- Email: [email protected]
Using the liquid nature of the stationary phase in countercurrent chromatography
Journal of Chromatography A, 1116 (2006) 143–148
Using the liquid nature of the stationary phase in countercurrent chromatography IV. The cocurrent CCC method Alain Berthod ∗ , Mahmoud Hassoun Laboratoire des Sciences Analytiques, Universit´e Claude Bernard-Lyon 1, Bat. CPE, 60622 Villeurbanne, France Received 27 October 2005; received in revised form 10 March 2006; accepted 13 March 2006 Available online 3 April 2006
Abstract The retention volumes of solutes in countercurrent chromatography (CCC) are directly proportional to their distribution coefficients, KD in the biphasic liquid system used as mobile and stationary phase in the CCC column. The cocurrent CCC method consists in putting the liquid “stationary” phase in slow motion in the same direction as the mobile phase. A mixture of five steroid compounds of widely differing polarities was used as a test mixture to evaluate the capabilities of the method with the biphasic liquid system made of water/methanol/ethyl acetate/heptane 6/5/6/5 (v/v) and a 53 mL CCC column of the coil planet centrifuge type. It is shown that the chromatographic resolution obtained in cocurrent CCC is very good because the solute band broadening is minimized as long as the solute is located inside the “stationary” phase. Pushing the method at its limits, it is demonstrated that the five steroids can still be (partly) separated when the flow rate of the two liquid phases is the same (2 mL/min). This is due to the higher volume of upper phase (72% of the column volume) contained inside the CCC column producing a lower linear speed compared to the aqueous lower phase linear speed. The capabilities of the cocurrent CCC method compare well with those of the gradient elution method in HPLC. Continuous detection is a problem due to the fact that two immiscible liquid phases elute from the column. It was partly solved using an evaporative light scattering detector. © 2006 Elsevier B.V. All rights reserved. Keywords: Countercurrent chromatography; Liquid stationary phase; Solute band broadening; Steroids; Chromatographic resolution
1. Introduction Countercurrent chromatography (CCC) is a liquid chromatography (LC) technique using a stationary phase that is also a liquid. No solid support is involved. All modern CCC instruments use centrifugal fields to hold the liquid stationary phase [1]. To avoid confusion, it is very important to point out from the beginning that there is no countercurrent circulation in CCC. The name of the technique comes from the Craig countercurrent distribution machine [2]. Indeed, CCC is not really a recent technique since it was introduced in the late 1960s by Ito et al. [3]. In almost four decades, an enormous amount of work was done in this field and the CCC naming got widely accepted [4–6]. Most of the published work proposes original and/or difficult separations and/or purifications of industrial, biological and especially,
∗
Corresponding author. Tel.: +33 472431434; fax: +33 472431078. E-mail address: [email protected] (A. Berthod).
0021-9673/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2006.03.031
natural products. Indeed, CCC is a preparative technique since the solutes have access to the volume of the liquid stationary phase and not just its surface. The stationary phase overload inconvenience in LC with a solid stationary phase is much less a problem in CCC with its liquid stationary phase [1,4–6]. We focused on the unique advantages of the liquid state of the stationary phase in CCC. In the first work of this series of articles, we demonstrated that it was possible to make complexation reaction within the liquid stationary phase to separate cations [7]. In a second article, we demonstrated that it was possible to use a CCC machine as a chemical reactor [8]. In a third article, we introduced the elution-extrusion method that consists in extruding the CCC column content to elute highly retained compounds [9]. In this work, the idea of a slow motion of the liquid stationary phase in the same direction as the mobile phase is revisited. The “slowly moving” stationary phase idea and part of its theory was introduced twenty years ago by Sutherland et al. [10]. The theory of the “cocurrent CCC” method was fully devel-
144
A. Berthod, M. Hassoun / J. Chromatogr. A 1116 (2006) 143–148
oped by Berthod [11] and used to measure with an acceptable precision the octanol–water partition coefficient of hydrophobic compounds [12]. An isolated co-current chromatography study using small bore tubing with a special high viscosity phase system managed to achieve a limited co-current separation by retaining the more viscous stationary phase by wetting the tubing wall in a non-centrifugal unit gravity application [13]. “True moving bed” chromatography was introduced with the solid stationary phase moving countercurrent to the mobile phase in a discontinuous way [14]. The cocurrent CCC method could be described also as “truly” moving bed chromatography since the stationary phase is slowly moving in the same direction as the mobile phase. This could be a first step in using the liquid character of the stationary phase in view of continuous separations following the “simulated” moving bed (SMB) method with classical solid stationary phase [14–16]. 2. Material and methods 2.1. Chemicals The biphasic liquid system was a mixture of methanol, ethyl acetate, water and heptane. All three organic solvents were obtained from SDS (Peypin, France, a Carlo Erba division). Water was deionized and distilled. It had a resistivity passing 18 M cm−1 . The composition used throughout the work was the mixture water/methanol/ethyl acetate/heptane 6/5/6/5 (v/v) also called Arizona System M [17]. This four solvent mixture splits in 44.5% of a heptane rich upper phase (density 0.768) and 55.6% of aqueous lower phase (d = 0.906). The upper phase composition is water/methanol/ethyl acetate/heptane 0.7/2.8/41.8/54.7% (v/v). The lower phase composition is water/methanol/ethyl acetate/heptane 46.9/36.7/16.3/0.07% (v/v) [16]. The test solutes were five steroid compounds: prednisone with a log Po/w of 1.62, prednisolone acetate (log Po/w = 2.40), testosterone (log Po/w = 3.32), estrone (log Po/w = 3.13) and cholesterol (log Po/w = 8.74). They were all five obtained from Sigma–Aldrich (L’Isles d’Abeau Chesnes, France). 2.2. Material The CCC apparatus was the SFCC 2000 of (Soci´et´e Franc¸aise de Chromatographie et Colonnes, Paris, France) whose production has been discontinued. It is a three-coil planet centrifuge following the classical Ito scheme IV [18]. It contains three multilayer spools connected in series and spinning around a central axis. Each spool is made by winding 26 m of 1/8 in. (3.2 mm O.D.) Teflon tubing (1.6 mm I.D.) in seven layers of 19 turns each making a volume of 53 mL per spool. The total apparatus volume is 158 mL but it can work with one spool only. The full cocurrent CCC set-up is shown by Fig. 1. Two pumps (Shimadzu LC6A, Kyoto, Japan) are needed: one to drive the mobile phase and the second one to drive the “stationary” phase. The two flows are joined before solute injection with a 5025 Rheodyne valve (Cotaty, USA). An evaporative light scattering detector, Cunow Model DDL 21 (Eurosep, Cergy Pontoise,
Fig. 1. The cocurrent CCC set-up. Two pumps are needed, one for each phase. The evaporative light scattering detector is able to give a signal with a nonhomogeneous biphasic liquid system.
