Effective Sample Preparation of Polyphenols in Wine Using Deep Eutectic Solvent-based Dispersive Liquid–Liquid Microextraction HPLC-UV Determination

Article abstract of DOI:10.1002/bkcs.12130

Polyphenols are phytochemicals that exist in grapes and are beneficial to human health. In this study, resveratrol, oxyresveratrol, and piceatannol in wine were extracted by deep eutectic solvent dispersive liquid–liquid microextraction (DES-DLLME), and a method was established for quantifying these polyphenols by high-performance liquid chromatography-UV/Vis (HPLC-UV/Vis). Several parameters pertaining to sample extraction, clean-up, and concentration were optimized and verified with central composite design (CCD) using Design Expert 11. The optimized sample preparation parameters are as follows: the DES extraction solvent, tributylmethylammonium chloride/decanoic acid (1:3 M ratio); basic solvent, 1.3 mL of 5% potassium bicarbonate; volume of acetic anhydride, 250 μL; derivatization time, 5 min; dispersive solvent, methanol; ratio of extraction and dispersive solvents, 1:5.5; and salt, 1.0 g. Chromatographic separation by HPLC/UV–Vis was performed on an ACME C18 (4.6 mm id × 150 mm length, 5 μm particle size) column in gradient elution mode using water and 70% methanol. Under the established extraction and HPLC-UV conditions, the limit of detection (LOD) and limit of quantitation (LOQ) of the three analytes in spiked samples ranged from 1.69 to 2.53 μg/L and 5.64 to 8.42 μg/L, respectively. Recovery studies were performed in low, medium, and high concentration ranges to establish a calibration curve, and the accuracy and precision in the working range were 95.1–108.0% and 1.3–6.7 RSD%, respectively. The calibration curves for quantitative analysis were obtained in the concentration ranges 5.6–56.4, 8.3–82.6, and 8.4–84.2 μg/L, with correlation coefficients (r²) ranging from 0.9947 to 0.9967. The proposed method was applied to the determination of polyphenols in wine samples.

Full text of DOI:10.1002/bkcs.12130

Article
DOI: 10.1002/bkcs.12130
BULLETIN OF THE
J. Noh et al.
KOREAN CHEMICAL SOCIETY
Effective Sample Preparation of Polyphenols in Wine Using Deep
Eutectic Solvent-based Dispersive Liquid–Liquid Microextraction
HPLC-UV Determination
*
Jongsung Noh, Seunghoon Song, and Seung-Woon Myung
Department of Chemistry, Kyonggi University, Suwon 16227, South Korea. *E-mail: swmyung@kgu.ac.kr
Received August 19, 2020, Accepted October 13, 2020
Polyphenols are phytochemicals that exist in grapes and are beneficial to human health. In this study, res-
veratrol, oxyresveratrol, and piceatannol in wine were extracted by deep eutectic solvent dispersive
liquid–liquid microextraction (DES-DLLME), and a method was established for quantifying these poly-
phenols by high-performance liquid chromatography-UV/Vis (HPLC-UV/Vis). Several parameters per-
taining to sample extraction, clean-up, and concentration were optimized and verified with central
composite design (CCD) using Design Expert 11. The optimized sample preparation parameters are as fol-
lows: the DES extraction solvent, tributylmethylammonium chloride/decanoic acid (1:3 M ratio); basic
solvent, 1.3 mL of 5% potassium bicarbonate; volume of acetic anhydride, 250 μL; derivatization time,
5 min; dispersive solvent, methanol; ratio of extraction and dispersive solvents, 1:5.5; and salt, 1.0 g.
Chromatographic separation by HPLC/UV–Vis was performed on an ACME C18 (4.6 mm id × 150 mm
length, 5 μm particle size) column in gradient elution mode using water and 70% methanol. Under the
established extraction and HPLC-UV conditions, the limit of detection (LOD) and limit of quantitation
(LOQ) of the three analytes in spiked samples ranged from 1.69 to 2.53 μg/L and 5.64 to 8.42 μg/L,
respectively. Recovery studies were performed in low, medium, and high concentration ranges to establish
a calibration curve, and the accuracy and precision in the working range were 95.1–108.0% and 1.3–6.7
RSD%, respectively. The calibration curves for quantitative analysis were obtained in the concentration
ranges 5.6–56.4, 8.3–82.6, and 8.4–84.2 μg/L, with correlation coefficients (r2) ranging from 0.9947 to
0.9967. The proposed method was applied to the determination of polyphenols in wine samples.
Keywords: Deep eutectic solvent, Dispersive liquid–liquid microextraction, Polyphenol, Wine,
HPLC-UV
Introduction
particular, it has a high concentration in grapes and foods
using grapes,6 and it has a higher concentration in red wine
grapes than it does in white wine grapes.7 When carbonate
is present in the winemaking process, the content of resver-
atrol may be reduced relative to that in wine produced in
the traditional manner.8
Interest in human health has increased as the quality of life
has increased in recent decades. Foods consumed by
humans can sustain good health, but people have also expe-
nded considerable effort to maintain healthy lives and pre-
vent aging. Accordingly, interest in nutrients with
biological activity has increased, as have healthy eating
habits.1
Polyphenols are contained in various nutrients exhibiting
biological activity. They are characterized by a variety of
structures with two or more acidic hydroxyl groups on the
aromatic rings (Figure 1), and it has been reported that,
the greater the number of hydroxyl groups in the molecule,
the stronger the antioxidant properties and the greater the
protection against free radicals. In this group,
oxyresveratrol, piceatannol, and resveratrol belong to the
stilbene family and are known to suppress cardiovascular
disease and inhibit brain degeneration and tumors.1,2 Res-
veratrol is produced when plants are exposed to bacteria,
fungi, and UV irradiation and are physically damaged.3–5 It
is found in peanuts, raspberries, blueberries, and so on. In
Several methods have been reported for the quantitative
analysis of stilbenes with high polarity. Analyses using
high-performance liquid chromatograph-ultraviolet detec-
tion
(HPLC-UV),9–11
high-performance
liquid
chromatograph-florescence detection (HPLC-FLD),12–14
and liquid chromatography-mass spectrometry (LC–MS)
have been utilized.15–17 The use of gas chromatography has
the disadvantage that derivatization is necessary because
the analytes are not sufficiently volatile, but it enables
detection at lower concentrations than does liquid
chromatography.18,19
Stilbenes exhibit low concentrations in foods, and the
matrix of food components is complex, so pretreatment is
essential for analysis. Classical extraction methods such as
liquid–liquid microextraction (LLE) require a long extrac-
tion time and use large volumes of harmful organic
Bull. Korean Chem. Soc. 2020
© 2020 Korean Chemical Society, Seoul & Wiley-VCH GmbH
Wiley Online Library
1

Article
BULLETIN OF THE
ISSN (Print) 0253-2964 | (Online) 1229-5949
KOREAN CHEMICAL SOCIETY
Table 1. HPLC-UV conditions for analysis of polyphenols.
Parameters
Conditions
Column
Mobile phase
ACME C18, 4.6 × 150 mm, 5 μm
A: 70% Methanol
B: Water
Gradient
Time
0
2
A(%)
10
10
B(%)
90
90
Figure 1. Chemical structure of polyphenols.
solvents, which can harm the environment and the human
body.6 To overcome this disadvantage, many different
extraction methods have been developed and applied. It has
been reported that a quick and effective pretreatment
method may be applied to use a microextraction method
and reduce the amount of solvent while enhancing selectiv-
ity. Applied microextraction methods include solid-phase
microextraction (SPME),16,17 single drop microextraction
(SDME),20 dispersion liquid–liquid microextraction
(DLLME),2 and stir bar adsorption extraction.21 Although
it is possible to reduce the amount of toxic organic solvents
used by applying a microextraction method, the methods
still involve exposure to harmful solvents. To eliminate
this, methods using supercritical fluids (SFs)22 and ionic
liquids (ILs)23 as extraction solvents have been proposed.
However, since supercritical fluids require specific equip-
ment to maintain the supercritical state, existing extraction
methods cannot be replaced by SFs immediately. ILs have
been touted as environmentally friendly solvents that
replace toxic organic solvents, but questions have arisen as
to whether they are truly environmentally friendly; some
ILs have exhibited environmental toxicity, high cost, and
limited biodegradability.24–26 Therefore, in this study, a dis-
persion liquid–liquid microextraction method (DLLME) is
used to enable desirable extraction and concentration per-
formance, even with small amounts of extraction solvent.
The toxic organic liquids normally used as extraction sol-
vents have been replaced with a deep eutectic solvent
(DES), so that an effective and environmentally friendly
analysis method for polyphenols may be established.
DESs are systems formed from a eutectic mixture of
Lewis or Brønsted acids and bases that can contain various
types of cationic and/or anionic species.26 The DES thus
formed gives a eutectic with a melting point lower than
those seen for the individual compounds. DESs can easily
be prepared by mixing a hydrogen bond donor (HBD) and
a hydrogen bond acceptor (HBA), and they exhibit desir-
able characteristics such as low conductivity, low vapor
pressure, high thermal stability, and high viscosity.27 In this
study, a DES was prepared by mixing an organic salt
(hydrogen bond acceptor) and a hydrogen bond donor. To
prepare a hydrophobic DES, quaternary ammonium salts
and decanoic acid, which contains a long alkyl chain, were
used, and the successfully prepared DES was used as an
extraction solvent.
12
35
1 mL/min
5 μL
304 nm
40
100
60
0
Flow rate
Injection volume
Wavelength
Experimental
Reagents and Materials.
Trans-Resveratrol, trans-
oxyresveratrol, and trans-piceatannol were purchased from
TCI (Seoul, Korea) and Toronto Research Chemicals
(Toronto, ON, Canada), respectively. The standard mate-
rials were dissolved in methanol to prepare a solution with
a concentration of 1000 μg/mL and stored in a refrigerator.
The stored solution was diluted with methanol when used.
Decanoic acid was used as an HBD, and ultra-high purity
reagent was obtained from JUNSEI (Tokyo, Japan). Tri-
butylmethylammonium chloride (N4441 Cl), methyltri-
octylammonium chloride (N8881 Cl), methyltrioctyl
ammonium bromide (N8881 Br), and tetraoctylammonium
bromide (N8888 Br) were used as HBA, and potassium
hydrogen phosphate, potassium carbonate, and potassium
bicarbonate were used as derivatization catalysts; all were
purchased from Sigma-Aldrich (St. Louis, MO). Acetic
anhydride was used as a derivatization reagent, and metha-
nol, ethanol, acetone, and acetonitrile were used as disper-
sion solvents, and all were HPLC grade reagents obtained
from Deoksan Co. (Gyeonggi, Korea). For studies of salt-
ing effects, sodium sulfate obtained from Daejeong
(Gyeonggi, Korea) was used. Purified water was passed
through a Synergy UV system (Millipore S.A.S, Molsheim,
France), and ultrapure water exhibiting a specific resistance
of 18.2 MΩÁcm was used.
Table 2. Factors and their levels for the central composite design.
Level
Factor
Low (−1) Medium (0) High (+1)
(x1) Volume of base
solution (mL)
(x2) Volume of acetic
anhydride (μL)
(x3) Ratio of Ext. and
Disp. solvent
0.7
1.0
400
1:4
1.5
1.3
250
550
1: 2.5
1.0
1: 5.5
2.0
(x4) Salt (Na2SO4, g)
Bull. Korean Chem. Soc. 2020
© 2020 Korean Chemical Society, Seoul & Wiley-VCH GmbH
www.bkcs.wiley-vch.de
2

BULLETIN OF THE
Article
Effective Sample Preparation of Polyphenols in Wine
KOREAN CHEMICAL SOCIETY
Figure 2. Acetylation mechanism of phenolic compound.
Apparatus and Instrumentation. The centrifuge used
was an MF80 instrument made by Hanil (Seoul, Korea),
and Design Expert 11 software from Stat Ease
(Minneapolis, MN) was used for statistical analysis. A vor-
tex mixer from Vision Scientific (Bucheon, Korea) was
used for mixing samples, and test tubes for centrifugation
were manufactured by Falcon (TX).
Figure 4. Effect of base solvent volume on peak area of
polyphenols.
optimization process was performed again by applying the
central composite design (CCD) method and using Design
Expert 11. The main experimental factors for considering
the interaction between experimental variables are (1) vol-
ume of derivatization catalyst (X1), (2) volume of derivati-
zation reagent (X2), (3) ratio of extraction solvent and
dispersion solvent (X3), and (4) the amount of salt (X4)
(Table 2), and these were applied to the CCD method.
HPLC Analysis. HPLC/UV–Vis experiments were per-
formed on an 1100 series HPLC system from Agilent (Palo
Alto, CA). An ACME C18 column from Phase Analytical
Technology, LLC (State College, PA), was used for the
simultaneous analysis of three polyphenols by liquid chro-
matography. The length and inner diameter of the column
were 150 mm and 2.1 mm, respectively, and the particle
size was 5 μm. For the HPLC mobile phase, a gradient elu-
tion method was used in which the ratio of water and 70%
methanol was maintained at 90:10 for the initial 2 min,
followed by 60:40 at 12 min, and 0:100 at 35 min. The
flow rate of the mobile phase and the injection volume
were 1 mL/min and 5 μL, respectively, and the detection
wavelength was 304 nm (Table 1).
Sample Preparation. The experimental procedure for the
DLLME is as follows. The 100 μL of DES as an extraction
solvent and 550 μL of dispersion solvent were added rap-
idly to the aqueous solution sample resulting after the
derivatization of the analyte with a base solvent as the cata-
lyst and acetic anhydride as derivatization reagent, and
1.0 g of sodium sulfate was added. The solution was then
gently shaken for 10 s. After centrifugation for 10 min at a
speed of 4000 rpm, the layers were separated, and 20 μL of
the extraction solvent comprising the upper layer was
removed and injected into the HPLC instrument.
Design of Experiments. Design of experiments was used
to develop experimental parameters, with the goal of
obtaining maximum information by identifying the main
variables affecting the experiment and minimizing the num-
ber of experiments. In this study, the optimal value for each
factor was selected by the one variable at a time (OVAT)
method that optimizes one experimental factor individually.
The OVAT method can easily be used to optimize each
experimental factor by fixing the other factors but has the
disadvantage that it is impossible to consider the effects of
those other experimental factors. Therefore, the
Results and Discussion
Optimization of Derivatization. A derivatization process
is required for the effective extraction and chromatographic
analysis of polyphenols contained in aqueous solution.
Through derivatization, hydrophilic properties can be
reduced to improve separation in
a reversed-phase
Figure 3. Effect of base solvent type on peak area of polyphenols:
(a) potassium phosphate dibasic, (b) potassium bicarbonate,
(c) potassium carbonate.
Figure 5. Effect of acetic anhydride volume on peak area of
polyphenols.
Bull. Korean Chem. Soc. 2020
© 2020 Korean Chemical Society, Seoul & Wiley-VCH GmbH
www.bkcs.wiley-vch.de
3

BULLETIN OF THE
Article
Effective Sample Preparation of Polyphenols in Wine
KOREAN CHEMICAL SOCIETY
Figure 6. Effect of derivatizing time on peak area of polyphenols.
Figure 8. Effect of ratio (1:1, 2:1, and 3:1) of hydrogen bond
donor to hydrogen bond acceptor on peak area of polyphenols.
chromatography system and to increase the extraction effi-
ciency of hydrophobic extraction solvents.2 For effective
derivatization, it is necessary to adjust the pH and establish
basic sample conditions. In basic solution, the hydroxyl
hydrogen of the polyphenols is easily dissociated, and
acetic anhydride, a derivatization reagent, is used to acety-
late the site (Figure 2). The type and volume of the deriva-
tization catalyst, volume of the derivatization reagent, and
derivatization time were considered as parameters in the
derivatization process, and their effects were investigated.
extraction efficiency may decrease. Therefore, it is impor-
tant to determine the optimal volume. Figure 4 shows a
graph comparing the areas of chromatograms resulting from
the volume of the potassium bicarbonate which is the deriv-
atization catalyst. The volume of the derivatization catalyst
was varied within the range of 0.5–4 mL to select the opti-
mal volume of the derivatization catalyst. The other condi-
tions were as follows: the derivatization catalyst used was
5% (w/v) potassium bicarbonate, the volume of derivatiza-
tion reagent (acetic anhydride) used was 800 μL, the deriv-
atization time was 5 min, the volume of extraction solvent
(N8881-Cl:DecA [1:2]) used was 100 μL, the volume of
dispersion solvent (acetone) used was 400 μL, and the
amount of salt (sodium sulfate) used was 1.5 g.
Amount of Acetic Anhydride. When the amount (volume)
of the derivatization reagent is small, derivatization does
not proceed completely, resulting in loss of sample. If the
amount of the reagent is large, the total volume of the sam-
ple increases, and this may decrease extraction efficiency.
Therefore, it is important to determine the appropriate
amount to achieve the optimal derivatization efficiency. To
select the amount of the derivatization reagent, the volume
was varied within the range of 50–1000 μL, the type and
volume of the derivatization catalyst were 5% (w/v) potas-
sium bicarbonate and 1.0 mL, respectively, the derivatiza-
tion time was 5 min, the volume of extraction solvent
(N8881-Cl:DecA [1:2]) used was 100 μL, the volume of
dispersion solvent (acetone) used was 400 μL, and the
amount of salt (sodium sulfate) used was 1.5 g. Figure 5
shows a graph comparing the areas of the chromatogram
peaks to the volumes of the derivatization reagent used,
Type of Catalyst. Dibasic potassium hydrogen phosphate,
potassium carbonate, and potassium bicarbonate (all at 5%
(w/v) concentration) were compared as basic catalysts for
derivatization. The conditions for selecting an optimized
derivatization catalyst are as follows: the volume of deriva-
tization catalyst used was 2.0 mL, the volume of derivatiza-
tion reagent (acetic anhydride) used was 800 μL, the
derivatization time was 5 min, the volume of extraction sol-
vent (N8881-Cl:DecA [1:2]) used was 100 μL, the volume
of dispersion solvent (acetone) used was 400 μL, and the
amount of salt (sodium sulfate) used was 1.5 g. Figure 3
shows a graph comparing the integrated areas of the chro-
matograms with the type of derivatization catalyst used.
Since the use of potassium bicarbonate resulted in the larg-
est peak area, potassium bicarbonate was selected as the
derivatization catalyst.
Volume of Base Solvent. If the volume of the derivatiza-
tion catalyst solution used is small, it may not have an
appropriate catalytic effect, whereas, when the volume is
large, the total volume of the sample increases and the
Figure 7. Effect of extraction solvent type on peak area of
Figure 9. Effect of dispersive solvent type on peak area of
polyphenols.
polyphenols.
Bull. Korean Chem. Soc. 2020
© 2020 Korean Chemical Society, Seoul & Wiley-VCH GmbH
www.bkcs.wiley-vch.de
4

BULLETIN OF THE
Article
Effective Sample Preparation of Polyphenols in Wine
KOREAN CHEMICAL SOCIETY
Figure 12. The HPLC chromatogram from the spike white wine
(LOQ concentration level).
Figure 10. Effect of ratio of extraction to dispersive solvent on
peak area of polyphenols.
hydrogen bond acceptors (HBA). The molar ratio of the hydro-
gen bond donor and the hydrogen bond acceptor was fixed as
2:1. The derivatization catalyst and volume were 5% (w/v)
potassium bicarbonate and 1.0 mL, respectively, the amount of
the derivatization reagent (acetic anhydride) was 400 μL, the
derivatization time was 5 min, the volume of the DES as an
extraction solvent was 100 μL, the volume of the dispersion
solvent (acetone) was 400 μL, and the amount of salt was
1.5 g. Figure 7 shows a graph comparing the extraction effi-
ciencies with the combination of hydrogen bond donor and
hydrogen bond acceptor by using the areas of peaks in the
chromatograms. Among the four hydrogen bond acceptors, tri-
butylmethylammonium chloride, which has the lowest hydro-
phobicity, resulted in the greatest chromatogram peak area for
the resulting extraction solvent (Figure 7).
and the optimal value for the volume of the derivatization
reagent is found to be 400 μL.
Derivatization Time. If the derivatization time is insuffi-
cient, the analytes may not be derivatized completely and
sample loss results. If the derivatization time is too long,
the time required for sample preparation will increase
unnecessarily. Therefore, it is important to shorten the sam-
ple preparation time by determining the required derivatiza-
tion time. The derivatization time was probed by changing
the time within the range of 5–80 min, the type and volume
of the derivatization catalyst were 5% (w/v) potassium
bicarbonate and 1.0 mL, respectively, the amount of the
derivatization reagent used was 400 μL, the volume of
extraction solvent (N8881-Cl:DecA [1:2]) used was
100 μL, the volume of dispersion solvent (acetone) used
was 400 μL, and the amount of salt (sodium sulfate) used
was 1.5 g. Figure 6 shows a graph comparing the areas of
the chromatogram peaks to the derivatization times; the
change with derivatization time was not very great, so the
derivatization time was set to 5 min for limiting the time of
the experiments.
Ratio of Hydrogen Bond Donor to Hydrogen Bond
Acceptor. Although DESs are prepared using the same
hydrogen bond donor and hydrogen bond acceptor, they
have different properties when the molar ratio of the donor
and acceptor is varied. Therefore, the extraction efficiency
was studied by changing the ratio of the hydrogen bond
donor (decanoic acid) and the hydrogen bond acceptor
(N4441 Cl) within the range 4:1 to 1:4, and the type and
volume of the derivatization catalyst were 5% (w/v) potas-
sium bicarbonate and 1.0 mL, respectively. The amount of
Synthesis of Deep Eutectic Solvent. To synthesize the opti-
mal DES for extracting polyphenols, optimization was per-
formed by comparing four HBA candidates. With decanoic
acid chosen as HBD, tributylmethylammonium chloride
(N4441 Cl), methyltrioctylammonium chloride (N8881 Cl),
methyltrioctylammonium
bromide
(N8881 Br),
and
tetraoctylammonium bromide (N8888 Br) were used as
Figure 11. Effect of salt amount on peak area of polyphenolstes.
Figure 13. Response surface plots.
Bull. Korean Chem. Soc. 2020
© 2020 Korean Chemical Society, Seoul & Wiley-VCH GmbH
www.bkcs.wiley-vch.de
5

BULLETIN OF THE
Article
Effective Sample Preparation of Polyphenols in Wine
KOREAN CHEMICAL SOCIETY
Table 3. Optimized DLLME conditions by using RSM.
Table 4. LODs and LOQs for polyphenols (μg/L).
A (mL)
B (μL)
C
D (g)
Compounds
LOD
LOQ
High
Optimized
Low
1.3
1.3
0.7
250
250
550
5.5
5.5
2.5
2.0
1.0
1.0
Resveratrol
Oxyresveratrol
Piceatannol
1.69
2.48
2.53
5.64
8.26
8.42
the derivatization reagent (acetic anhydride) was 400 μL,
the derivatization time was 5 min, the volume of the extrac-
tion solvent was 100 μL, the volume of the dispersion sol-
vent was 400 μL, and the amount of salt was 1.5 g. When
the molar ratio of the hydrogen bond donor and the hydro-
gen bond acceptor was 4:1, it exists as a solid at room tem-
perature and could not be used as an extraction solvent
(Figure 8). The highest extraction efficiency was obtained
when the molar ratio was 1:3.
The parameters used for selecting the optimal dispersion
solvent are as follows: the type and volume of the derivati-
zation catalyst were 5% (w/v) potassium bicarbonate and
1.0 mL, respectively, the amount of the derivatization
reagent was 400 μL, the derivatization time was 5 min, the
volume of the extraction solvent was 100 μL, the molar
ratio of the hydrogen bond donor and the hydrogen bond
acceptor was 3:1, the volume of the dispersion solvent was
400 μL, and the amount of salt was 1.5 g. The dispersion
solvent showing the optimum extraction efficiency was
identified from the area of a peak in the chromatogram, and
the use of methanol resulted in the largest peak area evenly
in three polyphenols (Figure 9).
Optimization of Dispersive Liquid–Liquid Micro-
extraction. After synthesis of the DES solvent suitable for
extracting polyphenols, the solvent was used in the disper-
sion liquid–liquid microextraction (DLLME) method. The
parameters for optimizing the extraction efficiency in the
method were (1) the type of dispersion solvent, (2) the vol-
ume ratio of the extraction solvent and the dispersion sol-
vent, and (3) the amount of salt. Therefore, experiments
were undertaken to optimize these parameters.
Ratio of Extraction and Dispersive Solvents. When the
volume of the dispersion solvent is small, the hydrophobic
extraction solvent cannot disperse adequately in the aque-
ous solution, and, when the volume is large, the analyte
may be more soluble in the aqueous solution than it is in
the extraction solvent. Therefore, it is important to
Type of Dispersive Solvent. In the DLLME, the dispersion
solvent must be miscible with both a hydrophilic sample
and a hydrophobic extraction solvent, and, when added to
an aqueous solution sample and an extraction solvent sys-
tem, the extraction solvent must be dispersible into fine
droplets. Extraction solvents dispersed in fine droplets can
increase extraction efficiency due to increased surface area.
It was confirmed that the DES prepared above mixed well
with methanol, ethanol, acetone, and acetonitrile, as well as
with aqueous solutions. Therefore, the extraction efficiency
was compared by selecting the various solvents as a disper-
sion solvent. If the polarity of the dispersion solvent is too
large or too small, it is difficult to disperse the extraction
solvent into an aqueous solution.
Figure 14. Comparison of OVAT and CCD.
Figure 15. Calibration curves for determination of polyphenols.
Bull. Korean Chem. Soc. 2020
© 2020 Korean Chemical Society, Seoul & Wiley-VCH GmbH
www.bkcs.wiley-vch.de
6

BULLETIN OF THE
Article
Effective Sample Preparation of Polyphenols in Wine
KOREAN CHEMICAL SOCIETY
Table 5. Working range, linear equation and r2.
Compounds
Working range(μg/L)
Linear equation
r2
Resveratrol
Oxyresveratrol
Piceatannol
5.6–56.4
8.3–82.6
8.4–84.2
y = 1.9356x + 10.895
y = 1.1889x + 3.6592
y = 1.0022x + 1.593
0.9947
0.9967
0.9952
determine the appropriate ratio of extraction solvent and
dispersion solvent. To determine this optimal ratio for
extraction solvent and dispersion solvent (methanol), ratios
of the two solvents were varied within the range 1:1 to
1:10, and the other parameters were set as in the experi-
ments described above. The best extraction efficiency
(maximum chromatographic peak area) was obtained when
the ratio of the extraction solvent to the dispersion solvent
was 1:4 (Figure 10).
the optimal extraction conditions by DLLME are as fol-
lows: the type and volume of the derivatization catalyst
were 5% (w/v) potassium bicarbonate and 1.0 mL, respec-
tively, the volume of derivatizing reagent (acetic anhydride)
was 400 μL, the derivatization time was 5 min, the volume
of the DES as an extraction solvent was 100 μL, the vol-
ume of the dispersion solvent was 400 μL, the molar ratio
of the hydrogen bond donor (decanoic acid) and the hydro-
gen bond acceptor (tributylmethylammonium) was 3:1, and
the amount of salt (sodium sulfate) was 1.0 g. Figure 12
shows the HPLC chromatogram analyzed after spiked with
three polyphenol standards at the level of LOQ concentra-
tion in 10% blank white wine and extracted by the
established method. Piceatannol, oxyresveratrol, and resver-
atrol were detected at 35.59, 35.99, and 37.17 min,
respectively.
Amount of Salt. The salting-out effect can reduce the solu-
bility of the analyte in the aqueous solution sample and
increase the solubility of the analyte in the extraction sol-
vent, and this is accomplished with a salt that is more solu-
ble in aqueous solution than is the analyte. However, the
addition of salt increases the ionic strength, so the extrac-
tion efficiency does not always increase. This is attributed
to changes in the physical properties of the Nernst diffusion
film, which reduce the rate at which the analyte diffuses
into the extraction solvent.28 To optimize the amount of salt
(sodium sulfate) used, extraction efficiencies were com-
pared by adding salt in amounts ranging from 0 to 2.0 g.
The type and volume of the derivatization catalyst were 5%
(w/v) potassium bicarbonate and 1.0 mL, respectively. The
volume of derivatizing reagent was 400 μL, the derivatiza-
tion time was 5 min, the volume of the extraction solvent
was 100 μL, the volume of the dispersion solvent was
400 μL, and the molar ratio of the hydrogen bond donor
and the hydrogen bond acceptor was 3:1. When the amount
of salt was less than 1.0 g, the extraction solvent did not
separate well from the aqueous solution sample, and, when
it was 1.0 g, the best extraction efficiency resulted
(as determined by the area of the peak in the chromato-
gram) (Figure 11).
Central Composite Design. The optimized experimental
factors obtained with the OVAT method are as follows: the
derivatization catalyst was 5% (w/v) potassium bicarbonate,
the volume of derivatization catalyst was 1.0 mL, the
amount of derivatization reagent was 400 μL, the derivati-
zation time was 5 min, the volume of the extraction solvent
was 100 μL, the volume of the dispersion solvent was
400 μL, the ratio of hydrogen bond donor to hydrogen
bond acceptor was 3:1, and the amount of salt was 1.0 g.
To overcome the problems resulting from experimental fac-
tors that cannot be considered in the OVAT method, the
optimization was performed again using the CCD method
and the optimization values obtained from the OVAT
method. The mutual effects of the four experimental factors
were obtained from six response surface curves
(Figure 13). As a result, the best result was obtained when
the volume of the derivatization catalyst was 1.3 mL, the
volume of the derivatization reagent was 250 μL, the vol-
ume ratio of the extraction solvent and the dispersion
As such, the optimal derivatization conditions for the
phenolic compounds present in the aqueous solution and
Table 6. Accuracy and precision of polyphenols (n = 3).
Compounds
Resveratrol
Concentration (μg/L)
Accuracy (%)
RSD (%)
5.6
31.0
56.4
8.3
45.4
82.6
8.4
98.9
108.0
99.1
5.5
4.6
2.8
6.7
4.3
2.4
6.1
3.9
1.3
Oxyresveratrol
Piceatannol
95.1
105.6
100.5
99.0
103.6
102.0
46.3
84.2
Bull. Korean Chem. Soc. 2020
© 2020 Korean Chemical Society, Seoul & Wiley-VCH GmbH
www.bkcs.wiley-vch.de
7

BULLETIN OF THE
Article
Effective Sample Preparation of Polyphenols in Wine
KOREAN CHEMICAL SOCIETY
Table 7. Intra- and inter-day accuracy and precision (n = 3).
Intra-day
Inter-day
Compds.
Conc (μg/L)
Accuracy (%)
RSD (%)
Accuracy (%)
RSD (%)
Resveratrol
5.6
31.0
56.4
8.3
45.4
82.6
8.4
95.9
101.3
91.2
86.1
94.2
87.3
80.5
91.1
89.8
2.3
4.2
3.2
4.1
4.3
3.3
5.8
5.0
3.2
104.5
107.2
96.6
97.8
103.5
96.7
90.1
100.5
98.8
4.0
4.3
3.5
4.7
4.9
3.7
3.7
4.4
4.1
Oxyresveratrol
Piceatannol
46.3
84.2
solvent was 1:5.5, and the amount of salt was 1.0 g
(Table 3). Therefore, somewhat better results were obtained
with optimization using the CCD method, as compared to
those obtained with optimizations using the OVAT method
(Figure 14).
5.64–8.42 μg/L, respectively (Table 4). The LOD and LOQ
in other literatures by HPLC-UV/Vis analysis was 0.4–950
and 1.l3–3160 ug/L,11,29–31 and our experimental results
were similar of better than these results.
Calibration Curves. The working range of the calibration
curve spanned analyte concentrations from the concentra-
tion limit to a concentration more than 10 times that of the
limit. The concentrations for the calibration curve ranged
from 5.6 μg/L, which is the LOQ concentration of the
analytes, to 84.2 μg/L, which is 15 times the LOQ concen-
tration (Figure 15). The correlation coefficient (r2) showed
good linearity and ranged from 0.9947 to 0.9967 (Table 5).
Method Validation. Using the optimized experimental
methods, detection limits, quantitation limits, calibration
curves, accuracy, precision, and so on were probed with
validation experiments described below.
LOD and LOQ: After assuming the limit of detection
(LOD), seven spiked aqueous samples were analyzed under
the optimized conditions and at an analyte concentration,
approximately three times that of the assumed LOD. The
standard deviation (σ) was calculated from the analytical
results for each sample. Another spiked sample, with
approximately twice the analyte level of the spiked sample,
was also analyzed, and calibration curves were plotted. The
slope (m) of each calibration curve was used to calculate
the LOD and LOQ. The LOD and LOQ values were calcu-
lated from 3 σ/m and 10 σ/m, respectively. The LOD was
required to have a S/N of 3 or higher, and the LOQ was
required to have a S/N of 3 or higher and a relative stan-
dard deviation (RSD) of less than 15%. If these conditions
were not satisfied, this experiment was performed with
increased analyte concentration. As a result, the LOD and
the LOQ exhibited ranges of 1.69–2.53 μg/L and
Accuracy and Precision. After spiking three polyphenol
standards for wine blank samples to establish a low concen-
tration, a medium concentration, and a high concentration
within the range of the calibration curve, accuracy, and pre-
cision determinations were performed for each concentra-
tion. The accuracy was expressed as a relative recovery and
ranged from 95.1 to 108.0%, showing good accuracy
(n = 3). The precision, which was expressed as a RSD, was
determined 3 times per sample and showed good reproduc-
ibility, with variations ranging from 1.3 to 6.7% (Table 6).
To measure the errors observed within a day, three experi-
ments for determining accuracy and precision were per-
formed at different times; the relative recovery ranged from
80.5 to 101.3% and the RSD from 2.3 to 5.8% (Table 7).
Accuracy and precision experiments were also carried out on
multiple days, and the relative recovery ranged from 90.1 to
107.2% and the RSD from 3.4 to 5.0% (Table 7).
Real Sample Monitoring. Analysis of polyphenols con-
tained in commercially available wines was performed
using the experimental methods established herein. The
experimental procedure for the real wine sample is as fol-
lows: 100 μL of extraction solvent (DES) and 550 μL of
dispersion solvent (methanol) were added rapidly to the
aqueous solution sample resulting after the derivatization of
the analyte with a base solvent (5% (w/v) potassium bicar-
bonate) as the catalyst and acetic anhydride as derivatiza-
tion reagent, and 1.0 g of sodium sulfate was added. The
solution was then gently shaken for 10 s. After centrifuga-
tion for 10 min at a speed of 4000 rpm, the layers were
Figure 16. Concentration of resveratrol in real wine sample: Num-
ber 1–7 wines from Chile, number 8–10 wines from Australia.
Bull. Korean Chem. Soc. 2020
© 2020 Korean Chemical Society, Seoul & Wiley-VCH GmbH
www.bkcs.wiley-vch.de
8

BULLETIN OF THE
Article
Effective Sample Preparation of Polyphenols in Wine
KOREAN CHEMICAL SOCIETY
6. J. I. Cacho, N. Campillo, P. Vinas, M. Hernandez-Cordoba,
J. Chromatogr. A 2013, 1315, 21.
7. L. Fremont, Life Sci. 2000, 66, 663.
8. B. Sun, A. M. Ribes, M. C. Leandro, A. P. Belchior, M. I.
Spranger, Anal. Chim. Acta 2006, 563, 382.
9. J. A. Baptista, J. F. da, P. Tavares, R. C. Carvalho, Food Res.
Int. 2001, 34, 345.
10. F. Y. Du, X. H. Xiao, G. K. Li, J. Chromatogr. A 2007,
1140, 56.
separated, and 20 μL of the extraction solvent comprising
the upper layer was removed and injected into the HPLC
instrument.
The 10 wines selected included nine red wines and one
white wine, and seven were from Chile, and three were
from Australia. Oxyresveratrol and piceatannol were not
detected, and resveratrol was detected in the range of
0.00–4000 mg/L in all wines. Due to the small number
of samples, no difference could be attributed to the country
of origin or wine type (white vs. red wine) (Figure 16).
11. J. Goncalves, J. S. Câmara, J. Sep. Sci. 2011, 34, 2376.
12. L. Mercolini, R. Mandrioli, M. A. Raggi, J. Pineal Res. 2012,
53, 21.
13. P. Vi~nas, C. López-Erroz, J. J. Marín-Hernández, M.
Hernández-Córdoba, J. Chromatogr. A 2000, 871, 85.
14. P. Vi~nas, N. Campillo, M. Hernández-Pérez, M. Hernández-
Córdoba, Anal. Chim. Acta 2008, 611, 119.
15. J. B. Jean-Denis, R. Pezet, R. Tabacchi, J. Chromatogr. A
2006, 1112, 263.
16. F. S. Mirnaghi, F. Mousavi, S. M. Rocha, J. Pawliszyn,
J. Chromatogr. A 2013, 1276, 12.
17. G. Sagratini, F. Maggi, G. Caprioli, G. Cristalli, M.
Ricciutelli, E. Torregiani, S. Vittori, Food Chem. 2012,
132, 1592.
18. R. Montes, M. Garcia-Lopez, I. Rodriguez, R. Cela, Anal.
Chim. Acta 2010, 673, 47.
19. O. Tokus, oglu, M.K. Unal, F. Yemis, J. Agric. Food Chem.
2005, 53, 5003.
20. P. Vi~nas, N. Martínez-Castillo, N. Campillo, M. Hernández-
Córdoba, J. Chromatogr. A 2011, 1218, 639.
21. E. Baltussen, P. Sandra, F. David, C. Cramers,
J. Microcolumn Sep. 1999, 11, 737.
22. A. Mouahid, C. Crampon, S.-A. A. Toudji, E. Badens,
J. Supercrit Fluids 2013, 77, 7.
Conclusion
In this study, an eco-friendly extraction solvent applicable
to DESME was synthesized, and the optimal conditions for
the simultaneous analysis of three polyphenols (resveratrol,
oxyresveratrol, piceatannol) contained in wine were
established. This method was applied to the analysis of real
samples. The LOD and LOQ for the three polyphenols
ranged from 1.69 to 2.53 μg/L and 5.6 to 8.4 μg/L, respec-
tively. The intra- and multi-day accuracies ranged from
80.5 to 101.3% and 90.1 to 107.2%, respectively, and the
precisions ranged from 2.3 to 5.8 RSD% and 3.3 to 4.9
RSD%, respectively. The correlation coefficient (r2)
exhibited linearity in the range 0.9947 to 0.9967.
In this study, a deep eutectic solvent (DES) was used as
an extraction solvent to establish a method that can effec-
tively analyze three types of polyphenols as an environ-
mentally friendly analysis method. In addition, in terms of
LOD and LOQ, results were almost similar or better than
those of other HPLC/UV–Vis analysis methods.
23. D. Zhao, Y. Liao, Z. Zhang, Clean 2007, 35, 42.
24. J. F. Brennecke, E. J. Maginn, AIChE Jo 2001, 47, 2384.
25. B. Tang, K. H. Row, Monatsh. Chem. 2013, 144, 1427.
26. E. L. Smith, A. P. Abbott, K. S. Ryder, Chem., Rev. 2014,
114, 11 060.
Acknowledgments. This work was supported by a Gradu-
ate Research Assistantship 2019 from Kyonggi University.
References
27. A. Abbott, G. Capper, D. Davies, R. Rasheed, V.
Tambyrajah, Chem Commun. 2003, 1, 70.
1. M. Fabjanowicz, J. P. Wasylka, J. Namiesnik, Trends Anal.
Chem. 2018, 103, 21.
28. E. Psillakis, N. Kalogerakis, Trends Anal. Chem. 2003,
22, 565.
2. T. R. Cabo, I. Rodriguez, R. Cela, J. Chromatogr. A 2012,
1258, 21.
29. C. Dominguez, D. Guillen, C. Barroso, J. Chromatogr. A
2001, 918, 303.
3. G. Arslan, N. Yilmaz, Asian J. Chem. 2013, 25, 1225.
4. J. Burns, T. Yokota, H. Ashihara, M. Lean, A. Crozier,
J. Agric. Food Chem. 2002, 50, 3337.
5. R. E. King, A. Joshua, A. Bomser, D. B. Min, Compr. Rev.
Food Sci. Food Saf. 2006, 5, 65.
30. A. Aresta, P. Cotugno, F. Massari, C. Zambonin, Food Anal.
Methods 2018, 11, 426.
31. C. Silva, J. Pereira, V. Wouter, C. Giro, J. Camara, Talanta
2011, 86, 82.
Bull. Korean Chem. Soc. 2020
© 2020 Korean Chemical Society, Seoul & Wiley-VCH GmbH
www.bkcs.wiley-vch.de
9