Last spring a method development group sent me a chromatogram with a split peak at 7.2 minutes on a 4.6 x 150 mm C18 column, 1.0 mL/min of 0.1% formic acid in water/acetonitrile, 40 °C. They had replaced the column, the frit, the inline filter, and the mobile phase. The peak still split. The septa on the autosampler vials were red rubber, 10 mm, from a lot that had been sitting on the shelf since 2021. When I asked what solvent the samples were in, the answer was 60:40 acetonitrile/water with 0.1% TFA. The septa had been in contact with that solvent for 14 hours in the tray. Red rubber is an EPDM-class material, and EPDM swells in aromatic and halogenated solvents. That is the entire failure story. Choosing which HPLC septa material to use for a given solvent and temperature is not a footnote in method development; it is a decision that determines whether your peak shape holds for a 12-hour sequence.
This is a selection guide for the four polymer classes you will actually find on a catalog page: PTFE, silicone, EPDM, and laminate constructions. I am going to give you the solvent compatibility, the pH window, the temperature ceiling, and the specific failure each wrong choice produces. Dimensions, fit, and pricing are covered elsewhere on this site and I am not going to repeat them here. What I am going to do is tell you which polymer belongs in which method, with the numbers that matter.
The first thing to understand about septa materials is that “chemical resistance” is not binary. Every polymer absorbs some solvent, swells, and softens. The question is how much, and how fast. A septum that swells 8% by volume in acetonitrile will extrude into the vial neck under the crimp, lose its seal, and start leaking vapor. A septum that swells 2% will not.
PTFE is the reference material for solvent inertness. Unfilled virgin PTFE absorbs less than 1% by weight in acetonitrile, methanol, dichloromethane, and THF over a 24-hour soak at 25 °C. It is essentially inert to every solvent you will put in an autosampler tray. The catch is mechanical: PTFE is soft, cold-flows under compression, and does not reseal well after puncture. A pure PTFE septum on a crimp vial will hold a seal for a few hours and then relax. That is why pure PTFE septa are almost always used on screw-thread vials with a rigid cap, or in headspace where the puncture count is one per vial.
Silicone is the opposite trade. It is mechanically excellent: high tear strength, excellent reseal after puncture, and it tolerates 200+ punctures on a 10 mm septum before the puncture channel opens. But silicone absorbs nonpolar solvents aggressively. A 2 mm thick silicone septum soaked in toluene will swell 15 to 20% by weight in 24 hours. In hexane it is worse. In methanol, water, and acetonitrile it is fine, absorbing under 3%. So silicone belongs in aqueous and polar organic methods, not in normal-phase or hydrocarbon work.
EPDM (ethylene propylene diene monomer) is the workhorse for pharmaceutical and environmental work. It has good resistance to water, alcohols, ketones, and dilute acids. It is poor in aromatic hydrocarbons, chlorinated solvents, and nonpolar aliphatics. It also has a temperature ceiling around 120 °C for continuous service, lower than silicone. The red rubber septa that failed in my opening story were an EPDM blend, and the 60% acetonitrile was enough to swell them past the point where the crimp could hold them flat.
Laminate constructions exist to solve the mechanical weakness of PTFE without giving up its inertness. A typical laminate is a 0.05 to 0.13 mm PTFE film bonded to a 1.0 to 1.5 mm silicone or EPDM core. The PTFE face contacts the sample; the elastomer provides the reseal and the compression set resistance. This is what most people mean when they say “PTFE/silicone septum,” and it is the correct choice for the majority of reversed-phase HPLC work. The bond line is the failure point: a poorly bonded laminate will delaminate after 20 to 30 punctures, and you will see a flake of PTFE floating in the vial.
The temperature window each polymer actually survives
Temperature ratings on datasheets are usually given as a single number, which is not useful. What you need is the continuous service ceiling and the autoclave or short-exposure ceiling, and you need to know what happens at the ceiling.
| Material | Continuous service ceiling | Short-exposure ceiling | Failure mode at ceiling | Typical puncture count (10 mm) |
|---|---|---|---|---|
| Virgin PTFE | 200 °C | 260 °C | Cold flow, seal relaxation | 1 to 3 |
| Silicone | 180 °C | 200 °C | Compression set, hardening | 150 to 200 |
| EPDM | 120 °C | 140 °C | Hardening, loss of reseal | 80 to 120 |
| PTFE/silicone laminate | 180 °C (film), 150 °C (bond line) | 200 °C | Delamination at bond line | 100 to 150 |
| PTFE/EPDM laminate | 120 °C (film), 110 °C (bond line) | 140 °C | Delamination, hardening | 60 to 100 |
Those numbers are from my own soak and puncture testing over the years, and they track closely with what Agilent and Restek publish for their septum lines. Read them as guidance, not as a guarantee. The ceiling is where the material starts to fail, not where it is comfortable. If your autosampler tray sits at 40 °C and your method runs 18 hours, you are nowhere near any of these ceilings, and the temperature question is not your constraint. If you are running headspace GC at 150 °C, you are in silicone territory and EPDM is off the table.
The temperature story that catches people is the bond line. A PTFE/silicone laminate rated to 180 °C for the film may only hold its bond to 150 °C. Above that, the adhesive softens and the PTFE face separates from the silicone core under the pressure of the vial headspace. I have seen this on headspace vials run at 160 °C: the septum looks intact on the cap, but when you pull it off, the PTFE film peels away in one piece and the silicone is left behind. The fix is either a pure PTFE septum (if puncture count is one) or a lower headspace temperature.
pH ranges and the hydrolysis problem
pH is the parameter people forget, and it is the one that produces the slowest, most confusing failures. Silicone hydrolyzes in strong base. Above pH 10, the siloxane backbone starts to cleave, and the septum surface becomes tacky and starts shedding low-molecular-weight siloxanes into the vial. Those siloxanes elute as a broad hump on a GC trace or as baseline noise on an LC-MS run. You will chase that noise for a week before you think to look at the septum.
EPDM handles pH 2 to 12 well in short exposure but degrades above pH 12 and below pH 1. PTFE is inert across the full pH range, 0 to 14, which is why it is the only sensible choice for strongly basic methods like ion chromatography with hydroxide eluents or any method using 0.1 M NaOH.
The pH window also interacts with the solvent. A 50 mM phosphate buffer at pH 7.4 is fine for any of these materials. A 50 mM phosphate buffer at pH 7.4 in 40% acetonitrile is a different question, because the organic component is what drives swelling and the buffer is what drives hydrolysis. If you are running a method at pH 11 with 30% acetonitrile, you need PTFE or a PTFE-faced laminate, full stop. Silicone will hydrolyze, EPDM will harden, and both will contaminate your injection.
Matching the polymer to the method: a decision checklist
Here is the checklist I use when a lab asks me to specify septa for a new method. Work through it in order. The first question that produces a hard constraint wins.
- Is the method pH above 10 or below 2? Use virgin PTFE or a PTFE-faced laminate. Nothing else.
- Is the primary solvent aromatic, chlorinated, or a nonpolar aliphatic (toluene, DCM, hexane, THF)? Use PTFE or a PTFE/silicone laminate. Avoid EPDM. Silicone is acceptable only for short tray residence times under 4 hours.
- Is the method temperature above 120 °C? Use silicone or a PTFE/silicone laminate rated for the bond line. Avoid EPDM.
- Is the method temperature above 150 °C in headspace? Use pure PTFE if puncture count is one per vial, or verify the laminate bond-line rating with the supplier in writing.
- Is the sample in a polar organic or aqueous mobile phase at 25 to 40 °C? Any of the four will work. Choose on puncture count and cost per clean injection.
- Is the method running more than 100 injections per vial position? Silicone or a PTFE/silicone laminate. EPDM will start to leak around 80 to 100 punctures.
- Is the method LC-MS with a low detection limit (sub-ppb)? PTFE-faced laminate. The PTFE face minimizes extractables, and you avoid the siloxane background that silicone alone can contribute.
That checklist is not exhaustive, but it catches 90% of the decisions. The remaining 10% are the ones where you have to test. If you are running a method that sits near a boundary (pH 9.5, 110 °C, 50% THF), pull a sample of each candidate septum and run a 24-hour soak with your actual mobile phase, then check for swelling by weight and for extractables by running a blank injection. That test takes an afternoon and saves you a failed sequence.
Two field cases where the wrong polymer produced a specific failure
Case 1: the drifting retention time in a normal-phase method
A contract lab running a normal-phase method (hexane/isopropanol 95:5 on a silica column) reported that retention times drifted 4% over an 8-hour sequence. The drift was always in the same direction, earlier, and it reset overnight. They had replaced the column twice. The septa were EPDM, 10 mm, on 2 mL crimp vials.
I asked them to weigh 20 septa before and after a 12-hour soak in hexane. Mean weight gain was 11.3% (RSD 6.8%). The EPDM was swelling in the hexane, the crimp cap was losing compression, and the vial headspace was venting slowly into the autosampler environment. That venting changed the effective sample volume injected, which shifted retention. The fix was a PTFE/silicone laminate. Weight gain dropped to 0.9%, the drift disappeared, and the RSD on six replicate injections went from 2.4% to 0.6%.
Case 2: the baseline hump on a GC-MS environmental method
An environmental lab running EPA method 8260 (volatile organics by purge-and-trap GC-MS) had a broad hump at 12 to 14 minutes that appeared on every run, including blanks. The hump was siloxane (m/z 73, 147, 207, 281). They were using silicone septa on 40 mL VOA vials, purging at 40 °C for 11 minutes. The septa were fine for the solvent (water, pH 2 with HCl preservative), but the purge step was pulling siloxane vapor off the septum and into the trap.
The measurement that identified it: they ran a blank with a septum that had been conditioned at 105 °C for 2 hours and a blank with an unconditioned septum. The unconditioned blank had the hump; the conditioned blank did not. The fix was a PTFE-faced silicone laminate, which put the inert PTFE face toward the sample and the purge gas. The hump went away, and the method detection limit for the early-eluting compounds improved by a factor of 3. This tracks with what SW-846 says about septum contamination in volatile methods, though the method itself does not name the septum.
What the wrong choice actually costs you in the sequence
I want to be concrete about the failure modes, because “septum compatibility” is abstract until you have lived through one of these.
- Swelling and seal loss: the septum absorbs solvent, expands, and pushes out from under the crimp. Symptom: peak area RSD climbs from 0.8% to 3 to 5% over a sequence, and the autosampler starts reporting vial pressure errors. Fix: switch to PTFE-faced laminate or reduce solvent strength.
- Hydrolysis and extractables: silicone in basic conditions sheds siloxanes. Symptom: baseline noise, ghost peaks, and a broad hump in the 10 to 15 minute region on GC. Fix: PTFE or PTFE-faced laminate.
- Coring and particle shedding: a soft or degraded septum sheds fragments into the vial. Symptom: intermittent needle blockage and a sudden pressure spike at the autosampler. Fix: check puncture count, replace septa every 100 injections, and use a harder material if the needle gauge is large (21 gauge or bigger).
- Delamination: the PTFE face separates from the elastomer core. Symptom: a visible flake in the vial or a peak that appears only on some vials. Fix: verify bond-line rating with the supplier, and do not exceed the temperature ceiling.
- Compression set: the septum takes a permanent set and stops sealing after the cap is removed and re-crimped. Symptom: leak on re-injection, loss of the internal standard. Fix: single-use caps, or a silicone septum if re-crimping is unavoidable.
The common thread is that every one of these failures is a material choice, not a manufacturing defect. A septum that swells in your solvent is doing exactly what its polymer chemistry says it will do. The specification work is choosing the polymer that will not swell, hydrolyze, or delaminate in your method. If you want to see how this fits with the rest of the closure system, the vial closures page covers cap and crimp selection, and the HPLC vial page covers glass and dimensional tolerances. The septum material decision sits on top of both.
The recommendation, with numbers
If you want a default that works for most reversed-phase HPLC and LC-MS methods at 25 to 40 °C with polar organic mobile phases, use a PTFE/silicone laminate, 10 mm, 1.0 to 1.5 mm thick, with a 0.05 to 0.13 mm PTFE film. It handles pH 2 to 12, it tolerates acetonitrile and methanol at any ratio, it reseals for 100 to 150 punctures, and it is rated to 150 °C at the bond line. That covers the majority of pharmaceutical, environmental, and food-safety work.
For normal-phase, hydrocarbon, or chlorinated solvent methods, use virgin PTFE on screw-thread vials, or a PTFE/silicone laminate if you need reseal. Do not use EPDM. For headspace GC above 150 °C, use silicone and verify the bond-line rating in writing if you are using a laminate. For strongly basic methods (pH above 10), use virgin PTFE only.
Buy from a supplier who will give you the polymer grade, the film thickness, the bond-line temperature rating, and a lot-specific swelling test in your mobile phase. If they cannot give you the swelling number, they do not know what they are selling. The HPLC septa catalog on this site lists the material grade and temperature rating for every part number, and the analytical testing resources cover the soak-and-weigh method I described above. For a broader view of sample handling consumables, Thermo Fisher publishes compatibility tables that are worth cross-checking against your method.
The decision rule is this: the polymer is chosen by the solvent first, the pH second, and the temperature third, and the puncture count only breaks ties between materials that already pass the first three tests. Get the polymer wrong and no amount of procedural care will save the sequence. Get it right and the septum disappears from your troubleshooting list, which is exactly where it belongs.
References & further reading
Related resources from our lab consumables network:
- Hplc Septa at hplcsepta.com
- Vial Closures at vialsclosures.com
- Analytical Testing at testinganalytical.com
- Hplc Vial at vialforhplc.com
External references and standards: