A customer audit in 2023 turned up a calibration certificate for a batch of 2 mL vials that cited a balance serial number, a calibration date, and nothing else. No statement of what the balance measured. No uncertainty. No link to the mass of an empty vial as opposed to the mass of the water used to check the balance. The certificate looked official. It was also worthless as evidence, because it proved a balance was calibrated, not that the fill volume in those vials was traceable to a national standard. That gap between “a certificate exists” and “the number on the label is traceable” is where ISO 17025 traceability for a 2ml vial usually breaks, and it is the subject of this article.

ISO/IEC 17025:2017 does not regulate vials. It regulates the laboratories that test them and the reports they issue. Clause 6.5 requires metrological traceability of measurement results to the SI through an unbroken chain of calibrations, each contributing a stated uncertainty. When a lab buys a 2 mL vial and reports a fill volume, a leak rate, or a septum puncture force, that report inherits whatever traceability the supplier could produce. If the supplier’s chain has a missing link, the lab’s result is not traceable, no matter how clean the certificate looks in the binder. Most labs I have audited can produce the first three or four links of that chain without effort. Almost none can produce the one that matters most.

For a dimensioned, filled consumable, an unbroken traceability chain looks like this. Each link has to name the reference standard, the method, the uncertainty, and the date.

• Link 1: the SI definition. The metre, the kilogram, the kelvin. For a vial this resolves to dimensional and mass references held by a national metrology institute (NIST in the US, NPL in the UK, PTB in Germany).
• Link 2: the working standard. A calibrated gauge block set or a certified mass set, itself calibrated against Link 1 with a stated uncertainty. Gauge blocks are usually Grade 0 or Grade 1 under ISO 3650. Mass sets are OIML Class E2 or F1.
• Link 3: the instrument. The optical comparator, the CMM, the analytical balance, the force gauge. Each has its own calibration certificate, and each certificate must reference the working standard from Link 2.
• Link 4: the measurement process. A written procedure that states the fixture, the temperature, the number of replicates, and the acceptance criterion. This is where most suppliers stop documenting.
• Link 5: the lot. The measured result for the specific lot shipped to the customer, with a sample size and a distribution, not a single pass/fail value.
• Link 6: the batch record. The link between the measurement and the physical vials in the box. Lot number, manufacture date, forming station, resin lot.
• Link 7: the material identity. The resin batch certificate, the additive package, the extractables profile. This is the link that almost nobody can produce.

Links 1 through 5 are routine for any supplier who has been through a customer audit before. Link 6 is where sloppy operations show up. Link 7 is where the chain usually dies, and I will spend most of this article on it because it is the one that most affects what your chromatography actually sees.

Link 7 is the one labs cannot produce, and it is the one that changes your chromatogram

A dimensional certificate tells you the vial is 32.0 mm tall and the ID is 12.0 mm. It says nothing about what the polymer will do at 40 °C in contact with acetonitrile for six hours. Those are different questions, and the second one has a different traceability chain: from the SI through the mass spectrometer’s calibration standards to the extractables result. A supplier who can hand you a dimensional certificate and nothing else has covered half the risk.

I have seen this fail in a way that took three weeks to diagnose. A contract lab running EPA Method 8260 by purge-and-trap started seeing a consistent 0.4 to 0.7 µg/L baseline rise in the C5 to C7 alcohol region across all blanks. The vials were the same catalog number as the previous lot. The septa were the same durometer. The dimensions were within tolerance. What had changed was the resin lot: the molder had switched to a polypropylene grade with a different slip-agent package, and the slip agent was bleeding into the headspace. The supplier’s certificate covered Links 1 through 5. It had nothing on Link 7. The fix was a resin-lot traceability requirement written into the purchase order, plus a supplier commitment to hold resin lots for the life of a customer’s validated method. That is a Link 7 control. You cannot get there from a dimensional certificate.

What Link 7 requires in practice is a certificate of analysis for the resin batch, not the molded part. It should name the resin grade, the manufacturer, the lot number, and the date of manufacture. For regulated work it should also include a statement of the additive package (antioxidant, slip agent, nucleating agent) and a mass-spec extractables screen in a defined solvent at a defined temperature. I ask for extractables in both water and acetonitrile at 40 °C for 24 hours, reported as a total ion chromatogram plus a list of any peak above 0.1 µg/mL. That is not a regulatory requirement. It is what I need to know before I sign off on a method transfer.

The same logic applies to the septum. A silicone-faced PTFE septum has its own material chain: the silicone cure system, the PTFE skive thickness, the bonding layer. A 10 mm septum with a 1.0 mm silicone layer and a 0.1 mm PTFE face is not the same barrier as a 0.2 mm PTFE face, and neither is the same as a pure PTFE/silicone laminate. If the supplier cannot name the PTFE thickness and the silicone durometer (Shore A), the chain stops at the invoice.

What a usable certificate of analysis actually contains

I have a one-page checklist I run against any new supplier before the first order. It is short, and it separates the suppliers who understand traceability from the ones who print certificates because customers ask for them.

The USP <661> reference is the one that trips people up. That chapter covers plastic components in a general sense and is a useful screen, but it is not a traceability document. It does not link your specific lot to a specific resin batch. If a supplier’s only material evidence is “USP Class VI,” you have a claim, not a chain. I say this as someone who has bought a lot of Class VI resin: the classification is real and worth having, but it is a property of the grade, not of the box in front of you. For the dimensional side, the relevant standards are ISO 11040 for the vial envelope and GPI 13-425 for the thread, and a competent supplier will name both on the drawing. For chromatographic consumables in general, the major vendors publish their own traceability documentation standards, and it is worth reading how Agilent and Thermo Fisher structure theirs before you write your own supplier requirement.

Calibration certificates that pass audit and certificates that fail

There is a specific pattern to certificates that fail a 17025 audit. The auditor does not look at whether a certificate exists. They look at whether the chain closes. Three questions do most of the work:

• Does the certificate name the standard it was calibrated against? “Calibrated against a certified mass” is not enough. It must name the mass set, its class, and its own certificate number.
• Is the uncertainty stated and is it fit for purpose? A fill-volume measurement with an uncertainty of ±0.05 mL is fine for a 2 mL nominal fill. The same uncertainty on a 100 µL residual-volume check is not. The auditor will do that arithmetic.
• Does the certificate cover the parameter you actually rely on? I have seen force gauges calibrated for tension when the puncture test is in compression. The certificate is valid, the calibration is real, and the number is meaningless for the application.

The fix is not more certificates. It is a written statement, from the supplier, of which parameter each instrument is calibrated for and what the resulting uncertainty is at the point of use. That statement is what you attach to your own method validation file. It is also what your customer’s auditor will ask for when they audit you.

Two field cases where the chain broke at a specific link

Case 1: the fill volume that drifted 1.8% across a lot

A pharma QC lab flagged a 1.8% downward drift in the reported fill volume of a 2 mL vial lot over four weeks. The vials were within the ±2% tolerance on the CoA. The lab’s own check, run on a calibrated balance, was giving a different number than the supplier’s. The measurement that identified the problem was a gravimetric check on 30 vials from a single cavity, using a NIST-traceable mass set at OIML E2. The lab’s balance was fine. The supplier’s fill nozzle had drifted, and the supplier’s own check used a balance calibrated against an internal check weight, not a certified mass. The chain broke at Link 2. The fix was a purchase-order clause requiring the supplier to calibrate the fill-station balance against an E2 mass set quarterly and to include the certificate number on the CoA. The drift stopped within one production cycle, because the supplier had never been measuring against a true reference. This is exactly the kind of gap that a 2ml vial supplier with a mature quality system will already have closed, and it is worth asking for the certificate before you place the order.

Case 2: the puncture force that passed at room temperature

An environmental lab running EPA SW-846 methods on an autosampler started getting intermittent coring after about 400 injections on a septum rated for 1,000 punctures. The septum was within dimensional spec. The puncture force was within spec at 22 °C. What the lab had not checked was the force at the autosampler’s actual operating temperature, which in that room ran 27 to 29 °C in summer. Silicone durometer drops with temperature, and the puncture force fell by roughly 15% across that range. The chain broke at Link 4: the measurement process did not match the use condition. The fix was a septum with a higher durometer (Shore A 50 instead of 45) and a written test at 30 °C. Coring stopped. The lesson is that a certificate is only traceable for the conditions it names, and the auditor will ask whether those conditions match your bench. For the general principles of septum and injection-port behavior, Restek publishes useful technical notes that are worth reading before you write your own test method.

