How to Read a Certificate of Analysis (COA) for a Lubricant Shipment: Verifying What You Actually Received

Last verified: 2026-07-20

Suppose a container has just come off the truck at your goods-inwards bay. Drums are stacked, labels visible, and somewhere in the paperwork folder is a Certificate of Analysis for the batch that's supposedly sitting in front of you. Someone asks: is this good to release to stores? The honest answer is that a certificate of analysis lubricant shipment document only tells you what a lab measured on one production batch, using specific test methods, before the goods ever left the plant — it does not, by itself, confirm the drums in front of you are that batch, that the numbers meet your contract, or that nothing happened to the oil in transit. Reading a COA properly means checking all three separately: the identity match, the method behind each figure, and the spec you're actually comparing against.

What is a COA actually telling you?

Three documents travel with most lubricant shipments, and buyers routinely conflate them.

A Technical Data Sheet (TDS) is the supplier's published statement of typical or target values for a product line — it describes what the formulation is designed to deliver, not what any specific drum contains. A Safety Data Sheet (SDS), built on the 16-section GHS format, covers hazard and handling information; it tells you nothing about product quality. The COA is different again: it's a set of lab-measured results tied to one production batch or lot, generated by testing an actual sample from that run.

That distinction is the whole exercise, and it's worth holding onto as a single sentence: a COA proves one batch, on the parameters listed, at the moment it was tested — and nothing beyond that.

It doesn't confirm condition on arrival. Transit — heat cycling, vibration, a compromised seal — happens after the COA was issued, not before. A separate Knowledge Hub piece on how TDS, COA, and SDS differ in a purchase order walks through that document trio in more depth if your team needs a shared reference before goods-inwards training.

Does the batch number on the paper match the drum in front of you?

Before you look at a single test result, check identity.

A COA is issued per production batch, not per shipment. A single shipment can easily span more than one batch, which means you should expect one COA per batch referenced, cross-checked against the drums, IBCs, or tank compartments that batch actually fills — typically via a drum serial range or lot code on the packing list. The exact physical marking format varies by supplier, so confirm it against your specific paperwork rather than assuming a universal layout.

If the batch or lot number stencilled on the physical container doesn't match the number on the COA in your folder, stop there. A COA for a different batch than what physically arrived proves nothing about the goods on your dock — no matter how clean the numbers on the page look.

Which test method is behind each number?

Every meaningful parameter on a lubricant COA sits behind a named test method, and the method matters as much as the value.

Kinematic viscosity (KV40, KV100) is measured under ASTM D445, the long-standing reference method; some labs now report it via ASTM D7042, an automated Stabinger-type instrument that produces dynamic viscosity, density, and calculated kinematic viscosity together. Either is legitimate — check the COA states which.

Viscosity Index (VI) is where the first real surprise sits. VI is not a direct measurement. ASTM D2270 is a calculation — a practice — that derives VI from the two viscosity readings above. A COA's VI figure is therefore only as trustworthy as its underlying KV40 and KV100 lines. If VI looks off, check the raw viscosity data before you question anything else.

Flash point carries a second, easy-to-miss trap. ASTM D92 uses a Cleveland open cup; ASTM D93 uses a Pensky-Martens closed cup. They are not interchangeable methods, they don't cover identical sample ranges, and — because the cup geometry differs — closed-cup testing reads a few degrees lower than open-cup on the same sample. A COA that states a flash point with no method named is an incomplete document, and a D92 result cannot be measured against a spec limit written for D93, or vice versa.

Pour point runs to ASTM D97.

Base Number (TBN), relevant on most engine oils, is tested by ASTM D2896 (potentiometric perchloric acid) or ASTM D4739 (potentiometric hydrochloric acid). These are not simply interchangeable alternatives: D2896 is the higher-precision method used to set and verify the base number of new, formulated oils, while D4739 is oriented to tracking base-number loss as an oil is used in service. A new-oil COA is therefore usually expected to report D2896. Their results on the same oil are not always numerically identical — so if your spec was written against one method and the COA reports the other, that's a method mismatch worth a question, not automatically an off-spec batch.

Acid Number (TAN) is tested by ASTM D664 (potentiometric) or ASTM D974 (colour-indicator, suitable only for lighter or transparent oils — not for dark, opaque samples).

Water content is tested by ASTM D6304, a coulometric Karl Fischer method built specifically for finished lubricating oils and capable of ppm-level precision. Some older paperwork still references ASTM D95, a Dean-Stark distillation method that's coarser and historically associated with crude oil and bitumen rather than finished lubricants. If a COA reports water via D95 on a finished lube, that's worth a question.

Density/specific gravity is typically ASTM D4052, a digital density-meter method that's the current standard for finished lubricants; ASTM D1298, a hydrometer method, is the older legacy approach.

Appearance/colour is read against the ASTM D1500 colour scale. Colour is a batch-consistency indicator, not a performance metric — a number on this line proves nothing about how the oil will perform.

A handful of industrial-oil parameters — foam (ASTM D892), demulsibility (ASTM D1401), oxidation stability or RPVOT (ASTM D2272, written for steam-turbine oils) — show up on some industrial COAs but aren't routine on a typical engine-oil certificate. Don't expect them unless your product category calls for them.

Why do these numbers move when something's wrong?

Reading a COA well means knowing what a deviation actually signals.

A flash point below spec usually points to a lighter or more volatile fraction than intended — a wrong-grade mix-up or solvent/fuel contamination — and it can also change how the product is classified for transport. Elevated water content, in ppm, generally traces back to contamination: condensation inside a poorly sealed drum or IBC, a compromised liner, or a container that previously held something else; water accelerates additive hydrolysis, particularly on ZDDP and ester-based systems, and it promotes rust. A VI reading that's lower than expected is a signal to re-check the two underlying viscosity lines first, since VI itself isn't measured. A TBN below the agreed floor — not merely below "typical," but below a genuine agreed minimum — signals a reduced acid-neutralization reserve, which matters more on some applications than others.

Colour shift between shipments that are supposed to be "the same product" is a consistency flag worth raising, since it can correlate with a base-oil or additive change, or with storage oxidation. On its own, though, colour proves nothing conclusive — it's a prompt to ask a question, not a verdict.

Typical or limit — which one are you actually comparing against?

This is where most COA disputes actually start, and it's avoidable.

A TDS "typical" figure is a representative number describing the formulation in general — it is not a pass/fail boundary. The number you should be comparing the COA against is the agreed spec: a minimum/maximum from your contract or purchase order, the min/max column on the TDS if one is provided, or a referenced OEM or industry specification such as an SAE J300 viscosity-grade band. Altonex Global, as a B2B trade platform connecting buyers and suppliers, does not set or arbitrate that spec — the buyer and supplier agree it directly, and the COA is checked against whatever they agreed, not against a marketing "typical" line.

That single confusion — treating a TDS typical value as if it were a contractual limit — accounts for a large share of avoidable disputes at goods-inwards.

A parameter is missing or out of range — now what?

Two situations call for the same discipline: write it down and raise it before you release the goods.

If a parameter material to your application is simply absent from the COA — TBN on an engine oil, water content on a hydraulic oil — treat that silence as a flag, not a pass. Raise it with the supplier before acceptance. A missing line is not evidence the batch is fine; it's evidence you haven't seen the number yet.

If a parameter is present but outside the agreed spec, document it precisely: cite the exact COA line, the spec limit it fails against, and the batch number involved. Raise it with the supplier under your own commercial terms. If the disagreement doesn't resolve on the paperwork alone, independent lab verification is the buyer's recourse — Altonex Global is not the seller in these transactions and does not adjudicate quality disputes; it connects you to the supplier so you can resolve it directly, or route the sourcing conversation through the RFQ center if you're evaluating an alternative.

Can you trust a supplier-issued COA under a letter of credit?

Here's the second real surprise, and it's a trade-finance mechanic that catches buyers off guard.

Under ICC UCP 600 — the rules governing documentary letters of credit, in force since 1 July 2007 — a COA issued by the seller (the beneficiary of the LC) is examined by banks like any other document in the set. UCP 600 does not require that COA to be independently issued unless the letter of credit itself explicitly names an independent issuer. In practice, that means a self-issued COA from the supplier's own QC lab satisfies the bank's document check on its own, unless your contract or LC wording specifically closes that gap.

A supplier COA is "what the seller's own lab measured." A third-party pre-shipment inspection certificate — issued by an independent accredited inspection firm with no stake in the transaction outcome — is "what an unrelated accredited party verified." Neither is inherently wrong; they answer different questions. If independence matters to your risk tolerance, that has to be written into the contract and, if applicable, the LC — it doesn't happen by default.

How do you verify a disputed COA?

When a result is genuinely in question, two verification paths exist, and they can run together.

The first is the lab's own retained or reference sample — the portion held back from the original batch that generated the COA, re-tested if there's a dispute about the original figures. The second is an independent sample drawn at receipt, following a recognized manual-sampling standard — ISO 3170 (Petroleum liquids — Manual sampling) or ASTM D4057 (Manual Sampling of Petroleum and Petroleum Products) — then tested at a lab accredited to ISO/IEC 17025:2017. Which sampling standard governs is a contract question; confirm which one your contract or LC specifies before you draw anything.

ISO/IEC 17025:2017 accreditation is scope-specific — it accredits a lab's competence for a defined set of methods, not the lab as a blanket authority. Check that the specific test in dispute falls inside the issuing lab's accredited scope; a lab can be 17025-accredited overall while a particular method sits outside what that accreditation covers. And accreditation says nothing about whether the product meets an OEM approval — that's a separate licensing status, verified independently (the API EOLCS directory is one example of where that kind of status is checked, not something Altonex Global confirms on a supplier's behalf).

When the COA, the TDS, and the purchase order disagree with each other, the agreed contract specification governs — and, where a letter of credit is involved, the spec stated in the credit governs. The TDS and COA are supporting, descriptive documents, not the controlling ones. This is general trade practice, not legal advice; for a specific dispute, your bank or counsel is the right party to consult.

Where does Incoterms fit into your inspection window?

One line, because it shapes your practical window rather than the COA content itself: the Incoterms 2020 delivery term on your contract determines where and when risk transfers, and that in turn shapes when inspection is practically meaningful. Under FOB, CFR, or CIF, risk transfers on board at the load port, which pushes serious inspection toward the loading stage. Under DAP, DPU, or DDP, risk transfers at the named destination, which extends your practical inspection window through to arrival. Incoterms allocate risk and cost — they don't grant inspection rights on their own; those live in the contract.

Goods just arrived — what's your next move?

Four outcomes exist at goods-inwards: accept, conditionally accept with a documented minor deviation, quarantine-hold, or reject. There's no universal percentage-deviation trigger that decides among them — thresholds are contract-specific, and inventing one here would be worse than useless.

What isn't optional is the paper trail. Generate a written non-conformance note every time something doesn't match: what was received, what was expected, what was actually found, and the disposition chosen. That note is what supports any later claim against the supplier, any LC discrepancy notice, or any insurance step — without it, a verbal disagreement six weeks later has nothing to stand on.

Forwardable asset: COA acceptance checklist

Save or route this to your goods-inwards team before the next shipment lands.

Columns worth building into your own template: parameter, method stated on the COA, COA batch value, agreed spec min/max, pass or flag.

Frequently asked questions