Why septa fail: the four mechanisms that matter

Last spring a QC analyst called me because her GC sequence had failed three runs in a row with a phantom peak at 8.2 minutes. We spent two hours chasing the inlet, the column, and the carrier gas purifier before I asked the obvious question: how old is the septum in that inlet? She checked the maintenance log. It had been punched 140 times over nine days because nobody had reset the counter after the last change. The phantom peak was a coring fragment sitting in the liner, slowly bleeding siloxanes every time the injector heated. That is the pattern I see over and over. Septa are the cheapest consumable in the lab and the most frequently misdiagnosed failure point. This article covers the failures I have actually seen in daily operation, what causes them, and the numbers that keep them from happening.

A septum is a small disc of elastomer, usually 11 mm or 12 mm in diameter and 1 mm thick for standard 2 mL vial closures, or 9.5 mm for 8 mm crimp and snap formats. It has one job: seal against a pressure differential while a needle passes through it hundreds of times. It fails through four distinct mechanisms, and each one has a different fix.

Mechanical coring and needle damage

When a needle punches through silicone, it cuts a small plug rather than stretching the material. That plug either falls into the vial (contaminating the sample) or stays lodged in the needle bore or the inlet liner (contaminating the system). The dominant variable is needle geometry. A beveled 23-gauge needle with a point style 2 (side-port) tip cores roughly 40% less than a blunt 22-gauge needle in the same silicone septa, based on our internal puncture testing at 20°C. Point style 3 (conical) is worse for septa because the tip concentrates force at a single point.

Needle diameter matters too. A 26-gauge needle (0.26 mm OD) creates a hole about 0.28 mm across; a 22-gauge (0.72 mm OD) creates one closer to 0.75 mm. Over 100 punctures, the 22-gauge hole has a much higher chance of failing to reseal under 10 psi inlet pressure. If you are running headspace or purge-and-trap at elevated pressure, use the smallest gauge your autosampler can reliably handle.

Thermal degradation and outgassing

Silicone rubber (polydimethylsiloxane) starts to degrade above its rated temperature. A standard silicone/PTFE septa rated to 200°C will outgas cyclic siloxanes (D3, D4, D5, D6) if the inlet sits at 250°C. Those siloxanes elute as a characteristic cluster of peaks at m/z 73, 147, 207, 281 in GC-MS, and they will show up in your blank runs. If you are running a 280°C inlet for high-boiling analytes, you need a high-temperature septa (typically a silicone with a molded-in PTFE facing, rated to 300°C or higher) or a septum purge flow of at least 3 mL/min to sweep the outgassing away from the column.

Chemical attack and swelling

Silicone swells in nonpolar solvents. Put a standard silicone septa in contact with toluene or hexane vapor for a week and it can grow 15 to 25% in volume, which pushes it out of the cap and breaks the seal. PTFE-faced septa resist this because the sample only contacts the PTFE face, but the PTFE is only 0.1 to 0.2 mm thick. If the needle scratches through the facing, the silicone underneath is exposed. For aggressive solvents (DCM, THF, DMSO at elevated temperature), specify a pure PTFE or PTFE/silicone laminate with a facing thickness of at least 0.13 mm.

Compression set and loss of reseal force

Every time you crimp a cap, you compress the septa by roughly 20 to 30%. Over time the elastomer takes a permanent set and stops pushing back against the needle hole. A septa that resealed perfectly on day one may leak at 5 psi on day thirty. This is why septa life is measured in punctures and thermal cycles, not calendar days.

A field guide to the failures I see most often

Failure 1: the phantom peak that is not in the method

Symptom: an extra peak in the chromatogram that shows up in blanks, moves slightly with inlet temperature, and disappears after you change the septum. Cause: a coring fragment in the liner or on top of the column. The fragment is silicone, so it pyrolyzes in the hot inlet and releases siloxane fragments that elute as peaks.

Fix: change the septum, then change the liner. A coring fragment will sit in the liner for weeks if you only change the septum. Inspect the liner under a loupe; a 0.5 mm translucent chip is easy to spot against the glass wool. If you find one, also check the top 2 cm of the column by cutting and re-installing, because fragments can migrate onto the stationary phase. Prevention is a septum change interval based on puncture count, not calendar time, plus a needle with a side-port tip.

