Facile Synthesis of Ferric-Modified Phosphomolybdic Acid Composite Catalysts for Biodiesel Production with Response Surface Optimization.

An attempt has been made to optimize the preparation of biodiesel from the transesterification of oleic acid with methanol over iron(III)-doped phosphomolybdic acid (H3PMo) catalysts. The prepared doped H3PMo salts were characterized using powder X-ray diffraction, Fourier transform infrared spectroscopy, thermogravimetric analysis, and scanning electron microscopy. The detailed characterization results demonstrated that the doped H3PMo salts have a strong interaction between the iron(III) ions and metal oxygen cluster, well preserving a typical Keggin structure of heteropolyacids and possessing good thermal stability. The effect of esterification reaction parameters was investigated and optimized using single-factor experiments method in combination with response surface methodology (RSM). The doped catalyst exhibited good catalytic activity, affording the oleic acid conversion of 89.2% with single factor optimization and 95.1% with RSM. More importantly, the catalyst was simply separated by decantation and exhibited good stability, with the oleic acid conversion of 70.2% after three consecutive cycles. Besides, this catalyst can also catalyze the esterification of other free fatty acids. Therefore, the doped H3PMo catalyst is a promising candidate for eco-friendly production of biodiesel in industry.

Introduction

The gradual depletion of the traditional fossil fuels increases the environmental pollution and global warming, so more and more attention is being paid to develop renewable fuels (e.g., biodiesel and biofuel).[1,2] Biodiesel production from plant oils,[3] animal fats,[4] waste cooking oils,[5,6] or free fatty acids (FFAs)[7] is regarded as a renewable green fuel owing to its clean, nontoxic, lower emissions of carbon monoxide, and carbon-neutral characteristics. On the other hand, it has similar physicochemical properties to conventional fossil diesel, such as comparable lubricity and flash point cetane value, and could be directly applied in engine without significant engine modifications.[8,9] In recent years, although the use of homogeneous alkali hydroxides or alcoholates and solid-base catalysts have become an effective method in the catalytic transesterification synthesis of biodiesel,[10] base-catalyzed transesterification of raw oils with FFAs results in serious saponification reaction, which reduces the biodiesel yield and influences the quality of production.[11] Therefore, conversion of FFAs into fatty acid methyl esters via the esterification reaction becomes a necessary and significant pretreatment process for raw oils with high concentrations of FFAs. Homogeneous acids (such as H3PO4, HCl, and H2SO4) are low cost and widely available and show excellent catalytic activity in esterification. However, they suffer from problems of corrosive to equipment, requirement of neutralization, and nonrecyclability associated with product separation and purification, leading to the increase in biodiesel production cost and relevant issues.[12,13] Instead, solid-acid catalysts can overcome these drawbacks in esterification, such as mixed-metal oxides,[14,15] zeolites,[16] metal complexes,[17] and sulfonic acid functionalized solid acid.[18] Unfortunately, they suffered from some drawbacks of low acidity, weak interaction between active component and support, complicated preparation process, and poor mass transfer.

Heteropolyacids (HPAs) of the Keggin series, in particular H3PW12O40 (H3PW), H3PMo12O40 (H3PMo), and H4SiW12O40 (H4SiW), have drawn significant attention as promising eco-friendly catalysts applied for various organic transformations such as esterification, transesterification, dehydration, etherification, oxidation, and acetylation reactions due to their strong acidity, thermostability, and good oxidizing ability.[19,20] However, HPAs is easily soluble in most of the organic solvents, which restrains the catalyst recycling and product purification. Very recently, the exchange of protons of HPAs with various cations (e.g., K+, Ag+, Cu2+, Mn2+) resulting in the formation of insoluble salts could be an effective strategy for the use of HPAs as solid-acid catalysts in green chemistry.[21] Li et al.[22] have reported the optimization of the biodiesel preparation process from Eruca sativa Gars vegetable oil catalyzed by Cs2.5H0.5PW12O40 heteropolyacid salt. The heteropolyacid salt shows an excellent catalytic activity for environmental biodiesel production with a high yield of 99%. Interestingly, the Cs2.5H0.5PW12O40 heteropolyacid salt can be reused for six cycles. Su et al.[23] reported that microwave-assisted the preparation of biodiesel from yellow horn oil using Cs2.5H0.5PW12O40 catalyst, and a high conversion yield (96%) could be achieved. In our previous study, Ag+-, NH4+-, and Cu2+-doped HPAs were successfully used as efficient catalysts in esterification reaction.[24,25] As far as we know, the use of iron(III)-doped phosphomolybdic acid (H3PMo) as a catalyst in esterification reaction for the production of biodiesel is still lacking.

Moreover, various parameters such as reaction time and temperature, the molar ratio of feedstock to methanol, and the catalyst dosage affecting the conversion are noteworthy. Compared to unplanned approaches, the response surface methodology (RSM) helps researchers gain a better insight into the interactions among experimental variables, offering an optimization process and also saving research time and costs.[26]

Therefore, in the present work, we focus on the design of iron(III)-modified H3PMo catalysts used for the production of biodiesel in the esterification reaction, and the modified catalysts were characterized by X-ray diffraction (XRD), Fourier transform infrared (FT-IR), thermogravimetric (TG), and scanning electron microscopy (SEM). Various esterification reaction variables were optimized by single-factor experiments and response surface experiments. Meanwhile, esterification of different free fatty acids with methanol is also studied to validate the compatibility of the doped H3PMo catalysts. Finally, the reusability of the doped catalyst was also profoundly investigated.

Results and Discussion

Characterization of the Catalyst

The powder XRD patterns of pristine H3PMo, Fe1/3H2PMo, Fe2/3H2PMo, and Fe1PMo are shown in Figure a. The pristine H3PMo shows principal diffraction 2θ angles at 7.6, 8.8, 25.9, 26.7, 28.0, 32.2, and 35.0°, which can be attributed to a body-centered cubic secondary structure of Keggin anions.[27,28] When iron(III) ions were added, the diffractograms of Fe1/3H2PMo, Fe2/3H2PMo, and Fe1PMo show diffraction lines that belong to those of triclinic symmetry for the crystal lattice. Besides, these samples also exhibited some new strong peaks around at 20.0 and 33.0° associated with iron species,[29−31] implying that the heteropolyacid salt was formed, and Villabrille et al. also found the same characteristic peaks for the iron-doped heteropolyacid.[32] The FT-IR spectra of pristine H3PMo, Fe1/3H2PMo, Fe2/3H2PMo, and Fe1PMo are displayed in Figure b. The peaks of the Keggin structure were clearly observed at approximately 1064 cm–1 (P–O in the central PO4 tetrahedron), 963 cm–1 (Mo=O in the MoO6 octahedron), 874 cm–1 (Mo–Oc–Mo bridge, Oc: corner-sharing oxygen), and 788 cm–1 (Mo–Oe–Mo bridge, Oe: edge-sharing oxygen) in all samples, indicating that the Keggin structure was well preserved in doped samples. Moreover, for the Fe1PMo sample, its characteristic peaks of Keggin structure were weaker than those of the Fe1/3H2PMo and Fe2/3H2PMo samples, this surprising result may be because the iron(III) ions fully substituted the acidic protons of phosphomolybdic acid that has a strong interaction between the iron(III) ions and metal oxygen cluster.[28,33] Therefore, the FT-IR results revealed that a good Keggin structure retention was achieved after the iron-doped heteropoly cage structure.

Figure 1

(a) Powder XRD patterns and (b) FT-IR spectra of iron(III)-doped H3PMo.

On the other hand, the catalytic data for the Fe1/3H2PMo, Fe2/3H2PMo, and Fe1PMo catalysts in the esterification are evaluated. Each reaction was performed with 10:1 of methanol-to-oleic acid molar ratio, and 5 wt % of the catalyst at 70 °C for 3 h. The results showed the conversion was 87.5, 89.8, and 89.2% with the Fe1/3H2PMo, Fe2/3H2PMo, and Fe1PMo catalysts, respectively, implying the iron(III) cations that fully or partially substituted H3PMo have better catalytic performance, with no significant difference in the esterification rate. Thus, based on the powder XRD and FT-IR, as well as catalytic activity of various catalysts, the Fe1PMo solid-acid catalyst prepared by a simple and convenient method was selected for the subsequent study.

TG analysis was used to test the stability of the catalysts. It is demonstrated that the Fe1PMo catalyst was thermally stable in the air up to 800 °C, as presented in Figure a. A slight weight loss (13.1%) due to the vaporization of physically adsorbed water and thermal decomposition of constitutional water (H3O+) from Fe1PMo in the range of 40–400 °C[34] was also observed. Then, no additional weight loss of Fe1PMo catalyst was observed at the temperature of >400 °C, presumably due to the high thermal stability of the Keggin structure similar to that of other heteropolyacid-based solid-acid catalysts.[35−37]

Figure 2

(a) TG profile of Fe1PMo catalyst. SEM images of (b) pristine H3PMo and (c) Fe1PMo catalyst.

SEM images of pristine H3PMo and Fe1PMo catalyst are given in Figure b,c. There was significant morphological difference between the two samples. The images exhibited that the pristine phosphomolybdic acid has a large irregular shape and the surface morphology is rough. For the Fe1PMo catalyst, the phosphomolybdic acid after doping with iron(III) ions acquired a multilayered laminar structure morphology but exhibited cubic shapes and some cracks. Thus, from the above observation, the changing surface morphology of the Fe1PMo catalyst was probably due to the exchange of protons with iron(III) ions, which is consistent with previous reports.[25]

Single-Factor Experiments for Esterification

Effect of Methanol-to-Oleic Acid Molar Ratio

Since the esterification of oleic acid with methanol is reversible, excess methanol is significant to shift the equilibrium toward the product side, thus improving the substrate conversion rate.[38,39] In Figure a, it is shown that the conversion of oleic acid is raised from 44.5 to 89.2% with the changing molar ratio of methanol to oleic acid from 2:1 to 10:1. This is because more amount of methanol could not only promote the dispersion of the catalyst but also consolidate forward reaction. However, further increase in the molar ratio of methanol to oleic acid to 18:1 makes no difference in increasing the oleic acid conversion for the limited equilibrium of esterification, and similar result was also found by Ezebor et al.[40] Taking energy consumption and operating cost into account, 10:1 was chosen as the optimum molar ratio for methanol to oleic acid.

Figure 3

Effects of molar ratio (a), catalyst dosage (b), temperature (c), and time (d) on the oleic acid conversion.

Effect of Catalyst Dosage

The relationship between catalyst dosage and oleic acid conversion is shown in Figure b. As exhibited in Figure b, the oleic acid conversion is slower without adding a catalyst, and 51.6% conversion was obtained with the addition of catalyst in the dosage of 0.5 wt %. When the catalyst dosage was increased from 1 to 5 wt %, the conversion was gradually increased from 68.2 to 89.2%. This reason for this phenomenon was that more catalyst could offer more active sites to activate the substrate. However, the oleic acid conversion does not further increase if the catalyst dosage keeps increasing to 6 wt % because of the equilibrium limit. Meanwhile, a large catalyst dosage might lead to the dispersion of active centers and worsen the partial coverage, and similar results were reported by Wan et al.[41] Therefore, we selected 5 wt % as the optimal catalyst dosage for the following studies.

Effect of Reaction Temperature

Esterification of oleic acid with methanol is the endothermic reaction, which could increase the conversion rate through raising the temperature. As exhibited in Figure c, low temperature severely hampers the esterification reaction and the conversion is 60.3% at 20 °C. Then, an increase of the reaction temperature from 30 to 70 °C distinctly heightens the oleic acid conversion from 67.7 to 89.2%. If the reaction temperature is further increased, the large amount of liquid methanol vaporized will reduce the methanol concentration in the reaction system, probably resulting in an unfavorable effect on conversion. Therefore, the optimum temperature can be set at 70 °C.

Effect of Reaction Time

The effect of reaction time on the oleic acid conversion was also studied, and the results are exhibited in Figure d. It can be seen from Figure d that the oleic acid conversion increases from 58.6 to 89.2% with the increment of reaction time from 0.5 to 3 h. However, there was no obvious increase in the conversion when the time was above 3 h, and the oleic acid conversion of 90.8% is obtained in 7 h. Therefore, the reaction time of 3 h was found to be optimum when taking the production cost into account.

Statistical Analysis for Optimization of Fe1PMo Catalyst for Esterification

Experimental Design and Evaluation

The preliminary single-parameter studies imply that catalyst dosage, molar ratio of methanol to oleic acid, and reaction time are significant variables for biodiesel production. The effect of the three variables was investigated by RSM, and a 3-level–3-factor Box–Behnken Design including seventeen experiments were performed by a second-order response surface. A second-order polynomial model (eq ) was correlated with independent variables as followswhere Cconversion is the oleic acid conversion to biodiesel, β0 is the intercept term, β is the linear coefficient, β is the quadratic coefficient, β is the interactive coefficient, and x, x are the coded independent variables (the reaction conditions).

The minimum and maximum values of the Box–Behnken Design process variables are shown in Table . The Box–Behnken Design for the three process variables X1, X2, and X3 with the experimental results and predicted responses based on this RSM model are given in Table . The quadratic model, describing the relationships between the predicted response variable (conversion of methyl oleate at 70 °C) and the reaction conditions, was selected for an in-depth statistical study. On the basis of the results, the quadratic regression model of the experimental data (eqs and 3) for the oleic acid conversion is given below

Table 1

Variables and Experimental Design Levels for Response Surface

		levels
variable	symbol	–1	0	1
reaction time (h)	X1	2	3	4
catalyst dosage (wt %)	X2	3	4	5
molar ratio of methanol to oleic acid	X3	6	10	14

Table 2

Experimental and Predicted the Oleic Acid Conversion Using RSM Box–Behnken Design

				conversion (%)
standard order	X1	X2	X3	experimental	predicted
1	–1	–1	0	84.70	84.51
2	1	–1	0	91.60	91.31
3	–1	1	0	85.70	85.99
4	1	1	0	93.30	93.49
5	–1	0	–1	72.50	71.66
6	1	0	–1	77.50	76.76
7	–1	0	1	84.40	85.14
8	1	0	1	93.50	94.34
9	0	–1	–1	74.50	75.53
10	0	1	–1	78.90	79.45
11	0	–1	1	93.70	93.15
12	0	1	1	93.90	92.88
13	0	0	0	90.60	90.44
14	0	0	0	90.40	90.44
15	0	0	0	89.30	90.44
16	0	0	0	90.20	90.44
17	0	0	0	91.70	90.44

Model Regression and Analysis

The analysis of variance (ANOVA) for the statistical significance of the model is shown in Table . From Table , the high value of R2 of 0.9894 for the response implied that the model explained that about 98% of the experimental data were compatible with the predicted data. The regression coefficient of the predicted R2 was 0.8846 (>0.80), and the adjusted R2 was 0.9758 was achieved; the actual conversion was in reasonable agreement with the predicted conversion. It has been summarized from the data that the model was precise and reliable. The result of oleic acid conversion shows the F-value of 72.65, and a very low probability value (<0.0001) indicated that this model was statistically significant. The F-value of the “lack of fit” was 2.44, indicating that the model was not significant. Besides, the p-value of X1, X3, X12, and X32 terms had significant effects on the oleic acid conversion. Other model terms were not significant. Based on the above analysis, X1, X3, and X32 have positive effects, whereas the other variables have negative effects.

Table 3

Analysis of ANOVA for Response Surface Second-Order Modela

source	sum of squares	df	mean square	F-value	probability (P) > F
model	785.04	9	87.23	72.65	<0.0001	significant
X1	102.25	1	102.25	85.16	<0.0001
X2	6.66	1	6.66	5.55	0.0507
X3	482.05	1	482.05	401.49	<0.0001
X1X2	0.12	1	0.12	0.10	0.7587
X1X3	4.20	1	4.20	3.50	0.1035
X2X3	4.41	1	4.41	3.67	0.0968
X12	25.17	1	25.17	20.96	0.0025
X22	2.90	1	2.90	2.42	0.1641
X32	152.59	1	152.59	127.09	<0.0001
residual	8.40	7	1.20
lack of fit	5.43	3	1.81	2.44	0.2047	not significant
pure error	2.97	4	0.74
cor total	793.44	16

Rp2red = 0.8846, Radj2 = 0.9758, R2 = 0.9894.

Interactions among Variables

Three-dimensional response surface plots (Figure a–c) show the interaction effect of the process-independent variables with their response. Figure a shows the combined effect of the amount of catalyst and the reaction time on methyl oleate conversion. Based on Figure a and as expected, the oleic acid conversion was found to decrease with increasing reaction time at a constant catalyst amount because of the reversibility of the esterification reaction, and these also clearly demonstrated that the effect of reaction time on oleic acid conversion was more than the catalyst dosage. Moreover, the interaction between the molar ratio (methanol to oleic acid) and reaction time on the conversion is given in Figure b. According to Figure b, an increase in both methanol to oleic acid molar ratio and reaction time at 4.0 wt % Fe1PMo catalyst results in the increase of methyl oleate conversion, implying the time and molar ratio had a positive influence on oleic acid conversion. However, a slight decrease in the conversion with the rise in methanol to oleic acid molar ratio was also observed because excess methanol led to a relative reduction in the active sites of the catalyst.[42]Figure c exhibits the interaction of methanol to oleic acid molar ratio and catalyst dosage on methyl oleate conversion. It was also observed that changes in the amount of methanol affect the oleic acid conversion at a constant catalyst dosage. Nonetheless, excess methanol was inadvisable because of the resulting dilution of the catalyst and oleic acid.

Figure 4

Three-dimensional response surface plots for the esterification of oleic acid with methanol with varying reaction parameters using Fe1PMo catalyst: effect of reaction time and the catalyst dosage (a); effect of molar ratio of methanol to oleic acid and the reaction time (b); and effect of molar ratio of methanol to oleic acid and the catalyst dosage (c).

Optimum Reaction Conditions

Optimization of the esterification process to maximize oleic acid conversion was performed using the Design-Expert software. Meanwhile, based on various factors' influence on the design issues such as energy consumption, the operating and circulation costs, the assurance safety devices, etc.[43] The optimum conditions were found to be 3.9 h, 5.0 wt % of catalyst, and 12.54 molar ratio of methanol to oleic acid, and these optimum conditions were validated by the triplication of the experiments. The results showed that the predicted and experimental values of oleic acid conversion were found to be 95.99 and 95.1%, respectively, indicating they were close to each other.

Reusability

The reusability and stability studies of heterogeneous catalyst are of great significance. To test the catalyst, the oleic acid esterification reaction was conducted under optimal conditions: reaction temperature of 70 °C, the molar ratio of methanol to oleic acid of 10:1, catalyst dosage of 5 wt %, and reaction time of 3 h. After the first production cycle, the catalyst was simply separated by decantation and subjected to direct reuse in a new oleic acid esterification reaction cycle for five times, and the results are exhibited in Figure a. As presented in Figure a, the catalyst keeps its activity for the third use (89.2% in first reuse versus 70.2% in third reuse). However, with continuous increment of reuse cycles, the oleic acid conversion drops sharply. After five cycles, only 41.5% conversion is obtained. Meanwhile, different variations are observed in the FT-IR spectra between the fresh catalyst and the fifth reused catalyst in Figure b, indicating that the significant decrease in oleic acid conversion in the fifth cycle compared to that in the first cycle can be mainly attributed to the change in the Keggin structure of the catalyst. This might occur due to the loss of catalyst and the blockage of the channel during each cycle. Thus, from the above observation, the Fe1PMo catalyst can show good stability during esterification.

Figure 5

Reuse of the catalyst (a). Reaction conditions: molar ratio of methanol to oleic acid = 10:1, reaction temperature = 70 °C, catalyst dosage = 5 wt %, and reaction time = 3 h. The comparison of FT-IR spectra of fresh catalyst and reused catalyst (b).

Catalytic Activity of Fe1PMo for the Esterification of Other Substrates

Esterification of different free fatty acids (high free fatty acid, nonedible oils) with methanol was performed over Fe1PMo to test the generality of this doped catalyst. The conversion of lauric acid, myristic acid, palmitic acid, and stearic acid was 92.4, 89.1, 92.7, and 90.9%, respectively, and the catalytic tests on high acid value Jatropha curcas crude oil with methanol showed the pre-esterification rate of 90.6%. These results show that the Fe1PMo catalyst has great potential to catalyze various esterification reactions for industrial biodiesel production.

Conclusions

The present work studied the preparation and utilization of iron(III)-doped H3PMo catalysts for the production of biodiesel. The findings revealed that the Fe1PMo catalyst gave good catalytic activity for the conversion of oleic acid with methanol to biodiesel. Moreover, the various esterification reaction parameters were then optimized to obtain maximal conversion, and a high conversion of 89.2% can be achieved with single-factor optimization and 95.1% with RSM. In addition, the stability of the as-synthesized Fe1PMo catalyst was studied, and it showed good reusability. Thus, the Fe1PMo catalyst could be an environmentally benign and efficient catalyst for the production of biodiesel at an industrial scale.

Experimental Section

Materials

Oleic acid (AR), methanol (AR, >99%), lauric acid (AR, 98%), myristic acid (AR, 98%), palmitic acid (AR, 98%), stearic acid (AR, 98%), phosphomolybdic acid (AR), and iron(III) chloride hexahydrate (FeCl3·6H2O, AR) were purchased from Sinopharm Chemical Regent Co., Ltd. All other chemicals were of analytical grade and used as received, unless otherwise noted. Jatropha curcas seeds were purchased from Luodian County, Guizhou Province, and the Jatropha curcas crude oil (JCCO) was obtained through mechanical expression.

Preparation of Iron(III)-Doped H3PMo Solid-Acid Catalyst

Iron(III)-doped phosphomolybdic acid catalyst was prepared according to the previous literature,[25,33,44] with slight modifications. Briefly, various calculated amounts of phosphomolybdic acid and iron(III) chloride were dissolved in distilled water. Then, the aqueous solution of iron(III) chloride was slowly added to the solution of phosphomolybdic acid under vigorous stirring at room temperature for over 1 h. Subsequently, the mixture was heated at 70 °C for 3 h and finally dried under vacuum at 110 °C for 12 h to obtain the product as doped catalysts; these as-prepared catalysts are designated as Fe1/3H2PMo, Fe2/3H2PMo, and Fe1PMo according to the following equations

The XRD patterns of the catalysts were obtained using a Rigaku D/max 2000 ultima plus diffractometer (monochromatic nickel filter, Cu Kα radiation) and a step scan technique at 2θ angles of 5–80°. The FT-IR spectra were scanned on a PerkinElmer spectrum 100 using the KBr disk technique (4000–400 cm–1). TG analysis was carried out with a Netzsch STA 449F3 instrument in dry air at a heating rate of 10 °C/min. The surface morphology of the catalyst was studied using a JEOL-6701F scanning electron microscope at 10.0 kV.

Procedure for Esterification

The catalytic activity of the prepared catalysts was tested through the esterification of oleic acid (other free fatty acids and oils). Esterification reaction of certain amounts of oleic acid, methanol, and Fe1PMo catalyst was performed in a 100 mL single-necked glass flask immersed in a thermostatic oil bath equipped with a reflux condenser and mechanical stirrer; no inert atmosphere was used and all reactions occurred in contact with air and at ambient pressure. After esterification, the used catalyst was simply separated from the liquid phase by decantation and methanol was then removed by reduced pressure distillation to purify the product. The acid value of the product was tested using the national standard method ISO 660-2009, and the conversion was calculated using eq . All data presented are averages of at least duplicate experiments.where AV1 is the acid value of raw materials and AV2 is the acid value of the catalyzed product.