Split, tailing and overlapping peaks: reading odd peak shapes

The textbook HPLC peak is a narrow, symmetrical bell. Real peaks often are not. Some odd shapes come from the sample, some from the instrument or method, and they matter because they change how the software measures area, and so the purity figure.

Tailing

Looks like: a sharp front and a long, dragging back edge.

Common causes: secondary interactions between the peptide and the column packing, such as basic residues sticking to exposed silica; a worn column; too much sample on the column.

Effect on purity: the tail can hide a small impurity eluting just after the main peak, or get cut off by the integration, depending on where the software ends the peak.

How it is measured: the tailing factor (or asymmetry factor). A value near 1.0 is symmetrical. Many methods set an upper limit as part of system suitability.

Fronting

Looks like: the reverse: a slow rise and a sharp back edge.

Common causes: too much sample, or a sample dissolved in a solvent stronger than the starting mobile phase.

Effect: as with tailing, it can mask a neighboring peak.

Split or doubled peaks

Looks like: one peak with a notch at the top, or two peaks right next to each other where one was expected.

Two very different explanations:

1. Two real species. For example a peptide and its deamidated or isomerized form, which are chemically very close. That is real chemistry and should be counted as separate peaks.
2. An artifact. A void or channel at the head of the column, a partly blocked frit, or a mismatch between the sample solvent and the mobile phase can split one compound into two peaks.

Telling them apart: an artifact usually affects every peak in the run, including standards. A real second species shows up only for that sample, and LC-MS will often give it a slightly different mass.

Shoulders

Looks like: a bump on the side of the main peak, not fully separated.

Usually means: a closely related impurity that the method only partly resolves.

Effect on purity: depends on integration. Drawn as part of the main peak, it inflates purity. Dropped as a separate peak, it lowers it. The report's peak table should show which.

Overlapping peaks and resolution

When two peaks run into each other, the measure of how well they are separated is resolution, Rs. A resolution of about 1.5 or more is generally taken as baseline separation, meaning the signal returns to baseline between the two peaks. Below that, area is shared between them, and where the software splits it is a judgment call.

What to look for on a report

• Is the main peak reasonably symmetrical?
• Are there shoulders, and how were they integrated?
• Does the peak table list peaks close to the main one, or only distant ones?
• Is the method's system suitability (tailing, resolution) stated?

None of these shapes is a failure by itself. They are reasons to read the purity figure alongside the chart, not on its own.