The Intrinsic Kinetics and Heats of Reactions for Cellulose Pyrolysis and Char Formation

Abstract

Thermochemical conversion of biomass starts with its pyrolysis. The intrinsic kinetics and heats of reactions for the three first reactions in cellulose pyrolysis are measured. The products from pyrolysis of cellulose can vary depending on the temperature of pyrolysis, with levoglucosan being formed at higher temperature. Furthermore, pyrolysis can be either exothermic or endothermic depending on the products formed. Cellulose is the most abundant organic polymer, comprising the largest fraction of lignocellulosic biomass. Recently, the conversion of biomass into biofuels has attracted significant interest due to the availability of biomass and its potential to provide sustainable liquid fuels.1 Pyrolysis, the thermal decomposition of a material without oxygen, is a promising route for converting solid biomass into liquid fuels.2 During pyrolysis, biomass undergoes different reactions to form three general classes of compounds: (1) noncondensable gases, (2) condensable vapors that can be condensed into a liquid product, and (3) solid char that consists of elemental carbon. Understanding the kinetics of char formation is critically important in designing more efficient thermochemical processes for biomass conversion.3 It is generally known that char is preferentially formed at low temperatures and slow heating rates.4 However, the kinetics and thermodynamics of char formation have been under debate in the literature.3b, 3d In this Communication we measure the intrinsic kinetics and heats of reaction of char formation from cellulose, and develop an intrinsic kinetic model for the pyrolysis of cellulose to anhydrosugars and char. Over the past few decades, several kinetic models describing cellulose pyrolysis have been developed.4b, 5 Most kinetic models of cellulose pyrolysis have been developed based on the decrease in sample weight loss and the characterization of product species. In contrast, little effort has been made to accurately quantify the heat released during pyrolysis. In the present study we have established an intrinsic kinetic model for cellulose pyrolysis that includes two parallel pathways for cellulose pyrolysis, to char and to anhydrosugars. We have also measured the heat released during the pyrolysis of cellulose, and performed both isothermal and dynamic experiments in a thermogravimetric analysis–differential scanning calorimetry (TGA–DSC) set-up that allows us to measure the heat of reaction as well as the reaction rate for each reaction step. We have previously shown that our data is collected in an intrinsic regime that is free from heat and transport effects.5j Our model is compared with experimental data to calculate the activation energies, pre-exponential factors, and heats of the individual reactions. At low temperatures (473 to 523 K), solid cellulose undergoes a reversible reaction to produce active cellulose.5j The exact structure of active cellulose is unknown but it probably is formed through an isomerization reaction in which hydrogen bonds within the cellulose are rearranged. When cellulose samples are exposed for longer periods of time, thermal degradation proceeds by two competetive reactions: solid char formation and volatile anhydrosugar formation.5c At fast heating rates the thermal decomposition of cellulose induces depolymerization of polymeric chains, reconfiguration of pyranose rings, and subsequent production of volatile anhydrosugars. As confirmed in a previous study,5j complete thermal decomposition under dynamic conditions at fast heating rates (1–150 K min−1 and 1 atm) produces small amounts of primary char (less than 5 %), which is material that contains primarily carbon. Together with the formation of char noncondensable gaseous products are released, including CO2, H2O, and H2. Char can also be formed from the repolymerization of volatilized anhydrosugars.6 Compared to pyrolysis in dynamic conditions, the complete decompositon of cellulose under isothermal conditions requires a relatively longer thermal degradation time, especially at the temperature near the lower limit of pyrolysis. The formation of char becomes a dominant reaction path at such slow heating rates and low temperatures.5b, 5c, 7 Equation (1) is the reaction of cellulose to form active cellulose, with forward and reverse rate constants k1 and k−1, respectively. Equation (2) is the reaction of active cellulose to form char, carbon dioxide, water, and hydrogen, with rate constant k2. Equation (3) is the reaction of active cellulose to form anhydrosugars, with rate constant k3. In addition to anhydrosugars, smaller C2 and C3 fragments such as hydroxyacetone and hydroxyacetaldehyde may be formed directly from cellulose, especially in the presence of alkali metals.2f All rate constants except k−1 are assumed to follow Arrhenius behavior, that is, ki=Aiexp(-EAi/RT), and first-order kinetics. The rate of the reverse reaction in Equation (1) was negligible compared to the other rates at all current experimental conditions for cellulose pyrolysis in this study, and therefore we did not include this rate constant in our modeling efforts. Our measured values for the kinetic and thermodynamical parameters are listed in Table 1, along with the values reported by Bradbury et al.5c Our kinetic parameters were calculated from isothermal experiments of cellulose pyrolysis from 473 to 573 K. These results are also consistent with previous cellulose pyrolysis studies done with a dynamic heating ramp with temperatures up to 1073 K.5j A TGA experiment was employed to measure the kinetics of each of these reactions. DSC was used to quantify the reaction heats for individual steps. Parameter Present work Reference[a] log10 A (A in s−1) EA [kJ mol−1] ΔH[b] (kJ equiv mol−1) log10 A (A in s−1) EA [kJ mol−1] k1 [s−1] 21.62±1.02 257.72±3.23 −1.5<ΔH1<0 19.45 242.67 k2 [s−1] 5.74±0.98 102.94±4.56 −170.17±7.27 10.12 153.13 k3[c] [s−1] 14.81±0.01 198.91±0.02 121.38±2.42 14.50 197.90 Figure 1 a shows a comparison of calculated and experimental weight changes for isothermal cellulose pyrolysis at temperatures from 513 to 573 K. The amount of char formation increases with decreasing temperature, as expected from the kinetic parameters in Table 1. At lower temperatures, longer times are required to reach the complete decomposition of cellulose. The kinetic parameters for the formation of levoglucosan (LGA) in these isothermal experiments are similar to the values previoulsy detemined from for formation of LGA in dynamic experiments.5j This result is not suprising because no significant amount of char is formed at such rapid heating rates, for which the global decomposition rates are estimated by thermal-lag models. The kinetic parameters for LGA formation are therefore fixed to reduce the degrees of freedom in the parameter estimation. We have previously shown that this data is collected in the absence of heat and mass transfer limitations.5j Weight changes for cellulose pyrolysis at isothermal conditions a) Experimental weight changes for total residue (symbols) and model estimation (curves), b) Predicted decomposition profiles at low temperature (513 K) and c) at high temperature (573 K). Curves are model estimations for the weight of total residue (solid lines), cellulose (dashed lines), active cellulose (dashed-dotted lines), and char (dotted line). Our kinetic model can be used to evaluate the weight fractions of chemical species in the residual mass, as shown for 513 and 573 K in Figure 1 b and c, respectively. At 513 K the conversion of cellulose to active cellulose occurs gradually. There is an initial induction period during which the cellulose is converted into active cellulose, followed by a period where weight loss occurs. In contrast, at 573 K the cellulose is almost instantenously converted into active cellulose, and then undergoes pyrolysis to anhydrosugars. The predicted active cellulose profile at 573 K quantitatively coincides with the experimental results by NMR.8 Figure 2 shows the calculated and measured final char amounts as a function of inverse temperature. As the reaction temperature increases the amount of char varies linearly with the inverse temperature (k3≫k2). The correlation expressed in Equation (4) is expected to be useful for the purpose of simulation of biomass pyrolytic reactors if they are operated without mass-transfer effects. Final amount of char formed by cellulose pyrolysis. Symbols are final residue weights measured by TGA. The solid line is the final char amount derived from the kinetic model and estimated parameters listed in Table 1. The heats of pyrolysis for the reactions shown in Equations (1–3) are also important to understand and model the thermal conversion of biomass. Several efforts have been made to quantify the enthalpy change during each step of biomass pyrolysis.3b, 3d, 6b, 9 However, the details on the reaction heats of individual routes (i.e., active cellulose, depolymerization, and char formation) are still controversial.3b, 10 There are significant differences between reported values. More surprisingly, it is not yet clear if char formation is endothermic or exothermic. In this Communication we have measured the enthalpies of reaction with DSC during cellulose pyrolysis under both dynamic and isothermal conditions. The resulting values were used to derive the individual enthalpy of pyrolysis for each step. DSC experiments were repeated for the same conditions with different initial sample amounts to eliminate extrinsic peaks in a DSC diagram and relative values are used for the estimation of reaction heat. In repeating experimental runs at the same temperature, the decomposition profiles for different amounts of initial sample loading are compared to confirm the absence of heat or mass transfer limitations. Measured sensible heats of initial cellulose and resulting char were considered to draw the baseline based on the corresponding decomposition profiles obtained by TGA. The DSC diagrams of cellulose decomposition at various temperatures are shown in Figure 3 a. Integrations of heat are presented in Figure 3 b as a function of the cellulose fraction f2 that is converted into char, CO2, H2O, and H2. In this plot, the fraction f2 was estimated from the final amount of char and total mole balance. In the dynamic pyrolysis conditions where higher heating rates are imposed, the overall enthalpy of pyrolysis was observed to be endothermic with a very small amount of char formed by thermal decomposition. a) Heat of cellulose pyrolysis at different temperatures. b) Enthalpy changes as a function of the fraction of cellulose converted into the char/gas products measured by isothermal (filled squares) and (open squares) dynamic experiments. The dashed6b and dotted3b lines are best fits of literature values. In cyclic DSC measurements between 323–523 K, it was previously observed that the partial conversion to active cellulose starts at around 473 K.5j A small exothermic peak was observed from the dynamic experimental conditions (β=1–15 K min−1) in the temperature range of 473–523 K prior to the onset of weight loss. Below 523 K, the weight loss was significantly delayed, as shown in the TGA diagram (Figure 1). Because the reaction rate of thermal decomposition is far slower compared to the conditions used, the initial exothermicity is thought to be of a different nature than the heats of char formation or depolymerization. We have therefore concluded that this initial exothermic peak is active cellulose formation. The integrated values were two orders of magnititude less than other thermal peak areas observed in the dynamic conditions. When the pyrolytic temperature increased in dynamic heating rates up to 773 K, only a negligible amount of residual mass was left in the sample pan. These integrated values for thermal peaks were assumed to be overall enthalpies with zero yield of char formation, as shown by the open symbols in Figure 3 b. The enthalpic contribution to cellulose decomposition without char/gas formation is limf2→0(ΔHtotal)=ΔH1+ΔH3. Because the reaction heat for active cellulose formation is relatively low, the reaction shown in Equation (1) is considered thermally neutral (or slightly exothermic) compared to the reactions shown in Equations (2) and (3). In the current study, ΔH2 and ΔH3 were measured as −170.17 kJ mol−1 (−1049.52 kJ kg−1) and 121.38 kJ mol−1 (748.59 kJ kg−1) respectively. Milosavljevic et al. reported a comparable value in their study of thermal effects on cellulose pyrolysis.3b They concluded that the slope of the overall heat of reaction changed by −20 kJ kg−1 per 1 % increase of char yield. For our experiments in this study, the heat changed by −51 kJ kg−1 per 1 % increase of char yield. The latent heat for the LGA phase transition (735 kJ kg−1) mostly contributed to their experimental results for the enthalpy of volatile product formation (538 kJ kg−1). Our experimental results indicate higher endothermicity for Equation (3), which is in better agreement with their measured values of the latent heat for LGA vaporization. Mok and Antal reported a similar result to ours which shows a higher exothermicity for the char formation (−36 kJ kg−1 per 1 % increase of char yield).6b The heat of reaction in their study was linearly dependent on char yield, consistent with our findings. Our intrinsic model for cellulose pyrolysis can now be incorporated into a reaction-transport model11 to model accurately the pyrolysis of larger cellulose particles in future work. The general conclusions of this work are: (1) At low temperature, in the range 513–573 K, the amount of char varied significantly. Isothermal experiments for cellulose pyrolysis can be modeled based on the B–S model and rate parameters reported in Table 1, which are consistent with other reported values in the literature. (2) Char formation has a lower activation energy than the competing anhydrosugar formation. The char yield therefore decreases with increasing temperature. (3) Char formation is exothermic, while anhydrosugar formation is endothermic. The isothermal experiments for cellulose pyrolysis were performed in a DSC–TGA (TA instruments SDT Q600 system) at the temperature range of 473 to 623 K, where a significant amount of char is formed. About 10–30 mg of initial cellulose (microcrystalline cellulose with 50 μm average particle size, Acros) was used for each run and weight changes were measured during the pyrolysis followed by preheating at 383 K for one hour. Then the sample was heated to a sample temperature at a rate of 150 K min−1. TGA was used to measure the weight changes of residual mass. Elemental analysis was performed to confirm the compositions of elements in the final solid char. DSC data were used to determine the reaction heats of pyrolysis. In order to eliminate the extrinsic thermal effect of the instrument, two experimental results for the same conditions but different initial weight were compared. In the analysis, their differential values of heat inclusion and weight change are used to estimate the heat of each reaction step. During the pyrolysis, the gaseous products were analyzed by GC (Shimadzu GC-2014 w/Supelco 80/100 HayeSep D packed and Restek Rtx-VMS capillary columns) at the exhaust of the DSC-TGA instrument to quantify the amount of gas and vaporized chemical species. In the analysis, CO2, H2O, and H2 were detected as gaseous chemical species when the cellulose is pyrolyzed under the helium carrier gas environment. We thank Professors Yu-Chuan Lin, Geoff Tompsetts, Phillip R. Westmoreland, and Paul Dauenhauer for very helpful discussions and help with experiments. This work was supported by the National Science Foundation through an EFRI grant. Detailed facts of importance to specialist readers are published as ”Supporting Information”. Such documents are peer-reviewed, but not copy-edited or typeset. They are made available as submitted by the authors. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.