What an HPLC purity figure means

An HPLC purity figure is a ratio. It compares the signal from the main peak to the signal from everything the detector saw and the software counted. That is all it is, and most confusion about purity comes from reading more into it.

The formula

Purity (area %) = main peak area / total integrated peak area x 100

The areas come from the peak table printed under the chromatogram. Each peak's area is the space under its curve, in detector units multiplied by time.

A worked example

A peak table with four integrated peaks:

Peak	Retention time (min)	Area	Area %
1	11.42	18	0.4
2	12.87	4,431	98.9
3	13.10	22	0.5
4	15.06	9	0.2
Total		4,480	100.0

Purity = 4,431 / 4,480 x 100 = 98.9%.

Notice what decides the answer: which peaks were integrated, and where the software drew the baseline under them. A fifth peak too small to clear the integration threshold would simply not be in the denominator.

What "99% pure" says

Under the stated method, 99% of the UV signal the detector recorded, in the peaks that were integrated, sat in the main peak.

What it does not say

• That 99% of the powder is peptide. Water and counterions are invisible to the detector at peptide wavelengths. That is a separate measurement, net peptide content.
• That the main peak is the right molecule. That needs mass spectrometry.
• That impurities were all seen. Anything that does not absorb at the chosen wavelength, co-elutes under the main peak, or leaves the column after the run ends is not counted.
• That another lab would get the same figure. A different column, gradient or wavelength can move it.

Relative, not absolute

Area percent is a relative measure: it compares peaks to each other within one run. An absolute measure would compare the sample to a reference standard of known amount, and that is what an "assay" figure usually is. Both can appear on one report, and they are not the same number.

Reading the figure with its method

A purity figure is most useful when the method is printed beside it: column, mobile phases, gradient, flow rate and detection wavelength. Without that, the number is hard to compare with anything, including another lot of the same material.