A 60 m x 0.25 mm x 0.25 µm DB-5 run on a soil extract came back with 14% area RSD on the internal standard across six injections, and the culprit was a 1 mm silicone-only septum in a crimp cap that had been on the vial for eleven hours before the autosampler got to it. The solvent was 1:1 dichloromethane:hexane. The septum had swollen, the injection port needle punched through a soft spot, and the resulting leak pulled air into the liner on every injection. That is a material selection failure, not an instrument problem. Choosing the right polymer for a given solvent and temperature is the single decision that most often gets made by habit rather than by data, and this guide is about making it deliberately.

Which gc vial material to use for a given solvent and temperature starts with the septum, not the glass

The glass body is almost always borosilicate 33-expansion, and for most methods it is not the variable. The variable is the closure: the septum polymer and the laminate that holds it. When people ask which gc vial material to use for a given solvent and temperature, they are really asking three questions at once. Does the polymer swell or dissolve in the solvent? Does it hold its shape and reseal at the autosampler temperature? And does it shed extractables that land on the column? Get any one wrong and you get a specific, diagnosable failure. Get all three right and the septum disappears from your troubleshooting list, which is where it belongs.

I run three polymer families through most GC work: PTFE (usually as a PTFE-faced laminate), silicone (the soft, high-puncture-count workhorse), and EPDM (ethylene propylene diene monomer, the chemically resistant rubber). Each has a solvent window, a pH window, and a temperature ceiling. The rest of this article is about those windows and the failure each wrong choice produces.

The four septum constructions you will actually be offered

Suppliers rarely sell you a bare polymer. They sell a laminate, and the laminate is what contacts your sample on one side and the autosampler needle on the other. The four you will see on a quote are:

• Silicone only. Softest, best reseal, highest puncture count (often 5 to 10 injections before coring risk climbs). Poor solvent resistance to chlorinated and aromatic solvents. Not recommended for long dwell times on the autosampler.
• PTFE-faced silicone. A thin PTFE film (typically 0.05 to 0.13 mm) bonded to a silicone core. The PTFE face gives solvent and pH resistance on the sample side; the silicone gives the reseal. This is the default for most GC work.
• PTFE/silicone/PTFE sandwich. PTFE on both faces. Better for headspace and for methods where the septum sees vapor on both sides. Slightly stiffer, so needle penetration force goes up.
• EPDM or EPDM-faced laminates. Rubber with good resistance to water, steam, ketones, and alcohols, and a higher temperature ceiling than silicone in some grades. Poor against aromatic and chlorinated solvents. Used where silicone extractables are a problem and the solvent is aqueous or alcoholic.

The gc vial you buy is a system: glass, septum, and cap. The cap material (usually aluminum for crimp, plastic for screw) matters for torque and re-use, but the septum is where the chemistry happens.

Solvent compatibility: the swelling table that decides your method

Swelling is the first failure mode. A silicone septum in dichloromethane can take up 20 to 40% of its weight in solvent within a few hours, and a swollen septum no longer seats flat in the cap. The result is a slow leak that shows up as drifting retention times and a falling internal standard response, usually on the later injections in a run. Here is how the three families behave against the solvents I see most often in GC.

Read that table as a starting point, not a guarantee. The supplier’s own compatibility data should be on the datasheet, and if it is not, ask. A datasheet without a test method behind it is a marketing brochure, and that applies to septa as much as to anything else. The Restek technical resources library has solvent-resistance charts that are worth bookmarking if you run mixed solvent classes.

pH windows: where each polymer starts to hydrolyse

pH failure is slower and nastier than solvent failure. Nothing swells visibly. The septum just starts to break down at the surface, and the breakdown products are siloxanes or rubber fragments that show up as a rising baseline or a cluster of siloxane peaks (m/z 73, 147, 207, 281 in MS) late in the run. The pH windows I work to:

• Silicone: roughly pH 3 to 9. Outside that, hydrolysis accelerates and the septum loses elasticity within days.
• PTFE-faced silicone: the PTFE face holds to about pH 12 on the sample side. The silicone core is still pH-limited, so a scratch through the face is a failure point.
• EPDM: roughly pH 2 to 13. Better base resistance than silicone, but only if the solvent is aqueous or alcoholic.

The classic mistake is a buffered aqueous extract at pH 10 in a plain silicone septum, run on a 24-hour autosampler queue. By injection 40 the septum has softened, the needle starts to core it, and you get a punched hole rather than a reseal. A cored septum is a punched hole, not a torn one, and once it is cored the leak is permanent. Switching to a PTFE-faced laminate fixes it. The HPLC autosampler vial side of the house has the same problem with aqueous mobile phases, and the same fix.

Temperature windows: the autosampler is usually hotter than you think

Silicone’s continuous service ceiling is around 200 °C for short excursions, but the reseal performance drops well before that. Above about 80 °C the compression set starts to matter: the septum takes a set, does not spring back, and the next needle punch is easier. PTFE is good to about 260 °C continuous, which is why it is the face material for high-temperature methods. EPDM sits around 150 °C continuous in the grades used for septa. But the temperature that actually matters is the autosampler tray temperature, and that is often 4 to 10 °C, not 200 °C. Cold is a different failure mode: a cold septum is stiff, penetration force goes up, and a soft silicone septum can tear on the way in. If your tray runs at 4 °C, a PTFE-faced silicone is the right compromise. If you are running a hot headspace method at 120 °C, you want PTFE on both faces.

For reference on how the major instrument makers specify septum temperature and puncture counts, the Agilent chromatography consumables catalogue and the Thermo Fisher sample handling pages both publish septum specifications by instrument family. Worth a look before you standardise on one part number across a lab.

The failure each wrong choice produces: a symptom-to-polymer map

This is the part I wish every new analyst got on day one. The symptom tells you which polymer is wrong. You do not need to guess.

• Rising baseline late in the run, siloxane peaks in MS. Silicone septum breaking down under solvent or pH attack. Fix: PTFE face, or EPDM if the solvent is aqueous.
• Drifting retention times, falling ISTD response on later injections. Swollen septum not reseating. Fix: PTFE-faced laminate, and shorten the dwell time on the autosampler.
• Visible hole in the septum after 3 to 5 injections. Septum too soft for the needle gauge or tray temperature. Fix: harder laminate, or a larger needle gauge.
• Ghost peaks at m/z 73, 147, 207, 281. Silicone bleed. Fix: PTFE face, or a lower-bleed septum grade.
• Needle bending or coring at 4 °C tray temperature. Stiff septum. Fix: silicone or a softer PTFE-faced grade, and check needle alignment.
• Sample loss over a 24-hour queue. Permeation through a thin silicone septum. Fix: PTFE face on the sample side, and a thicker laminate.

Two field cases make this concrete. The first: a contract lab running EPA method 8260 on a 24-hour queue kept seeing the internal standard drop 8 to 12% by injection 60. They had tried new septa, new caps, new needles. The measurement that identified it was a simple weigh-back: septa pulled from the autosampler after 12 hours had gained 18% mass in a DCM matrix. The fix was a PTFE-faced silicone laminate and a 12-hour maximum dwell. The RSD went from 14% to under 4%. The U.S. EPA SW-846 test methods for waste and materials do not prescribe a septum polymer, but they do prescribe holding times, and the septum has to survive the holding time.

The second: a pharmaceutical QC lab running a residual solvents method in DMF at 80 °C autosampler temperature saw ghost peaks at m/z 207 and 281 on every chromatogram. They had standardised on a plain silicone septum because it was the cheapest part number. The measurement was an injection-port blank with a new septum: the ghost peaks were still there, so it was not the inlet. Swapping to a PTFE/silicone/PTFE sandwich eliminated the peaks within one run. The lesson: at 80 °C, silicone bleed is a real source of interference, and the sandwich is the fix.

What the laminate actually does at the interface

It helps to think about the septum as a three-layer system even when it is sold as one part. The sample-side face is the chemical barrier. The core is the mechanical spring that reseals after the needle withdraws. The cap-side face is the compression interface that seals against the cap. Each layer has its own job, and a failure in any one shows up differently.

If the sample-side face is PTFE, the solvent sees PTFE and the core never touches the sample. That is why a PTFE-faced silicone survives DCM where plain silicone does not. If the face is scratched (a bent needle will do it), the solvent reaches the silicone core and you are back to swelling. If the core is too soft, the septum takes a compression set and stops resealing. If the cap-side face is too thin, the septum extrudes into the cap opening under crimp force and the seal is compromised. This is why I do not buy septa on price alone. The sample vial price is a number, but the cost of a failed run is a different number entirely.

Torque, crimp force, and why the septum grade changes the cap spec

A septum is only as good as the seal it sits in. For a 9 mm crimp cap, the crimp force has to compress the septum enough to seal but not so much that it extrudes. I work to a crimp that leaves the cap skirt at about 0.1 to 0.2 mm below the vial shoulder, checked with a crimp gauge, and I re-check the gauge every 500 caps. A soft silicone septum needs less crimp force than a stiff PTFE sandwich. If you change septum grade without re-checking the crimper, you will get leaks that look like a septum problem but are actually a crimp problem. For screw caps, the torque spec is usually 1.0 to 1.5 N·m for a plastic cap on a 9 mm thread, and the cap should not be re-torqued after the first seal.

The GPI thread standard (GPI 9 mm, GPI 13-425, GPI 15-425, GPI 18-400, GPI 24-400) defines the thread, not the septum. But the septum grade determines how much torque the seal will tolerate before it extrudes. A PTFE-faced laminate tolerates more torque than a plain silicone. A reagent bottle with the same GPI thread has the same trade-off, which is why the closure spec and the septum spec have to be written together.

A decision checklist you can tape to the autosampler

Here is the rule set I use when specifying a septum for a new method. It takes about five minutes and saves weeks.

• Solvent is chlorinated or aromatic? PTFE-faced silicone. No exceptions.
• Solvent is aqueous or alcoholic and pH is between 3 and 9? Silicone is fine if dwell time is under 8 hours.
• pH above 9 or below 3? PTFE face on the sample side, or EPDM if the solvent is aqueous.
• Autosampler tray above 60 °C? PTFE on both faces.
• Autosampler tray at 4 °C? Silicone or a soft PTFE-faced grade; check needle alignment.
• Queue longer than 12 hours? PTFE face, and cap the dwell time in the method.
• MS detection with siloxane interference? PTFE face, and run an injection-port blank with a new septum to confirm.
• Puncture count expected above 5? Silicone core, and replace on a schedule, not on failure.

If you cannot answer one of these questions, the method is not fully specified. That is a finding, not a nuisance.

What to put on the purchase order

Write the septum grade, the face material, the core material, the thickness, and the temperature rating on the PO. Do not write “9 mm septum, silicone” and expect the right part. A good PO line reads something like: 9 mm crimp, PTFE-faced silicone, 1.0 mm thickness, rated to 200 °C, 5-puncture minimum. That is a spec a supplier can be held to. It also gives you something to compare when a lot drifts, which it will. When a lot drifts, the first thing I check is the septum thickness with a micrometer (a 0.05 mm variation is enough to change the crimp seal) and the second is a weigh-back swell test in the actual method solvent. Both take under an hour and both tell you whether the problem is the part or the method.

The decision rule is simple. Match the septum face to the solvent, match the core to the temperature and puncture count, and match the crimp or torque to the laminate stiffness. If any one of those three is out of step, the failure will show up as a chromatographic problem, and you will spend a week looking at the inlet before you look at the cap. Start at the cap.

References & further reading

Related resources from our lab consumables network:

• Gc Vial at gcvialfactory.com
• Reagent Bottle at reagentbottleforsale.com
• Hplc Autosampler Vial at hplcautosamplervials.com
• Sample Vial at samplevialprice.com

External references and standards:

• Agilent chromatography consumables and columns
• Thermo Fisher chromatography and sample handling
• U.S. EPA SW-846 test methods for waste and materials
• Restek technical resources for chromatography