Energy & Fuels 2002, 16, 443-450 443

Biodiesel Fuels from Vegetable Oils: Transesterification

of Cynara cardunculus L. Oils with Ethanol
J. M. Encinar,* J. F. González, J. J. Rodrı́guez, and A. Tejedor
Departamento de Ingenierı́a Quı́mica y Energética, Universidad de Extremadura, Avda. de
Elvas s/n, 06071 Badajoz, Spain

Received July 17, 2001. Revised Manuscript Received November 9, 2001

A study was made of the transesterification reaction of Cynara cardunculus L. oil by means of
ethanol, using sodium hydroxide and potassium hydroxide as catalysts. The objective of the work
was to characterize the ethyl esters for use as biodiesels in compression ignition motors. The
operation variables employed were temperature (25-75 °C), catalyst type (sodium hydroxide and
potassium hydroxide), catalyst concentration (0.25-1.5 wt %), and ethanol/oil molar ratio (3:1-
15:1). Oil mass (200 g), reaction time (120 min), and alcohol type (ethanol) were fixed as common
parameters in all the experiments. The evolution of the process was followed by gas chromatog-
raphy, determining the concentration of the ethyl esters at different reaction times. The biodiesel
was characterized by determining its density, viscosity, high heating value, cetane index, cloud
and pour points, characteristics of distillation, and flash and combustion points according to ISO
norms. The biodiesel with the best properties was obtained using an ethanol/oil molar ratio of
12:1, sodium hydroxide as catalyst (1%) and 75 °C temperature. This biodiesel has very similar
properties to those of no. 2 diesel fuel.

Introduction The heating value of vegetable oils is similar to that

Diesel fuels have an essential function in the indus-
trial economy of a country. They are used in city buses,
locomotives, electric generators, etc. The diesel fuel
consumption of developed countries has been increasing
steadily over the last few decades and looks set to
continue into the future. An alternative diesel fuel must
be technically feasible, economically competitive, envi-
ronmentally acceptable, and readily available.1 Given
these requirements, triglycerides (vegetable oils/animal
fats) and their derivatives may be considered as viable
alternatives for diesel fuel.2-5 Vegetable oils are widely
available from a variety of sources, and they are
renewable. Also, these fuels are easily biodegradable,
they have a practically null sulfur content, and their
transport and storage offer no problems. Moreover,
vegetable oils sequester more carbon dioxide from the
atmosphere during their production than they add to it
when they are burned. Therefore, they help to alleviate
the increasing carbon dioxide content of the atmosphere.
Their main drawback is price, which is higher than
petroleum-derived diesels. In consequence, their use
must be accompanied by a policy oriented toward their
total tax exemption.
* To whom correspondence should be addressed. Phone: 34 924
289672. Fax: 34 924 271304. E-mail: jencinar@unex.es.
(1) Srivastava, A.; Prasad, R. Renewable Sustainable Energy Rev.
2000, 4, 111-133.
(2) Shay, E. G. Biomass Bioenergy 1993, 4, 227-242.
(3) Schwab, A. W.; Bagby, M. O.; Friedman, B. Fuel 1987, 66, 1372-
1378.
(4) Freedman, B.; Butterfield, R. O.; Pryde, E. H. J. Am. Oil. Chem.
Soc. 1986, 63, 1375-1380.
(5) Freedman, B.; Pryde, E. H.; Mounts, T. L. J. Am. Oil. Chem.
Soc. 1984, 61, 1638-1643.
of fossil diesel, but their use in direct injection diesel
engines is limited by some of their physical properties,
especially their viscosity. Vegetable oil viscosity is 10
times greater than that of diesel oil, with consequent
poor fuel atomization, incomplete combustion, carbon
deposition on the injectors, and fuel build-up in the
lubricant oils. Therefore the result can be serious engine
deterioration so that it is absolutely necessary to subject
the vegetable oils to treatments that diminish their
viscosity. There are four possible treatments: dilution,
microemulsification, pyrolysis, and transesterification,1,3
among which the last has doubtless been the most
commonly employed.
The transesterification reaction,6 also called alcoholy-
sis, is the displacement of alcohol from an ester by
another alcohol in a process similar to hydrolysis, except
than an alcohol is used instead of water. The result is
that triglyceride molecules (90-98% of the oil), which
are long and branched, are transformed into smaller
esters whose size and properties are similar to those of
diesel oils.7 The process also forms glycerin as byprod-
uct. The overall transesterification reaction can be
represented by
catalyst
triglyceride (TG) + 3ROH 98
3R′CO2R + glycerin (1)
Transesterification consists of a sequence of three
consecutive reversible reactions. The first step is the
(6) Otera, J. Chem. Rev. 1993, 93, 1449-1470.
(7) Quick, G. R.; Woodmore, P. J.; Wilson, B. T. Engine Evaluations
of Linseed Oil and Derivatives. In Vegetable Oils Diesel Fuel: Seminar
III, ARM-NC-28; Bagby, M. O., Pryde, E. H., Eds.; U.S. Department
of Agriculture: Peoria, IL, 1983; p 138.

10.1021/ef010174h CCC: $22.00 © 2002 American Chemical Society

Published on Web 01/23/2002
444 Energy & Fuels, Vol. 16, No. 2, 2002
conversion of triglycerides to diglycerides, followed by
the conversion of diglycerides to monoglycerides, and
finally monoglycerides into glycerin, yielding one ester
molecule from each glyceride at each step. Stoichiomet-
rically, three moles of alcohol are required per mole of
triglyceride, but in practice a higher ratio is employed
in order to displace the equilibrium to greater ester
production. Although the esters are the desired products
of the reactions, glycerin recovery also is important due
to its numerous applications in different industrial
processes. Vegetable oils also contain free fatty acids
(generally 1-5%), phospholipids, phosphatides, car-
otenes, tocopherols, sulfur compounds and traces of
water.8 Consequently some of these components also can
be present in the final products.
The alcohols most frequently employed are short-
chain alcohols, such as methanol, ethanol, propanol, and
butanol. These alcohols present few differences with
respect to the kinetics and final yield of esters. Freed-
man et al.,5 using methanol, ethanol, and butanol in a
process of transesterification of soybean oil, obtained
yields from 96 to 98% after an hour of reaction. The use
of other types of oils modified the final yield: they
obtained 93% of methyl ester using cottonseed oil and
98% using soybean oil. These yields were independent
of the type of alcohol. Therefore, the eventual selection
of one of these three alcohols will be based on cost and
performance considerations. In general, both methanol
and ethanol can be easily obtained from plant materials.
Because methanol is also cheaper, it is the most com-
monly employed alcohol. Nevertheless, since ethanol, as
extraction solvent is preferable to methanol because of
its much superior dissolving power for oils, it also is
often used as an appropriate alcohol for the transes-
terification of vegetable oils.9 Producing ethyl esters
rather than methyl esters is of considerable interest
because it allows production of an entirely agricultural
fuel, and the extra carbon atom brought by the ethanol
molecule slightly increases the heat content and the
cetane number. Another important advantage in the use
of ethanol is that the ethyl esters have cloud and pour
points that are lower than the methyl esters. This fact
improves the starts cold.
Alkaline alkoxides and hydroxides are the most effec-
tive transesterification catalysts compared to the acid
catalysts. Transesterification occurs at a faster rate in
the presence of an alkaline catalyst than in the presence
of the same amount of acid catalyst.10 Sodium alkoxides
(sodium methoxide for example), KOH, and NaOH are
the most efficient catalysts used for this purpose. They
require only short times (30 min) to complete the
reaction even at room temperature, while acid catalysts,
such as sulfuric acid, require higher temperatures (100
°C) and longer reaction times (3-4 h)5,11,12
The commercial development of a transesterification
process requires abundant and inexpensive raw mate-
(8) Marckley, K. S. Fatty Acids, 2nd ed.; Interscience: New York,
1960.
(9) Lago, R. C. A.; Szpiz, R. P.; Jablonka, F. H.; Pereira, D. A.;
Hartman, L. Oleagineux 1985, 40, 147-152.
(10) Formo, M. W. J. Am. Oil. Chem. Soc. 1954, 31, 548-559.
(11) Nye, M. J.; Southwell, P. H. Esters from Rapeseed Oil as Diesel
Fuel. In Vegetable Oils Diesel Fuel: Seminar III, ARM-NC-28; Bagby,
M. O., Pryde, E. H.; Eds.; U.S. Department of Agriculture: Peoria, IL,
1983; p 78.
(12) Harrington, K. J.; D’Arcy-Evans, C. Ind. Eng. Chem. Prod. Res.
Dev. 1985, 24, 314-318.
Encinar et al.
rial to be available. One possibility is to use foodstuff
oils such as sunflower, soybean, etc., but these oils are
very expensive. Another is to produce the oils (the
biomass in general) directly, i.e., to cultivate nonfood
crops for energy. This latter possibility has been con-
sidered by the authorities of the European Union as a
solution to the abandonment of traditional croplands.
One of the crops used to produce woody biomass and
oil is the cardoon, Cynara cardunculus L. This plant is
being studied experimentally in Spain in order to
determine its optimal growth conditions. The cardoon
is a thistle native to the Mediterranean area. It belongs
to the Asteraceae family (Compositae), which also
includes the artichoke, the sunflower, and the safflower.
In the Iberian Peninsula it is well adapted to the climate
and can be found growing wild. Total cultivated biomass
production can reach 20-30 tons dry matter per ha per
year, including 2000-3000 kg of seeds. It is used as
fodder, or as biomass for energy production by combus-
tion, pyrolysis, and gasification. Its seeds have a high
protein content, close to 15%, and an oil content of
around 25%. This oil is similar in composition to
sunflower oil,13 the four principal fatty acids being
palmitic (11-14%), stearic (3%), oleic (25%), and linoleic
(56%). At this time, the cultivation of Cynara in Spain
is under development in several regions. The Spanish
Government (Official Bulletin of the State 121 of May
20, 2000), and the Commission of Agriculture of the
European Union (Regulation CE 246/99, code NC
06029059), have authorized the cultivation of Cynara
in substitution of traditional croplands.
With these considerations, and as the continuation
of previous work,14 we carried out a study of the
transesterification process of refined Cynara carduncu-
lus L. oil using ethanol, to characterize the ethyl esters
obtained with a view to their use as biodiesel for
compression ignition engines.
Experimental Section
Materials. The Cynara cardunculus L. oil was obtained in
our laboratory by mechanical compression and then refined.
Anhydrous ethanol, sodium hydroxide, and potassium hydrox-
ide were supplied by Panreac. Ethyl esters of palmitic, stearic,
oleic, linoleic, linolenic, and erucic acids (employed as stan-
dards in the chromatographic determination) were supplied
by Merck and Sigma. All reagents were of analytical grade.
Transesterification. The reaction of transesterification
was carried out in a 500 mL spherical reactor, provided with
thermostat, mechanical stirring, sampling outlet, and conden-
sation system. This arragement is basically the same as that
described in the literature.15 The procedure followed (modifica-
tion of that used in previous work14) is described next. The
system was preheated to 75 °C, to eliminate moisture, and
then 200 g of Cynara cardunculus L. oil was added. When the
system reached 75 °C again, the ethanol and the catalyst were
added, in the amounts established for each experiment, and
the stirring system was connected, taking this moment as time
zero of the reaction. At evenly spaced intervals, 2 cm3 of sample
were withdrawn for later chromatographic analysis. Each
experiment was prolonged for 120 min, by which time the
(13) Benjelloun-Mlayah, B.; López, S.; Delmas, M. Ind. Crops Prod.
1997, 6, 233-236.
(14) Encinar, J. M.; González, J. F.; Sabio, E.; Ramiro, M. J. Ind.
Eng. Chem. Res. 1999, 38, 2927-2931.
(15) Ma, F.; Clements, L. D.; Hanna, M. A. Ind. Eng. Chem. Res.
1998, 37, 3768-3771.
Biodiesel Fuels from Vegetable Oils Energy & Fuels, Vol. 16, No. 2, 2002 445

conversion to esters was complete. After cooling, 25% of

glycerin based on the weight of oil was added, which resulted
in the formation of an upper phase consisting of ethyl esters
and a lower phase containing the liberated and added glycerin,
the excess of ethanol, the unreacted sodium, or potassium
hydroxide together with the soaps formed during the reaction
and some entrained ethyl esters and partial glycerides. After
separating the two layers by sedimentation, the ethyl esters
were purified by distilling the residual ethanol at 80 °C. The
remaining catalyst was extracted by successive rinses with
distilled water. Finally, the water present was eliminated with
CaCl2 followed by filtration. The lower phase was acidified with
a calculated amount of sulfuric acid, to neutralize any unre-
acted sodium or potassium hydroxide and to decompose the
soaps formed during the transesterification. The resulting
mixture was subjected to a distillation at 80 °C under a
moderate vacuum (absolute pressure of 150 mmHg) to recover
the excess of ethanol. This facilitated the separation of glycerin
from entrained ethyl esters and fatty acids derived from soaps.
Analysis. The European Organization for Normalization
(CEN) specifies the criteria that should be satisfied by a
biodiesel of high quality, or diesel and biodiesel mixtures, for
use in motor vehicles.16 The analytical methods used in this
work to determine the characteristics of the biodiesel are
basically those recommended by the cited organization. The
ethyl ester content was assayed by gas chromatography in an
HP 5890 chromatograph equipped with a FID, employing a
silica capillary column of 50 m length and 0.22 mm i.d.
(phase: BPx70). Hexane was used as solvent and the carrier
gas was nitrogen. Ethyl esters of palmitic, stearic, oleic, Figure 1. Ester concentrations vs time. Influence of temper-
linoleic, linolenic, and erucic acids were analyzed by this ature ([NaOH] ) 1%, ethanol/oil ) 9:1).
procedure. The following parameters were determined in the
final biodiesel product: density (pycnometry), viscosity (Brook-
field digital viscosimeter), high heating value (Parr-1351 bomb
calorimeter, ISO 1928 norm), cetane index (ASTM D 976
norm), cloud point (ISO 3015 norm), pour point (ISO 3016
norm), Ramsbottom carbon residue (ISO 4262 norm), distil-
lation characteristics (ISO 3405 norm), and flash and combus-
tion points (ASTM D-92 norm).

Results and Discussion

The operating variables employed were temperature
(25-75 °C), catalyst type (sodium hydroxide and potas-
sium hydroxide), catalyst concentration (0.25-1.5 wt %
referred to the total mixture, i.e., oil and ethanol), and
ethanol/oil molar ratio (3:1 to 15:1). Oil mass (200 g),
reaction time (120 min), and alcohol type (ethanol) were
fixed as common parameters in all the experiments.
Influence of Temperature. Alkaline alcoholysis of
vegetable oils is normally performed near the boiling
point of the alcohol. Nevertheless, the reaction may be
carried out at room temperature.1,5 We studied the
ethanolysis of the refined Cynara cardunculus oils at
25, 50, and 75 °C in order to determine the effect of
reaction temperature on the ethyl ester formation. In
all experiments, an ethanol/oil molar ratio of 9:1 and
1% NaOH (as catalyst) were used. Figure 1 shows the
temporal evolution of the results. After 2 min, the esters
present in the 75, 50, and 25 °C runs were 79, 76.5, and
75.5%, respectively, showing the influence of tempera- Figure 2. Evolution of ethyl ester concentrations with time
ture on ester conversion. However, after 2 h, ester (T ) 75 °C, [NaOH] ) 1%, ethanol/oil ) 9:1).
formation was basically identical for the three runs Figure 2 show the evolution of the six main ethyl
(93.1, 92.5, and 91.6%, respectively). Kinetically, the esters over the course of the reaction in an experiment
reaction was very fast. As can be observed in Figures carried out at 75 °C. The main ethyl ester is the linoleate
1, the final ethyl ester concentration was almost reached with a percentage of 55%. Oleate (25%), palmitate (10%),
in 10 min and the curves have an asymptotic tendency and stearate (5%) follow in importance. Finally, the
with time. ethyl esters corresponding to the linolenic and erucic
(16) Mordret, F. Ol., Corps Gras, Lipides 1994, 1, 23-24. acids are present at percentages less than to 1%. This
446 Energy & Fuels, Vol. 16, No. 2, 2002 Encinar et al.

Table 1. Influence of Temperature on Characteristic obtained. Hence, the same yield may be obtained at
Parameters of the Process ([NaOH] ) 1%, room temperature as at higher temperatures by simply
Ethanol/Oil ) 9:1)
extending the reaction time. Industrially, this could
temperature, °C mean energy savings and lower process costs.
parameter 25 50 75 Influence of Catalyst Type and Concentration.
density (25 °C), kg m-3 870 870 870 As was noted above, alkaline metal alkoxides and
viscosity, cSt hydroxides are the most effective transesterification
15 °C 8.61 8.48 8.30 catalysts, although for economic reasons sodium hy-
20 °C 7.47 7.04 7.35
25 °C 6.89 6.32 5.77 droxide is more often used.5 An alkaline catalyst con-
30 °C 6.32 5.60 4.57 centration in the range of 0.5-1 wt % yields 94-99%
40 °C 4.59 4.17 3.43 conversion of vegetable oil into esters.20 Further in-
HHV, MJ kg-1 39.9 40.0 40.0 crease in catalyst concentration does not increase the
cloud point, °C -3 -4 -3
pour point, °C -7 -7 -6 conversion and leads to extra costs because it is neces-
Flash point, °C 185 185 188 sary to remove it from the reaction medium at the end.
combustion point, °C 196 190 194 As has been indicated, acid catalysts, such as sulfuric
yield of esters, % 91.6 92.5 93.2 acid, require higher temperatures and longer reaction
cetane index 49.2 49.2 49.1
distillation, °C times. There is one clear advantage in using an acid
0% 295 285 295 catalyst instead of an alkaline catalyst: if the vegetable
10% 345 310 346 oil had a free fatty acid content of >1%, the alkaline
20% 348 348 350 catalyst would be destroyed, whereas the acid catalyst
30% 349 351 352
40% 350 353 353 would still be effective. Nevertheless, for refined oil such
50% 351 355 354 as that employed in this work, with free fatty acid
Ramsbottom residue, % 0.32 0.33 0.29 contents of <1%, alkaline catalysts are preferred for the
sulfur (% mass) <0.01 <0.01 <0.01 indicated reasons. Moreover in the literature there are
examples of uses of other catalysts such as p-toluene
composition is very similar to that of soybean oil,17,18 sulfonic acid,21 heterogeneous catalysts consisting of
with the exception of the linolenate whose percentage mixtures of CaO, MgO, and ZnO,22 and natural catalysts
in that oil is 8%. On the other hand, rapeseed oil19 has such as coconut ash or palm ash.23
a very different composition, with the oleate (60%) being In a previous work,14 where sodium hydroxide, potas-
the main ester. It appears that the reaction is very rapid sium hydroxide, and sodium methoxide were compared,
because the ethyl ester conversion is near the maximum it was demonstrated that the results obtained with
value after only 5-10 min, and stabilizes at the maxi- these three catalysts were basically the same. In the
mum value after 20-30 min. These results agree with present work, on account of the economic arguments,
those obtained by Fillières et al.19 Those authors only two catalysts were tested: sodium and potassium
indicate that at equilibrium there are di- and monogly- hydroxides, varying their concentration in the range
cerides with concentrations of 2 and 4%, respectively. between 0.25 and 1.5 wt %. Also, an experiment without
This equilibrium can be slightly displaced if the glycerin catalyst was carried out, although in this case the ester
layer is withdrawn from the reaction medium after 10- conversion was 0% after 120 min so that the properties
15 min and new ethanol and catalyst are added. of the final oil were the same as those of the initial oil.
After the initial period where the reaction was very In Figure 3, the percentages of esters obtained are
fast, there was a second period, much longer than the plotted versus reaction time for the experiments in
first, in which the composition evolved slowly toward which the concentration of sodium hydroxide was
equilibrium. In consequence, as was indicated above, varied. As can be observed, the best results were
this shows that the transesterification may take place achieved for a concentration of 1.0%. For higher values
satisfactorily at room temperature, in agreement with the yields were lower. This fact seems to be related to
results described by other authors.1,5 the free acidity of the oil. When there is a large free
Table 1 lists the results for the influence of temper- fatty acid content, the addition of more sodium hydrox-
ature on the most representative parameters of the ide, or any other alkaline catalyst, compensates this
process. As was to be expected, since the yields in the acidity and avoids catalyst deactivation.4,5 Nevertheless,
three experiments were very similar, the differences for refined oil with free fatty acid contents of <1%, the
between the values of the parameters were not very addition of an excessive amount of alkaline catalyst
significant. There was a sharp fall in viscosity and an gives rise to the formation of an emulsion, which
increase in yield as the temperature nose, while the increases the viscosity and leads to the formation of gels.
other variables basically remained constant. Therefore These hinder the glycerin separation, and as a result
we may state that the temperature is linked to reaction there is a dilution of the ethyl esters, and the apparent
time and that the final properties of the product ester yield diminishes. The result of these two opposing
depended on the degree of conversion, i.e., on the time effects is an optimal catalyst concentration that in this
required to obtain a given percentage of ester rather
than on the temperature at which this percentage was (20) Krisnamgkura, K.; Simamaharnnop, R. J. Am. Oil Chem. Soc.
1992, 69, 166-169.
(17) Chang, D. Y. Z.; Van Gerpen, J. H.; Lee, I.; Johnson, L. A.; (21) Fuls, J.; Hugo, F. J. C. Third International Confernce of Energy
Hammond, E. G.; Marley, S. J. J. Am. Oil. Chem. Soc. 1996, 73, 1549- Use Management, Berlin, 1981; Pergamon Press: Elmsford, NY, 1981;
1555. p 1595.
(18) Clark, S. J.; Wagner, L.; Schrock, M. D.; Piennaar, P. G. J. Am. (22) Peterson, G. R.; Scarrah, W. P. J. Am. Oil Chem. Soc. 1984,
Oil. Chem. Soc. 1984, 61, 1632-1638. 61, 1593-1597.
(19) Fillieres, R.; Benjelloun-Mlayeh, B.; Delmas, M. J. Am. Oil. (23) Graille, J.; Lozano, P.; Pioch, D.; Geneste, P. Oléagineux 1985,
Chem. Soc. 1995, 72, 427-432. 40, 271-276
Biodiesel Fuels from Vegetable Oils Energy & Fuels, Vol. 16, No. 2, 2002 447

Figure 3. Ester concentrations vs time. Influence of [NaOH]

(T ) 75 °C, ethanol/oil ) 9:1). Figure 4. Ester concentrations vs time. Influence of [KOH]
(T ) 75 °C, ethanol/oil ) 9:1).
Table 2. Influence of NaOH Concentration on
Characteristic Parameters of the Process (T ) 75 °C, of the process. Basically, it can be observed that as the
Ethanol/Oil ) 9:1)
catalyst concentration increased there was a big reduc-
% NaOH by weight tion in both viscosity and cloud point. There also was a
parameter 0 0.25 0.50 0.75 1.0 1.5 reduction, although to a lesser extent, in density. The
density (25 °C), kg m-3 917 890 880 870 870 877 minimum for these three parameters occurred for a
viscosity, cSt concentration of 1.0% NaOH. Also, the maximum yield
15 °C 115 14.72 10.29 8.58 8.30 8.48
20 °C 95 12.85 9.71 8.00 7.35 7.75
in ethyl esters occurred for a concentration of 1% NaOH.
25 °C 85 11.30 9.00 7.43 5.77 6.78 The pour, flash, and combustion points, cetane index,
30 °C 74.9 9.60 7.88 6.86 4.57 5.72 high heating value, and Ramsbottom residue did not
40 °C 57 7.83 6.14 5.00 3.43 4.51 vary significantly. On the other hand, the values of the
HHV, MJ.kg-1 33.0 39.2 39.7 39.8 40 39.7
cloud point, °C -7 2 0 -1 -3 -2 previous parameters were very significant in the run
pour point, °C -10 -6 -7 -8 -6 -7 performed without catalyst (0% of NaOH). Among these
flash point, °C 350 186 187 192 188 190 parameters, the viscosity had values 10 times higher
combustion point, °C 355 191 192 197 194 196
yield of esters, % 0 77.5 84.5 85 93.2 88.9
that in the other experiments. In summary, in the case
cetane index 51 49.3 49.2 49.4 49.1 49.2 of sodium hydroxide the best concentration was 1.0%.
distillation, °C The trends found when varying the concentration of
0% 245 298 300 304 295 301 KOH from 0% to 1.5 wt % were basically the same as
10% 269 346 348 349 346 347
20% 276 349 350 353 350 351 those found for sodium hydroxide. The optimal concen-
30% 280 350 351 354 352 353 tration for this catalyst also was 1%. Figure 4 shows
40% 283 351 352 355 353 354 the temporal evolution of the results. Qualitatively this
50% 285 352 353 355 354 354 evolution was very similar to that commented on for
Ramsbottom residue, % 0.42 0.32 0.32 0.30 0.29 0.31
sulfur (% mass) <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 Figure 3. Nevertheless, on this occasion, the yields
reached for the concentrations of 0.25 and 0.5% of
case is 1.0% of NaOH. One way or another, the mini- catalyst were very low. The other parameters evolved
mum concentration of catalyst required is hard to in a similar way to that indicated in Table 2 for the case
evaluate because a compromise must be made between of NaOH.
duration of reaction and catalyst concentration. Indeed, Table 3 lists the biodiesel parameters obtained with
the curves of Figure 3 suggest that at very long reaction the optimal concentration of the two catalysts, including
times the same equilibrium conversion would be reached for comparison the characteristics of a typical diesel (no.
in all runs. However, the time of reaction would increase 2 diesel) according to Schwab et al.’s data3 and a
when the catalyst concentration decreases. These re- biodiesel obtained by means of transesterification of
sults were qualitatively similar to those obtained in the Cynara oil with methanol.14 As can be observed, there
ethanolysis of rapeseed oil.19 was a good approximation of the different biodiesels to
Table 2 lists the results for the influence of NaOH the characteristics of no. 2 diesel. The greatest differ-
concentration on the most representative parameters ences were in the cloud and pour points, flash, and
448 Energy & Fuels, Vol. 16, No. 2, 2002 Encinar et al.

Table 3. Influence of Catalyst Type on Characteristic Parameters of the Process: Comparison with a No. 2 Diesel and
Methyl Esters of Cynara Oil
ethyl esters of ethyl esters methyl esters
parameter Cynara oil [NaOH])1% of Cynara oil [KOH])1% of Cynara oila no. 2 diesel
density (25 °C), kg m-3 870 870 880 847
viscosity (40 °C), cSt 3.43 4.59 3.56 2.39
HHV, MJ kg-1 40.0 40.0 39.8 45.2
cloud point, °C -3 -2 -1 -19
pour point, °C -6 -8 -3 -23
flash point, °C 188 185 175 85
combustion point, °C 194 193 179 92
yield of esters, % 93.2 90.4 94
cetane index 49.1 48.3 48.6 46
distillation, °C
0% 295 308 315 258
20% 350 350 345
40% 353 353 348
50% 354 355 350
Ramsbottom residue, % 0.29 0.28 0.26 0.14
sulfur (% mass) <0.01 <0.01 <0.01 0.25
a Catalyst: sodium methoxide; concentration of catalyst ) 1 wt %; methanol/Cynara oil ) 5:1.

combustion points, the viscosity, and Ramsbottom resi- Table 4. Comparison of Ethyl Esters of Cynara,
due which were all lower than no. 2 diesel. On the Sunflower, Rape, Groundnut, and Soybean Oils
contrary, the cetane index was higher in the biodiesels. sun- ground- soy-
This parameter guarantees that there will be good parameter Cynara flower rape nut bean
control of combustion, increasing performance and density (25 °C), kg m-3 870 873 902 906 881
improving cold starts, which will lead to less contami- viscosity, cSt
nanting exhaust gases.24 In the production of diesel 20 °C 7.35 6.97
21 °C 6.60
fuels, the cetane number is an important indicator of 38 °C 14.1 17.0
the quality of the product and is usually measured by 40 °C 3.43 4.55 4.41
using a standard engine test (ASTM D613). However, HHV, MJ kg-1 40.0 38.9 38.7 38.3 40.0
it is relatively difficult to measure and has rarely been cloud point, °C -3 -1 8 23 1
pour point, °C -6 -8 -4
determined for vegetable oils and fatty acid esters. In flash point, °C 188 187 191 195
this present work, the ASTM Standard D976 was combustion point, °C 194 192
applied, using boiling point and density for the calcula- cetane index 49.1 49.4 48.2
tion of the cetane index. This method is very well suited distillation, °C
to the routine testing of small volumes or when rapid 40% 353 352 336
50% 354 353 344
indications are required of the quality of a diesel fuel, 60%
and yields similar results to those obtained using more 90%
sophisticated methods.25,26 The energy content of the Ramsbottom residue, % 0.27 0.30 0.69
esters, based on their heating value, was around 90% sulfur (% mass) <0.01 <0.01 0.1
of no. 2 diesel taking the mass as reference, and 95%
taking the volume as reference. This not very significant and the low sulfur content, are better than no. 2 diesel
difference was because the esters contained 10 wt % of oil from the viewpoint of handling safety and the
oxygen that could be used in the combustion process, emission of contaminant gases during combustion.
diminishing the external air requirements. Comparison Moreover, due to their plant origin, these fuels would
of the two ethyl esters with the methyl ester shows that contribute to maintaining atmospheric CO2 equilibrium.
all three had very similar characteristics. Only slight Influence of Ethanol/Oil Molar Ratio. Another
differences were observed in the flash and combustion important variable affecting the yield of ester is the
points, which were lower in the methyl ester, and in molar ratio of alcohol to vegetable oil. As noted above,
the pour point, which was higher in the methyl ester. the reaction stoichiometry requires three moles of
These characteristics represent advantages of ethyl alcohol per mole of triglyceride to yield three moles of
esters over methyl esters since they improve cold starts fatty esters and one mole of glycerin. Because this is
and safety in handling. an equilibrium reaction, an excess of alcohol will
Table 4 lists the properties of sunflower, rape, ground- increase the ester conversion by shifting this equilibri-
nut, and soybean ethyl esters3,9,24 together with the um to the right. However, the higher molar ratio of
results obtained in this work employing 1% of sodium alcohol to vegetable oil interferes with the separation
hydroxide as catalyst. As can be seen, the results are of glycerin because there is an increase in solubility. The
very homogeneous, and the comparison with those of excess of alcohol seems to favor conversion of di-to
no. 2 diesel oil (Table 3) makes it clear that any of these monoglycerides, but there is also a slight recombination
ethyl esters might be used as an alternative to fossil of esters and glycerin to monoglycerides because their
fuels. Some characteristics, such as the high flash point concentration keeps increasing during the course of the
reaction, in contrast to reactions conducted with low
(24) Reed, T. B.; Graboski, M. S.; Gaur, S. Biomass Bioenergy, 1992, molar ratios.19 Krisnangkura and Simamaharnnop20
3, 111-115. also observed that when glycerin remained in solution
(25) O’Connor, C. T.; Forrester, R. D.; Scurrell, M. S. Fuel 1992,
71, 1323-1327. it helped drive the equilibrium back to the left, lowering
(26) Krisnangkura, K. J. Am. Oil. Chem. Soc. 1986, 63, 552-553. the yield of esters. In the present work, we used molar
Biodiesel Fuels from Vegetable Oils Energy & Fuels, Vol. 16, No. 2, 2002 449

