Kinetics of non-catalytic and ZnL2-catalyzed esterification of lauric acid with ethanol

Kinetics of non-catalytic and ZnL2-catalyzed esterification of lauric acid with ethanol

Fuel 117 (2014) 125–132 Contents lists available at ScienceDirect Fuel journal homepage: www.elsevier.com/locate/fuel Kinetics of non-catalytic and...

658KB Sizes 10 Downloads 72 Views

Fuel 117 (2014) 125–132

Contents lists available at ScienceDirect

Fuel journal homepage: www.elsevier.com/locate/fuel

Kinetics of non-catalytic and ZnL2-catalyzed esterification of lauric acid with ethanol Eduardo Jose Mendes de Paiva a, Valeria Graeser b, Fernando Wypych a, Marcos L. Corazza b,⇑ a b

CEPESQ, Research Center in Applied Chemistry, Department of Chemistry, Federal University of Paraná, P.O. Box 19081, Curitiba, PR 81531-980, Brazil Department of Chemical Engineering, Federal University of Paraná, CEP 81531-990, Curitiba, PR, Brazil

h i g h l i g h t s  We used zinc laurate (ZnL2) as catalyst to alkyl esters production with ethanol.  Auto-thermal and ZnL2-catalyzed conversion were well modeled.  Higher fatty acid conversions were observed.

a r t i c l e

i n f o

Article history: Received 7 July 2013 Received in revised form 18 August 2013 Accepted 6 September 2013 Available online 21 September 2013 Keywords: Biodiesel Fatty acid esterification Heterogeneous catalyst Thermodynamic analysis Lamellar compounds

a b s t r a c t This work report experimental kinetic data of ethyl esters production from auto-thermal esterification of lauric acid with ethanol and also using zinc laurates (ZnL2) as catalyst. The experiments were performed in a mechanically stirred reactor evaluating the effect of temperature (120–250 °C), ethanol to lauric acid molar ratio (MR) (3:1–9:1) and catalyst amount (2–10 wt%). Chemical equilibrium calculation using a non-ideal mixture approach was applied to perform a thermodynamic analysis of this esterification reaction at the equilibrium conditions. The UNIFAC-LV model was used for the activity coefficient calculations. The results showed that relative high lauric acid conversions can be obtained (around 92% of lauric acids ethyl esters – LAEE) even employing hydrated ethanol. Furthermore, the results showed that the catalyst was able to drive the reaction to the equilibrium in a relative short time (about 100 min). Ó 2013 Published by Elsevier Ltd.

1. Introduction In 1895, Emil Fischer discovered that esters are formed simply by heating a carboxylic acid in alcohol solution containing a small amount of strong acid catalyst [1–3]. Esters have played a significant role in daily living and chemical industry, such as plasticizes, fragrance, adhesive and lubricants [4–9] and recently had been matter of extensive research to production of renewable fuels and high value materials. In this context, biodiesel (alkyl esters mixture) production from free fatty acid is highlighted. As mentioned by Brahmkhatri and Patel [9], traditionally the fatty acid esters are produced using strong acids like sulphuric acid and high cost separation, large energy demand and polluting by-products are intrinsic characteristic of theses process. The development of new technologies enabling to employ waste raw material such fried oils and high fatty acids material sub-products from woody and oil industry could be a possible solution to achieve the energy goal and enhance the food and ⇑ Corresponding author. Tel.: +55 41 3361 3587. E-mail address: [email protected] (M.L. Corazza). 0016-2361/$ - see front matter Ó 2013 Published by Elsevier Ltd. http://dx.doi.org/10.1016/j.fuel.2013.09.016

cosmetics production. However, this route cannot be treated in the conventional way, i.e., alkali-catalyzed homogeneous system, to convert the high content fatty material in free fatty acids. As biodiesel consists of fatty ester of short chain alcohols, triacylglycerols should be first submitted to a pre-treatment (hydrolysis) before submitted to an esterification reaction [10]. The generic esterification reaction of a carboxylic acid with ethanol producing ethyl esters and water is schematically shown in the following equation:

RCOOH þ CH3 CH2 OH ¢ RCOOCH2 CH3 þ H2 O ðFAÞ

ðEtOHÞ

ðFAEEÞ

ð1Þ

ðWÞ

The vast majority of esters can be prepared using esterification reaction in the chemical engineering industry. Esterification has acquired further improvement from engineering side and this mainly depends on the research of esterification kinetics. On the other hand, the need to control chemical reactions at the molecular level, which depends on the catalytic mechanism, is rapidly increasing [8,11]. Zeng and co-workers [8] presented an extensive review about various types of catalyst employed in esterification reactions such as inorganic acid, Lewis acid, metallic compounds, solid acids,

126

E.J.M. de Paiva et al. / Fuel 117 (2014) 125–132

ion-exchange resin and others. The authors discussed objectively and quantitatively the various catalysts available and the mechanism involved and conclude that the most effective way to select them to esterification and process design is studying the intrinsic kinetics. The use of Lewis acid catalyst to (trans)esterification has been proposed by some authors and by our group as well. Macierzanka and Szelag [12] studied the esterification of glycerol in the presence of zinc carboxylates, the authors found that this catalyst has significant influence over monoglycerides (MAGs) conversions and its presence increase the glycerol-fatty acid interfacial area (microemulsions formation). Di Serio et al. [13] studied the influence of metallic stearates and acetates salts in the trans(esterification), the authors conclude that bivalent cations are catalysts for both reactions and its strength is related to the molecular structure and the type of anion, better results are achieved with stearates because their high solubility in the oil phase. Cordeiro et al. [14,15] employed a layered zinc hydroxide nitrate as catalysts to trans(esterification) with me(ethanol), the best results were obtained with esterification of lauric acid with methanol (97.4% FAME at 140 °C, MR 1:4 and 4 wt% of catalyst). Besides, the reuse of this catalyst reached 11 times without any significant activity loss and the authors also found that zinc hydroxide nitrate structure was ‘‘in situ’’ changed into zinc laurate during the first catalyst use. Lisboa et al. [16] also investigated the catalytic activity of various layered laurates (Cu, Mn, La and Ni). The best results were achieved with manganese laurate at mild conditions (90% FAME, 70% FAEE at 140 °C and lauric acid to alcohol molar ratio of 1:6), cupper and lanthanum laurates seems not to be effective to methyl esters production when compared to auto-thermal conversion. In this context, the esterification of lauric acid with ethanol, a greener and renewable reactant compared to methanol, from auto-thermal (non-catalyzed) kinetics and ZnL2-catalyzed (zinc laurate) kinetics reactions were experimentally investigated. In this first approach, lauric acid was chosen as model to be a pretty stable and reliable source of saturated fatty acid, which is very important in kinetic modeling. Furthermore, similar thermodynamics properties of these compounds support that results can be reproducible to other saturated FFA (e.g. Gibbs energy of formation at ideal gas state for palmitic, myristic, stearic and lauric acid are 260.0, 278.0, 243.0 and 293.0, kJ mol1, respectively) [17]. Also, it is worthy to mention that babassu oil, a non-edible oleaginous with about 44% of lauric acid has being studied in Brazil to biodiesel production purposes [18]. In this work, a systematic study of the auto-thermal reaction was carried out before the ZnL2 influence was accounted. Also a thermodynamic framework was surveyed in order to compare experimental and theoretical results of this fatty acid conversion. Thus, a kinetic model able to predict this esterification under different ethanol/lauric acid MR, ethanol assays (hydrated and anhydrous), catalyst amount, auto-thermal conversions, and temperature was surveyed.

computation starts from the time which the desired temperature was reached inside the vessel, which was about 20 min for the R01 reactor and 45 min for the R02 reactor. The same procedure to temperature rising was employed in all experiments and for both reactors. In all experiments the vessel volume was set around 2/3 of the reactor. For each isotherm and sample, the LA conversion was measured. The catalyst was separated by centrifuge method (10,500 rpm for 30 min). After this step, the solid material was separated from the samples and the remaining phase was filtered in a 45 micrometer PTFE filter, after that the excess of ethanol was recovered by rotatory evaporation under reduced atmosphere (450 mmHg at 80 °C). The LAEE conversions were both evaluated by AOCS official method Ca-5a-40 to determination of free fatty acids and proton nuclear magnetic resonance – H1 NMR spectroscopy. The NMR evaluations was taken at 400 MHz using CDCl3 as solvent and 1% TMS as internal standard. A correlation between these methods was recently made by our group [19]. 2.1. Material The esterification kinetics were performed with ethanol (Sigma 99.75 wt% and Vetec 94.8 wt%), lauric acid (C12H24O2 – Sigma, 99 wt%). All chemical were used without further purification and in a reagent grade. Zinc laurate were synthesized through precipitation or metathesis in alcoholic medium starting by neutralization of lauric acid with sodium hydroxide followed by precipitation through slow addition of aqueous solution of zinc salt [16]. The catalyst synthesized was then characterized. X-ray diffraction (XRD) patterns were recorded with a Shimadzu XDR-6000 instrument using Cu Ka radiation (k = 1.5418 Å), dwell time of 2° min1, current of 30 mA and tension of 40 kV. The samples were placed and oriented by hand pressing after filling top-loading aluminum or neutral glass sample holders. The Fourier transform infrared (FTIR) spectra were recorded with a Bio-Rad FTS 3500GX instrument, using approximately 1% of sample in 100 mg of spectroscopic grade KBr, the pellets being pressed at 10 tonnes. The measurements were performed by transmission mode with accumulation of 32 scans and recorded with a nominal resolution of 4 cm1. The catalyst recovering was evaluated by simply weighting the catalyst recovered after thoroughly washing with acetone/hexane (1:1), filtration and drying to the constant mass. Previously to the filtration process, the entire reaction medium was taken and submitted to rotary evaporation under reduced atmosphere. Samples were taken (1 mL) and filtrated with PTFE micro filters (22 and 44 micrometers, respectively). The filtered samples were diluted to 50 mL of a solution of nitric acid 5 M and then injected and analyzed by Thermo Scientific ICP OES spectrometer model ICAP 6000 in order to quantify the zinc metal traces. 2.2. Kinetic modeling

2. Material and methods In this study kinetic experiments at various temperatures (150– 250 °C), ethanol/lauric acid (LA) molar ratio (MR) and catalyst (ZnL2) content were systematically varied to obtain lauric ethyl esters (LAEE). The evaluated parameters are presented and discussed in the next section. All the kinetics experiments were performed in two stainless steel ParrÒ batch reactors (R01: 50 mL, model 4592HP and R02: 300 mL, model 4561) with PID control. The pressure inside the reaction vessel corresponded to the vapor pressure of ethanol at reaction temperature settings. The reaction zero time to kinetics

Initially, non-catalyzed (or auto-thermal) kinetic data were carried out at different MR (3:1 to 9:1), temperatures (120–250 °C) and reaction time varying from 0, 15, 30, 45, 60, 90, 120, 180, 240 until 300 min. After that, a systematic kinetic study was carried out with the reactants in the presence of zinc laurate, changing catalyst concentration (2 and 10 wt% related to lauric acid mass) and temperature (150 °C and 180 °C). In these catalyzed reactions the ethanol to lauric acid molar ratio (MR) was fixed at 8:1, based on our group previous studies [14–16]. The experimental conditions studied in this work are presented in Table 1. In this work, in an attempt to represent the experimental kinetic data obtained from lauric acid esterification with ethanol

127

E.J.M. de Paiva et al. / Fuel 117 (2014) 125–132 Table 1 Experimental conditions used for kinetic measurements of lauric acid esterification with ethanol. Run

T (°C)

MR (mol/mol)a

Water (mol/mol)b

ZnL2 (mol/mol)c

C1 C2 C3 C4 C5 C6 C7* C8 C9R C10* C11 C12R C13R C14 C15

150 200 120 150 180 200 250 180 180 180 180 150 180 180 120

3.00 3.00 6.00 6.00 6.00 6.00 6.00 8.03 8.93 3.00 6.94 7.95 7.95 8.02 8.01

0 0 0 0 0 0 0 1.19 0.06 0.02 0.05 1.18 1.18 1.19 1.18

0 0 0 0 0 0 0 0 0 0 0 17.98E04 44.97E04 9.10E04 9.10E04

FO ¼

k2

ð2Þ

where FA, Et, FAEE and H2O are related one mol of fatty acid (lauric acid), ethanol, ethyl ester (ethyl laurate) and water, respectively. A reversible kinetic model was assumed to this reaction considering a second order reaction of direct and reverse reactions and then the molar rate of lauric acid (dNA/dt) can be written as:

dNA ¼ k1 NA NB þ k2 NC ND dt

ð3Þ

where Ni (mol) is the number of moles of reactants ‘‘i’’ and A represents the fatty acids (FA); B ethanol (Et); C fatty acid ethyl esters (FAEE) and D the water, k1 and k2 are the direct and reverse reaction constants, respectively. The molar balance can be written as function of lauric acid conversion (XA), and then the Eq. (3) can be rewritten by:

dX A k1 NA NB  k2 NC ND ¼ dt N A0

ð4Þ

The number of moles of each component ‘‘i’’ in the reaction can be written as a function of the reaction coordinate (n), as presented in the following equation:

Ni ¼ Ni0 þ mi  n

ð5Þ

where mi is stoichiometric coefficient of each reactant ‘‘i’’ (see Eq. (1)), and Ni0 is the initial number of moles of component ‘‘i’’ in the reaction. The set of equations generated by Eqs. (4) and (5) was solved using MatlabÒ ‘‘ode23’’ subroutine. In this work the Arrhenius’ equation was used to calculate the kinetic constants considering the direct (k1) and reversible (k2) reaction equations (Eq. (6)), as follows:

  Eaj kj ¼ k0j exp RT

ð7Þ

Calc In Eq. (6), X Exp A;i and X A;i are the experimental and calculated values of lauric acid conversion (%) and nobs is related to experimental set. It is worth to mention that the kinetic parameters were obtained from a global estimation by regressing of all experimental kinetic data for all conditions presented in Table 1, except the kinetic set for run C9 and C7, which were missed during the parameter estimation and were used to validate the model. On the whole, 14 experimental data sets using five temperatures, comprising 112 data points of lauric acid conversion were used for the kinetic model fitting.

related to non-catalytic and ZnL2-catalyzed systems the following stoichiometric equation was considered: k1

nobs X Calc 2 ðX Exp A;i  X A;i Þ i

Runs C10 to C14 were performed in 300 mL vessel with mechanical stirring at 350 rpm. Runs C1 to C9 were performed in 50 mL vessel with mechanical stirring at 260 rpm. * Not used in the model fitting; Superscript R is for runs performed in replicate (duplicate). a mol of ethanol by mol of lauric acid. b mol of water by mol of ethanol plus mol of lauric acid. c mol of catalyst (molar mass of 439.38 g gmol1) by mol of ethanol plus mol of lauric acid.

FA þ Et ¢ FAEE þ H2 O

obtained in this work with the proposed kinetic model (Eqs. (4)– (6)). The objective function (OF) presented in Eq. (7) was minimized by using the fmincon subroutine from MatlabÒ.

2.3. Chemical equilibrium calculations In order to get a background to analyses and compare the kinetics at the stationary condition, i.e. reaching the thermodynamic equilibrium conversions, and also measuring the effects of the catalyst on the considered esterification of lauric acid with ethanol, thermodynamics calculations were performed using the approach described in our previous work [20]. Briefly, the reaction equilibrium was calculated considering a non-ideal mixture in a liquid phase and the activity coefficients were determined by UNIFACLV model [21]. Thermodynamic parameters for pure components were obtained from the literature or estimated from group contribution methods, as presented in Tables 2 and 3. 3. Results and discussion As early mentioned, kinetic experiments were performed using two different reactor models (R01 and R02 as early presented), then, in order to check the reproducibility of kinetic results, the same experimental condition was used with both reactor systems and the kinetics were compared. In Fig. 1(A) the experimental values of lauric acid conversion for non-catalyzed lauric acid esterification obtained using the reactor R01 are plotted against the conversion values obtained from the reactor R02 at same condition (run C9 in Table 1). It can be seen from Fig. 1(A) that lauric acid conversion present similar values for all experimental points obtained and as one can see are under 95% of statistical confidence (dashed lines in Fig. 1(A)). Also, two other kinetics were performed in duplicate (refers to Table 1). Similar statistical trends were obtained with kinetics results for runs C13 and C12, as it can be observed in Fig. 1(B) and (C), respectively. From results presented in Fig. 1, it is showed that the experimental procedure and apparatus are reproducible and reliable. No significant differences were observed when different reactor-types were used, i.e., the kinetic results obtained from two different size reactors were statistically identical.

Table 2 Thermodynamics properties of pure compounds formation at ideal gas state (298.15 K, 1 atm). Component Lauric acid Ethyl laurate Ethanol Water

ð6Þ

The kinetic parameters (pre-exponentials, k0j, reduced activation energy, Eaj/R and j represent the direct and reverse reaction) were obtained by correlating the kinetic experimental data

a b

DIPPR [17]. Constantinou and Gani [22].

DG0f ;i (kJ/mol) a

293.1 239.77a 167.8b 228.6a

DH0f ;i (kJ/mol) 640.0a 650.5a 235.1b 241.8a

128

E.J.M. de Paiva et al. / Fuel 117 (2014) 125–132

Table 3 Vapor pressure and heat capacities of pure components. Component

Model

A

B

C

D

E

T range (K)

Vapor pressure Lauric acid Ethyl laurate Ethanol Water

b a c c

201.56 7.20 8.11 7.764

20454.0 2399.38 1592.86 1.4583

24.33 194.45 226.18 2.775

8.05x1018 – – 1.23

6 – – –

316.9–743.0 278.1–712.0 50–1000 50–1000

Heat capacity Lauric acid⁄ Ethyl laurate⁄ Ethanol Water

d d e e

3.42 15.57 19.85 32.218

1.1946 1.28 20.95  102 0.192  102

0.7  103 0.6  103 10.37  105 1.055  105

2.0  107 1.0  107 20.04  109 3.593  109

273–1500 273–1000 273–1000 273–1800

a: Ceriani and Meirelles [23]. b: DIPPR [19]. c: Poling et al. [24]. 1 ðJ mol K1 Þ ¼ ai þ bi T þ ci T þ di T 3 and it was fitted using values obtained from group contribution method of Joback and Reid [26]. d: Equation used was Cpig i e: Sandler [25].

The kinetic results considering the catalyzed and non-catalyzed reaction of lauric acid esterification with ethanol at different temperatures, MR and catalyst amount, as well the model fitting are presented in the next sections. 3.1. Catalyst characterization The X-ray powder diffraction patterns show that compounds have a layered structure, as can be seen from the typical basal peak sequence between 3° and 19° in Fig. 2. The stacking direction is along the ‘‘a’’ axis (h00 direction). The basal spacing calculated for the synthesized zinc laurate was de 29.47 Å. The others peaks from 19° refers to perpendicular and diagonals planes to the lamella. The FTIR spectra of the compounds present vibration modes characteristic of saturated fatty acid salts and of the way the carboxylate groups are connected to the metals building the layers. From Fig. 3, the coordination of the laurate to the metals can be verified by the presence of stretching bands between 430 and 500 cm1, which can be attributed to M–O bonds. The difference between bands at region 1538 and 1338 cm1 (140 cm1), related to asymmetric and symmetric stretching of carboxyl groups, suggest that this compound is structured in lamellas. The asymmetric and symmetric stretching of the methyl groups can be seen between 2951 and 2847 cm1, respectively. The methylenic asymmetric and symmetric stretching (–CH2) can be observed at 2916, 1463 and 721 cm1, respectively. Also, a bunch of peaks can be seen between 1390 and 1100 cm1. In many cases, this is taken as fingerprint of this class of compounds. 3.2. Kinetic of auto-thermal esterification of lauric acid with ethanol The kinetic parameters obtained from correlating the proposed model for auto-thermal esterification (catalyst absence) of lauric acid are presented in Table 4. The kinetics, regarding the conditions C1–C6, C8, C9 and C11, in which ZnL2 amount was zero, were employed to the model fitting. The root mean square deviation (RMSD) obtained considering conditions C1 to C11 was RMSD 3.33% in terms of fatty acid conversion (%). From the low RMSD value it can be observed that the proposed auto-thermal reaction kinetic model is able to successfully predict experimental data. Similar results were reported by Jiputti and co-worker [27], in which the authors investigated the kinetic of dodecanoic acid (lauric acid) with 2-ethylexanol in the presence of sulfated zirconia at temperature range of 60 °C 333 to 170 °C and reported an activation energy of 55.5 kJ/mol. In Fig. 4, the kinetic results at different conditions of ethanol/ lauric acid MR are showed. In Fig. 4(A) is depicted the kinetic of

3:1 ethanol/lauric acid MR, at three temperatures (C1, C2 and C10). In Fig. 4(B) the kinetic results at 6:1 are showed. It can be observed from this figure that the model was able to predict kinetic behavior of the reaction for all experimental conditions. It can be seen from Fig. 4(A) that employment of MR 3:1 and higher temperatures leads to better LAEE conversions reaching values around 65% at 200 °C in relatively short time (2 h). When the MR of 6:1 was used (C3 to C6) the reaction rates were increased and equilibrium conversion of 80% at 200 °C and 75% at 180 °C were observed (Fig. 4(B)). At 250 °C it can be observed that conversion around 85% was reached in a relative short time (60 min). In addition, in the Fig. 5 (180 °C, different RM and water content) shows that an increase in the MR (C8 and C9) leads to an enhance in the equilibrium conversion and an increasing in the initial water content (C8) leads to a decreasing in the equilibrium conversion, as expected to this kind of reactions. It is worthy to emphasize that at mild temperatures conditions (120 and 150 °C) a long period of time it is required until the equilibrium can be reached and at higher temperatures conditions (180–250 °C) the equilibrium is reached within approximately 1 or 2 h. Moreover, model shows that reaction rate increase with the temperature reinforcing the endothermic characteristic of this reaction. Another important finding that must be pointed out is that the water content induces fatty acid formation. As example, runs performed with high water contents (94–95 wt%.) had lower conversions.

3.3. ZnL2-catalyzed esterification of lauric acid with ethanol After the auto-thermal modeling the ZnL2 influence was studied. The kinetic parameters for ZnL2-catalyzed reaction (k1,cat and k2,cat) were obtained by using the same computational calculation employed with non-catalyzed reaction. However, the kinetic constants were modified by using the Eq. (8), that consider the catalyst to lauric acid plus ethanol molar ratio (ZnL2, mol/mol), as presented in Table 1. The kj,cat were obtained using the Arrhenius equation (Eq. (6)). The obtained parameters are presented in Table 4. The RMSD obtained were 3.09% in terms of lauric acid conversion. The robustness of the model was checked by its ability to predict LAEE conversions at different reaction vessels, temperature, MR and amount of catalyst.

kj ¼ kj;cat ð1000  ZnL2 Þ

ð8Þ

It can be seen from the experimental data that the ZnL2 catalyst presence in the reaction medium affect the equilibrium conversion, besides increasing the reaction rate as expected. This behavior can be clearly seemed from data showed in Fig. 6.

129

E.J.M. de Paiva et al. / Fuel 117 (2014) 125–132

(A)

80

30000

70 60 50 40 30

25000 20000 15000

10

5000

10

20

30

40

50

60

70

80

0

90

5

Conversion (%) (run C9 Reactor R01)

(B)

90

10

15

20

25

30

35

40

2θ (Theta) Fig. 2. X-ray powder diffraction patterns of the layered zinc laurate.

85 130

80

Zinc laurate

120

75

Fingerprint (all trans)

55

60

65

70

75

80

85

90

95

Conversion (%) (run C13) 90

60 50

(C)

80

ν MO

1463 1398

70

-

νsCOO

50 50

80

νs CH2

55

90

2847

60

100

2916

65

445

110

70

Transmitancy (%)

Conversion (%) (run C13 duplicate)

95

Conversion (%) (run C12 duplicate)

2θ = 18,06° d = 29,47 A n=6

2θ = 6,06° d = 29,17 A n=2

10000

20

0

2θ = 3,06° d = 28,87 A n=1

35000

Intensity (cps)

Conversion (%) (run C9 Reactor R02)

90

40 30 3500

70

3000

2500

2000

1500

1000

500

-1

Wave Number (cm ) 60

Fig. 3. FTIR spectra of the layered zinc laurate.

50 40

Table 4 Kinetics parameters obtained from the auto-catalytic esterification of lauric acid with ethanol.

30 20

Parameters

10 0

Kinetic constants 0

10

20

30

40

50

60

70

80

90

Conversion (%) (run C12) Fig. 1. Results (mean and standard deviation) of lauric acid conversion considering different runs from Table 1: (A) auto-thermal conversions evaluated in the 50 mL and 300 mL batch reactors (run C9); (B) run C13; and (C) run C12.

As mentioned, it can be observed from the Fig. 6 that ZnL2 is able to displace the reaction equilibrium leading to conversions gain up to 40 p.p. compared to auto-thermal reactions at same conditions (run C8 in Fig. 5 and C2 in Fig. 4(A)). At 120 min, time stated by some authors [14–16] to reaction completeness, the gain is about 30%p.p. which is in good agreement with the results presented here. We believe that this equilibrium displacement in the reaction system is linked to the surfactant and high hydrophobicity of this catalyst. Thus, the surface energy of the whole system is greatly affected by ZnL2 presence. This assumption is supported by similar reports [12,13,28,29] evidencing the changes in the surface energy of the whole reaction system and microemulsions formation. A similar results were already observed by Bresler and Samsonov [30], however in that early study the authors performed esterification only with films.

k0 (mol1 min1)

Ea/R (K)

Non-catalyzed k1 k2

447.40 9.42

5567.99 3562.42

ZnL2-catalyzed k1,cat k2,cat

118288.11 15141.89

8174.33 7734.66

The root mean square deviation (RMSD) obtained was 3.33% (in % FFA conversion) for non-catalyzed and 2.73% for ZnL2-catalyzed.

In a general way, it must be pointed out that of the majority of the other studies involving similar FFA have been developed with methanol. For comparison purposes, Bassan et al. [31] investigated the esterification of FFA with various alcohols under the presence of Niobium phosphate as catalyst, these authors reported that the best results with lauric acid and methanol was 80% (MR 1:10, 10 wt% of catalyst after 4 h of reaction). Carmo Jr. and co-workers [32] carried out esterification of palmitic acid with Al-MCM-41 (0.6 wt% catalyst, MR 1:60 acid/alcohol, 130 °C and 2 h) the best results employing ethanol was under 70%. Rattanaphra et al. [33] performed the esterification of myristic acid with methanol in the presence of sulfated zirconia, conversions about

130

E.J.M. de Paiva et al. / Fuel 117 (2014) 125–132 100

80

(A)

70

90 80

Conversion (%)

Conversion (%)

60 50 40 30

60 50 40 30

Runs C1 C2 C10

20 10

70

0

50

100

150

200

250

Runs

20

C14 C13 C12

10 0

300

0

50

100

time (min) 100

200

250

300

350

400

time (min)

(B)

90

150

Fig. 6. Kinetics of ZnL2-catalyzed lauric acid esterification with ethanol at different conditions.

70 3.4

60 3.3

50

1.4 k 1,cat / k 2,cat k 1 / k2

1.3 1.2

3.2

40

1.1

20 10 0

3.1

Runs C7 C6 C5 C4 C3 0

50

100

150

200

250

300

time (min)

1.0

3.0

0.9

2.9

0.8

k1 / k2

30

k1,cat / k2,cat

Conversion (%)

80

0.7

2.8

0.6 2.7

Fig. 4. Experimental and model adjusted conversions to lauric acid esterification at different temperatures and different MR of 3:1 (A) and 6:1 (B) (non-catalyzed).

0.5 2.6 2.5 100

90

0.4 120

140

160

180

200

220

240

260

0.3 280

o

Temperature ( C)

80

Fig. 7. Kinetic constant evaluation for the model fitted in this work (non-catalyzed reaction).

Conversion (%)

70 60 50 40 30 Runs C8 C9 C11

20 10 0

0

50

100

150

200

250

300

350

400

time (min) Fig. 5. Experimental and model adjusted conversions to lauric acid esterification at different MR and 180 °C and varying water initial content.

95% (3 wt% of catalyst, MR 1:20 acid/alcohol, 170 °C) was achieved with the experiments. This papers show that similar and better results can be obtained with saturated FFA, even using hydrated ethanol in less than 2 h. At this point, similar considerations made in the Section 3.2 can also be made here: (i) the model suitable fit most of the experimental sets (lower RMSD value); (ii) adjusted model is in good agreement with experimental data; (iii) higher temperatures lead to higher conversions. Remarkably, ZnL2 affects the lauric acid conversions into ethyl ester (ethyl laurate) in the beginnings of the reaction and after 100 and 200 min (180 °C, 44.97  104 mol/

mol and 9.10  104 mol/mol of ZnL2, respectively). At 150 °C the reaction rate as the equilibrium conversion slowly changed when compared with two others conditions. However, the conversion at this condition is improved when compared with the non-catalyzed one. It must be pointed out that Zinc laurate showed a great tolerance to water content (5.2% of water). It can be an important issue to the process development of esterification reactions, mainly in the biodiesel industry because this catalyst is quite cheap and could be easily recovered. Reuse tests was already performed by our group [14,16] and reuse cycle reaches 11 times without significant catalyst activity loss [15]. Besides, Reinoso and co-workers [34] also studied the ZnL2 reuse cycles in the transesterification at 100 °C and also concluded that were no important changes in the activity of the catalyst, supported by their X-ray and FTIR studies. Further, the catalyst recovering and leaching will be discussed in the Section 3.4. In Fig. 7, the kinetic constants ratios of catalyzed (k1,cat/k2,cat) and non-catalyzed (k1/k2) are presented. Considering the non-catalyzed reaction it can be seen that reversible reaction is more favorable that the direct. From data presented in Table 4, one can argue that both k1 and k2 are higher with higher temperatures. Also, there is threshold point (nearly 240 °C) after which the direct reaction becomes more favorable than inverse but at this high temperature the products decomposition could take place and additional studies are required.

131

E.J.M. de Paiva et al. / Fuel 117 (2014) 125–132

3.4. Chemical equilibrium of esterification of lauric acid with ethanol As described in the Section 2, calculations were performed to predict the equilibrium conditions to lauric acid esterification with ethanol. The conversions obtained at equilibrium as a function of temperature ethanol/lauric acid MR (related to 1 mol of lauric acids) are presented in Fig. 9. It can be observed from Fig. 9, at constant temperature conditions, that the ethanol to lauric acid molar ration has great influence on the FFA conversion as expected, mainly at values of MR up to 8:1. Besides, the MR effect on thermodynamic equilibrium conversion is more accentuate according to the temperature is increased. Predicted values of equilibrium conversion are in agreement with the experimental lauric acid conversions or even more reliable (by using extrapolated values by means of the kinetic model solved at stationary state). This statement is evident considering the MR of 9:1, in which the thermodynamic model suitable describe the equilibrium condition for this reaction. Experimental results with MR of 9:1 are around 83% and UNIFAC-LV predicts the equilibrium conversion at the same condition with only 3% of deviation. Moreover, others qualitative information can be draw from Fig. 9, considering UNIFAC-LV model, the best MR to be employed is around 9:1 of ethanol to lauric acid molar ratio and the endotermic characteristic of the reaction shows that higher temperatures leads to higher conversions. It must be emphasized that the equilibrium conversions, using the UNIFAC-LV model to correct the non-ideality of liquid phase, can be a useful tool to drive the analysis, design and optimization of the lauric acid esterification, as well to predict the theoretical conversions before a catalyst could be tested. 3.5. Catalyst leaching and recovering The results presented in the Table 5 shows that zinc losses are under 0.00035%. The mass of catalyst recovered in the first test was 94% of the initial mass. Probably, zinc is leached in oxide form

100 Run C4 ( , exp; Run C12 ( , exp; Run C3 ( ,exp ; Run C15 ( , exp;

90 80

Model) Model) Model) Model)

Conversion (%)

70 60 50 40 30 20

100 250 oC

90

Equilibrium conversion (%)

The Fig. 8 presents a comparison between kinetics behavior at lower temperatures (120 °C and 150 °C), experimental and predicted data using the fitted kinetic model are plotted. It is important to mention that the experimental kinetic ZnL2-catalyzed data obtained at 120 °C was not used in the kinetic parameter adjustment. It can be seen from this figure that at temperatures below 150 °C the catalysts not affect the reaction rate. Probably, is due to the fact that its melting point is around 125–133 °C [14,16].

200 oC 180 oC 150 oC

80

120 oC

70 60 50 40 30 20

0

2

4

6

8

10

12

14

Ethanol to lauric acid molar ratio (MR) Fig. 9. Simulated values of equilibrium conversion of lauric acid esterification with ethanol at different temperatures. Calculated values using UNIFAC-LV model for activities coefficients of components in liquid phase.

Table 5 Zinc loss evaluated through atomic absorption.

Filtration 22 lm Filtration 45 lm

Sample mass (mg)

Zinc loss (mg)

Zinc loss (%)

503 475

0.001520 0.001645

0.00030 0.00035

(ZnO) and filters tested are able to retain catalyst particles ranging from 22 to 45 lm. These tests suggest that a filtration would be able to efficiently recover the catalyst in an industrial scale. Moreover, as mentioned before, this catalyst can be recycled to reactor without further treatment [15,16]. 4. Conclusions This work reports for the first time a systematic kinetic study of non-catalyzed and ZnL2-catalyzed lauric acid esterification with ethanol. Appreciable reaction yields of lauric acid ethyl esters were obtained in absence of catalyst (auto-thermal reaction) at 180 °C, 8:1 ethanol/lauric acid MR and reaction time ranging from 60 to 120 min (around 80%). As expected, there is a strong auto-catalytic behavior from esterification reactions with ethanol, blank reactions show that high conversions can be obtained employing higher temperatures and also there is a strong reversible reaction (hydrolysis). However, the energy costs and products thermal stability must be evaluated. On the other hand, the reaction rate and the equilibrium conversion are enhanced by the presence of ZnL2 (around 92% of conversion at 180 °C, 8:1 of ethanol to lauric acid molar ratio, 5% water content related to ethanol (m/m) and 110 min of reaction). Zinc laurate has showed a great tolerance to water content (5.2% of water), this catalyst is quite cheap and could be easily recovered. Besides, conversions obtained with ethanol were similar to those reported with methanol. The kinetic approach employed showed satisfactory agreement between experimental data and model used thus allowing a better understanding of reaction kinetics, pointing a promising route to be further investigated and improved to use in process optimization for biodiesel production via fatty acid esterification.

10 0 0

50

100

150

200

250

300

time (min) Fig. 8. Comparison between catalyzed and non-catalyzed kinetics at lower temperatures.

Acknowledgements The authors are grateful to MEC/Reuni, CAPES, CNPq and Fundação Araucária-Paraná (Brazilian governmental agencies) for the financial support and scholarships.

132

E.J.M. de Paiva et al. / Fuel 117 (2014) 125–132

References [1] Zhang HL, Feng GY, Wei TJ. New development and kinetics of esterification. Henan Chem Ind 1995;2:5–8. [2] Ronnback R, Salmi T, Vuori A, Haario H, Lehtonen J, Sundqvist A, et al. Development of a kinetic model for the esterification of acetic acid with methanol in the presence of a homogeneous acid catalyst. Chem Eng Sci 1997;525:3369–81. [3] Otera J, Nishikido J. Esterification: methods, reactions and applications. Vch Pub.; 2009. ISBN 3527322892. [4] Joseph T, Sahoo S, Halligudi SB. Bronsted acidic ionic fluids: a green, efficient and reusable catalyst system and reaction medium for Fischer esterification. J Mol Catal A: Chem 2005;234:107–10. [5] Mbaraka IK, Shanks BH. Conversion of oils and fats using advanced mesoporous heterogeneous catalysts. JAOCS 2006;83:79–91. [6] Krause P, Hilterhaus L, Fieg G, Liese A, Bornscheuer U. Chemically and enzymatically catalyzed synthesis of C6–C10 alkyl benzoates. Eur J Lipid Sci Technol 2009;111:194–201. [7] Martinez M, Oliveros RN, Aracil J. Synthesis of biosurfactants: Enzymatic esterification of diglycerol and oleic acid. 1. Kinetic modeling. Ind Eng Chem Res 2011;50:6609–14. [8] Zeng Z, Cui L, Xue W, Chen J, Che Y. Recent developments on the mechanism and kinetics of esterification reaction promoted by various catalysts. In: Patel Vivek, editor. Chemical kinetics. InTech; 2012. [9] Brahmkhatri V, Patel A. Esterification of lauric acid with butanol-1 over H3PW12O40 supported on MCM-41. Fuel 2012;102:72–7. [10] Tesser R, Casale L, Verde D, Di Serio M, Santacesaria E. Kinetics of free fatty acids esterification: batch and loop reactor modeling. Chem Eng J 2009;154:25–33. [11] Salciccioli M, Stamatakis M, Caratzoulas S, Vlachos DG. A review of multiscale modeling of metal-catalyzed reactions: mechanism development for complexity and emergent behavior. Chem Eng Sci 2011;66:4319–55. [12] Macierzanka A, Szelag H. Esterification kinetics of glycerol with fatty acids in the presence of zinc carboxylates: preparation of modified acylglycerol. Ind Eng Chem Res 2004;43:7744–53. [13] Di Serio M, Tesser R, Dimiccoli M, Cammarota F, Nastasi M, Santacesaria E. Synthesis of biodiesel via homogeneous Lewis acid catalyst. J Mol Catal A: Chem 2005;239:111–5. [14] Cordeiro CS, Arizaga GGC, Ramos LP, Wypych F. A new zinc hydroxide nitrate heterogeneous catalyst for the esterification of free fatty acids and the transesterification of vegetable oils. Catal Commun 2008;9:2140–3. [15] Cordeiro CS. Lamellar compounds as heterogeneous catalyst to transesterification reactions with methanol and ethanol. Doctoral Thesis – Universidade Federal do Paraná/Brasil, 2008 [in portuguese]. [16] Lisboa FS, Gardolinski JEFC, Cordeiro CS, Wypych F. Layred metal laurates as active catalysts in the methyl/ethyl esterification reactions of lauric acid. J Braz Chem Soc 2012;23:46–56.

[17] DIPPRÒ. Information and Data Evaluation Manager, Public Version 1.2.0, 2000. [18] Paiva EJM, Silva MLCP, Barboza JCS, Oliveira PC, Castro HF, Giordani DS. Nonedible babassu oil as a new source for energy production – a feasibility transesterification survey assisted by ultrasound. Ultrason Sonochem 2013;20:833–8. [19] Zatta L, Nepel A, Barison A, Wypych F. Modified montmorillonite as a heterogeneous catalyst in (m)ethyl esterification reaction of lauric acid. Quim Nova 2012;35:1711–8 [in Portuguese]. [20] Graeser V. Esterificação não-catalítica de ácido laurico etanol anidro: cinética e conversão de equilíbrio. M.Sc. Thesis. Department of Chemical Engineering, Federal University of Paraná, Curitiba-PR, Brazil, 2013 [in Portuguese]. [21] Fredenslund A, Jones RL, Prausnitz JM. Group-contribution estimation of activity coefficients in non-ideal liquid mixtures. AIChE J 1978;21:1086–99. [22] Constantinou L, Gani R. New group contribution method for estimating properties of pure compounds. AIChE J 1994;40:1697–710. [23] Ceriani R, Meirelles AJA. Predicting vapor–liquid equilibria of fatty systems. Fluid Phase Equilib 2004;215:227–36. [24] Poling BE, Prausnitz JM, O’Connell PJ. The properties of gases and liquids. 5th ed. New York, USA: McGraw-Hill; 2001. [25] Sandler SI. Chemical, biochemical, and engineering thermodynamics. 4th ed. John Wiley and Sons; 2006. [26] Joback KG, Reid RC. Estimation of pure-component properties from group contributions. Chem Eng Commun 1987;57:233–43. [27] Jitputti J, Kitiyanan B, Rangsunvigit P, Bunyakiat K, Attanatho L, Jenvanitpanjakul P. Transesterification of crude palm kernel oil and crude coconut oil by different solid catalysts. Chem Eng J 2006;116:61–6. [28] Osipow LI, Rosenblatt W. Micro-emulsion process for the preparation of sucrose esters. J Am Oil Chem Soc 1967;44:307–9. [29] Kaufman VR, Garti N. Organic reactions in emulsions: preparation of glycerol and polyglycerol esters of fatty acids by transesterification reaction. J Am Oil Chem Soc 1982;59:471–4. [30] Bresler SE, Samsonov GV. Displacement of chemical equilibrium in a surface layer. Institute of Macromolecular Compounds – USSR Academy of Sciences, 1953. [31] Bassan IAL, Nascimento DR, San Gil RAS, da Silva MIP, Moreira CR, Gonzalez WA, et al. Fuel Proc Technol 2013;106:619–24. [32] Carmo Jr AC, de Souza LKC, da Costa CEF, Longo E, Zamian JR, da Rocha Filho GN. Production of biodiesel by esterification of palmitic acid over mesoporous aluminosilicate Al-MCM-41. Fuel 2010;88:461–8. [33] Rattanaphra D, Harvey V, Thanapimmetha A, Srinophakun P. Kinetic of myristic acid esterification with methanol in the presence of triglycerides over sulfated zirconia. Renew Energy 2011;36:2679–86. [34] Reinoso DM, Damiani DE, Tonetto GM. Zinc carboxylic salts used as catalyst in the biodiesel synthesis by esterification and transesterification: study of the stability in the reaction medium. Applied Catalysis A: General 2013;374:221–7.