Renewable Energy 68 (2014) 618e624

Contents lists available at ScienceDirect

Renewable Energy
journal homepage: www.elsevier.com/locate/renene

Optimization of biodiesel production from waste sh oil

Pedro J. Garca-Moreno a, *, Mohriam Khanum b, Antonio Guadix a, Emilia M. Guadix a
a
b

Department of Chemical Engineering, University of Granada, 18071 Granada, Spain

Department of Chemical Engineering, University College London, WC1E6BT London, United Kingdom

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 4 October 2013
Accepted 3 March 2014
Available online

The present study deals with the production of biodiesel using waste sh oil. The research assesses the
effect of the transesterication parameters on the biodiesel yield and its properties, including temperature (40e60
C), molar ratio methanol to oil (3:1e9:1) and reaction time (30e90 min). The experimental results were tted to complete quadratic models and optimized by response surface
methodology. All the biodiesel samples presented a FAME content higher than 93 wt.% with a maximum,
95.39 wt.%, at 60
C, 9:1 of methanol to oil ratio and 90 min. On the other hand, a maximum biodiesel
yield was found at the same methanol to oil ratio and reaction time conditions but at lower temperature,
40
C, which reduced the saponication of triglycerides by the alkaline catalyst employed. Adequate
values of kinematic viscosity (measured at 30
C) were obtained, with a minimum of 6.30 mm2/s obtained at 60
C, 5.15:1 of methanol to oil ratio and 55.52 min. However, the oxidative stability of the
biodiesels produced must be further improved by adding antioxidants because low values of IP, below
2.22 h, were obtained. Finally, satisfactory values of completion of melt onset temperature, ranging from
3.31
C to 3.83
C, were measured.
2014 Elsevier Ltd. All rights reserved.

Keywords:
Fish waste oil
Biodiesel
Alkaline transesterication
Kinematic viscosity
Oxidative stability
Completion of melt onset temperature

1. Introduction
The sh industry still generates large quantities of by-products,
mainly because of sh processing and discards [1]. Although these
wastes are generally employed to obtain low value products such as
sh silage or fertilizer, highly valued commodities such as sh oil
can be also produced from them [2,3]. Fish oil containing a high
content of omega-3 PUFA (>20 wt.%) has gained an increasing interest in the pharmaceutical and functional food industries, due to
the recognized health benets of eicosapentaenoic (C20:5n-3) and
docosahexaenoic acids (C22:6n-3) [4,5]. However, sh oil having a
lower amount of EPA and DHA leads to very low yield of omega-3
concentrates which considerably reduces its applications on both
elds [6]. In addition, sh oil extracted from sh wastes may not
meet the quality criteria required for edible purposes [7].
In this framework, alternative uses for poor quality sh oil, for
example as a fuel, need to be investigated. Although crude sh oil
can be directly employed as combustible in boilers or furnaces, it
may results in large particulate emissions and carbon deposits on

* Corresponding author. Tel.: 34 958241329; fax: 34 958248992.

E-mail address: pjgarcia@ugr.es (P.J. Garca-Moreno).
http://dx.doi.org/10.1016/j.renene.2014.03.014
0960-1481/ 2014 Elsevier Ltd. All rights reserved.

the fuel injectors [8]. As a solution, crude sh oil can be converted

into biodiesel which is a renewable diesel fuel with a reduced
viscosity and a low emissions prole [9]. Furthermore, using sh
waste oil as feedstock for biodiesel has the potential to reduce the
production costs drastically [10]. Biodiesel is a mixture of fatty acid
alkyl esters mostly produced from the transesterication of fats and
oils. In this process, the triglycerides contained in the oil react with
an alcohol, commonly methanol, and a catalyst, usually sodium
hydroxide (NaOH), to yield fatty acid methyl esters (FAME) [11].
This reaction is highly inuenced by temperature, molar ratio of
methanol to oil, reaction time, catalyst and free fatty acids content
of the oil [12]. Although considerable research has been conducted
on the optimization of biodiesel production from vegetable oils by
this method, only a few studies have been carried out when
employing sh oil as raw material.
In light of the above, the aims of this work were: a) to evaluate
the inuence of temperature, methanol to oil ratio and reaction
time on the yield and fuel characteristics of biodiesel produced
from crude sh oil via transesterication using NaOH with methanol, and b) to nd, by using response surface methodology, the
processing conditions which lead to optimum values for yield, kinematic viscosity, oxidative stability and cold ow properties of the
biodiesel obtained.

P.J. Garca-Moreno et al. / Renewable Energy 68 (2014) 618e624

2. Materials and methods

2.1. Materials
The sh oil employed contained 90 wt.% of salmon oil and the
remaining 10 wt.% as a mixture of other sh oil in season. It was
purchased from Industrias Anes, S.L. (Vigo, Spain) and presented
the following properties: acid value of 6.6 mg KOH/g oil, peroxide
value of 6 meq O2/kg oil and a composition of 5.7 wt.% in EPA and
7.6 wt.% in DHA. Methanol (99.8%), sodium hydroxide (99%),
sodium sulfate (99%), n-heptane (99%) and methyl heptadecanoate (99%) were purchased from SigmaeAldrich, whereas sulfuric acid (95%) was provided by VWR International.
2.2. Production process
Due to the fact that the free fatty acids (FFA) content of the crude
sh oil is higher than 2 mg KOH/g oil, which would negatively affect
to the biodiesel yield because of soap formation, a two-step process
was performed as suggested by El-Mashad et al. [7]. The rst step
consisted of an acid-catalyzed pre-treatment to esterify the FFA
which permitted to reduce the acid value below 2 mg KOH/g oil
before the alkaline transesterication. In brief, 200 g of salmon oil
were introduced with methanol (molar ratio methanol to oil of 6:1)
and 1 wt.% sulphuric acid as catalyst into a at-bottomed ask. The
samples were stirred at 300 rpm and kept at 60
C for 60 min using
a hot plate controlled by a thermo-regulator. After completion of
the reaction, the mixture was transferred to a separating funnel to
settle for 1 h and the bottom layer containing the esteried fatty
acids and the unconverted oil was separated.
In the second step, the unconverted sh oil from the rst step
was subjected to alkaline transesterication. In short, 200 g of oil
were introduced with a given methanol amount and 1 wt.% sodium
hydroxide as catalyst into a at-bottomed ask. The samples were
stirred at 300 rpm and kept at the selected temperature for a
desired time. The three process parameters considered for this
study were temperature, molar ratio methanol to oil and reaction
time. A factorial experimental design comprising 27 runs was carried out, in which each input variable was set at three levels; 40, 50,
60
C for temperature; 3:1, 6:1 and 9:1 for methanol to oil molar
ratio; and 30, 60, 90 min for reaction time. After the reaction was
completed, the content of the reactor was poured to a funnel to
settle for 1 h. The alkyl esters were separated and puried by
washing two times with distilled water at 50
C to remove
unreacted catalyst, free glycerol and the remaining methanol. Then,
the methyl esters were vacuum ltered through sodium sulphate to
eliminate residual moisture. The biodiesel samples were stored
under nitrogen at 4
C in amber bottles until analysis.
2.3. FAME content and yield
The FAME content of the biodiesel samples was determined
according to the European standard EN 14103 [13] with some
modications. The methyl esters of biodiesel samples were
analyzed according to the method described by Camacho-Paez et al.
[14] with an Agilent 7890A gas chromatograph (Agilent Technologies, S.A., Santa Clara, California, USA) connected to a capillary
column of fused silica Omegawax (0:25 mm  30 m, 0:25 mm
standard lm; Supelco, Bellefonte, PA) and a ame-ionization detector. Nitrogen was used as the carrier gas and the total column
ow was 44 ml/min. The oven temperature program was initially
set at 150
C for 3 min, then increased at a rate of 10
C/min until
240
C and maintained at this temperature for 12 min. Matreya
(Pleasant Gap, PA) n-3 PUFAs standard (catalog number 1177) was
used for the qualitative fatty acid determination and methyl

619

heptadecanoate was employed as internal standard for quantitative

determination.
Once the FAME of the biodiesel samples were identied, the
peak areas were employed to determine the FAME content of each
sample by Eq. (1):

A  AEI CEI  VEI

 100

AEI
m

(1)

where C fatty acid methyl ester content (%); SA total peaks

area; AEI area of the peak corresponding to methyl heptadecanoate; CEI concentration of methyl heptadecanoate solution in
heptane (mg/ml); VEI volume of methyl heptadecanoate solution
(ml); m mass of biodiesel sample (mg).
The yield of the produced biodiesel was expressed as follows,
Eq. (2):

Yield

MBiodiesel  C
 100
MOil

(2)

where MBiodiesel is the mass of puried methyl esters obtained, MOil

is the mass of oil employed and C is the fatty acid methyl ester
content determined as described above.
2.4. Determination of biodiesel characteristics
2.4.1. Kinematic viscosity
Viscosity of the biodiesel samples were measured using a
rotational viscometer Haake model VT500 and NV sensor system
(Fisher Scientic, Aalst, Belgium). Each sample was placed between
the two coaxial cylinders and set at 30
C employing a water bath.
Shear stress (s) was measured at varying shear rates (g) from 0 to
3000 s1. Then, the dynamic viscosity (m, Pa s) of each sample was
determined as the slope of the straight line resulting when plotting
shear stress (s, Pa) versus the shear rate (g, s1). Measurements
were performed in duplicate.
The density determinations involve heating the biodiesel sample at 30
C and then simply weighing the sample, at that specic
temperature, to determine the mass of the sample with a known
volume. Measurements were carried out in duplicate.
Once the dynamic viscosity and the density were measured,
they were used to determine the kinematic viscosity at 30
C for
each sample. It was expressed in mm2/s.
2.4.2. Oxidative stability
The Rancimat test was employed to determine the oxidative
stability of the biodiesel samples according to the European standard EN 15751 [15]. A Metrohm Rancimat model 743 (Methrom
Instruments, Herisau, Switzerland) was utilized. A stream of
ltered, cleaned and dried air at a rate of 10 L/h was bubbled into
7.5 g of biodiesel samples contained in reaction vessels. These
vessels were placed in an electric heating block, which was set at
110
C. Efuent air containing volatile organic acids from the oil
samples were collected in a measuring vessel with 60 mL of
distilled water. The conductivity of the water was continuously
recorded and the induction period (IP) was automatically determined using the apparatus. The measurements were carried out in
duplicate. Rancimat IP was expressed as resistance time (in hours)
of the oil to oxidation.
2.4.3. Differential scanning calorimetry
DSC curves are commonly employed to obtain information
about the cold ow properties of biodiesel [16]. In this work, DSC
heating curves were obtained for each biodiesel sample in order to
determine their completion of melt onset temperature (COM).

620

P.J. Garca-Moreno et al. / Renewable Energy 68 (2014) 618e624

DSC analyses were conducted using a differential scanning

calorimeter Metler Toledo DSC 1 (Mettler Toledo, Barcelona, Spain).
For each scan, approximately 10 mg of sample were hermetically
sealed in an aluminium pan and tested against an identical empty
pan. To determine the phase transition of the biodiesel samples, a
ramp rate of 5
C/min was used under nitrogen atmosphere (50 ml/
min). Samples were cooled and held isothermally at 80
C for
10 min, then heated to 100
C.
2.4.4. Acid value
The acidity of the biodiesel samples was determined according
to the European standard EN 14104 [17]. The method used is based
on titration of the sample, diluted with ethanol-diethyl ether mixed
solvent, with potassium hydroxide solution in ethanol employing
phenolphthalein as indicator to detect the end point. The results
are expressed in milligrams of potassium hydroxide per gram of
sample (mg KOH/g oil). The samples were measured in duplicate.
2.5. Statistical analysis
The Statgraphics software (version 5.1) was used to relate each
output variable (FAME content, yield, viscosity, IP and COM) to the
experimental factors (temperature, methanol to oil ratio and reaction time) by means of a quadratic model as follows, Eq. (3):

Y b0

3
X
i1

bi X i

3
X
i1

bii Xi2

3
X

bij Xi Xj

(3)

Table 1
Experimental design and measured values for the response variables.
Exp#

Temp.
(
C)

Methanol:Oil
ratio

Time
(min)

FAME
(wt.%)

Yield
(%)

Viscosity
(mm2/s)

IP (h)

COM
(
C)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27

40
40
40
40
40
40
40
40
40
50
50
50
50
50
50
50
50
50
60
60
60
60
60
60
60
60
60

3
3
3
6
6
6
9
9
9
3
3
3
6
6
6
9
9
9
3
3
3
6
6
6
9
9
9

30
60
90
30
60
90
30
60
90
30
60
90
30
60
90
30
60
90
30
60
90
30
60
90
30
60
90

93.12
93.26
93.72
93.98
93.65
93.76
93.58
93.35
93.53
93.08
93.27
93.43
93.87
93.91
94.19
93.60
93.59
95.14
94.19
93.95
94.65
94.05
94.10
93.56
93.15
94.78
95.39

74.49
76.14
72.16
78.94
80.16
80.02
83.06
77.71
82.40
71.67
72.47
76.19
76.50
73.25
80.77
74.51
78.24
81.01
79.31
76.43
72.31
77.83
71.19
72.97
79.92
79.95
75.22

6.66
6.45
6.41
6.48
6.45
6.57
6.51
6.35
6.56
6.61
6.38
6.46
6.33
6.40
6.42
6.37
6.43
6.43
6.37
6.39
6.40
6.45
6.05
6.51
6.40
6.49
6.62

1.29
1.43
1.29
1.21
1.30
1.27
1.10
1.31
1.36
1.80
1.05
1.25
1.40
1.21
1.38
1.53
1.44
1.27
1.78
1.86
1.21
1.83
1.16
1.20
2.22
1.50
1.23

3.70
3.71
3.72
3.70
3.75
3.83
3.76
3.79
3.82
3.70
3.73
3.55
3.69
3.75
3.48
3.73
3.82
3.58
3.72
3.74
3.44
3.54
3.74
3.37
3.31
3.69
3.77

i<j

where the coefcients bi and bii are related to the linear and
quadratic effects, respectively, of each input factor on the output
variable and the cross-product coefcients bij represent the interactions between two input variables.
Secondly, the analysis of variance (ANOVA) was carried out. The
signicance of all terms in the models was judged statistically by
computing the p-value at a condence level of 95%. The regression
coefcients were then used to generate contour maps and to nd
the processing conditions which lead to an optimum (maximum or
minimum) value for a given response variable. The optimum solution should be obtained by a combination of the experimental
factors inside their ranges of application.
3. Results and discussion
3.1. Characterization of biodiesel samples
Table 1 shows the experimental values of FAME content, yield,
kinematic viscosity, IP and COM for the biodiesels produced.
3.1.1. FAME content and yield
The biodiesel purity was assayed by determining its FAME
content. A good conversion of triglycerides to FAME was obtained
for all the samples because their FAME content were higher than
93 wt.% (Table 1). Although the inuence of the processing conditions was not clear, it was seen that when maintaining constant the
methanol to oil ratio at 9:1 and the reaction time at 90 min, the
FAME content increased with temperature (from 93.53 wt.% at
40
C to 95.39 wt.% at 60
C). Besides, the FAME content improved
at larger reaction time when employing a methanol to oil ratio of
9:1 at 50
C and 60
C but not at 40
C. Hence, the maximum
conversion, 95.39 wt.%, was obtained at 60
C with a 9:1 methanol
to oil ratio and a reaction time of 90 min. Nevertheless, it should be
mentioned that this value is slightly lower than the minimum
required by the European standard (EN 14214) for use as 100%
diesel fuel (96.5 wt.%).

The fatty acids of the biodiesel samples were composed primarily of 28.7 wt.% oleic acid (C18:1n-9), 11.8 wt.% palmitic acid
(C16:0), 10.2 wt.% linoleic acid (C18:2n-6), 7.6 wt.% DHA (C22:6n-3),
5.7 wt.% EPA (C20:5n-3) and 5.0 wt.% palmitoleic acid (C16:1n-7).
The content of saturated fatty acids (SFA) and polyunsaturated fatty
acids (PUFA) amounted to 17.7 wt.% and 33.2 wt.%, respectively. This
was in agreement with the SFA content of waste cooking oil biodiesel (19.8 wt.%) and with the PUFA content of marine oil biodiesel
(37.3 wt.%) [18].
The biodiesel yields obtained ranged from 71.19 to 83.06%
(Table 1). It was observed that the highest percentage yields were
found at low temperature (40
C) and high methanol to oil ratio
(9:1). Although the reaction time had a positive effect on yield at
50
C, since at constant methanol to oil ratio the yield increases
with longer contact time, this trend was not seen at 40
C or at
60
C. These yield values are in line with those reported for biodiesel produced with oil extracted from sh canning industry
wastes, 73.9% using 1 wt.% of acid catalyst [19] and with menhaden
oil, 86.2% employing 2 wt.% of alkaline catalyst [20]. However, they
were considerably lower than those found for biodiesel obtained
with salmon oil, 99% [7] and with heckel sh oil, 96% [21]. It may be
due to the fact that the latter two studies employed a lower concentration of alkaline catalyst, 0.5 wt.%. According to Fadhil and Ali
[21], an alkaline catalyst (NaOH or KOH) concentration beyond
0.5 wt.% enhanced the formation of soaps by triglyceride saponication. Owing to their polarity, soaps are dissolved into the glycerol phase increasing also the methyl ester solubility in the glycerol,
which is a cause of yield loss [22].
3.1.2. Viscosity
Viscosity is an important fuel property since it affects the injection and atomization of a fuel. The higher the viscosity, the
greater the tendency of a fuel to cause operational problems such as
engine deposits [23]. The kinematic viscosity of the biodiesel
samples varied between 6.05 and 6.66 mm2/s at 30
C (Table 1).
Thus, the transesterication process was effective at reducing the

P.J. Garca-Moreno et al. / Renewable Energy 68 (2014) 618e624

viscosity of the crude sh oil, which was 45.34 mm2/s at 30
C.

According to Knothe [23], biodiesel viscosity increases with chain
length and degree of saturation of fatty acids. As a result, the biodiesels produced in this work (17.7 wt.% SFA and 33.2 wt.% PUFA)
showed a lower viscosity than biodiesel from a different marine
sh oil (37.06 wt.% SFA and 37.3 wt.% PUFA) which exhibited a kinematic viscosity of 7.2 mm2/s at 40
C [18]. The biodiesels obtained
were close to meet the European standard (3.5e5 mm2/s at 40
C)
since reduced viscosities would have been obtained at 40
C.
3.1.3. Oxidative stability
Oxidative stability is a key fuel characteristic determining the
shelf life of biodiesel. Biodiesel oxidation leads to an increase in
viscosity due to the formation of high-molecular-weight products
from polymerization reactions. This may results in the production
of insoluble substances which can clog fuel lines and pumps [24].
The fatty acid composition of biodiesel is the main factor inuencing the oxidative stability. In this sense, sh oil biodiesel is
highly susceptible to oxidation due to its high content of PUFA
containing more allylic methylene positions [25]. In this study, the
Rancimat IP of the biodiesels samples was determined as a measure
of biodiesel resistance to oxidation. In Table 1, it was observed that
the IP values varied between 1.05 and 2.22 h. These results are
signicantly higher than the IP (0.1 h) reported for biodiesel produced using oil from sh canning industry wastes [19], but
considerably lower than those found for high oleic sunower oil
biodiesel (IP of 5.1 h) and palm oil biodiesel (IP of 8.1 h) [26]. This
fact is attributed to the superior PUFA content of sh oil when
compared with vegetable oils. Therefore, the use of antioxidants
[27] or the mixture of sh oil with more stable oils before transesterication [19] are required in order to obtain oxidative stabilities which satisfy the European standard (minimum IP of 6 h).
3.1.4. Cold ow properties
One of the main disadvantages of biodiesel is its poor cold ow
properties. As ambient temperature decreases, the long-chain
saturated fatty acids start to form small crystals nuclei in biodiesel fuels which can become large enough to plug fuel lines [24].
Thermal analytical methods such as DSC have been previously used
to determine melting characteristics of biodiesel [10,16,28]. In this
work, similar DSC heating curves were obtained for the 27 biodiesel
samples, with three distinct endothermic peaks (Fig. 1). A rst small
peak was found between 73.60
C and 60
C. It was not very
sharp and may be associated to the presence of methyl esters of
polyunsaturated fatty acids with very low melting point such as
EPA and DHA. A larger peak was observed in the range of 55
C to
30
C, which is attributed to the presence of methyl esters of

Fig. 1. DSC heating curve of biodiesel (Exp. 15).

621

unsaturated fatty acids. The last visible peak was obtained between
20
C and 3.83
C, which may mainly consist of methyl esters of
SFA. Hence, the melting point of the biodiesel samples ranged from
73.60
C to 3.83
C. Similar endothermic peaks were found by
Garcia-Prez et al. [28] for biodiesel produced from poultry fat,
with a melting point between 69.5
C and 14.2
C. Subsequently,
DSC curves were analyzed to determine the COM for all the biodiesel samples (Table 1). COM is dened as the temperature at
which the melting phenomenon nishes and it ranged from 3.31
C
to 3.83
C. This small variation is due to the fact that transesterication does not alter the fatty acid composition of the
feedstocks and thus, it is the fatty acid prole of the raw oil which
determines the melting properties of the biodiesel [24]. A similar
value was found by Chiou et al. [10] for salmon oil methyl esters,
3.1
C; but a lower one was reported by the same authors for corn
oil methyl esters, 1.7
C, which is explained by its lower saturated
fatty acids content.
3.1.5. Acidity
The acid value of biodiesel is considered as a quality control
parameter given that it determines its corrosiveness and long-term
stability [29]. In this study, all the biodiesel samples showed an acid
value below 0.31 mg KOH/g oil after the transesterication step,
which is in accordance with the European standard (<0.5 mg KOH/
g oil).
3.2. Statistical modeling
The experimental data for each output variable were tted to a
complete second order model. Table 2 shows the polynomial coefcients calculated by multiple regressions for each response
variable and their associated p-value. It was observed that the
linear effect of temperature was statistically signicant for FAME
content, IP and COM. The linear effect of reaction time was also
signicant for FAME content and IP. On the other hand, the linear
effect of methanol to oil ratio was found to be statistically signicant only for yield. Quadratic effects were signicant only for reaction time in the case of viscosity and COM. The p-values for the
remaining effects indicated that only the interaction between
temperature and reaction time was statistically signicant in the
case of IP. Nevertheless, it should be mentioned that the proposed
quadratic models seems to not explain the variability of the data to
a large extent. The highest coefcient of determination was obtained for IP, R2 0.7076, whereas values above 0.5 were found for
the rest of the output variables (Table 2).
3.3. Optimization by response surface methodology
For the optimization of the biodiesel production, the yield was
chosen as measured variable showing the efciency of the process,
whereas viscosity, IP and COM were selected as variables indicating
the fuel quality of the biodiesel produced. The goodness of the t
for these variables was also proved by plotting the measured values
against the predicted ones for yield (Fig. 2a), viscosity (Fig. 2b), IP
(Fig. 2c) and COM (Fig. 2d). The data were correlated by means of a
regression line whose equation was inserted in each gure.
Although a low coincidence between the regression line and the
diagonal was observed for the four variables, most of the experimental points were placed inside the region delimited by the
dotted lines at both sides of the diagonal, which represent a deviation of 10% between predicted and measured values. Subsequently, by means of the second-order models obtained above and
using response surface methodology, contour maps were generated
where the yield, viscosity, IP and COM were plotted against time of
reaction and methanol to oil molar ratio, as shown in Fig. 3aed.

622

P.J. Garca-Moreno et al. / Renewable Energy 68 (2014) 618e624

Table 2
Polynomial coefcients and p-values for the response variables.
FAME (wt.%)

Yield (%)

Viscosity (mm2/s)

IP (h)

COM (
C)

Coefcient

p-Value

Coefcient

p-Value

Coefcient

p-Value

Coefcient

p-Value

Coefcient

p-Value

Constant
A: Temperature,
C
B: Methanol:Oil ratio
C: Time, min
AA
AB
AC
BB
BC
CC

9.61E01
8.88E02
1.39E02
4.70E02
8.78E04
4.26E04
5.19E04
8.33E03
2.15E03
1.41E04

e
0.0139
0.1249
0.0415
0.6760
0.9310
0.3007
0.7207
0.2025
0.5478

8.86E01
8.10E01
2.28
2.97E02
1.17E02
3.70E02
4.07E03
4.80E04
5.51E03
1.12E03

e
0.1371
0.0053
0.8066
0.3565
0.2185
0.1785
0.9975
0.5762
0.4226

8.65
4.20E02
1.72E01
2.23E02
2.33E04
1.24E03
1.16E04
5.96E03
6.43E04
1.08E04

e
0.0784
0.9442
0.6527
0.5692
0.2067
0.2370
0.1985
0.0569
0.0268

9.86E01
2.53E03
2.02E01
2.09E02
4.69E04
9.44E04
6.94E04
1.23E02
1.11E04
6.79E05

e
0.0061
1.0000
0.0029
0.5172
0.5989
0.0010
0.1555
0.8401
0.4250

3.65
8.33E04
3.01E02
1.29E02
2.22E05
1.03E03
5.56E05
3.58E03
7.22E04
1.25E04

e
0.0072
0.5938
0.5523
0.9621
0.3567
0.6152
0.4933
0.0621
0.0254

R2

0.5221

0.5171

Temperature was set at the optimum value obtained for each

output variable: 40
C for yield and 60
C for viscosity, IP and COM.
These contours plot permitted to nd optimum values for each
objective variable, which were marked as circles in Fig. 3.
3.3.1. Optimization of biodiesel production
The maximum value for yield, 83.6%, was located at the lowest
temperature assayed, 40
C, and the highest methanol to oil ratio
and reaction time, 9:1 and 90 min respectively (Fig. 3a). Due to the
high amount of catalyst employed, 1 wt.%, the saponication of
triglycerides by NaOH before alcoholysis is reduced at low

0.5294

0.7076

0.5599

temperature, which results in a higher FAME yield [21,30].

Although the stoichiometric methanol to oil ratio is 3:1, increasing
this ratio permits to drive the transesterication reaction at a faster
rate. Similarly to our result, Ramadhas et al. [30] also obtained a
maximum of biodiesel yield produced from rubber seed oil at a
methanol to oil molar ratio of 9:1. These authors also stated that
increasing the molar ratio beyond 9:1 did not result in a signicant
increase of yield. With regards to reaction time, it was found to
have a positive inuence on the biodiesel yield since a higher reaction time allows longer for the completion of transesterication
of triglycerides.

Fig. 2. Correlation between predicted and measured values of (a) Yield, (b) Viscosity, (c) IP and (d) COM.

P.J. Garca-Moreno et al. / Renewable Energy 68 (2014) 618e624

623

Fig. 3. Contour plots for (a) Yield at 40
C, (b) Viscosity, (c) IP and (d) COM at 60
C.

3.3.2. Optimization of biodiesel properties

Optimizing the fuel properties of the biodiesels obtained implies
to minimize viscosity, in order to improve the biodiesel ow, and
COM, to reduce the crystallization temperature of the fuel; while
maximizing IP, which implies a longer biodiesel shelf life.
The minimum value for kinematic viscosity, 6.30 mm2/s, was
obtained at the highest value of temperature assayed, 60
C, and at
medium values of methanol to oil ratio and reaction time, 5.15:1
and 55.52 min respectively (Fig. 3b). For a same feedstock, biodiesel
viscosity is mainly related to its purity [19]. These authors attributed higher biodiesel viscosities to a lower FAME content. Thus, the
higher temperature assayed, 60
C, which resulted in the maximum
FAME content was also preferred to obtain reduced viscosities. The
optimum for methanol to oil ratio and reaction time at medium
values may be due to the fact that not relevant differences in FAME
content (less than 1.3 wt.%) were found when using higher values
for these variables.
Fig. 3c shows the maximum value obtained for IP, 1.99 h, which
was found at 60
C, methanol to oil ratio of 9:1 and 30 min for
reaction time. Although increasing temperature favors the oxidation of FAME, it may also enhance the mixture of the produced
FAME and excess of methanol which could reduce the oxidative
deterioration of the FAME during the production process. Thus, the
highest ratio of methanol to oil was also preferred. Considering
reaction time, the lowest value of time assayed was found as optimum value because it implies a shorter contact time of FAME with
oxygen at the processing temperature.
Finally, in Fig. 3d it is observed the minimum for COM, 3.48
C,
which was obtained at 60
C, 3.74:1 methanol to oil ratio and
90 min. The COM value of biodiesel mainly depends on the SFA
composition of the fuel [23]. In this regard, no signicant differences were found among the SFA content of the 27 samples

obtained (less than 0.42 wt.%). Therefore, it cannot explain the

location of the optimum at these processing conditions values.
4. Conclusions
Oil obtained from sh waste, having a low content of omega-3
PUFA, was proven to be a promising raw material for the production
of biodiesel. Due to the high FFA content of the crude sh oil,
expressed as 6.6 mg KOH/g oil, an acid-catalyzed esterication was
performed before the transesterication with methanol and 1 wt.%
sodium hydroxide as catalyst. Then, it was evaluated the inuence of
the transesterication processing conditions such as temperature,
methanol to oil ratio and reaction time on yield and fuel properties of
the biodiesel produced. It was found that all the biodiesels samples
presented a FAME content higher than 93 wt.%, with a maximum of
95.39 wt.% at 60
C, 9:1 of methanol to oil ratio and 90 min. By
response surface methodology optimization, a maximum yield,
83.60%, was obtained at 9:1 methanol to oil ratio, 90 min of reaction
time and low temperature, 40
C, at which the saponication of triglycerides by NaOH before alcoholysis is reduced. Acceptable values of
kinematic viscosity were found, with a minimum of 6.30 mm2/s obtained at 60
C, 5.15:1 of methanol to oil ratio and 55.52 min. The
oxidative stability of the biodiesels produced was low due to their
high PUFA content, with a maximum IP of 2.22 h. Thus, it must be
further improved through the use of antioxidants. Finally, satisfactory
melting points, from 73.60
C to 3.83
C, with COM values ranging
from 3.31
C to 3.83
C, were provided for the sh waste oil biodiesels.
Acknowledgments
This work was supported by the Spanish National Plan I D i
(projects CTQ2008-02978 and CTQ2011-23009). P.J. Garca-Moreno

624

P.J. Garca-Moreno et al. / Renewable Energy 68 (2014) 618e624

acknowledges an FPI grant from the Spanish Ministry of Science

and Innovation (BES-2009-013871).

[15]

References
[16]
[1] Rustad T. Physical and chemical properties of protein seafood by-products. In:
Shahidi F, editor. Maximising the value of marine by-products. Cambridge:
Woodhead Publishing Limited; 2006. pp. 3e21.
[2] Aidos I, Kreb N, Boonman M, Luten JB, Boom RM, van der Padt A. Inuence of
production process parameters on sh oil quality in a pilot plant. J Food Sci
2003;68:581e7.
[3] Wu TH, Bechtel PJ. Salmon by-product storage and oil extraction. Food Chem
2008;111:868e71.
[4] Uauy T, Valenzuela A. Marine Oils: the health benets of n-3 fatty acids.
Nutrition 2000;16:680e4.
[5] Ward OP, Singh A. Omega-3/6 fatty acids: alternative sources of production.
Process Biochem 2005;40:3627e52.
[6] Bimbo AP. Current and future sources of raw materials for the long-chain
omega-3 fatty acid market. Lipid Technol 2007;19:176e9.
[7] El-Mashad HM, Zhang R, Avena-Bustillos RJ. A two-step process for biodiesel
production from salmon oil. Biosyst Eng 2008;99:220e7.
[8] Ma F, Hanna MA. Biodiesel production: a review. Bioresour Technol 1999;70:
1e15.
[9] Helwani Z, Othman MR, Aziz N, Fernando WJN, Kim J. Technologies for production of biodiesel focusing on green catalytic techniques: a review. Fuel
Process Technol 2009;90:1502e14.
[10] Chiou B, El-Mashad HM, Avena-Bustillos RJ, Dunn RO, Bechtel PJ, McHugh TH,
et al. Biodiesel from waste salmon oil. Trans ASABE 2008;51:797e802.
[11] Gerpen JV. Biodiesel processing and production. Fuel Process Technol
2005;86:1097e107.
[12] Meher LC, Vidya Sagar D, Naik SN. Technical aspects of biodiesel production
by transesterication e a review. Renew Sust Energ Rev 2006;10:248e68.
[13] UNE-EN 14103. Fat and oil derivatives-Fatty Acid Methyl Esters (FAME)determination of ester and linolenic acid methyl ester contents. European
Committee for Standardization.
[14] Camacho-Paez B, Robles-Medina A, Camacho-Rubio F, Gonzlez-Moreno P,
Molina-Grima E. Production of structured triglycerides rich in n-3 polyunsaturated fatty acids by the acidolysis of cod liver oil and caprylic acid in

[17]
[18]
[19]
[20]
[21]

[22]

[23]
[24]
[25]
[26]

[27]
[28]

[29]
[30]

a packed-bed reactor: equilibrium and kinetics. Chem Eng Sci 2002;57:

1237e49.
UNE-EN 15751. Automotive fuels-Fatty acid methyl ester (FAME) fuel and
blends with diesel fuel-Determination of oxidation stability by accelerated
oxidation method. European Committee for Standardization.
Dunn RO. Thermal analysis of alternative diesel fuels from vegetable oils. J Am
Oil Chem Soc 1999;76:109e15.
UNE-EN 14104. Oil and fat derivatives-Fatty Acid Methyl Esters (FAME)Determination of acid value. European Committee for Standardization.
Lin C, Li CR. Fuel properties of biodiesel produced from the crude sh oil from
the soapstock of marine sh. Fuel Process Technol 2009;90:130e6.
Costa JF, Almeida MF, Alvim-Ferraz MCM, Dias JM. Biodiesel production using
oil from sh canning industry wastes. Energ Convers Manage 2013;74:17e23.
Hong IK, Park JW, Lee SB. Optimization of sh-oil-based biodiesel synthesis.
J Ind Eng Chem 2013;19:764e8.
Fadhil AB, Ali LH. Alkaline-catalyzed transesterication of silurus triostegus
heckel sh oil: optimization of transesterication parameters. Renew Energ
2013;60:481e8.
Vicente G, Martnez M, Aracil J. Integrated biodiesel production: a comparison
of different homogeneous catalysts systems. Bioresour Technol 2004;92:297e
305.
Knothe G. Dependence of biodiesel fuel properties on the structure of fatty
acid alkyl esters. Fuel Process Technol 2005;86:1059e70.
Knothe G, Gerpen JV, Krahl J. The biodiesel handbook. 1st ed. Illinois: AOCS
Press; 2005.
Knothe G. Some aspects of biodiesel oxidative stability. Fuel Process Technol
2007;88:669e77.
Serrano M, Bouaid A, Martnez M, Aracil J. Oxidation stability of biodiesel from
different feedstocks: Inuence of commercial additives and purication step.
Fuel 2013;113:50e8.
Lin C, Lee J. Oxidative stability of biodiesel produced from the crude sh oil
from the waste parts of marine sh. J Food Agric Environ 2010;8:992e5.
Garcia-Perez M, Adams TT, Goodrum JW, Das KC, Geller DP. DSC studies to
evaluate the impact of bio-oil on cold ow properties and oxidation stability
of bio-diesel. Bioresour Technol 2010;101:6219e24.
Fan X, Burton R, Austic G. Preparation and characterization of biodiesel produced from sh oil. Chem Technol Fuel Oil 2010;46:287e93.
Ramadhas AS, Jayaraj S, Muraleedharan C. Biodiesel production from high FFA
rubber seed oil. Fuel 2005;84:335e40.