Bioresource Technology 214 (2016) 826–835

Contents lists available at ScienceDirect

Bioresource Technology
journal homepage: www.elsevier.com/locate/biortech

Review

A novel process for low-sulfur biodiesel production from scum waste

Huan Ma a,b, Min M. Addy b, Erik Anderson b, Weiwei Liu b,c, Yuhuan Liu d, Yong Nie e, Paul Chen b,
Beijiu Cheng a, Hanwu Lei f, Roger Ruan b,d,⇑
a
School of Life Sciences, Anhui Agricultural University, 130 Changjiang West Road, Hefei 230036, China
b
Center for Biorefining and Department of Bioproducts and Biosystems Engineering, University of Minnesota, 1390 Eckles Avenue, St. Paul, MN 55108, USA
c
School of Engineering, Anhui Agricultural University, 130 Changjiang West Road, Hefei 230036, China
d
MOE Biomass Engineering Research Center, Nanchang University, Jiangxi 330047, China
e
Separation Engineering Institute, Zhejiang University of Technology, Hangzhou 310014, China
f
Department of Biological Systems Engineering, Washington State University, Tri-Cities Campus, Richland, WA 99354, USA

h i g h l i g h t s

A new process was developed to convert scum to low sulfur biodiesel.

Solvent extraction combined with acid washing was used to clean the scum oil.

Activated carbon was integrated with reflux distillation to reduce sulfur content.

70% filtered and dried scum was converted to biodiesel with sulfur content 615 ppm.

Biodiesel from new process showed better properties than that from other processes.

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

Article history: Scum is an oil-rich waste from the wastewater treatment plants with a high-sulfur level. In this work, a
Received 19 March 2016 novel process was developed to convert scum to high quality and low sulfur content biodiesel. A combi-
Received in revised form 9 May 2016 nation of solvent extraction and acid washing as pretreatment was developed to lower the sulfur content
Accepted 10 May 2016
in the scum feedstock and hence improve biodiesel conversion yield and quality. Glycerin esterification
Available online 13 May 2016
was then employed to convert free fatty acids to glycerides. Moreover, a new distillation process integrat-
ing the traditional reflux distillation and adsorptive desulfurization was developed to further remove sul-
Keywords:
fur from the crude biodiesel. As a result, 70% of the filtered and dried scum was converted to biodiesel
Scum
Biodiesel
with sulfur content lower than 15 ppm. The fatty acid methyl ester profiles showed that the refined bio-
Solvent extraction diesel from the new process exhibited a higher quality and better properties than that from traditional
Adsorptive desulfurization process reported in previous studies.
Glycerolysis Ó 2016 Elsevier Ltd. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 827
2. Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 827
2.1. Material and reagent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 827
2.2. Process design and experiment design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 828
2.2.1. Basic process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 828
2.2.2. Experimental design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 830
2.3. Analytical methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 830
2.3.1. Acid value (AV) determination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 830
2.3.2. Fatty acid methyl ester (FAME) analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 831
2.3.3. Sulfur analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 831

⇑ Corresponding author at: Center for Biorefining and Department of Bioproducts

and Biosystems Engineering, University of Minnesota, 1390 Eckles Avenue, St. Paul,
MN 55108, USA.
E-mail address: ruanx001@umn.edu (R. Ruan).

http://dx.doi.org/10.1016/j.biortech.2016.05.029
0960-8524/Ó 2016 Elsevier Ltd. All rights reserved.
 H. Ma et al. / Bioresource Technology 214 (2016) 826–835 827

3. Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 831

3.1. Effect of solvent type on the sulfur removal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 831
3.2. Process comparison for low-sulfur biodiesel conversion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 831
3.2.1. Sulfur removal performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 831
3.2.2. Oil yield comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 832
3.2.3. FAME comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 832
3.3. Fuel properties of the scum-derived biodiesel from different processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 833
3.3.1. Cetane number (CN) and Kinematic viscosity (KV) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 833
3.3.2. Cloud point and acid value . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 834
3.4. Future work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 834
4. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 834
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 834
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 834

1. Introduction biodiesel with higher energy conversion efficiency was proposed

Biodiesel production has attracted considerable attention in
recent years due to the global energy crisis and awareness in envi-
ronmental pollution. To date, biodiesel conversion from vegetable
oils has been the primary route for biodiesel production
(Sivakumar et al., 2011). However, production of vegetable oil
based biodiesel competes with food production by engrossing agri-
cultural cropland. A major obstacle to the commercialization of
biodiesel production is the high cost associated with the oil feed-
stock, which makes up nearly 75–85% of the total production costs
(Demirbas, 2009). Therefore, low-cost feedstocks such as used
cooking oil, spent oil, grease, or low-quality oil are regarded
promising high potential feedstocks for biodiesel production
(Abedin et al., 2016; Li et al., 2014; Wang et al., 2016).
Scum is a floatable waste by-product from the wastewater
treatment process. It is an oil-rich material containing grease, ani-
mal fat, vegetable oil, food wastes, plastic material, soaps, wax, and
many metallic and non-metallic impurities. The oil content of
scum could be as high as 60% (Bi et al., 2015), making it a promis-
ing alternative feedstock for biodiesel production. Attempts have
been made to convert scum to biodiesel (Anderson et al., 2016;
Bi et al., 2015; Mu et al., 2016). Bi et al. (2015) proposed a novel
six-step method to produce scum-based biodiesel. In this process,
a combination of acid washing with acid catalyzed esterification
was used to remove impurities and convert free fatty acid to
methyl esters, followed by base-catalyzed transesterification to
produce fatty acid methyl ester (FAME). Converting scum to bio-
diesel is potentially a more profitable process than scum combus-
tion and anaerobic digestion. It was reported that scum to biodiesel
conversion yielded 29% and 104% more profit than scum combus-
tion and anaerobic digestion, respectively (Mu et al., 2016).
Although scum holds a great potential as a biodiesel feedstock,
one of its major issues is of its high sulfur content ranging from 600
to 1000 ppm. The sulfur compound presented in the scum primar-
ily comes from sulfur-containing organics in the wastewater from
restaurants, households and other facilities. The sulfur goes to the
atmosphere as sulfur dioxide during the combustion of biodiesel,
causing significant environmental and health problems (Bu et al.,
2011; Jung and Jhung, 2015). To regulate sulfur in biodiesel, the
American Society for Testing and Materials (ASTM) D6751 (2015)
specifies the sulfur content in biodiesel to be less than 15 ppm.
In a previous study (Bi et al., 2015), the scum-based biodiesel con-
tained 33.6 ppm sulfur, which is twice the sulfur limit allowed by
the ASTM specifications. In addition, there was a drawback in the
acid-esterification which was used to convert free fatty acid
(FFA) into methyl esters in the previous study. The resulting wet
methanol from the acid-esterification must be dried for reuse,
which causes two times higher thermal energy consumption in
biodiesel plants than the process without drying (Anderson et al.,
2016). Therefore, a novel process to lower the sulfur content in
in this study.
In the present work, solvent extraction incorporated with glyc-
erolysis and desulfurization distillation was investigated. Solvent
extraction combined with acid washing can deep clean the scum
oil by removing slightly polar and non-oil impurities, including a
part of organosulphur compounds. Heptane, hexane, and petro-
leum ether were investigated and compared for sulfur reduction
in the solvent washing step. Instead of traditional acid esterifica-
tion, glycerin esterification or so called ‘‘glycerolysis” was used to
reduce FFA in low-quality scum. Glycerolysis does not consume
any acid and methanol, therefore the vacuum stripping required
for methanol recycle can also be eliminated. In addition, a new dis-
tillation system integrating the traditional reflux distillation and
adsorptive desulfurization was developed to further remove sulfur
impurities from the biodiesel. Sludge-derived activated carbon was
used as the adsorbent in the new distillation system. The activated
carbon was packed on the top of the distillation column and steel
wool was used to fill the rest of the column. Consequently, the
adsorptive effect was enhanced due to increased surface contact
when fatty acid methyl esters (FAME) were in vapor phase during
the reflux distillation. For comparison, two other processes were
explored to prepare scum-based biodiesel, both of which used
the same reflux distillation system without addition of adsorbents
in the column, with one process including solvent extraction step
and the other excluding the solvent extraction step. The change
in sulfur-content in each step of the three processes was recorded
and compared in order to understand desulfurization performance
of the new process and evaluate the sulfur reduction potentials of
these three different processes. In addition, the oil yields of differ-
ent steps in three routes were determined to assess the conversion
efficiency of each process. Furthermore, the properties such as
cetane number, kinematic viscosity, and cloud point of the final
biodiesel from different processes were evaluated based on their
FAME profiles.
2. Method
2.1. Material and reagent
The scum samples were collected from the Metropolitan
Wastewater Treatment Plant in St. Paul, MN. Sulfuric acid (98%,
AR) was purchased from Mallinckrodt Baker, Inc., Paris, Kentucky,
butanetriol (1000 mg/ml), tricaprin (8000 mg/ml), and N-methyl-
N-(trimethylsilyl) trifluoroacetamide (MSTFA) from Sigma–
Aldrich, Inc., sodium methoxide (30% in methanol), methanol
(anhydrous, 99.8%), n-heptane (HPLC grade), potassium hydroxide
concentrate (0.1 N), 50 micron filter paper, and glycerol (99.9%)
from Thermo Fisher Scientific, Inc., distilled water from Premium
Waters, Inc., MN, USA, and Supelco 37 component fatty acid
methyl ester (FAME) mix from Sigma–Aldrich, Inc.
828 H. Ma et al. / Bioresource Technology 214 (2016) 826–835

Fig. 1. Process flow diagram for scum to biodiesel conversion.

2.2. Process design and experiment design
2.2.1. Basic process
The general process for the scum to biodiesel conversion is
illustrated in Fig. 1 and the main process can be summarized as
the following six steps:
Filtering ? Acid hydrolysis ? Solvent extraction ? Glyceroly-
sis ? Base catalyzed transesterification ? Fractional distillation
losic biomass and other solid particles were removed and a mix-
ture of liquid oil and water was recovered. The water in the
mixture was separated via gravitational settling and the upper liq-
uid scum oil was collected. Water content in the raw scum and
scum oil were determined according to ASTM E1756-08 (2008).
As reported in the previous study, scum oil contained metal and
non-metal impurities, such as Ca, Fe, Al, Mg, Zn, N, P, and etc.
which can be removed using acid washing (Bi et al., 2015).

2.2.1.1. Filtering. For better separation of the scum oil from bioso- 2.2.1.2. Acid washing. Acid washing was used to convert the soap in
lids, raw scum was heated to 315 K to reduce the viscosity. The liq- the scum to free fatty acids and remove metallic and non-metal
uefied scum was then filtered using a polyester mesh filter bag impurities. Based on the weight of scum oil from the last step, a
with a pore size of 100 micron. After filtration, the plastics, cellu- combination of 5.0% wt/wt of 98% sulfuric acid and 75% wt/wt of
 H. Ma et al. / Bioresource Technology 214 (2016) 826–835 829

water was charged to the oil. The resulting mixture was placed into
a flask with an electric agitator on top and maintained at 338 K for
1 h. After the reaction, the mixture was allowed to settle for an
additional 1 h under 338 K. After two distinct phases were present,
the acidic water and white precipitate were decanted from the bot-
tom of the separating funnel. The top layer of scum oil was sepa-
rated for the next process.
2.2.1.3. Solvent washing. Solvent washing was employed to further
remove impurity from the scum oil and improve biodiesel quality.
It also made the glycerin esterification and transesterification
easier due to less impurity interference with the reactants. Based
on the weight of scum oil obtained after acid washing, 50% wt/
wt of heptane and 30% wt/wt of water were added to the oil. The
mixture was maintained in a 3-necked round-bottom flask at
338 K for 1 h, and the flask was affixed with a condenser, an agita-
tor and a thermometer. After 30 min of settling, the bottom acidic
water phase and middle slightly polar hydrophobic regions were
decanted as wastes while the top layer of heptane and oil was
transferred to a rotary evaporator for solvent recovery. The hep-
tane evaporation was carried out at 373 K and 2.5 mmHg. Under
the high temperature and vacuum condition, 95% heptane could
be easily recycled.
2.2.1.4. Glycerin esterification. Instead of acid catalyzed esterifica-
tion, glycerin esterification or glycerolysis was used to convert
FFA to glycerides in scum oil. Glycerolysis was carried out in a jack-
eted reactor equipped with an agitator, a condenser and a nitrogen
inlet. Before glycerolysis, the oil was heated to 473 K under an inert
nitrogen environment to remove residual heptane and water. Once
the temperature reached to 511 K, 18% wt/wt of glycerin was
charged to the oil. The reactor was maintained at 511 K for 3 h.
The acid value (AV) of the oil phase was monitored every 10 min
during the first hour, and after 180 min. After glycerolysis, the oil
would undergo a base-catalyzed transesterification, effectively
releasing free glycerin and allowing it to physically separate from
the methyl esters.
2.2.1.5. Base-catalyzed transesterification. Base-catalyzed transes-
terification is a promising important method for biodiesel produc-
tion from low water feedstock (water content less than 0.05% wt)
with low free fatty acid (AV less than 2). It typically uses sodium
methoxide (CH3NaO) as catalyst to increase the reaction rate. The
scum oil derived materials after glycerolysis would meet the
requirement as a feedstock for base catalyzed esterification. 30%
wt/wt of methanol and 0.625% wt/wt of sodium methoxide were

Fig. 2. Schematic diagram of the vacuum distillation apparatus. (1) Heating mantle, (2) distilling flask, (3) temperature sensor, (4) distillation column, (5) thermometer, (6)
water inlet, (7) water outlet, (8) to vacuum pump, (9) condenser, (10) reflux ratio controller, (11) distillate collection flask.
830 H. Ma et al. / Bioresource Technology 214 (2016) 826–835
added into the corresponding amount of oil with vigorous and con-
stant agitation. The temperature was maintained at 347 K for 1 h.
In general, two phases will form after the reaction and settling.
The less dense top phase will include biodiesel and methanol,
while the denser bottom phase will contain glycerin, methanol
and catalysts. However, when using scum oil as feedstock, the
two phases were difficult to distinguish due to the similar dark col-
ors between the two layers. In addition, the inter-solubility of
methyl ester and glycerol was increased duo to presence of soap
and other organic compounds (such as surfactants or other impu-
rities). Therefore, a process called ‘glycerol washing’ was used to
improve FAME/glycerol separation. The sodium methoxide catalyst
in the glycerol layer must be neutralized before methanol stripping
by the addition of concentrated sulfuric acid. Otherwise, the mix-
ture would undergo a reverse reaction and re-saponification. After
removal of catalysts, the residual methanol in the mixture was
evaporated using a rotary evaporator. Then, glycerol (glycerol:
oil = 1:1 by weight) was charged into the mixture oil and mixed
thoroughly by magnetic stirring at 333 K for 10 min. After settling
for 1 h, the upper layer (crude biodiesel) containing mostly FAME
was separated for the subsequent distillation step. The lower layer,
composed of glycerol, traces of biodiesel and other impurity, could
also be distilled and recycled to improve plant economics.
2.2.1.6. Distillation apparatus. The crude biodiesel obtained from
base-catalyzed transesterification process was transferred to a cus-
tomized vacuum rectification distillation apparatus (Fig. 2). The
distillation column (60 cm  5 cm) was packed with steel wool
or the mixture of steel wool and activated carbon derived from
sludge pyrolysis. During distilling, the bottom temperature was
increased from 413 to 513 K, while the top temperature was
increased from 393 to 433 K, and the vacuum was maintained
between 0.3 and 0.4 mmHg. The reflux ratio was kept at 1:1 when
the bottom temperature was below 483 K and it was set to 1:3
after the temperature was above 483 K. The refined FAMEs were
collected at different temperatures in the range of 463–513 K.
The samples were analyzed with a GC–MS technique.
2.2.1.7. Preparation of activated carbon. Sewage sludge pellet (2–
3 mm dia.) was provided by the Blue Lake Wastewater Treatment
Plant in Minnesota, USA. The activated carbon was prepared via
Fig. 3. Comparison of three process routes for low-sulfur biodiesel production from
scum waste.
the pyrolysis carbonization of sewage sludge. In a typical synthesis,
sludge pellet (5 g) was maintained in a vacuum furnace at a con-
stant heating rate of 10 °C/min from room temperature to the
pyrolysis temperature of 500 °C and then held at this temperature
for 2 h. The Brunauer–Emmett–Teller (BET) surface area of acti-
vated carbon derived from sludge pellet was 495 m2/g and total
pore volume was 0.298 cm3/g.
2.2.2. Experimental design
2.2.2.1. Comparison of different solvents for solvent washing. Solvent
washing with three different solvents, namely heptane, hexane,
and petroleum ether was studied to evaluate the effectiveness of
these solvents. The same procedure described in Section 2.2.1.3
was used. After the solvent was distilled off, the sulfur concentra-
tions in the oil samples were measured using ultraviolet fluores-
cence as described in Section 2.3.3. The desulfurization rate
(DS%) was calculated as follows:
DSð%Þ ¼ ðS0  S1 Þ=S0  100% ð1Þ
where S0 and S1 are the sulfur concentrations of the oil samples
before and after solvent washing, respectively.
2.2.2.2. Process comparison. To reduce the sulfur content and meet
the ASTM standard (Sulfur content 615 ppm), three oil rendering
processes were used and compared as illustrated in Fig. 3. The
scum liquid after acid washing was divided into three aliquots to
go through three different process routes, dubbed R I, R II and R
III (Fig. 3). For R I, the raw scum was subjected to the entire base
conversion procedure from filtration to distillation, except heptane
washing. For R II and III, the samples were pretreated with heptane
but distilled using different columns. As shown in Fig. 3, the distil-
lation column was packed with steel wool in R I and II, while it was
filled with the mixture of steel wool (at the bottom of the column)
and sludge-activated carbon (10% of oil weight and on the top of
the column) in R III. The refined FAMEs obtained from all three pro-
cedures were collected within the same temperature ranges (463–
473 K, 473–483 K, 483–493 K, and 493–518 K) to determine and
compare the sulfur contents. The FAME in all samples was ana-
lyzed using GC–MS, and the sulfur contents were determined using
ultraviolet fluorescence.
2.3. Analytical methods
2.3.1. Acid value (AV) determination
Acid value (mg KOH g1 oil) is an indicator of the free fatty
acid content in the scum oil. The acid value of fatty acids was
Fig. 4. Effects of three solvent extractions on desulfurization rates of the scum oil.
 H. Ma et al. / Bioresource Technology 214 (2016) 826–835 831

determined according to the standard test method ASTM D1980- The gray middle layer formed in heptane washing process may
87 (1998). Variations in acid value results were minor with a stan- contain a large quantity of organosulphur compounds and could
dard deviation of less than 1%. In glycerolysis reaction, changes in be decanted as a waste. Therefore, higher desulfurization rate can
acid value were measured to monitor the oil conversion state. be achieved through the heptane washing. In contrast, there was
no middle layer formed when using hexane and petroleum ether
2.3.2. Fatty acid methyl ester (FAME) analysis extraction, indicating that a higher proportion of slightly polar,
FAME in all samples was analyzed with GC–MS using the non-oil impurities remained in the top hexane/oil or petroleum
method from Li et al. (2014). Oil samples (about 100 mg) were ether layer/oil, due to the higher non-polar affinity of hexane and
weighed into 10 ml volumetric flasks and diluted with dichloro- petroleum ether. Moreover, the benefits of heptane extraction also
methane to 10 ml. The GC–MS (Agilent 7890-5975C, USA) include easing the downstream process of biodiesel production
equipped with a HP-5 MS capillary column and a mass detector (e.g., the glycerol and FAME separation), increasing final biodiesel
was used. The carrier gas used was Helium with a flow rate of yield, and improving the biodiesel quality. Without the heptane
1.2 mL/min. The heating program was: initial temperature, 40 °C extraction, impurities would have to be carried through the whole
(3 min); rate, 5 °C/min to 290 °C. 1 lL of sample was injected into process, causing unwanted side reaction and emulsions and high
the GC with a split ratio of 1:10. The temperatures of injector and sulfur content in the refined biodiesel, which will be described
detector were maintained at 250 °C and 230 °C, respectively. Chro- and demonstrated in Section 3.2.
matographic data were recorded and addressed by a built-in Agi-
lent data analysis software. Components were identified in 3.2. Process comparison for low-sulfur biodiesel conversion
National Institute of Standards and Technology (NIST) Mass Spec-
tral Database. Quantification was carried out by comparing the In order to develop a low-sulfur biodiesel conversion, three pro-
peak area with that of the FAME standard mixtures (Supelco 37 cesses were designed to prepare biodiesel from the scum (Fig. 3).
component, Sigma–Aldrich, Co.). The scum oil after acid washing was divided into three aliquots
which went through three different routes as described above. In
R I, the scum oil was not subjected to the solvent washing, but
2.3.3. Sulfur analysis
the crude biodiesel was refined using the same column packed
Oil samples were submitted to Iowa Central Fuel Testing Labo-
with steel wool as R II. For comparison, heptane was chosen as
ratory in Iowa for elemental sulfur analysis. The ASTM D5453-12
the solvent to wash scum oil and removed a part of sulfur in R II
(2012) method titled ‘‘Standard Test Method for Determinations
and III. In addition, a combination of reflux distillation and adsorp-
of Total Sulfur in Light Hydrocarbons, Spark Ignition Fuel, Diesel
tive desulfurization was developed and used in R III to reduce sul-
Engine Fuel, and Engine Oil by Ultraviolet Fluorescence” was used
fur content in biodiesel. The distillation column used in R III was
to determine sulfur concentrations.
filled with the sludge-activated carbon (10% of oil weight) on the
top and steel wool at the bottom, which was different from R I
3. Results and discussion and II. The detailed comparison and the influences of the three pro-
cesses on scum to biodiesel conversion were described below.
3.1. Effect of solvent type on the sulfur removal
3.2.1. Sulfur removal performance
Three solvents, petroleum ether, hexane and heptane were Sulfur contents of the samples obtained using the three pro-
employed in the solvent extraction process. The desulfurization cesses were monitored to determine if any of the processes tested
rates were calculated based on the changes in sulfur content in could realize the low-sulfur biodiesel conversion from scum with
the scum liquid before and after solvent extraction. As can be seen high sulfur level. As seen in Table 1, nearly no change occurred
from Fig. 4, almost 17% of sulfur-containing compounds were in the sulfur content of the samples before (852.2 ppm) and after
removed when using heptane as a solvent, which was much higher acid washing (851.6 ppm), indicating that acid washing step did
than using hexane (10%) and petroleum ether (1.2%). This may be not reduce the sulfur content effectively because the sulfur existed
attributed to the different polarities of the solvents. As non-polar in the organosulfur form in scum oil. Despite of this, it was more
solvents, the polarity of heptane (0.2) is much higher than that of effective for removing inorganic impurities such as calcium, iron,
hexane and petroleum (0.06 and 0.01). Therefore, the low polarity aluminum, magnesium and zinc etc. as reported by Bi et al.
impurities in scum, not polar enough to be water soluble, but also (2015). In comparison with R I, the sulfur content of the samples
not miscible with heptane, were trapped in the middle layer in R II and III were markedly reduced 17% from 852.2 ppm to
between the top heptane/oil layer and the bottom water layer. 707.1 ppm after heptane washing. Moreover, the process including
the heptane washing step (in R II and III) yielded nearly 57% of sul-
fur reduction (from 852.2 ppm to 362.7 ppm) in scum oil after
Table 1
transesterification step, while only 33.2% (from 852.2 ppm to
Effects of different processes on the change of sulfur contents.
568.8 ppm) was achieved by the process without the heptane
Samples Sulfur contents (ppm) washing (R I). In addition, it was found that the FAME/glycerol sep-
RI R II R III aration after transesterification in R I was more difficult to carry
Raw scum 852.2 852.2 852.2 out than that in R II and III. This was mainly attributed to
After acid washing 851.6 851.6 851.6 unwanted emulsifications and increased miscibility of FAME and
After heptane washing – 707.1 707.1 glycerol caused by the organic impurities, which could be allevi-
After glycerin esterification 762.7 504.5 504.5 ated by the heptane extraction.
After transesterification 568.8 362.7 362.7
Noticeably, the greatest reduction of sulfur occurred during the
After distillation reflux distillation stage for all three processes. The sulfur contents
Biodiesel (180–200 °C) 66.3 33.4 48.5
Biodiesel (200–210 °C) 12.1 11.7 8.9
of all the samples obtained in low temperature range (180–200 °C)
Biodiesel (210–220 °C) 7.4 12.3 9.6 or high temperature range (220–246 °C) were much higher than
Biodiesel (220–246 °C) 21.2 47.9 18.2 that yielded in middle temperature ranges (200–210 °C and 210–
Biodiesel (180–246 °C) 24.2 (Fail) 21.8 (Fail) 13.3 (Pass) 220 °C). This reveals that most of sulfur may be bound to the light
ASTM limit 615
cut and heavy cut of crude biodiesel. As expected, the sulfur
832 H. Ma et al. / Bioresource Technology 214 (2016) 826–835

Table 2 3.2.2. Oil yield comparison

The oil yield in the process of biodiesel production.
Step in process Oil yield (%)
RI R II
Acid washinga 89.6 89.6
Heptane washinga – 98.5
Glycerolysis reactiona 98.0 98.3
Transesterificationa 89.3 93.0
Distillationa 72.3 88.5
Total biodiesel yieldb (%) 56.5 71.4
a
Oil yield was calculated based on the dry weight of oil before and after the
corresponding step in different Res.
b
Total biodiesel yield was calculated based on the dry weight of the filtered scum
oil before acid washing.
content (13.3 ppm) of the final biodiesel (mixed FAME obtained
from 180 °C to 246 °C) in R III passed the ASTM specification
(<15 ppm). While for the R I and II, the sulfur contents were
24.15 ppm and 21.8 ppm, which were 1.6 times and 1.45 times
of the ASTM limit, respectively. This demonstrates that the combi-
nation of heptane extraction, reflux distillation and adsorptive
desulfurization is an effective way to reduce sulfur in scum-
derived biodiesel. Compared with R II, nearly 40% of sulfur was
removed in R III by sludge carbon adsorption (from 21.8 ppm to
13.3 ppm) during distillation stage. This is mainly attributed to
the porous structure, high surface area and large amounts of active
adsorption sites of the activated carbon. There are many oxygen-
containing functional groups on the surface of activated carbon,
which are able adsorb sulfur-containing compounds by means of
oxygen-sulfur interactions (Jung and Jhung, 2015). Simultaneously,
the adsorptive desulfurization could be enhanced when FAME liq-
uid was in vapor phase under the low pressure and high tempera-
ture condition in the reflux distillation stage. This is mainly
because the FAME vapor was more easily to access to the adsorp-
tion sites in the adsorbents than as in a liquid phase.
On the other hand, heptane washing was also found to be an
important step for the whole the conversion process. Although
only 17% of sulfur was removed from the scum oil in heptane
washing step, it could alleviate the burden of sulfur removal in
the subsequent processing stages, as previously discussed. More-
over, the color of biodiesel in R I was darker than that in R II and
III, indicating that the heptane washing helped improve the biodie-
sel quality, which was also confirmed in the FAME analysis,
detailed below. Additionally, the result from previous research
(Bi et al., 2015) also indicated that the scum oil pretreated only
by acid washing without heptane extraction resulted in 33.6 ppm
sulfur content in final biodiesel, which was two times of the ASTM
limit.
The total oil yields and biodiesel yields of different steps in the
three routes were calculated and compared in Table 2. The oil
R III yields of acid washing (89.6%) and glycerolysis reaction (98%,
98.3%) in all three routes were nearly the same. In R I, transester-
89.6
98.5
ification resulted in an 89.3% yield and distillation step resulted
98.3 in a 72.3% yield. R II and III showed a similar yield of 93.6% during
93.0 the transesterification reaction and a slightly different oil yield in
86.8 distillation step (88.5%, 86.8%, respectively). As a result, the total
69.8
biodiesel yield from the dried and filtered scum oil in R I (56.5%)
was much lower than that of R II (71.4%) and R III (69.8%). All of
these results further revealed that heptane washing played an
important role in scum to biodiesel conversion. Without heptane
washing, the impurities in scum oil would be carried through the
process, which has shown to cause negative effect on transesterifi-
cation reaction by adsorbing or bonding on reactants (Anderson
et al., 2016). Consequently, this could hinder the monoglycerides
and di-glycerides from accessing methanol and could weaken the
conversion of glycerides into esters, yielding a low oil yield in R
I. On the other hand, it was suspected that some organic impurities
(for instance, slightly polar hydrophobic impurities and soap)
could increase the miscibility of glycerol and methyl ester. There-
fore, when scum oil was not subjected to heptane washing, the
separation of FAME and glycerol after transesterification turned
to be more difficult. The incomplete removal of glycerol and base
catalyst will cause the reverse reaction and re-saponification of
the FAME mixture during distillation process, which generated a
low oil yield of 56.5% in distillation step in R I. R III showed a
slightly lower total biodiesel yield (69.8%) and oil yield in distilla-
tion step (86.6%) compared to R II (71.4% and 88.5%, respectively).
This may be attributed to the adsorptive effect of sludge-activated
carbon in the reflux distillation stage. Despite of this, the combina-
tion of reflux distillation and adsorptive desulfurization exhibited
more efficient and competitive ability to remove sulfur than the
simple reflux distillation.
3.2.3. FAME comparison
The FAME profiles of the final biodiesel obtained in different
processes were summarized in Table 3. The nomenclature ‘C m:n’
characterizes a FAME by the number of carbon atoms m and the
number of double bonds n of the fatty acid. It can be observed that
FAMEs formed with different fatty acids in scum were mainly C14,
C16, and C18, but there were significant difference in their weight
percentages. A large amount of C18 (C18:0, C18:1 and C18:2) were
found in the refined biodiesel with 58.35% in R III, followed by
46.12% in R II and 35.58% in R I. The weight percentages of C16

Table 3
FAME comparison of the refined biodiesel obtained in different processes.

Component Weight percentage (wt%)

RI R II R III SBCa
C 14:0 methyl myristate 9.51 2.98 3.47 5.63
C 14:1 methyl myristoleate 0.53 0.19 0.21 –
C 16:0 methyl palmitate 21.39 22.25 21.56 32.45
C 16:1 methyl palmitoleate 17.88 14.07 13.17 2.27
C 18:0 methyl stearate 7.69 10.13 8.90 15.59
C 18:1 methyl oleate 18.87 27.24 37.71 41.35b
C 18:2 methyl linoleate 9.02 8.75 11.74
Total 84.89 85.61 96.76 97.29
a
The scum-to-biodiesel conversion process reported by Bi et al. (2015) consisted
of six steps, including filtering, acid washing, acid esterification, base transester-
ification, glycerol washing and oil refining. This process is denoted by SBC.
b
The total weight percentage of C 18:1 and C 18:2. Fig. 5. FAME comparison of refined biodiesel with different processes.
 H. Ma et al. / Bioresource Technology 214 (2016) 826–835
(C16:0 and C16:1) in the three routes showed only slight differ-
ences with 34.77% in R III, 36.32% in R II and 39.27% in R I. In com-
parison with R II and III, the refined biodiesel of R I contained
10.03% C14 (C14:0 and C14:1), which was 3.2 times and 2.7 times
higher than that of R II and III, respectively. Most of C14 was found
in the light cut of biodiesel distilled in the temperature range from
180 °C to 200 °C. Interestingly, the sulfur content of the light cut in
R I (66.3 ppm) was also higher than that in R II (33.4 ppm) and R III
(48.5 ppm), as described in Section 3.2.1. Therefore, it was sus-
pected that a part of myristic acid (C14:0) bonded with impurities
such as organosulphur compounds might be removed in heptane
washing step, resulting in the relative low contents of C14 and sul-
fur in the light cut of R II and III. The total amount of main FAMEs in
R III accounted for 96.76%, which was higher than that of R I
(84.89%) and II (85.61%). This is probably because that a lot of
impurities mixed in the FAME vapor were removed by the sludge
activated carbon, when the FAME mixture was refined under the
reflux distillation condition.
Compared with R III, the previous process (denoted by SBC)
reported by Bi et al. (2015) displayed a similar FAME profile as fol-
lows: 5.63% C14, 34.73% C16, 56.94% C18 and 97.29% for total
amount. However, distinct differences were noticed in terms of
the saturated FAMEs and unsaturated FAMEs. As shown in Table 3,
the FAME mixture in SBC exhibited much higher weight percent-
ages of C16:0 (32.45%) and C18:0 (15.59%) than those of the three
routes in the present work; but the C16:1 in SBC (2.27%) was sig-
nificantly lower than that in the present work (13.17–17.88%).
Accordingly, as illustrated in Fig. 5, SBC process generated more
saturated FAMEs than unsaturated FAMEs, while R I, II and III
yielded more unsaturated FAMEs than saturated FAMEs. A great
amount of unsaturated FAMEs accounted for 62.83% was observed
in R III, followed by 50.25% in R II, 46.3% in R I and 43.57% in SBC. In
contrast, SBC process had a much higher saturated FAMEs
accounted for 53.67% than all three process routes in the present
work (33.92–38.60%). As suggested by previous studies (Knothe
et al., 2015; Lanjekar and Deshmukh, 2016; Mandotra et al.,
2016), biodiesel properties are directly affected by the structure
of fatty acid methyl esters, such as chain length, number of double
bonds, and the relative concentration of the saturated and unsatu-
rated FAME components. Thus, several properties of the refined
biodiesel obtained in this work were calculated based on the FAME
profiles in Table 3 or determined below, including cetane number
(CN), kinematic viscosity, cloud point and acid value.
3.3. Fuel properties of the scum-derived biodiesel from different
processes
Numerous studies indicate that the fuel properties of biodiesel
are largely dependent upon the fatty acid methyl ester composi-
tion (Knothe et al., 2015; Lanjekar and Deshmukh, 2016;
Table 4
CNs and KVs (40 °C; mm2/s) of neat FAMEs reported in the previous literatures.
Number of C in fatty acid CNa KVb
chain Number of double Number of double
bonds bonds
0 1 2 0 1 2
C 14 72 3.30 2.73
C 16 85.9 52 32 4.38 3.67
(D9)
C 18 101 59.3 38.2 5.85 4.51 3.65
(D9)
a
Data for C 14, C16 and C18 were from reports by Freedman et al. (1990) and
Knothe (2012).
b
Data for C 14, C16 and C18 were from Knothe and Steidley (2005, 2011).
833
Table 5
Prediction of CNs and KVs (40 °C; mm2/s) of biodiesel derived from scum in present
work and the specifications in biodiesel standard.
Property Scum-to-biodiesel conversion Biodiesel standard
R Ia R IIa R IIIa SBCb ASTM D6751c EN 14214d
CN 56.93 58.31 63.69 – 47 min 51 min
KV 3.55 3.74 4.19 4.94 1.9–6.0 3.5–5.0
a
KV values for R I, II and III were predicted using Eq. (2).
b
The scum-to-biodiesel conversion process reported by Bi et al. (2015) is
denoted by SBC. KV value for SBC was determined according to ASTM D445.
c
Data was from ASTM D6751 (2015).
d
Data was from EN 14212 (2009).
Mandotra et al., 2016). Some critical properties of pure FAME could
be used to predict the corresponding properties of biodiesel com-
prising FAME mixture (Freedman et al., 1990; Knothe and
Steidley, 2011). Therefore, the relevant equations and models were
developed to calculate or predicate the fuel properties, including
cetane number and kinematic viscosity. In this study, the CN and
kinematic viscosity were calculated to evaluate the effect of differ-
ent processes on the biodiesel qualities; the effect of saturated
fatty esters on cold flow properties was discussed and acid value
of biodiesel was determined.
3.3.1. Cetane number (CN) and Kinematic viscosity (KV)
The cetane number (CN), being related to ignition and combus-
tion behavior, is a prime indicator of the quality of diesel fuels,
including biodiesel derived from renewable resources. Cetane
numbers of FAME mixtures can be generally calculated from the
equation (Knothe, 2012; Knothe et al., 2015) as follows:
X
CNmix ¼ AC  CNC ð1Þ
where CNmix is the CN of the mixture, AC = the relative amount (%) of
an individual neat ester in the mixture, and CNC is the CN of the
individual neat ester. Table 4 summarized the determined CN val-
ues of neat FAMEs from previous literatures. According to Eq. (1),
FAME profiles illustrated in Table 3 and CN data given in Table 4,
the CNs of biodiesel samples from different Res were calculated
and listed in Table 5.
As seen in Table 5, the calculated CN of biodiesel from R III was
approximately 63.69, while that of R II and R I were 58.31 and
56.93, respectively. Compared with R I and R II, the relatively
higher CN value of biodiesel from R III was attributed to its higher
percentages of C18:1 and C18:2 (FAME). The CN of biodiesel from
SBC process in previous report could not be accurately calculated
because of its unknown percentages of C 18:1 and C 18:2. In spite
of this, it can be predicted that the biodiesel from SBC process may
have a much higher CN than that from this work, which was
mainly due to its great amount of saturated FAME components
(53.67%). As prescribed in biodiesel standard, the CN in ASTM D
6751 (2015) is 47 and in EN 14214 is 51 (2009). Therefore, the
CN in all the biodiesel samples from the three routes meet the
two standards.
Kinematic viscosity (KV; at 40 °C) is another parameter speci-
fied in biodiesel standards. Similar to the discussion of CNs above,
the KV of FAME mixture could also be calculated based on the
viscosity data of neat FAMEs. The KV of mixtures Vmix can be calcu-
lated using the equation (Knothe and Steidley, 2005, 2011) as
follows:
X
V mix ¼ AC  V C ð2Þ
where Vmix is the kinematic viscosity of the biodiesel sample (FAME
mixture), and VC is the kinematic viscosity of the individual com-
pounds in the mixture. The corresponding determined KV values
834 H. Ma et al. / Bioresource Technology 214 (2016) 826–835
of neat FAMEs were given in Table 4. It appears that the viscosity of
the neat FAME increases with the chain length of fatty acid, but
decreases with increasing number of cis double bonds, which was
similar to that of CNs. The KV of biodiesel samples obtained in this
work was calculated according Eq. (2), and then summarized in
Table 5 along with the relevant specifications from the biodiesel
standards ASTM D6751 and EN 14214.
It can be seen that the KV values of biodiesel samples from the
three routes were in the range of 3.5–4.2 mm2/s (Table 5), which
was within the range prescribed in ASTM D6751 (1.9–6.0 mm2/s)
and EN 14212 (3.5–5.0 mm2/s). The KV of biodiesel from SBC pro-
cess could not be accurately calculated because the detailed per-
centage of C18:1 and C18:2 were not given in the previous study.
However, it can be estimated that the viscosity value for SBC pro-
cess would likely be in the range of 4.11–4.47 mm2/s based on the
total percentage of C18:1 and C18:2 shown in Table 3. The corre-
sponding experimental viscosity value for SBC process was
4.94 mm2/s (Table 5). As expected, all the estimated viscosity val-
ues of biodiesel samples from this work (in the range of 3.55–
4.19 mm2/s) were lower than both the determined viscosity value
(4.94 mm2/s) and predicated value (4.11–4.47 mm2/s) of SBC pro-
cess. This is mainly attributed to the higher percentage of saturated
FAME (both C16:0 and C18:0) in the FAME mixture from SBC pro-
cess (in Table 3 and Fig. 5). The lower viscosity value could allevi-
ate operational problems such as poor fuel atomization and engine
deposits, as suggested by previous studies (Knothe and Steidley,
2005, 2011).
3.3.2. Cloud point and acid value
The cloud point (CP) is one critical parameter to evaluate the
cold flow properties of biodiesel. Poor cold flow properties may
cause misfire, delayed startup and the plugging of filters and fuel
lines in engine. The saturated FAMEs have higher melting point
and poor cold flow properties than that of unsaturated FAMEs.
Therefore, the cold flow properties depend largely on the nature
and amount of the higher-melting saturated FAMEs (Abedin
et al., 2016; Lanjekar and Deshmukh, 2016; Moser et al., 2009).
For SBC process, the CP of the biodiesel with approximately
32.45% methyl palmitate (C16:0) and 15.95% methyl stearate
(C18:0) was reported as 13 °C (Bi et al., 2015); while that of soy
biodiesel with approximately 10% methyl palmitate and 5% methyl
stearate was reported as 0 °C (Foglia et al., 1997). As a result, it can
be assumed that the CP of scum biodiesel samples in this work
would be likely between the value of soy and scum biodiesel from
SBC process, because of their moderate saturated FAME contents
(33.92–38.60% as shown in Fig. 5) within the range of soy to scum
biodiesel of SBC process.
Total acid number (TAN) is mainly an indication of the degree of
oxidation and hydrolysis in biodiesel (Wang et al., 2008). The max-
imum TAN value of biodiesel specified in ASTM D675 (2015) is
0.5 mg KOH g1. In this study, the TAN of biodiesel in all three
routes were determined to be 0.05 mg KOH g1 oil, which is much
lower than the ASTM limit as well as that of SBC process
(1.43 mg KOH g1). Biodiesel with a low TAN is considered ‘‘safe’’
for storage and transportation, whereas those with high TAN above
the ASTM specification may cause severe operational problems and
corrosion during storage.
3.4. Future work
The sulfur content of scum largely depends on the source of
wastewater and may also vary with season. Even in the same
plants, scum from different batches may possess different physical
and chemical characteristics. Therefore, to ensure the stable con-
version of biodiesel from high-sulfur content scum, more specific
modifications of the process shown in this work need to be made
in the future. For example, the reflux ratio, the weight and height
of adsorbents in distillation column could be optimized in order
to enhance the adsorptive interaction between the FAME vapors
and adsorbents. In addition, modifications of both the structure
and compositions in adsorbents need to be researched to increase
the sulfur-adsorption capacity and selectivity of activated carbon.
For instance, the sulfur-adsorption capacity could be improved
by introducing more oxygen-containing functional groups on the
surface of activated carbon.
4. Conclusion
Several novel processes to lower the sulfur content in the bio-
diesel converted from high-sulfur content scum were developed
and compared. The results show that the desulfurization perfor-
mance of these processes was improved through combined effect
of heptane extraction and the new distillation system. Almost
70% of the filtered and dried scum was converted to biodiesel hav-
ing low-sulfur content and meeting the ASTM specification. More-
over, the FAME profiles indicated that the refined biodiesel from
the new processes exhibited a higher quality and better properties
than that from the process reported in previous studies.
Acknowledgements
This work was financially supported by grants from the
Legislative-Citizen Commission on Minnesota Resources (LCCMR),
MCES, University of Minnesota Center for Biorefining, the Natural
Science Foundation of Higher Education Institutes of Anhui Pro-
vince, China (Grant No. KJ2014A073) and Anhui Province Natural
Sciences Foundation, China (Grant No. 1508085SQE213).
References
Abedin, M.J., Kalam, M.A., Masjuki, H.H., Sabri, M.F.M., Rahman, S.M.A., Sanjid, A.,
Fattah, I.M.R., 2016. Production of biodiesel from a non-edible source and study
of its combustion, and emission characteristics: a comparative study with B5.
Renewable Energy 88, 20–29.
Anderson, E., Addy, M., Xie, Q., Ma, H., Liu, Y., Cheng, Y., Onuma, N., Chen, P., Ruan, R.,
2016. Glycerin esterification of scum derived free fatty acids for biodiesel
production. Bioresour. Technol. 200, 153–160.
ASTM Standard D1980-87, 1998. Standard Test Method for Acid Value of Fatty Acids
and Polymerized Fatty Acids.
ASTM Standard D5453-12, 2012. Standard Test Method for Determinations of Total
Sulfur in Light Hydrocarbons, Spark Ignition Fuel, Diesel Engine Fuel, and Engine
Oil by Ultraviolet Fluorescence.
ASTM Standard D6751, 2015. Standard Specification for Biodiesel Fuel Blend Stock
(B100) for Middle Distillate fuels.
ASTM Standard E 1756-08, 1756. Standard Test Method for Determination of Total
Solids in Biomass.
Bi, C.-H., Min, M., Nie, Y., Xie, Q.-L., Lu, Q., Deng, X.-Y., Anderson, E., Li, D., Chen, P.,
Ruan, R., 2015. Process development for scum to biodiesel conversion.
Bioresour. Technol. 185, 185–193.
Bu, J., Loh, G., Gwie, C.G., Dewiyanti, S., Tasrif, M., Borgna, A., 2011. Desulfurization
of diesel fuels by selective adsorption on activated carbons: competitive
adsorption of polycyclic aromatic sulfur heterocycles and polycyclic aromatic
hydrocarbons. Chem. Eng. J. 166 (1), 207–217.
Demirbas, A.H., 2009. Inexpensive oil and fats feedstocks for production of
biodiesel. Energy Educ. Sci. Technol. A – Energy Sci. Res. 23 (1–2), 1–13.
EN Standard 14214, 2009. Automotive Fuels-Fatty Acid Methyl Esters (FAME) for
Diesel Engines-Requirements and Test Methods.
Foglia, T.A., Nelson, L.A., Dunn, R.O., Marmer, W.N., 1997. Low-temperature
properties of alkyl esters of tallow and grease. J. Am. Oil Chem. Soc. 74 (8),
951–955.
Freedman, B., Bagby, M.O., Callahan, T.J., Ryan, T.W., 1990. Cetane numbers of fatty
esters, fatty alcohols and triglycerides determined in a constant volume
combustion bomb. SAE Tech. Paper, 0148–7191.
Jung, B.K., Jhung, S.H., 2015. Adsorptive removal of benzothiophene from model
fuel, using modified activated carbons, in presence of diethylether. Fuel 145,
249–255.
Knothe, G., 2012. Fuel properties of highly polyunsaturated fatty acid methyl esters.
Prediction of fuel properties of algal biodiesel. Energy Fuels 26 (8), 5265–5273.
Knothe, G., Steidley, K.R., 2005. Kinematic viscosity of biodiesel fuel components
and related compounds. Influence of compound structure and comparison to
petrodiesel fuel components. Fuel 84 (9), 1059–1065.
 H. Ma et al. / Bioresource Technology 214 (2016) 826–835
Knothe, G., Steidley, K.R., 2011. Kinematic viscosity of fatty acid methyl esters:
prediction, calculated viscosity contribution of esters with unavailable data, and
carbon–oxygen equivalents. Fuel 90 (11), 3217–3224.
Knothe, G., de Castro, M.E.G., Razon, L.F., 2015. Methyl esters (biodiesel) from and
fatty acid profile of Gliricidia sepium seed oil. J. Am. Oil Chem. Soc. 92 (5), 769–
775.
Lanjekar, R.D., Deshmukh, D., 2016. A review of the effect of the composition of
biodiesel on NOx emission, oxidative stability and cold flow properties. Renew.
Sustain. Energy Rev. 54, 1401–1411.
Li, M., Zheng, Y., Chen, Y., Zhu, X., 2014. Biodiesel production from waste cooking oil
using a heterogeneous catalyst from pyrolyzed rice husk. Bioresour. Technol.
154, 345–348.
Mandotra, S.K., Kumar, P., Suseela, M.R., Nayaka, S., Ramteke, P.W., 2016. Evaluation
of fatty acid profile and biodiesel properties of microalga Scenedesmus abundans
under the influence of phosphorus, pH and light intensities. Bioresour. Technol.
201, 222–229.
835
Moser, B.R., Knothe, G., Vaughn, S.F., Isbell, T.A., 2009. Production and evaluation of
biodiesel from field pennycress (Thlaspi arvense L.) Oily. Energy Fuels 23 (8),
4149–4155.
Mu, D., Addy, M., Anderson, E., Chen, P., Ruan, R., 2016. A life cycle assessment and
economic analysis of the scum-to-biodiesel technology in wastewater
treatment plants. Bioresour. Technol. 204, 89–97.
Sivakumar, P., Anbarasu, K., Renganathan, S., 2011. Bio-diesel production by alkali
catalyzed transesterification of dairy waste scum. Fuel 90 (1), 147–151.
Wang, H., Tang, H., Wilson, J., Salley, S.O., Ng, K.Y.S., 2008. Total acid number
determination of biodiesel and biodiesel blends. J. Am. Oil Chem. Soc. 85 (11),
1083–1086.
Wang, Y., Feng, S., Bai, X., Zhao, J., Xia, S., 2016. Scum sludge as a potential feedstock
for biodiesel production from wastewater treatment plants. Waste Manag. 47
(Pt A), 91–97.