Engineering Electronic Structure of Single-Atom Pd Site on Ti _0.87 O ₂ Nanosheet via Charge Transfer Enables C–Br Cleavage for Room-Temperature Suzuki Coupling

Abstract

Open AccessCCS ChemistryRESEARCH ARTICLE1 Jun 2021Engineering Electronic Structure of Single-Atom Pd Site on Ti0.87O2 Nanosheet via Charge Transfer Enables C–Br Cleavage for Room-Temperature Suzuki Coupling Yangxin Jin†, Fei Lu†, Ding Yi†, Junmeng Li, Fengchu Zhang, Tian Sheng, Fei Zhan, Ya'nan Duan, Gaochao Huang, Jinyang Dong, Bo Zhou, Xi Wang and Jiannian Yao Yangxin Jin† Department of Physics, School of Science, Beijing Jiaotong University, Beijing 100044 , Fei Lu† Department of Physics, School of Science, Beijing Jiaotong University, Beijing 100044 , Ding Yi† Department of Physics, School of Science, Beijing Jiaotong University, Beijing 100044 , Junmeng Li School of Chemical Engineering and Technology, Molecular Plus and Collaborative Innovation Center of Chemical Science and Engineering, Tianjin University, Tianjin 300072 , Fengchu Zhang Department of Physics, School of Science, Beijing Jiaotong University, Beijing 100044 , Tian Sheng College of Chemistry and Materials Science, Anhui Normal University, Wuhu 241000 , Fei Zhan Chemistry and Chemical Engineering Guangdong Laboratory, Shantou 515031 , Ya'nan Duan Chemistry and Chemical Engineering Guangdong Laboratory, Shantou 515031 , Gaochao Huang Chemistry and Chemical Engineering Guangdong Laboratory, Shantou 515031 , Jinyang Dong Department of Physics, School of Science, Beijing Jiaotong University, Beijing 100044 , Bo Zhou Chemistry and Chemical Engineering Guangdong Laboratory, Shantou 515031 , Xi Wang *Corresponding authors: E-mail Address: [email protected] E-mail Address: [email protected] Department of Physics, School of Science, Beijing Jiaotong University, Beijing 100044 and Jiannian Yao *Corresponding authors: E-mail Address: [email protected] E-mail Address: [email protected] Chemistry and Chemical Engineering Guangdong Laboratory, Shantou 515031 Key Laboratory of Photochemistry, Beijing National Laboratory for Molecular Sciences (BNLMS), Institute of Chemistry, Chinese Academy of Sciences, Beijing 100190 https://doi.org/10.31635/ccschem.020.202000388 SectionsSupplemental MaterialAboutAbstractPDF ToolsAdd to favoritesTrack Citations ShareFacebookTwitterLinked InEmail The palladium (Pd)-catalyzed Suzuki reaction is widely applied in the pharmaceutical industry, where constructing highly active and low-cost Pd sites are impendent. Here, we report the fabrication of a heterogeneous Pd/Tio2 catalyst via engineering of an electronic structure of a single Pd1 atom on monolayered Ti0.87O2 nanosheet (Pd1-Ti0.87O2). This catalyst motivated the kinetically sluggish C–Br cleavage, thus boosting the Suzuki reaction at room temperature. Pd1-Ti0.87O2 exhibited an outstanding activity with turnover frequency (TOF) of 11,110 h−1, exceeding that of PdCl2 and Pd(OAc)2 catalysts by a factor of >200. Various in situ techniques were employed to investigate the C–Br activation process, which showed that Pd1 kinetic-feasibly dissociated the chemisorbed bromobenzene, especially the C–Br bond cleavage. Theoretical calculations further revealed that the improved activity is ascribed to the optimized charge state of Pd1 within the Pd1O4 realm via charge transfer. Download figure Download PowerPoint Introduction The discovery of the palladium (Pd)-catalyzed Suzuki–Miyaura cross-couplings reaction is a breakthrough due to its vital value for pharmaceutical industries,1,2 since the first report in 1979s.3 Homogenous Pd catalysts have been well studied for Suzuki coupling. However, they generally suffer from high cost, catalyst residue formation, nonrecyclability, and sluggish reactivity for aryl halides.4–8 With regards to supporting Pd complexes on solid carriers, the catalysts could be recycled, but they give worse performances after each cycle, which is mainly due to the aggregation of Pd.9,10 The third choice is supported Pd nanoparticle catalysts; however, they do not have enough activity under mild conditions for their determined, dynamic structures.11–20 Inspiringly, heterogeneous single-atom catalyst emerged as the most potential candidate to settle these challenges, precisely, by providing a bridge between heterogeneous and homogeneous catalysis.21,22 Informatively, Chen et al.23 reported a promising heterogeneous single-atom catalyst that anchor Pd atoms on exfoliated graphitic carbon nitride (Pd-ECN) with a turnover frequency (TOF) value as high as 549 h−1 and robust stability, surpassing homogeneous catalysts in a continuing Suzuki reaction system. Last but not the least, the defects within supported along with the metal–support interactions have been employed to stabilize the single atoms and manipulate the electronic structure, thus regulating the catalytic activity.24,25 In this study, a monolayered cation-deficient Ti0.87O2 nanosheet was engineered as support. The specific Ti vacancies were used toward Pd cation anchorage, where strong electrostatic interaction between Pd cations and Ti vacancies ensured the atomic dispersion of Pd atoms on the Ti0.87O2 nanosheet surface (Pd1-Ti0.87O2). Then the catalyst was employed in room-temperature Suzuki reaction, showing an extraordinarily high activity (TOF up to 11,110 h−1). This TOF value is over 200 times higher than that of traditional homogenous Pd catalysts [e.g., PdCl2 and Pd(OAc)2]. The refined atomic structure of the Pd atoms in Ti0.87O2 was studied by high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) and X-ray absorption fine structure (XAFS). Various in situ characterizations, such as XAFS and Fourier transform infrared spectroscopy (FTIR), were used to observe the C–Br bond cleavage process. These results suggested that the Pd1-Ti0.87O2 catalyst indeed promoted hybridization between aryl halides and Pd single site. Theoretical calculations showed that there was a preferred charge transfer within the Pd1O4 realm, compared with the homogeneous Pd-related catalysts [PdCl2 and Pd(OAc)2]. As a result, the increased electronic density of the Pd1 site boosted the C–Br cleavage dramatically. This study indicates that the engineered electronic structure of a single-atom site within a specific realm plays a crucial role in catalytic activity. Experimental Methods Preparation of the monolayered Pd1-Ti0.87O2 nanosheet catalyst The monolayered titania nanosheet (Ti0.87O2) was synthesized according to a delamination process of stacked titanate materials, and the nonstoichiometry formula indicated the existence of Ti deficiencies.26–28 Typically, a stoichiometric chemical mixture of Li2CO3, K2CO3, and TiO2 (anatase) with a molar ratio of 0.14∶0.4∶1.73 was fully ground and then calcined at 1173 K for 1 h. After cooling to room temperature, the mixture was further calcinated at 1273 K for 20 h, followed by another grinding step. As a result, a precursor titanate of K0.8Ti1.73Li0.27O4 was produced. Then the composite titanate metal generated was stirred homogeneously in 1 mol dm−3 HCl solution at room temperature for 2 days through a protonation process. The acid liquor was renewed daily with a fresh one. The protonated titanate solids were gleaned after successive centrifugal separation process, washed several times with ddH2O to remove the residual acid, and dried in a vacuum drying oven at 353 K overnight. Thereafter, the gathered titanate was scattered sufficiently in an even tetrabutylammonium hydroxide [(C4H9)4NOH; TBAOH, 1.5M] aqueous solution to get the monolayered nanosheet. The molar ratio of proton and TBAOH is 1∶1. A typical impregnation approach was farther executed to attain the final Pd single-atom catalyst Pd1-Ti0.87O2: quantificational Pd precursor [tetraamminepalladium(II) hydrogen carbonate] was added into the colloidal suspensions of monolayered titania nanosheets. Other Pd precursors such as PdCl2 and Pd(OAc)2 were also applied in this impregnation process, but their poor solubility in water made it difficult to provide enough free Pd2+. Finally, the single-atom catalyst Pd1-Ti0.87O2 was prepared successfully by freeze-drying of compounded suspensions. The actual loading quantity of Pd was 0.10 wt%, determined by inductively coupled plasma optical emission spectrometer (ICP-OES). In situ FTIR experiment FTIR was recorded over a model Nicolet Avatar 360 (Thermo, NYC, USA) FTIR spectrophotometer equipped with a highly sensitive mercury cadmium telluride (MCT) detector. The C–Br activation process was observed by applying the attenuated total reflectance (ATR) method. Bromobenzene was added dropwise to the Pd1-Ti0.87O2 sample at room temperature. Then the mixture was subjected immediately to an infrared (IR) absorption by the FTIR spectrometer to gather spectral information of the reaction changes. In situ XAFS test X-ray absorption near-edge structure (XANES) and extended XAFS (EXAFS) measurements at the Pd K-edges were performed in fluorescence mode at the beamline BL14W1. PdO and Pd foils were used as reference samples. Quantitative curve fittings were carried out in k2-weighted EXAFS oscillation in the range of 0–6 Å using the ARTEMIS module of the IFEFFIT software package. In XAFS fit, the lattice parameters were optimized. With respect to the in situ investigation, the fresh Pd1-Ti0.87O2 sample was tested, serving as a contrast, then two drops of bromobenzene were added to the catalyst surface. Compared with the fresh one, a noticeable change was observed at the pre-edge peak in a short time. Density functional theory calculation Density functional theory (DFT) calculations were carried out using the Vienna ab-initio simulation package (VASP, University of Vienna, Austria). The Perdew–Burke–Ernzerhof generalized gradient approximation exchange-correlation potential (PBE-GGA) was used, and the electron–core interactions were treated by the projector augmented wave (PAW) method.29–32 The on-site Hubbard U term was added on Ti 3d orbitals at the value of 3.5 eV. Structures were optimized until the atomic forces were smaller than 0.01 eV Å−1 with a kinetic cutoff energy of 400 eV. Reaction barriers were determined with the nudged elastic band method. The optimized lattice constants in x–y plane of the monolayered TiO2 unit cell, including two Ti atoms and four O atoms, were 3.867 × 3.057 Å, which were in good agreement with the experimental results. The monolayered TiO2 nanosheet was modeled by 40 Ti atoms and 80 O atoms. One Ti atom and two neighboring O atoms were removed to create one hole that is occupied by the Pd atom. All atoms, including adsorbed H2O, were relaxed during geometry optimizations. The vacuum layer in the z-direction was 25 Å. A 2 × 2 × 1 Monkhorst–Pack k-point sampling was used, and the dipole correction was added along the z-direction. The adsorption energy in this study was defined as follows: E(ad) = E(ad/surf) − E(surf) − E(ad), where E(ad/surf), E(ad), and E(surf) were the total energies of the adsorbate binding to the surface, free adsorbate in the gas phase, and clean surface, respectively. The free energy of species was obtained from G = E + ZPE − TS, where E is the total energy of species, TS is the entropy at room temperature, and ZPE is the zero-point energy. All the vibrational frequencies, vi (Hz), were calculated based on the harmonic oscillators approximation.33 Results and Discussion Generally, Ti deficiency rich unilamellar titania nanosheet (Ti0.87O2) was synthesized by a soft-chemical exfoliation method (Figure 1a and Supporting Information Figure S1). Holistically, the Ti0.87O2 obtained featured negative charge, which enabled the divalent Pd ions (Pd2+) to attach to Ti deficiencies due to the strong electrostatic interaction.34 The powder X-ray diffraction pattern (PXRD) of Pd1-Ti0.87O2 ( Supporting Information Figure S2) retained consistency with the pure Ti0.87O2, with no signs of Pd nanoparticle, indicating the high dispersion of the Pd species. The Pd loading was estimated to be 0.10 wt %, which was quantified by ICP-OES. As depicted in the atomic force microscopy (AFM) image (Figure 1b), the Pd1-Ti0.87O2 catalyst exhibited a monolayered nanosheet feature with a uniform thickness of ∼1.1 nm (Figure 1c). The scanning electron microscopy (SEM) images ( Supporting Information Figure S3) also confirmed its flake-like morphology (2–10 μm in diameter). An energy-dispersive X-ray spectroscopy (EDS) element mapping (Figure 1d) revealed that the Pd element was dispersed homogeneously in Pd1-Ti0.87O2 nanosheet. Each bright spot (in red circles) on the HAADF-STEM image of the Pd1-Ti0.87O2 catalyst represented a Pd single-atom dispersed on the nanosheet (Figure 1e and Supporting Information Figure S4). Thus, the Pd single atoms were confirmed to be evenly dispersed on the monolayered Ti0.87O2 nanosheet. Figure 1 | (a) Schematic diagram of the preparation process of the monolayered Pd1-Ti0.87O2 catalyst; the red, gray, green, blue, and yellow balls indicate O, Ti, Li and K, Pd, and N [(C4H9)4N+] respectively. (b) AFM image and (c) corresponding height profiles of monolayered Pd1-Ti0.87O2 nanosheets. (d) Representative STEM and EDS elemental mapping of the Pd1-Ti0.87O2 sample. (e) HAADF-STEM image of the Pd1-Ti0.87O2 sample. AFM, atomic force microscopy; STEM, scanning transmission electron microscopy; EDS, energy-dispersive X-ray spectroscopy; HAADF-STEM, high-angle annular dark-field scanning transmission electron microscopy. Download figure Download PowerPoint We depicted the accurate location of the Pd single atom within the Pd1-Ti0.87O2 nanosheet by further investigation of the refined scanning transmission electron microscopy (STEM) images (Figure 2a). Ti deficiencies (in blue cycles) were observed in monolayered nanosheet, and Pd single atoms (in red cycles) were exclusively located at Ti sites, which was demonstrated by the relevant intensity variation along the x–y line scan profile (Figure 2b). The local coordination environment of the Pd single atoms was also studied by XAFS spectrometry. As shown in the XANES spectra (Figure 2c), the white line of Pd1-Ti0.87O2 was slightly lower than that of PdO but higher than that of Pd foil (the inset at the lower right in Figure 2c), verifying a moderate charge state of Pd1 rather than a metallic state in Pd foil and Pd(II) in PdO, which indicated that the metal–support interactions induced a charge transfer between Pd1 and Ti0.87O2 substrate.25,35 Meanwhile, the near-edge structural characteristics of Pd1 in Pd1-Ti0.87O2 was similar to PdO, in which the Pd localized in a square-planar coordination environment of the PdO4 structure. Thus, the Pd in Pd1-Ti0.87O2 should have a PdO-like tetra-coordinate configuration, as shown in the inset of Figure 2c (top left). Evidently, the only one scattering path at 1.5 Å for Pd1-Ti0.87O2 was attributable to the Pd–O rather than Pd–Pd ( Supporting Information Figure S5a), excluding the possible existence of metallic Pd clusters. Figure 2 | (a) The refined HAADF-STEM image of Pd1-Ti0.87O2. (b) The x–y line scan profile, measured from (a), the blue cycles indicate the Ti vacancies. (c) XANES spectra at Pd K-edge, along with PdO and Pd foil as references. (d) Ti 2p XPS spectra of Pd1-Ti0.87O2 and pure Ti0.87O2. (e) Charge density difference of Pd-substituted Ti0.87O2 slab, yellow and blue colors represent electron increment and decrement, respectively, with an isosurface value of 0.002 e Bohr−3. HAADF-STEM, high-angle annular dark-field scanning transmission electron microscopy; XANES, X-ray absorption near-edge structure; XPS, X-ray photoelectron spectroscopy. Download figure Download PowerPoint The average coordination number (CN) of Pd atoms was obtained via fitting the peak of the first shell of Pd–O (red line, Supporting Information Figure S5b), and the result of the relative parameters was displayed in Supporting Information Table S1. The CN (Pd–O) value was 3.64 in Pd1-Ti0.87O2, indicating that Pd1 was medially bonded with four oxygen atoms from the contribution of Ti0.87O2. The missing Pd–Pd scattering at ∼3.0 Å in Pd1-Ti0.87O2 further signified the atomical dispersion of Pd in the Pd1-Ti0.87O2. According to the analyses above, the Ti vacancies in Pd1-Ti0.87O2 act as anchoring centers to capture and stabilize atomically dispersed Pd atom.36 The configurationally, unsaturated oxygen atoms in Ti0.87O2 were exposed to fix individual Pd atoms, leading to the formation of a thermodynamically stable PdO4 coordination configuration.37,38 X-ray photoelectron spectroscopy (XPS) was executed to unveil the interaction between atomically Pd and Ti0.87O2. As shown in Figure 2d and Supporting Information Figure S6, the peaks of Ti 2p of Pd1-Ti0.87O2 shift to high binding energy, compared with pure Ti0.87O2. Accordingly, the Pd(II) 3d peaks of Pd1-Ti0.87O2 shifted to low binding energy, compared with normal Pd(II). The results indicated that there is an augmentation on the electron density of anchored Pd from Ti0.87O2, which was beneficial to promote the C–Br bond cleavage.39–41 The DFT calculation was used further to uncover the electronic property of Pd1. The chosen theoretical model was based on a monolayered TiO2 structure, where the Pd single-atom coordinated with four O atoms of TiO2, according to the experimental observations (Figure 2c). In the inspiration of these basal assays, a charge density difference calculation was conducted based on the Pd1-Ti0.87O2 slab, showing the change of charge distribution before and after Pd doping (Figure 2e). Considering the incompletely filled 4d states of Pd, O 2p orbitals might be bonded strongly due to its relatively significant overlap with the Pd 4d orbitals. This viewpoint has been verified by the projected density of states (PDOS) calculation of Pd 4d orbitals in the Pd-substituted TiO2 slab ( Supporting Information Figure S7). The shaded area in the region between −1.5 and 1.0 eV indicated the bonding hybridization between Pd and O. Moreover, judging from the charge distribution analysis, there was indeed electron accumulation of Pd1 (electron cloud in yellow). These results were consistent with the XANES and XPS analyses, suggesting that the interaction between Pd and Ti0.87O2 enriched the electron density of Pd to an appropriate extent, which was propitious to the chemisorption of the reactant.41 The catalytic performances of Pd1-Ti0.87O2 and the commonly used homogenous Pd catalysts toward the Suzuki reaction were evaluated (Table 1). p-Bromotoluene and excessive phenylboronic acid were employed as the model substrates to evaluate the selectivity and conversion of the title reaction. It came out that only trace 4,4′-dimethylbiphenyl was detected, excluding the possibility for the Ullman-coupling of p-bromotoluene.42 Noticeably, Pd1-Ti0.87O2 catalyst was proved to be highly efficient and exhibited a remarkable TOF of 11,110 h−1 (Entry 1, Table 1). Pd(PPh3)4 only yielded trace amounts of the target product (Entry 2, Table 1). Pd(OAc)2 and PdCl2, the two most frequently used Pd catalysts, afforded the cross-coupling product in an impressive yield (99.9% and 99.9%) under our reaction conditions (Entries 3 and 4, Table 1), which was consistent with published results.43,44 However, they showed far lower catalytic activity than Pd1-Ti0.87O2. Then we conducted DFT calculations on Pd(OAc)2 and PdCl2 ( Supporting Information Figure S8), specifically on the charge density of Pd, which increased in the following sequence: Pd1-Ti0.87O2 > Pd(OAc)2 > PdCl2. Deductively, the engineered electronic structure of Pd in Pd1-Ti0.87O2, at the most favorable degree, played a significant role in promoting catalytic performance. Table 1 | Catalytic Performance of Various Pd-Based Catalysts in the Suzuki–Miyaura Coupling Reaction of p-Bromotoluene with Phenylboronic at Pd TOF 1 Pd1-Ti0.87O2 2 Pd(PPh3)4 3 Pd(OAc)2 PdCl2 Reaction catalysts, phenylboronic acid, K2CO3, 3 and 3 as a to the represent the Pd of the applied Pd catalysts, to a by gas on the reaction of by based on conversion of in two Pd catalysts and were also using the Suzuki–Miyaura reaction (Figure was by with phenylboronic acid to the target product in the of Pd1-Ti0.87O2 catalyst in at room temperature, only trace product was by using and Figure that the catalytic of Pd1-Ti0.87O2 was well after suggesting the robust Pd1 in Pd1-Ti0.87O2. This by Pd1-Ti0.87O2 also indicated that no free residue of the homogeneous catalyst was but the catalyst was after the reaction. give of its structural stability, XANES of fresh and used catalysts was and no significant difference was between ( Supporting Information Figure As shown in Supporting Information Figure there is no Pd–Pd bond the first of Pd spectra ( Supporting Information Figure it could be that the of Pd was slightly after an test was performed ( Supporting Information Figure to the of Pd1-Ti0.87O2 by which could the catalytic by the of it in this The results indicate that the catalytic of the Pd1-Ti0.87O2 catalyst was not Meanwhile, the catalyst was to settle after the reaction for further ( Supporting Information Figure Figure 3 | (a) Catalytic of PdO and Pd1-Ti0.87O2 in the first reaction between and phenylboronic acid to at room temperature. (b) performances of Pd1-Ti0.87O2 catalyst Download figure Download PowerPoint The of the Pd1-Ti0.87O2 catalyst was for the Suzuki–Miyaura reaction under the aryl and with phenylboronic acid in high to with selectivity at room temperature. As shown in Table 2, the property of the in bromobenzene exhibited a negative on the Suzuki reaction by Pd1-Ti0.87O2. and yielded results with phenylboronic acid (Table 2, and The in that property of the C–Br bond substrates with and with phenylboronic acid and showed results (Table 2, and in the of Pd1-Ti0.87O2, attributable mainly to the C–Br bond in these two Table 2 | Reaction aryl aryl 3 and 3 as a Pd1-Ti0.87O2 K2CO3, at room temperature. The catalytic performances of Pd1-Ti0.87O2 could be ascribed to its to the chemisorbed aryl Here, we on the C–Br bond cleavage which is generally as the unveil the process of C–Br bond cleavage over the Pd1-Ti0.87O2 catalyst was studied by in situ FTIR As shown in Figure the peak at the of the of the C–Br during the reaction. This variation indicated that the C–Br cleavage could in a short time. The peak at a of the mode of bond in the was chosen as a reference for its during the reaction. As shown in Figure the increased intensity ratio of the peak confirmed the C–Br bond cleavage. Figure | (a) In situ FTIR spectra of the and (b) the corresponding of the area of the to C–Br (c) In situ XANES spectra at Pd for Pd1-Ti0.87O2 catalyst and in situ (d) Schematic of the possible structural during the reaction, according to Figure infrared spectroscopy; XANES, X-ray absorption near-edge structure. Download figure Download PowerPoint was supported by in situ XAFS to the variation of the local electronic and of this an was to as a of bromobenzene was added on the catalyst surface at temperature. As in Figure line in an pre-edge peak eV was This indicated that the interaction between C–Br bond and Pd the local of Pd and the of the Meanwhile, the electron of Pd from to 4d orbitals indicated the hybridization between C–Br and Pd The Pd spectra of the Pd1-Ti0.87O2 catalyst after reaction were shown by XAFS ( Supporting Information Figure and the fitting were in Supporting Information Table It is that the coordination of Pd has from the to the final This process could be in Figure in which indicates on the Pd1O4 into structure; room-temperature C–Br cleavage over the Pd1-Ti0.87O2 The on C–Br cleavage was further by DFT in Figure the adsorbed bromobenzene was adsorbed the surface of Pd1-Ti0.87O2, and the bond was TS the where an bond was in a was to C–Br bond cleavage and to the to TS, the energy was relatively low which that the reaction over Pd1-Ti0.87O2. This that the process from to was and that the C–Br bond cleavage process The theoretical calculations were consistent well with the in situ FTIR and XANES observations (Figure the hybridization between Pd and C–Br bond in the density of states of the TS was as revealed in Figure was a overlap peak at eV (in which signified the hybridization between C–Br and Pd This well with the calculated charge distribution of the TS, as shown in Figure judging from the charge distribution of TS, it could be that the interactions of C–Br bond and bond were the bond was This result that Pd1 kinetically dissociated the chemisorbed bromobenzene, based on the electronic This of engineering an electronic structure of active sites with promising catalytic could the of specific catalysts, with regards to engineering Inspiringly, based on could also be further to the orbitals and of active sites such as charge, and which could into charge, and Figure | (a) profiles of the C–Br activation process over Pd1-Ti0.87O2 indicates the chemisorbed C–Br bond of bromobenzene on Pd indicates the of the C–Br bond by the of TS indicates the state in the C–Br