Reading a Certificate of Analysis Without Taking It on Trust

A certificate of analysis arrives with almost every consignment of research material, and most of them are filed unread. That is understandable. The document looks technical, it carries numbers to two decimal places, and it appears to close the question of what is in the vial. It does not. A certificate is a summary of tests that somebody chose to run, presented in a format that somebody chose to use. Reading one properly means asking what was tested, by which method, by whom, and what was left out.

For laboratories, research groups and resellers buying at volume, that scrutiny is not pedantry. The certificate is the only paper trail connecting a physical batch to a claim about its contents. If the document cannot survive a few direct questions, the material behind it should not be assumed to either. All of this concerns compounds supplied strictly for laboratory research use.

What the document is actually claiming

A certificate of analysis reports specific analytical tests performed on a defined batch, on a defined date, using a defined sample. It is not a guarantee of performance and it is not a specification. Two documents can carry the same heading and represent entirely different amounts of work. One might report a single chromatographic run. Another reports identity, chromatographic purity, water content, residual solvents, counter ion content and appearance, each with the method named and an acceptance criterion beside it.

The useful mental model is narrow. The certificate says what was measured and what was seen, for one batch, once. Every extension beyond that is an assumption the buyer is making rather than a claim the supplier has signed.

Start with the header, not the numbers

Before reading a single figure, check the administrative details. The batch or lot number on the certificate should match the number printed on the container in front of you. Mismatches happen more often than the sector likes to admit, usually through clerical error rather than intent, and one makes the document unusable for that container regardless of how good the data look.

Then establish who produced the certificate. In-house quality control, the original manufacturer, and an independent contract laboratory are all legitimate sources, but they carry different weight and incentives. A certificate that names the testing laboratory, the analyst or approver, and the date of issue is doing more work than one carrying a logo and little else. Note the issue date. A certificate written long before the material reached your bench describes the batch as it was then, not as it is now.

Purity depends entirely on the method

A purity figure with no method attached is decoration. Chromatographic purity by area percentage reports the proportion of detectable material that eluted as the main peak at a chosen wavelength. Anything that does not absorb at that wavelength is invisible to the measurement. Water, salt and residual solvent are not counted, which is why a chromatographic purity figure and a mass balance can diverge considerably for the same sample.

Ask for the conditions: column, gradient, detection wavelength, run time. Ask whether the chromatogram itself is available. The trace shows what a single number cannot: whether the main peak is clean or shouldered, whether the baseline is stable, and whether integration has been drawn generously around a cluster of closely eluting species. Treat suspiciously round figures with care. The same value on every certificate for every line in a catalogue suggests a template rather than a measurement.

Identity is a separate question

Purity describes how much of the sample is the main component. Identity describes whether that component is the one named on the label. These are different tests and a serious certificate carries both. Mass spectrometry is the standard identity check, and the result should show the observed mass alongside the theoretical mass for the stated sequence, with the tolerance made explicit.

Reading that section means checking the arithmetic rather than accepting the presence of a number. A miscounted residue, a protecting group that was never fully removed, or an oxidised residue each shifts the observed mass in ways an attentive reader can spot. For buyers moving into bulk and trade supply, identity data matters more rather than less, since a single mislabelled batch distributed across many downstream customers is slow and expensive to unwind.

The tests that quietly go missing

Much of what a buyer needs is often absent. Common omissions include:

• Water content. Lyophilised material is hygroscopic and can hold a meaningful percentage of water, distorting any calculation based on gross weight.
• Counter ion and residual solvent. Purification leaves residues behind. The quantity varies by process, and it occupies mass a buyer may assume is product.
• Appearance and solubility observations. Cheap to record, and genuinely useful the day a later batch behaves differently.
• Endotoxin or bioburden. Relevant only to some workflows, but where relevant, absence means the test was not performed, not that it passed.

The principle is simple. Absence of a test is not evidence of a result. If a parameter matters to the work, it belongs in your own written specification, and the certificate should be checked against that rather than admired on its own terms.

Turning the document into a decision

Request certificates before ordering rather than after delivery, and ask for the actual batch document rather than a representative example. Write your own specification, keep it separate from whatever the supplier sends, and compare the two line by line. Where a material underpins a long programme, periodic independent verification is worth its few hundred pounds, if only to confirm that paperwork and physical material have stayed aligned.

Raise discrepancies in writing and keep the correspondence. A supplier that answers method questions plainly, sends chromatograms without fuss and states clearly what was not tested is showing you something more durable than any printed percentage. That behaviour, repeated across batches and years, is what confidence in documentation should rest on.