Two-Step Esterification Process of Palm Fatty Acid Distillate Using Soaking Coupled with Ultrasound: Process Optimization and Reusable Solid Acid Catalysts Authors: Jarernporn Thawornprasert, Krit Somnuk Published: 2024 ABSTRACT AbstractIn this study, high-intensity ultrasound (1000 W ultrasonic power and 18 kHz frequency) was employed in conjunction with heterogeneous acid catalysts to reduce the free fatty acid (FFA) content in palm fatty acid distillate (PFAD). The aim was to convert the FFAs in PFAD to methyl esters (MEs) through an esterification process. The use of Amberlyst-15 as a heterogeneous acid catalyst offers environmental advantages over homogeneous catalysts and the potential for reuse in the biodiesel production process. The ME purity in the esterified oil was optimized by varying four independent variables: methanol content (25–65 wt %), Amberlyst-15 loading (20–60 wt %), soaking time (30–130 min), and sonication time (30–630 s). Furthermore, the efficiency of reused solid acid catalysts in reducing FFA over multiple cycles was investigated. The results indicated that under optimal conditions of 36.73 wt % methanol content, 60 wt % Amberlyst-15 loading, 130 min soaking time, and 505 s sonication time at a reaction temperature of 60 °C, the ME purity reached 89.44 wt %. Moreover, the potential of the recovered solid catalyst to reduce FFA in fresh PFAD under optimal conditions was examined. It was found that recovered Amberlyst-15 could be reused for at least two cycles with ultrasound to achieve an ME purity of over 80 wt %. Introduction Biodiesel is a sustainable fuel produced from plant oils such as palm, soybean, coconut, and rapeseed.1,2 Biodiesel can be blended with petroleum diesel and directly used in diesel engines due to its physical and chemical properties being quite similar to diesel fuel.1 The adoption of biodiesel presents a way to reduce dependence on fossil fuels and mitigate the impact of petroleum price fluctuations on the global market. Furthermore, many countries have provided incentives and subsidies to encourage biodiesel production and consumption, resulting in lower greenhouse gas emissions, increased energy security, and measures to address climate change. Biodiesel significantly reduces emissions of nitrogen oxides, hydrocarbons, carbon dioxide, carbon monoxide, and particulate matter compared to petroleum diesel.2−5 Therefore, using biodiesel–diesel blended fuels can help reduce greenhouse gas emissions in the transportation sector.6 In addition, biodiesel has a greater cetane number than petroleum diesel, which improves the fuel blend’s properties and enhances engine combustion and performance.7 In Southeast Asia, crude palm oil (CPO) is a commonly used feedstock in biodiesel production. Thailand, for example, produces biodiesel (B100) by blending CPO with diesel. However, the use of edible palm oil as feedstock may negatively impact the agricultural industry due to the increasing demand for biodiesel feedstock. To address this issue and ensure price stability for palm oil while maintaining an adequate supply for domestic consumption, careful consideration is given to increasing the proportion of biodiesel derived from CPO in diesel fuel, taking into account the oil palm harvesting periods.8 For the second generation of feedstock for biodiesel production, palm fatty acid distillate (PFAD) is generated as a byproduct during the physical refining process of CPO to produce biodiesel from nonedible oil. PFAD from CPO is estimated to have a free fatty acid (FFA) content of approximately 3–5%.9−11 PFAD exists in a light-yellow wax phase at room temperature (30 °C), with its main component being palmitic acid.9,12 The animal feed, soap, detergent, and oleochemical industries all frequently utilize PFAD as a raw material.9,10 PFAD biodiesel has the potential to be used in diesel engines without requiring the engine to undergo any major modifications.9,10,12 However, the high FFA content in PFAD poses a significant challenge for biodiesel production using an alkali catalyst in the transesterification reaction, as it can lead to saponification reactions. This can result in decreased biodiesel yield when significant saponification occurs during the transesterification process.13,14 Using an acid catalyst in the FFA reduction process may effectively address the issue of soap generation resulting from saponification reactions, which occur when FFAs in the oil convert into esters during the esterification process.14,15 Consequently, biodiesel production from oils with a high FFA content requires a two-step process. Initially, the FFAs in the feedstock are converted to esters through an esterification reaction using an acid catalyst. Subsequently, the esterified oil is treated with a base catalyst to convert the remaining glycerides into esters and glycerol.15,16 Sulfuric acid is often used as a homogeneous acid catalyst in the esterification reaction due to its cost-effectiveness and ability to accelerate the reaction.17 However, the use of sulfuric acid in the esterification process can lead to equipment corrosion and increased maintenance costs.16,17 Furthermore, a homogeneous acid catalyst cannot be reused for multiple cycles after the FFA reduction process.14,17 Several studies have highlighted the advantages of employing heterogeneous catalysts in biodiesel production, including reduced corrosion in reactor tanks and lower maintenance costs.17 Moreover, heterogeneous acid catalysts in reactant mixtures can be reused for subsequent cycles in the esterification process.14,17 To solve these issues, heterogeneous acid catalysts offer a better alternative. Heterogeneous acid catalysts reduce equipment corrosion, can be easily separated from the mixture, and are reusable, lowering the catalyst cost. The heterogeneous acid catalysts including titanium dioxide, zirconium dioxide, zeolite, tin dioxide, and ion-exchange resins containing sulfonic acid groups (Nafion SAC-13, Nafion-NR50, Amberlyst-15, and Amberlyst-35) were used to produce biodiesel. Ion exchange resins and mesostructured silica modified with sulfonic acid groups are often utilized in biodiesel synthesis due to their high catalytic activity during the esterification process.18 Amberlyst-15 is a strong Brønsted acid catalyst. Its sulfonic acid groups provide the protons required for catalysis, making it especially beneficial for acid-catalyzed processes.19 This catalyst was chosen for reducing FFA in PFAD because it is a concentrated strong acidic resin with sulfonic acid groups. It is commonly used as a heterogeneous acid catalyst in esterification processes to produce biodiesels. Its strength and strong acid stability make it suitable for the conversion of FFA to ester. A study has reported on the use of sulfonic acid catalysts for catalyzing reactions and the reuse of catalysts to produce biodiesel in the esterification process. Lima et al.20 investigated the potential of mining kaolin waste to reduce FFA in oil during the esterification reaction by synthesizing a heterogeneous catalyst (Al-MCM-41) from silicon and aluminum. They added sulfonic acid groups to this material, and then, they synthesized Al-MCM-41 to obtain AM41–2H–O, which was then used as a catalyst in the esterification processes. They found that the ester purities of 99% and 98% were achieved for oleic acid and waste generated after the palm oil deodorization process (DDPO), respectively, under the conditions of a 1:30 molar ratio of fatty acid to methanol and 5% catalyst loading for a 2 h reaction time at 130 °C. In addition, the methyl ester (ME) purity was investigated at various cycles using the reused AM41–2H–O catalyst. The results clearly showed that reusing the catalyst for at least two reaction cycles converted the FFA in DDPO to an ester with over 72% ME purity. The primary objective of this study was to use ultrasound in combination with heterogeneous acid catalysts to accelerate the esterification process. Ultrasound generates high-frequency waves that can result in significant heat and mass transfer, leading to the formation of cavitation bubbles during the chemical reaction.21 Ultrasound reactors enhance reaction rates by improving the effective mass transfer and increasing the contact area between mixtures through cavitation. Therefore, the advantages of using ultrasound in biodiesel production compared with conventional agitation methods included a reduced reaction time, enhanced mass transfer, and more efficient mixing. In addition, ultrasound can enhance the mass transfer and reaction rates in homogeneous and heterogeneous catalysts.22 Moreover, most low-acoustic-energy-density (AED) ultrasonic baths are less efficient and powerful than ultrasonic probes, which require a longer reaction time and operate in combination with a mechanical stirrer. As previously mentioned, introducing the probe ultrasound directly into the mixture will result in a stronger ultrasonic wave and a more even distribution compared to using an ultrasonic bath.23,24 Trinh and Yusup25 investigated the use of a continuous ultrasonic reactor with heterogeneous catalysts for the production of biodiesel from rubber seed oil. They found that ultrasound enhances the mass transfer intensity between solid and liquid phases. Their results showed that the ultrasonic reactor was able to convert 95.95% of the FFAs in rubber seed oil within 3.5 min, whereas the batch reactor required over 30 min to achieve 98% FFA conversion through the esterification process. In the transesterification process utilizing a solid catalyst (K2CO3/nano-MgO), an 89.23% ester was obtained within 2.3 min using an ultrasonic reactor, compared to 11.5 min required by the batch reactor to achieve an 89.03% ester yield. Additionally, Boffito et al.26 investigated the use of ultrasound in the esterification process to reduce FFAs in tobacco and canola oils using Amberlyst-46 heterogeneous catalysts. They found that ultrasound enhances the esterification reaction rate by improving the mass transfer between methanol and adsorbed organic molecules on heterogeneous catalysts. By comparing ultrasound with mechanical stirring, they found that ultrasound converted 68% of the FFAs, while mechanical stirring converted only 23%. In order to optimize biodiesel production from oils with high FFA contents, the cavitation phenomenon in ultrasound reactors enhances mass transfer, significantly reducing reaction times and improving FFA conversion rates compared with conventional batch processes. The literature review above has demonstrated the effectiveness of ultrasonic reactors coupled with heterogeneous catalysts in FFA conversion, showing their potential to increase the esterification reaction rate for biodiesel production. Because of the high concentration of FFAs in PFAD, converting them to ester using a two-step process of esterification and transesterification poses a significant challenge for producing biodiesel. However, there are environmental and equipment issues associated with the esterification process and the large amounts of homogeneous acid catalysts, such as sulfuric acid, used in biodiesel production. Amberlyst-15 catalyst was selected for the esterification process due to its high efficiency in FFA reduction, low corrosive equipment, reusability, and cost-effectiveness. Furthermore, the porous structure of this catalyst benefits from the soaking process, which speeds up FFA conversion. Consequently, optimizing the soaking time was also a key factor in this experiment. Therefore, this study aims to address the research gap in the esterification process by enhancing the conversion of FFA to ME in PFAD through the utilization of high-intensity ultrasound and soaking with a heterogeneous acid catalyst. An experiment was designed to evaluate the production of ME from PFAD to determine the optimal conditions by using response surface methodology (RSM). To convert FFAs in PFAD into MEs via the heterogeneous esterification process, four variables were varied: methanol concentration (25–65 wt %), Amberlyst-15 loading (20–60 wt %), soaking time (30–130 min), and sonication time (30–630 s). Furthermore, the number of cycles for FFA reduction was examined by using reused solid acid catalysts. Materials and Methods Materials PFAD, which exists as a yellow solid wax at 30 °C, is a byproduct of the palm oil refining process, transitioning from a solid wax phase to a liquid phase at 43 °C. PFAD was purchased from a large commercial palm oil refining facility located in eastern Thailand. The esterification reaction can produce biodiesel from high FFA content in PFAD, as shown in eq 1.27 The composition of PFAD includes 98.52 wt % FFA, 1.35 wt % diglyceride (DG), and 0.13 wt % monoglyceride (MG). In this process, the commercial-grade methanol used had a purity of 99.7%. Meanwhile, Amberlyst-15 functioned as a sulfonic acid-functionalized polymer capable of exchanging protons (H+) with other cations found in the mixed reactants.28Figure 1 shows the wax phase of PFAD at 30 °C, the liquid phase of PFAD at 43 °C, and the esterified oil of PFAD (EPFAD) after the heterogeneous esterification process. Table 1 summarizes the properties and compositions of PFAD and EPFAD. Raw materials and products from the heterogeneous esterification processes are (a) the wax phase of PFAD at 30 °C, (b) the liquid phase of PFAD at 43 °C, and (c) EPFAD. Table 1Properties and Compositions of PFAD and EPFAD ResultBiodiesel standard of ThailandPropertyPFADEPFADFor commercial-based biodieselaFor agricultural enginesbMethodaDensity (kg/L)0.869 at 60 °C0.872 at 15 °C0.860–0.900 at 15 °C0.860–0.900 at 15 °CASTM D 1298Viscosity at 40 °C (cSt) 7.083.5–5.01.9–8.0ASTM D 445Flash point (°C) 157<120<120ASTM D 93Cloud point (°C) 15 Pour point (°C) 16 Acid value (mg KOH/g)227.8722.43<0.50<0.80ASTM D 664Water content (wt %) 0.08<0.0005 EN ISO 12937ME purity (wt %)0.0089.4496.5 EN 14103Triglyceride (wt %)0.000.00<0.20 EN 14105FFA (wt %)98.529.38 DG (wt %)1.351.07<0.20 EN 14105MG (wt %)0.130.12<0.70 EN 14105Higher heating value (MJ/kg)40.2238.99 Lower heating value (MJ/kg)37.7636.28 aDepartment of Energy Business.29bDepartment of Energy Business.30 Properties and Compositions of PFAD and EPFAD Department of Energy Business.29 Department of Energy Business.30 Equipment In the esterification process, the reaction between PFAD, methanol, and Amberlyst-15 catalyst is accelerated by the high-intensity ultrasound generated by an ultrasonic horn (model: AKHGZ-50420K, ACME-KORN, Thailand). This high-intensity ultrasound reactor was operated at a fixed frequency of 18 kHz, and the ultrasound power was set to a maximum of 1000 W to optimize the ME purity using a heterogeneous catalytic esterification process. In general, low frequencies of ultrasound are more effective for mass transfer between immiscible components than high frequencies. This is a result of the formation of larger cavitation bubbles, which enhance the mixing and release more energy.31 Furthermore, the frequency of this ultrasonic horn is nonvariable due to the unique design of the horn shape, and it is recommended to vibrate at its fixed natural frequency. Ultrasonic horn dimensions: a stainless steel ultrasonic horn with a diameter of 29 mm and a length of 150 mm was immersed in the reaction mixture inside the ultrasonic reactor. The reactor itself had the following dimensions: a 40 mm inner diameter, 70 mm height, and 3 mm wall thickness. The gap between the horn tip and the bottom of the reactor was fixed at 15 mm. A timer was integrated into the ultrasonic generator to regulate the duration of sonication. Furthermore, a cooling fan on top of the ultrasonic homogenizer was used to cool the piezoelectric ceramic transducer during the continuous operation. When the horn tip is used to generate ultrasound waves, their intensity is calculated by the ultrasonic power intensity (IUS). The unit of measurement for ultrasonic power intensity is watts of ultrasonic power transmitted to a mixture (PUS) per square millimeter of surface area at the horn tip (A), as shown in eq 2.32 The AED represents the ultrasonic power transmitted to a mixture (PUS) per total volume of the mixture (Vtotal), expressed in watts per volume, as calculated by eq 3.32 In this study, the ultrasound power intensity at the horn tip was 1.5 W/mm2, and the acoustic energy density was approximately 20 W/mL. Experimental Procedure for Esterification Process Using Ultrasound The impact of various parameters (methanol content, Amberlyst-15 loading, soaking time, and sonication time) on the ME purity from PFAD was investigated in the esterification process. The process began with the initial step of soaking and proceeded to ultrasonication in the second step. To prepare for the esterification process, 20 g of PFAD was weighed and heated in a reactor until it reached a temperature of 60 °C. Once PFAD turned into a liquid phase, the required concentrations of methanol and Amberlyst-15, as outlined in Table 2, were added to the reactor. The reactor lid had to be closed immediately after pouring the mixture to prevent methanol evaporation. The reaction time was immediately recorded based on the different soaking times at a constant temperature of 60 °C, as indicated in Table 2. The reaction temperature was controlled at approximately 60 °C to prevent the increasing temperature from ultrasonic cavitation during the esterification process, since the boiling point of methanol was 64.7 °C. Following the soaking step, the mixtures were subjected to immediate sonication using an ultrasonic homogenizer for varying sonication times, as shown in Table 2. After the ultrasonic esterification process, the sample was collected and poured through a mesh filter to separate the Amberlyst-15 catalyst from the crude EPFAD. The crude EPFAD, without the solid catalyst, was immediately placed in a cold box to prevent any forward or reverse reactions. The crude EPFAD underwent a washing process with water to remove any remaining methanol and impurities in the sample until the wastewater’s hydrogen ion potential (pH) was equal to 7. Subsequently, the EPFAD was dried at 105 °C to ensure minimal residual moisture content in the samples. Finally, the composition of triglyceride (TG), DG, MG, FFA, and ME in the EPFAD, in terms of weight percentages, was analyzed using thin-layer chromatography coupled with a flame ionization detector (TLC-FID). Each experimental condition in Table 2 was repeated three times, following the procedure described above. The results from these three repetitions were then averaged to ensure accuracy. Figure 2 outlines the experimental procedure for the esterification process using ultrasound in combination with solid acid catalysts. Table 2Experimental Design of Coded Factor Levels and Actual Results of Esterification Processa ME purity (wt %) RunM (wt %)A (wt %)S (min)T (s)actualpredictedError125(−2)40(0)80(0)330(0)44.5645.11–0.55235(−1)30(−1)55(−1)180(−1)24.8524.630.22335(−1)30(−1)55(−1)480(+1)27.9729.38–1.41435(−1)30(−1)105(+1)180(−1)31.7033.80–2.10535(−1)30(−1)105(+1)480(+1)41.1738.552.62635(−1)50(+1)55(−1)180(−1)62.4260.681.74735(−1)50(+1)55(−1)480(+1)65.5665.440.12835(−1)50(+1)105(+1)180(−1)70.5969.850.74935(−1)50(+1)105(+1)480(+1)74.7774.610.161045(0)20(−2)80(0)330(0)14.2613.031.231145(0)40(0)30(−2)330(0)48.7948.490.301245(0)40(0)80(0)30(−2)45.4546.55–1.101345(0)40(0)80(0)330(0)54.6055.38–0.781445(0)40(0)80(0)330(0)54.6855.38–0.701545(0)40(0)80(0)330(0)55.7355.380.351645(0)40(0)80(0)330(0)55.6555.380.271745(0)40(0)80(0)630(+2)56.1856.060.121845(0)40(0)130(+2)330(0)61.8262.27–0.451945(0)60(+2)80(0)330(0)78.9381.14–2.212055(+1)30(−1)55(−1)180(−1)28.4529.32–0.872155(+1)30(−1)55(−1)480(+1)33.0934.08–0.992255(+1)30(−1)105(+1)180(−1)33.8633.93–0.072355(+1)30(−1)105(+1)480(+1)39.8238.691.132455(+1)50(+1)55(−1)180(−1)63.8061.372.432555(+1)50(+1)55(−1)480(+1)66.1166.13–0.022655(+1)50(+1)105(+1)180(−1)67.0965.981.112755(+1)50(+1)105(+1)480(+1)69.8770.74–0.872865(+2)40(0)80(0)330(0)45.5045.93–0.43aM is methanol, A is Amberlyst-15, S is the soaking time, T is the sonication time, and ME is the methyl ester purity. Experimental Design of Coded Factor Levels and Actual Results of Esterification Processa M is methanol, A is Amberlyst-15, S is the soaking time, T is the sonication time, and ME is the methyl ester purity. Experimental procedure for biodiesel production from PFAD using ultrasound coupled with heterogeneous acid catalysts. Analysis Methods The composition percentages of TG, FFA, DG, MG, and ME in PFAD and EPFAD were determined using TLC-FID (model: Iatroscan MK-6s, Mitsubishi Kagaku Iatron Inc., Tokyo, Japan). The physical properties of the PFAD raw material and EPFAD product were evaluated by measuring the density using a hydrometer according to ASTM D1298-12b and measuring the viscosity using a viscometer bath (model: Julabo MD-16G, Labortechnik GmbH, Seelbach, Germany) according to ASTM D445-17a. Moreover, the flash point of the oil was analyzed using a Pensky–Martens semiautomatic device (Walter Herzog GmbH, Germany) according to ASTM D93-16a. Cloud point and pour point analyses were conducted using the Herzog CPP 97-2 device from Germany, referencing ASTM D2500 and ASTM D97 standards. The acid value was determined using titration techniques according to ASTM D664-09. The water content of EPFAD was measured using a Coulometric Karl Fischer Titrator (model: 831 KFC, Metrohm, Switzerland) according to EN ISO 12937. The low and high heating values were calculated from organic composition analysis using a CHNS/O analyzer (model: Flash 2000 Series, Thermo Scientific, Italy). The structure and surface of fresh and recovered catalysts were examined using scanning electron microscopy (SEM, model: SU 3900, Hitachi, Japan). Moreover, the surface area of Amberlyst-15 was investigated using the Brunauer–Emmett–Teller (BET) method with a surface area and porosity analyzer (model: ASAP2460, Micromeritics, USA). Experimental Design Table 2 shows the purities of ME in EPFAD obtained from the TLC-FID results under different conditions. The purities were determined using a multiple regression tool in Excel software to identify the optimal conditions for achieving the highest ester purity utilizing the RSM with a central composite circumscribed (CCC) design. To this end, the experimental design matrix included four independent variables: methanol content, Amberlyst-15 loading, soaking time, and sonication time, each set at five levels with rotatability encoded as −α, −1, 0, +1, and +α. Here, −α and +α represent the star points at the lowest and highest values of each parameter, respectively. The star points (αx) of the rotatable CCC are positioned at a distance α from the center, determined by the number of factors (k) in the design. In this study, the EPFAD production process was investigated by analyzing four independent variables. With four factors in the experimental design, the starting point is equal to two when the number of factors is set to four in eq 4. Therefore, the experiments were designed with five levels for each independent variable, with coded factor levels set to −2, −1, 0, +1, and +2, resulting in a total of twenty-eight experimental conditions for ME purity, as listed in Table 2. As shown in Table 2, the ranges of the four parameters assessed were methanol content (25–65 wt %), Amberlyst-15 loading (20–60 wt %), soaking time (30–130 min), and sonication time (30–630 s). To determine the optimal conditions for converting FFA in PFAD to high purities of ME in EPFAD after the esterification process using ultrasound coupled with a solid acid catalyst, the experimental design involved varying the four parameters using a five-level CCC method. The ME purities in EPFAD under different conditions were modeled using a second-degree polynomial and a multiple regression model, as expressed in eq 5). where Y represents the dependent variable, xi and xj are the uncoded independent variables, β0, is the constant coefficients, βi, βii, and βij are the coefficients of the independent variables, k is the number of factors, and ε is the error term. Results and Discussion Experimental Results and Statistical Analysis for EPFAD A prediction model was developed to estimate the optimal conditions for achieving the maximum ME purity through ultrasound-assisted esterification with a solid acid catalyst. The relationship between the independent and dependent variables was investigated using multiple regression analysis in order to identify the optimal conditions with careful consideration given to the statistical data. Table 2 presents the twenty-eight experimental conditions that were tested. The results indicated that the purity of ME ranged from 14.26 to 78.93 wt %. These experimental results were analyzed to establish a prediction model that could fit the experimental results by using multiple regression analysis at a 95% confidence level. The relationship between ME purity and the independent variables was modeled, as shown in eq 6. where ME is the ME purity of EPFAD (wt %), M is the methanol content (wt %), A is the Amberlyst-15 loading (wt %), S is the soaking time (min), and T is the sonication time (s). The analysis of variance (ANOVA) of the quadratic model affecting the ME purity is presented in Table 3. The probability of error (p value) is used to assess the statistical significance of the coefficient (β) for each term in the ME purity prediction model. Coefficients (β) with a p value <0.05 were considered significant for the prediction model, with the smallest p-value indicating the most significant influence on ME purity. Conversely, coefficients (β) with a p value >0.05 indicated that the term is not significant enough to fit the model and can be removed from the prediction model. For the correlation prediction analysis, the coefficient of the term β2A had the lowest p value, indicating that the Amberlyst-15 content had the most significant impact on the ME purity in the esterification process. Ridwan et al.33 reported that the conversion increased with an increasing catalyst content due to more active sites in the reaction mixture, resulting in an increased reaction rate. The sulfonic acid group on Amberlyst-15 acts as an ion exchanger and catalyzes the reaction with FFA present in the raw material, thus reducing FFA content and improving the quality of the biodiesel.34 The methanol content represented by the term β1M had the second most significant effect on ME purity, followed by terms β7M2 and β8A2. Methanol content plays an important role in the mass transfer of mixtures using ultrasound during the esterification process. Ultrasonic waves passing through the methanol form small gas bubbles known as cavitation bubbles.35 Cavitation is more likely to occur in methanol than in oil due to its lower viscosity and surface tension. In addition, methanol’s higher vapor pressure compared to oil makes it easier to convert into the vapor phase at low pressure, leading to the formation of cavitation bubbles.36,37 The explosion of these bubbles creates intense shock waves, generating shear forces and turbulence around the bubbles.38 This enhances the mixing and mass transfer between methanol and oil and improves contact between the methanol/oil mixtures and the solid catalyst. Furthermore, the influence of sonication time and soaking time, expressed in terms of β4T and β3S, respectively, had the fifth and sixth most significant effects on ME purity. In terms of the ion exchange rate when a solid catalyst is used, the sulfonic acid groups on Amberlyst-15 resin attract and exchange ions with FFAs in the raw material, reducing FFA levels and improving biodiesel quality. The ion exchange duration increases with increasing contact time between Amberlyst-15 and the raw materials, leading to a greater conversion of FFA to ME. Su and Guo34 reported that the reaction rate of the heterogeneous acid catalyst decreased because of the poor miscibility with alcohol when the heavy phase of the raw materials was used. Therefore, a longer reaction time will result in higher biodiesel purity due to an increase in mass transfer between the solid and liquid phases.27,34 The coefficient of multiple determination (R2) and the adjusted coefficient of multiple determination (R2adjiusted) were 0.995 and 0.993, respectively, indicating the high predictability of the model. The adequacy of the prediction model was assessed by testing for a lack of fit (LOF). The LOF assessment resulted in a value of 6.56, which is below the F-distribution value of 8.70 (F0.5,15,3) for the correlation prediction model of ME purity. The result indicates that there is no significant LOF because the inherent error is small compared with the experimental data. Therefore, it is acceptable to identify a correlation in the prediction model. The F-test was used to assess the hypothesis and consider the variation in the sample. The F-test resulted in an F0 value of 407.64, above the critical F value of 2.46 (F0.5,9,18). Thus, the statistical significance of the prediction model for ME purity has been proven. However, the number of studies was adequate to investigate the effect of varied variables on the ME purity. As a result, the fitted model suggested that the heterogeneous catalytic reaction employing ultrasound could convert FFA in PFAD to high ME purity via the esterification process. Figure 3 shows the relationship between predicted and experimental ME purities for heterogeneous catalytic reactions using ultrasound in the esterification process, with the data predicted by an empirical model. Table 3Values of each Coefficient and ANOVA of the Prediction modelaTerm of coefficientValuep valueβ0–152.2838560.000000000146β13.0023360.000000021126β23.8102920.000000000071β30.3430000.000060711832β40.0456830.000047092340β5–0.0100000.012570290442β6–0.0045600.005418491861β7–0.0246330.000000080967β8–0.0207200.000000939961β9–0.0000450.002208340052R20.995 R2adjiusted0.993 SourceSum of squaresMean squareF0FsignifDegrees of freedomRegression7639.80848.8670407.642.469Residual37.482.0824 18LOF error36.372.42496.568.7015Pure error1.110.3696 3Total7677.29 27aNote: F0 is the coefficient of ANOVA and Fsignif is the critical value of the F-distribution at the 0.05 significance level. Values of each Coefficient and ANOVA of the Prediction modela Note: F0 is the coefficient of ANOVA and Fsignif is the critical value of the F-distribution at the 0.05 significance level. Relationship between predicted ME purity and actual experimental ME purity. Optimal Conditions of ME Purity in EPFAD The optimal conditions for producing high-purity ME in EPFAD using ultrasound with a solid catalyst were determined using the model in eq 6 for the esterification process. As indicated by the p-value in the previous section, this study revealed that Amberlyst-15 had the most significant effect on ME purity, followed by the methanol content, sonication time, and soaking time. After optimizing the prediction model using Excel Solver, the highest purity of ME achieved was 91.11 wt % under the following conditions: 36.73 wt % methanol content (methanol to PFAD molar ratio of 3.1:1), 60 wt % Amberlyst-15 loading, 130 min soaking time, and 505 s sonication time at a reaction temperature of 60 °C. The experimental results validated the model’s predictions of ME purity and demonstrated the reliability of the optimal conditions. Based on the results of the actual experiments, it was determined that the actual ME purity was 89.44 wt %, deviating only by 1.83% from the prediction model. Under these optimal conditions, the experimental results showed that the FFA content in PFAD could be reduced from 98.52 to 9.38 wt %. The properties and compositions of EPFAD are listed in Table 1. Effect of Independent Variables on ME Purity of EPFAD The relationship between the dependent variable (ME purity of EPFAD) and the independent variables (methanol content, Amberlyst-15 loading, soaking time, and sonication time) is presented in the response surface plot shown in Figure 4. Figure 4a–c displays response surface plots illustrating the relationship between Amberlyst-15 loading and the variables of the methanol content, soaking time, and sonication time that impact ME purity. The response surface plot reveals that the relationship of Amberlyst-15 loading with other variables (methanol content, soaking time, and sonication time) significantly influences ME purity in EPFAD. The purity of ME increased significantly as the Amberlyst-15 loading increased due to the presence of more active sites in the mixture from the solid catalyst loading, enhancing catalytic activity.33 The sulfonic acid groups on Amberlyst-15 promote greater ion exchange between the mixture and the catalyst, thereby accelerating the reaction in biodiesel production.34 In addition, achieving high purity of ME in EPFAD production through esterification depends on crucial reaction parameters, such as methanol content, soaking time, and sonication time. In Figure 4a, it is observed that a methanol content range of 35–50 wt % and an Amberlyst-15 loading range of 58–60 wt % can produce EPFAD with over 80 wt % ME purity. This suggests that the purity of ME increases with an increasing methanol-to-oil ratio. Therefore, the methanol concentration in the oil plays a critical role in catalyzing the reaction in the esterification process. Lower methanol concentrations may lead to incomplete esterification reactions, resulting in lower ME purity. Conversely, excessive methanol concentrations can dilute the mixture, leading to a slower catalytic rate.27 Binnal et al.39 reported similar findings, indicating that the esterification rate decreases with an excessive methanol concentration. An increased methanol concentration in the mixture reduces the FFA levels, prompting a reverse reaction in the esterification process and consequently leading to a decreased purity of ME. Figure 4b displays the response surface plots that illustrate the impact of Amberlyst-15 loading and soaking time on the ME purity. The results showed that ME purity exceeds 80 wt % when Amberlyst-15 loading is within the range of 55–60 wt % and soaking time is over 80 min. As shown in Figure 4c, adding more Amberlyst-15 and soaking it for longer periods of time improves the purity of the ME, while a longer sonication time speeds up the esterification reaction. Over ME purity of 80 wt % on the surface plot is achieved under the parameter ranges of Amberlyst-15 loading in the range of 58–60 wt % and a sonication time exceeding 300 s. Figure 4e shows the response surface plot showing the impact of the methanol content and sonication time on ME purity. The range of methanol content of 35–55 wt % and the sonication time of 400–600 s can provide the ME purity with over 55 wt %. An appropriate methanol content range enhances mixing and mass transfer between methanol and oil because cavitation occurs more easily in methanol when ultrasound is used.35,38 However, when the methanol content exceeded 55 wt %, the AED decreased as the total volume in the ultrasonic reactor increased, resulting in a slight reduction in the ultrasonic intensity during mixing. Conversely, when the methanol content dropped to 35 wt %, the ultrasonic reactor achieved a high AED, which proved to be more effective than a low ultrasonic intensity during the conversion of FFA to ester by the esterification process. As methanol serves as a medium for cavitation bubble formation, a lower methanol content will reduce the mixing turbulence and cavitation bubbles. In Figure 4d, it is shown that ME purity was higher than 60 wt % when the soaking time exceeds 110 min within the methanol content range of 35–50 wt % because a longer soaking time enhances mass transfer between the solid catalyst and the mixture.27,34 Amberlyst-15, with its porous structure, benefits from soaking, as it allows the methanol/oil mixture to pass through more pores, thereby increasing mass transfer and accelerating the conversion of FFAs in the esterification process. Figure 4f shows the response surface plot illustrating the effects of the soaking time and sonication time on the ME purity. The results indicate that the purity of ME was more than 60 wt % when the soaking time exceeds 130 min and the sonication time is within the range of 400–600 s. Three-dimensional response surface plots of the four variables on ME purity in EPFAD: (a) Amberlyst-15 loading and methanol content, (b) Amberlyst-15 loading and soaking time, (c) Amberlyst-15 loading and sonication time, (d) methanol content and soaking time, (e) methanol content and sonication time, and (f) soaking time and sonication time. Reusability of Amberlyst-15 Catalyst Considering the reusability of Amberlyst-15 in the esterification process, it can be utilized for multiple cycles to produce biodiesel from fresh PFAD. Therefore, using heterogeneous catalysts as reused catalysts in the esterification process can help decrease costs in biodiesel production including treatment processes, waste disposal, toxic chemicals, and catalyst expenses. This study investigated the reusability of Amberlyst-15 in producing EPFAD from fresh PFAD during the esterification process. The Amberlyst-15 catalyst was recovered by washing it with methanol solvent to eliminate any residual biodiesel and impurities on the surface of the solid catalyst after each esterification process before reusing it in the subsequent cycle of EPFAD production. After washing, Amberlyst-15 was dried at 80 °C for 6 h in an oven to ensure the elimination of any residual solvent. Figure 5 shows the SEM images of the fresh and recovered catalysts (first to fifth cycles) at 50 and 5000× magnification. Figures 5a (50× magnification) and 5g (5000× magnification) show that fresh Amberlyst-15 appeared to have a smooth spherical surface. After the first cycle, Amberlyst-15 still appeared to have no noticeable difference compared to fresh Amberlyst-15, as shown in Figures 5b (50× magnification) and 5h (5000× magnification). After the second cycle, small cracks appeared on the surface of the recovered catalyst, as shown in Figure 5c (50× magnification) and 5i (5000× magnification). However, the recovered catalyst still maintained its spherical shape. Following the third cycle, some of the recovered catalyst showed signs of catalytic cracking, as depicted in Figure 5d (50× magnification). Figure 5j (5000× magnification) shows larger cracks on the surface of the recovered catalyst compared to the second cycle. Wang et al.40 noted that ultrasonic vibration and mechanical impact can lead to the breaking of solids. The propagation of high-frequency ultrasonic waves on liquids induces the formation of cavitation bubbles. The static stress from cavitation bubble formation results in a significant force of impact on the solid. Stress exceeding the solid’s strength can cause microcracks on its surface, ultimately leading to its breakage. In addition, solid catalysts are more prone to cracking when they swell from prolonged soaking in solvents.41 Upon examination of the SEM images shown in Figure 5e,f (50× magnification), it was observed that the recovered catalyst exhibited increased cracking with each recycling cycle. Despite this, the catalyst retained its ability to catalyze esterification reactions. Notably, the cracks on the surface of the recovered catalyst became more prominent, as shown in Figure 5k,l (5000× magnification). The cracking of recovered catalysts results in a decrease in catalyst efficiency.42 Therefore, the purity of biodiesel was reduced due to cracking of the recovered catalyst. In addition, the reduced surface area of the catalyst led to incomplete conversion of FFAs to esters. In this study, the surface area of the catalyst was evaluated using the BET method with a surface area and porosity analyzer. The catalytic activity of the catalyst was found to decrease with decreasing BET surface area and pore volume.34 The BET surface area of fresh Amberlyst-15 was 36.52 m2/g, with an average adsorption pore diameter of 357.03 Å. Conversely, for the recovered catalyst after five cycles, the BET surface area increased to 42.57 m2/g, while the average adsorption pore diameter decreased to 330.31 Å. The increase in the BET surface area of the recovered catalyst was attributed to the swelling of the catalyst as a result of prolonged soaking in the solvent. Amberlyst-15 is a polymer resin that is sensitive to solute adsorption and may swell upon contact with liquids, such as alcohol. Swelling occurs when alcohol penetrates the catalyst structure, which may cause expansion of existing pores or formation of new pores, resulting in more surface area for the structure. Therefore, BET analysis may show higher surface area values due to the increased surface area of the structure. Related similar descriptions were reported by Cabral et al.43 and Van de Steene et al.44 They reported that alcohol promotes the swelling of catalysts through solvent adsorption and the resulting BET surface area increases. In addition, the Amberlyst-15 catalyst is not strong enough to allow high-shear mixing and high-intensity ultrasound during the reaction. These issues led to an increase in cracks on the surface of both the reused catalyst and the BET surface area.42 The catalyst’s cracking formation may generate new surfaces and pores, thereby expanding its total surface area. The value of the BET increases when there is a greater number of pores available for gas adsorption. In addition, the decrease in the adsorption average pore diameter of the recovered catalyst may be due to the deposition of residual products from the process on the catalyst such as fatty acids, resulting in smaller pores.45 Ji-Yeon et al.46 reported that the outer surface of Amberlyst-15 is somewhat porous. The increased surface area of the catalyst after esterification from high FFA oil may be due to the blockage of the fatty acids in the pores, which inhibits the esterification process. A similar result was explored by Su and Guo,34 they found that the loss of catalytic activity in heterogeneous acid catalysts is related to the blockage and deposition of large fatty acid molecules in the small pores of the catalyst. Similar results were obtained by Burmana et al.,45 they reported that the pore size of the catalyst decreased after the esterification reaction because the byproduct residues adhered to the catalyst, resulting in a smaller pore size. This is because washing the catalyst with methanol has a limitation in cleaning these impurities from the catalyst. Similar results have been reported by Huang et al.47 they found that the biodiesel yield began to decrease after the number of catalyst reuses increased due to the accumulation of some biodiesel molecules on the catalyst surface covering the active sites. In addition, the decomposition of sulfonic groups on the catalyst also decreased the biodiesel yield. Consequently, the conversion of FFAs when using the recovered catalyst in the ultrasonic esterification process was lower than when using the fresh catalyst. Lastly, the reusability of the recovered Amberlyst-15 catalyst was evaluated in terms of ME purity after each cycle. The results indicated that the purity of ME in EPFAD decreased from the first to the fifth cycle, with purity levels of 90.85, 81.10, 65.01, 63.60, and 62.40 wt %, as shown in Figure 6. Based on these findings, it is recommended to utilize the recovered catalyst for at least two cycles to achieve ME purity levels exceeding 80 wt % for EPFAD production. In this study, the Amberlyst-15 catalyst was washed to remove the oil film on the Amberlyst-15 catalyst using a simple regeneration method. The methanol washing procedure is simple, inexpensive, and successful in recovering the reused catalyst. However, washing the catalyst with methanol has limitations in catalyst recovery because the byproduct residues adhere to the catalyst surface, leading to a decrease in ME purity. In addition to increasing the catalytic activity of the reused catalyst by washing with methanol followed by drying in oven,19 regenerating catalysts may be a solution to help restore the catalytic activity. Kim et al.48 studied the regeneration of reused Amberlyst-15. They reported that reused Amberlyst-15 was removed from the process using a stainless steel sieve. The reused catalyst was washed with deionized water and dried overnight. The reused catalyst was then immersed in H2SO4 for 5 h of stirring before cleaning with deionized water. Burmana et al.49 studied the recycling of catalysts using a catalyst regeneration system. They reported that the continuous regeneration of the catalyst with methanol can be supported by stirring with a pressurized N2. The stirring under pressurized N2 assisted the formation of a turbulent flow, which improved the removal of impurities in the pores of the catalyst. Furthermore, the activation of a reused catalyst can be performed by restoring the catalyst’s acidic properties by recharging sulfonic acid groups.47 Scanning electron microscopy images: (a, g) fresh catalyst at 50 and 5000× magnifications, (b–f, h–l) recovered catalysts after first to fifth cycles at 50 and 5000× magnifications. ME purity (wt %) in EPFAD from the first cycle to the fifth cycle. Conclusions In this study, the high FFA content in PFAD was converted into MEs using high-intensity ultrasound with heterogeneous acid catalysts. An ME purity of over 89.44 wt % in EPFAD was achieved under optimal conditions of 36.73 wt % methanol, 60 wt % Amberlyst-15, a soaking time of 130 min, and a sonication time of 505 s at a reaction temperature of 60 °C. The ultrasound accelerated the esterification reaction by enhancing mass transfer between oil, methanol, and solid catalyst through the cavitation phenomenon during sonication. However, small cracks were observed after the third cycle, when the recovered catalyst was tested for reusability in the esterification process. Despite a reduction in catalytic properties due to cracking and residue deposition, the recycled catalyst maintained its catalytic activity in the esterification process. It is recommended to use the recovered catalyst for up to two cycles to achieve ME purity levels exceeding 80 wt % in the production of EPFAD. These findings confirm that high-intensity ultrasound induces an effective heterogeneous esterification reaction and rapid mass transfer, leading to a rapid increase in ME purity after the soaking process. For further research and development, regeneration methods including solvent washing, treatment with concentrated acid solutions, and restoration of acidic properties will be studied to activate catalyst reuse. Moreover, future studies can also investigate the utilization efficiency of heterogeneous acid catalysts, including zirconium dioxide, tin dioxide, titanium dioxide, zeolites, and other sulfonic ion exchange resins, to evaluate their biodiesel production efficiency and catalyst reuse. Acknowledgments This research was supported by Prince of Songkla University under the Postdoctoral Fellowship Program. The authors declare no competing financial interest. References