The traceability chain has seven links, and ISO 17025 only names four

A lot of 500 crimp-top vials failed incoming inspection last March on a single number: 0.4 mm. That was the difference between the certificate’s stated seal dimension and the caliper reading on the finished, crimped assembly. The supplier’s paperwork was immaculate. ISO 9001 certificate, batch record, material declaration. None of it could explain the 0.4 mm, because none of it traced the aluminum shell back to a heat-treat furnace, or the septum back to a specific calendered sheet. This is the gap that ISO 17025 traceability for vial closures exposes, and it is a gap most purchasing departments do not know they have until an auditor asks the wrong question.

I have sat on both sides of that table: as the manufacturer producing certificates, and as the buyer being audited. The traceability chain for a consumable like a vial closure is not the same animal as traceability for a balance or a reference standard. It has more links, several of them are outside your control, and exactly one of them is the one almost nobody can produce on demand. That link is the subject of this article.

ISO/IEC 17025:2017 clause 6.5 requires that measurement results be metrologically traceable to the SI, through an unbroken chain of calibrations each with a stated uncertainty. That language was written with instruments in mind. A vial closure is not an instrument, so the laboratory has to construct its own chain, and the chain has seven links in practice:

  • Link 1: the base material. Resin lot, elastomer batch, aluminum alloy heat number.
  • Link 2: the processing step. Molding cavity, cure schedule, annealing furnace run.
  • Link 3: the dimensional measurement itself, traceable to a calibrated gauge.
  • Link 4: the functional test (seal integrity, penetration force, fragmentation count), traceable to a calibrated load cell or pressure standard.
  • Link 5: the certificate of analysis that ties links 1 through 4 to a specific lot number.
  • Link 6: the receipt and storage record at the laboratory, including the date the bag was opened.
  • Link 7: the linkage from that lot to the specific analytical result that used it.

Links 3 and 4 are where labs spend their audit preparation, because those are the numbers an auditor can see on a calibration certificate. Links 1, 2 and 5 are where the chain actually breaks. I have watched a lab produce a beautiful calibration history for its crimper torque tester and then be unable to show which molding cavity produced the septa in the vials that generated a failing sequence. The instrument was traceable. The consumable was not.

The reason this matters for vial closures specifically, more than for most consumables, is that the closure is a composite. It is at least two materials bonded or assembled together, and the bond is a process, not a material. You cannot trace a process by tracing its inputs. That distinction is where most supplier certificates quietly give up.

What an unbroken chain looks like for a molded elastomer part

Let me describe what a genuinely complete chain looks like, because most labs have never seen one and therefore do not know what to ask for. I will use a 20 mm chlorobutyl septum as the example, since it is the most common closure component in pharmaceutical and environmental work.

From the compound batch to the sheet

The chain starts at the compound batch, not the finished part. A chlorobutyl formulation has a batch number assigned at the mixer, and that number follows the material through calendering. A 2.0 mm ± 0.10 mm calendered sheet has a lot number that ties back to the compound batch and to the cure state. If the supplier buys sheet from a converter rather than calendering in-house, the chain has an extra handoff, and that handoff is where lot identity most often gets lost. Ask for the converter’s lot number, not just the molder’s.

From the sheet to the cavity

Die cutting or compression molding assigns a cavity number. A 32-cavity press produces 32 septa per cycle, and each cavity has its own dimensional drift over a production run. A complete record identifies the cavity, not just the lot. I have seen a 0.15 mm thickness drift isolated to cavity 19 of a 32-cavity tool, invisible in the lot average, and responsible for a penetration-force outlier that showed up only on one HPLC autosampler position. If the supplier cannot tell you the cavity, they cannot tell you the drift.

From the cavity to the certificate

The certificate of analysis should carry the compound batch, the sheet lot, the cavity range, the cure date, and the dimensional and functional test results with the gauge identification used to produce them. That last part is the one auditors should be asking for and rarely do: which calibrated instrument produced the number on this certificate? A certificate that says “thickness: 2.01 mm” without a gauge ID is a data point, not a traceable measurement.

From the certificate to your bench

Then the chain has to survive your own receiving dock. The lot number on the certificate has to match the lot number on the bag, the bag has to be logged with a receipt date, and the first-use date has to be recorded because elastomer properties drift with storage. A chlorobutyl septum stored at 25 °C and 50% RH in an unopened bag is stable for the stated shelf life. The same septum in an opened bag on a bench for six months is a different part. If your receiving record does not capture the open date, link 6 is broken and no supplier certificate can repair it.

The link most labs cannot produce is the cavity-to-lot record

Here is the opinionated part. Of the seven links, six can be reconstructed after the fact from paperwork that most suppliers already generate. The cavity-to-lot record is the one that is almost never produced, because it requires the molder to track production at the tool level rather than the batch level, and most molders do not, because most customers do not ask.

When an auditor asks “show me the traceability for this septum,” the supplier can usually produce the compound batch and the finished lot. What they cannot produce is the specific cavity, the cure cycle for that cavity on that date, and the dimensional data from that cavity at the time of production. The result is a chain that looks complete at the lot level and is hollow at the level where the actual variation lives.

I have a working rule for procurement: if a supplier’s certificate of analysis does not name a cavity or a tool, the traceability chain stops at the lot, and you should treat the lot as a black box. That is not necessarily a disqualifier. It is a fact you need to know before you write your own traceability statement into your quality manual. Do not claim cavity-level traceability you cannot actually obtain.

Why the cavity record is the one that breaks

Three reasons, in order of frequency. First, the molder runs multiple cavities per cycle and records the cycle, not the cavity. Second, the tool is often a multi-cavity family tool shared across part numbers, so cavity identity is not even stable across production runs. Third, the dimensional inspection is done on a sample from the lot, not per cavity, so even if the cavity is known, the measurement is not.

The practical consequence is that a lab’s traceability statement for vial closures should be written to the level the supply chain can actually support. If you claim lot-level traceability and your supplier can only give you batch-level, you have written a nonconformity into your own manual. I have seen this exact finding raised in an ISO 17025 assessment: the lab’s quality manual claimed traceability to the manufacturing lot, and the certificate named only a production date range. The assessor did not care about the septum. The assessor cared that the manual said something the records could not support.

Calibrated gauges are necessary and not sufficient

Every lab I have audited has a calibrated caliper and a calibrated torque tester. Almost none of them has connected the gauge to the closure specification in a way an auditor can follow. The gap is usually the specification itself: the drawing says “seal to 0.5 N·m ± 0.1 N·m” without stating which gauge, which anvil, or which crimper, and the calibration certificate for the torque tester is filed in a different binder from the closure specification.

The fix is a traceability map, one page, that links each closure specification to the gauge that measures it, the calibration due date, and the acceptance criterion. For a 20 mm crimp closure, that map typically has four rows:

Specification Gauge / method Typical acceptance Traceability source
Crimp seal diameter Calibrated caliper, go/no-go ring gauge 20.0 mm ± 0.2 mm finished Gauge block, NIST-traceable
Crimp torque Calibrated torque tester with 20 mm anvil 0.4–0.6 N·m Dead weight, NIST-traceable
Seal integrity Pressure decay or dye ingress < 0.5 psi/min decay at 15 psi Calibrated pressure standard
Penetration force Load cell, 21G needle 2.5–4.5 N first puncture Calibrated load cell

Note the fourth row. Penetration force is the specification most labs never write down, and it is the one that predicts coring and needle damage better than any visual inspection. A septum that takes 6 N to penetrate on the first puncture is a septum that will core on the tenth. If your traceability map does not include a penetration-force criterion, you are tracing dimensions and ignoring the failure mode that actually produces particulates in the vial. The HPLC vial side of the problem gets most of the attention, but the closure is where the puncture happens.

Two field cases where the chain broke at a specific link

Case 1: the seal dimension that drifted 0.4 mm and nobody could trace it

Symptom. A pharmaceutical QC lab saw intermittent autosampler errors on one position of a 120-vial tray. The needle was failing to penetrate cleanly on roughly one vial in forty, and the resulting chromatograms showed a split peak consistent with a partially seated vial.

Measurement that identified it. Finished crimp diameter, measured with a calibrated caliper on 50 vials from the suspect lot, ranged from 19.6 mm to 20.3 mm. The specification was 20.0 mm ± 0.2 mm. The lot average was in spec; the spread was not. The caliper was traceable. The crimper torque tester was traceable. What was missing was a record of which crimper jaw set had been used, and whether the jaw had been replaced after its service interval.

Fix. The lab added a jaw-set identifier to the crimper maintenance log and a first-article crimp-diameter check at the start of each shift, recorded against the jaw-set ID. The traceability chain now runs from the finished vial back to the jaw set, and the jaw set back to its replacement record. The supplier was not at fault. The break was on the lab’s own bench, at link 2, and no incoming certificate could have caught it.

Case 2: the certificate that named a lot the supplier could not find

Symptom. During an ISO 17025 internal audit, a lab pulled the certificate for a lot of 13 mm butyl septa used in a volatile organics method and found the lot number on the certificate did not match the lot number on the bag. The certificate named a production date range; the bag named a specific lot.

Measurement that identified it. There was no measurement. The identification was a document comparison, and it took two hours to establish that the supplier’s certificate was issued at the batch level, not the lot level, and the bag label was applied by a distributor. The chain had an extra handoff between molder and distributor that neither party had documented.

Fix. The lab moved to a distributor that relabels with the molder’s lot number and provides a certificate that names both the molder lot and the distributor lot, with a stated equivalence. The lab also added a receiving check that compares the three numbers (certificate, bag, purchase order) before the material is released to the bench. This is the kind of break that analytical testing labs hit constantly and manufacturers rarely see, because it happens in the distribution layer.

A decision checklist for evaluating a supplier’s traceability claim

When you evaluate a vial closure supplier against ISO 17025 traceability requirements, ask these questions in this order. A “no” at any of the first four is a finding you need to record, even if you decide to accept the risk.

  • Does the certificate name a compound batch and a sheet or resin lot, or only a finished-goods lot?
  • Does the certificate name a cavity or tool, or state explicitly that cavity-level records are not maintained?
  • Does the certificate name the calibrated gauge that produced each dimensional and functional result?
  • Does the certificate state a penetration force and a fragmentation or coring result, not just dimensions?
  • Does the supplier provide a lot-to-batch equivalence statement if a distributor relabels?
  • Does the supplier state a shelf life and storage condition (25 °C, 50% RH, unopened) that matches your own storage?
  • Can the supplier produce the cure schedule for the specific lot on request within five business days?

If the answer to the second question is no, which it will be for most suppliers, write your traceability statement to the lot level and stop there. That is an honest statement, and an honest statement survives an audit. An aspirational one does not.

What to write in your quality manual, and what to leave out

The most common ISO 17025 traceability finding I see on vial closures is not a missing record. It is an overclaim. A lab writes “all consumables are traceable to the manufacturer’s lot” and then cannot produce a lot-level certificate for the septa because the supplier only issues batch-level documentation. The assessor does not need to understand chromatography to find that. They just read the manual and ask for the record.

Write the statement to the level you can actually support, and name the exceptions. Something like: “Vial closures are traceable to the manufacturer’s production lot where the certificate of analysis names a lot; where the supplier issues batch-level documentation, the batch is recorded and the limitation is noted in the receiving record.” That sentence is defensible. It is also the sentence most labs are afraid to write, because it admits a limitation. An auditor respects a stated limitation far more than an unstated one.

Then close the loop on your own side. The receiving record has to capture the lot, the open date, and the storage condition, because links 6 and 7 are entirely within your control and are the ones you can fix without asking anyone. If your receiving log does not have an open-date field for elastomer closures, add one this week. It is the cheapest traceability improvement available to you.

The decision rule

Trace vial closures to the level your supplier can document, not the level your manual aspires to, and put the cavity-level question to every supplier before you sign. If they can name the cavity, the gauge, and the penetration force, you have a chain worth writing into your quality manual. If they can only name a lot, write lot-level traceability and record the limitation at receiving. The chain that survives an audit is the one that matches the paperwork, not the one that sounds most rigorous. For the closure-specific specifications behind these numbers, the reference material at chromatography consumables and the method guidance in U.S. EPA SW-846 are good places to start, and the technical notes from Restek, Agilent, and Thermo Fisher are worth reading against your own receiving criteria.

References & further reading

Related resources from our lab consumables network:

External references and standards: