2.5. Update on Moving Bed Technology
2.5.3. Supercritical Fluid-SMB Chromatography
The idea to use pressure gradient in SMB operation derived from fact that each zone of the unit plays different separation role and tuning elution strength in different zones in the unit can enhance separation performance. The working principle in SF-SMB operation is decreasing pressure gradient along the four zones of a SMB unit under supercritical condition to decrease elution strength. It is desired to apply highest pressure in zone I, whose task is to ensure complete regeneration of desorbent, because maximum desorbent strength is needed to elute the more retained component from the mobile phase. The
QR
QI
QD1 QF
Step 1
QR
QX
QD2
Step 2
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
constant of solute decreases as density of supercritical desorbent increases (Luffer et al., 1990) and for non-linear competitive condition (i.e. overloaded column) in which density of supercritical desorbent increases as adsorbed loading decreases. SF-SMB operation enables the use of pressure gradient throughout specific zones of the simulated moving bed thus adsorption is not depending on mobile phase only but also on the applied pressure.
Mostly CO2 is used as mobile phase in subcritical (Fuchs et al., 1992b) and/or supercritical chromatography due to its low viscosity resulting higher column efficiency and ability to perform separation at increased flow rates. Other reasons include its low economical value, non-toxicity and non-flammability. CO2 possess the potential to substitute organic solvents which is relatively more harmful in processing products related to human use i.e. food and pharmaceutical industry. SF-SMB operation also enables product withdrawal free of desorbent as the gaseous desorbent can be recycled after liquefaction in a condenser.
The coupling of supercritical fluid chromatography with SMB chromatography, however, is started by Clavier et al. (1996) when they successfully separate γ-linolenic ethyl ester (GLA) and docosahexaenoic ethyl ester (DHA) on C18 bonded silica by applying pressure gradient from 174 bar (in zone I) to 138 bar (in zone IV). In their operation, it is desired to apply the maximum elution strength in zone I, in which desorption of the more adsorbed component takes place, and minimum elution strength in zone IV, in which the less adsorbed component must be retained by the stationary phase.
They varied the adsorption strength of the mobile phase by playing with pressure in the zones of SMB to enable higher feed load therefore increasing productivity. This is
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
because steeper fronts of internal concentration profile are achieved by pressure gradient operation.
An important factor to evaluate the choice of mobile phase for SF-SMB operation is the dependence of the solvation power to temperature and pressure. In supercritical region, the solvating power is a function of temperature and pressure and so is the affinity of a given solute to that particular supercritical fluid. The affinity of that solute for a given stationary phase is also a function of temperature (Perrut, 1994). The SF-SMB system is judged to be less feasible, relative to isocratic, if the solvation power is independent of temperature and pressure (for linear system) or pressure and modifier concentration (for non linear system). Denet et al. (2001) have shown this phenomena when they use SF- SMB to separate the isomers of tetralol (1,2,3,4-tetrahydro-1-naphtol, a chiral pharmaceutical intermediate. They observed strong impact on separation performance even though the absolute change in selectivity, defined as the ratio of Henry’s constants of the two enantiomers, is relatively small. This is due to low separation factor (about 1.1) assuring SF-SMB is absolutely useful for this kind of separation. Productivity can be increased by almost three times, in comparison with isocratic operation, by rationalizing the unit behavior using triangle theory under linear and non-linear condition.
Rationale between SF-SMB with triangle theory is performed under linear (Mazzotti et al., 1997c) and non-linear systems (Di Giovanni et al., 2001). For linear system, it was found that the pure separation regime for supercritical system is no longer of triangular shape but either a truncated or full rectangle. The size of the regime has been shown to increase indicating that pressure gradient mode is in favor of separation compared to isocratic mode. The coordinates of optimum point (in terms of productivity, solvent
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
consumption, recovery and enrichment) has also shifted from isocratic optimum point in the direction much further away from the diagonal (Mazzotti et al., 1997c). For non-linear systems, different adsorption isotherm must be used in different zones of the unit. Linear isotherm of Henry’s law type is used for low fluid phase concentration while Langmuir competitive isotherm is used for elevated concentration. In the frame of equilibrium theory, it was found that the linear and non-linear pure separation regions have different intersections with the diagonal. The pure separation regime for non-linear condition follows the triangular shape of an isocratic operation, especially at higher feed concentration, while the pure regime for linear condition still has the shape of a rectangle.
The size of this region shrinks with increasing feed concentration as expected indicating that the separation task becomes increasingly difficult at higher feed load (Di Giovanni et al., 2001).
The limitation of this approach is the limited solvating power for elution of polar and large molecules (Schulte and Strube, 2001), particularly when CO2 is used. These molecules are even difficult to be eluted on common stationary phases i.e. those based on silica. Only lipophilic compound exhibit high solubility in pure CO2 but this problem can be handled by adding polar modifier such as alcohols (Fuchs et al., 1992a) although this attempt doesn’t solve the entire problem. The presence of this modifier affects the mobile and stationary phase as it may increase the solvating power of the supercritical fluid while at the same time it can cause deactivation of the most active sites of the adsorbent which is responsible for solute retention. This phenomenon might affect solute retention time and peak shape under supercritical condition (Wright and Smith, 1986).
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
Another example of SF-SMB separation is the separation of phytol (3,7,11,15- tetramethyl-2-hexadecen-l-ol) isomers with trans-isomer is used as fixer in perfume indus- try (Depta et al., 1999). They used a dynamic model for SMB simulation to predict the region of complete separation taking into account different column configuration (4 and 8 total column) and compressibility of the mobile phase. Later, Peper et al. (2002) optimize the separation of R- and S- Ibuprofen using numerical model. The operating condition for the model is based on prior experiment using 40º C and pressure ranges from 17 – 14 MPa. 4.5% (wt) 2-propanol is used as modifier. They are able to increase productivity as high as 504 gracemate/kgsolid/day with 99% raffinate purity. Johannsen et al. (2002) compared the process performance of bi-naphtol enantiomers separation on 10 different stationary phases. They found that the Kromasil CHI-DMB and Chiralcel OJ phase were the most suitable for the bi-naphtol system. Numerical optimization revealed that 6- column configuration (1/2/2/1) was sufficient for this separation. Further simulation and optimization study on phytol isomers using triangle theory leads to enhanced productivity up to 54 gfeed/lsolid/h. The earlier work of Pirkle et al. (1996) in examining some chiral stationary phases extensively over a wide range of temperature and mobile phase additives under sub/supercritical condition leads to better prospect of this system.