Journal of Chromatography A, 1388 (2015) 119–125

Contents lists available at ScienceDirect

Journal of Chromatography A
journal homepage: www.elsevier.com/locate/chroma

Isolation of ␤-carotene, ␣-carotene and lutein from carrots by

countercurrent chromatography with the solvent system
modifier benzotrifluoride夽
Michael Englert, Simon Hammann, Walter Vetter ∗
Institute of Food Chemistry (170b), University of Hohenheim, Garbenstr. 28, D-70599 Stuttgart, Germany

a r t i c l e i n f o a b s t r a c t

Article history: A carotenoid purification method with dual-mode countercurrent chromatography (CCC) for ␤-carotene,
Received 31 October 2014 ␣-carotene and lutein from a fresh carrot extract was developed. The fluorinated liquid benzotrifluo-
Received in revised form 6 February 2015 ride (IUPAC name: (trifluoromethyl)benzene) was used as a novel modifier in the non-aqueous ternary
Accepted 8 February 2015
solvent system n-hexane/benzotrifluoride/acetonitrile. The ternary phase diagram of the type I solvent
Available online 16 February 2015
system was used to select two-phase solvent mixtures which enabled an efficient preparative separa-
tion of ␣-carotene, ␤-carotene and lutein from concomitant pigments in crude carrot extract. By means
Keywords:
of the modifier, high separation factors (˛ ≥ 1.2) were obtained, allowing baseline resolution between
Carotenoids
Countercurrent chromatography
␣-carotene and ␤-carotene due to specific chemical interactions such as ␲–␲ molecular interactions.
Carrots After optimizing the injection step with a pseudo-ternary phase diagram, 51 mg of ␤-carotene, 32 mg
Benzotrifluoride of ␣-carotene and 4 mg of lutein could be isolated from 100.2 mg crude carrot extract in a short time
(trifluoromethyl)benzene and with high purities of 95% and 99% by using dual-mode CCC, respectively. Temperatures >22 ◦ C had a
negative impact on the separation of ␣-carotene and ␤-carotene.
© 2015 Elsevier B.V. All rights reserved.

1. Introduction
Carotenoids are naturally occurring pigments that are either
polyprenoid hydrocarbons (carotenes) or derivatives thereof, con-
taining oxygen, for example, in the form of hydroxyl, methoxyl
or keto groups (xanthophylls) [1]. They are biosynthesized pre-
dominantly in all photosynthetic plants including fruits, vegetables
and to some extent in microorganisms and in a few selected
invertebrates [2].
Numerous reports on human health benefits linked to high
dietary intake and consumption of fruits and vegetables which
contain natural pigments such as carotenoids are available [3].
Carotenoids have a broad range of functions and a significant
impact on the prevention of several degenerative and chronic
human health issues including coronary heart disease, oxidative
tissue damage, macular degeneration and cancer [4]. Also, they act
as antioxidants preventing lipid oxidation by scavenging reactive
夽 Presented at the 8th International Conference on Countercurrent
Chromatography—CCC 2014, 23–25 July 2014, Uxbridge, United Kingdom.
∗ Corresponding author. Tel.: +49 711 459 24016; fax: +49 711 2459 4377.
E-mail address: walter.vetter@uni-hohenheim.de (W. Vetter).
oxygen species and are responsible for the photoprotection of light
exposed tissues and cell layers [5].
Accordingly, there is a steadily growing demand for carotenoids
in pharmaceutical uses for clinical animal or human studies of
immune response and cancer. Hitherto mostly crude carotenoid
extracts were utilized and tested for this purpose because there
is a lack of individual and sufficiently specified compounds which
are, furthermore, very expensive [6]. This also applies for authen-
tic reference standards for the analysis and quantification of food
samples [7].
Suitable methods for the purification of carotenoids include
preparative thin-layer chromatography (TLC) and normal- and
reverse-phase high-performance liquid chromatography (HPLC)
after pre-concentration of crude extracts with solid-phase extrac-
tion [8–11]. However, most procedures remain laborious and bear
the risk of isomerization and sample degradation caused by the
sensitivity of carotenoids to light, oxygen, high temperatures and
chemicals [11,12].
Countercurrent chromatography (CCC) is representing a
liquid–liquid separation technique benefiting from several advan-
tages in comparison to TLC or HPLC for the preparative isolation
of natural products [13]. Irreversible adsorption or denaturation
of sample material on solid supports is avoided and crude mix-
tures can be separated without the risk of plugged inlet frits or

http://dx.doi.org/10.1016/j.chroma.2015.02.020
0021-9673/© 2015 Elsevier B.V. All rights reserved.
120 M. Englert et al. / J. Chromatogr. A 1388 (2015) 119–125
a)
b)
OH
c)
HO
Fig. 1. Chemical structures of (a) ␤-carotene, (b) ␣-carotene and (c) lutein.
plugged packings. Furthermore, CCC does not require expensive
stationary phases and has a generally low solvent consumption
combined with a much larger sample capacity which speaks in favor
of applying this technique [14].
Different CCC methods were already used for the purification
of selected xanthophylls, e.g. lutein [7,15,16], zeaxanthin [17],
violaxanthin [7] and neoxanthin [7,18] as well as water-soluble
carotenoids such as crocin from a fruit extract [19]. However, only
very few efforts were made to purify carotenes (e.g. lycopene)
mainly because of their stronger lipophilic nature and the lack of
established and suitable non-polar/non-aqueous solvent systems
for CCC [20,21].
The objective of our study was to develop a purification method
which is suitable to separate structurally similar ␣- from ␤-
carotene (Fig. 1) in a crude carrot extract. Both isomers differ only in
the position of one double bond in the cyclic ionone end group. For
this purpose we examined the applicability of the various solvent
systems already described for the purification of carotenoids and
report the first utilization of benzotrifluoride (IUPAC name: (tri-
fluoromethyl)benzene) in CCC as a modifier in the solvent system
n-hexane/acetonitrile. By means of dual-mode CCC, consisting of
switching the role of the mobile and the stationary phase after a
specific run time of the separation, the highly retained lutein could
also be obtained in short time and with high purity.
2. Experimental
2.1. Sample and chemicals
Fresh carrots were purchased from a local supermarket in
Stuttgart, Germany. They served as a cheap and easily available
natural source of ␣- and ␤-carotene as well as lutein. Acetonitrile,
n-hexane, methanol (all HPLC gradient grade) and dichloromethane
(purity >99.8%) were from Th. Geyer (Renningen, Germany). Anhy-
drous benzotrifluoride (␣,␣,␣-trifluorotoluene, purity >99%) was
from Sigma–Aldrich (Steinheim, Germany). Technical grade ace-
tone (BASF, Ludwigshafen, Germany) and tert-butylmethylether
((MTBE), Fisher Scientific, Leicestershire, Great Britain) were dis-
tilled before use. Butylated hydroxytoluene (BHT) and butylated
hydroxyanisole (BHA) were both from Fluka (Steinheim, Germany).
Anhydrous sodium sulfate (Na2 SO4 ) was from Carl Roth (Karlsruhe,
Germany).
2.2. Sample preparation
Amber glass was used in all preparation steps to avoid sample
degradation or isomerization of carotenoids. About 1 kg of fresh car-
rots were washed with water, diced and homogenized with an IKA
A11 basic analytical mill (IKA, Staufen, Germany). After grinding,
the sample was divided into two portions of 500 g and carotenoids
were extracted from each portion of the pureed carrots in a 2 L
flask at room temperature under stirring with a mixture of 1000 mL
n-hexane and 1000 mL acetone containing 100 mg/L of BHT and
BHA to prevent oxidation. After filtration (GE Whatman analytical
filter, 40 grades) of the solutions, solvents were evaporated to dry-
ness at 25 ◦ C and each residue was dissolved in 200 mL n-hexane.
After recombination of the two portions the solution was washed
three times with 800 mL of deionized water in a 1 L shake-flask
to remove accompanying polar sugars and proteins in the extract.
The n-hexane extract was separated and dried over anhydrous
Na2 SO4 to remove contaminating water followed by evaporation
of the organic solvent under vacuum yielding 505 mg of crude car-
rot carotenoid extract. The dried residues were stored in amber
glass flasks under nitrogen atmosphere in a refrigerator at 4 ◦ C until
required. Portions thereof were used for shake-flask tests and CCC
separations.
2.3. Determination of ternary phase diagram
The ternary phase diagram for n-hexane/benzotrifluoride/
acetonitrile was determined by adding specific volumes of each
of the three solvents to a 20 mL glass-stoppered graduated cylin-
der followed by vigorously shaking and equilibration for 15 min at
22 ◦ C in a water bath. Visual examination of a meniscus allowed the
determination of the biphasic solution region of the solvent sys-
tem and the measurement of the volume ratio of upper to lower
phase. Exact phase compositions of the utilized solvent system
were determined by split-injection gas chromatography (GC) cou-
pled with a flame ionization detector (FID) as described elsewhere
[22].
2.4. CCC apparatus
A CCC-1000 instrument (Pharma-Tech Research, Baltimore,
MD, USA) was employed for the separations. The apparatus
holds three serially connected semi-preparative multilayer-coils
equipped with PTFE tubing (1.6 mm bore diameter) mounted sym-
metrically on a rotating frame with a total column capacity Vc of
approximately 325 mL. The ˇ-value (ˇ = r/R where r is the distance
between the coil holder shaft and the coil axis and R is the dis-
