Chemineral — Petrochemicals from Oman
Chemineral
All Insights
Guide13 min read18 Nov 2026

How to Read a Certificate of Analysis: What Buyers Should Check Before Accepting a Cargo

What a certificate of analysis contains, how it differs from a specification and a data sheet, which fields to check, how to spot weak or suspicious certificates, and how to compare the result with your own test.

By Chemineral Trade Desk
Quality control laboratory bench with sample jars

Almost every shipment of chemicals, fertiliser or polymer comes with a certificate of analysis. Many buyers file it as soon as it arrives, as proof that someone tested the cargo. That is a missed chance, because the certificate is the supplier's own statement of what is in the shipment, and it is the document a buyer will lean on if the cargo later turns out to be wrong. A certificate that is vague, incomplete or inconsistent is a warning, and one that is clear and traceable is a solid base for acceptance and for any claim.

This guide explains what a certificate of analysis is, what each part tells you, how to check it against your contract, what weak certificates look like and how to compare it with your own results. It is general guidance. The specification in your contract and the test methods named in it decide what a certificate must show. For the other documents in a shipment, see our guide to export shipping documents.

What a certificate of analysis is

Answer first: a certificate of analysis, usually shortened to COA, is a document that reports the results of tests on a specific batch or lot of a product, set against the limits in its specification, and is issued by the producer, the supplier or a laboratory.

It answers a narrow question: for this batch, what were the measured values? That makes it different from three documents it is often confused with.

  • The specification states the limits a product is supposed to meet. It describes what should be true of any batch. See our guides to the urea specification and base oil specification.
  • The technical data sheet describes the product's typical properties and uses. Its numbers are typical values, not guarantees for a batch.
  • The safety data sheet describes hazards and handling. It does not report quality. See how to read a safety data sheet.

The COA is the only one of the four that ties results to a particular lot. It is evidence, and its value depends on who produced it, how the sample was taken, which methods were used and whether it matches the cargo.

Who issues it

  • The producer's laboratory. Most common. The producer tests its own output and issues the certificate with the shipment.
  • The supplier or trader. Often a copy of the producer's certificate, sometimes re-issued on the trader's paper. Ask for the original.
  • An independent laboratory or inspector. Tests a sample drawn from the cargo, and issues its own certificate. This carries more weight in a dispute, because the tester has no stake in the result. See our guide to choosing an independent inspector or laboratory.

A cargo can have several certificates: the producer's batch certificate, an independent loading certificate and a discharge-port result from your own laboratory. They should tell the same story.

Shelves of sealed retained samples
Shelves of sealed retained samples

The parts of a good certificate

Identification

  • Product name and grade, matching your order and the specification.
  • Batch, lot or tank number, and how it relates to the cargo, for example which vessel, tank or containers it covers.
  • Production or test date. Old dates, or dates that do not fit the shipment timeline, deserve a question.
  • Producer or manufacturing site, so you can match it to the approved source.
  • Quantity the certificate covers.

Tests and limits

  • A list of parameters, each with its unit.
  • The specification limit for each, as a minimum, a maximum or a range.
  • The result for this batch.
  • The test method used, by name or standard reference.
  • A pass or conforming statement, sometimes per line and sometimes overall.

Sampling

A good certificate says how and when the sample was drawn, and by whom. A sample taken from a tank before blending, from the end of a production run or from a single bag says less about the cargo than one drawn during loading. See our guide to representative sampling.

Authorisation

  • The issuing laboratory or company named, with contact details.
  • A signature or stamp from an identifiable, authorised person.
  • Where relevant, a statement of the accreditation of the laboratory and the scope of that accreditation. Accreditation to a recognised laboratory standard means the lab is assessed for technical competence, but it applies to specific tests, so check that the tests on your certificate are within the scope.
  • A certificate number, so it can be verified and referred to.

How to check it, step by step

Answer first: compare the certificate with the contract specification line by line, check that the identification matches the cargo and the other documents, and look for gaps.

1. Match the product and batch

Does the product name and grade match the order, the bill of lading and the invoice? Does the batch or tank number match what was shipped? A certificate for a different grade, or a lot that was not loaded, does not prove anything about your cargo.

2. Compare every parameter with the specification

Take the contract specification and go through it in order.

  • Is every parameter in the specification on the certificate? A certificate that omits a contractual parameter is incomplete, whatever it says about the others.
  • Is the limit on the certificate the same as the contract limit? Some certificates quote the producer's own, looser limits.
  • Is each result inside its limit?
  • Are the units the same, and the same basis, for example per cent by weight, dry or as received?

3. Check the test methods

The method matters as much as the number. A moisture figure from one method can differ from another. The method on the certificate should be the one named in the contract, or an agreed equivalent. If the contract names a method and the certificate uses a different one, ask why and whether the results are comparable. For instrumental tests, the certificate may name the standard but not the conditions, so ask the laboratory if the conditions matter.

4. Look at how results are presented

  • Real numbers, not just "pass" or "conforms". A certificate that only says "passes" gives you nothing to compare.
  • Sensible precision. Results reported to many more digits than the method can support, or all exactly at the limit, look manufactured.
  • Natural variation. Real batches vary. A series of certificates from the same source with identical results to the last digit is a red flag.
  • Consistent rounding and units.

5. Check dates and logic

Is the test date before the shipment date, and close enough to be meaningful? Does the production date make sense for the cargo, given storage? For a long storage period, consider asking for a retest. See shelf life and storage time.

6. Check the authorisation

Is the issuer identifiable, and a real laboratory or producer? Can you verify the certificate number with the issuer? Is the signature from a named person with a title? A certificate with no issuer name, a generic logo or a blurred stamp should be verified before you rely on it.

7. Cross-check with other documents

Weights, marks, vessel name, product description and dates should agree across the COA, bill of lading, packing list and invoice. Mismatches are classic reasons for a bank to refuse documents under a letter of credit. See our letter of credit guide and the guide to bills of lading.

What a certificate does not tell you

Answer first: a certificate reports tests on a sample. It does not guarantee that every part of the cargo is the same, that the cargo was not changed afterwards, or that the tests you care about were done.

  • Representativeness. A good result on a well-chosen sample suggests, but does not prove, that the cargo is uniform.
  • After-test changes. Moisture can be gained in transit, material can be contaminated during loading and products can degrade in storage. The certificate does not cover what happens after sampling.
  • Untested parameters. If a property that matters to you is not on the certificate, there is no evidence either way. Ask for it, or add it to the specification.
  • Compliance in your market. A certificate shows technical results. It does not by itself show that a product meets food, feed, drinking water or other regulatory requirements. Check those separately.
  • Origin and sanctions status. Those need their own documents. See certificates of origin and sanctions compliance.
Bulk carrier loading with a sampling station on the conveyor
Bulk carrier loading with a sampling station on the conveyor

Weak and suspicious certificates

Treat these as reasons to ask questions before accepting the cargo or paying:

  • No batch or lot number, or one that does not match the cargo.
  • No test methods, or methods that do not match the contract.
  • Missing contractual parameters.
  • Results shown only as "pass" or "OK".
  • Results that are identical across batches, or all exactly at a limit.
  • An unknown issuer, no accreditation where you expected it, or contact details that do not work.
  • Layouts that look edited, with misaligned fonts, different typefaces in the same table, or pasted logos.
  • A certificate dated before the product could have been produced, or long after the shipment.
  • A certificate issued by the trader with no reference to the producer's original.
  • Reluctance from the seller to share the original or to allow independent sampling.

None of these proves fraud. Many are simple carelessness. But a seller who cannot answer straightforward questions about a certificate is telling you something. Our guide to verifying a petrochemical supplier covers wider checks on the counterparty.

Comparing the certificate with your own results

Answer first: differences between laboratories are normal. What matters is whether the difference is within what the method can reproduce, and whether the cargo meets the contract limit under the agreed method.

Expect some difference

Two laboratories testing the same sample will not report identical numbers. Each method has a repeatability, the expected spread for the same laboratory, and a reproducibility, the expected spread between laboratories, and the second is usually larger. Before you call a result a failure, look at the gap and compare it with what the method allows. A result a little outside the limit, within the method's reproducibility, is a conversation. A result well outside it is a problem.

Agree the rules in advance

Good contracts say:

  • Whose result is final: the producer's certificate, the independent loading result or the discharge result.
  • Which laboratory settles disputes, and which sample it tests.
  • How differences are treated, for example by averaging, or by sending a retained sample to an umpire laboratory.
  • The time limit for raising a claim.

See key clauses in a sale contract and quality claims and disputes.

Test what matters, and keep samples

You do not need to repeat every test. Choose the parameters that matter most to your process and those most likely to change in transit, such as moisture. Keep sealed retained samples from loading and discharge, because they let you settle a disagreement with evidence. Chemineral can arrange SGS or Intertek pre-shipment inspection, which includes sampling and, where agreed, testing.

Certificates for different products

The principles are the same, but the important lines differ.

  • Urea and other fertilisers: nitrogen content, moisture, size distribution, biuret and crushing strength. See our storage and caking guide.
  • Polymers: melt flow, density, and for some grades additive-related tests, and visual contamination. See polymer resin quality checks.
  • Bitumen: penetration, softening point, flash point, viscosity and others set by the specification. See bitumen quality tests.
  • Base oil: viscosity, viscosity index, pour point, flash point and colour.
  • Caustic soda: assay and impurities such as carbonate and chloride.
  • Sulphur: purity, ash, moisture and acidity. See the granular sulphur buyers' guide.
  • Calcium carbonate: calcium carbonate content, whiteness, particle size and moisture.
  • Activated carbon: adsorption indices, hardness, moisture and ash.

A worked example, for illustration only

The product and numbers below are invented to show the process. They are not real results.

A buyer receives a producer's certificate for a cargo before shipment. The contract specification has six parameters, each with a method. The buyer lays the two side by side.

  • Parameters. The certificate lists five of the six. The missing one, a moisture limit, matters to the buyer's process. The buyer asks for it before the cargo sails, and the seller supplies a result on a revised certificate.
  • Limits. Two limits on the certificate are looser than the contract limits, because the producer quoted its own range. The buyer notes that the results still fall inside the tighter contract limits, but writes to confirm that the contract limits apply.
  • Methods. One test used a different method from the one in the contract. The buyer asks the laboratory whether results from the two methods are comparable, and agrees to treat the contract method as final, with the independent loading sample tested by that method.
  • Dates and batch. The batch number matches the bill of lading draft, and the test date is a week before loading, which fits the production schedule.
  • Plausibility. Results differ slightly from the previous three certificates from the same producer. That is what real batches look like.

At loading, an independent inspector draws samples during loading and the laboratory reports a moisture value slightly above the producer's figure. The gap is smaller than the method's reproducibility, and both are inside the limit. The buyer accepts the cargo, keeps the retained samples and files all three documents together.

Now change one thing: suppose the loading result had been well outside the limit and the producer's certificate had shown no moisture figure at all. The buyer would have the missing parameter, the independent result and the retained sample in hand, which is exactly the evidence needed to raise a claim in time. The checks cost an hour. They are what make either outcome manageable.

What to ask for in the contract

  1. A batch COA with every contractual parameter, the limit, the result and the method.
  2. Which laboratory issues it, and that its accreditation covers the tests where required.
  3. Sample details: when, where and by whom.
  4. The original certificate, not only a trader's copy.
  5. An independent loading certificate where the value justifies it.
  6. A rule for which result is final, and how differences are handled.
  7. A retained sample, sealed, with agreed retention time.
  8. Delivery of the certificate before shipment, so that it can be checked in time, and before payment where terms allow.
  9. The right to reject or claim if the COA is missing or inconsistent with the cargo.

A short checklist

  • Product, grade, batch and quantity match the cargo and other documents.
  • Every contractual parameter is present, with limit, result and method.
  • Methods match the contract.
  • Results are numbers, plausible and within limits.
  • Sampling is described.
  • The issuer is identifiable and can be verified.
  • Dates are logical.
  • The certificate is the original, or traceable to it.
  • Independent confirmation exists where it matters.

Common mistakes

  • Filing the certificate unread.
  • Accepting "pass" with no numbers.
  • Not comparing to the contract limits, only the producer's.
  • Ignoring test methods.
  • Accepting a trader's re-typed certificate with no original.
  • Treating the certificate as proof of the whole cargo with no sampling evidence.
  • Calling every difference a failure, with no regard to reproducibility.
  • Not keeping a retained sample.
  • Raising doubts too late, after the claims period has ended.

Frequently asked questions

What is the difference between a COA and a specification?

The specification lists the limits a product should meet. The certificate of analysis reports the measured results for one batch against those limits.

Is a certificate from the producer enough?

For low-risk products it may be, together with your own sample test. For critical uses or large values, add independent sampling and testing at loading, and keep retained samples.

What should I do if my test result differs from the certificate?

Compare the gap with the method's reproducibility. If it is within that range, the results are consistent. If it is larger, ask for the retained samples to be tested by the independent laboratory named in the contract, and follow the claims process.

Why do certificates list test methods?

Because different methods can give different numbers for the same sample. The method defines what the number means, so the certificate and the contract should use the same one.

How can I tell whether a certificate is genuine?

Verify the certificate number and issuer directly with the laboratory, using contact details you have found yourself, check that the batch and dates match the cargo, and look for the warning signs listed above.

Next step

If you want to agree the certificate, test methods and inspection before a first order, send the trade desk your product, specification and destination through the contact page. We quote Suez and Cape routing side by side for shipments from Sohar and Salalah, and firm offers are valid for 72 hours.

Certificate of AnalysisCOAQualitySpecificationDocumentsGuide
Trade with us
Act on this analysis.

Our trade desk can turn the views in this report into firm offers on today's indicative prices — valid 72 hours, with complete documentation.

Request a quote
Chat with our trade desk