Purification of 2-monoacylglycerols using liquid CO2 extraction

Article abstract of DOI:10.1007/s11746-012-2035-9

The fatty acid moiety of 2-monoacyl-sn-glycerol (2-MAG) undergoes spontaneous acyl migration to the sn-1(3) position, resulting in a thermodynamic equilibrium of approximately 1:9 of 2-MAG to 1(3)-monoacyl-sn-glycerol (1-MAG). Spontaneous acyl migration is an impediment to synthesizing and isolating specific 2-MAG for use as intermediates in the synthesis of structured triacylglycerols. 2-Monooleoyl-sn-glycerol was synthesized by the enzymatic ethanolysis of triolein and isolated by liquid CO₂ extraction. The resultant MAG, diacylglycerol, and fatty acid ethyl esters were quantified by ¹H NMR and supercritical fluid chromatography. The low polarity of the CO₂ and mild extraction temperature (25 °C) resulted in very low spontaneous acyl migration rates, allowing the MAG to be isolated in 80% yield and in a very high 2-MAG:1-MAG ratios of ≥93 mol%.

Full text of DOI:10.1007/s11746-012-2035-9

J Am Oil Chem Soc (2012) 89:1529–1536
DOI 10.1007/s11746-012-2035-9
ORIGINAL PAPER
Purification of 2-Monoacylglycerols Using Liquid CO2 Extraction
•
David L. Compton Fred J. Eller
•
•
Joseph A. Laszlo Kervin O. Evans
Received: 4 November 2011 / Revised: 2 February 2012 / Accepted: 2 February 2012 / Published online: 26 February 2012
Ó AOCS (outside the USA) 2012
Abstract The fatty acid moiety of 2-monoacyl-sn-glycerol
(2-MAG) undergoes spontaneous acyl migration to the
sn-1(3) position, resulting in a thermodynamic equilibrium
of approximately 1:9 of 2-MAG to 1(3)-monoacyl-sn-glyc-
erol (1-MAG). Spontaneous acyl migration is an impediment
to synthesizing and isolating specific 2-MAG for use as
intermediates in the synthesis of structured triacylglycerols.
2-Monooleoyl-sn-glycerol was synthesized by the enzy-
matic ethanolysis of triolein and isolated by liquid CO2
extraction. The resultant MAG, diacylglycerol, and fatty
acid ethyl esters were quantified by 1H NMR and supercrit-
ical fluid chromatography. The low polarity of the CO2 and
mild extraction temperature (25 °C) resulted in very low
spontaneous acyl migration rates, allowing the MAG to be
isolated in 80% yield and in a very high 2-MAG:1-MAG
ratios of C93 mol%.
Introduction
The spontaneous migration of mono- and diacylglycerol
(MAG and DAG) fatty acid moieties between the three
regiospecific sn-positions of the glycerol backbone has
been investigated since the 1920s [1]. Although consensus
has attributed the acyl migration mechanism to a five-
member intramolecular ketal-ring intermediate initiated by
a nucleophilic attack of a primary hydroxyl oxygen on the
secondary acyl carbonyl group [1–3], the phenomenon is
far from being understood and continues to be a subject
worthy of investigation [4, 5]. Dependent on temperature,
solvent polarity, water activity, fatty acid chain length,
acid/base impurities, and time [2, 5–8], spontaneous acyl
migration of MAG eventually reaches a thermodynamic
equilibrium of *1:9 ratio of 2-monoacyl-sn-glycerol
(2-MAG) to 1(3)-monoacyl-sn-glycerol (1-MAG), and
DAG eventually reaches a thermodynamic equilibrium of
*1:2 ratio of 1(3),2-diacyl-sn-glycerol (1,2-DAG) to 1,3-
diacyl-sn-glycerol (1,3-DAG) [6–9]. This spontaneous shift
of fatty acid moieties from the center glycerol position to
the peripheral glycerol positions complicates the synthesis
of structured lipids for which specific fatty acid moieties
are desired at the sn-2 position.
Keywords Acyl migration Á Liquid CO2 Á
2-Monoacylglycerol Á NMR Á Structured lipids
Mention of trade names or commercial products in this publication is
solely for the purpose of providing specific information and does not
imply recommendation or endorsement by the US Department of
Agriculture. The USDA is an equal opportunity employer.
The preferential physiological absorption and metabo-
lism of certain lipids by the mammalian gut and liver have
led to the design of structured lipids containing polyun-
saturated fats (e.g., eicosapentaenoic, docosahexaenoic) in
the sn-2 position flanked by medium-chain fatty acids (e.g.,
caprylic) in the sn-1 and 3 positions for nutritional studies
[10]. Similarly, human milk fat substitutes have been
produced from 2-MAG from tripalmitin and olive oil that
are esterified with medium chain fatty acids (e.g., capric,
caprylic) [11, 12]. Acidolysis of TAG enriched with a long
chain or polyunsaturated fatty acid of preference in the sn-2
D. L. Compton (&) Á J. A. Laszlo Á K. O. Evans
Renewable Product Technology Research Unit, National Center
for Agricultural Utilization Research, Agricultural Research
Service, US Department of Agriculture, Peoria, IL 61604, USA
e-mail: david.compton@ars.usda.gov
F. J. Eller
Functional Foods Research Unit, National Center for
Agricultural Utilization Research, Agricultural Research
Service, US Department of Agriculture, Peoria, IL 61604, USA
123

1530
J Am Oil Chem Soc (2012) 89:1529–1536
position with medium-chain fatty acids via a 1,3-specific
lipase is the most direct route to the medium-long/unsatu-
rated-medium-chain structured lipids. However, due to
hydrolytic side reactions and acyl migration the yield of the
desired long chain or polyunsaturated fatty acid at the sn-2
position of the structured lipid can be as low as 20% [13].
Enzymatic interesterification of medium-chain and long
chain TAG also produces medium-long-medium-chain
structured lipids, but random enzymatic distribution of fatty
acids among the three regiospecific positions and acyl
migration result in a mixture of products. The highest
incorporation of desired fatty acid at the sn-2 position in
these types of reactions ranges from 40 to 65% [14, 15]. A
more efficient route to sn-2 fatty acid specific structured
lipids is the 1,3-specific enzymatic alcoholysis of a sn-2-
long chain or polyunsaturated fatty acid enriched TAG,
isolation of the resultant long chain or polyunsaturated fatty
acid enriched 2-MAG, followed by 1,3-specific enzymatic
esterification of the 2-MAG with the appropriate medium-
chain fatty acid. This method results in sn-2 incorporation
of the desired long-chain fatty or polyunsaturated fatty acid
into the structured lipid at up to 95% [16].
resultant 1,2-DAG required further purification with vola-
tile organic solvents and extractions [8]. We investigated
the use of liquid CO2 in a fractionation column to remove
fatty acid ethyl esters (FAEE) and DAG from 2-MAG
synthesized from the enzymatic alcoholysis of 1,2,3-tri-
oleoyl-sn-glycerol (triolein) with excess ethanol. The
reaction products were characterized and quantified by
supercritical fluid chromatography (SFC) and spontaneous
acyl migration was quantified by 1H NMR spectroscopy [7].
Materials and Methods
Materials
1,2,3-Trioleoylglycerol (triolein) was purchased from
Sigma-Aldrich (St. Louis, MO, USA). Novozym 435,
Candida antarctica lipase B immobilized on a macropo-
rous acrylic resin, was obtained from Novozymes A/S
(New York, NY, USA). Ethanol (200 proof) was purchased
from Decon Laboratories, Inc. (King of Prussia, PA, USA).
Supercritical fluid chromatography (SFC)/supercritical
fluid extraction (SFE) grade CO2 (99.8% purity) was pur-
chased from ILMO Products Co. (Peoria, IL, USA). All
other reagents and solvents were purchased from Sigma-
Aldrich.
The key to the more efficient, two-step enzymatic route
to the sn-2 fatty acid specific structured lipids is the iso-
lation of the 2-MAG intermediate. The 2-MAG, typically
prepared by the enzymatic alcoholysis of TAG in excess
alcohol [7] must be separated from residual solvent, fatty
acid esters and DAG co-products while limiting acyl
migration of the fatty acid to the sn-1(3) position. Isolation
of 2-MAG by low-temperature crystallization involves
dissolving the crude 2-MAG in large volume of a binary
solvent system (e.g., n-hexane:methyl-t-butyl ether) and
cooling to -25 °C for at least an hour, resulting in crystals
of 2-MAG [16]. After repeated recrystallizations, 2-MAG
is isolated in 70–90% yield at 85% purity relative to
1-MAG [16]. 2-MAG can also be isolated by liquid–liquid
extraction [7]. In this case, crude 2-MAG is taken up in a
large volume of a binary solvent system (e.g., acetoni-
trile:water) that is then extracted by a series of increasingly
polar solvents (e.g., hexane, then chloroform, then
water:ethanol) [7]. 2-MAG purified by liquid–liquid
extraction is typically isolated in 80–90% yield at 85–90%
purity relative to 1-MAG [7, 10].
Methods
Synthesis of 2-Monooleoyl-sn-glycerol (2-MAG)
Time-dependent reactions were conducted in four identical
analytical trials. Triolein (0.200 g, 225 lmol) was dis-
solved in excess ethanol (0.800 g, 17.4 mmol) in a 100-mL
round bottom flask. Novozym 435 (0.1 g) was added and
the reaction mixture was rotated at room temperature for 1,
2, 3, or 4 h. The enzyme was removed from the reaction
mixture by filtration through a Whatman Anotop 25
(0.2 lm) inorganic membrane syringe filter into a 10 ml,
conical flask. Excess ethanol was removed via rotoevapo-
1
ration, and the resultant oily residue was analyzed by H
NMR. The analytical trials were replicated three times.
For quantities to be purified by liquid CO2 extractions,
triolein (10.0 g, 11.3 mmol) was dissolved in excess eth-
anol (40.0 g, 868 mmol) in a 100-mL round bottom flask.
Novozyme 435 (5.0 g) was added and the reaction mixture
was shaken (250 rpm) on an orbital shaker at room tem-
perature for 4 h. The reaction mixture was filtered through
a Spectrum Laboratories (Rancho Dominguez, CA, USA)
Spectra/Mesh nylon filter (20 lm) to remove the enzyme.
The excess ethanol was removed by rotoevaporation at
25 °C to yield *11 g of pale yellow oil, which was ana-
Both low-temperature crystallization and liquid–liquid
extraction purification methods result in moderately high
yields of 2-MAG while limiting spontaneous acyl migration
to *10%. In the present work, we attempted to improve
upon the liquid–liquid extraction technique by eliminating
all volatile organic solvents and reducing the purification to
a single extraction, requiring no subsequent washings or
precipitations. Liquid CO2 has been shown to be a viable
solvent for removing fatty acid esters while limiting acyl
migration during the isolation of 1,2-DAG, although the
1
lyzed by H NMR and SFC.
123

J Am Oil Chem Soc (2012) 89:1529–1536
1531
Liquid CO2 Extraction of 2-MAG
pressure of 31.4 MPa (310 atm), which was then held for
11 min. The column temperature, 100 °C, was isothermal
for the duration of the analysis. The FID was set at 350 °C.
Crude 2-MAG was purified using liquid CO2 extraction
with the apparatus illustrated in Fig. 1. A stainless-steel
column (20.3 cm 9 0.8 cm inner dia.) kept at ambient
temperature (25 °C) was packed with protruded stainless
steel-packing material (0.4 cm Pro-Pak, Scientific Devel-
opment Co., State College, PA, USA) resulting in 94%
column void volume. Crude 2-MAG (2.0 g) was loaded
onto the packing material at the bottom of the column.
Liquid CO2 was pumped via an ISCO 260D syringe pump
(Lincoln, NE, USA), jacketed at 5 °C, at 8.62 MPa (1,250
psi) constant pressure into both the top and bottom of the
column to fill the void volume. Liquid CO2 was then
pumped upward from the bottom of the column at
8.62 MPa and a rate of 1.5 mL/min. The CO2 and extract
exited through a heated micrometering valve (Autoclave
Engineers, Erie, PA, USA) set at 60 °C. Extracted material
was collected in a downstream vial and the total CO2 used
(36:1 wt:wt CO2 to crude 2-MAG ratio) was measured by
an American Meter Co. DTM-200A flow totalizer (Phila-
delphia, PA, USA). The raffinate (purified 2-MAG) was
collected from the column packing material by centrifu-
gation in a conical vial. The extractions were performed
three times.
NMR Spectroscopy
1H- and 13C-NMR spectra were obtained on a Bruker
(Fremont, CA, USA) Avance 500 spectrometer (500 MHz
1H/125.77 MHz 13C) using a 5 mm BBI probe. All samples
were dissolved in CDCl3, and all spectra were acquired
at 27 °C. Chemical shifts are reported as ppm from tetra-
methylsilane calculated from the lock signal (ND =
15.350609%).
Results and Discussion
SFC and NMR Analysis
SFC was a qualitative method for monitoring the alcohol-
ysis reaction products. SFC separated the reaction products
based on the solute’s propensity to partition into super-
critical CO2 [17–19]. Retention times of the reaction
products were confirmed versus standards. Without estab-
lishing calibration curves to correlate the reaction products’
concentrations to peak areas SFC could not be used to
determine absolute yields for the alcoholysis reaction
products. SFC was useful, however, for determining rela-
tive quantities of the crude reaction mixture products and
the purified products. The SFC chromatogram of the crude
alcoholysis reaction mixture (Fig. 2) showed that FAEE
was *83% of the total peak area relative to MAG (*16%)
and DAG (*1%). TAG accounted for \0.1% of the
product mixture (Table 1). SFC did not resolve the MAG
Supercritical Fluid Chromatography (SFC)
Reactions were analyzed by SFC using a Selerity Tech-
nologies’ (Salt Lake City, UT, USA) Series 4000 chro-
matograph with CO2 as the mobile phase. Samples
(100 lL) taken from the reactions were diluted in 1.0 mL of
hexane and 10 lL used for injection. The column was a SB-
methyl-100 SFC column (PN AE002, 50 lm 9 10 m x
0.25 lm film, Selerity Technologies) starting at 10.1 MPa
(100 atm) and 100 °C. The starting pressure was held for
5 min and ramped at 15.2 MPA (15 atm)/min to a final
1
and DAG positional isomers, so H NMR was the analyt-
ical method of choice for determining mol% yields of the
reaction products.
Fig. 1 Diagram of the high pressure apparatus used for liquid CO2 separation of reaction products from the Novozym 435 catalyzed alcoholysis
of triolein in excess ethanol
123

1532
J Am Oil Chem Soc (2012) 89:1529–1536
200
150
100
50
FAEE
was recovered after removing the unreacted ethanol. Ide-
ally, in this crude reaction mixture, the oleoyl fatty acids
were removed from the sn-1,3 positions of the TAG pro-
ducing 33 mol% 2-MAG and 66 mol% FAEE. The reac-
MAG
1
DAG
16
tion products as quantified by H NMR showed that the
raffinate
extract
crude
TAG was converted to 26 mol% MAG and 71 mol%
FAEE (Table 1). A low amount of DAG (\2 mol%) and
no TAG were found in the 4-h reaction mixture. The dis-
crepancy between the theoretical MAG mol% yield and the
actual mol% yield is discussed below.
0
11
12
13
14
15
17
Minutes
Fig. 2 SFC chromatograms of the reaction products from the
enzymatic alcoholysis of triolein with excess ethanol: crude reaction
mixture (bottom trace), extract from liquid CO2 extraction (middle
trace), and raffinate from liquid CO2 extraction (top trace)
The acyl migration of 2-MAG to form 1-MAG during
the 4-h reaction was determined as the mol% of 2-MAG
relative to 1-MAG [7]. The acyl migration of the very low
quantities of 1,2-DAG to 1,3-DAG was similarly deter-
1H NMR has been used to identify and quantify the
regiospecific isomers of MAG and DAG [7, 8]. Structures
of the relevant acylglycerols and fatty acid esters investi-
gated in the present study are shown in Fig. 3. The inte-
gration of the fatty acid x-carbon, methyl group protons
(0.89 ppm) was used as the reference value for the total
fatty acid groups in the reaction mixtures. The ratio of the
integrated b-proton signals of the MAG (4.93 ppm), DAG
(5.09 ppm), TAG (5.30 ppm) and the -O-CH2-CH3 protons
signals (4.13 ppm) of the FAEE relative to the x-carbon
methyl group protons was used to calculate the mol% of
each species in the sample [7, 8].
mined [8]. After the 4-h reaction, H NMR showed that of
1
the MAG formed 97% was 2-MAG and 3% was 1-MAG.
Similarly, of the DAG formed 97% was 1,2-DAG and 3%
was 1,3-DAG (Table 2). This amount of 2-MAG acyl
migration was significantly lower than the amount of acyl
migration previously reported for other purification meth-
ods, resulting in 85–90% 2-MAG relative to 1-MAG [10,
20]. It was possible that the degree of 2-MAG acyl
1
migration was less because of the rapidity of H NMR
analysis method compared to GC and HPLC [7]. Although
the crude reaction mixture contained very high purity
2-MAG, 97%, relative to 1-MAG, it was necessary to
determine what effect, if any, liquid CO2 extraction of the
FAEE had on acyl migration (see below).
2-MAG Synthesis
Quantitative conversion of TAG was achieved in 4 h by the
Novozym 435-catalyzed alcoholysis of triolein in excess
ethanol. The enzymatic reaction was conducted under dry
conditions to limit hydrolytic side reactions. Ninety eight
wt% of the theoretical mass of the crude reaction mixture
Unavoidable Formation of Glycerol
The alcoholysis reaction as prescribed in the literature [10,
20] supposedly yields 80–90% MAG with very high ratios
1
of 2-MAG (*90%) relative to 1-MAG. H-NMR analysis
Table 1 Composition of products obtained from the ethanolysis of triolein after 4 h and subsequent purification with liquid CO2
Species
Compositiona
Crude
Raffinate
Extract
NMRb (mol%)
SFCc (%)
NMR (mol%)
SFC (%)
NMR (mol%)
SFC (%)
FAEE
MAG
71.3 0.7
25.9 0.4
25.2 0.4
0.7 0.1
1.4 0.5
1.3 0.5
0.1 0.0
\0.1
83.1 0.6
1.5 0.5
97.2 1.5
91.0 0.5
6.2 1.0
5.3 1.8
4.7 1.5
0.6 0.3
\0.1
0.4 0.4
96.4 0.0
2.8 0.5
2.6 0.5
0.2 0.0
0.4 0.2
0.4 0.2
0.0 0.0
\0.1
97.6 0.1
15.8 0.4
93.0 0.5
2.3 0.1
2-MAG
1-MAG
DAG
–
–
–
–
–
–
1.1 0.3
6.7 0.3
0.1 0.0
1,2-DAG
1,3-DAG
TAG
–
–
–
–
–
–
\0.1
\0.1
\0.1
a
Data are the average of n = 3 trials with standard deviations
b
mol% of each glycerol species was determined based on integration values of their respective b-protons and mol% of FAEE was determined
based on the integration values of the –O-CH2-CH3 protons. The MAG and DAG mol% were the sum of their respective positional isomers
c
Percent compositions were determined based on integrated peak areas. SFC did not resolve the sn-regiospecific glycerol isomers
123

J Am Oil Chem Soc (2012) 89:1529–1536
1533
Fig. 3 Structures of reaction
products from the Novozym 435
catalyzed alcoholysis of triolein
in excess ethanol and scheme of
their separation by liquid CO2
extraction
Crude Reaction Mixture
O
O
O
O
OCR
OCR
OCR
OH
OH
OCR
O
O
O
O
OCR
OCR
OH
OCR
OCR
OH
O
O
OH
OCR
OCR
OH
TAG
1,3-DAG
1-MAG
1,2-DAG
FAEE
2-MAG
R= -(CH2)7-CH=CH-(CH2)7-CH3
Liquid CO2
(25 °C, 8.62 MPa)
less soluble
more soluble
Extract
FAEE
Raffinate
2-MAG 1-MAG
1,2-DAG 1,3-DAG
TAG
ignores glycerol formation and MAG yields are reported
relative only to the residual DAG and TAG [10, 20]. We
thought it prudent to quantify the amount of glycerol
formed during the alcoholysis reaction and to determine if
the alcoholysis could be optimized to limit glycerol
formation.
Table 2 Effect of liquid CO2 extraction on the composition of
1
2-MAG and 1,2-DAG as determined by H NMR
Acylglycerol
Composition (mol%)a
Crude
Raffinate
Extract
2-MAG
1,2-DAG
a
97.4 0.3
96.9 0.2
96.3 3.1
92.6 4.5
92.9 2.7
99.5 0.6
The mol% of glycerol formed during alcoholysis was
indirectly determined by 1H NMR as the difference
between the theoretical mol% yield of MAG less the mol%
of MAG, DAG, and TAG formed. Glycerol quantities were
independently confirmed by HPLC-ELSD (data not shown)
and were in good agreement with the quantities indirectly
determined by 1H NMR. This confirmed the expediency of
1H NMR as a method for quantitatively determining
acylglycerol species and their byproducts.
Data are the average of n = 3 trials with standard deviations. Mol%
of 2-MAG and 1,2-DAG are reported relative to 1-MAG and 1,3-
MAG, respectively
showed that after 4 h the alcoholysis reactions resulted in
*71 mol% FAEE, 26 mol% MAG, 2 mol% DAG, and
\0.1 mol% TAG. The product ratios revealed that the
FAEE were formed in quantities greater than the theoreti-
cal 66 mol% yield. The logical explanation to account for
the excess FAEE and lower than predicted amounts of
MAG was the unwanted enzymatic deacylation of MAG to
form glycerol. Glycerol could not be directly measured by
1H NMR or SFC. Indeed, to our knowledge, the literature
The mol% of glycerol was monitored over the course of
the 4 h alcoholysis reaction, and as shown in Fig. 4,
glycerol was present throughout the reaction. After the first
hour, only 85% of the TAG was converted. The quantity of
DAG formed, 12 mol%, was approximately equal to the
123

1534
J Am Oil Chem Soc (2012) 89:1529–1536
Similar to other liquid–liquid extraction methods for
purifying 2-MAG, [7, 10] liquid CO2 is used to extract the
FAEE byproduct formed during the alcoholysis of TAG.
The solubility or miscibility, reported as the partition coef-
ficient, of a particular solute in liquid or supercritical CO2
under specific pressures and temperatures is a function of the
solute’s polarity, molecular weight, and vapor pressure [18,
19, 22]. Earlier discrepancies in the literature as to the exact
values of partition coefficients for common lipid-type sol-
utes (e.g., triolein, diolein, monoolein, ethyl oleate) were
due to the purity of sample used, accuracy and limitation of
extraction techniques, and if the solutes are studied indi-
vidually or as binary, tertiary, or quaternary mixtures [22].
More recently, the partition coefficient values of these lipid
solutes have been more rigorously determined [23, 24].
Partitioning coefficient values can also be estimated for pure
substances using the Chrastil equation [25]. Additionally,
King and Friedrich [26] predicted the theoretical solubility
of solutes as the ratio of the solubility parameter of the
solvent gas to that of the dissolved solute based on the sol-
ute’s molecular structure. These theoretical values, how-
ever, are typically ideal and can differ significantly from
solubilities and partition values in the presence of other
solutes [19, 22, 24, 27]. Regardless of the absolute value of
their partitioning coefficients, the Chrastil equation and
King and Friedrich’s reduced solubility parameter predict
and data found in the literature agree that the general trend
for increasing partitioning of solutes relevant to the current
study into liquid CO2 (8.6 MPa, 25 °C) is: glycerol \ tri-
olein \ diolein B monoolein \ oleic acid \\ ethyl oleate
[18, 19, 22, 26, 27].
100
80
60
40
20
0
TAG
DAG
MAG
FAEE
Glycerol
0
1
2
3
4
Time (h)
Fig. 4 Product yields from the Novozym 435 catalyzed alcoholysis
of triolein in excess ethanol. Dashed line represents the theoretical
mol% yield of MAG and the dashed-dotted line represents the
theoretical mol% yield of FAEE. Data are presented as the mean of
n = 3 trials with error bars denoting standard deviations
MAG formed, 9 mol%, and even at this earliest time point
glycerol was formed in 1–2 mol%. By the end of the 4 h
reaction, all the TAG had been consumed and the acyl-
glycerols in the reaction mixture were mostly MAG
(26 mol%) with a very small amount of DAG (*2 mol%).
The amount of glycerol formed, *6 mol%, however, was
significantly higher. After 3 h, it appeared that the alco-
holysis had reached equilibrium at *25 mol% MAG, and
that there was no need to continue the reaction for an
additional hour, during which the glycerol content
increased from 5 to 6 mol%. However, at 3 h there was
still a residual amount of TAG (*0.5 mol%) present, and
the DAG content was twice that found after 4 h. Addi-
tionally, the experimental error for the glycerol content
determined at 3 h was within that found after 4 h. There-
fore, it was decided that the 4 h reaction was optimal,
producing the maximum amount of MAG and the lowest
amount of DAG and TAG (Table 1) without producing
significantly more unavoidable glycerol.
Liquid CO2 was used to extract FAEE from the crude
alcoholysis reaction mixtures. Figure 3 illustrates the
extraction scheme based on the partitioning trend of the
solutes. The extract, more soluble in the liquid CO2, is
carried from the crude reaction mixture and removed from
the extraction column (Fig. 1). The raffinate, which is less
soluble in liquid CO2, is left in the column. At 8.6 MPa and
25 °C, liquid CO2 removed nearly all of the FAEE from the
crude reaction mixture. After liquid CO2 extraction, the
raffinate contained [95 mol% MAG (Table 1). The raffi-
nate contained\2 mol% FAEE and *5 mol% DAG. Very
little of the MAG (\3 mol%) was carried over into the
liquid CO2 extract (Table 1). The near quantitative removal
of the FAEE from the raffinate and the slight carryover of
MAG into the extract are illustrated by the SFC analyses in
Fig. 2.
Liquid CO2 Purification of 2-MAG
The critical aspect of the TAG to 2-MAG synthesis is the
1,3-specific alcoholysis of the TAG fatty acids while lim-
iting spontaneous acyl migration of the remaining fatty acid
to form 1-MAG. Novozym 435 is reported in the literature
to be a non-regiospecific lipase [21]; however, under the
conditions used herein the lipase has demonstrated excel-
1
lent 1,3-specificity [7, 20]. H NMR analysis of the crude
The theoretical mass yield of 2-MAG based on complete
conversion of TAG was 36 wt% and the theoretical yield
of FAEE was 64 wt%. Correcting for the amount of
glycerol formed after 4 h, and ignoring the very small
residual quantity of DAG and TAG, the raffinate should
have accounted for 31 wt% and the extract 68 wt% of the
reaction mixture showed that the 2-MAG was formed in
97% purity relative to the 1-MAG (Table 2). It was nec-
essary to determine if this impressive 2-MAG:1-MAG ratio
found in the crude reaction mixture was maintained after
purification with liquid CO2.
123

J Am Oil Chem Soc (2012) 89:1529–1536
1535
mass of the crude reaction mixture before separation. Near
quantitative, 97%, recovery of the crude reaction mixture’s
mass was achieved after liquid CO2 extraction. The mass of
the raffinate recovered from the column was 23 wt% of the
initial mass of the crude reaction mixture and the extract
accounted for 74 wt%. It should be noted that recovery of
the raffinate from the stainless steel mesh by centrifugation
was not perfect, and some of the raffinate was lost. How-
ever, the recovered masses were in good agreement with
the predicted wt% and confirmed that the 2-MAG isolated
by liquid CO2 extraction contained the *6 mol% glycerol
Conclusions
Liquid CO2 showed itself to be a viable solvent for the
liquid–liquid extraction purification of 2-MAG obtained
from the ethanolysis of TAG. Liquid CO2 preferentially
extracted the FAEE leaving the purified 2-MAG (80%),
along with glycerol (18%) and very small quantities of
DAG (2%). A careful kinetic study of the alcoholysis
reaction showed that glycerol was formed at the earliest
stages of the reaction and was unavoidable if the reaction
was taken to the point of complete TAG consumption.
While the isolated 2-MAG may have contained a signifi-
cant amount of glycerol, the purity of 2-MAG relative to
1-MAG was exceptionally high, 93–99%. If required for
subsequent syntheses, the glycerol present could be
removed by an additional liquid–liquid solvent extraction
using a more polar solvent.
1
as predicted by H NMR and corroborated by HPLC.
Liquid CO2 extraction resulted in the isolation of
2-MAG in *80% yield relative to TAG, containing *18%
glycerol. This 2-MAG yield was towards the lower end of
yields reportedly obtained from similar, volatile organic
solvent, liquid–liquid extraction methods (80–90%) [10].
Because previous studies do not determine the amount of
glycerol and report 2-MAG yields relative only to other
acylglycerol species formed, these reported yields could
actually be considerably lower.
Acknowledgments The authors thank Jeffrey Teel for his excellent
technical assistance and Karl Vermillion for acquisition of all NMR
spectra.
Effect of Liquid CO2 Extraction on 2-MAG
Acyl Migration
References
1. Fischer E (1920) Wanderung von Acyl bei den Glyceriden. Ber
Deutschen Chem Gesellschaft 53:1612–1633
Liquid CO2 is an attractive solvent for the purification of
2-MAG. It is volatile, allowing for easy separation from the
acylglycerol products and is not considered a hazardous
solvent. Also, its polarity is low enough, similar to that of
hexane, to selectively remove FAEE. While these consid-
erations proved true, the low polarity of the CO2 was
predicted to result in higher rates of unwanted acyl
migration [4, 28]. It has been postulated that less polar
solvents decrease the energy of activation through the five-
member, ketal-ring intermediate to form the more ther-
modynamically stable, unbranched 1-MAG isomer [4]. It
was, therefore, possible that liquid CO2 extraction of
2-MAG could have resulted in significantly lower purity of
2-MAG relative to 1-MAG than that obtained from other
liquid–liquid extraction methods using more polar solvents,
85–90% [7, 10, 20].
2. Fureby AM, Virto C, Adlercreutz P, Mattiasson B (1996) Acyl
groupmigrationsin2-monoolein. BiocatalBiotransform14:89–111
3. Kodali DR, Tercyak A, Fahey DA, Small DM (1990) Acyl migra-
tion in 1,2-dipalmitoyl-sn-glycerol. Chem Phys Lipids 52:163–170
4. Li W, Du W, Li Q, Li Rw, Liu D (2010) Dependence on the
properties of organic solvent: study on acyl migration kinetics of
partial glycerides. Bioresour Technol 101:5737–5742
5. Li W, Du W, Li Q, Sun T, Liu D (2010) Study on acyl migration
kinetics of partial glycerides: dependence on temperature and
water activity. J Mol Catal B: Enzym 63:17–22
6. Boswinkel G, Derksen J, van’t Riet K, Cuperus F (1996) Kinetics
of acyl migration in monoglycerides and dependence on acyl
chainlength. J Am Oil Chem Soc 73:707–711
7. Compton DL, Vermillion KE, Laszlo JA (2007) Acyl migration
1
kinetics of 2-monoacylglycerols from soybean oil via H NMR.
J Am Oil Chem Soc 84:343–348
8. Laszlo JA, Compton DL, Vermillion KE (2008) Acyl migration
kinetics of vegetable oil 1,2-diacylglycerols. J Am Oil Chem Soc
85:307–312
1H-NMR analysis of the raffinate obtained from the
liquid CO2 extraction showed that 2-MAG was isolated in
93–99% purity relative to 1-MAG. This was an improve-
ment compared to other liquid–liquid extraction methods.