The purchase-order language that closes the chain

You cannot audit your way to traceability after the fact. You have to specify it before the order. Here is the language I use, and I have had it accepted by suppliers on three continents:

• Dimensional: “Each lot shall be measured against ISO 3650 Grade 0 gauge blocks, n ≥ 30 vials across ≥ 3 cavities, uncertainty stated at k=2. Certificate to name the gauge block set and calibration date.”
• Fill mass: “Fill weight shall be verified gravimetrically against an OIML E2 mass set. Balance calibration certificate number to appear on the CoA.”
• Resin: “CoA shall name the resin grade, manufacturer, and lot number. Resin lots shall be held for the life of the customer’s validated method.”
• Septum: “PTFE face thickness and silicone durometer (Shore A) shall be stated. Puncture force shall be measured at 30 °C, not 22 °C.”
• Extractables: “Supplier shall provide a total ion chromatogram in water and acetonitrile at 40 °C for 24 h, with any peak above 0.1 µg/mL identified.”
• Lot linkage: “Lot number on the shipping carton shall match the CoA. CoA shall not cover more than one production lot.”

That last clause is the one suppliers push back on hardest, because it forces them to run a certificate per lot instead of per quarter. Hold the line. A certificate that covers a production period is not a traceability document, it is a marketing document. If you are sourcing through a distributor rather than the molder, the distributor needs a written agreement with the molder to pass these certificates through. A HPLC consumables distributor who cannot produce a per-lot CoA from the molder is a link in your chain that you cannot audit. For a broader view of how consumable traceability fits into a supplier qualification program, a chromatography supplier with an ISO 17025-accredited calibration lab on site is usually the safer bet, because the chain is shorter. If you are specifying a private-label vial, the same language goes into the OEM vial specification, and the molder’s own certificates become yours to hold.

Troubleshooting: where to look when the chain does not close

When a customer audit or an internal review turns up a traceability gap, work backwards from the reported number. The checklist below is ordered by how often each link fails in practice, based on the last twenty or so supplier reviews I have run.

• Start with Link 7 (material). Can the supplier name the resin lot and the septum durometer? If not, stop here. Everything downstream is unverifiable.
• Then Link 2 (working standard). Is the balance or comparator calibrated against a certified reference, or an internal check weight? The certificate number tells you.
• Then Link 4 (process). Does the test condition match the use condition? Temperature, solvent, puncture rate, torque. A 22 °C certificate for a 30 °C bench is a gap.
• Then Link 6 (lot). Does the CoA cover one lot or a period? Pull a shipping carton and compare lot numbers. They should match exactly.
• Then Link 5 (sample size). Is the result a distribution or a single value? A single value on a dimensioned part is not a measurement, it is an anecdote.
• Finally Link 3 (instrument). Is the instrument calibrated for the parameter you rely on? Force in compression, not tension. Volume at the fill temperature, not room temperature.

If you get through all six and the chain still does not close, the problem is usually at Link 1: the supplier is citing a standard they have not actually calibrated against. Ask for the certificate of the certificate. A supplier who can produce that is a supplier you can build a long relationship with.

The decision rule

Do not accept a certificate that documents the instrument and stops. Require a per-lot CoA that names the resin lot, the septum durometer, the reference standard for each measurement, the uncertainty at k=2, and the sample size. If the supplier cannot produce Link 7, the material identity, treat the vial as untraceable regardless of what the dimensional data says, and either qualify a second source or write the resin-lot hold requirement into the purchase order before the next shipment. The chain is only as strong as the link you cannot see, and for a 2 mL vial that link is almost always the polymer, not the micrometer.

References & further reading

Related resources from our lab consumables network:

• 2Ml Vial at 2mlsamplevial.com
• Hplc Consumables at hplcconsumable.com
• Chromatography Supplier at chromatographysupplier.com
• Oem Vial at oemsamplevials.com

External references and standards:

• Agilent chromatography consumables and columns
• Thermo Fisher chromatography and sample handling
• U.S. EPA SW-846 test methods for waste and materials
• Restek technical resources for chromatography