Effect of Phosphorus Precursor, Reduction Temperature, and Support on the Catalytic Properties of Nickel Phosphide Catalysts in Continuous-Flow Reductive Amination of Ethyl Levulinate.

Levulinic acid and its esters (e.g., ethyl levulinate, EL) are platform chemicals derived from biomass feedstocks that can be converted to a variety of valuable compounds. Reductive amination of levulinates with primary amines and H2 over heterogeneous catalysts is an attractive method for the synthesis of N-alkyl-5-methyl-2-pyrrolidones, which are an environmentally friendly alternative to the common solvent N-methyl-2-pyrrolidone (NMP). In the present work, the catalytic properties of the different nickel phosphide catalysts supported on SiO2 and Al2O3 were studied in a reductive amination of EL with n-hexylamine to N-hexyl-5-methyl-2-pyrrolidone (HMP) in a flow reactor. The influence of the phosphorus precursor, reduction temperature, reactant ratio, and addition of acidic diluters on the catalyst performance was investigated. The Ni2P/SiO2 catalyst prepared using (NH4)2HPO4 and reduced at 600 °C provides the highest HMP yield, which reaches 98%. Although the presence of acid sites and a sufficient hydrogenating ability are important factors determining the pyrrolidone yield, the selectivity also depends on the specific features of EL adsorption on active catalytic sites.

Crystal Structure Representation for Neural Networks using Topological Approach.

In the present work we describe a new approach, which uses topology of crystals for physicochemical properties prediction using artificial neural networks (ANN). The topologies of 268 crystal structures were determined using ToposPro software. Quotient graphs were used to identify topological centers and their neighbors. The topological approach was illustrated by training ANN to predict molar heat capacity, standard molar entropy and lattice energy of 268 crystals with different compositions and structures (metals, inorganic salts, oxides, etc.). ANN was trained using Broyden-Fletcher-Goldfarb-Shanno (BFGS) algorithm. Mean absolute percentage error of predicted properties was ≤8 %.