Article pubs.acs.org/EF
Thermal Degradation Kinetic Study of Rubber Seed Oil and Its Methyl Esters under Inert Atmosphere Ali Shemsedin Reshad, Pankaj Tiwari,* and Vaibhav V. Goud* Department of Chemical Engineering, Indian Institute of Technology Guwahati, Guwahati, Assam 781039, India S Supporting Information *
ABSTRACT: Nonedible vegetable oil feedstocks are promising for sustainable production of biodiesel. Thermal decomposition characteristics of the feedstocks and their biodiesel are crucial for handling and quality control. Thermal degradation of rubber seed oil (RSO) and rubber seed oil methyl esters (ROME) was investigated with the help of thermogravimetry. The samples were pyrolyzed from 30 to 800 °C at heating rates of 10 °C/min to 50 °C/min with a 10 °C/min increment under a nitrogen atmosphere. The temperature window for thermal degradation of RSO and ROME was shifted toward a higher range as the heating rate increased from 10 °C/min to 50 °C/min. A transesterification reaction leads to a decrease in the molecular weight of triglycerides present in the sample (RSO), and this causes a lower thermal stability of the produced product (ROME). Fourier transform infrared (FT-IR) analysis of evolved gaseous products during pyrolysis revealed the formation of water, carbon dioxide, carbon monoxide, and saturated (alkanes) and unsaturated (alkenes) aliphatic hydrocarbons. Friedman (FRD), Flynn−Wall− Ozawa (FWO), modified Coat−Redfern (MCR), and Kissinger (KM) methods and Avrami theory were applied to calculate the values of activation energy (E), order of reaction (n), and enthalpy (ΔH). Furthermore, the pre-exponential factor (A), entropy (ΔS), and Gibbs free energy (ΔG) were also calculated.
■
(FT-IR) with TG.10,14 The analysis of evolved gas by using such a coupled technique has been used in various research fields to estimate sample structure and composition. Li et al.10 employed thermogravimetry analysis (TGA) to estimate thermal behavior of biodiesel samples derived from peanut oil, palm oil, and waste cooking oil. TGA results showed that the onset decomposition and peak temperatures for palm oil methyl esters are higher than both peanut oil and waste cooking oil methyl esters due to a lower content of unsaturated fatty acid components in palm oil methyl esters. The real time analysis of evolved products using attached FT-IR revealed the formation of alkanes and cyclic and aromatic compounds along with CO, CO2, and H2O. Santos et al.13 reported the thermal decomposition of sunflower oil and its biodiesel using nonisothermal thermogravimetric analysis under an inert atmosphere. The values of kinetic parameters have been estimated as 155.62−200.12 kJ/mol (E) and 0.95−1.82 (n) for sunflower oil and 61.32−115.35 kJ/mol (E) and 0.69−1.89 (n) for its biodiesel. Souza et al.15 evaluated the thermal and kinetic behavior of cotton oil and its biodiesel under air and nitrogen atmospheres. The values of activation energy (E) for cotton oil have been found higher than that of cotton-oil-based biodiesel for both air and nitrogen atmospheres. Investigations on thermal degradation of rubber seed oil and its biodiesel have been rarely reported. In the present work, the thermal degradation behavior of rubber seed oil and its biodiesel under a nitrogen atmosphere was studied. There are varieties of kinetic methods available in the open literature to deduce the kinetic parameters. However, the Friedman model (FRD), Flynn−Wall−Ozawa (FWO), and Coats−Redfern (CR) kinetic methods are considered more reliable and widely used. Hence, in
INTRODUCTION Biodiesel has gained worldwide attention as a partial substitute for fossil-based diesel fuel. Chemically, biodiesel is a mixture of long chain fatty acid alkyl esters derived from triglyceride present in vegetable oil, animal fat, and waste cooking oil through the transesterification process. The transesterification process improves the physicochemical, thermal, and flow properties of the feedstock. Soybean oil, sunflower oil, palm oil, coconut oil, corn oil, rapeseed oil, and olive oil are the most widely used first generation feedstocks for the production of biodiesel. First generation biodiesel feedstocks are usually categorized as a part of the food chain and account for around 60%−80% of the total biodiesel production cost.1,2 The challenge for the higher cost of biodiesel production can be overcome by using nonedible feedstocks. Jatropha curcas L. (Jatropha),1,2 Ricinus communis (Castor),3,4 Hevea brasiliensis (rubber tree),5,6 Pongamia pinnata L. (Karanja), Calophyllum inophyllum L. (Polanga),7 Croton megalocarpus (Musine),7 Cocos nucifera (coconut),7 and Mesua ferrea (Nahor)1,4 are sustainable, nonedible raw materials for biodiesel synthesis. Besides edibility of the feedstock, thermal degradation characteristics of both the feedstock and product are of great concern for scientific applications.8−11 Thermal degradation analysis can be used to estimate the thermal properties such as activation energy, enthalpy, Gibbs free energy, entropy, and heat capacity as well as the quality of the produced biodiesel.12,13 The thermogravimetry (TG) technique monitors the physical and chemical changes of a sample happening with changes in temperature. Thermal degradation of vegetable oils and their methyl esters mainly involves physicochemical processes of volatilization and decomposition. The gaseous products evolved during thermal decomposition of the sample can be identified in real time by coupling Fourier transform infrared spectroscopy © 2017 American Chemical Society
Received: August 1, 2017 Published: August 24, 2017 9642
DOI: 10.1021/acs.energyfuels.7b02249 Energy Fuels 2017, 31, 9642−9651
Article
Energy & Fuels
tube and gas cell were heated up to 250 °C to prevent condensation of the evolved products. Thermal Degradation Study. The values of activation energy (E) for the thermal decomposition of RSO and ROME were calculated using Friedman (FRD),18 Flynn−Wall−Ozawa (FWO),19 Kissinger (KM),20 and modified Coats−Redfern (MCR) methods.11,21 The thermal decomposition rate, considering Arrhenius temperature dependency for a constant heating rate, can be expressed with eq 1:
the present study, Friedman (FRD), Flynn−Wall−Ozawa (FWO), modified Coats−Redfern (MCR), and Kissinger (KM) kinetics methods were applied to estimate the value of activation energy (E) and enthalpy (ΔH) for thermal degradation reactions.
■
MATERIALS AND METHODS
Materials. Rubber seeds, collected from Assam India, were deshelled manually, and the obtained kernels were subjected to oil extraction. Rubber seed oil (RSO) and rubber seed oil methyl esters (ROME) were obtained with a Soxhlet extractor and ultrasonic- assisted transesterification, respectively. The details can be found in our previous studies.16,17 The extracted oil and produced biodiesel under optimum conditions were used for thermal analysis. The physicochemical properties of obtained RSO and ROME are presented in Table 1.16
β · dα /dT = f (α)· A exp(− E /RT )
where β, f(α), A, α, E, T, and R refer to the heating rate (°C/min), reaction mechanism model, pre-exponential factor (1/min), degree of decomposition of the samples (conversion), activation energy (J/mol), temperature (K), and gas constant (8.314 J/mol K), respectively. The value of α at the appropriate time/temperature can be calculated using TG data (eq 2):
α = (w0 − wt )/(w0 − wf )
Table 1. Physico-chemical Properties of Rubber Oil (RSO) and Its Methyl Esters (ROME) Samples properties specific gravity @ 24 °C kinematic viscosity @ 24 °C kinematic viscosity @ 40 °C calorific value iodine value acid value saponification value refractive index @ 24 °C cloud point pour point cetane no. moisture content flash point fire point
unit
a
b
ROME
biodiesel
0.91
0.883
mm2/s
30
5.82
0.86− 0.9d 1.9−6d
1.9−4.1d
mm2/s
13.13
3.81
1.9−6
1.9−4.1
diesel
0.846d
MJ/kg
39.34
39.53
35 min
g I2/100g mg KOH/ g oil mg KOH/ g oil
113 24
114 0.4
120 max 0.5 max
45.62− 46.48 N/A 0.35
235.28
190
N/S
N/A
1.47
1.45
N/S
N/S
3 2e 44 0.27 273 282
2.5 −3e 49.9 0.33 131 146
−3 to 12 −15 to 10 47 min 0.05 max 130 min N/S
−15 to 5 −20 40 min 0.05 max 52−96 N/S
°C °C % °C °C
(2)
where w0, wt, and wf are the initial weight, weight at time t, and final weight, respectively. The Friedman model (FRD; eq 3) is the first and general isoconversional method on the basis of the model free differential technique and is obtained by taking the natural logarithm on both sides of eq 118 and becomes eq 3.
c
RSO
(1)
ln(dα /dt )i , j = ln(f (α)·A)− E /RTi , j
(3)
18
21
Aspects of the Friedman and Coats−Redfern methods can be combined for an estimation of kinetic parameters for multiple heating rates. The general expression for the modified Coats−Redfern for the nth order is as follows (eq 4):22−24 ln(β /T 2)i , j = ln(A · R /E ·g (α))i , j − E /R· Ti , j
(4)
where g(α) is the integral form of the reaction model. Besides the differential approaches, the fundamental rate expression (eq 1) can also be used by the integral method of the Flynn−Wall− Ozawa model (eq 5) using Doyle approximation to estimate the kinetic parameters.19
ln βi , j = C − 1.052E /RTi , j
(5)
Kissinger proposed that the maximum rate occurs when d(dα/dt)/dt is zero. Therefore, the differentiation of the fundamental Arrhenius expression (eq 1) for a constant heating rate at which the maximum rate occurs (at peak temperature Tmax) is equal to zero. The simplified Kissinger model for the first order thermal decomposition at peak temperature (Tmax) expression is as follows (eq 6):
a
Rubber seed oil methyl ester. bStandard ASTM test methods. cFossil fuel. dMeasured at 40 °C. eDSC method. N/S, not specified; N/A, not applicable. The 13C nuclear magnetic resonance (NMR) spectra of rubber seed oil and its methyl esters (biodiesel) are depicted in Figures S1 and S2, respectively. The signals at 69 and 62 ppm in the 13C NMR spectrum of rubber seed oil are due to the carbonyl methylene groups (69 ppm (H− C−O−) and 62 ppm (CH2−C−O−); Figure S1), while the signal is absent in the biodiesel (ROME; Figure S2). It can be clearly seen that the glyceride backbone of triglyceride is totally absent in the ROME sample. The methoxy carbon of methyl esters of ROME illustrates the signal at 51.49 ppm. The unsaturation signal (−CC−) obtained between 133 and 120 ppm in 13C NMR is due to the presence of linoleic (polyunsaturated), linolenic (polyunsaturated), and oleic (monounsaturated) fatty acids and ester in RSO (Figure S1) and ROME (Figure S2), respectively. Methods. Thermogravimetric Analysis. Thermal degradations of rubber seed oil (RSO) and rubber seed oil methyl esters (ROME) were evaluated using a TG analyzer (Netzch STA449F300) at various heating rates (10, 20, 30, 40, and 50 °C/min) under a nitrogen atmosphere. The samples were weighed at ∼10 mg in an alumina crucible. TGA experiments were conducted from 30 to 800 °C for respective heating rates, where the nitrogen gas (99.999% purity) flow rate was set at 60 mL/min. The evolved gas products during thermal decomposition at a heating rate of 40 °C/min were continuously monitored and measured using a PerkinElmer TGA and FT-IR coupled system. The gas transfer
⎛ β ⎞ ⎛ AR ⎞ ⎛ E ⎞⎛ 1 ⎞ ln⎜ 2 ⎟ = ln⎜ ⎟ − ⎜ ⎟⎜ ⎟ ⎝ E ⎠ ⎝ R ⎠⎝ Tmax ⎠ ⎝ Tmax ⎠ j j
(6)
On the basis of the same degree of thermal degradation (i) at different heating rates (j), linear plots of ln(dα/dt)i,j versus 1/Ti,j (FRD, eq 3), ln(βj/T2i,j) versus 1/Ti,j (MCR, eq 4), ln(βj) versus 1/Ti,j (FWO, eq 5), and ln(β/Tmax2) versus 1/Tmaxj (KM, eq 6) were constructed. The slope of the straight line was used to calculate the value of activation energy. Thermodynamic parameters such as enthalpy (ΔH; eq 7), Gibbs free energy (ΔG), and entropy (ΔS) of RSO and ROME samples were calculated at the maximum peak temperatures.10,13,25 For this purpose, the value of the pre-exponential factor (A) for RSO and ROME were calculated using eq 8.10,25 The obtained pre-exponential factors were used to calculate ΔG (eq 9) and ΔS (eq 10).26−28
ΔH = E − RT ⎛ E ⎞ β·E ⎟ ·exp⎜ ⎝ RT ⎠ R·T 2
(8)
⎛ K ·T ⎞ ΔG = E + RT ln⎜ B ⎟ ⎝ h·A ⎠
(9)
A=
9643
(7)
DOI: 10.1021/acs.energyfuels.7b02249 Energy Fuels 2017, 31, 9642−9651
Article
Energy & Fuels
Figure 1. TGA and DTG profiles (A, B) for RSO and (C, D) for ROME samples.
Table 2. TG Characteristic Properties for Active Pyrolysis Stages of RSO and ROME Samples heating rate (°C/min) 10
20
30
40
50
stage sample
propertiesa
I
II
I
II
I
II
I
II
I
II
RSO
T0, °C Tf, °C Δwt, % Tmax, °C wmax, %/min T0, °C Tf, °C Δwt, % Tmax, °C wmax, %/min
223 309 20 271 3.26 120 407 97.2 250 15.9
353 471 69.5 416 12.1
242 334 20 296 5.9 165 456 97.2 268.5 29.1
373 480.0 69.5 424 35.2
249 345.6 20 310 8.8 172 470 97.2 277.5 41.4
382 483 69.5 438 44.6
253 350 20 318 11.7 181 474 97.2 283.5 55.2
384 487 69.5 443 60
264 365 20 328 13.6 193 480 97.2 287.9 76.2
395 488 69.5 446 76.7
ROME
a
To: Initial temperature for the main mass loss (°C). Tf: Final temperature for the main mass loss (°C). Tmax: Temperature for maximum rate of mass loss (°C). Δwt: change of mass loss. wmax: Maximum mass loss rate (%/min).
ΔS =
ΔH − ΔG T
At a particular degradation temperature, T, the points of ln(−ln(1 − α)) versus ln β at various heating rates could be fitted to a linear line. The reaction order (n) can be calculated from the slope of the linear line.
(10)
■
where KB is the Boltzmann constant (1.3806 × 10−23 m2 kg s−2 K−1) and h is the Planck constant (6.6261 × 10−34 m2 kg s−1). Besides the values of activation energy and pre-exponential factor of thermal degradation, the reaction order is also an important index. Avrami theory was applied to calculate the order of thermal degradation of both RSO and ROME at various temperatures using (eq 11):10,13,29 ⎛ A · exp(− E /RT ) ⎞ α = 1 − exp⎜ ⎟ βn ⎠ ⎝
RESULTS AND DISCUSSION TGA Analysis. The thermal decomposition behavior of both RSO and ROME samples was investigated. TGA and DTG profiles for RSO and ROME at various heating rates are shown in Figure 1. Volatilization occurs during the early stage, when the lighter components evolved. During decomposition, the heavier components break to low molecular weight components. Further, the evolved products, low molecular weight components during decomposition, go through the volatilization process. Thus, in the active thermal degradation stage, both phenomena take place. A change in the slope of the TG profile was considered the beginning of a new stage. However, only the
(11)
Equation 11 can be simplified to the following expression (eq 12):
ln(− ln(1 − α)) = ln A −
E − n ln β RT
(12) 9644
DOI: 10.1021/acs.energyfuels.7b02249 Energy Fuels 2017, 31, 9642−9651
Article
Energy & Fuels Table 3. Value of Activation Energy Deduced from FRD, FWO, and MCR Methods activation energy (kJ/mol) at conversion (%)a sample RSO
stage
method #
10
20
30
40
50
60
70
80
90
average
I
FRD FWO MCR FRD FWO MCR FRD FWO MCR
72.8 83.8 79.3 164.2 147.8 144 71.8 70.5 65.9
70.7 81.6 76.8 196.9 162.8 160 89.6 78.4 74.0
69.8 80.2 75.2 220.7 177.9 176.3 93.9 84.6 80.3
73.5 79.6 74.4 212.1 188.9 187.3 93.3 88.1 83.8
74.1 79.4 74.1 231.2 197.2 196 90.9 89.9 85.6
75.2 79.5 74.1 261.1 209.4 209.5 87.1 90.4 86.1
77.72 79.9 74.3 283.1 225.4 225.3 83.9 90.1 85.5
83.7 81.1 75.5 328.6 250.9 252.1 78.2 88.1 83.3
92.7 83.7 78.1 433.3 319.2 324.1 100.9 89.7 84.2
76.7 80.9 75.8 258.9 208.8 208.3 87.7 85.5 80.9
II
ROME
I
a Stage by stage conversion (α) of active pyrolysis of the samples, # average R2 was found to be ∼0.994, and overall average activation energy (E) for RSO decomposition was found to be 167.8 kJ/mol (FRD), 144.9 kJ/mol (FWO), and 152.5 kJ/mol (MCR).
Figure 2. FT-IR spectra for evolved products during thermal decomposition of (A) RSO and (B) ROME at a heating rate of 40 °C/min.
compounds also takes place within this stage. The active decomposition of RSO in the second stage (RSO-II) is due to the degradation and volatilization of triglyceride. At a heating rate of 20 °C/min, the thermal decomposition of RSO and its methyl ester (ROME) occurred in the temperature range of 242−480 °C (95.7 wt %) and 165−456 °C (97.2 wt %), respectively. The active thermal decomposition and volatilization of the RSO sample were started at relatively higher temperature values, started around 242 °C and completed around 480 °C, compared to its methyl ester (ROME; i.e., started around 165 °C and completed around 456 °C). This is due to the fact that RSO has higher molecular weight compounds and a stronger intermolecular force (higher viscosity) as compared to ROME. The value of onset temperature for ROME (165 °C) obtained in the present study is approximately similar to that of palm oil methyl
active thermal degradation stage was subjected to kinetic analysis. Therefore, overall, three stages were found for both the samples, RSO and ROME (Figure 1). Active pyrolysis stages for RSO (stage I, IIa, and IIb) and for ROME (stage I and II) were selected for kinetic analysis. The two split peaks of RSO (stage IIa and IIb) and ROME (stage Ia and Ib) were considered a single stage (II for RSO and stage I for ROME) for kinetic analysis. The main (active) thermal decomposition of the ROME sample at all the heating rates considered (10, 20, 30, 40, and 50 °C/min) occurred in single stage that describes the decomposition and volatilization phenomena. However, two-stage thermal decomposition was observed for the RSO sample (Table 2). The first stage (RSO-I) showed the presence of higher free fatty acids in the sample. In addition to free fatty acid decomposition, mass loss due to moisture removal and degradation of light volatile 9645
DOI: 10.1021/acs.energyfuels.7b02249 Energy Fuels 2017, 31, 9642−9651
Article
Energy & Fuels
Figure 3. Formation of evolved products during thermal decomposition of (A) RSO and (B) ROME at a heating rate of 40 °C/min.
ester (164.5 °C), peanut oil methyl ester (155.8 °C), and waste cooking oil methyl ester (142.2 °C) as reported by Li et al.10 The onset temperature values for palm oil methyl ester (164.5 °C),10 peanut oil methyl ester (155.8 °C),10 and waste cooking oil methyl ester (142.2 °C)10 were approximately similar to ROME (165 °C). It can be seen from Figure 1A−D that the onset temperature and the temperature at which the rate of mass loss is a maximum (Tmax) were shifted toward higher temperatures with increasing heating rate. This is due to low heat distribution (heat transfer limitation).10,22,30 With respect to heat transfer phenomena, the initial thermal degradation temperature values for rubber seed oil in the first stage (RSO-I) and second stage (RSO-II) and for rubber seed oil methyl esters (ROME) were shifted from 223 to 264 °C, 253 to 295 °C, and 120 to 193 °C as the heating rate increased from 10 °C/min to 50 °C/min, respectively (Table 2). Furthermore, the values of peak temperature, Tmax, were also changed from 271 to 328 °C, 416 to 446 °C, and 250 to 287.9 °C for thermal degradation of RSO-I, RSO-II, and ROME, respectively. Similarly, some other parameters such as Tf and wmax values were also increased (Table 3). From Figure 1B and D, it can also be clearly observed that the heating rate has a significant effect on the rate of thermal degradation of the samples. Fourier Transform Infrared (FT-IR) Analysis of Evolved Products. FT-IR spectra obtained for RSO and ROME were found with similar characteristics due to the similar nature of the chemical structure of the functional group present in the samples (Figure S3, Supporting Information). However, the signals specific to the hydroxyl group of free fatty acid can be observed only in the spectrum of RSO at 3480 cm−1. In addition, a single peak at 1456 cm−1 was observed for the bending vibration of CH2 and CH3 groups only in the RSO sample. Furthermore, a signal peak specific to the ester functional group for triglycerides for RSO and methyl esters of ROME was clearly observed at wavenumber 1735−1740 cm−1. Figure 2 shows FT-IR signatures of evolved products at various mass loss temperatures during thermal volatilization and decomposition of RSO and its methyl esters for a heating rate of 40 °C/min. Absorbance peaks corresponding to gaseous and liquid water molecules can be seen only for RSO thermal degradation at 3500−3950 cm−1 and 3400−3500 cm−1, respectively.31 This shows that the moisture content for RSO is higher than for ROME and the oxygencontaining group in ROME (i.e., R1COOR2) mainly decom-
posed into CO− and C−O.10,12 The characteristic infrared absorption peaks for volatile component of the functional group are clearly shown in Figure 2. Symmetric and asymmetric stretching vibrations of the −CH− and −CH3 asymmetric deformation vibration in the range of wavenumbers 3000−2700 cm−1 and 1475−1000 cm−1, respectively, revealed the presence of alkanes in the evolved products during the thermal degradation of RSO and ROME. Carbonyl groups of aldehydes and ketones were also observed as H−CO− and −CO− in plane bending vibrations at 1720−1740 cm−1 and 1735−1750 cm−1 (Figure 2), respectively. Furthermore, C−O−C (stretching vibration at 1000−1300 cm−1), CO (bending vibration at 2250−2400 cm−1 and 580−730 cm−1), CO (stretching vibration at 2000−2250 cm−1), and C−O (stretching vibration at 2200−2100 cm−1) were observed in the evolved gaseous product (Figure 2).14 Taking the absorbance of identified volatile compounds such as alkanes, alkenes, aldehydes, ketones, ethers, and CO2, the intensity of the evolved compounds with increasing thermal decomposition temperature is presented in Figure 3. As the thermal decomposition temperature increased, ether and aldehydes were formed due to deoxygenation of ester.12 The presence of CO2 in the evolved product revealed that the decarboxylation of ester and triglyceride occurred during the thermal decomposition of ROME and RSO.10 It can be seen from Figure 3 that the maximum absorbance characteristic peaks for alkenes, alkanes, aldehydes, and ether occurred at the same temperature (448 °C) and, for ketones and CO2, at temperature of 536 °C during thermal degradation of RSO. Similarly, during the thermal degradation of ROME, the maximum rate of alkenes, alkanes, aldehydes, and ether production occurred at a temperature of 307 °C, while for ketones and CO2, the maximum rate values were found at 571 °C. As can be observed in TGA profiles (Figure 1) and FT-IR spectra (Figure 2), beyond 500 °C, the mass losses during thermal decomposition of ROME (1.44 wt %) and RSO (1.19 wt %) are mainly due to the formation of CO2 and ketone (Figure 3). Taking the Lambert− Beer law into consideration, the concentrations of alkanes in evolved products were found to be at a maximum during the thermal decomposition of RSO and ROME.12,14 The appearance of absorbance profiles of RSO and ROME with temperature (Figure 3) were found similar to that of DTG curves (Figure 1). Kinetic Parameter Calculation. It can be seen from Figure 4 that the degree of conversion of both the samples was greatly 9646
DOI: 10.1021/acs.energyfuels.7b02249 Energy Fuels 2017, 31, 9642−9651
Article
Energy & Fuels
Figure 4. TGA mass conversion for (A) RSO-I, (B) RSO-II, and (C) ROME samples.
Figure 5. Regression plots based on (A) FRD for RSO-I, (B) FRD for RSO-II, (C) FWO for RSO-I, and (D) FWO for RSO-II.
varied with temperature and heating rate. To estimate the dependency of activation energy on temperature and degree of conversion during the active decomposition (pyrolysis) process, nine conversion fractions from 0.1 to 0.9 with an increment of 0.1 were selected at all heating rates. On the basis of FRD, FWO, and MCR isoconversional methods, the activation energies (E) for selected conversions were calculated from the slopes of linear regression (Figure 5−7). The deduced regression lines for RSO-I, RSO-II, and ROME are presented in Figures 5−7, and the values of activation
energies (E) are summarized in Table 3. Parallel lines shown in Figures 5−7 indicate that the values of activation energy (E) for thermal degradation of the samples (RSO and ROME) for free fatty acids (RSO-I), triglycerides (RSO-II), and fatty acid methyl esters (ROME) in the respective sample follow the same reaction rate or intensity. In other words, the values of activation energy (E) obtained at different degrees of conversion and temperatures were approximately similar for RSO-I and ROME. This suggests that a single mechanism or unification of multiple reactions mechanism was followed in the active thermal degradation of 9647
DOI: 10.1021/acs.energyfuels.7b02249 Energy Fuels 2017, 31, 9642−9651
Article
Energy & Fuels
Figure 6. Regression plots based on (A) MCR for RSO-I, (B) MCR for RSO-I,I and (C) KM for RSO-I and RSO-II.
Figure 7. Regression plots based on (A) FRD, (B) FWO, (C) MCR, and (D) KM for ROME thermal degradation.
RSO-I and ROME.10,11 However, for RSO-II, fitted lines at different conversions were slightly not parallel to each other, and this shows the change in activation energy at different degrees of conversions due to multiple and parallel reactions during thermal decomposition of triglycerides of RSO. Changes in the slope of the lines at different conversions show that the rates of thermal decomposition differ due to multiple reactions occurring. The higher values of activation energies were found at a later stage of conversion (α = 0.9). The values of R2 were found to be more
than 0.99 for the selected conversions (0.1 to 0.9), which show the fitness of the methods considered, FRD, FWO, and MCR. The values of activation energy (E) for RSO-I and RSO-II vary from 69.8 to 92.7 kJ/mol and 144 to 433.3 kJ/mol, respectively. The difference in the values of activation energy at lower (α = 0.1) and higher (α = 0.9) degrees of conversion for RSO-II reveals that incomplete decomposition of free fatty acids occurred in the first stage. Further, it also suggests that the thermal decomposition of RSO samples is a complex reaction which involves several parallel, competitive, and consecutive 9648
DOI: 10.1021/acs.energyfuels.7b02249 Energy Fuels 2017, 31, 9642−9651
Article
Energy & Fuels
Figure 8. Regression plots to calculate the order of reaction proposed by Avrami theory for (A) RSO-I, (B) RSO-II, and (C) ROME. (D) Enthalpy change for RSO and ROME based on FRD and MCR methods.
Table 4. Value of Reaction Order for Thermal Decomposition of RSO and ROME Calculated Using Avrami Theory RSO-I
RSO-II
ROME
n
R2
temp (°C)
n
R2
temp (°C)
n
R2
267 272 282 292 297 307
1.59 1.41 1.31 1.26 1.26 1.42
0.99 0.99 0.99 0.99 0.98 0.98
402 407 417 427 432 442 452 457
1.41 1.24 1.16 1.06 0.88 0.69 0.50 0.38
0.99 0.99 0.99 0.99 0.99 0.98 0.98 0.99
227 237 247 257 267 277 287 297 307 317 327
0.99 0.99 0.99 0.99 0.99 0.99 0.98 0.98 0.98 0.99 0.99
average (n)
1.37
0.92 0.96 1.04 1.10 1.05 0.86 0.61 0.39 0.24 0.19 0.19 0.69
temp (°C)
0.92
the Kissinger method (E; 67.89 kJ/mol, RSO-I; 183.85 kJ/mol, RSO-II) are lower than those obtained by FRD (76.7 kJ/mol, RSO-I; 258.9 kJ/mol, RSO-II), FWO (80.9 kJ/mol, RSO-I; 208.8 kJ/mol, RSO-II), and MCR (75.8 kJ/mol, RSO-I; 208.3 kJ/mol, RSO-II) methods. Table 3 demonstrates that the average values of activation energy (E) of ROME were estimated as 87.7, 85.5, and 80.9 kJ/ mol calculated using FRD, FWO, and MCR approaches, respectively. The values are lower than that of RSO-II. The higher molecular weight of triglycerides of RSO requires high energy for thermal decomposition and volatilization.13 The differences in the values of activation energy (E) of RSO and ROME revealed that the decomposition and/or volatilization mechanism of the samples occurs in a different manner, and RSO was chemically modified through the transesterification process (Table 3). However, the activation energy range for the first stage decomposition of RSO was found to be similar to the ROME thermal degradation with 5−12% error. Figure 7D shows the
types of reactions. Activation energy (E) is the critical energy barrier to be overcome to generate a chemical reaction, and also it represents the minimum energy required to break the chemical bond between atoms.29 Hence, the higher values of activation energy for RSO-II as compared to RSO-I indicate that more difficult reactions have taken place during the secondary stage (RSO-II). This is may be due to a higher free fatty acid content in the RSO sample. Considering all three methods (FRD, FWO, and MCR), the overall average values of activation energy (E) for thermal decomposition of RSO were found to be 167.8 kJ/mol (FRD), 144.85 kJ/mol (FWO), and 142.05 kJ/mol (MCR). The obtained overall average values of activation energy for RSO thermal decomposition are in good agreement with those of sunflower oil (170−210 kJ/mol),13 soybean (146.6−160.2 kJ/ mol),11 karanja seed oil (156.5−160.7 kJ/mol),11 and mustard seed oil (142.4−148.1 kJ/mol).11 Figure 6C shows the linear plots for the Kissinger method for thermal decomposition of RSO-I and RSO-II. The values of activation energy obtained by 9649
DOI: 10.1021/acs.energyfuels.7b02249 Energy Fuels 2017, 31, 9642−9651
Article
Energy & Fuels
values of ΔH, considering the active thermal degradation for the above two methods, were determined as 146.26 and 78.69 kJ/ mol for RSO and ROME, respectively. The values of ΔH are positive, which indicates endothermic reactions. High values of ΔH for RSO infer the high degree of endothermicity.34,35 Li et al. 12 evaluated the average values of ΔH for thermal decompositions of peanut oil (118.54 kJ/mol) and its biodiesel (48.08 kJ/mol) using the application of different kinetics methods. Oliveira et al.33 reported ΔH values for palm (90.53 kJ/mol) and babassu (80.38 kJ/mol) oil biodiesel. Additionally, ΔS and ΔG at To and Tmax for RSO and ROME thermal degradation were evaluated, and the obtained data are presented in Table 5. The negative values of ΔS and positive values of ΔG
fitness of the Kissinger method for ROME thermal decomposition, and the activation energy was found to be 92.5 kJ/mol. Considering all the above methods, the average activation energy for ROME was obtained as 86.65 kJ/mol. The calculated average activation energy is in good agreement with that for sunflower oil methyl ester.13 However, it is higher than that for peanut oil methyl ester (49.71 kJ/mol), waste cooking oil methyl ester (50.07 kJ/mol), and palm oil methyl ester (54.09 kJ/mol).10 This may be due to the fact that the physicochemical-thermal properties of the parent feedstocks and produced biodiesels differ. Order of Reactions for RSO and Its Methyl Ester. Most of the investigations assumed zero-order or first-order reaction for thermal decomposition of oil and biodiesel samples.11 In the present study, the dependency of order of the reaction (n) on temperature for thermal degradation of RSO and ROME was evaluated through the Avrami theory. The regression plots of RSO (stage I and II) and ROME are shown in Figure 8A−C, and the calculated values of order of reaction for thermal degradation of the samples are presented in Table 4. On the basis of the R2 values (Table 4), the Avrami theory is suitable and well fitted to estimate the values of n for thermal degradation of RSO and ROME. As the decomposition temperature increased from 267 to 307 °C (within stage I), the reaction order of RSO first decreased from 1.59 to 1.26 and then increased to 1.42. Further increasing the decomposition temperature for RSO, 402 to 457 °C (within stage II), the value of n decreased from 1.41 to 0.38. The average order of reaction for RSO-I (1.37) was found higher than that of RSO-II (0.92). The signatures of different compounds in FTIR spectra of RSO-I, RSO-II, and ROME also suggest that the values of overall order of the reactions may differ significantly. The values of n for sunflower oil reported by Santos et al.13 vary in the range of 0.95 to 1.82. Similarly, for overall decomposition of RSO, the n value is within the range 0.39 to 1.59. When the decomposition temperature is varied for ROME, the reaction order (n) initially increased from 0.92 (227 °C) to 1.10 (257 °C) and then decreased to 0.19 (317 °C). Li et al.10 and Santos et al.13 reported that the values of order reactions (n) varied from 1.6 to 1.68 (269 °C−277°), 1.77 to 1.95 (257 °C−277 °C), 0.71 to 1.13 (257 °C−277 °C), and 0.69 to 1.89 for peanut oil, palm oil, waste cooking oil, and sunflower methyl esters, respectively.10,13 Similarly, the order of reaction (n) was found to be varied from 0.96 to 1.1 (237−267 °C) for ROME thermal degradation. Furthermore, the average value of the order of reaction (n) for ROME (0.69) is in good agreement with that of soybean (0.5),32 higuereta (0.7),32 babassu (1.4),33 and palm (0.4)33 oil ethyl ester thermal decomposition. Thermodynamic Parameter Calculation for RSO and Its Methyl Esters. In addition to the values of activation energy and order of reaction, important thermodynamic parameters (ΔH, ΔG, and ΔS) for thermal decomposition of RSO and its biodiesel were calculated using eqs 7−10. It can be seen from Figure 8D that all the calculated ΔH values are positives; thermal decompositions of RSO and ROME within the active degradation stages are endothermic processes. Due to the high molecular weight of triglycerides of vegetable oil as compared to that of their fatty acid esters, ΔH for vegetable oil is higher than that of biodiesel,12 similar to E values. The values of ΔH calculated by FWO and MCR methods for RSO within a conversion interval of 0.1 to 0.9 fractions are higher than that of ROME. The values of ΔH by FWO and FRD were found within the range of 69.27−79.34 kJ/mol (RSO-I), 138.39−318.01 kJ/ mol (RSO-II), and 61.76−84.52 kJ/mol (ROME). The average
Table 5. Thermodynamic Parameters of RSO and ROME Samples at To and Tmax for Heating Rate of 10 °C/min sample
temp (°C)
log A (s−1)
ΔH (kJ/ mol)
ΔG (kJ/ mol)
ΔS (kJ/mol K)
RSO-I
To = 223 Tmax = 271 To = 353 Tmax = 416 To = 120 Tmax = 250
4.89 5.67 7.46 11.83 1.01 7.2
+63.75 +78.02 +122.33 +178.11 +22.89 +89.44
+145.01 +159.41 +195.35 +201.37 +115.73 +152.19
−163.82 −149.62 −116.64 −33.76 −236.22 −119.97
RSO-II ROME
confirm that the thermal decompositions of both the samples are nonspontaneous process.10,12,25,33 This was expected since the samples were subjected to forced thermal decomposition by nonisothermal conditions. The values of ΔS obtained at To and Tmax for RSO and ROME thermal degradation are negative, which indicates that the activated complex has a more ordered structure than the reactants, and that the reactions are slower.34−36 The absolute values of ΔS indicate that higher energy is required to reduce the degree of disorder at To as compared to disorder degree at Tmax for RSO-I, RSO-II, and ROME (Table 5). The higher value of ΔG reveals lower favorability of a reaction.10,12,25,37,38 The favorability orders for the thermal degradation process of RSO and ROME were found to be 145.01−201.37 kJ/mol and 115.73−152.19 kJ/mol, respectively. The higher value of ΔG for RSO indicates that a larger amount of heat is required for thermal decomposition as compared to its biodiesel, which is similar to the average activation energy values (E).
■
CONCLUSIONS Thermal decomposition of RSO and its methyl esters (ROME) under an inert atmosphere took place in two and one stage, respectively. The thermal stability of RSO was found to be greater than that of its biodiesel. The rate of maximum weight loss (wmax) was increased from 12.1%/min to 76.7%/min (RSO) and 15.9%/min to 76.2%/min (ROME) as the heating rate increased from 10 °C/min to 50 °C/min. Similarly, the active thermal degradation temperature ranges were also shifted from 223−471 to 264−488 °C for RSO and 120−407 to 193−480 °C for ROME. The values of activation energy (E) and enthalpy (ΔH) of RSO, calculated by several methods (FRD, FWO, MCR, and KM) were found to be greater than that of ROME. The values of the average order of reaction obtained using Avrami theory were found to be 1.14 and 0.69 for RSO and ROME, respectively. The values of order of favorability (ΔG) for thermal decomposition of RSO and ROME were 145.01−201.37 kJ/mol and 115.73−152.19 kJ/mol, respectively. Furthermore, 9650
DOI: 10.1021/acs.energyfuels.7b02249 Energy Fuels 2017, 31, 9642−9651
Article
Energy & Fuels the positive value of ΔG and negative value of ΔS at initial (To) and maximum (Tmax) thermal degradation temperatures indicate that thermal decompositions for RSO and ROME are nonspontaneous processes. The absorbance peaks for alkanes, alkenes, aldehydes, ketones, ethers, water, carbon dioxide, and carbon monoxide were detected in the evolved products. The absence of an absorbance peak for water in ROME degradation shows the quality of produced ester from RSO through transesterification. Thermal decomposition of RSO and ROME after 500 °C (∼1.19 wt % loss for RSO and 1.44 wt % loss for ROME) was mainly due to the formation of ketones and CO2. From TGA-FTIR, it can be concluded that absorbance characteristics for the formation of evolved products such as alkenes, alkane ethers, and aldehydes were at a maximum within a temperature range of 400−500 and 200−350 °C for RSO and ROME, respectively.
■
(9) Zhao, H.; Cao, Y.; Orndorff, W.; Cheng, Y.; Pan, W. J. Therm. Anal. Calorim. 2012, 109, 1145−1150. (10) Li, H.; Niu, S.; Lu, C.; Cheng, S. Energy Convers. Manage. 2015, 98, 81−88. (11) Volli, V.; Purkait, M. K. Fuel 2014, 117, 1010−1019. (12) Li, H.; Niu, S.; Lu, C.; Wang, Y. Energy Fuels 2015, 29, 5145− 5153. (13) Santos, A. G. D.; Caldeira, V. P. S.; Farias, M. F.; Araujo, A. S.; Souza, L. D.; Barros, A. K. J. Therm. Anal. Calorim. 2011, 106, 747−751. (14) Ma, Z.; Chen, D.; Gu, J.; Bao, B.; Zhang, Q. Energy Convers. Manage. 2015, 89, 251−259. (15) Souza, A. G.; Danta, H. J.; Silva, M. C. D.; Santos, I. M. G.; Fernandes, V. J.; Sinfronio, F. S. M.; Teixeira, L. S. G.; Novak, C. J. Therm. Anal. Calorim. 2007, 90, 945−949. (16) Reshad, A. S.; Panjiara, D.; Tiwari, P.; Goud, V. V. J. Cleaner Prod. 2017, 142, 3490−3499. (17) Reshad, A. S.; Tiwari, P.; Goud, V. V. Fuel 2015, 150, 636−644. (18) Friedman, H. L. J. Polym. Sci.: Part C 1964, 6, 183−195. (19) Flynn, J. H.; Wall, L. A. J. Res. Natl. Bur. Stand., Sect. A 1966, 70A, 487−523. (20) Kissinger, H. E. Anal. Chem. 1957, 29, 1702−1706. (21) Ebrahimi-Kahrizsangi, R.; Abbasi, M. H. Trans. Nonferrous Met. Soc. China 2008, 18, 217−221. (22) Niu, S.; Liu, M.; Lu, C.; Li, H.; Huo, M. J. Therm. Anal. Calorim. 2014, 115, 73−79. (23) Niu, S.; Han, K.; Lu, C.; Sun, R. Appl. Energy 2010, 87, 2237− 2242. (24) Vyazovkin, S.; Sbirrazzuoli, N. Anal. Chim. Acta 1997, 355, 175− 180. (25) Kim, Y. S.; Kim, Y. S.; Kim, S. H. Environ. Sci. Technol. 2010, 44, 5313−5317. (26) Sunitha, M.; Reghunadhan Nair, C. P.; Krishnan, K.; Ninan, K. N. Thermochim. Acta 2001, 374, 159−169. (27) Straszko, J.; Olszak-Humienik, M.; Mozejko, J. Thermochim. Acta 1997, 292, 145−150. (28) Pourmortazavi, S. M.; Kohsari, I.; Teimouri, M. B.; Hajimirsadeghi, S. S. Mater. Lett. 2007, 61, 4670−4673. (29) Gai, C.; Dong, Y.; Zhang, T. Bioresour. Technol. 2013, 127, 298− 305. (30) Ko, K.; Rawal, A.; Sahajwalla, V. Energy Convers. Manage. 2014, 86, 154−164. (31) Liu, Z.; Qi, N.; Luan, Y.; Sun, X. Adv. Mater. Sci. Eng. 2015, 2015, 1−8. (32) Rodriguez, R. P.; Sierens, R.; Verhelst, S. J. Therm. Anal. Calorim. 2009, 96, 897−901. (33) Oliveira, L. E.; Giordani, D. S.; Paiva, E. M.; De Castro, H. F.; Da silva, M. L. C. P. J. Therm. Anal. Calorim. 2013, 111, 155−160. (34) Sokoto, M. A.; Singh, R.; Krishna, B. B.; Kumar, J.; Bhaskar, T. Heliyon 2016, 2, e00172. (35) Aravindakshan, K. K.; Muraleedharan, K. Thermochim. Acta 1990, 159, 101−107. (36) Markovska, I. G.; Bogdanov, B.; Nedelchev, N. M.; Gurova, K. M.; Zagorcheva, M. H.; Lyubchev, L. A. J. Chin. Chem. Soc. 2010, 57, 411− 416. (37) Mo, Y.; Zhao, L.; Chen, C.; Tan, G. Y. A. T.; Wang, J. W. J. Therm. Anal. Calorim. 2013, 111, 781−788. (38) Vecchio, S.; Cerretani, L.; Bendini, A.; Chiavaro, E. J. Agric. Food Chem. 2009, 57, 4793−4800.
ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.energyfuels.7b02249. Figures S1−S3 (PDF)
■
AUTHOR INFORMATION
Corresponding Authors
*Tel.: +91 361 2582263/2272. Fax: +91 361 2582291. E-mail:
[email protected]. *Tel.: +91 361 2582263/2272. Fax: +91 361 2582291. E-mail:
[email protected]. ORCID
Pankaj Tiwari: 0000-0003-2578-3462 Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS The authors would like to acknowledge the Central Instruments Facility (CIF) and Center for Energy at Indian Institute of Technology Guwahati (IITG) for providing the characterization facility to conduct the sample analyses. The authors also would like to acknowledge the Center of Excellence on Sustainable Polymers at IITG for providing access to the TGA-FTIR instrument.
■
REFERENCES
(1) Chhetri, A. B.; Tango, M. S.; Budge, S. M.; Watts, K. C.; Islam, M. R. Int. J. Mol. Sci. 2008, 9, 169−180. (2) Atabani, A. E.; Silitonga, A. S.; Ong, H. C.; Mahlia, T. M. I.; Masjuki, H. H.; Badruddin, I. A.; Fayaz, H. Renewable Sustainable Energy Rev. 2013, 18, 211−245. (3) Saez-Bastante, J.; Pinzi, S.; Jimenez-Romero, F. J.; Luque de Castro, M. D.; Priego-Capote, F.; Dorado, M. P. Energy Convers. Manage. 2015, 96, 561−567. (4) No, S. Y. Renewable Sustainable Energy Rev. 2011, 15, 131−149. (5) Ahmad, J.; Yusup, S.; Bokhari, A.; Kamil, R. N. M. Energy Convers. Manage. 2014, 78, 266−275. (6) Reshad, A. S.; Barman, P.; Chaudhari, A. J.; Tiwari, P.; Kulkarni, V.; Goud, V. V.; Sahoo, N. Energy Fuels 2015, 29, 5136−5144. (7) Atabani, A. E.; Badruddin, I. A.; Mahlia, T. M. I.; Masjuki, H. H.; Mofijur, M.; Lee, K. T.; Chong, W. T. Energy Technology 2013, 1, 685− 694. (8) Math, M. C. Energy Sustainable Dev. 2007, 11, 100−104. 9651
DOI: 10.1021/acs.energyfuels.7b02249 Energy Fuels 2017, 31, 9642−9651
