Fuel 194 (2017) 144–156

Contents lists available at ScienceDirect

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

Full Length Article

Comparative study of NOx emissions of biodiesel-diesel blends from

soybean, palm and waste frying oils using methyl and ethyl
transesterification routes
Marcelino Aurélio Vieira da Silva a, Beatriz Lagnier Gil Ferreira a,⇑, Luiz Guilherme da Costa Marques b,
Aurélio Lamare Soares Murta c, Marcos Aurelio Vasconcelos de Freitas b
a
Transport Engineering Program (PET/COPPE), Federal University of Rio de Janeiro, Rio de Janeiro, Brazil
b
International Virtual Institute of Climate Change (IVIG), Federal University of Rio de Janeiro, Rio de Janeiro, Brazil
c
Fluminense Federal University, Niterói, Brazil

h i g h l i g h t s

20% soybean methyl ester presented lowest emissions compared to the reference fuel.

Blends with palm ethyl ester were amongst blends with best emissions at all loads.

20% palm methyl ester had the highest NOx emissions.

Waste frying oils increase NOx emissions comparing to diesel.

Ethyl esters blends showed lower NOx emissions than methyl esters blends.

a r t i c l e i n f o a b s t r a c t

Article history: The research for renewable and less polluting fuels has focused on biodiesel. This fuel can derive from
Received 12 September 2016 vegetable, animal or waste oils, and despite its potential to decrease atmospheric pollutants and green-
Received in revised form 21 December 2016 house gases, its influence on NOx emissions is still uncertain. It is believed that biodiesel emissions, espe-
Accepted 24 December 2016
cially NOx vary depending on the feedstock, blend percentage and transesterification route. A better
Available online 9 January 2017
understanding of these factors can help choosing the best blend. In this context, this article aims at eval-
uating how the variation of these factors affects NOx emissions. Tests are carried out in a stationary inter-
Keywords:
nal combustion engine with 20% and 50% blends of methyl and ethyl esters made from soybean oil, palm
NOx emissions
Palm oil
oil and waste frying oil (collected in the University Campus). The analysis of the results with Tukey’s test
Soybean oil compare their means and lead to the conclusion that, when considering the route, ethyl blends have
Waste frying oil lower NOx emissions, and palm ethyl ester blends had the best results. Also when all factors are taken
Methyl ester into account, B20 from soybean methyl ester has the lowest emissions of NOx. We recommend that
Ethyl ester future studies test the effect of antioxidants in NOx emissions, as well as test higher blend ratios.
Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction Biodiesel, as an alternative diesel, is obtained from the transes-

Concerns with the exhaustion of oil resources and environmental
issues linked to the combustion of its products have motivated
research for alternative and cleaner fuels from renewable sources
[1]. Oil products account for 40% of the total energy consumed
worldwide and, among them, diesel fuel is the most used. Its com-
bustion is responsible for emitting air pollutants and greenhouse
gases (GHG), prejudicial to human health and the environment [2,3].
⇑ Corresponding author.
E-mail address: beatrizlgf@pet.coppe.ufrj.br (B. Lagnier Gil Ferreira).
terification of vegetable oils, such as soy, palm, and sunflower, and
from animal fats, such as pork lard, beef tallow and fish oil. The use
of biodiesel is interesting since it is highly biodegradable, has low
toxicity and can replace diesel in many applications such as boilers
and internal combustion engines without major modifications to
the engine or losses in performances. Moreover, in comparison
with regular diesel, biodiesel is capable of reducing hydrocarbons
(HC), carbon monoxide (CO) and particulate matter (PM) emissions
and practically eliminating the emission of sulfates and CO2 (when
its entire life cycle is considered) [1,4].
However, international literature does not agree on the effects
of biodiesel on nitrogen oxides (NOx) emissions, which vary

http://dx.doi.org/10.1016/j.fuel.2016.12.084
0016-2361/Ó 2016 Elsevier Ltd. All rights reserved.
 M.A. Vieira da Silva et al. / Fuel 194 (2017) 144–156
depending on the feedstock, blend percentage, transesterification
route (depending on the alcohol used in the reaction) and engine
load [1,5–9]. On that note, it is important to comprehend how this
factors influence emissions in order to determine which combina-
tion constitutes the best alternative to diesel fuel.
Considering the hypothesis that these factors influence biodie-
sel NOx emissions, this article aims to evaluate how varying feed-
stock, transesterification routes, blend percentage and engine load
affect these emissions. With that purpose, we conducted tests with
biodiesel produced from soybean oil, palm oil and waste frying oil
(collected from restaurants in the university’s campus), using
methyl and ethyl transesterification routes, to generate electrical
energy using stationary internal combustion engines.
It will be possible to see in the bibliographical review section
that, among the papers selected in this research, the majority
focused only in testing methyl esters and very few conducted tests
with ethyl esters. Furthermore, the number of papers that com-
pares both transesterification routes is also limited. In light of this,
by studying more than one route, and comparing them, this paper
contributes with this existing research gap.
Starting from this introduction, the current work is divided into
six sections. The second section provides a bibliographical review
about biodiesel, its feedstock and NOx emissions. The third section
presents the materials and methods used in this study. The
obtained results are described in the fourth section and analyzed
in the following section. Lastly, the sixth section presents the con-
clusions, limitations and suggestions for new studies.
2. Bibliographical review
Biodiesel is a fuel composed of long-chain fatty acid alkyl esters,
from vegetable oils or animal fats. Among the methods for obtain-
ing biodiesel, the transesterification process is commonly used due
to its simplicity and low cost [1,10].
This method promotes a reaction between oils or fats with an
alcohol, forming esters (biodiesel) and glycerol, which is a by-
product that can be burned for heating purposes or used as raw
material in the cosmetics industry. Methanol and ethanol are the
alcohols most used in the process, forming methyl esters or ethyl
esters, respectively. However, methanol is used more frequently
due to its lower costs and physicochemical advantages, since its
shorter chain allows for an easier separation process between
esters and glycerol [1,10].
The oils and fats used to obtain biodiesel can come from differ-
ent feedstocks and can be classified as [1,8,11]:

Edible vegetable oil: sunflower, rapeseed, rice bran, soybean,
coconut, corn, palm, olive, canola, almond, etc.

Non-edible vegetable oil: jatropha, karanja, polanga, tobacco
seed, cotton seed, linseed, mahua, rubber seed, etc.

Waste or recycled oil: waste frying oil and sewage sludge.

Animal fats: chicken fat, pork lard, beef tallow and fish oil.
Table 1 lists the main feedstocks used in the six largest biodiesel
producing countries in the world [1,3]. Brazil, the fourth largest
producer, uses as its feedstock 76.9% of soybean oil, 19.8% of ani-
mal fats and 2.2% of cotton seed oil. The remaining 1.1% refers to
palm oil, peanut oil, oilseed radish oil, sunflower oil, castor oil,
sesame oil, waste frying oil and other oily products [12].
Biodiesel has similar properties to diesel and can be used in
internal combustion engines without major modification to its
characteristics. It can replace diesel entirely or blended to it in
any proportion, with minor losses in performance [1,8,10].
According to Hoekman and Robbins [13], the interest in biodie-
sel arises mainly from its domestic origin and renewable origin,
145
Table 1
Largest biodiesel producers and its main feedstocks. Source: [1,3,11].
Country Feedstock
1 United States Soybeans/waste oil/peanut
2 Germany Rapeseed
3 Argentina Soybeans
4 Brazil Soybeans/cotton oil/animal fat/palm oil
5 Indonesia Palm oil/jatropha/coconut
6 France Rapeseed/sunflower
which potentially decreases GHG when considering the whole life
cycle (including cultivation, production of oil, conversion to biodie-
sel and transportation), since the CO2 produced by the fuels com-
bustion, will be consumed by the crops via photosynthesis. As it
is a renewable energy source, its role in providing energy require-
ments for transportation is expected to increase significantly [1,9].
Compared to diesel, using biodiesel (net or blended with diesel)
lowers the emissions of HC, CO and PM. Moreover, it is highly
biodegradable and non-toxic, has no aromatic compounds and a
lower sulfur content. However, using biodiesel may lead to higher
emissions of NOx, a local pollutant that causes respiratory prob-
lems [1,8,10,13].
NOx emissions are worrisome as they represent a group of reac-
tive gases containing nitric oxide (NO) and nitrogen dioxide (NO2),
which are harmful to human health, contribute to acid rain, caus-
ing damages in structures and increasing the acidity of water
resources [14].
To further comprehend biodiesel’s NOx emissions, we con-
ducted a bibliographical review with 35 papers, published from
2003 to 2017, with experimental results for NOx emissions. Papers
that only contemplated bibliographical reviews were not included.
The review results are shown in Table 2, where it can be seen which
feedstock, blend, load and transesterification routes were used.
Among the gathered papers, only 1 [15] tested biodiesel made
from both ethyl and methyl transesterification routes. 33 papers
tested methyl esters while 1 tested only ethyl esters. This validates
the contribution of this study that carries out experiments with
both routes.
Amidst the papers that studied ethyl esters, the only feedstocks
used in these were mahua oil [29] and rapeseed oil [15]. All found
reductions in NOx, except when testing B75 and B100 of rapeseed
biodiesel [15]. Regarding the papers that tested methyl esters, 75%
reported an increase in NOx emissions.
Among the papers contemplated in this bibliographical review,
71% showed an increase in NOx emissions when using biodiesel,
while 23% presented decreases in emissions. The remaining 2%
had results that increased and decreased emissions when varying
blend percentage. Makareviciene and Janulis [15] tested biodiesel
made from rapeseed oil and found lower emissions for with blends
ratios of 25% and 50% while higher emissions for 75% and 100%.
Serrano et al. [35] tested biodiesel made from a mixture of soybean
and palm oils and identified lower percentages for B20 and higher
emissions for B7.
Considering the feedstocks used in this study we gathered six
[6,7,23,24,37,45] papers that studied palm methyl ester, in which
all of them detected higher emissions, except for Abu-Hamdeh
and Alnefaie [6] and Ng et al. [24]. Abu-Hamdeh and Alnefaie [6]
discovered that at constant engine speed and various torques,
B10, B30 and B50 had lower emissions than diesel; and Ng et al.
[24] identified 5% reductions in NOx emission for B100, under
reduced steady-state emissions test cycle, representative of on-
road driving condition, with constant speed and load values.
As for soybean methyl ester, four papers testing this fuel were
found [14,22,31,41] and only one [31] reported reducing in NOx
emissions by 7% for B100.
146 M.A. Vieira da Silva et al. / Fuel 194 (2017) 144–156

Table 2
Results of the bibliographical review.

Reference Feedstock Blend Route Test variation Emissions Reference

diesel
NOx vs. NOx
variation
[14] Soybean oil B5, B20 and B100 Methyl 2800, 2000, 1800, 1600, 1400, 1200 ; " Conventional
and 1000 rpm
[16] Waste oil B20 Methyl 0/1.3/2.6/3.9/5.2 bmep (bar) " " Conventional
[17] Waste frying palm oil B100 Methyl Full load – " Conventional
Canola oil – "
[5] Rice bran oil B10, B20 and B50 Methyl 0%, 40%, 47%, 55%, 64%, 82% and 97% " ; Mineral (with
15% EGR)
[6] Almond oil B10, B30 and B50 Methyl Torque: 4, 8, 12, 16 and 20 N m " ; Conventional
Palm ;
[18] Waste oil B100 and B20 Methyl 150 N m, Medium - 300 Nm, ; " Conventional
450 N m and High - 600 N m
[19] Jatropha B5, B10, B20, B30 and B100 Methyl 0%, 20%, 40%, 60%, 80% and 100% " " Conventional
[20] Rubber seed oil B10, B20, B50, B75 and B100 Methyl Various loads " " –
[21] Cotton seed oil B10, B20 and B30 Methyl Torque: between 5 and 25 N m " " Conventional
[22] Soybean crude oil B30, B50, B80 and B100 Methyl 15–90% rated engine load " " Conventional
[23] Palm (blended with B5, B20 and B40 Methyl Driving cycles: NEDC, UDC and ADC – " Conventional
coconut oil)
[24] Palm oil B50 and B100 Methyl Phase 1: 0.5–2.5 kW, 1800–2800 rev/ " – –
min
Palm oil B10, B20, B30, B40, B50, B60, B70, B80, Methyl Phase 2: constant speed and load – ; Conventional
B90 and B100
[25] Waste cooking palm B5 Methyl 1500–3500 rpm " " No. 2
oil
[7] Palm oil B20 Methyl 1. 5–37.5 kW at constant speed of " Conventional
2400 rpm
B30 2. Speed varying between 2000 and "
B40 2800 rpm at full load "
[26] Tobacco seed oil B17.5 Methyl 50%, 75% and 100% " " Conventional
3000, 2500, 2200, 2000 and
1500 rpm
[27] Mahua oil B20, B40, B60 and B100 Methyl Variable loads " " Standard
Diesel
[28] Karanja B5, B10, B15, B20, B25, B50 and B100 Methyl 50%, 75%, 90% and 100% " " Standard
Diesel
[29] Mahua oil B100 Ethyl Different BMEP " ; Standard
Diesel
[30] Mahua oil B100 Methyl Between 0 and 4 kW " ; No. 2
[31] Soybean B100 Methyl Variation at start of injection ; Ultra low
sulfur (BP15)
[32] Polanga oil B20, B40, B60, B80 and B100 Methyl 0%, 20%, 40%, 60%, 80%, and 100% " " High speed
[33] Mahua oil B20, B40, B60, B80 and B100 Methyl 25%, 50%, 75%, and 100% " " High speed
[34] Rice bran oil B20 Methyl 0%, 25%, 50%, 75% and 100% " " No. 2
[35] Soybean (84%) and B7 Methyl Different driving cycles – " Conventional
palm (16%) oil B20 Methyl – ;
[36] Jatropha oil B100 Methyl 0–6 kW " " Conventional
[37] Palm oil B20 and B100 Methyl 1000–4000 rpm, for each 500 rpm at " " Conventional
full load
[38] Sunflower oil and B10, B20, B50 Methyl 0.01/0.95/1.90/2.85/3.80 kW " ; Marine
olive oil
[15] Rapeseed oil B25, B50 Methyl – – ; Conventional
B75 and B100 " Conventional
B25, B50 Ethyl – – ; Conventional
B75 and B100 " Conventional
[39] Jatropha oil B5, B10, B20, B50 and B100 Methyl 0.10/0.26/0.51/0.77 MPa " " Conventional
[40] Mahua oil B85, B90, B95 and B100 (with the Methyl 0%, 20%, 40%, 60%, 80% and 100% " " Conventional
additive di-methyl carbonate)
[41] Soybean oil B10, B20, B50 and B100 Methyl 1300, 1500, 2000, 2500, 3000 rpm " " Conventional
[42] Waste cooking B5, B10, B20, B30, B50, B70 and B100 Methyl Different BMEP " ; Conventional
[43] Mango seed B25, B50, B75 and B100 Methyl 0%, 25%, 50%, 75% and 100% " " Conventional
[44] Rapeseed oil B10 and B20 Methyl 1800, 2200, 2600 and 3000 rpm, 75% ; " Conventional
and 100% load
[45] Palm oil B5 Methyl 1000, 1500, 2000, 2500 and ; " Conventional
3000 rpm

Finally, a total of five papers studied waste oil methyl ester
[16–18,25,42] in which all reported increased emissions except
for Attia and Hassaneen [42] who found an average of 6% decrease
for all blends.
According to Abu-Hamdeh and Alnefaie [6] and Mofijur et al.
[9], NOx emissions are positively correlated with the tempera-
ture of combustion, which is directly influenced by the
engine load. As a result, it is believed that increasing the load
leads to higher emissions. 74% of the selected papers presented
higher NOx emissions when the load was increased, while 14%
of them reported lower emissions, the rest didn’t consider this
variation.
 M.A. Vieira da Silva et al. / Fuel 194 (2017) 144–156 147

Koc and Abdullah [14], Leevijit and Prateepchaikul [7], Reza
Miri et al. [44] and Yasin et al. [45] obtained lower emissions while
varying engine speed, while in Alptekin et al. [18] this behavior
was seen for constant engine speed and various torques.
3. Material and methods
due to a forced ventilation system that dissipates generated heat. As
a result, we verified that the total load (resistance + ventilator acti-
vation) for each level was 18 kW, 33 kW, 48 kW, 65 kW and 82 kW,
the latter being over the maximum supported by the generator
(80 kW) and therefore was excluded from the test. On that note,
the tests were performed with levels 18 kW, 33 kW, 48 kW and
65 kW. Fig. 1 illustrates the experimental setup of the test rig.

3.1. Experimental setup

To conduct the experiments, we used an engine-generator set
composed of a diesel internal combustion engine (brand Perkins,
model 1104, with direct injection, 4-cylinder and 4.4 L, power band
of 50–106 kW), coupled with an adapted generator with maximum
power of 100 kV A, 200 V, triphasic, 60 Hz, 1800 rpm. A power fac-
tor of 0.8 was used, and consequently, the maximum power of the
generator was considered as 80 kW (total power  power factor).
To ensure uniform test conditions, and thus allowing the com-
parison between different fuels, it was used a fixed electronical
load bank with the following characteristics: maximum power of
500 kW, 200 V, triphasic, with 9 circuit breakers selecting the resis-
tive loads, each adding or taking 15 kW.
As the maximum active power of the generator-engine set was
80 kW and the loads were multiples of 15 kW, we initially chose
the following nominal loads: 15 kW, 30 kW, 45 kW, 60 kW and
75 kW. However, the experimental set was also connected to a
Power Quality Monitor (brand HOMIS, model 303), that indicated
that the actual load increase was higher than the bank load addition,
3.2. Tested fuels
In this study, we tested biodiesel blends from three different
feedstocks: palm oil, soybean oil and waste frying oil (collected
from restaurants on the university’s campus). The biodiesels were
obtained by a transesterification reaction between the oils and
methanol (forming methyl esters); and ethanol (forming ethyl
esters).
We tested diesel-biodiesel blends with 50% diesel and 50% bio-
diesel (B50), and 80% diesel and 20% biodiesel (B20). B50 was cho-
sen because it is an intermediate situation between the reference
fuel and B100, and B20 was selected to verify the behavior of
NOx emissions between the reference fuel and B50. In this study,
the reference fuel used was conventional diesel with 4% of biodie-
sel (B4), blend legally enforced and sold in Brazilian territory at the
time of the tests.
Table 3 summarizes the notation for the different fuels tested,
and Table 4 shows the physicochemical characteristics of each bio-
diesel (B100).

Fig. 1. Schematic diagram of the test rig.

Table 3
Notation of the tested fuels.

Methyl ester Ethyl ester

20% blend 50% blend 20% blend 50% blend
Soybean oil BSM BSM BSE BSE
20 50 20 50
Palm oil BPM
20 BPM
50 BPE
20 BPE
50
Waste frying oil BFM
20 BFM
50 BFE
20 BFE
50
148 M.A. Vieira da Silva et al. / Fuel 194 (2017) 144–156

Table 4
Properties of the biofuels (B100).

Units Methyl ester Ethyl ester Limits ASTM Method

Soybean Palm Waste Soybean Palm Waste
Specific mass g/ml 0.8819 0.8716 0.8871 0.8764 0.8767 0.8838 0.85 and 0.90 D4052
Kinematic viscosity mm2/s 4.208 4.454 4.912 4.434 4.624 – 3.00 and 6.00 D445
Acidity index mgKOH/g 0.2 0.1 0.3 0.2 0.3 0.4 max. 0.50 D664
Flash point °C 169.0 179.0 178.0 179.0 166.0 177.0 min. 100.0 D93
Sulphated ash Mass% 0.016 0.0027 0.0057 0.0021 0.0004 0.0044 max. 0.20 D874
Total sulfur mg/Kg 9.15 4.16 34.1 9.54 – 30.0 max. 50.0 D5453
Pour point °C 3 11.2 0 24 6 6 max. 19.0 D664
Carbon residue mass% 0.02 0.02 0.05 0.04 0.04 0.05 max. 0.05 D4530
Copper strip corrosion (3 h/50 °C) Degree of corrosivity 1st 1st 1st 1st 1st 1st max. 1st D130

A universal gas analyzer measured the NOx emissions based on hypothesis of equality (H0), and conclude that the means are
a non-dispersive infrared method (PC-MULTIGÁS) with the follow- different.
ing specifications:
4. Results

NOx Measurement Scale: 0–5000 ppm

Uncertainty: 0.6 ppm The results obtained in the experiments will be analyzed for

Lambda: 0–9.99 each load level in the following subsections. Figs. 2–5 represent

AFR: 0–99.99 the box-plots1 of NOx emissions for each type of fuel tested. All tests

Mist eliminator were carried out for a 5% significance level (a).
Tables 5, 7, 9 and 11 present the tests’ statistical results, with
The engine-generator set consumed each fuel for approximately the number of observations (number of measurements for each
2 h, enough time to collect consistent data on the gas emissions, fuel), mean (in ppm), standard deviation, standard error of the
which were measured by the gas analyzer at least four times per mean, maximum and minimum value of the measurements
cycle. and the difference (in percentage) between the blends and the
reference fuel emissions. These tables also show a classification
of the fuels by groups, indicated by letters, which represent
3.3. Method for comparison of means
the similarity of the results between the samples. For groups
with the same letter, the hypothesis of equality cannot be
To compare the emissions of each blend, we carried out com-
rejected hence, statistically, their emissions are considered to
parisons tests between means using Tukey’s test, with the objec-
be equivalent.
tive to test the following hypothesis:
Tables 5, 7, 9 and 11 also present the results of the repeatability
study (r) which indicated that no observation needed to be
H0 : li ¼ lj excluded from the analysis. Tables 6, 8, 10 and 12 show the results
of the p-value obtained by Tukey’s test to compare the means of
H1 : li – lj the emissions of the fuels. Values higher than 0.05 indicate that
the hypothesis of equality cannot be rejected between the com-
for all i – j [46]. Tukey’s test compares all pairwise differences while pared means.
controlling the probability of making one or more Type I errors,
which is defined as rejecting a true null hypothesis, such as making 4.1. 18 kW load
a confident claim about the direction of the difference when there is
either no difference or when the difference is in the opposite direc- For B4, BSE FE FE SM PM
20 , B50 , B20 , B20 and B50 , the hypothesis of equality can-
tion of the claimed difference [47]. not be rejected since they belong to groups with the same letter
To apply the method, we must first calculate the least signifi- and the p-value is higher than 0.05. Consequently, it is possible
cant difference (LSD), as presented in Eq. (1). to state that, in comparison with B4, the blends NOx emissions
did not show significant variations.
s
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

1 1 RMS NOx emissions from BPM FM SM FM SE
20 , B20 , B50 , B50 and B50 , belonging to
LSD ¼ qd;k;a  þ  ; ð1Þ groups C, B and A, were higher than B4. It is not possible to reject
ni nj 2
the hypothesis of equality between BSM FM SE
50 , B50 and B50 since they
belong to a same group.
in which:

qd;k;a : tabulated value which varies according to the degrees of BPE PE
20 and B50 had lower emissions than B4, even though they all
belong to a group with the same letter, they presented marginal
freedom of the residues (d), the number of treatments (k) and
but significant variations. Additionally, the hypothesis of equality
the significance level (a);

RMS: residual mean square; between BPE PE
20 and B50 is not rejected.

ni , nj : number of repetitions of each treatment.

Then, we compare the difference between pairwise means (d)
with the value for LSD. If d > LSD (and thus if p-value > a) we
accept the equality hypothesis (H0), and conclude that the means
are equal. If d 6 LSD (and thus if p-value < a) we reject the
1
A box-plot displays the minimum, the maximum, the lower and upper quartiles
(the 25th percentile and the 75th percentile, respectively), and the median (the 50th
percentile) on a rectangular box aligned either horizontally or vertically. The box
extends from the lower quartile to the upper quartile, and a line is drawn through the
box at the median. Lines extend from the ends of the box to the minimum and
maximum values [46].
 M.A. Vieira da Silva et al. / Fuel 194 (2017) 144–156 149

320

300

280

NOx emissions (ppm)

260

240

220

200

180

160

140
B20FE B20FM B20PE B20PM B20SE B20SM B50FE B50FM B50PE B50PM B50SE B50SM B4

Fig. 2. NOx emissions for the 18 kW power level.

Table 5
Statistical results for the 18 kW power level.

Fuel Mean Observations SDa SEMb rc Max Min Groups Variation to B4

BPM
20
297.00 7 11.33 10.48 45.31 311 284 A 74.0%
BFM
20
229.67 6 6.44 6.76 25.11 237 219 B 34.6%
BSM
50
202.11 9 4.99 3.83 21.44 209 197 C 18.4%
BFM
50
193.00 6 8.81 9.24 34.36 202 184 C 13.1%
BSE 192.17 6 13.38 14.04 52.17 206 179 C 12.6%
50
BPM
50
184.67 3 1.53 3.79 4.28 186 183 CD 8.2%
BSE 175.13 8 1.73 1.44 7.25 178 173 DE 2.6%
20
BFE
50
170.90 10 2.60 1.86 11.45 174 167 DE 0.1%
B4 170.69 16 14.03 7.47 64.53 186 153 DE –
BFE
20
165.83 6 7.25 7.61 28.28 174 158 DE 2.8%
BSM
20
165.50 6 3.83 4.02 14.95 169 162 DE 3.0%
BPE
20
162.50 8 4.54 3.79 19.05 169 157 E 4.8%
BPE
50
160.83 6 13.39 14.05 52.23 173 147 E 5.8%

a
Standard deviation.
b
Standard error of the mean.
c
Repeatability to exclude any observation.

Table 6
Results of p-value for Tukey’s test for the 18 kW power level.

BFE
20 BFM
20 BPE
20 BPM
20 BSE
20 BSM
20
B4 BFE
50 BFM
50 BPE
50 BPM
50 BSE
50
10
BFM
20
1.58  10
BPE
20
1.00 1.58  1010
10
BPM
20
1.58  10 1.58  1010 1.58  1010
BSE 0.790 1.58  1010 0.226 1.58  1010
20
10
BSM
20
1.00 1.58  10 1.00 1.58  1010 0.749
B4 0.996 1.58  1010 0.670 1.58  1010 0.995 0.992
BFE
50
0.997 1.58  1010 0.756 1.58  1010 0.999 0.994 1.00
BFM
50
9.91  105 4.10  108 1.27  106 1.58  1010 0.024 7.68  105 1.21  104 6.17  104
BPE
50
0.999 1.58  1010 1.00 1.58  1010 0.170 1.00 0.541 0.626 1.83  106
BPM
50
0.162 3.84  108 0.027 1.58  1010 0.937 0.144 0.417 0.515 0.984 0.020
BSE 1.86  104 2.00  108 2.66  106 1.58  1010 0.040 1.45  104 2.57  104 0.001 1.00 3.64  106 0.993
50
BSM 2.99  109 9.24  106 1.63  1010 1.58  1010 1.99  106 2.21  109 2.58  1010 4.62  109 0.784 1.81  1010 0.182 0.673
50

The bold numbers are values higher than 0.05 (significance level), which indicate that the equality hypothesis cannot be rejected.
150 M.A. Vieira da Silva et al. / Fuel 194 (2017) 144–156

4.2. 33 kW load BSM

20 was the only one who showed a significant decrease (12.1%) in
NOx emissions for the 33 kW load.
The BPE PE
20 and B50 blends belong to the same group as B4 and have NOx emissions increased for the other blends, which were
a p-value higher than 0.05 when compared to B4, thus the hypoth- divided into two groups, A and B, and the hypothesis of equality
esis of equality between the fuels cannot be rejected. Despite being between them and B4 is rejected. Among these fuels, BPM
20 showed
in the same group as B4 and presenting a p-value higher than 0.05, the highest emission increase (92.6%).

550

500

450
NOx emissions (ppm)

400

350

300

250

200
B20FE B20FM B20PE B20PM B20SE B20SM B50FE B50FM B50PE B50PM B50SE B50SM B4

Fig. 3. NOx emissions for the 33 kW power level.

Table 7
Statistical results for the 33 kW power level.

Fuel Mean Observations SDa SEMb rc Max Min Groups Variation to B4

BPM
20
503.50 6 27.99 29.37 109.17 530 471 A 92.6%
BFM
20
348.29 7 28.81 26.65 115.26 385 323 B 33.2%
BSM
50
339.63 8 27.72 23.18 116.44 372 314 B 29.9%
BSE 329.55 11 43.61 29.30 196.26 398 298 B 26.1%
20
BFE
50
321.90 10 11.75 8.41 51.71 338 307 B 23.1%
BFE
20
315.29 7 21.16 19.57 84.62 340 292 BC 20.6%
BSE 312.38 8 31.17 26.06 130.92 344 282 BC 19.5%
50
BFM
50
311.67 6 20.53 21.54 80.07 332 291 BCD 19.2%
BPM
50
279.14 7 19.13 17.69 76.50 301 259 CDE 6.8%
BPE
20
267.00 6 24.17 25.36 94.25 291 244 DEF 2.1%
BPE
50
266.38 8 15.17 12.69 63.73 286 255 EF 1.9%
B4 261.43 14 7.93 4.58 36.48 274 250 EF –
BSM 229.83 6 3.66 3.84 14.26 234 225 F 12.1%
20

a
Standard deviation.
b
Standard error of the mean.
c
Repeatability limit.

Table 8
Results of p-value for Tukey’s test for the 33 kW power level.

BFE
20 BFM
20 BPE
20 BPM
20 BSE
20 BSM
20
B4 BFE
50 BFM
50 BPE
50 BPM
50 BSE
50

BFM
20
0.362
BPE
20
0.031 2.84  106
BPM
20
4.64  1010 4.64  1010 4.64  1010
BSE 0.991 0.927 1.49  104 4.64  1010
20
BSM
20
7.25  107 4.72  1010 0.298 4.64  1010 6.63  1010
B4 4.56  104 1.58  109 1.00 4.64  1010 4.05  108 0.289
BFE
50
1.00 0.592 2.14  103 4.64  1010 1.00 7.32  109 2.86  106
10 6
BFM
50
1.00 0.266 0.090 4.64  10 0.963 6.20  10 0.004 1.00
BPE
50
0.011 3.06  107 1.00 4.64  1010 1.72  105 0.228 1.00 4.07  104 0.044
BPM
50
0.229 5.35  105 1.00 4.64  1010 0.003 0.024 0.933 0.031 0.450 0.998
BSE
50
1.00 0.196 0.043 4.64  1010 0.948 8.33  107 5.69  104 1.00 1.00 0.016 0.301
BSM 0.770 1.00 2.22  105 4.64  1010 1.00 5.16  1010 1.02  108 0.943 0.643 2.64  106 4.20  104 0.564
50
 M.A. Vieira da Silva et al. / Fuel 194 (2017) 144–156 151

650

600

550

NOx emissions (ppm) 500

450

400

350

300
B20FE B20FM B20PE B20PM B20SE B20SM B50FE B50FM B50PE B50PM B50SE B50SM B4

Fig. 4. NOx emissions for the 49 kW power level.

Table 9
Statistical results for the 49 kW power level.

Fuel Mean Observations SDa SEMb rc Max Min Groups Variation to B4

BSE 630.00 6 9.72 10.20 39.77 642 618.00 A 88.8%

20
BPM
20
585.00 6 17.66 18.54 72.29 605 566.00 B 75.3%
BSM
50
451.83 6 30.33 31.83 124.15 480 408.00 C 35.4%
BFM
20
444.50 6 15.02 15.76 61.46 459 424.00 C 33.2%
BFM
50
366.13 8 2.85 2.38 10.01 369 360.00 D 9.7%
BSE
50
361.43 7 4.72 4.37 17.46 367 354.00 DE 8.3%
BPE
50
360.00 8 6.52 5.45 22.91 367 352.00 DE 7.9%
BPM
50
354.83 6 2.79 2.92 11.41 358 352.00 DEF 6.3%
BFE
50
348.36 11 10.44 7.02 31.57 360 335.00 DEFG 4.4%
BPE
20
345.14 7 1.07 0.99 3.95 347 344.00 DEFG 3.4%
BFE
20
342.50 6 6.60 6.92 26.99 351 333.00 EFG 2.6%
BSM 334.17 6 0.75 0.79 3.08 335 333.00 FG 0.1%
20
B4 333.77 13 16.67 10.08 46.35 350 309.00 G –
a
Standard deviation.
b
Standard error of the mean.
c
Repeatability limit.

Table 10
Results of p-value for Tukey’s test for the 49 kW power level.

BFE
20 BFM
20 BPE
20 BPM
20 BSE
20 BSM
20
B4 BFE
50 BFM
50 BPE
50 BPM
50 BSE
50
10
BFM
20
1.10  10
BPE
20
1.00 1.10  1010
BPM
20
1.10  1010 1.10  1010 1.10  1010
SE 1.10  1010 1.10  1010 1.10  1010 1.32  106
B20
BSM
20
0.994 1.10  1010 0.931 1.10  1010 1.10  1010
10
B4 0.969 1.10  10 0.765 1.10  1010 1.10  1010 1.00
BFE
50
0.999 1.10  1010 1.00 1.10  1010 1.10  1010 0.566 0.199
BFM
50
0.039 1.10  1010 0.078 1.10  1010 1.10  1010 0.001 1.00  105 0.124
BPE
50
0.330 1.10  1010 0.524 1.10  1010 1.10  1010 0.014 0.001 0.726 0.999
BPM
50
0.885 1.10  1010 0.972 1.10  1010 1.10  1010 0.193 0.049 0.998 0.899 1.00
BSE 0.259 1.10  1010 0.428 1.10  1010 1.10  1010 0.011 0.001 0.619 1.00 1.00 0.999
50
BSM 1.10  1010 0.998 1.10  1010 1.10  1010 1.10  1010 1.10  1010 1.10  1010 1.10  1010 1.10  1010 1.10  1010 1.10  1010 1.10  1010
50

The bold numbers are values higher than 0.05 (significance level), which indicate that the equality hypothesis cannot be rejected.

4.3. 49 kW load p-value higher than 0.05, thus the hypothesis of equality is also
not rejected.
The hypothesis of equality cannot be rejected between fuels The other groups increase emissions and the hypothesis of
from group G (BFE PE FE SM
50 , B20 , B20 and B20 ) when compared with the ref-
equality between the blends and B4 is rejected. The fuels with
erence fuel. Despite not belonging to group G, BPM the highest increase were BSE PM
20 and B20 .
50 showed a
152 M.A. Vieira da Silva et al. / Fuel 194 (2017) 144–156

1000

900

NOx emissions (ppm) 800

700

600

500

400
B20FE B20FM B20PE B20PM B20SE B20SM B50FE B50FM B50PE B50PM B50SE B50SM B4

Fig. 5. NOx emissions for the 65 kW power level.

Table 11
Statistical results for the 65 kW power level.

Fuel Mean Observations SDa SEMb rc Max Min Groups Variation to B4

BSE 920.17 6 38.03 39.91 155.66 960 880 A 94.5%

20
BPM
20
756.33 6 86.14 90.39 352.54 828 613 B 59.9%
BFM
20
598.17 6 32.71 34.33 133.89 629 552 C 26.5%
BSM 584.29 7 46.64 43.13 172.54 630 524 C 23.5%
50
BSE
50
572.83 6 23.46 24.62 96.00 595 546 C 21.1%
BFM
50
508.67 6 19.91 20.89 81.47 527 481 D 7.5%
BPE
50
502.17 6 13.48 14.15 55.18 517 489 D 6.2%
BPM
50
491.22 9 16.87 12.97 55.77 511 463 D 3.9%
BPE
20
491.00 6 15.94 16.73 65.23 508 468 D 3.8%
BFE
50
485.67 9 20.04 15.41 66.25 509 448 D 2.7%
BFE
20
483.00 6 19.09 20.03 78.13 501 459 D 2.1%
B4 473.00 15 17.17 9.51 43.73 493 450 D –
BSM 471.83 6 4.67 4.90 19.09 476 466 D 0.2%
20

a
Standard deviation.
b
Standard error of the mean.
c
Repeatability limit.

Table 12
Results of p-value for Tukey’s test for the 65 kW power level.

BFE
20 BFM
20 BPE
20 BPM
20 BSE
20 BSM
20
B4 BFE
50 BFM
50 BPE
50 BPM
50 BSE
50
6
BFM
20
1.10  10
BPE
20
1.00 7.07  106
11
BPM
20
2.50  10 5.50  1011 2.50  1011
BSE 2.50  1011 2.50  1011 2.50  1011 3.24  1011
20
BSM
20
1.00 7.66  108 0.998 2.50  1011 2.50  1011
B4 1.00 2.77  1010 0.993 2.50  1011 2.50  1011 1.00
BFE
50
1.00 1.62  107 1.00 2.50  1011 2.50  1011 1.00 0.999
BFM
50
0.971 0.0003437 0.999 2.50  1011 2.50  1011 0.720 0.501 0.975
BPE
50
0.998 8.59  105 1.00 2.50  1011 2.50  1011 0.905 0.786 0.999 1.00
BPM
50
1.00 6.90  107 1.00 2.50  1011 2.50  1011 0.994 0.976 1.00 0.998 1.00
BSE 3.21  104 0.974 0.002 2.51  1011 2.50  1011 2.85  105 4.12  107 9.60  105 0.038 1.30  102 3.51  104
50
BSM 1.13  105 1.00 7.20  105 2.53  1011 2.50  1011 7.68  107 2.51  109 1.84  106 3.12  103 8.29  104 8.02  106 1.00
50

The bold numbers are values higher than 0.05 (significance level), which indicate that the equality hypothesis cannot be rejected.

4.4. 65 kW load 5. Discussion

For this load the results presented a more evident group divi- 5.1. Feedstocks
sion. For those belonging to group D, the hypothesis of equality is
not rejected, i.e. the variations in NOx emissions are not expressive. The following sections provide an analysis of the results of NOx
Once again, BSE PM
20 and B20 showed the largest variations in com-
emissions for each feedstock, in addition to a pairwise comparison
parison with B4. 94.5% and 59.9%, respectively. between them.
 M.A. Vieira da Silva et al. / Fuel 194 (2017) 144–156 153

Waste frying oil

650
600
550

NOx emissions (ppm)

500 B20FM
450 B20FE
400 B50FE
350 B50FM
B4
300
250
200
150
18kW 33kW 49kW 65kW
Load
Fig. 6. NOx emissions (ppm) for biodiesel from waste frying oil.

Palm oil
900

800

700
NOx emissions (ppm)

B20PM
600 B20PE
B50PE
500
B50PM
400 B4
300

200

100
18kW 33kW 49kW 65kW
Load
Fig. 7. NOx emissions (ppm) for biodiesel from palm oil.

5.1.1. Waste frying oil ester presented the highest emissions among the tested fuels,
Fig. 6 shows the results of NOx emissions versus load variation being the only one for which the hypothesis of equality with B4
for biodiesel from waste frying oil. By analyzing the figure along is rejected for all loads, in addition to rejecting the hypothesis for
with Tables 5–12, we verify that the BFM 20 is the only fuel to reject
all other fuels.
the hypothesis of equality with B4 under all test loads, with higher On the other hand, the blend with 20% of ethyl ester did not
emissions than all other fuels tested, except for the 33 kW load. For reject the hypothesis of equality with B4. Moreover, BPE
20 did not
this load, the hypothesis of equality between all blends is not reject the hypothesis with BPE PM
50 for all test loads and with B50 in
rejected, but for them and B4 is. the 33, 49 and 65 kW loads.
The ethyl blends presented results for which the hypothesis of The 50% blends showed similar behavior, in which the hypoth-
equality with B4 is not rejected for the 18, 49 and 65 kW loads. esis of equality with B4 was only rejected for the 49 kW load,
As for BFM
50 , the hypothesis of equality is not rejected except for where a slight increase was observed for BPE PM
50 (7.9%) and B50 (6.3%).
the 65 kW load. For the other loads, we observed an increase in
NOx emissions.
5.1.3. Soybean oil
Among the fuels, the hypothesis of equality is not rejected in
Fig. 8 shows the results of NOx emissions versus load variation
the following cases: (a) between ethyl blends for all loads; (b)
for biodiesel from soybean oil. By analyzing it along with Tables 5–
between blends with 50% of biodiesel for the 33, 49 and 65 kW
12 we verify that, differently from the other studied feedstocks, the
loads; (c) between BFM FE
50 and B20 for the 65 kW load. 20% methyl ester blend is the only one for which the hypothesis of
equality with B4 is not rejected for all loads. In other words, the
5.1.2. Palm oil emissions from both fuels are very similar.
Fig. 7 displays NOx emissions results versus load variation For the 18 kW load, all tested fuels had similar emissions, and it
for palm oil biodiesel. By analyzing the figure along with is not possible to reject the hypothesis of equality between BSM SE
20 , B20
Tables 5–12 we verify once again that the blend with 20% methyl and B4, neither between blends BSM SE
50 and B50 .
154 M.A. Vieira da Silva et al. / Fuel 194 (2017) 144–156

Soybean oil
950

850

NOx emissions (ppm)

750 B20SM
650 B20SE
B50SE
550
B50SM
450 B4
350

250

150
18kW 33kW 49kW 65kW
Load

Fig. 8. NOx emissions (ppm) for fuels made from soybean oil.

Table 13
Comparison soybean  palm oil.

Route Blend % 18 kW 33 kW 49 kW 65 kW
M 20 ; ; ; ;
50 NR " " "
E 20 NR " " "
50 " " NR "

Note: M: Methyl; E: Ethyl; NR: The hypothesis of equality is not rejected.

Table 14
Comparison soybean  waste frying oil.

Route Blend % 18 kW 33 kW 49 kW 65 kW
M 20 ; ; ; ;
50 NR NR " "
E 20 NR NR " "
50 " NR NR "

Note: M: Methyl; E: Ethyl; NR: The hypothesis of equality is not rejected.

For the 33 kW load, the hypothesis of equality between BSE SE

20 . B50
and BSM
with B4 was rejected, and NOx emissions increased for
50
these three blends. Moreover, the hypothesis of equality was not
rejected between these same blends.
Next, we highlight the larger variation as function of load of the
BSE
20 blend, which, starting from the 49 kW load, has the hypothesis
of equality rejected with all other soybean oil fuels. Moreover, it
shows a considerable increase in NOx emissions, being the blend
with the highest emissions among all tested fuels for the 49 and
65 kW loads.
5.1.4. Pairwise comparisons
Table 13 shows the comparison between emissions from soy-
bean biodiesel and palm biodiesel, while Table 14 compares NOx
emissions from soybean biodiesel with waste frying oil biodiesel.
The NOx emissions from soybean biodiesel were smaller than
palm oil biodiesel and waste frying oil biodiesel only for the blend
with 20% methyl ester, while the contrary was observed for the
ethyl blend, in which BSE PE FE
20 had higher emissions than B20 and B20 .
For all other blends, soybean biodiesel had higher emissions
than biodiesel blends from palm oil and waste frying oil, except
for the following cases in which it was not possible to verify the
hypothesis of equality: (a) 50% methyl ester, for the 18 kW load;
(b) 20% ethyl ester, for the 18 kW load; (c) BFM FE FE
50 , B20 and B50 , for
the 33 kW load; (d) 50% ethyl ester, for the 49 kW load.
Table 15 compares biodiesel from palm oil with those made
from waste frying oil. Among the methyl blends, B20 from palm
oil had higher emissions than B20 from waste frying oil. For the
50% blends, the hypothesis of equality was not rejected for all
loads, except for the 18 kW load, where the emissions from palm
oil biodiesel decreased.
As for the ethyl blends, the hypothesis of equality was not
rejected for B20 and B50 for all loads, except for load 33 kW, in
which NOx emissions from palm oil biodiesel decreased in com-
parison with waste frying oil biodiesel.
5.2. Transesterification route
Table 16 shows the results of the comparison between emis-
sions from methyl and ethyl esters.
The 20% methyl blends from waste frying oil and palm oil had
higher emissions than ethyl blends for all test loads. On the other
hand, BSM SE
20 had lower emissions than B20 (except for the 18 kW load,
for which the hypothesis of equality is not rejected).
 M.A. Vieira da Silva et al. / Fuel 194 (2017) 144–156 155

Table 15
Comparison palm oil  waste frying oil.

Route Blend % 18 kW 33 kW 49 kW 65 kW
M 20 " " " "
50 ; NR NR NR
E 20 NR ; NR NR
50 NR ; NR NR

Note: M: Methyl; E: Ethyl; NR: The hypothesis of equality is not rejected.

Table 16
Overview of the comparison between methyl and ethyl esters.

18 kW 33 kW 49 kW 65 kW
F 20 " " " "
50 " NR NR NR
P 20 " " " "
50 NR NR NR NR
S 20 NR ; ; ;
50 NR NR " NR

NR: The hypothesis of equality is not rejected; F: Waste frying oil; P: Palm oil; S: Soybean oil.

For B50, it was not possible to verify any differences, except for
palm biodiesel for the 18 kW load, and for soybean biodiesel for
the 49 kW load, in which the use of a methyl blend caused higher
emissions than the ethyl blend.
6. Conclusion
This article achieved its goal of evaluating the effects of varying
feedstocks, transesterification routes, blend percentages and
engine load in NOx emissions, combining these factors and deter-
mining emissions for each case. As a result, it was possible to con-
clude that the variation of these factors affects the emissions of the
pollutant.
Firstly, we highlight that NOx emissions increase as the test
load becomes higher, as observed in the bibliographical review
and in the performed tests. Even though some papers
[7,14,18,44,45] obtained lower emissions while varying tests con-
ditions such as engine speed or torques, all papers that varied elec-
trical loads obtained higher emissions as well.
BSM
20 is amongst the blends with best performance, having either
lower emissions than B4 or maintaining the same emission level as
B4, which could also be seen for soy blends in Armas et al. [31] as
opposed to Koc and Abdullah [14], Özener et al. [41] and Qi et al.
[22].
As in Saravanan et al. [16], Ozsezen et al. [17], Alptekin et al.
[18] and Kalam et al. [25] all waste frying blends showed increases
in NOx emissions or could not be differentiated with diesel.
The blends with 20% and 50% of palm ethyl ester were between
the best results at all loads, having either lower or approximately
the same emissions than diesel. In contrast, the palm methyl esters
blends did not show the same satisfactory results. As BPE
20 had the
second lowest level of NOx emissions, BPM 20 had the highest emis-
sions of all fuels, increasing between 60 and 93% when comparing
to the reference fuel. Leevijit and Prateepchaikul [7] conducted
similar tests as this study (varying load) and also found BPM
20 to
increase when compared to diesel. On the other hand, BPM 50 had
either marginals increases (3–8%) or couldn’t be differentiated
with B4.
The fact that palm blends occupy both extremes, highlights the
influence of the transesterification route in the behavior of the
blend. Few papers from literature compare this influence, as most
of them only study methyl biodiesel. As a result, we single out the
innovative aspect of this work, which allowed us to conclude that,
in general, ethyl blends have lower emissions than methyl blends.
As a recommendation for future studies, further tests can be
made using antioxidants, which have the potential to reduce NOx
emissions [48]. We also recommend testing blends with more than
50% biodiesel until B100, for the purpose of studying emission
behavior when biodiesel percentage in the blend increases. And
finally, compare the blends NOx emissions with CO2, PM, HC and
fuel consumption.
References
[1] Atabani AE, Silitonga AS, Badruddin IA, Mahlia TMI, Masjuki HH, Mekhilef S. A
comprehensive review on biodiesel as an alternative energy resource and its
characteristics. Renew Sustain Energy Rev 2012;16:2070–93. http://dx.doi.
org/10.1016/j.rser.2012.01.003.
[2] International Energy Agency. Key World Energy Statistics; 2015.
[3] U.S. Energy Information Administration. International Energy Statistics; 2012.
[4] Dai Y-M, Wu J-S, Chen C-C, Chen K-T. Evaluating the optimum operating
parameters on transesterification reaction for biodiesel production over a
LiAlO2 catalyst. Chem Eng J 2015;280:370–6. http://dx.doi.org/10.1016/j.
cej.2015.06.045.
[5] Agarwal D, Sinha S, Agarwal AK. Experimental investigation of control of NOx
emissions in biodiesel-fueled compression ignition engine. Renew Energy
2006;31:2356–69. http://dx.doi.org/10.1016/j.renene.2005.12.003.
[6] Abu-Hamdeh NH, Alnefaie KA. A comparative study of almond and palm oils as
two bio-diesel fuels for diesel engine in terms of emissions and performance.
Fuel 2015;150:318–24. http://dx.doi.org/10.1016/j.fuel.2015.02.040.
[7] Leevijit T, Prateepchaikul G. Comparative performance and emissions of IDI-
turbo automobile diesel engine operated using degummed, deacidified mixed
crude palm oil–diesel blends. Fuel 2011;90:1487–91. http://dx.doi.org/
10.1016/j.fuel.2010.10.013.
[8] Mofijur M, Atabani AE, Masjuki HH, Kalam MA, Masum BM. A study on the
effects of promising edible and non-edible biodiesel feedstocks on engine
performance and emissions production: a comparative evaluation. Renew
Sustain Energy Rev 2013;23:391–404. http://dx.doi.org/10.1016/j.
rser.2013.03.009.
[9] Xue J, Grift TE, Hansen AC. Effect of biodiesel on engine performances and
emissions. Renew Sustain Energy Rev 2011;15:1098–116. http://dx.doi.org/
10.1016/j.rser.2010.11.016.
[10] D́Agosto MDA, Vieira da Silva MA, de Oliveira CM, Franca LS, da Costa Marques
LG, Soares Murta AL, et al. Evaluating the potential of the use of biodiesel for
power generation in Brazil. Renew Sustain Energy Rev 2015;43:807–17.
http://dx.doi.org/10.1016/j.rser.2014.11.05.
[11] Kumar N, Varun, Chauhan SR. Performance and emission characteristics of
biodiesel from different origins: a review. Renew Sustain Energy Rev
2013;21:633–58. http://dx.doi.org/10.1016/j.rser.2013.01.00.
[12] Agência Nacional de Petróleo. Feedstock used for biodiesel production (B100)
in Brazil; 2014.
156 M.A. Vieira da Silva et al. / Fuel 194 (2017) 144–156
[13] Hoekman SK, Robbins C. Review of the effects of biodiesel on NOx emissions.
Fuel Process Technol 2012;96:237–49. http://dx.doi.org/10.1016/
j.fuproc.2011.12.036.
[14] Koc AB, Abdullah M. Performance and NOx emissions of a diesel engine fueled
with biodiesel-diesel-water nanoemulsions. Fuel Process Technol
2013;109:70–7. http://dx.doi.org/10.1016/j.fuproc.2012.09.039.
[15] Makareviciene V, Janulis P. Environmental effect of rapeseed oil ethyl ester.
Renew Energy 2003;28:2395–403. http://dx.doi.org/10.1016/S0960-1481(03)
00142-3.
[16] Saravanan S, Nagarajan G, Lakshmi Narayana Rao G, Sampath S. Theoretical
and experimental investigation on effect of injection timing on NOx emission
of biodiesel blend. Energy 2014;66:216–21. http://dx.doi.org/10.1016/j.
energy.2014.01.003.
[17] Ozsezen AN, Canakci M, Turkcan A, Sayin C. Performance and combustion
characteristics of a DI diesel engine fueled with waste palm oil and canola oil
methyl esters. Fuel 2009;88:629–36. http://dx.doi.org/10.1016/
j.fuel.2008.09.023.
[18] Alptekin E, Canakci M, Ozsezen AN, Turkcan A, Sanli H. Using waste animal fat
based biodiesels–bioethanol–diesel fuel blends in a DI diesel engine. Fuel
2015;157:245–54. http://dx.doi.org/10.1016/j.fuel.2015.04.067.
[19] Chauhan BS, Kumar N, Cho HM. A study on the performance and emission of a
diesel engine fueled with Jatropha biodiesel oil and its blends. Energy
2012;37:616–22. http://dx.doi.org/10.1016/j.energy.2011.10.043.
[20] Ramadhas AS, Muraleedharan C, Jayaraj S. Performance and emission
evaluation of a diesel engine fueled with methyl esters of rubber seed oil.
Renew Energy 2005;30:1789–800. http://dx.doi.org/10.1016/j.
renene.2005.01.009.
[21] Nabi MN, Rahman MM, Akhter MS. Biodiesel from cotton seed oil and its effect
on engine performance and exhaust emissions. Appl Therm Eng
2009;29:2265–70. http://dx.doi.org/10.1016/j.applthermaleng.2008.11.009.
[22] Qi DH, Chen H, Geng LM, Bian YZ. Experimental studies on the combustion
characteristics and performance of a direct injection engine fueled with
biodiesel/diesel blends. Energy Convers Manage 2010;51:2985–92. http://dx.
doi.org/10.1016/j.enconman.2010.06.042.
[23] Karavalakis G, Alvanou F, Stournas S, Bakeas E. Regulated and unregulated
emissions of a light duty vehicle operated on diesel/palm-based methyl ester
blends over NEDC and a non-legislated driving cycle. Fuel 2009;88:1078–85.
http://dx.doi.org/10.1016/j.fuel.2008.11.003.
[24] Ng J-H, Ng HK, Gan S. Characterisation of engine-out responses from a light-
duty diesel engine fuelled with palm methyl ester (PME). Appl Energy
2012;90:58–67. http://dx.doi.org/10.1016/j.apenergy.2011.01.028.
[25] Kalam MA, Masjuki HH, Jayed MH, Liaquat AM. Emission and performance
characteristics of an indirect ignition diesel engine fuelled with waste cooking
oil. Energy 2011;36:397–402. http://dx.doi.org/10.1016/j.energy.2010.10.02.
[26] Usta N. An experimental study on performance and exhaust emissions of a
diesel engine fuelled with tobacco seed oil methyl ester. Energy Convers
Manage 2005;46:2373–86. http://dx.doi.org/10.1016/j.
enconman.2004.12.002.
[27] Godiganur S, Suryanarayana Murthy CH, Reddy RP. 6BTA 5.9 G2-1 Cummins
engine performance and emission tests using methyl ester mahua (Madhuca
indica) oil/diesel blends. Renew Energy 2009;34:2172–7. http://dx.doi.org/
10.1016/j.renene.2008.12.035.
[28] Lahane S, Subramanian KA. Effect of different percentages of biodiesel–diesel
blends on injection, spray, combustion, performance, and emission
characteristics of a diesel engine. Fuel 2015;139:537–45. http://dx.doi.org/
10.1016/j.fuel.2014.09.03.
[29] Puhan S, Vedaraman N, Sankaranarayanan G, Ram BVB. Performance and
emission study of Mahua oil (madhuca indica oil) ethyl ester in a 4-stroke
natural aspirated direct injection diesel engine. Renew Energy
2005;30:1269–78. http://dx.doi.org/10.1016/j.renene.2004.09.010.
[30] Puhan S, Vedaraman N, Ram BVB, Sankarnarayanan G, Jeychandran K. Mahua
oil (Madhuca Indica seed oil) methyl ester as biodiesel-preparation and
emission characteristics. Biomass Bioenergy 2005;28:87–93. http://dx.doi.org/
10.1016/j.biombioe.2004.06.002.
[31] Armas O, Yehliu K, Boehman AL. Effect of alternative fuels on exhaust
emissions during diesel engine operation with matched combustion phasing.
Fuel 2010;89:438–56. http://dx.doi.org/10.1016/j.fuel.2009.09.022.
[32] Sahoo PKK, Das LMM, Babu MKGKG, Naik SNN. Biodiesel development from
high acid value polanga seed oil and performance evaluation in a CI engine.
Fuel 2007;86:448–54. http://dx.doi.org/10.1016/j.fuel.2006.07.02.
[33] Raheman H, Ghadge SV. Performance of compression ignition engine with
mahua (Madhuca indica) biodiesel. Fuel 2007;86:2568–73. http://dx.doi.org/
10.1016/j.fuel.2007.02.019.
[34] Saravanan S, Nagarajan G, Lakshmi Narayana Rao G, Sampath S. Combustion
characteristics of a stationary diesel engine fuelled with a blend of crude rice
bran oil methyl ester and diesel. Energy 2010;35:94–100. http://dx.doi.org/
10.1016/j.energy.2009.08.029.
[35] Serrano L, Lopes M, Pires N, Ribeiro I, Cascão P, Tarelho L, et al. Evaluation on
effects of using low biodiesel blends in a EURO 5 passenger vehicle equipped
with a common-rail diesel engine. Appl Energy 2015;146:230–8. http://dx.doi.
org/10.1016/j.apenergy.2015.01.063.
[36] Paul G, Datta A, Mandal BK. An experimental and numerical investigation of
the performance, combustion and emission characteristics of a diesel engine
fueled with jatropha biodiesel. Energy Proc 2014;54:455–67. http://dx.doi.org/
10.1016/j.egypro.2014.07.288.
[37] Rizwanul Fattah IM, Masjuki HH, Kalam MA, Mofijur M, Abedin MJ. Effect of
antioxidant on the performance and emission characteristics of a diesel engine
fueled with palm biodiesel blends. Energy Convers Manage 2014;79:265–72.
http://dx.doi.org/10.1016/j.enconman.2013.12.02.
[38] Kalligeros S, Zannikos F, Stournas S, Lois E, Anastopoulos G, Teas C, et al. An
investigation of using biodiesel/marine diesel blends on the performance of a
stationary diesel engine. Biomass Bioenergy 2003;24:141–9. http://dx.doi.org/
10.1016/S0961-9534(02)00092-2.
[39] Tan P, Hu Z, Lou D, Li Z. Exhaust emissions from a light-duty diesel engine with
Jatropha biodiesel fuel. Energy 2012;39:356–62. http://dx.doi.org/10.1016/j.
energy.2012.01.002.
[40] Nayak SK, Pattanaik BP. Experimental investigation on performance and
emission characteristics of a diesel engine fuelled with mahua biodiesel using
additive. Energy Proc 2014;54:569–79. http://dx.doi.org/10.1016/
j.egypro.2014.07.298.
[41] Özener O, Yüksek L, Ergenç AT, Özkan M. Effects of soybean biodiesel on a DI
diesel engine performance, emission and combustion characteristics
2014;115:875–83. http://dx.doi.org/10.1016/j.fuel.2012.10.081.
[42] Attia AMA, Hassaneen AE. Influence of diesel fuel blended with biodiesel
produced from waste cooking oil on diesel engine performance. Fuel
2016;167:316–28. http://dx.doi.org/10.1016/j.fuel.2015.11.064.
[43] Vijayaraj K, Sathiyagnanam AP. Experimental investigation of a diesel engine
with methyl ester of mango seed oil and diesel blends. Alexandria Eng J
2016;55:215–21. http://dx.doi.org/10.1016/j.aej.2015.12.001.
[44] Reza Miri SM, Mousavi Seyedi SR, Ghobadian B. Effects of biodiesel fuel
synthesized from non-edible rapeseed oil on performance and emission
variables of diesel engines. J Clean Prod 2017;142:3798–808. http://dx.doi.org/
10.1016/j.jclepro.2016.10.082.
[45] Yasin MHM, Paruka P, Mamat R, Yusop AF, Najafi G, Alias A. Effect of low
proportion palm biodiesel blend on performance, combustion and emission
characteristics of a diesel engine. Energy Proc 2015;75:92–8. http://dx.doi.org/
10.1016/j.egypro.2015.07.145.
[46] Montgomery DC. Design and analysis of experiments. 8th ed. John Wiley &
Sons, Inc.; 2013.
[47] Salkind NJ. Encyclopedia of research design. SAGE Publications Inc.; 2010.
[48] Velmurugan K, Sathiyagnanam AP. Impact of antioxidants on NOx emissions
from a mango seed biodiesel powered DI diesel engine. Alexandria Eng J
2016;55:715–22. http://dx.doi.org/10.1016/j.aej.2015.10.004.