HPLC Peak Integration: How a Purity Percentage Is Actually Calculated (2026)

The purity percentage on a Certificate of Analysis feels like a measured quantity — as if an instrument looked at the vial and read out "98.4% pure." It did not. That number is calculated from the geometry of a chromatogram, and the calculation involves real choices that can move the result by more than a percentage point. Our visual chromatogram guide covers how to spot a fake or low-quality trace; this guide goes the other direction and explains how a legitimate number is actually computed — and why understanding the math makes you a far harder buyer to mislead.

This is research-use educational content. Nothing here is a dosing recommendation or human-use claim.

The number is a ratio of areas, not a measurement of mass

Start with the core fact that surprises most people: HPLC purity is almost always reported as area-percent. The detector traces a signal over time, every component shows up as a peak, and the software integrates — measures the area under — each peak. Purity is then the target peak's area divided by the sum of all peak areas, times 100.

A 98.5% result means the main peak accounts for 98.5% of the combined integrated area of everything the detector saw. It does not mean the vial is 98.5% peptide by mass. The two usually track closely, but they are not the same statement, and the distinction matters: area-percent assumes every component absorbs UV light similarly, which is a reasonable approximation for peptides and their close relatives but is still an assumption baked into the number.

Why area, not height

A natural question: why integrate area at all, rather than just compare peak heights? Because height ignores width, and width carries information.

Imagine two peaks of identical height. One is a sharp, narrow spike; the other is a broad, low hill that happens to reach the same apex. The broad one contains far more material — its area is much larger — even though they are the same height. Reporting purity from height would systematically undercount broad components and overcount sharp ones. Area integration captures the full footprint of each component, which is why it is the standard basis for the calculation. Peak height still gets reported sometimes, but as a secondary descriptor, not the basis for purity.

Where the choices live: baseline, peak bounds, and overlap

Here is the part most COAs never show and most buyers never think about. Integration is not automatic and objective — it is software guided by settings, and three of those settings can change the answer.

Baseline placement. Before the software can measure area, it has to decide where "zero" is for each peak — where the baseline runs underneath it. On a flat, quiet trace this is easy. On a trace with drift or noise, the analyst (or the software's algorithm) chooses where to draw the baseline, and a baseline drawn slightly high or low changes every area it sits under.

Peak start and end points. The software marks where each peak begins and ends. For a clean, well-resolved peak this is unambiguous. For a peak with a dragging tail, deciding where the tail stops and the baseline resumes is a judgment call — and that judgment determines how much area gets attributed to the peak versus discarded as baseline.

Splitting versus merging overlapping peaks. When two peaks overlap, the software has to decide how to divide the shared region between them. Drop a vertical line at the valley between them? Fit and subtract modeled peak shapes? Each approach assigns the overlap differently, and the impurity peak's area — hence the purity number — shifts accordingly.

The consequence is the single most useful thing to internalize here: the same raw data can yield different purity numbers depending on integration settings. This is not fraud; it is the normal latitude of chromatographic integration. It is also exactly why the visible chromatogram and the stated method matter more than the headline figure. A number with no trace behind it hides every one of these choices.

Resolution: can the method even separate the impurities?

A purity number is only meaningful if the method actually resolves the impurities it is supposed to count. Resolution quantifies how completely two adjacent peaks are separated. High resolution means two compounds come off the column as cleanly distinct peaks with a clear valley between them. Low resolution means they overlap, and the integration software is left guessing where one ends and the next begins.

The failure mode this creates is important: when an impurity is poorly resolved from the main peak — eluting so close that they partially merge — the impurity can be swept into the main peak's integrated area. The result is a purity number that is too high, because material that should have been counted as impurity got counted as target. A method that cannot resolve a known related impurity from the target is, for that impurity, blind — and the purity figure inherits that blindness. This is one reason method validation, which includes demonstrating that the method resolves likely impurities, is so central to a credible result (covered in our look inside independent labs).

Tailing and peak symmetry: a quiet inflator

The tailing factor describes peak symmetry. A perfectly symmetric peak scores 1.0; a peak that drags out on its trailing edge scores higher. Tailing is partly a consequence of column condition and method, and a modest amount is normal.

Why a buyer should care: a badly tailing main peak can swallow a small impurity eluting just behind it. The tail of the big peak and the small impurity blur together, the software can't cleanly separate them, and the impurity's area gets absorbed into the main peak — inflating apparent purity. So a high tailing factor is not just an aesthetic flaw; it is a signal that the integration near the trailing edge may be hiding something. When a chromatogram shows a clean, symmetric main peak, the purity number rests on firmer ground than when the main peak smears into its baseline.

Putting the math to work as a reader

None of this requires running an instrument. It changes how you read a COA:

• Treat the purity percentage as a calculated ratio of peak areas, not a direct measurement — and therefore as something that depends on visible, checkable inputs.
• Insist on seeing the chromatogram, because the integration choices that produced the number are only auditable if the trace is present.
• Glance at peak shape and separation: a clean, symmetric, well-resolved main peak supports its number; a broad, tailing, or poorly separated peak undermines it regardless of how high the percentage reads.
• Remember that an impurity hidden under a tailing or poorly resolved main peak makes the number too generous, not too harsh — the error runs in the flattering direction.

For where this fits into actual sourcing and which compounds justify close reading, see the catalog entries for BPC-157 and semaglutide, the research methodology page, and our companion piece on what the impurity peaks themselves reveal.

Bottom line

A purity percentage is the output of a calculation — area-percent normalization over a set of integrated peaks — and that calculation rests on choices about baseline, peak boundaries, and how overlapping peaks are divided. Resolution and tailing decide whether those areas can even be measured cleanly, and both failure modes tend to inflate the apparent number by hiding impurities under the main peak.

The headline figure is real, but it is downstream of the chromatogram, not above it. Read the trace, check that the main peak is resolved and symmetric, and treat any purity number with no visible chromatogram behind it as a claim about a calculation you were never allowed to see.

For laboratory research use only. Not for human consumption.

Related guides:

• Reading an HPLC Chromatogram — spotting fake and low-quality traces
• What Is HPLC? — the method from the ground up
• Peptide Impurity Profiles Explained — what the impurity peaks mean
• Inside Third-Party COA Labs — method validation and why it matters