Communication pubs.acs.org/JACS
Confined Pyrolysis within Metal−Organic Frameworks To Form Uniform Ru3 Clusters for Efficient Oxidation of Alcohols Shufang Ji,†,○ Yuanjun Chen,†,○ Qiang Fu,‡,○ Yifeng Chen,§ Juncai Dong,∥ Wenxing Chen,† Zhi Li,† Yu Wang,⊥ Lin Gu,# Wei He,§ Chen Chen,† Qing Peng,† Yu Huang,†,∇ Xiangfeng Duan,∇ Dingsheng Wang,*,† Claudia Draxl,‡ and Yadong Li*,† †
Department of Chemistry, Tsinghua University, Beijing 100084, China Institut für Physik and IRIS Adlershof, Humboldt-Universität zu Berlin, Berlin 12489, Germany § School of Pharmaceutical Sciences, Tsinghua University, Beijing 100084, China ∥ Beijing Synchrotron Radiation Facility, Institute of High Energy Physics, Chinese Academy of Sciences, Beijing 100049, China ⊥ Shanghai Synchrotron Radiation Facility, Shanghai Institute of Applied Physics, Chinese Academy of Sciences, Shanghai 201204, China # Beijing National Laboratory for Condensed Matter Physics, Institute of Physics, Chinese Academy of Sciences, Beijing 100190, China ∇ Department of Materials Science & Engineering, University of California, Los Angeles, California 90095, United States ‡
S Supporting Information *
Metal−organic frameworks (MOFs) are a unique type of porous crystalline materials with fine-tunable and uniform pore structures.4 As a subclass of MOFs, zeolitic imidazolate frameworks (ZIFs) possess interconnected large three-dimensional molecular-scale cages that are accessible through small apertures.5 They can act as templates to obtain nitrogen-doped porous carbon with abundant nitrogen species.6 By using the striking features of ZIFs, we have developed a novel approach to synthesize atomically dispersed uniform clusters via a cageseparated precursor preselection and pyrolysis strategy. Our strategy includes two steps: (1) encapsulating and separating preselected metal cluster precursors by molecular-scale cages of ZIFs followed by (2) pyrolysis to remove ligands of the metal cluster precursors and form nitrogen-doped porous carbon to stabilize the clusters. To illustrate this strategy, the preparation procedure is shown in Scheme 1. ZIF-8, possessing large sod cages (cage diameter
ABSTRACT: Here we report a novel approach to synthesize atomically dispersed uniform clusters via a cage-separated precursor preselection and pyrolysis strategy. To illustrate this strategy, well-defined Ru3(CO)12 was separated as a precursor by suitable molecular-scale cages of zeolitic imidazolate frameworks (ZIFs). After thermal treatment under confinement in the cages, uniform Ru3 clusters stabilized by nitrogen species (Ru3/CN) were obtained. Importantly, we found that Ru3/CN exhibits excellent catalytic activity (100% conversion), high chemoselectivity (100% for 2-aminobenzaldehyde), and significantly high turnover frequency (TOF) for oxidation of 2-aminobenzyl alcohol. The TOF of Ru3/CN (4320 h−1) is about 23 times higher than that of small-sized (ca. 2.5 nm) Ru particles (TOF = 184 h−1). This striking difference is attributed to a disparity in the interaction between Ru species and adsorbed reactants.
Scheme 1. Illustration of the Ru3/CN Preparation Process
S
upported metal cluster and single-atom catalysts have attracted considerable research interest because they possess unique electronic structures and extraordinary catalytic properties and can bridge heterogeneous and homogeneous catalysis.1 However, atomic dispersion of metal species is not easy to control under preparation and reaction conditions, which usually results in the coexistence of multiple forms of species including nanoparticles, clusters, and even single atoms.2 This phenomenon brings difficulties for identifying active species. Moreover, atomically dispersed catalysts with uniform active sites bear great potential to provide fundamental insight leading to a deep understanding of the correlation between the structure and catalytic mechanisms at the molecular scale.3 It appears that the synthesis of uniformly dispersed metal species is significantly desirable but still remains a grand challenge. © 2017 American Chemical Society
dc = 11.6 Å) accessible through narrow six-ring pores (pore diameter dp = 3.4 Å), was selected as a host for the preselected precursor Ru3(CO)12 with an optimal size (molecular diameter dm ≈ 8.0 Å).7 Since dm is larger than dp and smaller than dc, one Ru3(CO)12 molecule can be accommodated by one cage and not be released. In the synthesis, Ru3(CO)12 was mixed with Received: May 15, 2017 Published: July 11, 2017 9795
DOI: 10.1021/jacs.7b05018 J. Am. Chem. Soc. 2017, 139, 9795−9798
Communication
Journal of the American Chemical Society
different regions (areas 1, 2, and 3 in Figure 1d). As shown in the intensity profiles (Figure 1f), the distances between two neighboring Ru atoms are in the range from 1.9 to 2.5 Å, consistent with the metal bond lengths in Ru3 triangular structures. Additionally, some Ru3 clusters were fragmented as a result of the influence of the high-energy or prolonged electron beam.9,11 To further identify the structure of the Ru3 clusters, X-ray absorption near-edge structure (XANES), extended X-ray absorption fine structure (EXAFS), and a wavelet transform (WT) were carried out. The Ru K-edge XANES curves (Figure 2a) show that the energy absorption threshold value of Ru3/
the precursors of ZIF-8 in one pot to be encapsulated in situ along with the assembly of Zn2+ and 2-methylimidazole during ZIF-8 crystallization. Upon thermal treatment, the isolated Ru3(CO)12 molecules within the cages of ZIF-8 (Ru3(CO)12@ ZIF-8) decomposed to obtain uniform Ru3 clusters stabilized by nitrogen species. The powder X-ray diffraction (PXRD) pattern of Ru3(CO)12@ZIF-8 matched well with that of ZIF-8 (Figure S1). The FT-IR spectrum of Ru3(CO)12@ZIF-8 (Figure S2) displayed characteristic νCO absorption bands at 2060 (s), 2025 (s), 2000 (w), and 1987 cm−1 (s), consistent with those of Ru3(CO)12, indicating that Ru3(CO)12 had been encapsulated in the cages of ZIF-8.8 Upon pyrolysis at 800 °C under a 5% H2/Ar atmosphere, ZIF-8 was transformed into nitrogen-doped porous carbon, and simultaneously the Ru3(CO)12 within the cages was decarbonylated, leading to the formation of Ru3 clusters anchored by nitrogen species. The broad shoulder peak at about 26° in the PXRD pattern of Ru3/CN corresponds to the C (002) plane (Figure S1). The X-ray photoelectron spectroscopy (XPS) analysis of C and N in Ru3/CN is shown in Figure S3. Brunauer−Emmett−Teller (BET) surface area analysis (Figure S4) revealed that Ru3/CN maintains porous structure, as supported by obvious voids in transmission electron microscopy (TEM) images (Figures 1a and S5). The
Figure 2. (a) Ru K-edge XANES spectra. (b) EXAFS Fourier transformed (FT) k2-weighted χ(k) function spectra of Ru3/CN and reference. (c) Wavelet transforms for the k2-weighted Ru K-edge EXAFS signals for the high-coordination shells in Ru3/CN. The colors in the contour figures indicate the moduli of the Morlet wavelet transform. (d) Comparison of the q-space magnitudes for FEFFcalculated k2-weighted EXAFS paths. (e) Corresponding EXAFS Rspace fitting curve for Ru3/CN. (f) Schematic model of Ru3/CN, with Ru in teal, N in blue, and C in gray.
Figure 1. (a) TEM images, (b) HAADF-STEM images, and (c) corresponding element maps showing distributions of Ru (red), N (blue), C (green), respectively. (d) Aberration-corrected HAADFSTEM images of Ru3/CN. (e) Corresponding intensity maps (left) and models (right) and (f) intensity profiles obtained in areas labeled 1, 2, and 3 in (d).
high-angle annular dark-field scanning TEM (HAADF-STEM) images and corresponding element maps (Figure 1b,c) indicate uniform distributions of Ru, C, and N throughout Ru3/CN. To discern the atomic structure of Ru3/CN, aberrationcorrected HAADF-STEM was used. In Figure 1d, several uniformly dispersed bright dots, corresponding to Ru atoms, can be clearly distinguished from the nitrogen-doped carbon matrix because of the Z contrast.9 As the HADDF images represent two-dimensional projections of samples along the incident beam direction, the configurations are different from those in the three-dimensional perspective.10 The observation of a group of three bright dots (marked with yellow and green circles) is consistent with triangular Ru3 structures, whereas that of two bright dots (marked with red circles) corresponds to Ru3 clusters that were not aligned in projections. As shown in Figure 1e, the corresponding models were constructed for three
CN is higher than that of Ru foil and lower than that of RuO2, indicating that the Ru3 clusters carry positive charge. The corresponding FT curve (Figure 2b) exhibits a main peak at about 1.4 Å (without phase correction), associated with the first shell of Ru−N scattering. Two higher-shell peaks at distances of 2.3 and 3.0 Å in Ru3/CN were observed and are ascribed to Ru−Ru scattering from Ru3 triangular metal frame structures and Ru−C scattering, respectively. To further investigate the nature of Ru3/CN, WT analysis of the Ru K-edge EXAFS oscillations was performed. In Figure 2c, the WT contour plot of Ru3/CN in the first coordination shell shows one intensity maximum at ∼4.2 Å−1, associated with Ru−N contribution. However, for the higher coordination shell, the WT contour plot shows two additional maxima at ∼2.3 and 7.0 Å−1, besides the intensity maximum at ∼4.2 Å−1, 9796
DOI: 10.1021/jacs.7b05018 J. Am. Chem. Soc. 2017, 139, 9795−9798
Communication
Journal of the American Chemical Society
Ru3/CN catalyst was characterized again (Figures S11−S15), confirming that uniformly dispersed structure of Ru3 clusters remained unchanged. To further test the scope of the reaction with Ru3/CN (Table S3), substituted benzyl alcohols with both electron-donating groups (2a, 2b) and electron-withdrawing groups (2d, 2e, 2g) were converted to the corresponding aldehydes in excellent yields (95−99%). Halogen-substituted benzyl alcohols (2c, 2f) were also converted into the corresponding aldehydes in 96% yield. Furthermore, allylic (2h) and N-heterocyclic (2i) benzyl alcohols also reacted smoothly to give the desired aldehydes in 93−94% yield with a longer reaction time. To understand the catalytic behavior of Ru3 clusters, other Ru species containing Ru single atoms (Ru1/CN) and smallsized Ru nanoparticles (Ru NPs/CN) as comparisons were prepared and characterized (Figures S16−S27 and Table S4). The activities of Ru1/CN, Ru NPs/CN, and commercial Ru/C are shown in Figure 3a. Lower conversions of 2-aminobenzyl alcohol for Ru1/CN (21%) and Ru NPs/CN (15%) compared with Ru3/CN (100%) were obtained after 0.5 h. The TOFs of Ru1/CN (416 h−1) and Ru NPs (184 h−1) are remarkably lower than that of Ru3/CN at 20% conversion (Figure 3b). Similarly, commercial Ru/C also showed low activity for this reaction, with a TOF of 103 h−1. These experimental results confirm that uniform Ru3 clusters are the major active species for the oxidation of 2-aminobenzyl alcohol, demonstrating that control over the formation of uniform sites is really an efficient way to enhance performance. Considering the striking difference in reactivity between Ru3/ CN, Ru1/CN, and small-sized Ru NPs/CN, first-principles calculations were performed to understand the mechanism. An embedded RuN4 model13 (Figure S26) and a four-layer Ru(0001) slab were employed to simulate Ru1/CN and Ru NPs, respectively. It is known that catalytic activity of transition-metal catalysts is determined by the strength of their interactions with reactant molecules.14 In Figure 4, we
