Protein A Column Sizing: Residence Time, Capacity, Cycles

The mass and geometry relations follow from conservation and cylindrical geometry:

For a fixed column, $V_{bed}=\pi D^2H/4$ and $N=\lceil M/(q_wV_{bed})\rceil$. All but the last cycle load to $q_wV_{bed}$; the final cycle takes the remaining feed. Sizing mode uses the selected integer cycle count directly. An unrounded minimum diameter is reported.

Residence time and capacity. Loading flow is $Q=V_{bed}/RT$ and superficial velocity is $u=H/RT$. RT is empty-bed contact time, not interparticle liquid transit time. Interpolate linearly between the two approximate PrismA QB10 values in Instructions 29262586 AC, Table 1, attached PDF p.4; no result is accepted outside 4–6 min in this example mode. For other residence times, change RT or select measured-capacity mode. Measured-capacity mode uses the user's QB10 at the selected RT. The cited vendor polyclonal IgG values are not universal mAb capacities. KA1965230418AN, attached PDF p.4, reports 77 g/L at 6 min for its particular experiment; this is distinct from the approximately 80 g/L typical instructions value.

Geometry and pressure applicability. The bore must be larger than the mean bead diameter to contain the packing. This tool also requires $D/d_p\ge 50$ before reporting its pressure estimate without a wall correction. Mehta and Hawley (1969), Wall Effect in Packed Columns, report significant confinement corrections below a diameter ratio of 50. The 50:1 cutoff is a conservative scope limit for this tool, not a universal guarantee of negligible wall effects or a validation for compressible agarose. A smaller column needs a wall-aware pressure model or measured pressure-flow data. Choose a larger existing column or reduce target cycles to increase the calculated bore.

Pressure. For spherical, uniform, incompressible beads,

Kramer et al. (2021), Eq.1, PDF p.3, gives this Carman–Kozeny form and the modified Reynolds criterion $Re_\epsilon=\rho u d_p/[\mu(1-\epsilon)]<2$. Maximum recipe flow is the greater of loading flow and the common buffer flow. The pressure screen uses that flow and the entered design viscosity. A low estimate does not establish hardware suitability.

Solutions and time. Each stage uses $V_j=N\,CV_jV_{bed}$. Total time is batch feed volume divided by loading flow, plus all solution volumes divided by buffer flow, plus N times additional hold and handling time. The same complete non-load recipe is counted for a partial final cycle.

Breakthrough illustration. A common Thomas-type logistic surrogate is written in applied mass coordinate $r=M_{applied}/(q_{10}V_{bed})$:

Here $r_{10}=1+\epsilon C_f/q_{10}$ and $r_{50}=r_{10}+w/2$, so y is 0.10 at $r_{10}$ and 0.90 at $r_{10}+w$. The void correction includes only interparticle hold-up. The width w is assumed, not measured. The spatial band is an illustrative traveling logistic front with the same outlet history; it is not a transport PDE solution. Thomas (1944) supplies the fixed-bed mass-action model background; this empirical logistic approximation is not a reproduction of his full solution. It does not feed back into sizing or estimate integrated loss.

• Constant clarified-feed titer, constant flow during each stage, equal resin capacity in every cycle, and one column used sequentially.
• Safety reserve 0.20 is a design assumption. Cytiva's capacity/performance note, PDF p.3, demonstrates loading to 80% of QB10 in specific methods; it does not validate that choice for another antibody.
• Bead diameter 60 µm is the PrismA nominal value (Instructions PDF p.4). Void fraction 0.40, density 1000 kg/m³ and viscosity 1 mPa·s are explicit planning assumptions. Use actual feed and packed-bed data.
• The default recipe uses 5 CV equilibration, 6 CV combined wash, 3 CV elution, 2 CV strip, 3 CV CIP, and 3 CV re-equilibration, modeled on the mAb1/mAb2 method in the capacity/performance note, PDF p.3. Recipes remain editable; the extra 15 min static hold and 10 min handling time are planning allowances.
• All non-load stages share one buffer-flow multiplier. The CIP hold is additional to pumping, so it is not silently treated as a total contact time.
• The default breakthrough width of 0.30 is illustrative. No kinetic constant, isotherm, dispersivity, product recovery, pool concentration, or resin-lifetime model has been fitted.
• Mass balances, cycle scheduling and geometry are derived here. The moving front is an original schematic; dimensions and colors are explanatory, not a vendor figure.

Use for preliminary single-column bind/elute capture planning with a Protein A binding target. The PrismA example is limited to 4–6 min and its source's polyclonal IgG data. In measured mode, verify QB10 at the actual RT, feed chemistry, temperature, antibody, lot and resin age. Changing RT does not automatically update a measured QB10.

Results are blocked for a bore no larger than the mean bead diameter and, separately, for a bore below 50 mean bead diameters because the pressure model has no wall correction. With the default 60 µm beads, the pressure-model minimum bore is 3 mm. This is a model applicability limit; it does not forbid experimentally validated smaller columns.

Carman–Kozeny assumes a clean, uniformly packed, rigid bed in laminar flow. Compressible agarose, particle size distribution, packing factor, wall effects, distributor resistance, bed aging and fouling can make actual pressure substantially different. The cited PrismA 300 cm/h guidance is specific to its stated packing and water-like viscosity conditions; it is not a universal limit. Source instructions discourage beds below 10 cm at large scale. Check actual resin, column and system ratings and the measured pressure-flow curve.

Solution totals exclude overfill, priming, system dead volume, transfer and neutralization solutions. Total time excludes column packing, preparation, tests and facility turnaround. No prediction of purity, impurity clearance, ligand leakage, sanitization efficacy, product recovery or allowable resin reuse follows from these estimates. The animation can continue beyond the planned loading stop solely to explain breakthrough.