Effect of pyrolysis temperature on composition, surface properties and thermal degradation rates of Brazil Nut shells

Effect of pyrolysis temperature on composition, surface properties and thermal degradation rates of Brazil Nut shells

Bioresource Technology 76 (2001) 15±22 E€ect of pyrolysis temperature on composition, surface properties and thermal degradation rates of Brazil Nut ...

559KB Sizes 1 Downloads 49 Views

Bioresource Technology 76 (2001) 15±22

E€ect of pyrolysis temperature on composition, surface properties and thermal degradation rates of Brazil Nut shells P.R. Bonelli, P.A. Della Rocca, E.G. Cerrella, A.L. Cukierman * PINMATE ± Departamento de Industrias, Programa de Investigaci on y Desarrollo de Fuentes Alternativas de Materias Primas y Energõa ± Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Ciudad Universitaria (1428) Buenos Aires, Argentina Received 8 September 1999; received in revised form 5 June 2000; accepted 6 June 2000

Abstract Changes in chemical and surface characteristics of Brazil Nut shells (Bertholletia excelsa) due to pyrolysis at di€erent temperatures (350°C, 600°C, 850°C) were examined. For this purpose, proximate and ultimate analyses, physical adsorption measurements of N2 ()196°C) and CO2 (25°C) as well as samples visualisation by scanning electronic microscopy (SEM) were performed. Appreciable di€erences in the residue characteristics, depending markedly on the pyrolysis temperature, were observed. Release of volatile matter led to the development of pores of di€erent sizes. Progressive increases in micropore development with increasing pyrolysis temperature took place, whereas a maximum development of larger pores occurred at 600°C. Furthermore, kinetics measurements of Brazil Nut shells pyrolysis from ambient temperature up to 900°C were performed by non-isothermal thermogravimetric analysis. A model taking into account the signi®cant changes in the residue during pyrolysis, through an increase in the activation energy with temperature and solid conversion, were found to properly ®t the kinetics data over the wide range of degradation investigated. Ó 2000 Elsevier Science Ltd. All rights reserved. Keywords: Biomass pyrolysis; Brazil Nut shells; Char characterisation; Pyrolysis kinetics

1. Introduction Pyrolysis of biomass has received special attention since it leads to useful products and simultaneously contributes to diminish environmental pollution arising from wastes accumulation and/or open ®eld burning. In particular, slow pyrolysis constitutes a suitable alternative when high yields of the carbon-enriched solid product (char) are required. Char can be directly used as a fuel or submitted to further processing to produce more value-added chemicals, such as activated carbons (Bridgwater and Bridge, 1991; Rashid Khan and Gorsuch, 1996). Pyrolysis of lignocellulosic materials is complex since their major constituents, namely cellulose, hemicellulose and lignin, show di€erent reactivities. Depending on temperature and overall conversion level, di€erent reactions associated to thermal decomposition of each constituent occur that in turn bring about changes in *

Corresponding author. Tel.: +54-11-4576-3383; fax: 54-11-45763366. E-mail address: [email protected] (A.L. Cukierman).

material properties. Interactions between constituents and minute amounts of mineral matter naturally present in whole biomass samples, that catalyze numerous reactions taking place during pyrolysis (Antal and Varhegyi, 1995; Caballero et al., 1997a), introduce additional factors of complexity, making it dicult to achieve a generalized knowledge of pyrolysis of any lignocellulosic material. Brazil Nuts (Bertholletia excelsa) belongs to the Lecythidaceae family and grow in a vast zone of South America, mostly in the region of Para (Brazil). Castanheiro do Para, the Brazilian name given to its tree, is found in many Amazonian States of Brazil, Peru, Colombia, Venezuela and Ecuador. The tree is enormous, frequently attaining a height of 50 m or more. The fruit is a large spherical woody capsule or pod. Inside each fruit pod there are 12±25 nuts with their own individual shell. The tree can produce approximately 300 or more of these fruit pods. The seed kernel contains 65±70% of oil, which can be used in soap manufacture and 17% protein. Thousands of tons of Brazil Nuts are exported each year. Today, the monetary value of Brazil Nut exportation from Amazonian Brazil is second only to

0960-8524/01/$ - see front matter Ó 2000 Elsevier Science Ltd. All rights reserved. PII: S 0 9 6 0 - 8 5 2 4 ( 0 0 ) 0 0 0 8 5 - 7

16

P.R. Bonelli et al. / Bioresource Technology 76 (2001) 15±22

rubber. In recent years, considerable shelling of nuts has been carried out prior to export generating large amounts of residues. Ecient conversion of the shells into useful products requires knowledge of relevant aspects concerning pyrolysis. Particularly, kinetics is of utmost importance for the proper design of large-scale pyrolysis reactors since it strongly depends on the residue reactivity (Cukierman et al., 1996; Della Rocca et al., 1999). Another important aspect related to the potential use of the shells, as precursors for activated carbons production by the wellknown ``physical'' activation two-stage method, concerns the e€ect of conditions employed in pyrolysis, namely the ®rst-stage, on composition and surface characteristics of the resulting char. Char properties are signi®cant for the quality of the ®nal product (Heschel and Klose, 1995; Horowitz et al., 1997; Tancredi et al., 1997). Studies on these aspects for pyrolysis of some shells and fruit stones have been published (Font et al., 1991; Balci et al., 1993; Heschel and Klose, 1995; Della Rocca et al., 1997; Caballero et al., 1997a,b; Encinar et al., 1997; Demirbas, 1998; Di Blassi et al., 1999). However, it remains dicult to predict with con®dence pyrolysis rates and char properties for a speci®c residue from data reported for others or based on its composition. In this context, the present work focuses on two different aspects on Brazil Nut shells pyrolysis: (1) evaluation of chemical and textural±morphological properties of nutshells during pyrolysis and (2) the thermal degradation of the shells over a wide range of temperatures.

2. Methods 2.1. Materials and samples preparation Shells of Brazil Nut (B. excelsa) were used as raw material. The shells, without previous treatment, were milled and screen-sieved. Fractions with particle diameters in the range of 1:2±1:4  10ÿ3 m were selected for the study. To examine the evolution of shell characteristics during pyrolysis, char samples were prepared at di€erent ®nal pyrolysis temperatures under nitrogen (¯ow rate 1:33  10ÿ3 L sÿ1 ). A ®xed bed reactor (0.02 m of inner diameter and 0.15 m of height) heated by an electric furnace was used for shell pyrolysis. Virgin shells were subjected to heat at a heating rate of 0.25°C sÿ1 up to ®nal temperatures of 350°C, 600°C and 850°C, respectively, and held at these temperatures for one hour. Afterwards, chars were cooled under nitrogen to room temperature and stored in sealed containers for sample characterization.

2.2. Chemical, textural and morphological characterization of the samples Proximate analyses of the shells and char samples obtained at the three di€erent temperatures were performed following the standard methods of the American Society for Testing and Materials (ASTM). A Carlo Erba EA 1108 Elemental Analyser (Carlo Erba Strumentazione, Milan, Italy) was employed to determine elemental composition of all the samples. The instrument was calibrated with the analysis of a suitable organic standard of a known elemental composition, provided by the manufacturer. The estimated errors in the determinations were less than 5% and they were obtained by repeating the analyses for all the samples at least three times. The main constituents of Brazil Nut shells, lignin, holocellulose (i.e. cellulose + hemicellulose) and solvent extractive components, were also determined. Extractives and lignin were respectively isolated according to the standard methods TAPPI T 204 om-88 and TAPPI T 222 om-88, as prescribed by the Technical Association of the Pulp and Paper Industry (TAPPI). Holocellulose was obtained by hydrolysis using HClO2 acid. Details on the procedure are described in Browning (1970). The reported results are mean values with a standard deviation of 5%. Surface characteristics of the nut shells and char samples were determined from physical adsorption measurements of N2 at ()196°C) and CO2 at 25°C by volumetric techniques. A Micromeritics Accusorb 2100 E sorption instrument (Micromeritics, Norcross, GA, USA) was employed for CO2 adsorption experiments, whereas adsorption measurements performed with N2 were carried out in a Micromeritics Gemini 2360 Surface Area Analyser (Micromeritics, Norcross, GA, USA). The samples were outgassed overnight at 120°C at a ®nal pressure of 1:33  10ÿ4 Pa (10ÿ6 mm Hg) prior to the measurements. Adsorption models were used to ®t the isotherms and evaluate surface properties. For the statistical tests, a 5% level of signi®cance was used. Changes in structure of virgin shells and char samples, due to the thermal conversion process, were examined by scanning electronic microscopy (SEM) using a 515 Philips microscope coupled with an 9100 energy dispersive X-ray spectrometer (Royal Philips Electronics, Amsterdam, The Netherlands). Photographs were taken on external surfaces and transverse cuts of the virgin shells and char samples. All the procedures used have been thoroughly detailed in Della Rocca (1998). 2.3. Kinetic measurements Kinetic measurements of Brazil Nut shells pyrolysis were performed by non-isothermal thermogravimetric analysis. A thermogravimetric balance Netzsch STA 409

P.R. Bonelli et al. / Bioresource Technology 76 (2001) 15±22

17

(Netzsch-Geratebau, Selb/Bayern, Germany) coupled with a nitrogen ¯ow device and a data acquisition system was used in the kinetics measurements. A preliminary set of experiments, reported in a previous work (Della Rocca, 1998), were carried out in order to select sample mass, particle diameter and gas ¯ow rate for which di€usional e€ects are negligible. Nitrogen ¯ow rate of 2:5  10ÿ3 L sÿ1 , sample masses of 10  0:2 mg and 3:7±4:4  10ÿ5 m particle diameter were employed for all the experimental runs. Experiments at di€erent heating rates, in the range 0.17±1.7°C sÿ1 , from room temperature up to 900°C were also carried out. No appreciable di€erences in mass loss vs temperature curves were found, pointing to negligible heat transfer e€ects (Della Rocca, 1998). Thermogravimetric experiments were duplicated. Di€erences between instantaneous mass fractions of the replicates were always less than 2%. Two di€erent models were used to correlate the experimental data. Model parameters were estimated by non-linear regression analysis. The sum of squares between the experimental and calculated instantaneous mass fractions was minimized. Standard deviation (S.D.) and variation coecient (VC) were calculated.

tures are shown in Table 1. Pyrolysis of the shells promotes signi®cant changes in their chemical features, which are re¯ected in the results obtained for the char samples and are strongly a€ected by the heat treatment temperature. As expected, thermal treatment of the shells brings about a decrease in the amount of volatile matter accompanied by an increase in ®xed carbon and ash content, with higher pyrolysis temperature. Elemental carbon and nitrogen contents increase with increasing pyrolysis temperature, while contents of hydrogen and oxygen decrease due to loss of volatile matter. These results are in accordance with others reported in the literature for di€erent lignocellulosic materials (Cukierman et al., 1996). The composition of the shells together with that of some of the food processing wastes as reported in the literature are listed in Table 2, for comparison purposes. As can be observed, Brazil Nut shells exhibited the highest percentage of lignin among all the residues examined.

3. Results and discussion

Textural transformations of the samples at the three pyrolysis temperatures were examined using adsorption isotherms with N2 ()196°C) and CO2 (25°C) as adsorbates. The complementary use of these two adsorbates is necessary to characterise carbonaceous materials with a complex network of pores of di€erent sizes (L opez

3.1. Variations in chemical composition of nut shells Proximate and ultimate analyses of the residue and char samples obtained at the three di€erent tempera-

3.2. Textural±morphological characteristics of the samples

Table 1 Chemical composition of Brazil Nut shells and char samples obtained at the di€erent pyrolysis temperatures (proximate and ultimate analysis)a Sample

Volatile matterb (%)

Fixed carbonb (%)

Ashb (%)

Cc (%)

Hc (%)

Nc (%)

Od (%)

Brazil Nut shells Char (T ˆ 350°C) Char (T ˆ 600°C) Char (T ˆ 850°C)

76.1 48.2 18.2 5.3

22.2 50.1 78.9 91.3

1.7 1.7 2.9 3.4

50.0 67.3 89.5 94.7

5.8 4.5 2.6 0.8

0.7 1.2 1.4 1.6

43.5 27.0 6.5 2.9

a

The reported values are averages of triplicate determinations, where the standard error did not exceed 5%. Dry basis. c Dry and ash free basis. d Estimated as the remaining % after subtraction of C, H and N from 100%. b

