I have sat through two ISO 17025 assessments where the assessor spent more time on the vial shelf than on the balance. Both times the finding was the same: the laboratory could trace the reference standard to NIST, but could not trace the vial that held it. The certificate of analysis for the glass was in a filing cabinet, undated, issued against a purchase order number that no longer existed in the ERP. That is a nonconformity under clause 6.5, and it is one of the easiest to avoid and one of the most common to earn.
This article is about what ISO 17025 actually asks of a vial, how traceability chains break in real laboratories, and what documentation you need in place before an assessor asks. I will use the clause numbers, the calibration hierarchy, and the specific records that survive scrutiny.
ISO/IEC 17025:2017 clause 6.5 is titled “Metrological traceability.” The requirement is that measurement results are traceable to the SI through an unbroken chain of calibrations, each contributing to the measurement uncertainty. The clause does not name vials. It does not name septa, caps, or any other consumable. What it does is establish a principle: anything that influences a measured value has to be accounted for in the traceability chain or excluded from it by evidence.
The question an assessor will ask is not “is your vial traceable to NIST?” Nobody calibrates a vial against a national standard. The question is whether the vial’s contribution to the measurement is characterized well enough that the laboratory can defend the uncertainty budget. For a 2 mL autosampler vial in a validated HPLC method, the vial contributes to uncertainty through four channels: extractables that co-elute with the analyte, adsorption loss on the glass surface, volume delivery from fill height, and septum coring that introduces particulates into the needle.
Clause 6.5.2 allows for “documented unbroken chain of calibrations” or, where that is not possible, “other means” including “certified reference materials” and “specified requirements.” That last phrase is the door through which vial specifications enter the system. A dimensional specification on a molded vial, a hydrolytic resistance class under USP <660>, a septum formulation identified by compound and cure, and a lot number that ties back to a manufacturing record are the “specified requirements.” They are not traceable in the metrological sense. They are traceable in the documentary sense, and that is what the assessor will test.
The distinction between calibration traceability and documentary traceability
I make this distinction explicit in our internal SOP because it saves arguments. Calibration traceability means: the pipette that filled the vial was calibrated against a balance, the balance was calibrated against NIST-traceable weights, the weights have a certificate with a stated uncertainty. Documentary traceability means: the vial itself can be identified to a manufacturer, a lot, a specification, and a date. The first is metrological. The second is what clause 7.5 (technical records) and clause 6.5.2 together demand for a component that touches the sample but is not itself a measuring instrument.
If you cannot produce a lot number for the vial that held the sample in a reportable result, you have a gap. It is not a gap the assessor will always find. It is a gap that will cost you the finding when they do.
Where the traceability chain breaks in a working lab
I have audited dozens of laboratories on consumable records. The chain breaks in the same five places almost every time. None of them are exotic. All of them are fixable in a week.
Break 1: the receiving inspection that never happened
The vials arrive in a case of 100 or 1,000. The shipping label is signed. The case goes on the shelf. Nobody records the lot number against the receiving date. Six months later, when an assessor asks which lot was in use during a specific analytical run, the answer is “the one on the shelf,” and the shelf has three lots on it. The fix is a receiving log with four fields: manufacturer, catalog number, lot number, date received. That is it. A paper log works. A spreadsheet works. What does not work is the assumption that the packing slip is the record, because packing slips get filed by accounts payable and never seen again.
Break 2: the certificate of analysis that lives in a PDF
Most vial manufacturers issue a certificate of analysis per lot. It states the glass class, the dimensions, the septum compound, and often the hydrolytic resistance result. It is issued, emailed, and filed in a folder named “CoAs” with 400 other files. When the assessor asks for the CoA for lot 22F-0417, someone searches for twenty minutes. The fix is a naming convention and an index. I use manufacturer_catalognumber_lotnumber_date.pdf and a one-page index in the quality manual that lists every consumable lot in use during the current reporting cycle.
There is a second problem with CoAs that is worth stating plainly. Many are template documents with the lot number changed and no test data behind them. An assessor who knows the industry will ask what was actually measured. A CoA that states “dimensions conform to specification” without a measured value is a statement of intent, not a record. Ask your supplier for the measured values on at least the critical dimensions: 12 x 32 mm body, 6 mm neck OD, 11 mm crimp or 9 mm screw thread, and the septa thickness. If they cannot provide them, you are relying on a claim.
Break 3: the mixed lot shelf
Vials get ordered in cases of 1,000 and used in a lab that runs 200 samples a week. A case lasts a month. Two cases overlap. The shelf has two lots. Nobody separates them. Under ISO 17025 clause 7.5.1, technical records have to be sufficient to identify “factors affecting the result.” If your shelf has two lots and your run sheet says “vial: 2 mL clear screw top,” you cannot identify the factor. Separate the lots physically and record which lot went into which run. This is the single highest-yield fix in the whole audit.
Break 4: the equivalent part substitution
A purchasing manager finds a cheaper vial with the same nominal dimensions. It is ordered. It goes on the shelf. Nobody updates the method documentation. The method was validated with a specific vial from a specific manufacturer, and the substitution changes the extractables profile, the septa formulation, and possibly the neck geometry. ISO 17025 clause 7.2.1.5 requires that methods be validated for their intended use, and clause 7.1.1 requires that deviations from documented methods be authorized, documented, and accepted by the customer. A vial substitution is a method deviation if the vial is named in the method. If it is not named, that is a separate problem, and I have written elsewhere about why the vial belongs in the method. The audit fix is a controlled substitution procedure: any change to a vial named in a validated method requires a re-validation of the affected performance characteristics, minimum three lots, minimum 20 injections per lot, and a comparison of peak area RSD and carryover.
Break 5: the expired shelf life nobody tracks
Vials do not have a regulatory expiry date the way a reference standard does. They have a shelf life the manufacturer states, typically 3 to 5 years from manufacture for unopened glass, less for septa, which are the limiting component. A silicone-faced septum with a PTFE film has a shorter working life once the package is opened, typically 12 months in a clean environment and much less in a lab with solvent vapors. If you cannot state the manufacture date of the lot on your shelf, you cannot state its shelf life. The fix is the same receiving log, with a manufacture date field, and a quarterly shelf sweep.
The documents that survive an assessment
I keep a document set for every vial lot in use. It is small. It is boring. It has never failed an audit.
- Receiving record: manufacturer, catalog number, lot number, manufacture date, date received, quantity, condition on receipt.
- Certificate of analysis: with measured values for dimensions and glass hydrolytic resistance class, not just conformance statements.
- Specification sheet: the manufacturer’s published dimensional drawing and material specification, controlled as a document under clause 7.5.2.
- Method reference: the section of the analytical method that names the vial, or the justification for why the vial is not named.
- Substitution record: if the lot changed from a previously validated vial, the re-validation data and the approval signature.
- Shelf life statement: the manufacturer’s stated shelf life and the date the lot entered service.
That is six documents. The vial itself is the seventh, and it is the one the assessor can hold.
What the specification actually has to pin down
A vial specification that survives an audit is not a marketing sheet. It is a set of measured, toleranced values. Here is what I require on the specification for any vial used in a validated method.
Dimensional specifications with tolerances
The dimensional tolerances on a molded vial are looser than most chemists assume. A 2 mL vial conforming to the 12 x 32 mm industry convention typically has a body OD tolerance of ±0.2 mm and a height tolerance of ±0.3 mm. The neck ID and the crimp or screw finish are the dimensions that matter for autosampler compatibility, and they are the ones most often left off the specification. For a 9 mm screw thread vial, the thread pitch and the finish dimensions are defined by the manufacturer’s drawing, and the closure torque is specified by the closure supplier, typically 1.0 to 1.5 N·m for a 9 mm screw cap and 2.5 to 3.5 N·m for an 11 mm crimp. Those are the numbers the assessor will want to see if you claim the closure is part of the method.
If you are running headspace, the dimensional specification tightens because the vial has to seat against a heated block and a pressure seal. A headspace vial typically needs a 20 mm crimp finish and a body OD held to ±0.15 mm so it seats concentrically. The tolerance is not cosmetic. It is the reason a headspace run drifts.
Material and glass class
The glass class has to be stated on the specification and tied to a test method. USP <660> defines hydrolytic resistance classes, and the powder test result for a Type I borosilicate vial is typically reported as a volume of 0.01 M HCl consumed per gram of glass, with the acceptance limit for Type I being below 0.10 mL/g. For a 2 mL vial that is a small absolute number, but it is the number that separates a low-extractable vial from a soda-lime vial with a surface treatment. If your method is sensitive to sodium or silica, this is the specification that matters.
For a GC vial used with a split/splitless inlet, the critical material property is often not the glass but the septa. A silicone septum with a PTFE film on the sample side is the standard for a reason: the PTFE is chemically inert and the silicone provides the reseal. The specification has to name the compound, the durometer (typically 45 to 55 Shore A), and the film thickness (typically 0.05 to 0.13 mm). A specification that says “silicone/PTFE septum” without those numbers is not a specification.
Septum puncture life
Puncture life is a real specification and it is measurable. A standard 11 mm silicone/PTFE septum in a crimp cap will typically survive 50 to 100 punctures with a 22-gauge needle before coring becomes visible, and fewer with a larger gauge or a blunt needle. For a 9 mm screw cap septum, the number is often lower because the septum is thinner. If your method runs 96 samples from a single vial in an autosampler tray, you are at the edge of the puncture life and the specification should say so. The assessor will not ask about puncture life. Your chromatograms will.
A decision checklist for selecting a traceable vial
This is the checklist I use when qualifying a new vial for a validated method. It has concrete thresholds. It is deliberately short.
- Dimensional conformance: body OD within ±0.2 mm of nominal, height within ±0.3 mm, neck finish matching the closure supplier’s drawing. Measured on a sample of at least 10 vials from the lot.
- Glass class: Type I borosilicate, USP <660> powder test below 0.10 mL/g of 0.01 M HCl. Certificate with the measured value.
- Septum formulation: compound named, durometer stated, PTFE film thickness stated. No “silicone/PTFE” without numbers.
- Puncture life: at least 50 punctures with the needle gauge used in the method, verified by a coring test (visual inspection of the septum face at 10x after the puncture count).
- Extractables: a blank injection of the vial filled with the method’s diluent, run through the method, with no peak above 3x the baseline noise at the analyte retention time.
- Carryover: a blank injected after a high-concentration standard, with carryover below the method’s LOQ, typically below 0.1% of the high standard peak area.
- Closure torque: verified with a torque tester at the supplier’s stated value, typically 1.0 to 1.5 N·m for a 9 mm screw cap, 2.5 to 3.5 N·m for an 11 mm crimp.
- Documentation: CoA with measured values, lot number on the case, manufacture date on the case, and a specification sheet controlled in the quality system.
If a supplier cannot meet items 1, 2, and 8, stop there. The rest is refinement. A vial without a measured CoA and a lot number is not traceable, and no amount of method validation will make it so.
Operational precautions that keep the chain intact
Documentation is half the problem. The other half is the physical handling that makes the documentation true. These are the precautions I enforce in our own lab and recommend to customers.
Separate lots physically and label the shelf
Two lots on one shelf is a finding waiting to happen. Use a labeled bin per lot, with the lot number and the date the lot entered service on the front. When a lot is exhausted, retire the bin and record the retirement date. This is a five-minute change that eliminates the most common traceability gap.
Record the lot on the run sheet, not just in the LIMS
Many LIMS systems have a consumable lot field. Most laboratories do not populate it because it is optional. Make it mandatory for any method where the vial is named. If the LIMS cannot enforce it, put a lot number field on the paper run sheet and transcribe it. The assessor will accept either. They will not accept “we can look it up.”
Qualify a new lot before it enters service
For a validated method, a new lot of the same catalog number is not automatically qualified. Run a system suitability test with the new lot against the previous lot: three injections of a mid-level standard, compare peak area RSD (acceptance typically ≤2.0% for a well-behaved method), retention time (typically ±0.02 min), and a blank for extractables. If the new lot fails, quarantine it and investigate. This is a two-hour exercise that has caught real problems in my lab, including a lot of septa with a film thickness at the low end of the tolerance that produced visible coring after 40 injections.
Control the substitution procedure
Any change to a vial named in a validated method is a method change. Document it, justify it, and get it approved by the quality manager before the new vial enters service. The justification should include the qualification data from the previous section. I have seen a laboratory substitute a vial with a different neck geometry and spend three weeks chasing a “carryover problem” that was a needle alignment issue. The vial was the cause. The method documentation would have caught it.
Troubleshooting: when the traceability chain is questioned
When an assessor or a customer questions the traceability of a vial, the conversation follows a predictable path. Here is the checklist I run before the meeting.
- Can you produce the lot number for the vial used in the questioned result? If yes, move to the next item. If no, the finding is already there. Fix the receiving log first.
- Can you produce the CoA for that lot? If the CoA is a conformance statement without measured values, ask the supplier for the measured data. If they cannot provide it, document the limitation and consider a different supplier.
- Is the vial named in the method? If yes, confirm the lot matches the named vial or that a substitution record exists. If no, be prepared to justify why the vial does not affect the result, with extractables and carryover data.
- Is the shelf life current? Check the manufacture date against the stated shelf life. A lot at 4 years 11 months on a 5-year shelf life is technically in date and practically a risk. Retire it.
- Was the lot qualified before service? Produce the system suitability data. If it does not exist, run it now and document the retrospective qualification, noting the date.
- Is the closure torque recorded? For a method where the closure is part of the specification, a torque record is the evidence that the closure was applied to specification. A torque tester costs less than a day of assessor time.
The pattern in every one of these is the same: the answer is a record, not an argument. An assessor will accept a record that shows a gap and a corrective action. They will not accept a verbal assurance that the vial was fine.
How the traceability chain connects to the rest of the workflow
A vial does not exist in isolation. It sits between the sample preparation and the analytical instrument, and the traceability chain has to run through both. If you are preparing a sample in a reagent bottle and transferring to a vial, the transfer step is part of the chain: the bottle lot, the transfer volume, and the vial lot all have to be recorded if the result is reportable. The same applies to a headspace vial, where the crimp torque and the septa lot are the two variables that most often change between runs.
The instrument side is where the external references matter. If your method is a U.S. EPA SW-846 method, the method number and revision are part of the traceability chain, and the vial specification has to be consistent with the method’s sample handling requirements. If you are running a method developed on an Agilent GC or LC platform, the autosampler’s vial specifications are published and should be cross-checked against your vial drawing. The same applies to Thermo Fisher sample handling systems, where the tray geometry and the vial height tolerance are the two specifications that decide whether the vial seats. Restek’s technical resources are a useful cross-check for septa and vial compatibility data when you are qualifying a new lot.
The point is not to collect links. The point is that the traceability chain for a vial runs from the manufacturer’s lot record, through your receiving log and CoA file, through the method documentation, and into the instrument’s sample handling specification. A break anywhere in that chain is a break the assessor can find.
What I would do first
If you are starting from zero, do not try to build the full document set in a week. Do these three things in this order, and you will close the majority of the gap.
First, start a receiving log today. Four fields: manufacturer, catalog number, lot number, date received. Add a manufacture date field if the case carries one. This is the foundation and it costs nothing.
Second, request measured CoAs from your vial supplier for the lots currently on your shelf. If they cannot provide measured values for dimensions and glass class, that is a supplier qualification issue, and you should be evaluating an alternative. A vial supplier that cannot produce a measured CoA is not a supplier you can trace.
Third, name the vial in your method documentation. If the vial is not named, add a section that specifies the catalog number, the lot qualification procedure, and the substitution rule. If the vial is named, confirm the lot on the shelf matches and that a qualification record exists. This is the step that turns a consumable into a controlled component, and it is the step an assessor will look for first.
The rest of the document set follows. The torque records, the puncture life data, and the extractables blanks are refinements that make the chain stronger, but they are not the foundation. The foundation is a lot number on a receiving log and a measured CoA in a file, and both of those are within reach of any laboratory this week.
ISO 17025 does not require you to calibrate a vial. It requires you to know which vial held the sample, what its specification was, and whether it met that specification. That is a documentary problem, not a metrological one, and it is solved with a log, a certificate, and a method reference. The laboratories that fail this assessment are not the ones without a calibration curve for glass. They are the ones that cannot find the lot number.
References & further reading
Related resources from our lab consumables network:
- Vial at vials-filters.com
- Headspace Vial at headspacevialsupplier.com
- Gc Vial at gcvialfactory.com
- Reagent Bottle at reagentbottleforsale.com
External references and standards: