Pyrolysis of Medium Density Fiberboard (MDF) wastes in a screw reactor

Energy Conversion and Management 92 (2015) 223–233 Contents lists available at ScienceDirect Energy Conversion and Management journal homepage: www.elsevier.com/locate/enconman Pyrolysis of Medium Density Fiberboard (MDF) wastes in a screw reactor Suelem Daiane Ferreira a,c,⇑, Carlos Roberto Altafini a, Daniele Perondi b,c, Marcelo Godinho b a Postgraduate Program in Mechanical Engineering, Professional Master Degree, University of Caxias do Sul, Caxias do Sul, Rio Grande do Sul, Brazil Postgraduate Program in Engineering Processes and Technologies, University of Caxias do Sul, Caxias do Sul, Rio Grande do Sul, Brazil c Postgraduate Program in Mining Engineering, Metallurgical and Materials, Federal University of Rio Grande do Sul (UFRGS), Porto Alegre, Rio Grande do Sul, Brazil b a r t i c l e i n f o Article history: Received 8 October 2014 Accepted 12 December 2014 Available online 8 January 2015 Keywords: Medium Density Fiberboard Pyrolysis Screw reactor and residence time a b s t r a c t Medium Density Fiberboard (MDF) wastes were undergoes via a thermal treatment through of a pyrolysis process. Pyrolysis was carried out in a pilot scale reactor with screw conveyor at two reaction temperatures (450 and 600 °C) and, for each one, three solid residence times (9, 15 and 34 min) were evaluated. Products (char/bio-oil/fuel gas) of the pyrolysis process were characterized and quantified. Results revealed that the products yields were influenced by pyrolysis temperature, as well as by solid residence time. Char yield ranged between 17.3 and 39.7 (wt.%), the bio-oil yield ranged between 23.9 and 40.0 (wt.%), while the fuel gas yield ranged between 34.6 and 50.7 (wt.%). The samples surface area at 450 and 600 °C in 15-min residence time were surprisingly high, 415 and 593 m2 g1, respectively, which are compatible with the superficial area of commercial activated carbons. Energetic efficiency of process was estimated from energetic content present in the reaction products and the energetic content of MDF wastes, and the following results were obtained: 41.4% (fuel gas), 35.5% (char) and 29.2% (bio-oil). The contribution of this work is the development of a detailed study of the MDF pyrolysis in a pilot reactor with screw conveyor that supports the biorefineries concept. Ó 2014 Elsevier Ltd. All rights reserved. 1. Introduction A biorefinery is a facility that integrates biomass conversion processes and equipment to produce fuels, power, and chemicals from biomass. Biorefineries combine the necessary technologies of the bio-renewable raw materials with those of chemicals intermediates and final products [1,2]. Bio-renewable feedstocks can be used as solid fuel, or converted into liquid or gaseous forms for the production of electric power, heat, chemicals, or gaseous and liquid fuels [3,4]. Accordingly, the pyrolysis is related to the biorefinery concept, characterized by the thermal decomposition of a solid fuel, generating valuable products on the market such as char, oil and fuel gas [5–7]. In Brazil, there are about 15 thousand companies in the furniture sector and the State of Rio Grande do Sul has 2700 companies making a profit of about U$ 9.3 billion, which represents 15.6% of the national income. Furniture industry generates a great amount of wood wastes; among them is the Medium Density Fiberboard ⇑ Corresponding author at: Postgraduate Program in Mining Engineering, Metallurgical and Materials, Federal University of Rio Grande do Sul (UFRGS), Porto Alegre, Rio Grande do Sul, Brazil. E-mail address: suelemferreira2006@yahoo.com.br (S.D. Ferreira). http://dx.doi.org/10.1016/j.enconman.2014.12.032 0196-8904/Ó 2014 Elsevier Ltd. All rights reserved. (MDF). It is estimated a monthly volume of 5160 m3 and about 332 tons of MDF wastes generated in Rio Grande do Sul. Bento Gonçalves city produces around 141.1 tonnes of MDF wastes monthly, it is considered the largest furniture center of Rio Grande do Sul and the second largest center of Brazil [8]. The total annual MDF waste generated by five major UK and Ireland producers was approximately 23,000 m3 (15.000 tons) in 2000 and the volume is rising rapidly due to the growing popularity of MDF architectural mouldings. Most of the MDF waste is currently sent to combustion on sites while a small proportion goes to landfills with cost for the manufacturers. As the demand for MDF architectural mouldings is rapidly growing, the fine sawdust also presents and increasing cost and environmental responsibility for the manufacturers. The dust is too fine and voluminous to be easily handled, stored and transported. It also contains free formaldehyde residue rendering it hazardousness through inhalation [9]. MDF is made with wood from planted forests of pine and eucalyptus. In addition, there are about 10 (wt.%) of resins in MDF composition, in particular urea formaldehyde [9]. The MDF wastes in Brazil are often sent to potteries, poultry and other locations, where it is burned as an energy source. In this context, the thermal treatment through pyrolysis can be an alternative to the MDF wastes final disposal. 224 S.D. Ferreira et al. / Energy Conversion and Management 92 (2015) 223–233 The reactor is one of the main components involved in the pyrolysis process. In accordance with Jahirul et al. [10], pyrolysis of biomass can be performed in various types of reactors, but the most common reactors are: fixed bed; bubbling fluidized bed and circulating fluidized bed. A special design of pyrolysis reactor is the semi-continuous type with screw conveyor (also named auger reactor). Auger reactor is externally heated, where the temperature and the rotation speed of the screw can be set separately [10,11]. Brown and Brown [12] stated that the auger reactor had high potential in technical auger (heat carrier temperature and mass flow rate, auger rotational speed and sweep gas volumetric flow rate) and market aspects, since it is relatively simple, applies low carrier gas flow and is suitable for large biomass particles. Previous studies investigated two types of auger reactor, namely single screw [13,14] and twin screw [12], in pyrolysis of biomass. Majority of the works are focused on single screw reactors, as it is simpler in design and operation. A similar single screw system without heat carrier (that is, heat is delivered externally through the reactor shell) is described by Puy et al. [13] using pine wood. These studies employ slightly different operating parameters (auger speed, biomass flow rate, temperature, sweep gas flow rate) and feedstocks. Results show that the greatest yields for biooil production (59%) and optimum product characterization were obtained at a 500 °C temperature and solid residence times longer than 2 min were applied. Brown and Brown [12] work in an auger reactor heated by an externally supplied heat carrier was used for thermal treatment (pyrolysis) of red oak wood biomass. A statistically designed set of experiments was performed in order to optimize the process conditions for maximizing bio-oil yield. This systematic, simultaneous investigation of multiple variables (heat carrier temperature and mass flow rate, auger rotational speed, and sweep gas volumetric flow rate) allowed the identification of interaction effects among the variables. It was concluded that the conditions for maximum bio-oil and minimum char yields were high flow rate of sweep gas (3.5 standard L min1), high heat carrier temperature (600 °C), high auger speeds (63 RPM) and high heat carrier mass flow rates (18 kg/h). Liaw et al. [14] investigated the effect of pyrolysis temperature on the yield and composition of bio-oils obtained from the auger pyrolysis of Douglas Fir wood. The maximum yield of bio-oil at 500 °C was about 59% while the yield of char was 13%. The results confirm that the auger reactor is able to achieve good yields of both bio-oil and bio-char, but the overall oil composition obtained will be affected by the slower heating rates achieved and the intensification of secondary reactions in gas phase. Gas generated in the pyrolysis process is commonly used as energy source to heat the pyrolysis reactor. Oils can be easily stored and transported to centralized or rural biorefineries where second generation transportation fuels and chemicals can be produced with economies of scale [15]. The char is rich in carbon and nutrients, which can be used as fuel [13] or as fertilizer replacement offering an advanced option for biological sequestration of carbon [16–18]. Char can be produced as activated carbon, and used to separate CO2 present in the biogas generated in the anaerobic digestion of manure. Few studies are reported on the thermal conversion of MDF wastes, but there are not studies about pyrolysis from these wastes with auger reactors. Yorulmaz and Atimtay [19] investigated the MDF combustion kinetics with thermogravimetric analysis. Gan et al. [9] investigated the conversion of MDF waste into chars and activated carbon using chemical activation and thermal carbonization processes (carried out in a carbolite LMF4 muffle furnace). Carbon produced from MDF sawdust presented in general a very low adsorption capacity toward the reactive dye, and the carbon physical characterization revealed that the conventional chemical activation and thermal carbonization process were ineffective in developing a microporous structure. The contribution of this work is the development of a detailed study about the pyrolysis of MDF wastes in a pilot reactor with screw conveyor (auger reactor), that supports the biorefineries concept. 2. Experimental section 2.1. Materials MDF samples used in this work come from several furniture companies. Initially, the MDF wastes received were in the form of chips, and subsequently shredded in a knife mill. Further, the shredded material was submitted to a Tyler sieves series to become a particle size less than 0.21 mm. MDF waste and the pyrolysis products underwent through various characterization tests, which will be identified as follows. 2.2. Characterization analysis Metals were determined using the 3111B and 3030E methods (Standard Methods for Examination of Water and Wastewater). Methodology for the phosphorus analysis is described in [20]. Experiments for phenol quantification were performed according to the C 3550 method (Standard Methods for the Examination of Water and Wastewater). To determine the formaldehyde the 3500-1994 NIOSH method was used. 2.3. Thermal analysis Thermogravimetric analysis (TGA/DTG) was performed using a Shimadzu thermobalance (TG-5050 model) under an inert atmosphere of N2 with a flow rate of 50 mL min1. Experiments were conducted under the following heating rates: 5, 10, 15, 25, 50, 75 and 100 °C min1 and diameter samples of less than 0.21 mm, with an initial mass of approximately 10 mg, from ambient temperature up to 800 °C. 2.4. Fourier Transform Infrared Spectroscopy (FTIR) analysis and Van Soest method FTIR analysis was performed on a Nicolet instrument, Thermo Scientific IS10Model. Samples analyzed by FTIR are originated in the thermobalance experiments. In each experiment about 10 mg of sample (MDF) were prepared under inert atmosphere (N2) at a flow rate of 50 mL min1 and a heating rate of 25 °C min1 up to a final temperature (250, 300, 400 °C), and finally were maintained at this temperature for 2 h. For evaluating the moisture, cellulose, hemicellulose and lignin present in the MDF samples, the Van Soest method [21] was used. This method is based on the fibers amount evaluation in neutral and acid detergents. 2.5. Scanning Electron Microscopy (SEM) SEM was used to analyze the MDF and char samples. Analysis was performed using a JOEL equipment JSM 6060 model, with an accelerating voltage of 10 kV. 2.6. Surface area, pore volume and pore radius Surface area, pore volume and pore radius were determined in a Nova 1200e Analyzer from Quantachrome Instruments. Samples S.D. Ferreira et al. / Energy Conversion and Management 92 (2015) 223–233 were subjected to measurement of nitrogen gas adsorption and desorption at temperature of 77.3 K. Samples under investigation were outgassed under vacuum at 350 °C for a period of 20 h prior to testing. The outgass procedure is carried out in order to remove moisture and volatile matter present in the sample, which can interfere the results. Surface areas were determined by Langmuir isotherms. 225  ’’ is where ‘‘HHVvol’’ is the higher heating value in kJ Nm3 and ‘‘h RP;i the i component’s combustion enthalpy in kJ kmol1. Higher heating value of the fuel gas in weight basis was determined by Eq. (4). HHV weight ¼ HHV v ol q ð4Þ where ‘‘HHVweight’’ is the higher heating value in kJ kg1. 2.7. Proximate and ultimate analysis 2.10. Semi-continuous pilot screw reactor Proximate analysis followed the ASTM/D-7582/2010 Standard, using a LECO Corporation Thermogravimetric Analyzer TGA 701. Ultimate analysis followed the ASTM/D-5373/2008 Standard for detection of carbon, hydrogen and nitrogen by infrared and thermal conductivity of the sample combustion, using a LECO Corporation TruSpec CHN. Sulfur detection by infrared followed the ASTM/ D-4239/2011 Standard, using a LECO Corporation TruSpec S. The higher heating value (HHV) determination followed the ASTM/D5865/2010 Standard, using a calorimetric pump Isoperibol, and the lower heating value (LHV) was calculated. For all these analysis the sample was dried in an oven of lamps with air circulation at temperature from 50 ± 5 °C up to constant weight, and after drying the sample, a milling at 0.21 mm in a ball mill was prepared. 2.8. Gas characterization Pyrolysis gas was collected when the production of fuel gas was constant. A 10 L volume FlexFoilÒ bag (SKC) was used for sampling of fuel gas. The analyses were carried out in a GC Auto System XL Perkin Elmer. The first procedure was done using N2 as carrier gas (flow rate 30 mL min1) and a Porapak Q packed chromatographic column of 3.6 m, and a thermal conductivity detector (TCD) at temperature 180 °C, and the oven temperature of 80 °C. Accordingly, it is possible to identify the following compounds: H2, CO2, CH4, CO + O2, C2H4, C2H6. CO and O2 are grouped into a single peak, i.e., the area corresponds to the sum of these compounds. At second procedure was also used N2 as carrier gas, and a Porapak Q column of 3.6 m with a flame ionization detector (FID) at temperature 70 °C, allowing the identification of following compounds: CH4, C2H4, C2H6, C3 and C4. At third procedure was employed Helium as carrier gas, a molecular 13 sieve and the thermal conductivity detector (TCD) at temperature 100 °C, allowing the identification of the compounds CH4, CO, O2, N2. For all procedures, the pyrolysis gas volume injected into the chromatograph was 0.3 mL. 2.9. Determination of higher heating value (HHV) of fuel gas Molar mass average (M) of fuel gas produced in the pyrolysis experiments was determined from the volumetric analysis (dry basis) of gas samples, and calculated by Eq. (1). M¼ n X vi  Mi ð1Þ i¼1 where ‘‘vi’’ is the molar fraction of the gas, i and ‘‘Mi’’ is the molar mass of the component i (g mol1). The fuel gas density (q) was determined by Eq. (2). p M q ¼ 0 R  T0 ð2Þ where ‘‘p0’’ is reference pressure (101.325 kPa) and ‘‘T0’’ is reference  is the universal gas constant temperature (298.15 K). ‘‘R’’ (8.3144 kJ kmol1 K). The fuel gas higher heating value (HHV) on volumetric basis at 298.15 K was calculated by Eq. (3). HHV v ol ¼ Pn vi  hRP;i qM i¼1 ð3Þ MDF pyrolysis process was carried out in a semi-continuous pilot screw reactor. Feeding of the MDF waste in each experiment ranged between 1.4 and 1.9 kg. These values correspond to the range of mass flow rate from 0.39 g s1 to 0.53 g s1, respectively, considering that the average time total of experiments was of 1 h. Reactor has its main parts made of in AISI 316 stainless steel. Screw conveyor has around 2 m long, 195 mm of diameter and 195 mm of pitch. Externally to the screw conveyor there is a cylindrical annular combustion chamber with internal diameter of 200 mm and outer diameter of 300 mm, in which flows combustion gases from LPG burner mounted at the opposite end to the feeding unit, i.e., the gases circulate externally to the reaction system and in opposite direction to the movement of the biomass which is being pyrolyzed. Burner with two stages of operation is controlled by a supervisory system, and it can be controlled by one of three K-type thermocouples (T1, T2 e T3) installed in contact with the outer wall of the annular combustion chamber. Combustion gases are led out of the reactor by an exhaust system, and externally, the combustion chamber is thermally insulated. The feeding biomass unit (waste silo) is sealed to prevent the entry of air into the reactor, ensuring that the thermal degradation reaction is carried out in the complete absence of oxygen. Near the inlet of the screw conveyor there is a cooling belt where water at ambient temperature circulates to prevent pyrolysis reaction already at the waste entrance. Screw conveyor is driven by a 2.2 kW AC motor coupled to a speed reducer, whose output speed is 14 rpm, reaching less than 1 rpm, while the electric motor is driven by a frequency converter that is operated from 1 to 60 Hz. At low frequencies the electric motor needs to be externally cooled. At the end of the screw conveyor there is an opening in the reactor bottom for discharging the char produced in the reaction. Char discharge is done inside a container (char collector) coupled to the reactor in whose entrance is installed a K-type thermocouple (T4), measuring the temperature closest of the reactor core. The gaseous phase generated in the pyrolysis reaction is conducted to a bio-oil separator that has externally cooling belts, internally a coil-type heat exchanger and in both it circulates water at ambient temperature. Condensing of higher molecular weight components (bio-oil) occurs in the bio-oil separator. Pyrolysis reaction is carried out under slightly negative pressure produced by a centrifugal fan installed on top of the bio-oil separator. Fuel gas is conducted to the centrifugal fan and through a tube to a flare installed within an exhaust hood, where it is burned. A complete scheme of the pyrolysis system used in this work is presented in Fig. 1. Experiments were performed at operational low frequencies of the electric drive motor screw conveyor, in particular, at 1, 2 and 3 Hz at temperatures 450 and 600 °C. Temperatures were monitored at the entrance of the char collector (T4). Solid residence time was determined from cold tests in the reactor, clocking the travel time from the wastes silo to the char collector. Table 1 reported the solid residence time obtained in the cold tests. Accordingly to Demirbas [6], this range of residence time refers to a conventional pyrolysis type. 226 S.D. Ferreira et al. / Energy Conversion and Management 92 (2015) 223–233 Fig. 1. Complete scheme of the pyrolysis system. To each reaction temperatures (450 and 600 °C) three experiments were conducted, one for each solid residence time (1, 2 and 3 Hz), i.e., a total of six experiments were performed. From this moment on the experiments are referenced by the reaction temperature and solid residence time. 2.11. Pyrolysis process yield products With the purpose of obtaining the pyrolysis process yield in terms of solid and liquid (bio-oil) produced from the MDF wastes pyrolyzed, their respective masses were measured. Fuel gas mass was obtained by difference between the mass of MDF waste feeded and the sum of char masses and bio-oil. It is also convenient to estimate separately the thermal conversion efficiency of the MDF pyrolyzed mass in each one of their products (char, bio-oil and fuel gas) on the experiments. Eqs. (5)–(7) report individually the thermal conversion efficiency of the pyrolysis reaction products. gchar ¼ mchar  HHV char mMDF  HHV MDF  gbio-oil ¼   100 ð5Þ   mbio-oil  HHV bio-oil  100 mMDF  HHV MDF ð6Þ gfuel-gas ¼ Frequency (Hz) Time for a complete rotation of the screw conveyor (s) Solid residence time (min) 1 2 3 355 156 94 34 15 9 ð7Þ where, mchar, mbio-oil, and mfuel-gas are char, bio-oil and fuel gas masses collected in the end of each experiment, respectively, both in kg; mMDF is the MDF mass pyrolyzed in kg; and HHVchar, HHVbio-oil, HHVfuel-gas, and HHVMDF, their respective High Heating Values. 2.12. Carbon closure on pyrolysis reaction Carbon closure is assessed by carbon balance Eq. (8). mc-MDF ¼ mc-char þ mc-bio-oil þ mc-fuel gas ¼ mc-products mMDF  yc-MDF ¼ mchar  yc-char þ mbio-oil  yc-bio-oil þ n   X yc-chemical speciei  ychemical speciei þmfuel gas  ð8Þ i¼1 where mc-MDF is the mass of carbon in the waste of MDF pyrolyzed (kg), mc-char is the mass of carbon in the char produced (kg), mc-bio-oil is the mass of carbon in the bio-oil (kg), mc-fuel-gas is the mass of carbon in the fuel gas produced (kg), mc-products is the mass of carbon in the products (kg), and yc is the mass fraction of carbon in the materials and in the chemical species involved in the analysis. The percentage difference in the carbon closure in the pyrolysis reaction is determined by Eq. (9). Diff :%c ¼ Table 1 Times involved in the pyrolysis experiments.   mfuel-gas  HHV fuel-gas  100 mMDF  HHV MDF   mc-MDF  mc-products  100 mc-MDF ð9Þ 3. Results and discussion Results of proximate, ultimate, cellulose, hemicelluloses, lignin and calorific value analysis of MDF wastes are presented in Table 2. S.D. Ferreira et al. / Energy Conversion and Management 92 (2015) 223–233 As noted in the values identified in Table 2, the MDF becomes attractive for the pyrolysis process due to its low ash and high volatile matter content. Low sulfur concentration minimizes the release of reduced sulfur compounds, while the low concentration of chlorine reduces the production of acid gas (HCl). Ultimate analysis indicated an oxygen/carbon mole ratio (O/C) of 0.52 and a hydrogen/carbon mole ratio (H/C) of 1.46, typical values to biomass (van Krevelen diagram) [22]. Biomasses are composed mainly of cellulose, hemicelluloses and lignin [23]. Variations in chemical composition can be attributed to variations between species, although there is significant variation in the same species due to their genetic and ecological growth conditions. Hemicelluloses to lignin ratio and the cellulose to lignin ratio were 1.61 and 3.46, respectively. Cellulose to lignin ratio shows that MDF sample is rich in cellulose. Table 3 presented the MDF waste characterization (concentrations of metals, minerals, phenols and formaldehydes). The concentration of metals, minerals and phenols in the MDF are similar to the results obtained by other authors [9,19]. The presence of alkali content in the biomass (MDF waste) which plays a catalytic role in its thermal decomposition [24]. The literature presents the concentration of urea formaldehyde resin around 10% [9]. In this work, the analyzed material had a formaldehyde concentration near to 0.2%. Fig. 2(a) and (b) show the images obtained from SEM with an increase of 200 and 100 of the MDF wastes. Analyzing these images it is observed the presence of irregularly fibers distributed along the sample. curves for different temperatures: (a) at 20 °C; (b) 250 °C; (c) 300 °C; and (d) at 400 °C. FTIR spectra at 20 °C showed the band at 3430 cm1, corresponding the OAH stretching vibration. At 250 °C there was a widening of OAH, possibly due to the increase in carboxylic acids, due to primary OAH oxidation and or hydrolysis of acetyl groups from hemicelluloses. At temperatures between 300 and 400 °C OAH stretching vibration decreases significantly. CAH bond stretching of aliphatic groups was observed between 2800 and 3000 cm1 at temperatures lower than 250 °C [25]. At higher temperatures it is not observed the presence of aliphatic CH group, indicating its thermal degradation. For lignin, the simultaneous decrease of the intensities aromatic CAH bands (1060 cm1) point out that there is a partial dehydration [26]. In All ranges of temperature evaluated it is possible to observe the band (1650 cm1) [27,28] correspond to C@C vibration of the aromatic skeleton of lignin indicating the presence of residual lignin, confirming the stability of aromatic ring. At temperature of 20 °C an intense peak (1230 cm1) CO bond (cellulose) was observed. The cellulose degradation occurs approximately between 225 and 375 °C. At higher temperatures (300 and 400 °C), a broadening is observed, indicating the breaking of this bond. Fig. 4 presents the thermal analysis (TGA/DTG) of MDF waste obtained from different heating rates (5, 10, 15, 25 e 50, 75, 100 °C min1) in the temperature range between 20 and 800 °C in inert atmosphere. TGA Fig. 4(a) has three distinct regions at all heating rates evaluated: First region which is between ambient temperature to 3.1. Thermal MDF analysis Fig. 3 shows the FTIR spectra for the samples originated in the TGA, accordingly to Section 2.4. In this figure are presented the Table 2 Proximate, ultimate and calorific value analysis. a MDF waste (wt.% on dry basis) Moisture Volatile Fixed carbon Ashes C H N Oa S Cl HHV (kJ kg1) LHV (kJ kg1) 7.26 78.30 21.17 0.53 52.68 6.43 3.26 37.46 0.03 0.14 19,005 17,623 Van Soest method Cellulose Hemicelluloses Lignin 57.0 26.5 16.5 By difference. Table 3 MDF waste characterization. Metal/mineral Value Al (g kg1) Ca (g kg1) Fe (g kg1) Mg (g kg1) Na (g kg1) K (g kg1) SiO2 (g kg1) Phenol (mg kg1) Formaldehyde (mg kg1) 0.0794 0.806 0.077 0.1854 10.17 3.19 0.27 <1.58 1681 227 Fig. 2. SEM images of the MDF waste (a) 200 and (b) 100. 228 S.D. Ferreira et al. / Energy Conversion and Management 92 (2015) 223–233 Fig. 3. MDF waste at different temperatures (FTIR): (a) 20 °C; (b) 250 °C; (c) 300 °C; and (d) at 400 °C. 200 °C, representing the volatilization of water and extractives present in wood; Second region in which occurs the releasing of volatile substances (between 225 and 380 °C); and the Third region (temperatures above 380 °C) at which occur reactions involving carbonaceous product (char) of the pyrolysis reaction [29]. Previous studies [23,25,26] indicated that the DTG curve of biomass has two characteristic peaks in the main region of devolatilization: one peak and one ‘‘shoulder’’ peak on the left side of the single peak. The shoulder peak on the left side corresponds to the decomposition of hemicellulose, while at higher temperatures, a single peak is the decomposition of cellulose. Because the ranges of cellulose and hemicellulose decomposition partially overlap, hemicellulose decomposition appears usually as a shoulder more or less pronounced rather than a sharp peak. Absence of shoulder peak (Fig. 4(b)) may be associated with high cellulose to hemicellulose ratio presented in the MDF waste. In the DTG curves it is observed the temperatures where the reaction rates are the highest. Based on Fig. 4(b), the region comprised within the range from 200 to 500 °C is that in which occurs the greatest mass loss. It can be noted that when there is an increase in the heating rate the thermal decomposition occurs at higher temperatures. Table 4 reports the temperature (Tmax) where the reaction rate is maximum for all heating rates tested. Maximum temperature (Table 4) of each heating rate has one extensive peak, they occurred between 327 and 371 °C. Maximum peaks were attributed to the decomposition of cellulose and hemicellulose. Higher heating rate was found to shift the DTG curve to a greater range of temperature. This was due to the fact that when the heating rate was increased, the sample retention time was shorter and the temperature required for organic matter to decompose was greater [30–33], causing the maximum temperature to move rightwards. These behavior was similar to those reported by Wang et al. [32] and Jeguirim and Trouve [31]. 3.2. Results in the pyrolysis screw reactor Product yields (char/bio-oil/fuel gas) are presented in Fig. 5. To determine the bio-oil yield, two phases were considered (oil phase and aqueous phase). From the mass balance, the fuel gas yield was obtained by difference. In Fig. 5(a) regarding the temperature of 450 °C, it is observed that with the increasing of solid residence time it increased the char yield from 24.9 to 39.7 (wt.%), while the bio-oil yield remains relatively constant. In Fig. 5(b) related to the temperature of 600 °C, there is a significant behavior change product yield, i.e., it increased the residence time, decreased the char yield (25.5–17.3 wt.%) and increased the bio-oil yield (23.9–40.0 wt.%). The solids residence time are the same in both experiments (15 min), the highest bio oil yield as well as char yield reduction, are associated with higher heating rate used in the experiment 600-15. A similar tendency was reported for yellow poplar wood waste at temperatures of 450 and 550 °C, where the increasing of temperature results in decreasing of char, while the bio-oil production grows [34]. Increased production of bio-oil at 600 °C is attributed to release of lignin, oligomers and hydrosugar from biomass solid matrix. At lower temperatures, these heavy compounds cannot be released [34]. Table 4 Maximum temperature. Fig. 4. Thermal analysis (TGA) figure (a) and DTG figure (b) of MDF waste under different heating rates. Heating rate (°C min1) Tmax (°C) 5 10 15 25 50 75 100 327.0 340.1 348.9 358.2 365.7 368.3 370.2 S.D. Ferreira et al. / Energy Conversion and Management 92 (2015) 223–233 229 Fig. 6. SEM images of the char produced at (a) 450-15 and (b) 600-15 (both 200). Fig. 5. Effect of the solid residence time on product yield: (a) 450 °C and (b) 600 °C. Fig. 6 shows the SEM of char generated in the experiments at 450 °C (a) and 600 °C (b) in 15 min residence time. Char produced in the screw reactor showed similar appearance in both experiments. Comparison between the original sample of MDF waste (Fig. 2) and pyrolyzed samples (Fig. 6) shows different morphology and structural changes after devolatilization (pyrolysis) at different temperatures. Char showed a broken fiber structure, unlike Fig. 2 that shows the fibers interwoven and well structured. Fig. 7 illustrates the SEM image of char generated in the experiment 600-15. In the experiment 600-15 it is possible to identify pores in the external surface, which are usually formed. Generally, the pore structure of solid materials can be divided into three classes: micropores (pore size smaller than 2 nm), mesopores (pore diameter between 2 and 50 nm) and macropores (pores wider than 50 nm) [35]. One of the methods for estimating the pores structure is by analyzing the adsorption/desorption isotherms. The amount of adsorbed nitrogen is an indicative of the char adsorptive capacity [36]. Fig. 8(a) and (b) show the adsorption/desorption isotherms for the char produced in experiments 450-15 and 600-15, respectively. At both experiments the isotherms may be considered type IV. Type IV isotherm are its hysteresis loop, which is associated with capillary condensation taking place in mesopores. At both experiments hysteresis persists down to very low pressures, although Fig. 7. SEM image of the char produced in the experiment 600-15 (2000). hysteresis in experiment 450-15 is more noticeable. Hysteresis in low pressures may be associated with the change in volume of the adsorbent, for example, the swelling of non-rigid pores or the irreversible uptake of molecules in pores of about the same width as that of the adsorptive molecule [37]. Hysteresis more pronounced at experiment 450-15 is an indicative of pores interconnectivity. In such systems, the distribution of pore sizes and the pore shape is not well-defined or irregular. If a pore is connected to the external vapor phase via a smaller pore, in many cases the 230 S.D. Ferreira et al. / Energy Conversion and Management 92 (2015) 223–233 Fig. 9. Pores distribution by BJH method. Fig. 8. (a) Adsorption/desorption isotherms for nitrogen at 77.3 K for the char produced at 450-15 and (b) char produced at 600-15. smaller pore acts as a neck, often referred to as an ’’ink-bottle’’ pore, hindering the desorption of the adsorbed molecules [38]. Fig. 9 shows the pores distribution by BJH method. It was observed that in both experiments (450-15 and 600-15) the unimodal pore formation occurred. Char produced in experiment 450-15 has only pores with radius smaller than 40 Å, while char produced in the experiment 600-15 also have pores larger than 40 Å. An increase in the temperature may have caused a collapse of pores which are close together, causing an increase in pore size. Surface area obtained by Langmuir isotherm were 415.4 and 593 m2 g1 for the experiments 450-15 and 600-15, respectively. This values of surface area are surprisingly high, and compatibles with the surface area of the commercial activated carbons. The commercial activated carbons typically has surface area between 300 and 1200 m2 g1 [39]. Gan and co-authors [9] evaluated the surface area and pore volume of char produced by carbonization at 550 °C for 30 min of untreated MDF and MDF treated with 85% H3PO4. The values found in surface area were 0.75 m2 g1 (untreated MDF) and 0.57 m2 g1 (treated with acid). Authors found there was no apparent presence of pores in the SEM images, and thus, through the BET (Brunauer– Emmett–Teller) analysis, the presence of pores with diameter greater than 50 nm was observed. Therefore, the authors concluded that the presence of urea–formaldehyde resin in the material may have avoided the porous structure proper development. Difference in surface area and pore size results is probably related to experimental conditions, mainly due to the type of reactor used by the authors and it is not associated to the presence of urea– formaldehyde resin. According to Apaydin-Varol and Putun [40] char is a complex material composed mainly of carbon fixed, volatile hydrocarbons, inorganic materials and a small amount of ash, favoring its use in thermal processes. Fig. 10 shows the behavior of volatile matter and fixed carbon present in the MDF and different chars in function of pyrolysis temperature (450-15 and 600-15). The volatile matter initially present in the char decreases significantly with increasing pyrolysis temperature, while there is an increase of fixed carbon, as expected. Fig. 10 shows that for 450-15 experiment the mass fraction of fixed carbon is about 57 (wt.%) and the mass fraction of volatile matter is approximately 37 (wt.%), while for 600-15 experiment there is an increase of carbon fixed mass fraction (78 wt.%) and a decrease of volatile matter (19 wt.%). This behavior occurred due to the release of a larger amount of volatile matter by an increase in the reaction temperature. Ultimate analysis (Fig. 11) shows that with the increasing of the reaction temperature the carbon content increases in the char, while the hydrogen and oxygen content decreases. A similar trend was reported by Kim et al. [41], Imam and Capareda [42], ApaydinVarol and Putun [40] and López et al. [43]. According to Imam and Fig. 10. Behavior of volatile matter and fixed carbon according to the pyrolysis temperature. 231 S.D. Ferreira et al. / Energy Conversion and Management 92 (2015) 223–233 Capareda [42], hydrogen losses are explained by breaking weaker bonds within the char structure making it highly carbonaceous at higher pyrolysis temperatures. Fig. 11 shows the behavior of the H/C and O/C (mol/mol) ratios in relation to reaction temperature. With the increasing of the reaction temperature, there is a significant decrease in both ratios, while the calorific value of the char increases with the increasing of the reaction temperature (29,915 kJ kg1 at 450-15 and 32,800 kJ kg1 at 600-15). Char produced in the experiment 600-15 have ratios (O/C and H/C) similar to anthracite (van Krevelen diagram) [22]. Bio-oil generated in the MDF pyrolysis (450 and 600 °C) had an appearance and color similar in all experiments. Separation of water was carried out using a rotary vacuum evaporator, and the average values in the experiments 450-15 and 600-15 were about 23.0 and 30.0 wt.%, respectively. Fig. 12 shows the bio-oil FTIR: (a) oil phase (600-15); (b) aqueous phase (450-15); and (c) aqueous phase (600-15). According to some authors [44,45], FTIR spectra obtained for the bio-oil produced showed the following bands: from 3300 to 3400 cm1 corresponds the axial vibrations of the OAH bond, attributed to the presence of phenols, alcohols, carboxylic and mainly the presence of water. The band (Fig. 12) to 2930 is related to the CAH bond of the aliphatic and aromatic groups. This band is present only in the oil phase. The band between 1600 and 1750 cm1 is attributed to C@O bond, confirming the presence of carbonyl, ketones, aldehydes and esters; at 1466 cm1, which is related to the deformations of the CAH bond, indicating the presence of groups CH3, CH2 and CH; at 1150 cm1 corresponding to angular deformation of the CAO bond of alcohols, phenols, esters and to axial vibrations of the skeleton (CAC bond). The absorption band around 1050 cm1 indicates the presence of lignin. Table 5 reports the ultimate analysis and heating value of biooil in oil phase (600-15). Characterization bio-oil is very similar to those found in the literature [40,45]. Table 6 presents the fuel gas composition obtained by gas chromatography (600–34; 600-15). In Table 6 it is observed a moderate change in the volumetric fuel gas composition between the experiments 600–34 and 600-15: CO, H2 and C2H4 decreased with the decreasing of solid residence time, while CH4, CO2 and C4H8 increased with the decreasing of solid residence time. Results reported by Imam and Capareda [42] for the fuel gas produced from the biomasses pyrolysis at 600 °C (%vol: H2 = 9.7; CO = 27.7; CH4 = 17.6; CO2 = 33.2; C2H4 = 4.3; and C2H6 = 7.0) were similar to those obtained in this work. Fig. 12. FTIR bio-oil analyzes (a) oil phase (600-15), (b) aqueous phase (450-15) and (c) aqueous phase (600-15). Table 5 Ultimate bio-oil analysis (wt.% – on db) and heating value. Element Value Carbon Hydrogen Nitrogen Sulfur Oxygen + halogens + ash HHV (kJ kg1) 60.97 7.61 5.87 0.08 25.47 26,963 Table 6 Fuel gas composition (vol.% – on db). Experiment Vol.% 600-34 600-15 H2 CO CH4 CO2 C2H4 C2H6 C4H8 8 7 51 41 7 10 23 32 6 4 4 4 1 2 Fuel gas generated in the pyrolysis process was burned in the flare. Based on the fuel gas composition (Table 6), the average values of molecular mass, density and higher heating value were determined at 25 °C and 101.325 kPa, whose values are presented in Table 7. Within the range studied is possible to conclude that the solid residence time does not influence significantly in the fuel gas proprieties. Thermal conversion efficiency of the MDF pyrolyzed mass in each one of their products (char, bio-oil and fuel gas) of the experiments 600–39 was estimated using Eqs. (5)–(7), and the results are shown in Table 8, in which are also shown the values obtained by Stals et al. [20] in the fast pyrolysis of hardwood waste at reaction temperature of 550 °C. In particular, the energy yield of fuel gas obtained in this study is a bit lower than to the yield found in the literature about gasification process. A reference to that yield is found in the work of Table 7 Fuel gas proprieties. Fig. 11. Behavior of the H/C and O/C ratios in relation to reaction temperature. 600-34 600-15 MM (g mol1) q (kg m3) HHV (kJ m3) HHV (kJ kg1) 29.62 31.24 1.2107 1.2770 17,320 17,095 14,305.2 13,386.8 232 S.D. Ferreira et al. / Energy Conversion and Management 92 (2015) 223–233 Table 8 Thermal efficiency of the pyrolysis reaction products. Thermal efficiency Present article (%) Stals et al. [20] (%) gchar gbio-oil gfuel-gas 35.5 29.2 41.4 17 32 51 Finally, the pyrolysis products (char/bio-oil), which have a market value, can be exploited for energy purposes and for the chemical industry (Biorefineries). References Table 9 Carbon closure for the experiments 600-15. mc-MDF (kg) mc-char (kg) mc-bio-oil (kg) mc-fuel-gas (kg) mc-products (kg) Difference (%) 0.537 0.179 0.128 0.256 0.563 4.8 Wander et al. [46] for the pine sawdust gasification process, whose average value was around 60%. It was assumed that the mass balance in all experiments was closed due to the mass of fuel gas produced in each experiment determined by difference between the MDF pyrolyzed mass and the sum of the masses of char and bio-oil, afterward, the carbon closure is assessed accordingly to Eq. (8). It is demonstrated by the experiment 600-15 from the characterization carried out for all materials involved in the pyrolysis reaction. Thus, the carbon masses were found in the MDF pyrolyzed, char, bio-oil and also in the fuel gas. For the experiments 600-15, the carbon balance results [Eq. (8)] and the percentage difference [Eq. (9)] are presented in Table 9. In Eq. (8), the term related to the carbon mass on fuel gas considers n = 6 (see Table 6). The carbon closure did not close, as it can be observed in Table 9 and it is attributed to the several uncertainties in the measurements made (mass and compositional analysis of the materials involved in this study). 4. Conclusions Through MDF characterization it can be observed similar characteristics of biomasses, with the main difference being the presence of the urea–formaldehyde resin. Char yield in the experiments carried out the 450 °C was higher in longer residence times (24.4–39.7 wt.%), while the 600 °C there was a decrease in char yield with increasing residence time (25.5– 17.3 wt.%). Bio-oil yield in the experiments carried out the 450 °C was relatively constant with increasing residence time (about 27 wt.%), while the 600 °C had an increase in the bio-oil production (23.9– 40 wt.%). Fuel gas yield ranging from 34.6% to 50.7% wt.%, was the highest yield obtained at higher temperatures. In the gas chromatography it was identified the presence of the main fuel chemical species in the fuel gas produced, resulting in the average higher heating value of 13,387 kJ kg1 in the experiment 600-15 and 14,305 kJ kg1in the experiment 600–34. Char calorific value found was 29,915 kJ kg1 (450-15) and 32,800 kJ kg1 (600-15), while the bio-oil calorific value was 26,963 kJ kg1 (600-15). The values of superficial area in both samples (450-15 and 600-15) were surprisingly high, 415 and 593 m2 g1, respectively, which are compatibles to the superficial area of commercial activated carbons. Auger reactor presented performance adequate for pyrolysis of MDF waste, and became an alternative for final disposition for these wastes. [1] Naik SN, Goud VV, Rout PK, Jacobson K, Dalai AK. Production of first and second generation biofuels: a comprehensive review. Renew Sustain Energy Rev 2010;35:1624–31. [2] Cherubini F. The biorefinery concept: using biomass instead of oil for producing energy and chemicals. Energy Convers Manage 2010;51:1412–21. [3] Demirbas A. Liquefaction of biomass using glycerol. Energy Sources, Part A 2008;6:30–1120. [4] Gonzalez JD, Kim MR, Buonomo EL, Boneli PR, Cukierman AL. Pyrolysis of biomass from sustainable energy plantations: effect of mineral matter reduction on kinetics and charcoal pore structure. Energy Sources, Part A 2008:30–809. [5] Onay O, Kockar OM. Fast and flash pyrolysis of rapeseed. Renew Energy 2003;28:2417–33. [6] Demirbas A. Biorefineries: current activities and future developments. Energy Convers Manage 2009;50:2782–801. [7] Abnisa F, Daud WMW. A review on co-pyrolysis of biomass: an optional technique to obtain a high-grade pyrolysis oil. Energy Convers Manage 2014;87:71–85. [8] Hillig EÉ, Schneider VE, Pavoni ET. Generation of wood waste and derivatives on furniture industry related to production’s variables. Produção 2009;19:292–303. [9] Gan Q, Allen SJ, Matthews R. Activation of waste MDF sawdust charcoal and its reactive dye adsorption characteristics. Waste Manage 2004;24:841–8. [10] Jahirul MI, Rasul MG, Chowdhury AA, Ashwath N. Biofuels production through biomass pyrolysis a technological review. Energies 2012;5:4952–5001. [11] Agirre I, Griessacher T, Rösler G, Antrekowitsch J. Production of charcoal as an alternative reducing agent from agricultural residues using a semi-continuous semi-pilot scale pyrolysis screw reactor. Fuel Process Technol 2013;106:114–21. [12] Brown JN, Brown RC. Process optimization of an auger pyrolyzer with heat carrier using response surface methodology. Bioresour Technol 2012;103:405–14. [13] Puy N, Murillo R, Navarro MV, Lópes JM, Rieradevall J, Aranguren FI. Valorisation of forestry waste by pyrolysis in an auger reactor. Waste Manage 2011;31:1339–49. [14] Liaw S, Zhou S, Wu H, Garcia-Perez M. Effect of pretreatment temperature on the yield and properties of bio-oils obtained from the auger pyrolysis of Douglas fir wood. Fuel 2013;103:672–82. [15] Elliott D. Advancement of bio-oil utilization for refinery feedstock. Presented at the Washington bioenergy research symposium; November 8, 2010. <http://www.pacificbiomass.org/documents/Elliott%20(C1).pdf>. [16] Strezov V, Evans TJ, Hayman C. Thermal conversion of elephant grass (Pennisetum Purpureum Schum) to bio-gas, bio-oil and charcoal. Bioresour Technol 2008;99:8394–9. [17] Lehmann J, Joseph S. Biochar for environmental management: science and technology. Sterling: Earthscan; 2009. p. 1–12. [18] Grierson S, Strezov V, Shah P. Properties of oil and char derived from slow pyrolysis of Tetraselmis chui. Bioresour Technol 2011;102:8232–40. [19] Yorulmaz S, Atimtay A. Investigation of combustion kinetics of treated and untreated waste wood samples with thermogravimetric analysis. Fuel Process Technol 2009;90:939–46. [20] Stals M, Carleer R, Reggers G, Schreurs S, Yperman J. Flash pyrolysis of heavy metal contaminated hardwoods from phytoremediation: characterisation of biomass, pyrolysis oil and char/ash fraction. J Anal Appl Pyrol 2010;89:22–9. [21] Van Soest PJ, Wine RH. Determination of lignin and cellulose in acid detergent fiber with permanganate. J Assoc Off Agric Chem 1968;51:780–5. [22] Van Krevelen DW. Typology, physics, chemistry, constitution. 3rd ed. Amsterdam; 1961. [23] Thanasit W, Nakorn T. Non isothermal characteristics of giant sensitive plants using thermogravimetric analysis. Bioresour Technol 2010;101:5638–44. [24] Nowakowski DJ, Jones JM, Brydson RMD. Potassium catalysis in the pyrolysis behaviour of short rotation willow coppice. Fuel 2007;86:2389–402. [25] Nasir M, Gupta A, Beg MDH, Chua GK, Kumar A. Fabrication of medium density fibreboard from enzyme treated rubber wood (Hevea brasiliensis) fibre and modified organosolv lignin. Int J Adhes Adhes 2013;44:99–104. [26] Bilba K, Quensanga A. Fourier transform infrared spectroscopic study of thermal degradation of sugar cane bagasse. J Anal Appl Pyrol 1996;38:61–73. [27] Esteves B, Marques AV, Domingos I, Pereira M. Chemical changes of heat treated pine and eucalypt wood monitored by FTIR. Food Sci Technol 2013;5:1–11. [28] Weiland JJ, Guyonnet R. Study of chemical modifications and fungi degradation of thermally-modified wood using DRIFT spectroscopy. Eur J Wood Wood Prod 2003;61:216–20. [29] Kercher AK, Nagle DC. TGA modeling of the thermal decomposition of CCA treated lumber waste. Wood Sci Technol 2001;35:325–41. [30] Ferdous D, Dalai AK, Bej SK, Thring RW. Pyrolysis of lignin: experimental and kinetic studies. Energy Fuels 2002;16:1405–12. S.D. Ferreira et al. / Energy Conversion and Management 92 (2015) 223–233 [31] Jeguirim M, Trouve G. Pyrolysis characteristics and kinetics of Arundo donax using thermogravimetric analysis. Bioresour Technol 2009;100:4026–31. [32] Wang G, Li W, Li B, Chen H. TG study on pyrolysis of biomass and its three components under syngas. Fuel 2008;87:552–8. [33] Hu S, Jess A, Xu M. Kinetic study of Chinese biomass slow pyrolysis: comparison of different kinetic models. Fuel 2007;86:2778–88. [34] Kim KH, Eom IY, Lee SM, Choi D, Yeoa H, Choi G. Investigation of physicochemical properties of biooils produced from yellow poplar wood (Liriodendron tulipifera) at various temperatures and residence times. J Anal Appl Pyrol 2011;92:2–9. [35] Guo S, Peng J, Li W, Yang K, Zhang L, Zhang S, et al. Effects of CO2 activation on porous structures of coconut shell-based activated carbons. Appl Surf Sci 2009;225:8443–9. [36] Fu P, Hu S, Xiang J, Yi W, Bai X, Sun L. Evolution of char structure during steam gasification of the chars produced from rapid pyrolysis of rice husk. Bioresour Technol 2012;114:691–7. [37] Zeid A, Alothman A. Review: fundamental aspects of silicate mesoporous materials. Materials 2012;5:2874–902. [38] Naumov S. Hysteresis phenomena in mesoporous materials (Master Dissertation). Universitat Leipzig; 2009. 95 pages. 233 [39] Geankoplis CJ. Transport processes and separation process principles: (includes unit operations). New Jersey: Prentice Hall; 2003. [40] Apaydin-Varol E, Putun E. Preparation and characterization of pyrolytic chars from different biomass samples. J Anal Appl Pyrol 2012;98:29–36. [41] Kim K, Kim HK, Su TC, Weon JC. Influence of pyrolysis temperature on physicochemical properties of biochar obtained from the fast pyrolysis of pitch pine (Pinus rigida). Bioresour Technol 2012;118:158–62. J.Y.. [42] Imam T, Capareda S. Characterization of bio-oil, syn-gas and bio-char from switch grass pyrolysis at various temperatures. J Anal Appl Pyrol 2012;93:170–7. [43] López AF, Centeno AT, Díaz IG, Alglacil FJ. Textural and fuel characteristics of the chars produced by the pyrolysis of waste wood and the properties of activated carbons prepared from them. J Anal Appl Pyrol 2013;104:551–8. [44] Jiang X, Ellis N, Shen D, Jiang J, Dai W, Zhong Z. Thermogravimetry-FTIR analysis of pyrolysis of pyrolytic lignin extracted from bio-oil. Chem Eng Technol 2012;35:827–33. [45] Zhang Y, Chen D, Zhang D, Zhu X. TG-FTIR analysis of bio-oil and its pyrolysis/ gasification property. J Fuel Chem Technol 2012;40:1194–9. [46] Wander PR, Altafini CR, Barreto RM. Assessment of a small sawdust gasification unit. Biomass Bioenergy 2004;27:467–76.