Tooling ownership is the first thing to check, and almost nobody does

A procurement manager once sent me two lots of 2 mL clear screw-thread vials, both 9-425, both from the same physical plant in Jiangsu, and asked why the certificate of analysis on one listed a hydrolytic resistance of 0.08 mL of 0.01 M HCl per gram while the other listed 0.12 mL. Same glass furnace. Same forming line. Different brand on the carton. That is the OEM versus private label vial difference in one number: the factory did not change, but who owned the drawing and the recipe did, and the recipe is what the CoA reports.

This article is about that gap. Not pricing structures, not contract terms, not MOQ negotiations. Those live elsewhere. Here I want to walk through the three things that actually differ when a vial leaves the same building under two different names: tooling ownership, recipe control, and who carries the liability when a lot drifts. If you buy vials, whether as a vial OEM or through a distributor’s private label, you need to know which of those three you actually hold.

A 9-425 screw-thread vial is not made on a generic machine with generic molds. The forming station uses a specific set of tooling: neck ring, blow molds, plunger, and the thread insert. That tooling costs real money. A typical 2 mL vial mold set for a 4-station IS machine runs somewhere between USD 18,000 and USD 45,000 depending on cavity count and whether it includes the cam-driven thread insert. That cost has to sit somewhere on a balance sheet.

In an OEM arrangement, the brand owner pays for the tooling, or pays an amortized tooling charge across the first orders, and the tooling is owned by the brand. It stays at the factory, but it is earmarked. It cannot be used to run another customer’s vials without written consent. If the relationship ends, the tooling is either bought out, moved, or scrapped. The brand controls the drawing that the tooling was cut to.

In a private label arrangement, the factory owns the tooling and runs it for whoever wants it. The distributor picks a stock mold off the shelf, applies their label, and the vial that ships is whatever that mold produces on that line on that day. There is nothing wrong with this model. It is how a large fraction of the world’s HPLC vials reach the market. But it means the distributor does not control the drawing. The factory does. If the factory decides to re-cut the neck ring to extend tool life, the distributor finds out when the next lot arrives.

What the drawing actually contains

A vial drawing is not a picture. It is a set of tolerances that the tooling was cut to hold. For a 9-425 vial the critical ones are:

• Neck OD at the thread crest: typically 9.00 mm ±0.10 mm. Out of tolerance and the cap either cross-threads or leaks.
• Thread pitch and start position: a 9-425 is a 9 mm nominal with a 1.0 mm pitch in most implementations, but the start angle matters more than the pitch. A 5-degree start offset and the cap tilts on the first turn.
• Neck ID: 6.00 mm ±0.10 mm for a standard 2 mL. This sets the septum compression.
• Body OD and concentricity: 11.6 mm ±0.15 mm, with concentricity under 0.20 mm TIR. Concentricity is what keeps the vial centered in the autosampler gripper.
• Bottom thickness and flatness: bottom thickness 1.1 mm ±0.15 mm, flatness under 0.10 mm. This is the one that decides whether the vial sits plumb in the tray.

If the brand owner holds the drawing, all five of those numbers are contractual. If the distributor holds a private label, those numbers are the factory’s internal standard, and they can drift inside the factory’s own AQL without anyone on the buyer’s side being notified. I have watched a private label lot come in with a body OD of 11.75 mm, inside the factory’s ±0.15 mm but 0.15 mm larger than the OEM lot from the same line. It jammed on a 54-position autosampler tray that had been running the OEM lot for two years.

The recipe is where the CoA number comes from, and it belongs to whoever wrote it

Type I borosilicate is a material classification, not a recipe. USP <660> and the European Pharmacopoeia set the hydrolytic resistance limit at 0.10 mL of 0.01 M HCl per gram of glass for Type I. ISO 4802 gives the same test method. A factory can hit that limit with a range of batch compositions: 33-expansion borosilicate with roughly 81% SiO2 and 13% B2O3 is the standard, but the alkali content, the fining agents, and the cullet fraction all move the number.

An OEM brand specifies the batch composition in the supply agreement. They specify the cullet ratio (I have seen OEM contracts cap it at 20%; private label runs often go to 40% to save furnace energy). They specify the annealing lehr profile, because residual stress shows up as a hydrolytic resistance drift after the vial is autoclaved. They specify the forming speed, because a faster IS machine cycle thins the wall and the wall thickness drives the crush strength.

Under private label, the factory’s internal recipe is the recipe. It is a good recipe, or the factory would not be in business. But it is optimized for yield and furnace life, not for the specific analytical method you are running. That is the trade. You get a lower unit price and you accept the factory’s internal quality target.

What changes when the brand changes but the factory does not

Here is the scenario that trips people up. You buy vials from Distributor A for three years. Distributor A sources from Factory X. Distributor A loses the account, or you switch to Distributor B for a better landed cost. Distributor B also sources from Factory X. You expect identical vials. You get a different lot behavior, and you cannot figure out why.

The answer is almost always one of four things:

• A different mold cavity set. Factory X runs 16-cavity molds. Distributor A’s OEM tooling was cavity-matched and certified. Distributor B’s private label runs on whichever mold is free that week, and cavity-to-cavity variation on a 16-cavity mold can be 0.08 mm on neck ID.
• A different cullet batch. The furnace runs continuously. The cullet fraction cycles over a 30 to 60 day window. Two lots six weeks apart can hit the same spec with a 0.02 mL/g difference in hydrolytic resistance.
• A different annealing profile. If the lehr was re-tuned for throughput, residual stress goes up. The vial passes incoming inspection and fails after two autoclave cycles at 121 °C.
• A different septum. This is the one people miss. The cap and septum are usually a separate supply chain from the glass. A private label cap assembly is often a stock 9-425 cap with a 1.0 mm PTFE-faced silicone septum from a third factory. The OEM version may be a 1.3 mm septum with a specific silicone durometer. Puncture force and reseal behavior differ.

None of these are defects. They are the natural consequence of the factory running to its own standard rather than yours. The vial is the same product category. It is not the same part.

The liability question nobody puts on the RFQ

When a vial fails in a regulated method, the question the auditor asks is not “was the vial defective.” It is “who signed the drawing and who holds the change control record.” Under OEM, the brand owner holds it. Under private label, the factory holds it, and the distributor is a pass-through.

This matters at exactly one moment: when a lot drifts and you need to prove the drift did not come from your side. Under OEM, the brand owner has the batch records, the furnace logs, the lehr profiles, and the tooling history, and they can produce them in a change control package within a week. Under private label, the distributor has to ask the factory, and the factory may or may not have kept the records in a form that satisfies your auditor.

I am not saying private label vials are unsafe. I am saying the evidence chain is shorter. If your method is validated under 21 CFR Part 211 or under an ISO 17025 scope, that evidence chain is part of your method. If you are running a research method or a screening assay, it is not. Match the sourcing model to the regulatory weight of the method, not to the price.

A comparison table you can actually use at the RFQ stage

Here is how I lay out the three sourcing models side by side. This is the table I put in front of a lab manager before they issue an RFQ, because it forces the conversation about ownership before the conversation about price.

Read the table by row, not by column. The row that matters most is change notification window. A 90-day written window is what lets you re-validate a method before the new lot hits the bench. A private label arrangement with no notification window means you find out at incoming inspection, or worse, at the QC failure.

Two field cases where the sourcing model was the actual root cause

Case 1: the autosampler pickup failure that was not a tolerance problem

A contract lab running a 120-position Agilent autosampler started seeing pickup errors on roughly 3% of positions in a 96-vial batch. They had been running private label 2 mL vials from a distributor for 14 months. The obvious suspect was vial concentricity. They measured 60 vials with a dial indicator and got a TIR of 0.14 to 0.22 mm, all inside the 0.20 mm spec, all inside the factory’s stated tolerance.

The actual cause was the cap. The private label distributor had switched cap suppliers six weeks earlier without notification. The new cap had a 0.15 mm taller skirt and a slightly different knurl. The autosampler gripper was grabbing the cap, not the vial body, and the taller skirt meant the gripper was seating 0.15 mm higher. On a vial with a body OD at the top of tolerance, the gripper lost its grip.

The fix was not to tighten the vial spec. It was to move to an OEM arrangement where the cap and septum were specified on the same drawing as the vial. That is the point. Under private label, the cap is a separate commodity. Under OEM, it is part of the part.

Case 2: the ghost peak that traced back to a cullet change

A pharma QC lab running a headspace method on a headspace vial saw a new peak at retention time 4.2 minutes appear in blank runs. It showed up in 4 of 10 blanks, always at low level, always the same RT. They spent three weeks on the GC, replaced the column, baked out the inlet, and changed the septum in the inlet. The peak stayed.

They then ran the blank with vials from an older lot still in the cupboard. No peak. The new lot was private label, sourced from the same factory as the previous OEM lot, but the factory had increased the cullet fraction from 20% to 35% in the intervening six months to reduce furnace energy. Higher cullet means more surface alkali on the cullet, and more surface alkali means a higher level of sodium and potassium species that can volatilize at the headspace oven temperature of 80 °C and elute as a ghost peak on a non-polar column.

The measurement that identified it was a simple one: they ran the same blank in a Type I vial from a different factory and the peak disappeared. Then they ran a hydrolytic resistance test on both lots. The private label lot came in at 0.11 mL HCl/g, over the USP Type I limit. The factory had shipped it anyway because their internal AQL sampled 20 vials per 100,000 and the drift sat inside the sampling error. The fix was to move the headspace method to an OEM vial with a contractual cullet cap. If you are running headspace, this is not a theoretical risk. It is a routine one, and the headspace vial specification should name the cullet cap explicitly.

The decision checklist: which model fits which method

I do not recommend OEM for everyone. I recommend it for the methods where the vial is part of the validated system. Here is how I sort it.

• Choose full OEM if: the method is validated under 21 CFR Part 211 or ISO 17025, the vial is named in the method, the autosampler is high-density (96 or 120 positions), or you run headspace at 80 °C or above. You need the drawing, the change notification, and the cap and septum on the same spec.
• Choose distributor-branded OEM (hybrid) if: the method is validated but the volume does not justify a full tooling charge. You get the drawing control without the tooling amortization, at a 10 to 20% premium over private label.
• Choose private label if: the method is a screening assay, a research method, or a method where the vial is not named. You accept the factory’s internal standard and you verify at incoming inspection with a sampling plan sized to the failure cost. A GC vial for a teaching lab does not need a 90-day change notification clause.
• Do not mix models within one method. The worst outcome I see is a lab that runs OEM vials for the validated method and private label vials for the same method on the second shift. The data is not comparable and nobody knows it.

On incoming inspection, whatever model you choose, sample to the failure cost. For a validated method, an AQL of 0.65 at a sample size of 80 vials per lot catches the drift that matters. For a screening method, an AQL of 2.5 at 32 vials is enough. The EPA SW-846 series sets container performance requirements that a private label vial can meet, but the sampling plan is yours to set, and the plan is what catches the drift before the method does.

What to ask the factory, and what to ask the distributor

The questions split cleanly by model. Ask the factory, whether OEM or private label:

• What is the cullet fraction cap in the batch recipe, and is it in writing?
• What is the lehr annealing profile, and what is the residual stress target in nm/cm?
• Which mold cavity set will run my lot, and what is the cavity-to-cavity variation on neck ID?
• What is the hydrolytic resistance target and the test method (ISO 4802 or USP <660>)?

Ask the distributor, under private label only:

• Who owns the tooling, and can I see the drawing?
• What is the change notification window, and is it contractual?
• Who supplies the cap and septum, and has that supplier changed in the last 12 months?
• Can you produce the batch records and furnace logs for a specific lot within five business days?

If the distributor cannot answer the last two, the private label arrangement is not a sourcing decision. It is a hope. The same discipline applies to a reagent bottle if you are buying those on the same PO. The ownership question does not change with the container format.

For reference specifications, the Agilent chromatography consumables catalog and the Thermo Fisher chromatography documentation both publish the dimensional tolerances they hold their own-brand vials to, and those numbers are a useful benchmark for what an OEM drawing should contain. Restek technical resources also publish septum and cap compatibility data that is worth reading before you specify the cap assembly.

The decision rule

Own the tooling and the recipe if the vial is named in a validated method, and buy private label only where you have accepted the factory’s internal standard in writing and sized your incoming sampling plan to catch the drift. The factory does not change when the brand changes. What changes is who holds the drawing, and that is the only thing the CoA number is really reporting.

References & further reading

Related resources from our lab consumables network:

• Vial at vials-filters.com
• Headspace Vial at headspacevialsupplier.com
• Gc Vial at gcvialfactory.com
• Reagent Bottle at reagentbottleforsale.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