Effects of deodorization temperature and time on

the formation of 3-MCPD, 2-MCPD, and glycidyl

esters and physicochemical changes of palm oil
Renan Tivanello , Maisa Capristo, Eduardo Vicente, Roseli Ferrari, Klicia Sampaio, and Adriana Arisseto

Abstract: This study verified the formation profile of esters of 3-monochloropropane-1,2-diol (3-MCPDE), 2-
monochloropropane-1,2-diol (2-MCPDE), and glycidol (GE), and evaluated the physicochemical changes (free-fatty
acid, acylglycerols, and colour) that occurred during the deodorization of palm oil in different conditions of time
(30, 60, 90, and 120 min) and temperature (210, 230, 250, and 270 °C). Levels of 3-MCPD and 2-MCPD esters
ranged from 1.91 to 2.70 mg/kg and 0.68 to 1.44 mg/kg, respectively, and were formed at the mildest tested condition
(210 °C, 30 min). No correlation was observed between these contaminants and physicochemical changes. GE levels
varied from 0.12 to 8.51 mg/kg and showed correlation with color and diacylglycerol content. While the temperature
had little influence on the formation of esters of 3-MCPD and 2-MCPD, the content of GE considerably raised above
250 °C.

1. INTRODUCTION Authority (EFSA) and the Joint FAO/WHO Expert Committee

Elucidating the formation of toxic compounds during heat pro- on Food Additives (JECFA) published contrasting opinions on the
cessing of foods is important to reduce risks to human health. establishment of a safe intake parameter for 3-MCPD. EFSA es-
Among emerging processing contaminants, the bound forms tablished a tolerable daily intake (TDI) of 0.8 µg/kg body weight
of chloropropanols, particularly esters of 3-monochloropropane- (bw) in 2016 while JECFA set a provisional maximum tolera-
1,2-diol (3-MCPDE) and 2-monochloropropane-1,3-diol (2- ble daily intake (PMTDI) of 4 µg/kg bw in 2017 (EFSA, 2016;
MCPDE), have been intensively studied in the past few years and FAO/WHO, 2017). In 2018, EFSA re-evaluated the available tox-
have been reported in several processed foods and food ingredients icological data and established a TDI for 3-MCPD of 2 µg/kg bw
(Arisseto, Marcolino, & Vicente, 2014; B. Craft, Chiodini, Garst, (EFSA, 2018).
& Granvogl, 2013; Svejkovská et al., 2004; Zelinková, Svejkovská, There are some proposed mechanisms available in the litera-
Velı́šek, & Doležal, 2006). The formation of glycidyl esters (GE) ture to explain the formation of 3-MCPDE and GE in refined
has similarly attracted worldwide attention (Craft, Chiodini, Garst, oils (Craft & Destaillats, 2014; Destaillats, Craft, Sandoz, et al.,
& Granvogl, 2013; Hrncirik & van Duijn, 2011). 2012; Rahn & Yaylayan, 2011; Zulkurnain et al., 2012), but the
Refined vegetable oils have been reported with the highest knowledge of the exact precursors and conditions involved in the
concentrations of these contaminants, especially palm oil. De- generation of such contaminants are still in need further research.
odorization is the refining step that favors the formation of these Moreover, data on the formation of 2-MCPDE as well as the
compounds, as demonstrated by many authors (Franke, Strijowski, physicochemical changes that occur simultaneously to the forma-
Fleck, & Pudel, 2009; Hrncirik & van Duijn, 2011; Zelinková tion of the contaminants are something that has not been explored
et al., 2006). Chlorides and acylglycerols, mainly triacylglycerols so far. Therefore, the focus of this study is to verify the formation
(TAG), occur naturally in palm oil and can act as precursors of 3- profile of 3-MCPDE, 2-MCPDE, and GE in palm oil, as well
MCPD esters (Destaillats, Craft, Sandoz, & Nagy, 2012; Ermacora as its relation with physicochemical changes along the deodor-
& Hrncirik, 2014; Šmidrkal et al., 2016). High concentrations of ization process conducted under varying conditions of time and
diacylglycerols (DAG) and monoacylglycerols (MAG) play an im- temperature.
portant role in the formation of GE (Ermacora & Hrncirik, 2014).
Toxicological studies have confirmed adverse effects associated 2. MATERIAL AND METHODS
with free 3-MCPD, mainly nephrotoxicity (Cho et al., 2008),
while glycidol may cause DNA damage and mutation, as demon- 2.1 Palm oil sample
strated by positive genotoxicity results in in vitro and in vivo assays A company in the region of Limeira–SP (Brazil) provided the
(EFSA, 2016; FAO/WHO, 2017). The International Agency for bleached palm oil (BPO), which was previously characterized in
Research on Cancer (IARC) classifies 3-MCPD in group 2B (pos- relation to free-fatty acids (FFA), composition of acylglycerols,
sible human carcinogen) and glycidol in group 2A (probable hu- color, and concentrations of 3-MCPDE, 2-MCPDE, and GE.
man carcinogen) (IARC, 2012; 2000). The European Food Safety The bleaching was carried out at a temperature of 103 °C using
0.8% of natural clay (pH 7.2) from Oil-Dri (Chicago, USA) and
addition of 500 mg/kg of citric acid before the process.
JFDS-2020-0277 Submitted 2/18/2020, Accepted 4/30/2020. Faculty of Food
Engineering, University of Campinas (UNICAMP), Rua Monteiro Lobato 80, 2.2 Solvents and reagents
Campinas, Sao Paulo, 13083-862, Brazil. Food Science and Quality Center, In- Heptane (purity ࣙ 99%) and acetone were acquired from
stitute of Food Technology (ITAL), Avenida Brasil 2880, C.P. 139, Campinas, Scharlau Brand (Barcelona, Spain), while methanol was purchased
Sao Paulo, 13070-178, Brazil. Direct inquiries to author Tivanello (E-mail: re-
[email protected]). from J. T. Baker (Mexico City, Mexico). Tetrahydrofuran (THF,
inhibitor-free, purity ࣙ99.99%), toluene (purity ࣙ99.9%), sodium
Toxicology & Chemical
Food Safety

C 2020 Institute of Food Technologists

 R

doi: 10.1111/1750-3841.15304 Vol. 85, Iss. 7, 2020 r Journal of Food Science 2255
Further reproduction without permission is prohibited
 MCPD and glycidyl esters in palm oil . . .

bicarbonate, sodium bromide (NaBr, purity ࣙ99.5%), and phenyl- 2.7 Free-fatty acids (FFA)
boronic acid (PBA, purity ࣙ97%) were acquired from Sigma- Free-fatty acids were determined by the official AOCS method
Aldrich (St. Louis, MO, USA). Sodium sulfate and sulfuric acid Ca 5a-40 (AOCS, 1998) by titration using standard sodium hy-
were delivered from Merck (Darmstadt, Germany) and phenolph- droxide and indicator phenolphthalein. FFAs were quantified ac-
thalein was obtained from Synth (Diadema, SP, Brazil). A Milli-Q cording to the mass of the sample, the volume of sodium hydrox-
Plus system from Millipore (Bedford, MA, USA) was used to ide, and the molecular weight of the major fatty acid. The results
obtain ultrapure water. were expressed as % of palmitic acid.

2.3 Standards 2.8 Color

The standards rac 1,2-bis-palmitoyl-3-chloropropanediol (pu- The color of palm oil before and after deodorization fol-
rity 98%), 1,3-dipalmitoyl-2-chloropropanediol (purity 98%), lowed the official AOCS method Cc 13e-92 (AOCS, 2017). A
glycidyl palmitate (purity 98%), glycidyl palmitate-d5 (chem- Lovibond R
colorimeter was used to measure yellow (Y) and red
ical purity 97% and isotopic purity 97.6%), and rac 1,2-bis- (R) colors from scales ranging from 0 to 70 with a one inch bucket.
palmitoyl-3-chloropropanediol-d5 (purity 98%) were purchased
from Toronto Research Chemicals Inc (Ontario, Canada). Mono- 2.9 Statistical analyses
glyceride (MAG), diglyceride (DAG), and triglyceride (TAG) stan- Analyses was conducted with IBM SPSS Statistic 25 software
dards were purchased from Sigma-Aldrich Co. (IBM Corporation, New York, USA). Using Shapiro-Wilk’s and
Levene tests, the ANOVA requirements, including the homo-
2.4 Deodorization process geneity of variances and a normal distribution of residues, were
investigated. All dependent variables were interpreted using one-
The deodorization process was performed in a laboratory-scale
way ANOVA. Welch’s correction was used depending on if the
batch deodorizer which consisted of a flow meter, hoses, adapters,
requirement of variances homogeneity was met or not. If a signif-
thermometer, condenser, cold trap, vacuum pump, heating mantle,
icant statistical effect was found, the means were compared using
oil sample flask, flow control, and a vacuometer. Four different
the Tukey multiple-comparison or Dunnett T3 test also depend-
temperatures (210, 230, 250, and 270 °C) and times (30, 60, 90,
ing on if the equality of variances could be assumed or not. The
and 120 min) were applied to deodorize 100 g of BPO based on
significance level chosen was 0.05.
the ideal and extrapolated conditions. Deodorization followed the
A possible correlation between the physicochemical parameters
recommendations from Bailey’s Industrial Oil and Fat Products (De
and the levels of the contaminants were analyzed. Depending on
Greyt & Kellens, 2005): a fixed pressure in the range of 1 to 4 mbar
the statistical distribution of the data and distribution chart, a
and steam injection set around 1%. A nondeodorized sample was
Pearson or Spearman correlation was performed.
used as control (time zero).

3. RESULTS AND DISCUSSION

2.5 Determination of 3-MCPDE, 2-MCPDE, and GE In our study, the figures of merit of the analytical method used
The analyses of the contaminants were carried out by the of- to determine the contaminants were those reported by Kamikata
ficial AOCS method Cd 29a-13 (AOCS, 2013). The method et al. (2019). The LOD and LOQ were 0.01 and 0.03 mg/kg for 3-
consisted of the conversion of GE to esters of 3-bromopropane- MCPDE while for 2-MCPDE and GE these values were 0.02 and
1,2-diol (3-MBPDE) in the presence of acidic sodium bromide 0.08 mg/kg, respectively. Recoveries ranged from 92% to 104%,
solution, followed by acid-catalyzed transesterification and neu- reproducibility varied from 4% to 9% and repeatability was be-
tralization with sodium bicarbonate. Subsequently, a salting-out of tween 3% and 6%. The linear ranges for 3-MCPDE, 2-MCPDE,
the fatty acid methyl esters in the presence of sodium sulfate and and GE were 0.03 to 8.11 mg/kg, 0.08 to 8.26 mg/kg, and 0.08
derivatization of 3-MCPD, 2-MCPD, and 3-MBPD released with to 20.34 mg/kg, respectively, with determination coefficients (R2 )
phenylboronic acid (PBA) were performed. The samples were an- of 0.997 for the MCPDEs and 0.993 for GE.
alyzed by gas chromatography coupled to a MSD 5975C mass The BPO used in the experiments presented 4.5% of FFA,
spectrometer (Agilent Technologies, model 7890A, New Castle, 88.2% of TAG, and 7.3% of DAG. The content of MAG was lower
DE, USA) using an HP1-MS (30 m × 0.25 mm, 1 µm, Agilent than the LOQ (<0.3%) and the red (R) and yellow (Y) colors were
Technologies) capillary column. The following ions were moni- 20 and 70, respectively. The levels of the contaminants were <0.01
tored: m/z 147, 196, and 198 for 3-MCPD derivative, m/z 150, mg/kg for 3-MCPDE and <0.02 mg/kg for 2-MCPDE and GE.
201, and 203 for the internal standard 3-MCPD-d5 derivative,
m/z 196 and 198 for 2-MCPD derivative, m/z 147 and 240 for 3-
3.1 Formation of 3-MCPDE, 2-MCPDE, and GE
MBPD derivative, and m/z 150 and 245 for the internal standard
The results for 3-MCPDE analyses are displayed in Table 1.
3-MBPD-d5 derivative.
Mean concentrations ranged from 1.91 to 2.70 mg/kg. In the
lowest applied temperature and time (210 °C, 30 min), a concen-
2.6 Composition of acylglycerols tration of 2.15 mg/kg was observed. Nonsignificant differences
The analyses were adapted from the official AOCS method were noted in all other times (60, 90, and 120 min), demonstrat-
Cd 22–91 (AOCS, 2009). High-performance size exclusion chro- ing that the major proportion of 3-MCPDE was already formed
matography with a refractive index detector (RID) was employed under the mildest evaluated condition. At the other temperatures,
to analyze TAG, DAG, and MAG using a column Jordi Gel DVB the levels of 3-MCPDE varied statistically considering different
100A (300 × 7.8 mm) and a Jordi Gel DVB 500A (300 × times. Despite the similarity between the values, a statistical dif-
10 mm) in series. BPO and refined palm oil samples were dis- ference between 90 and 120 min was found at a thermal condition
solved in THF, which was used as mobile phase at a flow rate of of 230 °C. At 250 and 270 °C, the main differences were observed
Toxicology & Chemical

0.5 mL/min. Quantification was expressed in percentage. between 30 and 60 min.

Food Safety

2256 Journal of Food Science r Vol. 85, Iss. 7, 2020

MCPD and glycidyl esters in palm oil . . .

Table 1–Concentrations of 3-MCPDE (mg/kg) under the applied conditions of time and temperature.

Temperature (°C)
Time (min) 210 230 250 270
30 2.15 ± 0.05a,B 2.02 ± 0.06a,b,A,B 1.91 ± 0.01a,A 2.29 ± 0.09a,C
60 2.31 ± 0.23a,A 2.03 ± 0.07a,b,B 2.31 ± 0.12b,B 2.57 ± 0.21b,B
90 2.07 ± 0.14a,A 2.09 ± 0.07b,A 2.15 ± 0.04b,c,A 2.70 ± 0.06b,B
120 1.97 ± 0.24a,A,B 1.93 ± 0.08a,A 2.35 ± 0.12c,B 2.67 ± 0.04b,C
a-c
Within the same column, mean values with different lowercase letters differ statistically (P < 0.05).
A-C
Within the same line, mean values with different capital letters differ statistically (P < 0.05). N = 6 (deodorizations were performed in duplicate and analyses in triplicate).

Data previously published in the literature report the low in-

fluence of the temperature on 3-MCPDE formation when the
deodorization process is carried out between 180 °C and 265 °C
during 1 hr (Hrncirik & van Duijn, 2011). Nevertheless, signifi-
cant variations were noted for each time while taking into account
the evaluated temperatures of the present study, and the highest
values were verified when the oil was deodorized at 270 °C, re-
gardless of the time applied.
The concentrations of 2-MCPDE are reported in Table 2. The
average levels ranged from 0.68 to 1.44 mg/kg. In general, esters
of 2-MCPD were half of the contents of 3-MCPDE, which was
similarly observed by (Seefelder, Scholz, & Schilter, 2011). Our
results show that 2-MCPD esters can also be verified at the mildest
tested condition (210 °C, 30 min). Although concentrations of 3-
MCPDE are always higher than those of 2-MCPDE, both com- Figure 1–FFA (% of palmitic acid) of palm oil before and after deodoriza-
pounds have similar formation profiles. Nevertheless, there is a tion. A-C Within the same temperature, mean values with different lower-
slight rise in the content of this contaminant with an increase in case letters differ statistically (P < 0.05). A-D Within the same time, mean
temperature, which was not observed for 3-MCPDE. To the best values with different capital letters differ statistically (P < 0.05). N = 6
of our knowledge, no study in the literature has emphasized the (deodorizations were performed in duplicate and analyses in triplicate).
ND, nondeodorized.
formation kinetics of 2-MCPDE so far.
Table 3 shows that the mean concentrations of GE varied from
The results have indicated that using less drastic conditions
0.12 to 8.51 mg/kg under the tested conditions. The maximum
to deodorize palm oil is not a suitable strategy to mitigate the
value observed was approximately eightfold higher than the limit
chlorinated contaminants and alternative methods should be used
established by the European Commission for vegetable oils used as
for their effective reduction. Unlike 3-MCPDE and 2-MCPDE,
ingredients in food or direct consumption (EU, 2018). At 230 and
the levels of GE can be minimized by changing deodorization
250 °C, no statistical difference was observed between 30, 90, and,
times and temperatures. At 210 and 230 °C, GE was formed
120 min of the process. This suggests that at lower temperatures,
at low concentrations, which were below the limit established
time has little influence on the formation of the contaminant. An
by the European regulation for vegetable oil (EU, 2018). How-
increase in the concentration of GE over time was observed only
ever, other quality parameters, such as % of FFA and color,
when a temperature of 270 °C was applied.
may not be satisfactory under these conditions and should be
On the other hand, the levels of GE always increased with the
investigated.
temperature, except at 210 and 230 °C in 30 min. This result
is consistent with the literature, which describes that tempera-
tures up to about 220 °C do not favor the formation of this 3.2 Physicochemical changes of palm oil during the
contaminant in a significant way. Temperatures above 240 °C are deodorization
necessary to raise the concentrations of these esters (Craft et al., Figure 1 shows the results obtained for FFA of palm oil be-
2013; Destaillats, Craft, Dubois, & Nagy, 2012; Šmidrkal et al., fore and after deodorization, which varied from 0.26% to 4.45%.
2011). This confirmed that the formation of GE depends directly Except for the treatment performed at 270 °C during 60 min,
on the temperature applied in the deodorization step (Cheng, Liu, all values were higher than 0.3%, the maximum limit suggested
& Liu, 2016; Destaillats, Craft, Dubois, et al., 2012; Hrncirik & by the Codex Alimentarius Commission (CAC) (CAC, 2015).
van Duijn, 2011). The temperature is a very important variable for the acidity decay,
Taking into account that GE is also considered a precursor of as observed for all tested times. The increase of the temperature
3-MCPDE (Rahn & Yaylayan, 2011), we can suggest that an ex- caused a significant reduction of the FFA content, especially in
cess of GE in the system at 250 and 270 °C may contribute to the the first 30 min of the deodorization process. After this time,
formation of 3-MCPDE. Under these extreme temperature con- no significant decrease was observed except for the treatment at
ditions, the highest values of 3-MCPDE were produced, especially 210 °C.
with extended times such as 90 and 120 min. The results support The impact of deodorization on the reduction of acidity is
the hypothesis that there is a difference in formation behavior supported by literature in this area. The efficient volatilization of
Toxicology & Chemical

between MCPDE and GE. FFA depends not only on the vapor pressure, but also on their
Food Safety

Vol. 85, Iss. 7, 2020 r Journal of Food Science 2257

 MCPD and glycidyl esters in palm oil . . .

Table 2–Concentrations of 2-MCPDE (mg/kg) under the applied conditions of time and temperature.

Temperature (°C)
Time (min) 210 230 250 270
30 0.74 ± 0.15a,b,A 0.86 ± 0.05a,A 1.03 ± 0.04a,B 1.22 ± 0.11a,C
60 0.85 ± 0.09a,b,A 1.04 ± 0.06b,B 1.24 ± 0.07b,C 1.41 ± 0.16b,D
90 0.68 ± 0.17a,A 1.07 ± 0.10b,B 1.09 ± 0.03a,B 1.44 ± 0.03b,C
120 0.93 ± 0.17b,A 1.00 ± 0.07b,A 1.24 ± 0.05b,B 1.43 ± 0.05b,C
a-b
Within the same column, mean values with different lowercase letters differ statistically (P < 0.05).
A-D
Within the same line, mean values with different capital letters differ statistically (P < 0.05). N = 6 (deodorizations were performed in duplicate and analyses in triplicate).

Table 3–Concentrations of GE (mg/kg) under the applied conditions of time and temperature.

Temperature (°C)
Time (min) 210 230 250 270
30 0.34 ± 0.06c,A 0.35 ± 0.04a,A 0.95 ± 0.07a,B 4.38 ± 0.09a,C
60 0.24 ± 0.04b,A 0.42 ± 0.04b,B 1.60 ± 0.35b,c,C 6.36 ± 1.98a,b,c,D
90 0.14 ± 0.03a,A 0.38 ± 0.14a,b,B 1.20 ± 0.22a,b,C 7.66 ± 0.15c,D
120 0.20 ± 0.05a,b,A 0.35 ± 0.03a,B 1.45 ± 0.43a,c,C 8.51 ± 0.24b,D
a-c
Within the same column, mean values with different lowercase letters differ statistically (P < 0.05).
A-D
Within the same line, mean values with different capital letters differ statistically (P < 0.05). N = 6 (deodorizations were performed in duplicate and analyses in triplicate).

concentration, the processing pressure and type of flow system

in the deodorization step (De Greyt, Kellens, & Huyghebaert,
1999). At the same conditions, other volatile compounds, such
as undesirable odors, are volatilized (Ceriani & Meirelles, 2004;
Petrauskaitè, Greyt, & Kellens, 2000; Sampaio, Ceriani, Silva,
Taham, & Meirelles, 2011). At the high temperatures used to
reduce FFA, a decrease of the process efficiency and stability of
the final product can also be observed (Ceriani & Meirelles, 2004;
De Greyt et al., 1999; Slew & Mohammad, 1989).
A correlation between the final % of FFA and the concentrations
of 3-MCPDE, 2-MCPDE, and GE was not observed. However,
other authors have reported a correlation between the formation
of MCPDE and the initial acidity of the oil (Ramli, Siew, Ibrahim,
Kuntom, & Razak, 2015; Šmidrkal et al., 2011).
Figure 2 presents changes in TAG and DAG concentrations
during the deodorization. These parameters were evaluated in
order to identify possible correlations with the formation of the
contaminants. TAG varied from 89.58% to 92.86% while DAG
ranged between 6.29% and 7.10%.
No statistical difference in the levels of TAG was observed over
time at 250 and 270 °C. For DAG, no statistical difference was
verified, except at 210 °C. An inversely proportional variation
of those compounds was verified at a temperature higher than
230 °C. This can be explained by the loss of FFA and acylglyc-
erols, especially DAG, by entrainment (Ceriani & Meirelles, 2004;
Petrauskaitè et al., 2000). The rise of the temperature from 210 to
270 °C contributed to the increase of TAG and the reduction of
DAG until 230 °C. Higher temperatures, such as 250 and 270°C,
could contribute to the increase of DAG by the hydrolysis of TAG. Figure 2–Concentration of TAG (%) and DAG (%) under the applied con-
The increase of DAG at high temperatures may have favored ditions of time and temperature. A-B Within the same temperature, mean
the formation of GE, as also observed in several studies (Cheng, values with different lowercase letters differ statistically (P < 0.05). A-C
Liu, Wang, & Liu, 2017; Cheng et al., 2016; Destaillats, Craft, Within the same time, mean values with different capital letters differ sta-
Sandoz, et al., 2012; Freudenstein, Weking, & Matthäus, 2013). tistically (P < 0.05). N = 4 (deodorizations were performed in duplicate
and analyses in duplicate). ND, nondeodorized.
Using the Spearman correlation test, it was found that, at the
lower temperatures (210 and 230 °C), the concentrations of DAG
and GE had a moderate correlation (ρ = 0.566). However, as the tionship was found between the concentration of acylglycerols and
temperature rises, the correlation decreases until it is lost under the the formation of 3-MCPDE and 2-MCPDE, according to other
condition of 270 °C (ρ = −0.167). On the other hand, no rela-
Toxicology & Chemical

authors (Hrncirik & van Duijn, 2011). It is worth mentioning

Food Safety

2258 Journal of Food Science r Vol. 85, Iss. 7, 2020

MCPD and glycidyl esters in palm oil . . .

4. CONCLUSIONS
3-MCPD and 2-MCPD esters were formed at the mildest con-
ditions evaluated in the present study, that is, 210 °C during
30 min. This is sufficient to exhaust the main route for the 3-
MCPDE formation and reach concentrations around 2.0 mg/kg.
2-MCPD esters were formed under all tested conditions and have
a similar formation profile compared to 3-MCPD esters. On av-
erage, the levels of 2-MCPDE were half the concentration of
3-MCPDE, except at 210 °C in which an average of 25% of the
concentration of 3-MCPDE was observed. Considering the de-
odorization parameters of temperature and time currently used
in the industry, it is not possible to mitigate the formation of
MCPD esters. Differently, GE levels were directly proportional to
the temperature used in the deodorization step and the highest
concentrations were observed at conditions above 250 °C. The
appropriate control of the deodorization step can produce low
levels of GE. The results also suggested that GE in excess could
act as intermediary compounds in the formation of 3-MCPDE
at higher temperatures. No correlation was verified between the
final FFA content and the formation of the contaminants. On the
other hand, DAG showed moderate and positive Spearman’s cor-
relation with GE at milder temperatures, losing this correlation at
higher temperatures. The color parameters presented correlation
only with GE.

ACKNOWLEDGMENTS
This study was supported by the São Paulo Research Founda-
tion (FAPESP) [Grant number 2016/23958-3 and 2014/21252-
0]; Coordenação de Aperfeiçoamento de Pessoal de Nı́vel Superior
Figure 3–The decay of the red (A) and yellow (B) colors during the deodor- (CAPES), Brazil [Financial code (001)]; and Conselho Nacional de
ization process of palm oil at various conditions. ND, nondeodorized.
Desenvolvimento Cientı́fico e Tecnológico (CNPq), Brazil [Grant
number 168441/2017-9].

that other studies suggest this correlation (Ermacora & Hrncirik, AUTHOR CONTRIBUTIONS
2014; Freudenstein et al., 2013). Adriana Arisseto, Eduardo Vicente, Klicia Sampaio, Maisa
Another relevant parameter for the deodorization of palm oil Capristo, Renan Tivanello, and Roseli Ferrari contributed to the
is color (De Greyt & Kellens, 2005). High temperatures and ex- conception and planning of the study. The development of the de-
tended time used in the deodorization step cause a reduction of odorization system was carried out by Roseli Ferrari and the exe-
color in palm oil, which is related to the degradation of carotenoids cution of the experiments by Renan Tivanello and Maisa Capristo.
(De Greyt & Kellens, 2005; De Greyt et al., 1999; Mba, Dumont, The contaminant analyzes were done by Renan Tivanello with
& Ngadi, 2015; Sampaio et al., 2013; Slew & Mohammad, 1989; the supervision of Eduardo Vicente. All authors participated in
Zulkurnain, Lai, Tan, Abdul Latip, & Tan, 2013). the writing of the manuscript, especially Renan Tivanello with
Figure 3 shows the decay of the parameters red (R) and yel- the supervision of Adriana Arisseto.
low (Y) over time at the different applied temperatures. There
was a constant drop in R only in the process performed at 210 CONFLICTS OF INTEREST
°C. At intermediate temperatures of 230 and 250 °C, there was The authors declare no potential conflict of interest.
little variation between values: all below 2.0. The lowest values
REFERENCES
were observed at 270 °C. Yellow (Y) presented a defined decay AOCS. (1998). Official Method Ca 5a-40 Approved 1998. Free fatty acids in crude and refined fats and
profile and varied according to the temperature. As can be seen, oils. Official methods and recommended practices of the AOCS. (6th ed.). Urbana, IL: AOCS.
the milder thermal condition (210 °C) did not cause a change AOCS. (2009). Official Method Cd 22–91 Approved 2009. Polymerized triglycerides by gel-permeation
HPLC. Official methods and recommended practices of the AOCS (6th ed.). Urbana, IL: AOCS.
in the parameter until 90 min. The rise of 20 °C in the process AOCS. (2013). Official Method Cd 29a-13 Approved 2013. Determination of bound
temperature caused a considerable reduction in the Y value with monochloropropanediol- (MCPD)–and bound 2,3-epoxy-1-propanol (glycidol-) by gas chro-
matography/mass spectrometry (GC/MS). Official Methods and Recommended Practices of the
60 min of deodorization. The increase of the temperature de- AOCS (6th ed.). Urbana, IL: AOCS.
creased the time needed for the reduction of yellow and red colors, AOCS. (2017). Official Method Cd 13e-92 Approved 2017. Lovibond method using colour
glasses calibrated in accordance with Lovibond tintometer colour scale. Official Methods and
but resulted in an increase of the GE concentrations. Recommended Practices of the AOCS, (7th ed.). Urbana, IL: AOCS.
Color reduction presented a negative Spearman correlation of Arisseto, A. P., Marcolino, P. F. C., & Vicente, E. (2014). Determination of 3-
monochloropropane-1,2-diol fatty acid esters in Brazilian vegetable oils and fats by an
ρ = −0.827 and ρ = −0.932 with GE for red and yellow, re- in-house validated method. Food Additives and Contaminants - Part A, 31(8), 1385–1392.
spectively. This correlation is intrinsically linked to the process https://doi.org/10.1080/19440049.2014.926400
Ceriani, R., & Meirelles, A. J. A. (2004). Simulation of batch physical refining and deodorization
conditions. The fast reduction of the color using a high tem- processes. Journal of the American Oil Chemists’ Society, 81(3), 305–312.
perature would lead to a larger formation of GE and vice versa. Cheng, W., Liu, G., & Liu, X. (2016). Formation of glycidyl fatty acid esters both
in real edible oils during laboratory-scale refining and in chemical model during
No correlation was observed between color and 3-MCPDE and high temperature exposure. Journal of Agricultural and Food Chemistry, (64), 5919–5927.
Toxicology & Chemical

2-MCPDE. https://doi.org/10.1021/acs.jafc.6b01520
Food Safety

Vol. 85, Iss. 7, 2020 r Journal of Food Science 2259

 MCPD and glycidyl esters in palm oil . . .

Cheng, W., Liu, G., Wang, L., & Liu, Z. (2017). Glycidyl fatty acid esters in refined edible oils: Kamikata, K., Vicente, E., Arisseto, A. P., Miguel, A. M. R. de O., Milani, R. F., & Tfouni, S.
a review on formation, occurrence, analysis, and elimination methods. Comprehensive Reviews A. V. (2019). Occurrence of 3-MCPD, 2-MCPD and glycidyl esters in extra virgin olive oils,
in Food Science and Food Safety, 16(2), 263–281. https://doi.org/10.1111/1541-4337.12251 olive oils and oil blends and correlation with identity and quality parameters. Food Control,
Cho, W., Han, B. S., Nam, K. T., Park, K., Choi, M., Kim, S. H., . . . Jang, D. D. (2008). 95, 135–141. https://doi.org/10.1016/j.foodcont.2018.07.051
Carcinogenicity study of 3-monochloropropane-1, 2-diol in Sprague–Dawley rats. Food and Mba, O. I., Dumont, M. J., & Ngadi, M. (2015). Palm oil: Processing, character-
Chemical Toxicology, 46(9), 3172–3177. https://doi.org/10.1016/j.fct.2008.07.003 ization and utilization in the food industry—A review. Food Bioscience, 10, 26–41.
Codex Alimentarius Commission (CAC). (2015). Standard for named vegetable Oils. CONDEX https://doi.org/10.1016/j.fbio.2015.01.003
STAN, 210–199. Petrauskaitè, V., Greyt, W. F. De, & Kellens, M. J. (2000). Physical refining of coconut oil: effect
Craft, B., Chiodini, A., Garst, J., & Granvogl, M. (2013). Fatty acid esters of monochloro- of crude oil quality and deodorization conditions on neutral oil loss. Journal of the American
propanediol (MCPD) and glycidol in refined edible oils. Food Additives & Contaminants: Part Oil Chemists’ Society, 77(6), 581–586.
A, 30(1), 46–51. https://doi.org/10.1080/19440049.2012.709196 Rahn, A. K. K., & Yaylayan, V. A. (2011). What do we know about the molecular mechanism
Craft, B. D., & Destaillats, F. (2014). Formation mechanisms. Processing contaminants in edible oils. of 3-MCPD ester formation? European Journal of Lipid Science and Technology, 113(3), 323–329.
Urbana, IL: AOCS Press. https://doi.org/10.1016/B978-0-9888565-0-9.50006-3 https://doi.org/10.1002/ejlt.201000310
De Greyt, & Kellens, K. (2005). Deodorization. In Bailey’s industrial oil and fat products (pp. Ramli, M. R., Siew, W. L., Ibrahim, N. A., Kuntom, A., & Razak, R. A. A. (2015). Other factors
341–338). to consider in the formation of chloropropandiol fatty esters in oil processes. Food Additives &
De Greyt, Kellens, M. J., & Huyghebaert, A. D. (1999). Effect of physical refining on selected Contaminants: Part A, 32(6), 817–824. https://doi.org/10.1080/19440049.2015.1032368
minor components in vegetable oils∗. European Journal of Lipid Science and Technology, 101(11), Sampaio, K. A., Ayala, J. V., Silva, S. M., Ceriani, R., Verhe, R., & Meirelles, A. J. A. (2013).
428–432. Thermal degradation kinetics of carotenoids in palm oil. Journal of the American Oil Chemists’
Destaillats, F., Craft, B. D., Dubois, M., & Nagy, K. (2012). Glycidyl esters in refined palm Society, 90(2), 191–198. https://doi.org/10.1007/s11746-012-2156-1
(Elaeis guineensis) oil and related fractions. Part I: Formation mechanism. Food Chemistry, 131, Sampaio, K. A., Ceriani, R., Silva, S. M., Taham, T., & Meirelles, A. J. A. (2011).
1391–1398. https://doi.org/10.1016/j.foodchem.2011.10.006 Steam deacidification of palm oil. Food and Bioproducts Processing, 89(4), 383–390.
Destaillats, F., Craft, B. D., Sandoz, L., & Nagy, K. (2012). Formation mechanisms of https://doi.org/10.1016/j.fbp.2010.11.012
monochloropropanediol (MCPD) fatty acids diesters in refined palm (Elaeis guineensis) oil Seefelder, W., Scholz, G., & Schilter, B. (2011). Structural diversity of dietary fatty esters of
and related fractions. Food Additives & Contaminants, 29(1), 29–37. chloropropanols and related substances. European Journal of Lipid Science and Technology, 113,
Ermacora, A., & Hrncirik, K. (2014). Influence of oil composition on the formation 319–322. https://doi.org/10.1002/ejlt.201000319
of fatty acid esters of 2-chloropropane-1,3-diol (2-MCPD) and 3-chloropropane-1,3- Slew, W. L., & Mohammad, Y. (1989). Effects of refining on chemical and physical properties
diol (3-MCPD) under conditions simulating oil refining. Food Chemistry, 161, 383–389. of palm oil products. Journal of Oleo Science, 66(7), 1116–1119.
https://doi.org/10.1016/j.foodchem.2014.03.130 Šmidrkal, J., Ilko, V., Filip, V., Doležal, M., Zelinková, Z., Kyselka, J., . . . Velisek, J. (2011).
European Food Safety Authority. (2018). Update of the risk assessment on 3- Formation of acylglycerol chloro derivatives in vegetable oils and mitigation strategy. Czech
monochloropropanediol and its fatty acid esters. EFSA Journal. 16(1): 5083. Journal of Food Sciences, 29(4), 449–456.
European Food Safety Authority. (2016). Risks for human health related to the presence of Šmidrkal, J., Tesařová, M., Hrádková, I., Berčı́ková, M., Adamčı́ková, A., & Filip,
3-monochloropropanediol (MCPD), and their fatty acid esters, and glycidyl fatty acid esters V. (2016). Mechanism of formation of 3-chloropropan-1,2-diol (3-MCPD) esters
in food. EFSA Journal, 14(5), 4426. under conditions of the vegetable oil refining. Food Chemistry, 211, 124–129.
Food and Agriculture Organization of the United Nations/World Health Organization https://doi.org/10.1016/j.foodchem.2016.05.039
(FAO/WHO). (2017). Evaluation of certain contaminants in food. World Health Organization Svejkovská, B., Novotný, O., Divinová, V., Reblová, Z., Dolezal, M., & Velisek, J. (2004).
Technical Report Series, 1002, 1–166. Esters of 3-chloropropane-1, 2-diol in foodstuffs. Czech Journal of Food Sciences, 22(5),
Franke, K., Strijowski, U., Fleck, G., & Pudel, F. (2009). Influence of chemical refining process 190–196.
and oil type on bound 3-chloro-1, 2-propanediol contents in palm oil and rapeseed oil. LWT– Zelinková, Z., Svejkovská, B., Velı́šek, J., & Doležal, M. (2006). Fatty acid esters of 3-
Food Science and Technology, 42(10), 1751–1754. https://doi.org/10.1016/j.lwt.2009.05.021 chloropropane-1, 2-diol in edible oils. Food Additives & Contaminants, 23(12), 1290–1298.
Freudenstein, A., Weking, J., & Matthäus, B. (2013). Influence of precursors on the https://doi.org/10.1080/02652030600887628
formation of 3-MCPD and glycidyl esters in a model oil under simulated deodor- Zulkurnain, M., Lai, O. M., Latip, R. A., Nehdi, I. A., Ling, T. C., & Tan, C. P.
ization conditions. European Journal of Lipid Science and Technology, 115(3), 286–294. (2012). The effects of physical refining on the formation of 3-monochloropropane-1, 2-
https://doi.org/10.1002/ejlt.201200226 diol esters in relation to palm oil minor components. Food Chemistry, 135(2), 799–805.
Hrncirik, K., & van Duijn, G. (2011). An initial study on the formation of 3-MCPD es- https://doi.org/10.1016/j.foodchem.2012.04.144
ters during oil refining. European Journal of Lipid Science and Technology, 113(3), 374–379. Zulkurnain, M., Lai, O. M., Tan, S. C., Abdul Latip, R., & Tan, C. P. (2013). Optimiza-
https://doi.org/10.1002/ejlt.201000317 tion of palm oil physical refining process for reduction of 3-monochloropropane-1,2-diol
International Agency for Research on Cancer (IARC). (2000). Glycidol. In IARC monographs (3-MCPD) ester formation. Journal of Agricultural and Food Chemistry, 61(13), 3341–3349.
on the evaluation of carcnogenic risks to humans. (Vol. 77, pp. 469–489). France: IARC. https://doi.org/10.1021/jf4009185
International Agency for Research on Cancer (IARC). (2012). Some chemicals present in
industrial and consumer products, food and drinking-water. IARC Monographs on the Evaluation
of Carcinogenic Risks to Humans, 101, 349–374.
Toxicology & Chemical
Food Safety

2260 Journal of Food Science r Vol. 85, Iss. 7, 2020