The auditor put a 40 mL VOC vial on the bench and asked a question that had nothing to do with the instrument: “Show me the certificate for the septa in this lot, and then show me the storage log for the box.” The lab had a valid calibration curve, a passing CCV at 98.7%, and a blank under the reporting limit. It still got written up, because the septa certificate was for a different lot than the vials on the shelf and the storage log stopped four months earlier. EPA method requirements for HPLC vials and the auditor expectations that follow them are mostly a documentation and handling problem. The chemistry is usually fine. The paper trail is where labs fail.
I have sat through enough of these audits to know the pattern. The auditor does not start with your chromatograms. They start with the container, the certificate, and the chain that connects the two to the sample you reported. This article covers what SW-846 and the method-specific language actually require of the container, and then the specific documents an auditor asks for first, in the order they ask for them.
SW-846 is not one method. It is a compendium, and the container requirements live in Chapter Three (inorganic analytes) and Chapter Four (organic analytes), then get restated or modified in each individual method. The chapters give you the floor: amber glass for most organic extracts, a minimum volume tied to the method’s required sample size, a closure that does not contribute analyte, and headspace rules for volatile work. Method 5030 and 5035 handle the volatile case; 3510 and 3520 handle aqueous and solid extraction; 8081 and 8270 handle the determinative step.
What SW-846 does not do is name a vial part number. It does not specify a septum durometer, a crimp style, or a residual seal force. That gap is deliberate, and it is exactly where the auditor expectations live. The method tells you the container must not contaminate the sample. It does not tell you how to prove it. You prove it with your own documentation, and that is what gets inspected.
The container language that actually shows up in the methods
Read the container paragraph in any 8000-series method and you will find the same three ideas repeated: the container material must be inert to the analyte, the closure must not introduce interference, and the sample must be preserved and held under conditions the method states. For volatile organics, Method 5035 pushes you toward either a zero-headspace vial or a field-preserved methanol extraction, and the choice has to be recorded. For semivolatiles, the language is looser: amber glass, PTFE-lined closure, refrigerate at 4°C ± 2°C, extract within 14 days for aqueous and 7 days if the sample was not preserved.
None of that names a septa material. That is on you. The practical reading is that your HPLC vial and its closure have to be justified as non-contributing, and the justification has to be in a document the auditor can pull. A certificate of analysis that lists extractable background at the method reporting limit is that document. A vendor datasheet with a generic “chemically inert” line is not.
Where the method stops and your SOP has to start
This is the part labs underestimate. SW-846 gives you the container type and the hold time. It does not give you the septa specification, the crimp torque, the storage temperature of the empty vial, or the maximum number of autosampler punctures before you change the cap. Those are your SOP decisions, and once you write them into the SOP, they become auditable. If your SOP says “crimp to 1.5 in-lb” and the auditor finds a crimper with no torque setting, that is a finding. The method did not require the torque value. Your SOP did, and the auditor checks you against your own document.
My rule: only put numbers in the SOP that you can actually measure and record. A torque spec you cannot verify is worse than no torque spec.
The paperwork an auditor asks for first, in the order they ask
Every 40 CFR Part 136 or state-equivalent audit I have been through follows roughly the same sequence. Knowing the order lets you stage the binder before they walk in.
- Certificate of analysis for the lot on the shelf. Not a sample certificate, not a spec sheet. The CoA with the lot number that matches the carton in front of them.
- Certificate for the septa and caps, matched to the same lot or to the closure lot referenced on the vial CoA. This is the one that trips people. Vials and closures often ship as separate lots, and the auditor will ask you to reconcile them. If the vial CoA does not reference the closure lot, you have a gap.
- Receipt and storage records. Date received, storage location, temperature if the SOP specifies one, and the person who logged it. An empty vial stored next to a solvent carboy is a finding waiting to happen.
- The extractable/background data. Either on the CoA or in a separate qualification report. The auditor wants to see that the container was shown not to contribute at or above your reporting limit.
- The SOP that maps the container to the method. One page that says: for Method 8270 we use this vial, this closure, this septa, this crimp, this hold time. Cross-referenced to the CoA.
- Training and handling records. Who crimps, who checks headspace, who logs the box. Usually a signature sheet, but it has to exist.
The pattern is that the auditor is testing whether your container is traceable to a document, and whether that document is current. A CoA from 2019 for a carton received last month fails on its face. So does a certificate that lists “typical” extractable values instead of a guaranteed limit.
In-spec and capable are not the same claim, and the CoA has to say which one it is
Here is the distinction that separates a CoA an auditor accepts from one they challenge. A guaranteed limit is a number the vendor commits to and will replace product against. A typical value is a number from a historical dataset that the vendor is not warranting. They look identical on a datasheet. They are not identical in an audit.
If your CoA says “extractable background: typical < 0.5 ppm” the auditor will ask what happens when it is 0.8 ppm. If the answer is “we have never seen that,” you have no control. If the CoA says “guaranteed < 0.5 ppm, tested per lot, out-of-spec lots rejected,” you have a control and a disposition path. That is the difference between a passing review and a finding.
The same logic applies to dimensional claims. A 12 x 32 mm vial with a “typical” neck OD of 6.0 mm and a “guaranteed” neck OD of 6.0 ± 0.1 mm are different products to an auditor. The guaranteed one has a tolerance band and a gauge. The typical one has a hope.
How to read the extractable section of a CoA in 90 seconds
Look for three things. First, the test method: was the background measured by the same determinative technique as your analyte class, or by a generic gravimetric or UV method? A UV absorbance number tells you nothing about a 2 ppb GC/MS response. Second, the reporting basis: per vial, per lot, or per square centimeter of wetted surface? Per-vial is the only basis that maps to your sample. Third, the disposition: what happens to a lot that exceeds the limit? If the CoA does not describe rejection, the limit is decorative.
I have rejected more than one incoming lot on that third point alone. The number was in spec, but the vendor could not tell me what they would do if it were not.
Two field cases where the container, not the instrument, caused the finding
Case one: the septa lot that drifted the siloxane baseline
A semivolatile lab running Method 8270 started seeing a Si-containing peak at retention time 14.2 min in the method blank at roughly 3x the reporting limit, intermittently. The instrument was clean. The column had been trimmed. The CCV passed. The pattern was that the blank only failed on days when a particular carton of caps was in use.
The measurement that identified it: they ran a solvent blank through a cap from each carton, six replicates each, and compared the siloxane peak area. The suspect carton gave a mean area with an RSD of 22% and a mean 4.1x the historical carton. The septa in that lot were a different silicone formulation with a higher low-molecular-weight siloxane fraction. The fix was a septa specification change: a HPLC septa with a documented siloxane extractable limit, tested per lot, plus a receiving test that runs one cap per carton through the blank before the carton goes to the bench. That receiving test takes 20 minutes and has caught two bad lots since.
The audit angle: the lab could not produce a septa CoA that matched the carton. That was the finding, separate from the blank failure.
Case two: the crimp that leaked at 4°C over 14 days
A water lab preserved volatiles in 40 mL vials and held them at 4°C for the 14-day hold. Over a month, four of 60 samples came back with low recovery on the volatile surrogate, clustered in one analyst’s batches. The instrument was fine. The samples were not.
The measurement: they weighed each vial at receipt and again at analysis. Four vials had lost 0.15 to 0.4 g over the hold, which is headspace loss through the closure. The cause was a hand crimper set too loose, giving a crimp that sealed at room temperature but relaxed at 4°C as the septum compressed. The fix was a torque-controlled crimper set to 1.5 in-lb, a daily crimp check with a torque gauge on one vial per batch, and a weight check at receipt that flags any vial more than 0.05 g below its filled weight. Recovery came back into the 95 to 105% band and stayed there.
Again, the chemistry fix was simple. The audit finding was that the SOP said “crimp securely” with no number, so there was nothing to check the analyst against.
What the auditor checks on the closure, and the numbers they want to see
The closure is where most container findings originate, because it has the most variables and the least documentation. Here is what I stage for the closure section of the binder, with the numbers I actually use.
| Audit item | What the auditor wants | Concrete target I use |
|---|---|---|
| Septa material and durometer | Specification with material and hardness | PTFE-faced silicone, 45 to 55 Shore A |
| Septa extractable limit | Guaranteed limit, per lot, per vial basis | < 0.5 ppm siloxane background by GC/MS |
| Crimp torque | Set point and verification method | 1.5 in-lb, gauge check daily |
| Puncture count before change | SOP limit and rationale | 3 punctures max, then new cap |
| Seal integrity | Weight check or leak test | No loss > 0.05 g over 14 days at 4°C |
| Closure-to-vial lot match | CoA cross-reference | Closure lot printed on vial CoA |
| Storage of empty vials | Location and temperature log | Sealed carton, 15 to 25°C, away from solvents |
| Headspace for VOC | Zero headspace or documented alternative | Zero headspace for 5035 aqueous |
The 3-puncture limit is one I will defend. Past three autosampler punctures, a 45 Shore A silicone septum starts to core and the leak rate climbs. I have measured it. The fourth and fifth puncture on a soft septum can open a path that shows up as a drifting internal standard late in a long sequence. If your SOP allows unlimited punctures, you have an unmeasured variable.
