You are on page 1of 8

Fuel 235 (2019) 1123–1130

Contents lists available at ScienceDirect

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

Full Length Article

Biodiesel purification by column chromatography and liquid-liquid T


extraction using green solvents

Sérgio S. de Jesus , Gabriela F. Ferreira, Maria Regina Wolf Maciel, Rubens Maciel Filho
Laboratory of Optimization, Design and Advanced Control, School of Chemical Engineering, University of Campinas, P.O. 6066, 13083-852 Campinas, SP, Brazil

A R T I C LE I N FO A B S T R A C T

Keywords: Biodiesel purification through acid-catalyzed in situ transesterification of dry biomass of Chlorella pyrenoidosa
Biodiesel was evaluated by liquid-liquid extraction and column chromatography, using the solvents hexane, 2-methylte-
Green solvent trahydrofuran (2-MeTHF), and cyclopentyl methyl ether (CPME). The purified biodiesel were characterized
Microalgae through their chemical composition and physicochemical characteristics. The results showed that the purified
Liquid-liquid extraction
biodiesel presented a purity level higher than 96.5%, and the purities obtained with CPME were very close to
Column chromatography
those obtained with hexane. Concerning physicochemical characteristics, biodiesel purified with green solvents
showed higher densities and viscosities, but compliant with the European and North American quality standards.
The difficulty to evaporate residual water from biodiesel purified with 2-MeTHF after the liquid-liquid extraction
process required that biodiesel was left in decantation for 24 h, which resulted in a biodiesel with water content
exceeding 0.05%mass. Biodiesel purified with green solvents, mainly by column chromatography using CPME as
eluent, show great potential to replace hexane in the biodiesel purification process, and the price of these
solvents is the only factor limiting their large-scale use.

1. Introduction biodiesel, as well as other impurities, is of great importance for the


production of a fuel that meets the established technical standards,
The concern regarding the replacement of petroleum-based fuels has ensuring higher performance in use, as well as protecting the engine
grown considerably in recent years. From the economic point of view, from possible damage. The most widely used purification procedure is
the threat of shortages and price instability, and from the environ- performed primarily through liquid-liquid extraction or column chro-
mental point of view, the need to reduce climate changes caused also by matography (at the level of laboratory) and involves the use of large
the effects of their use and energy sustainability have promoted the amounts of solvent, hexane is usually used in this step due to presenting
search for alternative renewable fuels that are less harmful to the en- greater solubility in biodiesel, easy recovery and low-cost [6]. However,
vironment [1]. As an alternative to fossil fuels, the production and use hexane is a solvent of fossil origin, highly toxic, classified as category 2
of biofuels from plant and animal raw materials has become a priority. (in a 1–10 scale) for human health and for the environment, being
Considered as the best alternative to meet a growing demand, biodiesel strongly recommended its replacement with other less toxic solvent
from microalgae cultivation gained wider prominence in research and [6,7]. As an alternative, a new class of solvents that are en-
government investments, mainly in the United States [2]. However, vironmentally friendly and comply with the principles of “green
large-scale production is still a limiting factor for the use of this biofuel, chemistry” has emerged recently. In addition to their high potential in
mainly because of some technical problems related to the downstream the use of organic syntheses, these solvents present properties that are
step. With the aim of reducing some steps of the process in order to very similar to the solvents used in lipid extraction process. 2-methyl-
make it cheaper, in situ or direct transesterification of biomass has be- tetrahydrofuran (2-MeTHF) and cyclopentyl methyl ether (CPME) are
come a promising and economically viable alternative [3–5]. In direct environmentally sustainable organic solvents and showed to be ex-
transesterification, extraction and transesterification occur simulta- cellent alternatives as solvents for lipid and protein extraction [8,9].
neously, alcohol acts as extraction solvent and esterification agent However, these solvents have not been tested for biodiesel purification
[3–5]. After the process, the biodiesel formed is separated from the and may present as an environmentally correct alternative, since the
catalyst (acid or base), from the excess alcohol, glycerol and glycerides, purification process requires a considerable volume of solvent, which
in addition to the residual biomass [4]. The separation of glycerol from generates a large amount of residue, even if part of the solvent used is


Corresponding author.
E-mail address: ssjesus@gmail.com (S.S. de Jesus).

https://doi.org/10.1016/j.fuel.2018.08.107
Received 23 May 2018; Received in revised form 19 August 2018; Accepted 24 August 2018
Available online 30 August 2018
0016-2361/ © 2018 Elsevier Ltd. All rights reserved.
S.S. de Jesus et al. Fuel 235 (2019) 1123–1130

reused. anhydrous sodium sulfate (∼1.0 g) to remove remaining water.


Biodiesel purification with the solvents 2-methyltetrahydrofuran Sodium sulfate was separated by vacuum filtration and then the
and cyclopentyl methyl ether is proposed in this work. Two different solvent was evaporated in a rotary evaporator (RV 10, IKA, Germany)
methods were chosen for purification of the mixture obtained after (Fig. 1D). Biodiesel was dried at 80 °C for 24 h.
acid-catalyzed direct transesterification: a) liquid-liquid extraction; b)
column chromatography. 2.2.3.2. Column chromatography. Purification of the crude extract
Purification with hexane was used for comparative purposes. obtained in direct transesterification reaction by column
chromatography consisted in adding the sample (4.5 ± 0.5 g) into a
2. Materials and methods chromatographic column packed with 10 g silica gel (0.035–0.07 mm,
60A) and 60 g alumina (aluminum oxide 90 neutral, 0.05–0.2 mm,
2.1. Materials 70–270 mesh ASTM) at 1:6 ratio, as eluent we used approximately
300 mL a mixture of hexane:ethyl acetate (19:1) or the green solvents:
Microalgae: Chlorella pyrenoidosa was purchased from Taiwan 2-MeTHF or CPME. The collected fractions were submitted to thin-layer
Chlorella Manufacturing (Taiwan, China). chromatography (TLC), using as mobile phase a mixture of petroleum
Chemicals: Hexane and ethyl acetate (Labsynth Ltda., Brazil), 2- ether, ethyl ether and glacial acetic acid (80:19:1), and the
methyltetrahydrofuran and cyclopentyl methyl ether (Sigma-Aldrich chromatogram was developed with iodine vapors. Oleic and
Co., Germany) were used as solvents for biodiesel purification. heptadecanoic were used as chromatographic standards [11]
(Fig. 1C). After separation, FAME was washed with 50 mL distilled
2.2. Methods water according to the procedure described in the previous section and
represented in Fig. 1C.
2.2.1. Determination of lipid content in dry microalgal biomass The yield of each process of separation was calculated according to
The lipid content in dry microalgal biomass (3.5 ± 0.2 wt% the equation:
moisture) refers to the total lipid extracted by petroleum ether after the
bound lipid hydrolysis [10]. Biodiesel mass (g)
Biodiesel yield (%) = × 100%
Boiling water (6 mL) was added to 1.00 ± 0.05 g of biomass, vortex microalgae mass (g) × oil content (%) (1)
mixed for 1 min and cooled; 1 mL of 25% ammonia solution was added
The samples were encoded as: HXLL, MeTHFLL, CPMELL, HXCC,
and vortexed for 2 min, followed by 7.5 mL methanol addition and
MeTHFCC and CPMECC, where: HX, MeTHF and CPME represent the
vortex for 2 min. Finally, 17 mL of diethyl ether was added, stirred
solvents hexane, 2-MeTHF and CPME respectively; LL for purification
vigorously, and 17 mL of petroleum ether was added for extraction, and
using liquid-liquid extraction and CC for purification by column chro-
stirred once more. The mixture was centrifuged and the top layer
matography.
containing lipids was separated from the aqueous phase. Solvent was
All experiments were performed in triplicate.
evaporated using a rotary evaporator (RV 10, IKA, Germany). The lipid
fraction was dried until constant weight in a vacuum oven at 100 °C.
The moisture content from the microalgae was determined grav- 2.2.4. Analytical techniques
imetrically with three replicates. 2.2.4.1. Gas chromatography. Fatty acids methyl esters were analyzed
by gas chromatography (GC/FID Agilent 6850, USA) equipped with
2.2.2. In situ transesterification capillary column DB-23 Agilent (50% cyanopropyl and 50%
Dry C. pyrenoidosa (50 g) was mixed with 5.4 mL catalyst (20% methylpolysiloxane), 60 m × 0.25 mm internal diameter, 0.25 μm film
sulfuric acid in relation to biomass) and 150 mL methanol and sub- thickness. The operating conditions of the gas chromatograph were:
mitted to ultrasound bath at 25 °C for 20 min. The mixture was heated 1 mL/min carrier gas flow (He), 24 cm/s linear speed, 280 °C detector
to 100 °C under reflux for 4 h. After transesterification, the product temperature, 250 °C injector temperature, 110–215 °C oven
obtained was left at rest for 1 h. The reaction mixture was filtered under temperature (5 °C/min heating), 215 °C for 24 min, 1.0 μL injection
vacuum in a büchner funnel. Residual biomass was resuspended in volume, and 1:50 split ratio. For identification and quantification, a
methanol and stirred for 1 h, the resulting mixture was separated and standard FAME mixture (Supelco® 37 component FAME mix, Sigma,
the filtrate was added to the raw product obtained previously for the Germany) and an internal C17:0 standard (Heptadecanoic – Sigma,
subsequent purification step. This procedure was performed two more Germany) were used. Ester content was calculated through all peaks
times with the purpose of recovering any trace of fatty acid methyl between C14 and C24 and relating it with the internal standard.
esters (FAME) contained in the residual biomass (Fig. 1A). Quantifications of free and total glycerol, mono, di and triglycerides
were conducted according to ASTM D6584, using butanetriol (Sigma,
2.2.3. Purification of crude methyl esters Germany) as internal standard for quantification of glycerol and trica-
Methyl esters produced after in situ transesterification process were prin (Sigma, Germany) as internal standard for quantification of mono,
purified with two different techniques, in order to verify the influence di and triglycerides.
of green solvents on the properties of purified biodiesel and its yield.
2.2.4.2. Chemical and physicochemical analyses. Density measurement
2.2.3.1. Liquid-liquid extraction. The resulting mixture obtained was performed at 15 °C using the hydrometer method.
previously (4.5 ± 0.5 g) was added in a separatory funnel, in Kinematic viscosity at 40 °C was tested according to standard ASTM
addition to 200 mL solvent (hexane or 2-MeTHF or CPME) and D445.Water content was determined by standard ASTM 6304 and acid
200 mL distilled water that was stirred and left at rest for 15 min. value through ASTM D664.
After this period the upper FAME-rich phase was separated from the Iodine number was calculated based on the composition of the re-
intermediate and lower phases (Fig. 1B), to these phases were added sidual oil obtained by GC. According to the official method of the
200 mL solvent (hexane or 2-MeTHF or CPME) and left at rest for American Oil Chemist’s Society (AOCS) Cd 1c-85, iodine number is
additional 15 min, the upper phase was separated and added to obtained through the following equation:
previously obtained fraction. This procedure was performed three
IN = (%C16: 1 × 0.950) + (%C18: 1 × 0.860) + (%C18: 2 × 1.732)
times. FAME-rich phase was washed with 50 mL distilled water (to
remove residues of catalyst and methanol) several times until the water + (%C18: 3 × 2.616) + (%C20: 1 × 0.785) + (%C22: 1 × 0.723)
pH was neutral. After separation, we added the organic fraction, (2)

1124
S.S. de Jesus et al. Fuel 235 (2019) 1123–1130

Fig. 1. Schematic representation of the in situ transesterification and purifications steps used in this study (Adapted from Ref. [4]).

where: IN is the iodine number, in centigrams of iodine absorbed per molar fractions of all fatty acids in the sample), 56.1 is the molecular
gram of sample; % Ca:b is the percentage of fatty acid with a carbons mass of KOH and 92.09 represents the molecular mass of glycerol.
and b unsaturations obtained by GC. The method for estimation of cetane index (CI) was determined
The saponification value was calculated based on the composition in using the values obtained for iodine number (IN) and saponification
fatty acids and according to method Cd 3a-94 of AOCS, as shown in Eq. value (SV) and according to the Krisnangkura equation [12]:
(3):
5458
CI = 46.3 + −0.225 × IN
3 × 56.1 × 1000 SV (4)
SV =
[(MMt × 3) + 92.09]−(3 × 18) (3)

where MMt corresponds to the mean molecular weight (sum of the

1125
S.S. de Jesus et al. Fuel 235 (2019) 1123–1130

2.2.5. Solubility parameters Purity


100 98.1 96.6 97.3 97.8 96.7 97.4
To predict the compatibility of the solvents under test (2-MeTHF
and CPME) with the solute (biodiesel), the Hansen solubility para-
meters (HSP) were studied for miscibility and solvation and were 80
Biodiesel
compared with the standard solvent (hexane, hexane:ethyl acetate).
The cohesive energy density is more conveniently treated in terms

Yields (%)
of the solubility parameter δ where δ2 = E/V, where E is the cohesive 60

energy and V is the molar volume. This gives us the classic formula for
Hansen Solubility parameters, where the total parameter, δt, is divided 40
into δD, δP and δH for dispersion, polar and hydrogen-bonding [13]:
δt2 = (δD )2 + (δP )2 + (δH )2 (5) 20
½
The units of solubility parameters are MPa .
For a solvents mixture, composite Hansen solubility parameter, 0
δi,mix, is calculated with: HXLL MeTHFLL CPMELL HXCC MeTHFCC CPMECC

δi, mix = φ1 δi,1 + φ2 δi,2 + … (6) Fig. 2. Effects of purification method on biodiesel yield and purity by in situ
transesterification of dried C. pyrenoidosa biomass.
Subscript i signifies one of the three components of the Hansen
solubility parameters, φ is the volumetric fraction of each solvent in the
mixture. purified product with water in order to remove any traces of con-
The HSP distance between two molecules (Ra) can be conveniently tamination by catalyst, this procedure was performed several times
viewed with a spherical representation. Hansen solubility parameters of until the neutral pH of the solution was reached, which resulted in a
the solute are in the center of the solubility sphere and the solubility small loss of product. As observed in Fig. 2 the yield was obtained in the
sphere radius (Ro) indicates the extent of the interaction for occurrence following order: CPME > 2-MeTHF > Hexane. The yield of the pur-
of solubilization [13,14]. The lower the Ra, the higher the probability ification process using these three solvents can be explained through
of them being compatible, that is, good solvents are within the solu- the Hansen solubility parameters (HSP). In this work, we used the data
bility sphere and bad solvents are outside it [13,14]. available for soy biodiesel [14] to represent the algae biodiesel. The Ro
value of 11.33 MPa1/2 [15] was used to indicate the marginal solubility
Ra2 = 4 × (δDS−δDP )2 + (δPS−δPP )2 + (δHS−δHP )2 (7) of the solute in the solvent and RED was calculated according to Eq. (8)
Subscript S is for the solvent and subscript P is for the solute. to predict the solubility of the solvent in the mixture of FAME. As ob-
If Ra < Ro, the solvent is within solubility sphere and the solute is served in Table 1, the three solvents used in this study showed RED
soluble in the solvent. The Relative Energy Difference (RED) is the value < 1, indicating that they are highly soluble in biodiesel, and the
numerical representation of this graphical visualization. The lower the RED values for 2-MeTHF and CPME were lower than those obtained for
RED value, the better the solvent is dissolved in the solute [13,14]. hexane and hexane:ethyl acetate (19:1). According to the theory ap-
plied to the HSP, the lower the RED value, the better the solvent for
RED = Ra/ Ro (8)
extraction/purification. By comparing the values obtained for RED
(Table 1) with the experimental results (Fig. 2) it can be observed that
2.2.6. Statistical analysis the “green solvents” presented better yields, in this case, the theoretical
All experiments were performed in triplicate. The standard devia- data were similar to the experimental data, especially when biodiesel
tion of the measures was calculated to verify the reliability of the re- was purified by column chromatography, with the percentage of pur-
sults. The differences in mean values were evaluated through the ification obtained for 2-MeTHF and CPME being very close. On the
method of analysis of variance (ANOVA). The differences associated other hand, it was observed that the use of hexane as eluent for chro-
with p < 0.05 were considered significant. matographic separation was not very effective, requiring increase in the
polarity of the eluent with the addition of ethyl acetate, which theo-
3. Results and discussion retically results in a lower value for Relative Energy Difference (RED)
from 0.9 to 0.8, which implies a greater capacity for solubilization with
3.1. Influence of purification methods on biodiesel yields and purity biodiesel. Concerning the difference between the yields obtained in the
liquid-liquid separation process, it was observed that after evaporation
Chlorella pyrenoidosa, with 12.6 ± 0.3 wt% of lipids and with
moisture content lower than 3.5 ± 0.2 wt%, was used for in situ
Table 1
transesterification reaction according to the protocol described in
Hansen solubility parameters for Hexane, 2-MeTHF, CMPE and soy biodiesel.
Section 2.2.1. The yields obtained after the transesterification and Ra was calculated with soy biodiesel as a solute and Ro = 11.33 MPa1/2.
purification process were calculated according to Eq. (1), and biodiesel
purity was defined as the percentage of mass of methyl esters in the Organic δD (MPa1/2) δP (MPa1/2) δH δt (MPa1/2) Ra RED
solvent (MPa1/ (MPa1/
final product after purification. Fig. 2 shows the yields obtained after 2
) 2
)
the transesterification process and the biodiesel purity level after pur-
ification. In all cases the percentage of purification and the purity were Hexane* 14.9 0.0 0.0 14.9 9.7 0.9
greater than 70% and 96%, respectively. Fig. 2 shows that liquid-liquid 2-MeTHF*** 16.4 4.8 4.6 25.8 5.2 0.5
CPME* 16.7 4.3 4.3 25.3 5.7 0.5
extraction provided higher yields than purification by column chro- Ethyl acetate* 15.8 5.3 7.2 28.3 – –
matography. In liquid-liquid extraction efficiency depends on the affi- Hexane:Ethyl 14.9 0.3 0.4 16.0 9.2 0.8
nity of the solute for the extraction solvent, on the ratio of the phases acetate
and on the number of extractions, which leads to lower loss, since the (19:1)
Soy 15.0 3.7 8.9 27.6 – –
procedure was repeated three times. In purification by column chro-
bio-
matography, part of the fractions containing the material of interest diesel**
(FAME) can be removed together with the fraction of unwanted ma-
terial. In the two techniques employed, we opted for washing the Retired from: *Ref. [13]; **
Ref. [15]; ***
Ref. [16].

1126
S.S. de Jesus et al. Fuel 235 (2019) 1123–1130

Table 2
Characterization of biodiesel obtained by in situ transesterification and purified by liquid-liquid extraction and column chromatography.
Property Purification by liquid-liquid extraction Purification by column chromatography

HXLL MeTHFLL CPMELL HXCC MeTHFCC CPMECC

Density at 15 °C [g/L] 0.87 ± 0.03 0.89 ± 0.06 0.89 ± 0.05 0.87 ± 0.05 0.89 ± 0.02 0.89 ± 0.04
Kinematic viscosity at 40 °C [mm2/s] 3.97 ± 0.50 4.23 ± 0.66 4.21 ± 0.05 3.96 ± 0.03 4.28 ± 0.06 3.98 ± 0.06
Acid value [mg KOH/g oil] 0.41 ± 0.06 0.45 ± 0.09 0.37 ± 0.08 0.39 ± 0.06 0.42 ± 0.01 0.37 ± 0.03
Iodine number [g I2/100 g] 119 127 125 74 117 120
Water content [%] 0.04 ± 0.01 0.09 ± 0.01 0.03 ± 0.01 0.04 ± 0.01 0.05 ± 0.01 0.03 ± 0.01
Saponification value [mg KOH/g oil] 199 200 199 203 198 197
Cetane index (CI) 47 45 47 57 48 47
Monoglycerides [m/m] 0.15 0.22 0.46 0.22 0.71 0.12
Diglycerides [m/m] 0.27 0.00 0.28 0.00 0.26 0.00
Triglycerides [m/m] 0.00 0.02 0.01 0.00 0.01 0.00
Free glycerol [m/m] 0.01 0.00 0.01 0.00 0.00 0.00
Total glycerol [m/m] 0.08 0.06 0.17 0.06 0.22 0.03

Fatty acid methyl ester composition, mass %


Myristic C14:0 1.18 1.18 1.01 1.61 1.14 1.20
Myristoleic C14:1 0.96 0.78 0.77 1.04 0.62 0.73
Palmitic C16:0 27.72 27.10 26.61 37.57 29.47 27.95
Palmitoleic C16:1 4.60 4.76 4.75 6.37 4.42 4.74
Stearic C18:0 6.80 1.28 0.69 0.93 6.81 7.11
Oleic acid C18:1n9c 11.04 10.66 10.91 14.9 11.04 11.09
Elaidic acid C18:1n9t ND 6.03 6.22 8.48 ND ND
Linoleic C18:2n6c 19.60 19.48 19.92 26.05 19.25 19.81
α-Linolenic C18:3n3 27.12 27.24 27.92 0.49 26.86 27.36
Arachidic C20:0 0.37 0.37 0.38 0.54 ND ND
Total saturated 36.07 29.93 28.69 40.65 37.42 36.26
Total unsaturated 63.32 68.95 70.49 57.33 62.19 63.73

ND: Not detected.

of 2-MeTHF (process described in Fig. 1D) a fraction of residual water selective for extraction of triglycerides [9,20], which can be considered
was dissolved in the mixture of FAME with the fraction of 2-MeTHF that an advantage for extraction of biodiesel, because triglycerides not
was not evaporated, and it was only discovered after drying the bio- converted to methyl esters and present in the mixture can be separated
diesel in oven at 80 °C for 24 h. This water was only separated effi- more efficiently through the use of this solvent. Depending on the
ciently after the sample decantation process, which was achieved in a concentration in which they can be present in biodiesel, non-reacted
period of about 24 h, in which there was formation of two distinct glycerides can increase the viscosity of the fuel and, consequently, re-
phases: the heavier containing water and the lighter composed of bio- duce the efficiency of combustion, clogging the fuel filter and forming
diesel. This additional step may have resulted in significant loss of deposits in engine parts such as pistons, valves and nozzles [21]. Bio-
FAME, leading to a yield lower than that obtained with CPME. Differ- diesel purity levels obtained with purifications using 2-MeTHF were not
ently from the other solvents used, 2-MeTHF shows solubility in water as high as those obtained with the other solvents, 2-MeTHF is a little
of 14 g water/100 g 2-MeTHF (against 1.1 g water/100 g CPME) and a selective solvent due to its borderline solubility and may be used for
different characteristic in relation to other solvents with respect to so- extraction of proteins, fatty acids and other types of lipids that do not
lubility, due to being “inversely soluble” in water [17]. That is, its so- contain fatty acids such as waxes, phospholipids, carotenoids and other
lubility decreases as temperature increases, which facilitates its eva- pigments [8,21], in addition studies have shown that 2-MeTHF is an
poration during the drying process. Although the theoretical results excellent solvent for extraction of mono, di and triglycerides, which
were similar to those obtained experimentally, the HSP characteriza- compromises the purity of biodiesel [6,22].
tions should be treated with caution, because algae biodiesel presents The values obtained for biodiesel purity are in accordance with the
some differences in composition in relation to soy biodiesel, which standard EN 14214, which stipulates a minimum acceptable purity
results in distinct values for δD, δP and δH. In addition, the limited level of 96.5% in methyl esters [21].
availability of the Hansen solubility parameters prevents the complete
characterization of the solubility of a given solvent for the desirable 3.2. Influence of purification methods on biodiesel properties
products (FAME) and undesirable products (glycerides and glycerol).
Fig. 2 shows the purity of the biodiesel obtained for each technique To ensure the quality of biodiesel, quality standards should be es-
and solvent employed in this study. The purity of the biodiesel re- tablished, aiming to establish limits on contaminants that will not im-
presents the percentage of triglycerides converted to methyl esters. pair the quality of emissions from burning, as well as the performance,
Selectivity is a term used to represent the efficiency of a solvent in engine integrity, and safety in transport and handling [21].
extracting the product of interest from other undesirable components. The chemical and physicochemical characteristics of algae biodiesel
The two purification techniques produced biodiesel with same purity depend mainly on the production and purification process. In addition,
level, and purification with hexane showed better results. Studies using some properties of biodiesel are related with the molecular structures of
hexane as solvent for lipid extraction showed that this solvent is highly its constituent alkyl esters [21].
selective due to its highly hydrophobic nature and its limited solubility
in water (0.01 g water/100 g hexane); however, this solvent is very
soluble in glycerides, which is undesirable for biodiesel purification 3.2.1. Physicochemical properties
[9,18,19]. Extractions using CPME also resulted in a biodiesel with high The density of biodiesel is directly associated with the molecular
level of purity (Fig. 2). CPME is a solvent that is little soluble in water structure of its molecules [21]. Biodiesels obtained in extractions with
and very efficient for extraction of hydrophobic proteins [8], and little green solvents showed the same density (Table 2). While biodiesels
purified with hexane presented densities that were a little lower than

