What a purity number does not tell you

01 · What the number actually is

A purity percentage on a peptide certificate is almost always area percent from a chromatogram: the detector traced a signal over time, peaks appeared, someone integrated the area under each, and the main peak’s share of the total was reported as a percentage.

The number is a summary of a picture. Given the picture you can see how many other peaks there are, how large each is, and where they sit relative to the main peak. Given only the summary you can see none of that.

Two lots can both report 99.1% and be materially different materials: one with a single distant impurity, one with a cluster of closely-eluting related substances the method barely separated.

Impurities eluting far from the main peak are typically solvents, reagents or unrelated contaminants — undesirable, visible, usually removable. Impurities eluting close to the main peak are typically related substances: deletion sequences, truncations, oxidised or deamidated forms of the peptide itself. Those are the difficult ones, and a bare percentage conceals them.

02 · Area percent is not mass percent

Reporting area percent assumes every species responds equally to the detector. With ultraviolet detection they do not. Response depends on what absorbs at the chosen wavelength, so a weakly absorbing species is under-represented and a strongly absorbing one over-represented.

This is not a criticism of the method. It is a caution against reading the figure as a statement about how many milligrams of what are in the vial. It is not that.

03 · A number without a method is not checkable

The same sample produces different purity figures under different conditions. A short run on a low-resolution column merges peaks a longer gradient would separate, and merged peaks report as one — inflating the number.

A certificate reporting purity without the column, gradient, detection wavelength and run time has published a result that cannot be reproduced or challenged.

Any number that cannot in principle be wrong is not a measurement. It is a claim wearing the clothes of one.

04 · A specification is not a result

“≥99%” and “99.1%” look similar and mean different things. The first is a specification — a threshold the supplier committed to. The second is a result — what this lot measured. A specification tells you what would have caused rejection. It does not tell you what the lot was.

Certificates carrying only the specification, with no measured value against it, are common. So are certificates printing the specification in the result column.

05 · What to ask for

The trace

The chromatogram image, not a transcription of its conclusion. This single request separates suppliers who ran the test from suppliers who have a document.

The method

Column, gradient, detection wavelength, run time. Enough for a second laboratory to attempt the same separation.

The lot number

Printed on the certificate and matching the vial. Explainer 04, and the check most often failed.

The date

After the manufacturing date. An undated certificate carries no information.

06 · What a purity figure does not address at all

Purity is silent on identity — whether the main peak is the intended molecule, which belongs to mass spectrometry and explainer 02. It is silent on endotoxin, residual solvents, water content, counterion and aggregation state.

Aggregation deserves specific mention. For some peptides, characterisation and aggregation are documented problems in the peer-reviewed literature. Where that is true, two laboratories reporting on the same nominal compound may not have handled identical material.

A high purity figure on a well-documented lot is a reason to take a supplier seriously. It is never a reason to take a compound seriously — that is settled by the evidence grade, and no certificate can move it.