The closure side also has to reference the vial. That is why I keep the closure lot number on the vial CoA and the vial lot on the closure CoA. If you buy vial closures from a different supplier than the vials, you own the reconciliation, and the auditor will ask you to show it.
The GPI and dimensional references that come up
For screw-thread vials, the closure thread is a GPI 8-425 or GPI 13-425 depending on the vial, and the auditor may ask you to confirm the thread match if they see cross-threading or a loose cap in the field. A 13-425 closure on an 8-425 vial will not seal, and it will not be obvious until the sample leaks. For crimp vials, the neck finish is 11 mm for the standard 2 mL vial, and the crimp dimension has to match the crimper. These are not SW-846 numbers. They are the physical facts your SOP has to encode.
If you want the reference data on septa and closure dimensions, the Restek technical resources are the ones I hand to new analysts. For the instrument-side fit, the Agilent consumables reference and the Thermo Fisher sample handling documentation give you the autosampler geometry you need to confirm the vial actually sits at the right depth.
The SOP language that survives an audit, and the language that does not
I have rewritten enough container SOPs to know which sentences get challenged. Here is the translation table I use.
| Weak language (gets a finding) | Audit-proof language |
|---|---|
| “Use clean, appropriate vials.” | “Use lot-traceable 2 mL amber vials with CoA on file, closure lot matched to vial CoA.” |
| “Crimp securely.” | “Crimp to 1.5 in-lb with calibrated crimper; verify with torque gauge daily.” |
| “Store vials in a cool, dry place.” | “Store sealed cartons at 15 to 25°C, away from solvent storage; log receipt date and location.” |
| “Replace septa as needed.” | “Replace cap after 3 punctures; record puncture count on run sheet.” |
| “Use appropriate septa for the method.” | “Use PTFE-faced silicone, 45 to 55 Shore A, extractable background < 0.5 ppm, per lot CoA on file.” |
The pattern is that every weak sentence has a number hiding inside it that the lab never wrote down. The auditor is not asking you to be more rigorous than the method. They are asking you to be as rigorous as your own SOP claims. Write the number, measure the number, log the number, and the container section of the audit goes quiet.
One more thing on storage. Empty vials are not inert packaging. A carton stored open next to a toluene carboy will pick up toluene on the septa and the vial interior, and it will show up as a blank hit weeks later. I have seen it. The fix is a sealed carton, a labeled shelf, and a storage log that the auditor can read in ten seconds.
What to put in the binder before the auditor arrives
Stage these in this order, tabbed, one page per item where possible. This is the sequence that has gotten me through every container-focused audit without a finding on the vial itself.
- Method-to-container mapping sheet: each method you run, the vial and closure you use for it, the septa spec, and the hold time.
- Current CoAs for every lot on the shelf, vial and closure, cross-referenced.
- Extractable/background qualification report, per lot, with the guaranteed limit and the disposition rule.
- Receiving and storage log, including the septa receiving blank test result per carton.
- Crimp torque verification log and the crimper calibration certificate.
- Puncture-count run sheet excerpt showing the 3-puncture limit in use.
- Training records for everyone who crimps, caps, or logs a carton.
If a vendor cannot give you the CoA cross-reference and the disposition rule, that is the finding, and it is a vendor problem you inherited. The EPA SW-846 test methods page is where the auditor will point when they say the method requires a non-contaminating container. Your job is to have the document that shows yours is.
The decision rule
Buy only vials and closures whose CoA states a guaranteed extractable limit on a per-vial basis, carries the closure lot number, and describes the rejection rule for an out-of-spec lot. Write the crimp torque, puncture limit, storage temperature, and receiving blank test into your SOP with numbers you can measure. Log every carton from receipt to use. If any of those four items is missing, the container is not auditable, no matter how clean the chromatogram looks. Fix the paperwork before the auditor asks, because the chemistry will not save you when the certificate does not match the box.
References & further reading
Related resources from our lab consumables network:
- Hplc Vial at hplcvialfactory.com
- Chromatography Products at chromatographyforsale.com
- Hplc Septa at hplcsepta.com
- Vial Closures at vialsclosures.com
External references and standards:
