Journal of Chromatography A, 1140 (2007) 95–100

Isolation of chlorophylls a and b from spinach by

counter-current chromatography
Carole Jubert a,∗ , George Bailey a,b,c,d
a Linus Pauling Institute, Oregon State University, 435 Weniger Hall, Corvallis, OR 97331, USA
b Environmental Health Sciences Center, Oregon State University, 1007 Agriculture and Life Sciences Building,
Corvallis, OR 97331, USA
c Marine Freshwater Biomedical Sciences Center, Oregon State University, 435 Weniger, Corvallis, OR 97331, USA
d Environmental and Molecular Toxicology, Oregon State University, 1007 Agriculture and Life Sciences Building,

Corvallis, OR 97331, USA

Received 12 August 2006; received in revised form 14 November 2006; accepted 17 November 2006
Available online 11 December 2006

Abstract
A method for the isolation of chlorophylls from spinach by counter-current chromatography was developed. An initial extraction protocol was
devised to avoid the notorious sensitivity of chlorophylls to degradation by light, heat, oxygen, acids and bases. Further purification and separation
of chlorophylls a and b were achieved using counter-current chromatography. Chlorophyll structures and purities were established by HPLC, fast
atom bombardment mass spectrometry and nuclear magnetic resonance. Purity was estimated to be >95% (100% by HPLC). Typical yields from
30 g of freeze-dried spinach were 300 mg of chlorophyll a and 100 mg of chlorophyll b.
© 2006 Elsevier B.V. All rights reserved.

Keywords: Counter-current chromatography; Chlorophylls a and b

1. Introduction ical transformations, which can occur during their extraction

Chlorophylls, responsible for the green color of leaves, are
the pigments of photosynthesis first described by Pelletier and
Caventou [1]. They named it from the Greek, chlorós – green
and phyllon – leaf. The first successful separation of chlorophyll
was reported by Tswett [2] giving birth to modern chromatog-
raphy techniques. The entire structure of chlorophylls a and
b was elucidated by Fischer and Wenderoth [3] and the first
chemical synthesis was reported by Woodward et al. [4]. In the
past 40 years, a considerable growth in chlorophyll research has
brought improvements in their separation, structure, chemistry
and analytical methods. A variety of chromatographic proce-
dures, including paper [5,6], thin-layer [7], conventional column
[8,9], and high-performance liquid chromatography [10,11]
have been developed and used for analytical and preparative sep-
arations of chlorophylls and their derivatives. It is well known
that chlorophylls are extremely susceptible to a number of chem-
∗ Corresponding author. Tel.: +1 541 737 4665; fax: +1 541 737 7966.
E-mail address: carole.jubert@oregonstate.edu (C. Jubert).
and separation. Many of the chlorophyll isolation procedures
described in the literature [12–14] suffer from low yields, tedious
and time-consuming purification steps, lack of separation of
chlorophylls a and b and excessive sample handling that can pro-
duce chlorophyll alteration products. Furthermore, commercial
preparations of chlorophylls a and b routinely contain epimers,
132 -hydroxychlorophylls and other allomers (oxidation prod-
ucts) and thus may require additional purification.
We have recently undertaken examination of counter-current
approaches as a means to provide highly purified chlorophylls
on a scale suitable for animal and human health research. Craig’s
counter-current distribution (CCD) was a breakthrough in sep-
aration science and became very popular in the 1940s [15–17].
CCD devices that filled whole rooms were capable of processing
liters of solvents and included up to several hundred partition ele-
ments that automated the manual partitioning of samples in sepa-
ratory funnels. The next generation, counter-current chromatog-
raphy (CCC), was first described by Ito in the early 1970s [18].
Since then, it has been widely used in the field of natural product
chemistry to effectively separate a large variety of compounds
[19–24]; yet, CCC techniques have been underutilized due to the

0021-9673/$ – see front matter © 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.chroma.2006.11.063
96 C. Jubert, G. Bailey / J. Chromatogr. A 1140 (2007) 95–100
lack of information available to the beginner. In a recent article
[25], Ito describes the basic technical details to quickly learn
optimal conditions for CCC separations. As CCC is an all-liquid
technique, it benefits from a number of advantages in compar-
ison with the more traditional liquid–solid separation methods,
such as column chromatography and HPLC. The main advantage
of CCC is that losses by adsorption or denaturation by contact
with a solid support are avoided. With a liquid stationary phase,
the compounds can diffuse into the entire volume of the station-
ary phase, rather than its surface only, as is the case with classical
solid stationary phase. As a result, the loading capability of
CCC is much higher than that of HPLC with an equal internal
volume.
There are few procedures that can be successfully adapted
from laboratory to production scale without difficulties. Prepar-
ative HPLC, for example, is not a linear scale-up as the product
can become hydrolyzed by or react with the column. CCC can be
scaled-up from analytical- to preparative-scale in a completely
straightforward manner avoiding these problems [26,27]. Most
parameters can be scaled up in proportion to the increase in
column volume.
The specific objective of our research was to develop a
method for the isolation of chlorophylls to be investigated as
chemoprotective agents. We describe an improved method for
the extraction, purification and separation of chlorophylls a and
b from spinach. Structure and purity of the isolated chlorophylls
were confirmed by high performance liquid chromatography,
fast atom bombardment mass spectrometry (FAB-MS) and
nuclear magnetic resonance spectroscopy.
2. Experimental
2.1. Materials
All solvents used were purchased from VWR, International
or Fischer Scientific. Light petroleum and denatured ethanol
were of reagent grade, whereas acetone, heptane, methanol, ace-
tonitrile and ethyl acetate were of HPLC grade. Spinach was
purchased from local venders and materials were as fresh as
possible and grown without using conventional pesticides and
fertilizers.
2.2. Sample preparation
To minimize degradation and isomerization of chlorophylls,
all work was performed in dim light. Fresh spinach leaves were
first separated from the mid-ribs and washed with cold water.
The leaves were frozen in 40 kg batches for freeze-drying (Ore-
gon Freeze-Dry, Albany, OR, USA). The freeze-dried material
was stored in bulk in plastic bags and refrigerated until use.
Each extraction used 30 g of freeze-dried spinach (equivalent
to 450 g of fresh spinach). The leaves were washed twice in
a blender with 500 mL of light petroleum (b.p. 30–60 ◦ C) to
remove a large portion of the carotenoids and waxes and then
extracted twice using 400 mL of methanol/light petroleum (3:1,
v/v). Filtration after each wash and extraction was done on a
fritted funnel in order to separate particulate spinach residues
from solvents. The combined extracts were transferred to a sep-
aratory funnel and washed with 250 mL of saturated sodium
chloride pulling the methanol into the aqueous and leaving the
light petroleum containing the dark chlorophyll pigment as the
organic layer. The aqueous layer was extracted with 200 mL of
light petroleum. The light petroleum layers were combined and
washed with 100 mL of saturated sodium chloride. The extract
was filtered on a clean fritted funnel a final time and evaporated
in vacuo (T < 30 ◦ C). This residue was dissolved with 50 mL
of acetone and left overnight at −20 ◦ C in order to precipitate
impurities. The acetone fraction was filtered using a clean frit-
ted funnel and evaporated in vacuo. The solvent-free extracted
product was enriched for chlorophylls (450 mg) but still con-
tained other pigments, such as carotenoids as well as oils, fats
and waxes derived from the spinach, which required additional
removal.
2.3. CCC
2.3.1. Apparatus
CCC was performed using the Ito multilayer-coil separator-
extractor produced by P.C., Potomac, MD, USA. The radius of
the column orbit was 10.2 cm and the coil was 18 cm in diameter
and 5 cm thick, with windings from a 10.2-cm hub (β = 0.5) to the
coil outer diameter (β = 0.85). The multilayer coil consisted of
a single piece of 2.7 mm I.D. PTFE tubing with a total capacity
of 400 mL. The coiling of the tubing was started with right-
handed fashion and the inner terminal was used as the inlet. The
revolution (forward or reversed) speed was adjusted from 0 to
1000 rpm with a speed controller (Bodine Electric, Chicago, IL,
USA). The CCC was equipped with a ReciPro Metering Pump
(Eldex Laboratories, Napa, CA, USA) for solvent delivery and
a V4 absorbance detector (Isco, Lincoln, NE, USA) for peak
detection.
2.3.2. Measurement of Kupper phase/lower phase value
A few milligrams of the crude extract was added to the
