English

• Catalytic hydrogenation is an essential part of the chemical production process, employed for the manufacture of fine chemicals, pharmaceuticals, fuels and other products. Hydrogenation reactions usually involve critical processes such as H₂ activation pathway, substrate-catalytic active center interaction and hydrogen transfer, in which, as the primary step, H₂ activation can be achieved through homolysis or heterolysis [1-4]. Homolysis hydrogenation is more common in heterogeneous catalysis, while hydrogen heterolysis needs to overcome higher energy barrier. However, hydrogen heterolysis is not only more favorable for the hydrogenation of polar unsaturated groups, but also achieves hydrogenation in the form of separated protons and electrons for substrates with high spatial resistance. Heterogeneous catalysts usually possess complex structures, which lead to diverse electronic structures and local coordination environments of the catalytic center, thereby inducing different heterolytic dissociation behaviors of H₂. Therefore, how hydrogen dissociation occurs and how it acts for subsequent hydrogenation steps have been an interesting but challenging topic for researchers of catalytic hydrogenation reactions.

  In recent years, synergistic combinations of frustrated Lewis acid-base pairs that are unable to form coordination bonds due to spatial site-barrier effects have been shown to induce heterolytic cleavage of H—H bonds in a "push-pull" manner, which has attracted enormous attention as soon as it was first proposed [5]. It is currently a cutting-edge research topic in the field of chemistry as it allows mild activation conditions and even avoids the involvement of metals [6-8]. While the majority of the pioneer FLP catalysts were homogeneous molecular catalysts of main-group elements, such as phosphine-borane complexes, the separation of their products and the recovery of catalysts were difficult [5,9-12]. The development of porous FLP-like heterogeneous catalysts is a promising avenue, offering homogeneous-like activity and expanding the diversity of FLP catalysts beyond solid surface FLP catalysts [13,14]. In addition, the activity of FLP catalysts can be enhanced by taking advantages of porous catalysts. For example, porous CeO₂ exposes more defect sites than its nonporous counterpart, which facilitates the construction of a higher density of FLP sites [6]. Particularly, metal-organic frameworks (MOFs) are three-dimensional crystalline materials composed of metal nodes or clusters connected by organic ligands, which have become one of the good candidate porous materials considering their enormous specific surface area, clearly defined structure, controllable pore characteristics and composition [15-17]. Currently, MOFs-based porous FLPs are mainly designed by three methods. The first method is to graft or immobilize the FLPs onto MOFs in a stepwise manner. Ma's group has made many attempts in this area. They focused on the "tailoring" of the pore environment of MOFs, using the open metal sites to anchor the N/B Lewis acid-base pairs, thus realizing the efficient hydrogenation of imine substrates by FLPs [18,19]. Subsequently, the computationally guided design and optimization of chiral FLP (CFLPs) were incorporated into the suitably regular pore. The efficient asymmetric hydrogenation effect synergistically induced by CFLP sites and MOF pore aggregation effects also expanded the application of MOFs-FLPs [20,21]. In contrast, Xu et al. anchored P/B-type FLPs within the mesopores of NU-1000, which can efficiently achieve the selective hydrogenation of nitrogen-containing heterocycles under mild conditions, preferring to N/B-type MOF-FLPs [22]. Recently, Liu et al. created O/M^δ+-like FLP sites in polyoxometalate-type MOFs by integrating coordination-defective Cu nodes and basic polyoxometalate with abundant oxygen atoms, precisely controlling the structures and spatial distances of LA and LB sites within MOFs to achieve excellent H₂ activation and catalytic hydrogenation activities [23]. The second is the synthesis of FLPs by in situ self-assembly in porous materials, by selecting suitable Lewis acid functionalized ligands or alkaline ligands. For example, porphyrin nitrogen atoms, and incorporating them into MOFs, which then encapsulate the corresponding acid-base molecules to form FLPs internally in situ by using pore effects to achieve catalytic hydrogenation of the substrate molecules [24,25]. The third approach is to build MOFs from the bottom up by creating ligands with FLPs, which is currently extremely challenging to design and synthesize [26,27]. The development of innovative porous FLPs is both scientifically and practically interesting. However, it is still an enormous challenge to create efficient FLP sites in situ on the backbones of MOFs, which are usually regarded as inert porous supports in hydrogenation processes.

  It is well known that in-situ constructed defects on reducible oxides are often originated from oxygen vacancies, which has become a research direction for solid surface FLPs, such as CeO₂ [28-33]. Ce-MOF, with the same metal-O structural motifs as CeO₂, and with a larger surface area, reversible Ce³⁺/Ce⁴⁺ redox pair and richer structure, provides a better research platform to construct FLP active sites [34-37]. Here, we constructed FLP sites consisting of Lewis acidic unsaturated Ce sites and basic sites Lewis basic vacuolar OH in situ in the unique lamellar structure Ce-UiO-66-F by a facile room-temperature synthesis approach and a convenient vacuum thermal activation step. The formation of oxygen vacancies or hydroxyl/metal ion pairs (ОH/Ce³⁺ FLPs) facilitates efficient hydrogen activation. The selective hydrogenation of dicyclopentadiene (DCPD) molecules was realized under mild conditions at 100 ℃ and a P(H₂) of 2.0 MPa, one of the steps in the preparation of Jet Propellant 10 (JP-10) fuel, which further breaks the fixed cognition that MOFs can only be used as a support material in the field of olefin hydrogenation, and also provides an in-depth study from the surface solid FLPs to the whole FLPs.

