Successful CPC purification is the result of a carefully developed method rather than a single experimental decision. After a promising biphasic solvent system has been identified, the next step is to understand how the target compound and its impurities behave within that system under realistic operating conditions.

Experimental determination of partition behavior, selection of the appropriate elution mode and optimization of key operating parameters all play critical roles in transforming a promising solvent system into a robust and reproducible separation method. By combining suitable mode selection with practical operating-window optimization, manufacturers can establish a reliable method development process that balances resolution, recovery, productivity and robustness.

After selecting candidate solvent systems (see post about Choosing the right solvent system), partition coefficients must be measured experimentally. The shake-flask method is the most common starting point because it provides a simple and rapid way to estimate how the compound of interest and its impurities distribute between the two phases.  

In a typical shake-flask experiment: 
1) prepare and equilibrate the biphasic solvent system
2) transfer equal volumes of the upper and lower phases into a vessel
3) add a known amount of crude sample
4) shake or vortex the mixture vigorously to establish equilibrium
5) allow complete phase separation
6) collect samples from both phases 
7) dilute, evaporate or otherwise prepare the phase samples if required for analytical compatibility
8) analyze both phases, usually by HPLC

The partition coefficient often expressed as K_d, is calculated from the concentration ratio of a compound between the two phases. It is important to define clearly how K_d is calculated, for example as the concentration in the upper phase divided by the concentration in the lower phase, or alternatively as the concentration in the stationary phase divided by the concentration in the mobile phase. For efficient CPC separation, the compound of interest should generally fall within a practical K_d window, often around 0.5–2.0. However, the ideal range depends on the K_d definition, the selected operating mode, the required resolution and the acceptable cycle time. If the K_d is too low, the compound elutes too quickly with limited retention. If the K_d is too high, retention becomes excessive, peaks broaden and cycle time increases. In addition to the K_d of the target compound, the selectivity between the CoI and impurities must also be evaluated, because purification depends not only on retention but on differential partitioning. A solvent system with a practical K_d for the target may still fail if the critical impurities show similar partition behavior.

Besides K_d determination, the shake-flask test also provides important practical information about the solvent system itself. Fast settling time, clear phase separation, sufficient sample solubility and a suitable phase ratio are essential, because slow phase separation, emulsion formation, incomplete sample dissolution or highly unbalanced phase volumes can indicate poor operational stability and may lead to stationary phase loss and inefficient CPC performance. 

The shake-flask method is therefore not a final proof of separation, but a critical decision-making tool. It helps identify which solvent systems are worth transferring into real CPC operation.

CPC can operate in two primary modes:
– Ascending mode (ASC)
– Descending mode (DSC)

The difference between these modes is based on which liquid phase is used as the mobile phase and which phase is retained in the rotor as the stationary phase.

Ascending mode (ASC)	Descending mode (DSC)
Less dense phase – mobile phase	Denser phase – mobile phase
Denser phase – stationary phase	Less dense phase – stationary phase
In ascending mode, the less dense phase is used as the mobile phase, while the denser phase is retained as the stationary phase. The mobile phase moves against the centrifugal field.	In descending mode, the denser phase is used as the mobile phase, while the less dense phase is retained as the stationary phase. The mobile phase moves in the same direction as the centrifugal field.
It is called ascending because the lighter mobile phase moves inward, against the centrifugal field, similarly to a lighter liquid rising through a denser phase.	It is called descending because the denser mobile phase moves outward, in the direction of the centrifugal field, similarly to a heavier liquid sinking through a lighter phase.
ASC is often selected when the target compound shows more practical elution behavior with the upper/lighter phase as mobile phase. It can also be advantageous when product recovery or solvent removal is easier from that phase.	DSC is often selected when the target compound shows more practical elution behavior with the lower/denser phase as mobile phase. It can be advantageous when partitioning, pressure behavior or product recovery favors lower-phase operation.

The choice between ASC and DSC should be guided by experimentally determined K_d values, phase-density difference, stationary-phase retention, system pressure and product-recovery considerations. Switching the mode changes which phase carries the compound through the rotor, and can therefore strongly influence elution time, peak shape, solvent-removal strategy and overall process practicality.

After a promising solvent system has been selected, CPC performance still depends on practical operating parameters, especially sample solubility, injection strategy, flow rate, rotor speed and stationary phase retention.

Key points to optimize include: 

Sample solubility: Poor solubility limits loading capacity, increases injection volume and may destabilize the run. The crude sample should dissolve sufficiently in the selected injection medium without precipitation, persistent emulsion formation or visible material at the phase boundary.
Sample-introduction strategy: If the crude sample does not dissolve well in the mobile phase, alternative strategies can be considered, such as injection in the stationary phase, biphasic sample injection or the use of a compatible third solvent. The selected strategy must be compatible with phase stability and should not disturb stationary phase retention during injection.
Flow rate: Higher flow rates may shorten cycle time but can reduce resolution or cause stationary phase loss if the system becomes unstable. Flow rate must therefore be optimized together with rotor speed and solvent system behavior.
Rotor speed: Rotor speed helps retain the stationary phase, but must be balanced with pressure, phase-density difference and equipment limits. 
Stationary phase retention: A robust method requires sufficient stationary phase retention throughout the run, because stationary phase loss can directly affect resolution, reproducibility and recovery. For this reason, stationary phase retention should be treated as a key method development parameter, not only as an operational observation.

The aim is to identify an operating window where solubility, loading capacity, resolution, cycle time, pressure, stationary phase retention and phase stability are all acceptable.

Once an appropriate solvent system and operating window have been established, the method can be transferred from screening to an actual CPC purification run. 

Typical steps include: 
1) prepare sufficient quantities of both equilibrated phases; 
2) select the appropriate operating mode, either ASC or DSC; 
3) prepare and dissolve the crude sample; 
4) fill and equilibrate the CPC system with the selected stationary and mobile phases; 
5) confirm stable stationary phase retention and pressure behavior;
6) inject the sample under the defined loading conditions; 
7) execute the CPC run; 
8) collect fractions; 
9) analyze the collected fractions by HPLC or another suitable analytical method.  

Following purification, fraction analysis enables determination of:
– product purity;
– recovery; 
– isolated yield; 
– loading capacity; 
– impurity distribution; 
– solvent consumption and cycle time,
– process robustness.

These results provide the basis for further optimization. Sample concentration, injected amount, flow rate, rotor speed, fraction collection strategy and solvent composition and operating mode can all be adjusted to improve productivity, recovery and overall process performance. 

Method development is inherently iterative and the data obtained during purification runs provide valuable guidance for further refinement. A well-developed CPC method not only improves separation performance but also creates a solid foundation for scale-up, routine purification and efficient process implementation.

In this sense, robust CPC purification is not defined by solvent selection alone. It is built by connecting partition behavior, phase stability, operating mode, sample handling and process parameters into one controlled and reproducible method.