Biomedical Engineering Reference
In-Depth Information
will be removed. For initial screening, the membrane should be selected with a
MWCO value three to five times lower than the molecular weight of the species to
be retained [18].
Species retention can be quantified by the rejection coefficient, R, [19]
C
p
C
r
R ¼
1
where C
p
¼
solute concentration in permeate and C
r
¼
solute concentration in
retentate.
The overall rejection coefficient for a given solute feed solution depends on (1) the
membrane retention and adsorption properties and (2) the gel layer properties. Hence, the
overall rejection coefficient is the combined rejection coefficient of membrane and gel
layer. The gel layer is dependent on the solute feed properties (concentration, diffusivity
coefficient, temperature), membrane-solute interactions and the operating conditions;
transmembrane pressure and cross-flow velocity (turbulence and shear) across the
membrane. Because the gel layer properties are dynamic and constantly changing,
steady state of the gel layer formation is required to accurately determine the overall
rejection coefficient. This is accomplished by keeping the solute feed concentration and
system operating conditions constant. The rejection coefficient can be determined
experimentally by operating the TFF system with the permeate open and recycling
back to the retentate vessel. After the system comes to steady state, the concentration of
the solute of interest is measured in the permeate and retentate streams. The rejection
coefficient is then calculated using the above equation.
To minimize product loss to the permeate, especially during diafiltration, a
membrane should provide a rejection coefficient R
0.90,
50% of the product will be lost to the permeate after just six diafiltration volumes.
Conversely, to maximize the removal of unwanted low molecular weight species, the
membrane should offer a low rejection coefficient of the contaminant, which will reduce
the number of diafiltration volumes required to remove the contaminant.
In addition to product “leakage” through the membrane, product can also be lost
due to nonspecific adsorption of the product to the membrane skin and substrate.
Fortunately, advances in improving the hydrophilic nature of TFF membranes have
generally removedmembrane adsorption as a primary cause of product loss. Proteins can
adsorb to membranes primarily through charge differences and hydrophobic interac-
tions. This results in fouling of the membrane and has the effect of lower product
recovery and degradation of filtrate flux and increases the cleaning difficulty membrane
performance recovery [20]. Loss due to adsorption can be assessed during membrane
selection. Recovery can be computed by performing a mass balance on the retentate and
permeate.
>
0.99. For example, at R
¼
6.5.1 QbD Principles for Membrane Selection
Maximize yield by selecting a membrane with high product retention and low adsorptive
losses.