Failure 2: the leaking vial that passes the pressure test but fails the run

Symptom: intermittent low response, especially in headspace or purge-and-trap, with no obvious chromatographic cause. Cause: a septa that has lost compression set and no longer seals against the cap. The classic tell is a vial that holds pressure for 30 seconds but leaks over five minutes.

Fix: check the crimp. A properly crimped 11 mm aluminum cap on a 20 mm crimp vial should be set to a crimping force that produces a cap skirt diameter of 19.7 to 20.0 mm and a flat top. If your crimper is set too loose, the septa is not compressed enough to seal; too tight, and you cut the septa. I have seen labs run crimpers 0.3 mm out of spec for months because nobody owned a crimp gauge. Buy one. They cost less than a column and they catch this failure in 30 seconds.

Failure 3: the needle that will not pierce

Symptom: autosampler error, bent needle, or a vial that gets pushed out of the tray. Cause: a septa that has hardened, usually from thermal aging or from a batch of caps stored in a hot warehouse. Silicone that has been held at 60°C for six months can harden by 10 to 15 Shore A points, which is enough to deflect a 26-gauge needle.

Fix: store septa and pre-capped vials below 30°C and away from UV. Check the durometer on incoming lots if you buy in bulk; a Shore A reading above 55 for a standard septa is a red flag. If you are running an autosampler vial with a lightweight needle carriage, use a pre-slit septa. The slit reduces penetration force by roughly half and eliminates the deflection problem.

Failure 4: the sample that evaporates before it is injected

Symptom: increasing concentration over a long autosampler run, or a systematic positive bias on volatile analytes. Cause: a septa that is permeable to the sample vapor, or a vial that is not sealed. Silicone is highly permeable to small molecules (methanol, acetaldehyde, DCM). A standard silicone septa on a 2 mL vial held at 25°C for 24 hours can lose 2 to 5% of a volatile analyte through the septa, depending on the compound and the headspace volume.

Fix: use a PTFE-faced septa (the PTFE is essentially impermeable to most solvents), keep the sample waiting time under 12 hours, and cap the vial immediately after filling. If your method uses a long autosampler queue, validate the hold time with a stability study at your actual storage temperature. I have seen methods fail validation because the analyst did not account for 18 hours of queue time on a volatile standard.

Failure 5: the septa that reacts with the sample

Symptom: disappearing analyte, new peaks, or a drifting baseline. Cause: chemical attack. Silicone contains residual curing agents and low-molecular-weight oligomers that can leach into aqueous samples, and it can adsorb hydrophobic analytes. For trace-level work (sub-ppb), a standard silicone septa is often the largest source of contamination in the vial.

Fix: for trace analysis, use a septa with a low-extractable silicone or a pure PTFE facing, and run a vial blank with the same septa lot. For reactive analytes (aldehydes, some amines, sulfur compounds), test the septa against the sample matrix before committing to a method. I keep a small library of septa types and run each one against a reactive standard when a new method comes in.

Comparison: choosing the right septa for the application

The table below is what I hand to new analysts. The numbers are typical values from our own testing and from manufacturer data sheets; always confirm against the lot certificate for critical work.

If you are running a standard HPLC vial with a 2 mL fill and a 9 mm screw cap, the default choice is a 1.0 mm silicone/PTFE septa. If your inlet runs above 250°C, step up to the high-temp version. If your sample is DCM or THF, go to pure PTFE. There is no universal septa, and pretending otherwise is how you end up with the phantom peak.

Operational precautions that actually prevent failures

Set puncture intervals by count, not by calendar

The single biggest cause of septa failure in daily use is a change interval based on time. I have seen labs change septa every Monday regardless of use, and I have seen labs run 400 injections on one septa because the maintenance log was on a clipboard nobody read. Set the interval by puncture count and log it.

• Standard silicone/PTFE, 23G needle, 10 psi inlet: change at 50 to 100 punctures. For critical trace work, change at 50.
• High-temp silicone/PTFE, 300°C inlet: change at 40 to 80 punctures. The thermal load reduces reseal life.
• Pre-slit septa: change at 100 to 150 punctures, but inspect for slit elongation every 50.
• Pure PTFE: change at 20 to 40 punctures. PTFE has poor elastic recovery and does not reseal well.

If your autosampler does not count punctures, use the sequence count from your chromatography data system and reset it at every septum change. A simple maintenance log with date, inlet, puncture count, and lot number takes 30 seconds and saves hours of troubleshooting.

Match the needle to the septa

Use a 23-gauge needle with a side-port (point style 2) tip for routine work. Avoid conical (point style 3) tips with silicone septa. If you must use a 26-gauge needle for low-volume injections, expect fewer punctures before failure and plan accordingly. For headspace, use a dedicated headspace needle with a side-port tip; the pressure cycling in headspace is harder on septa than liquid injection.

Control the crimp or the screw torque

For 11 mm crimp caps on 20 mm vials, the cap skirt should measure 19.7 to 20.0 mm after crimping, and the crimp should be flat with no visible tilt. Check with a crimp gauge every 50 caps. For 9 mm screw caps, hand-tighten plus a quarter turn is the rule; a torque of 0.8 to 1.2 N·m is typical for a 9 mm screw cap and is enough to compress the septa without cutting it. Over-torquing is common and it deforms the septa, which shortens reseal life.

Store septa properly

Store below 30°C, out of direct sunlight, in the original sealed bag. Silicone absorbs volatiles from the lab atmosphere, so do not store septa next to solvent bottles. If you buy pre-capped vials in bulk, check the lot date and use the oldest first. A septa that has been in a warehouse at 40°C for a year is not the same product as one that has been stored at 20°C.

Run a septa blank

Before committing a new lot to a critical method, run a blank vial with the same septa and cap, injected under the method conditions. Look for siloxane peaks (m/z 73, 147, 207, 281 in GC-MS) and for any peak that grows with inlet temperature. A lot that fails the blank should be rejected, not “worked around” with a longer purge.

Troubleshooting checklist

When a septa-related problem appears, work through this list in order. It takes about 15 minutes and catches most failures.

1. Check the puncture count. If the septum is past its interval, change it before doing anything else.
2. Inspect the old septum. Look for a visible hole, a coring fragment, a swollen edge, or a hardened surface. A hole larger than 0.5 mm means the needle is too large or the septum is past life.
3. Check the liner. A coring fragment in the liner is the most common cause of phantom peaks. Replace the liner if you find one.
4. Check the crimp or torque. Measure the cap skirt with a gauge for crimp vials, or check the torque for screw caps. A loose cap will leak; an over-tight cap will cut the septum.
5. Run a septa blank. If the blank shows siloxane peaks, the septum is outgassing. Switch to a high-temp or PTFE-faced septum, or increase septum purge flow to at least 3 mL/min.
6. Check the needle. Inspect the tip under a loupe. A bent or burred tip will core the septum. Replace the needle if the tip is damaged.
7. Check the storage conditions. If the septum is hardened (Shore A above 55 for standard silicone), the lot may have been stored too warm. Reject the lot.
8. Check for chemical attack. If the septum is swollen or discolored, the sample matrix is attacking the silicone. Switch to a PTFE-faced or pure PTFE septum.

What I recommend

For most labs running routine HPLC and GC, the default should be a 1.0 mm silicone/PTFE-faced septum, 11 mm for crimp vials or 9 mm for screw vials, rated to 200°C, changed at 50 to 100 punctures. Use a 23-gauge side-port needle, check the crimp with a gauge, and log the puncture count. That combination eliminates about 80% of the septa failures I see.

For high-temperature inlets (above 250°C), step up to a high-temp silicone/PTFE septum rated to 300°C with a 0.20 mm PTFE facing, and change at 40 to 80 punctures. For aggressive solvents or trace analysis, use pure PTFE or a low-extractable PTFE-faced septum and validate it with a blank. For high-throughput autosamplers with light needle carriages, use pre-slit septa and inspect the slit every 50 punctures.

If you are sourcing septa, ask for the lot certificate, the durometer, the PTFE facing thickness, and the rated temperature. A supplier who cannot provide those numbers is selling you a commodity, and a commodity is what you will get. The major instrument vendors publish their septa specs; Agilent’s consumables catalog and Thermo Fisher’s sample handling line are both good references for what a proper spec sheet looks like. Restek’s technical resources have useful guidance on septum purge and inlet maintenance, and the EPA SW-846 methods specify septa requirements for regulatory work that you should check against your method.

The takeaway is simple. Septa fail through coring, thermal degradation, chemical attack, and compression set, and each failure has a specific fix. Count your punctures, match the needle to the septum, control the crimp, store the product properly, and run a blank before you trust a new lot. Do those five things and the phantom peak at 8.2 minutes stops showing up in your sequence.