Start with the workflow, not the catalog

I have watched a lab lose three weeks of stability data because someone swapped a PTFE-faced silicone septum for a plain silicone one on a 40 mL volatile organic analysis vial. The samples sat in an autosampler tray at ambient for six hours, the silicone absorbed the analytes, and the recovery dropped below 70% across the board. Nobody caught it until the audit. That is the kind of failure a $0.09 component can cause, and it is why I treat septum selection as a specification exercise, not a shopping decision.

This guide covers what actually matters when you buy and use vial septa: the elastomer chemistry, the dimensions that have to match your crimper and your autosampler, the temperature and puncture limits that govern when you throw a septum away, and the torque and crimping numbers that keep a seal tight without cracking a neck. I will give you the numbers I use, the failure modes I have dealt with, and a decision checklist you can hand to a procurement team.

Every septum decision comes down to four variables: the sample chemistry, the temperature the vial will see, how many times a needle will pierce it, and the closure system you already own. Get those four right and the catalog narrows to two or three part numbers. Get one wrong and you get ghost peaks, lost volatiles, coring, or a bent needle.

Before you open a supplier page, write down:

  • Sample matrix: aqueous, organic, volatile, pH extremes, or biological.
  • Maximum temperature: autosampler tray (ambient), GC inlet (up to 300 °C), headspace oven (40 to 200 °C), or a hot block for derivatization.
  • Puncture count: single injection, or 10 to 50 injections from one vial on a multi-injection autosampler.
  • Closure: 11 mm crimp, 11 mm snap, 13 mm crimp, 18 mm crimp (headspace), or 9 mm screw thread.
  • Needle gauge: most autosamplers use 23 gauge (0.64 mm OD) or 22 gauge (0.72 mm OD); some use 26 gauge for small-volume work.

That list is the whole game. When a lab calls me with a chromatography problem, the first question is always which septum is on the vial, and the answer usually reveals which of these five they never wrote down.

The elastomer layer: what each material does and does not do

A standard septum is a sandwich: an elastomer disc that provides the reseal, and often a thin facing layer on the sample side that provides chemical resistance. The facing matters more than most buyers realize. Here is how the common materials behave.

Silicone

Excellent reseal, very good puncture recovery, wide temperature range (roughly -60 °C to 200 °C for standard grades). The problem is permeability. Silicone absorbs nonpolar volatiles, and it lets oxygen and moisture through. For a volatile organic analysis (VOA) sample or a headspace vial, bare silicone is the wrong choice. For aqueous samples at ambient temperature with a single injection, it is fine and cheap.

PTFE-faced silicone

This is the workhorse for HPLC and most GC autosamplers. The PTFE facing (typically 0.05 to 0.13 mm thick) gives chemical inertness against acids, bases, and most organic solvents, while the silicone underneath provides the reseal. Temperature limit is set by the silicone, so 200 °C is a realistic ceiling. If you run reversed-phase HPLC with aqueous buffers and acetonitrile, this is your default. The facing also stops the silicone from absorbing your analytes.

PTFE/silicone/PTFE (three-layer)

Facing on both sides. Useful when the septum contacts liquid on both faces (for example, a vial inverted in a tray, or a two-sided closure). Costs more, reseals slightly less well because the needle has to push through two PTFE films. I use it for corrosive samples and for long-term storage where the top face sees solvent vapor.

Butyl rubber

Low permeability to gases and moisture, which is why it shows up in pharmaceutical vial closures and lyophilization. Temperature limit is lower than silicone, around 100 to 120 °C. Not a good choice for repeated autosampler puncture; the reseal degrades faster than silicone.

Bromobutyl and chlorobutyl

Pharmaceutical-grade butyl variants with tighter extractables profiles. If you are doing USP <661> container closure testing or working under EPA SW-846 methods with strict blank requirements, the extractables matter. Ask the supplier for the extractables data sheet, not just the material name.

EPDM

Good chemical resistance, moderate temperature limit (about 135 °C), and lower cost than butyl. Common in aqueous and mildly acidic applications. Not for nonpolar solvents.

Natural rubber and red rubber

Cheap, high permeability, and a source of sulfur and accelerator extractables. I avoid these for anything analytical. They belong on sample storage vials, not on an instrument.

Dimensions: the numbers that must match

Septum dimensions are not interchangeable, and the tolerances matter. A septum that is 0.2 mm too thick will not seat in a crimp cap; one that is 0.3 mm too thin will leak under inlet pressure. Here are the standard sizes I keep on the shelf.

  • 11 mm crimp: septum OD 11.0 mm, thickness 1.0 mm (standard) or 1.5 mm (thick). Fits 11 mm aluminum crimp caps and 11 mm snap caps. The most common GC autosampler format.
  • 11 mm snap: same 11.0 mm OD, but the cap geometry differs; snap caps need a septum sized for the snap ring, usually 11.0 mm x 1.0 mm with a specific durometer.
  • 13 mm crimp: septum OD 13.0 mm, thickness 1.0 to 1.5 mm. Larger neck, used on some older GC systems and on 15 mL vials.
  • 18 mm crimp (headspace): septum OD 18.0 mm, thickness 1.5 to 3.0 mm. The thick septum is deliberate: it must hold inlet pressure up to 2 to 3 bar in a headspace sampler and reseal after a 23 gauge needle.
  • 9 mm screw thread: septum OD 9.0 mm, thickness 1.0 mm. Fits the 9 mm short-thread screw caps common on 2 mL HPLC vials.

The closure thread standard you will see referenced is the GPI 9 mm and GPI 13 mm (Glass Packaging Institute) screw thread, and for pharmaceutical closures the ISO 8362-1 vial and ISO 8362-3 closure standards. If a supplier cannot tell you which thread standard their cap meets, that is a red flag. Mismatched thread pitch is a common cause of a cap that “feels tight” but leaks.

Thickness is not cosmetic. A 1.5 mm septum in an 11 mm crimp cap reseals better after repeated puncture than a 1.0 mm septum, but it also sits higher and can interfere with some autosampler grippers. If your autosampler vial tray has tight gripper clearances, measure before you switch to thick septa.

Temperature ratings: the real limits, not the marketing ones

Suppliers quote a temperature range, and buyers treat it as a hard limit. It is not. It is the point at which the material starts to degrade, not the point at which it stops working. I derate every quoted maximum by 20 to 30 °C for anything that has to hold pressure or reseal repeatedly.

  • Silicone, PTFE-faced: quoted 200 °C, use up to 180 °C for repeated injection. Above 200 °C the silicone starts to outgas and you get siloxane peaks on a GC-MS baseline.
  • Butyl: quoted 120 °C, use up to 100 °C.
  • EPDM: quoted 135 °C, use up to 120 °C.
  • PTFE facing alone: stable to 260 °C, but it does not reseal, so it is only useful as a facing, not a standalone septum.

If your GC inlet is at 280 °C and the septum sits in the inlet body, you need a high-temperature septum, and you need to change it on a schedule. A septum that has been at 280 °C for 200 injections is a source of ghost peaks, and no amount of column conditioning will fix it.

Puncture counts: when to throw it away

This is the number suppliers rarely publish and operators rarely track. A septum has a finite number of reseals before the needle track leaks. Typical values I work to:

  • Standard 1.0 mm PTFE/silicone, 11 mm: 50 to 100 punctures before leakage becomes detectable, assuming a 23 gauge needle and ambient temperature.
  • Thick 1.5 mm PTFE/silicone, 11 mm: 100 to 150 punctures.
  • 18 mm headspace septum, 3.0 mm: 50 to 75 punctures at 2 bar, fewer if the sampler pre-pressurizes aggressively.
  • Bare silicone: 30 to 50 punctures.

Those numbers drop fast with a larger needle, a higher temperature, or a solvent that swells the elastomer. A 22 gauge needle at 150 °C will halve the puncture count. If you run a multi-injection method, put a puncture counter in your SOP and replace the septum on the vial, not just the cap, at the stated limit. I have seen labs reuse a septum for 300 injections because “it looked fine,” and the only symptom was a slow drift in peak area that got blamed on the detector.

Crimping and torque: the numbers that prevent leaks

A septum only seals if the crimp is right. Under-crimp and the vial leaks at the neck; over-crimp and you crack the glass or deform the septum so it never reseals. The numbers I use:

  • 11 mm aluminum crimp: crimp with a manual or pneumatic crimper set to a cap height of 5.5 to 5.7 mm after crimping. The crimp should be tight enough that the cap does not rotate by hand but loose enough that the vial does not chip. A good test is to crimp a vial with water and invert it; no leak at the seal after 30 seconds.
  • 13 mm crimp: cap height 6.0 to 6.3 mm after crimping.
  • 18 mm headspace crimp: cap height 7.5 to 7.8 mm. Headspace vials are less forgiving; a 0.3 mm error shows up as a pressure leak at 2 bar.
  • 9 mm screw cap: torque 0.6 to 0.9 N·m (about 5 to 8 in·lb). Hand-tight plus a quarter turn is a rough guide, but a torque wrench is better. Over-torque deforms the septum and cracks the thread.

I keep a go/no-go gauge on the bench for each crimp size. It takes ten seconds per vial and catches the crimper drift that happens after a few thousand cycles. If you buy pre-crimped vials, ask the supplier for their crimp height tolerance; a supplier who cannot quote it is not controlling the process.

Chemical compatibility: the short version

You do not need a 40-page compatibility chart. You need to know the four failure modes:

  • Silicone swells in nonpolar solvents (toluene, hexane, dichloromethane). The septum grows, the crimp loosens, and the seal fails. Use a PTFE facing.
  • Butyl and EPDM degrade in strong acids and oxidizers. Use PTFE-faced silicone or a PTFE-lined closure.
  • Natural rubber leaches sulfur and accelerators into aqueous samples. Never use it for trace analysis.
  • Any elastomer absorbs volatiles. For VOA and headspace work, the facing layer is doing the real work, and a thicker facing (0.13 mm vs 0.05 mm) reduces absorption.

For method-specific guidance, the EPA SW-846 methods (particularly 5030 and 8260 for volatiles) spell out the container and septum requirements, and the Restek technical resources have compatibility tables that are more practical than most supplier charts. The Agilent GC accessories pages list the septum part numbers matched to each inlet, which is a useful cross-check when you are specifying a new instrument.

Decision checklist: matching septum to workflow

Here is the table I hand to procurement. It converts the workflow into a part specification with numbers.

  • Routine reversed-phase HPLC, aqueous/organic, ambient tray, single injection: 9 mm or 11 mm PTFE/silicone, 1.0 mm thick, 200 °C rated. Replace at 50 punctures or 7 days in tray.
  • GC autosampler, 250 to 280 °C inlet, 23 gauge needle: 11 mm PTFE/silicone, 1.0 mm, high-temperature grade, inlet septum changed every 100 injections or weekly.
  • Headspace, 2 bar, 18 mm crimp: 18 mm PTFE/silicone, 3.0 mm thick, crimp height 7.5 to 7.8 mm, replace at 50 punctures.
  • VOA / volatile organics, EPA 8260: 18 mm or 20 mm PTFE-faced silicone with 0.13 mm facing, no headspace, crimp checked with go/no-go gauge.
  • Pharmaceutical stability, USP <661>: butyl or bromobutyl, extractables data sheet required, ISO 8362-3 closure, torque 0.6 to 0.9 N·m for screw caps.
  • Corrosive samples (strong acid, strong base): PTFE/silicone/PTFE three-layer, 1.5 mm thick, replace at 30 punctures.
  • Long-term storage, ambient: butyl or PTFE-faced silicone, 1.5 mm, check seal monthly.

That list is not exhaustive, but it covers 90% of the requests I get. If your workflow is not on it, the answer is almost always “PTFE-faced silicone, 1.0 or 1.5 mm, and derate the temperature.”

Buying: what to ask, what to check

Price per septum ranges from about $0.05 for a plain silicone disc to $0.40 for a three-layer PTFE-faced 18 mm headspace septum. The spread is large enough that procurement will push for the cheap end, and that is where the failures come from. Here is what I ask before I approve a supplier.

Ask for the certificate of analysis

A real CoA lists the lot, the material, the durometer (typically Shore A 40 to 55 for silicone septa), the thickness tolerance (±0.05 mm for good suppliers), and the extractables profile. If the CoA is a generic PDF with no lot number, you are buying on faith.

Check the facing thickness

PTFE facings come in 0.05 mm and 0.13 mm. The 0.13 mm facing costs more and resists absorption better. For trace analysis, it is worth it. For general HPLC, 0.05 mm is fine. Ask which one you are getting; the catalog often does not say.

Confirm the durometer

Softer silicone (Shore A 40) reseals better but extrudes under pressure. Harder silicone (Shore A 55) holds pressure but reseals less well. For headspace at 2 bar, go harder. For repeated injection at ambient, go softer.

Check the packaging

Septums ship in bags of 100 or 1000. If the bag is not cleanroom-compatible or is not resealable, you will pick up dust and fibers, and those show up as particles in your HPLC. I specify individually bagged or cleanroom-packed septa for anything going into a mass spec.

Troubleshooting checklist

When a method goes bad, run this list before you touch the column.

  • Ghost peaks on GC-MS: septum bleed. Change the inlet septum, then bake the inlet. If siloxane peaks (m/z 73, 207, 281) persist, the septum is the source.
  • Lost volatiles / low recovery: wrong facing or too-thin facing. Check that the septum is PTFE-faced and that the facing is on the sample side.
  • Leak at the neck: crimp height out of spec or wrong septum thickness. Check with the go/no-go gauge and re-crimp a test vial.
  • Cracked vial necks: over-crimp or over-torque. Reduce crimper setting or torque to 0.6 to 0.9 N·m.
  • Bent autosampler needles: septum too thick or coring. Switch to a thinner septum or a higher-durometer grade, and check needle alignment.
  • Coring (a plug of septum pushed into the vial): dull needle or a septum that is too soft. Replace the needle and move to Shore A 50 or higher.
  • Particles in HPLC: septum fragmentation or dust from packaging. Switch to cleanroom-packed septa and check the needle for burrs.
  • Slow drift in peak area over a run: septum leaking after too many punctures. Count punctures and replace on schedule.

For instrument-specific part numbers, the Thermo Fisher chromatography sample handling pages and the Agilent accessories pages are the two I check first, because they list the septum by inlet model rather than by generic size.

Where to buy and how to standardize

The vial septa catalog at vialsepta.com covers the 9 mm, 11 mm, 13 mm, and 18 mm formats in silicone, PTFE-faced silicone, butyl, and EPDM, with the thickness and durometer called out per part. If you are buying the whole closure (vial, septum, cap), the vial range at vials-filters.com is the companion piece, and it is worth buying vial and septum from the same supplier so the crimp height and thread pitch are matched.

Standardize on two or three part numbers per lab. One for routine HPLC, one for high-temperature GC, one for headspace. Write the puncture limit and the replacement interval into the SOP. Then audit it: pull ten vials from the autosampler at the end of a run and check the crimp height with the gauge. The labs that do this spend less on septa and far less on repeat analysis.

Takeaway

Septum selection is a specification problem with four inputs: chemistry, temperature, puncture count, and closure size. Match the facing to the sample, derate the quoted temperature by 20 to 30 °C, replace at the published puncture limit rather than when it looks worn, and control the crimp height to within 0.2 mm. A PTFE-faced silicone septum at 1.0 or 1.5 mm covers most workflows; butyl and EPDM are for specific pharmaceutical and aqueous cases; bare silicone belongs on storage vials, not on an instrument. Buy on the CoA and the facing thickness, not the price per bag, and the ghost peaks and low recoveries mostly go away.

References & further reading

Related resources from our lab consumables network:

External references and standards: