General Immobilization of Ultrafine Alloyed Nanoparticles within Metal-Organic Frameworks with High Loadings for Advanced Synergetic Catalysis.

The development of a general synthesis approach for creating fine alloyed nanoparticles (NPs) in the pores of metal-organic frameworks (MOFs) shows great promise for advanced synergetic catalysis but has not been realized so far. Herein, for the first time we proposed a facile and general strategy to immobilize ultrafine alloyed NPs within the pores of an MOF by the galvanic replacement of transition-metal NPs (e.g., Cu, Co, and Ni) with noble-metal ions (e.g., Pd, Ru, and Pt) under high-intensity ultrasound irradiation. Nine types of bimetallic alloyed NPs of base and noble metals were successfully prepared and immobilized in the pores of MIL-101 as a model host, which showed highly dispersed and well-alloyed properties with average particle sizes ranging from 1.1 to 2.2 nm and high loadings of up to 10.4 wt %. Benefiting from the ultrafine particle size and high dispersity of Cu-Pd NPs and especially the positive synergy between Cu and Pd metals, the optimized Cu-Pd@MIL-101 exhibited an extremely high activity for the homocoupling reaction of phenylacetylene under unprecedented base- and additive-free conditions and room temperature, affording at least 19 times higher yield (98%) of 1,4-diphenylbuta-1,3-diyne than its monometallic counterparts. This general strategy for preparing various MOF-immobilized alloyed NPs potentially paves the way for the development of highly active metal catalysts for a variety of reactions.

Since the concept of alloyed nanoparticles (NPs) catalysts was proposed in the 1977s,[1] alloyed NPs have always attracted considerable attention in both industrial and fundamental research due to their unique magnetic, optical, and catalytic properties in a wide range of applications, which normally cannot be achieved by their corresponding monometallic NPs.[2−4] In particular, alloying noble metals with low-cost transition metals, such as Cu, Co, or Ni, can not only reduce the usage of expensive noble metals, but also enable the tailoring of atomic and electronic structures as well as available active sites for synergetic catalysis.[5−7] Meanwhile, in addition to the chemical activity of each metal, the preferred catalytic properties of these alloyed NPs are also highly determined by their dispersity, stability, and especially particle size.[8,9] Since small alloyed NPs often suffer from inhomogeneous alloying and serious aggregation due to their high surface energy and the difficulty of assembling metals with very different physicochemical properties into one NP,[10] the synthesis of ultrafine well-alloyed NPs to maximize their catalytic efficiency is one of the primary goals in modern catalysis.[11−14]

In this regard, a most common and efficient approach is to immobilize metal NPs within a special porous material, where the strong interaction and nanoconfinement effect can stabilize these metal NPs and prevent them from aggregation during both synthesis and catalytic reaction processes.[15,16] However, the metal NPs of supported catalysts prepared by conventional methods (such as impregnation,[17] coprecipitation,[18,19] and deposition precipitation[20,21]) still suffer from many disadvantages such as large particle sizes, irregular size distributions, weak affinities to supports, and poor alloying when involving multiple metals.[22] Hence, there is no doubt that the development of an efficient strategy to prepare highly dispersed, ultrafine, and well-alloyed NPs immobilized in a well-chosen porous material is of great importance especially for catalytic applications.

Recently, metal–organic frameworks (MOFs), a new class of porous materials with a number of unique properties including high surface areas, uniform pore sizes, and tailorable structures,[23−29] have emerged as promising hosts for the immobilization of metal NPs.[30−33] Similar to mesoporous silica and zeolite, the well-defined pore structures of MOFs can be used as good stabilizers and homogenizers to limit the growth and/or aggregation of metal NPs.[34−38] Thus, encapsulation of various metal NPs into the pores of MOFs and their catalytic performance have been of great interest during the past decade.[30,39−42] Some examples are the incorporation of Cu, Ru, Pd, Pt, Au, Ag, and Ir NPs into different MOF cavities by using various strategies, such as incipient wetness impregnation, the double-solvent method, solid grinding, and chemical vapor deposition.[31] However, to the best of our knowledge, there are only very few examples of MOF-incorporated alloyed NPs with controlled sizes, and so far the development of a general approach for immobilizing various alloyed NPs into the MOF pores with high loadings and narrow size distributions around 2 nm is still a great challenge.[11,43,44]

Results and Discussion

Herein, we propose a general synthesis strategy to immobilize ultrafine alloyed NPs of transition and noble metals within the framework of an MOF. Our approach relies on the galvanic replacement of base metal NPs (e.g., Cu, Co, and Ni) with noble metal ions (e.g., Pd, Ru, and Pt) within an MOF structure under high-intensity ultrasound irradiation. We choose MIL-101 as a model host to demonstrate our method, and nine types of bimetallic alloyed NPs have been successfully prepared and immobilized within its pores, all of which show highly dispersed and well-alloyed properties with average particle sizes from 1.1 to 2.2 nm and high loadings of up to 10.4 wt %. As a proof of concept, the synergetic catalytic performance of Cu–Pd@MIL-101 is demonstrated by its much higher catalytic activity for the homocoupling reaction of phenylacetylene under an unprecedented base- and additive-free condition and room temperature as compared to its monometallic counterparts and other alloyed catalysts.

The synthesis procedure of ultrafine Cu–Pd NPs immobilized within MIL-101 (denoted as Cu–Pd@MIL-101) is illustrated in Figure A. MIL-101, a chromium-based MOF with constitutional formula of Cr3F(H2O)2O[(O2C)C6H4(CO2)]3·nH2O (n ≈ 25), has been used as a host matrix to immobilize alloyed NPs owing to its large surface area (the Langmuir surface area can reach 5900 m2 g–1), high mechanical stability, and large pore sizes (with diameters of 2.9 and 3.4 nm) and window sizes (two pore windows of 1.2 and 1.6 nm), which provide good opportunities to disperse various metal NPs for efficient catalysis.[5,45−48] Briefly, Cu2+ was first incorporated into MIL-101 via an impregnation method and in situ reduced by NaBH4 to obtain Cu/MIL-101. Then, the Cu NPs (the relative reduction potential of Cu2+/Cu is 0.34 V vs NHE) in MIL-101 were efficiently transformed into highly dispersed, ultrafine, and MIL-101-anchored Cu–Pd NPs through their galvanic replacement reaction with Pd2+ (Cu0 + Pd2+ → Cu2+ + Pd, the relative reduction potential of Pd2+/Pd is 0.92 V vs NHE) under high-intensity ultrasound irradiation (Figure B). We first examined the effect of Cu loadings (Cu/MIL-101) and the ultrasound intensity on the properties of the resultant Cu–Pd NPs. As shown in Figure S1, the average size of the resultant Cu–Pd NPs decreases from 3.98 to 2.16 nm when the relative ultrasound intensity is increased from 78% to 88%. However, further increasing the relative ultrasound intensity to 98% produces no obvious change in particle size of the resultant Cu–Pd NPs. Therefore, we choose 88% as the optimal relative ultrasound intensity for the following investigations. Besides, the loading of Cu NPs has also a great effect on the sizes of Cu–Pd NPs and total metal contents. For example, with the Cu loading increasing from 3.3 wt % to 5.7 wt %, the resultant Cu–Pd NPs show similar particle sizes, but the total metal loading of Cu–Pd NPs in MIL-101 increases from 3.9 wt % to 7.6 wt % (Figure S2 and Table S1). Further increasing the Cu loading to 8.7 wt %, the resultant Cu–Pd NPs start to aggregate, leading to nonuniform NPs that are poorly dispersed in MIL-101 (Figure S2). Therefore, 5.7 wt % is the most favorable Cu content to obtain desirable Cu–Pd@MIL-101 with both high dispersity and ultrafine particle size. Note that the Cu loading decreases from 5.7 wt % for Cu/MIL-101 to 2.5 wt % for Cu–Pd/MIL-101, while the Pd loading of Cu–Pd@MIL-101 is detected to be 5.1 wt % (Table S1). These results correlate well with the galvanic replacement reaction between Cu0 and Pd2+, implying that almost all Cu–Pd NPs are successfully immobilized within the MIL-101 structure under high-intensity ultrasound irradiation.

Figure 1

Schematic illustration of the immobilization of Cu–Pd alloyed NPs within MIL-101. (A) Synthesis route of Cu–Pd@MIL-101. (B) Sonochemical system used in this work to prepare Cu–Pd@MIL-101.

Transmission electron microscopy (TEM) and high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) were performed on Cu/MIL-101, Cu–Pd@MIL-101 and an ultrathin cut from Cu–Pd@MIL-101 (Figures , S3, and S4). The typical TEM (Figure A) and HAADF-STEM (Figure B) images of Cu/MIL-101 reveal that the Cu NPs are unevenly distributed on MIL-101. The statistic histogram for 100 randomly selected particles in Figure A shows that the average size of these Cu NPs is 8.25 ± 2.53 nm with a broad size distribution (inset of Figure A). In sharp contrast, the resultant Cu–Pd NPs are uniformly immobilized in the MIL-101 framework with a very narrow size distribution and a small average size of 2.16 ± 0.57 nm (Figure C and its inset). The corresponding high-resolution TEM (HRTEM) image (Figure D) shows the well-resolved lattice fringes with an interplane distance of about 0.21 nm, which can be attributed to the (111) plane of the face-centered cubic (fcc) structure of Cu–Pd.[49] The energy-dispersive X-ray spectroscopy (EDX) line-scan profile (Figure E and its inset) across a single Cu–Pd NP reveals the similar distribution of Cu and Pd elements in NPs, directly confirming the well-alloyed structure. The HAADF-STEM and TEM images of a representative ultrathin slice from Cu–Pd@MIL-101 clearly confirm the presence of well-dispersed Cu–Pd NPs immobilized within the MIL-101 framework with an average size of about 2 nm (Figure F–H and S4). The corresponding STEM-EDX elemental mapping images (Figure I–L) indicate that all the Cr, Cu, and Pd are homogeneously dispersed in Cu–Pd@MIL-101, which again supports the excellent dispersity and well-alloyed merit of Cu–Pd NPs in MIL-101. These results prove the successful immobilization of highly dispersed Cu–Pd NPs into MOFs with high loadings.

Figure 2

Characterization of MIL-101-immobilized Cu–Pd NPs. (A, B) TEM and HAADF-STEM images of Cu/MIL-101. (C) HAADF-STEM image of Cu–Pd@MIL-101. (D) HRTEM image of an individual Cu–Pd NP. (E) HAADF-STEM image of Cu–Pd@MIL-101. (F–H) TEM and HAADF-STEM images of an ultrathin cut from Cu–Pd@MIL-101 and (I–L) its corresponding elemental mappings of Cr, Cu, Pd, and Cu + Pd. Insets in (A, C) are the corresponding particle size distribution histograms; inset in (E) is the elemental line-scan profiles of a selected Cu–Pd NP in E.

Powder X-ray diffraction (PXRD) analysis was then used to prove the preservation of MIL-101 structure after the intense ultrasound process. As shown in Figure A, there is no obvious crystallinity loss for Cu/MIL-101 and Cu–Pd@MIL-101 after loading Cu NPs and Cu–Pd NPs as proven by their PXRD patterns similar to that of MIL-101. However, new weak diffraction peaks at 2θ = 36.5°, 42.3°, and 43.6° were observed for Cu/MIL-101 (inset of Figure A). The weak peak located at 2θ = 43.6° is indexed to the (111) plane of Cu (JCPDS file 65-9743), while the peaks at 2θ = 36.5° and 42.3° can be assigned to the (111) and (200) planes of Cu2O (JCPDS file 65-3288), which can be attributed to the easy oxidation of Cu0 to Cu+ by air in the preparation and storage processes. The PXRD pattern of Cu–Pd@MIL-101 shows a feeble peak at around 2θ = 40.9°, which can be indexed to (111) diffraction peak of fcc structure of Cu–Pd (JCPDS file 48-1551), consistent with the measured interplane distance of lattice fringes in HRTEM image (Figure D). It is noteworthy that although the loading of Cu–Pd NPs is as high as 7.6 wt %, the XRD diffraction peak of Cu–Pd alloyed phase is still very weak and wide, further confirming the ultrafine particle size of Cu–Pd NPs in Cu–Pd@MIL-101.

Figure 3

Confirming the structure properties of Cu–Pd@MIL-101. (A) PXRD patterns and (B) N2 adsorption-desorption isotherms of MIL-101, Cu/MIL-101, and Cu–Pd@MIL-101. (C) the Pd 3d region of the XPS spectra of Pd/MIL-101 and Cu–Pd@MIL-101. (D) the Cu 2p region of the XPS spectra of Cu/MIL-101 and Cu–Pd@MIL-101.

The pore characters of various samples were measured by N2 adsorption and desorption at 77 K. As expected, the BET surface areas and pore volumes for both Cu/MIL-101 (2174 m2 g–1 and 1.01 cm3 g–1) and Cu–Pd@MIL-101 (2399 m2 g–1 and 1.12 cm3 g–1) are diminished as compared to those of parent MIL-101 (2714 m2 g–1 and 1.29 cm3 g–1), which is caused by the high loadings and pore occupation of the immobilized Cu or Cu–Pd NPs (Figure B, Tables S2 and S3). In addition, a slight decrease in the pore diameter was observed as compared with the parent MIL-101 after the loading of Cu NPs or Cu–Pd NPs (Figure S5). Interestingly, Cu–Pd@MIL-101 shows a higher BET surface area and pore volume relative to Cu/MIL-101, indicating that the ultrafine particle size of Cu–Pd NPs would be more beneficial for facilitating the mass transport, since the average particle size of Cu–Pd NPs (2.16 ± 0.57 nm) is smaller than the inherent pore sizes of MIL-101 (with diameters of 2.9 and 3.4 nm). The electronic properties of Cu–Pd@MIL-101 were then investigated by X-ray photoelectron spectroscopy (XPS). As shown in Figure C, Cu–Pd@MIL-101 exhibits two prominent bands at 342.9 eV for Pd 3d3/2 and 337.7 eV for Pd 3d5/2 in its Pd 3d spectrum, both of which are close to but higher by 1.6 eV as compared with those of monometallic Pd/MIL-101 synthesized by a conventional impregnation method, revealing that the Pd exists in its metallic state in Cu–Pd@MIL-101. The increased binding energy of Pd0 for Cu–Pd@MIL-101 suggests the electron transfer from Pd atoms to Cu atoms, which can be indicative of the formation of Cu–Pd alloyed NPs and the existence of strong metallic synergic effects.[50] The Cu 2p spectrum of Cu–Pd@MIL-101 demonstrates that the Cu element in this sample mainly exists in the forms of Cu0 and Cu2+ (Figure D).[51] The presence of Cu2+ is due to the inevitable oxidation of Cu0 to CuO by air in the preparation and storage processes.[13] Besides, the Cu+ in Cu–Pd@MIL-101 that has been detected by XRD (Figure A) cannot be distinguished from Cu0 by XPS because of their overlapping Cu 2p spectra.[14] Furthermore, a similar peak shift of about 1.0 eV but to lower binding energy was also observed for the Cu0 2p1/2 of Cu–Pd@MIL-101 as compared to that of Cu/MIL-101, further confirming the electron transfer from Pd to Cu and thus the successful formation of Cu–Pd alloyed NPs.[52]

The above results confirm that our strategy can efficiently fabricate ultrafine and highly dispersed alloyed NPs within an MOF. Next, to further understand the evolution of the alloyed NPs in our system, a time-dependent experiment was conducted under the same condition but with a different ultrasound irradiation time. The representative intermediates at five different stages were then examined by TEM and HAADF-STEM. As shown in Figures A1–C1 and S6, the Cu NPs were transformed immediately into core–shell Cu@Pd NPs in the first 30 s of ultrasound irradiation due to the quick galvanic displacement of Cu0 atoms on the surface of Cu NPs by Pd2+ via a similar Kirkendall diffusion mechanism.[53] After 10 min of irradiation time, the core–shell Cu@Pd NPs were already transformed into well-alloyed Cu–Pd NPs as revealed by the similar element distributions of Pd and Cu in a selected NP, but most of them were agglomerated with a poor dispersity (Figure A2–C2). Interestingly, when the irradiation time was further extended to 20 min and then to 30 min, all large and agglomerated NPs gradually disappeared, and the number of ultrafine Cu–Pd NPs with a mean size of ca. 2.1 nm increased sharply with the size distribution also becoming homogeneous (Figure A3–C3, A4–C4). Meanwhile, the Pd and the total metal content of alloyed NPs increased with the displacement and decrease of Cu0 as suggested by the elemental line-scan curves and atomic absorption spectroscopy (AAS) results (Table S4 and Figure S7). As expected, further increasing the irradiation time to 45 min produced no obvious changes in the metal composition, particle size, and element distribution of the alloyed Cu–Pd NPs. For comparison, we also performed the same preparation procedure as Cu–Pd@MIL-101 but without high-intensity ultrasound irradiation. As shown in Figures S8A and S8B, although ultrafine Cu–Pd NPs with a mean size of ca. 2 nm can also be obtained, most of them are agglomerated into large particles that are poorly dispersive. These results confirm the imperative role of ultrasound irradiation treatment in dispersing the ultrafine Cu–Pd NPs produced by the galvanic replacement reaction.

