Kinetic study of the pyrolysis of pine cone shell through non-isothermal thermogravimetry: Effect of heavy metals incorporated by biosorption

Renewable Energy 96 (2016) 613e624

Contents lists available at ScienceDirect

Renewable Energy journal homepage: www.elsevier.com/locate/renene

Kinetic study of the pyrolysis of pine cone shell through nonisothermal thermogravimetry: Effect of heavy metals incorporated by biosorption
zquez, A. Ronda, M. Calero M.A. Martín-Lara*, G. Bla Department of Chemical Engineering, University of Granada, 18071 Granada, Spain

a r t i c l e i n f o

a b s t r a c t

Article history: Received 15 July 2015 Received in revised form 3 May 2016 Accepted 6 May 2016

The study concerns the pyrolysis kinetics of exhausted pine cone shell after its use as biosorbent of copper and lead from aqueous solutions in a fixed bed column. First, breakthrough curves of biosorption process were obtained. Main dynamic biosorption parameters were determined and analyzed. Then, non-isothermal thermogravimetric experiments were carried out with raw and metal-loaded biomass in a thermobalance under nitrogen atmosphere at different heating rates. A comparative study was performed. The activation energy dependent on the conversion rate was estimated by FlynneWalleOzawa (a free integral or iso-conventional method) and a mechanistic model (an integral or model-fitting method that considers three independent parallel reactions). The fluctuation of activation energy in Flynn-Wall Ozawa model can be considered the result of thermal degradation reactions of different pseudo-components of the lignocellulosic material (hemicellulose, cellulose and lignin). For raw pine cone shell and metal-loaded-pine cone shell, best fit parameters were determined according to a three independent parallel reactions scheme. The copper and lead present in metal-loaded samples did not modify values of determined parameters which describe the pyrolysis process. Finally, chemical analysis of the chars indicated that about 95% and 99% of copper and lead presented on original waste was recovered in generated chars. © 2016 Elsevier Ltd. All rights reserved.

Keywords: Biosorption Heavy metals Kinetics Pine cone shell Pyrolysis Thermogravimetric analysis

1. Introduction Solid waste treatment techniques such as thermal treatment with energy recovery could be applied as an alternative for the final disposal of the exhausted biosorbent. This is an essential issue not only from a waste treatment perspective but also from the use of alternative fuel to energy perspective. Nowadays, different countries are addressing the need to achieve energy security for energy generation in the future. In addition, global warming is already limiting the use of fossil fuels. Concretely, the EU aims to get 20% of its energy from renewable sources by 2020. Therefore, renewable energies have been deeply developed in recent years and are currently under investigation. Among these, biomass has a great potential to contribute to get energy from renewable resources. At present, an extensive research is carrying out for converting biomass into both energy and chemicals [1e5].

* Corresponding author. E-mail addresses: [email protected] (M.A. Martín-Lara), [email protected]
zquez), [email protected] (A. Ronda), [email protected] (M. Calero). (G. Bla http://dx.doi.org/10.1016/j.renene.2016.05.026 0960-1481/© 2016 Elsevier Ltd. All rights reserved.

Some of the challenges for the industrial application of biosorption process are the final disposition of the exhausted biosorbent and pollutant desorption (including recovery of loaded pollutants when they are valuable and simultaneous regeneration of the solid material for recycle). The utility of a particular biosorbent depends on its biosorption capacity (higher biosorption capacity, better biosorbent), but also on its possibility of easy regeneration and recycling, its availability and requirement or not of pre-processing and so on [6]. However, most researchers have evaluated only the biosorption capacity of studied biosorbent, some of them have considered its possibility of regeneration but very few authors have studied the disposition of the exhausted biosorbent required for industrial applications [7]. The methods used for desorption and recovery of loaded pollutants can be destructive or non-destructive, and the selection of the most appropriate method depends on factors such as the intensity of the pollutant binding, the possibility of recovery, and reuse of the pollutant, the loss of sorption capacity of the sorbent material, and the total cost of the operation. Generally, if cheap biomass is used as biosorbent, destructive recovery of valuable pollutants would be economically

614

M.A. Martín-Lara et al. / Renewable Energy 96 (2016) 613e624

feasible [8]. To the use of the exhausted biosorbent as an energy recovery the role of metal during the thermal treatment is a very important issue. Thus, although several researchers studied pyrolysis of pine cone shell [9e11], there are not studies on the effects of Cu(II) and Pb(II) on the mechanism of thermal decomposition of pine cone shell in an inert environment. The results of this work may be useful for the design and operation of pyrolysis units fed by lignocelullosic wastes even after its use as biosorbent of heavy metals in a fixed bed column. Pyrolysis technology is receiving great attention from the scientific community, because by the pyrolysis process at fast heating rate, the biomass can be converted into an easily-stored and easilytransportable liquid fuel (bio-oil) of high energy content. Moreover, pyrolysis is an important substage of gasification technology influencing the producer gas for further electricity supply [12,13]. In this paper a preliminary study on the possible use of exhausted biosorbent in pyrolysis reactors is presented, comparing the kinetics of thermal decomposition in an inert atmosphere of native and exhausted pine cone shell. Concretely, pyrolysis of the exhausted biosorbent is investigated in this work through nonisothermal themogravimetric experiments.

in up-flow mode. Biosorption tests were performed in duplicate at best operational conditions which were found by performing a factorial design changing bed depth, flow rate and inlet metal concentration. Table 1 presents operational data of the laboratory scale fixed-bed column and a schematic illustration about the general experimental setup is presented on Supplementary Material. Column effluent samples were collected at frequent time intervals and analyzed for effluent metal concentration. The Cu(II) and Pb(II) in the residual solution were analyzed in an Atomic Absorption Spectrometer (Perkin-Elmer, model AAnalyst 200).

2.3. Pyrolysis experiments 2.3.1. Thermogravimetric experiments Runs for pyrolysis were carried out on a Perkin Elmer thermobalance model STA 6000. Dynamic experiments were carried out in duplicate, under heating rates of 5, 10 and 20 ºC/min, from 30  C (303 K) up to 800  C (1073 K), including in this way the entire range of decomposition. The flow rate of the carrier gas (N2; purity  99.5%) was 20 mL/min and the sample had a weight of approximately 40 mg.

2. Materials and methods 2.1. Raw material The samples used in this study were pine cone shells (PCS) of Mediterranean pine (Pinus halepensis) supplied by a factory of Granada (Spain), Carsan Biocombustibles S.L. Previous to the runs; the samples were cut into smaller particles and sieved into fractions below 1 mm to be used in the biosorption tests and thermogravimetric analysis following EN 14780:2012 standard. 2.2. Biosorption experiments 2.2.1. Preparation of Cu(II) and Pb(II) solutions Metal solutions of Cu2þ and Pb2þ (initial concentration of 100 mg/L) were prepared by dissolving the necessary amount of nitrate salts (2.28 g of Cu(NO3)2$3H2O and 0.96 g of Pb(NO3)2, respectively) provided by Panreac (Barcelona, Spain) in 6 L of distilled water. 2.2.2. Fixed bed column tests For continuous biosorption experiments, a covering glass column (length of 23 cm, internal diameter of 1.5 cm) was packed with a 15 g of PCS. The metal solution (100 mg/L of copper or lead) was driven at a constant flow rate (6 cm3/min) using a peristaltic pump Table 1 Operational data of the laboratory scale fixed-bed column. Biosorbent mass, g Bed length (Z), cm Column internal diameter (D), cm Cross sectional area of the column (S), cm2 Bed length/Internal diameter Bed volume (Vb), cm3 Biosorbent particles density (rp), g/cm3 Bed porosity (3), g/cm3 Packing density or bulk density (rb), g/cm3 Volumetric flow rate (Q), cm3/min Initial solute concentration (Ci), mg/L Biosorbent particle diameter (dp), mm Superficial velocity (ysuperficial), cm/min Interstitial velocity (yinterstitial), cm/min Residence time (t), min Mode of operation

15 13.4 1.5 1.77 8.93 23.68 1.44 0.560 0.633 6 100 <1 3.40 6.06 2.2 Up-flow

2.3.2. Determination of copper and lead on raw PCS, metal-loaded PCS and chars An energy dispersive X-ray analyzer (PHILIPS Magix Pro (PW2440)) was used for qualitative detection and distribution of elements present in the solid samples (raw PCS, metal-loaded PCS and chars generated in pyrolysis process). The determination of amount of copper and lead presented in chars generated in pyrolysis process was performed following European Standard EN 15297 (Solid biofuels. Determination of minor elements).

3. Theoretical background 3.1. Mathematical description of biosorption in a fixed-bed column The performance of the fixed-bed columns is described using the breakthrough curve concept. The breakthrough curves show the loading behavior of metal to be removed from solution in a fixed bed and are usually expressed in terms of a normalized concentration, defined as the ratio of effluent metal concentration to inlet metal concentration (C/Ci) versus flow time (ttotal) or volume of effluent (Veff) for a given fixed-bed depth [14]. From a practical perspective, service or breakthrough time (tb) is set when the effluent concentration reaches its maximum tolerable level of discharge and exhaustion time (tex) is set when the concentration in the effluent exceeds 90%e95% of the inlet metal concentration. The volume of the effluent at breakthrough, exhaustion or total flow time, can be calculated through the following equation,

Veff ¼ Q $ti

(1)

where Veff is the volume of effluent (mL), ti is the breakthrough, exhausted or total flow time (min) and Q is the volumetric flow rate, (mL/min). The total mass of metal biosorbed is equal to the area under the breakthough curve (plot of the biosorbed Cu(II) or Pb(II) concentration CR (mg/L) versus time (min)), for a defined feed concentration and flow rate is calculated from following equation:

M.A. Martín-Lara et al. / Renewable Energy 96 (2016) 613e624

qtotal

Zt¼ti

Q ¼ 1000

CR dt

(2)

t¼0

where qtotal is the total mass of metal biosorbed (mg) and CR is the concentration of metal removal (mg/L). The total amount of metal ions sent to the column until breakthrough, exhaustion or total flow time, can be calculated from the following equation,

mtotal ¼

Ci $Q $ti 1000

(3)

where mtotal is total amount of metal ions sent to the column (mg) and Ci is the inlet metal concentration (mg/L). The total metal removal (%R) at breakthrough, exhaustion or total flow time can be calculated from the ratio of metal mass biosorbed (qtotal) to the total amount of metal ions sent to the column (mtotal) as follows,

  E kðTÞ ¼ A$exp  R$T

(4)

In the same way as occurs in batch processes, column studies require knowledge of biosorption capacity, qe (mg of sorbated metal/g of sorbent), and on metal concentration Ce (mg/L), which remains in the solution and can be determined by the following expressions,

qe ¼

qtotal m

(5)

Ce ¼

mtotal  qtotal $1000 Veff

(6)

where m represents the mass of the biosorbent (g).

  da da E $f ðaÞ ¼ b$ ¼ A$exp  R$T dt dT

The decomposition reaction of PCS is a typical solid state decomposition reaction. The conversion of PCS can be calculated as

m0  mt m0  m∞

(7)

where a is the conversion degree, m0 is the initial mass of PCS, mt is the mass at time t, and m∞ is the final mass of the sample after the reaction. In non-isothermal experiments, the mass of the sample is recorded as a function of temperature (which varies over time following a heating gradient). According to non-isothermal kinetic theory, the iso-conversional method can be applied to a solid state reaction. Therefore, the reaction rate under linear temperature increasing condition can be described as

da ¼ kðTÞ$f ðaÞ dt

(10)

where b is the heating rate (K/min). Integrating Eq. (10) by separation of variables gives:

gðaÞ ¼

Za

da A ¼ $ f ðaÞ b

ZT T0

  Z∞ E A$E $dT ¼ $ u2 $eu $du exp  R$T b$R x

A$E $PðxÞ ¼ b$R (11) where x ¼ E/RT. Different iso-conversional methods differs by approach carried out for the numerical determination of the function P(x) which is a function that has no an exact solution. The FWO method is one of the models commonly used for the determination of kinetic parameters of thermal decomposition of a material. The FWO method linearizes the temperature integral in Eq. (11) using Doyle's empirical approximation [15,16].

3.3. Kinetic evaluation of thermal decomposition by using a three independent parallel reactions mechanistic model

3.2. Kinetic evaluation of thermal decomposition by using FlynnWall-Ozawa (FWO) method (a model-free method)

a¼

(9)

where A is the pre-exponential or frequency factor (min1), E is the activation energy (kJ/mol) and R is the universal gas constant (kJ/ K$mol). Also, differential kinetic equation can be expressed in function of temperature and, therefore, equation (8) can be re-written as follows:

0

q %R ¼ total $100 mtotal

615

(8)

where t is the reaction time (min), T is the temperature (K), k(T) is the temperature-dependent rate constant, and f(a) is the reaction model which is a function of conversion. In addition, if k(T) follows Arrhenius equation:

In this work, the pyrolysis process is also described by a kinetic scheme consisting of three independent nth-order parallel reactions. In general, if the sample consists of more than one chemical component and each component decomposes independently from the others, a separate reaction can be considered for each component. Therefore, a three independent nth-order parallel reactions model is considered as the most realistic approach in the case of lignocellulosic materials according to different researchers [17,18]. According to this model, the thermal decomposition of biomass is associated with the decomposition of their main pseudocomponents cellulose, hemicellulose and lignin [17,19e21]. The kinetic of pyrolysis is evaluated using the following equations:

si Solidi /ri Chari þ ð1  ri ÞVolatilesi

(12)

where “Solidi” refer to different components of the original material (i ¼ 1e3 cellulose, hemicellulose and lignin), ‘‘Volatilesi’’ are the gases and condensable volatiles developed in the corresponding reactions (i ¼ 1e3), and ‘‘Chari’’ is the char formed in the decomposition of each solid. On the other hand, the “si” represents the initial fraction of solidi on the entire solid material (data reported in Table 1) and the “ri” the yield coefficients of each reaction that represents the maximum mass fraction of char obtainable by the reaction i from solidi. And if an n-th order kinetic decomposition is considered, the kinetic equations can be written as followed:

616

M.A. Martín-Lara et al. / Renewable Energy 96 (2016) 613e624

 d 

wSolid;i si

qffiffiffiffiffiffiffi



dt

  wSolid;i ni ¼ ki $ si

fit ð%Þ ¼ (13)

O:F: N

 $100 max wexp

(17)

where N is the number of experimental points. where wSolid,i, is the mass fraction of the solidi at time t and ni is the reaction order. And expressing the results in terms of conversion (reacted fraction) for each component (ai)

dai ¼ ki $ð1  ai Þni dt

(14)

where the conversion represents the fraction of solidi that was decomposed. Finally, if is considered that kinetic constants following the Arrhenius equation, the Equation (15) was obtained,

  dai E ¼ ki0 $exp  i $ð1  ai Þni dt R$T

(15)

In order to obtain a single set of kinetic parameters capable of reproducing the thermogravimetric curves at different heating rates, all experimental data were fitted simultaneously minimizing the objective function presented in the Eq. (16), using the optimization method of the function Solver of Excel program. The differential equations were integrated numerically using Euler's method. The objective function (O.F.) to minimize can be expressed as:

2 XX   4 wexp  wcal 2 þ O:F: ¼ j

i



3    !2 dw dw 5  dt exp dt cal (16)

where ‘i’ represents the experimental data at time ’t’ in the experiment with a heating rate ’j’, wexp and wcal are the experimental and calculated mass fraction and (dw/dt)exp and (dw/dt)cal are the derivate of experimental and calculated mass fraction respect to time. Finally, to measure the goodness of the fit, the deviation between experimental and calculated curves is defined in accordance with a previous publication of Conesa and Domene [18] as:

4. Results and discussion 4.1. Physical and chemical characteristics The physic-chemical characteristics of raw PCS are given in Table 2. One of the most main properties is density, it is an important characteristic because it determines transport and storage costs [22,23]. Bulk density of PCS is higher than that of wood chips (328 kg/m3 for beech specie or 223 kg/m3 for spruce and fir specie) and comparable to bulk density of wood pellets (620e650 kg/m3) [24]. Therefore, PCS could be processed without densification and so, reducing the pre-processing cost of the material. Also, PCS has a good calorific value similar to that of wood chips (12.1e15.3 MJ/kg) although lower than wood pellets (17.1 MJ/ kg) [24]. Therefore, due its good properties as fuel, the PCS could be considered as an interesting biomass for energy purposes. In Table 3 the biochemical, proximate and ultimate analysis of some lignocellulosic biomass reported in literature are presented. As seen in Table 3, proximate analysis results for PCS were similar to other common biomass used in pyrolysis installations. PCS has a high amount of volatile matter content, 74.85% and a low moisture content that could be considered appropriate for thermochemical processes. High-moisture content biomass can be used in fermentations or other wet/aqueous conversion processes that includes biological reactions, while biomasses as presented in Table 3 (dry biomasses) are more appropriate to thermo-chemical processes, such as combustion, gasification or pyrolysis [22]. Another important feature, ash content, was adequate in PCS. High ash content is handicap because it can cause fouling or aggregation, can affect quality of combustion, gasification or pyrolysis process and result in an increase of operating costs. Furthermore, high carbon and low oxygen in biomass as compared to coal (73.1% of carbon and 8.7% of oxygen) are favorable for combustion applications [23]. A higher proportion of carbon (relative to hydrogen and oxygen content) increases the energy content of a fuel because the energy contained in carbon-carbon

Table 2 Main characteristics of raw PCS. Property Ultimate analysis

Empirical formula

Proximate analysis

Biochemical analysis (on dry basis)

Density Calorific value *By difference.

Raw PCS Carbon, % Hydrogen, % Nitrogen, % Sulfur, % Oxygen*, % H/C N/C O/C CH1.91N0.005O0.73 Volatile matter, % Fixed carbon, % Ash, % Moisture, % Cellulose, % Hemicelluloses, % Lignin, % Bulk density, kg/m3 Real density, kg/m3 Lower heating value, MJ/kg

46.81 7.44 0.27 <0.1 45.43 1.91 0.005 0.73 74.85 16.73 0.53 7.90 38.89 21.54 30.64 633 1441.4 14.09

M.A. Martín-Lara et al. / Renewable Energy 96 (2016) 613e624

617

Table 3 Comparison of the proximate, ultimate and biochemical analysis of PCS with some biomass. Property

HH

Ultimate analysis

Proximate analysis (On dry basis)

Biochemical analysis (On dry basis)

Carbon, % Hydrogen, % Nitrogen, % Sulfur, % Oxygen*, % Volatiles, % Fixed carbon, % Ash, % Moisture, % Cellulose, % Hemicelluloses, % Lignin, %

42.61 5.51 1.13 0.14 50.62 73.86 20.87 5.27 7.24 34.5 20.6 35.1 [25]

Reference

NA NA NA NA NA 71.72 17.07 1.98 9.23 75.44 5.99 44.40 [26]

PKS

CS

WS

PWS

SC

PCS

51.56 6.31 0.70 0.10 41.33 75.14 22.05 2.81 12.69 33.03 23.82 45.59 [27]

41.67 6.15 0.72 <0.1 45.28 70.83 14.87 5.53 8.67 NA NA NA [28]

43.56 5.87 0.71 <0.1 39.37 67.69 15.53 9.58 7.20 NA NA NA

45.5 6.26 1.04 <0.1 47.20 75.54 11.15 3.70 9.60 44.75 16.73 30.72 [29]

43.90 6.20 0.50 <0.1 49.40 71.30 13.00 6.20 9.50 34.20 32.90 9.60 [30]

46.81 7.44 0.27 <0.1 45.43 74.85 16.73 0.53 7.90 21.47 38.86 30.55 In this work

* By difference. NA ¼ Not Available. HH¼ Hazelnut husk. PKS¼ Palm kernel shell. CS¼ Corn straw. WS¼ Wheat straw. PWS¼ Poplar wood sawdust. SC¼ Smooth cordgrass.

bonds is greater than that of carbon-oxygen and carbon-hydrogen bonds [22,31]. Also, nitrogen and sulfur content in the biomass is typically lower than the coal [23,25]. A low content of nitrogen implies fewer emissions of toxic nitrogen oxides during combustion and a low content of sulfur is important because prevent possibility of corrosion of equipment and generation of contaminant emissions [25,32]. Finally, the proportions of cellulose and lignin in biomass are especially important in biochemical conversion processes. For example, if the objective is obtaining biofuel ethanol, a biomass with a high content of cellulose and hemicellulose is preferred in order to get a high production of ethanol (lignin has very low biodegradability properties) [22]. In general, a thermochemical processing is preferred for biomass fuels with high content of lignin. 4.2. Biosorption experiments Fig. 1 shows breakthrough curves for Cu and Pb respectively

1.0

when tests were performed in single metal solution. From these experimental data the most significant parameters of breakthrough curves were calculated according to Equations (1)e(6) and the value of these parameters are presented in Table 4. The biosorption capacity of the column (qe,b) defined as the ratio of the amount of metal ion adsorbed prior to the breakthrough point and the quantity of biomass used in the packed bed was 1.29 mg Cu/g and 6.08 mg Pb/g. Also, biosorption capacities at exhaustion or total flow time were higher for lead than for copper. In addition, a longer breakthrough time (tb) and exhaustion time (tex) was observed when the column was working with the lead solution. The breakthrough time and exhaustion time were observed to be 33 and 420 min for copper and 152 and 750 min for lead biosorption. It resulted in high volume of metal solution treated (Veff) and high percentage of metal removal. The length of unused bed at breakthrough was 12.35 and 10.68 cm for copper and lead biosorption, respectively. As described above, all samples were submitted to the same biosorption process using the same procedure but affinity of PCS for lead is better, resulting in higher metal retentions. In following subsections the effects of the presence of these heavy metals in PCS samples has been studied on terms of pyrolysis kinetics. 4.3. Pyrolysis experiments

0.8

C/Ci

0.6

0.4 Cu Pb

0.2

0.0 0

100

200

300

400

500

600

700

t, min Fig. 1. Breakthrough curves at 298 K and pH 5 for Cu(II) and Pb(II) biosorption by PCS as a function of time (Ci, inlet metal concentration, 100 mg/L; C, outlet metal concentration).

4.3.1. Thermal decomposition of raw PCS, Cu-PCS and Pb-PCS Experimental TG (ThermoGravimetric) curves for thermal degradation of pine cone shell under an atmosphere of nitrogen at heating rates of 5, 10 and 20 ºC/min are showed in Fig. 2. The weight fraction (“w”) represents the mass fraction of the solid presented at each temperature respect to initial mass of solid sample and comprises char formed and non-reacted initial solid. Fig. 3 shows the experimental DTG (Derivative ThermoGravimetric) plots for PCS pyrolysis (nitrogen). The pyrolysis process consisted of four main stages. First, there is an initial small evolution of volatiles on TG that was mainly attributed to the release of water vapor due to drying of the sample. It corresponds to the first peak on the DTG curves. Other peaks on DTG curves observed at higher temperatures can be attributed to genuine pyrolysis process. The second decomposition step in the temperature range of 150e325  C could be attributed to decomposition of hemicelluloses. Later, the third stage was mainly attributed to the degradation of the cellulose.

618

M.A. Martín-Lara et al. / Renewable Energy 96 (2016) 613e624

Table 4 Values of the most significant parameters of breakthrough curves. Breakthrough point Metal

tb min

Cu 33 Pb 152 Exhaustion point Metal

tex min

Cu 340 Pb 750 Total flow time-point

Veff,b cm3

qtotal,b mg

mtotal,b mg

Rb %

qe,b mg/g

Ce,b mg/L

198 912

19.42 91.19

19.46 91.20

99.78 99.99

1.29 6.08

0.22 0.01

Veff,ex cm3

qtotal,ex mg

mtotal,ex mg

Rex %

qe,ex mg/g

Ce,ex mg/L

2040 4500

88.55 221.09

200.53 450.00

44.16 49.13

5.90 14.74

54.89 50.87

Metal

tt min

Veff,t cm3

qtotal,t mg

mtotal,t mg

Rt %

qe,t mg/g

Ce,t mg/L

Cu Pb

420 750

2520 4500

90.98 221.09

247.72 450.00

36.73 49.13

6.07 14.74

62.20 50.87

Nomenclature: tb, breakthrough time; tex, exhaustion time; tt, total flow time; Veff, volume of the effluent at breakthrough (Veff,b), exhaustion (Veff,ex) or total flow time (Veff,t); qtotal, the total mass of metal biosorbed at breakthrough (qtotal,b), exhaustion (qtotal,ex) or total flow time (qtotal,t); mtotal, the total amount of metal ions sent to the column until breakthrough (mtotal,b), exhaustion (mtotal,ex) or total flow time (mtotal,t); R, the total metal removal expressed as percentage at breakthrough (Rb), exhaustion (Rex) or total flow time (Rt); qe, biosorption capacity, mg of sorbated metal/g of sorbent at breakthrough (qe,b), exhaustion (qe,ex) or total flow time (qe,t); Ce, the metal concentration which remains in the solution at breakthrough (Ce,b), exhaustion (Ce,ex) or total flow time (Ce,t).

Compared to hemicellulose, cellulose pyrolyzed at higher temperature range (about 325e400  C). The two weight loss peaks of hemicellulose and cellulose in DTG curve (shown in Fig. 2) not are significantly differentiated. The hemicellulose decomposition frequently is observed as a more or less marked “shoulder” and could be attributed to decomposition of hemicelluloses but also to partial decomposition of cellulose and lignin [11,26]. However, cellulose peak usually appears as a well-defined peak at higher temperatures. For PCS samples a great peak accompanied with an extra ‘‘shoulder’’ peak is observed and in Pb-PCS the hemicelluloseshoulder is better visible. This is agreed with other biomass, corn straw or wheat straw [28]. This also might be caused by the lower content of lignin in PCS. According to Ma et al. [27] biomass with a higher content of lignin presents a slower thermal degradation and allows the loss of mass of hemicellulose and cellulose be separated clearly. This result was also confirmed by other authors as Ceylan and Topçu [25] evaluating hazelnut husk or Asadullah et al. [33] studying thermal decomposition of palm kernel shell. Finally, the fourth stage (a peak at the tail of the curve) was mainly attributed to lignin. In contrast to the sharper DTG peak of cellulose, lignin pyrolyzed slowly in a very broad range of temperature (100e800  C) providing a gently sloping baseline to the DTG curve. A comparison of the thermal behavior of the different samples (raw PCS, Cu-PCS and Pb-PCS) is presented in Fig. 4. Copper and lead ions at concentrations of this work did not seem to affect the mass loss rate during pyrolysis of PCS since the differences between TG and DTG curves of raw PCS, Cu-PCS and Pb-PCS during pyrolysis process are small. In literature, only the influence of metal ions in the wood on its pyrolytic behavior has been reported. Our findings are agree with some investigations that found similarities in the shapes of the weight loss curves of wood impregnated with metal ions indicating that metals of ion-exchanged wood did not change the main degradation pathways in solid phase [34e36]. However, most of published papers present significant difference in mass degradation of wood with the presence of metals. These differences with our results may be due to method of preparation of impregnated samples or type of salt use (acetates, chlorides, sulfates, nitrates, etc.). Table 5 presented main observations in thermogravimetric curves of pyrolysis when different individual metal ions have been incorporated into biomass (mainly wood) reported in literature. Finally, it can be seen from Fig. 4a to c that increase in heating rates only shifts the peak temperature to higher value without

changing thermal profile of decomposition (the main pyrolysis sections and the maximum weight loss rates are similar for different heating rates). This fact could be due to the decreased in efficiency of heat transfer at high heating rates. Some authors have reported that heating biomass particles occurs more progressively and carries a better heat transfer to the internal parts of the biomass at lower heating rates [32,42,43]. Finally, it is important to remark that it is assumed that it is slow pyrolysis that is being considered as TGA in this work and results could not represent fast or flash pyrolysis. 4.3.2. Kinetic parameters calculated using different kinetic models The determination of the kinetic parameters that describing the pyrolytic decomposition of a lignocellulosic residue as PCS is complex due to the large number of reactions involved; therefore, simplified kinetic models are generally applied.  The Flynn-Wall-Ozawa iso-conversional method The FlynneWalleOzawa (FWO) iso-conversional method determines the activation energy without making any assumptions about the reaction mechanisms, assuming that the reaction rate only depends on the temperature. In order to apply the FWO iso-conversional model, the activation energies at various conversion values (from 0.1 to 0.9) for each biomass sample were calculated. The activation energies and correlation coefficients calculated by FWO method are listed in Table 6. The activation energy ranged from 42.92 to 202.30 kJ/mol for raw PCS, from 21.87 to 143.56 kJ/mol for Cu-PCS and from 40.89 to 192.48 kJ/mol for Pb-PCS. Three trends of activation energy can be identified for raw PCS (see Fig. 5): 1) From values of conversions 0.1 to 0.2, activation energy increases although has low values (stage of drying of sample), 2) at high conversions (0.3e0.7) activation energy increases and reaches higher values corresponding to hemicellulose and cellulose degradation, 3) at higher conversions (0.8e0.9), activation energy increased and then rapid decreased. Moreover, the difference between the lowest and highest activation energy value for all samples is too high. These variations on activation energy as a function of degree of conversion suggest that pyrolysis of PCS has multi-step kinetics with various apparent activation energies [44]. Similar activation energy distribution was obtained for the whole pyrolysis process for corn straw and wheat straw by Chen et al. [28]. Following a more sophisticated model was used to evaluate

M.A. Martín-Lara et al. / Renewable Energy 96 (2016) 613e624

Fig. 2. TG curves for the pyrolysis of raw PCS (a), Cu-PCS (b) and Pb-PCS (c) at heating rates of 5, 10 and 20 K/min (T, temperature; w, weight solid fraction, comprises char formed and non-reacted initial solid).

619

Fig. 3. DTG curves for the pyrolysis of raw PCS (a), Cu-PCS (b) and Pb-PCS (c) at heating rates of 5, 10 and 20 K/min.

620

M.A. Martín-Lara et al. / Renewable Energy 96 (2016) 613e624

Fig. 4. Comparative TG and DTG curves for the pyrolysis of raw PCS and metal-loaded PCS at heating rates of 5 K/min (a), 10 K/min (b) and 20 K/min (c).

M.A. Martín-Lara et al. / Renewable Energy 96 (2016) 613e624

621

Table 5 Main observations in thermogravimetric curves of pyrolysis when different individual metal ions have been incorporated into biomass (mainly wood). Biomass

Metal ions (counter ion)

Content, Main observations in thermogravimetric curves of pyrolysis mg/g

 Wood

Kþ (CH3COO) Ca2þ (CH3COO)

2.12 1.40

 Wood  Bagasse

Cu2þ (Cl) Zn2þ (Cl) Fe2þ (SO2 4 )

1.90 2.60 3.70

 Pure microcrystalline Naþ (Cl) Fe2þ (SO2 Avicel cellulose 4 ) Zn2þ (Cl)  Bagasse

3.60 9.40 8.40

 Wood

Cr6þ (O2) Cu2þ (O2) As5þ (O2)

19.10 7.10 16.30

 S. bicolor shoots

Ni2þ (SO2 4 ) Zn2þ (SO2 4 )

1.22 0.18

 Wood

18.73 Pb2þ (NO 3) Cd2þ (NO 5.37 3) Fe2þ (NO 4.82 3) Zn2þ (NO 3.54 3) Ca2þ (NO 1.82 3) Naþ (NO 0.68 3) Mg2þ (NO 0.38 3) 2þ 2 Cu (SO4 ) 21.36 Cr3þ (Kþ, 6.02 SO2 4 ) 9.7 B3þ (Hþ, O2)

 Wood

 The replacement of metal ions present in the wood sample by potassium and calcium ions showed that, in the presence of potassium ions, catalysis of pyrolysis process occurs in the majority of the thermal decomposition range.  Calcium does not act as a catalyst in the pyrolysis process. The behavior of calcium-exchanged wood was similar to that of acid-washed wood.  In the wood sorbed with cupric chloride, the inflection point of the DTG curve at low temperature (which corresponds to the pyrolysis of the hemicelluloses) increased, depending on the amount of salt added, while the main peak of the curve DTG (associated to the pyrolysis of cellulose) decreased.  The sugar cane bagasse sorbed with zinc chloride provided a first peak in the DTG curve that would be too large to be due to pyrolysis of the hemicelluloses.  The most intense effects are shown in pyrolysis of wood sorbed with ferrous sulfate. A catalyzed pyrolysis associated with acid sulfate ion is produced.  In the decomposition of cellulose treated with NaCl two parts of the thermogravimetric curve can be represented with the same kinetic parameters obtained with untreated cellulose; while the other third part may be represented with lower activation energy.  Treatment of cellulose with FeSO4 and ZnCl2 changed significantly all the thermal decomposition of native sample. DTG curves were well described assuming a catalytic reaction. The kinetic parameters were lower than in the case of pure Avicel.  In the bagasse samples, the presence of metal ions used in this work reduced the activation energy for the pyrolytic decomposition of this waste. Therefore, the metal ions act as catalysts.  Treatment with chromate-copper-arsenate had an important effect on the thermal behavior of the wood sample, which was also more noticeable at higher concentrations of the chemical agents.  Temperatures to which the thermal decompositions occur, in the pyrolysis of the wood, decreased.  First, the decomposition of hemicellulose is catalyzed by metals present in the wood sample due to the effectuated treatment with metals. After, the decomposition of cellulose continues very fast and is also catalyzed by remained metals in the solid matrix.  For the untreated sample a large peak with two shoulders is observed in DTG curves corresponding to the decomposition of the hemicellulose (temperature values approximately from 200 to 250  C). The peak corresponding to the degradation of cellulose appears as a single peak at a temperature of 336  C. Instead, for impregnated samples, the peak corresponding to hemicellulose appears as a single one (about 250  C and without shoulders) and the main peak corresponding to the cellulose appears to a slightly lower temperature (329  C).  The similarities in the shapes of weight loss curves indicate that metals did not change the main routes of degradation of wood, however, some of the added metal ions modified maximum temperatures of cellulose degradation.  Specifically, iron and zinc appear to catalyze the degradation of cellulose, while calcium and lead inhibited such degradation. These ions slightly changed the main degradation peak temperature of cellulose. Finally, sodium, magnesium and cadmium ions show no significant changes in thermal degradation curves during pyrolysis.  Significant effects on thermal degradation of wood were found when wood was treated with metal ions. Important differences in thermogravimetric curves were observed for untreated and treated samples. The temperature for which the pyrolysis stars was around 200  C and 150  C for untreated and treated wood, respectively. In addition, the treated wood decomposed fast at lower temperature.

kinetics of PCS pyrolysis.  A three independent parallel reactions model In the present work, the pyrolysis process is also modeled by a kinetic scheme consisting of three independent nth-order parallel reactions. In general, if the sample consists of more than one chemical component and each component decomposes independently from the others, a separate reaction can be considered for each component. Therefore, for dry lignocellulosic biomass a three independent nth-order parallel reactions model is considered as the most realistic approach [38,45e47]. The first reaction is associated to the thermal decomposition of hemicellulose, which was reactive at low temperatures, the second to the thermal decomposition of cellulose fraction and the third to the thermal decomposition of lignin which took place over a wide temperature range [48]. The pyrolysis characteristics (the estimated values for each kinetic parameter) of raw PCS, Cu-PCS and Pb-PCS are summarized in Table 7. The parameters to be optimized are activation energies (three, one for each pseudo-component), pre-exponential factors (three, one for each pseudo-component) and reaction orders (three, one for each pseudo-component). The model adequately

Reference [34,35]

[37]

[38]

[39]

[40]

[36]

[41]

reproduces the pyrolysis process since there is a very good accordance between experimental and calculated thermogravimetric curves (fit % is lower than 0.4 for raw and metal-loaded PCS). Also, similar parameters values were obtained for all samples. Activation energies for hemicellulose decomposition ranged from 233 to 238 kJ/mol (the highest values), whereas, for lignin decomposition, energy activations were the lowest (42e44 kJ/mol). In addition, the values of kinetic parameters are within the ranges that have been reported for the pyrolysis of lignocellulosic materials in previous studies considering the possible variations in material composition and the different experimental conditions for pyrolysis runs [9,18]. For example, Varhegyi and Antal [38] presented values of activation energy from 101 to 111 kJ/mol for peak of hemicellulose in a kinetic evaluation of the bagasse pyrolysis with a model of independent parallel reactions, or Antal and Varhegyi [46] found activation energy values from 197 to 239 kJ/mol for pure cellulose and substrates consisting mainly of cellulose in their review about cellulose pyrolysis kinetics. However, for lignin, the interval of kinetic constants activation energies, pre-exponential factors, and reaction orders, found in literature, is very extensive [49e54]. Some studies report activation energies of 18e65 kJ/mol [50,54] while others of approximately 180 kJ/mol for pyrolysis [55].

622

M.A. Martín-Lara et al. / Renewable Energy 96 (2016) 613e624

Table 6 Kinetic parameters obtained for the pyrolysis of raw PCS, Cu-PCS and Pb-PCS by Flynn eWalleOzawa method.

a Raw PCS 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 Average Cu-PCS 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 Average Pb-PCS 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 Average

Slope

Intercept

R2

E (kJ/mol)

5429 7090 10187 12143 14783 16346 18015 25587 15699

17.28 16.27 20.42 22.96 26.58 28.46 30.56 41.25 23.88

0.999 0.945 0.915 0.912 0.933 0.942 0.942 0.921 0.743

42.92 56.06 80.54 96.00 116.87 129.23 142.43 202.30 124.12 110.05

2766 4971 10932 13734 15717 16656 17021 16025 18159

9.87 11.78 21.29 25.09 27.56 28.47 28.55 26.44 27.79

0.948 0.992 1.000 0.996 0.992 0.980 0.999 0.981 1.000

21.87 39.30 86.43 108.58 124.26 131.68 134.57 126.70 143.56 101.88

5172 17813 17810 19026 20795 20444 21083 19595 24346

15.99 35.06 32.90 33.72 35.69 34.33 34.82 31.98 36.71

0.942 0.898 0.975 0.995 0.997 0.992 0.992 0.998 0.960

40.89 140.83 140.81 150.42 164.41 161.63 166.68 154.92 192.48 145.90

Nomenclature: a, conversion degree; R2, the coefficient of determination; E, the apparent activation energy.

Table 7 Kinetic parameters obtained for the pyrolysis of raw PCS, Cu-PCS and Pb-PCS by a mechanistic model that considers three independent parallel reactions. Pseudo-component

k0 (min1)

Raw PCS Solid 1 (Cellulose) 2.56$1017 Solid 2 (Hemicellulose) 1.89$106 Solid 3 (Lignin) 4.20 O.F. 8.54$103; N ¼ 1911; fit (%) ¼ 0.21 Cu-PCS Solid 1 (Cellulose) 2.51$1017 Solid 2 (Hemicellulose) 1.72$106 Solid 3 (Lignin) 3.07 O.F. ¼ 1.60$102; N ¼ 2416; fit (%) ¼ 0.26 Pb-PCS Solid 1 (Cellulose) 2.79$1017 Solid 2 (Hemicellulose) 2.51$106 Solid 3 (Lignin) 4.85 O.F. ¼ 1.07$102; N ¼ 844; fit (%) ¼ 0.36

E (kJ/mol)

n

r

233.82 93.91 44.43

1.42 1.21 2.29

0.40 0.06 0.02

237.03 94.41 42.69

1.47 1.58 2.70

0.25 0.08 0.05

238.39 97.73 44.09

1.47 0.70 2.60

0.19 0.26 0.09

Nomenclature: k0, the temperature-independent rate constant; E, the apparent activation energy; n, the reaction order; r, the maximum mass fraction of char obtainable by the reaction from pseudo-component; O.F., objective function; N, number of experimental points; fit, way to measure the goodness of the fit.

analysis of the chars indicated that about 95% and 99% of copper and lead presented on original waste was recovered in generated chars, respectively. This may be the reason why the kinetics of thermal decomposition in an inert atmosphere is not affected by the presence of copper or lead in the PCS sample. Furthermore, this result is very interesting because pyrolysis could be considered as an environmentally interesting solution for the valorization of loaded-PCS wastes generated during the decontamination of industrial effluents containing copper and lead by biosorption, closing the cycle of exploitation of this agroforestry waste. Reddy et al. [56] obtained similar results when Cu-loaded barks generated during decontamination of water solutions containing copper were incinerated and pyrolysed. In the case of pyrolysis, the solid char contained about 99% of metallic copper presented on original waste before pyrolysis. Also, Chouchene et al. [57] studied the behavior of olive solid waste as biosorbent of copper and nickel from aqueous solutions and found that about 96% of each metal was recovered in residual ashes after combustion of metal-loaded samples.

5. Conclusions

Fig. 5. Activation energy distribution at different conversions for raw PCS, Cu-PCS and Pb-PCS (a, conversion degree; E, the apparent activation energy).

