Synthesis of biodiesel from Jatropha curcas L . seed oil using artificial zeolites loaded with CH 3 COOK as a heterogeneous catalyst

An environmentally benign process was developed for the transesterification of Jatropha curcas L. seed oil with methanol using artificial zeolites loaded with potassium acetate as a heterogeneous catalyst. After calcination for 5 h at 823 K, the catalyst loaded with 47 wt.% CH3COOK exhibited the highest efficiency and best catalytic activity. The easily prepared catalysts were characterized by means of X-ray diffraction and IR spectroscopy, as well as Hammett indicator titration. The results revealed a strong dependence of catalytic activity on basicity. The optimum reaction conditions for transesterification of J. curcas oil were also investigated. The methyl ester content in the biodiesel product exceeded 91% after 4h reaction at reflux temperature in the presence of 2% solid catalyst and no water washing process is needed during workup.


INTRODUCTION
Biodiesel is a biodegradable and non-toxic renewable alternative to diesel fuel that is composed of mono-alkyl esters of long-chain fatty acids derived from vegetable oils or animal fats.Biodiesel is increasing in importance because of its benign impact on the environment.[1,2] Biodiesel is produced mainly through the transesterification of vegetable oils using short-chain alcohols, typically methanol or ethanol [3], because these are cheap and readily available from syngas, in which methanol is usually preferred.[4,5] Transesterification of vegetable oils with methanol, also called methanolysis, is typically carried out in the presence of homogeneous base or acid catalysts.Homogeneous base catalysts, including potassium hydroxide, sodium hydroxide and potassium and sodium alkoxides, such as NaOCH 3 [1], have higher catalytic activity than acid catalysts.Furthermore, since acid catalysts are more corrosive than base catalysts, base catalysis is usually preferred in commercial processes.[6] In the conventional homogeneous reaction, removal of the base after reaction is a major problem, since aqueous quenching results in the formation of stable emulsions and saponification, making the separation of methyl esters difficult and producing waste water.[7] However, the use of solid base catalysts can not only overcome those disadvantages, but also confer some advantages, such as the elimination of a quenching step (and associated contaminated water waste) during the work-up process and the scope to operate in continuous mode.[8,9] Therefore, environmentally friendly heterogeneous catalysts are promising for biodiesel production because of environmental constraints and potential simplification of existing processes.
Recently, heterogeneous catalysts used to catalyze the transesterification of vegetable oils to prepare fatty acid methyl esters (FAME) have attracted considerable attention.Wilson et al. [5] prepared a series of Li-promoted CaO catalysts with Li loading in the range 0.26-4.0wt.% to catalyze the transesterification of glyceryl tributyrate and methanol.They found that the optimum loading correlated with the formation of an electron-deficient surface Li + species and associated-OH species at defect sites on the support.Lee et al. [10] reported a process for the production of biodiesel from vegetable oils using an Na/NaOH/γ-Al 2 O 3 heterogeneous catalyst.Under optimized reaction conditions, this Na/NaOH/γ-Al 2 O 3 heterogeneous base catalyst showed almost the same activity as a conventional homogeneous NaOH catalyst.However, the base catalyst had to be prepared in special apparatus equipped with a nitrogen flow line and cold circulating water flow.These disadvantages restrict potential industrial applications.Xie et al. [11] developed a type of Al 2 O 3 -loaded KNO 3 solid base catalyst for the transesterification of soybean oil.The catalytic activity was ascribed to the presence of K 2 O and K-O-Al groups derived from KNO 3 or other potassium compounds during high-temperature calcination.
Zeolites have attracted much attention in the preparation of solid base catalysts.[12,13,14,15] The basicity and catalytic activity of these zeolites can be modulated through ion exchange of alkali and the occlusion of alkali metal oxides in zeolite cages by decomposition.However, the preparation procedures for these modified zeolites have some disadvantages.For ion-exchanged zeolites, the exchanged samples need to be washed with distilled, deionized water to remove excess alkali remaining in the ETS-10 zeolite cages.[13] Furthermore, the treatment process requires a long time and produces much waste water.
In the present study, a modified artificial zeolite for catalysis of the transesterification of Jatropha curcas L. seed oil was developed using a simple preparation process.This catalyst preparation technique has many advantages: (1) compared with ETS-10 and NaX zeolites, artificial zeolites are cheaper; (2) artificial zeolites are used directly as a support without any pretreatment; and (3) after impregnation with potassium acetate, artificial zeolite samples were dried and then simply calcined in a muffle furnace.Unlike NaX occluded catalysts, [13,14] artificial zeolite-supported potassium acetate was thermally decomposed in an uncontrolled manner and the catalyst sample was prepared easily.To the best of our knowledge, this is the first time that a catalyst of artificial zeolites loaded with CH 3 COOK has been adopted for biodiesel production from J. curcas oil, a plant belonging to the Euphorbiaceae family with seed oil that is non-edible and a good material for industrial biodiesel production.[16] After calcination for 5 h at 823 K, the artificial zeolite catalyst loaded with 47 wt.%CH 3 COOK exhibited the highest basicity and the best catalytic activity for the reaction.The catalytic activity was evaluated in terms of the methyl ester content of the product after transesterification of J. curcas oil.The optimum reaction conditions were determined using an orthogonal experimental method.

Catalyst Preparation
Artificial zeolites (Na 2 OAl 2 O 3 xSiO 2 yH 2 O) were obtained from Shanghai Qingxi Chemical Science and Technology Ltd. (Shanghai, China).CH 3 COOK/artificial zeolite supported catalysts were prepared by impregnating artificial zeolites with an aqueous solution of potas-sium acetate.Samples with various CH 3 COOK loadings were impregnated for 12 h to ensure that the CH 3 COOK diffused and dispersed thoroughly over the surface and through the pores of the artificial zeolites.The samples were then dried overnight at 383 K and calcined (typically at 823 K) in air for 5 h to yield readily usable catalysts.In experiments, catalyst samples with loading amounts of 9,23,33,41,44,47 and 50 wt.%(relative to the total mass of the catalyst before calcination) were designated as 9%, 23%, 33%, 41%, 44%, 47% and 50% CH 3 COOK/artificial zeolites, respectively.[11]

Catalyst Characterization
The basic strength (H_) of the solid base was assessed using Hammett indicators.[17,18,19] In our experiments, the following Hammett indicators were used: bromthymol blue (H_=7.2),phenolphthalein (H_=9.8),2, 4-dinitroa-niline (H_=15.0),and nitroaniline (H_=18.4).Approximately 50 mg of the catalyst sample was shaken with 10 mL of anhydrous ethanolic solution of Hammett indicator and left to equilibrate for 2 h.[20] Then the color change of the solution was observed.When the solution exhibits a color change, this indicates that the basic strength of the catalyst is stronger than the indicator used.However, when the solution produces no color change, the basic strength of the catalyst is weaker than that of the indicator used.The basicity of the catalysts was determined by titration with a Hammett indicator and a 0.02 mol/L anhydrous ethanolic solution of benzene carboxylic acid.[6] It should be noted that Hammett indicator titration can only give qualitative information about the basic properties of catalysts.
X-Ray diffraction (XRD) patterns of the samples were recorded on a Rigaku D/MAX-2200 powder X-ray diffractometer with Cu K  (λ=0.154nm)radiation using an acceleration voltage of 40 kV and a current of 30mA, over a 2 range of 0-65° with a step size of 0.04° at a scanning speed of 3°/min.The data were processed using DiffracPlus software.The phases were identified using the Power Diffraction File (PDF) database (JCPDS, International Center for Diffraction Data).
IR spectra of the samples were measured using the KBr pellet technique.Spectra were recorded on a Shimadzu IR-Prestige-21 spectrometer with resolution of 4 cm -1 over the range 400-4000cm -1 .

Transesterification Reaction
Jatropha curcas oil was prepared by squeezing seeds of J. curcas from Luodian county, Guizhou Province, southwest China.The crude oil obtained was then further purified by filtering out solid impurities and refined to reduce the water content.According to gas chromatography (GC) analysis, the fatty acids of J. curcas oil consisted of: palmitic acid (12.47%), palmitoleic acid (2.10%), stearic acid (6.42%), oleic acid (32.04%), and linoleic acid (42.47%).The acid value of the oil was approximately 2.63mg KOH/g, and an average molecular weight of 880g/mol was calculated from the saponification value (Sv=194mg KOH/g).
A 250-mL three-necked glass flask equipped with a water-cooled condenser, thermometer and magnetic stirrer was charged with 110.0g (125mmol, calculated from the average molecular weight) of J. curcas oil, different volumes of methanol and various amounts of catalyst.The mixture was vigorously stirred and refluxed for the required reaction time.After completion of methanolysis, the mixture was filtered and excess methanol was recovered by rotary evaporation.The liquid phase was transferred into a separatory funnel and allowed to settle; the upper layer, the biodiesel product, was analyzed by GC.

Basic Strength of the Catalyst
Table 1 shows the basic strength of the parent artificial zeolite (entry 1) and various CH 3 COOK/artificial zeolite catalysts calcined at different temperatures.The basic strength of catalyst samples with CH 3 COOK loading of ＜33wt.%and calcined at 823 K are in the range 7.2<H_ ＜9.8 (entries 2-4).The basic strength increased to 15.0＜ H_＜18.4 when the CH 3 COOK loading exceeded 33 wt.% (entries 5-＜9).This indicates that there are at least two types of active base sites in the supported catalysts.
For loading of <33wt.%, the amount of CH 3 COOK is not enough to cover all of the support surface and incorporation of potassium ions into the vacancies of the support is mainly through strong salt-support or oxidesupport interaction.[20] New active base species emerge when CH 3 COOK loading exceeds 33wt.%,indicating that (CH 3 COOK) n -support interaction plays an important role in this process.However, the basic strength of 47 wt.%CH 3 COOK/artificial zeolite calcined at a temperature less than 623 K showed no increase compared with the artificial zeolite (entries 1, 10, 11).This is likely because CH 3 COOK loaded onto the artificial zeolite is not decomposed at temperature less than 623 K.The catalysts calcined between 623 and 973 K exhibit similar basic strength (entries [12][13][14][15].According to the definition of Tanabe, [21] these catalyst samples, with base strength in the range 15.0＜H_＜18.4,can be regarded as strong bases.).This is similar to the findings of Xie et al. [11] The calcination temperature also affects catalyst basicity.It is evident from Table 2 that catalysts calcined at 823K had the highest basicity (entries 7, 9-14).It was noted that 47% CH 3 COOK/artificial zeolite without calcination had higher basicity than 23% CH 3 COOK/artificial zeolite calcined at 823 K (0.2079 vs. 0.1393 mmol KOH/g, entries 9 and 4).).This is similar to the findings of Xie et al. [11] The calcination temperature also affects catalyst basicity.It is evident from Table 2 that catalysts calcined at 823 K had the highest basicity (entries 7, 9-14).It was noted that 47% CH 3 COOK/artificial zeolite without calcination had higher basicity than 23% CH 3 COOK/artificial zeolite calcined at 823 K (0.2079 vs. 0.1393 mmol KOH/g, entries 9 and 4).

Lnfluence of CH 3 COOK Loading
The catalytic activity of the catalysts was evaluated by comparing the FAME content of the transesterification products.The effect of CH 3 COOK loading on the cata-lytic activity is shown Fig. 1.As the CH 3 COOK loading increases from 9% to 47%, the methyl ester content increases.The highest methyl ester content (91.58%) was obtained for 47 wt.%CH 3 COOK loading.However, the FAME content decreased slightly for a further increase in CH 3 COOK loading, which may be due to partial covering of basic sites by K 2 O species from the excess CH 3 COOK on the surface of the composite.It is important to point out that the conversion decreased significantly when potassium salt loading exceeded the critical limit for the transesterification of soybean.[11]

Influence of Calcination Temperature
Figure 2 shows the influence of calcination temperature on the catalytic activity in the temperature range from 523 to 923 K.The optimal calcination temperature observed was 823 K.At this temperature, the methyl ester content reached 91.58%.However, when the calcination temperature increased to 923 K, the catalyst activity decreased.Such results are in line with the basicity properties shown in Table 2, indicating that higher basicity results in higher conversion and higher methyl ester content.However, the FAME content obtained using 47 wt.%CH 3 COOK/artificial zeolite calcined at 523 K is lower than that obtained using 23 wt.%CH 3 COOK/artificial zeolite calcined at 823 K, although the basicity of the former is higher than that of the latter (0.2540 vs. 0.1393 mmol KOH/g).This implies that basicity is the most important, but not the only factor affecting the activity of these supported catalysts.

FTIR Analysis
IR spectroscopy was used to investigate the CH 3 COOK/ artificial zeolite catalysts.The results are shown in Fig. 3.
For the parent artificial zeolite, two absorption bands at 3440 and 1650cm -1 were attributed to stretching and bending vibrations of physically absorbed water, respectively.CH 3 COOK/artificial zeolites catalysts calcined at high temperature (curves b, c, d, f and g in Fig. 3) showed intense absorption at approximately 3440 cm -1 , which could be assigned to ν OH stretching vibrations of hydroxyl groups attached to the support.Such hydroxyl groups are mainly formed by the reaction of surface-absorbed water with the support during the activation by calcination.[22] When the catalyst samples were calcined at higher temperatures (curves d, f and g in Fig. 3), there was very little absorbed water on the support surface.However, there was still a strong absorption peak at1650cm -1 .This indicates that the peak at 1650cm -1 was mainly due to surface hydroxyl groups and that these surface OH groups were possibly active sites, as reported by Xie and Huang.[23] Furthermore, the absorption broad band at 3440cm -1 could be partly assigned to stretching vibrations of Al-O-K or Si-O-K groups.[24,25,26] According to Stork and Pott, [27] K + ions may replace the protons of hydroxyl groups attached to the support during activation.Thus, K + ions from CH 3 COOK could form Al-O-K or Si-O-K groups by replacing the protons of hydroxyl groups attached to the artificial zeolites during activation by calcination, and can probably be considered to be another active basic species of this catalyst in transesterification.
As observed from Fig. 3, the absorption intensity at 3440cm -1 increased with the CH 3 COOK loading (curves b-d), indicating that an increase in basic sites resulted in an increase in catalytic activity, which is in accordance with the results shown in Fig. 1.On the other hand, the intensity of the abso0rption at 3440cm -1 decreased at higher calcination temperatures (curves d, f and g), indicating a decrease in basic sites, in line with Hammett basicity measurement (Table 2).The absorption peak at 1552cm -1 can be assigned to asymmetric vibration of COO - [28] of CH 3 COOK (curve e), which disappeared after calcination at 823K (curve d), demonstrating that CH 3 COOK decomposed completely after calcination at 823 K.In addition, there were some other absorption peaks at 400-1400cm -1 for all samples, which can be attributed to Al-O (or Si-O) symmetric stretching and ring vibration in Al-O and Si-O tetraheda formed through the oxygen atom.

XRD Analysis
XRD patterns of artificial zeolites and CH 3 COOK/artificial zeolite samples with different CH 3 COOK loading are shown in Fig. 4. The XRD pattern of artificial zeolites is irregular, which may be related to their structure.After calcination of artificial zeolites loaded with CH 3 COOK, more regular and distinct diffraction peaks (* and □ in Fig. 4) appeared in the XRD patterns.This is possibly due to the reaction of CH 3 COOK with the support during activation, resulting in the more regular and stable structure necessary for catalysis.
As shown in Fig. 4, when the CH 3 COOK loading was below 47wt.%(curves b and c), the XRD patterns contained only diffraction peaks (2 =20.5°,22.3°,28.6°,34.6°,40.6°)assigned to new species formed by reaction between CH 3 COOK and the support during the activation process.These diffraction peaks can probably be ascribed to two new species, KAlSiO4 (Kalsilite; *) and K 0.85 Na 0.15 AlSiO 4 (□), formed by the movement of K + from CH 3 COOK into the crystal lattices of the support and subsequent reaction.There were no characteristic peaks for CH 3 COOK or K 2 O, probably due to good dispersion of K + derived from CH 3 COOK on artificial zeolites in the form of various compounds, while a K 2 O phase undetectable by XRD may have dispersed onto the artificial zeolite surface as a monolayer.[29] In addition, 2 diffraction peaks at 32.8° and 58.6° can probably be ascribed to K 2 Al 2 O 4 (•), also obtained by reaction between K + and the support.
However, when the CH 3 COOK loading increased to 47 wt.%, the characteristic XRD peaks of K 2 O (2= 25.8°,41.9°)were detected (curve d in Fig. 4).According to the results shown in Fig. 1 and Table 2, this K 2 O species may account for the high catalytic activity and basicity of the catalyst, because 47% CH 3 COOK/artificial zeolites calcined at 823K for 5h exhibited the highest catalytic activity and basicity.Furthermore, when the calcination temperature increased to 923K, the characteristic XRD peak at 2 =32.8° vanished, presumably because K 2 Al 2 O 4 (•) was destroyed at higher calcination temperatures.
Taken together, the characterization results indicate that K 2 O (derived from CH 3 COOK) and surface hydroxyl groups, as well as Al-O-K (or Si-O-K) groups, were probably the active sites mainly responsible for the transesterification of J. curcas oil with methanol.

Optimization of the Transesterification Reaction
For the transesterification reaction, orthogonal experiments were carried out to determine the optimum reaction conditions.The effects of various factors on the reaction were also studied.The orthogonal scheme chosen and the data obtained are shown in Table 3.
As shown in Table 3, the extent to which the transesterification of J. curcas oil was affected in terms of the range (R) value was A＞B＞C, namely, oil/methanol ratio (A) first, followed by catalyst amount (B) and reaction time (C).Taking the FAME content into account, the optimum reaction conditions were A 2 B 2 C 2 ; namely, oil/methanol ratio, 1:10; catalyst amount, 2wt.%; and reaction time, 4h.
The molar ratio of J. curcas oil to methanol is the most important factor affecting the transesterification to methyl esters (Table 3).Theoretically, three moles of methanol are required for one mole of triglyceride to yield one mole of glycerol and three moles of FAME.However, a slight excess of methanol is required to drive the equilibrium to the product side, because the transesterification reaction is reversible.[1] With an increase in the oil/methanol molar ratio, the conversion increased considerably.However, further addition of methanol not only had no significant effect on the conversion, but also seriously affected glycerine separation due to the increase in solubility.[30] Thus, the optimum molar ratio of J. curcas oil to methanol to produce methyl esters was 1:10.The amount of catalyst is another important factor that affects the reaction.With no catalyst, transesterification does not occur.When the amount of catalyst was insufficient, the FAME content was very low (entries 1, 4, 7).Saponification took place when the amount of catalyst was increased to 3%, leading to product emulsion and making separation difficult.The FAME content and product yield were also influenced (entries 3, 6, 9).As revealed in orthogonal experiments, 2 wt.% catalyst was appropriate for this transesterification reaction.
The final factor affecting transesterification is the reaction time.When the reaction time is too short, the conversion of vegetable oil to methyl esters is incomplete, and conversion increases with the reaction time.However, when the reaction reaches equilibrium, prolonged reaction time does not increase the FAME content of the product, but increases the cost for biodiesel production.In homogeneous base-catalyzed transesterification, the reaction time is usually less than 1 h.For heterogeneous transesterification the reaction time needs to be longer, because the system components, the base catalyst, methanol and J. curcas oil in the present study, require a longer contact time.According to orthogonal experiments, an optimum reaction time of 4 h was chosen.
An experiment was carried out under above optimum reaction conditions, a product yield of 94.27% and methyl ester content of 91.58% were obtained after transesterification.

CONCLUSIONS
Easily prepared solid-base catalysts using artificial zeolites as a support were developed for the transesterification of J. curcas oil with methanol to produce biodiesel.The CH 3 COOK/artificial zeolites solid catalysts exhibited high catalytic activity in the transesterification process.The methyl ester content exceeded 91% when the catalyst with 47 wt.%CH 3 COOK was calcined at 823 K for 5 h.Catalyst characterization revealed that K 2 O, surface hydroxyl groups and Al-O-K (or Si-O-K) groups are the main basic sites.Furthermore, the activity of the catalysts depends strongly on their basicity.Or-thogonal experiments revealed the following optimum reaction conditions: oil/methanol ratio, 1:10; catalyst amount, 2wt.%; and reaction time, 4h.No water washing step is needed thus no waster water was produced during this process.The information mentioned above establishes certain basis for industrial biodiesel production from J. curcas seed oil using the heterogeneous catalysts.Further investigations on the reaction mechanism and active sites are under way in our laboratory.

Table 2
summarizes the basicity of a series of CH 3 COOK/artificial zeolite catalysts calcined at different temperatures, as determined using Hammett indicators.As shown in

Table 2 ,
the basicity of catalysts calcined at 823 K first increases and then decreases with the increasing CH 3 COOK loading (entries 2-8), with 47 wt.%CH 3 COOK/artificial zeolite exhibiting the highest

Table 1 .
Basic strength of various CH 3 COOK/artificial zeolite catalysts calcined at different temperatures.

Table 2 .
Basicity of various CH 3 COOK/artificial zeolite catalysts calcined at different temperatures.

Table 3 .
Results of orthogonal experiment L9_3_4 for basecatalyzed transesterification and range analysis.