Two Analysts, One Chromatogram, Two Purity Figures
Treating 98.4 percent and 98.7 percent as meaningfully different values is nearly always an error. Hand one chromatogram to two analysts and the purity figures that come back need not agree, even though the recorded trace is identical for both of them. Peak areas are not read off the paper; they are constructed. Judgment enters the measurement at that point, and none of it is visible on the finished certificate.
The shoulder that may or may not be a peak
Start with the most consequential decision. When an unresolved impurity shows itself as a bulge on the flank of the main peak, somebody has to decide whether that bulge becomes its own entry or gets swallowed by the peak it sits on. Swallowing it always pushes purity upward. This rarely amounts to dishonesty – integration software applies a default, and the default is frequently allowed to stand without anyone examining the result.
Worse is the species so poorly resolved that no bulge appears at all. Every convention in use will credit that material to the target. The techniques for finding it are covered in co-elution.
Two ways to divide a valley
Where a pair of peaks overlap only partly, their shared area must be apportioned. Two conventions are standard, and they do not return the same numbers.
| Drop-line | A vertical is dropped from the valley floor to the baseline, cutting the overlapping region in half. Easy to apply, and it reliably credits a slice of each peak to its neighbor. |
| Valley-to-valley | The baseline is run from one valley to the next, so each peak is treated as though perched on the tail of the one beside it. The usual effect is to reduce the area awarded to the smaller peak. |
Neither convention is right in any absolute sense; both are agreements, and any method worth the name states which agreement it follows. With a shoulder riding on a main peak, that choice alone can shift a reported purity by a noticeable fraction of a percent.
Where the software puts the floor
Underneath all of this sits the baseline, which the software has to project beneath each peak before any area exists. Across a flat, quiet stretch of trace nobody argues about it. Under gradient conditions the trace climbs as organic content increases, and now the projection is a genuine choice.
Put that floor a little too high and the area comes out small; a little too low and it comes out large. For a tall, cleanly separated main peak the distortion hardly registers. For a minor impurity hovering around the reporting threshold it settles the question of whether the impurity is recorded at all. Those thresholds are the subject of detection and quantitation limits.
Why the trace beats the number
Among everything contributing uncertainty to a stated purity, integration variability ranks near the top, which is the arithmetic behind the caution at the start of this piece. Measurement uncertainty sets that arithmetic out.
It is also the best reason to publish a chromatogram instead of a percentage. Nobody can re-integrate a number. A trace is open to reworking by anyone who questions the conclusion drawn from it.
Outside integration’s reach
Proportions are what integration governs, not identity. Whether the mass observed corresponds to the mass calculated for the claimed sequence is untouched by any of it. So are sterility and endotoxin, which are distinct tests lying wholly outside what a chromatographic method can address. That boundary is drawn in endotoxin, sterility and the limits of research grade.
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.
Third-party HPLC and purity tested, matched to certificates by cap and crimp colour.
Research-grade peptides dispatched from inside the USA, usually within 1 business day.
BPC-157 10mg
NAD+ 1000mg
MOTS-c 10mg
Selank 10mg