Sip

IB

a

AA

KPICCGI

1

(Acid

tptpteaf

MclAiHt

0d

Journal of Chromatography A, 1216 (2009) 8787–8792

Contents lists available at ScienceDirect

Journal of Chromatography A

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

cale-up of counter-current chromatography: Demonstration of predictablesocratic and quasi-continuous operating modes from the test tube toilot/process scale

an Sutherland ∗, Peter Hewitson, Svetlana Ignatovarunel Institute for Bioengineering, Brunel University, Uxbridge, Middlesex UB8 3PH, UK

r t i c l e i n f o

rticle history:vailable online 19 March 2009

a b s t r a c t

Predictable scale-up from test tube derived distribution ratios and analytical-scale sample loading opti-misation is demonstrated using a model sample system of benzyl alcohol and p-cresol in a heptane:ethyl

eywords:rocess scale

ntermittent counter-current extractionounter-current chromatographyCC

acetate:methanol:water phase system with the new 18 L Maxi counter-current chromatography cen-trifuge. The versatility of having a liquid stationary phase with its high loading capacity and flexibleoperating modes is demonstrated at two different scales by separating and concentrating target com-pounds using a mixture of caffeine, vanillin, naringenin and carvone using a quasi-continuous techniquecalled intermittent counter-current extraction.

UESSCcE

. Introduction

One of the key attractions of counter-current chromatographyCCC) is its predictable scale-up from the test tube to pilot scale.nother is its versatility, in that it can be used as an isocratic batchhromatography process as well as a continuous extraction process,ntermittent counter-current extraction (ICcE), as this paper willemonstrate.

Invented by Ito et al. [1] in the 1960s, the technology has gonehrough many stages before emerging today as a robust technologyoised to become an important addition to industry’s separationechnology portfolio. The advantages of having a liquid stationaryhase are numerous [2] – 100% sample recovery; high capacity dueo the large volume of stationary phase; the ability to handle crudextracts and particulates; significantly reduced solvent usage [3,4]nd a reduction in the number of processing steps – to name just aew.

One typical example of process-scale separation using the 4.6 Laxi-CCC centrifuge is the production of glucoraphanin, a cancer

hemoprotective agent derived from broccoli seeds [5,6], whiched to the manufacture of kilogram batches for research purposes.

nother is the separation of honokiol from magnolol, regioisomers

solated from Magnolia officianalis Rehd. et Wils. (Chinese nameoupu) [7], which subsequently formed the basis for a new syn-

hesized product and treatment regime for cancer biotherapy [8].

∗ Corresponding author.E-mail address: ian.sutherland@brunel.ac.uk (I. Sutherland).

021-9673/$ – see front matter © 2009 Elsevier B.V. All rights reserved.oi:10.1016/j.chroma.2009.03.040

© 2009 Elsevier B.V. All rights reserved.

This paper describes how simple test tube partitioning testscombined with sample loading optimisation at the analytical scale,can be transferred directly and predictably to pilot/process scale.It also demonstrates, with a model sample system containing amixture of four compounds, how the process can be used as aquasi-continuous counter-current chromatography process (ICcE)to concentrate one target compound while the others are strippedaway from it.

2. Experimental

2.1. Reagents and materials

All solvents used were of analytical grade and purchased fromFisher Chemicals (Loughborough, UK), deionised water and HPLCwater was purified from a Purite Select Fusion pure water system(Thame, UK). Benzyl alcohol (99%) and p-cresol (4-methylphenol,99%) were supplied by Sigma–Aldrich (Gillingham, UK). Com-pounds caffeine and umbelliferone were supplied by Fisher, whileferulic acid and vanillin were supplied by Sigma–Aldrich.

2.2. Preparation of sample and phase systems

2.2.1. Phase systemsA HEMWat phase system was used for the benzyl alcohol/p-

cresol separation, containing n-heptane, ethyl acetate, methanoland water in a volume ratio of 14:1:5:10. This system is moderatelyhydrophobic and annotated as 4A by Brunel [9].

8 atogr

pvtu

2

odptca

2

icaavAaHtrta

2

2

(3a6as

2

tHoa1a

2

samtatc

788 I. Sutherland et al. / J. Chrom

The GUESSmix separation required two different, more polar,hase systems depending on the separation—HEMWat-15 (2:3:2:3,/v/v/v) or HEMWat-16a (4:5:4:5, v/v/v/v). Due to the large quan-ities of phase system required to carry out the various runs thepper and lower phases were made up separately [10].

.2.2. Sample systemsTwo test mixtures were used in this study. The first one was a mix

f benzyl alcohol and p-cresol dissolved in the 4A lower phase atifferent concentrations. The second was based on the GUESSmix1

roposed by Friesen [11]. The sample used for ICcE separations onhe Midi and the 4.6 L Maxi was a mix of caffeine (C), vanillin (V),arvone (O) and naringenin (N) dissolved in 50/50 (v/v) of uppernd lower phases of either the HEMWat-15 or 16a phase systems.

.3. Test tube partitioning measurements

The procedure used for determination of distribution ratios (Kd)s well established and described in detail by Garrard [12]. A con-entrated solution of the test mixture was made up in methanolnd 30 �L was added to each vial containing 4 mL each of uppernd lower phase of the chosen phase system. The vial was shakenigorously until equilibrium had been established in both phases.fter the phases settled, 0.2 mL of each phase was pipetted into sep-rate HPLC vials, diluted 5-fold with methanol and injected into thePLC column. The distribution ratio/partition coefficient of a par-

icular compound in reversed-phase mode was calculated as theatio of HPLC peak area in the upper (stationary) phase to that ofhe lower (mobile) phase. All Kd values given in the paper are theverage of four sets of experiments.

.4. Apparatus

.4.1. Maxi-CCC centrifugesTwo Maxi-CCC centrifuges supplied by Dynamic Extractions

Slough, UK) were used in the study. Both have a rotor radius of00 mm, have dual bobbins (columns) which can operate in par-llel or series with 10 mm bore tubing and both can rotate up to00 rpm (121 × g). The first has a capacity of 4.6 L and the secondcapacity of 18 L [13]. Typical flow rates are 100–1000 mL/min in

eries or 200–2000 mL/min in parallel.

.4.2. Midi-CCC centrifugeThe Midi-CCC centrifuge (Dynamic Extractions, Slough, UK) and

he setup used for the ICcE study were described in detail byewitson et al. [14]. The Midi-CCC centrifuge has a rotor radiusf 110 mm, tubing bore of 4 mm and two bobbins (columns) withtotal capacity of 912 mL. The Midi can rotate up to a speed of

400 rpm (241 × g), has a typical flow range of 10–100 mL/min andmean ˇ value of 0.75 where ˇ is the ratio of planet to rotor radius.

.4.3. Milli-CCC centrifugeThe Milli-CCC centrifuge used for analytical separations and the

ample loading studies was previously described in detail by Jan-way et al. [15]. It has a rotor radius of 50 mm, tubing bore of 0.8 mm,ean ˇ of 0.74 and a single bobbin of 5.4 mL capacity with a coun-

erweight. The Milli can rotate up to a speed of 2100 rpm (246 × g)

nd has a typical flow range for most organic/aqueous phase sys-ems of 0.5–2 mL/min for a separation, but the Milli column canope with flow rates up to 10 mL/min (w.r.t. pressure) for refilling.

1 Generally useful estimate of sample system (GUESS) in CCC.

. A 1216 (2009) 8787–8792

2.5. Operating system

A fully automated liquid handling pumping and valving system[13] supplied by Armen Instrument (Vannes, France) was used withthe Maxi centrifuge. The system controls the flow of phase systemsand sample solutions while monitoring the pressure; detects the UVabsorbance of the eluent and facilitates remote fraction collection.The four 3-headed HPLC pumps on the system give a non-pulsatileflow of the phases. The system is controlled from a visual displaywhich allows the setting of all parameters (operating pressures,flow rates, flow path – series or parallel, operation mode – iso-cratic normal phase, reversed phase or intermittent, flow switchingtiming and UV detection wavelength) for an isocratic CCC or ICcErun.

When the 4.6 L Maxi was used to run ICcE the sample was loadedusing a separate preparative Knauer K-1800 HPLC pump (Berlin,Germany) to minimize the delay switching between normal andreversed phase operation. The valving was also changed to allow theeluent in reversed phase to pass through a second Knauer K-2501spectrophotometer with a preparative flow cell.

The Midi setup to run ICcE is described in [14]. The Millicentrifuge was set up to run reversed phase counter-current chro-matography [15].

2.6. Operating procedure

2.6.1. Maxi isocratic operating procedure2.6.1.1. Filling the columns and establishing hydrodynamic equilib-rium (4.6 L Maxi). All lines in the Armen liquid handling systemwere initially primed with the appropriate phase system using the“prime mode” of the operating system. The columns were filledwith upper phase at 2000 mL/min (1000 mL/min through each col-umn) from tail (periphery) to head (centre) with the columns inparallel to maintain balance during rotation. After 3 min, the cen-trifuge was rotated at 200 rpm to expel trapped air from the systemfrom the “head” end of the column. The rotational speed wasincreased to its operational speed of 600 rpm (121 × g). Hydrody-namic equilibrium was then established in reversed phase mode,flowing the lower phase from head (centre) to tail (periphery), at1600 mL/min through the columns in parallel (800 mL/min througheach column). Once break through of the lower phase was observedthe flow path through the columns was switched to series and theflow rate was set to 850 mL/min. Lower phase was pumped until nomore loss of stationary upper phase (carry over) was observed. Thecentrifuge was equilibrated at 30 C.

2.6.1.2. Filling the columns and establishing hydrodynamic equi-librium (18 L Maxi). Filling the 18 L Maxi and establishinghydrodynamic equilibrium was carried out in a similar manner tothe 4.6 L centrifuge, except in the filling stage, the column was filledwith upper phase at 2500 mL/min in parallel (to fill it more quickly)with the centrifuge rotated at only 75 rpm. Once full the rotationalspeed was increased to 600 rpm (121 × g) and hydrodynamic equi-librium was established in reversed phase. The flow rate of lowerphase was increased from 800 mL/min initially to 1200 mL/minafter 4.5 min and to 1600 mL/min after 7 min with the flow paththrough the columns in parallel. With such a long high-resolutioncolumn, experience has shown that better stationary phase reten-tion is achieved if mobile phase flow is ramped up gradually. Oncebreak through had occurred the path through the columns wasswitched to series and the flow rate was set to 850 mL/min. Lower

phase was pumped until no more carry over of stationary phasewas observed.

2.6.1.3. Separation and fraction collection (4.6 L and 18 L Maxi). Oncestable hydrodynamic equilibrium had been established the sample

atogr

wt(sea1

22rwlcwhtdtrtt5

2ccprn4tntiafwt4

2

wwaatfito

2

trspwwtrg

I. Sutherland et al. / J. Chrom

as injected through the Armen sample pump at 850 mL/min. Oncehe correct volume of sample had been pumped into the centrifugeby measuring the change of volume in the supply cylinder) theystem was switched to pump lower phase at 850 mL/min. The elu-nt was monitored with a UV detector (set at 230 nm) for all runsnd fractions were collected at 30 s intervals for the 4.6 L Maxi andmin intervals for the 18 L Maxi for analysis by HPLC.

.6.2. Maxi ICcE operating procedure

.6.2.1. Filling the columns and establishing hydrodynamic equilib-ium for ICcE (4.6 L Maxi). The columns were filled, from empty,ith upper phase as described above for isocratic operation. Equi-

ibration was performed in reversed phase mode, aiming to fill theolumns with a 50/50 mix of upper and lower phase. The centrifugeas rotated at 350 rpm (41 × g) and lower phase was flowed fromead (centre) to tail (periphery) at 500 mL/min with the flow pathhrough the bobbins in parallel. Once the volume of upper phaseisplaced was near 50% of the column volume, the centrifuge rota-ional speed was increased to 600 rpm (121 × g) and then the flowate was increased to 1200 mL/min to initiate break through. Theotal volume of upper phase displaced was measured to confirmhe initial volume of each phase in the columns was approximately0%.

.6.2.2. Intermittent counter-current extraction (ICcE) and fractionollection (4.6 L Maxi). The principle of intermittent counter-urrent extraction has been described previously [14]. The Armenumping system allows for the switching between normal andeversed phase to be automatically set up. The ICcE run started inormal phase with the upper phase mobile at 250 mL/min. Aftermin the flow was automatically switched to reversed phase with

he lower phase mobile at 250 mL/min for 4 min. This cycle betweenormal and reversed phase was then repeated for another 14 cycleso give a total run time of 120 min. While the system was runningn normal phase the sample dissolved in upper phase was loadednd vice versa, both at 50 mL/min. Sample loading was carried outor the first 20 min of the run. The upper and lower phase eluents

ere monitored with UV detectors set at 230 and 220 nm, respec-ively. Fractions were collected, alternately from each end, everymin throughout the extraction for analysis by HPLC.

.7. Theoretical model

The theoretical model developed by de Folter and coworkers [16]as used. Test tube distribution ratios of the compounds togetherith the value of stationary phase retention, obtained for the oper-

ting conditions used during the CCC operating procedure, are useds input parameters. The number of equivalent mixing and settlingransfers can be estimated in one of two ways—either by curvetting to the chromatogram produced by a standard of known dis-

ribution ratio or matching the resolution between two compoundsf known distribution ratios.

.8. Theoretical plates

In this paper the number of Craig counter-current distributionransfers [17] has been used in preference to theoretical plates. Theeason for this is that the number of theoretical plates changesignificantly with the volume ratio (Vm/Vs). In conventional solidhase chromatography, where Vm/Vs ∼ 13, a plate count of >11,000

ould be needed to achieve a resolution of 1.5 for two compoundsith an alpha of 2, whereas this could be achieved with about 300

heoretical plates in CCC where Vm/Vs was unity (Sf = 50%). In aecent review of CCC by Berthod [18] he illustrates how resolutionoes down markedly as retention of the stationary phase reduces

. A 1216 (2009) 8787–8792 8789

from 90 to 50%, whereas an analysis of the theoretical plates showsthat these values can go up or remain predominantly constant.

2.9. HPLC fraction analysis

The first HPLC method was developed for the analysis of ben-zyl alcohol and p-cresol mix. HPCCC fractions from the Maxi 4.6 Land 18 L centrifuges collected every 30 s and 1 min, respectivelywere diluted 25 times in both cases for further analysis. HPLC wasconducted using a Waters 2695 HPLC system with a Waters 2996photodiode array detection (DAD) system. Isocratic HPLC methodwas run with 45% aqueous ACN as the mobile phase, at 1 mL/minfor 2.5 min, thermostatted at 40 C on a Symmetry C18 column(75 mm × 4.6 mm I.D. 3.5 �m).

Two HPLC methods were developed for the GUESSmix analysis.The first method is described before in [14] for the ICcE Midi run.The second HPLC method, for the ICcE runs on Maxi, is a short-ened version of the previous method. It used a 50–90% gradientof aqueous methanol in 5 min at 1 mL/min on Sunfire C18 column(150 mm × 4.6 mm I.D., 5 �m) thermostatted at 30 C, with the samedilutions used in the first method.

3. Results and discussion

3.1. Milli-CCC sample loading studies

Test tube studies show that the distribution ratios for ben-zyl alcohol and p-cresol were 0.3 and 1.1, respectively, with theHEMWat 4A phase system. Sample loading studies were performedusing the 5.4 mL capacity Milli analytical CCC at 1 mL/min first vary-ing sample volume at a concentration of 10.5 mg/mL benzyl alcoholand 5 mg/mL p-cresol. It was found that resolution started drop-ping off quickly when sample volumes exceeded 5% of the columnvolume (>270 �L) and so 5% was chosen as the optimum percent-age of the column volume loaded. Sample concentration was thenincreased at this percentage column volume until the resolutionfell below 1.5. A final sample loading of 5% column volume and42 mg/mL benzyl alcohol and 20 mg/mL p-cresol was chosen forpilot-scale Maxi runs. A summary of these test results is given inTable 1.

3.2. Scale-up from 5.4 mL capacity Milli centrifuge to 4.6 L Maxicentrifuge

The volume ratio between the Milli (5.4 mL) and Maxi (4.6 L)centrifuges is approximately 850–1. The flow rate used on the Maxiwas therefore chosen to be 850 mL/min based on Wood’s volumet-ric scale-up principle [19] with all other parameters kept the same.Note that sample volume is normalised by injecting 5% of the col-umn volume, the same percentage as used for the Milli-CCC. Fig. 1shows the UV chromatogram obtained from scaling up this modelseparation with the analytical chromatogram inset. While the peakretention times for benzyl alcohol and p-cresol are similar (4.5 and6.5 min, respectively) the resolution between them fell from 1.54(Milli 5.4 mL) to 1.01 (4.6 L Maxi). Note that there was some lossof stationary phase as the first peak eluted which made the chro-matogram very “spiky” for the first 4 min after which it stabilised.The position of the actual chromatogram is shown dotted.

3.3. Check on chosen sample loading strategy

Before scaling up to the 18 L centrifuge, a check was run onwhether it was better to load a given mass in a low volume but highconcentration or in a high volume but low concentration. The resultof this experiment is shown in Fig. 2, where the chromatogramsfrom the HPLC fraction analysis are shown. Note that the relative

8790 I. Sutherland et al. / J. Chromatogr. A 1216 (2009) 8787–8792

Table 1Resolution between benzyl alcohol and p-cresol as sample loading is optimised using the milli-CCC centrifuge. Operating conditions: column volume 5.4 mL; tubing bore0.8 mm; rotor speed 2100 rpm (246 × g); mobile phase flow 1 ml/min; reverse phase.

Sample concentration Resolution (Rs)

BA (mg/mL) PC (mg/mL) Sample volume (percentage of column volume)

1.26% 2.5% 5.1% 9.9% 12.5% 15.7%

10.5 5 1.81 1.83 1.71 1.43 1.33 1.2121 10 1.6436.5 15 1.6242 20 1.5452.5 25 1.42

F the Mfl s for tc at (14:

saafv9ibds

Fth9t

ig. 1. 4.6 L Maxi separation of benzyl alcohol and p-cresol scaled up 850× fromow 850 mL/min; sample load, 290 mL (6.3% Vc) 12.2 g BA, 5.8 g PC. Run conditiononcentration and proportion of column volume as for Maxi. Phase system: HEMW

ize of the benzyl alcohol and p-cresol peaks is different from Fig. 1s HPLC peak area is now being measured instead of UV absorbancend the relative sensitivities for benzyl alcohol and p-cresol are dif-erent. When the chromatogram from the low concentration/higholume loading (12.8 mg/mL benzyl alcohol + 6.1 mg/mL p-cresol in

60 mL) is compared to the high concentration/low volume load-

ng (42 mg/mL benzyl alcohol + 20 mg/mL p-cresol in 290 mL) it cane seen that the peaks are much broader and the resolution hasropped from 1.01 to 0.83. The high sample concentration, lowample volume strategy was therefore chosen.

ig. 2. 4.6 L Maxi separation of benzyl alcohol and p-cresol comparing high concen-ration low volume (42 mg/mL benzyl alcohol + 20 mg/mL p-cresol in 290 mL) withigh volume low concentration (12.8 mg/mL benzyl alcohol + 6.1 mg/mL p-cresol in60 mL) with approximately the same mass loaded in each case. Speed and flow arehe same as for Fig. 1.

illi-CCC separation (inset). Operating conditions for 4.6 L Maxi: speed 600 rpm,he Milli-CCC: speed 2100 rpm, flow 1 mL/min; sample loading condition the same1:5:10, v/v/v/v). Sf before injection 80.0%, after separation 40.8% for the Maxi run.

3.4. Theoretical prediction from test tube distribution ratios andstationary phase retention

Once distribution ratios of target compounds have beenobtained from HPLC test tube studies (Sections 2.2 and 3.1) itis possible to predict peak elution profiles once stationary phaseretention (Sf) and resolution (Rs) are known [16,20]. This model hasbeen applied in Fig. 3a and compared with the HPLC chromatogramfrom the high concentration/low volume loading (42 mg/mL ben-zyl alcohol + 20 mg/mL p-cresol in 290 mL) run on the 4.6 L Maxicentrifuge. The number of equivalent mixing and settling transfersbased on Craig’s counter-current distribution [17] is 36. When scal-ing this up to the larger capacity 18 L Maxi centrifuge, the numberof transfers was increased by the ratio of the column volumes (4×)to 144. Fig. 3b shows that this gives a very good match to the resultobtained in practice.

3.5. Intermittent counter-current extraction optimisation usingthe Midi centrifuge

Hewitson et al. [14] have already described the process of ICcE.They give an example of an optimisation process using the Midi-CCCcentrifuge with the GUESSmix (C, V, N and O) using the HEMWat-

16a phase system.

The Midi optimisation process reduced run times from 280 to86 min and increased throughputs from 46 mg/h to 7.9 g/h. Thisfinal optimisation run (Fig. 4a) had the sample loaded over fourcycles for a period of 16 min alternatively loaded in upper and lowerphase.

I. Sutherland et al. / J. Chromatogr. A 1216 (2009) 8787–8792 8791

Fig. 3. Scale-up from the 4.6 L to the 18 L Maxi: (a) 4.6 L Maxi separation of benzylalcohol and p-cresol comparing theory with practice. The theoretical prediction isbased on distribution ratios from test tube tests and the stationary phase retentionfor the operating conditions of Fig. 1; (b) 18 L Maxi separation of benzyl alcohol and p-cresol comparing theory with practice. Operating conditions: speed 600 rpm, flow850 mL/min; sample concentration 42 mg/mL benzyl alcohol, 20 mg/mL p-cresol;sfp4

3M

ibT5tttwl

iwptwHnttrtttle

Fig. 4. Quasi-continuous counter-current-chromatography scale-up from (a) 912 mLMidi to (b) 4.6 L Maxi. Operating conditions for Midi [14]: speed, 1250 rpm, upperand lower phase flow, 60 mL/min, time interval, 4 min; sample load 11.2 g total in

contaminants are stripped away from it.Another strategy is demonstrated in Fig. 5, where a slight change

of phase system to heptane:ethyl acetate:methanol:water (4:5:4:5,v/v/v/v) keeps caffeine being continuously extracted with the lower

ample volume 290 mL. The theoretical prediction is based on distribution ratiosrom test tube tests and the stationary phase retention (Sf) for the given mobilehase flow rate. Sf values before injection and at the end of the run were (a) 80.0,0.8% and (b) 83.8, 52.7%, respectively.

.6. Intermittent counter-current extraction scale-up to the 4.6 Laxi centrifuge

Direct scale-up to the 4.6 L Maxi-CCC was performed by increas-ng the maximum flow rate (elution pump + sample loading pump)y the volume ratio between the two centrifuges (i.e. a factor 5).his set the elution pump to 250 mL/min and the sample pump to0 mL/min. For the Midi run the phase system flow was reduced ashe sample was loaded to keep the flow rate at 60 mL/min. Withhe Maxi-CCC run it was not possible in the control programmeo change the flow while running, so the flows were 300 mL/minhile the sample was loaded and these reduced to 250 mL/min after

oading.Fig. 4b shows the resulting chromatogram from the Maxi-CCC

ntermittent counter-current extraction run. This compares wellith the Midi run except vanillin is retained longer in Maxi com-

ared to Midi and about 55% of the vanillin was retained inhe column. For all the runs using the GUESSmix compounds,hich have relatively high separation factors (for example, for theEMWat 16a system ˛c–v = 6.1, ˛v–n = 2.3 and ˛n–o = 5.9), it wasoted that the peak resolutions achieved were significantly lesshan might be expected for an isocratic run. It is speculated thathis may be due to the disruption of the hydrodynamic equilib-ium after each switch of flow direction within the columns and

he time then required re-establishing the equilibrium. If this ishe case maximising the time between switching would minimisehe time for which the hydrodynamic equilibrium was disturbed. Aonger column would also allow the switching times to be furtherxtended. Of course if the system is being used to split compounds

14 min. Operating conditions for Maxi: speed, 600 rpm, upper and lower phase flow,250 mL/min; sample loading 40.5 g in 20 min. Phase system: HEMWat (4:5:4:5,v/v/v/v). Distribution ratios: caffeine (0.09); vanillin (0.55); naringenin (1.25) andcarvone (7.39).

into two eluent streams then this decrease in peak resolution is lessimportant.

This demonstrates, if tuned properly, ICcE can be used very effec-tively to concentrate a compound of interest in the column while

Fig. 5. Quasi-continuous counter-current-chromatography separation on the 4.6 LMaxi centrifuge demonstrating the elution of one compound from the “tail” whileall other compounds are eluted from the “head”. Operating conditions for Maxi:speed, 600 rpm, upper and lower phase flow, 250 mL/min; sample loading 40.5 gin 20 min. Phase system: HEMWat (2:3:2:3, v/v/v/v). Distribution ratios: caffeine(0.14); vanillin (1.21); naringenin (3.82) and carvone (14.8).

8 atogr

pr

4

csccatvtAt

A

(1cacstMtLd

[

1449.[17] L.C. Craig, O. Post, Anal. Chem. 21 (1949) 500.

792 I. Sutherland et al. / J. Chrom

hase from the “tail” while carvone, naringenin and vanillin areemoved with the upper phase from the “head” end of the column.

. Conclusions

These preliminary feasibility studies have shown that the pro-ess is scalable from test tube distribution ratios and sample loadingtudies performed at an analytical scale. They have also shownonventional two-bobbin CCC instruments can be used for quasi-ontinuous counter-current extraction, increasing the versatilitynd range of operation of this exciting new technology. Intermit-ent counter-current extraction though introduces another set ofariables: the flow ratio, the flow difference and the respectiveime intervals between switching from normal to reversed phase.

detailed in depth study is now needed to give this exciting newechnology the knowledge base it needs to be commercialised.

cknowledgements

The authors would like to thank the Research Councils EPSRCGrant ref: GR/R03143/01) and BBSRC (Grant refs: BBD524583/1;92/SBRI9675) for their support of research on the scale-up ofounter-current chromatography and helping to equip the centre,nd Pfizer for supporting the research on the intermittent counter-urrent extraction (ICcE). Thanks also to Brunel University for itsupport via West Focus, SRIF2 and HEIF4 which has helped fund

he development of the Advanced Bioprocessing Centre and the

axi-CCC centrifuges. Finally the authors are extremely gratefulo Dynamic Extractions Ltd. and in particular Dr. Philip Wood andee Janaway for their dedication in producing such an innovativeesign of centrifuge which is the first of its kind in the world.

[

. A 1216 (2009) 8787–8792

References

[1] Y. Ito, M. Weinstein, I. Aoki, R. Harada, E. Kimura, K. Nunogaki, Nature 212 (1966)985.

[2] A. Berthod, Counter-Current Chromatography—The Support-Free Liquid Sta-tionary Phase (Comprehensive Analytical Chemistry, vol. 38), Elsevier,Amsterdam, 2002.

[3] A.S. Graham, I.F. McConvey, P. Shering, J. Liq. Chromatogr. Relat. Technol. 24(2001) 1811.

[4] D. Keay, Manuf. Chemist (September 2008) 72.[5] K.L. Wade, K.K. Stephenson, F.E. Chou, J. Chromatogr. A 996 (2003) 85.[6] D. Fisher, I.J. Garrard, R. van den Heuvel, I.A. Sutherland, F.E. Chou, J.W. Fahey, J.

Liq. Chromatogr. Relat. Technol. 28 (2005) 1913.[7] L. Chen, Q. Zhang, G. Yang, L. Fan, I. Garrard, S. Ignatova, D. Fisher, I.A. Sutherland,

J. Chromatogr. A 1142 (2007) 115.[8] Y. Luo, Y. Xu, L. Chen, H. Luo, C. Peng, J. Fu, H. Chen, A. Peng, H. Ye, D.-C. Xie, A.

Fu, J. Shi, S. Yang, Y. Wei, J. Chromatogr. A 1178 (2007) 160.[9] I.A. Sutherland, A. Booth, L. Brown, B. Kemp, H. Kidwell, D. Games, A.S. Graham,

G.G. Guillon, D. Hawes, M. Hayes, L. Janaway, G. Lye, P. Massey, C. Preston, P.Shering, T. Shoulder, C. Strawson, P. Wood, J. Liq. Chromatogr. Relat. Technol. 24(2001) 1533.

[10] I. Garrard, L. Janaway, D. Fisher, J. Liq. Chromatogr. Relat. Technol. 30 (2007)151.

[11] J.B. Friesen, G.F. Pauli, J. Liq. Chromatogr. Relat. Technol. 28 (2005) 2777.12] I.J. Garrard, J. Liq. Chromatogr. Relat. Technol. 28 (2005) (1923).

[13] I.A. Sutherland, P. Hewitson, S. Ignatova, J. Chromatogr. A 1216 (2009)4201–4205.

[14] P. Hewitson, S. Ignatova, H. Ye, L. Chen, I. Sutherland, J. Chromatogr. A 1216(2009) 4187–4192.

[15] L. Janaway, D. Hawes, S.N. Ignatova, P. Wood, I.A. Sutherland, J. Liq. Chromatogr.Relat. Technol. 26 (2003) 1345.

[16] I.A. Sutherland, J. de Folter, P. Wood, J. Liq. Chromatogr. Relat. Technol. 26 (2003)

[18] A. Berthod, Adv. Chromatogr. 47 (2009) 323.[19] P. Wood, S. Ignatova, L. Janaway, D. Keay, D. Hawes, I. Garrard, I.A. Sutherland, J.

Chromatogr. A 1151 (2007) 25.20] J. de Folter, I.A. Sutherland, J. Chromatogr. A 1216 (2009) 4218–4224.