A modified kinetic analysis method of cellulose pyrolysis based on TG–FTIR technique

Highlights

• An improved reaction model of cellulose pyrolysis was proposed.

• The random scission model is used for cellulose pyrolysis.

• The gaseous products and functional groups releasing mechanism was included.

• Random scission model is proper to describe the cellulose pyrolysis kinetics.

• Same kinetic parameters could describe non-isothermal and isothermal data.

Abstract

The pyrolysis characteristics of cellulose was investigated using TGA and TG–FTIR. An modified reaction model was proposed to study cellulose pyrolysis. In this model, the active cellulose ($E=137\mathrm{}\mathrm{k}\mathrm{J}/\mathrm{m}\mathrm{o}\mathrm{l}$) and dehydrocellulose ($E=91.3\mathrm{}\mathrm{k}\mathrm{J}/\mathrm{m}\mathrm{o}\mathrm{l}$) were formed by parallel mechanism. The degradation of the active cellulose was followed by two competitive pathways, this is, one that formed levoglucosan ($E=159.2\mathrm{}\mathrm{k}\mathrm{J}/\mathrm{m}\mathrm{o}\mathrm{l}$) and another that formed furan and some low molecular weight products ($E=203.5\mathrm{}\mathrm{k}\mathrm{J}/\mathrm{m}\mathrm{o}\mathrm{l}$). The random scission model was used to describe the kinetics of cellulose pyrolysis. The same kinetic parameters were used for different linear heating rates (10–100 °C/min) and isothermal segment conditions. By successfully modeling non-isothermal and isothermal segment data, the model was proved to be credible in describing cellulose pyrolysis.

Introduction

Energy shortage and global warming have paved the way for increased utilization of biomass as a feedstock for thermochemical degradation processes. Cellulose is the most essential element in biomass due to its large proportion [1,2]. Its pyrolysis is a complex process that involves multiphase reactions, complex chemical pathways, and heat and mass transfer effects. In order to control product distributions by manipulating reaction conditions, such as heating rates, particle size, and product retention time, understanding the pyrolytic behavior of cellulose is fundamental and essential to biomass thermochemical conversions [3].

An evident variation exists in the magnitude of cellulose pyrolysis kinetics and in the measurement approaches used in previous studies. Broido and Nelson firstly proposed a global cellulose pyrolysis reaction model based on TGA, which involved two competitive reactions occurring directly from the cellulose, that is, one that yielded char with low molecular weight at low temperatures and another resulted in tar at high temperatures [4,5]. Bradbury and Shafizadeh proposed a modified model that cellulose first undergoes depolymerization to produce an intermediate, called “active cellulose.” This empirical Bradbury–Shafizadeh (B–S) model had been widely accepted by many researchers, except in few arguments about the existence of active cellulose [[6], [7], [8]]. However, the B–S model does not describe the decomposition mechanism in detail. Piskorz proposed a modified version, namely, the Waterloo scheme in consideration of a large amount of hydroxyacetaldehyde (HAA). The initial stage considered the competitive formation of char and active cellulose, followed by ring fragmentation (high activation energy$E$), which led to HAA and depolymerization (low activation energy) to yield levoglucosan (LG) [9]. Banyasz proposed a similar scheme, except the char was initiated from an intermediate called dehydrocellulose [10]. Shen et al. discussed the secondary reactions and formation pathways of typical products by using TGA–FTIR and Py–GC–MS analytical techniques and interpreted that the sequential and competitive mechanisms occurred between the LG and other products, such as HAA, hydroxyacetone (HA), and 5-hydroxymethyl-furfural (5-HMF) [11]. Lu et al. intensively investigated the formation mechanism of HAA. They claimed that the HAA and LG mainly formed via competing parallel pyrolytic pathways and that only a small amount of HAA would be derived from the secondary cracking of the LG [12].

Furthermore, product distribution and thermal transfer resistance are significantly influenced by heating rates. Milosavljevic and Suuberg reported the diverse kinetic parameters derived by various groups based on different heating rates [13]. Generally, the higher the heating rates, the lower the active energy was found [14]. Antal et al. proposed the secondary reactions of primary tar, which produces char and gas, and claimed that a universal rate law had not been observed due to the thermal lag that interfered with kinetic measurement [15]. More recently, Lin et al. confirmed the conclusion of Antal et al. (that is, the shift in pyrolytic temperature under different heating rates is due to the effect of thermal lag), coupled the first-order reaction model with two thermal lag estimations, and obtained the intrinsic kinetic parameter at $\mathrm{}E=198\mathrm{}\mathrm{k}\mathrm{J}/\mathrm{m}\mathrm{o}\mathrm{l}$ [16].

In the present work, a modified reaction model was proposed for cellulose pyrolysis, which was applicable for describing cellulose pyrolytic pathway. In order to obtain the kinetic parameters and to validate the model, TGA experiments were performed under different heating rates and retention temperatures. The volatile products under different retention temperatures was tested by real time TG–FTIR.