Figure 4

Influence of ultrasound irradiation time on the properties of the resultant Cu–Pd NPs immobilized within MIL-101. (A) TEM images and (B) HAADF-STEM images of various catalysts. (C) Cu–Pd particle size distribution histograms of Cu–Pd@MIL-101 obtained at five different times: 1, 30 s; 2, 10 min; 3, 20 min; 4, 30 min; 5, 45 min. Insets in (B) are the elemental line-scan curves of a selected Cu–Pd NP from B.

More interestingly, our strategy can be further extended to the preparation of a variety of other MIL-101-immobilized alloyed NPs by simply changing CuCl2·2H2O and PdCl2 to other transition-metal and noble-metal precursors, respectively. For example, the galvanic replacement of Cu, Co, or Ni NPs with Pd2+, Ru3+, or Pt2+ ions can produce eight other types of bimetallic alloyed NPs (Table S5), which are immobilized within MIL-101 (Figure and S9; see the Supporting Information for details). As shown in Figure S10, all of monometallic Cu, Co, and Ni NPs are distributed on MIL-101 with large average particle sizes and wide particle size distributions. Surprisingly, all the resultant bimetallic alloyed NPs show highly dispersed and well-alloyed properties with narrow particle size distributions and small average particle sizes ranging from 1.1 to 2.0 nm as revealed by their TEM, STEM, and HRTEM images, elemental line-scan profiles, particle size distribution histograms, and XRD patterns (Figures , S11–S18, and S19A–C). The XPS results confirm that all noble metals (Pd, Ru, or Pt) exist in their metallic states in these bimetallic alloyed NPs (Figure S19D–F), in which strong intermetallic synergic effects can be reflected by the increased binding energies of M0 (M = Pd, Ru, or Pt) relative to their corresponding monometallic counterparts. More importantly, all of these alloyed NPs have high metal loadings on MIL-101 (>8 wt %), and the NPs can still exhibit high dispersity and ultrafine size even when the metal loading is up to 10.4 wt % for the case of Ni–Pt@MIL-101. It is also worth noting that such high loadings of ultrafine alloyed NPs without aggregation in our study have rarely been observed in other previous literature studies,[11,43,44] directly confirming the efficient confinement effect of MIL-101 to prevent alloyed NPs from aggregation. In sharp contrast, all of the M-N-MIL-101 prepared without ultrasound irradiation show very broad particle size distributions with large particle sizes up to dozens of nanometers (Figure S8). Furthermore, we also synthesized a series of Cu–Pd@MIL-101 with different Pd/Cu molar ratios as examples to further explore the tunability of our strategy. As shown in Figure S20, it is obvious that the Pd/Cu molar ratios of Cu–Pd@MIL-101 can be flexibly tuned by simply changing the dosage of PdCl2 (Figure S20, Table S6). However, it is also worth noting that the particle size of Cu–Pd NPs has been found to slightly increase when the Pd contents were too high or too low. These results undoubtedly demonstrate the good generality and high efficiency of our strategy for immobilizing highly dispersed and well-alloyed NPs within MOFs with ultrafine particle sizes and very high loadings.

Figure 5

Characterization of other M-N@MIL-101: (A) Cu–Ru@MIL-101, (B) Cu–Pt@MIL-101, (C) Co–Pd@MIL-101, (D) Co–Ru@MIL-101, (E) Co–Pt@MIL-101, (F) Ni–Pd@MIL-101, (G) Ni–Ru@MIL-101, and (H) Ni–Pt@MIL-101 (1, HRTEM, 2, HAADF-STEM, and 3, HAADF-STEM images). Insets in (3) are the corresponding elemental line-scan profiles of a selected M-N NP.

Considering their aforementioned properties, these M-N@MIL-101 are supposed to find wide uses in synergetic catalysis. Here, we evaluated their promising application as catalysts for the homocoupling reaction of terminal alkynes to produce 1.3-diynes, which has been identified as important building blocks for medical industry and polymer chemistry.[54,55] In general, this reaction can be efficiently catalyzed by Pd-based and/or Cu-based salts under homogeneous conditions, which usually requires the addition of large amounts of bases and/or additives with attendant difficulties in catalyst recovery and recycling.[56−58]Alternatively, as summarized in Table S7, various solid-supported catalysts have also been applied to this reaction due to their good reusability, enhanced stability, and easy handling. However, various bases and/or other additives are still indispensable in these systems to activate and accelerate the reaction, which not only complicates the subsequent product separation but also easily gives rise to various environmental problems.[59,60] Herein, we proposed that the synergetic catalysis of Cu and Pd metals in ultrafine Cu–Pd alloyed NPs can be used as a solution to all of the above problems. As shown in Figure A, Cu–Pd@MIL-101 is highly active for the homocoupling reaction of phenylacetylene under base- and additive-free conditions and room temperature, affording a 98% yield of 1,4-diphenylbuta-1,3-diyne, which is at least 49 times higher than those of other alloyed catalysts, directly revealing its highly synergetic catalysis for this transformation. Then, we carefully investigated the effect of Pd/Cu molar ratios on the catalytic performance for this reaction. As shown in Figure B and Table S8, the yield of 1,4-diphenylbuta-1,3-diyne exhibits a typical volcano-shaped profile as a function of Pd/Cu molar ratios. When the Pd/Cu ratios increased or decreased from 1.22:1, the yields of 1,4-diphenylbuta-1,3-diyne decreased sharply, directly revealing that the optimum synergistic catalysis can be achieved by using a Pd/Cu molar ratio of 1.22:1 for this transformation. Furthermore, with phenylacetylene as a model substrate, the effect of various solvents on the catalytic performance was investigated, and acetonitrile was proven to be the best one for Cu–Pd@MIL-101 (Table S9). This might be because the polarity of acetonitrile matches well with that of the products and thus drives the reaction forward more efficiently, in accordance with the results over homogeneous catalysts reported previously for this reaction.[61,62] The influence of catalyst dosages on the yields of 1,4-diphenylbuta-1,3-diyne was also evaluated. As shown in Figure C, when the catalyst dosages increase from 1 mol % to 5 mol % (based on Pd), the yields of 1,4-diphenylbuta-1,3-diyne are also rapidly enhanced from 3% to 98% after 48 h of reaction time. However, further increasing the catalyst dosages to 10 mol %, the reaction rate is almost unchanged, and no obvious increase in the yields of 1,4-diphenylbuta-1,3-diyne was observed. That is to say, beyond a certain catalyst loading, further increasing the dosages of catalyst is considered to be unnecessary, similar to the reaction phenomenon reported previously.[63] For comparison, the activities of Cu/MIL-101, Pd/MIL-101, the mixture of Cu/MIL-101 and Pd/MIL-101, and Cu–Pd-MIL-101 synthesized prepared without high-intensity ultrasound irradiation were also measured (Figure D). Clearly, all these catalysts show much lower catalytic activities than Cu–Pd@MIL-101 under the same condition, giving 5%, 0%, 18%, and 20% yields of 1,4-diphenylbuta-1,3-diyne after 48 h of reaction, respectively. Actually, to the best of our knowledge, Cu–Pd@MIL-101 can represent the first example reported so far that can catalyze efficiently this reaction under base- and additive-free conditions and room temperature (Table S7). Considering the same support but different metal NP properties of these catalysts, the superior catalytic activity of Cu–Pd@MIL-101 can be directly attributed to its ultrafine Cu–Pd particle size with highly exposed active sites and the strong synergistic effect of Cu and Pd elements for this reaction. Furthermore, a control experiment by carrying out this reaction under N2 atmosphere shows that only a 9% yield of 1,4-diphenylbuta-1,3-diyne was obtained, suggesting that Cu–Pd@MIL-101 could not achieve a high activity for this reaction in the absence of O2. In addition, after Cu–Pd@MIL-101 was removed from the reaction mixture after 12 h of reaction time, no further homocoupling of phenylacetylene was observed, revealing the heterogeneous feature of Cu–Pd@MIL-101 for this reaction. AAS analysis of the reaction solution revealed that the Cu and Pd contents in the solution were below the detection limit, implying no significant leaching of the NPs occurred during the homocoupling process. On the basis of our results as well as the previously reported mechanism under homogeneous condition, we proposed a plausible mechanism for this reaction over Cu–Pd@MIL-101 (Figure E). As already proven by XRD results (Figure A), the Cu0 in Cu–Pd@MIL-101 can be easily oxidized partly into Cu+ (in the form of Cu2O) when being exposed to air. Thus, the generated Cu+ can react easily with phenylacetylene by deprotonation to generate Cu-phenylacetylene and H+.[64] As also already revealed by the XPS analysis, a strong intermetallic synergetic effect exists between the Pd and Cu atoms in Cu–Pd@MIL-101 with electron transfer from Pd atoms to Cu atoms. Thus, the Pd in Cu–Pd@MIL-101 can play a similar role as previously reported Pd2+ under homogeneous conditions, which can react with double the Cu-phenylacetylene by transmetalation to afford the dialkynylpalladium(II) species.[65] Finally, the reductive elimination of dialkynylpalladium(II) spontaneously produces 1,4-diphenylbuta-1,3-diyne with the O2 and proton also being consumed to generate water to complete the catalytic cycle, since the relative reduction potential of O2/H2O is 1.23 V vs NHE, much larger than that of Pd2+/Pd (0.92 V vs NHE).[66] One of the advantages for most of solid catalysts is their easy separation from the reaction mixture and thus good recyclability. So, the stability and reusability of Cu–Pd@MIL-101 were evaluated for the homocoupling of phenylacetylene. As shown in Figure F, Cu–Pd@MIL-101 can be recovered and reused for at least six runs in the subsequent reaction without significant decreases in catalytic activity after each run. XPS results show that the Pd 3d binding energies of the reused Cu–Pd@MIL-101 are lowered by ∼0.3 eV than those of the fresh Cu–Pd@MIL-101, which may be caused by the slight aggregation of Pd–Cu NPs after six runs (Figure S21) as revealed by its TEM and HAADF-STEM images (Figure S22). Obviously, the excellent stability of Cu–Pd@MIL-101 can be ascribed to the nanoconfinement effect of MIL-101 pores, which can stabilize these Cu–Pd NPs and prevent them from aggregation during the reaction. Furthermore, we also investigated the general applicability of Cu–Pd@MIL-101 for the homocoupling of a variety of terminal acetylene under the optimized reaction conditions. As summarized in Table S10, our Cu–Pd@MIL-101 can exhibit good tolerance for a wide scope of aromatic alkynes substrates with either electron-donating (−OCH3, −CH3, −tBu) or electron-withdrawing groups (−CF3, −F). Besides, the reaction conditions are also compatible with the heterocyclic substrates that contain pyridine and aliphatic alkynes. However, the presence of nitro and amino groups significantly decreases its reactivity toward the homocoupling reaction, leading to relatively lower yields of 1,4-bis(4-nitrophenyl)buta-1,3-diyne and 4,4′-(buta-1,3-diyne-1,4-diyl)dianiline (Table S10, entry 7 and entry 8).

Figure 6

Synergistic catalysis of Cu–Pd@MIL-101 for homocoupling reaction. (A) The catalytic activities of various alloyed catalysts for the homocoupling of phenylacetylene. (B) The influence of Cu/Pd molar ratio of Cu–Pd@MIL-101 on the homocoupling of phenylacetylene. (C) The influence of the catalyst dosages of Cu–Pd@MIL-101 on the homocoupling of phenylacetylene. (D) The yields of 1,4-diphenylbuta-1,3-diyne over various catalysts as a function of reaction time. (E) Proposed mechanism for the homocoupling of phenylacetylene catalyzed by Cu–Pd@MIL-101. (F) Reusability test of Cu–Pd@MIL-101 for the homocoupling of phenylacetylene.

In summary, we have developed a universal strategy for the immobilization of well-alloyed NPs within MIL-101. On the basis of the galvanic replacement of transition-metal NPs with noble-metal ions under high-intensity ultrasound irradiation, various alloyed NPs are successfully prepared and highly dispersed in the framework of MIL-101 with ultrafine particle sizes from 1.1 to 2.2 nm and high loadings of up to 10.4 wt %. As an example for the homocoupling reaction of phenylacetylene, the optimized Cu–Pd@MIL-101 exhibits an extremely high activity with 98% yield of 1,4-diphenylbuta-1,3-diyne, which is much more active than its monometallic counterparts and other alloyed catalysts. Our study not only opens up a new avenue for the design of highly dispersed and well-alloyed NPs immobilized in MOF with high loading and small particle size, but also directly demonstrates the efficient synergetic catalysis of Cu and Pd metals for the homocoupling reaction of terminal acetylenes under unprecedented base- and additive-free conditions and room temperature.

Safety Statement

Caution should be taken when using HF during MIL-101 preparation, because HF has very high toxicity and corrosivity.