397

Journal of Oleo ScienceCopyright ©2018 by Japan Oil Chemists’ Societydoi : 10.5650/jos.ess17210J. Oleo Sci. 67, (4) 397-406 (2018)

New Insights on Degumming and Bleaching Process Parameters on The Formation of 3-Monochloropropane-1,2-Diol Esters and Glycidyl Esters in Refined, Bleached, Deodorized Palm OilBiow Ing Sim1, Halimah Muhamad2, Oi Ming Lai3, Faridah Abas4, Chee Beng Yeoh2, Imededdine Arbi Nehdi5, Yih Phing Khor1 and Chin Ping Tan1＊

1 Department of Food Technology, Universiti Putra Malaysia, Faculty of Food Science and Technology, 43300 UPM Serdang, Selangor Darul Ehsan, MALAYSIA

2 Analytical and Quality Development Unit, Malaysian Palm Oil Board, No. 6, Persiaran Institusi, Bandar Baru Bangi, 43000, Selangor Darul Ehsan, MALAYSIA

3 Department of Bioprocess Technology, Universiti Putra Malaysia, Faculty of Biotechnology and Biomolecular Sciences, 43300 UPM Serdang, Selangor Darul Ehsan, MALAYSIA

4 Department of Food Science, Universiti Putra Malaysia, Faculty of Food Science and Technology, Malaysia, 43300 UPM Serdang, Selangor Darul Ehsan, MALAYSIA

5 Department of Chemistry, King Saud University, College of Science, Riyadh, SAUDI ARABIA

1 INTRODUCTIONFatty acid esters of 3-MCPD and glycidol are now classi-

fied as a new class of food contaminants after they have been widely found in the food chain. Refined palm oil was found to contain a remarkably high average level of 3-MCPD esters（3-MCPDE）and glycidyl ester（GE）in com-parison with other refined vegetable oils such as rapeseed, soy, and sunflower oils1－3）. The occurrence of these pro-cessed contaminants is of great concern because they are

＊Correspondence to: Chin Ping Tan, Department of Food Technology, Universiti Putra Malaysia, Faculty of Food Science and Technology, 43300 UPM Serdang, Selangor Darul Ehsan, MALAYSIAE-mail: tancp@upm.edu.myAccepted November 22, 2017 (received for review September 26, 2017)Journal of Oleo Science ISSN 1345-8957 print / ISSN 1347-3352 onlinehttp://www.jstage.jst.go.jp/browse/jos/　　http://mc.manusriptcentral.com/jjocs

believed to be hydrolyzed completely by lipases in the body into their free forms, namely 3-MCPD and glycidol, which are carcinogenic and can jeopardize human health. Both of these compounds have different toxicological properties. 3-MCPD has been classified as a non-genotoxic threshold carcinogen in which the toxicity primarily targets kidneys and male fertility, whereas glycidol is known to be a geno-toxic, non-threshold carcinogen in animal studies. It was shown to be mutagenic in a range of in vitro and in vivo

Abstract: This paper examines the interactions of degumming and bleaching processes as well as their influences on the formation of 3-monochloropropane-1,2-diol esters (3-MCPDE) and glycidyl esters in refined, bleached and deodorized palm oil by using D-optimal design. Water degumming effectively reduced the 3-MCPDE content up to 50%. Acid activated bleaching earth had a greater effect on 3-MCPDE reduction compared to natural bleaching earth and acid activated bleaching earth with neutral pH, indicating that performance and adsorption capacities of bleaching earth are the predominant factors in the removal of esters, rather than its acidity profile. The combination of high dosage phosphoric acid during degumming with the use of acid activated bleaching earth eliminated almost all glycidyl esters during refining. Besides, the effects of crude palm oil quality was assessed and it was found that the quality of crude palm oil determines the level of formation of 3-MCPDE and glycidyl esters in palm oil during the high temperature deodorization step of physical refining process. Poor quality crude palm oil has strong impact towards 3-MCPDE and glycidyl esters formation due to the intrinsic components present within. The findings are useful to palm oil refining industry in choosing raw materials as an input during the refining process.

Key words: 3-MCPD esters, glycidyl esters, crude palm oil, D-optimal design, water degumming, acid activated bleaching earth

B. I. Sim, H. Muhamad, O. M. Lai et al.

J. Oleo Sci. 67, (4) 397-406 (2018)

398

genotoxicity tests, and it was able to induce a broad spec-trum of genotoxic effects, including gene mutations, chro-mosomal aberrations, sister chromatid exchange and un-scheduled DNA synthesis4）.

At present, due to the increasing of world population, there is a growing demand for edible oil. The refining process is inevitably becoming the most essential step adopted by oil processors for producing edible oil with im-proved taste, appearance and keepability for fulfilling the surging demand. Physical refining is extensively used in the palm oil industry. It involves degumming, bleaching and finally steam distillation, in which free fatty acids and other volatile components are distilled off from the oil using an effective stripping agent（usually steam）under reduced pressure and elevated temperature5）. Unfortunately, 3-MCPDE and GE are mainly formed during this high-tem-perature deodorization step, although some formation might also occur at other stages of refining. According to Pudel et al.6）, degumming and bleaching treatments reduce the formation of 3-MCPDE and GE during deodorization. Conversely, others findings show that acidity from phos-phoric acid during the degumming process and acid-acti-vated bleaching earth play a significant role concerning the generation of 3-MCPDE and GE7, 8）. It is believed that the acid sites that carried over from the phosphoric acid and acid-activated bleaching earth will render the protonation of acylglycerol which will then favor the formation of 3-MCPDE and GE under the presence of chloride. On the other hand, Schurz9）and Shimizu et al.10）reported the use of high-performance acid-activated bleaching earth on the reduction of 3-MCPDE and GE, respectively. This led to a different conclusion regarding the effect of acidity of bleaching earth on the formation of these process contami-nants.

Refined palm oil is the major dietary source of both 3-MCPDE and GE, as it is widely used as a common ingre-dient or medium in various food products11）, especially infant formulas12, 13）. Thus, it is crucial to have more knowl-edge about the effects of precursors and processing on the formation of the esters. With this critical issue in mind, D-optimal design was opted to evaluate the formation of 3-MCPDE and GE in the palm oil physical refining process. This method enables a more comprehensive study on the effects and interactions of different degumming and bleaching treatments on their formation with a reduced number of experimental sets. Moreover, crude palm oil（CPO）quality parameters were examined to investigate the relationship of minor components such as free fatty acids（FFA）, diacylglycerols（DAG）and monoacylglycerols（MAG）as the putative precursors for 3-MCPDE and GE formation.

2 EXPERIMENTAL2.1 Materials

The standards, 1,2-Dipalmitoyl-3-chloropropanediol（PP-3-MCPD）, glycidyl palmitate（Gly-P）, pentadeuterated 1,2-dipalmitoyl-3-chloropropanediol（PP-3-MCPD-d5）and pentadeuterated glycidyl palmitate（Gly-P-d5）were pur-chased from Toronto Research Chemicals Inc.,（North York, ON, Canada）. Phenylboronic acid of purity ≥ 97.0％ was purchased from Fluka（Shanghai, China）. All other chemi-cals and reagents were of analytical grade. CPO was ob-tained from Sime Darby Golden Jomalina Sdn. Bhd.（Teluk Panglima Garang, Selangor）. Acid-activated bleaching earth（acidic pH and neutral pH）and natural bleaching earth were provided by PT. Clariant Adsorbents Indonesia（Bogor, Indonesia）.

2.2 Laboratory-scale re�ningApproximately 100 g of homogenized CPO were de-

gummed and bleached in a 250 mL three-neck flask with a stirrer and heating mantle equipped with a temperature probe under a nitrogen blanket. The CPO was heated to 90℃ before 0.50 mL of 20％ w/w phosphoric acid was added for degumming for 10 min. Subsequently, bleaching was performed at 105℃ for 20 min with a 1％ dosage of bleaching earth to obtain bleached palm oil（BPO）. Without any delay, the mixed hot oil containing bleaching earth was filtered into a pre-heated 500 mL Buchner flask through the appropriately sized Buchner funnel using Whatman No.1 filter paper. An aliquot of 90 g of the BPO sample was collected for the deodorization process at 260℃ for 20 min to obtain refined bleached and deodorized（RBD）palm oil in a 250 mL three-neck flask that was connected with a vacuum pump（Woosung Vacuum Co. Ltd., Incheon, Korea）. After deodorization, the RBD palm oil was then cooled to a temperature of 60℃ under vacuum conditions before it is kept frozen at －18℃ for analysis.

2.3 Design of experimentsThe effects of the phosphoric acid percentage（X1）and

different types of bleaching earth（X2）on the 3-MCPDE and GE levels of RBD palm oil was evaluated and modelled by using D-optimal design. Both degumming and bleaching processes on Standard Quality I CPO were studied concur-rently using three phosphoric acid dosages（0 – 2.5％ w/w）and three different bleaching earths at levels of 1％ w/w, namely natural bleaching earth（NBE）, acid-activated bleaching earth（acidic pH）（AAA）and acid-activated bleaching earth（neutral pH）（AAN）.

D-optimal design was made up of 16 experimental runs with three factorial and four center points（Refer to Table 3）. Design Expert version 6.0.6（Stat-Ease, Inc., Minneapo-lis, MN USA）were used to analyze data for all experimental runs with the level of significance set at a 95％ confidence level. F-test, ANOVA, and lack of fit test were used to eval-

Formation of 3-MCPDE and GE during Physical Refining

J. Oleo Sci. 67, (4) 397-406 (2018)

399

uate the adequateness of the models and the overall pre-dictive capability of the model was elucidated by the coef-ficient of determination R2.

The experimental results in this study were analyzed using the MINITAB statistical software（Version 14, Minitab Inc., PA, USA）. All data were expressed as the means±standard deviations of duplicate analyses. A one-way analy-sis of variance（ANOVA）at the 5％ significance level was used to determine significant differences（p＜0.05）between the means.

2.4 Determination of 3-MCPDE and GE by using GC-MSThe determination of 3-MCPDE and GE in CPO and RBD

palm oil was carried out according to AOCS Official Method Cd 29a-1314）. This is an indirect method that is based on the conversion of GE into 3-monobromopropanediol（3-MBPD）monoesters in an acidic solution containing bromide salt. The solution then undergoes acidic-catalyzed transesterification, extraction and discarding of the fatty acid methyl esters（FAME）followed by derivatization of free diols（MCPD and MBPD）with phenylboronic acid（PBA）prior to GC-MS analysis. A quantitative analysis of the phenylboronic derivative of 3-MCPD and 3-MBPD com-pounds on an Agilent Technologies 7890A gas chromato-graph coupled with a quadrupole Agilent 5875C mass se-lective detector and data processing system（Santa Clara, California, USA）was carried out by monitoring quantifier ions at m/z 147 and m/z 150 of 3-MCPD and its isotopically labelled internal standard（3-MCPD-d5）, respectively. Ions at m/z 196, 198（3-MCPD）and m/z 201（3-MCPD-d5）were used as qualifier ions. For GE, quantifier ions at m/z 147 and qualifier ions at m/z 240 were monitored for phenylbo-ronic derivatives of 3-MBPD; whereas for phenylboronic derivative of 3-MBPD-d5, the quantifier ion and qualifier ions were at 150 and 245, respectively. Both 3-MCPD and 3-MBPD contents were determined from a matrix-based calibration graph that was constructed by plotting the ratio of the amount of standard to the amount of internal stan-dard on the x-axis against the ratio of the corresponding peak areas on the y-axis. Ions at m/z 147（3-MCPD or 3-MBPD）and 150（3-MCPD-d5 or 3-MBPD-d5）were used for quantification. The method was validated based on several parameters: limit of detection（LOD）, limit of quan-titation（LOQ）, precision and recovery. Both the LOD and LOQ were determined by using the signal-to-noise ratio method. The precision of the method was evaluated by measuring three concentration levels in the low（0.5 mg/kg – PP-3-MCPD; 1.2 mg/kg – Gly-P）, medium（2.0 mg/kg – PP-3-MCPD; 4.0 mg/kg – Gly-P）and high（5.0 mg/kg – PP-3-MCPD; 7.1 mg/kg – Gly-P）regions, prepared in three repli-cations, of the analytical range of the study. The recovery, precision（％ RSD）and repeatability were determined from this data set. The recovery was calculated as follows:

Percent Recovery＝Conc.of spiked sample－Conc.of unspiked sample

Theoritical Conc.of spiked sample×100

The linearity of response was evaluated between 0–5.2 mg/kg for PP-3-MCPD and 0–11.9 mg/kg for Gly-P, pre-pared and injected in triplicate.

2.5 Acylglycerol composition analysis using HPLCAnalysis of triacylglycerides（TAG）, diacylglycerides

（DAG）, monoacylglycerides（MAG）of CPO were performed using Alliance e2695 Separation Modules High performance liquid chromatography（HPLC）（Waters, USA）coupled with an evaporative light scattering（ELS）2424 detector（Waters, USA）. The samples were prepared by adding 5％ w/v oil in acetone. The samples were filtered through 0.47 μm micron nylon filters fitted through a syringe. An aliquot of 10 μL of sample was injected into a Purospher® Star RP-18e column（5 μm, 250 mm×4 mm, Merck, Germany）with a temperature of 35℃. The flow gradient of the mobile phase comprised a mixture of acetone（solvent A）and acetonitrile（solvent B）with an initial composition of 90％ solvent B at 0 min, 85％ B at 8 min, 10％ B at 40 min, 90％ B at 50 min, and finally 90％ B at 52 min with a flow rate of 1.0 mL/ min. The drift tube temperature, nebulizer power and gas pressure of the ELS detector were set at 45℃, 60％（heating）and 35 psi, respectively.

2.6 Oil quality analysisThe FFA levels, carotene content and deterioration of

bleachability index（DOBI）for CPO were conducted ac-cording to MPOB test methods15）. All analyses were per-formed in triplicate, and the mean values and standard de-viations are reported.

2.7 pH measurement of bleaching earthThe pH values of the bleaching earths were determined

using a method as described in literature16）. A portable pH meter/mV meter（Sartorius AG, Goettinger, Germany）was used to measure the suspension of 2％（by weight）bleach-ing earth in deionized water that has undergone 30 min constant stirring by magnetic stirrer.

2.8 Physical Characterization of bleaching earthThe surface morphologies of the bleaching earths were

analyzed using a field emission scanning electron micro-scope（FESEM）（FEI Quanta SEM Model 400 F）equipped with an energy dispersive X-ray（EDX）feature. The nitro-gen isotherms were measured with the Thermo Finnigan Sorptomatic 1990 series nitrogen adsorption/desorption analyzer which was carried out at the temperature of 77 K. The pore volume and area were calculated using the Bar-ett-Joyner-Halenda（BJH）method17）.

B. I. Sim, H. Muhamad, O. M. Lai et al.

J. Oleo Sci. 67, (4) 397-406 (2018)

400

3 RESULTS AND DISCUSSION3.1 Impact of CPO quality on the formation of 3-MCPDE

and GE CPO is usually sorted into different grades according to

their FFA contents. In this study, four different grades of CPOs, namely Premium Quality（PQ）, Superior Quality（SQ）, Standard Quality I（SQ I）and Standard Quality II（SQ II）were subjected to a laboratory-scale physical refining process, as described in section 2.2. The quality character-istics of CPO, including their 3-MCPDE and GE contents, were examined and compared, as shown in Table 1. For monitoring purposes, AOCS Official Method Cd 29a-13 was validated in-house for quantitation of 3-MCPDE and GE in palm oil The positive validation results in terms of sensitiv-ity, reliability, recovery and precision proved that the method is reliable and suitable for the routine monitoring of 3-MCPDE and GE in palm oil（refer to Table 2）.

3-MCPDE and GE were not detected in CPO from all qualities, except that trace amounts of 3-MCPDE were found in the worst grade of CPO, namely SQ II CPO. On the

other hand, all RBD palm oil was found to contain both 3-MCPDE and GE. This clearly indicates that high deodor-ization temperature is one of the main culprits for the un-desirable formation of 3-MCPDE and GE. RBD palm oil from SQ II CPO contains the highest level of 3-MCPDE and GE, followed by SQ I（third-grade CPO）, and SQ（second-grade CPO）, whereas PQ（first-grade CPO）has a signifi-cantly low level of 3-MCPDE and GE（p＜0.05）. Similar trends were observed for FFA contents. FFA is commonly recognized as the main indicator of CPO quality, where the higher the FFA contents, the poorer the quality. DOBI is another good quality indicator for the oxidative status of CPO, as well as its ease of refining. A low DOBI value in CPO is undesirable because it might lead to poor oil bleachability and keepability. This result shows that poor quality CPO has a high tendency towards 3-MCPDE and GE formation due to the intrinsic components present within18, 19）.

FFA, DAG and MAG are widely known as the by-prod-ucts of enzymatic hydrolytic degradation of TAG by lipases,

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

Quality characteristic Premium Quality (PQ)

Superior Quality (SQ)

Standard Quality I

Standard Quality II

3-MCPD esters of CPO [mg/kg]b <LOD <LOD <LOD <LOQGlycidyl esters of CPO [mg/kg]c <LOD <LOD <LOD <LOD3-MCPD esters of RBD palm oil [mg/kg] 0.89±0.02C 1.09±0.03B 1.14±0.02B 1.75±0.07A

Glycidyl esters of RBD palm oil [mg/kg] 0.39±0.03C 0.43±0.02C 0.55±0.03B 0.72±0.02A

Free Fatty Acid (FFA) [%] 0.96±0.03D 1.21±0.04C 2.67±0.01B 4.42±0.03A

DOBId 3.67±0.05A 3.49±0.08B 3.24±0.05C 2.88±0.04D

β-carotene content [ppm] 586±0.12C 545±0.24D 649±0.30A 611±0.22B

Diacylglyceride (DAG) [%] 1.00±0.16B 1.54±0.18AB 1.72±0.11A 2.09±0.15A

Monoglyceride (MAG) [%] 0.0C 0.0C 0.06±0.18B 0.10±0.14A

a Standard deviations between batches are given in parentheses; results represent the means of two replicate trials. For each column, means with the same letter do not differ significantly at p < 0.05.

b LOD for 3-MCPDE = 0.05 mg/kg LOQ for 3-MCPDE = 0.10 mg/kgc LOD for GE = 0.15 mg/kgd DOBI, Deterioration of bleachability index

Table 2　 Performance characterist ics of 3-MCPD and glycidol determination in palm oil samples using AOCS Cd 29a-23 method.

Performance Characteristic 3-MCPD equiv. Gly equiv.Limit of Detection (LOD) (mg/kg) 0.05 0.10Limit of Quantitation (LOQ) (mg/kg) 0.15 0.20Linear Range (μg) 0.03 – 0.52 0.06 – 1.19Linearity (R2) 0.9993 0.9996Recovery (%) 99.1 – 101.8 93.8 – 101.9Precision (%RSD) 0.42 – 4.80 0.55 – 1.03

Formation of 3-MCPDE and GE during Physical Refining

J. Oleo Sci. 67, (4) 397-406 (2018)

401

and the process can be accelerated by harvesting, trans-portation and storage under elevated temperatures and harsh conditions20）. Freudenstein et al. reported a positive correlation between the amount of DAG and MAG in CPO and the 3-MCPDE and GE content in deodorized palm oil, in which DAG appears to be a stronger precursor than MAG29）. Another finding also shows that amounts of DAG exceeding 4％ in crude oil will result in higher amounts of esters after heat processing, and this phenomenon is ex-ceptionally apparent for GE21－23）. Although the direct cor-relation between DAG and GE levels is hardly concluded by comparing between RBD palm oil from different CPOs in this study, we observed that the higher DAG content in CPO makes oils more susceptible to 3-MCPDE and GE for-mation upon deodorization. MAG may be a potent precur-sor for the formation of esters, however, due to its very low level in palm oil（～0.3％）24）, and it is known to be signifi-cantly removed during oil degumming21）and volatilized during high temperature deodorization, MAG will not be considered as a critical factor involved in the formation 3-MCPDE and GE. The occurrence of 3-MCPDE and GE on the RBD palm oil derived from MAG-absence CPO（i.e., SQ and PQ CPO）reiterates the hypothesis.

3.2 Effects of degumming and bleaching on the reduction of 3-MCPDE and GE

Palm oil is commonly refined by physical refining due to its high efficiency, high yield and relatively small environ-ment impact. Physical refining is a continuous processing comprising a two-step operation of pre-treatment, namely degumming and bleaching processes, followed by deodor-ization. The deodorization step is well known as the main factor contributing to the undesirable formation of 3-MCPDE and GE. However, it is worth noting that degum-ming and bleaching processes may positively or negatively influence their formation under different conditions25）.

Table 3 shows the responses of D-optimal design, namely 3-MCPD（Y1）, glycidol（Y2）and carotene content（Y3）of RBD palm oil. The 3-MCPD response was significantly fitted to reduced quadratic polynomial models（p＜0.05）, and the other two responses were fitted to cubic responses. The ANOVA results of the final reduced models, the p-value of significant terms, the lack of fit test, the F-test and the de-termination coefficient R of the responses are shown in Table 4.3.2.1 Effects of degumming on the reduction of 3-MCPDE

and GEThe degumming process generally results in a reduction

Table 3　Matrix of the D-optimal design with three response variables.

No. of runIndependent Variables Response variables

Percent acid (X1, %)c

Types of bleaching earth (X2)

a3-MCPD (mg/kg)

Glycidol (mg/kg)

Carotene content (mg/kg)

1 0 NBE 0.73±0.02 0.66±0.01 511.4±0.812 0 AAA 0.59±0.01 0.60±0.01 366.8±0.343 0 AAN 0.71±0.02 0.56±0.02 431.1±0.274 0.63 NBE 1.61±0.03 0.52±0.02 455.3±0.205 0.63 AAA 1.16±0.03 0.37±0.01 348.5±0.346 0.63 AAN 1.06±0.03 0.37±0.02 402.4±0.207 1.25 NBE 1.42±0.01 0.59±0.02 481.5±0.148 1.25 AAN 1.26±0.02 0.41±0.02 417.4±0.079 1.25 NBE 1.38±0.01 0.75±0.00 464.0±0.2010 1.25 AAA 1.02±0.01 0.64±0.01 361.6±0.1411 2.50 AAN 0.99±0.01 0.62±0.01 395.6±0.0712 2.50 AAA 0.88±0.02 <LOQb 395.4±0.0013 2.50 NBE 1.35±0.05 0.68±0.01 479.7±0.2014 2.50 AAA 0.86±0.02 <LOQ 381.9±0.4115 2.50 AAN 1.10±0.01 0.65±0.01 397.5±0.0016 2.50 NBE 1.36±0.07 0.67±0.02 486.2±0.07

a For factor X2, NBE: Natural bleaching earth; AAN: Acid activated bleaching earth (Neutral pH); AAA: Acid activated bleaching earth (Acidic pH).

b LOQ for GE = 0.20 mg/kgc 0% acid represents water degumming at levels of 1% (w/w) water

B. I. Sim, H. Muhamad, O. M. Lai et al.

J. Oleo Sci. 67, (4) 397-406 (2018)

402

of 3-MCPDE and GE during further processing26）. Water degumming significantly reduced the 3-MCPDE from 1.14 mg/kg to 0.59 mg/kg, which is a percent reduction of nearly 50％, as depicted in Fig. 1（a）. Water is believed to remove the precursors that may involve the formation of 3-MCPDE, especially the chlorinated compounds that are polar in nature. Polar chloride substances were demonstrated to be the main source of chloride ions for the formation of MCPD esters in palm oil, although it still remains unclear whether it was due to their higher abundance in the oil or their greater reactivity27）. The chlorinated compounds may de-compose under high deodorization temperatures into hy-drogen chloride（gas）and react with acylglycerols to form 3-MCPDE28）. Conversely, this phenomenon is not obvious for GE, as shown in Fig. 1（b）. This finding corroborates with the other authors, who reported that the 3-MCPDE level is significantly influenced by chlorine content but not on the generation of GE29, 30）. The GE content is exception-ally low when 2.50％ of phosphoric acid was added during the dry degumming process together with the auxiliary effect of acid activated bleaching earth during the bleach-ing process. This can be explained by the instability of GE that tends to degrade under acidic solutions due to its re-active epoxide binding structure2, 31）.3.2.2 Effects of different types of bleaching earth on the

reduction of 3-MCPDE and GEThe bleaching process lessens the formation of

3-MCPDE and GE during deodorization6）. Bleaching earth is the most commonly-used adsorbent in the bleaching process due to its low cost but relatively high absorption performance for color pigments. In the studies, three types of bleaching earth were used. As shown in Table 5, NBE has the highest pH value, and AAA has the most acidic pH. According to Ramli, Siew, Ibrahim, Hussein, Kuntom, Abd. Razak and Nesaretnam8）, there are positive correlations between the acidity of bleaching earth and the levels of 3-MCPDE in oils, where the higher the acidity is, the greater the formation of 3-MCPDE, regardless of the type

of degumming（acid or water degumming）. Thus, the author recommended the use of natural bleaching clays or acid activated clays with a more neutral pH as a possible avenue for processing CPO with lower 3-MCPDE. In addi-tion, HED, Johansson and Mellerup32）also reported that acid activated bleaching earth may favor the formation of 3-MCPD compounds, whereas with the use of neutral bleaching earth may effectively reduce the amount of GE and at the same time keep the 3-MCPDE at low concentra-tions. Contrary to the expectation, our findings showed that AAA exerts a significant effect on 3-MCPDE reduction compared with NBE and AAN（p＜0.05）, as depicted in Fig. 1（a）. AAA gives a much higher adsorption capacity and better performance than NBE after acid activation treatment. As shown in Table 4, AAA possesses the highest average pore volume and pore size. The FESEM images, as shown in Figs. 2（a）-（c）, also illustrate that AAA and AAN have rough surfaces compared with NBE, which has uniform smooth surfaces, supporting the fact that acid treatment enhances the structural and surface properties of bleaching earth33）. Bleaching earth with greater pore volumes yields the best adsorption capacities34－36）. Thus, it is believed that the particular precursors that are responsi-ble for 3-MCPDE formation are being adsorbed from BPO prior to deodorization9）. Moreover, the carotene content in BPO constitutes additional supporting evidence that types of bleaching earths elicited larger effects than the degum-ming process, wherewith AAA was observed to remove most of the carotenes while NBE retained most of the caro-tenes, as shown in Fig. 1（c）. Carotenoid is postulated as one of the oil-soluble chlorinating agents that will induce the formation of 3-MCPDE by bringing chloride ions close to the lipids, resulting in chlorination through reaction with acyloxonium ions37）. Therefore, it is important to note that the performance of the bleaching earth is the key determi-nant for 3-MCPDE removal rather than its acidity. The effect of the acidity of the bleaching earth on 3-MCPDE formation is compensated by its high adsorption capacity.

Table 4　 The ANOVA, p-value, lack of fit test, F-test and the determination coefficient, R2, of three response variables of the final reduced model.

3-MCPD esters(mg/kg, Y1)

Glycidyl esters(mg/kg, Y2)

Carotene content(ppm, Y3)

Model Quadratic Cubic Cubicp-value < 0.0001a 0.0022a < 0.0001a

F-value 20.84 13.99 53.11R2 0.9124 0.9545 0.9789Adjusted R2 0.8687 0.8863 0.9605Lack of fit (p-value) 0.1064 0.1259 0.2606Lack of fit (F-value) 11.63 3.64 1.99

a Significant (p < 0.05).

Formation of 3-MCPDE and GE during Physical Refining

J. Oleo Sci. 67, (4) 397-406 (2018)

403

Fig. 1　 Interaction plots showing the effects of degumming and bleaching treatments on （a）3-MCPD levels in refined bleached deodorized（RBD）palm oil; （b）Glycidol levels refined bleached deodorized（RBD）palm oil; and（c）carotene content in bleached palm oil（BPO）.

Table 5　Properties of three types of bleaching earthsa.

Types of bleaching earth pH Total Pore volume (cm3/g) Total Pore area (m2/g)Acid Activated (Acidic) 3.18±0.01 0.380138 340.363Acid Activated (Neutral) 7.27±0.06 0.296388 177.953Natural 8.61±0.01 0.286729 120.564

a The total pore volume and area reported refers to pores with a diameter of 13.0 to 210.0 Å.

B. I. Sim, H. Muhamad, O. M. Lai et al.

J. Oleo Sci. 67, (4) 397-406 (2018)

404

On the other hand, GE was almost completely removed when acid-activated（acidic pH）bleaching earth was used after the degumming process with 2.50％ phosphoric acid. Shimizu et al.38）also reported the thorough removal of GE by treating the oil with acid activated bleaching earth. Instead of through adsorption, the elimination of GE was mainly through the modification of GE into their respective glycerol compounds and trace amounts of free glycerol and fatty acids. The possible mechanisms were proposed to

occur by epoxide ring opening of GE to form glycerol monoester by a reaction with water under acidic conditions and subsequently by interesterification reactions between the glycerol monoesters to generate simple acylglycerols and free glycerol.

Fig. 2　 FESEM images of（a）acid-activated bleaching earth（acidic pH）; （b）acid-activated bleaching earth（neutral pH）; and（c）natural bleaching earth.

Formation of 3-MCPDE and GE during Physical Refining

J. Oleo Sci. 67, (4) 397-406 (2018)

405

ConclusionThe CPO quality influences the tendency of 3-MCPDE

and GE formation in RBD palm oil. This signifies the crucial role of the composition of the CPO on the potential of ester formation during the refining process. By applying good manufacturing practices and by keeping the time period between harvesting and palm fruit milling as short as pos-sible, the CPO should be able to arrest the rise of partial acylglycerols and free fatty acids in the CPO and keep them to a minimum. Water degumming is able to signifi-cantly reduce 3-MCPDE by removing the potent precursors that might be involved in the formation, most notably polar chlorinated compounds. GE was almost completely removed when acid-activated bleaching earth was used after the high dosage of phosphoric acid（2.50％）in the de-gumming process. The elimination of GE was believed to take place via structural modification of GE into their re-spective glycerol compounds under acidic conditions, due to its reactive epoxide binding structure. Moreover, acid-activated bleaching earth had a greater effect on 3-MCPDE reduction compared with natural bleaching earth, indicat-ing that the performance of the bleaching earth is the key determinant for 3-MCPDE and GE removal rather than its acidity level.

AcknowledgementThe work was supported by the Putra Grant, Universiti

Putra Malaysia（project number GP-IBT/2013/9419500）. The authors extend their appreciation to the International Scientific Partnership Program ISPP at King Saud Univer-sity for funding this research work through ISPP# 0055.

References1） Weißhaar, R.; Perz, R. Fatty acid esters of glycidol in

refined fats and oils. Eur. J. Lipid Sci. Technol. 112, 158-165（2010）.

2） Matthäus, B.; Pudel, F.; Fehling, P.; Vosmann, K.; Freudenstein, A. Strategies for the reduction of 3-MCPD esters and related compounds in vegetable oils. Eur. J. Lipid Sci. Technol. 113, 380-386（2011）.

3） Arisseto, A.P.; Marcolino, P.F.C.; Vicente, E. Determi-nation of 3-monochloropropane-1,2-diol fatty acid es-ters in Brazilian vegetable oils and fats by an in-house validated method. Food Addit. Contam. 31, 1385-1392（2014）.

4） Bakhiya, N.; Abraham, K.; Gürtler, R.; Appel, K.E.; Lampen, A. Toxicological assessment of 3-chloropro-pane-1,2-diol and glycidol fatty acid esters in food. Mol. Nutr. Food Res. 55, 509-521（2011）.

5） Gibon, V.; De Greyt, W.; Kellens, M. Palm oil refining.

Eur. J. Lipid Sci. Technol. 109, 315-335（2007）.6） Pudel, F.; Benecke, P.; Fehling, P.; Freudenstein, A.;

Matthäus, B.; Schwaf, A. On the necessity of edible oil refining and possible sources of 3-MCPD and glycidyl esters. Eur. J. Lipid Sci. Technol. 113, 368-373（2011）.

7） Zulkurnain, M.; Lai, O.M.; Tan, S.C.; Abdul Latip, R.; Tan, C.P. Optimization of palm oil physical refining process for reduction of 3-monochloropropane-1,2-di-ol（3-MCPD）ester formation. J. Agric. Food Chem. 61, 3341-3349（2013）.

8） Ramli, M.; Siew, W.L.; Ibrahim, N.; Hussein, R.; Kun-tom, A.; Abd Razak, R.; Nesaretnam, K. Effects of de-gumming and bleaching on 3-MCPD esters formation during physical refining. J. Am. Oil Chem. Soc. 88, 1839-1844（2011）.

9） Schurz, K.D. Pat., EP2361298 B1（2010）.10） Shimizu, M.; Moriwaki, J.; Shiiba; D.; Nohara, H.; Kudo,

N.; Katsuragi; Y. Elimination of glycidyl palmitate in diolein by treatment with activated bleaching earth. J. Oleo Sci. 61, 23-28（2012）.

11） Crews, C.; Chiodini, A.; Granvogl, M.; Hamlet, C.; Hrnčiřík, K.; Kuhlmann, J.; Lampen, A.; Scholz, G.; Weisshaar, R.; Wenzl, T.; Jasti, P.R.; Seefelder, W. Ana-lytical approaches for MCPD esters and glycidyl esters in food and biological samples: A review and future perspectives. Food Addit. Contam. 30, 11-35（2012）.

12） Koo, W.W.K.; Hockman, E.M.; Dow, M. Palm olein in the fat blend of infant formulas: Effect on the intesti-nal absorption of calcium and fat, and bone mineraliza-tion. J. Am. Coll. Nutr. 25, 117-122（2006）.

13） Nelson, S.E.; Rogers, R.R.; Frantz, J.A.; Ziegler, E.E. Palm olein in infant formula: Absorption of fat and minerals by normal infants. Am. J. Clin. Nutr. 64, 291-296（1996）.

14） Àmerican Oil Chemists’s Society（AOCS）. AOCS Offi-cial Method Cd 29a-13, Determination of 2- and 3-MCPD fatty acid esters and glycidol fatty acid esters in edible oils and fats by acid transesterifications. Offi-cial Methods and Recommended Practices of the AOCS, 6th ed.（2013）.

15） Malaysia Palm Oil Board. MPOB Test Methods. PORIM Test Methods Vol 1 Palm Oil Research Institute of Malaysia, Malaysia（2005）.

16） Ng, S. Characteristics and efficiency of bleaching earth in the physical refining of palm oil. Master degree the-sis. University Putra Malaysia, Serdang, Malaysia（2006）.

17） Barrett, E.P.; Joyner, L.G.; Halenda, P.P. The determi-nation of pore volume and area distributions in porous substances. I. Computations from nitrogen isotherms. J. Am. Chem. Soc. 73, 373-380（1951）.

18） Zulkurnain, M.; Lai, O.M.; Latip, R.A.; Nehdi, I.A.; Ling, T.C.; Tan, C.P. The effects of physical refining on the

B. I. Sim, H. Muhamad, O. M. Lai et al.

J. Oleo Sci. 67, (4) 397-406 (2018)

406

formation of 3-monochloropropane-1,2-diol esters in relation to palm oil minor components. Food Chem. 135, 799-805（2012）.

19） Craft, B.D.; Nagy, K.; Seefelder, W.; Dubois, M.; Des-taillats, F. Glycidyl esters in refined palm（Elaeis guineensis）oil and related fractions. Part II: Practical recommendations for effective mitigation. Food Chem. 132, 73-79（2012）.

20） Siew, W.L.; Ng, W.L. Diglyceride content and composi-tion as indicators of palm oil quality. J. Sci. Food Ag-ric. 69, 73-79（1995）.

21） Craft, B.D.; Chiodini, A.; Garst, J.; Granvogl, M. Fatty acid esters of monochloropropanediol（MCPD）and gly-cidol in refined edible oils. Food Addit.Contam. 30, 46-51（2012）.

22） Masukawa, Y.; Shiro, H.; Kondo, N.; Kudo, N. General-ized method to quantify glycidol fatty acid esters in edible oils. J. Am. Oil Chem. Soc. 88, 15-21（2011）.

23） Masukawa, Y.; Shiro, H.; Nakamura, S.; Kondo, N.; Jin, N.; Suzuki, N.; Ooi, N.; Kudo, N.; A new analytical method for the quantification of glycidol fatty acid es-ters in edible oils. J. Oleo Sci. 59, 81-88（2010）.

24） Goh, E.M.; Timms, R.E. Determination of mono- and diglycerides in palm oil, olein and stearin. J. Am. Chem. Soc. 62, 730-734（1985）.

25） Pudel, F.; Benecke, P.; Fehling, P.; Freudenstein, A.; Matthäus, B.; Schwaf, A. On the necessity of edible oil refining and possible sources of 3-MCPD and glycidyl esters. Eur. J. Lipid. Sci. Technol. 113, 368-373（2011）.

26） Matthäus, B.; Pudel, F. Mitigation of 3-MCPD and glyc-idyl esters within the production chain of vegetable oils especially palm oil. Lipid Technol. 25, 151-155（2013）.

27） Ermacora, A.; Hrncirik, K. Influence of oil composition on the formation of fatty acid esters of 2-chloropro-pane-1,3-diol（2-MCPD）and 3-chloropropane-1,2-diol（3-MCPD）under conditions simulating oil refining. Food Chem. 161, 383-389（2014）.

28） Destaillats, F.; Craft, B.D.; Sandoz, L.; Nagy, K. Forma-

tion mechanisms of Monochloropropanediol（MCPD）fatty acid diesters in refined palm（Elaeis guineensis）oil and related fractions. Food Addit. Contam. 29, 29-37（2011）.

29） Freudenstein, A.; Weking, J.; Matthäus, B. Influence of precursors on the formation of 3-MCPD and glycidyl esters in a model oil under simulated deodorization conditions. Eur. J. Lipid Sci. Technol. 115, 286-294（2013）.

30） Shimizu, M.; Weitkamp, P.; Vosmann, K.; Matthäus, B. Influence of chloride and glycidyl-ester on the genera-tion of 3-MCPD- and glycidyl-esters. Eur. J. Lipid Sci. Technol. 115, 735-739（2013）.

31） Hogan, P.; Lee, J.I.; Matlock, M.; Solheim, L.; Wicklund, L.E. Pat., EP2506720 A1（2014）.

32） HED, K.; Johansson, M.; Mellerup, J. Pat., EP2874501 B1（2014）.

33） Zschau, W. Bleaching of edible fats and oils. Eur. J. Lipid Sci. Technol. 103, 505-551（2001）.

34） Christidis, G.; Scott, P.; Dunham, A. Acid activation and bleaching capacity of bentonites from the islands of Milos and Chios, Aegean, Greece. Appl. Clay Sci. 12, 329-347（1997）.

35） Lomić, G.A.; Kiš, E.E.; Dimić, E.B.; Romanić, R.S. In-vestigation of activated al-pillared clay efficiency in vegetable oil purification. Acta Period. Technol. 35, 31-36（2004）.

36） Lin, S.W.; Yoo, C.K. Adsorption isotherms for removal of iron, copper, phosphorus and oxidation products from crude palm oil using natural and acid-activated clays. J. Oil Palm Res. 19, 356（2007）.

37） ILSI, Workshop on ME in Food Products. Brussels, Belgium. http://ilsi.org/event/workshop-on-mcpd-and-glycidyl-esters-in-food-products/. Accessed 12 April 2017.

38） Shimizu, M.; Vosmann, K.; Matthäus, B. Generation of 3-monochloro-1,2-propanediol and related materials from tri-, di-, and monoolein at deodorization tempera-ture. Eur. J. Lipid Sci. Technol. 114, 1268-1273（2012）.