(206a) Scale-up and Optimization of Filtration Processes: Small Scale Pressure Filtration to Centrifugal Filtration

Mitchell, N. - Presenter, Process Systems Enterprise
Burcham, C. L., Eli Lilly and Company
Polster, C. S., Eli Lilly and Company
Girard, K., Pfizer, Inc
Bermingham, S., Process Systems Enterprise Limited

The separation
of solid particles from liquids via filtration is a key unit operation in the
pharmaceutical sector. This processing step commonly occurs after the isolation
of an Active Pharmaceutical Ingredient (API) via crystallization and may
involve a number of wash cycles in order to deliqour the cake and remove any residual
solvents below specified limits. For the purposes of filtration, pressure or
vacuum filters and centrifuges are most commonly employed.

1: Filtration scale-up workflow -
pressure filter to centrifuge

In this work we
detail a model based workflow for the scale-up of filtration unit operations,
as summarized in Figure 1 above. This workflow begins with the estimation of the following
cake properties from small scale pressure filter experiments:

Specific cake resistance

Medium resistance

Compressibility index of the cake

Reference pressure drop for compressible cakes

The above
parameters are then regressed from experimental data from the lab scale, namely
the change in filtrate volume as a function of time for a range of conditions,
as follows:

Pressure drop

Particle size distribution of particles in cake

These cake
properties were subsequently employed within a centrifuge model to predict the
filtration behavior of the system at plant scale. Additionally, the schedule of
operation of the centrifuge was optimized with an objective of achieving the
maximum product throughput through the centrifuge. The various trade-offs
observed in the optimization objective will also be outlined. This filtration
workflow for scale-up outlined above was applied in this work for a number of
industrial cases, involving compressible cakes. This model-based approach
allows the filtration unit step to be scaled-up and optimized in a fast and efficient
manner, with minimal experimental effort and consumption of valuable materials.