How to Read a Certificate of Analysis
What to look for on a COA, how it differs from a marketing spec, and how to sanity-check a batch without treating the PDF like a talisman.
A Certificate of Analysis (COA) is the paper that says what was in the vial when it was tested. Plenty of people read the purity line, skim the logo, and file it away. If you are trying to decide whether a batch is what it claims to be, that habit leaves a lot on the table.
What a serious COA actually lists
You want the compound named the same way the label names it, a test method that matches what you care about (HPLC for purity is common), a batch or lot number, a test date, and a lab you can look up. Names, addresses, and a way to reach the lab matter. Missing pieces do not prove fraud on their own, but they make verification harder, and that is worth noticing.
COA vs. product spec sheet
A COA is tied to one batch at one time. A generic spec sheet might only repeat marketing ranges. If the document never names a lot, never lists a test date, and reads like a brochure, treat it as background, not proof of what you are holding.
Identity and purity are different questions
Identity answers "is this the molecule we think it is?" Purity answers "how much of the sample is that molecule versus everything else?" You can fail either test in a loud way. A clean purity read on the wrong structure is still the wrong structure.
In-house testing vs. a third-party lab
In-house means the same company that sold the vial ran the assay. Third-party means an outside lab issued the report. Neither label is magic, but third-party work removes one obvious conflict of interest. If you call, ask whether they can confirm they tested the lot on the COA. Some labs will help; some will not. A flat refusal is data too.
Verifying a batch without being weird about it
Start with what is on paper: lot, date, lab, tests. Match those fields to the vial label. If something does not line up, stop and ask the vendor before you assume it is a typo. If the lab picks up the phone, keep the question narrow: did you issue this report for this lot? You are not doing detective cosplay; you are checking basic paperwork.
Signals that should slow you down
Watch for COAs that never tie to a specific lot, labs with no footprint you can find, files that look pasted together outside of normal LIMS exports, and purity that lands on the same tidy number batch after batch. Real assays wiggle. Perfection repeated forever is less convincing than a boring 98.7% next to a 99.1%.
How to read a purity line without over-reading it
Context still wins. For many research-grade peptides, an HPLC purity in the high nineties is ordinary. Pharma-grade material is often held tighter. A single number never replaces storage history, handling, or whether the identity test matches the label. Use the COA as one input, not a verdict.