Technical publication · Downstream chromatography
Scope and design question
For a clarified harvest containing a known antibody mass, what minimum Protein A resin volume and packed-bed diameter can process the batch in an integer number of cycles, at a qualified dynamic binding capacity and residence time, without exceeding the available downstream window?
This is a first-pass, single-column bind-and-elute calculation. It covers load capacity, ideal cylindrical geometry, flow, solution volumes and elapsed time. It does not predict pressure drop, breakthrough shape, pool quality, resin aging, cleaning effectiveness or continuous/PCC scheduling. When that fidelity is needed, experimentally parameterized general-rate models can represent loading, wash and elution across defined conditions and scales [9].
Dynamic binding capacity depends on residence time, feed concentration, molecule, stationary phase and operating conditions. Experiments have shown lower Protein A DBC at high mAb titer and short residence time, consistent with intraparticle mass-transfer limits [4]. Use the DBC measured for the intended breakthrough definition and feed—not a universal resin number.
Nomenclature
On small screens, swipe the table horizontally to compare every column.
02 · Interactive engineering calculator
First calculate the antibody mass that actually reaches capture. Apply an explicit utilization fraction α to a measured qDBC,x; do not treat α as recovery. If column volume is fixed instead, calculate n = ⌈MB/(qworkVC)⌉ and retain the discrete ceiling in facility and buffer calculations.
Calculation workflow
Calculate MB = VharvestCmAbYclarification on a consistent sample and volume basis. Use distributions or low/base/high cases when harvest volume, titer or clarification recovery varies. Keep total protein, target mAb and assay uncertainty separate.
Time-align the inlet before integration: Cin,a(t) = Cin(t − Vdelay/Q), then qDBC,x = (Q/VC)∫0tx[Cin,a(t) − Cout,target(t)]dt. This area form counts pre-threshold product loss; Protein A stationary-phase work distinguishes it from the simpler assumption that all product loaded before 10% breakthrough is bound [7]. A packed-bed Protein A study defines DBC10 using breakthrough volume minus an offset volume and held that offset-delay basis constant between experiments [10]. Measure Vdelay on the declared flow path and flow—for example from a validated conductivity front—and determine target breakthrough with a target-specific assay or a verified non-binding-signal correction. State the threshold, such as Cout,target/Cin,a = 0.10. A frontal-analysis procedure explicitly separates system delay, non-binding offset and target breakthrough [11]. Comparative studies show that capacity and process performance depend on the selected medium and conditions [3].
Residence time is τ = VC/Q and superficial velocity is u = H/τ. Published experimental comparisons found stationary-phase-dependent changes in DBC with operating flow and residence time [2]. Empirical qDBC(τ) fits have been evaluated across molecule-specific residence-time ranges [5]; use them only inside tested conditions, never as a universal resin law.
Convert VC to m3 and H to m before calculating diameter in metres. The resulting diameter and flow are only candidates: Protein A scale-up depends on operational and system parameters [1], while comparative stationary-phase data show distinct pressure-flow, mass-transfer and DBC behavior [7]. Select available hardware, recalculate actual volume and loading, then verify compression, distributor performance, packing method, system limits and scale-dependent dispersion.
Assume 6.00 kg mAb reaches capture, measured DBC10 is 50 g L−1 at τ = 4 min, and α = 0.80, so qwork = 40 g L−1. Three cycles require 50.0 L resin. At H = 0.200 m, the ideal diameter is 0.564 m and Qload = 50/4 = 12.5 L min−1. These are synthetic screening values, not resin or customer performance claims.
05 · Integer-cycle sensitivity
On small screens, swipe the table horizontally to compare every scenario.
Each scenario uses 10 load CV plus 18 non-load CV at 4 min per CV and 20 min of fixed transitions per cycle. Ideal non-load solution demand remains 2,700 L because nVC is constant at fixed capacity; real prime, flush, heel and skid volumes are not constant. Download the scenario dataset.
Time, buffer and yield
With Cload = 4 g L−1, each nominal cycle receives 2.00 kg in 500 L, equal to 10 CV. At 12.5 L min−1, load time is 40 min. If load concentration changes DBC, recalculate capacity rather than changing only load time.
An illustrative sum of 18 CV per cycle may include equilibration, wash, elution, strip and regeneration on a declared recipe basis. Three cycles at 50 L use 2,700 L before skid hold-up, line prime, sampling, recovery, preparation overage and cleaning-solution demand.
At a common 4 min/CV screen, 28 CV of flow occupies 112 min per cycle. Add 20 min per cycle for transitions and holds: 3 × 132 min = 6.60 h. A separately assumed 95% step yield gives 5.70 kg output; DBC10 does not imply 90% recovery.
Use phase-specific flow limits rather than one residence time when equilibration, wash, elution, strip and regeneration differ. Add gradient volumes, system dwell volume, collection delays, cleaning, integrity checks, packing or unpacking, sampling and quality holds where they occupy the same facility window. When yield, column utilization and sorbent productivity must be balanced against loading-time and pressure constraints, use an affinity-process design framework rather than this single capacity equation [6].
Limitations
On small screens, swipe the table horizontally to compare every control.
Practical design checklist
Clarified volume; titer distribution; target-mass assay; clarification recovery; turbidity and particle burden; HCP and DNA; pH, conductivity, temperature and viscosity; molecule class; allowable breakthrough, yield and pool-quality limits.
Resin identity and lot; DBC curve versus residence time and feed concentration; breakthrough definition; bed height; compression; available column sizes; distributor; packing method; pressure-flow curve; skid flow, pressure and UV limits; lifetime and CIP evidence.
Phase-specific CV and flow; line and skid hold-up; prime, flush and overage; transition and sampling time; elution pool volume; viral-inactivation interface; buffer vessels; operator and room occupancy; process window; cycle rounding; contingency and evidence owner.
Primary technical sources
Technical FAQ
Divide antibody mass entering capture by the product of whole-number cycles, qualified dynamic binding capacity and the permitted capacity-utilization fraction: VC,min = MB/(nαqDBC). Select an available column, then recalculate actual loading and batch time.
DBC10 is the target mass bound per resin volume when the defined outlet signal reaches 10% of the corrected inlet signal under stated test conditions. The integration, delay correction, residence time, feed and assay must be documented; it is not a universal resin constant.
For an ideal cylindrical packed bed, D = √[4VC/(πH)]. Convert resin volume to cubic metres and bed height to metres. Then select available hardware and verify its actual packed volume, pressure-flow and distributor limits.
A shorter residence time increases flow for a fixed bed, but may lower DBC when mass transfer is limiting and can increase pressure drop. Determine a product- and feed-specific DBC-versus-residence-time relationship, then screen capacity, time and hardware constraints together.
No. At fixed working capacity and a recipe expressed only in CV, ideal solution volume nVCΣCV is unchanged because resin volume falls inversely with cycles. Real line prime, flush, heel, cleaning and overage can increase with each additional cycle, while the batch window becomes longer.