The auditor asked for the insert lot number, and the lab manager handed over the vial lot number instead. That single substitution, logged in the inspection notes at 14:20 on day two of a SW-846 audit, triggered a 40-minute records review and a corrective action request that outlasted the actual finding. The insert was a 250 µL pulled-point glass insert sitting inside a 2 mL vial, and nobody had written down which bag it came from. The vial was traceable. The insert was not. If you run EPA methods and you cannot answer the insert traceability question cold, this article is about the paperwork and the physical requirements the auditor will actually check, not the general case for using inserts.
SW-846 is not a single method. It is a compendium of test methods, and the SW-846 test methods for waste and materials set is organized by chapter and method number, each with its own sample handling language. The relevant container requirements live in Method 5030 (purge and trap), Method 3510/3520 (separatory funnel and continuous liquid-liquid extraction), Method 8000 series determinative steps, and the Chapter Three “Inorganic Analytes” and Chapter Four “Organic Analytes” quality assurance sections. None of them name a vial insert. What they name is the sample container and the requirement that the container not alter the sample.
That is the hook the auditor uses. Under SW-846 the sample container is part of the analytical system. If the insert is in the autosampler path, the insert is part of the container. The EPA method requirements for vial insert compliance therefore reduce to one question: can you demonstrate the insert did not change the sample between collection and injection? Everything else is a records exercise.
The specific method language that gets cited
Method 5030C requires that samples be collected in containers that minimize headspace and that the sample be analyzed without transfer where possible. Method 8260 (volatile organics by GC-MS) specifies a 40 mL VOA vial with a PTFE-lined septum and zero headspace. For a 2 mL autosampler vial with a 250 µL insert, the insert is the only way to get the sample volume down to what the injector needs while keeping headspace to a minimum. The auditor will ask you to show that the insert volume and the transfer step did not introduce a bias against the early-eluting volatiles. That is a calibration question, not a container question, but the container feeds it.
Method 8270 (semivolatile organics) and the 8000 series determinative methods are less prescriptive on container geometry but require that the extract be stored and handled in a way that preserves analyte integrity. The insert is where that preservation either holds or fails. If your insert is a vial insert with a pulled point and a 5 µL residual volume, the extract sits in a geometry the method never anticipated, and the auditor knows it.
The paperwork an auditor asks for first, in the order they ask
I have sat through three SW-846 audits and two state-level inspections in the last six years. The records request is remarkably consistent. It is not a fishing expedition. The auditor has a checklist and the insert traceability question is item four or five on it, right after the vial lot log.
The insert traceability sheet
You need a document that ties every insert lot used in an EPA method run to the sample batch, the instrument, and the date range. The minimum fields are: insert manufacturer, part number, lot number, quantity received, date opened, date exhausted, and the run IDs the lot served. If your 2 mL vial and insert come from different suppliers, the sheet has two lot columns. If they come from the same supplier as a matched set, one column and a note. Either is acceptable. A blank column is not.
The failure mode I see most often is a lab that tracks the vial lot but treats the insert as a consumable with no identity. That works until the auditor asks to see the insert certificate and you produce a generic CoA that covers the glass, not the dimensional spec. The dimensional spec is what matters, because the insert dimension is a sampler specification. A 250 µL insert with an outer diameter of 5.0 mm ± 0.1 mm and a length of 31 mm ± 0.5 mm is what your autosampler was validated against. A lot that drifts to 5.2 mm OD will still fit the vial and will still inject, and it will change the needle depth and the residual volume, and nobody will notice until a QC check fails three weeks later.
The calibration and verification record
The auditor will want to see that the insert volume is reflected in your calibration. If you validated a 250 µL insert and you are now running a 150 µL insert, the calibration curve is built on a different sample volume and the method detection limit shifts. The record that closes this is a simple table: insert part number, nominal volume, verified volume by weight (fill with water, weigh, convert at 20 °C), and the date of verification. The verified volume should be within ± 5% of nominal. A 250 µL insert that verifies at 265 µL is a finding. A 250 µL insert that verifies at 248 µL is fine.
The contamination and blank record
Method blanks are where insert problems surface as data. If your insert is not cleaned or is a single-use part that got reused, the blank will show it. The auditor will pull three method blanks from the last quarter and look at the baseline. A blank that shows a consistent low-level siloxane or a phthalate at 2 to 5 times the reporting limit is a container problem until proven otherwise. The record that closes it is a blank log with the insert lot number on each blank, so you can rule the insert in or out.
| Audit item | What the auditor asks for | Acceptable answer | Common failure |
|---|---|---|---|
| Insert traceability | Lot number to run ID mapping | Sheet with lot, part number, date range, run IDs | Vial lot only, no insert identity |
| Dimensional spec | OD and length against sampler spec | CoA with 5.0 ± 0.1 mm OD, 31 ± 0.5 mm length | Generic glass CoA, no dimensions |
| Volume verification | Verified volume vs nominal | Gravimetric check within ± 5% | No verification, nominal only |
| Blank association | Insert lot on each method blank | Blank log with lot number | Blank log with date only |
| Change control | Supplier or lot change notification | Documented revalidation trigger | Silent supplier switch |
Three field cases where the insert was the finding
These are not hypothetical. Each one came out of an audit or a QC investigation I was part of, and each one has a measurement that identified it.
Case 1: the 5.2 mm lot that passed every visual check
Symptom: a 12% drop in response for the first two analytes in a 8260 run, appearing on one autosampler tray and not the other. The lab had changed insert suppliers three weeks earlier to save $0.04 per insert. Measurement: we pulled 20 inserts from the new lot and measured OD with a digital micrometer. Mean 5.21 mm, range 5.14 to 5.28 mm. The old lot measured 5.02 mm mean. The new inserts were tight enough in the vial neck that the needle was hitting the insert wall on the way down, and the septum puncture was off-center. Fix: rejected the lot, returned to the prior supplier, added an incoming OD check of 10 inserts per lot against the 5.0 ± 0.1 mm spec. The response recovered to within 2% of the original curve.
Case 2: the insert that was not in the traceability sheet
Symptom: an auditor asked for the insert lot used on a specific set of 8270 extracts and the lab could not produce it. The inserts had been purchased on a blanket PO and decanted into unlabeled bins. Measurement: we back-calculated from the purchase order dates and the run dates and narrowed it to two lots, then verified by comparing blank baselines from the two periods. One lot showed a phthalate at 3.2 µg/L in the method blank, the other at 0.4 µg/L. The higher lot was the one in use during the audited runs. Fix: the lab moved to a single-supplier insert with lot number printed on the bag and a log sheet at the bench. The corrective action was accepted, but the finding stood on the record.
Case 3: the reused insert that looked clean
Symptom: a semivolatile method showing a rising baseline over a 40-injection sequence, with the last five injections failing the continuing calibration check. Measurement: we ran a solvent blank through a fresh insert and through an insert that had been rinsed and reused. The reused insert showed a carryover peak at the retention time of the internal standard, 0.8% of the prior injection. The fresh insert showed 0.05%. Fix: single-use policy for inserts on that method, and a note in the SOP that a rinsed insert is not a clean insert. The carryover dropped below the 0.1% acceptance threshold.
How to build the insert records package before the auditor arrives
The good news is that this is a one-time build. Once the package exists, maintaining it is a 10-minute-per-lot task. The bad news is that retrofitting it during an audit does not work, because the auditor will spot a document created the week of the inspection.
The five documents that close the insert question
- Insert specification sheet with OD, length, nominal volume, glass type, and the sampler model it was validated against. One page per part number.
- Lot traceability log linking insert lot to run ID, with the date opened and date exhausted. Keep it in the same binder as the vial log.
- Incoming inspection record for OD, length, and volume on a sampling basis. Ten inserts per lot is enough for a lot of 1,000.
- Blank association log with the insert lot number on every method blank.
- Change control record for any supplier, part number, or lot change, with the revalidation trigger documented.
I would add a sixth: a one-page memo that states, in plain language, that the insert is a component of the sample container for the purposes of the method, and that the lab treats it as such. Auditors respond well to a lab that has thought about the question and written down the answer. They respond poorly to a lab that has not.
Where the consumables supply chain fits
If you buy inserts through a distributor, the lot traceability chain has one more link and one more place to break. Ask the distributor for the manufacturer lot number on the bag, not the distributor’s internal SKU. If they cannot provide it, that is a supply chain finding waiting to happen. A chromatography supplier that handles EPA-method labs will have the lot number on the outer bag and on the CoA. If you are consolidating HPLC consumables and GC consumables under one PO, make sure the insert line carries the same traceability requirement as the column line. Columns get lot numbers. Inserts should too.
For the dimensional and material specs themselves, the manufacturer data sheets are the reference. Agilent publishes insert dimensions and glass specifications in its GC columns and accessories catalog, and Thermo Fisher documents its chromatography and sample handling insert tolerances in the same way. Restek’s technical resources include insert selection guides with the OD and length numbers you need for the specification sheet. Pull the numbers from the manufacturer, not from memory, and put them on your own sheet with the source cited. That is what the auditor wants to see.
The dimensional and material parameters that go in the spec sheet
If you are writing the insert spec sheet from scratch, these are the numbers that matter for an EPA-method lab. They are not negotiable per method, but they are the baseline against which a lot change gets evaluated.
| Parameter | Typical value | Why it matters to the method |
|---|---|---|
| Outer diameter | 5.0 mm ± 0.1 mm (for 2 mL vial) | Needle alignment and septum puncture centering |
| Length | 31 mm ± 0.5 mm | Needle depth and residual volume |
| Nominal volume | 150 µL, 250 µL, or 350 µL | Calibration curve and MDL |
| Glass type | Type I borosilicate, clear or amber | Analyte adsorption and light-sensitive methods |
| Bottom geometry | Pulled point, 5 µL residual | Sample recovery at low volume |
| Temperature rating | Up to 60 °C for autosampler trays | Thermal expansion and septum seal |
| Puncture count (septum) | 1 to 3 per insert, single-use | Carryover and coring risk |
| Crimp torque | Hand crimper, no more than 2 full turns past contact | Septum compression and re-seal |
The puncture count is the one people argue about. A septum that has been punctured four times will still seal, but the re-seal force drops and the evaporation loss climbs. For a 250 µL insert at 25 °C, a 24-hour hold in a tray can lose 3 to 5% of the volume through a twice-punctured septum and 8 to 12% through a four-times-punctured one. If your method has a 24-hour hold before injection, that is a calibration bias you will not see in the QC check. Single-use inserts and single-use septa are the only defensible answer.
What to do when the auditor asks the question you cannot answer
You will not always have the record. The insert lot will be missing from the log for a two-week window, or the CoA will be for the wrong part number. The wrong answer is to produce a document that fills the gap after the fact. The right answer is to state the gap, state the compensating control, and state the corrective action. Auditors accept a documented gap with a fix in progress. They do not accept a fabricated record, and they are very good at spotting one.
The compensating control for a missing insert lot is usually a blank review. If the method blanks from the gap window are clean and the continuing calibration checks passed, you can argue the insert did not affect the data. That argument needs the blank log, the calibration records, and a written statement that the insert lot could not be confirmed. Put it in the corrective action file and move on.
The decision rule
Treat the insert as a lot-controlled component of the sample container, not as a consumable, and the audit question closes itself. That means a lot number on the bag, a traceability sheet at the bench, a dimensional check on incoming lots against 5.0 ± 0.1 mm OD and 31 ± 0.5 mm length, a gravimetric volume check within ± 5%, and a single-use policy with the puncture count capped at three. If you cannot produce the insert lot number for any EPA-method run in the last 12 months, you have a finding waiting for you, and the fix is a binder and a log sheet, not a new supplier.
References & further reading
Related resources from our lab consumables network:
- Vial Insert at hplcinsert.com
- 2Ml Vial at 2mlsamplevial.com
- Hplc Consumables at hplcconsumable.com
- Chromatography Supplier at chromatographysupplier.com
External references and standards: