Nanocavity in hollow sandwiched catalysts as substrate regulator for boosting hydrodeoxygenation of biomass-derived carbonyl compounds.

Hydrodeoxygenation of oxygen-rich molecules toward hydrocarbons is attractive yet challenging in the sustainable biomass upgrading. The typical supported metal catalysts often display unstable catalytic performances owing to the migration and aggregation of metal nanoparticles (NPs) into large sizes under harsh conditions. Here, we develop a crystal growth and post-synthetic etching method to construct hollow chromium terephthalate MIL-101 (named as HoMIL-101) with one layer of sandwiched Ru NPs as robust catalysts. Impressively, HoMIL-101@Ru@MIL-101 exhibits the excellent activity and stability for hydrodeoxygenation of biomass-derived levulinic acid to gamma-valerolactone under 50°C and 1-megapascal H2, and its activity is about six times of solid sandwich counterparts, outperforming the state-of-the-art heterogeneous catalysts. Control experiments and theoretical simulation clearly indicate that the enrichment of levulinic acid and H2 by nanocavity as substrate regulator enables self-regulating the backwash of both substrates toward Ru NPs sandwiched in MIL-101 shells for promoting reaction with respect to solid counterparts, thus leading to the substantially enhanced performance.

Generalized Rapid Synthesis of Supported Nanocluster Catalyst for Mild Hydrogenation of Phenol toward KA Oil.

Homogeneous and nanometric metal clusters with unique electronic structures are promising for catalysis, however, common synthesis techniques for metal clusters suffer from large size and even metal nanocrystals attributing to their high surface energy and unsaturated configurations. Herein, a generalized rapid annealing strategy for synthesizing a series of supported metal clusters as superior catalysts is developed. Remarkably, TiO2 supported platinum nanoclusters (Pt NC/TiO2 ) exhibits the excellent catalytic activity to realize phenol hydrogenation under mild conditions. The complete phenol conversion rate and 100% selectivity toward KA oil are achieved in aqueous solution at room temperature and normal pressure. Semi-continuous scale up production of KA oil is successfully performed under mild conditions. Such excellent performance mainly originates from the partial reconstruction of Pt NC/TiO2 in aqueous phenol solution. Considering that the phenol can be produced from lignin, this study underpins a facile, sustainable, and economical route to synthesize nylon from biomass.

Fast and Selective Semihydrogenation of Alkynes by Palladium Nanoparticles Sandwiched in Metal-Organic Frameworks.

The semihydrogenation of alkynes into alkenes rather than alkanes is of great importance in the chemical industry. Unfortunately, state-of-the-art heterogeneous catalysts hardly achieve high turnover frequencies (TOFs) simultaneously with almost full conversion, excellent selectivity, and good stability. Here, we used metal-organic frameworks (MOFs) containing Zr metal nodes ("UiO") with tunable wettability and electron-withdrawing ability as activity accelerators for the semihydrogenation of alkynes catalyzed by sandwiched palladium nanoparticles (Pd NPs). Impressively, the porous hydrophobic UiO support not only leads to an enrichment of phenylacetylene around the Pd NPs but also renders the Pd surfaces more electron-deficient, which leads to a remarkable catalysis performance, including an exceptionally high TOF of 13835 h-1 , 100 % phenylacetylene conversion 93.1 % selectivity towards styrene, and no activity decay after successive catalytic cycles. The strategy of using molecularly tailored supports is universal for boosting the selective semihydrogenation of various terminal and internal alkynes.

Feature selection for genomic data sets through feature clustering.

A subset selected by a supervised feature selection method may not be a good one for unsupervised learning and vice versa. We propose a novel Feature Selection algorithm through Feature Clustering, FSFC. FSFC does not need the class label information in the data set and is suitable for both supervised learning and unsupervised learning. We test FSFC on some biological data sets for both clustering and classification analysis and the results indicates that FSFC algorithm can significantly reduce the original data sets without scarifying the quality of clustering and classification.