- Email: [email protected]
Pyrolysis behavior and kinetics of biomass derived materials
Journal of Analytical and Applied Pyrolysis 62 (2002) 331– 349
www.elsevier.com/locate/jaap
Pyrolysis behavior and kinetics of biomass derived materials Travis Fisher 1, Mohammad Hajaligol *, Bruce Waymack, Diane Kellogg Philip Morris Inc., Research Center, Richmond, VA 23261, USA Received 5 May 2000; accepted 21 November 2000
Abstract From previous biomass decomposition studies, it is well established that thermolysis generally occurs between 200 and 400 °C. For most materials, this temperature range constitutes up to 95% of total degradation; nonetheless, secondary decomposition reactions continue to occur in the solid matrix above 400 °C. The extent of these reactions, as indicated by the material loss above 400 °C, is small, and in the past has been either ignored or included in the primary degradative step. However, this latter step (reactions above 400 °C) exhibits many unique characteristics that differentiate it from the primary pyrolysis step and therefore needs to be treated separately. Additionally, it is widely accepted that primary decomposition of biomass material ( B 400 °C) consists of a degradative process, whereas the secondary thermolysis ( \ 400 °C) involves an aromatization process. In this study, It is shown that the latter step can be deconvoluted from the primary decomposition step, particularly for materials with limited aromaticity, such as cellulose and other carbohydrates. A thermogravimetric analyzer (TGA) coupled with a differential scanning calorimeter (DSC) and mass spectrometer (MS) was used to pyrolyze materials such as cellulose, xylan and other carbohydrates; and a pyroprobe interfaced with a gas chromatograph (GC) and mass spectrometer for product identification. In addition to monitoring the major products for each step, one also monitored and compared the temperatures corresponding to the maximum rate. From this analysis, the DTG results at 20 °C min − 1 heating rate show that the temperature differences between the peak temperatures of the two decomposition steps are approximately 70, 190, and 200 °C for cellulose, pectin, and xylan, respectively. Furthermore, TGA data were used to calculate sets of biomass-specific kinetic parameters for these two steps. © 2002 Elsevier Science B.V. All rights reserved. Keywords: Pyrolysis; Biomass; Kinetics; Cellulose; Pectin; Xylan * Corresponding author. Tel.: + 1-804-2742419; fax: +1-804-2741994. 1 Present address: Department of Pharmacy, Medical College of Virginia Campus of Virginia Commonwealth University, Richmond, VA 23298, USA. 0165-2370/02/$ - see front matter © 2002 Elsevier Science B.V. All rights reserved. PII: S0165-2370(01)00129-2
332
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349
1. Introduction Past literature reviews have woven a concise and intricate tale from the extensive data generated over decades of research regarding the thermo-chemical degradation and pyrolytic conversion of biomass substrates [1– 10]. The majority of biomass materials are comprised of approximately 50% by weight of cellulose, which has contributed to the extensive kinetic and thermal breakdown data available for cellulose [1– 5]. The thermolytic behavior of biomass materials primarily depends on its chemical composition and structure. The degree of crystallinity and polymerization of the starting materials are integral in defining their respective thermal degradation behavior. In addition to the above parameters, experimentally derived kinetic parameters also depend on the specific pyrolysis conditions that include temperature, heating rate, pressure, particle size, ambient gas environment, and the presence of ash or mineral deposits within the substrates [11– 19]. Additionally, transport processes, such as inter- or intra-particle heat and mass transfer, could also affect the degrading process, and the exclusion of these phenomena would limit the prediction of thermolytic behavior and kinetic parameters [20,21]. The reaction mechanisms and chemistry of biomass decomposition have been thoroughly elucidated in previous literature reviews; consequently, these topics are excluded from the current discussion. Likewise, one only briefly reviews the kinetics of cellulose, leaving the bulk of the kinetic analysis for specific biomass chars. By the mid 1970s, Briodo et al. [22] established the basic kinetic model for cellulose pyrolysis, which as shown in Scheme 1, involves a competitive, multistep reaction mechanism. This proposed model has now expanded to include many other biomass materials, with some minor modifications by Shafizadeh [3,12], Richards [23], and Antal [7,8], as well as many others [9]. These modifications account for the slight differences in chemical structure and composition between biomass materials. It is widely accepted that the primary pyrolysis of biomass materials generally takes place in the range of 200– 400 °C, which results in bulk product volatilization, as well as the formation of a solid char residue. Once the temperature increases above 400 °C, products continue to slowly evolve, as the char residue undergoes further chemical and physical transformations. This occurs more slowly than during primary pyrolysis. This may not be important in the overall scheme for the decomposition of biomass materials; yet, it may be very important in terms of understanding the chemistry behind aromatic formation at temperatures as low as 450 °C. For example, the gasification or combustion of biomass materials yields
Scheme 1.
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349
333
Fig. 1. Schematic of the TGA/DSC/MS system used in this investigation.
solid particles and gas phase products that undergo similar processes, which limits the usefulness of assessing the early char formation. However, smoldering processes are somewhat different, whereby the vapor products are separated from the solid particles at a lower temperature. This separation of solid and vapor products is key to unlocking the formation mechanism(s) for certain distinct hydrocarbons at higher temperatures. In this work, the goal is to investigate and broaden the understanding of the latter steps of biomass pyrolysis, which Broido identified as reaction steps 3 and 4 in Scheme 1 [22]. Kinetic parameters were calculated to describe the thermolytic behavior observed for small samples of cellulosic materials, such as cellulose, hemi-cellulose (xylan) and pectin. The primary emphasis is on the volatilization and solid reactions occurring above 400 °C. The pyrolysis conditions are selected to substantially reduce the effects of heat and mass transfer on the experimental results. 2. Experimental The experimental work was carried out using a Netzsch thermogravimetric analyzer (TGA) coupled with a differential scanning calorimeter (DSC) and a quadrupole mass spectrometer (MS), a platinum coil CDS pyroprobe interfaced with a GC/MS, or an environmental scanning electron microscope (E-SEM). Samples were also sent for elemental analysis. The TGA/DSC/MS facilitated the acquisition of weight loss, heat flow, and mass spectra data as a function of heating rate and temperature (or time). The detailed description of the experimental setup is provided in detail by Hajaligol et al. [24]. However, a brief summary of the setup is as follows: a small sample of starting material (2–15 mg) is weighed and spread evenly in a sample cup; the cup is then placed on the balance sample holder; the startup protocol is initiated; and finally, the sample is heated at a constant heating rate (1–60 °C min − 1) to a desired temperature (700–1000 °C) using helium as the carrier gas at a constant flow rate of 150 ml min − 1. A schematic of the setup is illustrated in Fig. 1. The CDS pyroprobe/GC/MS was used for the further identification and quantification of detected masses.
334
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349
To observe the physical behavior of the starting materials undergoing heating, an E-SEM was used. A small sample ( 1 mg) of each biomass material was separately placed on the hot stage of an E-SEM. The stage was then heated at a rate of approximately 5 °C min − 1 to around 500 °C under an Argon/Hydrogen (70/30) atmosphere and a low pressure of 1.5–4.0 Torr. Frequent scanning with the E-SEM on the hot stage allowed visual characterization of the physical changes occurring at the surface of the heated particles. Small samples of cellulose, hemi-cellulose (xylan) and pectin were used in this study; their chemical structures are depicted in Fig. 2. The cellulose samples were from Avicel (PH-102), with a particle size of roughly 40 mm or less and an ash content of B 0.0008% by weight as determined by elemental analysis (Table 1). The hemi-cellulose samples were from a hardwood birch xylan by Fluka comprised of 95% a-D-xylose units, with methyl glucuronic acid side-chain saccharide units and about 2.5% ash content (per manufacturer). It is a fine crystalline solid, with particle sizes ranging from a few microns up to about 100 mm. The pectin sample was purchased from Acros; it is a polymer consisting of galacturonic acid units, with 67– 71% degree of esterfication and only about a 0.5% ash content (per manufacturer). CPC and NMR analysis of the pectin showed that it contained 35% glucose. In an effort to more accurately detect and characterize the higher temperature reactions and the evolved products of the aforementioned substrates, pre-charred samples of the starting materials were prepared between 350 and 360 °C for 10–15 min under a Helium flow. The organic elemental analysis of the samples used in this study is given in Table 1. In order to illustrate the reaction differences above and below 400 °C, specific masses were monitored with the mass spectrometer coupled to the TGA. Precharred samples of the starting materials were used to accurately monitor mass evolution above 400 °C. Pre-charred samples were used given the fact that the amount of material remaining after primary pyrolysis (typically 5 1 mg) was insufficient to allow accurate mass detection. Therefore, a precharred sample was prepared in a tube furnace, and 10 mg samples of precharred material were used to assess secondary pyrolysis. The evolution of the following masses was monitored as a function of temperature for the heating rates 1, 5, 20, and 60 °C min − 1: 2, 15, 78, 92, 106, 116, 118, 128, 166 and 178, which primarily and respectively represent Table 1 Elemental analysis of materials investigated in this study Wt.%
C
H
O
N
C/H
C/O
Ash
Cellulose Cellulose char Xylan Xylan char Pectin Pectin char
43.99 61.33 40.96 64.26 42.16 64.27
6.20 4.83 6.36 4.40 5.85 3.98
49.79 29.77 47.30 21.73 51.70 27.89
B0.001 B0.001 B0.5 B0.5 0.25 0.57
7.10 12.70 6.44 14.60 7.23 16.15
0.88 2.10 0.87 2.96 0.82 2.30
B0.008 0.008 4.53 13.95 0.50 0.95
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349
Fig. 2. Schematic of chemical structures of the materials used in this investigation.
335
336
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349
hydrogen (2), methane (15), benzene (78), toluene (92), xylene (106), indene (116), benzofuran (118), naphthalene (128), fluorene (166), and anthracene/phenanthrene (178).
3. Results and discussion In this section, one first describes the physical changes in the samples of cellulose, xylan, and pectin as a function of temperature as samples are heated and decomposed. Then, the thermogravimetric data and mass spectra collected for the different materials are analyzed to confirm the multi-step pathway in their thermodegradative process; and finally, the decomposition kinetics are discussed and the kinetic parameters for the samples are estimated. The E-SEM and high resolution SEM were used to observe the qualitative physical changes in the solid particles during sample degradation. Fig. 3(a–c) shows photomicrographs of starting and pyrolyzed (400 °C) samples for cellulose, xylan, and pectin particles, respectively. The cellulose particles retain their mechanical integrity and shape; although, they shrink slightly as they are heated and charred. The cellulose particles are fibrous and form a network of fibers that barely contact the heated crucible; therefore, heating under vacuum may not uniformly heat the particles. Unlike cellulose, particles of pectin and xylan show signs of softening or melting (150 and 200 °C, respectively), as well as bubble formation as they are heated. As shown in Fig. 3, xylan and pectin particles soften at lower temperatures and form a porous structure at higher temperatures. These temperatures correspond to the maximum temperature peaks observed for melting in TGA/DSC. Although the heating rate is slow and the sample weight is small, it is observed that smaller particles soften faster and coalesce with larger particles, resulting in the formation of large globular particles. Examination of the char sample morphology using SEM shows the rupture of the bubbles that are formed during pyrolysis. Bubbles are formed within the particles, and as the bubbles burst, volatile products are released into the gas stream. This observation provides an interesting insight into the role of transport processes in the study of solid organic materials. For non-melting (non-softening) substances, such as cellulose, particles retain their original shape; thus, estimation of heat and mass transfers becomes more readily predictable. However, melting substrates, such as xylan and pectin, coalesce and form round or spherically shaped liquid particles. The estimation of intra-molecular heat and mass transfers requires more accurate knowledge of process dynamics. The observation shows that the melting and bubbling processes become more vigorous as the heating rate increases. Fig. 4(a– c) shows the respective melting and bubbling phenomena for cellulose, xylan, and pectin when particles are heated at a rate of several thousand degrees per second. Under these conditions, even cellulose particles exhibit melting and bubbling (Fig. 4(a)). The derivatives of the thermogravimetric weight loss data (DTG) are shown as a function of temperature at specific heating rates (5, 20, and 60 °C min − 1) in Fig. 5(a,b) through Fig. 7(a,b) for cellulose, xylan, pectin and their chars, respectively.
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349
337
Fig. 3. SEM photomicrographs of: (a) starting and pyrolyzed (400 °C) cellulose samples; (b) starting and pyrolyzed (400 °C) xylan samples; (c) starting and pyrolyzed (400 °C) pectin samples.
338
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349
Fig. 4. SEM photomicrographs of pyrolyzed materials at high heating rate ( \ 1000 °C s − 1): (a) cellulose, (b) xylan, and (c) pectin.
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349
339
Fig. 5. Pyrolysis rates (DTG) of cellulose samples: (a) starting material and (b) char prepared around 350 °C for 10 min.
Depending on the heating rate and starting material, the major decomposition of the starting substrate is completed below 400 °C (Figs. 5(a), 6(a) and 7(a)). For xylan and pectin, the thermo-degradative temperature is below 300 °C. Above 400 °C, there was no appreciable weight loss observed, which would have suggested the existence of another major reaction. This remained true for even the highest heating rate (60 °C min − 1) used in this study. This is better illustrated by the curves in Fig. 8(a– c); the weight loss seen with a 20 °C min − 1 heating rate is extremely small above 400 °C. This observation created the perception that no
340
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349
major reactions existed above 400 °C; and thus, suggested that studies above 400 °C would be unnecessary. However, analysis of chars prepared at temperatures (350–360 °C for 10– 15 min) exceeding their primary decomposition temperature has shown the contrary. Further heating of these chars above 400 °C reveals a secondary thermolytic process with a significant weight loss, as illustrated for cellulose, xylan, and pectin in Figs. 5(b), 6(b) and 7(b), respectively. For additional comparison, Fig. 8(a– c) show the representative and respective weight loss and derivative of weight loss curves for samples of cellulose, xylan, pectin, and their chars at a heating rate of 20 °C min − 1.
Fig. 6. Pyrolysis rates (DTG) of xylan samples: (a) starting material and (b) char prepared around 350 °C for 10 min.
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349
341
Fig. 7. Pyrolysis rates (DTG) of pectin samples: (a) starting material and (b) char prepared around 350 °C for 10 min.
Weight loss data for Avicel cellulose pyrolysis are presented in Fig. 8(a). At a heating rate of 20 °C min − 1, cellulose rapidly decomposes between 300 and 350 °C; this corresponds to a weight loss of around 91%. Following this sharp weight loss, there is a gradual, but small, weight loss (additional 5% of starting cellulose) up to about 700 °C. This gradual weight loss can be more readily observed and detected as another step in decomposition if the pre-charred sample of Avicel cellulose is heated separately. As can be seen from Fig. 8(a), most of the weight loss for the pre-charred cellulose takes place between 350 and 700 °C. The derivatives (DTG) of the two weight loss curves are also given in Fig. 8(a). The DTG data provide two distinct Tmax, one around 335 °C for primary reactions and
342
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349
one around 420 °C for secondary reactions. The latter Tmax depends on the heating rate and the conditions under which pre-charred samples are prepared, and could be as high as 500 °C. The breakdown of pre-charred cellulose has been cited in the literature as a small exotherm occurring after the larger endothermic decomposition [9,22]. The data suggest that this second step reaction is most likely thermo-neutral [25], depending on the condition under which pre-charred samples are prepared. This observation
Fig. 8. Thermogravimetric data (TG and DTG) of different samples at 20 °C min − 1 heating rate: (a) cellulose and pre-charred cellulose, (b) xylan and pre-charred xylan, and (c) pectin and pre-charred pectin.
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349
343
Fig. 8. (Continued)
is depicted in Fig. 9. All samples are initially pre-charred under various heating (temperature and time) conditions. The figure represents the pyrolysis of precharred samples that have undergone their primary reactions to a variable extent. The weight loss and corresponding heat of reaction of these samples measured by TGA/DSC are shown in Fig. 9. If a pre-charred sample is observed to give a weight loss of up to 70% in the TGA, the net heat of reaction is almost zero (thermo-neutral). Conversely, if the pre-charred sample gives greater than 70% weight loss in the
Fig. 9. Heat of pyrolysis of cellulose samples as a function of decomposition percent.
344
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349
TGA, then it is still undergoing net endothermic reactions. The DSC data for cellulose pyrolysis shows an endothermic heat of reaction of about 450 J g − 1. There is no indication of any exotherm during pyrolysis of cellulose or pre-charred cellulose samples. Regardless of the cellulose sample size (2–10 mg), no change is observed in the DSC peak temperature or the amount of reaction endothermicity. Weight loss data for xylan and its char at 20 °C min − 1 are given in Fig. 8(b). The xylan sample shows two distinct reactions below 300 °C that can be seen from the uncharred TG and DTG curves; this has an attributed weight loss of approximately 70%. Similar to cellulose and pectin, the weight loss and corresponding rate of weight loss above 400 °C appear insignificant. However, when pre-charred samples of xylan are heated in Helium at a 20 °C min − 1 heating rate, a reaction (or a series of reactions) can be observed to take place between 400 and 600 °C, as well as another one above 600 °C. During this secondary pyrolysis, an additional 10% weight loss was observed. The high temperature reaction(s) is believed to arise from the inorganic content [19,24] of the xylan (4.5% ash content), as seen in Table 1. The endotherm associated with the decomposition of the current sample of xylan is spread over a broad temperature range, and its integration over this temperature range would introduce a major error in the estimated values. Therefore, it can only be concluded that one endotherm starts at around 150 °C and ends near 300 °C, which coincides with the completion of decomposition reactions. The thermogravimetric data of pectin and pectin char are presented in Fig. 8(c). At a heating rate of 20 °C min − 1, pectin (shown in Fig. 8(c)) rapidly decomposes (approximately 70%) around 250 °C. The difference between the pectin and cellulose decomposition temperatures is believed to be due to greater molecular structure stability, and that pectin has side groups that undergo elimination reactions at lower temperatures. Analogous with the observations for cellulose, the weight loss for pectin appears insignificant above 400 °C. When pre-charred samples of pectin are heated, an additional weight loss of around 10% is observed above 400 °C. Comparison of the DTG data provides two distinct Tmax, one for the starting pectin at around 235 °C, and one for pre-charred sample at around 420 °C. In contrast to the cellulose samples, pectin decomposition is not only less endothermic, but the endotherm is also followed by a small exotherm. Pectin first melts at around 150 °C with a heat of melting of about 75 J g − 1, and then it undergoes primary decomposition above 200 °C, corresponding to an endothermic reaction of about 90 J g − 1. In order to illustrate the reaction differences above and below 400 °C, specific masses were monitored with the mass spectrometer coupled to the TGA. The masses represent what evolves during secondary pyrolysis. The activation energies for each mass correspond nicely with that for the entire secondary thermolytic process. Below 400 °C, the major evolved products are from the primary decomposition reaction and are widely validated in the literature [26– 28]. Pre-charred samples of the starting materials were used to accurately monitor mass evolution above 400 °C. The presence of the aforementioned hydrocarbons from the char was confirmed by the Py-GC/MS, which showed that these compounds generally account for the majority of ascribed masses. However, there are other compounds
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349
345
Fig. 10. Cumulative amounts of different masses as a function of temperature for pre-charred xylan samples.
that contribute to different extents to the specific masses monitored during the TG/MS experiment. For instance, several oxygenated benzoid compounds contribute to mass 78; likewise, there are also significant oxygenated compounds (pyrones) contributing to mass 128. Additionally, pyrolysis of pre-charred samples eliminates the risk of mistaking secondary reactions of the primary products as being part of the secondary decomposition, as well as focuses on the specific temperature ranges where hydrocarbons evolve from these materials. One has previously used masses 78, 92, 106, 116, 118, 128, 166, and 178 to follow evolution profiles of aromatics from the samples of cellulosic materials [24]. Experimental results show that all the hydrocarbons from char samples evolve over a similar temperature range (illustrated for xylan in Fig. 10) with a few exceptions; the evolution of methane and benzene (second peak) occurs above 500 °C and hydrogen evolves at temperatures above 600 °C. These results indicate that substituted benzene, as well as larger aromatic and aliphatic hydrocarbons, evolve first, and as they are depleted, there is a seemingly continuous evolution of methane and benzene. As sources of CH4 and benzene are depleted above 600 °C, hydrogen production predominates. At this point, a highly carbonaceous char has been produced. This is consistent with previous work [29–31], where the aliphatic and aromatic concentrations of the cellulosic char were studied as a function of temperature. For example, Shafizadeh [29] did not measure the evolved gases in his study; yet, his data on solid char concentrations is in good agreement with the data on the evolved gases over similar temperature ranges. The relative amounts of some aromatics from cellulose chars were measured using the Py-GC/MS, for which, the results are tabulated in Table 2. In this table, the abundance of the masses is for the individual compounds represented by their respective mass number.
346
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349
4. Reaction kinetics The thermogravimetric data (TG and DTG) was used to characterize the decomposition reactions of the starting biomass materials, as well as provide estimates of their kinetic parameters. Furthermore, the TG/DTG data from precharred samples was used to characterize product formation reactions and evolution. Initially, a Netzsch kinetic software package was used to estimate the kinetic parameters for the reactions. This kinetics software operates based on a leastsquares regression model and provides an optimal value for the parameters based on the minimum mean of residuals. Cellulose has a relatively simple chemical structure, which makes it the ideal model compound for investigating the pyrolytic behavior of biomass materials. This has led to a plethora of available literature deriving various kinetic models to calculate and explain relevant decomposition parameters for cellulose [4,5]. The data is in agreement with these previous works, which further validates the approach to kinetic parameter estimation. However, xylan and pectin are structurally and chemically more complex (Fig. 2(b,c)), which dictates their unique physical and chemical properties that are often quite different from cellulose (Fig. 2(a)). These differences account for the paucity of previous literature pertaining to kinetic parameter estimations for xylan and pectin, and serve as the basis for the electing to use the free model estimate approaches by Friedman [32] and OzawaFlynn-Wall [33,34]. These approaches were used for the preliminary prediction of the number of reaction steps, as well as the reaction sequence for the overall reaction pathways. The kinetic software was then used to estimate the kinetic parameters of pectin and xylan decomposition based upon these free model estimates. The primary decomposition of cellulose is not discussed because of the aforementioned abundance of existing literature [1– 5]. The kinetic analysis produces a set of parameters for cellulose decomposition (with a single first order reaction) similar to those reported in the literature (Ea = 210 kJ mol − 1, and log A=16.1 s − 1). For xylan, as can be seen from the TG and DTG data depicted in Fig. 8(b), a two-step consecutive first order reaction model should fit the experimental data better. The estimated activation energies and pre-exponential factors for these two steps are E1 =168 kJ mol − 1, log A1 =15.4 s − 1 and E2 = 104 kJ mol − 1, log A2 = 7.3 s − 1. These data agree with those previously reported [35,36]. Unlike xylan, the experiTable 2 Relative abundance of different masses from pyrolysis of cellulose and pectin Sample
Pectin Avicel Cellulose
Relative mass abundance per gram of starting material 78
92
106
116
118
128
166
178
329 806 396 587
397 626 391 114
109 806 178 299
70 275 73 485
27 596 122 750
83 737 177 861
9815 23 774
21 281 82 180
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349
347
Table 3 Global kinetic parameters for the formation of different masses from pyrolysis of xylan Mass
Activation energy (kJ mol−1)
Pre-exponential factor (K)
2 (1st) 2 (2nd) 15 78 92 128 178
188 218 206 267 246 238 234
3.18×1011 8.38×1011 5.95×1013 2.93×1017 3.95×1016 7.16×1016 2.09×1016
mental data for pectin (Fig. 8(c)) can be accurately described by a single first order reaction. The activation energy and pre-exponential factor for pectin decomposition are 120 kJ mol − 1 and 9.98 s − 1, respectively. Yet, based on the similarities between pectin and cellulose (both have one primary DTG decomposition peak), these estimated kinetic parameters appear to be reasonable. When the TGA data for the pre-charred samples was fitted using the kinetics software package, very low activation energies and nonsensical pre-exponential factors were estimated. The numbers for activation energies are in the range of 50–60 kJ mol − 1 corresponding to a reaction order greater than 2. Thus far, it has not been possible to identify the most appropriate reaction scheme to model the precharred samples. Therefore, an alternative technique, called the Tmax-model [37], was used to estimate both activation energies and corresponding pre-exponential factors for the product evolution reactions taking place within the char. The Tmax was measured from the DTG at different heating rates (Fig. 5(b) through Fig. 7(b)) for each char sample. The relationship between heating rate (i) and the Tmax can be presented by the following equation: ln(i/T 2max) =ln(AR/Ea) −Ea/R(1/Tmax) where Ea is the activation energy and A is the pre-exponential factor for the reaction. The activation energies for the secondary reactions in the char are 151, 244, and 297 kJ mol − 1 for cellulose, xylan, and pectin, respectively. These values serve as good approximations; yet, there are over simplifications in the predictions, which may affect the validity of the parameter estimates. This discrepancy is primarily due to the complexity of the reactions taking place within the char matrix as the char is heated. Therefore, it would be useful to further examine the char data using a more sophisticated reaction scheme, such as parallel reaction mechanisms. However, determination of kinetic parameters for the evolved products using the Tmax-model provides reasonable initial parameter approximations until further investigation. The activation energies and pre-exponential factors for the monitored masses from xylan char are given in Table 3, and the cumulative ion counts for the corresponding masses are illustrated in Fig. 10. The data for the evolved products from xylan char are consistent with the data presented earlier for natural cotton containing about 5000-ppm potassium ions [24]. This is expected based upon the
348
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349
high (4.5%) ash content of the xylan. It is anticipated that these metal ions catalyze the char reactions, which accounts for the parallel product formation observed between xylan and natural cotton in similar temperature ranges. Conversely, the use of relatively pure compounds, such as Avicel cellulose or pectin, would have similar product evolution over a wider temperature range, which would yield slightly different and less consistent kinetic parameters if a simple model like the Tmaxmodel was used.
5. Conclusions It was demonstrated that the primary decomposition reactions of xylan and pectin involve a softening and melting step, whereas cellulose typically only experiences vaporization. However, the use of an E-SEM and a HR-SEM established that at a relatively high heating rate (\1000 °C s − 1) all the tested compounds showed signs of softening and swelling. A reasonable set of kinetic parameters for the primary decomposition of cellulose, xylan, and pectin were also calculated using the Tmax-model, as well as the Netzsch kinetic software. Furthermore, it was validated that secondary solid phase reactions in char are characteristically unique as compared to the primary reactions. Global kinetic parameters for secondary char reactions and product evolution for xylan above 400 °C chars were also estimated. There continues to be a great number of unanswered questions and unexplained phenomena pertaining to biomass char formation and product evolution that one will continue to explore and help to resolve.
Acknowledgements The authors wish to thank Philip Morris USA for their support and encouragement in conducting this research. Technical assistance from my colleagues Dr. M. Krauss, Ms. V. Baliga, Dr. V. Oja, Dr. R. Sharma and Mr. J. Edgar was greatly appreciated.
References [1] F. Shafizadeh, Appl. Polym. Symp. No. 28, Wiley, New York, 1975, p. 153. [2] F. Shafizadeh, In: Proceedings of the Specialists Workshop on Fast Pyrolysis of Biomass, SERI/CP, 1980, p. 622. [3] F. Shafizadeh, The Chemistry of Solid Wood, American Chemical Society, USA, 1984, p. 489. [4] M.J. Antal Jr, in: K. Boer, J. Duffie (Eds.), Advances in Solar Energy, vol. 1, American Solar Energy Society, Boulder, Colorado, 1982, p. 61. [5] M.J. Antal Jr, in: K. Boer, J. Duddie (Eds.), Advances in Solar Energy, vol. 2, American Solar Energy Society, Plenum, New York, 1985, p. 175. [6] T. Milne, Biomass Gasification: Principles and Technology, Noyes Data Corp, Park Ridge, NJ, 1981.
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349 [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37]
349
G. Varhegyi, J. Emma, M.J. Antal Jr, Energy Fuel 8 (1994) 1345. M.J. Antal Jr, G. Varhegyi, Ind. Eng. Chem. Res. 34 (1995) 703. I. Milosavljevic, V. Oja, E.M. Suuberg, Ind. Eng. Chem. Res. 35 (1996) 653. E.M. Suuberg, I. Milosavljevic, V. Oja, Twenty-Sixth Symposium (International) on Combustion, The Combustion Institute, 1996, p. 1515. M.R. Hajaligol, J.B. Howard, J.P. Longwell, W.A. Peters, Ind. Eng. Chem. Process Des. Dev. 21 (1982) 457. A.G.W. Bradberry, Y. Sakai, F. Shafizadeh, J. Appl. Polym. Sci. 23 (1979) 3271. W. Mok, M.J. Antal Jr, Thermochim. Acta 68 (1983) 155. W. Mok, M.J. Antal Jr, Thermochim. Acta 68 (1983) 165. M.R. Hajaligol, J.B. Howard, W.A. Peters, Combustion Flame 95 (1993) 47. D.F. Arseneau, Can. J. Chem. 49 (1971) 632. S.B. Martin, Presented at Tenth Symposium (International) on Combustion, The Comb. Inst., 1965, p. 877. J.C. Arthur Jr, O. Hinojosa, Textile Res. J. 36 (1966) 385. P.J. Baugh, O. Hinojosa, J.C. Arthur Jr, J. Appl. Polym. Sci. 11 (1967) 1139. R. Narayan, M.J. Antal Jr, Ind. Eng. Chem. Res. 35 (1996) 1711. M.R. Hajaligol, W.A. Peters, J.B. Howard, Energy Fuels 2 (1988) 430. A. Briodo M. Weinstein, In: Weidenman (Ed.), Proceedings of the 3rd International Conference on Thermal Analysis, Birkhauser Verlag, Basel, 1971, p. 285. G.N. Richards, J. Anal. Appl. Pyrol. 10 (1987) 251. M. Hajaligol, B. Waymack, D. Kellogg, In: Fuel Chem. Preprint, Vol. 44, No. 2, Fuel Chem. Div., ACS, Anaheim, California, 1999. B. Waymack, M.R. Hajaligol, T. Fisher, Pyrolysis Kinetics of Cellulosic Materials, The Comb. Inst., Eastern State Section, Raleigh, North Carolina, 1999. A.D. Pouwels, G.B. Eijkel, J.J. Boon, J. Anal. Appl. Pyrol. 14 (1989) 237. A.D. Pouwels, A. Tom, G.B. Eijkel, J.J. Boon, J. Anal. Appl. Pyrol. 11 (1987) 417. A.D. Pouwels, J.J. Boon, J. Anal. Appl. Pyrol. 17 (1990) 97. F. Shafizadeh, Y. Sekiguchi, Combustion Flame 55 (1984) 171. J.B. Boon, I.P. Botto, P.W. Arisz, Biomass and Bioenergy 7 (1 – 6) (1994) 25. R.A. Wind, G.E. Maciel, J.B. Wooten, Characterization of Carbonized Cellulose by Solid State ESR and NMR, Philip Morris USA Internal Report, 1990. H.L. Friedman, J. Polymer Lett. 4 (1966) 323. J.H. Flynn, J. Thermal Anal. 27 (1983) 95. T. Ozawa, J. Thermal Anal. 7 (1975) 601. G. Varhegyi, M.J. Antal Jr, T. Szekely, P. Szabo, Energy Fuel 3 (1989) 329. S.M. Ward, J. Braslaw, Combustion Flame 61 (1985) 261. R.L. Braun, A.K. Burnham, Energy and Fuel 1 (1987) 153.
.
www.elsevier.com/locate/jaap
Pyrolysis behavior and kinetics of biomass derived materials Travis Fisher 1, Mohammad Hajaligol *, Bruce Waymack, Diane Kellogg Philip Morris Inc., Research Center, Richmond, VA 23261, USA Received 5 May 2000; accepted 21 November 2000
Abstract From previous biomass decomposition studies, it is well established that thermolysis generally occurs between 200 and 400 °C. For most materials, this temperature range constitutes up to 95% of total degradation; nonetheless, secondary decomposition reactions continue to occur in the solid matrix above 400 °C. The extent of these reactions, as indicated by the material loss above 400 °C, is small, and in the past has been either ignored or included in the primary degradative step. However, this latter step (reactions above 400 °C) exhibits many unique characteristics that differentiate it from the primary pyrolysis step and therefore needs to be treated separately. Additionally, it is widely accepted that primary decomposition of biomass material ( B 400 °C) consists of a degradative process, whereas the secondary thermolysis ( \ 400 °C) involves an aromatization process. In this study, It is shown that the latter step can be deconvoluted from the primary decomposition step, particularly for materials with limited aromaticity, such as cellulose and other carbohydrates. A thermogravimetric analyzer (TGA) coupled with a differential scanning calorimeter (DSC) and mass spectrometer (MS) was used to pyrolyze materials such as cellulose, xylan and other carbohydrates; and a pyroprobe interfaced with a gas chromatograph (GC) and mass spectrometer for product identification. In addition to monitoring the major products for each step, one also monitored and compared the temperatures corresponding to the maximum rate. From this analysis, the DTG results at 20 °C min − 1 heating rate show that the temperature differences between the peak temperatures of the two decomposition steps are approximately 70, 190, and 200 °C for cellulose, pectin, and xylan, respectively. Furthermore, TGA data were used to calculate sets of biomass-specific kinetic parameters for these two steps. © 2002 Elsevier Science B.V. All rights reserved. Keywords: Pyrolysis; Biomass; Kinetics; Cellulose; Pectin; Xylan * Corresponding author. Tel.: + 1-804-2742419; fax: +1-804-2741994. 1 Present address: Department of Pharmacy, Medical College of Virginia Campus of Virginia Commonwealth University, Richmond, VA 23298, USA. 0165-2370/02/$ - see front matter © 2002 Elsevier Science B.V. All rights reserved. PII: S0165-2370(01)00129-2
332
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349
1. Introduction Past literature reviews have woven a concise and intricate tale from the extensive data generated over decades of research regarding the thermo-chemical degradation and pyrolytic conversion of biomass substrates [1– 10]. The majority of biomass materials are comprised of approximately 50% by weight of cellulose, which has contributed to the extensive kinetic and thermal breakdown data available for cellulose [1– 5]. The thermolytic behavior of biomass materials primarily depends on its chemical composition and structure. The degree of crystallinity and polymerization of the starting materials are integral in defining their respective thermal degradation behavior. In addition to the above parameters, experimentally derived kinetic parameters also depend on the specific pyrolysis conditions that include temperature, heating rate, pressure, particle size, ambient gas environment, and the presence of ash or mineral deposits within the substrates [11– 19]. Additionally, transport processes, such as inter- or intra-particle heat and mass transfer, could also affect the degrading process, and the exclusion of these phenomena would limit the prediction of thermolytic behavior and kinetic parameters [20,21]. The reaction mechanisms and chemistry of biomass decomposition have been thoroughly elucidated in previous literature reviews; consequently, these topics are excluded from the current discussion. Likewise, one only briefly reviews the kinetics of cellulose, leaving the bulk of the kinetic analysis for specific biomass chars. By the mid 1970s, Briodo et al. [22] established the basic kinetic model for cellulose pyrolysis, which as shown in Scheme 1, involves a competitive, multistep reaction mechanism. This proposed model has now expanded to include many other biomass materials, with some minor modifications by Shafizadeh [3,12], Richards [23], and Antal [7,8], as well as many others [9]. These modifications account for the slight differences in chemical structure and composition between biomass materials. It is widely accepted that the primary pyrolysis of biomass materials generally takes place in the range of 200– 400 °C, which results in bulk product volatilization, as well as the formation of a solid char residue. Once the temperature increases above 400 °C, products continue to slowly evolve, as the char residue undergoes further chemical and physical transformations. This occurs more slowly than during primary pyrolysis. This may not be important in the overall scheme for the decomposition of biomass materials; yet, it may be very important in terms of understanding the chemistry behind aromatic formation at temperatures as low as 450 °C. For example, the gasification or combustion of biomass materials yields
Scheme 1.
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349
333
Fig. 1. Schematic of the TGA/DSC/MS system used in this investigation.
solid particles and gas phase products that undergo similar processes, which limits the usefulness of assessing the early char formation. However, smoldering processes are somewhat different, whereby the vapor products are separated from the solid particles at a lower temperature. This separation of solid and vapor products is key to unlocking the formation mechanism(s) for certain distinct hydrocarbons at higher temperatures. In this work, the goal is to investigate and broaden the understanding of the latter steps of biomass pyrolysis, which Broido identified as reaction steps 3 and 4 in Scheme 1 [22]. Kinetic parameters were calculated to describe the thermolytic behavior observed for small samples of cellulosic materials, such as cellulose, hemi-cellulose (xylan) and pectin. The primary emphasis is on the volatilization and solid reactions occurring above 400 °C. The pyrolysis conditions are selected to substantially reduce the effects of heat and mass transfer on the experimental results. 2. Experimental The experimental work was carried out using a Netzsch thermogravimetric analyzer (TGA) coupled with a differential scanning calorimeter (DSC) and a quadrupole mass spectrometer (MS), a platinum coil CDS pyroprobe interfaced with a GC/MS, or an environmental scanning electron microscope (E-SEM). Samples were also sent for elemental analysis. The TGA/DSC/MS facilitated the acquisition of weight loss, heat flow, and mass spectra data as a function of heating rate and temperature (or time). The detailed description of the experimental setup is provided in detail by Hajaligol et al. [24]. However, a brief summary of the setup is as follows: a small sample of starting material (2–15 mg) is weighed and spread evenly in a sample cup; the cup is then placed on the balance sample holder; the startup protocol is initiated; and finally, the sample is heated at a constant heating rate (1–60 °C min − 1) to a desired temperature (700–1000 °C) using helium as the carrier gas at a constant flow rate of 150 ml min − 1. A schematic of the setup is illustrated in Fig. 1. The CDS pyroprobe/GC/MS was used for the further identification and quantification of detected masses.
334
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349
To observe the physical behavior of the starting materials undergoing heating, an E-SEM was used. A small sample ( 1 mg) of each biomass material was separately placed on the hot stage of an E-SEM. The stage was then heated at a rate of approximately 5 °C min − 1 to around 500 °C under an Argon/Hydrogen (70/30) atmosphere and a low pressure of 1.5–4.0 Torr. Frequent scanning with the E-SEM on the hot stage allowed visual characterization of the physical changes occurring at the surface of the heated particles. Small samples of cellulose, hemi-cellulose (xylan) and pectin were used in this study; their chemical structures are depicted in Fig. 2. The cellulose samples were from Avicel (PH-102), with a particle size of roughly 40 mm or less and an ash content of B 0.0008% by weight as determined by elemental analysis (Table 1). The hemi-cellulose samples were from a hardwood birch xylan by Fluka comprised of 95% a-D-xylose units, with methyl glucuronic acid side-chain saccharide units and about 2.5% ash content (per manufacturer). It is a fine crystalline solid, with particle sizes ranging from a few microns up to about 100 mm. The pectin sample was purchased from Acros; it is a polymer consisting of galacturonic acid units, with 67– 71% degree of esterfication and only about a 0.5% ash content (per manufacturer). CPC and NMR analysis of the pectin showed that it contained 35% glucose. In an effort to more accurately detect and characterize the higher temperature reactions and the evolved products of the aforementioned substrates, pre-charred samples of the starting materials were prepared between 350 and 360 °C for 10–15 min under a Helium flow. The organic elemental analysis of the samples used in this study is given in Table 1. In order to illustrate the reaction differences above and below 400 °C, specific masses were monitored with the mass spectrometer coupled to the TGA. Precharred samples of the starting materials were used to accurately monitor mass evolution above 400 °C. Pre-charred samples were used given the fact that the amount of material remaining after primary pyrolysis (typically 5 1 mg) was insufficient to allow accurate mass detection. Therefore, a precharred sample was prepared in a tube furnace, and 10 mg samples of precharred material were used to assess secondary pyrolysis. The evolution of the following masses was monitored as a function of temperature for the heating rates 1, 5, 20, and 60 °C min − 1: 2, 15, 78, 92, 106, 116, 118, 128, 166 and 178, which primarily and respectively represent Table 1 Elemental analysis of materials investigated in this study Wt.%
C
H
O
N
C/H
C/O
Ash
Cellulose Cellulose char Xylan Xylan char Pectin Pectin char
43.99 61.33 40.96 64.26 42.16 64.27
6.20 4.83 6.36 4.40 5.85 3.98
49.79 29.77 47.30 21.73 51.70 27.89
B0.001 B0.001 B0.5 B0.5 0.25 0.57
7.10 12.70 6.44 14.60 7.23 16.15
0.88 2.10 0.87 2.96 0.82 2.30
B0.008 0.008 4.53 13.95 0.50 0.95
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349
Fig. 2. Schematic of chemical structures of the materials used in this investigation.
335
336
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349
hydrogen (2), methane (15), benzene (78), toluene (92), xylene (106), indene (116), benzofuran (118), naphthalene (128), fluorene (166), and anthracene/phenanthrene (178).
3. Results and discussion In this section, one first describes the physical changes in the samples of cellulose, xylan, and pectin as a function of temperature as samples are heated and decomposed. Then, the thermogravimetric data and mass spectra collected for the different materials are analyzed to confirm the multi-step pathway in their thermodegradative process; and finally, the decomposition kinetics are discussed and the kinetic parameters for the samples are estimated. The E-SEM and high resolution SEM were used to observe the qualitative physical changes in the solid particles during sample degradation. Fig. 3(a–c) shows photomicrographs of starting and pyrolyzed (400 °C) samples for cellulose, xylan, and pectin particles, respectively. The cellulose particles retain their mechanical integrity and shape; although, they shrink slightly as they are heated and charred. The cellulose particles are fibrous and form a network of fibers that barely contact the heated crucible; therefore, heating under vacuum may not uniformly heat the particles. Unlike cellulose, particles of pectin and xylan show signs of softening or melting (150 and 200 °C, respectively), as well as bubble formation as they are heated. As shown in Fig. 3, xylan and pectin particles soften at lower temperatures and form a porous structure at higher temperatures. These temperatures correspond to the maximum temperature peaks observed for melting in TGA/DSC. Although the heating rate is slow and the sample weight is small, it is observed that smaller particles soften faster and coalesce with larger particles, resulting in the formation of large globular particles. Examination of the char sample morphology using SEM shows the rupture of the bubbles that are formed during pyrolysis. Bubbles are formed within the particles, and as the bubbles burst, volatile products are released into the gas stream. This observation provides an interesting insight into the role of transport processes in the study of solid organic materials. For non-melting (non-softening) substances, such as cellulose, particles retain their original shape; thus, estimation of heat and mass transfers becomes more readily predictable. However, melting substrates, such as xylan and pectin, coalesce and form round or spherically shaped liquid particles. The estimation of intra-molecular heat and mass transfers requires more accurate knowledge of process dynamics. The observation shows that the melting and bubbling processes become more vigorous as the heating rate increases. Fig. 4(a– c) shows the respective melting and bubbling phenomena for cellulose, xylan, and pectin when particles are heated at a rate of several thousand degrees per second. Under these conditions, even cellulose particles exhibit melting and bubbling (Fig. 4(a)). The derivatives of the thermogravimetric weight loss data (DTG) are shown as a function of temperature at specific heating rates (5, 20, and 60 °C min − 1) in Fig. 5(a,b) through Fig. 7(a,b) for cellulose, xylan, pectin and their chars, respectively.
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349
337
Fig. 3. SEM photomicrographs of: (a) starting and pyrolyzed (400 °C) cellulose samples; (b) starting and pyrolyzed (400 °C) xylan samples; (c) starting and pyrolyzed (400 °C) pectin samples.
338
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349
Fig. 4. SEM photomicrographs of pyrolyzed materials at high heating rate ( \ 1000 °C s − 1): (a) cellulose, (b) xylan, and (c) pectin.
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349
339
Fig. 5. Pyrolysis rates (DTG) of cellulose samples: (a) starting material and (b) char prepared around 350 °C for 10 min.
Depending on the heating rate and starting material, the major decomposition of the starting substrate is completed below 400 °C (Figs. 5(a), 6(a) and 7(a)). For xylan and pectin, the thermo-degradative temperature is below 300 °C. Above 400 °C, there was no appreciable weight loss observed, which would have suggested the existence of another major reaction. This remained true for even the highest heating rate (60 °C min − 1) used in this study. This is better illustrated by the curves in Fig. 8(a– c); the weight loss seen with a 20 °C min − 1 heating rate is extremely small above 400 °C. This observation created the perception that no
340
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349
major reactions existed above 400 °C; and thus, suggested that studies above 400 °C would be unnecessary. However, analysis of chars prepared at temperatures (350–360 °C for 10– 15 min) exceeding their primary decomposition temperature has shown the contrary. Further heating of these chars above 400 °C reveals a secondary thermolytic process with a significant weight loss, as illustrated for cellulose, xylan, and pectin in Figs. 5(b), 6(b) and 7(b), respectively. For additional comparison, Fig. 8(a– c) show the representative and respective weight loss and derivative of weight loss curves for samples of cellulose, xylan, pectin, and their chars at a heating rate of 20 °C min − 1.
Fig. 6. Pyrolysis rates (DTG) of xylan samples: (a) starting material and (b) char prepared around 350 °C for 10 min.
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349
341
Fig. 7. Pyrolysis rates (DTG) of pectin samples: (a) starting material and (b) char prepared around 350 °C for 10 min.
Weight loss data for Avicel cellulose pyrolysis are presented in Fig. 8(a). At a heating rate of 20 °C min − 1, cellulose rapidly decomposes between 300 and 350 °C; this corresponds to a weight loss of around 91%. Following this sharp weight loss, there is a gradual, but small, weight loss (additional 5% of starting cellulose) up to about 700 °C. This gradual weight loss can be more readily observed and detected as another step in decomposition if the pre-charred sample of Avicel cellulose is heated separately. As can be seen from Fig. 8(a), most of the weight loss for the pre-charred cellulose takes place between 350 and 700 °C. The derivatives (DTG) of the two weight loss curves are also given in Fig. 8(a). The DTG data provide two distinct Tmax, one around 335 °C for primary reactions and
342
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349
one around 420 °C for secondary reactions. The latter Tmax depends on the heating rate and the conditions under which pre-charred samples are prepared, and could be as high as 500 °C. The breakdown of pre-charred cellulose has been cited in the literature as a small exotherm occurring after the larger endothermic decomposition [9,22]. The data suggest that this second step reaction is most likely thermo-neutral [25], depending on the condition under which pre-charred samples are prepared. This observation
Fig. 8. Thermogravimetric data (TG and DTG) of different samples at 20 °C min − 1 heating rate: (a) cellulose and pre-charred cellulose, (b) xylan and pre-charred xylan, and (c) pectin and pre-charred pectin.
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349
343
Fig. 8. (Continued)
is depicted in Fig. 9. All samples are initially pre-charred under various heating (temperature and time) conditions. The figure represents the pyrolysis of precharred samples that have undergone their primary reactions to a variable extent. The weight loss and corresponding heat of reaction of these samples measured by TGA/DSC are shown in Fig. 9. If a pre-charred sample is observed to give a weight loss of up to 70% in the TGA, the net heat of reaction is almost zero (thermo-neutral). Conversely, if the pre-charred sample gives greater than 70% weight loss in the
Fig. 9. Heat of pyrolysis of cellulose samples as a function of decomposition percent.
344
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349
TGA, then it is still undergoing net endothermic reactions. The DSC data for cellulose pyrolysis shows an endothermic heat of reaction of about 450 J g − 1. There is no indication of any exotherm during pyrolysis of cellulose or pre-charred cellulose samples. Regardless of the cellulose sample size (2–10 mg), no change is observed in the DSC peak temperature or the amount of reaction endothermicity. Weight loss data for xylan and its char at 20 °C min − 1 are given in Fig. 8(b). The xylan sample shows two distinct reactions below 300 °C that can be seen from the uncharred TG and DTG curves; this has an attributed weight loss of approximately 70%. Similar to cellulose and pectin, the weight loss and corresponding rate of weight loss above 400 °C appear insignificant. However, when pre-charred samples of xylan are heated in Helium at a 20 °C min − 1 heating rate, a reaction (or a series of reactions) can be observed to take place between 400 and 600 °C, as well as another one above 600 °C. During this secondary pyrolysis, an additional 10% weight loss was observed. The high temperature reaction(s) is believed to arise from the inorganic content [19,24] of the xylan (4.5% ash content), as seen in Table 1. The endotherm associated with the decomposition of the current sample of xylan is spread over a broad temperature range, and its integration over this temperature range would introduce a major error in the estimated values. Therefore, it can only be concluded that one endotherm starts at around 150 °C and ends near 300 °C, which coincides with the completion of decomposition reactions. The thermogravimetric data of pectin and pectin char are presented in Fig. 8(c). At a heating rate of 20 °C min − 1, pectin (shown in Fig. 8(c)) rapidly decomposes (approximately 70%) around 250 °C. The difference between the pectin and cellulose decomposition temperatures is believed to be due to greater molecular structure stability, and that pectin has side groups that undergo elimination reactions at lower temperatures. Analogous with the observations for cellulose, the weight loss for pectin appears insignificant above 400 °C. When pre-charred samples of pectin are heated, an additional weight loss of around 10% is observed above 400 °C. Comparison of the DTG data provides two distinct Tmax, one for the starting pectin at around 235 °C, and one for pre-charred sample at around 420 °C. In contrast to the cellulose samples, pectin decomposition is not only less endothermic, but the endotherm is also followed by a small exotherm. Pectin first melts at around 150 °C with a heat of melting of about 75 J g − 1, and then it undergoes primary decomposition above 200 °C, corresponding to an endothermic reaction of about 90 J g − 1. In order to illustrate the reaction differences above and below 400 °C, specific masses were monitored with the mass spectrometer coupled to the TGA. The masses represent what evolves during secondary pyrolysis. The activation energies for each mass correspond nicely with that for the entire secondary thermolytic process. Below 400 °C, the major evolved products are from the primary decomposition reaction and are widely validated in the literature [26– 28]. Pre-charred samples of the starting materials were used to accurately monitor mass evolution above 400 °C. The presence of the aforementioned hydrocarbons from the char was confirmed by the Py-GC/MS, which showed that these compounds generally account for the majority of ascribed masses. However, there are other compounds
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349
345
Fig. 10. Cumulative amounts of different masses as a function of temperature for pre-charred xylan samples.
that contribute to different extents to the specific masses monitored during the TG/MS experiment. For instance, several oxygenated benzoid compounds contribute to mass 78; likewise, there are also significant oxygenated compounds (pyrones) contributing to mass 128. Additionally, pyrolysis of pre-charred samples eliminates the risk of mistaking secondary reactions of the primary products as being part of the secondary decomposition, as well as focuses on the specific temperature ranges where hydrocarbons evolve from these materials. One has previously used masses 78, 92, 106, 116, 118, 128, 166, and 178 to follow evolution profiles of aromatics from the samples of cellulosic materials [24]. Experimental results show that all the hydrocarbons from char samples evolve over a similar temperature range (illustrated for xylan in Fig. 10) with a few exceptions; the evolution of methane and benzene (second peak) occurs above 500 °C and hydrogen evolves at temperatures above 600 °C. These results indicate that substituted benzene, as well as larger aromatic and aliphatic hydrocarbons, evolve first, and as they are depleted, there is a seemingly continuous evolution of methane and benzene. As sources of CH4 and benzene are depleted above 600 °C, hydrogen production predominates. At this point, a highly carbonaceous char has been produced. This is consistent with previous work [29–31], where the aliphatic and aromatic concentrations of the cellulosic char were studied as a function of temperature. For example, Shafizadeh [29] did not measure the evolved gases in his study; yet, his data on solid char concentrations is in good agreement with the data on the evolved gases over similar temperature ranges. The relative amounts of some aromatics from cellulose chars were measured using the Py-GC/MS, for which, the results are tabulated in Table 2. In this table, the abundance of the masses is for the individual compounds represented by their respective mass number.
346
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349
4. Reaction kinetics The thermogravimetric data (TG and DTG) was used to characterize the decomposition reactions of the starting biomass materials, as well as provide estimates of their kinetic parameters. Furthermore, the TG/DTG data from precharred samples was used to characterize product formation reactions and evolution. Initially, a Netzsch kinetic software package was used to estimate the kinetic parameters for the reactions. This kinetics software operates based on a leastsquares regression model and provides an optimal value for the parameters based on the minimum mean of residuals. Cellulose has a relatively simple chemical structure, which makes it the ideal model compound for investigating the pyrolytic behavior of biomass materials. This has led to a plethora of available literature deriving various kinetic models to calculate and explain relevant decomposition parameters for cellulose [4,5]. The data is in agreement with these previous works, which further validates the approach to kinetic parameter estimation. However, xylan and pectin are structurally and chemically more complex (Fig. 2(b,c)), which dictates their unique physical and chemical properties that are often quite different from cellulose (Fig. 2(a)). These differences account for the paucity of previous literature pertaining to kinetic parameter estimations for xylan and pectin, and serve as the basis for the electing to use the free model estimate approaches by Friedman [32] and OzawaFlynn-Wall [33,34]. These approaches were used for the preliminary prediction of the number of reaction steps, as well as the reaction sequence for the overall reaction pathways. The kinetic software was then used to estimate the kinetic parameters of pectin and xylan decomposition based upon these free model estimates. The primary decomposition of cellulose is not discussed because of the aforementioned abundance of existing literature [1– 5]. The kinetic analysis produces a set of parameters for cellulose decomposition (with a single first order reaction) similar to those reported in the literature (Ea = 210 kJ mol − 1, and log A=16.1 s − 1). For xylan, as can be seen from the TG and DTG data depicted in Fig. 8(b), a two-step consecutive first order reaction model should fit the experimental data better. The estimated activation energies and pre-exponential factors for these two steps are E1 =168 kJ mol − 1, log A1 =15.4 s − 1 and E2 = 104 kJ mol − 1, log A2 = 7.3 s − 1. These data agree with those previously reported [35,36]. Unlike xylan, the experiTable 2 Relative abundance of different masses from pyrolysis of cellulose and pectin Sample
Pectin Avicel Cellulose
Relative mass abundance per gram of starting material 78
92
106
116
118
128
166
178
329 806 396 587
397 626 391 114
109 806 178 299
70 275 73 485
27 596 122 750
83 737 177 861
9815 23 774
21 281 82 180
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349
347
Table 3 Global kinetic parameters for the formation of different masses from pyrolysis of xylan Mass
Activation energy (kJ mol−1)
Pre-exponential factor (K)
2 (1st) 2 (2nd) 15 78 92 128 178
188 218 206 267 246 238 234
3.18×1011 8.38×1011 5.95×1013 2.93×1017 3.95×1016 7.16×1016 2.09×1016
mental data for pectin (Fig. 8(c)) can be accurately described by a single first order reaction. The activation energy and pre-exponential factor for pectin decomposition are 120 kJ mol − 1 and 9.98 s − 1, respectively. Yet, based on the similarities between pectin and cellulose (both have one primary DTG decomposition peak), these estimated kinetic parameters appear to be reasonable. When the TGA data for the pre-charred samples was fitted using the kinetics software package, very low activation energies and nonsensical pre-exponential factors were estimated. The numbers for activation energies are in the range of 50–60 kJ mol − 1 corresponding to a reaction order greater than 2. Thus far, it has not been possible to identify the most appropriate reaction scheme to model the precharred samples. Therefore, an alternative technique, called the Tmax-model [37], was used to estimate both activation energies and corresponding pre-exponential factors for the product evolution reactions taking place within the char. The Tmax was measured from the DTG at different heating rates (Fig. 5(b) through Fig. 7(b)) for each char sample. The relationship between heating rate (i) and the Tmax can be presented by the following equation: ln(i/T 2max) =ln(AR/Ea) −Ea/R(1/Tmax) where Ea is the activation energy and A is the pre-exponential factor for the reaction. The activation energies for the secondary reactions in the char are 151, 244, and 297 kJ mol − 1 for cellulose, xylan, and pectin, respectively. These values serve as good approximations; yet, there are over simplifications in the predictions, which may affect the validity of the parameter estimates. This discrepancy is primarily due to the complexity of the reactions taking place within the char matrix as the char is heated. Therefore, it would be useful to further examine the char data using a more sophisticated reaction scheme, such as parallel reaction mechanisms. However, determination of kinetic parameters for the evolved products using the Tmax-model provides reasonable initial parameter approximations until further investigation. The activation energies and pre-exponential factors for the monitored masses from xylan char are given in Table 3, and the cumulative ion counts for the corresponding masses are illustrated in Fig. 10. The data for the evolved products from xylan char are consistent with the data presented earlier for natural cotton containing about 5000-ppm potassium ions [24]. This is expected based upon the
348
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349
high (4.5%) ash content of the xylan. It is anticipated that these metal ions catalyze the char reactions, which accounts for the parallel product formation observed between xylan and natural cotton in similar temperature ranges. Conversely, the use of relatively pure compounds, such as Avicel cellulose or pectin, would have similar product evolution over a wider temperature range, which would yield slightly different and less consistent kinetic parameters if a simple model like the Tmaxmodel was used.
5. Conclusions It was demonstrated that the primary decomposition reactions of xylan and pectin involve a softening and melting step, whereas cellulose typically only experiences vaporization. However, the use of an E-SEM and a HR-SEM established that at a relatively high heating rate (\1000 °C s − 1) all the tested compounds showed signs of softening and swelling. A reasonable set of kinetic parameters for the primary decomposition of cellulose, xylan, and pectin were also calculated using the Tmax-model, as well as the Netzsch kinetic software. Furthermore, it was validated that secondary solid phase reactions in char are characteristically unique as compared to the primary reactions. Global kinetic parameters for secondary char reactions and product evolution for xylan above 400 °C chars were also estimated. There continues to be a great number of unanswered questions and unexplained phenomena pertaining to biomass char formation and product evolution that one will continue to explore and help to resolve.
Acknowledgements The authors wish to thank Philip Morris USA for their support and encouragement in conducting this research. Technical assistance from my colleagues Dr. M. Krauss, Ms. V. Baliga, Dr. V. Oja, Dr. R. Sharma and Mr. J. Edgar was greatly appreciated.
References [1] F. Shafizadeh, Appl. Polym. Symp. No. 28, Wiley, New York, 1975, p. 153. [2] F. Shafizadeh, In: Proceedings of the Specialists Workshop on Fast Pyrolysis of Biomass, SERI/CP, 1980, p. 622. [3] F. Shafizadeh, The Chemistry of Solid Wood, American Chemical Society, USA, 1984, p. 489. [4] M.J. Antal Jr, in: K. Boer, J. Duffie (Eds.), Advances in Solar Energy, vol. 1, American Solar Energy Society, Boulder, Colorado, 1982, p. 61. [5] M.J. Antal Jr, in: K. Boer, J. Duddie (Eds.), Advances in Solar Energy, vol. 2, American Solar Energy Society, Plenum, New York, 1985, p. 175. [6] T. Milne, Biomass Gasification: Principles and Technology, Noyes Data Corp, Park Ridge, NJ, 1981.
T. Fisher et al. / J. Anal. Appl. Pyrolysis 62 (2002) 331–349 [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37]
349
G. Varhegyi, J. Emma, M.J. Antal Jr, Energy Fuel 8 (1994) 1345. M.J. Antal Jr, G. Varhegyi, Ind. Eng. Chem. Res. 34 (1995) 703. I. Milosavljevic, V. Oja, E.M. Suuberg, Ind. Eng. Chem. Res. 35 (1996) 653. E.M. Suuberg, I. Milosavljevic, V. Oja, Twenty-Sixth Symposium (International) on Combustion, The Combustion Institute, 1996, p. 1515. M.R. Hajaligol, J.B. Howard, J.P. Longwell, W.A. Peters, Ind. Eng. Chem. Process Des. Dev. 21 (1982) 457. A.G.W. Bradberry, Y. Sakai, F. Shafizadeh, J. Appl. Polym. Sci. 23 (1979) 3271. W. Mok, M.J. Antal Jr, Thermochim. Acta 68 (1983) 155. W. Mok, M.J. Antal Jr, Thermochim. Acta 68 (1983) 165. M.R. Hajaligol, J.B. Howard, W.A. Peters, Combustion Flame 95 (1993) 47. D.F. Arseneau, Can. J. Chem. 49 (1971) 632. S.B. Martin, Presented at Tenth Symposium (International) on Combustion, The Comb. Inst., 1965, p. 877. J.C. Arthur Jr, O. Hinojosa, Textile Res. J. 36 (1966) 385. P.J. Baugh, O. Hinojosa, J.C. Arthur Jr, J. Appl. Polym. Sci. 11 (1967) 1139. R. Narayan, M.J. Antal Jr, Ind. Eng. Chem. Res. 35 (1996) 1711. M.R. Hajaligol, W.A. Peters, J.B. Howard, Energy Fuels 2 (1988) 430. A. Briodo M. Weinstein, In: Weidenman (Ed.), Proceedings of the 3rd International Conference on Thermal Analysis, Birkhauser Verlag, Basel, 1971, p. 285. G.N. Richards, J. Anal. Appl. Pyrol. 10 (1987) 251. M. Hajaligol, B. Waymack, D. Kellogg, In: Fuel Chem. Preprint, Vol. 44, No. 2, Fuel Chem. Div., ACS, Anaheim, California, 1999. B. Waymack, M.R. Hajaligol, T. Fisher, Pyrolysis Kinetics of Cellulosic Materials, The Comb. Inst., Eastern State Section, Raleigh, North Carolina, 1999. A.D. Pouwels, G.B. Eijkel, J.J. Boon, J. Anal. Appl. Pyrol. 14 (1989) 237. A.D. Pouwels, A. Tom, G.B. Eijkel, J.J. Boon, J. Anal. Appl. Pyrol. 11 (1987) 417. A.D. Pouwels, J.J. Boon, J. Anal. Appl. Pyrol. 17 (1990) 97. F. Shafizadeh, Y. Sekiguchi, Combustion Flame 55 (1984) 171. J.B. Boon, I.P. Botto, P.W. Arisz, Biomass and Bioenergy 7 (1 – 6) (1994) 25. R.A. Wind, G.E. Maciel, J.B. Wooten, Characterization of Carbonized Cellulose by Solid State ESR and NMR, Philip Morris USA Internal Report, 1990. H.L. Friedman, J. Polymer Lett. 4 (1966) 323. J.H. Flynn, J. Thermal Anal. 27 (1983) 95. T. Ozawa, J. Thermal Anal. 7 (1975) 601. G. Varhegyi, M.J. Antal Jr, T. Szekely, P. Szabo, Energy Fuel 3 (1989) 329. S.M. Ward, J. Braslaw, Combustion Flame 61 (1985) 261. R.L. Braun, A.K. Burnham, Energy and Fuel 1 (1987) 153.
.