Preparation and characterization of Pd NPs supported on MIXMOFs
Compared with MIL-53(Al)-NH2, MIXMOFs 1, 2, and 3 (Scheme 1) are more stable and economical to synthesize owing to the use of cheap terephthalic acid.41 Powder X-ray diffraction (XRD) studies indicate that all three compounds are highly crystalline and of the same structures as in MIL-53(Al) and MIL-53(Al)-NH2 (Figure 1).41, 44, 49 Nitrogen sorption studies show that the Brunauer–Emmett–Teller (BET) surface area of MIL-53(Al) (928 m2 g−1) and MIL-53(Al)-NH2 (654 m2 g−1) are lower than that of 2 (997 m2 g−1; Figure 2). The results demonstrate that the properties of MIXMOFs are not simple linear sums of those of the pure components, and thus support the notion that the sequence of functionalities within MIXMOFs can enhance a specific property.34
Scheme 1. Synthesis of MIXMOFs 1 (x=0.1), 2 (x=0.5), and 3 (x=0.9) based on MIL-53(Al). BDC=benzene-1,4-dicarboxylate, ABDC=2-aminobenzene-1,4-dicarboxylate.
Download figure to PowerPoint
The high surface area, porous structure, and stability in water and common organic solvent make the amine-functionalized MIXMOFs attractive catalyst supports. Recently, Férey and co-workers reported an anion-exchange method for the encapsulation of noble metals over the ethylenediaminegrafted MIL-101(Cr) (Cr3(F,OH)(H2O)2O[(O2C)-C6H4-(CO2)]3⋅nH2O (n≈25)).29 However, it is difficult to control the concentration of the strong base ethylenediamine; an excess amount of strong base may destroy the frameworks. Meanwhile the amine-functionalized MIXMOFs based on MIL-53(Al) can be obtained directly by replacing terephthalic acid with aminoterephthalic acid, and thus avoids introducing the strong base ethylenediamine to the MOFs. Herein, we use the ion-exchange approach for preparation of Pd NPs supported on amine-functionalized MIXMOFs based on MIL-53(Al). The surface amine groups of 1, 2, and 3 (see the Experimental Section for details) were first neutralized by aqueous HCl, followed by ionic reactions of the positively charged ammonium groups on the surface with anionic noble metal salts [PdCl4]2−.25b, 29, 52 Finally, the obtained material containing palladium salts were reduced with NaBH4 at low temperature.
After the loading of palladium, there is no apparent loss of crystallinity according to X-ray diffraction patterns, thus indicating that the framework of 2 is well maintained (Figure 3).52 The characteristic peak of Pd (111) at 2θ=40.1 ° is indistinguishable owing to the low Pd loading and small diameter Pd NPs.15 The Pd/2 exhibits a diminished BET surface area of 982 m2 g−1, whereas that of MIXMOFs 2 is 997 m2 g−1 (Figure 2). The decrease of surface area indicates that the cavities of 2 may be occupied and/or blocked by the palladium NPs located at the surface.15 The TEM images of Pd/2 (Figure 4) show that the Pd NPs are well dispersed on the outer surface of the support and no aggregation is observed, which are similar to the Pd NPs supported on MIL-53(Al)-NH252 and MOF-5.15 The mean diameter of Pd NPs is approximately 3.2 nm (Figure 4), meanwhile the palladium nanoparticles supported on the amino-free MIL-53(Al) aggregate immediately.52 The presence of amino group on the functionalized linker has been proven to be beneficial for the stabilization of Pd species.39 It may not be precluded that there exists coordination interactions between palladium nanoparticles and amino groups. X-ray energy-dispersive spectroscopy (EDS) further confirm the presence of the Pd in the Pd/2 samples (Figure 4). In the X-ray photoelectron spectroscopy (XPS) trace (Figure 5), the 3d5/2 and 3d3/2 peaks of Pd0 appear at 336.0 and 341.3 eV, respectively, and no obvious peak of Pd2+ is observed, thus indicating that palladium is in the reduced form.30
Figure 4. TEM and HRTEM images of a) and b) Pd/2 (0.96 wt % of Pd) before the reaction, c) and after five cycles. d) Size distribution of Pd NPs; mean size 3.2 nm, standard deviation 0.69. Inset in (a) is the EDS pattern of Pd/2.
Download figure to PowerPoint
Heck reactions catalyzed by Pd NPs
The Heck reactions were first performed using a low amount of palladium (a molar ratio: aryl halide/Pd=2000) and an excess of olefin with respect to aryl halide (molar ratio=1.5) to avoid of the formation of homocoupling by-products.3 Pure coupling products were obtained by column chromatography and characterized by NMR spectroscopy. The compounds were selectively obtained with trans configuration.
Initially, we used Pd/MIL-53(Al)-NH252 to investigate the effects of base and solvent for the Heck reaction of brombenzene and styrene at 120 °C (Table 1). The results showed that organic base triethylamine gave higher yields than that of inorganic bases such as Na2CO3, K3PO4, and NaOAc (Table 1, entries 1–4). No product was obtained in the absence of the base (Table 1, entry 5). The solvent also had a significant influence on the catalyst activity. Compared with NMP and o-xylene (Table 1, entries 6 and 7), higher TOF ((mol of prod) per (mol of Pd per hour)) was obtained in DMF (Table 1, entry 4). It was noted that no Pd black was observed under such rigorous conditions during the reaction. Therefore, the amino-stabilized Pd NPs are air-stable up to 120 °C and their catalytic activity is virtually independent of the presence of air. The amount of catalyst is also very important for the reaction. The reaction could not happen in the presence of only MIL-53(Al)-NH2. Although the use of 0.005 mol % and 0.01 mol % of Pd catalyst gave higher TOF, it gave only 28 % (Table 1, entry 10) and 59 % (Table 1, entry 9) yield of the desired products, respectively. It is very interesting that the bromobenzene can be activated effectively by only 0.05 mol % of Pd catalyst and very high yield was obtained (81 %; Table 1, entry 4). The product yield was only 83 % upon increasing the Pd catalyst to 0.1 mol % (Table 1, entry 11).
Table 1. Influences of the base and solvent on the Heck reaction of brombenzene (1 mmol) with styrene (1.5 mmol).[a]
|Entry||Base||Solvent||Yield [%][b]||TOF [h−1]|
Because the type of support is important for Pd NPs catalysis, Pd NPs supported on different supports (MIL-53(Al), 1, 2, and 3) as well as commercially available Pd/C were used in Heck reactions (Table 2). Notably, no product was obtained without the use of Pd catalyst (Table 2, entry 7). The Pd NPs supported on amine-functionalized MOFs (1, 2, 3, and MIL-53(Al)-NH2) showed significantly higher reactivity (up to 76 % yield for the brombenzene substrates) than that of Pd/MIL-53(Al) and Pd/C. These results might be attributed to the Pd NPs agglomeration (Figure S1).29 Interestingly, the different amount of amines in the supports also affected the activity of the Pd catalyst. By increasing of the amount of the amine groups, the activity of the catalyst increases first (up to 93 % yield with Pd/2, 50 % NH2; Table 2, entry 3), and then declines (81 % yield with MIL-53(Al)-NH2; Table 2, entry 5). It may not preclude that there exists coordination interactions between palladium and amino groups.39, 52 A certain number of amino groups on the functionalized linker proved to be beneficial for the immobilization of Pd species, which prevent the agglomeration of Pd NPs. However, the strong interaction between the excess amine groups and Pd species may lead to weakening of the palladium activation for the substrate.53, 54 It may be that the presence of distinct sequences of functionalities in the MIXMOFs leads to a complex chemical environment for catalysis.34
Table 2. Influence of the supports on the Heck reaction of brombenzene (1 mmol) with styrene (1.5 mmol).[a]
|Entry||Catalyst[b]||Yield [%][c]||TOF [h−1]|
|2||1, Pd/MIL-53(Al) and 10 % NH2||76||254|
|3||2, Pd/MIL-53(Al) and 50 % NH2||93||310|
|4||3, Pd/MIL-53(Al) and 90 % NH2||86||287|
Next, Pd/2 as the most efficient catalyst was applied to examine the scope of different substrates in the Heck reaction. A variety of aryl halides were coupled with different olefins in the presence of Pd at low loading (0.05 mol % of Pd)6 under the optimized reaction conditions (Table 3). The Heck reaction of styrene with iodobenzene proceeded easily at 120 °C resulting in trans-stilbene in excellent yields after 30 minutes. Owing to the higher bond energy of the CBr and CCl bonds, aryl bromide and chlorobenzene compounds are difficult to activate. Therefore, the reactions of aryl bromides (or chlorobenzene) with different olefins required harsh conditions and extended reaction times. Interestingly, as seen from Table 3, the Heck reactions of a variety of aryl bromide derivatives with styrene could also proceeded smoothly at 120 °C and gave the corresponding products in high yields after 6 h (Table 3, entries 2, 7–13). The electron-withdrawing groups (-NO2, -CN, -CF3, and -COCH3) in the para position relative to bromine can weaken the CBr bond,3 so these substituted bromobenzene compounds reacted more rapidly than bromobenzene (Table 3, entries 7–10). Reactions between styrene and a variety of electron-rich substrates such as 4-methy- and 4-methoxyl-substituted aryl bromides, also proceeded smoothly and gave the coupling products in good yields (Table 3, entries 11 and 12). Notably, the electron-donating and sterically hindered 2-methoxyl-substituted bromobenzene also gave good yields (Table 3, entry 13). However, the most challenging, yet readily accessible, chlorobenzene had much lower reactivity even after 24 hours (Table 3, entry 3). In addition, various olefin derivatives with different ester substituents were applied to this reaction (Table 3, entries 4–6). It was found that the cross-coupling of bromobenzene with these ester-substituted olefins afforded the corresponding trans-configured products in excellent yields. Interestingly, the efficiency of the tert-butyl acetate-substituted olefin was not affected by its steric hindrance and gave high yield (Table 3, entry 6).
Table 3. Results for the Heck reaction catalyzed by Pd/2.[a]
|Entry||R1||X||R2||Yield [%][b]||TOF [h−1]|
Notably, one of the disadvantages related to the use of supported Pd catalysts in the Heck reaction is leaching of the palladium into the solution, and results in contamination of the product with metal that is difficult to remove along with loss of the expensive Pd catalyst.3–5, 55 Therefore, the leaching of the metal from Pd/2 catalyst was examined. After the workup, inductively coupled plasma (ICP) analysis showed that the amount of Pd leached into the reaction mixture was very low (0.1 ppm). Pleasingly, the low amount of Pd satisfies specifications required by the pharmaceutical industry regarding the final purity of the products (Pd <2 ppm).3 A hot-filtration experiment clearly indicated that the reaction stopped when the filtrate was further subjected to the same reaction (Table 3, entry 14).
Advantages of the Pd/2 material were that the catalytic reactions could be conveniently carried out in air, and the separation of this heterogeneous catalyst could be achieved easily by filtration. Less active 4-methoxy-1-bromobenzene and styrene were used for reusability experiments. The results indicate that after being recycled for five runs, the Pd catalyst still showed a remarkable activity (a reduction from 81 % to 76 % yield; Table 4). The TEM image of the reused catalyst (Figure 5) reveals that the mean diameter of the nanoparticles is 3.4 nm, which is very similar to that before the catalysis reaction. The results illustrate that the loss of palladium active sites is negligible, thus resulting in the preservation of catalytic activity.30 Moreover, the Pd catalysts supported on amine-functionalized MOFs was comparably active after being stored in air for three months (Table 3, entry 15). These features are obvious improvements from existing air- and moisture-sensitive homogeneous palladium-phosphine catalysts.10
Table 4. Reusability of the Pd/2 catalyst in the Heck reaction of 1-bromo-4-methoxyenzene and styrene[a].
|Cycle||Yield [%][b]||Cycle||Yield [%][b]|
|3rd||82|| || |