Fuel 164 (2016) 329–338

Contents lists available at ScienceDirect

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

Biokerosene and green diesel from macauba oils via catalytic

deoxygenation over Pd/C
Larissa Noemí Silva, Isabel C.P. Fortes, Fabiana P. de Sousa, Vânya M.D. Pasa ⇑
Chemistry Department, Universidade Federal de Minas Gerais, Av. Antônio Carlos 6627, CEP: 31270-901 Belo Horizonte, MG, Brazil

h i g h l i g h t s g r a p h i c a l a b s t r a c t

Macauba Pulp and kernel oils were

deoxygenated to produce drop in
biofuels on Pd/C.

FTIR was used as an efficient assay to
Crude almond Crude pulp
quantify the synthesized macauba oil macauba oil
biohydrocarbons.

Higher hydrocarbon percentages Hydrolysis
were obtained with the previously
hydrolysed oils. Catalyst
Deoxygenaon Pd/C

Higher H2 pressure favored
deoxygenation of pulp oil but not Centrifugaon
kernel derivatives.

The produced hydrocarbons can be
used for green diesel or biojet fuel Biogasoline Biojet fuel Green diesel
formulation.

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

Article history: Macauba is a palm tree with good oil productivity (6.2 ton/ha), is native to Central and South America and
Received 8 July 2015 is not exploited for food purposes. Its oils have been noted as an important alternative for the production
Received in revised form 26 September of biofuels, especially for aviation. This study investigated deoxygenation catalysed by 5% w/w of palla-
2015
dium on charcoal (Pd/C) reduced in situ using crude and previously hydrolysed macauba pulp and almond
Accepted 29 September 2015
Available online 9 October 2015
oils with different compositions and acidity values. The effect of the fatty composition and nature of the
feedstock, reaction pressure, atmosphere, presence of stirring and use of the Pd/C catalyst were studied.
The results indicated high selectivity, with a predominance of saturated linear hydrocarbons that corre-
Keywords:
Deoxygenation
spond to green diesel, followed by biojet fuel hydrocarbons. Oxygen removal was favoured for free fatty
Macauba oil acids with long carbon chains, for which decarbonylation and/or decarboxylation predominates. The
Biohydrocarbons highest content of hydrocarbons (85% w/w) was obtained in the reaction of hydrolysed macauba almond
Biofuel oil at 10 bar of H2, 300 °C, 5 h of reaction and stirring at 700 rpm. This green product has potential appli-
Pd/C cations as a drop-in substitute for fossil fuels.
Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction portation sector, which is the second-fastest-growing sector in

The increase in global energy demand has drawn the attention
of companies, researchers and politicians to the search for
alternative energy sources to replace petroleum because of the lat-
ter’s scarcity, price instability and pollutant emission. The trans-
⇑ Corresponding author. Tel.: +55 31 34097577; fax: +55 31 34096650.
E-mail address: vanya@ufmg.br (V.M.D. Pasa).
terms of energy consumption [1], has stood out with its intense
increase in emissions between 1990 and 2011, especially in the
road (52% increase in the period), sea and air (a total increase of
80%) transportation sectors [2]. Large investments are being
applied to obtain biofuels from plant oils. Because of the high
productivity in terms of oil per hectare planted, palm trees are con-
sidered to be the most promising and viable feedstock for this
sector [3].

http://dx.doi.org/10.1016/j.fuel.2015.09.081
0016-2361/Ó 2015 Elsevier Ltd. All rights reserved.
330 L.N. Silva et al. / Fuel 164 (2016) 329–338
Biodiesel and bioethanol are already well-established commer-
cial biofuels. However, because of their oxygenated structure,
which is different from that of petroleum (which is basically
formed by hydrocarbons), engine adaptations are necessary or
the use of these fuels has been restricted to small proportions
blended with petroleum derivatives. In the case of biojet fuels
(BioJF), only hydrocarbon-based synthetic fuels will be accepted
by the market because such fuels guarantee high heats of combus-
tion and compatibility with turbines and polymeric conexions and
already-existing distribution systems. Additionally, the specifica-
tions for jet fuel in terms of cold performance are extremely rigid:
a maximum freezing point of 47 °C has been established [4,5]
because of the extreme conditions under which jet fuel is used.
Thus, the presence of oxygenated compounds in the composition
of this fuel can make its product specification impracticable, espe-
cially for mixtures with high proportions of biofuel. To date, the
content of biohydrocarbons used has been limited to a maximum
of 50% v/v, regardless of whether they are produced by the
Fisher–Tropsch (FT) route or triglycerides deoxygenation (HEFA).
However, by fermentation of sugars, a maximum of 10% v/v of bio-
hydrocarbons (SIP) are permitted [6,7].
Hydrocarbon-based fuels, i.e., those with similar compositions
to that of petroleum, are considered drop-in fuels because they
have the advantage of not requiring extra infrastructure for their
storage and distribution and can be used pure or in blends with
fossil derivatives at any ratio, with no alterations to engines and
their feeding circuits. Because of their easier inclusion in the
market, hydrocarbon-based fuels have stimulated studies to
develop routes and catalysts for the conversion of fatty materials
into hydrocarbons. The main deoxygenation routes include dehy-
dration, which eliminates H2O molecules and forms hydrocarbons
with the same number of carbon atoms as the original fatty
acid - Reaction (1); decarboxylation, which eliminates CO2 - Reac-
tion (2); and decarbonylation, which eliminates CO - Reaction (3).
During decarboxylation and decarbonylation, the hydrocarbons
that are formed have one less carbon atom than the precursor fatty
acid [8].
R-COOH þ 3 H2 ! RCH3 þ 2 H2 O ð1Þ
R-COOH ! RH þ CO2 ð2Þ
R-COOH þ H2 ! RH þ CO þ H2 O ð3Þ
Studies have investigated many catalysts with the purpose of
obtaining high conversions of oils, monoesters and/or free fatty
acids to hydrocarbons, including NiMo/Y-Al2O3 [8], Ni/SiO2Al2O3
[9], Al2O3 impregnated with H3PO4 and without H3PO4 [10], Nb2O5
and NbOPO4 [11]. Palladium and platinum-based catalysts have
been widely used [12–17] and generally yield satisfactory
conversions. However, the tests have frequently been performed
with model compounds. Consequently, the results may differ from
those obtained with natural feedstocks [18–20], which have higher
complexity but would be used in actual industrial applications.
Among the feedstocks, palm trees have high productivity and have
been suggested for use in biofuel production. Palm trees that are
not used for food production, such as the macauba [21], which is
the focus of the present study, are especially promising.
Macauba is a native palm tree of tropical forest and savannahs
that is present in tropical America from the south of Mexico to
Argentina [22]. It stands out among the palm trees because of its
high resistance [23], its high oil production (approximately
6.2 ton of oil/ha) and the fact that the whole plant can be used,
such as for the production of charcoal, animal nutritional flours,
bioresins and biofuels [24]. Additionally, there is no food market
established for macauba oils. Moreover, the palm can be planted
in degraded pastures, i.e., lands that do not have potential use for
food crops, and its planting has the potential to recover these lands
[23]. The macauba palm can be intercropped with livestock pasture
and other plant species, thereby making its large-scale production
more sustainable, as required by the bioenergy sector. Oils from
this palm tree are little studied or explored in the literature, and
no studies regarding their oils deoxygenation for the production
of green diesel and other drop-in biofuels were found in a litera-
ture review. The objective of this study is to investigate the cat-
alytic activity of palladium on charcoal (Pd/C) under different
conditions for deoxygenation of macauba pulp and almond oils
to obtain drop-in biofuels with potential applications for partial
substitution of diesel and jet fuel.
2. Materials and methods
2.1. Materials
The crude macauba pulp and almond oils, that were used as
feedstock in this study were obtained from the Vegetal Extrac-
tivism Cooperative of Monte Claros, Minas Gerais, Brazil and were
named CPO (crude pulp oil)and CAO (crude almond oil), respec-
tively. Fatty acids derived from these oils were also used. They
were obtained by alkaline hydrolysis of the crude oils, which pro-
duced PFA (pulp fatty acid) from the pulp and AFA (almond fatty
acid) from the macauba almond.
The following reagents, which were analytical degree and did
not undergo any type of purification prior to their use, were used:
palladium supported on activated charcoal, Pd/C (Sigma–AldrichÒ,
5% by weight); dodecane (Sigma–AldrichÒ, min. 99%); Linde hydro-
gen, helium, and nitrogen gases (min. 99.999%); N-methyl-N-
(trimethylsilyl) trifluoroacetamide (Fluka, min. 98.5%); deuterated
chloroform (D, 99.8% + 1% v/v TMS + Silver foil–Cambridge Isotope
Laboratories, Inc.); potassium biphthalate (Vetec, min. 99.5%);
ethyl ether (Vetec, min. 99.0%); hydrochloric acid (Vetec, min.
37%); sodium chloride (Cinética, min. 99%); potassium hydroxide
(Vetec, min. 97%); and ethyl acetate (Reagen Quimibrás Indústrias
Químicas S.A).
To create the analytical curves for quantification of the hydro-
carbon content by infrared spectroscopy, a commercial diesel oil
sample with 10 ppm of sulphur (S10) free of biodiesel that was
provided by Petrobras S/A was used.
2.2. Characterisation of the Pd/C catalyst
The 5% Pd/C catalyst was characterised by performing X-ray
diffraction on Pd/C powder. The analyses were carried out with a
Shimadzu XRD-6000 X-ray diffractometer using a graphite crystal
as a monochromator to select the Cu Ka1 radiation with wave-
length of 1.5406 Å and rate of 0.02° s1. The identification of the
peaks was performed by comparing the diffractogram obtained
from the database of the Crystallographica Search-Match version
2.1.1.1 software package with published records.
The surface area of the Pd/C catalyst was measured by the BET
method (Brunauer, Emmett and Teller) and the pore diameter was
calculated by the BJH method (Barret, Joynerand Halenda). For this
purpose the isotherm of N2 adsorption/desorption was obtained
using a Autosorb iQ2 Quantachrome equipment. The sample was
heated at 200 °C for 5 h and N2 was used as adsorbate.
2.3. Characterisation of the crude macauba oils
Crude macauba pulp and almond oils were used in the reactions
with Pd/C studied here. Characterisations of the fatty composition
and physico-chemical parameters were performed according to the
 L.N. Silva et al. / Fuel 164 (2016) 329–338
procedure below, with the purpose of better understanding the
characteristics and potential of this material.
2.3.1. Fatty acid profile
The analyses were conducted using a gas chromatographer with
a flame ionisation detector (GC-FID). For injection of the samples,
hydrolysis was performed, followed by methylation of the oils.
For this purpose, approximately 10 mg of oil was dissolved in
100 lL of 95% ethanol/5% 1 mol/L potassium hydroxide solution
in a cryogenic tube. The sample was mixed on vortex for 10 s. Next,
it was heated for 60 min at 90 °C in a thermostatic bath [adapted
from 25]. After cooling, 400 lL of 20% hydrochloric acid and
600 lL of ethyl acetate were added. After centrifuging for 10 s
and resting for 5 min, a 300 lL aliquot of the organic phase was
removed, placed in a cryogenic tubeand dried by evaporation, thus
obtaining the free fatty acids [adapted from 26]. The free fatty acids
were methylated with 100 lL of BF3/14% methanol and heated for
10 min in a water bath at 80 °C. Next, they were diluted with
400 lL of methanol and analysed in the GC-2010 Shimadzu GC-
FID. A 60 m  0.25 mm SP2340 column with a film thickness of
0.20 lm was used. The selected temperature gradient was main-
taining 140 °C for 5 min, heating at 4 °C/min to 240 °C, and remain-
ing at this temperature for 9 min. The temperatures of the injector
and detector were 260 °C and 300 °C, respectively. Helium was
used as the carrier gas, with a linear velocity of 78 cm/s. The vol-
ume of injection was 1 lL, and the split was 1/50. The identifica-
tion of the peaks was performed by comparison with the
SUPELCO37 methylated fatty acid standards.
2.3.2. Acidity value
The acidity value of the macauba pulp and almond oils were
determined via acid-base volumetric titration using a 0.1 mol/L
NaOH aqueous solution standardised with a potassium biphthalate
solution and using a 1% w/v phenolphthalein ethanol solution as
indicator. Approximately 2 g of sample were weighed and solu-
bilised in a 2:1 ethyl ether:ethanol mixture (v/v) [27]. The result
was calculated and is expressed in mg KOH/g.
2.3.3. Water content
The water contents of the macauba oils were determined using
an 831 KF METROHM coulometer for coulometric titration accord-
ing to ASTM-D6304.
2.3.4. Kinematic viscosity
The kinematic viscosity (kinem. visc.) was measured in a Pre-
citech Ostwald viscometer and Thermo Haake-Phoenix silicone
thermal bath at 40 °C, according to ASTM D445.
All tests were performed in triplicate.
2.4. Pre-treatment of oils by alkaline homogeneous hydrolysis
Some deoxygenation reactions were performed with the mix-
ture of free fatty acids derived from the macauba pulp and almond
oils, with the purpose of evaluating the effect of the feedstock form
on the yields and compositions of the reaction products. For this
purpose, hydrolysis of the crude oils was performed, using basic
homogeneous catalysis, yielding mixtures of free fatty acids.
The reactions were performed for 30 min using 1.67 mol L1 of
KOH ethanolic solution at a ratio of 1:5 w/v of crude oil in relation
to the solution, ethanol reflux temperature and mechanical stirring
at 200 rpm. The resulting product was heated under vacuum to
approximately 60–70 °C until complete evaporation of the ethanol.
The fatty product of the evaporation process was then cooled at
room temperature and acidified with H2SO4 in the form of an aque-
ous solution with a volume ratio of approximately 1:1. The mixture
was vigorously stirred to favour formation of free fatty acids. This
331
product was washed and the pH of the wash water was monitored.
The organic phase, which was composed of a mixture of free fatty
acids, was centrifuged for separation and subsequent removal of
the excess water. This product was named PFA (pulp fatty acid)
or AFA (almond fatty acid), depending on the feedstock used.
2.5. Characterisation of the hydrolysed macauba oils
To better evaluate their properties, some parameters of the
hydrolysed oils were determined. The water content and kinematic
viscosity were determined as previously described for the crude
oils.
2.6. Catalytic deoxygenation reactions with Pd/C
The batch reactions were performed in a Parr 4348 reactor with
a capacity of 100 mL. First, 2.9 g of dodecane were added to 0.5 g of
the Pd/C catalyst to protect it, and the reduction of the metal was
promoted in a pressurised system with 5 bar of H2 at 200 °C for 2 h.
Then, the system was cooled and depressurised. Next, 10 mL of the
fatty feedstock was added to the reactor, and the system was pres-
surised with H2 or N2 at different pressures (10 or 19 bar). The
reaction proceeded for 5 h at 300 °C with and without stirring
(700 rpm). The deoxygenation products were coded, as described
in Table 1.
One of the reactions with PFA was reproduced in the absence of
the catalyst to evaluate the effect of its presence on the reactions.
The reaction without catalyst occurred for 5 h under the following
conditions: 10 bar of H2, 300 °C and no stirring. The reaction pro-
duct was named Blank-PFA.
2.7. Quantification and characterisation of the products
To determine the percentage of feedstock converted into hydro-
carbons under the different synthesis conditions, Fourier transform
mid-infrared spectroscopy (FTIR) was performed based on the
European standard method EN 14078. To plot the analytical curves,
mixtures were prepared with 50-95% diesel oil (hydrocarbons) by
mass with the respective feedstock in increments of approximately
5%. The readings were performed in an ARIS-ZONE ABB
Bomem_MB Series spectrometer with diamond cell ata resolution
of 4 cm1 with 16 scans and a background was made with the die-
sel oil used in the mixtures. The reduction of the bands corre-
sponding to the C@O stretching of carboxylic acid carbonyls
(1710–1715 cm1) and/or esters (1740–1745 cm1) present in
the starting material was monitored. To verify the reliability of
the curves, validation samples, whose hydrocarbon percentages
were known, were analysed, and the deviation between the real
and calculated value was determined based on the equation of
the curve. Hydrogen nuclear magnetic resonance (1H NMR) spec-
troscopy analyses were also used as an auxiliary qualitative
method to monitor the reduction of carbonyl by decreasing the
intensity of the signal associated with the alpha-carbonyl H (at
approx. 2.30 ppm).
Other parameters of the 1H NMR spectra were also evaluated to
determine the sample characteristics. For these analyses, a Bruker
AVANCE DPX 200 spectrophotometer with the TOPSPIN1.3TM –
update level 10 software package was used. The chemical displace-
ments were recorded in parts per million (ppm) using tetramethyl-
silane as the internal reference standard. The spectra were
obtained at 23 °C using a spectral width of 4.14 kHz with 16 acqui-
sitions and digitalisation of the free induction decay of 32 k. The
processing involved the Fourier transform with a band widening
of 0.3 Hz with 32 k points, interactive phase adjustment and auto-
matic baseline adjustment. The integrals were corrected in terms
of the inclination (slope) and horizontal displacement (bias). The
332 L.N. Silva et al. / Fuel 164 (2016) 329–338
Table 1
Coding of the products of the deoxygenation reaction catalysed by Pd/C and performed at 300 °C for 5 h.
Catalyst Sample code Started material
5% w/w Pd/C CPO-10H,mix Crude pulp oil (CPO)
CPO-19H,mix
CAO-10H,mix Crude almond oil (CAO)
CAO-19H,mix
PFA-10H,mix Pulp fatty acid (PFA)
PFA-19H,mix
PFA-10H,no mix
PFA-19H,no mix
AFA-10H,mix Almond fatty acid (AFA)
AFA-19H,mix
AFA-10 N,mix
No catalyst Blank-PFA Pulp fatty acid (PFA)
spectra were edited with the ACD/NMR Processor Academic Edi-
tion: Processor Window software package.
2.8. Determination of the compositions of the products
The species present in the reaction products (except the gas
products) were investigated by gas chromatography coupled with
mass spectrometry (GC–MS) operated in the electronic impact ion-
isation mode. The analyses were performed with previous silyla-
tion by adding 100 lL of N-methyl-N-(trimethylsilyl) trifluoroac
etamide to 10 ± 1 mg of sample. The mixture was left to rest for
30 min at room temperature or under heating (in the case of solid
samples, at room temperature) and then diluted with 600 lL of
dichloromethane. The analyses were performed in a Shimadzu
QP5050 GC–MS with a 100 m  0.25 mm Pona 100 column. The
temperature gradient used was as follows: a constant temperature
of 50 °C for 10 min, heating at 2.5 °C/min to 185 °C, heating at the
rate of 5 °C/min to 250 °C, and remaining at this temperature for
20 min. The injector and detector temperatures were 280 °C and
260 °C, respectively. Helium was used as carrier gas, with a linear
velocity of 17.7 cm/s. The injection volume was 1 lL, and the split
was 1/50. The identification of the peaks was performed using the
GCMS Solution software package by comparing the spectrum
obtained with a database of different libraries.
Based on the results obtained, the percentages of hydrocarbons
in the fractions that correspond to automotive gasoline (hydrocar-
bons from C5 to C10), jet fuel (hydrocarbons from C9 to C15) and
diesel fuel (hydrocarbons from C10 to C25) were calculated accord-
ing to the intervals described by Do Brasil et al. [28]. For a more
efficient comparison of percentage products, the total hydrocar-
bons content was standardised. Similarly, the contents of each
hydrocarbons formed in the product were calculated. The data
analysis performed with the GC–MS results was only performed
for the products whose hydrocarbon content obtained by this tech-
nique, which was determined as a function of the percentage area
of the peaks, was greater than 65% or, in cases in which it was less
than 65%, if the content of hydrocarbons obtained by FTIR agreed
with the result of the GC–MS technique, with a maximum differ-
ence of 10%.
2.9. Performance at low temperature
To evaluate the cold behaviour of the biofuel with the highest
content of liquid hydrocarbons and its blends with jet fuel, differ-
ential scanning calorimetry (DSC) analyses were performed using a
DSC TA Instruments model Q20 calorimeter. The sample was sta-
bilised at 25 °C and then cooled at the rate of 5 °C/min to 75 °C.
Then, 5 to 10 mg of the sample was used in a hermetically closed
aluminium pan with an N2 flow, at rate of 50 mL/min.
Pressure and atmosphere Stirring (rpm)
10 bar, H2 700
19 bar, H2
10 bar, H2
19 bar, H2
10 bar, H2
19 bar, H2
10 bar, H2 No
19 bar, H2
10 bar, H2 700
19 bar, H2
10 bar, N2
10 bar, H2 No
3. Results and discussion
3.1. Characterisations of the Pd/C catalyst
The X-ray diffractogram of the Pd/C catalyst without any previ-
ous treatment is shown in Fig. 1. A wide peak centred at 2h = 23°
that is associated with the carbon phase was observed; this feature
indicates an amorphous carbonaceous material. Characteristic pal-
ladium peaks were observed at 2h = 40°, 47.5° and 68°, along with
graphite carbon signals [29–32].
The low intensity of the signals corresponding to palladium
allows the inference that they occurred in particles of small size
that were highly dispersed in the support [33], characteristics that
favour the activity of a heterogeneous catalyst.
The isotherm of N2 adsorption/desorption obtained for the Pd/C
catalyst is shown in Fig. 2. Pronounced presence of micropores was
observed, considering the large volume of adsorbed gas at low
pressures. This observation is supported by the high surface area
of the material (1068 m2/g). This result is consistent with literature
data. Mäki-Arvela et al. [15] obtained surface areas between 715
and 1126 m2/g for 5% w/w Pd/C catalyst, while Snåre et al. [16]
published a value of 936 m2/g. Additionally, the presence of hys-
teresis in the isothermal indicates the presence of mesopores.
Using calculations from BJH method, it was found mesopores with
diameter of, predominantly, about 36–40 Å.
3.2. Characterisation of the crude macauba oils
The fatty profiles of the macauba pulp and almond oils were
analysed by GC-FID through hydrolysis followed by methylation.
The results are presented in Table 2. They indicate the predomi-
nance of fatty acids of unsaturated carbon chains in the pulp oil,
especially oleic acid, and saturated short-chain fatty acids, such
as lauric acid, in the almond oil. These results are in agreement
with data that were previously reported in the literature
[24,34,35].
Table 3 presents other characteristics of the crude oils used in
this study. The results are presented with 95% reliability.
The higher acidity value of pulp oil is caused by its higher con-
tent of free fatty acids, which can be explained by its higher degree
of unsaturation. These parameters are explained by the greater
exposure of the mesocarp to air, humidity, and soil fungi and bac-
teria, because the fruit is collected after it falls, which favours oxi-
dation and generation of carboxylic acids in the pulp oil. Although
it is more saturated, the almond oil has shorter carbon chains and
lower acidity, thereby resulting in a slightly lower viscosity than
that of pulp oil (because of the smaller number of van der Waals
interactions between chains) and fewer hydrogen bonds,
respectively.
 L.N. Silva et al. / Fuel 164 (2016) 329–338 333

Table 2
Fatty profiles of macauba pulp and almond oils.

Fatty acid Abbreviation CPO (%) CAO (%)

Caprylic acid C8:0 5.29
Capric acid C10:0 3.67
Lauric acid C12:0 0.59 41.85
Myristic acid C14:0 0.28 10.24
Pentadecylic acid C15:0 0.14
Palmitic acid C16:0 18.78 9.36
Palmitoleic acid C16:1 3.2
Stearic acid C18:0 2.62 3.34
Trans-oleic acid C18:1 trans 2.36 1.54
Cis-oleic acid C18:1 cis 53.46 21.29
Trans-linoleic acid C18:2 trans 0.09
Cis-linoleic acid C18:2 cis 15.98 3.41
Arachidic acid C20:0 0.11
Gadoleic acid C20:1 1.18
11,14-Eicosadienoic acid C20:2 1.22
Short-Chain Saturated (6 C16:0) 19.79 70.41
Long-Chain Saturated (> C16:0) 2.73 3.34
Fig. 1. X-ray diffractogram of commercial 5% Pd/C catalyst.
Unsaturateda 77.49 26.24
a
Fatty acids with carbon chains longer than or equal to C16:1.

The deviations in relation to the real hydrocarbon (diesel oil)

value in the verification samples were always less than 1% by
weight, thus indicating that the results obtained using the method-
ology developed in this study are very suitable and efficient for cal-
culating the hydrocarbon content generated in the deoxygenation
reactions.

3.5. Catalytic deoxygenation reactions

3.5.1. Quantification of deoxygenation products

Fig. 2. N2 adsorption/desorption isotherms and specific area calculated by BET
method for commercial 5% w/w Pd/C catalyst.
3.3. Characterisation of the hydrolysed macauba oils
The characterisation of the macauba pulp and almond oils
hydrolysed in the laboratory, PFA and AFA, was performed. The
results are presented in Table 3 and are expressed with 95%
reliability.
Similarly to the macauba crude oil, although it is more satu-
rated, AFA has shorter carbon chains, thus causing its viscosity to
be lower than that of PFA because of the smaller number of van
der Walls interactions between chains.
3.4. Analytical curves for hydrocarbon quantification
The equations of the analytical curves obtained by FTIR for each
feedstock and their respective linear correlation coefficients (R2),
along with the results for the verification samples, are presented
in the Supplementary Material. Note that plotting two analytical
curves was only necessary for the CPO because of the occurrence
of two carbonyl signals from the feedstock, one from the oil
(approx. 1745 cm1) and the other from the free fatty acids
(approx. 1710 cm1), that were present in high concentration in
the matrix (acidity value approx. 76 mg KOH/g or 38% w/w in oleic
acid).
Using the calibration curves for each feedstock obtained by FTIR,
the percentages of formed hydrocarbons were determined (Fig. 3).
Fig. 3 indicates that the use of Pd/C as a catalyst allowed forma-
tion of hydrocarbons from AFA at 10 and 19 bar of H2, with better
performance than that of other feedstocks. This process is espe-
cially interesting for obtaining liquid hydrocarbons under milder
pressure conditions (10 bar), for which the conversion efficiency
was very high (85% w/w).
The better results obtained with AFA may have been influenced
by the lower content of unsaturated components in almond oil
because unsaturated species tend to deactivate and/or decrease
the activity of the catalyst [13,15,17]. The greater formation of
hydrocarbons from predominantly saturated feedstocks (CAO and
AFA), using a lower concentration of H2 in the medium, is in accor-
dance with the conclusions of Immer et al. [36], who used stearic
acid as the starting material. It is also in accordance with the
results of Kubičková et al. [14], who used stearic acid, ethyl stea-
rate and tristearin and observed that higher partial pressures of
H2 inhibited decarboxylation.
In the case of the feedstock derived from the macauba pulp, PFA
and CPO, which are characterised by a composition rich in unsatu-
rated carbon chains, similar conversion efficiencies were obtained
at 10 bar (18% and 23%, respectively). Higher hydrocarbon percent-
ages were achieved with the use of a higher pressure for both feed-
stocks, especially in the case of PFA, for which the conversion
efficiency doubled with the increase in pressure. These results sug-
gest that the deoxygenation kinetics of unsaturated feedstocks is
favoured by higher pressures, this result is consistent with that
of Mäki-Arvela et al. [15]. These authors demonstrated that in reac-
tions with ethyl stearate at 320 °C and 17 bar, higher H2 content in
the atmosphere favoured increased conversion to hydrocarbons.
A comparison of the same fatty acid profile (PFA versus CPO and
AFA versus CAO) demonstrated that the previous hydrolysis of oils
334 L.N. Silva et al. / Fuel 164 (2016) 329–338
Table 3
Characterisation of crude and hydrolyzed macauba pulp and almond oils.
Parameter Method/Reference
Acidity value (mg KOH/g) Acid-base titration
Water content (mg/kg) ASTM D6304
Kinem. Visc. at 40 °C (mm2 s1) ASTM-D445
NA: not applicable.
Fig. 3. Percentage of hydrocarbons formed after the macauba crude oils (CPO and
CAO) and macauba free fatty acids (AFA and PFA) deoxygenation reaction catalysed
with Pd/C under different synthesis conditions.
to fatty acids generally facilitated the formation of hydrocarbons in
reactions catalysed by Pd/C. This effect may be associated with the
presence of interfering compounds in the crude matrix, such as b-
carotene, vitamins, and others [37], that block the catalytic site and
thus the access of reagents. It may also be associated with elimina-
tion of the conversion of triglyceride to mono-chains (free fatty
acids or monoesters), which tends to occur before deoxygenation
[9]. Additionally, it is necessary to consider that smaller chains
have lower viscosity and higher diffusivity, thus favouring the
access of reagents to the catalytic sites.
Fig. 3 also evidences the importance of the catalyst for the occur-
rence of the deoxygenation reaction under the proposed conditions,
because in its absence (Blank-PFA), the dodecane, which was added
to the system before the synthesis, was practically the only hydro-
carbon that was present in the analysed product. The total or almost
total absence of hydrocarbons formed in the non-catalysed reaction
is consistent with the result obtained by Snåre et al. [33] after 6 h of
reaction of stearic acid at 300 °C and 6 bar of helium.
An evaluation of the influence of the reaction atmosphere indi-
cated higher activity of the Pd/C catalyst in H2 atmosphere (85%
conversion) compared with the N2 atmosphere (45% conversion)
under the same synthesis conditions. This difference can be attrib-
uted to the higher activity of the catalyst in its reduced metal form,
which is favoured in the presence of H2. Thus, with liberation of
oxygenated species during the reaction, especially CO2, the metal
tends to remain in the form of Pd2+, which is inactive, thus explain-
ing the result. Additionally, as suggested by Mäki-Arvela et al. [15]
and Immer et al. [36], H2 also has the ability to promote desorption
of organic contaminants at the active surface of the catalyst,
thereby conserving its activity for longer in reactions with the
H2-rich atmosphere.
3.5.2. Spectroscopic characterisation of deoxygenation products
The FTIR and 1H NMR spectra of the deoxygenation products
and their respective feedstocks were analysed, and the most
CPO CAO PFA AFA
75.65 ± 2.71 3.95 ± 0.03 NA NA
2284.2 ± 70.6 1051.7 ± 21.6 953.1 ± 69.0 4827.8 ± 222.5
32.9 ± 1.2 29.6 ± 1.5 18.6 ± 4.7 10.9 ± 0.5
relevant signals were identified for both cases. The percentages
of conversion to hydrocarbons were also determined in the FTIR
spectra. Fig. 4 shows the results for the crude oils and their prod-
ucts, and Fig. 5 shows the results for the hydrolysed oils and their
products.
It can be observed that in the case in which the feedstock was a
crude oil (Fig. 4), there was a reduction in the intensity of ester car-
bonyl (triglycerides), at approximately in 1745 cm1, and an
increase in the intensity of carboxylic acid carbonyl, at approxi-
mately 1712 cm1, thereby indicating that conversion of the oil
to free fatty acid occurred before the hydrocarbon formation step,
as previously reported [8,9]. Note the reduction or disappearance
of the bands associated with the hydroxyl group at approximately
930 and 1410 cm1 (Fig. 5), which indicates its consumption and
the consequent formation of hydrocarbons. Another important
point is the hydrogenation of carbon chains, which was indicated
by the disappearance of the signal that corresponds to alkenes
(3010 cm1) in the crude pulp oil derivatives (Fig. 4a) and con-
firmed by the 1H NMR results (see allyl H and bis-allylic H peaks
in Fig. 4b). This conversion was expected, considering the known
catalytic activity of palladium for hydrogenation, and it can be con-
sidered partially positive for biofuels in terms of oxidative stability.
However, in terms of cold performance, long saturated carbon
chains can be damaging because they have high solidification
temperatures.
Fig. 5a presents a comparison of the FTIR spectra of AFA and its
deoxygenation products with Pd/C under different atmospheres
while keeping the other reaction conditions constant. The spectra
confirmed high consumption of fatty acids (bands associated with
hydroxyl at 930 and 1410 cm1 and carbonyl at 1715 cm1) when
an H2 atmosphere was used and worse conversion in an N2
atmosphere.
Fig. 5c shows a comparison of the PFA products at 10 bar of H2
in the presence and absence of the catalyst. There was a reduction
in the fatty acid content only in the catalysed reaction. However,
two characteristics of the Blank-PFA spectra stand out: (1) the dis-
appearance of the alkene signal, near 3010 cm1, which indicated
that the synthesis conditions were capable of removing unsatura-
tions, even without the hydrogenating activity of palladium. (2)
The appearance of the signal at 970 cm1 that corresponds to
trans-configuration double bonds, which indicated the occurrence
of cis/trans isomerisation in the feedstock, considering the small
amounts of such species in natural products, and the absence of
the signal at 970 cm1 in the PFA spectrum.
In the case of the 1H NMR spectra, significant reductions in the
a- and b-carbonyl hydrogen signals compared with the starting
material (2.3 ppm and 1.6 ppm, respectively) were observed in
all cases (Figs. 4 and 5), thereby corroborating the infrared spec-
troscopy results and confirming the removal of carbonyl. Note also
the disappearance of the H-allylic signal at 2.0 ppm in all reactions
in the presence of the Pd/C catalyst, which corroborates the cata-
lyst’s hydrogenating activity. In the absence of this species, the
intensity of the signal was reduced, but it did not disappear,
thereby indicating that the hydrogenation was due to the reaction
conditions along with the action of palladium. In the products
derived from the macauba pulp, in which there were significant
 L.N. Silva et al. / Fuel 164 (2016) 329–338
Fig. 4. (a) FTIR spectra of CPO and its products, (b) 1H NMR (200 MHz; in CDCl3) spectra of CPO and its products, (c) FTIR spectra of CAO and its products, and (d) 1H NMR
(200 MHz; in CDCl3) spectra of CAO and its products.
amounts of fatty acids with two unsaturations, the signal that cor-
responded to bis-allylic hydrogen (2.7 ppm) in the feedstock also
disappeared, thereby indicating the saturation of the compounds
in the reactions both with and without the catalyst.
3.5.3. Product composition
The composition of samples whose hydrocarbon contents,
determined by GC–MS, were greater than 65% were analysed in
detail. Additionally, when the content was less than 65% but it is
in agreement with the result obtained by FTIR (a difference of less
than 10% between the two techniques), the composition of the
sample was also studied. In these cases, the hydrocarbon percent-
age of the products was normalised, and the percentage of hydro-
carbons that can be used for production of automotive biogasoline,
biojet fuel (BioJF) and green diesel were compared. The results are
shown in Fig. 6. For the biogasoline fraction, the hydrocarbon inter-
val between C5 and C10 was considered. For BioJF, the fraction
between C9 and C15 was used. For green diesel, the fraction from
C10 to C25 was took into account [28], as previously mentioned.
Note that the sum of the fuel fraction percentage for each product
is not 100% because of the occurrence of chains with the same
number of carbons that can be presented in more than one kind
of biofuel.
An occurrence of small or null percentages of hydrocarbons that
correspond to automotive gasoline (i.e. lighter hydrocarbons) was
335
observed. Biogasoline was the least-favoured product under the
synthesis conditions tested with 5% Pd/C as the catalyst. However,
there were significant contents of BioJF and green diesel. With the
exception of CPO-19H,mix, all of the products analysed can be used
in their totality or near-totality to obtain alternative diesel oil,
always with percentages of BioJF greater than 50%, reaching a max-
imum in the CAO-10H,mix product. In the case of CPO-19H,mix,
almost all of the hydrocarbons formed correspond to kerosene,
thereby indicating that a highly oxidised oil of low aggregated
value exposed to pressures of approximately 19 bar of H2 in the
presence of a catalyst can be used to obtain BioJF components.
Note that once again, the presence/absence of stirring during the
reaction with PFA did not affect the hydrocarbon fractions
obtained. The BioJF fraction must likely still be subjected toan iso-
merisation step because that the product obtained in the present
study is linear and saturated and therefore should have inadequate
cold properties.
Table 4 presents the contents of the main hydrocarbons of each
of the samples analysed. In all cases, the hydrocarbons formed
were always linear and saturated; the latter characteristic is com-
patible with the hydrogenating activity of palladium and with the
need for a saturation step prior to deoxygenation, a fact that is
already known in the literature [15,17].
The distribution of non-oxygenated products displayed in
Table 4 evidences the higher formation of hydrocarbons with
336 L.N. Silva et al. / Fuel 164 (2016) 329–338
Fig. 5. (a) FTIR spectra of AFA and its products, (b) 1H NMR (200 MHz; in CDCl3) spectra of AFA and its products, (c) FTIR spectra of PFA and its products and (d) 1H NMR
(200 MHz; in CDCl3) spectra of PFA and its products.
Fig. 6. Hydrocarbon fractions for automotive biogasoline, BioJF and green diesel in
deoxygenation with Pd/C products.
12 and 17 carbons. The predominance of dodecane, however, is
caused by the addition of this compound to the system before
the reaction; thus, it is not completely a reaction product. Based
on the literature [16,38], the predominant route in deoxygenation
reactions of fatty materials catalysed by Pd/C is decarboxylation,
which involves elimination of CO2 and results in hydrocarbons
with one less carbon atom than the fatty acid of origin, i.e.,
hydrocarbons with odd numbers of carbons. Considering the fatty
composition of the feedstock used (Table 1), it could be inferred
that the main expected products for the pulp derivatives are
heptadecane (approx. 75% of the fatty acids with 18 carbons), fol-
lowed by pentadecane (approx. 22% of the fatty acids with 16
carbons). In the case of the almond oil, the predominance of
undecane (approx. 42% of the fatty acids with 12 carbons) is
expected, followed by heptadecane (approx. 30% of the fatty acids
with 18 carbons). The results presented in Table 4, except for the
percentages of dodecane, are in accordance with these numbers
for the pulp derivatives. In the case of the almond oil, however,
heptadecane and undecane predominate, with the content of
the former higher than the latter, in contrast with what was
expected. This result makes it possible to infer that the deoxy-
genation reaction occurred preferentially in species with longer
carbon chains, which disagrees with the results of Mäki-Arvela
et al. [15]. It is important to note that this explanation cannot
be based on a difference in the unsaturation degree because
hydrogenation dominates deoxygenation [15,17,36]; thus,
removal of oxygen always occurs when the carbon chains are
already saturated.
 L.N. Silva et al. / Fuel 164 (2016) 329–338
Table 4
Normalised percentage content of the main hydrocarbons obtained in the products of deoxygenation catalysed by Pd/C.
Sample code C11H24 C12H26
CPO-10H,mix 0 65
CPO-19H,mix 0 95
CAO-10H,mix 1 99
CAO-19H,mix 15 27
PFA-19H,mix 0 58
PFA-10H,no mix 0 50
PFA-19H,no mix 0 59
AFA-10H,mix 20 32
3.6. Evaluation of the cold performance
The behaviour of the AFA-10H,mix synthesised biofuel was
evaluated at low temperatures, in its pure form and mixed with
commercial jet fuel with ratios ranging from 1% (B1) and 5% v/v
(B5), using thermal analysis (DSC). The results are presented in
Fig. 7.
As demonstrated by Fig. 7, the chosen AFA product catalysed by
Pd/C comprised various crystallisation fractions, which generated
different peaks throughout the analysis, with the first peak
appeared at +13.8 °C, whereas the pure JF started crystallisation
at 62.2 °C. None of the product fractions whose peaks were
observed by DSC have freezing temperatures that are similar to
that of jet fuel, thereby indicating that even after refinement, the
product would likely not be able to be used in its pure form. How-
ever, the blend of 5% of AFA-10H,mix with jet fuel changed the ini-
tial solidification value to 46.8 °C, which is a very satisfactory
result that can be improved by refining the product and using
antifreezing agents. This result indicates the potential application
of this mix as a partial substitute for fossil fuels, especially in cases
with less strict cold performance requirements, such as green die-
sel fuel. For the blend with 1% of AFA-10H,mix in jet fuel, the crys-
tallisation temperature almost equalled that of fossil jet fuel,
thereby indicating the possible commercial use of this mixture at
this proportion. The addition of the volumetric content of 1% of
synthetic hydrocarbons to kerosene is a possibility that has already
been considered by the airplane manufacturer Boeing [39] and
would represent a savings of more than 3.0 billion litres of fossil
jet fuel per year, based on the worldwide consumption of this fuel
in 2010 [40].
Fig. 7. DSC curves for jet fuel, AFA-10H,mix and blends with 1 and 5% v/v of AFA-
10H,mix in jet fuel.
337
C15H32 C16H34 C17H36 C18H38 Other
0 0 35 0 0
1 0 4 0 0
0 0 0 0 0
5 0 47 0 7
6 0 36 0 0
7 5 39 0 0
5 0 35 0 0
6 0 29 1 11
4. Conclusions
This study demonstrated that the Pd/C catalyst reduced in situ
was efficient at promoting deoxygenation and hydrogenation of
macauba pulp and almond oils in the presence of hydrogen. In gen-
eral, higher hydrocarbon percentages were obtained with the pre-
viously hydrolysed oils, especially almond oil under 10 bar of H2,
for which the formation of heptadecane predominated, followed
by undecane. This result indicated that deoxygenation occurred
preferentially in fatty acids with longer carbon chains. It was
observed that deoxygenation occurred with the release of carbon
oxides, starting from triglycerides and free fatty acids, because
hydrocarbons with odd numbers of carbon atoms predominated.
Higher H2 concentrations, i.e., higher pressures, favoured deoxy-
genation of pulp derivatives but not almond derivatives, thereby
indicating the significant effect of the feedstock on the final com-
position of the product. The hydrocarbons produced were satu-
rated linear species, which improved the oxidative stability but
raised the crystallisation temperatures of these products. However,
the study of the cold performance using DSC indicated promising
results for the blends with 5% and especially 1% v/v of AFA-10H,
mix in commercial jet fuel, whose solidification started at 47
and 60 °C, respectively, even though the product did not undergo
a refinement step and no antifreezing additives were added. Hence,
it is concluded that the macauba palm tree, which is native to
South America, has high potential as feedstock for industrial pro-
duction of green diesel and biojet fuel.
Acknowledgements
We acknowledge financial support from Agência Nacional do
Petróleo, Gás Natural e Biocombustíveis – ANP, the Financiadora
de Estudos e Projetos – FINEP and the Ministério da Ciência, Tec-
nologia e Inovação – MCTI through the Programa de Recursos
Humanos of ANP for the Oil and Gas Sector– PRH –ANP/MCTI).
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.fuel.2015.09.081.
References
[1] Industry Market trends. The damage done in transportation – which energy
source will lead to the greenest highways? April 30, 2012, <http://news.
thomasnet.com/imt/2012/04/30/the-damage-done-in-transportation-which-
energy-source-will-lead-to-the-greenest-highways> [accessed 17.11.14].
[2] IEA statistics. CO2 emission from fuel combustion – highlights. 2013 ed. France.
Organisation for Economic Co-operation and Development (OECD); 2013.
[3] Suarez PAZ, Santos ALF, Rodrigues JP, Alves MB. Biocombustíveis a partir de
óleos e gorduras: Desafios tecnológicos para viabilizá-los. Quím Nova
2009;32:768–75.
[4] ANP-Agência Nacional do Petróleo, Gás Natural e Biocombustíveis. Resolution
ANP N° 37, de 1°.12.2009 – DOU 2.12.2009.
338 L.N. Silva et al. / Fuel 164 (2016) 329–338
[5] ASTM D1655-15. Standard specification for aviation turbine fuels.
[6] ANP-Agência Nacional do Petróleo, Gás Natural e Biocombustíveis. Resolution
ANP N° 63, de 5.12.2014 – DOU 8.12.2014.
[7] ASTM D7566-13. Standard specification for aviation turbine fuel containing
synthetized hydrocarbons.
[8] Filho GNR, Brodzki D, Djega-Mariadassou G. Formation of alkanes,
alkylcycloalkanes and alkylbenzenes during the catalytic hydrocracking of
vegetable oils. Fuel 1993;72:543–9.
[9] Kim SK, Han JY, Lee H-s, Yum T, Kim Y, Kim J. Production of renewable diesel
via catalytic deoxygenation of natural triglycerides: comprehensive
understanding of reaction intermediates and hydrocarbons. Appl Energy
2014;116:199–205.
[10] De Araújo LRR, Scofield CF, Pastura NMR, Gonzalez WA. H3PO4/Al2O3 catalysts:
characterization and catalytic evaluation of oleic acid conversion to biofuels
and biolubricant. Mater Res 2006;9:181–4.
[11] Reguera FM, De Araujo LRR, Picardo MC, Bello FO, Scofield CF, Pastura NMR,
et al. The use of niobium based catalysts for liquid fuel production. Mater Res
2004;7:343–8.
[12] Gong S, Shinozaki A, Shi M, Qian EW. Hydrotreating of jatropha oil over
alumina based catalysts. Energy Fuels 2012;26:2394–9.
[13] Simakova I, Rozmysłowicz B, Simakova O, Mäki-Arvela P, Simakov A, Murzin
DY. Catalytic deoxygenation of C18 fatty acids over mesoporous Pd/C catalysts
for synthesis of biofuels. Top Catal 2011;54:460–6.
[14] Kubičková I, Snåre M, Eränen K, Mäki-Arvela P, Murzin DY. Hydrocarbons for
diesel fuel via decarboxylation of vegetable oils. Catal Today
2005;106:197–200.
[15] Mäki-Arvela P, Kubičková I, Snåre M, Eränen K, Murzin DY. Catalytic
deoxygenation of fatty acids and their derivatives. Energy Fuel 2007;21:
30–41.
[16] Snåre M, Kubičková I, Mäki-Arvela P, Eränen K, Murzin DY. Heterogeneous
catalytic deoxygenation of stearic acid for production of biodiesel. Ind Eng
Chem Res 2006;45:5708–15.
[17] Snåre M, Kubičková I, Mäki-Arvela P, Chicova D, Eränen K, Murzin DY. Catalytic
deoxygenation of unsaturated renewable feedstocks for production of diesel
fuel hydrocarbons. Fuel 2008;87:933–45.
[18] Han J, Duan J, Chen P, Lou H, Zheng X, Hong H. Nanostructured molybdenum
carbides supported on carbon nanotubes as efficient catalysts for one-step
hydrodeoxygenation and isomerization of vegetable oils. Green Chem
2011;13:2561–8.
[19] Choudhary TV, Phillips CB. Renewable fuels via catalytic hydrodeoxygenation.
Appl Catal A 2011;397:1–12.
[20] Choudhary TV, Parrott S, Johnson B. Unravelling heavy oil desulfurization
chemistry: targeting clean fuels. Environ Sci Technol 2008;42:1944–7.
[21] Lanes ECM, Costa PMA, Motoike SY. Alternative fuels: Brazil promotes aviation
biofuels. Nature 2014;511:31.
[22] Acrotech – Macauba. <http://acrotech.com.br/> [accessed 27.07.11].
[23] Motoike SY, Kuki KN. The potential of macaw palm (Acrocomia aculeata) as
source of biodiesel in Brazil. IRECHE 2009;1:632–5.
[24] Pires TP, Souza ES, Kuki KN, Motoike SY. Ecophysiological traits of the macaw
palm: a contribution towards the domestication of a novel oil crop. Ind Crops
Prod 2013;44:200–10.
[25] Guo H, Hu C, Qian J. Determination of underivatized long chain fatty acids
using hplc with an evaporative light-scattering detector. J Am Oil ChemSoc
2011;89:183–7.
[26] Christie WW. Gas chromatography and lipids: a practical guide. Oxford
(UK): Pergamon Press; 1989.
[27] IAL-Instituto Adolfo Lutz. Métodos físico-químicos para análise de alimentos.
4th ed. São Paulo; 2008.
[28] Do Brasil NI, Araújo MAS, De Souza ECM. Processamento de petróleo e gás. Rio
de Janeiro: LTC; 2012.
[29] Zhang Y, He X, Ouyang J, Yang H. Palladium nanoparticles deposited on
silanized halloysite nanotubes: synthesis, characterization and enhanced
catalytic property. Sci Rep 2013;3:2948–53.
[30] Nassor ECO. Catalisadores de Pd metálico modificados por líquidos iônicos
para a hidrogenação seletiva de compostos poliinsaturados. 2012. Dissertation
(Doctorate in Inorganic Chemistry) – Universidade Federal de Minas Gerais
[Federal University of Minas Gerais], Belo Horizonte; 2012. 128p.
[31] Samavat F, Ahmad PT, Mahmoodi F, Samavat MF, Tavakoll MH. The effect of
annealing on the size and morphology of palladium nanoparticles. Am J
Condens Mat Phys 2012;2:73–6.
[32] Tejos M, Schrebler R, Díaz F, Del Valle MA. Preparation of palladium thin films
and their usefulness as modified electrodes. Bol Soc Chil Quim
2002;47:113–21.
[33] Castro CS, Ferreti C, Di Cosimo JI, Assaf JM. Support influence on the basicity
promotion of lithium-based mixed oxides for transesterification reaction. Fuel
2013;103:632–8.
[34] Fortes ICP, Baugh PJ. Study of analytical on-line pyrolysis of oils from macauba
fruit (Acrocomia sclerocarpa M) via GC/MS. J Braz Chem Soc 1999;10:469–77.
[35] Silva LN. Síntese e caracterização de biodiesel a partir dos óleos da Macaúba
para usos especiais incluindo blendas com querosene de aviação. 2013. Thesis
(Master’s in Organic Chemistry) – Universidade Federal de Minas Gerais, Belo
Horizonte; 2013. 95p.
[36] Immer JG, Kelly MJ, Lamb HH. Catalytic reaction pathways in liquid-phase
deoxygenation of C18 free fatty acids. Appl Catal A – Gen 2010;375:134–9.
[37] Fréty R, Da Rocha MGC, Brandão ST, Pontes LAM, Padilha JF, Borges LEṔ, et al.
Cracking and hydrocracking of triglycerides for renewable liquid fuels:
alternative processes to transesterification. J Braz Chem Soc 2011;22:1206–20.
[38] Wang W-C, Thapaliya N, Campos A, Stikeleather LF, Roberts WL. Hydrocarbon
fuels from vegetable oils via hydrolysis and thermo-catalytic decarboxylation.
Fuel 2012;95:622–9.
[39] Reuters Brasil. Boeing pretende buscar aprovação para usar ’diesel verde’ como
combustível; jan. 2014. <http://br.reuters.com/article/businessNews/
idBRSPEA0S1XP20140114> [accessed 20.11.14].
[40] IEA-International Energy Statistics, <http://www.eia.gov/cfapps/ipdbproject/
iedindex3.cfm?tid=5&pid=63&aid=2&cid=ww,&syid=2007&eyid=2012&unit=
TBPD> [accessed 2.2.15].