A certificate of analysis is a document reporting what an analytical laboratory measured in a specific lot of material. That last part carries most of the weight. A COA is not a statement about a product line, a supplier, or a compound in general. It describes one batch, tested on one date, by one method.
Certificates circulate freely in this market, and a large share of them do not support the claim they are being used to make. Here is what a complete one contains, and the four failure modes worth checking for.
Anatomy of a complete certificate
A certificate that can actually be relied on identifies all of the following:
The material. Compound name, and ideally CAS number and molecular formula. Sequence, for a peptide.
The lot. A batch identifier that appears on both the certificate and the physical vial. If those two numbers cannot be matched, the certificate is describing something you cannot prove you are holding.
The laboratory. A named, contactable, independent facility — not "our lab," not a logo. Accreditation status where applicable.
The methods. Not just results. Which technique, and under which conditions: column chemistry, mobile phase, gradient, detection wavelength for HPLC; ionisation mode for MS.
The results, with their units and basis. "99.1% by RP-HPLC at 214 nm, relative peak area" is a result. "99%+ pure" is a marketing statement.
The date. Material changes over time. A certificate describes the lot as it was when assayed.
A signatory. A named person taking responsibility for the report.
The four ways certificates get faked
1. The recycled certificate
One genuine certificate, reused across many batches — sometimes for years, sometimes across different compounds entirely.
Tells: the lot number on the certificate does not match the vial in front of you. The assay date is old relative to the material's manufacture. The identical certificate image appears on multiple products, or on multiple vendors' sites — reverse image search is unreasonably effective here.
2. The drawn chromatogram
A trace produced in graphics software rather than by an instrument. These have become more common and more convincing, but they tend to be too clean.
Tells: a perfectly flat baseline with no low-level noise. No solvent front near the start of the run. Peaks that are exactly symmetrical — real peaks usually show some tailing. No integration table, no retention-time labels, no axis units. An impossibly round purity number. Real instrument output carries the software's own formatting, headers and timestamps; a bare image without any of that furniture deserves scrutiny.
3. The in-house certificate wearing third-party clothes
Testing performed by the seller and presented, through omission rather than outright claim, as independent verification.
Tells: no laboratory is named anywhere on the document. No address, no contact, no accreditation number. The phrase "third-party tested" appears in the marketing copy but no third party appears on the certificate. This is the most common failure of the four, and the easiest to check — if you cannot identify who performed the test, no one has.
4. The method-free result
A purity figure with no methodology behind it. "99% pure" answers none of the questions that determine what the number means: pure by which technique, detected how, and relative to what.
Tells: a results table with no methods section. Purity stated without a detection wavelength. No mention of what was not tested.
The question most certificates do not answer
Purity and identity are different measurements, and a certificate showing only one is incomplete.
HPLC reports how much of the detected material is a single component — a relative figure. Mass spectrometry confirms the component's mass matches the intended sequence. A certificate with a high HPLC number and no mass confirmation demonstrates that a sample is homogeneous. It does not demonstrate that the sample is the compound named at the top of the page.
Both belong on a complete certificate, and their absence is worth noticing.
What to ask a supplier
Three requests separate suppliers who test from suppliers who say they test:
- The certificate for the specific lot you are receiving — not a representative example, not the most recent one on file.
- The raw chromatogram, not a summary table transcribed into a PDF.
- The name of the laboratory, so the result can in principle be verified with the party that produced it.
A supplier who can answer all three quickly is operating a real quality process. A supplier who cannot answer any of them has not tested the material in a way that means anything.
For research use only · Not for human or veterinary use · No compound discussed here is FDA-approved for any indication