Abstract
The aim of this work was to thermally characterize the renewable lignocellulosic bioresources derived from palm trees in order to highlight their energy potential. Pyrolysis and combustion behaviours of date stones (DS) agricultural by-products were tested by thermo-gravimetric analysis, and the main chemical compositions were analyzed. The work has also been conducted to identify their most important physical characteristics. The study of the sizes and heating rate effects constitute the first part of the experimental work. Inert atmosphere and three heating rates: 10, 20, and 50 °C/min, were applied to various particle sizes of DS. In the second part, tests were carried out in an oxidizing atmosphere (21% O2) by varying the size of the DS. The kinetic parameters such as pre-exponential factor and activation energy were determined. Increasing the particle sizes and the heating rates didn’t have an appreciable influence on the global weight losses. However, degradation rates were significant with the porous structure of the DS. Weight losses in inert and oxidizing atmospheres were found to occur in two stages (drying and devolatilization) and in three stages (drying, devolatilization, and oxidation of the char).
Download PDF
Full Article
THERMOGRAVIMETRIC INVESTIGATION AND THERMAL CONVERSION KINETICS OF TYPICAL NORTH AFRICAN AND MIDDLE EASTERN LIGNOCELLULOSIC WASTES
Imen Abed,a,b,* Maria Paraschiv,a Khaled Loubar,a Fethi Zagrouba,b and Mohand Tazerout a
The aim of this work was to thermally characterize the renewable lignocellulosic bioresources derived from palm trees in order to highlight their energy potential. Pyrolysis and combustion behaviours of date stones (DS) agricultural by-products were tested by thermo-gravimetric analysis, and the main chemical compositions were analyzed. The work has also been conducted to identify their most important physical characteristics. The study of the sizes and heating rate effects constitute the first part of the experimental work. Inert atmosphere and three heating rates: 10, 20, and 50 °C/min, were applied to various particle sizes of DS. In the second part, tests were carried out in an oxidizing atmosphere (21% O2) by varying the size of the DS. The kinetic parameters such as pre-exponential factor and activation energy were determined. Increasing the particle sizes and the heating rates didn’t have an appreciable influence on the global weight losses. However, degradation rates were significant with the porous structure of the DS. Weight losses in inert and oxidizing atmospheres were found to occur in two stages (drying and devolatilization) and in three stages (drying, devolatilization, and oxidation of the char).
Keywords: Renewable energy; Pyrolysis; Combustion; Thermogravimetric analysis; Lignocellulosic by-products; Kinetics
Contact information: a: GEPEA, UMR 6144 CNRS/ Université de Nantes/Ecole des Mines de Nantes/ENITIAA, DSEE, 4 Rue Alfred Kastler, BP 20722, 44307 Nantes Cedex 3, France; b: Institut Supérieur des Sciences et Technologies de l’Environnement de Borj Cédria CP 1003 Hammam-Lif, Tunis, Tunisie; * Corresponding author: Email: imenabed@hotmail.com
INTRODUCTION
The recognised effects of climate change have prompted an urgent appraisal of realistic alternative power generation options. There are various forms of alternative energy sources, such as solar, wind, biogas/biomass, tidal, geothermal, fuel cell, hydrogen energy, etc. The attraction of these sources lies primarily in their abundance and ready access. The Renewable Energy Technologies (RETs) for exploiting these sources include biogas plants, solar lanterns, solar home lighting systems, improved cook stoves, improved kerosene lanterns, solar water pumping systems, solar water heating systems, and water mills. Currently, these options cannot replace the totality of fossil fuel exploitation, but the basic concept of alternative energy relates to issues of sustainability, renewability, and pollution reduction.
In this context, one of these options is to replace a part of fossil fuel with a renewable energy source such as lignocellulosic biomass. Fossil fuels – first coal, then petrol, diesel, and natural gas – have dominated our fuel market. In spite of this, biomass in the form of fuelwood continues to be a major source of energy to this day.
Nearly 70% of India’s cooking energy requirement and 32% of its primary energy requirement is met with biomass (MNRE 2009). Globally, fuelwood in its various forms accounts for about 64% of the estimated total world supply of combustible renewable energy sources (Demirbas et al. 2009). The estimates of worldwide annual generation of electricity from biomass amount to about 185 TWh, of which nearly three quarters are produced from solid biomass, 14% from biogas, and 12% from municipal solid waste. If we consider the portion contributed by biomass to the world’s total energy production, it comes to only less than 2% (World Energy Council 2007).
For some regions in the world, namely North African and Middle Eastern countries, fossil fuel sources of oil, gas, and coal are predictably running out. Pressures exerted by population and economic growth as well desalination needs increase demand for electricity (Rogner and Abdel-Hamid 2008). In fact, Tunisia and Morocco are already energy importers, while resources in Algeria, Libya, and Egypt are likely to be exhausted in a couple of generations. These countries must research other sources of energy recovery, with a special emphasis on the usage of agricultural waste.
The date palm tree has been considered an important resource for thousands of years. Dates are used as a raw material in many food industries, as well as a constituent in some animal feeds. In addition to the fruit itself, almost every part of the tree is put to use. Wastes of palm trees, which are produced in huge amounts in North Africa and the Middle East, have proved to be excellent biofuels (Hamada et al. 2002). Haimour and Emeish (2006) found that date stones constitute approximately 10% of the fruit’s weight, while Al-Badri and Lafta (1989) found that about 14% of the fruit’s weight is waste material. Adapted from a number of literature references according to Food and Agriculture Organization (FAO) to agricultural services, carbohydrates are the major components of date stones (Barreveld 1993). Indeed, these palm trees by-products are composed of cellulose (42%), hemicellulose (18%), lignin (11%), ash (4%), sugar and other compounds (25%).
Energy processes based on lignocellulosic biomass can be divided into two broad categories: biological (fermentation and anaerobic digestion) and thermochemical (combustion, gasification, and pyrolysis). These latter processes can significantly and immediately reduce the mass and volume of wastes, and allow for energy recovery.
In order to evaluate the thermal decomposition of the material, thermoanalytical techniques, in particular thermogravimetry (TGA) and derivative thermogravimetry (DTG), provide basic information in a relatively simple and straightforward manner. Generally, neither a systematic classification of biomass fuels, based on thermogravi-metric analysis, nor a general mechanism to interpret such measurements are available, especially for such specific types of lignocellulosic matter.
The aim of the present work was to study the pyrolytic and combustion behaviour of lignocellulosic material (palm by-products) and to obtain a variety of pyrolysis data by changing some factors, namely particle size, heat flow, and reaction atmosphere, with the ultimate purpose of using this kind of biomass for energy recovery and to promote the possibility of developing an uncomplicated thermal conversion technology system for co-pyrolysis, co-combustion, or co-gasification with wood. Thereafter, it will be interesting to develop new strategies on a larger scale, such as setting up facilities near small farms, making it possible to investigate the practical usage of lignocellulosic bioresources.
For the purpose of this research, date stones from palm trees were investigated: (named hereafter DS). Palm trees produce also another type of lignocellulosic by-product, which is palm stalk (PS). These agricultural wastes originated from Djerid, an area located in the south of Tunisia. A particular attention has been paid to investigate thermal analyses for DS after confirming the close similarity of these two lignocellulosic by-products. After drying at a temperature that didn’t exceed 50 °C during 72 hours, these palm by-products were ground to a desirable particle size and sieved to a powder of less than 1mm in diameter and kept in a desiccator to protect them from recapturing the hygroscopic moisture. DS was concerned for TGA/DTG analysis and kinetic parameter determination.
EXPERIMENTAL SETUP AND CHARACTERIZATION
In order to characterize DS and PS in a thermal setting, different analyses were conducted (Table 1).
Proximate Analyses
Moisture was determined by steaming at 105 °C for 24 hours, until a constant mass is achieved. The mineral matter was determined by calcinations at 550°C during 15 min, and the fat matter was quantified by the Soxhlet method.
The higher heating value of ground samples was measured using an adiabatic oxygen bomb calorimeter (Prolabo model) under 25 bars. The energy content of the dry matter was then calculated.
Ultimate Analyses
Carbon, hydrogen, and nitrogen compositions were determined by total combus-tion at 1050 °C under an oxygen flow. Carbon and hydrogen were transformed respect-tively into quantified carbon dioxide and water, either by catharometry or by specific infrared detectors. Nitrogen from the samples was transformed into various oxides of nitrogen, which were reduced in molecular nitrogen before being quantified by thermal conductemetry. Sulphur content was determined by an analyzer type LECO (model SC 144) performing total combustion at 1350 °C under an oxygen flow. It was transformed into sulphur dioxide, quantified using an infrared-specific detector.
Chlorine was determined by mineralization in an oxygen atmosphere by a combustion type SCHONIGER. Quantitative obtaining of the element in the form of chlorine allows proportioning with silver nitrate by an ordinary potentiometer. Oxygen was determined by the difference. All the experimental tests were carried out in triplicate.
Structural Characterization
The morphology and the chemical distribution of the elements on the surface of raw material were examined using scanning electron microscopy, combined with the ability to generate localised chemical information through Energy dispersive X-ray spectroscopy (EDX) under ultra-vacuum (Fei Quanta 200).
X-ray patterns were recorded with a Panalytical X’Pert PROMPD diffractometer with a wavelength of 1.789 °A. The X-ray patterns were recorded in the 2θ range 0 to 70° with a scan rate of 0.017° min-1. All the samples were films with similar thickness of 0.3 mm.
Thermogravimetric Analyses
The first part of TGA/DTG experiments was performed using a SETARAM/ SETSYS Evolution balance. Approximately 27 to 53 mg of DS samples were placed in the pan of the TGA microbalance, which was operated under a constant purge flow rate of 20 ml/ min of Argon under a pressure of 1032 mbar.
Three heating rates were applied (10, 20, and 50 °C/min) for 1.18 mm, 2.36 mm, and 4.75 mm of DS and the powdered form. Residual weight of these samples and the derivative of weight, with respect to time and temperature (differential thermogravimetry analysis, DTG), were recorded using TGA software. Temperature instructions went from 40 to 1000 °C.
The second part of the TGA/DTG experiments was carried out with a thermo standard balance SETSYS 1750 devoted to combustion under 21% of 02 at 10 °C/min.
RESULTS AND DISCUSSION
Ultimate and Proximate Analyses
The higher heating values of DS and PS were respectively 22.27 MJ.kg-1 and 19.66 MJ.kg-1, which were higher than the energy content of a multiple agricultural by-product such as rapeseed straw (17.64 MJ.kg-1) (Karaosmanoglu et al. 2001), pine chips (18.98 MJ.kg-1) (Sensoz et al. 2002), rice husk (14.42- 18.31 MJ.kg-1) (Mansaray and Ghaly 1997a), and wood (19.10 MJ.kg-1) (Kastanaki and Vamvuka 2006). Moisture related to the rough weight is about 8.00% and 5.8% for DS and PS, compared to 20% for dried wood and 45% for fresh wood. Table 1 summarizes the composition of DS and PS.
Table 1. Proximate and Ultimate Analyses of Palm Trees By-products
Scanning Electron Microscopy/ EDX Ultra-Vacuum
Scanning electron micrographs of raw DS and PS are shown in Fig. 1. The examination of the initial structure of the DS shows the presence of a few macropores (Fig. 1a) of various size and geometry on the surface. However, it is a fibrous form of PS (Fig. 1b).
Fig. 1. Scanning electron micrographs of (a) DS and (b) PS samples
More than one zone of the surface was analysed, since the chemical elements distribution is heterogeneous (Table 1).
These surfaces were the first contact field for many thermal transfer interactions during the thermal conversion process of the matter. With regard to the mineral content (Tables 2 and 3), it was found that the external surfaces of DS show appreciable amounts of carbon, and oxygen. Among the elements present in a few amounts in the DS matter are cobalt, aluminium, silicon, etc. Iron, chlorine, phosphorus, magnesium and others element were considered as traces.
Table 2. EDAX Quantification of DS and PS surface (Powder)
Table 3. EDAX Quantification of DS and PS surface (Particle)
Similar to the DS sample, external surfaces of PS show appreciable quantities of carbon, and oxygen, rather a significant amounts of chlorine and potassium. Silicon, calcium, phosphorus and potassium are known to be plant nutrients and soil improvement agents (Steenari and Lindqvist 1997). The distribution of chemical elements shows that the raw materials are highly heterogeneous depending on the form of the ground DS and PS. The potassium ion is one of the most abundant ions that naturally occur in native material.
It is important to identify the inorganic element distributions since they have an important effect on the pyrolysis temperature and on the products’ composition (char, tar, and gases).
On the other hand, when some research opts for a large scale application to investigate this lignocellulosic matter, it is necessary to highlight that high amounts of inorganic constituents in some lignocellulosic biomasses contribute to adverse impact on the different elements of the conversion systems. The chlorine and potassium in biomass are water-soluble, they can largely be removed through leaching, thus mitigating their impact on high temperature conversion devices. Water washing is a means to eliminate some metals, thus improving behaviour during thermal treatments (Jenkins et al. 1996).
Table 4. Elemental Compositions of the Raw Materials
Ash constituents, especially potassium, sodium, and calcium, act as catalysts for the decomposition process and favour char formation (Jensen et al. 1998 ; Fahmi et al. 2007). Chemical analyses of inorganic elements in the raw materials are given in Table 4. It can be noticed that this type of by-product shows an absence of heavy metals.
XRD Analysis
X-ray diffraction patterns of natural DS and PS are given in Fig. 2. The diffractogram of natural DS does not exhibit a horizontal basic line. This shows that the major part of the matter is amorphous. However, a few diffraction peaks emerge from the basic line, indicating the presence of a small amount of crystalline matter.
Fig. 2. XRD patterns of (a) (DS) and (b) (PS)
The XRD pattern of natural date pits has been compared to those of native cellulose (C6 H12 O6), xylane dehydrate (C10 H12 O9.2 H2 O), or hemicellulose dehydrate given in the JCPDF crystallographic data base (Table 5) (Zhao et al. 2007).
Table 5. Bragg Diffraction Angle (°2Th) and Reticular Distance (d) of DS and their Corresponding Compounds Identified in the JCPDF Crystallographic Data Base (N.Cl : native cellulose, HCel : hemicellulose dehydrate, C : carbon)
Characterization studies on biomass samples are quite important to judge suitability of feedstock for thermochemical conversion. The two lignocellulosic materials presented and characterized above showed a very close similarity. However, because of the important higher heating value of DS sample, we have chosen DS to investigate thermal gravimetric analyses and to show more detail for the thermal degradation behaviors.
Inert Atmosphere
TGA analyses were operated under a constant purge flow rate of 20 mL/min of Argon. Three heating rates were applied at 10, 20, and 50 °C/min.
Typical representations of TGA/DTG diagram for DS are shown in Fig. 3 under inert atmosphere and with a heating rate of 10 °C/min. The raw DS diagram shows three stages; drying, fast pyrolysis, and slow pyrolysis.
The first stage of weight loss ranged from 47 °C to around 160 °C (Fig. 3). The derivative plot (DTG) had a separate peak for this step of thermal process, which was attributed to the moisture loss and probably some volatile compounds in the biomass sample so that the low moisture content resulted in low weight loss during this stage.
The second stage corresponding to fast pyrolysis had three zones. Following this stage, the first zone (Z 1) showed a negligible weight loss (≈ 5%) in the temperature range of 135 °C to 230 °C, this phase was identified as the beginning of the decomposition of cellulose and hemicellulose. The second zone (Z 2) started around 230 °C and extended to 330 °C. Weight loss during this stage was rapid; it was the active pyrolysis step so the second stage was designated as the fast pyrolysis.
Zone 3 of the fast pyrolysis stage showed a rapid weight loss, but relatively less important (20%) than the previous stage, which probably corresponds to the first step of lignin decomposition; this indicates that the amount of lignin is significant in the raw DS.
It was concluded that the most distinguished peak in fast pyrolysis stage is related to the complete pyrolysis of hemicellulose, which took place until 350 °C, while the pyrolysis of cellulose occurred from 250 °C to 500 °C (Fig. 3). The same results have been reported in the literature concerning thermal degradation of rice husks (Manasary and Ghaly 1998c; Varhegyi et al. 1997).
Kilzer and Broido (1965) have proposed a general kinetic scheme for the pyrolysis of pure cellulose. Its decomposition would occur according to two competitive reactions occurring directly from cellulose. The first one (473 to 553 °C) is a slightly endothermic reaction of dehydratation, followed by an exothermal process that produces char and light gaseous species. In the second one (553 to 613 °C), cellulose is postulated to be transformed into an intermediate and unstable compound. A similar scheme has been proposed by Arsenau (1971).
The derivative plot of the region between 230 °C and 500 °C showed two observable peaks but a slow degradation of biomass continued until 900 °C, at which point complete devolatilization occurred.
It can be seen that the lignin degradation was slow and occurred over a wide range of temperatures (Vamvuka et al. 2003a). Following this stage, there was a continuous and slow weight loss from 500 °C to 1000 °C; this stage was so-called “slow pyrolysis,” and could be a thermal continuity of decomposition of lignin and other compounds with high molecular weights.
Fig. 3. Typical TGA diagram of DS (powder) in an Argon atmosphere (10 °C/min)
Such thermal behaviour can be explained by the structural differences in lignin, cellulose, and hemicellulose. Cellulose and hemicellulose are basically composed of polysaccharides. On the other hand, lignins are complex racemic polymers derived mainly from three hydroxycinnamyl alcohol monomers that differ in their degree of methoxylation and are heavily cross-linked. Lignin polymerization is characterized by some typical inter-unit linkages.
Thus, lignin has a high thermal stability and it is difficult to decompose (Yang et al. 2006). Lignin pyrolysis takes place at a pyrolysis temperature range from 150 °C to 900 °C, and no sharp weight loss peak appears (Li et al. 2004). Since the cellulosic compounds have the structure of branching a chain of polysaccharides and no aromatic compounds, that’s why their volatilization is easy (Gani and Naruse 2007).
Park, Atreya, and Baum (2010) reported that wood consists of three major components: hemicellulose (25-35%), cellulose (40-50%), and lignin (16-33%), a composition range similar to DS. Each component displays different characteristics during its thermal decomposition. Hemicellulose decomposition occurs at 200-260 °C and produces acetic acid. Cellulose decomposes to levoglucosan and dehydrocellulose at 240-350 °C. Lignin decomposes over a broad temperature range of 280-500 °C and produces more char than the other two components. These temperature ranges for degradation also describe the thermal behaviour of DS, so it is interesting to plan for co-pyrolysis of wood and DS at a large scale.
Effect of heating rate during pyrolysis of date stones
Figures 4 and 5 illustrate the influence of the heating rate on the degradation process of DS (˂100 µm and initial weight ˂ 40 mg). Few workers have discussed testing the effect of heating rate on the devolatilization behaviour of biomass. The most common ramp rate reported in the literature for TGA characterisation of biomass appears to be 10 °C/min (Vamvuka et al. 2003b), 20-30 °C/min (Peralta et al. 2002), 50 °C/min (Moghtaderi et al. 2004), or 100 °C/min (Pan et al. 1996), with relatively few using lower heating rates. It is therefore useful to investigate different heating rate profiles. Heating rate had an effect on the temperature range of the stages of the reaction (Fig. 4). Increasing the heating rate appeared to increase the range of temperatures of each stage, which may have been due to the heat transfer limitations.
Three temperatures were recorded: Tini is the temperature where weight loss first reached 0.5 mg/min; Tpeak is where the weight loss reached its maximum; and Tend is where the weight loss fell below 0.5 mg/min. The limit of 0.5 mg/min was established using actual profiles as an arbitrary limit for the start and end for peaks (Fig. 4).
Fig. 4. Effect of heating rate on DTG during pyrolysis of DS
Fig. 5. Effect of heating rate on TGA during pyrolysis of DS
It was expected that the peak temperature would increase with an increasing heating rate. However, in this study, for 50 °C/min there was a lower peak temperature than that registered for 10 °C/min. In fact, high heating rates and large mass loading tend to deflect the measured temperature from the actual sample temperature. It is clear that by increasing the heating rate, a longer time may be required for the purging gas to reach equilibrium with the temperature of the furnace or the sample because of the heat transfer limitations.
Char characterization
Characterization studies on biomass samples are quite important to express suitability of feedstock for thermochemical conversion. High volatile matter content of biomass with low ash and sulphur content is the main criterion for pyrolysis conversion. From the main characteristics of the raw material (Table 1), it appears that the high volatile content of lignocellulosic biomass favours pyrolysis conversion.
Scanning electron microscopy (SEM) was used to characterize the size and the shape of the char particles, as well as their porous surface structure (Fig. 6). Pyrolysis temperature and heating rate influenced the size and shape of particle due to a general increase in size and proportion of voids and a decrease in cell wall thickness. For DS biomass it seemed to be more favourable to apply pyrolysis at 500 °C, where the most highly porous and spherical particles were observed. Fast pyrolysis of DS does not favour the thermal conversion of the matter. At low pyrolysis temperature (Fig. 6b) and low heating rate, no porous structure was seen; the particles were thick-walled and covered by tar agglomerates. The wood sample presented a slightly porous surface at 600 °C and at a slow heating rate; wood samples maintain a rigid aspect, which is due to a natural lignin-developed structure. The diffractogram of DS having undergone pyrolysis under inert gas (Fig. 7) showed significant effects of pyrolysis: several peaks disappeared due to the decomposition of cellulose and hemicellulose during thermal conversion.