Last month I weighed 120 vials from a 48-hour evaporation stress run and found 17 of them had lost more than 1.2% of their fill mass. The rest averaged 0.3%. Same tray, same oven, same solvent. The difference came down to which closures had been torqued by hand and which had been seated with a calibrated capper. If you are chasing vial closures seal failure, evaporation and sample loss, the first measurement you need is not a pressure decay curve. It is a gravimetric map of which vials actually lost mass, and by how much, over a known interval.
The leak path is rarely a hole. It is a microchannel that opens after the needle leaves
When a closure leaks, most people picture a torn septum or a cracked cap. Those happen, but they are the easy cases. The harder case is a closure that passes a visual check, passes a dye ingress test at time zero, and still loses 0.8% of a volatile mobile phase over a 72-hour run. What you are looking at is a re-seal failure: the septum does not return to a gas-tight state after puncture because the elastomer has taken a compression set or because the cap no longer applies enough residual force to close the needle track.
A needle puncture creates a cylindrical channel through the septum. In a butyl or bromobutyl formulation, that channel is supposed to close by elastic recovery. The recovery is never perfect. At 23°C, a bromobutyl septum with a Shore A hardness of 50 will recover roughly 85 to 90% of its original sealing force within 60 seconds after needle withdrawal. The remaining 10 to 15% is permanent deformation. If the cap was under-torqued, or if the septum was already compressed beyond its working limit from a previous puncture, that residual channel stays open enough for vapor to escape at a rate you can measure on a four-place balance.
Compression set is the number that predicts re-seal behaviour, not hardness
Hardness tells you how the septum feels. Compression set tells you how much of its sealing force it gives up after being squeezed. The test is ASTM D395 Method B, run at the closure’s service temperature. A bromobutyl septum with a compression set of 18% at 70°C will re-seal better after three punctures than a silicone septum with a compression set of 35% at the same temperature, even though the silicone feels softer and more compliant in your fingers.
Here is the field rule I use. If a septum specification sheet does not list compression set, I assume the supplier has not controlled it. That means the re-seal behaviour of lot 14 can differ from lot 22 by enough to change evaporation loss by a factor of two. For any closure that will see multiple punctures from an autosampler, I require a compression set value at or below 20% after 22 hours at 70°C for bromobutyl, and at or below 30% after 22 hours at 70°C for silicone. Anything higher and I know the third puncture will not seal the same way the first one did.
Compression set also explains why a closure that passes a one-time leak test can fail on a 100-vial overnight run. The septum sits under compression for hours, not minutes. At 40°C in a heated autosampler tray, bromobutyl takes a set faster than at room temperature. The cap torque that was adequate at 22°C is no longer enough at 40°C because the septum has relaxed. If you are running a HPLC vial with a low-volatility mobile phase, you may not notice. Switch to acetonitrile with a 2 µL injection and a 12-hour sequence, and the drift shows up as retention time shifts and peak area changes that mimic column problems.
Measure evaporation loss gravimetrically before you blame the instrument
The cleanest way to separate closure loss from injection loss or detector drift is a gravimetric evaporation test. You need a balance that resolves 0.1 mg, a controlled-temperature chamber or a room with stable temperature, and a set of vials filled to a known mass with the actual solvent you use, not water. Water has a lower vapor pressure than methanol, acetonitrile, or dichloromethane, so a closure that looks fine with water will fail with acetonitrile.
My standard protocol is this. Fill each vial to 80% of nominal volume with the working solvent. Cap it with the closure you are testing. Record the mass of each vial to four decimal places. Place the vials in the chamber at your run temperature, which for an HPLC autosampler tray is often 25°C or 40°C. Weigh the vials again at 24 hours and 72 hours. Calculate percent mass loss per vial. A closure that loses less than 0.1% per 24 hours for a volatile solvent like acetonitrile is acceptable for most quantitative work. Above 0.5% per 24 hours, you will see measurable concentration drift in a long sequence.
Run the same test after one puncture, after three punctures, and after five punctures. This is the part most people skip. A closure that loses 0.05% per 24 hours unpunctured can lose 0.4% per 24 hours after three punctures. That is the re-seal failure showing up as a number. If you only test unpunctured vials, you will approve a closure that fails in the field.
| Closure type | Material | Compression set at 70°C | Re-seal after 3 punctures | Typical 72h loss, acetonitrile, 40°C |
|---|---|---|---|---|
| 9mm screw cap with bonded septum | Bromobutyl, PTFE-faced | 16-20% | Good | 0.08-0.25% |
| 11mm crimp cap with septum | Bromobutyl, PTFE-faced | 18-22% | Good to fair | 0.15-0.40% |
| 11mm snap cap with septum | Silicone, PTFE-faced | 28-35% | Fair to poor | 0.40-1.10% |
| 13mm screw cap with septum | Natural rubber, PTFE-faced | 25-40% | Poor | 0.60-1.50% |
| 9mm screw cap with pre-slit septum | Silicone, PTFE-faced | 30-38% | Poor after 2 punctures | 0.50-1.20% |
The numbers in that table are from my own testing and from supplier data I have collected over the years. They are not universal. But the pattern holds: bromobutyl re-seals better than silicone after multiple punctures, and pre-slit septa give up sealing force faster than solid septa when punctured repeatedly at the same location. If you are buying vial closures for a method that runs overnight sequences, the compression set and re-seal numbers matter more than the unit price.
Two field cases where the closure was the last thing anyone checked
Case one: a contract lab running EPA Method 8260 purge-and-trap volatiles on a HPLC vial platform started seeing internal standard responses drift downward by 12% over a 40-vial sequence. The column was new, the purge flow was stable, and the internal standard solution was freshly prepared. The lab manager suspected the autosampler syringe. I asked him to weigh ten vials before and after a 24-hour period at room temperature, capped but unpunctured. Average loss was 0.09%. Then I asked him to weigh ten vials after the autosampler had punctured each one three times. Average loss jumped to 0.52%. The septum was a silicone formulation with a compression set of 34%. The fix was switching to a bromobutyl septum with a PTFE face and a compression set of 19%, and the internal standard drift dropped to under 3% over the same sequence length.
Case two: a pharmaceutical QC lab noticed that vials of a methanol-based standard lost mass unevenly across a tray. The tray had 100 positions. The vials in the back row, closest to the autosampler heater, lost twice as much as the vials in the front row. The closures were 11mm crimp caps applied with a manual crimper. The operator had been crimping by feel, and the crimp force varied from 45 N to 85 N across the tray. The vials with the lowest crimp force were the ones losing the most mass. The fix was not a different septum. It was a calibrated crimping tool set to 70 N ± 5 N and a torque audit every 20 vials. Evaporation loss evened out to within 0.1% across the tray.
Cap torque and crimp force are part of the sealing system, not a packaging detail
A screw cap closure seals because the cap compresses the septum between the vial lip and the cap’s inner shoulder. The compression creates a radial contact stress that blocks vapor from escaping along the glass-septum interface. If the cap is under-torqued, that contact stress is too low and the leak path runs along the glass rim, not through the septum. If the cap is over-torqued, you can crack the glass or extrude the septum into the cap threads, which reduces the sealing area.
For a 9mm screw cap on a 12x32mm vial, the working torque range is 0.6 to 0.9 N·m. Below 0.5 N·m, I see evaporation loss increase by a factor of three or more with acetonitrile. Above 1.1 N·m, I start seeing stress cracks in the vial neck, especially with Type 33 expansion glass. For an 11mm crimp cap, the target is 70 to 80 N of crimp force. A manual crimper with no force gauge will drift over a day of use. I have measured a 30% drop in crimp force over 200 crimps as the operator fatigues.
If you are sourcing chromatography consumables for a regulated method, the closure torque specification should be part of the method documentation, not a note in the lab notebook. USP chapter 381 covers elastomeric closures for injections, but it does not set a torque value for your specific vial and cap combination. That is your job. The U.S. EPA SW-846 test methods for volatiles, for example, require vials to be sealed in a way that prevents loss of analytes, but they leave the torque setting to the lab. The right number comes from gravimetric testing, not from a catalog.
The PTFE face is a barrier, not a seal. Do not confuse the two
A PTFE-faced septum has a thin fluoropolymer layer on the side that contacts the sample. That layer is there to prevent the elastomer from leaching extractables into the solvent. It is not there to re-seal after puncture. PTFE does not have elastomeric recovery. Once a needle pierces the PTFE face, the hole in the fluoropolymer layer stays open. The re-seal comes from the elastomer underneath pushing back against the hole. If the elastomer has a high compression set, the PTFE face will not save you.
This is why a pre-slit PTFE-faced silicone septum can be a disaster for volatile samples. The slit in the PTFE face is a permanent opening. The silicone underneath is supposed to close it, but silicone has a high compression set and poor gas barrier properties compared to bromobutyl. A pre-slit silicone septum on a 9mm screw cap can lose 0.8% of an acetonitrile fill in 72 hours at 40°C even without any punctures, because the slit itself is a leak path. For volatile methods, I use solid bromobutyl septa with a PTFE face, not pre-slit septa, unless the autosampler needle is too fragile to pierce a solid septum repeatedly.
A troubleshooting checklist for evaporation loss traced to the closure
When a run shows concentration drift, retention time shifts, or internal standard loss, and you have already ruled out the column, the injector, and the detector, work through this list before you start changing methods.
- Weigh the vials. Fill ten vials with the working solvent, cap them, weigh them, and re-weigh after 24 hours at run temperature. If average loss exceeds 0.1% per 24 hours, the closure is the prime suspect.
- Weigh after punctures. Re-weigh the same vials after one, three, and five needle punctures. A jump in loss rate after puncture means re-seal failure, not a cap problem.
- Check the torque. Use a calibrated torque meter on screw caps. If the applied torque is below 0.6 N·m for a 9mm cap, re-torque and re-weigh. For crimp caps, use a force gauge and verify 70 to 80 N.
- Check the septum material. If the septum is silicone or natural rubber, check the compression set spec. Above 30% at 70°C, expect re-seal problems after multiple punctures.
- Check the PTFE face. If the septum is pre-slit, test a solid septum of the same elastomer. If the loss drops, the slit is the leak path.
- Check the vial lip. A chipped or uneven vial lip will prevent the septum from seating evenly. Run a finger around the lip or inspect under magnification. Replace any vial with a visible defect.
- Check the temperature. If the autosampler tray is heated, the closure’s compression set accelerates. A closure that passes at 25°C may fail at 40°C. Test at the actual run temperature.
For a deeper look at how closure materials behave under analytical conditions, the Agilent chromatography consumables and columns resource has application notes on closure selection for GC and HPLC, and Thermo Fisher chromatography and sample handling publishes compatibility data for septa and caps. Restek technical resources for chromatography also has a useful reference on septum bleed and re-seal behaviour. For independent verification of closure performance, analytical testing labs can run compression set, dye ingress, and gravimetric loss tests on your actual vials and closures.
The decision rule: buy closures with a compression set spec and verify it with a balance
If you only remember one thing from this article, make it this: a closure that has not been tested for compression set and re-seal behaviour after multiple punctures is an unknown variable in your method. The fix is not to buy the most expensive septum. It is to require a compression set value below 20% for bromobutyl or below 30% for silicone, to torque screw caps to 0.6 to 0.9 N·m or crimp to 70 to 80 N, and to verify evaporation loss gravimetrically with your actual solvent at your actual run temperature, before and after punctures. Do that once, document the numbers, and the next time a sequence drifts, you will know within an hour whether the closure is the cause or whether you need to look elsewhere.
References & further reading
Related resources from our lab consumables network:
- Vial Closures at vialsclosures.com
- Analytical Testing at testinganalytical.com
- Hplc Vial at vialforhplc.com
- Chromatography Consumables at chromatographyconsumables.com
External references and standards: