Statistical optimization of biodiesel production from para rubber seed oil by SO₃H-MCM-41 catalyst

2.1 Synthesis of SO₃H-MCM-41 catalyst

A combination of alkyl- and mercaptopropyl-functionalized MCM-41 was synthesized via the one-step co-condensation of Tetraethyl orthosilicate (TEOS, 99% Aldrich) and 3-mercaptopropyl (methyl) dimethoxysilane (MPMDS, 95% Aldrich) in the presence of Cetyltrimethylammonium bromide (CTAB, 98% APS) following the procedure described in the literature (Díaz et al., 2003). The molar composition of the mixture was as follows: (1 − x)TEOS: xMPMDS: 0.12CTAB: 0.27TMAOH: 18.8CH₃OH: 77.7H₂O. First, 2.2 g of CTAB was dissolved in 70.0 g of deionized water and 32.0 g of methanol, and then stirred for 30 min at room temperature. To this was added a mixture of 7.4 g of TEOS and 2.6 g of MPMDS dropwise, followed by 1.0 g of Tetramethylammonium hydroxide (TMAOH, 25 wt.% in water, Aldrich). The reaction was stirred at room temperature for 16 h to evaporate all the methanol. After that, hydrothermal treatment was applied at 95 °C for 48 h and the solid product obtained was separated by filtration and dried in air at 100 °C. The CTAB template was then removed by refluxing 1.5 g of solid with 205 mL ethanol/20 mL 37 wt.% HCl at 70 °C for 24 h. Finally, 1.0 g of the obtained SH-MCM-41 solid was oxidized by 20 mL of 30 wt.% H₂O₂ to obtain the SO₃H-MCM-41 catalyst.

2.2 Characterization of the catalyst

Each catalyst was characterized by its physical and chemical properties. The textural properties of the catalysts in terms of the surface area and pore characteristics were quantified by N₂ adsorption–desorption using an Autosorb-iQ instrument (Quantachrome, USA). The surface area was calculated from the isotherm data using the Brunauer-Emmett-Teller (BET) model, while the pore volume and pore diameter were calculated by Barrett–Joiner–Halender (BJH) model. The structural ordering was substantiated by Small-angle X-ray powder diffraction (SAXD) on a Rigaku TTRAX III X-ray diffraction spectrophotometer operating at a low angle using Cu Kα radiation in the 2θ angle range of 1–10° with a resolution of 0.02°. The presence of sulfhydryl and sulfonic groups on the MCM-41 surface was confirmed by Fourier transform infrared spectroscopy (FT-IR) using a Perkin Elmer Spectrum 100 FT-IR spectrometer. Thermal stability of catalyst was characterized by Perkin-Elmer TGA7 within the temperature range 100–800 °C at a heating rate of 10 °C/min. The amount of SO₃H-groups on the MCM-41 surface was determined by back titration (Hofena et al., 2011).

2.3 2^k factorial design

A 2^k factorial design was used to screen the factors that significantly affect the efficiency of the fatty acid methyl ester (FAME) (biodiesel) production process. The catalyst loading (5 and 15 wt.%), reaction time (12 and 48 h), reaction temperature (100 and 140 °C) and molar composition of MPMDS (0.29 and 0.45) were defined as factors A, B, C and D, respectively. The calculation was performed by using MS-Excel. From the 2⁴ factorial design, 16 combinations were performed (Table 1) with the percentage FAME yield as the response in each case.

2.4 Optimization by the Box–Behnken design

After the significant factors were verified (Section 2.3), optimization for FAME production was studied by response surface methodology using the Box–Behnken design to maximize the FAME yield. The quadratic regression model from the Box–Behnken design was used to predict the optimum condition, as shown in Eq. (1);

2.5 Transesterification reaction of RSO with methanol

Biodiesel was produced by transesterification of RSO at a 16:1 M ratio of methanol: oil, as previously reported (Samart et al., 2009). The mixture 5.0 g of RSO, 3.0 g of methanol and catalyst (wt.% of oil) were placed inside a 100 mL round-bottom flask equipped with a condenser while the cooling water was set at 10–15 °C using a thermostatic bath (Oliveira et al., 2010; Yin et al., 2012). The reactions were performed at stirring condition of 750 rpm, time and temperature condition were set up. At the end of the reaction time, the reaction was terminated by quenching in an ice-water bath. The excess methanol and catalyst were separated by evaporation and filtration, respectively. The FAME content was analyzed by gas chromatography (GC) Shimadzu GC-17A using DB-WAX (30 m length, 0.25 mm internal diameter and 0.25 μm film thickness) following the standard method of EN 14103. The FAME yield (% content) was calculated by Eq. (2);

2.6 Catalyst reusability

The spent catalyst was separated from the reaction mixture by filtration and washing by methanol followed by n-hexane. The catalyst was then dried at 100 °C and reused at the optimal conditions.

3.1 Characterization of catalyst

The catalyst was measured the textural properties by N₂ adsorption desorption. All samples were degassed at 150 °C for 8 h under flow nitrogen prior to measurement. The analysis was carried with 40 sorption pressures which could enough to provide all information of mesoporous characteristics. The BET surface areas and pore characteristics were quantified and the results are summarized in Table 3. The surface area and pore diameters were slightly decreased 1.3- and 1.14-fold, respectively at the higher MPMDS molar ratio because the excess MPMDS hindered the formation of the mesoporous structures. Moreover, the pore volume was decreased (3.2-fold) due to pore filling by the propyl sulfonic acid groups. The N₂ adsorption–desorption isotherms (Fig. 1A) indicated a IUPAC type IV isotherm, which corresponds to the mesoporous structure. The BJH pore size distribution (Fig. 1B) of both samples was uniformed. Moreover, pore diameter was decreased when the molar composition of MPMDS increased. The acid capacity of the catalyst was slightly (1.02-fold) enhanced by increasing the molar composition of MPMDS because the acid site depended on the amount of sulfonic groups (Table 3).

Close

Figure 1 (A) N₂ adsorption–desorption isotherms and, (B) BJH pore size distribution and (C) SAXD patterns of (a) 0.29-SO₃H-MCM-41 and (b) 0.45-SO₃H-MCM-41.

Export to PPT

The ordered mesoporous structure of catalyst was confirmed by SAXD analysis, where three diffraction peaks at a 2θ of 2.2°, 4.4° and 5.1°, which were designated as the (1 0 0), (1 1 0) and (2 0 0) planes, respectively, were observed (Fig. 1C). This supports that the hexagonal structure of the SO₃H-MCM-41 catalyst was successfully synthesized. However, the decreasing intensity of these three peaks indicated the presence of the sulfonic group, since the presence of the alkylsulfonic molecule hindered the highly ordered pore arrangement (Rác et al., 2006).

The functional groups on the catalyst surface were evaluated by FT-IR analysis, where wave numbers at 450 cm⁻¹ (SiO₄, tetrahedron vibration), 800 cm⁻¹ (Si—O—Si, symmetric vibration), a large band between 1000 and 1260 cm⁻¹ (Si—O—Si, asymmetric vibrations), and 3400 cm⁻¹ (SiO—H) were observed (Fig. 2). The wave numbers 1458 and 2989 cm⁻¹ were assigned to —CH vibrations. SH-MCM-41 (spectra a and b) presented an additional peak at wave number 2570 cm⁻¹ that was assigned to the SH-group, supporting that mercaptopropyl groups were introduced to the surface of MCM-41 successfully. After SH-MCM-41 was oxidized by H₂O₂ to SO₃H-MCM-41 (spectra c and d) a new peak at wave number 1350 cm⁻¹ (—SO₃ group) was evident while that at 2570 cm⁻¹ (SH-group) was lost supporting the complete transformation of SH-MCM-41 to SO₃H-MCM-41.

Close

Figure 2 Infrared spectra of the different sulfonic acid-MCM-41 catalysts: (a) 0.29-SH-MCM-41, (b) 0.45-SH-MCM-41, (c) 0.29-SO₃H-MCM-41, (d) 0.45-SO₃H-MCM-41 and (e) spent 0.29-SO₃H-MCM-41.

Export to PPT

Thermogravimetric analyses carried out for thermal stability verification of catalyst. According to DTG profile, both samples show two weight loss profiles at 380 °C and 470 °C (Fig. 3), which was the decomposition of sulfonic functional group and main chain of hydrocarbon of Propyl (methyl) sulfonic acid on MCM-41 (Díaz et al., 2000; Díaz et al., 2004). From thermogravimetric result, the decomposition of sulfonic acid did not present in transesterification temperature.

Close

Figure 3 TG-DTG profiles (a) 0.29-SO₃H-MCM-41 and (b) 0.45-SO₃H-MCM-41.

Export to PPT

3.2 2^k factorial design analysis of biodiesel production

The results of the 2⁴ factorial designs (16 combinations of catalyst loading, reaction time, reaction temperature and molar composition of MPMDS, with the FAME yield as the response; shown in Table 1) were used to calculate the estimated effects and plotted against the probability (pK) (Fig. 4). It was clear that all factors, except the molar composition of MPMDS (factor D), and their interactions had a significant effect on the percentage FAME content. The main effect of the catalyst loading (factor A) and reaction time (factor B) equally influenced the FAME yield because the acid catalysis in the transesterification presented a very low catalytic activity. Therefore, the reaction time and catalyst loading strongly affected the obtained FAME yield, as did their interaction (factor AB) since there are limitations on the catalyst loading where an excess catalyst loading induces phase separation and so reduces the catalytic activity. The significant factors were then expressed as a linear regression model, as shown in Eq. (3);

Close

Figure 4 Normal probability plot of the estimate effects for the 2⁴ factorial designs. (A = catalyst loading (wt.%), B = reaction time (h) and C = reaction temperature (°C), and see Table 1). Plots that lie on the line are those for factor D (Molar composition of MPMDS) and its combinatorial interactions with A, B and C, plus interaction BC.

Export to PPT

The reliability of the linear regression model was tested by the normal probability plot of residuals, and found to reasonably well fit a linear (straight line) pattern (R² = 0.9601). Thus, the residuals obtained from the regression model followed the normality assumption (Fig. 5).

Close

Figure 5 Normal probability plot of the residuals from the linear regression model showing the best fit linear regression line and correlation coefficient.

Export to PPT

3.3 Optimization of FAME (biodiesel) production using the Box–Behnken design

Table 2 shows the Box–Behnken design with 15 experiments and the experimentally obtained FAME yield. The quadratic regression model of Box–Behnken design was presented in Eq. (4),

Close

Figure 6 Parity plot of the predicted versus experimental FAME yields and the best fit linear regression line and correlation coefficient.

Export to PPT

The theoretical FAME yields with different reaction parameters were then calculated from Eq. (4) and the results presented as response surface and contour plots (Fig. 7). The effect of the catalyst loading and reaction time on the FAME content at a constant reaction temperature of 120 °C revealed an increasing FAME yield was predicted as the catalyst loading and reaction times were increased (Fig. 7A and B), giving a maximum FAME yield (81.9%) at a catalyst loading and reaction time of 14 wt.% and 48 h, respectively. At a constant reaction time of 30 h, the FAME yield increased with increasing catalyst loading level or reaction temperature (Fig. 7C and D), while at a constant catalyst loading of 10 wt.% the FAME content increased with increasing reaction time and temperature (Fig. 7E and F), but never exceeded that at 120 °C for 48 h with a 14 wt.% catalyst loading.

Close

Figure 7 (A, C, E) Response surface and (B, D, F) contour plots of the FAME yield (% content) at various (A, B) catalyst loadings (wt.%) and reaction times (h), (C, D) catalyst loadings (wt.%) and reaction temperatures (°C), and (E, F) reaction times (h) and reaction temperatures (°C).

Export to PPT

The optimal conditions for obtaining the highest FAME yield were calculated by maximization of the quadratic regression model, and found to be a catalyst loading of 14.5 wt.%, and a reaction time and temperature of 48 h and 129.6 °C, giving a maximum FAME yield of 84.0%. This was experimentally supported, where three replicated experiments at these optimal conditions gave a FAME yield of 83.10 ± 0.39%.

3.4 Catalyst reusability

The spent catalyst after use under the optimal set of conditions (Section 3.3) was then washed with solvents (see Section 2.7) and then reused under the same optimal conditions. The FAME yield obtained with successive reuse of the catalyst is shown in Fig. 8, where the FAME yield obtained after two, three and four uses of the catalyst was 1.03-, 1.08- and 1.09-fold lower than that for the unused catalyst. Thus, the SO₃H-MCM-41 catalyst has a relatively high stability and may potentially be reusable. The spent catalyst after four reuse was presented the weak C⚌O spectra at wave number 1738 cm⁻¹ from carbonyl containing compounds (Fig. 2) which was attracted by sulfonic group to reduce the catalytic activity of spent catalyst (Rubio et al., 2010).

Close

Figure 8 Catalytic activity (in terms of % FAME yield) of the fresh SO₃H-MCM-41 catalyst (1st use) and that of the regenerated catalyst after two, three and four uses in the transesterification of a 1:16 M ratio RSO: methanol mixture at a catalyst loading of 14.5 wt.% and a reaction temperature and time of 129.6 °C and 48 h per cycle.

Export to PPT