Fuel 212 (2018) 132–139

Contents lists available at ScienceDirect

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

Full Length Article

Optimization of direct liquid-liquid extraction of lipids from wet urban MARK

sewage sludge for biodiesel production
⁎
Cécile Kecha, , Anne Galloya, Christophe Frippiata, Albert Pielb, Damien Garotc
a
Cellule de Chimie Organique, Direction des Laboratoires d’Analyse, Institut Scientifique de Service Public, Rue du Chéra 200, 4000 Liège, Belgium
b
Direction des Technologies Environnementales, Institut Scientifique de Service Public, Rue du Chéra 200, 4000 Liège, Belgium
c
Centre de Recherches Métallurgiques, Avenue du Bois Saint-Jean 21, B27, Quartier Polytech 4, 4000 Liège, Belgium

A R T I C L E I N F O A B S T R A C T

Keywords: Efficient lipid extraction is one of the main remaining challenges for full-scale production of biodiesel from
Sewage sludge urban sewage sludge, a low-cost feedstock. The common approach is to extract lipids from dewatered or dry
Liquid-liquid extraction sludge by means of organic solvents. However, sludge dewatering and drying are energy-consuming and re-
Lipids present more than 50% of the total production cost. Direct liquid-liquid extraction of lipids has been poorly
Biodiesel
explored; only extraction using hexane as solvent is described in the literature. This study focuses on the opti-
mization of the extraction of lipids from wet raw urban sewage sludge at room temperature. This optimization is
conducted varying solvent, pre-treatment, number of extraction steps, contact time and ratio sludge to solvent.
The optimized method gives larger lipid amount (34.5%wt dry matter (DM)) than using hexane extraction
(32.8%wt DM). Finally, the potential of lipid fractions extracted from eight sludge samples for biodiesel pro-
duction was evaluated through the determination of acid value and free fatty acid (FFA) content, saponification
number and fatty acid composition.

1. Introduction for the production of biodiesel is thus twofold: it is in line with EU

The third target of the 2020 climate change and energy package
adopted by the European Union (EU) is to reach a share of 20% of
renewable fuels in the gross final consumption of energy in the EU in
2020. In that context, the Directive 2009/28/CE establishes a manda-
tory target of at least 10% of renewable fuel in the transport sector in
2020 [1]. Biodiesel is one of the most promising renewable fuels. It
mainly consists of a mixture of fatty acid alkyl esters obtained through
esterification/transesterification of lipid sources with alcohol and in the
presence of a catalyst. At present, the main obstacle to a widespread
production and commercialization of biodiesel is the need to use costly
edible oil (like soybean, rapeseed, sunflower, palm, cottonseed and
peanuts oil) as feedstock, representing between 70% and 85% of the
total production cost. Furthermore, the cultivation of crop-for-fuel
competes with human food production [2–4].
Urban sewage sludge generated in wastewater treatment plants
(WTTPs) is a low-cost lipid feedstock available in large quantities. In
Belgium, Walloon urban WWTPs produced around 49,000 tons dry
matter (DM) of sludge in 2014 [5]. Moreover, sludge management re-
presents a major cost in WWTPs operation (between 50% and 60% of
the total operating costs). The advantage of using urban sewage sludge
Directive 2009/28/CE and its associated targets on renewable fuels,
and it provides new valorisation options for urban sewage sludge
generated as waste in WTTPs.
The main challenge associated to the production of biodiesel from
sludge is an efficient lipid extraction. Sludge water content can account
for up to 95–98%wt. Dewatering and drying involve high energy inputs,
potentially representing more than 50% of the total biodiesel produc-
tion costs [6]. In addition, water removal by thermal drying results in
the loss of valuable organic compounds. It probably causes the loss of
lipids, decreasing biodiesel production yield [7]. Hence, other extrac-
tion methods need to be identified in order to be able to efficiently
produce biodiesel from wet sludge.
The common approach for extracting lipids from sludge is to use
organic solvents, usually in mixtures. The Bligh and Dyer method
(chloroform/methanol/water) is the most popular method for lipid
extraction from several materials [8]. This method uses the properties
of ternary mixtures: when solvents are mixed in an appropriate ratio,
they form a monophasic solution improving contact between lipids and
the organic solvent where lipids are soluble. Addition of more water
and/or chloroform produces a biphasic system from which it is easy to
isolate the lipids-containing chloroform layer. The advantage of the

⁎
Corresponding author.
E-mail address: c.kech@issep.be (C. Kech).

http://dx.doi.org/10.1016/j.fuel.2017.10.010
Received 5 April 2017; Received in revised form 21 June 2017; Accepted 5 October 2017
0016-2361/ © 2017 Elsevier Ltd. All rights reserved.
C. Kech et al.
Bligh and Dyer method is the presence of water, avoiding the need for
prior sludge drying. However, the maximum ratio of water allowed in
the sludge in order to reach a monophasic solution is 80%wt, implying
that the sludge has to be at least partially dewatered [3]. In addition,
the use of a chlorinated solvent results in the production of large
amounts of chemical hazardous waste that is highly toxic to the en-
vironment.
Smedes and co-workers studied the possibility of replacing chloro-
form by a non-chlorinated solvent for the determination of total lipid in
marine tissues [9]. They tried several solvent combinations and the
mixture of cyclohexane/isopropyl alcohol/water (Smedes mixture)
gave the best results. Pastore and co-workers describe the extraction of
lipids from a dewatered primary sludge (dry matter content (DMC)
14.4%wt) with hexane (ratio 1:1) under acid media (sulphuric acid) at
38 °C for 1 h (three consecutive extractions). These conditions gave a
lipid yield of 20.0%wt (on the basis of DM) [10]. Only one reference
reports the direct liquid-liquid extraction of lipids from wet primary
sludge (DMC 3.4%wt) by mean of hexane [7]. The extraction under acid
media (hydrochloric acid) at room temperature for 1 h using a 2:1
sludge/hexane ratio (three consecutive extractions) gave 26.7%wt of
lipids (on the basis of DM). The advantage of this method is that sludge
drying or dewatering was unnecessary. Other solvents were not tested.
Extraction from sludge may be influenced by many variables such as
sludge composition, type and amount of solvent, extraction time,
temperature, stirring rate, type of micro-organisms present in the
sludge, etc. [11]. As a preliminary research for the development of a
feasible and economical biodiesel production method using wet sewage
sludge, this paper aims at improving the liquid-liquid extraction
method at room temperature proposed by Olkiewicz and co-workers, as
described above [7]. Different solvents or combination of solvents for
lipid extraction from a wet sludge (DMC < 5%wt) were explored, and
the method was optimized varying (i) sludge pre-treatment, (ii) number
of consecutive extraction steps, (iii) contact time, and (iv) sludge to
solvent ratio. In a second step, the most favourable conditions were
applied to compare lipid content before and after sludge drying. In a
third step, the lipid content of eight sludge samples (four primary
sludge samples and four secondary sludge samples) was determined.
Finally, in order to verify the potential of the extracted lipids for bio-
diesel conversion, the lipid fractions obtained from all eight sludge
samples were characterized for acid value and free fatty acid (FFA)
content, saponification number, and fatty acid composition.
It must be noted that lipid content is not strictly measured in this
study. Instead, groups of substances with similar physical character-
istics and that are not volatilized during the test, are determined
quantitatively on the basis of their common solubility in the organic
solvent used for extraction. They should thus be referred to as “ex-
tractable oil and grease”. However, for the sake of brevity, the term
“lipid content” will be used throughout this paper.
2. Materials and methods
2.1. Reagents and instruments
2.1.1. Reagents
Experiments were conducted using methanol, ethanol, isopropyl
alcohol, hexane, cyclohexane, diethyl ether, tert-butyl methyl ether,
standardized potassium hydroxide 0.1 N in methanol, standardized
hydrochloric acid 0.1 N, sodium methoxide 10% in methanol, fuming
hydrochloric acid, sulphuric acid, phenolphthalein, thymolphthalein,
anhydrous sodium sulphate and sodium chloride (all of analytical-re-
agent grade).
The standard used for calibration of Karl Fisher coulometric titrator
was supplied by Riedel-de Haën (Hydranal®-Water Standard 1.00).
The standard used for identification of fatty acid methyl esters was
supplied by Supelco (37 Component FAME Mix).
All chemicals were used as received.
133
Fuel 212 (2018) 132–139
Water was obtained with a MilliQ Advantage A10 (Millipore).
2.1.2. Instruments
Three-dimensional shaker mixer T2C (WAB), Multi Reax Vortex
Mixer (Heidolph™), 870 KF Titrino plus (Metrohm), 8510 Ultrasonic
Cleaner MTH (Branson), centrifuge BR4-I (Jouan), automated eva-
poration system TurboVap®II (Zymark), oven EU53 (Jouan), oven
FD240 (Binder), analog dry block heater (VWR®), ThermoQuest Trace
GC 2000 equipped with a Trace MS analyzer (Interscience), BP5-MS
capillary column (30 m × 0.25 mm I.D., 0.25 µm film thickness, SGE
Analytical Science), Software Xcalibur (Thermo Scientific).
2.2. Sludge sampling and handling
Four primary and four secondary sludge samples were collected
before (raw sludge) or after partial dewatering at the urban WWTP
located in Wegnez (Liège, Belgium, population equivalent of 110,000)
from August 2015 to October 2016. This facility uses an activated
sludge process that produces three sludge types: greasy, primary and
secondary sludges.
Sludges were poured into 5000 mL plastic buckets, transported in a
refrigerated car to the laboratory and stored at 4 °C prior to sub-
sampling (maximum storage time 2 days). The manual subsampling was
realized in 250, 500 and 1000 mL plastic bottles (HDPE). The homo-
geneity of the subsampling was checked by comparing lipid content
values obtained from four 250 mL bottles of the same primary sludge
sample using the optimized method described in this paper
(RSD = 9.7%).
All the bottles were frozen to prevent degradation of biological
material and modification of sludge properties. The bottles were un-
frozen one at a time at 4 °C for direct analysis avoiding long time of
conservation (maximum shelf life: 15 days at 4 °C).
2.3. Sludge characterization
Prior to optimization of the liquid-liquid extraction, sludge samples
were analyzed in order to determine DMC and water content. After
optimization, the liquid-liquid extraction method was used to de-
termine lipid content of the sludge samples.
2.3.1. Dry matter content (DMC)
DMC was determined according to a drying method described in ISO
11465 [12]. The results were expressed as weight of residual solid after
thermal treatment per unit weight of initial wet sludge.
2.3.2. Water content
Water content WS (%) was determined according to Karl-Fisher ti-
tration (external dilution method). Each sample was measured in tri-
plicate.
2.3.3. Lipid content
The initial sludge sample must contain maximum 5% of lipids be-
cause high lipid concentrations may affect the nature of the organic
phase and have a measurable effect on the result [9]. 5 g of wet sludge
previously treated 30 min in an ultrasonic bath (in order to improve
extraction efficiency), was weighed in a 50 mL PE centrifugation tube
and acidified until pH 2 by the addition of 1 mL of fuming hydrochloric
acid. The mixture was then shaken for 2 min at room temperature
(1900 rpm). 8 mL of isopropyl alcohol and 10 mL of cyclohexane (ratio
of 1:1.6:2) were added and accounting for the amount of water in the
sample, VWATER mL of water was added to obtain a total of 11 g of
water. VWATER is calculated with Eq. (1):
VWATER = 11−mHCl−(mS × WS)/100 (1)
where mS (g) is the mass of the sludge sample, mHCl (g) is the mass of
hydrochloric acid (≅hydrochloric acid volume (mL)) and WS (%) is the
C. Kech et al.
sludge water content.
The mixture was shaken during 60 min at room temperature
(1900 rpm). The layers were separated by centrifugation (10 min at
3000 rpm) and the upper organic layer was transferred into a flask.
For the second and the third extraction, 10 mL of a mixture of cy-
clohexane containing 13%wt isopropyl alcohol were added to the
sludge and shaken during 60 min (1900 rpm), centrifuged (10 min;
3000 rpm) and the upper organic layers were combined with the one of
the first extraction. The organic layers were dried on anhydrous sodium
sulphate and evaporated at 35 °C under a stream of nitrogen to a vo-
lume of 1 mL. The residue was transferred quantitatively into a weighed
wide-mouth cup. The residual solvent was evaporated to dryness under
a stream of nitrogen, and the cup was placed in an oven at 105 °C for at
least 2 h. The dry residue weight was recorded after cooling in a de-
siccator. The result was expressed as weight of lipids per unit weight of
wet sludge and per unit weight of DM.
2.4. Lipid characterization
The lipid fractions extracted from the eight sludge samples were
characterized for acid value and FFA content, saponification number
and fatty acid composition.
2.4.1. Acid value and FFA content
Acid value wAV (mg/g) and FFA content wFFA (%wt) were de-
termined via a colorimetric titration method described in NBN EN ISO
660 [13].
Due to the predominance of palmitic acid in the oil, wFFA was ex-
pressed as equivalent to palmitic acid (molecular weight of 256).
2.4.2. Saponification number (Is)
The saponification number was determined using a colorimetric
back titration method described in NBN EN ISO 3657 [14] with some
modifications. About 250 mg of sample dissolved into 1 mL of diethyl
ether were neutralized by addition of exactly 20 mL of standardized
potassium hydroxide 0.1 N in methanol. The solution was heated at
85 °C during 4 h in a closed vial. After cooling, the solution was brought
to 25 mL with methanol. 10 mL of this solution were diluted by 200 mL
of 9:9:2 diethyl ether/ethanol/water medium and titrated by standar-
dized hydrochloric acid 0.1 N and thymolphthalein as indicator. The
endpoint corresponds to a colour change from green to yellow (original
colour of the sample into the medium). The same procedure was ap-
plied to a blank solution to determine the volume of potassium hy-
droxide consumption for the saponification of the sample.
The saponification number is the mass of potassium hydroxide in
milligrams that is required to saponify 1 g of sample. It is a di-
mensionless number calculated by means of Eq. (2).
IS = ((V0−V1) × CHCl × 56.1)/m (2)
where V0 (mL) is the volume of hydrochloric acid needed for the blank
neutralization, V1 (mL) is the volume of hydrochloric acid needed for
the sample neutralization, CHCl (mol/L) is the certified concentration of
the hydrochloric acid solution, 56.1 is the molecular weight of po-
tassium hydroxide and m (g) is the mass of sample test portion.
2.4.3. Fatty acid composition of sludge
The fatty acid composition of the extracted lipids was determined by
gas chromatography after derivatization through basic and acid cata-
lysis esterification/transesterification: esterification of FFAs to obtain
fatty acid methyl esters, and transesterification of triglycerides and
phospholipids.
2.4.3.1. Fatty acids derivatization method. About 15 mg of extracted
lipids were dissolved into 1 mL of tert-butyl methyl ether into a 10 mL
glass tube. 500 µL of sodium methoxide 10% solution in methanol were
added. A blank solution was prepared in the same way with 1 drop of
134
Fuel 212 (2018) 132–139
phenolphthalein to control the pH value of the solution during the
derivatization.
The tubes were heated 20 min in a dry block heater at 80 °C. After
cooling, sulphuric acid 1 M in methanol was added gradually to the
blank tube until phenolphthalein becomes colourless. The same volume
plus 200 µL was added to the other tube. The tubes were heated 30 min
in a heating block at 80 °C.
After cooling, 2 mL of saturated sodium chloride in water were
added and the fatty acid methyl esters were extracted with 2 mL of
hexane. The hexane phases were collected and dried by passing through
anhydrous sodium sulphate cartridges. The cartridges were rinsed with
1 mL of hexane and the final solutions were diluted two times before
gas chromatography analysis.
2.4.3.2. Fatty acid methyl ester gas chromatography analysis. The fatty
acid composition was determined using a gas chromatography coupled
to a mass spectrometry analyzer (GC-MS). Separation was achieved
using helium as a carrier gas (constant flow 1 mL/min). The injection
volume of the sample was 2 µL with a split ratio 10:1. The oven
temperature programme began at 40 °C, holding for 4.25 min, increased
by 10 °C/min to 150 °C, increased by 3 °C/min to 193 °C, holding for
8 min, increased by 3 °C/min to 230 °C, increased by 7 °C/min to 315 °C
and holding 6 min. The injector temperature was set at 250 °C for the
duration of the analysis. Chromatographic data were processed by
computer.
Fatty acid methyl esters were identified both by comparison of their
retention times with the 37 components Supelco standard and by mass
spectra with the MS analyzer operating at 70 eV according to the GC-MS
library. The area of each pic was multiplied by a response factor and the
area percentage was calculated on the basis of the total area of all pics.
At this stage, structure of compounds which could not be identified by
means of the Supelco standard weren’t confirmed. However, a lot of
FAMEs different from linear saturated chains and well-known un-
saturated chains could be present in sludge (like hydroxylated chains,
branched chains…). The choice has been made not to investigate fur-
ther on the structure of these compounds. There are presented as others.
Compounds with area percentage of less than 0.5% were neglected.
3. Results and discussion
3.1. Sludge characterization
The results of the characterization of the eight sludge samples from
the urban WWTP of Wegnez (Liège, Belgium) are presented Table 1.
Arithmetic averages of values of independent experiments are reported,
with error corresponding to one standard deviation (Average ± 1 SD).
At least duplicates were performed.
For some samples, water contents determined by Karl-Fisher titra-
tion are inconsistent with DMC determined by a drying method. This
can be perhaps explain by the fact than (i) the heating at 105 °C elim-
inates some compounds still present in the Karl-Fisher procedure (ii)
the measurement uncertainty of the Karl-Fisher procedure is high.
The results for lipid content are discussed below (3.4.)
3.2. Optimization of liquid-liquid extraction
After selection of the solvent, the optimization of the liquid-liquid
extraction was carried out varying (i) sludge pre-treatment (acidifica-
tion or not), (ii) number of consecutive extraction steps, (iii) contact
time and, (iv) ratio sludge to solvent. The primary sludge sampled in
August 2015 (sample number (1)) was used for the optimization ex-
periments.
All extraction tests were carried out as following, varying one
parameter at a time: (5 ± 0.5) g of wet sludge were previously treated
30 min in an ultrasonic bath, acidified until pH 2 by the addition of
1 mL of fuming hydrochloric acid and extracted by shaking three times
C. Kech et al. Fuel 212 (2018) 132–139

Table 1
Sludge characterization results from the eight sludge samples from urban WWTP located in Wegnez (Liège – Belgium).

Sample number Sludge Sampling date DMC (%wt) WS (%) Lipid content (%wt wet sludge) Lipid content (%wt DM)

(1) Primary (dewatered) August 2015 4.16 ± 0.04 92.9 ± 1.8 1.46 ± 0.13 34.5
(2) Primary (dewatered) March 2016 6.63 ± 0.08 94.7 ± 0.8 1.59 ± 0.03 24.1
(3) Primary (raw) October 2016 1.03 ± 0.02 92.8 ± 0.6 1.00 ± 0.03 -*
(4) Primary (dewatered) October 2016 2.19 ± 0.03 90.2 ± 0.6 1.22 ± 0.04 55.6

Average for primary sludges 92.7 1.32

(5) Secondary (raw) August 2015 1.95 ± 0.06 n.d. 0.95 ± 0.05 47.6
(6) Secondary (raw) March 2016 4.20 ± 0.05 97.2 ± 1.7 1.20 ± 0.06 28.6
(7) Secondary (raw) October 2016 0.81 ± 0.05 90.3 ± 1.3 0.98 ± 0.01 -*
(8) Secondary (dewatered) October 2016 2.45 ± 0.02 98.9 ± 0.9 1.09 ± 0.02 44.6

Average for secondary sludges 95.5 1.06

* Inconsistent results due to very low DMC.

60 min at room temperature with organic solvent (1900 rpm). 50

The results are represented under the form of a bar graph corre- 45

Lipid content (%wt DM)

sponding to the arithmetic average of values of independent experi- 40
ments, with error bars corresponding to one standard deviation 35
(Average ± 1 SD). At least duplicates were performed. 30 26.3
22.4
25
3.2.1. Determination of the extraction solvent 20
Starting from the methodology proposed by Olkiewicz and co- 15
workers [7], different solvents or combinations of solvents were tested 10
with a wet sludge (DMC 4.16%wt): hexane, chloroform, Bligh and Dyer
5
(chloroform/methanol/water) and Smedes (cyclohexane/isopropyl al-
0
cohol/water) mixtures.
chloroform/methanol/water Cyclohexane/isopropyl
The experiments were conducted as described above (3.2) except for alcohol/water
the Bligh and Dyer mixture for which no acidification and only one
Fig. 2. Optimization of liquid-liquid extraction from wet primary sludge: Bligh and Dyer
extraction of 60 min was achieved for practical reason. The same ex-
and Smedes mixtures comparison (1 extraction of 60 min – room temperature – shaking at
periment (no acidification and one extraction of 60 min) was operated
1900 rpm).
with Smedes mixture to compare the results.
The organic solvents volumes were 20 mL of hexane and chloroform
(sludge to solvent ratio 1:4), 5 mL of chloroform and 10 mL of methanol ((34.5 ± 2.9)%wt DM). This result is slightly higher than the one ob-
for Bligh and Dyer mixture (sludge to solvent ratio 1:1:2), and 8 mL of tained with hexane. This can be explained because a mixture of solvents
isopropyl alcohol and 10 mL of cyclohexane for Smedes mixture (sludge with different polarities can extract lipids of different polarities, like
to solvent ratio 1:1.6:2). polar fatty acids and phospholipids, and non-polar triglycerides, di-
Average lipid contents (%wt DM) obtained with each solvent are glycerides, monoglycerides and fatty acids of higher hydrocarbon [11].
presented in Figs. 1 and 2. The results are compared with those ob- Hexane alone is not able to extract polar lipids. The other advantages of
tained with hexane. this solvent combination are: low toxicity, non-olfactory, non-explosive,
The highest lipid content was obtained with chloroform lower density and cost effective.
((37.9 ± 1.2)%wt DM). This result is higher than the one obtained
with hexane ((32.8 ± 1.0)%wt DM). 3.2.2. Effect of sludge acidification
The lipid content obtained with the Bligh and Dyer mixture is lower Pastore and co-workers [10] performed a full characterization of the
than with the Smedes mixture ((22.4 ± 1.1)%wt DM and lipidic fraction extracted from sludge with hexane. They found that the
(26.3 ± 0.1)%wt DM respectively). main components are soaps (mainly calcium fatty acid salts). It can
For environmental reasons, the Smedes mixture was selected as the therefore be expected that the extraction step can be improved by using
most favourable combination for lipid extraction from wet sludge a strong acid to produce FFAs that are more soluble in the organic phase
than starting calcium soaps, increasing the extracted lipidic fraction.
50 Olkiewicz and co-workers evaluated the effect of sludge acidifica-
45 tion on the liquid-liquid extraction of lipids from wet sludge with
Lipid content (%wt DM)

37.9 34.5
40 hexane [7]. They concluded that sludge acidification highly increased
32.8 lipid yield (26.6%wt DM and 13.0%wt DM for acidified and non-acid-
35
30 ified samples, respectively).
25 The effect of acidification on the lipid extraction with Smedes
20
mixture was studied in this study according to the procedure described
above (3.2) (except that only two extractions of 60 min were per-
15
formed). The results are presented Fig. 3.
10
The results with and without acidification are respectively of
5
(33.5 ± 0.6)%wt DM and (31.8 ± 0.6)%wt DM. These values show
0 that acidification allows only a slightly better extraction yield. The lipid
n-hexane chloroform cyclohexane/isopropyl
extraction with Smedes mixture has the advantage over the extraction
alcohol/water
with hexane alone that acidification is not essential to obtain a high
Fig. 1. Optimization of liquid-liquid extraction from wet primary sludge: solvents com- lipid yield. But according to Pastore and co-workers [10], the presence
parison (3 extractions of 60 min – room temperature – shaking at 1900 rpm – acid media).
of soaps was expected and all the optimization tests were performed

135
C. Kech et al. Fuel 212 (2018) 132–139

50 50

Lipid content (%wt DM)

45 45
35.7
Lipid content (%wt DM)

40 40
33.5 31.1 33.2
35 31.8 35
28.9
30 30
25 25
20 20
15 15
10 10
5 5
0 0
With acidification Without acidification 30 min 40 min 50 min 60 min

Fig. 3. Optimization of liquid-liquid extraction from wet primary sludge: effect of sludge Fig. 5. Optimization of liquid-liquid extraction from wet primary sludge: determination
acidification (3 extractions of 60 min with Smedes mixture – room temperature – shaking of contact time (3 extractions with Smedes mixture – room temperature – shaking at
at 1900 rpm). 1900 rpm – acid media).

with prior acidification. 50

45
3.2.3. Determination of the number of extraction steps 40

Lipid content (%wt DM)

Lipid contents after two, three and four consecutive extraction steps 35 32.5
were compared according the method previously described (3.2) using 30
the Smedes mixture. The results obtained are presented Fig. 4. 23.2
25
The yield of extracted lipids increases with number of extraction
steps ((33.5 ± 0.6), (34.5 ± 2.9), (37.0 ± 1.2)%wt DM respec- 20
tively). 15
The optimization was limited to four extraction steps to avoid a too 10
long procedure, which would hinder transferability to a full-scale 5
system. 0
5g 10 g
3.2.4. Determination of contact time
Fig. 6. Optimization of liquid-liquid extraction from wet primary sludge: determination
Lipid contents after 30, 40, 50 and 60 min of contact time in each of sludge to solvent volume ratio (3 extractions of 60 min with Smedes mixture – room
extraction step were compared according the method described above temperature – shaking at 1900 rpm – acid media).
(3.2). The results are presented Fig. 5.
The lipid yield grows as the contact time increases in each extrac-
water is constant too. Therefore, when the amount of DM increases, the
tion step, reaching the highest result after 60 min (35.7%wt DM).
lipid yield decreases.
Although the results do not seem to stabilize to a threshold value, the
optimization was limited to 60 min as longer extraction times would
lead to a too long procedure. 3.2.6. Selection of the optimized procedure for the liquid-liquid extraction
The following criteria were adopted when selecting a final lipid
3.2.5. Determination of sludge to solvent volume ratio extraction method from wet sludge
The lipid contents obtained starting with 5 g and 10 g of wet sludge
and with a constant volume of Smedes mixture (ratio of 1:1.6:2 and • Solvent with low adverse effects on the environment,
2:1.6:2) were compared according the method described above (3.2). • Minimized solvent volume,
The results are presented Fig. 6. • Time to perform the procedure not exceeding one day of work,
The lipid yield is lower when the amount of wet sludge increases for • High extraction efficiency.
a constant volume of solvent. According to the procedure, amount of
According to the results obtained above, the proposed optimized
50 method is the following:
45
34.5 37.0 • Extraction with Smedes mixture,
Lipid content (%wt DM)

40
33.5 • Acidification of the sludge prior to extraction (not mandatory),
• Three extraction steps of 60 min each,
35
30
• Sludge to solvent (isopropyl alcohol/cyclohexane) ratio of 1:1.6:2
25 for the first extraction.
20
15 The lipid content obtained employing the above procedure was
10 compared with the results of previous studies (Table 2). The lipid yield
is higher using the combination of a non-polar and a polar solvent than
5
using hexane only and the extraction is effective at room temperature.
0 The method proposed by Olkiewicz and co-workers uses less solvent
2 steps 3 steps 4 steps
compared to the amount of wet sludge (sludge to solvent ratio of 2:1),
Fig. 4. Optimization of liquid-liquid extraction from wet primary sludge: number of ex- but the lipid yield obtained in the same conditions with a ratio of 1:2 is
traction steps (Extractions of 60 min with Smedes mixture – room temperature – shaking not higher than the yield obtained with the ratio of 2:1. Hence, based
at 1900 rpm – acid media).
on the samples tested in this study, the optimized method contributed

136
C. Kech et al. Fuel 212 (2018) 132–139

Table 2
Comparison of lipid contents obtained in this study with methods proposed in previous studies.

References Sludge DMC Solvent Ratio wet sludge/solvent vol. Temp. Time Steps Lipid content
(%wt) (°C) (h) (%wt DM)

[10] Primary (dewatered) 14.4 Hexane 1:1 38 1 3 20.0

[7] Primary (dewatered) 3.4 Hexane 2:1 Rt 1 3 26.7
This study Primary (dewatered) 4.2 Smedes mixture 1:1.6:2 Rt 1 3 34.5

50 mainly composed of biologic matter. Hence, the lipids extracted from

45 the latter are more polar because they are mainly originating from
37.2 microbial cells, which contain polar phospholipids in their cell walls
Lipid content (%wt DM)

40
35 32.5
30
25
20
15
10
5 The potential of the extracted lipids for biodiesel conversion was
0 evaluated for the eight sludge samples tested here using acid value and
Wet sludge 4.2 DM %wt Dry sludge 92.5 DM %wt
Fig. 7. Optimization of liquid-liquid extraction from primary sludge: comparison with dry
sludge (3 extractions of 60 min with Smedes mixture – room temperature – shaking at
1900 rpm – acid media).
to improve liquid-liquid extraction from wet sludge by a factor of about
1.3, with lipid yield above 34%wt DM.
3.3. Comparison with dry sludge
In order to compare the results obtained from wet sludge with those
obtained from dry sludge (conventionally used as the starting material
for oil and grease extraction), the optimized method was carried out on
a sample of wet sludge (DMC 4.2%wt) and a sample of the same sludge
previously dried in an oven and crushed (DMC 92.5%wt). The starting
mass of dried sludge was calculated to obtain the same amount of DM
(%wt) than the wet one. The results obtained are presented Fig. 7.
The lipid content in the dry sludge is slightly higher (37.2%wt DM)
than that in the wet sludge (32.5%wt DM). The method should be in-
dependent from the solid matter content of the sludge because a con-
stant water content is used in the extraction method. The higher results
are probably due to the fact that the dry sludge was crushed before
extraction, increasing the availability of the lipids for the solvent mix-
ture, while the wet sludge was previously treated in an ultrasonic bath.
A study of the pretreatment methods of the wet sludge should be ne-
cessary to optimize the extraction.
3.4. Comparison of primary and secondary sludges
[15]. The solvent mixture used for the extraction allows extracting li-
pids of different polarities. It can be therefrom concluded that lipids are
probably partially degraded during the biological treatment of the
WWT plant, resulting in a lower content in the final secondary sludge.
3.5. Evaluation of the potential for biodiesel conversion
FFA content, saponification number and fatty acid compositions.
Results are presented in Tables 3 and 4.
Acid values and FFA contents are higher for primary sludge than for
secondary sludge. This can be explained considering that primary
sludge contains soaps converted into FFAs upon exposure to an acidic
environment during the extraction process [16]. Lipids extracted from
both sewage sludges contain a large fraction of FFAs (> 25%), dis-
missing conventional base-catalysed conversion process. The reason is
that these FFAs are converted into soaps in a basic medium giving rise
to an emulsion making the separation from water arduous. So only a
specific acid-catalysed process must be used to the esterification/
transesterification reaction.
The saponification number is a measure of the amount of extracted
lipids that could be converted into biodiesel. It expresses the amount of
base that will react with 1 g of sample when heated in a specific
manner. It gives an estimation of the amount of acids present in the
sample, including any free acids originally present plus any combined
acids (for example, in esters) that have been converted to metal soaps
during the heating procedure. Saponification numbers of primary and
secondary sludges are similar and are on average 180.
The fatty acid composition of the lipids is essential when testing the
suitability of any potential feedstock for biodiesel production, since it
directly influences the properties of the produced fuel [17]. The fatty
acid compositions of the eight samples were determined in triplicates.
Results are presented in Table 4. Both sludge types have a significant
amount of palmitic acid (C16:0), stearic acid (18:0) and linoleic acid
Table 3
Extracted lipids characterization results from the eight sludge samples from urban WWTP
located in Wegnez (Liège – Belgium).

The optimized method was applied on the eight sludge samples. The Sample Sludge Sampling date Acid FFA content Saponification
results are presented Table 1. number value (%) number
Lipid content is expressed as weight of lipids per unit weight of wet (mg/
g)
sludge and per unit weight of DM.
The values show that lipid content is dependent on the DMC: sludge (1) Primary August 2015 148 67 252
samples with higher DMC allow higher amounts of lipids extracted. (2) Primary March 2016 124 56 193
When calculating the average lipid content (on the basis of wet sludge) (3) Primary October 2016 58 26 143
(4) Primary October 2016 104 48 166
by sludge type, it can be observed lipid content is higher in primary
sludge than in secondary sludge. Average for primary sludges 109 49 188
(5) Secondary August 2015 86 39 251
The primary sludge is produced from the mechanical settling of the
(6) Secondary March 2016 68 31 167
incoming wastewater. The secondary sludge is produced from the me- (7) Secondary October 2016 42 19 138
chanical settling after biological treatment of the wastewater. The (8) Secondary October 2016 50 23 152
nature of both sludges is thus different: primary sludge samples are Average for secondary sludges 61 28 177
mainly composed of organic matter, and secondary sludge samples are

137
C. Kech et al. Fuel 212 (2018) 132–139

Table 4
Fatty acids composition of the eight sludge samples from urban WWTP located in Wegnez (Liège – Belgium) (%wt).

Fatty acid Primary sludge Secondary sludge

August 2015 (1) March 2016 (2) October 2016 (3) October 2016 (4) August 2015 (5) March 2016 (6) October 2016 (7) October 2016 (8)

C12:0 – – – 1 – – – –
C14:0 3 3 1 3 1 2 2 2
C15:0 1 1 1 1 1 2 2 1
C16:1 1 – 3 1 10 7 7 9
C16:0 39 53 13 35 18 14 14 16
C17:0 1 1 – 1 – 1 1 1
C18:2 9 6 9 5 7 7 7 10
C18:3n3 – – – – – – – 1
C18:1n9c 7 4 7 5 6 6 6 7
C18:1n9t 2 2 4 2 5 4 4 6
C18:0 25 17 15 31 16 20 20 11
C20:1 – 1 – 1 – – – –
C20:0 1 1 1 1 1 1 1 1
C22:0 1 1 2 1 2 1 1 1
C24:0 2 2 7 2 6 5 5 6
Others 9 10 37 12 26 31 31 29

(C18:2). The amount of C16:0 is lower in secondary sludges than in
primary sludges. On the other hand, secondary sludges contain higher
amounts of palmitoleic acid (C16:1) than primary sludges, and the
amount of “others” is higher too than in primary sludges. Linolenic acid
(C18:3) is absent from both sludge types. These results confirm those
obtained in the study of Olkiewicz and co-workers [16], and it can be
concluded that sludge is an appropriate feedstock for biodiesel pro-
duction. The high content in saturated and monounsaturated fatty acids
will give biodiesel with a high stability, but with the disadvantage of
perhaps presenting poor cold flow properties.
From saponification number and fatty acid composition, we can
estimate the amount of saponifiable matter into the extracted lipids.
According to the annex B of NBN EN ISO 3657 [14], we can calculate an
average relative molecular mass from the fatty acid composition we
have determined for each sample (taking into account identified fatty
acids only), and from the saponification number, we can calculate the
amount of matter that could be transesterified per gram of lipids ex-
tracted, which is a theoretical amount of biodiesel that could be ob-
tained. It’s an approximate value that take into account only a part of
the fatty acids present in the extracted lipids, but it gives us a relatively
good idea of the potential for biodiesel conversion. On average, the
conversion into biodiesel of lipids extracted from primary sludges could
be at least of 75%, and the conversion of lipids extracted from sec-
ondary sludges could be at least of 63%. These results are close to those
obtained in the study of Olkiewicz and co-workers [7] after transes-
terification of primary sludge (72% on the basis of lipids).
These yields of conversion are limited because lipids extracted from
sludge cannot be totally converted into biodiesel because some non-
saponifiable lipids co-extracted with the solvent mixture cannot be in-
volved in the esterification/transesterification reaction.
4. Conclusions
The goal of this paper was to propose a method for lipid extraction
at room temperature from wet sludge in order to convert them into
biodiesel. Avoiding energy intensive steps of sludge dewatering and
drying could permit the lowering of the total biodiesel production cost.
The liquid-liquid extraction using cyclohexane/isopropyl alcohol/
water mixture (Smedes mixture) is an alternative improving the ex-
traction yield of lipids ((34.5 ± 2.9)%wt DM) compared to hexane
((32.8 ± 1.0)%wt DM). This mixture is environment-friendly and cost-
effective, and can be used with both primary and secondary sludges as
polar and non-polar lipids can be extracted. Prior acidification increases
method efficiency and the results are independent of sludge solid
matter content because a constant water content is used.
The optimized method can be applied at a laboratory scale on a
work day and with standard equipments. Further studies remain how-
ever required to assess large-scale feasibility of this extraction method.
Sludge is a very suitable feedstock for the production of biodiesel, as
it is mainly composed of fatty acids that are essential for this purpose.
The conversion into biodiesel from sludge extracted-lipids will have
to involve an acid-catalysed process because of the presence of high
amount of FFAs in primary and secondary sludges.
Acknowledgements
The authors wish to thank the Walloon public company SPGE
(Société Publique de Gestion de l’Eau) and the WWTP of Wegnez (Liège,
Belgium) for their collaboration and for the supply of sludge samples.
Financial support was provided by ISSeP (Institut Scientifique de
Service Public) under the Moerman funding mechanism (BioBoS pro-
ject).
References
[1] Directive 2009/28/EC of the European Parliament and of the Council of 23 April
2009 on the promotion of the use of energy from renewable sources and amending
and subsequently repealing Directives 2001/77/EC and 2003/30/EC, 2009.
[2] Bhuiya MM, Rasul MG, Khan MM, Ashwath et N, Azad AK. Prospects of 2nd gen-
eration biodiesel as a sustainable fuel – Part: 1 selection of feedstocks, oil extraction
techniques and conversion technologies. Renew Sustain Energy Rev
2016;55:1109–28.
[3] Sangaletti-Gerhard N, Cea M, Risco et V, Navia R. In situ biodiesel production from
greasy sewage sludge using acid and enzymatic catalysts. Bioresour Technol
2015;179:63–70.
[4] Siddiquee et MN, Rohani S. Lipid extraction and biodiesel production from muni-
cipal sewage sludges: a review. Renewable Sustainable Energy Rev
2011;15:1067–72.
[5] Société Publique de Gestion de l'Eau (SPGE), Available: http://www.spge.be; [Last
access May 2016].
[6] Dufreche S, Hernandez R, French T, Sparks D, Zappi et M, Alley E. Extraction of
lipids from municipal wastewater plant microorganisms for production of biodiesel.
J Am Oil Chem Soc 2007;84:181–7.
[7] Olkiewicz M, Caporgno MP, Fortuny A, Stüber F, Fabregat A, Font et J, et al. Direct
liquid–liquid extraction of lipid from municipal sewage sludge for biodiesel pro-
duction. Fuel Process Technol 2014;128:331–8.
[8] Bligh et EG, Dyer WJ. A rapid method of total lipid extraction and purification. Can
J Biochem Physiol 1959;37:911–7.
[9] Smedes F. Determination of total lipid using non-chlorinated solvents. Analyst
1999;124:1711–8.
[10] Pastore C, Lopez A, Lotito et V, Mascolo G. Biodiesel from dewatered wastewater
sludge: a two-step process for a more advantageous production. Chemosphere
2013;92:667–73.
[11] Siddiquee et MN, Rohani S. Experimental analysis of lipid extraction and biodiesel
production from wastewater sludge. Fuel Process Technol 2011;92:2241–51.
[12] ISO 11465 Soil quality. Determination of dry matter and water content on a mass
basis. Gravimetric method, 1993.

138
C. Kech et al.
[13] NBN EN ISO 660 Animal and vegetable fats and oils – Determination of acid value
and acidity, 2009.
[14] NBN EN ISO 3657 Animal and vegetable fats and oils – Determination of saponi-
fication value (ISO 3657:2013), 2013.
[15] Zhu F, Zhao L, Jiang H, Zhang Z, Xiong Y, Qi et J, et al. Comparison of the lipid
content and biodiesel production from municipal sludge using three extraction
139
Fuel 212 (2018) 132–139
methods. Energy Fuels 2014;28:5277.
[16] Olkiewicz M, Fortuny A, Stüber F, Fabregat A, Font et J, Bengoa C, et al. Effects of
pre-treatments on the lipid extraction and biodiesel production from municipal
WWTP sludge. Fuel 2015;141:250–7.
[17] Knothe G. Dependence of biodiesel fuel properties on the structure of fatty acid
alkyl esters. Fuel Process Technol 2005;86:1059–70.