“98% by HPLC” is the most quoted and least examined phrase in ingredient procurement. It describes a detector output at a point in time. It does not describe what the other two percent is, and it does not describe what else is in the drum.
Area percentage is a relative number
Area normalisation adds up every peak the detector sees and reports the analyte as a share of that total. It assumes two things: that all peaks are related to the analyte, and that they respond to the detector in proportion to their mass. Neither assumption is safe with a UV detector at 254 nm.
An impurity with no chromophore at that wavelength is invisible. A co-eluting isomer can be counted as product. A salt or a carrier does not appear at all. The number can be high and the material can still be diluted, contaminated or misidentified.
Four things hiding behind the percentage
- Carriers and diluents. A 98% extract diluted 1:5 with maltodextrin can still be reported as 98% if the assay is run on the concentrate before blending — or if the assay is calculated on the marker rather than the whole powder. Ask for the dilution statement and the as-is assay.
- Water. Karl Fischer moisture is not part of the HPLC area, but it is part of the weight you buy. A 98% HPLC result with 6% water is not 98% of what arrives.
- Isomers and related substances. This is where “purity” is actually defined. Ask for the impurity table, not a single number: which peaks were identified, which were not, and what the reporting threshold was.
- Residual solvent and ash. Neither appears in a UV chromatogram, and both count against your formulation weight.
The three questions that fix the ambiguity
What is the method? Column, mobile phase, wavelength, flow, run time and standard. A method reference such as “in-house” is not a method.
What basis? Dried, anhydrous or as-is. This alone changes the effective actives you receive by several percent.
What is the impurity specification? Any single unspecified impurity above 0.1%, total impurities, and the identity of the largest two. A supplier who cannot answer this has not characterised the material — they have run a chromatogram.
What a written specification looks like
Both of our P0 peptide pages show the full release table rather than a headline figure: Tirzepatide ≥ 98.0% HPLC with peptide content ≥ 85% and total impurities ≤ 2.0%, and Copper tripeptide-1 (GHK-Cu) ≥ 99.0% HPLC with copper content bounded to 2.4–3.0%. Every field is released against, not estimated. The same logic applies across the catalogue.