Table 5. Influence of Ethanol/Oil Molar Ratio on

Characteristic Parameters of the Process (T ) 75 °C,
[NaOH] ) 1%)
ethanol/oil molar ratio
parameter 3:1 6:1 9:1 12:1 15:1
kg.m-3
density (25 °C), 893 870 870 870 872
viscosity, cSt
15 °C 15.35 11.44 8.30 8.01 9.65
20 °C 13.91 10.30 7.35 7.01 8.50
25 °C 11.25 8.58 5.77 6.29 7.10
30 °C 10.32 7.44 4.57 5.15 6.30
40 °C 8.43 5.72 3.43 4.58 4.69
HHV, MJ kg-1 39.6 39.8 40 39.9 39.8
cloud point, °C 1 -1 -3 -3 -3
pour point, °C -7 -9 -6 -7 -6
flash point, °C 195 184 188 190 187
combustion point, °C 199 191 194 197 193
yield of esters, % 62.1 88.2 93.2 94.5 89.8
cetane index 49.2 49.0 49.1 48.4 49.0
distillation, °C
0% 283 301 295 297 300
10% 325 347 346 344 345
20% 327 351 350 348 349
30% 328 353 352 351 352
40% 332 354 353 354 354
50% 334 355 354 354 355
Ramsbottom residue, % 0.32 0.30 0.29 0.31 0.31
sulfur (% mass) <0.01 <0.01 <0.01 <0.01 <0.01

to ethyl esters. This multistage transesterification

Figure 5. Ester concentrations vs time. Influence of ethanol/
oil molar ratio (T ) 75 °C, [NaOH] ) 1%).
ratios of ethanol to Cynara oil between 3:1 and 15:1.
Figure 5 shows the ester percentages obtained versus
the reaction time for the five experiments carried out.
As can be observed, with a stoichiometric amount of
alcohol, the conversion to esters was near 62% after 2h.
Under these conditions, the separation of the glycerin
layer was fast, compared to reactions conducted with a
large excess of ethanol where the separation takes time,
is not complete, and increases the cost of alcohol
recovery. The ester yields increased as the molar ratio
increased up to a value of 12:1. The best results were
for molar ratios between 9:1 and 12:1. For molar ratios
less than 6:1, the reaction was incomplete. For a molar
ratio of 15:1 the separation of glycerin is difficult and
the apparent yield of esters decreases because part of
the glycerin remains in the biodiesel phase. Conse-
quently, molar ratios higher that 9:1 do not increase the
ester yields considerably. Therefore, it does not seem
recommendable-to-use ratios of 12:1 since the extra
consumption of alcohol is not compensated by the
increase in reaction yield. Moreover, an increase in
alcohol complicates recovery of the glycerin, since this
is partially soluble in ethanol. Also, the separation of
the esters and the recovery of the alcohol would be more
expensive. Therefore, the molar ratio 9:1 seems to be
the most appropriate.
As has been indicated, the reaction is very fast. Values
near the final conversion were reached in the first 10
min. As is seen in Figure 5, further displacement of the
equilibrium toward the formation of esters became very
slow, so that, even when there was an excess of alcohol,
this did not improve the equilibrium conversion sub-
stantially. One way of improving the process would be
to withdraw the glycerin from the reaction medium as
soon as it is formed, and to carry out successive
transesterifications until reaching complete conversion
process would allow one to work with smaller ethanol/
oil molar ratios.
Table 5 lists the results for the influence of the
ethanol/oil molar ratio on some of the most character-
istic parameters of the process. As can be observed, the
lowest viscosity values were obtained for an ethanol/oil
molar ratio between 9:1 and 12:1. With regard to the
cloud and pour points, the latter remained almost
constant while the former markedly improved as the
ethanol/oil molar ratio increased. However, the high
heating values, and flash and combustion points were
basically independent of the molar ratio. Therefore, the
behavior of these parameters also seems to ratify that
the molar ratio 9:1 is the most appropriate for the
transesterification of cardoon oil.
Final Considerations. In this work, the variables
affecting the ethanolysis of Cynara cardunculus L. oil
were studied. The study was carried out under labora-
tory-scale conditions. Consequently, it is hard to choose
the optimum operation process conditions. An ethanol/
oil molar ratio 9:1 or 12:1, 1% of NaOH, and 75 °C gave
the best results. Nevertheless, these conditions are not
necessarily the same for larger-scale operations, and
economic compromises must be made in choosing the
parameters for operation.
A very important variable in the transesterification
processes is the ethanol/oil molar ratio. This molar ratio
affects the separation and recovery of the glycerin. A
molar ratio of 6:1 is generally considered the most
appropriate, although in this present work, as has been
indicated, we found a molar ratio between 9:1 and 12:1
to be the best. Nevertheless, the results are quantita-
tively very similar to those of the literature.19 For
example, in the ethanolysis of rapeseed oil, yields
between 91 and 94% are obtained for ethanol/oil molar
ratio between 12:1 and 6:1. We have obtained yields
between 88.2 and 94.5 for molar ratios between 6:1 and
12:1. The ethanol/oil molar ratio also affects the distri-
bution of the catalyst between the ester layer and
450 Energy & Fuels, Vol. 16, No. 2, 2002
glycerin layer. Junek and Mittelbach27 show that for a
3:1 stoichiometric ethanol/oil molar ratio most of the
catalyst is drawn into the glycerin layer in contrast with
an excess of alcohol where the catalyst is evenly
distributed in the ester layer and glycerin layer.
The objective of this work was to characterize the
ethyl esters for use as diesel fuel substitute. One
important property in biodiesel fuel is the content of
unreacted vegetable oil in the form of glycerin or
glycerides. The determination of these products is
complex. As has been indicated, the literature19 reports
that in the final equilibrium of the ethanolysis of
rapeseed oil there is a 2 and 4% of di- and monoglyc-
erides when the ethyl ester yield is 94%. Although in
this work Cynara cardunculus oil has been used, the
quantities of di- and monoglycerides should be very
similar. In this situation the equilibrium can be shifted
to the right by carrying out a multistage transesterifi-
cation process. This idea is the basis of the Henkel
industrial process for methyl ester production, which
is carried out in a two-stage reaction with separation
of the glycerin after each stage.28
Conclusions
The main conclusions of the present work are the
following:
(a) Cynara cardunculus L. oil has suitable character-
istics for transesterification. The process yields a biodie-
sel with similar properties to those of diesel no. 2.
Consequently, cultivation of this plant can be considered
as a potential alternative for abandoned cropland, and
is a serious candidate for a renewable energy source.
(27) Junek, H.; Mittelbach, M. French Patent FR 2,603, 296, 1987.
(28) Assmann, G.; Blasey, G.; Gutsche, B.; Jeromin, L.; Rigal, J.;
Armengaud, R.; Cormary, B. World Patent WO 91 05,034, 1991.
Encinar et al.
(b) The influence of temperature is important kineti-
cally. The best results were obtained at 75 °C, very near
to the ethanol boiling point. Nevertheless, the transes-
terification progresses satisfactorily even at room tem-
perature, which could be very interesting for industrial
scale production due to the energy savings that it would
imply.
(c) Of the catalysts tested, sodium hydroxide gave the
highest ester yield, although the yields with potassium
hydroxide were not much lower. The optimum concen-
tration for both catalysts was near 1%. The addition of
an excessive amount of catalyst, however, gave rise to
the formation of an emulsion, which increased the
viscosity and led to the formation of gels. The addition
of lower quantities lead to a decrease in the yield.
(d) The ethanol/oil molar ratio is one of the variables
that had most influence on the process. Within the
range of molar ratios employed in the present work, the
best results were obtained for a 9:1 ratio. Smaller ratios
diminish the conversion. Larger ratios cause problems
in the separation of glycerin and the recovery of ethanol.
(e) Of the parameters determined in the biodiesels,
the density, viscosity, cetane index, and high heating
value were similar to, or even better than, those of diesel
nο. 2. However, the cloud point and pour points were
higher, which might point to potential difficulties in cold
starts. On the other hand, the flash and combustion
points, which were higher than those of diesel nο. 2,
constitute a guarantee of safety from the point of view
of handling and storage.
Acknowledgment. The authors express their grati-
tude to the “Junta de Extremadura” for the financial
support received to perform this study, by means of the
project IPR98B038.
EF010174H