4.3.3. Study of the presence of copper and lead on generated chars XRD figures of the solids generated by the pyrolysis of loaded samples (figures showed in Supplementary Material) revealed the presence of metallic copper or lead (depending of pyrolysed sample). As shown in Table 8, both amounts of copper and lead that have been quantified in chars are similar to those amounts of copper and lead which were retained by PCS solid during biosorption process in a fixed bed column. Specifically, chemical

During the thermal degradation of raw and metal-loaded PCS under inert atmosphere three major losses occurred due to volatilization of cellulose, hemicellulose and lignin. There was a substantial effect of heating rate on the mass loss and mass loss rate. The TG shifted to higher temperature ranges on increasing the heating rate. The pyrolysis process of PCS samples have been fitted using two kinetic models. A mechanistic model based on three independent parallel reactions reproduces adequately the pyrolysis process on the heating rates investigated, the first independent reaction corresponded to decomposition of hemicellulose fraction, the second one to cellulose and the third to lignin, although lignin decomposed on the whole range of temperatures. On the other hand, calculated kinetics parameters did not vary substantially with the presence of copper or lead on solid samples. In the observation of solid compounds after the thermal decomposition under atmosphere of nitrogen, the metals could still be predominantly bound to the char. Therefore, pyrolysis could be considered as an environmentally interesting solution for the valorization of loaded-PCS wastes generated during the decontamination of industrial effluents containing copper and lead by

M.A. Martín-Lara et al. / Renewable Energy 96 (2016) 613e624

623

Table 8 Determination of amount of copper and lead presented on raw PCS, metal loaded-PCS and chars obtained by their pyrolysis. Sample Raw

Cu-loaded Pb-loaded

Amount of sample, g

Mg metal (acid digestion)

Mg metal (theoretical)

%Recovery

Original waste

4.018

NA

NA

Char

0.884

NA

NA

Original waste Char Original waste Char

4.027 0.836 4.013 0.866

0.010 (Cu) 0.005 (Pb) 0.004 (Cu) 0.001 (Pb) 23.993 22.859 60.004 59.581

24.444 NA 59.152 NA

NA 95.27 NA 99.30

biosorption. Acknowledgement The authors gratefully acknowledge the financial support received from University of Granada via the concession of the project (PP20I5-12) “Characterization and management of byproducts generated in the gasification process of a metal-loaded waste after its use as heavy metal bioadsorbent”. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.renene.2016.05.026. References [1] A.T. McCurdy, A.J. Higham, M.R. Morgan, J.C. Quinn, L.C. Seefeldt, Two-step process for production of biodiesel blends from oleaginous yeast and microalgae, Fuel 137 (2014) 269e276.
rez, Y. Casas-Ledo
n, W. Prins, L. Radovic, Thermodynamic [2] L.E. Arteaga-Pe predictions of performance of a bagasse integrated gasification combined cycle under quasi-equilibrium conditions, Chem. Eng. J. 258 (2014) 402e411. [3] I. Backlund, L. Karlsson, L. Mattsson, U. Bergsten, Biorefinery product potentials using tree biomass as feedstock - a survey on opportunities and threats to the new wood products industry, Biomass Bioenerg. 70 (2014) 207e216. [4] A. Bazargan, S.L. Rough, G. McKay, Compaction of palm kernel shell biochars for application as solid fuel, Biomass Bioenerg. 70 (2014) 489e497. [5] M.J.R. Hinojosa, A.P. Galvín, F. Agrela, M. Perianes, A. Barbudo, Potential use of biomass bottom ash as alternative construction material: conflictive chemical parameters according to technical regulations, Fuel 128 (2014) 248e259. [6] N.R. Bishnoi, A. Garima, Fungus-an alternative for bioremediation of heavy metal containing wastewater: a review, J. Sci. Ind. Res. 64 (2005) 93e100. [7] K. Vijayaraghavan, Y.-S. Yun, Bacterial biosorbents and biosorption, Biotechnol. Adv. 26 (2008) 266e291. [8] A.P. McHale, S. McHale, Microbial biosorption of metals: potential in the treatment of metal pollution, Biotechnol. Adv. 12 (1994) 647e652.
, M. Mun ~ oz, Kinetics of pyrolysis and combustion [9] R. Font, J.A., Conesa, J. Molto of pine needles and cones, J. Anal. Appl. Pyrolysis 85 (2009) 276e286. [10] H.Q. Xie, Q.B. Yu, W.J. Duan, K. Wang, X.H. Li, X.B. Shi, Pyrolysis characteristics and kinetics of lignin derived from three agricultural wastes, J. Renew. Sustain. Energy 5 (2013), http://dx.doi.org/10.1063/1.4841215. [11] E. Pǎrpǎrit¸ǎ, M. Brebu, Azhar Uddin Md, J. Yanik, C. Vasile, Pyrolysis behaviors of various biomasses, Polym. Degrad. Stabil. 100 (2014) 1e9. [12] Z.Q. Ma, Y.M. Zhang, Q.S. Zhang, Y.B. Qu, J.B. Zhou, H.F. Qin, Design and experimental investigation of a 190 kW(e) biomass fixed bed gasification and polygeneration pilot plant using a double air stage downdraft approach, Energy 46 (2012) 140e147. [13] M. Narobe, J. Golob, D. Klinar, V. Francetic, B. Likozar, Co-gasification of biomass and plastics: pyrolysis kinetics studies, experiments on 100 kW dual fluidized bed pilot plant and development of thermodynamic equilibrium model and balances, Bioresour. Technol. 162 (2014) 21e29.
inz, G. Bla
zquez, G. Tenorio, M. Calero, Sorption of [14] M.A. Martín-Lara, F. Herna Cr(VI) onto olive Stone in a packed bed column: prediction of kinetic parameters and breakthrough curves, J. Environ. Eng. ASCE 136 (2010) 1389e1397. [15] C.D. Doyle, Kinetic analysis of thermogravimetric data, J. Appl. Polym. Sci. 5 (1961) 285e292. [16] C.D. Doyle, Series approximations to equation of thermogravimetric data, Nature 207 (1965) 290e291.
rhegyi, C. Di Blasi, Thermogravimetric analysis and devo[17] M.G. Gronli, G. Va latilization kinetics of wood, Ind. Eng. Chem. Res. 41 (2002) 4201e4208. [18] J.A. Conesa, A. Domene, Biomasses pyrolysis and combustion kinetics through n-th order parallel reactions, Thermochim. Acta 523 (2011) 176e181.

[19] J.A. Caballero, J.A. Conesa, R. Font, A. Marcilla, Pyrolysis kinetics of almond shells and olive stones considering their organic fractions, J. Anal. Appl. Pyrolisis 42 (1997) 159e175.
rhegyi, M.J. Antal, E. Jakab, P. Szabo
, Kinetic modeling of biomass py[20] G. Va rolysis, J. Anal. Appl. Pyrolisis 42 (1997) 73e87. ~o, F.J.A. Antunes, J.L. Figueiredo, Pyrolysis kinetics of lignocellulosic [21] J.J.M. Orfa materials three independent reactions model, Fuel 78 (1999) 349e358. [22] P. McKendry, Energy production from biomass (part 1): overview of biomass, Bioresour. Technol. 83 (2002) 37e46. [23] H.H. Sait, A. Hussain, A.A. Salema, F.N. Ani, Pyrolysis and combustion kinetics of date palm biomass using thermogravimetric analysis, Bioresour. Technol. 118 (2012) 382e389. [24] V. Francescato, E. Antonini, L.Z. Bergomi, Wood Fuels Handbook, AIEL (Italian Agriforestry Energy Association), Limena (Padova), 2008. [25] S. Ceylan, Y. Topçu, Pyrolysis kinetics of hazelnut husk using thermogravimetric analysis, Bioresour. Technol. 156 (2014) 182e188. [26] M.C. Celebi, N. Karatepe, An investigation of thermal decomposition behavior of hazelnut shells, Int. J. Green Energy 12 (2015) 93e97. [27] Z. Ma, D. Chen, J. Gu, B. Bao, Q. Zhang, Determination of pyrolysis characteristics and kinetics of palm kernel shell using TGAeFTIR and model-free integral methods, Energy Conv. Manag. 89 (2015) 251e259. [28] D.Y. Chen, Y. Zheng, X.F. Zhu, In-depth investigation on the pyrolysis kinetics of raw biomass. Part I: kinetic analysis for the drying and devolatilization stages, Bioresour. Technol. 131 (2013) 40e46. [29] X.L. Gu, X. Ma, L.X. Li, C. Liu, K.H. Cheng, Z.Z. Li, Pyrolysis of poplar wood sawdust by TGAeFTIR and PyeGC/MS, J. Anal. Appl. Pyrolysis 102 (2013) 16e23. [30] Y.G. Liang, B. Cheng, Y.B. Si, D.J. Cao, H.Y. Jiang, G.M. Han, X.H. Liu, Thermal decomposition kinetics and characteristics of Spartina alterniflora via thermogravimetric analysis, Renew. Energy 68 (2014) 111e117. [31] P. Pimenidou, V. Dupont, Characterisation of palm empty fruit bunch (PEFB) and pinewood bio-oils and kinetics of their thermal degradation, Bioresour. Technol. 109 (2012) 198e205. [32] T. Damartzis, D. Vamvuka, S. Sfakiotakis, A. Zabaniotou, Thermal degradation studies and kinetic modeling of cardoon (Cynara cardunculus) pyrolysis using thermogravimetric analysis (TGA), Bioresour. Technol. 102 (2011) 6230e6238. [33] M. Asadullah, N.S. Ab Rasid, S.A.S.A. Kadir, A. Azdarpour, Production and detailed characterization of bio-oil from fast pyrolysis of palm kernel shell, Biomass Bioenerg. 59 (2013) 316e324. [34] W.P. Pan, G.N. Richards, Influence of metal ions on volatile products of pyrolysis of wood, J. Anal. Appl. Pyrolysis 16 (1989) 117e126. [35] W.P. Pan, G.N. Richards, Volatile products of oxidative pyrolysis of wood: influence of metal ions, J. Anal. Appl. Pyrolysis 17 (1990) 261e273. [36] Z.A. Mayer, A. Apfelbacher, A. Hornung, A comparative study on the pyrolysis of metal- and ash-enriched wood and the combustion properties of the gained char, J. Anal. Appl. Pyrolysis 96 (2012) 196e202. [37] G.N. Richards, G. Zheng, Influence of metal ions and of salts on products from pyrolysis of wood: applications to thermochemical processing of newsprint and biomass, J. Anal. Appl. Pyrolysis 21 (1991) 133e146. [38] G. Varhegyi, M.J. Antal, Kinetics of the thermal decomposition of cellulose, hemicellulose, and sugar cane bagasse, Energy Fuels 3 (1989) 329e335. [39] L. Helsen, E. Van den Bulck, Kinetics of the low-temperature pyrolysis of chromated copper arsenate-treated wood, J. Anal. Appl. Pyrolysis 53 (2000) 51e79. [40] Z. Al Chami, N. Amer, K. Smets, J. Yperman, R. Carleer, S. Dumontet, J. Vangronsveld, Evaluation of flash and slow pyrolysis applied on heavy metal contaminated Sorghum bicolor shoots resulting from phytoremediation, Biomass Bioenerg. 63 (2014) 268e279. [41] S.E. Kinata, K. Loubar, A. Bouslamti, C. Belloncle, M. Tazerout, Influence of impregnation method on metal retention of CCB-treated wood in slow pyrolysis process, J. Hazard. Mater 233e234 (2012) 172e176. [42] S.S. Idris, N.A. Rahman, K. Ismail, A.B. Alias, Z.A. Rashid, M.J. Aris, Investigation on thermochemical behaviour of low rank Malaysian coal, oil palm biomass and their blends during pyrolysis via thermogravimetric analysis (TGA), Bioresour. Technol. 101 (2010) 4584e4592. [43] R.S. Chutia, R. Kataki, T. Bhaskar, Thermogravimetric and decomposition kinetic studies of Mesua ferrea L. deoiled cake, Bioresour. Technol. 139 (2013) 66e72.

624

M.A. Martín-Lara et al. / Renewable Energy 96 (2016) 613e624

[44] J.M. Cai, L.S. Bi, Kinetic analysis of wheat straw pyrolysis using isoconversional Methods, J. Therm. Anal. Calorim. 98 (2009) 325e330. [45] S.M. Ward, J. Braslaw, Experimental weight loss kinetics of wood pyrolysis under vacuum, J. Combust. Flame 61 (1985) 261e269. [46] M.J. Antal, G. Varhegyi, Cellulose pyrolysis kinetics: the current state of knowledge, Ind. Eng. Chem. Res. 34 (1995) 703e717. [47] K. Raveendran, A. Ganesh, K.C. Khilar, Pyrolysis characteristics of biomass and biomass components, Fuel 75 (1996) 987e998. [48] R.A. Naranjo, J.A. Conesa, E.F. Pedretti, O. Romero, Kinetic analysis: simultaneous modelling of pyrolysis and combustion processes of dichrostachys cinerea, Biomass Bioenerg. 36 (2012) 170e175. [49] E.M. Suuberg, W.A. Peters, J.B. Howard, Product composition and kinetics of lignite pyrolysis, Ind. Eng. Chem. Proc. Des. Dev. 17 (1978) 37e46. [50] W.K. Tang, US Forest Service Research. Paper FPL 71 Forest Prods. Lab, Department of Agriculture, Madison, WI, 1967. [51] G.E. Domburg, V.N. Sergeeva, Thermal degradation of sulphuric acid lign1ns of hard wood, J. Therm. Anal. 1 (1969) 53e62.

[52] H.F.J. Wenzl, The Chemical Technology of Wood, Academic Press, New York, 1970. [53] T.R. Nunn, J.B. Howard, J.P. Longwell, W.A. Peters, Product compositions and kinetics in the rapid pyrolysis of milled wood lignin, Ind. Eng. Chem. Proc. Des. Dev. 24 (1985) 844e852. [54] J.A. Caballero, R. Font, A. Marcilla, A.N. García, Flash pyrolysis of klason lignin in a pyroprobe 1000, J. Anal. Appl. Pyrolisis 27 (1993) 221e244. [55] J. Zhang, T. Chen, J. Wu, A novel Gaussian-DAEM-reaction model for the pyrolysis of cellulose, hemicellulose and lignin,, RSC Adv. 4 (2014) 17513e17520. [56] B.R. Reddy, N. Mirghaffari, I. Gaballah, Removal and recycling of copper from aqueous solutions using treated Indian barks, Resour. Conserv. Recycl. 21 (1997) 227e245.
, Biosorption performance, combustion [57] A. Chouchene, M. Jeguirim, G. Trouve behavior and leaching characteristics of olive solid waste during the removal of copper and nickel from aqueous solutions, Clean. Technol. Environ. Policy 16 (2014) 979e986.