A 96-position run on a Waters Alliance came back with a creeping baseline and three split peaks in the mid-polarity region. The method used a 0.1% formic acid / acetonitrile gradient at 40 °C, and the vials were capped with a bargain silicone/PTFE pre-slit septum I had approved to save four cents a unit. The instrument log showed nothing. The vial log showed everything: the silicone face had swollen, the slit had opened into a 1.2 mm tear by injection 60, and the septum was bleeding cyclic siloxanes into every vial in the tray. That is the whole problem with picking a septum material by habit. The polymer has to match the solvent, the pH, and the temperature of the method, or the failure shows up as a chromatography problem that you spend two days chasing in the wrong place.
This guide is about choosing the autosampler vial material for a specific solvent and temperature window. It is not about dimensions, neck finish, or price per case. Those live elsewhere on this site. What follows is the polymer-by-polymer decision: PTFE, silicone, EPDM, and the laminates built from them, what each survives, and the exact failure each wrong choice produces on the chromatogram.
Polytetrafluoroethylene is the most chemically inert polymer you will put near a sample. It resists essentially every HPLC solvent, from hexane to DMSO, across the full pH range from 0 to 14, and it holds up to 260 °C continuous. That is why almost every septum you buy has a PTFE face on the sample side. The mistake is treating a PTFE-faced septum as a PTFE septum.
Pure PTFE has no elasticity. It does not reseal a needle puncture. A solid PTFE disc will let a 22-gauge needle through and then sit there with a hole in it, and the hole does not close. If you are running a single injection per vial on a non-volatile method, that is fine. If you are pulling four injections from one vial, or the method involves any headspace pressure, a solid PTFE septum leaks from the second puncture onward. The autosampler vial that holds a volatile method needs a resealing layer behind that PTFE face, and that layer is where the chemistry decisions get made.
When PTFE alone is the right call
Solid PTFE or a PTFE/silicone laminate with the PTFE facing the sample is correct for aggressive solvents (chlorinated, aromatic, high-percentage DMSO or DMF), for strong acids and bases, and for any method where extractables from an elastomer would land in your analyte window. The trade is resealability. Plan on one to two punctures per vial, use pre-slit only if the method tolerates the slightly larger initial opening, and do not push a solid PTFE septum past two punctures no matter what the certificate says.
Silicone reseals beautifully and bleeds siloxanes into your method
Silicone (polydimethylsiloxane) is the workhorse backing layer because it is soft, it reseals a puncture cleanly for five to ten injections depending on needle gauge, and it tolerates -60 to 200 °C. It is also, chemically, a soup of low-molecular-weight siloxanes that migrate into the headspace and into the sample. On a GC method with a mass-selective detector, silicone bleed shows up as the classic m/z 73, 147, 207, 281 series. On an HPLC with UV detection it is quieter, but on a method with a low-wavelength channel (210 nm) you can watch the baseline lift across a tray as the septa warm up.
Silicone is also the wrong choice for non-polar solvents. Toluene, hexane, and dichloromethane swell silicone measurably; a septum that starts at 3.0 mm can grow past 3.6 mm in a hot toluene headspace, and a swollen septum is a septum that will not sit flat in the cap and will not reseal. The failure mode is a slow leak that looks exactly like a bad crimp.
The laminate is the actual product you buy
Almost every “silicone septum” on the market is a laminate: a thin PTFE film (typically 0.05 to 0.13 mm) bonded to a silicone core (1.0 to 2.0 mm). The PTFE face handles the chemistry; the silicone handles the reseal. The quality variable is the bond. A poorly bonded laminate delaminates at the puncture, the PTFE film peels back, and you get silicone exposed to the sample after the first injection. That is a silent failure: the first injection is clean and the fourth is contaminated. Ask your supplier for the bond peel strength and the film thickness, not just “PTFE/silicone.”
EPDM is the answer for aqueous, basic, and high-temperature methods
Ethylene propylene diene monomer is the polymer most labs underuse. It is a saturated-backbone rubber with no silicone in it, so it does not produce the siloxane bleed that ruins low-level GC/MS work. It handles aqueous buffers, dilute bases, and the pH 1 to 12 range comfortably, and it holds temperature to about 150 °C continuous. For a USP method running a phosphate buffer at pH 7.4 with UV detection at 220 nm, an EPDM-backed septum will outrun a silicone laminate on both bleed and reseal life.
Where EPDM loses is aromatic and chlorinated solvents. Toluene and dichloromethane swell EPDM badly, and the swell is not reversible. A method that switches between an aqueous mobile phase and a DCM extraction step should not use EPDM. The failure is a septum that has grown into the cap and either pops the cap off under headspace pressure or extrudes into the vial neck and blocks the needle path.
Temperature is a hard line, not a suggestion
Every polymer has a temperature above which it stops being the material on the certificate. Silicone is fine to 200 °C but starts to outgas aggressively above 180 °C. EPDM above 150 °C begins to harden and lose reseal. PTFE is the only one that goes to 260 °C without complaint, which is why high-temperature GC methods use a PTFE-faced septum with a PTFE or high-temperature silicone backing and nothing else. If your method runs the injector at 250 °C, the septum decision is already made for you.
Match the polymer to the solvent, the pH, and the temperature at once
The three variables interact. A method at pH 12 rules out silicone on chemical grounds before temperature even enters. A method at 220 °C rules out EPDM and silicone on thermal grounds before solvent enters. The table below is the decision I use, and it is built from the failure modes rather than from a supplier’s compatibility chart.
| Method condition | Right polymer | Wrong choice and the failure it produces |
|---|---|---|
| Aqueous buffer, pH 2 to 8, 30 to 60 °C, UV 210 to 254 nm | EPDM or PTFE/silicone laminate | Silicone-only: baseline lift and siloxane peaks at low UV |
| Dilute base, pH 10 to 12, ambient to 60 °C | EPDM or PTFE-faced laminate | Silicone: hydrolysis of the backing, tacky septum, needle drag |
| Aromatic or chlorinated solvent (toluene, DCM), ambient | PTFE-faced laminate, PTFE to sample | EPDM: irreversible swell, cap pop-off, needle path block |
| Headspace GC, 80 to 150 °C, volatile analytes | PTFE/silicone laminate, pre-slit, PTFE to sample | Solid PTFE: no reseal, pressure loss, poor area RSD |
| High-temperature GC injector, 200 to 260 °C | PTFE-faced high-temperature backing | Silicone or EPDM: thermal degradation, bleed, seal failure |
| Strong acid or strong base, pH 0 to 1 or 13 to 14 | PTFE face, inert backing | Silicone or EPDM: chemical attack, extractables, short reseal life |
Two operational numbers to keep on the bench. First, a standard crimp or screw cap should be torqued to a consistent value, and for a typical 11 mm crimp that is roughly 0.5 to 0.8 N·m on a calibrated crimper; over-torque deforms the septum and starts the failure before the first injection. Second, plan reseal life by polymer: PTFE-faced silicone 5 to 10 punctures, EPDM 3 to 6, solid PTFE 1 to 2. If your method runs more than that per vial, the polymer is not the problem, the sampling plan is.
Two field cases where the wrong polymer was the whole story
Case 1: the siloxane ghost in a pesticide panel
A contract lab running an SW-846 organochlorine pesticide panel on GC/MS kept seeing a peak at m/z 281 that co-eluted with a target compound and pushed the ion ratio out of the ±20% acceptance window. The instrument was clean, the column was new, and the standards were fresh. The measurement that found it: they ran a tray of blank vials with the same septa and saw the 281 peak appear at injection 40 and grow through injection 96. The septa were a silicone laminate with a thin PTFE film, and the film had been pierced through in the tray. The fix was a switch to an EPDM-backed septum, which has no siloxane to bleed; the 281 peak fell below the reporting limit and the ion ratios came back into tolerance. That single change saved a re-run of a 96-vial batch.
Case 2: the DCM swell that looked like a crimp fault
An environmental lab extracting semi-volatiles in dichloromethane saw intermittent low recoveries on the last three vials of every tray and a handful of caps sitting loose in the autosampler. The first assumption was a bad crimper. The measurement that identified it: they weighed a capped vial before and after a 30-minute DCM exposure and found the EPDM septum had gained 18% mass and grown enough to lift the cap. The DCM was swelling the EPDM, and the swell was worst on the vials that sat longest in the tray. The fix was a PTFE-faced laminate with the PTFE to the sample, which does not swell in DCM; the loose caps stopped and the recovery spread tightened from about 12% RSD to 4%. The crimper was never the problem.
The compatibility chart on the box is not the compatibility chart you need
Supplier compatibility charts rate a polymer as “good” or “excellent” against a solvent, and that rating usually reflects a 24-hour immersion test on a molded slab. Your septum is a 1.5 mm laminate under compression in a cap, pierced by a needle, sitting in a heated autosampler tray for eight hours. The failure modes that matter (swell under compression, reseal after puncture, bleed at temperature) are not what the slab test measures. Treat the chart as a starting filter, then confirm with a tray blank and a weigh-before/weigh-after on a few capped vials in the actual solvent at the actual temperature. That test takes an hour and catches the failures the chart cannot.
This is also where the incoming material check earns its place. Ask for the polymer grade (a generic “silicone” can be anything from a 40 Shore A to a 70 Shore A compound, and the softer one reseals better but swells faster), the film thickness on laminates, and the lot’s cure data. A supplier who cannot tell you the Shore hardness of the backing is selling you a septum you cannot predict. The same discipline applies across the vial range, including the headspace vial line where the septum is the entire pressure boundary, and it is worth reading how the vial and septum are specified together before you standardize on one laminate across every method.
A troubleshooting checklist for septum-driven failures
When a run misbehaves and the instrument checks out, work this list before you touch the method. Each item points at a specific polymer failure.
- Siloxane peaks at m/z 73/147/207/281 on GC/MS: silicone backing exposed or pierced through. Switch to EPDM or confirm the PTFE film is intact.
- Baseline lift at 210 nm across a tray: silicone bleed warming in the tray. Move to EPDM or a thicker PTFE face.
- Loose or popped caps: EPDM or silicone swell in aromatic/chlorinated solvent, or over-torque. Weigh a capped vial before and after solvent exposure.
- Low recovery on late-tray vials: evaporation through a failed reseal. Count punctures against the polymer’s reseal life.
- Needle drag or bent needles: septum extruded into the neck after swell or thermal softening. Check the polymer’s temperature rating against the tray temperature.
- Poor headspace area RSD (above 5%): solid PTFE with no reseal, or a pre-slit that has torn. Move to a PTFE/silicone laminate.
- Ghost peaks that appear only after injection 30: delamination of a cheap laminate. Check bond peel strength with the supplier.
If you run a mixed workload, the practical answer is not one septum for everything. Standardize on two: a PTFE/silicone laminate for the general reversed-phase and headspace work, and an EPDM-backed septum for the aqueous, basic, and low-level GC/MS methods where siloxane bleed is disqualifying. That covers the majority of methods without forcing a compromise. For the GC side specifically, the GC vial selection follows the same polymer logic, and the autosampler vial range on the main site lists the laminate options by method type. If you want the manufacturer’s own compatibility data as a cross-check, the Agilent consumables and Thermo Fisher sample handling references are useful starting points, and Restek’s technical resources have septum bleed data by polymer that is closer to real tray conditions than a slab chart.
The decision rule
Pick the septum by the harshest of the three conditions, not the average. If the solvent is aromatic or chlorinated, the sample side is PTFE and the backing is PTFE-tolerant, full stop. If the method is aqueous and basic or runs low-level GC/MS, the backing is EPDM because silicone will bleed into your window. If the temperature exceeds 150 °C, only a PTFE-faced high-temperature construction survives. Everything else is a PTFE/silicone laminate with the PTFE to the sample, a verified bond, and a puncture budget of ten or fewer per vial. Confirm the choice with one tray blank and one weigh-before/weigh-after in your actual solvent at your actual temperature before you commit a production method to it. The polymer is the cheapest variable in the method and the most expensive one to get wrong.
References & further reading
Related resources from our lab consumables network:
- Autosampler Vial at autosamplerhplc.com
- Vial at vials-filters.com
- Headspace Vial at headspacevialsupplier.com
- Gc Vial at gcvialfactory.com
External references and standards: