I have read more septum datasheets than I care to admit, and the one that taught me the most was a two-page PDF from a supplier we no longer buy from. Every number on it was accurate. Every number was also irrelevant to the failure we were chasing: a slow, creeping leak on a 24-vial autosampler tray that only showed up after the tray had been sitting overnight at 8 °C. The datasheet said the septum was good for 50 punctures at 25 °C. It did not say a single word about what happens at 8 °C, or about the 4 mm of headspace between the septum underside and the sample surface. That silence cost us three weeks and one very confused customer. This article is about reading that silence.
A septum datasheet is not a performance specification in the sense a chromatographer would want. It is a manufacturing conformance document. It exists so the supplier can say, in writing, “the batch you receive matches the batch we qualified,” and so that you, the buyer, have something to point at if it does not. That is a legitimate and useful purpose. It is not the same purpose as “this septum will work in your method.”
Once you understand this, the datasheet becomes much easier to read. Almost everything on it is traceable to a test the manufacturer ran on a lot of maybe 10,000 pieces, using a test rig that has very little to do with your autosampler. The numbers are real. They are just answering a narrower question than the one you are asking.
The three sections you will find on almost every sheet
Nearly every septum datasheet I have seen, from the major suppliers down to regional converters, organizes itself into three blocks:
- Identity and material: base polymer, formulation grade, any additives, colorant, and sometimes a durometer reading or a specific gravity value.
- Physical dimensions: diameter, thickness, and tolerance on each, plus sometimes a flatness or concentricity note.
- Performance claims: puncture count, temperature range, chemical compatibility, sometimes a bleed figure and a re-seal statement.
Of these three, the first block is the most under-read and the most useful. The third block is the most over-read and the least useful in isolation. I will take the identity block first because it is where the real information lives.
What the material block tells you that the performance block does not
When a datasheet says “silicone,” it is telling you almost nothing. Silicone is a family, not a compound. The properties that matter for a septum are set by the crosslink density, the filler loading, and the post-cure. A 50 Shore A silicone and a 45 Shore A silicone can differ by a factor of two in compression set, and compression set is the single property that most directly predicts whether your septum will still seal after a week in the tray.
So when I read a datasheet, I look for these specifics, in this order:
Durometer, and whether it is stated as a range or a target
A durometer of 50 ± 5 Shore A is a working specification. A durometer of “medium” is not. If the sheet gives no durometer at all, the supplier either does not test it or does not want you to know the batch-to-batch variation. Both are reasons to ask. For HPLC septa, 45 to 55 Shore A is the common band; below 45 the septum is soft enough to extrude into the injection port under pressure, and above 55 it starts to resist the needle and generates more coring debris. If you are buying for a high-pressure injection port, ask for the upper end of that band and ask for it by number.
Compression set, which almost never appears on the sheet
Compression set is the percentage of thickness a material fails to recover after being held compressed at temperature for a defined time. It is tested to ASTM D395, usually Method B, at 22 hours and a stated temperature. A good septum silicone will show compression set under 25% at 100 °C. A mediocre one will show 35% or more. The datasheet usually omits this number entirely, which is a tell. If a supplier cannot give you a compression set figure, they are selling you a shape, not a material.
Why does it matter? Because a septum in an autosampler cap is under compression from the moment the cap is crimped until the moment it is removed. If the material takes a permanent set, the seal force drops, and you get the slow leak I described at the top of this article. The leak is not a puncture failure. It is a material failure that the puncture count never had a chance to predict.
Post-cure and volatile extractables
Silicone that has not been adequately post-cured will continue to off-gas low-molecular-weight siloxanes for weeks. In GC this shows up as a rising baseline and ghost peaks. In HPLC it is subtler: you get small, drifting UV absorbance at low wavelengths, and you blame the mobile phase. A datasheet that states a post-cure schedule (for example, 4 hours at 200 °C) is telling you the supplier understands the problem. A datasheet that lists “low bleed” as a bullet point without a number is telling you they have a marketing department.
If you need a number, ask for extractables by USP <88> or a comparable gravimetric extractables test. Under 1% is typical for a good septum; under 0.5% is what you want for trace analysis at low UV wavelengths.
What the dimensions block hides behind its tolerances
This is where I see the most expensive misreadings, and it is not because the numbers are wrong. It is because the tolerances are quoted in a way that hides the interaction between septum and cap.
A typical HPLC septum is 11 mm diameter by 1.0 mm thick for a standard 2 mL crimp or snap vial, with a diameter tolerance of ± 0.1 mm and a thickness tolerance of ± 0.05 mm. Those numbers look tight. They are. But the datasheet does not tell you what the septum does when it is compressed into an 11 mm cap with a 7.5 mm aperture, which is the standard GPI 11 mm closure. Under crimp, the septum is forced to flow outward into the cap skirt, and the effective sealing diameter is not 11 mm. It is whatever the cap geometry allows. Two septa that both measure 11.00 mm on the bench can behave very differently in the cap if one is 1.02 mm thick and the other is 0.96 mm.
The thickness tolerance is the one to watch
Thickness drives three things at once: the compression force the cap applies, the dead volume above the sample, and the distance the needle has to travel before it enters the vial. A 0.05 mm thickness variation across a lot is small in absolute terms, but it is a 5% variation in compression, and it is enough to change the seal force on a snap cap. If you are running a method that is sensitive to headspace (most headspace GC methods are), this matters. If you are running a straightforward reverse-phase HPLC method with a 10 µL injection, it does not. Match the tolerance to the method, not to the catalog.
What is not on the sheet: the cap interaction
No septum datasheet I have ever seen describes the cap it is meant to sit in. That is a gap you have to fill yourself. The vial closures you buy should specify the aperture diameter, the crimp or snap style, and the recommended crimp force. If they do not, ask. A septum is only half a seal. The other half is the closure, and the datasheet is silent on it by design.
For context on how the septum and the HPLC vial interact at the neck, the vial drawing is often more informative than the septum sheet, because it shows the sealing land and the neck finish. Read them together.
The performance block: puncture count, temperature, and the two numbers that mislead
Puncture count and temperature range are the two headline numbers on every septum datasheet, and they are the two that most often lead buyers astray. Here is what each one actually measures.
Puncture count
A puncture count of 50 or 100 is generated on a test rig, usually with a single needle of a stated gauge, at a stated speed, at room temperature, with the septum held in a defined fixture. The number is the point at which the septum either cores through or fails a leak test. It is a comparative figure between materials, and it is useful for that. It is not a prediction of your autosampler’s behavior, because your autosampler does not puncture the way the rig does.
Three variables move the real number far from the datasheet number:
- Needle gauge and bevel. A 22-gauge blunt-tip needle and a 26-gauge bevelled needle do very different things to a silicone septum. The bevel slices; the blunt tip displaces. A datasheet that does not state the needle is not telling you which failure mode the count refers to.
- Puncture pattern. Real autosamplers puncture the same septum repeatedly in nearly the same spot. The rig usually punctures a fresh spot each time. The same-spot case is far harsher, and it is the case you actually run.
- Dwell time between punctures. A septum that reseals in 2 seconds may not reseal in 2 seconds after 200 punctures. The datasheet number is usually a fresh-septum number.
If you need a working number, halve the datasheet count and treat that as your change-out interval. That is conservative, and in my experience it is about right for a 22-gauge needle at 10 µL injection volumes.
Temperature range
The datasheet will say something like -40 °C to 200 °C. That range is the temperature at which the material does not visibly degrade. It is not the range at which it seals. Sealing performance starts to fall off well before the upper limit, because compression set rises with temperature and the material softens. A silicone rated to 200 °C is a poor choice for a 150 °C headspace method if the same septum has to hold pressure for 30 minutes, because at 150 °C it is already losing seal force.
The practical rule I use: for any method above 100 °C, ask for compression set at the method temperature, not at the datasheet’s stated maximum. If the supplier cannot provide it, treat the upper half of the temperature range as marketing.
Chemical compatibility, which is always qualitative
Every datasheet has a compatibility table, and every compatibility table is a grid of “excellent / good / fair / poor” with no test conditions attached. That grid is useful for ruling out obviously wrong materials (silicone in contact with non-polar solvents for extended periods, for example) and almost useless for anything else. If your method uses a solvent the table marks “good,” that does not mean the septum will not swell. It means it will swell less than the “fair” column. Ask for a swell figure in the specific solvent, or run a 24-hour immersion test yourself.
A worked comparison: three septa, three datasheets
To make this concrete, here is how I would read three hypothetical datasheets for a standard 11 mm HPLC septum, all sold at similar prices. None of the numbers are invented; they reflect the range I see in real supplier documentation.
- Septum A: silicone, 50 ± 5 Shore A, 11.0 ± 0.1 mm × 1.0 ± 0.05 mm, 100 punctures, -40 to 200 °C, “excellent” chemical resistance, no compression set, no extractables figure.
- Septum B: silicone, 48 ± 3 Shore A, 11.0 ± 0.05 mm × 1.0 ± 0.03 mm, 50 punctures, -40 to 180 °C, compression set 22% at 100 °C, extractables 0.4% by USP <88>.
- Septum C: silicone/PTFE laminate, 55 ± 5 Shore A, 11.0 ± 0.1 mm × 1.0 ± 0.05 mm, 30 punctures, -20 to 150 °C, compression set 18% at 100 °C, no extractables figure but a stated post-cure.
On the headline numbers, A looks best: 100 punctures, widest temperature range, cheapest per piece. On the numbers that predict failure, B is the better buy for most HPLC work: tighter tolerances, a real compression set figure, and a stated extractables value. C is the right choice only if you are running a method where the sample contacts the septum and you need the PTFE face to stop silicone extractables, and you can live with 30 punctures and a lower temperature ceiling.
This is the whole exercise in one table. The datasheet that looks strongest on the performance block is often the weakest on the identity block, and the identity block is the one that determines whether the septum still seals on day four.
Decision checklist with concrete thresholds
When I qualify a new septum supplier, I run through this list and will not move to a trial without answers to the first six:
- Durometer stated as a range? Accept 45 to 55 Shore A for standard HPLC. Reject “medium.”
- Thickness tolerance at or below ± 0.05 mm? Accept. Above ± 0.1 mm, ask why.
- Compression set figure at 100 °C? Accept under 25%. Above 30%, do not use for long-dwell methods.
- Extractables figure? Accept under 1% gravimetric; under 0.5% for low-UV work.
- Post-cure stated? Accept a schedule; reject silence.
- Puncture test conditions stated (needle gauge, speed, fixture)? Accept; otherwise treat the count as comparative only.
- Cap and closure specification available? If not, source closures from a supplier who publishes it. See the vial closures range for examples of what a complete closure spec looks like.
- Lot traceability? Accept a lot number on the bag; reject unmarked bulk bags.
What the datasheet hides about handling and installation
Even a good datasheet stops at the loading dock. It does not tell you how the septum behaves once it is in your hands, and three of the most common failure modes I see are installation errors that no specification can prevent.
Crimp force
For an 11 mm aluminum crimp cap, the recommended crimp force is typically in the range of 1.5 to 2.5 kN, set by the crimper and verified with a crimp-height gauge. Too little force and the septum is not compressed enough to seal; too much and you deform the septum and crush the vial neck. The datasheet gives you a septum thickness, not a crimp force. If you are buying HPLC septa in bulk, ask the supplier for the recommended crimp setting for their thickness tolerance, and then verify it on your own crimper with a gauge. I have seen a 0.1 mm thickness difference between two lots turn a good crimp into a leaking one.
Needle alignment and same-spot puncturing
An autosampler that is out of alignment by even 0.3 mm will puncture the same spot every time, and the septum will core through in a fraction of its rated count. This is not a septum defect. It is a mechanical one. Check needle alignment against the vial center at the start of every campaign, and if you can, rotate the tray position or use a septum with a slightly larger diameter to spread the punctures.
Storage and shelf life
Silicone septa absorb volatiles from their packaging and from the lab air. A bag left open on a bench next to a solvent bottle will pick up enough contamination to show up as ghost peaks in a sensitive method. Store sealed, in a cool dark place, and use within the shelf life stated on the bag. If the datasheet does not state a shelf life, assume 24 months from manufacture and ask for the manufacture date on the lot.
Troubleshooting checklist: when the septum is the suspect
When a method starts drifting and you suspect the septum, work through this list in order. In my experience the septum is the cause about a third of the time, and the other two thirds are the cap, the needle, or the sample.
- Ghost peaks at low UV? Run a blank injection with a fresh septum and a sealed vial. If the peaks disappear, the septum is off-gassing. Check post-cure and extractables.
- Slow pressure decay on a headspace vial? Check compression set and crimp force before blaming the septum. A 30% compression set septum will leak at 80 °C even with a perfect crimp.
- Coring debris in the vial? Check needle gauge and bevel, and check same-spot puncturing. A 26-gauge bevelled needle will core a soft septum faster than a 22-gauge blunt one.
- Inconsistent injection volumes? Check septum thickness tolerance and cap aperture. A 0.1 mm thickness variation changes the headspace volume enough to move a headspace GC result.
- Septum swells or softens? Check solvent compatibility with an immersion test, not the datasheet grid.
- Seal fails only after several days in the tray? This is compression set, almost every time. Ask for the ASTM D395 figure at your storage temperature.
If you want an independent read on a suspect lot, send a sample to a lab that runs analytical testing to USP <88> or an equivalent extractables protocol. It costs less than a week of troubleshooting and it gives you a number to take back to the supplier.
Where to look for the numbers the datasheet leaves out
When a supplier’s datasheet is thin, the gap is usually fillable from public sources. The major instrument vendors publish sample handling guidance that covers septum selection for their autosamplers; Agilent’s chromatography consumables documentation is a good starting point for GC-side septum behavior and bleed. Thermo Fisher’s chromatography and sample handling pages cover autosampler geometry and needle specifications. For method-level requirements that imply a septum specification, the U.S. EPA SW-846 test methods are worth reading for the headspace and volatile methods, because they state the vial and closure requirements explicitly. And Restek’s technical resources have some of the clearest published guidance on septum bleed and compression set that I have found outside a supplier’s own lab.
None of these replace a supplier’s own data. What they do is give you a benchmark. If a supplier’s compression set figure is far from what the public literature says is achievable for the material they claim, ask them to re-test.
The buying decision, stated plainly
If you take one thing from this article, take this: the datasheet is a conformance document, not a performance guarantee. Read it for the identity block first, the dimensions second, and the performance claims last. Ask for compression set, extractables, and post-cure by name, because those three numbers predict failure better than puncture count and temperature range ever will. Match the thickness tolerance to your method’s sensitivity to headspace, and match the durometer to your injection port pressure. Then verify the crimp force on your own equipment and check needle alignment before you blame the septum for a leak.
The septum that fails on day four is almost never the one with the lowest puncture count on the datasheet. It is the one whose datasheet was silent on compression set. Buy from suppliers who publish the numbers that predict failure, and treat the rest as a starting point for a conversation, not a specification.
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: