A septum lot that fails at injection 43 instead of the qualified 100 leaves you with a fragment in the vial, a split peak, and a supplier asking for a lot number you cannot tie to anything but a packing slip. That gap is not a QC problem. It is a documentation problem, and it gets decided at the point of purchase. Below is a buyer’s list of the documents that make an HPLC septa certificate of analysis worth the paper, what each certificate actually attests, and how to reconstruct a lot after a failure using nothing but its number.
None of this replaces method text or your audit retention schedule. It sits upstream of both: the paperwork you demand before the first vial is capped.
Buyers often ask for “the CoA” as if it were one artifact. It is not. A septum shipment can carry four distinct document types, and each answers a different question.
Certificate of analysis (CoA)
The CoA reports measured results for a defined test set against defined limits, for a specific lot. For a silicone/PTFE septum that set typically includes durometer (Shore A), specific gravity, compression set, extractables by gravimetric or chromatographic method, and dimensional checks on thickness and diameter. What it attests: this lot’s measured values fell inside the specification window. What it does not attest: that the specification window is tight enough for your method, or that the test methods are the ones you would choose. Read the method column, not just the pass/fail column.
Certificate of conformance (CoC)
The CoC is a statement, not a measurement. It says the lot was manufactured and inspected to a named specification. It carries no numeric results. A CoC is acceptable for commodity items where the specification is public and stable. For a septum that will see 100 punctures at 40 °C in a method-critical application, a CoC alone tells you almost nothing about the lot in your hand.
Certificate of origin and material declarations
Origin certificates name the country of manufacture and, where relevant, the country of final processing. Material declarations cover what the compound contains: silicone base, platinum cure catalyst, any pigment, and the PTFE film if present. These matter for extractables risk and for customs. They do not speak to dimensional quality.
Safety data sheet (SDS)
The SDS is a hazard communication document, not a quality document. It travels with the product because regulation requires it. Do not treat its presence as evidence of anything about the septum’s performance.
| Document | What it attests | Numeric content | When it is sufficient alone |
|---|---|---|---|
| CoA | Lot results vs. limits, per test method | Yes, per test | Never for method-critical septa; always pair with a spec |
| CoC | Conformance to a named specification | No | Commodity septa in non-critical methods |
| Origin / material declaration | Country of manufacture, compound composition | No | Customs and extractables screening only |
| SDS | Hazard classification | No | Shipping and handling only |
The practical rule: a CoA without a specification document is a scorecard with no rules printed on it. Demand both, and check that the CoA’s test list matches the specification’s test list line for line.
A CoA is only as good as the specification it points to
Ask a supplier what their septum CoA tests and you will usually get a list. Ask which revision of which specification the limits come from and the conversation slows down. That pause is the buying decision.
A usable specification for an HPLC septum names, at minimum:
- Durometer with tolerance, typically ±3 to ±5 Shore A, and the test standard (for silicone, ASTM D2240 is the common reference).
- Thickness and diameter with tolerance, for example 9 mm nominal ±0.2 mm on diameter and 1.0 mm ±0.1 mm on thickness for a standard 9 mm closure.
- Compression set at a stated temperature and duration, expressed as a percentage.
- Extractables with the solvent, temperature, and time stated, and a gravimetric or chromatographic limit.
- Puncture resistance or a functional test at a stated needle gauge and count.
If the CoA omits the method column, you cannot compare two suppliers’ numbers. A durometer of 50 Shore A measured on a stacked specimen is not the same reading as one measured on a single 1 mm sheet. The number is identical; the part is not.
This is where the HPLC septa selection work and the paperwork converge. The specification is the contract; the CoA is the receipt. Neither is complete without the other.
What a lot number must let you reconstruct after a failure
The lot number is the join key between the septum in your autosampler and every document that describes it. If it cannot reconstruct the chain below, it is a label, not a traceability identifier.
The five links a lot number should open
- Compound batch: which silicone compound lot went into the press, and its CoA from the compounder.
- Molding run: press, date, cure temperature and time, and the operator or shift record.
- Post-processing: any post-cure, washing, or PTFE lamination step, with its parameters.
- QC release: the test results that cleared the lot, with the instrument and analyst.
- Distribution: which customers received which portion of the lot, so a recall can be scoped.
Ask the supplier to walk one lot number through all five. Most can do the first three. The fourth is usually in a QC database. The fifth is where the chain often breaks, and it is the link that matters most when a failure forces a containment decision.
Date code versus lot code
A date code tells you when something was made. A lot code tells you what it was made from. They are not interchangeable, and a supplier that offers only a date code has told you something about their traceability system. For septa, demand a lot code that resolves to a compound batch. If the code format is opaque, ask for the decode table in writing.
The traceability chain for a consumable runs through the closure, not just the septum. If you buy septa pre-assembled into caps, the vial closures documentation should share the same lot scheme, or you will be reconciling two numbering systems during an investigation.
What the certificate should say about extractables and bleed
Extractables are the most method-relevant number on a septum CoA, and the most often underspecified. A line that reads “extractables: pass” is not a result. A usable line reads something like: gravimetric extractables, 2.0 mg per septum maximum, after 24 h at 40 °C in acetonitrile. The solvent, temperature, time, and limit are all part of the attestation.
Two cautions when reading this section:
- Extractables results are method-specific. A lot that passes in one solvent can fail in another. Check that the CoA solvent matches your mobile phase chemistry, or ask for the data in your solvent.
- Bleed behavior at temperature is a separate property from extractables at room temperature. If your method runs the autosampler tray warm, ask for a bleed profile at the tray temperature, not just an ambient extractables number.
Where a supplier cannot provide solvent-specific extractables data, a third-party analytical testing laboratory can run the extractables screen on a retained sample. That is a cost, but it is a bounded one, and it converts an unknown into a number you can put in a specification.
Dimensional attestation: the numbers that actually move between lots
Most of a septum’s dimensional print is stable. Two or three features are not, and the CoA should report them explicitly rather than burying them in a full dimensional sheet. For a 9 mm septum the features to watch are:
- Thickness at the puncture zone: typically 1.0 mm ±0.1 mm. Thickness drives needle penetration force and reseal.
- Outer diameter: 9.0 mm ±0.2 mm for a standard 9 mm closure. Diameter drives seating in the cap.
- Durometer: 45 to 55 Shore A is a common band for silicone septa; the CoA should state the measured value, not just the band.
Ask for a process capability index (Cpk) on the two dimensions that move, not a full dimensional report. A Cpk of 1.33 is a common minimum for a controlled process; below 1.0, the process is not capable and lot-to-lot variation is being shipped to you. If the supplier cannot produce a Cpk, they are not measuring the feature often enough to know.
A septum that meets every dimensional callout can still fail if the vial it seats into is out of tolerance. The HPLC vial neck dimensions and the septum’s seating geometry are a matched pair; document both or document neither.
Reading a CoA against a supplier’s own specification: a worked comparison
Two suppliers, same nominal part, two CoAs. The comparison below is illustrative and not drawn from any specific vendor’s data. It shows the kind of mismatch that survives a casual review.
| Attribute | Supplier A CoA | Supplier B CoA | What the difference means |
|---|---|---|---|
| Durometer | 48 Shore A, ASTM D2240, single sheet | 50 Shore A, “in-house method” | B’s number is not comparable to A’s without the method |
| Thickness | 1.02 mm, Cpk 1.41 | 1.00 mm nominal, no Cpk | B reports a target, not a capability |
| Extractables | 1.6 mg/septum, 24 h, 40 °C, ACN | “Pass” | B’s line carries no usable number |
| Puncture test | 100 punctures, 22 G needle, reseal pass | Not reported | B has not attested the failure mode you care about |
| Lot traceability | Lot resolves to compound batch and press run | Date code only | B cannot scope a recall |
Supplier A’s CoA is longer, but the extra length is the point. Each line is a claim you can test. Supplier B’s CoA is short because it makes fewer claims, and a short CoA is not a lean document; it is a thin one.
Two diagnostic examples: what a thin document set costs you
Both examples below are hypothetical. They are constructed to show how the document set, not the septum alone, determines how fast you can act.
Example 1: the lot that failed early
Symptom: a septum lot begins coring visibly at injection 40, against a qualified life of 100 punctures. Fragments appear in the vial, and a peak develops a shoulder.
Useful measurement: pull the CoA and check the durometer and thickness values against the specification. In this constructed case, the CoA reports 43 Shore A against a 45 to 55 band, with a note that the result was “within re-test tolerance.” That note is the finding: the lot was released outside the band on a re-test.
Corrective action: quarantine the lot, request the compound batch CoA from the compounder, and check whether the compound itself was soft or the press drifted. If the compound was in band, the failure is a molding-run issue and the supplier’s process control is the target. If the compound was out of band, the failure originates upstream and the supplier’s incoming inspection is the target. Without the compound batch CoA, you cannot tell which, and you will replace the septum when the problem is the compound.
Example 2: the lot you cannot scope
Symptom: a customer complaint about a seal failure traces back to a septum lot. You need to know which other shipments carried the same lot.
Useful measurement: the lot code on the bag. If it resolves to a press run and a date, the supplier can list every customer who received product from that run. If it resolves only to a month, the containment scope is a month of production, which is a much larger and more expensive recall.
Corrective action: if the supplier’s distribution record is incomplete, treat the entire month as suspect and test retained samples from adjacent lots. The cost of that testing is the price of a weak traceability system, and it is paid by the buyer, not the supplier.
What to demand, in order, before the next purchase order
The checklist below is the decision rule. It is ordered by leverage: the first items prevent the most expensive failures, and the last items are paperwork hygiene.
- Specification document with revision number. The CoA must cite it. No revision, no order.
- CoA with method column populated for durometer, thickness, diameter, extractables, and puncture test.
- Cpk on thickness and diameter at or above 1.33. Below 1.0 is a stop.
- Solvent-specific extractables data at your method’s temperature, not a generic “pass.”
- Lot code that resolves to a compound batch and a press run, with a written decode table.
- Distribution record the supplier can query by lot, so containment can be scoped.
- Retained sample policy with a stated retention period, so you can test the actual lot after a failure.
Items 1 through 3 are non-negotiable for any septum in a method-critical application. Items 4 through 7 are negotiable only against a written commitment to supply them on request within a stated time.
For context on what the underlying test methods look like, the U.S. EPA’s SW-846 test methods and the technical libraries from instrument vendors such as Agilent, Thermo Fisher, and Restek show how extractables and bleed are characterized in practice. They are useful for writing a specification; they are not a substitute for the supplier’s own lot data.
The decision rule
Buy the septum whose CoA cites a revision-controlled specification, reports method-specific numbers on the two dimensions that move, and carries a lot code that resolves to a compound batch and a press run. If any one of those three is missing, the certificate of analysis is a shipping document with a quality letterhead, and the lot number on the bag will not help you when the peak splits at injection 43. Demand the three, in writing, before the next order.
References & further reading
Related resources from our lab consumables network:
- Hplc Septa at hplcsepta.com
- Vial Closures at vialsclosures.com
- Analytical Testing at testinganalytical.com
- Hplc Vial at vialforhplc.com
External references and standards: