Optimization of Microwave-Assisted Alkali Pretreatment for Enhancement of Delignification Process of Cocoa Pod Husk

In this study, the optimization of microwave-assisted alkaline (MAA) pretreatment is performed to attain the optimal operating parameters for the delignification of cocoa pod husk (CPH). The MAA performance was examined by heating the CPH solid with different particle sizes (60–120 mesh) and NaOH solution with a different sample to a solvent (SS) ratio (0.02–0.05 g/L), for short irradiation time (1–4 min). Box-Behnken Design (BBD) was utilized to optimize the percentage of lignocellulose composition changes. The results show that by enlarging particle size, the content of lignin and cellulose decreased while hemicellulose increased. By prolong irradiation time, the content of lignin and hemicellulose decreased while cellulose elevated. On the other hand, increasing the SS ratio was not significant for hemicellulose content changes. From FTIR and SEM characterization, the MAA drove the removal of lignin and hemicellulose of CPH and increased cellulose slightly. Supported by kinetic study which conducted in this work, it was exhibited that MAA pretreatment technology is an effective delignification method of CPH which can tackle the bottleneck of its commercial biofuel production. Copyright © 2021 by Authors, Published by BCREC Group. This is an open access article under the CC BY-SA License (https://creativecommons.org/licenses/by-sa/4.0).


Introduction
Lignocellulosic biomass is considered as the most promising raw material to produce green energy to reduce dependence on fossil fuels [1]. Conversion of the abundant, renewable, and non-edible lignocellulosic biomass into energy 10.94%, and 11.64%, respectively [5]. Despite having potential, however, the rebellious nature of lignocellulosic biomass harshly hinders the further biological process, namely enzymatic hydrolysis and fermentation, becoming a significant drawback of lignocellulose utilization for biofuel production [6]. The development of cost-effective technology for effective conversion is urgently needed for the biofuel commercialization [7].
The pretreatment process is crucial to damage the complex structure of lignin, facilitating the release of monomeric sugars in biochemical conversions into bioenergy products [8][9][10][11]. Some pretreatment processes, for example, physical and chemical, Physico-chemical, biological processes, have been developed previously or combinations of these processes [12]. Each of these processes has advantages and limitations that depend on the nature of the biomass [13]. Among these methods, alkaline pretreatment exhibited a more effective and beneficial delignification performance because it uses inexpensive and widely available chemicals, compared to other chemicals such as deep eutectic solvents (DES) and ionic liquid (IL) [6,9,12,14]. However, the long delignification time is a drawback in this method. The application of microwaves has considered an alternative to conventional heating of alkali pretreatment [15].
Microwave-assisted delignification has been demonstrated to be a promising method with maximum removal of lignin and less time processing required [8,16]. Microwave facilitated indirect absorption of energy by molecules, increased heat transfer, and accelerated the rupture of biomass structure through the molecular collision triggered by the dielectric polarization. Polar particle movement produces fast heating resulting in lower process temperatures and times [17]. Despite some previous works using microwave heating for biomass pretreatment has been reported, the application of alkaline in microwave heating for lignin fractionation and depolymerization was still limited with a different operational condition for the different varieties of lignocellulosic feedstock. From literature studies, there are no works reported before about the combination of the microwave with the chemical process as a potential method for the delignification of CPH.
From the problems mentioned above, in this study, the optimization of microwave-assisted alkaline pretreatment was carried out to obtain the optimal operating condition for enhancing the delignification of CPH. The physicochemical properties were conducted by using SEM and FTIR analysis. It was hypothesized that the alkaline pretreatment combined with microwave irradiation gives a higher efficiency of lignin depolymerization than conventional heating.

Materials
CPH was collected from the plantation area in Jember District, East Java, Indonesia. It was firstly dried under the sunlight for 2 days, then milled and screened to obtain a particle size of 60−120 mesh. Analytical grade of natrium hydroxide (NaOH) was purchased from Sigma Aldrich, Japan.

Microwave-Assisted Alkaline (MAA) Pretreatment
The MAA process was run by mixing 5 g of cocoa pod husk and 3% NaOH solution with a sample to a solvent ratio of 0.02-0.05 g/L in the beaker glass. Then, the mixture was heated in the microwave (100 W) during particular irradiation time, 1−4 min. The same process was conducted for the different particle sizes of 60-120 mesh. After the pretreatment process, the pretreated solid was washed, filtered, and dried in the oven at 45 °C for 1 day. Finally, the sample was stored at 4 °C before analyzed. The process was run under batch operation in the triplicate experiment.

Analytical Methods
The compositions of lignin, hemicellulose, and cellulose at the untreated and MAAtreated conditions were analyzed using the Chesson method [18]. The solid of native and MAA treated were characterized by Scanning Electron Microscopy (SEM) (Evo MA 19, Carl Zeiss, England) analysis and Fourier Transform Infrared Spectroscopy (FTIR) (Nicolet iS10, Thermo Scientific, USA) analysis. The analysis was employed to investigate chemical composition changes after the MAA pretreatment applied.

Design of experiment
In this study, a Box-Behnken design (BBD) was utilized to optimize the operating conditions of MAA delignification towards the percentage of lignocellulose composition changes. gated. Experimental design of BBD of MAA pretreatment was shown in Table 1. According to statistical modeling generated by Design Expert 10 software (Stat-Ease), 17 runs were required to analyze the data. Each run was carried out in duplicate and data was showed in average. Table 1 shows the experimental design of this work with a randomized run. The second-order nonlinear polynomial equation was fitted for the experimental data as Eq. (1) follows: (1) where Y denotes the lignocellulose percentage as response, 0 denotes the intercept term, n denotes the number of variables, i denotes the liner effect term, ii means the square effect term, ij means the interaction effect term, Xi and Xj means the independent variables, and ε means the random error.

Kinetics analysis
The kinetics of the lignocellulosic components conversion was modelled as the equation of reaction rate. The reaction rate was expressed as the simple function of substrate concentration (CA) with the certain reaction rate constant and reaction order as the following Eq. (2). (2) The order of the reaction and the reaction the reaction rate constant was determined by frac-tional time method. The data was plotted as graphic and the required time to remain 80% (t0.8) substrate was observed from the plot for each concentration data. The fractional time was defined as the following Eq. (3): (3) Then, the required data was plotted linearly with log tF-as the y-axis and log CA0 as the xaxis. (4) From the linearized plot as Eq. (4), the value of reaction order and reaction rate constant was calculated.

Experimental Design and Statistical Analysis
Statistical analysis for the regression model of lignin, cellulose, and hemicellulose using Box-Behnken Design is shown in Tables 2-4. From the ANOVA result of the regression model of lignin (Table 2), there is a 72.17% chance of data could occur due to noise. The model pvalue of 0.7217 indicates that the model was statistically insignificant and demonstrates the unreliability of the models. As shown in Table  2 Table 4. Analysis of variance for the regression model of cellulose content.

Bulletin of Chemical Reaction Engineering & Catalysis, 16 (1), 2021, 35
Copyright © 2021, ISSN 1978-2993 AC, BC), and quadratic coefficients (A 2 , B 2 , C 2 ), had no significant effect (p > 0.05). The lack of fit p-value of 0.8017 means the lack of fit is not significant relative to the pure error. As shown in Tables 3 and 4 for ANOVA of hemicellulose and cellulose content, the Model p-value of 0.0703 and 0.098, respectively, revealed that model terms are insignificant when the significance level of 95% was used. However, when the significance level of 90% was applied, the terms are significant. In Table  3, the variable of exposure time (C) is a significant model term. In Table 4, the quadratic term of ratio (B²) is a significant model term. The lack of fit p-value of 0.7954 and 0.9206 demonstrated that the lack of fit of hemicellulose and cellulose, respectively, is not significant relative to the pure error. The R-squared of the hemicellulose and cellulose model are 0.8039 and 0.7796, respectively. They indicate that the models have a quite good agreement with the experimental data. It also represents that the models are acceptable and can satisfy 80.39% and 77.96%, respectively, of the variability in the delignification process.

Fitting of Models
In this work, a second-order polynomial mathematical equation presented the relationship between the independent variables and the responses. The developed model equations are provided in Equations (5-7), where A represents particle size (mesh), B represents the ratio (g/L), and C represents time (min).

Diagnostic of Model Adequacy
The percentage of normal probability plot of residuals was typically scattered near along the straight line, which denotes a tolerable devia-tion of variance. A good agreement between the model and the experimental data observed can be qualitatively justified by using a Predicted vs. Actual plot (parity plot) [19]. The accuracy of the model can be understood by evaluating the normal distribution with the predicted and actual results, as exhibited in Figure 1 In microwave heating, different from the heat transfer concept of conventional heating, heat transfer was induced throughout with more selectivity to higher polar parts. This condition generates hotspots within the complex structures of lignocellulosic biomass [20].
For the ratio effect parameter, lignin content decreases with increasing ratio when interacting feed to volume ratio variables (see Figure 2a). Conversely, lignin content increases with increasing ratio when interacting with time variables, as shown in Figure 2c. It was revealed that the delignification process was strongly affected by solvent volume. The delignification increased linearly as the increase of liquid to the solid ratio (LSR). This phenomenon may be attributed to the larger of the biomass swelling in larger solvent volume. This condition facilitates the absorption of microwaves straightly on the matrix, leading to degradation of the lignin compound [16]. The large portion of lignin content underwent the delignification process by breaking of aryl ether linkages after being subjected by alkaline solution [21]. Furthermore, hydrogen bonds that connect between lignin with cellulose and hemicellulose were ruptured. Sodium and hydroxide ions are formed during the process, then the ions strike ester-carbohydrate bonds in the lignocellulose [22]. Increasing the concentration of alkaline to a certain level may have a positive impact on the swelling of biomass. This swelling increases the surface area of the complex nature of biomass. The instability of holocellulose at a higher alkaline concentration in microwave heating may be caused by the dissolution of polysaccharides and lignin simultaneously [1]. However, a further increase in ratio had no significant enhancement during delignification. This phenomenon may be due to the difficulty of attaining the activation energy as the solvent volume escalates, subsequently leading to lower efficiency. This condition may due to the solution gets saturated with the substrate, thereby give a significant reduction in the mass transfer rate [16].
In this study, the highest delignification performance of MAA was obtained at mild condition, namely particle size of 90 mesh, SS ratio of 0.035, and irradiation time of 2.5 min. This condition resulted in the lignin content of 1.698%. Before the delignification process, the lignin content in cocoa pod husk was 21.4%. After the delignification process, the lignin content decreased to 0.099%, as shown in Figure  2c. Gazliya and Aparna [1] reported a high delignification result, i.e. from 20.11% to 9.43% at 5 N after MAA treatment.
On the irradiation time consideration, the lignin content rises to the optimum point and then drops dramatically by prolonging the irradiation time (Figure 2b). This trend may occur due to the collision phenomena between the molecules. Moreover, the disruption of intermolecular forces between the molecules stimulated by electromagnetic irradiation occurred. Furthermore, the prolonging irradiation time assisted the thermal accumulation, which caused a rapid breakdown of lignin [23,24]. Alexander et al. [16] reported that the delignification continuously elevated with extend in exposure time, yielding 73.21% delignification after 4.40 min.
Furthermore, conceptually, longer reaction times give an increase in the rate of reaction, which was presented by higher delignification yield. This condition provides a higher possibility of a sufficient collision frequency among the molecules which generate immense intramolecular heat energy [16]. However, there were no significant differences in delignification yields between 2 min and 3 min and reached a plateau in 2.2 min (See Figure 2a). This phenomenon may be attributed to kinetic factors, indicating that from 2.2 min to higher reaction times, the amount of extracted lignin is nearly constant. The prolonged delignification process beyond 3 min of irradiation time had an insignificant effect. From the other work utilizing microwave for biomass pretreatment, Gazliya and Aparna [1] reported that MAA pretreatment gave higher delignification results (i.e. 53.10% vs. 48.63%) using lower NaOH concentration (i.e. 5 N vs. 7.5 N) compared to alkaline pretreatment using conventional heating. This fact indicated that MAA gave a better result from performance and environmentally friendly aspects. Moreover, the physicochemical properties observed after the delignification process showed that MAA pretreatment increased cellulose content and surface area compared to alkaline pretreatment. Kamalini et al. [15] also found that MAA was sufficient for the biomass delignification process. It was reported that lignin content decreased from 19.33% to 14.59% after MAA applied.
In comparison with alkaline pretreatment, the delignification of softwood pellets using diluted sulfuric acid was performed by Zhou et al. [25]. It was obtained a lignin yield of 82.31 % at 190 °C for 10 min. By comparing this result with those reported studies, this work obtained delignification with comparable and higher yield under milder alkaline conditions and shorter time. Figure 3 shows a three-dimensional response plot of the effect of variables on the changes of hemicellulose composition after MAA pretreatment. Before the pretreatment process, the hemicellulose content in cocoa pod husk was 3.8%. After the pretreatment process, the most effective hemicellulose content decreased from 3.8% to 1.195% by 68.55%, which can be seen in Figures 3a-c. The most effective reduction of hemicellulose in the experiments was carried out at 80 mesh particle size, a ratio of 0.035, and a time of 2.5 min. Kamalini et al. [15] reported that the hemicellulose underwent lower solubilization during MAA pretreatment, which declined from 32.30% to 27.15%. Thus, the decrease in hemicellulose levels in this study can be said to be better than previous studies. Hemicellulose content increases significantly with increasing particle size (Figures  3a and b). On the other hand, changes in hemicellulose content were not too significant with an increase in the ratio (see Figures 3a and c).

The Effect of variable on hemicellulose content
Hemicellulose content decreases significantly with the longer heating time (Figures 3b and  c). Liu et al. [11] found that the more prolonged chemical treatment significantly gave a positive contribution to the hemicellulose separation. Moreover, the higher temperature also Figure 3. The effect of ratio and particle size (a), time and particle size (b), and time and ratio (c) on the hemicellulose content. provided higher hemicellulose yield than the low temperature during the alkaline treatment process. This phenomenon might be attributed to the synergistic work between microwave irradiation and alkaline treatment. The straight interaction of biomass with the electromagnetic field volumetrically and rapidly occurred when the microwave irradiated the biomass. This interaction significantly removed hemicelluloses from the substrate, leading to the high hemicelluloses yield [11]. The successful performance of microwave heating for xylan removal was also reported by Panthapulakkal et al. [26]. It was reported that the xylose removal was more significant 6.2% than the conventional heating. For the severe heating conditions, the combined microwave and alkaline treatment contributed positive effects on the escalation of removal of lignin and hemicelluloses. The separation was presented by high hemicelluloses yield and improvement on hemicelluloses molecular weight [11].

The effect of variables on cellulose content
As shown in Figure 4, Cellulose content decreased significantly with increasing particle size (see Figures 4a and b). The lowest cellulose content was obtained at the midpoint of the condition (see Figures 4a and c). Cellulose content increases with prolonged reaction time. The highest cellulose content is in a severe condition (Figures 4b and c). From this work, the cellulose concentration was discovered to be elevated following the MAA process. Due to the effect of microwave heating, the cellulose content was remarkably increased during the alkaline treatment. The microwave irradiation cracked the compact structure of biomass, simultaneously stimulated the alkaline penetration into the inner biomass structure, leading to the promotion of the cellulose content [11].

Fourier Transform Infrared Spectroscopy (FTIR)
FTIR characterization was carried out to determine the functional groups contained in the detached and delignified cocoa shells. In general, the functional groups found in the two types of cocoa husk arise from lignin, hemicellulose, and cellulose [27]. Typical FTIR peaks that appear include 3410.  Cocoa depectination (without delignification test) contains lignin, cellulose, and hemicellulose. As shown in Figure 5, the results of the FTIR spectrum of the detected cocoa had O−H groups at wave number 3410.15 cm −1 , group C=O at wave number 1743.65 cm −1 , and double bonds (C=C alkene and C=C aromatics) with each wave number is 1651.07 cm −1 and 1527.62 cm −1 . This result indicates that the detached cocoa (without delignification test) contains lignin compounds. Besides, depectination cocoa also has cellulose content which is characterized by the emergence of CH alkane and C−H alkene groups with wave numbers 2924.09 cm −1 and 900 cm −1 and the appearance of wave numbers 1242.16 cm −1 indicating the presence of C−O ester groups and the presence of C−O ester groups aromatic CH group with a wavenumber of 856.39 cm −1 . These results indicate that depectination cocoa contains cellulose and lignin compounds that bind to each other.
The content of lignin, hemicellulose, and cellulose in cacao depectination was changed when the sample underwent the delignification process. One of the changes is that there are a shift and decrease in peak intensity in the functional group. Delignified cocoa showed a shift in absorption peaks from several functional groups such as O−H group from 3410.15 cm −1 to 3450.65 cm −1 , which indicated that lignin levels increased, significantly decreased the peak of C=O group (1743.65 cm −1 ), as well as the decrease in C=C of the alkene in the delignification process to 1639.49 cm −1 . Also, the loss of C−H alkanes and C=C aromatic groups indicates a change in hemicellulose. The cellulose content in the depectinated cocoa, which has undergone a delignification process, also decreased the peak absorption of the C−O ester group (1238 cm −1 ). So, it can be said that the delignified cocoa has decreased lignin and hemicellulose content (see in Figures 2 and 3, respectively), but there is a slight increase in cellulose content. FTIR results in this work were found in agreement with the increase in cellulose content as shown in Figure 4.

Scanning Electron Microscopy (SEM)
SEM is a method used to analyze pore structure, morphology and can be used to analyze changes in particle size in samples. SEM on cocoa depectination and delignification shows a significant difference in particle size. Figure 6 illustrates the morphology and particle of cocoa before being delegated (depectination) and cocoa that has been delegated. Morphology shows that delignified cocoa has a smaller cavity compared to depectative  cocoa, in Figure 2b, having a thicker and rougher surface to protect the lignocellulosic content contained in fiber [28]. The significant difference in the two samples lies in the particle size. Delignified cocoa (Figure 2a) has a much smaller particle size and is barely noticeable, in contrast to depectination cocoa, which has massive particle and cavity sizes. The smaller the particle size, the more cellulose is produced, and the proportion of lignin decreases as the particle size decreases. This condition may because lignin binds cellulose in acidic conditions so that it will condense and precipitate [29]. So, delignification cocoa has lower lignin content compared to depectination cocoa.

Kinetic Study
Lignocellulosic contains majorly of lignin, cellulose, and hemicellulose. In this study, the kinetics of microwave assisted alkali (MAA) pretreatment were studied for lignin, as the major target of delignification process which would degrade to inhibitor during fermentation without any pretreatment. Kinetics of cellulose conversion in the pretreatment was also studied as the hexose, fermentable sugars were derived from cellulose. Meanwhile, hemicellulose was not discussed because the pentoses as the main product, was not preferable as fermentable sugar by many type of strain [30].
In this study, the alkali was assumed to be exceeding reactant, so the concentration of lignocellulosic component was the driving force, and the reaction was independent to each other. The result from the kinetics analysis was shown in Eqs. (8-9) below.
(8) (9) This result was obtained using fractional time method with 80% fraction. The R 2 value of the fractional time linear plot was 0.999 and 0.997 for lignin and cellulose, respectively. These values showed the fit sufficiency of this model to the existing data from the experiment. Using MAA, the pretreatment of CPH performed faster conversion rate for lignin with klignin of 1.241 L 0.264 /mol 0.264 s and nlignin of 1.264, compared with cellulose with lower kcellulose of 1.029 L 0.312 /mol 0.312 s and only slightly higher order kcellulose of 1.312. This process effectively degraded lignin and simultaneously favouring cellulose degradation which both may be beneficial to reduce the requirement of next processing after delignification, saccharification.
The kinetics of delignification was not yet studied for CPH using MAA. Keshwani et al. [31] studied the MAA for switchgrass using time dependent kinetics modelling, resulting reaction constant of 0.499 for lignin and 0.011 for cellulose. This study for CPH resulted in higher reaction constant, means that this process could be performed faster than previous study mentioned. For further research, it is still crucial to conduct the economic review of MAA pretreatment for application in the pilot and industrial scale.

Conclusions
In this work, the alkaline pretreatment under microwave heating was sufficient for the delignification process within a less time irradiation process. The optimal results were obtained on the mild condition; particle size of 90 mesh, SS ratio of 0.035, and a delignification time of 2.5 min. The integrated treatment was characterized by decreased lignin and hemicellulose content, but there is a slight increase in cellulose content. In this work, the best condition of MAA delignification was attained at mild condition of particle size, SS ratio, and irradiation time, namely 90 mesh, 0.035, and 2.5 min, respectively, resulted in the 1.698% decrease of lignin content. From kinetic study, using MAA, the pretreatment of CPH performed faster conversion rate for lignin with kinetic constants of lignin (klignin) and cellulose (kcellulose) are 1.241 and 1.029, respectively. Furthermore, kinetic order found for lignin (nlignin) and cellulose (ncellulose) were 1.264 and 1.312, respectively. This study presents that MAA process is an effective pretreatment method to degrade lignin and simultaneously favouring cellulose degradation.

Acknowledgment
Authors thank the Directorate General of Resources for Science, Technology, and Higher education, Ministry of Research, Technology, and Higher Education of Republic Indonesia for funding this work.