HPLC and how to read a chromatogram

Most purity figures on peptide certificates come from one technique: high-performance liquid chromatography, or HPLC, with an ultraviolet (UV) detector. The figure is only as useful as your understanding of what the instrument saw. This page explains the method and how to read its main output, the chromatogram.

What HPLC does

A pump pushes a liquid, called the mobile phase, through a narrow steel column packed with very small particles, called the stationary phase. A small, measured portion of the sample is loaded into the flowing liquid at the column inlet. Each component of the sample is held back by the particles to a different degree, so the components leave the column at different times. A detector at the outlet records them as they pass.

For peptides the usual mode is reversed-phase HPLC. The particles carry a water-repelling coating, commonly C18 chains (18-carbon chains bonded to silica). The mobile phase starts as mostly water with a little acid, often trifluoroacetic acid (TFA), and the share of an organic solvent, usually acetonitrile, rises during the run. That rising share is called the gradient. Peptides leave the column roughly in order of increasing overall hydrophobicity: the more water-repelling a molecule is, the later it comes off.

The detector measures how much UV light the liquid leaving the column absorbs. Peptide bonds absorb in the far ultraviolet, below about 230 nm, so peptides are usually detected between 210 and 220 nm. The certificate should state the wavelength, for example 214 nm.

Reading the chromatogram

The chromatogram plots the detector’s signal against time.

• The horizontal axis is time, in minutes, from the moment the sample was loaded.
• The vertical axis is the detector response: how much light was absorbed at the chosen wavelength.
• The baseline is the flat line recorded while nothing that absorbs is passing the detector.
• Each peak is something that absorbs at that wavelength leaving the column.

The time at the top of a peak is its retention time. Under the same method (same column, mobile phase, gradient, flow rate and temperature) the same substance comes off at the same time, so retention time is reproducible. It is not proof of identity, though. Different substances can share a retention time, and a retention time only means something when compared with a reference run under identical conditions. Identity is a job for mass spectrometry; see Mass spectrometry: identity, not purity.

Area percent

The software draws a baseline under each peak and measures the area between the peak and that baseline. Area-percent purity is the area of the main peak divided by the total area of all the integrated peaks, times 100. If the main peak has an area of 991 units and all the others together add up to 9, the purity is 991 ÷ 1,000, or 99.1%.

Peaks that also appear in a blank run, and the disturbance at the very start of the run, are normally left out of the calculation. A good report says which regions were integrated and which were excluded.

Why area percent is not mass percent

Area percent is a ratio of detector signals, not of masses. It can differ from the true share by mass for four reasons.

1. Different substances absorb differently. Each has its own signal per unit of mass at a given wavelength, its response factor. Area percent counts every peak as if it had the same response factor as the main peak. An impurity that absorbs more strongly looks bigger than it is; one that absorbs weakly looks smaller.
2. Some things are invisible. Anything that absorbs little or no light at the detection wavelength, such as water or inorganic salts, leaves no peak, so it cannot lower the purity figure. This is how a powder can read 99% by HPLC and still be only 70 to 90 percent peptide by mass, with the rest counterion and water; see Purity, net content, net peptide content and salt forms.
3. Some things are not separated. An impurity that stays on the column, leaves after the run ends, or comes off at the same time as the main peak is not counted on its own. A shoulder, or a peak that is not symmetrical, can be a sign that two substances overlap.
4. Integration involves choices. Where the baseline is drawn, how small a peak is still counted, and which regions are excluded all move the final number. Two labs can integrate the same trace differently.

So HPLC area percent is a good measure of how much of the UV-visible, separated material is the main peak, under that method. It is not a measure of how much of the powder is peptide.

A clean trace and an impure one

The example traces show the difference. The clean trace has one tall, narrow, symmetrical main peak on a flat baseline. Its two small peaks, 0.5% and 0.4% of the total area, are almost invisible at full scale and appear only when the vertical axis is magnified. That is normal, and it is why the peak table matters more than a glance at the picture.

The impure trace has a shorter main peak (86.2%), a shoulder on its front edge, and separate peaks before and after it. In reversed-phase HPLC, peaks before the main one are more polar than it, and peaks after it are more hydrophobic. Closely related substances, such as chains missing one amino acid, often come off near the main peak, which is why they are hard to separate; How peptides are made explains where they come from. HPLC alone cannot say what each peak is.

What to look for in a lab report

• The chromatogram itself, not only the number, with both axes labeled.
• The method: column type, mobile phase, gradient, wavelength.
• A peak table listing each peak’s retention time and area percent.
• Which regions were integrated, and which were excluded.
• A main-peak retention time that agrees with a reference run, where one was made.

An HPLC purity figure answers one question well. It does not tell you the identity of the main peak, how much peptide the powder holds, or anything about water, salts, solvents or microbes. Those need other tests, and a certificate either reports them or it does not.