Fatty acids residue from palm oil refining process as feedstock for lipase catalyzed monoacylglicerol production under batch and continuous flow conditions

Article abstract of DOI:10.1016/j.molcatb.2012.01.008

Free fatty acids are used in many cases for the production of soaps, candles and assist processing of rubber products, but we believe that new process technology should be developed to produce products with higher added value. Monoacylglycerols (MAGs) are nonionic surfactant, highly hydrophobic and has been used as controlled release systems for drugs. The results presented here for the lipase-catalyzed MAG production show that both batch and continuous flow conditions can lead to the desired product in short reaction time and high yield (70-95%) but the use of packed bed reactors (PBR) shows higher efficiency when compared to batch reactors.

Full text of DOI:10.1016/j.molcatb.2012.01.008

Journal of Molecular Catalysis B: Enzymatic 77 (2012) 53–58
Contents lists available at SciVerse ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
Fatty acids residue from palm oil refining process as feedstock for lipase catalyzed
monoacylglicerol production under batch and continuous flow conditions
Ivaldo I. Juniora,b,1, Marcela C. Floresa,2, Felipe K. Sutilia,c,3, Selma G.F. Leitec,4
,
Leandro S. de M. e Mirandad, Ivana C.R. Lealb,5, Rodrigo O.M.A. de Souzaa,∗
a Biocatalysis and Organic Synthesis Group, Chemistry Institute, Federal University of Rio de Janeiro, CEP 22941-909, Rio de Janeiro, Brazil
b Faculdade de Farmácia, Campus Macaé, Federal University of Rio de Janeiro, CEP 22941-909, Rio de Janeiro, Brazil
c Escola de Química, Federal University of Rio de Janeiro, CEP 20270021, Rio de Janeiro, Brazil
d Instituto Federal de Educac¸ ão Ciência e Tecnologia do Rio de Janeiro, Campus Maracanã, CEP 22941-909, Rio de Janeiro, Brazil
a r t i c l e i n f o
a b s t r a c t
Article history:
Free fatty acids are used in many cases for the production of soaps, candles and assist processing of
rubber products, but we believe that new process technology should be developed to produce products
with higher added value. Monoacylglycerols (MAGs) are nonionic surfactant, highly hydrophobic and has
been used as controlled release systems for drugs. The results presented here for the lipase-catalyzed
MAG production show that both batch and continuous flow conditions can lead to the desired product
in short reaction time and high yield (70–95%) but the use of packed bed reactors (PBR) shows higher
efficiency when compared to batch reactors.
Received 15 September 2011
Received in revised form
17 December 2011
Accepted 6 January 2012
Available online 16 January 2012
Keywords:
Flow chemistry
Lipases
© 2012 Elsevier B.V. All rights reserved.
1,2-O-isopropylidene glycerol
Packed bed reactor
Fatty acids
Palm oil
1. Introduction
Palm oil has a wide range of applications, about 80% is used
for food applications (cooking oil), while the rest is raw material
The Twelve Principles of Green Chemistry reflect the need for
definitive action with regard to the development of processes,
whether in universities and industry [1]. In this context, the avail-
ability of cheap raw material and renewable energy from nature
is the basis for industrial sustainability, as well as a better use of
industrial waste. In this sense, strategies to develop clean tech-
nologies for chemical processes aim to balance economic and
environmental aspects [2].
In recent years, the group of four main countries that grow soy-
beans comprising Brazil, Argentina, Bolivia and Paraguay recorded
92% increase in production and 66% increase in planted area with
soybeans and palm oil, which represents several tons fiber and
by-products of the refining procedure.
for a series of non-food applications [3]. Among food uses, olein
refined, bleached and deodorized is mainly used in cooking and
frying oils, fats and margarines. Many mixtures have been devel-
oped to produce solid fats with zero content of trans-fatty acids
[4].
The refining process removes free fatty acids, phosphatides,
odoriferous matter, water and impurities from the crude palm oil
to produce high quality edible oil that meets industry standards. To
achieve this chemical or physical objective, refining can be used as
shown in Scheme 1 and in both cases free fatty acids are obtained
as a byproduct of the refining processes.
Free fatty acids are used in many cases for the production of
soaps, candles and assist processing of rubber products, but we
believe that new process technology should be developed to pro-
duce products with higher added value [5]. Monoacylglycerols
(MAGs) are important molecules that find applications in differ-
ent fields. In food industry are widely used in the preparation of
bakers, pastry and margarine [6]. They are also used in pharmaceu-
tical [7] and cosmetic industry as drug carriers and to increase the
consistency of creams and lotions [8].
∗
Corresponding author. Fax: +55 2125627001.
Fax: +55 2125627001/2125627807.
Fax: +55 2125627001.
Fax: +55 2125627001/2125627005.
Fax: +55 2125627005.
Fax: +55 2125627807.
1
2
3
Monoacylglycerols (MAGs) are nonionic surfactant, highly
hydrophobic and has been used as controlled release systems for
4
5
1381-1177/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcatb.2012.01.008

54
I.I. Junior et al. / Journal of Molecular Catalysis B: Enzymatic 77 (2012) 53–58
Crude Palm Oil
Physical Refining
Degumming
Chemical Refining
Alkali
Neutralizaꢀon
Soap Stock
Free Faꢁy Acids
Earth Bleaching
Deodorising
Earth Bleaching
Deodorising
Free Faꢁy Acids
Refined Oil
Scheme 1. Palm oil refining process.
drugs [7]. Its application is mainly due to hydrophobic–hydrophilic
balance in its structures. In addition, plasticizers and lubricants
properties of monoacylglycerols allow its use in textile processes
and formulation of oils for different types of machines. These com-
pounds are also used as stabilizers and antifoam. It is estimated that
mixtures of these with MAG and Diacylglycerol (DAG) represent-
ing 75% of world production of emulsifiers [9]. MAG is also of great
interest in synthetic organic chemistry, which is used as synthetic
intermediate and chiral building blocks. In addition, monoacylglyc-
erols are considered safe by the Food and Drug Administration
(FDA), increasing the interest in its use and development of new
synthetic methods.
In order to add value to this residue of fatty acids from palm oil
industry, it was decided to make a comparative study on the esteri-
fication reaction catalyzed by lipase between 1,2-O-isopropylidene
glycerol and free fatty acids derived from palm oil refining (acid
residue) under batch and continuous flow conditions using the sur-
face response methodology (RSM) [10] for optimization of reaction
conditions. For industrial purposes, the continuous flow system is
preferred to batch reactors due to its greater process control, high
productivity and improvement of quality/purity and yield [11,12].
Several types of reactor can be used in continuous operation, among
those reactors, packed bed reactors (PBR) are the most popular due
to its high efficiency, low cost and ease of construction, operation
and maintenance [13–15].
considered critical variables (independent) of process and there-
fore assessed in planning. The reaction time was not considered a
significant variable in experimental designs, since past experience
shows that after 4 h of reaction no progress in reaction yield was
achieved.
For each enzyme a 24 CCD was applied with triplicate central
points to calculate the experimental error. The four variables and
their coded and real levels for the two enzymes evaluated in batch
reactions are presented in parentheses on Table 1.
In the batch reaction experiments, the reactions were carried
out using a stock solution containing 1,2-O-isopropylidene glycerol
and acid residue at 1:1 (equivalents) ratio, dissolved in n-heptane.
In 2 mL-ependorfs was added 1 mL reaction medium followed by
the addition of appropriate enzyme (1%, w/w) (Scheme 2). Conver-
sion was monitored using a modified method of Lowrey and Tinsley
[17] and confirmed by GC–MS analysis as described in detail in
Section 4.
Results of the experimental design with coded and real values
for both lipases evaluated are presented in Table 1. The first 16
entries in Table 1 are sufficient for determining the mathematical
model and the experiments of entries from 17 to 19 contain the
central point triplicate.
Both enzymes were able to catalyze the 1,2-O-isopropylidene
glycerol esterification with the acid residue at high conversions, as
shown in entries 8 and 16 (Table 1) for T. lanuginosus (TL IM) and R.
miehei (RMIM), respectively.
At this stage, there was a small influence of stirring for both
lipases, since small variations in conversion percentages were
found between reactions carried out under the same reaction tem-
perature. This phenomenon can be more easily understood by
examining the estimated effects (Table 2) that shows stirring as
the smallest effect among variables studied (−2.17 and −3.08 for
R. miehei (RMIM) and T. lanuginosus (TL IM), respectively). Another
important effect is the amount of enzyme used in each reaction,
since reactions with higher amounts of catalyst always lead to
major improvements in reaction conversions, as shown in Table 1
(entries 10 and 12).
It was also found that the estimated effects for isolated vari-
ables for both enzymes were all statistically significant as shown
by (p) values (Table 2). Variables showed similar effects for both
enzymes in relation to its grandeur and influences. The substrate
concentration showed the most positive effect among variables
studied for R. miehei (RMIM) and T. lanuginosus (TL IM) (11.1868 and
2. Results and discussion
In order to evaluate the best biocatalyst for the esterification
of 1,2-O-isopropylidene glycerol with acid residue from palm oil
refining process (Fig. 1), immobilized lipase from Rhizomucor miehei
(RMIM) and Thermomyces lanuginosus (TL IM) were chosen, which
are recognized in the literature by having relative specificity against
a large number of substrates and remarkable esterification activity
[16].
First, to identify variables with important effect in our reaction
system under batch conditions we use design of experiments with a
central composite design (CCD), composed of four variables, vary-
ing on two levels, which was used to maximize the acid residual
esterification catalyzed by lipases from R. miehei (RMIM) and T.
lanuginosus (TL IM). Thus, the need for a greater amount of experi-
ments is reduced without compromising the results. Temperature,
amount of enzyme, substrate concentration and stirring were

I.I. Junior et al. / Journal of Molecular Catalysis B: Enzymatic 77 (2012) 53–58
55
Fig. 1. Chromatogram of the fatty acid residue obtained from the Palm Oil refining process (fatty acid residue composition: 44% of palmitic acid, 42% of oleic acid and 14% of
stearic acid).
Table 1
Experimental design and results of the factorial CCD for the esterification reaction between 1,2-O-isopropylidene glycerol and acid residue catalyzed by Rhizomucor miehei
(RMIM) and Thermomyces lanuginosus (TL IM).
Entry
Temperature (◦C)
[E] (%)a
[S] (mM)
Stirring (rpm)
Conversion (%)
RMIM
TLIM
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
−1 (40)
1 (60)
−1 (0.1)
−1 (0.1)
1(1)
−1 (50)
−1 (50)
−1 (50)
−1 (50)
+1(100)
+1 (100)
+1 (100)
+1 (100)
−1 (50)
−1 (50)
−1 (50)
−1 (50)
+1 (100)
+1 (100)
+1 (100)
+1 (100)
0 (75)
−1 (50)
−1 (50)
−1 (50)
−1 (50)
−1 (50)
−1 (50)
−1 (50)
−1 (50)
+1 (250)
+1 (250)
+1 (250)
+1 (250)
+1 (250)
+1 (250)
+1 (250)
+1 (250)
0 (150)
0 (150)
0 (150)
58
62
75
78
82
87
82
94
56
54
62
67
80
90
78
96
72
72
72
60
58
69
70
77
87
85
95
52
54
63
71
75
89
81
93
75
74
75
−1 (40)
+1 (60)
−1 (40)
+1 (60)
−1 (40)
+1 (60)
−1 (40)
+1 (60)
−1 (40)
+1 (60)
−1 (40)
+1 (60)
−1 (40)
+1 (60)
0 (50)
1(1)
−1 (0.1)
−1 (0.1)
+1(1)
+1(1)
−1 (0.1)
−1 (0.1)
+1(1)
+1(1)
−1(0.1)
−1(0.1)
+1(1)
+1(1)
0(0.55)
0(0.55)
0(0.55)
0 (50)
0 (50)
0 (75)
0 (75)
a
By weight of total system.
23.0262, respectively) having a great influence on the conversion
rate.
As observed that R. miehei (RMIM), showed overall better results
than T. lanuginosus (TL IM), we decided to use this enzyme under
continuous flow conditions since the use of flow reactors can help
in improving the process under development.
The factorial design is presented in parentheses in Table 3 and
shows variables with their respective levels used. To adjust a second
order model, extra points (axial points) with the same distance from
the central point was added to the matrix for this project.
However, the optimization for the batch process, experimental
designs and analysis of results were performed using the software
To determine the optimal conditions for acid residue esterifica-
tion and identify variables with important effect on the reaction
catalyzed by immobilized lipase from R. miehei (RMIM) under
flow conditions we use a complete factorial design with two
independent variables: flow rate and substrate concentration, vary-
ing in two levels and three repetitions of the central point. The
temperature used for these reactions was the same used in the best
reaction profile obtained in batch, 60 ◦C (entry 16, Table 1).
Table 2
Estimated parameter effect for enzymes studied in CCD (batch process).
Variable
Effect
RMIM
p-Value
TLIM
RMIM
TLIM
Mean
Temperature (T)
Amount of enzyme (E)
Substrate concentration (S)
75.1584
3.4956
3.9081
11.1868
−2.1756
1.2943
2.1781
0.5231
−2.4843
−1.0693
1.9493
74.2894
7.0137
9.2487
23.0262
−3.0887
0.6862
4.3237
2.2487
−2.9362
−0.0412
1.3212
<0.0001a
0.0008a
0.0006a
<0.0001a
0.0022a
0.0062a
0.0022a
0.0364a
0.0017a
0.0090a
0.0027a
<0.0001a
0.0037a
0.0021a
0.0003a
0.0187a
0.2508
0.0097a
0.0345a
0.0206a
0.9321
0.0912
Stirring (St)
T × E
O
O
O
Lipase
R
OH
O
T × S
O
O
+
HO
R = palmitic, oleic and stearic
R
T × St
O
E × S
E × St
S × St
Scheme 2. Esterification reaction between 1,2-O-isopropylidene glycerol and free
fatty acids residue from palm oil refining process catalyzed by lipases.
a
Statistically significant at 95% of confidence level.

56
I.I. Junior et al. / Journal of Molecular Catalysis B: Enzymatic 77 (2012) 53–58
Table 3
Table 4
Experimental factorial design 22 and results for the esterification reaction between
the solketal and acid residue catalyzed by Rhizomucor miehei (RMIM) under contin-
uous flow conditions.
Estimated effect of the parameter 22 experimental design for the esterification reac-
tion between solketal and acid residual catalyzed by Rhizomucor miehei (RMIM)
under continuous flow.
Entry
Substrate (mM)
Flow (mL/min)
Conversion (%)
Variables
Effect
p-Value
1
2
3
4
5
6
7
8
9
−1 (71.9)
−1 (71.9)
+1(95.3)
+1(95.3)
−1.41 (67.5)
+1.41 (100)
0(83.7)
0 (83.7)
0 (83.7)
0 (83.7)
0 (83.7)
−1 (0.4)
+1(0.8)
−1 (0.4)
+1(0.8)
0(0.6)
56
26
53
42
73
20
28
41
37
37
36
Mean
36.9000
−28.8883
11.6500
7.6462
<0.0001a
<0.0001a
0.0006a
0.0011a
0.2763
Flow
Flow2
Substrate concentration
Substrate concentration2
Flow × substrate concentration
−0.4500
0 (0.6)
9.4000
0.0015a
−1.41 (0.3)
+1.41 (0.9)
0(0.6)
0(0.6)
0(0.6)
a
Statistically significant at 95% of confidence level.
10
11
Table 4 shows the estimated effects for the 22 experimental
design and p values. The variable flow, flow2, substrate concen-
tration and interaction between flow and substrate concentration
showed p < 0.05 being significant in the process. The variable flow
showed negative effect (−28.8883) as shown in Table 4, indicating
that the better low flow is the substrate conversion to products.
Eqs. (1)–(3) represent the experimental models of conversion
rate of acid residue to monoacylglicerols by enzymes R. miehei
(RMIM) and T. lanuginosus (TL IM) in batch and continuous flow con-
ditions, respectively, depending on variables studied. The model fit
was performed by analysis of variance and parameter R2.
Statistica 6.0 (Statsoft, Inc., USA). The level of significance was set
as 95% for the mathematical model and surface response. The sig-
nificance of regression coefficients and associated probabilities p(t)
were determined by the Student’s t test, the model equation sig-
nificance was determined by Fisher’s F test. The variance explained
by the model is given by the coefficients of multiple determination,
R2.
Experimental and matrix data for experimental designs for these
reactions are presented in Table 3. Experiments presented in entries
1–4 show combinations of individual variables in the process;
entries 5–8 refer to the axial points for constructing the quadratic
model and inputs 9–11 are triplicates of the central point for obtain-
ing the experimental error.
The highest conversion is shown in experiment 5 with 73% con-
version. This value is considerably lower compared with results
obtained in batch reactions, but the reaction time in the first exper-
iments was much higher. In continuous flow equipment used,
reaction proceeds in a microenvironment with 0.6 mL total volume.
In entry 5 was applied 0.6 mL/min, which means that the contact
time between substrate and enzyme was only 1 min, against 4 h by
batch reactions. Thus, it was possible to obtain excellent conversion
with reduced times under continuous flow conditions.
By using higher flow rates, the residence time for substrate is
not enough to lead to good conversions and only moderate to poor
results were obtained (entries 2 and 8, Table 3). It is important to
note that substrate concentration is also a problem that must be
taken into account. In this work high concentrations of substrates
can lead to poor conversions, even at low flow rates (entries 6 and
11, Table 3) which may be related to enzyme inhibition by the sub-
strate and also fatty acids deposition inside the channel as observed
for some experiments.
Y = 75.1584 + 3.4956T + 3.9081E + 11.1868S − 2.1756St
+ 1.2943TxE + 2.1781TxS + 0.5231TxSt − 2.4843ExSt
− 1.0693ExSt + 1.9493SxSt
(1)
(2)
Y = 74.2894 + 3.5068T + 4.6243E + 11.5131S − 1.5443St
+ 2.1618TxS + 1.1243TxSt − 1.4681ExS
Y = 36.6882 − 14.4442Q + 5.8912Q2 + 3.8231S + 4.6500QxS (3)
where Y is the conversion percentage and T, S, E, St and Q, not
encoded values of temperature, substrate concentration, enzyme
content, mechanical stirring and flow, respectively. Statistical tests
of models were performed by Fisher’s statistical test for ANOVA
(Table S1, see supporting information).
F values calculated for all mathematical models were highly
significant, being higher than those tabulated (Table S1, see sup-
porting information). The suitability of a model can be verified by
the coefficients of determination (R2) and (R) correlation. The coef-
ficients of determination R2 = 0.98 for R. miehei (RMIM) in batch,
R2 = 0.99 for T. lanuginosus (TL IM) in batch and R2 = 0.91 for R. miehei
Fig. 2. Response surface figures for batch and continuous flow process [18].

I.I. Junior et al. / Journal of Molecular Catalysis B: Enzymatic 77 (2012) 53–58
57
(RMIM), reactions under continuous flow conditions imply that the
sample variation sample of 98%, 99% and 91% for esters production
is attributed to the independent variables and can be explained by
the model accuracy.
R values as 0.99 for batch reaction to R. miehei (RMIM), 0.99 for
batch reactions to T. lanuginosus (TL IM) and 0.95 for continuous
flow conditions of R. miehei (RMIM) suggest a satisfactory repre-
sentation of the process model and a good correlation between
experimental results and theoretical values predicted by the model
equation. The closer to 1 the R value (correlation coefficient) indi-
cates the best correlation between experimental and predicted
values.
(including the buffer solutions) was added to a tube containing
0.6 mL of n-heptane and 1 mL of cupric acetate-pyridine (5%, w/v,
pH 6.0). The final solutions were vigorously mixed for 30 s in vor-
tex, and the upper organic phase was measured by a UV/visible
spectrophotometer at 715 nm. Each reaction was analyzed in
triplicate, and content conversion, calculated according to the
percentage difference for the absorbance shown by the stock solu-
tion.
4.4. Continuous flow reaction procedure
A 1 L HPLC bottle was equipped with desired reaction mixture
in heptane and a stir bar. The starting mixture was stirred for 5 min,
while the X-Cube (ThalesNano) instrument was equipped with the
packed bed reactor containing immobilized lipase (0.6 mL volume,
70 mm × 4 mm). The reaction parameters/temperature (40–60 ◦C),
0.1–3.0 mL/min flow rate and pressure (10 bar) were selected on
the flow reactor, and processing was started, whereby only pure
solvent (heptane) was pumped through the system until the instru-
ment had achieved the desired reaction parameters and stable
processing was assured. At that point the inlet tube was switched
from the solvent flask to the 1 L HPLC bottle containing the freshly
prepared reaction mixture. After processing through the flow reac-
tor, the inlet tube was dipped back into the flask containing pure
heptane and processed for 10 min further, thus washing from the
system any remaining reaction mixture. The excess of heptane was
removed under vacuum, and the product was obtained and ana-
lyzed by GC.
Fig. 2A and B shows the surface response graphs for batch reac-
tions of R. miehei (RMIM) and T. lanuginosus (TL IM), respectively.
The surface is framed in the reaction under continuous flow condi-
tions catalyzed by the enzyme R. miehei (RMIM), Fig. 2C shows the
negative effect of variable flow that was discussed.
The final step for monostearin synthesis was accomplished by
1,2-O-isopropylidene cleavage using boric acid as standard proce-
dure described over literature.
3. Conclusion
In conclusion, we developed an efficient biocatalytic method
under continuous flow conditions that add value to an acid residue
of oil palm refining process. Results presented show that both batch
and continuous flow conditions can lead to the desired product,
but the process flow productivity is better than the batch process,
since good yields can be obtained (73%) with short residence times
(1 min).
4.5. Batch reactions procedure
4. Experimental
The fatty acid residue was melted in a water bath at 60 ◦C until
form a homogeneous oil mixture. Then, a stock solution was pre-
pared in order to avoid pipeting errors. In a flask, the appropriate
amount of fatty acid residue and R,S-1,2-O-isopropylidene glycerol
in proportions of 1:1 (75 mM) were dissolved in n-heptane, enough
to complete 100 mL. The esterification reactions were performed
in 2 mL cryotubes, whose were poured 1 mL of stock solution,
followed by the addition of appropriate enzyme (1%, w/w). Each
reaction was done in triplicate. The cryotubes were finally incu-
bated in shaker at 250 rpm and 60 ◦C during 4 h. The conversion
were measured by a modification of the Lowry and Tinsley method
and then analyzed by GC–MS to be confirmed.
4.1. Materials
Heptane was purchased from Tedia Co., (R,S)-1,2-
isopropylidene glycerol from Sigma–Aldrich as well as all
chromatographic standards. Stearic acid (>98%) was purchased
from Vetec Ltda. Immobilized lipase (triacylglycerol hydrolase, EC
3.1.1.3; Lipozyme IM-20, 25 BIU/g) from R. miehei, supported on
a macroporous weak anionic exchange resin beads and Lipozyme
TL IM (lipase from T. lanuginosus, 35 BIU/g) were purchased from
Novozymes.
4.2. GC–MS analysis
GC–MS analysis was performed by using modified method from
EN 14105. Free fatty acids and (R,S)-1,2-isopropylidene glycerol
were transformed into more volatile silylated derivatives in pres-
ence of pyridine and N-methyl-N-trimethysilyltrifluoroacetamide
(MSTFA). All GC–MS measurements were carried out in duplicate
using a DB 5–HT (Agilent, J & W. Scientific®, USA) capillary col-
umn (10 m × 0.32 mm × 0.1 ␮m). The quantifying was done based
on calibration curves with internal standards. The GC–MS sam-
ples were prepared by dissolving 0.1 g of the final product on 1 mL
of n-heptane. 100 ␮L of this solution and pyridine solutions of
butanetriol (1 mg/mL) and tricaprine (8 mg/mL), used as internal
standards, were added on a flask forward by an addition of 100 ␮L
of MSTFA. After 15 min, these reactants were dissolved on 8 mL de
n-heptane. 1 ␮L of this sample was then injected into a Shimadzu
CG2010 equipment.
4.6. Statistical analysis
The experimental designs and results analysis were carried out
using the software Statistica 6.0 (Statsoft, Inc., USA), according
with the significance level established to obtain the mathemati-
cal model. The significance of the regression coefficients and the
associated probabilities, p(t), were determined by Student’s t test;
the model equation significance was determined by Fisher’s F test.
The variance is given by the multiple determination coefficients,
R2.
Acknowledgments
We thank CAPES (Coordenac¸ ão de Aperfeic¸ oamento de Pessoal
de Nível Superior), CNPq (Conselho Nacional de Desenvolvimento
Científico e Tecnológico), FAPERJ (Fundac¸ ão Carlos Chagas Filho de
Amparo a Pesquisa do Estado do Rio de Janeiro) and FINEP (Agência
Financiadora de Estudos e Projetos) for financial support. We also
thanks Companhia Refinadora da Amazônia (Agropalma) for the
acidic residue from palm oil refining process.
4.3. Lowry–Tinsley analysis
The esterification rate was also measured using a modifi-
cation of the Lowry and Tinsley assay. The depletion of fatty
acid was monitored as follows: 0.30 mL of the reaction solution

58
I.I. Junior et al. / Journal of Molecular Catalysis B: Enzymatic 77 (2012) 53–58
Appendix A. Supplementary data
(e) I. Johansson, M. Svevensson, Curr. Opin. Colloid Interface Sci. 6 (2001)
178–188.
[6] (a) M. Pernetti, K.F. Van Malssen, E. Floter, A. Bot, Curr. Opin. Colloid Interface
Sci. 12 (2007) 221–231;
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.molcatb.2012.01.008.
(b) E. Boyle, J.B. German, Crit. Rev. Food Sci. Nutr. 36 (1996) 785–805.
[7] T. Watanabe, N. Kosenki, H. Kyushiki, WO2004022091-A1, AU2003261950-A1.
[8] E. Martin, Trav. Chim. Aliment. Hyg. 79 (1988) 406–410.
[9] (a) S. Saadi, A.A. Ariffin, H.M. Ghazali, M.S. Miskandar, S.M. Abdulkarin, H.C. Boo,
J. Food Sci. 76 (2011) C21–C31;
References
(b) P. Pacivova, I. Buresova, H. Bilkova, J. Sci. Food Agric. 90 (2010) 2282–2288;
(c) H. Szelag, W. Zwierzykowski, Colloids Surf. A 155 (1999) 349–357;
(d) P. Kohler, W. Grosch, J. Agric. Food Chem. 47 (1999) 1863–1869.
[10] E.R. Gunawan, M. Basri, M.B.A. Rahman, A.B. Salleh, R. Rahman, Enzyme Microb.
Technol. 37 (2005) 739.
[11] C.G. Frost, L. Mutton, Green Chem. 12 (2010) 1687–1703.
[12] A. Pohar, I. Plazl, Chem. Biochem. Eng. Q. 23 (2009) 537–544.
[13] P. Watts, C. Wiles, Eur. J. Org. Chem. 10 (2008) 1655–1671.
[14] A. Kirschning, S. Ceylan, J. Wegner, Chem. Commun. 47 (2011) 4583–4592.
[15] (a) M.P. Dudukovic, F. Larachi, P.L. Mills, Chem. Eng. Sci. 54 (1999) 1975–1977;
(b) G.M. Whitesides, Nature 442 (2006) 368–370.
[1] S.L.Y. Tang, R.L. Smith, M. Poliakoff, Green Chem. 7 (2005) 761–762.
[2] (a) M.J. Hernaiz, A.R. Alcantara, J.I. Garcia, J.V. Sinisterra, Chem. Eur. J. 16 (2010)
9422–9437;
(b) Y.L. Gu, F. Gerome, Green. Chem. 12 (2010) 1127–1138;
(c) P. Anastas, N. Eghbali, Chem. Soc. Rev. 39 (2010) 301–312;
(d) J.H. Clark, F.E.I. Deswarte, T.J. Farmer Biofuels, Bioprod. Biorefin. 3 (2009)
72–90.
[3] A. Salamiah, Non-Food Uses of Palm Oil and Palm Kernel Oil, in: MPOPC Palm
Oil Information Series, Kuala Lumpur, 2000, 24 pp.
[4] K.G. Berger, Food Uses of Palm Oil, in: MPOPC Palm Oil Information Series, Kuala
Lumpur, 1996, 25 pp.
[16] (a) R. Fernandez-Lafuente, J. Mol. Catal. B 62 (2010) 197–212;
(b) R.C. Rodrigues, R. Fernandez-Lafuente, J. Mol. Catal. B 64 (2010) 1–22.
[17] R.R. Lowry, I.J. Tinsley, J. Am. Oil. Chem. Soc. 53 (1976) 470.
[18] R. Hess, U. Bornscheuer, A. Capewell, T. Scheper, Enzyme Microb. Technol. 17
(1995) 725.
[5] (a) I. Kralova, J. Sjoblom, J. Dispersion Sci. Technol. 30 (2009) 1363–1383;
(b) J.P. Jain, M. Sokolsky, N. Kumar, A.J. Domb, Polym. Rev. 48 (2008) 156–191;
(c) M.A.R. Meier, J.O. Metzger, U.S. Schubert, Chem. Soc. Rev. 36 (2007)
1788–1802;
(d) J.C. Mol, Green Chem. 4 (2002) 5–13;