Synthesis of biodiesel via homogeneous Lewis acid catalyst

Article abstract of DOI:10.1016/j.molcata.2005.05.041

Nowadays, most biodiesel (fatty acids methyl esters, FAME) is produced by the transesterification of triglycerides (TG) of refined/edible type oils using methanol and an homogeneous alkaline catalyst. However, production costs are still rather high compared with the ones of petroleum-based diesel fuel. To lower costs and make biodiesel competitive less-expensive feedstocks such as waste fats or non-edible type oils could be used. The use of homogeneous alkaline catalysts in the transesterification of such types of fats and oils poses great difficulties due to the presence of large amounts of free fatty acids (FFA). This paper studies the use of carboxylic salts as a possible alternative, because these catalysts are active also in the presence of high FFA concentrations. The most active catalyst (Cd, Mn, Pb, Zn carboxylic salts) have been individuated and a correlation of the activities with the cation acidity has been found.

Full text of DOI:10.1016/j.molcata.2005.05.041

Journal of Molecular Catalysis A: Chemical 239 (2005) 111–115
Synthesis of biodiesel via homogeneous Lewis acid catalyst
a
a
a
a,b
M. Di Serio , R. Tesser , M. Dimiccoli , F. Cammarota ,
b
a,∗
M. Nastasi , E. Santacesaria
a
Universit a` di Napoli Federico II, Dipartimento di Chimica, via Cintia, 80126 Napoli, Italy
b
ASER srl, Italy
Received 6 April 2005; received in revised form 25 May 2005; accepted 26 May 2005
Available online 19 July 2005
Abstract
Nowadays, most biodiesel (fatty acids methyl esters, FAME) is produced by the transesterification of triglycerides (TG) of refined/edible
type oils using methanol and an homogeneous alkaline catalyst. However, production costs are still rather high compared with the ones of
petroleum-based diesel fuel. To lower costs and make biodiesel competitive less-expensive feedstocks such as waste fats or non-edible type
oils could be used. The use of homogeneous alkaline catalysts in the transesterification of such types of fats and oils poses great difficulties
due to the presence of large amounts of free fatty acids (FFA). This paper studies the use of carboxylic salts as a possible alternative, because
these catalysts are active also in the presence of high FFA concentrations.
The most active catalyst (Cd, Mn, Pb, Zn carboxylic salts) have been individuated and a correlation of the activities with the cation acidity
has been found.
©
2005 Elsevier B.V. All rights reserved.
Keywords: Biodiesel; Transesterification; Esterification; Homogeneous catalysts; Waste oil
1
. Introduction
However, production costs are still rather high, compared
to petroleum-based diesel fuel production costs [1,2]. To
lower costs and make biodiesel competitive with petroleum-
based diesel, less-expensive feedstocks such as waste fats
or non-edible type oils, could be used [1–3]. Homogeneous
alkaline catalysts in the transesterification of such types of
fats and oils cannot directly be used due to the presence of
large amounts of free fatty acids (FFA) [1–3]; in fact, for the
use of these catalyst the FFA concentration should be less
than 0.5% (w/w) [4].
Biodiesel (fatty acids methyl esters, FAME) has recently
become very attractive, because of its environmental bene-
fits and the fact that it is produced from renewable sources
[1]. Nowadays, most biodiesel is produced by the trans-
esterification of triglycerides of refined/edible type oils
using methanol and an alkaline catalyst (NaOH, NaOMe)
[1–3]:
Several methods have been proposed to solve these
problems:
(
a) alkali refining method [5]: an alkali such as caustic soda
is added to transform free fatty acids into fatty acid soaps
and these are then removed by washing with water;
(
b) excess addition of catalyst method [5]: in addition to the
amount of catalyst sufficient for transesterification, an
excess of alkali is added in relation to the FFA amount;
∗
(c) solvent extraction method [5]: free fatty acids are
removed by extraction with a selective solvent;
Corresponding author. Tel.: +39 081674027; fax: +39 081674026.
E-mail address: santacesaria@chemistry.unina.it (E. Santacesaria).
1
381-1169/$ – see front matter © 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2005.05.041

1
12
M. Di Serio et al. / Journal of Molecular Catalysis A: Chemical 239 (2005) 111–115
(
d) distillation refining method [5]: free fatty acids are dis-
acetate as catalyst [7]. They showed that this catalyst, operat-
◦
−3
tilled away from the oil;
ing in a temperature range 200–250 C and with 5 × 10 :1
weight ratio of catalyst to oil, can be used in the presence of a
high FFA concentration, obtaining after purification, a crude
ester phase a with very low FFA concentration and high con-
version of triglycerides. However, in other transesterification
reactions several acetates of bivalent metals have been used as
catalysts [8]. Since the activity of the catalysts is a function of
both metal acidity and ester and alcohol molecular structures,
every ester–alcohol couple will have a specific metal that will
give a maximum of activity [9]. Starting from these consid-
erations, this paper reports the results of a catalytic screening
of acetates of the following metals: Ca, Ba, Mg, Cd, Mn, Pb,
Zn, Co, Ni in the transesterification of the TG with methanol.
Moreover, the stearates of all the above metals were synthe-
sized and the influence of the molecular weight of catalyst
anion on the reaction have also been investigated. Further-
more, the influence of the presence of water and of FFA
respectively on the activity has been studied. Finally, we will
show the possibility to perform TG transesterification and
FFA esterification in a single step, using a low catalyst con-
centration. Amethodforremovingthemetalcatalysthasbeen
suggested.
(
e) pre-esterification method [2,5]: FFA are firstly esterified
to FAME by using an acid catalyst:
RCOOH + MeOH ꢀ RCOOMe + H2O
and then the transesterification is performed, as usual, by
using an alkaline catalyst;
f) use of Bronsted acid catalysts [3,6]: an esterification and
transesterification step are both promoted by a Bronsted
acid catalyst at high temperature.
(
All the first four methods (a–d) result in a loss of prod-
uct (biodiesel). Moreover, methods (a) and (b) give place
to problems during phase separation, method (a) before the
transesterification, method (b) after it because of the forma-
tion of emulsion due to the presence of soaps. Method (c)
requires a high solvent/oil weight ratio while methods (d)
and (e) require high energy use.
Methods (e) and (f) seems to be more attractive. The
acid-catalyzed pre-esterification of FFA – method (e) – is
a common practice in decreasing FFA levels in high FFA
feedstocks, before performing the base-catalyzed transester-
ification [3]. Zhang et al. [6] recently showed, by a techno-
logical assessment of different continuous processes, that the
acid-catalyzed process using waste oil is technically feasible
and less complex than two a step process (pre-esterification
with homogeneous acid catalyst and alkali catalyzed stages),
although no commercial biodiesel plants since the time of
their work have been reported to use the acid-catalyzed pro-
cess.
2. Experimental
2.1. Methods, techniques and reagents
The catalytic screening was performed in small stainless
steel vial reactors. The reaction was made by introducing
reagents (methanol and soybean oil with an acidity of 0.2%,
w/w) and a catalyst in each reactor. All reactor sets (6) were
then heated in a ventilated oven. The temperature of the
oven was increased from room temperature at a fixed rate
The main drawback of the pre-esterification method (e)
consists in the necessity to remove the homogeneous acid
catalyst from the oil after pre-esterification. This problem
can be solved with the use of an heterogeneous acid catalyst
◦
[3,4]. However, the necessity to eliminate the water formed in
(20 C/min) until reaching the reaction temperature of the
◦
FFA esterification still remains because the presence of water
may favour ester saponification under alkaline conditions
runs (150–200 C). Then, after 55 min the temperature was
quickly diminished by putting the vials in a cold bath. Exper-
imental runs were also made with the addition of water and
FFA to the reactants.
Acetates and the synthesized stearate of the following met-
als: Ca, Ba, Mg, Cd, Mn, Pb, Zn, Co, Ni were tested as
catalysts.
[
7].
The disadvantage of the acid-catalyzed process – method
f) – is that the reaction, normally, requires a high methanol
(
to oil molar ratio and high acid catalyst concentration. Zang
et al. [6] obtained a 97% oil conversion to FAME within
2
0
40 min, using a 50:1 molar ratio of methanol to oil and a
.14:1 weight ratio of sulphuric acid to oil at 80 C. How-
The syntheses of metal stearates were performed in a
magnetically stirred glass reactor. The jacketed reactor was
kept isothermal by recirculating a thermostated oil at 180 C.
The reactions were made by introducing a metal acetate
and stearic acid in stoichiometric amounts. The acetic acid
formed during the reaction (3 h) was stripped by a nitrogen
stream with flow rate 2 l/h and collected in a condenser. The
conversion obtained in the performed syntheses, reported in
Table 1, were calculated by measuring the amount of acetic
acid recovered.
◦
◦
ever, this process gives rise to problems linked with the
corrosive action of the liquid acid catalyst and to the high
quantity of obtained by-products [2]. Actually, a plant of
1
0,000 t/y of biodiesel co-produces 2000 t/y of CaSO4 [5]
with the sulphuric acid being neutralized with CaO. A solid
acid catalyst could to eliminate these problems but nowadays
the proposed catalysts have not yielded satisfactory results,
and much greater research efforts are necessary [2].
Basu and Norris proposed, as a possible solution to the
production of biodiesel from oil with a high FFA concen-
tration, the use of a mixture of calcium acetate and barium
The FAME yields, in the catalytic tests, were determined
using the H-NMR technique (Bruker 200 MHz) [10], i.e.,
measuring the area of H-NMR signal related to methoxylic

M. Di Serio et al. / Journal of Molecular Catalysis A: Chemical 239 (2005) 111–115
113
Table 1
Conversion obtained in the synthesis of stearates
Table 2
◦
List of runs performed with acetates as catalyst at 200 C
Run
Stearates
Conversion (%)
Run
Catalyst
M2+ (mmol)
λGly
1.1
2.1
3.1
4.1
5.1
6.1
7.1
8.1
9.1
Ba(OOC(CH2)16CH3)2
Ca(OOC(CH2)16CH3)2
Mg(OOC(CH2)16CH3)2
Cd(OOC(CH2)16CH3)2
Mn(OOC(CH2)16CH3)2
Pb(OOC(CH2)16CH3)2
Zn(OOC(CH2)16CH3)2
Co(OOC(CH2)16CH3)2
Ni(OOC(CH2)16CH3)2
73
73
72
89
62
58
64
81
66
1.2
2.2
3.2
4.2
5.2
6.2
7.2
8.2
9.2
Ba(OOCCH3)2
Ca(OOCCH3)2
0.0571
0.0577
0.0573
0.0578
0.0575
0.0580
0.0583
0.0577
0.0574
0.0587
0.11
0.31
0.39
0.85
0.67
0.81
0.67
0.20
0.07
0.38
Mg(OOCCH3)2·4H2O
Cd(OOCCH3)2
Mn(OOCCH3)2
Pb(OOCCH3)2
Zn(OOCCH3)2·2H2O
Co(OOCCH3)2·4H2O
Ni(OOCCH3)2·4H2O
Ca(OOCCH3)2, Ba(OOCCH3)2
1
0.2
(Ca/Ba = 3/1, w/w)
(A1) and methylenic groups (A2), respectively:
Soybean oil 2.0 g (FFA 0.2% w/w), methanol 0.88 g.
A1/3
YFAME =
As can be seen, all the considered acetates except nickel
acetatehaveshowncatalyticactivity. Inparticular, theactivity
of barium acetate is lower than the activity of all the other
considered metal acetates, while calcium acetate is higher
only than barium and cobalt acetate. The catalyst mixture
A2/2
In the case of non-acid oil, the FAME yield corresponds also
to the conversion of glyceride groups (λGly):
λ
Gly
= YFAME
(Ca(OOCH3)2/Ba(OOCH3)2 = 3/1, w/w) proposed by Basu
To have the conversion of glyceride groups, when the reac-
tion is performed on acid oil, the initial FFA weight fraction
and Norris [7] shows a little cooperative effect giving place
to a slight increase in the obtained conversions, compared to
the performance of the single acetates.
(wFAME) and FFA conversion (λFAME) must be considered,
because FAME are also produced in the FFA esterification.
In this case, we have:
The activity of the tested metal acetates in the transester-
ification reaction is probably related to the Lewis acidity of
the metal [8], and there is, in general, an optimal value of
this acidity for every ester–alcohol couple employed in the
reaction [9]. In Fig. 1, the conversion of the glyceride groups
obtained with the different metal acetates is reported as a
function of the logarithm of their stability constants in the for-
mation of the corresponding complex with dibenzoilmethane
ꢀ
ꢁ
PMoil
wFAME
λ
Gly
= YFAME 1 + 3
PMOA 1 − wFAME
PMoil
wFAME
−
λFAME
3
PMOA 1 − wFAME
FFA conversion is determined by measuring the residual FFA
concentration by titration [11].
(log βDBM) [9,13]. The values of these constants can be con-
sidered as a relative measure of metal acidity [9,13–15]. As
can be seen, in the oil transesterification with methanol as for
other transesterification reactions [9,13], the activity rises to
In order to confirm the results obtained in the vials, dif-
3
ferent runs were performed in a 1 dm autoclave reactor at
◦
2
10–220 C. These runs were performed by introducing reac-
tants and catalyst into the autoclave and heating until the fixed
reaction temperature was achieved. The temperature profiles
of heating were roughly linear: the reaction temperature was
achieved in about 80 min. Samples of the reaction mixture
were withdrawn at different times and analyzed by both H-
NMR spectroscopy and titration.
3
. Results and discussion
First of all, a catalytic screening of different metal acetates
in the transesterification of a “neutral” soybean oil (FFA con-
centration = 0.2%, w/w) in the temperature range adopted by
◦
Basu and Norris [7] (200 C) was performed. Table 2 reports
both the other related operative conditions and the results.
Since at high temperatures the stainless steel internal sur-
face of the vials can catalyze the transesterification reaction
[
12], also a run without catalyst was performed. The obtained
◦
value for the conversion of glyceride groups at 200 C in the
absence of catalyst was about 8%.
Fig. 1. Conversion of glyceride groups vs. log βDBM. Oil (FFA 0.2%,
2
+
w/w) = 2.0 g, methanol = 0.88 g, M = 0.057 ± 0.002 mmol.

1
14
M. Di Serio et al. / Journal of Molecular Catalysis A: Chemical 239 (2005) 111–115
reach a maximum range with the increase in metal acidity
and then falls. In our case, the maximum is in the range of
the log βDBM between 8.60 (corresponding to cadmium) and
1
0.23 (corresponding to zinc).
However, in the transesterification reactions of TG with
methanol other factors must also be considered. For example,
the reaction mixture giving rise to a multiphase reaction sys-
tem leads to the partition of the catalytic species between the
two phases. In order to investigate the influence of this aspect,
as the catalyst phase partition can be related to the molecular
weight of the catalyst anion, runs were also performed by
using metal stearates instead of acetates in the same condi-
tion. In Fig. 1, the obtained conversions of glyceride groups
are reported. The results of Fig. 1 clearly show the influence
of the used anion. In all cases we observed an increase in the
catalytic activity for stearate metals with respect to acetates,
with the exclusion of Ni that remained inactive in both cases.
In particular, barium increases the conversion of glyceride
groups by nearly four times, while calcium and cobalt by
nearly two times.
Fig. 2. Influence of water and FFA on glyceride groups and FFA conver-
◦
sions for different acetates used (T = 200 C, oil = 2.0 g, methanol = 0.88 g,
The nature of the anion, when it gives place to a strong
interaction with the cation, as in the case of the catalysts
reported by Abreu et al. [16], can also change the order
of metal activity. In fact, by using metal complexes of 3-
hydroxy-2-methyl-4-pyrone, Abreu et al. found the following
activity order:
2
+
M
= 0.057 ± 0.002 mmol).
◦
2+
only the oil acidity (T = 200 C, M = 0.057 ± 0.002 mmol,
oil (FFA = 10.88%, w/w) = 2.0 g, methanol = 0.88 g). The
acid oil was prepared by adding oleic acid to the soybean
oil. A run without catalyst in the presence of high FFA con-
centration was done, to measure the conversion of glyceride
groups and FFA under these reaction conditions (see Fig. 2).
It must be pointed out that the conversion of glyceride
groups in the run without catalyst increases a little in the case
of transesterification of acid oil with respect to the transes-
terification of neutral oil. This increase is probably due to the
catalytic effect exerted by the presence of FFA in the reaction
medium (see Fig. 2). Together with transesterification, also a
non-catalytic esterification reaction occurs. In fact, at the end
of the reaction an FFA conversion of nearly 0.6 is reached
(see Fig. 2).
Sn2 ꢀ Zn2+ > Pb = Hg
+
2+
∼
i.e., in this case the catalyst activities followed their metal
acid strength monotonically.
Runs where then performed at a lower temperature
◦
(
150 C) using acetates, to study the influence of this parame-
ter. AscanbeseeninFig. 1, temperaturehasastronginfluence
◦
on catalytic activity. As a matter of fact, at 150 C barium and
calcium acetates show no activity while the yields obtained
with the metal having log βDBM in the range 8.60–10.23 are
in the range 18–44%. It is important to point out that the
influence of temperature is linked not only to the values of
the activation energy of the reaction in the presence of the
different metals, but also with the influence of temperature
on catalyst phase distribution.
To investigate the effect of water on catalytic activ-
ity, three runs were carried out with the presence of an
excess of water (water/methanol = 0.07, w/w) under the same
The effect of water on the catalytic activity of tested metals
is related to its concentration. As matter of fact, the conver-
sion of glyceride groups is poorly affected by FFA presence,
notwithstanding the presence of water formed by the esteri-
fication reaction at the same time (see Fig. 2).
The above-mentioned catalysts also showed a catalytic
activity in the esterification reaction. In fact, the FFA conver-
sion is higher than that obtained without catalyst (see Fig. 2).
The results described were confirmed with runs performed
◦
◦
conditions of the runs performed at 200 C (T = 200 C,
M
2
+
= 0.057 ± 0.002 mmol, oil (FFA = 0.2%, w/w) = 2.0 g,
methanol = 0.88 g) using calcium acetate and lead acetate as
the catalyst, respectively. Water has a strong depressive effect
on the activity of all the investigated metals, as can be seen
in Fig. 2, where the obtained conversions of glyceride groups
are reported. This effect can be attributed to the interaction of
water with the cation of the catalysts which decreases their
acid strength.
3
in a 1 dm autoclave reactor. The runs were performed at a
◦
temperature of 210–220 C with acid oil – oil (FFA = 7.5%,
w/w) = 250 g, methanol = 114 g – in the absence of cata-
lyst, in the presence of 1 g of lead acetate and of the Basu
and Norris catalytic mixture (Ca/Ba acetate) [7], respec-
tively. The obtained results are reported in Fig. 3. As can
be seen, the activity of lead acetate is higher than that of the
Basu and Norris catalytic mixture, in the transesterification
reaction.
The activities of these catalysts in the presence of water
were tested also in the presence of high FFA concentrations.
The runs were performed under the same conditions changing

M. Di Serio et al. / Journal of Molecular Catalysis A: Chemical 239 (2005) 111–115
115
Notwithstanding the fact that the activity of the best
catalyst Pb(Ac)2 is lowered by the water formed during
esterification of FFA, by using a low catalyst concentration
−
4
(
4 × 10 :1 weight ratio of catalyst to oil) it is possible to
obtain high FAME yields (96%) and a low final FFA concen-
tration (<1%), in a relatively short reaction time (200 min).
We can conclude, therefore, that this class of catalyst could
be used for a single stage process [17], in which the final ester
phase could be purified by FAME distillation, as suggested
by Basu and Norris [7] and by Zhang et al. [6]. The residual
(triglycerides, diglycerides, monoglycerides, catalyst) can
be re-circulated to the esterification–transesterification
reactor, while the glycerol is obtained, after treatment with
exchanges resin to eliminate metal traces. Glycerol obtained
is clearer than the one obtained with the classical technology
based on the use of alkaline catalysts giving place to soaps
forming emulsion difficult to remove. The proposed catalyst
have greater activity than that proposed by Basu and Norris
Fig. 3. Yields of FAME (%) and FFA concentration over the time
◦
(T = 210–220 C, oil = 250.0 g, methanol = 114 g).
[7]. Moreover, with respect to the process suggested by
Zhang et al. [6] using sulphuric acid for promoting both
esterification and transesterification this process has the
advantages of using a lower molar methanol/oil ratio (12/1
instead 50/1) and not necessitate an heavy neutralization
process of the acid catalyst.