two mutually equilibrated solvent phases (1–2 mL each; hep-
tane/ethanol/acetonitrile/water, 10:8:1:1, v/v) in a test tube, and
mixed to equilibrate. After settling, equal volumes of the upper
and lower phases were transferred into separate test tubes and
diluted each with an equal volume of acetone. Each phase
was assessed by HPLC and the area under each peak was
used to determine the KU/L values for the carotenoids and
chlorophylls.
2.3.3. Solvent system
The two phases (heptane/ethanol/acetonitrile/water, 10:8:1:1,
v/v) were mutually saturated by shaking in a separatory funnel
and separated immediately before use. In CCC, either phase
can serve as the mobile or stationary phase depending on the
direction of the column rotation. For this study, the aqueous
bottom layer served as the CCC mobile phase to obtain highly
pure chlorophyll a, while the top layer (rich in heptane) served
as the CCC mobile phase to obtain highly pure chlorophyll b.
For either protocol, the crude chlorophyll samples consisting
of approximately 450 mg of chlorophyll were resuspended in
 C. Jubert, G. Bailey / J. Chromatogr. A 1140 (2007) 95–100
12 mL of stationary phase. As a general rule, the sample size
can be up to about 5% of the total volume without altering the
partition coefficient.
2.3.4. Chlorophyll isolation
The stationary phase (heptane for chlorophyll a and ethanol
for chlorophyll b) was loaded into the inlet of the coil
in the absence of rotation, followed by the dissolved sam-
ple. Rotation (forward/counterclockwise for chlorophyll a and
reversed/clockwise for chlorophyll b) was then started at a rev-
olution speed of 700 rpm and the mobile phase (ethanol for
chlorophyll a and heptane for chlorophyll b) was pumped into
the column at a flow rate of 5 mL/min. The chromatographic
runs were monitored at 440 nm and the purified fractions were
collected from the spectroscopy outlet.
2.4. HPLC
A Waters (Milford, MA, USA) HPLC system, which con-
sisted of a separations module (model 2690) and a photo diode
array detector (model 996), was used for analyzing purified
chlorophyll samples. The instrument was controlled by a Dell
personal computer using Waters Millennium software. Chro-
matographic analyses were performed using an Alltima C18
column (250 mm × 4.6 mm I.D.) (Alltech, Deerfield, IL, USA)
at 35 ◦ C and a flow rate of 1 mL/min. For the separation of the
chlorophylls, eluent A consisted of methanol/0.5 M ammonium
acetate (4:1, v/v) and eluent B of methanol/acetone (9:1, v/v)
according to Jeffrey [11]. The two eluents were step-changed
after 3 min of flow.
2.5. NMR spectroscopy
The 1 H NMR spectra of the chlorophylls were recorded on
a Bruker 400 MHz spectrometer with deuterated acetone as the
solvent (δ 2.05).
2.6. Mass spectrometry
MS analysis was performed on a JEOL MSRoute (JMH-600)
magnetic sector mass spectrometer. A FAB positive ion mass
spectrum was recorded using Xenon and the sample matrix was
3-nba (3-nitrobenzyl alcohol).
Fig. 1. Step-isocratic HPLC separation of crude spinach extract. UV absorbance (λ = 440 nm) profile obtained during a run in which solvent A (methanol:0.5 M
ammonium acetate, 80:20 (v/v) pH = 7.1) was pumped for 3 min then step-changed to solvent B (methanol:acetone, 90:10 (v/v) and ran for 30 min. The injection
volume was 10 ␮L. Carotenoids (12%) eluted between 8 and 11 min, followed by chlorophylls b/b’ (18%) at 13.75 and 14.5 min and chlorophylls a/a’ (70%) at 18
and 20 min.
97
2.7. Quantification
Quantification was performed on a DU-70 UV-Vis spec-
trophotometer (Beckman, Fullerton, CA, USA). The corrected
equations for the determination of chlorophylls a and b in
methanol were used as determined by Porra et al. [28]. The
specific equations for chlorophyll concentrations in ␮g/mL
are as follows where A is the absorbance at the specified
wavelength:
Chla = 16.29A665.2 − 8.54A652.0
Chlb = 30.66A652.0 − 13.58A665.2
Chlsa + b = 22.12A652.0 + 2.71A665.2
3. Results and discussion
3.1. Crude spinach extract
Chlorophyll, being a highly reactive molecule, needs to be
handled with extreme care during isolation procedures. It is
especially susceptible to degradation by light, heat, oxygen,
acids and bases. Typical reactions are allomerization (oxida-
tion), epimerization at C-132 , de-metallation, de-phytylation,
trans-esterification and decarboxymethylation at C-132 (forma-
tion of ‘pyro’ derivatives) [29]. Our crude extract was free from
these derivatives, which simplified the purification. The HPLC
method gave good separation of chlorophylls a and b and their
epimers a’ and b’ (Fig. 1).
3.2. Partition coefficient and suitable solvent system
All chromatographic separations are based upon the differ-
ence in the equilibrium distribution of an analyte between a
mobile and a stationary phase. In CCC, the partition coefficient
(KU/L ) is expressed as the amount of solute in the upper phase
divided by that of the lower phase. The choice of a two-phase
solvent system for optimal separation is the main challenge in
CCC. In general, the more soluble a compound is in the mobile
phase, the faster it will elute. If the compound is evenly dis-
persed in the two phases (KU/L = 1), it will elute in one column
98 C. Jubert, G. Bailey / J. Chromatogr. A 1140 (2007) 95–100
volume no matter which phase is used as the mobile phase;
therefore, the recommended KU/L values for CCC are in the
range 0.5 ≤ KU/L ≤ 1.0. It is also important to measure KU/L val-
ues of impurities as well as target compounds. By calculating
the ratio of KU/L values between all analytes, the degree of res-
olution between peaks can be predicted. A ratio of 1.5 or more
will provide good separation. The measured KU/L values for
the carotenoids, chlorophylls b and a were 0.12, 0.47 and 1.61,
respectively.
Since CCC has no solid support, high peak resolution depends
on the retention of the stationary phase in the column. Ideally
at least 50% of the stationary phase should be retained while
the mobile phase passes through the system. Ito [25] notes that
the retention of the solvents is highly correlated to their settling
times after vigorous shaking and suggests less than 20 s. Our
solvent system settled in less than 5 s and gave nearly equal
volumes of phases post equilibration. In our runs, approximately
50% of the stationary phase was retained.
Another consideration in the choice of solvents is that
the analytes should be stable and soluble in the solvent sys-
tem. Initial trials with methanol revealed several chlorophyll
degradation products (isomers and allomers). These were not
observed with ethanol, which also gave adequate separation dur-
ing CCC purification. An added advantage is that ethanol forms
an azeotrope with water enabling easier solvent removal post
isolation.
3.3. CCC conditions for optimal separation
Specific parameters to consider for CCC operation are the
choice of the mobile phase and its flow rate, and the rotation
speed of the apparatus. Either phase can be used as the mobile
phase and the decision should be based on the KU/L value. A
high KU/L value (1.0 ≤ KU/L ≤ 2.0) should favor the upper phase
as the mobile phase while a low KU/L value (0.5 ≤ KU/L ≤ 1.0)
should favor the lower phase as the mobile phase. For the general
isolation of chlorophylls a and b, the lower aqueous phase was
used for the mobile phase generating a run of 2 h and 30 min.
Fig. 2. Counter-current chromatographic separation of carotenoids and chlorophylls. Solvent system, heptane-methanol-water-acetonitrile, 10:8:1:1. CCC column
volume, 400 mL; flow rate, 5 mL/min; rotation, 700 rpm. Detection, UV 440 nm; the actual units of absorbance were not measured as they were out of the linear
range of Beer’s law.
Chlorophylls (b, b’, a and a’) were isolated, giving an exception-
ally good separation for chlorophyll a. For the isolation of pure
chlorophyll b, we set up the elution in a tail to head mode [25]
to promote retention of the lower ethanol phase. Use of an upper
organic phase as the mobile phase facilitated the removal of sol-
vent from the collected fractions, and was especially useful in
accommodating chlorophyll sensitivity to heat and oxygenated
solvents. The flow rate of the mobile phase determines the sep-
aration time, the amount of stationary phase retained in the
column, and therefore the peak resolution. A lower flow rate
usually gives higher retention level of the stationary phase [30]
improving the peak resolution although it requires a longer sep-
aration time. The optimum revolution speed for a preparative
column with 2.6 mm I.D. is 600 to 800 rpm [25]. A flow rate
of 5 mL/min and a rotation speed of 700 rpm were used giving
a reasonable retention of the stationary phase (50%) and good
peak resolution.
3.4. Identification and purity
The chromatogram of the CCC separation of the crude extract
from spinach (heptane as the stationary phase) is given in Fig. 2
and shows the separation of carotenoids (peak I), chlorophylls
b/b’ (peak II), chlorophyll a (peak III) and chlorophyll a’ (peak
IV). The respective HPLC chromatograms of the isolated chloro-
phylls are shown in Fig. 3. The chromatographic behavior of
the isolated chlorophylls was identical with authentic standards
on reversed-phase HPLC. The exact mass for chlorophyll a
(892.5317 u) for the molecular ion was within 4 ppm of the the-
oretical value (892.5353 u). Furthermore, a low resolution, wide
mass range FAB-MS of chlorophyll a showed the significant
mass peaks at 481 (55%), 555 (40%), 614 (100%), and 893
(20%). All the peaks had complex isotope patterns implying
the presence of magnesium (Fig. 4). The 1 H-NMR of the iso-
lated chlorophyll a was determined in deuterated acetone. The
spectrum was characteristic of chlorophyll a (Fig. 5). A minor
impurity was detected at δ 5.40 consisting of 0.35 protons upon
integration and the only other extra peaks were the NMR solvent
 C. Jubert, G. Bailey / J. Chromatogr. A 1140 (2007) 95–100 99

Fig. 3. HPLC of chlorophyll fractions post CCC separation. Fractions collected from the CCC injection loop (peaks II, III and IV) were injected on HPLC with the
same conditions as Fig. 1. All runs were for 30 min, but no peaks eluted after 20 min.

(δ 2.05) and the two signals due to H2 O and HOD (δ 2.78 and
2.75). The presence of any residual solvents from the extrac-
tion/purification was ruled out upon analysis of the spectrum
as no characteristic solvents peaks were observed. In conclu-
sion, the proton chemical shifts of the isolated chlorophyll a
were in agreement with published spectra [31]. Purity was esti-
mated to be >95% (100% by UV–vis) compared to chlorophyll
a standards (Sigma). We note that these were routinely found
to be 90–92% pure based on spectroscopic measurements. In
summary, our protocol for the isolation of chlorophylls gave
us not only excellent purity but also high recovery. The yields
from 30 g of freeze-dried spinach were 300 mg for chlorophyll

•+
Fig. 4. FAB positive ion mass spectrometry of chlorophyll a. The loss of the phytyl chain (indicated as fragment ion M 614.3) was the most abundant sample ion.
100 C. Jubert, G. Bailey / J. Chromatogr. A 1140 (2007) 95–100

Fig. 5. 1 H NMR of chlorophyll a, 1.6 mM in [2 H6 ]acetone, recorded on a Bruker 400 MHz spectrometer. A minor impurity was detected at δ 5.40 consisting of 0.35
protons upon integration and the only other extra peaks were the NMR solvent (δ 2.05) and the two signals due to H2 O and HOD (δ 2.78 and 2.75).

a and 100 mg for chlorophyll b (1.5% based on the dry weight References
of spinach).
[1] J. Pelletier, J.B. Caventou, J. Pharm. 3 (1817) 486.
[2] M. Tswett, Ber. Deutsch. Bot. Ges. 24 (1906) 384.
4. Conclusion [3] H. Fischer, H. Wenderoth, Annalen 537 (1939) 170.
[4] R.B. Woodward, W.A. Ayer, J.M. Beaton, F. Bickelhaupt, R. Bonnett, P.
This work presents a successful application of counter- Buchschacher, G.L. Closs, H. Dutler, J. Hannah, F.P. Hauck, S. Ito, A.
current chromatography to the preparative purification of Langemann, E. LeGoff, W. Leimgruber, W. Lwowski, J. Sauer, Z. Valenta,
chlorophylls. There is no loss or degradation of sample by H. Voltz, J. Am. Chem. Soc. 82 (1960) 3800.
[5] H.H. Strain, W.A. Svec, in: E. Heftmann (Ed.), Chromatography, third ed.,
irreversible adsorption or hydrolysis. Analytical data of the Van Nostrand-Reinhold, New York, 1975, pp. 744–758.
chlorophylls prepared in our laboratory indicate that the degree [6] Z. Sestak, Photosynthetica 14 (1980) 239.
of purity of our isolated chlorophylls exceeds that of commercial [7] Z. Sestak, Photosynthetica 16 (1982) 568.
chlorophylls. CCC presents a number of advantages over cur- [8] H.H. Strain, B.T. Cope, W.A. Svec, Methods Enzymol. 23 (1971) 452.
rent technologies. It is possible to purify chlorophylls in one step [9] T. Omata, N. Murata, Plant Cell Physiol. 24 (1983) 1093.
[10] S. Roy, J. Chromatogr. 391 (1987) 19.
from a crude extract with high loading capacity without the need [11] S.W. Jeffrey, S.W. Wright, M. Zapata, Marine Freshwat. Res. 50 (1999)
for sample clean-up. The procedure described in this manuscript 879.
has several advantages: it is not time-consuming, chlorophylls a [12] K. Iriyama, M. Yoshiura, M. Shiraki, Seikagaku 52 (1980) 236.
and b are separated during the same isolation, yields are high and [13] H. Brockmann, N. Rische, Chlorophylls (1991) 103.
no alteration products are generated during isolation. Through [14] J. Yang, Y. Tao, Zhiwu Shenglixue, Tongxun 38 (2002) 156.
[15] L.C. Craig, J. Biol. Chem. 150 (1943) 33.
this work, we have demonstrated the unique appeal of CCC [16] L.C. Craig, J. Biol. Chem. 155 (1944) 519.
for chlorophyll purification. The technique offers high capacity, [17] N. Kresge, R.D. Simoni, R.L. Hill, J. Biol. Chem. 280 (2005) e4.
superior resolution, excellent sample recovery (no absorption [18] Y. Ito, R.L. Bowman, Science 167 (1970) 281.
on solid phase), possible solvent recycling, a mild method (does [19] Y. Wei, T. Zhang, G. Xu, Y. Ito, J. Chromatogr. A 929 (2001) 169.
not harm fragile compounds with harsh conditions), versatility [20] L. Lei, F. Yang, T. Zhang, P. Tu, L. Wu, Y. Ito, J. Chromatogr. A 912 (2001)
181.
(many solvent combinations possible), and ease of achieving [21] Q. Du, Z. Li, Y. Ito, J. Chromatogr. A 923 (2001) 271.
normal phase or reversed-phase separation. [22] A. Degenhardt, P. Winterhalter, Spec. Publ. R. Soc. Chem. 269 (2001) 143.
[23] J.A. Armbruster, R.P. Borris, Q. Jimenez, N. Zamora, G. Tamayo-Castillo,
Acknowledgments G.H. Harris, J. Liquid Chromatogr. Relat. Technol. 24 (2001) 1827.
[24] K.A. Alvi, J. Liquid Chromatogr. Relat. Technol. 24 (2001) 1765.
[25] Y. Ito, J. Chromatogr. A 1065 (2005) 145.
This study was supported by grants P01 CA90890, ES00210, [26] Y. Ito, J. Chromatogr. 538 (1991) 3.
and ES03850 from the National Institutes of Health. We thank [27] I. Sutherland, D. Hawes, S. Ignatova, L. Janaway, P. Wood, J. Liquid
Dr. Walter Conway for his excellent technical assistance con- Chromatogr. Relat. Technol. 28 (2005) 1877.
cerning CCC maintenance and performance. We also wish to [28] R.J. Porra, W.A. Thompson, P.E. Kriedemann, Biochim. Biophys. Acta,
Bioenerg. 975 (1989) 384.
thank Oregon Freeze Dry for the use of their equipment and [29] S.B. Brown, J.D. Houghton, G.A.F. Hendry, Chlorophylls (1991) 465.
Tammie McQuistan and Gina Miller for their technical help with [30] Q. Du, C. Wu, G. Qian, P. Wu, Y. Ito, J. Chromatogr. A 835 (1999) 231.
the extractions and CCC runs. [31] K.M. Smith, D.A. Goff, R.J. Abraham, Org. Magn. Reson. 22 (1984) 779.