Sterilising a custom culture device: five methods, and what each one does to your part

Sterilisation is the last line in a device quote and the first thing that goes wrong. It is specified as a step applied to a finished part, when in practice it is a process that changes the part — the dimensions, the surface, the mechanical properties, or all three — and the change is often invisible until an experiment behaves differently from the one before it.

The decision has to be made at design time, because the sterilisation method constrains the material, and the material constrains the sterilisation method.

The five routes

The first thing to notice is that the two methods available in a normal laboratory — autoclave and ethanol/UV — sit at opposite ends of the reliability spectrum. The autoclave is a true sterilisation process and it is the one most likely to damage the part. Ethanol and UV preserve the part and do not achieve terminal sterility.

What sterile actually means on a catalogue part

Vendors state this three different ways, and they are not equivalent:

• “Shipped sterile from cleanroom production.” eNUVIO’s position on the OMEGA line: the part is made in a cleanroom and never contaminated, rather than made and then decontaminated. No terminal process, no residuals, no dose.
• “X-ray irradiated.” Kugelmeiers on the SP5D. A named terminal process, applied through the PET-G tray and Tyvek cover.
• “Sterile, SAL 10⁻³.” Corning’s stated sterility assurance level for the Elplasia 12K. A probability statement: no more than one non-sterile unit in a thousand. This is the only one of the three that is quantitative, and 10⁻³ is the normal level for research consumables — implants and invasive devices work to 10⁻⁶.

When a fabricator says a custom device will be “supplied sterile”, ask which of those three sentences they mean. The answer determines whether you have a validated process, a cleanroom assertion, or a hope.

Material by material

PDMS. The best-documented case, and the guidance is unusually clear. Autoclaving, dry heat and gamma radiation are all effective but may alter the mechanical properties of PDMS; published protocol work therefore recommends ethanol immersion and UV irradiation specifically because they do not affect mechanical integrity, with oxygen plasma as a third option that also preserves mechanical properties where the equipment exists.

Why this matters more than it sounds: for any device whose function depends on PDMS stiffness — a post-and-pillar force readout, a deflecting membrane, a valve — the modulus is the calibration. Sylgard 184’s elastic modulus ranges from roughly 1 to 4 MPa depending on cure temperature between 25 °C and 200 °C, so a material whose stiffness is that sensitive to thermal history should not be put through repeated 121 °C cycles without re-checking the force calibration. A device that reads force through post deflection and has been autoclaved three times is not the device that was calibrated.

Thermoplastics — COP, COC, PMMA, PS. The binding constraint is thermal. Autoclaving at 121 °C approaches or exceeds the useful service temperature of common grades of PMMA and polystyrene, and COP/COC grades vary widely in glass transition temperature, so “COP is autoclavable” is not a statement that can be made about the material class, only about a specific grade. For these materials the realistic routes are gamma/X-ray or EO, both of which are contract services — which is one of the hidden costs of choosing a thermoplastic for the right reasons (low absorption, scalable moulding) as set out in the material guide.

Glass and PTFE. Autoclave freely. This is one of the reasons glass persists in applications where the fabrication cost is otherwise indefensible.

3D-printed resin parts. A special case, dealt with on the 3D printing page: heat can continue to cure or distort a printed part, and the leachables question is entangled with the sterilisation question rather than separate from it.

The ethylene oxide residual problem

EO is chosen for one reason — extensive material compatibility — and it carries one cost, which is that it leaves residues in the part.

Three residues form:

• Ethylene oxide (EO) — what remains after processing.
• Ethylene chlorohydrin (ECH) — forms when EO meets free chloride ions.
• Ethylene glycol (EG) — forms when EO meets water. Not separately limited, on the reasoning that if EO residues are controlled, biologically significant EG residues are unlikely.

The published limits under AAMI/ISO 10993-7 for patient-contacting devices give the scale of the concern:

Read the scope carefully. The standard states that devices without patient contact are not required to meet it, and a research-use culture device is not a patient-contacting device. The limits are reproduced here not because they bind you, but because they establish that residual EO is a real quantity in real parts at milligram scale — and your cells are considerably more sensitive than the irritation threshold those limits were set against. There is no published allowable EO residual for a cell culture device, which means the honest position is: if you EO sterilise a device that will hold cells, the aeration is your problem to specify and your risk to carry.

The published drivers of high residuals are worth knowing because they are all design and packaging decisions:

• Material. Some plastics absorb and retain gas heavily.
• Packaging. EO needs breathable packaging — Tyvek or medical-grade paper. A large adhesive label over the breathable area obstructs both gas entry and gas escape.
• Load configuration. Dense loads off-gas poorly.
• Aeration. The industry average is 6 to 24 hours in heated aeration; some devices need more.

If you go this route, specify aeration explicitly and ask for the residual test method. A steriliser can extend aeration, add nitrogen washes at deeper vacuum, or raise the process temperature to drive off-gassing — but only if asked.

Sterilisation destroys surface treatment

This is the interaction that catches people, and it is worth stating on its own.

Hydrophobic recovery of plasma-treated PDMS is accelerated by bonding and heat treatment. An autoclave cycle is a heat treatment. So the sequence “plasma treat → bond → autoclave → ship → store for three weeks → use” delivers a device whose surface is, functionally, untreated PDMS.

The consequences are concrete: channels that will not fill, air locks in narrow features, uneven seeding. And because the geometry is perfect and the part looks identical, the diagnosis usually takes a batch or two.

The rule: if the device depends on a treated surface, either sterilise before treating, or use a method that does not heat the part, or plan to re-treat after sterilisation. Whichever you choose, the sequence belongs in the specification. See surface treatment and coating for what survives what.

Choosing, in order

1. Can the part survive 121 °C? If yes, and the surface state does not matter, autoclave. This is by far the cheapest answer and it is available in your building.
2. Is the part single-use and made in volume? Then terminal sterilisation through final packaging — gamma, X-ray or EO — is the right industrial answer and it is a contract service with a lead time to add to your schedule.
3. Is the material thermally limited but geometrically simple and open? Ethanol immersion with UV is what most laboratories actually do for research-use parts. Be honest in the methods section about what it is: high-level disinfection, not validated sterilisation.
4. Is there an enclosed channel? UV cannot reach it. Plasma cannot reach it. You are down to heat, gas, radiation, or a flow-through chemical treatment with a validated rinse.
5. Does anything about the device’s function depend on polymer mechanical properties? Then avoid heat and radiation, or re-calibrate afterwards and say so.

What to write into the specification

• Method, by name, and whether it is validated to a named standard (ISO 17665 for moist heat, ISO 11137 for radiation, ISO 11135 for EO) or is a non-validated laboratory process.
• Sterility assurance level, if claimed. If nobody will state one, the part is “aseptically produced”, not “sterile”, and the two words should not be used interchangeably.
• Where sterilisation sits in the sequence relative to bonding, surface treatment and any coating.
• Packaging, and whether it is compatible with the method — breathable for EO, sealed for everything else.
• Shelf life of the sterile barrier, separate from the shelf life of the part. eNUVIO publishes a two-year shelf life on the OMEGA-MP stored at room temperature away from UV; a custom part with no stated shelf life is an open question you have silently inherited.
• Whether the part can be re-sterilised by the user, how many times, and what changes each time. For any device with a mechanical function, “how many times” has a real answer and it is often one.
• Residual testing, if EO is used.

How we can help

We do not sterilise anything. What we do is make sure the sterilisation questions are in the specification before the quotes come back, so you can see which fabricators have a validated process, which will ship from a cleanroom, and which are quietly leaving the problem with you. If you already have a device and are trying to work out why the second batch behaved differently from the first, sterilisation and surface treatment are the first two places to look, and we can put that question to shops that have the process capability to answer it. Send us the specification.