A QC lab logged 14 needle-bend faults in one overnight sequence. The autosampler kept flagging the same vial position, and the operator assumed the needle guide had drifted. It had not. The septum lot was 0.010 in thicker than the drawing, and the datasheet listed only a nominal thickness with no tolerance band. Nobody caught it because nobody knew which numbers on that sheet were measured and which were copied from a template. That is the core problem with septa datasheet specifications: most of them mix tested values with nominal values, and the difference is where failures hide.
Every septa datasheet has a line for hardness. It usually says 45 Shore A, or 50 Shore A, or 55 Shore A. That number is almost always nominal. It comes from the compound supplier’s typical value for that silicone or PTFE grade, not from a durometer reading on the finished septum you are buying. A hardness value without a test method reference, without a tolerance, and without a batch number is a decoration.
But there is one line that separates vendors who test from vendors who copy: compression set. Compression set is the percentage of thickness a septum fails to recover after being squeezed under a defined load at a defined temperature for a defined time. ASTM D395 Method B is the common reference. A real datasheet will say something like “Compression set, 22 h at 70°C, 25% max.” That is a tested value with a method and a limit. A nominal-only sheet will omit compression set entirely or list “low” as the value. If compression set is missing, the vendor has not characterized the septum’s resealing behavior after an autosampler needle has been through it 40 times. That is the number that predicts whether your 100-injection sequence holds pressure or starts leaking at injection 60.
Nominal versus tested: the four lines that matter on a septa sheet
When I read a septa datasheet, I sort every line into one of two columns. Tested values have a method, a tolerance, and a pass/fail limit. Nominal values are typical values from a compound data sheet or a CAD drawing, and they carry no guarantee. Most buyers cannot tell the difference because the sheet does not say which is which. Here is the practical split.
Thickness: nominal unless a tolerance is printed next to it
A septum listed as “0.125 in thick” is nominal. The real question is whether the vendor holds ±0.005 in or ±0.010 in. For a 9 mm autosampler septum, a ±0.010 in variation changes the effective crimp compression and the puncture force. One lot at 0.135 in and the next at 0.115 in will both fit the same cap, but the needle sees different resistance, and the seal behaves differently. A tested thickness line looks like “0.125 ±0.005 in, measured per ASTM D3767.” If there is no tolerance, assume the vendor has not measured it lot by lot. I have rejected three lots in five years for thickness drift that a proper tolerance would have caught at incoming inspection.
Diameter: nominal unless it references a dimensional standard
Septa diameters are cut from sheet or molded to a nominal size. A 9.0 mm septum may run 8.8 to 9.2 mm depending on the cutting die wear and the compound’s shrinkage. The datasheet that says “9.0 mm” without a tolerance is telling you the die size, not the part size. A tested diameter references something like ISO 3601 or a customer drawing with a ±0.1 mm tolerance. The difference shows up in cap assembly: an undersized septum shifts during crimping and creates an off-center puncture path. That is a fit problem, but the datasheet number is the first place to catch it.
Hardness: nominal unless it lists a durometer scale and a tolerance
Shore A hardness on a septum sheet is the most over-read number in the industry. A 45 Shore A silicone septum and a 50 Shore A silicone septum may come from the same compound batch with normal variation. A tested hardness line says “Shore A 45 ±5, ASTM D2240, 6 mm sample thickness.” Without the sample thickness, the reading is meaningless because durometer readings change with material thickness. Without the tolerance, the number is a marketing average. I have seen two lots with identical “50 Shore A” on the sheet measure 44 and 57 on a calibrated durometer. Both were within the compound supplier’s typical range. Neither was within a useful specification.
Puncture count or reseal rating: tested only if the sheet names the needle gauge and the leak threshold
Some datasheets claim “reseals after 50 punctures.” That claim means nothing without three pieces of information: the needle gauge used, the puncture speed and angle, and the leak detection method. A real reseal specification reads like this: “Minimum 50 punctures with a 23-gauge, 22° bevel needle at 10 mm/s, no leak >0.01 mL/min at 15 psi after puncture 50.” That is a tested value. If the sheet says only “resealable,” it is a nominal claim copied from the material’s general reputation. PTFE-faced silicone septa reseal differently than pure silicone, and a 26-gauge SPME needle tears less than a 19-gauge blunt needle. The gauge changes the puncture count by a factor of three or more.
How each line is actually measured, and what the sheet rarely tells you
Once you know which lines are tested, the next question is how. The method determines whether the number means anything for your application. Here is what happens on the test bench, and what a careful vendor will disclose.
Thickness measurement: contact versus non-contact
Thickness is usually measured with a dead-weight micrometer per ASTM D3767. The problem is that septa are compressible. A micrometer with a 0.25 in diameter foot and a 50 g load will read a soft silicone septum 0.002 to 0.004 in thinner than a non-contact laser gauge. A datasheet that says “measured per ASTM D3767” but does not state the foot diameter or load is hiding a variable. For firm PTFE-faced septa, the difference is small. For soft, low-durometer silicone, it is not. I specify contact measurement with a 0.25 in foot and a 50 g dead load for incoming inspection, and I reject lots that drift more than ±0.005 in from the drawing.
Hardness measurement: sample stack height changes the number
ASTM D2240 requires a minimum 6 mm sample thickness for Shore A. A single septum is often 1 to 3 mm thick, so the lab stacks several septa or measures a molded plaque of the same compound. A stacked sample reads differently than a molded plaque because the interfaces between layers absorb some of the indenter force. A vendor who measures a molded plaque and reports that number on a finished septum sheet is giving you a nominal value, not a part value. The honest approach is to report the plaque value and label it as such, or to report a stacked-septum value with the stack height noted. If the sheet does not say, assume it is a plaque value.
Compression set: the test that predicts reseal failure
Compression set is measured by compressing a sample to a fixed deflection, usually 25%, holding it at temperature for 22 hours, releasing it, and measuring the recovery after 30 minutes. ASTM D395 Method B is the standard. For a septum, a compression set above 30% means the material has permanently deformed under the crimp and the needle path. That permanent deformation is what causes the septum to leak after the autosampler has cycled 50 or 100 times. A tested compression set line on a datasheet is the single best predictor of long-sequence performance. A sheet without it is guessing.
Puncture force: the number that protects your needle
Puncture force is measured by driving a needle of known gauge and bevel through the septum at a controlled speed and recording the peak force. A typical value for a 23-gauge needle through a 0.125 in PTFE-faced silicone septum is 1.5 to 3.5 N. A value above 5 N will bend or dull needles in high-throughput autosamplers. The datasheet that lists puncture force at all is rare. The one that lists the gauge, bevel angle, speed, and peak force limit is rarer. If you run 500 injections per day, a 1 N difference in puncture force is the difference between a clean sequence and a stack of needle-bend errors. This is the number I ask for first when qualifying a new septum lot.
How to spot a sheet that hides its test method
Most septa datasheets are one page. They list dimensions, material, hardness, temperature range, and a few compatibility notes. The test method is usually absent. That absence is a signal, not an accident. A vendor who tests will tell you how. A vendor who does not test will give you a clean, confident number with no method attached. Here are the tells.
Tell one: no tolerance bands. A number with no tolerance is a nominal value. “Thickness: 0.125 in” is a CAD dimension. “Thickness: 0.125 ±0.005 in” is a specification. The tolerance is the difference between a drawing and a guarantee.
Tell two: no ASTM, ISO, or USP reference. A tested value references a method. A nominal value references nothing. If the sheet says “hardness 50 Shore A” with no ASTM D2240, no sample thickness, no tolerance, the vendor has not measured it on the finished part.
Tell three: the temperature range is suspiciously wide. A silicone septum listed as “-60°C to 250°C” is quoting the raw silicone compound’s general range, not the septum’s functional range. The septum in a crimp cap under autosampler cycling will not perform at 250°C for 100 injections. A tested temperature rating comes from a specific application test, like “100 injections at 250°C inlet, no leak >0.01 mL/min.” If the sheet gives a wide range with no test context, it is a compound data point, not a septum data point.
Tell four: no lot traceability. A tested value is meaningless without a lot number. If the vendor cannot tell you which lot the datasheet values came from, they are not running lot-by-lot testing. They are running a type test once and reprinting the sheet. A type test is useful for qualification, but it does not protect you from lot-to-lot drift. Ask for the certificate of analysis for the specific lot you are buying. If they cannot produce one, the datasheet numbers are nominal.
The two field cases that taught me which numbers to trust
Case one: the drifting hardness lot that passed every incoming check
A pharma QC lab running 120 samples per day on two GC systems started seeing split peaks on late-eluting compounds. The inlet pressure was stable, the column was new, and the liner was clean. The septum was a 11 mm PTFE-faced silicone part rated at 50 Shore A. The datasheet said 50 Shore A, no tolerance. We pulled 10 septa from the current lot and measured them on a calibrated durometer with a 6 mm stack. The average was 57 Shore A, and the range was 54 to 61. The previous lot averaged 49. The harder lot required higher crimp force to seal, and the autosampler’s standard crimp setting was under-compressing the septum. The fix was not a new septum design. It was a revised incoming inspection that measured hardness on every lot with a tolerance of ±5 Shore A. The vendor’s datasheet was not wrong. It was just nominal, and nominal does not protect you from a compounding drift of 8 points.
Case two: the puncture count claim that ignored needle gauge
An environmental lab running EPA SW-846 methods on a purge-and-trap autosampler was replacing septa every 40 injections because of carryover. The septum datasheet claimed “reseals after 100 punctures.” The lab used a 22-gauge needle with a blunt side-port design, not the 26-gauge standard the vendor had used for its puncture test. The blunt 22-gauge needle tore a larger hole per puncture, and the septum lost seal integrity after 35 to 45 cycles. The datasheet’s 100-puncture claim was tested, but with a needle that did not match the lab’s application. The fix was a septum with a thicker PTFE facing and a higher-durometer core, plus a revised specification that required the vendor to test puncture count with the lab’s actual needle geometry. The lesson: a tested number is only useful if the test conditions match your conditions. Ask for the needle gauge, bevel, and speed before you trust a reseal rating.
A comparison table for reading septa datasheet lines
Here is a decision checklist I use when reviewing a septa datasheet. It sorts the common lines into tested, nominal, and missing categories, with the concrete numbers I expect to see.
| Datasheet line | Nominal version (what you usually see) | Tested version (what to ask for) | Red flag if missing |
|---|---|---|---|
| Thickness | 0.125 in | 0.125 ±0.005 in, ASTM D3767, 0.25 in foot, 50 g load | No tolerance band |
| Diameter | 9.0 mm | 9.0 ±0.1 mm, ISO 3601 or customer drawing | No tolerance, no drawing reference |
| Hardness | 50 Shore A | 50 ±5 Shore A, ASTM D2240, 6 mm stack, 1 s reading | No scale, no tolerance, no sample thickness |
| Compression set | “Low” or omitted | 25% max, 22 h at 70°C, ASTM D395 Method B | Line absent entirely |
| Puncture force | Omitted | 2.5 ±1.0 N, 23-gauge, 22° bevel, 10 mm/s | No gauge or speed specified |
| Reseal rating | “Resealable” or “100 punctures” | 100 punctures min, 23-gauge, 22° bevel, no leak >0.01 mL/min at 15 psi | No needle geometry, no leak threshold |
| Temperature range | -60°C to 250°C | 100 injections at 250°C inlet, no leak >0.01 mL/min | No application test context |
| Lot traceability | None | Certificate of analysis per lot with actual measured values | No C of A available on request |
This table is the core of my incoming inspection workflow. If a vendor cannot fill the right-hand column for at least three of the eight lines, I treat the datasheet as a nominal document and run my own qualification before approving the lot for production use.
What to ask the vendor before you trust a single number
When I qualify a new septum supplier, I send a short list of questions. The answers tell me within one email whether the datasheet is tested or nominal. Here is the list.
- Which lot was used to generate the values on this datasheet? Can you send the certificate of analysis for that lot?
- What is the thickness tolerance, and what instrument and load did you use to measure it?
- Is the hardness value from a molded plaque or a stacked finished septum? What sample thickness and durometer scale?
- Do you test compression set on finished septa? What method, time, and temperature?
- What needle gauge, bevel angle, and speed did you use for the puncture force and reseal ratings?
- What is your lot-to-lot variation for thickness, hardness, and compression set over the last 12 months?
A vendor who answers all six with specific numbers and method references is testing. A vendor who answers with “our material is high quality” or “we follow industry standards” is not. The difference shows up in your autosampler within two lot changes.
The operational precautions that follow from reading the sheet correctly
Once you know which numbers are tested and which are nominal, your incoming inspection changes. You stop accepting a one-page datasheet as a quality document. You start requiring a certificate of analysis with the actual measured values for the lot you are buying. You set your own tolerance bands based on your application, not on the vendor’s drawing. For a 9 mm autosampler septum, I hold thickness to ±0.005 in, hardness to ±5 Shore A, and compression set to 30% max. For an 11 mm GC inlet septum running at 250°C, I add a puncture force limit of 5 N max with a 23-gauge needle and require a reseal test at the actual inlet temperature, not at room temperature.
You also start tracking lot numbers in your autosampler logs. When a sequence fails with needle-bend or leak errors, you check the septum lot first. A lot that drifts 0.008 in thicker or 6 points harder will show up as a cluster of errors in one overnight run. That is the signal. The datasheet will not tell you the lot drifted. Your data will.
For sourcing, I work with suppliers who publish tested values on their septa pages and who can produce a C of A on request. The same discipline applies to vial septa for headspace and LC applications, where thickness and hardness drift causes crimp failures. When I specify HPLC vial consumables, I ask for the same compression set and puncture force data, because an LC autosampler needle cycling through a cap at 10,000 injections per month sees the same cumulative damage as a GC inlet. And for autosampler vial setups running overnight sequences, the septum is the one component that touches the needle every single injection. The datasheet numbers are the first line of defense, but only if you know which ones were measured.
Where the standards stop and your own testing has to start
The external references help, but they do not close the gap. Agilent publishes application notes for their inlet septa that include puncture counts and temperature limits, but those numbers apply to their specific septum geometry and compound. Thermo Fisher documents septum compatibility for their autosampler lines, but the compatibility tables do not tell you lot-to-lot variation. EPA SW-846 methods set performance criteria for purge-and-trap and headspace systems, but they do not specify septum hardness or compression set. Restek technical resources include septum selection guides with temperature and puncture data, but again, those are type-test values for their catalog parts.
The gap is always the same: a standard or a vendor guide tells you what a septum should do, not what the specific lot in your drawer will do. The only way to close that gap is to demand tested values with method references and tolerances, and to verify those values on arrival. I run a 10-piece incoming inspection on every septum lot: thickness on five points across the part, hardness on a 6 mm stack, and a compression set check on three samples. It takes 30 minutes. It has caught three bad lots in five years. That is cheap insurance against an overnight sequence that dies at 3 a.m. because a nominal number was not a real one.
The decision rule is simple. If a septa datasheet line has no tolerance, no method reference, and no lot traceability, treat it as a nominal value. Use it for screening, not for acceptance. If it has all three, it is a tested value, and you can build your incoming inspection and your autosampler parameters around it. The difference between the two is the difference between a consumable that runs 100 injections clean and one that bends needles at injection 40.
References & further reading
Related resources from our lab consumables network:
- Septa at septacaps.com
- Vial Septa at vialsepta.com
- Hplc Vial at hplcgc.com
- Autosampler Vial at autosamplerhplc.com
External references and standards:
