1-s2.0-s0926860x09002981-main.pdf
TRANSCRIPT
8/11/2019 1-s2.0-S0926860X09002981-main.pdf
http://slidepdf.com/reader/full/1-s20-s0926860x09002981-mainpdf 1/6
Reaction kinetics of the catalytic esterification of acrylic acidwith propylene glycol
Mehmet R. Altıokka *, Elif Odes
Department of Chemical Engineering, Anadolu University, 26470, Eskisehir, Turkey
1. Introduction
Various acrylic esters are useful chemicals. Their excellentclarity, toughness, color retention, UV stability and chemicalinertness make acrylic ester emulsion polymers prime paint
binders. Acrylics are widely used in all types of paint formulations.Large volumes of acrylic emulsion polymers are used as binders forfiberfill and nonwoven fabrics, textile bonding or laminating,flocking, back coating and pigment printing binders. These are alsoused for leather finishing, textile and fiberfill bonding and as
adhesives [1].Among the other many acrylates, hydroxypropyl acrylates are
used in production of acrylate polymers [2–4] and as a co-
monomer in adhesives, elastomers [5], inks, radiation curable [6],automotive top coating, nanocomposites materials [7], photo cureresins and oil additives and as a viscosity reducer [8].
Hydroxypropyl acrylate contains both vinyl and hydroxyl
functionality. Vinyl allows copolymerization with a variety of other vinyl monomers in aqueous and nonaqueous systems. Wideco-monomer selection provides easy control of physical andchemical properties such as glass transition temperature and
solution viscosity. On the other hand hydroxyl functionality of hydroxypropyl acrylate provides the reaction readily witholigomeric di- or polyisocyanates, anhydrides, or epoxides. The
resulting adducts carry pendant vinyl unsaturation making theseresinsradiation curable by ultraviolet light or an electron beam [9].
Esterification of acrylic acid with alcohol has been performedcommercially by using homogeneous catalysts such as sulfuricacid, hydrofluoric acid and para-toluenesulfonic acid; but these aretoxic, corrosive and often hard to remove from the products. Thus
it is keenly desirable to develop new types of solid acid catalysts toreplace them, because the solid acid is less toxic and facilitates therecovery and recycling of catalysts [10,11].
No study has been reported in the literature for the esterifica-
tion of acrylic acid with propylene glycol. However, a number of studies related to the catalytic esterification reactions of acrylicacid with some other alcohols have been reported. Although the
catalytic activities of the various types of catalysts were comparedwith each other, kinetic information has not been given in thesestudies [1,12–18].
Esterification reaction between acrylic acid and 2-ethylhexanol
was investigated in an isothermal semi-batch reactor by usingsulfuric acid as a homogeneous catalyst. The reaction kinetics wasfound to be second order [14].
Esterification of acrylic acid with 1-butanol has been studied in
the presence of Cs2.5H0.5PW12O40, solid oxides and organic resinsas heterogeneous catalysts. In this solid–liquid reaction system itwas found that Cs2.5H0.5PW12O40 exhibited the highest catalyticactivity, in the unit of catalyst weight, among the solid oxide
catalysts, while the activity Cs2.5H0.5PW12O40 was less than thoseof organic resins such as Nafion and Amberlyst-15 [16].
In the study of transesterification of cyclohexyl acrylate withn-butanol and 2-ethylhexanol, it was found that while the acid
Applied Catalysis A: General 362 (2009) 115–120
A R T I C L E I N F O
Article history:
Received 8 January 2009
Received in revised form 12 March 2009Accepted 15 April 2009
Available online 23 April 2009
Keywords:
Esterification
Hydroxypropyl acrylate
Heterogeneous catalysis
Kinetic modeling
A B S T R A C T
The kinetics of esterification of acrylic acid with propylene glycol has been studied in the presence of
Amberlyst-15 as a heterogeneous catalyst. The reaction was realized in a batch reactor at different
temperatures and initial reactant mol ratios. The simultaneous dimerization/polymerization of acrylic
acid and products, in addition to the reversible esterification reaction, was proposed as the reaction
mechanism. Kinetic equations corresponding to the reaction mechanism have also been developed.
Temperature dependence of the reaction rates and activation energies was determined.
2009 Elsevier B.V. All rights reserved.
* Corresponding author. Tel.: +90 222 321 35 50/6505; fax: +90 222 323 95 01.
E-mail address: [email protected] (M.R. Altıokka).
Contents lists available at ScienceDirect
Applied Catalysis A: General
j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / a p c a t a
0926-860X/$ – see front matter 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2009.04.028
8/11/2019 1-s2.0-S0926860X09002981-main.pdf
http://slidepdf.com/reader/full/1-s20-s0926860x09002981-mainpdf 2/6
treated clay, Engelhard F-24, was an effective catalyst for thetransesterification reaction at moderate temperatures, tetrabutyltitanates as catalyst increase cyclohexanol selectivity at highertemperatures. It was also shown that transesterification reaction
follows LHHW mechanism in the presence of Engelhard F-24 [1].In another study it has been reported that heteropoly acids,
such as phosphorous tungstic acid, show higher activities than theconventional acids in the esterification of methacrylic acid with
tripropylene glycol [17].The role of Mn andFe as promoters of sulfated zirconiahas been
studied in the reaction of acrylic acid esterification by 1-butene tosec-butyl acrylate at 343 K. It was reported that Mn and Fe did not
improve the catalytic activity and selectivity to sec-butyl acrylatecompared with Amberlite resins or sulfuric acid but they stronglyresist deactivation [18].
In the present work, Amberlyst-15 and Amberlyst-36 andcesium salt of phosphorous tungstic acid were used as theheterogeneous catalysts for the esterification of acrylic acid with
propylene glycol. The best catalyst among them was to bedetermined in point of activity and selectivity towards hydro-xypropyl acrylate. Thereafter, the kinetic study of esterificationwasto be studied at the present of theselectedcatalysts among thecatalysts tested.
2. Experimental
2.1. Materials
Acrylic acid (>99.5%) and propylene glycol (>99%) wereobtained from Acros. Toluene, as a solvent, was supplied bySigma–Aldrich. Properties of catalysts used in this study are givenin Table 1. The catalysts were dried at 80 8C for overnight prior to
the experiment. Phenothiazine used as polymerization inhibitorwas obtained from Fluka.
2.2. Equipments
The reactor consisted of a two-necked spherical Pyrex flask of 250 ml capacity fitted with a spiral coil condenser and a sampledevice. The temperature was controlled within 0.1 K by circulat-
ing water from a thermostat into the water jacket of the reactor. The
reaction mixture was magnetically stirred at about 600 rpm.
2.3. Experimental procedure
In a typical run, toluene as a solvent and one reactant were
placed in the reactor. A known amount of catalyst and inhibitor
was added and the reactor contents were mixed well. After asteady value of desired temperature was attained, the second
reactant was added and this was taken as zero time for a run. Onemilliliters of the liquid sample was withdrawn from the reactor atregular intervals for analysis and immediately transferred to acrucible in an ice bath in order to ensure that no further reaction
took place.
2.4. Analysis
Thesamples were analyzedwith a gaschromatograph(HP 5890Series II) equipped with a flame ionization detector (FID) and acapillary column (TRB-FFAP). The column oven was held at 313 Kfor 3 min, ramped at 10 K/min to 333 K held for 3 min, followed by
10 K/min ramp-up to 433 K, held for 4 min and then ramped at10 K/min to 453 K, held for 2 min. The concentration of water inthe reaction mixture was measured by Karl Fischer titration(Metrohm KF-784). After verifying that the measured water
content corresponds to the calculated values based on stoichio-metric equation, the amount of water was determined fromstoichiometric mass balance equation.
3. Results and discussion
The effects of parameters, such as catalyst type and loading,
temperature and reactant mol ratio, on the reaction rate werestudied.
Swelling effect of the catalyst was neglected since preliminaryexperiment showed that the swelling process reaches the
equilibrium in a few minutes that the reactions were realized inthe presence of the reasonably stable catalyst. Therefore, theproposed model in this work comprises the effect of catalystswelling.
3.1. Catalyst selection
In the experiments, two types of ion exchange resins
(Amberlyst-15 and Amberlyst-36) as well as cesium salt of phosphorous tungstic acid exhibiting self-inhibiting effect forpolymerization [16], have been tested as the heterogeneouscatalysts. Experiments were carried out at 353 K and reactant mol
ratio of 1/1 with the same amount of catalyst andinhibitor. Resultsare given in Table 2. As seen from Table 2 that Amberlyst-36accelerates the reaction rate most. But its product selectivity isconsiderably lower than that of Amberlyst-15 at the same
conversion level of 0.45. On the other hand, althoughits selectivityis very high (100%), the activity of cesium salt of phosphoroustungstic acid is very low compared to that of the others and itshould also be noticed that its selectivity is given at very low
conversion of 0.04. It is believed that this high selectivity cannot besustained in entire range of conversion. Therefore Amberlyst-15
was chosen as a heterogeneous catalyst in this study.Effect of inhibitor, phenothizine,loading was also studied in the
presence of 0.1, 0.3 and 0.5 (wt%) in the reaction mixture. The bestratio was found to be 0.3 (wt%). This ratio was also proposed inliterature [12–14].
3.2. Effect of catalyst loading
Experiments were carried out in the presence of 3.11, 5.06 and8.46 wt% of Amberlyst-15 on dry basis keeping the otherparameters constant. At the conversion levels less than 10% the
initial reaction rate can be safely calculated by using Eq. (1):
r A0 ¼C A0 X A
t
(1)
Nomenclature
C concentration (mol/L)
X conversion
t time (min)
r reaction rate (mol/L min)
CsHPA cesium salt of phosphorous tungstic acid
T absolute temperature (K)k reaction rate constant (L/mol min)
K equilibrium constant
E activation energy (kJ/mol)
Subscripts
A, P, T and W acrylic acid, propylene glycol, hydroxypropyl
acrylate and water, respectively
M.R. Altıokka, E. O des / Applied Catalysis A: General 362 (2009) 115–120116
8/11/2019 1-s2.0-S0926860X09002981-main.pdf
http://slidepdf.com/reader/full/1-s20-s0926860x09002981-mainpdf 3/6
A plot of the reaction rate, calculated from Eq. (1), vs. catalystloading is given in Fig. 1. As seen in Fig. 1, the reaction rateincreases linearly, as expected, with catalyst loading since theactive surface area is proportional to the amount of catalyst.
Intersection of the line with ordinate will give the uncatalyzedreaction rate at given parameters. The mathematical expressionrelating the initial reaction rate to the catalyst loading can be
derived from Fig. 1 as follows:
r A0 ¼ 0:0002þ 0:0004ð%wtÞmol=L :min (2)
where(wt%)is theweight percent of the catalyst on drybasis in thereaction mixture. It should also be kept in mind that Eq. (2) is valid
only at the given parameters of 0.3 (wt%) inhibitor, 1/1 reactantmol ratio and 353 K at which the experiments were performed.However, it does not alter the conclusion that the general reactionrate increases linearly with the catalyst loading.
3.3. Effect of initial concentration
Initial reaction rate was determined by altering the concentra-tion of the component under investigation while keeping those of
the others constant. The results are shown in Figs. 2–5. Theordinate valuesof thesefigures were obtained from Eq.(1) by usingthe experimental data. It is evident from these figures that all
reactants and products are adsorbed on the catalyst since theinitial reaction rate is becoming nearly independent of theconcentrations as their corresponding values increase [19].However, the initial reaction rate, as seen in Figs. 2 and 4, is not
completely independent of the concentrations of acrylic acid andproduct. This is attributed to their dimerization/polymerizationreaction.
3.4. Kinetic modeling
As seen in Figs. 2–5, reaction rate is highly temperaturesensitive. It is almost doubling in every temperature rising of 10 K.Therefore, it is reasonable to accept that the overall reaction iscontrolled by the surface reaction since internal and external
Table 1
Properties of catalysts reported by manufacturer.
Amberlyst-15 Amberlyst-36 Cs2.5H0.5PW12O40
Manufacturer Rohm & Haas Co. Rohm & Haas Co. Nippon Organic Colour
& Chemical Company
Matrix Styrene-divinylbenzene
(macroreticular)
Styrene-divinylbenzene
(macroreticular)
–
Standard ionic form H+ H+ –
Surface area (m2/g) 50a 33 116a
% Moisture 48 55 –Particle size (mm) 0.23–0.6 0.6–0.85 –
Cation exchange capacity (mmol/g) 4.7 5.4 0.15a
Maximum operating temperature (K) 393 423 –
a Ref. [16].
Table 2
Effect of catalyst type on the reaction rate and product selectivity at 353 K and in the presence of 0.3 (wt%) inhibitor.
Catalyst Catalyst
loading (wt%)
Conversion
time (min)
Initial
concentration (M)
Conversion r AðavÞ ¼C A0 X A
t
(mol/L min 103)
Selectivity
C A0 C P0
CsHPAa 4.73 180 3.3 2.9 0.04 0.736 1.00
Amb–36 5.57 240 2.7 2.2 0.45 5.047 0.494
Amb–15 5.54 420 2.8 2.3 0.45 3.003 0.736
a Inhibitor is not used since it has self-inhibiting effect [16].
Fig. 1. Effectof catalystloading onthe reactionrate at353 K,0.3 (wt%)inhibitor,and
1/1 reactant mol ratio.
Fig. 2. Effectof acrylic acidconcentrationon theinitial reaction rate (C P0 = 0.83 mol/
L, 5 (wt%) catalyst on dry basis, and 0.3 (wt%) inhibitor).
M.R. Altıokka, E. O des / Applied Catalysis A: General 362 (2009) 115–120 117
8/11/2019 1-s2.0-S0926860X09002981-main.pdf
http://slidepdf.com/reader/full/1-s20-s0926860x09002981-mainpdf 4/6
diffusions, as the physical steps, are not strongly temperaturedependent. This conclusion was further confirmed by Eq. (2),
verifying the Madon Boudart test stating that reaction rate isproportional to the catalyst loading [20].
It is also reported that the influence of external and internaldiffusions can be neglected for most of the reactions catalyzed by
the Amberlyst series resins [21–29].
As concluded, from Figs. 2–5, that both reactants and productsare adsorbed on the catalyst surface. Under these assumptions the
general reaction stoichiometry and corresponding reaction rateexpression, in the form of Langmuir–Hinshelwood–Hougen–Watson model, can be written as:
CH2¼CHCOOHðAÞ
þCH3CHOHCH2OHðPÞ
, CH2¼CHCOOC3H6OHðTÞ
þH2OðWÞ
(3)
r 00A ¼ kðC AC P ðC TC W=K ÞÞ
ð1 þ K AC A þ K PC P þ K TC T þ K WC W þ K UC UÞ2
(4)
respectively, where U is solvent.Eq. (4) involves seven arbitrary constants that must be
evaluated. This makes it rather impractical. For this reason, some
simplified rate expressions, fitting experimental data, wereproposed in literature [30]. Furthermore, Eq. (3) will never beheld due to simultaneous polymerization of acrylic acid and
products.Taking intoaccount the general esterificationreactionas well as
polymerization of acrylic acid and products, the overall reaction
mechanism is proposed to be:
Aþ P,k1
k2
Tþ W; ðesterificationreactionÞ (5a)
2A)k3
dimer (5b)
2T)k4
dimer (5c)
Some of the dimer molecules formed here may transform intopolymer.
Assuming that each reaction step is elementary, the corre-
sponding rate expression can be written as follows:
dC Adt ¼ k1C AC P þ k2C TC W k3C 2A (6a)
dC Pdt ¼ k1C AC P þ k2C TC W (6b)
dC Tdt ¼ k1C AC P k2C TC W k4C 2T (6c)
dC Wdt ¼ k1C AC P k2C TC W (6d)
In Eqs. (6s), concentrations were used instead of activities sinceactivity can be defined as:
ai ¼ g ic i (7)
The coefficientg i canreadilybe combined with thereaction rateconstants in Eqs. (6s). Furthermore, to use concentrations insteadof activities is more practical.
The reaction rate constants, k1, k2, k3 and k4, in Eqs. (6s) weredetermined by applying nonlinear regression analysis in MATLAB7.0 program to the experimental data for each temperature.Results together with their confidence limits are given in Table 3.
3.5. Effect of temperature
Applying the Arrhenius equation to the values in Table 3, the
temperature dependency of the constants as well as theircorresponding activation energies was found to be:
k1 ¼ expð22:17 9667=T Þ L =mol min; DE ¼ 80:37kJ=mol (8a)
Fig. 3. Effect of propylene glycol concentration on the initial reaction rate
(C A0 = 0.96 mol/L, 5 (wt%) catalyst on dry basis, and 0.3 (wt%) inhibitor).
Fig. 4. Effect of hydroxypropyl acrylate concentration on the initial reaction rate
(C A0 = 1.06 mol/L, C P0 = 0.84 mol/L, 5 (wt%) catalyst on dry basis, and 0.3 (wt%)
inhibitor).
Fig. 5. Effect of water concentration on the initial reaction rate ( C A0 = 0.96 mol/L,
C P0 = 0.71 mol/L, 5 (wt%) catalyst on dry basis, and 0.3 (wt%) inhibitor).
M.R. Altıokka, E. O des / Applied Catalysis A: General 362 (2009) 115–120118
8/11/2019 1-s2.0-S0926860X09002981-main.pdf
http://slidepdf.com/reader/full/1-s20-s0926860x09002981-mainpdf 5/6
k2 ¼ expð13:83 7797=T Þ L =molmin; DE ¼ 64:82kJ=mol (8b)
k3 ¼ expð62:63 24; 710=T Þ L =mol min; DE
¼ 205:44kJ=mol (8c)
k4 ¼ expð30:56 12;340=T Þ L =mol min; DE
¼ 102:59kJ=mol (8d)
where T is absolute temperature in K.The equilibrium constant, depending on temperature, for
reaction (5a), can also be calculated from the values of k1 and k2
given in Eqs. (8a) and (8b) respectively, as follows:
K ¼k1
k2¼ expð1870=T þ 8:33Þ (9)
The heat of reaction, assuming independent of temperatureranging from 333 to 358 K, calculated to be
DH ¼ 1:870 8:314 ¼ 15:55kJ=mol (10)
This value is in good agreement with the literature value of 15.5 kJ/mol calculated by the method described in DIPPR 801 [31].
3.6. Comparison of model and experimental data
Using the numerical values of the reaction rate constants givenin Eqs. (8s), Eqs. (6s) was solved simultaneously by applyingRunge–Kutta method for the chosen temperature. Thus concen-
tration–time curves based on the model were obtained undergiven reaction conditions. These curves, together with theexperimental data, are shown in Figs. 6–8.
As seen in Figs. 6–8, there is a reasonably good agreementbetween calculated curves and experimental points.
4. Conclusions
The kinetics of esterification of acrylic acid with propylene
glycol has been studied batchwise in the presence of Amberlyst-15as heterogeneous catalyst. Phenothiazine (0.3 (wt%)) was also usedas inhibitor to reduce the polymerization of acrylic acid andproduct. Taking into account the general esterification reaction as
well as polymerization of acrylic acid and products, the overallreaction mechanism can be given by Eqs. (5s). Four rate constantsin Eqs. (6s) were also determined from the experimental dataapplying MATLAB 7.0 program. Variations of these constants with
temperature are presented in Eqs. (8s). It was also concluded thatthe selectivity of hydroxypropyl acrylate is significantly low at
high conversion levels of acrylic acid. Therefore, to work at low
Table 3
The average values of rate constant with 95% confidence interval at different temperatures (5 (wt%) catalyst on dry basis, and 0.3 (wt%) inhibitor).
T (K) k1 (104 L/mol min) k2 (104 L/mol min) k3 (104 L/mol min) k4 (104 L/mol min)
333 11.21 0.81 0.74 0.56 0.09 0.015 14.95 3.09
348 30.46 3.13 1.60 1.08 2.60 0.92 78.99 5.33
358 88.84 8.13 3.94 1.42 15.48 3.08 196.48 26.16
Fig. 7. Experimental points and calculated curve from Eqs. (6s) and (8s) (T = 348 K,
reactant mol ratio; 1/1, 5 (wt%) catalyst on dry basis, and 0.3 (wt%) inhibitor).
Fig. 6. Experimental points and calculated curve from Eqs. (6s) and (8s) (T = 333 K,
reactant mol ratio; 1/1, 5 (wt%) catalyst on dry basis, and 0.3 (wt%) inhibitor).Fig. 8. Experimental points and calculated curve from Eqs. (6s) and (8s) (T = 358 K,
reactant mol ratio; 1/1, 5 (wt%) catalyst on dry basis, and 0.3 (wt%) inhibitor).
M.R. Altıokka, E. O des / Applied Catalysis A: General 362 (2009) 115–120 119
8/11/2019 1-s2.0-S0926860X09002981-main.pdf
http://slidepdf.com/reader/full/1-s20-s0926860x09002981-mainpdf 6/6
conversion with a proper recycle of unreacted stream, afterproduct separation, is recommended in industrial usage.
References
[1] B. Saha, M. Streat, React. Funct. Polym. 40 (1999) 13.[2] D. Chistova,R. Velichova, W. Loos, E.J.Goethals, F.D.Prez, Polymer 44 (2003) 2255.[3] E. Ruckenstein, Y. Yuan, Polymer 38 (1997) 3855.[4] G. Moad, E. Rizzardo, S.H. Thang, Polymer 49 (2008) 1079.
[5] J. Mueller, U.S. Patent 6,512,059 (2003).[6] G.W. Borden, U.S. Patent 3,943,103 (1976).[7] M.J.Percy,J.I. Amalvy, D.P.Randall,S.P. Armes, S.J.Greaves,J.F. Watts,Langmuir 20
(2004) 2184.[8] C.D. Vo, J. Rosselgong, S.P. Armes, N.C. Billingham, Macromolecules 40 (2007)
7119.[9] The Dow Chemical Company, Product Information, Hydroxypropyl Acrylate.
[10] M.R. Altıokka, A. Cıtak, Appl. Catal. A 239 (2003) 141.[11] M.R. Altıokka, H.L. Hosgun, Ind. Eng. Chem. Res. 46 (2007) 1058.[12] V.C. Malshe, S.B. Chandalia, Chem. Eng. Sci. 32 (1977) 1530.[13] G.A. Chubarov, S.M. Danov, V.I. Logutov, T.N. Obmelyukhina, J. Appl. Chem. USSR
57 (1984) 192.[14] V.A. Fomin, I.V. Etlis, V.I. Kulemin, J. Appl. Chem. USSR 64 (1991) 1811.
[15] P. Dupont, J.C. Vedrine, E. Paumard, G. Hecquet, F. Lefebvre, Appl. Catal. A 129(1995) 217.
[16] X. Chen, Z. Xu, T. Okuhara, Appl. Catal. A 180 (1999) 261.[17] S. Shanmugam, B. Vieswanathan, T.K. Varadarajan, J. Mol. Catal. A. 223 (2004)
143.[18] N. Essayem, V. Martin, A. Riondel, J.C. Vedrine, Appl. Catal. A 326 (2007) 74.[19] H.S. Fogler, Elementsof Chemical Reaction Engineering, Prentice Hall, NewJersey,
1999.[20] R.J. Madon, M. Boudart, Ind. Eng. Chem. Fundam. 21 (1982) 438.[21] P. Delgado, M.T. Sanz, S. Beltran, Chem. Eng. J. 126 (2007) 111.[22] W.T. Liu, Ch.S. Tan, Ind. Eng. Chem. Res. 40 (2001) 3281.
[23] V.J. Cruz, J.F. Izquierdo, F. Cunill, J. Tejero, M. Iborra, C. Fite, R. Bringue, React.Funct. Polym. 67 (2007) 210.[24] M.P. Titus, M. Bausach, J. Tejero, M. Iborra, C. Fite, F. Cunill, J.F. Izquierdo, Appl.
Catal. A 323 (2007) 38.[25] S.H. Ali, A. Tarakmah, S.Q. Merchant, T.Al-Sahhaf, Chem. Eng. Sci. 62 (2007) 3197.[26] W. Mao, X. Wang, H. Wang, H. Chang, X.Zhang, J. Han, Chem. Eng. Proc. 47 (2008)
761.[27] G.D. Yadav, H.B. Kulkarni, React. Funct. Polym. 44 (2000) 153.[28] O. Darge, F.C. Thyrion, J. Chem. Technol. Biotechnol. 58 (1993) 351.[29] S. Dassy, H. Wiame, F.C. Thyrion, J. Chem. Technol. Biotechnol. 59 (1994) 149.[30] O. Levenspiel, Chemical Reaction Engineering, Wiley, Canada, 1972.[31] Design Institute for Physical Properties Research (DIPPR), American Institute of
Chemical Engineers, Project 801, 2006.
M.R. Altıokka, E. O des / Applied Catalysis A: General 362 (2009) 115–120120