1127
S.S. de Jesus et al. Fuel 235 (2019) 1123–1130

those obtained with green solvents, probably because they present maximum value accepted in standard EN 14214 which is 120 g I2/
higher level of purity. The fact that the density does not vary sig- 100 g. Algae biodiesel presents in its composition (as well as its oils)
nificantly with the purification method and solvent shows that the high concentration of unsaturated fatty acids, mainly linoleic and li-
density of biodiesel depends mainly on the origin of the raw material nolenic, which results in a biodiesel with high iodine number [5,18].
[23]. Concerning quality standards, the ASTM standard does not con- Water, in addition to promoting the hydrolysis of biodiesel resulting
sider it relevant, whereas the European standard EN 14214 establishes in free fatty acids, is also associated with the proliferation of micro-
density values from 0.86 to 0.90 g/L. Biodiesels obtained in this work organisms, corrosion in storage tanks with sediment deposition. Since
presented densities lower than 0.90 g/L, being compliant with the biodiesel presents a certain degree of hydroscopicity, water content
European quality standards. should be monitored during storage [21]. Both the European standard
Biodiesel viscosity is another important factor that should be taken (EN 14214) and the North American standard (ASTM D6751) establish
into consideration, because high viscosity causes heterogeneity in bio- a maximum limit of 0.05% (i.e., 500 mg/kg) of water in biodiesel. The
diesel combustion, due to decreased efficiency of atomization in the percentage of water obtained in this study ranged from 0.03 to 0.09%.
combustion chamber, causing the deposition of residues in the internal Table 2 shows that the moisture content in the biodiesel purified by
parts of the engine [21]. Therefore, these contaminants can be mon- liquid-liquid extraction with 2-MeTHF was higher than 0.05%. The
itored indirectly through the determination of kinematic viscosity at difficulty to separate the water after the extraction process, as men-
40 °C. Standard EN 14214 (analytical method EN ISO 3104) establishes tioned in Section 3.1, had great influence on the high content of
an acceptable range of 3.5–5.0 mm2/s for viscosity, while standard moisture, thus being noncompliant with the quality standards of EN
ASTM D6751 (analytical method D445) allows a little wider range of 14214 and ASTM D6751. According to Predojević [23], the use of an-
1.9–6.0 mm2/s. Table 2 shows that the biodiesel obtained for each si- hydrous sodium sulfate to remove residual water, after the biodiesel
tuation studied is compliant with the standards EN 14214 and ASTM washing step, is not a very effective method, which results in a high
D6751. However, we can observe that the purification with green sol- water content in the purified biodiesel.
vents resulted in a more viscous biodiesel, especially when purified The saponification value represents the milligrams of potassium
with 2-MeTHF. Table 2 shows that purification by column chromato- hydroxide required for saponification of one gram of fat or oil [23].
graphy using 2-MeTHF as eluent produced a biodiesel with higher This value is inversely proportional to the mean molecular weight of the
viscosity (4.28 ± 0.06 mm2/s) mainly due to its composition pre- fatty acids of the present triglycerides. This value is important to de-
senting a high percentage of non-reacted glycerides (mono-, di- and monstrate the presence of oils and fats of high proportion of fatty acids
triglycerides) and a higher percentage of impurities (Fig. 2) compared and low molecular mass, in mixtures with other oils and fats. No
with the other solvents used in this study. On the other hand, we can standard establishes the minimum or maximum limits for saponifica-
observe that in the liquid-liquid extraction the viscosity of the biodiesel tion value, for this reason, we hardly find relevant studies with these
obtained with CPME was very close to 2-MeTHF; the fact that CPME values for biodiesel. In this work the saponification value was de-
presents in its composition a high percentage of glycerides (data termined in order to calculate the cetane index, which was calculated
available in Table 2) may have been the most significant factor for the according to the methodology of Krisnangkura [12]. Table 2 shows that
viscosity being relatively equal to that of the biodiesel purified with 2- the saponification value ranged from 197 to 203 mg KOH/g. Higher
MeTHF. saponification values mean that the biodiesel is composed mostly of
fatty acids of lower molecular mass. However, the correlation of the
3.2.2. Chemical properties saponification value with the quality of biodiesel is not yet clear.
Monitoring the acidity in biodiesel is of great importance during The cetane number indicates the time of delay in fuel ignition for
storage, and changes in values in this period could mean the presence of diesel cycle engines, therefore, it reflects the quality of fuel ignition. A
water [21]. Acidity is measured in terms of quantity of KOH needed to higher cetane number means a shorter ignition delay time and more
neutralize the sample [23]. The method recommended by EN 14214 is complete combustion of the fuel charge in the combustion chamber.
EN 14104, and ASTM recommends the potentiometric method D664 The cetane number increases with the length of the unbranched carbon
[21,23]. All standards described above established maximum limits for chain, thus the hexadecane (cetane) is established as default value of
acidity of 0.5 mg KOH/g. Table 2 shows that the acidity ranged from 100 in the range of cetane number, while for 2,2,4,4,6,8,8-hepta-
0.37 to 0.45 mg KOH/g, thus being compliant with the limits estab- methylnonane (isocetane) is assigned the value 15 [21]. Compared with
lished by the standards EN 14214 and ASTM D6751. It can be observed fossil diesel, biodiesel provides higher values of cetane number. In
that the purification of FAMEs by column chromatography resulted in a Europe, both for diesel and biodiesel, the minimum acceptable cetane
biodiesel with lower acidity. Predojević [23] observed that when using number is fixed at 51 (method EN ISO 5165). In the North American
silica gel to purify biodiesel from waste frying oils the acidity decreased quality standard, the cetane number for diesel is 40, while for biodiesel
by 65%, and this method was more effective than using warm distilled the cetane number is established at 47 (method D613). In this work, we
water or 5% w/w phosphoric acid. Biodiesel purified using 2-MeTHF calculated the cetane index, as an alternative to the cetane number. The
presented higher levels of acidity, which leads us to conclude that in its cetane index is also an indicator of quality for fuel ignition and is re-
composition there is a higher percentage of free fatty acids, compared lated with the time between the injection of fuel into the cylinder and
with the other solvents used. the start of ignition [24]. The methodology described by Krisnangkura
The number of unsaturations not only has effect on the values of [12], in which the cetane index (CI) is estimated, is based on saponi-
density and viscosity of biodiesel, but is also of great importance in fication values and iodine numbers, being used only for methyl esters.
oxidative stability [21]. Standard EN 14214 adopted the iodine number Table 2 shows that the cetane index was higher than 47, except for the
to determine the number of unsaturations, while North American biodiesel obtained through purification by liquid-liquid extraction with
standard ASTM D6751 does not mention the maximum allowed value. 2-MeTHF, thus being noncompliant with both EN 14214 and ASTM
The iodine number depends on the origin of the raw material and has D675. The biodiesel obtained through the process of purification by
major influence on the tendency for oxidation of the fuel. Although the column chromatography using as eluent hexane:ethyl acetate (19:1)
official analytical method of standard EN 14214 is EN ISO 1411, there was the only one that met the two quality standards. However, it is
are different methods to determine the iodine number; in this work we worth noting, the results were calculated for the cetane index, through
used the official method adopted by the AOCS Cd 1c-85. As observed in the equation of Krisnangkura who used data from vegetable oils [12].
Table 2 the iodine numbers are in the range of 74–127 g I2/100 g, Glycerol is a by-product from the reaction of transesterification of
purification through liquid-liquid extraction provided higher iodine oils and fats. The determination of residual glycerol serves as parameter
numbers, and extractions with green solvents were higher than the to evaluate the efficiency of the process of purification of biodiesel.

1128
S.S. de Jesus et al. Fuel 235 (2019) 1123–1130

Table 3 the presence of C18:1n9t was not observed. In the purification by


. Solvent properties of hexane, 2-MeTHF and CPME. column chromatography, it was observed that, using hexane:ethyl
Solvent Hexane 2-MeTHF CPME acetate, the biodiesel presented higher percentages of C16:0 and C18:1
than other solvents studied, in addition to a significant percentage of
Price ($/kg) 0.5 5.0 11.0 C18:1n9t, which was not observed in the purification by liquid-liquid
Toxicity indication 6 4 4
extraction. However, there was a decrease in the percentage of C18:0
Boiling point [°C] 69 80 106
Solubility of solvent in water (23 °C) [g/100 g] 0.01 14 1.1
and C18:3n3, these being lower than 1%. Table 2 shows that the bio-
Solubility of water in solvent (23 °C) [g/100 g] 0.01 4.4 0.3 diesel presented mainly the methyl esters C16:0, C18:1n9c, C18:2n6c
Azeotropic point with water [°C] – 71a 83b and C18:3n3 (except in purification by column chromatography with
Evaporate 1 kg solvent [kWh] 0.121 0.126 0.132 hexane:ethyl acetate). Studies of in situ transesterification using Chlor-
Masse CO2 (g) generated by kg of solvent 96 101 106
ella vulgaris obtained a biodiesel with higher percentages of C16:1,
Log P 3.94 0.82 1.59
C18:0, C18:1n9c and C18:3n3 [5], on the other hand, in studies with
Azeotrope (wt%). wet biomass of C. vulgaris and under subcritical conditions, the com-
Adapted from Refs. [20,22,26]. position of the FAME was mostly composed of C16:0, C16:1, C18:2n6c
a
Composition = 2-MeTHF: 89.4 H2O:10.6. and C18:3n3 [25]. Another study, using the microalgae C. pyrenoidosa,
b
Composition = CPME: 83.7 H2O:16.3. obtained higher percentages of C16:0, C18:2n6c and C18:3n3 [18]. The
composition of the FAMEs may vary according to the methodology
High concentrations of glycerol in biodiesel cause storage problems, employed and the conditions during the transesterification reaction, as
because when biodiesel is mixed with petroleum diesel there is se- well as to the methodology used for purification. However, microalgae
paration of glycerol in storage tanks [21]. Problems such as formation biodiesel is composed mainly of methyl esters C16:0, C16:1, C18:1n9c,
of deposits, clogging of engine nozzles and emissions of aldehydes are C18:2n6c and C18:3n3.
also related with the high concentration of glycerol in biodiesel [21]. Finally, we can observe that all methods employed produced a
Residual free glycerol can be easily removed by washing biodiesel. biodiesel with a percentage higher than 50% of long-chain methyl es-
Although it is practically insoluble in biodiesel, glycerol can be found ters of fatty acids with low degree of unsaturation (C14:0, C16:0, C16:1
dispersed in the form of droplets. The presence of residual soaps can and C:18:1), which produces a biodiesel with better quality.
interfere, increasing the concentration of glycerol in biodiesel due to Table 3 presents the main characteristics of the solvents used in this
the formation of emulsions [21]. In Europe and in the United States, the study. Price, boiling point, partition coefficient (log P), toxicity cate-
maximum permitted content of free glycerol in biodiesel is 0.02%mass gory, and energy required for solvent evaporation are very important
and its determination is performed through the chromatographic parameters for the choice of extraction solvent in the industry [22].
methods EN ISO 14105 and EN ISO 14106, established by the European Table 3 shows that green solvents present great potential to replace
standard, and ASTM D6584, by the North American standard. Table 2 hexane in extraction processes, because they present boiling point
shows the absence of free glycerol in the biodiesel obtained for each lower than 130 °C, which is ideal for industrial use, moreover, they
situation studied. consume only 5%–10% more energy and produce only 5%–10% more
Mono-, di- and triglycerides result from the incomplete reaction of carbon dioxide than hexane [22]. In addition, these three solvents are
glycerides, thus this is an important parameter that can be used to lipophilic, with Log P > 0. However, although these solvents present
evaluate the efficiency of the conversion of oils and fats into biodiesel. high potential for the extraction and purification of biodiesel, their
Depending on their concentration in biodiesel, non-reacted glycerides prices are still prohibitive; while 1 kg of hexane costs $0.5, 2-MeTHF
can increase the viscosity of the fuel [21]. The European standard sets and CPME are 10 and 22 times more expensive. These values can also
maximum limits of 0.8, 0.2 and 0.2%mass for mono-, di- and triglycer- mean a great barrier in the large-scale use of these solvents.
ides respectively. While the ASTM standard does not stipulate values for However, this technique shows a few advantages compared to other
these glycerides. Total glycerol can be calculated through the con- biodiesel purification methods that have been recently developed, not
centrations of free glycerol and mono-, di- and triglycerides. The Eur- only in avoiding the use of fossil-based solvents, but also reducing costs.
opean standard establishes a maximum limit of 0.25%mass, while in the In comparison with techniques currently available, purification with
United States the limit is 0.24%mass. Table 2 shows that all purified deionized water shows several drawbacks such as high water con-
biodiesel presented in their composition low concentrations of mono- sumption, which generates large quantities of waste water and high
glycerides; however, in all situations studied we observed limits lower production costs, such as deionization of water. Also, biodiesel obtained
than 0.8%mass. On the other hand, the biodiesel purified by liquid-liquid has high moisture content, not meeting the quality standards [27,28].
extraction with hexane and with CPME, in addition to the biodiesel Although the membrane process generates few residues, the energy
purified by column chromatography using 2-MeTHF as eluent, pre- expenditure and the cost of equipment installation are barriers to this
sented a percentage of diglycerides above that established by the Eur- technique use on a large scale [28]. Purification by dry washing results
opean standard. The presence of mono- and diglycerides mean that not in biodiesel with a significant amount of methanol that does not meet
all glyceride present in the sample was reacted and that they were the standard EN 14214 [29], in addition to the intensive use of energy.
formed as intermediates of the esterification reaction. The presence of Also, the purification method Magnesol among other absorbers, de-
these glycerides is often caused by the experimental conditions, such as mands further costs with absorber, which is not regenerated after the
molar ratio of reactants and reaction time, thus it is not a problem purification process [30,31]. Thus, the purification process developed
caused mainly by the extraction method. In this study, significant in this work is more attractive than those currently used.
amounts of triglycerides were not observed. Total glycerol concentra- It should be noted that although the price of green solvents is still
tion was lower than that established by the standards EN 14214 and prohibitive, once they are not yet produced on a large scale, the de-
ASTM D6751. mand is still small because it is a recent technology, still under devel-
The FAMEs obtained in this study are presented in Table 2. As can opment. Energy expenditure for solvent evaporation is required only in
be observed, the biodiesel purified with hexane presented a different the final stage of the process, in the drying step, where biodiesel is
composition compared with those obtained with 2-MeTHF and CPME. heated to 75 °C and an exact amount of absorbent is added to the bio-
Depending on the purification technique the composition of the FAMEs diesel and left for a period until later filtration [30]. Fixed-bed pur-
varied significantly. Purification by liquid-liquid extraction using ification techniques also show a great disadvantage in relation to the
hexane as solvent resulted in a biodiesel with higher percentage of purification proposed in this work, because the biodiesel separation is
C18:0 than that obtained with 2-MeTHF and CPME; on the other hand, carried out in a column that is kept under heating and requires vacuum

1129
S.S. de Jesus et al. Fuel 235 (2019) 1123–1130

application for the purification to occur [27,28,30]. Chlorella vulgaris. Fuel 2012;94:544–50.
[6] Sicaire A-G, Vian M, Fine F, Joffre F, Carré P, Tostain S, et al. Alternative bio-based
solvents for extraction of fat and oils: solubility prediction, global yield, extraction
4. Conclusion kinetics, chemical composition and cost of manufacturing. Int J Mol Sci
2015;16:8430–53.
Algae biodiesel purification using green solvents was evaluated as to [7] Henderson RK, Jiménez-González C, Constable DJC, Alston SR, Inglis GGA, Fisher
G, et al. Expanding GSK's solvent selection guide – embedding sustainability into
its physicochemical characteristics and chemical composition. Biodiesel solvent selection starting at medicinal chemistry. Green Chem 2011;13:854–62.
produced through acid-catalyzed in situ transesterification was purified [8] Tenne S-J, Kinzel J, Arlt M, Sibilla F, Bocola M, Schwaneberg U. 2-
through liquid-liquid extraction and column chromatography, using the Methyltetrahydrofuran and cyclopentylmethylether: two green solvents for efficient
purification of membrane proteins like FhuA. J Chromatogr B 2013;937:13–7.
green solvents: 2-methyltetrahydrofuran (2-MeTHF) and cyclopentyl [9] Probst KV, Wales MD, Rezac ME, Vadlani PV. Evaluation of green solvents: oil ex-
methyl ether (CPME). The solvent hexane was used for comparison. traction from oleaginous yeast Lipomyces starkeyi using cyclopentyl methyl ether
According to the results it was possible to conclude that the biodiesel (CPME). Biotechnol Progress 2017;33:1096–103.
[10] AOAC. Official methods of analysis of the association of official analytical chemists.
purified with green solvents showed results that were similar those of
In: Horwitz W, editor. Method 989.05. 15th ed.Arlington: Association of Official
biodiesel purified with hexane. However, it was observed that in the Analytical Chemists; 2009.
process of purification by liquid-liquid extraction using 2-MeTHF there [11] Geris R, dos Santos NAC, Amaral BA, de Souza Maia I, Castro VD, Carvalho JRM.
was need for an additional step of purification, which consisted in Biodiesel from soybean oil – experimental procedure of transesterification for or-
ganic chemistry laboratories. Quim Nova 2007;30:1369–73. [in Portuguese].
leaving the material in decantation for 24 h to remove residual water [12] Krisnangkura K. A simple method for estimation of cetane index of vegetable oil
contained in the biodiesel. This additional step can be a limiting factor methyl esters. JAOCS 1986;3:552–3.
for the use of this solvent in the liquid-liquid extraction process. [13] Hansen CM. Hansen solubility parameters: a user's handbook. 2nd ed. Boca Raton:
CRC Press; 2007.
Although biodiesels purified with green solvents present higher visc- [14] Halim R, Danquah MK, Webley PA. Extraction of oil from microalgae for biodiesel
osities and densities, they are all compliant with the quality standards. production: a review. Biotechnol Adv 2012;30:709–32.
Concerning purity, some specific improvements in the purification [15] Batista MM, Guirardello R, Krähenbühl MA. Determination of the solubility para-
meters of biodiesel from vegetable oils. Energy Fuels 2013:7497–509.
process, especially when using 2-MeTHF as solvent, can be im- [16] Damergi E, Schwitzguébe J-P, Refardt D, Sharma S, Holliger C, Ludwi C. Extraction
plemented so biodiesel purity is equal to or greater than that obtained of carotenoids from Chlorella vulgaris using green solvents and syngas production
using hexane. Some chemical properties obtained in this study may be from residual biomass. Algal Res 2017;25:488–95.
[17] He M, Zhou D-Q, Ge H-L, Huang M-Y, Jiang Y-Y. Caytalytic behavior of wool-Rh
related to the methodology employed for in situ transesterification re- complex in asymmetric hydrogenation of 2-methyl furan. Polym Adv Technol
action, of which we can mention the concentration above 0.2%mass of 2003;14:273–7.
diglycerides in biodiesel purified through liquid-liquid extraction with [18] D’Oca MGM, Viêgas CV, Lemões JS, Miyasaki EK, Morón-Villarreyes J, Primel EG,
et al. Production of FAMEs from several microalgal lipidic extracts and direct
hexane and CPME, and by column chromatography using the eluent 2-
transesterification of the Chlorella pyrenoidosa. Biomass Bioenergy
MeTHF. Finally, we can conclude that 2-MeTHF and mainly CPME used 2011;35:1533–8.
as eluent in purification by column chromatography have great po- [19] Wahlen BD, Willis RM, Seefeldt LC. Biodiesel production by simultaneous extrac-
tential to replace hexane in the biodiesel purification process, and the tion and conversion of total lipids from microalgae, cyanobacteria, and wild mixed-
cultures. Bioresour Technol 2011;102:2724–30.
price of these solvents is the only factor limiting their large-scale use. [20] Single Probst K. cell oil production using Lipomyces starkeyi: fermentation, lipid
analysis and use of renewable hemicellulose-rich feedstocks [PhD thesis] Kansas
Acknowledgments State University; 2014.
[21] Lôbo IP, Ferreira SLC, da Cruz RS. Biodiesel: quality parameters and analytical
methods. Quim Nova 2009;32:1596–608. [in Portuguese].
This study was supported by the Coordination for Improvement of [22] Breil C, Meullemiestre A, Vian M, Chemat F. Bio-based solvents for green extraction
Higher Education Personnel (PNPD-CAPES), the State of São Paulo’s of lipids from oleaginous yeast biomass for sustainable aviation biofuel. Molecules
2016;21:1–14.
Research Foundation (FAPESP, grant process no. 2010/03764-3 and [23] Predojević AJ. The production of biodiesel from waste frying oils: a comparison of
2015/19609-0), National Council of Technological and Scientific different purification steps. Fuel 2008;87:3522–8.
Development (CNPq, grant process no. 166844/2017-9) and Espaço da [24] Knothe G. Depedence of biodiesel fuel properties on the structure of fatty acid alkyl
esters. Fuel Process Technol 2005;86:1059–70.
Escrita (Coordenadoria Geral UNICAMP) for the language services
[25] Tsigie YA, Huynh LH, Ismadji S, Engida AM, Ju Y-H. In situ biodiesel production
provided. from wet Chlorella vulgaris under subcritical condition. Chem Eng J
2012;213:104–8.
[26] Specialty chemicals – cyclopentyl methyl ether (CPME). http://www.zeon.co.jp/
References
business_e/enterprise/spechemi/spechemi5-13.html; 2018 [accessed 21 April
2018].
[1] Milano J, Ong HC, Masjuki HH, Chong WT, Lam MK, Loh PK, et al. Microalgae [27] Atadashi IM, Aroua MK, Aziz AA. Biodiesel separation and purification: a review.
biofuels as an alternative to fossil fuel for power generation. Renewable Sustainable Renewable Energy 2011;36:437–43.
Energy Rev 2016;58:180–97. [28] Bateni H, Saraeian A, Able C. A comprehensive review on biodiesel purification and
[2] Demirbas A, Fatih Demirbas MF. Importance of algae oil as a source of biodiesel. upgrading. Biofuel Res J 2017;15:668–90.
Energy Convers Manage 2011;52:163–70. [29] Atadashi IM. Purification of crude biodiesel using dry washing and membrane
[3] Lewis T, Nichols PD, McMeekin TA. Evaluation of extraction methods for recovery technologies. Alexandria Eng J 2015;54:1265–72.
of fatty acids from lipid producing microheterotrophs. J Microbiol Methods [30] Faccini CS, Cunha MED, Moraes MSA, Krause LC, Manique MC, Rodrigues MRA,
2000;43:107–16. et al. Dry washing in biodiesel purification: a comparative study of adsorbents. J
[4] Ehimen EA, Sun ZF, Carrington CG. Variables affecting the in situ transesterification Braz Chem Soc 2011;22:558–63.
of microalgae lipids. Fuel 2010;89:677–84. [31] Manique MC, Faccini CS, Onorevoli B, Benvenutti EV, Caramão EB. Rice husk ash as
[5] Velasquez-Orta SB, Lee JGM, Harvey A. Alkaline in situ transesterification of an adsorbent for purifying biodiesel from waste frying oil. Fuel 2012;92:56–61.

1130

You might also like