COA & Laboratory Testing Explained
A Certificate of Analysis is genuinely useful evidence — and routinely over-read. This is a guide to what analytical testing actually establishes, what it doesn't, and how to tell the difference when you're looking at a real one.
What Is a Certificate of Analysis?
A Certificate of Analysis (COA) is a document issued by a laboratory reporting the results of specific tests performed on a specific sample. It's a record of what was tested, how, and what was found — not a general-purpose stamp of approval. A COA is only as meaningful as the tests it actually lists.
- Laboratory
- [Illustrative Lab Name]
- Sample ID
- EX-000000
- Batch / Lot No.
- LOT-EXAMPLE-01
- Test date
- YYYY-MM-DD
- Analytical method
- HPLC-UV (example)
- Identity
- Consistent with reference standard (example result)
- Purity (HPLC, area %)
- 99.X% (example result)
- Content / Assay
- [Not shown unless separately tested]
This mock-up shows typical fields, not typical numbers — a real COA's content depends entirely on what that lab actually tested.
Identity vs. Purity vs. Quantity
Per ICH Q6A (the international guideline on drug substance specifications), identity, assay (content), and purity/impurities are three separate, universal categories of test — not three words for the same thing.
Identity
Confirms the sample's chemical identity — typically via a method like mass spectrometry or infrared spectroscopy, compared against a known reference. Answers "what is this," not "how much" or "how clean."
Purity
Usually reported as a percentage from an analytical method like HPLC — a relative measure of the target compound against detected impurities, under that specific method's conditions.
Quantity / Content
An absolute measurement (e.g. milligrams), requiring a quantitative assay against a reference standard of known concentration — a different test to either of the two above.
Why "99% purity" isn't "exactly the labelled amount"
A 99% purity result from an HPLC area-percent calculation says that 99% of the detected peak area corresponds to the target compound — relative to whatever else that method detected. It says nothing by itself about whether the vial contains 5mg, 4.2mg, or 6mg of that compound. Confirming the labelled quantity requires a separate quantitative assay, not just a purity figure. These are routinely conflated — they are not the same test.
Understanding HPLC
High-Performance Liquid Chromatography — the method behind most purity figures you'll see on a COA.
Conceptually, HPLC works by pushing a liquid sample through a narrow column packed with material that different molecules pass through at different speeds. Each compound in the sample takes a characteristic amount of time to come out the other end — its retention time. A detector records each compound as it exits, producing a peak: the bigger and more concentrated that compound, the larger the peak's area.
How area-percent purity is typically calculated: the target compound's peak area is divided by the total area of all detected peaks. This is fast and doesn't require a reference standard — but it rests on an assumption worth knowing: that every compound present responds to the detector equally per unit of mass. That assumption doesn't always hold, which is one reason the same sample can report different purity figures under different methods.
Why methodology and conditions matter: column type, detection wavelength, and run method all affect what shows up and how it's measured. A compound invisible to one detection method isn't necessarily absent — it may simply not respond to that particular detector.
Limitations: HPLC only reports what it's set up to detect. It won't flag a compound it wasn't configured to look for, and area-percent purity, as above, is a relative figure, not an absolute content measurement.
Understanding Mass Spectrometry
What it measures
The mass-to-charge ratio (m/z) of ionised molecule fragments — essentially, how heavy the molecule and its characteristic breakdown pieces are.
Why that confirms identity
A compound's molecular mass and fragmentation pattern act like a fingerprint. Matching that fingerprint against a known reference is strong identity evidence.
What it doesn't tell you
Identity confirmation isn't quantity confirmation. Mass spec can confirm what a compound is without establishing exactly how much of it is present in a given vial.
Why methods are often paired
HPLC and mass spectrometry are frequently used together precisely because they answer different questions — separation/relative quantity versus identity confirmation — not because either alone is sufficient.
What Does "99% Purity" Actually Mean?
What it may tell you
- That, under the method used, roughly 99% of the detected peak area corresponds to the target compound.
- That the lab didn't detect large amounts of other substances using that method.
- A reasonable signal of manufacturing/synthesis quality control, if the method is disclosed and credible.
What it does NOT automatically tell you
- The exact quantity of compound in the vial — that needs a separate assay.
- That nothing undetectable to that method is present.
- That every vial in the batch matches this exact figure.
- That the product is safe, sterile, or suitable for any particular use.
- Anything about a different batch, even of the "same" product.
Batch Traceability
A batch (or lot) number identifies a specific production run. A meaningful COA is tied to a specific sample and batch number, with a test date — and ideally a manufacturing date — so a reader can connect the document to the actual product in front of them.
The traceability problem worth watching for: if a supplier displays the same COA, unchanged, across restocks and inventory changes over an extended period, that document may no longer correspond to what's currently being sold. A COA's value comes specifically from being tied to the batch it was tested from — reused indefinitely, it stops functioning as evidence for current stock.
Traceability strengthens documentation; it doesn't eliminate a more fundamental limitation covered below — sampling.
Independent Testing
Manufacturer testing
Testing performed by whoever made the compound. Advantage: direct access to the production batch. Limitation: an inherent conflict of interest, since the same party has a commercial stake in the result.
Supplier-provided documentation
A COA passed along by a reseller, often originating from the manufacturer or a lab the manufacturer engaged. Worth checking whether it's tied to the specific batch being sold, not just "a" COA for "this product."
Independent third-party testing
A lab with no commercial relationship to the seller testing the product. Reduces (doesn't eliminate) conflict-of-interest concerns — still depends on sample handling, method quality, and which batch was actually sampled.
None of these three is automatically fraudulent or automatically reliable. Each carries different incentives and different limitations — worth evaluating on those terms rather than assuming "independent" always means "correct" or "manufacturer-provided" always means "suspect."
How to Read a COA
A practical sequence to work through on an actual document.
If a document is missing several of these, that's not proof of anything wrong — but it does mean you're reading a less complete picture than the sequence above would give you.
COA Red Flags
Reasons to look closer — not automatic proof of a problem.
No identifiable laboratoryIf you can't tell who actually performed the testing, you can't evaluate their credibility.
Missing test dateWithout one, you can't tell how current the result is.
No sample or batch identifierMakes it impossible to confirm the report matches the product in front of you.
Report doesn't match the batch being discussedA mismatch between the document and what's actually for sale.
Analytical method unclearWithout knowing the method, you can't judge what the result can and can't tell you.
Purity presented as proof of exact quantityThe conflation this whole page is about — see Section 5.
Cropped screenshotsRemoving context can hide exactly the fields that would let you judge the result properly.
Results shown without enough method detailA number with no context about how it was produced is hard to evaluate.
Same report reused indefinitelySee the traceability problem in Section 6.
What a COA Cannot Prove
- Safety for human use.
- Clinical effectiveness.
- Regulatory approval of any kind.
- Sterility — unless a sterility test (e.g. USP ⟨71⟩) was actually performed and reported.
- Endotoxin status — unless specifically tested (e.g. USP ⟨85⟩, the Bacterial Endotoxins Test).
- The exact quantity in a vial — unless a quantitative assay, not just a purity figure, supports it.
- That every unit in the batch is identical to the tested sample.
- That future batches will share the same analytical characteristics.
Testing Beyond Purity
Depending on the material, several other test categories may exist. "Tested" means nothing on its own — the question is always: tested for what?
Content / Assay
Quantitative measurement of how much compound is actually present — distinct from purity, as covered above.
Water / Moisture
Typically measured by Karl Fischer titration (USP ⟨921⟩) — relevant because moisture can affect stability.
Residual Solvents
Checks for leftover solvents from manufacturing, governed by guidelines such as ICH Q3C.
Endotoxin Testing
Specifically detects bacterial endotoxins (USP ⟨85⟩) — not implied by a purity or identity result.
Sterility Testing
A dedicated test (USP ⟨71⟩) for the absence of viable microorganisms — a different question entirely from chemical purity.
Microbial Testing
Broader screening for microbial contamination, separate from endotoxin-specific testing.
Not every test category applies to every research material, and a reasonable COA won't necessarily include all of them. The point isn't that every test is always required — it's that "tested" is only a meaningful claim once you know specifically what was tested.