1654 Eur. J. Lipid Sci. Technol.

2014, 116, 1654–1663

Research Article
Molecular distillation and characterization of diacylglycerol‐
enriched palm olein

Chiou Moi Yeoh1,3, Eng Tong Phuah2, Teck Kim Tang2, Wai Lin Siew3, Luqman Chuah Abdullah4
and Thomas Shean Yaw Choong4,5

1
Department of Chemical Engineering, Faculty of Engineering and Science, Universiti Tunku Abdul Rahman,
Kuala Lumpur, Malaysia
2
Institute of Bioscience, Universiti Putra Malaysia, Selangor, Malaysia
3
Malaysian Palm Oil Board No. 6, Persiaran Institusi, Selangor, Malaysia
4
Department of Chemical and Environmental Engineering, Faculty of Engineering, Universiti Putra Malaysia,
Selangor, Malaysia
5
INTROP, Universiti Putra Malaysia, Selangor, Malaysia

Lipase‐catalyzed glycerolysis of palm olein was used to produce a mixture of acylglycerols with

34‐wt% of DAG. The reaction conditions were 5‐wt% of Lipozyme TLIM at 55°C and 8 h of
reaction time. For commercial purposes, it is required to purify the product up to 80‐wt% DAG and
with free fatty acids (FFA) content below 0.1‐wt%. A single‐step distillation process was not sufficient
to meet this product requirement. Two distinct 2‐step short path distillation approaches were then
studied. First scheme involved the removal of TAG by initial distillation step at 250°C, followed by
separation of the MAG and FFA from distillate obtained at 180°C during second distillation step at
vacuum pressure of 0.1 Pa. Second scheme involved the removal of MAG and FFA in first step at
180°C prior to purification of DAG from residue at 250°C during second distillation step at vacuum
achieved up to 0.1 Pa. The results suggested that the first scheme of 2‐step distillation operation was
able to achieve 89.9‐wt% of DAG purity without exceeding the limit of 0.1‐wt% of FFA. A final yield of
21.5‐wt% and DAG recovery of 47.8% were obtained using the first scheme. A detailed DAG profile
was identified and product characterizations such as fatty acid composition, slip melting point, and
solid fat content profile were also investigated. It was observed that purified‐DAG product showed
lower iodine value and higher slip melting point than raw material palm olein. The final product
had 1134  10 ppm tocols content.

Practical applications: This paper has two main practical applications: (i) Enables production of
highly purified DAG‐based palm olein via appropriate processing method and processing conditions.
(ii) Provide knowledge and understanding of the physicochemical properties of DAG‐enriched palm
olein fraction, which is a crucial aspect in food applications.

Keywords: DAG / Glycerolysis / Palm olein / Short path distillation

Received: December 22, 2013 / Revised: June 12, 2014 / Accepted: June 17, 2014
DOI: 10.1002/ejlt.201300502

1 Introduction
Correspondence: Prof. Thomas Shean Yaw Choong, Department of
Chemical and Environmental Engineering, Universiti Putra Malaysia, In the last decade, the search for fat‐based substitute to
43400 Serdang, Selangor, Malaysia replace the conventional edible oils and fats is gaining
E‐mail: csthomas@upm.edu.my, tsyc2@yahoo.co.uk
momentum worldwide as overconsumption of TAG‐based
Fax: þ60 3 8656‐7150
fats and oils is claimed to be the main culprit for the alarming
Abbreviations: :FFA, free fatty acid; G, glycerol; IV, iodine value; MD, rise in obesity rate and weight‐related diseases. The recent
molecular distillation; SFC, solid fat content; SMP, slip melting point; SPD, disclosure of DAG‐enriched oils has lightened up the world
short path distillation especially the obese community. Intense research works

ß 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.ejlst.com

Eur. J. Lipid Sci. Technol. 2014, 116, 1654–1663 Distillation and characterization of diacylglycerol 1655

conducted to date revealed that the nutritional health benefit areas is well‐published, for instance, purification of structured
of DAG‐oil lies in its ability to reduce serum TAG lipids [12, 15], separation of MAG from acylglycerol
concentration and exhibit high postprandial lipid oxidation mixtures [5, 14], extraction of valuable FFA from vegetable
activity, resulting in a decrease of both body weight and oil deodorizer distillate [16], recovery of polyphenols and
visceral fat mass, probably attributable to its distinct tocopherols from deodorizer distillate [17], enrichment of
metabolism pathway in contrary to TAG despite both alkylglycerols [18], etc.
provide the same energy contribution without sacrificing Single or double‐step SPD is normally applied in the
the desirable appearance, flavor, aroma, texture and mouth purification of DAG component in which the former
feel of food [1–4]. Moreover, DAG is generally regarded as approach has turned up to be ineffective as compared to
safe (GRAS) and has been used in tandem with MAG as the latter. However, the 2‐step distillation process reported to
anionic emulsifiers in food, cosmetic as well as pharmaceuti- date are normally designed to remove FFA together with
cal industries [5]. In view of its extensive applications, DAG‐ MAG in the initial distillation step, followed by DAG
enriched oil has been introduced in United States and enrichment from residue in the second distillation step [5,
in Japan as functional cooking oil containing 80% DAG 11]. In present study, different 2‐step SPD processing scheme
under the brand names of “Enova Oil” and “Healthy Econa was proposed and investigated wherein DAG was initially
Cooking Oil,” respectively. Commercialization of DAG‐ distilled off at high temperature in the first distillation step
enriched oil has been successful with total annual sales of and the FFA and MAG are eliminated from distillate in the
approximately US $200 million in Japan since its introduction second distillation step at lower temperature. Comparisons
in the late 90’s till 2009 [6]. were performed to evaluate the efficiency of both conven-
Chemical glycerolysis, in the presence of chemical catalyst tional and proposed strategies. Moreover, the properties of
at elevated temperature, is the common industrial practice the DAG‐rich palm olein fractions remain scarce. The
employed to produce DAG and MAG [7]. This approach is objectives of this study were to develop an effective distillation
yet an energy intensive process, which upsurges both capital process for purification of DAG‐oil and to characterize the
investment and operational cost. Moreover, severe reaction properties of DAG‐rich oil, which provides an indication of
conditions often result in undesired by‐products formation, suitability for use in other foods.
which requires further purification. The recent rapid
development in enzyme technologies have successfully
provided an alternative pathway to produce DAG in milder 2 Materials and methods
operating conditions, improved regioselectivity and less
environmental impact. DAG can be produced via enzymatic 2.1 Materials
esterification of free fatty acid (FFA) with glycerol (G) [8],
enzymatic glycerolysis of glycerol with edible oil [5, 9] as well Palm Olein IV56 was purchased from Sime Darby, Malaysia;
as enzymatic partial hydrolysis of edible oil [10, 11] in which glycerol Anhydrous (99.9%) was from Mallinckrodt Baker,
lipase‐catalyzed glycerolysis approach outperforms other USA; and Lipozyme TLIM was given by Novozyme A/S,
methods due to its abundant and cheap glycerol supply Denmark. Acetone from Fisher Scientific UK Limited, UK
owing to emerging of biodiesel industry. and acetonitrile from Mallinckrodt Baker, USA were HPLC
The main components of the reaction products from grade. All other chemicals used were reagent grade. Stand-
lipase‐catalyzed glycerolysis process are mostly DAG, MAG, ards for HPLC (1,2‐dipalmitin, 1,3‐dipalmitin, 1,2‐diolein,
FFA, unreacted TAG, and negligible amount of G in which 1,3‐distearin, 1,2‐dimyristin, 1,3‐dilinolein, 2‐oleoylglyerol,
the content of each component varies, depending on the 1,2‐dipalmitoyl‐3‐oleoyl‐rac‐glycerol) were purchased from
operating parameters. In order to obtain concentrated Sigma Chemical, USA, while other standards (1,2‐diolein,
DAG‐oil containing 80‐wt% DAG and 0.1‐wt% FFA, 1,3‐diolein, glycerol‐1‐palmitate‐3‐oleate, glycerol‐1‐palmi-
it is crucial to develop an effective method for separation and tate‐2‐oleate) were purchased from Larodan, Sweden.
purification of DAG from the reaction mixtures. Short path Monopalmitin and monoolein were purchased from Tokyo
distillation (SPD) or commonly known molecular distillation Kasei, Japan.
(MD) is a promising thermal separation technique that can
be employed to intensify DAG content from acylglycerol 2.2 Glycerolysis reaction
mixtures without significant alteration in oil quality as
reported in previous literatures [5, 11–13]. SPD is character- Enzymatic glycerolysis reaction was conducted in solvent‐free
ized by reduced residence time of the distilled liquid to system for DAG production as described in earlier work with
the operating temperature, formation thin film layer by slight modification [9]. The reactions were performed by
strong centrifugal force which improves heat and mass reacting 1000 g palm olein with 170 g glycerol being absorbed
transfer, a sufficiently low pressure inside distillator, and onto 170 g silica gel 60G in the presence of 5‐wt% Lipozyme
relatively short distance between the evaporator and the TLIM in a 2000 mL jacketed batch reactor. The reaction
condenser [5, 14]. Application of this technology in lipid mixtures were heated at 55°C and agitated at 200 rpm. The

ß 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.ejlst.com

1656 C. M. Yeoh et al. Eur. J. Lipid Sci. Technol. 2014, 116, 1654–1663

reaction mixtures were then collected after 8 h and centri- curves were constructed, and the results are given as the
fuged to remove the enzyme particles to hinder further weight percentage of total glycerides [19, 20].
enzymatic reaction. In this study, water produced during the
reaction was reduced as much as possible to promote further 2.6 Fatty acid composition

