Fuel 115 (2014) 500506

Contents lists available at ScienceDirect

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

Impact of fatty ester composition on low temperature properties

of biodieselpetroleum diesel blends q
Bryan R. Moser
United States Department of Agriculture, Agricultural Research Service, National Center for Agricultural Utilization Research, 1815 N. University St., Peoria, Illinois IL 61604, USA
Bio-Oils Research Unit, National Center for Agricultural Utilization Research, Agricultural Research Service, United States Department of Agriculture, 1815 N. University St.,
Peoria, IL61604, USA

h i g h l i g h t s
 Seven biodiesels with different fatty acid proles were blended with diesel fuel.
 Inuence of biodiesel type on cold ow properties of blends was determined.
 Fatty acid prole did not affect cold ow properties at low (15%) blend levels.
 Cold ow properties at B1 to B2 were equivalent regardless of fatty acid prole.
 Mathematical relationships were noted between cold ow properties and blend ratio.

a r t i c l e

i n f o

Article history:
Received 5 November 2012
Received in revised form 17 July 2013
Accepted 18 July 2013
Available online 1 August 2013
Keywords:
Biodiesel
Blend
Cold ow properties
Diesel
Fatty acid methyl esters

a b s t r a c t
Several biodiesel fuels along with neat fatty acid methyl esters (FAMEs) commonly encountered in biodiesel were blended with ultra-low sulfur diesel (ULSD) fuel at blend levels permitted by ASTM D975
(B1B5) and cold ow properties such as cloud point (CP), cold lter plugging point (CFPP) and pour point
(PP) were measured. The objective was to determine whether or not the fatty acid composition of biodiesel affects cold ow properties of blends at levels such as B1 to B5. Statistical methods such as least
squares regression and one-way analysis of variance coupled with Tukeys Studentized Range test were
applied to the resulting cold ow property data. Statistical analysis revealed that fatty acid prole did not
affect cold ow properties at low blend levels unless the biodiesel sample contained a high percentage
(>48%) of long-chain saturated FAMEs. Other important conclusions were that variances in cold ow
property data at low blend levels (B1 to B2) were minimal and generally statistically equivalent regardless of fatty acid composition. Lastly, application of least-squares statistical regression to CP, CFPP and PP
data revealed distinct mathematical relationships between cold ow properties and blend ratio. Specifically, CP was best t to an exponential decay model whereas PP exhibited linearity and CFPP provided
the highest R2 values when tted to polynomial equations. In summary, this study demonstrated that
in most cases feedstock selection for biodiesel fuel had minimal impact on cold ow properties at the
blend levels permitted by ASTM D975, the US standard specication for diesel fuel oils.
Published by Elsevier Ltd.

1. Introduction
Abbreviations: AOCS, American Oil Chemists Society; ASTM, American Society
for Testing and Materials; AV, acid value; CEN, European Committee for Standardization; CFPP, cold lter plugging point; CME, canola oil methyl esters; CP, cloud
point; FA, fatty acid; FAME, fatty acid methyl ester; FP, ash point; FPME, eld
pennycress oil methyl esters; kV, kinematic viscosity; mp, melting point; PME, palm
oil methyl esters; PP, pour point; SG, specic gravity; SBME, soybean oil methyl
esters; SFME, sunower oil methyl esters; ULSD, ultra-low sulfur (<15 ppm) diesel;
WCME, waste cooking oil methyl esters; YGME, yellow grease methyl esters.
q
Disclaimer: Mention of trade names or commercial products in this publication
is solely for the purpose of providing specic information and does not imply
recommendation or endorsement by the US Department of Agriculture. USDA is an
equal opportunity provider and employer.
Tel.: +1 309 681 6511; fax: +1 309 681 6524.
E-mail address: Bryan.Moser@ars.usda.gov

0016-2361/$ - see front matter Published by Elsevier Ltd.

http://dx.doi.org/10.1016/j.fuel.2013.07.075

Biodiesel must meet the requirements of fuel standards such as

ASTM D6751 or EN 14214 before it may be blended with petrodiesel. Currently, blends up to B5 (5 vol% biodiesel in petrodiesel) and
B7 are permitted in ASTM D975 and EN 590, the US and European
diesel fuel standards [1,2]. Fuel properties depend on the fatty acid
(FA) composition of the lipid from which biodiesel is derived, and
as a result biodiesel prepared from different feedstocks will have
different FA proles and consequently different fuel properties
[18]. The ve most commonly encountered FAs in plant oils are
palmitic (C16:0; hexadecanoic), stearic (C18:0; octadecanoic), oleic
(C18:1; 9Z-octadecenoic), linoleic (C18:2; 9Z,12Z-octadecadienoic),

501

B.R. Moser / Fuel 115 (2014) 500506

and linolenic (C18:3; 9Z,12Z,15Z-octadecatrienoic) acids. Other

FAs, such as erucic (C22:1; 13Z-docosenoic) acid, are found in oils
from the Brassicaceae family, of which rapeseed and eld pennycress [9] are examples. Biodiesel fuels containing less common
FAs include capric (C10:0; decanoic) acid-containing cuphea [10],
lauric (C12:0; dodecanoic) acid-containing coconut [11], palm kernel [11], and babassu [12], petroselenic (C18:1; 6Z-octadecenoic)
acid-containing coriander [13], and 5Z-eicosenoic (C20:1) acidcontaining meadowfoam [14].
Cold ow is characterized as the ability of a fuel to resist solidication at sub-ambient temperatures. The length of the hydrocarbon tail, nature of the ester group, as well as the location,
orientation and number of double bonds impact melting points
(mps) of FA methyl esters (FAMEs) [1,37]. Consequently, mp increases with increasing hydrocarbon tail length, as indicated by
the mps of FAMEs from palmitic (28 C), stearic (38 C), arachidic
(C20:0; eicosanoic; 46 C), and behenic (C22:0; docosanoic;
53 C) acids (Table 1). Additionally, the decreasing mps of FAMEs
from stearic (38 C), oleic ( 20 C) linoleic ( 43 C) and linolenic
( 52 C) acids illustrate the inuence of unsaturation on mp
(Table 1) [7].
The consensus distilled from several structureproperty relationship studies is that monounsaturated FAMEs are ideal components of biodiesel from a fuel property perspective [3,4,6]. Such
constituents provide a satisfactory balance between cold ow
and oxidative stability as well as between kinematic viscosity
(kV) and cetane number. As such, production of biodiesel enriched
in monounsaturated FAMEs such as oleic acid has been reported
and encouraged [8,1517]. The objective of the current study
was to determine whether or not FA composition impacts cold ow
properties of low-level (B5 or below) blends in ultra-low sulfur
(<15 ppm S) diesel (ULSD) fuel. In other words, does it matter what
the FA composition of biodiesel is at low blend levels? To answer
this question, several biodiesel fuels spanning a wide range of FA
compositions were blended with ULSD (B1B5) and cold ow
properties such as cloud point (CP), cold lter plugging point
(CFPP) and pour point (PP) were determined.
Fuel properties of biodiesel blends with petrodiesel have been
reported extensively. For instance, properties of blends prepared
from camelina [18], canola [19], cottonseed [19,20], distillers
grains [21], microalgae [22], palm [23], poultry fat [20], soybean
[15,19,23,24], and sunower [19] were reported, among numerous
others. Others investigated the inuence of blend ratio (B0B100)
on fuel properties such as oxidative stability [25], kV [26,27],

density [27], specic gravity (SG) [28,29], and cold ow [25] and
have determined that kV, density and SG all increase as the percentage of biodiesel is increased whereas oxidative stability and
cold ow improve signicantly as the biodiesel component is decreased. However, none have determined whether or not FA composition played a role in fuel properties of blends, especially at
blend ratios permitted by ASTM D975. The current investigation
addresses this issue.
2. Materials and methods
2.1. Materials
Fungible ULSD was donated by a petrochemical company.
Waste cooking oil methyl esters (WCME) and yellow grease methyl
esters (YGME) were donated by commercial producers. Rened,
bleached and deodorized canola (low-erucic acid rapeseed), palm,
soybean and sunower (high oleic) oils were purchased from KIC
Chemicals, Inc. (New Platz, NY). Field pennycress seeds were collected from a wild population in Peoria County, IL and oil was expelled as described previously [9]. FAMEs (99.9%) were purchased
from Nu-Chek Prep, Inc. (Elysian, MN). All other chemicals were
obtained from SigmaAldrich Corp (St. Louis, MO). All materials
were used as received.
2.2. Preparation of fatty acid methyl esters
Methanolysis was accomplished with 0.5 mass% sodium methoxide (with respect to oil) and a 6:1 mol ratio of methanol to oil
at 60 C for 1.0 h. Removal of glycerol by gravity separation and
methanol by rotary evaporation (10 mbar; 30 C) was performed
after the mixture cooled to room temperature. FAMEs were
washed with water until a neutral pH was achieved and dried with
MgSO4 to yield canola oil methyl esters (CME; 98 mass%), eld
pennycress oil methyl esters (FPME; 96 mass%), palm oil methyl
esters (PME; 98 mass%), soybean oil methyl esters (SBME;
99 mass%), and sunower oil methyl esters (SFME; 97 mass%).
2.3. Fatty acid composition
FAMEs prepared as described previously [8] were analyzed
using a Varian (Walnut Creek, CA) 8400 GC equipped with an FID
detector and SP2380 (Supelco, Bellefonte, PA) column

Table 1
FAME composition (area%) of CME, FPME, PME, SBME, SFME, WCME, and YGME along with melting points (mps) of individual FAMEs.a

FAME

mp (C)b

CME

FPME

PME

SBME

SFME

WCME

YGME

C12:0
C14:0
C16:0
C16:1 9c
C18:0
C18:1 9c
C18:2 9c, 12c
C18:3 9c, 12c, 15c
C20:0
C20:1 11c
C22:0
C22:1 13c
Unknown (sum)
R satc
R monounsatc
R polyunsatc
Unknown (sum)

4.3
18.1
28.5
34.1
37.7
20.2
43.1
N/R
46.4
7.8
53.2
3.1

4.5 (0.1)
0.3 (0.1)
2.2 (0.2)
62.7 (0.3)
20.6
9.7

0
6.7
63.0
30.3
0

3.1 (0.1)
0.2
0.5
12.6 (0.1)
22.4 (0.1)
11.8 (0.1)
2.4
9.8
0.5
35.0 (0.2)
1.7
6.5
57.6
34.2
1.7

1.2
42.8 (0.3)
0.2
4.4 (0.1)
40.1 (0.4)
10.2
0.2
0.4

0.5
48.8
40.3
10.4
0.5

10.5

4.7
22.7 (0.1)
53.4 (0.1)
8.2 (0.1)

0.5
15.2
22.7
61.6
0.5

6.3 (0.1)

4.3
80.4 (0.1)
7.7
0.3
1.0 (0.1)

0
11.6
80.4
8.0
0

10.9 (0.1)
0.6 (0.1)
4.0
38.1 (0.1)
40.5 (0.1)
4.7
0.4

0.8
15.3
38.7
45.2
0.8

0.5
6.6 (0.1)
0.5 (0.1)
4.3 (0.1)
66.1 (0.2)
17.2 (0.1)
2.0
0.4
0.5
0.5

1.4
12.3
67.1
19.2
1.4

For example, C18:1 9c signies an 18 carbon FA with one cis (c) double bond at carbon 9 (methyl oleate); = not detected; Values in parentheses are standard deviations
(n = 3); where not indicated, standard deviation was zero; N/R = not reported.
b
From [18].
c
Sat = C12:0 + C14:0 + C16:0 + C18:0 + C20:0 + C22:0; R monounsat = C16:1 + C18:1 + C20:1 + C22:1; R polyunsat = C18:2 + C18:3.

502

B.R. Moser / Fuel 115 (2014) 500506

(30 m  0.25 mm i.d., 0.20 lm lm thickness). Carrier gas was He

at 1 mL/min. The oven temperature was initially held at 150 C
for 15 min, then increased to 210 C at 2 C/min, followed by an increase to 220 C at 50 C/min. The injector and detector temperatures were set at 240 C and 270 C, respectively. FAME peaks
(triplicates, means reported) were identied by comparison to
the retention times of reference standards.
2.4. Fuel properties
All data were collected following American Oil Chemists
Society (AOCS), American Society for Testing and Materials (ASTM)
and European Committee for Standardization (CEN) standard
test methods using instrumentation described previously
[8,9,13,14,18,21]: acid value (AV, mg KOH/g), AOCS Cd 3d-63; CP
(C), ASTM D5773; CFPP (C), ASTM D6371; ash point (FP, C),
ASTM D93; free and total glycerol (mass%), ASTM D6584; moisture
content (ppm), ASTM D6304; PP (C), ASTM D5949. For greater precision, PP was measured with a resolution of 1 C instead of the
specied 3 C increment.
2.5. Preparation of biodieselpetrodiesel blends
Precisely measured amounts of CME, FPME, PME, SBME, SFME,
WCME, YGME, methyl erucate, methyl laurate, methyl linoleate,
methyl myristate (C14:0; methyl tetradecanoate), methyl oleate,
methyl palmitate, and methyl stearate were splash blended at
room temperature (2224 C) with ULSD with continuous stirring
to ensure homogeneity. The mixtures were based on volume with
B1.0, B2.0, B3.0, B4.0 and B5.0 blends prepared.
2.6. Statistical analysis
Least-squares statistical regression and one-way analysis of
variance (ANOVA) were conducted using KaleidaGraph version
4.1.1 for Windows by Synergy Software (Reading, PA). CP, CFPP
and PP were compared at each blend level using Tukeys Studentized Range (HSD) test. Results were considered statistically different if p 6 0.05. Each measurement was conducted in triplicate with
mean values along with standard deviations reported.
3. Results and discussion
3.1. Preparation and composition of fatty acid methyl esters
Homogeneous base-catalyzed transesterication afforded
FAMEs employing classic methanolysis conditions elucidated previously [30]. High yields (>95 mass%) were observed, which was
anticipated as CME, PME, SBME, and SFME were prepared from
commercially-available rened oils with low AVs (<0.50 mg KOH/
g) that were readily amenable to base-catalyzed transesterica-

tion. FPME was produced from laboratory-extracted crude oil with

a low AV (0.61 mg KOH/g).
FA proles are presented in Table 1. Methyl oleate was the predominant FAME detected in CME (62.7%), SFME (80.4%) and YGME
(66.1%) whereas methyl linoleate was the principal component of
SBME (53.4%) and WCME (40.5%). The primary constituents in
PME and FPME were methyl palmitate (42.8%) and methyl erucate
(35.0%), respectively. The remainder of the FA proles consisted
primarily of the ve common FAMEs discussed previously. Saturated FAMEs were less than 16% of the overall FA distribution with
the exception of PME (48.8%). Monounsaturated FAMEs were the
most abundant FAMEs in CME (63.0%), FPME (57.6%), SFME
(80.4%), and YGME (67.1%). Polyunsaturated FAMEs were the most
common FAMEs in SBME (61.6%) and WCME (45.2%).
3.2. Fuel properties of biodiesel samples
Presented in Table 2 are the AVs, moisture contents, FPs and
free and total glycerol data for the biodiesel samples along with
a comparison to biodiesel standards ASTM D6751 and EN 14214.
All FAMEs provided AVs signicantly below the maximum threshold of 0.50 mg KOH/g specied in ASTM D6751 and EN 14214. The
free and total glycerol contents of the FAMEs were within the limits specied in the biodiesel standards. FP and moisture content of
all samples were within the specications listed in the biodiesel
standards, where applicable.
Previous studies [5,31] determined that relatively small percentages of saturated FAMEs have a disproportionate effect on
low temperature properties of biodiesel as a result of their high
mps (Table 1). As seen in Table 3, PME exhibited the highest CP
(>20 C), CFPP (12.3 C) and PP (13.7 C) values due to its high content of saturated FAMEs (48.8%; Table 1). CME and FPME, which
contained the lowest levels of saturated FAMEs (<7%), exhibited
the lowest CP, CFPP and PP values. Intermediate cold ow properties were noted for SBME, SFME, WCME and YGME, which contained 1116% saturated FAMEs. Due to the wide variability in
FA compositions, CP, CFPP and PP values differed by as much as
37 C. For example, differences of 36.7, 29.0 and 31.7 C were noted
between the CP, CFPP and PP values of FPME and PME.
3.3. Least-squares statistical regression of biodiesel-ULSD blend data
As observed in previous studies, addition of biodiesel to petrodiesel resulted in a deterioration of cold ow properties relative
to unblended petrodiesel [18,24,25]. Specically, CP, CFPP and PP
increased with increasing concentration of biodiesel in ULSD. For
example, CP increased from 14.6 C to 13.7 C as the percentage
of CME in petrodiesel was increased from B1 to B5 (Table 3). Furthermore, PP increased from 25.0 C (B1) to 23.7 C (B5). Lastly,
an increase in CFPP from 17.0 C to 15.0 C was noted as the
content of CME increased from B1 to B5. Such trends also generally
held true for neat FAMEs blended with ULSD (Table 4).

Table 2
Acid value, ash point, free and total glycerol content, and moisture level of biodiesel samples along with a comparison to the corresponding specications listed in ASTM D6751
and EN 14214.a

a
b

ASTM D6751

EN 14214

CME

FPME

PME

SBME

SFME

WCME

YGME

Acid value (mg KOH/g)

0.50 max

0.50 max

0.01 (0.01)

0.04 (0.01)

0.08 (0.01)

0.09 (0.02)

0.09 (0.01)

0.17 (0.01)

0.17 (0.01)

Glycerol content:
Free (mass%)
Total (mass%)
Flash point (C)
Moisture (ppm)

0.020 max
0.240 max
93 min
b

0.020 max
0.250 max
101 min
500 max

0.001
0.137
183
177 (6)

0.001
0.148
189
226 (7)

0.011
0.066
179
201 (4)

0.009
0.093
181
368 (5)

0.009
0.199
182
244 (4)

0.001
0.123
185
301 (7)

0.001
0.105
184
255 (4)

Values in parentheses are standard deviations from the reported means. For free and total glycerol as well as ash point, n = 1.
Not specied.

503

B.R. Moser / Fuel 115 (2014) 500506

Table 3
Cold ow properties (C) of biodiesel (B100), ULSD (B0) and blends of biodiesel (B1B5) in ULSD along with one-way ANOVA of B1B5 blends.a
Vol%
ULSD:
B0

CP

CFPP

15.7 (0.5)

PP

17.0 (0)

CP

CFPP

PP

25.3 (0.6)

Canola oil methyl esters:

B1
14.6 (0.2)a
B2
14.1 (0.2)b
B3
13.9 (0.2)b
B4
13.8 (0.1)b
B5
13.7 (0.1)b
B100
1.2 (0.1)

17.0
16.3
15.7
15.3
15.0
12.0

(0)a
(0.6)ab
(0.6)bc
(0.6)bc
(0)c
(0)

25.0
24.7
24.3
24.0
23.7
10.3

(0)a
(0.6)ab
(0.6)ab
(0)ab
(0.6)b
(0.6)

Field pennycress oil methyl esters:

14.4 (0.2)a
16.7 (0.6)a
14.1 (0.2)ab
16.7 (0.6)a
13.9 (0.2)b
16.0 (0)ab
13.9 (0)b
15.7 (0.6)ab
13.7 (0.2)b
15.0 (0)b
11.6 (0.1)
12.7 (0.6)

25.3
25.0
24.7
24.0
23.7
17.0

(0.6)a
(0)ab
(0.6)abc
(0)bc
(0.6)c
(0)

Palm oil methyl esters:

B1
14.3
B2
14.1
B3
13.7
B4
13.5
B5
13.1
B100
>20

16.3
15.3
14.3
13.7
13.0
12.3

(0.6)a
(0.6)ab
(0.6)bc
(0.6)c
(0)c
(0.6)

24.7
24.0
22.3
21.0
20.0
13.7

(0.6)a
(0)a
(0.6)b
(0)c
(0)d
(0.6)

Soybean oil methyl esters:

14.4 (0.1)a
14.3 (0.2)a
14.1 (0.1)ab
13.9 (0.1)bc
13.8 (0.2)c
1.5 (0.2)

Sunower oil methyl esters:

B1
14.4 (0.1)a
B2
14.2 (0.1)ab
B3
14.0 (0.3)ab
B4
13.9 (0.3)ab
B5
13.8 (0.3)b
B100
3.6 (0.1)

16.7
16.0
15.7
15.3
15.0
0.7

(0.6)a
(0)ab
(0.6)ab
(0.6)b
(0)b
(1.5)

25.0
24.7
24.3
24.0
24.0
2.0

(0)a
(0.6)ab
(0.6)ab
(0)b
(0)b
(0)

Waste
14.5
14.4
14.0
13.9
13.7
0.6

Yellow grease methyl esters:

B1
14.4 (0.1)a
B2
14.2 (0.1)ab
B3
13.9 (0.1)bc
B4
13.8 (0.2)c
B5
13.7 (0.2)c
B100
2.5 (0.1)

17.0
16.3
16.0
15.7
15.0
0.3

(0)a
(0.6)ab
(0)b
(0.6)bc
(0)c
(0.6)

25.0
24.7
24.3
24.0
23.7
5.0

(0)a
(0.6)ab
(0.6)ab
(0)ab
(0.6)b
(0)

(0.3)a
(0.1)ab
(0.1)b
(0.2)bc
(0.2)c

17.0
16.3
16.0
15.7
15.0
3.0

(0)a
(0.6)ab
(0)b
(0.6)bc
(0)c
(0)

25.3
24.7
24.3
24.0
23.7
0.0

(0.6)a
(0.6)ab
(0.6)ab
(0)ab
(0.6)b
(0)

cooking oil methyl esters:

(0.2)+
16.7
(0.2)a
16.3
ab
(0.2)
16.0
(0.3)ab
15.3
(0.2)b
15.0
(0.2)
4.0

(0.6)a
(0.6)ab
(1.0)ab
(0.6)b
(0)b
(0)

25.3
24.7
24.7
24.3
23.7
2.0

(0.6)a
(0.6)ab
(0.6)ab
(0.6)ab
(0.6)b
(0)

a
Mean cold ow properties (with standard deviations in parenthesis) within a column followed by the same letter(s) are not signicantly different based on Tukeys HSD
test at p 6 0.05 with 14 error degrees of freedom.

Cold ow property data presented were subjected to regression

in an effort to elucidate the mathematical relationship(s) between
low temperature behavior and blend ratio. Depicted in Table 5
are the best t equations obtained from regression of blends data
(B0B5) along with the corresponding coefcients of correlation
(R2). Regression was performed on CP, PP and CFPP blend
data (B0B5) of biodiesel and neat FAME samples. Because blend
levels covered by ASTM D975 were of principal interest in this
study, B100 samples were not included in the statistical regressions.
The best t obtained for CP data after regression was with the
exponential decay equation (y = m1 + m2 + em3x). In all cases, R2 values were in excess of 0.95. Such a curve is indicative of signicant
changes in the dependent variable (CP) at near-zero levels (B0B1)
of the independent variable (blend ratio) with the effect diminishing as the value for the independent variable increased. In other
words, addition of 1% biodiesel (B1) to ULSD resulted in a signicantly higher DCP (DCP = CPBX CPBX 1) relative to DCPs observed
between B1 and B5. The reason postulated for this difference was
that enough biodiesel was present in the B1 blends to provide sufcient crystal growth of FAMEs at higher sub-ambient temperatures to be visible to the naked eye. Further addition of biodiesel
up to B5 had a smaller effect on CP because the B1 sample already
contained FAMEs that underwent crystal growth at temperatures
higher than that of unblended ULSD.
Regression of PP data resulted in highly linear ts for the biodiesel samples. However, only C16:0 and C18:0 provided linear ts for
the neat FAME samples. The PPs of the remaining neat FAME samples were largely unchanged from B0 to B5, thus they could not be
subjected to statistical regression. In cases where regression was
applicable, R2 > 0.92 were noted. The linearity of PP revealed a directly proportional relationship between PP and blend ratio. As the
content of biodiesel in blends was increased, the accumulation of
crystals resulting in restricted owability increased.

The best t obtained for CFPP data after regression was with a
polynomial expression (y = mo + m1x + m2x2). With regard to the
biodiesel samples, R2 values in excess of 0.96 were noted. Lower
values were calculated (R2 > 0.86) for the neat FAMEs, although
C22:1, C16:0 and C18:0 yielded R2 in excess of 0.91.
3.4. Statistical signicance of blend ratio on cold ow properties of
blends
Although least squares statistical regression clearly revealed
mathematical relationships between cold ow property data and
biodiesel blend ratio, the issue of statistical signicance was not
addressed. Tables 3 and 4 contain one-way ANOVAs along with Tukeys HSD tests which address statistical signicance. Clearly, statistically signicant differences were observed for CP, PP and
CFPP between B0 and B1, as well as between B5 and B100. The
principal question that will be addressed in this section is: within
a given biodiesel fuel type, does cold ow property data (CP, PP,
CFPP) vary signicantly between B1 and B5?
Statistically signicant differences were noted for CP data as the
blend ratio varied between B1 and B5 for all biodiesel fuels. However, in several cases statistically equivalent CPs were obtained for
B1 through B3 (SBME) and B1 through B4 (SFME and WCME). For
example, the CPs of the B1 and B4 blends of SFME in ULSD were
14.4 0.1 and 13.9 0.3 C, respectively, which were not statistically different according to Tukeys HSD test where p 6 0.05
(n = 3; 14 error degrees of freedom) was needed to indicate significant difference. With regard to neat FAMEs, CPs for the B1B5
blends of C22:1, C12:0 and C18:1 were statistically equivalent.
Both C18:2 and C16:0 yielded statistically equivalent results for
B1 through B3 with C14:0 affording equivalent CPs from B1
through B4. Only 18:0 provided statistically different CPs at each
blend level.

504

B.R. Moser / Fuel 115 (2014) 500506

Table 4
Cold ow properties (C) of neat FAMEs blended (B1B5) with ULSD along with one-way ANOVA of B1B5 blends.a
Vol%

CP

ULSD:
0

CFPP

15.7 (0.5)

PP

CP

CFPP

PP

17.0 (0)

25.3 (0.6)

16.3
16.0
15.7
15.7
15.3

(1.2)
(0)
(0.6)
(0.6)
(0.6)

25.0
25.0
25.0
25.0
24.7

(0)
(0)
(0)
(0)
(0.6)

Methyl laurate (C12:0):

14.6 (0.2)
14.5 (0.1)
14.4 (0.3)
14.3 (0.1)
14.2 (0.3)

16.3
15.7
15.7
15.7
15.0

(0.6)a
(0.6)ab
(0.6)ab
(0.6)ab
(0)b

25.0
25.0
25.0
24.7
24.7

(0)
(0)
(0)
(0.6)
(0.6)

Methyl linoleate (C18:2):

B1
14.7 (0.1)a
B2
14.6 (0.1)ab
B3
14.5 (0.2)ab
B4
14.4 (0.1)b
B5
14.4 (0.2)b

16.0
16.0
15.7
15.7
15.3

(0)
(0)
(0.6)
(0.6)
(0.6)

25.7
25.0
25.0
25.0
24.7

(0.6)a
(0)ab
(0)ab
(0)ab
(0.6)b

Methyl myristate (C14:0):

14.6 (0.1)a
14.5 (0.3)ab
14.3 (0.2)ab
14.2 (0.3)ab
14.1 (0.1)b

16.0
15.7
15.7
15.7
15.3

(0)
(0.6)
(0.6)
(0.6)
(0.6)

25.0
25.0
25.0
24.7
24.7

(0)
(0)
(0)
(0.6)
(0.6)

Methyl oleate (C18:1):

B1
14.6
B2
14.6
B3
14.5
B4
14.4
B5
14.4

16.3
15.7
15.7
15.7
15.0

(0.6)a
(0.6)ab
(0.6)ab
(0.6)ab
(0)b

25.0
25.0
25.0
24.7
24.7

(0)
(0)
(0)
(0.6)
(0.6)

Methyl palmitate (C16:0):

14.4 (0.1)a
14.2 (0.1)a
14.2 (0.1)a
13.8 (0.2)b
13.7 (0.1)b

16.3
16.0
15.7
15.7
14.7

(0.6)a
(0)ab
(0.6)ab
(0.6)ab
(0.6)b

23.3
21.3
20.3
19.3
18.0

(0.6)a
(0.6)b
(0.6)bc
(0.6)cd
(0)d

13.7
13.3
13.0
10.0
9.0

(0.6)a
(0.6)a
(0)a
(0)b
(0)c

20.3
16.0
14.0
12.0
9.7

(0.6)a
(0)b
(0)c
(0)d
(0.6)e

Methyl erucate (C22:1):

B1
14.6
B2
14.4
B3
14.4
B4
14.3
B5
14.2

(0.3)
(0.2)
(0.1)
(0.2)
(0.2)

(0.1)
(0.1)
(0.1)
(0.2)
(0.2)

Methyl stearate (C18:0):

B1
13.4 (0.1)a
B2
12.3 (0.1)b
B3
11.1 (0.1)c
B4
7.5 (0)d
B5
6.3 (0.2)e

a
Mean cold ow properties (with standard deviations in parenthesis) within a column followed by the same letter(s) are not signicantly different based on Tukeys HSD
test at p 6 0.05 with 14 error degrees of freedom. Those columns with no letters following the mean values were not signicantly different, as they had p > 0.05.

Table 5
Least-squares statistical regression of CP, PP and CFPP data of biodiesel and neat
FAME blend sets (B0B5). Depicted are the best ts along with the corresponding
coefcients of correlation (R2).a,b

CME
FPME
PME
SBME
SFME
WCME
YGME
C22:1
C12:0
C18:2
C14:0
C18:1
C16:0
C18:0

CP
Best t

CP
R2

CFPP
Best t

CFPP
R2

PP
Best t

PP
R2

ED
ED
ED
ED
ED
ED
ED
ED
ED
ED
ED
ED
ED
ED

0.9643
0.9903
0.9697
0.9685
0.9872
0.9669
0.9852
0.9882
0.9794
0.9879
0.9762
0.9814
0.9516
0.9709

Poly
Poly
Poly
Poly
Poly
Poly
Poly
Poly
Poly
Poly
Poly
Poly
Poly
Poly

0.9637
0.9773
0.9955
0.9702
0.9889
0.9888
0.9702
0.9613
0.8947
0.8769
0.8661
0.8947
0.9115
0.9229

Linear
Linear
Linear
Linear
Linear
Linear
Linear
N/A
N/A
N/A
N/A
N/A
Linear
Linear

0.9982
0.9272
0.9768
0.9670
0.9623
0.9227
0.9982

0.9757
0.9563

a
ED = exponential decay (y = m1 + m2  e m3x); linear: y = mx + b; N/A = no t
(slope = 0); poly = polynomial (y = m0 + m1x + m2x2).
b
C12:0, methyl laurate; C14:0, methyl myristate; C16:0, methyl palmitate;
C18:0, methyl stearate; C18:1, methyl oleate; C18:2, methyl linoleate; C22:1,
methyl erucate.

Statistically equivalent CFPPs were noted for B1 and B2 blends

of CME, PME, SBME, and YGME. FPME yielded equivalent CFPPs
from B1 through B4 whereas SFME and WCME provided equivalent
CFPPs from B1 to B3. In all biodiesel fuel cases, the CFPP values of
B1 blends varied signicantly from B5 blends. However, C22:1,
C18:2 and C14:0 among the neat FAMEs yielded statistically equivalent values for CFPP from B1 to B5. Most of the remaining neat
FAMEs (C12:0, C18:1 and C16:0) provided equivalent CFPPs from
B1 through B4 with only the B5 blends affording statistically different values. As was the case with the CP data, C18:0 provided the
greatest number of signicantly different CFPPs as the blend ratio

varied between B1 and B5, although equivalent CFPPs were noted

from B1 to B3.
With regard to PP data, statistical equivalence was more common than with CP or CFPP due to the greater standard deviations
obtained for PP data. For instance, among the neat FAMEs, C22:1,
C12:0, C14:0, and C18:1 yielded equivalent CFPP values from B1
to B5 whereas B1 to B4 blends were equivalent for C18:2. Only
18:0 provided statistically different CFPPs at each blend level. With
regard to the biodiesel samples, nearly all provided equivalent
CFPPs from either B1 to B3 (FPME, SFME) or B1 to B4 (CME, SBME,
WCME, and YGME). PME afforded equivalent CFPP values at B1 and
B2 while the remaining blends exhibited statistically different
values.
In summary, statistically equivalent results for CP, PP and CFPP
were generally noted at low blend levels (B1 and B2). Statistical
equivalence beyond the B2 blend level was also observed in many
cases. These results indicated that variances in cold ow property
data at low blend levels (B1 to B2) were minimal. Furthermore,
neat FAMEs tended to vary to lesser degrees than the biodiesel
samples. The reason for this reduced variability among the neat
FAMEs may be due to the complex compositions of the biodiesel
samples, as they were comprised of mixtures of several FAMEs.
The neat FAMEs had greater compositional uniformity and consequently behaved in a more orderly fashion than the more complex
biodiesel samples.
3.5. Statistical signicance of biodiesel type on cold ow properties of
blends
The previous two sections established mathematical relationships between cold ow property data and biodiesel blend ratio
as well as the statistical signicance of differences observed in
CP, PP and CFPP data at low blend levels. The objective of this section was to determine whether or not variations in biodiesel fuel
type result in statistically different cold ow property data at

505

B.R. Moser / Fuel 115 (2014) 500506

Table 6
One-way ANOVA of cold ow property data of B1, B2, B3, B4, and B5 blends of biodiesel fuels in ULSD.a
CP

CFPP

PP

CP

CFPP

PP

B1:
CME
FPME
PME
SBME
SFME
WCME
YGME

14.6
14.4
14.3
14.4
14.4
14.5
14.4

(0.2)
(0.2)
(0.3)
(0.1)
(0.1)
(0.3)
(0.1)

17.0
16.7
16.3
17.0
16.7
16.7
17.0

(0)
(0.6)
(0.6)
(0)
(0.6)
(0.6)
(0)

25.0
25.3
24.7
25.3
25.0
25.3
25.0

(0)
(0.6)
(0.6)
(0.6)
(0)
(0.6)
(0)

B2:
CME
FPME
PME
SBME
SFME
WCME
YGME

14.1
14.1
14.1
14.3
14.2
14.4
14.2

(0.2)
(0.2)
(0.1)
(0.2)
(0.1)
(0.2)
(0.1)

16.3
16.7
15.3
16.3
16.0
16.3
16.3

(0.6)
(0.6)
(0.6)
(0.6)
(0)
(0.6)
(0.6)

24.7
25.0
24.0
24.7
24.7
24.7
24.7

(0.6)
(0)
(0)
(0.6)
(0.6)
(0.6)
(0.6)

B3:
CME
FPME
PME
SBME
SFME
WCME
YGME

13.9
13.9
13.7
14.1
14.0
14.0
13.9

(0.2)
(0.2)
(0.1)
(0.1)
(0.3)
(0.2)
(0.1)

15.7
16.0
14.3
16.0
15.7
16.0
16.0

(0.6)b
(0)b
(0.6)a
(0)b
(0.6)b
(0)b
(0)b

24.3
24.7
22.3
24.3
24.3
24.7
24.3

(0.6)b
(0.6)b
(0.6)a
(0.6)b
(0.6)b
(0.6)b
(0.6)b

B4:
CME
FPME
PME
SBME
SFME
WCME
YGME

13.8
13.9
13.5
13.9
13.9
13.9
13.8

(0.1)
(0)
(0.1)
(0.1)
(0.2)
(0)
(0.1)

15.3
15.7
13.7
15.7
15.3
15.3
15.7

(0.6)b
(0.6)b
(0.6)a
(0.6)b
(0.6)b
(0.6)b
(0.6)b

24.0
24.0
21.0
24.0
24.0
24.3
24.0

(0)b
(0)b
(0)a
(0)b
(0)b
(0.6)b
(0)b

B5:
CME
FPME
PME
SBME
SFME
WCME
YGME

13.7
13.7
13.1
13.8
13.8
13.7
13.7

(0.1)b
(0.2)b
(0.2)a
(0.2)b
(0.1)b
(0.2)b
(0.1)b

15.0
15.0
13.0
15.0
15.0
15.0
15.0

(0)b
(0)b
(0)a
(0)b
(0)b
(0)b
(0)b

23.7
23.7
20.0
23.7
24.0
23.7
23.7

(0.6)b
(0.6)b
(0)a
(0.6)b
(0)b
(0.6)b
(0.6)b

a
Mean cold ow properties (with standard deviations in parenthesis) within a column followed by the same letter(s) are not signicantly different based on Tukeys HSD
test at p 6 0.05 with 20 error degrees of freedom. Those columns with no letters following the mean values were not signicantly different, as they had p > 0.05.

Table 7
One-way ANOVA of cold ow property data of B1, B2, B3, B4, and B5 blends of neat FAMEs in ULSD.a,b
CP

CFPP

PP

B1:
C12:0
C14:0
C16:0
C18:0
C18:1
C18:2
C22:1

14.6
14.6
14.4
13.4
14.6
14.7
14.6

(0.2)b
(0.1)a
(0.1)a
(0.1)a
(0.1)a
(0.1)b
(0.3)b

16.3
16.0
16.3
13.7
16.3
16.0
16.3

(0.6)a
(0)a
(0.6)a
(0.6)a
(0.6)a
(0)b
(1.2)b

25.0
25.0
23.3
20.3
25.0
25.7
25.0

B3:
C12:0
C14:0
C16:0
C18:0
C18:1
C18:2
C22:1

14.4
14.3
14.2
11.1
14.5
14.5
14.4

(0.3)bc
(0.2)bc
(0.1)b
(0.1)a
(0.1)c
(0.1)c
(0.1)bc

15.7
15.7
15.7
13.0
15.7
15.7
15.7

(0)b
(0.6)b
(0.6)b
(0)a
(0.6)b
(0.6)b
(0.6)b

B5:
C12:0
C14:0
C16:0
C18:0
C18:1
C18:2
C22:1

14.2
14.1
13.7
6.3
14.4
14.4
14.2

(0.3)c
(0.1)bc
(0.1)b
(0.2)a
(0.2)c
(0.2)c
(0.2)c

15.0
15.3
14.7
9.0
15.0
15.3
15.3

(0)b
(0.6)b
(0.6)b
(0)a
(0)b
(0.6)b
(0.6)b

CP

CFPP

PP

(0)c
(0)c
(0.6)a
(0.6)a
(0)c
(0.6)c
(0)c

B2:
C12:0
C14:0
C16:0
C18:0
C18:1
C18:2
C22:1

14.5
14.5
14.2
12.3
14.6
14.6
14.4

(0.1)a
(0.3)a
(0.1)a
(0.1)a
(0.1)b
(0.1)b
(0.2)b

15.7
15.7
16.0
13.3
15.7
16.0
16.0

(0.6)a
(0.6)a
(0)a
(0.6)a
(0.6)b
(0)b
(0)b

25.0
25.0
21.3
16.0
25.0
25.0
25.0

(0)c
(0)c
(0.6)a
(0)a
(0)c
(0)c
(0)c

25.0
25.0
20.3
14.0
25.0
25.0
25.0

(0)c
(0)c
(0.6)b
(0)a
(0.6)c
(0.6)c
(0)c

B4:
C12:0
C14:0
C16:0
C18:0
C18:1
C18:2
C22:1

14.3
14.2
13.8
7.5
14.4
14.4
14.3

(0.1)c
(0.3)c
(0.2)b
(0)a
(0.2)c
(0.2)c
(0.2)c

15.7
15.7
15.7
10.0
15.7
15.7
15.7

(0.6)b
(0.6)b
(0.6)b
(0)a
(0.6)b
(0.6)b
(0.6)b

24.7
24.7
19.3
12.0
24.7
24.7
25.0

(0.6)c
(0.6)c
(0.6)b
(0)a
(0.6)c
(0.6)c
(0)c

24.7
24.7
18.0
9.7
24.7
24.7
24.7

(0.6)c
(0.6)c
(0)b
(0.6)a
(0.6)c
(0.6)c
(0.6)c

a
Mean cold ow properties (with standard deviations in parenthesis) within a column followed by the same letter(s) are not signicantly different based on Tukeys HSD
test at p 6 0.05 with 20 error degrees of freedom.
b
C12:0, methyl laurate; C14:0, methyl myristate; C16:0, methyl palmitate; C18:0, methyl stearate; C18:1, methyl oleate; C18:2, methyl linoleate; C22:1, methyl erucate.

similar blend ratios. Presented in Tables 6 and 7 are CP, CFPP and
PP data of blends with ULSD organized by blend ratio (as opposed
to Tables 3 and 4 where the data was organized by fuel type). Also
depicted in Tables 6 and 7 are the results of one-way ANOVAs coupled with Tukeys HSD tests to establish statistical signicance. Results were considered statistically different (n = 3; 20 error degrees
of freedom) if p > 0.05 (p values not shown).
Statistically equivalent values for CP, CFPP and PP were noted at
both B1 and B2 for the biodiesel samples (Table 6). At B3 and B4, all

biodiesel samples provided equivalent CPs but PME yielded statistically different CFPP and PP values when compared to the other
samples. All of the cold ow properties of PME were statistically
different at the B5 level when compared to the other samples, with
all other samples yielding equivalent values. Based on these results, feedstocks with saturated FAME contents below 16% provide
statistically equivalent values for CP, CFPP and PP at all blend ratios
in the B1 to B5 range. As seen in Table 1, CME, FPME, SBME, SFME,
WCME, and YGME contained between 6.5% and 15.3% combined

506

B.R. Moser / Fuel 115 (2014) 500506

saturated FAME content. PME, with a saturated FAME content of

48.8%, consistently exhibited statistically different results for CFPP
and PP at the B3 blend level and above. However, at low blend levels such as B1 and B2 even PME with its comparatively high saturated FAME content provided CP, CFPP and PP values that were
statistically equivalent to the other biodiesel fuels. The reason for
the disparate behavior of PME relative to the other samples is postulated to be due to the higher mps of saturated FAMEs relative to
constituents more commonly encountered in the other biodiesel
samples (unsaturated FAMEs).
The ANOVA results presented in Table 7 for the neat FAMEs concurred with the trends observed for the biodiesel samples. Specifically, long-chain (C16 and above) saturated FAMEs (C16:0 and
C18:0) provided statistically different cold ow properties from
the other neat FAMEs at all blend levels. Shorter-chain saturated
FAMEs such as C12:0 and C14:0 were not statistically different
from the unsaturated FAMEs (C22:1, C18:1 and C18:2) due to their
lower mps relative to C16:0 and C18:0. However, C16:0 and C18:0
were statistically different from C12:0 and C14:0 as well as being
generally different from each other. At all blend levels, C18:0 provided statistically different results from C12:0, C14:0, C18:1, C18:2
and C22:1 for CP, CFPP and PP. C16:0, with its lower mp versus
C18:0, afforded statistically equivalent CFPPs to the remainder of
neat FAMEs (aside from C18:0) at all blend levels but yielded different PPs at all blend levels. Consequently, C18:0 was the most
different from the remainder of the neat FAMEs followed by
C16:0, which paralleled the mp behavior of these constituents.
4. Conclusions
FA prole did not affect cold ow properties at low blend levels
(B5 and below) unless the biodiesel sample contained a high percentage of long-chain (C16 and above) saturated FAMEs. Other
important conclusions were that variances in cold ow property
data at low blend levels (B1 to B2) were generally statistically
equivalent regardless of the feedstock. Lastly, application of
least-squares statistical regression to CP, CFPP and PP data revealed
distinct mathematical relationships between cold ow properties
and blend ratio. Specically, CP was best t to an exponential decay model whereas PP exhibited linearity and CFPP provided the
highest R2 values when tted to polynomial equations. In summary, these results should be useful to biodiesel producers and
blenders who are concerned with cold ow property issues, as this
study demonstrated that in most cases feedstock selection for biodiesel fuel had minimal impact on cold ow properties at low
blend levels permitted by ASTM D975.
Acknowledgement
The author acknowledges Benetria N. Banks for excellent technical assistance.
References
[1] Moser BR. Biodiesel production, properties, and feedstocks. In vitro Cell Dev
Biol-Plant 2009;45:22966.
[2] Moser BR. Biodiesel from alternative oilseed feedstocks: camelina and eld
pennycress. Biofuels 2012;3:193209.

[3] Knothe G. Dependence of biodiesel fuel properties on the structure of fatty acid
alkyl esters. Fuel Process Technol 2005;86:105970.
[4] Knothe G. Designer biodiesel: optimizing fatty ester composition to improve
fuel properties. Energy Fuel 2008;22:135864.
[5] Moser BR. Inuence of blending canola, palm, soybean, and sunower oil
methyl esters on fuel properties of biodiesel. Energy Fuel 2008;22:43016.
[6] Knothe G. Improving biodiesel fuel properties by modifying fatty ester
composition. Energy Environ Sci 2009;2:75966.
[7] Knothe G, Dunn RO. A comprehensive evaluation of the melting points of fatty
acids and esters determined by differential scanning calorimetry. J Am Oil
Chem Soc 2009;86:84356.
[8] Moser BR, Vaughn SF. Efcacy of fatty acid prole as a tool for screening
feedstocks for production of biodiesel. Biomass Bioenergy 2012;37:3141.
[9] Moser BR, Knothe G, Vaughn SF, Isbell TA. Production and evaluation of
biodiesel from eld pennycress (Thlaspi arvense L.) oil. Energy Fuel
2009;23:414955.
[10] Knothe G, Cermak SC, Evangelista RL. Cuphea oil as a source of biodiesel with
improved fuel properties caused by high content of methyl decanoate. Energy
Fuel 2009;23:17437.
[11] Schober S, Steidl I, Mittelbach M. Ester content evaluation in biodiesel from
animal fats and lauric oils. Eur J Lipid Sci Technol 2006;108:30914.
[12] Silva RB, Neto AFL, Santos LSS, Oliveira Lima JR, Chaves MH, Santos JR, et al.
Catalysts of Cu(II) and Co(II) ions adsorbed in chitosan used in
transesterication of soybean and babassu oils a new route for biodiesel
synthesis. Bioresource Technol 2008;99:67938.
[13] Moser BR, Vaughn SF. Coriander seed oil methyl esters as biodiesel fuel:
unique fatty acid composition and excellent oxidative stability. Biomass
Bioenergy 2010;34:5508.
[14] Moser BR, Knothe G, Cermak SC. Biodiesel from meadowfoam (Limnanthes alba
L.) seed oil: oxidative stability and usual fatty acid composition. Energy
Environ Sci 2010;3:31827.
[15] Tat ME, Wang PS, Van Gerpen JH, Clemente TE. Exhaust emissions from an
engine fueled with biodiesel from high-oleic soybeans. J Am Oil Chem Soc
2007;84:8659.
[16] Graef G, LaVallee BJ, Tenopir P, Tat M, Schweiger B, Kinney AJ, et al. A higholeic-acid and low-palmitic-acid soybean: agronomic performance and
evaluation as a feedstock for biodiesel. Plant Biotechnol J 2009;7:41121.
[17] Pighinelli ALMT, Ferrari RA, Miguel AMRO, Park KJ. High-oleic sunower
biodiesel: quality control and different purication methods. Grasas Aceites
2011;62:17180.
[18] Moser BR, Vaughn SF. Evaluation of alkyl esters from Camelina sativa oil as
biodiesel and as blend components in ultra low-sulfur diesel fuel. Bioresource
Technol 2010;101:64653.
[19] Alptekin E, Canakci M. Characterization of the key fuel properties of methyl
ester-diesel fuel blends. Fuel 2009;88:7580.
[20] Tang H, Salley SO, Ng KYS. Fuel properties and precipitate formation at low
temperature in soy-, cottonseed-, and poultry fat-based biodiesel blends. Fuel
2008;87:300617.
[21] Moser BR, Vaughn SF. Biodiesel from corn distillers dried grains with solubles:
preparation, evaluation, and properties. Bioenergy Res 2012;5:43949.
[22] Chen YH, Huang BY, Chiang TH, Tang TC. Fuel properties of microalgae
(Chlorella protothecoides) oil biodiesel and its blends with petroleum diesel.
Fuel 2012;94:2703.
[23] Moser BR, Cermak SC, Isbell TA. Evaluation of castor and lesquerella oil
derivatives as additives in biodiesel and ultralow sulfur diesel fuels. Energy
Fuel 2008;22:134952.
[24] Moser BR, Williams A, Haas MJ, McCormick RL. Exhaust emissions and fuel
properties of partially hydrogenated soybean oil methyl esters blended with
ultra low sulfur diesel fuel. Fuel Process Technol 2009;90:11228.
[25] Sharafutdinov I, Stratiev D, Shishkova I, Dinkov R, Batchvarov A, Petkov P, et al.
Cold ow properties and oxidation stability of blends of near zero sulfur diesel
from Ural crude oil and FAME from different origin. Fuel 2012;96:55667.
[26] Tat ME, Van Gerpen JH. The kinematic viscosity of biodiesel and its blends with
diesel fuel. J Am Oil Chem Soc 1999;76:15113.
[27] Alptekin E, Canakci M. Determination of the density and the viscosities of
biodieseldiesel fuel blends. Renew Energy 2008;33:262330.
[28] Dunn RO. Specic gravity and API gravity of biodiesel and ultra-low-sulfur
diesel (ULSD) blends. T ASABE 2011;54:5719.
[29] Moser BR. Efcacy of specic gravity as a tool for prediction of biodiesel
petrodiesel blend ratio. Fuel 2012;99:25461.
[30] Freedman B, Pryde EH, Mounts TL. Variables affecting the yields of fatty esters
from transesteried vegetable oils. J Am Oil Chem Soc 1984;61:163843.
[31] Dunn RO. Crystallization behavior of fatty acid methyl esters. J Am Oil Chem
Soc 2008;85:96172.