Extreme Scale-Down Approaches for Rapid Chromatography Column Design and Scale-Up During Bioprocess Development - Measurement, Monitoring, Modelling, and Control of Bioprocesses - page 115

3.2 Empirical Scale-Up Example
This example illustrates the use of equations to adjust elution peaks generated in a
millilitre-scale column separation to predict the height, width and retention times
of larger-scale elution profiles. Two key differences between the scales are the
dispersion and retention volumes; For example, band broadening occurs inside and
outside a chromatography column, and as column sizes reduce, extra-column
effects become increasingly important in determining overall peak spreading.
Failing to account for this can lead to incorrect selection of cut points and/or
inaccurate peak volume estimation during scale-up, thus complicating the opera-
tion of subsequent steps; e.g. an unexpectedly dilute eluate may require longer
downstream processing times. Correcting for the impact of retentive and dispersive
effects upon an elution profile during scale-up is crucial and is especially important
for columns of around 1 mL in size, since this packed bed volume may be similar
to or smaller than the extra-column volume. Hutchinson et al. [ 11 ] describe how
experimental correction factors derived from conductivity changes can be used to
correct small-scale dispersion and retention effects and thus enable accurate pre-
diction of the shape, size and position of larger laboratory and pilot elution profiles
from a Protein A column. The approach was exemplified by using data from a
1-mL Protein A column challenged with a chimeric monoclonal antibody to
predict the elution peaks from 3-mL laboratory and 18.3-L pilot columns. Tran-
sition analysis was conducted in 5-mm-diameter columns with bed heights ranging
between 20 and 205 mm. The transitions were brought about as a high-conduc-
tivity equilibration buffer was replaced by a low-conductivity elutant, and thus the
relationship between total dispersion and column height was determined. This was
used to correct the small-scale elution profiles for dispersion when predicting the
larger packed bed outcomes. A simple mathematical approach was also used to
correct for retention effects. The corrected 1-mL data provided good predictions of
elution profiles of both the 3-mL and 18.3-L columns. Such information could then
be used to determine the most suitable adjustments to apply at extreme scale-down
to achieve the same eluate volume and concentration as obtained at scale-up and
which would thus be suitable for experimentation in small-scale mimics of sub-
sequent unit operations.
3.3 Mechanistic Scale-Up Example
Mechanistic understanding has also been used for scale-up predictions; For example,
Gerontas et al. [ 9 ] developed methods to derive process understanding from a highly
minimal set of column runs to predict larger-scale elution profiles generated by
varying the salt concentration. Computational fluid dynamics (CFD) and 1-mL
column data were used to make the predictions. This involved using the general rate
model (GRM) to account systematically for properties such as axial dispersion, film
mass transfer, intra-particle diffusion and binding kinetics. GRM parameters were