Tangential flow filtration pushes feed across a membrane rather than through it, so the retained material sweeps the surface instead of blinding it, and a step can concentrate or exchange buffer on a volume a dead end filter would stop on in minutes. What a buyer actually decides is format, membrane area and how the step is controlled, and those three follow the product rather than the catalogue.
tff systems, and what a tff system includes
A crossflow skid is a pump, a feed vessel, the holder or cartridge, pressure sensors on feed, retentate and permeate, valves and a controller that holds transmembrane pressure and crossflow rate while the volume falls. Benchtop units add a balance for automatic diafiltration; process systems add flow meters, conductivity and single use flow paths. What to check on any quote is the minimum working volume, because hold up decides how much product a small batch leaves behind, and whether the controller logs the run in a form your batch record can use.
tff cassettes, and what a cassette gives you
A flat sheet cassette stacks membrane and screen layers into a plate and frame holder, which gives a high area in a small footprint, a well defined channel and the shear a screened path produces. That combination gives high flux on proteins and is why antibody concentration and diafiltration run on cassettes. The costs are the holder and its torque specification, a channel that can foul on particulates, and a hold up volume per cassette that matters at small scale. Screen type, open or suspended, is chosen from the feed's own solids.
A tff filter, and what the cutoff really means
A membrane's nominal molecular weight cutoff is a nominal retention rather than a sieve: convention puts it where about ninety per cent of a test solute is retained, and shape, charge and concentration all move the real figure. The practical rule is to choose a cutoff several times smaller than the product so it is retained cleanly, three to five times for a protein, and to prove retention on your own feed rather than on the specification. Membrane chemistry matters as much, since polyethersulfone, regenerated cellulose and PVDF bind differently.
hollow fiber tff, and where fibres beat cassettes
A hollow fibre module is a bundle of tubes with the feed inside the lumen, which gives an open unobstructed channel. That is what suits cells, lysates and anything with solids: a fibre module tolerates a feed that would block a screened cassette, runs at low shear, which keeps fragile cells and viral particles intact, and can be backflushed. It costs area per volume and gives lower flux on clean protein feeds, so perfusion, cell harvest and virus work run on fibres and protein polishing runs on cassettes.
ufdf, and what the two letters do
UFDF is ultrafiltration followed by diafiltration, the step that finishes most biologics processes. Ultrafiltration concentrates by removing water and small solutes until the product is at the target concentration; diafiltration then exchanges the buffer by adding formulation buffer at the rate permeate leaves, counted in diavolumes, until the old buffer is washed out to a stated fraction. Doing it in that order saves buffer and time, and the constraint is the product's own solubility and viscosity at the final concentration.
tff purification, and what it can and cannot separate
Crossflow separates by size alone, so it concentrates, exchanges buffer and clears solutes well below the cutoff, and it does not resolve two species of similar size. That places it around chromatography rather than instead of it: clarification and concentration before capture, buffer exchange between steps, and the final concentration into formulation. Where it does act as a purification is virus and particle work, where the product is orders of magnitude larger than the impurities, and there the recovery figure is the whole argument.
A tff column comparison, and why the shape differs
People reach for the word column because chromatography is the familiar format, and the shapes are not comparable. A chromatography column is a packed bed the feed passes through once, retaining by chemistry, and its capacity is per millilitre of resin. A crossflow module is a membrane the feed passes across many times in a recirculating loop, retaining by size, and its capacity is per square metre of area and per hour. Sizing arithmetic therefore has nothing in common: one is bed volumes, the other is area and flux.
A tff membrane area, and sizing it from flux
Area is sized from the flux the feed sustains and the time you will allow. Measure flux in litres per square metre per hour on a small module at the transmembrane pressure and crossflow the process will use, then divide the volume to be processed by that flux and the hours available, and add margin for the fall in flux as concentration rises. Scaling between modules of the same path length is by area; changing path length changes the hydraulics and needs its own measurement rather than arithmetic.
Questions people ask about tff filtration
Cassette or hollow fibre?
Fibres for cells, lysates, virus and anything with solids, because the lumen is an open channel at low shear. Cassettes for clean protein feeds, because the screened channel gives higher flux per square metre.
How is membrane area chosen?
From measured flux on your own feed at the operating pressure and crossflow, divided into the volume and the time available, with margin for the flux fall as concentration rises.
What does UFDF mean on a process flow?
Ultrafiltration to concentrate, then diafiltration to exchange buffer at constant volume, counted in diavolumes until the original buffer is washed out to a stated fraction.