Applied Catalysis A: General 378 (2010) 160–168 Contents lists available at ScienceDirect Applied Catalysis A: General journal homepage: www.elsevier.com/locate/apcata Biodiesel production process by homogeneous/heterogeneous catalytic system using an acid–base catalyst Anastasia Macario a,∗ , Girolamo Giordano a , Barbara Onida b , Donato Cocina b , Antonio Tagarelli c , Angelo Maria Giuffrè d a Department of Chemical Engineering & Materials, Università della Calabria, Rende (CS), Italy b Department of Materials Science & Chemical Engineering, Politecnico di Torino, Torino, Italy c Department of Chemistry, Università della Calabria, Rende (CS), Italy d Department of Biotechnology, M.A.A., Università Mediterranea di Reggio Calabria, Reggio Calabria, Italy article info Article history: Received 1 January 2010 Received in revised form 9 February 2010 Accepted 10 February 2010 Available online 18 February 2010 Keywords: Biodiesel Transesterification Esterification Zeolites Acid–base catalysis abstract The transesterification of triglycerides contained in waste oilseed fruits with methanol has been studied in heterogeneous/homogeneous systems using acid and base catalysts. The acid catalysts (strong acid catalysts: USY, BEA, FAU-X, and weak acid catalysts: MCM-41 and ITQ-6 with Si/Al= ∞) were prepared by hydrothermal synthesis procedures. In order to obtain acid–base catalysts, potassium was loaded on different materials by ionic exchange (obtaining K-MCM-41, K-ITQ-6,). XRD, ICP-MS, IR after CO and CO 2 adsorption, thermal analyses and N 2 adsorption/desorption techniques have been used for catalysts characterization. The highest triglycerides conversion and biodiesel yield values were achieved by K-ITQ- 6 catalysts, after 24 h of reaction at 180 ◦ C. Deactivation of this catalyst occurs for potassium leaching, but its regeneration and reuse are feasible and easy to perform. A possible continuous biodiesel production process has been proposed. © 2010 Elsevier B.V. All rights reserved. 1. Introduction The transestrification of vegetable oils, catalyzed by either acids or bases, consists in three consecutive and reversible reactions in which triglycerides are converted first to diglycerides, then to monoglyceride and finally to glycerin. In each steps, one ester is formed. In the overall reaction, using methanol as alcohol, 3 mol of methyl esters are produced for each mole of triglyceride. Industrial processes use 6 mol of methanol for each mole of triglyceride obtaining methyl esters as biodiesel mixture (FAME =Fatty Acid Methyl Esters). This large excess of methanol ensures that the reaction is driven in the direction of methyl esters. The methyl esters mixture (or Biodiesel fuel) has similar proper- ties of fossil diesel fuel (cetane number, kinematic viscosity) but it does not contain petroleum products and sulfur compounds. Fur- thermore, it possesses a higher flash point (>130 ◦ C) with respect the conventional diesel. For these reasons biodiesel is recognized as “green fuel”. Finally, it is almost neutral towards CO 2 emission because its renewable sources (vegetable oils and biomass). Industrial current production of biodiesel is carried out by homogeneous alkali-catalyzed transesterification of vegetable oils ∗ Corresponding author. Tel.: +39 0984 49 66 67; fax: +39 0984 49 66 55. E-mail address: macario@unical.it (A. Macario). with methanol, using sodium hydroxide, potassium hydroxide or potassium methoxide as catalyst [1]. The homogeneous basic transesterification shows a very fast kinetic of reaction but also a collateral saponification reaction that reduces the biodiesel pro- duction efficiency. To prevent the biodiesel yield loss due to the saponification reaction, oil and alcohol must be dry and the oil should have a minimum amount of free fatty acids (less than 0.1 wt%). Biodiesel is finally recovered by repeated washing with water to remove glycerol, soap and the excess of methanol. By contrary, the acid transesterification allows to obtain a biodiesel production without formation of by-products. The drawbacks of an acid homogeneous transesterification are the corrosive catalyst (H 2 SO 4 ,H 3 PO 4 , and HCl) and the slow reaction rate. This may be increased at high temperature and pressure, involving larger costs [2]. Methanol and oil are poorly soluble, so the reaction mixture contains two liquid phases. Others alcohols can be used, but being more expensive. Moreover, an acid pre-treatment is often carried out in the homogeneous alkaline-transesterification of oils having more than the 5 wt% of free fatty acids, in order to improve the biodiesel efficiency production [1–3]. Recent works, also, demon- strate as a heterogeneous enzymatic catalyst represents a potential solution to produce biodiesel from very low-quality triglycerides source [4,5], but in these cases the cost of the enzymes has to be considered. All these aspects suggest that an environmental- friendly and cheaper biodiesel production process could be carried 0926-860X/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.apcata.2010.02.016 A. Macario et al. / Applied Catalysis A: General 378 (2010) 160–168 161 out using acid or basic heterogeneous catalysts or, better, hetero- geneous catalyst with both acid and basic properties. They could combine the advantagesof the alkalineand acid transesterifications with those of heterogeneous catalytic process. The suitable catalyst should possess high activity and selectivity, high water-tolerance, high stability, it should be inexpensive and its production process should beenvironmental friendly. The activity and selectivity prop- erties of catalyst generally depend on the amount and the strength of acid or basic sites. Towards organic reaction (like transesterifica- tion), catalysts with high hydrophobic surface area are preferable because otherwise water can interact with active sites preventing the adsorption of organic reactants.As it is well known,zeolites and related materials are suitable materials for these purposes because they can be easy synthesized and modified in order to affect the acidity, basicity and hydrophobicity of their surface. In this contest, our work focus on the possibility to use strong acid zeolites (USY,BEA, FAU-X), weakacid materials (materialswith high density of silanol groups: pure silica mesoporous materials and delaminated zeolites) and materials with acid–base proper- ties (K-MCM-41, K-ITQ-6) as catalysts for the transesterification of waste fruit oilseeds with high amount of free fatty acids (FFAs), higher than 5 wt% (as low-quality and cheap triglycerides sources), with methanol to biodiesel. There are many different materials developed as catalysts for the transesterification of triglycerides to biodiesel, such as solid acids [6–11] (Amberlyst-15, H-Zeolites, Cs-heteropoly acids) or solid basic [9,11–19] (KOH-NaX, KI-Al 2 O 3 , Na/NaOH/␥-Al 2 O 3 , ETS-10, CaO, NaCs-X, KOH-Al 2 O 3 ). All these materials show interesting results but only in respect with acid or basic catalytic aspect. Moreover, in order to achieve good catalytic performance, some of them can used only in strong conditions, such as high temperature, high methanol content or in presence of extracting co-solvent, or starting from high quality triglycerides source. While, the conversion of biological feedstocks to biodiesel using mesoporous calcium silicate mixed oxides, as heterogeneous catalysts with acids and basics, is recently published and patented by Lin et al. [20,21]. Besides the usual zeolites characterization techniques (XRD, TG-DTA, ICP-MS, N 2 adsorption), the study of the adsorption of CO and CO 2 on the catalysts has been carried out by IR spec- troscopy, in order to identify the nature and the strength of the catalytic active sites. The reaction conditions were not optimized for the highest reaction yield: all catalytic tests provided to com- pare the activities of different catalysts. The catalyst showing the best catalytic performances (triglycerides and FFAs conversion and biodiesel yield) was regenerated and used for more than one reac- tion cycle. The efficiency of catalyst regeneration procedure has been evaluated in order to perform a possible continuous biodiesel production process. Chemical composition (free fatty acids, mono- glycerides, diglycerides and triglycerides content) and total acidity of the biodiesel final product have been measured. 2. Materials and methods 2.1. Materials The silica sources used for catalysts preparation were: pre- cipitated silica (BDH) and silica fumed (Aerosil 200 (Degussa) or silica fumed (Aldrich)). The structure directing agents used were: cetyltrimethylammonium bromide (CTABr, Aldrich) for meso- porous materials. The mineralizing agents were: sodium hydroxide (Carlo Erba), tetramethylammonium hydroxide (TMAOH, 25%, Fluka) and tetraethylammonium hydroxide (TEAOH, 40%, Fluka). Aluminum hydroxide (98%, Aldrich) was used as metal sources for Al–BEA catalyst preparation,while sodium aluminate(NaAlO 2 , 99%, Carlo Erba) was used as aluminum source for FAU-X preparation. Waste fruit oilseeds with oleic acid (C18:1, 39 wt%) and linoleic acid (C18:2, 30 wt%) as main FFAs content (total free acidity 5.58%) is used as triglycerides source. 2.2. Catalysts preparation Pure silica MCM-41 type material was prepared starting from a gel with the following molar composition: 1SiO 2 –0.26TMAOH–0.12CTABr–40H 2 O The crystallization time and temperature were, respectively, 24 h at 140 ◦ C in autoclave. Delaminated zeolite ITQ-6 (Si/Al = ∞) sample was prepared by swelling the laminar pure silica PREFER according to the procedure described by Corma et al. [22]. USY zeolite has been supplied by UOP Molecular Sieves. BEA catalyst was prepared by hydrothermal synthesis starting from the following molar gel: 50SiO 2 –10TEAOH–2Na 2 O–0.5Al 2 O 3 –350H 2 O time and temperature of crystallization were, respectively, 5 days at 150 ◦ C in autoclave. The molar composition of the gel for hydrothermal synthesis of FAU-X zeolites was the following: 1SiO 2 –0.6Na 2 O–0.1Al 2 O 3 –40H 2 O After the synthesis, the solid phases of all syntheses were recov- ered by filtration and washed with distilled water. The samples were calcined in air flow at 550 ◦ C for 8 h (heating rate: 5 ◦ C/min). Finally, samples containing potassium (K-MCM-41, K-ITQ-6) were obtained by ionic exchange carried out on correspondent cal- cined material at 60 ◦ C, for two times, with KCl 1 M solution and a ratio solid/solution equal to 0.01 g/ml. After ionic exchange, the dried samples were activated at 300 ◦ C for 8 h. Commercial anhy- drous potassium silicate in powder was supplied by Alfa Aesar Co. 2.3. Catalysts characterization Powder X-ray diffraction (XRD) data were recorded using a Phillips PW 1710 diffractometer with CuK ␣ radiation. The samples were scanned in the range of 2Â from 1 ◦ to 8 ◦ (for mesoporous materials) or from 5 ◦ to 45 ◦ (for microporous materials) in steps of 0.005 ◦ with a count time of 1 s at each point. XRD analyses of all synthesized samples show the characteristicdiffraction peaks (pat- terns not reported), confirming that the expected phases have been obtained for all materials. BET surface area and physical properties of samples were evaluated by N 2 adsorption/desorption isotherms carried out at 77 K on a Micromeritics ASAP 2020 sorption ana- lyzer. Thermal decompositions of as-synthesized samples were investigated by SHIMADZU DTG-60 instrument, between 20 ◦ C and 850 ◦ C, at a ramp of 5 ◦ C/min in air with a flow rate of 5ml/min. Chemical composition of samples was evaluated by an Elan DRC-e ICP-MS instrument (PerkinElmer SCIEX). Samples were introduced by means of a quartz nebulizer. The ICP torch was a standard torch (Fassel type torch) with a platinum injector. For the quantitative analysis, calibration curves were built on six different concentra- tions in a calibration range of 1–5000 ␮g/l and having composition similar to that of solution samples. Standard solutions were pre- pared by diluting a solution of Na and K (1000 mg/l). For IR measurements,self-supporting waferswereprepared and activated under dynamic vacuum (10 −4 Torr) for 1 h at 573K, in an IR cell allowing in situ thermal treatments, gas dosage and IR measurements to be carried out both at room temperature and at a nominal temperature of 77K, presumably in fact around 100 K. 162 A. Macario et al. / Applied Catalysis A: General 378 (2010) 160–168 Spectra were collected on a Bruker IFS 55 Equinox instrument equipped witha MCT cryodetector working with 2 cm −1 resolution. Difference spectra are obtained after subtraction of the spectrum of the naked sample. 2.4. Transesterification reaction The alcoholysis of triglycerides with anhydrous methanol (99.9%, Sigma–Aldrich) performed in batch Teflon-steel autoclaves immersed in a temperature-controlled bath. The reaction temper- ature was varied from 100 ◦ C to 180 ◦ C. In a typical experiment 5.0 g of oil were mixed with 2.74 ml of methanol (molar ration oil:methanol equal 1:20) and 0.15 g of catalyst (5wt% respect to oil). Then the mixture was transferred in the autoclave and stirred (650 rpm). The autoclave was close and heated in the bath until the desired temperature.The reaction was been carriedout underauto- geneous pressure in the batch-autoclave system for the required time. To analyze the reaction progress, products and reactant were separated from catalyst and glycerin by centrifugation in hexane (95%, Sigma–Aldrich). To analyze the chemical composition of the biodiesel produced we used theASTM D-6584-00 method.Through this method, it is possible to quantify triglycerides (TG), diglyc- erides (DG), monoglycerides (MG), free fatty acids and methyl ester by unique chromatogram. The Rtx ® -Biodiesel fused silica capil- lary column (10m × 0.32 mm× 0.10␮m) was used in an Agilent 6890 GC instrument equipped with FID detector. Before analysis, trimethylsilyl-trifluoroacetamide (98%+Acros Organics) wasadded to the mixture as silyating agent. Tricaprin (Sigma–Aldrich, puriss. p.a. standard forGC) was added as external standard. 1␮l of sample was injected and analyzed by the following oven temperature pro- gram: 1min at 140 ◦ C, 10 ◦ C/min to 360, 8min at 360 ◦ C. To correct the area obtainedfrom the GC,inorder tocalculatethe exact triglyc- erides conversion and biodiesel yield, the response factors for each compounds havebeencalculated usingthe correspondent standard compound (puriss. p.a. standard for GC): oleic acid, methyl oleate, monoglycerides, diglycerides and triolein. The biodiesel yield (in percentage) was calculated according to the following equation: B y =  A Bio /R Bio [(A Bio /R Bio ) + (A FFA /R FFA ) + 2(A DG /R DG ) + 3(A TG /R TG )]  ×100 where A i is the area of the peak correspondents to each compounds and R i is the related response factor. Total acidity value of final biodiesel mixturewas calculated bythe StandardEuropean Method EN14104 and compared with the European Quality Standard Spec- ification of Biodiesel EN14214-UNI10946. 2.5. Recovery and regeneration of catalyst At the end of each reaction cycle, two centrifugation processes were carried out in order to separate the catalyst from the reaction mixture and the glycerol to biodiesel mixture. The recovered cata- lyst has been washed with ethanol, in order to remove all organic compounds (triglycerides and esters tracks), and then with water, in order to remove ethanol. After drying at 120 ◦ C for 1 night, the catalyst was reused for another reaction cycle. In order to check the presence of organic compounds, a thermal analysis has been carried out on the regenerated catalysts. 3. Results and discussion 3.1. Catalytic test results: screening of all tested catalysts For a first comparison among all catalysts tested, the same reaction conditions were employed for the starting experiments: Fig. 1. Catalytic performance (triglycerides conversion and biodiesel yield) obtained by all catalysts tested at 100 ◦ C, at 5wt% respect to the oil, for 24 h and with a molar ratio between oil:methanol equal to 1:20. 100 ◦ C, 24h of reaction, 650 rpm, 5 wt% of catalyst respect to the triolein, molar ratio equal to 1:20 between oil and methanol. As observed from Fig. 1, the strong acid catalysts (USY, BEA and FAU-X) do not reach an appreciable triglycerides conversion in the reaction conditions. Commercial potassium silicate, appears as the best catalysts showing a triglycerides conversion of the 82% and biodiesel yield of the 79%. Among the synthesized catalysts reported in Fig. 1, the most promising are K-MCM-41 and K-ITQ-6. Their triglycerides conversion is similar, 58% and 64% respectively, whereas the biodiesel yield differs significantly, being 56% for K- ITQ-6 and only 3% for K-MCM-41. The pure silica mesoporous material, Si-MCM-41, has not been used as catalyst in the mentioned reaction due to very weak acidity of its surface. While, concerning the pure silica delaminated zeolite ITQ-6, due to its high weak acid silanols groups density, its catalytic activity at 100 ◦ C until 24 h of reaction can be detected but, in any case, has been low: the triglycerides conversion was closed to 20%. Starting from these results, in order to improve the catalysts per- formance, time and temperature of reaction have been increased. Operating at 180 ◦ C and following the reaction for 72 h, for acid catalysts (FAU-X, BEA, USY) an increasing of triglycerides conver- sion has not been observed. Finally, large increase of triglycerides conversion and biodiesel yield have been observed for K-MCM-41 and K-ITQ-6 catalysts, maintaining the temperature at 100 ◦ C and increasing the reaction time until 72h (see Fig. 2(a)). In particu- lar, K-ITQ-6 catalyst gave a triglycerides conversion of 98% and a biodiesel yield of 74% (Fig. 2(b)). The triglycerides conversion of K-MCM-41 reached a value of 90%, but the main products were the free fatty acids (32%) and monoglycerides (42%) whereas the biodiesel yieldgrew only up to the 15%. Insteadfor K 2 SiO 3 catalyst a faster reaction kinetic has been observed: after 17 h of reaction, the triglycerides conversion was 73% (Fig. 2(a)) and the biodiesel yield was the 68% (Fig. 2(b)). After almost 30 h of reaction, the potassium silicate shows its best catalytic performance (83% of triglycerides conversion and 80% of biodiesel yield) and they do not change for longer reaction time. A further increase of the reaction temperature (up to 180 ◦ C) for K-ITQ-6 caused a faster kinetic of transesteri- fication (Fig. 3). At 180 ◦ C, the K-ITQ-6 catalyst gave the 97% of triglycerides conversion and the 80% of biodiesel yield after 24 h of reaction. Following the reaction until 96 h, the triglycerides con- version was observed to grow up to 99%, the biodiesel yield and the FFAs conversion to increases up to 90%. The composition of the biodiesel mixture at different reaction time is reported in Table 1. At 180 ◦ C, the amount of free fatty acids, tri-, di- and monoglyc- erides decreases and simultaneously the amount of methyl esters increases when reaction time increases up to 48h. In these con- A. Macario et al. / Applied Catalysis A: General 378 (2010) 160–168 163 Fig. 2. Triglycerides conversion and biodiesel yield, as function of time, obtained at 100 ◦ C, with the 5 wt% of catalyst respect to the oil and with a molar ratio between oil:methanol equal to 1:20. Fig. 3. Triglycerides conversion, biodiesel yield and FFAs conversion, as function of time, obtained using K-ITQ-6 catalyst at 180 ◦ C, at 5 wt% respect to the oil and 1:20 molar ratio between oil:methanol. ditions, the triglycerides conversion is almost complete, whereas the biodiesel yield only slightly increases, because the reaction proceeds with the conversion of mono- and diglycerides to free fatty acids (see Table 1). The temperature increasing does not affect appreciably the catalytic performance of pure silica ITQ-6 material: the triglycerides conversion grew to38% but the biodiesel yieldwas closed to 3%. This means, most probably, that the weak acid silanol groups of the ITQ-6 surface are able to esterificate the free fatty acids but they are not able to transesterificate the triglycerides. Fig. 4. Comparison between K-ITQ-6 and commercial KOH catalysts. Conditions for KOH: 24 h at 70 ◦ C, methanol:oil molar ratio 10:1—catalyst amount: 5 wt%; condition for K-ITQ-6: 48 h at 180 ◦ C, methanol:oil molar ratio 20:1—catalyst amount: 5 wt%. Further investigations have not been applied to this catalyst due to its low catalytic performance. Due to the promising results obtained by K-ITQ-6 catalyst, a comparison of its catalytic performance with those of commercial KOH homogeneous basic catalyst could be interesting. For this pur- pose, in the Fig. 4 the biodiesel yield, the FFAs and the triglycerides conversion obtained for K-ITQ-6 and KOH catalysts are reported. The main result noticeable by this comparison is, even if the time and temperature of reaction is different for two catalysts, when the triglycerides conversion is complete, K-ITQ-6 permits to obtain higher biodiesel yield and higher FFAs conversion than KOH cat- alyst. Moreover, no soap products are detectable in the reaction mixture usingK-ITQ-6 catalysts by GC analysis (results notshowed) Table 1 Triolein conversion, biodiesel yield and chemical composition of biodiesel mixture obtained using 5wt% of K-ITQ-6 catalyst at 180 ◦ C and with triolein:methanol molar ratio 1:20. Time [h] Triglycerides conversion [%] Biodiesel yield a [%] Methyl esters b [%] Free fatty acids [%] MG c [%] DG c [%] TG c [%] 17 94 65 67.2 12.5 15.6 2.1 2.6 24 96.7 79.7 82.4 6.08 9.23 1.14 1.13 48 98 87 88.7 2.41 7.97 0.32 0.63 a B y = {(A Bio /R Bio )/[(A Bio /R Bio )+(A FFA /R FFA )+2(A DG /R DG )+3(A TG /R TG )]} × 100. b Weight percentage of methyl esters, obtained directly by GC, after correction with response factors. c EN14214-UNI10946 values [wt%]: MG = 0.8; DG = 0.20; TG = 0.20. 164 A. Macario et al. / Applied Catalysis A: General 378 (2010) 160–168 Fig. 5. N 2 adsorption isotherms of mesoporous (a) and delaminated ITQ-6 (b) catalysts, before and after ionic exchange. and three clear phases are visible after centrifugation: catalyst, glycerol and biodiesel, and their separation is easy. Finally, the total acidity of biodiesel mixture, measured by the European Standard method (EN14104), is equal to 0.23 mg KOH/g after 48 h of reaction, when the amount of free fatty acids is 2.42 wt%. This is a good value of biodiesel final acidity, considering that the maximum value accepted by the European Quality Standard Specification of Biodiesel (EN14214-UNI10946) is 0.5 mgKOH/g (0.8 mgKOH/g for the American ones ASTM6751). Moreover, also the maximum amount of diglycerides and triglyc- erides are respected, while only the amount of monoglycerides is higher (see legend in Table 1). 3.2. Catalysts characterization Main physico-chemical properties of tested catalysts are sum- marized in Table 2. For the microporous materials (USY, FAU-X and BEA) the low triglycerides conversion may be ascribed to the fact that the higher density of active acid sites is inside to the catalyst structure. Then, the big molecule of triglycerides can- not enter inside the microporous channels of the catalysts and the very low concentration of acid sites on the external surface of these catalysts is not enough to obtain an appreciable triglyc- Fig. 6. TG curves of K-ITQ-6 and K-MCM-41 catalysts. erides conversion. In fact, the pore dimension of acid microporous materials used (FAU-X, USY and BEA) does not exceed 6Å (see Table 2) while the spherical molecular diameter of triolein (triglyc- eride of oleic acid C18), calculated by empirical correlation [23],is 14 Å. This aspect is also confirmed by a previous study [9] where the activity of a strong acid zeolites (BEA) has been increased (of the 10%) by lantanium ionic exchange, due to the fact that the La species are able to increase the presence of “external” basic sites. In any case, the more acid La-BEA catalyst are not able to reach a triglycerides conversion upper to 50%, meaning that the strong acid zeolites are not suitable as heterogeneous catalyst for biodiesel production. Moreover, the high hydrophilicity of the sur- face of these acid catalysts can be another reason of their inactivity towards the transesterification reaction. The Si/Al molar ratio of these catalysts, in fact, ranging between Si/Al= 5–50 in the start- ing synthesis gel. Then, the surface of final catalyst can be very hydrophilic, due to presence of aluminum and the water can covers the surface of the acid solid, preventing the adsorption of organic substrate. Further investigations have been carried out only on the samples that have showed the best or an appreciable catalytic performance, that is the mesoporous materials and delaminated zeolites. Therefore, textural properties of these catalysts have been ana- lyzed by ICP-MS, N 2 adsorption/desorption and thermogravimetric thermal analyses. After ionic-exchange with K + the amount of potassium loaded is 0.22 wt% and 1.59wt% for MCM-41 and ITQ-6, respectively (Table 2). The N 2 adsorption/desorption isotherm, BET surface area and pore diameter of the mesoporous sample after potassium loading by ionic exchange, has been reported in Fig. 5(a). Both BET value and pore size of mesoporous structures decrease after K + loading, indicating thepresence ofthe cation.For ITQ-6delaminated zeolite, the ionic exchange, carried out to load potassium cation on the highly hydroxilated surface of this material, strongly reduces the final BET value (Fig. 5(b)), indicating a great amount of potassium loaded (as confirmed by ICP elemental analysis). The hydrophobicity of mesoporous samples and delaminated zeolites were estimated from the weight loss of physisorbed water at 200 ◦ C. The K-MCM-41 loses the 6.80% and the K-ITQ-6 loses only the 3.21% (Fig. 6). These results indicate that the delaminated zeolite containing potassium is more hydrophobic than K-MCM-41 (thermogravimetric analysis results). This can be one explanation why K-MCM-41 shows a lower catalytic activity towards the trans- esterification reaction, respect to the K-ITQ-6 catalyst. A. Macario et al. / Applied Catalysis A: General 378 (2010) 160–168 165 Fig. 7. IR spectra related to the adsorption of CO at 77 K on K-ITQ-6 outgassed at 573 K. Section (a) OH stretching region; section (b) CO stretching region. 3.3. Infrared spectroscopy analyses results Due to the good catalytic performance showed by K-ITQ-6, we have characterized the catalyst by FT-IR analysis, in order to have additional information on the nature of the acid and basic catalytic sites. 3.4. Adsorption of CO at 77 K Fig. 7 reports IR spectra related to increasing coverage of CO adsorbed at 77 K ontoK-ITQ-6, previously outgassed at 573K. In the OH stretching region (section (a)), the decrease of the isolated SiOH stretching mode at 3750 cm −1 is observed, accompanied by the increase of the broad band at 3660 cm −1 , due to the same species interacting via H-bonding with CO. The shift, which is an indirect measure of the Brønsted acidity, is about −90 cm −1 , i.e. the value measured for isolated silanols in all silica ITQ-2 and amorphous sil- ica Aerosil [24]. The stretching mode of CO molecules interacting with silanols is observed at 2156 cm −1 (section (b)), as expected [24]. Another band increases at 2138 cm −1 , due to “liquid-like” CO, probably mainly physisorbed in ITQ-6 microporosity. Fig. 8 reports the IR spectra related to increasing coverage of CO adsorbed at 77 K onto K-MCM-41. The inset in section (a), which refers to the OH stretching region, reports the spectrum of the sample outgassed at 573 K before adsorption of CO. Besides the band at 3748 cm −1 , due to isolated silanols, a broader band isobserved at 3698cm −1 , which is ascribed to silanols perturbed by hydrocarbon moieties, due to templates residues. Upon adsorption of CO, the band at 3698cm −1 is not perturbed (section (a), body of the figure). The band due to isolated silanols, instead, decreases and that due to H-bonded silanols increases at 3660 cm −1 . The measured shift is some −90 cm −1 , revealing the same Brønsted acidity of silanols as in the case of K-ITQ-6. The stretching mode of CO molecules interacting with silanols is observed at 2156cm −1 (section (b)), as in the case of K-ITQ-6. At low coverage, a shoulder is hardly discernible at 2162cm −1 , due to CO interacting with K + ions [25]. The band due to physisorbed “liquid-like” CO is observed at 2138cm −1 and its relative intensity with respect to the band at 2156 cm −1 is weaker than in the case of K-ITQ-6. This is ascribed to the absence of microporosity in the MCM-41 sample, at variance with K-ITQ-6. 3.5. Adsorption of CO 2 at room temperature Fig. 9 reports the IR spectra related to the adsorption of CO 2 at room temperature on K-ITQ-6 (section (a)) and K-MCM-41 (sec- tion (b)), previously outgassed at 573K. In the case of K-ITQ-6, a composite band between 1600 cm −1 and 1700cm −1 is observed, accompanied by a second one at 1349 cm −1 . These bands are not depleted by outgassing at RT and they are ascribed to carbonate species, formed by interaction of CO 2 with basic oxygen sites. They are not formed upon adsorption of CO 2 on K-MCM-41 (section (b)). In conclusion, only on the surface of ITQ-6 exchanged with potassium, both acids and basics sites are present, which are responsible of its higher catalytic activity, with respect to the K- MCM-41. Table 2 Main physico-chemical properties of catalysts tested. Catalyst type Si/Al gel molar ratio BET surface area [m 2 /g] Pore diameter [Å] K + loaded [wt%] USY a 15 720 6 – Al–BEA 50 545 5.6 – H-FAU-X 5 764 5.9 – Si-MCM-41 ∞ 969 47 – K-MCM-41 ∞ 677 42 0.22 ITQ-6 ∞ 900 – – K-ITQ-6 ∞ 288 – 1.59 a Si/Al content measured in the crystals. 166 A. Macario et al. / Applied Catalysis A: General 378 (2010) 160–168 Fig. 8. IR spectra related to the adsorption of CO at 77 K on K-MCM-41 outgassed at 573 K. Section (a) OH stretching region; section (b) CO stretching region. Fig. 9. Difference IR spectra related to the adsorption of CO 2 at RT on K-ITQ-6 (sec- tion (a)) and K-MCM-41 (section (b)) outgassed at 573 K. Curve 1: in contact with CO 2 (p = 47 mbar); curve 2: outgassed at RT after contact with CO 2 . 3.6. Proposed reaction mechanism In the Fig. 10 the reaction mechanism proposed for the reaction catalyzed by acid–basic K-ITQ-6 catalyst is reported. As it is possible to notice, the presence of potassium loaded by ionic exchange to the silanol groups of the pure silica delam- inated zeolite allows the formation of homogeneous methoxide (Fig. 10(a)) that promotes the carbonyl carbon atom attack of triglycerides, and the homogeneous Brønsted basic catalysis takes place [26]. Simultaneously, the free weak Brønsted acid silanol groups of the catalyst can protonate the carbonyl group of TG and FFAs, increasing their electrophilicity and rendering they more susceptible to the alcohol nucleophilic attack, allowing the transesterification and esterification reaction by a well known homogeneous Brønsted acid catalysis mechanism [27]. The acid esterification of free fatty acid byweak acid Brønsted silanol groups of catalyst prevents the saponification reaction and the biodiesel separation and purification can be carried out more easily. For this reason the FFAs conversion is higher than the value obtained by commercial KOH (as showed in the Fig. 4). Moreover, the simul- taneous presence of basic and acidic sites on the K-ITQ-6 catalyst surface could also explain why for an interesting heterogeneous basic catalyst (the ETS-10), reported by a Suppes et al. [16], the high free fatty acids concentration (about the 30 wt%) inhibit the basic solid catalyst. 3.7. Reusability of catalyst The reusability was studied only for K-ITQ-6 and K 2 SiO 3 , due to the fact that these catalysts having best performance. For K- ITQ-6 catalysts the amount of recovered material was about 80% of the initial quantity (calculated weighting the dried sample before and after reaction). While, concerning commercial K 2 SiO 3 , even if during the first reaction cycle the catalytic performance are better that that of K-ITQ-6, the solid is partially dissolved in the reaction media and only less than the 40 wt% can be recovered. Moreover, the biodiesel yield after the second cycle strongly decreases to 63%. Triglycerides conversion achieved during the second reaction cycle by K-ITQ-6 was only the 56%, whereas the biodiesel yield was A. Macario et al. / Applied Catalysis A: General 378 (2010) 160–168 167 Fig. 10. (a) Proposed potassium methoxide mechanism formation. (b) Proposed transesterification Brønsted acid mechanism. 35%, after 24h of reactionat180 ◦ C (5 wt%of catalyst andmolarratio oil:methanol equal to 1:20). Comparing these results with those obtained after the first reaction cycle, a clear potassium leaching has been taken. Moreover, the results of ICP analysis carried out on the used catalyst show that, after the first use, the amount of potassium on K-ITQ-6 zeolite decreases from 1.59wt% to 0.51 wt%, with a leaching of potassium of the 78%. This means that the maximum amount of potassium in the final biodiesel does not reach the 0.1 wt% (result obtained by mass bal- ance on potassium presents on K-ITQ-6 catalyst before and after reaction). However this metal content can be easily reduced by a bland washing of biodiesel with distilled water. 3.8. Regeneration of catalyst The thermal analysis carried out on the regenerated catalyst, after one reaction cycle, shows that the regeneration procedure allows to obtain a free catalyst surface: no mass loss is detected (Fig. 11) a part of physisorbed water. By 29 Si NMR analysis car- ried out on regenerated catalyst the amount of terminal silanol groups measured are quite the same of the amount detected on the fresh ITQ-6: 27 ± 3% and 28 ± 5%, respectively. This result sug- gested to use the regenerated catalyst for a new ionic exchange cycle in order to prepare new catalyst (K-ITQ-6) for biodiesel pro- duction. After this ionic exchange, the same amount of potassium was loaded (1.62 ± 4 wt%) and the same catalytic performance of new K-ITQ-6 catalyst has been obtained: 92% of biodiesel yield and complete triglycerides conversion after 48 h at 180 ◦ C. A possible flow-sheet of continuous process for biodiesel production using heterogeneous acid–basic K-ITQ-6 catalyst is reported in the Fig. 12. Two reactors are need: the first one for transesterification–esterification process and the second one for catalyst regeneration. Each fixed-bed reactor is supplied by catalyst regeneration lines, consisting of hexane feed for organic com- pound removing, water feed for hexane removing and, finally, KCl 1 M solution feed for potassium exchange and fresh catalyst preparation. When the catalyst is exhaust, the reactor stops to work and the catalyst regeneration process starts, simultaneously, the second reactor, terminated the catalyst regeneration, starts to work for a new biodiesel production process with the fresh cata- lyst. Fig. 11. TG and DTA curves of regenerated ITQ-6 catalysts. 168 A. Macario et al. / Applied Catalysis A: General 378 (2010) 160–168 Fig. 12. Flow-sheet of proposed continuous process for biodiesel production. 4. Conclusion Microporous acid catalysts are not suitable for transesterifica- tion reaction of triglycerides because of the diffusion limitation of reactants inside micropores. Modification of ITQ-6 surface with K + ions, carried out by ionic-exchange, produced a catalyst with weak Brønsted acid sites (silanols) and basic sites, responsible of catalytic activ- ity in transesterification of triglycerides and esterification of FFAs. The catalyst is active towards the conversion of waste oil with high free acidity (5.58%) without formation of unde- sired by-product (soap-compounds) and promotes a simultaneous homogeneous/heterogeneous and acid/base catalysis. This may open the way to use of low-quality oil to perform a cheaper biodiesel production by acid–basic solid catalysts. The heteroge- neous catalyst can be easy and efficiently regenerated and reused for new reaction cycle. This opens the possibility to perform a con- tinuous catalytic process for biodiesel production. References [1] F. Ma, M.A. Hanna, Bioresour. Technol. 70 (1999) 1–15. [2] H. Fukuda, A. Kondo, H. Noda, J. Biosci. Bioeng. 92 (2001) 405–416. [3] L.C. Meher, D. Vidya Sagar, S.N. Naik, Renew. Sust. Energy Rev. 10 (2006) 248–268. [4] A. Macario, M. Moliner, A. Corma, G. Giordano, Micropor. Mesopor. Mater. 118 (2009) 334–340. [5] A. Macario, G. Giordano, L. Setti, A. Parise, J.M. Campelo, J.M. Marinas, D. Luna, Biocatal. Biotrans. 25 (2007) 328–335. [6] E. Lotero, Y. Liu, D.E. Lopez, K. Suwannakarn, D.A. Bruce, J.G. Goodwin, Ind. Eng. Chem. Res. 44 (2005) 5353–5363. [7] A.A. Kiss, A.C. Dimian, G. Rothenberg, Adv. Synth. Catal. 348 (2006) 75–81. [8] G.D. Yadav, J.J. Nair, Micropor. Mesopor. Mater. 33 (1999) 1–48. [9] Q. Shu, B. Yang, H. Yuan, S. Qing, G. Zhu, Catal. Commun. 8 (2007) 2159– 2165. [10] Y.M. Park, D W. Lee, D K. Kim, J S. Lee, K Y. Lee, Catal. Today 131 (2008) 238–243. [11] A. Brito, M.E. Borges, R. Arvelo, F. Garcia, M.C. Diaz, N. Otero, Int. J. Chem. React. Eng. 5 (2007) A104. [12] W. Xie, H. Li, J. Mol. Catal. A: Chem. 255 (2006) 1–9. [13] W. Xie, X. Huang, H. Li, Bioresour. Technol. 98 (2007) 936–939. [14] H J. Kim, B S. Kang, M J. Kim, Y.M. Park, D K. Kim, J S. Lee, K Y. Lee, Catal. Today 93–95 (2004) 315–320. [15] M. Di Serio, R. Tesser, L. Pengmei, E. Santacesaria, Energy Fuels 22 (2008) 207–217. [16] G.J. Suppes,M.A. Dasari, E.J. Doskocil, P.J. Mankidy, M.J. Goff, Appl. Catal. A: Gen. 257 (2004) 213–223. [17] X. Liu, H. He, Y. Wang, S. Zhu, X. Piao, Fuel 87 (2008) 216–221. [18] E. Leclercq, A. Finiels, C. Moreau, JACOS 78 (2001) 1161–1165. [19] K. Noiroj, P. Intarapong, A. Luengnaruemitchai, S. Jai-In, Renew. Energy 34 (2009) 1145–1150. [20] V.S Y. Lin, J.A. Nieweg, C.Kern,B.G.Trewyn,J.W.Wiench, M. Pruski, Prep. Symp. Am. Chem. Soc. Div. Fuel Chem. 51 (2006) 426–427. [21] V.S Y. Lin, J.A. Nieweg, J.G. Verkade, C.R. Venkat Reddy, C. Kern, Patent WO/2008/013551. [22] A. Corma, U. Diaz, M.E. Domine, V. Fornés, J. Am. Chem. Soc. 122 (2000) 2804–2809. [23] Perry’s Chemical Engineers’ Handbook, McGraw-Hill International Editions, sez.17-20, Table 17-10. [24] B. Onida, L. Borello, B. Bonelli, F. Geobaldo, E. Garrone, J. Catal. 214 (2003) 191–199. [25] S. Bordiga,G. Turnes Palomino, C. Paze’, A. Zecchina, Micropor. Mesopor. Mater. 34 (2000) 67–80. [26] E. Lotero, J.G. Goodwin Jr., D.A. Bruce, K. Suwannakarn, Y. Liu, D.E. Lopez, Catal- ysis, vol. 19, Royal Society of Chemistry, 2006, pp. 41–83. [27] Y. Liu, E. Lotero, J.G. Goodwin Jr., J. Catal. 243 (2006) 221–228. . heterogeneous/homogeneous systems using acid and base catalysts. The acid catalysts (strong acid catalysts: USY, BEA, FAU-X, and weak acid catalysts: MCM-41 and ITQ-6 with Si/Al= ∞) were prepared by hydrothermal. www.elsevier.com/locate/apcata Biodiesel production process by homogeneous/heterogeneous catalytic system using an acid base catalyst Anastasia Macario a,∗ , Girolamo Giordano a , Barbara Onida b , Donato Cocina b , Antonio. simultaneous homogeneous/heterogeneous and acid/ base catalysis. This may open the way to use of low-quality oil to perform a cheaper biodiesel production by acid basic solid catalysts. The heteroge- neous catalyst can be easy and