synthesis. Molecular sieve (3 A, 20 g) was therefore added
into the reaction mixtures at the beginning of the reaction to The fatty acid composition was determined using Hewlett‐
remove the water produced. Packard 6890 Series GC. The analysis was done according to
ISO 5508:1990 Animal and vegetable fats and oils—analysis
2.3 Short path distillation by GC of the methyl esters of fatty acid. The oil sample was
first dissolved in n‐hexane before methylation using sodium
The distillation was performed using a laboratory short‐path methoxide. The solution was then diluted with distilled water
distillation VKL 70, from VTA GmbH, Germany. The major and allowed to settle for 5 min. The upper layer of fatty acid
part of the equipment was constructed from glass. The main methyl ester (FAME) was collected and decanted for GC
components were an integral condenser, a cooling trap for analysis.
filling with liquid nitrogen and a vacuum system. The vacuum The GC was equipped with a FID, electronic integrator
system included a diffusion pump and a rotary vane pump. and data processor. The flow rate of the helium (carrier gas)
The evaporator surface was 0.04 m2, with vacuum achieved of was positioned at 0.8 mL/min with a split ratio of 1:100.
up to 0.1 Pa. The heating of the evaporator was provided by A fused silica capillary column (DB‐23, 60 m  0.25 mm,
the jacket circulated with heated thermo oil. 0.25 mm film (J&W Scientific, Folsom, USA)) was used in the
The oil sample was preheated to 80°C. The weight of GC. The FID and injector temperature were set at 240°C.
round‐bottom flasks for residue and distillate were mea- The column was initially conditioned by setting at ambient
sured and recorded as initial weight. The oil sample was then temperature for an hour before programmed to 220°C at a
heated to the required temperature. The temperatures were rate of 3°C/min. This temperature was maintained for about
180, 230, 250°C for single‐step purification. For the first 16 h until a stable baseline was obtained. The temperature
approach of double‐step purification, the evaporator was then set to 180°C and remained isothermal. FAME
temperature was set at 250°C. The collected distillate was was quantified by comparing the retention time and the peak
then purified at 180°C. For the second approach of double‐ areas with a RM‐6 methyl ester standard.
step purification, the evaporator temperature was set at
180°C. The collected residue was then purified at 250°C. 2.7 Slip melting point
The weight of the residue and distillate with bottle after
experiment were recorded. The mass flow rate of the feed Slip melting point (SMP) is the temperature at which a
was 500 g/h. column of fat, of specified length, rises in an open capillary
tube under specific conditions of test. The SMP was
2.4 Free fatty acid determined according to the MPOB test method p.4.2
[21]. Capillary glass tubes which are thin walled, uniformly
The FFA content was measured by a titration method defined bored glass tubes open at both ends with an internal diameter
in AOCS Official Method Ca 5a‐40. The oil sample was of 1.1–1.3 mm, external diameter of 1.4–1.7 mm and length
dissolved in iso‐propanol with phenolphthalein as indicator. of 50–60 mm, were used. The capillary tube was dipped into
The mixtures were then titrated with sodium hydroxide liquid sample until column of fat of 10 mm high was obtained
solution until a permanent faint pink color appeared and in the tube. The fat column was chilled until the fat was
persisted for at least 1 min. The FFA content was calculated solidified. The tube was placed in test tube, held in a beaker of
as a percentage of palmitic acid. water and equilibrated at 10  1°C for 16 h.
After 16 h, the capillary tube was removed from the test
2.5 Acylglycerol composition analysis tube and attached to a thermometer with a rubber band so
that the lower ends of the tubes was leveled with the bottom of
Acylglycerol compositions were determined by Gilson the mercury bulb of the thermometer. The thermometer was
(France) HPLC equipped with a refractive index detector then suspended in a beaker containing 400 mL of boiled
model 2410 from Waters, USA, using two commercially distilled water so that the lower end of the thermometer was
packed LiChrospher1100 RP‐18 columns end‐capped immersed in the water to a depth of 30 mm. The starting
(250 mm) with 5 mm particle size in series. The acylglycerol temperature of the batch was adjusted to 8–10°C below the
compositions were eluted with acetone/acetonitrile (70:30) expected slip point of the sample. The water bath was agitated
for TAG determination and acetone/acetonitrile (50:50) for with magnetic stirrer and heat was supplied to increase the
FFA, MAG, and DAG determination, both at a flow rate of temperature at a rate of 1°C/min. The heating rate was
1 mL/min. The identification of the TAG and DAG was reduced to 0.5°C/min as the slip point was reached. The
according to Swe (1995) and Ghazali (1995). Calibration observation was recorded based on two measurements.

ß 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.ejlst.com

Eur. J. Lipid Sci. Technol. 2014, 116, 1654–1663 Distillation and characterization of diacylglycerol 1657

2.8 Solid fat content When oil was fed to the SPD, it was split into two streams.
Part of the feed oil (F) was collected in the distillate stream
The solid fat content (SFC) of oil was measured using pulsed (D) while the rest remained in the residue stream (R).
NMR spectrometry (Bruker Minispec PC120, Rheistetten‐ The split ratio D =R was an important parameter to evaluate
Forchheim, Germany). The MPOB p4.9 test method was the performance of SPD process. It was calculated as the
used [21]. SFC of the sample before and after reactions were following equation.
measured. The samples were melted at 70°C and mixed
thoroughly before filled into sample tube (180  10 mm) to a D mass of distillate
split ratio; ¼ ð1Þ
height of 3 cm. The samples were melted at 70°C for 30 min, R mass of residue
chilled at 0°C for 90 min prior to measurement. Melting,
chilling, and holding of samples were carried out in pre‐ Previous literatures revealed that the separation efficiency
equilibrated thermostated water bath. The values of % SFC in SPD was affected by several operating parameters such as
were recorded based on three measurements. feed flow rate, vacuum pressure and evaporator tempera-
ture [15]. In this study, low operating pressure of 0.1 Pa was
2.9 Total tocopherols and tocotrienols contents employed in order to reduce the boiling point of acylglycerols
analysis in the reaction mixtures, thereby enhancing the evaporation
rate of acylglycerols during separation process at reduced
Tocopherols and tocotrienols contents in the oil were analyzed temperature [15]. Feed flow rate is another critical operating
according to AOCS Official Method Ce 8‐89. The sample was variable in SPD, which requires proper monitoring. High feed
dissolved in n‐hexane before being injected to Gilson (France) flow rate reduces the processing duration at sacrifice of
HPLC. The HPLC was equipped with fluorescence detector acylglycerol purity whereas low flow rate promotes efficient
(Perkin Elmer, USA) using a 250  4.6 mm Si column separation process with increase in operating time. Prelimi-
commercially packed with 5 mm particle size (Phenomenex, nary experiments at 180°C indicated that mass flow rate of
USA). Wavelength used for excitation and emission was 290 500 g/h was adequate to provide effective separation (data not
and 330 nm, respectively. The samples were eluted with iso‐ shown). One of the objectives of this work was to investigate
propanol at a flow rate of 1 mL/min. The tocopherols and the effect of evaporator temperature on separation efficiency.
tocotrienols were identified according to Standard Tocopherols The recovery of a certain component was determined by
and Tocotrienols (Isomer Kits of ChromaDex, USA). using Eq. (2):

2.10 Statistical analysis componentN ðwt%Þ  feedN ðwt%Þ

% recovery ¼
component ðwt%Þ  feedðwt%Þ
One‐way ANOVA was applied to analyze the data statistically  100% ð2Þ
using Minitab 16.2 system software. Significant differences
between means were determined by using Tukey’s test at 5% Where the component is referred to each MAG, DAG, or
significance level. TAG and N is referred to either distillate or residue of the
process.

3 Results and discussion 3.1 Single‐step purification

Lipase‐catalyzed glycerolysis of palm olein (temperature ¼ 55°C, Single‐step purification was initially conducted to study the
agitation speed ¼ 200 rpm, enzyme load ¼ 5‐wt% Lipozyme effect of evaporator temperature on the separation of
TLIM, glycerol to silica gel ratio ¼ 1, and reaction time ¼ 8 h) acylglycerols. Three evaporator temperatures were consid-
resulted in a mixture of acylglycerols consisting of
57‐wt% ered (180, 230, and 250°C) and the results were presented in
TAG,
34‐wt% DAG,
6‐wt% MAG,
3‐wt% of FFA, Table 1.
and negligible amount of G. Due to cost consideration, only Results demonstrated that the split ratio increased from
5‐wt% Lipozyme TLIM was used in the present study. To 0.06 to 0.50 in tandem with increasing evaporator tempera-
achieve product specification of at least 80‐wt% DAG and ture from 180 to 250°C. Low DAG purity (1.82‐wt%) was
FFA of not more than 0.1‐wt%, SPD has been proven to be a obtained in the distillate at 180°C but increased drastically to
favorable technique for the separation and enrichment of 61.7‐wt% at elevated temperature of 250°C. Meanwhile the
DAG by means of high vacuum condition. Owing to the recovery of the DAG in the distillate was also found to
distinct glyceride chain length and vapor pressure of different increase from 1.82% to 61.70%. At low evaporator
acylglycerols, the easiness of removal was expected to follow temperature (180°C), most of the DAG molecules were
the order as shown below. retained in the residue stream due to their high molecular
weight and vaporization point. When higher operating
FFA  MAG < DAG < TAG temperature (230°C) was applied, considerable amount

ß 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.ejlst.com

1658 C. M. Yeoh et al. Eur. J. Lipid Sci. Technol. 2014, 116, 1654–1663

Table 1. Split ratio and DAG recovery for different temperatures at with previous studies [22, 23]. Xu and others (1998) revealed
vacuum pressure of 0.1 Pa by single‐step purification that high evaporator temperature during purification pro-
moted acyl migration in which 1,3‐DAG was converted to 1,2
180°C 230°C 250°C (2,3)‐DAG that was thermodynamically unstable intermedi-
Distillate yield (wt%)a 6.1 16.5 33.4 ate compound. Therefore, there is a possibility that two
Residue yield (wt%)a 93.9 83.5 66.6 molecules of migrated DAGs (1,2‐DAG) resynthesized into
Split ratio (D/R)a 0.06 0.20 0.50 both TAG and MAG via interesterification at relatively high
DAG purity in the 9.1  1.3a 45.7  1.2b 61.7  0.9c temperature. The hypothesis was further supported by the
distillate (wt%)b increase in both TAG and MAG contents after distillation
DAG purity in the 31.6  0.5c 22.1  0.3b 10.0  0.5a process (Table 2).
residue (wt%)b The acylglycerol compositions, yield and recovery of all
DAG recovery in the 1.82 24.48 61.70 components for different temperature studied were shown in
distillate (%)a
Table 2. Most of the FFA and MAG was distilled off while
DAG recovery in the 97.29 59.91 19.94
DAG and TAG remained in the residue at 180°C. Results
residue (%)a
Loss of DAG (%)a 0.89 15.61 18.36
indicated that DAG content increased slightly to 31.6‐wt% in
the residue at evaporator temperature of 180°C as compared
a
Results are means of two duplicates.
to initial 30.5‐wt% DAG in the feed. At 230°C, a bigger
b
Values in the same row with different letters are significantly portion of DAG was distilled off together with FFA and
different (p  0.05). MAG. A 45.7‐wt% of DAG was obtained in the distillate with
12.1‐wt% of FFA and 40.7‐wt% of MAG. At elevated
temperature of 250°C, most of the DAG was distilled off with
of DAG was entrained into the distillate stream due to the fact FFA and MAG while most of the TAG remained in the
that the boiling point of DAG compounds is reached, residues. A 61.7‐wt% of DAG was obtained in the distillate
suggesting that high temperature should be applied to distill with 7.2‐wt% of FFA and 28.1‐wt% of MAG. By virtue of
off high‐boiling components from acylglycerol mixtures. its low molecular weight, most FFA was concentrated in
Wang and coworkers (2010) also suggested that bubbling of the distillate at evaporator temperature of 230 and 250°C.
the acylglycerols with low boiling points might contribute to The single‐step purification process demonstrated that the
the entrainment of DAG into the distillate. However, loss of separation of partial acylglycerols increased in parallel with
DAG was reported to increase from 0.89% to 18.36% with evaporator temperature. While this process was not able to
the increase of the evaporator temperature. It was hypothe- achieve product specification of at least 80‐wt% of DAG with
sized that acyl migration during the purification of DAG by FFA content of less than 0.1‐wt%, nevertheless this gave deep
distillation led to this phenomenon, which was consistent insight into the distillation mechanism as temperature shifted.

Table 2. Acylglycerol compositions, yield, and recovery of all components for different temperatures vacuum pressure of 0.1 Pa by the
single‐step purification

Temperature 180°C 230°C 250°C

Components Feed Distillate Residue Feed Distillate Residue Feed Distillate Residue

Weight (g) 100.0 6.1 93.9 100.0 16.5 83.5 100.0 33.4 66.6
Acylglycerol TAG 63.2 0.0 67.2 62.4 1.5 77.8 58.2 3.0 90
compositions (wt%) DAG 30.5 9.1 31.6 30.8 45.7 22.1 33.4 61.7 10
MAG 4.0 60.2 0.2 4.6 40.7 0.1 5.9 28.1 0
FFA 2.3 29.7 0.0 2.2 12.1 0.0 2.5 7.2 0
Yield (%) TAG 63.2 0.0 63.1 62.4 0.2 65.0 58.2 1.0 59.9
DAG 30.5 0.6 29.7 30.8 7.5 18.5 33.4 20.6 6.7
MAG 4.0 3.7 0.2 4.6 6.7 0.1 5.9 9.4 0.0
FFA 2.3 1.8 0.0 2.2 2.0 0.0 2.5 2.4 0.0
Recovery (%) TAG 0.0 99.8 0.4 104.1 1.7 103.0
DAG 1.8 97.3 24.5 59.9 61.7 19.9
MAG 91.8 4.7 146.0 1.8 159.1 0.0
FFA 78.8 0.0 90.8 0.0 96.2 0.0
Total recovery of DAG (%) 99.1 84.4 81.6

Results are means of two duplicates.

ß 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.ejlst.com

Eur. J. Lipid Sci. Technol. 2014, 116, 1654–1663 Distillation and characterization of diacylglycerol 1659

3.2 Double‐step purification Distillate, D2 (17.8 g)

Since single‐step purification was not able to achieve the

desired product specification, double‐step purification was
therefore applied. Preliminary studies in single‐step distilla-
Distillate, D1 (39.3 g)
tion indicated that recovery of DAG was the lowest in the Stage 2
180o
distillate at evaporator temperature of 180°C and the highest
recovery of DAG in the distillate stream was obtained when
evaporator temperature was raised to 250°C. Consequently,
Feed, F (100 g) Product stream
evaporator temperature of 180°C could be applied to remove Stage 1
250oC
FFA while 250°C could be applied to separate DAG from Residue, R2 (21.5 g)
TAG.
Two schemes considered at the present study which were:

a) First scheme—removal of TAG before FFA Residue, R1 (60.7 g)

b) Second scheme—removal of TAG after FFA
(a)
The process overview for the first and second scheme was
shown in Fig. 1. In the first scheme, the reaction product was
Distillate, D1 (21.8 g)
first purified at 250°C where most of DAG was recovered at
the distillate. The distillate of the first step was then further
purified at 180°C to remove the FFA and MAG. On the other
hand, second scheme involved the initial purification of Feed, F (100 g) Distillate, D2 (28.7 g)
reaction product at 180°C to remove both FFA and MAG Stage 1
180oC Product stream
where most of the DAG was recovered in the residue. The
residue of the first step was further purified at 250°C to
remove TAG in order to enhance the DAG purity. The
overall process for both the schemes showed that the second Stage 2
scheme had a larger throughput compared to the first scheme. Residue, R1 (78.2 g)
250o

Table 3 shows the comparison of the split ratio,

acylglycerol compositions, yield and recovery of all compo-
nents for both first scheme and second scheme. For the first
scheme, the split ratio was 0.65 for the first step at 250°C and
Residue, R2 (49.5 g)
0.83 for the second step at 180°C. For the second scheme, the
split ratio was 0.28 for the first step at 180°C and 0.58 for
the second step at 250°C. For the first scheme, a fraction of (b)
39.3‐wt% distillate with 54.4‐wt% of DAG was obtained at
Figure 1. Double‐step DAG purification in short path distillation
the first step of purification at 250°C. This fraction was (SPD) on a feed of 100 g basis: (a) first scheme involves the removal
further purified at 180°C where a fraction 54.7‐wt% residue of TAG at 250°C before FFA at 180°C under vacuum pressure of
with 89.9‐wt% of DAG was obtained. For the second scheme, 0.1 Pa and (b) second scheme involves the removal of TAG at
a fraction of 78.2‐wt% residue with 50.3‐wt% of DAG was 250°C after FFA at 180°C under vacuum pressure of 0.1 Pa.
obtained from the purification of the first step at 180°C. This
fraction was further purified at 250°C where a fraction of migration. High evaporator temperature (250°C) in the first
36.7‐wt% distillate with 81.8‐wt% of DAG was obtained. step induced significant acyl migration and thereby gave rise
Both products achieved more than 80‐wt% DAG from the to the conversion of unstable DAG isomers to both undesired
two schemes. The yield obtained for the required product was by‐products of TAG and MAG as described earlier. The
higher in the second scheme (28‐wt%) as compared to the DAG loss was similar to the results obtained from single‐step
first scheme (21.5‐wt%). purification where high loss in DAG was observed under
During separation at 250°C for the first scheme, 52.92% extreme temperature. The total loss of DAG for the first
of the DAG was recovered at the distillate where 25.9% of scheme was 27.89%.
DAG was lost. After the second step purification at 180°C, a For the separation of the second scheme, 85.71% of DAG
90.41% of DAG was recovered in the residue where 3.76% was recovered in the residue for the purification at 180°C.
of DAG was lost. The DAG loss was high at the first step Approximate 11.48% of DAG was lost and only 2.81% of
as compared to the second step, indicating that strong DAG was recovered in the distillate. The residue from the
positive correlation between evaporator temperature and acyl first step was then purified at 250°C where 59.80% of DAG

ß 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.ejlst.com

1660 C. M. Yeoh et al. Eur. J. Lipid Sci. Technol. 2014, 116, 1654–1663

Table 3. Split ratio, acylglycerol compositions, yield, and recovery of all components for the first and second scheme by double‐step
purification at vacuum pressure of 0.1 Pa

First scheme Second scheme

1st step, 250°C 2nd step, 180°C 1st step, 180°C 2nd step, 250°C

Feed Distillate Residue Distillate Residue Feed Distillate Residue Distillate Residue
Component
Weight (g) 39.3 60.7 17.8 21.5 21.8 78.2 28.7 49.5
100.0 100.0
Split ratio (D/R) 0.65 0.83 0.28 0.58

Acylglycerol TAG 46.9 1.6 85.7 0.0 7.6 40.9 0.0 49.1 3.9 91.1
compositions DAG 40.4 54.4 14.1 7.0 89.9 45.8 5.9 50.2 81.8 8.8
(wt%) MAG 9.4 36.1 0.2 76.1 2.5 10.8 70.8 0.6 13.9 0.1
FFA 3.3 7.9 0.0 16.9 0.1 2.5 23.3 0.1 0.4 0.0
Yield (%) TAG 46.9 0.6 52.0 0.0 1.6 40.9 0.0 38.4 1.1 45.1
DAG 40.4 21.4 8.6 1.2 19.3 45.8 1.3 39.3 23.5 4.4
MAG 9.4 14.2 0.1 13.5 0.5 10.8 15.4 0.5 4.0 0.0
FFA 3.3 3.1 0.0 3.0 0.0 2.5 5.1 0.1 0.1 0.0
Recovery (%) TAG 1.34 110.92 0.00 259.86 0.00 93.88 2.92 117.45
DAG 52.92 21.18 5.83 90.41 2.81 85.71 59.80 11.10
MAG 150.93 1.29 95.48 3.79 142.91 4.34 850.23 10.55
FFA 94.08 0.00 96.89 0.69 203.18 3.13 146.80 0.00

Total recovery 72.11 63.58

DAG (%)

Results are means of two duplicates.

was recovered in the distillate and as high as 29.1% of DAG significant amount of FFA being detected as a consequence.
was lost. The total loss for the second scheme was 36.42%. The first scheme, on the other hand, involved first the
The DAG recovery was increased by the first approach as separation of DAG from TAG at 250°C and a second
compared to the second approach. purification step at 180°C as a refinement to remove low
For the first scheme, the distillate from the first distillation boiling point acylglycerols such as MAG and FFA, producing
step (250°C) contained 7.9‐wt% of FFA and 54.4‐wt% of highly concentrated DAG with FFA less than 0.1‐wt%.
DAG which it was further purified at 180°C to remove FFA.
The DAG content in end‐product (residue of the second 3.3 Characterization of purified DAG
step) was very satisfactory being 89.9‐wt% and almost no
FFAs were detected which met the product specification of A detail characterization had been carried out on purified
not more than 0.1‐wt% of FFA. For second scheme, the DAG product that had been obtained from the first approach
residue of the first step (180°C) which contained 50.3‐wt% of of double‐step purification.
DAG and trace amount of FFA was further purified at 250°C.
The end‐product (distillate of the second step) contained 3.4 DAG profile by high‐performance liquid
81.8‐wt% of DAG and significant amount of FFA (0.4‐wt%). chromatography
The FFA was concentrated and exceeded the product
specification limit of less than 0.1‐wt% of FFA, indicating A detailed DAG profile for the product was identified using
the necessity of an additional purification step. This HPLC with available DAG reference standard. DAG
experimental data provided valuable information on suitabil- reference standard, which was not available commercially
ity of the first scheme of double‐step distillation process to was predicted using equivalent carbon number (ECN) as
be employed in DAG purification even though a larger presented in the Eq. (3). The ECN related to both total
throughput was obtained in the second scheme. The inability number of double bonds (DB) and number of carbon atoms
of the second approach in obtaining distilled DAG with (CN) in the fatty acids was given in Table 4.
desired product specifications could be ascribed to the fact
that the residue stream still contained trace amount of ECN ¼ CN 2ðDBÞ ð4Þ
FFA after the first purification step at 180°C. The following
purification step at 250°C therefore resulted in the entrain- DAG profile based on the available reference standards
ment of the FFA into the distillate stream, leading to ECN prediction was shown in Fig. 2. From the injection of

ß 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.ejlst.com

Eur. J. Lipid Sci. Technol. 2014, 116, 1654–1663 Distillation and characterization of diacylglycerol 1661

Table 4. DAG compositions of purified DAG products longer time to elute out and the presence of DB in DAG
would shorten the elution period.
DAG
CN DB ECN

LL 36 4 28
3.5 Fatty acid composition and iodine value
ML 32 2 28
MM 28 0 28 The fatty acid composition of the purified DAG product was
OL 36 3 30 compared with that of palm olein IV 56 which was the raw
PL 34 2 30 material for the glycerolysis reaction (Table 5). Results
MO 32 1 30 showed that the amount of palmitic acid in the product
MP 30 0 30 increased from 41.43  0.55 to 47.5  0.50‐wt% while the
OO 36 2 32 oleic acid reduced from 41.27  0.38 to 36.07  0.21‐wt%.
SL 36 2 32 Acyl migration of fatty acids with different chain lengths and
PO 34 1 32
saturation was proposed to contribute to this circumstance.
PP 32 0 32
Unsaturated fatty acid located in the 1,3‐positions was
MS 32 0 32
SO 36 1 34
suggested to migrate to 2‐position during distillation process
PS 34 0 34 at high temperature to form unstable 1,2‐DAG molecules,
SS 36 0 36 which were later resynthesized to produce TAG. The
resultant TAG was subsequently distilled off in SPD

operation, resulting in lower unsaturated fatty acid content.

L, linoleic; O, oleic; S, stearic; P, palmitic; M, myristic.
Furthermore, the reaction mixtures produced via glycerolysis
was expected to have lower amount of unsaturated fatty acid
reference standard, it was found that 1,3‐DAG eluted faster or oleic acid because Lipozyme TLIM displayed sn‐1,3
than 1,2‐DAG. It was also worth mentioned that the elution selectivity towards TAG (palm olein) and glycerol. It was
of the DAG was in group based on ECN. Same ECN would worth noting that saturated fatty acids in palm olein took sn‐1
elute out at almost the same time whereas higher ECN took position with higher probability [11, 24]. Iodine value (IV)

Figure 2. DAG profiles by HPLC where L, linoleic; M, myristic; O, oleic; P, palmitic; S, stearic acid.

ß 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.ejlst.com

1662 C. M. Yeoh et al. Eur. J. Lipid Sci. Technol. 2014, 116, 1654–1663

Table 5. Fatty acid composition and iodine value of palm olein and 80
purified DAG products 70 Palm Olein IV 56

Solid Fat Content (%)

60 Glycerolysis Product
Fatty acid Purified DAG 50 High DAG Product
composition Palm olein product 40
30
12:0 0.33  0.06a 0.83  0.12b 20
14:0 1.13  0.06a 1.70  0.10b 10
16:0 41.43  0.55a 47.50  0.50b 0
16:1 0.3  0.00a 0.27  0.06a 0 10 20 30 40 50 60 70
18:0 4.03  0.21a 3.43  0.32a Temperature (oC)
18:1 41.27  0.38b 36.07  0.21a
18:2 11.17  0.55b 9.67  0.25a Figure 3. Solid fat content for palm olein IV56, Glycerolysis Product
18:3 0.37  0.06a 0.3  0.00a and purified‐DAG products.
20:0 0.03  0.06a 0.17  0.06b
IV 56.67  1.16b 47.67  0.58a
which is higher than that of palm olein. Purified DAG product
showed the highest SFC. Both purified DAG product and
Values in the same row with different letters are significantly different
(p  0.05). glycerolysis reaction mixtures demonstrated similar SFC
profile pattern but the former showed higher SFC than that
of the glycerolysis product. The high DAG product was
was also observed to be lower in purified DAG (IV
48) as completely melted at 65°C while the glycerolysis product
compared to the corresponding reactant palm olein (IV
56), was at 45°C. For palm olein, it had a sharp slope and was
which was in accordance with the fatty acid profiles of purified fully melted at 25°C. Based on the results obtained, purified
DAG product. DAG‐oil may be incorporated as hard stock to edible oil blend
to produce margarine in order to impart its beneficial health
3.6 Slip melting point and tocols content effects.

The SMP and the tocols content were given in Table 6. It

was found that the SMP increased to 40.4  0.2°C after 4 Conclusions
glycerolysis reaction. After purification, SMP was further
increased to 52.1  0.1°C. It was expected as the palmitic acid Double‐step distillation process is a potent separation approach
in the product was found to increase. Tocols content was not that can be employed to purify DAG from lipase‐catalyzed
significantly reduced after the glycerolysis reaction. However, glycerolysis reaction mixtures. The two schemes studied
after purification process, the amount of tocols content was revealed that the first processing approach is more efficient
increased in the purified product and it was favorable as the in producing purified DAG‐oil which meets the product
product would be less susceptible to oxidization and give specification of more than 80‐wt% DAG and at most 0.1‐wt%
nutritional benefit to human body. FFA. During the first step of SPD, TAG was completely
removed from the acylglycerol mixtures at elevated temperature
3.7 Solid fat content of 250°C and DAG purity of the distillate was 54.4‐wt%.
Second step of SPD involved the removal of FFA and MAG at
The SFC profiles for palm olein, glycerolysis reaction lower evaporator temperature of 180°C, yielding final concen-
mixtures and purified DAG product were given in Fig. 3. trated DAG‐oil in the residue with 89.9‐wt% purity. Purified
After glycerolysis reaction, it was found that the SFC was DAG‐oil was characterized by slightly lower IV of 48, high
reduced by 15% but it tends to melt completely at 15°C, SMP of 52.1°C, enriched with tocols (1134  10 ppm) and
high SFC, which melts completely at 65°C.
Table 6. Slip melting point and tocols content of palm olein,
interesterified palm olein, and purified DAG products The authors would like to thank the Director General, MPOB for
permission to publish this paper.
Sample/temperature SMP Tocols content
(°C) (°C) (ppm)
The authors have declared no conflict of interest.
Palm olein IV 56 24.7  0.1a
634  12 a

IE palm olein IV 56 40.4  0.2b 620  8a

Purified DAG product 52.1  0.1c 1134  10b References
Values in the same column with different letters are significantly [1] Flickinger, B., Matsuo, N., Nutritional characteristics of
different (p  0.05). DAG oil. Lipids 2003, 38, 129–132.

ß 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.ejlst.com

Eur. J. Lipid Sci. Technol. 2014, 116, 1654–1663 Distillation and characterization of diacylglycerol 1663

[2] Kamphuis, M. M., Mela, D. J., Westerterp‐Plantenga, M. S., [13] Yang, T., Zhang, H., Mu, H., Sinclair, A., Xu, X.,
Diacylglycerols affect substrate oxidation and appetite in Diacylglycerols from butterfat: Production by glycerolysis
humans. Am. J. Clin. Nutr. 2003, 77, 1133–1139. and short‐path distillation and analysis of physical properties.
[3] Maki, K. C., Davidson, M. H., Tsushima, R., Matsuo, N., J. Am. Oil Chem. Soc. 2004, 81, 979–987.
Tokimitsu, I., Umporowicz, D. M., Dicklin, M. R., Foster, [14] Fregolente, L. V., Fregolente, P. B. L., Chicuta, A. M.,
G. S., Ingram, K. A., Anderson, B. D., Frost, S. D., Bell, M., Batistella, C. B., Maciel Filho, R., Wolf‐Maciel, M. R., Effect
Consumption of diacylglycerol oil as part of a reduced‐energy of operating conditions on the concentration of monogly-
diet enhances loss of body weight and fat in comparison with cerides using molecular distillation. Chem. Eng. Res. Des.
consumption of a triacylglycerol control oil. Am. J. Clin. 2007, 85, 1524–1528.
Nutr. 2002, 76, 1230–1236. [15] Arzate‐Marti’nez, G., Jime’nez‐Gutie’rrez, A., Garci’a, H. S.,
[4] Taguchi, H., Omachi, T., Nagao, T., Matsuo, N., Experimental analysis and modeling of the separation of
Tokimitsu, I., Itakura, H., Dietary diacylglycerol suppresses triacylglycerol and free fatty acid mixtures using molecular
high fat diet‐induced hepatic fat accumulation and micro- distillation. Ind. Eng. Chem. Res. 2011, 50, 11237–11244.
somal triacylglycerol transfer protein activity in rats. J. Nutr. [16] Martins, P. F., Ito, V. M., Batistella, C. B., Maciel, M. R. W.,
Biochem. 2002, 13, 678–683. Free fatty acid separation from vegetable oil deodorizer
[5] Fregolente, P., Pinto, G., Wolf‐Maciel, M., Filho, R., distillate using molecular distillation process. Sep. Purif.
Monoglyceride and diglyceride production through lipase‐ Technol. 2006, 48, 78–84.
catalyzed glycerolysis and molecular distillation. Appl. [17] Meyer, F., Eggers, R., Oehlke, K., Harbaum‐Piayda, B.,
Biochem. Biotechnol. 2010, 160, 1879–1887. Schwarz, K., Siddiqi, M. A., Application of short path
[6] Lo, S. K., Tan, C. P., Long, K., Yusoff, M., Lai, O. M., distillation for recovery of polyphenols from deodorizer
Diacylglycerol oil—Properties, processes and products: A distillate. Eur. J. Lipid Sci. Technol. 2011, 113, 1363–
review. Food Bioprocess Tech. 2008, 1, 223–233. 1374.
[7] Sonntag, N., Glycerolysis of fats and methyl esters—Status, [18] Tenllado, D., Reglero, G., Torres, C. F., A combined
review and critique. J. Am. Oil Chem. Soc. 1982, 59, 795A– procedure of supercritical fluid extraction and molecular
802A. distillation for the purification of alkylglycerols from shark
[8] Lo, S. K., Baharin, B. S., Tan, C. P., Lai, O. M., liver oil. Sep. Purif. Technol. 2011, 83, 74–81.
Diacylglycerols from palm oil deodoriser distillate. Part 1— [19] Ghazali, H., Hamidah, S., Che Man, Y., Enzymatic trans-
Synthesis by lipase‐catalysed esterification. Food Sci. Technol. esterification of palm olein with nonspecific and 1,3‐specific
Int. 2004, 10, 149–156. lipases. J. Am. Oil Chem. Soc. 1995, 72, 633–639.
[9] Yeoh, C. M., Choong, T. S. Y., Abdullah, L. C., Yunus, R., [20] Swe, P., Man, Y., Ghazali, H., Composition of crystals of
Siew, W. L., Influence of silica gel in production of palm olein formed at room temperature. J. Am. Oil Chem.
diacylglycerol via enzymatic glycerolysis of palm olein. Soc. 1995, 72, 343–347.
Eur. J. Lipid Sci. Technol. 2009, 111, 599–606. [21] Ainie, K., MPOB Test Methods. Kajang, Selangor,
[10] Cheong, L. Z., Tan, C. P., Long, K., Affandi Yusoff, M. S., Malaysia: Malaysia Palm Oil Board. 2004.
Arifin, N., Lo, S. K., Lai, O. M., Production of a [22] Xu, X., Skands, A., Adler‐Nissen, J., Purification of specific
diacylglycerol‐enriched palm olein using lipase‐catalyzed structured lipids by distillation: Effects on acyl migration. J.
partial hydrolysis: Optimization using response surface Am. Oil Chem. Soc. 2001, 78, 715–718.
methodology. Food Chem. 2007, 105, 1614–1622.
[23] Xu, X., Skands, A. R. H., Høy, C. E., Mu, H., Balchen, S.,
[11] Wang, Y., Zhao, M., Song, K., Wang, L., Han, X., Tang, S., Adler‐Nissen, J., Production of specific‐structured lipids by
Wang, Y., Separation of diacylglycerols from enzymatically enzymatic interesterification: Elucidation of acyl migration
hydrolyzed soybean oil by molecular distillation. Sep. Purif. by response surface design. J. Am. Oil Chem. Soc. 1998, 75,
Technol. 2010, 75, 114–120. 1179–1186.
[12] Xu, X., Jacobsen, C., Nielsen, N. S., Heinrich, M. T., Zhou, [24] Che Man, Y. B., Haryati, T., Ghazali, H. M., Asbi, B. A.,
D., Purification and deodorization of structured lipids by Composition and thermal profile of crude palm oil and its
short path distillation. Eur. J. Lipid Sci. Technol. 2002, 104, products. J. Am. Oil Chem. Soc. 1999, 76, 237–242.
745–755.

ß 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.ejlst.com