Chromatography pushes a mixture through a column that different components travel through at different speeds.
Chromatography pushes a mixture through a column that different components travel through at different speeds. What arrives at the far end arrives in order, separated. Everything a purity figure says afterwards depends on that separation having worked, which is why the method matters as much as the number.
The instrument records a signal over time, producing peaks. A purity percentage is calculated by comparing the area of the target peak with the total area of all peaks. The percentage is a derived figure, and the trace behind it carries information the percentage discards.
Two components that travel at the same speed arrive together and appear as one peak. The calculation then counts an impurity as part of the target, and the reported purity is too high. Nothing in the number reveals this; only the method, or a second method, can.
Confirming a result with a method that separates on a different principle makes co-elution far less likely, because two unrelated techniques are unlikely to merge the same pair. A result confirmed twice by different means is substantially better evidence than a result measured twice the same way.
Detectors respond to particular properties, and a component the detector does not respond to is invisible regardless of how well it was separated. A purity figure is therefore a percentage of what the detector could see, which is a narrower claim than it sounds.
Deciding where a peak starts and ends sets the area, and overlapping peaks make that boundary a choice. Software automates it and analysts can override it. Two competent analysts can produce slightly different percentages from one trace, which is normal and worth knowing.
A published chromatogram lets a reader see the number of peaks, their relative sizes and whether any are poorly resolved. A seller publishing only the summary percentage is withholding the part that would let the figure be checked.
Which technique, which detector, whether a second method confirmed it, and whether the trace itself is available. A figure that cannot answer those is a number without a method attached.
The claims on this page about the technique behind most purity figures, in terms a buyer can check can be checked against sources outside this site: