A customer sent me two datasheets for the same 250 µL glass insert last March. One listed “capacity: 250 µL.” The other listed “capacity: 250 µL (nominal), verified fill volume 210 µL at 6 mm drawdown.” Same part number family, same outer diameter, same price within 4 cents. The first sheet told me nothing I could verify. The second told me exactly where the needle would sit when the autosampler ran dry. That gap, between a number that was measured and a number that was typed, is the whole game when you read a vial insert datasheet specifications sheet. Most buyers never see it because they read the top of the sheet and stop.
I have been on both sides of this document. I have written spec sheets for insert lots and I have rejected supplier sheets that could not survive a five-minute interrogation. What follows is how I read one now, line by line, and which lines I treat as promises versus decoration.
Every number on an insert datasheet falls into one of three buckets, and the sheet should label which bucket each value sits in. If it does not, assume the worst bucket.
Guaranteed (or “spec,” “min,” “max,” “≤,” “≥”) means the supplier will replace the lot if a measured part falls outside the stated window. This is a contractual number. It should be tied to a sampling plan and an AQL. When I see “OD: 6.00 ± 0.10 mm” with no qualifier, I ask what the AQL is and how many pieces per lot get measured. A guaranteed tolerance with no sampling statement is a wish.
Typical (or “nom,” “average,” “typical value”) means the middle of the distribution, with no promise about the tails. A “typical” wall thickness of 0.45 mm tells you the mean. It tells you nothing about the 3-sigma low side, which is the number that decides whether the insert survives a 2,000 g centrifuge spin or a crimper jaw.
Nominal is the design intent. It is the number on the drawing, not the number on the part. A “nominal 250 µL” capacity is the geometric volume of the drawn glass envelope. The usable volume is smaller because the cone at the bottom and the meniscus at the top eat into it. When a sheet says “capacity: 250 µL” with no method, it is almost always nominal, and the real working volume is 200 to 220 µL depending on where your needle draws.
Here is the decision table I use. Print it, tape it to the bench, and hold the next datasheet against it.
| Datasheet line | What it usually is | What it should be | What to ask |
|---|---|---|---|
| Capacity (µL) | Nominal, geometric | Verified fill volume at stated drawdown | “At what needle depth and what residual volume?” |
| Outer diameter (mm) | Typical, mean of a run | Guaranteed ± tolerance with AQL | “What is the sampling plan and the AQL?” |
| Wall thickness (mm) | Typical | Min value, guaranteed | “What is the 3-sigma low side?” |
| Bottom cone angle | Nominal drawing value | Guaranteed range | “How is it measured, optically or by gauge?” |
| Chemical inertness | Marketing claim | Test method + acceptance criterion | “Which USP chapter, which extraction condition?” |
| Puncture resistance | Rarely stated | Count to failure, method named | “Needle gauge, septum, force, pass criterion?” |
| Temperature rating | Often absent | Continuous and excursion limits | “Is that glass transition or practical use?” |
Outer diameter is the one number that must be guaranteed, and usually is not
OD is the line I read first because it is the line that decides whether the insert drops into the 2 mL vial without jamming and whether the autosampler gripper can lift it. A 6.00 mm nominal insert in a 6.20 mm ID vial leaves 0.10 mm of radial clearance per side. That is fine. A 6.10 mm insert in the same vial leaves 0.05 mm per side, and now a single out-of-round piece at the high end of the distribution will wedge.
When a sheet gives “OD: 6.0 mm” with one decimal and no tolerance, I assume it is a nominal. Ask for the guaranteed window. A serious sheet will say something like 6.00 ± 0.05 mm, 100% gauge-checked, AQL 1.0. A sheet that says “6.0 mm” and nothing else is telling you the drawing, not the part.
How OD is actually measured
There are three methods and they give three different answers. A go/no-go ring gauge tells you pass or fail against a hard limit. A laser micrometer tells you the diameter at one angular position. An optical comparator tells you the profile. If the sheet does not name the method, the number is unverifiable. I have seen lots that passed a two-point micrometer check at 6.02 mm and failed a ring gauge because the piece was oval by 0.08 mm. The micrometer missed it because it measured across the long axis. This is why I ask for the method, not just the number.
Wall thickness is where “typical” quietly becomes a liability
Wall thickness controls two things: puncture resistance and thermal mass. A thicker wall resists needle damage but takes longer to equilibrate in a temperature-controlled autosampler tray. A thinner wall equilibrates fast but cracks at the cone when a needle deflects off-center.
Most sheets list wall thickness as a typical value, say 0.45 mm. That is the mean. The number that matters is the low side. If the distribution runs 0.38 to 0.52 mm, the thin pieces are the ones that fail first under repeated puncture. I want the sheet to state a minimum, not a typical. If it only gives a typical, I ask for the Cpk or the 3-sigma low limit. A supplier who cannot produce that number has not characterized the process.
Capacity claims need a drawdown depth or they are fiction
This is the most common misleading line on an insert datasheet. “Capacity: 250 µL” is a geometric volume. The volume your autosampler can actually aspirate depends on three things: the cone geometry at the bottom, the needle depth setting, and the residual volume you are willing to leave behind.
A 250 µL insert with a 6 mm drawdown might deliver 210 µL of usable sample. Set the needle to 2 mm from the bottom and you get 235 µL but risk hitting the cone wall. Set it to 10 mm and you get 180 µL with a clean draw. The sheet should state the drawdown depth alongside the capacity. If it does not, the number is decorative.
I once audited a sheet that claimed “300 µL capacity” for a 6 x 29 mm insert. The geometric volume of that envelope is about 300 µL if you fill it to the rim. Usable volume at a 5 mm drawdown was 245 µL. The buyer had sized a method around 280 µL injections and could not understand why the last 40 injections of every tray ran dry. The fix was a larger insert, not a different needle. That is a datasheet failure, not a method failure.
Chemical inertness lines are almost always untested marketing
You will see “chemically inert borosilicate glass” on nearly every sheet. Borosilicate is inert to most HPLC solvents, but the claim as written is not a test result. The tested version of that line names a method and an acceptance criterion. The one I trust reads something like: USP <661.1> Type I glass, hydrolytic resistance confirmed per lot, extraction test per USP <1663> at 40 °C for 24 h, no peak above 0.5 mAU at 210 nm.
If the sheet cites a chapter, ask which one and what the acceptance criterion was. A sheet that says “USP compliant” without a chapter number is using the standard as a logo. The real standards that reach an insert are narrow, and a supplier who knows them will name them precisely.
The puncture resistance line is missing from most sheets, and it is the one I want
Inserts fail mechanically more often than they fail chemically. The failure mode is needle puncture at the cone. A septum-rated needle punctures the insert bottom thousands of times across a tray. The glass either survives or it does not, and the datasheet should tell you which.
A useful puncture line reads: ≥ 500 punctures with a 22-gauge blunt needle at 5 mm/s, 2 N force, no through-crack at 50x magnification. That is a test with a needle gauge, a speed, a force, and a pass criterion. A sheet that says “high puncture resistance” without those four parameters is telling you nothing.
The number varies enormously by wall thickness and cone geometry. I have seen 0.40 mm walls fail at 180 punctures and 0.50 mm walls survive 900. If your method runs 96-well trays with a single needle, 180 is fine. If you run a 2,000-injection sequence on a multi-draw autosampler, 180 is a mid-run failure waiting to happen. Ask for the count. If the supplier cannot give it, ask them to run it. A competent glass supplier can produce that number in a day.
Temperature ratings hide a glass transition behind a practical limit
Most insert sheets either omit temperature or state a single number like “up to 200 °C.” That number is usually the practical use limit, not a material property. Borosilicate glass softens around 820 °C, but the practical limit for an insert in an autosampler is set by the septum, the crimp, and the thermal expansion mismatch with the vial, not the glass.
What I want on the sheet is two numbers: a continuous use limit and a short-excursion limit, each with a duration. Something like continuous 150 °C, excursion to 200 °C for 30 min, no dimensional change beyond 0.02 mm. That tells me the insert survives a hot injection cycle without cracking at the cone. A single “200 °C” with no duration is a guess.
Spotting a sheet that hides its test method
There are five tells. I check for all five before I approve a new supplier.
- No method column. If the sheet has a value column but no method column, every number is nominal until proven otherwise.
- One-decimal dimensions with no tolerance. “6.0 mm,” “0.4 mm,” “29 mm” are drawing values. A tested sheet carries tolerances.
- Standards cited without chapter numbers. “USP compliant” and “ISO certified” are logos. “USP <661.1> Type I” and “ISO 4796-1” are citations.
- No sampling plan. A guaranteed tolerance with no AQL and no sample size is a claim without a mechanism. Ask how many pieces per lot get measured.
- No lot traceability. If the sheet cannot be tied to a lot number and a certificate of analysis, the values are marketing copy, not quality data.
I also cross-check against the supplier’s own HPLC consumables catalog. If the catalog and the datasheet disagree on OD or capacity, the datasheet is the one that gets audited, and the disagreement itself is a finding.
Two field cases where the sheet and the part disagreed
Case 1: the 0.08 mm oval that jammed a gripper
A pharma lab on a 1,500-injection-per-week method started getting “insert not found” errors on one autosampler arm, about 12 per 1,000 injections. The datasheet said OD 6.00 ± 0.05 mm. We pulled 50 inserts from the suspect lot and measured each one at four angular positions with a laser micrometer. The mean was 6.01 mm, within spec. But 7 of the 50 measured 6.06 mm across one axis and 5.98 mm across the other, an ovality of 0.08 mm. The two-point micrometer check the supplier used measured only the long axis, which passed. The gripper, which grips on the short axis, saw a 6.06 mm piece and slipped. The fix was a change to a 100% ring-gauge check at the supplier and a tighter ovality spec of 0.03 mm. Error rate dropped to 0.4 per 1,000. The datasheet had not been lying about the tolerance. It had been hiding the measurement method.
Case 2: the “250 µL” insert that held 214
A contract lab validated a method on a 250 µL insert and sized the injection volume at 200 µL with a 5 mm drawdown. Halfway through a 2,000-injection sequence, the autosampler began flagging short draws. We measured the actual usable volume on 20 inserts from the lot by gravimetric fill and drawdown simulation. Mean usable volume was 214 µL, not 250. The cone geometry had drifted by about 0.3 mm in depth across the lot, which ate 12 µL of usable volume. The sheet’s “250 µL” was nominal, and the supplier had never stated a verified fill volume. The fix was a new spec line: verified fill volume ≥ 240 µL at 5 mm drawdown, gravimetric, per lot. The supplier could meet it. They just had not been asked to.
What a datasheet should say before you sign a supply agreement
Here is the checklist I put into a supply agreement now. If a supplier cannot meet all of it, the sheet is not ready to govern a production method.
- OD: guaranteed tolerance (e.g., 6.00 ± 0.05 mm), 100% gauge-checked, AQL 1.0, ovality ≤ 0.03 mm.
- Wall thickness: guaranteed minimum (e.g., ≥ 0.42 mm), method named.
- Capacity: verified fill volume at a stated drawdown depth (e.g., ≥ 240 µL at 5 mm), gravimetric, per lot.
- Puncture resistance: count to failure with needle gauge, speed, and force named (e.g., ≥ 500 punctures, 22-gauge blunt, 5 mm/s, 2 N).
- Chemical: USP <661.1> Type I, extraction per USP <1663> at 40 °C/24 h, no peak > 0.5 mAU at 210 nm.
- Temperature: continuous and excursion limits with durations (e.g., 150 °C continuous, 200 °C for 30 min).
- Traceability: lot number, CoA, and retained samples on every shipment.
For method context, the extraction conditions in EPA SW-846 are a useful cross-check on whether a supplier’s inertness claim survives a real solvent matrix, and the technical notes from Restek and Agilent’s consumables documentation show how the major instrument makers word their own insert specs, which is a fair benchmark for what a supplier should be able to state. Thermo Fisher’s sample handling pages are another reference point when you need to compare a supplier’s claimed OD against an instrument maker’s stated vial cavity.
When you are sourcing, the vial insert page on this site is where the dimensional and material questions live, and a chromatography supplier who can produce a method-backed sheet will usually do so on request. The ones who cannot are the ones whose sheets you should stop reading.
The decision rule
Read the datasheet from the bottom up, not the top down. Start with the traceability and method statements. If there is no method column, no sampling plan, and no lot-linked CoA, the numbers above them are nominal and you should price the part accordingly, which usually means it is not cheap enough to justify the risk. If the method column is there, verify that the three lines that actually govern your method (OD with ovality, verified fill volume at your drawdown depth, and puncture count at your needle gauge) are guaranteed values with stated test conditions. Everything else on the sheet is context. Those three are the contract.
References & further reading
Related resources from our lab consumables network:
- Vial Insert at hplcinsert.com
- 2Ml Vial at 2mlsamplevial.com
- Hplc Consumables at hplcconsumable.com
- Chromatography Supplier at chromatographysupplier.com
External references and standards: