(560ag) Main Effects Observed in the Conversion of Fructose in HMF

5-hydroxymethyl-2-furfural (HMF) is a furanic compound considered as an important chemical platform once it is obtained from renewable raw materials and it has wide application in the industry, acting as an intermediate in the production of fine chemicals, polymer materials, biofuels and pharmaceuticals. This work aimed an experimental study of the reaction of HMF formation by heterogeneous catalysis, varying catalysts and reagents, with determination of yield and selectivity for optimization, as well as kinetic data. Experiments in aqueous medium were carried out in a Parr reactor using a combination of a substrate (glucose or fructose) and a catalyst (macroporous or supergel ion exchange resins). The maximum HMF yield was 20 % in 3.5 h and the conversion was 84 % after 7.5 h in the reaction of fructose catalyzed by the supergel resin (named SGC650H) at 110 °C, which indicated a poor performance towards HMF using water as solvent. This result is attributed to the HMF rehydration under high temperature conditions to produce levulinic and formic acids, so the best alternative to increase HMF selectivity is to use anhydrous solvents. Using dimethylsulfoxide (DMSO) as solvent, the yields significantly increased to almost 100 % in 3 h reacting at 110 °C. Besides conducted in batch reactions, experiments in continuous packed bed reactors were also performed with reaction systems containing fructose, DMSO and SGC650H resin. Analyzing the reactions performed with this system in both reactors, the increase in temperature and the reduction of the particle size of the catalyst favors the HMF formation, in contrast to the fructose feed concentration. Considering a pseudo first order kinetic model, the available data allowed to calculate rate constants and Arrhenius parameters, the latter varied with the concentration of reagent fed. This variation was linear in the Cremer-Constable diagram, which means that the model was insufficient to represent the reaction. Langmuir-Hinshelwood rate equations (LHHW) models were also tested by the rate determining step method, but none had adjusted to the data, and thus a rate law was not determined by the methods applied.