0.22 µm Syringe Filters and Luer Fittings: Membrane Choice, Adsorption Loss and What to Stock

For a dilute aqueous peptide solution the membrane is chosen for low non-specific binding first and chemical compatibility second, which narrows the practical field to hydrophilic polyethersulfone (PES) or a hydrophilic surface-modified polyvinylidene fluoride (PVDF). Nylon and unmodified PTFE are the wrong answers, for different reasons. The harder half of the question is the half most selection charts skip: the filter keeps some of what passes through it. A 0.22 µm rating is a bacterial-retention specification, not a guarantee that everything smaller travels through unimpeded. The internal surface of a microporous membrane is very large relative to the housing that holds it, and a peptide at low concentration begins coating that surface before the first drop reaches the receiving vessel.

What follows is about hardware: what the membrane, the housing and the fitting do to recovery, and what the ratings printed on the packet actually certify. It is deliberately not about what time and temperature do to the molecule — how long a peptide holds up once it is in solution is a separate problem with its own literature. The register throughout is laboratory practice, for material supplied for research use only.

What a 0.22 µm rating actually certifies

The number is a performance classification, not a measured hole. ASTM F838 defines the test: a challenge with Brevundimonas diminuta, an organism roughly 0.3 µm across, at a concentration of at least 1 × 10⁷ colony-forming units per square centimetre of effective filtration area, with a sterile filtrate required for the device to pass ^[1]. The U.S. Food and Drug Administration uses that same definition of a sterilising-grade filter in its aseptic processing guidance ^[2]. The challenge level sits far beyond any realistic bioburden on purpose, so the test detects oversized pores rather than merely confirming ordinary performance ^[1].

"0.2 µm" and "0.22 µm" denote the same grade; the printed digit is manufacturer nomenclature, not a demonstrated performance gap. What genuinely differs between two filters carrying the same number is what was validated, and against what. FDA expects retention validation with the actual product under worst-case conditions, including maximum use time and pressure ^[2]. The European Medicines Agency asks for the same evidence on the actual solution and — the part that matters here — for data on product-filter compatibility, explicitly including adsorption of product components onto the membrane and extractables released from it ^[3]. USP General Chapter 1229.4 states the dependency plainly: removal depends on the bioburden, the properties of the solution, the filtration conditions and the filter itself, not on the rating alone ^[4].

Three things the rating does not cover. Bacterial endotoxin is not retained — it is a lipopolysaccharide that passes a 0.22 µm membrane freely, so filtration is not a depyrogenation step; establishing endotoxin status takes its own assay. Mycoplasma, at roughly 0.15 to 0.3 µm and highly deformable, can traverse 0.2 µm rated membranes, which is why a separate 0.1 µm class exists and is challenged with Acholeplasma laidlawii instead ^[5]. And integrity is a different measurement from retention: the bubble-point method in ASTM F316 infers maximum limiting pore diameter from the pressure needed to displace a wetting liquid from the largest pore, and a disposable syringe filter is almost never integrity-tested by its user ^[6].

Where the peptide goes: adsorption on the membrane

The loss mechanism is not filtration in any useful sense — it is adsorption. A microporous membrane presents an internal surface area orders of magnitude greater than its visible face, and all of it is available for non-specific binding. Aldo Pitt's 1987 study in the Journal of Parenteral Science and Technology framed the problem the way it still needs framing, measuring non-specific binding and product recovery for four representative proteins — porcine insulin, human chorionic gonadotropin, bovine serum albumin and sheep immunoglobulin G — across common microporous membranes, and showing recovery to be a function of the polymer the membrane is made from ^[7].

Peptide-specific evidence points the same way and is blunter about magnitude. Quantifying by HPLC, Kristensen and colleagues reported that 90% or more of three cationic membrane-active peptides could be lost from solution through rapid adsorption to ordinary glass and plastic containers ^[9]. That is in vitro work on containers rather than filters — but a filter housing is a container with vastly more internal surface, and the chemistry of peptide lost to a container wall is the chemistry at work inside a membrane. Goebel-Stengel and colleagues incubated radiolabelled peptides in glass and plastic tubes over 48 hours and found binding to be both peptide-dependent and surface-dependent, concluding that the combination has to be checked empirically rather than predicted from the polymer name ^[8].

Two further studies push this into the fluid path itself. Van Wanseele and colleagues measured recovery of three neuromedin-like peptides delivered from glass and plastic syringes, and found both the syringe material and the perfusion solvent changed how much peptide emerged ^[10]. On filters directly, a 2020 study in Mycotoxin Research found that passing a complex mixture through syringe filters caused analyte losses that varied with membrane material — enough to alter the in vitro genotoxic readout of the filtrate ^[11]. Those were small-molecule mycotoxins in a cell assay, not peptides and not an animal or human study; the transferable finding is the mechanism, not the percentages.

Be plain about what does not exist. There is no published, peer-reviewed head-to-head comparison of PES against hydrophilic PVDF covering the research peptides in circulation. Most comparative recovery data offered to buyers comes from manufacturer application notes run on model proteins such as albumin or myoglobin, which are not peptides and do not behave like them. A single recovery percentage quoted for "peptides" as a class is an extrapolation.

Membrane materials and what each one is for

Selection is a two-axis problem: will the solvent attack the polymer, and will the polymer hold on to the solute. For aqueous work the second axis dominates, because water attacks none of these materials.

PES or PVDF: what the evidence supports

Both are legitimate, and both are marketed as "low protein binding". That phrase is not defined by any standard: no threshold, no test method, no unit. Two manufacturers can apply it to membranes that differ measurably in recovery. Treat it as a category label, and treat a datasheet offering it without a measured figure as having said nothing.

The real differences: PES gives higher flow at a given pressure than most alternatives and tolerates a wide pH range, which is why it is the aqueous default; its liability is a known tendency to release UV-absorbing extractables, which matters when the filtrate feeds a low-wavelength UV measurement. Hydrophilic-modified PVDF is more chemically resistant and copes with mixed aqueous-organic systems that would soften cellulose esters; its liability is that the surface treatment is what makes it low-binding, so the hydrophobic version of the same polymer behaves entirely differently and will simply refuse an aqueous sample.

Which of the two recovers more of a given peptide depends on that peptide's net charge and hydrophobicity, on its concentration, and on the solvent — all three shown to be decisive in the cited work ^[8][9][10]. The defensible procedure is not to declare a winner but to fix the variable: one membrane, one device diameter, one supplier, lot recorded, and any change treated as a change of method that resets comparability with earlier work.

Extractables, pre-rinsing and the first fraction

Everything in the fluid path leaches something. USP General Chapter 665, official since 1 May 2026, formalises evaluation of plastic components used in pharmaceutical manufacture precisely because polymeric contact surfaces release process-equipment-related leachables into the fluid they touch ^[13]. A syringe filter is a small instance of the same problem — membrane, housing, support layer and adhesive all contribute — where contact time is short but the surface-to-volume ratio is extreme.

• Pre-wet and pre-rinse with the same solvent the sample is in, then discard that volume. It flushes loosely bound extractables and begins occupying the binding sites that would otherwise take the leading edge of the sample ^[7].
• Discard the first fraction of filtrate rather than the last. Adsorption is front-loaded: the earliest liquid through meets the most unoccupied surface ^[7].
• Run a filtered blank — solvent alone through an identical device from the same lot — whenever the filtrate feeds a measurement an extractable could confuse ^[13].
• Record membrane, diameter, manufacturer and lot alongside the material lot. Regulators treat filter-product compatibility as data rather than assumption, and a filtration step not written down is not a controlled step ^[3].

Hold-up volume: the loss you can calculate

Adsorption is hard to predict. Hold-up volume is not. Every device retains liquid in the membrane, the support and the outlet channel after the plunger bottoms out, and the figure is stated on the datasheet. Across common devices the published ranges run roughly as follows: a 13 mm filter, with something like 0.6 to 1.0 cm² of effective filtration area, holds up on the order of 10 µL; a 25 mm filter, at roughly 3.5 to 4.5 cm², on the order of 75 to 125 µL; a 33 mm device more again. Read the number from the packet, not from this paragraph — it varies by manufacturer, membrane and housing design.

The arithmetic is unforgiving at small volumes. Pushing 1 mL through a 25 mm device can strand roughly a tenth of the liquid inside the housing before adsorption takes anything at all, and that loss is not recovered by pushing air through afterwards: air clears the outlet channel but does not displace liquid held in the wetted membrane. Hence the rule — use the smallest diameter that will pass the intended volume without blinding, and size up only when the membrane clogs before the syringe empties. The larger device costs volume every time; the smaller one costs only when it actually blocks.

Luer lock or Luer slip, and why the standard changed

The fitting is standardised, and the standard is newer than most people assume. ISO 594-1 and ISO 594-2, which governed conical Luer fittings for decades, have been withdrawn and superseded for intravascular and hypodermic connections by ISO 80369-7, whose current edition was published in 2021 ^[12]. It specifies dimensions and functional performance for the 6% Luer taper connectors used on syringes, needles and the devices that join them. The wider ISO 80369 series exists to make different connector families physically non-interconnectable, which is why its tolerances are tighter than those of the standard it replaced ^[12].

Geometrically, a syringe filter is a female inlet and a male outlet. The inlet accepts the male tip of the syringe; on a lock version it also carries an internal thread that engages the syringe's outer collar. The outlet is usually a male Luer slip, so filtrate can be directed into a receiving tube. Slip is a friction fit between two matched tapers and nothing more; lock adds a threaded collar that resists axial separation. The distinction stops being cosmetic the moment the membrane begins to resist.

• Choose Luer lock when the membrane is likely to blind. Pressure rises as a filter clogs, and a slip joint under pressure can separate abruptly; the lock exists for that failure mode.
• Choose slip only where resistance is low and the connection is momentary. It is faster and there is nothing to cross-thread.
• Match the syringe to the filter. A slip-tip syringe seats in a lock inlet but cannot engage the thread, leaving a friction fit that feels secure and is not.
• Do not improvise adapters between connector families. The purpose of the current standard is that families which should not connect physically cannot ^[12].
• Read the stated maximum operating pressure and stop when the plunger stops moving easily. Forcing a blinded membrane past its rating is how housings crack and seals extrude.

What to stock

• One low-binding membrane as the house default — hydrophilic PES or hydrophilic-modified PVDF — in a single diameter from a single supplier, so recovery behaviour stays constant across a series of work.
• The smallest sensible diameter for the volumes actually handled, with a larger diameter held in reserve for viscous or particle-laden solutions.
• Sterile, individually packaged devices where sterility of the filtrate matters. Bulk non-sterile devices are a different product and are cheaper for a reason.
• A record of the sterilisation method: gamma-irradiated and ethylene-oxide-sterilised devices are both common, and the latter are subject to allowable residual limits for ethylene oxide and its reaction products under ISO 10993-7 ^[14].
• A separate, clearly labelled hydrophobic PTFE option if organic-solvent work happens on the same bench, so it never reaches an aqueous sample by accident.
• The paperwork: stated membrane, rating, effective filtration area, hold-up volume, maximum operating pressure and lot number, filed where the filtration step is recorded ^[3].

Consumables are the least glamorous line in a laboratory's records and among the most consequential, because they sit between the material and every number taken from it. A device that adsorbs a third of a dilute peptide does not announce itself; it quietly makes every downstream measurement wrong by a constant nobody wrote down. The compounds these consumables are used with are supplied for research use only. They are not medicinal products, they are not intended for administration to humans or animals, and nothing above should be read as guidance for anything other than handling laboratory material under the conditions its documentation describes.