413
Indonesian Journal of Science & Technology 8.
413-428 Indonesian Journal of Science & Technology Journal homepage: http://ejournal.
edu/index.
php/ijost/ Multi-Distributed Activation Energy Model for Pyrolysis of Sugarcane Bagasse: Modelling Strategy and Thermodynamic Characterization Siti Jamilatun1.
Muhammad Aziz2.
Joko Pitoyo1* Department of Chemical Engineering.
Faculty of Industrial Technology.
Universitas Ahmad Dahlan.
Indonesia Institute of Industrial Science.
The University of Tokyo 4-6-1 Komaba.
Meguro-ku.
Tokyo 153-8505.
Japan *Correspondence: E-mail: joko2107054001@webmail.
ABSTRACT
The multi-distributed activation energy model .
ulti-DAEM) is the most effective approach for outlining the kinetics model of biomass pyrolysis.
The purpose of this study is to identify the optimal number and shape of the DAEM for sugarcane bagasse pyrolysis and to discuss its thermodynamic characteristics using the combination of multi-DAEM and differential thermal analysis (DTA).
The heating rate of 10, 30, and 50 CC/min was The results revealed that the multi-DAEM with five pseudo components and Weibull distribution shape gave the lowest relative root mean of the squared error (RRMSE) of 66% and 0.
41%, respectively.
Kinetic and thermodynamic studies showed that the 1st and 4th pseudo components which represent lignin, have activation energy (E.
6 and 180.
kJ/mol, and less endothermic or possibly exothermic Meanwhile, the 2nd, 3rd, and 5th pseudo components which represent cellulose, hemicellulose, and moisture have activation energy (E.
1, 152.
2, and 145.
kJ/mol, respectively, and endothermic properties.
A 2023 Tim Pengembang Jurnal UPI
ARTICLE INFO
Article History:
Submitted/Received 24 Dec 2022
First Revised 12 May 2023
Accepted 05 Jul 2023
First Available online 06 Jul 2023
Publication Date 01 Dec 2023
___________________
Keyword:
Bagasse pyrolysis.
Kinetic.
Multi-DAEM.
TG-DTA.
Thermodynamic.
Jamilatun et al.
Multi-Distributed Activation Energy Model for BagasseA 414
INTRODUCTION
One of Indonesia's major sources of farming waste is bagasse.
(Pradana et al.
Bagasse can be used as fuel in boilers and power generators (Ordonez-Loza et al.
It is a lignocellulosic biomass that Several processes .
ncluding pyrolysi.
can convert into fuel.
(Jamilatun et al.
, 2022.
Pradana et al.
, 2.
Pyrolysis is a thermal decomposition without an oxidizing agent .
ir or oxyge.
(Guedes et al.
, 2018.
Pitoyo et al.
, 2.
It is considered a superior conversions because of its versatility in choosing raw materials, wider temperature operational .
-600 CC), possible operation at atmospheric pressure, and its ability to produce three valuable products .
olid, liquid, and ga.
simultaneously (Jamilatun et , 2019.
Terry et al.
, 2.
Several research studies have been carried out related to the mechanism, operating parameters, and kinetics model of pyrolysis (Hameed et al.
, 2019.
Kaczor et al.
, 2020.
Wang et al.
, 2.
There are two fundamental mathematical procedures to experimentally determine the kinetics of biomass pyrolysis's parameters:
model-free and model-fitting methods (Cai et al.
, 2.
The former method, also called the iso-conventional method, assumes that the conversion value affects kinetic parameters like the frequency factor and activation energy (Aboyade et al.
, 2.
This method includes Miura differential method.
Miura-Maki integral method.
Coats-Redfern.
Flynn-Wall-Ozawa.
Kissinger, and Kissinger-Akahira-Sunose.
This method is easier because it only requires linear regression (Bonilla et al.
, 2019.
Sukarni.
Zhao et al.
, 2.
However, it has several areas for improvement, such as requiring a minimum of three experiments with different heating rates and not being suitable for multiple reactions (Vyazovkin et al.
, 2.
Sometimes it is difficult to find conversion derivatives to the activation energy due to significant variations in conversion to the activation energy (Cai et al.
, 2.
The model-fitting method can be grouped into single-reaction and multi-reaction Multi-reaction models include the lumped kinetic model and DAEM.
The lumped kinetic model assumes several parallel reactions, each with their individual activation energy.
At the same time.
DAEM assumes that the decomposition mechanism involves multiple independent parallel reactions with various activation energies .
ee https://w.
edu/sis/centers/car bon/removal.
Sonobe & Worasuwannarak.
Vyazovkin et al.
, 2.
DAEM explains the kinetics of biomass pyrolysis, the mechanism of thermally degradable materials, and complex chemical systems such as coal pyrolysis (Quan et al.
However, the model-fitting method is weak, as the obtained kinetic parameters provide accurate data fitting at only one heating rate (Vyrhegyi et al.
, 2.
The distributional shape of DAEM describes different behavior and kinetic The exact shape of the distribution of activation energy is unknown.
Its shape can be grouped into two types, symmetrical and asymmetrical.
Symmetric distribution shapes include Gaussian.
Gumbel.
Cauchy, and Logistic, with the Gaussian distribution being the most widely applied (Dhaundiyal & Singh, 2016.
Tran et , 2.
However, the Gaussian distribution has a weakness as it is symmetrical, while the actual distribution is asymmetrical (Burnham & Braun, 1.
This makes asymmetric distributions such as the Weibull and Gamma distributions more attractive to be implemented (Cai & Liu.
Kuo-Chao et al.
, 2.
In addition to the shape of the distribution, the number of distributions or pseudo-components is an important factor.
It determines the accuracy of a simulation and the complexity of Too few distributions reduce DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 415 | Indonesian Journal of Science & Technology.
Volume 8 Issue 3.
December 2023 Hal 413-428 the accuracy of a model, while too many require high computational costs.
Three distribution numbers have been broadly employed to describe the pyrolysis of biomass (Cai et al.
, 2.
The use of a distribution number of four or five has also been employed to describe the thermal decomposition of plastic waste, marine biomass, and also lignocellulosic biomass (Burra & Gupta, 2018.
Kristanto et al.
, 2021.
Lin et al.
, 2.
Choosing the correct number and shape of the model can simplify the complexity of the calculations while increasing the prediction ability and accuracy of the simulations performed.
Thermogravimetric differential thermal analysis (TG-DTA) is an analysis method generally used to study the thermal decomposition of biomass (Viju et al.
, 2.
TG-DTA can simultaneously calculate weight change and differential heat flow as a function of time and temperature.
DTA
measurement is based on the difference between the reference and sample The properties of the pseudo component have a significant correlation with the reaction enthalpy profile (Kristanto et al.
, 2.
determination of the number and shape of DAEM is rarely investigated, while the use of the relationship between multi-DAEM and DTA profile to explain the thermodynamic properties of the involved pseudo components has never been performed.
This paper aims to determine the number and shape of the best DAEM that can accurately describe the decomposition process in bagasse pyrolysis and, at the same time, discuss the thermodynamic properties of each pseudo component based on the relationship between multiDAEM and DTA profile.
METHODS
Materials Bagasse was obtained from PT Madukismo Yogyakarta.
Samples were washed to remove the impurities, then oven dried for 24 h.
After drying, the sample was crushed and sieved to get a grain size of 60 Table 1 lists the results of the proximate, ultimate, and compositional Thermogravimetric Pyrolysis Pyrolysis was carried out using a simultaneous TG-DTA Hitachi STA-200RV at atmospheric pressure in an inert nitrogen The sample is weighed as much as 6-10 mg and placed in a platinum pan, a small sample weight .
-10 m.
is taken to reduce mass and heat transfer barriers ( Y.
Lin et al.
, 2009.
Vyrhegyi et al.
, 2.
Constant heating rates of 10, 30, and 50 CC/min were utilized to heat the sample from 30 to 900CC.
N2 gas with high purity .
99%) is flown at 100 mL/min to obtain an inert condition.
In TG-DTA, there is a thermal lag at high heating rates between the thermocouple reading and the sample's actual temperature.
Therefore 10 CC/min heating rate was selected to evaluate the thermodynamic properties based on the DTA profile to minimize the occurrence of thermal lag (Jr & Grynli, 2.
Table 1.
Proximate and elemental analysis of bagasse.
Ultimate Value .
Proximate Moisture Volatile matter Fixed carbon Ash Value .
Composition Cellulose Hemicellulose Lignin Value .
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 Jamilatun et al.
Multi-Distributed Activation Energy Model for BagasseA 416 Kinetic Modeling Kinetic modeling is important to determine the optimal kinetic parameters, pre-exponential factor (A), and activation energy (E).
In this study, the nth-order reaction model and DAEM were used.
The nth-order reaction model is shown by Equation .
T ya yu(T) = 1 Oe .
Oe .
Oe yc.
OT0 exp (Oe yu ya ) yccycN] ycIycN .
Oeyc.
where yu is the heating rate.
T is the temperature at time t.
T0 is the initial temperature.
R is the universal gas constant, and n is the reaction order.
Moreover, the degree of conversion is formulated by Equation .
yco Oeyco yu(T) = ycoycn OeycoycN yce where mi is the initial mass of the sample, mT is the mass of the sample at temperature T, and mf is the final mass of the sample.
The equation for single-DAEM is shown in Equation .
T ya ya yu(T) = 1 Oe O0 exp [Oe OT0 yu exp (Oe ycIycN) yccycN ] yce.
Biomass is a complex chemical material consisting of multiple pseudo components that do not interact with each other during the thermal decomposition process.
multi-DAEM is needed to accurately describe the decomposition process (Vyazovkin et al.
, 2.
To explain the contribution of each pseudo component, or contribution factor.
cj was introduced (Kristanto et al.
, 2.
, and the equation for multi-DAEM is shown in Equation .
T ya ya yu ycIycN yu(T) = 1 Oe OcycAycc exp (Oe j=1 ycayc O0 exp [Oe OT0 ) yccycN ] yceyc .
yccya Here Nd is the number of distributions, and fj(E) is the distribution function or the shape of DAEM.
In this study, seven shapes of DAEM were used: Gaussian.
Logistic.
Gumbel.
Cauchy.
Weibull.
Gamma, and Rayleigh.
Table 2 shows the distribution function of each DAEM.
There is no exact analytical solution to the problem since it has double integrals, inner integrals dT, and outer integrals dE.
(Mcguinness et al.
, 1999.
yerfyo, 2007.
Tran et al.
, 2.
In this study, the temperature integral estimation suggested by Cai et al.
was applied to solve the inner integral dT and the trapezoidal integration rule to solve the outer integral dE.
The pre-exponential factor (A.
was determined by setting the initial optimization value close to the values in the literature, namely 1014.
13, 1013.
71, 1013.
90, and 67x1013/s for hemicellulose, cellulose, respectively (Vyrhegyi et al.
, 2.
The upper limit of outer integral dE is 500 kJ/mol (Gyne & Gyne, 1.
Table 2.
Type of distribution function used in this research.
Distributions Gaussian Logistic Gumbel Cauchy Weibull Gamma Rayleigh Distribution Function, f(E) .
aOeya0 )2 exp [Oe yuaOo2A 2yua 2 .
aOeya0 ) yce.
= sech 4yua 2yua yaOeya0 yaOeya yce.
= exp [Oe ( exp {Oe ( .
})] yua yua yua Oe1 yaOeya0 2 yce(E) = .
( ) ] Ayua yua ycoOe1 yco ya ya yco yce(E) = yce.
= ( ) exp [Oe ( ) ] yuI yuI yuI yaycoOe1 ya yce(E) = exp (Oe ) yu.
yuEyco yuE yce.
= ya ya2 exp (Oe .
yua2 2yua Mean value of E Standard deviation of E ya0 ya0 ya0 ya0 yco yco yco yuIyu .
) yuI2 .
) Oe .
)} ] ycoyuE ycoyuE 2 yuaOo yuU 4OeyuU ) yua2 DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 417 | Indonesian Journal of Science & Technology.
Volume 8 Issue 3.
December 2023 Hal 413-428 The objective function in Equation .
was minimized to achieve the kinetic The objective function was determined by summing the square errors between the experimental and the DAEM simulation data.
The objective function was minimized using the Matlab software package, i.
, the fmincon function using the sequential quadratic programming (SQP) This optimization used a maximum iteration of 1000 and a maximum function evaluation of 5000 for all distributed activation energy models (DAEM).
The range of values used in the literature is the basis for the optimization's constraints (Vyrhegyi et al.
, 2.
The value of the E0 was taken from 100 to 350 kJ/mol, the standard deviation (E) was taken from 0 to 70 kJ/mol, the pre-exponential factor (A.
was selected from 1010 to 1020/s, and the contribution factor .
was selected from 0 ycuycy SSE = Oci=1.
uyceycuycy,ycn (T) Oe yuycoycuyccyceyco,ycn (T)] .
exp,i and model,i are the degree of conversion of the experimental data and the model, respectively, and np is the number of data points.
The quality of fitting was evaluated using the relative root mean of the square error (RRMSE) and the coefficient of determination (R.
, as shown in Equations .
(Feng et al.
, 2.
RRMSE =
1 ycuycy Oc .
u (T)Oeyuycoycuyccyceyco,ycn(T)] ycuycy i=1 yceycuycy,ycn Oo I yceycuycy,ycn yu y 100% ycuycy R2 = 1 Oe Oci=1.
uyceycuycy,ycn (T)Oeyuycoycuyccyceyco,ycn (T)] ycuycy Oci=1.
uyceycuycy,ycn (T)Oeyu I yceycuycy,ycn (T)] .
where yuIyceycuycy,ycn (T) is the average value of the experimental data.
Figure 1 shows the optimization algorithm to determine the optimal kinetic parameters.
Figure 1.
Kinetic parameters calculation algorithm.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 Jamilatun et al.
Multi-Distributed Activation Energy Model for BagasseA 418
RESULTS AND DISCUSSION
TG-DTA of Bagasse Figure 2 presents the results of the TGDTA analysis of bagasse at a heating rate of 10, 30, and 50 CC/min.
The heating rate dramatically affects the decomposition Considering the TGA's characteristics, the high heating rate reduced the time needed for decomposition 27 min at a heating rate of 10 CC/min to 17.
98 and 12.
09 min at 30 and 50 CC/min, respectively.
On the other hand, a high heating rate reduced the total solids conversion from 97.
14% at a heating rate of 10 CC/min to 89.
83 and 91.
48% at a heating rate of 30 and 50 CC/min, respectively.
This is likely because the high heating rate triggers a secondary decomposition reaction that converts volatile materials into char (Guedes et al.
, 2.
Based on the characteristics of DTG (Figure 2.
, the peak temperature was moved to the right by increasing the heating rate from 332.
58CC at a heating rate of 10CC/min to 350.
87 and 17CC at heating rates of 30 and 50 CC/min, respectively.
Based on the characteristics of the DTA, the high heating rate causes the loss of the peak appearance on the DTA curve.
This is probably brought on by the thermal delay between the sample temperature and thermocouple measuring (Jr & Grynli, 2.
This suggests the use of low heating rates in studying biomass decomposition kinetics to avoid the loss of peaks on the DTA or DTG curves.
The DTG curve shows a peak at 332.
58 CC, a shoulder at 300 CC, and tailings at a temperature range of 365-550 CC.
The peak is associated with cellulose decomposition, the shoulder decomposition, and the tailings at the end of the pyrolysis temperature are associated with lignin decomposition.
Figure 2.
The curve of several parameters: .
DTG at different temperatures, .
TG-DTA at a heating rate of 10 CC/min, .
TG-DTA at a heating rate of 20 CC/min, and .
TG-DTA at a heating rate of 30 CC/min.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 419 | Indonesian Journal of Science & Technology.
Volume 8 Issue 3.
December 2023 Hal 413-428 Determination Number of Distribution Figure 3 depicts how the number of distributions has an impact on fitting quality using five distribution numbers Gaussian.
One Gaussian shows the most significant deviation among the five distribution numbers, and this is because of the limitations of one Gaussian in describing the multiple reactions that occur in biomass Figure 3a shows that two, three, four, and five Gaussian provide overlapping data fittings.
From Figure 3b, it can be seen clearly that the five Gaussians show the lowest RRMSE.
Table 3 shows that using five Gaussian gives one distribution .
he 3rd pseudo componen.
with a reasonably low contribution factor.
Using a distribution number of more than five will result in several pseudo components with a relatively low contribution factor, which has no significant role in the reaction, and instead results in high computational complexity.
Thus, it is not optimal.
From the figure, moreover, it can be seen that the number of peaks on the DTG curve can indicate the number of distributions which can be used in the DAEM (Kristanto et al.
, 2.
Determination Distribution Shape The The evaluation was performed on various shapes of DAEM such as Gaussian.
Logistic.
Gumbel.
Cauchy.
Weibull.
Gamma.
Rayleigh, and Reaction orders .
on-DAEM) to obtain optimal conditions .
roviding a total RRMSE is minimu.
Figure 4 shows that the Gamma.
Rayleigh, and Reaction order models provide a high value of total RRMSE.
Meanwhile, the Gaussian.
Logistic.
Gumbel.
Cauchy, and Weibull models provide accurate results, as shown by the overlap of the graphs in Figure 4a and the low total RRMSE in Figure 4b.
Table 3 shows a good fitting quality of the five models with a coefficient of determination Ou 0.
Weibull best fits the five models (Kuo-Chao et al.
, 2.
characterized by the lowest RRMSE value.
The reason is that the actual energy distribution in the biomass' thermal decomposition is asymmetrical, especially during the initial and final stages.
Weibull model is an asymmetrical distribution model, so it can provide good fitting, especially at the beginning and end of the thermal decomposition (Cai & Liu, 2007.
Kuo-Chao et al.
, 2.
, as shown in Figure 6.
Figure 3.
Identification of different numbers of Gaussian DAEM: .
fitting data model and .
calculated RRMSE.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 Jamilatun et al.
Multi-Distributed Activation Energy Model for BagasseA 420 Table 3.
Kinetic and statistic parameters for bagasse pyrolysis.
Shape of DAEM Weibull Gaussian Logistic Gumbel Cauchy Kinetics RRMSE (%) RRMSE (%) RRMSE (%) RRMSE (%) RRMSE (%) .
Pseudo component .
Figure 4.
Identification of different shapes of multi-DAEM: .
fitting data model and .
calculated RRMSE.
Sensitivity Analysis of DAEM Kinetic Parameter Figure 5 reveals the local sensitivity analysis of the parameters obtained from DAEM.
Local sensitivity analysis is evaluated on specific parameters (Sciacovelli & Verda.
Sensitivity analysis is applied to assess the parameters' robustness in different input data.
The x-axis shows deviations from optimal parameters, the yaxis shows the kinetic parameters of DAEM, and the different colors in contours show DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 421 | Indonesian Journal of Science & Technology.
Volume 8 Issue 3.
December 2023 Hal 413-428 the RRMSE values.
At the same value of x, the closer to the yellow area the higher the RRMSE value and the more sensitive.
It can be seen from Figure 5 that the standard deviation (E) is the most sensitive parameter among the existing parameters, marked by the number of contour areas with the yellow Meanwhile, the activation energy (E.
and pre-exponential factor (A) have relatively excellent robustness.
The E0 and A not sensitive when deviated from the optimal value indicated by the low value of RRMSE.
Kinetic Study of Bagasse Pyrolysis Figure 6 shows the DTG curve formed from the differentiation of experimental TGA data and the DAEM simulation using five pseudo components.
Matching the number and shape of peaks between the DTG experiment and the DAEM simulation is needed to accurately describe the kinetics of the decomposition reaction (Kristanto et al.
It can be seen from Figure 6 that all DAEMs exhibit four major pseudo components .
he 1st, 2nd, 4th, and 5th pseudo component.
and one minor pseudo component .
he 3rd pseudo componen.
Based on the Weibull model, at a temperature range of less than 250 CC, a minor pseudo component .
rd pseudo componen.
was decomposed with an E0 of 60 kJ/mol and a contributing factor of 0101, which probably represents the decomposition of bound moisture and light In the range of 215-325CC, the 5th pseudo component decomposition occurred with an E0 of 152.
25 kJ/mol, a standard deviation of 19 kJ/mol, and a contributing factor of 17, which probably represents the hemicellulose decomposition.
In the range of 245-370CC, the 2nd pseudo component decomposition occurred with an E0 of 06 kJ/mol, a standard deviation of 0.
kJ/mol, and a contributing factor of 0.
which probably represents the cellulose In the range of 280-525 CC, the 1st and 4th pseudo component decomposition occurred with E0 of 189.
61 and 181.
16 kJ/mol, standard deviations (E) of 26.
72 and 60.
kJ/mol, and contributing factors of 0.
1375, respectively.
This represents the decomposition of lignin and char.
The appearance of several pseudo components in lignin decomposition was also reported in a previous study (Kristanto et al.
, 2.
Figure 7 shows the activation energy distribution of the five pseudo components during bagasse pyrolysis.
The order of appearance of the peaks in the figure corresponds to the appearance of the pseudo components on the DTG curve.
From the figure, the narrowest activation energy distribution range is seen in the 2nd pseudo component as reported by Huber et al.
, with an activation energy of 172-178 kJ/mol and a standard deviation of 0.
85 kJ/mol.
This range is included in cellulose's activation energy distribution range (Quan et al.
The distribution of the 5th pseudo componentAos activation energy is more comprehensive than that of the 2nd pseudo component, with an activation energy distribution of 146-155 kJ/mol and a standard deviation of 1.
19 kJ/mol, belonging to the hemicellulose distribution range (J.
Zhang et al.
, 2.
At last, the 1st and 4th pseudo components have the widest distribution of the activation energy, in the range of 155200 kJ/mol, and standard deviations of 71 and 60.
30 kJ/mol .
This indicates that the two components have a complex structure, and the decomposition arises over a broad temperature range, as in lignin decomposition (Jiang et al.
, 2010.
Wang et , 2.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 Jamilatun et al.
Multi-Distributed Activation Energy Model for BagasseA 422 Figure 5.
The local sensitivity analysis of multi-DAEM.
Figure 6.
Experimental and simulation DTG of bagasse with multi-DAEM at heating rate of 10 CC/min using: .
Gaussian, .
Logistic, .
Gumbel, .
Cauchy, .
Weibull, and .
Weibull distribution at 30 CC/min.
Figure 7.
Activation energy distribution of multi-DAEM for pyrolysis of bagasse.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 423 | Indonesian Journal of Science & Technology.
Volume 8 Issue 3.
December 2023 Hal 413-428 Thermodynamic characterization The DTA instrument can determine heat flow in the reaction based on the temperature difference between the sample and reference for a fixed amount of heat input (Zhang et al.
, 2.
The sample temperature remains constant for the endothermic reaction, so the heat flow (DTA) value is higher, whereas, in the exothermic reaction, the heat flow (DTA) value is lower.
Using the relationship between DTA and the distribution of pseudo components in DTG, it is possible to understand the thermodynamic properties of each pseudo component and biomass decomposition behavior.
Figure 8 shows the relationship between TGA.
DTA, and multi-DAEM simulation, as also several stages for thermodynamic Each stage shows different temperature ranges, conversions, pseudocomponent contributions, heat flow, and thermodynamic properties, as shown in Table 4.
At the temperature of 200-250 CC or stage I of 3.
7 V, a slight increase in the DTG curve is accompanied by a slight increase in DTA value, representing the decomposition reaction of the 3rd pseudo component.
Figure 8.
Comparison of TGA.
DTA, and calculated DTG.
Table 4.
Thermodynamic characterization of bagasse pyrolysis.
Pseudo component Contribution (%)** Stage Temperature (CC) Conversion (%) i VII Vi 2nd pseudo: 11.
5th pseudo: 88.
2nd pseudo: 29.
5th pseudo: 70.
2nd pseudo: 74.
5th pseudo: 22.
4th pseudo: 3.
2nd pseudo: 95.
4th pseudo: 4.
2nd pseudo: 76.
4th pseudo: 23.
2nd pseudo: 90.
4th pseudo: 9.
1st pseudo: 5.
4th pseudo: 94.
1st pseudo: 73.
4th pseudo: 26.
1st pseudo: 100.
1st pseudo: 100.
Heat flow (V ) Thermodynamic 31-58-35.
More endothermic Less endothermic/ possibly exothermic Less endothermic/ possibly exothermic Exothermic Endothermic Less endothermic/ possibly exothermic Note: **At the final temperature of each stage DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 Jamilatun et al.
Multi-Distributed Activation Energy Model for BagasseA 424 This reaction may be related to the dehydration of active cellulose or bound moisture, which is endothermic.
A sharp increase in DTA value of 56.
03 V occurred at a temperature range of 250-330 CC or stage II-IV, after the presence of the 2nd pseudo component.
This indicates an decomposition of the 2 and 5 pseudo components representing cellulose and hemicellulose, respectively.
The higher 2nd pseudo component contribution than the 5th reveals the more endothermicity of the 2nd pseudo component.
The endothermicity of the 2nd pseudo component is related to the depolymerization of cellulose.
There is a fluctuation in the DTA value between 352550 CC or stages VI-IX.
This indicates that the decomposition of the 1st and 4th pseudo exothermic and endothermic reactions (Kristanto et al.
, 2021.
Yang et al.
, 2.
with the exothermic tendency of the 4th pseudo component.
Thermodynamic parameters such as A.
E0, enthalpy (OIHA), entropy (OISA), and Gibbs free energy (OIGA) are important for understanding the behavior of a chemical or physical process (Khajehzadeh et al.
, 2.
The thermodynamic parameter is obtained using the equation provided by Kim et al.
(Kim et al.
, 2.
at the peak temperature of each pseudo component because that temperature gives the highest reaction rate (Aamer et al.
, 2.
A high A value improves both the reaction rate and the frequency of molecular collisions.
The E0 and OIHA values indicate the minimum energy needed for a reaction and the low E0 and OIHA values increase the reaction rate.
A high OISA indicates a high degree of disorder which has implications for increasing spontaneous reactions, high reactivity, and increasing reaction rates.
Meanwhile, the high OIGA decreases the spontaneous reaction.
Thermodynamic parameters of the bagasse pyrolysis are summarized in Table 5.
Based on Table 5, the 3rd pseudo component has the lowest E0.
OIHA, and OIGA values and the highest A and OISA values.
hence, it has a high tendency for the reaction to occur spontaneously.
The 5th pseudo component, which represents hemicellulose, has relatively low E0.
OIHA, and OIGA values and relatively high A and OISA Therefore, it has a relatively high tendency for spontaneous reactions to occur but is still weaker than the 3rd pseudo The 2nd pseudo component, which represents cellulose, has relatively low E0.
OIHA, and OIGA values, while A and OISA values are relatively high.
Hence, it has a fairly high tendency for spontaneous reactions to occur but is still weaker than the 3rd and 5th pseudo components.
Meanwhile, the 1st and 4th pseudo components, which represent lignin, have high E0.
OIHA, and OIGA values but low A and OISA values, so spontaneous reactions have a low tendency to occur (Xu & Chen, 2.
Table 5.
Thermodynamic parameters for bagasse pyrolysis.
Pseudo E0 .
J/mo.
A .
OIHA .
J/mo.
OISA .
J/mo.
OIGA .
J/mo.
09594E 13
03117E 15
30756E 15
08759E 14
04291E 14
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 425 | Indonesian Journal of Science & Technology.
Volume 8 Issue 3.
December 2023 Hal 413-428
CONCLUSION
TG-DTA pyrolysis of bagasse has been investigated using a DAEM to determine the optimal number and shape of DAEM.
The combination of DAEM and DTA can be used to study the thermodynamic properties of bagasse pyrolysis.
The results show that the multi-DAEM with five pseudo components gave the lowest RRMSE of 0.
Based on the shape of the multi-DAEM, the Weibull distribution gives the lowest average RRSME value of 0.
Based on the kinetic and thermodynamic studies, the 1st and 4th pseudo components have E0 of 189.
6 and 6 kJ/mol and OIGA of 191.
7 and 169.
kJ/mol, representing lignin decomposition.
The 2nd pseudo component represents cellulose with an E0 of 176.
1 kJ/mol and OIGA 9 kJ/mol.
The 5th pseudo component represents hemicellulose with an E0 of 152.
kJ/mol and OIGA of 142.
1 kJ/mol.
The 3rd pseudo component represents the bound moisture or light volatile with an E0 of 145.
kJ/mol and OIGA of 125.
6 kJ/mol.
The combination of multi-DAEM and DTA indicates that the thermal decomposition reactions in the 2nd, 3rd, and 5th pseudo components are endothermic, the 1st pseudo component is exothermic, and the 4th pseudo component is endothermic or possibly exothermic.
ACKNOWLEDGMENT
The author is grateful for the research grant provided by "The Directorate of Research.
Technology, and Community Service from the Ministry of Education.
Culture.
Research, and Technology" as part of the National Competitive Basic Research (PDKN): plan for the Fiscal Year 2022.
Number 001/PB.
PDKN /BRIn.
LPPM/VI/2022.
AUTHORSAo NOTE The authors declare that there is no conflict of interest regarding the publication of this article.
The authors confirmed that the paper was free of plagiarism.
REFERENCES