How to read a certificate of analysis

Start with the sample, not the headline

A certificate of analysis, or COA, summarises tests performed on a particular sample or batch. It is not a universal seal of quality. First locate the sample identifier, batch number, laboratory, test dates and methods. Check that these details match the material being discussed. A report for a different batch cannot establish the quality of the batch in front of you.

Look for actual measured results alongside acceptance limits. “Pass” alone leaves important questions unanswered: what was measured, with which method, and against what specification? Sampling and storage also matter. Even a competent laboratory can only describe the sample it received.

HPLC purity: a percentage with boundaries

High-performance liquid chromatography separates components before a detector measures their signals. A reported purity of 99% commonly means the main peak accounts for 99% of the integrated peak area under the stated method. It does not automatically mean 99% of the powder’s mass is the intended substance.

Different compounds can produce different detector responses. Some impurities may be invisible at the selected wavelength, fall outside the integration window, or overlap the main peak. Read the chromatogram, detector conditions and integration notes. Ask whether the number is area percent or a calibrated quantitative assay; those are different measurements.

Mass spectrometry: evidence for identity

Mass spectrometry measures mass-to-charge ratios. The expected signal, isotope pattern and, where appropriate, fragmentation can support an identity assignment. Read the calculated and observed values and the stated ion or charge state. An unexplained mismatch deserves investigation rather than a rounded-away explanation.

A matching mass alone is not a complete identity proof. Different structures can share the same mass, and mixtures can complicate interpretation. Identity is strongest when complementary methods agree. Mass spectrometry does not, by itself, certify purity, concentration or safety.

Water and endotoxin: separate questions

Water content, often measured by Karl Fischer titration, can affect the relationship between gross sample weight and the amount of dry material. “Loss on drying” can include volatile substances other than water. Check which method was used and whether assay results are reported on an as-is, dried or anhydrous basis.

Endotoxin testing measures bacterial endotoxin under a specified method, usually with results in endotoxin units per mass or volume. Read the units, detection limit and any interference or recovery checks. “Not detected” means below that method’s detection limit, not absolute zero. An endotoxin result is not a sterility test; neither substitutes for a complete suitability assessment.

Red flags on an unreliable report

Watch for missing batch identifiers, impossible date sequences, inconsistent units, cropped chromatograms, unexplained peak removal, and claims unsupported by the listed tests. Identical traces across supposedly different batches warrant questions. A logo or a polished PDF does not authenticate laboratory work.

Verify the report number through the laboratory’s independently established contact route, rather than trusting contact details printed only on the document. Ask for the full report and relevant method details. Laboratory accreditation is useful context, but check that its scope covers the particular test; it is not a guarantee about every claim on a page.

Read a COA as a set of bounded measurements. Good evidence tells you what was tested, how it was tested, and what remains unknown.