Reading the Peaks: A Beginner’s Guide to HPLC

A chromatogram turns a separation into a picture. Understanding the journey behind its peaks reveals what the picture can—and cannot—say about a sample.

An imaginary laboratory report shows three peaks. One dominates the page, labeled “90% area.” That percentage describes a share of recorded detector response; interpreting it as a share of material requires additional assumptions.^[4] The interesting part of the report is the chain of events that produced the number. Following that chain makes the trace much easier to read.

A journey becomes a line

High-performance liquid chromatography, or HPLC, separates components of a liquid sample before they reach a detector. A flowing liquid, called the mobile phase, carries the sample through a column. Inside, components interact differently with the stationary phase. Those interactions can make some components travel more slowly than others, allowing them to emerge at different times.^[1]

The chromatogram records detector signal vertically and elapsed time horizontally. A peak appears as a band of material passes the detector. Its retention time runs from the start of sample introduction to the peak’s highest point, or apex. Height measures the distance above the baseline; area includes the space between the peak and that baseline.^[1] These features provide different pieces of the story. A retention time helps locate an event; an area describes how much signal the event contributed.

The detector has a particular view

Separation and detection are distinct jobs. An ultraviolet detector observes light absorption, so its usefulness depends on the compounds being measured. Shimadzu’s detector guide explains why detector selection follows the properties of the analyte—the substance of interest—and the analytical conditions. A direct UV measurement needs a detectable response from the target under the chosen analytical conditions.^[2]

That makes a chromatogram a view through a particular window, rather than a complete inventory. An apparently empty stretch of a UV trace means that this measurement did not produce a visible UV feature there. It does not, by itself, establish the absence of every possible substance.^[2] The detector label belongs alongside the graph for the same reason that a map needs a legend: it explains what the marks represent.

The baseline is part of the measurement

The neat area printed beside a peak is calculated using boundaries. Integration is the process of assigning the region to be measured, including where the peak starts and ends and how its baseline runs underneath it. Agilent’s OpenLab documentation distinguishes several ways of drawing that baseline, including valley boundaries, forced horizontal baselines and manual integration.^[3]

Consider the outline of a hill drawn on graph paper. Measuring its area requires deciding where the ground lies. Moving that ground line changes the enclosed area even though the hill’s outline stays put. This is a geometric illustration, not a laboratory result. It explains why a chromatogram and its integration settings should be considered together. Agilent’s baseline codes make some of those processing choices visible in the results, including flags for manually integrated peaks.^[3]

A clean peak can still conceal a mixture

Two components can emerge together: a problem called coelution. A diode-array detector, often abbreviated DAD, adds spectral information to the time trace. Agilent’s UV impurity check compares spectra across a peak with the spectrum at its apex. Differences can reveal overlapping components with sufficiently different UV spectra.^[5]

The qualification matters. A screen based on spectral differences cannot establish the absence of a component that contributes too little distinguishable spectral information. Likewise, a strong match to a reference spectrum is supporting evidence for identity, rather than a complete molecular identification.^[5] A smooth outline and a favorable spectral comparison are useful observations, but they address particular questions.

ICH Q2(R2) asks whether a procedure distinguishes the analyte from relevant interference, recommending an additional procedure when needed.^[6] The goal is to test an explanation for the peak, rather than simply admire its shape.

Three peaks on an imaginary report

These invented, unitless areas return to the opening puzzle; they are not Chudik Scientific measurements. Agilent defines area percentage as each peak’s area divided by the summed peak area, without applying response factors.^[4]

Hypothetical feature	Invented area	Share of the total	Interpretation
Early peak A	50	5%	A smaller share of the included signal
Middle peak B	900	90%	The largest included signal contribution
Late peak C	50	5%	The same area as A, without an assumed identity or mass
Total	1,000	100%	The denominator for these area percentages

The arithmetic is straightforward: 900 divided by 1,000 gives 90%. The chemical interpretation is conditional. Agilent explains that area percentages approximate relative component amounts when detector responses are sufficiently alike.^[4] The table therefore does not assign milligrams to B. Nor does it name A or C. Leaving those cells unnamed is part of reading the evidence correctly, rather than a failure to extract enough information from the picture.

Identity and quantity need their own evidence

Mass spectrometry supplies another kind of information. David Bunk and Mark Lowenthal’s NIST chapter describes molecular-mass measurements, peptide mapping and tandem mass spectrometry sequencing as qualitative evidence about proteins and peptides.^[7] These investigate identity differently from the UV signal that drew the chromatographic peak.

Quantity requires relating response to known concentration or content over a suitable range, with accuracy and precision evaluated for the procedure.^[6] A calibrated amount and an uncorrected area percentage answer different questions, even when a report places both beside the same graph.

The most revealing way to read a chromatogram is to move from the picture back toward its construction: separation, detector, integration, and then the evidence for the stated interpretation. The opening 90% is still useful. Its value becomes clearer once it is understood as a defined observation, with the additional questions kept visible.

Research and attribution note: This article synthesizes manufacturer documentation, ICH guidance and the official abstract of Bunk and Lowenthal’s NIST chapter. The table and geometric analogy are original teaching illustrations. No Chudik Scientific experiment is reported.

For the broader measurement discussion, read Inside the Vial.