Similarly, the small amount of DAG in the raffinate was
almost exclusively the 1,2-isomer, 88–96%. The very small
quantities of MAG and DAG carried over into the extract
(Table 1) also possessed very high ratios of the sn-2
positional isomers relative to the sn-1 positional isomers,
[90% (Table 2). These data show that the low polarity of
CO2 had little influence on the spontaneous acyl migration
of the sn-2 positional isomers.
9. Mattson FH, Volpenhein RA (1962) Synthesis and properties of
glycerides. J Lipid Res 3:281–296
´
´
10. MdM Mun˜ıo, Robles A, Esteban L, Gonzalez PA, Molina E
(2009) Synthesis of structured lipids by two enzymatic steps:
ethanolysis of fish oils and esterification of 2-monoacylglycerols.
Process Biochem 44:723–730
11. Ilyasoglu H, Gultekin-Ozguven M, Ozcelik B (2011) Production
of human milk fat substitute with medium-chain fatty acids by
lipase-catalyzed acidolysis: optimization by response surface
methodology. LWT - Food Sci Technol 44:999–1004
´
´
´
12. Jimenez MJ, Esteban L, Robles A, Hita E, Gonzalez PA, Mun˜ıo
MM, Molina E (2010) Production of triacylglycerols rich in
palmitic acid at position 2 as intermediates for the synthesis of
human milk fat substitutes by enzymatic acidolysis. Process
Biochem 45:407–414
123

1536
J Am Oil Chem Soc (2012) 89:1529–1536
13. Lai OM, Low CT, Akoh CC (2005) Lipase-catalyzed acidolysis
of palm olein and caprylic acid in a continuous bench-scale
packed bed bioreactor. Food Chem 92:527–533
14. Maduko CO, Akoh CC, Park YW (2007) Enzymatic production
of infant milk fat analogs containing palmitic acid: optimization
of reactions by response surface methodology. J Dairy Sci 90:
2147–2154
15. Maduko CO, Akoh CC, Park YW (2007) Enzymatic interesteri-
fication of tripalmitin with vegetable oil blends for formulation of
caprine milk infant formula analogs. J Dairy Sci 90:594–601
16. Pfeffer J, Freund A, Bel-Rhlid R, Hansen CE, Reuss M, Schmid
RD, Maurer SC (2007) Highly efficient enzymatic synthesis of
2-monoacylglycerides and structured lipids and their production
on a technical scale. Lipids 42:947–953
17. Turner C, King JW, Mathiasson L (2001) Supercritical fluid
extraction and chromatography for fat-soluble vitamin analysis.
J Chromatogr A 936:215–237
synthesis of symmetrically structured tag. J Am Oil Chem Soc 79:
879–883
21. Jackson M, King J (1997) Lipase-catalyzed glycerolysis of soy-
bean oil in supercritical carbon dioxide. J Am Oil Chem Soc
74:103–106
¨
22. Gu¨c¸lu¨ U, Temelli F (2000) Correlating the solubility behavior of
fatty acids, mono-, di-, and triglycerides, and fatty acid esters in
supercritical carbon dioxide. Ind Eng Chem Res 39:4756–4766
23. Gupta RB, Shim J–J (2007) Solubility in supercritical carbon
dioxide. CRC Press, Boca Raton
24. Kassim KM, Davarnnejad R (2007) Supercritical fluid extraction
of binary systems of monoolein, diolein, and triolein using carbon
dioxide solvent. Int J Chem Reactor Eng 5:A45–A49
25. Chrastil J (1982) Solubility of solids and liquids in supercritical
gases. J Phys Chem 86:3016–3021
26. King JW, Friedrich JP (1990) Quantitative correlations between
solute molecular structure and solubility in supercritical fluids.
J Chromatogr A 517:449–458
27. Sahle-Demessie E (1997) Fractionation of glycerides using
supercritical carbon dioxide. Ind Eng Chem Res 36:4906–4913
28. Fureby AM, Tian L, Adlercreutz P, Mattiasson B (1997) Prepa-
ration of diglycerides by lipase-catalyzed alcoholysis of triglyc-
erides. Enzyme Microb Technol 20:198–206
18. Nilsson W, Gauglitz E, Hudson J (1991) Solubilities of methyl
oleate, oleic acid, oleyl glycerols, and oleyl glycerol mixtures in
supercritical carbon dioxide. J Am Oil Chem Soc 68:87–91
19. King JW, Sahle-Demessie E, Temelli F, Teel JA (1997) Thermal
gradient fractionation of glyceride mixtures under supercritical
fluid conditions. J Supercrit Fluids 10:127–137
20. Irimescu R, Iwasaki Y, Hou CT (2002) Study of tag ethanolysis
to 2-MAG by immobilized Candida antarctica lipase and
123