France) was used for continuous detection. Nitrogen was used at a flow rate of 2 L/min to vaporize continuously the biphasic mobile phase. The temperature of the drying chamber was set at 80 ◦ C. The high voltage of the photomultiplier tube was set at 620 V. The signal was collected and processed by a Shimadzu CR5A integrator or a computer. 3. The cocurrent CCC method 3.1. The general idea The method presented here is a chromatographic method with a stationary and a mobile phase. The cocurrent term arises due to the slow motion of the “stationary” phase. The method can be compared to the moving walkways in airports. A person walking quietly on the moving walkway can pass somebody trotting along the sidewalk. His slow walking speed adds to the speed of the moving walkway and the resulting speed can be higher that that of the passenger jogging on the sidewalk. The cocurrent CCC method uses the fact that the stationary phase is a liquid. It is possible to push it slowly in the same direction as the mobile phase. The result will be that no compound can be trapped inside the column. The most retained compound sticks to the stationary phase. It will eventually elute since the stationary phase slowly moves toward the column exit. 3.2. Theoretical foundations It was demonstrated that the cocurrent CCC retention equation was [11]: VM + KD VS VR = (FM + FS )tR = (FM + FS ) (1) FM + K D FS in which tR is the solute retention time; F is the flow rate of the mobile phase (subscript M) and “stationary” phase (subscript S) (since there is actually no stationary phase, S could stand for slower phase); VR , VM and VS are respectively the solute retention volume and the mobile and “stationary” (slower) phase volumes inside the column; KD is the solute distribution ratio
A. Berthod, M. Hassoun / J. Chromatogr. A 1116 (2006) 143–148
145
between the “stationary” (slower) phase and the mobile phase. KD is very often called the solute partition coefficient. It can be expressed as: KD =
tR FM − VM . VS − t R FS
(2)
Since the two liquid phases are not miscible, a biphasic liquid system enters and exits the CCC column. Continuous detection becomes a serious problem. It was first tried to add a clarifying agent miscible in both phases and homogenizing them. 2-Propanol was used as a clarifying agent with the heptanemethanol-water system [12] needing a third pump. The evaporative light scattering detector uses a nebulizer to spray the liquid entering the detector. The spayed mobile phase is dried by heat in the evaporative chamber and the dried solid solutes can be detected by light scattering. This detector is not sensitive to mobile phase composition changes. However, it is sensitive to change in the solute concentration in the eluting phase. 3.3. Band broadening in cocurrent CCC In the development of the elution extrusion method, it was demonstrated that the broadening of the solute bands inside the column was dependent only on the location of the band inside the column [9,19]: Vem Wins = 4VC (3) VRi N In Eq. (3), Wins is the width of the band (peak width at base expressed in volume units) at a given position inside the column; VC is the CCC “column” volume (VC = VM + VS ); Vem is the volume of mobile phase used to push the solute at the given position inside the column; VRi is the classical solute retention volume (VRi = VM + KD VS ) and N is the column plate number. In cocurrent CCC, the stationary phase is also in motion. It can be considered that all solutes will see a column shorter than it actually is; exactly like the traveler on the moving walkway does not feel that the corridor is as long as it actually is. Eq. (3) will give Wins , the width of the solute band inside but at the outlet of the column taking Vem as the mobile phase volume used to move the solute up to the column outlet. Next, Wb , the solute band width outside the column, will be a combination of Wins , the solute band width inside the column, with M%, the mobile phase, and S%, the “stationary” phase volume needed to elute it: Wb = Wins [S% + M% (1 + (KD − 1)Sf)]
(4)
Fig. 2 shows the change in retention volume (top) and peak width (bottom) as a function of the “stationary” phase flow rate, FS , for five test compounds with a wide range of polarity, 0.12 (prednisone)
Fig. 2. Change in steroid retention volumes (top) and peak width at base (bottom) when the “stationary” phase flow rate change. Liquid system: water/methanol/ethyl acetate/heptane 6/5/6/5 (v/v). Mobile phase: lower aqueous phase at a constant flow rate 2 mL/min. “Stationary” phase: upper phase at the indicated FS flow rate. Machine volume VC = 53 mL. Rotor rotation speed: 800 rpm. Detection by ELSD.
retention volume and peak width, respectively. Solutes having a high affinity for the organic upper phase are highly retained in classical CCC (e.g. cholesterol, KD = 40, VR = 1.46 L, Table 1). As soon as the “stationary” upper phase moves, their retention volumes decrease dramatically [12]. The positive point added by this work is that the peak width decreases as well. This allows maintaining a very acceptable resolution decreasing drastically the separation time. Fig. 3 shows five computer-smoothed chromatograms of the same steroid test mixture obtained with five different upper phase flow rates. 3.4. Amount of each phase inside the machine The main problem in CCC is to retain the maximum amount of liquid stationary phase inside the column. All modern CCC “columns” use centrifugal forces to maintain the stationary phase [1,4–6]. The amount of stationary phase retained inside hydrodynamic machine is measured by the Sf factor: Sf =
VS VC
(5)
the British Brunel CCC team extensively studied the variation of Sf, the stationary phase retention factor, versus many parameters such as mobile phase flow rate, rotor rotation speed, coiled
Upper (mobile) phase flow rate: 2 mL/min; machine volume: 53 mL; rotor rotation: 800 rpm; VR : retention volume; tR : retention time, S: percentage of the VR retention volume made of “stationary” upper phase; Wb : peak width at base; N = 16 (VR /Wb )2 ; Rs: resolution factor.
150 270 440 680 2200 8.3 29.6 51.2 77.5 96.8 1.7 1.6 2.0 1.8 0 0 0 0 0 0.12 0.56 1.4 4.6 40 Prednisone Prednisolone acetate Testosterone Estrone Cholesterol
21.3 37.2 67 183 1460
10.6 18.6 33.5 91.5 730
8.6 12 22.5 60 460
140 150 140 150 160
1.55 1.9 2.8 6.2
25.9 40.7 62.4 106 166
10.3 16.3 25.0 42.5 66.2
2.9 12.3 26 53.5 91
7 12 17 27 36
220 180 220 250 340
34.2 45.8 57.5 71.8 82.3
9.8 13.1 16.4 20.5 23.5
11 11 11 11 7
N plates Wb (mL) S (%) tR (min) VR (mL) Rs N plates S (%) KD Solute
VR (mL)
tR (min)
Wb (mL)
N plates
Rs
VR (mL)
tR (min)
S (%)
Wb (mL)
FS = 1.5 (mL/min) FS = 0.5 (mL/min) Classical elution (no S flow) Upper phase flow rate
Table 1 Chromatographic figures of merit for the separation of steroid compounds by cocurrent CCC
1.5 1.2 1.2 0.7
A. Berthod, M. Hassoun / J. Chromatogr. A 1116 (2006) 143–148 Rs
146
Fig. 3. Chromatograms of the steroid mixture obtained with five different upper phase flow rates. Liquid system: water/methanol/ethyl acetate/heptane 6/5/6/5 (v/v). The peak order is: (1) prednisone; (2) prednisolone acetate; (3) testosterone; (4) estrone and (5) cholesterol. Mobile phase: lower aqueous phase at a constant flow rate 2 mL/min. “Stationary” phase: upper phase at the indicated FS flow rate. Machine volume VC = 53 mL. Rotor rotation speed: 800 rpm. Detection by ELSD.
tube diameter and machine volume [20–22]. However, it may be anticipated that the cocurrent CCC method may change the amount of “stationary” phase retained inside the CCC machine. The Sf retention factor was studied with a constant aqueous (mobile) phase flow rate of 2 mL/min and different organic (stationary) upper phase flow rates. In classical elution CCC (true stationary phase), Wood et al. firmly established that the Sf factor was linearly decreasing with the square root of the mobile phase flow rate. We found that Sf, the amount of organic upper phase retained in our CCC machine was also related to the square root of Ftot , the total flow inside the CCC machine that is the “stationary” phase FS flow rate added to the constant FM flow rate (2 mL/min). Table 2 lists the slopes, intercepts and regression coefficient of the Sf versus (FM + FS )1/2 measured at different rotor speeds. Fig. 4 shows the Sf retention factor plotted versus the square of the rotation speed. This value is somewhat proportional to the centrifugal force inside the rotating coil. It is recalled that the centrifugal field inside the coil planet centrifuge CCC column is not constant. However, a faster rotor rotation induces necessarily a higher centrifugal field. From a practical point of view, it makes sense that more upper organic phase be retained inside the “column” compared to the classical CCC elution scheme since Table 2 Slope, intercept and regression coefficient of the organic upper phase retention factor, Sf, versus the square root of the total flow rate Rotor rotation (rpm)
Slope
Intercept
r2
400 600 700 800 900
27.0 19.5 16.0 11.2 4.8
12.7 30.5 39.1 50.2 62.0
0.994 0.981 0.973 0.988 0.996
Liquid system: water/methanol/ethyl acetate/heptane 6/5/6/5 (v/v); lower phase flow rate: 2 mL/min; upper phase flow rate varied between 0 to 2 mL/min, both phase entering in the column head.
A. Berthod, M. Hassoun / J. Chromatogr. A 1116 (2006) 143–148
147
Fig. 4. Amount of organic upper phase inside the CCC machine, expressed as Sf (VS /VC ), plotted versus ω2 , the square of the rotor rotation speed related to the mean centrifugal field. Lower phase flow rate = 2 mL/min, upper phase flow rate as indicated by the right values in the figure caption.
there is a second pump continuously feeding the CCC apparatus with that phase. However, it was found that the amount of upper organic phase becomes almost independent of the upper organic phase flow rate when a sufficiently high centrifugal field is applied to the CCC coil. In our experiments, the Sf factor was 69% at 900 rpm in classical CCC elution with a true upper organic stationary phase (FS = 0) and only 71% when the upper organic stationary phase was pushed at 2 mL/min, the same flow rate as the lower aqueous phase (Fig. 4). This later point deserves some comments. Fig. 3 (bottom chromatogram) shows that it is possible to separate partly five steroids in a CCC machine when the two immiscible phases enter and leave the “column” at exactly the same flow rates (2 mL/min). This is due to the ability of the CCC column to retain the upper organic phase inside the coil (Sf = 71% at 800 rpm). Inside the CCC column, the linear speeds of the two phases are different. The column content is 37.6 mL upper organic phase (71% of VC = 53 mL) and 15.4 mL of lower aqueous phase. Then the upper organic phase needs 18.8 min to move from the head of the column to its tail (37.6 mL over 2 mL/min) when the lower aqueous phase needs only 7.7 min to travel the same distance. Inside the CCC column, the aqueous lower phase is more than twice as fast as the organic upper phase. This explains why solutes separation can occur as experimenttally observed (Fig. 3) when the flow rates of the two phases are exactly the same. 3.5. The detection problem Fig. 5 is the scan of an actual experiment. It is observed that the later eluting peaks are detected in a dramatically increasing noise. In cocurrent CCC, two immiscible liquid phases exit out of the column. Most classical chromatographic detectors cannot handle this situation. In previous work, a clarifying agent was added post-column. Propanol [12] or methanol [18] were used as solvent added post-column to solubilize the two liquid phases exiting the CCC column. Solvent recycling was not possible. Only detectors evaporating the mobile phase solvents
Fig. 5. Actual chromatogram of the separation of 5 steroids by cocurrent CCC. Liquid system: water/methanol/ethyl acetate/heptane 6/5/6/5 (v/v). Mobile phase: lower aqueous phase, flow rate 2 mL/min; “stationary” phase: upper phase at 0.5 mL/min flow rate. Machine volume VC = 53 mL. Rotor rotation speed: 800 rpm. Detection ELSD (80 ◦ C, 620 V). The peak order is: (1) prednisone (0.32 mg); (2) prednisolone acetate (0.34 mg); (3) testosterone (0.42 mg); (4) estrone (1.5 mg) and (5) cholesterol (1.1 mg). Injection volume 200 L of the steroids in lower phase.
can be used for continuous detection and solvent recycling in cocurrent CCC. The mass spectrometer with electrospray ionization and the evaporative light scattering detection (ELSD) system fulfill this condition. The later ELSD system was used to obtain Fig. 5. With no exception, the ELSD signal was relatively stable when only the two immiscible phases exited the column (nothing injected). After a solute injection, the beginning of the chromatograms was relatively noiseless. However, the most retained compounds (estrone and cholesterol) were always detected in an important noise (Fig. 5). An electronic smoothing software was used to obtain the Fig. 3 chromatograms. The noise is due to local apolar solute heterogeneity in the eluting biphasic system. The apolar solute (KD 1) is essentially located in the upper organic phase. During the evaporation process, there is no solute in the lower aqueous phase (=no signal) and all the solute concentrated in the upper phase (=maximum signal). The observed noise corresponds in fact to the succession of spikes each corresponding to an upper phase microdroplet containing the solute. This detection problem was also encountered with the octanol/water biphasic liquid system [12].
148
A. Berthod, M. Hassoun / J. Chromatogr. A 1116 (2006) 143–148
4. Conclusion A mixture of steroids with widely differing polarities can be separated by cocurrent CCC as it would be separated using a gradient of mobile phase composition in HPLC. It was demonstrated that it is possible to use the difference in the two immiscible phase flow rates to separate solutes with a significant resolution. The gain in separation duration and solvent consumption can be tremendous. The possibilities of the method can even be extended by varying the flow rate of the slower phase during the experiment. If the detection problem can be overcome, it can be possible to start a difficult separation using the classical way (true stationary phase) and then to use a gradient of “stationary” phase flow rate to maximize the resolution of the separation of the most retained compounds. References [1] Countercurrent Chromatography, the Support Free Liquid Stationary Phase, in: A. Berthod (Ed.), in: Comprehensive Analytical Chemistry, vol. 38, Elsevier, Amsterdam, 2002. [2] Y. Ito, Foreword in: A., Berthod, (Ed.), Countercurrent Chromatography, the Support Free Liquid Stationary Phase. (Comprehensive Analytical Chemistry, vol. 38), Elsevier, Amsterdam, 2002, pp. xix–xx. [3] Y. Ito, M.A. Weinstein, I. Aoki, R. Harada, E. Kimura, K. Nunogaki, Nature 212 (1966) 985–987. [4] N.B. Mandava, Y. Ito, CCC, Theory and Practice (Chromatographic Science Series), vol. 44, M. Dekker, New York, 1985.
[5] W.D. Conway, Countercurrent Chromatography, Apparatus, Theory & Applications, VCH, Weinheim, 1990. [6] Y. Ito, W.D. Conway (Eds.), High Speed Countercurrent Chromatography, Chemical Analysis, vol. 132, Wiley, New York, 1996. [7] K. Talabardon, M. Gagean, J.M. Mermet, A. Berthod, J. Liq. Chromatogr. Rel. Technol. 21 (1998) 3003–3019. [8] A. Berthod, K. Talabardon, S. Caravieilhes, C. De Bellefon, J. Chromatogr. A. 828 (1998) 523–530. [9] A. Berthod, M.J. Ruiz-Angel, S. Carda-Broch, Anal. Chem. 75 (2003) 5886–5894. [10] I.A. Sutherland, D.H. Waddington, T.J. Peters, J. Liq. Chromatogr. 7 (1984) 363–376. [11] A. Berthod, Analusis 18 (1990) 352–357. [12] A. Berthod, R.A. Menges, D.W. Armstrong, J. Liq. Chromatogr. 15 (1992) 2769–2785. [13] C.A. Lucy, B.P. Hausermann, Anal. Chim. Acta 307 (1995) 173–183. [14] E.R. Francotte, P. Richert, J. Chromatogr. A. 769 (1997) 101–107. [15] E.R. Francotte, Simulated Moving Bed: Basics and Applications, European Meeting, I.N.P.L., L’Institut National Polytechnique de Lorraine, Nancy, 1993. [16] T. Yun, G. Zhong, G. Guiochon, AIChE J. 43 (1997) 2970–2983. [17] A. Berthod, M. Hassoun, M.J. Ruiz-Angel, Anal. Bioanal. Chem. 383 (2005) 327–340. [18] Y. Ito, J. Chromatogr. 214 (1981) 122–125. [19] A. Berthod, M. Hassoun, G. Harris, J. Liq. Chromatogr. Rel. Technol. 28 (2005) 1851–1866. [20] P.L. Wood, D. Hawes, L. Janaway, I.A. Sutherland, J. Liq. Chromatogr. Rel. Technol. 26 (2003) 1373–1396. [21] P.L. Wood, D. Hawes, L. Janaway, I.A. Sutherland, J. Liq. Chromatogr. Rel. Technol. 26 (2003) 1417–1430. [22] P.L. Wood, I.A. Sutherland, J. Liq. Chromatogr. Rel. Technol. 24 (2001) 1699–1710.
Using the liquid nature of the stationary phase in countercurrent chromatography IV. The cocurrent CCC method Alain Berthod ∗ , Mahmoud Hassoun Laboratoire des Sciences Analytiques, Universit´e Claude Bernard-Lyon 1, Bat. CPE, 60622 Villeurbanne, France Received 27 October 2005; received in revised form 10 March 2006; accepted 13 March 2006 Available online 3 April 2006
Abstract The retention volumes of solutes in countercurrent chromatography (CCC) are directly proportional to their distribution coefficients, KD in the biphasic liquid system used as mobile and stationary phase in the CCC column. The cocurrent CCC method consists in putting the liquid “stationary” phase in slow motion in the same direction as the mobile phase. A mixture of five steroid compounds of widely differing polarities was used as a test mixture to evaluate the capabilities of the method with the biphasic liquid system made of water/methanol/ethyl acetate/heptane 6/5/6/5 (v/v) and a 53 mL CCC column of the coil planet centrifuge type. It is shown that the chromatographic resolution obtained in cocurrent CCC is very good because the solute band broadening is minimized as long as the solute is located inside the “stationary” phase. Pushing the method at its limits, it is demonstrated that the five steroids can still be (partly) separated when the flow rate of the two liquid phases is the same (2 mL/min). This is due to the higher volume of upper phase (72% of the column volume) contained inside the CCC column producing a lower linear speed compared to the aqueous lower phase linear speed. The capabilities of the cocurrent CCC method compare well with those of the gradient elution method in HPLC. Continuous detection is a problem due to the fact that two immiscible liquid phases elute from the column. It was partly solved using an evaporative light scattering detector. © 2006 Elsevier B.V. All rights reserved. Keywords: Countercurrent chromatography; Liquid stationary phase; Solute band broadening; Steroids; Chromatographic resolution
1. Introduction Countercurrent chromatography (CCC) is a liquid chromatography (LC) technique using a stationary phase that is also a liquid. No solid support is involved. All modern CCC instruments use centrifugal fields to hold the liquid stationary phase [1]. To avoid confusion, it is very important to point out from the beginning that there is no countercurrent circulation in CCC. The name of the technique comes from the Craig countercurrent distribution machine [2]. Indeed, CCC is not really a recent technique since it was introduced in the late 1960s by Ito et al. [3]. In almost four decades, an enormous amount of work was done in this field and the CCC naming got widely accepted [4–6]. Most of the published work proposes original and/or difficult separations and/or purifications of industrial, biological and especially,
∗
Corresponding author. Tel.: +33 472431434; fax: +33 472431078. E-mail address: [email protected] (A. Berthod).
0021-9673/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2006.03.031
natural products. Indeed, CCC is a preparative technique since the solutes have access to the volume of the liquid stationary phase and not just its surface. The stationary phase overload inconvenience in LC with a solid stationary phase is much less a problem in CCC with its liquid stationary phase [1,4–6]. We focused on the unique advantages of the liquid state of the stationary phase in CCC. In the first work of this series of articles, we demonstrated that it was possible to make complexation reaction within the liquid stationary phase to separate cations [7]. In a second article, we demonstrated that it was possible to use a CCC machine as a chemical reactor [8]. In a third article, we introduced the elution-extrusion method that consists in extruding the CCC column content to elute highly retained compounds [9]. In this work, the idea of a slow motion of the liquid stationary phase in the same direction as the mobile phase is revisited. The “slowly moving” stationary phase idea and part of its theory was introduced twenty years ago by Sutherland et al. [10]. The theory of the “cocurrent CCC” method was fully devel-
144
A. Berthod, M. Hassoun / J. Chromatogr. A 1116 (2006) 143–148
oped by Berthod [11] and used to measure with an acceptable precision the octanol–water partition coefficient of hydrophobic compounds [12]. An isolated co-current chromatography study using small bore tubing with a special high viscosity phase system managed to achieve a limited co-current separation by retaining the more viscous stationary phase by wetting the tubing wall in a non-centrifugal unit gravity application [13]. “True moving bed” chromatography was introduced with the solid stationary phase moving countercurrent to the mobile phase in a discontinuous way [14]. The cocurrent CCC method could be described also as “truly” moving bed chromatography since the stationary phase is slowly moving in the same direction as the mobile phase. This could be a first step in using the liquid character of the stationary phase in view of continuous separations following the “simulated” moving bed (SMB) method with classical solid stationary phase [14–16]. 2. Material and methods 2.1. Chemicals The biphasic liquid system was a mixture of methanol, ethyl acetate, water and heptane. All three organic solvents were obtained from SDS (Peypin, France, a Carlo Erba division). Water was deionized and distilled. It had a resistivity passing 18 M cm−1 . The composition used throughout the work was the mixture water/methanol/ethyl acetate/heptane 6/5/6/5 (v/v) also called Arizona System M [17]. This four solvent mixture splits in 44.5% of a heptane rich upper phase (density 0.768) and 55.6% of aqueous lower phase (d = 0.906). The upper phase composition is water/methanol/ethyl acetate/heptane 0.7/2.8/41.8/54.7% (v/v). The lower phase composition is water/methanol/ethyl acetate/heptane 46.9/36.7/16.3/0.07% (v/v) [16]. The test solutes were five steroid compounds: prednisone with a log Po/w of 1.62, prednisolone acetate (log Po/w = 2.40), testosterone (log Po/w = 3.32), estrone (log Po/w = 3.13) and cholesterol (log Po/w = 8.74). They were all five obtained from Sigma–Aldrich (L’Isles d’Abeau Chesnes, France). 2.2. Material The CCC apparatus was the SFCC 2000 of (Soci´et´e Franc¸aise de Chromatographie et Colonnes, Paris, France) whose production has been discontinued. It is a three-coil planet centrifuge following the classical Ito scheme IV [18]. It contains three multilayer spools connected in series and spinning around a central axis. Each spool is made by winding 26 m of 1/8 in. (3.2 mm O.D.) Teflon tubing (1.6 mm I.D.) in seven layers of 19 turns each making a volume of 53 mL per spool. The total apparatus volume is 158 mL but it can work with one spool only. The full cocurrent CCC set-up is shown by Fig. 1. Two pumps (Shimadzu LC6A, Kyoto, Japan) are needed: one to drive the mobile phase and the second one to drive the “stationary” phase. The two flows are joined before solute injection with a 5025 Rheodyne valve (Cotaty, USA). An evaporative light scattering detector, Cunow Model DDL 21 (Eurosep, Cergy Pontoise,
Fig. 1. The cocurrent CCC set-up. Two pumps are needed, one for each phase. The evaporative light scattering detector is able to give a signal with a nonhomogeneous biphasic liquid system.
France) was used for continuous detection. Nitrogen was used at a flow rate of 2 L/min to vaporize continuously the biphasic mobile phase. The temperature of the drying chamber was set at 80 ◦ C. The high voltage of the photomultiplier tube was set at 620 V. The signal was collected and processed by a Shimadzu CR5A integrator or a computer. 3. The cocurrent CCC method 3.1. The general idea The method presented here is a chromatographic method with a stationary and a mobile phase. The cocurrent term arises due to the slow motion of the “stationary” phase. The method can be compared to the moving walkways in airports. A person walking quietly on the moving walkway can pass somebody trotting along the sidewalk. His slow walking speed adds to the speed of the moving walkway and the resulting speed can be higher that that of the passenger jogging on the sidewalk. The cocurrent CCC method uses the fact that the stationary phase is a liquid. It is possible to push it slowly in the same direction as the mobile phase. The result will be that no compound can be trapped inside the column. The most retained compound sticks to the stationary phase. It will eventually elute since the stationary phase slowly moves toward the column exit. 3.2. Theoretical foundations It was demonstrated that the cocurrent CCC retention equation was [11]: VM + KD VS VR = (FM + FS )tR = (FM + FS ) (1) FM + K D FS in which tR is the solute retention time; F is the flow rate of the mobile phase (subscript M) and “stationary” phase (subscript S) (since there is actually no stationary phase, S could stand for slower phase); VR , VM and VS are respectively the solute retention volume and the mobile and “stationary” (slower) phase volumes inside the column; KD is the solute distribution ratio
A. Berthod, M. Hassoun / J. Chromatogr. A 1116 (2006) 143–148
145
between the “stationary” (slower) phase and the mobile phase. KD is very often called the solute partition coefficient. It can be expressed as: KD =
tR FM − VM . VS − t R FS
(2)
Since the two liquid phases are not miscible, a biphasic liquid system enters and exits the CCC column. Continuous detection becomes a serious problem. It was first tried to add a clarifying agent miscible in both phases and homogenizing them. 2-Propanol was used as a clarifying agent with the heptanemethanol-water system [12] needing a third pump. The evaporative light scattering detector uses a nebulizer to spray the liquid entering the detector. The spayed mobile phase is dried by heat in the evaporative chamber and the dried solid solutes can be detected by light scattering. This detector is not sensitive to mobile phase composition changes. However, it is sensitive to change in the solute concentration in the eluting phase. 3.3. Band broadening in cocurrent CCC In the development of the elution extrusion method, it was demonstrated that the broadening of the solute bands inside the column was dependent only on the location of the band inside the column [9,19]: Vem Wins = 4VC (3) VRi N In Eq. (3), Wins is the width of the band (peak width at base expressed in volume units) at a given position inside the column; VC is the CCC “column” volume (VC = VM + VS ); Vem is the volume of mobile phase used to push the solute at the given position inside the column; VRi is the classical solute retention volume (VRi = VM + KD VS ) and N is the column plate number. In cocurrent CCC, the stationary phase is also in motion. It can be considered that all solutes will see a column shorter than it actually is; exactly like the traveler on the moving walkway does not feel that the corridor is as long as it actually is. Eq. (3) will give Wins , the width of the solute band inside but at the outlet of the column taking Vem as the mobile phase volume used to move the solute up to the column outlet. Next, Wb , the solute band width outside the column, will be a combination of Wins , the solute band width inside the column, with M%, the mobile phase, and S%, the “stationary” phase volume needed to elute it: Wb = Wins [S% + M% (1 + (KD − 1)Sf)]
(4)
Fig. 2 shows the change in retention volume (top) and peak width (bottom) as a function of the “stationary” phase flow rate, FS , for five test compounds with a wide range of polarity, 0.12 (prednisone)
Fig. 2. Change in steroid retention volumes (top) and peak width at base (bottom) when the “stationary” phase flow rate change. Liquid system: water/methanol/ethyl acetate/heptane 6/5/6/5 (v/v). Mobile phase: lower aqueous phase at a constant flow rate 2 mL/min. “Stationary” phase: upper phase at the indicated FS flow rate. Machine volume VC = 53 mL. Rotor rotation speed: 800 rpm. Detection by ELSD.
retention volume and peak width, respectively. Solutes having a high affinity for the organic upper phase are highly retained in classical CCC (e.g. cholesterol, KD = 40, VR = 1.46 L, Table 1). As soon as the “stationary” upper phase moves, their retention volumes decrease dramatically [12]. The positive point added by this work is that the peak width decreases as well. This allows maintaining a very acceptable resolution decreasing drastically the separation time. Fig. 3 shows five computer-smoothed chromatograms of the same steroid test mixture obtained with five different upper phase flow rates. 3.4. Amount of each phase inside the machine The main problem in CCC is to retain the maximum amount of liquid stationary phase inside the column. All modern CCC “columns” use centrifugal forces to maintain the stationary phase [1,4–6]. The amount of stationary phase retained inside hydrodynamic machine is measured by the Sf factor: Sf =
VS VC
(5)
the British Brunel CCC team extensively studied the variation of Sf, the stationary phase retention factor, versus many parameters such as mobile phase flow rate, rotor rotation speed, coiled
Upper (mobile) phase flow rate: 2 mL/min; machine volume: 53 mL; rotor rotation: 800 rpm; VR : retention volume; tR : retention time, S: percentage of the VR retention volume made of “stationary” upper phase; Wb : peak width at base; N = 16 (VR /Wb )2 ; Rs: resolution factor.
150 270 440 680 2200 8.3 29.6 51.2 77.5 96.8 1.7 1.6 2.0 1.8 0 0 0 0 0 0.12 0.56 1.4 4.6 40 Prednisone Prednisolone acetate Testosterone Estrone Cholesterol
21.3 37.2 67 183 1460
10.6 18.6 33.5 91.5 730
8.6 12 22.5 60 460
140 150 140 150 160
1.55 1.9 2.8 6.2
25.9 40.7 62.4 106 166
10.3 16.3 25.0 42.5 66.2
2.9 12.3 26 53.5 91
7 12 17 27 36
220 180 220 250 340
34.2 45.8 57.5 71.8 82.3
9.8 13.1 16.4 20.5 23.5
11 11 11 11 7
N plates Wb (mL) S (%) tR (min) VR (mL) Rs N plates S (%) KD Solute
VR (mL)
tR (min)
Wb (mL)
N plates
Rs
VR (mL)
tR (min)
S (%)
Wb (mL)
FS = 1.5 (mL/min) FS = 0.5 (mL/min) Classical elution (no S flow) Upper phase flow rate
Table 1 Chromatographic figures of merit for the separation of steroid compounds by cocurrent CCC
1.5 1.2 1.2 0.7
A. Berthod, M. Hassoun / J. Chromatogr. A 1116 (2006) 143–148 Rs
146
Fig. 3. Chromatograms of the steroid mixture obtained with five different upper phase flow rates. Liquid system: water/methanol/ethyl acetate/heptane 6/5/6/5 (v/v). The peak order is: (1) prednisone; (2) prednisolone acetate; (3) testosterone; (4) estrone and (5) cholesterol. Mobile phase: lower aqueous phase at a constant flow rate 2 mL/min. “Stationary” phase: upper phase at the indicated FS flow rate. Machine volume VC = 53 mL. Rotor rotation speed: 800 rpm. Detection by ELSD.
tube diameter and machine volume [20–22]. However, it may be anticipated that the cocurrent CCC method may change the amount of “stationary” phase retained inside the CCC machine. The Sf retention factor was studied with a constant aqueous (mobile) phase flow rate of 2 mL/min and different organic (stationary) upper phase flow rates. In classical elution CCC (true stationary phase), Wood et al. firmly established that the Sf factor was linearly decreasing with the square root of the mobile phase flow rate. We found that Sf, the amount of organic upper phase retained in our CCC machine was also related to the square root of Ftot , the total flow inside the CCC machine that is the “stationary” phase FS flow rate added to the constant FM flow rate (2 mL/min). Table 2 lists the slopes, intercepts and regression coefficient of the Sf versus (FM + FS )1/2 measured at different rotor speeds. Fig. 4 shows the Sf retention factor plotted versus the square of the rotation speed. This value is somewhat proportional to the centrifugal force inside the rotating coil. It is recalled that the centrifugal field inside the coil planet centrifuge CCC column is not constant. However, a faster rotor rotation induces necessarily a higher centrifugal field. From a practical point of view, it makes sense that more upper organic phase be retained inside the “column” compared to the classical CCC elution scheme since Table 2 Slope, intercept and regression coefficient of the organic upper phase retention factor, Sf, versus the square root of the total flow rate Rotor rotation (rpm)
Slope
Intercept
r2
400 600 700 800 900
27.0 19.5 16.0 11.2 4.8
12.7 30.5 39.1 50.2 62.0
0.994 0.981 0.973 0.988 0.996
Liquid system: water/methanol/ethyl acetate/heptane 6/5/6/5 (v/v); lower phase flow rate: 2 mL/min; upper phase flow rate varied between 0 to 2 mL/min, both phase entering in the column head.
A. Berthod, M. Hassoun / J. Chromatogr. A 1116 (2006) 143–148
147
Fig. 4. Amount of organic upper phase inside the CCC machine, expressed as Sf (VS /VC ), plotted versus ω2 , the square of the rotor rotation speed related to the mean centrifugal field. Lower phase flow rate = 2 mL/min, upper phase flow rate as indicated by the right values in the figure caption.
there is a second pump continuously feeding the CCC apparatus with that phase. However, it was found that the amount of upper organic phase becomes almost independent of the upper organic phase flow rate when a sufficiently high centrifugal field is applied to the CCC coil. In our experiments, the Sf factor was 69% at 900 rpm in classical CCC elution with a true upper organic stationary phase (FS = 0) and only 71% when the upper organic stationary phase was pushed at 2 mL/min, the same flow rate as the lower aqueous phase (Fig. 4). This later point deserves some comments. Fig. 3 (bottom chromatogram) shows that it is possible to separate partly five steroids in a CCC machine when the two immiscible phases enter and leave the “column” at exactly the same flow rates (2 mL/min). This is due to the ability of the CCC column to retain the upper organic phase inside the coil (Sf = 71% at 800 rpm). Inside the CCC column, the linear speeds of the two phases are different. The column content is 37.6 mL upper organic phase (71% of VC = 53 mL) and 15.4 mL of lower aqueous phase. Then the upper organic phase needs 18.8 min to move from the head of the column to its tail (37.6 mL over 2 mL/min) when the lower aqueous phase needs only 7.7 min to travel the same distance. Inside the CCC column, the aqueous lower phase is more than twice as fast as the organic upper phase. This explains why solutes separation can occur as experimenttally observed (Fig. 3) when the flow rates of the two phases are exactly the same. 3.5. The detection problem Fig. 5 is the scan of an actual experiment. It is observed that the later eluting peaks are detected in a dramatically increasing noise. In cocurrent CCC, two immiscible liquid phases exit out of the column. Most classical chromatographic detectors cannot handle this situation. In previous work, a clarifying agent was added post-column. Propanol [12] or methanol [18] were used as solvent added post-column to solubilize the two liquid phases exiting the CCC column. Solvent recycling was not possible. Only detectors evaporating the mobile phase solvents
Fig. 5. Actual chromatogram of the separation of 5 steroids by cocurrent CCC. Liquid system: water/methanol/ethyl acetate/heptane 6/5/6/5 (v/v). Mobile phase: lower aqueous phase, flow rate 2 mL/min; “stationary” phase: upper phase at 0.5 mL/min flow rate. Machine volume VC = 53 mL. Rotor rotation speed: 800 rpm. Detection ELSD (80 ◦ C, 620 V). The peak order is: (1) prednisone (0.32 mg); (2) prednisolone acetate (0.34 mg); (3) testosterone (0.42 mg); (4) estrone (1.5 mg) and (5) cholesterol (1.1 mg). Injection volume 200 L of the steroids in lower phase.
can be used for continuous detection and solvent recycling in cocurrent CCC. The mass spectrometer with electrospray ionization and the evaporative light scattering detection (ELSD) system fulfill this condition. The later ELSD system was used to obtain Fig. 5. With no exception, the ELSD signal was relatively stable when only the two immiscible phases exited the column (nothing injected). After a solute injection, the beginning of the chromatograms was relatively noiseless. However, the most retained compounds (estrone and cholesterol) were always detected in an important noise (Fig. 5). An electronic smoothing software was used to obtain the Fig. 3 chromatograms. The noise is due to local apolar solute heterogeneity in the eluting biphasic system. The apolar solute (KD 1) is essentially located in the upper organic phase. During the evaporation process, there is no solute in the lower aqueous phase (=no signal) and all the solute concentrated in the upper phase (=maximum signal). The observed noise corresponds in fact to the succession of spikes each corresponding to an upper phase microdroplet containing the solute. This detection problem was also encountered with the octanol/water biphasic liquid system [12].
148
A. Berthod, M. Hassoun / J. Chromatogr. A 1116 (2006) 143–148
4. Conclusion A mixture of steroids with widely differing polarities can be separated by cocurrent CCC as it would be separated using a gradient of mobile phase composition in HPLC. It was demonstrated that it is possible to use the difference in the two immiscible phase flow rates to separate solutes with a significant resolution. The gain in separation duration and solvent consumption can be tremendous. The possibilities of the method can even be extended by varying the flow rate of the slower phase during the experiment. If the detection problem can be overcome, it can be possible to start a difficult separation using the classical way (true stationary phase) and then to use a gradient of “stationary” phase flow rate to maximize the resolution of the separation of the most retained compounds. References [1] Countercurrent Chromatography, the Support Free Liquid Stationary Phase, in: A. Berthod (Ed.), in: Comprehensive Analytical Chemistry, vol. 38, Elsevier, Amsterdam, 2002. [2] Y. Ito, Foreword in: A., Berthod, (Ed.), Countercurrent Chromatography, the Support Free Liquid Stationary Phase. (Comprehensive Analytical Chemistry, vol. 38), Elsevier, Amsterdam, 2002, pp. xix–xx. [3] Y. Ito, M.A. Weinstein, I. Aoki, R. Harada, E. Kimura, K. Nunogaki, Nature 212 (1966) 985–987. [4] N.B. Mandava, Y. Ito, CCC, Theory and Practice (Chromatographic Science Series), vol. 44, M. Dekker, New York, 1985.
[5] W.D. Conway, Countercurrent Chromatography, Apparatus, Theory & Applications, VCH, Weinheim, 1990. [6] Y. Ito, W.D. Conway (Eds.), High Speed Countercurrent Chromatography, Chemical Analysis, vol. 132, Wiley, New York, 1996. [7] K. Talabardon, M. Gagean, J.M. Mermet, A. Berthod, J. Liq. Chromatogr. Rel. Technol. 21 (1998) 3003–3019. [8] A. Berthod, K. Talabardon, S. Caravieilhes, C. De Bellefon, J. Chromatogr. A. 828 (1998) 523–530. [9] A. Berthod, M.J. Ruiz-Angel, S. Carda-Broch, Anal. Chem. 75 (2003) 5886–5894. [10] I.A. Sutherland, D.H. Waddington, T.J. Peters, J. Liq. Chromatogr. 7 (1984) 363–376. [11] A. Berthod, Analusis 18 (1990) 352–357. [12] A. Berthod, R.A. Menges, D.W. Armstrong, J. Liq. Chromatogr. 15 (1992) 2769–2785. [13] C.A. Lucy, B.P. Hausermann, Anal. Chim. Acta 307 (1995) 173–183. [14] E.R. Francotte, P. Richert, J. Chromatogr. A. 769 (1997) 101–107. [15] E.R. Francotte, Simulated Moving Bed: Basics and Applications, European Meeting, I.N.P.L., L’Institut National Polytechnique de Lorraine, Nancy, 1993. [16] T. Yun, G. Zhong, G. Guiochon, AIChE J. 43 (1997) 2970–2983. [17] A. Berthod, M. Hassoun, M.J. Ruiz-Angel, Anal. Bioanal. Chem. 383 (2005) 327–340. [18] Y. Ito, J. Chromatogr. 214 (1981) 122–125. [19] A. Berthod, M. Hassoun, G. Harris, J. Liq. Chromatogr. Rel. Technol. 28 (2005) 1851–1866. [20] P.L. Wood, D. Hawes, L. Janaway, I.A. Sutherland, J. Liq. Chromatogr. Rel. Technol. 26 (2003) 1373–1396. [21] P.L. Wood, D. Hawes, L. Janaway, I.A. Sutherland, J. Liq. Chromatogr. Rel. Technol. 26 (2003) 1417–1430. [22] P.L. Wood, I.A. Sutherland, J. Liq. Chromatogr. Rel. Technol. 24 (2001) 1699–1710.