Applied Catalysis A: General 353 (2009) 203–212

Contents lists available at ScienceDirect

Applied Catalysis A: General

journa l homepage: www.e lsev ier .com/ locate /apcata

Simultaneous transesterification and esterification of unrefined or wasteoils over ZnO-La2O3 catalysts

Shuli Yan, Steven O. Salley, K.Y. Simon Ng *

Department of Chemical Engineering and Materials Science, Wayne State University, Detroit, MI 48202, USA

A R T I C L E I N F O

Article history:

Received 10 June 2008

Received in revised form 17 October 2008

Accepted 27 October 2008

Available online 11 November 2008

Keywords:

Biodiesel

Unrefined and waste oils

Transesterification

Esterification

Hydrolysis

A B S T R A C T

A single-step method was developed for biodiesel production from unrefined or waste oils using a series

of heterogeneous zinc and lanthanum mixed oxides. Effects of metal oxide molar ratio, free fatty acids

(FFA) and water content in feedstock, molar ratio of methanol and oil, and reaction temperature on the

yield of biodiesel were investigated. A strong interaction between Zn and La species was observed with

enhanced catalyst activities. Lanthanum promoted zinc oxide distribution, and increased the surface acid

and base sites. The catalyst with 3:1 ratio of zinc to lanthanum was found to simultaneously catalyze the

oil transesterification and fatty acid esterification reactions, while minimizing oil and biodiesel

hydrolysis. A reaction temperature window of 170–220 8C was found for the biodiesel formation. A high

yield (96%) of fatty acid methyl esters (FAME) was obtained within 3 h even using unrefined or waste oils.

Published by Elsevier B.V.

1. Introduction

Biodiesel, a renewable fuel with similar combustion propertiesto fossil diesel, is normally produced by transesterification ofhighly refined oils with short-chain alcohols. Biodiesel cansignificantly decrease the exhaust emission of CO2, SOx andunburned hydrocarbons from motor vehicles [1,2]. Biodiesel isenvironmentally beneficial, and therefore, is a promising alter-native to fossil diesel [3].

Transesterification reaction of triglycerides for the productionof biodiesel is as follows:

* Corresponding author. Tel.: +1 313 577 3805; fax: +1 313 578 5814.

E-mail address: sng@eng.wayne.edu (K.Y. Simon Ng).

0926-860X/$ – see front matter . Published by Elsevier B.V.

doi:10.1016/j.apcata.2008.10.053

A conventional operation usually uses strong basic or acidicsolutions (i.e., NaOH, KOH and H2SO4) as catalyst and food-gradevegetable oils as raw material. These homogeneous catalysts arequite sensitive to free fatty acids (FFA) and water in the oilfeedstocks and alcohols. FFA reacts with the basic catalyst(NaOH, KOH) and forms soaps. This soap formation complicatesthe glycerol separation, and drastically reduces the methyl esteryield. Water in the feedstock leads to hydrolysis of oils and fattyacid methyl esters (FAME) in the presence of strong basic oracidic catalysts. Thus, some inexpensive oils, such as crudevegetable oils, waste cooking oil, and rendered animal fats,

which generally contain a high content of FFA and water, cannotutilize homogeneous catalysts directly. Furthermore, the watercontent in alcohols is also an important issue in traditionalprocesses. Since alcohols are hygroscopic, the recovered alcohols

Table 1Fatty acid methyl esters, FFA and water content of food-grade soybean oil, crude

soybean oil, crude palm oil and waste cooking oil.

Fatty acid

components

Food-grade

soybean

oil (%)

Crude

soybean

oil (%)

Crude

palm

oil (%)

Waste

cooking

oil (%)

C 14:0 0 0.27 0.21 0

C 16:0 11.07 13.05 41.92 11.58

C 16:1 0.09 0.39 0.23 0.18

C 18:0 3.62 4.17 3.85 4.26

C 18:1 20.26 22.75 42.44 24.84

C 18:2 57.60 52.78 11.30 53.55

C 18:3 7.36 6.59 0.04 5.60

FFA content 0.02 3.31 0.24 3.78

Water content 0.02 0.27 0.04 0.06

S. Yan et al. / Applied Catalysis A: General 353 (2009) 203–212204

must be dried thoroughly to remove the water from the azeo-tropes. For conventional processes using homogenous catalysts,the FFA content in the feedstock must be lower than 0.50(wt%) [4] and water content lower than 0.06 (wt%) [5]. Usually,highly refined oils are used in conventional methods for bio-diesel production. According to the calculation of Haas et al. [6],the cost of oil feedstock accounts for up to a total of 88% ofbiodiesel production cost in traditional processes. With recentincreases in refined oil prices, the cost of oil feedstockaccounts for an even higher fraction of the total productioncost. Thus, it is important to develop new catalytic processeswhich can handle unrefined and waste oils directly to lower thecost of biodiesel.

Recently, an acid- and alkali-catalyzed two-step method forbiodiesel production using some unrefined or waste oils as rawmaterials has been reported [7,8]. In this method, an acidic catalyst(H2SO4, HCl) is initially used to convert FFA to the esters, and thenin the second stage, transesterification of oil is performed using analkaline catalyst (NaOH, KOH). Although this method can utilizeunrefined or waste oils for biodiesel production, the processrequires multiple reactions, washing, and separation stages. Thestrong acidic or basic catalysts used are highly corrosive, and mustbe removed from the biodiesel product by multiple washing. Thus,a significant amount of waste water is generated, together with aloss of catalyst [9,10].

Therefore, it is advantageous to develop a new class ofheterogeneous catalysts, which has a higher tolerance to waterand FFA in oils, and can simultaneously catalyze both theesterification and transesterification reactions. There are reportsof heterogeneous catalytic esterification of fatty acids [11,12]and transesterification of highly refined oils [13–16]. Moreover,Omota et al. [12] reported a process of fatty acid esterificationusing zirconium salts as catalysts. Suppes et al. [13] reportedthat some salts and oxides of magnesium, calcium, and zinc areactive in transesterification. Serio et al. [14] used Mg–Al calcinedhydrotalcites in the tranesterification of refined soybean oil,which contained 0.1 (wt%) FFA, at high temperature. They foundthat this type of heterogeneous catalyst is more tolerant towater and the catalytic performance was not affected by thepresence of 1 (wt%) of water in oils. Recently, Yan et al. [17]reported that supported CaO catalysts directly catalyzed thereaction of some crude oils into biodiesel when water content inthe oils was less than 2 (wt%) and FFA content less than 3.5(wt%). However, further increases of FFA and water content inoils could inhibit the transesterification. The mechanism for theimproved tolerance to water and FFA of this type of catalystswas not fully elucidated. Bournay and Hillion [18] stated in apatent that oils with high content of FFA can be directly usedwith a Lewis acid catalyst (zinc aluminate) for biodieselproduction. However, that catalyst was quite sensitive to waterand the limitation of water content in oils is 0.15 (wt%).Sreeprasanth et al. [19] prepared other types of Lewis acidcatalysts containing zinc and iron and used them in transester-ifying unrefined and used oils. He found that the catalyst isactive in both esterification and transesterification. However,the catalyst is also active in the hydrolysis of FAME to FFA, thusthe overall yield of biodiesel is decreased. In this paper, weprepared and characterized a new class of ZnO-La2O3 hetero-geneous catalysts with different Zn:La ratios. The effects ofmetal oxide ratio on the Lewis acid and base sites, and on thetransesterification of triglyceride, esterification of fatty acid,and hydrolysis of triglyceride and fatty acid methyl esters,were investigated. These mixed oxide catalyst shows verypromising results for processing unrefined and waste oils dire-ctly into biodiesel.

2. Experiments

2.1. Materials

Food-grade soybean oil was purchased from Costco warehouse(Detroit, MI), crude soybean oil from BDI (Denton, TX), crude palmoil from Malaysia Palm Oil Board (Selangor, Malaysia) and wastecooking oil was obtained from a local restaurant. The fatty acidcompositions of these four kinds of oil were determined by GC–MS(Table 1). Oleic acid and methyl alcohol were obtained from theMallinckrodt Chemicals (Phillipsburg, NJ), with water contents of0.02% and 0.03%, respectively. Zinc nitrate hexahydrate (98%),lanthanum nitrate hydrate (98%), and urea (99%) are of analysisgrade, and were purchased from Sigma–Aldrich Company (St.Louis, MO).

2.2. Catalyst preparation and characterization

A homogeneous-coprecipitation method [20] was used toprepare catalyst samples. First, 2 M Zn(NO3)2 and 1 M La(NO3)3

were prepared with distilled water. Then, solutions with varyingratios of Zn–La (1:0, 1:1, 3:1, 9:1, 0:1) were mixed with a 2 M ureasolution. The mixture was boiled for 4 h, and then dried at 150 8Cfor 8 h, followed by step-rising calcination at 250 8C, 300 8C, 350 8C,400 8C, finally at 450 8C for 8 h. The catalysts are noted as Zn10La0,Zn1La1, Zn3La1, Zn9La1, Zn0La10 according to their catalystcompositions.

Powder X-ray diffraction (XRD) patterns were taken with aRigaku RU2000 rotating anode powder diffractometer (TheWoodlands, TX) equipped with CuKa radiation (40 kV, 200 mA).Scanning electron microscopy (SEM) and energy dispersivespectrometer (EDS) were taken with a Hitachi S-2400 ScanningElectron Microscope (San Jose, CA). Maximum operating voltageused was 25 kV. N2 adsorption and desorption isotherms weremeasured at 77 K with a Quantachrome AS-1MP volumetricadsorption analyzer. Before adsorption measurements, all thesamples were outgased for 12 h at 300 8C.

XPS analysis was performed with a PHI 5500 system(PerkinElmer, Wellesley, MA), using a monochromatic Al Ka X-ray radiation source (1486.6 eV) and AugerScan system control(RBD Enterprises, Bend, OR). Elemental concentration on thesample surface was measured by XPS multiplex scan (spot size:�1 mm diameter).

2.3. Biodiesel reactions and product analysis

Catalytic transesterification, esterification and hydrolysis reac-tions were conducted in a 500 mL stainless steel stirred reactor(Parr 4575 HT/HP Reactor). For transesterification, 126 g of oil,180 g of methanol, and 3 g of catalyst were used. During the

Fig. 2. Transesterification activities of Zn10La0, Zn9La1, Zn3La1, Zn1La1 and

Zn0La10 at 200 8C. Reaction mixtures are 126 g of oil, 180 g of methanol, 3 g of

catalyst. Zn3La1 shows the highest activity compared to other catalysts.

S. Yan et al. / Applied Catalysis A: General 353 (2009) 203–212 205

reaction, 15 mL of reaction mixture was collected via a two-stageliquid sampler at different time intervals. When the reaction wascompleted, the catalyst was filtered out. The liquid productobtained was vaporized to remove excessive methanol, and thensettled in a separating funnel. The upper layer in the separatingfunnel (mainly containing fatty acid methyl esters) was char-acterized by a GC–MS spectrometer (Clarus 500 MS System,PerkinElmer, Shelton, CT) equipped with a capillary column (Rtx-WAX Cat. no.12426) (Bellefonte, PA). Methyl arachidate (Nu-ChekPrep Inc, Elysian, MN) was used as an internal standard.

Esterifications were performed with 126 g of oleic acid, 180 g ofmethanol, and 3 g of Zn3La1 catalyst. Yield of oleic acid methylester were determined by GC–MS. And oleic acid concentrationwas determined using a Brinkman/Metrohm 809 titrando (West-bury, NY) according to ASTM D 664.

Hydrolysis of oil and hydrolysis of biodiesel were conductedwith 285 g of food-grade soybean oil or soybean biodiesel (WackerOil, Manchester, MI), 15 g of water, and 2.3 (wt%) of Zn3La1catalyst. Yield of FFA was determined by Brinkman/Metrohm 809titrando (Westbury, NY) according to ASTM D 664. Water contentwas analyzed using a Brinkman/Metrohm 831 KF Coulometer(Westbury, NY) according to ASTM D 6304-00.

3. Results and discussion

3.1. Catalytic activity of zinc lanthanum mixed oxides in oil

transesterification

Triglycerides are the major component of vegetable oils. In thisstudy, a food-grade soybean oil, which contained 99.5 (wt%) oftriglycerides, was used as the model for most of the studies.

A step-rising heating method (heating rate of 2 8C/min, and heldat the target temperature for 1 min) was used to compare thecatalytic activities of mono-metal oxide (Zn10La0, Zn0La10) andmixed metal oxides (Zn3La1) for the oil transesterification reaction(Fig. 1). All three catalysts show activities in catalyzing thetransesterification reaction when the reaction temperature ishigher than 170 8C; while without any catalyst, the transester-ification reaction starts at 220 8C. The mixed oxides catalyst clearlyshows an enhanced activity as compared to the pure oxides.

Fig. 1. Transesterification activities of Zn10La0, Zn3La1 and Zn0La10 as a function of

temperature. Reaction mixtures were heated at 2 8C/min and maintained at target

temperatures for 1 min. Zn3La1 shows the highest activity compared to Zn10La0

and Zn0La10.

To investigate the effect of catalyst composition on transester-ification, mixed oxides with different molar ratio of zinc tolanthanum were tested at 200 8C (Fig. 2). The times needed to reachchemical equilibrium were about 60 min, 80 min, 120 min,150 min and 200 min for Zn3La1, Zn0La10, Zn1La1, Zn9La1 andZn10La0, respectively. The catalyst with 3:1 molar ratio of zinc tolanthanum showed the highest activity in oil transesterification.

The effect of reaction temperature on Zn3La1– catalyzedtranseseterification process was shown in Fig. 3. Since themethanol concentration was kept in excess, a power rate lawmodel can be written as:

�g ¼ k1Caoil

where r is the reaction rate (for <50.00% FAME yield), k1 is theapparent reaction rate constant, Coil is the oil concentration, and ais the reaction order. Based on the data in Fig. 3, the apparentreaction order to the oil was found to be 1.2. The apparentactivation energy Eapp was 90.9 KJ mol�1, which was significantlyhigher than the reported activation energy (19.9 KJ mol�1) using

Fig. 3. Transesterification with different molar ratio of methanol to oil with Zn3La1

at 200 8C. Reactants are 3 g of catalyst and 300 g of oil and methanol mixture. 36–42

of molar ratio of methanol to oil shows high yield of FAME.

Fig. 4. Transesterification at different reaction temperatures with Zn3La1. Reactant

mixtures are 126 g of oil, 180 g of methanol, 3 g of catalyst.

S. Yan et al. / Applied Catalysis A: General 353 (2009) 203–212206

NaOH catalyst [21]. The higher activation energy is expected of aheterogeneous catalytic system.

The effect of molar ratio of methanol to oil was shown in Fig. 4.In traditional homogeneous processes, a 6:1 molar ratio ofmethanol to oil is usually used for biodiesel production. However,in Fig. 4 it gave poor yields of FAME in Zn3La1 catalyzed process,and higher molar ratios of methanol to oil (36:1 to 42: 1) led toyields higher than 95% within 60 min. The optimal molar ratio issuggested to be 36:1. The excess methanol promoted thetransesterification reaction forward, and extracted products(FAME and glycerin) from reactants to renew the catalyst surface[17].

3.2. Catalytic activity of zinc lanthanum mixed oxides in fatty acid

esterification

In unrefined or waste oils, FFA concentration can be very high(0.5–30%) [22]. Therefore, developing a catalyst which cansimultaneously esterify FFA and transesterify triglyceride intobiodiesel is important. As fatty acids have similar chemicalproperties, oleic acid was used as a model FFA for the esterificationstudy. The esterification reaction with methanol is as follows:

Fig. 5a presents the results of using Zn3La1 in esterifying pureoleic acid with methanol using a step-rising heating method(heating rate of 2 8C/min, and held at the target temperature for1 min). When the temperature was higher than 140 8C, esterifica-tion activity was observed. At 200 8C, a 96.7% yield of oleic acidmethyl ester can be obtained within 10 min (Fig. 5b).

Since FFA can exist in considerable amount in unrefined orwaste oil feedstock, investigation of esterifying FFA with methanolin the presence of triglycerides is important. A mixture oilcontaining 5.4 (wt%) of oleic acid and 94.6 (wt%) of food-gradesoybean oil was prepared and reacted with excess methanol at200 8C. The results are shown in Fig. 6. Oleic acid content in mixedoil quickly decreased in the first 10 min. The yield of FAMEapproached 95.3% after 60 min. Fig. 6 suggests that, on Zn3La1, theesterification reaction is rapid and occurs simultaneously with thetransesterification of triglycerides in a single-step process.

Fig. 5. Esterification activity of Zn3La1 in pure oleic acid (a) yield of oleic methyl est

maintained at target temperatures for 1 min. Zn3La1 shows the esterification activity

3.3. Catalytic activity of zinc lanthanum mixed oxides in biodiesel and

oil hydrolysis

Unrefined or waste oils generally contain a high content ofwater. Thus, fatty acid methyl esters and triglycerides hydrolysisare undesirable side reactions that may decrease the yield of FAMEand increase the acidity of reaction mixtures. Fig. 7 illustrates that,without zinc and lanthanum mixed oxides catalyst hydrolysisreactions were not observed even for reaction temperature up to250 8C. In the presence of Zn3La1, hydrolysis occurs at tempera-tures above 220 8C. As hydrolysis is not desirable, a reactiontemperature lower than 220 8C is necessary. A final yield of 95.0%FAME can be obtained with a mixture of 5.3 (wt%) water and 94.7(wt%) food-grade soybean oil at 200 8C (Fig. 8). The water contentsin the reactant mixture were found to maintain at about 5.2%during the process. The FFA contents were very low (<0.01%),

ers as a function of temperature. Reaction mixtures were heated at 2 8C/min and

over 140 8C; (b) yield of oleic methyl esters at 200 8C.

Fig. 7. Hydrolysis of biodiesel and oil in presence or absence of Zn3La1 catalyst.

Reaction mixtures are 285 g of oil or biodiesel and 15 g of distilled water.

Fig. 6. The FAME yield and oleic acid content as a function of time at 200 8C. Reaction

mixtures are 126 g of the oil containing 5.4 (wt%) oleic acid and 94.6 (wt%)

triglycerides, 180 g of methanol, 3 g of Zn3La1 catalyst.

S. Yan et al. / Applied Catalysis A: General 353 (2009) 203–212 207

which are beyond the detection limit of the titrator. Fig. 8 suggeststhat at 200 8C and in the presence of Zn3La1, there was little or nohydrolysis of biodiesel and triglyceride, while a high yield of FAMEwas obtained. The triglyceride hydrolysis (1) and FAME hydrolysis(2) reactions are:

3.4. Effects of FFA and water on transesterification

FFA and water are usually considered as poisons to bothhomogeneous acidic and basic catalysts in traditional biodieselproduction processes [4,5]. In order to examine the effects of FFAand water, the activity of Zn3La1 catalyst on food-grade soybeanoils with 5.20%, 10.13%, 15.21% and 30.56% of oleic acid, with 1.03%,3.12% and 5.07% of water, and with 5.16% of oleic acid and 3.10% ofwater were examined. Fig. 9a also shows that the addition of oleicacid accelerated the reaction rate and shortened the time to a highyield of FAME. For example, without FFA addition, 60 min is neededto reach completion; while for oil with 5.20% FFA addition, the timeto reach completion is only about 20 min. With further increase inFFA addition, the time to reach completion decreased. In Fig. 9b,the effect of FFA addition on the equilibrium yields of FAME usingH2SO4, NaOH and Zn3La1 as catalyst is compared. A sharp decreasein FAME yield was observed in the processes that use H2SO4 orNaOH as catalyst [23–25]. Fig. 9b shows that even with 5.20% FFA,FAME yield decreased to 78% and 88% in NaOH catalyzed andH2SO4 catalyzed processes, respectively. However, FAME yield wasmaintained at 96.6% in Zn3La1 catalyzed process. More differencecan be observed for 30.56% FFA addition, FAME yield decreased to10% and 60% in NaOH catalyzed and H2SO4 catalyzed processes,respectively; while FAME yield was still maintained at 96.0% inZn3La1 catalyzed process. The result indicates that in comparisonwith NaOH and H2SO4, Zn3La1 has a remarkable tolerance to FFA inthe transesterification reaction. Thus, this class of mixed metaloxides is very promising for direct conversion of acidic oils to FAMEin a single-step process.

Effects of water on FAME yield using Zn3La1 as catalyst areshown in Fig. 9c and d. It was found that the addition of waterprolonged the time to reach completion. For example, with 3.12%water added to oil, the time to completion was prolonged from60 min to 90 min. Further increase of water led to the increase ofreaction time to equilibrium. However, the equilibrium yield ofFAME was around 97% regardless of the water content. On theother hand, for H2SO4 and NaOH catalyzed processes, watershowed considerable effect on transesterification activity (Fig. 9d)[23–25]. When water addition in oils was 5.20%, FAME yieldsdecreased to 78% and 11% for NaOH and H2SO4 catalyst,respectively.

3.5. Single-step conversion of unrefined and waste oils

Zn3La1 was demonstrated to be active in the transesterificationreaction above 170 8C, in FFA esterification above 140 8C, and inbiodiesel and oil hydrolysis over 220 8C. Thus, in order to obtain ahigh FAME yield, the reaction temperature in this system should belimited to the range of 170–220 8C to enhance simultaneous

(1)

(2)

Fig. 8. The FAME yield and water content as a function of time at 200 8C. Reaction

mixtures are 126 g of the oil containing 5.3 (wt%) oleic acid and 94.7 (wt%)

triglycerides, 180 g of methanol, 3 g of Zn3La1 catalyst.

S. Yan et al. / Applied Catalysis A: General 353 (2009) 203–212208

transesterification and esterification, while limiting hydrolysisreactions. It should be noted that in a fixed bed reactor con-figuration, the optimal operating temperature will be a functionof contact time with the catalyst. Therefore, it is necessary to

Fig. 9. Effect of FFA and water additions on transesterification (a) yield of FAME in the p

FAME; (c) yield of FAME in the presence of different water addition; (d) effect of water add

sulfuric acid amount is 3%, molar ratio of methanol to oil is 6:1, reaction temperature is 60

molar ratio of methanol to oil is 6:1, reaction temperature is 25 8C, and reaction time is 8

molar ratio of methanol to oil is 36:1, reaction temperature is 200 8C, and reaction tim

re-examine the temperature windows for the three reactions infixed bed reactors to optimize the FAME yield. Fig. 10 [26,27]illustrates the possible reaction pathway in converting unrefinedand waste oils into biodiesel in the presence of Zn3La1 at 200 8C.FAME is formed through triglyceride transesterification and FFAesterification reactions; while being consumed through FAMEhydrolysis reaction. Triglyceride hydrolysis can also occur whichlead to lower FAME yield and higher total acid number of theresulting products. At 200 8C, the reaction rates of FAME hydrolysisand triglyceride hydrolysis were not significant.

Several unrefined and waste oils without any pretreatmentwere converted directly using the Zn3La1 catalyst at 200 8C. Fig. 11illustrates the FAME yields from waste cooking oil, unrefinedsoybean oil, unrefined palm oil, food-grade soybean oil with 5.2%oleic acid and 3.1% water, and food-grade soybean oil. The FFA andwater contents of these oils are shown in Table 1. It is remarkablethat the equilibrium yield of the different oils were all very higharound 96%. For this Zn3La1 catalyst, the presence of FFA and waterdid not significantly affect the equilibrium yield.

3.6. Effect of metal oxide composition on catalyst structure

3.6.1. XRD, SEM, EDS and BET

The XRD patterns and EDS of zinc and lanthanum metal oxidesare given in Fig. 12 and Table 2. The XRD pattern of Zn10La0corresponds with hexagonal wurtzite structure of zinc oxide. The

resence of different FFA addition; (b) effect of FFA content on equilibrium yield of

ition on equilibrium yield of FAME. Reaction conditions: (1) acidic catalysis process,

8C, and reaction time is 96 h [19]; (2) alkaline catalysis process, NaOH amount is 1%,

h [20]; (3) heterogeneously catalytic process, catalyst amount of Zn3La1 is 2.3 (wt%),

e is 1.5 h.

Fig. 10. Reactions pathways of transesterification, esterification, and hydrolysis of unrefined and waste oils.

Fig. 11. FAME yield of crude palm oil, crude soybean oil, waste cooking oil, food-

grade soybean oil and food-grade soybean oil with 3.1% water and 5.2% oleic acid

addition. Reaction mixtures are 126 g of oil, 180 g of methanol, 3 g of catalyst. Note

that 96% yield can be obtained within 3 h.

Fig. 12. XRD patterns of pure and mixed zinc and lanthanum mixed metal oxides.

Note the transition of bulk ZnO structures to mixed ZnO-La2O3 structures as La

content increases.

S. Yan et al. / Applied Catalysis A: General 353 (2009) 203–212 209

pattern of Zn0La10 shows a mixture of La2CO5 and LaOOH. Thediffraction patterns observed for Zn9La1, Zn3La1 and Zn1La1 showlower intensity than Zn10La0, and mixed ZnO, La2CO5 and LaOOHare found in Zn9La1, Zn3La1 and Zn1La1.

The mean grain size of ZnO in Zn9La1, Zn3La1 and Zn1La1was calculated by the Deby-Scherrer equation based on thereflection peak of ZnO (1 0 1) in Fig. 12. The bulk molar ratios ofzinc to lanthanum were determined by EDS (Table 2). The meangrain size of ZnO was found to decrease with the addition of La(Table 2), suggesting that a strong interaction between theLa and Zn species enhances the ZnO dispersion. Figs. 1 and 2show that the mixed ZnO-La2O3 catalysts (Zn9La1, Zn3La1 and

Table 2Specific surface area, mean grain size and lattice constants of zinc lanthanum oxides as a f

content increase.

Catalyst XRD structure Mean grain

size of ZnO (nm)

Lattice constants fo

a (A) c (A)

Zn10La0 ZnO >100 3.25 5.21

Zn9La1 ZnO 27.6 3.25 5.36

Zn3La1 ZnO, La2CO5, LaOOH 17.1 3.25 5.23

Zn1La1 ZnO, La2CO5, LaOOH 9.8 3.33 5.10

Zn0La10 La2CO5, LaOOH – – –

Zn1La1) have higher activities than pure zinc oxide (Zn10La0),which correlate well with the effects of lanthanum on enhancingthe dispersion of ZnO.

The XRD pattern of Zn3La1 shows a mixture of ZnO, La2CO5 andLaOOH phases. Various polar crystal planes of ZnO, such as ZnO(1 0 2), (1 0 3) and (1 1 2), could be observed. Methanol moleculesprefer to stick on ZnO (1 0 2) and (1 0 3) which contain higherconcentration of oxygen atoms than zinc. While on ZnO (1 1 2),where concentration of zinc atoms is higher, adsorption ofcarbonyl groups are favored [28,29]. Therefore, these polarsurfaces can be attributed as the active centers for transesterifica-tion and esterification reactions.

unction of Zn:La ratio. Note the Zn:La ratio. Note the decrease of ZnO grain size as La

r ZnO phase Zn:La

(bulk molar ratio)

Specific

surface area (m2/g)Vol (A3) Density (C)

47.63 5.68 1.0: 0 16.3

48.62 5.56 8.9: 1.0 16.8

47.81 5.65 3.5: 1.0 15.7

49.12 5.50 1.2: 1.0 14.9

– – 0: 1.0 12.2

Fig. 13. A representative SEM image of Zn3La1 showing aggregations of catalyst

particles.

Table 3The binding energy and surface percentage of Olat, Zn2+, La3+ as a function of Zn:La ratio. N

catalyst.

Samples Binding energy (eV) Surface percentage (at.%

La3d Olat Oad Zn2p Olat Zn2+ La3+

Zn10La0 – 528.9 530.6 1021.9 19.9 10.2 –

Zn9La1 836.8 530.2 531.8 1021.6 23.2 24.4 2.1

Zn3La1 835.7 530.2 531.5 1021.4 30.5 26.8 2.8

Zn1La1 835.3 530.1 531.4 1021.1 32.9 4.2 2.9

Zn0La10 835 530.3 531.7 – 37.5 – 7.9

Fig. 14. Schematic representation of possible mechani

S. Yan et al. / Applied Catalysis A: General 353 (2009) 203–212210

The XRD patterns of Zn10La0, Zn9La1, Zn3La1 and Zn1La1 showa shift of the diffraction peaks of ZnO (1 0 0), (0 0 2), (1 0 1) basedon MDI’s JADE and DATASCAN. Lattice constants of ZnO crystalwere calculated (Table 2). The increase of lattice constants in a andc directions suggests partial incorporation of La 3+ ion with ZnOcrystal resulting in lattice distortion [30]. Thus, crystal growth ofwurtzite ZnO was inhibted and ZnO was highly dispersed [31].

Specific surface area of catalysts ranges from 12.2 m2/g to16.8 m2/g (Table 2), and do not show a direct correlation withlanthanum loading. The particle size and morphology of Zn3La1are shown in Fig. 13. Small particles and some big aggregations areobserved.

3.6.2. XPS

The XPS data of Zn, La and O elements on the surface of Zn10La0,Zn9La1, Zn3La1, Zn1La1 and Zn0La10 are shown in Table 3. Thebinding energy of 1021.9 eV, 835.0 eV, 530.6 eV and 528.9 eV canbe attributed to Zn 2+, La 3+, Oad (adsorbed oxygen) and Olat (latticeoxygen), respectively. For metal oxides, lattice oxygens on the

ote the highest total surface percentage of Olat, Zn2+ and La3+ observed for the Zn3La1

) Surface atom ratio

Zn2+ and La3+ Olat, Zn2+ and La3+ Zn2+:La3+ (Zn2+ + La3+):Olat

10.2 30.1 – 0.5

26.5 49.7 11.6 1.1

29.6 60.1 9.6 1.0

7.1 40.0 1.4 0.2

7.9 45.4 – 0.2

sm for esterification of fatty acid with methanol.

Fig. 15. Schematic representation of possible mechanism for transesterification of triglyceride with methanol.

S. Yan et al. / Applied Catalysis A: General 353 (2009) 203–212 211

surface are Lewis base sites and metal ions are Lewis acid sites [32].Base sites are considered catalytic sites for transesterificationreactions [33,34], and acid sites are considered as the active sitesfor esterification reactions [34,35]. Therefore, it is interesting to seeif there is a correlation between surface concentrations of Olat,Zn2+, La3+ and the activities of esterification and transesterificationreactions.

The effect of La3+ concentration on the fraction of basic and acidsites is shown in Table 3. Percentages of surface lattice oxygen(base site) increased with the lanthanum content. On the otherhand, surface content of Zn2+ and La3+ in Zn3La1 are higher thanZn10La0, Zn9La1, Zn1La1 and Zn0La10, suggesting that there is anoptimal La loading to maximize surface acid sites. Combining boththe acid and base sites, Zn3La1 again show the highest total surfacepercentage of Olat, Zn2+ and La3+, as compared to other catalysts.This correlates well with the transesterification activities (Figs. 1and 2) and XRD findings (Table 2).

Shifts of binding energies of Olat, Zn2p and La3d are presented inTable 2. Binding energies of Olat in Zn9La1, Zn3La1 and Zn1La1 arehigher than Zn10La0, while Zn2p in Zn9La1, Zn3La1 and Zn1La1are lower than Zn10La0. Moreover, binding energies of La3d

decrease with the La content. The shift of binding energy can beattributed to the electron transfer from lattice oxygen atoms tometal atoms. This suggests that La3+ acts as an electron donor,enhancing the interaction of reactant molecules with catalystsurfaces [36].

3.7. Reaction mechanism of simultaneous transesterification and

esterification

The reaction mechanism of simultaneous esterification andtransesterification is suggested in Figs. 14 and 15. The esterifica-tion takes place between the adsorbed fatty acids and freemethanol. The interaction of the carbonyl oxygen of fatty acid withthe Lewis acidic site (L+) of the catalyst forms carbocation. Thenucleophilic attack of alcohol to the carbocation produces a

tetrahedral intermediate (Fig. 13). During esterification thetetrahedral intermediate eliminates water molecule to form onemole of methyl ester. The transesterification takes place betweenthe adsorbed methanol and triglyceride. The transesterificationmechanism can be extended to di- and mono-glyceride. Methanolis adsorbed on the Lewis base site (B�) of the catalyst and formsoxygen anion. The nucleophilic attack of alcohol to the estersproduces a tetrahedral intermediate (Fig. 14). Then the hydroxylgroup breaks and forms two kinds of esters. In both cases,esterification and transesterification use of excess methanol favorsforward reaction and thus maximizes the FAME yield.

4. Conclusion

The synthesis of FAME from unrefined and waste oils wasinvestigated using a series of zinc and lanthanum mixed oxidescatalysts. There was a strong interaction between zinc andlanthanium species, and the catalyst with 3:1 of zinc to lanthanumwhich has shown a higher activity than pure metal oxides. Wehave found that at 200 8C, Zn3La1 was highly tolerant to FFA andwater, active in both transesterification and esterificationreactions, and with no hydrolysis activity. Within 3 h, 96% yieldof FAME was obtained even with crude palm oil, crude soybean oil,waste cooking oil, food-grade soybean oil with 3% water and 5%oleic acid addition. A temperature window was suggestedbetween 170 8C and 220 8C to change unrefined and waste oilsinto FAME based on Zn3La1 catalyst. Lanthanum promoted ZnOdispersion, increased the surface amounts of acid and base sites,thus enhanced the catalyst ability in both transeseterification andesterification reaction. The zinc and lanthanum mixed oxidescatalyst allows the direct use of unrefined and waste oils forbiodiesel production. Using this class of catalysts, which isrelatively inexpensive because of low raw materials andmanufacturing cost, significantly simplifies the oil pretreatmentprocess and product purification process, and greatly decreasesthe production cost of biodiesel.

S. Yan et al. / Applied Catalysis A: General 353 (2009) 203–212212

Acknowledgements

Financial support from the Department of Energy (grantDEFG36-05GO85005) and Michigan’s 21st Century Job Fund isgratefully acknowledged.

References

[1] S.J. Clark, L. Wagner, M.D. Schrock, J. Am. Chem. Soc. 61 (1984) 1632–1638.[2] P.R. Muniyappa, S.C. Brammer, H. Noureddini, Bioresour. Technol. 6 (1996) 19–24.[3] E. Crabbe, H.C. Nolasco, G. Kobayashi, K. Sonomoto, Process Biochem. 37 (2001)

65–71.[4] B. Freedman, E.H. Pryde, T.L. Mounts, J. Am. Oil Chem. Soc. 61 (1984) 1638–1643.[5] F. Ma, L.D. Clements, M.A. Hanna, Trans. ASAE 41 (1998) 1261–1264.[6] M.J. Haas, A.J. McAloon, W.C. Yee, T.A. Foglia, Bioresour. Technol. 97 (2006) 671–

678.[7] H.L. Ngo, N.A. Zafiropoulos, T.A. Foglia, Energy Fuels 22 (2008) 626–634.[8] M. Canakci, J. Gerpen, Trans. ASAE 44 (2001) 1429–1436.[9] B. Narendra, D.C. Drown, K. Roger, H. Dwight, U.S., 1995.

[10] G. Vicente, M. Martınez, J. Aracil, Bioresour. Technol. 92 (2004) 297–305.[11] K. Ishihara, M. Nakayama, S. Ohara, H. Yamamoto, Tetrahedron 58 (2002) 8179–

8188.[12] F. Omota, A.C. Dimian, A. Bliek, Chem. Eng. Sci. 58 (2003) 3175–3185.[13] G.J. Suppes, D. M. A., E.J. Doskocil, Appl. Catal. A 257 (2004) 213–223.[14] M.D. Serio, M. Ledda, M. Cozzolino, Ind. Eng. Chem. Res. 45 (2006) 3009–3014.

[15] C.V. McNeff, L.C. McNeff, B. Yan, D.T. Nowlan, A.E. Gyberg, B.J. Krohn, R. Fedie, T.R.Hoye, M. Rasmussen, Appl. Catal. A: Gen. 343 (2008) 39–48.

[16] T. Lacome, G. Hillion, B. Delfort, R. Revel, S. Leporg, G. Acakpo, US, 2005.[17] S.L. Yan, H.F. Lu, B. Liang, Energy Fuels 22 (2007) 646–651.[18] L. Bournay, G. Hillion, European (2003).[19] P.S. Sreeprasanth, R. Srivastava, D. Srinivas, Appl. catal. A 314 (2006) 148–159.[20] B. Guiffard, M. Troccaz, Mater. Res. Bull. 33 (1998) 1759–1768.[21] H. Noureddini, D. Zhu, J. Am. Oil Chem. Soc. 74 (1997) 1457–1463.[22] J.M. Marchetti, V.U. Miguel, A.F. Errazu, Renewable Sustainable Energy Rev. 11

(2007) 1300–1311.[23] M. Canakci, J.V. Gerpen, Trans. ASAE 42 (1999) 1203–1210.[24] H. Fukuda, A. Kondo, H.J. Noda, Biosci. Bioeng. 92 (2001) 405–416.[25] M.G. Kulkarni, A.K. Dalai, Ind. Eng. Chem. Res. 45 (2006) 2901–2913.[26] D. Kusdiana, S. Saka, Bioresour. Technol. 91 (2004) 289–295.[27] S.L. Yan, School of Chemical Engineering, Sichuan University, Chengdu, 2007, p.

115.[28] R.A. Santen, M. Neurock, Molecular Heterogeneous Catalysis, WILEY-VCH Verlag

GmbH & Co. KGaA, 2006.[29] I. Chorkendorff, J.W. Niemantsverdriet, Concepts of Modern Catalysis and

Kinetics, WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim, 2003.[30] J. Wahlena, D.E.D. Vos, P.A. Jacobsa, J. Catal. 249 (2007) 15–23.[31] Z. Han, Q. Yang, G.Q. Lu, J. Solid State Chem. 177 (2004) 3709–3714.[32] V.E. Henrich, P.A. Cox, The Surface Science of Metal Oxides, Cambridge University

Press, Cambridge, 1994.[33] J.V. Gerpen, Fuel Process. Technol. 86 (2005) 1097–1107.[34] H.E. Hoydonckx, D.E. Vos, S.A. Chavan, Top. Catal. 27 (2004) 83–96.[35] J. Otera, Angew. Chem. 113 (2001) 2099–2106.[36] L.Q. Jing, X.J. Sun, B.F. Xin, J. Solid State Chem. 177 (2004) 3375–3382.