Applied Catalysis A: General 366 (2009) 154159

Contents lists available at ScienceDirect

Applied Catalysis A: General

journal homepage: www.elsevier.com/locate/apcata

Optimization of the activity of CaO/Al2O3 catalyst for biodiesel production using

response surface methodology
Masoud Zabeti, Wan Mohd Ashri Wan Daud *, Mohamed Kheireddine Aroua
Chemical Engineering Department, Faculty of Engineering, University Malaya, 50603, Kuala Lumpur, Malaysia

A R T I C L E I N F O A B S T R A C T

Article history: In this work the response surface methodology (RSM) in conjunction with the central composite design
Received 26 February 2009 (CCD) were used to optimize the activity of CaO/Al2O3 solid catalysts for the production of biodiesel. In
Received in revised form 29 June 2009 order to measure the catalyst activity, we used palm oil as a representative raw material for the
Accepted 30 June 2009
conversion to biodiesel. The biodiesel production was carried out in a batch laboratory scale reactor.
Available online 9 July 2009
The results showed that both the calcination temperature and the amount of calcium oxide loaded on the
support had signicant positive effects on the biodiesel yield. The maximum basicity and biodiesel yield
Keywords:
obtained were about 194 mmol/g and 94%, respectively. Overall, the catalyst showed high performance
Alumina
Calcium oxide
at moderate operating conditions and its activity was maintained after two cycles.
Optimization 2009 Elsevier B.V. All rights reserved.
Solid catalyst
Transesterication

1. Introduction therefore can be easily separated from the product, regenerated

and reused and the catalyst also can be applied in a continuous
Biodiesel is a mixture of fatty acid alkyl esters. It can be process. Nevertheless, the processes using heterogeneous catalysts
produced from sources such as vegetable oils and animal fats via a cannot be performed under mild conditions and normally need
transesterication reaction with an alcohol [1]. The work on high molar ratios of alcohol to oil, large amounts of catalyst
biodiesel has gained a lot of interest, thanks to its biodegradability, content, high temperatures and long reaction times to overcome
exceptional cetane numbers (as high as diesel), high ash point, diffusion problems owing to the formation of three phases of the
low viscosity, lower combustion emission proles, ability to be reactants [7,8]
blended with fossil-based diesel at any proportions and ability to The existence of three phases with alcohol and oil leads to
be used in conventional diesel engines with no further modica- diffusion limitations, thus lowering the rate of reaction. One of the
tions [2]. Homogeneous acid and base catalysts such as potassium ways to overcome this mass transfer problem is to use structure
hydroxide [3], sodium hydroxide [4] and sulfuric acid [5] have been promoters or catalyst supports which can provide more specic
conventionally used for biodiesel production. The process is surface area and pores for active species, where they can be
normally performed under mild conditions and produces high anchored and react with large triglyceride molecules.
amount of methyl esters in a short reaction time due to the Gamma aluminum oxide has been widely used as a support of
formation of two phases in the reaction medium. However, a active species in industrial processes such as reforming, hydro-
purication step is required to remove the catalyst from products genation and steam reforming due to its high thermal stability and
and this is difcult because of its solubility in the reaction medium. specic surface area (as high as 300 m2/g), mesopore size of 5
This process leads to generation of large amount of wastewater 15 nm, pore volume of 0.6 cm3/g and ability to be shaped into
containing liquids of high basicity or acidity which are not pellets and extruded [9]. In case of biodiesel production, alumina
environmentally benign [6]. was applied as a support for catalysts such as sodium hydroxide
Solid base catalysts have been considered as an alternative to [10] and potassium oxide [11].
homogeneous base catalysts due to several advantages: the Calcium oxide has been reported to be used in transesterica-
catalysts are not dissolved or consumed in the reaction and tion reactions since it possesses basicity as high as barium and
strontium oxide and its sources such as calcium carbonate and
calcium acetate are cheaper and less toxic [1214]
* Corresponding author. Tel.: +60 3 79675297; fax: +60 3 79675319.
In this work calcium acetate as an environmentally friendly and
E-mail addresses: [email protected] (M. Zabeti), cheap precursor of calcium oxide was used to synthesis a high
[email protected] ( Daud), [email protected] (M.K. Aroua). performance solid base catalyst that consists of calcium oxide

0926-860X/$ see front matter 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2009.06.047
 M. Zabeti et al. / Applied Catalysis A: General 366 (2009) 154159 155

supported on g-alumina. The activity of the catalyst for (TCD). Three hundred milligrams of sample was treated at 300 8C
transesterication reaction was optimized using response surface for 1 h by passing helium at a ow rate of 20 ml/min. Then, the
methodology (RSM). According to the literature, the activity of a sample was saturated with carbon dioxide at a ow rate of 20 ml/
solid catalyst is highly correlated with the active site concentration min and at 100 8C for 1 h. Subsequently, the system was ushed
of the catalyst. From the literature it was also found that the with helium (20 ml/min) at 100 8C for 1 h to eliminate physisorbed
basicity of the catalyst is strongly inuenced by the calcination CO2. Desorption was carried out from room temperature to 800 8C
temperature and by the loading amount of active species on the at a heating ramp of 5 8C/min.
support [1517]. As a consequence, a set of experiments was Leaching of the catalyst into the reaction medium was
designed in the specied region of the two variables mentioned determined using the inductively coupled plasma mass spectro-
and several catalysts were synthesized in order to investigate the metry (ICP-MS) technique according to ASTM standard (D 6751).
conditions which give maximum activity for biodiesel production.
2.5. Analysis of the biodiesel
2. Materials and methods
Analysis of biodiesel samples were carried out by using a HP
2.1. Chemicals 6890 Gas Chromatogram (GC) equipped with a ame ionization
detector (FID) and capillary column DB23 (60 m  0.25 m 
Palm oil was obtained from a local market and used as the 0.15 mm). One hundred mg/ml n-hexane solutions of the biodiesel
feedstock for transesterication reactions. Other materials used samples were injected by an auto injector at an oven temperature
were g-alumina (anhydrous, specic surface area according to of 50 8C, which then was heated up to 230 8C. Helium was used as
BET: 120190 m2/g) and methanol (for synthesis, >99.5%); normal the carrier gas and the injector temperature and the detector
hexane (GC grade, >99%); other standards for GC column temperature were 250 8C and 280 8C, respectively. The yields were
calibration were obtained from Merck. Calcium acetate mono- calculated from the equation below:
hydrated (A.C.S reagent) was supplied by SigmaAldrich.
weight of biodiesel produced  total wt:% of FAME
Yield %
2.2. Catalyst preparation weight of oil
 100
All catalysts were prepared using an impregnation method of (1)
aqueous solution of calcium acetate on alumina support. In all
cases, 8 g of g-alumina in powder form was preheated at 600 8C for here FAME means: fatty acid methyl esters.
1 h to remove physisorbed water. For each catalyst, a desired
amount of precursor was dissolved in 50 ml of distilled water and 2.6. Design of experiments and optimization method
introduced onto the alumina in a glass ask; the mixture was
stirred for 4 h at room temperature. The obtained precipitation was Applications of central composite design (CCD) and response
heated overnight in an oven in order to remove water through surface methodology (RSM) techniques to design and optimize the
evaporation. Prior to the reaction, all catalysts were calcined in air chemical and environmental processes have been recently
at desired temperatures for 5 h. reported in the literature and their reliability to generate a model
equation and to calculate optimum conditions have been proven
2.3. Reaction procedure [1822]. Accordingly, in this work, experiments were designed by
Design Expert software (version 7.1) using a ve-level-two-factor
The activities of the catalysts were measured based on the CCD to investigate the effect of the parameters on the catalyst
results of transesterication reactions of palm oil with methanol. activity for biodiesel production and to obtain a good model
All experiments were carried out using a 150 ml glass-jacketed equation to predict the optimum catalyst preparation conditions.
reactor equipped with a digital magnetic stirrer and a water-cooled The factors selected for the optimization of the catalyst activity
condenser. Fifty grams of oil was charged into the reactor, followed were: calcination temperature and dosage of CaO precursor.
by methanol with a ratio of 12:1 to oil; then 3.5 wt.% of the catalyst Basicity of the catalyst and biodiesel yields were chosen as
(on the basis of the weight of oil) were added and the mixture was responses. Experiments were carried out in randomized order with
rigorously stirred for 5 h. The mixture was heated up to 65 8C by ve replications of centre point in order to obtain a good
means of a water bath. After the specied reaction time was estimation of errors. The model equation of all signicant variables
reached, the catalyst was separated by centrifuging and the was generated by the software and RSM was used to predict
glycerol was separated in a separation funnel; nally the residual optimum points. The actual and coded values of each parameter
alcohol was evaporated by means of heating. The methyl ester are presented in Table 1. The coded values were designated by 1
yields were determined and calculated using results from gas (minimum), 0 (medium), +1 (maximum), a and +a. Alpha is
chromatography. dened as a distance from the centre point which can be either
inside the range or outside, with the maximum value of 2n/4,
2.4. Catalyst characterization where n is the number of factors [23]. Hereby the value of alpha is
set at 0.5.
Basicity of the catalysts was measured by using temperature Compared to the classical approach, which is one factor at the
programmed desorption (TPD) method of CO2 on Thermo Finnigan time, this technique is able to determine the interaction effects of
(TPDRO 1100) equipped with a thermal conductivity detector variables, in addition to giving good estimations of errors.

Table 1
Actual and coded values of the variables.

Factor Units Low level High level Low axial High axial Mean

CaO precursor dosage wt.% 80 (1) 120 (+1) 90 (a) 110 (+a) 100 (0)
Calcination temperature 8C 600 (1) 800 (+1) 650 (a) 750 (+a) 700 (0)
156 M. Zabeti et al. / Applied Catalysis A: General 366 (2009) 154159

Table 2
Experimental design layout and experimental results of the responses (experiments were carried out according to the run order).

Run Type Factor A Factor B Factor A Factor B Response 1 Response 2

Actual value (wt.%) Actual value (8C) Coded value Coded value Yield (wt.%) Basicity (mmol/g)

10 Fact 80 600 1 1 5.23 58.88

11 Fact 120 600 +1 1 8.83 61.32
13 Fact 80 800 1 +1 38.49 122.85
6 Fact 120 800 +1 +1 39.43 100
5 Axial 90 700 0.5 0 86.06 172.87
8 Axial 110 700 +0.5 0 91 186.89
4 Axial 100 650 0 0.5 81.13 161.34
3 Axial 100 750 0 +0.5 90 179.42
1 Center 100 700 0 0 94.37 202.63
2 Center 100 700 0 0 92.15 190.89
7 Center 100 700 0 0 93 195.84
9 Center 100 700 0 0 82 187.89
12 Center 100 700 0 0 90.37 191

Note: A, calcium acetate dosage; B, calcination temperature.

Moreover, the time and cost of experiments could be reduced as Table 3

Statistical analysis of variance for the two responses.
the total number of trials is minimized.
A preliminary study was carried out to investigate the levels of Source p-value, Prob. > F
each factor. The lower level of calcination temperature was 600 8C,
Yields Basicity
because below this temperature the catalyst was almost inacti-
vated and dominated by CaCO3, which is the reason for the catalyst Model <0.0001, signicant <0.0001, signicant
A: calcium precursor dosage 0.5447 0.4217
deactivation. The upper limit of calcination temperature, 800 8C,
B: calcination temperature 0.0004 0.0002
was determined by preliminary study; results showed that methyl AB 0.8052 0.1314
ester yields reduced dramatically beyond this temperature. In A2 0.0295 0.0573
addition, at temperatures greater than 800 8C, the gamma phase of B2 0.0063 0.0019
Lack of t 0.3941, not signicant 0.1981, not signicant
alumina which is highly porous is transformed to theta phase and
consequently to alpha phase, which is non-porous [24]. Limits of R-squared 0.9857 0.9879
calcium oxide precursor dosage were assigned to be 80 wt.% and Adj. R-squared 0.9756 0.9792
Pred. R-squared 0.8887 0.7732
120 wt.% (based on the initial support) since preliminary study
Adeq. precision 24.549 26.70
revealed that, for dosages outside of this range, insignicant
amount of yields were obtained. Adeq.: adequate; Adj.: adjusted; Pred. predicted; Prob.: probability.

3. Results and discussions

tness of the model as well as the effects of signicant individual
3.1. Data analysis terms and their interactions on the chosen responses. From the
table it can be seen that the model was signicant for both
The experimental points, in coded and actual values, and the responses with p-values less than 0.0001 (model and term p-value
response values are represented in Table 2. The experiments were <0.05 indicates the model is signicant for 95% condence
performed in random order to minimize errors due to effects of intervals) to predict the response values. Calcination temperature,
lurching factors, such as time, on the response. quadratic term of calcium oxide precursor dosage and quadratic
A quadratic polynomial equation was generated to t the term of calcination temperature were signicant model terms for
experimental data using response surface methodology for both responses with p-values less than 0.05. In addition, effects of
optimization of experimental parameters, as shown below: CaO precursor dosage and calcination temperaturecalcium oxide
Equations in terms of coded factors: precursor dosage interactions were insignicant (term p-value
>0.100 indicates the model terms are not signicant). Insignicant
Basicity 191:49  2:98A 24:82B  6:32AB  34:05A2 terms can be manually removed from the model to improve the
 72:05B2 (2) regression model and optimization results. However, in this
optimization, removing interaction term of AB from the model
resulted in small changes in the optimization results which can be
Yields 92:57 1:56A 15:18B  0:66AB  28:66A2
neglected. Therefore, we preferred to keep the interaction term of
 40:52B2 (3) AB in the model in order to discuss the effect of the interaction term
on the responses. Moreover, the individual term of A cannot be
Equations in terms of actual factors: removed from the model since it is required to support the model
hierarchy.
Basicity 4570:70947 19:09139A 10:652B  3:16125
Lack of t, which is the weighted sum of squared deviations
 103 AB  0:085137A2  7:20548  103 B2 (4) between the mean response at each factor level and the
Yields 2746:7671 14:641A 5:85792B  3:325 corresponding tted value, is not signicant for both responses
 104 AB  0:071652A2  4:05207  103 B2 (5) (lack of t p-value >0.05 is not signicant); this indicates that the
model is tted to all data. (Non-signicant lack of t is good).
where A: CaO precursor dosage in wt.%; B: calcination Predicted R-squared, which is a measure of how good the model
temperature in 8C. predicts values for the response, is within 0.20 of adjusted R-
Table 3 shows the results of a statistical analysis of variance squared, for both responses; this conrms that there is no problem
(ANOVA) which was carried out to study the signicance and with either experimental data or the model. Adequate precision is
 M. Zabeti et al. / Applied Catalysis A: General 366 (2009) 154159 157

a measure of signal to noise ratio and as requirement of the model a of catalyst precursor dosage becomes less signicant. Therefore,
ratio of greater than 4 is desirable. In this model the ratios of 24.54 the maximum amount of basicity achieved is around the middle
and 26.70 for biodiesel yields and basicity, respectively, are much range of the corresponding variables.
greater than 4 and the result indicates adequate model discrimi-
nation. 3.3. Effect of the parameters on the biodiesel yields

3.2. Effect of the parameters on the basicity of the catalyst Fig. 1(b) and Eq. (3) indicate trends of variables effectiveness on
biodiesel yields which is nearly identical to the basicity data of the
A-three-dimensional graph of second order predicted model of catalyst. In this context, individual terms of catalyst precursor
the catalyst basicity response is shown in Fig. 1(a). As can be seen dosage and calcination temperature have positive inuences on
from the gure, the basicity of the catalyst increased when the the methyl ester yields, while the parameters interactions and
values of the catalyst precursor dosage and calcination tempera- their second order terms have negative effects on the methyl ester
ture increased to around their mid-range due to the fact that these yields. In fact, by increasing the values of each parameter to
factors have positive effects on the response (Eq. (2)). However, approximately the middle of the corresponding range, the yields
this trend is more obvious with regard to the calcination reached the maximum amount; after that they began to drop
temperature which has more signicant inuence on the response. gradually. From Fig. 1(a) and (b) a good correlation between the
An increase in the value of variables beyond a certain amount of basicity and activity of the catalyst can be observed.
the mid-range leads to a slight reduction in basicity of the catalyst.
This is owing to the fact that the quadratic term of the parameters 3.4. Investigation of the optimum operating conditions
is more signicant with a negative inuence on the basicity.
As shown in Fig. 1(a) and Eq. (2), the calcination temperature The optimum operating conditions are those points at which
catalyst precursor dosage interaction is insignicant, but its effect the maximum amounts of basicity of the catalyst and consequently
on the basicity of the catalyst is negative. As the calcination the maximum biodiesel yields are achieved. The contour plots of
temperature increases to nearly the middle of its range, the effect the second order model for each response at the experimental
ranges of each factor were represented in Fig. 2(a) and (b) in
addition to the overlay plot of the responses (Fig. 3). According to
Fig. 2(a) and (b), each contour represents pairs of points at which
the same results of basicity or yield is obtained. For instance, by
loading 95% of calcium oxide precursor on the support and
calcination temperature of 690 8C the yield of 88% is obtained.
Fig. 3 demonstrates a marked area of variable ranges at which the
basicity and yields of more than 170 mmol/g and 80% are obtained,
respectively. The optimum operating conditions to achieve the
most active catalyst, according to Fig. 3, are loading 100.54 wt.% of
calcium acetate on alumina support and calcination temperature
at 718 8C. In this regard, the amount of calcium incorporated on
alumina was measured to be 8.84 wt.% by ICP-MS. An increase in
calcium oxide loading onto the support beyond 100% could lead in
partial blockage of active sites, thus reducing the basicity of the
catalyst and consequently the activity of the catalyst [15].
Moreover, the gamma phase of alumina which is the highest
porous form of alumina is transformed to theta phase with less
porosity and specic surface area when the calcination tempera-
ture rises beyond 800 8C and accordingly the amount of active site
concentration and activity of the catalyst are declined [24].
An experiment was carried out using the catalyst which was
synthesized by calcination at 718 8C and loading 100.54 wt.% of
calcium precursor on alumina to evaluate the accuracy of the
second order equations. The experimental and predicted results
are shown in Table 4.

3.5. Evaluation of the catalyst reusability

The reusability of the optimum catalyst was tested by applying

it for two subsequent cycles; after the second cycle, the
performance of the catalyst was sustained with nearly 91% of
methyl ester yields. The slight decrease in the activity of the
catalyst might be due to the reaction between calcium oxide and
glycerol in the transesterication reaction and to the formation of
calcium diglyceroxide which has been described by Kouzu et al.
[25]. After the second cycle, the catalyst was collected by ltration
and washed with methanol for several times, followed by
calcination in air at 718 8C. The collected catalyst was applied in
Fig. 1. Response surface plot in the studied range of the variables for (a) the a transesterication reaction with operating conditions described
predicted basicity (b) the predicted yields. Dos.: dosage; Pre.: precursor; Temp.: in section 2.3 and the methyl ester yields obtained were 94.23%
temperature. which is close to the activity of the fresh catalyst. This can be
158 M. Zabeti et al. / Applied Catalysis A: General 366 (2009) 154159

Fig. 3. Overlay plot of the basicity and yields greater than 170 mmol/g and 80%. Dos.:
dosage; Pre.: precursor; Temp.: temperature.

Table 4
Experimental and predicted results of the basicity and yields responses for the
optimum catalyst (calcination temperature, 718 8C; calcium precursor dosage,
100.54 wt.%).

Response Experimental results Prediction value

Basicity 190 mmol/g 193.483 mmol/g

Yields 95% 94.0043%

biodiesel production from palm kernel oil; despite the reusability

of the catalyst, a large amount of methanol/oil molar ratio, 65:1,
and the catalyst concentration of 10 wt.%, was required to achieve
the methyl ester yields of 94% after 3 h in addition, a substantial
decrease in the activity of the catalyst was observed when the
catalyst was synthesized at calcination temperatures greater than
450 8C, which limits its applications for lower temperatures.

4. Conclusions

Fig. 2. Contour plot in the studied range of variables for (a) the predicted basicity (b) A response surface methodology was applied to investigate the
the predicted yields. Dos.: dosage; Pre.: precursor; Temp.: temperature. optimum conditions for catalyst activity. Loading amount of
catalyst precursor on the support and calcination temperature
were parameters that affected the basicity of the catalyst and
explained by the fact that during the calcination calcium consequently the activity of the catalyst. However, calcination
diglyceroxide is transferred to calcium oxide. Moreover, the temperature had more signicant effect on the catalyst activity. A
amount of calcium in the biodiesel product after each cycle was good correlation was found between the basicity of the catalyst
measured by ICP-MS and 31 ppm and 12 ppm of calcium after rst and the amounts of yields obtained. The optimum catalyst activity
and second cycle were detected, respectively. The results indicate for transesterication reaction was achieved by loading
that the amount of the catalyst leaching into the reaction medium 100.54 wt.% of the calcium oxide precursor on alumina, followed
is insignicant. by calcination at 718 8C. The catalyst was reusable for two cycles;
The catalyst which was reported in this study provided high based on the ICP-MS results, no leaching of the active species were
performance towards transesterication reaction using mild detected in the product.
operating conditions, while an insignicant amount of the catalyst
was observed in the obtained biodiesel. However, when Sr(NO3)2/ Acknowledgement
ZnO [15] and SrO [26] were used in transesterication reaction,
signicant amounts of catalysts were leached into the reaction This study was carried out with the aid of a research grant from
medium. Benjapornkulaphong et al. [27] used Ca(NO3)2/Al2O3 for University Malaya Research Fund.
 M. Zabeti et al. / Applied Catalysis A: General 366 (2009) 154159 159

References D.C.S. Azevedo, C.L. Cavalcante Jr., P. Maireles-Torres, Appl. Catal., A: Gen. 334
(2008) 3543.
[1] A.K. Dalai, M.G. Kulkarni, L.C. Meher, IEEE EIC Climate Change Tech. Conf. EICCCC [13] A. Demirbas, Energy Convers. Manage. 48 (2007) 937941.
2006 art. no. 4057358. [14] M.L. Granados, M.D.Z. Poves, D.M. Alonso, R. Mariscal, F.C. Galisteo, R. Moreno-
[2] D.E. Lopez, J.G. Goodwin Jr., D.A. Bruce, E. Lotero, Appl. Catal., A: Gen. 295 (2005) Tost, J. Santamaria, J.L.G. Fierro, Appl. Catal., B: Environ. 73 (2007) 317326.
97105. [15] Z. Yang, W. Xie, Fuel Process. Technol. 88 (2007) 631638.
[3] S. Baroutian, M.K. Aroua, A.A.A. Raman, N.M.N. Sulaiman, J. Chem. Eng. Data 53 [16] W. Xie, X. Huang, Catal. Lett. 107 (2006) 5359.
(2008) 877880. [17] S. Ramu, N. Lingaiah, B.L.A. Prabhavathi Devi, R.B.N. Prasad, I. Suryanarayana, P.S.
[4] G. Arzamendi, I. Campo, E. Arguinarena, M. Sanchez, M. Montes, L.M. Ganda, Sai Prasad, Appl. Catal., A: Gen. 276 (2004) 163168.
Chem. Eng. J. 134 (2007) 123130. [18] A. DCruz, M.G. Kulkarni, L.C. Meher, A.K. Dalai, JAOCS 84 (2007) 937943.
[5] L. Bournay, D. Casanave, B. Delfort, G. Hillion, J.A. Chodorge, Catal. Today 106 [19] A.K. Tiwari, A. Kumar, H. Raheman, Biomass Bioenergy 31 (2007) 569575.
(2005) 190192. [20] S. Ghafari, H.A. Aziz, M.I. Hasnain, A.K. Zinatizadeh, J. Hazard. Mater. 163 (23)
[6] T.F. Dossin, M.F. Reyniers, R.J. Berger, G.B. Marin, Appl. Catal., B: Environ. 67 (2006) (2009) 650656.
136148. [21] A.A.G. Lima, M. Nele, E.L. Moreno, H.M.C. Andrade, Appl. Catal., A: Gen. 171 (1)
[7] I.K. Mbaraka, B.H. Shanks, JAOCS 83 (2006) 7991. (1998) 3143.
[8] T.F. Dossin, M.-F. Reyniers, R.J. Berger, G.B. Marin, Appl. Catal., B: Environ. 67 [22] W.D. Yang, K.M. Hung, C.S. Hsieh, Mater. Sci. Eng., A 333 (12) (2002) 123
(2006) 136148. 133.
[9] I. Chorkendorff, J.W. Niemantsverdriet, Concepts of Modern Catalysis and [23] G. Vicente, M. Martinez, J. Aracil, Bioresour. Technol. 98 (2007) 17241733.
Kinetics, second ed., Wiley-VCH, Germany, 2003. [24] J. Kuo, D.L. Bourell, J. Mater. Sci. 32 (1997) 26872692.
[10] H.-J. Kim, B.-S. Kang, M.-J. Kim, Y.M. Park, D.-K. Kim, J.-S. Lee, K.-Y. Lee, Catal. [25] M. Kouzu, T. Kasuno, M. Tajika, S. Yamanaka, J. Hidaka, Appl. Catal., A: Gen. 334
Today 9395 (2004) 315320. (2008) 357365.
[11] W. Xie, H. Peng, L. Chen, App. Catal., A: Gen. 300 (2006) 6774. [26] X. Liu, H. He, Y. Wang, S. Zhu, Catal. Commun. 8 (2007) 11071111.
[12] M.C.G. Albuquerque, I. Jimenez-Urbistondo, J. Santamaria-Gonzalez, J.M. [27] S. Benjapornkulaphong, C. Ngamcharussrivichai, K. Bunyakiat, Chem. Eng. J. 145
Merida-Robles, R. Moreno-Tost, E. Rodriguez-Castellon, A. Jimenez-Lopez, (2009) 468474.