  Ce(Ⅳ)-MOFs have become attractive candidates for catalyzing hydrogenation reactions by virtue of their comparatively low cost, high redox activity and tunable functionality. Typical Ce(Ⅳ)-MOFs are usually synthesized under solvothermal conditions with toxic solvents such as N,N-dimethyl-formamide (DMF), where their relatively low product yields or undesired Ce(Ⅲ) phases are more troubling issues [38,39]. Recently, the water-based room temperature synthesis (RTS) approach has demonstrated the ability to regulate the effective synthesis of highly redox-active Ce(Ⅳ)-MOFs, with very high temporal and spatial yields, and thus this approach confers a great potential for a more sustainable industrial production with lower energy losses and safer conditions [40]. Importantly, RTS is highly versatile for controlling the crystal defect pathway, the so-called low-temperature induced defect method [41,42]. Thus room temperature synthesized Ce(Ⅳ)-MOFs will exhibit varying degrees of defects, especially for the 12-connected phase, while maintaining the chemical stability of the parent MOFs. Meanwhile, Ce-UiO-66-F showed better thermal stability relative to Ce-UiO-66, while the use of fluorine-functionalized ligands was able to increase the Lewis acidity of the metal nodes. Based on the above, we synthesized Ce-UiO-66-F in one step using this RTS method, specifically by adding ceric ammonium nitrate, 2-fluoroterephthalic acid and formic acid to water and stirring at room temperature for 5 h, in which formic acid was used as a moderator. Then, the activated samples were obtained and named Ce-UiO-66-F-X, after being treated at different pyrolysis temperatures for 3 h under vacuum condition (Fig. 1 and Fig. S1 in Supporting information).

  Figure 1

  

  Figure 1.  The preparation processes of Ce-UiO-66-F via room temperature synthesis (RTS) approach followed by vacuum-activated treatment.

  DownLoad: Full-Size Img PowerPoint

  As shown in Fig. 2, TEM image of Ce-UiO-66-F exhibits a lamellar interlayer morphology, which is not significantly changed after pyrolysis treatment at different pyrolysis temperatures (Figs. 2a and b, Figs. S2 and S3 in Supporting information). Particularly, compared with the irregular bulk morphology of Ce-UiO-66-H synthesized under room temperature condition (Fig. S4 in Supporting information), the presence of the neighboring functional group-F was capable of regulating the MOF growth process. The successful synthesis of Ce-UiO-66-Br, which has a non-uniform octahedral structure, also proved that the presence of functional groups with electron-drawing effects in the adjacent site can significantly alter the morphology of MOFs (Fig. S5 in Supporting information).

  Figure 2

  

  Figure 2.  SEM and TEM images for (a) Ce-UiO-66-F and (b) Ce-UiO-66-F-200. EDS elemental mapping images of (c) Ce-UiO-66-F-200. HR-TEM images of (d) Ce-UiO-66-F and (e) Ce-UiO-66-F-200.

  DownLoad: Full-Size Img PowerPoint

  It is likely that the precise modulating effect will be influenced by the characteristics of the nearby atoms, such as radius and electronegativity. The EDS mapping image displays the Ce, O, and F elements in various colors, indicating that Ce-UiO-66-F was effectively obtained (Fig. 2c). The catalyst was characterized by PXRD as displayed in Fig. 3a. The PXRD pattern of Ce-UiO-66-F is identical to the crystalline phase of Ce-UiO-66 synthesized by a previous solvothermal method, and the peak intensities remain nearly unchanged after the pyrolysis treatment. Comparatively the Ce-UiO-66-H showed phase separation after vacuum pyrolysis treatment at 200 ℃, i.e., the appearance of a new phase of Ce-MOF (CSUST-1) [43] with mixed valence (Fig. S6a in Supporting information), thus Ce-UiO-66-F has a strong thermal stability. Meanwhile, we subsequently immersed the synthesized Ce-UiO-66-F in boiling water for 12 h, which showed excellent hydrolytic stability (Figs. S7 and S8 in Supporting information). It is possible that during pyrolysis, the F atom may be involved in coordination after partial removal of the ligand, which would contribute to the stability [44-46]. The edges of the Ce-UiO-66-F nanosheets display lattice fringes of (444) and (622), corresponding to lattice spacings of 0.306 nm and 0.321 nm, respectively (Fig. 2d). Relatively speaking, the lattice of Ce-UiO-66-F-200 is more disordered, with concave-convex lattice points and pores visible in the orange circle area. Furthermore, from the lattice arrangement with relative networks, the crystal plane spacings obtained by Fourier transform are 0.309 nm and 0.324 nm respectively (Fig. 2e). From the above results, it can be seen that the thermal activation step has caused the crystals to suffer lattice strain to some extent.

  Figure 3

  

  Figure 3.  Characterization of the Ce-UiO-66-F and active Ce-UiO-66-F-X catalyst: (a) XRD, (b) FTIR, (c) N₂ sorption isotherms and (d) pore size distributions.

  DownLoad: Full-Size Img PowerPoint

  The presence of various functional groups in the Ce-UiO-66-BDC-X compounds can be identified using the infrared (IR) spectra from their distinctive bands. As shown in Fig. 3b, antisymmetric (1580–1496 cm^-1) and symmetric (1415–1380 cm^-1) stretches of the ligand carboxylic acid group appeared in the spectra of all four samples, respectively. Three peaks at 2962, 2856 and 2928 cm^-1 can be attributed to ν(C−H) stretching vibration and δ(C−H) + ν_as(OCO) of the coordinated formate. In addition, a clearly C-F stretching vibration is exhibited at 1225 cm^-1, 520 cm^-1 belong to the Ce-O stretching vibrations, confirming the successful coordination formation of Ce-UiO-66-F. The typical N₂ adsorption and desorption isotherms are utilized to investigate the surface area and pore-size distribution of Ce-MOFs. Generally, the Ce-UiO-66 prepared with the conventional solvothermal approach features a large number of micropores, with the pore size generally concentrated around 0.8 nm. All the Ce-UiO-66-F-X synthesized by room temperature synthesis method show typical type-Ⅳ isotherms, except for the 0.8 nm micropores, also appeared mesopores in the range of 3–7 nm, and further vacuum thermal activation slightly shifted the mesopores to larger sizes (Figs. 3c and d). Meanwhile, the surface area and total pore volume as well as the average pore diameter of Ce-UiO-66-F-X gradually enhanced with the thermal activation temperature. However, as the pyrolysis temperature approaches 250 ℃, the skeleton structure may partially collapse due to further ligand elimination, resulting in a loss of specific surface area and pore volume to some extent (Table S1 in Supporting information). Similarly, Ce-UiO-66-H/Br also showed changes in specific surface area, pore size and other properties after thermal activation at 200 ℃, which further also demonstrated the universality and simplicity of the vacuum thermal activation method (Figs. S6 and S9, Table S2 in Supporting information). The surface states of Ce and O elements in the catalysts were analyzed by XPS technique (Fig. 4 and Fig. S10 in Supporting information). The fitted O 1s spectra of Ce-UiO-66-F and Ce-UiO-66-F-X have three peaks at 529.6–530.0, 531.3–531.6, and 532.2–532.8 eV attributed to the lattice oxygen Ce-O-Ce, carboxylate group oxygen Ce-O-C, adsorbed oxygen (O_ad/-OH_ad), or adsorbed water, respectively. For the Ce-UiO-66-F sample, the fitted peaks at 885.7 eV and 903.8 eV correspond to Ce³⁺, while the other fitted peaks correspond to Ce⁴⁺. The shift in the binding energy of O 1s after vacuum pyrolysis treatment of Ce-MOF is attributed to the reduction of Ce⁴⁺ to Ce³⁺, accompanied by the formation of defects and vacancies, which leads to the unsaturation of the coordination state around the atoms, as well as the generation of defective sites to adsorb oxygen species. Moreover, with the increase of pyrolysis temperature, the peak of F 1s moves toward lower binding energy, indicating a tendency of electron transfer from Ce metal atoms to F atoms, which indirectly suggests the deficiency in coordination of O atoms in the original skeleton. At the same time, the ratio of Ce³⁺/(Ce³⁺+Ce⁴⁺) varied with increasing pyrolysis temperature to 23.83%, 24.91%, 26.92%, and 26.31%. In addition, the ratios of O_ad/O_total for Ce-UiO-66-F and Ce-UiO-66-F-X exhibited ratios of 23.58%, 27.65%, 35.17%, and 26.70% (Fig. 4d), respectively, which were related to the catalytic activity as discussed below. The defect/oxygen vacancies in the structure lead to a spatial redistribution of the local electron density, where the electron donor binding energy increases (Lewis basic sites of oxygen) and the electron acceptor binding energy decreases (Lewis acidic sites of Ce³⁺). The σ orbitals of H—H interact with Ce sites and the σ* orbitals interact with the terminal O sites, leading to hetero-cleavage of the H—H bond, thus the difference in the electron cloud density between Ce and O will be suitable for the H—H bond activation and H-transfer in the DCPD hydrogenation process. Using CO as a probe molecule, we undertook an infrared spectroscopic study on Ce-UiO-66-F-200, a representative sample, to better validate the oxidation state of the porous Ce-MOF material and its local environment. In addition to the physical adsorption of CO (≈2141 cm^-1), the strong splitting peaks at 2171 cm^-1 for the bridging adsorption of Ce⁴⁺ with CO and the adsorption peak of Ce³⁺ with CO at 2119 cm^-1 are further evidence of the unsaturated coordination state of the samples (Fig. S11 in Supporting information) [47,48].

  Figure 4

  

  Figure 4.  (a) Ce 3d, (b) O 1s, (c) F 1s XPS, the ratio of O_ad/O_total and Ce³⁺/(Ce³⁺+Ce⁴⁺) of the Ce-UiO-66-F and (d) active Ce-UiO-66-F-X catalyst.

  DownLoad: Full-Size Img PowerPoint

  The hydrogenation of DCPD to endo-tetrahydrodicyclopentadie (endo-THDCPD) was employed as a probe molecule. Due to the fact that endo-THDCPD is a crucial intermediate product of JP-10 fuel with a greater volumetric energy content to offer more propulsive energy, it is particularly valuable for research [49,50]. Hydrogenation of DCPD is usually carried out with metal catalysts, such as Ni, Pd, and Pt, the easy loss and toxicity of noble metals and the high cost constrain its practical application. In addition, DCPD is a dimer of cyclopentadiene (CPD) with two unsaturated double bonds, namely the norbornene ring (NB-bond) and the cyclopentene ring (CP-bond) (Fig. 5a). Therefore, the activation of the two double bonds as well as the dynamic process of hydrogenation has also been the direction of research in the last decade. In the catalytic tests, Ce-UiO-66-F-200 exhibited the optimal catalytic performance for DCPD, with a conversion of 96.9% at 10 h (H₂ pressure 2 MPa, temperature 100 ℃). As a contrast, the conversion of unactivated Ce-UiO-66-F was about 10% in the same reaction time, and the great difference in performance indicates that the thermal activation step is critical for the exposure of the active site (Fig. 5b). On the other hand, unactivated Ce-UiO-66 then has negligible conversion, so the fluorine functionalized groups affect the growth of MOF to some extent, and Ce-UiO-66-F with a unique two-dimensional morphology will expose more surface and edge active sites relative to the irregular bulk Ce-UiO-66. Similarly, Ce-UiO-66-Br with electron-withdrawing effect exhibits modest improvement in catalytic activity after activation at 200 ℃ (conversion from 4% to 30%), demonstrating that simple activation is useful for triggering catalytically active sites. However, unfortunately, its better activation effect cannot be reached at this lower energy. For Ce-UiO-66-H, treatment at 200 ℃ leads to drastic change of crystal structure (or structure collapse and destruction) with the conversion is only 14%, and thus the activation effect is negligible (Fig. 5c). The Ce-UiO-66–200 sample maintains a good stability of properties in the first three times, and starts to show a more obvious decrease in the fourth time (Fig. S12 in Supporting information). The XRD and XPS results show that the crystalline phase of Ce-UiO-66-F-200 stays unchanged after the repeated hydrogenation tests (Fig. S13 in Supporting information). However, there is a slight increase in the proportion of Ce³⁺ due to reduction but a significant decrease in the amount of adsorbed oxygen, which may be attributed to the presence of -OH occupying the unsaturated sites during hydrogenation, as well as structural collapse due to temperature and pressure.

  Figure 5

  

  Figure 5.  (a) Reaction pathways for hydrogenation of DCPD to endo-THDCPD. Conversion and selectivity for (b) Ce-UiO-66-F-X and (c) Ce-UiO-66-H/Br-X.

  DownLoad: Full-Size Img PowerPoint

  During this hydrogenation process, 8,9-dihydrodicyclopentadiene (8,9-DHDCPD), which is an intermediate product after saturating the NB bond, is selectively obtained. This is likely due to the higher energy required for the hetero-cleavage mode of the FLPs on the Ce-MOF as opposed to the homo-cleavage mode of H₂ on the transition metal surface, as well as the higher energy needed to activate the CP bond. This discrepancy in hydrogenation reactivity can be attributed to the structural disparities between the two double bonds, primarily in their connectivity to neighboring rings. Similarly, based on the energy difference between the activation of the C≡C and C=C, we can also utilize the Ce-UiO-66-F-200 to perform validation experiments for the selective hydrogenation to styrene (Fig. S12c). To the best of our knowledge, the utilization of the intrinsic active sites of MOF skeletons for C=C hydrogenation of olefins is still seldom reported, therefore, it is of great importance to investigate how to improve intrinsic activity of MOFs for the research and development of new materials as well as for the further understanding of the hydrogenation mechanisms.

  Ce-UiO-66-F's typical thermogravimetric (TGA) plot illustrates three key stages. Desolvation, which typically takes place at temperatures below 150 ℃, is the process of removing water molecules from the porous structure. The second step corresponds to the removal of structural water molecules, partially compensated ligands and linked ligands in the temperature range of about 150–300 ℃. In the final step, the organic ligands of the Ce-MOF are completely decomposed, ultimately yielding CeO₂ (Figs. 6a and b). To exclude the possibility of metal cluster loss, we conducted elemental content testing on the samples, which revealed that the metal content remained almost unchanged after thermal decomposition (Table S3 in Supporting information). The ideal Ce-UiO-6-F is defect-free in the absence of auxiliary ligands competing for coordination, and all Ce nodes are fully coordinated. The molecular weight ratio of the six BDC-F linkers to CeO₂ is 96/100, and the chemical formula in the fully desolvated state should be Ce₆O₆(BDC-F)₆. However, the addition of formic acid competing ligand resulted in Ce₆O_6+x(BDC-F+HCOO)_6-x [51,52]. Based on the nuclear magnetic resonance (NMR) results of Ce-UiO-66-F showing HCOOH: BDC-F = 0.315, the ideal molecular formula can be calculated to be Ce₆O_6.25(BDC-F)_4.37(HCOO)_1.38. The ligand connectivity of Ce-UiO-66-F synthesized at room-temperature is 11.5, which is less than the perfect 12 linkage number, corroborating the reports of low-temperature-induced defects. Given that TGA demonstrates that the Ce₆ node is less connected than the ideal structure, the number of connections of the samples after pyrolysis treatments at 150, 200, and 250 ℃ were calculated to be 9.88, 8.44 and 7.92, respectively, based on the ratios of the ligand decomposition content to the residual CeO₂ as well as the NMR data (Fig. 6c). As a representative, the molecular formula of Ce-UiO-66-F-200 can be calculated to be Ce₆O_7.78(BDC-F)_3.98(HCOO)_0.30, showing that mainly 80% of the formic acid molecules competing for coordination are removed from the coordination structure of the inorganic node after thermal activation at 200 ℃.

  Figure 6

  

  Figure 6.  Characterization of the Ce-UiO-66-F and Ce-UiO-66-F-X catalyst: (a) TG, (b) 2^nd derive of TG (determine the temperature at which the ligand begins to lose weight), (c) EPR spectra, (d) ¹H NMR spectra, (e) DRS UV–vis spectra and (f) DRIFTS spectra of Ce-UiO-66-F and active Ce-UiO-66-F-X.

  DownLoad: Full-Size Img PowerPoint

  To further identify defects in Ce-MOF, the electron paramagnetic resonance (EPR) was employed to study these catalysts. As displayed in Fig. 6d, EPR spectroscopy of Ce-UiO-66-F-200 reveals a characteristic signal at g = 2.13, which is increased and broadened when compared to the unactivated Ce-UiO-66-F, implying that thermal activation generates a high number of unpaired electrons or oxygen vacancies. As shown in Fig. 6e, the UV–vis diffuse reflectance spectroscopy (DRS) of all samples display a broad band between 250 nm and 500 nm, which is attributed to a ligand-to-metal charge transfer, i.e., from the 2p valence band of O^2- to the 4f levels of Ce⁴⁺. As the structure becomes relatively disordered after thermal excitation, the energy difference between the orbitals occupied before and after the electron leaps becomes smaller, the probability of electron leaps is larger, and the value of the molar absorption coefficient (ε) is larger, yielding what is known as the hyperchromic effect (increase in absorbance) in analytical chemistry. In particular, the absorption edge shows a slight red shift with increasing pyrolysis temperature, which is due to the increase in Ce³⁺ and oxygen vacancies [48,53,54]. As well, the color changes shown in the optical image in Fig. S14 (Supporting information) further validate the effect of structural changes on the absorption of light. As we all know that defect sites can be capped by hydroxyl groups, water, as well as residual moderator or formate anions with charge-compensating ability, diffuse reflectance infrared Fourier-transform spectroscopy (DRIFTS) has been collected in order to compare hydroxyl group stretching at the nodes (Fig. 6f). For the defect-free UiO-66 which theoretically should have only one peak in the hydroxyl region corresponding to μ₃—OH, the presence of shoulder peak also appeared at lower frequencies in the Ce-UiO-66-F system, indicating the presence of defects. Moreover, with further pyrolysis, the main μ₃—OH signal at 3645 cm^-1 weakens, and the other two broad peaks concentrated at 3634 and 3625 cm^-1 gradually become more prominent, which can be attributed to the hydroxyl groups close to the defect sites and the other hydroxyl groups in Ce-SBU are affected, which also indicates the formation of more defects in Ce clusters [55].

  The electronic structure and coordination environment of Ce atoms in defective Ce-MOF materials were further analyzed by using extended X-ray absorption fine structure (XAFS) in transmission mode [56]. As shown in Fig. S15a (Supporting information), the X-ray absorption near edge structure (XANES) spectrum of Ce L₃ edge of Ce-UiO-66-F is similar to CeO₂, and there are mainly two absorption features at 5726.1 and 5735.2 eV at the absorption edge, which are divided into transitions from 2p to 4f¹5d and 4f⁰5d respectively [57]. In particular, the energy 5726.1 eV of the main edge transition in the spectrum shows an obvious red shift relative to the 5728.9 characteristic peak of CeO₂, which is also higher than the characteristic peak 5724.7 eV of Ce³⁺, proving that Ce-UiO-66-F involves a mixed state of Ce³⁺/Ce⁴⁺. Following vacuum thermal activation, the sample Ce-UiO-66-F-200 showed a further shift to a lower binding energy, indicating a decrease in valence state. Meanwhile, the intensity of the white peaks of Ce-MOF becomes wider and weaker after thermal activation, which may be related to the change of the local environment around Ce atoms, probably accompanied by the deformation and distortion of Ce clusters. It is also further revealed that the higher atomic orbital splitting degree (charge density delocalization distribution) after thermal activation potentially indicates the presence of more structural disorder and defects. The trend observed in the extracted EXAFS signals also confirms the distortion of the clusters, with a decrease in the oscillation amplitude after thermal activation (Fig. S15b in Supporting information). The defect structure can be better understood from the pseudo-radial distribution function obtained from the Fourier transform of EXAFS signals. The Fourier transform of the first shell layer indicates Ce-O distances, with the first peak representing the Ce-O₁ distance from the bridging oxygen of the cluster and the other peak attributable to the Ce-O₂ distance from the Ce atom bound to the carboxylic acid ligand (Fig. S15c in Supporting information) [58,59]. After thermal activation, the height of the peak at the same position decreases significantly, indicating that the strength of the Ce-O shell layer decreases upon pyrolysis (decrease in coordination number), suggesting that pyrolysis triggers a loss of linkers (defects) in the MOF. Wavelet transform spectra (WT-EXAFS) also confirmed this assignment, as shown in the three-dimensional contour plots in Figs. S15d-f (Supporting information). Therefore, it is also confirmed that the elongation of the Ce-O distance is consistent with the structural delocalization redistribution of charge density when oxygen vacancies are formed

  For Lewis acidic MOFs with defective nodes, there is a relatively high correlation between the dicarboxylate electron density and Ce₆O₄(OH)₄ Lewis acidity. The unsaturated coordination metal site in MOFs can be effectively bound by the strong coordination electron lone pair of the high emission fluorescent dye N-methylacridone (NMA), which results in a significant shift in the fluorescence maximum (λ_max) [60]. As illustrated in Fig. S16a (Supporting information), the emission peaks of NMA at 413.0 and 433.0 nm shifted to 435.2 and 465.8 nm, respectively, upon binding to Ce-UiO-66. Especially, Ce-UiO-66-F-200 with a maximum wavelength of 468.4 nm, shows the greatest shift tendency of the NMA emission wavelength (Fig. S17 in Supporting information). The position shift indicates the binding affinity of the fluorescent dye molecule for the metal and can be utilized for estimating the Lewis acidity. In light of this, the fluorine-functionalized ligands and defect constructions synergistically enhance the Lewis acidity of Ce-MOFs. To further evaluate the H₂ activation ability of Ce-UiO-66-F-200, the basicity and acidity of the material were measured by temperature-programmed desorption (TPD) profiles (Figs. S16b and c in Supporting information). As shown, there are two distinct bulges in the NH₃-TPD curve, corresponding to the Lewis acid sites on the MOF backbone, which belong to the unsaturated Ce³⁺ centers. After further increasing the vacuum thermal activation temperature to 250 ℃, although it exhibited a slightly higher desorption temperature, its adsorption capacity was significantly reduced, which was strongly correlated with the structural damage (Fig. S18 in Supporting information). Meanwhile the characteristic peak in the CO₂-TPD profile of Ce-UiO-66-F-200 located at about 90 ℃ reveals a Lewis base site provided by a vacant adjacent O atom, however the Lewis base is weak. Thus, the coexistence of Lewis acidic and basic sites observed in the Ce-UiO-66-F-200 backbone may indicate H₂ activation capacity, and this has been revealed by their catalytic activity in the hydrogenation of DCPD.

  Infrared spectroscopy is a powerful technique to analyze the properties of UiO MOFs, especially the μ₃—OH stretching frequency and the carboxylate stretching frequency. For the hydroxyl vibration of Zr-MOFs, the characteristic peak at 3670 cm^-1 is compatible with the Zr-OH/H₂O vibration newly produced from H₂ activation via dissociation with terminal OH, protonation of the terminal oxygen atom. Relevant studies demonstrate that when the Zr metal node in the cluster is replaced by Ce, the characteristic infrared wavelength of μ₃—OH will be significantly blue-shifted, and when all the metals in the cluster are replaced by Ce, its infrared wavelength is shifted to 3645 cm^-1 [55,61]. Furthermore, the fluorination modification will also have a considerable impact on the properties and concentration of OH groups. As illustrated by the FT-IR spectra of the hydroxyl region of Fig. S16f (Supporting information), Ce-UiO-66-F-200 appeared to have a newly generated characteristic peak of -OH approximately at 3640 cm^-1 after 1 h of H₂ exposure, and its signal gradually became weaker with air exposure time increased and completely disappeared after 50 min. Comparatively, Ce-UiO-66-F showed only a weak signal at 3640 cm^-1 after 1 h of H₂ exposure, and this characteristic peak vanished after only 5 min of air exposure (Fig. S16e in Supporting information). Particularly, this characteristic peak generated by hydrogen cleavage was hardly observed for Ce-UiO-66-H after treatment under hydrogen (Fig. S16d in Supporting information). Hence, this also confirms that vacuum activation can increase the ability of MOF hydrogen cleavage and also change the binding strength of hydrogen protons on metal clusters (surface oxygen vacancies can enhance the stability of H species).

  Based on the above discussion, the oxygen vacancies caused by ligand deficiency play a crucial role in the hydrogenation reaction, and we list the stereograms of the parameters such as ligand connectivity (LC) number, Ce³⁺ ratio, and adsorbed oxygen ratio in relation to the catalytic performance to facilitate us to understand the variations more clearly, and thus we conclude that more suitable ligand deficiencies show the optimal hydrogenation performance (Fig. S19 in Supporting information). Finally, we propose the reaction process of activated Ce-MOF in hydrogen activation as well as NB-bond activation of DCPD (Fig. S20 in Supporting information). Unquenched acid-base pairs consisting of coordinated unsaturated Ce and adjacent OH with oxygen vacancy spacers possess LA^…LB center distances that match H—H bonds, increasing the overlap with the H₂ electron cloud and inducing heterolytic cleavage of the H—H bond in a "push-pull" manner, which effectively improves the activation efficiency of H₂ at low temperatures. Thus, the heterolytic H₂ dissociation at the porous FLP sites of Ce-UiO-66-F-X forms protons (O-H^δ+) and hydrides (Ce-H^δ-). Secondly, due to the nucleophilic characteristics of DCPD molecules, the (C=C) NB bond preferentially reacts with O-H^δ+ to produce carbon positive ions, which then adsorb onto the unsaturated Ce sites. Finally, the saturated (C-C) NB bond is obtained by hydrogen transfer from Ce-H^δ- and desorption occurs, realizing the hydrogenation process.

  Based on our experimental observations, Ce-MOFs are more susceptible to the conditions of the heat treatment process and thus can achieve thermal activation at lower temperatures. Therefore, we have realized the construction of a three-dimensional porous FLP catalyst by structurally modulating intrinsic Ce-UiO-66-F metal nodes and the pore environment by utilizing a vacuum thermally activated vacancy-mediated defect strategy. Specifically, different vacuum thermal activation temperatures (T = 150, 200, 250 ℃) modulated the valence and proportion of metal Ce in Ce-UiO-66-F, as well as the number of Ce cluster (SBU) connections, with the effects on the pore structures. The catalyst performs highly efficient heterogeneous selective hydrogenation of DCPD molecules involving H₂ heterolytic dissociation. These results, in our opinion, have the potential to improve MOFs-based catalyst design and provide important new understandings of related catalytic systems. Therefore, as a pioneering academic finding to replace/reduce the use of transition metals, the intrinsic clearly defined spatial structure of FLP sites of intrinsic Ce-MOF builds a firm groundwork for the H₂ activation mode, which opens up a new canvas for hydrogenation catalysis.

  Declaration of competing interest

  The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

  CRediT authorship contribution statement

  Danfeng Zhao: Writing – review & editing, Writing – original draft, Investigation, Formal analysis, Data curation. Jing Lin: Writing – review & editing, Visualization, Validation, Formal analysis. Rushuo Li: Writing – review & editing, Validation, Supervision, Formal analysis, Data curation. Liang Chu: . Zhaokun Wang: Formal analysis, Data curation. Xiubing Huang: Writing – review & editing, Visualization, Supervision, Resources, Methodology, Conceptualization. Ge Wang: Writing – review & editing, Resources, Project administration, Investigation, Funding acquisition, Conceptualization.

  Acknowledgments

  This work was supported by the National Key Research and Development Program of China (No. 2021YFB3500700), the National Natural Science Foundation of China (No. 51972024), Natural Science Foundation of Guangdong Province (No. 2022A1515010185), Fundamental Research Funds for the Central Universities (No. FRF-EYIT-23–07).

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2024.110172.

  
  
• 1. [1]

     W. Gao, S. Liu, G. Sun, et al., Small 19 (2023) 2300956.

  2. [2]

     K. Chang, Y. Dong, X. Xu, Chem. Comm. 55 (2019) 12777–12780. doi: 10.1039/c9cc06676d

  3. [3]

     L. Zhang, Z. Chen, Z. Liu, et al., Nat. Commun. 12 (2021) 6574.

  4. [4]

     G.J. Kubas, Science 314 (2006) 1096–1097. doi: 10.1126/science.1135430

  5. [5]

     G.C. Welch, R.R.S. Juan, J.D. Masuda, et al., Science 314 (2006) 1124–1126. doi: 10.1126/science.1134230

  6. [6]

     S. Zhang, Z.Q. Huang, Y. Ma, et al., Nat. Commun. 8 (2017) 15266.

  7. [7]

     Z.Q. Huang, L.P. Liu, S. Qi, et al., ACS Catal. 8 (2018) 546–554. doi: 10.1021/acscatal.7b02732

  8. [8]

     Y. Ma, S. Zhang, C.R. Chang, et al., Chem. Soc. Rev. 47 (2018) 5541–5553. doi: 10.1039/c7cs00691h

  9. [9]

     A.R. Jupp, D.W. Stephan, Trends Chem. 1 (2019) 35–48.

  10. [10]

      J. Lam, K.M. Szkop, E. Mosaferi, et al., Chem. Soc. Rev. 48 (2019) 3592–3612. doi: 10.1039/c8cs00277k

  11. [11]

      D.W. Stephan, Acc. Chem. Res. 48 (2015) 306–316. doi: 10.1021/ar500375j

  12. [12]

      D.W. Stephan, Chem 6 (2020) 1520–1526.

  13. [13]

      Y. Zhang, P.C. Lan, K. Martin, et al., Chem. Catal. 2 (2022) 439–457.

  14. [14]

      A. Dai, S. Li, T. Wang, et al., Chin. Chem. Lett. 34 (2023) 107559.

  15. [15]

      D. Zhao, X. Li, K. Zhang, et al., Coord. Chem. Rev. 487 (2023) 215159.

  16. [16]

      Q. Liu, Y. Li, Y. Fan, et al., J. Mater. Chem. A 8 (2020) 11442–11447. doi: 10.1039/d0ta01845g

  17. [17]

      Y. Zhong, P. Liao, J. Kang, et al., J. Am. Chem. Soc. 145 (2023) 4659–4666. doi: 10.1021/jacs.2c12590

  18. [18]

      Z. Niu, W.D.B. Gunatilleke, Q. Sun, et al., Chem 4 (2018) 2587–2599.

  19. [19]

      Z. Niu, W. Zhang, P.C. Lan, et al., Angew. Chem. Int. Ed. 58 (2019) 7420–7424. doi: 10.1002/anie.201903763

  20. [20]

      Y. Zhang, S. Chen, A.M. Al-Enizi, et al., Angew. Chem. Int. Ed. 135 (2023) e202213399.

  21. [21]

      J. Zhang, Y. Shao, Y. Li, et al., Chin. Chem. Lett. 29 (2018) 1226–1232.

  22. [22]

      Z.M. Xu, Z. Hu, Y. Huang, et al., J. Am. Chem. Soc. 145 (2023) 14994–15000. doi: 10.1021/jacs.3c04929

  23. [23]

      Y. Liang, Z. Zhang, X. Su, et al., Angew. Chem. Int. Ed. 135 (2023) e202309030.

  24. [24]

      S. Shyshkanov, T.N. Nguyen, F.M. Ebrahim, et al., Angew. Chem. Int. Ed. 58 (2019) 5371–5375. doi: 10.1002/anie.201901171

  25. [25]

      S. Shyshkanov, T.N. Nguyen, A. Chidambaram, et al., Chem. Comm. 55 (2019) 10964–10967. doi: 10.1039/c9cc04374h

  26. [26]

      J. Ye, J.K. Johnson, ACS Catal. 5 (2015) 6219–6229. doi: 10.1021/acscatal.5b01191

  27. [27]

      J. Ye, J.K. Johnson, ACS Catal. 5 (2015) 2921–2928. doi: 10.1021/acscatal.5b00396

  28. [28]

      J. Jia, C. Qian, Y. Dong, et al., Chem. Soc. Rev. 46 (2017) 4631–4644.

  29. [29]

      L. Wang, T. Yan, R. Song, et al., Angew. Chem. Int. Ed. 58 (2019) 9501–9505. doi: 10.1002/anie.201904568

  30. [30]

      K. Werner, X. Weng, F. Calaza, et al., J. Am. Chem. Soc. 139 (2017) 17608–17616. doi: 10.1021/jacs.7b10021

  31. [31]

      L. Li, W. Liu, R. Chen, et al., Angew. Chem. Int. Ed. 61 (2022) e202214490.

  32. [32]

      X. Huang, K. Zhang, B. Peng, et al., ACS Catal. 11 (2021) 9618–9678. doi: 10.1021/acscatal.1c02443

  33. [33]

      D. Yang, Y. Xu, K. Pan, et al., Chin. Chem. Lett. 33 (2022) 378–384.

  34. [34]

      Y. Guo, Z. Liu, F. Zhang, et al., ChemCatChem 13 (2021) 874–881. doi: 10.1002/cctc.202001531

  35. [35]

      Ü. Kökçam-Demir, A. Goldman, L. Esrafili, et al., Chem. Soc. Rev. 49 (2020) 2751–2798. doi: 10.1039/c9cs00609e

  36. [36]

      Y. Liang, C. Li, L. Chen, et al., Chin. Chem. Lett. 32 (2021) 787–790.

  37. [37]

      X. Sun, K. Yuan, Y. Zhang, J. Rare Earths 38 (2020) 801–818.

  38. [38]

      M. Lammert, M.T. Wharmby, S. Smolders, et al., Chem. Comm. 51 (2015) 12578–12581.

  39. [39]