Over time the glycerides conversion reaches a maxi-
mum value then decreases, in the case of catalyzed runs.
This behaviour was also observed by Abreu et al. [16] who
explained the observed decrease over time as the intervention
of a reverse reaction favoured by the formation of a two-phase
system, with its consequent different catalyst partition. How-
ever, much greater research efforts will be done to explain
this behaviour, in the future.
Acknowledgement
An other run was conduced in autoclave to verify the
performance of the best catalyst (Pb stearate) founded in
Thanks are due to ASER srl for financial support and col-
laboration.
the screening. The runs was performed at a temperature of
◦
2
10–220 C with acid oil – oil (FFA = 7.4%, w/w) = 250 g,
methanol = 114 g – in the presence of only 0.1 g of lead
stearate. The autoclave was kept at 210–220 C for 200 min
References
◦
and then cooled at room temperature. The FAME phase was
separated by glycerine phase by means of a separator funnel.
After methanol distillation, we obtained 0.96 glyceride con-
version and 0.74% FFA final concentration. The comparison
of these results with those reported in Fig. 3, taking also into
account that lead stearate concentration is 10 times less than
lead acetate, confirms that Pb stearate is the best catalyst in
the transesterification of acid oils.
[1] M. Frangrui, M.A. Hanna, Bioresour. Technol. 10 (1990) 1.
[2] E. Loreto, Y. Liu, D.E. Lopez, K. Suwannakaran, D.A. Bruce, J.G.
Goodwin Jr., Ind. Eng. Chem. Res., Web Release Date: January 25,
2
005.
3] A.S. Ramadhas, S. Jayaraj, C. Muralledharan, Bioresour. Technol.
5 (2002) 253.
[4] L. Jeromin, E. Peukert, G. Wollomann, US Patent 4,698,186 (1987).
[
8
[5] T. Ono, K. Yoshiharu, US Patent 4,164,506 (1979).
[
6] Y. Zhang, M.A. Dub e` , D.D. McLean, M. Kates, Bioresour. Technol.
89 (2003) 1.
[
[
7] H.N. Basu, M.E. Norris, US Patent 5,525,126 (1996).
8] G.W. Parshall, S.D. Ittel, Homogeneous Catalysis, II ed., Wiley, New
York, 1992, p. 270.
4
. Conclusions
[9] M. Di Serio, B. Apicella, G. Greco, P. Iengo, L. Fiocca, R. Po, E.
As has been seen, bivalent cations are catalysts for
both transesterification and esterification reactions. Catalytic
activities are related to the Lewis acid strength of the metals
Santacesaria, J. Mol. Cat. A: Chem. 130 (1998) 233.
10] G. Gelbard, O. Br e` s, R.M. Vargas, F. Vielfaure, U.F. Schuchardt,
JAOCS 72 (1995) 1239.
11] ASTM D803-82 (colourimetric method).
[
[
[
(
which must have an optimal intermediate value) and to the
12] M.A. Dasari, M.J. Goff, G.J. Suppes, JAOCS 80 (2003) 189.
molecular structure of the anion. The best catalytic perfor-
mances were obtained with cation metals having a complex
stability constant with dibenzoilmetane (log βDBM) [9,13] in
the range between 8.60 (corresponding to cadium) and 10.23
[13] K. Tomita, H. Ida, Polymer 16 (1975) 185.
[14] G. Van Uitert, W.C. Fernelius, J. Am. Chem. Soc. 75 (1953) 2736.
[
[
15] G. Van Uitert, W.C. Fernelius, J. Am. Chem. Soc. 75 (1953) 3862.
16] F.R. Abreu, D.G. Lima, E.H. Ham u` , S. Einloft, J.C. Rubim, P.A.Z.
Suarez, JAOCS 80 (2003) 601.
(corresponding to zinc). Then the stearates have better per-
[
17] D. Siano, M. Di Serio, R. Tesser, M. Dimiccoli, F. Cammarota,
E. Santacesaria, L. Siano, M. Nastasi, Italian Patent Application
ITMI2004A001323, 2004.
formances than acetates because their higher solubility in the
oil phase where the reaction occurs.