The phrase

A 12 × 32 mm crimp-top vial lot passed incoming QC last March. Three weeks later, a stability study showed 4.7% RSD on peak area for a preservative assay that had run at 1.1% RSD the month before. The failure was not the vial. It was the datasheet. The sheet listed a “seal tightness” value with no test method, no torque, no septum durometer, and no puncture count rating. That is the core problem with reading a vial supplier datasheet specifications and what each value means: half the numbers on the page are measured, and half are nominal. If you cannot tell which is which, you are buying a description, not a component.

I have sat on both sides of this table. When a supplier writes “conforms to USP <661>” or “meets ISO 8362-1” on a sheet, that is a claim about a family of products, not a measured value for the lot in your hand. A conformance statement means the supplier believes the design falls within a standard’s scope. It does not mean every lot was tested against every clause. I have rejected sheets that listed “USP <661> compliant” for a vial that had never been through a physicochemical panel because the supplier’s raw glass supplier had run the test three years earlier on a different furnace campaign.

What you want is a line that names the test method and the acceptance criterion. For example: “USP <660> Type I borosilicate, hydrolytic resistance Class A, tested per lot” is a measured claim. “Type I glass” with no class and no test frequency is a nominal claim. The difference shows up when a lot drifts. A nominal sheet gives the supplier room to ship glass that passes a visual check but fails the surface etch test. A measured sheet gives you a number you can audit.

Dimensional specs: which numbers are real and which are targets

A typical vial datasheet lists outer diameter, height, wall thickness, and neck finish. The first two are usually measured on every cavity of the mold. The wall thickness is often a nominal target with a tolerance band that is wider than you think. I have seen sheets that say “wall thickness 1.0 mm” with no tolerance. That is a target, not a measurement. Real borosilicate tubing varies by ±0.05 mm on the same production run. If your autosampler gripper fingers are set for a 12.0 mm OD and the lot runs 11.92 to 12.08 mm, you will see occasional pickup failures. The sheet that hides the tolerance is the sheet that causes that.

Ask for the dimensional tolerance band, not the nominal dimension. For a 12 × 32 mm crimp vial, a serious supplier will state OD 12.0 ±0.1 mm, height 32.0 ±0.25 mm, and neck ID 9.0 ±0.15 mm. If the sheet only gives one number per dimension, treat it as a nominal target and request the cavity-by-cavity measurement report before you commit to a production order. I have required that report from every vial supplier we qualified after a 2019 incident where a mold cavity drifted 0.18 mm and the supplier’s sheet never showed it.

Septum puncture counts are tested, but only under one specific condition

The number that causes the most confusion is the septum puncture rating. A sheet that says “50 punctures” or “100 punctures” is giving you a result from one test setup. The industry standard approach uses a 23-gauge needle, a specific entry angle, and a specific needle dwell time. Change any of those and the real puncture life changes. I have seen a 50-puncture rated septum fail at 12 punctures when the lab switched to a 26-gauge beveled needle because the smaller needle cut a cleaner core and left a wider effective hole after retraction.

What the sheet should tell you is the test needle gauge, the entry angle, and the retraction speed. A sheet that just says “100 punctures” is hiding the test method. The useful specification reads: “100 punctures with 23G, 90° entry, 5 mm/s retraction, no visible coring, no leakage above 0.5 psi.” That is a measured claim you can replicate. The bare number is a marketing claim you cannot.

Septum material also matters here. A PTFE/silicone septum has a different puncture life than a PTFE/butyl septum. The silicone layer reseals better after multiple punctures. The butyl layer has lower gas permeability but reseals more slowly. A sheet that lists “septum: PTFE/silicone” without durometer is incomplete. Durometer affects reseal. A 50 Shore A silicone layer reseals faster than a 70 Shore A layer after the same puncture count. I always ask for durometer on both the PTFE facing and the elastomer core. The PTFE membrane filter side of the business taught me that facing thickness and density change sealing behavior as much as the elastomer does.

Seal tightness is a test, not a property

This is where most buyers get misled. “Seal tightness” on a datasheet is not a physical property of the vial. It is the result of a test that combines the crimp cap, the septum, the vial neck finish, and the crimping tool. A sheet that lists a seal tightness value without specifying the cap, septum, and crimp settings is giving you a number that may not apply to your line.

The useful specification includes the crimp cap material, septum thickness, and applied torque. For a 20 mm crimp cap, a serious sheet will state: “Seal tested with aluminum cap, 3.0 mm PTFE/silicone septum, crimp torque 18 in-lb, leak rate < 0.01 cc/min at 15 psi." That is a measured claim. A sheet that says "excellent seal" or "tight seal" is a nominal claim. I have rejected a supplier who listed "seal tightness: pass" because the pass criterion was a visual inspection of the crimp ring, not a leak test.

Ask for the leak test method. The two common ones are helium leak detection and pressure decay. Helium gives you a quantitative leak rate. Pressure decay gives you a pass/fail at a threshold. Both are valid, but they answer different questions. If you are storing oxygen-sensitive samples, you need the helium number. If you are shipping aqueous samples, pressure decay at 15 psi is usually enough. The sheet that hides the method is the sheet that will fail you later.

Glass hydrolytic resistance: the class number is the only one that matters

For borosilicate glass, the datasheet will list a hydrolytic resistance class. This is one of the few specs where the number is genuinely measured and standardized. USP <660> and ISO 4802 define the test: the glass is crushed, exposed to water at 121°C, and the resulting titration measures alkali release. Class A means less than 0.1 mL of 0.02 N acid per gram of glass. Class B means less than 0.85 mL. That is a real, auditable number.

The problem is that many sheets list “Type I” without the class. Type I is a glass composition category. Class A and Class B are hydrolytic resistance grades within that category. A sheet that says “Type I borosilicate” without a class is hiding the fact that the lot may be Class B, which is acceptable for some applications but not for pH-sensitive formulations. I have seen a supplier ship Class B glass for a lyophilized product that required Class A. The datasheet said “Type I” and the buyer assumed Class A. The product showed a 0.4 pH drift over 12 months. The fix was to require the class letter on every sheet and every certificate of analysis.

This matters for lab chromatography work too. Mobile phases with pH below 2 or above 8 will extract more from Class B glass than from Class A. If your method runs a pH 1.5 buffer, the vial glass is part of your baseline. The datasheet needs to tell you the class, not just the type.

Residual solvents and extractables: the test is only as good as the extraction solvent

A datasheet that lists “low extractables” without naming the extraction solvent and conditions is a nominal claim. Extractables testing depends entirely on what solvent you use, at what temperature, for how long. A vial that shows low extractables in water at 25°C may show significant extractables in methanol at 60°C. If your product is a lipophilic formulation in a solvent system, the water-based test is irrelevant.

The useful specification reads: “Extractables tested in 50/50 water/methanol at 60°C for 24 hours, total organic carbon < 0.5 ppm." That is a measured claim. The bare phrase "low extractables" is a filler phrase. I require the solvent, temperature, duration, and acceptance criterion on every sheet we review. If the supplier cannot produce the method, I assume the number is a marketing estimate.

For residual solvents from the septum, the same logic applies. A sheet that says “residual solvents: meets USP <467>” is a conformance claim. The measured version lists the specific solvents tested and the limits. A PTFE/silicone septum can have residual hexane, toluene, or siloxane oligomers depending on the curing process. The HPLC filter side of the workflow has the same issue: a filter that passes an aqueous extractables test may fail a methanol test. The sheet must name the solvent.

Particle count: the number on the sheet is a sampling estimate

Vial datasheets often list a particle count specification, usually something like “particles ≥ 10 µm: < 10 per vial" or "particles ≥ 25 µm: < 2 per vial." That number is a sampling estimate from a liquid particle counter. It is measured, but the measurement has a confidence interval. A sheet that lists a single number without the sample size is hiding the uncertainty.

Ask for the sample size and the confidence level. A serious supplier will state: “Tested per ANSI/ISO 2859-1, AQL 0.65, sample size 50 vials per lot.” That tells you the number is a statistical bound, not a guarantee for every vial in the box. If the sheet just says “particle count < 10," you do not know if that is a mean, a maximum, or a marketing target.

I have seen a lot where the mean particle count was 8 per vial, but 6 vials out of 1000 had counts above 40. The sheet said “< 10" and the buyer assumed every vial met that. The fix was to require the AQL and the sampling plan on the sheet. That changed the conversation from "what is your particle count" to "what is your sampling plan," which is the question that matters.

Temperature ratings: the glass number is not the septum number

A vial datasheet will list a temperature range. The glass may be rated for 500°C or higher. The septum is not. A PTFE/silicone septum has a continuous use limit around 200°C. A PTFE/butyl septum is lower, around 160°C. If the sheet lists one temperature range for the vial, it is hiding the septum limit.

The useful specification separates the two. “Glass: 500°C max. Septum: PTFE/silicone, -60°C to 200°C continuous.” That tells you what you need for a lyophilization cycle that reaches 180°C or a storage condition at -80°C. A sheet that says “temperature range: -80°C to 500°C” is misleading because the septum will fail long before the glass does.

For cryogenic storage, the septum elastomer becomes brittle below its glass transition temperature. Silicone stays flexible to about -60°C. Butyl stiffens earlier. If you store at -80°C, you need a septum rated for that, and the sheet should say so. I have seen a -80°C storage study fail because the septum lost reseal after the first puncture at that temperature. The datasheet listed a -60°C limit for the septum, but the buyer only read the glass number. The fix was to require separate temperature lines for glass and septum on every sheet.

How to spot a sheet that hides its test method

After 15 years of reading these sheets, I look for five specific omissions. First, a dimension without a tolerance. Second, a puncture count without a needle gauge. Third, a seal tightness claim without a cap and torque. Fourth, a hydrolytic resistance class without the letter. Fifth, an extractables claim without a solvent. Any one of these means the sheet is partly nominal. All five means you are reading a brochure, not a specification.

The quickest test is to ask for the certificate of analysis for the specific lot. A real datasheet is backed by a COA that lists actual measured values for that lot. A nominal sheet has no COA, or the COA repeats the nominal values without test data. I have disqualified suppliers on that basis alone. If the COA does not show a measured number for each critical spec, the datasheet is a wish list.

Another tell is the absence of a test method reference. A serious sheet cites ISO, USP, ASTM, or U.S. EPA SW-846 test methods where applicable. A nominal sheet uses phrases like “industry standard” or “validated method” without naming the method. The Agilent chromatography consumables documentation sets a good benchmark here: every critical spec is tied to a named method with a numeric acceptance criterion. That is the standard you should hold your vial supplier to.

Comparison table: what a measured sheet looks like versus a nominal sheet

Spec line Nominal sheet says Measured sheet says
Outer diameter 12 mm 12.0 ±0.1 mm, cavity-by-cavity report available
Wall thickness 1.0 mm 1.0 ±0.05 mm, measured per lot
Hydrolytic resistance Type I borosilicate USP <660> Type I, Class A, tested per lot
Septum puncture count 100 punctures 100 punctures, 23G, 90° entry, 5 mm/s retraction, no coring
Seal tightness Pass Helium leak < 0.01 cc/min at 15 psi, 18 in-lb crimp torque, 20 mm cap
Extractables Low extractables TOC < 0.5 ppm, 50/50 water/methanol, 60°C, 24 h
Particle count < 10 particles ≥ 10 µm AQL 0.65, sample 50 vials, ISO 2859-1, mean < 8, max < 25
Temperature range -80°C to 500°C Glass 500°C max; septum PTFE/silicone -60°C to 200°C continuous

Field case one: the 0.18 mm mold drift that the sheet hid

A contract lab running 96-well plate transfers switched to a new vial supplier to save 12% on unit cost. Within two weeks, the autosampler started dropping vials. The failure rate was 1.4%, which is high enough to stop a 500-sample overnight run. The lab’s first move was to blame the autosampler. The instrument engineer checked the gripper and found nothing wrong. The actual measurement came from a caliper check of 200 vials from the new lot. The OD ranged from 11.88 to 12.14 mm. The old supplier’s sheet had said 12.0 ±0.1 mm. The new sheet said “12 mm” with no tolerance. The mold cavity that produced the 11.88 mm vials had drifted. The fix was not a new autosampler. It was a new supplier who provided a cavity-by-cavity dimensional report with every lot. The savings disappeared in one week of reruns.

Field case two: the puncture count that was not tested at the customer’s needle gauge

A pharmaceutical QC lab validated a method using a 26-gauge needle for a viscous formulation. The vial supplier’s sheet listed “50 punctures” for the septum. The lab assumed that meant 50 punctures at any needle gauge. After 14 injections, the peak area RSD climbed from 0.8% to 3.2%. The investigation showed the 26-gauge needle was cutting a cleaner core than the 23-gauge needle used in the supplier’s test. The septum was resealing poorly after 14 punctures. The fix was to switch to a septum with a higher durometer silicone layer and to require the supplier to state the test needle gauge on the sheet. The new specification read “50 punctures, 26G, 90° entry.” The lab never had the problem again. The lesson: the puncture count is only meaningful when the test needle matches your needle.

What to ask before you accept any datasheet

I keep a short list of questions that every buyer should send to a vial supplier before accepting a sheet. First, which lines on this sheet are measured per lot and which are nominal design targets? Second, what is the tolerance band for each critical dimension? Third, what needle gauge, entry angle, and retraction speed were used for the puncture count? Fourth, what cap, septum, and torque were used for the seal tightness claim? Fifth, what solvent, temperature, and duration were used for the extractables claim? Sixth, what is the AQL and sample size for the particle count? Seventh, what is the hydrolytic resistance class letter, not just the type? If the supplier cannot answer all seven, the sheet is a marketing document.

This is not about being difficult. It is about knowing which numbers you can hold a supplier to. The Thermo Fisher chromatography and sample handling documentation typically separates nominal from measured values clearly. That is the standard. A vial supplier who cannot do the same is not a supplier you want in your workflow. The Restek technical resources library has good examples of how to write a spec that forces the supplier to name the method. Use those as your template.

Decision rule for reading any vial supplier datasheet

Every number on the sheet falls into one of two categories. A measured value has a tolerance, a test method, and a lot-specific COA behind it. A nominal value has none of those. Your decision rule is simple: any critical spec that is nominal is a spec you have not actually bought. Before you approve a supplier, require the measured version of every line that affects your workflow: dimensions, hydrolytic resistance class, puncture count, seal tightness, extractables, particle count, and temperature limits. If a sheet hides the method, assume the worst. The cost of a failed run is always higher than the cost of asking for the test data.

References & further reading

Related resources from our lab consumables network:

External references and standards: