A Certificate of Analysis is a statement about one production lot, made by one laboratory, using one set of methods, on one day. Everything that goes wrong when people read these documents comes from forgetting at least one of those four constraints. Read in the right order, a COA is a short and honest document. Read as a grade or a marketing badge, it becomes a way of believing things nobody actually asserted.
Start at the header, because the header decides whether the rest applies
Before a single number is worth reading, three things in the header block have to agree with the object in your hand: the material name, the lot or batch identifier, and — where printed — the pack format. If the vial says one lot and the PDF says another, stop. You are not looking at a discrepancy to be resolved by judgement; you are looking at a document for different material. Match it through COA verification or the batch verification portal rather than by eye alone, and record the result either way.
The issue date matters too. A COA describes the lot at the time of analysis. It is not a running guarantee, and it says nothing about what happened to the vial in transit or on your shelf afterwards.
The method line changes what every number means
Underneath or beside each result there is usually a method: HPLC, mass spectrometry, Karl Fischer, and so on. That line is not decoration. A purity value produced under one chromatographic method is not directly comparable with a value produced under another, because the separation conditions determine what gets resolved into a separate peak and what stays hidden under the main one. Understanding HPLC explains this in more depth, and the shorter version lives in the HPLC glossary entry.
If a number is printed with no method, record it as a number with no method. Do not supply the missing information from memory.
Three columns that look interchangeable and are not
Well-built documents keep the expected range, the measured value, and the method's lower reach in separate columns. Collapsing them is the most common reading error in a receiving log.
- The specification is the range the lot was expected to fall within. It is a rule, not an observation.
- The result is what this lot measured. It is an observation, not a promise about the next lot.
- The limit of detection or quantitation is the floor below which the method sees nothing. "Not detected" describes the instrument's reach, not the absence of the substance.
Copy each into its own field. A single merged "purity: pass" line destroys information you will want back later.
Fields that get over-read
Appearance is a qualitative note and belongs alongside your own visual check, not above it. Water or moisture content describes residual water under the stated method; it is not a storage instruction. A residual solvent panel is only as informative as the list of solvents that was actually run — a short panel is not a clean result, it is a short panel. And a missing field is a missing field. Nothing on the document converts research material into anything other than research material, which is the point made plainly in the research use policy.
Filing it so it is findable in two years
The value of a COA is realised at the moment somebody questions a result, which is usually long after the vial is gone. Store the document itself rather than a link to a portal, name the file so a stranger could match it to an inventory row, and keep it attached to the lot rather than to the order. The broader chain is described in documentation traceability, and the general practice is collected in the documentation centre. The field-by-field walkthrough in understanding COAs is worth keeping open the first few times somebody new does receiving.
FAQ
Does every lot have a published COA?
No. Records are published when they have been uploaded for that specific lot. If a record is not yet online, ask for it against the lot number rather than the SKU.
Is purity the same as assay?
Not necessarily. Where both appear, they are separate determinations made by separate methods. They can move independently, and averaging them produces a number that describes nothing.
Can I apply a COA to a second vial of the same product?
Only if that vial carries the same lot. Documents are lot-scoped, and a same-SKU different-lot substitution breaks the record silently.
Does a high purity figure change how the material may be used?
No. The catalogue is positioned research-use-only regardless of what any analytical figure says. Materials described here are supplied for in vitro laboratory research only.