Table 2 Composition of Brazil Nut shells and other food processing wastes

a b

Raw material

Holocellulosea (%)

Lignina (%)

Extractivesb (%)

Reference

Brazil Nut shells Hazelnut shells Hazelnut shells Almond shells Almond shells Olive stones

48.5 70.6 66.5 69.1 68.9 63.5

59.4 39.5 33.5 27.7 31.1 30.0

3.4 2.2 8.3 3.2 11.7 5.5

This work Della Rocca (1998) Balci et al. (1993) Caballero et al. (1997a) Balci et al. (1993) Caballero et al. (1997a)

Dry and extractive free basis. Fraction of components that are soluble in ethanol±benzene, expressed on a dry basis.

18

P.R. Bonelli et al. / Bioresource Technology 76 (2001) 15±22

plied to ®t nitrogen adsorption isotherms and evaluate the speci®c surface area (SN2 ) of the samples following the conventional procedure detailed by Gregg and Sing (1982). Straight lines over the range of relative pressures 0.05±0.30 …r2 P 0:99† for which the BET model is valid were found in all the cases. From the BET plots, the characteristic parameters were estimated by regression analysis. The speci®c surface areas were evaluated from: S N2 ˆ am

Fig. 1. N2 ()196°C) adsorption isotherms of Brazil Nut shells and char samples obtained at the di€erent pyrolysis temperatures, p being the equilibrium pressure and p0 the adsorbate saturation pressure at the working temperature.

Peinado et al., 1985; Magnaterra et al., 1994; Della Rocca et al., 1999). The adsorption isotherms for N2 obtained for the residue and char samples providing information about sample larger pores, mainly macropores, mesopores and larger micropores, are illustrated in Fig. 1. In this ®gure, are shown nitrogen adsorbed volumes expressed in standard conditions of temperature and pressure (STP) per sample mass unit, Va , as a function of the relative pressure (p/p0 ). The raw shell residue and the char at 350°C exhibit similar isotherm shapes between that of type I and II isotherms, according to the International Union of Pure and Applied Chemistry (IUPAC) classi®cation, indicating a predominantly microporous surface (Gregg and Sing, 1982; Byrne and Marsh, 1995). This suggests that only slight changes in the shell structure occur under pyrolysis at relatively low temperatures, up to 350°C, at least in the range of porosity evaluated by this technique. In contrast, pronounced changes in the shape of the isotherms resembling features of type II as well as enhanced N2 volumes were observed in the samples pyrolyzed at higher temperatures (>350°C). These changes may be attributed to the development of larger pores. The well-known multilayer adsorption model developed by Brunauer, Emmett and Teller (BET) was ap-

Vm NA ; vM

…1†

where am is the area occupied by a nitrogen molecule, am (N2 ) ˆ 16.2 10ÿ20 m2 , Vm the monolayer volume, NA the Avogadro number and vM is the molar volume (STP) of the adsorbate. Estimated values of SN2 together with standard deviations are listed in Table 3. Sample micropore (<2 nm) volumes were evaluated from the CO2 adsorption isotherms at 25°C using the Dubinin±Radushkevich (DR) equation. The application of this equation to the low-pressure part of the isotherms is particularly suitable to investigate solid microporous textures or those in which micropores contribution is signi®cant (L opez Peinado et al., 1985). The DR equation, based on the Polanyi thermodynamical theory of adsorption, considers that adsorption in very ®ne pores involves a volume ®lling process rather than layer-by-layer adsorption on the pore walls. It is given by (Gregg and Sing, 1982) log / ˆ log /0 ÿ D log2 …p0 =p†;

…2†

where / is the amount adsorbed expressed as a liquid volume, /0 the total micropore volume of the sample and D is the Dubinin coecient. CO2 adsorption data, plotted according to Eq. (2), are shown in Fig. 2. The CO2 adsorption appears to be linear over the relative pressure range for the shells and derived chars, indicating that the DR equation satisfactorily describes present data. For each sample, the total micropore volume (/0 ) was obtained from the ordinate at the origin of the DR plots (Fig. 2). Although the DR concept involves pore volume ®lling, /0 is often converted to apparent surface area values (Gutierrez et al., 1988) following Eq. (3) SCO2 ˆ

/0 NA am ; v0M

…3†

Table 3 Nitrogen speci®c surface areas (SN2 ), total micropore volumes (/0 ) and apparent carbon dioxide surface areas (SCO2 ) as estimated from N2 and CO2 adsorption data applying BET and DR equations Sample

SN2 ´ 10ÿ3 (m2 kgÿ1 )

/0 ´ 103 (m3 kgÿ1 )

SCO2 ´ 10ÿ3 (m2 kgÿ1 )

SCO2 =SN2

Brazil Nut shells Char (T ˆ 350°C) Char (T ˆ 600°C) Char (T ˆ 850°C)

0.12 ‹ 0.01 0.87 ‹ 0.04 4.2 ‹ 0.2 3.3 ‹ 0.2

0.030 ‹ 0.001 0.034 ‹ 0.002 0.14 ‹ 0.01 0.20 ‹ 0.02

118 131 430 624

1180 146 108 208

P.R. Bonelli et al. / Bioresource Technology 76 (2001) 15±22

Fig. 2. DR plots for CO2 (25°C) adsorption on Brazil Nut shells and char samples obtained at the di€erent pyrolysis temperatures.

where am (CO2 ) ˆ 21.75 10ÿ20 m2 and v0M is the liquid molar volume of the adsorbate. The estimated values of the total micropore volumes and CO2 surface areas (SCO2 ) together with the (SCO2 =SN2 ) ratios for raw shells and derived chars are also presented in Table 3. Nitrogen surface area (SN2 ) reached a maximum at 600°C and slightly decreased at the highest temperature (850°C). Increases in SN2 may be attributed to release of volatiles, which favours the development of new pores that become accessible to N2 . At 850°C pores widening and/or the coalescence of neighbouring pores seem to predominate, leading to the decrease in SN2 . On the other hand, successive increases in SCO2 values for the shells treated with higher temperatures indicate progressive development of micropores. Nitrogen surface areas were considerably lower than those calculated from CO2 adsorption (SCO2 ) indicating very narrow micropores or obstructions of pore entrances restricting N2 di€usion. Nitrogen cannot penetrate pores slightly larger than the size of a N2 molecule due to an activated di€usion process that strongly depends on the temperature. CO2 , on the other hand, can di€use into smaller pores since its adsorption is carried out at higher temperatures and due to its ability to diffuse rapidly through the solid by a solubilization mechanism (L opez Peinado et al., 1985; Amarasekera et al., 1995). Thus, di€erences between values of SN2 and SCO2 suggest prevalence of micropores over meso and macropores in the nut shells and char samples obtained at all the temperatures. Further information about pore development may also be inferred from the SCO2 =SN2 ratio. This ratio decreases from a high value for the shells to lower ones for the char samples, indicating that the microporous structure of the raw samples drastically changes to be-

19

come more porous upon thermal treatment. This may be attributed to widening and/or coalescence of preexistent micropores, especially up to 600°C. The further increase in SCO2 =SN2 at 850°C is likely due to development of very large pores or cavities and, therefore, the relative contribution of micropores becomes again more predominant. Typical SEM photographs of nut shells and char samples at the di€erent pyrolysis temperatures are illustrated in Fig. 3(a)±(f). The external surface of the shell shows isodiametric polygonal ¯attened cells with well-de®ned lumen arranged in a pattern that resembles a honeycomb (Fig. 3(a)). In Fig. 3(b) are shown the SEM of shells at a higher magni®cation, where ordered compact cell structures are also observed. Pyrolysis causes noticeable changes in the shells' structure. Microphotographs of the external surfaces of the char samples obtained at 600°C (Fig. 3(c)) and 850°C (Fig. 3(e)) indicate progressive destruction of the cell lumen as the pyrolysis temperature increases; the surface of the char sample obtained at 850°C appeared mostly wrinkled. Figs. 3(d) and (f) display transverse sections of the char samples at 600°C and 850°C, respectively. The original cellular structure remained almost the same at 600°C, whereas the char at 850°C shows signs of structural destruction and walls rupture. This char sample presents deep cavities and small pores on the walls of the cavities (Fig. 3(f)). Moreover, for both char samples, particles of irregular shape were observed. These particles might be due to pyrolytic carbon deposits resulting from hydrocarbons cracking. During carbonization, hydrocarbons release from the samples as volatile matter. Collisions of these compounds with pore walls might bring about cracking and carbon deposition (Kumar and Gupta, 1995; Della Rocca et al., 1999). Removal of cell contents and consequently, the opening-up of cellular structures give rise to porous networks of distinctive features, which markedly depend on the pyrolysis temperature. SEM observations are consistent with results arising from N2 adsorption measurements. In addition, analyses by energy disperse X-ray spectroscopy (EDS) indicate high silicon and potassium contents in shell and char surfaces. Other elements present in lower proportion in the raw shells and chars are aluminium, iron, calcium, magnesium and chlorine, whereas no Na was detected in any of the samples. In particular, a relative increase in potassium content was found for the char obtained at the highest temperature, suggesting that K migrates from the inner to the surface of the shells upon pyrolysis. Migration of inorganic constituents to the surface, as a consequence of thermal conversion processes, has been also reported for other types of biomass (Wornat et al., 1995; Della Rocca, 1998).

20

P.R. Bonelli et al. / Bioresource Technology 76 (2001) 15±22

Fig. 3. SEM photographs of Brazil Nut shells and the char samples obtained at increasing pyrolysis temperatures: (a) external surface of nutshells (´200); (b) tranverse section of nutshells (´1000); (c) external surface of the char at 600°C (´200); (d) transverse section of the char at 600°C (´1000); (e) external surface of the char at 850°C (´200); (f) transverse section of the char at 850°C (´1000). Scale bars are indicated on the photographs.

P.R. Bonelli et al. / Bioresource Technology 76 (2001) 15±22

21

ÿdw=dt ˆ k…w ÿ w1 †

…5†

and k ˆ k0 exp…ÿEa =RT †;

Fig. 4. Thermogravimetric curve of Brazil Nut shells pyrolysis. Comparison between experimental and predicted values according to the deactivation model (DM).

3.3. Kinetics modelling A typical thermogravimetric (TG) curve at a heating rate of 0.83°C sÿ1 , namely values of sample weight loss (w ˆ m/mo ) as a function of temperature (T), is shown in Fig. 4. The TG curve shows that pyrolysis of the shells becomes substantial at 170°C, as indicated by the largest weight loss taking place up to around 420°C and which may be mainly due to holocellulose degradation. At higher temperatures, the slower degradation of the high lignin content in the shells (Table 2), known to show more resistance to pyrolysis than holocellulose (Antal, 1982), seems to become predominant. For models application, ®tting of TG curves (w vs T) was carried out taking into account the measured linear relationship between temperature and time, that is given by T ˆ Tamb ‡ vt;

…4†

where T is the instantaneous temperature, Tamb , ambient temperature, v the heating rate and t is the reaction time. As mentioned in Section 2.3, from preliminary experiments it was veri®ed that di€erent heating rates in the range 0.17±1.7°C sÿ1 led to similar w vs T curves. A simple model (SM) that assumes pyrolysis as a ®rst-order overall decomposition was used. The reaction rate is given by

…6†

where k is the speci®c rate constant, k0 the preexponential factor, Ea the activation energy, w1 the residual mass fraction, R the universal gas constant and T is the absolute temperature. The SM satisfactorily describes experimental data but only for a restricted range of temperatures, up to around 420°C. Estimated values of the model parameters together with the standard deviation and variation coecient are listed in Table 4. In order to ®t experimental data over the whole range of degradation temperatures investigated, namely up to 870°C, a deactivation model (DM) proposed in the literature (Balci et al., 1993) was used. The DM model assumes a ®rst-order overall decomposition, as given by Eq. (5), and considers that the signi®cant physical and chemical changes which take place within the solid as pyrolysis proceeds cause solid deactivation. This fact a€ects the reaction rate constant (kapp ) and is taken into account through an increase of the activation energy with the temperature and the solid conversion according to kapp ˆ k0 exp fÿEa0 …1 ‡ bTzc †=RT g; where z is the normalized fractional conversion …8†

z ˆ 1 ÿ w=1 ÿ w1

and Ea0 the initial activation energy, for z ˆ 0 and b, c, ®tting parameters (b the deactivation rate, c the order with respect to z). In the regression analysis, the values of preexponential factor and activation energy obtained for the SM, as detailed in Table 4, were used and kept constant. Thus, Ea0 in Eq. (7) constitutes an initial value of the activation energy; only, b and c were estimated. DM model parameters as well as S.D. and VC are reported in Table 4. The DM predicts increasing activation energy values comprised between 47.2 and 82 kJ molÿ1 for the range of temperatures from 25°C to 870°C. The deactivation model was ®tted to experimental results as shown in Fig. 4. The DM seems to properly describe experimental data over the whole range of Brazil Nut shells' decomposition.

Table 4 Estimation of model characteristic parameters Model

Estimated parameters

SM

k0 ˆ 5.7 ´ 102 sÿ1 Ea ˆ 47.2 kJ molÿ1 w1 ˆ 0.51 b ˆ 0.00083°Cÿ1 cˆ2

DM

…7†

Temperature range (°C)

S.D. (%)

VC (%)

25±440

0.5

0.6

25±870

1.4

2.2

22

P.R. Bonelli et al. / Bioresource Technology 76 (2001) 15±22

4. Conclusions Pyrolysis of Brazil Nut shells leads to signi®cant changes in chemical and textural±morphological characteristics of lignocellulosic waste that depend markedly on temperature. Release of volatile matter brings about the opening-up of the residual cellular structure and consequently, the development of porosity. The development of pore networks of predominantly microporous character was largely dependent on pyrolysis temperature. Progressive increases in micropores surface with temperature take place, even though a maximum development of larger pores arises at 600°C suggesting enlargement of preexistent micropores or coalescence between neighboring ones. At the highest temperature (850°C), union of larger pores seems to occur and micropores development becomes strongly predominant. Kinetics of Brazil Nut shells pyrolysis over the wide range of temperatures from 25°C to 900°C has been studied. Only a model which considers the signi®cant changes taking place within the solid as pyrolysis proceeds through an increase in activation energy with temperature and residue conversion can adequately represent the experimental data over the broad range of temperatures investigated. Acknowledgements The authors gratefully acknowledge Consejo Nacional de Investigaciones Cientõ®cas y Tecnicas (CONICET), Universidad de Buenos Aires (UBA) and Agencia Nacional de Promoci on Cientõ®ca y Tecnol ogica (ANPCYT) from Argentina for ®nancial support. References Antal Jr., M.J., 1982. Biomass pyrolysis: a review of the literature. Part 2 ± Lignocellulose pyrolysis. In: Boer, K.W., Due, J.A. (Eds.), Advances in Solar Energy, vol. 2. American Solar Energy Society, Boulder, CO. Antal Jr., M.J., Varhegyi, G., 1995. Cellulose pyrolysis kinetics: the current state of knowledge. Ind. Eng. Chem. Res. 34 (3), 703±717. Amarasekera, G., Scarlett, M.J., Mainwaring, D.E., 1995. Micropore size distributions and speci®c interactions in coals. Fuel 74 (1), 115± 118. Balci, S., Dogu, T., Y ucel, H., 1993. Pyrolysis kinetics of lignocellulosic materials. Ind. Eng. Chem. Res. 32, 2573±2579. Bridgwater, A.V., Bridge, S.A., 1991. A review of biomass pyrolysis and pyrolysis technologies. In: Bridgwater, A.V., Grassi, G. (Eds.), Biomass pyrolysis liquids upgrading and utilisation. Elsevier Applied Science, London and New York (Chapter 2). Browning, B.I., 1970. Methods of wood chemistry, vol. 2, John Wiley, New York (Chapter 19). Byrne, J.F., Marsh, H., 1995. Introductory overview. In: Patrick, J.W. (Ed.), Porosity in carbons. Halsted Press, New York (Chapter 1).

Caballero, J.A., Conesa, J.A., Font, R., Marcilla, A., 1997a. Pyrolysis kinetics of almond shells and olive stones considering their organic fractions. J. Ann. Appl. Pyrolysis 42, 159±175. Caballero, J.A., Marcilla, A., Conesa, J.A., 1997b. Thermogravimetric analysis of olive stones with sulphuric acid treatment. J. Ann. Appl. Pyrolysis 44, 77±88. Cukierman, A.L., Della Rocca, P.A., Bonelli, P.R., Cassanello, M.C., 1996. On the study of thermochemical biomass conversion. Trends in Chemical Engineering 3, 129±144. Della Rocca, P.A., Horowitz, G.I., Bonelli, P.R., Cassanello, M.C., Cukierman, A.L., 1997. Olive stones pyrolysis: chemical, textural and kinetics characterization. In: Bridgwater, A.V., Boocock, D.G.B. (Eds.), Developments in Thermochemical Biomass Conversion. Blackie, London, pp. 176±190. Della Rocca, P.A., 1998. Study on biomass thermal conversion processes. Ph.D. thesis, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires. Della Rocca, P.A., Cerrella, E.G., Bonelli, P.R., Cukierman, A.L., 1999. Pyrolysis of hardwoods residues: on kinetics and chars characterization. Biomass and Bioenergy 16, 79±88. Demirbas, A., 1998. Kinetics for non-isothermal ¯ash pyrolysis of hazelnut shell. Biores. Technol. 66, 247±252. Di Blassi, C., Signorelli, G., Russo, C., Rea, G., 1999. Product distribution from pyrolysis of wood and agricultural residues. Ind. Eng. Chem. Res. 38 (6), 2216±2224. Encinar, J.M., Beltr an, F., Ramiro, A., Gonz alez, J.F., 1997. Catalyzed pyrolysis of grape and olive bagasse, In¯uence of catalyst type and chemical treatment. Ind. Eng. Chem. Res. 36 (10), 4176± 4183. Font, R., Marcilla, A., Verdu, J., Devesa, J., 1991. Thermogravimetric kinetic study of the pyrolysis of almond shells and almond shells impregnated with CoCl2 . J. Anal. Appl. Pyrolysis 21, 249±264. Gregg, S.J., Sing, K.S.W., 1982. Adsorption, surface area and porosity. Academic Press, London. Gutierrez, M.C., Cukierman, A.L., Lemco€, N.O., 1988. Study of subbituminous coal chars: e€ect of heat treatment temperature on their structural characteristics. J. Chem. Tech. Biotech. 41, 85±93. Heschel, W., Klose, E., 1995. On the suitability of agricultural byproducts for the manufacture of granular activated carbon. Fuel 74 (12), 1786±1791. Horowitz, G.I., Lombardi, G., Cassanello, M.C., Cukierman, A.L., 1997. Gasi®cation of an agricultural by-product: a study towards activated carbons preparation. In: Proceedings of Fourth Asian Paci®c International Symposium on Combustion and Energy Utilisation. Chulalongkorn University, Bangkok, pp. 757±761. Kumar, M., Gupta, R., 1995. Scanning electron microscopic study of acacia and eucalyptus wood. J. Mater. Sci. 30, 544±551. L opez Peinado, A., Rivera-Utrilla, J., L opez-Gonz alez, J.D., MataArjona, A., 1985. Porous texture characterization of coals and chars. Adsorption Sci. Technol. 2, 31±38. Magnaterra, M.R., Fusco, J.R., Ochoa, J., Cukierman, A.L., 1994. Kinetic study of the reaction of di€erent hardwood sawdust chars with oxygen, chemical and structural characterization of the samples. In: Bridgewater, A.V. (Ed.), Advances in thermochemical biomass conversion, Blackie, London, pp. 116±131. Rashid Khan, M., Gorsuch, C., 1996. Utilization of waste materials: examples of business successes. In: Rashid Khan, M. (Ed.), Conversion and utilization of waste materials. Taylor & Francis, Washington, DC, pp. 3±13. Tancredi, N., Cordero, T., Rodrõguez-Mirasol, J., Rodrõguez, J.J., 1997. Activated carbons from eucalyptus wood, In¯uence of the carbonization temperature. Sep. Sci. Technol. 32 (6), 1115± 1126. Wornat, M.J., Hurt, R.T., Yang, N.Y.C., Headley, T.J., 1995. Structural and composition transformations of biomass chars during combustion. Combustion and Flame 100, 131±143.