tance between central axis of the centrifuge and the coil holder
shaft) varied from 0.50 (internal terminal) to 0.75 (external termi-
nal). Solvents were delivered with a 655A-12 HPLC Pump (Merck,
Darmstadt, Germany) in constant flow mode.
The effluent was continuously monitored with an 87 UV–vis
spectral photometric detector (Knauer, Berlin, Germany) oper-
ated at 450 nm. A DI-160 data logger (Dataq, Akron, OH, USA)
was employed for data acquisition. A type 50 Teflon rotary valve
(Rheodyne, Cotati, CA, USA) equipped with a 10 mL sample loop
was used for manual sample injection into the column. For
sample collection an Isco Retriever 500 (Teledyne Isco, Lincoln,
NE, USA) with 20 mL amber glass tubes was used and frac-
tionation was carried out in accordance with the online UV–vis
chromatogram.
2.5. Preparation of solvent system and sample solution
A two-phase solvent system with n-hexane/benzotrifluoride/
acetonitrile (10:3.5:6.5, v/v/v) was prepared by adding 800 mL
n-hexane, 280 mL benzotrifluoride and 520 mL acetonitrile in a
2 L separation funnel followed by vigorously shaking for 2 min
and equilibration for 24 h. Before performing CCC the upper and
lower phases were separated and degassed for 10 min in an ultra-
sonic bath. Sample solutions for CCC were prepared by dissolving
100.2 mg of the crude carrot extract in 10 mL of a mixture of upper
and lower phase (1:1, v/v).
 M. Englert et al. / J. Chromatogr. A 1388 (2015) 119–125
2.6. Separation procedure
Dual mode CCC separation was performed starting at ambient
laboratory temperature (22 ◦ C). Tail-to-head mode was selected
and the separation column was entirely filled with the lower phase
which initially represented the stationary phase at a flow rate
of 10 mL/min. After adjustment of the flow rate to 2.0 mL/min
the rotation of the centrifuge was started and the rotation speed
was increased to 1010 ± 10 rpm. Then the upper phase (initial
mobile phase) was pumped into the apparatus with a flow rate
of 2.0 mL/min and the effluent was collected in a 100 mL graduated
cylinder. When the mobile phase emerged at the outlet (indicated
by visual examination of a meniscus and a clear formation of two
phases) the displaced volume of stationary phase was measured.
After these preparatory measures, the sample injection was car-
ried out by introducing the sample solution into the sample loop
through the injection valve followed by turning the injection switch
to the inject position. The separation was conducted for 140 min
(↔ 280 mL upper phase) until complete elution of the target com-
pound ␣-carotene according to the UV–vis chromatogram. In order
to reverse column inlet and outlet, head-to-tail mode was selected
at the corresponding rheodyne valve of the apparatus. Then the
lower phase was pumped into the head inlet for 60 min with a flow-
rate of 4 mL/min (↔ 240 mL lower phase) until elution of the target
compound lutein was achieved. A scheme of the dual-mode CCC
separation and the behavior of the compounds can be found in the
supplementary information (Fig. S1.).
Fractions of ␤-carotene were collected from 95 to 105 min (1
fraction, 20 mL), fractions of ␣-carotene were collected from 115
to 130 min at 2.5 min/tube (6 fractions, 5 mL each) and fractions
of lutein were collected from 160 to 172 min at 2.0 min/tube (6
fractions, 8 mL each). Fractions collected where brought to dry-
ness by an AVC 2-33 IR speed-vac concentrator (Christ, Osterode,
Germany) set at 25 ◦ C and 10 mbar and the sample weight was
determined gravimetrically. Afterwards samples were re-dissolved
in MTBE/MeOH/H2 O (81:15:4, v/v/v).
2.7. Shake-flask-tests
In each case about 1–2 mg of the crude carrot extract were trans-
ferred into 4 mL amber glass vials, followed by the addition of 2 mL
of each upper and lower phase of the equilibrated solvent systems.
The vials were shaken for 30 s and equilibrated in a water bath at
temperatures of 10, 15, 20, 22, 25 and 30 ◦ C for 30 min, respectively.
After equilibrium was reached, 100 ␮L of each phase was separated
into 1.5 mL amber glass vials. Solvents were evaporated to dryness
under a gentle stream of nitrogen at room temperature and the
residues were re-dissolved in 1 mL of MeOH/MTBE/H2 O (81:15:4,
v/v/v).
Then the upper and lower phases were separately analyzed by
HPLC (Section 2.8). Peak areas obtained by UV–vis detection served
for the calculation of the partition coefficient K of ␣- and ␤-carotene
as well as lutein. The separation factor ˛ between ␣-carotene and
␤-carotene was determined according to Eq. (1):
K(␤-Carotene)
˛= (1)
K(␣-Carotene)
Systems were also tested for short settling times (ts < 15–20 s)
and adequate retention of the stationary phase Sf .
2.8. Analysis of extract and CCC fractions by HPLC
The crude carrot extract and collected fractions of dual-mode
CCC separation were analyzed by HPLC-UV–vis with a ProStar
system (Varian, Darmstadt, Germany) consisting of a model 230
pump, a model 325 UV–vis detector (single wavelength mode
121
III
150
absorbance (450 nm)
