First author: Insoo Ro, Ji Qi, Seungyeon Lee
Corresponding author: Phillip Christopher
Communications Unit: University of California, Santa Barbara, USA
Thesis DOI: https://doi.org/10.1038/s41586-022-05075-4
Full text at a glance
It is well known that metal catalytic reactions are usually assumed to be carried out at bifunctional active sites, so co-positioned active substances promote different basic steps in the catalytic cycle. The researchers have established bifunctional active sites on homogeneous binuclear organometallic catalysts. Empirical evidence suggests the similar presence of bifunctional active sites on supported metal catalysts, e.g., at the metal-oxide carrier interface. However, due to the distribution of potentially active site structures, dynamic reconfiguration, and the required non-mean field kinetics description, elucidating the bifunctional reaction mechanism on the supported metal catalyst is challenging. A supported, single-atom dispersed rhodium-tungsten oxide (Rh-WOx) dual(pair) site catalyst was synthesized to overcome these limitations. The relative simplicity of the site structure and the full description of the mean field modeling make experimental dynamics associated with first-principles-based microdynamic simulations. Rh-WOx alignment sites catalyze ethylene hydroformylation through a bifunctional mechanism, including Rh-assisted WOx reduction, ethylene transfer from WOx to Rh, and H2 dissociation at the Rh-WOx interface. In the gas-phase ethylene carbonylation reaction, the site catalyst exhibited a selectivity of >95% at a yield of 0.1 gpropanal cm−3 h−1. The results show that the paralocation site of the oxide loading can achieve a bifunctional reaction mechanism, which is highly active and selective for industrial reactions using homogeneous catalysts.
Background
This study characterized bifunctional active sites at the heterogeneous Rh-WOx interface with structures dispersed from atoms to three-dimensional domains, and each species exhibited structure-related reactivity. The authors focused on hydroformylation of olefins, in which olefins, CO, and H2 react to add formyl (CHO) groups, and H generates aldehydes on the C=C double bond. Industrial olefin hydroformylation is performed on a homogeneous Rh catalyst with near-quantitative selectivity for the formation of aldehydes. Oxide-loaded Rh catalysts will have unwanted olefin hydrogenation reactions and by-product formation in gas-phase flow reactors; Moreover, during hydroformylation, the catalyst is poisoned by adsorbed CO, thereby limiting their reaction rate. In addition, WOx interacts strongly with olefins during the recompression process, which occurs at temperatures similar to hydroformylation. Therefore, the authors propose the hypothesis that co-localized Rh and WOx species can interact with different reactants during hydroformylation of olefins and overcome the limitations of rh catalysts.
Graphic and text analysis
Figure 1. Atomically dispersed Rh is coordinated with WOx with a different structure. a, in the case of different W loads, Rh/xW structure schematic. Element colors are as follows: red, O; Green, Rh; Blue, W; Gray, C. b, a representative STEM image of Rh/0.7W. Atomic-scale dispersed Rh/W pairs are circled in yellow. c,b Figures are pseudo-color STEM images of the Chinese box area. Pseudocolor coloring is based on scattering intensity. STEM images (d) and EDS element mapping images (e) for d, Rh/24.5W. Images are collected after preprocessing of a non-in situ CO of 523 K. f, CO probe molecular FTIR spectra collected under saturated CO coverage under Ar gas flow at 298 K, after in situ pretreatment at 523 K. The spectrum normalizes to the maximum strength of CO tensile. g, the relative strength of the ASYMMETRICAL stretch of CO during testing of TPD at a linear slope of 20 K min-1 in an Ar gas stream of 100 sccm, with temperatures from 323 to 673 K.
Figure 2. Rh-WOx behaves differently on sites. a, the apparent activation energy (Eapp) formed by propionaldehyde is a function of the W load in Rh/xW-Al2O3; In the temperature range of 393–434 K, the total pressure is 1 bar. The error estimated by Eapp is within the fitted 95% confidence interval. b, at 423 K, propionaldehyde forms a function of tow as a function of toco and C2H4 reaction series of TOF. Error bars are standard deviations obtained in three measurements by rejoining fresh samples into the reactor. c, at 423 K, propanol forms a relationship between TOF and the total pressure of the reactor. d, e, propionaldehyde selectivity (d) and propionaldehyde form TOF (e); The barometric pressures are 1 bar and 10 bar and the temperature is 403 K. For a-d, the 60 mg Rh/xW-Al2O3 catalyst was diluted with 1 g SiO2 and reduced to 1 h in CO at 523 K. It was then exposed to H2, C2H4, and CO in a 1:1:1 molar ratio with a total flow rate of 30 sccm. f, at 373 K and 10 bar, ethylene conversion and propionaldehyde selectivity at Rh/0.7W over time. A portion of the 420 mg Rh/0.7W-Al2O3 catalyst was mixed with 4 g SiO2, reduced to 1 h at CO and 523 K, and then activated by H2, C2H4 and CO at 1:1:1 molar ratio and total flow rate of 30 sccm (at 423 K and 1 bar). After activation, the catalyst was exposed to H2, C2H4 and CO with molar ratios of 1:1:1 and a total flow rate of 15 sccm at a total pressure of 373 K and 10 bar.
Figure 3. Mechanism for site activation. a, b, during CO reduction with 10 sccm CO and 5 sccm He as inert standards (a), and during hydroformylation at 10 sccm CO, 10 sccm H2, 10 sccm C2H4 and 5 sccm He and after subsequent CO reduction (b), Rh/Al2O3, Rh/0.7W and 0.7W-Al2O3 CO2A3. The inset figure in each figure shows the temperature program used during the analysis. c, MECHANISM AND ENERGY OF CO REDUCTION SURFACE MONOOXYGENATED W6+. Rh(CO)2 species are coordinated with a third CO molecule, which enables monooxygenated W6+ species to be converted to dioxometric geometry, which can easily reduce to W5+ through the precipitation of carbon dioxide. The coordination of ethylene with W5+ after CO2 formation and desorption is also shown. The free energy values for each state are displayed in kJ mol-1 below each image.
Figure 4. Mechanism and kinetic simulation of hydroformylation. a, the mechanism of CO reduction of W6+, and the catalytic cycle of ethylene hydroformylation and hydrogenation at the Rh-WOx pair. In contrast to Rh/Al2O3, the activation of the catalyst does not require desorption of co ligands, and H2 activation occurs at the Rh-W interface. b, catalyst activation state after ethylene migration from W to Rh, formation of Rh-W bond (Rh-W ≈ 2.7 Å) and coordination of H2 at the Rh-W bond interface (left); Synergistic dual-core H2 dissociation transition state (medium); Dissociated 2H* (right) on the Rh atom and Rh-W interfaces. c, Comparison of kinetic parameters from experimental and microdynamic simulations: Apparent activation energy, Eapp (top left); The number of reaction progressions corresponding to the CO formation products (top right); At 1 and 10 bar total pressures, rh/Al2O3 and Rh-WOx are selectable for propionaldehyde at 403 K ., at 403 K.
Summary and outlook
This study is a key step beyond the single-atom catalyst (SAC) that has been actively studied for the past decade. SAC consists of a single metal atom or chemical substance (active site) dispersed on a non-catalytic solid (carrier). The catalyst synthesized here is similar to SAC, but involves two types of active sites that can facilitate different steps in the catalytic cycle and are collectively localized to form isolated pairs called pairs of sites. Thus, each pair of pairs of sites is bifunctional: its two active centers work together to catalyze different steps of the reaction.
The work has several important implications. First, it shows that pair sites consisting of metals and metal oxides can achieve a bifunctional catalytic mechanism that provides high reaction rates and product selectivity. Second, it shows that pair sites with a well-defined structure can be prepared, which can then be used as a model system to elucidate the bifunctional catalytic mechanism. With this in mind, the understanding of heterogeneous catalyst mechanisms has improved considerably when SAC with a well-defined structure was developed. The emergence of model-pair catalysts is expected to provide a similar breakthrough. Third, the authors' controlled synthetic Rh-WOx characterization of catalysts at sites, atomic levels, and the bifunctional mechanism strategies developed by the authors provide a reference for future catalyst designs.