Fuel 91 (2012) 102111

Contents lists available at ScienceDirect

Fuel
journal homepage: www.elsevier.com/locate/fuel

Predicting cetane number, kinematic viscosity, density and higher heating value
of biodiesel from its fatty acid methyl ester composition
Luis Felipe Ramrez-Verduzco , Javier Esteban Rodrguez-Rodrguez, Alicia del Rayo Jaramillo-Jacob
Programa de Procesos de Transformacin, Instituto Mexicano del Petrleo, Eje Central Norte Lzaro Crdenas 152, C.P. 07730 Mxico D.F., Mexico

a r t i c l e

i n f o

Article history:
Received 24 June 2010
Received in revised form 30 May 2011
Accepted 27 June 2011
Available online 21 July 2011
Keywords:
Biodiesel
Fatty acid methyl ester
Prediction
Mixing rules
Physical properties

a b s t r a c t
Biodiesel is a renewable bio-fuel derived from natural fats or vegetable oils, and it is considered as a
promising alternative to substitute diesel fuels. Cetane number, viscosity, density, and higher heating
value are important properties to affect the utilization of biodiesel fuels, because they are involved in
the denition of fuel quality and are required as input data for predictive engine combustion models. This
work presents the characterization of two biodiesel samples made from beef tallow and soybean oil
through their fatty acid methyl esters (FAMEs) prole. Empirical equations were developed to estimate
four physical properties of methyl esters; and an average absolute deviation (AAD) of 5.95%, 2.57%,
0.11% and 0.21% for the cetane number, kinematic viscosity, density, and higher heating value were
founded. Cetane number, viscosity, and higher heating value increases because of the increase of molecular weight and these physical properties decrease as the number of double bonds increases. Unlike that
of above properties, density decreases as molecular weight increases and density increases as the degree
of unsaturation increases. Two general mixing rules and ve biodiesel samples were used to study the
inuence of FAMEs over the physical properties of biodiesel. The prediction of the cetane number, kinematic viscosity, density and higher heating value of biodiesel is very close to the experimental values.
2011 Elsevier Ltd. All rights reserved.

1. Introduction
Decline of available oil reserves and more stringent environmental regulations have motivated the global interest in renewable energy sources. Biodiesel is considered as an attractive
alternative to replace diesel fuels. Biodiesel consists of a mixture
of fatty acid alkyl esters that can be obtained from animal fats or
vegetable oils, mainly by trans-esterication reactions. Biodiesel
has several advantages because it is renewable, biodegradable,
non-toxic, and friendly with environment. Currently, the main
disadvantage could be the high costs of some types of raw materials. However, as the petroleum oil reserves become scarce, this
behavior could be changed. The main differences between biodiesel and diesel are that the former usually has higher density, viscosity, cloud point and cetane number; and lower volatility, gross
caloric value and sulfur content. Cetane number (/), viscosity
(g), density (q) and higher heating value (d) are four important
properties of fuels. Cetane number is widely used to evaluate
the ignition quality of fuels. Viscosity and density have a direct
effect on the atomization process during combustion. Higher
heating value also known as the gross caloric value or gross energy is the amount of heat released during the combustion of one
Corresponding author.
E-mail address: lframir@imp.mx (L.F. Ramrez-Verduzco).
0016-2361/$ - see front matter 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.fuel.2011.06.070

gram of fuel to produce CO2 and H2O at its initial temperature,

this property is usually used to dene the energy content of fuels
and thereby their efciency.
Efforts have been made to nd useful methods to estimate
physical properties of fatty acid methyl esters (FAMEs) from
parameters related to its chemical structure. For instance, in the
middle 1960s Gouw and Vlugter [1] used the Smittenberg relation
to estimate the density of saturated methyl esters at 20 C and
40 C. Allen et al. [2] proposed empirical correlations to estimate
the viscosity of saturate and unsaturated FAMEs as a quadratic
function of their molecular weight. A difference comes from
Krisnangkura et al. [3] who calculated the viscosity of saturated
FAMEs as a function of temperature and the number of carbons
that fatty acids contain. Another option for viscosity was proposed
by Ceriani et al. [4] who used a group contribution methodology to
estimate the viscosity of FAMEs. The Sugdens parachor method
was used by Allen et al. [5] and Ejim et al. [6] to predict the surface
tension of saturated and unsaturated FAMEs. Liquid heat capacity
and absolute entropy at 25 C of FAMEs was calculated by van
Bommel et al. [7]. Klopfenstein, for his part, was proposed a nonlinear relationship between carbon chain length and cetane number
of methyl esters [8].
On the other hand, the prediction of biodiesel properties using
mixing rules has widely been studied. Allen et al. [2], Benjumea
et al. [9], Goncalves et al. [10] and Krisnangkura et al. [3] were used

L.F. Ramrez-Verduzco et al. / Fuel 91 (2012) 102111

103

Nomenclature
AD
AAD
FAME
Mi
n
N
BT
CO
CT
SB

absolute deviation
average absolute deviation
fatty acid methyl ester
molecular weight of the ith FAME
number of FAMEs
number of double bounds
biodiesel made from beef tallow
biodiesel made from corn oil
biodiesel made from cottonseed oil
biodiesel made from soybean oil

