framework caused significant improvement in that field-selectivity increased to 28.7%. However, the introduction of Ce to the UiO-66 does not influence the rate of MeOH synthesis, which was found to be 278 µmol MeOH /g Cu /h for UiO-66(Zr) and 275 µmol MeOH /g Cu /h for UiO-66(Ce/Zr). The best performance in catalytic CO 2 hydrogenation showed the Cu-loaded materials. The methanol STY after 25 h of test run was 468 µmol MeOH /g Cu /h for Cu/UiO-66(Ce/Zr) and 397 µmol MeOH /g Cu /h for Cu/UiO-66(Ce/Zr), whereas the selectivity to MeOH reached 59 and 56%, respectively ( Figure 10). Our results are much better than those reported by Kobayashi et al. [19] who obtained 67.8 and 114.2 µmol MeOH /g Cu /h over UiO-66(Zr) and Cu/UiO-66(Zr), respectively-over 4 times less than for corresponding catalysts discussed in this work. Nevertheless, obtained in their work Cu/UiO-66(Zr) revealed over 90% selectivity to MeOH, which is much better than that compared to our work. It must be also mentioned here, that they performed the catalytic test at higher T and lower pressure than we did (T = 220 • C, p = 2 Atm., H 2 /CO 2 of 5:1). Those condition could have an impact on the catalysts performance since the reaction favours lower temperature and higher pressure. analyses performed during the tests revealed that the only organic by-product of the reaction was dimethyl ether. According to [18], the adsorption and activation of CO2 and H2 takes place on the unsaturated Zr and Cu sites, respectively. But in our conditions the CO2 activation occurs on the Zr and Ce (if applicable) sites. The poorest selectivity to MeOH, of only 3.5%, was noticed for UiO-66(Zr), but the introduction of cerium to the framework caused significant improvement in that field-selectivity increased to 28.7%. However, the introduction of Ce to the UiO-66 does not influence the rate of MeOH synthesis, which was found to be 278 µmolMeOH/gCu/h for UiO-66(Zr) and 275 µmolMeOH/gCu/h for UiO-66(Ce/Zr). The best performance in catalytic CO2 hydrogenation showed the Cu-loaded materials. The methanol STY after 25 h of test run was 468 µmolMeOH/gCu/h for Cu/UiO-66(Ce/Zr) and 397 µmolMeOH/gCu/h for Cu/UiO-66(Ce/Zr), whereas the selectivity to MeOH reached 59 and 56%, respectively ( Figure 10). Our results are much better than those reported by Kobayashi et al. [19] who obtained 67.8 and 114.2 µmolMeOH/gCu/h over UiO-66(Zr) and Cu/UiO-66(Zr), respectively-over 4 times less than for corresponding catalysts discussed in this work. Nevertheless, obtained in their work Cu/UiO-66(Zr) revealed over 90% selectivity to MeOH, which is much better than that compared to our work. It must be also mentioned here, that they performed the catalytic test at higher T and lower pressure than we did (T = 220 °C, p = 2 Atm., H2/CO2 of 5:1). Those condition could have an impact on the catalysts performance since the reaction favours lower temperature and higher pressure.  We have anticipated that Ce introduction to the Cu/UiO-66(Zr) would improve its performance, but unlike it was expected, the MeOH production over Cu/UiO-66(Ce/Zr) was lower than for analyses performed during the tests revealed that the only organic by-product of the reaction was dimethyl ether. According to [18], the adsorption and activation of CO2 and H2 takes place on the unsaturated Zr and Cu sites, respectively. But in our conditions the CO2 activation occurs on the Zr and Ce (if applicable) sites. The poorest selectivity to MeOH, of only 3.5%, was noticed for UiO-66(Zr), but the introduction of cerium to the framework caused significant improvement in that field-selectivity increased to 28.7%. However, the introduction of Ce to the UiO-66 does not influence the rate of MeOH synthesis, which was found to be 278 µmolMeOH/gCu/h for UiO-66(Zr) and 275 µmolMeOH/gCu/h for UiO-66(Ce/Zr). The best performance in catalytic CO2 hydrogenation showed the Cu-loaded materials. The methanol STY after 25 h of test run was 468 µmolMeOH/gCu/h for Cu/UiO-66(Ce/Zr) and 397 µmolMeOH/gCu/h for Cu/UiO-66(Ce/Zr), whereas the selectivity to MeOH reached 59 and 56%, respectively ( Figure 10). Our results are much better than those reported by Kobayashi et al. [19] who obtained 67.8 and 114.2 µmolMeOH/gCu/h over UiO-66(Zr) and Cu/UiO-66(Zr), respectively-over 4 times less than for corresponding catalysts discussed in this work. Nevertheless, obtained in their work Cu/UiO-66(Zr) revealed over 90% selectivity to MeOH, which is much better than that compared to our work. It must be also mentioned here, that they performed the catalytic test at higher T and lower pressure than we did (T = 220 °C, p = 2 Atm., H2/CO2 of 5:1). Those condition could have an impact on the catalysts performance since the reaction favours lower temperature and higher pressure.  We have anticipated that Ce introduction to the Cu/UiO-66(Zr) would improve its performance, but unlike it was expected, the MeOH production over Cu/UiO-66(Ce/Zr) was lower than for We have anticipated that Ce introduction to the Cu/UiO-66(Zr) would improve its performance, but unlike it was expected, the MeOH production over Cu/UiO-66(Ce/Zr) was lower than for Cu/UiO-66(Zr) (Figure 10). The Cu-containing catalysts were pre-reduced in H 2 /Ar prior to the tests to provide Cu 0 active sites for H 2 dissociation, but CuO reduction might have not been completed and except for Cu 0 , the cationic copper (e.g. Cu 2 O) could have been also present on the catalyst surface [38]. It is known that the combination of Cu species plays important role in the catalyst performance: the Cu 0 is responsible for H 2 dissociation, whereas Cu + stabilizes the formate intermediates. We expect that the decrease in MeOH production in the case of Cu/UiO-66(Ce/Zr) (when compared to Cu/UiO-66(Zr)), can be due to strong interaction between Cu 0 and Ce 4+ cluster that facilitates the oxygen spill-over from Ce 4+ to Cu 0 , destabilizing (i.e., oxidizing) the active site for H 2 adsorption and dissociation. It was found by the HAADF-STEM with the EDS elemental mapping, that copper was well-dispersed in Cu/UiO-66(Zr) and Cu/UiO-66(Ce/Zr) samples. It is known, that the formation of nanostructured copper-ceria composites characterizes with the strong electronic interactions between nanoparticles [40], and the transition between Ce 3+ and Ce 4+ takes place by the equilibration of mobile oxygen vacancies at the Cu-Ce interface [41]. In the presented case, that strong Cu-Ce interaction can be also proven by the significant decrease in thermal stability of Cu/UiO-66(Ce/Zr) compared to UiO-66(Ce/Zr) (Figure 8). Such a strong Cu 0 and Ce 4+ interaction can significantly inhibit the MeOH formation because H 2 dissociation is required for hydrogenation of the surface formate, which is the rate determining step. From the other hand, zirconium is non-reducible and stabilizes Cu 0 active sites. We may assume that Cu/UiO-66(Zr) possess more Cu 0 active sites than Cu/UiO-66(Ce/Zr).
Also, different character of Cu-Zr and Cu-Ce interactions explains the difference in MeOH selectivity on Cu/UiO-66(Zr) and Cu/UiO-66(Ce/Zr). On those catalysts the reaction takes place at the Cu-Zr 4+ and the Cu-Zr 4+ / Cu-Ce 4+ interfaces. But the mechanism of CO 2 hydrogenation to MeOH is different for Zr-and Ce-containing systems, which results in different product selectivity. It was found by Wang et al. [42] that CO 2 adsorption on Cu-ZrO 2 promotes formation of bicarbonate species that are later hydrogenated to hydrocarboxyl species (COOH), and those readily decompose to CO, which decreases MeOH selectivity. Whereas the Cu-CeO 2 assists formation of surface carbonates that are first hydrogenated to formate species (HOOC) and further hydrogenated to MeOH. However, our results show that the presence of cerium in Cu/UiO-66(Ce/Zr) does not improve MeOH selectivity compared to Cu/UiO-66(Zr).
