You are on page 1of 6

Fuel 86 (2007) 2383–2388

www.fuelfirst.com

A kinetic study of the esterification of free fatty acids (FFA)


in sunflower oil
M. Berrios, J. Siles, M.A. Martı́n *, A. Martı́n
Departamento de Quı́mica Inorgánica e Ingenierı́a Quı́mica, Universidad de Córdoba, Campus Universitario de Rabanales,
Edificio Marie Curie, 14071 Córdoba, Spain

Received 27 July 2006; received in revised form 2 February 2007; accepted 6 February 2007
Available online 7 March 2007

Abstract

The kinetics of the esterification of free fatty acids (FFA) in sunflower oil with methanol in the presence of sulphuric acid at concen-
trations of 5 and 10 wt% relative to free acids as catalyst and methanol/oleic acid mole ratios from 10:1 to 80:1 was studied. The exper-
imental results were found to fit a first-order kinetic law for the forward reaction and a second-order one for the reverse reaction.
The influence of temperature on the kinetic constants was determined by fitting the results to the Arrhenius equation. The energy of
activation for the forward reaction decreased with increasing catalyst concentration from 50 745 to 44 559 J/mol.
Based on the experimental results, a methanol/oleic acid mole ratio of 60:1, a catalyst (sulphuric acid) concentration of 5 wt% and a
temperature of 60 C provided a final acid value for the oil lower than 1 mg KOH/g oil within 120 min. This is a widely endorsed limit for
efficient separation of glycerin and biodiesel during production of the latter.
 2007 Elsevier Ltd. All rights reserved.

Keywords: Biodiesel; Esterification; FFA; Kinetics; Sunflower

1. Introduction an alternative fuel and as an additive for petroleum diesel.


The prefix ‘‘bio’’ denotes its biological origin and renew-
Transportation is the activity with the most uncertain able character; by contrast, the better known, non-renew-
future as regards energy demands in the European Union. able petroleum diesel is obtained from petroleum
In fact, about 67% of the current oil demand in the EU is distillates.
estimated to be required by this sector. Because the EU At present, FAME are obtained by reacting triglycerides
relies heavily on oil imports, fuel consumption is among with lower alcohols such as methanol or ethanol in the
its primary concerns. presence of a strong base as a catalyst; the reaction yields
Liquid fuels of agricultural origin are being increasingly glycerin as a by-product. The triglycerides used come from
considered as alternatives to gasoline and gas oil as sources a variety of oils including rapeseed, sunflower, soybean,
of energy. Specifically, biodiesel has a substantial potential palm and peanut oil, as well as from animal fat [1–3].
to reduce oil imports and ensure continuity in the energy The use of these materials, however, is restricted by their
supply. high cost, which has led some authors [4] to recommend
Biodiesel, which consists of long-chain fatty acid methyl employing used oil and some animal fats as alternative
ethers (FAME) obtained from renewable lipids such as raw materials.
those in vegetable oils or animal fat, can be used both as According to several researchers Freedman et al. [5], Liu
[6] and Mittelbach et al. [7], the oil or fat used in alkaline
transesterification reactions should contain no more than
*
Corresponding author. Tel.: +34 957218586; fax: +34 957218625. 1% free fatty acids (FFA), which is equivalent to
E-mail address: iq2masam@uco.es (M.A. Martı́n). 2 mg KOH/g triglyceride. If the FFA level exceeds this

0016-2361/$ - see front matter  2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.fuel.2007.02.002
2384 M. Berrios et al. / Fuel 86 (2007) 2383–2388

threshold, saponification hinders separation of the ester system consisted of a jacketed reactor through which water
from glycerine and reduces the yield and formation rate from a thermostated system at the required temperature
of FAME. was circulated.
Canakci and Van Gerpen [8] reduces the recommended
acidity to below 1 mg KOH/g triglyceride (i.e. a 0.5% FFA 3. Results and discussion
content).
Van Gerpen [9] found the FAME yield from unrefined The elimination of FFA relies on the well-known ester-
oil to decrease from 93–98% to 86–87% in the presence ification reaction
of 6.66% of FFA and ascribed it to phospholipids destroy-
ing the catalyst. R1 –COOH + R2 –OH ¢ R1 –COO–R2 + H2 O ð1Þ
According to Meher et al. [10], the FFA and moisture
contents are the two variables dictating the feasibility of which is catalysed by acids. In this work, R1 was a linear
transesterifying vegetable oils. chain of 11–17 carbon atoms containing a variable number
Others researchers [11] studied the transesterification of of unsaturations depending on the particular origin of the
beef tallow catalysed by NaOH in the presence of free fatty raw material, and R2 was a methyl radical.
acids and water. The yield of beef tallow methyl esters By means of essays, it has been verified that the esterifi-
(BTME) was highest when no FFA or water was added cation reaction does not take place in catalyst absence
to the reaction mixture. On the other hand, the addition (Fig. 1).
of 0.6% of FFA resulted in the lowest BTME apparent A test has been made with oil which acid value is known,
yield (less than 5%), whichever the amount of water a 5% w/w catalyst concentration and a molar ratio metha-
supplied. nol/oleic acid of 240:1. The experiment has been carried
The economic and environmental interest in using oils out in the same conditions of temperature and agitation
and fats involved in the production of biodiesel, the that the others tests, for 5 min so that the reaction of ester-
unavoidable need to remove FFA and the interest in char- ification does not pass in an appreciable length. Next, the
acterizing the kinetics of the reaction with a view to opti- mixture was settled in order to separate the two phases.
mizing the reactor dimensions led us to undertake the The acidity of the methanol phase (methanol and catalyst)
present study with a view to examining the influence of before and after reaction has been determined: 5,78 and
operational variables on the kinetics of esterification of 5,27 mg KOH/g sample, respectively. The smaller obtained
FFA in sunflower oil using sulphuric acid as catalyst. value confirms that fatty acids do not dissolve in the meth-
anol phase and the part of catalyst could dissolve in the oil
2. Materials and methods phase. Therefore, the chemical reaction takes place in the
oil phase.
The raw material used in the esterification tests was sun- This reaction is heterogeneous, methanol is only spar-
flower oil the acidity of which was raised by addition of sely soluble in sunflower oil; so it requires agitation in
fatty acids obtained in its physical refining. The acid value order to avoid mass transfer taking control over the pro-
of the samples ranged from 5 to 7 mg KOH/g oil, which is cess [3].
slightly higher than the levels typically found in frying oils. Tests were conducted at variable agitation speeds from
The sample contained the following major fatty acids: 200 to 600 rpm – above which this variable was found to
palmitic (6%), stearic (3%), oleic (17%) and linoleic (72%); have no further effect on the reaction rate (Fig. 2). A speed
all were determined by gas chromatography in accordance
with UNE-EN 14331. 2.8
The acid value was determined as per the UNE-55001, 2.6
following washing of the sample with distilled water to
2.4
remove any residual catalyst.
Acid value (mg KOH/g oil)

The esterification catalyst used was sulphuric acid at 2.2

concentrations of 5 and 10 wt% relative to total acidity 2.0


(as oleic acid).
1.8
Free fatty acids (FFA) were esterified with anhydrous Reaction Non-catalyzed
1.6 Reaction Catalyzed
methanol, using a methanol/oleic acid mole ratio of 20:1–
80:1 and 10:1–114:1 with a catalyst concentration of 5% 1.4

and 10%, respectively. 1.2


The influence of temperature on the esterification rate
1.0
was studied at a fixed methanol/oleic acid ratio of 60:1,
the two catalyst concentrations and 30, 40, 50 and 60 C. 0.8
0 50 100 150 200
The experimental set-up used was identical with that
Time (min)
previously employed by [12], batch stirring reactor of 2 l
of capacity with magnetic agitation, except that the heating Fig. 1. Comparison of the non-catalysed and catalysed reaction.
M. Berrios et al. / Fuel 86 (2007) 2383–2388 2385

3.5 6
10:1
200 rpm 60ºC; 10% H2SO4 20:1
300 rpm
5 40:1
3.0 400 rpm
60:1

Acid value (mg KOH / g oil)


500 rpm
Acid value (mg KOH/g oil)

80:1
600 rpm
4 114,6:1
2.5

2.0
2

1.5
1

1.0 0
0 20 40 60 80 100 120 140
0 20 40 60 80 100 120 140
Time (min)
Time (min)
Fig. 4. Variation of the acid value with the methanol/oleic acid mole ratio
Fig. 2. Comparison of the agitation speeds.
at a sulphuric acid concentration of 10 wt%.

of 600 rpm was therefore subsequently used to examine the Once a methanol/oleic acid mole ratio of 60:1 was
influence of the catalyst concentration, methanol/fatty acid adopted as optimal, the influence of temperature at the
mole ratio and temperature on the acid value. two catalyst concentrations studied was examined. As
Being the density of the fluid of 913 kg/m3, the diameter expected, raising the temperature increased the esterifica-
of the agitator 4 cm, the speed of agitation of 600 rpm and tion rate (see Figs. 5 and 6). The highest rate was obtained
the power consumed in the process is of 6.6 W, the tests at 60 C, which is close to the boiling point of methanol at
were carried out with a power number of 20 (Np). atmospheric pressure. A higher temperature obviously
Figs. 3 and 4 show the variation of the acid value with increased the rate further, but required using a pressure
the methanol/oleic acid mole ratio at a sulphuric acid con- above atmospheric level or a more sophisticated experi-
centration of 5% and 10%, respectively. In both cases, the mental set-up such as that employed by Kocsisová et al.
reaction rate increased with increasing mole ratio. Above [13].
60:1, however, the increase was negligible, so this ratio A reaction time of 120 min, a methanol/oleic acid mole
was adopted as the experimental limit in order to avoid ratio of 60:1, a sulphuric concentration of 5% and a tem-
unnecessarily increasing operational costs and the reactor perature of 60 C provided an acid value below
size. On the other hand, an increased catalyst concentra- 1 mg KOH/g oil, which is the maximum recommended
tion increased the esterification rate. Thus, at a metha- level for ensuring a good yield of FAME while facilitating
nol/oleic acid mole ratio of 20:1, a catalyst concentration their separation from glycerin in the transesterification
of 5% provided an acid value of 4 mg KOH/g oil within reaction. In fact, higher acid values result in saponification
120 min, whereas one of 10% gave 2 mg KOH/g oil in the by reaction with the catalyst, which decreases the surface
same time. tension and facilitates emulsification [8].

8 7

20:1 60ºC
7 60ºC; 5% H2SO4 40:1 5%HH 2SO
SO 50ºC
6 5% 2 44
60:1 40ºC
Acid value (mg KOH / g oil)
Acid value (mg KOH / g oil)

6 80:1 30ºC
5

5
4
4
3
3

2
2

1 1

0 0
0 20 40 60 80 100 120 140 0 20 40 60 80 100 120 140

Time (min) Time (min)

Fig. 3. Variation of the acid value with the methanol/oleic acid mole ratio Fig. 5. Variation of the acid value with temperature at a sulphuric acid
at a sulphuric acid concentration of 5 wt%. concentration of 5% and a methanol/oleic acid mole ratio of 60:1.
2386 M. Berrios et al. / Fuel 86 (2007) 2383–2388

7 where
30ºC
40ºC
s ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

10% H2SO4
6 50ºC K2
60ºC a¼ þ K  A0 ð5Þ
4
Acid value (mg KOH / g oil)

5
a K1
4 b¼ and K ¼ ð6Þ
K K2
3 K1 and K2 were determined by trial and error, using vari-
able b values until a plot of the right-hand side of Eq. (4)
2
consisting of a straight line with a negligible intercept
1
was obtained. By way of example, Fig. 7 shows the fitting
of the experimental data obtained at a methanol/oleic acid
0 mole ratio of 10:1, a catalyst concentration of 10% and a
0 20 40 60 80 100 120 140
temperature of 60 C. Similar plots were constructed under
Time (min)
all other experimental conditions.
Fig. 6. Variation of the acid value with temperature at a sulphuric acid Fig. 8 shows the kinetic constants obtained as described
concentration of 10% and a methanol/oleic acid mole ratio of 60:1. above. As expected, K1 increased with increasing metha-
nol/oleic acid mole ratio and catalyst concentration; on
3.1. Kinetic model the other hand, K2 was negligible irrespective of the cata-
lyst concentration, which indicates that the hydrolysis reac-
The kinetic model used in this work relied on the follow- tion hardly took place. Based on the small value of K2,
ing assumptions: some authors have proposed fitting experimental data to
a first-order kinetic law [13,15].
(a) The esterification reaction was a reversible heteroge- The influence of temperature on the specific reaction
neous process the rate of which under the operating rate was determined by fitting K1 and K2 to the Arrhenius
conditions used was controlled by that of the chemi- equation,

cal reaction. DE


K ¼ A  exp ð7Þ
(b) The rate of the non-catalysed reaction was negligible RT
relative to the catalysed reaction.
(c) The chemical reaction occurred in the oil phase. using plots of ln K as a function of the reciprocal tempera-
(d) The methanol/oleic acid mole ratio used was high ture (Figs. 9 and 10). However, both the frequency factor,
enough for the methanol concentration to remain A, and the energy of activation, DE, were obtained by non-
constant throughout the process. linear regression, using the software Mathcad v. 12. The re-
sults are shown in Table 1.
Under these conditions, the reaction was assumed to be The energy of activation for the forward reaction
pseudo-homogeneous, first-order in the forward direction decreased with increasing catalyst concentration: from
and second-order in the reverse direction, and hence to 50 745 J/mol with 5% sulphuric acid to 44 559 J/mol with
conform to the following kinetic law: a 10% concentration of the acid.

d½A
¼ K 1  ½A  K 2  ½C  ½D ð2Þ 1.6
dt
where [A] denotes the concentration of FFA in mg KOH/g 1.4
Ln{[A0+E(β-0.5)] / [A0-E(β+0.5)]}

oil (i.e. the acid value); [C] and [D] are the concentrations 1.2
of FAME and water, respectively, formed during the reac-
tion, both also referred to 1 g of oil; and K1 and K2 are the 1.0

kinetic constants for the forward and reverse reaction, 0.8


respectively. If [D] and [C] are assumed to be zero at the
start (t = 0), and A = A0  E (E being the acidity re- 0.6

moved), then, according to Carberry [14], 0.4 y = 4,765·10-3 + 0,012·x

dE r ² = 0,998
¼ K 1  ðA0  EÞ  K 2  E2 ð3Þ 0.2
dt
0.0
where A0 is the initial concentration of FFA. 0 20 40 60 80 100 120 140

Integration of Eq. (3) yields Time (min)


  
A0 þ E  b  12 Fig. 7. Determination of the kinetic constants by using Eq. (4) at a
2  K 2  a  t ¼ ln    ð4Þ methanol/oleic acid mole ratio of 10:1, a catalyst concentration of 10%
A0  E  b þ 12 and a temperature of 60 C.
M. Berrios et al. / Fuel 86 (2007) 2383–2388 2387

0.07 Table 1
k1 (min-1) (5%)
Energies of activation and frequency factors for the esterification of FFA
0.06
k1 (min-1) (10%)
in sunflower oil

0.05
k2 (g/mg·min) (5%) A E (J/mol) R2
k2 (g/mg·min) 10%)
+6
K1 (5%) 2.869 · 10 50745.2 0.999
0.04
K2 (5%) 37.068 31007.3 0.932
K1 (10%) 3.913 · 10+5 44558.8 0.997
k

0.03
K2 (10%) 707.166 42761.4 0.932
0.02

0.01
8
0.00 5% H2SO4

Calculate acid value (mg KOH/ g oil)


10% H2SO4 +10%
0 20 40 60 80 100 120
6
Molar ratio (Methanol/Oleic Acid)
-10%
Fig. 8. Variation of the rate constants with the methanol/oleic acid mole
ratio. 4

-2
5% H2SO4
2
ln k1
-4

0
-6
0 1 2 3 4 5 6 7
ln k

ln k2 Experimental acid value (mg KOH/g oil)


-8
Fig. 11. Comparison of the experimental acid value and that predicted by
Eq. (4).
-10

Finally, the goodness-of-fit of the experimental data to


-12 the proposed model was assessed by comparing the exper-
2,95e-3 3,00e-3 3,05e-3 3,10e-3 3,15e-3 3,20e-3 3,25e-3 3,30e-3 3,35e-3
imental acid value with the theoretical prediction of Eq. (4)
-1
1/T (K ) at a methanol/oleic acid mole ratio of 60:1 (see Fig. 11).
Fig. 9. Influence of temperature on the specific reaction rate (Arrhenius Three-fourths of the experimental data were reproduced
equation) at a sulphuric acid concentration of 5 wt%. with errors not greater than 10%.

-2 4. Conclusions
10% H2SO4
ln k1
-4
Free fatty acids (FFA) in sunflower oil can be effectively
removed by esterification with methanol, using a 5% sul-
-6
phuric acid concentration relative to FFA, a methanol/
oleic acid mole ratio of 60:1, a temperature of 60 C and
ln k

-8
ln k2 agitation at a speed of 250 rpm or higher.
Under these conditions, the acidity removed, E,
-10
expressed as the acid value (mg KOH/g oil), as a function
of time (in min) conforms to the following equation:
-12
A0 ðe2K 2 at  1Þ

-14 bð1 þ e 2 at Þ þ 0:5ðe2K 2 at
2K  1Þ
2,95e-3 3,00e-3 3,05e-3 3,10e-3 3,15e-3 3,20e-3 3,25e-3 3,30e-3 3,35e-3

1/T (K-1) where a and b have the same meaning as in Eqs. (5) and (6),
respectively, and K1 and K2 are calculated by using the data
Fig. 10. Influence of temperature on the specific reaction rate (Arrhenius
equation) at a sulphuric acid concentration of 10 wt%.
of Table 1 in Eq. (7).

The frequency factor for the forward reaction varied Acknowledgements


similarly to the energy of activation and increased from
2.869 · 106 to 3.913 · 106 in raising the catalyst concentra- The authors are grateful to the firm Biodiesel de And-
tion from 5% to 10%. alucı́a for funding this work.
2388 M. Berrios et al. / Fuel 86 (2007) 2383–2388

References Proceedings of the alternative energy conference. Michigan


(USA): American Society of Agricultural Engineers; 1994. p. 74.
[1] Antolin G, Tinant FV, Briceño Y, Castaño V, Pérez C, Ramirez AI. [8] Canakci M, Van Gerpen J. Biodiesel production from oils and fats
Optimization of biodiesel production by sunflower oil transesterifi- with high free fatty acids. Trans ASAE 2001;44(6):1429–36.
cation. Bioresour Technol 2002;83:111–4. [9] Van Gerpen J. Biodiesel processing and production. Fuel Process
[2] Al-Widyan MI, Al-Shyoukh AO. Experimental evaluation of the Technol 2004.
transesterification of waste palm oil into biodiesel. Bioresour Technol [10] Meher LC, Vidya Sagar D, Naik SN. Technical aspects of biodiesel
2002;85:253–6. production by transesterification – a review. Renew Sust Energy Rev
[3] Ma F, Clements LD, Hanna HA. The effect of mixing on 2006;10(3):248–68.
transesterification of beef tallow. Bioresour Technol 1999;69:289– [11] Ma F, Clements LD, Hanna MA. The effect of catalyst, free fatty
93. acids and water on transesterification of beef tallow. Trans ASAE
[4] Zhang Y, Dube MA, Malean DD, Kates M. Biodiesel production 1998;41(5):1261–4.
from waste cooking oil: process design and technological assessment. [12] Alcantara R, Amores J, Canoira L, Fidalgo E, Franco MJ, Navarro
Bioresour Technol 2003;89:1–16. A. Catalytic production of biodiesel from soy-bean oil, used frying oil
[5] Freedman B, Pryde EH, Mounts TL. Variables affecting the yields of and tallow. Biomass Bioenergy 2000;18:515–27.
fatty esters from transesterified vegetable oils. JAOCS 1984;61: [13] Kocsisová T, Cvengros J, Lutisan J. High-temperature esterification
1638–43. of fatty acids with methanol at ambient pressure. Eur J Lipid Sci
[6] Liu K. Preparation of fatty acid methyl esters for gas-chromato- Technol 2005;107:87–92.
graphic analysis of lipids in biological materials. JAOCS 1994;71(11): [14] Carberry JJ. Chemical and catalytic reaction engineering. New
1179–87. York: McGraw-Hill; 1976.
[7] Mittelbach M, Pokits B, Silberholz A. Diesel fuel derived from [15] Sendzikiene E, Makareviciene V, Janulis P, Kitrys S. Kinetics of free
vegetable oils, IV: production and fuel properties of fatty acid methyl fatty acids esterification with methanol in the production of biodiesel
esters from used frying oil. Liquid fuels from renewable resources, fuel. Eur J Lipid Sci Technol 2004;106:831–6.

You might also like