Liquid–liquid extractions (LLE)

Using the software, we can:

i) develop a methodology to anticipate emulsion formation;

ii) Perform LLE troubleshooting to improve yield and cycle time.

The background of liquid liquid mixing:

• Two immiscible liquids – a dispersed phase (drop) in a matrix (continuous) phase.
• Mixing impacts the equilibrium between drop coalescence vs. drop breakup.
• Lab-scale vessels provide unrealistic hydrodynamics that can mask coalescence and dispersion suspension problems.

 

The moon-shot: To develop an efficient and reliable methodology to anticipate emulsion formation and phase separation times at manufacturing.

Phase separation times in liquid–liquid extractions

Stage 1 – Developing in-house knowledge

Literature search of liquid mixing process:

• Define relevant parameters
o Mean drop size
o Turbulent shear rate at the blade circulation flow rate
o Relative residence time
o Maximum energy dissipation

• Prior work/knowledge
o Laboratory and/or manufacturing observations

Experimental apparatus:
• Reactor geometry
• Impeller type
• Liquid–liquid system
DoE approach:
• Exclude correlated variables
• Perform full-factorial tests

Phase separation times in liquid–liquid extractions
Stage 1 – Developing in-house knowledge (bench scale DoE)

 

Higher maximum energy dissipation (MED) leads to longer

separation times.

•  For the same MED, different impellers yielded similar separation times.
•  Interaction with other mixing variables?

High maximum energy dissipation (MED) leads to larger
separation times.

• High residence time in the zone of MED leads to centrifugal separation –
  swirl effects
• Power and geometry effects can now be anticipated!

 

Stage 2 –Applying knowledge at manufacturing

 

 

Scaled-down manufacturing reactor to a 2 L bench reactor:
• A correlation exists between the maximum energy dissipation and the amount of
residual emulsion at the end of the procedure.

The new optimized mixing conditions yielded the desired
product. Besides stirring speed, no changes are made to the manufacturing
technique.

Lessons learned:

1. Software was used to study the effect of geometric
   variables in oil-inwater systems.
   Key mixing parameters are identified!
2. A reasonable acceptable prediction model was generated and integrated in a
   QbD methodology.
   Increased know-how! Right first time!
3. The software was used for the optimization of an important manufacturing
   liquid–liquid
   extraction process.
4. Increased batch yield and cycle-time! $

Taken from the Second VisiMix Forum by Nuno Lousa - Hovione