      M. Lammert, C. Glißmann, H. Reinsch, et al., Cryst. Growth Des. 17 (2017) 1125–1131. doi: 10.1021/acs.cgd.6b01512

  40. [40]

      S. Dai, E. Montero-Lanzuela, A. Tissot, et al., Chem. Sci. 14 (2023) 3451–3461. doi: 10.1039/d2sc05161c

  41. [41]

      S. Dai, F. Nouar, S. Zhang, et al., Angew. Chem. Int. Ed. 133 (2021) 4328–4334. doi: 10.1002/ange.202014184

  42. [42]

      G. Huang, Y. Wang, T. Liu, Chin. Chem. Lett. 30 (2019) 2309–2312.

  43. [43]

      H.G. Jin, M. Wang, J.X. Wen, et al., ACS Appl. Mater. Interfaces 13 (2021) 3899–3910. doi: 10.1021/acsami.0c18899

  44. [44]

      J.P. Vizuet, M.L. Mortensen, A.L. Lewis, et al., J. Am. Chem. Soc. 143 (2021) 17995–18000. doi: 10.1021/jacs.1c10493

  45. [45]

      A. Cadiau, Y. Belmabkhout, K. Adil, et al., Science 356 (2017) 731–735. doi: 10.1126/science.aam8310

  46. [46]

      M.S. Christian, K.J. Fritzsching, J.A. Harvey, et al., JACS Au 2 (2022) 1889–1898. doi: 10.1021/jacsau.2c00259

  47. [47]

      S. Rojas-Buzo, P. Concepción, J.L. Olloqui-Sariego, et al., ACS Appl. Mater. Interfaces 13 (2021) 31021–31030. doi: 10.1021/acsami.1c07496

  48. [48]

      S. Rojas-Buzo, D. Salusso, F. Bonino, et al., Mater. Today Chem. 27 (2023) 101337.

  49. [49]

      D. Jia, J. Zhao, Z. Tao, et al., Carbon 184 (2021) 855–863.

  50. [50]

      C. Wu, X. Chen, J. Liang, et al., Chem. Eng. J. 462 (2023) 142141.

  51. [51]

      G.C. Shearer, S. Forselv, S. Chavan, et al., Top. Catal. 56 (2013) 770–782. doi: 10.1007/s11244-013-0027-0

  52. [52]

      X. Feng, H.S. Jena, C. Krishnaraj, et al., J. Am. Chem. Soc. 143 (2021) 21511–21518. doi: 10.1021/jacs.1c05357

  53. [53]

      S. Rojas-Buzo, B. Bohigues, D. Salusso, et al., ACS Catal. 13 (2023) 9171–9180. doi: 10.1021/acscatal.3c00502

  54. [54]

      B. Choudhury, A. Choudhury, Mater. Chem. Phys. 131 (2012) 666–671.

  55. [55]

      T. Islamoglu, D. Ray, P. Li, et al., Inorg. Chem. 57 (2018) 13246–13251. doi: 10.1021/acs.inorgchem.8b01748

  56. [56]

      S. Smolders, K.A. Lomachenko, B. Bueken, et al., ChemPhysChem 19 (2018) 373–378. doi: 10.1002/cphc.201700967

  57. [57]

      X. Kan, G. Zhang, J. Ma, et al., Adv. Funct. Mater. 34 (2024) 2312044.

  58. [58]

      K.A. Lomachenko, J. Jacobsen, A.L. Bugaev, et al., J. Am. Chem. Soc. 140 (2018) 17379–17383. doi: 10.1021/jacs.8b10343

  59. [59]

      M. Ronda-Lloret, I. Pellicer-Carreño, A. Grau-Atienza, et al., Adv. Funct. Mater. 31 (2021) 2102582.

  60. [60]

      P. Ji, T. Drake, A. Murakami, et al., J. Am. Chem. Soc. 140 (2018) 10553–10561. doi: 10.1021/jacs.8b05765

  61. [61]

      C. Atzori, K.A. Lomachenko, J. Jacobsen, et al., Dalton Trans. 49 (2020) 5794–5797. doi: 10.1039/d0dt01023e

• 

  Figure 1  The preparation processes of Ce-UiO-66-F via room temperature synthesis (RTS) approach followed by vacuum-activated treatment.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  SEM and TEM images for (a) Ce-UiO-66-F and (b) Ce-UiO-66-F-200. EDS elemental mapping images of (c) Ce-UiO-66-F-200. HR-TEM images of (d) Ce-UiO-66-F and (e) Ce-UiO-66-F-200.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  Characterization of the Ce-UiO-66-F and active Ce-UiO-66-F-X catalyst: (a) XRD, (b) FTIR, (c) N₂ sorption isotherms and (d) pore size distributions.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  (a) Ce 3d, (b) O 1s, (c) F 1s XPS, the ratio of O_ad/O_total and Ce³⁺/(Ce³⁺+Ce⁴⁺) of the Ce-UiO-66-F and (d) active Ce-UiO-66-F-X catalyst.

  下载: 全尺寸图片 幻灯片
  

  Figure 5  (a) Reaction pathways for hydrogenation of DCPD to endo-THDCPD. Conversion and selectivity for (b) Ce-UiO-66-F-X and (c) Ce-UiO-66-H/Br-X.

  下载: 全尺寸图片 幻灯片
  

  Figure 6  Characterization of the Ce-UiO-66-F and Ce-UiO-66-F-X catalyst: (a) TG, (b) 2^nd derive of TG (determine the temperature at which the ligand begins to lose weight), (c) EPR spectra, (d) ¹H NMR spectra, (e) DRS UV–vis spectra and (f) DRIFTS spectra of Ce-UiO-66-F and active Ce-UiO-66-F-X.

  下载: 全尺寸图片 幻灯片
•