A

Aa

b

a

AA

KCRCEC

1

noaspipaoit[

ril

0d

Journal of Chromatography A, 1216 (2009) 4169–4175

Contents lists available at ScienceDirect

Journal of Chromatography A

journa l homepage: www.e lsev ier .com/ locate /chroma

dvantages of a small-volume counter-current chromatography column

lain Berthoda,∗, Svetlana Ignatovab, Ian A. Sutherlandb

Laboratoire des Sciences Analytiques, CNRS, Université de Lyon, Bat. CPE-308, 69622 Villeurbanne, FranceBrunel Institute for Bioengineering, Brunel University, Uxbridge, Middlesex UB8 3PH, UK

r t i c l e i n f o

rticle history:vailable online 12 November 2008

eywords:ounter-current chromatographyetention theoryhromatographic resolutionfficiencyolumn volume

a b s t r a c t

Counter-current chromatography (CCC) works with a support-free liquid stationary phase. This allowsfor preparative separations and purifications. However, there are serious technical constraints becauseof the need to keep a liquid stationary phase in a column. Centrifugal fields are used. A new com-mercial hydrodynamic 18 mL column made with a narrow-bore 0.8 mm Teflon tubing was evaluatedby comparing it with older hydrodynamic CCC columns and a similar 19 mL column but made with1.6 mm Teflon tubing. A small-volume CCC column allows for reliable and fast solute partition coef-ficient determination. When resolution is required, both high efficiency and liquid stationary phaseretention are needed. Unfortunately, these two requirements bear technical contradictions. A columncoiled with a narrow tubing bore will provide a high chromatographic efficiency while a column con-

taining wider tubing bore will achieve higher stationary phase retention. In all cases, increasing themagnitude of the centrifugal field also increases the stationary phase retention. The solution is to buildcentrifuges able to produce high fields that will provide acceptable liquid phase retention with narrow-bore tubes. The new 18 mL 0.8 mm tubing bore column is able to rotate as fast as 2100 rpm generatinga 240 × g field. The two older CCC columns cannot compete with the new one. However, the small19 mL column with 1.6 mm bore tubing can be useful when fast results are desired without top reso-

cc

V

TriaC[tt

n

lution.

. Introduction

Counter-current chromatography (CCC) is a separation tech-ique that uses two liquid phases, one is the mobile phase, and thether liquid phase is the stationary phase [1–4]. The advantages ofsupport-free liquid stationary phase are tremendous. Since the

ample components can access the volume of the liquid stationaryhase, and not only the surface of the solid stationary phases used

n classical liquid chromatography (LC), overload is much less of aroblem than it is in classical preparative LC. Also, problems suchs irreversible adsorption or loss of sample in the column cannotccur with a properly handled CCC column. If part of the samples retained too long in the CCC column, extrusion of the liquid sta-ionary phase allows for quick recovery of the retained components5].

The other great advantage of CCC is the very simple soluteetention mechanism: only liquid–liquid partitioning is involvedn solute retention. The retention volume, VR, depends on theiquid–liquid distribution coefficient, KD, also called partition

∗ Corresponding author.E-mail address: berthod@univ-lyon1.fr (A. Berthod).

speHoUa2e

021-9673/$ – see front matter © 2009 Elsevier B.V. All rights reserved.oi:10.1016/j.chroma.2008.11.013

© 2009 Elsevier B.V. All rights reserved.

oefficient, of the solute and on the phase volumes inside theolumn:

R = VM + KDVS (1)

he subscripts M and S stand for mobile and stationary phases,espectively. CCC is used to determine solute partition coefficientsn biphasic liquid systems [2,4]. The very simple retention mech-nism allows for a straightforward direct scale-up from a smallCC column (tens of mL) to a large prep size unit (up to several L)6]. These advantages are establishing CCC as a modern preparativeechnique able to purify large amounts of material with competitivehroughputs [7].

