Key points are not available for this paper at this time.
Bio-oil obtained from fast pyrolysis of biomass has the potential to be upgraded to renewable, drop-in fuel intermediates. The quality of bio-oil produced is critical to downstream process development and is determined by a number of factors like reaction timescale, extent of primary and secondary reactions, and the presence of minerals. In this work, we provide a mechanistic understanding of the various competing reactions in fast pyrolysis of cellulose and other glucose-based carbohydrates through a unified microkinetic model. The model incorporates the reactions of the cellulose chain and of the glucose intermediate to form a variety of bio-oil components, which are confirmed by either experiments or theoretical calculations reported in the literature. The majority of the Arrhenius rate parameters were based on existing studies, while a small number were fitted to match the experimental product distribution over a wide range of temperatures. The model yields of all the major primary fast pyrolysis products, viz., levoglucosan, formic acid, glycolaldehyde, 5-hydroxymethyl furfural, furfural and char, match reasonably well with the experimental data over the temperature range of 400–550 °C. The model predicted that the time required for complete conversion corresponds to 2.5–3 s. Net rate analysis identified key pathways to levoglucosan via either mid-chain initiation, or depropagation and end-chain initiation that were dominant over the timescales of 0–100 ms and 100–1000 ms, respectively, at 500 °C. Unlike the existing lumped kinetic models, our model provides a detailed picture of levoglucosan formation that involves mid-chain glycosidic bond cleavage followed by unzipping of levoglucosan from the chain ends. The model, utilizing the same set of rate coefficients, was able to predict the dominant products of fast pyrolysis of maltohexaose, cellobiose and glucose that also are in good agreement with experimental data. The initial chain length of cellulose was found to have a profound effect on the product yields for chain lengths shorter than 50, while it had no effect on the total reaction time.
Vinu et al. (Sun,) studied this question.