Article

pubs.acs.org/IECR

Esterification of Free Fatty Acids in Used Cooking Oil Using Ion-

Exchange Resins as Catalysts: An Efficient Pretreatment Method for
Biodiesel Feedstock
Sumaiya Zainal Abidin,†,‡ Kathleen F. Haigh,† and Basudeb Saha*,§
†
Department of Chemical Engineering, Loughborough University, Loughborough, Leicestershire, LE11 3TU, United Kingdom
‡
Faculty of Chemical and Natural Resources Engineering, Universiti Malaysia Pahang, Lebuhraya Tun Razak, 26300 Gambang,
Kuantan, Pahang Darul Makmur, Malaysia
§
Department of Applied Science, Faculty of Engineering, Science and the Built Environment, London South Bank University,
London, SE1 0AA, United Kingdom

ABSTRACT: The esterification of used cooking oil (UCO) with methanol was studied using different types of ion-exchange
resins, that is, Purolite D5081, Purolite D5082, and Amberlyst 36. Several catalyst characterization analyses (elemental analysis,
surface area measurement, particle size distribution analysis, scanning electron microscopy analysis, true density measurement,
and acid capacity analysis) have been conducted in the screening stage. Of all of the catalysts investigated, Purolite D5081 resin
showed the best catalytic performance and was selected for further experimental studies. The esterification process was carried
out in a jacketed stirred batch reactor for 8 h. Elimination of mass transfer resistances and the effect of catalyst loading (0.5−1.5%
w/w), reaction temperature (50−65 °C), and methanol to UCO feed mole ratio (4:1−12:1) on the conversion of FFAs were
investigated. The highest FFAs conversion was found to be 92%, at a catalyst loading of 1.25% w/w, 60 °C reaction temperature,
6:1 methanol to UCO molar ratio, and stirring speed of 475 rpm. During the reusability study, the conversion of catalyst dropped
by 8−10% after each reutilization cycle. Several experiments have been conducted through the homogeneous contribution study,
and the results confirmed that both resin pore blockage and sulfur leaching are dominant factors that decrease the catalytic
performance of Purolite D5081 ion-exchange resin.

1. INTRODUCTION convert the triglycerides into a mixture of esters and glycerine,

A majority of the world’s energy is supplied through
petrochemical sources, coal, and natural gases. However, with
a high energy demand from various sectors, global warming, an
unstable price, and finite availability of petroleum-based oil as
well as environmental pollution due to the widespread use of
fossil fuels make the petroleum-based energy unreliable. The
world energy forum has also predicted that, in less than 10
decades, fossil-based oil, including coal and natural gases, will
be depleted.1 Therefore, it is increasingly necessary to develop
renewable energy resources to replace the traditional sources.
Vegetable oil has become one of the most popular feedstocks
for renewable fuel nowadays and is widely available from a
variety of sources. Biodiesel, in particular, has the following
advantages over diesel fuel: it produces less smoke and
particulates, has higher cetane numbers, produces lower carbon
monoxide and hydrocarbon emissions, is biodegradable and
nontoxic, and provides better performance in engine lubricity as
compared to conventional diesel fuels.2 Conversely, depending
on the sources of feedstock, they present other technical
challenges such as low volatility, high pour points, high cloud
points and high cold filter plugging temperatures, elevated NOx
emissions, and incomplete combustion.3,4 There is also a major
concern regarding the unfavorable cold flow properties of
biodiesel as it tends to gel (freeze) at lower temperature, thus
affecting the driveability of biodiesel vehicles.4
Transesterification is the most popular method to produce
biodiesel. It follows three consecutive equilibrium reactions that
in the presence of a catalyst and alcohol. In recent years, the
major obstacle for biodiesel’s commercialization from vegetable
oil is the high cost of raw feedstock. Therefore, cheap nonedible
vegetable oil, animal fats, and waste oils have been found to be
effective feedstock replacements to reduce the cost of
biodiesel.5−10 According to Balat,11 the price of used cooking
oil (UCO) is 2.5−3.5 times cheaper than that of virgin
vegetable oils; thus it can significantly reduce the production
cost of biodiesel. However, these oils contain a significant
amount of free fatty acids (FFAs); for example, crude mahua oil
contains about 13−20% FFAs,12,13 crude jatropha oil contains
about 12−14% FFAs,14,15 crude tobacco seeds oil contains
about 17% FFAs,16 crude Zanthoxylum bungeanum seed oil
(ZSO) contains about 25% FFAs,17 and waste oil contains 6−
15% by weight of FFAs.18,19 Oils and fats with a high FFAs
content cannot be directly used in a base-catalyzed trans-
esterification reaction. A study by Naik et al.20 showed a
significant drop in the conversion of ester when the FFAs
content was above 2%. The FFAs react with the alkali catalyst
to form soap. The saponification process forms a stable
emulsion and leads to difficulties in process separation and
reutilization of catalyst. In addition, it also creates a significant
Received: March 21, 2012
Revised: October 4, 2012
Accepted: October 17, 2012
Published: October 17, 2012

