Pyrolysis kinetics and product distribution of α-cellulose: Effect of potassium and calcium impregnation

Highlights

• Thermogravimetric analysis and kinetic behavior of α-cellulose.

• The parameters of kinetic triplet in model reflect single-step mechanism.

• The predominant reaction pathway was A (α-cellulose) into B (liquid).

• Effect of AAEMs on kinetic and product distribution.

Abstract

Cellulose accounts for the largest proportion of lignocellulosic biomass. Herein, experimental and simulation studies are used to deeply understand the kinetic characteristics of the thermal decomposition of α-cellulose. The simulated data is in good agreement with the experimental data in the aspects of the conversion and the conversion rate versus temperature. The decomposition of α-cellulose, mainly occurring at 270–420 °C, induced an apparent activation energy ranging from 175.42 kJ/mol to 197.73 kJ/mol at a conversion of 10–90%. With 0.1–0.2 wt% K or Ca impregnation into the α-cellulose, the mean activation energy for pyrolysis was lowered (from 181.47 kJ/mol (for α-cellulose) to 141.11 kJ/mol (for 0.2 wt% K/α-cellulose) and 159.46 kJ/mol (for 0.1 wt% Ca/α-cellulose)) and higher amounts of liquid and gas products were produced. Furthermore, the addition of potassium and calcium increased the production of lower molecular weight components, such as furfural and its derivatives. The kinetic parameters of the α-cellulose pyrolysis were determined based on a nonlinear least-squares regression of the experimental data assuming first-order kinetics and correlated with the simulated result. The kinetic rate constants indicate that the predominant reaction pathway is from α-cellulose into a liquid product, rather than from α-cellulose into a gas product.

Introduction

Biofuel derived from biomass is being increasingly considered as an alternative and renewable fuel source owing to the growing energy shortage and the consequences of carbon dioxide emission [1,2]. The main components of lignocellulosic biomass include cellulose (40–60 wt%), hemicellulose (20–40 wt%), and lignin (10–25 wt%) [3,4]. Given that cellulose comprises the highest proportion of biomass, it has been extensively investigated, especially for pyrolysis kinetics [[4], [5], [6]].

The kinetic triplet, including the activation energy, frequency factor, and reaction model, plays an important role in the study of biomass pyrolysis. The activation energy is usually derived from different isoconversional methods such as Friedman method, integral Kissinger-Akahira-Sunose (KAS) method, Ozawa-Flynn-Wall (OFW) method, or the Kissinger method [7]. The accuracy of the Friedman, KAS, and OFW methods is higher than that of the Kissinger method because they estimate the activation energy at different conversional levels, whereas the Kissinger method considers E_a over the entire reaction. Indeed, based on the expression representing for Kissinger method, the linear regression is only fitted for entire process [8]. Specifically, one decomposition peak temperature ($T_p$) is determined at each heating rate ($\beta$) and one single value of activation energy for whole process is calculated from the slope of graph of $\ln \left(\beta /T_p^2\right)$ vs. $1/T_p$ [8]. Biomass generally comprises of various components and thus the activation energies for their thermal decomposition are different. Kissinger method showing the single activation energy could not meet the demand of complex kinetics. Unlike Kissinger method, the methods of Friedman, KAS, and OFW show the specific temperature with respect to certain conversion degree (X). Through this, the linear graphs ($\ln \left(\beta /T_X^2\right)$ vs. $1/T_X$ in KAS, $\ln \left(\beta /T_X^2\right)$ vs. $1/T_X$ in FWO, and $\ln \left(dX/dt\right)$ vs. $1/T_X$) in Friedman) are plotted at certain conversion degree (X). Thereby, the dependency of activation energy on degree of conversion is revealed, explains more reliably the decomposition kinetic compared to Kissinger method. According to expressions of the isoconversional methods [8], the frequency factor is determined based on the identification of reaction model which describes reaction mechanism in study on solid-state reaction kinetics. Reaction model is the function of conversion which can be identified by master plot method [7]. Pyrolysis, thermal decomposition in an oxidant-free environment, is a type of deaccelerating reaction and may occur according to three mechanisms: a diffusion-based model, contracting geometry model, or an order-based model [7,9]. Previous studies calculated the frequency factor for the 0^th, 1st, and 2nd order reaction models [10,11]. However, calculations based on the assumption of a reaction model order do not ideally represent the course of pyrolysis. As a result, it is necessary to determine a precise pyrolysis reaction model in order to describe the kinetics of the pyrolysis process more accurately.

A majority of biomass contains trace amounts of alkali and alkaline earth metals (AAEMs) such as potassium, calcium, sodium, and magnesium, which strongly affect the pyrolysis of biomass [12,13]. As a crucial mechanism for biomass industrialization, pyrolysis is commonly used for the conversion of lignocellulosic biomass among various thermochemical conversion processes. Therefore, the impact of AAEMs on the pyrolysis of cellulose is necessary for understanding and improving the pyrolysis process.

Several studies have investigated the effects of metallic ions on the pyrolysis of cellulose or biomass [12,14,15]. Patwardhan et al. [12] used various inorganic salts (NaCl, KCl, MgCl₂, CaCl₂, Ca(OH)₂, Ca(NO₃)₂, CaCO₃ and CaHPO₄) and switchgrass ash which were impregnated on pure cellulose. The study shows that primary pyrolysis reaction products are formed via competitive reactions resulting distribution of (a) low molecular weight species – formic acid, glycolaldehyde and acetol; (b) furan ring derivatives – 2-furaldehyde and 5-hydroxymethyl furfural; (c) anhydro sugars – levoglucosan. Hu et al. [15] investigated effects of inherent alkali and alkaline earth metallic species (AAEMs) on biomass pyrolysis at different temperatures. Increasing temperature could also promote depolymerization and aromatization reactions of active tars, forming heavier polycyclic aromatic hydrocarbons, leading to decrease of tar yields and species diversity. According to a parallel reaction scheme based on the Broido-Shafizadeh model [16], the decomposition of cellulose can be divided into two stages: (1) the formation of reaction intermediates, i.e. active cellulose, through depolymerization reactions and (2) their subsequent degradation. Inside of those stages, there are involving numerous reactions such as dehydration, depolymerization, isomerization, and fragmentation to produce char and volatile products. During the pyrolysis process, cellulose degradation was enhanced by the addition of AAEMs and increased decomposition of anhydro sugars for the formation of low molecular weight (LMW) species and char was observed [[17], [18], [19], [20], [21]]. Within the LMW species produced, the formation of hydroxyl acetaldehyde and acetone can be promoted by alkaline metals, while furans formation can be enhanced by alkaline earth metals [12]. To the best of our knowledge, most studies focus on the effect of AAEMs on cellulose pyrolysis have focused on examining elemental compositions of the pyrolytic products. However, the effects of AAEMs using a quantitative kinetics model of cellulose pyrolysis have not been well reported.

In this study, simulation and experimental approaches are used in concert to investigate the pyrolysis conditions and to determine akinetic model for cellulose pyrolysis. Simulation steps with a rigorous conceptual flowsheet are utilized to determine the kinetic triplet of α-cellulose pyrolysis, while experimental data are collected in a micro-tubing reactor to investigate the pyrolysis reaction time and temperature as well as the product distribution. More specifically for the simulation protocol, a master plot method was applied to identify the reaction model through a comparison of the calculated reaction model curves with the experimental data. Furthermore, the effects of K and Ca as AAEMs loading in α-cellulose on the pyrolysis kinetics and product distributions were systematically investigated. This step can reflect the effects of AAEMs and thus, control the reaction pathways of cellulose pyrolysis to the particular product target.