CoA, Batch Numbers and Retained Samples for Adhesive Orders

A speaker adhesive certificate of analysis is the only document that connects the drum on your dock to the batch you qualified six months ago. Everything else in the file — the quotation, the packing list, the safety data sheet, the company certificates — describes a product in general. The CoA describes this lot, with numbers, against limits. If you are buying adhesives from China for a driver line, this is the single piece of paperwork worth negotiating hardest, and it is routinely the one nobody asks for until something has already gone wrong.

The reason it matters more for adhesives than for most bought-in parts is that you cannot inspect the thing you care about. A cone is measurable on arrival. A drum of centring adhesive looks identical whether it is the formulation you approved or a cost-reduced version of it, and the difference shows up eight weeks later as rub-and-buzz on a batch of finished drivers. Chemistry is the one category where the paperwork genuinely substitutes for inspection — which is precisely why it has to be real paperwork.

What a CoA is, and the three documents people confuse it with

A certificate of analysis reports measured results, for a named batch, against a stated specification, using stated test methods. Four elements, and if any of them is missing the document is decorative.

It is not a certificate of conformity, which asserts that a product meets a standard without necessarily showing the numbers. It is not a test report, which reports results from a one-off exercise — a type test on a formulation, say, or the halogen-free measurement discussed below — rather than from routine batch release. And it is emphatically not a management-system certificate.

That last confusion is the expensive one. A management-system certificate such as ISO 14001 speaks to documented process discipline at an organisation. It says nothing about the contents of your drum. Suppliers offer it because it is the certificate they have; buyers accept it because it looks like assurance. If you are shown one, check the issuing body, the holder name and the validity date before it goes in your file — and then ask for the batch paperwork separately, because that is the document that answers your actual question. The distinctions are worked through in the guides to reading a Chinese certificate of conformity and what ISO certifications from Chinese manufacturers actually cover.

A concrete example. Aosibo Adhesives, a Huangjiang, Dongguan supplier whose range is built around loudspeaker assembly, declared ISO 14001 on its onboarding application — but no certificate was provided at review, so nothing is claimed about it, and its corporate-registry read was still open when the profile went live. That is a normal starting position with a new chemical supplier, and exactly the situation in which batch documents matter more than company documents. Ask for the certificate if you want it; ask for the CoA because you need it.

What an adhesive CoA should actually report

"Appearance: conforms" is not a result. A usable CoA carries numeric values with units, a specification range, and the test method for each line. Which lines belong on it depends on the chemistry, and this is where you should push back on a supplier's default template.

Three practical details that separate a CoA you can use from one you cannot.

The specification limits on the certificate are the supplier's, not yours. A default template will carry wide internal limits — viscosity 8,000–20,000 mPa·s, say — inside which the product can vary far more than your process tolerates. Negotiate the limits that appear on your CoA down to the range your qualification actually supports, and reference the qualified batch number in the purchase agreement so there is an anchor.

Ask for the CoA before shipment, not with the shipping documents. A certificate that arrives in the courier pack with the bill of lading cannot stop a failing batch from being loaded. A certificate issued at batch release, emailed for your review, and made a condition of the shipping instruction can. For dangerous-goods cargo where rebooking is slow and expensive, this is worth the friction.

Insist on the test method, not just the number. Viscosity measured on different spindles at different shear rates gives different numbers for the same material. Without the method, a viscosity figure cannot be compared across batches, let alone across suppliers — which makes it useless for exactly the purpose you wanted it for. The same logic applies to reading any adhesive datasheet, and the documentation set that travels with imported chemistry follows the same rules.

Batch numbers, and the repacking problem

A batch number is only useful if you know what it identifies. There are two candidates, and suppliers are frequently vague about which one their code refers to.

The compounding batch is a single mixing run. Everything in it is chemically the same material. This is the number that matters for consistency, because it is the unit within which variation is bounded.

The pack or fill lot is a quantity drawn off and put into containers. One compounding batch can be split across many fill lots over weeks, and — more awkwardly — a repacker can fill from drums that came from several compounding batches.

That distinction is not academic when the supplier's registered premises are an upper-floor office unit rather than a ground-level chemical plant, which is a common and entirely normal arrangement in Guangdong for blending, repacking and trading operations with compounding done elsewhere. It is the situation described on the Dongguan profile referenced above: the registered address is a sixth-floor room, the business scope had not been read against the corporate registry, and whether compounding happens in-house was not established. The right response is not suspicion but specificity — ask where the compounding is done, under whose licence, whether that site can be visited, and which of the two numbers the code on your pack refers to. The blender-versus-compounder question is worth settling before the batch conversation, because it determines whose records you are actually asking for.

What to require in writing:

• A code key. What the digits mean, so you can read the compounding date and batch off a drum without emailing anyone. A code like 2609A means nothing until someone tells you it does.
• The batch number on every individual pack, not only the outer carton or the pallet label. This is the same fight as the date-of-manufacture marking discussed in the remaining shelf life clause, and you should win both in one sentence.
• One batch per consignment where possible, or a packing list that states which packs carry which batch. A mixed-batch pallet with no map is a pallet you cannot trace.
• A retention period for production records — the batch sheet, the raw material lots consumed, the release test data — stated in months and long enough to outlive a field failure that surfaces a year later.

Retained samples: the argument-settler

A retained sample is a sealed portion of the released batch, kept by the supplier and — ideally — mirrored by you. Its entire purpose is to make a future dispute testable rather than rhetorical. Six months after a scrap spike, the only honest way to answer "was the adhesive different?" is to pull the retain of the qualified batch, pull the retain of the suspect batch, and run them side by side.

Most buyers who ask for retains ask for the policy and stop there. Four parameters decide whether the retain is worth anything:

1. Volume. A 20 ml retain will not run a set of peel panels. Specify enough material to repeat the incoming tests and at least a partial bond-strength check — for most speaker grades that means a few hundred millilitres or more, and for a two-part system it means both components in ratio.
2. Retention period. It must exceed the declared shelf life plus your own consumption cycle, and ideally your product's warranty exposure. A retain destroyed at twelve months does not help with an eighteen-month field failure.
3. Storage conditions. A retain kept in an unconditioned warehouse in a Guangdong summer proves nothing about the batch, only about the warehouse. Specify the storage condition for retains explicitly.
4. Access. Written commitment that a retain will be released to you or to a third-party laboratory of your naming, at your cost, on request. Without that clause the retain exists for the supplier's benefit, not yours.

Keep your own retain in parallel. It costs a sealed container and a shelf, and it removes the single most common deadlock in a chemical dispute, which is that each party only has access to the sample the other one kept.

Your own incoming check

Documents narrow the risk; they do not eliminate it. A short incoming routine, run on every consignment, catches the things a CoA cannot: transit damage, freeze exposure, wrong lot picked, mislabelled pack. Fifteen minutes per consignment is enough.

1. Read the pack. Batch number, date of manufacture, and remaining shelf life against your contract term. Photograph one pack face per batch and file it.
2. Match the CoA to the pack. Batch numbers must agree. If they do not, quarantine before opening anything.
3. Appearance and separation. Phase separation, grit, crystals in an epoxy hardener, thickening. Two minutes, and it catches most transit and storage failures.
4. One functional number. Viscosity if you have a viscometer, otherwise fixture time or gel time on a scrap coupon at your line's cure profile. One number per batch, logged, builds a trend line that predicts trouble long before a scrap spike does.
5. A bonded coupon. For the critical joints — spider and surround — bond one coupon pair per batch on the real substrates and pull it. Compare against your qualification control, not against an absolute target.
6. Log it. Batch, date received, results, and the driver lots it was consumed on. When a field failure appears, this log is what turns a guess into a traceable answer.

If you run pre-shipment inspection on other categories, note that a third-party inspector is of limited use on chemistry — they can verify pack counts, labelling, batch marking and pallet condition, which is worth doing, but they will not tell you what is in the drum. The pre-shipment inspection guide covers what an inspector can and cannot confirm.

Common questions

The supplier says they do not issue a CoA. Now what?

That is common with blending and repacking operations, and it is negotiable rather than fatal. Two routes: ask them to obtain and pass through the compounder's batch certificate, or agree a short release-test panel they will run and report per batch — viscosity, solids, and one functional property is often enough. If neither is possible, price in your own incoming testing and treat every batch as a partial requalification, which is a real cost you should put in the comparison when you evaluate the quote.

Does a certificate of analysis prove the formulation has not changed?

No, and this is the most common misreading. A CoA proves the batch met the stated limits on the stated properties. A reformulated product can meet identical viscosity and solids limits and behave differently on your substrates. Consistency over time comes from the CoA plus a written formulation-change notification commitment — see the second-sourcing and requalification matrix for what a change actually forces you to redo.

How does the undefined unit of measure affect batch traceability?

More than you would expect. A stated minimum order quantity of 1,000 units means nothing until you know whether a unit is a bottle, a syringe, a cartridge or a kilogram — and the pack size determines how many individual packs must carry a batch mark, and how large a fill lot is. Settle the unit and the pack size before you write the marking clause.

Who should hold the retained sample if I buy through a trading company?

Whoever compounded the material, with a contractual right of access flowing to you. A retain held by an intermediary who did not make the product, and who may not be in the chain in two years, is the weakest version of this control. Name the compounding site in the agreement.

Is a halogen-free claim covered by the CoA?

Usually not, and it should not be. Halogen content is normally a type-test property measured on the formulation rather than a routine batch-release test. Ask for the test report separately, along with the laboratory, the method and the threshold it was measured against — the claim is the number, not the phrase.

What to ask the supplier next

One message, and read the reply for specificity rather than warmth:

• A sample CoA for any recent batch of a grade close to yours, with the specification limits and test methods visible.
• Whether a CoA is issued per compounding batch or per fill lot, and which number appears on the pack.
• The batch code key in writing.
• Retained-sample policy: volume, retention period, storage condition, and whether a retain will be released to a laboratory you name.
• Production-record retention period for batch sheets and raw material lots.
• A written formulation-change notification commitment with a defined notice period.
• The compounding site, by name and address, and whether it can be visited.
• The halogen-free test report and its threshold, if you need that grade.
• Pack size and unit of measure, since it governs the marking clause.

None of these were available at review for the supplier used as the example here — its listing publishes what was stated on an application and flags what was not verified, including an unsighted ISO 14001 certificate and an open registry check, which you can read on the Aosibo Adhesives factory profile. A supplier whose range maps bond by bond onto a loudspeaker driver is usually capable of answering all nine; the point of asking is to find out before a purchase order, not after a scrap report.

Get the CoA specification agreed, the batch marking specified and the retains commitment in writing, and a chemical purchase stops being an act of trust and becomes something you can audit.