Development and validation of a novel multi-reaction model for hydrothermal carbonization of biogenic materials

Abstract

The increasing demand for energy, rapid depletion of fossil fuels, and global concern on climate change and sustainability make it imperative to find alternative energy sources, prefer- ably from renewable sources. Biomass to biofuels is one of the leading contenders to replace traditional fuels, which may follow either thermal or biochemical conversion pathways. Among the thermal pathways, hydrothermal carbonization (HTC) has gained much traction in recent years for its high carbon efficiency. Despite the increasing research on HTC, the mechanism and kinetic of the process are not fully understood. Most of the kinetic models of HTC assume the reaction rate as a single first-order reaction defined by the Arrhenius equation. The models developed so far used single model compounds of biomass constituents and cannot directly be applied for other biogenic materials. In this study, a novel kinetic model was developed for pseudo-components of biomass to be used in a multi-reaction model for HTC. Batch experi- ments for kinetics were conducted at temperatures of 220, 235, and 250°C over time periods ranging from 5 minutes to 1 hour. The HTC kinetics of hemicellulose (xylan), cellulose, and lignin were approximated using the Arrhenius equation. The formation of solid char after HTC of hemicellulose was modeled as a first-order reaction kinetic process, with an activation en- ergy of 46.79 kJ/mol and an Arrhenius pre-exponential factor of 46.99 s−1. For cellulose, the char formation kinetics were modeled using two first-order reactions in series. The activation energies were 175.83 kJ/mol and 181.80 kJ/mol, with pre-exponential factors of 2.03 × 1014 s−1 and 2.06 × 1016 s−1. The kinetics of lignin were modeled as a zero-order reaction, with an activation energy of 44.24 kJ/mol and a pre-exponential factor of 7.88 molL−1s−1. These kinetic models were then used for different mass compositions of cellulose and lignin (1:1, 3:1, 1:3) and xylan, cellulose, and lignin (1:1:1, 1:3:1), and successful predictions were obtained by the model. However, deviations from the model were observed when lignin was present in high amounts (1:1:3) in the presence of inorganics. Yields for most lignocellulosic biomass from the literature were successfully predicted by the kinetic approach developed in this work. Addition- ally, results from the organic fraction of municipal solid waste (OFMSW) at low temperatures were successfully approximated by the model.