Considering the effect of the concentration of structure defects in UiO-66 on its performance in MeOH synthesis via CO 2 hydrogenation, it was noticed (likewise reported in [19]) that there is no correlation between those two parameters. Rungtaweevoranit et al. [9] observed that the most important is Cu contact with the support. By encapsulation of copper inside the UiO-66 framework they increased the MeOH production, when compared to Cu deposited over UiO-66. It was explained with higher number of active Cu sites surrounded by the Zr oxide SBU in the case of metal encapsulation, and thus, strong metal-support interaction.

Characterization After Catalytic Tests
The XRD analyses of spent catalyst ( Figure 11) exhibited that their crystal structures changed after reaction carried out under increased pressure (1.8 MPa) and temperature (200 • C) for 25 h. It can be seen that cerium insertion into the UiO-66 framework stabilizes it to some extent (vide two characteristic reflexions at 2θ = 7.38 and 8.52 • on the diffractograms of UiO-66(Ce/Zr) and Cu/UiO-66(Ce/Zr), corresponding to (111) and (200) planes in UiO-66). The diffractograms of UiO-66(Zr) and UiO-66(Ce/Zr) also display the symmetry forbidden (110) reflection at 2θ = 6.57 • that indicates the presence of the reo unit cell, which is equivalent to missing an entire Zr 6 O 6 (BDC) 12 12− SBU; thus, the existence of open Zr centers. In such a case, the unit cell has a 12+ net positive charge that has to be compensated by negatively charged ligands [43]. Cliffe et al. [44]  of Cu in the sample increased. The changes in the samples structures after catalytic tests influenced their textural properties, i.e., the SBET of all samples decreased and the mean pore size increased ( Table  4). The most significant deterioration of textural properties can be observed for UiO-66(Zr).   The FT-IR spectra of spent catalyst are presented in Figure 12. The decrease of intensities of bands at 664, 473, and 551 cm −1 corresponding to µ3-O, µ3-OH, and Zr-(OC) in the Zr6O4(OH)4(-CO2)12 SBU is observed. It implies some degradation of the [Zr6O4(OH)4] clusters caused by the extended reaction at elevated pressure and temperature. It can be associated to the formation of the whole cluster defects and the reo unit cell that was previously detected with the XRD (Figure 11). The existence of coordinated BDC linker in the UiO-66 is proven by the bands at 1507 and 744 cm −1 associated with the vibration of the C=C in benzene ring , and bands at 1550−1650 cm −1 and 1335-1446 cm −1 regions corresponding to asymmetric and symmetric ν(OCO)as stretching vibrations.   The FT-IR spectra of spent catalyst are presented in Figure 12. The decrease of intensities of bands at 664, 473, and 551 cm −1 corresponding to µ 3 -O, µ 3 -OH, and Zr-(OC) in the Zr 6 O 4 (OH) 4 (-CO 2 ) 12 SBU is observed. It implies some degradation of the [Zr 6 O 4 (OH) 4 ] clusters caused by the extended reaction at elevated pressure and temperature. It can be associated to the formation of the whole cluster defects and the reo unit cell that was previously detected with the XRD (Figure 11). The existence of coordinated BDC linker in the UiO-66 is proven by the bands at 1507 and 744 cm −1 associated with the vibration of the C=C in benzene ring, and bands at 1550−1650 cm −1 and 1335-1446 cm −1 regions corresponding to asymmetric and symmetric ν(OCO) as stretching vibrations. of Cu in the sample increased. The changes in the samples structures after catalytic tests influenced their textural properties, i.e., the SBET of all samples decreased and the mean pore size increased ( Table  4). The most significant deterioration of textural properties can be observed for UiO-66(Zr).  The FT-IR spectra of spent catalyst are presented in Figure 12. The decrease of intensities of bands at 664, 473, and 551 cm −1 corresponding to µ3-O, µ3-OH, and Zr-(OC) in the Zr6O4(OH)4(-CO2)12 SBU is observed. It implies some degradation of the [Zr6O4(OH)4] clusters caused by the extended reaction at elevated pressure and temperature. It can be associated to the formation of the whole cluster defects and the reo unit cell that was previously detected with the XRD (Figure 11). The existence of coordinated BDC linker in the UiO-66 is proven by the bands at 1507 and 744 cm −1 associated with the vibration of the C=C in benzene ring , and bands at 1550−1650 cm −1 and 1335-1446 cm −1 regions corresponding to asymmetric and symmetric ν(OCO)as stretching vibrations.