The liquid stationary phase is the best part of the CCC tech-ique. It is also its major problem: it is very difficult to maintain aupport-free liquid phase stationary when a liquid mobile phase isumped through it. From the beginning, Ito, the inventor and mostfficient developer of the CCC technique, used centrifugal fields [8].e developed numerous CCC columns mostly based on variations

f the coil planet centrifuge hydrodynamic CCC column [1,3]. AK small-volume hydrodynamic column able to provide reliablend stable high centrifugal field was firstly presented at the CCC002 symposium [9]. The interest of such a low volume column isvaluated here.

4 atogr.

2

2

eC

tt(wadietfnpa

vvnw7l

pSTi(r5m∼t

rismi

waG

2

bw(1dSi2

2

othoAti1I

Fp(

170 A. Berthod et al. / J. Chrom

. Materials and methods

.1. Counter-current chromatography columns

Two different J-type CCC instruments were used. Each wasquipped with different spools to allow working with four differentCC columns.

The Mini HPCCC is a modern hydrodynamic CCC column ofhe coil planet centrifuge type. It contains a single 7 cm diame-er coil mounted in a cantilever rotor containing a counterweightFig. 1). It is produced by Dynamic Extractions (Slough, UK,ww.dynamicextractions.com). The system volume was measured

s 19.6 mL including the column, its connecting tubing and UVetector. Therefore, the coil volume is only 18 mL obtained by wind-

ng 35 m of 0.8 mm I.D. Teflon tubing in about 180 turns in the 7 cmxternal diameter bobbin of 4.5 cm length. The distance betweenhe central rotor axis and the coil axis is 5.0 cm. The ˇ-value rangesrom 0.5 to 0.76. Very stable rotation is obtained with minimumoise using a cantilever design previously fully described [9] andatented [10]. The maximum rotation speed is 2100 rpm producing240 × g field.

Another hydrodynamic CCC column was used. It has a similarolume but obtained coiling with a wider-bore tubing. The systemolume was measured as 20.8 mL including the column, its con-ecting tubing and detector. The coil volume is 20 mL obtained byinding 9.5 m of 1.6 mm I.D. Teflon tubing in about 50 turns in acm external diameter bobbin. The maximum rotation speed is

imited to 1800 rpm producing a 180 × g field.The second CCC device used is a SFCC2000 three-spool coil

lanet centrifuge that was built between 1988 and 1992 by theociete Francaise de Colonnes Chromatographiques (Paris, France).his sturdy machine is still in use daily in our lab. However, its no longer possible to use the initial 1500 rpm rotation speed200 × g). Due to mechanical wearing, the maximum reliable rotor

otation is reduced today to only 800 rpm producing a limited4 × g centrifugal field. This unit has three identical 8 cm O.D. coilsade by winding 24 m of 1.65 mm I.D. Teflon tubing producing a50 mL volume each. The ˇ-value ranges from 0.56 to 0.78. The

hree coils are mounted symmetrically in the rotor (R = 6 cm). The

ahcui

ig. 1. The Mini HPCCC hydrodynamic column of Dynamic Extractions (Slough, UK, wwrotected by a silicon coat and the labeled counterweight. Two coils were used: one contai20 mL).

A 1216 (2009) 4169–4175

otor self-balances when all three coils are used in series giv-ng a column volume of 156 mL. The SFCC2000 can work with aingle coil when the two others being by-passed, thus giving aeasured 54 mL CCC column due to some internal connecting tub-

ngs.The pump was a LC6A Shimadzu (Lyon, France). The detector

as a Shimadzu SPD6A UV detector and the data were collectednd processed using the Azur chromatography software (Datalys,renoble, France).

.2. Chemicals

The solvents heptane, ethyl acetate and methanol were providedy Sigma (Sigma–Aldrich, L’Isles d’Abeau Chesnes, France). Wateras deionized (Aquadem, Veolia Water, Paris, France) and distilled

Simplicity, Millipore, Molsheim, France). The test solutes were the2 compounds selected for the wide polarity range (Table 1) andefined as the GUESS mixture [11]. They were all obtained fromigma and used as received. The concentrations of the solutes weren the g/L range adjusted so that the peaks are easily detectable at54 nm.

.3. Solvent systems

The different biphasic liquid systems used were all compositionsf the heptane–ethyl acetate–methanol–water system belongingo the Arizona scale [12]. The compositions of the Arizona scaleave exactly the same heptane over ethyl acetate and methanolver water ratios. They are referred to by alphabetic letters from

to Z hence the Arizona naming from the postal code AZ ofhis American state. The first most polar composition of the scales referred as A and is simply the binary ethyl acetate–water:1 (v/v) mixture. The least polar composition is referred as Z.

t is the heptane–methanol 1:1 (v/v) mixture. The intermedi-

te N composition made with the 1:1:1:1 (v/v) proportion ofeptane, ethyl acetate, methanol and water was the most usedomposition. Once a composition is selected, the four solvent vol-mes are measured and mixed together in a graduated cylinder

ntroducing water, the denser liquid, last. The biphasic system is

w.dynamicextractions.com). The inset shows the 7 cm coil with its flying leadsns 35 m of 0.8 mm I.D. tubing (18 mL), the other contains 9.5 m of 1.6 mm I.D. tubing

A. Berthod et al. / J. Chromatogr. A 1216 (2009) 4169–4175 4171

Table 1Structures and some physicochemical properties of GUESS test solutes.

Code Solute Structure Formula m.w. pKa log Ko/w

r New coccine red C20H11N2O10S3Na3 604.5 Trianion −5.1

C Caffeine C8H10N4O2 194.2 14.0 −0.07

V Vanillin C8H8O3 152.2 10.2 1.21

M Coumarin C9H6O2 146.2 n.i. 1.39

Q Quercetin C15H10O7 302.2 10.3 1.48

F Ferulic acid C10H10O4 194.2 4.4 1.51

U Umbel liferone C9H6O3 162.2 10.3 1.58

Z Salicylic acid C7H6O3 138.1 2.9 2.26

N Naringenin C15H12O5 272.3 10.1 2.52

O Carvone C10H14O 150.2 n.i. 3.07

E Estradiol C18H24O2 272.4 10.5 3.94

b �-Carotene C40H56 537.0 n.i. 17.6

S d b wea ere o

tiaudn

3

olutes selected to form the GUESS mixture spanning a wide polarity range [11]. r annd total retention (KD = ∞). n.i. = non ionizable; the octanol–water log Ko/w values w

ransferred to another identical cylinder to mix its components. It

s sonicated for 5 min to remove excess dissolved gases and usedfter 30 min thermal equilibration without separating the two liq-id phases. The entering filter of the pump is positioned in theesired upper or lower liquid phase according to the experimentaleed.

utt

re selected as colored red and yellow markers for, respectively no retention (KD = 0)btained using the KowWin software (www.syrres.com/esc/est kowdemo.htm).

. Partition coefficient determination

Eq. (1) shows the direct link between the solute retention vol-me and its partition coefficient. Different methods can be usedo determine solute partition coefficients in a biphasic liquid sys-em. The shake-flask method is most often used [13]. A weighted

4172 A. Berthod et al. / J. Chromatogr.Ta

ble

2R

eten

tion

volu

mes

ofth

eG

UES

Ste

stso

lute

sin

fou

rd

iffe

ren

tC

CC

colu

mn

sw

ith

two

dif

fere

nt

liqu

idsy

stem

s.

CC

Cco

lum

nSF

CC

3-co

ilSF

CC

1-co

ilM

ini1

.6m

mI.D

.M

ini0

.8m

mI.D

.C

CC

colu

mn

SFC

C3-

coil

SFC

C1-

coil

Min

i1.6

mm

I.D.

Min

i0.8

mm

I.D.

�rp

m

Ari

zon

aL:

2/3/

2/3

(v/v

)

800

800

180

021

00

Ari

zon

aN

:1/

1/1/

1(v

/v)

800

800

180

021

00

Fm

L/m

in4

45

24

45

2V

Mm

L45

18.5

7.1

12.2

3514

5.8

10.6

VS

mL

111

35.5

13.7

7.4

121

4015

9V

Cm

L15

654

20.8

19.6

156

5420

.819

.6

Cod

eSo

lute

KD

VR

(mL)

VR

(mL)

VR

(mL)

VR

(mL)

KD

VR

(mL)

VR

(mL)

VR

(mL)

VR

(mL)

rN

ewco

ccin

ere

d0

4518

.57.

112

.20

3514

5.8

10.6

CC

affe

ine

0.16

6324

9.3

13.5

0.09

4517

711

.5V

Van

illi

n0.

7913

246

.518

180.

3578

2811

14M

Cou

mar

in2.

3030

010

038

.529

0.80

132

4618

18Q

Qu

erce

tin

1.23

182

6224

210.

1553

208

12F

Feru

lic

acid

0.39

8932

12.5

150.

1250

197.

512

UU

mbe

llif

eron

e0.

7312

64

417

180.

2161

22.5

912

.5Z

Sali

cyli

cac

id2.

8335

911

946

331.

5822

677

2925

NN

arin

gen

in3.

1939

913

251

360.

3881

2912

14O

Car

von

e10

.17

1174

380

146

873.

0940

913

852

38E

Estr

adio

l4.

2251

416

865

431.

0516

256

2120

Liqu

idsy

stem

s:h

epta

ne–

ethy

lac

etat

e–m

eth

anol

–wat

er.R

ever

sed

-ph

ase

mod

e:th

eor

gan

icu

pp

erp

has

eis

the

stat

ion

ary

ph

ase;

the

aqu

eou

slo

wer

ph

ase

isth

em

obil

ep

has

efl

own

inth

eh

ead

tota

ild

irec

tion

inth

eC

CC

colu

mn

.Th

ep

arti

tion

coef

fici

ent

valu

esli

sted

are

the

aver

age

valu

esob

tain

edw

ith

the

fou

rse

tsof

exp

erim

ents

,sta

nd

ard

dev

iati

ons:

liqu

idsy

stem

L,4%

;li

quid

syst

emN

2.5%

.

mcttttteoto

tv

V

A

V

pattIc

ciutttdttitc

staTpcflv

vvwtmpC

fitecvr(

A 1216 (2009) 4169–4175

ass of solute, typically in the �g range, is introduced in a test tubeontaining an equal amount of the two liquid phases, typically inhe mL range. The test tube is shaken so that the solute distribu-ion reaches equilibrium between the two phases. HPLC is usedo quantitate the amount of solute in each phase [13]. However,here are always systematic biases, small changes in the liquid sys-em composition, interfering impurity and/or HPLC mobile phaseffect, which can produce errors in the shake-flask determinationf partition coefficients for use in CCC. The direct measurement ofhe CCC solute retention volume is often more accurate than mostther methods [14].

Since there are only the mobile and the stationary phase insidehe CCC column of volume VC, the relationship between the twoolumes is:

M + VS = VC (2)

lso, Eq. (1) can be rearranged as:

R = VC + (KD − 1)VS (3)

Eq. (3) shows that the solute retention volume is directly pro-ortional to its partition coefficient. It is also proportional to themount of stationary phase retained in the CCC column and tohe column volume itself. A CCC column with a small volume con-ains necessarily a small volume of stationary phase since VS < VC.t will produce reduced retention volumes hence faster partitionoefficient measurements.

KD accurate determination requires an exact knowledge of theolumn volume, that is constant, and stationary phase volume, thats highly dependent on experimental conditions and liquid systemsed. Solute r (Table 1) is a red colored dye that is not retained inhe reversed-phase mode with the aqueous lower mobile phase inhe head to tail flowing direction. Its partition coefficient can beaken as zero so that its retention volume is a precise experimentaletermination of VM, the mobile phase volume. Once VM is known,he retention volume of the other solutes allows for the computa-ion of their partition coefficients (Eqs. (1) and (3)). The solutes werenjected one by one to avoid any identification problem. Table 2 listshe results obtained with two liquid systems and four different CCColumns.

The four columns provided similar partition coefficients for allolutes. The solute partition coefficient depends on the liquid sys-em and not the CCC column used to measure it. Table 2 lists theverage value of the KD coefficient obtained with the four columns.he Arizona N system (1:1:1:1, v/v) gave more accurate resultsrobably because it was better retained than the L system by allolumns (higher and more stable Sf at the same rotation speed andow rate compared to L) and easier to prepare using the exact sameolume of the four solvents.

There is no discussion: a shorter CCC column needs less solventolume to elute a solute (Eqs. (1) and (3)). The less hydrophilic car-one solute needed more than a liter of L mobile phase to eluteith the 3-coil SFCC 154 mL column (Table 2). This volume is about

hree times lower dropping to 380 mL when a single coil is usedaking a 54 mL CCC column. 146 and 87 mL were enough L mobile

hase volumes to elute carvone with the two small-volume MiniCC columns.

The volume of solvent needed to complete an experiment is therst important parameter; the second important parameter is theime needed to complete the measurement. Fig. 2 compares the

xperiment durations for the Table 2 results (10 compounds, 4 CCColumns, Arizona N system). The two Mini columns have similarolumes. However, the 1.6 mm I.D. column gives significantly fasteresults than the 0.8 mm I.D. column. This is due to a faster flow rate5 mL/min) possible with the larger-bore Mini column. Flow rate

A. Berthod et al. / J. Chromatogr.

Fig. 2. Retention time of the GUESS solutes (Table 1) in four CCC columns withtC5(

a

4

a

S

wb

S

T(re

B

icRuaw

dsrb(ttrit

trii

bfcrtobc(1

5

bniaGcgc

ow5cCt

iia

R

Wnumber by:

N = 16(

VR

W

)2(8)

Fig. 3. Stationary (upper) phase retention ratio in percentage of the column volume

he Arizona N liquid system (heptane/ethyl acetate/methanol/water 1:1:1:1, v/v).olumn volumes (and flow rates) were: SFCC 3-coil 156 mL (4 mL/min); SFCC 1-coil4 mL (4 mL/min); Mini 1.6 mm I.D. 20.8 mL (5 mL/min); Mini 0.8 mm I.D. 19.5 mL2 mL/min).

nd stationary phase retention are related.

. Flow rate and stationary phase retention

It was established that the stationary phase retention percent-ge ratio, Sf,

f = VS

VC(4)

as directly related to the square root of the mobile phase flow ratey [15,16]:

f = A − B√

F (5)

he A and B constants are, respectively the Sf ratio at zero flow rateA is close to 1 or 100% [16]) and B is the slope or gradient of theelationship. Wood et al. established that the B gradient could bexpressed by

= 800

d2C

√�M

KRω2(�L − �U)(6)

n which dC is the internal diameter of the tubing used to make theoil, �M is the mobile phase viscosity, ω is the rotor rotation speed,is the rotor radius, �L is the lower phase density and �U is the

pper phase density and K is a constant that include the � termnd 2 value found in the original Wood equation [16]. This resultas independently confirmed by different authors [17–19].

Table 2 presented lower Sf values for the same experimental con-itions (column, flow rate and rotation speed) and for the Arizona Lystem compared to the Arizona N values. These lower Sf values areelated (Eq. (6)) to a smaller density difference (�� = 0.130 g/mL)etween the two phases of the L system than that of the N�� = 0.146 g/mL [12]). Fig. 3 shows the plots of the Sf factor versushe square root of the flow rate for the four CCC columns tested withhe same liquid system (Arizona L, 2:3:2:3, v/v) at their maximumotation speed. The regression equations as well as the correspond-ng regression coefficients are listed facing their respective lines onhe right side of the figure.

The Wood equation (Eq. (6)) predicted that the results of thewo SFCC columns (1-coil and 3-coil) would be similar at the sameotor rotation (800 rpm) and with the same liquid system. The tub-ng diameter is the same in the two columns and the tubing lengths not a parameter of Eq. (6). The big difference in tubing diameter

pvI8(B

A 1216 (2009) 4169–4175 4173

etween the two Mini CCC columns of similar volume is responsibleor the weak stationary phase retention of the small-bore (0.8 mm)olumn. As soon as the flow rate passes 1 mL/min, this columnetains less than 50% of its volume as stationary phase (Fig. 3). Alsohe retention capability of the large-bore Mini is comparable to thatf the SFCC columns having spools coiled with the similar tubingore. The higher rotor rotation speed of the Mini (ω = 1800 rpm)ompensates for its smaller rotor size (R = 5 cm) and tubing bore1.6 mm) compared to the respective value of 800 rpm, 6 cm and.65 mm for the SFCC columns.

. Tubing bore and resolution

Since Eq. (6) predicts more difficult stationary phase retentiony a smaller-bore (dC) hydrodynamic CCC column, it seems that it isot wise to make CCC columns with small tubing bore. The answer

s not that simple. Tubing bore is related to column efficiency. This isgeneral trend in all types of chromatography early pointed out byiddings [20,21]. Smaller silica particles in LC and smaller capillaryolumn bores in GC give much better efficiency in the chromato-raphic process. CCC is no exception. Smaller tubing bore in CCCoils also give enhanced efficiency with sharper peaks.

Fig. 4 compares the caffeine and vanillin chromatogramsbtained with the two Mini columns. The caffeine peak efficiencyas calculated as 3500 plates for the small-bore column and only00 plates for the large-bore column. Similarly, the vanillin plateounts for the two columns were, respectively 800 and 150 plates.learly, the small-bore column has a much better efficiency thanhe large-bore column.

However, the useful performance parameter in chromatographys not efficiency but resolution. The resolution factor is a compos-te parameter of retention differences (thermodynamic parameter)nd average peak width (kinetic parameter):

s = VR2 − VR1

(W2 + W1)/2(7)

is the peak width measured at base and related to the plate

lotted versus the square root of the lower mobile phase flow rate. CCC columnolumes and rotor rotations: Mini 1.6 mm I.D. 20.8 mL and 1800 rpm; Mini 0.8 mm.D. 19.5 mL and 2100 rpm; SFCC 1-coil 54 mL and 800 rpm; SFCC 3-coil 156 mL and00 rpm. Liquid system Arizona L, heptane–ethyl acetate–methanol–water 2:3:2:3v/v), head to tail flowing direction. The regression equations give A intercepts andslopes (Eq. (5)). The R2 regression coefficient is listed below its equation.

4174 A. Berthod et al. / J. Chromatogr. A 1216 (2009) 4169–4175

Fig. 4. Separation of caffeine and vanillin with the Arizona L heptane–ethyl acetate–methanol–water 2:3:2:3 (v/v) liquid system. Reversed-phase mode: organic uppers tion. T2 i 20.8v nm.u

C

R

t

R

sppbttiralr

ytwTu

alficprafir(

cmtabpCb(b

tationary phase and aqueous lower mobile phase flown in the head to tail direc100 rpm, 2 mL/min. Sf = 38%, Rs = 2.0. Bottom chromatograms: column HPCCC Minolume 0.5 mL, concentration 1 mg/mL in lower mobile phase. Detection UV at 254nits.

ombining Eqs. (1), (4), (7) and (8), Rs can be expressed as

s =√

N

4KD2 − KD1

(1 − Sf )/Sf + (KD2 + KD1)/2(9)

rivially rearranged in:

s = Sf

√N

4KD2 − KD1

1 − Sf [1 − (KD2 + KD1)/2](10)

Eq. (10) shows that the resolution factor increases with Sf, thetationary phase retention ratio and with the square root of N, thelate number. At constant volume, the tubing length is inverselyroportional to the square of the tubing diameter. The plate num-er is likely to increase with the tubing length [22] so this seemso say that the resolution factor will be inversely proportional tohe tubing diameter. This is true at constant Sf only. However, its always possible to run a given CCC column adjusting the flowate so that the desired Sf value is obtained. If time is not a factor,small-bore CCC column will produce a better resolution than a

arger-bore column of the same volume running it at a slow flowate.

Unfortunately time is often an important factor in chemical anal-

sis. The Wood equation (Eq. (6)) predict a much better hold ofhe stationary phase by wider-bore coils than by narrow-bore coilshich means higher flow rate capability with a large-bore coil.

his point was checked injecting the GUESS mixture (minus fer-lic acid and narigenin that, respectively co-eluted with quercetin

utf

s

op chromatograms: column HPCCC Mini 19.6 mL with a 0.8 mm I.D. tubing coil,mL with a 1.6 mm I.D. tubing coil, 1800 rpm, 5 mL/min, Sf = 66%, Rs = 2.0. InjectionLeft chromatograms shown in volume units. Right chromatograms shown in time

nd vanillin) in the four CCC columns using the same Arizona Niquid system. The SFCC column in its best resolutive 3-coil con-guration could separate neither the less retained r, C, Q and Uompounds nor the M and E compounds. The 1-coil configurationrovided even worse resolution. The stationary phase retentionatio Sf could reach significant values as high as 80% at 2 mL/minnd 800 rpm (Fig. 3) but the best observed efficiency (3-coil con-guration) was below 300 plates, too low to produce a competitiveesolution when compared to the Mini modern “high performance”HP) CCC columns [23].

Fig. 5 compares the chromatogram obtained with the 0.8 mmolumn (Fig. 5 top) and a 2 mL/min flow rate and 2100 rpm maxi-um rotor rotation and the similar chromatogram obtained with

he 1.6 mm column (Fig. 5 bottom) and 5 mL/min and 1800 rpmlso maximum rotation. These two chromatograms were selectedecause they both generate similar resolution factors betweeneaks. The insets expand the low resolution region (early eluting, Q, U and V compounds) clearly showing the similar resolutionetween peaks. The 0.8 mm column gives resolution factors of Rs

M, E) 1.1; Rs (E, Z) 2.2 and Rs (Z, O) 3.8. These values are slightlyetter than the resolution factors obtained with the 1.6 mm col-

mn: Rs (M, E) 1.0; Rs (E, Z) 1.6 and Rs (Z, O) 2.6 (Fig. 5). Howeverhe 1.6 mm column chromatogram was completed in 13 min, twiceaster than the 0.8 mm chromatogram (22 min).

In this particular example, the efficiencies experimentally mea-ured on the chromatograms ranged from 1200 to 4000 for the

A. Berthod et al. / J. Chromatogr.

Fig. 5. Separation of nine GUESS solutes with the Arizona N heptane–ethylacetate–methanol–water 1:1:1:1 (v/v) liquid system. Reversed-phase mode: organicupper stationary phase and aqueous lower mobile phase flown in the head to taildirection. Top chromatograms obtained in 22 min: column HPCCC Mini 19.6 mLwith a 0.8 mm I.D. tubing coil, 2100 rpm, 2 mL/min. Sf = 46%. Bottom chromatogramsobtained in 13 min: column HPCCC Mini 20.8 mL with a 1.6 mm I.D. tubing coil,1(e(

07mcsow

6

wIccpSw

stsSsutduiSaprc

A

tsab

R

[

[[

[[[[

[

[

[

800 rpm, 5 mL/min, Sf = 72%. Injection volume 0.5 mL, concentration ∼1 mg/mLeach) in lower mobile phase. Detection UV at 254 nm. The inset chromatograms arenlargements of the early eluting peaks. Peak identification: see Table 1. CarvoneO) retention time is indicated.

.8 mm Mini column (Fig. 5 top) and were only between 150 and00 for the 1.6 mm Mini column (Fig. 5 bottom). In term of chro-atographic resolution, similar resolution factors between the two

hromatograms of Fig. 5 means that the gain in efficiency (four toix times higher efficiencies) obtained by the narrow-bore tubing isffset by the loss of stationary phase retention (respective Sf valuesere 45% versus 72%) as predicted by Eq. (10).

. Conclusion

A small-volume CCC column should be part of all laboratoriesanting to use CCC for preparative separations and purifications.

t allows for the rapid determination of the solute partition coeffi-

ients. Two factors are desirable for a good CCC column. First, theolumn should be able to retain a significant amount of stationaryhase (high Sf values) that will give it a significant resolution power.econd, it should have a high efficiency (high plate number N) thatill even increase the column resolution power. Theory demon-

[

[[

[

A 1216 (2009) 4169–4175 4175

trates that the first requirement is fulfilled using a significant-boreubing, larger than 1 mm, to prepare the coils. However, the oppo-ite, narrow-bore tubing, is required to obtain a high efficiency.ince theory also shows that a high centrifugal field increases thetationary phase retention, the solution is to work with a CCC col-mn of low volume made with coils containing a narrow-boreubing and capable to rotate at very high speed. This is the exactescription of the 0.8 mm I.D. 18 mL Mini hydrodynamic CCC col-mn, commercially called HPCCC [23]. However, in routine work, it

s known that high rotation speeds are prone to mechanical failures.o the 1.6 mm I.D. 19 mL Mini CCC column can also be considered asvery useful machine able to produce a very acceptable resolutionower giving results rapidly because it can work with higher flowates and lower rotation speed compared to the 0.8 mm I.D. CCColumn.

cknowledgements

AB thank University Claude Bernard, Lyon 1, and the French Cen-re National de la Recherche Scientifique for continuous financialupport through UMR5180 (P. Lanteri), Dynamic Extractions Ltd.nd Brunel Institute for Bioengineering for the loan of a 0.8 mmore “HPCCC Mini” CCC column.

eferences

[1] Y. Ito, in: J. Cazes (Ed.), Encyclopedia of Chromatography, Marcel Dekker, NewYork, 2001, p. 241.

[2] Y. Ito, W.D. Conway, High-Speed Countercurrent Chromatography (ChemicalAnalysis, vol. 132), Wiley, New York, 1996.

[3] Y. Ito, Sep. Purif. Rev. 34 (2005) 131.[4] A. Berthod, Countercurrent Chromatography, The Support-free Liquid Station-

ary Phase (Comprehensive Analytical Chemistry, vol. 38), Elsevier, Amsterdam,2002.

[5] A. Berthod, M.J. Ruiz-Angel, S. Carda-Broch, Anal. Chem. 75 (2003) 5886.[6] L.J. Chen, Q. Zhang, G. Yang, L. Fan, J. Tang, I. Garrard, S. Ignatova, D. Fisher, I.A.

Sutherland, J. Chromatogr. A 1142 (2007) 115.[7] I.A. Sutherland, Curr. Opin. Drug Discov. Dev. 10 (2007) 540.[8] 1.Y. Ito, M. Weinstein, I. Aoki, R. Herada, E. Kimura, K. Ninogaki, Nature 212

(1966) 985.[9] L. Janaway, D.W. Hawes, S. Ignatova, P.L. Wood, I.A. Sutherland, J. Liquid Chro-

matogr. Relat. Technol. 26 (2003) 1345.10] D.W. Hawes, L. Janaway, I.A. Sutherland, P.L. Wood, Centrifuge, PCT Int. Appl.,

2003, 27 pp.11] J.B. Friesen, G.F. Pauli, J. Liquid Chromatogr. Relat. Technol. 28 (2005) 2777.12] A. Berthod, M. Hassoun, M.J. Ruiz-Angel, Anal. Bioanal. Chem. 383 (2005)

327.13] A. Berthod, S. Carda-Broch, J. Chromatogr. A 995 (2003) 55.14] S.J. Gluck, E.J. Martin, J. Liquid Chromatogr. 13 (1990) 3559.15] Q. Du, C. Wu, G. Quian, P. Wu, Y. Ito, J. Chromatogr. A 835 (1999) 231.16] P.L. Wood, D. Hawes, L. Janaway, I.A. Sutherland, J. Liquid Chromatogr. Relat.

Technol. 26 (2003) 1373.17] P.S. Fedotov, V.A. Kronrod, T.A. Maryutina, B. Ya Spivakov, Russ. J. Anal. Chem.

57 (2002) 24.18] P.S. Fedotov, V.A. Kronrod, T.A. Maryutina, B. Ya Spivakov, Russ. J. Anal. Chem.

57 (2002) 146.19] C.H. He, C.X. Zhao, AIChE J. 53 (2007) 1460.

20] J.C. Giddings, Dynamics of Chromatography, Marcel Dekker, New York, 1965, p.

293.21] J.C. Giddings, Unified Separation Science, Wiley, New York, 1991.22] I.A. Sutherland, J. De Folter, P. Wood, J. Liquid Chromatogr. Relat. Technol. 26

(2003) 1449.23] H. Guzlek, P. Wood, L. Janaway, J. Chromatogr. A, this issue.