owing to the Ru−C contribution. As the WT maximum corresponds to the same location of the maximum in the qspace magnitude, a complementary comparison of the q-space magnitudes for FEFF-calculated k2-weighted EXAFS paths was constructed (Figure 2d). The Ru−N(C) path shows one maximum near 4.2 Å−1, while the Ru−Ru path shows one maximum near 8.7 Å−1 associated with a shoulder around 3.8 Å−1, which is well-compatible with the WT of Ru foil (Figure S6). Moreover, with the Debye−Waller factors, σ2, increased from 0.005 to 0.015 Å2, an obvious shift of the Ru−Ru path maximum to 7.0 Å−1 is clearly exhibited, suggesting that the WT maximum at 7.0 Å−1 in Figure 2c can be ascribed to Ru− Ru scattering. Taken together, the results further well suggest the existence of Ru−Ru bonds. Thus, the Ru−N coordination and Ru−Ru metallic bonds of the Ru3 triangular structures were detected, indicating that uniform Ru3 clusters were atomically dispersed and stabilized by nitrogen species, as further confirmed by time-of-flight secondary ion mass spectrometry (TOF-SIMS) analysis (Figure S7). Quantitative EXAFS curve fitting was carried out, and the results for Ru3/CN are displayed in Figures 2e and S8. For comparison, the fitting results for Ru foil and RuO2 are shown in Figures S9 and S10. Moreover, the corresponding fitting parameters are listed in Table S1. The coordination number for the first shell of Ru−N (Ru−Ru) is 2.2 ± 0.7 (1.6 ± 0.5), and the corresponding mean bond length is 1.99 ± 0.02 Å (2.53 ± 0.02 Å). First-principles calculations were performed to elucidate the possible structure of Ru3/CN. The CN framework was simulated using a simplified model of a nitrogen-doped graphene monolayer, where the triangular Ru3 cluster was embedded within a hole that is surrounded by four pyrrole subunits (Figure 2f). The Ru−N bond lengths are in the range from 1.95 to 2.04 Å (Table S2), consistent with the EXAFS measurement. Bader charge analysis revealed that the in-plane and the out-of-plane Ru atoms exhibit positive charges of 0.62 and 0.46, respectively, consistent with the XANES results. The selective oxidation of primary alcohols to aldehydes is one of the most fundamental processes in organic synthesis.12 The oxidation of 2-aminobenzyl alcohol was chosen as a model reaction to evaluate the catalytic performance of Ru3/CN. As shown in Figure 3a, the Ru3/CN catalyst exhibited high conversion of 2-aminobenzyl alcohol (up to 100%) after 0.5 h, and products of excessive oxidation (2-aminobenzoic acid or 2nitrobenzaldehyde) were not detected even when the reaction time was extended, certifying high chemoselectivity. The calculated turnover frequency (TOF) of Ru3/CN was as high as 4320 h−1 (Figure 3b). When the reaction was over, the used
Figure 4. Geometries of 2-aminobenzyl alcohol on Ru3/CN (middle), Ru1/CN (left), and Ru NPs (right). The blue and red triangles and black circles mark the adsorption energies of the molecule bound via the amino group, the hydroxyl group, and both groups, respectively. Positive values represent exothermic adsorption. The gray, blue, red, white, and teal spheres represent C, N, O, H, and Ru atoms, respectively.
display adsorption energies of the 2-aminobenzyl alcohol reactant on the three systems. On Ru(0001), the interaction is so strong that the reactant molecules are expected to block active sites at the surface, which impairs the catalytic activity. On Ru3/CN and Ru1/CN, the limited number of Ru atoms results in a significant reduction in the adsorption energy, which
Figure 3. (a) Conversion (%) of 2-aminobenzyl alcohol vs time for Ru3/CN (red), Ru 1/CN (blue), Ru NPs/CN (yellow), and commercial Ru/C (green). The inset shows the scheme of oxidation of 2-aminobenzyl alcohol to 2-aminobenzaldehyde. (b) Corresponding TOFs obtained at 20% conversion of 2-aminobenzyl alcohol. 9797
DOI: 10.1021/jacs.7b05018 J. Am. Chem. Soc. 2017, 139, 9795−9798
Communication
Journal of the American Chemical Society
21390393, U1463202, 21471089, 21671117). We thank the Hefei Light Source, Beijing Light Source, and Shanghai Light Source for use of the instruments. Q.F. and C.D. acknowledge support from the German Research Foundation (DFG) through the Collaborative Research Center SFB-658.
brings about improved TOFs compared with Ru NPs. Comparing Ru3/CN and Ru1/CN, the difference in their TOFs is attributed to a variation of the bonding pattern between the reactant molecules and Ru species. Among the various adsorption configurations in Figure 4, those merely bound via an amino group are not expected to be effective, since in aerobic oxidation the amino group behaves as a spectator and needs to be far from the active site. On Ru1/CN, such a structure is 0.58 eV more stable than that bound via a hydroxyl group, meaning that the reactants prefer to adopt the ineffective configuration and block the Ru1 active sites. On Ru3/ CN, by contrast, the reactants can adopt another geometry where one Ru atom connects with both the hydroxyl and amino groups. Such a configuration, whose adsorption energy is close to that of the ineffective structure, is also effective for the reaction, thus explaining why Ru3/CN behaves better than Ru1/CN. In summary, we have prepared uniform Ru3 clusters by a novel cage-separated precursor preselection and pyrolysis strategy. The structure of the Ru3/CN catalyst was characterized by HAADF-STEM and XAFS, and the catalytic performance was evaluated using the oxidation of 2-aminobenzyl alcohol as a model reaction. This catalyst shows 100% conversion, 100% selectivity, and an unexpected high TOF (4320 h−1), which are superior to those of Ru single atom and small-sized Ru particle catalysts. First-principles calculations revealed that the dramatic difference comes from a disparity in the interaction between Ru species and adsorbed reactant molecules, further confirming that uniform Ru3 clusters are efficient sites of enhanced performance. This work should provide a new strategy for the generation of atomically dispersed uniform cluster materials and pave a new avenue for further research in the application of cluster catalysts.
■
■
ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/jacs.7b05018. Detailed experimental section and supporting figures and tables (PDF)
■
REFERENCES
(1) (a) Yang, X. F.; Wang, A.; Qiao, B.; Li, J.; Liu, J.; Zhang, T. Acc. Chem. Res. 2013, 46, 1740. (b) Tyo, E. C.; Vajda, S. Nat. Nanotechnol. 2015, 10, 577. (c) Liu, P.; Zhao, Y.; Qin, R.; Mo, S.; Chen, G.; Gu, L.; Chevrier, D. M.; Zhang, P.; Guo, Q.; Zang, D.; Wu, B.; Fu, G.; Zheng, N. Science 2016, 352, 797. (d) Wei, H.; Liu, X.; Wang, A.; Zhang, L.; Qiao, B.; Yang, X.; Huang, Y.; Miao, S.; Liu, J.; Zhang, T. Nat. Commun. 2014, 5, 5634. (2) (a) Corma, A.; Concepcion, P.; Boronat, M.; Sabater, M. J.; Navas, J.; Yacaman, M. J.; Larios, E.; Posadas, A.; López-Quintela, M. A.; Buceta, D.; Mendoza, E.; Guilera, G.; Mayoral, A. Nat. Chem. 2013, 5, 775. (b) Liu, L.; Diaz, U.; Arenal, R.; Agostini, G.; Concepcion, P.; Corma, A. Nat. Mater. 