Synthesis of UiO-66 and UiO-66(Ce/Zr)
The UiO-66(Zr) was synthesized according to the procedure described by Katz et al. [45]. In the first flask, 1 g of zirconium (IV) chloride was dissolved in 40 mL of DMF. Next, 8 mL of concentrated HCl was added. In the second flask 0.984 g of terephthalic acid (H 2 BDC) was dissolved in 80 mL of DMF. Both solutions were poured into the Teflon-lined autoclave, mixed together for 30 min and heated at 80 • C for 24 h. The obtained white product was filtered under reduced pressure and washed three times with 40 mL of DMF. Then the powder was suspended in 100 mL of ethanol for three days in order to exchange the residual DMF to EtOH. In the final step the product was filtered again and dried at 80 • C for 12 h. All the chemicals were purchased from Sigma Aldrich.
The bimetallic UiO-66(Ce/Zr) was obtained by dissolving 1.73 g of ammonium cerium (IV) nitrate and 0.74 g of zirconium (IV) chloride in 15 mL DMF and mixed together for 20 min. The solution of organic linker was prepared by dissolving 1 g of H 2 BDC in 68 mL of DMF under continuous stirring for 30 min. Both the solutions were transferred to Teflon-lined autoclave and heated at 120 • C for 48 h. Obtained yellowish white product was filtered, washed with 40 mL of DMF, and suspended in 100 mL of ethanol for three days. After DMF to EtOH exchange the powder was filtered again and dried at 80 • C for 12 h.

Synthesis of Cu/UiO-66
The Cu/UiO-66(Zr) and Cu/UiO-66(Ce/Zr), with the nominal loading of Cu of 10 wt.%, were obtained via the incipient wetness impregnation (IWI) method. The ethanolic solution of copper nitrate trihydrate solution was slowly instilled into the flasks with UiO-66(Zr) or UiO-66(Ce/Zr) powders and stirred continuously. Obtained green powders were left overnight at room temperature to dry slowly and then heated at 160 • C for 12 h.

Characterization
The N 2 adsorption-desorption experiments were carried out using the Autosorb 1C apparatus (Quantachrome Instruments, Boynton Beach, FL, USA) at 77 K. Before measurements the samples were outgassed in vacuum at 100 • C for 12 h. The specific surface areas (S BET ) were determined using the multipoint BET method. The specific total pore volume was evaluated from the N 2 uptake at a relative pressure of P/P 0 = 0.99.
The powder X-ray diffraction (XRD) was conducted using the X'Pert Pro apparatus (PANalitycal, Malvern, UK) using Ni-filtered Cu Kα radiation (λ = 1.54056 Å) equipped with specific optics to perform analysis by small-angle X-ray scattering (SAXS) in transmission geometry) with X'Pert HighScore Plus program used for display and analysis of the diffraction patterns. The scanning was performed from 5 to 80 • with a rate of 0.03 • /step.
The powder X-ray diffraction (XRD) was conducted using the X'Pert Pro apparatus (PANalitycal, Malvern, UK) using Ni-filtered Cu Kα radiation (λ = 1.54056 Å) equipped with specific optics to perform analysis by small-angle X-ray scattering (SAXS) in transmission geometry) with X'Pert HighScore Plus program used for display and analysis of the diffraction patterns. The scanning was performed from 5 to 80 • with a rate of 0.03 • /step.
The thermogravimetric analysis (TGA) of the catalyst was performed using Autosorb iQ apparatus (Quantachrome, Boynton Beach, FL, USA). The change of sample mass was registered under flowing air at temperature increasing from 25 to 900 • C with a 10 • C/min heating rate.
The X-ray photoelectron spectroscopy (XPS) was performed using a Thermo-Scientific K-ALPHA spectrometer (Waltham, MA, USA) equipped in Al-K radiation (1486.6 eV) monochromatized by a twin crystal monochromator, yielding a focused X-ray spot with a diameter of 400 µm, at 3 mA × 12 kV when charge compensation was achieved with the system flood gun that provides low energy electrons and low energy argon ions from a single source. The alpha hemispherical analyser was operated in the constant energy mode with survey scan pass energies of 200 eV to measure the whole energy band, and 50 eV in a narrow scan to selectively measure the particular elements. An estimation of the intensities was done after a calculation of each peak integral. S-shaped background subtraction and fitting the experimental curve to a combination of a Lorentzian (30%) and Gaussian (70%) lines. Binding energies, referenced to the C1s line at 284.6 eV, have an accuracy of ± 0.1 eV.
The microscopic observations of the samples were carried out using scanning electron microscope (SEM, Zeiss Supra 35, Zeiss, Oberkochen, Germany) with field emission gun, equipped with TEAM™ Trident Analysis system. The secondary electron imaging and energy-dispersive X-ray spectroscopy (EDX) were used to determine the morphology of the analyzed materials. The transmission electron microscopy (TEM) was performed using Cs-corrected transmission electron microscope S/TEM Titan 80-300 (FEI), operating at an accelerating voltage of 300 kV, and equipped with Schottky-type field emission gun, CETCOR Cs-probe corrector (CEOS, Heidelberg, Germany), a Gatan Energy Filter (Tridiem 863, Cambridge, UK) and EDS detector. The scanning transmission electron microscopy (STEM) and high-resolution transmission electron microscopy (HRTEM) imaging were applied as primary research techniques. The chemical composition of studied samples was examined with energy dispersive spectrometry (EDS).

Catalytic Tests
The activity tests of synthesised UiO-66 and Cu/UiO-66 materials were performed in a non-gradient reactor of a 300 cm 3 volume using a H 2 /CO 2 mixture of a 3:1 ratio and a flow rate of 80 mL/min. The tests were carried out at 200 • C and p = 1.8 MPa. For each test, a catalyst sample of 0.2 g was used.
The concentration of reagents at the reactor outlet was measured by gas chromatography throughout the experiment that lasted 25 h.

Conclusions
In this work the UiO-66 and Cu/UiO-66 materials were obtained, characterized and tested in CO 2 hydrogenation to methanol. An approach to exchange half of Zr 4+ cations with Ce 4+ in the UiO-66 framework was made. The impact of Ce and Cu presence in the UiO-66-based catalytic system on the methanol production and selectivity to methanol was determined.
All of obtained materials were crystalline and possessed developed surface area; however, some decrease in S BET was noticed after Ce introduction to the UiO-66 framework and after copper deposition. As was proven by the XPS and STEM/EDS, the concentration of cerium in UiO-66(Ce/Zr) was lower than assumed, i.e., less than 50% of Zr was replaced with Ce. Moreover, the dispersion of Ce in the UiO-66 framework was uniform. In the obtained UiO-66(Ce/Zr) and Cu/UiO-66(Ce/Zr) samples cerium occurred in its +4 and +3 oxidation state, but the former predominated. The EDS elemental mapping indicated good Cu dispersion over the UiO-66(Zr) and UiO-(Ce/Zr) samples. Strong Cu-Zr and Cu-Ce interactions caused significant decrease in thermal stability of the UiO-66(Zr) and UiO-66(Ce/Zr) samples. Besides, strong Cu-Ce interaction in Cu/UiO-66(Ce/Zr) facilitated oxygen transfer at the Cu + -Ce 4+ interface, resulting in the reduction of Ce 4+ to Ce 3+ and oxidation of Cu + to Cu 2+ .
The composition of UiO-66-based catalytic system, especially the character of Cu-Zr and Cu-Ce interactions, had an impact on MeOH production and MeOH selectivity. The partial exchange of Zr with Ce in UiO-66 did not influence the MeOH production but significantly increased the selectivity to MeOH. It was due to the formation of surface carbonates during CO 2 adsorption on Ce sites, and their further hydrogenation first to formate species, and finally, to methanol. The CO 2 hydrogenation over Zr nodes in UiO-66(Zr) was occurring with the formation of surface bicarbonates whose hydrogenation to hydrocarboxyl species is followed with CO formation that decreases MeOH selectivity. The above phenomena did not apply in the case of the Cu/UiO-66 catalytic systems, i.e., the selectivity to MeOH was very similar for both Cu/UiO-66(Zr) and Cu/UiO-66(Ce/Zr). In addition, the former showed higher MeOH production. Both catalysts suffered from some degradation of the structure after catalytic tests. The defects formed under reaction conditions because of degradation of [Zr 6 O 4 (OH) 4 ] clusters.
It caused the decrease in the S BET and the increase of pore size. Moreover, partial destruction of UiO-66 framework led to agglomeration of copper.