the GrunbergNissan equation to estimate the viscosity of biodiesel

from their individual viscosity of FAMEs. Surface tension of biodiesel
was calculated by means of the Dalton-type mass average equation
by Allen et al. [5] and Shu et al. [11]. The effect which specic gravity
of FAMEs has on the specic gravity of biodiesel was considered by
Ejim et al. [6]. Ramos et al. [12] proved that the cetane number of biodiesel can be calculated using their fatty acid methyl composition
and cetane number of FAMEs. Yuan et al. [13] suggested that biodiesel can be regarded as near-ideal solution and once the vapor pressure of each individual FAME is known, the vapor pressure of
biodiesel could be calculated by a simple mixing rule.
This work was developed by the need for simple and reliably calculation methods to estimate some physical properties of bio-fuels,
which could avoid experimental work, which we know it is difcult,
costly, and time consuming. Thus, this paper deals with the prediction of four properties of biodiesel from their fatty acid methyl esters
and the use of mixing rules. For this purpose, two biodiesel samples
from beef tallow and soybean oil are characterized by gas chromatographymass spectrometry (GCMS) in order to obtain its FAME
prole; subsequently, these FAME distributions are used to study
the inuence of FAMEs over the properties of biodiesel samples. In
addition and taking on that a better understanding of connections
between molecular structure and physical properties of FAMEs is
important to establish the quality of biodiesel, our work aims to develop four new empirical correlations to estimate the cetane number, viscosity, density and higher heating value of FAMEs as a
function of their molecular weight and degree of unsaturation.
2. Experimental methods
2.1. Materials
The biodiesel samples used in this study were made from beef
tallow (BT) supplied by Total Energy Services; and from soybean
oil (SB) certied and purchased by Chevron Phillips. Because the
samples were in compliance with the ASTM 6751-08a specication
(only there were minor deviations in both the ash point of BS as in
the total sulfur of BT), no attempt was made to purify them further.
The properties of biodiesel samples are given in Table 1.
2.2. Biodiesel FAMEs characterization
Biodiesel samples studied in this work were analyzed by gas
chromatography, using a 6890 Agilent Technologies device coupled to a mass detector (GCMS) and an Agilent capillary column
(HP-88; 100 m  0.25 mm i.d.  0.20 lm lm). The FAME analysis
was carried out by electron ionization mass in the full scan mode
and transfer line at 250 C, the injection temperature was 225 C,
whereas the sample volume injected was 1 lL. GC-grade helium
was used as carrier gas with a ow rate of 1.2 mL/min. The oven
program was given as follows: initial temperature, hold time

SF

biodiesel made from sunower oil

Greek symbols
/
cetane number
g
kinematic viscosity
q
density
d
higher heating value

No. 1, heating rate, nal temperature and hold time No. 2 was
100 C, 4 min, 3 C/min, 240 C and 15 min, respectively. In order
to determine the retention times of the fatty acid methyl esters,
a standard mixture of 37 FAMEs solved in methylene chloride
(with a standard concentration of 30 mg/mL) was run (see Table 2
and Fig. 1). The standard sample that was used in this work has a
certicate of RESTEK Co. (ISO 9001 registered quality system).
Fig. 1 shows the chromatogram obtained by GCMS of the standard
with 37 FAMEs.

2.3. Cetane number measurements

The cetane number was determined by comparing its combustion characteristics in a test engine with those for blends of reference fuels of known cetane number under standard operating
conditions. This is accomplished using the bracketing handwheel
procedure which varies the compression ratio (handwheel reading)
for the sample and each of two bracketing reference fuels to obtain
a specic ignition delay permitting interpolation of cetane number
in terms of handwheel reading (ASTM D 613-05 method). The measurements were carried out with a CFR F-5 equipment (cetane
method diesel fuel rating unit) that was calibrated with two standard samples of low cetane number (45.5) and high cetane number
(60.0) supply by Chevron Phillips. The repeatability (the difference
between successive results obtained by the same operator with the
same apparatus under constant operating conditions on identical
test material) and reproducibility (the difference between two single and independent results obtained by different operators working in different laboratories on similar material) were estimated
approximately in 0.9 and 4.3, respectively. The experimental
determination of this property is usually difcult and expensive,
and it requires large volume of sample for testing (approximately
500 mL in the case of the ASTM 613-05 method).

2.4. Kinematic viscosity measurements

The ASTM D 445-04e was used to measure the kinematic viscosity of biodiesel. It was measured the time for a volume of liquid
to ow under gravity through a calibrated glass capillary viscometer at 40 C. The repeatability and reproducibility were 0.11% and
0.65%, respectively.

2.5. Density measurements

A digital analyzer (APPAR) was used to measure the density of
biodiesel at 20 C, this equipment consisting of a U-shaped, oscillating sample tube, a system for electronic excitation, and frequency counting. The repeatability and reproducibility were
0.0001 g/mL and 0.0005 g/mL, respectively.

104

L.F. Ramrez-Verduzco et al. / Fuel 91 (2012) 102111

Table 1
Specication of BT and SB biodiesel samples.
Fuel property

Units

ASTM test method

BT

SB

Limits for biodiesel according to the ASTM 6751-08a

Flash point
Water and sediment
Kinematic viscosity, 40 C
Total sulfur
Cupper strip corrosion
Cetane number
Cloud point
Acid number
Free glycerin
Total glycerin
Distillation temperature, 90% recovered (T90)

C
vol.%
mm2/s
ppmw

C
mg KOH/g
wt.%
wt.%
C

D
D
D
D
D
D
D
D
D
D
D

156
<0.05
4.36
17
No. 1a
49.2
3
0.26
0.005
0.197
357.4

92
<0.05
4.07
1.4
No. 1a
51.7
0
0.23
0.004
0.178
355.0

93 min.
0.05 max.
[1.96.0]
15 max.
No. 3 max.
47 min.
Report
0.50 max.
0.020 max.
0.240 max.
360 max.

93-02a
1796-04
445-04e
5453-98
130-04
613-05
2500-05
664-95
6584
6584
1160-06

Table 2
Retention times obtained by GCMS of the C4:0-C22:6 methyl esters from the standard.
No. peak

Retention time (min)

Fatty acid

Compound

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
28
29
30
31
32
33
34
35
36
37

9.73
11.74
15.48
20.81
23.76
26.72
29.62
32.44
34.38
35.11
37.03
37.69
39.21
40.12
41.62
42.46
43.31
43.74
44.68
45.63
46.80
46.97
47.79
48.02
48.83
49.83
50.79
51.11
51.88
52.00
52.08
52.72
53.85
54.30
54.74
56.08
60.16

C4:0
C6:0
C8:0
C10:0
C11:0
C12:0
C13:0
C14:0
C14:1
C15:0
C15:1
C16:0
C16:1
C17:0
C17:1
C18:0
C18:1
C18:1
C18:2
C18:2
C20:0
C18:3
C20:1
C18:3
C21:0
C20:2
C22:0
C20:3
C22:1
C20:3
C20:4
C23:0
C22:2
C20:5
C24:0
C24:1
C22:6

Methyl
Methyl
Methyl
Methyl
Methyl
Methyl
Methyl
Methyl
Methyl
Methyl
Methyl
Methyl
Methyl
Methyl
Methyl
Methyl
Methyl
Methyl
Methyl
Methyl
Methyl
Methyl
Methyl
Methyl
Methyl
Methyl
Methyl
Methyl
Methyl
Methyl
Methyl
Methyl
Methyl
Methyl
Methyl
Methyl
Methyl

2.6. Higher heating value measurements

The higher heating value of biodiesel was measured by the Leco
AC-350 equipment that was calibrated with benzoic acid and the
ASTM D-240-02 method was followed. The results were expressed
in MJ/kg, where the repeatability and reproducibility was
0.13 MJ/kg and 1.66 MJ/kg, respectively.

3. Theoretical basis
Four new empirical correlations to calculate the cetane number,
kinematic viscosity, density, and higher heating value of fatty acid
methyl esters were derived in this work. These equations were obtained by correlating a set of experimental data compiled from dif-

Butyrate
Hexanoate
Octanoate
Decanoate
Undecanoate
Laurate
Tridecanoate
Myristate
Myristoleate (cis-9)
Pentadecanoate
Pentadecenoate (cis-10)
Palmitate
Palmitoleate (cis-9)
Heptadecanoate
Heptadecenoate (cis-10)
Stearate
Elaidate (trans-9)
Oleate (cis-9)
Linoelaidate (trans-9.12)
Linoleate (cis-9,12)
Arachidate
c-Linolenate (cis-6,9,12)
Eicosenoate (cis-11)
Linolenate (cis-9,12,15)
Heneicosanoate
Eicosadienoate (cis-11,14)
Behenate
Eicosatrienoate (cis-8,11,14)
Erucate (cis-13)
Eicosatrienoate (cis-11,14,17)
Arachidonate (5,8,11,14)
Tricosanoate
Docosadienoate (cis-13,16)
Eicosapentaenoate (cis-5,8,11,14,17)
Lignocerate
Nervonate (cis-15)
Docosahexaenoate (cis-4,7,10,13,16,19)

ferent sources [1,1425], Eqs. (1)(4) was adjusted using 21, 46, 36,
and 22 experimental values, respectively. The parameters that are
presented in this work were estimated numerically by using the
software Design-Expert, the variation of physical properties with
molecular weight and degree of unsaturation were properly tted
by means of statistical regression. In all of these empirical correlations, Mi represents the molecular weight of the ith FAME, and N is
the number of double bonds in a given FAME, therefore N is equal
to zero, one, two, and three for the methyl ester C18:0, C18:1,
C18:2 and C18:3, respectively.
The cetane number of each fatty acid methyl ester is obtained
from

/i 7:8 0:302  M i  20  N
where /i is the cetane number of the ith FAME.

L.F. Ramrez-Verduzco et al. / Fuel 91 (2012) 102111

105

Fig. 1. Chromatogram obtained by GCMS of the C4:0-C22:6 methyl esters from the standard.

The kinematic viscosity expressed as a function of Mi and N is

Ingi 12:503 2:496  InM i  0:178  N

where gi is the kinematic viscosity at 40 C of the ith FAME

in mm2/s.
The expression for the density of saturated and unsaturated
FAMEs is

qi 0:8463

4:9
0:0118  N
Mi

where qi is the density at 20 C of the ith FAME in g/cm3.

The higher heating value of methyl esters can be calculated
from

1794
di 46:19 
 0:21  N
Mi

where di is the higher heating value of the ith FAME in MJ/kg.

As mentioned earlier, numerous studies have shown that some
physical properties of biodiesel can be estimated from the individual physical properties of FAMEs, and using appropriate mixing
rules [2,3,5,6,913]. For this purpose, we should use the following
general expression.

fb

n
X

zi  fi

i1

where f is a function that represents any physical property (the subscripts b and i refer to the biodiesel and the pure ith FAME, respectively), zi is the mass or mole fraction of the ith FAME. The function
fb must be replaced by the variables /b, ln(gb), qb and db in order to
specify the cetane number, natural logarithm of kinematic viscosity,
density and higher heating value of biodiesel, where as the function
fi must be interchanged by the variables /i, ln(gi), qi and di in order
to specify the properties of the individual ith FAME.
We are postulating that both mass and mole fraction could be
numerically similar for each ith FAME of specic biodiesel due to
the following facts:

(A) Biodiesel is composed of a relative small number of FAMEs

that can be varied from 4 to 12 depending on the type of
fat or vegetable oil used to produce it.
(B) The average molecular weight of biodiesel is very close to
the molecular weight of FAMEs (for the biodiesel samples
studied in this work the average molecular weight of biodiesel was in the range of 288.5638293.4029 g/mol, as it can
be seen in Table 3, and the difference in molecular weight
between the FAME with greater and lesser carbon atoms
was 110.1968 g/mol).
Eq. (6) provides a simple method for estimating the mole fraction (xi) of individual FAMEs from its mass fraction (wi)
composition.

xi

wi
Mi

n
P wi 
Mi
i1

where xi and wi are the mole and mass fraction, whereas Mi is the
molecular weight of the ith FAME.
Comparisons between experimental and calculated values of
physical properties of FAMEs or biodiesel can be made by means
of the average absolute deviation (AAD) dened by

!
  X
n 
fj;exp  fj;cal  100
1

AAD


n
fj;exp
j1

where n is the number of experimental points that are being compared, whereas fj,exp and fj,cal are functions that represent the experimental and calculated individual FAME or biodiesel physical property.
4. Results and discussion
4.1. Biodiesel characterization
The molecular weight of each FAME (Mi) and concentration of
ve biodiesel samples are shown in Table 3. The composition reported in Table 3 for the rst two biodiesel samples made from
beef tallow (BT) and soybean (SB) was determined experimentally

106

L.F. Ramrez-Verduzco et al. / Fuel 91 (2012) 102111

Table 3
Average molecular weight (M) and concentration (mass fraction and mole fraction in parenthesis) of ve biodiesel samples.
ith FAME

Methyl ester

Mi (g/mol)

wi (xi)

1
2
3
4
5
6
7
8
9
10

C14:0
C16:0
C16:1
C18:0
C18:1
C18:2
C18:3
C20:0
C20:1
C22:1

242.3975
270.4507
268.4348
298.5038
296.4879
294.4721
292.4562
326.5570
324.5411
352.5943
P
M ni1 xi  M i

BT

SB

SF [2]

CO [2]

CT [2]

0.005(0.006)
0.159(0.171)
0.009(0.010)
0.085(0.083)
0.317(0.311)
0.365(0.361)
0.046(0.046)
0.003(0.003)
0.008(0.007)
0.003(0.002)

0.000(0.000)
0.101(0.109)
0.000(0.000)
0.045(0.044)
0.243(0.240)
0.531(0.528)
0.072(0.072)
0.004(0.004)
0.000(0.000)
0.004(0.003)

0.001(0.001)
0.060(0.065)
0.000(0.000)
0.059(0.059)
0.160(0.158)
0.714(0.711)
0.006(0.006)
0.000(0.000)
0.000(0.000)
0.000(0.000)

0.000(0.000)
0.099(0.107)
0.000(0.000)
0.031(0.031)
0.291(0.287)
0.568(0.564)
0.011(0.011)
0.000(0.000)
0.000(0.000)
0.000(0.000)

0.008(0.010)
0.229(0.244)
0.000(0.000)
0.031(0.030)
0.185(0.180)
0.542(0.531)
0.005(0.005)
0.000(0.000)
0.000(0.000)
0.000(0.000)

291.0834 g/mol

292.6725 g/mol

293.4029 g/mol

292.5832 g/mol

288.5638 g/mol

BT: beef tallow oil methyl ester, SB: soybean oil methyl ester, SF: sunower oil methyl ester, CO: corn oil methyl ester, CT: cottonseed oil methyl ester.

in this work, whereas the composition of the following three biodiesel samples (SF: sunower oil methyl ester, CO: corn oil methyl
ester and CT: cottonseed oil methyl ester) was reported previously
by Allen et al. [2]. The analysis technique for the identication of
FAMEs developed in this work allowed a clear identication of
the cis and trans isomers of C18:1 and C18:2 methyl esters in
the BT and SB biodiesel samples, because one of the design features
of the polysiloxane polymer capillary column (Agilent HP-88) is
the separation of cis:trans FAMEs. The ratio cis/trans was 6.4/
93.6 for the C18:1 methyl ester in the BT sample, and only the
presence of the cis isomer C18:2 methyl ester was detected; in
the case of the SB sample only the presence of the cis isomer for
both C18:1 and C18:2 methyl esters was conrmed (none trace
of the trans isomer was founded), the concentrations of C18:1
and C18:2 methyl esters presented in Table 3 encompass the contribution of both isomers. In both samples BT and SB, only the cis
isomer for the C18:3 methyl ester was identied. A possible explanation for the predominance of the cis isomer in the soybean biodiesel may be found if we consider the natural biosynthetic process
of plants, where cis double bounds are inserted by desaturase enzymes in two carbon units at specic positions of the fatty acids
chains. The FAME that had the highest concentration in all biodiesel samples given in Table 3 was the methyl linoleate (C18:2
methyl ester). The soybean biodiesel reported by Allen et al. [2]
has a very similar composition with their homologous biodiesel
SB measured in this work.
The mole fraction for the ve biodiesel samples reported in
Table 3 was derived using Eq. (6). The mole fraction is numerically
similar to the mass fraction for each methyl ester in the ve biodiesel samples, mainly due the molecular weight of FAMEs is restricted to a narrow range.

Table 4
Deviations between the experimental data reported in the literature and results
obtained in this work for the soybean biodiesel characterization.
Methyl ester

AD1

AD2

AD3

C14:0
C16:0
C18:0
C18:1
C18:2
C18:3
C18:4
C20:0
C22:1

0.000
0.039
0.005
0.003
0.011
0.072
0.060
0.004
0.004

0.001
0.002
0.002
0.018
0.010
0.011
0.000
0.004
0.004

0.000
0.019
0.015
0.013
0.019
0.012
0.000
0.004
0.004

AD1: Comparison between Srivastava and Prasad [26] and this work.
AD2: Comparison between Allen et al. [2] and this work.
AD3: Comparison between Goering et al. [27] and this work.

Table 4 gives a comparison through the absolute difference (AD)

between the FAME composition obtained in this work and the
experimental data published by three different Refs. [2,26,27]. Biodiesel (BS) made from soybean oil was used for this purpose because as mentioned above, it was certied by Chevron Phillips.
Srivastava and Prasad [26] were reported the presence of the
C18:4 methyl ester instead of the C18:3 methyl ester that was
identied either by Allen et al. [2], Goering et al. [27] or this work.
The average absolute deviation (AAD) was 0.006 and 0.010 when
we compared our results with the results given by Allen et al. [2]
and Goering et al. [27], respectively. It is important to mention that
the estimation of the ADD was performed without taking on the
concentration of the C16:1, C17:0 and C20:1 methyl esters, as in
all cases the concentration of these compounds was zero. Then,
our results for the FAMEs characterization are in good agreement
with the results published in the literature, considering that there
are many factors that can affect the prole of FAMEs in biodiesel,
mainly related to the quality of raw material (for the vegetable oils
mainly depends on the soil and climatic conditions used during the
culture) and processing conditions during the synthesis of biodiesel. Although the composition of FAMEs in biodiesel made from
vegetable oil of the same origin can be varied, it can be established
that if the biodiesel is of good quality these variations should be
small, and so it is possible to make comparisons between them.
4.2. Physical properties of FAMEs
The prediction of cetane number, kinematic viscosity, density,
and higher heating value for fourteen FAMEs is reported in Table 5.
These values were estimated by means of Eqs. (1)(4). When we
compare the experimental data versus the estimated values of
these properties, we obtained an average absolute deviation
(AAD) for the cetane number, kinematic viscosity, density, and
higher heating value of 5.95%, 2.57%, 0.11% and 0.21%, respectively;
which leads to establish that the empirical correlations here proposed are accurate to reproduce the properties of FAMEs.
In general, we can observe from Table 5 that cetane number,
viscosity, and higher heating value increase because of the increases of molecular weight and all of these properties decrease
as the number of double bonds increases. Unlike that of previous
properties, density increases with decreasing molecular weight
and increasing degree of unsaturation.
In contrast to the prediction method proposed by Klopfenstein
[8], in which the cetane number of each FAME is proportional to
the square of its molecular weight, this paper proposes that the
functionality between the cetane number and the molecular
weight of each FAME is linear, see Eq. (1). This is due mainly to
two reasons, rst we wanted to develop a mathematical expression

107

L.F. Ramrez-Verduzco et al. / Fuel 91 (2012) 102111

Table 5
Comparison between measured and predicted physical properties for fatty acid methyl esters.
/
(reported)

/
(calculated)

g (mm2/s)

g (mm2/s)

q (g/cm3)

q (g/cm3)

(reported at 40 C)

(calculated at 40 C)

(reported at 20 C)

(calculated at 20 C)

C8:0

33.6a
39.7b

40.0

1.20b
1.17e
1.2150f
1.16g

1.15

0.8771e
0.8771f
0.8775g
0.8769k
0.87754l

C10:0

47.2a
51.6b

48.5

1.71b
1.69e
1.7258f
1.69g
1.72h

1.72

C12:0

61.4a
60.8c

56.9

2.43b
2.38e
2.4331f
2.28g
2.43h

C14:0

66.2a

65.4

C16:0

74.5a
74.3c

C16:1

Methyl ester

a
b
c
d
e
f
g
h
i
j
k
l
m
n
o

d (MJ/kg)
(reported)

d (MJ/kg)
(calculated)

0.8773

34.907b
34.72o

34.85

0.8724e
0.8723f
0.8730g
0.8724k
0.87301l

0.8726

36.674b
36.50o

36.56

2.45

0.8691e
0.8698f
0.8695g
0.8694k
0.86952l

0.8692

37.968b
37.87o

37.82

3.3381f
3.23g
3.30h

3.33

0.8671g
0.8671k
0.86713l

0.8665

38.89o

38.79

73.9

4.38b
4.4136f
4.32g
4.38h

4.37

0.8650m

0.8644

39.449b
39.45o

39.56

51.0a

53.3

3.67b
3.67h
3.0642i

3.59

0.8690i

0.8764

39.293b
39.29o

39.30

C18:0

86.9a
75.6c

82.3

5.85b
5.8675f
5.61g
5.85h

5.59

0.8643n

0.8627

40.099b
40.07o

40.18

C18:1

59.3a
56.5c

61.7

4.51b
4.5728f
4.45g
4.51h

4.60

0.8741f
0.875g
0.8740k
0.87362l

0.8746

40.092b
39.91o

39.93

C18:2

38.2a

41.1

3.79

0.8862f
0.890g
0.8866k

0.8865

39.698b
39.70o

39.68

C18:3

22.7b
22.7c
20.6d

20.5

3.65b
3.7028f
3.64g
3.65h
3.14b
3.27g
3.14h
3.2980i

3.11

0.9019i
0.8979k

0.8985

39.342b
39.34o

39.43

C20:0

100.0c

90.8

7.2097j

7.00

0.8632j
0.8633n

0.8613

40.61o

40.70

C20:1

64.8c

70.2

5.77h
5.9108i

5.77

0.8738i

0.8732

40.61o

40.45

C22:1

76.0c

78.7

7.21g
7.33h
6.9556i

7.09

0.8707i
0.8706k
0.87024l

0.8720

40.99o

40.89

Knothe [14].
Knothe [15].
Tong et al. [16].
Papadopoulos et al. [17].
Liew et al. [18].
Pratas et al. [19].
Yuan et al. [20].
Knothe and Steidley [21].
Pratas et al. [22].
Extrapolate value from measurements at higher temperatures given by Pratas et al. [22].
Gouw and Vlugter [1].
Hammomd and Lundberg [23].
Extrapolate value from measurements at higher temperatures given by Bonhorst et al. [24].
Derived from empirical correlation proposed by Gouw and Vlugter [1].
Freedman and Bagby [25].

as simple as possible, and secondly, the experimental results reported by Knothe [14] for saturated FAMEs show a linear behavior
(see Fig. 2). In addition, it is also proposed though Eq. (1) that the

cetane number decreases 20 units for each increase in the number

of double bonds of the corresponding FAME, it relies on the behavior shown by the values reported in Table 5, where a decrease of

108

L.F. Ramrez-Verduzco et al. / Fuel 91 (2012) 102111

Fig. 4. Spatial geometry of the methyl stearate or C18:0 methyl ester molecule.
White, gray and red spheres represent hydrogen, carbon and oxygen atoms,
respectively. (For interpretation of the references to color in this gure legend, the
reader is referred to the web version of this article.)

Fig. 5. Spatial geometry of the methyl elaidate or C18:1 (trans-9) methyl ester
molecule. White, gray and red spheres represent hydrogen, carbon and oxygen
atoms, respectively. (For interpretation of the references to color in this gure
legend, the reader is referred to the web version of this article.)
Fig. 2. Cetane number (/) as a function of molecular weight (M) for methyl esters
from C8:0 to C18:0.

cetane number between pair of FAMEs (C16:0 and C16:1, C18:0

and C18:1, C18:1 and C18:2) can be observed.
In a homologous series of saturated FAMES, viscosity increases
directly with chain length (number of carbon atoms) as it can be
seen in Table 5, which it is expected considering that van der Waals
forces grew with the length of the non-polar part of FAMEs. In this
work, it was founded a linear relationship between kinematic viscosity of saturated FAMEs and its molecular weight in a loglog
plot (see Fig. 3). This evidence led us to the proposal given by Eq.
(2) where the ln(gi) is a function of ln(Mi), which contrast with
the proposal made by Allen et al. [2], who states that viscosity of
methyl esters is dependent on molecular weight in a quadratic
form.
It can be seen from Table 5 that for the same number of carbon
atoms, the unsaturated FAMEs have lower viscosities than the saturated FAMEs. The reason for this phenomenon can be found by a
careful consideration of spatial geometries. The tetrahedral bond
angles on carbon results in a molecular geometry for saturated
FAME that is relatively linear although with zigzags (Fig. 4). This
molecular structure allows to FAME molecules to be rather closely
together, where van der Waals intermolecular interactions between sp3 hybridisation atoms are present. On the other hand,

Fig. 6. Spatial geometry of the methyl oleate or C18:1 (cis-9) methyl ester
molecule. White, gray and red spheres represent hydrogen, carbon and oxygen
atoms, respectively. (For interpretation of the references to color in this gure
legend, the reader is referred to the web version of this article.)

Fig. 7. Spatial geometry of the methyl linoleate or C18:2 (cis-9,12) methyl ester
molecule. White, gray and red spheres represent hydrogen, carbon and oxygen
atoms, respectively. (For interpretation of the references to color in this gure
legend, the reader is referred to the web version of this article.)

Fig. 3. Viscosity (g) at 40 C as a function of molecular weight (M) for methyl esters
from C8:0 to C18:0 in a loglog plot.

the introduction of one or more double bonds in the hydrocarbon

chain in unsaturated FAMEs results in non-bends in the trans-conguration molecules (Fig. 5); and one or more bends in the cis-conguration molecules (Figs. 68). The geometry of the double bond
is almost always a cis-conguration in the FAMEs that biodiesel
contains, and these molecules do not stack very well, in other
words, the cis-conguration is like a coil that hinder the approach
of the sp2 atoms from the double bond of neighboring molecules.
Then, the intermolecular interactions are weaker in unsaturated

L.F. Ramrez-Verduzco et al. / Fuel 91 (2012) 102111

109

Fig. 8. Spatial geometry of the methyl linolenate or C18:3 (cis-9,12,15) methyl ester
molecule. White, gray and red spheres represent hydrogen, carbon and oxygen
atoms, respectively. (For interpretation of the references to color in this gure
legend, the reader is referred to the web version of this article.)

than saturated molecules. As a result, the viscosity is lower for

unsaturated FAMEs.
In regard to the density, Gouw and Vlugter [1] suggested that
the density of fatty acids is inversely proportional to their number
of carbon atoms. According to this proposal it is expected that an
increase in the number of carbon atoms causes a decrease in the
density, and to some extent the decrease is marginal when the
number of carbon atoms is larger. Following a similar idea to that
of Gouw and Vlugter [1], we propose that the density can be estimated from parameters related to the chemical structure of FAMEs.
In particular, we chose the molecular weight and the number of
double bonds. Thus, we propose trough Eq. (3) that the density is
inversely proportional to molecular weight and directly proportional to the degree of unsaturation of FAMEs. Fig. 9 shows the variation of the density of FAMEs as a function of molecular weight,
and it can be seen to t quite well to a reciprocal model, where
the density has more signicant changes when the molecular
weight changes between 153.2380 g/mol and 186.2912 g/mol that
when changes between 298.5038 g/mol and 326.5570 g/mol. Also,
we propose that the density of FAMEs increases in 0.0118 g/cm3 for
each additional double bound in the chemical structure of molecule. Generally, biodiesel is rich in unsaturated compounds
(mainly C18:1 and C18:2 methyl esters, see Table 3) and this feature can explain that biodiesel has a greater density than diesel.
On the other hand, higher heating value increases with chain
length (number of carbon atoms), but the increase is more mean-

Fig. 9. Density (q) at 20 C as a function of molecular weight (M) for methyl esters
from C8:0 to C20:0.

Fig. 10. Higher heating value (d) as a function of molecular weight (M) for methyl
esters from C8:0 to C22:0.

ingful when molecular weight change from 158.2380 g/mol to

186.2912 g/mol than from 326.5570 g/mol to 354.6101 g/mol, as
can be seen in Fig 10. Due the above behavior, we propose that a
hyperbolic t can establish a relationship between higher heating
value and molecular weight of FAMEs (see Eq. (4) and Fig. 10).
From Table 5 we can observe that unsaturated esters have lower
energy content on a weight basis. For example, methyl stearate
has a higher heating value of 40.07 MJ/kg, which is 0.41% higher
than that of methyl oleate (39.91 MJ/kg), which in turn is 0.53%
higher than that of methyl linoletate (39.70 MJ/kg). In this work
it is proposed that the higher caloric values should decrease
0.21 MJ/kg for each increase in the degree of unsaturation of
FAMEs, as can be seen in Eq. (4).
4.3. Physical properties of biodiesel
Cetane number is a relative measure of the delay time between
injection and auto-ignition of fuels. High cetane numbers guarantee good cold start behavior and a smooth run of the engine. In
contrast, fuels with low cetane numbers tend to increase gaseous
and particulate exhaust emissions because of incomplete combustion. The experimental cetane number obtained in this work in the
BT and SB samples are according to the ASTM 6751-08a specication that requires a minimum value of 47 (Table 1). It can be seen
from Table 6 that estimated values of cetane number for these biodiesel samples are close to the experimental results considering
their reproducibility (4.3 in the case of the BT and SB samples
measured in this work). The ADD for the cetane number was
3.33% and 3.36% when the mass and mole fraction were used,
respectively. The maximum prediction error was 5.86% for the cottonseed biodiesel. The predictions errors of this work were minor
that the previously reported by Bamgboye and Hansen [31], who
claim that their methodology is able to predict the cetane number
of biodiesel with an accuracy of 88%.
Viscosity is an important fuel parameter because it affects the
atomization quality, size of drop and penetration. Fuel with high
viscosity tend to form larger droplets on injection which can cause
poor fuel atomization during the spray, increases the engine
deposits, needs more energy to pump the fuel, and wears fuel
pump elements and injectors. High viscosity consequently leads
to poor combustion, increased exhaust smoke and emissions. In
Table 6 the kinematic viscosity at 40 C is given for ve biodiesel
samples.

110

L.F. Ramrez-Verduzco et al. / Fuel 91 (2012) 102111

Table 6
Cetane number and kinematic viscosity at 40 C in ve biodiesel samples. Comparison between experimental and calculated values.
Biodiesel
BT
SB
SF
CO
CT

/b
(reported)

P
/b ni1 wi  /i
(calculated)

P
/b ni1 xi  /i
(calculated)

gb (mm2/s)

gb exp

Pn

(reported)

(calculated)

(calculated)

53.8a, 57.8b
51.7a, 49.0c
50.0c
53.0c
51.2d

56.2
50.2
48.7
51.4
53.8

56.3
50.3
48.8
51.6
54.2

4.36a
4.07a, 3.97e, 3.99f
4.03e
4.18e
4.06e

4.29
4.10
4.05
4.13
4.12

4.28
4.10
4.05
4.13
4.12

i1 wi

 Ingi (mm2/s)

Pn

gb exp

i1 xi

 Ingi (mm2/s)

BT: beef tallow oil methyl ester, SB: soybean oil methyl ester, SF: sunower oil methyl ester, CO: corn oil methyl ester, CT: cottonseed oil methyl ester.
a
Experimental results obtained in this work.
b
Zheng and Hanna [28].
c
Ramos et al. [12].
d
Knothe [15].
e
Alptekin and Canakci [29].
f
Tate et al. [30].

Table 7
Density and higher heating value in ve biodiesel samples. Comparison between experimental and calculated values.
Biodiesel

qb

Pn
3
i1 wi  qi (g/cm )
(calculated at 20 C)

qb

db (MJ/kg)
(reported)

0.8850B,b,

0.8774
0.8810

0.8772
0.8809

0.8840B,c
0.8840B,c
0.8836B,c

0.8820
0.8803
0.8784

0.8819
0.8802
0.8781

39.57a
39.61a, 39.80d,
41.28e
40.00d, 41.33e
41.14e
40.15d, 41.18e

qb (g/cm3)
(reported)

BT
SB
SF
CO
CT

0.8795A,a
0.8812A,a,
0.8845B,c
0.8840B,b,
0.8840B,b,
0.8840B,b,

Pn
3
i1 xi  qi (g/cm )
(calculated at 20 C)

P
db ni1 wi  di (MJ/kg)
(calculated)

P
db ni1 xi  di (MJ/kg)
(calculated)

39.77
39.74

39.77
39.74

39.74
39.75
39.70

39.74
39.75
39.70

BT: Beef tallow oil methyl ester, SB: soybean oil methyl ester, SF: sunower oil methyl ester, CO: corn oil methyl ester, CT: cottonseed oil methyl ester.
A
Density at 20 C.
B
Density at 15 C.
a
Experimental results obtained in this work.
b
Alptekin and Canakci [29].
c
Alptekin and Canakci [32].
d
Krisnangkura [33].
e
Demirbas [34].

The experimental values of viscosity are in the range of 3.97

4.36 mm2/s, which are into the minimum (1.9 mm2/s) and maximum (6.0 mm2/s) range required by the ASTM 6751-08a specication. Table 6 shows that the prediction of the viscosity of biodiesel
samples is very close to the experimental values, because an AAD
of 1.65% and 1.68% was founded in calculations when the mass
and mole fractions of FAMEs were used respectively.
Density affects the mass of fuel injected into the combustion
chamber and the airfuel ratio. This is because fuel injection pumps
meter fuel by volume not by mass and a denser fuel contains a greater mass in the same volume. Table 7 summarizes the comparison between experimental and calculated density in ve biodiesel
samples. The absolute difference in density was 0.0021 g/cm3 and
0.0002 g/cm3 for the BT and SB samples, when mass fraction was
used in the mixing rule, and a close result was founded for the mole
fraction. Due densities of FAMEs were estimated at 20 C by Eq. (3),
this implies that the prediction of biodiesel density through mixing
rules also remains evaluated at this temperature. Of course, in order
to compare the experimental and calculated density of biodiesel it is
necessary to have the same condition in temperature. Although
Table 7 presents a comparison between the experimental and
calculated density of sunower, corn, and cottonseed biodiesel at
different temperatures [29,32], with a difference of 5 C, there would
still be observed that there is a very good concordance in both. On
the other hand, Tat and van Gerpen [35] studied the variation of specic gravity with respect to the temperature for a soybean biodiesel,
and found that the specic gravity increases approximately 0.0033
units for each 5 C decreasing in temperature. Then, if our calculated
density (for the SF, CO and CT samples in Table 7) are corrected by
the effect of temperature, that means, add 0.0033 g/cm3 to the results reported in Table 7, it can be observed that the experimental
and calculated values agree very well.

Higher heating value is the amount of heat produced by the complete combustion of a unit quantity of fuel, this property is obtained
when all products of the combustion are cooled down to the temperature before the combustion and the water vapor formed during combustion is condensed. Higher heating value of biodiesel is
approximately 12% less than that of petro-diesel (39.5741.33 MJ/
kg (Table 7) compared to 46 MJ/kg). The higher heating value of a
fuel increases with increasing carbon number in fuel molecules and
also increases as the ratio of carbon and hydrogen to oxygen and
nitrogen increases [36]. In regard to the prediction, the estimated high
caloric values are in a very close range, with a minimum and maximum of 39.70 MJ/kg and 39.77 MJ/kg, respectively. The AAD between
experimental and calculated high caloric values was 1.92%, when
the mass and mole fractions were used in the mixing rule.
In general, it appears that in the calculation of the four physical
properties (cetane number, kinematic viscosity, density, and higher
heating value) for biodiesel given by Eq. (5), the terms related with
the FAMEs C16:0, C18:0, C18:1, C18:2 and C18:3 have a more signicant contribution in the sum, due the concentration of these
FAMEs is greater with respect to the remainder (FAMEs C14:0,
C16:1, C17:0, C20:0, C20:1, and C22:1).
It is important to mention that the methodology that was developed for the prediction of some physical properties has some
restrictions:
a. Because in order to derive the empirical correlations of
this work was not considered the methyl ricinoleate within
our database, the methodology should not be used in
the biodiesel of castor oil, it must remembered that the
castor oil consists mainly (approximately 90%) of the ricinoleic acid (12-hydroxy-9(Z)-octadecenoic acid) according
to Knothe [15].

L.F. Ramrez-Verduzco et al. / Fuel 91 (2012) 102111

b. It only must be used to predict the density at 20 C and the

kinematic viscosity at 40 C, since the data series that were
compiled in this work were limited to these temperatures.
Efforts for future work could be targeted to the inclusion
of functionality with the temperature for these thermophysical properties.
c. No distinction is made between the cis isomers and trans in
the case of the unsaturated FAMEs.
5. Conclusions
We develop a useful experimental methodology for the identication of FAMEs in biodiesel using a gas chromatography technique and two biodiesel samples made from beef tallow and
soybean were characterized.
We probe that in the case of biodiesel follows that the mass and
mole fraction for each FAME have similar numerical values because
of the narrow range of molecular weight.
Our work develops four new empirical correlations to estimate
the cetane number, kinematic viscosity, density, and higher heating value of FAMEs from two structural features of molecules
(molecular weight and degree of unsaturation). We found that
there is a good agreement between experimental and calculated
physical properties.
A calculation procedure was probed to estimate four physical
properties of biodiesel with accuracy sufcient for engineering
applications; it was found that simple mixing rules based on mass
and mole fraction composition were suitable to connect the physical properties of individual FAMEs according to the biodiesel.
Acknowledgment
We express our gratitude to the Instituto Mexicano del Petrleo for the nancial support during the course of the Project
D.00479.
References
[1] Gouw TH, Vlugter JC. Physical properties of fatty acid methyl esters. I. Density
and molar volume. JAOCS 1964;41:1425.
[2] Allen CAW, Watts KC, Ackman RG, Pegg MJ. Predicting the viscosity of biodiesel
fuels from their fatty acid ester composition. Fuel 1999;78:131926.
[3] Krisnangkura K, Yimsuwan T, Pairintra R. An empirical approach in predicting
biodiesel viscosity at various temperatures. Fuel 2006;85:10713.
[4] Ceriani R, Paiva FR, Goncalves CB, Batista EAC, Meirelles AJA. Densities and
viscosities of vegetable oils of nutritional value. J Chem Eng Data
2008;53:184653.
[5] Allen CAW, Watts KC, Ackmamn RG. Predicting the surface tension of biodiesel
fuels from their fatty acid composition. JAOCS 1999;76:31723.
[6] Ejim CE, Fleck BA, Amirfazli A. Analytical study for atomization of biodiesels
and their blends in a typical injector: surface tension and viscosity effects. Fuel
2007;86:153444.
[7] van Bommel MJ, Oonk HAJ, van Miltenburg JC. Heat capacity measurements of
13 methyl esters of n-carboxilic acids from methyl octanoate to methyl
eicosanoate between 5 K and 350 K. J Chem Eng Data 2004;49:103642.
[8] Klopfenstein WE. Effect of molecular weights of fatty acid esters on cetane
numbers and diesel fuels. JAOCS 1985;62:102931.

111

[9] Benjumea P, Agudelo J, Agudelo A. Basic properties of palm oil biodiesel-diesel

blends. Fuel 2008;87:206975.
[10] Goncalves CB, Ceriani R, Rabelo J, Mafa MC, Meirrelles AJA. Viscosities of fatty
mixtures: experimental data and predictions. J Chem Eng Data
2007;52:20006.
[11] Shu Q, Wang J, Peng B, Wang D, Wang G. Predicting the surface tension of
biodiesel fuels by a mixture topological index method, at 313 K. Fuel
2008;87:358690.
[12] Ramos MJ, Fernandez CM, Casas A, Rodriguez L, Perez A. Inuence of fatty acid
composition of raw material son biodiesel properties. Bioresour Technol
2009;100:2618.
[13] Yuan W, Hansen AC, Zhang Q. Vapor pressure and normal boiling point
predictions for pure methyl esters and biodiesel fuels. Fuel 2005;84:
94350.
[14] Knothe G. Dependence of biodiesel fuel properties on the structure of fatty
acid alkyl esters. Fuel Process Technol 2005;86:105970.
[15] Knothe G. Designer biodiesel: optimizing fatty ester composition to improve
fuel properties. Energy Fuel 2008;22:135864.
[16] Tong D, Hu C, Jiang K, Li Y. Cetane number prediction of biodiesel from the
composition of the fatty acid methyl esters. J Am Oil Chem Soc
2011;88:41523.
[17] Papadopoulos CE, Lazaridou A, Koutsoumba A, Kokkinos N, Christoforidis A,
Nikolaou N. Optimization of cotton seed biodiesel quality (critical properties)
through modication of its FAME composition by highly selective
homogeneous hydrogenation. Bioresour Technol 2010;101:18129.
[18] Liew KY, Seng CE, Oh LL. Viscosities and densities of the methyl esters of some
n-alkanoic acids. JAOCS 1992;69:1558.
[19] Pratas MJ, Freitas S, Oliveira MB, Monteiro SC, Lima AS, Coutinho JAP. Densities
and viscosities of fatty acid methyl and ethyl esters. J Chem Eng Data
2010;55:398390.
[20] Yuan W, Hansen AC, Zhang Q. Predicting the temperature dependent viscosity
of biodiesel fuels. Fuel 2009;88:11206.
[21] Knothe G, Steidley KR. Kinematic viscosity of biodiesel fuel components and
related compounds. Inuence of compound structure and comparison to
petrodiesel fuel components. Fuel 2005;84:105965.
[22] Pratas MJ, Freitas S, Oliveira MB, Monteiro SC, Lima AS, Coutinho JAP. Densities
and viscosities of minority fatty acid methyl and ethyl esters present in
biodiesel. J Chem Eng Data 2011;56:217580.
[23] Hammond EG, Lundberg WO. Molar refraction, molar volume and refractive
index of fatty acid esters and related compounds in the liquid state. J Am Oil
Chem Soc 1954;31:42732.
[24] Bonhorst CW, Althouse PM, Triebold HO. Esters of naturally occurring fatty
acids. Ind Eng Chem 1948;40:237984.
[25] Freedman B, Bagby MO. Heats of combustion of fatty esters and triglycerides.
JAOCS 1989;66:16015.
[26] Srivastava A, Prasad R. Triglycerides-based diesel fuels. Renew Sust Energy Rev
2000;4:11133.
[27] Goering CE, Schwab AW, Daugherty MJ, Pryde EH, Heakin AJ. Fuel properties of
eleven oils. Trans ASAE 1982;25:147283.
[28] Zheng D, Hanna MA. Preparation and properties of methyl esters of beef
tallow. Bioresour Technol 1996;57:13742.
[29] Alptekin E, Canakci M. Characterization of the key fuel properties of methylester fuel blends. Fuel 2009;88:7580.
[30] Tate RE, Watts KC, Allen CAW, Wilkie KI. The viscosities of three biodiesel fuels
at temperatures up to 300 C. Fuel 2006;85:10105.
[31] Bamgboye AI, Hansen AC. Prediction of cetane number of biodiesel fuel from
the fatty acid methyl ester (FAME) composition. Int Agrophys 2008;22:
219.
[32] Alptekin E, Canakci M. Determination of the density and the viscosities of
biodiesel-diesel fuel blends. Renew Energy 2008;33:262330.
[33] Krisnangkura K. Estimation of heat of combustion of triglycerides and fatty
acid methyl esters. JAOCS 1991;68:568.
[34] Demirbas A. Relationships derived from physical properties of vegetable oil
and biodiesel fuels. Fuel 2008;87:17438.
[35] Tat ME, van Gerpen JH. The specic gravity of biodiesel and its blends with
diesel fuel. JAOCS 2000;77:1159.
[36] Demirbas A. Calculation of higher heating values of biomass fuels. Fuel
1997;76:4314.