Tuesday, February 5, 2013
In a laboratory scale column the column wall offers support to the column
bed and contributes to the stability of the column. However, when the column
is scaled up and its diameter increases, the wall support contribution
to bed stability starts to decrease. For column diameters greater than 25–30
cm, the lack of wall support may become an issue and could cause redistribution
of packing particles and settling of the bed. The total drag force on
the packing particles is a function of the liquid velocity, the liquid viscosity,
and the bed height. The supporting force that keeps the particles in
place decreases with increasing column diameter as a smaller fraction of
the particles are supported by the column wall. Therefore, for large-scale
columns with compressible packings, the maximum velocities are restricted
and decrease with increasing column diameter under identical bed height
and pressure drop, and the situation worsens with duration of column use
[38–41]. This phenomenon is more prominent for nonrigid gel materials and is often reversible within limits but almost always with a marked hysteresis
[39,42]. As illustrated in Fig. 6, the issue of physical stability of the
column bed becomes particularly significant at large scale because the bed
height is larger than in the lab and the column is put to more frequent reuse
during commercial manufacturing. The bottom of the column is most vulnerable
because it feels the pressure drop across the column as well as that
due to the weight of the column itself, which can be appreciable for large
columns. These effects must be taken into account for robust column design.
For cases where bed compression is a problem and the maximum permissible
bed height is smaller than the minimum required to obtain satisfactory
separation, a suggested solution is to use stacked columns [42]. This
reduces the pressure difference across any single section of the column and
permits the use of smaller particles and nonrigid gels to obtain enhanced
resolution.
2. Bed Stability (Chemical)
Chemical stability of the packing material includes any factors that may result
in deterioration of the column performance over a period of use. It may be the
leaching of ligands into the mobile phase as often in affinity chromatography
destruction of the matrix in the mobile phases used for column operation,
regeneration, or storage (e.g., silica packings at high pH) or irreversible binding
at the packing surface [41]. The issue of chemical stability of the column
bed becomes particularly significant when the column is reused many times
during commercial manufacturing.
3. Product Loading
The product loading (milligrams of product loading per milliliter of resin) is
generally held constant during scale-up. In most cases the resolution is found
to decrease with increasing product loading after the loading has reached a
certain level. Further, this behavior is more prominent when the paricle size
is small. To ensure a successful scale-up and successful operation at large
scale, studies must be conducted at lab scale to determine the maximum product
loading with which satisfactory resolution can still be achieved. It is common
to operate the column at 80–90% of this maximum loading.
4. Gradient Separations
Gradient elution is widely used owing to its ability to provide higher efficiency,
reduced process times and solvent consumption, and a concentrated product stream. However, as process scale increases, buffer volumes increase
also, and it becomes increasingly difficult to form accurate and reproducible
gradients. Thus, either the possibility of performing step gradients should be
explored or the ability of the chromatography skid to perform adequate buffer
mixing and form controlled gradients must be evaluated [15].
5. Flow Distribution
For large diameter columns, uniform flow distribution at the column head may
become difficult to achieve. This may result in deviations from the desired
plug flow and lead to peak tailing. Use of a flow distributor at the column
inset is generally found to be the most effective way of ensuring uniform flow
distribution in the column [41]. The rational design of inlet and outlet headers
to ensure uniform distribution is discussed at length in Chapter 2.
6. Packing Quality
Packing large columns such that the resulting packed column is homogeneous
is very critical for obtaining uniform flow distribution. Channeling inside the
column often leads to peak tailing and/or peak splitting. A variety of approaches
have been developed by the different chromatographic equipment
vendors to alleviate this problem. The most popular technique at preparative
scale is axial compression and the use of self-packing columns [35]. Some
degree of compression has been known to enhance resolution [41]. The connection
between packing and the phase ratio used in quantitative scale-up was
touched upon in the previous section.
7. System Design
The system dead volume arising from the piping and other support equipment
for chromatography columns such as the valves, flow meters, air sensors, and
tubings is much larger at pilot and particularly manufacturing scale than at
lab scale. This leads to dilution effects and higher pressure drops as well as
additional band broadening, so the impact these factors may have on the overall
column performance must be evaluated. The general guidelines are to keep
the system dead volume to the minimum, to have bypasses through devices
such as air traps and filters for use during sample load, and to choose tubing
diameter to achieve turbulent flow so as the reduce undesirable axial mixing
[43]. Further, the chromatographic system should be designed such that all
inlet sources are at or above the level of the column, whereas all outlet sinks are at or below the level of the column. This ensures that the column is not
being operated against a hydrostatic pressure head.
8. Fraction Collection
The peak width and shape as seen in the chromatogram depends on several
factors such as the column dimensions, extracolumn effects, operating conditions,
and sample volume [42]. Thus, even if the scale-up is carried out following
a well-thought-out methodology, it is still very likely that the peak width
and shape may differ from that obtained at lab scale. Thus, the fraction collection
strategy must be revisited at preparative scale based on column performance
at that scale. The critical monitoring device, e.g., the UV detector,
should be placed as close to the fraction collector as possible to ensure good
representation of the process stream. Also, it is good to have a fraction collector
that can collect fractions based on several process parameters, such as UV absorbance,
conductivity, time, volume, first or second derivative of the signal, etc.
9. Media Availability
Although selectivity for a separation may be the primary criterion for selection
of a resin for analytical separation, several other factors need to be considered
before a resin is selected for preparative separation. These include the availability
of large quantities of the resin, cost, continuity of supply, batch-tobatch
consistency, column lifetime, and support documentation to aid in regulatory
filing [44].
10. Costing
The cost of the feedstock is generally not given adequate consideration when
the process optimization is carried out at bench scale. However, as the process
is scaled up, the process should be modeled and raw material and facility costs
must be examined. Resin costs usually account for the biggest contribution
to the raw material costs. Thus, although a certain resin may offer the best
selectivity at bench scale, it may be too expensive, under linear scale-up, at
process scale [8].
11. Sample Pretreatment
Often cells or inclusion bodies undergo a rupture step by homogenization or
some other mechanism prior to a chromatography step with the purpose of
capturing the product from the cell culture or fermentation broth. In these
situations the process stream is replete with lipids, nucleic acids, proteins, and other macromolecular contaminants. These impurities affect the passage of
the product through the column and have a very significant effect on column
performance. Because most cell-rupture operations are more effective at large
scale than at bench scale, pretreatment of the process stream at large scale
to remove all the interfering contaminants may become crucial to ensuring
satisfactory and robust column performance [8].
12. Scale-Down
Although our focus is on scale-up, it is worth mentioning that some issues
related to validation are often addressed by scaling down. Thus, using a smallscale
study for viral clearance is often faster, safer, and cheaper while being at
least as accurate as larger scale studies [45]. Many of the scale-up approaches
discussed above apply, mutatis mutandis, to scale-down.
REFERENCE
Scale-Up and Optimization in
Preparative Chromatography
Principles and Biopharmaceutical Applications
edited by
Anurag S. Rathore
Pharmacia Corporation
Chesterfield, Missouri, U.S.A.
Ajoy Velayudhan
Oregon State University
Corvallis, Oregon, U.S.A.
YOURS CHORMATOGRAPHICALLY,
KAUSHIK ZALA
