How a vial causes a pressure spike in the first place

I spent the better part of a Tuesday last March chasing a pressure spike that appeared on every sixth or seventh injection on a 7697A. The baseline would drift up about 4 to 6 pA over the course of an hour, then snap back after a blank. The obvious suspects were the transfer line, the loop, the carrier regulator. All clean. The actual cause was a box of 20 mL headspace vials with a crimp finish that sat about 0.25 mm proud of the nominal seat, and a batch of aluminium crimp caps whose septa had been bonded slightly off-centre. The autosampler was seating them, pressurising them, and leaking a few microlitres of carrier past the septum at every injection. The leak was small enough that the pressure controller could compensate without throwing a fault, so no error was logged. The only visible symptom was a baseline that wandered.

That is the pattern I want to walk through here. Pressure spikes and baseline drift on a headspace GC rarely announce themselves as vial problems. They read as instrument faults, and people spend days on the instrument before they think to look at the consumable. This article is a field procedure for tracing those symptoms back to the vial, the closure, or the crimp, with the specific measurements that separate a vial defect from an instrument fault.

A headspace autosampler runs a pressure-controlled loop fill. The vial is pressurised to a setpoint (typically 10 to 20 psi above the loop pressure, or a fixed value like 15 psi depending on method), the loop is filled for a fixed time, the valve switches, and the loop is swept to the column. The pressure controller in the sampler is a closed-loop device. It will chase a leak by opening the carrier valve wider until the setpoint is met, or until it hits a limit and throws an alarm.

Two failure modes emerge from this. The first is a slow leak across the septum or around the crimp. The controller compensates, the pressure at the loop is technically correct, but the flow through the loop during fill is higher than the method assumes. The result is a slight dilution of the headspace sample and a baseline that drifts because the carrier flow to the column is not what the method was validated against. The second is an intermittent seal: the vial holds pressure at 40 °C but leaks at 80 °C, or holds on one injection and leaks on the next because the septum re-seals slightly differently each time the needle withdraws. Intermittent seals produce the classic spike-then-recover pattern in the pressure trace and a baseline that steps rather than drifts.

Neither of those is the instrument’s fault. Both are vial or closure faults, and both are diagnosable with a leak-down test and a few calipers.

What the pressure trace actually shows you

If your sampler logs pressure (the Agilent 7697A and the Thermo TriPlus both do, at least during method runs), pull the trace for the last 50 injections. Look for:

• Staircase drift upward over 20 to 50 injections followed by a reset: this is a slow leak that the controller is compensating for until it hits a limit or the method re-zeroes. Vial septum or crimp.
• Single-injection spikes that recover immediately: intermittent seal, usually a septum that has been punctured slightly off-centre or a crimp that is loose on one side.
• Spikes that correlate with vial position in the tray: a mechanical problem with the tray or the transfer needle, not the vial.
• Spikes that correlate with sample batch (all vials from one box): a vial lot problem.

The correlation test is the fastest discriminator. If the spikes follow the vial lot, the vial is the cause. If they follow the tray position, the sampler is the cause. If they follow the operator who crimped the caps, the crimper is the cause.

The dimensional stack that decides whether a vial seals

A 20 mm crimp headspace vial has four dimensions that matter for sealing: the crimp bead diameter, the bead height, the neck diameter below the bead, and the mouth flatness. The relevant standard is the GPI 20 mm finish (Glass Packaging Institute), and the tolerances are not generous. A vial that is 0.2 mm over on bead diameter will not seat in a standard crimp cap. A vial that is 0.3 mm under will seat but the cap will not grip, and the seal will fail under pressure.

I keep a go/no-go gauge on the bench for exactly this reason. It costs about £40 and it saves a day of troubleshooting every quarter. The gauge checks bead diameter at 20.0 ± 0.2 mm and bead height at 6.0 ± 0.3 mm for a standard 20 mm crimp vial. If a vial fails the gauge, it does not go in the tray. That single rule has eliminated about 80% of the pressure spikes I used to see on crimp-finish vials.

The other dimension that matters is mouth flatness. The top of the vial mouth is where the septum sits. If the mouth is not flat within about 0.1 mm, the septum will not seat evenly and the crimp will compress one side more than the other. This is a moulding defect, and it is the one I find most often on cheap vials. You cannot see it by eye. You can see it by placing the vial on a flat surface and rocking it, or by checking the seal with a leak-down test at the method temperature.

Crimp versus screw thread: the pressure question

Screw-thread headspace vials (the 18 mm and 20 mm magnetic or screw caps) are more forgiving on dimensional tolerance because the cap torque is set by the autosampler or by a torque-limited capper. But they have their own failure mode: under-torqued caps leak under pressure, and over-torqued caps deform the septum and cause coring. The typical torque window for an 18 mm screw cap is 1.5 to 2.0 N·m. For a 20 mm magnetic cap, the autosampler handles it, but the cap has to be seated squarely or the magnet will not pull it down evenly.

My rule: if the method runs above 80 °C or above 15 psi vial pressure, use crimp. If the method is below 80 °C and the lab is doing fewer than 50 samples a day, screw thread is fine and less operator-dependent. The headspace vial page on our site lists the dimensional tolerances for both finishes, and it is worth checking your current lot against them before you blame the instrument.

Septum behaviour under pressure and temperature

The septum is the part of the closure that does the actual sealing, and it is the part that fails most often. A standard headspace septum is a silicone rubber disc, often faced with PTFE on the sample side, typically 3.0 mm thick and 20 mm diameter for a 20 mm cap. The silicone provides the reseal, the PTFE provides the chemical barrier.

Three things go wrong with septa under pressure:

• Compression set. Silicone takes a permanent deformation under sustained compression at temperature. At 100 °C, a standard silicone septum will take a set of about 15 to 20% after 30 minutes. Once it has set, it no longer reseals after the needle withdraws, and the vial leaks. This is why septum temperature rating matters: a septum rated to 120 °C is not the same as one rated to 200 °C, and the rating is about compression set, not about melting.
• Coring. A blunt or burred needle punches a plug out of the septum instead of slicing through it. The plug can lodge in the needle, in the loop, or in the transfer line, and it produces exactly the kind of intermittent pressure spike that looks like an instrument fault. I have pulled 0.8 mm silicone plugs out of a transfer line that had been causing a 3 psi spike every 40 injections.
• Chemical attack. The PTFE facing is not a guarantee. It is a barrier, and it is thin (typically 0.05 to 0.1 mm). If the sample contains a solvent that swells silicone (toluene, xylene, some chlorinated solvents), the silicone under the PTFE will swell and the PTFE will wrinkle. A wrinkled septum does not seal. This is the cause of a lot of “baseline drift” that gets blamed on column bleed.

The puncture count matters too. A standard silicone/PTFE septum is good for about 3 to 5 punctures at the same point before the reseal degrades. If your method uses a single vial for multiple injections (which is common in equilibrium headspace), rotate the injection point or accept that the seal is degrading from injection 4 onward. This is one of the reasons a GC vial specification has to state puncture count explicitly. It is not a marketing number; it is a method parameter.

Temperature rating is a compression-set number

When a septum datasheet says “rated to 200 °C”, what it means is that the silicone compound retains enough elasticity after 30 minutes at 200 °C to reseal. It does not mean the septum is inert at 200 °C, and it does not mean the PTFE facing will stay intact. For methods at 150 °C or above, I use a higher-grade silicone or a butyl septum, and I check the compression set after the run by looking at the needle track. If the track has not closed, the septum is done.

For most headspace methods (40 to 90 °C), a standard silicone/PTFE septum rated to 120 °C is fine. The pressure spikes I see in that range are almost never septum temperature failures. They are dimensional failures at the crimp, or coring from a worn needle.

A leak-down test you can run in ten minutes

This is the single most useful diagnostic I know for pressure spikes, and almost nobody runs it. The procedure:

1. Take a vial from the suspect lot, cap it with the suspect cap, and crimp it with the suspect crimper.
2. Pressurise it to 20 psi with dry nitrogen through a septum-piercing needle connected to a gauge.
3. Withdraw the needle, and watch the gauge for 60 seconds.
4. A good vial holds pressure within 0.5 psi over 60 seconds. A vial that drops more than 2 psi has a seal problem.
5. Repeat at the method temperature (put the vial in a block or oven at the method temperature, then pressurise and test).

If the vial fails at room temperature, the problem is dimensional (crimp, bead, mouth flatness). If it passes at room temperature and fails at temperature, the problem is the septum (compression set or chemical attack). That single distinction tells you whether to look at the vial or the closure.

I have run this test on three lots of vials in the last year. Two lots failed at room temperature on bead diameter (both were 0.3 mm over, and both were from the same supplier). One lot passed at room temperature and failed at 80 °C, and the cause was a septum with a PTFE facing that had been applied too thin. The Thermo Fisher chromatography and sample handling documentation has a similar procedure in its headspace troubleshooting guide, and the Restek technical resources site has a good write-up on septum compression set that is worth reading if you want the material science behind it.

Baseline drift: separating vial causes from column and carrier causes

Baseline drift on a headspace GC has at least five plausible causes: column bleed, carrier gas contamination, detector drift, septum bleed, and vial leakage. The vial cause is the one people check last, so here is how to move it up the list.

Column bleed produces a monotonic rise that correlates with oven temperature programme, not with injection number. If the baseline rises during the temperature ramp and falls during cooldown, it is the column, not the vial.

Carrier contamination produces a rise that correlates with total carrier volume used, not with sample. If the drift is the same on blanks as on samples, it is the carrier.

Detector drift produces a rise that is independent of injection. Run a no-injection sequence and watch the baseline.

Septum bleed produces a rise that correlates with septum temperature and with the number of punctures. It shows up as siloxane peaks (m/z 73, 147, 207 on MS, or a rising baseline on FID) and it is worse at higher temperatures.

Vial leakage produces a rise that correlates with vial lot and with the pressure trace. This is the one that gets misdiagnosed. If the drift disappears when you switch to a known-good lot of vials, and reappears when you switch back, the vial is the cause. That is a two-hour test and it is worth running before you replace a column.

The blank-vial control

Run a sequence of 10 blank vials (empty, capped, crimped) through the method, and watch the pressure trace and the baseline. A blank vial cannot produce sample carryover, so any drift or spike in the blank sequence is either the vial/closure or the instrument. If the blank sequence is clean and the sample sequence drifts, the cause is in the sample matrix (chemical attack on the septum) or in the sample preparation. If the blank sequence drifts, the cause is the vial, the closure, or the instrument.

I run this control every time I change vial lots. It takes 40 minutes and it has caught three bad lots in the last two years.

Decision checklist: vial or instrument?

Here is the checklist I use on the bench. It is not exhaustive, but it separates the vial cause from the instrument cause in about 90% of cases.

The two tests that matter most are the go/no-go gauge on bead diameter and the leak-down test at method temperature. Those two catch the majority of vial-caused pressure spikes. The blank-vial control catches the rest.

Specifications that prevent the problem

If you are specifying a headspace vial for a method that runs above 60 °C or above 10 psi, here are the numbers I would put in the specification:

• Finish: GPI 20 mm crimp, bead diameter 20.0 ± 0.2 mm, bead height 6.0 ± 0.3 mm.
• Mouth flatness: ±0.1 mm. Ask for the inspection method, not just the number.
• Glass: Type I borosilicate, 1.0 mm wall nominal, USP <660> compliant.
• Septum: silicone/PTFE, 3.0 mm thick, rated to at least 20 °C above the method temperature, puncture count specified as 3 to 5 at a single point.
• Crimp: aluminium, 0.2 mm wall, crimped to a torque that produces a leak-down of less than 0.5 psi per minute at 20 psi.
• Lot release: dimensional gauge on 10 vials per lot, leak-down on 3 vials per lot at method temperature.

That last item is the one most labs skip, and it is the one that would have caught the box that cost me a day in March. A supplier who will not run a leak-down test at method temperature is a supplier who does not understand what the vial does under pressure.

For methods below 60 °C and below 10 psi, the specification can be looser, but the dimensional stack still matters. A vial that does not seat in the crimper will not seal regardless of temperature. The HPLC autosampler vial specification pages on our sister site cover the equivalent dimensional stack for liquid injection, and the principle is the same: the vial is a mechanical component in a pressure system, and it has to be specified as one.

Operational precautions that reduce the failure rate

Three habits cut the pressure-spike rate in my lab by more than half:

1. Crimp with a calibrated crimper, and check it monthly. A crimper that is 0.2 mm out of adjustment will produce a 20% failure rate on a good lot of vials. I keep a reference vial and a go/no-go gauge next to the crimper, and I check the crimp height on the first vial of every batch. The crimp height for a 20 mm aluminium cap is typically 7.5 to 7.9 mm measured from the vial base to the top of the cap. Outside that window, the seal is unreliable.

2. Rotate the injection point. If your method uses a single vial for multiple injections, move the needle position by 1 mm between injections, or use a sampler that does. A septum that is punctured three times at the same point will leak under pressure. Three punctures at three points will not.

3. Replace the transfer needle on a schedule, not on failure. A needle that has done 500 punctures is blunt enough to core a septum. I replace at 500 regardless of appearance. The cost of a needle is trivial compared to the cost of a day chasing a pressure spike.

The U.S. EPA SW-846 test methods for headspace analysis (Method 5021A and Method 8260D) specify vial and septum requirements that are worth reading if you want a regulatory frame for the specification. They do not go into the dimensional tolerances, but they do specify the temperature and pressure ranges, and those are the numbers that drive the septum choice. For reagent and solvent handling that feeds into the same method, the reagent bottle specification follows the same logic: the container is part of the method, not an accessory to it.

What I would change if I were starting over

I would put a go/no-go gauge and a leak-down rig on the bench on day one, and I would write the lot-release criteria into the purchasing contract. The pressure spikes and baseline drifts I have chased over the years have almost all come down to one of three things: a vial that was 0.2 mm out of tolerance, a septum that had taken a compression set at temperature, or a crimper that was out of adjustment. None of those is visible on the instrument, and none of them throws a fault code. They just produce a baseline that wanders and a pressure trace that steps, and they cost days.

The fix is not complicated. Specify the dimensional stack, test it on receipt, and test the seal at method temperature. A supplier who can give you the bead diameter tolerance and a leak-down number is a supplier worth keeping. A supplier who cannot is a supplier who will cost you a day in March.

The next time you see a pressure spike, pull the pressure trace before you pull the column. If the spike correlates with vial lot, the vial is the cause. Run the leak-down test at method temperature, check the bead diameter with a gauge, and check the crimper against a reference vial. In my experience, that sequence finds the cause in under an hour, and the cause is almost never the instrument.

References & further reading

Related resources from our lab consumables network:

• Headspace Vial at headspacevialsupplier.com
• Gc Vial at gcvialfactory.com
• Reagent Bottle at reagentbottleforsale.com
• Hplc Autosampler Vial at hplcautosamplervials.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