© 2012 American Chemical Society 14653 dx.doi.org/10.1021/ie3007566 | Ind. Eng. Chem. Res. 2012, 51, 14653−14664
Industrial & Engineering Chemistry Research
amount of toxic wastewater through the neutralization process.
Several studies on the utilization of waste oils as raw material
have been conducted,21,22 and it has been found that high yield
is achieved using a two-step synthesis of biodiesel. The first step
is the pretreatment step (esterification process) required to
reduce the amount of FFAs in the feedstock before being
subjected to a transesterification reaction, that is, the second
step.
Currently, most biodiesel processes use homogeneous acid
catalysts (e.g., sulphuric acid, hydrochloric acid) to esterify
FFAs in waste oils because they are effective and soluble in
methanol. These homogeneous catalysts have a number of
drawbacks: corrosion of equipment, existence of side reactions,
generation of a substantial amount of wastewater, a high
production cost due to additional equipment/reagents for
separation/neutralization processes, and difficulties in catalyst
recovery.23,24 The use of heterogeneous catalysts has proven to
simplify the production and purification processes because they
can be easily separated from the reaction mixture. This allows
multiple usage of the catalyst through regeneration, resulting in
a reduction in waste and a reduction in the environmental
impact.
Heterogeneous catalysts can be classified into several
different categories, single and mixed metal oxides, hydro-
talcites, heteropolyacids, ion-exchange resins, zeolites, sulph-
ated metal oxides, and organic acid catalysts. Ion-exchange
resins are more popular because they can catalyze the reaction
using mild conditions due to the high concentration of acid
sites.25 Most of the ion-exchange resins used in biodiesel
production are based on cross-linked polystyrene-divinyl
benzene copolymers bearing the sulfonic acid group
(−SO3H). In the past few years, several studies have been
reported on the development of ion-exchange resins as catalyst
in the biodiesel process.16,26,27 Cation-exchange resins are
preferable in esterification reaction as they offer better
selectivity toward the desired products.28,29 Ö zbay et al.30
reported the esterification of FFAs in waste cooking oil using
various types of ion-exchange resin catalysts, but the FFAs
conversion was relatively very low, with a conversion of about
45.7% at the optimum conditions (20% v/v methanol, 60 °C,
2% w/w catalyst, 180 min). On the other hand, a study by Feng
et al.26 showed a 90% conversion of FFAs when a cation-
exchange resin, NKC-9, was used as the catalyst in biodiesel
esterification. The optimized conditions were achieved in 3 h
using 18% w/w catalyst and 66 °C reaction temperature.
In this study, the utilization of ion-exchange resins in the
pretreatment stage of the UCO has been explored, using novel
catalysts Purolite D5081 and Purolite D5082. These catalysts
were chosen because they were developed specifically for the
esterification of edible oil for biodiesel production. This work
has been conducted in collaboration with Purolite International
Limited for possible commercialization of these catalysts.
Previously, Amberlyst BD20 resin has been proposed as the
comparison catalyst, but due to commercial reasons, Rohm and
Haas were not able to supply it and recommended Amberlyst
36 as an alternative. The main goal of this research work is to
identify the best catalyst for esterification process and to
determine the optimum conditions for the esterification process
prior to the transesterification. Several catalyst characterization
analyses (elemental analysis, surface area measurement, particle
size distribution analysis, scanning electron microscopy analysis,
true density measurement, and acid capacity analysis) were
conducted, and comparison between various ion-exchange
14654
Article
resins on the FFAs conversion was carried out in the screening
stage. Ion-exchange resin with the highest conversion of FFAs
was chosen for further experimental studies. Mass transfer
resistances as well as the effect of catalyst loading, reaction
temperature, and methanol to UCO molar ratio on the
conversion of FFAs have been studied. Analysis of FFAs has
been carried out using the ASTM D974 standard titration
method.
2. EXPERIMENTAL METHODS
2.1. Materials. The UCO was supplied by Greenfuel Oil
Co. Limited, UK. The UCO has an acid value of approximately
12.2 mg KOH/g and a FFAs value of approximately 6.1%.
Methanol (>99.5% purity), 2-propanol, toluene, 0.1 M
volumetric standard hydrochloric acid, 0.1 M of volumetric
standard sodium hydroxide, 0.1 M standardized solution of
potassium hydroxide in 2-propanol, p-naphtholbenzein, methyl
red, and n-hexane (≥99% purity) were purchased from Sigma
Aldrich, UK and Fisher Scientific, UK and were used as
received. Pure standards of fatty acid methyl ester (methyl
palmitate, methyl heptadecanoate, methyl stearate, methyl
oleate, methyl linoleate, methyl linolenate) were purchased
from Sigma Aldrich, UK with over than 99% purity. Ion-
exchange resin catalysts (Purolite D5081 and Purolite D5082)
were supplied by Purolite International Limited, and Amberlyst
36 was purchased from Sigma Aldrich, UK. These resins were
pretreated before being used as the reaction catalysts. The
pretreatment method used in this research work is a common
method used by most researchers in esterification of FFAs
using ion-exchange resins.31−35 Basically, dry catalyst does not
require any pretreatment as the active sites are already free and
available on the resin’s surface and inside the pores, whereas the
wet catalyst usually goes through the preconditioning process
with solvent, is rinsed, and then dried in an oven for a certain
period of time.31 All wet resins were immersed in methanol
overnight and pretreated with the same solvent under
ultrasonic condition. The process takes about 8−10 cycles of
rinsing steps to ensure that all impurities are totally removed
(the color of filtrate turned from light brownish to colorless
solution). During this washing process, the conductivity of the
residual solution was recorded, and the process continues until
the conductivity of the residual solution is approximately the
same with the solvent (methanol). Finally, resins were dried in
a vacuum oven at 100 °C for 6 h to completely remove water
and methanol. All other chemicals used were analytical reagent
grade.
2.2. Used Cooking Oil (UCO) and Product Character-
ization. 2.2.1. Average Molecular Mass Determination:
Analysis of Fatty Acid Composition in the UCO. The fatty
acids bonded to the glycerine backbone vary depending on the
oil type, and as a result an average molecular mass is generally
determined on the basis of the fatty acid composition of the oil.
Average molecular mass was calculated by multiplying the mass
fractions of fatty acids presence in the oil with the individual
molecular mass of each fatty acids involved. The determination
of fatty acid composition was done by converting the
triglycerides to glycerine and fatty acid methyl ester (FAME)
through a methylation or hydrolysis process.36 Derivatization
through the methylation process has been widely used to
characterize lipid fractions in fats and oil.37,38 It is a well-
accepted characterization method because of the robustness
and reproducibility of the chromatographic data. These
dx.doi.org/10.1021/ie3007566 | Ind. Eng. Chem. Res. 2012, 51, 14653−14664
Industrial & Engineering Chemistry Research
methods are also cheaper in terms of reagent usage and do not
require expensive equipment.
In this study, the sample was prepared using methods in
European Union (EU) Regulation,39 and the results were
verified using British Standard Institution (BSI) standard
method, EN ISO 12966-2:2011.40 Derivatisation process begins
by weighing 100 mg of UCO in a 20 mL screw-cap test tube or
reaction vial. Next, the UCO was dissolved in 10 mL of n-
hexane. 100 μL of 2 N potassium hydroxide was added to the
reaction vial together with 100 mL of methanol. The tube or
vial was closed and mixed vigorously for 60 s. The sample was
then transferred into a conical bottom tube for centrifugation
process. This process takes about 10 min with 16 000 rpm
rotational speed. The upper layer, the clear supernatant, was
analyzed by GC−MS (GC−MS Hewlett-Packard HP-6890),
with DB-WAX (J&W Scientific) capillary column, 30 m length
and 0.25 mm inner diameter, and packed with polyethylene
glycol (0.25 μm film thickness). Helium was used as a carrier
gas at a constant flow rate of 1.1 mL/min. The amount of
sample injected was 1 μL. The temperature of the injector and
detector was 250 °C. The initial oven temperature was 70 °C
held for 2 min, programmed at 40 °C/min to 210 °C,
programmed at 7 °C/min to 230 °C, final temperature held for
11 min. Methyl heptadecanoate was used as the internal
standard.
2.2.2. Free Fatty Acids (FFAs) Analysis. The acid value of the
UCO was determined using ASTM D974, which was
specifically used to determine the acidic or basic constituents
in highly colored fats and oils. The determination of acid or
base number was carried out by dissolving the sample in a
mixture of toluene, 2-propanol, and a small amount of water to
form a single phase solution. The resulting solution was then
titrated at room temperature with an alcoholic base solution,
and the end point is detected by changes of the indicator, p-
naphtholbenzein, color from orange to green. The conversion
of FFAs was calculated using the following equation:
CA 0 − CA
conversion of FFAs, XA =
CA 0
where CA0 is the initial concentration of FFAs of the mixture
(i.e., at time t = 0) and CA is the concentration of FFAs at any
time t.
2.3. Catalyst Characterization. 2.3.1. Elemental Analysis.
Carbon, hydrogen, and nitrogen determination was carried out
using an Exeter Analytical CE440 C, H, N analyzer. It works by
combusting the sample in pure oxygen with tin as a catalyst.
Any sulfur, phosphorus, or halogens were removed and excess
oxygen was reduced, leaving only CO2, H2O, and N2 as the
combustion products. These gases are then analyzed for
quantity using a system of detectors and scrubbers. The sulfur
determination was carried out separately using oxygen flask
combustion analysis, followed by a titration. All of the results
were reported in weight percentage of carbon, nitrogen,
hydrogen, and sulfur, and the residual was assumed to be
oxygen. All samples for elemental analysis were analyzed in
duplicate, and the error was less than 3%.
2.3.2. Surface Area Measurement. Surface area, pore
volume, and average pore diameter were determined from the
nitrogen adsorption and desorption isotherms using a Micro-
meritics ASAP 2020 surface analyzer. The samples were
degassed at a two-stage temperature ramping, followed by
sample analysis at −195.15 °C using nitrogen gas. The
14655
Article
Brunauer−Emmett−Teller (BET) method was used to
calculate the surface area.
2.3.3. True Density Measurement. The true density (ρt) of
the sample was measured using a Micromeritics helium
pycnometer 1305. Samples with known mass (m) were placed
into the pycnometer cell. After that, a constant flow of helium
gas was introduced into the cell several times to ensure that
void spaces between the particles were filled by helium gas. The
actual volume of sample (v) was calculated on the basis of the
known parameters, that is, cell volume, expansion medium
volume, and the charge pressure before and after expansion.
Finally, the true density of particles was determined using the
standard formula: true density, ρt = (m/v). All samples were
measured in duplicate, and the error was less than 1%.
2.3.4. Particle Size Distribution (PSD) Analysis. Particle size
distribution (PSD) analysis was performed with a Coulter LS
130 Particle Analyzer with particle size measurement over the
range 0.1−1000 μm. The Fraunhofer optical model was used
for the measurement of the distribution pattern. The
equipment operating parameters were set to record the
background measurement for 60 s followed by sample
measurement for 60 s. The samples were then introduced
into the dispersion module with isopropyl alcohol as the
solvent, and a total of two measurement cycles were carried out
for each sample and the average was taken as the final result.
2.3.5. Acid Capacity Analysis. Acid capacity determination
was conducted using a conventional titration method. The
analysis took place in a 50 mL Erlenmeyer flask with the ion-
exchange resins (∼0.5 g) contacted with 50 mL of 0.1 M of
volumetric standard sodium hydroxide. The flasks were sealed
with Parafilm and shaken for 72 h in an orbital shaker. After
that, the solution was filtered, and 5 mL aliquots were back-
titrated with 0.1 M volumetric standard hydrochloric acid using
methyl red as an indicator. All samples were analyzed in
duplicate, and the error was less than 3%.
2.3.6. Scanning Electron Microscopy (SEM) Analysis. A
scanning electron microscope (Carl Zeiss (Leo/Cambridge)
Stereoscan360) was used to study the morphology of the ion-
exchange resins. The pretreated catalysts were preconditioned
(drying at 100 °C for 6 h and stored in desiccators) prior to
analysis to eliminate moisture. After that, the samples were
mounted on a metal stub using a carbon conductive pad, and
silver metals were placed at both sides of the samples to act as a
conductor. Finally, the samples were coated with gold under
vacuum condition in an argon atmosphere prior to observation.
2.4. Experimental Method. The esterification process was
carried out in a four-neck 1000 mL cylindrical jacketed-glass
reactor, equipped with a mechanical stirrer, sampling outlet,
and reflux condenser to prevent the loss of reactant due to
vaporisation. Heating was achieved by circulating water from a
water bath and through the reactor, and a thermocouple was
used for temperature monitoring. Figure 1 shows the
experimental setup of the reaction process, and Schemes 1
and 2 shows the reaction scheme of the process.
A specified amount of UCO and methanol was added to the
reactor, and the stirring and heating of the reaction mixtures
were started. When the reactor reached the required temper-
ature, catalyst was added, and this point was taken as the zero
time for the reaction. The samples were periodically taken from
the reactor for FFAs analysis. After 8 h, the reaction mixture
was transferred to a separation funnel and allowed to settle
overnight to form two layers; the top layer consisted of excess
methanol and its impurities, whereas the bottom layer was
dx.doi.org/10.1021/ie3007566 | Ind. Eng. Chem. Res. 2012, 51, 14653−14664
Industrial & Engineering Chemistry Research
Figure 1. Experimental setup for the reaction process.
mainly unreacted UCO together with traces of methanol,
glycerine, esters, and the remaining catalyst. The bottom layer
was withdrawn from the separating funnel together with the
catalyst, and the retained catalyst was washed, dried, and stored
for further experimental work. Studies on the mass transfer
resistance as well as the effect of methanol to UCO molar ratio
(4:1, 6:1, 9:1, 12:1), catalyst loading (0.5, 0.75, 1, 1.25, 1.5% w/
w), and reaction temperature (50, 55, 60, 62, 65 °C) have been
conducted.
In terms of the reproducibility of experimental data, selected
experiments have been repeated three times, and it was found
that there was a 1−2% difference in the results. Therefore, it
was assumed that a similar error applies to all results. In
addition, a blank run without catalyst has been performed, and
there was no conversion of FFAs after 8 h of reaction time.
Therefore, it was concluded that the esterification of FFAs
occurs only due to the presence of the catalyst.
2.5. Reusability Study Including an Investigation of
Homogenous Contribution. A detailed reusability study of
Purolite D5081 catalyst with the highest conversion was
conducted to determine the catalyst life span. The first phase of
this work was to investigate the effect of reusing the same batch
of a catalyst for a number of cycles on the conversion of FFAs.
For this purpose, used catalyst was washed with methanol with
the aid of an ultrasonic bath until there were no traces of oil
and colorless solution was obtained. During this washing
process, the conductivity of the residual solution was recorded,
and the process continues until the conductivity of the residual
solution is approximately the same with the solvent
(methanol). The washed catalyst was filtered and dried in a
vacuum oven at 100 °C for 6 h. The reusability study was
carried out under the optimum process conditions. The fresh
Scheme 1. General Reaction Scheme for Biodiesel Productiona
a
R1, R2, and R3 represent the fatty acids groups attached to the backbone of triglycerides.
14656
Article
Scheme 2. Reaction Scheme for the Esterification Process:
Conversion of FFAs to FAMEa
a
R1 represents the fatty acid group.
and used catalysts were tested for SEM, SEM-EDX, acid
capacity, and elemental analyses.
An investigation into the homogeneous contribution was
carried out to establish if sulfonic acid groups were leached
during the reaction process. The process flow diagram for the
reusability study is summarized in Figure 2. It was done by
reacting fresh methanol with unused pretreated catalyst for 8 h
at 60 °C and 475 rpm stirring speed. After that, the catalyst was
filtered, and the solution was used in the subsequent reaction
with fresh UCO (6:1 methanol:UCO molar ratio), in the
absence of any catalyst. This is referred to as the first cycle of
homogeneous contribution study. The filtered catalyst was
dried in the vacuum oven for 6 h (100 °C), and this catalyst
was then used for an experiment with fresh methanol and fresh
UCO at optimum condition to monitor the conversion trend as
a result of the catalyst being treated with methanol. This result
was plotted as the first cycle of the methanol treated catalyst
study.
The spent catalyst from the previous methanol treated
catalyst experiment was washed thoroughly with methanol
using an ultrasonic bath until there was no evidence of UCO
contamination. The catalyst was filtered and dried in a vacuum
oven at 100 °C for 6 h. The experimental work proceeded to
the second homogeneous cycle, where the reaction took place
between the fresh methanol and used catalyst. The used catalyst
was the same catalyst used in the first cycle of homogeneous
contribution study and the first cycle of methanol treated
catalyst study. The filtered solution from the methanol−catalyst
reaction was added to fresh UCO (6:1 methanol:UCO molar
ratio) without the presence of catalyst, and the FFAs
conversion was monitored. The result was plotted as the
second cycle of the homogeneous contribution study. The used
catalyst obtained from the methanol−catalyst reaction was
dried (100 °C, 6 h) and introduced to fresh UCO and fresh
methanol. By using the optimum reaction parameters, the
esterification was conducted, and the second cycle of methanol
treated catalyst data was plotted.
dx.doi.org/10.1021/ie3007566 | Ind. Eng. Chem. Res. 2012, 51, 14653−14664
Industrial & Engineering Chemistry Research
Figure 2. Process flow diagram for catalyst reusability study.
Figure 3. Chromatogram of the derivatized UCO.
3. RESULTS AND DISCUSSION
3.1. GC−MS Analysis of Derivatized Used Cooking Oil
(UCO). Figure 3 shows a typical chromatogram of the
derivatized UCO. From the chromatogram, six different
components (including the reference standard) were identified.
14657
Article
The retention times for each individual component are as
follows: methyl palmitate (C16:0) appeared at 8.324 min
retention time, methyl heptadecanoate (C17:0, reference
standard) at 8.950 min, methyl stearate (C18:0) at 9.712
min, methyl oleate (C18:1) at 9.914 min, methyl linoleate
dx.doi.org/10.1021/ie3007566 | Ind. Eng. Chem. Res. 2012, 51, 14653−14664
Industrial & Engineering Chemistry Research Article

(C18:2) at 10.349 min, and finally methyl linolenate (C18:3) at Table 3. Properties and Characteristics of Various Ion-
11.005 min. As the methylation process converts fatty acids to Exchange Resins
methyl esters through derivatization method, it could be
catalyst properties Purolite D5081 Purolite D5082 Amberlyst 36
concluded that there were five main components present in
UCO, the palmitic acid, stearic acid, oleic acid, linoleic acid, and physical appearance black spherical black spherical black spherical
beads beads beads
linolenic acid.
matrix hyper-cross- hyper-cross- macroporous
The response factor for each component was determined linked linked
using the calibration of individual standards. The predeter- cross-linking level highly cross- highly cross- medium cross-
mined response factor was then used to determine the fatty acid linked linked linked
composition of the sample. By calculating the mass fractions of particle size
each fatty acid, the average molecular mass of the fatty acids distribution (μm)
could be easily calculated and will be further used to determine 10%> 638.6 806.8 741.1
the methanol to UCO molar ratio.17 The fatty acids 50%> 496.5 644.8 575.8
composition is summarized in Table 1. These data clearly 90%> 396.4 482.1 341.1
nitrogen BET
BET surface area 514.18 459.62 30.0
Table 1. Percent Composition of Fatty Acid in Used (m2/g)
Cooking Oil (UCO) total pore volume 0.47 0.36 0.19
(cm3/g)
component % composition (w/w)
average pore 36.9 31.4 254.2
palmitic acid (C16:0) 11.34 diameter (Å)
stearic acid (C18:0) 3.18 true density (g/cm3) 1.311 1.375 1.568
oleic acid (C18:1) 43.95 acid capacity 1.59 1.79 5.00
linoleic acid (C18:2) 36.44 (mmol/g)
linolenic acid (C18:3) 5.09 swelling (media: negligible negligible negligible
methanol)

show that oleic and linoleic acids comprise more than 80% of
the total fatty acids in the sample. Using the European Union
(EU) Regulation39 method, the average molecular mass of fatty
acids is 278.11 g/mol. The reliability of the previous standard
method was verified by analyzing the sample according to the
British Standard Institution (BSI) method, EN ISO 12966-
2:2011.40 It was found that the results were very similar and the
average molecular mass of fatty acids obtained using the second
method was 277.93 g/mol. Therefore, it can be concluded that
both methods are comparable.
3.2. Catalyst Characterization. 3.2.1. Elemental Analysis.
The results for elemental analysis are presented in Table 2. For
Purolite D5082 (459.6 m2/g) and Amberlyst 36 (30 m2/g).
Table 3 also shows that Purolite D5081 has the largest pore
volume (0.47 cm3/g) among all of the resins analyzed.
3.2.3. True Density Measurement and Particle Size
Distribution Analysis. The measured true densities for all
resins are as follows: 1.311, 1.375, and 1.568 g/cm3 for Purolite
D5081, Purolite D5082, and Amberlyst 36, respectively.
Purolite D5082 and Amberlyst 36 have a similar distribution
of particle size, whereas Purolite D5081 has slightly smaller
particles, and this information is shown in Figure 4. Amberlyst

Table 2. Elemental Analysis for Ion-Exchange Resinsa

catalyst %C %H %N %S % Ob
Purolite D5081 77.04 5.32 0.95 4.09 12.61
Purolite D5082 68.87 4.44 0.13 5.92 20.65
Amberlyst 36 42.18 4.10 0.10 18.27 35.35
a
All percentages are in w/w %. bOxygen by difference.

all ion-exchange resins, it was assumed that the only other

species present was oxygen. Oxygen cannot be measured by
elemental analysis, and, therefore, the percentage of oxygen
content was determined by the difference from the total weight
percentage of other elements (i.e., carbon, hydrogen, nitrogen,
and sulfur). For cation-exchange resins, the polymer structure
consists of carbon, hydrogen, sulfur, and oxygen elements.
There was an unexpected presence of nitrogen in the cation-
exchange resins, and the value was less than 1%. In this case,
nitrogen was assumed to be a contaminant in the sample. It can
be seen from Table 2 that Amberlyst 36 has the highest
percentage of oxygen and sulfur as compared to the other
resins.
3.2.2. Surface Area Measurement. The results for surface
area, average pore diameter, and total volume are presented in
Table 3. Purolite D5081 catalyst shows the highest Brunauer−
Emmett−Teller (BET) surface area (514.2 m2/g), followed by
14658
Figure 4. Particle size distribution of ion-exchange resins catalysts.
36 has the widest size distribution, with particle ranges between
100 and 900 μm and a secondary peak. The two resins from
Purolite have narrower distributions, where the range lies
between 375 and 900 μm for Purolite D5082 and 280 and 800
μm for Purolite D5081. Chemical and physical characteristics
for all resins are summarized in Table 3.
3.3. Optimization of Esterification Process in a
Jacketed Batch Reactor. 3.3.1. Screening of Ion-Exchange
Resins. To identify the best of the three ion-exchange resins for
further experimental work, all three resins, Purolite D5081,
dx.doi.org/10.1021/ie3007566 | Ind. Eng. Chem. Res. 2012, 51, 14653−14664
Industrial & Engineering Chemistry Research Article

Purolite D5082, and Amberlyst 36, were evaluated under the reach the equilibrium. However, these catalysts are highly
same reaction conditions, that is, at 1% w/w of catalyst loading, porous, and as a result this effect is very small when compared
6:1 methanol to UCO molar ratio, 60 °C reaction temperature, to the effect of the catalysts pore volume.
and 350 rpm impeller stirring speed. From Figure 5, it can be Although Amberlyst 36 has the highest sulfur content and
the largest average pore diameter, the result from the
esterification reaction (Figure 5) shows that Amberlyst 36 has
the lowest conversion as compared to Purolite D5081 and
D5082 resins. This is because Amberlyst 36 has the lowest
specific surface area and lowest pore volume as compared to the
Purolite resins, and, therefore, there are less active catalytic sites
available for the reaction to occur. The level of DVB cross-
linking also contributes significantly to the level of FFAs
conversion. From Figure 5, it could be seen that resins with
high DVB cross-linking (Purolite D5081 and D5082) result in
higher FFAs conversion as compared to lower DVB cross-
linking resin (Amberlyst 36). The FT-IR analysis has also been
carried out, and there was no noticeable change of functional
groups for all of the resin samples, and, therefore, these results
are not presented in this Article. The FT-IR results also indicate
Figure 5. Effects of different types of ion-exchange resins on FFAs that sulfur was present only in sulfonic acid functional group.
conversion (experimental conditions: molar ratio (methanol:UCO), Because Purolite D5081 showed the best catalytic performance,
6:1; catalyst loading, 1% w/w; stirring speed, 350 rpm; reaction
it was used for further experimental work.
temperature, 60 °C).
3.3.2. Investigation on the Effect of Mass Transfer
seen that, after 8 h, Purolite D5081 resin achieved the highest Resistance. There are two types of mass transfer resistances
FFAs conversion of ∼88%, while Purolite D5082 and involved in ion-exchange catalysis. The first is external mass
Amberlyst 36 achieved FFAs conversion of ∼78% and ∼30%, transfer resistance, which takes place across the solid−liquid
respectively. interface, while the second is the internal mass transfer
The differences in the properties of various resins, that is, resistance, associated with the differences in particle size
surface area measurement, elemental analysis, and acid capacity distribution of the catalysts. Mixing is one of the key factors to
analysis, can be used to explain the differences in catalytic optimize the production of biodiesel as it increases the
activity (Tables 3−5). These analyses show that even though interaction between the reactants (methanol and UCO) and
the catalyst, predominantly at the early stage of the reaction.
Table 4. Elemental Analysis for Fresh and Used Ion- However, after the reaction mixture reaches the stage where the
Exchange Resinsa reactant and the catalyst are well-mixed (e.g., there is sufficient
contact between the catalyst and the reactant), there is no
catalyst %C %H %N %S % Ob additional benefit from increasing the stirring speed. This
fresh D5081 77.04 5.32 0.95 4.09 12.61 phenomenon is due to the external mass transfer resistances.
used D5081c 77.41 5.69 0.93 3.32 12.66 To investigate the influence of external mass transfer
fresh D5082 68.87 4.44 0.13 5.92 20.65 resistance, three different stirring rates were investigated for
used D5082c 69.07 4.53 0.03 5.41 20.97 the reaction process, and the conversion patterns were
fresh Amberlyst 36 42.18 4.10 0.10 18.27 35.35 observed. Figure 6 shows the trend for FFAs conversion of
used Amberlyst 36c 42.18 4.19 0.10 18.17 35.36 Purolite D5081 for three levels of impeller agitation speed, that
a
All percentages are in w/w %. bOxygen by difference. cWashing after is, 350, 475, and 600 rpm. The FFAs conversion between the
first cycle reaction.

Table 5. Acid Capacity for Fresh and Used Ion-Exchange

Resins
catalyst fresh resin (mmol/g) used resin (mmol/g)
D5081 1.59 1.39
D5082 1.79 1.59
Amberlyst 36 5.00 4.99

Purolite D5081 has the lowest sulfur content (Table 4), it

exhibits the highest BET surface area and total pore volume,
which means that there are more accessible active sites for the
reaction to occur, and hence it reaches the equilibrium at a
faster rate. In addition, the particle size distribution results
(Figure 4) show that Purolite D5081 has the smallest average Figure 6. Effect of different stirring speed on the FFAs conversion −
particle size, which leads to a larger external surface area as external mass transfer resistance (experimental conditions: catalyst,
compared to the other resins. A larger surface area may Purolite D5081; molar ratio (methanol:UCO), 6:1; catalyst loading,
contribute to a faster reaction rate and shorten the time to 1.25% w/w; reaction temperature, 60 °C).

14659 dx.doi.org/10.1021/ie3007566 | Ind. Eng. Chem. Res. 2012, 51, 14653−14664

Industrial & Engineering Chemistry Research Article

selected stirring rates gave almost an identical trend with less

than 2% difference when the agitation speed increased from
350 to 600 rpm. As the stirring speed only has a small impact
on the FFAs conversion, it was confirmed that there was no
evidence of external mass transfer resistance in the esterification
reaction.
Internal mass transfer resistance refers to resistance of
movement of reactant inside the pores of the catalyst. To
investigate the occurrence of internal mass transfer resistance, a
set of experiments were conducted using various ranges of
particle sizes, and this method has been widely used in the
catalytic reaction process.18,35,41−44 Figure 7 shows the effect of

Figure 8. Effect of different catalyst loading on the FFAs conversion

(experimental conditions: catalyst, Purolite D5081; stirring speed, 350
rpm; reaction temperature, 60 °C; molar ratio (methanol:UCO), 6:1).

the catalyst loading increases, the number of active catalytic

Figure 7. Effect of different resin size on the FFAs conversion −
internal mass transfer resistance (experimental conditions: catalyst,
Purolite D5081; stirring speed, 475 rpm; catalyst loading, 1.25% w/w;
reaction temperature, 60 °C; molar ratio (methanol:UCO), 6:1).
different PSD on the FFAs conversion. From Figure 7, it can be
seen that 280−800 μm particle range gives a slightly higher
conversion as compared to the other two particle ranges. This is
because this range, that is, 280−800 μm, covers a wide range of
particles, and therefore the possibility of having smaller particles
is higher. This leads to a higher conversion of FFAs as smaller
particles have shorter pore channels and a larger area of the
pores is accessible by the reactant molecules. To support this
argument, a comparison was made between 600−710 and
425−500 μm range catalyst particles, and it was found that the
conversion for 600−710 μm particle size was slightly lower
than 425−500 μm particle size in the first 5 h of reaction.
Nevertheless, the conversions of FFAs for all three particle
ranges were approximately the same at the end of the 8 h
reaction. This finding proved the previous theory that smaller
catalyst particles contribute to a larger accessible surface area
and thus increase the FFAs conversion. However, the difference
between each range was very small (less than ±2%), and the
final conversions were approximately the same. Hence, the
effect of internal mass transfer limitation can be eliminated. As a
conclusion, stirring speed of 350 rpm and above was selected
for subsequent experiments as there was no external mass
transfer resistance above this speed limit and the Purolite
D5081 resin was used without sieving.
3.3.3. Effect of Catalyst Loading. The effect of catalyst
loading is shown in Figure 8. The reaction temperature was set
at 60 °C with 6:1 methanol to UCO molar ratio, and the
reaction took approximately 8 h to reach equilibrium. The
result shows that the conversion is greatly influenced by the
amount of catalyst, especially at the early stage of the
esterification reaction. This is attributed to the fact that as
sites increases, and a shorter time is required to reach
equilibrium. Variation in catalyst loading also has an effect on
the final conversion of FFAs with the difference between the
highest and the lowest catalyst loading of about 10%. At a low
catalyst concentration (<1% w/w), the difference in the
conversion was significant. However, as catalyst loading was
increased (≥1.25% w/w), the trends were less significant. This
phenomenon indicates that as the percentage of resin increases,
the conversion increases. However, as the number of catalytic
sites increases (≥1.25% w/w catalyst loading), the benefits are
reduced as there are sufficient active catalytic sites in the
reactant molecules to catalyze the reaction. In the case of the
system investigated, it was found that once the catalyst loading
increased to 1.25% w/w, there were sufficient catalyst sites for
this particular reaction. As a result, 1.25% w/w loading was
chosen as the optimum catalyst loading and was used for all
further experimental work.
3.3.4. Effect of Reaction Temperature. The choice of
temperature range was determined by the findings of Leung
and Guo.45 They investigated the effect of temperature on the
biodiesel process, and the results showed that any temperatures
higher than 50 °C have a negative impact on the virgin oils but
had a positive impact for waste oils. Esterification was
investigated at different temperatures (50, 55, 60, 62, 65 °C),
and the results are shown in Figure 9. It has been observed that
Figure 9. Effect of different reaction temperatures on FFAs conversion
(experimental conditions: catalyst, Purolite D5081; stirring speed, 350
rpm; catalyst loading, 1.25% w/w; molar ratio (methanol:UCO), 6:1).

14660 dx.doi.org/10.1021/ie3007566 | Ind. Eng. Chem. Res. 2012, 51, 14653−14664

Industrial & Engineering Chemistry Research
increasing temperatures lead to reduction in the viscosity of
UCO, which enhances the contact between the methanol and
UCO leading to a higher conversion of FFAs. The highest
conversion was obtained at 65 °C. A decrease in the volume of
the reaction mixture was observed when the temperature
reached the boiling point of methanol, 64.7 °C. It was expected
that there will be some changes to the system when this
temperature is reached, with more methanol present in the
headspace of the reactor as vapor. Liu et al.46 also claimed that
beyond 65 °C, methanol started to vaporize rapidly, forming a
large number of bubbles to form foam and resulting in a
decrease in FFAs conversion. Generally, in a typical biodiesel
reaction process, low temperature will result in lower
conversion, while higher temperatures lead to excessive
methanol loss due to evaporation. After consideration of the
safety issues, cost implications, and the conversion trends for
each temperature, the optimum reaction temperature was
found to be 60 °C.
3.3.5. Effect of Methanol to UCO Molar Ratio. Figure 10
shows the effect of methanol to UCO molar ratio on the
Figure 10. Effect of different molar ratio (methanol:UCO) on the
FFAs conversion (experimental conditions: catalyst, Purolite D5081;
catalyst loading, 1.25% w/w; stirring speed, 350 rpm; reaction
temperature, 60 °C).
conversion of FFAs. The reaction temperature was set at 60 °C,
at 350 rpm stirrer speed with 8 h reaction time. The molar
ratios were calculated on the basis of the molecular mass of
average fatty acid composition. From Figure 10, it can be seen
that the conversion increased slightly with an increase in
methanol to UCO molar ratio. However, the conversion
differences between the molar ratios were less noticeable, giving
about 2% increments in conversion as the molar ratio increases.
In comparison, a ratio of 4:1 has a comparably lower initial
conversion as compared to the other three molar ratios. This
will be due to insufficient methanol in the mixture to force the
reaction, leading to longer reaction time to reach equilibrium.
The reaction time can be shortened using a higher methanol to
UCO molar ratio; however, the costs will be higher due to
higher methanol recovery requirements. A high molar ratio also
leads to difficulties in the separation process and thus hinders
separation by gravity.47 Furthermore, this situation, that is, high
molar ratio, could also increase the solubility of glycerine, and
thus reduces the yield of FAME as glycerol remains in the
pretreated UCO phase. Excess methanol also could drive the
combination of methyl ester and glycerine to monoglycerides,
which increases the viscosity of the reaction mixture.46 Taking
into account the safety issues and the capital and operating
14661
Article
costs, a molar ratio of 6:1 (methanol:UCO) was chosen as the
optimum molar ratio for the esterification reaction.
3.4. Catalyst Reusability Study. A series of reactions,
using the same batch of Purolite D5081 catalyst, were carried
out at optimum process conditions, that is, 1.25% w/w catalyst
loading, 333 K reaction temperature, 6:1 methanol to UCO
feed mole ratio, and 475 rpm stirring speed, to determine the
catalyst life span. The results are shown in Figure 11, and from
Figure 11. Reusability study on Purolite D5081 ion-exchange resins
(experimental conditions: catalyst, Purolite D5081; stirring speed, 475
rpm; catalyst loading, 1.25% w/w; reaction temperature, 60 °C; molar
ratio (methanol:UCO), 6:1).
these data it can be seen that the conversion decreased by
approximately 8−10% per cycle (e.g., fresh catalyst gives ∼92%
conversion, the first cycle gives ∼84% conversion, and the
second cycle gives ∼73% conversion). Potential reasons for the
loss of activity include contamination of the external or internal
surface of the catalyst and sulfur leaching.
Scanning electron microscopy (SEM) analysis was used to
investigate the effect of cleaning regimes on the surface of the
ion-exchange resin catalysts. Purolite D5081 was cleaned by
placing the catalyst in a flask of methanol, which was
subsequently placed in an ultrasonic bath, and the method is
detailed in section 2.5. To determine the effect of ultra-
sonication, Purolite D5082 was cleaned using methanol but
without ultrasonication. A series of SEM images are shown in
Figure 12, which show the comparison of fresh and used
catalyst (after the first run) as well as comparing the effect of
cleaning regimes. Figure 12a and b shows a sample of Purolite
D5081 before and after esterification, and it can be seen that
there is no trace of UCO on the surface of the used catalyst in
Figure 12b. Figure 12c and d shows a sample of Purolite D5082
before and after esterification, and in this case it can be seen
that there is UCO present on the surface of the catalyst,
indicating incomplete washing of the catalyst. This observation
fits with the results of the elemental analysis for used D5082
(Table 4), which shows there is a slight increase in residual
carbon. On this basis it was decided that ultrasonication was
needed as part of the cleaning process to ensure UCO was
removed from the surface of the catalyst.
Contamination of the internal surface of the catalyst could
possibly result from the resin pore blockage, either by the
presence of metal ions in the UCO or by the blockage by the
UCO residue itself. An energy-dispersive X-ray spectroscopy
(EDX) analysis of fresh and used Purolite D5081 catalysts has
been conducted, and a trace amount of metal was found on the
catalyst. The amount of sodium and iron content was very
dx.doi.org/10.1021/ie3007566 | Ind. Eng. Chem. Res. 2012, 51, 14653−14664
Industrial & Engineering Chemistry Research
Figure 12. SEM analysis of Purolite catalysts taken at 5000×
magnification: (a) and (b) show Purolite D5081 before and after
the esterification process, and (c) and (d) show Purolite D5082 before
and after esterification process.
similar for both fresh and used Purolite D5081 catalysts
(∼0.20% and ∼0.15%). In the used Purolite D5081 catalyst, a
trace amount of calcium ions was found (approximately 0.012−
0.10%). Given that only a trace amount of metal was found on
the surface of the catalyst, it has been concluded that the metal
ion impurities in UCO do not contribute to the deactivation of
Purolite D5081 catalyst.
It was found that the mass of catalyst increased slightly after
each reaction cycle, and it has been assumed that this was due
to the presence of triglycerides, proteins, trace amount of
phospholipids, and other impurities in UCO that could
potentially foul Purolite D5081 catalyst. The BET analysis
showed that Purolite D5081 resin has the highest total pore
volume of 0.47 cm3/g, which was slightly higher than other
conventional ion-exchange resins. This means it is possible for
the triglycerides, proteins, and a trace amount of phospholipids
molecules to be retained within the resin catalyst, and they may
contribute to pore blockage. This may also reduce the
accessibility of the active catalyst sites and may lead to higher
internal mass transfer resistance and finally contribute to
decreasing catalytic activity.
Sulfur leaching occurs due to the detachment of sulfonic acid
group from the polymer matrix. Water is one of the products
formed during the esterification reaction, and in theory it could
hydrolyze sulfonic acid groups to form homogeneous sulphuric
acid. An elemental analysis was carried out to determine the
level of sulfur within various resin samples, and the results are
shown in Table 4. Fresh D5081 resin had a sulfur content of
4.1%, and after the first reaction cycle this decreased by nearly
20% to 3.3%. Ion-exchange capacity results in Table 5 show a
decrease of acid capacity for fresh and used catalyst of
approximately 13%. This indicates that sulfur leaching
contributes to the reduction in catalytic activity of Purolite
D5081 catalyst.
To further investigate the leaching of homogeneous species
in the reaction mixture and UCO blockage within the resin
pores, several sets of experiments have been carried out. The
experimental work and the process flow diagram have been
detailed in section 2.5 and Figure 2, respectively. Figure 13
shows the conversion of FFAs during the uncatalyzed reaction
between the treated-methanol solution and UCO, and this
14662
Article
Figure 13. Study on the homogeneous contribution of Purolite D5081
ion-exchange resins (experimental conditions: catalyst, Purolite
D5081; stirring speed, 475 rpm; reaction temperature, 60 °C; molar
ratio (methanol:UCO), 6:1).
result clearly shows there was a significant FFAs conversion for
the first cycle of homogeneous contribution study. This
confirms the occurrence of sulfonic acid group leaching,
believed to be due to the detachment of sulfonic acid group
from the catalyst surface, followed by the hydrolysis of sulfonic
acid species with water to form homogeneous species. Data for
the second cycle show after the first homogeneous contribution
cycle that conversion is very low, indicating that there is no
further leaching of the sulfonic acid group into methanol in
subsequent cycles.
In addition, the catalyst deactivation has also been
investigated using the catalyst that was previously filtered
from the methanol solution. The results are summarized in
Figure 14. From the figure, it could be seen that the conversion
Figure 14. Study on the methanol treated catalyst deactivation
(experimental conditions: catalyst, Purolite D5081; catalyst loading,
1.25% w/w; stirring speed, 475 rpm; reaction temperature, 60 °C;
molar ratio (methanol:UCO), 6:1).
for the second cycle was slightly lower as compared to the first
cycle. This was solely due to the blockage of large molecules of
UCO molecules as the data from Figure 13 show that the
leaching of sulfonic acid was negligible during the second cycle
of the reaction.
Figure 15 shows the comparison between the reusability
study and methanol treated catalyst study. The difference in
FFAs conversion between the fresh catalyst and first methanol
treated catalyst cycle is believed to be due to leaching of
sulfonic acid groups. This is because both experiments were
conducted at the same experimental parameters; the only
dx.doi.org/10.1021/ie3007566 | Ind. Eng. Chem. Res. 2012, 51, 14653−14664
Industrial & Engineering Chemistry Research
Figure 15. Comparison between the reusability study and the
methanol treated catalyst study (experimental conditions: catalyst,
Purolite D5081; stirring speed, 475 rpm; catalyst loading, 1.25% w/w;
reaction temperature, 60 °C; molar ratio (methanol:UCO), 6:1).
difference was the condition of the catalyst, where the former
experiment used fresh catalyst and the latter one used
methanol-treated-catalyst. This finding was verified using the
homogeneous contribution study shown in Figure 13.
On the other hand, the first cycle of methanol treated catalyst
study gives a conversion close to the first cycle from the
reusability study. For the former experiment, the resin itself was
not introduced to any UCO mixture, so there was no possibility
of any triglyceride blockage during the reaction, and this means
the conversion was slightly higher. So, it could be inferred that
the 6% reduction in conversion was solely due to blockage of
large UCO molecules in the catalysts’ pores. The same trend
was observed during the second cycle for both cases where the
calculated conversion difference was about the same. Therefore,
it could be concluded that the reduction of FFAs conversion in
subsequent cycles (either for the reusability study or for the
methanol treated catalyst study) is largely due to the
progressive blockage of the large triglycerides molecules in
the pores of the resin, because there is a steady reduction in the
FFAs conversion for each cycle. Furthermore, the possibility of
sulfonic acid groups leaching throughout the process was small.
When both conditions (reusability study and methanol treated
catalyst study) were evaluated individually, both of them
showed sulfur leaching and pore blockage simultaneously and
show a reduction in FFAs conversion with increasing cycle
number.
4. CONCLUSIONS
Esterification pretreatment of UCO using various types of ion-
exchange resins has been investigated. Among the catalysts
investigated, Purolite D5081 resin showed the best catalytic
performance as compared to the other two resins, Purolite
D5082 and Amberlyst 36. This is probably due to the catalytic
properties of this resin, as it has the highest specific surface area
and largest total pore volume. At the optimum reaction
condition of 60 °C reaction temperature, 6:1 methanol to UCO
molar ratio, 1.25% w/w catalyst loading, and 475 rpm stirring
speed, Purolite D5081 achieved an FFAs conversion of 92%.
During the reusability study, the conversion of catalyst dropped
by 8−10% after each cycle. Several experiments have been
conducted through the homogeneous contribution study, and
the results confirmed that resin pore blockage and sulfur
leaching were two dominant factors that decrease the catalytic
performance of the polymeric resin.
14663
■
Article
AUTHOR INFORMATION
Corresponding Author
*Tel.: +44 (0) 2078157190. Fax: +44 (0) 2078157699. E-mail:
b.saha@lsbu.ac.uk.
Notes
The authors declare no competing financial interest.
■ ACKNOWLEDGMENTS
We gratefully acknowledge Purolite International Ltd. (late Dr.
Jim Dale and Mr. Brian Windsor) for kindly supplying the
catalysts for this research work and GreenFuel Oil Co. Ltd. for
supplying the UCO. We also would like to thank EPSRC
funding for the Ph.D. Scholarship to K.F.H. and Universiti
Malaysia Pahang and Malaysian Government for the Ph.D.
scholarship to S.Z.A.
■
REFERENCES
(1) Sharma, Y. C.; Singh, B.; Upadhyay, S. N. Advancements in
Development and Characterization of Biodiesel: A review. Fuel 2008,
87, 2355.
(2) Boehman, L. Biodiesel Production and Processing. Fuel Process.
Technol. 2005, 86, 1057.
(3) Lotero, E.; Liu, Y.; Lopez, D. E.; Suwannakarn, K.; Bruce, D. A.;
Goodwin, J. G. Synthesis of Biodiesel via Acid Catalysis. Ind. Eng.
Chem. Res. 2005, 44, 5353.
(4) Gui, M.; Lee, K.; Bhatia, S. Feasibility of Edible Oil vs. Non-
edible Oil vs. Waste Edible Oil as Biodiesel Feedstock. Energy 2008,
33, 1646.
(5) Stamenković, O. S.; Velićković, A. V.; Veljković, V. B. The
Production of Biodiesel from Vegetable Oils by Ethanolysis: Current
State and Perspectives. Fuel 2011, 90, 3141.
(6) Behzadi, S.; Farid, M. M. Review: Examining the Use of Different
Feedstock for the Production of Biodiesel. Asia-Pac. J. Chem. Eng.
2007, 2, 480.
(7) Canakci, M. The Potential of Restaurant Waste Lipids as
Biodiesel Feedstocks. Bioresour. Technol. 2007, 98, 183.
(8) Ramadhas, A. S.; Jayaraj, S.; Muraleedharan, C. Biodiesel
Production from High FFA Rubber Seed Oil. Fuel 2005, 84, 335.
(9) Supple, B.; Holward-Hildige, R.; Gonzalez-Gomez, E.; Leahy, J. J.
The Effect of Steam Treating Waste Cooking Oil on the Yield of
Methyl Ester. J. Am Oil Chem. Soc. 2002, 79, 175.
(10) Tomasevic, A. V.; Silver-Marinkovic, S. S. Methanolysis of Used
Frying Oil. Fuel Process. Technol. 2003, 81, 1.
(11) Balat, M. Potential Alternatives to Edible Oils for Biodiesel
Production − A Review of Current Work. Energy Convers. Manage.
2011, 52, 1479.
(12) Shashikant, V. G.; Hifjur, R. Process Optimization for Biodiesel
Production from Mahua (Madhuca Indica) Oil using Response
Surface Methodology. Bioresour. Technol. 2006, 97, 379.
(13) Jena, P. C.; Raheman, H.; Kumar, G. V. P.; Machavaram, R.
Biodiesel Production from Mixture of Mahua and Simarouba Oils with
High Free Fatty Acids. Biomass Bioenergy 2010, 34, 1108.
(14) Srivastiva, A.; Prasad, R. Triglycerides-Based Diesel Fuels.
Renewable Sustainable Energy Rev. 2000, 4, 111.
(15) Worapun, I.; Pianthong, K.; Thaiyasuit, P. Two-Step Biodiesel
Production from Crude Jatropha Curcas L. Oil using Ultrasonic
Irradiation Assisted. J. Oleo Sci. 2012, 61, 165.
(16) Veljković, V. B.; Lekićević, S. H.; Stamenković, O. S.; Todorović,
Z. B.; Lazic, M. L. Biodiesel Production from Tobacco (Nicotiana
Tabacum L.) Seed Oil with a High Content of Free Fatty Acids. Fuel
2006, 85, 2671.
(17) Zhang, J.; Jiang, L. Acid-Catalyzed Esterification of Zanthoxylum
Bungeanum Seed Oil with High Free Fatty Acids for Biodiesel
Production. Bioresour. Technol. 2008, 99, 8995.
(18) Zheng, S.; Kates, M.; Dubé, M. A.; McLean, D. D. Acid-
Catalyzed Production of Biodiesel from Waste Frying Oil. Biomass
Bioenergy 2006, 30, 267.
dx.doi.org/10.1021/ie3007566 | Ind. Eng. Chem. Res. 2012, 51, 14653−14664
Industrial & Engineering Chemistry Research Article

(19) Russbueldt, B. M. E.; Hoelderich, W. F. New Sulfonic Acid Ion- (43) Tesser, R.; Casale, L.; Verde, D.; Di Serio, M.; Santacesaria, E.
Exchange Resins for the Preesterification of Different Oils and Fats Kinetics of Free Fatty Acids Esterification: Batch and Loop Reactor
with High Content of Free Fatty Acids. Appl. Catal., A 2009, 362, 47. Modeling. Chem. Eng. J. 2009, 154, 25.
(20) Naik, M.; Meher, L. C.; Naik, S. N.; Das, L. M. Production of (44) Silva, V. M. T. M.; Rodrigues, A. E. Kinetic Studies in a Batch
Biodiesel from High Free Fatty Acid Karanja (Pongamia Pinnata) Oil. Reactor using Ion Exchange Resin Catalysts for Oxygenates
Biomass Bioenergy 2008, 32, 354. Production: Role of Mass Transfer Mechanisms. Chem. Eng. Sci.
(21) Ngo, H. L.; Zafiropoulos, N. A.; Foglia, T. A.; Samulski, E. T.; 2006, 61, 316.
Lin, W. Efficient Two-Step Synthesis of Biodiesel from Greases. Energy (45) Leung, D. Y. C.; Gou, Y. Transesterification of Neat and Used
Fuels 2008, 22, 626. Frying Oil: Optimization for Biodiesel Production. Fuel Process.
(22) López, D. E.; Goodwin, J.; James, G.; Bruce, D. A.; Lotero, E. Technol. 2006, 87, 883.
Transesterification of Triacetin with Methanol on Solid Acid and Base (46) Liu, Y.; Wang, L.; Yan, Y. Biodiesel Synthesis Combining Pre-
Catalysts. Appl. Catal., A 2005, 295, 97. Esterification with Alkali Catalyzed Process from Rapeseed Oil
(23) Yan, S.; DiMaggio, C.; Mohan, S.; Kim, M.; Salley, S. O.; Ng, K. Deodorizer Distillate. Fuel Process. Technol. 2009, 90, 857.
Y. S. Advancements in Heterogeneous Catalysis for Biodiesel (47) Encinar, J. M.; González, J. F.; Rodríquez-Reinares, A. Biodiesel
Synthesis. Top. Catal. 2010, 53, 721. Preparation and Characterization. Fuel Process. Technol. 2007, 88, 513.
(24) Chai, F.; Cao, F.; Zhai, F.; Chen, Y.; Wang, X.; Su, Z.
Transesterification of Vegetable Oil to Biodeisel using a Heteroge-
neous Solid Catalyst. Adv. Synth. Catal. 2007, 348, 1057.
(25) Kitakawa, N. S.; Honda, H.; Kuribayashi, H.; Toda, T.;
Fukumura, T.; Yonemoto, T. Biodiesel Production using Anionic Ion-
Exchange Resin as Heterogeneous Catalyst. Bioresour. Technol. 2007,
98, 416.
(26) Feng, Y.; He, B.; Cao, Y.; Li, J.; Liu, M.; Yan, F.; Liang, X.
Biodiesel Production using Cation-Exchange Resin as Heterogeneous
Catalyst. Bioresour. Technol. 2010, 101, 1518.
(27) Yun, L.; Ling, W. Bio-Diesel Preparation from Waste Oil using
Cation Exchange Resin as Heterogeneous Catalyst. Chem. Technol.
Fuels Oils 2009, 45, 417.
(28) Liu, Y.; Lotero, E.; Goodwin, J. G., Jr. A Comparison of the
Esterification of Acetic Acid with Menthanol using Heterogeneous
versus Homogeneous Acid Catalysis. J. Catal. 2006, 242, 278.
(29) Grob, S.; Hasse, H. Reaction Kinetics of the Homogeneously
Catalysed Esterification of 1-Butanol With Acetic Acid in a Wide
Range of Initial Compositions. Ind. Eng. Chem. Res. 2006, 45, 1869.
(30) Ö zbay, N.; Oktar, N.; Tapan, N. A. Esterification of Free Fatty
Acids in Waste Cooking Oils (WCO): Role of Ion-Exchange Resins.
Fuel 2008, 87, 1789.
(31) Bianchi, C. L.; Boffito, D. C.; Pirola, C.; Ragaini, V. Low
Temperature De-Acidification Process of Animal Fat as a Pre-Step to
Biodiesel Production. Catal. Lett. 2010, 134, 179.
(32) Tesser, R.; Di Serio, M.; Casale, L.; Sannino, L.; Ledda, M.;
Santacesaria, E. Acid Exchange Resins Deactivation in the
Esterification of Free Fatty Acids. Chem. Eng. J. 2010, 161, 212.
(33) Park, J. Y.; Kim, D. G.; Lee, J. S. Esterification of Free Fatty
Acids using Water-Tolerable Amberlyst as a Heterogeneous Catalyst.
Bioresour. Technol. 2010, 101, S62.
(34) Klepacova, K.; Mravec, D.; Bajus, M. tert-Butylation of Glycerol
Catalysed by Ion-Exchange Resin. Appl. Catal., A 2005, 294, 141.
(35) Masato, K.; Atsuko, N.; Jyu-suke, H. Pre-Esterification of FFA in
Plant Oil Transesterified into Biodiesel with the help of Solid Acid
Catalysis of Sulfonated Cation-Exchange Resin. Appl. Catal., A 2011,
405, 36.
(36) David, F.; Sandra, P.; Vickers, A. K. Column Selection for the
Analysis of Fatty Acid Methyl Esters. Agilent Application Notes 5989-
3760EN, 2005; p 1.
(37) Dowd, M. K. Gas Chromatographic Characterization of
Soapstocks from Vegetable Oil Refining. J. Chromatogr., A 1998,
816, 185.
(38) Knothe, G.; Steidley, K. R. A Comparison of Used Cooking
Oils: A Very Heterogeneous Feedstock for Biodiesel. Bioresour.
Technol. 2009, 100, 5796.
(39) EU Regulation 2568/91. Official Journal EU, document L248 5/
9/1991, 1991.
(40) EN 1S0 12966-2. Animal and Vegetable Fats and Oils-Gas
Chromatography of Fatty Acid Methyl Esters: Part 2: Preparation of
Methyl Esters of Fatty Acids, 2011; p 1.
(41) Helfferich, F. Ion Exchange; McGraw-Hill: New York, 1962.
(42) Yadav, G. D.; Kulkarni, H. B. Ion-Exchange Resin Catalysis in
the Synthesis of Isopropyl Lactate. React. Funct. Polym. 2000, 44, 153.

14664 dx.doi.org/10.1021/ie3007566 | Ind. Eng. Chem. Res. 2012, 51, 14653−14664