2017, 16, 132. (3) (a) Li, G.; Jin, R. Acc. Chem. Res. 2013, 46, 1749. (b) Lv, C.; Cheng, H.; He, W.; Shah, M. I. A.; Xu, C.; Meng, X.; Jiao, L.; Wei, S.; Li, J.; Liu, L.; Li, Y. Nano Res. 2016, 9, 2544. (c) Yang, M.; Liu, J.; Lee, S.; Zugic, B.; Huang, J.; Allard, L. F.; Flytzani-Stephanopoulos, M. J. Am. Chem. Soc. 2015, 137, 3470. (4) Furukawa, H.; Cordova, K. E.; O’Keeffe, M.; Yaghi, O. M. Science 2013, 341, 1230444. (5) (a) Zhang, J. P.; Zhang, Y. B.; Lin, J. B.; Chen, X. M. Chem. Rev. 2012, 112, 1001. (b) Liu, X.; He, L.; Zheng, J.; Guo, J.; Bi, F.; Ma, X.; Zhao, K.; Liu, Y.; Song, R.; Tang, Z. Adv. Mater. 2015, 27, 3273. (6) (a) Cao, X.; Tan, C.; Sindoro, M.; Zhang, H. Chem. Soc. Rev. 2017, 46, 2660. (b) Aijaz, A.; Fujiwara, N.; Xu, Q. J. Am. Chem. Soc. 2014, 136, 6790. (c) Shen, K.; Chen, X.; Chen, J.; Li, Y. ACS Catal. 2016, 6, 5887. (d) Oar-Arteta, L.; Wezendonk, T.; Sun, X.; Kapteijn, F.; Gascon, J. Mater. Chem. Front. 2017, DOI: 10.1039/ C7QM00007C. (e) Sun, J.-K.; Xu, Q. Energy Environ. Sci. 2014, 7, 2071. (7) (a) Shi, Q.; Chen, Z.; Song, Z.; Li, J.; Dong, J. Angew. Chem., Int. Ed. 2011, 50, 672. (b) Karagiaridi, O.; Lalonde, M. B.; Bury, W.; Sarjeant, A. A.; Farha, O. K.; Hupp, J. T. J. Am. Chem. Soc. 2012, 134, 18790. (8) Moggi, P.; Predieri, G.; Di Silvestri, F.; Ferretti, A. Appl. Catal., A 1999, 182, 257. (9) Kulkarni, A.; Chi, M.; Ortalan, V.; Browning, N. D.; Gates, B. C. Angew. Chem., Int. Ed. 2010, 49, 10089. (10) Li, Z. Y.; Young, N. P.; Di Vece, M.; Palomba, S.; Palmer, R. E.; Bleloch, A. L.; Curley, B. C.; Johnston, R. L.; Jiang, J.; Yuan, J. Nature 2008, 451, 46. (11) (a) Batson, P. E.; Dellby, N.; Krivanek, O. L. Nature 2002, 418, 617. (b) Houk, R. J. T.; Jacobs, B. W.; Gabaly, F. E.; Chang, N. N.; Talin, A. A.; Graham, D. D.; House, S. D.; Robertson, I. M.; Allendorf, M. D. Nano Lett. 2009, 9, 3413. (12) (a) Hanson, S. K.; Wu, R.; Silks, L. A. Org. Lett. 2011, 13, 1908. (b) Guan, M.; Wang, C.; Zhang, J.; Zhao, Y. RSC Adv. 2014, 4, 48777. (13) Kattel, S.; Wang, G. J. Phys. Chem. Lett. 2014, 5, 452. (14) (a) Nørskov, J. K.; Bligaard, T.; Rossmeisl, J.; Christensen, C. H. Nat. Chem. 2009, 1, 37. (b) Nørskov, J. K.; Studt, F.; Abild-Pedersen, F.; Bligaard, T. Fundamental Concepts in Heterogeneous Catalysis; John Wiley & Sons: Hoboken, NJ, 2014.
AUTHOR INFORMATION
Corresponding Authors
*
[email protected] *
[email protected] ORCID
Qiang Fu: 0000-0002-6682-8527 Chen Chen: 0000-0001-5902-3037 Xiangfeng Duan: 0000-0002-4321-6288 Dingsheng Wang: 0000-0003-0074-7633 Claudia Draxl: 0000-0003-3523-6657 Yadong Li: 0000-0003-1544-1127 Author Contributions ○
S.J., Y.C., and Q.F. contributed equally.
Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This work was supported by the China Ministry of Science and Technology under Contract of 2016YFA (0202801) and the National Natural Science Foundation of China (21521091, 9798
DOI: 10.1021/jacs.7b05018 J. Am. Chem. Soc. 2017, 139, 9795−9798
