98.4 Percent Is a Range Reported as a Point

Two certificates arrive for what is nominally the same compound. One reads 98.4%, the other 98.9%, and the order goes to the second. That choice was made on noise. Purity gets quoted to one decimal place across this industry, the precision such a figure implies is hardly ever earned, and reading those two documents as a disagreement ranks among the commonest errors in supplier comparison.

Four places the number shifts before anyone reads it

What a certificate calls purity is an area ratio lifted off an integrated chromatogram. Every stage feeding that division contributes a wobble of its own, and the wobbles do not cancel out.

Begin at the injector. Inject one sample repeatedly and the areas will not match; a relative standard deviation in the low tenths of one percent is ordinary on a healthy system, and that is a floor rather than a typical case. Integration adds the next layer: somebody or something must decide where the baseline runs beneath a tailing peak, and that decision moves the area. Hand one trace to two capable analysts and their answers routinely diverge, by more as the peak grows more asymmetric.

Detector response comes third and is the least intuitive. An area percentage quietly assumes every species answers the detector identically at whichever wavelength was chosen, which is false. An impurity without the target’s chromophore gets under-counted; one carrying a stronger chromophore gets over-counted. Because the peptide bond dominates absorbance around 214 nm, the assumption largely survives there. Shift detection to 280 nm and it can collapse outright. Resolution supplies the fourth: whatever the method fails to separate is tallied as target, so this error pushes a result upward instead of scattering it both ways. Co-elution covers that mechanism.

How wide the band actually is

Stack the four together and a figure from a properly run method carries roughly this much doubt:

Accreditation obliges a laboratory to be capable of stating that quantity. Certificates accompanying research peptides mostly never do. The working rule follows: two documents separated by a few tenths of one percent are telling you the same thing, and only a gap of several percent is worth raising with whoever issued them.

Compare the documents, not the decimals

Sorting vendors by the final digit measures how precisely their integration software rounds, not how good the contents of the vial are. Worse, two numbers generated on different columns at different wavelengths under different methods were never on a common scale to begin with, a problem that exists before uncertainty enters.

Attention repays itself better on what each document contains: whether a wavelength is stated, whether the chromatogram is reproduced rather than summarized, whether net peptide content appears, and whether the batch identifier on the page matches the vial in hand. Why certificates of analysis disagree works through the mechanisms, net peptide content explained takes up the figure that usually goes missing, and choosing a research peptide supplier in the USA sets out what to ask for.

A missing interval is not a verdict

Uncertainty is neither error nor defect. Where a laboratory publishes one, it demonstrates that it understands the method it ran. Where none appears, nothing is established about the result beyond its precision going unstated. The scope of the exercise stays narrow either way: composition is the whole of what a purity certificate describes. Sterility is not in it, endotoxin is not in it, and neither is any statement of suitability for a person or an animal.

All material is supplied for laboratory research use only. It is not a drug, not a supplement, and not for use in humans or animals.