A customer sent back 1,200 vials last March because their BTEX run on a purge-and-trap concentrator started showing a methylene chloride ghost at 4.2 µg/L on every blank. The septa were silicone-backed, the method ran at 40 °C for 11 minutes, and the liner was clean. The contamination came from the adhesive layer between the silicone and the PTFE film, which had been slowly off-gassing into the headspace of every vial since the lot was opened. That is a material selection failure, not a manufacturing defect. Choosing which voa vial material to use for a given solvent and temperature is the single decision that determines whether your blanks stay clean or your QC charts drift for a month.
The material in a VOA vial is not one polymer. It is a laminate, a septum, a liner, and sometimes a coating, and each layer has its own solvent compatibility, pH window, and temperature ceiling. Get the laminate right and pick the wrong septum, and you still fail. Get both right and run the method at 5 °C above the rated ceiling for six hours, and you fail again. What follows is the working map I use when a lab asks which construction belongs under which method.
There are four material classes you will encounter on a VOA vial spec sheet: PTFE, silicone, EPDM, and the laminate constructions that combine them. They are not interchangeable, and the published chemical resistance charts that vendors circulate are almost always generated at 23 °C for 24 hours, which is not how anyone runs a 72-hour headspace method.
PTFE: the reference standard for aggressive solvents
PTFE is the default for a reason. It is essentially inert to every organic solvent you will run through a purge-and-trap or headspace autosampler, including methylene chloride, toluene, hexane, and the chlorinated solvents that eat everything else. It handles the full pH range from 0 to 14 without measurable degradation. Its continuous service temperature is 260 °C, far above any VOA method ceiling.
The catch is mechanical, not chemical. A pure PTFE septum is stiff. It does not reseal well after a needle puncture, and on a 23-gauge autosampler needle you get 2 to 3 clean reseals before the puncture channel starts leaking. Under a 30 psi headspace pressurization cycle, that leak shows up as a low recovery on the first injection after the vial sits. The fix is not to abandon PTFE. It is to use PTFE as the sample-contact face of a laminate, not as the whole septum.
One more PTFE failure mode worth naming: cold flow. Under a crimp cap torqued to 1.5 N·m and stored for two weeks, a pure PTFE disc will slowly deform into the neck opening. When you pull the cap you find a dimple on the underside. That dimple is a stress point that will crack on the next puncture. If you are storing vials more than 48 hours before analysis, use a laminate, not a solid PTFE disc.
Silicone: excellent reseal, poor solvent resistance
Silicone has the best reseal memory of the four classes. A 2 mm silicone septum will take 15 to 20 punctures at 23-gauge and still hold pressure. That is why it appears on almost every general-purpose cap in the catalog. It is also why it fails on aromatic and chlorinated solvents.
Silicone swells in toluene, xylene, and methylene chloride. The swelling is not cosmetic. A swollen septum pushes the sample-contact face of the laminate into the vial neck, and on a 20 mm crimp finish that can lift the PTFE film off the glass rim by 0.1 to 0.2 mm. That gap is where your analyte escapes during the 40 °C equilibration. I have measured this: a silicone-backed septum in a 1,000 mg/L toluene standard lost 12% of headspace concentration in 90 minutes compared to a PTFE/silicone laminate with the PTFE face down. The difference was the swell, not the permeability.
Silicone is fine for aqueous samples, alcohols, and low-aromatic matrices at or below 60 °C. It is the wrong choice for anything in EPA SW-846 Method 5030 or 5035 with a chlorinated solvent list. The SW-846 methods do not name a septum material, but the recovery criteria in 8260 will fail you if the septum is the wrong one.
EPDM: the pH and aqueous workhorse, not a solvent septum
EPDM is a terpolymer that shows up in VOA vials less often than PTFE or silicone, but it has a specific niche: aqueous samples at moderate pH with no chlorinated solvents. It handles pH 1 to 12 and continuous service to 120 °C. It is also cheaper than a PTFE laminate, which is why some distributors push it.
The failure mode is solvent. EPDM swells badly in aliphatic hydrocarbons and aromatics. A hexane extract on an EPDM septum will show 30 to 40% swell within 4 hours, and the septum will extrude out from under the cap. If you see a septum that looks like it has been pushed sideways on a crimped vial, that is EPDM in the wrong method. EPDM belongs on aqueous metals samples, not on a solvent extraction.
Laminate: the working default, and where the adhesive bites you
The laminate is what most labs should be buying for VOA work. A typical construction is a 0.010 in PTFE face bonded to a 0.060 in silicone backing, with the PTFE face toward the sample. The PTFE handles the solvent and the silicone handles the reseal. It is the best compromise available for a 23-gauge autosampler running a multi-injection sequence.
The failure mode is the adhesive. The bond between the PTFE film and the silicone is a thin layer of proprietary chemistry, and it is the first thing to fail at elevated temperature or in prolonged solvent contact. That is the March return I opened with. The symptom is not a visible defect. It is a slow rise in blank response on the chlorinated solvent channel over the life of the lot, typically 2 to 5 µg/L per week on a 8260 blank, until the blank exceeds the reporting limit. The measurement that identifies it is a septum blank run: puncture a fresh septum with clean reagent water, incubate at method temperature for the full cycle, and inject. If the adhesive is degrading, the septum blank will show the ghost.
The fix is to specify a laminate with a no-adhesive or co-cured bond, or to move to a pure PTFE-faced silicone where the PTFE is mechanically captured rather than glued. Ask your supplier for the bond type in writing. If they cannot tell you, they do not know.
Matching material to solvent class: a working matrix
The table below is the one I keep taped inside the cabinet. It is built from the solvent classes that actually appear in VOA methods, not from a generic chemical resistance chart. Temperature is the method ceiling, not the polymer ceiling, because the polymer ceiling is almost never the limiting factor.
| Solvent class / matrix | Recommended material | Method temp ceiling | pH window | Failure if wrong |
|---|---|---|---|---|
| Aqueous, no organics (metals, anions) | EPDM or PTFE/silicone laminate | 90 °C | 1 to 12 | EPDM is fine; silicone will leach siloxanes into IC |
| Alcohols, ketones (low aromatic) | PTFE/silicone laminate | 60 °C | 2 to 12 | Silicone alone swells 8 to 12% in acetone |
| Aromatics (BTEX, PAHs) | PTFE-faced laminate, PTFE face down | 80 °C | 0 to 14 | Silicone alone loses 10 to 15% headspace in 90 min |
| Chlorinated solvents (DCM, TCE, PCE) | Pure PTFE or co-cured PTFE laminate | 60 °C | 0 to 14 | Adhesive-backed laminate ghosts 2 to 5 µg/L/week |
| Acidified preservative (pH < 2, HCl) | PTFE laminate | 60 °C | 0 to 14 | EPDM degrades at pH < 1 over 14 days |
| Caustic preservative (pH > 12, NaOH) | PTFE laminate | 60 °C | 0 to 14 | Silicone hydrolyzes at pH > 12 over 7 days |
| High-temp headspace (> 100 °C) | Pure PTFE with mechanical capture | 150 °C | 0 to 14 | Silicone backing hardens and cracks at 120 °C |
Two things to note about that table. First, the temperature ceiling is the method ceiling, not the polymer rating. A PTFE septum rated to 260 °C is still the wrong choice at 150 °C if the crimp cap cannot hold torque at that temperature, and most aluminum caps relax above 120 °C. Second, the pH windows are for continuous contact over the method cycle, not for a single injection. A silicone septum that sees pH 13 for 11 minutes is fine. The same septum in a pH 13 preserved sample stored for 14 days before analysis is not.
Temperature windows: where each polymer actually stops working
Temperature is the variable that gets the least attention and causes the most silent failures. The published ratings are for continuous service in air, not for a septum under compression against a hot sample. Here is what the ratings mean in practice.
PTFE above 200 °C
Pure PTFE is rated to 260 °C continuous and will not degrade chemically below that. But a VOA vial septum is a thin disc under compression, and above 200 °C the PTFE begins to cold-flow faster than the crimp can compensate. You get a slow leak at the crimp interface, not at the puncture. If you are running a high-temperature headspace method (some polymer additive methods go to 180 °C), specify a PTFE disc with a mechanical capture ring rather than a crimped cap. The crimp is the weak link, not the polymer.
Silicone above 120 °C
Silicone hardens above 120 °C. The hardening is permanent: once a silicone septum has seen 130 °C for a full method cycle, its reseal memory drops from 15 punctures to about 4. If your method runs at 140 °C and you are reusing vials, you are reusing hardened septa. The symptom is a needle that punches through cleanly on injection 1 and tears the septum on injection 3, leaving a fragment in the vial. That fragment is the debris your autosampler finds before you do.
EPDM above 120 °C
EPDM is rated to 120 °C continuous, but the practical ceiling for a compressed septum is closer to 100 °C. Above that, the compression set becomes permanent and the septum will not reseal after the first puncture. EPDM is not a high-temperature material, and no amount of vendor marketing changes the compression set curve.
Field case 1: the ghost that came from the adhesive
A mid-size environmental lab running EPA 8260 on a purge-and-trap concentrator started seeing a methylene chloride response on every method blank at 3.8 to 4.5 µg/L against a reporting limit of 5 µg/L. The blanks were clean when the lot was opened and drifted up over three weeks. The lab had already replaced the trap, baked the lines, and changed the purge vessel. None of it helped.
The measurement that found it was a septum blank. The lab punctured a fresh septum from the same lot with clean reagent water, capped it, and ran it through the full 11-minute purge at 40 °C. The septum blank came back at 4.1 µg/L methylene chloride. The adhesive layer between the PTFE and silicone was the source. The fix was a switch to a co-cured PTFE laminate with no adhesive layer, and the blanks dropped to <0.5 µg/L within one lot change. The lab also added a septum blank to every new lot acceptance, which is the practice I recommend to every VOA lab I work with.
Field case 2: the crimp that relaxed at 130 °C
A polymer additives lab was running a headspace method at 130 °C for 25 minutes on a PTFE/silicone laminate with an aluminum crimp cap. Recovery on the internal standard dropped from 98% to 71% over two weeks, and the drop was worse on the later injections in each sequence. The autosampler log showed no needle errors. The leak was not at the puncture.
The measurement that found it was a torque check. The lab pulled ten crimped vials from the autosampler tray after the run and measured cap torque with a calibrated torque tester. Vials that had been through the 130 °C cycle read 0.6 to 0.8 N·m against a target of 1.5 N·m. The aluminum crimp had relaxed at temperature, and the septum was no longer under enough compression to seal. The fix was a switch to a magnetic cap with a PTFE-faced septum rated to 150 °C, which holds compression without a crimp, plus a torque check on every tenth vial at the start of each sequence. Recovery returned to 97 to 99%.
Selection checklist: what to specify before you order
When you write the spec for a VOA vial order, the material line should carry five pieces of information. If any one is missing, the supplier will fill it with whatever is cheapest.
- Face material and orientation. PTFE face toward sample, not silicone. State it explicitly. A surprising number of “PTFE-lined” caps ship with the silicone face down.
- Bond type. Co-cured or mechanically captured, not adhesive. Ask for the bond chemistry in writing.
- Thickness. 0.010 in PTFE face over 0.060 in silicone backing is the working default for 23-gauge autosamplers. Thinner PTFE cold-flows faster; thicker silicone reduces puncture count.
- Temperature rating at compression. Not the polymer rating. Ask for the compression-set rating at your method temperature, which is usually 20 to 40 °C below the published continuous rating.
- Lot-level septum blank. Require a septum blank certificate with each lot, run at your method temperature and cycle time. This is the only way to catch an adhesive change before it reaches your blanks.
The VOA vial itself is a glass and closure system, and the closure is where the material decision lives. If you are sourcing the closure separately, the septa and vial septa catalogs are where the material grade is actually specified. For method-specific guidance on which material a given instrument vendor recommends, the Agilent GC accessories and Thermo Fisher sample handling documentation are worth reading against your own method, and Restek’s technical resources publish septum blank data that is more honest than most vendor charts. If you also run HPLC on the same samples, the HPLC vial material guidance follows the same logic but with a different temperature ceiling.
The decision rule
Pick the material from the solvent class first, then check the temperature against the compression-set rating, not the polymer rating. Chlorinated solvents and anything stored more than 48 hours get a co-cured PTFE laminate with the PTFE face down. Aromatics get the same laminate at or below 80 °C. Aqueous samples with no organics can use EPDM if cost matters and the method stays below 100 °C. Silicone alone belongs only on short-cycle, low-aromatic methods where reseal count matters more than solvent resistance. And every lot gets a septum blank at method temperature before it goes on the instrument, because the adhesive layer is the failure you cannot see and the blank is the only measurement that catches it.
References & further reading
Related resources from our lab consumables network:
External references and standards: