Cerium-based metal organic frameworks with UiO-66 architecture: synthesis, properties and redox catalytic activity †

A series of nine Ce( IV )-based metal organic frameworks with the UiO-66 structure containing linker molecules of diﬀerent sizes and functionalities were obtained under mild synthesis conditions and short reaction times. Thermal and chemical stabilities were determined and a Ce-UiO-66-BDC/TEMPO system was successfully employed for the aerobic oxidation of benzyl alcohol

Here we report our successful determination of the conditions to stabilize the [Ce 6    All compounds were obtained as microcrystalline powders; therefore structures were confirmed from PXRD data. The structures of Ce-UiO-66-Fum and -BDC were confirmed by Rietveld refinement (Table 1, Figure S1-2). Ce-UiO-66-BDC 10 and -Fum exhibit structures isoreticular with their Zr analogues 6,14 and crystallize in the space group ‫3݉ܨ‬ ത ݉ and ܲ݊3 ത , respectively ( Figure 1 and S5). In Ce-UiO-66-BDC the [Ce 6 O 4 (OH) 4 ] 12+ clusters are organized in a cubic close-packed arrangement and bridged by twelve different BDC 2molecules, to 15 give the structural formula [Ce 6 O 4 (OH) 4 (BDC) 6 ] ( Figure S3). SEM measurements of the particle morphology of Ce-UiO-66-BDC showed that the compound forms as agglomerates of mostly spherical particles, with diameters in the range of 100-500 nm ( Figure S4). Unit cells of Ce-UiO-66-NDC and -BPDC were 20 confirmed by Le Bail profile fitting ( Figure S6-7). Crystallographic details for all four compounds are given in the Supporting Information (Table S1).
The oxidation state of cerium in Ce-UiO-66-BDC was determined by XANES spectroscopy (Figure 2). L III -edge 25 XANES features typical for Ce(III) are significantly different from those of Ce(IV). Ce (III) displays a very intense single peak (5726 eV), whereas Ce(IV) exhibits two well-separated maxima of lower intensity (5729 and 5739 eV). 31 The XANES spectra unambiguously demonstrate that Ce-UiO-66-BDC contains 30 Ce(IV) without any detectable trace of Ce(III).
The synthesis of functionalized phase-pure samples of Ce-UiO-66-BDC-X (X=F, CH 3 , Cl, NO 2 , COOH) was accomplished under identical synthetic conditions as those used for the Ce-UiO-66-BDC. The PXRD patterns and the lattice parameters as 35 determined by Le Bail fitting in the space group ‫3݉ܨ‬ ത ݉ are presented in Figure S8-S13 and Table S2.
Chemical stability of Ce-UiO-66-BDC was proven by stirring in different solvents for 24 h at room temperature. The compound is stable in a variety of organic solvents as well as in water, 40 although some peak broadening was observed. Ce-UiO-66-BDC decomposes only in acidic (2 M HCl) and basic (2 M NaOH) media ( Figure S14). Ce-UiO-66-Fum is similarly stable in organic solvents, though it is more susceptible to degradation ( Figure S15). The longer linker containing Ce-UiO-66-NDC and 45 -BPDC are stable in aprotic organic solvents. However in water, ethanol and under air both compounds show a slow continuous loss of intensity in the diffraction pattern. Interestingly, it was possible to recover the crystallinity of the compounds by heating in 1 ml DMF for 5 min at 100 °C, with crystallinity confirmed by 50 PXRD measurements (Figure S16-S17). Ce-UiO-66-BDC collapses on heating above 300 °C, with a 55 weight loss of 30.3 wt%. This framework collapse is clearly observed in the VT-PXRD data ( Figure S19 and S20), with few changes occurring in the diffraction patterns over the range 40-240 °C; from 320-520 °C the reflections broaden dramatically to result in a rather amorphous final product. The observed weight 60 loss of framework collapse is 4 wt% lower than expected (expected 34.2 wt%); this discrepancy is attributed to structural defects, arising from missing BDC linker molecules, as previously reported for Zr containing UiO-66. 7,10,11 Based on the TGA results, it is assumed that on average the [Ce 6 O 4 (OH) 4 ] 12+ 65 clusters are coordinated by 11 linkers instead of 12.
Ce-UiO-66-Fum was also studied by TGA and shows a similar pattern of weight losses ( Figure S21). Decomposition occurs at a significantly lower temperature than for the Ce-UiO-66-BDC compound and also the reported Zr-Fum. 14 TGA studies have also 70 been performed on the functionalized Ce-UiO-66-BDC-X compounds. The -NO 2 functionalized compound is approximately as stable as the analogous unfunctionalized compound ( Figure  S22), however compounds bearing other functionalities showed significantly lower stability. 75 Solution 1 H-NMR was used to confirm the incorporation of the functionalized terephthalate linkers without modification to their functional groups ( Figure S25-31). N 2 sorption measurements were performed to evaluate the porosity of the Ce-UiO-66 compounds. The activation temperature and results are presented 80 in Table 2 and Figure S32-33. All sorption isotherms show the characteristic Type I adsorption isotherm curve shape. 32 Ce-UiO-66-BDC has a specific BET surface area of 1282 m 2 g -1 . We expect Ce-UiO-66-BDC to exhibit a smaller specific BET surface area than Zr-UiO-66, since Ce is approx. 50 % heavier than Zr. 85 Thus, compared with a specific surface area for defect rich Zr-UiO-66 reported as 1580 m 2 g -1 , 9 it seems likely that the relatively high surface area of Ce-UiO-66-BDC results from the presence of missing linker molecules, in agreement with the observations from the TGA study. Ce-UiO-66-Fum demonstrates a specific BET surface area of 732 m 2 g -1 . This value is in accordance with the BET surface area reported for the analogous Zr-fumarate (856 m 2 g -1 ). 14 Sorption isotherms for Ce-UiO-66-NDC and -BPDC could not be measured as both compounds decompose 5 during activation under reduced pressure, even at temperatures as low as 100 °C. The N 2 sorption isotherms of the functionalized Ce-UiO-66-BDC-X (X= F, CH 3 , Cl, NO 2 ) show a decrease in the specific BET surface area with increasing weight and size of the functional group (Table S3 and Figure S33). For Ce-UiO-66- 10 BDC-COOH no N 2 sorption isotherm could be obtained, because the compound decomposes during activation at 100 °C.  15 PXRD patterns collected after the N 2 sorption experiments indicate that all other samples remain intact after activation, although for Ce-UiO-66-Fum and -BDC-Cl some peak broadening was observed ( Figure S34). Given the well-known redox chemistry of cerium oxides, 33 and 20 more specifically the previously reported stoichiometric oxidation of 1,4-benzenediol with a similar hexanuclear Ce-benzoate cluster, 28 we tested Ce-UiO-66-BDC as a catalyst in the aerobic oxidation of benzyl alcohol (Scheme 1, Table 3).
Scheme 1 Aerobic oxidation of benzyl alcohol. 25 Using only Ce-UiO-66-BDC activated at 180 °C, a modest yield of 8 % benzaldehyde was achieved, which is significantly more than for the uncatalyzed blank reaction (2 %) or for the reaction employing nanoparticulate CeO 2 (7 %). Based on existing literature combining ((NH 4 ) 2 Ce(NO 3 ) 6 ) and TEMPO 30 (2,2,6,6-tetramethylpiperidin-1-yl)oxyl) as a co-catalyst, 34 we devised an analogous system with Ce-UiO-66-BDC. Addition of TEMPO to Ce-UiO-66-BDC resulted in a benzyl alcohol conversion of 29 %. Upon raising the activation temperature of the framework to 220 °C, a strong increase in activity was 35 observed, with 88 % conversion of benzyl alcohol and complete selectivity to benzaldehyde. No benzoic acid formation was found via GC-MS of the silylated reaction mixture. The marked influence of the activation temperature is attributed to the removal of strongly adsorbed guest molecules and possible 40 cluster dehydration as evidenced from the TGA data ( Figure  S18), creating open coordination sites analogous to the situation in Zr-UiO-66-BDC. A reaction using only TEMPO as catalyst resulted in a benzaldehyde yield of 7 %, clearly indicating a synergetic effect between Ce-UiO-66-BDC and TEMPO. 45 Strikingly, such a synergism is not observed at all between Zr-UiO-66-BDC and TEMPO (see Table 3). Finally, Ce-UiO-66-BDC proved stable under the applied reaction conditions, as evidenced by PXRD. ICP analysis determined the amount of Ce in solution to approximately 2 ppm, which together with a hot-50 filtration test further proves the heterogenous nature of the reaction ( Figure S39-40).
To gain some mechanistic insight, the reaction was performed in absence of O 2 (N 2 atmosphere), and only a low benzaldehyde yield of 11 % was found. In such conditions, a buildup of 1-55 hydroxy-2,2,6,6-tetramethylpiperidine (TEMPOH), the reduction product of TEMPO, was detected by GC-MS, with ~40 % of the original TEMPO being converted to TEMPOH ( Figure S35-36). 35 Under O 2 , only 3 % of the initial TEMPO is found as TEMPOH, as the latter is prone to a fast reoxidation to TEMPO. From these 60 observations, we propose a basic catalytic cycle ( Figure S37). First, TEMPO undergoes a one-electron oxidation at the surface of Ce-UiO-66-BDC to form its oxoammonium counterpart, with a concomitant reduction of Ce 4+ to Ce 3+ . Due to its size, TEMPO is unable to enter the pores of Ce-UiO-66-BDC, as was verified 65 by additional adsorption experiments ( Figure S38), leaving only Ce 4+ close to the particle surface available for oxidation. The oxoammonium species reacts with the alcohol to form the aldehyde while being reduced to TEMPOH. 36 The latter spontaneously oxidizes back to TEMPO under O 2 , but could 70 alternatively react with an oxoammonium cation to form two TEMPO molecules. Finally, we hypothesize that reoxidation of Ce 3+ by dioxygen regenerates the MOF co-catalyst. Using longer linkers is a viable option to increase the catalytic activity by allowing the reactants access to the internal pore voids. This is 75 shown by the conversion increase from 29 to 80 % for respectively BDC and 2,6-naphthalenedicarboxylate based materials, both activated in air at 180 °C.  90 Ce 4+ ions. Ce-UiO-66-BDC shows the highest chemical and thermal stability and initial experiments revealed it can be used as a co-catalyst with TEMPO in alcohol oxidations. Further investigations are currently carried out to extend the number of Ce-based MOFs to other topologies, to get a deeper understanding of the catalytic process and to study the properties 5 of the Ce-UiO-66-type MOFs in other catalytic reactions.
We acknowledge the support of Bordiga, Braglia, Bouchevreau, Lamberti, and Lillerud, for the collection of the XANES spectra in Lund. The travel to Lund of Lamberti and Lomachenko was supported by the Russian Mega-grant No. 10 14.Y26.31.0001.