II
100
[mV]
I
50
0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
time [min]
Fig. 2. HPLC/UV–vis chromatogram (450 nm) of crude carrot extract showing
the three principal compounds (I) lutein, (II) ␣-carotene and (III) ␤-carotene.
Conditions: polymeric C30 column (250 × 4.6 mm i.d., 5 ␮m particle size); linear gra-
dient from 100% solvent A: MeOH/MTBE/H2 O (81:15:4, v/v/v) to 100% solvent B:
MeOH/MTBE/H2 O (6:90:4, v/v/v) in 40 min; flow-rate: 1.0 mL/min.
using 450 nm) and a Dynamax automatic sample injector Model
AI-A1, equipped with a 100 ␮L sample loop. Samples were injected
in partial loop fill mode with 10 ␮L injection volumes and sam-
ple concentration ranges of 0.02–0.30 mg/mL. Chromatographic
separation of carotenoids was achieved using a 250 × 4.6 mm i.d.
polymeric C30 column with 5 ␮m particle size (YMC, Wilmington,
MA, USA) equipped with a guard column 10 × 4.6 mm i.d. of the
same stationary phase. Separations for the general detection of
carotenes and xanthophylls were performed at ambient labora-
tory temperature with 1.0 mL/min flow rate using a linear gradient
from 100% A to 100% B in 40 min with solvent A: MeOH/MTBE/H2 O
(81:15:4, v/v/v) and solvent B: MeOH/MTBE/H2 O (6:90:4, v/v/v).
2.9. Determination of a solubility isotherm
Mixtures of stationary phase and mobile phase with eight
pre-defined ratios (w/w) of the selected solvent system n-
hexane/benzotrifluoride/acetonitrile (10:3.5:6.5, v/v/v) were
added to 10 mg of the crude carrot extract in a glass-stoppered
measuring cylinder and weighed out by means of an analytical
balance at 22 ◦ C in order to determine the mass quantities. Specific
masses of stationary phase or mobile phase were successively
added until the distinct layers of a two-phase system appeared.
Afterwards a solubility isotherm or pseudo-ternary-phase diagram
with stationary phase, mobile phase and extract (w/w/w) was
plotted in orthogonal form [23].
3. Results and discussion
3.1. HPLC/UV–vis analysis of crude carrot extract
The HPLC/UV–vis chromatogram of the crude carrot extract
showed peaks of the three principal compounds (I) lutein, (II) ␣-
carotene and (III) ␤-carotene, with a different signal intensity at
450 nm (Fig. 2). Additionally, several smaller side peaks compris-
ing less than 10% of the total carotenoid content were detected but
their identification was not pursued. They presumably were caused
by various hydrocarbon carotenoids, including chlorophyll degra-
dation products, such as phytofluene, phytoene and neurosporene
which are regularly found in carrot extracts [24].
3.2. Selection of the solvent system
Because of the structural similarity between ␣- and ␤-carotene
(Fig. 1) and their lipophilic character, only non-aqueous, two-
phase solvent systems composed mainly of n-hexane, acetonitrile
and dichloromethane as modifier produced an acceptable partition
coefficient range (0.4 < K < 2.5), but, with non-ideal separation fac-
tor values (˛ < 1.2) between ␣- and ␤-carotene (data not shown).
Other lipophilic solvent systems such as n-hexane/ethanol/water,
122 M. Englert et al. / J. Chromatogr. A 1388 (2015) 119–125
Fig. 3. (a) Chemical structure of benzotrifluoride and (b) ternary phase diagram of n-
hexane/benzotrifluoride/acetonitrile with two-phase region at 22 ◦ C shown in grey
shading. Volumetric solvent composition for the CCC separation is represented as a
red triangle (50% n-hexane, 17.5% benzotrifluoride, 32.5% acetonitrile, 10:3.5:6.5).
The tie-lines indicate that benzotrifluoride partitioned evenly between n-hexane
and acetonitrile. The composition of the upper and lower phases varied extensively
with the amount of benzotrifluoride in the system. (For interpretation of the color
information in this figure legend, the reader is referred to the web version of the
article.)
n-hexane/acetonitrile or n-hexane/chloroform/acetonitrile at vari-
ous volume ratios produced a non-ideal partition coefficient range
for ␣- and ␤-carotene and a low separation factor ˛ between them.
Consequently, the search for a new solvent system modifier acting
as a bridge solvent for the n-hexane/acetonitrile based system was
initiated. It was taken into account that the modifier should interact
with the conjugated system of the carotenoids to gain selectivity
between them and that it must also distribute between n-hexane
and acetonitrile in a manner that the lower phase also exhibits non-
polar characteristics, in order to achieve partition coefficients in
the ideal range. In addition, the sample should be easily dissolved
in the two-phase solvent system and the modifier should be more
environmentally friendly and less harmful than dichloromethane.
This led us to study benzotrifluoride (Fig. 3a), which is rel-
atively unknown as a solvent in general and which was not
previously utilized in CCC. Benzotrifluoride possesses physico-
chemical properties which makes it attractive for CCC as a modifier
for non-aqueous and non-polar solvent systems [25]. These prop-
erties include (i) high density (1.19 g/mL at 22 ◦ C), (ii) moderate
boiling point (102 ◦ C), (iii) chemical stability, (iv) compatibility with
PTFE material, (v) a short atmospheric lifetime and (vi) the capabil-
ity of dissolving a wide range of organic compounds [26]. Due to the
trifluoromethyl group benzotrifluoride is more polar than benzene
or toluene and is intermediate between dichloromethane and chlo-
roform. Accordingly, it partitions more readily into the lower, polar
phase containing acetonitrile, thus, allowing it to bridge the gap
between both phases [26]. Because benzotrifluoride has a low price
(65D /L, February 2015) is relatively non-toxic and offers a low envi-
ronmental impact, usage in industrial processes and large-scale
applications is also conceivable, ideally with the waste streams of
benzotrifluoride being reclaimed by distillation.
For the closer characterization of the system n-
hexane/benzotrifluoride/acetonitrile, the ternary phase diagram
was obtained at 22 ◦ C (Fig. 3b) under normal ambient pressure. It
represents a type I two-phase solvent system since n-hexane and
acetonitrile are both immiscible while benzotrifluoride is miscible
with both n-hexane and acetonitrile.
Table 1
Partitioning coefficients K of ␣-carotene, ␤-carotene and lutein in different volu-
metric mixtures of the two-phase solvent system n-hexane/benzotrifluoride/
acetonitrile determined by HPLC/UV–vis (mean values after duplicate
determination).
Solvent Volume ratio K (␣-carotene) K (␤-carotene) K (lutein)
system (v/v/v)
A 10:3.5:6.5 1.46 2.04 0.17
B 14:5.5:6.5 1.34 1.61 0.47
C 10:4.5:6.5 1.40 1.58 0.70
The volumetric percentage of each solvent in a specific com-
position of the solvent system inside the triangular area can be
obtained by the grid which represents a 10% incremental increase
in each solvent. The biphasic region (Fig. 3b, grey background),
defined by the binodal curve is approximately one-fourth of the
total triangular area and it indicates that up to about 50% of ben-
zotrifluoride can be added to the binary two-phase solvent system
n-hexane/acetonitrile before the system becomes monophasic. By
determining the exact phase compositions of five different vol-
umetric solvent mixtures, tie-lines could be drawn, which were
useful to understand how various compositions changed the prop-
erties of the system. The tie-lines indicated that (i) benzotrifluoride
distributed evenly in n-hexane and acetonitrile and (ii) that the
composition of the upper and lower phases varied extensively with
the amount of benzotrifluoride in the system. Since the tie-lines are
aligned almost parallel to each other, the use of gradient elution
mode CCC may be excluded.
In order to obtain partition coefficients K for ␣-carotene and ␤-
carotene in a reasonable range, three different volumetric mixtures
of the biphasic region of the solvent system were tested in shake-
flask tests (Table 1). The settling time for all systems was below 20 s
which is a suitable range for CCC, generally resulting in acceptable
Sf values. All three solvent systems provided acceptable K values for
␣-carotene and ␤-carotene, ranging from 1.34 to 2.04, with solvent
system A showing the highest separation factor (˛ = 1.40) between
␣-carotene (K = 1.46) and ␤-carotene (K = 2.04). However, for a reg-
ular CCC separation, the K value of lutein (0.17) in solvent system A
was not in an ideal range, in comparison to systems B and C, which
produced K values of 0.47 and 0.70, respectively.
The selected solvent system A for the CCC separation n-
hexane/benzotrifluoride/acetonitrile (10:3.5:6.5, v/v/v) (Fig. 3b,
shown as red triangle) had a nearly equal volume ratio of 51% lower
phase to 49% upper phase. The exact volumetric phase composi-
tion of the equilibrated solvent system, as determined by GC, was
56.7% acetonitrile, 29.0% n-hexane and 14.3% benzotrifluoride for
the lower phase, and 76.0% n-hexane, 12.5% acetonitrile, and 11.6%
benzotrifluoride for the upper phase. These data indicate that direct
production of upper and lower phase of the solvent system without
equilibration can be achieved, thus reducing solvent consumption
and time. Waste streams can be reclaimed by distillation and after
GC analysis the required amounts of each solvent can be added to
obtain equilibrated solvent mixtures of upper or lower phase.
3.3. Preliminary optimization
The retention of the stationary phase Sf (86%) of the
new solvent system A (n-hexane/benzotrifluoride/acetonitrile,
10:3.5:6.5, v/v/v) was similar to that of the robust system n-
hexane/acetonitrile (Sf = 85%), which was also tested with the same
conditions (tail-to-head mode, rotation speed of 1010 ± 10 rpm,
flow rate of mobile phase 0.5 mL/min). A linear relationship
between Sf and the mobile phase flow rate could be observed
for both solvent systems (Fig. 4). However, if the flow rate was
increased, it could be observed by means of the respective slopes of
the straight lines that the system containing benzotrifluoride was
 M. Englert et al. / J. Chromatogr. A 1388 (2015) 119–125 123

90 tail-to-head (140 min) head-to-tail (60 min)

stationary phase retention SF (%)

85
500
80
y = -3.15x + 87.4 β-carotene

absorbance (450 nm) [mV]

75
70 lutein
y = -4.33x + 81.2
65 α-carotene
60 n-hexane/BTF/ACN (10:3.5:6.5, v/v/v)
55 n-hexane/ACN (1:1, v/v)
50
0 1 2 3 4 5
mobile phase flow rate (mL/min)
0 20 40 60 80 100 120 140 160 180 200
Fig. 4. Retention of the stationary phase Sf of the solvent systems n-
hexane/acetonitrile (1:1, v/v) and n-hexane/benzotrifluoride/acetonitrile time [min]
(10:3.5:6.5, v/v/v) in dependence of the mobile phase flow rate. The rotation
speed was set at 1010 ± 10 rpm and tail-to-head mode was selected. Fig. 6. CCC-UV–vis chromatogram (450 nm) of the separation of crude carrot extract
with dual-mode utilization (mobile phase and elution mode reversal). Conditions:
CCC-1000 with three serially connected multilayer coils (1.6 mm i.d. PTFE tubing)
and a total column capacity of 325 mL and 10 mL sample loop; rotation speed:
less stable at higher flow rates due to the more negative slope of
1010 ± 10 rpm; solvent system: n-hexane/benzotrifluoride/acetonitrile (10:3.5:6.5,
−4.33 in comparison to −3.15 for the pure n-hexane/acetonitrile v/v/v); sample size: 100.2 mg; separation: dual mode CCC separation with tail-to-
system. Within this application, a flow rate of 2 mL/min during head mode for 140 min at a flow rate of 2 mL/min and head-to-tail elution for 60 min
the CCC separation was considered as a compromise between at a flow rate of 4 mL/min.
separation time, stationary phase retention Sf and the resulting
chromatographic resolution between ␣-carotene and ␤-carotene.
3.4. CCC separation
A solubility isotherm or so-called pseudo-ternary phase dia-
gram of the crude carrot extract was determined in the selected
The separation of the crude carotenoid extract was conducted
solvent system because in preliminary tests the solubility of the
in dual-mode CCC, beginning at 22 ◦ C in tail-to-head elution for
carrot extract was smaller than assumed. Eight mass quantities
140 min with 2 mL/min after the sample injection of 100.2 mg,
of stationary phase, mobile phase and extract along the binodal
under the optimized conditions (Fig. 6). The retention of the station-
curve are presented in orthogonal form (Fig. 5). The reason for this
ary phase Sf was 72% at the beginning of the separation which was
investigation was to optimize the injection step and the possible
∼20% lower than during the preliminary investigations (Sf = 86%).
mass load of the extract under the condition of an equilibrated
The elution order in CCC was the reverse of the separation on the
two-phase solvent system. The diagram was nearly symmetrical
polymeric C30 HPLC column with ␤-carotene eluting at 115 min
with a theoretical maximum sample load of only 2.1% carrot extract
and ␣-carotene at 125 min, suggesting the elution was by order
with 50% mobile phase and 47.9% stationary phase. It was therefore
of decreasing polarity of the carotenoids. Baseline resolution was
concluded that the carrot extract was not readily soluble in high
achieved between structurally similar ␣-carotene and ␤-carotene.
quantities and that it should be dissolved in optimal mass amounts
Fractions of ␤-carotene could be collected between 95 and 105 min
of mobile and stationary phase to prevent any precipitation inside
and fractions of ␣-carotene could be collected between 115 and
the CCC column. The densities of the upper ( = 0.71 g/cm3 ) and
130 min. Subsequent HPLC/UV–vis analysis of the combined frac-
lower phase ( = 0.75 g/cm3 ) of the selected solvent system were
tions showed that 51 mg of ␤-carotene and 32 mg of ␣-carotene
used to approximate the amount of upper (5.8 mL) and lower phase
were collected both with a purity of 95% (Fig. 7a,b). In tail-to-head
(4.2 mL) and the sample load (∼100 mg) necessary for 10 mL of the
mode, stationary phase bleeding was observed during the elution
optimized sample solution.
of the main compounds ␣-carotene and ␤-carotene but no contam-
ination of the collected fractions was observed. Apart from this, a
5 continuous, slight stationary phase bleeding occurred due to the
increase in the apparatus temperature after a separation time of
extract mass load maximum 60 min. This was mainly attributable to the different miscibility
4
of the solvents in dependence of the temperature (Section 3.5) in
the selected non-aqueous solvent system, leading to an unbalanced
extract (%w)

3 state in both phases.

To reduce the overall separation time and solvent consumption,
2 as well as to improve the productivity, dual mode was selected. It
provides a suitable method for the separation of compounds that
1 have a wide range of polarity [27]. The role of mobile and station-
ary phase and the elution mode were reversed (head-to-tail) after
140 min and the flow rate was increased to 4 mL/min. In this case
0
lutein with its high affinity for the lower phase of the two-phase sol-
0 10 20 30 40 50 60 70 80 90 100
mobile phase (%w) vent system could be eluted 20 min after the reversal (Fig. 6). The
separation time until the elution of lutein would have been con-
Fig. 5. Solubility isotherm or pseudo-ternary phase diagram of the carrot siderably longer in the range of hours if the elution mode would
extract and mobile phase (%, w/w) in the selected solvent system n- have not been changed. The fractions collected between 160 and
hexane/benzotrifluoride/acetonitrile (10:3.5:6.5, v/v/v) at 22 ◦ C. The carrot extract
170 min, which had a combined weight of 4 mg (the last fraction
mass load maximum is represented as a red triangle at the extremum of the binodal
curve. (For interpretation of the color information in this figure legend, the reader from 170 to 172 min was discarded because it contained consider-
is referred to the web version of the article.) able amounts of impurities), were analyzed by HPLC/UV–vis, with
124 M. Englert et al. / J. Chromatogr. A 1388 (2015) 119–125

a) III
Table 2
150 51 mg Partitioning coefficients K of ␣-carotene and ␤-carotene and separation fac-
absorbance (450 nm)

tor ˛ between them at different temperatures with the two-phase solvent

purity ≥ 95%
system n-hexane/benzotrifluoride/acetonitrile (10:3.5:6.5, v/v/v) determined by
100 HPLC/UV–vis (mean values after duplicate determination).
[mV]

Temperature K (␣-carotene) K (␤-carotene) ˛

50 10 1.27 1.89 1.49
15 1.33 1.95 1.47
20 1.42 2.01 1.42
0 22 1.46 2.04 1.40
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
b) 150 time [min] 25 1.87 2.12 1.13
II 30 2.01 2.16 1.07
32 mg
absorbance (450 nm)

purity ≥ 95%
100
[mV]

1.50

50

selectivity factor α between

α-carotene and β-carotene
1.40

0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 1.30
c) 150 I time [min]
4 mg
absorbance (450 nm)

purity ≥ 97% 1.20

100
[mV]

1.10
50

1.00
0 10 15 20 25 30
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
time [min] temperature [°C]

Fig. 7. HPLC/UV–vis chromatograms (450 nm) of collected fractions of the CCC sep-
aration of the crude carrot extract with a) (III) ␤-carotene b) (II) ␣-carotene and c) (I)
lutein. Conditions: polymeric C30 column (250 × 4.6 mm i.d., 5 ␮m particle size); lin-
ear gradient from 100% solvent A: MeOH/MTBE/H2 O (81:15:4, v/v/v) to 100% solvent
B: MeOH/MTBE/H2 O (6:90:4, v/v/v) in 40 min; flow-rate: 1.0 mL/min.
Fig. 8. Influence of temperature in the range from 10 to 30 ◦ C on the separation
factor ˛ between ␣-carotene and ␤-carotene with the critical value for baseline
separation of ˛ = 1.2 in CCC drawn as a horizontal line.

the results showing a lutein purity of over 97% (Fig. 7c). The sep-
aration performance is summarized in Table S-1 (supplementary
data).
3.5. Influence of temperature on reproducibility
The influence of a temperature change in the solvent system was
investigated in shake-flask-tests since this can be a critical issue
in non-aqueous solvent systems. The mutual solubility of organic
solvents generally increases with temperature while the viscosity
decreases. Thus, the stationary phase retention Sf and the resolution
power can be influenced by temperature changes [18,28].
As the CCC apparatus used in this work was not equipped
with a temperature regulation unit it was not possible to obtain
reproducible results at room temperatures above 22 ◦ C. Extended
separation times above 1 h generated too much frictional heat
inside the apparatus, due to the mechanical rotation of the gears
and rotating tubing connections. In this case, a fast elution of ␣-
and ␤-carotene was observed immediately after the solvent front
emerged. Both compounds could not be baseline separated and the
peaks overlapped.
The shake-flask-tests verified that the partition coefficients of
the compounds, and furthermore, the separation factor ˛ between
␣- and ␤-carotene of the selected solvent system were affected by
temperature changes in the range from 10 to 30 ◦ C (Table 2, Fig. 8).
No significant change of the separation factor ˛ could be observed
between 10 ◦ C (˛ = 1.49) and 22 ◦ C (˛ = 1.40). However, when the
temperature increased to 25 and 30 ◦ C, ˛ significantly dropped to
1.13 and 1.07, respectively. At these elevated temperatures a base-
line separation could not be expected, as this requires generally ˛
>1.2 in CCC [22].
4. Conclusions
Benzotrifluoride, as a solvent system modifier, was introduced
for the n-hexane/acetonitrile system in CCC, resulting in a novel
type I solvent system with unique and mainly satisfying features.
In particular, the system can be useful for the isocratic separa-
tion of lipophilic compounds such as carotenoids as it provides
partition coefficients in a suitable range for CCC. Under optimized
conditions even a separation between the structurally similar ␣-
carotene and ␤-carotene in carrot extract was achieved through
specific chemical interactions such as ␲–␲ molecular interactions
between the compounds and the modifier benzotrifluoride. Draw-
backs of the novel n-hexane/benzotrifluoride/acetonitrile system in
the presented method were the rather low solubility of the crude
carrot extract and the influence of temperature on the separation
factor ˛ between ␣-carotene and ␤-carotene.
However, it is already apparent that using benzotrifluoride as a
modifier in CCC offers possibilities to expand the usage of lipophilic
two-phase solvent systems with suitable partition coefficients even
for very non-polar compounds. Further investigation into the field
of applications and in large scale applications are worth consider-
ation.
Conflict of interest
The authors declare no conflict of interest.
Acknowledgement
We are grateful to the state of Baden-Württemberg for providing
a stipend grant to Michael Englert.
 M. Englert et al. / J. Chromatogr. A 1388 (2015) 119–125
Appendix A. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/j.chroma.
2015.02.020.
References
[1] F. Delgado-Vargas, A.R. Jiménez, O. Paredes-López, Natural pigments:
carotenoids, anthocyanins, and betalains—characteristics, biosynthesis,
processing, and stability, Crit. Rev. Food Sci. Nutr. 40 (2000) 173–289.
[2] D. Noviendri, R.F. Hasrini, F. Octavianti, Carotenoids Sources, medicinal prop-
erties and their application in food and nutraceutical industry, J. Med. Plants
Res. 5 (2011) 7119–7131.
[3] R. Ãlvarez, B. Vaz, H. Gronemeyer, Ã.R. de Lera, Functions, therapeutic appli-
cations, and synthesis of retinoids and carotenoids, Chem. Rev. 114 (2014)
1–125.
[4] H.N. Basu, A.J. Del Vecchio, F. Flider, F.T. Orthoefer, Nutritional and potential
disease prevention properties of carotenoids, J. Am. Oil Chem. Soc. 78 (2001)
665–675.
[5] D.B. Rodriguez-Amaya, Quantitative analysis, in vitro assessment of bioavail-
ability and antioxidant activity of food carotenoids—a review, J. Food Compos.
Anal. 23 (2010) 726–740.
[6] Y. Liu, J. Liu, X. Chen, Y. Liu, D. Di, Preparative separation and purification of
lycopene from tomato skins extracts by macroporous adsorption resins, Food
Chem. 123 (2010) 1027–1034.
[7] R. Aman, R. Carle, J. Conrad, U. Beifuss, A. Schieber, Isolation of carotenoids
from plant materials and dietary supplements by high-speed counter-current
chromatography, J. Chromatogr. A 1074 (2005) 99–105.
[8] J. Oliver, A. Palou, Chromatographic determination of carotenoids in foods, J.
Chromatogr. A 881 (2000) 543–555.
[9] B. Tan, Analytical and preparative chromatography of tomato paste
carotenoids, J. Food Sci. 53 (1988) 954–959.
[10] M. Isaksen, G.W. Francis, Preparative HPLC of carotenoids, Chromatographia 27
(1989) 325–327.
[11] Q. Su, K.G. Rowley, N.D.H. Balazs, Carotenoids: separation methods applicable
to biological samples, J. Chromatogr. B 781 (2002) 393–418.
[12] D.B. Rodriguez-Amaya, A Guide to Carotenoid Analysis in Foods, ILSI Press,
Washington, DC, 2001, pp. 64.
[13] Y. Ito, Golden rules and pitfalls in selecting optimum conditions for high-speed
counter-current chromatography, J. Chromatogr. A 1065 (2005) 145–168.
[14] A. Berthod, T. Maryutina, B. Spivarov, O. Shpigun, I.A. Sutherland, Countercur-
rent chromatography in analytical chemistry (IUPAC technical report), Pure
Appl. Chem. 81 (2009), 355-338.
125
[15] Y. Wei, T. Zhang, G. Xu, Y. Ito, Application of CCC for the separation of lutein from
a crude extract of marigold flower petals, J. Liq. Chromatogr. Relat. Technol. 26
(2003) 1659–1669.
[16] H.-B. Li, F. Chen, T.-. Zhang, F.-. Yang, G.-. Xu, Preparative isolation and
purification of lutein from the microalga Chlorella vulgaris by high-speed
counter-current chromatography, J. Chromatogr. A 905 (2001) 151–155.
[17] F. Chen, H. Li, R.N. Wong, B. Ji, Y. Jiang, Isolation and purification of the
bioactive carotenoid zeaxanthin from the microalga Microcystis aeruginosa by
high-speed counter-current chromatography, J. Chromatogr. A 1064 (2005)
183–186.
[18] S. Baldermann, A. Reinhard, N. Köhler, P. Fleischmann, Application of high-
speed counter-current chromatography for the isolation of 9’-cis-neoxanthin
from fresh spinach, J. Chromatogr. A 1151 (2007) 183–186.
[19] H. Oka, Y. Ikai, S. Yamada, J. Hayakawa, K.-I. Harada, M. Suzuki, H. Nakazawa, Y.
Ito, Separation of gardenia yellow components by high-speed countercurrent
chromatography, in: W.D. Conway, R.J. Petroski (Eds.), Modern Countercur-
rent Chromatography, ACS Symposium Series 593, American Chemical Society,
Washington, DC, 1995, pp. 92–106.
[20] Y. Wei, T. Zhang, G. Xu, Y. Ito, Application of analytical and preparative high-
speed counter-current chromatography for separation of lycopene from crude
extract of tomato paste, J. Chromatogr. A 929 (2001) 169–173.
[21] S. Baldermann, K. Ropeter, N. Köhler, P. Fleischmann, Isolation of all-
trans lycopene by high-speed counter-current chromatography using a
temperature-controlled solvent system, J. Chromatogr. A 1192 (2008)
191–193.
[22] M. Englert, W. Vetter, Solvent systems with n-hexane and/or cyclohexane
in countercurrent chromatography—physico-chemical parameters and their
impact on the separation of alkyl hydroxybenzoates, J. Chromatogr. A 1342
(2014) 54–62.
[23] L. Marchal, O. Intes, A. Foucault, J. Legrand, J. Nuzillard, J.-H. Renault, Ratio-
nal improvement of centrifugal partition chromatographic settings for the
production of 5-n-alkylresorcinols from wheat bran lipid extract. I. Flooding
conditions—optimizing the injection step, J. Chromatogr. A 1005 (2003) 51.
[24] C.Y. Lee, Changes in carotenoid content of carrots during growth and post-
harvest storage, Food Chem. 20 (1986) 285–293.
[25] A. Ogawa, D.P. Curran, Benzotrifluoride: a useful alternative solvent for organic
reactions currently conducted in dichloromethane and related solvents, J. Org.
Chem. 62 (1997) 450.
[26] J. Maul, P. Ostrowski, G. Ublacker, B. Linclau, D. Curran, Benzotrifluoride and
derivatives useful solvents for organic synthesis and fluorous synthesis, in: P.
Knochel (Ed.), Modern Solvents in Organic Synthesis, Springer, Berlin, Heidel-
berg, 1999, pp. 79–105.
[27] M. Agnely, D. Thiébaut, Dual-mode high-speed counter-current chromatogra-
phy: retention, resolution and examples, J. Chromatogr. A 790 (1997) 17.
[28] T.A. Maryutina, S.N. Ignatova, I.A. Sutherland, J. Liq. Chromatogr. Rel. Technol.
26 (2003) 1537.