Most filtration decisions in monoclonal antibody production are made at the wrong level: teams compare membranes when the real question is which process step the filter has to serve. The two operations look similar on a datasheet and behave very differently in a process line.
The Short Answer: Match the Step, Not the Membrane
Choose microfiltration when the task is to remove cells and debris from harvest fluid before capture chromatography. Choose ultrafiltration when the task is to concentrate or buffer-exchange a purified antibody. The process step decides the membrane, the format and the validation package — not the supplier’s catalogue.
That distinction matters commercially as well as technically. A clarification filter and a concentration membrane differ in pore size, format, operating mode, cleaning regime and documentation. Buying one to do the other’s job usually shows up later as fouling, low flux, or a validation gap that blocks release.
Where Each Step Sits in an mAb Process
A typical monoclonal antibody process moves through harvest, capture, viral inactivation, polishing, concentration and formulation, then sterile filtration before fill. Microfiltration dominates the early, high-solids end; ultrafiltration dominates the late, high-value end where the product is already purified and concentration matters.
| Process step | Typical operation | MF or UF | What it achieves | What it does not do |
|---|---|---|---|---|
| Harvest and clarification | Removal of cells, debris and large particulates | Microfiltration | Protects the capture column and reduces turbidity | Does not remove viruses, host-cell proteins or small aggregates |
| Capture | Protein A or similar chromatography | Neither (chromatography) | Captures and concentrates the antibody | Not a filtration step |
| Viral inactivation and polishing | Low-pH hold, ion exchange, polishing filters | Usually neither, sometimes MF | Reduces viral and impurity load | Not a substitute for validated viral clearance |
| Concentration and buffer exchange | Ultrafiltration / diafiltration (TFF) | Ultrafiltration | Raises product concentration and exchanges buffer | Does not reliably separate proteins of similar size |
| Sterile filtration | 0.2 µm filtration before fill | Microfiltration (validated) | Removes microorganisms when validated for the product | Is not a clarification step and is not automatically sterile |
For the regulatory context of sterilising filtration and aseptic processing, the FDA guidance on sterile drug products produced by aseptic processing sets out what a sterilising step has to demonstrate.
What Microfiltration Removes — and What It Cannot
Microfiltration is conventionally described as retaining particles in the roughly 0.1 to 10 micrometre range, which is why it fits harvest fluid full of cells, cell debris and colloidal material. In practice most mAb processes combine depth filtration for the heavy solids load with a membrane step for the finer fraction, because a single membrane fouls quickly when it meets whole harvest fluid.
What it cannot do is just as important. Microfiltration does not remove viruses, does not clear host-cell proteins, and does not sterilise a solution unless the specific filter has been validated for bacterial retention with the actual product. Teams that treat a 0.2 micrometre filter as a universal safety step create a documentation gap rather than a control. Bacterial retention testing follows ASTM F838, and aseptic processing is addressed by ISO 13408-1.
- Dead-end mode: the feed passes through once; capacity is limited by solids loading.
- Tangential flow mode: feed recirculates across the membrane; better for high-solids streams but adds pump, tubing and hold-up volume.
- Depth plus membrane: the common harvest configuration, where each layer does a different job.
What Ultrafiltration Does: Concentration, Diafiltration and MWCO
Ultrafiltration retains the antibody while water, salts and small molecules pass through. It is normally run in tangential flow mode so that the membrane surface is swept continuously, and it is used for two distinct jobs: raising concentration, and exchanging the product into the formulation buffer — usually called diafiltration when buffer is added while filtrate is removed.
The selection variable is the molecular weight cut-off rather than a pore size in micrometres. Viral safety, which ultrafiltration supports but does not by itself establish, is covered by ICH Q5A. Because an IgG molecule is around 150 kDa, the cut-off is chosen well below the product size so the antibody is retained while buffer components pass. Cut-off choice is a balance: tighter membranes retain the product more reliably but generally cost flux and can concentrate impurities along with it.
| Decision | What it controls | What to ask the supplier |
|---|---|---|
| Molecular weight cut-off | Retention of product versus passage of buffer components | Retention data for a molecule of similar size, not only a nominal cut-off |
| Membrane material | Fouling behaviour, cleaning, compatibility | Chemical compatibility list and cleaning validation approach |
| Format | Cassette, hollow fibre or single-use flow path | Hold-up volume, scale-down model and linearity of performance |
| Operating mode | Flux, transmembrane pressure, recirculation | Recommended operating window and the consequences of running outside it |
Selection Criteria That Actually Decide It
Once the process step is fixed, the remaining criteria are practical rather than theoretical. They concern the feed stream, the scale, the cleaning regime and the evidence the supplier can provide. A filter that performs well in a bench trial can still fail qualification if its documentation cannot support the claims made for it.
| Criterion | Question to settle | Why it changes the choice |
|---|---|---|
| Purpose | Clarification, concentration, buffer exchange or sterilising filtration? | Determines membrane type, format and validation route |
| Feed stream | Solids load, turbidity, viscosity, product concentration | Drives fouling, capacity and flux, and therefore sizing |
| Scale | Bench, pilot or production; single lot or campaign | Decides single-use versus reusable and the skid configuration |
| Cleaning | Cleaning and sanitisation regime, storage between runs | Affects material compatibility and validation evidence |
| Documentation | Integrity test, retention data, extractables, change control | Determines whether the unit can be qualified at all |
| Supply | Lead time, lot consistency, spare parts and consumables | Determines whether the process can keep running |
Equipment and Consumable Decisions
The membrane is only one part of the purchase. Listings for laboratory and process equipment are organised in the HHG store by category, which is the quickest way to see what is currently available. A filtration step also needs a holder or housing, a pump, pressure and flow instrumentation, tubing and connectors, and — at production scale — an automation layer that records the run. Each of these is a separate item with its own specification, and mixing them into a single “system” price is where comparisons usually go wrong.
It also helps to separate three things that are often quoted together: the device (pump, holder, skid), the consumable (membrane, capsule, cassette) and the spare part (seal, sensor, tubing set). They have different lifetimes, different documentation and different replacement economics.
For a used skid, add the questions that apply to any pre-owned capital equipment: service history, calibration status, software and firmware version, availability of consumables and spare parts, and whether the original validation package travels with the unit.
Records and Documentation to Request
In a regulated process the documentation is part of the product. Extractables and leachables expectations for plastic components are set out in USP <665>, and sterile filtration practice is covered by PDA technical reports. A membrane that works but cannot be qualified is not usable, and a skid without service records is a capital risk. The table below is the minimum a buyer should request before committing, whether the equipment is new or pre-owned.
| Document | What it proves | Why it matters at purchase |
|---|---|---|
| Membrane specification and lot data | Material, cut-off or pore rating, lot traceability | Confirms the unit matches the process description |
| Integrity test record | The filter or membrane is intact before use | Required for release and for deviation investigation |
| Bacterial retention data | Whether a sterilising filter retains organisms under test conditions | Supports any sterility claim; a pore size alone does not |
| Extractables and leachables information | What the product contact materials may release | Needed for toxicological risk assessment |
| Cleaning and sanitisation validation | The reuse regime is supported by data | Determines whether reusable formats are viable |
| Change control notification | The supplier will tell you before the product changes | Prevents silent changes to a validated process |
| Service and calibration history (used skids) | How the unit has been maintained | Supports acceptance testing and residual value |
Five Selection Mistakes
Most filtration problems that reach procurement are not membrane failures; they are selection failures that were locked in at the quotation stage. These five recur often enough to check before signing.
- Choosing by pore size alone. Two membranes with the same rating can differ in material, format and fouling behaviour.
- Using a sterilising filter as a clarification filter. It will foul early, and the sterility claim will not survive the run.
- Ignoring viscosity and solids load. Sizing on clean buffer data understates the membrane area needed for real feed.
- Mixing single-use and reusable cost models. Compare on cost per run at your actual campaign length, not on unit price.
- Buying without the records. If the integrity, retention and cleaning evidence does not exist, qualification becomes a project of its own.
Frequently Asked Questions
What is the difference between microfiltration and ultrafiltration?
Microfiltration removes particles such as cells and debris, typically in the 0.1 to 10 micrometre range. Ultrafiltration retains dissolved molecules such as antibodies while water and salts pass, and is defined by molecular weight cut-off rather than pore size.
Which one is used first in mAb production?
Microfiltration comes first, at harvest clarification, to protect the capture chromatography column. Ultrafiltration comes later, after purification, to concentrate the antibody and exchange it into formulation buffer.
Does a 0.2 micrometre filter sterilise a solution?
Only when the specific filter has been validated for bacterial retention with that product and process conditions. Pore size alone is not a sterility claim; the retention study and integrity test are what support it.
What records should I request when buying filtration equipment?
Ask for membrane specification and lot data, integrity test records, retention data where a sterility claim is made, extractables information, cleaning validation, change control notification, and — for used skids — service history, calibration records and spare parts availability.
If you are buying a used skid rather than a new one, the same documentation logic applies as for any capital purchase; our guide to the medical equipment procurement process sets out the sequence. More guides on sourcing, records and lifecycle cost are in the Knowledge section.
If you are specifying a filtration step or evaluating a pre-owned skid, send the process step and the documentation you already hold, and we will tell you what is missing before you commit.
Related articles
Remote Monitoring of Medical Equipment: What You Can and Cannot SeeKnowledge
HL7 and Device Integration When Buying Used Equipment: Compatibility Before PriceKnowledge
Transferring Medical Software Licences: What Moves With the HardwareKnowledge
Isolating Legacy Medical Devices on a Network: Compensating Controls That Hold UpKnowledge