SW-846 chapter 3 is where the container requirements actually live

The auditor asked for the container log before she asked for the calibration records. That detail still bothers me, because the lab had spent three weeks cleaning up instrument qualification paperwork and had not touched the box of sample container receipts. The finding was specific: four 1 L amber glass bottles used for a SW-846 Method 8082 PCB extraction had no lot traceability, no certificate of analysis on file, and no record of the solvent rinse before first use. Three of the four bottles were fine. The fourth came from an unopened case that had sat in a warehouse for eleven months, and the septa on the caps had hardened enough that the closure torque had drifted below the point where the liner compressed. The lab lost two weeks of sample holding time and had to re-collect six groundwater wells. That is the cost of treating the bottle as an afterthought in an EPA method audit.

What follows is the working version of EPA method requirements for reagent bottle selection and handling, written from the perspective of someone who has sat through the audit, written the corrective action, and rebuilt the container program from the ground up. If you are preparing for an EPA method audit or a state-delegated program review, the container requirements and the paperwork trail are the two places auditors spend the most time, and they are the two places most labs are least prepared.

Most analysts know SW-846 as a collection of numbered methods (8082, 8270, 6010, 3015, and so on). The container requirements are not in those methods. They are in Chapter Three, “Inorganic Analytes” and Chapter Four, “Organic Analytes,” of the SW-846 compendium, plus the individual method sections that reference them. The U.S. EPA SW-846 test methods page is the authoritative index, and Chapter Three (Section 3.1.3 and following) is where you find the container material guidance by analyte class.

The short version, which auditors will check against your actual bottles:

• Volatile organics (Method 5030/5035/8260): glass, with a PTFE-lined septum cap, filled to zero headspace. Two containers per sample for 5035 (one for low-level, one for high-level, or one with a methanol preservation and one with an acid preservation depending on the option chosen).
• Semi-volatile organics (3510/3540/8270): amber glass, 1 L minimum, PTFE-lined screw cap.
• Pesticides and PCBs (3510/8081/8082): amber glass, 1 L, PTFE-lined cap, no plastic.
• Metals (3010/3050/6010/6020): plastic (HDPE or LDPE) is acceptable and often preferred, with the exception of mercury, which requires glass or fluoropolymer to limit adsorption and diffusion losses.
• Cyanide and sulfide: plastic or glass, but the preservation chemistry (NaOH to pH > 12) drives the choice toward HDPE with a tight-sealing cap.
• Hexavalent chromium (3060/7196): plastic, because glass can reduce Cr(VI) at the surface over holding time.

The rule an auditor applies is not “did you use amber glass.” It is “did you use a container whose material is consistent with the analyte class in the method you cited on the chain of custody.” A 1 L amber glass bottle used for total metals is a finding waiting to happen, even if the result is fine, because the method does not authorize it.

What “PTFE-lined” actually means in the audit

The liner is where I see the most avoidable findings. A PTFE-faced silicone septum is not the same as a solid PTFE liner. The silicone core is permeable to some volatile organics over holding time, and the PTFE face is thin enough (typically 0.003 to 0.010 in) that a cap torqued too hard will extrude it past the bottle bead. Auditors who know what they are looking at will pick up a cap, look at the liner, and ask you to state the liner material and thickness from the specification sheet. If your answer is “it’s the standard cap,” you have already lost the exchange.

For volatile work, I specify a solid PTFE liner, 0.010 to 0.020 in thick, in a GPI 24-400 or GPI 22-400 closure depending on bottle size. The GPI thread finish is a published standard (Glass Packaging Institute), and the cap and bottle must match it. A 24-400 cap on a 24-410 bottle will thread on and feel tight, but the liner will not seat evenly and the headspace seal will fail under the pressure of a 5035 sample that has been frozen and thawed. That is a real failure mode, and it shows up as low recoveries on the volatiles run, not as a visibly leaking cap.

What the auditor asks for first: the container log and the COA file

In my experience across state-delegated EPA audits and third-party data reviews, the first container-related request is the container receipt log, followed immediately by the certificate of analysis or certificate of conformance for each lot in use. The log needs to show, at minimum:

• Receipt date and lot number for every case received
• Manufacturer and catalog number
• Container material (glass type, plastic resin) and nominal volume
• Closure type, liner material, and GPI finish
• Cleaning or pre-treatment performed before first use (solvent rinse, acid soak, oven bake), with the SOP reference
• Date first placed into service and the date the case was opened

The COA is the second ask. For glass bottles, the useful COA states the glass type (Type I borosilicate, Type III soda-lime), the nominal capacity and the actual capacity at the fill line, the neck finish (GPI number), and the results of any extractable testing the manufacturer performs (often a metals screen by ICP-MS at the ppb level). For plastic, it states the resin, the presence or absence of mold release agents, and the extractables profile. A COA that says only “meets specifications” is not enough for a method audit, and I have seen labs written up specifically because their COA file contained no analyte-relevant extractables data.

This is where the distinction between a reagent bottle and a sample container matters in the paperwork. The reagent bottle you use to hold a prepared calibration standard or a preservation reagent is not the same container as the sample bottle, and the auditor will expect separate logs. The reagent bottle log needs the same lot traceability plus the preparation record (what was in it, when it was made, who made it, and the expiration date). If you are using the same bottle type for both, your log must distinguish them by service, or the auditor will assume you are not tracking either.

The chain of custody is a container document

Chain of custody forms are usually treated as sampling paperwork. They are also container paperwork, because the form records the container type, the number of containers, the preservation, and the temperature on receipt. An auditor will cross-check the COC container count against the bottle count in the receipt log and against the bottle count in the extraction batch. A mismatch of one bottle, explained, is fine. A mismatch of one bottle, unexplained, is a data qualification. I have watched a lab lose an entire batch of 8270 data because the COC said “2 × 1 L amber glass” and the extraction bench record showed three bottles received and two extracted, with no note on the third.

Holding time, headspace, and the numbers auditors measure

Holding time is set by the method and by the preservation chemistry, and the container is the variable that determines whether you can meet it. The numbers I work to, drawn from SW-846 Chapter Three and the individual method tables:

Headspace is the one auditors physically check. For volatile work, they will hold the bottle up to the light and look for a bubble. The standard is no visible headspace, and the practical test is that a bubble smaller than a pea is acceptable only if the method allows it. It does not. A 6 mm bubble in a 40 mL VOA vial is enough to fail the check, and I have seen it fail on a bottle that was filled correctly but capped with a septum that had been pierced twice already. The puncture count on a septum matters: a silicone-PTFE septum is good for one or two punctures before the reseal integrity drops, and a butyl or PTFE-faced septum is good for more but degrades faster under autosampler needle friction.

If you are running headspace or purge-and-trap from the same vials, the septum choice interacts with the autosampler. The HPLC autosampler vial and the VOA vial are different animals, and I have seen labs try to run volatiles from a 2 mL autosampler vial with a pre-slit septum and wonder why the recoveries drift. The pre-slit septum is designed for HPLC injection, not for a sealed headspace sample. Match the vial to the method, not to the autosampler tray.

Two field cases where the container was the finding

Case 1: the amber bottle that wasn’t amber enough

A lab running 8082 on groundwater started seeing elevated PCB congener results on blanks and low-level samples, with a pattern that did not match the site chemistry. The measurement that identified it: the lab ran a set of method blanks through the same extraction batch and found 2,4-DDT at 0.8 µg/L in a blank that should have been below the reporting limit. The source was the bottle. The lab had switched suppliers to save on freight, and the new “amber” bottles were a lighter amber (a lower iron oxide content in the glass batch) that transmitted more UV than the previous lot. The fix was to specify the bottle against a transmittance specification (typically < 10% transmittance at 400 nm for the amber glass used in pesticide work) and to add a bottle blank to the batch QC. The bottles were not defective. They were not the right glass for the method, and the COA did not state the transmittance because the manufacturer did not test for it.

Case 2: the cap that torqued to spec and still leaked

A different lab had a metals program with HDPE bottles and a documented torque of 15 in-lb on the 28 mm caps. The bottles passed the torque check, but the acid preservation was drifting above pH 2 on receipt. The measurement: the lab weighed the bottles before and after a 30-day hold and found a 0.4 to 0.7 g loss on about 10% of the bottles. The caps were the right GPI finish, but the liner was a foam-backed PE liner that compressed permanently after the first torque cycle, so the seal relaxed within a week. The fix was to change to a PTFE-faced silicone liner and to re-torque to 18 in-lb with a calibrated torque wrench, and to add a post-fill weight check on a 5% sample. The pH drift stopped. The lesson is that torque is a number, but the liner is the mechanism, and a torque spec without a liner spec is not a seal.

The paperwork auditors ask for after the container log

Once the container log and COA file are reviewed, the auditor moves to the records that tie the bottle to the data. The sequence I have seen repeatedly:

• Container preparation records: the solvent rinse or acid soak log, with the reagent lot and the analyst initials. A bottle that was “rinsed with methanol” needs a record of which methanol lot and when.
• Batch QC records: method blank, laboratory control sample, matrix spike, and the bottle blank if you run one. The bottle blank is not required by SW-846, but it is the single best defense against a container-sourced finding, and I recommend it on every batch where the reporting limit is below 5× the expected container background.
• Holding time records: the receipt date, the extraction date, and the analysis date, with the elapsed time calculated. An auditor will do this arithmetic independently.
• Corrective action records: if a container failed (leak, headspace, broken seal), the record of what was done, when, and who approved the re-collection or the data qualification.

The corrective action record is the one most labs skip. If a bottle arrives with headspace and the lab proceeds anyway, the correct record is a nonconformance with the disposition (re-collect, qualify, or reject). If there is no record, the auditor assumes the lab did not notice, which is a worse finding than the original problem.

A decision checklist for the container program before the audit

I use this list as a pre-audit walkthrough. Every item is something an auditor has asked me for at least once.

• Is the container material matched to the analyte class in the cited method? Amber glass for organics, HDPE for metals (glass for mercury), HDPE for Cr(VI).
• Is the closure liner specified by material and thickness, not just “PTFE-lined”? Solid PTFE 0.010 to 0.020 in for volatiles; PTFE-faced silicone for preserved aqueous samples.
• Is the GPI finish recorded for every bottle and cap pairing? 24-400, 22-400, 28-410, and so on, with the cap and bottle finish matching.
• Is the torque value documented and measured? A calibrated torque wrench, a stated value (I use 15 to 18 in-lb for 24 mm and 28 mm caps), and a post-fill weight check on a sample of bottles.
• Is there a bottle blank in the batch QC? Especially for low-level organics and trace metals.
• Is the holding time calculated and recorded, not assumed? Receipt to extraction to analysis, with the method limit stated.
• Is the container receipt log complete for every case in the building? Including the unopened cases in the back room, which is where the expired liner lives.

For the analytical side, the consumables you use downstream of the bottle (autosampler vials, septa, syringe filters) carry their own traceability requirements, and the sample vial lot should be logged with the same discipline as the bottle lot. The chromatography products supply chain is deep, and the auditor will follow the chain from the bottle to the injection. If you are running GC or LC methods, the instrument vendor’s consumable specifications (for example, Agilent chromatography consumables and columns or Thermo Fisher chromatography and sample handling) are useful references for the septum and vial specifications that pair with your bottle choice, and the method validation resources at Restek technical resources are where I check septum bleed and liner inertness data before I commit to a lot.

The recommendation: qualify the bottle the way you qualify the method

The decision rule is this: treat the reagent bottle and sample container as a method component with a specification, a lot, and a verification record, not as a commodity. That means a written container specification per method (material, volume, GPI finish, liner material and thickness, transmittance for amber glass, extractables limit), a COA on file for every lot, a bottle blank in every low-level batch, and a torque and weight check on a sample of filled bottles. If you cannot produce those four items on demand, you are not ready for an EPA method audit, regardless of how clean your instrument records are. The bottle is the first thing the sample touches and the last thing the auditor checks, and it is the cheapest place to fix a problem before it becomes a data qualification.

References & further reading

Related resources from our lab consumables network:

• Reagent Bottle at reagentbottleforsale.com
• Hplc Autosampler Vial at hplcautosamplervials.com
• Sample Vial at samplevialprice.com
• Chromatography Products at chromatographyselling.com

External references and standards:

• Agilent chromatography consumables and columns
• Thermo Fisher chromatography and sample handling
• U.S. EPA SW-846 test methods for waste and materials
• Restek technical resources for chromatography