Available online at www.sciencedirect.com
Fuel 87 (2008) 2286–2295 www.fuelfirst.com
Transesterification of Jatropha curcas oil glycerides: Theoretical and experimental studies of biodiesel reaction Neyda C. Om Tapanes a, Donato A. Gomes Aranda a,*, Jose´ W. de Mesquita Carneiro b, Octavio A. Ceva Antunes c a
Laborato´rio GREENTEC, Escola de Quı´mica, Universidade Federal do Rio de Janeiro, Av. Brigadeiro Trompowsky s/n, Centro de Tecnologia E-211, Ilha do Funda˜o, Rio de Janeiro, CEP 21.949-900, Brazil b Departamento de Quı´mica Inorgaˆnica, Instituto de Quı´mica, Universidade Federal Fluminense, Outeiro de Sa˜o Joa˜o Batista, s/n, Nitero´i, Rio de Janeiro, CEP 24020-150, Brazil c Departamento de Quı´mica Inorgaˆnica, Instituto de Quı´mica, Universidade Federal de Rio do Janeiro, Av. Brigadeiro Trompowsky s/n, Centro de Tecnologia E-637, Ilha do Funda˜o, Rio de Janeiro, CEP 21.949-900, Brazil Received 23 October 2006; received in revised form 22 November 2007; accepted 14 December 2007 Available online 14 January 2008
Abstract Vegetal oil, also known as triglycerides, is a mixture of fatty acid triesters of glycerol. In the triglycerides alkyl chains of Jatropha curcas oil, predominate the palmitic, oleic and linoleic fatty acids. The process usually used to convert these triglycerides to biodiesel is called transesterification. The overall process is a sequence of three equivalent, consecutive and reversible reactions, in which diand monoglycerides are formed as intermediates. Semi-empirical AM1 molecular orbital calculations were used to investigate the reaction pathways of base-catalyzed transesterification of glycerides of palmitic, oleic and linoleic acid. The most probable pathway and the rate determining-step of the reactions were estimated from the molecular orbital calculations. Our results suggest the formation of only one tetrahedral intermediate, which in a subsequent step rearranges to form the products. The rate determining-step is the break of this tetrahedral intermediate. Ó 2007 Elsevier Ltd. All rights reserved. Keywords: Transesterification; Semi-empirical calculations; Reaction mechanism
1. Introduction Fatty acid methyl or ethyl esters derived from renewable sources, such as vegetable oils, have gained importance as an alternative fuel for diesel engines. Edible oils such as soybean oil in USA, rapeseed oil in Europe and palm oil in countries with tropical climate, such as Malaysia, are being used for the production of biodiesel to their compression ignition engines. In other countries, the use of edible oils for engine fuel is not usual; however, there are several non-edible oil seed species which could be utilized as a source for oil production. Among these, Jatropha curcas *
Corresponding author. Tel.: +55 21 2562 7657; fax: +55 21 2562 7567. E-mail address:
[email protected] (D.A. Gomes Aranda).
0016-2361/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.fuel.2007.12.006
is a multipurpose species with many attributes and considerable potential. The oil from the seeds is potentially the most valuable end product, with properties like: low acidity, good oxidation stability as compared to soybean oil, low viscosity as compared to castor oil and better cold properties as compared to palm oil. In addition, viscosity, free fatty acids and density of the oil and the biodiesel are stable within the period of storage [1]. Jatropha curcas is a drought-resistant tree belonging to the Euphorbiaceae family, which is cultivated in Central and South America, South-east Asia, India and Africa. This highly drought-resistant species is adapted to arid and semi-arid conditions. It grows almost anywhere, even on gravelly, sandy and saline soils and is often used for erosion control [2].
N.C. Om Tapanes et al. / Fuel 87 (2008) 2286–2295
The fatty acid profile of Jatropha curcas oil, determined by chromatography analysis, is listed in Table 1 [2]. Typically 1% of the vegetable oils are unsaponifiable compounds (carotenoids, phospholipids, tocopherols or tocotrienols and oxidation products). The commonly used method for production of biodiesel is the transesterification of vegetable oils. Transesterification, also called alcoholysis, is the reaction of triglycerides with alcohols to produce for examples methyl or ethyl esters and glycerol as a by-product. A catalyst is usually used to improve the reaction rate and yield. The reaction requires excess of alcohols to increase the efficiency of the transesterification process [3]. The transesterification reaction is represented by the general equation show in Scheme 1. It consists in three equivalent, consecutive and reversible reactions. The triglyceride is converted stepwise to diglyceride, monoglyceride and finally glycerol. At each reaction step, one molecule of methyl or ethyl ester is produced for each molecule of methanol or ethanol consumed. The three stages of the transesterification reaction are indicated in Scheme 2. The base-catalyzed transesterification has a long story of development. Biodiesel fuel produced by this method is in the market in some countries such as North America, Brazil and mainly in some European countries like Germany and France. However, although base-catalyzed transesteri-
Table 1 Fatty acid profile of Jatropha oil Fatty acid profile
Content % (m/m)
Myristic acid Palmitic acid Palmitoleic acid Stearic acid Oleic acid Linoleic acid Linolenic acid Arachidic acid Gadoleic acid
0.38 16.0 max 1–3.5 6–7.0 42–43.5 33–34.4 >0.80 0.20 0.12
Scheme 1. General equation for transesterification of triglycerides.
2287
fication represents the best alternative to produce biodiesel, it still has some disadvantages. Low free fatty acid content and anhydrous reagents are required due to the saponification possibility. This soap formation lowers the ester yields and can hinder the stages of separation and purification of ester and glycerol as well as the washing stage. Additional disadvantages are that the catalyst recovery process is slow and expensive, increasing the operating costs. Therefore, the process still needs to be optimized. In order to understand and control the transesterification reaction it is necessary to know the reaction mechanism. The several steps of the mechanism of the transesterification reaction are poorly understood. In particular, doubts still exist about the pathway of the process involving formation and breaking of the tetrahedral intermediate. Two types of mechanisms describing the base-catalyzed transesterification reactions have been proposed by various research groups: The first mechanism proposes the formation of two tetrahedral intermediates [3,4]. Other mechanism suggests the formation of only one tetrahedral intermediate which, in a subsequent step, rearranges to form the products [5,6]. However, the lack of detailed information on the actual molecular species, the rate determining-step in the overall reaction mechanism, and the configuration of the transition state complex have severely hampered the quantitative understanding of the reaction kinetics. Computational chemistry methodologies have been used as a powerful tool to study the mechanisms and kinetics of several chemical reactions. Therefore, by means of theoretical calculations it may be possible to better understand the mechanism and the kinetics of the transesterification reactions, with the objective of relating the computer simulations with the results obtained experimentally [7–10]. Such a detailed knowledge would provide valuable guidelines for a deeper understanding of the role played by the catalyst in the formation and breaking of the tetrahedral intermediates in the base-catalyzed transesterification. The purpose of this paper is to study the reaction mechanism of base-catalyzed transesterification of the glycerides of the Jatropha curcas oil, to provide relevant results of kinetic properties, rate determining-step and catalytic effects, to define the catalyst and other experimental conditions that allow maximizing the reaction yield. The results obtained in the computational simulations are compared with experimental Jatropha curcas oil transesterification data.
2. Materials and methods
Scheme 2. Stages of transesterification of triglycerides.
The procedures and methods used in the quantum studies and in the transesterification experiments are given in this section. The methods used in each stage, as well as
2288
N.C. Om Tapanes et al. / Fuel 87 (2008) 2286–2295
the equipment, reagents and methods of characterization are also described. 2.1. Quantum calculations For the large molecules studied in the present work, with a high degree of conformational freedom, the first point that should be determined is the preferred conformations, which may be used in the simulation of the reaction mechanism. This was done with the Titan package [11], using the conformer distribution subroutine, and the molecular mechanics MMFF method [12,13]. Among the stable conformations, we decide to continue simulations using only the most stable conformation. Simulation of the reaction mechanism was done calculating several points along the reaction coordinate in each case. The point of highest energy along the reaction path was taken as the transition structure. These calculations were done with the semi-empirical AM1 method [14] using the Gaussian 03 W software [15]. The choice between any of the large variety of theoretical methods available is always a matter of balance between required accuracy and computer time. Semiempirical calculations are widely used in homogeneous catalysis [16,17]. The semi-empirical methods simplify the solution of the Schro¨dinger’s equation by parametrization of relevant integrals. This considerably reduces the computational cost while the quality of the results is still acceptable. Therefore even for such large molecules calculations can be done at a reasonable time [18–20]. Among the several available semi-empirical methods we choose the AM1 parametrization. Previous calculations of reaction mechanism using the AM1 method have been reported and they demonstrate the reliability of the method, at least for reactions similar to those studied in the present work (reactions involving only the organic elements C, H, O, N) [17,21–23]. The transesterification reaction can be described as a three stage process involving consecutive, reversible and equivalent stages, with the same chemical transformation. From the mechanistic point of view there should be no difference among process 1, 2 and 3 stages Due to the similarity in the mechanism for transesterification of the three stages and the computational complexity of the problem (calculation with large molecules), we choose to simulate only the last step: monoglyceride transesterification, since it is the simplest stage, besides obtaining the two final products (methyl ester and glycerin) [24,25]. We therefore assume that the mechanism and activation barriers for the other stages should be similar to that simulated here. The monoglycerides used in the theoretical simulation were 1- and 2-palmitic acid monoglyceride (1-MGpalm and 2-MGpalm, respectively), the 1-oleic acid monoglyceride (1-MGoleic) and 1-linoleic acid monoglyceride (1MGlinol). These are the fatty acid radicals predominant in the Jatropha curcas oil. We choose the 1-MGpalm to study the mechanism of the base-catalyzed transesterification.
2.2. Transesterification reactions Refined, bleached, and deodorized Jatropha curcas oil was obtained from Science Academy of Cuba. Methanol, ethanol, potassium hydroxide and sodium methoxide of analytical grade were obtained from Fisher Scientific Co. (Pittsburgh, PA). Reactions were performed in a bath reactor. The reactor was immersed in a constant-temperature water bath equipped with a temperature controller. Agitation was set at a constant speed throughout the experiment (300 rpm). Initially, the reactor was filled with 100 g Jatropha curcas oil and heated to 45 °C. The catalytic solution, prepared with 0.8% of the catalyst dissolved in the alcohol (molar ratio of alcohol/oil = 9), was then added to the reactor at which point the reaction was assumed to have started. After completion 60 min of reaction, glycerol was separated by gravity. The catalyst and unreacted alcohol were removed by hot water washings. In order to avoid the formation of emulsion after the transesterification reaction, 0.1% aqueous citric acid was used as a washing solution to remove catalyst. Finally, esters were dried in a conventional dryer vessel, under a nitrogen stream, for 1 h at 130 °C. Determination of free glycerol and mono-, di- and triglyceride content were based on EN 14105 standard [26]. The ester content was evaluated by EN 14103 standard [27]. The Origin and the Statistica 6.0 softwares were used for statistical modeling. 3. Results and discussion 3.1. Mechanisms of the base-catalyzed transesterification 3.1.1. Mechanism with formation of two tetrahedral intermediates [3,4] This reaction mechanism for the base-catalyzed transesterification was formulated as three steps. In a pre-step the basic catalyst (NaOH or KOH) reacts with the alcohol, producing an alkoxide anion. The first step is a nucleophilic attack of the alkoxide anion on the carbonyl group of the glyceride to form a tetrahedral intermediate (intermediate I). In the second step, the tetrahedral intermediate reacts with a second alcohol molecule (methanol) to regenerate the anion of the alcohol (methoxide), and form another intermediate (intermediate II). In the last step, rearrangement of the tetrahedral intermediate results in the formation of a fatty acid ester and glycerin. All these steps are reversible. In the pre-step the basic catalyst is mixed with the alcohol, the actual catalyst, and the alkoxide group is formed. Scheme 3 summarizes the proposed mechanism. Mulliken atomic charge densities on the oxygen atoms in the optimized structures of the tetrahedral intermediate I of 1-MGpalm and 2-MGpalm (Scheme 4) indicate that the oxygen supposed to be protonated by a second methanol molecule to form the tetrahedral intermediate II has the smallest negative charge. This result may represent a possible contradiction of this mechanism since protonation
N.C. Om Tapanes et al. / Fuel 87 (2008) 2286–2295
OH- + CH3 OH
CH3O - + H2O
H
C H
+ CH O 3 CH3
C CH 2
O
Step 2
Step 1
O H
14
H H
H
C CH 2
O
C
H
CH3
14
H H
H
C
C CH 2
CH3
14
CH 3O
+
OCH 3
C C OH
H OH
H
C
OH
C OH
OH
+ O H
H
OCH 3
C C
H
O
O
OH H
2289
Intermediate II
Intermediate I
Step 3 H H
OH
C
O +
C
H
C
OH H
C
CH 2
OCH 3
14
CH3
OH
H
Scheme 3. The mechanism of base-catalyzed transesterification of 1-MGpalm with formation of two tetrahedral intermediates.
-0.3964
O
CH 2
14
CH3
- 0.6884
O
OH
C
H
C CH 2
O
OCH 3
C
H
C
H -0.3654
C H
OH
C O
C
H
- 0.3875
C
H
H H
H
-0.7115
H
14
CH3
OCH 3
OH
-0.4044
OH
Scheme 4. Charge distribution on the oxygen atoms in the optimized structures of the tetrahedral intermediate I of 1-MGpalm and 2-MGpalm.
should occur in the most negative oxygen atom, which is indeed the carbonyl oxygen. Note, however, that protonation of the oxygen atom bonded to the glycerin group leads directly to bond breaking, as show in Table 2. The theoretical bond length between the oxygen atom of the glycerin and the carbon atom of the carbonyl group of the 1-MGpalm in the tetrahedral intermediates I and II are given in Table 2. Observe that the bond length for the tetrahedral intermediate II is a breaking-bond length (larger ˚ ). These results suggest that the tetrahedral interthan 3.5 A mediate II is not formed in the reaction, therefore the break of the C–O bond take place before, or simultaneously with the H+ attack. This is not surprising, since the results obtained in this theoretical paper agrees better with results obtained experimentally [5,6] based on an alternative mechanism, as discussed below. 3.1.2. Mechanism with formation of one tetrahedral intermediate In this mechanism the tetrahedral intermediate, generated in step 1, is broken in step 2, forming the alkyl ester
Table 2 Bond length between the oxygen atom of the glycerin and the carbon atom of the carbonyl group (–) in the optimized structures of the neutral molecule and the tetrahedral intermediates I and II of 1-MGpalm and 2-MGpalm Compounds
1-MGpalm
2-MGpalm
Structure
Neutral molecule
O H2 C
O
C
Bond length 0 ˚) (A
Structure
1.37
H2 C OH
O
HC
C
C15H31
H C OH
O
Bond length 0 ˚) (A 1.37 C15H31
H2C OH
H2C OH O
Intermediate I
H2 C
O
H C OH
C
C15H31
1.47
OCH3
H2 C OH
O
HC
C
O
C15H31
H2C OH
OCH3
H2 C OH
O
1.46
H2C OH O
Intermediate II
H2 C
+
O H H C OH
H2C OH
C
C15H31
OCH3
3.96
HC
+
O H H2C OH
C
C15H31
OCH3
3.90
2290
N.C. Om Tapanes et al. / Fuel 87 (2008) 2286–2295
OH- + CH3 OH
CH3O - + H2O Step 1
Step 2
O H
C
C
H
C H
14
CH3
+ CH O 3
OH H
CH 2 14
O OCH 3
C C
H
O
C
O
C
C
H H
OH
H
H
C
H H
CH 2
O
H
O
OH OH
CH3
C
H
OH
+
C
H
C OCH 3
CH 2
14
CH3
OH
H
Step 3
CH 3OH H
OH
C
H
C
H H
+ CH3O
C
H
OH OH
Scheme 5. The mechanism of base-catalyzed transesterification of 1-MGpalm with formation of only one tetrahedral intermediate.
and the glyceride anion. In step 3, the active specie is regenerated, allowing to start a new catalytic cycle. The mechanism is shown in Scheme 5. Our results show that this mechanism seems more probable for the base-catalyzed transesterification of vegetable oils than the previous one. Therefore, it will be used in the quantum simulations for the kinetic study.
This is not surprising, since this step occurs rather quickly. Fig. 2 shows the calculated activation energy of steps 1, 2 and 3 for the reactions of the palmitic, oleic and linoleic acid monoglycerides with methanol and ethanol. Similar results were observed in all cases. Based on the molecular calculations, the following relevant consequences may be derived:
3.2. Kinetic study 3.2.1. Quantum simulation Activation energies were determined by the semi-empirical simulations of each step of the reaction mechanism, not considering the pre-step, which is common for all of the reactions. Energy profiles along the fully optimized reaction steps 1, 2 and 3 are illustrated in Fig. 1a–c, respectively. Fig. 1a describes the changes in the bond length of the oxygen atom of the methoxide anion to the carbon atom 0 ˚ ) along step 1. of the carbonyl of 1-MGpalm (3.6–1.4 A In this approach the reaction pass through an initial complex and after a transition state (TS1 in Fig. 1a) with low activation energy (4.19 kcal/mol) forms the tetrahedral intermediate as the main product. In the next step, Fig. 1b, the tetrahedral intermediate dissociates via a second transition state (TS2, show in Fig. 1b) with a larger activation energy (15.82 kcal/mol), leading to the methyl ester and the glyceride ion. This step of the simulation was done by studying the energy as a function of the separation between the carbon atom of the carbonyl of the 1-MGpalm and the oxygen atom of 0 ˚ ). These results suggest that the glycerin (1.48–3.48 A rupture of the tetrahedral intermediate to form the final product may play a key role in the base-catalyzed transesterification of glycerides. The final step, Fig. 1c, involves a third transition state TS3 with small activation energy (10.24 kcal/mol).
– The first step, which involves formation of the tetrahedral intermediate, is the fastest one. The methoxide anion, formed in the pre-step by the action of the basic catalyst has high reactivity and reacts quickly with the monoglyceride to form an initial complex of low energy (initial complex), which than leads to the tetrahedral intermediate. In the absence of the catalyst, this step could be the slowest one. – The second step involves the rupture of the tetrahedral intermediate. This step has to overcome a large energy barrier. Results suggest that this step is the rate determining-step and controls the rate of the Transesterification reaction. – When comparing the reactions of transesterification of saturated monoglycerides and unsaturated monoglycerides the following facts emerge: Decomposition of the tetrahedral intermediates of the unsaturated monoglycerides involves a higher energy than the decomposition of the intermediate from saturated monoglyceride (using methanol: Ea1-MGoleic = 18.13 kcal/mol; E1-MGlinol = 17.82 kcal/mol; E1-MGpalm = 15.82 kcal/mol; and using ethanol E1-MGoleic = 18.15 kcal/mol and E1-MGlinol = 17.53 kcal/mol; E1-MGpalm = 15.50 kcal/mol). Decomposition of the tetrahedral intermediate of 1MGoleic involves the largest energy, indicating higher stability than in the case of the intermediate from of 1-MGlinol.
N.C. Om Tapanes et al. / Fuel 87 (2008) 2286–2295
2291
O 1.23 1.37
-330,00
CH3
+
O
C
H2CO HC O H H2C O H
C15H33
Reactants O
-340,00
1.38
E (kcal/mol)
δ −O
CH3
1.24
C
H2C O HC O H H2C O H
-350,00
2.23
C15H33
TS1
δ−
CH3
O
O
1.29
-360,00
1.47
Initial Complex
Ea
1.43
C
H2C O HC O H H2C O H
-370,00
C15H33
IntermediateI
-380,00
CH3
O
O
1.24
-355,00
1.38
C
δ−
H2C O HC O H H2C O H
2.23
C15H33
TS
-360,00
Product
E (kcal/mol)
O
+
1.23
-365,00
1.36
H3C
-370,00
O
C
H2C O HC O H H2C O H
C15H33
-375,00 Intermediat
-380,00
O
O
1.29 1.47
1.43
C
H2C O HC O H H2C O H
-210
δ−
CH3
H 2C O HC O H H 2C O H
C15H33
+
CH3
0.97
H
O O
-215 1.31
E (kcal/mol)
-220
H 2C O HC O H H 2C O H
-225
H
TS3
CH3
1.17
H 2C O H HC O H H 2C O H
+
CH3 O
Products
-230 -235
Ea
-240 -245
Initial Complex
Fig. 1. Energy profiles for the reaction between 1-MGpalm and methanol, (a) step 1, (b) step 2 and (c) step 3.
2292
N.C. Om Tapanes et al. / Fuel 87 (2008) 2286–2295
Energy of Activation, kcal/mol
20 15,5 15 15,82 12,87 10,24
10
5
1-MGpalm Met
4,57 4,19
1-MGpalm Et
0 Step 1
Step 2
Step 3
20
20 17,53
18,13 15 12,87 10,24
10
5
1-MGoleic Met
4,38
1-MGoleic Et 2,51
Energy of Activation, kcal/mol
Energy of Activation, kcal/mol
18,15
17,82
15
12,87 10,24
10
6,01 1-MGlinol Met
5
1-MGlinol Et
3,23 0
0 Step 1
Step 2
Step 3
Step 1
Step 2
Step 3
Fig. 2. Activation energy of steps 1, 2 and 3 for reactions of 1-MGpalm, 1-MGoleic and 1-MGlinol with methanol and ethanol.
– The activation energies of the determining-steps for the reaction with methanol and ethanol show approximately the same values. As discussed below the reaction using ethanol as catalyst is somewhat slower than the corresponding reaction with methanol. This means that the main difference between reaction rates must be in the pre-step, where the alkoxide is formed. This seems reasonable since ethoxide is a stronger base than methoxide. Therefore, in the equilibrium process leading to formation of the alkoxide the methoxide ion is more easily formed than the ethoxide one. The optimized molecules and the geometrical data of the intermediates and transition states of all reactions studied are show in Figs. 3–5 and in Table 3, respectively.
3.2.2. Transesterification reactions The reaction rate of base-catalyzed transesterification of triglycerides with homogeneous catalyst is given in Eq. (1): a
b
d
ðrTG Þ ¼ k 1 ½C TG ½C ROH þ k 2 ½C RxCOOR þ ½C GL
c
ð1Þ
(rTG): Triglycerides rate (mol/(volume/time)) a, b, d and c: reaction order of triglycerides, alcohol, biodiesel and glycerol, respectively. k1 and k2: specific velocity of reaction. CTG and CROH: molar concentration of triglycerides and alcohol. Considering the terms b, d and c as zero [3–5] Eq. (1) is simplified to Eq. (2):
Fig. 3. Optimized structures of the tetrahedral intermediates TI (a) and TS2 (b) for the reaction between 1-MGpalm and methanol.
N.C. Om Tapanes et al. / Fuel 87 (2008) 2286–2295
2293
Fig. 4. Optimized structures of the tetrahedral intermediates TI (a) and TS2 (b) for the reaction between 1-MGoleic and methanol.
Fig. 5. Optimized structures of the tetrahedral intermediates TI (a) and TS2 (b) for the reaction between 1-MGlinol and methanol.
Table 3 ˚ ; angles in degrees), for the tetrahedral intermediate (TI) and the transition structure TS1, TS2 and TS3 for the Optimized parameters (bond length in A reactions between 1-MGpalm, 1-MGoleic and 1-MGlinol, and methanol 1-MGpalm
C(1)–O(3) C(1)–O(4) C(1)–O(5) C(1)–C(2)
