5 Unit Operations That Define 90% of All Biotech Purification Trains

Biotech processes vary enormously in detail. The core toolkit doesn’t.

Whether you’re purifying a monoclonal antibody, a recombinant enzyme, a small-molecule fermentation product, or a viral vector — most purification trains are built from the same five unit operations. Different orders. Different parameters. Same concepts.

The Five Operations

1
Centrifugation — Remove Cells First

Every fermentation and cell culture process starts with the same problem: your product is in a broth full of cells, cell debris, proteins, lipids, DNA, and media components. Step one: remove the solids.

Centrifugation exploits density differences. Cells and dense particles sediment faster than liquid under spin force. The clarified liquid — the supernatant — moves forward. The pellet is waste.

Use it when
Feed has whole cells (always, if it does), cell density >5 g/L dry cell weight, need to process large volumes quickly.
Key parameters
G-force, residence time, temperature, feed solid content.
Common pitfall
Too low a G-force for the cell size leaves significant debris in the supernatant that fouls downstream columns. Mammalian cells are fragile; yeast need much higher forces. Depth filtration often used alongside — captures sub-micron particles centrifuges miss.
2
Ultrafiltration / Nanofiltration — Separate by Size

UF and NF membranes have defined pore sizes — molecular weight cut-off (MWCO). Molecules above MWCO are retained in the retentate; molecules below pass through in the permeate.

Use it when
Concentrating product, diafiltration to wash out small molecules while retaining large target, separating two molecules with ≥5–10× MW difference.
Key parameters
MWCO, transmembrane pressure, crossflow velocity, diafiltration volumes.
Diafiltration
10 diavolumes removes ~99.995% of a permeable impurity. Primary tool for buffer exchange and removing residual reagents.
Common pitfall
Choosing MWCO near target MW. If protein is 50 kDa, a 30 kDa membrane risks product loss. Use cutoff at 10–30% of target MW for reliable retention.
3
Ion Exchange Chromatography — Separate by Charge

Anion exchange (AEX): positively charged resin, negatively charged molecules bind. Cation exchange (CEX): negatively charged resin, positively charged molecules bind.

Molecule charge depends on isoelectric point (pI) and operating pH. Above pI: negatively charged. Below pI: positively charged. The difference between pI values of product and impurities determines selectivity.

Use it when
Target and main impurities have different pI values, need to remove host cell proteins, DNA, or endotoxins, want an orthogonal polishing step.
Key parameters
pH, ionic strength, buffer composition, column loading density, elution gradient slope.
Common pitfall
Loading IEX with a feed that still has particulates. Particles foul columns and destroy resin capacity. Always clarify before loading.
4
Affinity Chromatography — The Selective Capture Step

An immobilized ligand specifically binds your target molecule. Everything else passes through in the flowthrough. Single-step purities above 95% are routine. Protein A for antibodies, IMAC for His-tagged proteins are the two most common examples.

Use it when
Molecule has a natural binding partner, purifying His-tagged proteins, need capture-level purity in one step and can justify the resin cost.
Key parameters
Ligand density, binding pH and conditions, elution condition, resin lifetime (number of cycles).
Common pitfall
Protein A elution requires low pH (~3.5), causing aggregation in pH-sensitive molecules. Always check aggregation propensity and neutralize immediately after elution.
5
Drying — Making Solid Final Forms

Spray drying: feed atomized into hot gas, water evaporates instantly, leaves dry powder. High throughput, continuous, suited to heat-stable molecules.

Freeze drying (lyophilization): feed frozen, water sublimed under vacuum. Low temperature throughout — standard for thermolabile biologics and injectables.

Vacuum tray drying: evaporation under reduced pressure, low capital cost, common for small-molecule fermentation products.

Use it when
Solid final form required, long-term stability needs low water activity, or shipping economics favor powder over liquid.
Key parameters
Inlet air temperature, atomizer speed (spray), shelf temperature and chamber pressure (freeze drying), feed solid content.
Common pitfall
Drying a dilute feed. A feed at 10 g/L costs 100× more to dry than the same mass at 1,000 g/L. Always concentrate first with UF.

Putting It Together: A Typical 3–5 Step Train

These five operations don’t just run independently — they compose into a logical sequence. Here’s what a typical purification train looks like:

Step Operation Purpose Property exploited
1 Centrifugation / depth filtrationRemove cells and debris Feed clarification Density / Size
2 UF concentration + diafiltrationConcentrate and buffer exchange Volume reduction + conditioning Molecular weight
3 Affinity or IEX captureHigh-selectivity capture of target Primary purification Binding / Charge
4 IEX or SEC polishingRemove remaining impurities Final purification Charge / Size
5 UF/DF or dryingConcentrate and formulate Final form + stability Molecular weight

Key insight: each step should exploit a different physical property — size, charge, hydrophobicity, specific binding. Two consecutive steps exploiting the same property don’t give additional separating power. They just add cost and yield loss.

How to Know if a Step is Doing Useful Work

Selectivity measures whether a unit operation is actually separating your product from impurities — or just concentrating everything together:

Selectivity (α) = (Product concentration factor) / (Impurity concentration factor)
where concentration factor = Cout / Cin

When α > 1, the operation enriches product more than impurities. When α = 1, it concentrates both equally — useful for volume reduction, useless for purification. When α < 1, it enriches the impurities. That step should be redesigned or removed.

A well-designed purification train has α > 1 at every step. Each step leaves the product more enriched than the step before.

The Yield Math You Can’t Ignore

Total process recovery is the product of per-step yields. At a typical 80–90% per step, a four-step train recovers 41–66% of the product. At 70%, only 24%. That math is fixed — it doesn’t improve when you scale up.

Four steps at 85% each = 52% total recovery. Nearly half the product — and every dollar spent producing it — discarded. This is why the number of steps matters as much as the yield at each one.

This is also why two operations with selectivity ≈1 are worse than one good one: they each take their yield loss without contributing purification.

Design your purification train before the constraints lock in

Drag-and-drop unit operations, real mass-balance calculations, selectivity tracking, and instant what-if analysis — all in the browser. See which five operations give you the best yield and purity for your molecule.

Try it free →