Food Chemistry 135 (2012) 799805

Contents lists available at SciVerse ScienceDirect

Food Chemistry
journal homepage: www.elsevier.com/locate/foodchem

The effects of physical rening on the formation of

3-monochloropropane-1,2-diol esters in relation to palm oil minor components
Musrah Zulkurnain a,b, Oi Ming Lai c, Razam Abdul Latip d, Imededdine Arbi Nehdi e, Tau Chuan Ling f,
Chin Ping Tan a,
a

Department of Food Technology, Faculty of Food Science & Technology, Universiti Putra Malaysia, 43400 UPM Serdang, Selangor, Malaysia
Department of Food Technology, School of Industrial Technology, Universiti Sains Malaysia, 11800 Minden, Penang, Malaysia
Department of Bioprocess Technology, Faculty of Biotechnology and Biomolecular Sciences, Universiti Putra Malaysia, 43400 UPM Serdang, Selangor, Malaysia
d
Sime Darby Research Sdn. Bhd., Golden Jomalina Food Industries Sdn. Bhd., Batu 9, Jalan Banting-Kelang, 42500 Teluk Panglima Garang Industrial Estate, Selangor, Malaysia
e
Department of Chemistry, College of Science, King Saud University, Riyadh 1145, Saudi Arabia
f
Institute of Biological Sciences, Faculty of Science, University of Malaya, 50603 Kuala Lumpur, Malaysia
b
c

a r t i c l e

i n f o

Article history:
Received 2 December 2011
Received in revised form 16 April 2012
Accepted 30 April 2012
Available online 11 May 2012
Keywords:
3-MCPD esters
Precursor
D-optimal design
Water degumming
Adsorbents
Crude palm oil

a b s t r a c t
The formation of 3-monochloropropane-1,2-diol (3-MCPD) esters in rened palm oil during deodorisation is attributed to the intrinsic composition of crude palm oil. Utilising D-optimal design, the effects
of the degumming and bleaching processes on the reduction in 3-MCPD ester formation in rened palm
oil from poor-quality crude palm oil were studied relative to the palm oil minor components that are
likely to be their precursors. Water degumming remarkably reduced 3-MCPD ester formation by up to
84%, from 9.79 mg/kg to 1.55 mg/kg. Bleaching with synthetic magnesium silicate caused a further 10%
reduction, to 0.487 mg/kg. The reduction in 3-MCPD ester formation could be due to the removal of
related precursors prior to the deodorisation step. The phosphorus content of bleached palm oil showed
a signicant correlation with 3-MCPD ester formation.
2012 Elsevier Ltd. All rights reserved.

1. Introduction
The presence of 3-monochloropropane-1,2-diol (3-MCPD) esters in most rened oils and fats is an important issue in the fat
and oil industry. The carcinogenicity of choropropanol shows
long-term effects in animal studies, including infertility and malfunction of certain organs, such as the kidney (Hwang, Yoon,
Kim, Jang, & Yoo, 2009). The 3-MCPD ester contents of various oils
have been reported by a number of research institutes and industry
laboratories. Rened palm oil was reported to have the highest level of 3-MCPD esters (Matthus, Pudel, Fehling, Vosmann, & Freudenstein, 2011; Zelinkov, Svejkovsk, Velek, & Dolezal, 2006).
The reason for the presence of 3-MCPD esters is unclear because
the mechanism of formation of chloroesters in a lipid environment
is not as straightforward as that in the aqueous phase.
A recent mechanism study revealed that the formation of 3MCPD esters in a hydrophobic environment, via acyloxonium ion
Abbreviations: 3-MCPD, 3-monochloropropane-1,2-diol; GCMS/MS, gas chromatography tandem mass spectrometry; CPO, crude palm oil; RBD PO, rened
bleached deodorised palm oil; DAG, diacylglycerol; MAG, monoacylglycerol; FFA,
free fatty acid; PV, peroxide value.
Corresponding author. Tel.: +60 3 89468418; fax: +60 3 89423552.
E-mail addresses: tancp@putra.upm.edu.my, tcping@yahoo.com (C.P. Tan).
0308-8146/$ - see front matter 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.foodchem.2012.04.144

formation, can activate lipids toward the nucleophilic attack of

chloride ions, which is applicable to the formation of chloroesters
during the rening of vegetable oils (Rahn & Yaylayan, 2011b).
The formation of acyloxonium ion from a related precursor is inuenced by the steric hindrance effect and the leaving group involved. Partial acylglycerols, phospholipids, carotenoids and
other lipid-related compounds are suspected to be the precursors
of chloroesters (Collier, Cromie, & Davies, 1991; Hamlet, Sadd, &
Gray, 2004; ILSI, 2009; Rahn & Yaylayan, 2011a).
The physical rening of palm oil consists of three main processing steps: degumming and bleaching steps (in the pre-treatment
step) and deodorisation, for the purication of crude palm oil into
edible oil. Dry acid degumming is usually practised, whereby concentrated phosphoric acid is used in combination with adsorbents
to remove gums and impurities. The pre-treatment step is the most
crucial step for sufcient removal of impurities and undesired
components that would adversely affect the quality of the endproduct when subjected to deodorisation.
The formation of 3-MCPD esters in oil rening is associated with
high-temperature processing, mainly during the deodorisation
step. However, 3-MCPD ester precursors could also be formed at
other stages of rening, prior to deodorisation. The degumming
process, using acid, could activate the formation of 3-MCPD ester

800

M. Zulkurnain et al. / Food Chemistry 135 (2012) 799805

precursors (Schurz, 2010). On the other hand, bleaching with acidactivated bleaching earth could generate the protonation of organic
compounds, such as acylglycerols, into 3-MCPD ester precursors,
imparting a strong correlation between bleaching earth acidity
and 3-MCPD ester formation during palm oil rening (Ramli et al.,
2011). However, a high dosage of activated clay can adsorb some
of the 3-MCPD ester precursors from the oil (Schurz, 2010).
In this study, the formation of 3-MCPD esters in the palm oil
rening process was assessed for a possible relationship with selected minor components of palm oil that are likely to be potential
3-MCPD ester precursors. Among the proposed precursors are partial acylglycerol (mono- and di-acylglycerol) (Rahn & Yaylayan,
2011b), phospholipids (Collier et al., 1991; Hamlet et al., 2004),
and carotenoids (ILSI, 2009). Use of a D-optimal design modelling
of the experiment (involving a categorical factor) is possible with
a reduced number of experimental sets (Hoskins, Colbourn, &
Montgomery, 2005). This method allows a detailed analysis of
the effect of related 3-MCPD ester precursors on the level of 3MCPD esters formed in the nal rened oil when different degumming and bleaching treatments are employed.
2. Materials and methods
2.1. Materials
Crude palm oil (CPO) was supplied by Sime Darby Golden Jomalina Sdn. Bhd. (Teluk Panglima Garang, Selangor). The chemical,
3-chloropropane-1,2-diol (3-MCPD), was purchased from Aldrich
(France), and isotopically labelled 3-chloropropane-1,2-diol (3MCPD-d5) internal standard was purchased from Cambridge Isotope Laboratories (CIL, Massachusetts, USA). Phenylboronic acid
(PBA) of 99.0% purity was purchased from Fluka (Shanghai, China). All other chemicals and solvents were of analytical grade.
Synthetic magnesium silicate, Magnesol R60 (Dallas Group of
America, Inc., New Jersey, USA), was provided by Global Speciality
Ingredients (M) Sdn. Bhd. (Port Klang, Selangor) as a Malaysian
representative of Dallas Group of America, Inc. (New Jersey,
USA). Activated carbon was provided by Efgen Carbon Sdn Bhd
(Kapar, Selangor). Taiko Supreme 1B activated clay (Taiko Group
Sdn. Bhd, Pasir Gudang, Johor), Taiko TPT 110 non-activated natural clay (Taiko Group Sdn. Bhd, Pasir Gudang, Johor) and phosphoric acid, 85% (Merck, Darmstadt, Germany), were supplied
by Sime Darby Golden Jomalina Sdn. Bhd. (Teluk Panglima
Garang, Selangor).
2.2. Laboratory-scale physical rening
Four different grades of CPO were rened with standard conditions that are commonly practised in the industry. Approximately
700 g of crude palm oil were degummed and bleached in a 2 l lab-

oratory-scale rening ask with a stirrer and heating mantle that

was connected to a vacuum pump (Shidae Electric Co., Ltd.,
Gyeonggi-do, Korea). Under vacuum, the CPO was heated to
70 C before 0.05% w/w phosphoric acid was added for degumming
for 20 min. Subsequently, bleaching was performed at 95 C for
30 min with a 1% dosage of activated bleaching earth to obtain
bleached palm oil (BPO). Subsequently, 100 g of the BPO sample
were withdrawn for a quality analysis before proceeding to the
deodorisation stage. For water degumming, 1% water was added,
and then the hydrated gums were removed by centrifugation at
3570 g for 15 min on Rotina 46 (Hettich Zentrifugen GmbH & Co.,
KG, Tuttlingen, Germany).
BPO was deodorised at 260 C for 90 min to obtain rened
bleached and deodorised (RBD) palm oil in a custom-made glass
deodoriser consisting of a 1 l ask equipped with a built-in sparging steam inlet, a thermocouple and a 1 cm diameter vapour duct.
The deodoriser was connected through three-stage water-jacketed
traps, two empty ask traps and a vacuum pump (Shidae Electric
Co., Ltd., Gyeonggi-do, Korea). Stripping steam was produced from
distiled water and evaporated in a glass steam generator that was
connected directly to the deodoriser steam inlet. Both the steam
generator and the deodorising ask were heated in a temperature-controlled heating mantle. The deodoriser was set to ve different temperatures, ranging from 180 to 270 C. Thereafter, the oil
was kept frozen at 18 C prior to analysis. The quality characteristics of RBD palm oil, such as colour, free fatty acid (FFA) and peroxide value, were analysed.
2.3. Experimental design and optimisation using a D-optimal design
D-optimal design was applied to model the effects of the phosphoric acid percentage (X1) and different types of bleaching adsorbents (X2) and palm oil minor components, namely diacylglycerol
(DAG), monoacylglycerol (MAG), carotene, phosphorus and free
fatty acid (FFA) content in BPO on the 3-MCPD ester levels of
RBD palm oil. The processes of degumming and bleaching were
studied simultaneously, using three phosphoric acid dosages (0
0.1%) and four different bleaching adsorbents at levels of 1%: activated clay (Taiko Supreme 1B), natural clay (Taiko TPT 110), activated carbon mesh size 20 (0.85 mm) and synthetic magnesium
silicate (Magnesol R60). During the degumming process, 0% acid
represents water degumming, while acid degumming was studied
with two different concentrations of phosphoric acid.
A quadratic model was selected for second-order polynomial
modelling because it is adequate to represent the responses in
the region of interest. D-optimal design consists of 15 experimental runs with ten factorial and ve centre points, as shown in Table 2. The data from the experiments were analysed using Design
Expert version 6.0.10 (Stat-Ease, Inc., Minneapolis, MN USA) with
the level of signicance set at a 95% condence level for all tests.

Table 1
Quality characteristics and composition of four different crude palm oils.
Quality characteristics

Premium quality (PQ)

Superior quality (SQ)

Standard quality I

Standard quality II

3-MCPD esters of CPO [mg/kg]

3-MCPD esters of RBD PO [mg/kg]
Free fatty acid (FFA) [%]
Peroxide value (PV) [meq/kg]
DOBI index
Phosphorus (P) [ppm]
b-Carotene content [ppm]
Diglycerol [%]
Monoglycerol [%]

<LOD
1.54 ( 0.08)a
1.19 ( 0.02)a
Nil
3.2 ( 0.04)b
3.8 ( 0.3)a
615 ( 1.0)b
3.72 ( 0.02)b
0.0a

<LOD
1.49 ( 0.05)a
1.34 ( 0.02)a
Nil
3.0 ( 0.04)b
4.4 ( 0.4)a
611 ( 1.0)b
3.65 ( 0.02)b
0.0a

<LOD
1.72 ( 0.22)a
3.52 ( 0.03)b
Nil
2.9 ( 0.08)b
4.8 ( 0.3)a
597 ( 1.4)b
5.63 ( 0.02)c
0.05 ( 0.02)a

0.061 0.01
5.93 ( 0.41)b
4.19 ( 0.01)b
0.90 0.01
2.4 ( 0.12)a
8.8 ( 0.3)b
476 ( 1.2)a
2.20 ( 0.02)a
0.54 ( 0.01)b

DOBI, Deterioration of bleachability index.

LOD = 0.006 mg/kg.
Standard deviations between batches are given in parenthesis; results represent the means of two replicate trials. For each column, means with the same letter do not differ
signicantly at p 6 0.05.

801

M. Zulkurnain et al. / Food Chemistry 135 (2012) 799805

Table 2
Matrix of the D-optimal design and six response variables.
No. of run

Independent variablesa
Percent acid
(X1, %)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
a

0
0
0
0
0
0.05
0.05
0.05
0.05
0.1
0.1
0.1
0.1
0.1
0.1

Bleaching
adsorbent (X2)

A
B
C
D
D
A
B
C
D
A
A
B
C
C
D

Response variables
Characteristic of RBD palm oil

Characteristics of bleached palm oil (BPO)

3-MCPD
(Y1, mg/kg)

Colour
(Y2, R)

DAG
(Y3, %)

MAG
(Y4, %)

Carotene
(Y5, ppm)

Phosphorus
(Y6, ppm)

1.55 0.05
1.75 0.08
4.32 0.08
0.46 0.05
0.51 0.07
9.79 0.05
10.3 0.05
14.0 0.06
8.24 0.04
9.28 0.07
10.5 0.1
10.1 0.04
11.7 0.1
9.75 0.04
7.50 0.05

4.0
4.4
5.9
4.2
4.3
3.1
3.4
5.8
3.2
2.0
2.0
2.5
5.2
5.2
2.3

3.38 0.02
2.55 0.03
2.21 0.04
2.50 0.03
2.45 0.02
2.95 0.02
2.45 0.03
2.26 0.02
2.45 0.03
2.36 0.01
2.35 0.01
2.21 0.02
2.32 0.02
2.35 0.02
2.68 0.03

0.35 0.02
0.18 0.01
0.16 0.02
0.14 0.01
0.12 0.01
0.31 0.01
0.25 0.01
0.19 0.01
0.15 0.01
0.42 0.01
0.39 0.01
0.19 0.02
0.18 0.02
0.20 0.02
0.18 0.01

436 1.0
408 0.9
425 1.0
454 2.0
461 2.0
389 2.6
391 2.0
405 2.0
443 2.7
375 2.1
382 1.9
380 2.5
405 1.4
408 1.9
438 1.5

4.52 0.08
4.66 0.10
7.77 0.09
3.01 0.09
3.02 0.09
5.98 0.08
8.26 0.05
16.1 0.14
6.78 0.10
7.77 0.12
7.70 0.10
10.7 0.08
20.8 0.09
22.0 0.10
6.51 0.07

For factor X2, A, activated clay; B, natural clay; C, activated carbon; D, magnesium silicate.

The responses were tted into the following quadratic polynomial

regression model:

Y i b0 bi X i bii X 2i bij X i X j ;

where Yi is the response and b0, bi, bii, and bij are the constant coefcients of intercept, linear, quadratic and interaction terms, respectively. Xi and Xj are independent variables. Each model was analysed
using ANOVA, a regression analysis and graphical plots. The adequacy of the models was evaluated by the lack of t test, F-test
and ANOVA, and the overall predictive capability of the model
was explained by the coefcient of determination R2.
All of the experimental results in this study are reported as the
means standard deviations (SD) of triplicate analyses. Signicance was established at p < 0.05 for Tukeys test for ANOVA, using
MINITAB statistical software (Version 14, Minitab Inc., PA, USA).
The Pearson correlation coefcients between different response
variables and the 3-MCPD content were analysed using the same
statistical software programme.
2.4. 3-MCPD ester determination using GCMS/MS
The relative 3-MCPD ester contents of CPO and RBD palm oil of
each experimental run were determined using the modied BfR
Method 008 (BfR. Federal Institute for Risk Assessment: BfR Method_82_FC-008-01, 2009). The determination of 3-MCPD esters was
carried out based on an indirect method using acid transesterication, which involves the release of 3-MCPD from its esters, purication by ammonium sulphate extraction, PBA derivatisation and
quantication using gas chromatography with tandem mass spectrometry detection (GCMS/MS). A quantitative analysis of the derivatised 3-MCPD compound on a Thermo TSQ Quantum XLS Triple
Quadruple GCMS/MS system (Waltham, MA, USA) was carried out
by monitoring characteristic ions (quantier) at m/z 147 of 3MCPD and m/z 150 of isotopically labelled internal standard (3MCPD-d5). Ions at m/z 91 and 196 (3-MCPD) and m/z 93 and 201
(3-MCPD-d5) were used as qualiers. The 3-MCPD content was calculated as the ratio of the peak area responses of 3-MCPD (m/z
147) over the internal standard (m/z 150) for the calibration standards, blank and samples. The respective content was determined
from a matrix-based calibration graph that was constructed by
plotting the peak area ratios for the calibration standards (y-axis)
against the amount of 3-MCPD (x-axis).

2.5. Oil quality analysis

The free fatty acid level, peroxide value and colour of CPO, BPO
and RBD PO of each run were determined according to the AOCS
(2009) Ca5a-40, Cd 8b-90 and Cc 13e-92, respectively. The deterioration of the bleachability index (DOBI) for CPO was conducted
according to MPOB (2005). All analyses were performed in triplicate, and the mean values are reported.
2.6. Minor component contents of BPO
2.6.1. Triacylglycerol composition analysis using HPLC
Analyses of monoacylglycerol (MAG), diacylglycerol (DAG) and
triacylglycerol (TAG) of BPO, in every run, were performed using
high-performance liquid chromatography (HPLC) with an evaporative light scattering detector (ELSD). The separation of TAGs, DAGs,
and MAGs was performed on a Purospher STAR RP-18 column
(10 cm; 2.1 cm, i.d.) (Merck, Darmstadt, Germany). An isocratic
binary solvent system of acetone (A) and acetonitrile (B) at a ratio
of 62.5:37.5 (v/v) was used at a ow rate of 1.5 ml/min. A total of
100 ll of the oil sample was dissolved in 900 ll chloroform, and
5 ll aliquots were injected for HPLC analysis. The levels of TAGs,
DAGs, and MAGs are given as % m/m of the total acylglycerol content in the lipid sample.
2.6.2. Carotene content
The carotene contents in the crude oil and BPO of each experimental run were measured with a Hewlett Packard 8452A diode
array UV/vis spectrophotometer, according to MPOB Test Methods
(2005). A total of 0.15 g BPO was weighed in a 25 ml volumetric
ask and dissolved in iso-octane. The absorbance of b-carotene at
446 nm was obtained for each sample and converted to b-carotene
content in ppm, using the following equation:Carotene content
(ppm) = 25  Abs  10 000/(Mass  2610)
Each sample was measured in triplicate, and the mean values
are reported.
2.6.3. Phosphorus content
The total phosphorus content was determined according to
MPOB Test Methods (2005). Approximately 5 g of sample were
ashed in the presence of 0.1 g of magnesium oxide, and the ash
was acid-digested with 5 ml of 6 M nitric acid. Yellow phosphovanadomolybdic complex was formed by adding a mixture of 10 ml

M. Zulkurnain et al. / Food Chemistry 135 (2012) 799805

of 5% ammonium molybdate solution and 10 ml of ammonium

vanadate solution and was then determined calorimetrically using
a UVvis spectrophotometer at 400 nm after a 20 min reaction. The
absorbance obtained for each sample was converted to phosphorus
content in ppm, using the following equation:

Phosphorus contentppm 25  Abs  CF=Mass;

where CF is the calibration factor obtained from the reciprocal of
the gradient of a calibration curve built from a series of concentrations of phosphate solution (4, 8, 12 and 16 lg/ml) reacted with a
20 ml mixture solution of ammonium molybdate and ammonium
vanadate solution and 6 M nitric acid. Each sample was analysed
in triplicate, and the mean values are reported.

3. Results and discussion

3.1. General
There are many factors that lead to the formation of 3-MCPD esters in oil rening (Matthus et al., 2011; Pudel et al., 2011; Hrncirik & van Duijn, 2011). In this study, the palm oil rening process
was scrutinised for 3-MCPD ester formation with regard to the processing parameters and in relation to the possible precursors of the
chloroester mechanism of formation. The formation of 3-MCPD esters at different levels in rened oils of various sources could be
attributed to the precursor contents found in different types of
raw materials (Matthus et al., 2011). Partial acylglycerides, phospholipids and carotenoids are among the suspected precursors
(Hamlet et al., 2004; ILSI, 2009) that are present in considerable
amounts in CPO.
3.2. Effects of raw material quality on 3-MCPD ester formation
The quality of CPO as a raw material of the palm oil rening process was studied for the tendency toward 3-MCPD ester formation.
Four CPO samples of different qualities were rened under standard conditions: degumming at 70 C for 20 minutes with 0.05%
phosphoric acid, bleaching at 100 C for 30 min with 1% activated
bleaching clay, and deodorisation at 260 C for 90 min (Gibon, De
Greyt, & Kellens, 2007). The CPO quality characteristics and 3MCPD ester level of RBD palm oil attained were compared, as
shown in Table 1. There was no signicant difference in the quality
characteristics between the Premium Quality (PQ) CPO, which is
the rst-grade CPO, and Superior Quality (SQ) CPO, which is the
second-grade CPO (p > 0.05). They possessed low FFA and phosphorus levels, an absence of monoglycerides, and high carotene
contents. The third-grade CPO, Standard Quality I CPO, contained
higher levels of these components, and the worst-grade CPO, Standard Quality II CPO, had far higher levels than had the Standard
Quality I CPO.
There were no signicant differences among the 3-MCPD ester
levels of RBD palm oil attained from CPO of the rst three quality
grades (p > 0.05). However, RBD palm oil from Standard Quality II
CPO contained 3-MCPD esters at ve times higher levels than those
of other CPO samples. The 3-MCPD esters were also detected in the
Standard Quality II CPO in trace amount. The MAG and phosphorus
contents of the CPO were very high compared to other samples.
This nding indicates that CPO of poor quality has a high tendency
toward 3-MCPD ester formation, even in its unrened form. The
intrinsic composition of the CPO could contribute to 3-MCPD ester
formation and requires further study. The relationships of these
minor components of palm oil to 3-MCPD ester formation were
studied by utilising different treatments of degumming and
bleaching processes in this study.

Further analysis was carried out to detect the effect of increasing deodorisation temperature on 3-MCPD ester formation in Standard Quality II CPO. As shown in Fig. 1, the formation of 3-MCPD
esters signicantly increased as the deodorisation temperature increased. The level of chloroesters reached 9 mg/kg when the sample was deodorised at 250 C. Further increases in temperature,
however, led to the degradation of 3-MCPD esters, as shown previously (Matthus et al., 2011; Pudel et al., 2011).
3.3. Effects of degumming and bleaching on the reduction of 3-MCPD
esters
3.3.1. General
The degumming and bleaching processes are steps that could
potentially reduce 3-MCPD ester formation by removing related
precursors prior to deodorisation. Utilising D-optimal design, the
effects of different acid dosages and bleaching adsorbents on the
reduction of 3-MCPD esters were scrutinised, using low-grade
CPO (Standard Quality II). The deodorisation temperature was set
at 250 C because the maximum formation of 3-MCPD esters occurred at this temperature.
3.3.2. Model tting
Table 2 shows the responses of D-optimal design, namely 3MCPD content (Y1), the colour (Y2) of RBD palm oil, and the BPO
DAG (Y3), MAG (Y4), carotene (Y5) and phosphorus (Y6) contents.
All responses were signicantly tted into reduced quadratic polynomial models (p < 0.05). The best tting models were determined
by a regression analysis with backward elimination in which insignicant factors were deleted from the initial model to yield the
most signicant nal reduced models. The ANOVA results of the nal reduced models, the p-value of signicant terms, the lack of t
test, the F-test and the determination coefcient R2 of the responses are shown in Table 3.
The 3-MCPD levels (Y1) were signicantly (p < 0.05) modelled
into a polynomial model with signicant quadratic terms. The lack
of t of the model was not signicant (p > 0.05). The R2 value was
0.974 with no observed outliers, suggesting that the data were well
tted to the described model. The computed model F-value of 27.9,
in mg/kg of 3-MCPD, was higher than the tabular value of
F0.05(6,3) = 8.94, implying that the model was signicant at the
95% condence level. For the other responses, the F-values of the
nal reduced model ranged from 35.46 to 255.23, and the R2 values
were close to one, implying that the models are signicant.
3.4. Effects of degumming and bleaching on the formation of 3-MCPD
esters
The formation of 3-MCPD esters during deodorisation was inuenced by the degumming and bleaching parameters, as depicted in
Fig. 2(a). Water degumming signicantly reduced 3-MCPD ester
3-MCPD esters (mg/kg)

802

12
d

10
6

b
a

4
2
0
180 C

200 C

230 C

250 C

270 C

Temperature (C)
Fig. 1. The 3-MCPD esters level of rened bleached deodorised palm oil (RBD PO) at
different deodorisation temperatures. (Same letters do not differ signicantly at
p 6 0.05).

803

M. Zulkurnain et al. / Food Chemistry 135 (2012) 799805

Table 3
The ANOVA, p-value, lack of t test, F-test and the determination coefcient, R2, of six response variables of the nal reduced model.
Final reduced model

3-MCPD (mg/kg, Y1)

Colour (%, Y2)

DAG (%, Y3)

MAG (%, Y4)

Carotene (ppm, Y5)

Phosphorus (ppm, Y6)

Model
p-value
F-value
X1
X2
X1X2
X12
R2
Adjusted R2
Predicted R2
Lack of t (p-value)
Lack of t (F-value)

Quadratic
0.0002a
36.45
<0.0001a
0.0081a
0.481
0.0001a
0.980
0.953
0.829
0.482
1.10

2FI
<0.0003a
27.72
<0.0001a
<0.0001a
0.0003a

0.997
0.995
0.982
0.053
8.77

2FI
<0.0001a
60.42
0.0001a
<0.0001a
<0.0001a

0.984
0.967
0.873
0.054
8.62

Linear
<0.0001a
35.46
0.0481a
<0.0001a

0.934
0.908
0.855
0.152
3.75

Quadratic
<0.0001a
42.57
0.0003a
<0.0001a
0.0093a
0.0182a
0.983
0.960
0.789
0.244
2.42

Quadratic
<0.0001a
138.36
<0.0001a
<0.0001a
0.0004a
0.0460a
0.995
0.987
0.942
0.207
2.85

2FI, two factors interaction.

For factor X1: Acid concentration (%); X2: bleaching adsorbent type.
a
Signicant (p < 0.05).

DESIGN-EXPERT Plot

Interaction graph

3-MCPD ester (mg/kg)

B: Absorbents

14.0

DESIGN-EXPERT Plot

X = A: Acid %
Y = B: Absorbents

Design Points
B1 Activated Clay
B2 Natural Clay
B3 Activ Carbon
B4 Mg silicate

6.5

393

-0.9

369
0.03

(a)

0.05

0.08

0.10

Interaction graph

0.05

0.08

0.10

Interaction graph

Color (R)

X = A: Acid %
Y = B: Absorbents

B: Absorbents

6.1

X = A: Acid %
Y = B: Absorbents

Phosphorus (ppm)

Design Points

17.5

Design Points

12.4

B1 Activated Clay
B2 Natural Clay
B3 Activ Carbon
B4 Mg silicate

4.0

2.9

2.3

1.9

0.03

0.05

A: Acid %

0.08

5.1

7.3

0.00

(b)

0.03

A: Acid %

DESIGN-EXPERT Plot

B: Absorbents

22.5

0.00

(c)

A: Acid %

DESIGN-EXPERT Plot

B1 Activated Clay
B2 Natural Clay
B3 Activ Carbon
B4 Mg silicate

417

2.8

0.00

Phosphorus (ppm)

441

Carotene (ppm)

10.3

Color (R)

B1 Activated Clay
B2 Natural Clay
B3 Activ Carbon
B4 Mg silicate

B: Absorbents

465

X = A: Acid %
Y = B: Absorbents

3-MCPD ester (mg/kg)

Design Points

Interaction graph

Carotene

0.00

0.10

(d)

0.03

0.05

0.08

0.10

A: Acid %

Fig. 2. Interaction plots showing the effects of degumming and bleaching treatments on: (a) 3-MCPD levels in rened bleached deodorised palm oil (RBD PO), (b) phosphorus
content and, (c) carotene content in bleached palm oil (BPO), and (d) colour of RBD PO.

M. Zulkurnain et al. / Food Chemistry 135 (2012) 799805

mation in the nal rened oil at a 3-MCPD ester level below

8 mg/kg (Matthus et al., 2011). However, the DAG content ranged
between 2.11% and 3.38%, resulting in a 3-MCPD ester level
exceeding 10 mg/kg. In addition, the DAG content showed a poor
correlation with the 3-MCPD ester level in RBD palm oil with a
non-signicant Pearsons correlation of -0.424 (p > 0.05), as depicted in Fig. 3(a). The MAG content also showed a non-signicant
Pearsons correlation of 0.297 (p > 0.05). This nding agrees with
the nding of Strijowski, Heinz, and Franke, (2011) that no direct
correlation was observed between the level of partial glycerides
and polar compounds with the level of 3-MCPD esters.
The phosphorus content exhibited a pattern of rening effects
similar to the 3-MCPD levels in RBD palm oil, as shown in
Fig. 2(b). This nding resulted in a signicant 70% correlation with
the 3-MCPD ester level (p < 0.001), as depicted in Fig. 3(b). Likewise, water degumming resulted in the lowest phosphorus

DESIGN-EXPERT Plot
Correlation: -0.393

10.6

7.2

3.9

0.5
2.21

(a)

2.50

2.79

3.09

3.38

Diacy gly cerol (%)

DESIGN-EXPERT Plot
Correlation: 0.683

3.5. Inuences of palm oil minor components on 3-MCPD ester

formation
Utilising D-optimal design, the palm oil minor components in
BPO were studied in relation to 3-MCPD ester formation. Small
variations in the DAG content from the initial value in CPO (2.2%)
were observed because the DAG content is not strongly affected
by the rening process (Goh & Timms, 1985). Different adsorbents
showed distinct patterns of DAG content at increasing levels of acid
degumming. Activated clay and natural clay showed a reduction in
the DAG content, while magnesium silicate and activated carbon
showed increasing DAG contents.
A threshold of 4% DAG in the raw material was suggested in a
previous study to exhibit a lower potential for 3-MCPD ester for-

14.0

3-MCPD ester (mg/kg)

formation from 9.79 to 1.55 mg/kg (p < 0.05). This reduction could
be because reactants such as chloride, which contributes to the formation of 3-MCPD esters, are washed off during water degumming
(Matthus et al., 2011). The phosphoric acid dosage had a greater
effect on the development of chloroesters than did the different
types of bleaching adsorbents. The increase in acid dosage had
no signicant effect on the 3-MCPD ester level (p > 0.05), with
the exceptions of the activated carbon and magnesium silicate
conditions.
Bleaching, using magnesium silicate (Magnesol R60), showed
the greatest reduction of 3-MCPD esters compared to activated
clay, natural clay and activated carbon, especially when water
degumming was used. Magnesium silicate possesses a relatively
large surface area and the most active basic sites compared to alumina, silica and activated carbon (Zhu, Yates, & Caldwell, 1994).
These properties may be responsible for the adsorption or degradation of particular precursors of 3-MCPD esters from BPO prior to
deodorisation.
Haines et al., (2011) indicated that 3-MCPD was unstable above
pH 6. High pH conditions may assist in the removal of related precursors and cause a reduction of 3-MCPD esters. However, when
acid was used during degumming, the pH conditions of bleaching
dropped and reduced the removal effect. The pH values of the
bleaching adsorbents were tested at 6.0, 9.0, 9.5 and 9.4 for activated clay, natural clay, activated carbon and magnesium silicate,
respectively. However, the effect of pH on 3-MCPD ester reduction,
as reported previously (Ramli et al., 2011), was only signicant for
magnesium silicate. Bleaching with natural clay slightly increased
3-MCPD ester formation compared to activated clay, which possessed a much lower pH. This nding may also suggest that acid
activation of bleaching earth, using sulphuric acid, does not contribute to the activation of 3-MCPD ester precursors during bleaching. On the other hand, bleaching with activated carbon resulted in
the formation of high amounts of MCPD esters, up to 14 mg/kg. A
similar observation was reported by Haines et al., (2011) but with
higher amounts of 3-MCPD esters because a direct method of 3MCPD ester detection was used. Although activated carbon possesses the highest pore volume, which allows greater adsorption
of precursors and impurities, the chemistry of the precursors and
their mechanisms of adsorption could be among the factors that
require further investigation.
The quality characteristics of RBD palm oils were tested and
compared. PV was not present for all rening treatments, indicating that all of the rening treatments sufciently remove oxidative
compounds from crude oil, and no signicant difference was observed for the FFA content (p > 0.05). Water degumming produced
a strong colour, in the range of 4.0 to 5.9 red, suggesting that colour
xation might occur during deodorisation due to an insufcient
degumming treatment. Activated clay resulted in excellent pigment removal, while magnesium silicate caused a slightly stronger
colour, as depicted in Fig. 2(d).

14.0

3-MCPD ester (mg/kg)

804

10.6

7.2

3.9

0.5
3.0

(b)

7.8

12.5

17.2

22.0

Phosphorus (ppm)

Fig. 3. Correlation of 3-MCPD levels with (a) diacylglycerol, and (b) phosphorus
content in bleached palm oil (BPO).

M. Zulkurnain et al. / Food Chemistry 135 (2012) 799805

content, and bleaching with magnesium silicate removed most

phospholipids, followed by activated clay, natural clay, and activated carbon. This nding suggests that phospholipids might be
one of the precursors of 3-MCPD ester formation, but this hypothesis has yet to be veried.
The total phosphorus content measures the total phosphatides
present in the oil, including phospholipids and some inorganic
phosphates (Goh, Tong, & Gee, 1984). In palm oil, inorganic phosphate accounts for most of the total phosphorus content, while
phospholipids are present at only 1030% of the total phosphorus
(Gibon et al., 2007). Water degumming efciently removed
hydratable phospholipids from 9.78 ppm to 1.55 ppm, while acid
degumming removed non-hydratable phospholipids. However,
employing concentrated phosphoric acid (>75%) increased the
residual phosphorus content to 9.9 ppm (Choukri, Kinany, Gibon,
Tirtiaux, & Jamil, 2001). Bleaching further removes non-hydratable
phospholipids by decomposition and adsorption on bleaching
earth (OBrien, 2009).
With respect to the carotene content, the bleaching adsorbents
imparted greater effects than did the degumming treatment,
whereby magnesium silicate retained most of the carotenes, and
bleaching clays removed most of the carotenes, as shown in
Fig. 2(c). Acid degumming reduced more carotene than did water
degumming. Carotenoid is proposed to act as an oil-soluble chlorinating agent that promotes 3-MCPD ester formation in palm oil
(ILSI, 2009). However, the carotene content showed a negative correlation with 3-MCPD ester levels, with a Pearsons correlation of 0.694 (p < 0.05).
4. Conclusions
The CPO quality determines the level of 3-MCPD ester formation in palm oil during deodorisation. CPO quality could be an initial indication of the chloroester level when the oil is subjected to
deodorisation under normal rening conditions. The intrinsic composition of palm oil is suspected to be one of the factors that lead to
this nding. A study of the relationships of minor components of
palm oil to the formation of 3-MCPD esters suggested a strong correlation with the BPO phosphorus content. Thus, future research
should consider the effect of phospholipids in the formation mechanism of 3-MCPD esters in palm oil.
The reduction of chloroester formation during deodorisation
could be achieved with water degumming instead of acid degumming and bleaching with adsorbents such as synthetic magnesium
silicate. These results are possibly due to the removal or inactivation of related 3-MCPD ester precursors prior to deodorisation,
caused either by the pH conditions or by other means. However,
the nal RBD palm oil produced is prone to strong colour
formation. The synergistic effects of the combined adsorbents of
magnesium silicate and activated clay could be benecial in reducing 3-MCPD ester formation and enhancing rened oil quality,
which is an important subject for future research.
Acknowledgments
Financial support of this work by the Ministry of Science, Technology and Innovation of Malaysia through Science Fund (05-0104-SF0384) and the Ministry of Higher Education through

805

Fundamental Research Grant Scheme (02-11-08-0619FR) is gratefully acknowledged.

References
AOCS (2009). Ofcial methods and recommended practices of the AOCS, (6th Ed.)
BfR. Federal Institute for Risk Assessment: BfR Method_82_FC-008-01 (2009).
Determination of 3-MCPD fatty acid esters in edible oils and solid fats by GC
MS-an indirect determination by detection of free 3-MCPD released from 3MCPD esters by acid hydrolysis and by derivatization with phenylboronic acid.
Dahlem, Germany
Choukri, A., Kinany, M. A., Gibon, V., Tirtiaux, A., & Jamil, S. (2001). Improved oil
treatment conditions for soft degumming. Journal of the American Oil Chemists
Society, 78, 11571160.
Collier, P. D., Cromie, D. D. O., & Davies, A. P. (1991). Mechanism of formation of
chloropropanols present in protein hydrolysates. Journal of the American Oil
Chemists Society, 68, 785790.
Gibon, V., De Greyt, W., & Kellens, M. (2007). Palm oil rening. European Journal of
Lipid Science and Technology, 109, 315335.
Goh, E. M., & Timms, R. E. (1985). Determination of mono- and diglycerides in palm
oil, olein, and stearin. Journal of the American Oil Chemists Society, 62, 730734.
Goh, S. H., Tong, S. L., & Gee, P. T. (1984). Total phospholipids in crude palm oilquantitative analysis and correlations with oil quality parameters. Journal of the
American Oil Chemists Society, 61, 15971600.
Haines, T. D., Adlaf, K. J., Pierceall, R. M., Lee, I., Venkitasubramanian, P., & Collison,
M. W. (2011). Direct determination of MCPD fatty acid esters and glycidyl fatty
acid esters in vegetable oils by LC-TOFMS. Journal of the American Oil Chemists
Society, 88, 114.
Hamlet, C. G., Sadd, P. A., & Gray, D. A. (2004). Generation of
monochloropropanediols (MCPDs) in model dough systems. 2. Unleavened
doughs. Journal of Agricultural and Food Chemistry, 52, 20672072.
Hoskins, D. S., Colbourn, C. J., & Montgomery, D. C. (2005). Software performance
testing using covering arrays: Efcient screening designs with categorical
factors. Proceedings of the second international workshop on software and
performance (WOSP 05). http://dx.doi.org/10.1145/1071021.1071034.
Hrncirik, K., & van Duijn, G. (2011). An initial study on the formation of 3-MCPD
esters during oil rening. European Journal of Lipid Science and Technology, 113,
374379.
Hwang, M., Yoon, E., Kim, J., Jang, D. D., & Yoo, T. M. (2009). Toxicity value for 3monochloropropane-1,2-diol using a benchmark dose methodology. Regulatory
Toxicology and Pharmacology, 53, 102106.
ILSI (2009). Workshop on 3-MCPD Esters in Food Products. 56 February, Brussels,
Belgium. http://www.ilsi.org/Europe/Documents/E2009MCPD-2.pdf. Accessed
April 2011.
Matthus, B., Pudel, F., Fehling, P., Vosmann, K., & Freudenstein, A. (2011). Strategies
for the reduction of 3-MCPD esters and related compounds in vegetable oils.
European Journal of Lipid Science and Technology, 113, 380386.
MPOB (2005). MPOB Test Methods. Palm oil research institute of Malaysia (PORIM),
(1995). PORIM Test Methods. Vol. 1. Palm oil research institute of Malaysia,
Malaysia.
OBrien, R. D. (2009). Fats and oils: Formulating and processing for applications. Boca
Raton: CRC Press.
Pudel, F., Benecke, P., Fehling, P., Freudenstein, A., Matthus, B., & Schwaf, A. (2011).
On the necessity of edible oil rening and possible sources of 3-MCPD and
glycidyl esters. European Journal of Lipid Science and Technology, 113, 368373.
Rahn, A. K. K., & Yaylayan, V. A. (2011a). What do we know about the molecular
mechanism of 3-MCPD ester formation? European Journal of Lipid Science and
Technology, 113(3), 323329.
Rahn, A. K. K., & Yaylayan, V. A. (2011b). Monitoring cyclic acyloxonium ion
formation in palmitin systems using infrared spectroscopy and isotope labelling
technique. European Journal of Lipid Science and Technology, 113, 330334.
Ramli, M. R., Siew, W. L., Ibrahim, N. A., Hussein, R., Kuntom, A., Abd. Razak, R. A., &
Nesaretnam, K. (2011). Effects of degumming and bleaching on 3-MCPD esters
formation during physical rening. Journal of the American Oil Chemists Society,
88, 18391844.
Schurz, K. (2010). Method for reducing the 3-MCPD content in rened vegetable
oils. Patent Cooperation Treaty Application (WO10063450).
Strijowski, U., Heinz, V., & Franke, K. (2011). Removal of 3-MCPD esters and related
substances after rening by adsorbent material. European Journal of Lipid Science
and Technology, 113, 387392.
Zelinkov, Z., Svejkovsk, B., Velek, M., & Dolezal, M. (2006). Fatty acid esters of 3chloropropane-1,2-diol in edible oils. Food Additives and Contaminants, 23,
12901298.
Zhu, Z. Y., Yates, R. A., & Caldwell, J. D. (1994). The determination of active lter aid
adsorption sites by temperature-programmed desorption. Journal of the
American Oil Chemists Society, 71, 189194.